vvEPA
Interagency Review Draft
EPA/635/R-21/286b
www.epa.gov/iris
Toxicological Review of Formaldehyde—Inhalation
Supplemental Information
[CASRN 50-00-0]
December 2021
Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
Supplemental Information for Formaldehyde—Inhalation
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of predissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and should
not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement of recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
CONTENTS
DISCLAIMER ii
CONTENTS iii
TABLES vi
APPENDIX A. INFORMATION IN SUPPORT OF HAZARD IDENTIFICATION A-l
A.l. Chemical Properties and Human Exposure A-l
A.l.l. Chemical Properties A-l
A.l.2. Human Exposure A-4
A.2. Toxicokinetics of Inhaled and Endogenous Formaldehyde A-15
A.2.1. Toxicokinetics of Inhaled Formaldehyde at the Portal of Entry (POE) A-16
A.2.2. Spatial Distribution of Tissue Uptake of Formaldehyde at the Portal of Entry A-17
A.2.3. Tissue Penetration of Formaldehyde Within the Upper Respiratory Tract A-22
A.2.4. Modifying Factors and Specific Uncertainties Regarding the Toxicokinetics of
Inhaled Formaldehyde Within the POE A-38
A.2.5. Conclusions Regarding the Toxicokinetics of Inhaled Formaldehyde Within the
POE A-4 5
A.2.6. Toxicokinetics of inhaled formaldehyde beyond the portal of entry A-46
A.2.7. Levels of Endogenous and Inhaled Formaldehyde in Blood and Distal Tissues A-46
A.2.8. Conjugation, Metabolism, and Speciation of Formaldehyde Outside the POE A-54
A.2.9. Elimination Pathways of Exogenous and Endogenous Formaldehyde A-54
A.2.10.Conclusions Regarding the Toxicokinetics of Inhaled Formaldehyde Outside of the
POE A-58
A.2.11.Toxicokinetics Summary A-58
A.2.12. Modeling Formaldehyde Flux to Respiratory Tract Tissue A-60
A.3. Reflex Bradypnea A-78
A.4. Genotoxicity A-85
A.4.1. Genotoxicity of Formaldehyde in Cell-Free Systems A-86
A.4.2. Genotoxicity of Formaldehyde in Prokaryotic Organisms A-88
A.4.3. Genotoxicity of Formaldehyde in Nonmammalian Systems A-94
A.4.4. Genotoxicity of Formaldehyde in in Vitro Mammalian Cells A-97
A.4.5. Genotoxicity of Formaldehyde in Experimental Animals A-118
A.4.6. Genotoxic Endpoints in Humans A-134
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
A.4.7. Supporting Material for Genotoxicity A-185
A.5. Support for Hazard Assessments of Specific Health Effects A-230
A.5.1. General Approaches to Identifying and Evaluating Individual Studies A-230
A.5.2. Sensory Irritation A-262
A.5.3. Pulmonary Function A-300
A.5.4. Immune-Mediated Conditions, Including Allergies and Asthma A-340
A.5.5. Respiratory Tract Pathology A-393
A.5.6. Mechanistic Evidence Related to Potential Noncancer Respiratory Health Effects A-433
A.5.7. Nervous System Effects A-581
A.5.8. Developmental and Reproductive Toxicity A-621
A.5.9. Carcinogenicity: Respiratory Tract, Lymphohematopoietic, or Other Cancers A-657
APPENDIX B. INFORMATION IN SUPPORT OF THE DERIVATION OF REFERENCE VALUES AND
CANCER RISK ESTIMATES B-l
B.l. Dose-Response Analyses for Noncancer Health Effects B-l
B.l.l. Evaluation of Model Fit Using BMDS models B-l
B.l.2. Noncancer Estimates from Observational Epidemiology Studies B-2
B.1.3. Noncancer Estimates from Animal Toxicology Studies B-17
B.2. Dose-Response Analysis for Cancer B-30
B.2.1. Cancer Estimates from Observational Epidemiology Studies B-30
B.2.2. Cancer Estimates from Animal Toxicology Studies Using Biologically Based Dose
Response (BBDR) Modeling B-33
B.2.3. Estimates of Cancer Risk Using DNA Adduct Data from Animal Toxicology Studies
and Background Incidence B-94
APPENDIX C. ASSESSMENTS BY OTHER NATIONAL AND INTERNATIONAL HEALTH AGENCIES C-l
APPENDIX D. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS AND EPA'S
DISPOSITION D-l
D.l.NRC FORMALDEHYDE PANEL SUMMARY RECOMMENDATIONS SPECIFIC TO
FORMALDEHYDE AND EPA RESPONSES D-l
D.2.RESPONSE TO PUBLIC COMMENTS D-41
APPENDIX E. SUMMARY OF PUBLIC COMMENTS AND EPA'S DISPOSITION E-l
E.l. Insert Appendix E here E-l
APPENDIX F. SYSTEMATIC EVIDENCE MAP UPDATING THE LITERATURE FROM 2016-2021 F-l
F.l. INTRODUCTION F-l
F.2. METHODS F-l
F.2.1. Specific Aims F-l
This document is a draft for review purposes only and does not constitute Agency policy.
iv DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F.2.2. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria and
Supplemental Material Tagging F-2
F.2.3. Literature Search and Screening Strategies F-3
F.2.4. Literature Inventory F-4
F.3. RESULTS F-7
F.3.1. Sensory Irritation Effects in Human Studies F-7
F.3.2. Pulmonary Function Effects in Human Studies F-9
F.3.3. Immune-Mediated Conditions in Humans, Focusing on Allergies and Asthma F-ll
F.3.4. Respiratory Tract Pathology in Human Studies F-14
F.3.5. Animal Studies of Respiratory Tract Pathology F-16
F.3.6. Site-specific Cancer in Human Studies F-19
F.3.7. Animal Studies of Respiratory Tract Cancer F-21
F.3.8. Animal Studies of Lymphohematopoietic Cancers F-23
F.3.9. Mechanistic Studies of Inflammation and Immune-Related Responses F-25
F.3.10. Mechanistic Studies of Respiratory Tract Cancer, Focusing on Genotoxicity F-35
F.3.11. Mechanistic Studies of Lymphohematopoietic Cancer, Focusing on Genotoxicity F-41
F.3.12. Nervous System Effects F-46
F.3.13. Reproductive and Developmental Effects F-49
APPENDIX G. QUALITY ASSURANCE FOR THE IRIS TOXICOLOGICAL REVIEW OF FORMALDEHYDE G-l
REFERENCES R-l
v DRAFT—DO NOT CITE OR QUOTE
This document is a draft for review purposes only and does not constitute Agency policy.
-------
Supplemental Information for Formaldehyde—Inhalation
TABLES
Table A-l. Physicochemical properties of formaldehyde A-2
Table A-2. Ambient air levels by land use category based on 2018 annual site averages A-6
Table A-3. Formaldehyde emission rates from various consumer products A-9
Table A-4. Studies on residential indoor air levels of formaldehyde A-10
Table A-5. Dosimetry and response of formaldehyde in experimental animals by indirect
measurements A-18
Table A-6. Comparison of formaldehyde uptake at the portal of entry with single or repeated
inhalation exposure A-21
Table A-7. ADH3 kinetics in human and rat tissue samples and cultured cells A-26
Table A-8. Levels of folate intermediates, activity of folate-dependent enzymes, and the rate of
oxidation of formate in the liver of various species A-28
Table A-9. Summary of endogenous and exogenous DNA-protein crosslinks in nasal tissues of
rats following inhalation exposure of 13CD2-labeled formaldehyde A-31
Table A-10. Summary of endogenous and exogenous DNA monoadducts in nasal tissue of
monkeys and rats following inhalation exposure of 13CD2-labeled formaldehyde A-34
Table A-ll. Summary of blood and tissue levels of total3 formaldehyde in humans and
experimental animals following inhalation exposure to formaldehyde A-48
Table A-12. Summary of endogenous and exogenous DNA monoadducts in distal tissues of
monkeys and rats following inhalation exposure of 13CD2-labeled formaldehyde A-51
Table A-13. Summary of endogenous and exogenous DNA-protein crosslinks in distal tissues of
monkeys and rats following inhalation exposure of 13CD2-labeled formaldehyde A-54
Table A-14. Summary of excretion study following exposure to formaldehyde by inhalation in
rats A-5 6
Table A-15. Measured levels of formaldehyde, methanol and ethanol in room air and exhaled
breath A-56
Table A-16. Formaldehyde respiratory depression (RD) values for several mouse strains and
exposure durations A-82
Table A-17. Formaldehyde respiratory depression (RD) values for several rat strains and
exposure durations A-82
Table A-18. Summary of genotoxicity of formaldehyde in cell-free systems A-87
Table A-19. Summary of genotoxicity of formaldehyde in prokaryotic systems A-89
Table A-20. Summary of genotoxicity studies for formaldehyde in nonmammalian organisms A-94
Table A-21. Summary of in vitro genotoxicity studies of formaldehyde in mammalian cells A-105
Table A-22. Summary of in vivo genotoxicity studies of formaldehyde inhalation exposure in
experimental animals A-125
Table A-23. Summary of in vivo genotoxicity studies of formaldehyde exposure by
intraperitoneal and oral routes of exposure in experimental animals A-130
Table A-24. Summary of genotoxicity of formaldehyde in human studies A-141
Table A-25. Summary of search terms for cancer mechanisms A-185
Table A-26. Evaluation of genotoxicity endpoints in epidemiology studies of formaldehyde
exposure A-188
Table A-27. Genotoxicity summary table A-224
Table A-28. Approach to evaluating observational epidemiology studies for hazard identification ...A-231
Table A-29. Approach to evaluating experimental animal studies for hazard identification A-233
Table A-30. Inhalation exposure quality: formaldehyde (Note: exposure deficiencies are shaded) ...A-240
This document is a draft for review purposes only and does not constitute Agency policy.
vi DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-31. Summary of search terms for sensory irritation A-262
Table A-32. Inclusion and exclusion criteria for studies of sensory irritation A-262
Table A-33. Criteria for categorizing study confidence in epidemiology studies of sensory
irritation A-265
Table A-34. Evaluation of studies examining sensory irritation in humans: residential studies A-267
Table A-35. Evaluations of studies examining sensory irritation in humans: school-based studies ....A-272
Table A-36. Evaluations of studies examining sensory irritation in humans: controlled human
exposure studies A-272
Table A-37. Evaluation of studies examining sensory irritation in humans: anatomy courses A-275
Table A-38. Evaluations of studies examining sensory irritation in humans: occupational studies A-282
Table A-39. Summary of epidemiology studies of laboratory exposures to formaldehyde and
human sensory irritation A-289
Table A-40. Summary of epidemiology studies of occupational exposures to formaldehyde and
human sensory irritation A-294
Table A-41. Summary of search terms for pulmonary function A-301
Table A-42. Inclusion and exclusion criteria for studies of pulmonary function A-301
Table A-43. Criteria for categorizing study confidence in epidemiology studies of pulmonary
function A-303
Table A-44. Evaluation of formaldehyde - pulmonary function epidemiology studies A-305
Table A-45. Formaldehyde effects on pulmonary function in controlled human exposure studies ....A-331
Table A-46. Study details for references depicted in Figures A-26 - A-28 A-339
Table A-47. Summary of search terms - allergy-related conditions, including asthma A-344
Table A-48. Inclusion and exclusion criteria for studies of allergy and asthma studies in humans A-344
Table A-49. Inclusion and exclusion criteria for studies of hypersensitivity in animals A-345
Table A-50. Criteria used to assess epidemiologic studies of respiratory and immune-mediated
conditions, including allergies and asthma, for hazard assessment A-354
Table A-51. Evaluation of allergy and asthma studies A-356
Table A-52. Evaluation of controlled acute exposure studies among people with asthma A-389
Table A-53. Summary of search terms for respiratory tract pathology in humans A-393
Table A-54. Inclusion and exclusion criteria for studies of repiratory pathology in humans A-394
Table A-55. Summary of search terms for respiratory tract pathology in animals A-396
Table A-56. Inclusion and exclusion criteria for studies of repiratory pathology in animals A-397
Table A-57. Criteria for categorizing study confidence in epidemiology studies of respiratory
pathology A-399
Table A-58. Respiratory pathology A-400
Table A-59. Evaluation of controlled inhalation exposure studies examining respiratory
pathology in animals A-409
Table A-60. Evaluation of controlled inhalation exposure studies examining cell proliferation
and mucociliary function in animals A-423
Table A-61. Supportive short-term respiratory pathology studies in animals A-428
Table A-62. Summary of supplemental literature search terms for mechanistic studies relevant
to potential noncancer respiratory health effects A-435
Table A-63. Inclusion and exclusion criteria for mechanistic studies relevant to potential
noncancer respiratory health effects A-437
Table A-64. Criteria and presentation of strength of the evidence for each mechanistic event
and for potential associations between events relating to potential respiratory
health effects A-441
This document is a draft for review purposes only and does not constitute Agency policy.
vii DRAFT—DO NOT CITE OR QUOTE
This document is a draft for review purposes only and does not constitute Agency policy.
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-65. Decision criteria for the evaluation of mechanistic studies relevant to potential
noncancer respiratory effects A-445
Table A-66. URT-specific structural modification, sensory nerve-related changes, or immune and
inflammation-related changes A-446
Table A-67. LRT (e.g., lung, trachea, BAL) markers of structural modification, immune response,
inflammation, or oxidative stress A-454
Table A-68. Changes in pulmonary function involving provocation (e.g., bronchoconstrictors;
allergens; etc.) A-469
Table A-69. Serum (primarily) antibody responses A-473
Table A-70. Serum markers of immune response (other than antibodies), inflammation, or
oxidative stress A-479
Table A-71. Effects on other immune system-related tissues (e.g., bone marrow, spleen,
thymus, lymph nodes, etc.) A-489
Table A-72. Effects on other tissues (data extracted for possible future consideration, but not
included in the current analyses) A-494
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from
formaldehyde exposure A-507
Table A-74. Mucociliary function studies in experimental animals A-515
Table A-75. Mucociliary function studies in humans A-518
Table A-76. Subchronic or chronic exposure cell proliferation studies in experimental animals A-524
Table A-77. Short-term exposure cell proliferation studies in experimental animals A-528
Table A-78. Summary of changes in the lower respiratory tract (LRT) as a result of formaldehyde
exposure A-544
Table A-79. Summary of changes in LRT cell counts and immune factors as a result of
formaldehyde exposure A-549
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde
exposure A-556
Table A-81. Summary of changes in blood cell counts and immune factors as a result of
formaldehyde exposure A-562
Table A-82. Summary of search terms for neurological effects A-582
Table A-83. Inclusion and exclusion criteria for studies of nervous system effects A-583
Table A-84. Evaluation of observational epidemiology studies of formaldehyde—neurological
effects A-587
Table A-85. Evaluation of human controlled exposure studies of formaldehyde - nervous system
effects A-595
Table A-86. Evaluation of controlled inhalation exposure studies examining nervous system in
animals A-598
Table A-87. Evaluation of studies pertaining to mechanistic events associated with nervous
system effects A-614
Table A-88. Summary of search terms for developmental or reproductive toxicity A-622
Table A-89. Inclusion and exclusion criteria for studies of reproductive and developmental
effects in humans A-624
Table A-90. Inclusion and exclusion criteria for studies of reproductive and developmental
effects in animals A-624
Table A-91. Criteria for categorizing study confidence in epidemiology studies of reproductive
and developmental effects A-630
Table A-92. Evaluation of observational epidemiology studies of formaldehyde - reproductive
and developmental outcomes A-631
This document is a draft for review purposes only and does not constitute Agency policy.
viii DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-93. Study quality evaluation of developmental and reproductive toxicity animal studies A-649
Table A-94. Summary of search terms for carcinogenicity in humans A-658
Table A-95. Inclusion and exclusion criteria for evaluation of studies of cancer in humans A-659
Table A-96. Summary of search terms for respiratory tract cancers in animals A-661
Table A-97. Inclusion and exclusion criteria for studies of nasal cancers in animals A-662
Table A-98. Summary of search terms for lymphohematopoietic cancers in animals A-664
Table A-99. Inclusion and exclusion criteria for studies of LHP cancers in animals A-665
Table A-100. Lymphohematopoietic and upper respiratory cancers: age-Adjusted SEER
incidence and U.S. death rates and 5-year relative survival by primary cancer
site A-669
Table A-101. Categorization of exposure assessmentmethods by study design A-670
Table A-102. Outcome-specific effect estimates classified with High confidence A-676
Table A-103. Outcome-specific effect estimates classified with Medium confidence A-676
Table A-104. Outcome-specific effect estimates classified as uninformative A-679
Table A-105. Evaluation of occupational cohort studies of formaldehyde and cancers of the URT
(NPC, SN, OHPC) and LHP (HL, MM, LL, ML) A-680
Table A-106. Evaluation of case-control studies of formaldehyde and cancers of the URT (NPC,
SN, OHPC) and LHP (HL, MM, LL, ML) A-711
Table A-107. Evaluation of controlled inhalation exposure studies examining respiratory tract
cancer or dysplasia in animals A-741
Table A-108. Evaluation of controlled inhalation exposure studies examining
lymphohematopoietic cancers in animals A-749
Table B-l. Concentration-response information for the central estimate of the effect extracted
from Hanrahan et al. (1984) B-4
Table B-2. Concentration-response information for the upper bound on the central estimate of
the effect extracted from Hanrahan et al. (1984) B-5
Table B-3. Benchmark dose modeling of sensory irritation using a BMR of 10% B-8
Table B-4. Parameter estimates for log-logistic model with BMC of 10% extra risk over an
assumed background of 3% and lower confidence limit for the BMCL for
prevalence of conjunctival redness and/or nose or throat dryness; data from
Andersen and Molhave (1983) B-10
Table B-5. Observed and estimated values and scaled residuals for log-logistic model with BMC
of 10% extra risk over an assumed background of 3% and lower confidence limit
for the BMCL for prevalence of conjunctival redness and/or nose or throat
dryness; data from Andersen and Molhave (1983) B-10
Table B-6. Parameter estimates for probit model with BMC of 10% extra risk and 95% lower
confidence limit for the BMCL for prevalence of eye irritation; data from Kulie et
al. (1987) B-ll
Table B-7. Observed and estimated values and scaled residuals for probit model with BMC of
10% extra risk and 95% lower confidence limit for the BMCL for prevalence of
eye irritation; data from Ku lie et al. (1987) B-ll
Table B-8. Modeled effect estimates for night-time symptoms of an asthma attack; data from
Venn et al., 2003 B-15
Table B-9. Benchmark dose modeling of rat respiratory histopathological effects B-19
Table B-10. Endpoints selected for dose-response modeling for reproductive and
developmental toxicity in animals B-24
This document is a draft for review purposes only and does not constitute Agency policy.
ix DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-ll. Summary of BMD modeling results for serum testosterone in male Wistar rats
exposed to formaldehyde by inhalation for 13 weeks (Ozen et al., 2005); BMR =
1 SD change from the control mean B-25
Table B-12. Summary of BMD modeling results for relative testis weight in male Wistar rats
exposed to formaldehyde by inhalation for 4 weeks (Ozen et al., 2002); BMR = 1-
SD change from the control mean B-26
Table B-13. Model predictions for relative testis weight (Ozen et al. 2002) B-28
Table B-14. Parameter estimates B-29
Table B-15. Table of data and estimated values of interest B-29
Table B-16. Likelihoods of interest B-29
Table B-17. Tests of interest B-29
Table B-18. Extra risk calculation for environmental exposure to 0.0550 ppm formaldehyde (the
LECooos for NPC incidence) using a log-linear exposure-response model based on
the cumulative exposure trend results of Beane Freeman et al. (2013), as
described in Section 2.2.1 B-31
Table B-19. Evaluation of assumptions and uncertainties in the CUT model for nasal tumors in
the F344 rat B-38
Table B-20. PBPK models for formaldehyde-DPX B-43
Table B-21. Parameter estimates for PBPK modeling B-47
Table B-22. Influence of control data in modeling formaldehyde-induced cancer in the F344 rat B-49
Table B-23. Variation in number of cells across nasal sites in the F344 rat B-57
Table B-24. Parameter specifications and estimates for clonal growth models of nasal SCC in the
F344 rat using alternative characterization of cell replication and death rates B-75
Table B-25. Parameter specifications and estimates for clonal growth models of nasal SCC in the
F344 rat using cell replication and death rates as characterized in Conolly et al.
(2003) B-76
Table B-26. Comparison of statistical confidence bounds on added risk for two models B-78
Table B-27. Summary of evaluation of major assumptions and results in CUT human BBDR model B-80
Table B-28. Extrapolation of parameters for enzymatic metabolism to the human in Conolly et
al. (2000) B-82
Table B-29. Mean formaldehyde-induced N2-hmdG adducts in rats and monkeys (Swenberg et
al., 2011) B-96
Table B-30. A/2-hmdG adduct levels (Lu et al., 2011) and rat tumor data (Monticello et al., 1996;
Subramaniam et al., 2007) B-102
Table C-l. Hazard conclusions and toxicity values developed by other national and international
health agencies C-l
Table F-l. Example of outcome-specific PECO: LHP cancer in animals F-2
Table F-2. Literature search strategy F-3
Table F-3. Studies of sensory irritation effects in humans F-8
Table F-4. Studies of pulmonary function effects in humans F-10
Table F-5. Studies of immune-mediated conditions in humans, focusing on allergies and asthma F-12
Table F-6. Studies of respiratory tract pathology in humans F-15
Table F-7. Animal studies of respiratory tract pathology F-17
Table F-8. Studies of site-specific cancer in humans F-20
Table F-9. Animal studies of respiratory tract cancers F-22
Table F-10. Animal studies of lymphohematopoietic cancer F-24
Table F-ll. Mechanistic studies relating to respiratory or systemic inflammatory and immune
responses F-26
This document is a draft for review purposes only and does not constitute Agency policy.
x DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-12. Mechanistic studies relating to respiratory tract cancers, focusing on genotoxicity F-36
Table F-13. Mechanistic studies relating to lymphohematopoietic cancers, focusing on
genotoxicity F-42
Table F-14. Studies of nervous system effects F-47
Table F-15. Studies of reproductive and developmental effects F-50
FIGURES
Figure A-l. Chemical structure of formaldehyde A-l
Figure A-2. Formaldehyde Ambient Concentrations Contribution by Sector A-6
Figure A-3. Range of formaldehyde air concentrations (ppb) in different environments A-13
Figure A-4. Schematic of the rat upper respiratory tract depicting the gradient of formaldehyde
concentration formed following inhalation exposure, both from anterior to
posterior locations, as well as across the tissue depth A-17
Figure A-5. Metabolism of formaldehyde A-25
Figure A-6. Compartmentalization of mammalian one-carbon metabolism A-28
Figure A-7. Metabolic incorporation and covalent binding of formaldehyde in rat respiratory
tract. 3H/14C ratios in macromolecular extracts from rat respiratory mucosa (A)
and olfactory mucosa (B) following 6-hour exposure to 14C- and 3H-labeled
formaldehyde (0.3, 2, 6, 10, and 15 ppm, corresponding to 0.37, 2.46, 7.38, 12.3,
18.42 mg/m3, respectively) A-32
Figure A-8. Endogenous and dietary sources of formaldehyde production A-42
Figure A-9. 3H/14C ratios in macromolecular extracts from rat bone marrow following 6-hour
exposure to 14C- and 3H-labeled formaldehyde (0.3, 2, 6, 10, and 15 ppm,
corresponding to 0.37, 2.46, 7.38, 12.3, 18.42 mg/m3, respectively) A-51
Figure A-10. Reconstructed nasal passages of F344 rat, rhesus monkey, and human A-62
Figure A-ll. Illustration of interspecies differences in airflow and verification of CFD simulations
with water-dye studies A-63
Figure A-12. Lateral view of nasal wall mass flux of inhaled formaldehyde simulated in the F344
rat, rhesus monkey, and human A-64
Figure A-13. Lateral view of nasal wall mass flux of inhaled formaldehyde simulated at various
inspiratory flow rates in a human model A-65
Figure A-14. Single-path model simulations of surface flux per ppm of formaldehyde exposure
concentration in an adult male human A-69
Figure A-15. Pressure drop versus volumetric airflow rate predicted by the CUT CFD model
compared with pressure drop measurements made in two hollow molds (CI and
C2) of the rat nasal passage (Cheng et al., 1990) or in rats in vivo (Gerde et al.,
1991) A-71
Figure A-16. Formaldehyde-DPX dosimetry in the F344 rat A-73
Figure A-17. Flux of highly reactive gas across nasal lining as a function of normalized distance
from nostril for 5 adults and 2 children A-75
Figure A-18. Left panel: Concentration-related hypothermia in mice exposed to an isocyanate
for 360 minutes A-78
Figure A-19. An oscillograph that compares the respiratory cycle for mice exposed to an URT
irritant (lower tracing) to an air control group (upper tracing) A-79
This document is a draft for review purposes only and does not constitute Agency policy.
xi DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Figure A-20. Formaldehyde effects on minute volume in naive and formaldehyde-pretreated
male B6C3F1 mice and F344 rats A-80
Figure A-21. This graph demonstrates the impact of RB on fetal development A-85
Figure A-22. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and sensory irritation in humans A-264
Figure A-23. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and pulmonary function in humans A-302
Figure A-24. Plots of change in FEF at 25-75% of FVC across a work shift or anatomy lab session
by study with study details A-334
Figure A-25. Plots of change in FEV1 across a work shift or anatomy lab session by study with
study details A-337
Figure A-26. Plots of change in FVC across a work shift or anatomy lab session by study with
study details A-339
Figure A-27. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and respiratory and immune-mediated
conditions A-346
Figure A-28. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and respiratory tract pathology in humans
(reflects studies identified in searches conducted through September 2016) A-395
Figure A-29. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and respiratory tract pathology in animals
(reflects studies identified in searches conducted through September 2016) A-398
Figure A-30. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and mechanistic data associated with
potential noncancer effects on the respiratory system (reflects studies identified
in searches conducted through September 2016; see Appendix F for literature
identification from 2016-2021) A-439
Figure A-31. Mechanistic events for respiratory effects of formaldehyde based on robust or
moderate evidence A-500
Figure A-32. Mechanistic events for respiratory effects of formaldehyde based on robust,
moderate, or slight evidence A-501
Figure A-33. Nasal cell proliferation in rats exposed to formaldehyde A-523
Figure A-34. Possible sequences of mechanistic events identified based on the most reliable
evidence available A-568
Figure A-35. Literature search documentation for sources of primary data pertaining to
formaldehyde exposure and nervous system effects (reflects studies identified
in searches conducted through September 2016) A-584
Figure A-36. Literature search documentation for sources of primary data pertaining to
formaldehyde exposure and developmental and reproductive toxicity A-626
Figure A-37. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and upper respiratory or
lymphohematopoietic cancers in humans through 2016 (see Appendix F for
details on the systematic evidence map updating the literature through 2021) A-660
Figure A-38. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and upper respiratory tract (nasal) cancers in
animals A-663
This document is a draft for review purposes only and does not constitute Agency policy.
xii DRAFT—DO NOT CITE OR QUOTE
This document is a draft for review purposes only and does not constitute Agency policy.
-------
Supplemental Information for Formaldehyde—Inhalation
Figure A-39. Literature search documentation for sources of primary data pertaining to
inhalation formaldehyde exposure and lymphohematopoietic (LHP) cancers in
animals A-666
Figure B-l. Regression of prevalence of "burning eyes" versus indoor formaldehyde
concentration (ppm) in mobile homes (approximately 1-hour air samples).
Dashed lines show upper and lower 95th percentile confidence intervals on
model results B-3
Figure B-2. Plot of the prevalence odds by residential concentration-response information from
Table 1 B-4
Figure B-3. Plot of the upper bound on prevalence odds by residential concentration-response
information from Table 2 B-6
Figure B-4. Log-logistic model with BMC of 10% extra risk over an assumed background of 3%
and lower confidence limit for the BMCL for prevalence of conjunctival redness
and/or nose or throat dryness; data from Andersen and Molhave (1983) B-9
Figure B-5. Probit model with BMC of 10% extra risk and 95% lower confidence limit for the
BMCL for prevalence of eye irritation; data from Kulie et al. (1987) B-ll
Figure B-6. Midsaggital section of rat nose showing section levels (Kerns et al., 1983) (nostril is
to the left) B-18
Figure B-7. Lateral view of contour plot of formaldehyde flux to the rat (on the top) and human
nasal lining (on the bottom) using CFD modeling (Kimbell et al., 2001) (nostril is
to the right) B-20
Figure B-8. Midsaggital section of rat nose showing section levels (Kerns et al., 1983) (nostril is
to the left) B-20
Figure B-9. Multistage (top panel) and log-logistic (bottom panel) model fit for Level 1
squamous metaplasia B-21
Figure B-10. Log-probit model fit for Level 1 squamous metaplasia B-21
Figure B-ll. Basal hyperplasia in Wistar rat (Woutersen et al., 1989): multistage model (k= 1) fit B-22
Figure B-12. Squamous metaplasia in Wistar rat (Woutersen et al., 1989): log-logistic (top panel)
and multistage (bottom panel) model fit B-23
Figure B-13. Plot of mean response (serum testosterone, Ozen et al., 2005) by concentration,
with the fitted curve for Exponential Model 2 with constant variance B-25
Figure B-14. Plot of mean response (relative testis weight, Ozen et al., 2002) by concentration,
with the fitted curve for a linear model with modeled variance. BMR = 1 SD
change from the control mean B-27
Figure B-15. Plot of mean response by concentration, with fitted curve for selected model;
concentration shown in mg/m3 B-28
Figure B-16. Dose response of normal (aN) and initiated (oti) cell division rate in Conolly et al.
(2003) B-3 6
Figure B-17. ULLI data for pulse and continuous labeling studies B-54
Figure B-18. Logarithm of normal cell replication rate aN versus formaldehyde flux (in units of
pmol/mm2-hour) for the F344 rat nasal epithelium B-57
Figure B-19. Logarithm of normal cell replication rate versus formaldehyde flux with
simultaneous confidence limits for the ALM B-58
Figure B-20. Logarithm of normal cell replication rate versus formaldehyde flux with
simultaneous confidence limits for the PLM B-59
Figure B-21. Various dose-response models of normal cell replication rate; N1 B-62
Figure B-22. Various dose-response models of normal cell replication rate; N2 B-63
Figure B-23. Various dose-response models of normal cell replication rate; N3 B-63
This document is a draft for review purposes only and does not constitute Agency policy.
xiii DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Figure B-24. Various dose-response models of normal cell replication rate; N4 B-64
Figure B-25. Various dose-response models of normal cell replication rate; N5 B-64
Figure B-26. Various dose-response models of normal cell replication rate; N6 B-65
Figure B-27. BBDR models for the rat—models with positive added risk B-72
Figure B-28. BBDR rat models resulting in negative added risk B-73
Figure B-29. Models resulting in positive added rat risk: Dose response for normal and initiated
cell replication B-74
Figure B-30. Models resulting in negative added rat risk: Dose response for normal and initiated
cell replication B-75
Figure B-31. Effect of choice of NTP bioassays for historical controls on human risk B-85
Figure B-32. Conolly et al. (2003) hockey-stick model for division rates of initiated cells in rats
and two modified models B-87
Figure B-33. Conolly et al. (2003) J-shaped model for division rates of initiated cells in rats and
two modified models B-88
Figure B-34. Very similar model estimates of probability of fatal tumor in rats for three models
in Figure F-2 B-89
Figure B-35. Cell proliferation data from Meng et al. (2010) B-91
Figure B-36. Graphs of the additional human risks estimated by applying these modified models
for at, using all NTP controls, compared to those obtained using the original
Conolly et al. (2004) model B-93
Figure B-37. Schematic of typical dose-response curves with axes shifted to include background
dose and risk B-101
Figure B-38. Endogenous A/2-hmdG adducts as a function of formaldehyde exposure B-103
Figure B-39. Endogenous A/2-hmdG adducts as a function of formaldehyde exposure
concentration B-104
Figure B-40. Total (endogenous plus exogenous) A/2-hmdG adducts as a function of
formaldehyde exposure B-104
Figure B-41. Endogenous and exogenous adduct levels (adducts per 107 dG) appear to be
correlated for data from animals exposed to 0.7 (left) and 2.0 ppm (right)
formaldehyde B-105
Figure B-42. Endogenous and exogenous adduct levels from individual animals appear to be
uncorrelated for exposures of 6 (left), 9 (right), and 15 ppm (bottom)
formaldehyde B-106
Figure B-43. Underestimation of slope of dose response using bottom up approach B-107
Figure D-l. Various assumed dose-response curves for initiated cell division rates (as function of
formaldehyde flux to tissue) D-54
Figure D-2. Logarithm of replication rate for normal cells (aN) versus formaldehyde flux (in units
of pmol/mm2/h) for the F344 rat nasal epithelium D-55
Figure D-3. Estimates of extra human risk of respiratory cancer from lifetime exposure to
formaldehyde D-56
Figure D-4. Decision tree for the use of mechanistic data and BBDR modeling [Abbreviations:
extrapol. = extrapolation; u-v = uncertainty-variability; POD=point of departure] D-57
Figure F-l. Sensory irritation literature tree (interactive version here) F-7
Figure F-2. Pulmonary function effects in humans literature tree (interactive version here) F-9
Figure F-3. Asthma and immune effects in humans literature tree (interactive version here) F-ll
Figure F-4. Human respiratory tract pathology literature tree (interactive version here) F-14
Figure F-5. Animal respiratory tract pathology literature tree (interactive version here) F-16
Figure F-6. Human cancer literature tree (interactive version here) F-19
This document is a draft for review purposes only and does not constitute Agency policy.
xiv DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Figure F-7. Animal respiratory tract cancer literature tree (interactive version here) F-21
Figure F-8. Animal lymphohematopoietic cancer literature tree (interactive version here) F-23
Figure F-9. Mechanistic inflammation and immune effects literature tree (interactive version
here) F-25
Figure F-10. Mechanistic respiratory tract cancer literature tree (interactive version here) F-35
Figure F-ll. Mechanistic lymphohematopoietic cancer literature tree (interactive version here) F-41
Figure F-12. Nervous system effects literature tree (interactive version here) F-46
Figure F-13. Reproductive and developmental effects literature tree (interactive version here) F-49
This document is a draft for review purposes only and does not constitute Agency policy.
XV DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
ABBREVIATIONS
a2u alpha 2u-globulin
ACGIH American Conference of Governmental
Industrial Hygienists
AIC Akaike's information criterion
ALD approximate lethal dosage
ALT alanine aminotransferase
AST aspartate aminotransferase
ATSDR Agency for Toxic Substances and
Disease Registry
BMD benchmark dose
BMDL benchmark dose lower confidence limit
BMDS Benchmark Dose Software
BMR benchmark response
BUN blood urea nitrogen
BW body weight
CA chromosomal aberration
CAS Chemical Abstracts Service
CASRN Chemical Abstracts Service Registry
Number
CBI covalent binding index
CHO Chinese hamster ovary (cell line cells)
CL confidence limit
CNS central nervous system
CPN chronic progressive nephropathy
CYP450 cytochrome P450
DAF dosimetric adjustment factor
DEN diethylnitrosamine
DMSO dimethylsulfoxide
DNA deoxyribonucleic acid
EPA Environmental Protection Agency
FDA Food and Drug Administration
FEVi forced expiratory volume of 1 second
GD gestation day
GDH glutamate dehydrogenase
GGT y-glutamyl transferase
GSH glutathione
GST glutathione-S-transferase
Hb/g-A animal blood:gas partition coefficient
Hb/g-H human blood:gas partition coefficient
HEC human equivalent concentration
HED human equivalent dose
i.p. intraperitoneal
IRIS Integrated Risk Information System
IVF in vitro fertilization
LCso median lethal concentration
LD50 median lethal dose
LOAEL lowest-observed-adverse-effect level
MN micronuclei
MNPCE
micronucleated polychromatic
erythrocyte
MTD
maximum tolerated dose
NAG
N-acetyl-p-D-glucosaminidase
NCEA
National Center for Environmental
Assessment
NCI
National Cancer Institute
NOAEL
no-observed-adverse-effect level
NTP
National Toxicology Program
NZW
New Zealand White (rabbit breed)
OCT
ornithine carbamoyl transferase
ORD
Office of Research and Development
PBPK
physiologically based pharmacokinetic
PCNA
proliferating cell nuclear antigen
POD
point of departure
POD[adj]
duration-adjusted POD
QSAR
quantitative structure-activity
relationship
RDS
replicative DNA synthesis
RfC
inhalation reference concentration
RfD
oral reference dose
RGDR
regional gas dose ratio
RNA
ribonucleic acid
SAR
structure activity relationship
SCE
sister chromatid exchange
SD
standard deviation
SDH
sorbitol dehydrogenase
SE
standard error
SGOT
glutamic oxaloacetic transaminase, also
known as AST
SGPT
glutamic pyruvic transaminase, also
known as ALT
SSD
systemic scleroderma
TCA
trichloroacetic acid
TWA
time-weighted average
UF
uncertainty factor
UFa
interspecies uncertainty factor
UFh
intraspecies uncertainty factor
UFs
subchronic-to-chronic uncertainty
factor
UFd
database deficiencies uncertainty factor
This document is a draft for review purposes only and does not constitute Agency policy.
xvi DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 APPENDIX A. INFORMATION IN SUPPORT OF
2 HAZARD IDENTIFICATION
3 A.l. CHEMICAL PROPERTIES AND HUMAN EXPOSURE
4 A.l.l. Chemical Properties
5 Formaldehyde (CASRN 50-00-0) is the first of the series of aliphatic aldehydes and is a gas
6 at room temperature. Its molecular structure is depicted in Figure A-l. It is noted for its reactivity
7 and versatility as a chemical intermediate. It readily undergoes polymerization, is highly
8 flammable, and can form explosive mixtures with air. It decomposes at temperatures above 150°C
9 fWHO. 20021.
H
\
c=o
/
H
Figure A-l. Chemical structure of formaldehyde.
10 At room temperature, pure formaldehyde is a colorless gas with a strong pungent,
11 suffocating, and highly irritating odor (NLM, 2015). Formaldehyde is readily soluble in water,
12 alcohols, ether, and other polar solvents fWHO. 20021. A synopsis of its physicochemical properties
13 is given in Table A-l.
14 Production, uses, and sources of formaldehyde
15 Formaldehyde has both commercial and industrial uses. Formaldehyde has been produced
16 commercially since the early 1900s and, in recentyears, has been ranked in the top 25 highest
17 volume chemicals produced in the U.S. (NTP. 20101 fATSDR. 19991. Based on EPA's Chemical Data
18 Reporting CDR) the national production volume for formaldehyde was 3.9 billion lb/yr in 2011 and
19 between 1 and 5 billion lbs/yr for the years 2012 through 2015
20 fhttps://chemview.epa.gov/chemview/#l.
This document is a draft for review purposes only and does not constitute Agency policy.
A-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Supplemental Information for Formaldehyde—Inhalation
Table A-l. Physicochemical properties of formaldehyde
Name
Formaldehyde
International Union for Pure and Applied Chemistry
name
Formaldehyde
Synonyms
Formic aldehyde
Methanal
Methyl aldehyde
Methylene oxide
Oxomethane
Oxymethylene
Chemical Abstracts Service Index name
Formaldehyde
Chemical Abstracts Service Registry Number
50-00-0
Formula
HCHO
Molecular weight
30.03
Density
Gas: 1.067 (air = 1)
Liquid: 0.815 g/mL at -20°C
Vapor pressure
3,883 mm Hg at 25°C
Log Kow
-0.75 to 0.35
Henry's law constant
3.4 x 10"7 atm-m3/mol at 25°C
2.2 x 10"2 Pa-m3/mol at 25°C
Conversion factors (25°C, 760 mm Hg)
1 ppm = 1.23 mg/m3(v/v)
1 mg/m3 = 0.81 ppm (v/v)
Boiling point
-19.5°C at 760 mm Hg
Melting point
-92°C
Flash point
60°C; 83°C, closed cup for 37%, methanol-free aqueous solution; 50°C
closed cup for 37% aqueous solution with 15% methanol
Explosive limits
73% upper; 7% lower by volume in air
Autoignition temperature
300°C
Solubility
Very soluble in water; soluble in alcohols, ether, acetone, benzene
Reactivity
Reacts with alkalis, acids and oxidizers
Sources: American Conference of Governmental Industrial Hygienists (ACGIH) (2002); World Health Organization
International Programme on Chemical Safety (WHO) (2002); (Gerberich and Seaman, 2013; ATSDR, 1999; Walker,
1975).
Approximately 55% of the consumption of formaldehyde is in the production of industrial
resins fNTP. 20101. Formaldehyde is a chemical intermediate used in the production of some
plywood adhesives, abrasive materials, insulation, foundry binders, brake linings made from
phenolic resins, surface coatings, molding compounds, laminates, wood adhesives made from
melamine resins, phenolic thermosetting, resin curing agents, explosives made from
hexamethylenetetramine, urethanes, lubricants, alkyd resins, acrylates made from
trimethylolpropane, plumbing components from polyacetal resins, and controlled-release fertilizers
made from urea formaldehyde concentrates WHO (1989. as cited in {ATSDR. 1999. 930871.
Formaldehyde is used in smaller quantities for the preservation and embalming of biological
specimens. It is also used as a germicide, an insecticide, and a fungicide in some products. It is
found (as an ingredient or impurity) in some cosmetics and personal hygiene products, such as
This document is a draft for review purposes only and does not constitute Agency policy.
A-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
some soaps, shampoos, hair preparations, deodorants, sunscreens, dry skin lotions, and
mouthwashes, mascara and other eye makeup, cuticle softeners, nail creams, vaginal deodorants,
and shaving cream CNTP. 2010: WHO. 2002: ATSDR. 19991.
Formaldehyde is commonly produced as an aqueous solution called formalin, which is used
in industrial processes and usually contains about 37% formaldehyde and 12-15% methanol.
Methanol is added to formalin to slow polymerization that leads eventually to precipitation as
paraformaldehyde. Paraformaldehyde has the formula (CH2O),,, where n is 8 to 100. It is
essentially a solid form of formaldehyde and therefore has some of the same uses as formaldehyde
fKiernan. 20001. When heated, paraformaldehyde sublimes as formaldehyde gas. This
characteristic makes it useful as a fumigant, disinfectant, and fungicide, such as for the
decontamination of laboratories, agricultural premises, and barbering equipment Long-chain
polymers (e.g., Delrin plastic) are less inclined to release formaldehyde, but they have a
formaldehyde odor and require additives to prevent decomposition.
The major sources of anthropogenic emissions of formaldehyde are motor vehicles, power
plants, manufacturing plants that produce or use formaldehyde or substances that contain
formaldehyde (i.e., adhesives), petroleum refineries, coking operations, incineration, wood burning
and tobacco smoke. Among these anthropogenic sources, the greatest volume source of
formaldehyde is automotive exhaust from engines not fitted with catalytic converters fNEG. 20031.
The Toxic Release Inventory (TRI) data for 2016 show total releases of 19.4 million pounds with
about 13 million to underground injection (EPA TRI Explorer,
https://enviro.epa.gov/triexplorer/tri release.chemicall.
Formaldehyde is formed in the lower atmosphere by photochemical oxidation of
hydrocarbons or other formaldehyde precursors that are released from combustion processes
fATSDR. 19991. Formaldehyde can also be formed by a variety of other natural processes, such as
decomposition of plant residues in the soil, photochemical processes in sea water, and forest fires
(National Library of Medicine, 2015).
The input of formaldehyde into the environment is counterbalanced by its removal by
several pathways. Formaldehyde is removed from the air by direct photolysis and oxidation by
photochemically produced hydroxyl and nitrate radicals. Measured or estimated half-lives for
formaldehyde in the atmosphere range from 1.6 to 19 hours, depending upon estimates of radiant
energy, the presence and concentrations of other pollutants, and other factors fATSDR. 19991.
Given the generally short daytime residence times for formaldehyde, there is limited potential for
long-range transport (WHO. 2002). In cases where organic precursors are transported long
distances, however, secondary formation of formaldehyde may occur far from the anthropogenic
sources of the precursors.
Formaldehyde is released to water from the discharges of both treated and untreated
industrial wastewater from its production and from its use in the manufacture of formaldehyde-
containing resins fATSDR. 19991. Formaldehyde is also a possible by-product from using ozone
This document is a draft for review purposes only and does not constitute Agency policy.
A-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
and/or hydrogen peroxide for drinking-water disinfection. In water, formaldehyde is rapidly
hydrated to form a glycol, and the equilibrium favors the glycol.
A.1.2. Human Exposure
General population exposure to formaldehyde can occur via inhalation, ingestion and
dermal contact Each of these pathways and associated media levels are discussed below.
Formaldehyde exposure can also occur occupationally via three main scenarios:
• The production of aqueous solutions of formaldehyde (formalin) and their use in the
chemical industry (e.g., for the synthesis of various resins, as a preservative in medical
laboratories and embalming fluids, and as a disinfectant).
• Release from formaldehyde-based resins in which it is present as a residue and/or through
their hydrolysis and decomposition by heat (e.g., during the manufacture of wood products,
textiles, synthetic vitreous insulation products, and plastics). In general, the use of
phenol-formaldehyde resins results in much lower emissions of formaldehyde than those of
urea- based resins.
• The pyrolysis or combustion of organic matter (e.g., in engine exhaust gases or during
fire fighting) (IARC. 2006a).
Occupational exposures occur not only during the production of products containing
formaldehyde, but also during the use of these products in construction and decoration (Kim etal..
2011). Industries with the greatest potential for exposure include health services, business
services, printing and publishing manufacture of chemicals and allied products, manufacture of
apparel and allied products, manufacture of paper and allied products, personal services,
machinery (except clerical), transport equipment, and furniture and fixtures (IARC. 1995).
Exposure levels for the workers of various professions in a selected number of studies range from
49 to 4,280 |ig/m:i (40 to 3,480 ppb), with plywood particle board production workers having the
highest exposures (Kim etal.. 2011).
In recent years, concerns have been raised regarding occupational exposures resulting from
the use semi-permanent professional hair straightening products. In 2010, responding to requests
from hair salon employees, the National Institute of Occupational Safety and Health (NIOSH)
conducted a study of hair smoothing treatment products marketed as formaldehyde free. McCarthy
et al. (2010) found that the formaldehyde content in a total of 105 samples of these products
ranged from 6.8 to 11.8%, with an average of 8.8%. Air samples taken in seven hair salons during
smoothing treatments showed 8-hour time-weighted average concentrations of formaldehyde
ranging from 7.4 |J.g/m3 (6 ppb) to 407.1 |J.g/m3 (331 ppb) (McCarthy etal., 2010). Air
concentrations vary depending on factors such as room ventilation, ceiling height, room size, and
duration of the treatment (McCarthy et al., 2010). Another study by Pierce etal. f20111 collected
air samples during the use of four commercially available hair smoothing products. The hair stylist
8-hour time-weighted average concentrations of formaldehyde ranged from 24.6 |ig/m:i (20 ppb) to
This document is a draft for review purposes only and does not constitute Agency policy.
A-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
196.8 n.g/m3 (160 ppb) for one treatmentper day and 61.5 ng/m3 (50 ppb) to 922.5 |J.g/m3 (750
ppb) for four consecutive treatments fPierce etal.. 20111. Time weighted average concentrations
decreased as the distance from the treatment location increased fPierce etal.. 20111.
Inhalation
Ambient air monitoring data for formaldehyde are available from EPA's Ambient
Monitoring Archive for HAPs which includes data from the Air Quality System database and other
data sources (https://www.epa.gov/amtic/amtic-air-toxics-data-ambient-monitoring-archive).
Measurement data are collected from National Air Toxic Trends Sites (NATTS) and other sites
across the country operated by state, local, and tribal agencies that are not part of the NATTS
network. Data for the year 2018, come from 100 monitors located in 27 states and the District of
Columbia. The annual means for these monitors range from 0.25-11.06 |J.g/m3 (0.20-9.01 ppb) and
have an overall average of 2.97 |J.g/m3 (2.42 ppb). The annual means were derived by EPA through
averaging all available daily data from each site that has at least three valid quarters for the year
(i.e., a valid quarter is a quarter that contains at least seven daily averages)
(https://www.epa.gov/system/files/documents/2021-08/annual-average-statistics-
documentation-2018.pdf). Table A-2 presents the data by land use category based on the annual
means from each site for 2018. The land use is established in the Air Quality System database from
the site description.
This document is a draft for review purposes only and does not constitute Agency policy.
A-5 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
2014 NATA Formaldehyde Ambient Concentrations
Contribution by Sector
¦ Point "Biogenic "Non-point
¦ Mobile Onroad "Secondary ¦ Mobile Nonroad
Figure A-2. Formaldehyde Ambient Concentrations Contribution by Sector.
Source: Based on data provided by M Woody (EPA/OAR)
Table A-2. Ambient air levels by land use category based on 2018 annual site
averages
Annual formaldehyde ambient air concentrations by category (ng/m3)
Agriculture
Commercial
Forest
Industrial
Mobile
Residential
Number of annual averages
5
31
4
n
6
43
Mean
2.02
2.88
1.98
3.42
3.80
3.00
Minimum
1.40
0.25
1.03
1.74
2.02
0.88
Maximum
2.61
4.84
3.40
8.25
5.71
11.06
Source: EPA's Ambient Monitoring Archive for HAPs which includes data from the Air Quality System and other
data sources at https://www.epa.gov/amtic/amtic-air-toxics-data-ambient-monitoring-archive.
1 In general, ambient levels of formaldehyde in outdoor air are significantly lower than those
2 measured in the indoor air of workplaces or residences fATSDR. 1999: IARC. 19951. Indoor sources
3 of formaldehyde in air include volatilization from pressed wood products, carpets, fabrics,
4 insulation, permanent press clothing, latex paint, and paper bags, along with emissions from gas
5 burners, kerosene heaters, and cigarettes. Kim et al. (20151 suggested that air fresheners, scented
This document is a draft for review purposes only and does not constitute Agency policy.
A-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
candles, and electric diffusers may also contribute to indoor concentrations of formaldehyde.
Indoor air levels are affected by the age of the source materials, temperature, humidity, and
ventilation rates fParthasarathv et al.. 2 011: IARC. 2006bl Release rates of formaldehyde from
consumer products have been published in the literature. Table A-3 presents a selected number of
products and their respective emission rates in [ig/m2-hr.
In general, the major indoor air sources of formaldehyde can be described in two ways: (1)
those sources that have the highest emissions when the product is new with decreasing emission
over time, as with the first set in the examples above; and (2) those sources that are reoccurring or
frequent such as the second set of examples above. Several studies were found in the literature that
investigated indoor air concentrations of formaldehyde in various housing types. Median indoor air
concentrations in various European countries in both commercial and residential buildings ranged
from 10 |J.g/m3 to 50 |ig/m:i fSarigiannis etal.. 20111. A summary of residential indoor air data in
the U.S. and Canada is provided in Table A-4. These are organized by manufactured (i.e., mobile
homes/trailers with wheels that are designed to be moved) and conventional housing and in
chronological order, beginning with the most recent studies. Results vary depending on housing
characteristics and date of study. In general, higher concentrations are found in manufactured
houses.
Even though formaldehyde levels in construction materials have declined, indoor inhalation
concerns still persist For example, as shown in Table A-4, studies have measured formaldehyde
levels in manufactured homes. ATSDR (2007) reported on air sampling in 96 unoccupied trailers
provided by the Federal Emergency Management Agency (FEMA) used as temporary housing for
people displaced by Hurricane Katrina (see Table A-4). Formaldehyde levels in closed trailers
averaged 1,279 ± 849 |ig/m3 (mean ± standard deviation [SD]) (1.04 ± 0.69 ppm), with a range of
12-4,500 |J.g/m3 (0.01-3.66 ppm). The levels decreased to an average of 480 ± 324 |ig/m3 (0.39 ±
0.27 ppm), with a range of 0.00-2,005 |ig/m3 (0.00-1.63 ppm) when the air conditioning was
turned on. Levels also decreased to an average of 111 ± 98 |ig/m3 (0.09 ± 0.08 ppm), with a range
of 12-603 |J.g/m3 (0.01-0.49 ppm) when the windows were opened. ATSDR (2007) found an
association between temperature and formaldehyde levels; higher temperatures were associated
with higher formaldehyde levels in trailers with the windows closed. They also noted that different
commercial brands of trailers yielded different formaldehyde levels.
In December 2007 and January 2008, the Centers for Disease Control and Prevention (CDC)
measured formaldehyde levels in a stratified random sample of 519 FEMA-supplied occupied travel
trailers, park models, and mobile homes ("trailers") (CDC, 2008). At the time of the study, sampled
trailers were in use as temporary shelters for Louisiana and Mississippi residents displaced by
hurricanes Katrina and Rita. The geometric mean level of formaldehyde in sampled trailers was 95
|ig/m3 (77 ppb), and the range was 3.7-726 |ig/m3 (3-590 ppb) (see Table A-4).
Another study by Maddalena et al. (2008) measured indoor air concentrations for a range of
volatile organic compounds (VOCs), including formaldehyde in four unoccupied temporary housing
This document is a draft for review purposes only and does not constitute Agency policy.
A-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
units (i.e., mobile homes) under steady state ventilation conditions. A morning and afternoon
measurements were taken for each unit. The overall average air concentration of formaldehyde for
the four mobile homes was 569 |ig/m3. This is consistent with values measured by ATSDR (2007)
and CDC (2008). Consistently higher air concentrations of formaldehyde were measured in the
afternoon samples.
Air concentrations of formaldehyde were lower for conventional housing as shown in Table
A-4. Mean values from studies published between 1980 and 2008 ranged from 6.2 to >1,230
|ig/m3. Although no conclusions could be drawn based on the age of the study alone, some of the
studies in Table A-4 suggests that air concentrations are influenced by the age of the house and
season of the year. Lower air concentrations were observed as the age of the house increased.
Higher concentrations were generally observed during the summer months.
Salthammer etal. f 20101 present a thorough review of formaldehyde sources and levels
found in the indoor environment. Based on an examination of international studies carried out in
2005 or later they conclude that the average exposure of the population to formaldehyde is 20 to 40
|ig/m3 under normal living conditions. Figure A-3 summarizes the range of formaldehyde air
concentrations in various environments. The dotted line represents the WHO guidelines of 100
|ig/m:i. More recently, Branco etal. f20151 measured hourly mean formaldehyde concentrations as
high as 204 |ig/m3 in nursery schools in Portugal.
Data on formaldehyde levels in outdoor and indoor air were collected under Canada's
National Air Pollution Surveillance program (WHO. 2002: Health Canada. 2001). The effort
included four suburban and four urban sites sampled in the period 1990-1998. A Monte Carlo
analysis applied to the pooled data [n = 151) was used to estimate the distribution of time-weighted
24-hour air exposures. This study suggested that mean levels in outdoor air were 3.3 |ig/m3 (2.7
ppb) and mean levels in indoor air were 35.9 |ig/m3 (29.2 ppb) fHealth Canada. 20011. The
simulation analysis also suggested that general population exposures averaged 33-36 |ig/m3
(27-30 ppb).
Since the early to mid 1980s, manufacturing processes and construction practices have
been changed to reduce levels of indoor formaldehyde emissions (ATSDR. 1999). A 2008 law
enacted by the California Air Resource Board (Final Regulation Order: Airborne Toxic Control
Measure to Reduce Formaldehyde Emissions from Composite Wood Products;
http://www.arb.ca.gov/regact/2007/compwood07/fro-final.pdf) has limited the amount of
formaldehyde that can be released by specific composite wood products (i.e., hardwood plywood,
particle board, and medium density fiberboard) sold, supplied, or manufactured for use in
California. For this reason the mean indoor air levels presented by Health Canada (2001) (based on
samples collected from 1989-1995) may overestimate current levels.
This document is a draft for review purposes only and does not constitute Agency policy.
A-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-3. Formaldehyde emission rates from various consumer products
Products
Emission Rate (ng/m2-hr)
Reference
Pressed wood products
ND-1,500
Pickrell (Pickrell et al., 1983)
New clothing
0.63-31.25
Pickrell (Pickrell et al., 1983)
Insulation products
2.17-25.83
Pickrell (Pickrell et al., 1983)
Paper plates and cups
3.13-41.67
Pickrell (Pickrell et al., 1983)
Fabrics
ND-14.58
Pickrell (Pickrell et al., 1983)
Carpets
ND-2.71
Pickrell (Pickrell et al., 1983)
Carpets with urethane foam backing
411-6a
Yu (Yu and Crump, 1998)
Textile carpet
83-36a
Yu (Yu and Crump, 1998)
Carpet with synthetic/PVC fibers
120-lla
Yu (Yu and Crump, 1998)
Carpet assembly
153,000-783a
Yu (Yu and Crump, 1998)
Carpet underlay
8,110-12a
Yu (Yu and Crump, 1998)
Vinyl/PVC flooring
22,280-9 la
Yu (Yu and Crump, 1998)
Linoleum flooring
220-223
Yu (Yu and Crump, 1998)
Vinyl tiles
91-45a
Yu (Yu and Crump, 1998)
Rubber floorings
l,400b
Yu (Yu and Crump, 1998)
Soft plastic flooring
590b
Yu (Yu and Crump, 1998)
Cork floor tiles
805-7a
Yu (Yu and Crump, 1998)
Mineral wool insulation batt
15-12b
Yu (Yu and Crump, 1998)
Glass wool fibrous insulation
4-0.08
Yu (Yu and Crump, 1998)
Extruded polystyrene thermal insulants
l,400-22a
Yu (Yu and Crump, 1998)
Extruded polyethylene duct and pipe insulants
0.8-0.28b
Yu (Yu and Crump, 1998)
Plastic laminated board
0.4b
Yu (Yu and Crump, 1998)
Vinyl and fiber glass wallpaper
300b
Yu (Yu and Crump, 1998)
PVC foam wallpaper
230
Yu (Yu and Crump, 1998)
PVC wall covering
100
Yu (Yu and Crump, 1998)
Vinyl coated wallpaper
95-20
Yu (Yu and Crump, 1998)
Vinyl wallpaper
40
Yu (Yu and Crump, 1998)
Wallpaper
100-31
Yu (Yu and Crump, 1998)
Vapor barriers (bituminous tar)
6.3°
Yu (Yu and Crump, 1998)
Black rubber trim for jointing
103
Yu (Yu and Crump, 1998)
Vinyl covering
46-30d
Yu (Yu and Crump, 1998)
Textile wall and floor coverings
l,600b
Yu (Yu and Crump, 1998)
Acoustic partitions
158-6a
Yu (Yu and Crump, 1998)
Office chair
l,060-100a
Yu (Yu and Crump, 1998)
Particle board
1,500-2,167e
200-283
Pickrell et al. (1984)
Yu (Yu and Crump, 1998)
Plywood
1,292-1,375e
1,450-44
Pickrell et al. (1984)
Yu (Yu and Crump, 1998)
Bare urea-formaldehyde wood products (%- %")
8.6-1,580f
Kellv et al. (1999)
This document is a draft for review purposes only and does not constitute Agency policy.
A-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Products
Emission Rate (ng/m2-hr)
Reference
Coated urea-formaldehyde wood products
<2.7-460f
Kellv etal. (1999)
Permanent press fabric
42-215f
Kellv etal. (1999)
Decorative laminates
4.2-51f
Kellv etal. (1999)
Fiberglass products
16-32f
Kellv etal. (1999)
Bare phenol-formaldehyde wood products
4.1-9.2f
Kellv etal. (1999)
Paper grocery bags
to
o
V
Kellv etal. (1999)
Paper towels
A
o
'°X
Kellv etal. (1999)
Latex paint
326-854b
Kellv etal. (1999)
Finger nail hardener
178,000-215,500b
Kellv etal. (1999)
Nail polish
20,700b
Kellv etal. (1999)
Commercially applied urea-formaldehyde floor finish
421-l,050,000b
Kellv etal. (1999)
a The first number in the range indicates initial emissions; the second number indicates emissions after some time
(e.g., hours, days, months).
b Values represent initial emissions.
c 124 days old.
d <98 days old.
e Range indicates different test conditions in temperature and relative humidity.
f Emission rates represent typical conditions, defined as 70 °F, 50% Relative Humidity, and 1 air change per hour.
Table A-4. Studies on residential indoor air levels of formaldehyde
Location (year measured)
Na
Concentration mean (range);
Hg/m3
Reference
Manufactured housing
LA & MS, FEMA-supplied temporary housing
units (Dec. 2007-Jan. 2008)
519b
95 (3.7-726)c
CDC, 2008
FEMA 4 temporary housing units (2007)
4b
569 (331-926)
Maddalena et alv
2008
Baton Rouge, LA, 96 FEMA-supplied temporary
housing units (2006)
Baselined
Ventilation with air conditioning and
bathroom vents only
Ventilation with open windows and vents
96
852
863
1,279(12-4,500)
480 (0-2,005)
111 (12-603)
ATSDR, 2007
Florida, new manufactured house (2000)
NR
95 (NR)
Hodgson et al., 2OO20
United States, East and Southeast (1997-98)
4
42 (26-58)
Hodgson et al., 20006
California, mobile homes (1984-85)
470
86-lll(NR)
(Sexton et al., 1989)'
United States (NR)
Complaint mobile homes
Newer mobile homes
Older mobile homes
>500
260
123-1,107 (0-5,166)
1,032
308
(Gammage and
Hawthorne, 1985)
Texas, mobile homes whose residents
requested testing (1979-82)
Homes < 1 yr old
Homes > 1 yr old
443b
NR (ND-9,840)
> 2,460 for 27% of homes
> 2,460 for 11.5% of homes
Norsted et al. 1985f
This document is a draft for review purposes only and does not constitute Agency policy.
A-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Location (year measured)
Na
Concentration mean (range);
Hg/m3
Reference
United States (NR)
430b
> 1,230 for 4% of samples
615-1,230 for 18% of samples
123-615 for 64% of samples
< 123 for 14% of samples
Breysse, 1984s
United States (NR)
431b
470 (12-3,599)
(Ulsamer et al.,
1982)g
United States (NR)
Complaint homes, WA, < 2 yr old
Complaint homes, WA, 2-10 yr old
Complaint homes, MN, < 2 yr old
Complaint homes, MN, 2-10 yr old
Complaint homes, Wl, < 2 yr old
Complaint homes, Wl, 2-7 yr old
Random sample, Wl, < 2 yr old
110b
77b
66b
43b
38b
9b
NR
950 (NR)
581 (NR)
1,041 (NR)
339 (NR)
891 (NR)
560 (NR)
661 (NR)
Stone et al., 1981s
Wisconsin, complaint homes, 0.2-12 yr old (NR)
65b
590h (NR)
(Dallv et al., 1981)g
Conventional housing or unspecified
Summer Field, CA (2006)
52b
36 (4.7-143.6)
Offerman et al., 2008
Quebec, Canada (2005)
96b
30 (9.6-90)
(Gilbert et al., 2006)
Prince Edward Island, Canada (winter 2002)
59b
39.0 (5.5-87.5)
(Gilbert et al., 2005)
Los Angeles, CA; Houston, TX, and Elizabeth, NJ
(summer 1999-spring 2001)
398
22 ± 7.1'
Weisel etal., 2005
New York City, NY(46 houses)(1999), Los
Angeles, CA (41 houses) (2000)
NYC (winter)
NYC (summer)
LA (winter)
LA (fall)
37
41
40
33
12 ±4.7 (5.2-22)
21 ± 11 (5.8-51)
21 ± 11 (7.9-59)
16 ±6.2 (8.2-32)
(Sax et al., 2004)
Canada (1989-1995)
Northwest Territories; Windsor, Ontario;
Hamilton, Ontario; Trois-Rivieres, Quebec;
Saskatoon, Saskatchewan
151
36 (12-144)
(Environment
Canada, 2000)
United States, East and Southeast, site-built
houses (1997-1998)
7
44j (17-71)
Hodgson et al., 20006
Arizona (Jun. 1995-Feb. 1998)
189
21h (max. 408)
(Graf etal., 1999)
Louisiana, 53 houses: 75% urban;25% rural (NR)
419
460 (ND-6,599)
Lemus et al., 19980
Boston, MA (1993)
winter, 4 residences
summer, 9 residences
14
26
13.7 (7.4-19.8)
19.8 (7.3-66.2)
(Reiss et al., 1995)0
Maryland (1995)
Newly build house
30 days after installation pressed wood
lb
<94
55
(Hare et al., 1996)
Colorado (1992-93)
Prior to occupancy
After occupancy for 5 months
9
26 (8.0-66)
49 (33.0-81.2)
Lindstrom et al.,
19950
New Jersey, 6 residential houses (1992)
36
67.1 (33-125)
Zhang et al., 1994
This document is a draft for review purposes only and does not constitute Agency policy.
A-11 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Concentration mean (range);
Location (year measured)
Na
Hg/m3
Reference
Arizona, houses (NR)
202b
32 (max. 172)
Krzvzanowski et al.
(1990)d
United States, residential, various locations
273
44.0h(NR)
Shah and Singh,
(1981-84)
1988b
San Francisco, CA, Bay Area (1984)
(Sexton et al., 1986)b
Kitchen
48
50 (NR)
Main bedroom
45
44 (NR)
United States (NR)
(Gammage and
Homes with UFFI
>1,200
62-148 (123-4,182)
Hawthorne, 1985)
Homes with UFFI
131
31-86 (12-209)
Pullman, WA, houses (NR)
NR
6.2-89 (NR)
Lamb et al., 1985f
United States (NR)
Breysse, 1984s
UFFI houses
244b
> 1,230 for 2.8% of samples
615-1,230 for 1.9% of samples
123-615 for 24.1% of samples
Non-UFFI houses and apartments
59b
< 123 for 71.2% of samples
> 1,230 for 1.8% of samples
615-1,230 for 1.8% of samples
123-615 for 36.3% of samples
< 123 for 60.1% of samples
United States (1982)
Hawthorne etal.,
Houses 0-30 yr old
40b
75.9 ± 95.0'
1983s
Houses 0-5 yr old
18b
103.0 ± 112.1'
Houses 5-15 yr old
llb
52.0 ± 52.0'
Houses > 15 yr old
llb
39.0 ± 52.0'
Houses 0-5 yr old
18b
107.0 ± 114.0'
spring
137 ± 125'
summer
58.o ± es.o1
autumn
Houses 5-15 yr old
llb
53.0 ± 49.0'
spring
60.0 ± 59.0'
summer
41.9 ±43.1i
autumn
Houses > 15 yr old
llb
44.o ± es.o1
spring
36.0 ± 46.0'
summer
32.0 ± 28.0'
autumn
United States (1983)
Grimsrud et al., 1983s
Energy-efficient new houses
20b
76 (NR)
Low-ventilation modernized houses
16b
37 (NR)
United States (1981)
(Ulsamer et al.,
Houses without UFFI
41b
40 (12-98)
1982)s
Houses with UFFI
636b
150 (12-4,200)
United States (1980-81)
Offerman et al., 1982s
Houses averaging 2 yr old
9b
air-tight construction
44 ± 22'
This document is a draft for review purposes only and does not constitute Agency policy.
A-12 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Location (year measured)
Na
Concentration mean (range);
Hg/m3
Reference
mechanical ventilation
Houses averaging 6 yr old (loose
construction)
ib
33 ± 20'
17 (NR)
United States (1978-79)
13b
120h(NR)
(Dallv etal., 1981)g
United States (1979)
Energy-efficient house
Unoccupied house without furniture
Unoccupied house with furniture
Occupied house
day
night
2b
98 (40-150)
81 ± 7.0'
225 ± lS.O1
263 ± 26.0'
141 ± 44.0'
Berk et al., 1980s
Note: Concentrations were converted from ppb to ng/m3 for consistency (1 ppb = 1.23 ng/m3).
ND = not detected; NR = not reported.
a Number of samples unless denoted with footnote (b).
b Number of houses.
c Geometric mean.
d Baseline refers to initial levels measured 4 days prior to intervention phase of the study during which
ventilation via air conditioning or open windows was provided.
e Cited in IARC {2006, 2825926@@author}
f Cited in ATSDR {1999, 93087@@author}.
s Cited in WHO {1989,1256168@@author}.
h Median.
1 Standard deviation.
Source: Adapted from NTP {2010,1041161@@author} and other sources as noted.
Remote Air
Rural Air
WHOGuidelines
Urban Air
Normal Indoor Air
I 1
Polluted Indoor Air
I 1
Extreme Conditions
h
o.i
10
100
1000
Figure A-3. Range of formaldehyde air concentrations (ppb) in different
environments.
This document is a draft for review purposes only and does not constitute Agency policy.
A-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for Formaldehyde—Inhalation
Notes: Graph is in logarithmic scale; "Normal indoor conditions," "polluted indoor conditions," and "extreme
conditions" were not defined.
Source: Salthammer et al. (2010).
In addition, the Canadian indoor air data may overestimate formaldehyde levels in U.S.
homes, because many residential homes in Canada use wood burning stoves more frequently and
have tighter construction (due to colder winters), leading to less dilution of indoor emissions. The
outdoor air levels, however, appear to have remained fairly constant over recent years, and the
median outdoor level from the Canadian study (2.8 [ig/m3) (2.3 ppb) is very similar to the median
of the U.S. monitoring data (2.83 [ig/m3) (2.3 ppb) in 1999.
Indoor air measurements combined with information about daily activity diaries have been
used as surrogate of personal exposures. A recent study conducted with 41 children ages 9-12
years old in Australia concluded that although indoor air measurements from stationary monitors
tended to slightly overestimate personal exposures, they were a good surrogate of personal
exposures to children (Lazenby et al.. 2012). The mean exposure from personal monitors ranged
from <5 to 34 |ig/m3 (<4-26.3 ppb) with a mean of 13.7 |ig/m3 (11.1 ppb) fLazenby etal.. 20121.
Ingestion
Limited U.S. data indicate that concentrations in drinking water may range up to
approximately 10 |ig/L in the absence of specific contributions from the formation of formaldehyde
by ozonation during water treatment or from leaching of formaldehyde from polyacetyl plumbing
fixtures (WHO. 2002). In the absence of other data, one-half this concentration (5 |ig/L) was judged
to be a reasonable estimate of the average formaldehyde in Canadian drinking water.
Concentrations approaching 100 |ig/L were observed in a U.S. study assessing the leaching of
formaldehyde from domestic polyacetal plumbing fixtures, and this concentration was assumed to
be representative of a reasonable worst case fWHO. 20021.
Formaldehyde has been used in the food industry for the preservation of dried foods, fish,
certain oils and fats, and disinfection of containers (ATSDR. 19991. Formaldehyde is a natural
component of a variety of foodstuffs (1995: WHO. 1989). However, foods may be contaminated
with formaldehyde as a result of fumigation (e.g., grain fumigation), cooking (as a combustion
product), and release from formaldehyde resin-based tableware flARC. 19951. Also, the compound
has been used as a bacteriostatic agent in some foods, such as cheese flARC. 19951. There have
been no systematic investigations of levels of formaldehyde in a range of foodstuffs that could serve
as a basis for estimation of population exposure (Health Canada. 2001). According to the limited
available data, concentrations of formaldehyde in food are highly variable. In the few studies of the
formaldehyde content of foods in Canada, the concentrations were within a range of
<0.03-14 mg/kg (Health Canada. 2001). Data on formaldehyde levels in food have been presented
by Feron etal. f 19911 and fWHO. 19891 from a variety of studies, yielding the following ranges of
measured values:
This document is a draft for review purposes only and does not constitute Agency policy.
A-14 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
• Fruits and vegetables: 3-60 mg/kg
• Meat and fish: 6-20 mg/kg
• Shellfish: 1-100 mg/kg
• Milk and milk products: 1-3.3 mg/kg
Daily intake of formaldehyde was estimated by WHO f 19891 to be in the range of 1.5-14 mg
for an average adult Similarly Fishbein (19921 estimated that the intake of formaldehyde from
food is 1-10 mg/day but discounted this on the belief that it is not available in free form. Although
the bioavailability of formaldehyde from the ingestion of food is not known, it is not expected to be
significant fATSDR. 19991. Using U.S. Department of Agriculture (USDA) consumption rate data for
various food groups, Owen etal. (1990) calculated that annual consumption of dietary
formaldehyde results in an intake of about 4,000 mg or approximately 11 mg/day.
A.l.1.1. Dermal Contact
The general population may have dermal contact with formaldehyde-containing materials,
such as some building products and cosmetics (see Section 1.2 for the details on these products).
Generally, though, dermal contact is more of a concern in occupations that involve handling
concentrated forms of formaldehyde, such as those occurring in embalming and chemical
production.
A.2. TOXICOKINETICS OF INHALED AND ENDOGENOUS FORMALDEHYDE
This chapter presents specific information on the toxicokinetics [absorption, distribution,
metabolism, and excretion (ADME)] of inhaled and endogenously-produced formaldehyde from
human and experimental animal studies. Although toxicokinetics is typically discussed in a
sequential manner [i.e., with absorption defined as delivery to the blood; distribution describing
delivery to the target tissue(s); metabolism outlining conversion to a more-or-less active chemical
species, often metabolism occurs in liver, target tissue elsewhere; and excretion documenting tissue
clearance and removal processes], the primary site of action of inhaled formaldehyde is at the
portal of entry (POE), specifically within the upper respiratory tract (URT). Therefore, this section
will first discuss the uptake (also referred to as "absorption" in the formaldehyde literature) of
inhaled formaldehyde into the URT tissue, and its transport, metabolism, and removal within the
POE. Following this is a description of what is known regarding the absorption of formaldehyde
from the POE into the blood and the potential for distribution of exogenous formaldehyde to
systemic sites, along with a discussion of formaldehyde metabolism and excretion processes that
may occur outside of the POE.
This document is a draft for review purposes only and does not constitute Agency policy.
A-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
Formaldehyde is produced endogenously during normal cellular metabolism and as a
byproduct of lipid peroxidation, or as a product in the catabolism of other chemicals introduced
through dietary environmental, or pharmaceutical sources. Therefore, discussions of inhaled
formaldehyde require a consideration of the potential impact of endogenous formaldehyde on its
toxicokinetics, as well as on its toxicity. The available evidence on the metabolism and kinetics of
endogenous formaldehyde is discussed within each of the following subsections specifically as it
pertains to the toxicokinetics of exogenous formaldehyde.
In the last subsections, the available toxicokinetic models of formaldehyde are presented.
A.2.1. Toxicokinetics of Inhaled Formaldehyde at the Portal of Entry (POE)
Formaldehyde is a highly reactive, highly water soluble, respiratory irritant, towards which
the human body has developed several detoxification and removal processes at the site(s) of first
contact (e.g., nasal passages for inhalation). Thus, this discussion of the toxicokinetics of inhaled
formaldehyde at the POE is organized according to the most likely sites of first contact between
inhaled formaldehyde and biological materials, in the context of the known anatomy and potential
elimination processes of the respiratory tract tissues. Several of the key considerations for
evaluating the toxicokinetics of inhaled formaldehyde at the POE in the rat nose are represented
schematically in Figure A-4. The respiratory tract is divided broadly as (1) upper respiratory tract
(URT), which includes the nasal cavity, pharynx, and larynx and (2) the lower respiratory tract
(LRT) comprising the trachea, bronchi, and lungs. Species differences in the structure of the
airways, as well as the composition of the surface epithelium at various nasal locations, are
important considerations to keep in mind when interpreting results in rodents and extrapolating
observations to humans. Nasal passages, starting from anterior to posterior, are lined by four
different types of epithelia: (1) squamous or keratinized, stratified (nasal vestibule); (2)
transitional or nonciliated cuboidal/columnar; (3) respiratory or ciliated pseudostratified
cuboidal/columnar (main chamber and nasopharynx); and (4) olfactory (dorsal and dorsoposterior
nasal cavity) (Harkema et al.. 2006). It is important to note that rodents and humans differ in the
distribution of nasal epithelial surfaces. For example, the olfactory epithelium in rats and mice
makes up approximately 50-52% and 45-47%, respectively, of the nasal cavity surface area,
whereas in humans, it makes up only 3% (Sorokin et al.. 1988: Gross etal.. 1982).
This document is a draft for review purposes only and does not constitute Agency policy.
A-16 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Rat nose [formaldehyde
(in red), squamous
epithelium (se),
respiratory epithelium
(re), nasal turbinates with
olfactory epithelium (NT-
oe) or with re (NT-re),
cribiform plate (CP), and
olfactory bulb (OB)]
Nasopharynx (to lower
respiratory tract)
Gradient ot
formaldehyde
concentration
Vialt*
Inspired air and
formaldehyde (red)
J Mucus
I Cilia
Epithelium
[epithelial cells
(EC) and goblet
cells (GC)]
Basement membrane
Lamina propria
[systemic circulation
(SC) and lymphoid
tissue (NALT)]
Figure A-4. Schematic of the rat upper respiratory tract depicting the gradient
of formaldehyde concentration formed following inhalation exposure, both
from anterior to posterior locations, as well as across the tissue depth.
Modeling based on observations in rodents predicts a similar pattern of distribution
in humans. Drawn based in part on images by NRC (2011) and Harkema et al.
(2006). Note: other components (e.g., naris; transitional epithelium) have been
omitted to increase clarity.
1 A.2.2. Spatial Distribution of Tissue Uptake of Formaldehyde at the Portal of Entry
2 The distribution of inhaled formaldehyde within the URT and LRT can provide information
3 useful to interpreting any potential toxicity. The nasal passages in humans are generally similar to
4 other mammalian species. One key difference, however, is that humans and nonhuman primates
5 have nasal passages adapted for both oral and nasal (oronasal) breathing as opposed to obligate
6 nasal breathing in rodents. A second key difference regards the shape and complexity of the nasal
7 turbinates, with relatively simple shapes in humans, and complex, folded patterns in rodents. In
8 general, these differences provide better protection of the rodent LRT against inhaled toxicants
9 than is provided to the human LRT (Harkema etal.. 2006).
This document is a draft for review purposes only and does not constitute Agency policy,
A-17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
Indirect measurement studies
Much of what is known regarding the uptake of formaldehyde is based on indirect
measurements of formaldehyde-induced changes and/ or molecular interactions, or removal of
formaldehyde from the air. This is because, in biological systems, formaldehyde exists as total or
analyzable formaldehyde, which includes free and reversibly bound (acid-labile) forms (Heck et al..
1982). Conventional methods cannot directly measure low levels of free formaldehyde with
certainty in tissues and body fluids. Additionally, carbonyl impurities such as acetone,
formaldehyde and acetaldehyde are present even in quartz distilled water and may interfere in the
measurements fEsterbauer etal.. 19821. Uptake of formaldehyde (defined as retention within the
respiratory tract tissue), based on rough estimates determined from the amount of formaldehyde
removed from the air, indicate that majority large percentage of formaldehyde is removed from
inhaled air by the URT.
Indirect estimates of formaldehyde uptake, based on interactions with cellular materials,
have been made in experimental animals, including monkeys (Casanova etal.. 1991: Monticello et
al.. 1989). dogs (Egle. 1972). and rats (Kimbell et al.. 2001b: Chang etal.. 1983: Heck etal.. 1983:
Kerns etal.. 19831 as shown in Table A-5.
Table A-5. Dosimetry and response of formaldehyde in experimental animals
by indirect measurements
Reference and
species
Exposure and analysis
Observations
Casanova et al
(1991):
0.86, 2.46, 7.38 mg/m3for 6-hr
[14C]CH20 from [14C]PFA.
Estimated the amount of DNA-
protein crosslinks (DPX) formed
in various tissues
DPX Levels
Area of the respiratory tract
Monkeys, rhesus;
male, n=9; 8.74 kg;
4.6 yr old
Highest
Middle turbinate mucosa
Lower
Anterior lateral wall/septum and nasopharynx
Very low
Larynx/trachea/carina
None
Maxillary sinuses and lungs
Monticello et al.,
(1989) Monkeys,
rhesus;
male,
n=9; 4-6 yrs; 6-7 kg
7.4 mg/m3, 6 hrs/d; 5 d/wk; 1 or
6 wk CH20 from PFA. Animals
injected with [3H]-Thd, sacrificed,
histoauto-radiography of cell
proliferation measured
Proliferation
Area of the respiratory tract
Significant
Nasal passages
Minimal
Lower respiratory tract
None
Maxillary sinuses
Egle, (1972)
Dogs/Mongrel;
Male and female;
n=4; 13-19 kg
150 to 350 mg/m3 CH20 vapors
from formalin: nose-only
inhalation from a respirometer;
animals preanesthetized;
Uptake at all ventilation rates and concentrations
Total respiratory tract (TRT)
=100%
URT- inhalation
100%
This document is a draft for review purposes only and does not constitute Agency policy.
A-18 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information for Formaldehyde—Inhalation
Reference and
species
Exposure and analysis
Observations
aldehydes analyzed by a
colorimetric method
URT- inhalation + exhalation
=100%
Heck et al.,
(1983);
Rats, Fischer;
Male,
n=3; 18-250 g
Radioactivity immediately after
6hr exposure to [14C]CH20 from
[14C]PFA, each averaging 3
exposures and 4 rats at 6.2,12.3,
18.5, or 29.5 mg/m3
Equivalents of [14C] in various tissues (nmol/g)a or
mg/m3
6.15
12.3
18.5
29.5
Nasal
Mucosa
0.59 ±0.18
1.15 ±0.29
1.78 ±0.4
2.28 ±0.61
Trachea
0.26 ±0.13
0.39 ±0.13
0.36 ± 0.09
0.40 ± 0.13
Plasma
0.05 ± 0.01
0.08 ±0.01
0.10 ± 0.04
0.11 ±0.05
aValues, representing mean ± SD, were extracted from graphical data using GrabIT software.
CH20, formaldehyde; PFA, paraformaldehyde; DPX, DNA-protein crosslinks.
As shown in Table A-5, Casanova et al., (1991) used DNA-protein crosslinks (DPX or DPC)
levels as a measure of regional dosimetry of formaldehyde in monkeys exposed to formaldehyde by
inhalation assuming that the rate of crosslink formation depends on the concentration of
formaldehyde delivered at the portal of entry tissues. They subjected rhesus monkeys to a single 6-
hr exposure of formaldehyde over a range (0.9-7.4 mg/m3) and concluded based on the observed
pattern of DPX formation that formaldehyde uptake primarily occurs in nasal passages involving
middle turbinates, to a smaller extent in the nasopharynx and trachea, but not in maxillary sinuses
or lungs (Casanova etal.. 1991). Monticello et al. (1989) predicted the uptake of formaldehyde
based on other indirect measures such as cell proliferation in monkeys repeatedly exposed to 7.4
mg/m3 formaldehyde, 6 hrs/day, 5 days/wk for 1 or 6 wks. They concluded that formaldehyde
uptake primarily occurs in nasal passages and middle turbinates, to a smaller extent in the
nasopharynx and trachea, with evidence of increased proliferation in proximal regions of the
bronchi, but no indication of effects in the maxillary sinuses. In dogs exposed to formalin vapors,
almost 100% of inhaled formaldehyde is retained in the URT, indicating that little, if any, inhaled
formaldehyde would reach the LRT, and this is independent of respiration rate, tidal volume, and
inhaled formaldehyde concentration (Egle. 1972).
Similarly, radiolabeling studies, exemplified by Heck et al. (1983) in rats show that the
majority of the labeled formaldehyde is retained within the nasal passages and, to a far lesser
extent, within the other parts of the URT and proximal LRT, with no evidence of significant
distribution into plasma. However, because formaldehyde is incorporated into the one-carbon (1C)
pool (see discussion later in this section), possibly facilitating its distribution in a toxicologically-
inactive form, neither the distribution of radiolabel nor the estimated retention are interpreted to
provide a clear picture of the spatial distribution of inhaled formaldehyde within the respiratory
tract tissues. Notably, long-term exposure of rats to formaldehyde for 30 months induced lesions in
the nasal cavity and proximal trachea fKerns etal.. 19831. Kimbell et al., (2001b) predicted the
This document is a draft for review purposes only and does not constitute Agency policy.
A-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
uptake of formaldehyde in the nasal passages of F344 rats, rhesus monkeys and humans to be
respectively 90%, 67% and 76% using the computational fluid dynamics (CFD) modeling. Similar
to these predictions for rats, Morgan et al., (1986c) demonstrated that rat nasal passages scrubbed
nearly all of the inhaled formaldehyde (on average *97%). In rats, the evidence suggests that
higher concentrations of formaldehyde are taken up in the respiratory mucosa as compared to the
olfactory mucosa (Casanova-Schmitz etal.. 1984b: Swenberg et al.. 1983a).
Extrapolation using fluid dynamic modeling
There are no studies available in the literature that directly addressed uptake of
formaldehyde into the respiratory tract of humans. However, a few modeling studies based on
findings in rodents report estimated uptake of inhaled formaldehyde in humans fKimbell etal..
2001b: Kimbell and Subramaniam. 2001: Overton etal.. 20011. Kimbell et al. (2001b), using a
three-dimensional, CFD model of the nose, predicted human nasal uptake of approximately 76% of
the inhaled formaldehyde at unidirectional steady-state nasal inspiratory flow corresponding to
sleeping activity, decreasing to 58% under heavy exercise activity. Overton et al. (2001) modeled
overall uptake in the entire respiratory tract and predicted that 95% of inhaled formaldehyde is
retained in the respiratory tract in general in any activity state. A detailed description of modeling
efforts in humans and monkeys (and rats) is provided in Appendix B.2.2. Overall, dosimetric
modeling studies in humans have shown close agreement with observations of exposed rodents:
namely, that 90-95% of inhaled formaldehyde is retained in the URT (Kimbell et al.. 2001b:
Overton etal.. 2001: Subramaniam et al.. 1998).
Relationship of formaldehyde uptake to endogenous levels and prior exposure
Heck et al (1982) developed a gas chromatography-mass spectrometry (GC-MS) method to
measure total or analyzable formaldehyde, which includes both free as well as reversibly bound
formaldehyde [hydrated formaldehyde bound to glutathione (GSH) and tetrahydrofolate (THF)].
However, this method does not measure irreversibly bound formaldehyde. Based on this method,
endogenous formaldehyde levels were 1.5-4.3 folds higher atthe POE (i.e., nasal mucosa; *12.6
|ig/g or 0.42 mM) than in other tissues (i.e., testes
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Supplemental Information for Formaldehyde—Inhalation
exposures in rats exposed to the same dose of formaldehyde for the previous 9 days (Heck etal..
19831. In a different study Chang et al. (1983) also observed similar uptake in preexposed as well
as naive rats; however, mice responded differently with naive mice exhibiting more radioactivity
uptake than preexposed mice (see Table A-6). The authors concluded that since mice tend to lower
their minute volume with repeated exposures to formaldehyde, they tend to have less absorption,
hence less radioactivity compared to naive mice. So comparing the results in rats, which do not
alter their minute volume as mice do, it was suggested that repeated exposure does not affect the
uptake of formaldehyde in nasal cavity of rats (Chang etal.. 19831.
Table A-6. Comparison of formaldehyde uptake at the portal of entry with
single or repeated inhalation exposure
Reference and
design
Exposure and analysis
Observations
Heck et al. (1982) Rats,
Fischer
Male, n=8
200-250 g
7.4 mg/m3 [13C] CH20 (from PFA) for 6 hours/d;
10-days exposure; chamber inhalation; CH20
measured as PFPH derivative by GC/MS
Nasal mucosa levels
total3 CH20 (|ig/gb)
Unexposed Exposed
12.6 ±2.7 11.7 ±3.6
Heck etal. (1983)
Rats, Fischer
Male, n=3;
180-250 g
Two groups: (a) preexposure: (b) naive: On Davs 1-9:
group a) received 18.5 mg/m3 CH?0 (from PFA):
whole bodv exposure. 6 hrs/dav: group b): no
preexposure. On Day 10: groups a and b received
[14C] CH20 (from PFA) for 6 hours, nose-only
exposure. Tissue homogenates counted with LSC for
14C02 trapped in ethanolamine in 2-methoxy-
ethanol counted for radioactivity.
Equivalents of 14C
in respiratory mucosa (pg/gc)
naive rats
67.5 ± 9.2
preexposed
64.4 ± 7.6
(No significant difference)
Chang et al. (1983)
Rats, Fischer;
Male, N=3;
180-200 g
i) oreexoosure:
7.4 or 18.4 mg/m3 unlabeled CH20 from PFA, 6
hrs/d, 4-days whole-body exposure; on 5th day
14CH20 from PFA, 6 hrs
ii) naive animals:
14CH20, 6 hrs from PFA
Radioactivity in nasal cavitv:
preexposed rats = naive rats
Radioactivity in nasal cavitv:
naive mice > pretreated mice
Mice, B6C3F1
Male, N=3; 26 g
aTotal formaldehyde includes free plus reversibly bound formaldehyde.
bData from Heck et al. (1982) given in |amols/g is converted to \xg/g by the equation: |amols x 30 = \xg/g (30 is the
molecular weight of formaldehyde).
cData from Heck et al. (Heck et al., 1983) given in nmols/g is converted to converted to ]ug/g by the equation:
(nmol/g /1,000) x 30 = ]Ltg/g) (30 is the molecular weight of formaldehyde).
CH20, formaldehyde; PFA, paraformaldehyde; PFPH, pentafluorophenylhydrazine; GC/MS, gas
chromatography/mass spectrometry; LSC, liquid scintillation counting; C02, carbon dioxide.
Summary of spatial distribution ofPOE uptake
To summarize, a majority of inhaled formaldehyde is rapidly absorbed and retained in the
URT based on CFD modeling studies in humans fKimbell etal.. 2001b: Kimbell and Subramaniam.
2001: Overton etal.. 2001: Subramaniam etal.. 19981. indirect or direct measurements in monkeys
(Monticello etal.. 1989: Casanova et al.. 1988). and direct measurements in dogs (Egle. 1972) and
rats (Kimbell et al.. 2001b: Chang etal.. 1983: Heck etal.. 1983: Kerns etal.. 1983). despite the
anatomical and physiological differences between species, such as obligate nose breathing in
This document is a draft for review purposes only and does not constitute Agency policy.
A-21 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
rodents (rats and mice) and oronasal breathing in primates (monkeys and humans) (Harkema etal..
2006: Schreider. 19861. As demonstrated in monkeys and rats, and as modeled in humans, a
concentration gradient of inhaled formaldehyde follows an anterior to posterior distribution, with
high concentrations of formaldehyde distributed to squamous, transitional and respiratory
epithelia, and less uptake by olfactory epithelium, and very little or no formaldehyde reaching more
distal sites such as the larynx or lung. Further, at inhaled concentrations as high as 7.4 mg/m3,
exogenous exposure does not appreciably change the levels of formaldehyde over the endogenous
levels in the nasal mucosa (Heck etal.. 1982). Also, repeated exposures to formaldehyde do not
alter the tissue formaldehyde levels in rats, but naive mice do show higher tissue uptake than
preexposed mice, which is attributed to species differences in minute volume and response to
irritant gases f Chang etal.. 19831.
A.2.3. Tissue Penetration of Formaldehyde Within the Upper Respiratory Tract
Within the URT, penetration of formaldehyde follows initial interaction with the
mucociliary apparatus followed by diffusion into the epithelial cell layer where it can be
metabolized. Important details to consider in evaluating formaldehyde nasal dosimetry and
toxicity are the differences in the types of epithelium lining the nasal surfaces. As described earlier,
there are striking differences in the amount of olfactory epithelium and respiratory epithelium
present between the noses of rats, which have a highly complex sense of smell, compared to
humans, who use the nose primarily used for breathing. In all species, air (and formaldehyde) must
first pass over squamous, transitional, and respiratory epithelium before coming in contact with
olfactory epithelium. This section will focus on the interaction and fate of inhaled formaldehyde in
the URT.
Formaldehyde interaction with the mucociliary layer
The mucociliary apparatus of the URT is the first line of defense against airborne agents in
that it may entrap, neutralize, and remove particulates and airborne chemicals from inspired air
(Morgan et al.. 1983). The mucociliary apparatus is comprised of three layers: a thick mucus layer
(epiphase) at the top, a watery fluid layer (hypophase) in the middle, and a ciliated epithelial layer
at the bottom fSchlosser. 19991. Inhaled formaldehyde must pass through the mucus layer
covering the URT before it can react with the cellular components in this region.
The respiratory mucus is composed of 97% water, 2-3% glycoproteins, 0.3-0.5% fats, and
about 0.1-0.5% soluble proteins (Bogdanffv etal.. 1987). Formaldehyde gas (unhydrated) is highly
soluble in water, in which ithydrolyzes to a reversible hydrated form called methanediol or
methylene glycol with a half-life of 70 milliseconds and with an equilibrium constant
[CH20]/[CH2(0H)2] of 4.5 x 10"4 at 22°C (Sutton and Downes. 1972). In aqueous solution, most of
the formaldehyde (99.9%) exists as methanediol in an equilibrium with free (0.1%) formaldehyde
fFox etal.. 19851. Thus, formaldehyde is first hydrated in nasal mucus to form methanediol, which
subsequently interacts with the nasal mucociliary apparatus fPriha etal.. 1996: Bogdanffv et al..
This document is a draft for review purposes only and does not constitute Agency policy.
A-22 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
19861. Physical-organic chemistry studies of the reaction of formaldehyde with amines (and
presumably other biological nucleophiles) have conclusively demonstrated that the unhydrated or
free form of formaldehyde, but not the hydrated form or methanediol is the reactive species
fAbrams and Kallen. 19761. Methanediol is either transported to the underlying tissue (presumably
by diffusion) or it is removed within nasal mucus by convective flow and subsequent ingestion.
Schlosser (1999) estimated that 22-42% of the absorbed formaldehyde in rodents is removed by
mucus flow.
Airborne pollutants and reactive gases have been shown to decrease mucus flow rates in
several animal models fas reviewed in Wolff. 19861. Degradation in the continuity or function of
this mucociliary apparatus can impair clearance of inhaled pollutants at the portal of entry. For
example, Morgan et al. (1983) have shown that a single exposure of 18.45 mg/m3 formaldehyde in
Fischer rats causes mucostasis (cessation or severe slowing of mucus flow) in several regions of the
nasoturbinates. Repeated exposure (6 hours/day for 1-9 days) results in ciliastasis (loss of ciliary
activity) occurring with greater frequency and across more regions of the nasoturbinates in
subsequent days of exposure. Thus, continued exposure would be expected to result in an
increased uptake, as well as an altered deposition of inhaled formaldehyde within the URT tissue.
Further, Morgan et al (1986c) also reported that rats exposed 6 hours daily for 3 weeks showed
increase in mucostasis extending from anterior to posterior regions at the 18.45 mg/m3 dose;
however, at lower doses (0.6-7.4 mg/m3) the effect was either undetectable or less severe. In
addition, Morgan et al. (1986c) showed an increase in mucus flow at lower concentrations after
4 days exposure, but not after 6 days to 0.6 mg/m3 formaldehyde. Thus, there are some
uncertainties regarding the occurrence of mucostasis at lower concentrations of formaldehyde
exposure.
In addition, as methanediol and free formaldehyde are transported through the mucociliary
apparatus, the free formaldehyde is known to bind to soluble proteins such as albumin in the nasal
mucus (Bogdanffv etal.. 1987). Similarly, the nasal lining fluid contains antioxidants, including the
thiol GSH with which formaldehyde is known to interact, likely eliciting a transient GSH depletion
during and following formaldehyde exposure. However, it is unclear to what extent inhaled
formaldehyde interacts with soluble and insoluble factors within the mucociliary layer and whether
reactive byproducts may be formed by these interactions. Importantly, endogenous formaldehyde
produced during normal cellular metabolism is unlikely to be present at appreciable levels in the
mucus, and thus, would not be expected to participate in similar reactions. Interactions with
soluble proteins are expected to further reduce the amount of formaldehyde available to react with
cellular materials. As such, alterations in the levels of soluble proteins within the mucus could
substantially affect tissue uptake.
Formaldehyde diffusion into the epithelial cell layer
The less reactive methanediol is better able to penetrate tissues, while the free
formaldehyde reacts with the macromolecules. However, when the free formaldehyde (~0.1%) is
This document is a draft for review purposes only and does not constitute Agency policy.
A-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for Formaldehyde—Inhalation
used up, a fraction of methanediol (from the 99.9%) will convert to free formaldehyde so that the
equilibrium of methanediol with free formaldehyde (i.e. 99.9:0.1 ratio) is maintained in the aqueous
media fFox etal.. 19851. However, several uncertainties exist regarding the transition of inhaled
formaldehyde from the mucociliary layer to the underlying epithelium. Although direct
experimental evidence is lacking, the biochemical properties of formaldehyde make it likely that
inhaled formaldehyde (in the hydrated or anhydrated form) undergoes passive transport, via
simple diffusion, across biological membranes. Thus, higher extracellular formaldehyde levels
would be expected to result in increased diffusion into the cell owing to the concentration gradient
formed. However, this concentration gradient may be affected by endogenous formaldehyde levels
because in humans, as in other animals, formaldehyde is an essential metabolic intermediate in all
cells fThompson et al.. 20091.
Enzymatic metabolism of formaldehyde within cells of the URT
Formaldehyde, either from exogenous sources (inhaled air) or endogenous sources
(enzymatic and nonenzymatic mechanisms as well as that released endogenously from metabolism
of xenobiotics), can be metabolized by several different enzyme pathways. Based on studies of
endogenous formaldehyde and in vitro enzyme inhibition experiments (Teng etal.. 20011. and as
summarized in Figure A-5, formaldehyde has been shown to be predominantly metabolized to
formate by GSH-dependent class III alcohol dehydrogenase (ADH3; also described as formaldehyde
dehydrogenase or FDH) and by a minor pathway involving mitochondrial aldehyde dehydrogenase
2 (ALDH2) which is GSH-independent Catalase may also be involved, to a minor extent, in
oxidizing formaldehyde, especially under conditions when hydrogen peroxide is formed (Uotila and
Koivusalo. 19741.
This document is a draft for review purposes only and does not constitute Agency policy.
A-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
Supplemental Information for Formaldehyde—Inhalation
Enzymatic mechanisms
Intermediary metabolism
Choline metabolism
Sterol metabolism
Amino acid metabolism
Histone lysine demethylation
Xenobiotic
.. .. Metabolism
Non-enzymatic
mechanisms Oxidative
Lipid Peroxidation demethylation of
Oxidativestress N-,o-,ands-
Methanol *rouPs
/ I
One-carbon C02
Metabolism (exhaled
breath)
Na" H COO-
tu rine)
Figure A-5. Metabolism of formaldehyde.
Abbreviations: C02, carbon dioxide; DPX, DNA-protein crosslinks; GSH, glutathione; H20, water; H202, hydrogen
peroxide; HMGSH, hydroxymethylglutathione; NAD+, nicotinamide adenine dinucleotide (oxidized); NADH,
nicotinamide adenine dinucleotide (reduced); Na+HCOO", sodium formate. Enzymes: a, alcohol dehydrogenase-3
(ADH3); b, aldehyde dehydrogenase 2 (ALDH2); c, catalase; d, S-formyl-GSH hydrolase.
Adapted from NTP (2010).
Both ADH3 and ALDH2 enzymes have been found across different species and in a broad
range of tissues, including the nasal mucosa (Reviewed inThompson etal.. 2009). In rodents, both
ADH3 and ALDH2 exhibit region-specific differences in the nose, in that the specific activity of
ADH3 is twice higher in the olfactory mucosa than in respiratory mucosa, while the specific activity
of ALDH2 is 5-8 times higher in respiratory than in olfactory tissue fBogdanffv et al.. 1986:
Casanova-Schmitz etal.. 1984al. In rats, higher levels of ADH3 activity have been reported in the
cytoplasm of the respiratory and olfactory epithelial cells and in the nuclei of olfactory sensory
cells, as compared to other regions of the nasal mucosa (Keller etal.. 1990). These enzymes are
enriched in the nasal tissues presumably to protect the underlying tissues against respired
toxicants. This highlights a significant barrier to the penetration of inhaled formaldehyde beyond
This document is a draft for review purposes only and does not constitute Agency policy.
A-25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information for Formaldehyde—Inhalation
the respiratory epithelium and a means by which these same cells can rapidly metabolize
formaldehyde produced endogenously within the cell fUotila and Koivusalo. 19741.
The ADH3-mediated pathway of formaldehyde oxidation involves a two-step enzymatic
reaction but is preceded by the rapid and reversible nonenzymatic binding of formaldehyde to GSH,
which results in the formation of S-hydroxymethylglutathione (HMGSH) or the glutathione
hemiacetal adduct In the first of a two-step enzymatic reaction, ADH3 converts HMGSH to
S-formylglutathione (S-formyl-GSH) in the presence of the co-factor, nicotinamide adenine
dinucleotide (NAD+). In the second step, another enzyme S-formyl-GSH-hydrolase converts S-
formyl-GSH to formate with the concomitant release of free GSH. Under physiological conditions,
cellular NAD+ levels are two orders of magnitude higher than NADH (reduced form of NAD+) and
intracellular GSH levels are high enough (in millimolar concentrations) to favor rapid oxidation of
HMGSH to formate fSvensson etal.. 1999: Meister and Anderson. 19831. Because of this rapid
metabolism, formaldehyde is likely to have a short half-life in biological systems. As previously
mentioned, and given the importance of this major detoxification pathway, individual variations in
GSH levels within the nasal mucosa are of particular importance in formaldehyde metabolism.
ADH3 shows comparable kinetics across rats and humans. As shown in Table A-7, the
affinity (Km) of purified human liver ADH3 for HMGSH is 6.5 |iM fUotila and Koivusalo. 19741 and
4.5 mM for rat liver fCasanova-Schmitz and Heck. 19831. Hedberg et al. (2000) demonstrated that
the kinetics of ADH3 in human buccal tissue lysates are in close agreement with those reported for
purified human liver ADH3 (Uotila and Koivusalo. 1974). This is comparable to the rat respiratory
and olfactory mucosal Km values in the presence of GSH as well as the Km of ADH3 from rat liver
soluble fraction (2.6 |iM) (Casanova-Schmitz etal.. 1984a). In contrast, the affinity of ALDH2,
presumably represented in the absence of GSH is several-fold lower than ADH3 fSiew etal.. 19761.
Thus, at lower concentrations of formaldehyde ADH3 is the dominant formaldehyde detoxification
pathway. The Km of ADH3 is in close agreement across species and tissue types, including the nasal
mucosa, all of which exhibit similar responses to GSH depletion (i.e., in the absence of GSH, ALDH
family members oxidize formaldehyde, which is associated with mitochondrial ALDH2). Both
ADH3- and ALDH2-mediated pathways oxidize formaldehyde to formic acid (formate). ADH3 is
also known to catalyze the NADP-dependent reduction of the endogenous nitrosylating agent S-
nitrosoglutathione (GSNO) and is also referred to as S-nitrosoglutathione reductase (GSNOR)
flensen etal.. 19981.
Table A-7. ADH3 kinetics in human and rat tissue samples and cultured cells
Source
Km (\M)
Vmax (nmol/mg
protein x min)
References
Purified human liver ADH3
6.5
2.77 ±0.12
Uotila and Koivusalo
(1974)
Rat respiratory mucosal homogenate (+GSH)
2.6 ±2.6
0.90 ±0.24
Casanova-Schmitz et
al. (1984a)
Rat respiratory mucosal homogenate (- GSH)
481 ± 88
4.07 ±0.35
This document is a draft for review purposes only and does not constitute Agency policy.
A-26 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
Source
Km (\M)
Vmax (nmol/mg
protein x min)
References
Rat olfactory mucosal homogenate (+GSH)
2.6 ±0.5
1.77 ±0.12
Rat olfactory mucosal homogenate (- GSH)
647 ± 43
4.39 ±0.14
Rat liver (+GSH)a
4.5 ± 1.9a
2.0 ±0.3
Human buccal tissue (+ GSH)
11 ±2
2.9 ±0.6
Hedberg et al. (2000)
Human buccal tissue (- GSH)
360 ± 90
1.2 ±0.7
aSoluble fraction of rat liver homogenate.
Formate can undergo three possible outcomes: (1) enter the one-carbon pool for use in the
synthesis of DNA and proteins (aka "metabolic incorporation"), (2) become further oxidized to C02
and eliminated in exhaled air, or (3) be excreted in urine (Figure A-5).
One-carbon metabolism
As summarized in Figure A-6, the tetrahydrofolate (THF)-mediated eukaryotic one-carbon
(1C) metabolism involves an inter-connected network which is highly compartmentalized between
the cytosol, mitochondria, and nucleus (Reviewed in Tibbetts and Appling. 201011. A majority of
the 1C metabolism takes place in the mitochondria followed by the cytosol and nucleus. In the
cytoplasmic 1C metabolism, de novo synthesis of purines and thymidylate, and remethylation of
homocysteine to methionine takes place. The 1C metabolism in the mitochondrial compartment
involves formylation of methionyl-tRNA, oxidation of one-carbon donors, such as serine, glycine,
sarcosine, and dimethylglycine (DMG). In addition, mitochondria contribute 1C units for
cytoplasmic 1C metabolism in the form of formate. The mitochondrial and cytoplasmic pathways
are connected by serine, glycine and formate which are the 1C donors. The nuclear compartment of
1C metabolism predominantly provides de novo synthesis of dTMP from dUMP.
Some of the steps in the cytosolic and mitochondrial 1C metabolism are common. Formate,
formed from the metabolism of formaldehyde, enters the 1C pool and is either oxidized to CO2 and
eliminated in exhaled breath or is used in protein and DNA synthesis. As shown in Figure A-6,
formate is combined with THF whereby its 1C group is transferred to THF forming 10-formyl-THF
(10-CHO-THF), mediated by the enzyme 10-HCO-THF-synthetase. The 10-CHO-THF is then
oxidized by CHO-THF dehydrogenase to C02 and H20 and eliminated in the exhaled breath, with the
release of THF which can be reused for binding with formic acid. Alternatively, 10-CHO-THF can
also be converted through two-steps of reversible reactions to 5,10-methenyl-THF (CH+-THF) to
5,10-methylene-THF (CH2-THF). Serine, derived from glycolytic intermediates, is the main source
of 1C units. Serine combined with THF is converted reversibly by the enzyme serine
hydroxymethyl transferase (SHMT) to glycine and CH2-THF. Further, the enzyme methylene
tetrahydrofolate reductase (MTHFR) converts CH2-THF to 5-methyl-THF (CH3-THF). The 1C
metabolism products -CH2-THF and CH3-THF utilize their one-carbon units, respectively, in DNA
(dTMP) and protein (methionine) biosynthetic pathways (metabolic incorporation).
This document is a draft for review purposes only and does not constitute Agency policy.
A-27 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Ado Met
Methyl cyck 7
AdoHcy ~ Hey
Cytoplasm
Folate
— Methionine ^
THF <+-
Mitochondria
CH3-THF
Nucleus
glycolysis
serine -
*¦ THF
CH2-THF (jUMP
glycine .
NADP+
dTMP f3
NADPH
CH+-THF
i2
10-cHo-THF
ADP + P,
*¦ serine
glycine '*4*' glycine
THF <
4m
co2+ nh3
fMet-tRNA
fV!
CH..-THF THF
3m
CH+-THF
2m
sarcosine
DMG -
THF *
betaine
t
t
choline*
THF ^ ATP
formate ¦*-
10-CHO-THF__
ADP + Pi^-st C°2
ATP^V"-- THF
—formate
- DMG
»betaine
Methionine
Hey
choline
Figure A-6. Compartmentalization of mammalian one-carbon metabolism.
The end products, donors, and activated units carried by tetrahydrofolate (THF) of
the 1C metabolism are shown in red, blue, and green, respectively. Note that
reactions 1-4 are common in both the cytoplasmic and mitochondrial (m)
compartments, while reactions 4 and 10 are present in the nucleus (n). Enzymes
catalyzing the reactions: 1: 10-formyl-THF synthetase; 2: 5,10-methenyl-THF (CH+-
THF) cyclohydrolase; 3: 5,10-methylene-THF (CH2-THF) dehydrogenase; 4, 4n, and
4m: serine hydroxymethyltransferase (SHMT); 5: glycine cleavage system; 6: 5,10-
methylene-THF reductase; 7: methionine synthase; 8: dimethylglycine
dehydrogenase (DMGDH); 9: sarcosine dehydrogenase (SDH); 10 and lOn:
thymidylate synthase; 11: 10-formyl-THF dehydrogenase (only the mitochondrial
activity of this enzyme is shown, but it has been reported in both compartments in
mammals); 12: methionyl-tRNA formyltransferase; 13: dihydrofolate (DHF)
reductase; 14: betaine-homocysteine methyltransferase. Abbreviations: AdoHcy, S-
adenosylhomocysteine; AdoMet, S-adenosylmethionine; Hey, homocysteine.
Source: Tibbetts and Appling (2010).
The rate of formate metabolism depends on the availability of dietary folic acid, which is the
main source of THF. It is also important to note that levels of folate intermediates and folate-
dependent enzymes show some differences in rats and primates (see Table A-8).
Table A-8. Levels of folate intermediates, activity of folate-dependent
enzymes, and the rate of oxidation of formate in the liver of various species
Folate intermediate/folate-dependent enzyme
Rat
Monkey
Human
10-formyl-THF (nmoles/g of liver)
4.6 ± 1.3
10.5 ±0.8
3.3 ±0.5
Tetrahydrofolate (nmoles/g of liver)
11.4 ±0.8
7.4 ±0.8
6.5 ±0.3
This document is a draft for review purposes only and does not constitute Agency policy.
A-28 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information for Formaldehyde—Inhalation
Folate intermediate/folate-dependent enzyme
Rat
Monkey
Human
5-CHs-THF (nmoles/g of liver)
9.3 ±0.6
7.6 ± 1.1
6.0 ±0.7
10-formyl-THF synthetase (nmoles of product/min/mg protein)
65.9 ±0.0
142 ± 16
75.0 ±8.7
10-formyl-THF dehydrogenase (nmoles of product/min/mg protein)
88.3 ± 1.7
33.0 ±4.0
23.0 ±2.2
5,10-CH2-THF reductase (nmoles of product/min/mg protein)
1.21 ±0.07
0.22 ± 0.02
0.42 ± 0.07
Serine hydroxymethyl transferase (nmoles of product/min/mg protein)
10.8 ±0.6
17.1 ±9.7
18.5 ±0.7
Dihydrofolate reductase (nmoles of product/min/mg protein)
19.8 ± 1.3
4.1 ±0.7
0.74 ±0.17
Methionine synthase (nmoles of product/min/mg protein)
0.09 ±0.007
0.09 ±0.012
0.10 ± 0.008
Rate of formate oxidation (mg/kg/hr)
78
40
0
Source: Skrzydlewska (2003)
As shown in Table A-8, the normal hepatic THF levels of monkeys and humans are 1.5 and
1.75-fold lower than the levels in rats. Also, the levels of 10-formyl-THF-dehydrogenase levels are
2.67- and 3.83-fold lower in monkeys and humans, respectively, compared to the levels in rat liver,
which might cause an accumulation of formate in primates since there is decreased oxidation of
formate to CO2. Thus, primates oxidize formate less efficiently than rats fSkrzvdlewska. 20031.
Interaction of formaldehyde with cellular macromolecules in the URT
As mentioned earlier, it has been shown that "free" formaldehyde (i.e., the 0.1% of total
formaldehyde that does not exist in the form of methanediol) reacts with macromolecules (Abrams
and Kallen. 19761. However, it is unclear whether methanediol in certain hydrophobic matrices
(e.g., crossing biological membranes, etc.) could be converted to a more reactive form and available
to interact with cellular materials. Inhaled formaldehyde interacts at the portal of entry with the
nasal passages, and these interactions can be either noncovalent (reversible) or covalent
(irreversible).
Noncovalent interactions:
Formaldehyde is reversibly bound to GSH and THF in the cells forming the glutathione
hemithioacetal adduct or hydroxymethylglutathione (HMGSH) adductand 5, IO-CH2-THF adducts.
Levels of the cellular antioxidant glutathione are abundant in the cell ~5 mM with which
formaldehyde readily forms the hemiacetal adduct The dissociation constant for the hemiacetal
and CH2-THF adducts are approximately 1.5 mM fUotila and Koivusalo. 19741 and «30 |a,M,
respectively {Kallen, 1966 #119}. Based on in vitro experiments formaldehyde has been shown to
reversibly bind to human and rat nasal mucus, in particular the fraction containing albumin
(Bogdanffv et al.. 19871.
Covalent binding
Formaldehyde covalently binds to protein, DNA, DNA and proteins forming protein adducts,
DNA adducts, DNA-protein crosslinks (DPX or DPC), and DNA-DNA crosslinks (DDX). A
This document is a draft for review purposes only and does not constitute Agency policy.
A-29 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
complication that has been explored in some of these studies is that inhaled formaldehyde can also
be metabolized and incorporated into DNA and proteins via the 1C pool.
Protein adducts
Formaldehyde has been shown to bind to histones and chromatin forming N6-formyllysine
(Edrissi etal.. 2013) and a major source of this adduct has been shown to result from endogenous
formaldehyde. Further, in rats exposed to various inhalation concentrations of 13C-labeled
formaldehyde (0.9-11.2 mg/m3), a concentration-dependent increase in 13C-labeled N6-
formyllysine, which was distinguished from endogenous N6-formyllysine, was detectable in the
total proteins as well as in protein fractions from different cellular compartments (cytoplasmic,
membrane, and nuclear) of the respiratory epithelium fEdrissi etal.. 20131.
DNA-protein Crosslinks
Formaldehyde-induced DNA-protein crosslinking occurs predominantly between the
epsilon-amino groups of lysine, especially the N-terminus of histones, and exocyclic amino groups
of DNA (Lu etal.. 2008a). Several analytical methods including radiolabeled formaldehyde have
been used to evaluate DPX formation in experimental animals. Earlier experiments have shown
that inhalation of F344 rats to 2.46-36.93 mg/m3 of 14C-formaldehyde (6 hours/day, 2 days) caused
a significant increase in the radioactivity of interfacial (IF) DNA1, representing DPX, observed in
tissue homogenates from respiratory but not olfactory epithelium at > 7.38 mg/m3 (Casanova-
Schmitz and Heck. 1983). Formaldehyde-induced DPX levels have been shown to have
concentration-dependence in both monkeys (0.86 to 7.37 mg/m3) (Casanova etal.. 1991) and rats
(0.37-12.1 mg/m3) (Casanova etal.. 1994: Casanova etal.. 1989). In both rodents and monkeys
there was a nonlinear concentration-response for DPX formation, which has been attributed to
saturation of detoxification enzymes at high concentrations fCasanova etal.. 1991: Casanova etal..
19891. In monkeys, the DPX distribution pattern in the nasal passages following formaldehyde
inhalation was in the order of middle turbinates > anterior lateral wall/septum > maxillary sinuses
and lungs (Casanova et al.. 1991). which corresponded to the location and proliferative response.
In rats the DPX distribution pattern was in the order of lateral meatus > medial and posterior
meatus (Casanova etal.. 1994). which corresponded to the high and low tumor incidence sites in
the respiratory tract fMonticello etal.. 19891. This is possibly due to the differences in the anatomy
of nasal passages and breathing patterns of these two species.
Recently, Lai etal. (2016) developed a method that distinguishes deoxyguanosine-methyl-
cysteine (dG-Me-Cys), a DPC formed from exogenous formaldehyde from that formed from
endogenous formaldehyde (see Table A-9). In monkeys exposed to 7.4 mg/m3 of 13C-labeled
1 During a typical DNA extraction of tissue homogenates, the DNA separated into aqueous phase is termed aqueous
(AQ) DNA, while the DNA trapped in the protein precipitate from the interphase (between aqueous and organic
phases) was washed, treated with protein kinase and reextracted to get the interfacial DNA (IF DNA).
This document is a draft for review purposes only and does not constitute Agency policy.
A-30 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
formaldehyde for 2 days, both exogenous and endogenous DPCs were detectable, with the levels of
exogenous DPCs being 2.8-fold less than the endogenous DPC adducts. In contrast, only
endogenous DPCs were detectable in air-exposed monkeys. In rats, a higher dose of 18.5 mg/m3
formaldehyde exposed for 1, 2, or 4 days was tested. DPC levels in nasal tissues were detected and
were comparable for endogenous and exogenous formaldehyde among rats exposed 1 or 2 days,
but at 4 days, DPC levels from exogenous formaldehyde had increased 5-fold above those from
endogenous formaldehyde. Similarly, DPC levels from exogenous formaldehyde increased between
7 days and 28 days in rats exposed to 2.5 mg/m3.
Using in vitro studies, Yu et al. (2015b) have shown that DPX such as, dG-CI-h-cysteine or
dG-CH2-GSH can undergo hydrolytic degradation to give rise to hm-dG monoadducts under
physiological pH and temperature conditions. These results provide a mechanism which explains
why formaldehyde-induced DPX are removed within 12.5-24 hrs in cultured human epithelial cell
lines (Ouievrvn and Zhitkovich. 2000) and lymphoblasts (Craft etal.. 1987). However, the in vivo
studies by Lai etal. (2016) did not replicate this phenomenon. These more precise studies have
shown that in rats exposed to 2.5 mg/m3 labeled formaldehyde for 28 days, at 1-week
postexposure, 87% of the exogenous DPC were retained in the nasal tissues, suggesting a slow
repair of these bulky adducts. The potential implications of this for dose-response modeling are
discussed in Appendix B.2.2.
Table A-9. Summary of endogenous and exogenous DNA-protein crosslinks in
nasal tissues of rats following inhalation exposure of 13CD2-labeled
formaldehyde
Reference
and design
Exposure and analysis
Exposure
duration
CH20
conc.
Observations
Lai et al.
0 (air control) or 7.4 mg/m3 [13CD2]-CH20
from PFA by inhalation; 6 hrs./d; for 2 d;
whole-body exposure; nasal tissue
collected; DNA extracted with DNAzol
reagent, dG-Me-Cys purified on HPLC and
analyzed by nano-LC/ESI/MS-MS.
(mg/m3)
Endogenous
adducts
Exogenous
adducts
(2016):
Monkeys,
cynomolgus;
A/=4-6.
dG-Me-Cys/108 dG
2 d
0
3.59 ± 1.01
ND
2 d
7.4
3.76 ± 1.50
1.36 ±0.20
Lai et al.
0 (air control) or 18.5 mg/m3 [13CD2]-
CH20 from PFA by inhalation; 6 hrs./d; for
1,2, or 4 d; whole-body exposure; nasal
tissue collected; DNA extracted with
DNAzol reagent, dG-Me-Cys purified on
HPLC and analyzed by nano-LC/ESI/MS-
MS.
Exposure
Duration
(mg/m3)
Endogenous
adducts
Exogenous
adducts
(2016): Rats.
F344; A/=4-6.
dG-Me-Cys/108 dG
4 d
0
6.50 ±0.30
ND
1 d
18.5
4.42 ± 1.10
5.52 ±0.80
2 d
18.5
4.28 ±2.34
4.69 ± 1.76
4 d
18.5
3.67 ±0.80
18.18 ±
7.23
Lai et al.
Rats inhalation exposure to 2.5 mg/m3
CH20 for 7 or 28 days and allowed to
recover for 1 or 7 days PE. Nasal tissue
collected and DNA extracted at the given
time points and analyzed for dG-Me-Cys
adducts as above.
Exposure
Duration
(mg/m3)
Endogenous
adducts
Exogenous
adducts
(2016): Rats.
F344; A/=4-6.
dG-Me-Cys/108 dG
7 d
2.5
4.78 ± 0.64
0.96 ±0.17
28 d
2.5
4.51 ±1.48
2.46 ± 0.44
This document is a draft for review purposes only and does not constitute Agency policy.
A-31 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Supplemental Information for Formaldehyde—Inhalation
Reference
and design
Exposure and analysis
Exposure
duration
CH20
conc.
Observations
28d + ld PE
2.5
3.78 ±0.69
2.12 ± 1.00
28 d + 7d PE
2.5
3.51 ±0.16
2.14 ± 1.02
Abbreviations: PFA, paraformaldehyde; LC, liquid chromatography; MS, mass spectrometry; HPLC, high
performance liquid chromatography; CH20, formaldehyde; DPC, DNA-protein crosslinks; dG-Me-Cys,
deoxyguanosine-methyl-cysteine; PBMC, peripheral blood mononuclear cell; ESI, electron spray ionization; PE,
post-exposure.
Distinguishing covalent binding of formaldehyde from metabolic incorporation
Few studies from the same research group addressed the issues of differentiating covalently
bound (i.e., DPX formation) versus metabolically incorporated formaldehyde in rats exposed to
formaldehyde by inhalation (Casanova and Heck. 1987: Casanova-Schmitz etal.. 1984b: Casanova-
Schmitz and Heck. 19831.
Casanova-Schmitz et al., (1984b 1 used dual isotope labeling as a way to partially distinguish
between covalent binding (DPX formation) and metabolic incorporation of formaldehyde. In this
approach, male F344 rats were exposed to a mixture of 3H- and 14C-labeled formaldehyde for 6
hours at exposure concentrations ranging from 0.37-18.42 mg/m3, a day after exposure to
nonradioactive formaldehyde with the same dose range. The IF DNA was extracted from
respiratory and olfactory mucosa, and the 3H/14C ratios of different phases of DNA extraction (i.e.,
AQ DNA and IF DNA) were measured. It is important to note that formaldehyde loses the hydrogen
atom during oxidation reactions (i.e., metabolic incorporation), but not during covalent binding to
DNA. Therefore, the 3H/14C ratio in a sample that contains adducts and crosslinks should be higher
than in a sample that primarily contains DNA with metabolically incorporated formaldehyde.
CH20 (ppm) CH20 (ppm)
Figure A-7. Metabolic incorporation and covalent binding of formaldehyde in
rat respiratory tract. 3H/14C ratios in macromolecular extracts from rat
respiratory mucosa (A) and olfactory mucosa (B) following 6-hour exposure to 14C-
and 3H-labeled formaldehyde (0.3, 2, 6,10, and 15 ppm, corresponding to 0.37, 2.46,
7.38,12.3,18.42 mg/m3, respectively).
Source: Adapted from Casanova-Schmitz et al.(1984b)
This document is a draft for review purposes only and does not constitute Agency policy.
A-32 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
As seen in panel A of Figure A-7, Casanova-Schmitz etal. (1984b) report that IF DNA from
nasal respiratory mucosa has a significantly higher 3H/14C ratio (Y-axis) than the aqueous phase
(AQ) DNA, with a nonlinear dose response of IF DNA at exposure concentrations equal to or greater
than 2.46 mg/m3. These data suggest that IF DNA has significantly more 3H, a phenomenon likely
explained by additional 3H-formaldehyde molecules present as DPXs prior to DNA extraction.
These crosslinks were due to exogenous formaldehyde that could be attributed to DPX. The 3H/14C
ratio was linearly increased for the organic fraction, suggesting covalent binding of formaldehyde to
respiratory mucosa proteins. In contrast, olfactory mucosa did not show increased 3H/14C ratio in
the IF DNA or AQ DNA or proteins phase as a function of formaldehyde concentration (panel B,
Figure A-7). In total, these data suggest that the radiolabeling observed following formaldehyde
exposure in rats results from both covalent binding and metabolic incorporation in the nasal
mucosa, but not the olfactory mucosa (Casanova-Schmitz etal.. 1984b], The respiratory mucosa
from unexposed rats appears to contain 15% of DNA as IF DNA (Casanova-Schmitz and Heck.
19831. possibly as endogenous DPX.
DNA monoadducts
Another form of formaldehyde-induced covalent DNA modifications is hydoxymethyl-DNA
(hm-DNA) adducts or DNA monoadducts. Five studies conducted in one laboratory used 13CD2-
formaldehyde in experimental rats and monkeys coupled with an LC/MS approach to distinguish
hm-DNA adducts formed by endogenous and exogenous formaldehyde (Yu etal.. 2015b: Lu etal..
2011: Moeller etal.. 2011: Lu etal.. 2010). as summarized in Table A-10. In this method, hm-DNA
adducts formed by exogenous 13CD2-formaldehyde are distinguished from unlabelled endogenous
hm-DNA adducts based on the differences in their typical m/z ratio fLu etal.. 2012b! As shown in
Table A-10, both exogenous and endogenous N2-hydroxymethyl-deoxyguanosine (N2-hm-dG)
adducts were detected in nasal tissues of cynomologous monkeys exposed to 2.34 or 7.5 mg/m3
13CD2-formaldehyde for two days, and across several rat studies testing exposures ranging from
0.9-18.7 mg/m3 formaldehyde for several hours up to 28 days (Yu etal.. 2015a: Yu etal.. 2015b: Lu
etal.. 2011: Lu etal.. 2010). Notably, however, these studies demonstrate that the levels of
endogenous N2-hm-dG adducts were several folds higher than corresponding exogenous adducts in
nasal tissue.
While these studies provide the first insights into the relationship between endogenous and
exogenous DNA monoadducts, further study may help to clarify some remaining uncertainties. For
example, the potential involvement of different types of DNA monoadducts, as well as their specific
toxicodynamic roles (e.g., for cancer development), remain poorly understood. Of the studies which
used inhalation exposure to 13C-labeled formaldehyde, only Lu et al., (2010) quantified other adduct
types; interestingly, while the authors detected 13CD2-labeled N2-hm-dG adducts and dG-CH2-dG
crosslinks, they did not detect N6-hydroxymethyl-deoxyadenosine (N6-hm-dA) adducts in the nasal
epithelium of rats exposed for 1 or 5 days (12.3 mg/m3) to exogenous formaldehyde. However, the
This document is a draft for review purposes only and does not constitute Agency policy.
A-33 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for Formaldehyde—Inhalation
same group reported the formation of both N2-hm-dG (most of the tissues) and N6-hm-dA
monoadducts (only in bone marrow) in rats that were dosed by gavage with 13C-labeled methanol,
which is a precursor of formaldehyde fLu etal.. 2012bl. Similarly a different research group
reported that rats dosed subcutaneously with nitrosamines fWang etal.. 20071. which are
precursors to formaldehyde, and smokers fWang etal.. 20091 both exhibit N6-hm-dA monoadducts
in peripheral tissues. Thus, additional sensitive evaluations of dA monoadducts, particularly
following longer term formaldehyde exposure and preferably in humans, may be informative. Also
of interest, it is important to keep in mind that the experiments conducted to date involve
comparisons of endogenous adduct levels, which would represent steady-state formaldehyde levels
after having built up over time from the continuous presence of endogenous formaldehyde, to
exogenous adduct levels resulting from short-term and/ or episodic (e.g., 6 hr/day) exposures. As
an illustration, with exogenous exposure for 6-hr/day, multiple weeks or longer could be needed to
reach steady-state levels, and, even so, those levels could be roughly expected to be four-fold lower
than if a continuous (24 hrs/d) exogenous exposure occurred at the same concentration. The
recent study by Yu et al. (Yu etal.. 2015b) begins to address this, noting that "quasi-steady-state"
levels appear to be nearing after 6hr-day exposure to 2.46 mg/m3 formaldehyde for 28 days;
however, exogenous adducts were still substantially increased with 28 days, as compared to 21
days of exposure, and exogenous adducts reached «37% of endogenous adducts (1.05 versus 2.82
adducts/107 dG, in contrast to the «14% observed after 7 days of exposure) under this scenario.
Considering these data at 2.46 mg/m3, the comparability of endogenous versus exogenous adducts
relevant to lifetime exposure scenarios would be informed by additional studies incorporating a
range of experiments and formaldehyde concentrations that span short, episodic exposures to more
constant, long-term exposures.
Table A-10. Summary of endogenous and exogenous DNA monoadducts in
nasal tissue of monkeys and rats following inhalation exposure of 13CD2-
labeled formaldehyde
Reference
and design
Exposure and analysis3
Portal of
entry tissues
CH20
exposure
conc.
(mg/m3)
Observations
Endogenous
adducts
Exogenous
adducts
Moeller et al
(2011);
Monkeys,
cynomolgus;
n=3
2.34 or 7.5 mg/m3 [13CD2]-
CH20; 6 hours/d; for 2 days
(whole-body exposure);
sacrificed immediately after
exposure; tissues collected.
Nasal
maxilloturbinates
N2-hm-dG/107dG
2.34
2.50 ±0.40
0.26 ± 0.04
7.5
2.05 ±0.54
0.41 ± 0.05
This document is a draft for review purposes only and does not constitute Agency policy.
A-34 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and design
Exposure and analysis3
Portal of
entry tissues
CH20
exposure
conc.
(mg/m3)
Observations
Endogenous
adducts
Exogenous
adducts
Yu et al.,
(2015b):
Monkeys,
cynomolgus;
n=4
0 (air control), 2.4 or 7.5
mg/m3 [13CD2]-CH20
generated from [13CD2]PFA;
nose-only exposure; 6
hours/d for 2 consecutive
days; Sacrificed immediately
after exposure;
maxilloturbinates (Animal #1)
and all other nasal tissues
(Animal #2) were collected.
Nasal
maxilloturbinates
2.4
2.50 ± 0.44
0.26 ± 0.04
7.5
2.05 ±0.54
0.41 ± 0.05
Nasal dorsal
mucosa
0
3.81 ± 1.19
ND
7.5
3.62 ± 1.28
0.40 ± 0.07
Nasal
nasopharynx
0
3.48 ±0.53
ND
7.5
3.62 ± 1.34
0.33 ±0.10
Nasal septum
0
3.75 ±0.32
ND
7.5
3.56 ±0.69
0.39 ±0.15
Nasal anterior
maxillary
0
4.21 ±0.53
ND
7.5
3.80 ±0.91
0.34 ±0.12
Nasal posterior
maxillary
0
3.95 ±0.74
ND
7.5
3.46 ± 1.05
0.36 ±0.16
Trachea carina
0
2.69 ±0.95
ND
7.5
2.33 ± 1.12
ND
Trachea proximal
0
2.35 ± 1.05
ND
7.5
2.35 ± 1.05
ND
Lu et al.,
(2010); Rats,
Fisher; Male,
n=5-8
12.28 mg/m3 [13CD2]-CH20
generated from [13CD2]PFA; 6
hours/day, 1 or 5 days; nose-
only exposure;
Sacrificed immediately after
exposure; tissues collected.
Nasal tissueb'c
Exposure
duration
Endogenous
adducts
Exogenous
adducts
N2-hm-dG/107 dG
1-day
2.63 ±0.73
1.28 ±0.49
5-days
2.84 ± 1.13
2.43 ± 0.78
N6-hm-dA/107 dA
1-days
3.95 ±0.26
ND
5-days
3.61 ±0.95
ND
dG-CH2-dG/107 dG
1-day
0.17 ±0.05
0.14 ± 0.06
5-days
0.18 ±0.06
0.26 ± 0.07
Lu et al.
(2011); Rats,
Fischer; n=5-6
[13CD2]-CH20 from [13CD2]PFA;
6 hours, nose-only exposure;
Sacrificed immediately after
exposure; tissue collected.
Nasal tissue
Exposure
concentration
(mg/m3)
Endogenous
adducts
Exogenous
adducts
N2-hm-dG adducts/107 dG
0.9 ±0.25
3.62 ± 1.33
0.039 ± 0.019
2.5 ±0.12
6.09 ±3.03
0.19 ±0.08
7.1 ±0.62
5.51 ± 1.06
1.04 ± 0.24
11.2 ± 2.71
3.41 ±0.46
2.03 ± 0.43
18.7 ± 2.58
4.24 ±0.92
11.15 ±3.01
This document is a draft for review purposes only and does not constitute Agency policy.
A-35 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
Reference
and design
Exposure and analysis3
Portal of
entry tissues
CH20
exposure
conc.
(mg/m3)
Observations
Endogenous
adducts
Exogenous
adducts
Yu et al.
(2015); Rats,
Fischer, male;
n=8-9
0 (air control) or 2.46 mg/m3
[13CD2]-CH20 from [13CD2]PFA;
nose-only exposure; 6
hours/d for 7,14, 21, or 28
consecutive days;
postexposure recovery for 6,
24, 72, and 168 hours.
Sacrificed immediately after
exposure at indicated time
points; tissues collected.
Nasal epithelium
Exposure
duration
Endogenous
adducts
Exogenous
adducts
N2-hm-dG/107 dG
Air control
2.84 ±0.54
ND
7 days
2.51 ±0.63
0.35 ±0.17
14 days
3.09 ±0.98
0.84 ±0.17
21 days
3.34 ± 1.06
0.95 ±0.11
28 days
2.82 ±0.76
1.05 ±0.16
6 hours PE
2.80 ±0.58
0.83 ±0.33
24 hours PE
2.98 ±0.70
0.80 ± 0.46
72 hours PE
2.99 ±0.63
0.63 ±0.12
168 hours PE
2.78 ±0.48
0.67 ± 0.20
aTissue DNA was extracted, reduced with sodium cyanogen borohydride (NaCNBH3), digested and analyzed by
nano-UPLC-MS/MS.
bNasal respiratory epithelium from the right and left sides of the nose and the septum.
Exogenous N6-hmdA adducts were not detected in any tissues; exogenous N2-hm-dG and dG-dG crosslinks were
detected only in nasal tissues.
Abbreviations: CH20, formaldehyde; D2, deuterium; MS, mass spectrometry; PE, postexposure; PFA,
paraformaldehyde; ND, not detected; N2-hm-dG, N2-hydroxymethyl-deoxyguanine; N6-hm-dA, N6-hydroxymethyl-
deoxyadenine; dG-CH2-dG, dG-dG crosslinks; UPLC, ultra-pressure liquid chromatography.
Unknown contribution of potential interactions with other nasal mucosa elements
Formaldehyde is likely to interact with other components of the nasal mucosa depending on
the concentration and duration of exposure. A small amount of inhaled formaldehyde, converted
predominantly to methanediol, is expected to penetrate the epithelial cell layer and react with the
basement membrane or with constituents of the lamina propria, including components of the
connective tissue/extracellular space, mucus gland components, lymphoid components, and
vascular components. Andersen etal. f20081 examined the gene expression in different tissue
compartments of male F344 rats exposed to formaldehyde concentrations ranging from 0.9-18.5
mg/m3 by inhalation exposure. They reported that at low concentrations (0.9-2.5 mg/m3)
formaldehyde is likely to react with the extracellular components of the cells at or near the cell
membrane, while at higher doses (7.5-18.5 mg/m3) responses are observed in both extracellular
and intracellular sites involving more genes in the response. The gene expression data from this
study suggests the possibility for a potential interaction of formaldehyde with other nasal mucosa
components.
Removal of inhaled formaldehyde from the POE
The main processes for removing inhaled formaldehyde from the URT involve clearance in
the mucus and metabolism to formic acid. Formic acid can enter the 1C pool and may either be
oxidized to C02 or incorporated metabolically into nucleic acids and proteins carrying the 1C units
This document is a draft for review purposes only and does not constitute Agency policy.
A-36 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
through THF derivatives. Formate can also be absorbed into circulation, reach the kidneys, and be
excreted in urine.
Summary of penetration, metabolism and removal of inhaled formaldehyde within the URT
tissue
In summary, as inhaled formaldehyde enters the URT it interacts with the mucociliary
apparatus which is the first line of defense. In nasal mucus, most of the formaldehyde is rapidly
converted to methanediol (*99.9%) and a minor fraction remains as free formaldehyde (*0.1%).
Inhaled formaldehyde induces mucostasis and ciliastasis in rat nasal mucociliary apparatus
extending from the anterior to posterior regions of nasal cavity depending on the concentration and
duration of exposure fMorgan et al.. 1986al However, as previously noted, uncertainties remain
regarding the pattern of induced mucostasis, or the complete lack thereof, at low levels of
formaldehyde exposure. Methanediol is assumed to be better able to penetrate the tissues, while
free formaldehyde reacts with the macromolecules. It is assumed that the equilibrium is rapid,
hence that the methanediol:free formaldehyde equilibrium ratio is maintained (Fox etal.. 1985).
However, uncertainties remain regarding the net impact of the transition of inhaled formaldehyde
from the mucociliary layer to the underlying epithelium due to the presence of endogenous
formaldehyde, which is a component of normal cellular metabolism. In the URT, formaldehyde is
predominantly metabolized by glutathione-dependent class III alcohol dehydrogenase (ADH3) and
by a minor pathway involving aldehyde dehydrogenase 2 (ALDH2) to formate. Formate can either
enter the one-carbon pool leading to protein and nucleic acid synthesis, or is further metabolized to
CO2 and eliminated in expired air or excreted in urine unchanged.
Formaldehyde can interact with macromolecules either by noncovalently binding to GSH,
THF, or albumin in nasal mucus or covalently forming DPX, DDX, hm-DNA adducts, or protein
adducts. In rats and monkeys, formaldehyde exposure results in a concentration-dependent
increase in DPX. Metabolic incorporation studies with 14C-formaldehyde have shown both covalent
binding and metabolic incorporation in nasal tissues (Casanova and Heck. 1987: Casanova-Schmitz
etal.. 1984b). Distribution patterns in the nasal passages correspond to the tumor incidence
locations in rats and to proliferative response patterns in both rats and monkeys. Hence, DPX has
been used as a surrogate biomarker of exposure for risk assessment. Inhaled formaldehyde
induces a concentration-dependent increase in N2-hm-dG adducts in the nasal passages of monkeys
and rats. Recently, analytical methods have been developed that can distinguish N2-hm-dG adducts
formed from exogenous sources from those formed from endogenous sources. Notably,
endogenous N2-hm-dG adduct levels are much higher than exogenous monoadduct levels in
animals, because formaldehyde is known to be produced continuously during normal cellular
metabolism. It has been suggested that N2-hm-dG adducts could be used as a marker of exposure in
risk assessment. However, this use might be compromised by several methodological issues in the
adduct isolation and analysis.
This document is a draft for review purposes only and does not constitute Agency policy.
A-37 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
A.2.4. Modifying Factors and Specific Uncertainties Regarding the Toxicokinetics of Inhaled
Formaldehyde Within the POE
Many factors could influence the uptake and removal of inhaled formaldehyde at the POE.
Distribution and tissue penetration of inhaled formaldehyde could both be significantly modified as
a result of changes in environmental factors or tissue alterations induced by prolonged exposure.
Similarly metabolic detoxification of formaldehyde and clearance from the URT are dependent
upon a number of cofactors and proteins that may be modified by changes to the environment or by
prolonged exposure. Finally modeling indicates that endogenous formaldehyde has the potential
to impact on the toxicokinetics of inhaled formaldehyde. This section will not include a description
of every potential modifying factor, but will attempt to highlight those interpreted to be most
important or controversial, particularly those that may be essential to interpreting differences
between experimental animals and humans.
Adjustments to account for reflex bradypnea in rodent studies
Reflex bradypnea (RB) is a protective reflex that allows rodents—but not humans—to
significantly reduce their inhalation exposures to URT irritants such as formaldehyde. When an
irritating concentration of formaldehyde triggers RB via the trigeminal nerve, rodents have an
immediate decrease in respiratory rate and minute volume, and thus a marked decrease in
formaldehyde exposure. Their RB persists until the exposure ends although the strength of the
response in the initial minutes after exposure begins can be much stronger than later in the
exposure. Kane and Alerie (19771 showed a maximal response in naive mice of 13.7% decreased
respiration rate from exposure to 0.55 ppm formaldehyde. This increased slightly to 15.6% in mice
preexposed for 3 days. Consequently, a rodent study may not be health protective for humans
unless the chamber concentrations or minute volume are adjusted to account for the rodents'
reduced formaldehyde exposure. However, existing models and dose-response analyses have not
accounted for this effect.
Unfortunately, it is not known if or when rodents develop a tolerance to formaldehyde and
resume normal breathing. Considering that Chang and Barrow (Chang and Barrow. 19841 reported
that F-344 rats experienced RB throughout 10 days of formaldehyde exposure, it may be
appropriate to adjust short-term rodent exposure concentrations to make them health protective
for humans. Because a long-term RB study has never been performed for formaldehyde or any
other URT irritant, there is no way of knowing whether similar adjustment is warranted for
subchronic and/or chronic rodent studies. This is a significant data gap.
Modification due to effects of exposure on nasal mucosa function
Several events reported to occur after inhalation exposure to formaldehyde have the
potential to modify the toxicokinetics of formaldehyde in the URT during subsequent exposure
scenarios. Important among these factors are dynamic tissue modeling changes in mucociliary
This document is a draft for review purposes only and does not constitute Agency policy.
A-38 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
clearance, reduction in minute volume, and changes in glutathione levels and glutathione-mediated
ADH3 activity.
Functional changes in the respiratory epithelium could have significant effects on the
subsequent uptake of inhaled formaldehyde. Squamous metaplasia, a tissue conversion that is an
adaptive response that occurs in nasal epithelium exposed to toxic levels of formaldehyde, has been
observed in rats exposed to >2.46 mg/m3 formaldehyde for longer than 18 months. This type of
dynamic tissue remodeling of nasal airways can affect formaldehyde dosimetry, as squamous
metaplastic tissue is known to absorb considerably less formaldehyde than other epithelial types
fKamata et al.. 19971. This is of critical concern for dosimetric modeling efforts, which typically rely
on results from simulations of acute, rather than prolonged, exposure. The highest flux levels of
formaldehyde in simulations of the rat nose in Kimbell et al. (2001b) are estimated in the region
just posterior to the nasal vestibule. A consequence of squamous metaplasia is to "push" the higher
levels of formaldehyde flux toward the more distal regions of the nose (Kimbell etal.. 1997).
Uncertainties in the modeling of formaldehyde dosimetry are presented by Subramaniam et al.
(2008) and are discussed in the PBPK Section (see Appendix B.2.2). A similar concern is raised
regarding the observation that exposure affects the integrity and/or function of the mucociliary
layer, as previously discussed (see Section A.2.3).
Exposure-induced changes to factors involved in the detoxification of formaldehyde could
also affect its toxicokinetics during a subsequent challenge. The enzyme ADH3 is central to the
metabolism of formaldehyde; however, exposure to formaldehyde in turn alters the activity of
ADH3-dependent critical metabolic pathways. For example, transcription of ADH3 correlates with
the proliferative states in human oral keratinocytes (Nilsson et al.. 2004: Hedberg etal.. 2000). In
rodent lung, an increase in ADH3 activity affects other ADH3 substrates involved in protein
modification and cell signaling fOue etal.. 20051. Other pathways of ADH3 include oxidation of
retinol and long-chain primary alcohols and reduction of S-nitrosoglutathione (GSNO). GSNO can
accelerate ADH3-mediated formaldehyde oxidation and, likewise, formaldehyde increases ADH3-
mediated GSNO reduction nearly 25-fold. Because GSNO is an endogenous bronchodilator and
reservoir of nitric oxide (NO) activity, ADH3-mediated reduction of GSNO can cause a deregulation
of NO (Reviewed in (Thompson et al.. 2010).
Similarly, glutathione is essential to detoxification of formaldehyde through the major
pathway. GSH is present in most cells at levels far in excess of formaldehyde. In humans, the
HMGSH levels are high since circulating GSH concentrations are «50 times higher than
formaldehyde (Sanghani etal.. 2000). It is estimated that ~50-80% of formaldehyde in animal cells
is reversibly bound to GSH (Uotila and Koivusalo. 1989) and to a minor extent bound reversibly to
tetrahydrofolate (Heck etal.. 1982). Inhaled formaldehyde is similarly expected to undergo
detoxification following reversible binding to GSH. Glutathione levels are unchanged in tissue
homogenates following acute exposures, but represent a possible adaptive response that may be
location-specific and changed with prolonged exposure. For example, repeated exposure to
This document is a draft for review purposes only and does not constitute Agency policy.
A-39 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
formaldehyde (18.45 mg/m3, 6 hrs/d for 9 days) did not affect either the GSH levels or the specific
activities of ADH3 and ALDH2 in the nasal mucosa F344 rats fCasanova-Schmitz etal.. 1984al.
Interfacial DNA levels can be increased by glutathione depletion. This was tested by Casanova and
Heck (1987) by exposing rats for 3 hours on two consecutive days to a range (1.11-12.3 mg/m3) of
formaldehyde by inhalation, on Day 1 to nonlabeled formaldehyde and on Day 2 to a mixture of [3H]
and [14C]-labeled formaldehyde. Two hours before the exposure on the second day the animals
were injected i.p. with 300 mg/kg phorone, a GSH depleting agent. The authors reported a 90-95%
decrease in GSH levels and significant decrease in metabolic incorporation in nasal respiratory and
olfactory mucosa and bone marrow of phorone-treated rats. In contrast, the 3H/14C ratios of IF DNA
were increased in a concentration-dependent manner for both phorone-treated and control groups
of rats, albeit the levels were slightly higher in phorone-treated rats compared to control rats.
Thus, depletion of GSH appeared to result in more unmetabolized formaldehyde available for
covalent binding (crosslink formation) following 3-hour exposure.
Specific uncertainties regarding the potential impact of endogenous formaldehyde
Since formaldehyde is produced through normal cellular metabolism, several uncertainties
exist which might impact the metabolism of exogenous formaldehyde in the body. This section
covers the sources of endogenous formaldehyde, comparisons about its concentration gradient, its
metabolism and reactivity, and the impact of inhaled formaldehyde on endogenous formaldehyde.
Sources of endogenous formaldehyde
Formaldehyde is endogenously produced through normal cellular metabolism from three
main sources. As detailed below and outlined in Fig. 5, these sources include: (1) enzymatic
reactions, (2) nonenzymatic reactions, and (3) as a metabolic byproduct of cellular metabolism of
xenobiotics (e.g., drugs, environmental contaminants) that enter the body.
(1) Enzymatic pathways that generate formaldehyde endogenously as a normal component
of cellular metabolism include four metabolic pathways: methylamine deamination, choline
oxidation, histone lysine demethylation, and amino acid metabolism (serine, glycine, methionine).
Formaldehyde can also be generated through endogenous generation from exogenous sources (e.g.,
methanol). These enzymatic sources are summarized in Figure A-8.
Methylamine is endogenously produced through amine catabolism, which upon
deamination carried out by the enzyme semicarbazide-sensitive amino oxidase (SSAO) gives rise to
formaldehyde. Choline oxidation is another endogenous metabolic process by which formaldehyde
is generated. Choline is converted to glycine through several intermediary steps (choline betaine
dimethylglycine (DMG) sarcosine glycine. The last two steps in this pathway are catalyzed
by dimethylglycine dehydrogenase (DMGDH) and sarcosine dehydrogenase (SDH), respectively,
using flavin adenine dinucleotide (FAD) as a cofactor. During these two steps the dehydrogenases
nonenzymatically condense tetrahydrofolate (THF) with formaldehyde generating 5,10-
methylene-THF (5,10-CH2-THF), also known as "active formaldehyde."
This document is a draft for review purposes only and does not constitute Agency policy.
A-40 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information for Formaldehyde—Inhalation
The other mechanism of endogenous formaldehyde production is through histone lysine
demethylation, which is carried out by two classes of enzymes near the nucleus in a cell. One is a
FAD-dependent amine oxidase, also known as lysine-specific demethylase 1 (LSD1/KDM1). The
other one belongs to the Jumonji C terminal (JmjC) domain-containing histone demethylase
(JHDM1/KDM2A). The LSD1 and JHDM1 enzymes act, respectively, on dimethyl lysine and
trimethyl lysine converting them to monomethyl- and dimethyl lysine with the liberation of
formaldehyde as an intermediary product (Shi et al.. 20041. Formaldehyde can also be generated
from methanol by either enzymatic or nonenzymatic pathways.
(2) Formaldehyde can also be formed nonenzymatically by the spontaneous reaction of
methanol with hydroxyl radicals, wherein intracellular hydrogen peroxide is converted to the
hydroxyl radical through the Fenton reaction fCederbaum and Oureshi. 19821. Another mechanism
of nonenzymatic production of formaldehyde is through lipid peroxidation of polyunsaturated fatty
acids (PUFA) (Shibamoto. 2006: Slater. 1984). It is known that a certain level of oxidative stress
and lipid peroxidation occurs in every individual, and these oxidative processes are likely to
contribute to endogenous formaldehyde production (Ozen etal.. 2008: Zararsiz et al.. 20061.
(3) Formaldehyde may also be produced intracellularly during microsomal cytochrome
P450 enzyme-catalyzed oxidative demethylation of N-, 0-, and S-methyl groups of xenobiotics
fATSDR. 20081 that enter the body through dietary, environmental, or medicinal exposures, as
shown in Figure A-8. Dhareshwar and Stella (2008) estimated that formaldehyde released from
prodrugs is ~2-100 mg. However, the authors point out that in humans with endogenous blood
levels of ~2-3 |ig/g of blood total formaldehyde (Heck etal.. 1985). the fraction of formaldehyde
released from xenobiotics may contribute a small fraction to the endogenous pool (Dhareshwar and
Stella. 20081.
This document is a draft for review purposes only and does not constitute Agency policy.
A-41 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Non-enzymatic mechanisms
PUFA
I
Consumption of fruits
& vegetables; alcohol
^ Oxidative stress
Methanol 'J/'
Lipid
peroxidation
"Jt
•OH<~
¦ H203
Enzymatic mechanisms
Steroid
biosynthesis
(Lanosterol)
Stress
I
H>Adrenaline
^Creatinine
Lecithin —
FORMALDEHYDE
Xenobiotic metabolism
(Oxidative demethylation of
W- O-, and S-metbyl groups)
Dichloromethane
N.N-dimethylaniline
NNK
Prodrugs
«—Methionine
Serine >. Glycine
THF ^
_5d r LUb I
Intermediary
metabolism
Methylamine 5'10jCH* thf
Sarcosine
THF
5,10-CIVTHF
Dirriethylglycine
Betaine
t
Choline
CI-pool*
Monornethyl lysine
HCHO
Carbinolamine
Dimethyl lysine
| hcho
Cationic amine Carbinolamine
t t
Dimethyl lysine Trimethyl lysine
JHDM1/KDM2A - LSD1/KDM1-
mediated mediated
Choline metabolism Histone lysine demethylation
Figure A-8. Endogenous and dietary sources of formaldehyde production.
Formaldehyde is generated in the body through (a) Enzymatic mechanisms - involving (i) Steroid biosynthesis -
from lanosterol, (ii) Intermediary metabolism - from methylamine (Yu and Zuo, 1996), (iii) Choline metabolism
(Binzak et al., 2000), (iv) Stress - through adrenaline (Yu et al., 1997), (v) histone lysine demethylation (Shi et al.,
2004) and (vi) Methanol metabolism (enzymatic) (Skrzvdlewska, 2003); (b) Nonenzymatic mechanisms - (i)
Methanol oxidation (Cederbaum and Qureshi, 1982) (ii) Lipid Peroxidation of polyunsaturated fatty acids or
PUFA (Shibamoto, 2006) and (iii) Oxidative Stress (Slater, 1984); (c) Xenobiotic metabolism - demethylation of
chemicals (ATSDR, 2008) and prodrugs (Dhareshwar and Stella, 2008).
Abbreviations: DMG: dimethyl glycine; CI: one carbon; NNK: 4-(methylnitrosamino)-l-(3-pyridyl)-l-butanone; THF:
tetrahydrofolate; LSD1/KDM1, lysine (K)-specific demthylase 1; JHDM1/KDM2A, JumonjiC-domain containing
histone demthylase 1.
Enzymes: a, alcohol dehydrogenase-1 (ADH1) in primates and ADH1 and catalase in rodents; b, semicarbazole-
sensitive amine oxidase; c, serine hydroxymethyl transferase; d, sarcosine dehydrogenase; e, dirriethylglycine
dehydrogenase.
1 The presence of comparatively high levels of endogenous formaldehyde in cells of the URT
2 presents an important uncertainty to evaluating the toxicokinetics of inhaled formaldehyde. Once
3 inhaled formaldehyde interacts with aqueous matrices such as mucus and is hydrated, the
4 biochemical interactions of inhaled formaldehyde and endogenous formaldehyde are assumed to be
5 very similar, given that there are no differences in chemical structure. However, other than in the
6 nucleus (i.e., the experiments detailing DNA adducts), no data are available to inform where and to
7 what extent endogenous and exogenous formaldehyde may be available to participate in these
8 reactions.
9 Although much is unknown regarding the impact of endogenous formaldehyde on the
10 formaldehyde uptake and metabolism as outlined in the sections above, uncertainties relevant to
This document is a draft for review purposes only and does not constitute Agency policy.
A-42 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
interpreting the potential for biological differences between inhaled formaldehyde and endogenous
formaldehyde are important to specify. Several of these uncertainties, which are essential to
consider when comparing the distribution and macromolecular binding of endogenous
formaldehyde versus inhaled formaldehyde, are outlined below.
Comparisons regarding the concentration gradient of endogenous formaldehyde
Endogenous formaldehyde is known to be produced within all cells of the URT. The specific
levels of endogenous formaldehyde within each type of cell, or even within the various components
of the nasal tissue (e.g., the respiratory mucosa lining the maxilloturbinates; the squamous
epithelium lining the luminal surface of the nasal vestibule), are likely to vary across individuals
and have not been experimentally defined. However, there is likely to be a general level (for which
estimates have been calculated) that could be applied homogenously across the URT tissue. With
formaldehyde inhalation, it does not appear that the general (endogenous) levels of formaldehyde
in the entire nasal mucosa are significantly altered (e.g.Heck etal.. 1983: Heck etal.. 1982). A
concern is raised when interpreting observed changes in the levels or macromolecular binding of
endogenous formaldehyde, as compared to those caused by inhaled formaldehyde. Specifically, a
consideration of the tissue region assayed needs to be incorporated. While endogenous
formaldehyde is produced within all regions of the nasal mucosa, uptake of inhaled formaldehyde
occurs at specific anatomic locations, primarily the squamous epithelium and respiratory mucosa in
anterior regions of the nose. Thus, comparisons of endogenous levels (or effects) in homogenates
containing isolates where all components are "target" tissues versus inhaled formaldehyde levels
(or effects) in homogenates containing both "target" and "nontarget" (e.g., olfactory epithelium)
isolates are difficult to interpret. Notably, the comparisons involving N2-hm-dG DNA adducts (Lu_et
al.. 2011: Moeller etal.. 2011: Lu etal.. 2010) addressed this concern. These authors compared
isolates of nasal respiratory mucosa and observed that dose-dependent increases in N2-hm-dG
adducts due to short-term, exogenous exposure do not reach the level of N2-hm-dG adducts due to
endogenous formaldehyde until exposure to >11 mg/m3 formaldehyde (Lu etal.. 2011): relatedly,
low levels of dG-CH2-dG adducts appeared to be higher with exogenous exposure to 12.3 mg/m3
formaldehyde for 5 days, as compared to adducts caused by endogenous formaldehyde (Lu etal..
20101. Similarly, the measurements by Heck et al. (1983; 19821 also appeared to quantify these
effects based on isolated respiratory mucosa.
A related concern, based on the decreasing concentration of inhaled formaldehyde reaching
deeper components of the nasal mucosa, is that exogenous formaldehyde is not expected to interact
to the same extent with all components (cellular and extracellular) of the nasal mucosa. Rather,
these interactions are highly enriched in the epithelial cells and associated cellular/extracellular
components along the apical surface of the respiratory mucosa. This is assumed to be in contrast
with endogenous formaldehyde, which is present (possibly at comparable levels) inside all cells of
the nasal mucosa. Although the respiratory epithelium would be expected to comprise the majority
of the cellular makeup of the isolated mucosa, contributions from cells in the lamina propria to
This document is a draft for review purposes only and does not constitute Agency policy.
A-43 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
measured levels and effects of endogenous formaldehyde would be expected to far outweigh those
same contributions attributable to exogenous exposure. Thus, this introduces an uncertain amount
of inequality to comparisons of the relative contributions of exogenous and endogenous
formaldehyde to macromolecular binding. It also highlights an important characteristic of the
levels of exogenous and endogenous formaldehyde in tissue isolates; namely that these levels do
not necessarily reflect, nor even approximate, the comparative levels in the target cells. However, it
would be methodologically arduous to isolate select portion(s) of the respiratory mucosa for
comparison, and as such, it does not appear that any studies have done so.
Comparisons regarding metabolism and reactivity of endogenous formaldehyde
As compared to exogenous formaldehyde, for which it is unknown how quickly it may be
detoxified by the normal cellular machinery, the production and subsequent detoxification of
endogenous formaldehyde appears to be kept under strict control. As mentioned earlier, the
majority of endogenous formaldehyde is reversibly bound to GSH at any time (Sanghani etal..
20001.
The regulation of endogenous formaldehyde appears to be imperfect, given the presence of
endogenous N2-HOCH2-dG (dG) adducts (Swenbergetal.. 20111. The endogenous adductlevels
reported by Swenberg et al. f20111 are about the same as the exogenous levels that would result
from a single 6-hour exposure to «10 ppm formaldehyde. Given that endogenous formaldehyde is
present continuously, the equivalent continuous exposure to exogenous formaldehyde that would
result in the same dG levels must be somewhat less than 10 ppm, perhaps 1 or 2 ppm (i.e., a
continuous exposure to 2 ppm could produce the same dG levels as a single, 6-hour exposure to
10 ppm; a much more detailed pharmacokinetic analysis would be required to exactly determine
the exact equivalent exposure). Toxicokinetic models that are calibrated or matched with
formaldehyde-induced DPX data and use the DNA-binding constant determined in vitro by Heck
and Keller (1988) can be used with reasonable reliability to predict induced tissue levels of
formaldehyde in the rat nose from exogenous exposure. For example, Georgieva et al. (2003)
predict an exogenous level in nasal tissue of around 17 [J.M from a 6-ppm exposure. Heck et al.
(1982) reported a total endogenous level in rat nasal tissue of 12.6 |ig/g or 420 [iM. But as
described just above, the dG adducts from endogenous formaldehyde correspond to an exposure of
less than 10 ppm, though the total amount of endogenous formaldehyde is over 20-times higher.
Hence, much, but not all, of the endogenous formaldehyde (measured by Heck et al., (1982)) must
be bound or sequestered in a way that reduces its ability to react with DNA, in comparison with
exogenous formaldehyde.
Impact of inhaled formaldehyde on the function of endogenous formaldehyde
Although formaldehyde inhalation does not appear to result in a measurable change in the
total level of formaldehyde in the nasal tissue of rats fHeck etal.. 19821. it has yet to be determined
whether exposure results in any changes to the normal functions of endogenous formaldehyde. For
This document is a draft for review purposes only and does not constitute Agency policy.
A-44 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
example, in the study by Lu et al., (20111. rats exposed to 13C-formaldehyde showed a
concentration-dependent increase in the exogenous hm-dG adduct levels, and the corresponding
endogenous N2-hm-dG adduct levels were highly variable at different exposure concentrations in
the nasal tissues. In addition to the potential "compartmentalization" differences mentioned above,
the endogenous DNA adduct levels, reflective of endogenous formaldehyde, do not appear to be
static. Possible effects of exogenous formaldehyde exposure on metabolism and distribution
processes of endogenous formaldehyde cannot be conclusively ruled out However, no appreciable
changes in the number of adducts formed as a result of interactions of endogenous formaldehyde
with cellular constituents have been noted, even in the presence of formaldehyde exposure fe.g.. Yu
et al.. 2015b],
Summary of potential modifying factors and specific uncertainties
The toxicokinetics of formaldehyde may be influenced by certain formaldehyde-related
effects, such as mucociliary clearance (Morgan etal.. 1983). reflex bradypnea (rodents only) and
reduction in minute volume (Chang etal.. 1983: Chang etal.. 1981). and dynamic tissue remodeling
fKamata et al.. 19971. which have the potential to modulate formaldehyde uptake and clearance.
For example, during repeated inhalation exposure to formaldehyde, mice but not rats lower their
minute volume thereby restricting the intake of the gas f Chang etal.. 1983: Chang etal.. 19811.
which may impact dosimetric adjustment if extrapolated to humans. Exposure to formaldehyde can
also cause a perturbation of ADH3-dependent pathways involved in cell proliferation (Nilsson et al..
2004: Hedberg etal.. 2000). protein modification and cell signaling (Que etal.. 2005). GSNO
metabolism, and deregulation of nitric oxide-dependent pathways (Thompson et al.. 2010). In rats
exposed by inhalation to formaldehyde, a rapid GSH depletion can result in more free formaldehyde
available for covalent binding and lowering metabolic incorporation (Casanova and Heck. 1987).
A.2.5. Conclusions Regarding the Toxicokinetics of Inhaled Formaldehyde Within the POE
Within the POE, a majority of inhaled formaldehyde is rapidly retained in the URT of
humans and experimental animals, irrespective of species differences in the anatomy, physiology,
and breathing patterns. Based on formaldehyde's molecular and biochemical properties, it can
reasonably be inferred that total formaldehyde levels are not significantly affected by exogenous
exposure. Also, one can conclude that following inhalation, formaldehyde levels are successively
reduced as formaldehyde from the air penetrates through the various components of the nasal
mucosa. Formaldehyde levels are reduced through interactions with components of the mucus and
through mucociliary clearance; through reactions with cellular materials at the plasma membrane
of the respiratory epithelium; via interactions with glutathione (GSH) and other macromolecules in
the intracellular and extracellular space; through localized metabolism and conjugation reactions;
and through reversible interactions with intracellular materials. This results in the formation of a
gradient of formaldehyde across the tissue space, with the greatest formaldehyde concentration at
the apical surface of the mucosa, and the lowest levels of formaldehyde at deeper components of
This document is a draft for review purposes only and does not constitute Agency policy.
A-45 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
the tissue, such as the nasal associated lymphoid tissues (NALT) and blood vessels. In the URT,
formaldehyde is metabolized by cytosolic ADH3 (major) and mitochondrial ALDH2 (minor)
enzymes to formate which is further metabolized to CO2 and eliminated in expired air of enter the
1C pool leading metabolic incorporation, or excreted in urine unchanged. The toxicokinetics of
formaldehyde may be influenced by several modifying factors in the nasal passages, which should
be considered for dosimetric adjustment when extrapolating to humans since these factors may
impact risk assessment
A.2.6. Toxicokinetics of inhaled formaldehyde beyond the portal of entry
Consistent with the previously described concentration gradient of inhaled formaldehyde
within the POE, multiple studies report that very little inhaled formaldehyde reaches the
vasculature of the respiratory tract to allow for absorption into the systemic circulation. Similarly,
there is very little evidence that inhaled formaldehyde is distributed to tissues such as the bone
marrow, liver, or brain. Studies examining the potential for direct interactions of inhaled
formaldehyde with cellular macromolecules at distal sites have also not reported any evidence of
these effects, despite observing that endogenous formaldehyde elicits such effects. Although the
evidence is not entirely conclusive, and some uncertainties remain to be explored, the currently
available data support an overall conclusion that appreciable amounts of inhaled formaldehyde are
not distributed outside of the URT. Formaldehyde produced endogenously through enzymatic and
nonenzymatic mechanism as well as that produced by the demethylation of xenobiotics (ATSDR.
2008). may pose some uncertainties for the exogenous formaldehyde metabolism.
A.2.7. Levels of Endogenous and Inhaled Formaldehyde in Blood and Distal Tissues
Using the detection methods employed by Heck et al. (1982), two studies from the same
group reported endogenous levels of total formaldehyde in blood to be 2.61 ± 0.14 ng/g of blood in
unexposed human subjects (Heck etal.. 1985). 2.24 ± 0.07 and2.71± 0.29 ng/g of blood in control
F344 (Heck etal.. 1985) and SD rats (Kleinnijenhuis etal.. 2013). respectively, and 2.42 ± 0.09 ng/g
of blood in unexposed rhesus monkeys (Casanova etal.. 1988). providing relatively consistent
measurements across species with an average blood level of «2.5 ng/g («0.1 mM) (see Table A-ll).
Levels of endogenous formaldehyde higher than in blood were also detected in other distal tissues
of rats, although the nasal tissue contained the highest levels fHeck etal.. 19821. The blood
formaldehyde levels were not significantly changed when tested during exposure or shortly after
exposure to formaldehyde concentrations ranging from 2.3 to 7.4 mg/m3 across the three species,
with varying durations of exposure (Casanova etal.. 1988: Heck etal.. 1985). The lack of increase in
the blood formaldehyde levels could also be due to the metabolism of formaldehyde in human
erythrocytes, which are known to contain the formaldehyde metabolizing enzymes ADH3 (Uotila
andKoivusalo. 19871 andALDH2 flnoue etal.. 19791.
The tissue levels of endogenous formaldehyde determined experimentally by Heck et al.
fHeck etal.. 19821 may be highly uncertain. Campbell Tr etal. f20201 assessed these values to be
This document is a draft for review purposes only and does not constitute Agency policy.
A-46 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
20x lower based upon their modeling estimates and attributed this discrepancy to the potential for
the Heck et al. measurement methodology to overestimate tissue formaldehyde levels. This is
addressed again in A.2.12 in a discussion of model derived estimates of the effects of endogenous
formaldehyde on formaldehyde dosimetry.
EPA notes that while these data indicate that inhaled formaldehyde is not absorbed into the
systemic circulation, a rough bounding calculation based on the human data indicates that the Heck
et al. (19851 experiment lacks the sensitivity needed to reach this conclusion. This bounding
calculation assumes that the 2.3 mg/m3 of inhaled formaldehyde completely mixes with the blood,
and because of its high solubility, it has a volume of distribution equal to that of all body water
(0.57 L/kg of body-weight; fGuvton. 19911. Using these parameters, the Heck et al (1985)
experiment is estimated to result in an increased blood formaldehyde concentration of 0.016 M-g/g2.
This quantity is one-half the experimental error of 0.03 |ig/mL. Hence, even if all of the 2.3 mg/m3
of inhaled formaldehyde completely mixes with the blood, under the experimental protocol above
for the human exposure, formaldehyde blood concentration would increase by 0.016 ng/g, a
quantity that cannot be detected by the Heck et al. (19851 experiment.3 Moreover, this quantity is
two orders of magnitude lower than the endogenous blood levels. Hence, these results are
consistent with a lack of 14C radiolabel increases in the plasma of rats exposed to 14C formaldehyde
fHeck etal.. 19831. as well as a lack of increase in total formaldehyde calculated following exposure
of rats to 13C formaldehyde (Kleinnijenhuis etal.. 2013). Altogether, the data argue that the amount
of inhaled formaldehyde absorbed into the blood is not likely to be significant, even if one assumes
that only 5% of the endogenous formaldehyde in blood is not sequestered.
A similar trend was observed in distal tissues. Heck et al. (1983) exposed rats to a range of
14C-formaldehyde concentrations (6.14-29.48 mg/m3 for 6 hours), and observed that the ratio of
tissue distribution relative to plasma radioactivity (|imole equivalents/g tissue) was not correlated
with the exposure concentration, except in the esophagus fHeck etal.. 19831. Mucociliary transport
from the nose and trachea may have led to these relatively higher esophageal levels. Overall, these
data also indicate that tissue distribution of formaldehyde levels were independent of the exposure
concentration and duration of exposure.
Overall, the published data demonstrate no significant increase in formaldehyde levels in
blood following formaldehyde inhalation. These data also report no significant differences in tissue
and blood formaldehyde levels between preexposed and naive animals. Such observations were
obtained from short-term experimental animal studies based on 14C-radiolabeling by GC-MS. The
2Heck et al. (1985) air concentration = 1.9 ppm = 1.9*1.23 mg/m3 = 2.34 mg/m3; t = 40/60 h; Inhalation Rate = 10-15
cubic m/day. Assuming 10 m3/24 hrs, we get 10/24 m3/h. Formaldehyde inhaled = 1.9 x 1.23 x (10/24) x 40/60 h =
0.649 mg. Body water = 40 kg for a 70-kg man (Guyton, 1991); concentration of HCHO = HCHO inhaled/body
water in mg/kg = 0.649/40 = 0.0162 mg/kg or ]ug/g.
3Even if one were to assume that formaldehyde stays only in the blood stream, this concentration increases to 0.12
Hg/g of blood, which is still within the experimental error.
This document is a draft for review purposes only and does not constitute Agency policy.
A-47 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 use of only this approach is problematic because there is no distinction as to whether the
2 formaldehyde measured in these studies is free, reversibly or irreversibly bound, measured as
3 formate, or part of the one-carbon pool. Nevertheless, taken together with the bounding
4 calculations and relative activity calculations described above, the lack of significance of exogenous
5 formaldehyde reaching distal tissues appears to hold even given the uncertainty.
Table A-ll. Summary of blood and tissue levels of total3 formaldehyde in
humans and experimental animals following inhalation exposure to
formaldehyde
Reference and
species
Exposure and analysis
Observations
Heck et al., (1985)
Human volunteers
Male, n=4; female, n=2
24-44 yrs old
2.34 ± 0.07 mg/m3 CH?0 (source not specified);
40 min exposure in a walk-in chamber; venous
blood collected before and after exposure; Total
CH20 measured as PFPH derivative by GC-
MS/SIM
Total0 formaldehyde (jug/g of blood)
Before exposure:
After exposure:
2.61 ±0.14
2.77 ±0.28
Casanova et al., (1988)
Monkeys, rhesus
Male, n=4;
200-250 g
7.37 mg/m3 CH20 (from PFA); 6 hrs/d, 4 d/wk, 4
wks; chamber inhalation; whole-body exposure;
pre- and postexposure blood collected; Total
CH20 measured as PFPH derivative by GC-
MS/SIM
Total0 formaldehyde (jug/g of blood)
Before exposure:
0 min. after exposure
40 min. after exposure:
2.42 ± 0.09
1.84 ±0.15
2.04 ± 0.40
Heck et al., (1985)
Rats, Fischer
Male, n=4,
232 ± 22 g
17.69 ± 2.95 ms/m3 CH?0 (source not
specified); 2-hours exposure: chamber
inhalation; nose-only; controls-no exposure;
Total CH20 measured as PFPH derivative by GC-
MS/SIM
Total0 formaldehyde (jug/g of blood
Before exposure:
After exposure:
2.24 ± 0.07
2.50 ± 0.07
Kleinnijenhuis et al.,
(2013)
Rats, Sprague Dawley
Male, n=10
12 wks-old
12.3 mg/m313CH20 (19.3% in aqueous solution:
source not specified): 6-hours exposure. Nose-
only chamber; Blood samples collected before,
during and after exposure; analyzed by HPLC-
MS/MS after derivatizing with 2,4-DNPH
Total0 formaldehyde (mg/L ofbloodh)
Before Exposure:
During Exposure (3 hrs):
During Exposure (6 hrs):
After Exposure (*6.2 hrs):
After Exposure (6.5 hrs):
2.71 ±0.29
2.63 ±1.12
2.01 ± 0.48
2.11 ±0.35
1.81 ±0.22
Heck et al., (1982)
Rats, Fischer
Male, n=8
200-250 g
7.37 mg/m313CH20 from PFA; 6 hours/d;
10-days exposure; chamber inhalation; CH20
measured as PFPH derivative by GC/MS
Rat tissue levels (mean ± SE) of total3 CH20
Unexposed
Exposed
Tissue
Hg/g
Mg/g
Nasal mucosa
12.6 ±2.7
11.7 ± 3.6
Liver
6.03 ±0.5
NR
Testes
8.40 ± 3.0
NR
Brain
2.91 ±0.42
NR
Heck et al., (1983)
Rats, Fischer
Male, n=3;
Two groups: (a) preexposure: (b) naive: On days
1-9: group a) received 18.42 mg/m3: CH?0 (from
PFA): whole body exposure, 6 hrs/dav: group b):
Animals
Exposed
Equivalents of 14C in tissues
(Mean ± SE)
no exposure. On day 10: groups a and b
naive rats
Nasal mucosa
Plasma
This document is a draft for review purposes only and does not constitute Agency policy.
A-48 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and
species
Exposure and analysis
Observations
180-250 g
received 14C-CH20 (from PFA) for 6 hours, nose-
only exposure. Tissue homogenates counted
with LSC for 14C02 trapped in ethanolamine in 2-
methoxy-ethanol counted for radioactivity
preexposed
2148 ±255
76 ± 11
2251± 306
79 ±7
Not significant
Not significant
Heck et al., (1983)
Rats, Fischer,
Male, n=12
Naive rats: dosed with 6.14,
12.28, 18.42 or 29.48
mg/m314C-CH20 (from
PFA); 6-hours nose-only;
sacrificed immediately after
exposure; tissue
homogenates counted with
LSC.
Tissue
(DPM/g
tissue)/(DPM/g
plasma)11
Tissue
(DPM/g
tissue)/(DPM/g
plasma)11
Esophagus
4.94 ± 1.23
Spleen
1.59 ±0.50
Kidney
3.12 ±0.47
Heart
1.09 ± 0.09
Liver
2.77 ±0.25
Brain
0.37 ±0.06
Intestine
2.64 ± 0.48
Testes
0.31 ±0.05
Lung
2.05 ±0.36
RBC
0.30 ± 0.08
includes free and reversibly bound formaldehvde(Heck et al., 1982).
Calculated concentration in blood and corrected for stability.
"Values (Meant SD) are ratios of concentrations (radioactivity) in tissues relative to plasma immediately after a 6-
hour exposure to 14C-formaldehyde averaged for four concentration groups (n = 12/concentration).
CH20, formaldehyde; GC, gas chromatography; LC, liquid chromatography; MS, mass spectrometry; HPLC-MS/MS,
high performance liquid chromatography/tandem mass spectroscopy; PFA, paraformaldehyde; SIM, selected ion
monitoring; DNPH, dinitrophenyl hydrazine; PFPH, pentafluorophenyl hydrazine; DPM, disintegrations per
minute; ND, not detected; UPLC, ultraperformance liquid chromatography; NaCNBH3, sodium cyanogen
borohydride.
1 Covalent binding of formaldehyde to macromolecules beyond POE
2 Formaldehyde has been shown to interact with the macromolecules in the blood or blood
3 cells, but not in other distal organs as described below.
4 Evidence of covalent binding of formaldehyde to blood proteins
5 Formaldehyde has also been shown to covalently bind to serum proteins such as the amino
6 acid valine in hemoglobin (Hb) forming N-methylvaline adducts in workers in plywood and
7 laminate factory workers with occupational exposure (Bono et al.. 20061. Also, with human serum
8 albumin (HSA) it forms formaldehyde-HSA complexes (Thrasher et al.. 19901. However, N6-
9 formyllysine, another formaldehyde-induced protein adduct that also occurs endogenously, was not
10 detectable in blood cells or in distal tissues (liver, lung and bone marrow) in rats exposed to
11 exogenous 13C-labeled formaldehyde fEdrissi etal.. 20131.
This document is a draft for review purposes only and does not constitute Agency policy.
A-49 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Evidence of DPX in the blood cells of formaldehyde exposed workers
DPXs have also been reported in the peripheral blood lymphocytes (PBLs) of formaldehyde-
exposed workers f Shah am et al.. 2 0 0 3: Shaham etal.. 1997: Shaham et al.. 19961. Shaham et al.
(1996) observed a statistically significant increase in DPX levels in PBLs compared to unexposed
subjects and reported a linear relationship between years of exposure and the amount of DPX.
Lack of experimental evidence of endogenous and exogenous DNA monoadducts and DNA-protein
crosslinks in blood and distal tissues
According to the available adduct studies, inhaled formaldehyde does not reach systemic
tissues in concentrations sufficient to elicit detectable interactions of formaldehyde with DNA. In
the bone marrow of monkeys (Moeller etal.. 2011). and in the bone marrow, liver, lung, spleen,
thymus, and blood of rats (Lu etal.. 2010). DNA monoadducts were formed by interactions with
endogenous formaldehyde, but adducts formed from exogenous formaldehyde were not found
(see Table A-12). It is important to note that Moeller et al. (2011) observed 6-8 times higher
endogenous N2-hm-dG adducts in the bone marrow compared to the nasal tissues of monkeys.
Although there were some limitations with the experimental methods, including a possible
overestimation of endogenous adducts due to reasons discussed (see Section A.2.3), the data
support a general lack of systemic distribution of inhaled formaldehyde.
As described for the POE tissues, efforts have been made to differentiate covalent binding
from metabolic incorporation in bone marrow. Male F344 rats were exposed to a mixture of 3H-
and 14C-labeled formaldehyde for 6 hours at 0.37-18.42 mg/m3 one day after exposure to
nonradioactive formaldehyde with the same exposure range (Casanova-Schmitz etal.. 1984b). The
authors extracted IF DNA from bone marrow (femur) and determined the 3H/14C ratios of different
phases of DNA (i.e., AQ DNA and IF DNA). As previously described, a sample that contains adducts
and crosslinks should be higher than in a sample that primarily contains metabolically incorporated
formaldehyde. In contrast to results in respiratory mucosa, bone marrow from the distal femur did
not show increased 3H/14C ratio in the IF DNA or AQ DNA or proteins phase as a function of
formaldehyde concentration (see Figure A-9). Therefore, the authors concluded that radiolabeled
metabolites of formaldehyde reached the distal site (femur bone marrow) and were subsequently
metabolically incorporated into macromolecules (see Figure A-7). In total, these data suggest that
the labeling of bone marrow macromolecules was likely due to metabolic incorporation rather than
due to covalent binding (Casanova-Schmitz etal.. 1984b).
Recently Lai etal. (2016) developed an ultrasensitive mass spectrometry method which
distinguishes unlabeled DPC from 13CD2-labeled DPCs induced respectively, from endogenous and
exogenous formaldehyde. The authors demonstrated that inhalation exposure of stable isotope
labeled (13CD2) formaldehyde to rats (18.45 mg/m3; 6 hours/day; 1-4 days) and monkeys (2.5
mg/m3; 6 hours/day; 2 days) induced exogenous DPCs in POE tissues such as nasal passages in
both species, but not in distal tissues, such as bone marrow and peripheral blood monocytes (rats
and monkeys) and liver (monkeys), although endogenous DPCs were detectable in all tissues
This document is a draft for review purposes only and does not constitute Agency policy.
A-50 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(see Table A-13). These observations further confirm the lack of experimental evidence of
formaldehyde distribution to distal tissues.
u
u
10
09
06
0 7
0 6
0 5
0 4
03
0 2
01
0
-B-
—fi-
- 0 Protein
b ,5...
U DNA(AQ)
RNA
(c) 8QNE MARROW
2 4 6 fi 10 I? 14 16
CH20 (ppm)
Figure A-9. 3H/14C ratios in macromolecular extracts from rat bone marrow
following 6-hour exposure to 14C- and 3H-labeled formaldehyde (0.3, 2, 6,10,
and 15 ppm, corresponding to 0.37, 2.46, 7.38,12.3,18.42 mg/m3, respectively).
Source: Adapted from Casanova-Schmitz et al.(1984b)
Table A-12. Summary of endogenous and exogenous DNA monoadducts in
distal tissues of monkeys and rats following inhalation exposure of 13CD2-
labeled formaldehyde
Reference
and design
Exposure and analysis3
CH20
conc.
Observations
Moeller et al.
(2011);
Monkeys,
cynomolgus;
n = 3
2.3 and 7.5 mg/m3 [13CD2]-CH20 from PFA; 6 hrs/d; for 2 d;
whole-body exposure; sacrificed immediately after exposure;
necropsied within 3 hrs; nasal mucosa and bone marrow
collected; tissue DNA extracted, reduced with NaCNBHs, digested
and analyzed by nano-UPLC/MS.
(mg/m3)
Bone marrow
Endogenous
adducts
Exogenous
adducts
DNA adducts/107dG
2.34
17.5 ±2.6
ND
7.5
12.4 ±3.6
ND
Yu et al.
(2015);
Monkeys,
cynomolgus;
0 (air control), 2.4 or 7.5
mg/m3 [13CD2]-CH20
from [13CD2]PFA; nose-
only exposure; 6 hrs/d
for 2 consecutive days;
Sacrificed immediately
after exposure; Tissue
DNA was extracted,
reduced with NaCNBHs,
digested and analyzed by
nano-UPLC-MS/MS
Distal tissue
N2-hm-dG/107 dG
Scrapped bone marrow (Animalttl)
2.4
17.5 ±2.6
ND
Scrapped bone marrow (Animal#2)
7.5
12.4 ±3.6
ND
Air control (Animal#2)
0
10.18 ± 1.35
ND
Scrapped bone marrow (Animal#2)
7.5
11.00 ±2.01
ND
Air control (Animal#2)
0
5.65 ± 2.12
ND
Saline extrusion bone marrow
(Animal#2)
7.5
4.41 ± 1.00
ND
This document is a draft for review purposes only and does not constitute Agency policy.
A-51 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and design
Exposure and analysis3
CH20
conc.
Observations
Air control (Animal#2)
0
3.64 ± 1.09
ND
White blood cells (Animal#2)
7.5
3.79 ± 1.19
ND
Lu et al. (2010);
Rats, Fisher;
Male, n=5-8
12.3 mg/m3 [13CD2]-CH20
from [13CD2]PFA; 6
hrs/day, 1 or 5 days;
nose-only exposure;
Sacrificed immediately
after exposure. Lung,
liver, spleen, bone
marrow, thymus, and
blood collected; tissue
DNA extracted, reduced
with NaCNBHs, digested
and analyzed by nano-
UPLC-MS/MS
Adduct
N2-hm-dG/107 dGa
Durations
1 day
5 days
Tissue
Endogenous
Exogenous
Endogenous
Exogenous
Lung
2.39 ± 0.16b
NDC
2.61 ±0.35
ND
Liver
2.66 ±0.53
ND
3.24 ±0.42
ND
Spleen
2.35 ±0.31
ND
2.35 ±0.59
ND
Bone marrow
1.05 ±0.14
ND
1.17 ±0.35
ND
Thymus
2.19 ±0.36
ND
1.99 ±0.30
ND
Bloodd
1.28 ±0.38
ND
1.10 ±0.28
ND
Adduct
N6-hm-dA/107 dAa
Durations
1 day
5 days
Distal Tissue
Endogenous
Exogenous
Endogenous
Exogenous
Lung
2.62 ±0.24
ND
2.47 ±0.55
ND
Liver
2.62 ±0.46
ND
2.87 ±0.65
ND
Spleen
1.85 ±0.19
ND
2.23 ±0.89
ND
Bone marrow
2.95 ±1.32
ND
2.99 ±0.08
ND
Thymus
2.98 ± 1.11
ND
2.48 ±0.11
ND
Bloodd
3.80 ±0.29
ND
3.66 ±0.78
ND
Adduct
dG-CH2-dG/107 dGa
Durations
1 day
5 days
Distal Tissue
Endogenous
Exogenous
Endogenous
Exogenous
Lung
0.20 ± 0.04®
ND
0.20 ±0.03
ND
Liver
0.18 ±0.05
ND
0.21 ±0.08
ND
Spleen
0.15 ±0.06
ND
0.16 ±0.08
ND
Bone marrow
0.09 ± 0.01
ND
0.11 ±0.03
ND
Thymus
0.10 ±0.03
ND
0.19 ±0.03
ND
Bloodd
0.12 ±0.09
ND
0.10 ±0.07
ND
0 (air control), 2.4 or 7.5
mg/m3 [13CD2]-CH20
from [13CD2]PFA; nose-
Formaldehyde
exposure duration
Rat bone marrow
Rat white blood cells
N2-OHMe-dG (adducts/107 dG)
This document is a draft for review purposes only and does not constitute Agency policy.
A-52 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and design
Exposure and analysis3
CH20
conc.
Observations
Yu et al.
(2015); Rats,
Fischer;
only exposure; 6 hrs/d
for 2 consecutive days;
Sacrificed immediately
after exposure; tissues
collected. Tissue DNA
was extracted, reduced
with NaCNBHs, digested
and analyzed by nano-
UPLC-MS/MS
Endogenous'
Exogenous
Endogenous'
Exogenous
Air control
3.58 ±0.99
ND
2.76 ±0.66
ND
7 days
3.37 ± 1.56
ND
2.62 ± 1.12
ND
14 days
2.72 ±1.36
ND
2.26 ±0.46
ND
21 days
2.44 ± 0.96
ND
2.40 ±0.47
ND
28 days
3.43 ± 2.20
0.34s
2.49 ±0.50
ND
28days + 6hrs PE
2.41 ±1.14
ND
2.97 ±0.58
ND
28 days + 24 hrs PE
4.67 ± 1.84
ND
2.57 ±0.58
ND
28 days + 72 hrs PE
5.55 ±0.76
ND
1.75 ±0.26
ND
28 days+168 hrs PE
2.78 ± 1.94
ND
2.61 ± 1.22
ND
Distal tissue
N2-OHMe-dG (adducts/107 dG)
Air control
28-day exposure
Endogenous
Exogenous
Endogenous
Exogenous
Thymus
0.78 ± 0.04
ND
0.63 ± 0.06
ND
TBLN
3.46 ± 1.24
ND
3.01 ±0.71
ND
Lymph nodes
2.99 ±0.85
ND
2.80 ± 1.38
ND
Trachea
3.18 ±0.72
ND
2.63 ±0.92
ND
Lung
2.29 ±0.24
ND
2.13 ±0.26
ND
Spleen
2.18 ±0.19
ND
1.83 ±0.25
ND
Kidneys
2.17 ±0.60
ND
1.99 ±0.09
ND
Liver
1.97 ±0.38
ND
1.80 ±0.02
ND
Brain
2.13 ±0.17
ND
2.35 ± 1.00
ND
aThe limit of detection for dG monoadducts, dA monoadducts, and dG-dG crosslinks was =240, -75, and =60 amol,
respectively.
bn = 4-5 tissues.
cNot detectable in 200 ng of DNA.
d60-100 ng of DNA was typically used for analysis of white blood cells isolated from blood.
en = 3.
fNo statistically significant difference was found using the 2-sided Dunnett's test (multiple comparisons with a
control).
gThe amount of exogenous N2-hm-dG adducts that was found in only 1 bone marrow sample analyzed by AB SCIEX
Triple Quad 6500.
Abbreviations: PFA, paraformaldehyde; UPLC, ultra-pressure liquid chromatography; MS, mass spectrometry; N2-
hm-dG, N2-hydroxymethyl-deoxyguanosine; N6-hm-dG, N6-hydroxymethyl-deoxyadenosine; dG-CH2-dG, dG-dG
crosslink; TBLN, tracheal bronchial lymph nodes; ND, not detected.
This document is a draft for review purposes only and does not constitute Agency policy.
A-53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
Table A-13. Summary of endogenous and exogenous DNA-protein crosslinks
in distal tissues of monkeys and rats following inhalation exposure of 13CD2-
labeled formaldehyde
Reference and
design
Exposure and analysis
Exposure
duration
CH20
conc.
Observations
Lai et al. (2016);
Monkeys,
cynomolgus;
0 (air control) or 7.4
mg/m3 [13CD2]-CH20 from
PFA; 6 hrs./d; for 2 d;
whole-body exposure;
PBMC, bone marrow and
liver collected; tissue DNA
extracted; dG-Me-Cys
purified on HPLC and
analyzed by nano-
LC/ESI/MS-MS.
Tissue
analyzed
(mg/m3)
Endogenous
adducts
Exogenous
adducts
dG-Me-Cys/108 dG
PBMC
2 d
0
1.34 ±0.25
ND
2 d
7.4
1.57 ±0.58
ND
Bone
marrow
2 d
0
2.30 ±0.30
ND
2 d
7.4
1.40 ± 0.46
ND
Liver
2 d
0
15.46 ± 1.98
ND
2 d
7.4
11.80 ±2.21
ND
Lai et al. (2016);
Rats, F344; A/=4-
6.
0 (air control) or 18.5
mg/m3 [13CD2]-CH20 from
PFA; 6 hrs./d; for 1,2, 4 d;
whole-body exposure;
PBMC, and bone marrow
collected; tissue DNA
extracted; dG-Me-Cys
purified on HPLC and
analyzed by nano-
LC/ESI/MS-MS.
Tissue
analyzed
Exposure
Duration
(mg/m3)
Endogenous
adducts
Exogenous
adducts
dG-Me-Cys/108 dG
PBMC
4 d
0
4.98 ±0.61
ND
1 d
18.5
3.26 ±0.73
ND
2 d
18.5
3.00 ± 0.98
ND
4 d
18.5
7.19 ± 1.73
ND
Bone
marrow
4 d
0
1.64 ± 0.49
ND
1 d
18.5
1.80 ± 0.47
ND
2 d
18.5
1.84 ±0.61
ND
4 d
18.5
1.58 ±0.38
ND
Abbreviations: PFA, paraformaldehyde; LC, liquid chromatography; MS, mass spectrometry; HPLC, high
performance liquid chromatography; CH20, formaldehyde; DPC, DNA-protein crosslinks; dG-Me-Cys,
deoxyguanosine-methyl-cysteine; PBMC, peripheral blood mononuclear cell; ESI, electron spray ionization.
A.2.8. Conjugation, Metabolism, and Speciation of Formaldehyde Outside the POE
Were inhaled formaldehyde to reach the blood or distal tissues, the same factors described
for POE effects, specifically those regarding metabolism, reactivity, and the role of endogenous
formaldehyde, would be relevant to other tissues. The majority of formaldehyde that reached these
systemic sites is expected to be in the form of methanediol which is not reactive with
macromolecules.
A.2.9. Elimination Pathways of Exogenous and Endogenous Formaldehyde
Elimination pathways of endogenous and exogenous pathways may not be different since
all tissues contain surplus GSH and NAD+. Endogenous formaldehyde is oxidized by ADH3 to
formate which is either eliminated as C02 in the exhaled breath or used in the cellular
macromolecular synthesis or excreted in urine. Similarly, the majority of inhaled formaldehyde is
metabolized in the URT by conversion to formate. Further, part of it may be metabolized to C02 or
utilized in the 1C pool. Since the available evidence does not show significant amounts of
This document is a draft for review purposes only and does not constitute Agency policy.
A-54 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
exogenous formaldehyde being transported into blood, the subsequent clearance of any exogenous
formaldehyde that does reach the blood should be similar to the handling of endogenous
formaldehyde.
Excretion of formaldehyde
Inhalation exposure to formaldehyde has not been shown to cause significant changes to the
tissue levels of formaldehyde in the nasal mucosa, the blood, or in the distal tissues. Thus, it is not
expected that formaldehyde and formaldehyde metabolite content in excretion products would be
altered by exposure. The data supporting this expectation are consistent in human and animal
studies.
Formate levels have been detected in both unexposed as well as formaldehyde-exposed
individuals. Gottschling et al. (1984) examined urinary formic acid levels of 35 veterinary medicine
students working in an anatomy lab before exposure and within two hours following 1-, 2-, or 3-wk
exposure to a mean formaldehyde concentration of <0.615 mg/m3. The authors did not observe
significant change in the pre- and postexposure levels of formic acid. Since co-exposure to
methanol may also contribute to the metabolism and excretion of formate, the fact that no
significant increase in urinary formate was seen even with that co-exposure further supports the
conclusion that the formaldehyde exposure does not significantly increase formate excretion.
Heck et al. (1983) determined the relative contributions of various elimination pathways in
F344 rats following inhalation exposure to 0.77 and 16.1 mg/m3 of 14C-formaldehyde. As shown in
Table A-14, the percentages of radioactivity in various fractions appear to be similar between the
two dose groups tested. Within 70 hours after a 6-hour formaldehyde exposure, nearly 40% of
radioactivity from inhaled 14C-formaldehyde appeared to be eliminated via expiration, probably as
14C02 (it should be recalled that nearly 100% of inhaled formaldehyde is taken up by the URT); and
«17 and 5% of radioactivity was eliminated in the urine and feces, respectively. Nearly 40% of
radioactivity remained in the carcass, which is presumably due to both covalent binding and
metabolic incorporation. Thus, in one form or another, 40% of the 14C from inhaled formaldehyde
is not eliminated, and is expected to persist in the tissue(s) for some time. Overall, the authors
concluded that, in rats, the relative elimination pathways for the remaining 60% of the 14C are
independent of exposure concentration, and followed the pattern of elimination in the order of
expired air > urine > feces.
Although not specifically demonstrated following exposure, assumptions based on the
known distribution and metabolism of formaldehyde and its detoxification products allow for
inferences to be drawn regarding how inhaled 14C reaches these elimination points. Approximately
one-third of inhaled formaldehyde is estimated to be removed in the URT mucus (Schlosser. 1999).
It is expected that the majority of this formaldehyde would be removed from the URT via
mucociliary clearance and excreted in urine in various forms. A large amount of inhaled
formaldehyde penetrating the mucociliary layer of the URT is metabolized in the nasal cavity, giving
This document is a draft for review purposes only and does not constitute Agency policy.
A-55 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
rise to formate, which can be excreted in urine. Part of this formate may also be further oxidized
and eliminated in the exhaled breath as CO2. Some formaldehyde is incorporated into the 1C pool.
Table A-14. Summary of excretion study following exposure to formaldehyde
by inhalation in rats
Reference and
species
Treatment and
analysis
Observations
Heck et al.
(1983)
Rats, Fischer
Male, n=4
210 g
0.77 and 16.1 mg/m3 HCHO for 6 hours;
rats sacrificed 70 hours after removal
from exposure chamber; tissues, urine,
feces collected; exhaled 14C02 trapped in
a solution of 5 M ethanolamine in 2-
methoxyethanol and % radioactivity
measured in LSC.
% Radioactivity (Mean ± SD) in various fractions
Source of radioactivity
Air borne CH20
0.77 mg/m3
16.1 mg/m3
Expired air:
39.4 ± 1.45
41.9 ±0.8
Urine:
17.6 ± 1.2
17.3 ±0.6
Feces:
4.2 ± 1.5
5.3 ± 1.3
Tissues3 and carcasses:
38.9 ± 1.2
35.2 ±0.5
aNasal mucosa, trachea, esophagus, lung, kidney, liver, intestine, spleen, heart, plasma, erythrocytes, brain, testes.
Levels of endogenous formaldehyde in exhaled human breath
Given that inhaled formaldehyde is almost entirely captured in the URT and is thus unlikely
to reach either the lower respiratory tract (LRT) or the systemic circulation to an appreciable
extent following exposure, and given that formaldehyde inhalation does not appreciably change
total formaldehyde levels in blood or any other tissue; it has been postulated that formaldehyde in
exhaled breath (measured in mouth-only exhalations) is expected to predominantly represent a
contribution from endogenous formaldehyde. However, it is important to understand the relative
amount of formaldehyde that is produced by the body and released in expired breath versus the
amount of formaldehyde in ambient air.
Table A-15 summarizes six studies that attempted to measure endogenous formaldehyde in
exhaled breath. All studies performed prior to 2010 are limited by their analytical methods, which
are subject to interference from other ions and isotopes that have the same m/z ratio [m/z = 31) as
formaldehyde (e.g., methanol, ethanol, and nitric oxide). Also, it was not possible to differentiate
between exogenous and endogenous formaldehyde in exhaled breath because the study subjects
inhaled room air containing formaldehyde (=11 |ig/m:i formaldehyde).
Table A-15. Measured levels of formaldehyde, methanol and ethanol in room
air and exhaled breath
Study
Analytical
Method
Sample
Formaldehyde c
(m/z 31) ng/m3
Methanol
Hg/m3
Ethanol
Hg/m3
Moser et al.
(2005 )a
PTR-MS
DL: NR
Room air:
"Negligible"
"Negligible"
"Negligible"
Exhaled breath:
5.24 (median)
198
NR
This document is a draft for review purposes only and does not constitute Agency policy.
A-56 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
Analytical
Method
Sample
Formaldehyde c
(m/z 31) ng/m3
Methanol
Hg/m3
Ethanol
Hg/m3
N = 344
1.49-89 (range)
Kushch et al.
(2008)
N = 370
PTR-MS
DL: NR
Room air:
NR
NR
NR
Exhaled breath:
6.39 (median,
nonsmokers)
5.53 (median, 81
smokers)
241 (median,
nonsmokers)
NR
Cap et al.
(2008)b
N = 34
SIFT-MS
DL: 3.68
Hg/m3 or
better
Room air:
11.79 ± 1.84
NR
NR
Exhaled breath:
2.46 (mean)
1.23 (median)
0-14.74 (range)
0 and 3.68 in 2 smokers
365 (mean)
232 (median)
125-2848
(range)
549 (mean)
101 (median)
33-12604
(range)
Turner et al.
(2008)
N = 5
SIFT-MS
DL: 6.14
Hg/m3 or
better
Room air:
ND
NR
NR
Exhaled breath:
ND
617 (mean)
549 (mean)
Wang et al.
(2008)
N = 3
SIFT-MS
DL: NR
Room air:
11.05 ±3.68
54 ± 11
124 ± 63
Exhaled breath:
6.51 (mean)
4.91-8.6 (range)
329 (mean)
185.46 (mean)
Riess et al.
(2010)
N~ 8
(nonsmokers)
N = 2
(smokers)
Acac
method
DL: <0.62
Hg/m3d
Charcoal
filtered air:
0
NR
NR
Exhaled breath:
<0.62 (nonsmokers), ND
<0.62 (2 smokers), ND
NR
NR
PTR-MS e
DL: =0.62
Hg/m3
Charcoal
filtered air:
0
NR
NR
Exhaled breath:
1.84 (mean; 0.86-2.82),
nonsmokers;
1.23 - 2.82, 2 smokers
NA
NA
aAuthors reported room air concentrations for 179 chemicals were "negligible." No smoker data were provided.
bSmoker data and formaldehyde ambient concentration provided by Dr. Spanel (personal communication).
cValues of formaldehyde in parts per billion (ppb) are converted as ng/m3 = ppb x 30 (m.w.)/24.45 or ppb x 1.23.
dThe acac method's limit of detection is 0.062 ng formaldehyde/m3, but the authors calculated a detection limit of
0.62 |Jg/m3 due to a slight periodically fluctuating background noise signal.
eAfter subtraction for methanol and NO product ions.
Abbreviations: DL = Detection Limit; NR = Not Reported; ND = Not Detected; NA = Not Applicable; PTR-MS = Proton
Transfer Reaction Mass Spectrometry; SIFT-MS -= Selected Ion Flow Tube Mass Spectrometry.
1 Riess et al. (20101. employed the acetyl acetone [acac) method4 to measure formaldehyde.
2 This method is superior to the PTR-MS method used in previous studies because it has a lower limit
3 of detection, exhibits no interference from other exhaled chemicals, and possesses the ability to
4 measure in dry or humid atmospheres. In addition, volunteers inhaled formaldehyde-free air. For
5 comparison, Riess et al. (2010) used both the acac method and the PTR-MS method and observed
4The acac method entails the cyclization of 2,4-pentanedione (acac), ammonium acetate, and formaldehyde to form
dihydropyridine 3, 5-diacetyl-l, 4-dihydrolutidine (DDL), which fluoresces at 510 nm after excitation at 412 nm.
This document is a draft for review purposes only and does not constitute Agency policy.
A-57 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
mean exhaled formaldehyde concentrations of 1.84 |ig/m3 in nonsmokers and 1.23-2.82 |ig/m3 in
smokers by the PTR-MS method, but no detectable formaldehyde in any subjects (including
smokers) by the formaldehyde-specific acac method (see Table A-15). A concentration of 5.13
|ig/m:i was detected by the acac method in a single smoker who was asked to smoke two cigarettes
immediately before the measurement This smoker's formaldehyde level declined below the level
of detection within 30 min. Formaldehyde levels were 1.47 to 2.09 |ig/m:i in subjects asked to
consume methanol-rich hard fruit liquor within 48 hours of the test (recall that methanol is
metabolized by alcohol dehydrogenase to formaldehyde throughout the body). So, even when
formaldehyde levels were intentionally elevated, very little endogenous formaldehyde was expelled
in exhaled breath and these elevations were transient
In summary, Riess et al. (2010), the only study to date which avoided the limitations of
previous studies, demonstrated that if endogenous formaldehyde exists in exhaled breath, it is
usually below their level of detection of <0.62 |ig/m3.
A.2.10. Conclusions Regarding the Toxicokinetics of Inhaled Formaldehyde Outside of the
POE
In summary, the published data demonstrate that endogenous formaldehyde blood levels
across species are approximately 0.1 mM and these levels do not change with exogenous
formaldehyde exposure, arguing that inhaled formaldehyde is not absorbed into blood. One
limitation of these studies is that these detection methods did not provide a clear distinction on the
nature of formaldehyde (e.g., free, reversibly or irreversibly bound, measured as formate, or part of
the 1C pool). Formaldehyde inhalation studies show metabolic incorporation, but not covalent
binding (e.g., hm-DNA adducts and DPCs) in bone marrow of rats which conclusively show that
exogenous formaldehyde is not transported to the distal tissues. Formaldehyde is likely to be
metabolized in a similar way in distal tissues since enzymes required for metabolism are expressed
in all the tissues. Endogenous levels of formaldehyde in exhaled breath analyzed by different
research groups are often limited due to the lack of specificity in analytical methods and
confounding by presence of formaldehyde in room air in these studies. Based on a recent improved
method, endogenous formaldehyde concentrations in exhaled air have been detected to be lower
than the study's detection limit of 0.62 |ig/m:i outside of exceptional circumstances (just after
smoking two cigarettes or ingesting something with a high level of methanol).
A.2.11. Toxicokinetics Summary
Formaldehyde is an endogenous chemical produced intracellularly by enzymatic and
nonenzymatic pathways during normal cellular metabolism and a relatively small fraction of free
formaldehyde is produced from metabolism of xenobiotics. Studies in experimental animals using
direct and indirect measurements and modeling studies in human subjects have clearly shown that
a majority of inhaled formaldehyde is rapidly absorbed in the URT despite anatomical and
physiological differences across species. Inhaled formaldehyde develops a concentration gradient
This document is a draft for review purposes only and does not constitute Agency policy.
A-58 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
with an anterior to posterior distribution in the nasal cavity. High concentrations of formaldehyde
are distributed to squamous, transitional, and respiratory epithelia; less formaldehyde uptake
occurs in the olfactory epithelium, and very little or no formaldehyde reaches the lower respiratory
tract, except possibly at very high exposure concentrations and/or during periods of high exertion
with oronasal breathing. Studies in rats show that single exposure to high levels of formaldehyde
or repeated exposure to varying concentrations does not appreciably change the tissue levels of
formaldehyde over the endogenous levels in the nasal mucosa.
Inhaled formaldehyde entering the nasal cavity interacts with the mucociliary apparatus
which is the first line of defense. The majority of formaldehyde is rapidly convered to methanediol
(=99.9%), with a minor fraction (=0.1%) remaining as free formaldehyde in the nasal mucus. A
rapid equilibrium is assumed such that the 99.9:0.1% ratio is maintained at all times. Methanediol
penetrates the tissues while free formaldehyde reacts with the macromolecules. Uncertainties
remain about formaldehyde transition to underlying epithelium owing to the presence of
endogenous formaldehyde, which is a component of normal cellular metabolism. Formaldehyde is
metabolized to formate predominantly by ADH3 and by a minor pathway involving mitochondrial
ALDH2. Formate can either enter the one-carbon pool leading to protein and nucleic acid synthesis,
or is further metabolized to CO2 and eliminated in expired air or excreted in urine unchanged.
Formaldehyde can interact with macromolecules either noncovalently (GSH, THF) or
covalently (DPX, DDX, hm-DNA monoadducts, protein adducts). In rats and monkeys, DPXs show
dose-response in the nasal cavity where DPX distribution corresponds to tumor sites (rats) and cell
proliferation (rats and monkeys), suggesting that DPX may be a good biomarker of exposure.
Formaldehyde also induces concentration-dependent increase in DNA monoadducts (e.g., N2-hm-dG
adducts) in the nasal passages of monkeys and rats which can be distinguished from endogenous
adducts by improved analytical methods. Higher levels of endogenous N2-hm-dG adducts are
detectable than the exogenous monoadducts, except at the highest exogenous exposure
concentrations.
The toxicokinetics of formaldehyde may be influenced by certain formaldehyde-induced
effects, such as modifications to mucociliary clearance, reflex bradypnea (rodents only) and
reduction in minute volume, and dynamic tissue remodeling (e.g., squamous metaplasia), which
have the potential to modulate formaldehyde uptake and clearance. For example, inhaled
formaldehyde induces mucostasis and ciliastasis in rat nasal mucociliary apparatus extending from
anterior to posterior regions of nasal cavity depending on the concentration and duration of
exposure. Thus, at least at higher concentrations (e.g., at low concentrations, formaldehyde does
not clearly cause mucostasis), estimates of tissue formaldehyde levels may be more uncertain.
Similarly, the differences observed in altered minute volumes in rats and mice during repeated
inhalation exposure to formaldehyde may impact dosimetric adjustment if extrapolated to humans.
Endogenous blood formaldehyde levels average around 0.1 mM across different species and
inhalation exposure to formaldehyde does not alter the blood formaldehyde levels, arguing that
This document is a draft for review purposes only and does not constitute Agency policy.
A-59 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
inhaled formaldehyde is not significantly absorbed into blood. Formaldehyde-induced exogenous
DNA monoadducts were detectable in nasal tissues but not in distal tissues of experimental animals
exposed by inhalation. This argues against systemic transport of formaldehyde to distal tissues.
Also, formaldehyde inhalation studies show metabolic incorporation, but not covalent binding in
bone marrow of rats, further supporting the lack of transport of formaldehyde (as opposed to
metabolites of formaldehyde) to the distal tissues.
Analysis of formaldehyde in exhaled breath can be confounded by interfering gases in the
analytical techniques or can be confounded by the presence of formaldehyde in the room air. With
improved techniques, endogenous formaldehyde concentrations in exhaled air have been detected
to be usually lower than the detection limit of 0.62 |ig/m:i. Overall, no evidence is available to
indicate that inhaled formaldehyde is systemically transported.
A.2.12. Modeling Formaldehyde Flux to Respiratory Tract Tissue
Formaldehyde is highly reactive and water soluble, thus its absorption in the mucus layer
and tissue lining of the respiratory tract is known to be significant. This absorption is highly
regional and the absorption patterns differ substantially across species. This section first provides
the motivation for developing detailed dosimetry models for the regional and species-specific
absorption of formaldehyde. It then discusses the computation of inhaled formaldehyde transport
in the upper (nose and mouth) and lower (lung and trachea) respiratory tract using fluid dynamic
models, and evaluates the level of confidence in these predictions. Finally, a revised dosimetry
model that incorporates estimates of endogenous formaldehyde is discussed.
Species differences in anatomy: consequences for gas transport and respiratory tract lesions
The regional dose of inhaled formaldehyde in the epithelial lining of the respiratory tract of
a given species depends on the amount absorbed at the airway-tissue interface, water solubility,
mucus-to-tissue phase diffusion, and chemical reactions, such as hydrolysis, protein binding and
metabolism, and on the amount of formaldehyde delivered by the inhaled air to the tissue lining.
This is a function of the major airflow patterns, air-phase diffusion, and absorption at the airway-
epithelial tissue interface. Formaldehyde-induced squamous cell carcinomas (SCC) and other
lesions that occur in the rat and monkey nasal passages and in the monkey lower respiratory tract
are seen to be localized, with the lesion distribution patterns also showing species-specificity. It
has been argued that the main determinant of these patterns and their differences among species is
regional dose (Moulin et al., 2002; Ibanes et al., 2996) (Bogdanffv et al.. 1999: Monticello etal..
1996: Monticello and Morgan. 1994: Morgan etal.. 1991).
The anatomy of the respiratory tract, in particular the upper part (see Figure A-10), and
airflow patterns in this region (see Figure A-ll) show large differences across species.
Furthermore, because of the convoluted nature of the airways (see Figure A-10), the uptake of
reactive and water-soluble gases such as formaldehyde in the upper respiratory tract (as seen in
various simulations, Figure A-12) is highly nonhomogeneous over the nasal surface. Thus, as
This document is a draft for review purposes only and does not constitute Agency policy.
A-60 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information for Formaldehyde—Inhalation
shown in Figure A-12, the spatial distribution of formaldehyde flux also shows strong species
dependence. These observations, when juxtaposed with the localized occurrence of lesions, suggest
that regional dose may be important in reducing uncertainty when extrapolating risk-related dose
across species. Kimbell etal. fl9931. Kepler etal. f!9981. and Subramaniam et al. f 19981 developed
anatomically realistic finite-element representations of the noses of F344 rats, rhesus monkeys, and
humans, and used them in physical and computational models (Kimbell etal.. 2001a: Kimbell etal..
2001b); see Figure A-10 and Figure A-ll). This assessment uses dosimetry derived from these
representations.
Formaldehyde dosimetry in the lower human respiratory tract (i.e., in the trachea and lung)
may also be important to consider. The upper respiratory tract is generally a good scrubber of
formaldehyde; as a result, there is less penetration into the lungs. However, the extent of this
scrubbing varies among species. The rat upper respiratory tract is extremely efficient with only
about 3% fractional penetration to the lower respiratory tract (Morgan et al.. 1986a): however,
penetration to the lung appears to be higher in the rhesus monkey (see Figure A-12). Accordingly,
while frank effects were seen only in the upper respiratory tract in rodents, DPX lesions induced by
exposure to 6 ppm formaldehyde were also present in the major bronchiolar region of the rhesus
monkey (see Section 1) whose respiratory tract morphology is somewhat similar to the human (see
Figure A-10 and Figure A-ll). Another factor is that humans are oronasal breathers, with a
significant fraction of the population breathing normally through the mouth (Niinimaa etal.. 1981).
while rats are obligate nose-only breathers. Oronasal breathing implies a much higher dose to the
lower respiratory tract, particularly at higher activity profiles (see Figure A-13 and Figure A-14
and Niinimaa et al. (1981). For all these reasons, the cancer dose-response assessment based upon
nasal tumors observed in the F344 rat includes an additional exercise involving the human lung,
even though the lung is not identified as a target organ in the hazard assessment. The dose-
response section evaluates the extent to which human risk estimates increase when formaldehyde
dose to the lower human respiratory tract is also considered. The dosimetry modeling for this
purpose uses an idealized single-path model of the lower respiratory tract developed by Overton
etal. (2001) discussed later Appendix B.2.2.
This document is a draft for review purposes only and does not constitute Agency policy.
A-61 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F344 Rat
Rhesus Monkey
Figure A-10. Reconstructed nasal passages of F344 rat, rhesus monkey, and
human.
Note: Nostril is to the right, and the nasopharynx is to the left. Right side shows the finite element mesh. Left-
hand side shows tracings of airways obtained from cross sections of fixed heads (F344 rat and rhesus monkey)
and magnetic resonance image sectional scans (humans). Aligned cross sections were connected to form a three-
dimensional reconstruction and finite-element computational mesh. Source: Adapted from Kimbell et al. (2001b).
Additional images provided courtesy of Dr. J.S. Kimbell, CUT Hamner Institutes.
Human
This document is a draft for review purposes only and does not constitute Agency policy,
A-62 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(A) F344 Rat
CFD-Model simulated inspiratory airflow
Streams observed in experimental water-dye studies
(B) Rhesus Monkey
CFD-Model simulated inspiratory airflow
Streams observed in experimental water-dye studies
(C) Human
CFD-Model simulated inspiratory airflow
Black bars = Axial velocities measured in hollow molds
White bars = CFD-simulated axial velocities
¦n ll ll I~1
'o 3 0 1
Plot 3
3.0
Plot 4
CO 2.0
¦i n r n ii,o
Jidilfliln.
o
S"
> 2.0 •
CO
x 10
<
Plot 5 3.Q ¦
i 111 1111 111
Plot 6
tfi n il il m
3.0
2.0 -
1.0 -
Plot 7 3.0
2.0 •
llllll 1 "
Plots
d\ fl H H In
Normalized airway diameter
Figure A-11. Illustration of interspecies differences in airflow and verification
of CFD simulations with water-dye studies.
Note: Panels A and B show the simulated airflow pattern versus water-dye streams observed experimentally in
casts of the nasal passages of rats and monkeys, respectively. Panel C shows the simulated inspiration airflow
pattern, and the histogram depicts the simulated axial velocities (white bars) versus experimental measurements
made in hollow molds of the human nasal passages. Dye stream plots were compiled for the rat and monkey
over the physiological range of inspiration flow rates. Modeled flow rates in humans were 15 L/minute.
Source: Adapted from Kimbell et al. (2001b).
This document is a draft for review purposes only and does not constitute Agency policy,
A-63 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F344 Rat
Key
pm ol /(m m2 -h r-ppm)
Rhesus Monkey
2000
1500
1000
500
0
Figure A-12. Lateral view of nasal wall mass flux of inhaled formaldehyde
simulated in the F344 rat, rhesus monkey, and human.
Note: This Is a rendering of a three-dimensional surface. Nostrils are to the right. Simulations were exercised in
each species at steady-state inspiration flow rates of 0.576 L/minute in the rat, 4.8 L/minute in the monkey, and
15 L/minute in the human. Flux was contoured over the range from 0-2,000 pmol/(mm'-hour-ppm) in each
species.
Source: Kimbell et al. (2001b).
This document is a draft for review purposes only and does not constitute Agency policy,
A-64 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
18 L/min
25.8 L/min
31.8 L/min
37 L/min
Figure A-13. Lateral view of nasal wall mass flux of inhaled formaldehyde
simulated at various inspiratory flow rates in a human model.
Note: This is a rendering of a three-dimensional surface, showing the right lateral view. Uptake is shown for the
nonsquamous portion of the epithelium. The front portion of the nose (vestibule) is lined with keratinized
squamous epithelium and is expected to absorb relatively much less formaldehyde.
Source: (Kimbell et al., 2001a).
1 Modeling formaldehyde uptake in nasal passages
2 Anatomical reconstruction and tissue types: The dose-response modeling results evaluated
3 and used in this document are based on several published computational models for air flow and
4 formaldehyde uptake in the nasal passages of a F344 rat5, rhesus monkey, and human, and in the
5 human lung (Kimbell et al.. 2001b: Overton etal.. 2001: Kepler etal.. 1998: Subramaniam et al..
6 1998: Kimbell etal.. 1993). The anatomical reconstructions for both computational and physical
5This strain of the rat is considered anatomically representative of its species and widely used experimentally, most
notably in bioassays sponsored by the National Toxicology Program.
This document is a draft for review purposes only and does not constitute Agency policy,
A-65 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
models were based on tracings of airways obtained from cross sections of fixed heads (F344 rat and
rhesus monkey) and magnetic resonance image sectional scans (human).
Formaldehyde-induced nasal SCCs in rats are observed to arise only from respiratory or
transitional epithelial cells in F344 rats and thought to be associated with the transformation of
these cell-types to a squamous epithelial type due to exposure to formaldehyde fMorgan etal..
1986a). Therefore the dosimetry calculations in Kimbell etal. (2001b) focused on predicting the
wall mass flux of formaldehyde (rate at which mass of formaldehyde is transported to unit area of
the nasal or lung lining prior to disposition within the body—mass/[area-time]) to regions lined by
respiratory or transitional epithelium and excluding squamous epithelial cells. An additional
distinction was made regarding these regions. Formaldehyde hydrolyses in water and reacts
readily with a number of components of nasal mucus, and was therefore assumed to be absorbed at
a higher rate by epithelial lining coated with mucus. The approximate locations of mucus-coated
and nonmucus coated respiratory/transitional epithelial cells were mapped onto the reconstructed
nasal geometry of the computer models. Types of nasal epithelium overlaid onto the geometry of
the models were assumed to be similar in characteristics across all three species (rat, monkey, and
human) except for thickness, surface area, location, and the extent of the nasal surface not coated
by mucus. These characteristics were estimated from the literature or by direct measurements
fConollvetal.. 2000: CUT. 19991.
The fluid dynamics modeling in the respiratory tract comprises two steps: (1) model airflow
through the airway lumen (solution of Navier-Stokes equations) and (2) using these solutions of the
airflow field as input, model formaldehyde flux to the respiratory tract lining (solution of
convective-diffusion equations). The local formaldehyde flux at the airway-to-epithelial tissue
interface was assumed to be proportional to the air-phase formaldehyde concentration adjacent to
the nasal lining. The proportionality constant is the mass transfer coefficient for the tissue phase,
specified as boundary conditions on the solutions, and takes different values in the model
depending on whether the tissue is coated with a mucus layer (km) or not (knm). Epithelium not
coated with mucus was considered similar to epidermal tissue, and a value available from the
literature for such tissue was used for knm- On the other hand, Kimbell et al. determined km
empirically for the rat by fitting the overall nasal uptake predicted by the CFD model to the average
experimental values obtained by Morgan et al. f!986al. The values of km and knm depend only on
the solubility and diffusivity of the gas in the tissue, the thickness of tissue, and the reaction rate of
the gas (Hanna et al., 2001@@author-year}. Tissue thickness varies across species, but because
formaldehyde is highly reactive and soluble, the primary kinetic determinant of interspecies
differences in the net mass transfer rate is likely the difference in air-phase resistance and not
tissue thickness. Therefore, Kimbell etal. (2001b) assumed that values for the tissue phase mass
transfer coefficients were the same for the human. EPA judges this assumption to be reasonable.
The air-phase resistance (which is the inverse of the air-phase mass transfer coefficient) on the
other hand would vary substantially between the rat and human on account of the substantial
This document is a draft for review purposes only and does not constitute Agency policy.
A-66 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
interspecies variations in airway geometry and airflow discussed earlier. Details of the boundary
conditions for air flow and mass transfer, are provided in Kimbell et al. (2001b; 2001: 19931 and
Subramaniam et al. f 19981.
For the rat, minute volumes were allometrically scaled to 0.288 L/minute for a 315 g rat
(Mauderly, 1986), and simulations were carried out at the steady-state unidirectional inspiratory
rate of 0.576 L/minute. For the human, simulations were carried out at the steady-state
unidirectional inspiratory rate of 15,18, 50, and 100 L/min, corresponding to half of the values for
the minute volumes associated with the activity patterns of sleeping, sitting, and light and heavy
exercise, respectively fICRP. 19941. Because formaldehyde is highly water soluble and reactive,
Kimbell etal. f2001bl assumed that uptake occurred only during inspiration. Thus, for each breath,
flux into nasal passage walls (rate of mass transport in the direction perpendicular to the nasal wall
per mm2 of the wall surface) was assumed to be zero during exhalation, with no backpressure to
uptake built up in the tissues. Overton etal. (2001) estimated the error due to this assumption to
be small, roughly an underestimate of 3% in comparison to cyclic breathing. Inspiratory airflow
was assumed to be constant in time (steady state). Subramaniam etal. (1998) considered this to be
a reasonable assumption during resting breathing conditions based on a value of 0.02 obtained for
the Strouhal number. Unsteady effects are insignificant when this number is much less than one.
However, this assumption may not be reasonable for light and heavy exercise breathing scenarios.
Kimbell etal. (2001b) partitioned the nasal surface by flux to facilitate the use of local
formaldehyde dose in dose-response modeling. Each of the resulting 20 "flux bins" was comprised
of elements of the nasal surface that receive a particular interval of formaldehyde flux per ppm of
exposure concentration (Kimbell etal.. 2001b). These elements were not necessarily contiguous.
The spatial coordinates of elements comprising a particular flux bin were fixed for all exposure
concentrations, with formaldehyde flux (pmol/(mm2-hour) in a bin scaling linearly with exposure
concentration (ppm), and therefore often expressed in terms of flux per ppm, that is,
pmol/(mm2-hour-ppm).
Mass flux was estimated for the rat, monkey, and human over the entire nasal surface and
over the portion of the nasal surface that was lined by nonsquamous epithelium (lateral wall mass
flux shown in Figure 12). Formaldehyde flux was also estimated for the rat and monkey over the
areas where cell proliferation measurements were made fMonticello etal.. 1991: Monticello etal..
19891 and over the anterior portion of the human nasal passages that is lined by nonsquamous
epithelium. Maximum flux estimates for the entire upper respiratory tract were located in the
mucus-coated squamous epithelium on the dorsal aspect of the dorsal medial meatus near the
boundary between nonmucus and mucus-coated squamous epithelium in the rat, at the anterior or
rostral margin of the middle turbinate in the monkey, and in the nonsquamous epithelium on the
proximal portion of the mid-septum near the boundary between squamous and nonsquamous
epithelium in the human see Kimbell etal. f2001al. The rat-to-monkey ratio of the highest site-
specific fluxes in the two species was 0.98. In the rat, the incidence of formaldehyde-induced SCCs
This document is a draft for review purposes only and does not constitute Agency policy.
k-67 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Supplemental Information for Formaldehyde—Inhalation
in chronically exposed animals was high in the anterior lateral meatus (ALM, Monticello et al.
(1996). Flux (per ppm of inhaled concentration) at this site in the rat was similar to that predicted
near the anterior or proximal aspect of the inferior turbinate and adjacent lateral walls and septum
in the human, with a rat-to-human ratio of 0.84.
Formaldehyde Uptake in The Lower Respiratory Tract
Unlike the nasal passages, the human lower respiratory tract lends itself to a more
simplified or idealized rendering. The one-dimensional (known as a "single-path" model) rendering
of the human lung anatomy by Weibel (1963), which captures the geometry of the airways in an
average or homogeneous sense for a given lung depth, is generally considered adequate unless the
fluid dynamics at locations of airway bifurcations need to be explicitly modeled. Such an
idealization of lung geometry has been successfully used in various models for the dosimetry of
ozone and particulate and fibrous matter.6 The single-path model was used to calculate
formaldehyde uptake in the human lower respiratory tract (Overton etal.. 2001: CUT. 1999). These
authors applied a one-dimensional equation of mass transport to each generation of an adult
human symmetric, bifurcating Weibel-type respiratory tract anatomical model. In order to achieve
consistency with the inhaled output from the CFD model of the upper respiratory tract in
Subramaniam et al. f 19981. Overton etal. f20011 augmented their model with an idealized upper
respiratory tract, and constrained their one-dimensional version of the nasal passages to have the
same inspiratory air-flow rate and uptake during inspiration as the CFD simulations. Results most
relevant to this assessment are shown in Figure A-ll.
6Such idealized representations are likely to be inappropriate for considering susceptible individuals, such as those
with chronic obstructive pulmonary disease.
This document is a draft for review purposes only and does not constitute Agency policy.
A-68 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
MODEL GENERATION
Figure A-14. Single-path model simulations of surface flux per ppm of
formaldehyde exposure concentration in an adult male human.
Source: Overton et al. (2001).
1 The primary predictions of the model were: more than 95% of the inhaled formaldehyde is
2 retained; formaldehyde flux in the lower respiratory tract increases for several lung airway
3 generations relative to flux in posterior-most segment of the nose; with further increase in lung
4 depth, formaldehyde flux decreases rapidly resulting in almost zero flux to the alveolar sacs.
5 Overton et al. (2001) also modeled uptake at high inspiratory rates. At a minute volume of
6 50 L/minute7 formaldehyde flux in the mouth cavity is comparable (but a bit less) to that occurring
7 in the nasal passages (see Figure A-14).8
7 Note: the ororiasal switch occurs at about 35 L/minute Niiriimaa et al. (1981).
8 Mouth breathers form a large segment of the population. Furthermore, at concentrations of formaldehyde where
either odor or sensory irritation becomes a significant factor, humans are likely to switch to mouth breathing even
at resting inspiration. Overton et al. (2001] did not model uptake in the oral cavity at minute volumes less than
This document is a draft for review purposes only and does not constitute Agency policy,
A-69 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
Level of confidence in formaldehyde uptake simulations
As mentioned earlier, the computational fluid dynamics simulations involved two steps, and
the confidence in each step is addressed separately below.
Confidence in predicted airflow profiles
To verify the CFD simulations of nasal airflow profiles, the authors constructed physical
models from the finite-element reconstructions used in the computational models. The simulated
streamlines of steady-state inspiration airflow predicted by the CFD model agreed reasonably well
with experimentally observed patterns of water-dye streams made in casts of the nasal passages for
the rat and monkey as shown in panels A and B in Figure A-ll. The airflow velocity predicted by
CFD model simulations of the human also agreed well with measurements taken in hollow molds of
the human nasal passages (see panel C, Figure A-ll) (Kepler etal.. 1998: Subramaniam et al.. 1998:
Kimbell etal.. 1997: Kimbell etal.. 1993). However, the accuracy and relevance of these
comparisons are limited. Because the airflow profiles were verified by only a simple video analysis
of dye streak lines observed in the physical molds this method can be considered reasonable for
only the major airflow streams. For the human, axial airflow velocities were also measured
experimentally in a physical cast, and these compared well with CFD simulations (see panel C in
Figure A-ll). However, the physical model used for the velocity measurements corresponds to
that of a different individual than the one for which the CFD simulations were carried out.
Another verification comes from measuring pressure gradients across the nasal cavity.
Plots of pressure drop versus volumetric airflow rate predicted by the CFD simulations compared
well with measurements made in rats in vivo fGerde etal.. 19911 and in acrylic casts of the rat nasal
airways (Cheng et al., 1990) as shown in Figure A-15. This latter comparison remains qualitative
due to differences among the simulation and experiments as to where the outlet pressure was
measured and because no tubing attachments or other experimental apparatus were included in
the simulation geometry. The simulated pressure drop values were somewhat lower, possibly due
to these differences.
Kimbell etal. (2001a) examined the extent to which their results were subject to errors in
mass balance and applied ad-hoc corrections to compensate for these errors. Because airflow and
uptake were simulated separately, they each contributed separately to the mass balance error;
however, the error component due to airflow was minimal (< 0.4%). The percent overall uptake of
formaldehyde was defined as 100% x (mass entering nostril - mass exiting outlet)/(mass entering
nostril), and its mass balance error was calculated as 100% x (mass entering nostril - mass
50 L/minute. However, since 0.55 of the inspired fraction is through the mouth for the normal nasal breathing
population (Niinimaa et al.. 1981) at an inspiratory rate of 50 L/min, we can make an indirect inference from
their result at this heavy breathing rate that average flux across the human mouth lining would be comparable to
the average flux across the nasal lining computed in Kimbell et al. (2001b; 2001) for mouth breathing conditions
at resting or light exercise inspiratory rates.
This document is a draft for review purposes only and does not constitute Agency policy.
A-70 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
absorbed by airway walls - mass exiting outlet)/(mass entering nostril). For the rat, monkey and
human the mass balance errors associated with simulated formaldehyde uptake from air into tissue
were less than 14% at resting minute volumes, and therefore, not a major concern, but these errors
increased to 27% at the highest human inspiratory rate corresponding to exercise conditions.
Kimbell etal. f2001al corrected for these errors by evenly distributing the lost mass over the entire
nasal surface in their simulation results.
800
700
600
£ 500
a.
2
I 400
300
200
100
0
0 100 200 300 400 500 600 700 800 900 1000
Airflow rata (ml/min)
Figure A-15. Pressure drop versus volumetric airflow rate predicted by the
CUT CFD model compared with pressure drop measurements made in two
hollow molds (CI and C2) of the rat nasal passage (Cheng et al., 1990) or in
rats in vivo (Gerde et al.. 1991).
Source: Kimbell et al. (1997).
Confidence in modeled flux estimates
Unlike the verification of the airflow simulations, it was not possible to evaluate the regional
formaldehyde flux calculations directly; however, there are several indirect qualitative and
quantitative lines of evidence that provide general confidence in the flux profiles predicted by
Kimbell et al. (2001b; 20011 for the F344 rat nasal passages when the flux is averaged over gross
regions of the nasal lining. This evidence is listed below.
In Kimbell etal. (2001b). the tissue-phase mass-transfer boundary conditions were set by
fitting overall (whole nose) formaldehyde uptake at various exposure concentrations to the
experimental data in fMorgan etal.. 1986al. Since this was the only data set available, it was not
possible to independently verily the model results for overall uptake. However, results from earlier
work by Kimbell et al. f 19931 are informative for this purpose because in this case the model was
This document is a draft for review purposes only and does not constitute Agency policy.
A-71 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
not calibrated by fitting model predictions to experimental data; instead, this model assumed an
infinite sink for absorption at the nasal lining on account of the highly reactive and soluble nature of
formaldehyde. Kimbell etal. f!9931 predict 99% uptake of inhaled formaldehyde in the rat nose,
which is slightly above the upper end of the range of 91-98% observed by Morgan et al. (1986a).
The utility of those simulations is however limited because the posterior portion of the nose was
not included in the model, and the assumption of infinitely absorbing nasal walls makes the
boundary condition less realistic than that used in Kimbell etal. f2001bl. Calculations based upon
Kimbell etal. (19931 are compared with various experimental observations below.
Morgan et al. f!9911 showed general qualitative correspondence between the main routes
of flow and lesion distribution induced by formaldehyde in the rat nose, and hypothesized that the
localized nature of the lesions must be related to the regional uptake of formaldehyde. This was
borne out by Kimbell etal. f!9931 who described similarities in patterns of computed regional mass
flux and lesion distribution due to formaldehyde. These authors reported on correlations in
patterns in the coronal section immediately posterior to the vestibular region (as discussed earlier,
the vestibular region is protected by keratinized epithelium and is therefore not likely to
significantly absorb formaldehyde); simulated flux levels over regions where lesions were seen,
such as the medial aspect of the maxilloturbinate and the adjacent septum, were an order of
magnitude higher than over other regions where lesions were not seen, such as the nasoturbinate.9
A reasonable level of confidence in flux predictions by Kimbell etal. (1993) is also attained
indirectly by comparing experimental data on formaldehyde-DPX concentration in the F344 rat
with modeled results in Cohen-Hubal etal. (1997): these authors used flux estimates generated by
the CFD model in Kimbell etal. (1993) in a physiologically-based pharmacokinetic (PBPK) model
for formaldehyde-DPX concentration in the F344 rat. This hybrid CFD-PBPK model was calibrated
by optimizing model predictions of DPX concentrations against DPX collected over the entire nose
in separate experiments by Casanova et al. f!991: 19891 on F344 rat noses exposed to
formaldehyde at 0.3, 0.7, 2.0, 6.0, and 10 ppm. The nasal regions were then separated into two
categories depending upon whether tumor incidence was high or low in a region, and model
predictions of DPX concentrations were compared with the experimental data considered only
from the high-tumor region, including additional DPX data from the high-tumor region at 15-ppm
exposure concentration which had not been used in model calibration. The predictions are seen to
compare well with experimental values (see Figure A-16). Such a comparison is not available for
the simulation of uptake patterns in the human.
9This 1993 CFD model differed somewhat from the subsequent model by Kimbell et al. (2001b) used in this
assessment. In the 1993 model, the limiting mass-transfer resistance for the gas was assumed to be in the air
phase; that is, the concentration of formaldehyde was set to zero at the airway lining. Furthermore, this same
boundary condition was used on the nasal vestibule as well, while in the more recent model, the vestibule was
considered to be nonabsorbing. Unfortunately, Kimbell et al. (2001b) did not report on correspondences
between flux patterns and lesion distribution.
This document is a draft for review purposes only and does not constitute Agency policy.
A-72 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
36.5
X
CL
Q
£ 0.5
HCHO Flux (jiM-mm/rn irt)
73.0 199.5 146.0 18215 219.0 255.5
2.5"
2-5 5J
HCHO Exposure Concentration (ppm)
HCHO Flux (jjM-mm/mln)
34.5 69,9 193.5 133.0 172.5 297.B 241.5
T
-r
(D
x
u_
~
HCHO Exposure Concentration (ppm)
Figure A-16. Formaldehyde-DPX dosimetry in the F344 rat.
Panel A: calibration of the PBPK model using data from high and low tumor incidence sites. Panel B: model
prediction compared against data from high tumor incidence site. Dashed line in panel A shows the extrapolation
outside the range of the calibrated data.
Source: Cohen-Hubal et al. (1997).
1 Effect of reflex bradvpnea on dosimetry
2 A source of uncertainty in the modeled human flux estimates arises because the value of the
3 tissue-phase mass-transfer coefficient used as a boundary condition in human simulations is the
4 same as that obtained from calibration of the rat model. As explained earlier, qualitatively this
5 appears reasonable; however, EPA is unable to quantitatively evaluate the impact of this
6 uncertainty.
This document is a draft for review purposes only and does not constitute Agency policy.
A-73 DRAFT-DO NOT CITE OR QUOTE
7.5 re.# 12.s is. e 17.5
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
The CFD simulations do not model reflex bradypnea, a protective reflex observed in
rodents. As discussed at length in Section A-3, it is reasonable to expect a range of 25% fChang et
al.. 19831 to 45% (Barrow etal.. 19831 decrease in minute volume in F344 rats at the exposure
concentration of 15 ppm. Explicit omission of this effect in the modeling is, however, not likely to
be a source of major uncertainty in the modeled results for uptake of formaldehyde in the rat nose
for the following reason: the CFD model for the F344 rat was calibrated to fit the overall
experimental result for formaldehyde uptake in the F344 rat at 15 ppm exposure concentration by
adjusting the mass transfer coefficient used as boundary condition on the absorbing portion of the
nasal lining. Thus, any reflex bradypnea occurring in those experimental animals is implicitly
factored into the value used for the boundary condition. Nonetheless, some error in the localized
distribution of uptake patterns may be expected, even if the overall uptake is reproduced correctly.
Modeling Interindividual Variability in the Nasal Dosimetry of Reactive and Soluble Gases
Garcia etal. (2009) used computational fluid dynamics to study human variability in the
nasal dosimetry of reactive, water-soluble gases in 5 adults and 2 children, aged 7 and 8 years. The
authors considered two model categories of gases, corresponding to maximal and moderate
absorption at the nasal lining. We focus here only on the "maximal uptake" simulations in Garcia et
al. f20091: note that this term for the simulations does not correspond to regions of maximum flux
but rather characterizes the gas category. In this case, the gas was considered so highly reactive
and soluble that it was reasonable to assume an infinitely fast reaction of the absorbed gas with
compounds in the airway lining. Although such a gas could be reasonably considered as a proxy for
formaldehyde, these results cannot be fully utilized to inform quantitative estimates of
formaldehyde dosimetry (and does not appear to have been the intent of the authors either). This
is because the same boundary condition corresponding to maximal uptake was applied on the
vestibular lining of the nose as well as on the respiratory and transitional epithelial lining on the
rest of the nose. This is not appropriate for formaldehyde as the lining on the nasal vestibule is
made of keratinized epithelium which is considerably less absorbing than the rest of the nose
(Kimbell etal., 2001).
Garcia etal. (2009) concluded that overall uptake efficiency, and average and maximum flux
levels over the entire nasal lining did not vary substantially between adults (1.6-fold difference in
average flux and much less in maximum flux), and the mean values of these quantities were
comparable between adults and children. These results are also in agreement with conclusions
reached by Ginsberg etal. (2005) that overall extrathoracic absorption of highly and moderately
reactive and soluble gases (corresponding to Category 1 and 2 reactive gases as per the scheme in
U.S. EPA (1994) is similar in adults and children. On the other hand Garcia etal. (2009) state that
their models predicted significant interhuman variability in flux levels at specific points on the
nasal wall; figure 6A of their paper (reproduced here as Figure A-17) indicates a 3- to 5-fold
difference among the individuals in the study when flux was plotted as a function of distance from
the nostrils normalized by the length of the septum. This observation needs to be accompanied by
This document is a draft for review purposes only and does not constitute Agency policy.
A-74 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
a caveat: because similar fluxes may correspond to different regions in individuals, it is possible
that this spread in values overestimates the actual variability in local flux in these individuals.
A 1.0E-07
«N
E
jn
" 1.0E-08
3
§
Adults - maximum uptake
Children - maximum uptake
1.0E-09 I , , i
-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
(Distance from nostril s)/( Length of septum)
Figure A-17. Flux of highly reactive gas across nasal lining as a function of
normalized distance from nostril for 5 adults and 2 children.
While the sample size in this study is too small to consider the results representative of the
population as a whole, various comparisons with the characteristics of other study populations add
to the strength of this study; for example, the surface area to volume ratio among the five adults
ranged from 0.87 to 1.12 mm-1 which compared well with a result of 1.05±0.23 obtained from
measurements in 40 adult Caucasians (Yokley, 2006), and the surface area ranged from 16,683 to
23,219 cm2 which compared well with a result of 18,300±2,200 cm2 obtained from measurements
in 45 adults (Guilmette et al., 1997). It is useful to note here that the nasal anatomy reconstructed
for modeling the dosimetry of formaldehyde in the human nose in Kimbell et al. fKimbell etal..
2001b: 2001) and discussed earlier was that of one of the individuals in the Garcia et al. study.
Models Estimating the Effects of Endogenous Formaldehyde on Dosimetry Predictions in Nasal
Tissues
Schroeter etal. T20141 developed a hybrid toxicokinetic fluid dynamic model for predicting
the uptake of inhaled formaldehyde that incorporates the production of endogenous formaldehyde
in nasal tissue, and estimated a net decrease in uptake of inhaled formaldehyde at the lowest
exposure concentrations based on modeling assumptions regarding the intracellular concentration
of endogenous formaldehyde. More specifically, due to endogenous formaldehyde production, the
model of Schroeter et al. (2014) predicts a net desorption of formaldehyde at zero exposure and
that an external exposure between 1.23 |a,g/m3 and 12.3 |a,g/m3 (0.001 and 0.01 ppm) is required
before there is sufficient air concentration to cause a net uptake of formaldehyde. However, any
This document is a draft for review purposes only and does not constitute Agency policy.
A-75 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 external exposure is predicted to cause some, albeit very small, increase in the tissue concentration,
2 since a nonzero air concentration reduces the net efflux of endogenous formaldehyde. While the
3 analysis of Schroeter et al. (2014) represents an important first step towards incorporating the
4 presence of endogenous formaldehyde into models estimating the flux (or uptake) of inhaled
5 formaldehyde, several uncertainties in the underlying assumptions have yet to be addressed:
6 • Endogenous formaldehyde levels were calculated based on blood concentrations. But Heck
7 and colleagues (1982) measured 12.6 jag/g total formaldehyde in rat nasal tissues and only
8 2.24 jag/g in rat blood (Heck etal.. 1985).
9 • Based on DNA-adduct measurements, it appears that the majority of formaldehyde is bound
10 to GSH in a manner that reduces its interaction with DNA and, presumably, other key
11 macromolecules (see A.l.1.3.3.3). The extent of GSH-binding could significantly reduce
12 diffusion across the epithelial cell membrane (i.e., between blood and nasal tissue), in which
13 case blood concentrations may not correlate well with tissue concentrations.
14 • Since nasal tissue levels of formaldehyde are higher than blood levels, it is likely that these
15 levels are produced by endogenous metabolism in situ, rather than entering the mucosa via
16 diffusion from a "blood" layer at a specific depth from the mucosa-air surface, the latter
17 being the assumption used by Schroeter et al. (2014).
18 • The tissue levels of formaldehyde predicted by the model of Schroeter et al. (2014) appear
19 to be orders of magnitude in excess of the levels that would be consistent with the observed
20 DPX levels (Heck etal.. 1983) and formaldehyde-DNA binding rate (Heck and Keller. 1988).
21 • While Schroeter et al. (2014) did not report exhaled breath levels, their results indicate that
22 uptake will exactly balance desorption in humans at about 1.23 |a,g/m3 (0.001 ppm or
23 1 ppb), from which one might assume this is the level their model would predict in exhaled
24 breath. In the study of Riess et al. (2010). exhaled breath levels for nonsmokers were found
25 to be below a detection limit of 0.62 |ig/m3, which corresponds to 0.5 ppb at 20°C. While
26 this is within a factor of two, an acceptable level of error for such an extrapolation, it is a
27 further indication that the assumed level of free endogenous formaldehyde in the Schroeter
28 et al. (2014) model is too high.
29 Despite these limitations, the efforts by Schroeter et al. (2014) highlight the fact that at
30 sufficiently low levels of exogenous formaldehyde, the contribution of endogenous formaldehyde
31 could become significant; accounting for this contribution would address a critical uncertainty for
32 interpreting the uptake of inhaled formaldehyde. Additional studies addressing the potential
33 contribution of endogenous formaldehyde are warranted. As discussed in the Toxicological Review
34 (see Section 2.2.1), the unit risk estimate for nasal cancers based on rat studies are not appreciably
35 altered if calculated using the revised formaldehyde estimates from Schroeter et al. (2014).
36 Campbell Tr et al. (2020) modified the original model by Andersen etal. (2010) using
37 exogenous and endogenous formaldehyde adduct data from Lengetal. (2019) (2 8 day study of 6
38 hrs/day exposures), Yu etal. (2015b) (28 day study of 6 hrs/day exposures), and Lu et al. (2011:
39 2010) (a single 6-hour exposure). The following major changes were made to the original model:
This document is a draft for review purposes only and does not constitute Agency policy.
k-76 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
a) The model simulates observed data for formaldehyde-induced DNA mono-adducts (N2-
hydroxymethyl-dG). The previous models simulated formaldehyde-induced DNA-protein
cross-links (DPX).
b) A zero-order term (VMMUC) was used to account for tissue clearance of inhaled
formaldehyde. This is a restriction on uptake from the air phase to the tissue compartment.
c) The rate of production of endogenous formaldehyde (Kp) was increased to nearly double
the original rate set by Andersen etal. (20101. The maximum rate of formaldehyde oxidase
metabolism (Vmax) was increased by over a factor of 10.
There are some notable observations from the data used in the modeling. Lengetal. f20191
showed no exogenous formaldehyde-induced DNA adducts in the nose at concentrations up to 0.3
ppm and no increase in endogenous formaldehyde-induced DNA adducts up to 0.3 ppm. Lu et al.
(20H; 20101 observed an increase in exogenous formaldehyde adducts in rat nasal tissue starting
at 0.7 ppm but no increase in endogenous adducts between 0.7 ppm-15 ppm (although there does
appear to be a perturbation in the mean and variance of endogenous adducts in this range). The
data at and above 0.7 ppm was used to re-optimize the cellular metabolic parameters. The data up
to 0.3 ppm by Lengetal. f20191 (which did not observe increased adducts) was used to visually
optimize the parameter defining the lower limit on uptake (VMMUC). Because of the abrupt change
in observed adduct levels between 0.3 ppm and 0.7 ppm there is model uncertainty within that
concentration range and below the limit of detection.
Key results from this work add to our characterization of uncertainties related to
endogenous formaldehyde levels and formaldehyde dose-response at low exposures. First, the
model estimated a non-zero value for VMMUC, indicating that the inhalation rate must exceed the
tissue clearance rate for formaldehyde to be absorbed by the tissue. The model was calibrated with
the restriction that formaldehyde absorption in the nose occurs only at exposure concentrations
above 0.3 ppm in the rat Secondly, Campbell Tr etal. f20201 assessed steady-state concentration of
free endogenous FA to be 20x lower than the value determined experimentally by Heck et al.
(1982) and 15 x lower than assessed by Andersen etal. (2010)). In Campbell Tr etal. (2020). the
estimate for free endogenous levels decreased from 0.31 mM to 0.020 mM and the basal
concentration of endogenous formaldehyde bound to sulfhydryl increased from 0.057 to 0.12mM
(2x higher). Campbell Tr etal. f20201 attributed this discrepancy to the potential for the Heck et al.
measurement methodology to overestimate tissue formaldehyde levels.
The original model Andersen etal. f 20101 did not adequately fit these new data, and
Campbell et al. justified changes to the Andersen et al. (2010) model parameters for cellular
metabolism on the grounds that data from Heck etal. (1982) are biased due to the method used to
measure tissue formaldehyde. However, it is possible that the cause of this model/data discrepancy
is inadequate model structure rather than a bias in the original data. As a result, there is inherent
model uncertainty in the revised model for cellular metabolism.
This document is a draft for review purposes only and does not constitute Agency policy.
A-77 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for Formaldehyde—Inhalation
Extrapolation of results in Campbell Tr et al. (2020) to humans is not possible because the
data and the model are specific to rats.
A.3. REFLEX BRADYPNEA
Reflex bradypnea (RB) is a protective reflex that allows laboratory rodents to minimize
their exposure to upper respiratory tract (URT) irritants such as aldehydes, ammonia, isocyanates,
and pyrethroids (Gordon et al.. 20081. This reflex is initiated by stimulation of trigeminal nerve
endings in the mucosa of the URT and the eyes. It is associated with the chemosensitive part of the
nociceptive system—the common chemical sense that detects noxious airborne exposures fNielsen.
1991).
The signs of reflex bradypnea: RB is manifest by immediate decreases in the metabolic
rate, CO2 production, and demand for oxygen. This is followed by rapid decreases in body
temperature (i.e., hypothermia; as much as 11°C in rats and 14°C in mice; Figure A-18), activity,
heart rate, blood pressure, respiratory rate (breaths/minute; Figure A-19), and minute volume (see
Figure A-20). RB also results in decreased blood pC>2 and pCO2 and increased blood pH (see Figure
A-21) {Pauluhn, 1989,; Pauluhn, 1996,; Pauluhn, 2003, ;Pauluhn, 2008 ; Gordon, 2008, 626432;
Jaeger, 1982, 42673}; Chang, 1984,10197}. Thus, the physiological effects and signs of RB maybe
misinterpreted as, for example, chemical-induced behavioral or developmental effects.
RB is regulated by a complex feedback response (Yokley. 2012). Gordon et al. (2008)
demonstrated that the extent of RB depends on the concentration of the irritant (see Figure A-18).
For example, after several hours of exposure to an isocyanate, mice exhibited concentration-
dependent changes with those in the high concentration group presenting a mean body
temperature of 23°C and approximately 90% decreases in respiratory rate and minute volume.
Figure A-18. Left panel: Concentration-related hypothermia in mice exposed
to an isocyanate for 360 minutes. Note the gradual recovery in body temperature
after exposure ceased.
This document is a draft for review purposes only and does not constitute Agency policy.
A-78 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Right panel: Concentration-related decreases in respiratory rate in mice exposed to an isocyanate. Note the
correlation between the curves for rectal temperature and respiratory rate over the course of 180 minutes.
Source: (Gordon et al., 2008).
Figure A-19. An oscillograph that compares the respiratory cycle for mice
exposed to an URT irritant (lower tracing) to an air control group (upper
tracing). The exposed animals have a characteristic pause before exhaling—a
bradypneic period—which results in a net decrease in the respiratory rate
(breaths/minute). Because the exposed group has a slightly greater tidal volume
(height of the tracings) but a much lower respiratory rate, the net result is a lower
minute volume and reduced exposure to the irritant
Source: Kane and Alarie (1977).
This document is a draft for review purposes only and does not constitute Agency policy,
A-79 ' DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
B6C3F., Mice ^AA Ratc
• Naive (6 ppm)
0 Pretreated (6 ppm)
¦ Naive (15 ppm)
D Pretreated (15 ppm)
20
1Q 1111111111
0 2 4 6 810
120
240
i
360 0 2 4 6 810
120
240
360
Figure A-20. Formaldehyde effects on minute volume in naive and
formaldehyde-pretreated male B6C3F1 mice and F344 rats. Pretreated animals
were exposed to 6.9 or 17.6 mg/m3 formaldehyde 6 hours/day for 4 days. Note that
the mice had a greater response than the rats, and the pretreated animals had a
greater response than the naive animals.
Source: Redrawn from Chang et al. (1983).
Figure A-20 demonstrates that the onset of RB after formaldehyde inhalation is immediate,
with a marked decrease in minute volume in mice and rats minutes after exposure begins. Because
reduced respiration lessens exposure to an irritating chemical, the toxicity is reduced and the
animal's survival is enhanced. This is important for the survival of rodents living in burrows and
confined spaces that may not be able to avoid exposure. Figure A-18 (left panel) demonstrates that
the effects of RB are reversible, but it can take several minutes to several hours for all physiological
parameters to return to preexposure conditions, depending on the extent of hypothermia (Barrow
etal.. 1983: laeger and Gearhart. 19821 {Pauluhn, 1996,}.
The physiological signs of RB in rodents can be striking but they are not signs of toxicity
and, as such, are not considered appropriate for defining an animal POD. Also, the signs of RB are
not relevant to humans since humans cannot experience RB. RB can only occur in small animals
such as mice and rats that can, because of their small size, rapidly lower their core body
temperatures when their metabolic rate reflexively decreases. Even a mild decrease in body
temperature can lessen the toxicity and metabolic activation of many chemicals, but it can also slow
the excretion of toxicants. Overall, the protection from cellular toxicity afforded by RB-induced
hypothermia outweighs the undesirable effect of a slower excretion rate f Gordon etal.. 20081.
This document is a draft for review purposes only and does not constitute Agency policy.
A-80 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for Formaldehyde—Inhalation
Even though RB has been reported in the literature since the 1960s, it is largely unknown to most
toxicologists. None of the rodent inhalation studies of formaldehyde, except for a few RB-specific
studies, attempted to identify or measure RB, including measures of body temperature and
respiration. As RB likely occurred in most, if not all, rodent inhalation toxicity studies involving
high level exposures to formaldehyde, this uncertainty is acknowledged and discussed in the
assessment, and for particular health outcomes it is specifically considered during study evaluation
(e.g., see description below regarding behavioral effects, since RB can affect activity).
Irritation, reflex bradypnea, and the RD50: A test for assessing sensory irritation was
developed by Yves Alarie in the 1960s. In an Alarie test, rodent respiration is measured before,
during and after exposure to one or more concentrations of an irritant, and then respiratory
depression (RD) is statistically quantified. RD is followed by a subscript that gives the percentage
of respiratory depression (e.g., RDo, RD20, RD50, RD70, etc.) The most commonly reported value in
Alarie tests is the RD50—the concentration of an irritating chemical that causes a 50% depression in
the respiratory rate {ASTM, 2012, }; fKane etal.. 19791.
"Irritation" refers to two distinct processes. The first process is sensory irritation of nerve
endings. URT irritation of the trigeminal nerve, which humans perceive as a burning or stinging
sensation, is what triggers RB in rodents. The second process relates to an inflammatory response
elicited by an irritating chemical, which is manifested by histopathologic changes such as local
redness, edema, pruritus, and cellular alterations. Sensory irritation may prevent histopathologic
damage through avoidance or through RB in rodents. Bos etal. (2002) found no correlation
between chemical concentrations that cause sensory irritation (as measured by the Alarie test) and
concentrations that induce histopathological changes. For a variety of irritants, the lowest
concentration that induces nasal histopathologic lesions can range from 0.3 x RD50 to more than 3 x
RDso-
Alarie tests are useful for (1) identifying chemicals which are URT sensory irritants, (2)
quantifying irritating concentrations, and (3) ranking chemicals for their irritancy potential. Alarie
(1981) proposed using0.03x RD50 values to predict threshold limit values (TLVs: typically used to
define workplace exposures that can be repeatedly encountered without adverse effects) for a
variety of irritants. More recently, {Nielsen, 2007, 992980 proposed the use of animal RD50 and RD0
values along with human data in a weight-of-evidence approach to predict acute or short-term
TLVs, the RDo being a threshold or NOEL for decreased respiratory rate.
Tables 16 and 17 present formaldehyde RD values from several Alarie studies for mice and
rats, respectively.10 No RD values exist for female mice or rats. Across the literature, there is fairly
good agreement on RD50 values for various strains of mice:
10Several studies cited in Tables 16 and 17 tested formalin, which means the animals were co-exposed to
formaldehyde and methanol. Considering that methanol's mouse RD50 of 54,963 mg/m3 (41,514 ppm) is 10,000
times greater than formaldehyde's mouse RD50, methanol was likely to have a negligible impact on the
formaldehyde RD values {Nielsen, 2007,}..
This document is a draft for review purposes only and does not constitute Agency policy.
A-81 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-16. Formaldehyde respiratory depression (RD) values for several
mouse strains and exposure durations
Study
Mouse strain
Exposure
(min)
RD50
(mg/m3)
RD10
(mg/m3)
RDo
(mg/m3)
Kane and Alarie (1977)
cf Swiss-Webster
10
3.8
O
In
Qj
0.31a
Nielsen et al. (1999)
cf BALB/c
10
4.9
0.4
Barrow et al. (1983)
cf B6C3F1
10
5.4
0.9*
0.49*
Chang et al. (1981)
cf B6C3F1
10
6.0
-
-
de Ceaurriz et al. (1981)
cf Swiss OFi
5
6.5
-
-
aValue derived from a graph.
1 Figure A-20 shows that rats are less responsive to URT irritants than mice, which is why
2 rats have higher RD50 values than mice:
Table A-17. Formaldehyde respiratory depression (RD) values for several rat
strains and exposure durations.
Study
Rat strain
Exposure
(min)
RD50
(mg/m3)
RD10
(mg/m3)
RDo
(mg/m3)
Cassee et al. (1996a)
cf Wistar
30
12.3
-
-
Barrow et al. (1983)
cf F-344
10
16.1
1.2a
-
Gardner et al. (1985)
cf Crl-CD
15
17.0
-
-
Chang et al. (1981)
cf F-344
10
39.0
-
-
aValue derived from a graph.
3 Tolerance: Nearly all rodent studies that assessed RB are acute Alarie tests lasting no more
4 than a few minutes or hours. There are no long-term studies that investigated whether-or-when
5 rodents develop a tolerance to formaldehyde or other irritants and eventually begin to breathe
6 normally. Mouse studies are a particular concern because mice have a greater RB response than
7 rats and are able to sustain bradypnea and hypothermia for a longer period than rats. The bulleted
8 short-term (4 days to 4 weeks) studies below examined the potential for rodents to develop
9 tolerance to formaldehyde and cyfluthrin. The formaldehyde studies show no sign of tolerance
10 over 10 days of exposure at concentrations as high as 18 mg/m3, but what happens after 10 days
11 remains unknown.
12 • Kane and Alarie (19771 observed a progressive decrease in respiratory rate (i.e., a
13 progressively greater RB response) over 4 days of formaldehyde exposure in Swiss-
14 Webster mice exposed to an RD50 of 3.8 mg/m3. A similar lack of tolerance was also seen in
15 mice exposed to acrolein (an aldehyde) at an RD50 of 3.9 mg/m3.
This document is a draft for review purposes only and does not constitute Agency policy.
A-82 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
• Chang etal. f!9831 exposed mice and rats to 6.9 or 17.6 mg/m3 formaldehyde (two of the
concentrations used in the Battelle carcinogenicity study) 6 hours/day for 4 days. On day 4,
both mice and rats showed concentration-related decreases in respiratory rate and minute
volume, but the decreases in mice were markedly greater (see Figure A-20).
• Chang and Barrow (19841 observed no tolerance in F-344 rats exposed to 18 mg/m3
formaldehyde for 10 days. Tolerance was observed in rats exposed over 4 days to a very
high formaldehyde concentration of 34 mg/m3, likely due to destruction or downregulation
of sensory trigeminal nerve endings or receptors, respectively.
• fPauluhn. 19981 exposed Wistar rats 6 hours/day, 5 days/week for 4 weeks to cyfluthrin, a
pyrethroid URT irritant, at the acute RD50 concentration of 47 mg/m3. Mean decreases in
respiratory rate were 45% at week 2 and 55% at week 4, that is, there was no sign of
tolerance. Since formaldehyde and cyfluthrin are both URT irritants, it is likely that similar
results might be seen with formaldehyde.
Reflex bradypnea and interpreting health effects data: Current testing guidelines do not
require examination of RB-related endpoints, and reduced inhaled rodent exposure may complicate
interpretations regarding inferences of potential human risk. For example, Battelle's
carcinogenicity study illustrates an apparent role of RB in long-term studies. The study authors
observed a disparity in formaldehyde-induced squamous metaplasia and inflammation between
B6C3F1 mice and F-344 rats. Both species were identically exposed in whole-body chambers at
analytical concentrations of 0, 2.5, 6.9, or 17.6 mg/m3. At comparable concentrations, nasal lesions
were much less severe in mice than in rats. In fact, incidences of squamous cell carcinoma were
similar in rats exposed at 6.9 mg/m3 and in mice exposed at 17.6 mg/m3—a difference in
concentration of more than 2-fold (Kerns etal.. 19831. Kerns et al. reasoned this 2-fold difference
between mice and rats may be due to "their physiological responses to formaldehyde inhalation,"
that is, due to RB. To support their hypothesis, they cited a 4-day Alarie test by Chang et al. f 1983:
described in the bullet above) in which the reduction in minute volume was 2-fold greater in mice
than in rats when exposed at 17.6 mg/m3 (see Figure A-20). In other words, the rats exposed at 6.9
mg/m3 and the mice exposed at 17.6 mg/m3 may have had similar lesion incidences because they
were exposed to approximately the same inhaled "dose" of formaldehyde due to RB.
The hypothesis offered by Kerns etal. (1983) that mice in the Battelle study inhaled about
half as much formaldehyde as rats at 17.6 mg/m3 due to RB, is logical and compelling, but there are
no long-term RB data to support it at this time. Thus, although it might be considered appropriate
to adjust a rodent POD to account for potential decreases in respiration (thus inferring that use of
the exposure levels and corresponding results of that study may not be health protective for
humans), this approach was not applied in this assessment. Overall, the lack of a long-term study to
determine whether-or when rodents eventually develop tolerance to formaldehyde or any other
URT irritant represents a significant data gap.
The potential impact of reflex bradypnea on behavioral studies: The normal
physiological effects of RB can complicate the interpretation of behavioral studies in rodents.
This document is a draft for review purposes only and does not constitute Agency policy.
A-83 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
Hypothermia causes reduced peripheral nerve conduction velocity due to an apparent reduced flux
of potassium and chloride ions across axon membranes. Hypothermia also causes prolonged
synaptic delay time at neuromuscular junctions. A progressive decrease in body temperature
results in ataxia, loss of fine motor control and reflexes, a reduction in cerebral blood flow and brain
function, and eventually a loss of consciousness {Mallet, 2002, }. Thus, what appear to be
chemically-induced behavioral effects may actually be partly attributable to RB-induced
hypothermia. Thus, the irritant effects were considered during evaluations of behavioral studies
(see Appendix A.5.7), including a preference for studies that allowed for a recovery time of at least
2 hours after exposure before testing, given the recovery parameters discussed above.
The impact of reflex bradypnea on developmental toxicity studies: Pregnant dams are
protected by RB, but their fetuses are not Fetuses can experience developmental delays or defects
due to impaired placental transfer of O2 (hypoxia) and CO2 (hypercapnia), fetal hypothermia, and
malnutrition. Fetuses do not tolerate hypothermia as well as adults {Pauluhn, 1989, }.
When dams experience RB, their fetuses may experience hypoxia due to (1) reduced
maternal respiration and (2) a left shift in maternal oxyhemoglobin affinity caused by an increase in
blood pH (respiratory alkalosis). Normal oxygen exchange to the fetus requires a gradient between
maternal and fetal oxyhemoglobin affinities. When pregnant dams experience RB, their blood pH
becomes more alkaline, resulting in a left shift in maternal oxyhemoglobin affinity. A maternal left
shift results in the affinities of maternal and fetal oxyhemoglobin being indistinguishable, which
impairs oxygen exchange to the fetus (hypoxia) and removal of CO2 (hypercapnia). {Rossant&
Cross, 2001, } describe hypoxia as a normal regulator of placental development in both humans and
mice.
When {Holzum, 1994,} exposed pregnant rats to cyfluthrin, they observed concentration-
related decreases in fetal weights (see Figure A-23); Holzum etal. also observed concentration-
related decreases in placental weights. Clearly, further studies on the impact of formaldehyde and
other URT irritants on the placenta and fetus are needed, but the results of Holzum et al. show how
RB has the potential to delay fetal growth. It should be noted that reductions in maternal feeding
and metabolism during periods of RB can result in reduced fetal glucose levels. It is also important
to emphasize that RB-induced developmental effects caused by fetal hypoxia, hypercapnia,
hypothermia, and malnutrition are not relevant to humans.
This document is a draft for review purposes only and does not constitute Agency policy.
A-84 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for Formaldehyde—Inhalation
Relative weight of placentas and fetuses
vehi 0.46 2.55 11.9 12.8+02
Concentration [mg Cyfluthrin/mJ]
Figure A-21. This graph demonstrates the impact of RB on fetal development.
It shows concentration-related decreases in placental and fetal weights in pregnant
dams exposed to cyfluthrin, a pyrethroid insecticide. Note that the decrements in
fetal and placental weights were lessened in the 12.8 mg/m3 group when the dams
were provided with oxygen-rich air (39% O2).
Source: {Holzum, 1994,}. Graph generated by Jurgen Pauluhn (Bayer Healthcare AG, Germany).
Summary: Reflex bradypnea (RB) is a protective response observed in rodents exposed to
formaldehyde and other upper respiratory tract irritants. The most notable signs of RB are
concentration-related decreases in body temperature, respiratory rate (breaths/minute), and
minute volume. Even though the effects of RB can be striking they are not relevant to humans. It is
likely that RB occurred in most, if not all, rodent inhalation toxicity studies testing high levels of
formaldehyde exposure, but the extent of RB in these studies cannot be ascertained since it was not
measured. In comparative studies, mice exhibit RB at a lower formaldehyde concentration than
rats and had a more pronounced and more sustained RB response than rats.
Because rodents experiencing RB have reduced minute volumes, they inhale less
formaldehyde and thus are expected to experience less toxicity than if they were breathing
normally. Several studies demonstrate that mice and rats do not develop tolerance to
formaldehyde over as much as 10 days of exposure; however, there are no long-term studies that
show whether-or-when rodents eventually develop a tolerance to formaldehyde. This is a
significant data gap. Thus, while RB is considered during study evaluation and during evidence
synthesis and integration, adjustments are not applied to account for the potential impact of RB on
long-term rodent health endpoints considered for use in dose-response analysis.
A.4. GENOTOXICITY
The evaluations of genotoxic effects of formaldehyde exposure included primary sources
from peer-reviewed literature and secondary sources of peer-reviewed reports by other federal
This document is a draft for review purposes only and does not constitute Agency policy.
A-85 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
agencies and non-federal institutions (see Section A.4.7), although a systematic literature search
was not conducted. In general, the following criteria were considered for making judgments about
evidence for the genotoxic and/or mutagenic potential of formaldehyde. These include but are not
limited to: (a) nature and type of tests, (b) degree of response, (c) number and performance of test
strains, (d) dose/concentration levels, (e) biological significance, (f) strength of evidence
(conflicting evidence in the same assay system for the same end point), and (g) evaluation of the
study results across the same end points. Studies of genotoxicity in exposed humans were
consistently evaluated using a structured set of criteria (see Section A.4.7).
The terms genotoxicity and mutagenicity differ depending on the effect seen on DNA.
Genotoxicity refers to potentially harmful effects caused either directly or indirectly to the genetic
material by chemical or physical agents, and these effects are not necessarily persistent and
transmissible and may or may not be associated with mutagenicity. Mutagenicity refers to the
induction of permanent, transmissible changes in the amount, chemical properties, or structure of
the genetic material. Mutations may involve a single gene or gene segment, a block of genes, parts
of chromosomes, or whole chromosomes and result in either structural and/or numeric changes.
Since mutagenicity is considered a subset of gentoxic effects, the term "genotoxic effects" will be
generally used through out the rest of the document unless the assay determines specific
mutations.
A variety of genotoxic effects have been demonstrated in both in vitro and in vivo test
systems as a result of exposure to formaldehyde (a Summary Table by Genotoxic Endpoint is
presented in Section A.4.7). Note that no single genotoxicity or mutagenicity test/system or study
is able to detect the entire spectrum of formaldehyde-induced genotoxic events. Therefore,
genotoxic endpoints are briefly discussed for cell free systems, prokaryotic organisms,
nonmammalian organisms, in vitro mammalian systems, in vivo experimental animals, and humans
[reviewed in fNTP. 2010: 2008: IARC. 2006a: Liteplo and Meek. 2003: Conawav etal.. 1996:1ARC.
1995: Ma and Harris. 1988: Auerbach etal.. 1977). In addition, the overall weight of evidence for
formaldehyde-induced mutations is considered in the context of the current EPA cancer guidelines
(U.S. EPA. 2005). Note that all studies from the available database have been depicted in several of
the following tables, but only the studies most relevant to this discussion are briefly described in
the text
A.4.1. Genotoxicity of Formaldehyde in Cell-Free Systems
Formaldehyde or formalin11 has been shown to form both hydroxymethyl DNA (hmDNA)
adducts and DNA-protein crosslinks (DPX or DPC) following treatment of various cell-free systems
with formaldehyde or formalin (see Table A-18). The formation of DNA-DNA crosslinks were
observed in calf thymus DNA fChaw etal.. 19801 and duplex DNA fHuang and Hopkins. 1993:
nStuclies that used formalin often contained 10-15% methanol as a stabilizing agent. Although formaldehyde is a
metabolic product of methanol, it is not genotoxic in in vitro reactions.
This document is a draft for review purposes only and does not constitute Agency policy.
A-86 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information for Formaldehyde—Inhalation
Huang etal.. 19921. Furthermore, DNA-protein crosslinks were seen in plasmid DNA, calf thymus
histones, and other acelluar systems fKuvkendall and Bogdanffv. 19921 Lu, 2009,1639318; Lu,
2008, 626083; Lu, 2010, 383598}. The formation of hmDNA adducts was observed following in
vitro reaction of formalin in solution with free DNA ribonucleoside fKennedv etal.. 19961.
deoxyribonucleosides and nucleotides f Cheng etal.. 2008: Cheng etal.. 2003: Mcghee and von
Hippel. 1975a. b), calf thymus DNA (Fennell. 1994b: Beland etal.. 1984: Von Hippel and Wong.
19711. human placental DNA (Zhong and Hee. 20041. and isolated rat liver nuclei (Fennell. 1994a:
Heck and Casanova. 19871. Cheng etal. (20081 also reported that nitrosamines which form
formaldehyde during their metabolism via formation of a-esters can react in vitro with
deoxyribonucleosides or calf thymus DNA and form the hmDNA adducts. Studies have shown that
N6-hydroxymethyl-deoxyadenosine (N6-hmdAdo) was the predominant adduct formed followed by
N2-hydroxymethyl-deoxyguanosine (N2-hmdGuo) and N4-hydroxymethyl-deoxycytidine (N4-
hmdCyd) when formaldehyde was reacted with calf thymus DNA f Cheng etal.. 2008: Beland etal..
1984) or human placental DNA (Zhong and Hee. 2004).
Table A-18. Summary of genotoxicity of formaldehyde in cell-free systems
Test system
Dose and Agent3
Results'5
Duration; Method
Reference
DNA-DNA crosslinks
Calf thymus DNA
0.17 mM 37% HCHO
+
40 days; RP-HPLC
Chaw et al., (1980)
Duplex DNA
25 mM HCHO
+
9 days; DPAGE
(Huang et al.,
1992)
Duplex DNA
25 mM HCHO
+
9 days; DPAGE
(Huang and
Hopkins, 1993)
DNA-protein crosslinks
Lysine or Cysteine and dG
50 mM 20% HCHO in H20
+
48 hours; RP-
HPLC/LC_MS
Lu et al. (2010)
Histone 4
50 mM 20% HCHO in H20
+
10 min; LC-MS
Lu et al. (2008a)
Plasmid DNA, calf thymus
histones
0.0015 mM HCHO
+
1 hr; filter binding assay
Kuykendall and
Bogdanffv, (1992)
Calf thymus DNA
0.5 mM HCHO
+
4 hours; ESI-MS/MS
(Lu. 2009)
DNA adducts
Guanosine
2400 mM 37% HCHO
+
48 hours
Kennedy et al. (1996)
Deoxyguanosine
2300 mM formalinc
+
20 hours
Cheng et al.
(2003)
Guanosine
0.001 mM HCHO
+
90 hours
Cheng et al.
(2003)
This document is a draft for review purposes only and does not constitute Agency policy.
A-87 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose and Agent3
Results'5
Duration; Method
Reference
DNA nucleosides/ nucleotides
50 mM formalin
+
72-120 hours
Mcghee and von
Hippel (1975a)
DNA nucleosides/ nucleotides
300 mM formalin
+
72-120 hours
Mcghee and von
Hippel (1975a)
Calf thymus DNA
0.001 mM formalin
+
90 hours
Cheng et al.
(2003)
Calf thymus DNA
0.167 mM formalin
+
48 hours
Beland et al., (1984)
Calf thymus DNA
0.4 mM formalin
+
4 hours
Fennell, (1994a)
Calf thymus DNA
200 mM formalin
+
20 hours
(Von Hippel and
Wong. 1971)
Calf thymus DNA or
deoxyribonucleosides
50 mM a-acetates of NDMA;
NNKand NNALd
+
1 or 90 hours
Cheng et al. (2008)
Human placental DNA
3.34 mM formalin
+
20 hours
Zhong and Hee
(2004)
Rat - Hepatic nuclei
0.1 mM HCHO (14C and 3H)
aqueous solution
+
0.5 hr
Heck and Casanova
(1987)
Rat - Hepatic nuclei
0.4 mM 14C-HCHO
+
4 hours
Fennell (1994a)
alowest effective concentration for positive results; highest concentration tested for negative or equivocal results.
b+ = positive, all experiments performed without exogenous activation.
cFormalin - all experiments with formalin contained 37% formaldehyde plus 10-15% methanol.
dthese nitrosamines are precursors to formaldehyde.
Abbreviations: HCHO, formaldehyde; NDMA, N-nitrosodimethylamine; NNK, 4-(methylnitrosamino)-l-(3-pyridyl)-l-
butanone; NNAL, 4-(methylnitrosamino)-l-(3-pyridyl)-l-butanol; DPAGE, denaturing polyacrylamide gel
electrophoresis; HPLC, high performance liquid chromatography; LC-ESI-MS, liquid chromatography electrospray
ionization mass spectrometry; LSC, liquid scintillation counting; MS, mass spectrometry; NMR, nuclear magnetic
resonance; RP-HPLC, reverse phase high performance liquid chromatography; UV, ultraviolet.
1 A.4.2. Genotoxicity of Formaldehyde in Prokaryotic Organisms
2 A number of reports describe the mutagenicity of formaldehyde in bacterial test systems
3 (Salmonella typhimurium and Eschericia coli) using reverse and forward mutation assays as well as
4 assays with specific E. coli strains for detecting deletions, insertions and point mutations
5 (see Table A-19).
6 Formaldehyde was mutagenic in the reverse mutation assay in all of the studies with the
7 Salmonella strains TA102 and TA104, and most of the studies with TA100 strains with and without
8 metabolic activation and in strains TA2638 and TA2638a without metabolic activation. Mixed
9 results were reported with TA97, TA98, and TA1537 strains, while most of the studies with the
10 TA1535 and TA1538 strains were negative with or without metabolic activation. (Sarrif etal..
This document is a draft for review purposes only and does not constitute Agency policy.
A-88 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
1997: Mtiller etal.. 1993: Tung etal.. 1992: Wilcox etal.. 1990: Marnett et al.. 1985) (Rvden etal..
2000: Le Curieuxetal.. 1993: O'Donovan and Mee. 1993: TemcharoenandThillv. 19831.
With respect to forward mutations, formaldehyde has been shown to induce these types of
mutations both in S. typhimurium fTemcharoen and Thillv. 19831 as well as in E. coli strains
fBosworth etal.. 1987: Temcharoen and Thillv. 19831. Temcharoen and Thilly (1983) showed that
formaldehyde induced both toxicity and mutagenicity in the Salmonella strain TM677 (8-
azaguanine sensitive), both with or without metabolic activation. On the other hand, Bosworth et
al. (19871 reported formaldehyde to be mutagenic in E. coli strain D494 uvrB, a more sensitive
strain to base-pair substitutions. Furthermore, formaldehyde has been shown to induce diverse
mutations in a forward mutation assay in E. coli strains GP120, GP120A, 7-2, and 33694, which
contained a xanthine guanine phosphoribosyl transferase {gpt) reporter gene fCrosby et al.. 19881.
In this study, formaldehyde tested at two different concentrations (4 and 40 mM) produced point
mutations (41%), deletions (18%), and insertions (41%) at low concentrations of exposure, while
the high-dose exposure resulted predominantly in point mutations (92%). The point mutations at
low-dose exposure were transversions at GC base pairs, while at high-dose exposure they were
transition mutations at a single AT base pair in the gpt gene fCrosbv etal.. 19881.
Wang etal. f20071 have also shown that formaldehyde causes dose-dependent increase in
microsatellite instability in E. coli. Exposure to 2.5 mM formaldehyde caused a 2- to 24-fold
induction in mutation frequencies of the complementary dinucleotide repeat microsatellites (GpT)
and (ApC) compared to in untreated controls. It is possible that microsatellite instability could
change the conformation of DNA to Z-DNA structure, making the DNA not amenable for DNA repair.
Table A-19. Summary of genotoxicity of formaldehyde in prokaryotic systems
Test system
Dose3
(ng/
plate)
Agentb
Resultsc'd
Comments
Reference
-S9
+S9
Reverse mutation
S. typhimurium
TA100
10, 25
35% HCHO sol.
+
+
PP method; values
visually determined
from graph; (T) at 37.5
(-S9) and 50 (+S9)
I^Lg/plate
(Orstavik and
Hongslo, 1985)
12
37% HCHO with
10% methanol
(+)
(+)
PI method
Schmid et al. (1986)
15, 7.5
HCHO/ml
+
+
Suspension method
Sarrif et al. (1997)
30
37% HCHO with
10-15% methanol
+
+
PI method; values
visually determined
from graph. Methanol
tested '-ve1 up to 500
]ug/plate (-S9 or +S9)
in the same study.
Connor et al. (1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-89 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose3
(ng/
plate)
Agentb
Resultscd
Comments
Reference
-S9
+S9
30
HCHO (form not
specified)
(+)
ND
PP method
Takahashi et al.
(1985)
39
37% HCHO with
10-15% methanol
"(T)
"(T)
PI method
De Flora (1981)
50
35% HCHO
+
+
PP method; dose
range 6.25-50
]ug/plate only
provided
Dillon et al. (1998)
75
HCHO (form not
specified)
-
+
PI method; -S9 data
<2-fold compared to
control
Sarrif et al, (1997)
80
37% HCHO with
10% Methanol
(+)
+
PP method
Schmid et al. (1986)
90
HCHO (form not
specified)
-
ND
PP method; (T): >90
I^Lg/plate
Marnett et al. (1985)
100, 50
37% aq.sol. HCHO
+,+
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee
(1993)
100
HCHO (form not
specified)
+
-
PP method; (T) >200
I^Lg/plate
Sarrif et al. (1997)
150
37% HCHO
+
ND
PP method;
Discrepancy in
mutagenic data
observed between
author's report and
the graph from the
citation (150 vs. =30
M-g/plate)
Fiddler et al. (1984)
333.3, 10
37% HCHO
-
+
PP method; (T): NR
Haworth et al. (1983)
500, 20
37% HCHO in
distilled water
(+)
+
PP method
(Connor et al.,
1985a)
5. typhimurium
TA102
10
HCHO/mL
+
ND
Fluctuation test; (T) at
30 |ag/m L
Le Curieux et al.,
(1993)
17.2
HCHO (in water)
+
ND
PP method
Ryden et al., (2000)
25
HCHO (form not
specified)
+
ND
PI method; (T) >100
I^Lg/plate
Wilcox et al., (1990)
50
HCHO (form not
specified)
(+)
(+)
PP method; values
visually determined
from graph
(De Flora et al.,
1984)
50
35% HCHO
+
+
PP method;'+' with
rat S9 and '±' with
mouse S9; Authors
show a dose range
6.25-50 ]ug/plate.
Dillon et al., (1998)
This document is a draft for review purposes only and does not constitute Agency policy.
A-90 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose3
(ng/
plate)
Agentb
Resultscd
Comments
Reference
-S9
+S9
90
HCHO (form not
specified)
+
ND
PP method; (T): >90
fig/plate
Marnett et al. (1985)
200, 100
37% aq.sol. HCHO
+/ +
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee
(1993)
200
HCHO (in water)
+
ND
PI method; (T) at 600
mg/plate
Watanabe et al. (1996)
5000
HCHO (form not
specified)
(+)
(+)
PI method; (+) by 1 lab
and '-ve1 by 2 labs
Jung et al., (1992)
5000
HCHO (form not
specified)
(+)
(+)
PI method; reported
'(+) by one lab and
'-ve1 by 2 labs
Muller et al. (1993)
5. typhimurium
TA104
50
35% HCHO
+
+
PP method; Authors
show a dose range
6.25-50 ]ug/plate.
Dillon et al. (1998)
90
HCHO (form not
specified)
+
ND
PP method; (T): >90
fig/plate
Marnett et al. (1985)
5. typhimurium
TA1535
39
formalin
"(T)
"(T)
PI method
De Flora (1981)
100
37% aq.sol. HCHO
~} ~
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee
(1993)
100
HCHO (form not
specified)
-
-
PI method; (T) at 150
fig/plate
Sarrif et al. (1997)
100
HCHO (form not
specified)
-
-
PP method; (T) >200
fig/plate
Sarrif et al. (1997)
333.3
37%HCHO
-
-
PP method; (T): NR
Haworth et al. (1983)
5. typhimurium
TA97
50
HCHO (form not
specified)
+
ND
PI method; (T) at 100
fig/plate
Sarrif et al. (1997)
90
HCHO (form not
specified)
-
ND
PP method; (T): >90
fig/plate
Marnett et al. (1985)
5. typhimurium
TA98
10, 25
35% HCHO sol.
+
+
PP method; values
visually determined
from graph; (T) at 37.5
(-S9) and 50 (+S9)
fig/plate
Oerstavik and Hongslo
(1985)
30
37% HCHO with 10-
15%
methanolMethanol
+
+
PI method; Methanol
tested up to 500
mg/plate (-S9 or +S9)
was '-ve1. Values
visually determined
from graph.
Connor et al., (1983)
30
HCHO (form not
specified)
(+)
ND
PP method
Takahashi et al.
(1985)
39
37% HCHO with 10-
"(T)
"(T)
PI method
De Flora, (1981)
This document is a draft for review purposes only and does not constitute Agency policy.
A-91 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose3
(ng/
plate)
Agentb
Resultscd
Comments
Reference
-S9
+S9
15% methanol
50, 100
37% aq.sol. HCHO
+,+
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee,
(1993)
50, 100
HCHO (form not
specified)
+
+
PP method; (T) >00
fig/plate
Sarrif et al., (1997)
75
HCHO (form not
specified)
-
+
PI method; -S9 data
<2-fold compared to
control
Sarrif et al., (1997)
90
HCHO (form not
specified)
-
ND
PP method; (T): >90
fig/plate
Marnett et al., (1985)
333.3, 10
37% HCHO
-
(+)
PP method; (T): NR
Haworth et al., (1983)
500
37% HCHO in
distilled water
"(T)
(+)
(T)
PP method
Connor et al.
(1985b)
5. typhimurium
TA1537
39
37% HCHO with 10-
15% methanol
"(T)
"(T)
PI method
De Flora, (1981)
50, 75
HCHO (form not
specified)
+
+
PI method
Sarrif et al., (1997)
100
37% aq.sol. HCHO
~} ~
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee,
(1993)
100
HCHO
-
-
PP method
Sarrif et al., (1997)
333.3
37%HCHO
-
-
PP method; (T): NR
Haworth et al., (1983)
5. typhimurium
TA1538
39
formalin
"(T)
"(T)
PI method
De Flora, (1981)
100
37% aq.sol. HCHO
~} ~
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee,
(1993)
5. typhimurium
TA2638
500
HCHO (in water)
+
ND
PI method; (T) at 1000
mg/plate
Watanabe et al.,
(1996)
5. typhimurium
TA2638a
17.2
HCHO (in water)
+
ND
PP method
Ryden et al., (2000)
S. typhimurium
UTH8413, UTH8414
500
37% HCHO with
10-15%
methanolMethanol
"(T)
"(T)
PI method; Methanol
tested '-ve1 up to 500
]ug/plate with/without
S9.
Connor et al., (1983)
500
37% HCHO in
distilled water
"(T)
"(T)
PP method
Connor et al.
(1985b)
E. coli WP2,
WP2uvrA, H/R30R,
Hs30R (uvrA)
420
HCHO (form not
specified)
+
ND
RM assay
Takahashi et al.
(1985)
E. coli NG30 (recA)
63
HCHO (form not
specified)
-
ND
RM assay; values
visually determined
from graph
Takahashi et al.
(1985)
This document is a draft for review purposes only and does not constitute Agency policy.
A-92 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose3
(ng/
plate)
Agentb
Resultscd
Comments
Reference
-S9
+S9
E. co//016 (polA)
52.5
HCHO (form not
specified)
-
ND
RM assay; values
visually determined
from graph
Takahashi et al.
(1985)
E. coli K12
(AB1886)/(uvrA); K12
(AB2480)/(recA/uvrA)
150
HCHO (form not
specified)
-
ND
RM assay
Graves et al. (1994)
E. coli K12
(AB1157)(WT)
1875
HCHO (form not
specified)
+
ND
RM assay
Graves et al. (1994)
E. coli WP2 (pkMlOl)
200
HCHO (form not
specified)
"(T)
ND
PI method
Wilcox et al., (1990)
200, 100
37% aq.sol. HCHO
-/ +
ND
Results by PI & PP
methods, respectively
O'Donovan and Mee,
(1993)
700
HCHO (in water)
+
ND
PI method
Watanabe et al.,
(1996)
E. coli WP2 uvrA
(pkMlOl)
150
HCHO (form not
specified)
+
ND
PI method; dose-
response from 10-300
I^Lg/plate
Wilcox et al., (1990)
200, 50
37% aq.sol. HCHO
(form not
specified)
+,+
ND
Results by Results by
PI & PP methods,
respectively
O'Donovan and Mee,
(1993)
400
HCHO (in water)
+
ND
PI method
Watanabe et al.,
(1996)
E. coli (Lac+
reversion) WP3104P
10
HCHO (form not
specified)
(+)
ND
RM assay
Ohta et al., (1999)
E. coli (Lac+
reversion) WP3101P,
WP3102P, WP3103P,
WP3105P, WP3106P
30
HCHO (form not
specified)
-
ND
RM assay
Ohta et al., (1999)
Forward mutation
S. typhimurium
TM677
0.167, 0.33
mM
37% HCHO with
10-15% Methanol
+
+
PP method
(Temcharoen and
Thillv. 1983)
E. coli D494uwB
(pGW1700)
6.0 |J.g/mL
HCHO (form not
specified)
+
ND
Ampicillin FM assay
Bosworth et al., (1987)
Deletions, Insertions and Point mutations
E. coli GP120,
GP120A, 7-2, 33694
4 mM
HCHO (form not
specified)
+
ND
gpt FM assay
Crosby et al., (1988)
Microsatellite Instability
E. coli J M109
2.5 mM
HCHO (form not
specified)
+
ND
Mutation frequency
analysis and
sequencing.
Wang et al., (Wang et
al.. 2007)
Methanol
alowest effective dose for positive results; highest ineffective dose tested for negative or equivocal results
This document is a draft for review purposes only and does not constitute Agency policy.
A-93 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
bsingle value indicates identical dose/concentration effective for both without (-S9) or with (+S9) metabolic
activation; for -S9 assay data showing two signs (+ or -) separated by a comma indicate respectively, use of PI and
PP methods.
c+ = positive; - = negative; (+) = weak positive; ND = test was not done; (T), toxic.
Abreviations: HCHO, formaldehyde; PI, plate incorporation (or standard plate); PP, pre-incubation plate; FM,
forward mutation; RM, reverse mutation; gpt, xanthine guanine phosphoribosyl transferase.
A.4.3. Genotoxicity of Formaldehyde in Nonmammalian Systems
Formaldehyde (commercial grade) or formalin (mostly containing 37% formaldehyde and
10-15% methanol) has been tested in several nonmammalian systems including yeast, molds,
plants, insects, and nematodes. As summarized in Table A-20, formaldehyde has been shown to
cause gene conversion, strand breaks, crosslinks, homozygosis and related damage in yeasts
(Saccharomyces cerevisiae); forward and reverse mutations in molds (Neurospora crassa);
micronuclei formation in spiderworts (Tradescantia pallida); DNA damage and mutations in several
plants; genetic cross-over or recombination, sex-linked recessive lethal mutations, dominant lethal
mutations, heritable translocations, and gene mutations in insects (Drosophila melanogaster); and
recessive lethal mutations in nematodes (Caenorhabditis elegans). Formaldehyde failed to show
micronuclei formation in newt larvae (Pleurodeleswaltl) (reviewed in (IARC. 2012: NTP. 2010:
IARC. 2006a). DNA protein crosslinks were observed in Saccaromyces cerevisiae and E. coli
(Magana-Schwencke and Ekert. 1978: Magana-Schwencke and Moustacchi 1980: Wilkins and
McCleod 19761.
Some of the nonmammalian studies compared the effects of formaldehyde in wild type and
DNA repair-deficient organisms. For example, Magana-Schwencke etal. T19781 showed that
excision repair-deficient Saccharomyces cerevisiae strains are more susceptible to formaldehyde-
induced lethal effects and have reduced capacity to form single strand breaks (SSBs) compared
with repair-proficient strains, suggesting that the repair process possibly involves SSB formation.
Also, formaldehyde is more mutagenic in repair-deficient Neurospora crassa compared to the
corresponding repair-proficient strains fde Serres and Brockman. 19991.
Table A-20. Summary of genotoxicity studies for formaldehyde in
nonmammalian organisms
Test system
Concentration315
Results"
Commentsd
Reference
DNA damage
Various plant and
fungal species®
1233 mM 3.7%
HCHO (at pH 3.0
and 7.0)
+
1.5 hours, PCR/GE,
(Douglas and Rogers,
1998)
DNA protein crosslinks
Saccharomyces
cerevisiae
17 mM HCHO
(form not
specified)
+
0.25 hours, DNA
extractability; (T) 90 & 60%
survival at 33 & 66 mM
(Magana-Schwencke
and Ekert. 1978)
This document is a draft for review purposes only and does not constitute Agency policy.
A-94 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration315
Results"
Commentsd
Reference
5. cerevisiae
33 mM HCHO
(form not
specified)
+
HCHO with 42 & 95% DNA
damage, respectively
(Magana-Schwencke
and Moustacchi,
1980)
E. coli
130 mM HCHO
(form not
specified)
+
10 min; alkaline sucrose
gradient centrifugation
(Wilkins and
Macleod, 1976)
DNA repair inhibition
S. cerevisiae
66 mM HCHO
(form not
specified)
+
0.25 hours, ASG; (T) 90 &
60% survival at 33 & 66
mM HCHO with 42 & 95%
(Magana-Schwencke
and Ekert. 1978)
DNA damage, respectively
Dominant lethal mutation
Drosophila
larval feeding method,
(Auerbach and
melanogaster
60 mM 36% HCHO
frequency of hatchability
Moser, 1953a);
in water
Auerbach and Moser
(1953b)
D. melanogaster
43 mM HCHO
Exposure duration NR,
(Sram, 1970)
(form not
+
frequency of dominant
specified)
lethal mutations
Forward mutation
Neurospora crassa
heterokaryon H-59
strain
3 mM formalin
+
3 hours, frequency of ad-3
mutations
{de Serres, 1988,
1311638; de Serres,
1999, 1311639)
N. crassa
heterokaryon H-12
strain
8 mM formalin
(+)
3 hours, frequency of ad-3
mutations
{de Serres, 1988,
1311638; de Serres,
1999, 1311639)
Gene conversion
S. cerevisiae
18 mM 30% HCHO
0.5 hour, frequency of
{Chanet, 1975,1311646}
strain D4
recombinants
Genetic crossing over or recombination
D. melanogaster
14 mM HCHO
larval feeding method
(Sram, 1970)
(form not
+
specified)
42 mM HCHO
duration of exposure NR,
(Sobels, 1957,
(form not
+
frequency of recombinant
1311647(®(®author-vear}
specified)
83 mM HCHO
(form not
+
duration of exposure NR,
frequency of cross overs
(Ratnavake, 1970)
specified)
Heritable translocation
D. melanogaster
14 mM HCHO
2 hours, frequency of
(Khan. 1967)
(form not
+
recombinants
specified)
83 mM HCHO
(form not
+
duration of exposure NR,
frequency of
(Ratnavake, 1970)
specified)
translocations
Homozygosis by mitotic recombination or gene conversion
This document is a draft for review purposes only and does not constitute Agency policy.
A-95 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration315
Results"
Commentsd
Reference
Saccharomyces
0.62 mM formalin
16 hours, frequency of
(Zimmermann and
cerevisiae
+
resistant colonies
Mohr. 1992)
Micronucleus
Pleurodeles waltl
0.17 mM HCHO
168 hours, Masson's
(Siboulet et al., 1984)
(form not
-
haemalum staining
specified)
Pleurodeles waltl
0.33 mM HCHO
12 hours, Masson's
(Le Curieux et al.,
larva
(form not
specified)
—
haemalum staining
1993)
Tradescantia pallida
8 mM HCHO (form
6 hours, acetocarmine
(Batalha et al., 1999)
not specified)
staining
Mutation
Plants (others)
NR
+
NR
(Auerbach et al.,
1977)
Reverse lethal mutation
Caenorhabditis
elegans
23 mM HCHO from
PFA
+
4 hours, frequency of
mutations
(Johnsen and Baillie,
1988)
Reverse mutation
Neurospora crassa
10 mM HCHO
4 hours, frequency of
(Jensen et al., 1951)
(form not
+
mutations
specified)
10 mM formalin
-
3 hours, frequency of
mutations
(Kolmark and
Westergaard, 1953)
24 mM HCHO
0.5 hours, frequency of
(Dickev et al., 1949)
(form not
-
mutations
specified)
Sex-linked lethal mutation
D. melanogaster
8 mM formalin
+
larval feeding method,
frequency of sex linked
lethals
(Stumm-Tegethoff,
1969)
14 mM HCHO
larval feeding method
(Alderson, 1967)
(form not
+
specified)
14 mM HCHO
2 hours, frequency of
(Khan. 1967)
(form not
+
progeny
specified)
33 mM formalin
+
duration of exposure NR,
frequency of eclosions
(Kaplan. 1948)
42 mM HCHO
Exposure duration NR,
(Sobels and van
(form not
specified)
+
frequency of sex-linked
lethals
Steenis, 1957)
60 mM 36% HCHO
in water
+
larval feeding method,
frequency of sex linked
lethals
(Auerbach and
Moser, 1953b)
67 mM HCHO
(form not
(+)
larval feeding method,
frequency of sex linked
(Ratnavake, 1968)
specified)
lethals
This document is a draft for review purposes only and does not constitute Agency policy.
A-96 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration315
Results"
Commentsd
Reference
73 mM HCHO
(form not
specified)
+
duration of exposure NR,
frequency of sex-linked
lethals
(Ratnavake, 1970)
Single strand breaks
S. cerevisiae
33 mM HCHO
(form not
specified)
+
0.25 hours, ASG; (T) 90 &
60% survival at 33 & 66
mM HCHO with 42 & 95%
DNA damage, respectively
(Magana-Schwencke
et al.. 1978)
indicates lowest effective concentration for positive results; highest concentration tested for negative or
equivocal results.
indicates that the multiple dose/concentration values reported correspond to order of the indicated test result(s)
(e.g., without activation; with activation). Identical doses/concentrations for multiple test results are indicated by
a single value; otherwise are seperated by commas,
indicates + = positive; - = negative; (+) = weak positive.
indicates the duration of exposure and the assay used to assess the endpoint, dose-response and toxicity (T) if
any.
indicates that authors tested the following species: Agaricus bisporus, Glycine max, Lycopersicon esculentum,
Pinus resinosa, Pisum sativum, Populus x euramericana, Viciafaba, and Zea mays.
Abbreviations: ad-3, adenine-3 locus; ASG, alkaline sucrose gradient; HCHO, formaldehyde; NR, not reported;
PCR/GE, polymerase chain reaction/gel electrophoresis; PFA, paraformaldehyde.
A.4.4. Genotoxicity of Formaldehyde in in Vitro Mammalian Cells
Formaldehyde has been tested for its genotoxic potential in several mammalian cell culture
systems originating from rodents (mice, rats, hamsters) and humans, mostly without metabolic
activation. In a majority of these systems, formaldehyde tested positive for: DNA reactivity
including DNA adducts, DPXs, and SSBs; cytogenetic changes such as sister chromatid exchanges
(SCEs), chromosomal aberrations (CAs), and micronuclei (MN); cell transformation and mutation
induction; and other genotoxic endpoints such as unscheduled DNA synthesis (UDS) and DNA
repair inhibition (summarized in Table A-21).
DNA Reactivity and Damage
DNA adducts
Formaldehyde has been shown to form hmDNA adducts in CHO cells (Beland et al.. 1984)
and rat and human nasal epithelial cells (Zhong and Que Hee. 2004) as shown in Table A-21.
Beland et al. (1984) first reported hmDNA adducts in CHO cells incubated with 1 mM of
radiolabeled formaldehyde. After a 2-hour incubation, small amounts of N6-hmdA were detected
with concomitant metabolic incorporation of formaldehyde (i.e., into DNA bases). Zhong and Que
Hee (2004) reported three types hmDNA adducts in human nasal epithelial cells exposed to varying
concentrations of formalin (10-500 ng/mL). In this study, the hmDNA adduct levels were in the
order of N6-hmdA > N2-hmdG > N4-hmdC. In HeLa cells exposed to [13CD2]-formaldehyde,
(Lu,2012,1254607@@author-year) detected both exogenous (13C-labeled) and endogenous
This document is a draft for review purposes only and does not constitute Agency policy.
A-97 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
(unlabeled) N2-hmdG adducts; however, this study detected endogenous but not exogenous N6-
hmdA adducts.
DNA-protein crosslinks
As summarized in Table A-21, DNA protein crosslinks have been reported in several
mammalian cell lines (primary and transformed) from rodents (mice, rats, hamsters) and humans,
(reviewed in (IARC. 2006a: Conawav etal.. 1996: IARC. 1995).
The lowest effective concentration of formaldehyde or formalin causing DPX formation
varied between different cell lines (see Table A-21). Among the animal cell lines, DPX formation
was observed at the in vitro concentrations of 0.125-0.25 mM in CHO cells and 0.01-0.2 mM in V79
cells. Several human cell lines (either primary cells or developed cells lines), including epithelial,
fibroblasts, buccallymphoblastoid, lymphoma, and peripheral blood lymphocytes, among others,
that were exposed to formaldehyde also formed DPXs (Emri etal.. 2004: Li etal.. 2004: Costa etal..
1997: Craft etal.. 1987). Selected studies have been briefly described below, although all available
and relevant studies are included in Table A-21).
Craft et al. (1987) analyzed DPXs in TK6 human lymphoblastoid cells immediately after a 2-
hour exposure (zero time) to 0-600 |j.M formaldehyde with a significant nonlinear increase in DPXs
above 50 |iM, which correlated with the onset of cytotoxicity. DPXs were completely repaired
within 24 hours after exposure.
DPXs were also detected in Epstein-Barr Virus (EBV)-human Burkitt's lymphoma cells
exposed to paraformaldehyde (which depolymerizes to release formaldehyde) at doses that were
cytotoxic (>0.003%) (Costa etal.. 1997). Grafstrom etal. (1986) reported that the number of DPXs
induced by 100 |iM formaldehyde in vitro in human bronchial epithelial cells and fibroblasts was
similar; although, DPX levels were several-fold higher than SSBs in the epithelial cells. In a different
study, the same authors fGrafstrom et al.. 19841 noted that formaldehyde exposure resulted in the
formation of DPXs at similar levels in bronchial epithelial cells and in DNA excision repair-deficient
xeroderma pigmentosum (XP) skin fibroblasts, and their removal rate was similar with a half-life of
2-3 hours, suggesting that the DPX are repaired independently of the excision repair. Further,
formaldehyde was only moderately cytotoxic to normal bronchial epithelial cells and fibroblasts at
concentrations that induced substantial DNA damage. Repair of the formaldehyde-induced DNA
SSBs and DPXs appeared to be inhibited by the continued presence of formaldehyde in the culture
medium fGrafstrom etal.. 19841.
A linear increase in DPX levels was observed in primary human skin fibroblasts and
keratinocytes from 25-100 |iM formaldehyde, as indicated by the ability of formaldehyde to reduce
DNA migration in the comet assay after methylmethane sulfonate (MMS) pretreatment {Emri,
2004, 626272}. Similar findings were also reported for primary human peripheral blood
lymphocytes (PBLs) and HeLa cells fLiu etal.. 20061. Peak response for SSBs was seen at 10 |j.M in
both cells, with higher concentrations resulting in crosslink formation fLiu etal.. 20061. DPX
formation was also observed in whole blood culture after exposure to 25 |iM, as indicated by the
This document is a draft for review purposes only and does not constitute Agency policy.
A-98 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
affect of formaldehyde on DNA migration in the comet assay after y-radiation fSchmid and Speit.
20071. The repair of DPX was complete 8 hours after an exposure to 100 [J.M formaldehyde, while
DPX formed at >200 mM were repaired within 24 hours.
Formaldehyde-induced DPXs are removed either through spontaneous hydrolysis or active
repair processes fOuievrvn and Zhitkovich. 20001. Inhibition of specific proteosomes (protein
complexes involved in degrading unwanted or damaged proteins) in xeroderma pigmentosum
(XP)-A cells inhibited DPX repair, thereby supporting the role of enzymatic degradation fOuievryn
and Zhitkovich. 20001. The average half-life of formaldehyde-induced DPXs in human epithelial cell
lines was 12.5 hours (range 11.6 to 13 hourslfOuievrvn and Zhitkovich. 20001.18 hours in HeLa
cells (Liu et al., 2006), and 24 hours in human lymphoblasts (Craft etal.. 19871. This difference was
primarily due to slower active repair of DPXs, with a t1/2 of 66.6 hours for human lymphocytes
compared to other human cell lines fOuievryn and Zhitkovich. 20001.
Speit et al., (2000) hypothesized that single peptides or small peptide chains cross-linked to
DNA are critical to formaldehyde-induced mutation. However, these authors did not find significant
differences in the induction and repair of DPXs in a normal human cell line (MRC4CV1), nucleotide
excision repair (NER)-deficient xeroderma pigmentosum (XP) fibroblast cell line, and a Fanconi
anemia (FA) cell line exposed to 125-500 |j.M formaldehyde for 2 hours. In contrast, these cells
showed increased susceptibility to formaldehyde-induced MN formation. It is suggested that the
NER pathway affects cytogenetic makers of genotoxicity rather than the cross-link repair (Speit et
al.. 20001.
DNA Single Strand Breaks (SSBs)
Formaldehyde has been shown to induce SSBs in a number of mammalian cell systems in
vitro (see Table A-21). Certain cell lines seem to be more sensitive for SSB formation than others.
For example, formaldehyde induced SSBs at concentrations ranging from 0.005-0.8 mM in human
primary cells including lung/bronchial epithelial cells (Grafstrom. 1990: Saladino etal.. 1985:
Grafstrom etal.. 1984: Fornace etal.. 1982). skin fibroblasts (Snyder and van Houten. 1986:
Grafstrom etal.. 1984). lymphocytes (Liu etal.. 2006). and in human cell lines A549 (Vock etal..
1999) and HeLa (Liu etal.. 2006) cells, and rathepatocytes (Demkowicz-Dobrzanski and
Castonguav. 19921. In many of these studies SSB induction was dose-dependent. However,
formaldehyde did not induce SSBs in human foreskin fibroblasts (Snvder and van Houten. 1986).
human skin keratinocytes exposed for 20 hrs {Emri et al., 2004, 626373}, mouse leukemia cells
(Ross etal.. 1981: Ross and Shipley. 1980) and hamster CHO cells (Marinari etal.. 1984) and V79
cells (Speitetal.. 2007b).
Formaldehyde induces more DPX than SSBs in normal human bronchial epithelial cells
(Grafstrom. 1990: Saladino etal.. 1985). Grafstrom et al. (1984) examined the kinetics of DNA
repair in nucleotide excision repair (NER)-proficient human bronchial epithelial cells and
fibroblasts and NER-deficient fibroblasts from XP patients by alkaline elution technique. They
reported comparable levels of DPX in all cell lines, suggesting non-involvement of NER in DPX
This document is a draft for review purposes only and does not constitute Agency policy.
A-99 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
removal. However, the SSB levels are higher than DPX in XP cells compared to the normal
fibroblasts, although both these DNA lesions are repaired at comparable rates, suggesting an
additional indirect mechanism of SSB formation possibly involving a different repair pathway. SSBs
in HeLa cells induced by 10 |iM formaldehyde were repaired by 90 minutes after cells were washed
to remove formaldehyde fLiu etal.. 20061.
Cytogenetic markers of genotoxicity
Clastogenic effects, including increased MN, CAs, and SCEs, have been reported in a variety
of in vitro systems as shown in Table A-21.
Micronucleus fMNl formation
Studies have shown MN formation either in V79 lung epithelial cell lines fSpeitetal.. 2007b:
Merk and Speit. 1998). in human fibroblasts with varying DNA repair backgrounds (Speit etal..
2000). or in whole blood cultures (Schmid and Speit. 2007). Speit et al. (2000) reported a higher
frequency of MN formation in XP and FA cell lines compared to normal human cell lines suggesting
the importance of NER and crosslink repair following formaldehyde exposure. In V79 cells, Speit et
al. (2007b)observed that MN frequency increased with repeated formaldehyde treatments
compared to a single treatment; however, such an increase was not observed if the treatment
interval was increased to 24 hours. An increase in micronucleus frequency was observed in mouse
erythropoietic cells (Ti etal.. 2014). human A549 lung epithelial cells (Speit etal.. 2011a). human
lymphoblasts {Ren, 2013,15783392}, and human whole blood cultures (Speitetal.. 2011a).
Schmid and Speit (2007) observed a statistically significant increase in MN formation at or
above a formaldehyde concentration of 300 |iM in human whole blood cultures treated with
formaldehyde 24 hours after the start of the culture and cytochalasin B (CytB) added 20 hours later
(44 hours after the start of the culture). This prompted the conclusion that the level of DPX
formation from formaldehyde exposure would need to be high for MN formation and the cells must
be exposed after the first mitosis (which is 24 hours). In examining MN formation more closely
with Fluorescence In Situ Hybridization (FISH), Schmid and Speit (20071 found that 81 percent of
the time, formaldehyde was inducing a micronuclei that was centromere negative indicating the
effect to be clastogenic rather than aneugenic (a centromere contained micronuclei).
Sister chromatid exchanges (SCEs)
Sister chromatid exchanges occur as a result of errors in replication process, where an
exchange in the chromatids between sister chromatids occurs during the anaphase. DPX are likely
to cause replication block and might stimulate SCEs in cells. Therefore, evaluation of SCEs is
important in assessing the genotoxicity of formaldehyde.
Formaldehyde has been shown to induce SCEs in most of the in vitro studies, both in rodent
and human cells. The available studies are summarized in Table A-21. Different cell types
responded differently for various concentrations for formaldehyde, particularly at low doses. For
This document is a draft for review purposes only and does not constitute Agency policy.
A-100 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
example, the lowest effective concentration (LEC) of formaldehyde in Chinese hamster embryo cells
was 0.01 mM, for CHO cells it was 0.03 mM, and V79 cells responded at a concentration of 0.06 mM,
while human lymphocytes required slightly higher concentrations (0.125 mM) to show any effect
Neuss and Speit (2008) observed a significant dose-dependent increase in SCE formation in V79
cells and A549 cells following a range of formaldehyde concentrations with 0.1 mM being the LEC
when BrdU was added immediately after formaldehyde exposure. However, when BrdU addition
was delayed by 4 hours the LEC increased to 0.2 mM suggesting DNA repair. In co-cultivation
experiments, the authors first treated A549 cells for 1 hr with 0.05 mM formaldehyde and then co-
cultured them with V79 cells with or witout changing the culture medium, SCEs were observed in
A549 cells in both situations, but in the co-cultured V79 cells, SCEs were observed only when the
medium was not changed, suggesting residual availability of formaldehyde in the medium to induce
SCEs in V79 cells and that formaldehyde which entered the A549 cells is either utilized or
inactivated. Miyachi and Tsutsui (2005) measured the induction of SCEs in Syrian hamster embryo
(SHE) cells at an LEC of 0.01 mM within an hour of formaldehyde exposure. Schmid and Speit
(2007) observed that SCEs were induced by 200 [J.M in lymphocytes from human whole blood
cultures, an effect apparently associated with cytotoxicity as indicated by a concomitant reduction
in the proliferative index.
Chromosomal aberrations fCAsl
Several studies have demonstrated formaldehyde-induced CAs in a variety of mammalian
cells, such as CHO cells (Garcia etal.. 2009: Nataraian et al.. 1983). Chinese hamster lung fibroblasts
(Ishidate etal.. 1981). Syrian hamster embryo (SHE) cells (Hagiwara etal.. 2006: Hikiba etal..
20051. mouse lymphoma cells fSpeitand Merk. 20021. human PBLs fDresp and Bauchinger. 1988:
Schmid etal.. 1986). and human fibroblasts (Levy etal.. 1983).
Hikiba et al., (2005.) used SHE cells to measure the induction of CAs following exposure to a
series of formaldehyde concentrations (0, 33, 66, and 99 |j.M) for 24 hours and observed the
percentages of aberrant metaphases to be 0, 6, 6, and 71, respectively. The aberrations were
predominantly chromosome gaps and chromosomal breaks and exchanges. The relative colony-
forming efficiency remained high (at least 85%). Dose-dependent increases in chromosomal
aberrations were observed when CHO cells were exposed to 0.15mM of commercial formaldehyde
(Garcia etal.. 2009). Chinese hamster lung fibroblasts, when exposed to 0.6 mM formalin induced
chromosomal aberration within 24h or exposure flshidate etal.. 19811. Note that formalin was
used in this study as a source of formaldehyde.
Dresp and Bauchinger (1988) exposed human lymphocytes to various concentrations of
formaldehyde. A dose-dependent increase in chromosomal aberrations was observed. Schmid et
al. (1986) used the same cell lines and exposed them to 0.25 and 0.5mM formaldhyde containing
10% methanol. Both chromatid breaks and gaps were observed. It should be recognized that the in
vitro studies used different forms of formaldehyde, including commercial grade formaldehyde,
This document is a draft for review purposes only and does not constitute Agency policy.
A-101 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
paraformaldehyce, formalin (formaldehyde containing 10-15% methanol) or methanol-free
formaldehyde.
Mutations and cell transformation
Mutations may occur as a result of the misrepair of formaldehyde-induced DNA damage
(DPXs, DNA adducts, SSBs, or clastogenic effects) or as a result of replication errors during
mitogenesis. The in vitro evidence for formaldehyde-induced mutations, as discussed below, is
strengthened by the correlation between these genotoxic and clastogenic events of formaldehyde
and the induction of mutations in other test systems. Numerous studies have demonstrated
formaldehyde-induced DNA mutations under a variety of experimental conditions (reviewed in
flARC. 2012: NTP. 2010: TARC. 2006a: T.iteplo and Meek. 2003: Conawav etal.. 1996: TARC. 1995:
Ma and Harris. 1988: Auerbach etal.. 19771.
Deletion and point mutations
Several studies demonstrated deletion mutations in cultured mouse lymphoma cells fSpeit
and Merk. 2002: Mackerer et al.. 19961. CHO cells and V79 lung epithelial cells at the hypoxanthine
phosphoribosyl transferase (hprt) locus (Merk and Speit. 1999.1998: Graves etal.. 1996: Grafstrom
etal.. 19931 as well as in human TK6 lymphoblast cells (Crosby etal.. 1988: Craft etal.. 1987:
Goldmacher and Thillv. 19831 as shown in Table A-21.
Craft et al., (1987) measured the induction of mutations in the thymidine kinase [tk) locus
or at the ouabain resistance (Ouar) locus in TK6 human lymphoblastoid cells. The mutagenesis at tk
locus can result from base-pair substitutions, small and large deletions, and chromosome exchange
events, while mutations at the Ouar locus require specific base-pair substitutions. Lymphoblostoid
cells were exposed to single (0,15, 30, 50,125, or 150 |a,M for 2 hours) or multiple treatments, that
is, 3, 5, or 10 treatments of 50, 30, or 15 |a,M, respectively, or 4 treatments of 150 |a,M for 2 hours
(treatments were spaced 2-4 days apart) with formaldehyde and mutations analyzed. The authors
observed a nonlinear increase in tk mutagenesis with sinlge treatment of formaldehyde with
increasing slope >125 |a,M. Although multiple treatments caused an increase in tk mutagenesis,
their combined effect was less than the single treatment of equivalent C x t (150 |a,M x 2 hours). No
mutations were observed at the Ouar locus in lymphoblasts that received four treatments of 150 |a,M
for 2 hours. Tk mutagenesis followed a similar exposure-response curve as DPX formation in this
study (Craft et al. 1987).
Using the same cell system, Crosby et al. (1988) showed that repetitive treatments of 150
|iM formaldehyde induced mutants at the X-linked hypoxanthine-guanine phosphoribosyl
transferase (HPRT) locus. Of these mutants, 14/30 of them contained partial or complete deletions
with most of the partial deletions showing unique deletion patterns, while only a third (5/15) of
spontaneous mutants had partial or complete deletions, indicating that formaldehyde can induce
large losses of DNA in human lymphoblast cells. This work was followed up by fLiber etal.. 19891.
who showed that HPRT mRNA from human lymphoblast mutants (16 formaldehyde-induced and
This document is a draft for review purposes only and does not constitute Agency policy.
A-102 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
10 spontaneous, both not showing deletions) contained a preferential AT to CG transversion at a
specific site fLiber etal.. 19891.
Formaldehyde has been shown to induce hprt mutations in CHO cells involving single-base
pair transversions mostly occurring at AT sequences fGraves etal.. 19961. Formaldehyde also
induced forward mutations in mouse lymphoma L5178Y tk± cells both in the absence and presence
of rat liver S9 (higher concentrations required for effect with S9). Both toxicity and mutagenicity
were abolished when formaldehyde dehydrogenase (FADH) was incorporated in the exposure
medium (Blackburn et al.. 19911. suggesting detoxification of formaldehyde.
A study by Merk and Speit (1998) indicated that formaldehyde-induced DPXs did not result
in direct gene mutations in the hprt locus of V79 Chinese hamster cells, suggesting that
formaldehyde was not mutagenic. However, the hprt mutation assay may be insensitive to deletion
mutations fMerk and Speit. 19981 because the hprt locus in the V79 cell line is primarily sensitive to
point mutations. Additionally, one study showed the formation of deletion mutations by
formaldehyde at the same locus in human lymphoblasts (Crosby etal.. 19881.
In the mouse lymphoma assay (L5178Y cells), Speit and Merk (20021 demonstrated that a
2-hour exposure to formaldehyde was mutagenic in a concentration-dependent manner. Mutation
was mainly attributed to a strong increase in small colony mutants suggestive of CAs.
Recombination or deletion of DNA from the tk locus was primarily responsible for the loss of
heterogeneity, thereby leading to the observed mutant phenotype. This mutagenic finding in the
L5178Y cell mouse lymphoma system, which is likely to occur by a clastogenic mechanism rather
than by point mutations (Speit and Merk. 20021. is consistent with that of Craft et al. (Craft etal..
19871. who demonstrated formaldehyde mutagenicity at the tk locus of TK6 cells, and also with the
findings of fGrafstrom et al.. 19841. who demonstrated increased SSB formation in formaldehyde-
exposed cell lines.
Transformation
Formaldehyde has also been shown to induce cell transformation in mouse embryo
fibroblasts fBoreiko and Ragan. 1983: Frazelle etal.. 1983: Ragan and Boreiko. 19811 and hamster
kidney cells fPlesner and Hansen. 19831 as shown in Table A-21. In mouse embryonic C3H/10TV2
cells, a single exposure to formaldehyde (0.003-0.083 mM) for 24 hours did not induce
transformation; however, when formaldehyde treatment was followed by continuous treatment
with 0.1 |a,g/mL with the tumor promoter 12-O-tetradecanoyl phorbol-13-acetate (TPA), a dose-
dependent increase in transformation was observed at low concentrations of 0.003 mM (Boreiko
and Ragan. 19831 or 0.017 mM (Ragan and Boreiko. 19811 formaldehyde. Ragan and Boreiko
(1981) have also shown that treatment of mouse embryo fibroblasts with varying doses of formic
acid (=2 to 22 mM) or methanol (=0.11 to 1.1 M) did not induce transformation either alone or
following TPA promotion in mouse embryo fibroblasts. The authors concluded that since
commercial formalin contains 10% methanol, and use of 105 times higher methanol concentrations
(=2.2 M) in this experiment ruled out the background interference of methanol (precursor to
This document is a draft for review purposes only and does not constitute Agency policy.
A-103 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
formaldehyde) or formic acid (a metabolic product of formaldehyde) with formaldehyde-induced
cell transformation. In a different study using the same cells, the ability of formaldehyde to act as a
tumor promoter was tested with repeated applications of formaldehyde following initiation with N-
methyl-N'-nitro-N-nitrosoguanidine (MNNG) by Frazelle et al. 19831 who observed a weak tumor
promoting activity of formaldehyde. Another study with a 3-hour exposure to formaldehyde (0.003
to 3.33 mM) with metabolic activation using S9 mix in baby hamster kidney (BHK) cells induced
dose-dependent increase in transformation (Plesner and Hansen. 1983).
Expression ofp53 mutation and cell death
Four cell lines derived from formaldehyde induced rat nasal squamous cell carcinomas
(SCCs) from a previous study (Recio etal.. 1992) were analyzed by Bermudez et al., (1994) for p53
mutations as shown in Table A-21. These cell lines were aneuploid overexpressing transforming
growth factor-a and epidermal growth factor, expression of which is a common feature of SCCs and
is frequently found in human tumors. Two each of these cell lines contained wild type DNA
sequences while two others possessed mutated p53 gene sequences, being point mutations, in
particular having transversions atcodons 132 (TTC->TTA) and 271 (CGT->CAT) ofthe p53 gene.
In order to understand the mechanism of transformed cell lines conveting to tumor phenotype, the
auhors injected either the the wild type or cells with mutant p53 sequnces into nude mice. They
observed that only cell lines expressing the p53 mutation were tumorigenic, suggesting
involvement of specific p53 mutations in the tumorigenicity of formaldehyde. Wong et al. (2012)
examined signal transduction pathways in response to formaldehyde exposure. The authors
studied p53 phosphorylation in human lung epithelial (H460 cells) and fibroblast cells exposed to
formaldehyde and compared the role of different protein kinases using specific inhibitors for ATR,
ATM, and DNA, measuring Serl5p53 and thr68-CHKl phosphorylation, p53 accumulation, and
induction of p21. At low doses, formaldehyde-induced DNA-protein crosslinks caused ATR-
mediated activation of p53 in human lung fibroblasts and epithelial cells. The S-phase of the cell
cycle seems to be specifically sensitive for this effect without the involvement of topoisomerase
binding protein 1 (topBPl). Other pathways, such as BER and NER, mismatch repairs were not
affected by p53 activation, suggesting that non-DPC adducts, including DNA-peptide and hmDNa
adducts, play a minor role in formaldehyde-induced p53 activation.
Other genotoxic endpoints
As summarized in Table A-21, in vitro formaldehyde exposure induces other genotoxic and
related effects in mammalian cells such as UDS and DNA repair inhibition.
Unscheduled DNA synthesis
UDS, which represents DNA repair activity following excision of DNA damage, has been
reported in rathepatocytes (Williams etal.. 1989b) and SHE cells (Hamaguchi and Tsutui. 2000)
exposed to formaldehyde. UDS was also observed in HeLa cells (Martin etal.. 1978). but not in
This document is a draft for review purposes only and does not constitute Agency policy.
A-104 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
human bronchial epithelial cells (Doolittle etal.. 19851 upon formaldehyde exposure. These studies
suggest that formaldehyde-induced DNA damage was followed by DNA repair.
DNA repair inhibition
Formaldehyde can inhibit DNA repair and induce cell transformation fEmri etal.. 2004:
Speit etal.. 2000: Grafstrom et al.. 1984: Boreiko and Ragan. 19831 as shown in Table A-21. Studies
have shown that formaldehyde causes DNA repair inhibition at a concentration range of 0.125 mM
to 10 mM in human bronchial epithelial cells fGrafstrom et al.. 19841 and skin fibroblasts or
keratinocytes fEmri etal.. 20041. DNA repair proficient or deficient cell lines (e.g., XP), or cell lines
hypersensitive to DNA-DNA crosslinks (e.g., FA) (Speit etal.. 2000). In a study using human
keratinocytes and fibroblasts, Emri et al. (2004) tested the formation of DNA SSBs induced by
ultraviolet (UV) irradiation by UVB or UVC with or without prior treatment with 10 |iM
formaldehyde. The authors reported that SSB induced by UV irradiation alone were repaired
within 3-6 hours of exposure, while cells with UV irradiation followed by formaldehyde exposure
had higher SSBs at the same time points due to increased chromosomal damage, suggesting that
formaldehyde exposure altered the repair kinetics in these cells.
Aneuploidy
Studies on aneuploidy in various in vitro and human cell systems have provided mixed
results as shown in Table A-21. For example, increase in aneuploidy was observed in hamster CHO
cells (Kumari etal.. 2012) and human erythropoietic stem cells (Ti etal.. 2014). However, no
increase in aneuploidy cells were observed in hamster V79 lung epithelial cells (Kuehner et al..
2012: Speit etal.. 2011a) or in human myeloid progenitor cells (Kuehner etal.. 2012).
Table A-21. Summary of in vitro genotoxicity studies of formaldehyde in
mammalian cells
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
p53 Mutations
Rat
Nasal tumor cell lines
NA
+
ND
cell lines derived from nasal
tumors of rats from 2-yr tumor
study; rats exposed to 18.5
mg/m3 HCHO, 6 hrs/day, 5
days/wk for 2 yrs
(Bermudez et al.,
1994)
Deletion mutations
Mouse
Lymphoma L5178Y
tk+/" cells
0.063 mM HCHO
(commercial)
+
ND
2 hrs; mouse lymphoma assay;
cytotoxic at 250 |j.M conc.
(Speit and Merk,
2002)
0.8 mM 37% HCHO
+ 10% methanol
ND
+
3 hrs; MF at TK locus; 40-50%
total growth at 0.8 mM dose
(Mackerer et al.,
1996)
Hamster
CHO cells/Hprt locus
0.3 mM HCHO (37%
w/w)
+
ND
1 hr; 6-TG resistant mutants;
dose-dependent J, in CFE and
(Grafstrom et al.,
1993)
This document is a draft for review purposes only and does not constitute Agency policy.
A-105 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
| in MF
0.5 mM HCHO
(commercial)
-
ND
4 hrs; HPRT assay; (T) by
relative CE > 0.125 mM
(Merk and Speit,
1998)
1 mM HCHO (40%
aq. Sol.)
+
ND
1 hr; 6-TG resistant colonies;
base transversions at AT base
pairs
(Graves et al.,
1996)
Hamster
V79 lung epithelial
cells
0.5 mM HCHO
(commercial)
-
ND
4 hrs; HPRT assay; (T) by
relative CE > 0.25 mM
(Merk and Speit,
1999)
Human
Bronchial
fibroblasts/epithelial
cells (HP/?7~ locus)
0.1 mM HCHO
(commercial)
+
ND
5 hrs; 6-TG resistant mutants
scored; MF nonlinear dose-
dependent |; (T) > 0.1 mM by
CFE
(Kilburn and Moro,
1985)
Human
Lymphoblast/TK6
0.03 mM 37%
HCHO + 10-15%
methanol
+
ND
2 hrs; MF at TK locus
measured; single exposure (0-
150 |am) nonlinear f in MF; (T)
at 0.125 mM
(Craft et al., 1987)
0.13 mM 37%
HCHO + 10-15%
methanol
+
ND
2 hrs; MF at TK locus; cell
survival was 15% at 0.15 mM;
cells treated for 2 hrs with
0.07 mM methanol were not
mutagenic, not cytotoxic
(Goldmacher and
Thillv. 1983)
0.15 mM HCHO
(commercial)
+
ND
8 exposures x 4 days, 2 hrs
dosing; MF at HPRT locus; MF
12.4-fold higher over
background; (T) 50% survival
each treatment
(Crosbv et al.,
1988)
Point mutations
Mouse
Lymphoma cell/ TK+/-
0.1 mM (-S9) and
0.5mM (+S9) 37%
HCHO +10%
methanol
+/-
+/-
NR; assay supplemented with
FDH and NAD+; MFattheTK
locus; results indicate without
and with FDH/NAD+,
respectively; 50% (T) at 0.1
mM (-S9) and 0.5 mM (+S9)
with FDH
(Blackburn et al.,
1991)
0.14 mM HCHO
form not specified
+
ND
4 hrs; MF at TK locus; highly
mutagenic but total growth is
very low
(Wangenheim and
Bolcsfoldi, 1988)
Hamster
CHO cells/Hprt locus
1 mM HCHO (40%
aq. Sol.)
+
ND
1 hr; 6-TG resistant colonies
had base transversions at AT
base pairs
(Graves et al.,
1996)
Human
Lymphoblast/TK6
0.15 mM HCHO
(commercial)
+
ND
2 hrs (8 times); sequence
analysis of HPRT mutants
showed base substitutions at
AT base pairs
(Liber etal., 1989)
This document is a draft for review purposes only and does not constitute Agency policy.
A-106 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
DNA-protein crosslinks
Mouse
Hepatocytes
0.5 mM [14C] HCHO
(aq. Sol.)
+
ND
2 hrs; nonlinear dose-
dependent | in DPX.
(Casanova et al.,
1997)
0.5 mM [14C] HCHO
(aq. Sol.)
+
ND
2 hrs; HPLC analysis of DNA
digest; Dose-dependent f in
DPX.
(Casanova and
Heck. 1997)
Mouse
L5178Ytk+/- Lymphoma
cells
0.031 mM HCHO
(commercial)
+
ND
2 hrs; DPX show dose-
response; cytotoxic at 250 |j.M
conc.
(Speit and Merk,
2002)
Mouse
Leukemia L1210 cells
0.125 mM 37%
HCHO
+
ND
1 hr; (T) at 0.3 |j.M conc.
(Ross et al., 1981)
0.2 mM 37% HCHO
+
ND
2.5 hrs; (T) >0.175 mM
(Ross and Shiplev,
1980)
Mouse
Bone marrow
mesenchymal cells
0.125 mM HCHO
(37%)
+
ND
12 hrs; Alkaline comet assay;
(T) from 0.175 mM to 0.2 mM
(She et al., 2013)
Rat
C18 tracheal epithelial
cell line
0.1 mM PFAin PBS
+
ND
1.5 hrs; DPX analyzed by
alkaline elution; (T) at 0.4 mM
(Cosma and
Marchok, 1988)
Rat
Aortic endothelial cells
0.5 mM HCHO
(commercial)
+
ND
1.5 hrs; K+/SDS assay; dose-
dependent | in DPC > 2 hrs;
(T) by LDH release at 2 mM
(Lin et al., 2005)
Rat
Primary tracheal
epithelial cells
0.05 mM PFA in
PBS
+
ND
1.5 hrs; DPX analyzed by
alkaline elution; (T) > 0.2 mM
{Cosma, 1988, 626327}
3.34 mM
HCHO/PBS
+
ND
3 hrs; dose-dependent f in
DPX
{Cosma, 1988, 626327}
Rat
Yoshida
lymphosarcoma cells
0.25 mM HCHO
(36% sol)
+
ND
4 hrs; alkaline elution assay;
(T) IDsoO.25 mM
(O'Connor and Fox,
1987)
Hamster
CHO cells
0.125 mM HCHO
(commercial)
+
ND
2 hrs; Brdll incorporation-FPG
technique; conc.-related 4,
DNA migration inhibition;
(Garcia et al.,
2009)
0.2 mM HCHO (NS)
+
ND
1.5 hrs; dose-dependent T* in
DPX up to 2 mM HCHO; values
visually determined from
graph
(Zhitkovich and
Costa, 1992)
0.25 mM HCHO
(NS)
+
ND
1.5 hrs; dose-dependent f in
DPX formation up to 2 mM
HCHO; values visually
determined from graph
(Olin et al., 1996)
0.5 mM HCHO
(commercial)
+
ND
1.5 hrs; alkaline elution assay;
DPX showed dose-dependent
1(0.5-4.5 mM); 82% viability at
{Marinari, 1984, 68819
This document is a draft for review purposes only and does not constitute Agency policy.
A-107 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
4.5 mM HCHO
Hamster
V79 lung epithelial
cells
0.01 mM 16%
HCHO (ultrapure
methanol free)
+
ND
1 hr; Comet assay; dose-
dependent I in DNA migration
at HCHO > 0.01 mM;
(Speit et al., 2007b)
0.025 mM 16%
HCHO (ultrapure
methanol free);
+
ND
4 hrs; Comet assay; dose-
dependent I DNA migration;
(T) at 0.2 mM by cell
counts/proliferation index;
(Speit et al., 2008a)
0.0625 mM HCHO
(commercial)
+
ND
4 hrs; Comet assay; dose-
dependent 1 migration
inhibition (0.0625-0.5 mM); (T)
by relative CE > 0.25 mM;
(Merk and Speit,
1999)
0.12 mM HCHO
(commercial)
+
ND
1 hr; method not specified
Swenberg et al., 1983
0.125 mM
HCHO (commercial)
+
ND
4 hrs; K-SDS assay; nonlinear
dose-dependent f in DPC
(values visually determined
from graph); HCHO (T) by
relative CE assay > 0.125;
(Merk and Speit,
1998)
Human
Nasal epithelial cells
0.2 mM 16% HCHO
(ultrapure
methanol free)
+
ND
1 hr; Comet assay; dose-
dependent 1" DPXfrom 0.05-
0.3 mM; (T) byCF>0.02mM;
(Speit et al., 2008b)
Human
A549 lung epithelial
cells
0.2 mM 16% HCHO
(ultrapure
Methanol free)
+
ND
1 hr & 4 hrs; Comet assay;
dose-dependent T* migration
inhibition from 0.1-0.3 mM;
(T) by CF > 0.02 mM;
(Speit et al., 2008b)
0.2 mM HCHO
(stabilized with
Methanol)
+
ND
3 hrs; KCI/SDS method; DPX
time-dependent T* up to 12
hrs; TVz 12.5 hrs; (T) > 0.2 mM
by CF assay,
(Quievrvn and
Zhitkovich, 2000)
0.2 mM 16% HCHO
aq. sol., methanol-
free
+
ND
1 or 3 x 24 hr intervals; comet
assay
{Speit, 2010, 1041161}
Human
Lung/bronchial
epithelial cells
0.1 mM
HCHO (commercial)
+
ND
1 hr; alkaline elution
technique; (T) 0.021 mM ID50
by growth inhibition
(Saladino et al.,
1985)
0.1 mM HCHO
(commercial)
+
ND
1 hr; alkaline elution
technique; (T) at 0.3 mM by
CFE
(Grafstrom et al.,
1986)
0.2 mM 37% HCHO
(w/w)
+
ND
1 hr; alkaline elution
technique; (T) at 1 mM
(Grafstrom et al.,
1984)
2 mM HCHO (Not
Specified)
+
ND
1 hr; Alkaline elusion
technique;
{Grafstrom, 1990,
891139}
0.39 mM HCHO
+
ND
4 hrs; KCI-SDS method
(Duan, 2011)
This document is a draft for review purposes only and does not constitute Agency policy.
A-108 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
0.8 mM 37% HCHO
+
ND
1 hr; alkaline elution;
(Fornace et al.,
1982)
Human
Bronchial epithelial
cells/fibroblasts
0.1 mM 37% HCHO
+
ND
1 hr; alkaline elution
technique;
(Grafstrom et al.,
1983)
Human
Fibroblasts
(diploid)/HF/SV40
0.2 mM HCHO +
Methanol)
+
ND
3 hrs; (T) > 0.2 mM by CF
assay; DPX half life is 12.5 hrs
(Quievrvn and
Zhitkovich, 2000)
Human
Fibroblast
(Bronchial/Skin)
0.25 mM
HCHO (NS)
+
ND
1.5 hrs; DPX dose-response
not prominent; values visually
determined from graph
(Olin et al., 1996)
Human
Skin keratinocytes/
fibroblasts
0.025 mM HCHO
(NS)
+
ND
8 hrs with subsequent
exposure to methyl methane
sulfonate (0.25 mM)
(Emri et al., 2004)
Human
XP fibroblasts
0.2 mM 37% HCHO
(w/w)
+
ND
1 hr; alkaline elution
technique; DPC TV2 2-3 hrs
(Grafstrom et al.,
1984)
Human
Normal, XPA and FA
repair deficient
fibroblasts
0.125 mM
HCHO (commercial)
+
ND
2 hrs; Comet assay; dose-
dependent DNA migration
inhibition; No migration
inhibition after 24 hrs;
(Speit et al., 2000)
Human
Fibroblasts/XP-F and
XP-A
0.2 mM HCHO
(stabilized with
Methanol)
+
ND
3 hrs; DPX removal XP-A = XP-
F cells; (T)>0.2 mM by CF
assay;
(Quievrvn and
Zhitkovich, 2000)
Human
Lymphocytes
0.05 mM 10%
formalin
+
ND
1 hr; comet assay; KCI/SDS
assay; nonlinear dose-
dependent | > 50 |j.M HCHO
(Liu et al., 2006)
0.1 mM; 0.3 mM
HCHO in water
+
-
3 hrs; (T) at 0.3 mM (+S9)
(Andersson et al.,
2003)
0.2 mM
HCHO + Methanol)
+
ND
3 hrs; KCI/SDS method; DPX
J1/218.1 hrs; (T) >0.2 mM by
CF assay,
(Quievrvn and
Zhitkovich, 2000)
Human
White blood cells
0.001 mM
HCHO (NS)
+
ND
1.5 hrs; Dose-dependent T* in
DPX formation up to 2 mM
HCHO; values visually
determined from graph
(Shaham et al.,
1996)
Human
Whole blood cultures
0.025 mM 16%
HCHO (ultrapure
Methanol free)
+
ND
exposure duration not
specified; Comet assay; dose-
dependent migration
inhibition; DPX >0.2 mM
persist for 24 hrs;
(Schmid and Speit,
2007)
Human
Lymphoblast/TK6
0.05 mM 37%
HCHO + 10-15%
Methanol
+
ND
2 hrs; MF atTK locus
measured; (T) at 0.125 mM
(Craft et al., 1987)
Human
0.1 mM 16% HCHO
+
ND
2 hrs; Comet assay with g-
(Kuehner et al..
This document is a draft for review purposes only and does not constitute Agency policy.
A-109 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
Lymphoblast/TK6
(ultrapure MetOH
free)
irradiation; DPX formation
dose-dependent; (T) at 0.1
mM 24 hrs by MTT assay
2013)
Human lymphoblasts
(PD20& PD20-D2)
0.125 mM 37%
HCHO
+
ND
24 hrs; Dose-dependent f in
DPC from 0.05-0.15 mM;
PD20>PD20-D2; (T) >0.15 mM
(Ren et al., 2013)
Human
EBV-Burkitt's
lymphoma cells
0.03% PFA in water
+
ND
18 hrs; Dose-dependent f in
DPX; (T) 0.01% PFA
(Costa et al., 1997)
Human
T-leukemia (Jurkat E6-
1) cells
1 mM HCHO
(commercial)
+
ND
2 hrs; SDS-PAGE; (T) > 1 mM
by cell death assay
(Saito et al., 2005)
Human
HeLa cells
0.05 mM
10% formalin
+
ND
1 hr; KCI/SDS precipitation
method; (T) > 100 mM by
absorbance after 12 hrs; dose-
dependent 1" in DPX; repaired
within 18 hrs after HCHO
removal
(Liu et al., 2006)
Human
Kidney cells/Ad293
0.2 mM
HCHO + Methanol
+
ND
3 hrs; KCI/SDS method; DPX
J1/212.5 hrs; (T) > 0.2 mM by
CF assay,
(Quievrvn and
Zhitkovich, 2000)
Human
Gastric mucosa cells
1 mM HCHO
+
ND
1 hr; (T) not reported
(Blasiak et al.,
2000)
DNA adducts
Hamster
CHO cells
1 mM [3H] 37%
HCHO/10-15%
Methanol
+
ND
2 hrs; (T)>2.5mM
(Beland et al.,
1984)
Human
Nasal epithelial cells
0.33 mM 37%
HCHO + 10%
Methanol
+
ND
24 hrs; hmdA and hmdG
adducts dose-dependent f .
Viability showed dose-
dependent from 10
500 mM;
(Zhong and Que
Hee. 2004)
Human
HeLa cells
0.5 mM
[13CD2]HCHO (20%
in heavy water)
+
ND
3 hrs; No (T) information
provided.
(Lu et al., 2012a)
Chromosomal aberrations (CA)
Hamster
CHO cells (AA8) and
their mutants (UV4,
UV5, UV61)
0.15 mM
HCHO (commercial)
+
ND
2 hrs; Brdll incorporation-FPG
technique; dose-dependent f
in Cas
(Garcia et al.,
2009)
Hamster
CHO cells
0.2 mM
PFA in water
+
+
2 hrs; Brdll incorporation;
dose-dependent T* in SCE +/-
S9;
Natarajan et al., 1983
Hamster
CHO cells mutants
0.2 mM
HCHO (commercial)
+
ND
2 hrs; Brdll incorporation-FPG
technique; dose-dependent f
(Garcia et al.,
This document is a draft for review purposes only and does not constitute Agency policy.
A-110 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
(KO40)
in CAs
2009)
Hamster
CHO cells
0.53 mM HCHO
(+)
(+)
8-12 hrs; Giemsa staining;
(Gallowav et al.,
1985)
Hamster
Lung fibroblasts
0.6 mM Formalin
+
ND
24 hrs; microscopic evaluation
(Ishidate et al.,
1981)
Hamster/Syrian
Embryo cells
0.033 mM 37%
HCHO + 7-13%
Methanol
+
ND
24 hrs; CA assay; 85% relative
CFEat 0.099 mM
(Hikiba et al., 2005)
Human
Fibroblasts
2 mM HCHO (NS)
+
ND
0.25 hr; Giemsa staining; dose-
dependent 1 in CA;
(Lew et al., 1983)
Human
Lymphocytes
Human lymphoblasts
(PD20& PD20-D2)
0.125 mM HCHO
(NS)
+
ND
1 hr; PCC technique; dose-
dependentf in CA
(Dresp and
Bauchinger, 1988)
0.125 mM 37%
HCHO
+
ND
24 hrs; Dose-dependent f in
CA from 0.05-0.15 mM;
PD20=PD20-D2; (T) >0.15 mM
(Ren et al., 2013)
Human
lymphocytes
0.25 mM, 0.5 M
37% HCHO + 10%
Methanol
+
+
1 hr; conc. Respectively, for
chromatid breaks and gaps;
proliferation inhibition at 1 M
(-S9) and 0.5 mM (+S9)
(Schmid et al.,
1986)
Micronucleus (MN)
Mouse
erythropoietic cells
0.025 mM HCHO
(37% + 10-15%
methanol)
+
ND
1 hr; Dose-dependent in MN
from 0.025-0.1 mM;
(Ji et al., 2014)
Hamster
V79 lung epithelial
cells
0.075 mM 16%
HCHO (ultrapure
Methanol free);
+
ND
2 hrs; MN test; MN> 0.075
mM; dose-dependent f in MN;
(Speit et al., 2007b)
0.1 mM 16% HCHO
(ultrapure
Methanol-free);
+
ND
4 hrs; MN test; dose-
dependent in MN; (T) at 0.2
mM by cell
counts/proliferation index;
(Speit et al., 2007b)
0.125 mM
HCHO (commercial)
+
ND
4 hrs; MN assay with AO
staining; nonlinear dose-
dependent 1" in MN (values
visually determined from
graph); (T) by relative CE >
0.125 mM;
(Merk and Speit,
1998)
Human
A549 lung epithelial
cells
0.15 mM 16%
HCHO (ultrapure,
methanol-free)
+
ND
2 hrs (0.3 mM) or 30 hrs (0.15
mM); CBMN assay; Mostly
centromere -ve by FISH
analysis
(Speit et al., 2011a)
Human
Normal, XPA and FA
repair deficient
fibroblasts
0.125 mM
HCHO (commercial)
+
ND
2 hrs; MN test; MN > 0.075
mM; dose-dependent T* in
MN; normal
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
Human lymphoblasts
(PD20& PD20-D2)
0.125 mM 37%
HCHO
+
ND
24 hrs; Dose-dependent f in
MN from 0.05-0.15
mM;PD20>PD20-D2; (T) >0.15
mM
(Ren et al., 2013)
Human
Whole blood cultures
0.3 mM 16% HCHO
(ultrapure,
methanol-free)
+
ND
27 hrs; CBMN assay; mostly
centromere negative by FISH
analysis
(Speit et al., 2011a)
Human
Whole blood cultures
0.3 mM 16% HCHO
(ultrapure
Methanol free);
+
ND
24 hrs; HCHO dosed 44 hrs
after culture; MN test; dose-
dependent | in MN (0.1-0.4
mM); (T) > 0.3 mM by NDI;
(Schmid and Speit,
2007)
Single strand breaks (SSB)
Mouse
Leukemia L1210 cells
0.125 mM
37% HCHO
-
ND
1 hr; (T) at 0.3 mM
(Ross et al., 1981)
0.2 mM 37% HCHO
(+)
ND
2.5 hrs; (T) >0.175 mM
(Ross and Shiplev.
1980)
Rat
Hepatocytes
1 mM HCHO (NS)
+
ND
4 hrs; HCHO cytotoxic >1.5
mM; dose-dependent f in SSB,
enhanced by GSH depletion
(Demkowicz-
Dobrzanski and
Castonguay, 1992)
Rat -tracheal epithelial
cell line
0.2 mM PFAin PBS
+
ND
1.5 hrs; SSB analyzed by
alkaline elution; HCHO toxic at
0.4 mM
{Cosma, 1988, 626327}
Rat
Yoshida
lymphosarcoma cells
0.25 mM HCHO
(36% sol)
+
ND
4 hrs; alkaline elution assay;
(T) IDsoO.25 mM
(O'Connor and Fox,
1987)
Hamster
CHO cells
4.5 mM HCHO
(commercial)
-
ND
1.5 hrs; 82% viability at 4.5
mM HCHO
(Marinari et al.,
1984)
Hamster
V79 lung epithelial
cells
0.2 mM 16% HCHO
(ultrapure
Methanol free)
-
ND
1 hr; Comet assay;
(Speit et al., 2007b)
Human
Bronchial epithelial
cell
0.1 mM 37% HCHO
+
ND
1 hr; alkaline elution
technique; (T) at 0.3 mM
(Grafstrom et al.,
1983)
0.3 mM 37% HCHO
(w/w)
+
ND
1 hr; SSB dose-dependent
SSB 3 times higher than XP
cells
(Grafstrom et al.,
1984)
Human
Lung/bronchial
epithelial cells
0.1 mM HCHO
(commercial)
+
ND
1 hr; alkaline elution
technique; (T) 0.021 mM ID5o
by growth inhibition
(Saladino et al.,
1985)
0.1 mM HCHO
(commercial)
+
ND
1 hr; alkaline elution
technique; (T) at 0.3 mM by
CFE
(Grafstrom et al.,
1986)
0.8 mM 37% HCHO
+
ND
1 hr; alkaline elution;
(Fornace, 1982)
This document is a draft for review purposes only and does not constitute Agency policy.
A-112 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
Human
Lung/bronchial
epithelial (A549) cells
1.0 mM HCHO
(commercial)
+
ND
8-72 hrs; Dose-dependent in f
DSB formation; DSB formed
when viability, determined by
MTT assay, was >60%
(Vock et al., 1999)
Human
Skin keratinocytes/
fibroblasts
0.1 mM HCHO (NS)
-
ND
20 hrs
(Emri et al., 2004)
Human
XP fibroblasts
0.3 mM 37% HCHO
(w/w)
+
ND
1 hr; SSB dose-dependent f
(Grafstrom et al.,
1984)
Human
Foreskin fibroblasts
0.1 mM 37% HCHO
+ 10% Methanol
+
ND
0.5 hr; nick translation assay;
low doses induce SSB
Snyder and Van
Houten, 1986
0.25 mM 37%
HCHO + 10%
Methanol
-
ND
0.5 hr; alkaline sucrose
sedimentation analysis; high
doses don't induce SSB
(Snvder and van
Houten, 1986)
Human
HeLa cells
0.005 mM 10%
formalin
+
ND
1 hr; Comet assay; (T) > 100
|j.M after 12 hrs; SSB repaired
within 90 min
(Liu et al., 2006)
Human
Lymphocyte,
peripheral blood
0.005 mM 10%
formalin
+
ND
1 hr; comet assay; KCI/SDS
assay; nonlinear dose-
dependent | > 50 |j.M HCHO
(Liu et al., 2006)
Sister chromatid exchanges (SCE)
Hamster
CHO cells
0.03 mM 37%
HCHO with 10%
methanol
+
ND
24 hrs; Brdll incorporation;
SCE dose-dependent f
(Obe and Beek,
1979)
0.04 mM HCHO
(commercial)
(+)
(+)
26 hrs; Brdll incorporation-
FPG technique
(Gallowav et al.,
1985)
0.2 mM PFA in
water
+
+
2 hrs; Brdll incorporation;
dose-dependent f in SCE +/-
S9;
(Nataraian et al.,
1983)
Hamster
CHO cells (AA8) and
their mutants (UV4,
UV5, UV61, KO40)
0.15 mM HCHO
(commercial)
+
ND
2 hrs; Brdll incorporation-FPG
technique; dose-dependent T*
in CAs
(Garcia et al.,
2009)
Hamster
Embryo cells
0.01 mM 37%
HCHO/7-13%
Methanol;
+
ND
24 hrs; Brdll incorporation;
dose-dependent f in SCE; (T)
by relative CE 68% at 0.033
mM
(Mivachi and
Tsutsui, 2005)
Hamster
V79 lung epithelial
cells
0.05 mM 16%
HCHO (ultrapure,
methanol-free)
+
ND
24 or 28 hrs exposure to HCHO
and Brdll; Aneuploidy and
Toxicity measured by SCE and
PI, respectively.
(Speit et al., 2011a)
0.06 mM 37%
HCHO with 10%
methanol
+
-
28 hrs; formalin + activation
with primary rat hepatocytes;
(T) at 0.54 mM (+S9) and 0.2
(Basler et al., 1985)
This document is a draft for review purposes only and does not constitute Agency policy.
A-113 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
mM (-S9)
0.1 mM 16% HCHO
(ultrapure
Methanol free);
+
ND
2 hrs; Brdll labeling; SCE > 0.1
mM; genotoxicity paralleled
cytotoxicity; (T) > 0.1 mM by PI
(Speit et al., 2007b)
0.1 mM 16% HCHO
(ultrapure
Methanol free);
+
ND
1 hr; Brdll labeling; SCE dose-
dependent ^(0.1-0.2 mM)
(Neuss and Speit,
2008)
0.1 mM 16% HCHO
(ultrapure
Methanol free);
+
ND
4 hrs; Brdll labeling; dose-
dependent in SCE; (T) at 0.2
mM by cell
counts/proliferation index;
(Speit et al., 2008a)
0.125 mM HCHO
(commercial)
+
ND
4 hrs; Brdll incorporation;
dose-dependent f in SCE; (T)
by relative CE > 0.125 mM
(Merk and Speit,
1998)
0.125 mM HCHO
(commercial)
+
ND
4 hrs; Brdll incorporation;
dose-dependent f in SCE; (T)
by relative CE > 0.25 mM
(Merk and Speit,
1999)
0.13 mM 37%
HCHO with 10%
methanol
+
ND
2 hrs; (T) at 0.54 mM
(Basler et al., 1985)
0.13 mM; 0.20 mM
37% HCHO with
10% methanol
+
-
3 hrs; (T) at 0.4 mM (-S9)
(Basler et al., 1985)
Human
A549 lung epithelial
cells
0.1 mM 16% HCHO
(ultrapure
Methanol free);
+
ND
1 hr; Brdll labeling; SCE dose-
dependent 1" (0.1-0.3 mM)
(Neuss and Speit,
2008)
Human
A549 + V79 (co-
cultivated)
0.05 mM 16%
HCHO (ultrapure
Methanol free);
+
ND
1 hr; Brdll labeling; SCE dose-
dependent 1" (0.05-0.2 mM);
treated A549 cells not washed
before adding V79 cells
(Neuss and Speit,
2008)
Human
A549 + V79 (co-
cultivated)
0.3 mM 16% HCHO
(ultrapure
Methanol free);
-
ND
1 hr; Brdll labeling; treated
A549 cells washed before
adding V79 cells
(Neuss and Speit,
2008)
Human
Lymphocytes
0.125 mM
37% HCHO + 10%
Methanol
+
+
1 hr; Brdll labeling;
proliferation inhibition at 1 M
(-S9) and 0.5 mM (+S9)
(Schmid et al.,
1986)
0.167 mM
37% HCHO + 10%
Methanol
+
ND
24 hrs; Brdll incorporation;
dose-dependent f in SCE
(Obe and Beek,
1979)
0.167 mM
formalin or PFA
+
ND
72 hrs; Brdll incorporation
with fluorescence + Giemsa
method; (T) >0.33 mM and
similar for formalin and PFA;
dose-dependent T* for
formalin reported
(Kreiger and Garrv.
1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-114 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
Human
Whole blood cultures
0.2 mM 16% HCHO
(ultrapure
Methanol free)
+
ND
72 hrs; Brdll labeling; no dose-
response; (T) at 0.2 mM by PI
(Schmid and Speit,
2007)
Unscheduled DNA synthesis (UDS)
Rat
Hepatocytes
400 mM HCHO (NS)
+
ND
18-20 hrs; [3H]dThd
incorporation and
autoradiography
(Williams et al.,
1989a)
Human
Bronchial epithelial
cells
0.1 mM 37% HCHO
(reagent grade sol.)
-
ND
22 hrs; [3H]dThd incorporation
and autoradiography; (T) > 1
mM
(Doolittle et al.,
1985)
Human
Foreskin fibroblasts
0.5 mM 37% HCHO
+ 10% Methanol
-
ND
0.5 hr; UDS
(Snyder and van
Houten, 1986)
Human
Bronchial fibroblasts
1 mM 37% HCHO
-
ND
1 hr; [3H-Thymidine]
incorporation.
(Grafstrom et al.,
1983)
Human
Embryo cells
0.1 mM HCHO (37%
sol)
+
ND
1 hr; [3H]dThd incorporation;
dose-dependent T* in UDS
(0.1-1 mM)
(Hamaguchi and
Tsutui, 2000)
Human
HeLa cells
0.001 mM HCHO
(commercial)
+
ND
2.5 hrs; [3H]dThd
incorporation
(Martin et al.,
1978)
DNA repair inhibition
Human
Skin
keratinocytes/fibrobla
sts
0.01 mM HCHO
(NS)
+
ND
0.5 hr after exposure to UVB
(Emri et al., 2004)
Human
Normal, XPA and FA
repair deficient
fibroblasts
0.125 mM HCHO
(commercial)
+
ND
2 hrs
(Speit et al., 2000)
Cell transformation
Mouse
Embryo
fibroblast/C3H10TV2
cells
0.003 mM HCHO
(37%)
+
ND
24 hrs; HCHO treatment
followed by TPA treatment,
transformation +ve and dose-
dependent; (T) > 0.017 mM
(Boreiko and
Ragan, 1983)
0.017 mM HCHO
(37% w/w)
exposure
+
ND
24 hrs HCHO, 6 wks to
medium ± TPA. HCHO+TPA
+ve, dose-dependent f (0.017-
0.34 mM); HCHO alone -ve
(0.083 mM); methano + TPA or
formic acid + TPA-ve. HCHO
cytotoxic at 0.033 mM
(Ragan and
Boreiko, 1981)
Mouse
Embryo
fibroblast/C3H10TV2
cells
0.033 mM HCHO
(37% w/w)
exposure;
[+]
ND
4 hrs initiation with 0.5 |J.g/mL
MNNG, promotion on days 5,
8, 15, 22, 29, 36 with HCHO
with change of medium
(Frazelle et al.,
1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-115 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Dose/
Concentration3
Results'5
Comments (duration;
endpoint method; toxicity)
Reference
-S9
+S9
Hamster
Kidney cell/BHK-
21/cl. 13
0.03 mM HCHO
37% aq.sol.
+
+
3 hrs; Style's cell
transformation assay;
transformation dose-
dependent T* (0.03-0.67 mM);
(T) >0.67 mM
(Plesner and
Hansen, 1983)
Aneuploidy
Hamster
CHO cells (WT&XPF-
deficient)
0.3 mM HCHO (Not
Specified)
+
ND
4 hrs; Wright's stain and G-
banding; +ve for tetraploidies
and polyploidies
(Kumari et al.,
2012)
Hamster
V79 lung epithelial
cells
0.05 mM HCHO,
16% ultra-pure,
methanol-free
ND
7 days exposure; FISH analysis;
(T) at 0.05 mM by CFA
(Kuehner et al.,
2012)
Hamster
V79 lung epithelial
cells
0.1 mM HCHO, 16%
ultra-pure,
methanol-free
ND
24 or 28 hrs exposure to HCHO
and Brdll; Aneuploidy and
Toxicity measured by SCE and
PI, respectively.
(Speit et al., 2011a)
Human
A549 lung epithelial
cells
0.05 mM HCHO,
16% ultra-pure,
methanol-free
ND
14 days exposure; FISH
analysis; (T) at 0.02 mM by
CFA
(Kuehner et al.,
2012)
Human
myeloid progenitor
cells
0.05 mM HCHO,
16% ultra-pure,
methanol-free
ND
9 days exposure; Aneuploidy
in chromosomes 6 7, and 8
tested by FISH analysis; (T) at
0.1 mM by CFA
(Kuehner et al.,
2012)
Human
erythropoietic stem
cells
0.05 mM HCHO
(37% +10-15%
methanol)
+
ND
5 days; FISH analysis;
Combined analysis of
monosomies or trisomies of 7
and 8 are positive.
(Ji et al., 2014)
aLowest effective concentration (LEC) for positive results or highest ineffective concentration tested (HIC) for
negative or equivocal results.
b+ = positive; - = negative; (+), equivocal.
6-TG, 6-thioguanine; CF, colony formation; FA, Fanconi anemia; FDH, formaldehyde dehydrogenase; FPG,
fluorescence plus Giemsa technique; HCHO, formaldehyde; hmdA, hydroxymethyl-deoxyadenosine; hmdG,
hydroxymethyl-deoxyguanosine; hmDNA, hydroxymethyl-DNA; HPRT, hypoxanthine phosphoribosyl transferase;
ID5o, HCHO concentration causing 50% growth inhibition compared to control cells; MF, mutation frequency; MN,
micronucleus; NAD, nicotinamide adenine dinucleotide; ND, not done; NDI, nuclear division index; NR, not
reported; NS, not specified; PFA, paraformaldehyde; PCC, premature chromosome condensation; PI, proliferation
index; SCC, squamous cell carcinoma; SCE, sister chromatid exchange; (T), toxicity or cytotoxicity; TK, thymidine
kinase; XP, xeroderma pigmentosum; AA8, parental CHO cells; CHO cell mutants deficient in nucleotide excision
repair (UV4 & UV5), or transcription-coupled repair (UV61) or crosslink repair-deficient (KO40).
1 Summary on in vitro genotoxicity of formaldehyde
2 In vitro genotoxicity of formaldehyde has been reported in several mammalian cell culture
3 systems (see Table A-21). Formaldehyde is mutagenic in several mouse lymphoma cells, Chinese
4 hamster ovary (CHO) and hamster lung epithelial (V79) cells, human lung epithelial carcinoma
5 (A549) cell line, fibroblasts, gastric mucosa cells, and human peripheral blood lymphocytes (PBLs)
This document is a draft for review purposes only and does not constitute Agency policy.
A-116 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
and lymphoblasts. As shown in Table A-21, several genotoxicity endpoints, such as DNA-protein
crosslinks, hydroxymethyl-DNA adducts, single strand breaks, cytogenetic markers, such as
micronucleus, chromosomal aberrations, and sister chromatid exchanges, and other genotoxic end
points, such as unscheduled DNA synthesis, DNA repair inhibition, and cell transformation have
been demonstrated in animal and human cell systems.
Cell lines derived from formaldehyde-induced rat nasal squamous cell carcinomas showed
p53 mutations and the mutant cells were tumorigenic when injected in nude mice, suggesting the
mutagenicity and carcinogenicity of formaldehyde. Further, formaldehyde induced deletions and
point mutations at the thymidine kinase [tk) locus in cultured mouse lymphoma cells and human
lymphoblasts or at the hypoxanthine phosphoribosyl transferase (hprt) locus in CHO and V79 cells,
and the mutations showed a dose-dependent increase. Further, these mutations contained base
substitutions at the AT base pairs at both these loci.
Evidence of formaldehyde-induced genotoxicity was observed in rodent and human cells
wherein a dose-dependent increase in DPX formation was reported over a range of formaldehyde
concentrations (0.01-0.0625 mM) (see Table A-21). DPX are formed within an hour of exposure
and removed within 24 hrs after formaldehyde removal in cultured human cells. The average half-
life (ti/2) of DPX is 2-3 hours in xeroderma pigmentosum (XP) fibroblasts, 12.5 hours inAd293
kidney cells and A549 cells, and 18.1 hours (range 1-60 hours) in PBLs. The higher removal time in
PBLs is either due to low levels of glutathione in lymphocytes or inefficient repair. Thus, the
existing data suggest that repair of DPX depends on the cell type. The removal of DPX is carried out
either by spontaneous hydrolysis or other DNA repair processes; however, no difference in DPX
removal has been observed between normal human fibroblasts and fibroblasts from XP or Fanconi
anemia cell line, suggesting a lack of involvement of nucleotide excision repair in the repair process.
In proliferating cells, unrepaired DPX can arrest DNA replication and lead to the induction of other
genotoxic effects such as SCEs. Further evidence of DNA reactivity was observed in CHO cells, HeLa
cells, and human nasal epithelial cells wherein formaldehyde induced hm-DNA adducts.
Among the other types of genotoxicity, formaldehyde induced SSBs in several mammalian
cell systems, including mouse leukemia cells; rat primary hepatocytes, tracheal epithelial cells, and
lymphosarcoma cells; and human lung/bronchial epithelial cells, A549 and HeLa cells, skin
fibroblasts, and PBLs, within an hour of exposure (see Table A-21). It has been shown that SSBs can
be formed directly in lung/bronchial epithelial cells with formaldehyde exposure, independent of
DNA repair.
Several studies have demonstrated formaldehyde-induced cytogenetic markers (CAs, MN
and SCEs) in different rodent and human primary cells and cell lines (see Table A-21). For example,
CAs are induced in CHO cells (normal and DNA repair deficient), V79 cells, and hamster embryo
cells, with a dose-dependent increase in human fibroblasts and lymphocytes. Further evidence
exists for formaldehyde-induced clastogenic effect as observed by MN induction in V79 cells and a
dose-dependent increase in MN induction in both human whole blood cultures and normal and
This document is a draft for review purposes only and does not constitute Agency policy.
A-117 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
repair deficient fibroblast cells. Furthermore, formaldehyde induced SCEs in CHO cells (normal and
repair-deficient) and V79 cells at various concentrations (0.01-0.5 mM). The dose-dependent
increase in SCE was higher in mutant CHO cells compared to the normal counterparts, suggesting
the importance of DNA repair in SCE removal. Exposure of A549 cells for 1 hour with formaldehyde
or co-culturing the exposed A549 cells with unexposed V79 cells beyond 1 hour induces SCE in both
cell types, suggesting that formaldehyde is active in the medium for a longer time and continues to
induce genotoxicity in spite of the high reactivity of formaldehyde with macromolecules.
In addition, formaldehyde induces DNA repair inhibition in normal as well repair-deficient
fibroblasts derived from XP and Fanconi anemia patients. In mouse embryo fibroblasts,
formaldehyde acts as a potential initiator with a dose-dependent increase in cell transformation but
acts as a weak promoter in hamster kidney cells. Overall, there is significant evidence that
formaldehyde is genotoxic and mutagenic in several human and rodent cell culture systems.
A.4.5. Genotoxicity of Formaldehyde in Experimental Animals
In experimental animals, formaldehyde has been shown to induce DNA adducts, DPCs,
DDXs, SSBs, cytogenetic alterations, such as, MN, SCEs, CAs, and mutations, as summarized in Table
A-22.
DNA reactivity and DNA damage
Formaldehyde is highly DNA reactive. Based on numerous experimental animal studies
across several species, exposure has been shown to cause damage at the site of contact and/or
portal of entry (POE), including the formation of DNA adducts, DPXs, DDXs, SSBs and other
cytogenetic effects (see Table A-22). In addition, some animal studies have reported evidence of
effects on DNA at sites distal to the POE; however, these observations were not highly consistent
across the available studies (acknowledging that the primary focus of most studies was the POE),
and interpretations are complicated by the frequent use of test articles presumed to introduce
methanol co-exposure (see Table A-22). This limitation is of significant concern for changes
observed outside of the POE.
DNA adducts
Beland et al. (1984) demonstrated the formation of hmDNA mono adducts (e.g., N6-hmdA)
from the in vitro reaction of formaldehyde with calf thymus DNA (see Section A.4.4). The hmDNA
adducts are labile in nature and hence they were detected as methylDNA (me-DNA) adducts after
chemically reducing them with NaBH:iCN followed by LC/MS analysis (Lu etal.. 2011: Moeller etal..
2011: Lu etal.. 2010: Wang etal.. 2009: Wang etal.. 2007). Using [13CD2]-formaldehyde inhalation
exposures or orally administered [13CD4]-methanol, one research group has reported the
development of an LC/MS method that distinguishes formaldehyde-induced hmDNA mono adducts
and DNA-DNA crosslinks originating from endogenous and exogenous exposures in different
tissues of rats fLu etal.. 2012b: Lu etal.. 2011: Lu etal.. 20101 and monkeys fMoeller etal.. 20111.
This document is a draft for review purposes only and does not constitute Agency policy.
A-118 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Lu et al., (20101 exposed F344 rats to a single dose of 12.3 mg/m313CD2-formaldehyde by
inhalation for 1 and 5 days. The authors detected three forms of endogenous DNA damage, i.e., the
N2-hmdG and N6-hmdA mono adducts and dG-CH2-dG crosslinks, in all tested tissues (nose, lung
liver, spleen, bone marrow, thymus, and blood). The exogenous N2-hmdG adductand dG-CH2-dG
crosslinks were detectable only in nasal tissue and their levels increased from 1 day to 5 days of
exposure. However, the exogenous N6-hmdAdo adducts were not detectable in any of the tissues
analyzed (Lu etal.. 20101.
The same group of investigators also exposed F344 rats to inhaled [13CD2]-formaldehyde
(0.9 to 18.7 mg/m3) for 6 hours and measured N2-hmdG adducts in the nasal epithelium fLu etal..
20111. While both the endogenous and exogenous hmDNA adducts were analyzed in exposed rats,
this study did not report the use of unexposed controls. Compared to the 13C-labeled exogenous
mono adducts formed by exposures up to 11.2 mg/m3, endogenous N2-hmdG adducts formed at
levels between 1.7 and over 90-fold higher, showing considerable variation in adduct levels across
doses. Although the exogenous N2-hmdG adducts exhibited a nonlinear increase over the range of
concentrations tested, their levels appeared to be above endogenous levels only at the highest
formaldehyde concentration tested.
Further, the same group of investigators studied the distribution of hmDNA adducts in
Cynomolgus monkeys that were exposed by inhalation to 2.34 or 7.5 mg/m3 of 13CD2-formaldehyde
(6 hours/day for 2 days) (Moeller etal.. 2011). Endogenous N2-hmdG mono adducts were detected
in the nasal maxilloturbinates and bone marrow, but exogenous DNA adducts were only detectable
in the maxilloturbinates. The endogenous tissue levels of hmDNA adducts were 5-10 fold higher
than corresponding exogenous adduct levels.
Recently, another study from the same research group examined endogenous and
exogenous hm-DNA adducts in rats exposed to low levels of [13CD2]-formaldehyde (1, 30, and 300
ppb) by nose-only inhalation for 28 days fLeng et al.. 20191. The authors reported detectable levels
of endogenous, but not exogenous hm-DNA adducts in several tissues including those in lower or
upper respiratory tract (nasal epithelium, trachea and lung), blood and bone marrow, and in tissues
other than respiratory tract, bone marrow and blood cells. Thus, any exogenous formaldehyde-
induced hm-DNA adducts are below the limit of detection for exposure concentrations up to 300
ppb fLengetal.. 20191.
In addition to inhalation exposures, hmDNA adducts have been measured after exposure to
chemicals (i.e., nitrosamines, methanol) that are metabolized to formaldehyde fLu etal.. 2012b:
Wang etal.. 2007). Wang et al (2007) have detected the N6-hmdA adduct in the liver and lung of
rats injected subcutaneously with the tobacco-specific nitrosamines, N-nitrosodimethylamine
(NDMA), or 4-(methylnitrosamino)-l-(3-pyridyl)-l-butanone (NNK) atO, 0.025, and 0.01 mmol/kg
b.w. doses. The N6-hmdA adduct showed a dose-response formation with both nitrosamines and
was also detected endogenously in saline controls, albeit at low levels. Compared to saline controls,
N6-hmdA levels in exposed rats were 4.5- to 15-fold higher in the liver, and 2.2- to 3.8-fold higher in
This document is a draft for review purposes only and does not constitute Agency policy.
A-119 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
the lung. Following gavage exposure with 500 and 2,000 mg/kg [13CD4]-labeled methanol, hmDNA
adducts were detectable in several tissues of Sprague-Dawley rats, including bone marrow fLu et
al.. 2012bl. In this study, the authors also analyzed an unexposed control group. A dose-dependent
increase in exogenous N2-hmdG adducts was reported in several tissues including bone marrow,
suggesting that exogenous methanol is transported to bone marrow where it is converted to
formaldehyde and results in the formation of exogenous hmDNA adducts that are identical to
endogenous formaldehyde mono adducts. Interestingly however, the levels of endogenous N2-
hmdG adducts, but not N6-hmdA adducts, in methanol-exposed animals were significantly increased
in several tissues compared to endogenous N2-hmdG adduct levels in the corresponding tissues of
unexposed controls. This observation suggests that exposure to exogenous methanol affects the
formation and/or persistence of the endogenous N2-hmdG, but not N6-hmdA adducts, which may
have also occurred in an earlier rat study that did not report the use of unexposed controls fLu et
al.. 2011). From these studies, it appears that hmDNA adducts are likely to be formed in distal
tissues when formaldehyde is produced as a metabolite of chemicals such as methanol fLu etal..
2012b) or from NNK and NDMA (Wang etal.. 2007). Thus, oral exposure to methanol, but not
inhaled formaldehyde, seems to produce formaldehyde-specific adducts in distal tissues of
experimental animals.
DNA-protein crosslinks
Several in vivo studies involving rodents and monkeys have demonstrated DPX formation
following inhalation exposure to formaldehyde (see Table A-22). In rats, several short- and long-
term inhalation exposures of formaldehyde have been shown to induce DPX formation in nasal
passages. For example, inhalation exposure to formaldehyde induced DPX in nasal mucosa with a
single 3-hour (Casanova and Heck. 1987: Heck and Casanova. 1987) or 6-hour exposure (Casanova
etal.. 1989: Lam etal.. 19851 or 6 hours daily exposure for 2 days fCasanova-Schmitz etal.. 1984b:
Casanova-Schmitz and Heck. 1983).
DPX levels have been measured from the nasal lateral meatus, medial meatus, and posterior
meatus (Casanova etal.. 1994) or the entire nasal cavity showing a nonlinear dose-response effect
at and above 0.37 mg/m3 dose (Casanova et al.. 1989) after inhalation of 14C-formaldehyde. These
sites have been shown to be associated with a high tumor incidence fMorgan etal.. 1986bl or
cellular proliferation (Monticello etal.. 1991: Monticello etal.. 1989) in chronic formaldehyde
exposure studies in rats.
Casanova-Schmitz and Heck (1983) have reported a significant increase in DPXs in
respiratory, but not olfactory mucosa, at >7.37 mg/m3 of formaldehyde exposure of rats with a
linear increase in the exposure range of 2.46-36.8 mg/m3. The inability of this study to detect DPXs
at lower levels of formaldehyde exposure is likely due to the protective mechanism of GSH, which
catalyzes the oxidative metabolism of formaldehyde to formate. Lam et al. (1985) have shown that
co-exposure of rats with 4.6 mg/m3 acrolein and 7.4 mg/m3 formaldehyde for 6 hours resulted in
higher DPX in the nasal mucosa of rats compared to the rats given formaldehyde alone, suggesting
This document is a draft for review purposes only and does not constitute Agency policy.
A-120 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
that GSH depletion by acrolein enhanced the macromolecule binding of formaldehyde. The same
group in a different study did not detect DPX formation in the olfactory mucosa and bone marrow
even at high exposure concentration of 18.42 mg/m3 fCasanova-Schmitz etal.. 1984bl.
Casanova and Heck (1987) reported that GSH depletion caused an increase in DPX
formation in the IF-DNA of the nasal mucosa of F344 rats when a dual-isotope (3H/14C) method was
used. The dual isotope method distinguished between metabolic incorporation and covalent
binding of formaldehyde. Formaldehyde is oxidized to formate, losing one hydrogen atom
(indicated by a decrease in the 3H/14C ratio), and becomes metabolically incorporated into
macromolecules. However, when GSH is not available (depleted), it leaves residual (unoxidized)
formaldehyde to covalently bind to DNA, forming DPX. However, the residual formaldehyde may
form adducts by reacting with deoxyribonucleosides in the DNA hydrolysates, which could also lead
to an overestimation of the amount of DNA-bound formaldehyde. Casanova et al. (1989) used an
improved method which is based on the determination of the total 14C-formaldehyde bound to DNA.
This study showed that formaldehyde was exclusively bound to IF DNA, indicating the formation of
DPXs. Hydrolysis of DPXs in different samples quantitatively released formaldehyde. DPX
formation was detectable at all concentrations (0.37-12.3 mg/m3 for 6 hours) of formaldehyde
exposure. Overall, these studies show that formaldehyde induces DPXs in nasal epithelial cells of
rodents. However, there are no published rodent studies that assess DPXs beyond the nasal
passages of the upper respiratory tract. Neuss et al., (2010b) did not detect a significant increase in
DPX formation, as determined by Comet assay in the bronchoalveolar lavage (BAL) cells of F344
rats exposed up to 18.45 mg/m3 formaldehyde by whole-body inhalation compared to controls.
DPXs were also found in the nasal mucosa and extranasal tissues of rhesus monkeys
exposed to 0.86, 2.45, or 7.36 mg/m3 formaldehyde 6 hours/day for 3 days fCasanova et al.. 19911.
These data were used as a basis for cross-species prediction of formaldehyde-induced DPXs in
humans. The presence of DPXs in rhesus monkeys confirms formaldehyde's DNA reactivity as a
general effect Additionally, DPXs were detected in the larynx/trachea/carina (pooled sample) and
in intrapulmonary airways of monkeys exposed to 2.5 or 7.4 mg/m3 formaldehyde. These data
demonstrate direct effects of formaldehyde on DNA of tissues that correspond to observed tumor
sites (e.g., nasal and nasopharynx) in humans.
Recent studies by Lai et al. (2016) have shown that DPXs formed by endogenous
formaldehyde were detectable in tissues at the portal of entry (nose) as well as at distal tissues
(e.g., blood cells, and bone marrow) in rats or monkeys. However, when either species was exposed
to [13CD2]-labeled formaldehyde, exogenous DPXs were detectable only in the respiratory tissues.
In rats, exogenous DPCs accumulated over a 28-day period of exposure and remained up to one
week after removal of exposure, suggesting that DPXs might be repaired slowly (see Table A-22).
Recently, another study from the same research group examined endogenous and
exogenous DPX adducts in rats exposed to low levels of [13CD2]-formaldehyde (1, 30, and 300 ppb)
by nose-only inhalation for 28 days fLeng etal.. 20191. The authors reported detectable levels of
This document is a draft for review purposes only and does not constitute Agency policy.
A-121 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
endogenous, but not exogenous DPXs in several tissues including those in lower or upper
respiratory tract (nasal epithelium, trachea and lung), blood and bone marrow, and in tissues other
than respiratory tract, bone marrow and blood cells. Thus, any exogenous formaldehyde-induced
DPX adducts are below the limit of detection for exposure concentrations up to 300 ppb fLengetal..
20191.
DNA-DNA crosslinks
There is limited evidence showing the formation of DNA-DNA crosslinks (DDX) induced by
inhalation exposure to formaldehyde. Lu et al. (2010) reported dG-CH2-dG crosslinks in the nasal
epithelium of F344 rats exposed to 12.3 mg/m3 formaldehyde for 1 or 5 days (6 hours/day).
However, roughly 65% of the dG-CH2-dG crosslinks were considered artifacts formed during
sample workup and storage. Wang et al. (2007) reported very low levels of dA-CH2-dA crosslinks
of formaldehyde in rats exposed to NDMA and NNK, but cautioned that these crosslinks may be
generated artifactually upon DNA storage. Thus, the DDX may not be a useful biomarker of
formaldehyde exposure.
DNA SSBs by alkaline elution
Formaldehyde has been shown to induce DNA SSBs in few studies involving mice (Wang
and Liu. 20061 and rats fSul etal.. 2007: Im etal.. 20061. as summarized in Table A-22.
Im et al. (2006) reported a dose-dependent increase in DNA damage as analyzed by the
comet assay in both PBLs and livers of Sprague-Dawley rats exposed by inhalation to 6.14 and 12.3
mg/m3 formaldehyde. In the same strain of rats, Sul et al (2007) also observed a dose-dependent
increase in SSBs in lung epithelial cells following inhalation exposure to 0, 6.15 and 12.3 mg/m3
formaldehyde for 2 weeks (6 hours/day, 5 days/wk). In a developmental toxicity study, pregnant
mice injected i.p. with formaldehyde from gestational days 6 to 19 exhibited DNA damage in
maternal as well as fetal liver at 0.2 and 1 mg/kg respectively fWang and Liu. 20061.
Cytogenetic markers of genotoxicity
Micronucleus
Few studies examined the effect of formaldehyde exposure on MN induction in rodents by
exposing the animals by inhalation, i.p. injection, or gavage as summarized in Table A-22.
Inhalation exposure studies in rats were negative, while studies that used formalin by gavage in
mice fWard etal.. 19831 and rats fMigliore etal.. 19891 were positive for MN formation. Speitand
coworkers did not observe MN formation in the peripheral blood cells (Speit etal.. 2009) and BAL
cells (Neuss etal.. 2010b) of F344 rats exposed to 0, 62,1.23, 7.38,12.3, and 18.45 mg/m3
formaldehyde. However, the Neuss et al (2010b) study did not report the use of a positive control
for MN induction, while in the other two studies, the use of cyclophosphamide as a positive control
did not appear to induce a high MN count or showed results within the range of control values
This document is a draft for review purposes only and does not constitute Agency policy.
A-122 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
(Speitetal.. 2011b: Speit etal.. 2009). Ward et al. (19831 observed aneuploidy and structural
chromosomal aberrations (e.g., breaks, exchanges, aberrant chromosomes with and without gaps)
in femoral bone marrow cells of mice dosed with formalin (100 mg/kg) or methanol (1000 mg/kg).
The cytogenetic effects seen in bone marrow suggest that the formalin or methanol given by gavage
was able to reach bone marrow and induce genotoxicity. Similarly, Migliore et al. (1989) observed
MN formation in the gastric epithelial cells of Sprague-Dawley rats exposed to a single dose of
formalin (200 mg/kg). Lastly, (Liu etal.. 2017) have shown that inhalation exposure to
formaldehyde in ICR mice for 20 weeks caused a significant increase in the ratio of polychromatic
erythrocytes/normochromatic erythrocytes, but not micronuclei induction in bone marrow (Liuet
al.. 20171.
Sister chromatid exchanges
Few studies examined the effect of formaldehyde exposure on SCEs in mice and rats. Two
of the three studies in rats were negative for SCEs in blood cells (Speitetal.. 2009: Kligerman etal..
1984). both of these studies used inhalation exposure to 18.45 mg/m3 formaldehyde for 6
hours/day, 5 d/wk for 4 weeks.
In an inhalation study, Brusick (1983) exposed CD-I mice to target concentrations of 0,
7.38,14.76 or 30.75 mg/m3 formaldehyde vapors for 6 hours/day for 4-5 days. Significantly high
levels of SCEs/cell were reported in the bone marrow of female mice both at the mid and high
concentrations, while the low-concentration group had levels that were not statistically significant
from the control group. Thus, formaldehyde exposure has provided equivocal results on the SCEs
in rodents.
Chromosomal aberrations
Few studies reported the effect of formaldehyde inhalation on CA induction in rodents and
these results were mixed (see Table A-22).
Kligerman et al. (1984) found no difference in the incidence of SCEs or CAs and mitotic
index in the PBLs of male and female F344 rats exposed to formaldehyde for 5 days up to 18.45
mg/m3 dose. Also, Dallas et al. (1992) reported no clastogenic effects in bone marrow of Sprague-
Dawley rats exposed at the same concentration of formaldehyde for 8 weeks. However, the authors
observed a modest, but statistically significant increase (1.7 to 1.8 fold) in CAs in pulmonary lavage
cells at the high dose (18.45 mg/m3) compared to controls, but not at lower doses (0.61 and 3.7
mg/m3 fDallas etal.. 19921.
Speit et al (2009) investigated the genotoxicity of formaldehyde in peripheral blood
samples of Fischer-344 rats exposed to 0 to 18.45 mg/m3 formaldehyde for 4 weeks (6 hours/day,
5 days/week). Compared to controls, the authors found no significant increase in genotoxicity
assays such as the comet assay (with or without y-irradiation of blood samples), the SCEs assay, and
micronucleus test Earlier studies by Casanova-Schmitz et al. (1984b) showed that formaldehyde
does not cause toxicity to bone marrow. Following formaldehyde exposure by i.p. injection in mice,
This document is a draft for review purposes only and does not constitute Agency policy.
A-123 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
data were negative for CAs in spermatocytes (Fontignie-Houbrechts etal.. 1982: Fontignie-
Houbrechts. 19811 and polychromatic erythrocytes fNataraian etal.. 19831. while Gomaa et al.
f20121 demonstrated an increase in chromosomal aberrations in bone marrow cells of adult male
albino rats exposed to formaldehyde at 0.2 mg/kg/day i.p injection for 4 weeks. Oral
administration of formaldehyde to rats showed positive results for CAs in the gastric epithelial cells
(Migliore etal.. 1989).
Since many leukemogens initiate leukemogenesis by directly damaging the hematopoietic
stem cells/hematopoietic progenitor cells (HSP/HPC), Zhao et al. (Zhao etal.. 20201 examined the
effect of formaldehyde exposure either in vivo or ex vivo. They exposed either BALB/c mice to 3
mg/m3 formaldehyde by inhalation for 2 weeks or by ex vivo to cells from bone marrow, lung, nose,
and spleen with 0, 50,100, and 400 |a,M formaldehyde for 1 hour. Using a myeloid progenitor colony
formation (MPCF) assay, they have shown that formaldehyde exposure caused a decrease in bust-
forming unit-erythroid (BFU-E) and colony-forming unit-granulocyte, macrophage (CFU-GM)
colonies in all the four tissues from both in vivo and ex vivo (up to 400 |j.M) exposure to
formaldehyde. The authors conclude that their study confirms the presence of HSP/HPC in mouse
lung and nose and hypothesize that following formaldehyde-induced DNA damage at the point of
entry these damaged stem cells possibly migrate to bone marrow and induce leukemia (Zhao etal..
20201. However, the formaldehyde used in this study was generated from 10% formalin which
contains methanol added as a stablizer; it is likely that methanol could also contribute to the
outcome, preventing attribution of the results to formaldehyde alone.
Overall, inhalation exposure to formaldehyde has produced mixed and equivocal results in
rodents for cytogenetic markers of genotoxicity. Formaldehyde did not induce MN in bone marrow
cells of male Sprague-Dawley rats (Dallas etal.. 19921 and caused no increase in the frequency of
SCEs or CAs and mitotic index in blood lymphocytes of F344 rats of either sex (Kligerman etal..
19841. However, a modest, but statistically significant, increase (1.7- to 1.8-fold) in CAs has been
observed in pulmonary lavage cells of Sprague-Dawley rats after exposure to 18.45 mg/m3 (Dallas
etal.. 19921 and a significant increase in CAs in bone marrow cells of female Wistar rats exposed to
1.5 mg/m3 formaldehyde (Kitaeva et al.. 1990): however, the latter finding involved methanol co-
exposure, reducing confidence in these results. Also, formaldehyde exposure by inhalation in CD-I
mice induced SCEs in bone marrow cells at «15 mg/m3 fBrusick. 19831. Thus, some studies show
that inhaled formaldehyde may be able to induce cytogenetic effects in distal tissues with repeated
exposures, possibly only at very high formaldehyde concentrations.
Mutations
Formaldehyde exposure has been shown to induce mixed results for mutations in several
test systems as summarized in Table A-22. The dominant lethal mutation test has been performed
using mice and rats, where males were exposed to formaldehyde or formalin vapors by inhalation
or i.p. injection, mated with females, and where mutations were then scored in the offspring. In two
This document is a draft for review purposes only and does not constitute Agency policy.
A-124 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
of these studies, formaldehyde injected i.p. to CD-I mice was negative for dominant lethal
mutations fEpstein etal.. 1972: Epstein and Shafner. 19681. while another study which used a
higher dose (50 mg/kg) of formaldehyde showed weakly positive results fFontignie-Houbrechts.
19811. Specific pathogen-free ICR mice exposed to inhaled formaldehyde were positive for
dominant lethal mutations fLiu etal.. 2009bl. In this study mutation rates were dose dependent
and mainly inherited from the paternal germ line.
Recio et al. (19921 demonstrated point mutations in the GC base pairs of the p53 tumor
suppressor gene in 45% (5 out of 11) of the primary nasal squamous cell carcinomas (SCCs) from
F344 rats that were chronically (2 yrs) exposed to 18.45 mg/m3 formaldehyde. Samples from this
study were further analyzed by Wolf et al. (1995) who demonstrated the presence of p53 tumor
suppressor protein which correlated with proliferating cell nuclear antigen (PCNA) but not TGF-
alpha in the nasal SCCs. However, Meng et al. (2010) failed to detect the p53 mutations in the nasal
mucosa of rats exposed to 0.86 to 18.42 mg/m3 formaldehyde for 13 weeks. It is likely that the
duration of exposure is important for the mutations to occur in these studies. In summary,
formaldehyde produced mixed results in the DLM test. Short-term (13-week) exposure of rats to
formaldehyde did not produce detectable mutations in the p53 tumor suppressor gene or Ha-ras
oncogene; however, a chronic 2-yr study resulted in SCC formation and mutations in the GC base
pairs of the p53 gene in rats.
Table A-22. Summary of in vivo genotoxicity studies of formaldehyde
inhalation exposure in experimental animals
Test system
Concentration3
Results'5
Comments
Reference
Mutation
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Rats/F344, nasal SCCs
18.45 mg/m3; HCHO
from PFAC
+
Inhalation, 6 hrs/day, 5
days/wk, 2 yrs
(Recio et al.,
1992)
Rats/F344, nasal SCCs
18.45 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, 5
days/wk, 2 yrs
(Wolf et al.,
1995)
Rats/F344, nasal
mucosa
18.45 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, 5
days/wk,13 wks; Cell
proliferation showed a
conc.-dependent f;
significant at 12.3 and
18.45 mg/m3 exposures
(Meng et al.,
2010)
Evaluations specific to genotoxicity to systems other than the respiratory tract, bone marrow, or blood cells
Rats/Strain not
specified - dominant
lethal test
1.47 mg/m3; HCHO
(not specified)
(+)
Inhalation, 4 hrs/day, for 4
wks
(Kitaeva et al.,
1990)
Mice/ICR, specific
pathogen-free
dominant lethal test
200 mg/m3; Formalin
(37% HCHO w/w
aq.sol.)
+
Whole-body inhalation
exposure of $ mice for 2
hrs; 6 wks postexposure cf
mated to $ at 1:1;
(Liu et al.,
2009b)
DNA-protein crosslinks
This document is a draft for review purposes only and does not constitute Agency policy.
A-125 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Results'5
Comments
Reference
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Monkey/Rhesus
nasal turbinates
0.86 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs; the LEC
"l^with the 1" in distance
from the portal of entry;
DPX levels show conc.-
dependent ^from
0.86-7.4 mg/m3, in the
order of middle turbinates
> lateral wall/septum,
nasopharynx >
larynx/trachea/carina.
(Casanova et
al.. 1991)
Monkey/Rhesus
nasal, larynx, trachea, &
carina
2.5 mg/m3; HCHO
from PFA
+
(Casanova et
al.. 1991)
Monkey/Rhesus
maxillary sinuses, lungs
7.4 mg/m3; HCHO
from PFA
+
(Casanova et
al.. 1991)
Monkeys/Cynomolgus
nose
7.4 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, for 2
days
(Lai et al., 2016)
Rats/F344
nasal mucosa
0.37 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs;
nonlinear conc.-
dependent|in DPX
between 0.37 to 12.1
mg/m3
(Casanova et
al.. 1989)
Rats/F344
nasal mucosa
0.86 mg/m3; HCHO
from PFA
+
Inhalation 6 hrs/day, 5
days/wk, 11 wk + 4 d + 3
hrs (preexposed); or 3 hrs
only (naive); fcell
proliferation > 7.48 mg/m3
(Casanova et
al.. 1994)
Rats/F344
nasal mucosa
2.5 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, for 2
days; cytotoxicity > 12.3
mg/m3
(Casanova-
Schmitz et al.,
1984a)
Rats/F344
nasal mucosa
2.5 mg/m3; HCHO
from PFA
+
Inhalation, 3 hrs/day, for 2
days
(Casanova and
Heck. 1987)
Rats/F344
nasal mucosa
2.5 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day; for 7
or 28 days
(Lai et al., 2016)
Rats/F344
nasal mucosa
7.4 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, for 2
days
'Casanova-
Schmitz and
Heck. 1983>b
Rats/F344
nasal mucosa
7.4 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs; co-
exposure to 2 ppm acrolein
caused a significant f in
toxicity and DPX formation
(Lam et al.,
1985)
Rats/F344
nasal mucosa
18.45 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day; for
1,2, and 4 days
(Lai et al., 2016)
Rats/F344
olfactory mucosa
18.45 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for 2
days
(Casanova-
Schmitz et al.,
1984a)
This document is a draft for review purposes only and does not constitute Agency policy.
A-126 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Results'5
Comments
Reference
36.9 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for 2
days
(Casanova-
Schmitz and
Heck. 1983)b
Rats/F344, nasal
epithelium, trachea,
lung
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Rats/F344
BAL cells
18.45 mg/m3; HCHO
from formalin vapors
-
Inhalation, 6 hrs/day, 5
days/wk, for 4 wks
{Neuss, 2010,
1578360
Mice/BalbC
lung
3.0 mg/m3; HCHO
vapor from 10%
formalin
Inhalation, nose-only; 8
hours/day for 7 days;
(Ye et al.,
2013b)
Evaluations specific to genotoxicity in cells of the blood and bone marrow
Monkeys/Cynomolgus
bone marrow, PBMC
7.4 mg/m3; HCHO
from PFA
"
Inhalation, 6 hrs/day, for 2
days
(Lai et al., 2016)
Rats/F344
bone marrow
12.43 mg/m3; HCHO
from PFA
"
Inhalation, 3 hrs/day, for 2
days
(Casanova and
Heck. 1987)
Rats/F344
bone marrow
18.45 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for 2
days
(Casanova-
Schmitz et al.,
1984a)
Rats/F344
bone marrow, PBMC
18.45 mg/m3; HCHO
from PFA
-
Inhalation, 6 hrs/day; for
1,2, and 4 days
(Lai et al., 2016)
Rats/F344, bone
marrow, PB MC
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
~
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Rats/F344
peripheral blood
18.45 mg/m3; HCHO
from formalin vapors
"
Inhalation, 6 hrs/day, 5
days/wk, for 4 wks
(Speit et al.,
2009)
Mice/BalbC
bone marrow
1.0 mg/m3; HCHO
vapor from 10%
formalin
+
Inhalation, nose-only; 8
hours/day for 7 days;
dose-dependent T* in DPC
(Ye et al.,
2013b)
Mice/BalbC
PBM cells
3.0 mg/m3; HCHO
vapor from 10%
formalin
+
Inhalation, nose-only; 8
hours/day for 7 days;
dose-dependent T* in DPX
(Ye et al.,
2013b)
Evaluations specific to genotoxicity in systems other than the respiratory tract, bone marrow or cells of the blood
Monkeys/Cynomolgus
liver
7.4 smg/m3; HCHO
from PFA
-
Inhalation, 6 hrs/day, for 2
days
Lai etal. f20161
Rats/F344, olfactory
bulbs, liver, hippo
campus, cerebellum
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Mice/Kunming
kidney & testes
0.5 mg/m3; HCHO
vapor from 10%
formalin
+
Inhalation, 72 hrs
continuous exposure
(Peng et al.,
2006)
This document is a draft for review purposes only and does not constitute Agency policy.
A-127 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Results'5
Comments
Reference
Mice/Kunming
liver
1.0 mg/m3; HCHO
vapor from 10%
formalin
+
Inhalation, 72 hrs
continuous exposure
(Zhao et al.,
2009; Peng et
al.. 2006)
Mice/BalbC
spleen, testes
1.0 mg/m3; HCHO
vapor from 10%
formalin
+
Inhalation, nose-only; 8
hours/day for 7 days;
dose-dependent T* in DPX
(Ye et al.,
2013a)
DNA adducts
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Monkey/Cynomologus
maxilloturninate
2.33 mg/m3; HCHO
(not specified)
+
Inhalation, 6 hrs/day, for 2
days; conc.-dependent f in
exogenous adducts
(Moeller et al.,
2011)
Monkeys/Cynomolgus -
nasal dorsal mucosa,
nasopharynx, nasal
septum, nasal posterior
maxillary
7.5 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, for 2
days;
(Yu et al.,
2015b)
Monkeys/Cynomolgus -
trachea carina, trachea
proximal
7.5 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for 2
days;
(Yu et al.,
2015b)
Rats/F344
nasal epithelium
0.86 mg/m3; HCHO
from PFA
+
Inhalation, for 6 hrs; conc.-
dependent f in exogenous
adducts
(Lu et al., 2011)
Rats/F344
nasal epithelium
2.46 mg/m3; HCHO
from PFA
+
Inhalation, 6 hrs/day, for 7,
14,21, or 28 days; recovery
for 6, 24, 72, or 168 hours;
exposure-dependent T*
hmdG mono adducts
(Yu et al.,
2015b)
Rats/F344 -nasal
epithelium
12.3 mg/m3; 20%
HCHO in water
+
Inhalation, 1 and 5 days;
exposure-dependent f in
exogenous hmdG adduct
and dG-dG crosslinks
(Lu et al., 2010)
Rats/F344
lung
12.3 mg/m3; HCHO
from PFA
-
Inhalation, 1 and 5 days
(Lu et al., 2010)
Rats/F344, nasal
epithelium, trachea,
lung
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Evaluations specific to genotoxicity in cells of the blood and bone marrow
Monkey/Cynomologus
bone marrow
2.33 mg/m3; HCHO
(not specified)
—
Inhalation, 6 hrs/day, for 2
days;
Moeller et al., 2011
Monkeys/Cynomolgus
bone marrow, white
blood cells
7.5 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for 2
days;
(Yu et al.,
2015b)
Rats/F344
white blood cells and
bone marrow cells
12.3 mg/m3; HCHO
from PFA
Inhalation, 1 and 5 days
(Lu et al., 2010)
Rats/F344, bone
marrow, PB MC
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Evaluations specific to genotoxicity in systems other than the respiratory tract, bone marrow or cells of the blood
This document is a draft for review purposes only and does not constitute Agency policy.
A-128 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Results'5
Comments
Reference
Rats/F344
thymus, lymph nodes,
trachea, lung, spleen,
kidney, liver, brain
2.46 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, for
28 days;
(Yu et al.,
2015b)
Rats/F344
liver, spleen, thymus
12.3 mg/m3; HCHO
from PFA
—
Inhalation, 1 and 5 days
(Lu et al., 2010)
Rats/F344, olfactory
bulbs, liver, hippo
campus, cerebellum
0.0012, 0.0369, 0.369
mg/m3 [13CD2]-HCHO
Inhalation, nose-only, 6
h/d, 28 days
(Leng et al.,
2019)
Chromosomal aberrations
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Rats/SD Pulmonary
lavage cells
18.45 mg/m3; HCHO
from PFA
+
Inhalation, whole body; 6
hrs/day, 1 or 8 wks
(Dallas et al.,
1992)
Evaluations specific to genotoxicity in cells of the blood and bone marrow
Rats/Wistar
Bone marrow
0.49 mg/m3; HCHO
(not specified)
+
Inhalation, 4 hrs/day, 4
months
(Kitaeva et al.,
1990)
Rats/SD
Bone marrow
18.45 mg/m3; HCHO
from PFA
"
Inhalation, whole body; 6
hrs/day, 1 or 8 wks
(Dallas et al.,
1992)
Rats/F344 Peripheral
blood cells
18.45 mg/m3; HCHO
from PFA
"
Inhalation, 6 hrs/day, 5
days/wk, for 4 wks
(Speit et al.,
2009)
Rats/F344
Lymphocytes
18.45 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, 5
days; no significant dose-
related effect on mitotic
activity
(Kligerman et
al.. 1984)
Mice/CD-I, male &
female, Bone marrow
cells
30.75 mg/m3; HCHO
from PFA
Inhalation, 6 hrs/day, 4-5
days;
(Brusick, 1983)
Mice/BALB/c, bone
marrow-
hematopoietic stem
and progenitor cells
3 mg/m3, HCHO from
10% formalin
+
Inhalation, 8 h/d, 5d/wk, 2
weeks
(Zhao et al.,
2020)
Micronucleus
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Rats/F344
BAL cells
18.45 mg/m3; HCHO
from formalin vapors
Inhalation, 6 hrs/day, 5
days/wk, for 4 wks;
positive control was not
used for the assay
(Neuss et al.,
2010a)
Evaluations specific to genotoxicity in cells of the blood and bone marrow
Rats/Outbred white
polychromatophylic
erythrocytes (bone
marrow)
12.8 mg/m3,
commercial
formaldehyde
+
Inhalation; whole-body
exposure; 4 hrs/day, 5
days/wk
(Katsnelson et
al.. 2013)
Rats/F344 -peripheral
blood
18.45 mg/m3; HCHO
from formalin vapors
"
Inhalation, 6 hrs/day, 5
days/wk, for 4 wks
(Speit et al.,
2009)
Mice/male ICR
bone marrow cells
20 mg/m3 36.5%-38%
HCHO in water
(formalin)
+
Inhalation, 2 hrs/day for 15
days
(Yu et al.,
2014a)
This document is a draft for review purposes only and does not constitute Agency policy.
A-129 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Results'5
Comments
Reference
Mice/ICR, bone marrow
1,10 mg/m3, HCHO
-
Inhalation, 2 h/d, 20
(Liu et al., 2017)
cells
source not reported
weeks; micronucleus
Single strand breaks
Evaluations specific to genotoxicity in the upper or lower respiratory tract
Rats/SD
lung epithelial cells
6.14 mg/m3; HCHO
(commercial)
+
Inhalation, 6 hrs/day, 5
days/wk for 2 wks;
j cytotoxicity (lipid
peroxidation & protein
carbonyl oxidation)
observed at 18.42 mg/m3
(Sul et al., 2007)
Evaluations specific to genotoxicity in blood cells
Rats/SD, PBLs
6.14 mg/m3; HCHO
+
Inhalation, 5 days/wk for 2
(Im et al., 2006)
(commercial)
wks
Evaluations specific to genotoxicity in systems other than the respiratory tract, bone marrow or blood cells
Rats/SD, liver
6.14 mg/m3; HCHO
+
Inhalation, 5 days/wk for 2
(Im et al., 2006)
(commercial)
wks
Sister chromatid exchanges
Evaluations specific to genotoxicity in cells of the blood and bone marrow
Rats/F344
18.45 mg/m3; HCHO
-
Inhalation, 6 hrs/day, 5
(Kligerman et
Lymphocyte
from PFA
days; no significant dose-
related effect on mitotic
activity
al.. 1984)
Rats/F344
18.45 mg/m3;
-
Inhalation, 6 hrs/day, 5
(Speit et al..
Peripheral blood cells
Formalin vapors
days/wk, for 4 wks
2009)
Mice/CD-I, male &
14.76 mg/m3; HCHO
-/ +
Inhalation, 6 hrs/day, 5
(Brusick, 1983)
female Bone marrow
from PFA
days; cf mice: -ve; 9 mice:
cells
+ve; conc.-dependent T*
in SCEs
Gray shading indicates experiments examining tissues or cells outside of the upper respiratory tract that are
assumed to have included co-exposure to methanol, and are thus may be less reliable.
aLowest effective concentration (LEC) for positive results or highest ineffective concentration tested (HIC) for
negative or equivocal results.
b+ = positive; - = negative; (+), equivocal.
Thermal depolymerization of paraformaldehyde (PFA) or freshly prepared formalin (no methanol) are the
preferred test article methods. Generation of formaldehyde from formalin, uncharacterized aqueous solutions
(noted as not specified), or an unspecified source (also noted as not specified) is assumed to involve co-exposure
to methanol, and the evidence is less reliable.
HCHO, formaldehyde; PFA, paraformaldehyde; hmDNA, hydroxymethylDNA; SCE, sister chromatid exchange; SCC,
squamous cell carcinoma; hmdA, hydroxymethyl deoxyadenosine; hmdG, hydroxymethyl deoxyguanosine; MN,
micronucleus.
Part of the data adapted from NTP (2010).
Table A-23. Summary of in vivo genotoxicity studies of formaldehyde
exposure by intraperitoneal and oral routes of exposure in experimental
animals
Test system
Concentration3
Results'5
Comments
Reference
Mutation
This document is a draft for review purposes only and does not constitute Agency policy.
A-130 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Resultsb
Comments
Reference
Rats/Albino
Spermatocyte; DLM
0.125 mg/kg; test
article: 37% HCHO (+
10% methanol)
+
i.p., $ given 5 daily doses and
mated to $; dose-dependent \ in
DLM index; effects greater with
shorter time gap postexposure
(Odeigah, 1997)
Mice/CD-I DLM test
20 mg/kg HCHO; test
article: Not Specified
i.p. injection to mated to $ and
autopsied 13 d past mid-wk of
mating
(Epstein and
Shafner, 1968)
DNA-protein crosslinks
Rats/F344
tracheal implants
0.01% HCHO in PBS;
test article: Not
Specified
+
instillation, twice weekly for 2, 4,
or 8 wks
(Cosma et al.,
1988)
Mice/NS
liver (Fetal) [Chinese
lang-English Abstract]
0.2 mg/kg; test article:
HCHO (not specified)
+
i.p. injection to pregnant mice
from GD 6 to 19
(Wang and Liu,
2006)
Mice/NS
Liver (maternal)
[Chinese lang-English
Abstract]
20 mg/kg; test article:
HCHO (not specified)
i.p. injection to pregnant mice
from GD 6 to 19
(Wang and Liu,
2006)
Chromosomal aberrations
Mice/CBA
femoral polychromatic
erythrocytes
25 mg/kg; test article:
HCHO (PFA in water)
i.p. injections (two) within 24 hr
interval; cells sampled 16 and 40
hrs post 2nd inj.
(Nataraian et
al.. 1983)
Mice/Q strain
Spermatocytes
50 mg/kg; test article:
HCHO (35% sol.)
i.p. injection, single
(Fontignie-
Houbrechts,
1981)
Mice/Q strain
Spermatogonia
30 mg/kg; test article:
HCHO (commercial)
i.p., 35% HCHO solution + 90
mg/kg H202
(Fontignie-
Houbrechts et
al.. 1982)
Rats/SD
gastric epithelial cells
(stomach, duodenum,
ileum, colon)
200 mg/kg; test article:
HCHO (in water)
+
p.o., 16, 24, or 30 hrs; time-
dependent 'X in CA in all tissues;
toxic at 30 hrs; no significant
change in mitotic index
(Migliore et al.,
1989)
Mice/B6C3Fl-bone
marrow
100 mg/kg; test article:
formalin; or 1,000
+
Gavage, single exposure; HCHO
and methanol showed 21- and
15-fold increase compared to
controls, respectively
(Ward et al..
mg/kg methanol
1983)
Rats (male albino),
bone marrow cells
0.2 mg/kg/day; test
article: HCHO (source
not specified)
+
i.p injection, single injection for 4
wks
(Gomaa et al.,
2012)
Micronucleus
Mice/CBA
femoral polychromatic
erythrocyte and spleen
cell
25 mg/kg; test article:
HCHO (PFA in water)
i.p. injections (two) of HCHO
solution within 24 hr interval; cells
sampled 16 and 40 hrs post 2nd
inj.
(Nataraian et
al.. 1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-131 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Supplemental Information for Formaldehyde—Inhalation
Test system
Concentration3
Resultsb
Comments
Reference
Mice/NMRI
bone marrow
30 mg/kg; test article:
HCHO (commercial)
—
i.p. injection, single
(Gocke et al.,
1981)
Mice/CD-I
reticulocytes
30 mg/kg; test article:
HCHO (35%)
—
i.v. two injections; sampled 24,48,
or 72 hrs after exposure
(Morita et al.,
1997)
Mice/CD-I
bone marrow or
peripheral blood
200 mg/kg; test article:
35% HCHO
Gavage twice (bone marrow) or
once (peripheral blood); all mice
killed at 300 mg/kg dose
(Morita et al.,
1997)
Rats/SD
gastric epithelial cells
(stomach, duodenum,
ileum, colon)
200 mg/kg; test article:
HCHO (in water)
+
p.o., 16, 24, or 30 hrs; time-
dependent 1" in MN in all tissues;
toxic at 30 hrs; no significant
change in mitotic index
(Migliore et al.,
1989)
aLowest effective concentration (LEC) for positive results or highest ineffective concentration (HIC) tested for
negative or equivocal results.
b+ = positive; - = negative; (+), equivocal.
Thermal depolymerization of paraformaldehyde (PFA) or freshly prepared formalin (no methanol) are the
preferred test article methods. Generation of formaldehyde from formalin, uncharacterized aqueous solutions
(noted as not specified), or an unspecified source (also noted as not specified) is assumed to involve co-exposure
to methanol, and the evidence is less reliable.
HCHO, formaldehyde; PFA, paraformaldehyde; DLM, dominant lethal mutation; i.p., intra peritoneal; i.v., intra
venous; GD, gestation day; MN, micronucleus;
Part of the data adapted from NTP (2010).
Summary of in vivo genotoxicity studies of formaldehyde by routes of exposure in experimental
animals
Formaldehyde reacts with cellular macromolecules at the portal of entry causing
genotoxicity. Genotoxicity of inhaled formaldehyde involves direct interaction with DNA inducing
DNA-protein crosslinks and/or hydroxymethylDNA adducts or DNA mono adducts, single strand
breaks, micronuclei, and chromosomal aberrations in nasal passages of experimental animals. DPX
are formed predominantly by crosslinking of the epsilon-amino groups of lysine and the exocyclic
amino groups of DNA, especially the N-terminus of histone. Due to the differences in the anatomy
of nasal passages and breathing patterns of rats and monkeys, the location of DPX formation differs.
Over a range of 0.86 to 7.37 mg/m3, formaldehyde-induced DPX levels showed concentration-
dependent increase in monkey respiratory tract in the order of middle turbinates > anterior lateral
wall/septum > maxillary sinuses and lungs. Thus, the lowest effective concentration (LEC) being
higher with increase in the anatomical distance from the portal of entry. Furthermore, these
anatomical sites are known to be associated with formaldehyde-induced proliferative response in
monkeys. In rats, DPX formation showed concentration dependence between 0.37-12.1 mg/m3
formaldehyde, which was nonlinear with a sharp increase above 4.9 mg/m3. With exposures up to
28 days, DPXs were shown to accumulate and persisted for an additional 7 days at a concentration
of 2.5 mg/m3. In addition, DPX formation was six-fold higher in the lateral meatus compared to the
medial and posterior meatus, corresponding respectively, to high and low tumor incidence sites in
rats. DPXs were not detected in olfactory mucosa, bronchoalveolar lavage (BAL) cells of rats or in
This document is a draft for review purposes only and does not constitute Agency policy.
A-132 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
lungs of mice exposed to formaldehyde. DPXs (from exogenous formaldehyde) also were not
detected in bone marrow and peripheral blood monocyte cells (rats and monkeys) and liver
(monkeys) following inhalation exposure. Since DPXs are likely to induce replication errors, they
have been considered to be a marker of mutagenicity. The repair of DPX in eukaryotes appears to
depend on the dose and duration of formaldehyde exposure. The overall evidence indicates that
the DPXs are markers of exposure as well as genotoxic endpoints.
HydroxymethylDNA adducts in experimental animals can result from DNA reacting with
endogenously-produced or exogenous formaldehyde. Mono adducts formed from endogenous
formaldehyde (produced during normal cellular metabolism) are distinguished from those formed
by exogenous exposure using stable isotope (13C)-labeled formaldehyde coupled with sensitive MS
techniques. Inhaled formaldehyde induces N2-hmdG adducts in the nasal epithelium of F344 rats,
but not in distal tissues, and the adduct levels are associated with concentration and duration of
exposure. In rhesus monkeys, formaldehyde induces N2-hmdG adducts in the maxilloturbinates,
and the mono adduct levels are associated with the exposure concentration of formaldehyde.
Endogenous N2-hmdG mono adducts and dG-dG crosslinks are also detected in rats and monkeys,
but in all experimental animals exposed exogenously to formaldehyde by inhalation, N2-hmdG
adducts were only elevated in nasal passages, not in tissues beyond the portal of entry. However,
formaldehyde-specific hmDNA adducts have been detected in rodent tissues distal to the portal of
entry when the animals were exposed to methanol or nitrosamines, which are known to release
formaldehyde as a metabolic intermediate in vivo. These studies suggest the lack of transport of
formaldehyde beyond the portal of entry when given by inhalation in animals. Although the
hmDNA adducts are considred to be genotoxic endpoints of formaldehyde exposure, their
mutagenicity has not been enstablished.
There is limited evidence about mutagenicity of formaldehyde in experimental animals.
Formaldehyde did not induce mutations in the nasal mucosa of rats with inhalation exposure to
18.5 mg/m3 for 13 weeks, but there are no available studies involving longer periods of exposure.
However, formaldehyde inhalation exposure caused other genotoxic endpoints, including
chromosomal aberrations and single strand breaks but not micronuclei in cells of respiratory
system.
Twelve out of 17 that analyzed formaldehyde-induced genotoxic endpoints in bone marrow
or blood cells were negative. Conflicting results have been obtained in terms of source of
formaldehyde. Formaldehyde derived from paraformaldehyde or commercial formalin was
negative for DPX formation in bone marrow and peripheral blood cells, although one recent study,
which used 10% formalin as a source of formaldehyde, induced DPX in bone marrow and
peripheral blood mononuclear cells. Formaldehyde did not induce hmDNA adducts in the bone
marrow of monkeys and rats, suggesting that inhaled exogenous formaldehyde may not be
transported to the tissues distal to the portal of entry. Formaldehyde failed to induce CAs in 4/5
studies in the bone marrow or peripheral blood cells of rats and mice (see Table A-22), although
This document is a draft for review purposes only and does not constitute Agency policy.
A-133 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
one study detected CAs in bone marrow of rats. Limited available evidence shows that inhaled
formaldehyde did not induce micronuclei in the peripheral blood cells of rats, but was positive for
inducing SSBs in peripheral blood and bone marrow cells and produced mixed results on SCE
formation. The above studies clearly indicate the complexicity of data analyses with contradicting
results in the same assay sytem, type of exposure, and/or methodology utilized.
Formaldehyde produced mixed results in tissues other than the respiratory and
hematopoietic systems (see Table A-23). Three studies demonstrated DPX formation in mouse
kidney, testes, liver and spleen when 10% formalin was used as a source of formaldehyde. Inhaled
formaldehyde did not induce hmDNA adducts in the liver, spleen, and thymus of rats, but SSBs were
detectable in the liver of rats following inhalation exposure.
Several studies evaluated the genotoxicity and mutagenicity of formaldehyde by routes
other than inhalation exposure and reported mixed results (see Table A-23), suggesting that
formaldehyde induced genotoxicity might depend on the route of exposure and formulation of
formaldehyde administered.
A.4.6. Genotoxic Endpoints in Humans
A large set of research studies in several countries, involving different exposure settings,
found that exposure to formaldehyde is associated with damage or changes to human DNA that
inform mechanisms of carcinogenesis. These studies have observed increased levels of DNA
damage, DNA-protein crosslinks, and chromosomal breaks in buccal and nasal epithelial cells, and
peripheral blood lymphocytes. Chromosomal damage, manifested as an increased frequency of
different types of chromosomal aberrations, has been reported. It has been shown that increased
frequency of chromosomal aberrations and micronuclei are associated with increased cancer
mortality, and these endpoints are considered by EPA to be highly relevant to the assessment of
genotoxicity in humans fBonassi etal.. 2011: Bonassi etal.. 2008: Bonassi etal.. 2007: U.S. EPA.
2005: Bonassi et al.. 2004b). Single strand breaks in DNA, indicating genetic instability also are
considered by EPA to be highly relevant to the assessment of genotoxicity for humans. However, an
increased level of sister chromatid exchange in peripheral lymphocytes has not been found to be
associated with cancer mortality in a large collaborative evaluation (Bonassi et al.. 2004a).
Although sister chromatid exchange is an indication of genotoxicity, this endpoint is considered to
be less relevant as a predictor of cancer risk.
EPA evaluated the studies, focusing on study design, comparison groups, assessment of
exposure and cytogenetic endpoints, and analytic methods. As discussed in this synthesis, although
the entire set of studies contributed to the assessment, those with the stronger study designs and
methods, and which provided adequate details, were given more weight Most of the studies
reporting on measures of genotoxicity did not describe the details of population selection,
recruitment, and participation, which makes it difficult to evaluate potential selection bias.
However, most did report the population source(s), and since knowledge of a person's status
This document is a draft for review purposes only and does not constitute Agency policy.
A-134 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
regarding these endpoints would not be a factor in his or her decision to participate, the reporting
deficiency is likely not a serious limitation.
Chromosomal Aberrations in Peripheral Blood Lymphocytes
A total of 16 studies were available that evaluated chromosomal aberrations in peripheral
blood lymphocytes (PBLs) or less differentiated subsets among individuals in a variety of exposure
settings, including students in anatomy and embalming courses, workers in industrial settings, and
workers in pathology laboratories (Table A-24). Average formaldehyde concentrations in these
occupational settings generally were above 0.1 mg/m3, although two studies evaluated
chromosomal aberrations among groups exposed to lower average concentrations (Santovito etal..
1239472: Pala etal.. 20081. Study results were heterogeneous, and the studies were variable in
their study designs and reporting detail. Several did not state whether sample analysis was blinded
with respect to exposure status, did not provide demographic information on exposed and referent
groups to support assertions of similarity, had extremely small sample sizes (N <15), or incubated
cells for longer than 48-50 hrs (thus not restricting to Mi metaphases, and/ or did not describe
their approach to data analysis: fGomaa etal.. 2012: Lazutka etal.. 1999: He etal.. 1998: Kitaeva et
al.. 1996: Vasudeva and Anand. 1996: Vargova et al.. 1992: Thomson etal.. 1984: Fleigetal.. 1982:
Suskov and Sazonova. 19821. Nine publications for 8 occupational groups provided detailed
descriptions of study methods and important attributes of the exposed and referent groups (Costa
etal.. 2015: Lan etal.. 2015: Santovito etal.. 2014: Musaketal.. 2013: Santovito etal.. 1239472:
Takab etal.. 2010: Zhang etal.. 2010: Pala etal.. 2008: Bauchinger and Schmid. 19851.
Formaldehyde was associated with a higher prevalence of chromosomal aberrations among
workers in pathology laboratories fCostaetal.. 2015: Musak etal.. 2013: Santovito etal.. 1239472:
Takab etal.. 20101: these effects included chromatid-type aberrations (Costa etal.. 2015: Takab etal..
20101. chromosome-type aberrations fCostaetal.. 2015: Musak etal.. 20131. chromosomal
exchange (Musak etal.. 2013). and premature centromere division (Takab etal.. 2010). Costa et al.
(2015) also reported an increase in aneuploidies and in the number of aberrant and multiaberrant
cells. In one study of paper makers, formaldehyde exposure was associated with dicentrics and
centric rings (Bauchinger and Schmid. 1985). Average 8-hour TWA formaldehyde concentrations
of 0.32, 0.47 and 0.9 mg/m3 were associated with a 1.7 - 1.9-fold increase in total chromosomal
aberrations among exposed groups (Costa etal.. 2015: Musak etal.. 2013: Takab etal.. 2010). An
increased mean number of chromosomal aberrations per cell was significantly associated with an
8-hour TWA concentration of 0.07 mg/m3 among pathologists compared to unexposed hospital
workers exposed to 0.04 mg/m3 by Santovito etal. (2011). One well-conducted study did not
observe associations (Pala etal.. 2008). possibly because the group of laboratory workers was
exposed to very low formaldehyde concentrations (75% of workers at < 0.026 mg/m3). Another
study in nurses found no differences with their referent group, although this group likely
experienced a wide variation in the intensity of their formaldehyde exposure, and no formaldehyde
measurements were conducted f Santovito etal.. 20141. An increased frequency of chromosomal
This document is a draft for review purposes only and does not constitute Agency policy.
A-135 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
aberrations or aberrant cells was also found in a few studies that incubated cell cultures for a
longer period (72 hours) fGomaa etal.. 2012: Lazutka etal.. 1999: Kitaeva et al.. 19961. but not by
all fVasudeva andAnand. 1996: Fleig etal.. 19821. Incubation times longer than required to achieve
first generation metaphase would be expected to result in greater heterogeneity in the aberration
frequencies detected.
Zhang etal. (2010). using fluorescence in situ hybridization techniques, observed an
increased level of chromosome aneuploidy (monosomy 7 and trisomy 8) in cultured CFU-GM
colony cells in a small group of highly exposed formaldehyde-melamine production workers
[n = 10) compared to a referent group matched by age and gender (n = 12). Although only a small
number of workers were evaluated, this report provided complete details on study design,
participation, population characteristics, exposure measurements, cytogenetic analyses, and data
analysis and results. Subsequently, a larger group of the same cohort (n = 29 exposed, n = 23
referent) were included in a chromosome-wide evaluation of aneuploidy, again using cultured CFU-
GM colony cells (Lan etal.. 20151. An elevated risk ratio for monosomy, trisomy, and tetrasomy
was found in several chromosomes, including chromosomes 5 and 7, a finding that was predicted a
priori. In addition, investigators reported an increased frequency of structural chromosome
aberrations in chromosome 5 (IRR 4.15, 95% CI 1.20-14.35). Gentry etal. f20131 reported on
analyses using data on the cohort studied by Zhang etal. f 20101 and noted that few of the DNA
analyses scored 150 or more cells per individual as specified by the study protocol. Although the
pilot study methods were criticized for not adhering to the assay protocol (Gentry etal.. 2013). a
clarification of the assay protocol was provided by the investigators with a description of how the
study adhered to it (Rothman etal.. 2017). The criticism by Gentry etal. (2013) applied to both the
exposed and unexposed groups; thus, no bias should have occurred. Analyzing fewer cells per
individual may have increased the variability in the prevalence estimates of aneuploidy, which may
have attenuated the measures of association. Although the chromosome anomalies may have
arisen either in vivo or during the in vitro cell culture period (Gentry etal.. 2013). there was a
significant increase in the exposed workers compared to the referent group, indicating a
formaldehyde-associated tendency toward aneuploidy or other chromosomal abberations. Median
formaldehyde concentrations measured in the exposed and referent groups were 1.7 mg/m3 and
0.032 mg/m3, respectively. Personal exposure monitoring was conducted for several other
chemical exposures, including chloroform, methylene chloride, tetrachloroethylene,
trichloroethylene, benzene, or other hydrocarbons, which were not detected. Statistical models
were adjusted for potential confounders including age, gender, recent infection, body mass index,
and current tobacco, alcohol, and medication use.
The differences in lymphocyte subset levels between exposed and unexposed workers
reported by Zhang et al. (2010) were challenged by (Mundtetal.. 2017) in a reanalysis who did not
find evidence of an exposure-response trend within the exposed group, although the difference
between unexposed and exposed subjects was reconfirmed. Rothman et al. f20171 also responded
This document is a draft for review purposes only and does not constitute Agency policy.
A-136 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
to the critique by Mundt explaining that the exposure levels in the exposed group were relatively
homogenous and the study was not designed to provide a range of exposures wide enough to
evaluate exposure-response relationships given the expected effect size and sample size in the
study. Overall, the evidence from the set of studies in which there is higher confidence are
consistent with the finding that formaldehyde exposure is associated with chromosomal
aberrations in peripheral blood lymphocytes.
Micronuclei
An increase in micronuclei in buccal mucosa, nasal mucosal cells and peripheral blood
lymphocytes (PBLs) was associated with formaldehyde exposure in a large number of studies
(Table A-24). Micronuclei were reported in a diverse set of exposed populations including plywood
production workers, formaldehyde production and other chemical workers, pathologists and other
laboratory workers, and anatomy and mortuary lab students, and were observed at average
concentrations of 0.1 mg/m3 (Wang etal.. 2019: Ballarin et al.. 19921. 0.2 mg/m3 (Costa etal.. 2019:
Ladeira etal.. 20111. and 0.5 mg/m3 (Costa etal.. 2013: Costa etal.. 2011: Costa etal.. 2008: Ying et
al.. 19971. Micronuclei in peripheral lymphocytes and exfoliated cells are considered biomarkers of
genotoxic events and chromosomal instability, including errors in DNA repair mechanisms,
dysfunction or lack of telomeres, and other failures during DNA replication and repair processes
fBonassi etal.. 20111. Micronuclei in PBL is a validated predictor of cancer risk in epidemiology
studies (Bonassi et al.. 2007). Studies of exposure to formaldehyde over a short duration found no
changes in micronucleus frequency in nasal mucosal cells (!!! INVALID CITATION !!! 1. buccal
mucosal cells (Speit et al.. 2007a 4-hr exposures for 10 days) or peripheral blood lymphocytes (Lin
et al.. 2013 8-hour cross-shift change!.
Measurements in exfoliated buccal cells (EBC) revealed a consistently increased frequency
of micronuclei or binucleated cells among exposed individuals f Costa etal.. 2019: Aglan and
Mansour. 2018: Peteffi etal.. 2015: Ladeira etal.. 2011: Viegas etal.. 2010: Burgaz etal.. 2002: Ying
etal.. 1997: Titenko-Holland etal.. 1996: Suruda et al.. 1993). Differences were reported using
various study designs, including changes in anatomy and embalming students before and after lab
courses and prevalence surveys comparing exposed workers and referent groups. Generally,
differences were observed at formaldehyde exposure levels averaging 0.2 mg/m3 and above.
Micronuclei frequencies were greater by 1.5 to 6-fold in exposed workers with mean formaldehyde
concentrations of 0.2 to 0.5 mg/m3 compared to referent groups fCosta etal.. 2019: Ladeira etal..
2011: Viegas etal.. 2010). Most of the studies of micronuclei frequency in buccal cells provided
detailed discussions of design, methods, and results; potential confounders and other exposures
that could pose a risk of genotoxicity were considered and excluded either in the design or data
analysis. Associations with exposure duration also were observed by some researchers. Aglan
(2018) analyzed micronuclei frequency in EBC from hair stylists who routinely conducted hair
straightening treatments and compared them to a group of hair stylists who did not conduct these
treatments. Formaldehyde concentrations can be high when hair straightening treatments are used,
This document is a draft for review purposes only and does not constitute Agency policy.
A-137 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
and 15-minute TWA concentrations greater than 1.9 mg/m3 were measured in this group. An
increase in MN frequency was observed between the referent group and exposed groups stratified
by exposure duration (below or above 5 years). However, there is more uncertainty in these results
because reporting deficiencies prevented analysis of the potential for selection bias. While Costa
(2019) reported a nonsignificant increase across tertiles of formaldehyde concentration above 0.2
ppm among anatomy/ pathology workers, the authors did not observe a trend in the frequency of
nuclear buds across exposure duration from less than 8 years to over 14 years. Other studies of
workers with mean exposure duration over 5 years also reported associations with exposure
duration fLadeira et al.. 2011: Viegas etal.. 20101.
Fewer studies are available that assessed micronuclei in nasal cells, but results were
generally consistent Significant differences in nasal micronuclei frequency were observed among
anatomy students after an 8-week course fYingetal.. 19971. pathology workers compared to
unexposed workers at the same institutions (Burgaz etal.. 2001). and between formaldehyde
production workers (Ye etal.. 2005) or plywood production workers (Ballarin et al.. 1992)
compared to their referent groups. Formaldehyde concentrations among exposed groups averaged
0.1->1.0 mg/m3. One study did not observe formaldehyde-related changes in nasal cells of
embalming students fSuruda etal.. 19931. but did report an increase in micronuclei with acentric
fragments (centromere negative micronuclei) using fluorescence in situ hybridization (FISH)
(Titenko-Holland etal.. 1996). These results suggest that the predominant damage in these cells
consisted of DNA and/or chromosomal breaks.
Most of a large set of studies that measured micronuclei in peripheral blood lymphocytes
reported increased levels among exposed participants working in diverse exposure settings and in
several countries fCostaetal.. 2019: Wang etal.. 2019: Aglan and Mansour. 2018: Souza and Devi.
2014: Bouraoui etal.. 2013: Costa et al.. 2013: Costa etal.. 2011: Ladeira etal.. 2011: Tiang etal..
2010: Viegas etal.. 2010: Costa etal.. 2008: Orsiere etal.. 2006: Ye etal.. 2005: He etal.. 1998:
Suruda etal.. 1993). Several of these studies included a large sample size, and all provided detailed
discussions of design, methods, and results, including how potential confounders and other
exposures that could pose a risk of genotoxicity were considered and excluded, either in the design
or data analysis. Costa etal. (2019) reported that the frequency of micronuclei in PBL and EBC
were correlated in their study population. A clear concentration-related response in micronucleus
frequency measured in peripheral blood lymphocytes was reported among plywood production
workers in two studies that evaluated effects across multiple exposure categories fTiangetal.. 2010:
Ye etal.. 2005). Micronuclei frequency (and centromeric micronuclei) increased with cumulative
exposure (Wang etal.. 2019: Suruda etal.. 1993) and the duration of exposure (Aglan and Mansour.
2018: Souza and Devi. 2014: Bouraoui etal.. 2013: Lin etal.. 2013: Ladeira etal.. 2011: Tiang etal..
2010: Viegas etal.. 2010). Observed effects were independent of confounding by age, gender, or
smoking status.
This document is a draft for review purposes only and does not constitute Agency policy.
A-138 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
A study of anatomy students did not observe changes in micronuclei in peripheral blood
lymphocytes after an 8-week course, although increased levels were observed in buccal and nasal
cells, suggesting that changes in lymphocytes may occur after a longer duration of formaldehyde
exposure fYing etal.. 19971. Lin etal. f20131 did not observe an increase in micronucleus
frequency across formaldehyde exposure categories among plywood workers in China. However,
the referent group was exposed to mean concentrations of 0.13 mg/m3, a level associated with
increased micronucleus frequency in another study of plywood workers (Tiang etal.. 20101.
The sensitivity of the micronucleus assay can be enhanced by probing cells with
pancentromeric DNA probes. A micronucleus that has a single centromere (CI + MN) suggests
chromosome migration impairment, and the presence of two or more centromeres (Cx + MN)
indicates centromere amplification, with both conditions indicating aneuploidy (Iarmarcoai et al.,
2006). Orsiere etal. f20061 and Bouraoui etal. f20131 evaluated micronuclei in lymphocytes using
FISH and a pancentromeric probe and found increased levels of centromeric micronulei, including
monocentromeric micronulei (CI + MN) and multicentromeric micronuclei (Cx + MN) among
exposed pathology and anatomy lab workers. The enhanced chromosome loss is consistent with
the increase in aneuploidy in lymphocytes reported by Zhang etal. f201011.
DNA Damage
Most studies of DNA single-strand breaks, DNA crosslinks, apurinic or apyrimidinic sites,
and sites with incomplete DNA repair using the Comet assay observed associations in peripheral
blood leukocytes with occupational formaldehyde exposure involving workers in plywood or
furniture manufacturing use of melamine resin and pathology laboratories (Zendehdel etal.. 2017:
Costa etal.. 2015: Peteffi etal.. 2015: Lin etal.. 2013: Gomaa etal.. 2012: Costa etal.. 2011: Tiang et
al.. 2010: Costa etal.. 20081 (Table-A24). A 1.5 to 3-fold difference was observed comparing
exposed groups to their referent groups at average concentrations as low as 0.09 mg/m3
(Zendehdel etal.. 2017). 0.14 mg/m3 (Tiangetal.. 2010) or 0.04-0.11 mg/m3 (Peteffi etal.. 2015). A
clear concentration-related response was observed in plywood plant workers (Lin etal.. 2013: Tiang
etal.. 2010). In addition to the cross-sectional comparisons, an increased level of damage to DNA,
indicated by increased tail moment levels in the Comet assay, was associated with formaldehyde
exposure over an 8-hour work shift fLin etal.. 20131 and after an exposure for 4 h/day for 5 days
during a controlled human exposure study (Zeller etal.. 2011). One study of workers in medium
density fiberboard manufacture did not observe increases in Comet assay measures in the exposed
group at a mean 8-hr TWA 0.25 ± 0.07 mg/m3 (Avdin etal.. 2013). The range of exposure levels
(0.12-0.41 mg/m3) was lower than most of the studies that evaluated DNA damage using the Comet
assay, and almost half of the exposed workers in this study reported using personal protective
equipment
An increased level of DPXs was associated with formaldehyde exposure in a few studies,
both across an 8-hour work shift fLin etal.. 20131. and in comparisons of formaldehyde-exposed
workers and their referent groups fShaham et al.. 2 0 0 3: Shaham et al.. 19971. Lin etal. f20131 also
This document is a draft for review purposes only and does not constitute Agency policy.
A-139 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
compared DPX rates between formaldehyde-exposed plywood workers and a referent group but
did not observe differences by exposure group. There was no trend across levels of exposure or
duration of employment, possibly because the comparison group had significant exposure to
formaldehyde (0.019-0.252 mg/m3) and workers had been employed only for a mean of 2.5 years.
Shaham et al. f20031 found higher DPX levels in peripheral lymphocytes among a group of
pathologists with a mean duration of exposure of 16 years compared to administrative workers
from the same hospitals. While DPX levels in the exposed group were comparable to the exposed
groups studied by Lin etal. (20131. DPX levels in the administrative workers were 60% less than
those measured in the referent group of woodworkers, perhaps reflecting their lower
formaldehyde exposure. Analyses ruled out potential confounding by age, gender, smoking,
education, and country of origin. Shaham et al. f20031 also observed higher levels of pantropic p53
protein (mutant plus wild-type protein) in serum in the exposed group compared to unexposed,
with a particularly strong association in males (pantropic p53 >150 pg/mL, adjusted OR = 2.0 (95%
CI 0.9-4.4)). Increased serum pantropic p53 levels (p53 >150 pg/mL) was associated with mutant
p53 content, and also with elevated DPX (OR = 2.5, 95% CI 1.2-5.4), suggesting a link between
increases in DPX and overexpression of mutantp53 protein, an indication of loss of tumor
suppressor gene capability.
Malondialdehyde-deoxyguanosine (MidG) adducts in DNA extracted from whole blood were
elevated in pathologists who spent time conducting tissue fixation (mean formaldehyde 0.212 ±
0.047 mg/m3) compared to workers and students in other science labs (Bono etal.. 2010). The
prevalence of MidG DNA adducts was increased in the entire group of pathologists compared to the
referent group among whom average formaldehyde concentrations were 0.028 mg/m3. Increased
levels also were observed among a subgroup exposed to 0.07 mg/m3 formaldehyde and higher.
This finding suggests the presence of formaldehyde-associated DNA damage concurrent with the
induction of oxidative stress. An increase in oxidative stress, indicated by elevated plasma levels of
malondialdehyde (MDA), was observed among employees at a cosmetic manufacturing company,
who also had higher plasma levels of p53 compared to a group of employees in a hospital
administrative department with no formaldehyde exposure (Attia etal.. 2014). Although no air
monitoring was conducted, the cosmetics workers had higher urinary formate levels compared to
the referent group. Both plasma MDA and plasma p53 levels were related to urinary formate levels
and also to each other. Regression analyses were adjusted for age and gender. Together, these two
studies suggest that formaldehyde may increase systemic oxidative stress, which may be related to
observed increases in peripherial white blood cell genotoxicity.
DNA Repair Protein Activity
06-alkylguanine DNA alkyl-transferase activity in peripheral blood lymphocytes of students
after 9 weeks or 3-months exposure to formaldehyde in embalming or anatomy labs was compared
to enzyme activity prior to the beginning of the courses. Although an association with decreased
This document is a draft for review purposes only and does not constitute Agency policy.
A-140 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
activity was indicated in one study of embalming students (Hayes etal.. 19971. this finding was not
confirmed by a subsequent study of anatomy students fSchlink etal.. 19991.
Susceptibility: Gene-Environment Interaction
A few studies of genotoxicity among formaldehyde-exposed groups also evaluated
differences in subgroups defined by polymorphic variants in genes coding for proteins involved in
the detoxification of xenobiotic toxic substances, including glutathione-S-tranferases (GSTM1,
GSTT1, GSTP1), CYP2E1, and specifically formaldehyde (alcohol dehydrogenase (ADH5)) (Table A-
24). Polymorphisms in DNA repair proteins also were studied includingthe X-ray repair cross-
complementing genes (XRCC1, XRCC2, XRCC3), RAD51, PARP1, and MUTYH. This included genes of
Fanconi anemia pathway (FANCA, BRIP1). The frequency of chromosomal aberrations in
lymphocytes was higher in a formaldehyde-exposed group but did not vary by GSTT or GSTM
polymorphism (Santovito etal.. 1239472). However, the GSTM1 null variant and the GSTP1 codon
105 Val allele was associated with an increased olive tail moment and MN frequency, respectively,
among exposed individuals, but not in the referent group (Tiang etal.. 2010). Costa etal. (2015) and
Costa (2019) also reported an increase in MN frequency in exfoliated buccal cells among exposed
individuals with the Val variant in the GSTP1 rsl695 polymorphism, whereas chromosomal
aberrations (CSAs) were more prevalent among the exposed group homozygous for the He allele.
This research group also reported an increase in nuclear buds in buccal cells among exposed
individuals with the A variant in the CYP2E1 rs6413432 polymorphism while exposed individuals
homozygous for the wildtype T allele had a higher % tDNA measured in the comet assay. These
associations were not observed in the referent group. In addition, the variant allele for the ADH5
Val309Ile polymorphism was associated with an increased frequency of micronuclei in
lymphocytes among exposed individuals, but not in the referent group (Ladeira etal.. 2013). The
frequency of nuclear buds was associated with formaldehyde exposure and among carriers of the
XRCC3 Met variant allele in both exposed and referent individuals, but effect modification was not
apparent (Ladeira etal.. 2013). Costa (2019) did not observe associations with the XRCC gene
polymorphisms and micronuclei frequency in EBC or PBL among formaldehyde exposed workers.
However, micronuclei frequency was increased in PBL among exposed individuals with the Ala
variant in the FANCA rs719823 variant. Therefore, genetic differences may alter susceptibility to
the cytogenetic effects of formaldehyde, but more definitive research is needed.
Table A-24. Summary of genotoxicity of formaldehyde in human studies
Reference and study
design
Exposure
Results
Chromosomal Damage and Induction of DNA repair
Prevalence Studies
This document is a draft for review purposes only and does not constitute Agency policy.
A-141 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Costa et al. (2015)
Exposure assessed via
Comparison of exposed (N=84) and referent
Portugal
air sampling and
(N=87), frequencies of chromosome aberrations
Prevalence study
deriving an 8-hr TWA
(CA), structural and numerical
Population: 84 anatomy
for each subject.
Aberration MRa 95% CI
pathology workers from 9
Total CA 1.91 1.44-2.53
hospital laboratories,
Exposure
CSAs 2.07 1.27-3.38
exposed to formaldehyde
concentration:
CTAs 1.86 1.39-2.48
for at least 1 year,
Mean: 0.38 ppm (0.47
Gaps 1.65 1.34-2.03
compared to 87
mg/m3)
Aneuploidies 1.64 1.36-1.98
unexposed employees
Range: 0.28-0.85 ppm
Aberrant cells 1.66 1.28-2.17
from administrative
(0.34-1.05 mg/m3)
Multi-aberrant 3.96 2.09-7.48
offices in same geographic
cells
area. Exclusions: cancer
Exposure duration
a MR - mean ratio; all models adjusted for age, gender and
history, radiation therapy
12.0 ±8.2 years
smoking habit, multi-aberrant cells MR also adjusted for fruit
or chemotherapy, surgery
consumption (# pieces eaten per day)
with anesthesia or blood
transfusion in last year.
No associations observed for models of formaldehyde
Exposed and referent
exposure as continuous variable, exposure duration or
similar for mean age 39
professional activity on genotoxicity endpoints (data not
years, 77% females, 25%
provided by authors)
smokers. Outcome:
Peripheral blood samples,
Mean SCE per cell in peripheral lymphocytes:
coded, analyses blinded to
ratio of exposed to referent
exposure status.
Ratio 95% CI
Chromosome aberrations
SCE/cell 1.27 1.10-1.46
structural and numerical),
Poisson regression adjusted for gender, smoking,
duplicates cultured 51
and age
hours (cited Roma-Torres
et al., 2006), 4% Giemsa
stain; scored 100
metaphases per person,
CTAs & CSAs according to
Savage et al., 1975; gaps
not included.
Exposed compared to
unexposed using Mann-
Whitney U-test for CA
measures; negative
binomial regression for
untransformed total-CAs,
CSAs, CTAs, gaps,
aneuploidies, & aberrant
cells; Poisson regression
for untransformed
multiaberrant cells.
Lan et al. (2015) China
Personal monitors for
Among all 24 chromosomes analyzed, elevated IRR for
Prevalence study
3 days over entire
monosomy found for chromosomes 1, 5, 7, 4,19,10,16, 21,
shift within a 3-week
2, 8,18,12, 20,13, 6, and 14 (p < 0.05, Table 2 in Lan et al.);
period.
elevated IRR for trisomy found for chromosomes 5,19, 21,1,
This document is a draft for review purposes only and does not constitute Agency policy.
A-142 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Population: 43
Formaldehyde
20, and 16; elevated IRR for tetrasomy found for
formaldehyde-melamine
concentration: 8 h
chromosomes 4,15,17,14, 3,18, 8,12,
2,10 and 6.
workers (95% employed
TWA
for >1 yr) compared to 51
Exposed
Selected Comparison of Chromosome Aberration
workers from other
Median: 1.38 ppm (1.7
Rates*
regional factories no
mg/m3)
Chromosome
IRR 95% CI
p-Value
formaldehyde exposure
10th & 90th percentile:
Monosomy
frequency-matched by age
0.78, 2.61 ppm (0.96,
1
2.31 1.61-3.31
6.02E-06
and gender; participation
3.2 mg/m3)
5
2.24 1.57-3.20
9.01E-06
rates exposed 92%,
7
2.17 1.53-3.08
1.57E-05
referent 95%; selected
Referent
4
2.02 1.40-2.90
0.00015
subset with scorable
0.026 ppm (0.032
19
1.74 1.29-2.34
0.00026
metaphases, high
mg/m3)
10
1.86 1.30-2.65
0.00064
formaldehyde levels
10th & 90th percentile:
16
1.54 1.12-2.12
0.0075
among exposed,
0.015, 0.026 ppm
Trisomy
comparable referents with
(0.019, 0.032 mg/m3)
5
3.40 1.94-5.97
1.98E-05
scorable metaphases (29
19
2.07 1.24-3.46
0.0055
exposed and 23 referent).
Formaldehyde LOD:
21
2.09 1.22-3.57
0.0071
Outcome: Chromosome-
0.012 ppm
Tetrasomy
wide aneuploidy in CFU-
4
1.64 1.21-2.21
0.0012
GM colony cells cultured
Personal sampling for
15
3.10 1.53-6.28
0.0017
for 14 days using
organic compounds
17
2.40 1.33-4.32
0.0036
OctoChrome FISH; scored
on 2 or more
* Chromosomes with IRR with p-values < 0.001
minimum 150 cells/
occasions. No
subject; analysis blinded
chloroform,
Increased frequency of structural chromosome aberrations
to exposure. Analyzed
methylene chloride,
in chromosome 5, IRR 4.15, 95% CI 1.20-14.35 (p = 0.024)
using negative binomial
tetra-chloroethylene,
regression controlling for
trichloro-ethylene,
age and gender; incidence
benzene, or
rate ratio (IRR). Also
hydrocarbons were
evaluated potential
detected; urinary
confounding from current
benzene at
smoking and alcohol use,
background levels and
recent infections, current
similar between
medication use, and body
groups
mass index (Supplemental
tables in the paper)
Related reference:
(Zhang et al., 2010)
Santovito et al. (2014)
All exposed used
Frequency of Chromosomal Aberrations and SCEs
Italy
protective equipment;
among nurses and referent (mean ± SE)
Prevalence study
no formaldehyde
# Nurses
Referent
Population: 20 female
measurements; nurses
CA/ NSM
20 0.025 ± 0.003
0.02 ± 0.003
nurses from 2 analogous
also exposed to
Cells with
20 0.025 ± 0.003
0.02 ± 0.003
departments in 2 hospitals
antibiotics, cytostatic
aberrations/
(mean age 37 yr); 20
drugs, anesthetics and
NSM
unexposed from
sterilants
SCEs/ NSM
20 6.55 ±0.033*
4.10 ±0.37
administrative
NSM: number of scored metaphases
*p <0.001
This document is a draft for review purposes only and does not constitute Agency policy.
A-143 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
departments of same
hospital (mean age 39.6
yr); all nonsmokers and
did not consume alcohol
Outcome: Peripheral
blood samples, coded.
Cultures incubated for 48
hr for CA and 72 hrfor
SCE; CA slides stained with
5% Giemsa, scored 200
metaphases per subject,
SCE 50 metaphases scored
per subject; Mean
frequencies compared,
Wilcoxon test
Employment duration:
Exposed 11.8 yr, range
1-28 yr; Referent 11.2
yr, range 7-20 yr
No association CAs or SCEs with age or duration
Costa et al. (2013)
Portugal
Prevalence study
Population: 35 pathology
workers from 4 hospital
laboratories, exposed to
formaldehyde for at least
1 year (88.6% female,
mean age 41.2 yr, 20%
smokers), compared to 35
unexposed employees
from same work area
(80% female, mean age
39.8 yr, 20% smokers).
Outcome: SCE, coding
and analysis blinded; stain
fluorescence plus Giemsa,
scored 50 M2 metaphases/
subject by one reader
Related references:
(Costa et al., 2011;
Costa et al., 2008)
Exposure assessed via
air sampling and
deriving an 8-hr TWA
for each subject.
Exposure
concentration:
Mean: 0.44 mg/m3
Range: (0.28-0.85)
mg/m3
Exposure duration
12.5 (1-30) yrs
Mean SCE per cell 1.3-fold higher in exposed workers
compared to controls (p<0.05, Student's t-test).
Univariate analyses presented in Figure 1 of Costa et al.
Mean SCE per cell in peripheral lymphocytes:
ratio of exposed to referent
Ratio 95% CI
SCE/cell 1.245 0.594 -1.897
Multivariate analysis adjusted for gender,
smoking, and age
Musak et al. (2013)
Slovakia
Prevalence study
Population: 105
technicians and
pathologists at hospital
labs (79% female, mean
age 41.7 yrs, 27.6%
smokers) compared to 250
other medical staff (89%
Air monitoring once
per year (no details
provided).
Exposure conc.:
Mean: 0.32 mg/m3
Range: 0.14-0.66
mg/m3
Exposure duration:
Mean: 14.7 ± 10.4 yrs
Range: NR
Chromosome aberrations in peripheral
lymphocytes
Aberration OR 95% CI
CA 1.70 1.6-2.72
CTA 1.37 0.85-2.19
CSA 1.57 0.98-2.53
Chromosomal 2.6 1.1-5.9
exchange
Binary logistic regression controlling for age, gender,
job type, and smoking
This document is a draft for review purposes only and does not constitute Agency policy.
A-144 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
female, mean age 36.2 yrs,
19.2% smokers), all
healthy.
Outcome: Differences in
frequency of
chromosomal aberration
in peripheral blood
lymphocytes, blinded
analysis, 100 mitoses
scored/ subject, 2 scorers
Gomaa et al. (2012)
Egypt
Prevalence study
Population: 30 workers in
pathology, histology and
anatomy laboratories at a
university (30% female,
mean age 42.5 yr)
compared to 15 referents
(46.7% female, mean age
39.3 yr). Source of
referent was not
described.
Outcome: Chromosome
aberrations in peripheral
blood lymphocytes,
cultured 72 hr, blinding
not described; mean # per
100 metaphases;
Difference between
exposed and referent,
Student's t-test
No formaldehyde
measurements;
exposure defined by
job type
Mean employment
duration 14.3 yr
Chromosomal aberrations in peripheral lymphocytes
Structural
Referent
Exposed
Chromatid gap
1.9 ±0.36
6.5 ±0.65*
& break
Chromatid
8.7 ±0.55
15.5 ±0.47*
deletion
Ring
5.5 ±0.33
16.4 ±0.29*
chromosome
Dicentric
0.9 ±0.41
9.0 ±0.54*
chromosome
Total
20.0 ±0.27
46.4 ±0.35
Numerical
Aneuploidy
0.2 ±0.12
0.7 ±0.10
Polyploidy
0.6 ±0.14
0.9 ±0.09
* Student's t-test, p <0.05; mean per 100 metaphases
±SE
No association with age or gender, ANOVA
Santovito et al. (2011)
Italy
Prevalence study
Population: 20 pathology
workers (70% female,
mean age 45.7 yr)
compared to 16 workers
from the same hospital
(43.8% female, mean age
42.1 yr). All subjects were
non-smokers and had not
consumed alcohol in 1
year.
Outcome: Frequency of
chromosome aberrations
per cell and mean % cells
Exposure cone:
Personal air sampling,
8-hour duration.
Referent: Mean: 0.036
± 0.002 mg/m3
Pathologists: Mean:
0.073 ± 0.013 mg/m3
LOD 0.05 mg/mL
Exposure duration:
Mean: 13 yrs
Range: 2-27 yrs
Chromosomal aberrations in peripheral
lymphocytes
Referent
Exposed
Mean CA/cell
0.011 ± 0.004
0.03 ± 0.004*
% of cells with
1.00 ± 0.342
2.50 ±0.286
aberrations
*p <0.001, Mann-Whitney U test
Effects of exposure on chromosomal aberrations
and cells with aberrations (coefficient (SE))
Exposure
p- Value
# CA
0.960 (0.275)
0.001
# cell with
0.838 (0.287)
0.004
aberrations
This document is a draft for review purposes only and does not constitute Agency policy.
A-145 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
with aberrations; Venous
Generalized linear models with Poisson error
blood sample collected at
distribution, adjusted for age
end of shift on same day
as formaldehyde
measurements, samples
coded and processed
within 4 hours of
collection, cells harvested
48 hr, 5% Giemsa stain,
scored 100
meta phases/subject
Jakab et al. (2010)
Exposure assessed via
Cytogenetic analysis in cultured peripheral
Hungary
records on area air
lymphocytes (mean ± SD)
Prevalence study
samples, measured
Unexposed Exposed
Population: 37 female
within 1-3 years of
Total CA 1.62 ±0.26 3.05 ± 0.62*
workers in 3 hospitals & 1
data collection.
Chromatid-type 1.00 ±0.20 2.35 ±0.46*
university pathology
aberrations
department (21 exposed
Exposure
Chromosome- 0.62 ±0.18 0.70 ±0.26
to formaldehyde alone
Concentration:
type
(mean age 43.3 yr, 23.8%
8-hr TWA: 0.9 mg/m3
aberrations
smokers), compared to 37
Range: 0.23-1.21
Aneuploidy 8.89 ±0.66 5.4 ±0.61*
healthy female unexposed
mg/m3
SCE (%/cell) 6.16 ±0.16 6.36 ± 0.26
health-service staff (mean
Exposure duration:
High frequency 3.76 ± 1.14 7.05 ±2.19
age 41.8 yr, 16.2%
Mean: 17.7 yrs
SCE
smokers).
Range: 4-34 yrs
PCD (%) 7.6 ± 0.84 13.65 ± 1.59*
Outcome: Peripheral
PCD (CSG) 5.57 ±0.66 8.8 ±1.07*
lymphocytes; CA, SCE,
*p <0.05, Student's t-test, compared to controls
premature centromere
SCE % and mean HF/SCE higher in referent and exposed
division (PCD), mitoses
smokers; mean SCE % associated with older age
with >3 chromosomes
with PCD (centromere
separation general (CSG)),
CA stain 5% Giemsa, cells
harvested 50 hr, scored
100 metaphases/ subject.
SCE fluorescence plus
Giemsa; scored 50 cells/
subject; analyses blinded
Zhang et al. (2010)
Personal monitors for
Leukemia-specific chromosome changes:
China
3 d within a 3-wk
Prevalence study
period.
Significant increase chromosome aneuploidy in cultured CFU-
Population: 43
Formaldehyde
GM colony cells among subset of high exposed (n =10)
formaldehyde-melamine
concentration: 8 h
compared to matched controls (n = 12)
workers (95% employed
TWA
Data provided in Figure 4 of Zhang et al. 2010.
for >1 yr) compared to 51
Exposed
workers from other
Median: 1.57 mg/m3
regional factories
This document is a draft for review purposes only and does not constitute Agency policy.
A-146 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
frequency-matched by age
and gender; participation
rates exposed 92%,
referent 95%; Analyzed
subset of exposed (n=10, 9
male, 1 female, mean age
31 yr) and referent (n =12,
11 male, 1 female, mean
age 32 yr)
Outcome: Chromosome
aberration in peripheral
blood cells, blinded to
exposure. Chromosome
aneuploidy in cultured
CFU-GM colony cells using
FISH; monosomy 7 and
Trisomy 8; scored
minimum 150 cells/
subject.
Related reference:
Mundt et al. (2017);
Lan et al. (2015);
Gentry et al. (2013)
10th & 90th percentile:
0.74, 3.08 mg/m3
Referent
0.039 mg/m3
10th & 90th percentile:
0.022, 0.039
Analyzed using negative binomial regression (exposed
compared to unexposed) controlling for age, gender, and
smoking
Mundt et al. presented individual data in graphs for
chromosome 7 and chromosome 8 (n = 10 exposed and n =
12 controls), noting smoking status and whether 150 or more
cells were evaluated. No patterns apparent.
Costa et al. (2008)
Exposure assessed via
air sampling at
breathing zone and
deriving an 8-hr TWA
for each subject
Concentration:
Mean: 0.54 mg/m3
Range: (0.05-1.94)
mg/m3
Duration: 11 yrs
Range: (0.5-27) yrs
Mean SCE per cell in peripheral lymphocytes
Portugal
Prevalence study
Population: 30 pathology
lab workers (4 hospitals),
(70% female, mean age 38
yr, 27% smokers)
compared to 30
administrative employees
matched by age, gender,
lifestyle, smoking habits
and work area (63.3%
female, mean age 37 yrs,
23% smokers).
Outcome: Peripheral
lymphocytes; blood
samples collected 10-11
am; processed
immediately; stain
fluorescence plus 5%
Giemsa, SCE/ cell 50 s
division metaphases
scored by one observer,
Scored blind to exposure
Controls Exposed
SCE/cell 4.49 ±0.16 6.13 ±0.29*
*p <0.05, Student's t-test
No association of SCE with gender or age. Smoking increased
SCE among referent group (smoking prevalence 23% in
referent, 27% in exposed.
No association of SCE with duration of exposure
This document is a draft for review purposes only and does not constitute Agency policy.
A-147 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
status. Effect of smoking
and gender also analyzed
Pala et al. (2008) Italy
Prevalence study
Population: 36 lab
workers (66.7% female,
mean age 40.1 yr, 16.7%
smokers)
Outcome: CA and SCE, in
peripheral lymphocytes
(blood sampled at end of
8-hour) Blinded analyses,
CA: cells harvested at 48
hr, 100 metaphases/
subject, SCE: harvest at 72
hr, 30 2nd division cells/
subject.
Personal air
monitoring (8-hour
sample)
High exposure group:
> 0.026 mg/m3, 75th
percentile (range
0.005-0.269 mg/m3)
and low-exposure
group: <0.026 mg/m3
Concentration:
Low (n = 27): 0.015
(0.005-0.0254) mg/m3
High (a? = 9): 0.056
(0.026-0.269) mg/m3
Frequency chromosome aberrations in
peripheral lymphocytes
CA SCE
< 0.026
mg/m3
> 0.026
mg/m3
Means ratio
(95% CI)
2.95 ± 1.79
("=19)
2.22 ± 1.27
[n=5)
0.83
(0.42-1.64)
6.57 ± 1.38
("=17)
5.06 ±0.76
(n=2)
0.81
(0.56-1.18)
Multivariate regression models adjusting for gender, age and
smoking; Poisson model for CA, SCE log-normal random
effects model
Authors did not use a referent group
Ye et al. (2005) China
Population: 18 workers at
a formaldehyde plant at
least 1 year (38.9%
female, mean age 29 yr,,
and 16 workers exposed
to indoor air
formaldehyde via building
materials (75% female,
mean age 22 yr) compared
to 23 students with no
known source of
formaldehyde exposure
(dormitories) (48% female,
mean age 19 yr); all
nonsmokers
Outcome: SCE in
peripheral lymphocytes,
time of sample not stated;
stain Giemsa solution,
analysis blinded, 30 M2
lymphocytes analyzed/
subject.
Area samples;
Exposure duration:
Workers 8.5 (1-15) yrs
Waiters 12 weeks
TWA Concentration
Controls
0.011 ± 0.0025 mg/m3
Max. 0.015 mg/m3
Wait staff
0.107 ±0.067 mg/m3
Max. 0.30 mg/m3
Workers
0.985 ± 0.286 mg/m3
Max. 1.694 mg/m3
SCE frequency by exposure group
Referent Wait Formaldehyde
Staff workers
Mean SCE 6.38 ±
0.41
6.25
8.24 ± 0.89*
*p <0.05, ANOVA. Values estimated from graph in Figure 2
of Ye et al.
(Shaham et al.. 2002)
Israel
Prevalence study
Population: 90 workers
from 14 hospital
pathology departments
Personal and area
samples, sampling at
different points in
work day, sampling
duration averaged 15
min
SCE frequency in peripheral lymphocytes by
exposure group and smoking status (mean ± SE)
Mean number
SCEs per
chromosome
Mean
proportion of
high frequency
cells
This document is a draft for review purposes only and does not constitute Agency policy.
A-148 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
(65 females, 25 males;
Exposure
Unexposed 0.19 ±0.004 0.44 ±0.02
mean age 44.2 yr, 34%
concentration:
Exposed 0.27 ±0.003* 0.88 ±0.01*
smokers) compared to 52
Low level exposure:
No smoking
administrative workers
Mean: 0.49 mg/m3
Low 0.28 ± 0.004 0.88 ± 0.015
from the same hospitals (8
Range: 0.05-0.86
High 0.26 ±0.021 0.86 ± 0.016
females, 44 males; mean
mg/m3
Smoking
age 41.7 yr, 46.9% active
Low 0.27 ± 0.007 0.89 ± 0.018
smokers, 53.1%
High level exposure:
High 0.28 ±0.006 0.92 ±0.021
nonsmokers)
Mean: 2.76 mg/m3
*p <0.01, ANOVA adjusting for age, gender, smoking
Outcome: SCE in
Range: 0.89-6.89
status, education years and origin (ethnicity)
peripheral lymphocytes;
mg/m3
Mean # SCEs per
No association with exposure duration (<15 years and >15
chromosome and
Exposure duration:
years) with adjustment for age, gender, smoking status,
proportion of high
Mean: 15.4 yrs
education years and origin (ethnicity)
frequency cells compared
Range: 1-39 yrs
between exposed and
referent. High frequency
cells defined as > 8 SCEs;
blinding not described,
stain fluorescence plus 5%
Giemsa, scored 30-32
cells/ subject.
Related references:
Shaham et al. (1997)
Lazutka et al. (1999)
Industrial hygiene
Frequency of chromosomal aberrations in peripheral
Lithuania
area measurements
blood lymphocytes by exposure (CA/ 100 cells ±
Prevalence study
reported by plants;
SEM)
Population: Carpet and
carpet plant,
# CA Frequency
plastic manufacturing;
formaldehyde 0.3-1.2
Carpet Workers
Carpet plant, exposed, 38
mg/m3, styrene
Exposed 79 3.79 ±0.32*
male, 41 female (age
0.13-1.4 mg/m3,
Referent 90 1.68 ±0.13
22-65 yr, 49% smokers);
phenol 0.3 mg/m3;
Plasticware
unexposed, 64 male, 26
plasticware plant,
workers
female, 30% smokers;
formaldehyde 0.5-0.9
Exposed 97 4.17 ±0.29*
Plastic plant, exposed 34
mg/m3, styrene
Referent 90 1.68 ±0.13
male, 63 female (age 28-
4.4-6.2 mg/m3,
*p < 0.0001; ANOVA adjusted for age
64 yr, 37% smokers);
phenol 0.5-0.75
Predominant types of damage were chromatid and
unexposed 64 males, 26
mg/m3
chromosome breaks
females
Outcome: CA in peripheral
Duration exposure,
Duration of exposure not associated with CA frequency; Age
blood lymphocytes;
carpet plant: 2 mo-21
and smoking (data not shown) were not associated with CA
fluorescence plus Giemsa
yr; plastic plant: 2
frequency
stain, cells harvested 72
mo-25 yr
hr, scored 100
metaphases/ subject on
coded slides.
Shaham et al. (1997)
Field and personal air
SCE (mean # per chromosome) in peripheral
Israel
sampling, sample
lymphocytes
This document is a draft for review purposes only and does not constitute Agency policy.
A-149 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Prevalence study
Population: 13 pathology
workers (mean age 42 yr,
23% smokers) compared
to 20 referent workers
matched by age (mean
age 39 yr, 30% smokers).
Outcome: SCE in
peripheral lymphocytes,
Mean # per chromosome,
stain fluorescence plus 5%
Giemsa, blinding not
described, mean of 30
cells/ individual,
Related references
(Shaham et al., 1996)
duration 15 minutes,
multiple times during
work-day (# not
reported).
Concentration:
Mean: not reported
Range: 1.7-1.97
mg/m3
Personal samples:
Range: 3.4-3.8 mg/m3
Exposure duration
mean 13 years (range
2-25 years)
Unexposed Exposed
SCE 0.186 ±0.035 0.22 ±0.039*
*p = 0.05, ANOVA adjusted for smoking status
years of exposure linearly correlated with mean number of
SCE per chromosome, adjusting for smoking
Kitaeva et al. (1996)
Russia (translated)
Prevalence study
Population: 15
formaldehyde production
workers (5 females, 10
males, mean age 38 yr),
anatomy instructors (6
female, 2 male), mean age
41 yr) compared to 6
unexposed (mean age
28.5 yr)
Outcome: Blood collection
in 1988. CA: cells
harvested at 72 hr;
blinding not described.
Unclear if statistical
analyses were performed.
No quantitative
exposure assessment
Exposure duration:
Formaldehyde
production 9.7 years
Anatomy instructors
17 years
CA (% aberrant metaphases) in peripheral
lymphocytes
Referent (n=6) Exposed
Workers (n=8)
% of 1.8 ± 0.6 (547 5.4 ± 1.9 (148
metaphases at metaphases metaphases
72 hours examined) examined)
lymphocyte
culture
No metaphases observed at 72 hours in lymphocyte cultures
from anatomy instructors
Authors reported that % CA was not dependent on age,
gender and length of employment
Vasudeva and Anand
(1996) India
Prevalence study
Population: 30 female
medical students exposed
15 months, compared to
30 age-matched
nonmedical students. All
17-19 years old
Outcome: chromosomal
aberrations in peripheral
blood samples, mean
Exposure not
quantified
Exposure conc.: < 1.23
mg/m3
Exposure duration:
15 months
No significant difference in chromosomal aberrations
between groups (p>0.5).
Mean frequency of aberrant metaphases
Exposed: 1.2%
Unexposed: 0.9%
No additional quantitative information available
This document is a draft for review purposes only and does not constitute Agency policy.
A-150 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
frequency aberrant
metaphases, cells
harvested at 72 hr, 100
cells/ subject; blinding not
reported.
Vargova et al. (1992)
Czechoslovakia
Prevalence study
Population: 20 wood
workers with at least 5
years of exposure (10
females, 10 males, mean
age 42.3 yr), compared to
19 workers from the same
plant with no known
occupational contact with
chemicals.
Outcome: CA frequency,
peripheral lymphocytes,
Giemsa staining, cells
harvested 48 hr, 100 cells/
subject. Blinding not
described.
Task-based air
sampling in breathing
zone over 8 hours
Exposure conc.:
Range: 0.55-10.36
mg/m3
Exposure duration:
5->16 yrs
Frequency of chromosomal aberrations in
peripheral lymphocytes by exposure group
Exposed
Unexposed3
% aberrant
cells
# breaks per
cell3
3.08
0.045
3.60
0.030
3 According to authors, both groups reported %
aberrant cell levels above normal range (1.2-2%)
Bauchinger and
Schmid (1985)
Germany
Prevalence study
Population: 20 male paper
makers exposed for at
least 2 years (mean age
40.8 yr, 30% smokers)
compared to 20
unexposed male workers
from the same factory
Outcome: Peripheral
lymphocytes, CA/ cell
(scored 500 cells/ subject),
cells harvested 48 hr,
Giemsa staining; SCE/ cell
(scored 50/ subject)
analyzed using coded
slides, SCE stratified by
smoking status.
Exposure assessment
based on air
monitoring and job-
function.
Exposure
concentration.: =1.47
mg/m3, plus 3.7
mg/m3 for 45 minutes
(supervisors) or 90
minutes (operators)
per 8 hours
Exposure duration
Mean: 14.5 yrs
Range: 2-30 yrs
Frequency of CA and SCE/cell (mean ± SE) in
peripheral lymphocytes
Referent
Exposed
% cell with CA
SCE/ cell
Aberrations/ cell
Chromatid
Acentric
fragments
Dicentrics
Centric rings
0.86 ±0.10
9.53 ± .0.35
0.0038 ±
0.0005
0.0046 ±
0.0006
0.0005 ±
0.0002
0.0001 ±
0.0001
0.87 ± 0.08
8.87 ±0.24
0.0042 ± 0.0005
0.0034 ± 0.0005
0.0013 ±
0.0003*
0.0003 ±
0.0001*
*p <0.05, Mann-Whitney rank U test
Frequency of SCE was not associated with exposure when
stratified by smoking
Thomson et al. (1984)
Great Britain
Prevalence study
Personal air
monitoring over 1-3
months before blood
samples
No significant difference in incidence of chromosome
aberrations or SCE frequency found between groups.
SCE frequency (mean per cell)
This document is a draft for review purposes only and does not constitute Agency policy.
A-151 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Population: 6 pathology
workers (2 female, 4 male,
mean age 33.5 yr)
compared to 5 referents
(3 female, 2 male, mean
age 27.8 yr) (study details
on referent not provided)
Outcome: CA frequency,
stain fluorescence plus
Giemsa technique (Perry
and Wolff. 1974), cells
harvested 48 hr, slides
coded and scored 100 1st
division metaphases/
subject; SCE frequency,
cells harvested 72 hr, 50
cells/ subject
Exposure conc.: TWA
Mean: 2.26 mg/m3
Range: 1.14-6.93
mg/m3
Exposure duration: 4-
11 years, 2-4 hr/day,
2-3 days/week
Exposed (N=6) 6.78 ±0.31
Referent (N=5) 6.44 ± 0.38
(individual data reported, analytic methods were not
described)
Fleig etal. (1982)
Germany
Prevalence study
Population: 15
formaldehyde-
manufacturing workers
(mean age 50 yr)
compared to 15 age-and
gender matched
unexposed workers from
same plant.
Outcome: Chromosome
aberrations in peripheral
blood lymphocytes cells
harvested 70-72 hours,
10% Giemsa stain; slides
coded; scored 100
metaphases/ subject.
Personal air sampling.
1946-1971: <6.15
mg/m3 (MAK)
1971-1982: <1.23
mg/m3 (MAK)
Duration:
Mean: 28yrs
Range: 23-35 yrs
Chromosomal aberrations in peripheral blood
lymphocytes
Unexposed Exposed
Mean % aberrant 3.33 3.07
cells including gaps
Mean % aberrant 1.07 1.67
cells excluding gaps
P >0.05, Fisher's exact text
Smoking habit not associated with CA (data not reported)
Suskov and Sazonova
(1982) Russia
Prevalence study
Population: 31 phenol-
formaldehyde workers
(mean age 39.1 yr)
compared to 74 referents
matched by gender,
smoking, alcohol
Workers exposed to
both phenol and FA.
Area samples
Exposure conc.:
Formaldehyde Mean:
0.5 mg/m3
Phenol mean: 0.3
mg/m3
Exposure duration:
Frequency of chromosomal aberrations by
exposure group
Mean % aberrant Referent Exposed
cells
Aberrant 2.4 ±0.22 5.0 ±0.40*
metaphases
Aberrant 0.024 ± 0.058 ±
chromosomes per 0.002 0.006*
cell
This document is a draft for review purposes only and does not constitute Agency policy.
A-152 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
1.26 ±0.076 1.27 ±0.044
consumption, and
medication
Outcome: Chromosomal
aberrations via mean
frequency of aberrant
metaphases, Buckton and
Evans method (1973);
cells harvested at 50 hr
4 months to 30 yrs
Chromosomal
breaks per aberrant
chromosome
*p <0.001, chi-square
Short-term Studies
(Ying et al.. 1999) China
Population: 23
nonsmoking anatomy
students (11 males, 12
females, age not reported)
exposed during 8-week
course, 3-hr session, 3
times/ wk.
Outcome: SCE in
peripheral blood
lymphocytes, assessed
before the start of the
course and at the end of
8-week period. Blinded
analysis of slides, one
observer with repeat by
second; 30 M2
lymphocytes per subject
analyzed; Lymphocyte
transformation rate (LTR)
Air sampling,
estimated TWA and
peak levels during
class and in the
dorms.
Anatomy labs:
Mean 3-hr TWA: 0.51
± 0.299 mg/m3, range:
0.07-1.28 mg/m3
Dormitories:
Mean TWA: 0.012 ±
0.003 mg/m3, range:
0.011-0.016 mg/m3
Duration: 8 wks
Frequency SCE and lymphocyte transformation rate
(%) (Mean+SEM), Change over 8 weeks
SCE
LTR
Before
exposure
6.383 ± 0.405
59.07 ±6.35
After exposure
6.613 ±0.786
56.92 ±8.64
*p <0.05, paired t-test
Levels in males and females were similar
He et al. (1998) China
Prevalence study
Population: 13 anatomy
students exposed during a
12-week course compared
to 10 students. Age and
gender similar between
groups, all nonsmokers
(data not shown).
Outcome: CA and SCE in
peripheral lymphocytes,
CA: modified fluorescence
plus Giemsa stain, cells
harvested 48 hr, scored
100 metaphases/ subject.
SCE: cells harvested 72 hr,
Breathing zone air
samples in location of
exposed students.
Concentration in
breathing zone: Mean
2.92 mg/m3
Duration:
12 weeks (10
hrs/week)
Frequency of SCE and chromosomal aberrations in
peripheral lymphocytes
Referent
Exposed
Mean SCE per
5.26 ±0.51
5.91 ±0.71*
cell
Lymphocyte CA
3.40 ± 1.57
5.92 ±2.40*
*p <0.05, analytic test not described
This document is a draft for review purposes only and does not constitute Agency policy.
A-153 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
50 metaphases/ subject.
Blinding not described
Suruda et al. (1993)
USA
Panel study
Population: 29 students
(with adequate samples)
(24.1% female, mean age
23.6 yr, 17.2% smokers)
exposed to formaldehyde
for 9 weeks during
embalming course, with
baseline samples taken.
Mean duration of
embalming 125 min.
Possible exposure prior to
course.
Outcome: SCE in
peripheral lymphocytes,
stain fluorescein plus
Giemsa, 50 s division
metaphases scored/
subject; blood samples
collected in morning
before 1st class and after 9
weeks; analysis of slides
blinded to exposure status
Personal sampling for
121 of 144
embalmings; Exposure
concentration: Mean:
1.72 mg/m3
Range: (0.18-5.29)
mg/m3
Duration:
9 wks (0.173 yrs)
Frequency of SCE before and after a 9-week
embalming course
SCE
Before
exposure
7.72 ± 1.26
After exposure
7.14 ± 0.89*
*p <0.01, difference in mean before and after
exposure, matched Student's t-test
(Yager et al.. 1986)
USA
Panel study
Population: 8 anatomy
students (1 male, 7
females, mean age 26 yr,
all nonsmokers) exposed
to formaldehyde during a
10 week course (2
sessions/week). No
occupational or lab
formaldehyde exposure
during previous year.
Outcome: Mean SCEs per
cell in peripheral
lymphocytes; before and
after 10 weeks, samples
coded and randomized
together for analysis
Ambient air and
breathing zone
monitoring.
Breathing zone
concentration:
Mean:1.5 mg/m3
Range: 0.9-2.4 mg/m3
Exposure duration:
10 weeks
Mean SCE per cell before and after 10-week course
(mean ± SEM)
Before
Mean SCE per 6.39 ±0.11
cell
After
7.20 ±0.33*
*p = 0.02, paired t-test
This document is a draft for review purposes only and does not constitute Agency policy.
A-154 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Zeller et al. (2011)
Germany
Controlled human
exposure study
Subjects: 41 healthy
volunteers exposed 4 hr/
day for 5 days, all male,
nonsmokers
Outcome: SCE in
peripheral lymphocytes:
method according to
Schmid and Speit (2007),
scored 30 cells/ sample.
Proliferation index (PI)
calculated from 1st, 2nd,
and 3rd mitoses in 100
metaphases. Analyzed
using Wilcoxon Sign Rank
test
12 groups of 2 to 4
persons in a chamber,
exposures randomly
assigned.
Formaldehyde
concentrations: 0 (i.e.,
background level of
0.01 ppm), 0.3 ppm
(0.37 mg/m3)a with
four peaks of 0.6 ppm
(0.74 mg/m3), 0.4 ppm
(0.49 mg/m3) with
four peaks of 0.8 ppm
(0.98 mg/m3) and 0.5
ppm (0.67 mg/m3) and
0.7 ppm (0.86 mg/m3),
peaks 15 min each, 4
15-min exercise
sessions during
exposure.
Frequency of SCE/ metaphaseand PI in
lymphocytes before and after 4-hour exposure (N ¦¦
40)
Lymphocytes
Before
After
SCE/
metaphase
6.1 ± 0.898a
6.1 + 0.938
PI
2.46+0.114
2.47+0.145
ap = 0.689
Chromosomal Breaks or Aneuploidy
Prevalence Studies
Aglan (2018) Egypt
Prevalence study, June
2015 - September 2016
Population: 60 hair stylists
who routinely conducted
hair straightening
compared to 60 stylists
who did not conduct this
treatment. Excluded
subjects with chronic
disease and /or regular
medications, family
history of cancer,
recurrent abortions,
smoking or pregnancy.
Ages 20-36 years.
Outcome: Blood collected
at end of 8-hour shift.
CB Micronucleus test in
lymphocytes. Replicate
cultures for each sample,
incubated 72 hours. 2,000
binucleasted cells from
coded slides (1,000 from
each replicate culture),
scored using criteria by
Passive air sampling
(Umex-100) at fixed
position in breathing
zone, 15-minute
samples during hair
straightening process;
15-minute TWA
Group 1 (work
duration < 5 years):
1.68 ± 0.27 ppm
Group 2 (work
duration > 5 years):
1.83 ± 0.16 ppm
MN frequency (%) in PBL and buccal cells by
duration of employment (< 5 and > 5 years)
PBL
EBC
MeaniSD MeaniSD
Referent (n=60) 0.22 ±0.42* 0.17+0.38*
<5 years 0.61 ±0.50 0.32+0.48
(n=31)
>5 years 1.66+0.48 0.94 + 0.58
(n=29)
p < 0.01, p < 0.001, Kruskal Wallis test
Between group differences statistically significant in PBL and
for EBC except between referent and < 5 year exposure
group (least significant difference test)
This document is a draft for review purposes only and does not constitute Agency policy.
A-155 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Fenech (2003). MN
frequency % altered cells.
MN in exfoliated buccal
cells. Cheeks scraped with
wooden spatula, fixed in
3:1 methanol/ acetic acid
and dropped onto slides,
stained with Feulgen/ Fast
Green, examined at 400x
according to Tolbert et al.,
1991. Analyzed
independently by 2
people, 1,500 cells scored
per person using criteria
by (Sarto et al.. 1987).
% altered cells.
Costa et al. (2019)
Portugal
Prevalence study
(extension of Costa et
al.. 2015) adding
outcomes)
Population: 85 anatomy
pathology workers from 9
hospital laboratories,
exposed to formaldehyde
for at least 1 year,
compared to 87
unexposed employees
from administrative
offices in same geographic
area. Exclusions: cancer
history, radiation therapy
or chemotherapy, surgery
with anesthesia or blood
transfusion in last year.
Exposed and referent
similar for mean age 39
years, 77% females, 25%
smokers. Outcome:
Peripheral blood samples,
coded, analyses blinded to
exposure status.
Exfoliated cells were
collected for each cheek
separately. Cytokinesis-
blocked MN test, (Costa
Exposure assessed via
air sampling and
deriving an 8-hr TWA
for each subject.
Exposure
concentration:
Mean: 0.38 ppm (0.47
mg/m3)
Range: 0.28-1.39 ppm
(0.34-1.72 mg/m3)
Exposure duration
12.0 ±8.2 years
MN frequency (%) in peripheral lymphocytes,
exposed relative to referent group, Mean Ratio
(MR)
Ratio
95% CI
Exposure
1.55**
1.2-1.99
Poisson regression models adjusted for age,
gender, smoking habits
**p< 0.01
MN frequency in exfoliated buccal cells, Mean
Ratio (MR)
Exposed:
MR
95% CI
Unexposed
MNB
63:69
4.08***
2.12-7.87
BNbud
63:69
2.88***
1.76-4.71
Poisson regression models adjusted for age,
gender, smoking habits; ***p < 0.001
Correlation between MNL and MNB: r = 0.359, p < 0.001
MN frequency in PBLand exfoliated buccal cells
by level and duration in exposed, Mean Ratio
(MR)
MNL
N MR
95% CI
BNbud
N MR
95% CI
Level
(ppm)
0.08-0.22 27 1.0 20 1.0
0.23-0.34 29 1.5** 1.12-2.00 16 1.42 0.64-3.14
0.35-1.39 28 1.37 1.04-1.81 17 1.96 0.91-4.24
Duration
years
This document is a draft for review purposes only and does not constitute Agency policy.
A-156 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
et al., 2008); culture
incubation 72 hr; stain 4%
Giemsa; scored 1,000
binucleated cells/subject,
criteria defined by Fenech
et al. (2007).
Buccal MN cytome assay.
2,000 differentiated cells
scored for frequency of
MN, nuclear buds and
nucleoplasm^ bridges
according to Tolbert et al.
1992 and Thomas et al.
2009.
T-Cell Receptor mutation
assay in mononuclear
leukocytes, flow
cytometry, minimum of
2.5 x 105 lymphocyte-
gated events were
acquired, # events in
mutation cell window
(CD3-CD4+ cells) divided
by total number of events
for CD4+ cells
<8 28 1.0 25 1.0
8-14 28 0.78 0.51-1.23 18 0.74 0.30-1.78
>14 28 0.68 0.40-1.15 20 1.00 0.37-2.74
Poisson regression models adjusted for age, gender, smoking
habits
*p < 0.05; **p< 0.01.
Wang et al. (2019)
Shanghai, China
Population: 100 male
chemical production
workers exposed to
formaldehyde > 1 year
through 4 work processes
(i.e., production
examination, glue
spraying, coating and
workplace inspection).
Unexposed group (n = 100
males) from the logistics
workshop in same factory.
Exposed and referent
were comparable for
mean age, smoking and
alcohol consumption.
Outcome: CBMN
according to Fenech et al.
(2000,1993). Blinded
analysis. Venous
peripheral blood cultured
Routine formaldehyde
monitoring by factory
Range of geometric
means (mg/m3):
Exposed: 0.06-0.25
Unexposed: 0.01
Cumulative dose
(mg/m3-yr)
determined for each
worker (C x T). C =
geometric mean of
concentration for a
year at a sampling
site, T = years.
Exposed: 0.90 (0.60-
1.78)
Referent: 0.06 (0.02-
0.10)
MN frequency (% per 1000, 95% CI) in PBLs
Exposed Referent
3.05 ±1.47 1.71 ±0.96
Poisson regression models adjusted for age,
gender, smoking habits
Micronucleus frequency (per 1000, frequency ratio
(FR)) in PBL)
CED (mg/m3- N Exposed FR (95% CI)
year)
0.01-0.06 45 1.36 ±0.86 1
0.06-0.125 55 1.87 ±0.92 1.38(1.00-
1.91)
0.125-0.9 46 2.50 ± 1.17 1.83(1.34-
2.52)
0.9-3.75 54 3.65 ± 1.40 2.67(1.99-
3.64)
Poisson regression models with adjustment for age,
smoking status and alcohol use
This document is a draft for review purposes only and does not constitute Agency policy.
A-157 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
for 44 hr, Cytochalasin-B
added to cultures, cells
harvested 28 hours later,
air dried slides stained
with Giemsa, MN
dectected at 400x with
confirmation at l,000x.
1,000 binucleated cells
scored/ subject
Peteffi et al. (2015)
Brazil
Prevalence study
Population: 46 workers in
furniture manaufacturing
facility (mean age 34.5 yr,
56.5% male, 1 smoker)
and unexposed group (n =
45) recruited from
employees and students
of local university with no
history of occupational
exposure to potentially
genotoxic agents or
substances metabolized to
formic acid, (mean age
35.4 yr, 33.3% male, 0
smokers)
Outcome: Oral buccal
epithelial cell samples
(scraped with endocervical
brush), micronucleus test,
DNA-specific Feulgen
staining and
counterstaining with Fast
Green according to
(Tolbert et al., 1992);
analyzed 2,000 cells/
person by 2 independent
observers (1,000 ea).
Monitoring in 7
sections in facility;
referent monitoring in
5 areas of university;
breathing zone 8 hr
samples collected on
same day as biological
samples. Urine
samples collected at
end of work day on 5th
day of work;
correlation of
formaldehyde
concentration in air
with urinary formic
acid concentration, r =
0.626, p<0.001
UV painting,
lamination/press,
packaging, edge
lamination 0.03-0.04
ppm (0.037-0.05
mg/m3)
Edge painting,
machining and drilling
center, board cutting
0.06-0.09 ppm
(0.07-0.11 mg/m3))
Referent mean (SD)
0.012 (0.008) ppm
(0.015 (0.01) mg/m3)
Formic acid median
Exposed 20.47 mg/L
Referent 4.57 mg/L
Exposure duration
5.76 yr
Comparisons of micronucleus frequency and other
DNA damage in buccal cells, median (interquartile
range)
Referent Exposed p-
Value
Micronuclei 0 0 0.08
Nuclear buds 0 0.24 0.126
(0-0.50) (0-0.63)
Binucleated 0.50 1.34 0.003
cells (0-1.38) (0.64-2.38)
Karyorrhexis 1.0 1.31 0.372
(0.49-2.04) (0.58-2.49)
Nonparametric tests used because data were not normally
distributed. Exposed and referent compared using Mann-
Whitney test
No differences between men and women for measures of
DNA damage in either exposed or referent
No correlation between urinary formic acid and measures of
DNA damage
This document is a draft for review purposes only and does not constitute Agency policy.
A-158 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Souza and Devi (2014)
No measurements
India
reported.
MN frequency in Lymphocytes by Exposure Group
Prevalence study
Duration exposure
(mean (SD))
Population: 30 male
Mean+SD 95% CI
workers in anatomy
mean 10.66 yr, range
Exposed (N= 9.5 + 3.23 8.29-10.7
departments (embalming)
1-30 yr
30)
in several medical colleges
Comparison 3.73 + 1.43 3.19-4.26
(mean age 39.9 yr, 50%
group (N = 30)
smokers); compared to 30
Difference in 5.76 4.47-7.063
male clerical workers in
means3
same facilities (mean age
aNo difference = 0
37.8 yr, 30% smokers).
Outcome: Total MN/
Association of MN frequency with exposure and smoking
1,000 cells in peripheral
evaluated using two-way ANOVA. Smoking was not
lymphocytes. Assays
associated with MN frequency.
conducted blinded.
Cytokinesis -blocked
Pearson's correlation test showed a positive correlation (r =
micronucleus assay
0.5, P = 0.02) between the duration of exposure and the
(Costa,2008,
frequency of MN in lymphocytes.
626187Costa et al..
2008)1,1,000 binucleated
cells/ subject.
Bouraoui et al. (2013)
Exposure assessed by
MN frequency in peripheral lymphocytes (Mean ±
Tunisia
job title and duration
SD)
Prevalence study
of employment.
Referent Exposed
Population: 31 pathology
Atmospheric air
MN (%/l,000 7.08 ±4.62 25.35 ± 6.28*
workers (60% female,
sampling performed in
binucleated
mean age 42, 9.6%
area of potential
cells)
smokers) compared to 31
exposure
FISH MN (%/ 6.12 ±4.24 23.25 ± 5.92*
unexposed administrative
Concentration:
2000 cells)
staff in same facility (60%
Means of 3 samplings:
C+MN 4.03 ±3.64 18.38 ± 5.94*
female, mean age 43 yr,
0.25 mg/m3
C-MN 2.09 ±0.74 4.87 ± 3.22
12.9% smokers).
2.21 mg/m3
Cl+MN 2.93 ±2.74 15.35 ± 6.03*
Outcome: MN peripheral
4.2 mg/m3
Cx+MN 1.1 ± 1.16 3.03 ±2.7*
lymphocytes: Cytokinesis-
*p <0.05, Student's t-test
blocked MN assay in
Duration:
combination with FISH
Mean 15.68 yrs (6.53
Duration of exposure was associated with all of the
using all-chromosome
± 0.7 hrs/day)
cytogenetic alterations.
centromeric probe (Sari-
Abbreviations: C +, C -, CI + MN, Cx + MN
Minodier et al., 2002);
stain 5% Giemsa, 2,000
binucleated cells scored/
subiect, (Fenech, 2000),
blinding not described
Costa et al. (2013)
Exposure assessed via
Univariate analyses presented in Figure 1 of the paper. MN
Portugal
air sampling and
frequency was 2.5-folds higher in exposed group compared
Prevalence study
to referent group.
This document is a draft for review purposes only and does not constitute Agency policy.
A-159 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Population: 35 pathology
workers from 4 hospital
laboratories, exposed to
formaldehyde for at least
1 year (88.6% female,
mean age 41.2 yr, 20%
smokers), compared to 35
unexposed employees
from same work area
(80% female, mean age
39.8 yr, 20% smokers).
Outcome: MN in
peripheral lymphocytes,
samples collected
between 10 & 11 am.
Cytokinesis-blocked MN
test (Teixeira et al..
2004). 1,000 cells
analyzed/ subject,
MN per 1,000 binucleated
cells, scored blindly by one
reader, criteria Fenech
(2007)
Related references: Costa
et al. (2011); Costa et
al. (2008)
deriving an 8-hr TWA
for each subject.
Exposure conc.:
Mean 0.44 mg/m3,
range 0.28-0.85
mg/m3
Exposure duration
12.5 ± 8.1 yrs, range
1-30 yr
MN frequency (%) in peripheral lymphocytes,
exposed relative to referent group
Ratio
95% CI
Exposure
2.1
1.025-3.174
Multivariate analysis, adjusted for gender,
smoking and age
Lin et al. (2013) China
Prevalence study
Population: 96 plywood
workers exposed to
formaldehyde (13.5%
female, mean age 33 yr,
30.2% smokers) compared
to referent group (N = 82)
(4% female, mean age 31
yr, 40% smokers).
Outcome: MN assay in
peripheral lymphocytes,
analyzed 1,000
binucleated cells/ subject,
scoring criteria Fenech
(1993), Fenech (2003),
blinded analysis
MN assessed by exposure
group and # years worked.
Personal air
monitoring and job
assignment.
Average
concentration:
High, N = 38 (making
glue): 1.48 mg/m3,
range 0.914-2.044
mg/m3
Low, N = 58 (sanding
boards, pressing wood
scraps with glue at
high temp): 0.68
mg/m3, range
0.455-0.792 mg/m3
Referent group, N=82
(grinding wood
scraps): 0.13 mg/m3,
range 0.019-0.252
mg/m3
MN Frequency in peripheral lymphocytes by
formaldehyde exposure level and work years
By Exposure levels
Referent Low
High
MN freq
2.05 ± 1.72 2.02 ±1.81
2.37 ± 1.79
(%)
ANOVA p-value = 0.455; Poisson regression p-value =
0.288
Number of Work Years
<1 (N= 57) 1-3 (N=
>3 (N= 57)
64)
MN freq
1.02 ± 1.10 2.25 ±
2.90 ±
(%)
1.56*
1.96*
ANOVA p-value < 0.001; Poisson regression p-value <
0.001
ANOVA and Poisson regression adjusting for age, gender,
smoking status, alcohol, duration of employment
This document is a draft for review purposes only and does not constitute Agency policy.
A-160 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Exposure duration:
2.52 yrs
Costa et al. (2011)
Portugal
Prevalence study
Population: 48 pathology
workers from 5 hospital
laboratories, exposed for
at least 1 year (28%
female, mean age 40 yr,
21% smokers), compared
to 50 unexposed
employees matched by
age, gender, lifestyle,
smoking habits and work
area (25% female, mean
age 37 yr, 14% smokers).
Outcome: MN in
peripheral blood
lymphocytes, (Teixeira
et al.. 2004); stain 4%
Giemsa; scored 1,000
binucleated cells/ subject,
scored blind by one
reader, criteria Fenech
(2007)
Exposure assessed via
air sampling in
breathing zone and
deriving an 8-hr TWA
for each subject.
Concentration:
Mean: 0.53 mg/m3,
range 0.05-1.94
mg/m3
Duration:
Mean: 13.6 yrs, range:
1-31 yr
MN frequency (%) in peripheral lymphocytes
MN
Referent
3.66 ±0.51
Exposed
6.19 ±0.62*
*p <0.05; Mann-Whitney U test and Kruskal-Wallis
test
Ladeira et al. (2011)
Portugal
Prevalence study
Population: 56 hospital
workers in histopathology
labs (66% female, mean
age 39.5 yr, 19.6%
smokers) compared to 85
administrative staff (64%
female, mean age 32.4 yr,
29.4% smokers).
Outcome: MN in
peripheral lymphocytes
and buccal cells. Samples
coded and analyzed
blinded. Lymphocytes,
cytokinesis-block
micronucleus cytome
assay, stain May-
Grunwald-Giemsa, 1,000
binucleated cells scored/
Personal air sampling,
6-8 hours, estimated
8-hr TWA
Exposure conc.:
Mean TWA 8h 0.2 ±
0.14 mg/m3
Mean ceiling value:
1.4 ± 0.91 mg/m3,
range 0.22-3.6 mg/m3
Exposure duration:
14.5 (1-33) yrs
MN frequency (Mean ± SD) by cell type
Lymphoctyes Buccal cells
Referent
Exposed
ORa
95% CI
0.81 ±0.172
3.96 ± 0.525*
9.67
3.81-24.52
0.16 ±0.058
0.96 ±0.277*
3.99
1.38-11.58
*p<0.002, Mann-Whitney test
aOdds ratio for risk of presence of MN; binary logistic
regression
MN frequency (Mean ± SD) by years of exposure
Years
N
Lymphocytes
Buccal cells
<5
8
2.75 ±0.940
0.63 ±0.625
6-10
19
3.05 ±0.775
0.63 ±0.326
11-20
12
5.50 ± 1.317
0.83 ±0.458
>21
15
5.00 ± 1.151
1.20 ±0.8
Evaluated potential confounding by age, gender,
smoking and alcohol, no major evidence of
confounding noted
This document is a draft for review purposes only and does not constitute Agency policy.
A-161 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
subject by 2 readers;
buccal mucosa cells, stain
Feulgen, 2,000 cells
scored/ subject, 2 readers
Related references:
Viegas et al. (2010);
(Speit et al.. 2012)
Jiang et al. (2010)
China
Prevalence
Population: 151 male
workers from 2 plywood
plants (mean age 27.4 yr,
52.3% smokers) compared
to 112 unexposed workers
at a machine
manufacturer in same
town (mean age 28.7 yr,
42.9% smokers).
Outcome: Cytokinesis-
block micronucleus (CB-
MN), Fenech et al.
(1993), scoring criteria
Fenech et al. (2003),
1,000 binucleated
lymphocytes/ subject,
blinded analysis
Exposure assessed by
job title and personal
air monitoring.
Exposure
concentration ppm
converted to mg/m3
by EPA.
Exposed:
1.08 mg/m3, range
0.1-7.75 mg/m3
Referent: <0.01
mg/m3 (LOD)
Duration:
Mean 2.51 yrs
Range: (0.5-25) yrs
Lymphocyte MN frequency by duration and
formaldehyde concentration
Duration
(yrs)
MNa
Cone.
(mg/m3)
MNb
0.6-1
1-3
3-25
4.33 ±2.81 0.0123°
5.84 ±3.63
5.84 ±
3.24*
0.1353
0.3444
0.4797
3.1488
2.67 ± 1.32
4.03 ± 2.40
5.74 ±
3.13*
6.76 ±
3.81*
8.25 +
3.53*
aANOVA, Dunnett-Hsu test, p =0.04, adjusted for
age, formaldehyde concentration, current smoking
status, alcohol
Haiuj> aiLUMUI
'ANOVA, p <0.05; Trend p <0.001
cReferent group
Viegas et al. (2010)
Portugal
Prevalence study
Population: 30
formaldehyde factory
workers and 50
pathology/anatomy lab
workers exposed for >1
year (40% female, mean
age 35.7 yr, 31.3%
smokers), compared to 85
unexposed individuals
(63.5% female, mean age
33.9 yr, 30.6% smokers)
Outcome: MN assay,
buccal mucosa cells and
peripheral lymphocytes.
Blinded coding and
analysis, Buccal cells,
Personal air sampling,
(N=2 in factory, N=29
in labs) 6-8 hours,
estimated 8-hr TWA
Exposure duration:
Factory workers:
6.2 (1-27) yr
Lab workers:
14.5 (1-33) yr
8-HrTWA
Concentration in:
Factory: 0.26 mg/m3,
range 0.25-0.27
mg/m3
Lab: 0.34 mg/m3,
range 0.06-0.63
mg/m3
Ceiling Concentrations
MN
Referent
Factory
Laboratory
Peripheral
1.17 ±
1.76 ± 2.07
3.7 ±3.86*
lymphocytes
1.95
Buccal cells
0.13 ±
1.27 ±
0.64 ±
0.48
1.55*
1.74*
*p <0.01, Spearman's correlation test
Years of exposure correlated with MN in peripheral
lymphocytes (r = 0.401, p <0.01), and MN in buccal cells (r =
0.209, p = 0.008); Spearman's test
No correlation between MN frequency and smoking or
gender, small magnitude of correlation with age (r = +0.194;
p <0.05 for blood lymphocytes, r = -0.168; p <0.05 for buccal
cells).
This document is a draft for review purposes only and does not constitute Agency policy.
A-162 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Feulgen stain, 2,000 cells
scored/ subject by 4
observers, scoring criteria
(Tolbert et al.. 1992),
peripheral lymphocytes,
stain May-Grunwald-
Giemsa, 1,000 binucleated
cells scored/ subject
Also discussed in (Viegas
et al.. 2013)
Factory: 0.64 mg/m3,
range 0.004-1.28
mg/m3
Lab: 3.1 mg/m3, range
0.03-6.18 mg/m3
Costa et al. (2008)
Portugal
Prevalence study
Population: 30 pathology
lab workers (4 hospitals),
(70% female, mean age 38
yr, 27% smokers)
compared to 30
administrative employees
matched by age, gender,
lifestyle, smoking habits,
and work area (63.3%
female, mean age 37 yrs,
23% smokers).
Outcome: MN in
peripheral lymphocytes
(Teixeira et al.. 2004),
stain 4% Giemsa; scored
1,000 binucleated cells/
subject, scored blind by
one reader, criteria (Caria
et al.. 1995)
Air sampling in
breathing zone,
derived an 8-hr TWA
for each subject
Concentration:
Mean: 0.54 mg/m3,
range: 0.05-1.94
mg/m3
Duration: 11 yrs
Range: (0.5-27) yrs
MN frequency in peripheral lymphocytes
Referent
Exposed
Lymphocyte
3.27 ±0.69
5.47 ±0.76*
MN
P=0.003, Mann-Whitney U-test and Kruskal-Wallis test.
Authors reported positive correlation between formaldehyde
exposure levels and MN frequency (r=0.384, p=0.001)
Pala et al. (2008) Italy
Prevalence study
Population: 36 lab
workers (66.7% female,
mean age 40.1 yr, 16.7%
smokers)
Outcome: Peripheral
lymphocytes (blood
sampled at end of 8-hour
shift), analysis blind to
exposure. MN using
modified cytokinesis-
blocked method, Fenech
et al. (1986); stain 3%
Personal air
monitoring (8-hour
sample);
Exposure categories:
High: > 0.026 mg/m3,
Low: < 0.026 mg/m3
Mean concentration:
Low (n = 25): 0.015
mg/m3 (range
0.005-0.0254)
High (a? = 9): 0.056
mg/m3 (range
0.026-0.269)
Duration of exposure:
NR
Micronuclei Frequency by Exposure Level (mean ±
SD)
<0.026 mg/m3
>0.026 mg/m3
MN
0.26 ±0.24
0.31 ±0.17
Means ratio (95% CI) 1.43 (0.26-7.81), Poisson regression
adjusted for gender, age, smoking and other exposures
This document is a draft for review purposes only and does not constitute Agency policy.
A-163 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Giemsa, 2,000 cells/
subject
Orsiere et al. (2006)
France
Prevalence
Population: 59 hospital
pathology workers from 5
labs (81% female, mean
age 44.7 yr, 20% smokers)
compared to 37
unexposed workers (76%
female, mean age 44 yr,
24% smokers).
Outcome: MN in
peripheral lymphocytes.
Subgroups selected
randomly from initial
groups. Assays conducted
blinded. Cytokinesis-
blocked micronucleus
assay (Sari-Minodier et
al.. 2002); stain 5%
Giemsa, scoring criteria
Fenech (2000), 1,000
binucleated cells/ subject;
FISH with a pan-
centromeric DNA probe,
same
operator scored exposed
and referent blinded
Related reference:
larmacovai et al. (2006)
Personal sampling;
Short-term: 15
minutes, Long-term 8
hours during typical
work-day.
Concentration1:
Mean 15-minute: 2.46
mg/m3, range
<0.12-25. 1 mg/m3
Mean 8-hour 0.123
(range <0.123-0.86
mg/m3
Duration exposure
13.2 years, range
0.5-34 years
Binucleated micronucleated cell rate (BMCR) in
peripheral lymphocytes (mean ± SD)
Unexposed (n=37) Exposed [n=59)
% BMCR
11.1 ±6.0
16.9 ± 9.3*
*Number BMCR per 1,000 binucleated cells, p<0.05,
Mann-Whitney U-test.
Linear regression of BMCR, increase of 0.263 per 1,000
binucleated cells in exposed, p =0.003, adjusting for gender,
age, smoking and alcohol.
FISH Analysis of MN in peripheral lymphocytes by
exposure (mean ± SD)
FISH
Results1
% BMCR
% MN
C+ MN (%)
C - MN (%)
C1+ MN (%)
Cx+ MN (%)
Unexposed
Exposed
p-Value
00
1
II
c
00
1
II
c
11.9 ±5.6
19.1 ±10.1
0.021
14.4 ±8.1
21.0 ± 12.6
0.084
10.3 ±7.1
17.3 ± 11.5
0.059
4.1 ±2.7
3.7 ±4.2
0.338
3.1 ±2.4
11.0 ±6.2
p<0.001
7.8 ±5.5
6.3 ±6.3
0.163
1Results expressed as frequency per 1,000
binucleated cells, mean ± SD; analyzed using Mann-
Whitney U-test
Linear regression of CI + MN, increase of 0.586 MN
containing one centromere per 1,000 binucleated cells in
exposed, <0.001, adjusting for gender, age, smoking and
alcohol
Ye et al. (2005) China
Prevalence study
Population: 18 workers at
a formaldehyde plant at
least 1 yr (38.9% female,
mean age 29 yr, and 16
workers exposed to indoor
air formaldehyde via
building materials (75%
female, mean age 22 yr)
compared to 23 students
with no known source of
formaldehyde exposure
(dormitories) (48% female,
Formaldehyde
sampling: TWA
Concentration
Controls
0.011 ± 0.0025 mg/m3
Max. 0.015 mg/m3
Wait staff
0.107 ±0.067 mg/m3
Max. 0.30 mg/m3
Workers
0.985 ± 0.286 mg/m3
Max. 1.694 mg/m3
Exposure duration:
Workers 8.5 (1-15) yrs
MN frequency in nasal cells
Referent
Wait Staff
HCHO
Workers
MN
1.25 ±0.65
1.75 ± 1.00
2.70 ±
1.50*
P <0.05, one-way ANOVA, values estimated from
figure
This document is a draft for review purposes only and does not constitute Agency policy.
A-164 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
mean age 19 yr); all
nonsmokers
Outcome: MN In nasal
cells, stain Wright's,
scoring criteria {Sarto,
2003, 2443662}, per 3,000
cells, blinding not stated.
Waiters 12 weeks
Burgaz et al. (2002)
Turkey
Prevalence study
Population: 28 pathology
workers (46.4% female,
mean age 29.7 yr, 43%
smokers) and 18
unexposed male
employees (mean age 31.1
yr, 25% smokers), may
overlap with study
population from Burgaz
et al. (2001) Outcome:
MN frequency in buccal
mucosal cells, stain
Feulgen's reaction plus
Fast Green, MN, 3,000
cells/ subject counted,
coded slides, scoring
criteria (Sarto et al.,
1987) and (Tolbert et
al.. 1992)
Concentration:
Range:2.46-4.92
mg/m3
Duration: 4.7 ± 3.33
(1-13) yrs
MN frequency (%) in buccal mucosal cells (mean ±
SD)
Referent Exposed
MNF Frequency 0.33 ±0.30 0.71 ±0.56*
*p <0.05, multifactorial ANOVA adjusting for age,
smoking, and gender
MN frequency was not associated with duration of exposure
Burgaz et al. (2001)
Exposure based on
occupation and
duration of
employment and
quantified via
stationary air
monitors
Exposure conc.:
2.46-4.92 mg/m3
(converted from ppm
by EPA)
Exposure duration:
Mean: 5.06 ± 3.47 Yrs
Range: (1-13) yrs
MN frequency (%) in nasal epithelial cells (mean ±
SD)
Referent Exposed
Turkey
Prevalence study
Population: 23 pathology
workers (12 male, 11
female) occupationally
exposed 5 days, 8 hours/
wk, mean age 30.6 yr, 39%
smokers compared to 25
male university and
hospital staff, mean age
35.4 yr, 76% smokers
Outcome: MN frequency
in nasal cells. Previously
coded slides, stain
Feulgen's reaction plus
Fast Green, MN, 3,000
cells/ subject counted,
MN frequency 0.61 ±0.27 1.01 ±0.62*
*p <0.05, nonparametric test
MN frequency was not associated with duration of exposure.
MN frequency higher in male exposed, similar between
smokers and nonsmokers in referent.
This document is a draft for review purposes only and does not constitute Agency policy.
A-165 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Results
MN frequency (%) in peripheral blood lymphocytes
(mean ± SD)
Referent Exposed
Lymphocyte 3.15 ± 1.46 6.38 ±2.50*
MN
*p <0.01, analytic test not described
MN frequency (%) in buccal mucosa cells
Referent Exposed
Reference and study
design
scoring criteria (Sarto et
al.. 1987) and (Tolbert
etal.. 1992)
He et al. (1998) China
Prevalence study
Population: 13 anatomy
students exposed during a
12-week course (10 hr/
wk) compared to 10
students from same
school. Age and gender
similar between groups,
all non-smokers.
Outcome: MN assay,
(Fenech and Morley.
1985), scored 1,000 cells
per individual, blinding not
described
Kitaeva et al. (1996)
Russia
Prevalence study
Population: anatomy
instructors (8 female, 5
male), mean age 41 yr)
compared to 6 female
unexposed (mean age
28.5 yr); students (6
female, 6 male)
Outcome: MN in buccal
cells, 1994-95. MN in
mucosal cells compared
between exposed and
referent instructors, and
before and after a 40-
minute exposure for
students at 24 and 48
hours. Blinding not
described, stain Feulgen
and light green, analyzed
2,000 cell/ subject
Ballarin et al. (1992)
Italy
Prevalence study
Population: 15 plywood
factory workers (46.7%
Exposure
Breathing zone air
samples during
dissection.
Measurements limited
to location of exposed
students.
Concentration in
breathing zone: Mean
3.17 mg/m3
Duration:
12 weeks (10 hrs/wk)
No quantitative
exposure assessment.
Duration of
employment among
instructors, females
23.6 years; males 25.6
years
17 years
40-minute exposures
Female 0.64 (N=6)
instructors
Before
Female 0.58
students
Male 0.77
students
2.94*
(N=8)
24 Hr Post 48 Hr Post
2.50** 2.64**
2.02* 1.86
*p <0.05, **p <0.01, Student's t-test
Personal sampling;
8-hr TWA (NIOSH,
1977)
Warehouse (N=3)
0.39 ± 0.20 mg/m3,
Mean frequency micronuclei per 1000 cells in nasal
mucosal cells by exposure group
Referent
Exposed
MN (%) (SD)
0.25 (0.22)
0.9 (0.47)"
*p<0.01, Mann-Whitney U test
This document is a draft for review purposes only and does not constitute Agency policy.
A-166 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
female, mean age 31 yrs,)
compared to 15 university
or hospital clerks matched
for age and sex (mean age
31 yr). All nonsmokers.
Outcome: MN in nasal
mucosal cells, stain
feulgen's plus Fast Green,
analysis blinded by one
reader, 6,000 cells/
subject, scoring criteria
(Sarto et al.. 1987).
range 0.21-0.6 mg/m3
Shearing-press (N=8)
0.1 ± 0.02 mg/m3,
range 0.08-0.14
mg/m3
Sawmill (N= 1), 0.09
mg/m3
Inspirable wood dust:
0.11-0.69 mg/m3,
0.73 in sawmill
Employment duration
6.8 yrs
Short-term Studies
Lin et al. (2013) China
Cross-shift change
Population: 62 plywood
workers (17.7% female,
mean age 34 yr, 17.7%
smokers)
Outcome: Peripheral
lymphocytes, cytokinesis-
block micronucleus assay,
Fenech (1993), analyzed
1,000 binucleated cells/
subject, scoring criteria
(Fenech, 1993), Fenech
(2003); blinded analysis
Air sampling and job
function.
Mean exposure: 0.27
± 0.20 mg/m3, range:
0.012-0.67 mg/m3
Mean exposure
duration 2.53 ± 2 yr
Frequency micronuclei in binucleated cells in
peripheral lymphocytes
Before
After exposure
exposure
MN (%)
2.29 ± 1.21
2.29 ± 1.65
p = 0.754, paired Wilcoxon test
Regression coefficients for formaldehyde level, before shift
0.73 (-0.46, 1.92); after shift -0.01 (-1.38, 1.35)
Poisson regression adjusted for age, gender, smoking, and
alcohol
Ying et al. (1997) China
Panel study
Population: 25 non-
smoking anatomy
students (13 males, 12
females, mean age 18.8
yr, Han nationality)
exposed during 8-week
course, 3-hour session, 3
times/ wk.
Outcome: MN Nasal and
Buccal cells, assessed
before the start of the
course and at the end of
8-week period. Blinded
analysis, one observer;
Wright's stain, scored
4,000 cells/ subject; MN
blood lymphocytes, stain
4% Giemsa, scored mean
Air sampling,
estimated TWA and
peak levels during
class and in the
dorms.
Anatomy labs:
Mean TWA: 0.51 ±
0.299 mg/m3, range:
0.07-1.28 mg/m3
Dormitories:
Mean TWA: 0.012 ±
0.003 mg/m3, range:
0.011-0.016 mg/m3
Duration: 8 weeks
Micronucleated Cell Frequency (Mean+SEM),
Change over 8 weeks
Before
After exposure
exposure
Oral Mucosa
0.57 ±0.32
0.86 ±0.56*
Nasal Mucosa
1.20 ±0.67
3.84 ± 1.48*
Lymphocytes
0.91 ±0.39
1.11 ± 1.54
*p <0.01, paired t-test
This document is a draft for review purposes only and does not constitute Agency policy.
A-167 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
of 2870-3167 cells/
subject; MN scoring
criteria (Sarto et al..
1987)
Titenko-Holland et al.
(1996)USA
Panel study
Population: same subjects
as in Suruda et al.
(1993); 35 mortuary
students intermittently
exposed for 90 days (28
students (with adequate
samples, 22 males, 6
females)), age 20-33
years.
Outcome: MN analysis on
buccal and nasal cells
using FISH; blinded
analysis
Related study: Suruda et
al. (1993), same subjects
See Suruda et al.
(1993)
Subjects with
complete MN data
from buccal mucosa
cells (n=19):
Lagged (7-10 days
before the last
sampling):
1.2 ± 2.1 ppm-hrs;
90-day cumulative (90
days):
14.8 ± 7.2 ppm-hrs;
Subjects with
complete MN data
from nasal cells
(n=13):
Lagged (7-10 days):
1.9 ± 2.5 ppm-hrs;
90-day cumulative (90
days): 16.5 ± 5.8 ppm-
hrs
Micronuclei before and after embalming class
(per 1,000 cells) by cell type
Preexposure Postexposure
Buccal Cells (N =
19)
MN Total 0.6 ±0.5
MN+ 0.4 ±0.4
MN" 0.1 ±0.2
Nasal Cells (N = 13)
MN Total 2.0 ±1.3
MN+ 1.2 ±1.3
MN" 0.5 ±0.5
2.0 ±2.0*
1.1 ± 1.3
0.9 ± 1.1*
2.5 ± 1.3
1.0 ±0.8
1.0 ±0.6*
*p <0.05, Wilcoxon sign-rank test, two-tailed
Association with 90-day cumulative exposure for change in
total MN frequency in buccal cells, r =0.44, p =0.06; no
association with 7-10 day lagged exposure, Spearman rank
order correlation
Suruda et al. (1993)
USA
Panel study
Population: 29 students
(with adequate samples)
(24.1% female, mean age
23.6 yr, 17.2% smokers)
exposed to formaldehyde
for 9 weeks during
embalming course, with
baseline samples taken.
Mean duration of
embalming 125 min.
Possible exposure prior to
course.
Outcome: MN assay,
nasal, buccal and
micronucleated peripheral
blood lymphocytes.
Personal sampling for
121 of 144
embalmings;
cumulative exposure
estimated using
sampling data and
time-activity data;
Continuous area
samples over
embalming tables for
short-term peaks;
Concentration1:
Mean: 1.72 mg/m3,
range 0.18-5.29
mg/m3
Duration: 9 weeks
Average cumulative
exposure 18.2
Micronuclei before and after embalming class (per
1000 cells)
Cell type
Before
exposure
After 9 weeks
Buccal 0.046 ±0.17
Nasal 0.41 ±0.52
Micronucleated 4.95 ± 1.72
lymphocytes
0.60 ± 1.27*
0.50 ±0.67
6.36 ±2.03*
*p <0.05, Wilcoxon sign-rank test
Buccal MN in males associated with cumulative exposure,
Spearman coefficient, not nasal MN or micronucleated
lymphocytes
This document is a draft for review purposes only and does not constitute Agency policy.
A-168 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Analysis blinded to
exposure status; MN assay
buccal and nasal cells,
Stich et al. (1982), stain
Feulgen/ Fast Green,
1,500 cell/ subject; MN
lymphocytes Fenech and
Morley (1985), stain
Feulgen 2,000 cells/
subject
mg/m3-hr, range
5.3-41.3 mg/m3-hr
8-hr TWA Mean 0.41
mg/m3, range 0.123 -
1.2 mg/m3
Measurements of
glutaraldehyde,
phenol, & methanol
all < LOD, isopropyl
alcohol < LOD or very
low.
Zeller et al. (2011)
Germany
Controlled human
exposure study
Subjects: 41 healthy
volunteers exposed 4 hr/
day for 5 days, all male,
nonsmokers
Outcome: MN in
peripheral blood
lymphocytes and nasal
mucosa cells assessed
before and after exposure.
Lymphocytes: CBMN test,
scored 1,000 binucleated
cells/ subject on coded
slides. Nuclear division
index (NDI) = # cells with 1
- 4 micronuclei/ Total cells
scored. Nasal cells: scored
2,000 cells/ subject on
coded slides. Difference in
means analyzed using
Cochran Mantel Haentzel
test and ANOVA.
12 groups of 2 to 4
persons in a chamber,
exposures randomly
assigned.
Formaldehyde
concentrations: 0 (i.e.,
background level of
0.01 ppm), 0.3 ppm
(0.37 mg/m3)a with
four peaks of 0.6 ppm
(0.74 mg/m3), 0.4 ppm
(0.49 mg/m3) with
four peaks of 0.8 ppm
(0.98 mg/m3) and 0.5
ppm (0.67 mg/m3) and
0.7 ppm (0.86 mg/m3),
peaks 15 min each, 4
15-min exercise
sessions during
exposure.
Frequency of micronuclei and NDI in lymphocytes
and nasal mucosa before and after 4-hour exposure
over 5 days (N = 40)
Cells with
micronuclei/
1000
Nuclear
Division Index
Lymphocytes
Before 6.5 + 3.226
After 5.7 + 3.339a
Nasal mucosab
Before 0.21 + 0.35
After 0.27+0.42
1-week after 0.24 + 0.43
2-weeks after 0.24 + 0.45
3-weeks after 0.17+0.41
2.0 + 0.232
2.0 + 0.176
ap = 0.11
bSeveral slides could not be analyzed, hence only
1000 cells scored for several individuals (9-13
subjects per sampling time).
Speit et al. (2007a)
Germany
Controlled human
exposure study
Subjects: 21 healthy
volunteers exposed to
formaldehyde for
4hrs/day for 10 days, 11
males, nonsmokers, aged
19-36 years.
Outcome: MN in buccal
mucosal cells assessed
Source: para-
formaldehyde.
Exposure duration:
10 consecutive days, 5
groups of 3-6 persons
in chamber, 4-hour
exposures, some
exposures masked
with ethyl acetate
(EA), 3 15-min
exercise sessions
during exposure.
MN Frequency (per 1000 cells) in Buccal Mucosa,
mean ± SD
Immediately
End of 10-day
before
exposure
exposure
Mean MN
0.86 ± 0.84
1.33 ± 1.45
p = 0.052, Wilcoxon signed rank test
This document is a draft for review purposes only and does not constitute Agency policy.
A-169 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
prior to controlled
exposure and then during
postexposure period.
Blinded analysis at end of
study by one person, stain
DAPI/ propidium iodide,
Analyzed 2,000 cells/
subject
Cumulative exposure
16.6 mg/m3 - hours;
Target concentrations:
0, 0.15, 0.3,0.5,0 +
EA, 0.3 + EA, 0.5 + EA,
0.3 + 4 x 0.6, 0.5 + 4 x
1.0, and 0.4 + 4 x 1.0 +
EA
DNA Damage
Prevalence Studies
Zendehdel et al.
(2017) Iran
Prevalence study
Population: Workers in 3
melamine dinnerware
manufacturing workshops
(n=49) and referents
matched by age and sex
(n=34) who worked in
food industries, # smokers
higher in referent (26%
versus 16%), >90% male.
Recruitment and
participation were not
described.
Outcome: Peripheral
blood cells, Comet assay,
alkaline conditions,
according to Tice et al.,
2000, blinding not
described; minimum of 50
randomly selected cells
per sample; tail moment
and Olive moment
Personal air sampling,
NIOSH method 3500,
whole shift for each
worker.
Median time weighted
average in three
workshops,
0.086 mg/m3; range,
0.02-0.22 mg/m3;
authors state that 2/3
of sample were
exposed to < 0.1
mg/m3
Work duration:
Exposed 2.5 (1-22)
years
Referent 2.0 (1-25)
years
Comparison of DNA damage (comet assay) between exposed
and referent
Olive moment Tail moment
Median (min-max) Median (min-
max)
Exposed 13 (7.4-36.7) 22.2 (12.3-65)
(N = 49)
Referent 8.4 (6.4-31.7) 14.8 (6.4-57.7)
(N = 34)
p value = 0.001; Mann-Whitney test
Costa et al. (2015)
Portugal
Prevalence study
Population: 83 anatomy
pathology workers from 9
hospital laboratories,
exposed to formaldehyde
for at least 1 year,
compared to 87
unexposed employees
from administrative
offices in same geographic
Exposure assessed via
air sampling and
deriving an 8-hr TWA
for each subject.
Exposure
concentration:
Mean: 0.38 ppm (0.47
mg/m3)
Range: 0.28-0.85 ppm
(0.34-1.05 mg/m3)
Comparison of % DNA in tail (comet assay) between
exposed and referent
Mean SD Mean Ratio (95% CI)
Exposed 11.67a 0.72 1.5 (1.14 - 1.96)b
(N = 83)
Referent 7.5 0.47 1.0
(N = 87)
aStudent's t-test, p<0.001
bmodel adjusted for age, gender, smoking habit, and fruit
consumption (# pieces consumed per day)
This document is a draft for review purposes only and does not constitute Agency policy.
A-170 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
area. Exclusions: cancer
Exposure duration
history, radiation therapy
12.0 ± 8.2 years
or chemotherapy, surgery
with anesthesia or blood
transfusion in last year.
Exposed and referent
similar for mean age 39
years, 77% females, 25%
smokers. Outcome:
Peripheral blood samples,
coded, analyses blinded to
exposure status.
Comet assay: alkaline
conditions according to
Singh et al., 1988; Scored
blind 100 cells/ donor
from two gels; % DNA in
comet tail.
Exposed compared to
unexposed using Student's
t-test for In % tDNA; linear
regression of In %tDNA
Peteffi et al. (2015)
Monitoring in 7
Comparisons of DNA damage (comet assay) in
Brazil
sections in facility;
peripheral blood cells, median (interquartile range)
Prevalence study
referent monitoring in
Referent Exposed p-
Population: 46 workers in
5 areas of university;
Value
furniture manaufacturing
breathing zone 8 hr
Damage index 2.0 6.5 0.007
facility (mean age 34.5 yr,
samples collected on
(0-4.0) (1.0-12.5)
56.5% male, 1 smoker)
same day as biological
Damage 2.0 6.0 0.003
and unexposed group (n =
samples. Urine
frequency (%) (0-4.0) (1.0-12.5)
45) recruited from
samples collected at
employees and students
end of work day on 5th
of local university with no
day of work;
No differences between men and women for measures of
history of occupational
correlation of
DNA damage in either exposed or referent
exposure to potentially
formaldehyde
genotoxic agents or
concentration in air
No correlation between urinary formic acid and measures of
substances metabolized to
with urinary formic
DNA damage
formic acid, (mean age
acid concentration, r =
35.4 yr, 33.3% male, 0
0.626, p<0.001
smokers)
Outcome: Peripheral
UV painting,
blood processed within 4
lamination/press,
hr. Comet assay, alkaline
packaging, edge
conditions according to
lamination 0.03-0.04
Tice et al. (2000); silver
ppm (0.037-0.05
nitrate staining according
mg/m3)
to Nadin etal. (2001); 100
Edge painting,
cells/ person read by two
machining and drilling
independent observers
center, board cutting
This document is a draft for review purposes only and does not constitute Agency policy.
A-171 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
(50 cells each), classified
by visual scoring according
to Anderson et al. (1994);
5 categories based on tail
migration (0—IV) and
frequency of damaged
cells (sum of 1—IV), damage
index (Pitarque et al.,
1999)
Nonparametric tests used
because data were not
normally distributed.
Exposed and referent
compared using Mann-
Whitney test
0.06-0.09 ppm
(0.07-0.11 mg/m3))
Referent mean (SD)
0.012 (0.008) ppm
(0.015 (0.01) mg/m3)
Formic acid median
Exposed 20.47 mg/L
Referent 4.57 mg/L
Correlation
formaldehyde
concentration and
formic acid r = -0.626,
p <0.001
Exposure duration
5.76 yr
(Avclin et al., 2013)
Turkey
Prevalence study
Population: 46 male
workers from 2 MDF
plants (mean age 33.4 yr,
39.1% smokers) compared
to 46 non-exposed male
workers in same area
(mean age 38.4 yr, 50%
smokers) (administrative
government offices and
maintenance services).
Half of workers used
personal protective
equipment.
Outcome: DNA damage,
Comet assay, tail intensity,
tail moment, and tail
migration, alkaline
conditions, 100 cells/
subject
24 area samples in
workplaces; personal
samples in breathing
zone over 8 hours.
Mean: 0.25 ± 0.07
mg/m3
Range (0.12-0.41)
Duration:
Mean: 7.3 yrs
Range (0.33-30)
Comparison of Comet assay results in peripheral
blood lymphocytes by exposure
Unexposed Exposed
Tail intensity 5.28 ±0.22 4.25 ±0.29*
Tail moment 0.816 ±0.002 0.624 ±0.003*
Tail migration 2.16 ±0.007 1.68 ±0.005*
*ANOVA, P <0.05
Comparisons by smoking strata indicate similar pattern
Lin et al. (2013) China
Prevalence study
Population: 96 plywood
workers exposed to
formaldehyde (13.5%
female, mean age 33 yr,
30.2% smokers) compared
to referent group (N=82)
(4% female, mean age 31
yr, 40% smokers).
Exposure assessed by
air monitoring and job
assignment.
Average
concentration:
High Exposure, N=38
(making glue): 1.48
mg/m3 (0.914 - 2.044)
Low exposure, N=58
(sanding boards,
Comparison of Comet assay results in peripheral
blood lymphocytes by exposure and duration of
employment.
By Exposure
Referent Low High
Tail 0.67 ± 0.88 ±0.55* 1.01 ±
moment 0.55 0.56*
(Ln)
*ANOVA p-value = 0.006; linear regression model,
trend p-value = 0.002, adjusted for age, gender,
This document is a draft for review purposes only and does not constitute Agency policy.
A-172 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Outcome: Blood
lymphocytes: DNA
damage, Comet assay,
olive tail moment, alkaline
conditions (pH=13), 50
cells/ sample, blinded
analysis.
pressing wood scraps
with glue at high
temp): 0.68 mg/m3
(0.455-0.792)
Referent group, N=82
(providing & grinding
wood scraps): 0.13
mg/m3 (0.019-0.252)
Exposure duration:
2.52 yrs
smoking status, alcohol consumption, duration of
employment
By Number of Work Years
<1 (N= 1-3 (N = 64) >3 (N = 57)
57)
Tail 0.76 ± 0.73 ±0.59 0.99 ± 0.52
moment 0.56
(Ln)
*ANOVA p-value = 0.131; trend p-value = 0.059,
Adjusted for age, gender, smoking status, alcohol
consumption, and formaldehyde levels
Gomaa et al. (2012)
Egypt
Prevalence study
Population: 30 workers in
pathology, histology and
anatomy laboratories at a
university (30% female,
mean age 42.5 yr)
compared to 15 referents
(46.7% female, mean age
39.3 yr). Source of
referent was not
described.
Outcome: Comet assay,
alkaline conditions
according to Singh et al.,
1988; tail length & tail
moment; blinding not
described; analyzed 50
cells per subject
No formaldehyde
measurements;
exposure defined by
job type
Exposure duration:
mean 14.3 yr
Comparisons of Comet assay results by exposure
Unexposed Exposed
Tail length (|am) 12.5 ± 1.5 47.3 ± 8.5*
(7.2-14.7) (16.5-74.2)
Tail moment 10.8 ± 1.2 56.1 ±16.5*
(5.8-13.6) (11.4-88.1)
*Student's t-test, p <0.05; Mean value per 50
comets ± SE, distribution in parentheses
Results comparable between males and females
Costa et al. (2011)
Portugal
Prevalence study
Population: 48 pathology
workers from 5 hospital
laboratories, exposed for
at least 1 year (28%
female, mean age 40 yr,
21% smokers), compared
to 50 unexposed
employees matched by
age, gender, lifestyle,
smoking habits, and work
Air sampling in
breathing zone;
8-hr TWA derived for
each subject.
Concentration: ppm
converted to mg/m3
by EPA.
Mean: 0.53 mg/m3
Range: (0.05-1.94)
Duration:
Mean: 13.6 yrs
Range: (1-31)
Comparisons of Comet assay results by
exposure
Unexposed Exposed
Tail length 42.00 ± 1.6 54.55 ±
2.02*
% DNA Tail 8.01 ±0.64 11.76 ±
0.74*
ANOVA, Student's t-test, p <0.05, compared to referent
group.
Tail length and % tail DNA did not vary by gender, age, or
smoking. Comet assay parameters were not associated with
exposure duration.
This document is a draft for review purposes only and does not constitute Agency policy.
A-173 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
area (25% female, mean
age 37 yr, 14% smokers).
Outcome: DNA damage,
comet assay, tail length
and % tail DNA; alkaline
conditions, 100 cells/
subject; analysis blind to
exposure
Jiang et al. (2010)
China
Prevalence study
Population: 151 male
workers from 2 plywood
plants (mean age 27.4 yr,
52.3% smokers) compared
to 112 unexposed workers
at a machine
manufacturer in same
town (mean age 28.7 yr,
42.9% smokers).
Outcome: Peripheral
blood lymphocytes, Comet
assay, olive tail moment,
alkaline conditions;
blinded analysis, analyzed
> 100 cells/ subject
Related reference: (Yu et
al., 2005) in Chinese
Exposure assessed by
job title and personal
air monitoring.
4 exposure groups
based on 8-hr TWA:
0.135, 0.344, 0.479,
3.141 mg/m3.
Concentration: ppm
converted to mg/m3
by EPA.
Mean: 1.02 mg/m3
Range: (0.1-7.75)
Duration:
Mean: 2.51 Yrs
Range: (0.6 - 25)
Comparison of Comet assay results in peripheral blood
lymphocytes by exposure and duration of employment
Ln tail moment (TM), geometric mean (95% CI)
Referent (n=112) 0.93 (95%CI: 0.78-1.10)
0.135 mg/m3 (n = 60) 2.85 (95%CI: 2.37-3.43)*
0.344 mg/m3 (n=35) 3.01 (95%CI: 2.48-3.64)*
0.479 mg/m3 (n=43) 4.37 (95%CI: 3.78-5.05)*
3.141 mg/m3 (n=13) 8.86 (95%CI:
6.50-12.07)**
*TM compared to referent group, ANOVA, p<0.05;
**TM compared to referent and other exposure
groups, ANOVA p<0.05
Tail moment by exposure history (years)*
0.6-1 (n=33) 2.27 (2.91-3.71)
1-2 (n=68) 2.69(3.50-4.13)
3-25 (n=50) 3.53 (4.11-4.78)**
*ANOVA, p = 0.03, adjusted for age, formaldehyde
exposure history and concentration, current smoking
status, alcohol consumption
**Dunnett-Hsu test, compared to 0.6-1 yr subgroup,
p = 0.01
Costa et al. (2008)
Portugal
Prevalence Study
Population: 30 pathology
lab workers (4 hospitals),
(70% female, mean age 38
yr, 27% smokers)
compared to 30
administrative employees
matched by age, gender,
lifestyle, smoking habits
and work area (63.3%
female, mean age 37 yrs,
23% smokers).
Outcome: Peripheral
lymphocytes; blood
samples collected 10-11
am; Scored blind to
Air sampling in
breathing zone, 8-hr
TWA derived for each
subject
Mean: 0.54 mg/m3
Range: (0.05-1.94)
Years employed:
Mean ± SD: 11 ± 7 yrs
Range: (0.5-27)
Comparisons of Comet assay results in peripheral
blood lymphocytes by exposure
Unexposed Exposed
Tail Length 41.85 ± 1.97 60.00 ± 2.31*
*p <0.05, Student's t-test
Tail length was also significantly longer among exposed
females compared to males. No difference noted by
smoking status
No difference by duration of exposure (data not provided)
This document is a draft for review purposes only and does not constitute Agency policy.
A-174 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
exposure status; Comet
assay, tail length, alkaline
conditions (pH=13), 100
cells/ subject
Short-term Exposure
(Lin et al., 2013) China
Cross-shift change
Population: 62 plywood
workers (17.7% female,
mean age 34 yr, 17.7%
smokers) assessed in
2011.
Outcome: Peripheral
blood lymphocytes,
change over 8-hr shift;
Comet assay, olive tail
moment, alkaline
conditions (pH=13),
blinded analysis, 50 cells/
subject.
Exposure assessed by
air sampling and job
function.
Mean exposure: 0.27
± 0.20 mg/m3
Range: 0.012-0.67
mg/m3
Comet assay results before and after work-shift
Before After exposure
exposure (n= (n= 62)
60)
Ln-transformed 1.47 ± 0.72 2.30 ± 1.28*
Tail moment
* p = < 0.001, paired t-test
Regression coefficients for formaldehyde level, before shift -
0.69 (-2.11, 0.73); after shift 3.64 (1.36, 5.92)
Zeller et al. (2011)
Germany
Controlled human
exposure study
Subjects: 41 healthy
volunteers exposed 4 hr/
day for 5 days, all male,
nonsmokers
Outcome: peripheral
lymphocytes. Comet
assay: alkaline conditions
(pH 13). Analyzed 100
cells/ subject on coded
slides.
12 groups of 2 to 4
persons in a chamber,
exposures randomly
assigned.
Formaldehyde
concentrations: 0,
0.37 mg/m3, with four
peaks of 0.74 mg/m3,
0.49 mg/m3 with four
peaks 0.98 mg/m3 and
0.67 mg/m3 and 0.86
mg/m3, peaks 15 min,
4 15-min exercise
sessions during
exposure.
Results of Comet assay in lymphocytes before and
after 4-hour exposure (N = 37)
Before After exposure
exposure
Tail Moment 0.30 ±0.117 0.33 ±0.118
Tail Intensity 2.28 ±0.492 2.66 ±0.646*
*p = 0.002, Wilcoxon signed rank test, compared to
preexposure level.
DNA Adducts
Bono et al. (2010) Italy
(Prevalence study)
Population: 20
pathologists from 3
pathology wards who
worked in tissue fixation
rooms (production rooms)
and 20 students and
workers from a
university's science labs
Personal sampling
over an 8-hour shift in
each subject; LOD
0.05 ng/m3;
questionnaire data on
job-specific work
(work in production
room where slides
were fixed or other
areas) & use of
Mean levels MidG adducts per 10s NNs by
exposure group
N Mean ± p-Value
SE
Referent 20 2.4 ± 0.3
Exposed 20 5.7 ± 1.3 0.0451
8-hr TWA
<22 ng/m3 13 2.3 ± 0.44
23-66 ng/m3 13 2.7 ±0.55 0.775
>66 ng/m3 13 7.3 ± 1.9 0.0182
This document is a draft for review purposes only and does not constitute Agency policy.
A-175 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Outcome: MidG adducts
in DNA extracted from
whole blood, methods
described in Van Helden et
al., 2009; compared mean
log-transformed MidG
adducts by exposure
tertile or exposure status,
using ANCOVA adjusting
for sex, age, smoking
personal protection
Mean formaldehyde
in production room
0.212 ± 0.047 mg/m3,
other areas 0.0324 ±
0.0061 mg/m3,
referents 0.028 ±
0.0025 mg/m3
1 compared to referent
2 compared to <22 ng/m3
DNA-Protein Crosslinks
Prevalence Studies
Lin et al. (2013) China
(Prevalence)
Population: 96 plywood
workers exposed to
formaldehyde (13.5%
female, mean age 33 yr,
30.2% smokers) compared
to referent group (N=82)
(4% female, mean age 31
yr, 40% smokers).
Outcome: Peripheral
blood lymphocytes: DNA-
protein cross-links (DPX),
KCI- SDS assay, blinded
analysis
Exposure categories
by air monitoring and
job assignment.
Average
concentration:
High exposure, N=38
(making glue): 1.48
mg/m3 (range
0.914-2.044)
Low exposure, N=58
(sanding boards,
pressing wood scraps
with glue at high
temp): 0.68 mg/m3
(range 0.455-0.792)
Referent group, N=82
(providing & grinding
wood scraps): 0.13
mg/m3 (range
0.019-0.252)
Exposure duration:
2.52 yrs
DPX levels in peripheral blood lymphocytes by
formaldehyde exposure and years of employment
DPX by Formaldehyde Level
Referent Low High
DPX 22.73 ± 22.53 ± 20.37 ±
(%) 21.47 22.26 20.52
*ANOVA p-value = 0.894; trend p-value = 0.682,
adjusted for age, gender, smoking status, alcohol use
and duration of employment
DPX by Number of Work Years
<1 (N= 57) 1-3 (N= 64) >3 (N= 57)
DPX 19.34 ± 22.10 ± 25.06 ±
(%) 20.77 20.98 20.57
ANOVA,a p-value = 0.577;b trend p-value = 0.376.
Adjusted for age, gender, smoking status, alcohol use,
formaldehyde exposure levels
b Calculated using linear regression models with
adjustment for age, gender, smoking status, alcohol
use and formaldehyde exposure levels.
Shaham et al. (2003)
Israel
Prevalence study
Population: 186 workers
from 14 hospital
pathology departments
(mean age 45.8 yr, 68.3%
female, 36.6% smokers)
compared to 213
administrative workers
from the same hospitals
(mean age 42.1 yr, 40.4%
Field and personal air
sampling, sample
duration 15 minutes,
multiple times during
work-day (# not
reported).
Concentration
Low exposure: 0.49
(range 0.049-0.86)
mg/m3
High exposure: 2.8
(range 0.89-6.9)
mg/m3
Comparison of DNA-protein crosslinks by exposure
Referent Exposed
Mean DPX/ 0.14 ±0.006 0.21 ±0.006**
total DNA ± SE
**p <0.01, adjusted for age, gender, smoking,
education and region of origin
Mean frequency DNA-protein crosslinks by level of
exposure
Referent Low High
Mean 0.14 0.19 0.20
DPX/total
DNA1
This document is a draft for review purposes only and does not constitute Agency policy.
A-176 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
female, 44.6% smokers).
Age distribution, gender,
origin (ethnicity), and
years of education
differed significantly
between the groups but
were adjusted for in the
analysis.
Outcome: peripheral
blood lymphocytes. Mean
percent DPX of total DNA
in quantity white blood
cells, K-SDS method,
double blinded.
Duration:
Mean: 15.9 yrs
Range: 1-51 yrs
1SE was not provided. Trend by exposure level was
not statistically significant.
Shaham et al. (1997)
Israel
Prevalence study
Population: 12 pathology
workers (mean age 44 yr)
compared to 8 age-
matched controls (mean
age 41 yr).
Outcome: Mean percent
DPX, K-SDS method,
double blinded
Related references
Shaham et al. (1996)
Field and personal air
sampling, sample
duration 15 minutes,
multiple times during
work-day (# not
reported).
Concentration:
Mean: NR
Range: 3.4-3.8 mg/m3
Exposure duration
mean 13 years (range
2 -31 years)
Frequency of DPX by Exposure
Unexposed Exposed
Mean DPX % 23 ± 7 29 ± 6*
*p = 0.03, ANOVA adjusting for smoking status
Years of exposure linearly correlated with DPX levels
Short-term Studies
Lin et al. (2013) China
Cross-shift change
Population: 62 plywood
workers (17.7% female,
mean age 34 yr, 17.7%
smokers)
assessed in 2011.
Outcome: Blood
lymphocytes: % cross links
measured before and
after 8-hour shift, blinded
analysis.
Air sampling and job
function.
Mean exposure: 0.27
± 0.20 mg/m3
Range: 0.012-0.67
mg/m3
DPX frequency before and after work-shift
Before After exposure
exposure (n= (n= 60)
62)
DPX (%) 27.22 ± 10.07 31.68 ±14.19*
*p = 0.019, paired t-test
Regression coefficients for formaldehyde level, before shift
1.70 (-17.84, 21.24); after shift -6.04 (-31.23,19.15)
DNA Repair
Schlink et al. (1999)
Germany
Population: Anatomy
students, Group 1,41
Personal sampling
near breathing zone
once per week,
MGMT activity change compared (U-test, paired data) before
and after exposure; as well as between exposure groups
(Wilcoxon, Mann and Whitney U-test)
This document is a draft for review purposes only and does not constitute Agency policy.
A-177 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
students from one
university course, 3-hr
labs, 2 times per week
(43.9% female, ages 21-30
yr, 39% smokers); Group
2,16 students from a
different university course
(50% female, ages 21-27
yr, 37.5% smokers), and
Referent, 10 unexposed
students (60% female,
ages 22-44 yr, 30%
smokers); no previous
formaldehyde exposure
Outcome: 06-alkylguanine
DNA alkyl-transferase
activity in peripheral blood
lymphocytes (modification
of Klein and Oesch, 1990),
expressed as fmol MGMT/
10s cells (LOD 1 fmol
MGMT/ 10s cells), blind to
period of sample (before
or after); Blood samples
collected before 1st class
and after days 50 and 111
sampling period not
reported,
formaldehyde
exposed, Mean ± SD,
0.2 ± 0.05 mg/m3,
0.14-0.3 mg/m3
Mean MGMT activity by exposure group (fmol
MGMT/ 106 cells)
N Day 0 Day 50 Day > 90
Group 1 41 133.2 131.11 128.21
Group 2 16 146.92
Referent 10 138.9
1p >0.05 compared to Day 0
2p >0.05 compared to referent
MGMT activity did not differ by gender, smoking, allergy
status, or alcohol consumption
Haves et al. (1997)
USA
Panel study
Population: 29 students
(with adequate samples)
exposed to formaldehyde
for 9 weeks during
embalming course 16
male, 7 females, 6
smokers. Mean duration
of embalming 125 min. 15
with previous embalming
exposure within previous
90 days
Outcome: Os-alkylguanine
DNA alkyltransferase
activity in peripheral
lymphocytes, expressed as
pmol AGT/ mg protein
(LOD 0.006 pmol AGT/ mg
protein), blind to period of
sample (before or after);
blood samples collected in
Personal sampling for
121 of 144
embalmings; Exposure
concentration: Mean:
1.72 mg/m3
Range: (0.18-5.29)
mg/m3
Duration:
9 weeks (0.173 yrs)
Total number of
reported embalmings
correlated with
estimated cumulative
formaldehyde
exposure (r = 0.59, p <
0.01).
Individual data pre- and postcourse AGT activity in peripheral
blood lymphocytes depicted in graphs by embalming
experience during previous 90 days (yes/ no), decreased in
17 students, increased in 6 students (ANOVA adjusting for
age, sex and smoking, p < 0.05).
This document is a draft for review purposes only and does not constitute Agency policy.
A-178 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
morning before 1st class
and after 9 weeks
Related reference:
Suruda et al. (1993)
P53 protein levels in blood
Attia et al. (2014)
Urine formic acid
Comparison of plasma p53 and plasma MDA
Egypt
according to Hopner &
concentrations in exposed and referent groups
Prevalence study
Knappe, 1974; unclear
Referent Exposed p-Value
Population: 40 employees
how to relate urine
Plasma 2.78 ± 13.34 ± <0.05
at cosmetic manufacturing
formic acid levels to
p53 0.48 4.67
company (23% male,
air concentrations
(U/ml)
mean age 25.8 yrs, 20%
Plasma 3.59 ± 9.73 ± 2.72 <0.05
smokers) randomly
Urinary formate
MDA 0.83
selected, compared to
Exposed: 53.4 ± 15.01
(nmol/ml)
referent (N=20) selected
mg/L
from hospital
Referent: 12.7 ± 4.57
Correlations in exposed group:
administrative
mg/L
Urinary formate & p53, r=0.912 <0.001
department with
P <0.05
Urinary formate & MDA, r =0.79, p <0.001
comparable SES & no
Plasma MDA & plasma p53, r =0.81, p <0.001
history of occupational
exposure to formaldehyde
Age and gender were not associated with plasma p53,
(35% male, mean age 34
plasma MDA or urinary formate
yrs, 15% smokers)
Outcome: Peripheral
blood; plasma MDA
(commercial kit), plasma
p53 (p53 enzyme-linked
immunosorbent assay kit).
Blinding not stated.
Statistical analyses of
coded data (blinded
assumed). Exposed
compared to referent,
means (Student's t-test),
correlation between
urinary formate and MDA
or p53 using linear
regression
Shaham et al. (2003)
Field and personal air
Comparisons of exposure, serum total p53, serum
Israel
sampling, sample
mutant p53 and DPXs (OR, 95% CI)
Prevalence study
duration 15 minutes,
Total Male Female
multiple times during
Total p53 protein > 150 pg/mLa
Population: 186 workers
work-day (# not
Referent 1.0 1.0 1.0
from 14 hospital
reported).
Exposed 1.6 2.0 0.8
pathology departments
Concentration
(0.8-3.1) (0.9-4.4) (0.2-2.7)
(mean age 42.1 yr, 59.6%
Total p53 protein > 150 pg/mLb
This document is a draft for review purposes only and does not constitute Agency policy.
A-179 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
male, 36.6% smokers)
compared to 213
administrative workers
from the same hospitals
(mean age 45.8 yr, 31.7%
male, 44.6% smokers).
Age distribution, gender,
origin (ethnicity), and
years of education
differed significantly
between the groups but
were adjusted for in the
analysis.
Outcome: p53 proteins
(wild type and mutant) in
serum, p53 quantitative
ELISA kit immunoassay,
mutant p53 in serum using
quantitative ELISA kit
immunoassay. Categorical
analysis of p53 levels
(>pg/mL), exposure
groups compared using
chi-square test; logistic
regression of p53 >150
pg/mL
Low exposure: 0.49
(range 0.049-0.86)
mg/m3
High exposure: 2.8
(range 0.89-6.9)
mg/m3
Duration:
Mean: 15.9 yrs
Range: (1-51) yrs
DPX< 0.187 1.0 1.0 1.0
b
DPX> 0.187 2.5 1.9 2.8
(1.2-5.4) (0.5-7.2) (1.1-7.1)
aLogistic regression models adjusted for sex, age and
smoking
bln the exposed group, logistic regression models adjusted
for sex, age and smoking
bDPX expressed as % of total DNA
Correlations:
Total p53 protein and mutant p53 protein, r =0.75, p <0.01
Proportion p53 > 150 pg/mL among
exposed
DPX< 0.187 33.3%
DPX > 0.187 55.7% (p<0.01)
Genetic Susceptibility
Costa et al. (2015);
2019 Portugal
Prevalence study
Population: 84 anatomy
pathology workers from 9
hospital laboratories,
exposed to formaldehyde
for at least 1 year,
compared to 87 non-
exposed employees from
administrative offices in
same geographic area.
Exclusions: cancer history,
radiation therapy or
chemotherapy, surgery
with anesthesia or blood
transfusion in last year.
Exposed and referent
similar for mean age 39
years, 77% females, 25%
smokers. Outcome:
Exposure assessed via
air sampling and
deriving an 8-hr TWA
for each subject.
Exposure
concentration:
Mean: 0.38 ppm (0.47
mg/m3)
Range: 0.28-0.85 ppm
(0.34-1.05 mg/m3)
Exposure duration
12.0 ±8.2 years
Effect modification by genetic polymorphisms on
associations of formaldehyde with markers of
genotoxicity (mean ratio, 95% CI)
Referent Exposed
N MR (95% CI) N MR (95% CI)
CYP2E1 rs6413432 (% tDNA)
T/T 53 1.00 51 1.61
(1.20-2.16)
T/A+ 15 0.84 7 0.42
A/A (0.54-1.30) (0.20-0.89)
GSTP1 rsl695 (CSAs)
lle/lle 32 1.00 37 5.43
(2.04-14.46)
lle/Val + 55 1.79 47 0.26
Val/Val (1.14-7.94) (0.97-3.27)
XRCC1 rsl799782 (% tDNA)
Arg/Arg 67 1.00 53 1.46
(1.10-1.93)
This document is a draft for review purposes only and does not constitute Agency policy.
A-180 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
Peripheral blood samples,
coded, analyses blinded to
exposure status.
Differences in genotype
distribution evaluated
using Pearson's chi-square
test, effect modification
by genotype in regression
models of exposure on In
% tDNA (comet assay) and
chromosome aberrations,
CYP2E1 rs6413432,
GSTM1 deletion, GSTT1
deletion, GSTP1 rsl695,
XRCC1rsl799782, XRCC1
rs25487, PARP1
rsll36410, MUTYH
rs3219489, XRCC3
rs861539
Arg/Trp 2 0.19 6 4.93
(0.06-0.57) (1.33-18.32)
PARP1 rsll36410 (Multiabberrant cells)
Val/Val 60 1.00 50 5.97
(2.34-15.25)
Val/Ala 8 3.00 9 0.09
(0.55-16.4) (0.01-0.95)
Regression models adjusted for age, gender, smoking habit,
and fruit consumption
Micronuclei frequency (%/1000 cells) by genetic
polymorphisms in formaldehyde exposed and
unexposed workers
Controls
Exposed
Gene site N MeaniSE
N
Mean (SE)
CYP2E1 rs6413432
BNbud
T/T 53 0.36 ± 0.077
51
0.80 ±0.12
T/A+ 15 0.20 ±0.11
7
1.57 ±0.20*
A/A
GSTP1rsl695
MNB
lle/lle 28 0.14 ±0.07
29
0.45 ±0.11
lle/Val + 41 0.20 ± 0.07
33
0.82 ±0.15*
Val/Val
FANCA rs7190823
MNL
Thr/Thr 9 2.33 ± 0.93
12
2.33 ±0.57
Thr/Ala + 77 2.84 ± 0.32
70
4.74 ± 0.44*
Ala/Ala
* p-values CYP2E1 rs6413432 A variant, 0.022; GSTP1
rsl695 Val variant 0.05; FANCA rs7190823 Ala variant
0.019
Ladeira et al. (2013)
Portugal
Prevalence study
Population: 54 hospital
workers in histopathology
labs compared to 82
administrative staff.
Outcome: Genotyping
XRCC3 Met241Thr, ADH5
Val309lle, ADH5
Asp353Glu; associations of
polymorphism with mean
Personal air sampling,
6-8 hours, estimated
8-hr TWA
Exposure conc.:
Mean TWA 8h 0.2 ±
0.14 mg/m3
Mean ceiling value:
1.4 ± 0.91 mg/m3,
range 0.22-3.6 mg/m3
Exposure duration:
14.5 (1-33) yrs
Frequency of micronuclei and nuclear buds (mean ±
SE) in lymphocytes by exposure and genotype
(number in parentheses)
Endpoint
Genotypes
MN
XRCC3
Met/Met
Thr/Met
Thr/Thr
Exposed
2.92 ±0.93
5.05 ±0.98
3.53 ±0.80
(p=0.372)
(13)
(22)
(19)
Referent
1.15 ±0.46
0.70 ±0.30
0.74 ±0.23
(p=0.621)
(20)
(27)
(35)
ADH5
This document is a draft for review purposes only and does not constitute Agency policy.
A-181 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
micronuclei,
nucleoplasm^ bridges and
nuclear buds In
lymphocytes and buccal
cells within exposed and
referent groups, Kruskal-
Wallis test
Related references:
Ladeira et al. (2011)
Val/Val
Val/lle
Exposed
2.57 ±0.65
4.91 ±0.75
(p=0.024)
(21)
(33)
Referent
0.97 ±0.28
0.75 ±0.23
(p=0.176)
(29)
(53)
ADH5
Asp/Asp
Asp/Glu
Exposed
4.08 ±0.91
3.93 ±0.67
(p=0.70
(24)
(30)
Referent
0.86 ±0.23
0.81 ±0.26
(p=0.211)
(35)
(47)
NBUD
XRCC3
Met/Met
Thr/Met
Thr/Thr
Exposed
0.38 ±0.18
1.5 ±0.33
0.21 ±0.12
(p=0.002)
(13)
(22)
(19)
Referent
0.2 ±0.09
0.04 ± 0.04
0.03 ±0.29
(p=0.045)
(20)
(27)
(35)
ADH5
Val/Val
Val/lle
Exposed
0.62 ±0.28
0.88 ±0.21
(p=0.274)
(21)
(33)
Referent
0.00 ± 0.0
0.11 ±0.04
(p=0.061)
(29)
(53)
ADH5
Asp/Asp
Asp/Glu
Exposed
0.71 ±0.23
0.83 ±0.25
(p=0.74)
(24)
(30)
Referent
0.06 ± 0.04
0.09 ± 0.04
(p=0.633)
(35)
(47)
No differences noted for nucleoplasms bridges or
micronuclei in buccal cells (data provided in article)
Santovito et al.
(2011)ltaly
Prevalence study
Population: 20 pathology
workers (mean age 45.7
yr) compared to 16
workers from the same
hospital (mean age 42.1
yr); similar age and gender
distribution. All subjects
were non-smokers and
had not consumed alcohol
in 1 year.
Outcome: Genotypes
GSTT, GSTM; associations
Exposure cone:
Personal air sampling,
8-hour duration.
Referent: Mean: 0.036
± 0.002 mg/m3
Pathologists: Mean:
0.073 ± 0.013 mg/m3
Exposure duration:
Mean: 13 yrs
Range: 2-27 yrs
Frequency of chromosomal aberrations per cell
(mean ± SE) in lymphocytes by exposure and
genotype (number in parentheses)
Exposed Referent
GSTT-pos 0.028 ± 0.003 (16)
GSTT-null 0.04 ± 0.015 (4)
GSTM-pos 0.031 ±0.004 (17)
GSTM-null 0.023 ± 0.003 (3)
0.01 ±0.004 (12)
0.013 ±0.009 (4)
0.01 ± 0.004 (10)
0.012 ± 0.008 (6)
No differences also were found for the % of cells with
chromosomal aberrations (data provided in article)
This document is a draft for review purposes only and does not constitute Agency policy.
A-182 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and study
design
Exposure
Results
of polymorphisms with CA
per cell and % of cells with
aberrations within
exposed and referent
groups; generalized linear
models with Posson
distribution errors
adjusted for gender and
age
Jiang et al. (2010)
Exposure assessed by
Frequency of olive TM (geometric mean (95% CI) in
China
job title and personal
lymphocytes by exposure and genotype (number in
Prevalence
air monitoring.
parentheses)
Population: 151 male
Exposure
Exposed
Referent
workers from 2 plywood
concentration ppm
GSTM1-
3.27 (2.83-3.78)
1.01(0.77-1.32)
plants (mean age 27.4 yr,
converted to mg/m3
pos
74)
(46)
52.3% smokers) compared
by EPA.
GSTM1-
3.86 (3.31-4.5)
0.87 (0.69-1.1) (66)
to 112 unexposed workers
1.08 mg/m3, range
null
(77)
at a machine
P =0.07
P =0.43
manufacturer in same
0.1-7.75 mg/m3
GSTT1-
3.72 (3.26-4.25)
1.04 (0.82-1.31)
town (mean age 28.7 yr,
Duration:
pos
(83)
(63)
42.9% smokers).
GSTT1-
3.36 (2.83-3.99)
0.8 (0.61-1.04) 49)
Outcome: genotypes
Mean 2.51 yrs
null
(68)
GSTM1, GSTT1, GSTP1;
Range: (0.5-25) yrs
P =0.47
P=0.11
associations with olive TM
GSTP1-
3.64 (3.19-4.16)
0.96 (0.74-1.23)
and CBMN frequency
lle/lle
(90)
(58)
within exposed and
GSTP1
3.43 (2.87-4.1)
0.89 (0.7-1.14) (54)
referent; ANCOVA
Val pos
(61)
adjusted for age, smoking
P = 0.49
P = 0.83
and alcohol
Frequency of In CBMN (mean ±
SD) in lymphocytes by
exposure and genotype (number in parentheses)
Exposed
Referent
GSTM1-
5.57 ± 3.45 (74)
2.91 ± 1.5 (46)
pos
GSTM1-
5.5 ±3.32 (77)
2.5 ± 1.15 (66)
null
P = 0.84
P = 0.18
GSTT1-
5.59 ±3.51 (83)
2.75 ±1.41 (63)
pos
GSTT1-
5.46 ± 3.22 (68)
2.57 ± 1.19 (49)
null
P = 0.70
P = 0.47
GSTP1-
5.01 ± 2.98 (90)
2.79 ± 1.36 (58)
lle/lle
GSTP1
6.32 ±3.78 (61)
2.54 ± 1.27 (54)
Val pos
P = 0.05
P =0.26
This document is a draft for review purposes only and does not constitute Agency policy.
A-183 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
ADH, alcohol dehydrogenase; AGT, 06-alkylguanine-DNA alkyltransferase; ANOVA, analysis of variance; C-,
centromere negative; C+, centromere positive; CA, chromosomal aberration; CB-MN or CBMN, cytokinesis block-
micronucleus; CFU-GM, colony forming unit-granulocyte/macrophage; CI, class interval; CSA, chromosome-type
aberration; CSG, centromere separation general; CTA, chromatid-type aberration; DAPI, diamidinophenylindole;
DPX/DPC, DNA-protein crosslink; EA, ethyl acetate; ELISA, enzyme-linked immunosorbent assay; FISH,
fluorescence in situ hybridization; GST, glutathione S-transferase; HCHO, formaldehyde; HF, high frequency; IRR,
incidence rate ratio; K-SDS/KCI-SDS, potassium chloride-sodium dodecyl sulfate; LOD, level of detection; LTR,
lymphocyte transformation rate; MidG, malondialdehyde-deoxyguanosine; MAK, maximum permissible
concentration (German); MDA, malondialdehyde; MGMT, Os-methylguanine methyl transferase; MN,
micronucleus; MR, mean ratio; NSM, number of scored metaphases; OR, odds ratio; PARP, poly (ADP-ribose)
polymerase; PCD, premature centrosome division; PI, proliferation index; SCE, sister chromatid exchange; SD,
standard deviation; SE, standard error; SEM, standard error of the mean; tDNA, tail DNA; TWA, total weighted
average; XRCC, X-ray repair cross complementing.
This document is a draft for review purposes only and does not constitute Agency policy.
A-184 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 A.4.7. Supporting Material for Genotoxicity
2 Literature Search Methods for Genotoxic Endpoints
3 A systematic evaluation of the literature database on studies examining potential genotoxic
4 endpoints in relation to formaldehyde exposure was not conducted. However, a consistent set of
5 search terms was used, initially in September 2012, with regular updates as described elsewhere.
6 These terms were intended to inform the broader topic of mode of action for either respiratory
7 tract or lymphohematopoietic cancers and the retrieved citations were screened for studies on
8 genotoxic endpoints. The search strings used in specific databases are shown in Table A-25.
9 Additional search strategies included:
10 • Review of reference lists in identified articles, and
11 • Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
12 EPA. 2010bl.
Table A-25. Summary of search terms for cancer mechanisms
Mechanisms for Repiratory Tract Cancers - Pubmed
1
(formaldehyde[tiab] OR formaldehyde[mh])
2
AND (nose[tiab] OR nasal[tiab] OR nasopharynx[tiab] OR nasopharyngeal[tiab] OR respiratory[tiab] OR
bronchial[tiab] OR "upper respiratory"[tiab] OR mucociliary[tiab] OR mononuclear[tiab] OR "nasal
mucosa"[tiab] OR "human bronchial"[tiab] OR "nasal cavity"[tiab] OR trachea[tiab] OR "oral mucosa"[tiab] OR
lymphoblasts[tiab] OR "endothelial cells"[tiab] OR "respiratory tract"[tiab] OR olfactory[tiab] OR "nasal
epithelia"[tiab] OR "nasal turbinates"[tiab] OR "nose"[mh] OR "nasopharynx"[mh] OR "trachea"[mh] OR
"smell"[mh])
3
AND (tumor[tiab] OR carcinoma[tiab] OR cancer[tiab] OR neoplastic[tiab] OR cytotoxic[tiab] OR
cytotoxicity[tiab] OR proliferation[tiab] OR "cell proliferation"[tiab] OR immunosuppression[tiab] OR
immune[tiab] OR genotoxicity[tiab] OR genotoxic[tiab] OR mutation[tiab] OR mutagenic[tiab] OR
epigenomic[tiab] OR epigenetic[tiab] OR microRNA[tiab] OR "micro RNA"[tiab] OR methylation[tiab] OR
"chromosome aberration"[tiab] OR "chromosomal aberration"[tiab] OR micronuclei[tiab] OR MN[tiab] OR
micronucleus[tiab] OR "sister chromatid exchange"[tiab] OR SCE[tiab] OR "single strand break"[tiab] OR
SSB[tiab] OR glutathione[tiab] OR oxidation[tiab] OR "oxidative damage"[tiab] OR inflammation[tiab] OR
"DNA-protein crosslink"[tiab] OR DPX[tiab] OR "DNA adduct"[tiab] OR clastogen[tiab] OR clastogenicity[tiab]
OR promotion[tiab] OR promoter[tiab] OR "DNA repair"[tiab] OR "immune activation"[tiab] OR
phagocyte[tiab] OR macrophages[tiab] OR cytogenetic[tiab] OR "regenerative cell proliferation"[tiab] OR
mutagenesis[tiab] OR "DNA-protein crosslinks"[tiab] OR "respiratory cancer"[tiab] OR "nasal cancer"[tiab] OR
"immune function"[tiab] OR "immune biomarkers"[tiab] OR "respiratory disease"[tiab] OR DPC[tiab] OR "DNA
damage"[tiab] OR irritation[tiab] OR bronchitis[tiab] OR "regenerative hyperplasia"[tiab] ORtoxicological[tiab]
OR adenomas[tiab] OR rhinitis[tiab] OR dysplasia[tiab] OR metaplasia[tiab] OR inhalation[tiab] OR
carcinogen[tiab] OR "chromosomal damages"[tiab] OR "nasal carcinoma"[tiab] OR toxicology[tiab] OR
toxicity [tiab] OR "DNA-DNA cross-link"[tiab] OR "respiratory epithelium"[tiab] OR SCC[tiab] OR "pathological
changes"[tiab] OR "histopathological nasal changes"[tiab] OR cilia [tiab] OR "nasal lesions"[tiab] OR "protein
oxidation"[tiab] OR "cellular immunity"[tiab] OR autoantibodies[tiab] OR tumour[tiab] OR "cell damage"[tiab]
OR "neoplasms"[mh] OR "carcinoma"[mh] OR "immunosuppression"[mh] OR "immune tolerance"[mh] OR
"mutation"[mh] OR "epigenomics"[mh] OR "methylation"[mh] OR "glutathione"[mh] OR "inflammation"[mh]
OR "phagocytes"[mh] OR "macrophages"[mh] OR "cytogenetics"[mh] OR "mutagenesis"[mh] OR "nose
neoplasms"[mh] OR "bronchitis"[mh] OR "adenoma"[mh] OR "rhinitis"[mh] OR "metaplasia"[mh] OR
This document is a draft for review purposes only and does not constitute Agency policy.
A-185 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Mechanisms for Repiratory Tract Cancers - Pubmed
"inhalation"[mh] OR "carcinogens"[mh] OR "toxicology"[mh] OR "toxicity"[Subheading] OR "cilia"[mh] OR
"autoantibodies"[mh] OR "immune system phenomena"[mh] OR "mutagens"[mh] OR "Cytotoxicity,
lmmunologic"[mh] OR "Cell Proliferation"[mh] OR "MicroRNAs"[mh] OR "Chromosome Aberrations"[mh] OR
"Sister Chromatid Exchange"[mh] OR "DNA Breaks, Single-Stranded"[mh] OR "DNA Adducts"[mh] OR
"Promoter Regions, Genetic"[mh] OR "DNA Repair"[mh] OR "Respiratory Tract Diseases"[mh] OR "DNA
Damage"[mh] OR "Respiratory Mucosa"[mh] OR "Immunity, Cellular"[mh])
4
NOT ("formalin test"[tiab] OR "formaldehyde fixation"[tiab] OR "formalin fixed"[tiab] OR "formaldehyde
fixed"[tiab] OR formalin-induced[tiab] OR formaldehyde-induced[tiab])
Mechanisms of LHP Cancers - Pubmed
1
(formaldehyde[tiab] OR formaldehyde[mh])
2
AND (blood[tiab] OR lymphocytes[tiab] OR "bone marrow"[tiab] OR hematopoietic[tiab] OR "hematopoietic
stem cells"[tiab] OR leukocytes[tiab] OR "white blood cell"[tiab] OR "NK cell"[tiab] OR "natural killer cell"[tiab]
OR b-lymphocyte[tiab] OR b-cell[tiab] OR t-lymphocyte[tiab] OR t-cell[tiab] OR leukemia[tiab] OR
lymphoma[tiab] OR myeloid[tiab] OR serum[tiab] OR albumin[tiab] OR adduct[tiab] OR genotoxic[tiab] OR
aneuploidy[tiab] OR pancytopenia[tiab] OR epigenomics[tiab] OR epigenetic[tiab] OR microRNA[tiab] OR
"micro rna"[tiab] OR methylation[tiab] OR "chromosome aberration"[tiab] OR "chromosomal
aberration"[tiab] OR micronucleus[tiab] OR "sister chromatid exchange"[tiab] OR glutathione[tiab] OR
oxidation[tiab] OR "oxidative damage"[tiab] OR inflammation[tiab] OR dna-protein-crosslink[tiab] OR "dna
adduct"[tiab] OR "immune activation"[tiab] OR "blood"[Subheading] OR "blood"[mh] OR "lymphocytes"[mh]
OR "lymphocyte count"[mh] OR "bone marrow"[mh] OR "hematopoietic system"[mh] OR "hematopoietic
stem cells"[mh] OR "leukocytes"[mh] OR "leukocyte count"[mh] OR "leukocytes"[mh] OR "killer cells,
natural"[mh] OR "killer cells, natural"[mh] OR "b-lymphocytes"[mh] OR "b-lymphocytes"[mh] OR "t-
lymphocytes"[mh] OR "t-lymphocytes"[mh] OR "leukemia"[mh] OR "lymphoma"[mh] OR "serum"[mh] OR
"albumins"[mh] OR "aneuploidy"[mh] OR "pancytopenia"[mh] OR "epigenomics"[mh] OR "epigenomics"[mh]
OR "micrornas"[mh] OR "micrornas"[mh] OR "methylation"[mh] OR "chromosome aberrations"[mh] OR
"chromosome aberrations"[mh] OR "sister chromatid exchange"[mh] OR "glutathione"[mh] OR
"inflammation"[mh] OR "dna adducts"[mh])
3
NOT ("formalin test"[tiab] OR "formaldehyde fixation"[tiab] OR "formalin fixed"[tiab] OR "formaldehyde
fixed"[tiab] OR formalin-induced[tiab] OR formaldehyde-induced[tiab])
Mechanisms of Respiratory Tract Cancers - WoS
1
Formaldehyde (Title only)
2
AND (nose OR nasal OR nasopharynx OR nasopharyngeal OR respiratory OR bronchial OR upper-respiratory
OR mucociliary OR mononuclear OR nasal-mucosa OR human-bronchial OR nasal-cavity OR trachea OR oral-
mucosa OR lymphoblasts OR endothelial-cells OR respiratory-tract OR olfactory OR nasal-epithelia OR nasal-
turbinates)
3
AND (tumor OR carcinoma OR cancer OR neoplastic OR cytotoxic OR cytotoxicity OR proliferation OR
immunosuppression OR immune OR genotoxicity OR genotoxic OR mutation OR mutagenic OR epigenomic OR
epigenetic OR microRNA OR micro-RNA OR methylation OR chromosome-aberration OR chromosomal-
aberration OR micronuclei OR MN OR micronucleus OR sister-chromatid-exchange OR SCE OR single-strand-
break OR SSB OR glutathione OR oxidation OR oxidative-damage OR inflammation OR DNA-protein-crosslink
OR DPX OR DNA-adduct OR clastogen OR clastogenicity OR promotion OR promoter OR DNA-repair OR
immune-activation-phagocyte OR macrophages OR cytogenetic OR regenerative-cell-proliferation OR
mutagenesis OR DNA-protein-crosslinks OR respiratory-cancer OR nasal-cancer OR immune-function OR
immune-biomarkers OR respiratory-disease OR DPC OR DNA-damage OR irritation OR bronchitis OR
regenerative-hyperplasia OR toxicological OR adenomas OR rhinitis OR dysplasia OR metaplasia OR inhalation
OR carcinogen OR chromosomal-damages OR bronchitis OR nasal-carcinoma OR toxicology OR toxicity OR
DNA-DNA-cross-link OR respiratory-epithelium OR SCC OR pathological-changes OR histopathological-nasal-
changes OR cilia OR nasal-lesions OR protein-oxidation OR cellular-immunity OR autoantibodies OR tumour
OR cell-damage)
This document is a draft for review purposes only and does not constitute Agency policy.
A-186 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
Mechanisms for Repiratory Tract Cancers - Pubmed
4
NOT (formalin-test OR formaldehyde-fixation OR formalin-fixed OR formaldehyde-fixed OR formalin-induced
OR formaldehyde-induced)
Mechanisms of LHP Cancers - WoS
1
Formaldehyde (Title only)
2
AND (blood OR lymphocytes OR bone-marrow OR hematopoietic OR hematopoietic-stem-cells OR leukocytes
OR white-blood-cell OR NK-cell OR natural-killer-cell OR b-lymphocyte OR b-cell OR t-lymphocyte OR t-cell OR
leukemia OR lymphoma OR myeloid OR serum OR albumin OR adduct OR genotoxic OR aneuploidy OR
pancytopenia OR epigenomics OR epigenetic OR microRNA OR micro-rna OR methylation OR chromosome-
aberration OR chromosomal-aberration OR micronucleus OR sister-chromatid-exchange OR glutathione OR
oxidation OR oxidative-damage OR inflammation OR dna-protein-crosslink OR dna-adduct OR immune-
activation)
3
NOT (formalin-test OR formaldehyde-fixation OR formalin-fixed OR formaldehyde-fixed OR formalin-induced
OR formaldehyde-induced)
Study Evaluations of Epidemiological Studies of Genotoxic Endpoints
Epidemiological studies examining genotoxic endpoints were evaluated for potential bias and other
issues using the same domains as were assessed for studies in other health effects categories (see
Table A-26). Rather than confidence conclusions of low, medium or high, an overall conclusion of
"no obvious bias" was used if no concerns were identified. For studies with a potential bias
identified, the potential bias or issue was summarized in the comment row. For each assay (e.g.,
chromosomal aberrations, CBMN, Comet assay), factors related to assay methods that could affect
the endpoint values were identified using published reviews from collaborations that compared
assay methods across epidemiological studies (Moller et al., 2020; Fenech et al., 2020; (Bonassiet
al.. 2011: Fenech etal.. 20111 Valverde et al., 2009; Bonassi et al., 2005). Such factors included
sample collection and processing flows, whether sample processing and analysis was blinded to
exposure status, cell culture details, details of scoring (number of scorers, criteria, staining number
of cells scored). An appropriate citation to a standardized assay protocol was considered
acceptable. These reviews noted that assay results have been found to vary by age, gender and
smoking status; studies that did not report assessing confounding by these factors were identified.
In the study evaluation table for each study, row cells have been given a grey fill for evaluation
domains with identified concerns about methods. Study evaluation concerns are discussed in the
syntheses of genotoxic endpoints if they may explain observed heterogeneity in study results.
This document is a draft for review purposes only and does not constitute Agency policy.
A-187 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-26. Evaluation of genotoxicity endpoints in epidemiology studies of formaldehyde exposure
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
Aglan (2018)
Passive air
Blood collected at end
60 female
Exposed
Comparisons
Unexposed n =
Reporting
(Egypt)
sampling (Umex-
of 8-hour shift on day
hairstylists selected
participants were
between
60
deficiencies result
Hair stylists
100) at fixed
hair straightening
between June 2015
comparable for
unexposed, group 1
Group 1
in some concern
position in
occurred, processed
and September
work tasks, number
and group 2 using
n = 31
about potential
breathing zone,
within 6 hours.
2016, aged 20-36
of clients and work
Kruskal Wallis test
Group 2
for selection bias.
15-minute samples
Cytokinesis block
years with
duration. Only
for nonnormally
n = 29
during hair
micronucleus test in
comparable work
nonsmokers were
distributed
Comparisons were
straightening
lymphocytes (Maffei
hours, number of
included, and all
variables (MNL and
for duration of
process;
et al, 2002). Replicate
clients, usual tasks
were female.
MNB) and least
exposure (greater
15-minute TWA
cultures for each
included hair
Exposed and
significant
or less than 5
Group 1 (work
sample, incubated 72
straightening and
unexposed were
difference.
years) and 15-min
duration < 5
hours, cytochalasin-B
no gaps in
"matched" for age,
Comparisons were
TWA
years): 1.68 ± 0.27
added for the last 28
employment.
residency,
across duration
concentrations
PPm
hours. 1,000
Excluded subjects
nutritional habits
(greater or less
also were
Group 2 (work
binucleated cells
with chronic
and SES.
than 5 years) and
statistically
duration > 5
examined per person.
disease and /or
15-min TWA
different in these
years): 1.83 ± 0.16
2,000 binucleated
regular
concentrations
groups.
PPm
cells from coded slides
(1,000 from each
replicate culture),
scored using criteria
by(Fenech et al.,
2003). MN frequency
medications, family
history of cancer,
recurrent
abortions, smoking
or pregnancy.
Comparison group
were higher in
Group 2 (p = 0.03, t
test).
% altered cells.
MN in exfoliated
buccal cells. Cheeks
scraped with wooden
spatula, fixed in 3:1
methanol/acetic acid
and dropped onto
was 60 healthy
female hair stylists
who did not
straighten hair
"matched age,
residency,
nutritional habits,
This document is a draft for review purposes only and does not constitute Agency policy.
A-188 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
slides. Air dried slides
stained with
Feulgen/Fast Green,
examined at 400x
according to Tolbert
et al., 1991. Analyzed
independently by 2
people, 1,500 cells
scored per person
using criteria by
(Sarto et al., 1987)
% altered cells.
and socio-
economic
standard."
Participation rates
not reported. No
data provided to
confirm asserted
comparability
between exposed
and referents.
Attia et al.
Urine formic acid
according to
Hopner & Knappe,
1974; unclear how
to relate urine
formic acid levels
to air
concentrations
Peripheral blood;
plasma MDA
(commercial kit),
plasma p53 (p53
enzyme-linked
immunosorbent assay
kit.
Blinding not stated,
but likely minimal bias
because
interpretation not
required
40 employees at
company randomly
selected compared
to referent (N = 20)
selected from
hospital
administrative
department with
comparable gender
and SES & no
history of
occupational
exposure to
formaldehyde
Age differed
between exposed
and referent, but
age and gender
were not associated
with formate levels,
MDA levels, or p53
levels
Analyses of coded
data (blinded
assumed)
Exposed compared
to referent, means
(Student's t-test),
correlation
between urinary
formate and MDA
or p53 using linear
regression
Exposed n = 40,
referent n = 20
No obvious bias
(2014) (Egypt)
Cosmetic
manufacture
(Avclin et al.,
2013) (Turkey)
24 area samples in
workplaces;
personal samples
in breathing zone
over 8-hour
period. 8-hour
TWA calculated
Peripheral blood
lymphocytes; samples
processed within 6 hr,
comet assay, tail
intensity, tail moment,
and tail migration,
alkaline conditions,
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported. 46
Exposed and
referent
comparable with
respect to age, sex,
lifestyle, and
smoking habit. No
history of
ANOVA or Kruskal-
Wallis H test
depending on test
for normality;
presented mean &
SD by exposure
Exposed N = 46
Referent N = 46
No obvious bias
Medium
density
fiberboard
plants
This document is a draft for review purposes only and does not constitute Agency policy.
A-189 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
(prevalence
study)
Singh et al., 1988, cells
lysed >1 hr,
electrophoresis 20
min, 100 cells/ subject
(2 replicates), image
analysis software.
Blinding not stated
male workers
compared to 46
nonexposed males
in same area
(administrative
government offices
and maintenance
services)
occupational
exposure to
formaldehyde or
other chemicals
group, stratified by
smoking status
Results of test for
normality were not
reported, comet
assay endpoints
were not In-
transformed
(Ballarin et
al.. 1992)
Personal samplers,
Sampling in
warehouse (N = 3)
shearing-press
(N = 8) & sawmill
(N = 1), sampled
formaldehyde and
wood dust
Calculated 8-hr
TWA, reference for
measurements
(NIOSH, 1977).
Nasal respiratory
mucosa cells, cell
collection using
endocervical brush,
smeared onto
previously coded
slides, stain Feulgen's
reaction plus Fast
Green, MN, analysis
blinded by one reader
for cytogenetic, 6,000
cells/subject, scoring
criteria (Sarto et al.,
1987)
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Referent from
different source
population:
university or
hospital clerks;
excluded heavy
drinkers
All nonsmokers,
matched to referent
for age and sex
Differences
analyzed using
Mann-Whitney test
Exposed n = 15;
Referent n = 15
Small sample
numbers; no
obvious bias
(Italy)
Plywood
factory
(Bauchinger
and Schmid,
Exposure
assessment based
on air monitoring
and job-function.
Sampling design
and duration was
not described.
Peripheral
lymphocytes, CA/ cell
(scored 500
cells/subject), Giemsa
staining; SCE/cell
(scored 50/subject)
analyzed using coded
slides
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Exposed and
All male,
Comparable for age,
more smokers
among referent; no
previous radiation
history or exposure
to other industrial
chemicals
Mann-Whitney
rank U test to
compare groups,
SCE analysis
stratified by
smoking
Exposed N = 20;
Referent N = 20
Possible bias
toward null
because no
adjustment for
smoking in CA
analysis
1985)
Germany
Papermaking
This document is a draft for review purposes only and does not constitute Agency policy.
A-190 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
referent worked at
same factory
Bono et al.
Personal sampling
over an 8-hour
shift in each
subject; LOD 0.05
Hg/m3;
questionnaire data
on job-specific
work (work in
production room
where slides were
fixed or other
areas) & use of
personal
protection
MidG adducts in DNA
extracted from whole
blood, methods
described in Van
Helden et al., 2009;
evaluated in 20 out of
40 exposed and 20
out of 32 referent
workers (selection
criteria were not
described)
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Recruited workers
from 3 pathology
labs and workers &
students from a
university lab with
no exposure to
formaldehyde
Mean
formaldehyde levels
varied by age,
smoking, and
exposure status
(referent, work in
production room,
work in other
areas); confounding
assessed in analysis
Formaldehyde
exposure tertiles
based on 8-hr
average
formaldehyde
concentration,
compared mean
log-transformed
MidG adducts by
exposure tertile or
exposure status,
using ANCOVA
adjusting for sex,
age, smoking;
evaluated multiple
comparisons using
Dunnett tests
Exposed N = 20
Referent N = 20
No obvious bias;
small sample size
especially for
analysis of effect
modification by
smoking
(2010) Italv
Pathology labs
Bouraoui et
Area sample in
macroscopic room,
diffuse radical
samplers
containing 2,4-
dinitrophenyl-
hydrazine, 24-hour
duration, 3
samplings.
Cytokinesis-blocked
MN assay in
peripheral
lymphocytes in
combination with FISH
using all-chromosome
centromeric probe
(Sari-Minodier et
al., 2002); cultured
Recruitment and
selection not
described.
Participation rates
not reported.
Excluded x-ray
history during
previous 6 months,
use of drugs
Comparison groups
were similar for
potential
confounders
Multivariate
regression of
genotoxic markers
with possible
confounders
excluding smokers;
age and gender
were associated
but exposure
groups were
comparable
Exposed n = 31
Referent n = 31
No obvious bias
al. (2013)
Tunisia
Anatomy/
pathology lab
in hospital
72 hr, smeared onto
slides, stain 5%
This document is a draft for review purposes only and does not constitute Agency policy.
A-191 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
Giemsa, 2,000
binucleated cells
scored/subject,
criteria (Fenech,
2000) blinding not
described.
Burgaz et al.
Stationary area
Nasal respiratory
Recruitment and
Higher proportion
Comparison of
Exposed n = 23,
Possible bias to
(2001)
(Turkey)
Anatomy/
pathology
departments in
hospital &
university
measurements;
mucosal cells;
selection not
of females in
means using
Referent n = 25
null because of
number of samples
collected using
described.
exposed (referent
nonparametric
age in referent
and duration not
endocervical brush,
Referents worked
was only male),
methods, two-
reported
cells smeared onto
in same hospital &
slightly older
tailed tests,
previously coded
university
individuals, and
stratified by
slides, stain Feulgen's
smokers (and heavy
smoking;
reaction plus Fast
smokers) in
correlation using
Green, MN, 3,000
cells/ subject counted,
scoring criteria (Sarto
et al., 1987) and
referent. Analyses
stratified by
smoking. Stated
that referents had
Spearman's test
(Tolbert et al.,
1992)
no occupational
exposure to
genotoxic agents.
Burgaz et al.
Stationary area
Buccal mucosal cells;
Recruitment and
Higher proportion
Comparison of
Exposed n = 28,
No obvious bias
(2002)
(Turkey)
Anatomy/
pathology
departments in
hospital &
university
measurements;
cells collected with
selection not
of females (referent
means using
Referent n = 18
number of samples
wooden spatula,
described.
was only male), and
nonparametric
and duration not
smeared onto slides,
Referents worked
smokers in referent.
methods (Mann-
reported
stain Feulgen's
in same hospital &
Age comparable.
Whitney test), two-
reaction plus Fast
university
Stated that
tailed tests,
Green, MN, 3,000
referents had no
correlation using
cells/ subject counted,
occupational
Spearman's test
coded slides, scoring
exposure to
Multifactorial
Possible
overlap with
criteria (Sarto et al..
genotoxic agents;
ANOVA adjusting
for smoking,
This document is a draft for review purposes only and does not constitute Agency policy.
A-192 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Burgaz et al.
(2001)
1987) and (Tolbert
et al.. 1992)
exposure and
gender and age
Costa et al.
(2008)
(Portugal)
Hospital
pathology
laboratories
(i = 4)
(prevalence)
Samples in
breathing zone,
NIOSH method
#3500. Sampling
duration, sample
number were not
given.
8-HOUR TWA
calculated for each
worker
Peripheral
lymphocytes; blood
samples collected
10-11 am; processed
immediately; Scored
blind to exposure
status; Comet assay,
parameter: tail length,
alkaline conditions
(pH=13), Singh et al.,
1988, lysis 1 hr, 20
min electrophoresis,
100 cells/ subject,
image analysis
software;
Cytokinesis-blocked
MN test, (Teixeira et
al.. 2004); culture
incubation 72 hr;
samples applied by
smears to slides, stain
4% Giemsa; scored
1,000 binucleated
cells/subject, scored
blind by one reader,
criteria (Caria et al..
1995); SCE/ cell, 50
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Unexposed worked
in administrative
offices in hospitals
in proximity to
pathology labs
Exposed matched
to unexposed by
age, gender,
lifestyle and
smoking habits;
unexposed worked
in same area in
administrative
offices
Demographic
information
provided
Analyses by one-
way ANOVA and
Student's t-test
Exposed n = 30;
Referent n = 30
No obvious bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-193 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
2nd division
metaphases scored by
one observer,
Scored blind to
exposure status
Costa et al.
Samples in
breathing zone,
NIOSH method
#3500. Sampling
duration, sample
number was not
given.
8-hr TWA
calculated for each
worker
Peripheral
lymphocytes; blood
samples collected
10-11 am; processed
immediately; scored
blind to exposure
status;
comet assay,
parameter: tail length
and % tail DNA;
alkaline conditions,
Singh et al., 1988,100
cells/subject, image
analysis software;
Cytokinesis-blocked
MN test (Teixeira et
al., 2004); culture
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Excluded exposed
with <1 yr
employment.
Unexposed worked
in administrative
offices in hospitals
in proximity to
pathology labs.
Exposed matched
to unexposed by
age, gender, and
smoking habits.
Demographic
information
provided
Comet assay:
normal
distribution,
analyses by one-
way ANOVA and
Student's t-test
MN: not normal
distribution, used
nonparametric
tests, Mann-
Whitney U test and
Kruskal-Wallis test
Exposed n = 48;
Referent n = 50
No obvious bias.
(2011)
(Portugal)
Hospital
pathology
laboratories
(n = 5)
(prevalence)
incubation 72 hr;
samples applied by
smears to slides, stain
4% Giemsa; scored
1,000 binucleated
cells/subject, scored
blind by one reader,
criteria Fenech
(2007)
This document is a draft for review purposes only and does not constitute Agency policy.
A-194 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Costa et al.
# samples and
duration not
reported. Air
sampling in
breathing zone.
Calculated 8-hr
TWA for each
subject; NIOSH
method # 3500
Peripheral blood
samples collected
between 10-11 am.
Samples processed
and
assays conducted
blinded. Cytokinesis-
blocked MN test
(Teixeira et al.,
2004). 1,000 cells
Included workers
with at least 1-year
employment in
4 hospital
pathology anatomy
labs; referent
worked in
administrative
offices in same
area & no
occupational
exposure history to
formaldehyde
Similar in gender
distribution, age,
BMI, and smoking
habit
Demographic
information
provided
Difference in
means, Student's t-
test; tested for
normal distribution
multivariate
analysis adjusted
for age, gender,
and smoking
Exposed n = 35;
referent n = 35
No obvious bias
(2013)
(Portugal)
Anatomy/
pathology lab
workers
analyzed/subject,
MN per 1,000
binucleated cells,
scored blindly by one
reader, criteria Fenech
(2007).
SCE, scored 50 M2
metaphases/ subject
by one reader
T-Cell Receptor
mutation assay in
mononuclear
leukocytes, # events in
mutation cell window
(CD3-CD4+ cells)
divided by total
number of events for
CD4+ cells
Costa et al.
Samples in
breathing zone for
periods during
formaldehyde-
Peripheral blood
samples collected
between 10-11 am.
Included workers
with at least 1-year
employment in
4 hospital
Similar distributions
by exposure group
for age, gender, and
smoking. Evaluated
Exposed compared
to unexposed using
Student's t test for
In % tDNA or
Exposed = 84;
Unexposed = 87
No obvious bias
(2015) Portugal
This document is a draft for review purposes only and does not constitute Agency policy.
A-195 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
Anatomy/
related tasks,
Samples processed
pathology anatomy
possible
Mann-Whitney Li-
pathology
NIOSH method
and
labs; referent
confounding by
test for CA
laboratories
#3500. Sampling
analyzed blinded.
worked in
other measures
measures; linear
duration, sample
Chromosome
administrative
(diet) and found
regression of In
number was not
aberrations (structural
offices in same
confounding by fruit
%tDNA; negative
given.
and numerical),
area & no
consumption for
binomial regression
8-hr TWA
duplicates cultured 51
occupational
frequency of
for untransformed
calculated for each
hours (cited Roma-
exposure history to
multiaberrant cells
total-CAs, CSAs,
worker
Torres et al., 2006),
formaldehyde;
and %tDNA.
CTAs, gaps,
4% Giemsa stain;
exclusions
aneuploidies, &
coded slides; scored
cancer/tumor
aberrant cells;
100 metaphases per
history, radiation
Poisson regression
person, 1250x
therapy or
for untransformed
magnification; CTAs &
chemotherapy
multiaberrant cells.
CSAs according to
treatments, last
Models adjusted
Savage et al., 1975;
year surgery with
for age, gender and
gaps not included.
anesthesia and
smoking plus actual
Comet assay: alkaline
blood transfusions.
confounders for
conditions according
specific
to Singh et al., 1988;
parameters.
Scored blind 100
Analyzed effect
cells/donor from two
modification by
gels; % DNA in comet
genotype
tail.
(homozygous
variant plus
heterozygous)
compared to
homozygous
wildtype, genotype
frequency
compared by
Pearson's chi-
square test
This document is a draft for review purposes only and does not constitute Agency policy.
A-196 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Costa et al.
(2019) Portugal
Anatomy/
pathology
laboratories
Samples in
breathing zone for
periods during
formaldehyde-
related tasks and
at other sites
"considered
relevant", NIOSH
method #3500.
Sampling duration
and number were
not given.
8-hr TWA
calculated for each
worker
Peripheral blood
samples collected and
processed and assays
conducted blinded.
Exfoliated cells were
collected for each
cheek separately.
Cytokinesis-blocked
MN test, (Costa et
al.. 2008); culture
incubation 72 hr;
samples applied by
smears to slides, stain
4% Giemsa; scored
1,000 binucleated
cells/subject, scored
blind by one reader,
criteria defined by
Fenech et al. (2007)
Buccal MN cytome
assay. Scored blind by
same reader, 2,000
differentiated cells
scored for frequency
of MN, nuclear buds
and nucleoplasm^
bridges according to
Tolbert et al. 1992 and
Thomas et al. 2009.
SCE/ cell, 50 2nd
division metaphases
This study analyzed
additional
endpoints using
blood and buccal
cell samples
collected in Costa
et al. (2015).
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Included workers
with at least 1-year
employment in
9 hospital
pathology anatomy
labs; referent
worked in
administrative
offices in same
area & no
occupational
exposure history to
formaldehyde.
Similar distributions
by exposure group
for age, gender, and
smoking. Exposed
smokers smoked
less than unexposed
smokers (11 versus
15 pack-years).
Evaluated possible
confounding by
other measures
(diet) and found
confounding by fruit
consumption for
frequency of
multiaberrant cells
and %tDNA. The
association of
exposure with
possible
confounders was
examined using
linear regression.
Dietary habits were
reported to be
parameter-specific
actual confounders
for white blood cell
counts.
Sample size varied
by endpoint
because of "sample
limitation and/or
technical losses,"
although
missingness likely
not associated with
exposure. Data
were log
transformed to
approximate
normal distribuion
forTCR-Mf and
Mann-Whitney U
test applied to MN
in lymphocytes and
buccal cells and
nuclear buds in
buccal cells.
Associations (mean
ratio (MR), 95% CI)
with SCE, MNB,
BNbud and log TCR-
Mf were assessed
using Poison
regression.
Untransformed
MNL also were
modeled using
negative binomial
regression. Models
adjusted for age,
MNL
Exposed = 84;
Unexposed = 87
SCE/cell
Exposed = 84;
Unexposed = 87
MNB
Exposed = 63;
Unexposed = 69
BNbud
Exposed = 63;
Unexposed = 69
TCR-Mf
Exposed = 61;
Unexposed = 64
No obvious bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-197 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
scored by one
observer,
Scored blind to
exposure status.
T-Cell Receptor
mutation assay in
mononuclear
leukocytes, flow
cytometry, minimum
of 2.5 x 105
lymphocyte-gated
events were acquired,
# events in mutation
cell window (CD3-
CD4+ cells) divided by
total number of
events for CD4+ cells
gender, smoking
habits and dietary
habits.
Effect modification
by genotype
analyzed using
Mann-Whitney U
test for specific
polymorphisms in
CYP2E1, GSTM1,
GSTT1, GSTP1,
SRCC1, PARP1,
MUTYH, RAD51
BRIP1 and FANCA.
Fleig et al.
Personal sampling,
8-hour shift,
number of
measurements or
people with
monitors not
reported.
Measurements
were not reported.
Provided
categories of
maximum
exposure as % of
MAK value for
25%, 60%, and
100% of MAK for
Chromosome
aberrations,
peripheral blood
lymphocytes cultured
70-72 hours, 10%
Giemsa stain; coded
slides.
Presented aberrant
cells/ individual both
including gaps and
excluding gaps
Recruitment and
selection of
participants not
described.
Referent group
from
administrative or
office staff at same
site with no
formaldehyde
exposure
Referent matched
to exposed by age
and gender; stated
smoking not
associated with CA
(data not reported)
Fisher-Yates exact
test
Exposed n = 15,
referent n = 15
Cell incubation
period 72 hours
(1982)
Germany
Formaldehyde
manufacturing
This document is a draft for review purposes only and does not constitute Agency policy.
A-198 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
two periods
(before and after
1971
Gomaa et al.
No formaldehyde
measurements
Chromosome
aberrations (structural
and numerical), cited
Verma (1998),
peripheral blood
lymphocytes cultured
72 hours, 5% Giemsa
stain; blinding not
described; scored
total CA and types,
analyzed 50-100
metaphases per
subject.
Comet assay, alkaline
conditions according
to Singh et al., 1988;
tail length & tail
moment; blinding not
described; analyzed
50 cells per subject.
Recruitment and
selection of
participants not
described.
Referent group
described to be
unexposed
Age comparable
between exposed
and referent; data
analysis by gender;
no evaluation of
smoking
Difference in mean
values between
exposed and
referent, Student's
t-test
Exposed n = 30,
referent n = 15
Cell incubation
period 72 hours;
blinding not
described; no
evaluation of
smoking
(2012) Egypt
Pathology,
histology and
anatomy
laboratories at
a university
Haves et al.
Personal sampling;
cumulative
exposure
estimated using
sampling data and
time-activity data;
continuous area
samples at head
Blood samples
collected in morning
before 1st class and
after 9 weeks; analysis
blinded to exposure
status; O6-
alkylguanine DNA
alkyl-transferase
Recruited
volunteers prior to
beginning of
course; reported
loss to follow-up.
15 students had
some prior
embalming
experience during
lifetime; exposure
to other chemicals
below LOD or very
Change in
individual;
Individual data pre-
and postcourse
AGT activity in
peripheral blood
lymphocytes
depicted in graphs
N = 29
No obvious bias,
small sample size
(1997) (USA)
Panel study, 9
weeks
embalming
course
This document is a draft for review purposes only and does not constitute Agency policy.
A-199 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Related to
Suruda et al.
height over
embalming tables
for short-term
peak
concentrations;
monitored for
other compounds:
glutaraldehyde,
methanol,
isopropyl alcohol,
and phenol
activity in peripheral
blood lymphocytes
(according to Klein
and Oesch, 1990),
expressed as pmol
AGT/mg protein (LOD
0.006 pmol AGT/ mg
protein), blind to
period of sample
(before or after)
low; confounding
not likely
by embalming
experience during
previous 90 days
(yes/ no), ANOVA
adjusting for age,
sex, and smoking.
(1993)
He et al.
Breathing-zone
samples during
dissection;
number, duration
of sampling not
described
Blood collection not
described. Assays
used whole blood.
Cytokinesis-blocked
MN assay, cultured 72
hr, cells processing
(Fenech and
Morlev. 1985),
Recruitment and
selection details
not described.
Demographic data
comparing exposed
and referent
groups were not
provided.
All nonsmokers, age
and sex similar
(data not reported)
Analytic method
not described
Exposed n = 13
Referent n = 10
(# in table
reported as 13)
Deficiencies and
inconsistency in
reporting, small
sample numbers.
(1998) (China)
Prevalence
Anatomy
students
blinding not described
(scored 1,000 cells per
individual), CA
analyzed 100
metaphases, modified
fluorescence-plus-
Giemsa stain; SCE
analyzed 50
metaphases, Giemsa
stain,
Blinding not described
This document is a draft for review purposes only and does not constitute Agency policy.
A-200 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Jakab et al.
Area samples,
records of
measurements
within 1-3 years of
study
8-hr TWA
determined
Venous blood
collection, timing not
stated, peripheral
blood lymphocytes
HPRT gene mutations,
unscheduled DNA
synthesis,
CA and SCE whole
blood samples,
cultures incubated 50
(CA) and 72 (SCE)
hours; CA stain 5%
Giemsa, SCE
fluorescence plus
Giemsa; analyses
blinded, for CA scored
100 metaphases/
subject.
Scored total CA and
types, SCE and high
frequency SCE, total
premature
centromere division
(PCD) and mitoses
with >3 chromosomes
with PCD
Recruitment and
selection of
participants not
described.
Participation rates
not reported.
Referent group
from health-service
staff in same
hospitals
Provided data on
demographic
characteristics; Age
comparable,
Formaldehyde only
group had higher
proportion of
smokers, more
cigarettes/day and
higher proportion
drinkers. Solvents
were ethyl alcohol,
acetone, and xylene
Exposure groups
compared,
student's t-test SCE
stratified by
smoking, CA
frequency analyses
not stratified
HCHO alone
A/= 21; HCHO
and solvents
N = 16; Referent
N = 37
Possible
confounding by
smoking on CA
association not
assessed.
Direction:
potential over-
estimation
(2010)
Hungary
Hospital and
university
pathology
department
Jiang et al.
Personal samples
in breathing zone;
3-5 workers from
each job title, 5
referent workers;
8 hour samples;
Blood lymphocytes;
blinded analysis;
comet assay (DNA
strand breaks),
lymphocytes isolated
within 2 hr after blood
draw, alkaline
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported. 263
Excluded subjects
with recent
exposure to known
mutagenic agents
(x-ray) chronic
conditions
(autoimmune
Ln-transformed
Olive TM and
CBMN frequency
ANOVA differences
by exposure group;
t-test for
differences in
Referent
N= 112
Exposed N = 151
No obvious bias
(2010) (China)
Woodworkers
(prevalence
study)
This document is a draft for review purposes only and does not constitute Agency policy.
A-201 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
calculated 8-hour
TWA
conditions, Singh et
al., 1988; slides
dessicated, shipped to
Beijing, >100 cells/
subject, image
analysis software.
MN: cytokinesis-block
micronucleus assay
(chromosome
damage), scoring
criteria (Fenech et
al.. 2003) 1,000
male workers all
Han Chinese; 151
exposed from two
plywood
industries; 112
referents from a
machine
manufacturing
plant in same town
disease), recent
antibiotic use.
Structured
questionnaire
collected info on
smoking, alcohol,
medical conditions,
occupational history
& house
redecoration in last
year. Evaluated
mean age and
frequency of
smoking and
alcohol by exposure
level.
means. ANCOVA
differences by
years of exposure
among exposed
adjusted for age,
formaldehyde
concentration,
smoking and
alcohol.
binucleated
lymphocytes/ subject
Kitaeva et al.
Exposure
definition by job
task, no
formaldehyde
measurements
MN assay in buccal
mucosal cells, blinding
not described, cell
collection using swab,
smeared onto slides,
stain Feulgen and light
green, analyzed 2,000
cell/ subject. CA in
peripheral blood
(blood from finger),
reported %
metaphases with
aberrations after 72-
hours culture; #
metaphases at 72
hours cultivation was
low (148), observed in
Recruitment and
selection not
described.
Referent group not
defined clearly.
Referents 10 years
younger than
exposed; Stated
that age and
smoking were not
related to MN or CA
frequency, gender
not related among
unexposed, Data
not shown.
Analysis using
Student method
with Freeman-
Tukey
transformation and
results were not
clearly presented
Female Exposed
n = 8
Female Referent
n = 7; Students
n = 12
Small numbers,
reporting
deficiencies for
details of study
design and results,
difficult to
evaluate
(1996)Russia
Translation
Formaldehyde
production and
anatomy lab
workers
This document is a draft for review purposes only and does not constitute Agency policy.
A-202 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
only 8 exposed
workers
(Kurttio et
al.. 1993)
No formaldehyde
measurements;
exposure defined
by task; 5 out of 15
exposed,
considered to be
exposed to
formaldehyde;
referent selected
from same town
employed at
municipal energy
plant, a loading
company, or a
health care center
Venous blood samples
cultured all on same
day; cultured for 48 hr
according to Jantunen
et al., 1986; slides
coded; analyzed 100
metaphases per
subject
Selection of
exposed and
referents not
described;
referents were
employed in other
industries
(potential for dis-
similarities)
All male, matched
on age, data
analysis excluded
one smoker
Structural
aberrations, mean
# per cell by
exposure, Mann-
Whitney U-test (2-
tailed)
Exposed n = 15;
Referent n = 15
5 out of 15
considered
exposed to
formaldehyde; no
formaldehyde-
specific data
analysis
Not informative
Finland
Wood
plywood/
veneer
manufacture
Ladeira et al.
Personal air
sampling, 6-8
hours, estimated
8-hr TWA (NIOSH
method 2541)
Ceiling values for
each task
Cell collection
between 10 am and
noon. Samples coded
and analyzed blinded.
Peripheral blood
lymphocytes,
cytokinesis-block
micronucleus cytome
assay, fresh samples,
cultured for 72 hr,
applied to slides with
cytocentrifuge, May-
Grunwald-Giemsa,
Recruitment and
selection not
described.
Participation rates
not reported.
Excluded history of
cancer, radio or
chemotherapy, use
of therapeutic
drugs, exposure to
diagnostic x-rays in
the past six
months, intake of
Exposed were older,
with lower
proportion of
drinkers and
smokers
Comparisons by
exposure group;
binary logistic
regression and
Mann-Whitney test
Stratified by
categories of age,
gender and
smoking
Exposed n = 56,
referent n = 85
No obvious bias
(2011)
(Portugal)
Histopathology
labs In 6
hospitals
This document is a draft for review purposes only and does not constitute Agency policy.
A-203 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
1,000 binucleated
cells scored/ subject
by 2 readers; buccal
mucosa cells,
collection using
endobrush, smeared
onto slides, stain
Feulgen, 2,000 cells
scored/ subject, 2
readers
vitamins or other
supplements like
folic acid (no one
was excluded)
Lan et al.
(2015) China
Formaldehyde-
melamine resin
production or
use
Bassig et al.
(2016);
related
study, Zhang
et al. (2010)
Personal monitors
for 3 days over
entire shift within
a 3-week period.
Formaldehyde
concentration: 8-h
TWA
Exposed
Median: 1.38 ppm
(1.7 mg/m3)
10th & 90th
percentile: 0.78,
2.61 ppm 0.96, 3.2
mg/m3)
Referent
0.026 ppm (0.032
mg/m3)
10th & 90th
percentile:
0.015, 0.026 ppm
(0.019, 0.032
mg/m3)
Postshift and
overnight peripheral
blood samples.
Metaphase spreads
from colony forming
unit granulocyte
macrophage(CFU-
GM) cultured for 14
days; chromosome-
wide aneuploidy
analysis using
OctoChrome FISH;
scored minimum 150
cells/subject; analysis
blinded to exposure.
Analyzed
aneuploidy among
subset with
scorable
metaphases, high
formaldehyde
among exposed
and existence of
comparable
referents.
Participation rates
exposed 92%,
referent 95%.
Referent from 3
workplaces in same
geographic region
as exposed,
engaged in
manufacturing
with similar
demographic and
SES; excluded
history of cancer,
Referents
frequency-matched
by age (5 yr) and
gender
Personal sampling
of volatile organic
compounds;
concentrations at
background, urinary
benzene at
background and
comparable
between groups
Analyzed using
negative binomial
regression
controlling for age
and gender. Also
evaluated potential
confounding from
current smoking
and alcohol use,
recent infections,
current medication
use, and body mass
index
(Supplemental
tables)
Exposed n = 29;
Referent n = 23
No obvious bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-204 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
LOD: 0.012 ppm
chemotherapy, and
radiotherapy,
previous
occupations with
exposure to
benzene,
butadiene, styrene,
and/or ionizing
radiation.
Lazutka et al.
Industrial hygiene
area
measurements
reported by plant;
carpet plant,
formaldehyde
0.3-1.2 mg/m3,
styrene 0.13-1.4
mg/m3, phenol 0.3
mg/m3;
plasticware plant,
formaldehyde
0.5-0.9 mg/m3,
styrene 4.4-6.2
mg/m3, phenol
0.5-0.75 mg/m3
Peripheral blood
samples; chromosome
aberrations, cells
cultured 72 hr,
differential staining
fluorescence-plus-
Giemsa, CA scored on
coded slides, >100
first mitotic division
cells per subject.
Recruitment and
selection not
described.
Participation rates
not reported.;
Source population
for nonexposed
referents not
described
Nonexposed were
"approximately"
matched to
exposed by age;
males and females,
smokers and
nonsmokers
included;
demographic
information
provided; unable to
distinguish between
formaldehyde and
styrene
ANOVA including
variable for
exposure and age,
no adjustment for
smoking or gender;
CA data
transformed using
average square
root
transformation
Carpet plant,
exposed 38
male, 41 female;
unexposed 64
male, 26 female
Plastic plant,
exposed 34
male, 63 female;
unexposed 64
males, 26
females
Cell incubation
period 72 hours;
unable to
distinguish
between
formaldehyde and
styrene effects
Direction:
potentially
overestimated
(1999)
Lithuania
Carpet and
plastic
manufacturing
Prevalence
study
Lin et al.
Prevalence: Area
samples (2 badges
in each of 5
workplaces with
differing tasks), 8-
hour samples on
two days.
Blood lymphocytes;
blinded analysis;
comet assay (DNA
strand breaks),
alkaline conditions
(pH=13), Olive and
Banath, 2006, lysis 2-
hr for N = 178&over-
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Exposed and
Excluded subjects
with exposure to
known mutagenic
agents in previous 3
months
(radiotherapy &
chemotherapy).
Structured
Natural log-
transformed olive
TM. Prevalence:
ANOVA differences
by exposure group
(control, low and
high), adjusting for
age, sex, smoking,
Referent N = 82
Low N = 58
High N = 38
Referent group
with significant
formaldehyde
exposure,
potential bias
toward null.
(2013)
(China)
Woodworkers
(prevalence
study) 2009
This document is a draft for review purposes only and does not constitute Agency policy.
A-205 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
(cross-shift)
2011
Change over work-
shift: badges in
breathing zone of
2-4 representative
workers
conducting
different job types
(8-hour samples).
Referent group
exposed, mean
0.13 mg/m3
(0.019-0.252)
night for N = 62, 50
lymphocytes/ sample,
image analysis
software; cytokinesis-
block micronucleus
assay, (Fenech,
1993) analyzed 1,000
referent from same
factory.
questionnaire
collected info on
smoking, alcohol,
medical conditions,
occupational
history, and house
redecoration in last
year.
alcohol, # work
years)
Regression for
trend across
exposure level
adjusting same as
above; Poisson
regression for MN
frequencies, linear
regression for
Ln(OTM )
Across-shift:
Paired Wilcoxon
text (MN freq) or
paired t-test (OTM
or DPX); regression
models for trend
with exposure
levels
binucleated cells/
subject, scoring
criteria (Fenech,
1993), (Fenech et
al., 2003); Zhitkovich
and Costa's KCI-SDS
assay (DNA-protein
crosslinks)
Marcon et al.,
2014 Italy
Population
living in
proximity to
chipboard
plants
Modeled outdoor
formaldehyde
concentrations at
residential address
based on data
from 62
monitoring sites in
district; 4 one-
week sampling
periods (2 each in
warm and cold
seasons);
calculated annual
average
Epithelial mucosal
cells using cytology
brush; comet assay,
alkaline conditions, 50
cells per subject; MN
2,000 cells per
subject, according to
Tolbert et al., 1991
Random sample of
participants in
previous survey
(93% of population
in Viadana District)
with children under
12 yrs, Italian
primary language,
and address
information;
invited stratified
random sample in
3 strata of distance
from wood
No adjustment for
indoor
formaldehyde
concentrations; co-
exposure with N02
Linear regression
for tail length, tail
intensity, tail
moment and
binucleated cells;
negative binomial
regression for
micronuclei and
nuclear buds;
models adjusted
for children's sex,
age, nationality,
parents' education,
parents' smoking,
N = 413;
Analysis
included only
complete
datasets for
comet assay,
n = 310 and MN
n = 374
Potential
exposure
misclassification;
no obvious bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-206 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
concentration of
formaldehyde and
N02; estimated at
each address using
ordinary Kriging;
formaldehyde 2.5
+ 0.3 |J.g/m3, N02
16.0 + 3.5 |J.g/m3,
factories (656
remaining in
district since 2006
of 750),
participation 63%,
participation was
not higher in
residents closest to
wood factories;
higher proportion
of nonparticipants
were of foreign
nationality and had
smoking parents
exposure to
tobacco smoking at
home, time with
windows open,
traffic near home,
orthodontic
appliance,
condition of teeth,
person who
collected cell
sample
Musak et al.
Air monitoring
once per year (no
details provided)
Chromosomal
aberration, peripheral
blood lymphocytes,
blinded analysis,
cultured 48 hr, 100
mitoses scored/
subject, 2 scorers
Recruitment and
selection of
participants not
described.
Participation rates
not reported.
Exposed and
referent all
employed in
hospitals
Exposed and
referent
comparable for age,
gender; % smokers
slightly higher in
exposed; analyses
adjusted for age,
gender, job type,
and smoking
Adjusted odds
ratios, Binary
logistic regression
controlling for age,
gender, job type,
and smoking
Exposed
N = 105;
Referent
N = 250
No obvious bias
(2013) Slovakia
Prevalence
study
Pathologists
Orsiere et al.
Personal sampling
near breathing
zone;
Short-term: 15
minutes, Long-
term 8 hours
during typical work
day.
Peripheral
lymphocytes, blood
samples taken preshift
and postshift;
processed within 6 hr,
assays conducted
blinded. Chemi-
luminescence
microplate assay;
Selection &
recruitment of
exposed and
referent not
described,
however
subgroups selected
randomly. Exposed
and referent
Groups similar for
gender, age, %
smokers. No
exposure to other
genotoxic
substances.
Excluded history of
radiotherapy or
chemotherapy and
Differences by
group analyzed
using
nonparametric
Mann-Whitney Li-
test; median DNA
repair across shift
analyzed using
Wilcoxon W-rank
Exposed n = 59;
referent n = 37;
Subgroups
Exposed n = 18;
referent n = 18
No obvious bias.
(2006) (France)
Hospital
pathology labs
(prevalence)
This document is a draft for review purposes only and does not constitute Agency policy.
A-207 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
cytokinesis -blocked
micronucleus assay
(Sari-Minodier et al.,
2002); cultured 72 hr,
smears on slides, stain
5% Giemsa, scoring
criteria (Fenech, 2000)
; 1,000 binucleated
cells/ subject; FISH
with a pan-
centromeric DNA
probe, same operator
scored exposed and
referent blinded
worked in same
institution.
use of therapeutic
drugs that were
known mutagens or
reproductive
toxicants
sum test. Analyzed
binucleated
micronucleated cell
rate (BMCR), and
MN measures using
multivariate
regression
adjusting for
smoking, drinking,
age, and gender.
Pala et al.
Personal samples,
one 8-hour shift;
75% exposed to <
0.026 mg/m3.
Peripheral blood
samples collected at
same time at end of
day; processed within
20 hr; analysis blind to
exposure.
CA, harvested after 48
hr, 100 metaphases/
subject
SCE, cultures
harvested at 72 hr,
analysis of 30 second-
division cells/subject;
MN: modified
cytokinesis-blocked
method, Fenech et al.,
1986; 72 hr
incubation, stain 3%
Selection &
recruitment of
exposed and
referent not
described.
Participation rates
not reported.
Statistical models
adjusted for gender,
age, and smoking
Multivariate
regression models
adjusting for
gender, age, and
smoking; Poisson
model for CA and
MN, SCE log-
normal random
effects model,
comparisons were
low and high
exposure groups,
below and above
26 ng/m3
N = 36
No obvious bias;
only 9 exposed
above 0.026
mg/m3.
(2008) (Italv)
Research
institute lab
(prevalence)
This document is a draft for review purposes only and does not constitute Agency policy.
A-208 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Giemsa, 2,000
cells/subject
Peteffi et al.
Monitoring in 7
sections in facility;
referent
monitoring in 5
areas of university;
breathing zone
8-hr samples
collected on same
day as biological
samples. Urine
samples collected
at end of work day
on 5th day of work;
correlation of
formaldehyde
concentration in
air with urinary
formic acid
concentration, r =
0.626, p <0.001
Peripheral blood
processed within 4 hr.
comet assay, alkaline
conditions according
to Tice et al., 2000;
silver nitrate staining
according to Nadin et
al., 2001; 100 cells/
person read by two
independent
observers (50 cells
each). Blinding not
stated, classified by
visual scoring
according to Anderson
et al., 1994; 5
categories based on
tail migration (0—IV)
and frequency of
damaged cells (sum of
1—IV), damage index
(Pitarque et al., 1999)
Oral mucosa samples
(scraped with
endocervical brush),
micronucleus test,
DNA-specific Feulgen
staining and
counterstaining with
Fast Green according
46 workers in
furniture
manufacturing
facility and
unexposed group
recruited from
employees and
students of local
university with no
history of
occupational
exposure to
potentially
genotoxic agents
or substances
metabolized to
formic acid
Exposed and
referent had
comparable
distributions for
age, smoking, and
alcohol; differed by
gender
Exposed 56.5%
male, referent
33.3% male; no
association of any
biomarkers with
gender (data not
shown)
Nonparametric
tests used because
data were not
normally
distributed.
Exposed and
referent compared
using Mann-
Whitney test;
Exposed n = 46,
referent n = 45
No obvious bias
(2015) Brazil
Furniture
manufacturing
This document is a draft for review purposes only and does not constitute Agency policy.
A-209 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
to (Tolbert et al.,
1992); analyzed
2,000 cells/ person by
2 independent
observers (1,000 ea)
Santovito et
All exposed used
protective
equipment; no
formaldehyde
measurements,
intensity and
frequency likely
highly variable
Peripheral blood
samples, coded,
processed within 2 hr
after collection.
Cultures incubated for
48 hr for CA and 72 hr
for SCE; CA slides
stained with 5%
Giemsa, scored 200
metaphases per
subject, gaps not
scored as CA; SCE 50
metaphases scored
per subject
20 female nurses
from 2 analogous
departments in 2
hospitals; 20
referents from
administrative
departments of
same hospital; all
nonsmokers and
did not consume
alcohol
Accounted for sex,
age, smoking, and
alcohol in design;
referents from
same hospitals
Nurses exposed to
other substances
Mean frequencies
compared,
Wilcoxon test;
regression analysis,
association of age
and exposure
duration on CA and
SCE
Exposed n = 20;
Referent n = 20
Potential for large
degree of
exposure
misclassification
and variation in
intensity of
exposure; bias
toward null; small
sample size
al. (2014)
Italy
Hospital nurses
Santovito et
Personal sampling
near breathing
zone, 8-hour
duration
Venous blood sample
collected at end of
shift, samples coded
and processed within
4 hr, same day
concentration
sampling conducted,
cultured 48 hours;
CA 5% Giemsa stain;
scored 100
metaphases/ subject
Recruitment and
selection of
participants not
described;
participation rates
not reported.
All nonsmokers,
nondrinkers, no
drug use 1 year
prior; no
information on
other exposures
(acetone, ethyl
alcohol, xylene)
Mean % of cells
with aberrations
and frequencies of
aberrations per cell
compared using
Mann-Whitney U
test, 2-tailed.
Generalized linear
models (Poisson
distribution)
adjusting for age,
gender,
polymorphisms,
Exposed n = 20;
Referent n = 16
No obvious bias
Small sample size
al. (2011)
Italy
Pathology
wards
This document is a draft for review purposes only and does not constitute Agency policy.
A-210 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
Cubic spline
regression of mean
% of cells with
aberrations and
frequencies of
aberrations per cell
with number years
exposed and age
Schlink et al.
Personal sampling
Blood samples
Recruitment and
Considered effects
MGMT activity
Exposed N= 41
No obvious bias,
(1999)
Germany
Anatomy
students
near breathing
collected before 1st
participation of
of age, sex,
change compared
Referent N= 10
small sample size
zone once per
class and after days 50
students were not
smoking, and
(U-test, paired
week, sampling
and 111; O6-
described. 41
alcohol
data) within
period not
alkylguanine DNA
students from one
categories of sex,
reported,
formaldehyde
exposed, Mean ±
SD, 0.2 ±0.05
mg/m3, 0.14-0.3
mg/m3
alkyl-transferase
activity in peripheral
blood lymphocytes
(modification of Klein
and Oesch, 1990),
expressed as fmol
MGMT/106 cells (LOD
1 fmol MGMT/ 106
cells), blind to period
of sample (before or
after)
university course,
16 students from a
different university
course, and 10
unexposed
students
smoking, allergy,
and alcohol; as well
as between groups
(Wilcoxon, Mann
and Whitney Li-
test)
Shaham et
Personal and
Peripheral
Selection &
Exposed and
Analyses by ANOVA
Exposed DPX:
Low sample
al. (1997)
(Israel)
anatomy/
pathology
departments
(prevalence)
"field" samples,
lymphocytes; DPX, K-
recruitment of
referent matched
adjusting for
N= 12 SCE:
numbers; no
duration 15
SDS method; double
exposed and
by age (matching
smoking; difference
N = 13 Referent
obvious bias.
minutes, multiple
blinded. SCE at 72
referent not
protocol not
in means, t-test;
DPX: N = 8
times during work
hours, mean of 30
described.
described). No
linear regression
SCE: W = 20
day (# not
cells/ individual,
Participation rates
exposure to other
for DPX levels or
reported).
blinding not described
not reported.
mutagens or
means SCE per
Referent group
substances known
to cause DPX in
chromosome by
This document is a draft for review purposes only and does not constitute Agency policy.
A-211 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
also reported in
Shaham et al.,
1996
worked at same
institution.
either exposed or
referent.
years of exposure
to formaldehyde
(Shaham et
al.. 2002)
Personal and area
samples, sampling
at different points
in work day,
sampling duration
averaged 15 min
SCE in peripheral
lymphocytes, blood
samples collected at
same time in morning;
blinding not
described, stain
fluorescence plus 5%
Giemsa, scored 30-32
cells/subject
Recruitment and
selection of
participants not
described.
Referent group
from
administrative
sections of same
hospitals
Authors presented
demographic data.
Exposed were
higher proportion
female, European/
American,
education >12yr,
and lower
proportion
smokers. No
exposures to other
chemicals linked to
SCE. Confounding
addressed in
analysis
Mean # SCEs per
chromosome and
proportion of high
frequency cells
compared between
exposed and
referent.
Difference between
means assessed
using ANOVA
(unbalanced
design) adjusting
for age, gender,
smoking, origin and
education years
Exposed n = 90;
Referent n = 52
No obvious bias
Israel
Hospital
pathology labs
Shaham et
Personal and
"field" samples,
duration 15
minutes, multiple
times during work
day (# not
reported).
Peripheral
lymphocytes; DPX,
same protocol as
Shaham et al.
Selection &
recruitment of
exposed and
referent not
described.
Exposed and
referent worked in
same institution.
Adjustment for age,
sex, smoking, origin,
and years of
education in
analysis. No
exposure to other
mutagens or
substances known
to cause DPX in
either exposed or
referent.
Analyses:
comparisons of
mean DPX adjusted
for sex, smoking,
age, origin, and
years education.
Comparison of
mean DPX by low
and high
formaldehyde
levels and by
duration of
exposure, Mann-
Whitney test
Exposed
N = 186;
Referent n = 213
No obvious bias.
al. (2003)
(Israel)
14 hospital
pathology
departments
(prevalence)
(1997); SCE;
pantropic p53
This document is a draft for review purposes only and does not constitute Agency policy.
A-212 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
(Souza and
Devi, 2014)
No formaldehyde
measurements
reported.
Total MN/1,000 cells
peripheral
lymphocytes. Assays
conducted blinded.
Cytokinesis -blocked
micronucleus assay
(Costa et al.,
2008); stain 4%
Recruitment and
selection of
participants not
described.
Participation rates
not reported.
Provided
characteristics of
exposure groups
(see Table 1). All
male, age
comparable, higher
prevalence smokers
in exposed.
Adjustment in
analysis. Excluded
frequent exposure
to x-rays or other
radiation, worked in
paint or pesticide
industries or history
of chemotherapy.
Frequency MN
compared by
exposure group
using Student's
t-test, and by
duration of
employment using
Pearson's
correlation.
Exposure and
smoking evaluated
together using two-
way ANOVA.
Exposed N = 30
Referent N = 30
No obvious bias
India
Prevalence
study Anatomy
Dept
(embalming)
Giemsa, scoring
criteria (Fenech,
2000), 1,000
binucleated cells/
subject. Frequency
MN compared by
exposure group using
Student's t-test, and
by duration of
employment using
Pearson's correlation.
Speit et al.
Generation using
para-
formaldehyde; 10
consecutive days,
5 groups of 3-6
persons in
chamber, 4 hour
exposures, some
exposures masked
with ethyl acetate,
3 15-min exercise
sessions during
exposure;
MN in buccal mucosal
cells—1 week before
start, at time=0, after
end of exposure, and
1, 2, and 3 weeks after
end of exposure; cells
collected with metal
spatula, smeared onto
slides, blinded analysis
at end of study by one
person, stain DAPI/
propidium iodide,
2,000 cells/ subject
Excluded severe
allergy, skin or
airways disease,
acute infection,
current smoking or
within last 3 years,
contact lenses or
glasses, > 50 g
alcohol per day,
present use of
psychotropic
agents, exposure
to ionizing
Within person
comparison
Post exposure
compared to
preexposure using
Wilcoxon ranked
sum test
N = 21
No obvious bias.
(2007a)
(Germany)
Controlled
human
exposure study
This document is a draft for review purposes only and does not constitute Agency policy.
A-213 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
randomized order
of concentration,
double blinded
radiation, or
cytostatic drugs
during the last 6
months
Suruda et al.
Personal sampling
for 121 of 144
embalmings;
cumulative
exposure
estimated using
sampling data and
time-activity data;
Continuous area
samples at head
height over
embalming tables
for short-term
peak
concentrations;
monitored for
other compounds:
glutaraldehyde,
methanol,
isopropyl alcohol,
and phenol
Nasal mucosa cells,
oral mucosa cells,
blood samples
collected in morning
before 1st class and
after 9 weeks;
processed on same
day, analysis of slides
blinded to exposure
status; pre- and
postslides from each
subject stained at
same time and read
together by one
reader, conducted a
blinded 10% recount
of slides; MN assay
buccal and nasal cells
Stich et al. (1982),
collected with
cytopathology
brushes, slides
prepared with
cytocentrifuge, stain
Feulgen/ Fast Green,
1,500 cell/ subject;
MN lymphocytes
(Fenech and
Recruited
volunteers prior to
beginning of
course; reported
loss to follow-up.
Excluded one
student with many
embalmings in
previous 90 days, &
one students who
chewed tobacco
during study
21 students had
some prior
embalming
experience during
lifetime; exposure
to other chemicals
below LOD or very
low, confounding
not likely
Change in
individual;
difference in mean
pre- and
postexposure,
matched Student's
t-test (SCE) or
Wilcoxon sign-rank
test (micronuclei);
Change with
cumulative
exposure
spearman's rank
correlation
coefficient & linear
regression (if
residuals were
normally
distributed)
N = 29
No obvious bias
(1993) (USA)
Panel study, 85
days
Embalming
course
This document is a draft for review purposes only and does not constitute Agency policy.
A-214 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Morlev, 1985), stain
Feulgen 2,000 cells/
subject;
SCE 50 s division
metaphases scored/
subject
Suskov and
Sazonova
(1982)USSR
Phenol-
formaldehyde
resin
production
Area samples, #
and duration not
reported
Cytogenetic analysis in
peripheral
lymphocytes;
Chromosomal
aberrations, blinding
not described,
Buckton and Evans
cytogenetic method,
1973
Recruitment and
selection not
described.
Average age in
exposed 39.1 yr,
referent 34 yr.
Matched for
gender, smoking,
alcohol, and
medication (data
not shown)
Compared
chromosome
aberration
frequency by
exposure group,
chi-square
Exposed n = 31;
Referent n = 74
Brief report,
minimal detail of
methods
Thomson et
al. (1984)
Great Britain
Pathology lab
Sampling in
breathing zone; 26
samples taken for
the duration of the
task involving
formaldehyde
exposure, over 1-3
months, sample
duration not
reported,
calculated TWA
Measured peaks in
breathing zone on
one day for
different tasks
CA frequency, stain
fluorescence plus
Giemsa technique
(Perrv and Wolff,
1974), cells
harvested 48 hr, slides
coded and scored 100
1st division
metaphases/ subject;
SCE frequency, cells
harvested 72 hr, 50
cells/subject; blinding
not reported
All exposed worked
in same laboratory;
characteristics of
referent not
provided.
Obtained smoking
histories
Data analysis not
described
Exposed n = 6;
referent n = 5
Reporting of study
methods and
group
characteristics not
adequate; low
sample numbers
This document is a draft for review purposes only and does not constitute Agency policy.
A-215 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Titenko-
See Suruda et al.
Buccal cells, Scored
previously unstained
and unanalyzed slides.
New method: FISH
with a centromeric
probe—differentiates
between clastogenic
vs aneuploidogenic
mechanism (total MN,
MN- and MN+);
<1,500 cells scored for
14 of 35 subjects;
scored pre- and
postexposure slides at
same time, blinded.
Frequency calculated
by dividing # cells with
MN by total # cells
counted, multiplying
by 1,000.
78% of preexposure
slides and 76% of
postexposure slides
were scorable; 10% of
slides were rescored
Subjects with
missing MN data
were compared to
those with
complete data by
Student's t-test;
comparable for
age, smoking, and
mean exposure
Change in
individual.
Exposure to other
chemicals below
LOD or very low,
confounding not
likely
Change in total
MN, MN- and MN+
frequency (per
1000 cells) and
change in mean
MN. Excluded
subjects with <500
epithelial cells
available for
analysis.
Difference scores
evaluated using
Wilcoxon sign-rank
test.
Association with
both formaldehyde
exposure metrics
via Spearman non-
parametric
correlation
coefficient, two-
sided p-values
Complete MN
data from
buccal mucosa,
n = 19
Complete MN
data from nasal
mucosa, n = 13
No obvious bias
Holland et al.
(1993)
(1996) Same
Calculated two
exposure periods:
1) Lagged 7-10
days before last
sampling to
account for lag in
development of
MN
2) 90-day
cumulative
subject as
Suruda et al.
(1993)
(USA)
Panel study, 90
days
Embalming
course
Vasuveda and
Anand,1996
India
Medical
student lab
<1 ppm, no data
reported to
support assertion
Peripheral blood
lymphocytes,
frequency of aberrant
metaphases; cell
culture 72 hr, Giemsa
staining, blinding not
reported
Recruitment and
selection of
participants not
described. No
demographic
information
provided.
Stated that
participants had
received no or
insignificant
radiation
treatments (no data
reported); exposed
and referents
Data analysis not
described
Exposed n = 30;
referent n = 30
Reporting of
methods, design
and results not
adequate to
evaluate; cell
incubation 72 hr
This document is a draft for review purposes only and does not constitute Agency policy.
A-216 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
matched by age, no
other potential
confounders
evaluated
Viegas et al.
Personal air
Buccal mucosa
Recruitment and
Presented
Correlation
Exposed,
No obvious bias
(2010)
(Portugal)
Formaldehyde
& resin
production,
pathology/
anatomy lab
workers
sampling, (N = 2 in
Peripheral blood
selection not
comparisons for
evaluated using
Produc-tion
factory, N = 29 in
lymphocytes, sample
described.
gender, age, and
Pearson or
n = 30, Lab
labs) 6-8 hours,
collection between 10
Participation rates
smoking.
Spearman
workers n = 50,
estimated 8-hr
am & noon. Blinded
not reported.
Difference by
correlation test
Referent n = 85
TWA (NIOSH
coding and analysis,
gender (higher
depending on
method 2541).
buccal cell MN cell
prevalence males in
distribution
Ceiling values for
collection using
exposed); genotoxic
each task
endobrush, smeared
endpoints were not
onto slides, Feulgen
associated with
Also discussed
in (Viegas et
al.. 2013)
stain, 2,000 cells
scored/ subject by 4
observers, scoring
criteria (Tolbert et
al., 1992), peripheral
smoking or gender,
and only slightly
with age
lymphocytes, samples
processed within 6 hr,
cultured for 72 hr,
applied to slides with
cytocentrifuge, stain
May-Grunwald-
Giemsa, 1,000
binucleated cells
scored/ subject
Wang et al.
Routine
CBMN according to
Recruitment and
Mean age and
MN frequency
Exposed
No obvious bias
(2019)
formaldehyde
Fenech et al.
selection of
frequency of
compared using
n = 100
Shanghai, China
monitoring by
factory with
(Fenech, 2000),
(Fenech, 1993).
participants not
described;
smoking and
alcohol use were
Poisson regression
and frequency ratio
Unexposed n =
100
This document is a draft for review purposes only and does not constitute Agency policy.
A-217 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
Exposure
participant
Consideration of
Analysis and
Reference
measures and
Outcome
selection and
likely
completeness of
and setting
range
classification
comparability
confounding
results
Study size
Comment
Chemical
sampling site
Blinded analysis.
participation rates
slightly higher in
(FR) as effect
production
selection using
Venous peripheral
not reported. 100
exposed. Work
estimate. Exposure
China national
blood cultured for 44
male workers
duration was higher
was analyzed with
standard for
hr, Cytochalasin-B
exposed to
in exposed. Age,
quartiles for
hazardous
added to cultures,
formaldehyde > 1
smoking status and
cumulatiave dose
substances air
cells harvested 28
year through 4
alcohol use were
and FA-HSA
sampling in the
hours later, air dried
work processes
adjusted in
concentration.
workplace.
slides stained with
(i.e., production
statistical models.
Cumulative dose
Cumulative dose
Giemsa, MN
examination, glue
(mg/m3):
determined for
dectected at 400x
spraying, coating
0.01-0.06
each worker (C x
with confirmation at
and workplace
0.06-0.125
T). C = geometric
l,000x. 1,000
inspection).
0.125-0.9
mean of
binucleated cells
Demographic
0.9-3.75
concentration for a
scored/ subject
information,
year at a sampling
smoking and
site, T = years.
alcohol, medical
Serum
and occupational
formaldehyde-
history (job types
albumin adducts
and # years)
(FA-HSA)
collected by
quantified in
questionnaire.
fasting venous
Unexposed group
peripheral blood.
(n = 100 males)
Geometric mean
from the logistics
range (mg/m3):
workshop in same
Exposed:
factory age
0.06-0.25
matched (likely
Unexposed: 0.01
frequency matched
since rates were
different)
Yager et al.
Area samples
Whole blood cultures;
Recruitment and
All nonsmokers,
Paired t-test of
N = 8
No obvious bias
(1986) USA
randomly
distributed
stain fluorescence
plus Giemsa
selection not
described.
7 female
before and after
samples
This document is a draft for review purposes only and does not constitute Agency policy.
A-218 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Anatomy
course, 10
weeks
(N = 13, 1-4/
week); breathing
zone samples on
30 individuals at
15 tables (N = 35,
2-8/week), mean
sampling duration
18 minutes
technique, Mean SCEs
per cell in peripheral
lymphocytes; before
and after samples
coded and
randomized together
for analysis, scored 80
cells/subject
Vargova et
8-hour sampling
duration in
breathing zone
CA frequency,
peripheral
lymphocytes, Giemsa
staining, cells
harvested 48 hr, 100
cells/ subject.
Blinding not
described.
CA frequency in both
exposed and referent
was higher than range
considered normal
Recruitment and
selection of
participants not
described;
participation rates
not reported.
Referents were
matched to
exposed (did not
report what
matching
parameters were),
no info on subject
characteristics was
reported
Authors stated
questionnaire data
suggested that
factors such as
smoking and
alcohol were
different between
exposed and
referent; analyses
were not adjusted.
Exposed and
referent compared
using student's
t-test and arcsin-sq
rt transformation
test
Exposed n = 20
(or 25?);
Referent n = 19
Reporting of study
methods and
group
characteristics not
adequate; #
exposed in text
did not match #
exposed in table II
in the paper. Lack
of adjustment for
confounding, bias
toward null
al. (1992)
Czechoslovakia
Woodworking
Ye et al.
Sampling
according to
NIOSH method;
Referent n = 6;
Waiters n = 18;
MN in nasal mucosa,
cell collection using
swab, cells smeared
onto slide, stain
Wright's, scoring
Recruitment and
selection not
described.
Included:
nonsmokers, no
Waiters and
workers older than
referent, % male
52% in referent,
25% in workers,
Analysis using one-
way ANOVA and
tested for multiple
comparisons. Data
presented in
Workers n = 18;
waiters n = 16;
referent n = 23
Possible bias away
from null; expect
higher frequency
of MN in older
(2005) (China,
1992)
This document is a draft for review purposes only and does not constitute Agency policy.
A-219 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Formaldehyde
exposure in
factory or
indoor air from
building
materials
Workers n = 36
criteria (Sarto et al.,
1987), per 3000 cells,
medicines for 3
weeks prior and
during study, no x-
ray history for 6
months prior, no
drug use;
comparison groups
were from
different sources:
industrial exposed,
wait staff (indoor
air exposed), and
unexposed student
volunteers
61% in wait staff; all
Han Chinese; no
adjustment for age
or gender in
analyses.
figures and values
estimated from
graph by EPA.
individuals. Small
sample numbers.
blinding not stated,
results reexamined by
another trained staff.
SCE in peripheral
lymphocytes, time of
sample not stated;
stain Giemsa solution,
analysis blinded, 30
M2 lymphocytes
analyzed/subject.
Ying et al.
NIOSH (1977)
method; 3-hr TWA
and peaks; sample
duration, number
and frequency not
described
Nasal mucosa cells,
oral mucosa cells,
blood samples
collected before 1st
class and after last
class; analysis of slides
by one blinded
observer with
reexamination by
another, nasal and
buccal cells collected
with swab, smeared
onto slides, MN Nasal
and Buccal cells,
Wright's stain, scored
4,000 cells/ subject;
MN blood
lymphocytes, stain 4%
Giemsa, scored mean
Included
nonsmokers,
students living in
dorms, disease-
free & no
medications prior 3
weeks, no x-ray
history prior 6
months
Mean age 18.8 ± 1.0
yr, all Han
nationality, all lived
in dorms, all
nonsmokers
Change in
individual over
time; paired t-tests
N = 25
No obvious bias,
small sample size
(1997); (Yine
et al.. 1999)
(China)
Panel study,
8-week class
Anatomy
students
This document is a draft for review purposes only and does not constitute Agency policy.
A-220 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
of 2,870-3,167 cells/
subject; MN scoring
criteria (Sarto et al.,
1987), SCE and LTR
(Zhao and Ying, 1994):
30 M2 lymphocytes
per subject analyzed
blind to exposure
Zendehdel et
al. (2017)
Iran
Melamine
dinnerware
manufacturing
Related
publication:
Zendehdel et
al. (2018)
Personal air
sampling, NIOSH
method 3500,
whole shift for
each worker.
Median time
weighted average
in three
workshops,
0.086 mg/m3;
range, 0.02-0.22
mg/m3; authors
state that 2/3 of
sample were
exposed to < 0.1
mg/m3
Comet assay, alkaline
conditions, according
to Tice et al., 2000.
Blood samples
collected same day as
air sampling; blinding
not described;
minimum of 50
randomly selected
cells per sample; tail
moment and Olive
moment
Workers in 3
melamine
dinnerware
manufacturing
workshops (n=49)
and referents
matched by age
and sex (n=34) who
worked in food
industries, #
smokers higher in
referent (26%
versus 16%), >90%
male. Recruitment
and participation
were not
described.
Data in Table 1 of
paper supported
comparability of
age, sex, and #
smokers in exposed
and referent
groups.
Normal distribution
assessed using
Kolmogorov-
Smirnov test.
Difference in mean
tested using
Student t-test or
Mann-Whitney test
Exposed
N = 49; Referent
N = 34
No obvious bias
blinding not
described;
Zhang et al.
(2010) China
Formaldehyde-
melamine resin
production or
use
Personal sampling
for full shift (>240
min) on 3 working
days over 3 weeks.
Exposed: at least 2
samples per
individual;
Postshift and
overnight peripheral
blood samples;
analysis blinded to
exposure.
Metaphase spreads
from cultured colony
Participation rates
exposed 92%,
referent 95%.
Referent from 3
workplaces in same
geographic region
as exposed,
Referents
frequency-matched
by age (5 yr) and
gender
Analyzed using
negative binomial
regression
(exposed compared
to unexposed)
controlling for age,
High N= 10
Low N = 12
Small sample
numbers, no
obvious bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-221 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
and setting
Exposure
measures and
range
Outcome
classification
Consideration of
participant
selection and
comparability
Consideration of
likely
confounding
Analysis and
completeness of
results
Study size
Comment
Related
publications:
Bassig et al.
(2016);
Gentry et al.
(2013);
(Mundt et
al.. 2017)
Reanalyses
Referent: Sampling
in subgroup on
one day.
Evaluated for
other known or
suspected
leukemogens
(benzene, phenol,
chlorinated
solvents), found
none. Analysis
blinded.
forming unit
granulocyte
macrophage(CFU-
GM); identified loss of
chromosome 7 and
gain of chromosome 8
using FISH
engaged in
manufacturing
with similar
demographic and
SES; excluded
history of cancer,
chemotherapy, and
radiotherapy;
previous
occupations with
exposure to
benzene,
butadiene, styrene
and/or ionizing
radiation.
gender, and
smoking
Mundt et al.
presented
individual data in
graphs for
chromosome 7 and
chromosome 8,
noting smoking
status and whether
150 or more cells
were evaluated.
Gentry et al.
reported that < 150
cells per individual
were analyzed for
several subjects.
Not expected to be
different between
exposed and
unexposed, impact
likely to increase
variability and
attenuate
association
This document is a draft for review purposes only and does not constitute Agency policy.
A-222 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
Summary Table by Genotoxicity Endpoint
A text summary of the available genotoxicity data that emphasizes genotoxicity studies
incorporating inhalation formaldehyde exposure and related experiments (i.e., given the known
toxicokinetics of inhaled formaldehyde) is provided in Section 1.2.5 (Evidence on Mode of Action
for Upper Respiratory Tract Cancers). The table below provides a summary of the most relevant
data organized by genotoxicity endpoint, as compared to the organization by test system in the
previous sections. In addition, when possible, this table separates the summary into investigations
of respiratory- versus nonrespiratory-related tissues or systems. Thus, observations of
genotoxicity in the upper respiratory tract (URT) and in peripheral blood lymphocytes (PBLs)
following inhalation exposure or in related in vitro systems are presented in Table 27 in order of
their importance and relevance to cancer risk beginning with gene mutations, DPXs and DDCs, DNA
adducts, CAs, MN, DNA strand breaks, SCE, and other effects. Overall, the evidence supports the
conclusion that formaldehyde is genotoxic. Particular weight is placed on the following
observations:
1) Consistent observations of mutations in exposed rodents and various in vitro systems;
2) Observations of CAs, MNs, and SSBs in exposed humans across a range of studies,
occupations, and exposure scenarios, with supporting, similar findings in exposed rodents
and in vitro systems; and
3) Consistent observations of DPX detected in multiple experimental systems, showing a
concentration-dependent increase, and concordance of DPX distribution with sites of
tumors in the nose.
This document is a draft for review purposes only and does not constitute Agency policy.
A-223 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-27. Genotoxicity summary table
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
Gene Mutations
Respiratory tract tissues
or in vitro systems
+(1/2) In vivo, rodent (inhalation); +
1/1 chronic; 0/1 subchronic studies
+ (5/5) In vitro, human cell lines,
acute studies
+(8/10) In vitro, rodent cell lines,
acute studies
+(13/17) Nonmammalian systems
In vivo rodent studies analyzed SCCs
from a chronic study and non-
neoplastic nasal mucosa from a
subchronic study at 18.45 mg/m3
All in vitro studies assume MeOH co-
exposure; cellular sources both POE
and systemic sites
Negative in vitro rodent data for HPRT;
+ results include colony formation and
mutation frequency
Mutations induced by
formaldehyde across a range
of in vitro systems. Mutations
observed in SSC in nasal
tissues of exposed rodents at
18.45mg/m3 in one chronic
inhalation study.
Observation of gene
mutations in nasal SSC in
one chronic-duration rodent
study (which only tested
high formaldehyde levels),
with confirmatory evidence
from in vitro test systems
Other tissues
+(1/2) in vivo, rodent (inhalation);
dominant lethal studies
+(1/2) in vivo, rodent (i.p.);
dominant lethal mutation studies
Formalin inhalation exposure at 200
mg/m3 prevents interpretation;
another inhalation study at 1.5 mg/m3
was equivocal
i.p. exposure with MeOH co-exposure
caused + DLM in rats (0.125 mg/kg),
but not in mice (20 mg/kg) at much
higher levels
Results are interpreted as
equivocal; the available studies
do not provide evidence of
mutations in other tissues
across several species. No
mutations in subchronic-
duration rodent study. No
studies of exposed humans
or primates.
Chromosomal
aberrations (CA)
Respiratory tract tissues
or in vitro systems
+(1/1) in vivo, rodent (inhalation):
short term study
+(4/4) In vitro, human cells/cell
lines, acute studies
+(5/6) In vitro, rodent cell lines,
acute studies
In vivo rat study at 18.45 mg/m3 with
4-wk exposure
In vitro studies assume co-exposure to
MeOH; cell sources both POE and
systemic sites
1 equivocal CA study in a rodent cell
line
CAs were observed in the only
in vivo rodent study, which is
supported by positive results
in human and rodent cells in
vitro.
Evidence from exposed
humans across several
different occupations is
consistent with the
induction of CAs. These
results are supported by
observations of CAs in the
only available in vivo rodent
study (4 weeks at high
levels), which was
consistent with findings
from multiple in vitro
studies of human and
rodent cells lines
This document is a draft for review purposes only and does not constitute Agency policy.
A-224 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
Other tissues
+(11/16) in vivo, human (inhalation):
PBLs
+(1/5) in vivo, rodent (inhalation):
short term studies
+(2/2) in vivo, rodent (gavage, p.o.):
acute studies
+(1/4) in vivo, rodent (i.p.): acute or
short term studies
In humans, half + CAs were observed in
pathologists and half among industrial
workers; often, these studies involved
relatively higher formaldehyde
exposure levels (e.g., average >0.2
mg/m3) and longer employment
duration (e.g., average >10 yr)
The only positive rodent inhalation
study involved MeOH co-exposure*;
4 studies used PFA
Oral exposure in rats and mice involved
MeOH co-exposure, although 1 study
indicated it takes >10x MeOH to cause
a similar level of CAs
The + i.p. study was in rat bone
marrow cells after 4-wk
exposure; - studies were acute, mice
studies
Most of the human studies
interpreted with higher
confidence observed increased
CA in PBLs; Lower exposure
levels may explain null
findings.
Rodent results are interpreted
as equivocal. The rodent
studies do not provide
evidence that CAs are induced
in other tissues; however, the
data suggest the possibility
that rats might be more
sensitive and that exposure
duration is important.
Micronuclei (MN)
Respiratory tract tissues
or in vitro systems
+(11/13) in vivo, human (inhalation);
+(0/1) in vivo, rodent (inhalation);
short term study
+(5/5) In vitro, human cell line;
acute study
+(4/4) in vitro, rodent cell lines;
acute studies
+(1/3) nonmammalian studies
MN reported in buccal and nasal cells,
occupational (average >0.5 mg/m3),
anatomy or embalming courses
(average >0.5 mg/m3 with intermittent
peaks). No increase after 5-10 days in
2 controlled human exposure studies,
In vivo rat study at 18.45 mg/m3for
4 wk (in BAL)
MN observed in primary human blood
cultures, and in 3 in vitro rodent
studies with no MeOH co-exposure;
remaining cell studies assume MeOH;
cellular sources both POE and systemic
sites
Consistently increased
frequency of MN or related
endpoint in buccal and/ or
nasal cells of exposed
individuals
Consistent evidence of MN
across a range of in vitro
mammalian cells, but not in a
short term rodent inhalation
study.
Available evidence suggests
increased MN levels
associated with cumulative
exposure; the pattern of
chromosomal loss
(monocentromeric and
multi-centromeric
micronuclei) was consistent
with aneuploidy in exposed
individuals
This document is a draft for review purposes only and does not constitute Agency policy.
A-225 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
Other tissues
+(11/16) in vivo, human (inhalation)
PBLs,
+(1/2) in vivo, rodent (inhalation);
short-term studies
+(1/5) in vivo, rodent (i.p., i.v., p.o.
or gavage); acute studies
MN reported in PBLs of workers from
plywood and formaldehyde production
industry, and pathology, anatomy, and
mortuary lab students, at exposure
concentrations of 0.1-0.5 mg /m3. Null
results in studies with low sensitivity.
No increase after 5 days in controlled
human exposure study. Prevalence
increases with longer exposure
duration.
In rodents, MN were in bone marrow
erythrocytes at 12.8 mg/m3with 10-wk
exposure, but not in peripheral blood
at 18.45mg/m3 with 4-wk exposure.
The + non-inhalation study was an oral
rat study of gastric epithelial cells;
all - studies were in mice
Most of a large set of studies
that measured MN in PBLs
reported increased levels
among exposed participants
working in diverse exposure
settings and in several
countries.
The two rodent inhalation
studies suggest the possibility
that MN induction may require
longer exposure duration, but
results were mixed; data
suggest the possibility that rats
might be more sensitive.
Respiratory tract tissues
or in vitro systems
+(1/3) In vitro, human cell lines;
short-term studies
+(1/3) in vitro, rodent cell lines;
short-term studies
All negative in vitro studies have co-
exposure with MeOH
Inconsistent results from in
vitro human or rodent cell
lines; Methanol co-exposure is
likely to influence the
aneuploidy in cultured cells
Chromosome aneuploidies
are consistent with study
findings of CA and mono-
centromeric and
multicentromeric
micronuclei in PBLs of
exposed humans
Aneuploidy
Other tissues
+(3/4) in vivo, human (inhalation)
+(1/3) in vitro, rodent cell lines
+(1/3) in vitro, human cell lines
An occupational study in humans
reported monosomy 7 and trisomy 8 in
cultured CFU-GM colony cells from
peripheral blood. Analysis of same
cohort with bigger sample size
detected aneuploidy in several
chromosomes.
Two in vitro studies each from rodent
and human cell lines used MeOH-free
HCHO, one positive study in human
cells has co-exposure with MeOH.
Significant increase in
chromosome aneuploidy in
cultured CFU-GM colony cells
among subset of highly
exposed workers compared to
matched controls
This document is a draft for review purposes only and does not constitute Agency policy.
A-226 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
DNA adducts
Respiratory tissues or in
vitro systems*
+(2/2) in monkeys (inhalation) hm-
DNA adducts
+(3/4) in rats (inhalation) hm-DNA
adducts
+(2/2) in vitro human cell lines, hm-
DNA adducts
+(1/1) in vitro rodent cell lines, hm-
DNA adducts
+(10/10) in cell-free systems, hm-
DNA adducts
No in vivo studies in humans showing
hm-DNA adducts with a direct
exposure to formaldehyde.
Detectable hm-DNA adducts in all nasal
passages, but not in lungs of rats.
High endogenous hm-DNA adduct
levels rats and monkeys, but monkeys
> rats
All tissues in nasal passages
demonstrated hm-DNA
adducts except lung tissue of
rodents.
Endogenous levels of hm-DNA
adducts are very high in both
rats and monkeys compared to
exogenous hm-DNA adducts.
Monkeys have much higher
endogenous hm-DNA adduct
levels compared to rats.
Formaldehyde readily forms
hm-DNA adducts in tissues
at POE. However, available
evidence does not show
their formation in distal
tissues.
Other tissues
+(1/1) in vivo, human, MiG adduct
+(0/2) in vivo, monkeys (inhalation),
acute studies
+(0/2) in vivo, rodent (inhalation),
acute studies
One study reported MiG adducts in
peripheral blood of pathologists,
uncertainties with regard to site of
DNA interactions. hm-DNA adducts
were not found in distal tissues of
exposed monkeys or rodents
Absence of hm-DNA adducts in
distal tissues suggest lack of
formaldehyde transport to
distal sites.
Limited evidence of
formaldehyde-induced
oxidative DNA damage.
DDC
Respiratory tissues or in
vitro systems*
+(1/1) in vivo, rat (inhalation), acute
study
+(3/3) in vitro, cell-free systems
Only one in vivo study reports DDC.
But DDC are unstable and could be
generated as an artifact.
Limited evidence of DDC
formation by formaldehyde in
vivo.
Limited evidence that
formaldehyde inhalation
results in DDC although
artifacts were not ruled out.
Other tissues
+(0/1) in vivo monkey (inhalation)
short-term study
+(0/1) in vivo rat (inhalation) short-
term study
DDC were not detectable in distal
tissues.
DDC have not been detected in
distal tissues
DNA-Protein
Crosslinks
Respiratory tissues or in
vitro systems*
+(1/1) in vivo, monkeys (inhalation),
acute study
+(7/11) in vivo, rodents (inhalation),
acute studies
+(30/30), in vitro, human cell lines,
acute studies
+(21/21) in vitro, rodent cell lines,
acute studies
+(3/3) nonmammalian systems
+(4/4) cell-free systems
Concentration-dependent increase in
DPX in rodents (0.37-12.1 mg/m3) and
monkeys (0.86-7.37 mg/m3); DPX
demonstrated in nasal mucosa of rats
but absent from olfactory mucosa and
lung; a negative study in BAL cells used
formalin vapors
Consistent evidence of DPX
across multiple test systems
(two species in vivo, different
cell lines, nonmammalian and
cell-free test systems)
Anatomical distribution of
DPX in rats corresponds to
sites of tumor incidence, cell
proliferation, and
cytotoxicity in the nose.
However, no mechanism is
identified for DPX formation
in PBLs of occupationally
exposed individuals.
This document is a draft for review purposes only and does not constitute Agency policy.
A-227 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
Other tissues
+(2/3) in vivo, human (inhalation)
PBLs
+(4/8) in vivo, rodent (inhalation)
Occupational settings, one null study
of plywood workers had low sensitivity
(referent group had high exposure), no
difference in prevalence by exposure
group, but increase in DPX was
observed over 8-hour shift.
Positive rodent studies have co-
exposure with MeOH.
In vivo human studies show
exposure duration-dependent
increase in DPX in PBLs, but
animal in vivo studies are
confounded by MeOH
coexposure.
DNA strand
breaks
Respiratory tissues or in
vitro systems*
+(1/1) in vivo, rodent (inhalation),
short-term study
+(10/12) in vitro, human cells, acute
studies
+(3/7), in vitro, rodent cells/cell
lines, acute studies
+(4/4) nonmammalian systems
Only one in vivo study and several cell
culture studies reports SSB formation,
but most of these studies have co-
exposure with MeOH.
Human cells were more sensitive to
SSB formation by HCHO exposure
(0.005-0.8 mM)
Excision-repair deficient yeasts were
more sensitive compared to repair-
proficient strains.
Single strand breaks in rat
study were positively
associated with concentration.
Some evidence for SSB with
dose-response in respiratory
tissues from an inhalation
study in rats, and consistent
evidence in PBLs from
several studies of human
exposure and from rodent
studies
Other tissues
+(8/9) in vivo, human (inhalation)
PBLs,
+(3/4) in vivo, rodent (inhalation),
short-term studies
Exposure settings were occupational
with means > 0.2 mg/m3, 1 controlled
human exposure study (4-hour
duration). Categorical analysis by one
study showed exposure-response trend
beginning at 2nd quintile (mean 0.14
mg/m3) Positive rodent in vivo studies
have co-exposure with MeOH.
Consistent evidence of SSB
formation in both human and
rodent in vivo studies
Sister chromatid
exchange (SCE)
Respiratory tissues or in
vitro systems*
+(6/6) in vitro, human cells/cell
lines, short-term studies
+(13/14) in vitro hamster cell lines,
short-term studies
Positive studies included mostly co-
exposure with MeOH, but several
studies in both human and animal cell
lines, which used methanol-free
formaldehyde, were also positive.
Consistent evidence of SCE
formation from in vitro human
and rodent cell lines
No in vivo studies in
animals, and less consistent
results in exposed humans
Other tissues
+(8/16) in vivo human (inhalation)
PBLs
+(0/3) in vivo, rat (inhalation) short-
term studies
Several studies of occupational
exposure showed increased SCE levels.
Although MeOH-free or MeOH-co-
exposed rat studies were negative,
male rats received MeOH-free
formaldehyde were positive in bone
marrow cells.
Evidence that SCE is induced in
some exposed human
populations, although the
results across studies are not
consistent
This document is a draft for review purposes only and does not constitute Agency policy.
A-228 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Genotoxicity
endpoint(s)
Experimental
system
Genotoxicity evidence (in
descending relevance)
Other relevant information
or limitations
Endpoint summaries
Endpoint conclusion
Other effects (cell
transformation;
DNA repair
inhibition;
unscheduled DNA
synthesis; gene
conversion,
crossing over and
translocation)
Respiratory tissues or in
vitro systems*
+(4/7) in vitro, human primary
cells/cell lines, (2/5 UDS) and (2/2
DNA repair inhibition, short-term
studies
+(4/5) in vitro, rodent cell lines,
short-term studies (1/1 UDS; 3/4 cell
transformation)
+(8/8) nonmammalian system;
[(1/1) DNA repair inhibition; +(2/2)
gene conversion; +(3/3) genetic
crossing over/recombination; +(2/2)
heritable translocation]
Although most of the in vitro and
nonmammalian studies were positive
for other genotoxic effects, these
studies had co-exposure with MeOH.
Available evidence suggests a
variety of other genotoxic
endpoints induced by
formaldehyde exposure, which
may play a supplemental role
in overall genotoxicity.
Many of the other genotoxic
endpoints support the
overall genotoxicity and
mutagenicity of
formaldehyde across
multiple experimental
systems.
Other tissues
+(1/2) in vivo human (inhalation)
Change in 06-alkylguanine DNA alkyl-
transferase activity in PBLs before and
after 2- to 3-month exposure in
embalming or anatomy labs
Evidence is inadequate to
conclude effect on DNA repair
inhibition
This document is a draft for review purposes only and does not constitute Agency policy.
A-229 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
A.5. SUPPORT FOR HAZARD ASSESSMENTS OF SPECIFIC HEALTH
EFFECTS
Supporting information is described for sensory irritation (A.5.2); pulmonary function
(A.5.3); respiratory and immune-mediated conditions, including allergies and asthma (A.5.4);
respiratory tract pathology (A.5.5); mechanistic evidence for potential noncancer respiratory health
effects (A.5.6); respiratory tract, lymphohematopoietic, and other cancers (A.5.9); nervous system
effects (A.5.7); and developmental and reproductive toxicity (A.5.8). The supporting information
includes documentation of literature search methods and specific considerations for evaluating
individual studies to determine their usefulness for assessing the health hazards of formaldehyde
inhalation. General approaches used in the identification and evaluation of individual studies are
summarized in Section A.5.1, with additional details outlined under each of the evaluated hazards.
Because formaldehyde exposure-related issues were a significant concern in this assessment, a
separate description of the considerations for judging exposure assessments in observational
epidemiology studies is included (A.5.1, Exposure Assessments for Observational Epidemiology
Studies), and all experimental studies considered for use in hazard identification, including
controlled exposure studies in both humans and animals, were separately evaluated to assess the
quality of the inhalation exposure protocols (A.5.1, Exposure Quality Evaluation: Animal Toxicology
and Controlled Human Exposure Studies). Quantitative methods (e.g., benchmark dose modeling)
applied to health effect studies considered for use in deriving reference values or cancer risk
estimates are presented in Appendix B.
A.5.1. General Approaches to Identifying and Evaluating Individual Studies
Literature Search Methods
Literature search strategies involved keyword-based queries of the following literature
databases: PubMed fhttps: //www.ncbi.nlm.nih.gov/pubmed/ ) and Web of Science
fhttps: / /apps. webofknowledge.com/ ), with many of the health effect-specific searches including
additional queries of Toxline fhttps://toxnet.nlm.nih.gov/newtoxnet/toxline.html and/or DART
fhttps: //toxnet.nlm.nih.gov/newtoxnet/dart.html. Updates to the computerized searches were
performed annually (i.e., either September or October) through 2016, after which point a separate
systematic evidence map was developed to capture newer literature. For searches through 2016,
the computerized search results were augmented by secondary search approaches, including
curation of reference lists in published reviews and other national or international health
assessments of formaldehyde. Studies were screened for relevance to this toxicological review
based on inclusion and exclusion criteria organized according to PECOO category (Population,
Exposure, Comparison, Outcome, and Other) considerations. This screening was performed using
title and abstract information or hand curation of the full text articles (when screening decisions
This document is a draft for review purposes only and does not constitute Agency policy.
A-230 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
could not be made based on the abstract) in Endnote libraries, and all of the screening decisions are
documented in the formaldehyde page of the U.S. EPA Health Effects and Research Online (HERO)
database fhttps: //hero.epa.gov/hero /). Studies identified as relevant to assessing the health
hazards of formaldehyde inhalation based on the criteria for the individual health effect searches
were evaluated for use in the assessment.
Evaluation of Individual Observational Epidemiology Studies
Epidemiology studies were evaluated for several aspects of bias and sensitivity that could
influence interpretation of study results, including population selection, exposure (measurement
and levels/range), outcome ascertainment, consideration of confounding, and analytic approach.
The potential for selection bias, information bias (relating to exposure and to outcome), and
confounding were evaluated, and an overall confidence classification was developed for each study
(or for a specific analysis within a study) (see Table A-3). The confidence classifications are "high,"
medium," "low," and "not informative." In some cases, sufficient information was available to allow
characterization of the potential direction of bias (i.e., a low confidence study with a likely over-
estimation of the effect estimate). For each study, the evaluations are recorded for each category,
and the confidence classifications for specific endpoints are depicted in a diagram with text
summarizing key limitations.
Table A-28. Approach to evaluating observational epidemiology studies for
hazard identification
High Confidence
(highly informative)
No concern for bias, AND
Study design is highly informative for the outcome in question,
AND
• Analyses were appropriate and robust
Medium Confidence
(informative, with limitations2)
Bias may be present but not expected to have strongly influenced
the effect estimates, AND
Study design and analyses were informative for the outcome in
question
Low Confidence
(minimally informative)
Methodological limitations are significant, but the study results
might still be of limited use (e.g., as support for observations from
other studies; to identify potential data gaps) AND/OR
Bias is apparent or other study aspects reduced sensitivity
Not Informative
(excluded as critically deficient)
Major concerns exist regarding methodological limitations that
increased risk of bias, OR
Description of methods and/ or results were not adequate to
enable a complete evaluation
This document is a draft for review purposes only and does not constitute Agency policy.
A-231 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
Confidence classifications were developed for each study by integrating the judgements for
each category of bias and sensitivity: population selection, information bias, confounding, analysis,
and other (sensitivity). Some considerations included in the expert evaluations included:
Population Selection: Recruitment, selection into study, and participation independent of
exposure status and reported in sufficient detail to understand how subjects were identified
and selected.
Information Bias: Validated instrument for data collection described or citation provided.
Outcome ascertainment conducted without knowledge of exposure status. Timing of
exposure assessment appropriate for observation of outcomes. Information provided on
the distribution and range of exposure with adequate contrast between high and low
exposure.
Potential for confounding: Important potential confounders addressed in study design or
analysis. Potential confounding by relevant co-exposures addressed.
Analysis: Appropriateness of analytic approach given design and data collected;
consideration of alternate explanations for findings; presentation of quantitative results.
Other considerations not otherwise evaluated: Sensitivity of study (exposure levels,
exposure contrast, duration of follow-up, sensitivity of outcome ascertainment).
Controlled human exposure studies were evaluated for important attributes of
experimental studies including randomization of exposure assignments, blinding of subjects and
investigators, and inclusion of a clean air control exposure and other aspects of the exposure
protocol. The evaluation of few individuals (n < 10) resulted in reduced confidence. Several studies
did not describe the measures used to control bias, resulting in a lower level of confidence in these
study results. However, some of these studies evaluated multiple dose levels, an important
strength for the hazard assessment Therefore, these studies were included with medium
confidence when reporting detail was the only identified limitation.
Evaluation of Individual Experimental Animal Studies
Experimental animal studies were evaluated and assigned the following confidence ratings:
High, Medium, or Low Confidence, or "Not Informative" based on expert judgement of each study's
experimental details related to predefined criteria within five study feature categories: exposure
quality, test subjects, study design, endpoint evaluation, and data considerations and statistical
analysis. These evaluations were conducted for each independent "experiment" (i.e., a cohort of
exposed animals assessed for an endpoint or set or related endpoints). Considerations for several
of the criteria can differ depending on what endpoint is being evaluated; thus, a study with multiple
experiments may be evaluated several times, with differing end results. The criteria were assessed
independent of the direction, magnitude, or statistical significance of the experimental results, and
they inform the reliability of the study findings regarding whether these findings are likely to be
This document is a draft for review purposes only and does not constitute Agency policy.
A-232 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information for Formaldehyde—Inhalation
caused by formaldehyde exposure alone. Notably, the criteria are evaluated with regard to the
study's ability to inform the health outcome being evaluated, which may differ from the author's
intended purpose. High to Low Confidence studies represent the most to least useful experiments
for the endpoint(s) in question, respectively, for use in hazard identification (see Table A-4).
Table A-29. Approach to evaluating experimental animal studies for hazard
identification
High Confidence
(highly informative)
• No notable methodological limitations, AND
• Experimental design is highly informative3 for the outcome in
question
Medium Confidence
(informative, with limitationsh)
• Minor concern regarding methodological limitations, AND/ OR
• Experimental design is informative for the outcome in question
Low Confidence
(minimally informative)
• Methodological limitations are apparent and significant, but the
study results might still be of limited use (e.g., as support for
observations from other studies; to identify potential data
gaps) AND/ OR
• Experimental design is minimally informative for the outcome
in question
Not Informative
(excluded as critically deficient)
• Major concerns exist regarding methodological limitations,
which are expected to be a driver of study results, OR
• Experimental design is noninformative for the outcome in
question
Considerations for whether the experimental design is informative include the value (e.g., sensitivity; specificity)
of the methodological approaches for informing the outcome in question, based on known or expected biology
and common practice. These considerations include, but are not limited to: appropriateness and sufficiency of
exposure timing and/or duration to allow for the outcome to be affected; sensitivity and specificity of the
endpoint assays regarding their ability to detect subtle changes in the outcome; and how well the tested animals
(e.g., based on what is known about insensitive species, strains, or sexes) are able to reveal the outcome (note:
the human relevance of the response is not considered at this point).
bAs the expectation is that experimental studies should attempt to control all variables, any study limitation
capable of influencing the data was considered to have negatively affected the reliability of the results. Studies
were categorized as Medium Confidence if they had specific issues which introduce a limited amount of
uncertainty regarding the interpretation of the results as solely attributable to formaldehyde inhalation exposure.
Documentation of the expert judgement evaluations within each of the study feature
categories generally emphasized the identification of observed or potential limitations that might
decrease confidence in the results, with less emphasis on documenting study-specific details that
were interpreted as sufficient for the criteria preferences. These category-specific judgements
were then used to assign the overall determinations of confidence (with the criteria most pertinent
to determining confidence clearly identified). In general terms (specifics are provided for each
hazard outcome evaluation in Appendix A.5.1-A.5.9), the five experimental feature categories
evaluated in experimental animal studies involved the following considerations:
This document is a draft for review purposes only and does not constitute Agency policy.
A-233 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Exposure Quality: Given the importance of the inhalation exposure paradigms used across
the available experimental animal studies, detailed evaluations of exposure quality were
separately performed for each study (see below, Exposure Quality Evaluation: Animal
Toxicology and Controlled Human Exposure Studies).
Test Animals: The species, sex, strain, and age are considered appropriate and sensitive for
testing the endpoint(s); sample size provides reasonable power to assess the endpoint(s);
overt systemic toxicity is absent or not expected at the tested concentrations, or it is
appropriately accounted for. Groups appear to be adequately matched at the onset of the
experiment
Study Design: The study design is appropriate and informative for evaluating the
endpoint(s), including a sufficient exposure duration and/or appropriate timing of endpoint
evaluations to allow for sensitive detection of the effect(s) of interest, and a lack of
additional variables introduced over the course of the study that would be expected to
modify the endpoint(s).
Endpoint Evaluation: The protocols used to assess the endpoint(s) are sensitive (able to
detect subtle changes in the health outcome of interest), complete (include the appropriate
protocol controls), discriminating (specific for the health outcome in question), and
biologically sound (note: this applies to evaluations of novel or unproven methods
regarding their ability to detect the changes in the endpoints of interest). The potential for
experimenter bias is minimized.
Data Considerations and Statistical Analysis: Data for all endpoints evaluated in the
study are presented with sufficient detail (e.g., variability is included) and in the preferred
form (e.g., arbitrary cut-offs were not applied to continuous data). Statistical methods and
the group comparisons analyzed appear to be completely reported, appropriate, and
discerning (note: when inappropriate statistical methods appear to have been used, EPA
sometimes performed additional comparisons).
Evaluation of Individual Mechanistic Studies
In general, studies relevant to mechanistic interpretations informing hazard identification
were not individually evaluated. Rather, the body of evidentiary support (or lack thereof) for
specific, influential mechanistic events (e.g., those known to be associated with the health outcome
of interest; those previously implicated in authoritative reviews as relevant to interpreting
formaldehyde exposure-induced health effects) were considered in totality, with judgements based
on overarching interpretations across sets of related studies.
However, in several instances where a reasonable number of studies were available but the
mechanistic interpretations were not well-established, the individual mechanistic studies were
systematically evaluated. For evaluations of individual mechanistic studies in experimental animal
studies (i.e., mechanistic studies related to respiratory effects; mechanistic studies related to
nervous system effects) the same general features evaluated for more apical measures of toxicity
This document is a draft for review purposes only and does not constitute Agency policy.
A-234 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
were considered (i.e., evaluations of exposure quality and study design were emphasized), although
the specific criteria were simplified to accommodate the increased heterogeneity of the available
mechanistic studies, as compared to more traditional apical measures of toxicity. Similarly, study
evaluations of individual human studies (i.e., mechanistic studies related to respiratory effects;
human studies of genotoxicity endpoints) emphasized consideration of exposure assessment, study
design, outcome ascertainment, and comparison groups for potential sources of bias and their
potential impact.
Evaluation of Exposure in Individual Studies
Exposure Assessments for Observational Epidemiology Studies
All residential or school-based studies with measures of formaldehyde exposure were
included in the hazard identification evaluation. Because the database of studies with direct
measurements is relatively large, residential studies with indirect measures of formaldehyde
exposure (e.g., based on age of building or presence of plywood) were not included. Most of the
included studies attempted to estimate average formaldehyde levels using area samples placed in
one or more locations, with measurement periods ranging from 30 minutes to 2 weeks. A few
studies included more than one sampling period (i.e., sampling on multiple days in different
seasons over the course of a year). Studies in adults and in children indicate that area-based (e.g.,
residential or school) samples are highly correlated with personal samples (Lazenbv etal.. 2012:
Gustafson et al.. 2005): therefore, the use of measures based on residential (e.g., bedroom) samples
rather than personal samples was not considered to be a limitation when evaluating a study.
Formaldehyde concentrations have been found to be uniform throughout the home in both
standing housing stock and mobile homes {Dally, 1981, 22217; Stock, 1987, 23226; Sexton, 1989,
31992; Quackenboss, 1989;Clarisse, 2003,195854}. Therefore, associations have generally been
analyzed using household average concentrations.
The validity of the measurement of average formaldehyde concentration was assessed by
reviewing the description of sampling methods provided in each study. Indoor average
formaldehyde measurements may be influenced by humidity and temperature, season, number of
rooms sampled, sample placement, ventilation, and specific sources of formaldehyde in the building
(Dannemiller etal.. 2013: Salthammer etal.. 2010). Longer sampling periods (e.g., 1- to 2-weeks
duration) were considered to be reflective of usual average exposure levels experienced by
occupants. Studies have shown that formaldehyde levels levels remain relatively stable over a
series of days or weeks {Gustafson, 2005,1512154; Stock, 1987, 23226; Hodgson etal., 2000},
although concentrations are also correlated with season, which reflects the influence of
temperature and humidity { Dannemiller, 2013,1949600; Jaernstroem et al., 2006; Clarisse, 2003,
195854}. Within-person variability increases with shorter sampling durations f Gustafson etal..
2005). However, indoor formaldehyde concentrations have not been found to be associated with
indoor combustion sources, such as active smoking or ETS exposure, and cooking with gas stoves or
This document is a draft for review purposes only and does not constitute Agency policy.
A-235 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
wood burning (Mullen etal.. 2015: Dannemiller et al.. 2013: Gustafson etal.. 2005: Clarisse etal..
2003: Stock. 1987: Hanrahan et al.. 1984: Dally etal.. 19811. Study evaluations looked for
information regarding factors that influence formaldehyde levels as well as quality control
measures and/or citations for exposure protocols. The following characteristics were examined to
assess the potential bias and informativeness of the exposure measures in the observation
epidemiology studies of formaldehyde in residences and schools:
• Duration of exposure measurement period and number of sampling occasions
• Consideration of temperature, relative humidity, and a discussion of quality control
• For shorter exposure periods (< 1 day), details regarding measurement protocol (e.g.,
shutting windows) and consideration of influence of sources of exposure (e.g., smoking or
appliances)
• Limit of detection (LOD) and percent
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
Exposure Quality Evaluation: Animal Toxicology and Controlled Human Exposure Studies
Inhalation toxicity studies are particularly challenging because of the inherent complexity of
generating and characterizing consistent chamber atmospheres. Poor study design, human error,
and problems with mechanical and electronic equipment can impair an inhalation exposure and
undermine the validity of a study. In experimental studies, there is an expectation that test subjects
in an inhalation chamber study will be exposed solely to a well-characterized test article under
conditions that are carefully regulated, frequently measured, and clearly reported. When a
chamber study is conducted under Good Laboratory Practice (GLP) standards, there is typically
greater confidence that all aspects of that study were properly performed and documented.
Inhalation studies were evaluated by scientists familiar with inhalation chamber operations
for seven key elements of exposure quality:
1) Generation Method: The equipment and method used to generate a chamber atmosphere
should be clearly described. If methods from another publication are cited, the methods in
the secondary article were evaluated (if accessible).
2) Test Article Characterization: The test article is the substance or mixture of substances
to which humans or animals are exposed. Any substances used to generate the test article
should be well characterized. For example, formaldehyde gas can be produced by heating
paraformaldehyde, formalin, UFFI insulation, or Delrin plastic. The test article description
should ideally include its physical nature (solid, liquid, gas, etc.), purity, CAS registry
number (if known), and physicochemical properties (including isomerization and
radiolabeling). Because inhaled methanol (but not formaldehyde) is systemically
distributed and can cause neurological and developmental effects, a methanol control
group is desirable for studies of commercial formalin. Only 2 of 84 studies known or
believed to have tested commercial formalin included methanol controls.
3) Analytical Method: The method used to measure test atmospheres should be clearly
described and suitable for the test chemical. There are specific methods (e.g., direct
sampling, adsorptive, or chemical reactive methods, and subsequent analytical
characterization such as HPLC, gas chromatography, etc.) and nonspecific methods such as
gravimetric filter analysis. In addition, a real-time monitoring device (e.g., an aerosol
photometer for aerosols or a total hydrocarbon analyzer for gases or vapors) may be used
to monitor the stability of chamber atmospheres.
4) Analytical Concentrations: Every chamber study should report three concentrations,
which are listed in the order of their usefulness:
• The analytical concentration is the analytically measured concentration of a substance to
which test subjects are exposed in their breathing zone. Because analytical concentrations
are recorded throughout the course of a chamber study, they can reveal generation
problems, fluctuations, analytical problems, and missed exposures. If analytical
concentrations are not reported for a study considered for use in quantitative analyses, an
effort should be made to acquire them from the study authors, as analytical concentrations
are preferred when deriving an RfC. The use of target or nominal concentrations to derive
This document is a draft for review purposes only and does not constitute Agency policy.
A-237 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
an RfC should be cited as a study limitation, although nominal concentrations are
considered accurate for gases (but not vapors).
• The nominal concentration is the mass of generated test article divided by the total
volume of air passed through the chamber. Nominal and analytical concentrations for gases
are usually quite close. Conversely, the nominal concentration for a vapor or aerosol is
typically greater than the analytical concentration (sometimes orders of magnitude greater)
due to test chemical clumping, precipitation, and/or deposition on chamber walls and
plumbing.
• The target concentration is the concentration the study director hopes to achieve in a
chamber study (e.g., 1, 3, and 10 mg/m3). Because a target concentration is a goal—not a
measurement—one should not assume that test subjects were actually exposed at the
precise target concentrations.
• Some fluctuation in analytical chamber concentration is expected, but concentrations
should deviate from the mean chamber concentration by no more than ±10% for gases or
vapors or ±20% for liquid or solid aerosols (GD 39; OECD, 2009). Excessive atmosphere
fluctuation is evidence of a test article generation problem.
5) Particle Size Characteristics: Particle median diameter, density, and distribution
(geometric standard deviation or ag) should be characterized whenever test subjects may
be exposed to an aerosol or to a vapor that may condense into inhalable aerosol particles.
Particle sizing is not necessary when testing a gas. The mass median aerodynamic
diameter (MMAD) is often calculated, but metrics such as physical diameter, median
particle number, or surface area may also be evaluated as the most relevant metric.
6) Chamber Type: Inhalation chambers are either dynamic or static. Dynamic chambers,
which include nose-only, head-only, and whole-body chambers, have a constant flow of
filtered air and consistent test article concentrations, but static chambers do not. EPA and
OECD inhalation test guidelines indicate use of a dynamic chamber. Static chamber studies
are not preferred for longer term hazard identification or exposure response analyses in
particular, as they can lead to a harmful buildup of by-products (e.g., CO2). Consideration
should also be given to whether the test article is best delivered by whole-body or nose-
only chambers. Animals exposed to an aerosol in a whole-body chamber may receive a
significant oral exposure due to preening of particles deposited on their fur. To prevent
this, nose-only chambers are recommended when testing aerosols and vapors that may
precipitate into particles.
7) Controls: A concurrent negative (air) control group should be used in inhalation toxicity
studies. The test chamber, itself, is considered an experimental variable that should be
controlled.
Inhalation study deficiencies are shaded in Table A-30 for easy recognition. A study's
exposure quality may be upgraded if a study author provides key missing data. Each study was
subjectively ranked as having Robust, Adequate, or Poor exposure characterization based upon
the number and severity of deficiencies it has:
This document is a draft for review purposes only and does not constitute Agency policy.
A-238 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Robust Exposure Characterization: There are no notable uncertainties or limitations
regarding exposure methodology.
Adequate Exposure Characterization: There are minor uncertainties or limitations
regarding exposure methodology.
Poor: There are serious uncertainties or limitations regarding exposure methodology.
This document is a draft for review purposes only and does not constitute Agency policy.
A-239 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-30. Inhalation exposure quality: formaldehyde (Note: exposure deficiencies are shaded)
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
Robust Exposure Characterization: there are no notable uncertainties or limitations regarding exposure methodology
Adams et al. (1987)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamicwhole-
Mouse
body
Ahmed et al. (2007)
Paraformaldehyde
NR
HPLC
Reported
NA
Dynamic whole-
Mouse
body
(Albert et al., 1982)
Rat
See (Sellakumar et al.,
Paraformaldehyde
—
—
1985)
(Andersen et al., 2010)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Rat
body
(Appelman et al., 1988)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Rat
whole-body
(Babiuk et al., 1985)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Rat
(and 7 other aldehydes)
body
(Bach et al., 1990)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Human
"climate
[Exposure parameters are
chamber"
inferred from coauthor using
same climate chamber in
Anderson and M0lhave,
(Andersen and Molhave,
1983)1
(Barrow, 1983)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry and
Reported
NA
Dynamic head-
Mouse and Rat
colorimetric method
only
(Battelle, 1981)
See Kerns et al. (1983)
Paraformaldehyde
—
(Kerns et al., 1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-240 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Berglund and Nordin,
Freshly prepared formalin
Evaporation
IR spectrophotometry;
Reported
NA
Dynamic
1992)
from paraformaldehyde
sodium bisulfite method;
olfactomer
Human
(no methanol)
acetyl acetone method
(Berglund et al., 2012)
Freshly prepared formalin
Evaporation
IR spectrophotometry;
Reported
NA
Dynamic
Human
from paraformaldehyde
(no methanol)
acetyl acetone method
olfactometer
(Casanova et al., 1994)
Paraformaldehyde,
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Rat
[14C]-paraformaldehyde
body
Cassee et al. (1996a,b)
Freshly prepared formalin
Evaporation
Formaldehyde analyzer
Reported
NA
Dynamic
Rat
from paraformaldehyde
(no methanol) and/or
acetaldehyde, acrolein
nose-only
(Cassee and Feron,
Freshly prepared formalin
Evaporation
IR spectrophotometry
Reported
NA
Dynamic nose-
1994a)
Rat
from paraformaldehyde
(no methanol).
only
Exposures were to PFA
only, ozone only, or to
both chemicals
(Chang et al., 1981)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry and
Reported
NA
Dynamic head-
Rat and mouse
colorimetric method
only
(Chang et al., 1983)
Paraformaldehyde and
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
[14C]-paraformaldehyde
body and
Rat and mouse
head-only
(1982)
See (Kerns et al., 1983)
Paraformaldehyde
—
—
NA
(Coon et al., 1970)
Freshly prepared formalin
Spray nozzle and
IR analyzer equipped with a
Reported
NA
Dynamic whole-
Rat, guinea pig, rabbit, dog,
monkey
(paraformaldehyde
added to hot distilled
water; 1.35% solution)
evaporation of solution
catalytic oxidizer
body
(Dalbev, 1982)
Paraformaldehyde
Thermal depolymerization
Colorimetric analysis
Within 5% of target
NA
Dynamic whole-
Hamster
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-241 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Dallas et al., 1989)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Rat
body
(Day et al., 1984)
UFFI off-gas products
Broken-up UFFI foam was
Chromotropic acid
Reported
NA
Dynamic whole-
Human
dampened with water,
then gases collected in
4500 L polyethylene
balloons.
body
(Dean et al., 1984)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Mouse
body
(Dinsdale et al., 1993)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Rat
body
Experiment 2
(See also Experiment 1-
Inadequate)
(Feron et al., 1988)
Paraformaldehyde
Thermal depolymerization
Colorimetric
Reported
NA
Dynamic whole-
Rat
body
(Fuiimaki et al., 2004b)
Paraformaldehyde
NR
HPLC
Reported
NA
Dynamic whole-
Mouse
body
(Green et al., 1987)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
body
(Green et al., 1989)
Paraformaldehyde
Thermal depolymerization
Colorimetric monitor
Reported
NA
Dynamic whole-
Human
body
(Groten et al., 1997)
Paraformaldehyde alone
Vaporization of freshly
Colorometric method
Reported (sampled
NA
Dynamic whole-
Rat
or in combination with
dichloromethane, aspirin,
di(2-ethylhexyl)-
phthalalate, cadmium
chloride, stannous
chloride, butyl
hydroxyanisol,
loperamide, and
spermine
made formalin
in the animals'
breathing zone)
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-242 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Havashi et al., 2004)
Paraformaldehyde
Thermal depolymerization
HPLC
Reported
NA
Dynamic whole-
body
Mouse
(Holmstrom et al.,
Paraformaldehyde with
Thermal depolymerization
Formaldehyde meter
Reported
NA
Dynamic whole-
1989b)
and without wood dust
body
Rat
(Jakab, 1992)
Paraformaldehyde;
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
Mouse
exposure was to
formaldehyde gas with or
without carbon black
aerosol
body
(Kamata et al., 1997)
Formalin with 10%
Sprayed into a bottle
Acetylacetone
Reported for
NA
Dynamic nose-
Rat
methanol
A methanol control group
was used
heated to 70°C
formaldehyde and
methanol
only
Kerns etal. (1983); CUT
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic whole-
(1982); Battelle Columbus
body
Laboratories (1981);
Swenberg et al. (1980)
(Kerns et al., 1983);
(1982); (Battelle. 1981);
(Swenberg et al., 1980a)
Rat and mouse
(Kulle et al., 1987a)
Paraformaldehyde
Thermal depolymerization
Toxic gas monitor,
Reported
NA
Dynamic
Human
(reference provided)
chromotropic acid
whole-body
(Kulle. 1993)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Human
(reference provided)
whole-body
(Kuper et al., 2011)
Probably freshly prepared
NR
IR spectrophotometry
Reported
NA
Dynamic
Rat
formalin (10.21% FA)
whole-body
(Larsen et al., 2013)
Polyacetal (a
Permeation tube in a Kin-
HPLC
Reported
NA
Dynamic head-
Mouse
formaldehyde polymer)
in permeation tubes
Tek gas standard
generator
only
This document is a draft for review purposes only and does not constitute Agency policy.
A-243 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
Martin (1989)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Rat
whole-body
(Monteiro-Riviere and
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Popp, 1986)
Rat
whole-body
(Monticello et al., 1991)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Rat
whole-body
(Monticello et al., 1996)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Rat
whole-body
(Monticello and
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Morgan, 1997)
Rat
whole-body
Based on (Monticello et
al.. 1996)
Morgan et al. (1986a)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
±5% of nominal
NA
Dynamic
Rat
head-only
Morgan et al. (1986c)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Rat
whole-body
(Mueller et al., 2012)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor,
Reported
NA
Dynamic
Human
HPLC
whole-body
(Mueller et al., 2013)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
Reported
NA
Dynamic
Human
HPLC
whole-body
(Ozen et al., 2002)
Paraformaldehyde
Thermal depolymerization
Gas chromatography and
Reported
NA
Dynamic
Rat
formaldehyde monitor
whole-body
(Reuzel et al., 1990)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
Reported
NA
Dynamic
Rat
whole-body
(Riedel et al., 1996)
Guinea pig
Formaldehyde gas
Pressurized bottles
Photometric
Reported
(in animals'
breathing zone)
NA
Dynamic
whole-body
(Roemer et al., 1993)
Paraformaldehyde
Thermal depolymerization
IR spectrophometry
Within 10% of
NA
Dynamic head-
Rat
nominal
only
This document is a draft for review purposes only and does not constitute Agency policy.
A-244 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Rusch et al., 1983)
Freshly prepared formalin
Air was bubbled through
Chromotropic acid
Reported
NA
Dynamic
Rat, monkey, hamster
(unstabilized 5% solution
with 0.03% methanol)
formalin
whole-body
Saldiva et al. (1985)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Rat
whole-body
(Sauder et al., 1986)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
(reference provided)
body
(Sauder et al., 1987)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
body
(Sellakumar et al., 1985)
Paraformaldehyde;
A slurry of PFA in paraffin
PFA: Chromotropic acid
Reported
NA
Dynamic whole-
and
exposure to
oil (kerosene) was
HCI: titration with NaOH
[NOTE: HCI is a
body
(Albert et al., 1982)
Rat
formaldehyde and/or
HCI. Co-exposure to
generated by thermal
depolymerization.
BCME: gas
chromatography/mass
powerful catalyst
for the
formaldehyde and HCI
forms bis(chloromethyl)-
ether (BCME), a
carcinogenic reaction
product.
HCI was from a
compressed gas tank.
spectrometry
polymerization of
FA into oligomers
(Bevington and
Norrish, 2012).
Unlike FA gas,
oligomer particles
may be respirable]
(Sheppard et al., 1984)
Freshly prepared formalin
Air was bubbled through
IR spectrophotometry
Reported
NA
Respiratory valve
Human
from paraformaldehyde
(methanol-free)
formalin
mouthpiece
(Songur et al., 2003)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
Reported
NA
Dynamic whole-
Rat
body
(Songur et al., 2008)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
Reported
NA
Dynamic whole-
Rat
body
(Sorg et al., 2001b)
Paraformaldehyde
Thermal depolymerization
Photoacoustic multi-gas
Reported
NA
Dynamic whole-
Rat
monitor
body
[Cited exposure parameters
from (Sorg et al., 1998)]
(Swenberg et al., 1980b)
See Kerns et al. (1983)
Paraformaldehyde
NA
—
This document is a draft for review purposes only and does not constitute Agency policy.
A-245 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Swiecichowski et al..
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
1993)
body
Guinea pig
{Tobe, 1985, 3574}
Formalin
Sprayed into a heated
Acetylacetone
Reported for
NA
Dynamic whole-
[Study report]
(w/10% methanol)
glass bath
formaldehyde and
body
Rat
A methanol control group
was used
methanol
(Tsukahara et al., 2006)
Paraformaldehyde
NR
HPLC
Reported
NA
Dynamic whole-
Mouse
body
(Usanmaz et al., 2002)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
Reported
NA
Dynamic
Mouse
Not described
(Vosoughi et al., 2013)
Paraformaldehyde
Thermal depolymerization
Photoionization detector
Reported
NA
Dynamic
Mouse
(Wood and Coleman,
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported.
NA
Dynamic whole-
1995)
Animals were able
body
Mouse
to stop irritating FA
exposure
(Woutersen et al., 1987)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic
Rat
whole-body
(Woutersen et al., 1989)
Paraformaldehyde
Thermal depolymerization
Colorimetric
Reported
NA
Dynamic whole-
Rat
body
Zeller et al. (2011)
Paraformaldehyde
Thermal depolymerization
HPLC and formaldehyde
Reported
NA
Dynamic whole
Human
monitor
body
(Zitting, 1982)
Polyacetal plastic
Oxidative
Visible absorption
Reported
NA
Dynamic whole-
Rat
(Delrin®)
thermodegradation
(250°C) to formaldehyde,
formic acid, and acrolein
spectrometry (NIOSH, 1972)
body
(Zwart et al., 1988)
Paraformaldehyde
Thermal depolymerization
Colorimetric
Reported
NA
Dynamic whole-
Rat
(Woutersen et al.,
1987)
body (reference
provided)
Adequate Exposure Characterization: there are minor uncertainties or limitations regarding exposure methodology.
This document is a draft for review purposes only and does not constitute Agency policy.
A-246 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
{Andersen, 1979, 6248301;
also described in Andersen
and M0lhave (1983)
Human
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Within 20% of
target
NA
Dynamic whole-
body
(Andersen et al., 2008)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry,
HPLC
Reported
NA
Dynamic whole-
body
Rat
(=30% variation in
atmospheres)
Andersen and Lundqvist
(1970) [book chapter]
Human
Described in (Andersen
and Molhave, 1983)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Within 20% of
target
NA
Dynamic
"climate
chamber"
(Andersen and Molhave,
1983) [book chapterl
Human
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Within 20% of
target
NA
Dynamic
"climate
chamber"
(Apfelbach and Weiler,
1991)
Rat
Paraformaldehyde
Thermal depolymerization
HPLC
NR
NA
NR
Exposures in
plexiglas holding
cages
(Asian et al., 2006)
Rat
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
NR
"Desired
concentrations
were prepared"
NA
Dynamic whole-
body
(Bender et al., 1983)
Human
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
NR13
NA
Dynamic smog
chamber with 7
sets of ports
(Boja et al., 1985)
Rat
Paraformaldehyde
Thermal depolymerization
Gas chromatography
NR
NA
Dynamic whole-
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-247 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Chang and Barrow,
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry and
NR
NA
Dynamic head-
1984)
colorimetric method
only
Rat
(Fuiimaki et al., 2004b)
Paraformaldehyde
NR
Formaldehyde monitor
NR
NA
Dynamic whole-
Mouse
(Secondary source not
body
[Exposure parameters in
found)
(Fujimaki et al., 2004a)]
Holmstrom et al. (1989b)
Paraformaldehyde
Thermal depolymerization
NR
Reported
NA
Dynamic whole-
Rat
body
(Horton et al., 1963b)
Paraformaldehyde
Thermal depolymerization
Method of Goldman and
NR
NA
Dynamic whole-
Mouse
Yagoda
(reference provided)
body
(Ito et al., 1996)
Formalin w/13%
Formalin was placed in
4-amino-3-hydrazino-5-
Reported
NA
Dynamic
Rat
methanol
50°C diffusion tubes
mercapto-l,2,4-triazole
NR for methanol
(not described)
A methanol control group
method; analytical method
was used
for methanol NR
James et al. (2002)
Human
Formaldehyde
Formaldehyde off-gassed
from various materials in a
spacecraft simulator
IR spectrophotometry,
chromotropic acid,
HPLC
Reported
(steady state
concentrations
were not achieved
until the last few
days of the study)
NA
Spacecraft
simulator
(Kulle and Cooper, 1975)
Rat
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
NR
NA
Dynamic
olfactometer
(Lang et al., 2008)
Paraformaldehyde
Thermal depolymerization
Dinitrophenylhydrazine and
NR
NA
"Quasi static
(and ethyl acetate as a
HPLC analysis
conditions"
Human
masking agent)
Formaldehyde monitor
(Meng et al., 2010)
Paraformaldehyde
Thermal depolymerization
IR Spectrophotometry
NR
NA
Dynamic
Rat
(not described)
(Moeller et al., 2011)
[13CD2]-formaldehyde
NR
NR
Reported
NA
Dynamic whole-
body
Monkey
This document is a draft for review purposes only and does not constitute Agency policy.
A-248 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Monticello et al., 1989)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
NR
NA
Dynamic whole-
Monkey
body
(Morgan et al., 1984)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry and
Within 20% of
NA
This is not an
Frog
An ex vivo study of frog
colorimetric assay
nominal
inhalation
palates exposed to
formaldehyde gas
chamber study
(Nielsen et al., 1999)
Paraformaldehyde
Thermal depolymerization
NR
NR
NA
Dynamic whole-
Mouse
body
National Toxicology Program
Paraformaldehyde
Thermal depolymerization
Formaldehyde meter
NR
NA
Dynamic whole-
(2017)
body
Mouse
Ozen et al. (2003a)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
NR
NA
Dynamic whole-
Rat
body
Ozen et al. (2003b)
Paraformaldehyde
Thermal depolymerization
Gas chromatography and
NR
NA
Dynamic whole-
Rat
formaldehyde monitor
body
(Ozen et al., 2005)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
NR
NA
Dynamic whole-
Rat
body
Sari et al. (2004a)
Mouse
Paraformaldehyde
NR
(Secondary source not
found)
"a chemical method"
and
Formtector XP-308
Reported
NA
Dynamic whole-
body
Sari et al., (2004b)
Paraformaldehyde
NR
"measured chemically"
Reported
NA
Dynamic whole-
Mouse
(Mice were exposed
(Secondary source not
and
body
Cited exposure parameters
intranasally to 500 ppm
found)
Formtector XP-308
from Sari et al. (2004a)
toluene/mouse 6 h/day
for 3 days prior to FA
exposure)
(Sari et al., 2005)
Paraformaldehyde
NR
"measured chemically"
Reported
NA
Dynamic whole-
Mouse
(Secondary source not
found)
and
Formtector XP-308
body
(Sarsilmaz et al., 1999)
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
NR
NA
Dynamic
Rat
(reference provided)
whole-body
This document is a draft for review purposes only and does not constitute Agency policy.
A-249 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Sarsilmaz et al., 2007)
Rat
[Assumed to be the same
cohort as (Asian et al.,
2006)1
Paraformaldehyde
Thermal depolymerization
(reference provided)
Formaldehyde monitor
NR
"Desired
concentrations
were prepared"
NA
Dynamic "prism-
shaped glass
covers"
(Schachter et al., 1986)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
(apparent co-exposure to
2-propanol)
over boiling
2-propanol
body
(Schachter et al., 1987)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
(apparent co-exposure to
2-propanol)
over boiling
2-propanol
body
Songur et al. (2005)
Rat
Paraformaldehyde
Thermal depolymerization
Formaldehyde monitor
NR
NA
Dynamic
(Sorg et al., 1998)
Paraformaldehyde
Thermal depolymerization
HPLC
Reported
NA
Dynamic whole-
Rat
44% decline in
concentration over
the course of the
experiment
body
Sorg et al. (2001b)
Rat
Experiment 2 and 3
(See also Experiment 1-lnadequate)
Paraformaldehyde
Thermal depolymerization
HPLC
(Sorg et al., 1998)
NR
NA
Dynamic whole-
body
(Sorg et al., 2004)
Paraformaldehyde with
Thermal depolymerization
Photoacoustic multi-gas
Reported
NA
NR
Rat
co-exposure to orange oil
(a known irritant)
monitor
(Sorg and Hochstatter,
1999)
Rat
Experiment 2
(See also Experiment 1-
Inadequate)
Paraformaldehyde
Thermal depolymerization
HPLC
(Sorg et al., 1998)
NR
NA
Dynamic whole-
body
(Wilmer et al., 1987)
Rat
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
NR
NA
Dynamic whole-
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-250 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Wilmeretal.. 1989)
Paraformaldehyde
Thermal depolymerization
IR spectrophotometry
NR
NA
Dynamic
Rat
Whole-body
(Witek et al., 1986)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
(apparent co-exposure to
2-propanol)
over boiling
2-propanol (82.5°C)
body
(Witek etal.. 1987)
Paraformaldehyde
Thermal depolymerization
Chromotropic acid
Reported
NA
Dynamic whole-
Human
(apparent co-exposure to
2-propanol)
over boiling
2-propanol (82.5°C)
body
Poor Exposure Characterization: there are serious uncertainties or limitations regarding exposure methodology.
Al-Saraj et al. (2009)
10% Formalin
Evaporation
Colorimetric method
Reported
NA
Dynamic whole-
Rabbit
No methanol control
[Pretreatment with
Ivermectin which can
cause cleft palate and
clubbed forelimbs in
rabbits]
(based on a reference)
Methanol not measured
(12 ppm)
body
Amdur (1960)
Formalin (37%)
Sintered glass bubbler
Colorimetric method and
Reported
NaCI
Dynamic whole-
Guinea pig
chromotropic acid
particles
measured
body
(Arican et al., 2009)
Paraformaldehyde
Thermal depolymerization
NR
NR
NA
Dynamic whole-
Rat
body
Bansal et al. (2011)
10% Formalin
Evaporation from open
NR
NR
NA
Open containers
Rabbit
40% Formalin
No methanol control
containers
Target and nominal
concentrations also
NR
of formalin were
placed below
cages
This document is a draft for review purposes only and does not constitute Agency policy.
A-251 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article
characterization
Analytical
Particle
Chamber
Study/species
and controls
Generation method
Analytical method
concentrations
size
Description
(Biagini et al., 1989)
Formalin w/10-15%
Injected into a GC injector
Formaldehyde monitor
Reported
NA
Dynamic whole-
Monkey
methanol
and heated to 220-230°C
Methanol not measured
body
No methanol control
[Anesthesia with
ketamine and xylazine,
which cause
bronchodilation, could
affect pulmonary
function measurements.]
(Bian et al., 2012)
Formalin
Evaporation
Formaldehyde meter
10.0 ±1.0 mL/m3
NA
Dynamic whole-
Rat
No methanol control
Methanol not measured
body
(Bhalla et al., 1991)
Paraformaldehyde
Thermal depolymerization
NR
NR
NA
Dynamic nose-
Rat
only
(Bokina et al., 1976)
Rabbit
NR
No methanol control
NR
NR
NR
NA
NR
(Buckley et al., 1984)
Formalin
NR
IR spectrophotometry
Reported
NA
Dynamic whole-
Mouse
(co-exposure to
methanol)
No methanol control
Methanol not measured
body
(Casset et al., 2006b)
Formalin
Evaporated from a Pyrex
HPLC
<10% of target
NA
Dynamic whole-
Human
(35% aqueous medicinal
solution of formaldehyde;
co-exposure to methanol)
No methanol control
boiler at 85°C
Methanol not measured
body with
subjects wearing
masks
(Chonglei et al., 2012)
Mouse
Mice were
simultaneously exposed
to formaldehyde,
benzene, toluene, and
xylene vapors.
The test article for
formaldehyde was NR
NR
Digital electrochemical
analyzer and gas
chromatography
NR
NA
Dynamic whole-
body
(airflow not
reported)
Cometto-Muniz et al. (1989)
NR
NR
Chromotropic acid
Reported
NA
Dynamic
Human
No methanol control
olfactometer
This document is a draft for review purposes only and does not constitute Agency policy.
A-252 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Day et al., 1984)
Solution of formalin in
Atomized and then
Chromotropic acid
Reported
NA
Dynamic whole-
Human
methanol.
No methanol control
evaporated on a hot plate.
Methanol not measured
body
De Ceaurriz et al. (1981)
Mouse
NR
No methanol control
NR
Colorimetric method
Methanol not measured
NR
NA
Dynamic whole-
body
(Dinsdale et al., 1993)
Rat
Experiment 1
(See also Experiment 2 -
Robust)
Formalin (co-exposure to
methanol)
No methanol control
Jet atomizer (Exp 1)
IR spectrophotometry
Methanol not measured
Reported
NA
Dynamic whole-
body
(Ezrattv et al., 2007)
Formalin
Thermal depolymerization
Semiconductor gas sensor
NR
NA
Dynamic whole-
Human
(co-exposure to
methanol)
No methanol control
Methanol not measured
body
(Falk et al.. 1994)
Human
Formalin
(co-exposure to
methanol)
No methanol control.
Evaporation from a heated
glass surface
Liquid chromatography
Reported for
treated and
negative control
groups
NA
Dynamic
Whole-body
(Gieroba et al., 1994)
Rabbit
38% Formalin
No methanol control
Evaporation
None
NR
NA
A tube delivered
FA vapor to
rabbits' nares
(Gofmekler, 1968)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
NR
(Gofmekler and
Bonashevskava, 1969)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
NR
(Golalipour et al., 2007)
Rat
NR but exposure would
have been to formalin
with co-exposure to
methanol
No methanol control
NR, but formaldehyde and
methanol would have off-
gassed from necropsy tubs
of formalin
Formaldehyde Draeger
tubes
Methanol not measured
Reported
NA
Not a chamber
study; rats
exposed in
dissection room
This document is a draft for review purposes only and does not constitute Agency policy.
A-253 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Guseva, 1973)
Rat
NR
No methanol control
NR
Rats were simultaneously
exposed by inhalation and
drinking water
Fuchsin sulfurous acid
method
Methanol not measured
NR
NA
Dynamic (not
described)
(Han et al., 2013)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
Static
(Harving etal.. 1990)
Alkaline solution of
Thermal depolymerization
Acetylacetone
Reported
NA
Dynamic whole-
Human
formalin; co-exposure to
methanol
No methanol control
Methanol not measured
body
(Silva Ibrahim et al..
Formalin (purity NR)
Ultrasonic nebulizer
NR
NR
0.5-1 nm
Dynamic whole-
2015)
Rat
A vehicle control group
MMAD NR
body
was exposed to water
No methanol control
(lonescu et al., 1978)
Rabbit
NR
(probably aerosolized
formalin)
No methanol control
NR
NR
Methanol not measured
NR
(target and nominal
concentrations also
NR)
NA
Static
(Jaeger and Gearhart,
1982)
Formalin
No methanol control
Aerosolization and
evaporation
IR spectrophotometry and
colorimetric method
Methanol not measured
Reported
NA
Dynamic whole-
body
(Mason jar)
Mouse and Rat
Kamata et al. (1996a)
Rat
Formalin (with 10%
methanol)
No methanol control
Formalin was sprayed and
heated to generate a
vapor
Acetylacetone
Methanol not measured
Reported
NA
Dynamic whole-
body
Kamata et al. (1996b)
Rat
Formalin with 10%
methanol
No methanol control
Sprayed into a bottle
heated to 70°C
Acetylacetone
Methanol not measured
Reported
NA
Dynamic nose-
only
(Kane and Alarie, 1977)
Formalin
Evaporation
Colorimetric method
Reported
NA
Dynamic head-
Mouse
No methanol control
Methanol not measured
only
This document is a draft for review purposes only and does not constitute Agency policy.
A-254 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Katsnelson et al., 2013)
Rat
NR
No methanol control
NR
NR
Methanol not measured
Reported
NA
Dynamic whole-
body
(Kimura et al., 2010)
Rat
37% Formalin with 15%
methanol
No methanol control
Dynamic gas generator
(evaporation)
4-amino-3-hydrazino-5-
mercapto-l,2,4-triazole
method
Methanol not measured
NR
NA
Dynamic whole-
body
(Kim et al., 2013b)
Mouse
NR
No methanol control
NR
HPLC
NR
NA
NR
(Kitaev et al., 1984)
NR
NR
Gravimetric (not described)
NR
NA
Dynamic
Rat
No methanol control
Methanol not measured
(not described)
(Krakowiak et al., 1998)
10% Formalin
Evaporation
Chromotropic acid
Reported
NA
Dynamic whole-
Human
No methanol control
Methanol not measured
body
(Kum et al., 2007a)
Rat
Formalin
No methanol control
NR
Gas detection pump
(reference provided)
Methanol not measured
NR
NA
Dynamic
whole-body
(Lee et al., 1984)
Guinea pig
4% Formalin w/1%
methanol
37% formalin w/10%
methanol
No methanol control
Aerosol generated by a
nebulizer
Formaldehyde:
chromotropic acid
Methanol: IR
spectrophotometry
NR for
formaldehyde or
methanol
NR
Dynamic whole-
body
(Liao et al., 2010)
Formalin
NR
Formaldehyde meter
NR
NA
Static
Rat
No methanol control
Methanol not measured
(Lino dos Santos Franco
et al., 2006)
Rat
Formalin (diluted to 1%;
with 0.32% methanol)
Ultrasonic nebulizer
NR for formaldehyde or
methanol
NR for
formaldehyde or
NR
Dynamic whole-
body
A methanol control group
was used.
methanol
(nominal
concentration NR)
(Lino dos Santos Franco
et al., 2009)
Rat
Formalin
No methanol control
Ultrasonic nebulizer
NR
NR
Methanol not
measured
NR
Dynamic
(probably whole-
body)
This document is a draft for review purposes only and does not constitute Agency policy.
A-255 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
Lino dos Santos Franco et al.
(2011)
Rat
Formalin (diluted to 1%;
with 0.32% methanol)
No methanol control
Ultrasonic nebulizer
NR
NR
Methanol not
measured
NR
NR
Liu et al. (2009)
Rat
Formalin (37%)
No methanol control
Evaporation from the inner
walls of the static chamber
Formaldehyde monitor
Reported
NA
Static
Liu et al. (2010)
Rat
Formalin (37%)
No methanol control
Evaporation from the inner
walls of the static chamber
Formaldehyde monitor
Reported
NA
Static
(Lu et al., 2008b)
Wood baseboard
(not described);
co-exposure to
unidentified chemicals
NR
NR
NR
NA
Dynamic
Mouse
Not described
(Maiellaro et al., 2014)
Rat
Formalin (source and
purity NR)
The vehicle control was
exposed to water
Ultrasonic nebulizer
NR
Methanol not measured
NR
Note: one exposure
level tested
Reported
Dynamic
(Malek et al., 2003c)
(Malek et al., 2003a)
(Malek et al., 2003b)
Rat
Formalin
No methanol control
Evaporation from a dish in
the chamber
Formaldehyde Draeger
tubes
Methanol not measured
Reported
NA
Static with holes
(Malek et al., 2004)
Mouse
Formalin
No methanol control
Evaporation from a dish in
the chamber
Formaldehyde Draeger
tubes
Methanol not measured
Reported
NA
Static with holes
(Maronpot et al., 1986)
Mouse
Formalin (9.2%w/v)
No methanol control
Nebulization and
evaporation
Chromotropic acid
Reported
NA
Dynamic whole-
body
(Matsuoka et al., 2010)
Mouse
Formalin
No methanol control
Evaporation
Cosmos® smell sensor
NR
NA
Dynamic whole-
body
(Monfared, 2012)
Mouse
NR
No methanol control
NR
NR
NR
NA
Dynamic whole-
body
(Morgan, 1983)
Rat
Paraformaldehyde
(reference provided)
Thermal depolymerization
NR
NR
NA
Dynamic whole-
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-256 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Nalivaiko et al., 2003)
Rabbit
Paraformaldehyde
Thermal depolymerization
None
NR
NA
A tube delivered
FA vapor to
rabbits' nares
(Ohtsuka et al., 1997)
Rat
NR
No methanol control
Aerosol generated by an
atomizer
NR
Methanol not measured
NR
NR
Dynamic whole-
body "test
room"
(Ohtsuka et al., 2003)
Rat
1% Formalin
No methanol control
Aerosol generated by an
atomizer
NR
Methanol not measured
NR
NR
Dynamic whole-
body "test
room"
(Pazdrak et al., 1993)
NR
No methanol control
NR
IR spectrophotometry
Reported
NA
Dynamic whole-
body
Human
(Pitten et al., 2000)
Rat
Formalin
No methanol control
Evaporation from a dish in
the chamber
Acetylacetone method and
photometric evaluation
Methanol not measured
Reported
NA
Static
(Pross et al., 1987)
Formalin
Evaporation of formalin
Formalin: chromotropic acid
NR
NA
Dynamic whole-
Human
No methanol control
aerosol
Methanol not measured
body
(Pross et al., 1987)
Human
Milled UFFI particles (4
Hm) contaminated with
heavy microbial growth
UFFI aerosol generation
not described
UFFI aerosol: gravimetric
filters and an aerodynamic
particle sizer
NR
NA
Dynamic whole-
body
(Pross et al., 1987)
Human
UFFI off-gas products.
UFFI off-gas generated by
passing air through beds of
fractured UFFI wetted with
water
NR
NR
NA
Dynamic whole-
body
(Pushkina et al., 1968)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
NR
(Sadakane et al., 2002)
Mouse
Formalin (0.5% solution
in saline
No methanol control
Aerosol generated by an
ultrasonic nebulizer
NR
Methanol not measured
NR
NR
NR
(Saillenfait et al., 1989)
Rat
Formalin w/10%
methanol
No methanol control
Air was bubbled through
formalin
IR spectrophotometry
Methanol not measured
Reported
NA
Dynamic
This document is a draft for review purposes only and does not constitute Agency policy.
A-257 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Sandikci et al., 2007b)
Rat
NR
No methanol control
NR
NR
(reference provided)
Methanol not measured
NR
NA
Dynamic whole-
body
(Sandikci et al., 2009)
Rat
NR
No methanol control
NR
Formaldehyde Draeger
tubes
NR
NA
Dynamic whole-
body
Sanotski et al. (1976)
Rat
NR
No methanol control
NR
Colorimetry (not described)
Methanol not measured
NR
NA
Dynamic
(not described)
(Schreiber et al., 1979)
Hamster
NR
No methanol control
NR
NR
NR
NA
NR
(Schuck et al., 1966)
Human
Formaldehyde
and other photooxidation
products
Formaldehyde was
generated during
propylene photooxidation
and ethylene
photooxidations in a
reaction chamber exposed
to high intensity UV light
(3000 A)
Chromotropic acid
Mean
concentrations
provided in a graph
NA
Reaction
chamber with
welding masks
attached for eye
exposure
(Senichenkova, 1991b)
Rat
NR
No methanol control
NR
Gravimetric (not described)
Methanol not measured
NR
NA
Dynamic
(not described)
(Senichenkova and
Chebotar, 1996)
Rat
NR
No methanol control
NR
Gravimetric (not described)
Methanol not measured
NR
NA
Dynamic
(not described)
(Sheveleva, 1971)
Rat
NR
No methanol control
NR
NR
(reference provided);
Methanol not measured
Reported
NA
Dynamic whole-
body
(Sorg et al., 1996)
Rat
Formalin
No methanol control
Air was bubbled through
formalin
NR
Methanol not measured
Reported
NA
Dynamic whole-
body
Sorg et al. (2001b)
Rat
Experiment 1
(See also Experiments 2 and 3-
Ad equate)
Formalin
No methanol control
Evaporation of formalin
NR
Methanol not measured
NR
NA
Dynamic whole-
body
This document is a draft for review purposes only and does not constitute Agency policy.
A-258 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
Sorg et al. (2002)
Rat
Formalin
No methanol control
Evaporation
None
NR
NA
Cotton swabs
containing
various formalin
dilutions were
placed in a maze
(Sorg and Hochstatter,
1999)
Rat
Formalin
No methanol control
Air was bubbled through
formalin
(Sorg et al., 1996)
NR
NR
NA
Dynamic whole-
body
Experiment 1
(See also Experiment 2-
Adequate)
(Speit et al., 2011b)
Rat
Formalin
No methanol control
Evaporation
NR
Methanol not measured
Reported
NA
Dynamic whole-
body
(Swenberg et al., 1983b)
[book chapter]
Rat and Mouse
[14C]- formaldehyde
NR
NR
NR
NA
NR
(Swenberg et al., 1986)
[book chapter]
Rat and Mouse
NR
No methanol control
NR
NR
NR
NA
NR
(Tani et al., 1986)
Rabbit
37% Formalin
No methanol control
Evaporation
4-amino-3-hydrazino-5-
mercapto-l,2,4-triazole
method
Methanol not measured
NR
NA
Direct exposure
to the upper and
lower
respiratory tract
via two T-tubes
(Tepper et al., 1995)
Mouse
Carpet containing volatile
organic compounds,
pesticide residues, and
microbiological flora
Heating of carpet
Gas chromatography
High resolution mass
spectrometry
Reported for FA and
9 other specific
organic chemicals
NR
Dynamic head-
only
(Tarkowski and Gorski,
1995)
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
NR
Mouse
This document is a draft for review purposes only and does not constitute Agency policy.
A-259 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study/species
Test article
characterization
and controls
Generation method
Analytical method
Analytical
concentrations
Particle
size
Chamber
Description
(Wang et al., 2012)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
Static
(not otherwise
described)
(Weber-Tschopp et al.,
Formalin (35%)
No methanol control
A syringe delivered
formalin to a heated
(120°C) Pyrex glass tube
Chromotropic acid
Reported
NA
Dynamic whole-
body
1977)
Human
Methanol not measured
(Xing et al., 2007)
Mouse
NR
No methanol control
NR
NR
NR
NA
NR
(Yang et al., 2001)
Human
Plywood (5 layers) which
off-gassed formaldehyde
and traces of C6-Cu
aldehydes.
The plywood was cut into
50- x 10-cm planks and
placed in a small chamber
to facilitate off-gassing.
Formaldehyde monitor
Reported for
formaldehyde, but
location of
measures NR;
concentrations of
other gases NR
NA
Eyes were
exposed via
modified swim
goggles
(Yorgancilar et al., 2012)
Rat
NR
No methanol control
NR
NR
NR
—
NR
(Yu and Blessing, 1997)
Rabbit
38% Formalin
No methanol control
Evaporation
None
NR
NA
A tube delivered
FA vapor to
rabbits' nares
(Yu and Blessing, 1999)
Rabbit
NR
No methanol control
NR
None
NR
NA
FA vapor puffed
in front of the
rabbits's nares
(Zhang et al., 2013)
Mouse
Formalin (10%)
No methanol control
NR
NR
NR
NA
Dynamic nose-
only
(Zhang et al., 2014b)
Rat
Formalin
No methanol control
Evaporation
NR
Reported but
questionable
NA
Static
(Zhou et al., 2006)
NR
No methanol control
NR
Formtector
NR
NA
NR
Rat
Methanol not measured
Zhou et al. (2011a)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
Static
Zhou et al. (2011b)
Rat
NR
No methanol control
NR
NR
Methanol not measured
NR
NA
Static
This document is a draft for review purposes only and does not constitute Agency policy.
A-260 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
FA - formaldehyde; HPLC - high performance liquid chromatography; IR - infrared; MMAD (og) - mass median aerodynamic diameter (geometric standard
deviation); NA - Not applicable; NR - not reported; PFA - paraformaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
A-261 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 A.5.2. Sensory Irritation
2 Literature Search
3 A systematic evaluation of the literature database on studies examining the potential for
4 sensory irritation in relation to formaldehyde exposure in humans was initially conducted in 2012,
5 with yearly updates (see A.1.1). The search strings used in specific databases are shown in
6 Table A-31. Additional search strategies included:
7 • A review of reference lists in the the articles identified through the full screening process
8 and
9 • A review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
10 EPA. 2010bl.
11 Symptoms of irritation in humans, primarily ocular, nasal, and throat symptoms, were the
12 focus of this review. Inclusion and exclusion criteria used in the screening step are described in
13 Table A-32. The search and screening strategy, including exclusion categories applied and the
14 number of articles excluded within each exclusion category, is summarized in Figure A-24. Based
15 on this process, 58 studies were identified and evaluated for consideration in the Toxicological
16 Review.
Table A-31. Summary of search terms for sensory irritation
Database,
search parameters
Terms
PubMed
No date restriction
(Formaldehyde[majr] OR paraformaldehyde[majr] OR formalin[majr]) AND
(irritation OR irritant OR irritants)
Web of Science
No date restriction
TS=(Formaldehyde OR paraformaldehyde OR formalin) AND TS=(irritation OR
irritant OR irritants)
Table A-32. Inclusion and exclusion criteria for studies of sensory irritation
Included
Excluded
Population
• Human
• Animals
Exposu re
• Indoor exposure via
inhalation to formaldehyde
• Measurements of
formaldehyde concentration
in air
• Not formaldehyde
• Dermal
• Exposure defined using job title/industry
• Outdoor exposure
Comparison
• Evaluated health outcomes
and associations with
formaldehyde exposure
• Case reports
• Surveillance analysis /Illness investigation
(no comparison)
This document is a draft for review purposes only and does not constitute Agency policy.
A-262 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Included
Excluded
Outcome
• Ocular, nasal and throat
symptoms
• Exposure studies/no outcome evaluated
• Studies evaluating other health outcomes
• Properties, uses
Other
• Reviews and reports (not primary
research), letters, meeting abstract, no
abstract, methodology paper, nonessential
article in a foreign language
This document is a draft for review purposes only and does not constitute Agency policy.
A-263 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Sensory Irritation (Human) Literature Search
"D
O
>¦
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
All articles identified for consideration in the literature search for sensory irritation were
evaluated to determine the degree of confidence in the reported results regarding the association of
formaldehyde inhalation with sensory irritation in humans. Observational epidemiology and
controlled human exposure studies were evaluated. The results of controlled human exposure
studies were considered to be relevant to the health assessment because irritation appears to be an
acute phenomenon rather than a time-dependent chronic response. Each study was evaluated for
precision and accuracy of exposure assessment, measurement of outcome, participant selection and
comparability, possibility of confounding, analysis and completeness of results, and study size.
Table A-33 provides criteria used to categorize the epidemiology studies. The accompanying tables
in this section document the evaluation. Studies are arranged alphabetically within each table.
Symptoms related to irritation in the eyes, nose, and throat were reported by most studies.
Generally, symptoms were ascertained via self-report or through interviews, both using a
standardized questionnaire (e.g., American Thoracic Society [ATS]). Generally, self-reported
symptoms will be influenced to some degree by recall bias if exposure is known to the responder,
although this is of less concern if an appropriate comparison is used. For some studies, there were
more serious concerns about selection or information bias related to the participants' knowledge of
their exposure or selection into a study based on presence of symptoms and concerns about
exposure, which could produce spurious findings (Salonen et al.. 2009: Ritchie and Lehnen. 1987:
Norsted etal.. 1985: Dally etal.. 1981)}(Wei etal., 2007; (Ritchie and Lehnen. 19851: Bracken et al.,
1985).
The time frame of the exposure assessment relative to the assessment of symptoms was an
important aspect of the evaluation of symptom prevalence. Questions about symptom occurrence
over an extended time period (weeks and months) that were separated in time from the exposure
assessment period were considered to be more limited by recall bias. This limitation was apparent
in some of the studies of anatomy students. The occupational studies generally ascertained the
prevalence of symptoms while at work via interview using standardized questionnaires.
Treatment of potential confounding by studies also was evaluated. EPA considered age,
gender, and smoking to be important confounders to evaluate for effects on sensory irritation. EPA
also looked for consideration of confounding by other irritants in the workplace, depending on the
occupational setting.
Table A-33. Criteria for categorizing study confidence in epidemiology studies
of sensory irritation
Confidence
Exposure
Study Design and Analysis
High
General population: Exposure measure
corresponds to appropriate time window for
outcome ascertainment (e.g., measures in
more than one season if time window covers
Instrument for data collection (e.g., ATS
questionnaire) described or reference provided.
Symptoms reported without knowledge of
exposure status. Assessment of symptoms
This document is a draft for review purposes only and does not constitute Agency policy.
A-265 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Confidence
Exposure
Study Design and Analysis
12 months, or addressed season in the
analysis). Exposure assessment designed to
characterize mean individual exposures
appropriate to analysis. Work settings:
Ability to differentiate between exposed and
unexposed, or between low and high
exposure.
timed concurrent with exposure assessment.
Analytic approach evaluating dose-response
relationship using analytic procedures that are
suitable for the type of data, and quantitative
results provided. Confounding considered and
addressed in design or analysis; large sample
size (number of cases).
Medium
General population: More limited exposure
assessment, or uncertainty regarding
correspondence between measured levels
and levels in the etiologically relevant time
window.
Work settings: Referent group may be
exposed to formaldehyde or to other
exposures affecting respiratory conditions
(potentially leading to attenuated risk
estimates)
Instrument for data collection less well
described. Symptoms reported without
knowledge of exposure status. Assessment of
symptoms timed concurrent with exposure
assessment. Analytic approach more limited;
confounding considered and addressed in
design or analysis but some questions regarding
degree of correlation between formaldehyde
and other exposures may remain. Sample size
may be a limitation.
Low
General population: Short (<1 day) exposure
measurement period without discussion of
protocol and quality control assessment.
High likelihood of confounding that prevents
differentiation of effect of formaldehyde from
effect of other exposure(s), limited data
analysis (or analysis that is not appropriate for
the data) or small sample size (number of
cases).
Not
informative
Exposure range does not allow meaningful
analysis of risks above 0.010 mg/m3; no
information provided.
Concern regarding selection bias with direction
away from null. Description of methods too
sparse to allow evaluation.
1 Controlled human exposure studies were evaluated for important attributes of
2 experimental studies, including randomization of exposure assignments, blinding of subjects and
3 investigators, and inclusion of a clean air control exposure and other aspects of the exposure
4 protocol. The evaluation of few individuals [n < 10) resulted in reduced confidence. Several studies
5 did not describe the measures used to control bias, resulting in a lower level of confidence in study
6 results. However, some of these studies evaluated multiple dose levels, an important strength for
7 the hazard assessment. Therefore, these studies were included with medium confidence when
8 reporting detail was the only identified limitation.
This document is a draft for review purposes only and does not constitute Agency policy.
A-266 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-34. Evaluation of studies examining sensory irritation in humans: residential studies
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Bracken et al.
(1985)
(Ontario)
Residential
(prevalence)
Exposed homes
randomly selected
from a group
currently being
monitored for
formaldehyde and
previously at
homeowner
request. Possible
selection bias.
Area samples; average of
3 hr samples; approx. 5
per home.
UFFI Mean 0.07, max
0.13 mg/m3; non-UFFI
Mean 0.06, max 0.12
mg/m3; Lab Mean 0.15,
max 7.2 mg/m3.
Limited sampling period,
details of sampling
protocol not provided.
Most samples may have
been below LOD (NIOSH,
1977, chromotropic)
Self-report,
ATS
question-
naire.
Response
was not
blinded to
presence of
UFFI.
Exposed: Homes
with UFFI,
Referent: non-
UFFI homes
from university
community; age
and smoking
prevalence
similar.
Symptom
prevalence
estimated from
graphs in Figures 1
and 2 in publication.
Compared
prevalence by
exposure group,
t-test
N = 54
exposed;
N = 26
referent
Overall
Confidence
Not
informative
Selection bias probable;
formaldehyde
concentration similar in
comparison groups
Dally et al.
(1981)
(Wisconsin)
Residential
(prevalence)
Survey of homes
reported to State
Division of Health
because of
symptoms;
potential for
selection bias
Area samples; average of
30-60 minute samples in
multiple locations. LOD
0.12 mg/m3
Mobile homes, Median
0.58, range <0.12 to 4.53
mg/m3.
Conventional, Median
0.12, range <0.12 to 1.34
mg/m3.
Limited sampling period.
Self-report,
questionnai
re.
Responses
blind to
formaldehy
de
measurem
ents.
No comparison
group; smoking
status was not
associated with
formaldehyde
concentration;
no adjusted
results provided
Symptom
prevalence among
exposed
N=256
SB IB Cf Oth
Overall
Confidence
Not
informative
N
No comparison group;
potential for selection bias;
limited statistical analyses
Hanrahan
et al.
(1984)
(Wisconsin)
Residential
(prevalence)
Recruited from a
randomly selected
list of mobile homes
in Wisconsin;
response rate 31%.
Concern is less
because
formaldehyde
concentrations, age,
Area samples; average of
1 hour samples from 2
rooms. Median 0.2
mg/m3, range <0.12 to
0.98 mg/m3
Limited sampling period
in closed residence with
no point formaldehyde
emissions; sampling and
Self-report,
questionnai
re, no
description
. Response
blind to
formaldehy
de
Logistic
regression
adjusting for
age, gender,
and smoking
status.
Logistic regression,
provided graph of
predicted mean
prevalence
normalized to mean
age, and upper and
lower 95% CI by
concentration from
regression model
N = 61
SB IB Cf Oth
Overall
Confidence
Medium
Limited sampling period;
Questionnaire not
described.
This document is a draft for review purposes only and does not constitute Agency policy.
A-267 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
and gender were
comparable to
nonrespondents,
and participants
blinded to
formaldehyde
concentration.
analytic protocols
referenced; LOD 0.12
mg/m3
measurem
ents.
Liu et al.
(1991):
Sexton et
al. (1986)
(California)
Residential
(prevalence)
Recruited from a
randomly selected,
age-stratified list of
mobile homes in
California; response
rate 44%. However,
the proportion of
respondents with
asthma was not
different from U.S.
prevalence in the
1980s (4.7% age-
adjusted; MMWR
Surveillance
Summaries; April
24, 1998 / 47(SS-
1); 1-28), suggesting
minimal concern for
selection bias.
Area samples using
passive monitors; 7-day
average in 2 rooms in 2
seasons. Mean summer
0.089 ppm, winter 0.088
ppm; TWA concentration
estimated using average
concentration multiplied
by # hours spent in the
home per day during the
week of sampling.
Validity study (Sexton
et al., 1986) reported
LOD of 0.01 ± 0.30 ppm;
range, LOD - 0.57 mg/m3
Self-report,
mailed
questionnai
re, no
description
Responses
blind to
formaldehy
de
measurem
ents.
Appropriat
e time
frame
relative to
exposure
measurem
ents.
Logistic
regression
adjusting for
age, gender,
smoking status,
status of chronic
respiratory
disease/allergy.
Logistic regression,
beta coefficients for
change in symptom
prevalence per
concentration
change were not
provided.
Prevalence
estimated from
graph of prevalence
by category of
formaldehyde TWA
exposure in
publication.
836
homes,
1096-
1394
individua
Is
SB IB
a oth
Overall
Confidence
Medium
h
Questionnaire not
described
Lovreglio et
al. (2009)
(prevalence)
Selection of 59
homes in city not
described.
24 hour samples in
kitchen in 59 homes;
reported mean, median,
range.
Self-report,
questionnai
re (onset of
symptoms
while in
kitchen).
Formaldehyde
and
acetaldehyde
concentrations
were correlated
(p=0.001).
Formaldehyde
concentrations
varied by
smoking status.
Data analyses
No data provided,
qualitative results
only.
182
subjects
living in
59
homes
SB IB
Cf Oth
Overall
Confidence
Not
informative
Results of data analysis
were not provided;
confounding by smoking or
co-exposure was not
addressed
This document is a draft for review purposes only and does not constitute Agency policy.
A-268 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
not described,
no adjustment
or stratification.
Main and
Hogan
(1983)
(prevalence)
Recruitment and
selection were not
described.
Three 1-hour area
samples using impingers
taken on 4 occasions
(August, September,
December, April) always
on a Monday. At least 1
sample was taken from
each office in both
trailers. Limited
sampling period in
closed residence with no
point formaldehyde
emissions; sampling and
analytic protocols
referenced; referent
group assumed to have
no exposure.
0.15-1.97 mg/m3
Self-report,
ATS
question-
naire,
symptom
history at
work
Potential
dissimilarity of
administrative
employees and
police officers
(healthier);
direction of bias
possibly away
from null; more
exposure to ETS
among referent;
possible
direction
toward null
Symptom
prevalence at work
compared between
exposed and
referent, chi-
square; small
sample size
Exposed
21,
Referent
18
SB IB
cf
Oth
Overall
Confidence
Low
¦
Potential dissimilarity
between comparison
groups; more exposure to
ETS among referent; small
sample size
Norsted et al.
(1985)
(Texas)
Residential
(prevalence)
Homes selected on
request of
residents; Possible
selection bias.
Sampling protocols not
described
Self-report;
symptom
reports not
blind to
exposure
status
No comparison
group; no
adjusted results
provided
Total # participants
in homes unknown.
443
mobile
homes
SB IB
Cf
Oth
Confidence
1 1
Not
n
informative
potential for selection bias;
Reporting deficiencies, no
comparisons
Olsen and
Dossing
(1982)
(Denmark)
Day care
center
workers in
Recruited from all
newly built mobile
day care centers in
2 boroughs (n = 7)
and 3 referent
centers selected at
random; response
rates 94% exposed,
Area samples; average of
2-hour samples in 2-4
locations, on 1 occasion.
Exposed mean 0.43,
range 0.24 to 0.55
mg/m3; referent mean
0.08, range 0.05 to 0.11
mg/m3; limited sampling
Self-report,
questionnai
re; linear
analogue
scale for
severity,
experience
within one
Referent
selected from
stationary child
care facilities in
same residential
area. Age and
smoking
prevalence
Prevalence and
severity presented
in graphs;
comparisons
between exposed
and referent groups
Exposed
= 66;
Referent
= 26
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Some uncertainties
regarding temporal
This document is a draft for review purposes only and does not constitute Agency policy.
A-269 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
mobile homes
(prevalence)
76% referent.
Responses similar in
exposed and
referent to 3
questions not
expected to be
related to
formaldehyde.
period in closed
residence with no point
formaldehyde emissions;
sampling and analytic
protocols referenced
month;
questionnai
re
described
and
citation
provided
similar in
exposed and
referent.
concordance of exposure
and symptom assessments
(Ritchie
and
Lehnen,
1987);
{Ritchie,
1985, 24726}
(Minnesota)
Residential
(prevalence)
Selection into
survey at request of
family physician;
potential for
selection bias;
however, health
responses were
blind to sampling
results
Area samples; average of
30-minute samples in 2
rooms.
Bedroom mean:
Mobile homes 0.43
mg/m3, Conventional
0.15 mg/m3, range 0.012
(LOD) to 6.79 mg/m3.
Limited sampling period
in closed residence with
no point formaldehyde
emissions; sampling &
analytic protocols
referenced;
Self-report,
interview;
symptoms
same day
as
exposure
measurem
ents,
respondent
s did not
know the
formaldehy
de
measurem
ent for
their
homes
Prevalence
stratified by
age, gender,
and smoking
status.
Presented graphs of
prevalence by
exposure (3
categories); tables
of prevalence (SE)
by type of home,
exposure category,
and smoking status
N =
2,000
residents
; 891
homes
SB IB Cf Oth
Overall
Confidence
Low
t
~
1 1
~
Potential for selection bias
Salonen et
al. (2009)
(Finland)
(prevalence)
Building selected
because of
complaints and
symptom reports of
occupants; possible
selection bias
Area sampling in 20 of
176 buildings selected
from database of Finnish
Institute of Occupational
Health, 2001 - 2006, N =
1 -12 per building;
during work hours 9-4
pm for 1-2 hours. LOD
0.5 ppb
Mean 0.011 mg/m3; Max
0.044 mg/m3.
Limited sampling period.
Self-report,
standardize
d
questionnai
re
No comparison
buildings
evaluated.
Compared
concentrations
to
recommended
indoor limit
(RIL)
Presented ratio of
average
concentration
divided by
recommended
indoor limit (based
on RD50 for
respiration rate in
mouse bioassay and
adjustment to 24
hours based on
Haber's Law.
20
buildings
SB IB Cf Oth
Overall
Confidence
Not
informative
Possible selection bias; no
comparison group
This document is a draft for review purposes only and does not constitute Agency policy.
A-270 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Thun et al.,
1991
(prevalence)
No information to
evaluate
No formaldehyde
measurements
Self-report,
questionnai
re; new
symptoms
over a one
year
period.
Exposed: Homes
with UFFI,
Referent:
homes without
UFFI. No
information to
compare
exposed and
referent
Data were not
provided,
qualitative results
with p-values
1,396
exposed,
1,395
referent
SB IE Cf Oth
Overall
Confide ncs
Not
informative
Inadequate reporting
detail; no formaldehyde
measurements
(Zhai et
al.. 2013)
Jan 2008-Dec
2009 (China)
(prevalence)
Provided criteria for
selection of homes
in defined area;
evaluated 186
homes in Shenyang,
China; homes were
decorated in last 4
years and occupied
within the last 3
years.
Cited Code for indoor
environmental pollution
control of civil building
engineering (GB50325-
2001); sampling period
not reported.
Samplers in breathing
zone in bedroom, living
room and kitchen; N =
558 in 186 homes;
exposure groups
polluted homes: >0.08
mg/m3, mean 0.09-0.13
mg/m3 in three rooms;
nonpolluted <0.08
mg/m3, mean
0.04-0.047 mg/m3.
Respiratory
symptoms
via
questionnai
re (ATS,
1978);
randomly
selected
one adult
from each
house, plus
82 children
(assisted by
parents)
Prevalence
ratios for
specific
symptoms/
disorders
unadjusted for
other variables,
characteristics
in two groups
not described;
regression
analyses of
combined
respiratory
symptoms were
adjusted
Compared symptom
prevalence for
children and adults
by exposure
category (reported
p-values);
multivariate logistic
regression of
respiratory system
symptoms (all) in
children and adults,
adjusting for age,
gender, smoking in
family, occupation,
education,
ventilation
frequency,
domestic pets,
house facing, family
history of allergy,
height, weight.
Polluted
homes
A/ = 119;
Nonpollu
ted
homes
N = 67
Symptom prevalence ratios
SB
IB
Cf Oth
Overall
Confidence
Medium
-
Sampling period not
reported
Analysis of combined
respiratory symptoms
Overall
SB 1
Cf
< >rh
Confidence
Medium
¦
1 1
This document is a draft for review purposes only and does not constitute Agency policy.
A-271 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-35. Evaluations of studies examining sensory irritation in humans: school-based studies
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness of
results
Size
Confidence
(Wantke et
al.. 1996b)
(Austria)
Schools
(panel,
intervention)
Children at school
where symptoms
were reported;
evaluated all
children attending
3 forms; low
concern for
selection
Area samples;
Sample number
and duration not
described; s.d. not
reported.
Concentration in 3
grades:
Before move:
0.053, 0.085,
0.092 mg/m3;
After move: 0.036,
0.028, 0.032
mg/m3
Symptoms
assessed before
and 3 months
after a move to a
different school
building.
Symptoms
reported by
parents in a
standardized
questionnaire.
Participants and
investigators not
blinded.
Comparison to self
before and after
removal from
exposure
Symptom prevalence
before and after
move; McNemartest
of difference
N = 62
SB IB a Oth
Overall
Confidence
Not
informative
Participants and
investigators not blinded;
Reporting deficiencies
Table A-36. Evaluations of studies examining sensory irritation in humans: controlled human exposure studies
Reference
Exposure assessment (quality
descriptor and exposures)
Outcome
classification
Consideration of
possible bias
(randomized exposure
order, blinding to
exposure)
Consideration
of likely
confounding
Resu Its
presentation
Size
(Andersen and
Molhave, 1983;
Andersen, 1979)
Confidence: Medium
Paraformaldehyde, dynamic
chamber, analytical
concentrations reported; 0.24,
0.4, 0.81,1.61 mg/m3
Self-report,
questionnaire;
symptom scores
Random assignment to
order of exposure, blinding
not described. 31.2%
smokers.
Within person
comparison
Provided
prevalence
N= 16
(Bender et al.,
1983)Confidence: Low
Paraformaldehyde, dynamic
chamber, analytical
concentrations not reported; 0,
0.43, 0.69, 0.86, 1.11, 1.23 mg/m3
Self-report
response (eye
only), time to 1st
response
Order of exposure
assignment not described,
blinding not described
Within person
comparison
Provided
prevalence
N=7
This document is a draft for review purposes only and does not constitute Agency policy.
A-272 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure assessment (quality
descriptor and exposures)
Outcome
classification
Consideration of
possible bias
(randomized exposure
order, blinding to
exposure)
Consideration
of likely
confounding
Resu Its
presentation
Size
Berglund et al.
(2012)
Confidence: High
Paraformaldehyde, analytical
concentrations reported; series of
18, 0.0078-1.23 mg/m3;
Nasal irritation (< 3
sec sniffs); Self-
report, forced
choice response
Exposure concentrations
randomly presented;
blinding not described.
Within person
comparison
Graph of
detection
prevalence by In
concentration
N= 31
Day et al., 1984
Not informative
Marginal; no clean air exposure,
1.23 mg/m3
Self-report,
questionnaire
Nonrandom exposure
assignment, blinding not
described
No comparisons
Provided
prevalence
00
T—1
II
£
Green et al. (1987)
Confidence: High
Paraformaldehyde, dynamic
chamber, analytical
concentrations reported; 0, 3.69
mg/m3
Self-report,
questionnaire;
symptom scores
Random assignment to
order of exposure, single
blinded.
Within person
comparison
Provided
prevalence &
statistical analyses
N=22
Green et al. (1989)
Confidence: High
Paraformaldehyde, dynamic
chamber, analytical
concentrations reported; 0, 3.69
mg/m3
Self-report,
questionnaire;
symptom scores
Random assignment to
order of exposure, double
blinded.
Within person
comparison
Provided score
data and statistical
analyses
graphically
'si-
rs
ii
£
James et al., 2002
Not informative
Emissions from materials in a
spacecraft simulator; analytical
concentrations reported; steady
state concentrations were not
achieved until end of study
?; 0.02 - 0.09 mg/m3
Self-report,
questionnaire
Nonrandom exposure
assignment, blinding not
described
Within person
comparison
?
N = A
Krakowiak et al.
(1998)
Not informative
Formalin, no methanol control;
analytic concentrations reported;
0.5 mg/m3
Self-report, diary;
symptom scores
Nonrandom exposure
assignment, single blinded.
Within person
comparison
Provided average
symptom scores
2
groups.
N = 10 in
each
Kulle (1993): Kulle et
al. (1987b)
Confidence: Medium
Paraformaldehyde, dynamic
chamber, analytical
concentrations reported; 1: 0,
0.62, 1.23, 2.46, II: 0,1.23 3.69
mg/m3
Self-report,
questionnaire;
symptom scores
Random assignment to
order of exposure, blinding
not described.
Within person
comparison
Regression
coefficients not
provided, only
p-values
1: N =10;
II: N =9
This document is a draft for review purposes only and does not constitute Agency policy.
A-273 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration of
possible bias
(randomized exposure
Consideration
Exposure assessment (quality
Outcome
order, blinding to
of likely
Resu Its
Reference
descriptor and exposures)
classification
exposure)
confounding
presentation
Size
Lang et al. (2008)
Paraformaldehyde, "quasi-static"
Self-report,
Random assignment to
Within person
Graphs/tables and
N=21
chamber conditions, analytical
questionnaire;
order of exposure, double
comparison
statistical analyses
Confidence: High
concentrations reported; 0, 0.19,
0.37, 0.62, peaks to 1.23 mg/m3
objective measures
blinded.
(Mueller et al., 2012)
Confidence: High
Paraformaldehyde, dynamic
chamber, analytical
concentrations reported; clean
air, 0.37 + 4 peaks of 0.74 mg/m3,
0.49 + 4 peaks of 0.98 mg/m3,
0.62 mg/m3 and 0.86 mg/m3
Self-report,
questionnaire;
objective measures
Exposure concentrations
randomly presented;
blinding not described.
Within person
comparison
Graphs of
difference
between pre- and
end of test values
N =41
Sauder et al. (1986)
Paraformaldehyde, dynamic
Self-report,
Nonrandom exposure
Within person
Provided average
N=9
Not informative
chamber, analytical
questionnaire;
assignment, blinding not
comparison
symptom scores &
concentrations reported; 0,3.69
mg/m3
symptom scores
described.
statistical analyses
Schachter et al.
Paraformaldehyde over boiling 2-
Self-report,
Random assignment to
Within person
Provided
N=15
(1986): Witeketal.
(1986)
propanol, dynamic chamber,
analytical concentrations reported
questionnaire;
symptom scores
order of exposure, double
blinded.
comparison
prevalence and
score
Confidence: Medium
Schachter et al.
Paraformaldehyde over boiling 2-
Self-report,
Random assignment to
Within person
Provided
N=15
(1987)
propanol, dynamic chamber,
analytical concentrations
questionnaire;
symptom scores
order of exposure, double
blinded. Participants had
comparison
prevalence and
scores
Confidence: Medium
reported.; 0, 2.46 mg/m3
routine occupational
formaldehyde exposure, N
= 2 smokers.
(Schuck et al., 1966)
Not informative
Propylene and ethylene
photooxidation with UV light; eye
exposure only; analytic
concentration reported
graphically; 0.12-1.23 mg/m3
Self-report,
questionnaire;
objective measures
Nonrandom exposure
assignment, blinding not
described
Within person
comparison
Graphs
N=12
Witeketal. (1987):
Witeketal. (1986)
Confidence: Medium
Paraformaldehyde over boiling 2-
propanol, dynamic chamber,
analytical concentrations
reported; 0, 2.46 mg/m3
Self-report,
questionnaire;
symptom scores
Random assignment to
order of exposure, double
blinded.
Within person
comparison
Provided
prevalence and
score
N=15
This document is a draft for review purposes only and does not constitute Agency policy.
A-274 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure assessment (quality
descriptor and exposures)
Outcome
classification
Consideration of
possible bias
(randomized exposure
order, blinding to
exposure)
Consideration
of likely
confounding
Resu Its
presentation
Size
(Yang et al., 2001) Not
informative
Plywood exposure; 2.03, 3.68, 5.3
mg/m3; eye exposure only;
Analytical concentrations
reported for formaldehyde but
not for other off gassed
compounds
Objective measure
Random assignment to
order of exposure, double
blinded. 25% smokers.
Within person
comparison
Graph of eye blink
frequency and
table of p-values
N= 8
Table A-37. Evaluation of studies examining sensory irritation in humans: anatomy courses
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
(Akbar-
Khanzadeh
et al.. 1994)
(Ohio)
Anatomy
students
(cross-sectional)
Participation not
reported.
TWA personal
breathing zone
samples obtained
on all exposed
subjects (9 days),
and 1 unexposed
(6 days).
Exposed mean
1.53, range 0.086
to 3.62 mg/m3.
Referent mean
0.12, range 0.09
to 0.17 mg/m3.
Self-report,
Medical Research
Council
standardized
questionnaire
No comparisons
reported.
Provided symptom
prevalence during
exposure, no
comparison to
baseline or to
unexposed; no
statistical data
analysis
34
exposed;
12
referent
SB IB a Oth
Overall
Confidence
Not
informative
No within person
comparison to baseline or
the referent; Reporting
deficiencies
(Chia et al..
1992)
(Singapore)
Anatomy
students
(cross-sectional)
Medical
students in 1st
year lab course
(92%
participation);
referent group =
3rd or 4th year
medical students
Area samples at
dissecting tables,
n=6, collected on
two occasions.
Personal
samples, n=14
students,
duration 2.5
Self-report,
modified MRC
standardized
questionnaire;
symptoms during
previous 4 weeks
of course (recall
Comparison to
referent
matched on age,
sex and
ethnicity
Symptom
prevalence in
exposed compared
to referent;
Referent activities
very different
Exposed
N = 150;
referent
N = 189
Overall
SB
IB
U
<>rh
Confidence
1
Low
¦
¦
I
Questions about dissimilarity
of 1st and 4th year students
and potential for recall bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-275 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
(participation
rate not
reported)
hours; mean
0.91, SD = 0.22
mg/m3, range
0.50 to 1.48
mg/m3, LOD =
0.062 mg/m3.
Assumed no
formaldehyde
exposure in
referent based on
activities (ward
rounds and
classroom).
accuracy
reduced?)
during previous 4 weeks of
course
(Fleisher.
1987)
Anatomy
students
(cross-sectional)
44% of 204
surveyed in
gross anatomy
course; of those
less than 50%
responded to
both
questionnaires.
Greater
motivation to
participate
among those
with symptoms?
Area samples in 6
labs, 1 day during
semester
(approximately 3
hours); Drager
tubes, 3 labs, LOD
1.23 mg/m3,
NIOSH method, 3
labs, LOD 0.02
mg/m3. Personal
breathing zone
for 2 instructors.
0.64, 0.18
mg/m3; probable
nondifferential
misclassification
due to sampling
method with low
sensitivity (3 labs)
and low
frequency of
sampling.
Adequate
differentiation
Self-report,
questionnaire;
data collection 1
month after end of
course; symptoms
all or some of the
time, rarely or
never, (temporal
gap reduced recall
accuracy?)
Within person
comparison:
symptoms
during lab with
exposure
compared to lab
with no
exposure to
formaldehyde.
Compared mean
symptom scores,
paired t-test
N = 38
SB IB a
Oth
Overall
Confidence
m
Low
Low response to both
questionnaires and selection
potential; temporal gap in
symptom response reduced
recall accuracy potential
This document is a draft for review purposes only and does not constitute Agency policy.
A-276 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
between
exposure groups
(Kriebel et
al.. 1993)
(Massachusetts)
Anatomy
students
(panel)
96%
participation
Personal samples
in the breathing
zone, 1-1.5
hours; multiple
days. Range
0.60-1.14
mg/m3,
geometric mean
= 0.9, SD 1.5
mg/mB
Self-report;
questionnaire
before, during and
immediately after
lab each day
Within person
comparison:
symptoms
during and after
lab compared to
prelab
symptoms.
Symptom
prevalence before,
during and after
lab. Mean prelab
and cross-lab
change over 10
weeks evaluated
using multivariate
linear regression
N=24
SB IB Cf Oth
Overall
Confidence
High
(Kriebel et
al.. 2001)
(Massachusetts)
Anatomy
students
(panel)
94.4%
participation;
attendance
declined from
n=37 to n=10
over 13 weeks
(better
attendance by
healthy
individuals?)
Individual TWA
using zone-
exposure matrix
based on
continuous
monitoring in six
homogenous
sampling zones
(LOD = 0.06
mg/m3). 12 min
work-zone
concentrations
calculated using
sampling data
and recorded
work; locations.
Mean 1.35, SD
0.69 mg/m3; 12
min peak 13.42
mg/m3
Self-report,
questionnaire;
symptom intensity
10-point scale
Within person
comparison:
symptoms
before and after
lab
Generalized
estimating
equation
regression
accounting for lack
of independence
of repeated
measures in
individuals;
symptom
intensity, %
change per ppm or
ppm-weeks
N=38
SB
IB Cf Oth
Overall
Confidence
Medium
N
This document is a draft for review purposes only and does not constitute Agency policy.
A-277 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
(Mori et al..
2016) (Japan)
Medical
students, 1st and
2nd year
Students (2nd
year) enrolled in
afternoon gross
anatomy classes,
April-July 2013,
mean age 22.9
yrs; compared to
nonexposed 1st
year students,
mean age 21.2
yrs. 75% males
Area sample, 5
locations during
class on same day
questionnaires
were completed.
Mean (SD) 0.1
(0.02) ppm
Questionnaire, 16
subjective
symptoms,
frequency never,
sometimes, or
often;
administered April
2013 before, May
2013 during, and
January 2014 6
months after
completion of
course.
Presented
characteristics
by exposure
group; adjusted
for age, sex and
allergy status in
regression
models.
Prevalence of
symptoms
compared,
Cochran's Qtest
and McNemar's
test; Regression of
presence or
absense of
symptoms in
relation to
exposure group on
day of survey,
controlling for
doctor-diagnosed
allergies, sex and
age
123
exposed
(98.4%);
114
unexpos
ed
(91.9%)
Overall
Confidence
High
(Saowakon
et al.. 2015)
(Tailand)
Medical students
and academic
staff
Students and
faculty in gross
anatomy
dissection labs;
Selection,
recruitment and
participation
was not
reported. Ages
19-21 yrs,
nonsmokers
with no history
of chronic
respiratory
disease or
symptomatic
illness
Personal
samplers (n=36
students, 4
instructors); area
samples, all
NIOSH-2016
method; 3-hr
samples over
duration of class,
3 classes,
January, August,
and October
Students:
Mean (SD) ppm
Class 1:
0.193 (0.120)
Class 2:
0.271 (0.159)
Class 3:
0.828(0.182)
Questionnaire, 20
symptoms,
completed before
start of dissection
and after chest
and abdominal
opening (classes 2
& 3); Severity
scale, 0-4.
Reported each
symptom as
percentage of
score for all
symptoms
averaged over all
classes; no
comparisons
N=36
students;
n=4
instruc-
tors
SB IB Cf Oth
Overall
Confidence
Not
informative
No within person
comparison to baseline or
the referent; reporting
deficiencies
This document is a draft for review purposes only and does not constitute Agency policy.
A-278 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
(Takahashi et
al.. 2007)
(Japan)
Medical students
(panel)
Did not report #
recruited versus
# that agreed to
complete
questionnaire.
Not clear if there
were refusals.
Area samples in 8
locations in lab, >
10 minutes;
Personal samples
(breathing zone)
on 18/143
students. Mean
3.0, SD = 0.60
mg/m3, range 2.2
to 4.6 mg/m3.
Self-report,
questionnaire
after 1st day and at
end of 2-month
course.
Within person
comparison:
symptoms after
1st day and at
end of course
Symptom
prevalence after
first day and after
lab at end of
course; McNemar
exact test
(estimated from
Figure 1 in
publication).
N=143
SB IB cf
Oth
Overall
Confidence
III
Medium
Large gap between symptom
ascertainment and exposure
measurements
(Takigawa et
al.. 2005)
(Japan)
Anatomy
students
(intervention)
Volunteers; 76%
completed
questionnaires
both before and
during lab
Area samples in 9
locations in lab, >
10 minutes.
Personal samples
on 24 of 78 in
phase I (2001)
(duration 42-962
minutes); median
3.3 mg/m3, range
2.2 to 8.9 mg/m3,
and on 46 of 79
in phase II (2004)
(duration
100-540
minutes); median
0.88 mg/m3,
range 0.40 to 3.4
mg/m3.
Self-report,
questionnaire
before and during
each course;
frequency (4-point
scale); score
change during
session
Groups similar
in age and %
male/female;
prevalence of
smoking not
reported.
Symptom change
index, 25
symptoms, by
phase of
intervention;
Mann-Whitney
test.
N = 78
Overall
Confidence
Medium
(Uba et al..
1989)
(California)
Anatomy
students
(panel)
78.6%
completed both
questionnaires
Personal
sampling
(impingers) in the
breathing zone
over 7 months;
multiple days;
TWA
concentration;
Self-report;
American Thoracic
Society
questionnaire;
symptoms after
lab on one day in
November (at
approx. 8-10
weeks); symptoms
Within person
comparison:
persistent
symptoms
beginning and
end of course (7
months); also
symptoms
during lab
Numbers with
symptoms in
exposed and
unexposed labs;
McNemar's test
paired samples,
OR, p-value.
N=81
SB IB Cf Oth
Overall
Confidence
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-279 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
range 0.06 to
1.14 mg/m3
before 1st day and
after last day (Sept
1984-Apr 1985)
session
compared to lab
with no
exposure to
formaldehyde.
(Wantke et
al.. 1996b)
(Austria)
Anatomy
students
(panel)
Volunteers;
participation
37.5% (45 of 120
students);
possibility of
selection bias
away from null
Area samples;
Continuous daily
measurements
for formaldehyde
at 2 locations
during 3-hour
lab, 5 days/ week
for 4 weeks.
Mean 0.15, range
0.07 to 0.27
mg/m3
Self-report,
standardized
questionnaire at
beginning
(symptoms during
3 months before
lab) and at end of
course (symptoms
over last 4 weeks),
(recall?)
Within person
comparison
Symptom
prevalence before
and during lab;
McNemar exact
test; multiple
measurements
during course
would be ideal
N =45
Overall
•SB
IB
: -
itth
Confidence
Low
¦
¦
1
Low participation, possibility
of selection bias away from
null; Potential recall issues -
symptoms for previous
weeks
Area samples;
Continuous daily
measurements
for formaldehyde
and phenol at 2
locations during
lab, exposures for
43 days. Mean
0.27, range 0.13
to 0.41 mg/m3
(Wantke et
al.. 2000)
Austria
Anatomy
students
(panel)
Selection was
not described;
27 of the 45
students in
Wantke et al.,
1996
Self-report,
questionnaire at
beginning, 5 weeks
and 10 weeks,
Daily symptom
cards during class.
Within person
comparison;
symptoms at
beginning and
during lab at
middle and end
of 10-week
course
Symptom
prevalence before,
middle and at end
of 10 week course;
McNemar exact
test
N = 27
SB
IB a Oth
Overall
Confidence
Medium
h
This document is a draft for review purposes only and does not constitute Agency policy.
A-280 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration
of participant
selection and
comparability
Exposu re
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
resu Its
Size
Confidence
(Wei et al..
2007)
Anatomy
students
(cross-sectional)
Volunteer, all
students present
on the day that
sampling was
conducted;
symptom
questionnaire
was not
completed
outside of class
so difference
may have been
influenced by
perception
relative to
symptoms in
class (possibly
resulting in
overestimation
of risk)
Area samples
near dissection
tables, 30 minute
samples, N = 12.
Measurements
before,
beginning, middle
and completion
of 3-month gross
anatomy class.
Geometric mean:
before 0.03,
beginning 0.89,
middle 0.76, end
0.24 mg/m3
(medium)
Self-report,
questionnaire on
sampling days
after 2 hours of lab
(medium)
Within person
comparison
(high)
Frequency of
symptoms during
class; prevalence
and severity scores
during class
compared to
"usual life
situation"; Walsh
test (inadequate
comparison)
N = 79-
94
SB IB Cf Oth
Overall
Confidence
Not
informative
This document is a draft for review purposes only and does not constitute Agency policy.
A-281 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-38. Evaluations of studies examining sensory irritation in humans: occupational studies
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
(Alexanders
son et al..
1982)
(prevalence)
All exposed workers
employed >1 yr;
evaluated
employees present
at work on study
day (both exposed
and referent);
Selection for
healthy survivors
TWA personal
sampling for
formaldehyde,
terpenes & dust,
N=31; 1 working
day, 6-7 hours
0.05-1.62
mg/m3; no
measurements
for referent
group; Although
no
measurements
in referent, high
concentration in
exposed allows
assumption of
an adequate
exposure
contrast for
comparison of
exposed and
referent
Self-report,
British Medical
Research Council
questionnaire;
symptoms at
work, same day
as exposure
assessment
Symptom
prevalence in
exposed
compared to
referent.
Exposed:
employees of
carpentry works;
referents were
not exposed to
formaldehyde or
other irritants in
same factory;
Similar % age,
height, sex, &
weight.
Prevalence
smoking 48% in
exposed, 40% in
referent.
Symptom
prevalence at work
compared between
exposed and
referent, chi-square
N=47
exposed;
N=20
referent
SB IB Cf Oth
Overall
Confidence
Low
Healthy survivor bias
(Alexanders
son and
Hedenstiern
a. 1989)
(prevalence,
follow-up of
(Alexanders
son et al..
1982)
Evaluated
employees who
participated in
previous study, 4 yr
follow-up
(Alexandersson
et al.. 1982); 13
exposed and 2
referents lost-to-
follow-up; 13
exposed transferred
to unexposed jobs
TWA using
personal
sampling, 3-4
15 minute
samples/person;
2 working days;
Mean 0.5
mg/m3; Mean
peak 0.69
mg/m3 limited
sampling period;
although no
measurements
Self-report,
British Medical
Research Council
questionnaire
Symptom
prevalence in
exposed
compared to
referent.
Exposed:
employees of
carpentry works;
referents were
not exposed to
formaldehyde or
other irritants in
same factory;
Change in symptom
prevalence at work
1980-1984, chi-
square
N=21
exposed;
N=18
referent
SB
IB
Cf
Oth
Overall
Confidence
Low
H P
Healthy survivor bias;
confounding by smoking
This document is a draft for review purposes only and does not constitute Agency policy.
A-282 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
possible survivor
bias
in referent, high
concentration in
exposed allows
assumption of
an adequate
exposure
contrast for
comparison of
exposed and
referent
Similar % age,
height, sex, &
weight.
Prevalence
smoking 50% in
exposed, 33% in
referent.
Moderate
concern for
confounding by
smoking
(direction of bias
unclear).
(Alexanders
son and
Hedenstiern
a. 1988)
(prevalence)
Selection for
healthy; evaluated
employees present
at work on study
day (both exposed
and referent)
TWA using
personal
sampling, 3-4
15 minute
samples/person;
1 working day,
no
concentration
reported for
referent
0.12-1.32
mg/m3
Although no
measurements
in referent, high
concentration in
exposed allows
assumption of
an adequate
exposure
contrast for
comparison of
exposed and
referent
Self-report,
standardized
questionnaire;
outcome
assessed same
day as exposure
Symptom
prevalence
among workers
exposed to acid-
hardening
lacquers;
referents were
"nonexposed"
employees at
same factory. All
male, exposed
slightly younger,
50% smokers;
referent: 33%
smokers.
Sampled for dust
and solvents:
authors
considered all
exposures to be
very low and not
confounders.
Moderate
concern for
confounding by
Symptom
prevalence at work
compared between
exposed and
referent, chi-
square; no
adjustment
N=38
exposed;
N=18
referent
SB IB Cf Oth
Overall
Confidence
Low
H H
Confounding and no
adjustment in analyses
This document is a draft for review purposes only and does not constitute Agency policy.
A-283 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
Participation >90%
in exposed, >80% in
referent; Healthy
survivor effect likely
similar among
exposed and
referent groups
smoking
(direction of bias
unclear).
(Herbert et
al.. 1994)
(prevalence)
TWA continuous
sample in
breathing zone;
5 sites, 2 days
0.09 - 0.33
mg/m3 referent
not reported;
sampled for
dust. Although
no
measurements
in referent,
formaldehyde
exposure not
expected for oil/
gas field
workers,
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Self-report,
Respiratory
symptoms
ascertained via
interview using
standardized
questionnaire
Possible
respiratory
irritants in
comparison
group (oil sands
workers); higher
prevalence of
smokers (52% vs
28%) and shorter
duration of
employment
among exposed,
(5 versus 10
years)
Symptom
prevalence
compared by
exposure group,
chi-square;
unadjusted analyses
N=99
exposed;
N=165
referent
SB IB
Cf
Oth
Overall
Confidence
Low
¦
Different prevalence
smoking and duration of
employment between
exposed and referent; no
adjustment in analyses
(Holmstrom
and
Wilhelmsso
n. 1988):
(Wilhelmsso
n and
Holmstrom.
100% participation;
healthy survivor
bias probable;
source populations
for exposed and
referent
(government clerks)
were different,
raising possible
unmeasured
confounding
Area samples in
one group,
1979-1984,
personal
samples (1-2
hours) in 1985
in all groups.
Sampling data in
referent.
0.05-0.5 mg/m3
Self-report,
questionnaire
Groups similar for
age and smoking,
87% and 93%
male in exposed,
56% male in
referent (gender
related
differences in
perception of
irritation?) No
exposure to
Compared
symptoms
prevalence across
exposure groups,
chi-square;
unadjusted analyses
N=70
Group 1,
N=100
Group 2;
N=36
referent
as
Overall
Confidence
Low
Healthy survivor bias;
groups selected from
different source
populations; Potential
confounding and no
adjustment in analyses
This document is a draft for review purposes only and does not constitute Agency policy.
A-284 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
1992)
(prevalence)
Adequate
exposure
contrast for
comparison of
exposed and
referent
solvents,
concentrations
for other
chemicals all <1%
of OEL (phenol,
ammonia,
epichlorhydrin,
methanol and
ethanol).
(Holmstrom
etal.. 1991)
(prevalence)
Details of
recruitment and
participation not
described. Healthy
survivor bias
probable; source
populations for
exposed and
referent were
different, raising
possible
unmeasured
confounding
Personal
exposure
measurements
stable through
year, average
0.2 - 0.3 mg/m3,
peaks seldom >
0.5 mg/m3
Formaldehyde
Concentration,
mean
MDF0.26
mg/m3,
wood dust 0.25
mg/m3,
referent 0.09
mg/m3;
adequate
exposure
contrast for
comparison of
exposed and
referent
Self-report,
questionnaire
MDF group
slightly older
(44.1 yr)
compared to
wood (39.3 yr)
and referent
(39.9 yr); % male
varied, smoking
less prevalent in
referent
Exposed groups
each compared to
referent;
prevalence rate
difference, 95%
confidence
intervals; no
adjustment
MDF:
N=16
Wood:
N=29
Referent:
N=36
KM
Overall
Confidence
Low
Healthy survivor bias;
groups selected from
different source
populations; Potential
confounding and no
adjustment in analyses
(Holness
and
Nethercott,
Minimal concern for
selection bias.
Recruitment source
was list provided by
funeral home
2 area samples
(impingers),
during
embalming, 30
to 180 minutes.
Self-report,
American
Thoracic Society
questionnaire;
Symptom
prevalence
compared
between exposed
(apprentice
Comparisons
between exposed
and referent,
logistic regression
adjusted for # pack-
N=84
exposed;
N=38
referent
SB IB
Cf
Oth
Overall
Confidence
Medium
¦
This document is a draft for review purposes only and does not constitute Agency policy.
A-285 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
1989)
(prevalence)
association, 86.6%
of eligible
participated.
Participation rate
among referents
was not given.
Gave
concentration
for referent.
0.1-1.0 mg/m3
Adequate
exposure
contrast for
comparison of
exposed and
referent
before and after
embalming
funeral service
workers) and
unexposed
(service
volunteers and
paid students),
probable
unmeasured
confounders.
Groups similar for
age, height, and
smoking status.
Source of
formaldehyde
exposure was
formalin (also
contained
methanol)
years smoked.
Provided data and
results of statistical
analyses
Groups selected from
different source
populations
(Horvath et
al.. 1988)
(Wisconsin)
Occupational
(prevalence)
71% participation in
exposed; 88%
participation in
referent. Age and
sex distribution in
participants similar
to entire workforce
in their respective
companies.
Evaluated and ruled
out survivor bias
using reasons for
leaving employment
among 54 former
employees;
evaluated
characteristics of
30/45
nonparticipants
8-hour TWA
using Personal
and area
sampling on day
of exam.
Exposed mean
1.04, range 0.32
to 4.48 mg/m3.
Referent mean
0.06, range
0.04-0.15
mg/m3;
adequate
exposure
contrast for
comparison of
exposed and
referent
Self-report,
American
Thoracic Society
questionnaire;
assessed same
day as exposure
assessment;
before and after
shift
Symptom
prevalence in
exposed workers
at a particleboard
manufacturing
plant compared
to referent
workers at 2 food
production
plants. Higher
proportion male
in exposed and
slightly older
average age
(expect bias
toward null for
symptoms).
Smoking and
mobile home
Symptom
prevalence during
work in exposed
and referent
compared;
prevalence at end
of shift using
multiple regression
with adjustment
N=109
exposed;
N=254
referent
Overall
Confidence
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-286 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
who were younger
and higher % male,
with similar %
smokers and mobile
home residency.
residency similar.
Particulate
exposure in
exposed and
referent
(different
sources), other
chemical
exposures were
not detectable or
below PEL.
Self-report,
questionnaire,
composite
experience for
previous months
or years (reduced
accuracy of recall,
possible recall
bias)
(Kilburn et
al.. 1985)
(prevalence)
97% participation
among exposed.
Environmental
samples for
formaldehyde,
xylene, toluene,
and chloroform
by regional
NIOSH
laboratory in 10
of 25 labs; 1-4
hours sampling
time; self-report
of duration of
exposure
(hours/day)
0.25-2.34
mg/m3;
adequate
exposure
contrast for
comparison of
exposed and
referent
Incomplete
matching: Among
76 exposed,
group of 40
matched to
referent on age,
cigarette
smoking, and
ethnicity;
multiple chemical
exposures;
evaluated effects
among
participants with
>4 hours
formaldehyde
exposure/day
stratified by 2
levels for xylene.
Prevalence by hours
formaldehyde
exposure and
xylene exposure;
results of statistical
analyses not shown
N=76
exposed;
N=56
referent
SB IB
Cf
Oth
Overall
Confidence
Low
Reduced accuracy of
recall; incomplete
matching
(Lofstedt et
al.. 2011)
(prevalence)
>90 % participation
in exposed and
referent; healthy
worker survival?
Individual
samples over a
single 8-hour
shift
Self-report,
questionnaire;
existence of
symptoms during
Referent from
the same
industry (not
workers in core
Logistic regression
models, symptoms
by referent, low and
high formaldehyde
N=43 of
48
exposed;
This document is a draft for review purposes only and does not constitute Agency policy.
A-287 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting and
design
Consideration of
participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration of
likely
confounding
Analysis and
completeness of
results
Size
Confidence
Higher proportion
of referents had
ever had asthma or
allergic symptoms
in childhood
0.013-0.19
mg/m3,
geometric mean
0.037 mg/m3;
subjects
categorized into
low and high
formaldehyde
using LOD; also
sampled MCA,
ICA and dust
prior week
(reduced recall
accuracy? and
potential for
recall bias)
production or die
casting),
comparable for
age; smoking
prevalence,
prevalence
female, and work
duration higher in
referent.
Symptom
prevalence
compared
between groups.
Co-exposures
measured but not
adjusted for in
analysis.
Independent
effect of
formaldehyde
could not be
determined
groups; no
adjustment for
other irritants
(isocyanic acid,
methyl isocyanate,
dust) which were
strongly associated
with symptoms.
Also restricted
analyses excluding
asthma or allergies,
females, or
smokerswith similar
results
N=69 of
84
referents
Could not distinguish
effect of formaldehyde
from those of other
irritants that were
strongly associated with
symptoms; Potential for
information bias (reduced
recall accuracy); potential
health worker survival
(Neghab et
al.. 2011)
(prevalence)
100% participation;
healthy worker
survival?
Area samples
(40 minutes,
N=7) in 7
workshops and
1 in office area.
Mean 0.96
mg/m3; SD 0.49
mg/m3;
adequate
exposure
contrast for
comparison of
exposed and
referent
Self-report,
interview &
American
Thoracic Society
questionnaire;
symptoms at
work
Referent from
the same
industry and
comparable for
socioeconomic
status, age,
smoking
prevalence (25%).
Symptom
prevalence
compared
between groups.
Symptom
prevalence
compared by
exposure group,
chi-square
N=70
exposed
N=24
referents
SB
IB Cf Oth
Overall
Confidence
Medium
h
Healthy survivor bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-288 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Supporting Material for Hazard Analyses of Sensory Irritation
Table A-39. Summary of epidemiology studies of laboratory exposures to
formaldehyde and human sensory irritation
(Reference), study design, exposure levels
Results
(Kriebel et al.. 1993) (Massachusetts)
Panel study, 24 clinical anatomy students dissecting cadavers
during 10-week lab once a week, 3 hours. Outcome: Symptoms
recorded before, during and after the lab; ATS questionnaire for
baseline and modified brief questionnaire during lab,
references provided.
Exposure: Personal samples in breathing zone (1- to 1.5-hour
duration).
Geometric mean 0.73 ppm (SD 1.22 ppm). Range 0.49-0.93
ppm (n=8). No trend in concentrations over semester.
Formaldehyde levels in three air samples in the cavities of the
cadavers were 3.0, 3.6 and 4.3 ppm.
Analysis: Multivariate linear regression models; mean prelab
and cross-lab change in symptoms analyzed using random
effects models.
Average symptom prevalence increased from
beginning to end of weekly lab session by 43%.
Prevalence (%) Before, Midway and After Lab
Session
Symptom
pre
mid-
Post
Eyes
16
66
59
Nose
46
75
67
Throat
25
45
40
Breathing
16
41
36
Cough
15
26
20
SB IB Cf Oth
Overall
Confidence
High
Analysis indicated that magnitude of increase in
symptom prevalence across lab session
decreased as semester advanced (In week: eye (5
-0.74, p = 0.002; throat (5 -0.39, p = 0.03; nose (5 -
0.64, p = 0.06).
No trend in prelab symptom severity over 10-
week course
(Uba et al.. 1989) (California)
Panel study, 1984-1985.
103 of 142 medical students in a 7-month anatomy class,
meeting twice a week for 4 hours (September 1984-April 1985),
mean age (range): 24.3 (21-33) years.
Outcome: Persistent symptoms: 103 students completed
respiratory questionnaire (ATS) at the beginning (September
1984) and end of course (April 1985). Acute symptoms: 81/103
students completed different questionnaire after anatomy lab
with formaldehyde exposure and after microanatomy lab (no
formaldehyde) during Nov 1984. Order of questionnaires
varied.
Exposure: Personal samplers (impingers) in the breathing zone.
TWA formaldehyde concentrations (N = 32 samples during
different class periods over 7 months). Short-term samples (N
= 16) for peak concentrations during dissection and
observation. Dissecting room ventilated 24 hours/day
TWA concentrations: range, < 0.05 (LOD) to 0.93 ppm (< 0.06 to
1.1 mg/m3).
During dissection: mean 1.9 ppm (2.3 mg/m3); range 0.1 to 5.0
ppm (0.12 to 6.1 mg/m3).
Symptoms during lab session: symptom
prevalence in anatomy lab (exposed)
compared with microanatomy lab
(unexposed) (N = 81)
Symptom
Ex- Unexposed Odds
posed Ratio
Itchy eyes
Watery
eyes
Burning
eyes
Burning
nose
Sore
throat
Sneezing
33
36
47
19
21
10
Rhinorrhea 13
Chest 4
tightness
Cough 5
Wheezing 2
33*
12*
infinite
infinite
5.3**
10**
4.3**
infinite
1.3
infinite
This document is a draft for review purposes only and does not constitute Agency policy.
A-289 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
When observing dissection: mean 1.2 ppm (1.5 mg/m3); range
0.2 to 2.0 ppm (0.25 to 2.5 mg/m3).
Monthly average in September, October, and May: 0.6, 0.8, 0.1
ppm (0.74, 0.98, and 0.12 mg/m3).
Analysis: Symptom prevalence at beginning of course
compared to end of course, paired analysis, McNemar's test;
symptom prevalence after lab with formaldehyde compared to
lab with no formaldehyde, odds ratios, McNemar's test paired
samples
Dyspnea
0
infinite
McNemar's test paired samples, * p<0.001;
**p<0.05
Persistent symptoms (Number reporting
symptoms only in September 1984 or only
in April 1985)
SB IB Cf Oth
Overall
Confidence
High
Symptom
Sept.
1984
April
1985
Odds
Ratio
Cough
1
8
00
o
*
Phlegm
4
9
2.3
Chronic
4
2
0.5
bronchitis
Chest
9
0
Q**
illnesses
Wheezing
37
1
0.03**
Wheezing
4
0
o***
with Dyspnea
Dyspnea on
0
0
-
exertion
McNemar's test paired samples, * p = 0.02;
**p<0.001; ***p = 0.05
(Mori et al.. 2016)
(Japan)
Cross-sectional study, Students (2nd year), n=123 (98.4%)
enrolled in afternoon gross anatomy classes, April-July 2013,
mean age 22.9 yrs; compared to nonexposed 1st year students,
n=114 (91.9%), mean age 21.2 yrs. 75% males
Outcome: Questionnaire, 16 subjective symptoms, frequency
never, sometimes, or often; administered April 2013 before,
May 2013 during, and January 2014 6 months after completion
of course.
Exposure: Area samples at breathing height, 5 locations during
class in May 2013 on same day questionnaires were
completed. Mean (SD) 0.123 (0.025) mg/m3 (conversion by
EPA).
Area sample, 5 locations during class on same day
questionnaires were completed.
Mean (SD) 0.1 (0.02) ppm
Analysis: Regression of presence or absense of symptoms in
relation to exposure group on day of survey, controlling for
doctor-diagnosed allergies, sex and age
Symptoms reported comparing exposed to
unexposed on a day during gross anatomy class
(OR (95% CI))
Symptom
OR
95% CI
Eye soreness
2.35
1.3-4.27
Eye strain
1.82
1.07-3.14
Itchy eye
0.75
0.43-1.31
Dry eye
1.11
0.63-1.96
Tearing
2.62
1.36-5.04
Itchy nose
1.76
1.01-3.06
Nasal
0.78
0.44-1.36
Runny nose
0.82
0.47-1.44
Sore throat
1.45
0.82-2.55
Dry throat
0.87
0.49-1.57
SB IB Cf Oth
Overall
Confidence
High
(Kriebel et al.. 2001) (Massachusetts)
Mean postlab intensity of eye, nose, and throat
irritation decreased over semester.
This document is a draft for review purposes only and does not constitute Agency policy.
A-290 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
Panel study, 38 anatomy students (of 54 total) during 12-week
class meeting once per week, 2.5 hours. Mean age 24.9 years,
23.7% male, 2 current smokers, 5 ex-smokers, 4 history of
asthma
Outcome: Symptom questionnaires before and after each lab
session. Scale of symptom intensity ranged from 0 (not at all)
to 10 (very, very much).
Exposure: Continuous monitoring in six homogenous locations
(LOD = 0.05 ppm [0.06 mg/m3). 12-minute work-zone
concentrations for each student calculated using sampling data
and recorded work locations.
Geometric mean concentration over all lab sessions and
participants: 0.7 ppm [0.9 mg/m3] (GSD 2.13)
Peak 12 min concentration 10.91 ppm (13.42 mg/m3)
Average ± SD concentration over all weeks and participants: 1.1
± 0.56 ppm (1.4 ± 0.69 mg/m3)
Concentrations decreased over 12-week semester.
Analysis: Generalized estimating equation regression model
accounting for lack of independence of repeated measures in
individuals.
Association of symptom intensity with
exposure during lab & interaction with
time (Percent change in intensity per
ppm or ppm-weeks)
Recent
Recent
exposure0 exposure
ln(week)c
Eye 1.22*
Irritation
Nose 1.09*
Irritation
Throat 0.81*
Irritation
-0.35*
-0.42*
-0.36*
SB
IB Cf Oth
Overall
Confidence
Medium
N
p <0.001 for significant deviation from
slope = 0
bMean concentration during 2.5-hour
lab
c Interaction between recent exposure
and natural log of week number,
indicating declining strength of
association with time.
Attendance declined from n=37 to n=10 over 13 weeks (better
attendance by healthy individuals?)
(Takahashi et al.. 2007) (Japan)
Panel study, 2002-2003.
143 medical students (68.5% male, 88.8% 20-24 years of age)
who dissected cadavers 15 hours per week for 2 months and 76
students who had taken same course 2 to 4 years earlier (68.4%
male, 77.6% 20-24 years of age).
Outcome: Symptom questionnaire administered after 1st day of
exposure and at end of course.
Exposure: Area formaldehyde samples (> 10 minutes, 8
locations in room), upon opening of thorax, mean 2.12 ppm (SD
0.23), range 1.7-2.44 ppm (2.6 ± 0.28 mg/m3, range 2.13-3.05
mg/m3). Breathing zone samples (18/143 students), mean 2.4
ppm (SD 0.49), range 1.79-3.78 ppm; (mean 3.0 ± 0.61 mg/m3,
range 2.24-4.72 mg/m3)
Analysis: Prevalence after first exposure and at end of course
compared, McNemar's test
Prevalence after first exposure and at end of
course estimated from Figure 1 in the paper.
Largest increase in symptoms (p<0.05) reported
for eye soreness (from about 35% to about 68%
on 1st day versus end of course), lacrimation (12%
to 60%), throat irritation (14% to 42%), eye
fatigue (28% to 44%), rhinorrhea (17% to 35%),
skin irritation (14% to 28%).
SB IB cf
Oth
Overall
Confidence
Medium
Large gap between symptom ascertainment and exposure
measurements
This document is a draft for review purposes only and does not constitute Agency policy.
A-291 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
(Takigawa et al.. 2005)
(Japan)
Intervention study, purpose: Evaluate installation of a
ventilation system between phases and effects on
formaldehyde concentrations and symptoms. 2 phases; 1st
phase: 78 volunteer anatomy students in 2001 (mean age 21.6
years); 2nd phase: 79 volunteer anatomy students 3 years later
in 2004 (mean age 21.7 years).
Outcome: Self-administered questionnaires on health
complaints before and during each two-month course.
Symptom frequency: 1 (never), 2 (scarcely), 3 (sometimes), and
4 (always). Symptom change index: Symptom frequency score
during session subtracted from score before course.
Exposure: Area formaldehyde samples (>10 minutes, 9
locations in room); upon opening of thorax (represents highest
concentration over 2 months).
Phase I: Median (range) 2.59 (2.1-3.0) mg/m3 (concentration
reported as 0.259 mg/m3 in Table 3 of the paper must be an
error).
Phase II: Median (range) 0.729 (0.291-0.971) mg/m3
Personal samples (measured with gas sampler on 24 students in
first phase (42-962 min) and 46 in second phase (100-540 min)):
Phase I: Median (range) 3.313 (2.238-8.909) mg/m3
Phase II: Median (range) 0.878 (0.396-3.386) mg/m3
Analysis: Symptom change index, 1st and 2nd phases compared;
Mann-Whitney test, p <0.05.
SB IB Cf Oth
Overall
Confidence
Medium
Large gap between symptom ascertainment and exposure
measurements
Symptom change indexes for 8 of 25 measured
symptoms were significantly less comparing the
second phase results with the first phase results.
Symptom Change Index
Symptom
1st
(N=78)
2nd
(N=79)
Skin
Eczema
0.13
-0.09
Eye
Itchy
0.74
0.27
Irritated
0.96
0.52
Watery
1.42
0.46
Poor vision
0.17
-0.27
Nose
Itchy
0.67
0.22
Changed
0.18
0.33
sense smell
Throat
Sore
0.69
0.22
(Wantke et al.. 2000) (Austria)
Panel study, 27 medical students, participants in Wantke et al.
(1996) enrolled in a 2nd dissection class, 55.6% male
Outcome: Symptoms standardized questionnaire at beginning,
in middle, and at end of 10-week course. Daily symptom cards
during class
Exposure: Continuous measurements for formaldehyde and
phenol at 2 locations during lab, exposures for 43 days
Formaldehyde Mean 0.265 ± 0.07 mg/m3, range 0.133-0.410
mg/m3,
Phenol Mean 4.65 ± 2.96 mg/m3, range 0.09-11.8 mg/m3
Analysis: Prevalence in November and December compared to
October, McNemar exact test
Symptom prevalence was not correlated with
smoking, or type I allergy, complaints of dizziness
occurred only in males
Prevalence of Symptoms at Beginning,
Middle (5 Weeks) and End (10 Weeks) of
Course
Symptoms Before Middle End
Burning
eyes
Sneezing
Nosebleed
Cough
Shortness
of breath
0.111 0.481*
0.074
0.185
0.074
0
0.037
0.111
0.148
0.185
0.333*
0.037
0.185
0.074
0.037
This document is a draft for review purposes only and does not constitute Agency policy.
A-292 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
SB
IB a Oth
Overall
Confidence
Medium
h
*p<0.05, **p<0.01
See Wantke et al., 1996
(Wantke et al.. 1996b) (Austria)
Panel study, 1995. 45 medical students enrolled in 1st dissection
class, 51.1% male, age 20.9 years,
3 hour sessions, 5 days/week for 4 weeks
Outcome: Symptoms, standardized questionnaire at beginning
and at end of 4-week course
Exposure: Continuous measurements for formaldehyde, 2
locations during lab; Mean 0.124 ± 0.05 ppm, range
0.059-0.219 ppm
No sampling for phenol
Analysis: Compared symptom prevalence during course to
before, McNemar exact test
Prevalence of Symptoms During 4 Week
Course
Symptoms Before During p-
Value
Burning
eyes
Sneezing
Nosebleeds
Cough
Shortness
of breath
0.133 0.289 <0.02
0.244
0.089
NS
0.244
0.044
NS
0.044
0
NS
0
0.022
NS
sa is cf
Oth
Overall
Confidence
III
Low
Symptom prevalence was not correlated with
gender, smoking, or type I allergy.
Low participation, possibility of selection bias away from null;
Potential recall issues - symptoms for previous weeks
(Chia et al.. 1992) (Singapore)
Cross-sectional study. 1st year medical students in anatomy lab,
150 of 164 total (91.5%); referent 189 3rd and 4th year medical
students, no recent formaldehyde exposure; matched on age,
sex, and ethnicity.
Outcome: Symptoms during previous 4 weeks of anatomy
course (twice per week, 2.5 hr (or other activities for referent),
assessed via a modified MRC standardized questionnaire
Exposure: Area samples at dissecting tables, n=6, collected on
two occasions, Mean (SD) 0.5 ppm (0.08), range 0.4-0.6 ppm
Personal samples, n=14 students, duration 2.5 hours, Mean (SD)
0.74 (0.18), range 0.41-1.2 ppm
LOD 0.05 ppm
Analysis: Symptom prevalence in exposed compared to
referent
Prevalence of Symptoms
Symptom Ex- Refer- p-
posed ent Value
(n = 150) (n
189)
Decreased 0.127
ability to
smell
Eye 0.8
irritation
Throat 0.313
irritation
Dry mouth 0.18
0.032 0.002
0.132
0.138
0.058
<
0.001
<
0.001
<
0.001
Overall
Confidence
Low
Questions about dissimilarity of 1st and 4th year students and
potential for recall bias during previous 4 weeks of course
No statistically significant difference for other
symptoms (cough with mucus, chest tightness,
chest pain, and breathlessness) (data were not
reported).
(Fleisher, 1987) (New York)
Cross-sectional study
1st year medical students (N = 89) (43.6% of total 204 surveyed)
(71% male) in gross anatomy course (formaldehyde exposed).
Symptoms prevalence (% reporting symptom all
or some of the time) among 38 students
responding to both questionnaires (N=38)
This document is a draft for review purposes only and does not constitute Agency policy.
A-293 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
Referent: Same students (n=60) (72% male) in
Symptom
Anatomy
Path/
pathology/microbiology laboratory six months later. 98.9% of
Micro
all students attended 75-100% of all lab sessions.
Eye Irritation
68.4*
21.0
Outcome: Symptoms questionnaire one month after end of
Nose Irritation
61.1*
13.1
course.
Sneezing
37.8*
15.8
Symptom frequency: all of the time, some of the time, rarely or
Tightness in
never.
chest
11.1
2.6
Exposure: Area formaldehyde measurements in 6 anatomy
Shortness of
labs, one day during semester, 1983; sampling time 188-222
breath
8.3*
0.0
minutes. Personal breathing zone samples (3M Diffusion), 2
Cough
28.6*
5.3
instructors, sampling time 180-190 minutes
Throat
Area samples:
Irritation
38.9*
7.9
Drager tubes (all labs): 2-year
history of exposure at a melamine-formaldehyde resin
producing plant (mean (SD) age: 38.2 (8.4) years; mean (SD)
work duration 13.2 (7.8) yrs. 24 male, healthy referent
employees with no current or history of exposure to
formaldehyde or other respiratory toxicants (mean (SD) age:
40.0 (8.2) years); mean (SD) work duration 14.5 (8.1) yrs. 100%
participation.
Outcome: Respiratory symptoms ascertained via interview using
standardized questionnaire (ATS).
Exposure: Area samples (40-minute sampling time) in 7
workshops (N=7) and offices (N=l)
Formaldehyde concentration: ppm, mean (SD):
Exposed: 0.78 (0.40) (0.96 (0.49) mg/m3)
Referent: nondetectable
Analysis: Symptom prevalence compared
Prevalence Respiratory Symptoms:
Symptom Exposed Referen
t
Cough 20%* 0%
Phlegm 28.6%* 0%
Chest tightness 52.9%* 0%
*p < 0.05
This document is a draft for review purposes only and does not constitute Agency policy.
A-294 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
SB
IB Cf Oth
Overall
Confidence
Medium
h
Concern for healthy worker survivor bias
(Holnessand Nethercott. 1989) (Toronto, Canada)
Prevalence survey, 84 of 97 selected funeral service apprentice
workers from funeral homes selected by the Metropolitan
District Funeral Director's Association (mean (SD) age 32.1 (11.1)
yrs, 89% male, work duration 8.2 yrs (SD 9.9)). 38 service
volunteers and paid student volunteers as referent subjects
similar in age to the apprentices (mean (SD) age 28.7 (12.7) yrs,
84% male, work duration 7.2 yrs (SD 11.9)).
Outcome: Questionnaires (ATS) administered before and after
an embalming procedure.
Exposure: Area samples (N=2) during each embalming
procedure, mean sampling duration (range): 85 minutes (30-180
minutes).
Mean (SD) formaldehyde: Exposed: 0.36 (0.19) ppm (0.44 (0.23)
mg/m3)a, range 0.08-0.81 ppm. Autopsied cases 0.44 ppm.
Average levels were 0.21 ppm when ventilation units were in
operation.
Referent: 0.02 ppm (0.025 mg/m3)a
Analysis: Differences evaluated using logistic regression analysis
controlling for smoking (pack-years).
Prevalence elevated for 12 of 13 eye, URT,
respiratory and cutaneous symptoms, but 5
were significantly higher compared with
referent: chronic bronchitis (20% vs. 3%, p =
0.035), shortness of breath (20% vs. 3%, p =
0.043), nasal (44% vs. 16%, p = 0.003) and eye
(42% vs. 21%, p = 0.026) irritation and past skin
problem (42% vs. 13%, p = 0.003).
SB IB
a
Oth
Oi/erall
Confidence
Medium
¦
Groups selected from different source populations
(Horvath et al.. 1988) (Wisconsin)
Prevalence survey, 109 of 159 workers at a particleboard
manufacturing plant (71% participation); 57% male; mean age
37.4, SD 11.7 years; Mean duration of employment 10.3 years (1
- 20 years); Referent: 254 of 300 workers at 2 food plants (44%
male; mean age (SD): 34.2 (10.6) years.
Outcome: Respiratory symptoms questionnaire (American
Thoracic Society, ATS) completed before and after monitored
work shift. Intensity assessed by subjects with visual analog
scale.
Exposure: Personal and area samples; Eight-hour, TWA
concentrations measured on each worker on the day of
examination. In the particleboard plant, TWA values averaged
1.04 mg/m3; range 0.26 to 4.4 mg/m3. In the food plants, TWA
values averaged 0.08 mg/m3, range 0.03 ppm to 0.12 ppm).
Other agents sampled in particleboard or molded products
plant.
Symptom Prevalence While at Work
Reported in Preshift Questionnaire:
Symptom Exposed Referen
t
Nose/ throat
irritation
Eye irritation
43.9%*
49.5%*
13.0%
24.0%
p < 0.05
Symptom Prevalence Reported at End
of Shift:
Symptom Exposed Referen
t
Throat
sore/burning
Cough
Phlegm
Nose burning
Stuffy nose
22.0%
34.9%*
26.6%*
28.4%*
33.9%*
3.9%
18.9%
9.8%
2.0%
14.2%
This document is a draft for review purposes only and does not constitute Agency policy.
A-295 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
Compound
Mean (Range)
Total particulates3
Respirable
particulates
Phenol
Carbon monoxide
Sodium hydroxide
Nitrogen dioxide
0.38 (0.25-4.4) mg/m3
0.11 (0.025-1.06) mg/m3
0.15 (0.11-0.26) ppm
7.35 (3.0-11.0) ppm
0.4 - 0.21 mg/m3
ND
Itching nose
Eyes burning
Eyes itching
21.1% 7.9%
39.5%* 9.1%
19.3%* 7.1%
aTotal particulates in food plants were 0.5 and 0.42 mg/m3.
Analysis: Prevalence compared using chi-square statistic. Dose-
response of end of shift symptoms evaluated using multiple
regression models.
SB IB
Of
Oth
Overall
Confidence
Medium
p <0.05
Intensity (visual analogue scale, 0 - 100) for
burning eyes, mean (SD) 47 (27)
Shortness of breath (8.3 vs. 5.1%), wheezing
(3.7 vs. 2.8%), and difficulty breathing (6.4 vs.
2.0%) were not significantly increased.
Dose-response: formaldehyde a significant
predictor of cough, chest complaints, phlegm,
burning nose, stuffy nose, burning eyes, itchy
nose, sore throat, and itchy eyes in multiple
regression models; coefficients were not
reported.
fLofstedt et al.. 20111
Prevalence survey. Sweden
3 brass foundries producing cores using Hot Box method. 43
of 48 exposed workers; 69 of 84 referents working outside
core-production and die-casting halls; not exposed to
chemicals. Prevalence of "ever" asthma or childhood allergy
lower in exposed than in referent (9% and 19%, respectively
versus 14% and 35%, respectively, p<0.05]
Outcome: Self-report, questionnaire; existence of symptoms
during prior week; nasal signs
Exposure: Individual measurements. Monoisocyanates: Mean
of 4-5 5-minute samples randomly distributed over entire
shift.
Formaldehyde: sampling over entire 8-hr shift
Categorized low and high using LOD as cut-point (LOD not
reported].
Mean 0.51 mg/m3, SD 0.049 mg/m3, range 0.013-0.19 mg/m3
Associations of ocular and nasal symptoms
within the previous week and nasal signs
with formaldehyde exposure
Referen
Low
High
t (n=68)
(n = 30)
(n = 12)
Any
1.0
4.3
4.7
nasal
(1.7-11.
(1.2-19.
symptom
c
2)
1)
Nasal
1.0
2.8
2.8
signs -
(1.1-6.9)
(0.8-10.
dry
2)
mucosa
Irritated
1.0
NR*
6.3
eyes
(1.4-28.
4)
NR: not reported
SB
IB
Cf
Oth
Overall
Confidence
Low
t
Could not distinguish effect of formaldehyde from those of
other irritants that were strongly associated with symptoms;
Potential for information bias (reduced recall accuracy);
potential health worker survival
Nasal symptoms included discharge, itch,
sneezing and congestion
ICA and MIC also associated with these nasal
endpoints, nasal symptoms OR 3.9 low and 5.0
in high exposed; nasal signs OR 4.5 low and 1.9
high exposed
(Alexandersson et al.. 1982); Alexandersson and
Hedenstierna, 1989) (Sweden)
Prevalence survey, 1980, Employees at carpentry works (N=47)
for > 1 year, regularly exposed to formaldehyde, and working on
the study day, mean age (± SE) 35 (1.8) years, 49% smokers,
Symptom Prevalence at Work, 1980
(%)
Exposed Referent
Eye
74
0
This document is a draft for review purposes only and does not constitute Agency policy.
A-296 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
duration employment 5.9 years. Referent (N=20) not exposed
to formaldehyde or other lung irritants, employed at the same
plant, mean age (± SE) 35.3 (2.3) years. Asthmatics excluded.
Follow-up 5 years later (1984), 34 exposed and 18 referents; 21
remained exposed, 13 transferred to tasks with no exposure to
irritants.
Outcome: Interviews using standardized questionnaire focused
on nose, eyes, upper airways, and lungs, chronic bronchitis
defined by British Medical Research Council.
Exposure: 1980 study: Personal samplers for formaldehyde,
terpenes, and dust, N=31, duration 6-7 hour/day;
Mean concentration (range): formaldehyde 0.47 mg/m3,
0.05-1.62 mg/m3, terpenes 0 (0-9) mg/m, dust 0.5 (0.3-0.7)
mg/m3
1984 study: 3-4 15 minute samples per person in the exposed
group, estimated TWA
Mean TWA concentration (± SD):
formaldehyde 0.50 (0.12) mg/m3
Mean Peak concentration (± SD): formaldehyde 0.69 ± 0.68 ppm
Analysis: Prevalence of symptoms while at work, change from
1980 to 1984, chi-square
Nose,
Throat
36
0
Symptom Prevalence at Work, 1984
(%)
Ex- Trans- Referent
posed ferred
Eyes
Smartin
45
30
0
g
Itching
40
20
17
Running
60
30
12
Nose
Running
30
10
12
Dryness
15
0
6
4/ Smell
0
0
0
Change from 1980 to 1984 not statistically
significant, p >0.05
S3
IB
Cf
Oth
Overall
Confidence
Low
H N
Healthy survivor bias
(Herbert et al.. 1994)
Prevalence survey, 99 oriented strand board (OSB) workers
(exposed, 98% participation), mean age 35.4 years, 51.5%
smokers; work duration 5.1 years; 165 oil/gas field plant workers
(not exposed to formaldehyde or oil and gas vapors) from same
geographic area (82% participation), mean age 34.9 years, 27.9%
smokers, work duration 10 years. Excluded 14 workers in
referent with hydrogen sulfide exposure.
Outcome: Respiratory symptoms ascertained via interview using
standardized questionnaire.
Exposure: Time weighted average formaldehyde and dust
concentrations based on 21-hour continuous sampling in the
breathing zone at 5 work sites on 2 separate days.
Formaldehyde: range 0.07-0.27 ppm (0.09-0.33 mg/m3). Dust
mean: 0.27 mg/m3, 2.5 nm diameter
Analysis: Symptom prevalence compared
Prevalence Respiratory Symptoms (relevant
to URT irritation):
Symptom Exposed Referent
Usual Cough 24.5%*
Usual Phlegm 31.3%*
Chest tightness 43.4%*
11.1%
13.3%
22.8%
*p< 0.05
Overall
SB
IB
a
Oth
Confidence
Low
Different prevalence smoking and duration of employment
between exposed and referent; no adjustment in analyses
This document is a draft for review purposes only and does not constitute Agency policy.
A-297 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
(Holmstrom et al.. 1991)
Sweden
Prevalence survey, Group 1: 16 persons exposed to medium
density fiberboard (MDF) dust for at least 30% of the workday,
mean age 44.1 yrs, 100% male, 38% smokers. Group 2: 29
exposed to other types of wood dust, mean age 39.3 yrs, 86.2%
male, 31% smokers. Group 3 (Referent), 36 governmental clerks
living in same village as chemical plant, mean age 39.9 yrs,
47.2% male, 28% smokers. (Groups 2 and 3 same as for
Holmstrom and Wilhelmsson, 1988)
Outcome: Symptom prevalence; Questionnaire and medical
examination
Exposure: Personal exposure measurements stable through
year, average 0.2-0.3 mg/m3, peaks seldom > 0.5 mg/m3,
Formaldehyde Concentration, mean
MDF 0.26 mg/m3, range 0.17-0.48 mg/m3
Wood dust 0.25 mg/m3, range 0.3-1.0 mg/m3
Referent 0.09 mg/m3
Analysis: Exposed compared to referent; prevalence rate
difference, 95% confidence intervals
Rate Difference (%) in Symptoms,
Exposed versus Referent
Sympto MDF
Wood Dust
m
%
95% CI
%
95% CI
Nasal
66
47, 85
3
-20, 26
Eye
38
13, 64
1
-1, 13
Throat
19
-3, 42
4
-8, 18
Lower
36
9, 63
3
-14, 21
airway
Relief from symptoms during weekends in 80%
in MDF group and 67% in wood dust group;
and during vacations.
SB
IB
a
Oth
Overall
Confidence
Low
H h
Healthy survivor bias; groups selected from different source
populations; Potential confounding and no adjustment in
analyses
(Alexandersson and Hedenstierna. 1988) (Sweden)
Prevalence survey, 38 exposed employees working with acid-
hardening lacquers for the previous 12 months (mean age (SD):
34 (10) years, mean duration employment 7.8 years) and at
work on the study day. 18 referent employees at the same
company (mean age (SD): 37 (9) years). Asthmatics excluded.
Outcome: Interviews regarding irritation of eyes, nose, throat,
lungs and bronchi were conducted using a standardized
questionnaire.
Exposure: Formaldehyde measurements in the breathing zone,
3-4 15 minute samples per person in the exposed group. No
formaldehyde measurements reported for referent group.
Formaldehyde TWA: 0.40 mg/m3, range: 0.12-1.32 mg/m3.
Peak concentration (15 minute): 0.70 mg/m3, range: 0.14-2.6
mg/m3.
Additional measurements of solvents and dust (4 hr)
Analysis: Group comparisons, chi-square statistic
Symptom Prevalence at Work
Exposed Referent
N (%) N (%)
Eye
Nose, Throat
Dyspnea
Chest
oppression
Cough
25 (65.8)
15 (39.5)
4(10.5)
4(10.5)
2(5.3)
3 (16.7)
0
0
0
0
SB
IB
Cf
Oth
Overall
Confidence
Low
H P
This document is a draft for review purposes only and does not constitute Agency policy.
A-298 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
Selection for healthy survivors; Potential confounding and no
adjustment in analyses
(Holmstrom and Wilhelmsson. 1988) (Wilhelmsson and
Holmstrom, 1992) (Sweden)
Prevalence survey, three test groups chosen by the Swedish
Board of Occ. Safety and Health. Group 1: 70 exposed to
formaldehyde at a chemical plant (resins and impregnation of
paper for laminate production), mean age 36.9 yrs, 87% male,
work duration 10.4 yr (SD 7.3), range 1-36 yr. Group 2:100
exposed to wood dust and formaldehyde, mean age 40.5 yrs,
93% male, work duration 16.6 yr (SD 11.3), range 1-45 yr. Group
3 (referent), 36 governmental clerks living in same village as
chemical plant, mean age 39.9 yrs, 56% male, work duration
11.4 (SD 5.4), 4-18 yr.
Outcome: Questionnaire and medical examination, excluding
upper airway infections. Atopics identified and analyzed
separately from nonatopics based on a laboratory test utilizing
the allergosorbent principle.
Exposure: Breathing zone (personal samplers, 1-2 hours), mean,
range 1985: Group 1: 0.26 (SD 0.17) mg/m3; 0.05-0.50 mg/m3.
Group 2: 0.25 (SD 0.05) mg/m3; 0.2-0.3 mg/m3 and 1.65 mg/m3
for wood dust.
Group 3 Referent: 0.09 mg/m3
Cumulative exposure (dose-years) based on JEM
No occupational exposure to solvents; other agents (phenol,
ammonia, epichlorhydrin, methanol, and ethanol) less than 1%
above PEL
Analysis: Compared symptom prevalence across exposure
groups, chi-square
Overall
Confidence
Low
*
Healthy survivor bias; groups selected from different source
populations; Potential confounding and no adjustment in
analyses
Significantly increased symptom prevalence
reported in formaldehyde exposed groups
Exposure Group
1 2
3
Nasal
64%*
53%*
25%
Eye
24%*
21%
6%
Deep
44%*
39%*
14%
airway
discomfort
*p < 0.05
No significant difference between atopics vs.
nonatopics in symptom prevalence.
Majority reported symptoms did not change
over time
(Kilburn et al., 1985) (Los Angeles)
Prevalence survey, 76 female histology technicians in 23
hospitals & 2 labs (exposed), 97% of eligible, mean (SD) age 40.8
(11.6) years, work duration 12.8 (9.3) years; 56 women in
referent (secretaries and clerks in same institutions) matched
with 40 of the technicians for age, cigarette smoking, and
ethnicity, mean (SD) age 39.5 (10.5) years.
Outcome: Questionnaire for symptoms; composite experience
for previous months or years
Exposure: Environmental samples for formaldehyde, xylene,
toluene, and chloroform by regional NIOSH laboratory in 10 of
25 labs; 1-4 hours sampling time.
Formaldehyde, xylene and toluene
concentrations were not correlated with
symptoms (data not shown).
Symptom Prevalence (%) by Duration of
Formaldehyde Exposure (hours)
>4 hours1
Formaldehyde
Symptom Ref (Hours)
0 1-3 >4
Xylene: #
Slides Cover
slipped
<100 <100
Number
22 47 27
20
This document is a draft for review purposes only and does not constitute Agency policy.
A-299 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Supplemental Information for Formaldehyde—Inhalation
(Reference), study design, exposure levels
Results
< odor2
5
14
32
32
22
45
Eye
20
28
59
66
63
70
Throat
12
14
36
49
37
65
Dry Mouth 20
43
50
47
41
55
Cough
Dry
9
14
23
34
22
50
Mucous
9
14
0
19
7
35
Blood
0
0
0
8.5
4
15
Chest
Tight
5
14
27
40
26
60
Pain
5
14
23
40
37
40
Collected information on exposures, work practices and
ventilation.
Tissue specimen preparation,
Formaldehyde 0.2-1.9 ppm (0.25-2.34 mg/m3)a; rooms with
tissue processors, xylene 8.9-12.6 ppm, chloroform 2-19.1 ppm;
Staining and cover-slipping, xylene 3.2-102 ppm, toluene
8.9-12.6 ppm.
Clerical offices Formaldehyde ND; xylene ND
Analysis: Prevalence by hours formaldehyde exposure and
xylene exposure (statistical analyses not provided).
SB IB
Cf
Oth
Overall
Confidence
Low
Reduced recall accuracy over extended period
1 Xylene exposure among those with >4 hours
exposure to formaldehyde
2 Decreased odor perception
CI = confidence interval; MDF = medium density fiberboard; OR = odds ratio; OSB = oriented strand board; SE =
standard error.
Concentrations reported by authors as ppm or ppb converted to mg/m3
A.5.3. Pulmonary Function
Literature Search
A systematic evaluation of the literature database on studies examining the potential for
effects on pulmonary function in relation to formaldehyde exposure was initially conducted in
November 2012, with yearly updates (see A.1.1). The search strings used in specific databases are
shown in Table A-41. Additional search strategies included:
• Review of reference lists in the the articles identified through the full screening process and
• Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
EPA. 201 Obi
This review focused on standard quantitative measures of pulmonary function including
spirometric measures, FEVi, FVC, and FEF25-75, as well as PEF measured using a flowmeter.
Inclusion and exclusion criteria used in the screening step are described in Table A-42. The search
and screening strategy, including exclusion categories applied and the number of articles excluded
within each exclusion category, is summarized in Figure A-25. Based on this process, 53 studies
were identified and evaluated for consideration in the Toxicological Review.
This document is a draft for review purposes only and does not constitute Agency policy.
A-300 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-41. Summary of search terms for pulmonary function
Database,
search parameters
Terms
PubMed
No date restriction
(Formaldehyde[majr] OR paraformaldehyde[majr] OR formalin[majr]) AND
("pulmonary function" OR "lung function" OR "spirometr*")
Web of Science
No date restriction
TS=(Formaldehyde OR paraformaldehyde OR formalin) AND TS=(pulmonary
function OR lung function OR spirometry)
Abbreviations: Majr= major topic (filter); TS= the requested "topic" is included as a field tag
Table A-42. Inclusion and exclusion criteria for studies of pulmonary function
Included
Excluded
Population
Human
Animals
Exposu re
• Indoor exposure via inhalation
to formaldehyde
• Measurements of formaldehyde
concentration in air, or
exposure during dissection or
embalming
• No formaldehyde specific analyses
• Job title/industry based analysis
• Dermal
• Outdoor exposure
Comparison
Evaluated outcome associations with
formaldehyde exposure
• Case reports
• Surveillance analysis /Illness
investigation (no comparison)
Outcome
• Reported measure of FVC, FEV,
FEF or PEF based on spirometry
or flowmeter
• Pulmonary function among
asthmatic subjects in controlled
human exposure studies (there
were evaluated in the section on
other respiratory conditions
including asthma
• Exposure studies/no outcome
evaluated
• Studies of other outcomes
Other
• Reviews and reports (not primary
research), letters, meeting abstract,
no abstract, methodology paper
This document is a draft for review purposes only and does not constitute Agency policy.
A-301 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Pulmonary Function (Human) Literature Search
t5
2 c
jq 0)
£ QJ
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
The American Thoracic Society has published guidelines for equipment performance
requirements, validation, quality control, test procedures, and reference equations for each type of
spirometric measurement (Miller et al., 2005; Miller et al., 2005), as well as the interpretation of
testing results (Pellegrino etal.. 2005). In addition to the use of conventional spirometric
equipment, peak expiratory flow has been measured in research settings using portable flow
meters operated by study participants trained in their use. Although it requires careful training
and monitoring, this method has the advantage in that it can be used in large epidemiological
studies and multiple measurements can be obtained over time. Studies of residential exposure to
formaldehyde were conducted in this way Krzyzanowski etal. f 19901.
Based on the evaluation of participant selection, exposure and outcome classification,
confounding and other limitations, a level of confidence in the study results, high, medium, low or
not informative was assigned to each study. Eight studies with one or more critical limitations
were classified as not informative.
Lung function varies by race or ethnic origin, gender, age, and height, and is best compared
when normalized to the expected lung function based on these variables fPellegrino etal.. 20051
fHankinson et al.. 19991. Analyses were considered to be limited if they did not adjust or otherwise
account for these variables. Lung function also has been associated with smoking status and
socioeconomic status (Chan etal.. 2000). These predictors of lung function were considered as
potential confounders in the evaluation of studies of formaldehyde exposure. FEVi and PEFR
exhibit diurnal variation and this complicates the interpretation of changes across a work shift or
during a laboratory session if no comparisons were made with an unexposed group fChan etal..
2000: Lebowitz etal.. 1997). Studies with no comparison were given less weight in evaluating
study results.
The healthy worker effect and survivor (lead time) bias was a concern for several cross-
sectional occupational studies, some of which had no other major limitations. Removal of
individuals more sensitive to the irritant effects of formaldehyde from jobs or tasks with
formaldehyde exposure likely occurred in industries with high formaldehyde exposures, and this
type of selection bias might result in an attenuation of risk estimates or a null finding if these
individuals also experienced effects on pulmonary function.
Table A-43. Criteria for categorizing study confidence in epidemiology studies
of pulmonary function
Confidence
Exposure
Study design and analysis
High
General population: For short-term
exposure, sampling period coincides with
pulmonary function measurements.
For long-term exposure, exposure measure
based on at least 3-day sample,
corresponding to appropriate time window
(e.g., measures in more than one season if
Population-based selection of participants or
selection of workers at beginning of exposures
(no lead time bias). Instrument for data
collection described or reference provided (e.g.,
ATS guidelines) and outcome measurement
conducted without knowledge of exposure
status. Analytic approach evaluating dose-
This document is a draft for review purposes only and does not constitute Agency policy.
A-303 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Confidence
Exposure
Study design and analysis
time window covers 12 months, or
addressed season in the analysis). Exposure
assessment designed to characterize mean
individual exposures appropriate to analysis.
Work settings: Ability to differentiate
between exposed and unexposed, or
between low and high exposure.
response relationship using analytic procedures
that are suitable for the type of data, and
quantitative results provided. Confounding
considered and addressed in design or analysis;
large sample size (number of cases).
Medium
General population: More limited exposure
assessment, or uncertainty regarding
correspondence between measured levels
and levels in the etiologically relevant time
window.
Work settings: Referent group may be
exposed to formaldehyde or to other
exposures affecting respiratory conditions
(potentially leading to attenuated risk
estimates)
Lead time bias may be a limitation for
occupational studies. Instrument for data
collection described or reference provided and
outcome measurement conducted without
knowledge of exposure status. Analytic
approach more limited; confounding
considered and addressed in design or analysis
but some questions regarding degree of
correlation between formaldehyde and other
exposures may remain. Sample size may be a
limitation.
Low
General population: Short (<1 day) exposure
measurement period without discussion of
protocol and quality control assessment.
Work settings: Short sampling duration (<1
work shift) without description of protocol.
Lead time bias may be a limitation for
occupational studies. High likelihood of
confounding that prevents differentiation of
effect of formaldehyde from effect of other
exposure(s), limited data analysis (or analysis
that is not appropriate for the data) or small
sample size (number of cases).
Not
informative
Exposure range does not allow meaningful
analysis of risks above 0.010 mg/m3; no
information provided.
Description of methods too sparse to allow
evaluation.
This document is a draft for review purposes only and does not constitute Agency policy.
A-304 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-44. Evaluation of formaldehyde - pulmonary function epidemiology studies
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Laboratory Students Studies
(Akbar-
Selection of
TWA personal
Pre- and
Within person
Mean (SD)
34
Khanzad
participants not
breathing zone
postlab
change across
absolute value at
expose
eh et al..
described.
samples
spirometry
one lab. Age (26
baseline and mean
d; 12
Medical
obtained on all
using ATS
vs. 32 yr), height
% difference
referent
1994)
students and
exposed
criteria on 1
and weight
across lab
s
(Cross-
instructors in
subjects, 9
day per
similar between
compared within
sectional)
anatomy lab;
days, and 1
student; all had
exposed and
and between
referents were
unexposed. 6
at least 6 weeks
unexposed; 21%
groups; t-test
nonmedical
days
of
with history of
students and
Range
formaldehyde
asthma in
instructors.
0.086-3.62
mg/m3
Also sampled
methanol
(mean 110
ppm) and
phenol (not
detected)
exposure at
time of
spirometry
exposed and
none in referent;
nonsmokers
Akbar-
Selection of
Personal
% predicted;
Variables
Mean cross-lab
50
Khanzadeh
participants not
(breathing
prelab and
expressed as a
change analyzed
expose
et al., 1997
described.
zone) (n = 44)
postlab
percentage of
within and
d; 36
(Cross-
and area (n =
spirometric
reference values
between groups
referent
sectional)
76)
variables; four
accounting for
using regression
s
formaldehyde
students
height, weight,
model and t-test
samples
assessed each
age, sex, and
Range
time
race; all
0.34-5.47
nonsmokers.
mg/m3
Since data
collection
Cross-lab change
SB IB a Oth
Overall
Confidence
Medium
Reporting deficiencies;
small sample size in
referent
Cross-lab change
Overall
¦SH
IB Cf Oth
Confidence
y
Low
1 1
Analyses did not account
for possible
acclimatization to
formaldehyde over time.
This document is a draft for review purposes only and does not constitute Agency policy.
A-305 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
occurred
throughout the
course, analyses
did not account
for
acclimatization to
formaldehyde
over time.
(Binawar
a et al..
2010)
(Cross-
sectional)
Excluded
Individuals with
symptoms,
stress, type-1
allergy,
respiratory
disease, and
smokers
First-year
medical
students in
anatomy lab
No
formaldehyde
measurements
Pre- and
postlab
spirometry, %
predicted, day
of course not
reported
Within person
change
Percent predicted
prelab compared
to postlab means
(SD), t-test; no
comparison group
N=80
Cross-lab change
SB IB Cf Oth
Overall
Confidence
Low
No comparison group
(Chia et
al.. 1992)
(Cross-
sectional)
Subjects
selected
randomly; all
agreed to
participate
Area samples at
dissecting
tables, n = 6,
collected on
two occasions.
Personal
samples, n=14
students,
duration 2.5
hours
Range
0.50-1.48
mg/m3
Spirometric
measures
(published
methods); once
before and
after
dissection, 1st
day after 2-
week vacation.
Within person
change; before
and after
dissection means
adjusted for age
and height,
stratified by sex.
Means, absolute
values adjusted for
age and height,
stratified by
gender; and p-
values; no SE; no
comparison group
N=22
Cross-lab change
SB IB Cf Oth
Overall
Confidence
Low
No comparison group;
Small sample size
This document is a draft for review purposes only and does not constitute Agency policy.
A-306 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Khaliq &
Tripathi,
2009
(Cross-
sectional)
Participants
randomly
selected;
excluded
students with
respiratory
illness or
previous
exposure to
formalin; all
nonsmokers
No
formaldehyde
measurements.
Formaldehyde
exposure
assumed for
dissection
classes
Pre- and
postlab
spirometry; 3
tests using best
value,
measured on
1st day of
exposure and
24 hours after
Within person
change
Mean absolute
value (SD)
compared pre- and
postlab, t-test; no
comparison group
N=20
Cross-lab change
SB IB Cf Oth
Overall
Confidence
Low
No comparison group;
Small sample size
(Kriebel
et al..
2001)
(panel
study)
94%
participation;
attendance
declined from
n=37 to n=10
over 13 weeks
(better
attendance by
healthy
individuals?)
Work-exposure
matrix from
sampling in 6
work zones,
multiple days,
and reported
time spent in
each zone
Average 1.35
mg/m3,10-
minute peak
13.42 mg/m3
Spirometric
measures
(ATS methods)
before and at
end of 13
weeks. PEF,
prelab and
across-lab
change every
weekly lab
session
Within person
change; multiple
measurements; 2
smokers and 7 ex-
smokers, PEF in
smokers no
different from
nonsmokers
PEF as fraction of
value before 1st lab
session; Individual
prelab and cross-
lab change data
analyzed together
in relation to
recent, average
and cumulative
formaldehyde in
single generalized
estimating
equations model.
GEE adjusted for
cold on lab day.
Cross-lab change:
no comparison
group
N=38 of
51 with
pre-
and
postlab
measur
es for
>1
week
Longitudinal
Overall
SB IB
11
l>rh
Confidence
Medium
4/
1
1 1
Decline in attendance,
association with
symptoms unknown
Cross-lab change
SS IB Cf Oth
Overall
Confidence
Lew
No comparison group
(Kriebel
et al..
1993)
96%
participation
Personal
samples in the
breathing zone,
1-1.5 hours of
PEF repeated
measures
Wright flow
meter;
Within person
change; multiple
measurements;
one smoker
Mean absolute
value (SD) prelab
and cross-lab
change in
N=20 in
analysis
out of 24
Longitudinal
SB IB Cf Oth
Overall
Confidence
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-307 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
(panel
study)
3-hour lab;
multiple days
Range
0.60-1.14
mg/m3
Pentachloro-
phenol
measured but
not detected.
measured 1-3
times during
each weekly lab
pulmonary
function analyzed
in separate models
using random
effects models
including asthma,
asthma*week, eye
and nose or throat
symptoms.
Provided data and
results of
statistical analyses;
Also showed
absolute value (SD)
and cross-lab
change (SD) at
weeks 1 and 2 and
9 and 10
Small sample size
Cross-lab change
SB IB Cf Oth
Overall
Confidence
Low
No comparison group
Mohamma
d 'pour
and
Maleki,
2011
(cross-
sectional)
30 veterinary
students, male
and female,
aged 18-20 yr,
nonsmokers;
selection of
participants not
described
No
formaldehyde
measurements
Inadequate
Pre- and
postlab
spirometry
Within person
change;
nonsmokers, age
comparable
Mean absolute
value (SD)
compared pre- and
postlab, ANOVA;
tested interaction
between sexes and
exposure
N=15
females
/
N=15
males
SB IB Cf Oth
Overall
Confidence
Not
informative
Exposure levels uncertain
and likely variable in this
occupational group
(Saowak
on et al..
2015)
(Tailand)
Medical
students
and
Students and
faculty in gross
anatomy
dissection labs;
selection,
recruitment,
and
Personal
samplers (n =
36 students, 4
instructors);
area samples,
all NIOSH-2016
method; 3-hr
Siblemed 120
protable
spirometer,
completed
before start of
dissection and
after end of
Within person
change; all
nonsmokers
Average change
over one 3-hr lab
session in the
exposed group
(Within person
change), paired
t-test. Uncertainty
N=36
student
s; n=4
instruct
ors
SB IB Cf Oth
Overall
Confidence
Low
No comparison group
This document is a draft for review purposes only and does not constitute Agency policy.
A-308 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
academic
staff
participation
were not
reported. Ages
19-21 yrs,
nonsmokers
with no history
of chronic
respiratory
disease or
symptomatic
illness
samples over
duration of
class, 3 classes,
January,
August, and
October
Students:
Mean (SD) ppm
Class 1:
0.193 (0.120)
Class 2:
0.271(0.159)
Class 3:
0.828 (0.182)
dissection lab,
maximum of
two readings
whether each
participant was
assessed more
than once.
(Uba et
al.. 1989)
(panel
study)
72.5%
participation
Personal
sampling
monitors
(impingers) in
the breathing
zone; multiple
days and during
3 different
months
TWA Range
0.06-1.14
mg/m3
Spirometric
measures
(ATS methods);
Absolute value
(SD) pre- and
postlab and
cross-shift
change before
Day 0 (before
exposure), at 2
weeks and 7
months
Within person
change; all
nonsmokers
Cross-shift change
in pulmonary
function analyzed
using repeated
measures ANOVA,
adjusted for sex;
change at 2 weeks
and 7 months
compared to the
baseline day.
Compared mean
values measured
at noon on
baseline day, 2
weeks and 7
months.
N=96
Longitudinal
SB IB Cf Oth
Overall
Confidence
High
Cross-lab change
SB IB Cf Oth
Overall
Confidence
High
Residential Studies and School Based Studies
(Bentave
b et al.,
Elderly (20
randomly
Measurements
in common
Assessed by
same team in
Adjusted for sex,
age, country,
General estimating
equations analysis,
N = 600
Pulmonary function
measures
This document is a draft for review purposes only and does not constitute Agency policy.
A-309 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
2015);
(Cross-
sectional),
2009-2011
selected per
home)
permanently
living In
randomly
selected
nursing homes
(8 per city) in
selected city in
7 countries.
Exclusion
criteria stated
(neurological or
psychiatric
disorders)
room; one
week samples;
also measured
particulates,
N02, ozone,
temperature,
humidity and
C02; range of 1
week averages
0.001-0.021
mg/m3, median
0.006 mg/m3;
categorical (low
and high) based
on median
concentration
in each nursing
home
all countries;
medical visit
and
standardized
questionnaire
(European
Community
Respiratory
Health Survey);
spirometry
(ATS/ European
Respiratory
Society
guidelines), %
predicted
BMI, highest
school level,
smoking, and
season
accounting for
correlations within
nursing homes;
adjusted OR (95%
CI); stratification
by presence or
ventilation
SB
IB
Cf
Oth
Overall
Confidence
Medium
Confounding by co-
exposures was not
assessed; range of
average concentrations
within low and high
exposure categories
associated with overall
effects is not known
Broder et
al.
(1988b,
1988c);
Broder et
al.
(1988a)
(Cross-
sectional)
Identification of
exposed
through
households
with UFFI
registered with
state consumer
agency;
referents
selected
randomly from
houses on
adjacent
streets;
concern for
Area samples
on 2 successive
days in hallway,
all bedrooms
and yard.
Median conc. in
rooms were
similar, Inside:
referent 0.035
ppm, range
0.006-0.112
ppm [0.043
mg/m3, range
0.007-0.138
mg/m3]. 90%
Spirometry
protocol
described
Adjustment for
important
confounders in
data analysis
Regression models
of spirometry
values between
and within each
exposure group,
analysis adjusted
for total hours
spent in
house/week,
outside
temperature,
gender, age,
height, smoking,
and race;
presented only
N=l,72
6
expose
d;
N=720
referent
SB IB Cf Oth
Overall
Confidence
Medium
For within group
analyses. Downgraded
from high because results
not presented for
formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
A-310 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
possible over-
reporting of
symptoms but
not for
pulmonary
function
0.061; UFFI
0.043 ppm,
range
0.007-0.227
[0.053 mg/m3,
range
0.009-0.279
mg/m3], 90%
0.073 ppm
Outside:
referent 0.005
ppm, UFFI
0.005 ppm
statistically
significant
regression
coefficients; no
data shown for
formaldehyde
associations
(Franklin
et al..
2000)
(Cross-
sectional )
Recruitment
through local
schools;
response rate
of participants
was not
described.
Participation
not expected
to be
influenced by
outcome or
exposure
3-4 day passive
samples in
bedroom and
main living area
Median (IQR)
0.019 (0.011,
0.035) mg/m3
(communicatio
n by author)
Spirometry
protocol (ATS),
measure-ments
in clinic
Children with
current or history
of upper or lower
respiratory tract
disease were
excluded. %
predicted based
on age, sex, and
height. Mean
eNOS levels by
exposure
category adjusted
for age and atopic
status
Mean absolute
value (SD) and %
predicted (SD) by
exposure group
(<50 and >50 ppb);
only 10 homes in
high exposure
group (data
provided by
author); no
demographic info
except for age
N=224
SB IB a Oth
Overall
Confidence
Medium
Limited exposure
contrast; few subjects in
high exposure group
Krzyzano
wski et
al.
(1990).
adults &
A stratified
random sample
of 202
households of
municipal
employees;
Two one-week
household
samples,
multiple
locations
Mean 0.032
PEF, Wright
flow meter
measured 4
times daily for
2 weeks
Potential
confounding
analyzed in
analysis
Random effects
model accounting
for repeated
measures,
adjusted for
asthma, acute
N=202;
repeate
d
measur
es
SB IB Cf Oth
Overall
Confidence
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-311 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
children
(cross-
sectional)
eligibility
criteria
described
mg/m3;
maximum
0.172 mg/m3
respiratory illness,
smoking, SES, N02,
time of day;
separate analyses
for 15 years and
younger, and over
15 years of age.
(Marks et
al.. 2010)
Schools and
classrooms
were selected
using a two-
stage process,
all students in
selected
classrooms
(grades 4, 5, or
6) were
recruited.
Participation:
418 subjects
(77%) of 543
students in
selected
classes.
One area
sample in each
classroom
2 days/week
for 6 weeks
Spirometry
protocol
described
Randomized
double blind
intervention
study of unflued
and flued gas
heaters, N02 and
formaldehyde
levels varied
together in same
direction
Analysis of effects
in relation to
heater use (flued
vs unflued),
correlated co-
exposures
N=400
Overall
SB
IB Cf Oth
Confidence
Not
informative
No quantitative analyses
specifically for
formaldehyde
Norback et
al., 1995
(Cross-
sectional)
Recruited from
154 randomly
selected
members of
general
population;
57%
participated.
Possibly not
Formaldehyde
(one 2-hour
sample) in the
bedroom at
pillow height.
Also measured
guanine in
bedroom
(house dust
Spirometry and
peak flow
protocol
described; FEVi
(percent
predicted
accounting for
age, sex, and
height).
Analysis did not
account for high
prevalence of
asthma
symptoms in
study group; VOC
concentrations
were correlated
and effects could
FEVi was percent
predicted
accounting for age,
sex, and height;
Kendall's rank
correlation test
N=88
SB IB Cf
Oth
Overall
Confidence
Low
Exposure: Most exposed
to concentration
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
representative
sample because
study design
selected 50%
subjects with
asthma
symptoms (may
respond
differently to
formaldehyde
exposure)
mites), and
room
temperature,
air humidity,
VOCs,
respirable dust,
and C02 in
living room and
bedroom.
Limited
sampling
period in closed
residence with
no point
formaldehyde
emissions;
sampling and
analytic
protocols
referenced
(Andersson et
al., 1981; LOQ
0.1 mg/m3);
Formaldehyde
and Range
<0.005-0.110
Hg/m3 (most
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
of participant
Exposure
Consideration
Analysis and
selection and
measure and
Outcome
of likely
completeness of
Reference
comparability
range
measure
confounding
results
Size
Confidence
72.7%
(exposed 6-7
regression
multiple
exposures in classroom
participation
hours/day); 24
hour samples, 2
samples per
analysis
controlled forSES
(education and
comparisons;
multiple regression
model, % change
that were also associated
with pulmonary function,
but correlation not
classroom, 2
occupation of
per 1SD increase
anticipated
seasons; all
parents,
in formaldehyde
students in
urban/rural, #
(value of SD not
class assigned
smokers at home.
reported).
the median
No adjustment
chemical
for other
concentration;
chemicals in
median 29.8
classroom. Do
Hg/m3 (range
not expect
6.5-136.5
correlation
Hg/m3
between
formaldehyde
and PBDE
congeners or
phthalates in dust
Occupational Studies
(Alexand
All exposed
TWA personal
Spirometric
Preshift variables
Preshift values
N=47
Preshift
workers
employed
sampling;
1 working day.
measures
(ATS methods);
compared to
reference
compared to
predicted based on
expose
d; N=20
ersson et
al.. 1982)
SB IB
Cf
Oth
Overall
Confidence
>1 yr,
recruitment
Range in
exposed
measured on
Monday
equations
age, height, and
gender evaluated
referen
t
H
Medium
from workers
0.05-1.62
morning and
within exposed
Concern for selection for
present on
mg/m3;
after work in
and referent
healthy. P-values were
study day
referent not
exposed;
groups. SD not
reported
(healthy worker
reported;
referents
reported;
effect).
although no
tested either in
difference across
Cross-shift
Referents
measurements
the morning or
shift, compared
selected from
in referent,
afternoon
mean values
plant
high
before and after
This document is a draft for review purposes only and does not constitute Agency policy.
A-314 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
concentration
in exposed
allows
assumption of
an adequate
exposure
contrast for
comparison of
exposed and
referent
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
shift in exposed
(paired t-test)
No comparison
group
Size
Confidence
employees not
exposed to
irritants;
participation
rate not
reported.
Cross-shift
change not
evaluated in
referent
SB
IB a Oth
Overall
Confidence
Low
N
No comparison group
Alexanders
son and
Hedenstier
na, 1989,
1982
Possible
selection for
healthy during
4-year follow-
up; 13 exposed
and 2 referents
lost-to-follow-
up; 13 exposed
transferred to
unexposed jobs
TWA using
personal
sampling
among all
exposed; 3-4
measurements
of 15 minute
periods during
2 working days.
Range in 1980
exposed
0.05-1.62
mg/m3;
referent not
reported;
Range in 1985
not reported.
Sampled for
dust. Although
no
measurements
in referent,
high
Spirometric
measures
(ATS methods);
measured on
Monday
morning across
shift in
exposed;
referents
tested either in
the morning or
afternoon
Values compared
to predicted
normal based on
age, gender, and
height; analyses
stratified by
smoking status.
Dust levels
considered to be
low.
Mean absolute
value (SD) before
work compared to
predicted normal
based on age,
gender, and height
in 1980 and 1984,
and mean
difference from
predicted (SD) in
1984 by smoking
status; 5-year
change corrected
for age-dependent
change; stratified
by smoking. Mean
change across shift
(SD) stratified by
smoking, no
comparison group
(low)
N=21
expose
d; N=18
referen
t
Preshift
SB
IB
a
Oth
Overall
Confidence
u
Medium
W
4-
Concern for selection for
lealthy; small sample
Cross-shift
SB
IB
a
Oth
Overall
Confidence
1
LOW
No comparison group
This document is a draft for review purposes only and does not constitute Agency policy.
A-315 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
concentration
in exposed
allows
assumption of
an adequate
exposure
contrast for
comparison of
exposed and
referent.
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
(Alexand
ersson
and
Hedensti
erna,
1988)
Selection for
healthy;
evaluated
employees
present at work
on study day
TWA using
personal
sampling, 3-4
15-minute
samples/
person; 1
working day.
Range in
exposed
0.12-1.32
mg/m3;
referent not
reported;
although no
measurements
in referent,
high
concentration
in exposed
allows
assumption of
an adequate
exposure
Spirometry on
Monday after
two days
unexposed and
again at end of
shift on second
day. Half of
referent tested
before, and half
tested after
shift
Referents were
"nonexposed"
employees at
same factory. All
male, exposed
slightly younger,
50% smokers;
referent: 33%
smokers.
Analyses
stratified by
smoking status.
Sampled for dust
and solvents:
Authors
considered all
exposures to be
very low and not
confounders
Mean values and
difference from
reference values
by exposure group,
and by smoking
status among
exposed. Change
over 2 days by
smoking status.
Mean comparisons
within exposure
groups, Student's
t-test
N=38
expose
d; N=18
referen
t
Preshift
SB
IB
a
Oth
Overall
Confidence
u
Medium
W
4-
Concern for selection for
lealthy, small samples
Cross-shift
SB
IB
a
Oth
Overall
Confidence
1
LOW
No comparison group
This document is a draft for review purposes only and does not constitute Agency policy.
A-316 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
contrast for
comparison of
exposed and
referent.
Gamble et
a I1976
Of 68 workers
exposed to
hexa-
methylene-
tetramine-
resorcinol
resin, 52 (77%)
completed
questionnaire
and lung
function testing
Area samples
Spirometry
protocol
described
Referent matched
by age, race, sex,
shift, and job;
Exposure to
multiple
chemicals
Exposure group
defined by use of
hexamethylene-
tetramine-
resorcinol resin,
not formaldehyde
N=19
expose
d; N=19
referen
t
SB
IB Cf Oth
Overall
Confidence
Not
informative
h
No quantitative analyses
specifically for
formaldehyde
(Herbert
et al..
1994)
Participation
98% in
exposed, 82%
in referent.
Excluded
accidental
hyd rogen
sulfide
exposure
(n=14). Cross-
shift change
not evaluated
in referent
TWA
continuous
sample in
breathing zone;
5 sites, 2 days.
Range in
exposed
0.09-0.33
mg/m3;
referent not
reported;
sampled for
dust. Although
no
measurements
in referent,
formaldehyde
exposure not
Spirometric
measures; best
of 5
maneuvers,
Snowbird
criteria (Ferris,
1978); at start
of work shift
and after 6
hours
Preshift
comparisons
adjusted for age,
height, and
smoking; not dust
levels, which
authors
considered to be
low
Exposed compared
to referent using
ANCOVA adjusting
for age, height,
and cigarette pack-
years. Presented
absolute values
and p-values from
ANCOVA.
Unconditional
logistic regression
of FEVi/FVC <75%
controlling for age
and cigarette pack-
years. Presented
odds ratios, 95% CI
by smoking
category.
N=99
expose
d;
N=165
referen
t
Preshift
SB
IB
a
Oth
Overall
Confidence
Medium
h
Selection for healthy in
prevalence study;
possible irritant exposure
in referent; co-exposure
to dust
Cross-shift
Overall
SB
IH
I*
Oth
Confidence
h
Low
No comparison group
This document is a draft for review purposes only and does not constitute Agency policy.
A-317 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
of participant
Exposure
Consideration
Analysis and
selection and
measure and
Outcome
of likely
completeness of
Reference
comparability
range
measure
confounding
results
Size
Confidence
expected for
Presented
oil/gas field
absolute values of
workers;
preshift and
adequate
postshift with t-
exposure
statistics and ]>
contrast likely
values; no
for comparison
comparison group
of exposed and
referent.
Holmstrom
100%
Area samples in
Spirometric
Values compared
Presented
N=70
Overall
and
participation;
one group,
measures (FVC,
to expected
observed and
Group
SB IB Cf Oth
Confidence
Wilhelmss
Possible
1979-1984,
FEVi/FVC)
normal based on
expected values by
i;
H H
Medium
on, 1988
differential
personal
percent of
age, sex, smoking,
exposure group,
N=100
imprecision of
samples (1-2
expected
height, and
SD not reported.
Group
cumulative
hours) in 1985
normal based
weight; respirable
Statistical
2; N=36
Medium Healthy
formaldehyde
in all groups.
on age, sex,
particulates
comparisons of
referen
workers; comparison
dose:
Estimated
smoking,
measured but not
observed and
t
groups selected from
formaldehyde
mean
height, and
adjusted for in
expected within
different source
levels
formaldehyde
weight.
analysis.
exposure group
populations
estimated prior
and dust
Comparison
(paired t-test);
to 1979 when
exposure of
groups:
analyzed
exposures were
every
Formaldehyde
correlation with
likely higher.
participant for
only,
duration of
Healthy
each year of
formaldehyde
exposure and
workers
employment,
dose-yrs.
Range in Group
#10.05-0.5
mg/m3, Group
#2 0.2-0.3
mg/m3;
referent mean
0.09 mg/m3;
and wood dust,
referent group.
Referent group
was composed of
administrative
workers who may
not be
comparable to
exposed.
cumulative dose
but did not provide
quantitative
results
This document is a draft for review purposes only and does not constitute Agency policy.
A-318 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Comparable
smoking status
between groups
(data NR)
(Holness
and
Netherco
tt. 1989)
Participants
recruited from
list of funeral
homes, 86.6%
participation;
79.8% of
embalmers
were active
embalmers
(healthy
workers);
community
referent less
similar?
2 area samples
(impingers),
during
embalming, 30
to 180 minutes.
Range in
exposed
0.10-1.0
mg/m3,
referent mean
0.025 mg/m3;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Lung function
as percent
predicted;
measured at
initial
assessment and
before and
after
embalming
procedure
among exposed
and before, and
after a 2-3
hour period in
referents.
Analyses adjusted
for age, height,
and pack-years
smoked, referent
may not be
comparable for
other possible
confounders
Mean percent
predicted (SD)
presented by
exposure group or
by active or
inactive
embalmers, p-
value from
regression model
adjusted for age,
height, and pack-
years smoked;
percent change
during embalming
N=84
expose
d; N=38
referen
t
Overall
SH
IB Cf Oth
Confidence
Medium
M
Comparison groups
selected from different
source populations
Change during
embalming
SB IB
Cf
Oth
Overall
Confidence
Medium
¦
comparison groups
selected from different
source populations
(Horvath
et al..
1988)
71%
participation in
exposed; 88%
participation in
referent. Age
and sex
distribution in
participants
8-hour TWA
using personal
and area
sampling on
day of exam.
Range in
exposed 0.32 to
4.48 mg/m3;
Spirometric
measures
(ATS methods);
% predicted
Adjusted for age,
sex, height, and
smoking in
analyses;
particulates
measured but not
adjusted for in
analysis. Smoking
Variables
evaluated as
percent of
predicted normal;
mean % predicted
(SD) compared
between exposure
groups, t-test;
N=109
expose
d;
N=254
referen
t
Preshift
SB IB Cf Oth
Overall
Confidence
High
Cross-shift
This document is a draft for review purposes only and does not constitute Agency policy.
A-319 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
similar to entire
workforce in
their respective
companies.
Evaluated and
ruled out
survivor bias
using reasons
for leaving
employment
among 54
former
employees;
evaluated
characteristics
of 30/45
nonparticipants
who were
younger and
higher % male,
with similar %
smokers and
mobile home
residency.
referent
0.037-0.15
mg/m3;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
prevalence 53%
in both groups;
mean total
particulates
somewhat higher
in referent.
Other co-
exposures not
detected or a
fraction of PEL
(respirable
particulates,
phenol, CO,
sodium
hydroxide, N02
and acrolein).
multiple regression
on log
concentration
adjusted for age,
sex, height, and
smoking; for cross-
shift change,
paired t-test
(before and after)
of percent
predicted values
SB IB Cf Oth
Overall
Confidence
High
Imbus and
Tochilin,
1988
76% and 84.5%
of employees
tested at each
plant
Area samples
of
formaldehyde
and wood dust
on same day as
pulmonary
testing.
Sampling
protocol(#
Spirometry
protocol
described
(ATS); cross-
shift change
Within person
change; values
presented as
percent
predicted;
descriptive data
on study group
were not given.
Provided data, no
statistical analyses
presented
Plant A
N=94;
Plant B
N=82
SB IB Cf Oth
Overall
Confidence
Not
informative
Reporting deficiencies
This document is a draft for review purposes only and does not constitute Agency policy.
A-320 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Lab workers in
college
anatomy and
histopathology
departments;
selected every
2nd person
from
occupational
list.
samples and
sampling
period) not
described.
Range in
exposed
<0.012-0.074
mg/m3
No unexposed
referent group.
Spirometry
protocol not
described;
measured on
Monday.
Selected every
second person
on list from
each exposure
group.
Khamagao
nkar and
Fulare,
1991
Multiple
30-minute area
samples in the
breathing zone
in exposed (N =
43) and
unexposed (N =
18) areas.
Range in
exposed
0.044-2.79
mg/m3;
referent mean
0.125 mg/m3,
range ND-0.64
mg/m3;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Comparison
group matched
by age and sex (N
= 74).
Comparable for
mean height and
weight; smoking
prevalence: 54%
exposed, 59%
referent.
Other exposures
in lab
Mean absolute
values (SD not
reported)
compared
between exposed
and referent; p-
values reported
N=37
expose
d; N=37
matche
d
referen
t
Overall
SH
IB Cf Oth
Confidence
Medium
i i
Possible exposures in
referent that affect
pulmonary function;
exposure to
formaldehyde in referent
labs
This document is a draft for review purposes only and does not constitute Agency policy.
A-321 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Kilburn et
a I1985
Concern for
selection bias
toward
overestimating
association.
41%
participation,
volunteers,
nonrandom
selection of
participants in
exposed.
Critical
deficiency
No
formaldehyde
concentration
measurements.
Critical
deficiency
Spirometry
protocol
described;
testing before
and after work
shift
Potential
noncomparability
of batt makers
and
administrative
employees,
calculated %
predicted using
reference
population.
Possible exposure
to other
contaminants
among batt
makers
Preshift absolute
values and percent
predicted, and
postshift absolute
values by smoking
status (SD not
reported) among
batt makers and
referent group
N=44
expose
d; N=26
referen
t
SB IB Of Oth
Overall
Confidence
Not
informative
Low participation and
nonrandom selection of
exposed; no
formaldehyde
measurements and
possible co-exposures
Kilburn et
al., 1989
Attendees at 4
national
conventions in
4 different
cities between
1982 and 1986,
compared to
lung function in
a Michigan
population.
Participation
<40%; not
clearly
presented
Formaldehyde
sampling in 10
labs in Los
Angeles (not
representative
of entire
sample); very
wide range of
concentration
Spirometry
protocol
described
(ATS); percent
of "referent"
value
Questionable
comparability to
Michigan referent
population;
exposure both to
formaldehyde
and solvents;
probable
confounding by
local air pollution
in Anaheim, CA
Exposure group
defined by
histology
technician; not
specific to
formaldehyde
N=280
SB
IB
Cf
Oth
Overall
Confidence
Not
informative
No quantitative analyses
specifically for
formaldehyde
Levine et
al., 1984
94%
participation
among
No sampling
measurements;
Rank order
Spirometric
measures
% predicted
based on age and
Regression model
of lung function in
relation to
N=90
This document is a draft for review purposes only and does not constitute Agency policy.
A-322 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
morticians
attending a
required
postgraduate
course
using reported
# embalmings.
Comparison to
funeral home
records for 5
persons
indicated #
embalmings
was over-
reported.
(ATS methods),
% predicted
height; all males
and Caucasian
exposure rank,
adjusted for age,
height, pack-years.
Table 6 in the
paper: mean %
predicted (SD)
comparing low and
high rank category
by smoking status,
low and high rank
matched by age,
Student's t-test
Overall
Confidence
Medium
Uncertainty regarding
assignment to exposure
rank
(Lofstedt
et al..
2009)
86%
participation in
exposed and
69%
participation in
referent.
Healthy
survivor effect
Personal
samples on all
exposed
participants
over a single 8-
hour shift on
same day as
lung function
testing. Range
in exposed
0.014-1.6
mg/m3;
referent not
reported;
major exposure
was to
isocyanates,
low correlation
with
formaldehyde
concentrations
Spirometry
protocol
described (ATS
methods),
cross-shift
change,
percent
predicted using
Swedish
reference;
testing on day
after 2
unexposed
days
Referent from the
same industry;
older age and
smoking
prevalence higher
in exposed.
Important
confounders
addressed in
analysis.
Regression models
of association of
change over shift
with log
formaldehyde level
among exposed,
adjusted for
smoking on test
day and co-
exposure to ICA or
MIC (in two
models);
compared mean
change in %
predicted across
shift between
exposed and
referent
N=64
expose
d;
N=134
referen
t
Cross-shift
SB IB
a
Oth
Overall
Confidence
1 1
Medium
¦
I 1
Healthy survivor effect.
This document is a draft for review purposes only and does not constitute Agency policy.
A-323 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
(Lofstedt
et al..
2011)
(follow-up
of Lofstedt
et al.,
2009)
90%
participation in
exposed and
referent.
Evidence of
survivor bias:
prevalence of
childhood
allergy lower
among
exposed in
2005 (4%
versus 31%).
Higher
prevalence of
nasal
symptoms
among
referents in
2005.
Personal
samples on all
exposed
participants
over a single 8-
hour shift on
same day as
lung function
testing. Range
in exposed in
2001:
0.014-0.44
mg/m3, range
in exposed in
2005:
0.01-0.19
mg/m3;
referent not
reported
Spirometry
protocol
described (ATS
methods),
cross-shift
change,
percent
predicted using
Swedish
reference;
testing on day
after 2
unexposed
days
Referent from the
same industry;
comparable for
age; smoking
prevalence and
work duration
higher in
referent.
Exposure to
formaldehyde,
MIC and ICA
among exposed;
correlation
between
formaldehyde
and isocyanates
low.
Analysis within
each exposure
group
Compared preshift
percent predicted
values (SD) from
2001 and 2005 and
change between
the years (SD)
within exposed
and referent
(Student's
t-test). Multiple
regression of
changes in percent
predicted across
shift adjusted for
MIC,
formaldehyde,
smoking (pack-
years), and
childhood allergy;
authors stated no
significant
association but
quantitative
results were not
reported.
N=25
expose
d;
N=55
referen
t
Preshift 2001 to 2005
SB IB
Cf
Oth
Overall
Confidence
Low
\
Limited sample size to
detect small changes
between 2001 and 2005;
concern for survivor bias;
Co-exposure to MIC & ICA
in exposed—unable to
differentiate for
comparisons of change
from 2001 to 2005.
Cross-shift
SB
IB Cf Oth
Overall
Confidence
Medium
N
Main and
Hogan,
1983
All
administrative
personnel
(exposed) and
all workers on
payroll (police
personnel) who
Three 1-hour
area samples
(impingers), 4
occasions
(August,
September,
December,
Spirometric
measures
(ATS methods);
Percent
predicted
Percent
predicted,
stratified by
smoking status;
potential
dissimilarity
between
Percent predicted
by exposure group
and smoking
status; t statistic
and p-value
presented
N=14
expose
d; N=17
referen
t
Preshift
Overall
SH
IK
•
isrh
Confidence
Low
¦
Comparison groups
selected from different
sources (possible
This document is a draft for review purposes only and does not constitute Agency policy.
A-324 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
did not work in
trailers
(referent) who
were still
employed at
end of
34-month
period.
Comparison
groups not
similar
April) always on
a Monday.
Range in
exposed
0.15-1.97
mg/m3; limited
sampling
period in closed
structure with
no point
formaldehyde
emissions;
sampling and
analytic
protocols
referenced;
referent not
reported
administrative
employees and
police officers;
ETS more
common among
referent
unmeasured
confounding), ETS in
referent; small sample
size (low sensitivity)
Malaka
and
Kodama,
1990
Participation
93%; current
workers.
Healthy
survivor effect
Personal and
area sampling,
duration not
reported; JEM
(cumulative
measure);
range in
exposed
0.27-4.28
mg/m3,
referent
0.004-0.09
mg/m3;
sampled for
dust; adequate
Spirometric
measures
(ATS methods);
% predicted
and absolute
values tested
on Monday and
cross-shift
Referent from
same company;
matched on age,
ethnicity and
smoking; analyses
adjusted for age,
height, weight,
cigarettes per
day, and dust.
Percent predicted
by category of
cumulative
exposure (none,
low, high) using
ANCOVA; Linear
regression of
absolute value on
cumulative
exposure adjusted
for age, height,
weight, cigarettes/
day, and dust.
Cross-shift change:
means of absolute
N=93
expose
d; N=93
referen
t
Preshift
Overall
SB IB
: -
(>rh
Confidence
1 1
Medium
1
1 1
Cross-shift
SB IB
: -
Confidence
1 1
Medium
¦
rn
This document is a draft for review purposes only and does not constitute Agency policy.
A-325 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
exposure
contrast likely
for comparison
of exposed and
referent.
values compared
before and afer
shift in exposed
and referent,
paired t-test
Milton et
a I1996
Evidence of
selection of
healthy
workers (some
refusals to
avoid working
in basement
area); direction
toward
underestimatio
n of effect
Personal
sampling on
each
participant
during 5-6 days
of PEF
measurement,
4 hours on
2 days, same
day as lung
function
testing;
calculated 8-
hourTWA.
Range in
exposed
0.0012-0.265
mg/m3
Spirometry
protocol
described (ATS
criteria); tested
before and
after work after
2 days off work
and 2 other
workdays. PEF
using mini-
Wright peak
flow meter,
measurements
5 per day
during and off
work, 6 days at
work and 4
days off. Self-
reported PEF
correlated with
spirometric PEF
(88 person-
days before (r =
0.91) and after
(r = 0.93) shift
Within person
change, cross-
over design, also
adjusted for night
shift and PEF at
home, multiple
exposures
including to
endotoxin,
phenol resin, and
formaldehyde.
Concentrations
were
correlated—
difficult to
differentiate
individual risk
PEF variability
(high minus low
for the day as
percent of mean
over all days).
Linear regression
of FEVi and FVC
and home
amplitude percent
mean PEF adjusted
for smoking, pack-
years of cigarettes,
and years since
start of exposure.
Cross-shift PEF and
overnight PEF,
logistic regression
of >5% decline in
PEF or linear
regression of
change in PEF on
natural log of
formaldehyde;
models were GEE
to account for
repeated
measures
N=37
SB
IB
Of
Oth
Overall
Confidence
Not
informative
N
Correlated co-exposures
This document is a draft for review purposes only and does not constitute Agency policy.
A-326 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Neghab et
a I2011
Participation
100%. Cross-
shift change
not evaluated
in referent.
Healthy
survivor effect
Area samples
(40 minutes, N
= 7) in 7
workshops and
1 area sample
in office area.
Range not
reported, mean
(SD) 0.96
(0.49); referent
not reported;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Spirometric
measures
(ATS methods);
testing before
and at end of
shift on first
working day of
the week;
percent
predicted
Referent from the
same industry
and comparable
socioeconomic
and demographic
status; %
predicted based
on age and
height; all male
Preshift values
(percent
predicted) (SD)
compared
between exposed
and referent
(Student's t-test),
Pre- and postshift
percent predicted
compared (paired
t-test); Regression
models of lung
function and
association with
duration of
exposure adjusted
for age, height,
weight, and
smoking
N=70
expose
d; N=24
referen
t
Preshift
Overall
SB IB
: *
1 >th
Confidence
1 1
Medium
4,
m
1 1
Healthy worker survival.
Obtained additional
information from author
to clarify results.
Cross-shift
SB
IB Cf Oth
Overall
Confidence
Low
h
No comparison group
Nunn et
al., 1990
Follow-up
complete
(1980-1985)
for 76% of
exposed and
74% of
referent.
Attempted to
include former
employees;
evidence of
survivor bias
Area samples
(1-6 hours)
1979-1985,
personal
samples for
representative
set of exposed
workers,
1985-1987,
estimated prior
to 1979. Range
in exposed
FEVi values
(FEVi/height3)
adjusted for
height
Referent group
from same
factory but
exposed to other
potential irritants
(phenolic and
epoxy resins,
carbon fibers)
and phenol- and
urea-
formaldehyde.
Regression of
FEVi/height3 on
time of screening
visit for each
worker, adjusting
for age in 1980,
smoking status in
1980 and 1985,
maximum and
mean exposure
rank, and total
duration of
N=125
expose
d; N=95
referen
t
SB IB Cf Oth
Overall
Confidence
1 | | |
Medium
¦¦
n i~~i
Concern for selection
bias: loss to follow-up
higher among exposed
with low lung function
compared to referent;
referent exposed to other
potential irritants.
This document is a draft for review purposes only and does not constitute Agency policy.
A-327 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
0.1-2.46
mg/m3 and
above.
Uncertainty
regarding
formaldehyde
levels in
referent not
reported
Stratified results
by smoking
exposure.
Presented mean
slope (95% CI) by
exposure (exposed
and referent), and
smoking status
Ostojic et
a I., 2006
16 physicians
and lab
technicians
exposed daily
in pathology/
anatomy lab
(employed
>4 yrs), source
of referent not
described (all
male, matched
for age and
height)
Assessment of
formaldehyde
exposure was
not described.
No
concentration
data reported;
exposed
defined by
work in
pathology/
anatomy lab
Spirometry
protocol
described;
morning
measurements;
percent
expected
Referent matched
by age and
stature, all
nonsmokers
Compared percent
predicted (mean,
SD) in exposed and
referent using
Student's t-test
N=16
expose
d; N=16
referen
t
SB
IB
Cf Oth
Overall
Confidence
Not
informative
¦
Reporting deficiencies.
Pourma-
habadian
et al., 2006
Selection and
participation of
study groups
not described.
Area samples,
8-hour average,
not measured
in referent
Spirometry
protocol not
described
Differences by
group for age,
length of service,
height, sex,
education, and
smoking; no
adjustment for
age, height, sex,
weight, or
smoking
Absolute values
preshift and
postshift (mean,
SD), and mean
difference across
shift (SD)
compared
between exposed
and referent using
t-test. No
adjustment for
N=124
expose
d; N=56
referen
t
SB
IB Cf Oth
Overall
Confidence
Not
informative
¦
¦
1
Reporting deficiencies;
concern for confounding.
This document is a draft for review purposes only and does not constitute Agency policy.
A-328 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
age, height, sex,
weight, or smoking
Schoenber
g and
Mitchell,
1975
Participation
94%; current
workers.
Healthy survival
effect
Formaldehyde
measurements
taken by
insurance
company
during same
month; 0.5-1
mg/m3; 3
breathing zone
samples,
10.6-16.3
mg/m3;
exposed
categorized by
duration;
additional
exposure to
phenol (5-10
mg/m3; OSHA
PEL 19 mg/m3).
Concentrations
for "never on
line" not
reported;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
Spirometric
measures;
measured
before and
after shift on
Monday and
Friday.
% predicted
based on age,
height, and
gender;
standardized for
15 pack-years
cigarette
smoking; multiple
exposures
(phenol)
Compared %
predicted
(adjusted for
cigarette smoking)
across categories
of duration
N=48
expose
d; N=15
referen
t
O/erall
SB IB
-
: :rh
Confidence
Medium
1
1 1
Healthy survival effect.
Multiple exposures:
formaldehyde, phenol.
Phenol is an irritant but
may not be associated
with pulmonary function
at these levels. Small
sample size.
This document is a draft for review purposes only and does not constitute Agency policy.
A-329 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
of participant
selection and
comparability
Exposure
measure and
range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness of
results
Size
Confidence
Sripaiboon
kij et al.,
2009
100% and 71%
participation in
exposed and
referent
Area samples;
#, dates and
protocol not
described
Spirometry
protocol
described
Models adjusted
for age, sex,
education,
smoking, and ETS.
Co-exposures to
other irritants
(glass
microfibers) and
sensitizers
(phenol resin,
mineral oils)
Exposure group
defined by glass
microfibers or
sensitizing agents;
not specific to
formaldehyde
N=19
expose
d;
N=159
referen
t
Not Informative
Owraii
Si
5 Q" Otfi
Conf derice
f4ot
irrfornvatrve
Tanveer et
al., 1995
49 male
workers
exposed to
formaldehyde
resins (mean
duration 15.6
yr) and 29 male
referents
(security and
administrative
staff).
Recruitment
and
participation
not described.
Healthy
survivor effect
possible
8-hr TWA 0.03
mg/m3;
exposure
protocols and
measurements
not described.
(concerned that
TWA value may
be a typo
because of
comment in
discussion
stated that
findings by
Dally et al. at
0.33-1.7 ppm
supported by
this study at
0.03 mg/m3)
Respiratory
questionnaire,
standardized
MRC, and
spirometry
(ATS protocol);
baseline in
morning and at
end of
workshift
(cross-shift
measured in 31
exposed and 22
referent)
Exposed and
referent
comparable for
age, height,
smoking, and
alcohol; co-
exposures not
discussed
Compared preshift
% predicted,
exposed and
referent, means,
by smoking status
and duration of
exposure,
Student's t-test;
compared cross-
shift change
N=49
expose
d; N=29
referen
t
SB IB Cf Oth
Overall
Confidence
Not
informative
Unable to assess
exposure assessment or
recruitment and selection
protocol; Concern for
selection for healthy
This document is a draft for review purposes only and does not constitute Agency policy.
A-330 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Supporting Material for Hazard Analyses of Pulmonary Function
Table A-45. Formaldehyde effects on pulmonary function in controlled human
exposure studies
Study and design
Results
Medium Confidence (Randomized, results fully reported)
References: Witek et al., 1986; Schachter et al., 1986
No decrements in percent change from
Population: N = 15 healthy, age 18 - 35 years, N=15 asthmatic,
baseline in resting protocol; FVC, FEVi,
age 22 ± 5 years, all nonsmokers.
MEF50% (shown below), MEF40% or Raw-
Exposure: 40 minutes; Clean air and 2 ppm
Exercise protocol showed decrement in
(2.46 mg/m3)a
MEF50% 30 min after exposure end.
Protocol: Random assignment to order of exposure, double
Percent Change from Baseline (MeaniSD)
blinded. Two dose levels, four exposure conditions, 2 days at
Clean Air
2 ppm
rest and 2 days with exercise segment (10 minutes, at 10
FVC (L)
During exposure (@ 40 min.)
minutes into the exposure period), separated by 4 days. Testing
rest
-1.14 ±4.8
-0.99 ± 3.5
at baseline, and at 4 times during 40-minute exposure, and 10
exercise
1.6 ±7.7
0.17 ±6.2
and 30 minutes postexposure. Change from baseline tested
FEVi (L)
using "standard test" and Bonferroni adjustment.
rest
-0.41 ± 5.0
1.65 ±4.5
exercise
4.87 ± 8.3*
4.56 ±5.3"
MEF50% (L/sec)
rest
2.74 ±4.4
7.4 ±5.0*
exercise
8.72 ± 12.6
8.8 ±8.1**
FVC (L)
30 min. postexposure
rest
0.31 ±5.1
1.75 ±3.5
exercise
-2.53 ± 5.4
-0.25 ± 5.6
FEVi (L)
rest
0.5 ±4.7
-1.15 ±5.3
exercise
-0.37 ± 4.5
1.76 ±4.91
MEF50% (L/sec)
rest
-0.87 ± 5.4
2.65 ±8.1
exercise
1.07 ±5.3
-5.74 ± 5.4**
*p<.05; **p<.01
Reference: Schachter et al., 1987
Population: N = 15 healthy hospital laboratory workers routinely
Percent Change from Baseline (MeaniSD)
exposed to HCHO as part of their job, age 32 ± 11.3 years, 33.3 %
Clean Air
2 ppm
male, N = 2 smokers.
FVC (L)
During exposure (@ 40 min.)
Exposure: 40 minutes; clean air and 2.0 ppm (2.46 mg/m3)a
rest
-1.64 ±5.67
-1.30 ±3.64
Protocol: Random assignment to order of exposure, double
exercise
-1.32 ±6.94
-1.60 ± 6.03
blinded.
FEVi (L)
Two dose levels, four exposure conditions, 2 days at rest and 2
rest
-1.25 ±5.25
-2.05 ±3.62
days with exercise. One 10-minute exercise segments at 5
exercise
-0.67 ± 6.33
-1.56 ±6.02
minutes into the 40-minute exposure period. Testing at
baseline, and at 4 times during exposure, and 10 and 30 minutes
postexposure. Percent change from baseline tested using one
FVC (L)
rest
30 min. postexposure
0.68 ±4.13 -0.54 ±2.51
sample t-test with Bonferroni adjustment.
exercise
0.30 ±4.58
-0.07 ± 4.25
FEVi (L)
rest
1.94 ±5.85
-0.95 ± 3.0
exercise
0.62 ±3.81
0.23 ±4.2
This document is a draft for review purposes only and does not constitute Agency policy.
A-331 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study and design
Results
Reference: Green et al., 1987
Population: n = 22, mean age 26.9 ± 3.6 year, nonsmoking, no
history of allergies or hay fever; gender not reported.
Exposure: 60 minute, clean air or 3.01 + 0.01 ppm [3.7 + 0.01
mg/m3]a
Protocol: Random assignment to order of exposure; single
blinded. Two 15-minute exercise segments at 15 and 45 minutes
into the 60-minute exposure period. Testing before and during
exposure period (approximate 15 minute intervals); paired t-test
comparing ratio of exposed value at time(n) to time(0) to ratio of
clean air value at time(n) to time(0).
Declines evident at 47 minutes, Statistically
significant decrements measured in several
endpoints at 55 minutes.
Absolute values at 55 minutes exposure
Clean air 3 ppm
FVC 5.04 ±0.15 4.92 ±0.15*
FEVi 4.29 ±0.12 4.15 ±0.13*
FEV3 4.93 ±0.15 4.80 ±0.15*
FEF25-75 4.74 ±0.25 4.56 ±0.29
*p < 0.02, paired t-test
Reference: Green et al., 1989
Population: N = 24,14 women and 10 men, age 18-35 years,
nonsmoking, no history of asthma, no medications, FVC >80%,
FEV/FVC >75%.
Exposure: 2 hour, clean air, 3 ppm [3.69 mg/m3]a, 0.5 mg/m3
ACA (activated aerosol carbon), 3 ppm plus 0.5 mg/m3 ACA.
Protocol: Randomized block design with 4 2-hour exposure
conditions, one per week; double blinded. Four 15-minute
exercise segments at 15, 45, 75, and 105 minutes into the 2-hour
exposure period. Spirometric testing before and during
exposure period (5 times). PEF at 2 hours, and hourly intervals
for 8-hours postexposure, and at 12 and 16 hours postexposure.
Results presented in graphs for FEVi, FVC,
FEF25-75, and FEV3. During exposure to
formaldehyde + ACA, statistically significant
changes were measured in FVC and FEV3 at
several intervals and decreased SGaw was
measured at the end of exposure;
magnitudes of the changes were less than
10% of baseline. No statistically significant (p
>0.05) effects were observed on FVC, FEVi, or
FEV3, at any of 5 intervals during 2-hour
exposures; for formaldehyde only exposure,
statistically significant decrements were
observed for FEF25-75 and SGaw at 50 and 80
minutes, magnitudes of the changes were
3-5%, compared with baseline.
Low Confidence (Incomplete reporting of results, or blinding not described with multiple exposure levels)
References: Andersen (1979). Andersen and Molhave (1983)
Population: N = 16 healthy students, age 30-33, 68.8 % male,
31.2% smokers
Exposure: 5 hours; 0.3, 0.5,1.0, and 2.0 mg/m3
Protocol: Formaldehyde exposure order determined by Latin
square design; blinding not described. Groups of 4 over 4 days;
testing before (during 2 hours clean air) and 2 times during
exposure. No exercise component.
No change in FVC, FEVi, or FEF25-75; data
presented in graphs
Visual inspection indicates decrease in VC at 1
and 2 mg/m3, FEF25-75 at 0.5 mg/m3 (not
statistically significant).
Reference: Kulle et al., 1987
Population: Group 1 (N = 10), Group 2 (N = 9), nonsmoking
healthy, age 26.3 ± 4.7 years, 53% male.
Exposure: 3 hour, Group 1: 0.0, 0.5,1.0, or 2.0 ppm at rest (0.0,
0.62,1.23, 2.46 mg/m3)a at rest, and an additional 2.0 ppm with
exercise; Group 2: 0.0,1.0, or 3.0 ppm (0.0,1.23, or 3.69
mg/m3)/ and an additional 2.0 ppm with exercise.
Protocol: Exposure order randomly assigned; blinding not
reported. 3-hour exposures each week, at same time on 5
occasions. 8-minute exercise segment every half hour during 2
ppm exposure. Pulmonary function tests (FVC, FEVi, FEF25-75 and
No change in pulmonary function (means by
testing time, no SD presented).
This document is a draft for review purposes only and does not constitute Agency policy.
A-332 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study and design
Results
SGaw) at 0, 30, 60, 90,120,150, and 180 minutes during
exposure, and 24 hours postexposure.
Reference: Lang et al (2008)
Population: N=21, age 19 - 39 years, nonsmoking, healthy
volunteers.
Exposure: 4 hours, clean air, 0.15, 0.3 and 0.5 ppm (0.0, 0.19,
0.37, and 0.62 mg/m3)a; additional 0.3 and 0.5 ppm with peaks
up to 1.0 ppm (1.23 mg/m3)a; additional 0.0, 0.3, and 0.5 ppm
with ethyl acetate to "mask" formaldehyde.
Protocol: Exposure order randomly assigned; double blinded.
Ten 4-hour exposure conditions, one per day, over 10 days.
Airway resistance (Rtot, PEF, FEVi, FEF25-75, and SGaw measured
on first exam and on first and last exposure day, pre and post
exposure. No exercise component.
No statistically different differences between
baseline Day 1 and postexposure on Day 10
(data not presented).
Low Confidence (No randomization; blinding not discussed)
Reference: Day et al (1984)
Population: 2 groups of 9 adults each. Group 1, N = 9, adversely
affected (nonrespiratory) by HCHO fumes emitted by urea foam
insulation (UFFI) in their homes. Group 2, N = 9, not affected by
UFFI present in their homes, or volunteer with no UFFI exposure.
Descriptive data on study subjects was not presented.
Exposure: 1.5 hours in chamber, 1.0 ppm (1.23 mg/m3)a, 0.5
hour under hood, 1.2 ppm (1.48 mg/m3)a; no clean air control.
Protocol: Testing before, after, and 6.5 hours after exposure. No
exercise component.
No change in FVC, FEVi, or FEF25-75 (mean ±
SD) paired t-test
Reference: Sauder et al (1986)
Population: n = 9, mean age 26 ± 3.6 years, healthy, non allergic
(for 6 weeks prior to test), nonsmokers.
Exposure: 3 hours; 0, 3 ppm (3.69 mg/m3)a
Protocol: Nonrandom assignment; blinding not described. 8-
minute bicycle exercise followed by spirometry measurements
after each 30-minute interval during 3 hour exposures. First day
clean air only, second day 3 ppm formaldehyde. Testing again
after 24 hours. Repeated measures ANOVA
Clean air
3 ppm
30 minutes
FVC
FEVi
FEF25-75
FVC
FEVi
FEF25-75
4.61
3.98
4.46
4.71
4.02
4.45
4.62
3.90*
4.16**
180 minutes
4.68
3.99
4.48
*p <0.05, ** p <0.01, paired t-test
Statistically significant decreases in FEVi (2%)
and FEF25%-75% (7%) after first 30 minutes;
Range in response:
FEVi -5% to +1%
FEF25-75 —14% to +2%
No other changes during exposure or 24
hours after.
Concentrations reported by authors as ppm or ppb converted to mg/m3.
1 Study summaries describing change in pulmonary function measures during a work shift or
2 anatomy lab session
3 Appendix Figures 26 -28 present study findings for three spirometry measures, FEF25-75,
4 FEVi, and FVC, and study details are summarized in Appendix Table A-46. For each measure, the
This document is a draft for review purposes only and does not constitute Agency policy.
A-333 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 mean difference across a work shift or lab session in exposed and referent groups (when reported)
2 is plotted with error bars depicting the standard error. Separate graphs depict the mean before and
3 after difference expressed as absolute value (e.g., FEVi in liters) or percent predicted. The third
4 plot shows results for studies that reported changes as a percent of the baseline value.
5
Reference
Setting
Referent Confidence
(Binawara et
Anatomy
No referent Low
al„ 2010). N =
lab
80
Reference
Setting
Referent
Confidence
Akbar-
Anatomy
N = 36
Low
Khanzadeh,
lab
1997,
N = 50
(Akbar-
Anatomy
N = 12
Medium
Khanzade
lab
h et al.,
1994).
N = 34
Reference
Setting
Referent
Confidence
Malaka, 1990,
Wood
N = 50
Medium
N = 55
products
fAlexandersso
Wood
Not
Low
n and
products
measured
Hedenstierna,
1989). N = 21
fHorvath et
Wood
N = 254
High
al., 1988). N =
products
109
fAlexandersso
Wood
Not
Low
n and
products
measured
Hedenstierna,
1988). N = 38
fAlexandersso
Wood
Not
Low
n et al., 1982),
products
measured
N =47
Khaliq, 2009,
Anatomy
No
Low
N = 20
lab
referent
Uba, 1989,
Anatomy
Week 2 vs
High
N = 96
lab
baseline
day
Malaka, 1990
Alexandersson, 1989
Horvath, 1988
Alexandersson, 19SS
Alexandersson, 1982
Kaliq, 2009
Uba, 1989
j—r
-0.07
1—1
-0.61
m
-0.1
0.01
—I
-0.18
HZ
-0.14
H
-0.32
0.18
m—i
tE
i i i i
-0.08
-0.09
I I I
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
FEF at 25-75% of FVC (Us)
d Exposed
~ Referent
Binawara, 2010
-8-6-4-2 0 2
FEF at 25-75% of FVC (% Predicted)
Akbar-Khanzadeh, 1997
Akbar-Khanzadeh, 1994
9.3 |—
2.2 I 1
2.31 H
2.5 | 1
I 1 1 1 1 1
-2 0 2 4 6 8 10
FEF at 25-75% of FVC (% Change)
Figure A-24. Plots of change in FEF at
25-75% of FVC across a work shift or anatomy lab session by study with study
details. The difference in reported means before and after shift or lab as either
liters/second or % predicted are shown, and percent change in FEF across the lab
This document is a draft for review purposes only and does not constitute Agency policy.
A-334 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
was reported by two studies (3rd panel). Mean difference or percent change and SE
are shown. These were calculated by EPA when not reported using SD for before
and after means.
This document is a draft for review purposes only and does not constitute Agency policy.
A-335 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Setting
Referent
Confidence
Malaka, 1990,
Wood
N = 50
Medium
N = 55
products
Herbert, 1989,
Wood
Not
Low
N = 99
products
measured
(Alexanderss
Wood
Not
Low
on and
products
measured
Hedenstierna
, 1989), N = 21
Horvath, 1988,
Wood
N = 254
High
N = 109
products
(Alexanderss
Wood
Not
Low
on and
products
measured
Hedenstierna
, 1988), N = 38
(Alexanderss
Wood
Not
Low
on et al..
products
measured
1982), N =47
Khaliq, 2009,
Anatomy
No referent
Low
N = 20,
lab
(Chia et al.,
Anatomy
Not
Low
lab
measured
1992),
N = 13
Uba, 1989,
Anatomy
Week 2 vs
High
N = 96
lab
baseline
day
Malaka, 1990
Herbert, 1989
Alexandersson, 1989
Horvath, 19SS
Alexandersson, 1983
Alexandersson, 1982
Kaliq, 2009
Chia, 1992
Uba, 1989
0.0S
0
m
I 1
m
h
r
"T~
T"
D-
-0.04
-0.05
-0.04
-0.04
-0.01
-0.17
-0.0S
-0.13
-0.03
-0.05
1
-0.3 -0.2 -0.1 0.0 0.1
Change Across Shift/Lab, FEV 1 sec (L)
Neghab, 2011
Lofstedt, 2009
Binawara. 2010
0.1
a
-10.5
-1.4
-19.7
Reference
Setting
Referent
Confidence
1
-20
I
-15
I I I
-10 -5 0 5
(Neghab et
Chemicals
Not
Low
Change Across thiTVLab, l-tv 1 sec
(Percent Predicted)
al., 2011). N
measured
= 70
(Lofstedt et
al., 2009), N
Chemicals
N = 134
Medium
1=1 Exposed
Ho In ess, 1989
1.45
1 1 n Referent
= 64
h
-0.03
(Binawara
et al..
Anatomy
No referent
Low
Akbar-Khanzadeh, 1997
6.2
1 1
lab
2.4
1—1
1
|-
2010). N =
Akbar-Khanzadeh, 1994
-0.03
80
Reference
Setting
Referent
Confidence
Holness, 1989,
Embalming
N = 13
Medium
N = 22
Akbar-
Anatomy
N = 36
Low
Khanzadeh,
lab
1997, N = 50
(Akbar-
Anatomy
N = 12
Medium
Khanzadeh
lab
et al..
-2 0 2 4 6
Change Across Shift/Lab, FEV 1 sec
(Percent Change)
This document is a draft for review purposes only and does not constitute Agency policy.
A-336 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Figure A-25. Plots of change in FEV1 across
a work shift or anatomy lab session by
study with study details. The difference in reported means before and after shift
or lab as either liters or % predicted are shown, or percent change in FEV1 across
the lab. Mean difference or percent change and SE are shown. These were
calculated by EPA when not reported using SD for before and after means.
1994), N =
34
Demographic information for Holness, 1989 are for
entire study groups.
This document is a draft for review purposes only and does not constitute Agency policy.
A-337 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Setting
Referent
Confidence
Malaka, 1990,
Wood
N = 50
Medium
N = 55
products
Herbert, 1989,
Wood
Not
Low
N = 99
products
measured
(Alexanderss
Wood
Not
Low
on and
products
measured
Hedenstierna
, 1989), N = 21
Horvath, 1988,
Wood
N = 254
High
N = 109
products
{Alexandersson
Wood
Not
Low
, 1988, 31634
products
measured
(Alexanderss
Wood
Not
Low
on et al..
products
measured
1982), N = 47
Khaliq, 2009,
Anatomy
No
Low
N = 20,
lab
referent
(Chia et al..
Anatomy
Not
Low
lab
measured
1992),
N = 13
Uba, 1989,
Anatomy
Week 2 vs
High
N = 96
lab
baseline
day
Reference
Setting
Referent
Confidence
(Neghab et
Chemicals
Not
Low
al., 2011), N
measured
= 70
(Lofstedt et
Chemicals
N = 134
Medium
al., 2009), N
= 64,
(Binawara et
Anatomy
No
Low
al., 2010), N
lab
referent
= 80
Reference
Setting
Referent
Confidence
Holness, 1989,
Embalming
N = 13
Medium
N = 22
Akbar-
Anatomy
N = 36
Low
Khanzadeh,
lab
1997, N = 50
(Akbar-
Anatomy
N = 12
Medium
Khanzadeh
lab
et al.,
1994), n =
34
Malaka, 1990
Herbert, 1989
Alexandersson, 1939
Horvath, 19S8
Alexandersson, 19SS
Alexandersson, 19S2
Kaliq, 2009
Chia, 1992
Uba, 1989
0.01
0.01
H=
0.01
l-C
-0.05
-0.03
-0.06
0
d—I
-0.05
-0.12
-0.12
-0.01
-0.04
-0.3
-0.2
-0.1
0.0
Change Across Shift/Lab, FVC (L)
Neghab, 2011
Lofstedt, 2009
Binawara, 2010 | [-
r
~r~
-10
H=E
"T~
-9.8
-0.7
-2
-16.1
-15 -10 -5 0 5
Change Across Shift/Lab, FVC (Percent Predicted)
Ho In ess, 1989
Akbar-Khanzadeh, 1997
Akbar-Khanzadeh, 1994
1.13
0.88
4.6
2.5
3T1
-0.3
-1.4
c Exposed
~ Referent
Change Across Shift'Lab, FVC (Percent Change}
This document is a draft for review purposes only and does not constitute Agency policy.
A-338 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Demographic information for Holness, 1989 are for
entire study groups.
Figure A-26. Plots of change in FVC across a
work shift or anatomy lab session by study
with study details. The difference in reported means before and after shift or lab
as either liters or % predicted are shown, or percent change in FVC across the lab.
Mean difference or percent change and SE are shown. These were calculated by EPA
when not reported using SD for before and after means.
Table A-46. Study details for references depicted in Figures A-26 - A-28
Study information
Group characteristics
Measures reported/ analysis
Occupational studies
(Neghab et al„ 2011)
Resin production
Confidence: Low (No comparison
group)
Exposed: N = 70, male, age 38 yr,
24% smokers; Referent: Not
measured
FEVi, FVC, FEVi/FVC, PEF
Mean values (percent predicted) before and after
shift compared (paired t-test) in exposed
(Lofstedt et al., 2009)
Chemical company
Confidence: Medium (Healthy
survivor effect)
Exposed: N = 64, 89% male, age 44
yr, 25% smokers; Referent: N = 134,
88% male, age 40 yr, 22% smokers
VC, FEVi
Compared mean difference across shift (percent
predicted) between exposed and referent
(regression); association with formaldehyde
adjusting for isocyanate levels and smoking
(regression)
Malaka and Kodama, 1990
Plywood manufacture
Confidence: Medium (healthy
survivors)
Exposed: N = 55, male, age 27 yr,
53% smokers; Referent: matched
by age, ethnicity and smoking; N =
50, male, age 29 yr, 53% smokers
FEVi, FVC, FEVi/FVC, FEF25-75
Mean values before and after shift compared
(paired t-test) in exposed and referent
Herbert et al., 1989
Particle board manufacture
Confidence: Low (No comparison
group)
Exposed: N = 99, sex NR, age 35 yr,
52% smokers; Referent: Not
measured
FEVi, FVC, FEVi/FVC
Mean values before and after shift compared
(paired t-test) in exposed
Alexandersson and
Hedenstierna, 1989
Cabinet manufacture, 5-year
follow-up of (Alexandersson et
al.. 1982)
Confidence: Low (No comparison
group)
Exposed: N = 21, male, age 37 yr,
48% smokers; Referent: Not
measured
FEVi, FVC, FEVi/FVC, FEF25-75
Mean values before and after shift compared,
stratified by smoking status (paired t-test) in
exposed
(Holness and Nethercott,
1989)
Funeral workers (embalming)
Confidence: Medium
(comparison groups selected
from different source
populations)
Exposed: N = 22, 89% male, age 32
yr, 50% smokers; Referent
(community volunteers): N = 13,
84% male, age 28 yr, 37% smokers
(Demographic information for are
for entire study groups)
FEVi, FVC, FEF50, FEF75
Compared mean percent change during
embalming (or after 2-3 hr) (percent predicted)
between exposed and referent (regression
adjusting for age, height, and pack-yr smoked
(Horvath et al., 1988)
Particle board manufacture
Confidence: High
Exposed: N = 109, 57% male, age 37
yr, 53% smokers; Referent (food
processing): N = 254, 44% male, age
34 yr, 53% smokers
FEVi, FVC, FEVi/FVC, FEF25-75, PEF
Mean values before and after shift (percent
predicted) compared (paired t-test) in exposed
and referent; correlation with formaldehyde
concentration
This document is a draft for review purposes only and does not constitute Agency policy.
A-339 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study information
Group characteristics
Measures reported/ analysis
(Alexandersson, 1988)
Wood products
Confidence: Low (No comparison
group)
Exposed: N = 38, male, age 34 yr,
50% smokers; Referent: Not
measured
FEVi, FVC, FEVi/FVC, FEF25-75
Mean values before shift on first day and after
shift on second day compared, stratified by
smoking status (paired t-test) in exposed
(Alexandersson et al„ 1982)
Cabinet manufacture
Confidence: Low (No comparison
group)
Exposed: N = 47, male, age 35 yr,
51% smokers; Referent: Not
measured
FEVi, FVC, FEVi/FVC, FEF25-75
Mean values before and after shift compared,
stratified by smoking status (paired t-test) in
exposed
Anatomy lab (dissection)
(Saowakon et al., 2015)
Anatomy course
Confidence: Low (No comparison
group)
N = 36, gender NR, age 19.8 yr,
nonsmokers; no referent
FVC, FEVi, FEVi/FVC, FEF25-75, PEF
Mean values compared before and after
dissection session (paired t-test) in exposed
(Binawara et al., 2010)
Anatomy course
Confidence: Low (No comparison
group)
N = 80, male, age 20 yr,
nonsmokers; referent: No referent
FEVi, FVC, FEVi/FVC, FEF25-75, PEF
Mean values (percent predicted) before and after
shift compared (paired t-test) in exposed
Khaliq and Tripathi, 2009
Anatomy course
Confidence: Low (No comparison
group; small sample size)
Exposed: N = 20, male, age 18 yr,
nonsmokers; no referent
FEVi, FVC, FEVi/FVC, FEF25-75, PEF
Mean values before and after lab compared
(repeated measure ANOVA) in exposed
Akbar-Khanzadeh et al., 1997
Anatomy course
Confidence: Low (Analyses did
not account for possible
acclimatization to formaldehyde
over time)
Exposed: N = 50, 50% male, age 24
yr, nonsmokers; referent
(physiotherapy students): N = 36,
24% male, age 24 yr, nonsmokers
FEVi, FVC, FEF25-75
Compared mean percent change (standardized
for baseline) over lab in exposed and referent
(paired t-test); compared difference between
groups (unpaired t-test)
(Akbar-Khanzadeh et al.,
1994)
Anatomy course,
Confidence: Medium
(Comparison groups dissimilar;
small sample size in referent)
Exposed: N = 34, 71% male, age 26
yr, nonsmokers; referent: N = 12,
67% male, age 31 yr, nonsmokers
FEVi, FVC, FEVi/FVC, FEF25-75
Compared mean percent change (standardized
for baseline) over lab in exposed and referent
(paired t-test); compared difference between
groups (unpaired t-test)
(Chia et al.. 1992)
Anatomy course
Confidence: Low (No comparison
group; small sample size)
Exposed: N = 13 male, n = 9 female,
age NR, smoking NR; referent: Not
measured
FEVi, FVC (means adjusted for age and height);
Mean values before and after lab compared (chi-
square statistic)
(Uba et al.. 1989)
Anatomy course
Confidence: High
Exposed: N = 96, 74% male, age 24
yr, nonsmokers; comparison: Cross-
lab change week 2 vs. baseline day
FEVi, FVC, FEVi/FVC, FEF25-75
Mean percent change over lab session at 2 weeks
compared to baseline (repeated measures
ANOVA, adjusted for sex)
1 A.5.4. Immune-Mediated Conditions, Including Allergies and Asthma
2 Literature Search
3 A systematic evaluation of the literature database on studies examining the potential for
4 respiratory and immume-mediated conditions, including allergies and asthma, in relation to
This document is a draft for review purposes only and does not constitute Agency policy.
A-340 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 formaldehyde exposure was initially conducted in October 2012, with yearly updates (see A.1.1).
2 The search strings used in specific databases are shown in Table A-47. Additional search strategies
3 included:
4 • Review of reference lists in the articles identified through the full screening process,
5 • Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
6 EPA. 2010b1. and
7 • Review of abstracts (initial title search for formaldehyde, then abstract review) from
8 2005-2014 presented at International Society of Environmental Epidemiology annual
9 meetings.
10 The focus of this review is on hypersensitivity (allergy) and on asthma; these are well-
11 developed areas of research with respect to immune-related effects of inhalation exposure to
12 formaldehyde. Within these areas, several different types of endpoints or outcomes have been
13 examined. EPA included the following outcomes in studies in humans in this review:
14 • Prevalence of current allergy symptoms (nasal, ocular, or dermatologic), incidence of
15 allergies, or skin prick tests in general population or occupational studies with inhalation
16 exposure measures;
17 • Incidence of asthma (based on parent- or self-report of physician-diagnosis), prevalence of
18 current asthma (based on various validated questionnaires or based on medical records),
19 asthma control among people with asthma (based on questionnaires developed to assess
20 markers of asthma morbidity such as symptoms, medication use and healthcare utilization);
21 and
22 • Pulmonary function (standard spirometry) and bronchial challenge-airway reactivity tests
23 among people with asthma; [pulmonary function studies in general (nonasthmatic)
24 populations were reviewed in the "Pulmonary Function" section],
25 EPA considered "ever had asthma" to be of limited use in this review, as the formaldehyde
26 measures available do not reflect cumulative exposures that could be related to cumulative risk,
27 and thus EPA did not include studies limited to "ever had asthma."
28 Case reports of occupational asthma were not systematically reviewed, but selected
29 references are included for illustration. Formaldehyde-specific antibodies were not examined, as
30 there has been little evidence of effects; selected references are included for illustration.
31 Based on the ultimate conclusion that the toxicity studies in animals were most
32 appropriately reviewed as mechanistic information (see Section 1.2.3 of the Toxicological Review),
33 the experimental studies identified as a result of this literature search are evaluated and described
34 as mechanistic studies related to noncancer respiratory health effects section (see Appendix A.5.6).
35 In regard to the experimental studies identified by this literature search, particular attention (and
This document is a draft for review purposes only and does not constitute Agency policy.
A-341 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 inclusion/exclusion criteria applied in the HERO database) emphasized the identification of studies
2 examining the following endpoints:
3 • Airway inflammatory responses to sensitizing antigens, such as bronchoconstriction and
4 airway hyperresponsiveness. (Studies describing the development of immunological or
5 allergy animal models were not included, however.)
6 • Biomarkers relating to potential mechanisms in animal toxicology studies, such as
7 eosinophil infiltration, immunoglobulins (e.g., total or anti-allergen-specific IgE or IgG), and
8 cytokines pertinent to hypersensitivity responses, and neurogenic mechanisms of airway
9 inflammation.
10 • Note: contact dermatitis is a well-established effect from dermal exposure and the effects of
11 dermal exposure are not a focus of this review; thus studies of contact dermatitis from
12 dermal exposures are excluded from this literature search (and the literature search in
13 Appendix A.5.6).
14 Inclusion and exclusion criteria for selection of studies are summarized in Table A-48 and
15 Table A-49, respectively, for human and animal studies.
16 After compilation into a single database and electronic removal of duplication citations, the
17 4,622 articles were initially screened within an EndNote library; the initial screening was based on
18 title (3,409 excluded), followed by screening by title and abstract (1,046 excluded). Most of the
19 exclusions at these stages were because the paper was not related to this review (e.g., studies of use
20 of formaldehyde in vaccines, or studies of other chemicals) or were secondary data sources
21 (reviews). Full text review was conducted on 167 identified articles. Most of the exclusions at this
22 stage were because the study did not examine any of the selected outcome measures or did not
23 conduct an analysis of formaldehyde. Four studies were excluded based on the aspects of the
24 "comparison" criteria (e.g., limited exposure range):
25 • Smedje etal.. 1997—limited exposure range with 54% less than LOD (LOD 0.005, range
26 <0.005 to 0.010 mg/m3) [The follow-up study of this cohort, described in Smedie and
27 Norback. 2001 was not excluded because it included an additional measurement period and
28 wider range of exposures.]
29 • Kim etal. f20071—limited exposure range, with large percentage less than LOD (LOD 0.006,
30 mean 0.007, maximum 0.016 mg/m3)
31 • Zhao etal. (2008)—limited exposure range. The LOD was not reported but the minimum
32 and maximum values were reported as 0.001 and 0.005 mg/m3; this maximum is lower
33 than the LOD in most studies. Technical difficulties led to the exclusion of measures from
34 14 of the 46 classrooms, but the authors did not comment on the unusual finding of higher
35 levels in outdoor compared to indoor measures. [The corresponding author did not respond
36 to an email inquiry asking for clarification regarding the exposure measures.]
This document is a draft for review purposes only and does not constitute Agency policy.
A-342 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 • Chatzidiakou etal. T20141—did not present an analysis of the effect of variability in
2 formaldehyde within either urban or suburban setting, and the design did not allow for
3 separation of effects of location from effects of formaldehyde.
4 The search and screening strategy, including exclusion categories applied and the number
5 of articles excluded within each exclusion category based on the full text screening, is summarized
6 in Figure A-29. Based on this process, 36 human studies and 16 animal-mechanistic studies were
7 identified and evaluated for consideration in the Toxicological Review.
This document is a draft for review purposes only and does not constitute Agency policy.
A-343 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-47. Summary of search terms - allergy-related conditions, including
asthma
Database,
Initial search date
Terms
PubMed
10/31/2012
No date restriction
formaldehyde and (asthma or wheeze or respiratory or allergy or immune or
sensitization) NOT ("formalin test" OR "formaldehyde fixation" OR "formalin fixation"
OR "formalin fixed" OR "formaldehyde fixed" OR "formalin-induced" OR "formalin-
evoked")
Web of Science
11/5/2012
No date restriction
(TS=formaldehyde and TS=asthma) OR (TS=formaldehyde and TS=allergy) OR
(TS=formaldehyde and TS=immune) OR (TS=formaldehyde and TS=respiratory) OR
(TS=formaldehyde and TS=sensitization) OR (TS=formaldehyde and TS=wheeze)
Toxline
11/2/2012
No date restriction
formaldehyde @AND @OR (immune allergy asthma respiratory wheeze sensitization)
Table A-48. Inclusion and exclusion criteria for studies of allergy and asthma
studies in humans
Included
Excluded
Population
0.01 Human
0.02 Animals
Exposu re
0.03 Indoor exposure via
inhalation to
formaldehyde,
measured in homes or
schools or by personal
monitors in general
population studies
0.04 Occupational
exposure settings (e.g.,
manufacture of pressed
wood products)
0.05 Not formaldehyde
0.06 Outdoor formaldehyde exposure
0.07 Dental-related exposures or cosmetic and other
dermal-related exposures
0.08 Exposure via dialysis
0.09 Formaldehyde as fixative
0.10 Intervention studies in which formaldehyde and
numerous other factors were simultaneously changed
Comparison
Analysis of variation in risk in
relation to variation in
formaldedhye, specifcially:
0.11 at exposures above
0.010 mg/m3
0.12 across exposure
range that spans at least
0.01 mg/m3 (e.g., from
0.02 to 0.03 mg/m3)
0.13 Case reports (selected references used for
illustration)
Outcome
0.14 Allergy symptoms3
0.15 Skin prick tests
0.16 Incidence of specific
allergies
0.17 Prevalence of
current asthma3
0.18 Incidence of asthma
0.21 Sick building syndrome, sick building symptoms,
chemical sensitivity studies
0.22 Contact dermatitis, eczema, or urticaria in studies of
worker populations with likely dermal exposure
0.23 Formaldehyde-specific antibodies (FA-Ig)
0.24 Pulmonary function in controlled exposure studies
in people without asthma [these studies are included in
Section A.5.3. Pulmonary Function]
This document is a draft for review purposes only and does not constitute Agency policy.
A-344 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Included
Excluded
0.19 Asthma control or
severity
0.20 Controlled exposure
pulmonary function
studies in people with
asthma
0.25 Lifetime prevalence of asthma ("Ever had asthma" or
"ever had wheezing episode")
Other
Reviews, reports, no abstract (title only), meeting abstract,
methodology paper, formaldehyde used in vaccine preparation,
other miscellaneous reasons—not on topic
aBased on the methods used in the American Thoracic Society questionnaire (Ferris, 1978) or subsequent
instruments that built upon this work, such as the International Study of Arthritis and Allergies in Children (ISAAC)
and European Community Respiratory Health Survey (ECHRS) questionnaires.
Table A-49. Inclusion and exclusion criteria for studies of hypersensitivity in
animals
Included
Excluded
Population
0.26 Animals
0.27 Humans
Exposu re
0.28 Inhalation route,
formaldehyde
0.29 Not formaldehyde
0.30 Oral or dermal exposure protocol
Comparison
0.31 One or more exposure
group compared to
control
0.32 No control group
Outcome
0.33 Bronchoconstriction
or airway
hyperresponsiveness
measures
0.34 Total or anti-allergen-
specific IgE or IgG
0.35 Eosinophil infiltration
in lung
0.36 Th2 cytokines (e.g., IL-
4, IL-5)
0.37 General chronic bioassay measures (e.g., organ
weight, tumor incidence)
0.38 Host resistance assays
0.39 Antibody responses not involving respiratory
sensitizers (e.g., sheep red blood cells, tetanus
toxoid)
0.40 Dermal sensitization measures
0.41 In vitro studies, measures of inflammation and
irritation (e.g., TNF-a, ROS), and formaldehyde-
specific antibody studies were identified using a
more specific search string in Section A.5.6.
Other
0.42 Reviews, reports, meeting abstract, no abstract
(title only), methodology paper
This document is a draft for review purposes only and does not constitute Agency policy.
A-345 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Immune - Allergy and Asthma (Human and Animal) Literature Search
Figure A-27. Literature search documentation for sources of primary data
pertaining to inhalation formaldehyde exposure and respiratory and immune-
mediated conditions.
This document is a draft for review purposes only and does not constitute Agency policy,
A-346 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
The selected studies were evaluated using a systematic approach to identify strengths and
limitations, and to rate the confidence in the results. Details of the evaluation considerations for the
observational epidemiology studies of allergic response based on history of specific conditions or
on skin prick tests, or asthma (current prevalentce, incidence, or asthma control) are described
below, followed by a summary of the evaluation of controlled human acute exposure studies.
Observational Epidemiology Studies
Ascertainment of allergic sensitization and allergies
EPA consulted with a group of experts14 regarding issues pertaining to ascertainment of
allergy sensitization and allergies in epidemiology studies. The group was given extracted
information regarding case ascertainment or outcome classification from 12 studies using
questionnaire-based measures or skin prick tests; descriptive information about the study
population (e.g., size, age, country) was also provided. The set included studies of formaldehyde
and of other exposures, but the material did not include any information regarding results.
The experts raised several points about the types of measures and interpretations of these
measures. The category includes allergic sensitization based on skin prick tests and history of
allergy-related symptoms. Sensitization may be present without clinical symptoms, and symptoms
may be present without a positive skin prick test Thus, these address different (but overlapping)
responses or conditions. The clinical expression of symptoms can be IgE-mediated or non-IgE
mediated; in most cases studies are not designed to make this distinction. The experts
recommended grouping the symptoms by site (i.e., nose and eyes; skin), and noted that food
allergies constitute a different type of group.
Questionnaire-based ascertainments of nasal and ocular symptoms have been developed
and widely used, for example in the International Study of Arthritis and Allergies in Children
(ISAAC) fAsher etal.. 19951. The additional ascertainment of seasonality and triggers can be
helpful in distinguishing between allergic and nonallergic basis of the symptoms. When comparing
specific types of self-reported allergies to specific types of positive skin prick tests, specificity of
self-report is relatively high (approximately 90% or higher), but sensitivity is lower (ranging from
30-70%) (see,for exampleLakwiiketal.. 1998: Braun-Fahrlander etal.. 1997: Dotterudetal.. 19951.
Limiting case ascertainment to physician-diagnosed allergies increases specificity but is considered
to have low sensitivity because self-treatment with nonprescription medications is common. For
studies of association, specificity is a more important consideration than sensitivity. It was also
noted that validation of the questionnaire-based instruments is more established in Europe and the
United States than in other populations.
Questionnaire-based ascertainments of atopic dermatitis or eczema have also been
developed (Williams etal.. 1996: Asher etal.. 1995). These questionnaires focus on the extent,
location, and itchiness of the rash and age at onset (typical onset before age 2 years). Specificity,
14Dr. Hasan Arshad, University of Southampton, Southampton^ United Kingdom; Dr. Peter Gergen, National
Institute of Allergy and Infectious Diseases, Bethesda, Maryland; Dr. Elizabeth Matsui, Johns Hopkins University,
Baltimore, Maryland; Dr. Dan Norback, Uppsala University, Uppsala, Sweden; Dr. Matthew Perzanowski, Columbia
University, New York City, NY.
This document is a draft for review purposes only and does not constitute Agency policy.
A-347 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
compared to physician diagnosis, was high (>0.95) in school-age children (Williams etal.. 19961
and in younger children fvon Kobvletzki et al.. 20131.
Based on the discussions with these experts, EPA made the following decisions:
• ISAAC questionnaires for rhinitis or rhinoconjunctivitis were considered to provide an
adequate basis for case ascertainment in studies in Europe and the United States; in studies
in other areas (i.e., areas that have not been included in ISAAC), specific mention of
validation of the questionnaire was needed to receive a high confidence rating. Although
the specificity of questions pertaining to rhinitis may be somewhat lower than the
specificity of questions pertaining to rhinoconjunctivitis fKim etal.. 20121. this difference
was not sufficient to conclude that the rhinitis questions should be viewed with lower
confidence.
• EPA had lower confidence in the symptom ascertainment in Matsunaga et al. (2008)
because this study was based on self-report of medical treatment (medication use) for
atopic eczema and for allergic rhinitis in the past year, without clarifying the type of
medication. EPA did not find studies examining the sensitivity or specificity of this
question-based assessment with respect to ascertainment of allergy history.
• EPA had lower confidence in allergy ascertainment in Fransman etal. f20031 because the
question included food as one of the types of allergies, and was not as specific regarding
symptoms as the ISAAC-based questionnaires.
• Skin prick test protocols in the set of studies ranged from 5 to 12 allergens; EPA did not
consider this difference to be sufficient to conclude that the protocols should be viewed
with different levels of confidence.
Longitudinal studies can examine the initial manifestation of the response (sensitization or
symptoms); cross-sectional studies can examine period-specific prevalence of allergies. Either
question can be relevant when thinking about the influence of environmental exposures. For
studies of incidence of allergies, the exposure measure should reflect a period before occurrence;
for studies of the prevalence of allergy symptoms, the exposure measure should reflect the same
period as the characterization of symptoms; for studies of allergy sensitization, the exposure
measure should reflect the period before or during which sensitization occurs.
• In the only study of incident allergies fSmedie and Norback. 20011. the baseline assessment
excluded children with a positive skin prick test. Measurements of formaldehyde in
classrooms were taken at baseline and again two years later; the end of the follow-up
period was two years after this measurement (4-year total follow-up). EPA considered this
protocol to reflect a relevant exposure period.
• Because of questions regarding the relevant time window of exposure, EPA had lower
confidence in skin prick test results for studies in adults than in children.
This document is a draft for review purposes only and does not constitute Agency policy.
A-348 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
Ascertainment of asthma
EPA also consulted with a group of experts15 regarding issues pertaining to ascertainment
of asthma in epidemiology studies. This group was given extracted information regarding case
ascertainment or outcome classification from 23 studies using questionnaire-based measures of
asthma, some of which included a validation component. As with the other group, descriptive
information about the study population (e.g., size, age, country) was also provided and the material
did not include any information regarding results for formaldehyde or other exposures.
The experts raised several points about the ascertainment of asthma and the terminology
used for different types of measures. Self- (or parent-) report of physician-diagnosed asthma can
be reliably used in epidemiological studies of incidence of asthma, although this method can miss
undiagnosed asthma. "Current" asthma, or prevalence of current asthma, is typically ascertained
through a set of questions pertaining to symptoms or medication use over of period of time (e.g.,
last 12 months). A similar, but usually expanded, set of questions can be used to assess asthma
control over a shorter period of time (e.g., 2-4 weeks). (Asthma control pertains to the extent to
which symptoms can be reduced or eliminated with medication.) Asthma exacerbation is a term
typically used in clinical trials and considers the need for using systemic corticosteroids. Most of
the studies identified in the formaldehyde literature are studies of prevalence of current asthma.
Most of the studies identified in this review used a classification scheme based on the
American Thoracic Society questionnaire (Ferris. 1978) or subsequent instruments that built upon
this work, including the ISAAC and European Community Respiratory Health Survey (ECHRS)
questionnaires. These questionnaire-based approaches have been found to have an adequate level
of specificity and positive predictive value for use in etiologic research fRavault and Kauffmann.
2001: Tenkins etal.. 1996: Burnev et al.. 19891. The questionnaires typically use several questions
to define current asthma based on symptoms relating to wheezing episodes or shortness of breath,
reported history of asthma attacks, or use of asthma medication. Using the question "Has a doctor
ever told you that you have asthma?" is a validated approach for the ascertainment of asthma
incidence. As noted in the discussion of ascertainment of allergies, the questionnaires have been
used in many studies but have not necessarily been validated in every population.
The age of study participants is an important consideration in the interpretation of various
measures. Specificity of symptom questions is reduced in the very young (<5 years) because
wheezing can occur with respiratory infections in infants and young children, and specificity is
reduced at older ages (e.g >75 years) because of the similarities in symptoms and medication use
for chronic obstructive pulmonary disease and asthma (Abramson et al.. 2014: Taffet etal.. 2014).
Asthma can be atopic (allergic) or nonatopic. In the United States 1988-1994 NHANES data,
56% of self-reported physician diagnosed asthma cases had at least one positive skin prick test
15Dr. Lara Akinbami, U.S. Centers for Disease Control, Atlanta, Georgia; Dr. Peter Gergen, National Institute of
Allergy and Infectious Diseases, Bethesda, Maryland; Dr. Christine Joseph, University of Michigan, Ann Arbor,
Michigan; Dr. Felicia Rabito, Tulane University, New Orleans, Louisiana; Dr. Carl-Gustaf Bornehag, Karlstad
University, Karlstad, Sweden.
This document is a draft for review purposes only and does not constitute Agency policy.
A-349 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
(Arbes et al.. 20051. Thus, the delineation of asthma into these different groups can reduce some of
the heterogeneity, but exclusion of either group may significantly reduce the sensitivity of case
ascertainment
Based on the discussions with these experts, EPA made the following decisions:
• ATS-based questionnaires or subsequent variations (ISAAC, ECHRS) for prevalence of
current asthma that include questions on medication use and symptoms were considered to
provide an adequate basis for case ascertainment in studies in Europe and the United
States; in studies in other areas (i.e., areas that have not been included in ISAAC), specific
mention of validation of the questionnaire was needed to receive this level of confidence.
• EPA had lower confidence in the asthma ascertainment in Matsunaga et al. (20081 because
this study was based on self-report of medical treatment (medication use) for asthma in the
past year. This ascertainment method may result in reduced sensitivity. The resulting
prevalence of asthma based on this definition was lower than found in a study by Mivake et
al. (20111. which was conducted in a similar population (women enrolled in a pregnancy
cohort in Japan) and used a broader definition based on symptoms and medication use
[asthma prevalence 2.1% and 5.5%, respectively, in Matsunaga etal. f20081 and Mivake et
al. (2011)]. With respect to specificity, this is a relatively young cohort (pregnant women,
median age approximately 30 years), suggesting that chronic obstructive pulmonary
disease would not be common.
• EPA had lower confidence in the asthma ascertainment in the study by Tavernier etal.
f20061 because of low specificity of the classification. The experts noted that three of the
five screening conditions were not specific to asthma (received more than three courses of
antibiotics for upper or lower respiratory symptoms in the past 12 months, have history of
fever or eczema, and family history of asthma in first degree relatives), and recommended
excluding this study. However, because the study did meet EPA's initial inclusion criteria,
EPA retained it but noted this limitation in the evaluation.
• Some studies included results for more than one asthma measure; in this assessment, EPA
based its evaluation on outcomes that were defined over a recent time period (e.g.,
symptoms in the past 12 months) and did not include outcomes defined over a lifetime (e.g.,
ever had asthma). Studies that did not clearly delineate the time period of ascertainment
were included, but EPA noted the lower confidence in these measures.
• Rumchev et al. f20021. a study of emergency room visits for asthma in children ages 6
months to 3 years was classified as not informative with respect to asthma. [NRC f20111
also recommended excluding Rumchev etal. (2002) on the basis of the age distribution.]
This study, in addition to two other studies that examined wheezing episodes among infants
(Roda etal.. 2011: Raaschou-Nielsen etal.. 2010). were thus excluded from the asthma
analysis, but are included in a separate section on lower respiratory tract symptoms in
infants and toddlers.
EPA also considered issues regarding the timing of the exposure with respect to the specific
outcome under study.
This document is a draft for review purposes only and does not constitute Agency policy.
A-350 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
• In the only study of incident asthma fSmedie and Norback. 20011. measurements of
formaldehyde in classrooms were taken at baseline and again two years later; the end of the
follow-up period was two years after this measurement (4-year total follow-up). EPA
considered this protocol to reflect a relevant exposure period.
• For studies of prevalence of current asthma (based on symptoms and medication use over
the past year), EPA looked for information that supported the suitability of the exposure
measure as a characterization of exposure during this time period. Examples include a
study that collected exposure measures in at least two seasons or that examined season in
the analysis.
• EPA considered exposure measures taken concurrently with completion of the asthma
questionnaire to reflect a relevant exposure period for studies of asthma control (symptoms
and medication use over the past 2-4 weeks).
• For results pertaining specifically to nighttime symptoms, EPA considered exposure
measures taken in the home to provide a more relevant exposure measure than school-
based exposures.
Exposure assessment
Based on the review of exposure assessments in the studies (see the general criteria for
Exposure Assessments for Epidemiological Studies, Appendix A.5.1), EPA made the following
decisions:
• EPA had lower confidence in the exposure measurements in two studies that used relatively
short sampling periods (30 minutes and two hours, respectively, in Dannemiller etal.. 2013:
Hsu etal.. 2012) and two studies in which the sampling time was not specified (Zhai etal..
2013: Choi etal.. 2009). (Neither of these two authors responded to an email inquiry from
EPA regarding this question.) Each of these four studies did contain some information
regarding the specifics of the sampling protocol or quality control procedures and
encompassed a wide range of exposures.
• Although Hwang etal. f20111 reported a geometric mean, this study did not provide more
complete information on distribution of exposure levels (e.g., 75th percentile, or maximum
value); thus, EPA also had lower confidence in the exposure description of this study.
• EPA also had lower confidence in the exposure measures of the study by Tavernier et al.
(2006). This study used a 7-day measurement period in two locations in the home, and
reported results by tertile of exposure. However, no information on the distribution of
exposure levels (e.g., cutpoints for the tertiles) was provided, so it is difficult to interpret the
results. The corresponding author did not respond to an email inquiry from EPA regarding
this information. [The paper by Gee etal. f20051 appears to be the same study; this paper
reported median levels of 0.03 and 0.04 ppm (0.037 and 0.049 mg/m3) in the living room
and bedroom samples.]
There was also variation in the exposure measurements used within the five occupational
studies identified in this search (Neghab etal.. 2011: Fransman etal.. 2003: Herbert etal.. 1994:
Malaka and Kodama. 19901 fHolness and Nethercott. 19891. with exposure assessments based on
This document is a draft for review purposes only and does not constitute Agency policy.
A-351 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
one or more area samples in specific task areas, personal samples, or a combination of both. For
hazard identification, an accurate characterization of "high" versus "low" exposure or "exposed"
versus "nonexposed" may be able to provide a sufficient contrast to examine associations, even if
there is considerable heterogeneity within the high exposure group. EPA considered the exposure
assessment in each of these five studies to be adequate for this purpose, but noted the relatively
high exposure [up to 0.08 mg/m3 in the "low" exposure group of the Fransman et al. (2003)] would
potentially result in an attenuated effect estimate.
Assessmen t of participan t selection
The process through which study participants are identified, recruited, and selected, in
addition to the participation rate, are important considerations in epidemiology studies. A
selection bias can be introduced if both the exposure and the outcome (disease status) is directly or
indirectly related to likelihood of participation. For the general population studies, EPA made the
following decisions:
• EPA had high confidence in recruitment strategies based on geographic-based or
population-based sampling frames (e.g., of residences or schools). However, EPA had lower
confidence for the studies with this design that also had very low participation rates
[(<20%)Billionnetetal. (2011) Hsu etal. (2012): Hwang etal. (2011) Matsunaga etal.
("200811.
• EPA also had lower confidence in clinic-based, case-control studies that did not report any
details of the recruitment of selection process. Choi etal.. 2009: Rumchev etal. f20021. and
in case-control designs that were not drawn from a defined population Garrett etal. f!999a.
b).
• EPA had low confidence in the selection process in the case-control study by Tavernier et al.
f2006I Although cases and controls were drawn from two primary care practices, 95 cases
were excluded because no age- and sex- matched control was identified.
A primary consideration regarding participant selection in the occupational exposure
studies was the recruitment of current workers, that is, workers who remained in a workplace for
some time (e.g., 2 or more years). This type of design could result in the "healthy worker effect,"
resulting in the potential loss of affected individuals from the workforce. EPA noted this as a
limitation in all of the occupational studies. The participation rate in one of these studies was 66%
(Fransman etal. (2003)). and ranged from 87-100% in the other four studies. EPA did not consider
this difference to be sufficient to conclude that the protocols should be viewed with different levels
of confidence.
Assessment of potential confounding and other analysis issues
EPA approached the evaluation of potential confounding by considering critically important
risk factors that could also be related to formaldehyde exposure (and are not in the causal
pathway). Age and sex were considered key demographic variables, although it is not likely either
This document is a draft for review purposes only and does not constitute Agency policy.
A-352 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 is associated with variability in indoor formaldehyde levels. EPA also examined information on
2 potential correlation between formaldehyde and other air pollutants associated with allergy or
3 asthma; the specific measures differed depending on the setting. The evaluation of the control for
4 confounding was not based on whether a particular variable was or was not included in a model;
5 rather a broader array of information was used, including the approach to modeling and
6 information on patterns of exposure in the specific study population.
7 Based on these considerations, EPA made the following decisions:
8 • EPA had low confidence in three studies because of evidence of confounding that could not
9 be addressed (Yeatts etal.. 2012: Choi etal.. 2009: Smedje etal.. 1997: Norback etal.. 1995).
10 Two of these studies could not distinguish between effects of formaldehyde and effects of
11 other exposures strongly correlated with formaldehyde fYeatts etal.. 2012: Smedie etal..
12 1997: Norback etal.. 19951. and the third fChoi etal.. 20091 did not address risk factors foi-
ls the outcomes that were shown to vary between cases and controls, and that could
14 reasonably be postulated to also be related to formaldehyde levels.
15 Reasons for different ratings within a study
16 • In some cases, different evaluation ratings were given for the different outcomes or
17 analyses included a study:
18 For Palczvnski et al. (1999). the difference in evaluation ratings for children and adults foi-
ls the skin prick test analyses is based on greater uncertainty regarding the timing of the
20 exposure measure in this outcome in these two groups.
21 For Garrett etal. f!999a. 1999bl. the inclusion of approximately 30% of the controls from
22 the same household as the asthma cases and the inability to distinguish between ever-
23 and current asthma resulted in a low confidence rating for the asthma analysis and a
24 medium confidence rating for the skin prick test analysis.
25 For Fransman etal. (2003). the ratings for allergies (low confidence) differed from that of
26 asthma (medium confidence), due to the uncertainty regarding the specificity of the
27 questions used to ascertain allergy history.
28 For Herbert etal. T19941 uncertainty about time window of exposure measurement
29 with respect to skin prick test results resulted in a "low" confidence rating for that
30 analysis and a "medium" confidence rating for the asthma analysis.
31 Summary of reclassification of studies
32 This evaluation process resulted in the refinement of the inclusion criteria for asthma: the
33 eligible population for asthma was changed from "humans" to "humans, age >4 years" because the
34 respiratory disorder occurring in infants and toddlers may be related to, but is distinct from,
35 asthma, which is more reliably diagnosed in school-aged children. As noted previously, four studies
36 that had been identified as asthma studies were thus reclassified as studies of "lower respiratory
37 tract symptoms in infants and toddlers." These studies, and the reasons for this reclassification,
38 are:
This document is a draft for review purposes only and does not constitute Agency policy.
A-353 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Supplemental Information for Formaldehyde—Inhalation
• Raaschou-Nielsen etal. f20101—limited to infants; outcome = wheezing episodes
• Roda etal. f20111—limited to infants; outcome = lower respiratory tract infection (with and
without wheeze episode)
• Rumchev et al. f20021—limited to ages 6-36 months; outcome = asthma based on
emergency room discharge data
Considerations of alternative classifications
This evaluation process necessarily results in the categorization of what is essentially a
continuous measure (confidence level). In some cases, different overall confidence levels could be
supported, depending on the emphasis that was placed on different strengths and limitations. In
these situations, EPA considered the impact of alternative classifications. For examples, Smedie and
Norback (2001) is the only study that examined incidence of allergies or asthma; the prospective
design is a considerable strength of the study. However, the exposure assessment (conducted in
classrooms in the baseline year and in Year 3 of the four-year follow-up) was limited by a high
prevalence of values below the detection limit (54% of 1993 samples and 24% of 1997 samples
were below 0.005 mg/m3; geometric mean 0.004 and mean 0.008 mg/m3), resulting in
uncertainties in interpreting the analysis conducted using formaldehyde as a continuous measure.
EPA classified this as a low confidence study because of the analysis, but also conducted a
sensitivity analysis using an alternative classification of medium confidence.
Summary of overall evaluation of confidence
Based on the considerations described above, EPA developed an overall evaluation of its
confidence in each study (or a specific analysis within a study), with high, medium, and low
confidence categories. Table A-50 describes the criteria used in this classification. Because the
exposure assessment was a primary consideration in this evaluation, it is presented as a separate
column, with other aspects of study design and analysis combined in another column. The
subsequent table in this section provides the more detailed documentation of the evaluation of
observational epidemiology (see Table A-51); studies are arranged alphabetically within this table.
Table A-50. Criteria used to assess epidemiologic studies of respiratory and
immune-mediated conditions, including allergies and asthma, for hazard
assessment
Overall
evaluation
Exposure assessment
Study design and analysis
High
confidence
General population: Exposure measure based
on at least 3-day sample, corresponding to
appropriate time window (e.g., measures in
more than one season if time window covers
12 months, or addressed season in the
analysis. For inferences above 0.050 mg/m3,
exposure range includes large enough sample
High specificity of outcome ascertainment;
participant selection based on population-
based sampling frame with high participation
rate; confounding considered and addressed in
design or analysis; analysis allows for
examination of variation in effect in relation to
variation in exposure level using analytic
This document is a draft for review purposes only and does not constitute Agency policy.
A-354 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Overall
evaluation
Exposure assessment
Study design and analysis
above 0.050 mg/m3 to allow for meaningful
analysis in this range.
Work settings: Ability to differentiate
between exposed and unexposed, or
between low and high exposure.
procedures that are suitable for the type of
data. Large sample size (number of cases)
Medium
confidence
General population: More limited exposure
assessment, or uncertainty regarding
correspondence between measured levels
and levels in the etiologically relevant time
window.
Work settings: Referent group may be
exposed to formaldehyde or to other
exposures affecting respiratory conditions
(potentially leading to attenuated risk
estimates)
Uncertainty regarding specificity of outcome
ascertainment or participant recruitment
process; confounding considered and addressed
in design or analysis but some questions
regarding degree of correlation between
formaldehyde and other exposures may
remain. Total sample size adequate but limited
in stratified analyses.
Low
confidence
General population: Short (<1 day) exposure
measurement period without discussion of
protocol and quality control assessment.
Low specificity of outcome ascertainment; high
likelihood of confounding that makes it unable
to differentiate effect of formaldehyde from
effect of other exposure(s), limited data
analysis (or analysis that is not appropriate for
the data) or small sample size (number of cases)
Excluded
(not
informative)
Exposure range does not allow meaningful
analysis of risks above 0.010 mg/m3
This document is a draft for review purposes only and does not constitute Agency policy.
A-355 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-51. Evaluation of allergy and asthma studies
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Annesi-
Maesano et
al. (2012)
(France)
Schools:
children
(prevalence
survey)
Schools randomly
selected from
defined
geographic area,
ages 9-10 years.
Participation rate
81% in initial
survey, 69% with
full protocol.
5-day samples
in classrooms;
sampling from
108 schools; all
classes of
specified grade
level per
school.
Median (75th
percentile)
0.027 (0.034)
mg/m3
(estimated
from figure).
Protocol
discussed.
ISAAC questionnaire
Allergy:
"sneezing and runny
nose accompanied by
itchy eyes out of cold in
the past year"
Asthma:
asthma in past year
(wheezing or whistling
in the chest or
wheezing or whistling
in the chest at night-
time or
taken asthma
treatment in the past
year)
Exercise induced
asthma based on
response to pulmonary
function testing after
exercise protocol.
Exposure measurement
blinded to outcome
classification
Adjusted for age,
gender, passive
smoking, and
paternal or
maternal history
of asthma and
allergic diseases.
Also examined
dampness, gas
appliances,
ethnicity,
socioeconomic
status, and
season.
Other measures
included: NOx,
PM2.5,
acetaldehyde,
acrolein
Generalized
estimating
equation
modeling,
accounting for
nonindependenc
e of observations
within-area
(schools)
environment,
including climate.
OR (95% CI) (CI
estimated from
figure). Models
took into account
within city
correlations
among
participants.
Additional
stratification of
asthma analysis
by atopy status.
Sensitivity
analysis: exercise
induced asthma
limited to
measures in
same week (n =
4,643)
6,683
Allergy
(rhinoconjunctivitis) and
Asthma
SB IB Of Oth
Overall
Confidence
High
No other pollutants were
associated with
rhinoconjunctivitis. PM2.5
and acrolein were
associated with asthma.
This document is a draft for review purposes only and does not constitute Agency policy.
A-356 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Billionnet
et al.
(2011)
(France)
Residences:
adults
(prevalence
survey)
October
2003-
December
2005
Nationally
representative
sample of
residences (Indoor
Air Quality
Observatory
study); 13.6%
participation rate
(567 of 4,165
households). Low
participation rate
1-week sample
in bedroom;
75th percentile
0.028 to
mg/m3.
Protocol
discussed.
ISAAC questionnaire:
Rhinitis based on self-
report of, in the past 12
months, sneezing,
running or blocked
nose without cold or
respiratory infection.
ECRHS: Asthma based
on one of following
criteria: (i) having an
asthma attack in the
last 12 months; (ii)
having been woken by
an attack of shortness
of breath in the last 12
months; and (iii)
currently using asthma
medicine. Exposure
measurement blinded
to outcome
classification
Covariates
chosen if
associated with
asthma or rhinitis
and affecting one
or more effect
estimates for
volatile organic
compound
exposure
measures by 20%
or more.
Adjusted for age,
gender, smoking,
education,
relative humidity,
time of survey,
pets, mold,
outdoor pollution
sources within
500 meters. Did
not specifically
address
correlation
between
formaldehyde
and other
exposures(other
than noting that
these were not
among the higher
correlations
seen).
Generalized
estimating
equation
modeling,
accounting for
nonindependenc
e of within-area
(dwellings)
observations. OR
(95% CI)
(estimated from
figure).
Additional
models took into
account within
dwelling
correlations
among
participants.
Compared
nonparticipants
(pollutant
measures but no
health
questionnaire)
and participants.
Sensitivity
analysis excluding
relatives.
1,012
Allergy (rhinitis) and
asthma
SB
IB Cf Oth
Overall
Confidence
Medium
N
Low participation rate but
potential for diffential
participation (by
formaldehyde exposure
and disease status)
unlikely.
Wheezing
Not informative
Analyses included ages 3 ¦
10 years of age
(Branco et
al.. 2020)
(Portugul)
A total of 1530
preschoolers
(n=648 3-5 years)
and primary
Daily exposure
based on time-
averaged air
concentration
The ISAAC
questionnaire was
completed by parents
or guardians, which
Potential
confounders
selected based
on previous
Multivariate
logistic regression
for each
individual
N = 1530
This document is a draft for review purposes only and does not constitute Agency policy.
A-357 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
School:
children
(prevalence
survey)
2013 - 2016
school children
(n=882 6-10 years)
were randomly
recruited from
urban and rural
nursery (n=17)
and primary
schools (n=8)
participating in
the INAIRCHILD
project. There
were two phases
in 2013/2014 and
2015/2016.
Children < 3 years
were excluded.
Participants
represented 39%
of the original
sample. No
comparisons of
participants and
nonparticipants.
42% were aged 3-
5 years, with less
specific asthma
diagnosis. Low
participation
raises concern for
selection bias. PFT
was only
conducted in the
49% who reported
wheezing or
asthma diagnosis
possibly
introducing bias in
and reported
time in specific
school
locations.
Continuous
monitoring in
each room (24
h to 9 days)
{Branco, 2019,
HERO}. Time-
activity
obtained from
parents' 24-
hour daily
diary, class
timetables and
teachers.
Inhaled daily
dose estimated
using time-
averaged
exposure,
inhalation rate
for each
activity {EPA,
2011, HERO}
and body
weight. Mean
HCHO
concentration
(SD) 35.3(43.1)
Mg/m3);
were validated by
physicians. Spirometry
measurements were
taken in participants
identified as asthmatic
from the questionnaire
responses or reporting
ever having one or
more asthmatic
symptoms (wheezing,
dyspnea, or nocturnal
cough with no upper
respiratory infection)
(of 763, missing or
failed in 269).
Spirometry before and
after bronchodilator
using ERS/ATS and
Global Initiative for
Asthma guidelines
conducted by pediatric
doctors with pulmonary
specialization. Methods
and OA described.
Asthma diagnosed
based on symptoms (>
1) and PFT results using
GINA guidelines. Skin
prick tests conducted
on children with PFT
results using several
aeroallergens (n=341,
missing or failed for
153).
Outcomes: reported
active wheezing in last
12 months (relevant to
experience and
included site
(urban, rural),
study phase, sex,
age group, BMI
and parental
history of
asthma. Also
controlled for
surrogates of
home indoor
exposure
including
mother's
education, living
with smoker.
Other covariates
for contact with
farm animals
during 1st year of
life, pets at home
in previous year
&/or 1st year of
life.
pollutant as
continuous
variable (per IQR)
or dichotomized
using median, or
regulatory
cutoffs. Models
also for all
pollutants
simultaneously.
Asthma diagnosis
SB
IB
Cf Oth
Overall
Confidence
Low
1
¦
Concern regarding
potential for selection
bias (low participation and
missing values) and
decreased specificity of
asthma diagnosis by
including very young
children (< 5 years)
This document is a draft for review purposes only and does not constitute Agency policy.
A-358 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
PFT endpoints.
Missing PFT data
for 269 of 763
selected (35%).
pre-schoolers);
reported asthma (does
child have or ever had
asthma?); diagnosed
asthma by study
physicians, FEV1/FVC
<0.90, reduced FEV1
(<80% predicted),
asthma diagnosed in
5.5%, asthma with or
without aeroallergen
sensitization, and no
asthma. (Inclusion of
notable proportion of
children aged <5 years
likely decreased
specificity of asthma
diagnosis.
Choi et al.
(2009)
(Korea)
Residences:
children (and
adults?)
(case-control
study)
March-June
2006
Conducted in
university
outpatient clinic;
recruitment
procedure for
cases or controls
not described.
Mean age cases
15.4 years (SD =
3.4; controls 16.2
years (SD = 4.1)
Household
sample in living
room at
location away
from sources
of VOCs
(sampling
period not
reported, but
closed
windows, no
smoking or use
of potential
sources, and
use of
duplicates).
Geometric
mean 0.043
mg/m3, 75th
Atopic dermatitis and
allergic asthma: based
on medical history, skin
prick test and IgE
(criteria not provided)
No information
on
socioeconomic
status; higher
percentage of
cases lived near
roads or in
industrial area
(21%, 34%, 44%
of controls,
dermatitis, and
asthma cases,
respectively).
Housing age <3
years old in 29%,
40%, and 58% in
controls,
dermatitis, and
asthma cases,
Nonparametric
(Mann-Whitney)
comparison of
formaldehyde by
group; geometric
mean, 25th, and
75th percentiles
reported.
50 atopic
dermatitis
cases, 36
asthma
cases, 28
controls
Allergy (atopic
dermatitis) and lower
respiratory tract
symptoms in infants and
toddlers
SB IB Cf Oth
Overall
Confidence
Low
¦
Selection and recruitment
process not reported;
sampling period not
reported and specific
criteria for case definition
not reported; potential
confounders (age and
type of housing and
location differed between
This document is a draft for review purposes only and does not constitute Agency policy.
A-359 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
percentile
0.115 mg/m3.
respectively; and
50%, 56%, and
72% of controls,
dermatitis cases
and asthma cases
lived in
apartments.
cases and controls, as
measure of
socioeconomic status) not
addressed. Limited
analysis.
Dannemill
er et al.
(2013)
(United
States)
Residences:
children
(asthma
control)
July 2008-
February
2010
Related
reference:
Sandel et
al. (2014)
Low-income
homes in Boston,
recruited from
past allergy
cohorts, asthma
clinics, newspaper
ads, and referrals
from other
participants.
(Boston Allergen
Sampling Study).
79% (37 out of 47)
participated in this
analysis. Mean
age 10.5 years.
Boston Allergen
Sampling Study.
30-minute
pumped air
sample in
kitchen.
Median 0.044
mg/m3;
31% >0.060
mg/m3;
maximum =
0.162 mg/m3.
Protocol
discussed;
analysis of
sources of
exposure
Asthma control (5
questions) [based on
validated
questionnaire];
symptoms and inhaler
use in past 4 weeks
Examined season,
temperature, and
relative humidity
(email from
Karen
Dannemiller to
Glinda Cooper,
May 6, 2015)
Logio-
transformed
formaldehyde;
t-tests.
37 asthma
cases(out
of 47
children in
study, 79%)
Asthma control
SB IB Cf Oth
H
Overall
Confidence
Medium
Recruitment was not from
a well-defined population.
Limited exposure
measurement period (but
quality control details
provided).
Fransman
et al.
(2003)
(New
Zealand)
Wood
workers
(prevalence
survey)
Plywood mill
workers,
participation rate
66%. Internal
comparison by
exposure level.
Mean duration 4.7
years in mill, 2.7
years in current
job. Workers'
knowledge of
Personal
samples (15-
minute
samples);
above 0.100
(geometric
mean 0.260
mg/m3). Limit
of detection
0.030 mg/m3.
Allergy symptoms:
self-report of sensitivity
to house dust, food,
animals or
grasses/plants.
Asthma:
Current asthma
medication use; past
12 months, asthma
attack or being woken
by shortness of breath
Adjusted for age,
gender, ethnicity,
and smoking for
comparisons
between high
and low exposure
within workplace.
Weaker
association seen
with terpenes.
Inhalable dust,
Logistic
regression, OR
(95% CI)
112
Allergy (allergy
symptoms)
SB IB Cf Oth
m
Overall
Confidence
Uncertain impact of
outcome classification
and uncertainty regarding
details of analysis; see
This document is a draft for review purposes only and does not constitute Agency policy.
A-360 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
formaldehyde
exposure not
discussed.
abietic acid, and
endotoxin also
measured but not
clear if these
were considered
in the analysis of
the allergy
symptoms data
asthma discussion for
other limitations
Asthma
Overall
SB IB
i;t
IH-h
Confidence
y
Medium
Selection out of the
exposed work force of
"affecteds" possible in this
type of prevalence study.
"Low" exposure group
exposed to levels of
formaldehyde up to 0.080
mg/m3. Either limitation
would result in reduced
(attenuated) effect
estimate.
Garrett et
al. (1999a,
1999b)
(Australia)
Residences:
children
(prevalence
survey)
Combined analysis
of cases and
controls from a
case-control study
of asthma in two
rural towns.
Recruitment
through schools
and medical
centers; additional
advertisement for
nonasthmatic
children. 30 of the
95 controls were
from same
households as
cases; the 65
other controls
4-day
household
samples (4
seasons),
multiple
locations; up to
0.139 mg/m3.
Protocol
discussed.
Separate paper
about
exposure
measures. 74%
of children had
lived in same
house for at
least 5 years.
Allergy:
12 allergen skin prick
test (cat, dog, grass mix
#7, Bermuda grass,
house dust, 2 dust mite,
5 fungi).
Asthma
Parent report of doctor-
diagnosed asthma.
Mean score 4.6 in
asthma cases, 0.7 in
controls on respiratory
symptom questionnaire
completed at last home
visit (symptom
frequency, 4 categories,
over past year of:
cough, cough in the
Adjusted for
parental asthma
history, sex;
other factors
examined but not
needed in final
model (passive
smoke, pets,
indoor N02,
fungal spores,
house dust mite
allergens)
Prevalence (n, %)
by exposure
group; logistic
regression, OR
(95% CI); figure
showing wheal
size and number
of positive
responses by
exposure group.
Evaluated
relation between
formaldehyde
and NOx, house
dust, fungal
spores, housing
age.
145 in
allergy
analysis; 53
cases, and
95 controls
in asthma
case-
control
analysis
Allergy (skin prick tests)
SB
IB
a oth
Overall
Confidence
Medium
*
Uncertainty about about
effect of recruitment
process and about time
window of exposure
measurement with
respect to skin prick test
results.
Asthma
This document is a draft for review purposes only and does not constitute Agency policy.
A-361 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
were from 37
households.
morning, shortness of
breath, waking due to
shortness of breath,
wheeze/ whistling,
asthma attacks, chest
tightness, and chest
tightness in the
morning).
Exposure measurement
blinded to outcome
classification.
SH
IK
Ct
<>rh
Confidence
LOW
¦
Uncertainty about asthma
definition (current asthma
or ever asthma?).
Uncertainty about effect
of recruitment process
and ability to fully address
household correlation of
cases and controls; could
result in attenuated effect
estimate. Incomplete
reporting of results
(adjusted results reported
as "not statistically
significant")
Herbert et
al. (1994)
(Canada)
Wood
workers
(prevalence
survey)
Related
reference:
Herbert et al.
(1995)
Oriented strand
board
manufacturing,
mean duration 5.1
years. Referent
group = oil field
workers, not
exposed to gas or
vapors, mean
duration 10.0
years.
Participation rate
98% in workers,
82% in
comparison group.
99 exposed, 165
referents.
Because both
Area samples.
21 hours
continuous
sampling on
two separate
days); range
0.090 to 0.330
mg/m3
Allergy:
6 allergen skin prick
test (wheat, rye,
Alternaria, cat, house
dust, birch).
Asthma:
International Union
Against Tuberculosis
and Lung Disease
(1986) questionnaire
(asthma; lower
respiratory tract
symptoms (list includes
woken by shortness of
breath; attacks of
wheeze, wheeze with
chest tightness.)
[increased prevalence
Adjusted for age
and smoking;
dust measured
and reported as
low, not included
in analysis
Logistic
regression, OR
(95% CI);
prevalence of
"outcome"
(positive
responders) not
reported
99
exposed;
165
referents
Allergy (skin prick tests)
SB IB Cf Oth
Overall
Confidence
Low
N
Uncertainty about time
window of exposure
measurement with
respect to skin prick test
results; some uncertainty
about referent group.
Asthma
SB
IB Cf Oth
Overall
Confidence
Medium
N
Selection out of the
This document is a draft for review purposes only and does not constitute Agency policy.
A-362 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
groups are
"exposed"
workers, healthy
worker effect
unlikely. Some
uncertainty about
effect of
exposures in the
referent group
of lower respiratory
tract symptoms
associated with lower
FEVi or FEVi/FVC in
these workers]. Time
frame of asthma
definition interpreted
to be relevant to
occupational exposure.
Exposure measurement
blinded to outcome
classification
exposed work force of
"affecteds" possible in
this type of prevalence
study, and some
uncertainty about
referent group.
(Holness
and
Nethercot
1.1989)
(Canada)
Funeral home
workers
(prevalence
survey)
Participants
recruited from list
of funeral homes,
86.6%
participation;
79.8% of
embalmers were
active embalmers
(healthy workers);
community
referent (service
organization and
students)—
potential
differences
(weight, smoking)
2 area samples
(impingers),
during
embalming, 30
to 180
minutes.
Range in
exposed 0.10-
1.0 mg/m3,
referent mean
0.025 mg/m3;
adequate
exposure
contrast likely
for comparison
of exposed and
referent.
American Thoracic
Society (1978)
questionnaire: wheeze
(no details of
questions)
Univariate
analysis; did not
consider other
variables
Frequency by
group and p-
value from a
logistic regression
N=84
exposed;
N=38
referents
SB IB Cf Oth
Overall
Confidence
Low
Uncertainty regarding
asthma definition.
Selection out of the
exposed work force of
"affecteds" possible in
this type of prevalence
study; would result in
reduced (attenuated)
effect estimate. No
consideration of potential
confounding
Hsu et al.
(2012)
(Taiwan)
Residences:
children
(case-control)
Initially recruited
through randomly
selected
kindergartens and
day care centers;
73% of
successfully
2-hour
household
sample
(probably
bedroom);
Median 0.076
Initial screening
through parent report
of history of 2 or more
diseases (asthma,
allergic rhinitis) or
symptoms (wheezing,
coughing at night,
None addressed
in analysis.
Similar season
distribution in
cases and
controls
Mann-Whitney U
test for case-
control
differences in
exposure
distribution.
Median, 25th and
48 allergic
rhinitis, 36
eczema, 9
asthma
cases, and
42 controls
Allergy (rhinitis, eczema)
and asthma
This document is a draft for review purposes only and does not constitute Agency policy.
A-363 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
August 2008-
September
2009
contacted agreed
to send
questionnaires to
families and 68%
of the
questionnaires
were completed.
Selected for
follow-up if had
not moved or
renovated house
since birth. Of the
980 potential
cases and 802
potential controls
selected, 267
(27%) and 89
(11%) participated
in clinical exam; 59
cases and 42
controls (22% and
47% of cases and
controls,
respectively,
completing exam)
also completed
home exposure
measures.
mg/m3; 75th
percentile
0.030 mg/m3.
Limited
sampling
period with no
information on
protocol.
eczema, sneezing,
runny or stuffy nose)
during last 12 months;
confirmation of asthma,
rhinitis, and eczema by
clinical examination.
Controls answered "no"
to all of the disease and
symptom questions.
Exposure measurement
blinded to outcome
classification
75th percentiles
given for cases
and controls. P-
values reported if
<0.10. No
additional
modeling of the
formaldehyde
data undertaken.
SB IB Cf Oth
Overall
Confidence
Low
Low and differential (at
various steps)
participation rate. Short
exposure sampling period
and no information on
protocol. Limited
analysis. Uncertainty
regarding distribution (%
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
2) Rural area;
nested case-
control study of
asthma (FERMA)
(rural sampling
fro regular
contact with
farm animals)
Examined
nonparticipants
nonindependenc
e of participants
in similar
neighborhood.
Assessed
collinearity with
other measures
(NOx, PM2.5)
(but 9 rural
and 7
urban
excluded,
unspecified
number
excluded
from
analysis
limited to
current
asthma
analysis of current
asthma).
Hwang et
al. (2011)
(Korea)
Residences:
children
(case-control)
May 2008
Case-control
study, drawn from
1,005 elementary
students (one
school, all grades)
(84% participation
rate). 33 cases
(out of 129?) and
40 controls (out of
unspecified
number) agreed to
participate in
environmental
measurement
study. Controls
selected from
respondents with
no asthma
symptoms or
diagnosis, age and
sex matched to
cases.
3-day
household
sample (2
rooms) and
personal
sample.
Geometric
mean,
controls: 0.036
mg/m3 (no
information on
upper
distribution
reported).
Self-report asthma
symptoms or physician-
diagnosed asthma
based on ISAAC
questionnaire
Adjusted for age,
gender, income,
parents'
education,
passive smoking
Log-transformed;
logistic
regression, OR
(95% CI)
33 cases,
40 controls
Asthma
SB
IB
a oth
Overall
Confidence
Low
*
Asthma definition does
not distinguish between
current asthma and ever
asthma. Uncertainty
regarding selection
processes [high
prevalence of family
history of asthma in cases
(86%) and controls (96%)];
uncertainty about analysis
and distribution
(Huang et
al.. 2017)
Participants in a
previous cross-
sectional study
Continuous
formaldehyde
sampling in
History of airway
diseases using
translated ISAAC
Covariates
considered in
models based on
Differences
between cases
and controls
N =409
Current rhinitis
This document is a draft for review purposes only and does not constitute Agency policy.
A-365 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
(Shanghai,
China)
Residences:
children
(case-control)
March 2013-
December
2014
(2011-2012)
selected from 88
kindergartens
located in 6
Shanghai districts
(note: references
for cross-sectional
study stated 72
kindergartens
selected in 5
districts, N =
14,884). Included
if homes were not
renovated in the
previous 2 years
and agreed to an
on-site home
inspection, N=454
residences, 4.5%
of cross-sectional
survey for 10,182
participants with
contact
information (409
of 454 residences
assessed), 5 -10
years old. Concern
for selection bias
since eligibility
was based on ever
asthma status and
home renovation.
child's
bedroom, 24
hours, in
breathing zone
(detection
range: 0.012-
0.08 mg/m3).
Monitors
calibrated
before
sampling.
Average
concentration
(|ig/m3), 24-hr
21.5 ±13; 6-hr
22.2 ±17.9
Range 6.0 -
60.0 Mg/m3,
with 2
bedrooms
higher
Short sampling
duration less
likely to
represent
concentrations
over the
previous year
questionnaire; cases
responded "yes" to
symptom/disease
question in either
phase (cross-sectional
or case-control phases)
from questionnaire.
Current rhinitis: In the
past 12 months, has
your child had a
problem with sneezing,
or a runny, or a blocked
nose when he/she did
not have a cold or the
flu?
literature and
previous
analyses,
included age, sex,
family history of
atopy, family
annual income
level, household
ETS, household
dampness-
related
exposures,
antibiotics
exposure during
1st year of life,
home decoration
around time of
birth, season of
sampling. Higher
proportion of
homes with
mechanical
ventilation
among current
rhinitis cases
compared to
controls (77.5%
versus 65%)
compared using
Kolmogorov-
Smirnov test.
Multiple logistic
regression
models per IQR
increment or
quartile of
formaldehyde
concentration.
SB
IB Cf
Oth
Overall
Confidence
Low
1
Concern for selection bias,
difference in ventilation
methods by case status
suggests uncontrolled
confounding, Low
formaldehyde
concentrations
(Isa et al.,
2020a)
(Malaysia)
8 randomly
selected schools in
Hulu Langat,
Selangor,
Malaysia,
Formaldehyde
concentrations
measured
during class
time using
Asthma & allergy
information and
symptoms within
defined period using
ECRHS and ISAAC
Regression
models
controlled for
atopy, sex,
doctor's
2-level hierarchic
multiple logistic
regression, OR
(95% CI).
Concerns for
N=470
Allergy (rhinitis, dermal,
skin prick tests)
This document is a draft for review purposes only and does not constitute Agency policy.
A-366 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Schools:
children
(prevalence
survey)
August-
November
2018 &
February
2019
randomly selected
students from 4
classes (Form two,
aged 14 years).
Excluded students
reporting smoking
in last 12 months
or treated with
antibiotics in last 4
weeks.
Participation not
reported.
PPM
Formaldemete
r(accuracy of
10% at 2 ppm).
Monitors 1
meter from
ground in
center, 4 one-
hour periods.
Concentration
(reported as
mg/m3, but
appears to
have been
Mg/m3) median
(IQR)
Urban 13.2
(9.3); Suburban
3.1 (5.2)
Uncertainty in
concentrations
given short
sampling
duration
questionnaires.
Responses were blind
to environmental data.
Allergy skin prick test
for mites, fungi and cat
allergens after 15
minutes measuring
wheal diameter (atopy
defined as > 3 mm).
Respiratory symptoms
in last 12 months:
wheezing, daytime
breathlessness,
nocturmal attacks of
breathlessness. Allergic
symptoms in last 12
months: rhinitis, skin
allergy.
diagnosed
asthma, parental
asthma/ allergic
and location of
schools.
No adjustment
for ETS.
Associations also
observed for N02
- unknown
impact of
confounding on
formaldehyde
associations.
choice of
exposure metric
(continuous
variable) with no
information
about
distribution
below the LOD.
SB
IB
Cf
Oth
Overall
Confidence
¦
Low
Low
Uncertainty in exnosure
concentrations and
distribution eiven short
samnline duration verv
low concentrations in half
the schools with unclear
nronortion of samnles less
than the I OD and analysis
usine concentration as a
continuous variahle
Particination details not
reported.
Kim et al.
(2011)
(Korea)
Schools:
children
(prevalence
survey)
November-
December
2004
12 schools, 2-3
randomly selected
classrooms per
school
Participation rate
96%; 450 excluded
based on missing
data)
7-day samples
in classrooms.
1 SD above
mean = 36
Mg/m3;
maximum = 47
Mg/m3.
Protocol
discussed,
closed
windows.
Current medication use
or had asthma attack in
past 12 months.
Exposure measurement
blinded to outcome
classification
Adjusted for age,
sex, self-reported
pet or pollen
allergy,
environmental
tobacco smoke at
home, other
home
environment
(indoor
dampness,
remodeling,
changing floor,
Logistic
regression, OR
(95% CI) per 10
Mg/m3 increase;
additional
modeling to
account for
within school and
within city
correlations.
2,365
Asthma
SB IB Cf Oth
Overall
Confidence
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-367 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
age of home). All
samples within
same season.
Krzyzanow
ski et al.
(1990)
(United
States,
Arizona)
Residences:
adults,
children
(prevalence
survey)
Related
references:
Quackenboss
et al. (1989a):
Quackenboss
et al. (1989b)
Selected from 202
households
(stratified sample
from municipal
employees). 2,322
completed
baseline survey;
subgroups
selected based on
housing
characteristics
(type, age,
remodeling).
Clusters within
similar outdoor
PM and pollen
levels.
Participation rate
not reported but
sampled
nonresponders:
higher proportion
of current
smokers among
refusals (35%
versus 27%)
Two one-week
household
samples
(different
seasons),
multiple
locations;
Mean 0.032
mg/m3;
maximum
0.172 mg/m3
(most <0.074,
only a few
above 0.110
mg/m3)
Protocol
discussed
(separate
paper).
Asthma: American
Thoracic Society (1978)
questionnaire; doctor-
diagnosed asthma (ever
and current) and
symptom questions:
wheezing apart from
colds, 2 or more attacks
of shortness of breath
with wheezing in last
year. Exposure
measurement blinded
to outcome
classification
Environmental
tobacco smoke.
Also examined
N02
Contingency
tables, stratified
by age group and
for children, by
environmental
tobacco smoke
exposure.
Adults: 613
Children:
298
Asthma, children and
adults
SB IB Cf Oth
Overall
Confidence
Medium
For children, relatively
small # in higher exposure
categories. For adults,
incomplete reporting of
results.
Lajoie et
al. (2014)
(Quebec,
Canada)
Intervention
study October
Asthmatic children
with exacerbation
requiring medical
care in the past
year referred by
physicians at
tertiary care
center, 3-12
Pre and post-
intervention.
Passive air
sampling for
formaldehyde
in bedroom, 6-
8 days, during
winter and
Variable number with
complete data for each
outcome. Participants
were not blinded,
although technicians
were.
Formaldehyde-specific
Intervention/Control
Potential
confounders for
asthma outcomes
were age,
gender,
parents' level of
education, and
eczema.
Power calculation
reported.
Multivariate
linear models
Formaldehyde
analyses used
results in
intervention
For ISAAC
questionnai
re,
interventio
n n = 43,
control =
39
Current asthma
symptoms
SB IB
Cf
Oth
Overall
Confidence
¦
Medium
Medium confidence
Small sample size
This document is a draft for review purposes only and does not constitute Agency policy.
A-368 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
Reference,
of participant
Exposure
Consideration
Analysis and
setting,
selection and
measure
of likely
completeness
and design
comparability
and range
Outcome measure
confounding
of results
Size
Confidence
2008-June
years old, (n=83,
summer
Proportion with > 1
Comparing
group only.
Other coexposures that
2011
71.5% of those
seasons. Other
episode of wheezing
baseline
Change from year
have been associated with
meeting inclusion
measurements
over last 12 months,
concentrations
1 to year 2 in
asthma symptoms also
criteria) in homes
for N02, VOCs,
ISAAC questionnaire
formaldehyde,
prevalence of
declined in intervention
with low
dust, house
administered to
N02, and dust
asthma
group (toluene,
ventilation rates
dust mites, cat
parents: 43/39;
mites were
symptoms and
ethylbenzene, styrene,
(<0.30 ACH).
and dog
Mean number of days
comparable,
medical care in
limonene, alpha-pinene,
Randomly
allergens,
with asthma symptoms
Toluene and
the past year
airborne mold spores,
assigned to
airborne mold
per 14 day period (> 1
mold spores were
associated with a
although formaldehyde
intervention to
spores
coughing, wheezing,
higher in
50% reduction in
reduction was greatest.
increase
chest tightness,
intervention
formaldehyde
ventilation rates
disturbed sleep or
group.
concentration
by 0.15 ACH
trouble breathing
Comparing year 1
analyzed using
(n=43) and control
Symptoms diary: 37/32;
to year 2,
mixed liner
(n=40).
administered to parents
reductions in
models with
2 weeks per month
formaldehyde,
repeated
from November-
toluene, styrene,
measures
March in 2010 and
limonene, and
2011;
alpha-pinene,
Asthma control over
airborne mold
one month, Asthma
spore
quiz: 31/25;
concentrations
were significantly
different for
intervention
group compared
to control. N02
concentrations
increased.
Allergens in
mattress and rugs
in bedroom did
not change.
(Li et al.,
Infants aged < 4
Air sampling
Baseline information
Potential
Cox regression in
N = 963
Time to onset of wheeze
months attending
(no2,
obtained using
confounders
entire sample;
event
2019)
14 maternal and
formaldehyde)
validated ISAAC
selected from
formaldehyde
(Hong Kong)
child health clinics
using
questionnaire
baseline
modeling as
This document is a draft for review purposes only and does not constitute Agency policy.
A-369 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
Reference,
of participant
Exposure
Consideration
Analysis and
setting,
selection and
measure
of likely
completeness
and design
comparability
and range
Outcome measure
confounding
of results
Size
Confidence
Birth cohort
between
standardized
completed by parents
characteristics
continuous
Overall
September
September 2013
diffusion
prior to age 4 months.
associated with
variable
SB IB Cf Oth
Confidence
2013 to April
to April 2014,
samplers at 6
Weekly respiratory
formaldehyde
¦ y
Low
2014
stratified by family
months of age.
health diary and
concentrations
¦ n
history of asthma,
N0210-14
monthly health
using log-rank
family history of
day sampling
telephone survey
test, p < 0.25.
1 n\M
allergy and no
period.
blinded to exposure
Stepwise
C.onnern for selection hias.
family history.
Formaldehyde
status until 18 months
adjustment, final
Partinination rate was
Included if locally
72 hour
of age. New onset
models adjusted
verv low of elimhle
born ethnic
sampling
wheeze (time to event)
for N02, sex,
agreed^ and of those
Chinese, age < 4
period using
measured from 6 to 18
neonatal
selected there was
months, Birth
ISO 16000-4
months of age. 120
respiratory
notahle data loss
. data
weight > 2.5 kg,
method.
(12.5%) infants had
illness, having a
was romnlete for 67%. No
gestation > 36
Concentrations
new onset wheeze at
sibling, family
romnarisons of
weeks, cared for
not reported.
an average of 13.2
history allergy or
nartirinants and
at home,
months.
asthma, pets, or
nonnartirinants and no
telephone
cooking fuel. No
desrrintive statistics
numbers available,
control for
nrovided for studv
mothers aged > 18
smoking or ETS.
samnle No rontrol for
years, Cantonese
smoking or ETS.
speaking.
Excluded if
congenital
disease, cared for
at child-care
center > 20
hours/week,
moving after
recruitment. Of
14,755 eligible,
4310 agreed to
participate (29%).
After stratification
by family history,
1434 were
recruited and data
were complete for
This document is a draft for review purposes only and does not constitute Agency policy.
A-370 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
963. 471 subjects
had been lost
because of invalid
outcome or air
samples or they
dropped out. No
comparisons of
participants with
nonparticipants.
No descriptive
statistics provided
for study sample.
(Liu et al..
2018a)
(China)
Hospital
based case-
control:
children
September
2016 to
March 2017
Recruited 180
children with an
asthma diagnosis
from hospital and
180 healthy
controls in same
city (Changchun)
during September
2016 to March
2017.
Administered
ISAAC
questionnaire,
validated for
children in Korea.
Asthma severity
assessed with
pulmonary
function tests.
Children excluded
if medical
treatment with
vitamins or
antibiotics within
3 month, severe
Indoor area
samplers
placed 1 -1.5
meters above
ground, doors
and windows
closed 12
hours prior.
HCHO sampled
in living room
and bedroom
with QC-2B
sampler,
Beijing
Municipal
Institute of
Labor
Protection
method.
Citation for
method
provided.
Sampling
period was 2
months.
Asthma diagnosis via
ISAAC responses (2 or
more incidents of
cough, wheezing, and
dyspnea for 3 or more
consecutive days). In
addition, FEVi
increased by >15% after
(3-agonist inhalation
and persistent asthma
was stable for 3 or
more months prior to
study.
History of allergy,
breast feeding,
ETS and indoor
plants were
associated with
asthma status.
Included in model
with PM2.5 and
HCHO. Sex, mean
age, mean BMI
and race were
comparable
between cases
and controls.
Associations with
pollutant
concentration
(quartiles)
analyzed with
multivariate
regression.
180 cases;
180
controls
Current asthma
symptoms
SB IB Cf Oth
Overall
Confidence
Medium
Medium
While reporting details
were brief, citations were
given and appropriate
methods for exposure and
outcome ascertainment
appear to have been used
and the sampling period
for HCHO was adequate.
Coexposures to PM and
N02 were simultaneously
controlled.
This document is a draft for review purposes only and does not constitute Agency policy.
A-371 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
organ failure
(heart, renal and
other serious
disorders).
Median (range)
Mg/m3 HCHO
Asthma 38.35
(12.04-
142.12)
Control 25.11
(12.26-94.34)
N02 and PM
also measured.
Madureira
et al.
(2016)
(Porto,
Portugal)
Children,
case-control,
October 2012
-April 2013
Random
recruitment of 38
residences among
asthmatic children
and 30 residences
among
nonasthmatic
children
previously
identified in a
cross-sectional
study (Madureira
et al., 2015).
Parents
volunteered to
respond to ISAAC
questionnaire for
n=1099 children
(aged 8-10 years,
69% of recruited).
Excluded
respondents with
a recent
renovation or who
had moved since
responding. No
information
comparing
Measurements
of VOC,
aldehydes,
PM2.5,
PM10,
bacteria, fungi,
carbon dioxide
(C02),
temperature
and relative
humidity levels
were
conducted
simultaneously
both indoors
and outdoors.
Sampling and
analysis
methods
described.
Continuous
passive
sampling for
formaldehyde
and other
VOCs and
aldehydes in
bedroom over
For asthma cases,
parents responded yes
to both of 2 questions
in ISAAC questionnaire:
1) Has your child ever
had asthma diagnosed
by a doctor? and 2) In
the past 12 months, has
your child had
wheezing or whistling
in the chest? Parents of
controls responded no
to both questions.
Higher
proportion of
cases were boys.
Comparable for
age, BMI and
parental
education level,
family history of
allergic disorders
and number
of siblings was
slightly higher in
cases. No other
chemical or
biological risk
factors differed
between groups
(except limonene
was higher in
control). Analyses
were not
adjusted for
potential
confounders.
Concentrations
(7-day means)
compared
between groups.
Cases n=38
Controls
n=30
Current Asthma
SB
IB
Cf
Oth
Overall
Confidence
¦
¦
Low
Low
Small sample size,
potential for selection
bias, no adjustment for
confounding and some
differences noted
between cases and
controls
This document is a draft for review purposes only and does not constitute Agency policy.
A-372 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
participants to
nonparticipants.
Potential exists for
selection bias with
greater
environmental
controls among
asthmatic families.
Although extent of
bias impact
unknown, TVOCs,
acetaldehyde and
ventilation rates
higher in control
homes, but not
PM or bacteria
and fungi counts..
7 days.
Formaldehyde
concentrations
all above the
detection limit.
Malaka
and
Kodama
(1990)
(Indonesia)
Wood
workers
(prevalence
survey)
Plywood mill
workers, random
sample of exposed
workers (based on
measurements),
stratified by
smoking, work
duration (<, > 5
years), (random
sampling process
not specified).
Random sample of
nonexposed
(defined based on
area measures
and job history),
matched to
exposed by age,
duration, and
smoking. 93%
Personal and
area samples
(duration not
reported);
above 200
(mean 910, up
to 3480
Mg/m3).
Nonexposed
areas based on
measure-
ments (e.g.,
warehouse,
saw mill)
American Thoracic
Society (1978)
questionnaire. Asthma
defined as "Ever had
attack of wheezing that
made you feel short of
breath?" or ever had
asthma and if so, do
you currently have
asthma? Also included
"occupational asthma"
(not defined). Since
purpose of study was
the impact of
occupational exposure,
asthma definition is
iinterpreted to be
relevant to current
status. [Increased
prevalence of asthma
Adjusted for age,
smoking, dust
Percent by
exposure status,
OR, p-value 95%
CI not reported
(but could be
calculated for
crude OR
estimate)
93
exposed;
93
referents
Asthma
SB IB Cf
Oth
Overall
Confidence
HH
Medium
Selection out of the
exposed work force of
"affecteds" possible in this
type of prevalence study.
"Unexposed" exposure
group exposed to levels of
formaldehyde up to
0.086mg/m3. Either
limitation would result in
reduced (attenuated)
effect estimate.
"Occupational asthma"
not defined and "ever"
asthma may differ from
current prevalence.
This document is a draft for review purposes only and does not constitute Agency policy.
A-373 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
participation rate
and mean
duration about 6
years in both
groups.
associated with lower
FEVi or FEVi/FVC in
these workers].
Matsunag
a et al.
(2008)
(Japan).
Residences:
adults
(Prevalence
survey)
Pregnancy cohort,
enrolled 2nd
trimester.
Recruited through
pregnancy clinics
and obstetrics
departments. 17%
of pregnant
women in the city
participated;
recruitment
extended to other
areas. Low
participation rate.
Internal
comparison group.
24-hour
personal
sample; 60th
percentile 33
mg/m3, 90th
percentile 58
mg/m3
Allergy:
Self-report of medical
treatment (medication
use) for atopic eczema
or allergic rhinitis in
past 12 months.
Exposure measurement
blinded to outcome
classification.
Asthma:
Self-report of medical
treatment (medication
use) for asthma in past
12 months.
Adjusted for age,
gestation, parity,
family history (of
asthma, atopic
eczema, allergic
rhinitis), smoking
status, current
passive smoking
at home and
work, mold in
kitchen, indoor
domestic pets,
dust mite antigen
level, family
income,
education, and
season of data
collection. Also
examined N02
Logistic
regression, OR
(95% CI) by 4
exposure
categories (30th,
60th and 90th
percentiles); also
presented
dichotomized at
90th percentile.
Results also
stratified by
family history of
allergies.
998
21 asthma
cases, 57
eczema,
140 rhinitis
cases
Allergy (atopic eczema,
rhinitis) and asthma
SB
IB
Cf Oth
Oi/erall
Confidence
Medium
w-
Low participation rate but
potential for diffential
participation (by
formaldehyde exposure
and disease status)
unlikely. For allergy, lack
of data pertaining to
sensitivity and specificity
of these questions.
Limited to one-day
exposure sample (but did
address season in
analysis). For asthma,
potential low sensitivity of
outcome the questions,
and in addition, small #
Mi et al.
(2006)
(China)
Schools:
children
(prevalence
survey)
10 schools, 3
classrooms (7th
grade) per school.
Participation rate
99%
4-hour (school
day) air
samples; some
information on
measurement
protocol.
Minimum =
0.003 mg/m3;
(unclear if this
ECRHS definition
Medication use or
asthma attack in past
12 months; additional
questions on lower
respiratory tract
symptoms (in past 12
months, wheeze or
whistling in the chest,
Adjusted for age,
gender, smoking,
observed water
leakage and
indoor moulds.
Also examined
temperature,
relative humidity,
indoor C02,
Logistic
regression, OR
(95% CI) per
0.010 mg/m3
increase.
1,414
Asthma
SB IB Cf Oth
Overall
Confidence
Medium
Uncertainty about
exposure distribution and
analysis (e.g., percent
This document is a draft for review purposes only and does not constitute Agency policy.
A-374 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
November-
December
2011
is 1/2 of LOD?; 1
SD above
mean = 18
Mg/m3;
maximum = 20
Mg/m3.
daytime breathlessness
attack at rest or after
exercise, nighttime
breathlessness attack).
Exposure measurement
blinded to outcome
classification
indoor 03, and
examined
collinearity of
exposures.
35 ug/m3)
and low (< 35
ug/m3) based on
the median.
Asthma-like symptoms,
Allergy-like symptoms
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Medium
Selection of schools was
part of a larger European
framework. Appropriate
methods for exposure
assessment and outcome
ascertainment
instruments appear to
have been used although
endpoint, asthma-like
symptoms, is not specific
to current asthma
definition.
Outcome definition for
allergy-like symptoms
using ISAAC questionnaire
included combined
symptoms of rhinitis
(nose), eye and skin
conditions.
This document is a draft for review purposes only and does not constitute Agency policy.
A-375 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
compared to definition
for current asthma
Neghab et al.
(2011)
(Iran)
Workers:
melamine-
formaldehyde
resin plant
(prevalence
survey)
Exposed:
melamine-
formaldehyde
resin plant
workers. Referent
group: office
workers from
same plant, no
present or past
exposure to
formaldehyde or
other respiratory
irritant chemicals.
Participation rate
100%. Duration
>2 years
Area samples
(40 minutes) in
7 workshops
and 1 area
sample in
office area.
Exposed (mean
±SD) 0.96
(±0.49) mg/3;
unexposed =
nondetectable.
American Thoracic
Society (1978)
questionnaire
(modified): wheezing
symptoms (no details of
questions)
No covariates
considered in the
symptom
analysis. Similar
in demographics
and current
smoking (but
smoking
frequency higher
among exposed)
Fisher's exact
test,
OR (p-value)
n = 70
exposed,
24
unexposed
Asthma
SB IB a
Oth
Overall
Confidence
HH
Low
Uncertainty regarding
asthma definition.
Selection out of the
exposed work force of
"affecteds" possible in
this type of prevalence
study; would result in
reduced (attenuated)
effect estimate.
This document is a draft for review purposes only and does not constitute Agency policy.
A-376 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Norback et al.
(1995)
(Sweden)
Residences:
adults (nested
case-control)
64% participation
rate for cases, 57%
for controls
2-hour
household
sample
(bedroom).
Limited
sampling
period in
closed
residence with
no point
formaldehyde
emissions;
sampling and
analytic
protocols
referenced
(Andersson
et al..
1981. LOQ
OA
mg/m3);
range reported
as <0.005 to
0.110 mg/m3,
thus most
were
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
Reference,
of participant
Exposure
Consideration
Analysis and
setting,
selection and
measure
of likely
completeness
and design
comparability
and range
Outcome measure
confounding
of results
Size
Confidence
14 years).
using diffusion
or nasal congestion.
for other indoor
diffusion
indoor formaldehyde
Participation 96%
samplers.
Cases defined by
chemical
sampling
concentrations
Samplers
reporting symptoms
exposures,
and pumped air
placed 2
weekly over a 3-month
personal factors
sampling),
meters above
period.
and home
personal factors
floor.
environment
(sex, race,
factors.
current
Mean
smoking, atopy,
concentrations
parental
formaldehyde
asthma/allergy)
indoor 4.2
and home
ug/m3, max
environment
18.0 ug/m3,
factors
100%>DL
(ETS,
Outside 5.5
dampness/mold,
ug/m3, max
recent indoor
6.0 ug/m3,
painting). 3-level
100%>DL
logistic regression
models (child,
school,
classroom)
including
significant
exposure
variables from
first model, all
personal factors
and all
environment
factors. No
results reported
for rhinitis and
formaldehyde
because it wasn't
significantly
associated with
This document is a draft for review purposes only and does not constitute Agency policy.
A-378 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
rhinitis in the first
model.
Palczvnski et
al. (1999)
(Poland)
Residences:
adults,
children
(prevalence
survey)
Random sample of
120 households
with children ages
5-1 5 years, built
10 years before
study.
Participation rate
not reported (i.e.,
were more than
120 households
originally
recruited?)
24-hour
household
sample, area
not specified;
up to 0.067
mg/m3(most
<0.050).
Calibration
0.005 to 0.100
mg/m3
Allergy:
5 allergen skin prick test
(dust, dust mites,
feathers, grasses);
serum IgE positive if >
0.35 kU/l RAST.
Asthma:
Bronchial asthma
diagnosis based on
American Thoracic
Society (1978) criteria
(additional details not
reported). Diagnosis
interpreted to be for
current status.
Exposure measurement
blinded to outcome
classification
Environmental
tobacco smoke
Contingency
table analysis,
prevalence (n, %)
by age (adult;
children)
exposure group,
and
environmental
tobacco smoke
exposure.
Highest exposure
group very
sparse.
278 adults,
186
children
Allergy (skin prick tests),
children
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Uncertainty about time
window of exposure
measurement with
respect to skin prick test
results.
Allergy (skin prick tests)
in adults
SB IB a Oth
Overall
Confidence
Low
Uncertainty about time
window of exposure
measurement with
respect to skin prick test
results (greater
uncertainty in adults than
in children)
Asthma, children and
adults
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Uncertainty regarding
asthma definition
All outcomes
This document is a draft for review purposes only and does not constitute Agency policy.
A-379 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Not informative above
0.050 mg/m3 because of
sample size (<5).
Raaschou-
Nielsen et
al. (2010)
(Denmark)
Infants (birth
cohort)
1998-2003
Copenhagen
Prospective Study
on Asthma in
Childhood. 378
out of 411 (92%)
participants at
18-month
follow-up; 343
with
formaldehyde
data.
Three 10-week
bedroom
sampling
periods from
birth to 18
months (aimed
for 6,12, and
18 months).
Median 0.018
mg/m3, 95th
percentile
0.037 mg/m3.
Within
individual
variance 69%
of total
variance
Daily diary kept by
parents on respiratory
symptoms. Training
and definitions
provided. Wheezing =
any symptom severely
affecting the child's
breathing, such as noisy
breathing (wheeze or
whistling sounds),
breathlessness,
shortness of breath, or
persistent, troublesome
cough). Reviewed by
study personnel every
6th month and after a 3-
day period of
respiratory symptoms.
Outcome defined as
"ever had at least one
symptom day";
sensitivity analysis
defined outcome as
three or more
consecutive days with
wheezing symptoms.
Adjusted for sex,
area of residence,
education of
mother, baseline
lung function
Logistic
regression of
"ever had at least
one symptom
day" (88% = yes)
and linear
regression of
number of
symptom days
(excluded 78 with
0 days). Analyzed
by quintile of
exposure
(reference =
<0.012 mg/m3)
343
Lower respiratory tract
symptoms in infants and
toddlers
SB IB a Oth
Overall
Confidence
Low
*
Analysis does not take
into account important
features of the data
(e.g., temporal
variations in symptoms
and large within
individual variability
formaldehyde); could
have masked an
association
Roda et al.
(2011)
(France)
Residences:
infants (birth
cohort)
Infants
(singletons, >2,500
g) from 5
maternity
hospitals in Paris.
N = 3840 out of
4,177(92%)
initially enrolled
Questionnaire
on home
characteristics
at baseline and
updated at 3,
6, 9, and 12
months. N =
196 randomly
Parent questionnaire at
1, 3, 6, 9, and 12
months:
•Upper respiratory
infections
•Lower respiratory
infections
Examined sex,
older sibling,
parental asthma,
history,
socioeconomic
status (4 levels,
based on parents'
occupation),
Exposure
prediction model
for high versus
low (based on
median):
sensitivity 72.4%
2,940
Lower respiratory tract
symptoms in infants and
toddlers
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
This document is a draft for review purposes only and does not constitute Agency policy.
A-380 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
2003-2006
completed 1 or
more
questionnaires;
2,940 had baseline
and 12 month
questionnaire
(70% of initial
enrollees; 76% of
those with 1 or
more
questionnaire)
selected for
predictive
modeling
analysis.
Based on 4 1-
week
measures at 1,
3, 6, and 9
months. LOD
0.008 mg/m3.
Median 0.020
mg/m3; IQR
0.014, 0.027
mg/m3.
Predictors
included
measures of
continuous
formaldehyde
exposure,
intermittent
exposure,
home
characteristics,
and air flow
•Eczema
•wheezing episodes
(frequency)
•At 12 months, also
includes shortness of
breath, dyspnea, dry
cough at night without
cold
Used to define lower
respiratory infections
with and without
wheeze
prenatal and
postnaltal
tobacco smoke
exposure,
dampness, breast
feeding <3
months, day care,
pets in home
specificity 73.6%.
Exposure
prediction model
by tertile:
sensitivity 57.4%
specificity 82.1%.
Outcome
examined as LRI
versus no LRI,
and as 3-level
variable in
multinominal
logistic regression
(LRI-with wheeze;
LRI-no wheeze,
no LRI)
Did not test predictive
model on separate sample
(may overestimate
sensitivity and specificity)
Rumchev
et al.
(2002)
(Australia)
Residences:
children
(case-control)
Related
reference:
Limited to ages 6-
36 months;
recruitment
process not
described for
cases or controls;
cases from
emergency room
and controls (age
matched) from
area health
department,
8-hour
samples,
bedroom and
living room,
two seasons.
Mean 0.030
and 0.28 and
maximum
0.224 and
0.190 mg/m3,
respectively, in
Emergency room
discharge diagnosis of
asthma, ages 6-36
months.
Adjusted or
considered age,
allergies, family
history of
asthma, dust
mites, relative
humidity,
temperature,
atopy,
environmental
tobacco smoke,
pets, air
Generalized
estimating
equation
modeling for
repeated
measures
88 cases,
104
controls
Lower respiratory tract
symptoms in infants and
toddlers
SB IB a
Oth
Overall
Confidence
III
Medium
Recruitment process not
described; uncertainty as
to what is included within
this case definition and
length of time between
This document is a draft for review purposes only and does not constitute Agency policy.
A-381 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Rumchev
et al.
(2004)
representing the
catchment area of
the hospital
bedroom and
living room.
conditioning, use
of gas appliances
emergency room visit and
subsequent exposure
measure.
Smedje
and
Norback
(2001)
(Sweden)
Schools:
children
(nested case-
control
design)
1993-1997
Related
reference:
Smedje et al.
(1997):
however, this
baseline study
of prevalence
of current
asthma used
measures
taken in 1993,
which ranged
from <0.005
to 0.010
mg/m3, with
>50% less
than LOD.
Thus, this
analysis did
Nested case-
control in school-
based cohort
study, 1st, 4th, and
7th grades at
baseline (1993);
follow-up in 1997.
Excluded if history
of allergy at
baseline. 78%
participation in
follow-up. Schools
randomly selected
in Uppsala,
Sweden; 2-5
classrooms
selected from
schools for
exposure
measures.
Participants
compared to
nonparticipants on
baseline
characteristics.
4-hour (school
day) samples,
2-5 rooms per
school (chose
frequently
used rooms),
1993 and
1995; <0.005
to 0.042
mg/m3. Mean
0.008,
geometric
mean 0.004
mg/m3
Allergy:
Parent report of
incident allergy to hay
fever/pollen or pet
dander.
Asthma:
Parent-report of
incident physician
diagnosis (validation
study: specificity >99%,
sensitivity 73%
compared with
physician's
assessment).
Exposure measurement
blinded to outcome
classification
Adjusted for age,
sex, history of
atopy (eczema) at
baseline, changes
in smoking
habits.
Collinearity
among measures
(including VOC,
mold) assessed;
did not attempt
adjustment for
multiple
exposures but
pattern of results
differed among
the exposures
examined.
Logistic
regression, OR
(95% CI) per
0.010 mg/m3
increase
[high proportion
below detection
limit of 0.005
mg/m3, 54% of
1993 samplesand
24% of 1997],
Results similar
when students
who were no
longer in the
school excluded
(about 2/3 left
the school at
mean of 1.5 years
before follow-up)
88 incident
pollen
allergy; 50
incident
pet allergy
cases; 56
incident
asthma
cases out
of 1,258 at
baseline.
Allergy (incidence of
allergies) and asthma
(incidence)
SB IB Cf Oth
m
Overall
Confidence
Low
Exposure measures in
only 2 of the 4 years;
uncertainty about
distribution; relatively
high percentage
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
not meet
EPA's
inclusion
criteria.
Tavernier
et al.
(2006)
(United
Kingdom)
Residences:
children
(case-control)
Related
reference:
Gee et al.
(2005)
Cases from two
primary care
practices, age- and
sex-matched
controls from
same practices.
Ages 4-17 years.
Participation rate
50%. 95 additional
cases excluded
because no
matching control
identified.
[Note: Gee et
al. (2005)
described the age
range as 4-16
years]
7-day sample
in living room
and bedroom.
Did not report
any
information on
exposure
distribution.
[Note: Gee
et al.
(2005)
described this
as a 5 day
sample;
median values
0.037 and
0.049 mg/m3 in
living room
and bedroom,
respectively]
Positive responses to
three questions on
screening
questionnaire: (1)
wheezed in the last 12
months; (2) woken at
night by cough in the
absence of a cold or
respiratory infection in
the last 12 months; (3)
received more than
three courses of
antibiotics for
respiratory symptoms
(both upper and lower
respiratory tract) in the
last 12 months; (4)
history of hay fever or
eczema; (5) family
history of asthma in
first degree relatives.
In validation study,
positive predictive
value 84% for meriting
trial for asthma
medication. Exposure
measurement blinded
to outcome
classification.
[Note: Gee et al.
(2005)described the
positive predictive
Adjusted for
measured
exposures (e.g.,
endotoxin, Der p
1, particulate
matter, N02, and
other risk factors.
Logistic
regression, OR
(95% CI) bytertile
(but exposure
levels by tertile
not reported)
105 cases,
95 controls
Asthma
SB
IB Cf
Oth
Overall
Confidence
1
¦
Low
1
Uncertainty regarding
selection process and loss
of almost half of the cases.
Outcome classification
includes questions that are
not specific to asthma.
Uncertainty as to exposure
range, particularly upper
tertile (no response from
email to corresponding
author).
This document is a draft for review purposes only and does not constitute Agency policy.
A-383 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
value from the
validation study as
79%]
Venn et al.
(2003)
(United
Kingdom)
Residences:
children
(case-control
and symptom
control
among cases)
October-May
1998
Related
reference:
Venn et al.
(2000)
Participants in air
pollution study
1993-1995, 85%
response rate; 835
potential cases
(positive wheeze
question) and 860
potential controls
recontacted in
1998; 54%
responded. From
this, 243 eligible
cases and 383
eligible controls
identified.
Participation rate
79% cases, 59%
controls.
3-day sample
in bedroom in
1998
concurrent
with data
collection on
outcomes;
median 22
pg/m3; 75th
percentile 32
Mg/m3
Asthma:
Parent report of
persistent wheeze
(1995-1996 and 1998);
validation by medical
record review of
prescription for asthma
medication.
Symptom frequency:
One month daily diaries
recording symptoms,
including daytime and
nighttime wheezing,
chest tightness,
breathlessness, and
cough, each measured
on 0 to 5 scale.
Exposure measurement
blinded to outcome
classification
Adjusted for age,
sex, Carstairs
deprivation index
(based on postal
code). Also
examined and
addressed other
variables,
including N02,
moisture, mold,
season
Logistic
regression, OR
(95% CI) by
quartile.
Examined effect
modification of
symptom
frequency by
atopy
190 cases,
214
controls
Asthma
SB
IB
a oth
Overall
Confidence
Medium
¦
Uncertainty about time
window of exposure
measure
Asthma control
SB IB Cf Oth
Overall
Confidence
High
Yeatts et
al. (2012)
(United Arab
Emirates)
Residences
(survey)
October 2009
to May 2010
Nationally
representative
sample of
households,
stratified by
geographic area
and population
density. 628
households,
household
participation rate
75%. Age-
stratified sample
7-day sample
in living room.
71%
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
selected from
households.
household
tobacco smoke
exposure.
separate analysis of
children and adults; only
29% above LOD—
analyzed as above versus
below LOD
Yon et al.
(2019)
(Seongnam
City, Korea)
Cross-
sectional
5th and 6th grade
students were
recruited from 22
randomly selected
classrooms at 11
elementary
schools (n = 620),
aged 10- 12 yr. A
total of 427
children
participated
(68.9%).
Formaldehyde
sampling in
each classroom
using monitors
with pumps
during the 1st
and 2nd half of
the school
year.
Mean 27.17 ±
7.72 Mg/m3; as
high as 60
Mg/m3 in some
classrooms.
Duration and
sampling
methods were
not described.
Current asthma or
rhinitis definition:
presence of
characteristic
symptoms and /or signs
during the previous 12
months using ISAAC
questionnaire, Self
report. Rhinitis severity
categorized using
Allergic Rhinitis and Its
Impact on Asthma
guidelines.
Current asthma n = 10
Rhinitis n = 246
Models for
asthma or rhinitis
adjusted for age
and sex apriori.
Also adjusted for
variables based
on statistical
significance in
model (p < 0.10).
Covariates were
BMI z-score,
height,
prematurity or
low birth weight,
home renovation,
environmental
tobacco smoke,
keeping a pet at
home, and
physician-
diagnosed atopic
dermatitis,
allergic rhinitis,
and parental
asthma
Analysis used
generalized linear
mixed models
with robust
variance
estimates and
post hoc
Bonferroni
correction.
Accounted for
classroom
(random effect)
N =427
Current asthma
SB
IB
Cf Oth
Overall
Confidence
Low
¦
Low
Few children with asthma
contributed to analyses
Rhinitis in last 12 months
and rhinitis severity
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Medium
Reporting deficiencies
raise concern for bias in
exposure measurement,
sampling duration and
methods not described.
(Yu et al..
2017)
(Hong Kong)
Birth cohort
702 of 2423 (29%)
eligible infants
aged < 4 months
attending 29
maternal and child
health centers
between
Air sampling
(N02,
formaldehyde)
using
standardized
diffusion
samplers at 6
Baseline information
obtained using
validated ISAAC
questionnaire
completed by parents
prior to age 4 months.
Weekly respiratory
Potential
confounders
selected from
baseline
characteristics
associated with
formaldehyde
Cox regression in
entire sample;
formaldehyde
modeling as
continuous
variable; effect
modification by
N = 535
New onset wheezing
Infants
SB IB Cf Oth
Overall
Confidence
Low
Low
This document is a draft for review purposes only and does not constitute Agency policy.
A-385 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
November
2009 to April
2011
November 2009 to
April 2011,
stratified by family
history of asthma,
family history of
allergy and no
family history.
Enrollment
numbers based on
power
calculations. A
total of 535 with
complete air
sampling for N02
and HCHO.No
comparisons of
participants with
nonparticipants.
months of age
in bedroom.
Mean (SD)
concentrations
N02 42.4
(30.97) |ig/m3;
HCHO 51.09
(74.94) ng/m3;
no details
regarding
sampling
methods or
duration.
health diary and
monthly health
telephone survey
blinded to exposure
status until 18 months
of age. New onset
wheeze measured from
6 to 18 months of age.
120 (11%) infants had
new onset wheeze at
an average of 11.4
months.
concentrations
using log-rank
test, p < 0.25.
Stepwise
adjustment, final
models adjusted
for N02, neonatal
respiratory
illness, having a
sibling, family
history allergy or
asthma, living
area, pets, or
cooking fuel.
family history
was analyzed.
No details provided for
exposure measurements;
concern for selection bias.
Participation rate was
very low (29% of eligible
agreed) and of those
selected there was
notable data loss, data
was complete for 76%. No
comparisons of
participants and
nonparticipants. No
control for ETS
Zhai et al.
(2013)
(China)
Residences
(survey)
January 2008
to December
2009
Provided criteria
for selection of
homes in defined
area; evaluated
186 homes in
Shenyang, China;
homes were
decorated in last 4
years and
occupied within
the last 3 years.
Participation rate
of households not
reported (i.e.,
were more than
186 households
originally
recruited?)
Cited Code for
indoor
environmental
pollution
control of civil
building
engineering
(GB50325-
2001); samples
in 3 rooms per
house
(bedroom,
living room,
kitchen);
sampling time
not specified
(no response
from email to
Asthma: based on
American Thoracic
Society (1978)
questionnaire
Univariate
analysis;
confounding
unlikely
explanation of
the results in
children
Univariate results
for asthma
outcome
[multivariate
modeling of
"respiratory
symptoms"; not
clear what is
included in this
category)
186 homes
186 adults,
82 children
Asthma
Children
SB IB
Cf
Oth
Overall
Confidence
Medium
¦
¦
Uncertainty regarding
exposure measurement
period. Although
potential confounders
were not considered in
asthma only analysis, the
magnitude of the results
is unlikely to be explained
by confounders.
Adults
This document is a draft for review purposes only and does not constitute Agency policy.
A-386 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting,
and design
Consideration
of participant
selection and
comparability
Exposure
measure
and range
Outcome measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Participants within
houses were
randomly selected
corresponding
author);
N=558 samples
in 186 homes.
Exposure
groups
"polluted"
homes: >0.08
mg/m3, mean
0.09-0.13
mg/m3 in three
rooms;
"nonpolluted"
<0.08 mg/m3,
mean
0.04-0.047
mg/m3. 64% of
the 186
houses, and
24% of the 82
houses with
children were
>0.08 mg/m3
("polluted")
Overall
SB IB Cf
in-h
Confidence
Low
m
1 1
See notes ahove for
children In addition for
adults small number of
positive responses.
This document is a draft for review purposes only and does not constitute Agency policy.
A-387 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Evaluation of Controlled Exposure Studies
2 The evaluation of controlled exposure studies examined four primary elements: the type of
3 exposure (paraformaldehyde preferred over formalin or undefined test articles), use of
4 randomization procedures to allocate exposure, blinding of the participant and of the assessor to
5 exposure, and the details regarding the analysis and presentation of results. The subsequent table
6 in this section provides the more detailed documentation of the evaluation of controlled human
7 exposure studies (see Table A-52); studies are arranged alphabetically within this table.
This document is a draft for review purposes only and does not constitute Agency policy.
A-388 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-52. Evaluation of controlled acute exposure studies among people with asthma
Consideration of
possible bias
(randomized
Consideration
Outcome
exposure order,
of likely
Results
Reference
Exposure assessment
classification
blinding to exposure)
confounding
presentation
Size
Confidence
Casset et al.
Formalin, 30 minutes,
Spirometry; FEVi,
Mild asthma, ages 19-35
Within-person
Individual data
19
Overall
(2006b)
0.032 (background) and
FEF25-75, PEF (protocol
years, no respiratory
values and t-tests
Confidence
0.092 mg/m3, achieved
not mentioned) and
infections for 2 weeks;
High
concentrations
bronchial challenge-
not in relevant allergy
analyzed.
airway reactivity test
season or living with a
Randomized, double
Includes allergy
(PDzoFEViDerpl)
pet if allergic.
blinded, detailed
challenge.
(standard protocol)
Random assignment to
data presentation;
Nose clipped during
Testing pre- and every
order of exposure (3
applies to mouth
exposure (mouth
hour up to 6 hours
weeks between
breathing
breathing)
postexposure.
experiments); double
blinded
Ezrattv et
Formalin, 60 minutes, 0
Spirometry; FVC, FEVi
Intermittent asthma
Within-person
Individual data
12
Overall
al. (2007)
and 0.500 mg/m3,
(ECRHS protocol), and
(dyspnea < twice per
values and
Confidence
achieved
bronchial challenge-
week and night
Wilcoxon sign
High
concentrations
airway reactivity test
symptoms < twice per
rank test
analyzed.
(PD15 FEVigrass)
month with PEF > 80%),
Randomized, double
Includes allergy
(standard protocol)
ages 18-45 years; not in
blinded, detailed
challenge
Testing pre- and every
allergy season.
data presentation
hour up to 6 hours
Random assignment to
postexposure.
order of exposure (2
weeks between
experiments); double
blinded.
This document is a draft for review purposes only and does not constitute Agency policy.
A-389 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure assessment
Outcome
classification
Consideration of
possible bias
(randomized
exposure order,
blinding to exposure)
Consideration
of likely
confounding
Results
presentation
Size
Confidence
Green et al.
(1987)
Paraformaldehyde,
60 minute, clean air and
3 ppm, achieved
concentrations
analyzed.
Spirometry; FVC, FEVi,
SGaw (ATS protocol),
testing pre- and
during exposure
period, =15 minute
intervals.
Asthma (clinical history),
no respiratory infection
for 2 weeks, age 19-35
years.
Random assignment to
order of exposure; two
15-minute exercise
segments in 60-minute
exposure period; single
blinded
Within person
Group means and
SE
16
Overall
Confidence
Medium
Randomized, single
blinded
Harving et
al. (1990)
Related
Reference:
Harving et
al. (1986)
Formalin, 90 minutes,
filtered air (8), 0.120
and 0.850 mg/m3,
achieved
concentrations
analyzed.
Spirometry; FEVi, Raw,
SGaw (protocol not
mentioned), testing
pre- and near end of
exposure period.
Bronchial challenge-
airway reactivity test,
immediately after
exposure
PEF by home peak
flowmeter every 2
hours after exposure
and next morning
Asthma (substantial
bronchial hyperreactivity
to histamine), age 15-36
years.
Random assignment to
exposure order (one per
week); double blinded
Within-person
Group means and
SD
15
Randomized, double
blinded, detailed
analysis
This document is a draft for review purposes only and does not constitute Agency policy.
A-390 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure assessment
Outcome
classification
Consideration of
possible bias
(randomized
exposure order,
blinding to exposure)
Consideration
of likely
confounding
Results
presentation
Size
Confidence
Krakowiak
etal. (1998)
Formalin, 2 hours, 0.5
mg/m3, achieved
concentrations
analyzed.
Spirometry
FEVi (testing 2 hours
pre- and immediately
after, 5 hr, and 24 hr)
PEF (testing at
beginning of
exposure, every hour
for 12 hours, 24 hours
after)
Formaldehyde-exposed
workers with asthma.
Order not randomized (1
week between
experiments); single
blinded
Within person
Group means (bar
graph)
10
Overall
Confidence
Low
Not randomi
single blindir
SD not repor
zed,
g, SE or
ted
Sauder et
al. (1987)
Paraformaldehyde,
3 hours, clean air and 3
ppm, achieved
concentrations
analyzed.
Spirometry; FVC, FEVi,
SGaw(ATS protocol),
testing at 0,15, 30,
60,120,180 minutes
during exposure.
Asthma (clinical history),
no respiratory infection
for 6 weeks, age 26-40
years.
Order not randomized;
clean air followed by
formaldehyde (one
week apart); blinding
not specified
Within person
Grouped means
and paired t-tests
for most
measures,
individual FEVi
data
9
Overall
Confidence
Low
Not randomi
blinding not
specified
zed,
Sheppard
etal. (1984)
Paraformaldehyde,
10 minutes, 0,1, and
3ppm, achieved
concentrations
analyzed.
Spirometry; SGaw,
testing before and 2
minutes after
exposure.
Asthma (clinical history),
age 18-37 years.
Randomization of order
not reported; two
protocols (at rest and
during exercise) >1 day
apart; blinding not
specified
Within person
Grouped means
and SD and paired
t-tests
7
Overall
Confidence
Low
Randomizati
blinding not
specified
an and
This document is a draft for review purposes only and does not constitute Agency policy.
A-391 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure assessment
Outcome
classification
Consideration of
possible bias
(randomized
exposure order,
blinding to exposure)
Consideration
of likely
confounding
Results
presentation
Size
Confidence
Witek et al.
(1987);
Witek et al.
(1986)
Paraformaldehyde
(with 2-propanol?), 40
minutes, 0 and 2ppm
Spirometry; FVC, FEVi,
Raw, testing during and
at 10 and 30 minutes
postexposure; PEFR
assessed from 1 to 24
hours post exposure.
Mild asthma (ATS
definition), age 18-35
years. Random
assignment to order of
exposure; two protocols
(at rest and during
exercise); double
blinded
Within person
Individual data
values and paired
t-test
15
R
b
n
at
P
in
P
Overall
Confidence
High
andomized,
inded;
Dnparametri
lalysis could
"eferred but
dividual dat
"ovided
double
c
be
a
This document is a draft for review purposes only and does not constitute Agency policy.
A-392 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Experimental Animal Studies
2 The experimental animal studies identified as a result of the literature search specific to this
3 section are evaluated as mechanistic information in Appendix A.5.6.
4 A.5.5. Respiratory Tract Pathology
5 Literature Search
6 Studies in Humans
7 A systematic evaluation of the literature database on studies examining the potential for
8 respiratory tract pathology in humans in relation to formaldehyde exposure was initially conducted
9 in September 2012, with regular updates as described elsewhere (including a separate Systematic
10 Evidence Map that updates the literature from 2017-2021 using parallel approaches; see Appendix
11 F). The search strings used in specific databases are shown in Table A-53. Additional search
12 strategies included:
13 • Review of reference lists in the articles identified through the full screening process and
14 • Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde fU.S.
15 EPA. 201 Obi
16 This review focused on histopathological endpoints and signs of pathology in nasal tissues.
17 Inclusion and exclusion criteria used in the screening step are described in Table A-54. The search
18 and screening strategy, including exclusion categories applied and the number of articles excluded
19 within each exclusion category, is summarized in Figure A-30. Based on this process, as of the last
20 literature search update, 12 studies were identified and evaluated for consideration in the
21 Toxicological Review.
Table A-53. Summary of search terms for respiratory tract pathology in
humans
1. Database,
2. Initial Search Date
3. Terms
PubMed
12/18/2012
No date limitation
(Formaldehyde[majr] OR paraformaldehyde[majr] OR formalin[majr]) AND
(Hyperplasia OR metaplasia OR nasal mucosa OR occupational diseases OR respiratory
tract diseases OR rhinitis OR mucociliary) AND (epidemiology OR epidemiological OR
epidemiologic OR cohort OR retrospective studies OR retrospective OR prospective
studies OR prospective OR cross-sectional OR case-control OR cross-sectional study OR
prevalence study OR occupational)
Web of Science
12/19/2012
No date limitation
TS=(Formaldehyde OR paraformaldehyde OR formalin) AND TS=(Hyperplasia OR
metaplasia OR nasal mucosa OR occupational diseases OR respiratory tract diseases
OR rhinitis OR mucociliary) and TS=(epidemiology OR epidemiological OR
epidemiologic OR cohort OR retrospective studies OR retrospective OR prospective
This document is a draft for review purposes only and does not constitute Agency policy.
A-393 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1. Database,
2. Initial Search Date
3. Terms
studies OR prospective OR cross-sectional OR case-control OR cross-sectional study OR
prevalence study OR occupational)
Toxline
05/03/2013
No date limitation
(Formaldehyde OR Paraformaldehyde OR Formalin) AND (Hyperplasia OR metaplasia
OR nasal mucosa OR occupational diseases OR respiratory tract diseases OR rhinitis OR
mucociliary) AND (epidemiology OR epidemiological OR epidemiologic OR ohort OR
retrospective studies OR retrospective OR prospective studies OR prospective OR
cross-sectional OR case-control OR cross-sectional study OR prevalence study OR
occupational)
Table A-54. Inclusion and exclusion criteria for studies of repiratory
pathology in humans
Included
Excluded
Population
• Humans
0.43 Animals
Exposu re
• Indoor exposure via
0.44 Not about formaldehyde
inhalation to
0.45Not inhalation (e.g., dermal exposure)
formaldehyde
• Measurements of
formaldehyde
concentration in air
Comparison
Evaluated outcome
• Case reports
associations with
formaldehyde exposure
• Surveillance analysis/Illness investigation (no
comparison)
Outcome
• Histopathology and
0.46 Other health endpoints
signs of pathology in
0.47 Nasal symptoms (e.g., rhinitis, mucous flow rate)
nasal tissues
0.48 Not a health study
0.49 Exposure studies/no outcomes evaluated
Other
Reviews and reports (not primary research), letters,
meeting abstract, no abstract, methodology paper,
nonessential article in a foreign language (e.g., after
review of title and abstract, if available, or consultation
with native speaker)
This document is a draft for review purposes only and does not constitute Agency policy.
A-394 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Respiratory Tract Pathology (Human) Literature Search
73
1-
o
0)
to
«£
¦g s
V
+-»
3
Q_
E
o
u
(/)
— *00
03 0
C
o 2
"O
nj
"O
Q.
D
2 c
(/) d)
T U
-------
Supplemental Information for Formaldehyde—Inhalation
1 Studies in Animals
2 A systematic evaluation of the literature database on studies examining the potential for
3 respiratory tract pathology in animals in relation to formaldehyde exposure was initially conducted
4 in September 2012, with regular updates as described elsewhere. The search strings used in
5 specific databases are shown in Table A-55. Additional search strategies included:
6 • Review of reference lists in the the articles identified through the full screening process,
7 • Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
8 EPA. 201 Obi, and
9 • Review of references in 6 review articles relating to formaldehyde and respiratory
10 pathology in animals, published in English, identified in the initial database search.
11 Inclusion and exclusion criteria used in the screening step are described in Table A-56.
12 After manual review and removal of duplication citations, the 1,631 articles were initially screened
13 within an EndNote library; title was considered first, and then abstract in this process. Full text
14 review was conducted on 105 identified articles. The search and screening strategy, including
15 exclusion categories applied and the number of articles excluded within each exclusion category, is
16 summarized in Figure A-31. Based on this process, 41 studies were identified and evaluated for
17 consideration in the respiratory tract pathology section of the Toxicological Review. An additional
18 35 studies related to MOA for pathology were considered in the overarching mechanistic evaluation
19 (see Appendix A.5.6).
Table A-55. Summary of search terms for respiratory tract pathology in
animals
Database,
initial search date
Terms
PubMed
10/18/2012
Search up through
9/30/2012
Formaldehyde* AND (animals OR dog OR dogs OR canine OR canines OR beagle OR beagles
OR "guinea pig" OR "guinea pigs" OR Cavia OR hamster OR hamsters OR Cricetinae OR
Mesocricetus OR mice OR mouse OR Mus OR monkey OR monkeys OR Macaca OR primate
OR primates OR rabbit OR rabbits OR hare OR hares OR rat OR rats OR Rattus OR Rana or
rodent OR rodents OR Rodentia) AND (alveol* OR bronchial OR bronchi OR buccal OR
laryngeal OR larynx OR lung OR mouth OR nasal OR nasopharyngeal OR nasopharynx OR
nose OR pharyngeal OR pharynx OR pulmonary OR respiratory OR sinonasal OR sinus OR
trachea*) AND (edema OR oedema OR cancer OR carcinogens OR carcinogenesis OR
carcinogenicity OR carcinoma OR "cell proliferation" OR cilia OR dysplas* OR epithelial OR
epithelium OR goblet OR histopath* OR hyperplas* OR hypertrophy* OR metaplas* OR
mucociliary OR mucos* OR mucous OR mucus OR necrosis OR neopla* OR olfactory OR
patholog* OR rhinitis OR squamous OR transitional OR tumor OR tumour OR turbinate OR
ulceration) NOT human
Web of Science
10/18/2012
Search up through
9/30/2012
Topic=Formaldehyde* AND (animals OR dog OR dogs OR canine OR canines OR beagle OR
beagles OR "guinea pig" OR "guinea pigs" OR Cavia OR hamster OR hamsters OR Cricetinae
OR Mesocricetus OR mice OR mouse OR Mus OR monkey OR monkeys OR Macaca OR
primate OR primates OR rabbit OR rabbits OR hare OR hares OR rat OR rats OR Rattus OR
This document is a draft for review purposes only and does not constitute Agency policy.
A-396 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Database,
initial search date
Terms
Rana or rodent OR rodents OR Rodentia) AND (alveol* OR bronchial OR bronchi OR buccal
OR laryngeal OR larynx OR lung OR mouth OR nasal OR nasopharyngeal OR nasopharynx OR
nose OR pharyngeal OR pharynx OR pulmonary OR respiratory OR sinonasal OR sinus OR
trachea*) AND (edema OR oedema OR cancer OR carcinogens OR carcinogenesis OR
carcinogenicity OR carcinoma OR "cell proliferation" OR cilia OR dysplas* OR epithelial OR
epithelium OR goblet OR histopath* OR hyperplas* OR hypertrophy* OR metaplas* OR
mucociliary OR mucos* OR mucous OR mucus OR necrosis OR neopla* OR olfactory OR
patholog* OR rhinitis OR squamous OR transitional OR tumor OR tumour OR turbinate OR
ulceration) NOT human
Toxline
10/21/2012
Search up through
9/30/2012
formaldehyde AND (animal OR "nasal cavity" OR nose OR "respiratory tract" OR "cell
proliferation" OR mucociliary OR histopathology OR pathology OR cancer OR tumor) NOT
(human OR humans OR epidemiology OR epidemiological OR occupation* OR work* OR
antinocicepti* OR nocicepti* OR pain OR sensory OR "formalin test" OR bacteria OR
bacterial)
(including synonyms and CAS numbers, but excluding PubMed records)
Table A-56. Inclusion and exclusion criteria for studies of repiratory
pathology in animals
Included
Excluded
Population
0.50 Animals
0.51 Irrelevant species/ matrix, or human studies
0.52
Exposu re
0.53 Inhalation exposure,
formaldehyde or test
article generating
formaldehyde
0.54 Not formaldehyde (or formaldehyde exposure not
quantified: full text screening only)
0.55 Dermal or oral exposure or other noninhalation
exposure
0.56 Endogenous properties
Comparison
Outcome
0.57 Respiratory tract
pathology
0.58 MOA for pathology
(note: these are evaluated
and discussed in the
overarching MOA section;
see A.1.6)
0.59 Assessment of formaldehyde exposure
0.60 Chemical properties
0.61 Formaldehyde use in methodology or treatement
0.62 Not related to respiratory tract pathology
Other
Reviews and reports (not primary research), letters, meeting
abstract, policy/ current practice paper, duplicate,
nonessential article in a foreign language
This document is a draft for review purposes only and does not constitute Agency policy.
A-397 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Respiratory Tract Pathology (Animal) Literature Search
OJ
C.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
Studies in Humans
Each study was evaluated for precision and accuracy of exposure assessment, measurement
of outcome, participant selection and comparability, possibility of confounding, analysis and
completeness of results, and study size (see Table A-57). The accompanying tables in this section
document the evaluation. Studies are arranged alphabetically within each table.
For studies that evaluated histopathological lesions in nasal biopsies, EPA looked for either
a detailed explanation of how tissues were evaluated and scored, or a citation for a standard
method. Cross-sectional studies among occupational cohorts likely were influenced by the
selection of the workforce toward individuals less responsive to the irritant properties of
formaldehyde, with a reduction in sensitivity. These studies were downgraded because of this
limitation. Treatment of potential confounding by studies also was evaluated. EPA considered age,
gender and smoking to be important confounders to evaluate for effects on pathological endpoints.
EPA also looked for consideration of confounding by other co-exposures in the workplace
depending on the occupational setting.
Table A-57. Criteria for categorizing study confidence in epidemiology studies
of respiratory pathology
Confidence
Exposure
Study design and analysis
High
Work settings: Ability to differentiate
between exposed and unexposed, or
between low and high exposure.
Selection of workers at beginning of exposures (no lead time
bias). Instrument for data collection described or reference
provided and outcome measurement conducted without
knowledge of exposure status. Analytic approach evaluating
dose-response relationship using analytic procedures that are
suitable for the type of data, and quantitative results
provided. Confounding considered and addressed in design or
analysis; large sample size (number of cases).
Medium
Work settings: Referent group may be
exposed to formaldehyde or to other
exposures affecting respiratory
conditions (potentially leading to
attenuated risk estimates).
Lead time bias may be a limitation for occupational studies.
Instrument for data collection described or reference provided
and outcome measurement conducted without knowledge of
exposure status. Analytic approach more limited;
confounding considered and addressed in design or analysis
but some questions regarding degree of correlation between
formaldehyde and other exposures may remain. Sample size
may be a limitation.
Low
Work settings: Short sampling duration
(<1 work shift) without description of
protocol. Missing values or values
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-58. Respiratory pathology
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
School Settings
(Norback et al..
2000) (cross-
sectional study)
Exposure
Interview for
Primary school
Multiple linear
Multiple linear
N = 234
measurements in 2
symptoms, nasal
personnel at 12
regression
regression
individuals,
randomly selected
lavage and acoustic
of 18 randomly
models adjusted
models;
but unit of
classrooms at each
rhinometry; use of
selected schools
for age, sex,
reported
analysis was
school on 2
both subjective
(out of 62) and
smoking, atopy,
regression
school
occasions;
and objective
with restriction to
and mean
coefficients and
means,
Measurements of
measures enabled
schools with
classroom
whether
N = 12
respirable dust, C02,
evaluation of
classes 1-6 and
temperature;
statistically
temperature,
information bias
no changes in
Co-exposure:
significant (p
humidity,
ventilation or
Nasal patency
<0.05);
formaldehyde (4-
redecoration
measures were
uncertainties in
hour sample),
during study
inversely
analysis: use of
airborne
period (March
associated with
school-based
microorganisms,
1993-March
dust, N02, and
mean
viable molds and
1995). 234
Aspergillus.
concentration as
bacteria, N02 (only
current
Elevations in
unit of analysis
in 1993); all staff
employees (84%)
nasal lavage
assigned school
working 20 hr/wk
biomarkers
mean concentration.
or more.
associated with
Formaldehyde
Excluded those
N02, Aspergillus,
concentration: mean
on sick leave or
and yeast;
0.0095 mg/m3; min-
otherwise off
correlation
max of means,
duty. High
between indoor
0.003-0.016 mg/m3;
participation
levels of
provided citation for
reduced
pollutants or
analysis; LOD 0.005
likelihood of
microbials not
mg/m3 (Smedje et
selection bias.
reported;
al., 1997)
correlated with
ventilation? No
SB IE
a
Oth
Overall
Confidence
¦
Low
Unknown correlation
between co-exposures
(dust, N02, and Aspergillus)
which also were inversely
associated with nasal
patency and biomarkers,
potential confounding;
some schools with mean <
LOD; less robust analytic
approach given unit of
analysis
This document is a draft for review purposes only and does not constitute Agency policy.
A-400 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
indoor sources
of combustion—
N02 levels higher
in schools near
traffic
Occupational Settings
Ballarin et al.
(1992)Prevalence
study
Personal sampling;
8-hr TWA (NIOSH,
1977)
Warehouse (N = 3),
0.39 ± 0.20 mg/m3,
range 0.21-0.6
mg/m3
Shearing-press (N =
8), 0.1 ±0.02 mg/m3,
range 0.08-0.14
mg/m3
Sawmill (N = 1), 0.09
mg/m3
Inspirable wood
dust: 0.11-0.69
mg/m3, 0.73 in
sawmill
Cytopathology
analysis of nasal
respiratory mucosa
cells by two trained
readers blinded to
exposure status;
scoring and
classification
analogous to
Torjussen et al.
(1979) and Edling
et al. (1988); most
severe score
present assigned.
Participant
selection and
recruitment not
described.
Nonsmokers in
plywood factory
(N = 15)
compared to
nonsmoking
university or
hospital clerks (N
= 15) matched by
age and sex.
Excluded heavy
drinkers. Use of
referent group
with different
occupations
results in less
similar
comparison
groups
Addressed
potential
confounding by
age and sex
through
matching and
smoking and
heavy alcohol
use by exclusion.
Mean
histological
scores in
exposed and
referent
compared using
Mann-Whitney
U test and
frequency by
classification
using chi-square
test
15 exposed/
unexposed
pairs
Overall
SB IB
::t
I >rh
Confidence
1 1
Medium
1
1 1
Inclusion only of current
workers raises possibility of
healthy worker survival
effect due to irritation
effects
Berke, 1987
Cross-sectional
study
Exposure
measurements since
the mid 1970s using
personal monitoring
(monitoring protocol
Clinical exam and
nasal cytology by
pathologist blind to
exposure or clinical
status. System for
Participant
selection and
recruitment not
described. 52
volunteers from
Mean age in
exposed higher
than employee
referent group,
comparable to
Exposed (Groups
1 and 2)
compared to
referent (Groups
3 and 4); chi-
42 exposed,
10
employee
referents,
28 white-
SB IB Cf Oth
==¦
Methods were not well
Overall
Confidence
Not
informative
This document is a draft for review purposes only and does not constitute Agency policy.
A-401 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
not described).
Group 1 ranging
from 0.02-1.3 ppm.
Group 2 plant
0.05-2.0 ppm
classifying atypical
and typical
metaplasia not
defined.
three paper
plants (currently
employed,
participation 95%
of available
exposed)42
exposed, 10
referent workers.
28 additional
referent white-
collar employees
(36% atypical
squamous
metaplasia in this
group)—not
representative?
additional white-
collar referent
group. Smoking
prevalence 60%
in Groups 1, 2,
and 3; 20% in
white-collar
referent.
Statistical
analysis
excluding
smokers
square test with
adjustment for
age and
smoking;
analysis of
combined
groups not
appropriate
(exposures
different and
very different
demographic
characteristics)
collar
referents
described. Comparisons of
dissimilar groups.
Nonstandard outcome
definition and analyses that
cannot be interpreted.
Inclusion of only current
workers and long duration
of employment (mean >15
years) raises possibility of
healthy worker survival
effect
(Boysen et al..
1990)
Cross-sectional,
study
Formaldehyde
monitoring
conducted after
1980. Before 1980,
exposure assigned
by plant health
officer with
knowledge of the
production process,
recent
measurements, and
worker sensations.
Range of
formaldehyde 0.5
ppm to >2 ppm
(0.62-2.5 mg/m3);
no measurements in
Slides evaluated by
two authors
blinded to clinical
or occupational
status. Histology:
Scoring and
classification of
histologic samples
per variation of
Tojussen (1979)
protocol.
Rhinoscopy:
Scoring according
to Boysen et al.
(1982)
37/74 volunteers
from a chemical
company
producing
formaldehyde
(50% of exposed
workforce).
Referents: 37 age
matched subjects
without overt
nasal disease
(office staff,
hospital
laboratory
personnel, and
EN&T
outpatients). Use
Exposed and
referent
comparable for
age, smoking, or
previous nasal
disease.
Comparison of
histological
results between
exposed and
referent groups
using Wilcoxon
rank sum test,
evaluated
associations
with age,
smoking,
intensity and
duration of
exposure;
comparison of
rhinoscopical
37 exposed,
37 referents
SB
IB Cf Oth
Overall
Confidence
Medium
N
Inclusion only of current
workers and long duration
of employment raises
possibility of healthy
worker survival effect due
to irritation effects
This document is a draft for review purposes only and does not constitute Agency policy.
A-402 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
referent; however,
exposure contrast
likely adequate.
of referent group
with different
occupations
results in less
similar
comparison
groups
results using chi-
square test
This document is a draft for review purposes only and does not constitute Agency policy.
A-403 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
Edling et al.
(1988.1987a)
Prevalence Study
Related studies:
(Odkvist et al..
1985)
Past TWA
formaldehyde
measurements by
plant Industrial
hygienists
sporadically
between 1975 and
1983. Levels of FA in
air ranged from
0.1-1.1 mg/m3, with
peaks up to 5
mg/m3. No
measurements
available before
1975, but estimated
levels higher during
the 1960s and early
1970s. No
measurements in
referent; however,
exposure contrast
likely adequate.
Rhinoscopy: Nasal
mucosa histological
grading by
pathologist blinded
to exposure using
Torjussen et al.
(1979) grading
system
75 of 104
exposed male
factory workers
from 3 plants
(72% of eligible).
Referents: 25
men with similar
age and no
known industrial
exposures to
formaldehyde;
source of referent
group not
described.
Evaluated
characteristics of
nonparticipants
at 1 plant, age
and exposure
time similar, %
with symptoms
higher in
nonparticipants,
% smokers lower
Exposed mean
age: 38 years;
35% smokers.
Referent mean
age: 35 years,
48% smokers.
Histological
score was higher
among exposed
smokers
compared to ex-
smokers and
nonsmokers;
possible
confounder
Exposed groups
compared to
referent group
using Wilcoxon
rank sum test,
no adjustment
for age or
smoking
75 exposed,
25 referents
SB IB Cf Oth
KM
Overall
Confidence
Medium
4,
Inclusion of only current
workers and long duration
of employment (mean 10.5
years) and high prevalence
of symptoms raises
possibility of healthy
worker survival effect due
to irritation effects
(Holmstrom et
al.. 1989c:
Holmstrom and
Wilhelmsson,
1988)
Cross-sectional
study
Personal sampling in
breathing zone for
1-2 hours in 1985.
Chemical Plant:
0.05-0.5 mg/m3,
mean 0.26 [SD0.17
mg/m3]. Furniture
Factory: 0.2-0.3
mg/m3, mean 0.25
Nasal symptoms
questionnaire,
nasal volume flow
rate using
rhinomanometry;
mucociliary
clearance using
green dye to
measure time for
Participant
selection and
recruitment
protocol not
reported;
excluded subjects
with upper
airway infections;
nasal specimens
Formaldehyde
exposed were
slightly younger
than
formaldehyde-
dust exposed or
referent;
smoking status
higher in
Compared
exposure groups
using 2-tailed
t-test for
symptoms, nasal
flow rate, and
histology, and
chi-square test
N = 62 of 70
Group 1, N =
89 of 100
Group 2, N =
32 of 36
Referent
SB
IB
a
Oth
Overall
Confidence
1
Medium
4-
Inclusion of only current
workers and long duration
of employment raises
possibility of healthy
This document is a draft for review purposes only and does not constitute Agency policy.
A-404 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
Exposure
of participant
Consideration
Analysis and
Size/
measures and
Outcome
selection and
of likely
completeness
estimated
Reference
range
classification
comparability
confounding
of results
power
Comments
[SD 0.05 mg/m3].
spot to reach
in 62 of 70
exposed; higher
for mucociliary
worker survival effect due
Referent 0.09 mg/m3
rhinopharynx.
formaldehyde
% male in
clearance
to irritation effects
formaldehyde. Total
Histological
exposed, 89 of
exposed groups.
dust and respirable
changes in nasal
100
Duration of
dust also measured.
mucosa graded by
formaldehyde/
exposure and
a pathologist blind
wood dust
smoking status
to exposure
exposed, and 32
were not
according to
of 36 referents.
correlated with
Torjussen (1979)
Apparent high
histology score,
participation and
therefore
outcome
confounding not
assessment
a concern
blinded to
exposure status
reduced
likelihood of
selection bias.
Use of referent
group with
different
occupations
results in less
similar
comparison
groups
This document is a draft for review purposes only and does not constitute Agency policy.
A-405 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Consideration
Exposure
of participant
Consideration
Analysis and
Size/
measures and
Outcome
selection and
of likely
completeness
estimated
range
classification
comparability
confounding
of results
power
Personal sampling
Nasal symptoms
43 exposed
Evaluated
Logistic
69
over a single 8-hour
and signs;
employees at 3
impacts of
regression,
unexposed,
shift. Formaldehyde
questionnaire;
brass foundries
confounding by
single-pollutant
30 low and
concentration, mean
examination by
producing cores
exclusion of
analyses, OR
12 high
(SD), range: 0.051
rhinologist blind to
using Hot Box
smokers,
(95% CI); cut-
exposure
(0.049) mg/m3,
exposure status
method (90%)
females, or
point for
0.013-0.190 mg/m3;
Referent: 82
asthmatic and
categories of
71.4% of exposed
assembly workers
allergic subjects
formaldehyde
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Exposure
measures and
range
Outcome
classification
Consideration
of participant
selection and
comparability
Consideration
of likely
confounding
Analysis and
completeness
of results
Size/
estimated
power
Comments
Controlled Human Exposure Studies
(Falk et al.,
1994)
Formalin exposure;
analytic
concentrations,
mean: Group 1:
0.021, 0.028, 0.073,
0.174;
Group 2: 0.023,
0.029, 0.067, 0.127
Nasal mucosa
swelling measured
using
rhinostereometry
(summary of
changes for both
turbinates)
Double blind
exposures,
exposure-order
stochastically
distributed and
separated by 2
days.
Within-person
comparison
Results
presented in
graphs
N = 6-7 per
group
Overall
Confidence
Medium
(Pazdrak et al.,
1993)
Test article
characterization and
exposure generation
method not
described;
clean air followed by
0.5 mg/m3
formaldehyde
Stage 1 evaluation
of symptoms,
morphological
changes, and
biochemical
changes in nasal
washings. Stage 2
clinical comparison
of nasal mucosa by
group.
Two-stage, single-
blind examination
with nonrandom
order of exposure
assignment.
Within-person
comparison
Results
presented with
statistical
analyses
N = 8-11
per group
Overall
Confidence
Low
Andersen and
Lundqvist from
Andersen and
Molhave, 1970
Paraformaldehyde.
Dynamic chamber;
analytic
concentrations;
clean air followed by
0.3, 0.5,1.0, and 2.0
mg/m3
formaldehyde.
Nasal airflow
resistance and
nasal mucocilliary
flow
Subjects assigned
to four groups,
each group with
four different
exposures over
four consecutive
days, order
decided by Latin
square design.
Within-person
comparison
Results
presented with
statistical
analyses
N = 16
Overall
Confidence
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-407 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
Studies in Animals
In addition to the general factors considered for all toxicology studies of formaldehyde
inhalation exposure (see Appendix A.5.1), factors specific to the interpretation of respiratory tract
pathology were considered when determining study confidence. These criteria reflect the large
database of well-conducted studies, and include: the use of too few test subjects (i.e., a sample size
of less than 10 was considered a significant limitation); a failure to report lesion incidence and/or
severity; the lumping of multiple lesions (e.g., squamous metaplasia and hyperplasia) together; a
failure to report quantitative incidences and/or statistical analyses; the use of insensitive sampling
procedures (multiple sections across multiple levels of the respiratory tract were preferred); and
use of an exposure duration or follow-up that is likely insensitive for detecting slow-developing
lesions (a duration of >1 year was preferred). Finally, somewhat in contrast to the available
experimental animal studies for other health effect sections, most studies of respiratory pathology
used paraformaldehyde or freshly prepared formalin as the test article, although some studies
tested commercial formalin. While co-exposure to methanol is a major confounding factor for
systemic endpoints, it is less of a concern ("+"; see below) when identifying effects of inhaled
formaldehyde on respiratory pathology. Most inhaled methanol bypasses the nose but is readily
absorbed in the lungs and distributed systemically. A discussion of the different test articles (i.e.,
paraformaldehyde, formalin, etc.) used for formaldehyde inhalation studies can be found in
Appendix A.5.1. Additional considerations that might influence the interpretation of the usefulness
of the studies during the hazard synthesis are noted, including limitations such as the use of only
one test concentration or concentrations that are all too high or too low to provide a spectrum of
the possible effects, as well as study strengths like very large sample sizes or use of good laboratory
practices (GLP); however, this information typically did not affect the study evaluation decisions.
Studies are grouped by exposure duration, and then organized alphabetically by first
author. If the conduct of the experimental feature is considered to pose a substantial limitation that
is likely to influence the study results, the cell is shaded gray; a "+" is used if potential issues were
identified but not expected to have a substantial influence on the interpretation of the experimental
results; and a "++" denotes experimental features without limitations that are expected to influence
the study results. Specific study details (or lack thereof) that highlight a limitation or uncertainty in
answering each of the experimental feature criteria are noted in the table cells. For those
experimental features identified as having a substantial limitation likely to influence the study
results, the relevant study details are bolded.
This document is a draft for review purposes only and does not constitute Agency policy.
A-408 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-59. Evaluation of controlled inhalation exposure studies examining respiratory pathology in animals
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utilitv for Hazard IDe
Criteria relevant
to evaluating the
experimental
details within
each
experimental
feature category
Exposure quality
evaluations (see B.4.1.2)
are summarized (++ =
"robust"; + =
"adequate"; gray box =
poor); relevance of the
tested exposure levels is
discussed in the hazard
synthesis
Sample size
provides
reasonable power
to assess
endpoint(s) in
question; species,
strain, sex, and age
relevant to
endpoint; no overt
systemic toxicity
noted or expected
Interpreting the
appropriateness,
reproducibility, and
informativeness of the
study design for
evaluating respiratory
tract pathology. Although
no studies designed
according to inhalation
guidelines were identified,
several GLP-compliant
studies were identified
and are highlighted below
The protocols used to
assess respiratory tract
pathology are sensitive,
complete, discriminating
(specific), and
biologically sound
(reliable); experimenter
bias minimized
Statistical
methods, group
comparisons, &
data/variability
presentation are
appropriate &
discerning
Expert judgement
based on conclusions
from evaluation of
the 5 experimental
feature categories
Respiratory Pathology—Chronic
(Appelman et
al.. 1988)
Rat
++
+
Small N (N=10)
++
+
Lesion severity provided
for 13-week but not 52-
week sacrifice
++
Medium
[small N; limited
reporting of lesion
severity]
(Dalbev. 1982)
Hamster
++
++
++
Note: single concentration
study
+
Lesion severities NR
++
Medium
[failure to report
lesion severities]
This document is a draft for review purposes only and does not constitute Agency policy.
A-409 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Holmstrom
et al., 1989c)
++
Note: high concentration
exposure (15.3 mg/m3-
day)
+
Small N
(N=16/group)
++
Note: single concentration
study
Lesion severities NR;
nonstandardized
histological
characterization makes
interpretation of effect
difficult
Incidence of
metaplasia and
dysplasia
reported together
Not Informative
[small N; failure to
report lesion
severities; incidence
of metaplasia and
dysplasia reported
together]
Rat
(Kamata et al..
+
Formalin; methanol
concentration was
reported and a
methanol control was
used.
+
Inadequate
number of animals
for interim
sacrifices (N=5)
++
+
Lesion severities NR;
prevalence of neoplastic
lesions complicates
assessment of
nonneoplastic lesions
++
Medium
[formalin; small N for
interim sacrifices;
failure to report
lesion severities]
1997)
Rat
(Kerns et al.,
1983)
++
+
Survival to 18
months was <33%
in all groups
(N>25)
++
Note: data from this study
based on a GLP study (CUT
1982)
Lesion severities NR;
incidence NR; only
three nasal sections (II,
III, and V) evaluated
++
Medium
[somewhat limited
sampling, high
mortality, and failure
to report lesion
incidence and
severities]
Mouse
See also
(Battelle,
1982) and
(Swenberg et
al.. 1980b)
This document is a draft for review purposes only and does not constitute Agency policy.
A-410 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Kerns et al.,
1983)
Rat
See also
(Battelle,
1982) and
(Swenberg et
al.. 1980b)
++
+
Transient viral
infection at weeks
52-53 was
considered
unlikely to
influence study
outcome because
of its short course
++
Note: data from this study
based on a GLP study (CUT
1982)
++
Note: incidence and
severity data by nasal
section extracted from
CUT(1982)
++
High
[Note: transient viral
infection]
(Monticello et
al.. 1996)
Rat
++
++
++
Lesion severities NR;
lesion incidence NR
Insufficient data
to verify
magnitude of
concentration-
response
Low
[Failure to report
lesion incidence and
severities; insufficient
data to verity
magnitude of
concentration-
response]
(Sellakumar et
al.. 1985)
Rat
see also (Albert
et al., 1982)
+
Formaldehyde was
generated by heating a
slurry of
paraformaldehyde in
paraffin oil (kerosene),
which could cause co-
exposure to paraffin oil.
[Note: high
concentration exposure
(18.2 mg/m3-day)]
++
++
Note: single concentration
study
+
Lesion severities NR
++
Medium
[Likely co-exposure to
paraffin oil
(kerosene); testing at
a single high
concentration; failure
to report lesion
severities]
This document is a draft for review purposes only and does not constitute Agency policy.
A-411 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Woutersen et
++
++
++
+
Lesion severities NR;
significant incidence of
lesions in controls
++
Statistical analyses
of lesions NR
High
[Failure to report
lesion severities]
al.. 1989)
Rat
Respiratory Pathology—Subchronic
(Andersen et
al.. 2010)
++
+
small N (N=8)
++
++
+
Data for levels III-
V NR; statistical
analyses of lesions
NR
Medium
[Small N; data for
levels lll-V NR]
Rat
(Arican et al.,
2009)
Analytical method and
concentrations NR
++
++
Note: single concentration
study
Lesion severities NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Failure to report
analytical method
and analytical
concentrations;
failure to report
lesion incidence and
severities; results
described
qualitatively]
Rat
(Casanova et
al.. 1994)
++
Small N (N=3)
++
Lesion severities NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Small N; failure to
report lesion
incidence and
severities; results
described
qualitatively]
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-412 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Coon et al.,
1970)
++
Small N (N=2)
Continuous exposure (22
hours/day)
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Small N; single
concentration tested;
failure to report
lesion incidence and
severities; results
described
qualitatively]
Dog
(Coon et al.,
1970)
++
++
Continuous exposure (22
hours/day)
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Single concentration
tested; failure to
report lesion
incidence and
severities; results
described
qualitatively]
Guinea pig
(Coon et al.,
1970)
++
Small N (N=3)
Continuous exposure (22
hours/day)
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Small N; single
concentration tested;
failure to report
lesion incidence and
severities; results
described
qualitatively]
Monkey
This document is a draft for review purposes only and does not constitute Agency policy.
A-413 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Coon et al.,
1970)
++
Small N (N=3)
Continuous exposure (22
hours/day)
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Qualitative
descriptions only
Not Informative
[Small N; single
concentration tested;
failure to report
lesion incidence and
severities; results
described
qualitatively]
Rabbit
(Coon et al.,
1970)
++
++
Continuous exposure (22
hours/day)
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Qualitative
descriptions only
Not informative
[Single concentration
tested; failure to
report lesion
incidence and
severities; results
described
qualitatively]
Rat
(Feron et al.,
1988)
++
Note: exposure in the
high concentration
group was excessive
(24.4 mg/m3-day)
++
++
+
No quantitative interim
sacrifice data to inform
lesions immediately
after exposure
++
Note: recovery
period data
informs
persistence of
lesions
High
[Note: only tested
high formaldehyde
levels]
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-414 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Horton et al.,
1963a)
+
Analytical
concentrations NR
Note: extremely high
concentration exposure
(200 mg/m3-day)
++
+
Early mortality in high
exposure group by 11th
day of exposure
Nose was not examined;
lesion severity NR
Note: lesions are of
questionable adversity
++
Low
[Analytical
concentrations NR;
early mortality in the
high concentration
group, which had an
extremely high
concentration; nose
was not examined;
failure to report
lesion severity]
Mouse
(Maronpot et
al.. 1986)
+
Formalin; methanol
concentration was not
reported and a
methanol control was
not used. [Note: high
concentration exposure
(49.2 mg/m3)]
+
Small N (N=10)
++
++
++
Medium
[Formalin; small N]
Mouse
(Rusch et al.,
1983)
++
Note: concentrations
tested were very low
(0.23-3.6 mg/m3-day),
and unlikely to elicit a
response
++
++
+
Lesion severity NR
incidence of
squamous
metaplasia and
hyperplasia
reported
together;
data reported for
only one nasal
section
Medium
[Failure to report
lesion severity;
incidence of
squamous metaplasia
and hyperplasia
reported together;
data reported for
only one nasal
section]
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-415 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Rusch et al.,
1983)
++
Note: concentrations
tested were very low
(0.23-3.6 mg/m3-day),
and unlikely to elicit a
response
++
++
+
Lesion severity NR
Incidence of
squamous
metaplasia and
hyperplasia
reported
together; data
reported for only
one nasal section
Medium
[Failure to report
lesion severities;
incidence of
squamous metaplasia
and hyperplasia
reported together;
data reported for
only one nasal
section]
Monkey
(Rusch et al.,
1983)
++
Note: concentrations
tested were very low
(0.23-3.6 mg/m3-day),
and unlikely to elicit a
response
++
+
Limited study design: only
endpoint evaluated was
squamous metaplasia
++
Specific incidence
data NR, so lack
of effect could not
be verified
Medium
[Specific incidence
data NR; note: only
squamous metaplasia
was evaluated]
Hamster
(Wilmer et al.,
1989)
+
Analytical
concentrations NR
++
++
+
Lesion severity NR
++
Medium
[Analytical
concentrations NR;
failure to report
lesion severities]
Rat
(Woutersen et
++
Note: high concentration
exposure (24.4 mg/m3-
day)
++
++
++
++
High
[Note: the high
concentration level
was excessive]
al.. 1987)
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-416 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Zwart et al.,
1988)
++
++
++
+
Lesion severity NR;
lesion incidence
incompletely reported
++
Medium
[Failure to completely
report lesion
incidence; severity
NR]
Rat
Respiratory Pathology—Short-term
(Andersen et
al.. 2008)
+
-30% variations in
chamber concentrations
+
Small N (N=8)
++
++
+
Statistical analyses
of lesions NR
Medium
[Small N; variation in
chamber
concentrations]
Rat
(Bhalla et al..
1991)
Analytical method and
concentrations NR
+
Small N (N=6)
+ +
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
++
Not Informative
[Failure to report
analytical method
and FA
concentrations; small
N, failure to report
lesion incidence and
severities]
Rat
(Bucklev et
al.. 1984)
+
Formalin; methanol
concentration was not
reported and a
methanol control was
not used
++
++
Note: single concentration
study
Lesion incidence NR
+
Statistical analyses
of lesions NR
Low
[Formalin; failure to
report lesion
incidence]
Mouse
This document is a draft for review purposes only and does not constitute Agency policy.
A-417 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Cassee and
Feron, 1994a)
++
++
++
Note: single concentration
study
+
Incidence and severity of
hyperplasia and
metaplasia reported
together
+
Statistical analyses
of lesions NR
Medium
[Incidence and
severities of
hyperplasia and
metaplasia were
reported together]
Rat
(Cassee et al.,
1996b)
++
+
Small N (N=6)
++
+
Data NRfor 7.9 mg/m3
group
+
Statistical analyses
of lesions NR
Medium
[Small N, failure to
report data for the
7.0 mg/m3 group]
Rat
(Chang et al.,
1983)
++
Sample size N
unclear
Note: single concentration
study; this study
measured reflex
bradypnea
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Low
[Sample size unclear,
failure to report
lesion incidence and
severity]
Rat
(Chang et al.,
1983)
++
Sample size N
unclear
Note: single concentration
study; this study
measured reflex
bradypnea
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Low
[Sample size unclear,
failure to report
lesion incidence and
severity]
Mouse
This document is a draft for review purposes only and does not constitute Agency policy.
A-418 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(lonescu et
al.. 1978)
Test article
characterization NR;
analytical
concentrations NR;
formaldehyde
generation method NR
Test subject strain
and number NR
++
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
++
Not Informative
[Analytical
concentrations NR;
test article
characterization NR;
FA generation
method NR; test
subject strain and
number NR; failure to
report lesion
incidence and
severity]
Rabbit
(Kamata et al.,
+
Formalin; no methanol
control or concentration
was reported. [Note:
high concentration
exposure (179.1 mg/m3)]
+
Small N (N=5)for
histo-pathology
++
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Low
[Formalin; small N for
histopathology;
failure to report
lesion incidence and
severities]
1996)
Rat
(Kuper et al.,
2011)
+
Appears to be freshly
made formalin; although
formaldehyde
generation method NR
+
Small N (N=8)
++
Note: GLP-compliant
study
++
++
High
[Small N]
Rat
(Kuper et al.,
2011)
+
Appears to be freshly
made formalin; although
formaldehyde
generation method NR
+
Small N (N=6)
++
Note: GLP-compliant
study
++
++
High
[Small N]
Mouse
This document is a draft for review purposes only and does not constitute Agency policy.
A-419 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Lima et al.,
2015)
Test article
characterization NR;
concentrations NR-
likely high levels
+
Small N (N=7);
males only
Short (20 min x 3) daily
exposures; controls did
not appear to be chamber
exposed. Note: 5 d
exposure
Lesion severity NR;
lesion incidence
(nonmorphometric
analyses) NR
Note: randomized, but
blinding NR
+
Statistical analyses
of lesions NR
Not Informative
[Failure to
characterize the test
article and report
levels; short
periodicity; lesion
data NR]
Rat
(Monteiro-
Riviere and
++
+
Small N (N=5;
note: only 3/
treated group
examined in
"detail")
++
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Medium
[Small N; lesion
incidence and
severity NR]
Popp, 1986)
Rat
(Monticello et
al.. 1989)
+
Analytical
concentrations NR
++
++
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
++
Medium
[Analytical
concentrations NR;
lesion incidence and
severity NR]
Monkey
(Murta et al.,
2016)
Test article
characterization NR;
concentrations NR-
likely high levels
+
Small N (N=7);
males only
Short (20 min x 3) daily
exposures note: 5 d
exposure
Lesion severity NR;
lesion incidence
(nonmorphometric
analyses) NR
Note: randomized, but
blinding NR
+
Statistical analyses
of lesions NR
Not Informative
[Failure to
characterize the test
article and report
levels; short
periodicity; lesion
data NR]
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-420 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(NTP. 2017)
Mouse
+
Analytic concentrations
NR
++
Note: "randomly
assigned"; Males
only; =25 mice/
group; genetically
modified (Trp53+/-
)
++
Note: 8 wk exposure
duration with 32 wk
follow up was not a
notable issue for these
outcomes as numerous
lesions found
+
Blinding NR; only 3 nasal
sections evaluated (and
1 larynx)
+
Statistical analyses
of lesions NR
Medium
[limited sampling and
minor reporting
limitations]
(Reuzel et al.,
1990)
Rat
++
++
++
++
+
Statistical analyses
of lesions NR
High
(Schreiber et
al.. 1979)
Hamster
Test article
characterization NR;
analytical
concentrations NR;
formaldehyde
generation method NR
Note: high concentration
exposure (307.5 mg/m3)
+
Small N (N=3 to 5)
++
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Not Informative
[Failure to
characterize the test
article, describe the
generation method,
and report analytical
concentrations;
failure to report
lesion incidence and
severities]
(Speit et al.,
2011b)
Rat
+
Formalin; methanol
concentration was not
reported and a
methanol control was
not used
+
Small N (N=6)
++
++
++
Medium
[Small N; formalin]
This document is a draft for review purposes only and does not constitute Agency policy.
A-421 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation0
Data
Considerations
and Statistical
Analvsisd
Overall Confidence
Rating Regarding
Utility for Hazard IDe
(Wilmer et al.,
1987)
+
Analytical
concentrations NR
++
++
Lesion severity NR;
lesion incidence NR
++
Note: intermittent
versus continuous
exposures
compared
Medium
[Analytical
concentrations NR;
failure to report
lesion incidence and
severities]
Rat
(Yorgancilar et
Test article
characterization NR;
analytical
concentrations NR;
formaldehyde
generation method NR
+
Small N (N=8)
+
Note: single concentration
study
Lesion severity NR;
lesion incidence NR
+
Statistical analyses
of lesions NR
Not Informative
[Failure to
characterize test
article; failure to
report analytical
concentrations and
generation method;
small N; failure to
report lesion
incidence and
severities]
al.. 2012)
Rat
NR = not reported; N/A = not applicable.
aGray = inadequate N (N= 1 or 2) or multiple less essential study details (e.g., sex, strain) NR; + = inadequate N (e.g., N= >2 to <10) or individual less essential
study details NR; ++ = adequate N (using guidance from OECD TG 452 and TG 413: chronic: >20 animals/sex/group; subchronic: 10 animals/sex/group,
respectively).
bGray = test protocols for assessing endpoints could not be evaluated or had critical flaws, timing of exposures expected to compromise the integrity of the
protocols, protocols completely irrelevant to human exposure; + = informative components of the protocol were NR/insufficiently assessed, limited human
relevance or single concentration study; ++ = protocol considered relevant to human exposure.
cGray = uncontrolled variables are expected to confound the results or lack of reporting for lesion incidence and severity; + = limited information provided for
observed lesions (i.e., incidence and/or severity) uncontrolled variables may significantly influence results; ++ = adequate reporting of data, no potential
confounding identified.
dGray = failure to report a sufficient amount of data to verify results; + = failure to report statistical analyses; ++ = adequate reporting of data.
designation for Utility for Hazard ID (i.e., confidence) based on EPA judgment regarding the five evaluated criteria, with multiple impactful "gray" categories
generally leading to a designation of "not informative."
This document is a draft for review purposes only and does not constitute Agency policy.
A-422 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-60. Evaluation of controlled inhalation exposure studies examining cell proliferation and mucociliary
function in animals
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint
Evaluation
Data Considerations
& Statistical
Analysis'*
Overall
Confidence
Rating Regarding
Utility for Hazard
ID®
Criteria relevant to
evaluating the
experimental
details within each
experimental
feature category
Exposure quality
evaluations (see
B.4.1.2) are
summarized (++ =
"robust"; + =
"adequate"; gray
box = poor);
relevance of the
tested exposure
levels is discussed in
the hazard synthesis
Sample size
provides
reasonable
power to assess
endpoint(s) in
question;
species, strain,
sex, and age
relevant to
endpoint; no
overt systemic
toxicity noted
or expected
Interpreting the
appropriateness,
reproducibility, and
informativeness of the
study design for
evaluating respiratory
tract pathology.
Although no studies
designed according to
inhalation guidelines
were identified,
several GLP-compliant
studies were identified
and are highlighted
below
The protocols used
to assess respiratory
tract pathology are
sensitive, complete,
discriminating
(specific), and
biologically sound
(reliable);
experimenter bias
minimized
Statistical methods,
group comparisons,
and data/variability
presentation are
appropriate and
discerning
Expert judgement
based on
conclusions from
evaluation of the
5 experimental
feature categories
Cell Proliferation
Andersen et al.
(2008)
Rat
+
=30% variations in
atmospheres
++
++
++
++
High
Andersen et al.
(2010)
Rat
++
+
Variable
sample size
(N=lto 8)
++
++
++
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-423 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint
Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence
Rating Regarding
Utilitvfor Hazard
ID®
(Casanova et
al.. 1994)
Rat
++
++
Relevance of exposure
scenario unclear
(Note: nasal regions
selected for analysis
may not be relevant to
humans)
++
++
Medium
Cassee and Feron
(1994)
Rat
++
+
Number of cells
analyzed NR
++
Note: single
concentration study
++
++
Qualitative data
only
Medium
(Cassee et al.,
1996b)
Rat
++
+
Small N (N=3to
5)
++
+
Data for 7.9 mg/m3
NR
++
High
Chang et al.
(1983)
Rat
++
+
Variable
sample size
(N=4to 9)
Unclear description of
study design
Note: single
concentration study
++
++
Medium
Chang et al.
(1983)
Mouse
++
+
Variable
sample size
(N=4 to 10)
Unclear description of
study design
Note: single
concentration study
++
++
Medium
(Kuper et al.,
2011)
Rat
++
Formaldehyde
generation method
NR
++
++
Note: GLP-compliant
study
++
++
High
(Kuper et al.,
2011)
Mouse
++
Formaldehyde
generation method
NR
++
++
Note: GLP-compliant
study
++
++
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-424 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endpoint
Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence
Rating Regarding
Utilitvfor Hazard
ID®
Meng et al.
(2010)
Rat
+
Analytical
concentrations NR
++
++
++
++
High
Monticello et al.,
1991
Rat
++
+
Variable
sample size
(N=4to 6)
++
++
++
High
(Monticello et
al.. 1989)
Monkey
+
Analytical
concentrations NR
++
+
Note: single
concentration study
+
Qualitative data
only for nasal region
++
Medium
(Monticello et
al.. 1996)
Rat
++
+
Variable
sample size
(N=3 to 8)
+
Nonstandard selection
of nasal regions; Note:
regions may not be
relevant to humans
++
+
Statistical analyses
of cell proliferation
NR
Medium
(Reuzel et al.,
1990)
Rat
++
++
++
++
+ +
High
Roemer et al.
(1993)
Rat
++
++
++
++
++
High
Speit et al. (2011)
Rat
+
Formalin exposure;
no methanol
controls and
concentration NR
++
++
++
++
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-425 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint
Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence
Rating Regarding
Utilitvfor Hazard
ID®
Wilmer et al.
(1987)
Rat
+
Analytical
concentrations NR
Small and
variable sample
size (N=l to 3)
++
++
++
Medium
Wilmer et al.
(1989)
Rat
+
Analytical
concentrations NR
++
++
++
++
High
Woutersen et al.
(1987)
Rat
++
Note: high
concentration
exposure (24.4
mg/m3-day)
Small N (N=2)
++
++
+
Statistical analyses
of cell proliferation
NR
Medium
Zwart et al. (1988)
Rat
++
++
++
++
+
Cell proliferation
data not readily
accessible from
graphic form
High
Mucociliary Function
Flo-Neyret et al.
(2001)
Frog
Not an inhalation
study. Exposure
based on immersion
into formaldehyde
solution (i.e.,
formalin)
+
frogs
Ex vivo amphibian
study; experiments
carried out three days
after sacrifice; mucus
removed from palate
during preparation
and returned to palate
for testing
++
++
Not Informative
This document is a draft for review purposes only and does not constitute Agency policy.
A-426 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (unbolded) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint
Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence
Rating Regarding
Utilitvfor Hazard
ID®
Morgan et al.
(1984)
Frog
+
Analytical
concentrations
within 20% of
nominal
+
frogs
Ex vivo amphibian
study; method of
sacrifice (anesthesia)
and palate harvest NR
+
Inter-animal
variation observed
at several
concentrations
++
Low
Morgan et al.
(1986a)
Rat
++
++
++
Note: mucociliary
function assessed
using dissected nasal
cavities
++
+
Statistical analyses
of mucociliary
function data NR
High
Morgan et al.
(1986c)
Rat
++
++
++
Note: mucociliary
function assessed
using dissected nasal
cavities
++
+
Statistical analyses
of mucociliary
function data NR
High
NR = not reported; N/A = not applicable.
aGray = inadequate N (N= 1 or 2) or multiple less essential study details (e.g., sex, strain) NR; + = inadequate N (e.g., N= >2 to <10) or individual less essential
study details NR; ++ = adequate N.
bGray = Test protocols for assessing endpoints could not be evaluated or had critical flaws, timing of exposures expected to compromise the integrity of the
protocols, protocols completely irrelevant to human exposure; + = informative components of the protocol were NR/insufficiently assessed, limited human
relevance or single concentration study; ++ = protocol considered relevant to human exposure.
cGray = uncontrolled variables are expected to confound the results; + = limited information provided for observations (e.g., qualitative data) or uncontrolled
variables may significantly influence results; ++ = adequate reporting of data, no potential confounding identified.
dGray = failure to report a sufficient amount of data to verify results; + = failure to report statistical analyses; ++ = adequate reporting of data.
designation for Utility for Hazard ID based on EPA judgment and the following criteria: gray = the presence of generally >2 gray boxes in the study feature
categories; low = failure in 2 categories; medium = failure in 1 category; high = no category failures; the presence of multiple +'s may demote tier level.
This document is a draft for review purposes only and does not constitute Agency policy.
A-427 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
Supporting Material for Hazard Analyses of Respiratory Tract Pathology
Supplementary materials relevant to evaluating the evidence for respiratory tract pathology
are described below. Cell proliferation and mucociliary function studies, which inform the potential
mode(s) of action for the induction of respiratory tract pathology following formaldehyde
inhalation, are described in Appendix A.5.6.
Supportive short-term respiratory tract pathology studies in experimental animals
Due to the abundance of high-quality, longer duration exposure studies on respiratory tract
effects in experimental animals, the results of supportive medium and high confidence short-term
studies that did not provide information that was unexamined or inadequately examined in the
longer term studies (i.e., species differences; the relative contribution of concentration and duration
to lesion development) are summarized below (note: the details of low confidence animal studies
are not described for respiratory pathology owing to the large number of high and medium
confidence studies available).
Table A-61. Supportive short-term respiratory pathology studies in animals
Reference and Study Design
Results
RAT
High Confidence
Reuzel et al. (1990)
Wistar rats; male; 10/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 22
hours/day for 3 days.
Test article: Paraformaldehyde.
Actual concentrations were 0,0.37 (±0.01),
1.4 (±0.0), and 3.8 (±0.1) mg/m3.1
This study also evaluated the combined
effects of ozone and FA mixtures on nasal
epithelium. Data presented here in the
Results column are for FA-only exposed
rats.
Histopathologic evaluation of the
respiratory tract included 6 standard
sections of the nose.
Concentration of FA
0 mg/m3
0.37 mg/m3
1.4 mg/m3
lla
llla
II
III
II
III
Disarrangement/loss of cilia without hyper/metaplasia
Minimal to slight
0/10
0/10
0/10
0/10
0/9
0/9
Moderate
0/10
0/10
0/10
0/10
0/9
0/9
Disarrangement/loss of cilia with hyper/metaplasia
Minimal to slight
0/10
0/10
1/10
0/10
2/9
0/9
Moderate
0/10
0/10
0/10
0/10
0/9
0/9
Marked
0/10
0/10
0/10
0/10
0/9
0/9
Keratinization
Minimal to slight
0/10
0/10
0/10
0/10
0/9
0/9
Moderate
0/10
0/10
0/10
0/10
0/9
0/9
Rhinitis
Minimal to slight
0/10
0/10
2/10
0/10
1/9
0/9
Moderate
aLevel in the nose e
0/10
?xamined.
Concentratio
0/10
n of FA
0/10
0/10
0/9
0/9
0 mg/m3
3.8 mg/m3
lla llla
II
III
Disarrangement/loss of cilia without
hyper/metaplasia
This document is a draft for review purposes only and does not constitute Agency policy.
A-428 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and Study Design
Results
Minimal to slight
0/10
0/10
0/10
0/10
Moderate
0/10
0/10
0/10
0/10
Disarrangement/loss
hyper/metaplasia
of
cilia
with
Minimal to slight
0/10
0/10
7/10
3/10
Moderate
0/10
0/10
3/10
5/10
Marked
0/10
0/10
2/10
0/10
Keratinization
Minimal to slight
0/10
0/10
7/10
0/10
Moderate
0/10
0/10
1/10
0/10
Rhinitis
Minimal to slight
0/10
0/10
0/10
0/10
Moderate
0/10
0/10
0/10
0/10
opening
— of ductus
pharyngeus
Figure 1 from Reuzel et al. (1990)
depicting cross levels of the rat nose
evaluated for histopathological lesions.
Main limitations: No major limitations.
aLevel in the nose examined.
Histopathological changes for Level I not reported.
Histopathological changes for Levels IV, V, and VI reported together.
Only change observed was minimal to slight rhinitis in rats (4/10) exposed
to 3.8 mg/m3 FA.
Medium Confidence
Andersen et al., 2008)
Fischer 344 rats; male; 8/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for up to 3weeks.
Rats sacrificed at end of single 6-hour
exposure (day 1), 18 hours after single 6-
hour exposure (day 1 recovery), at end of 5
days of exposure (day 5), at end of 6 days
of exposure (day 6), 18 hours after 6 days
of exposure (day 6 recovery), and at end of
15 days of exposure (day 15).
Test article: Paraformaldehyde.
Actual concentrations were determined on
a daily basis and reported in the Results
column. Target concentrations were 0,
0.9, 2.5, 7.4, and 18.5 mg/m3.1
This study also evaluated the effects of a
single FA instillation (40 nL, 400 mM per
nostril). Data presented here in the
Results column are for inhalation
exposures.
Histopathologic evaluation of the
respiratory tract included nasal sections at
levels I (front of nose), II (anterior lateral
meatus, anterior septum, medial aspect
maxilloturbinate), and III (posterior lateral
meatus, posterior septum).
Target and Actual FA Concentrations3
Target concentration
(mg/m3)
Day 1
(mg/m3)
Day 5
(mg/m3)
Day 6
(mg/m3)
Day 15
(mg/m3)
0
0±0
0±0
0±0
0±0
0.9
0.74±0.23
0.79±0.15
0.75±0.16
0.7±0.11
2.5
2.08±0.46
2.14±0.43
2.26±0.49
2.2±0.31
7.4
5.83±1.73
6.43±0.76
6.00±1.25
6.14±0.97
18.5
17.7±5.7
NA
NA
NA
3Daily means ± SD.
Histopathology Incidence
FA (mg/m3)
0
0.9
2.5
7.4
Time point
lnla
Inl
EH
Inl
EH
Inl
EH
SM
Day 1
0b
1
0
6
0
8
0
0
Day 1 Rc
4
2
1
1
3
7
8
0
Day 5
1
1
0
5
3
8
8
7
Day 6
5
2
0
4
1
7
8
0
Day 6 R
6
1
0
3
2
7
8
0
Day 15
3
1
0
0
2
5
7
0
0 ppm: EH and SM were ND; 0.7 ppm: SM was ND; 2 ppm SM was ND
alnl = inflammatory infiltrate; EH = epithelial hyperplasia; SM = squamous
metaplasia.
bNumber of animals with the lesion (n = 8).
cRecovery group.
Histopathological Incidence
FA (mg/m3)
0
18.5
This document is a draft for review purposes only and does not constitute Agency policy.
A-429 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and Study Design
Level 1
Level II
Main limitations: small sample size;
Time point
lnla
Inl
UcL
EH
Inl
UcL
EH
somewhat high variability in chamber
Day 1
0b
8
NR
NR
7
2
1
Results
concentrations.
0 ppm: UcLwas NR, EH was ND
alnl = inflammatory infiltrate; UcL = ulcerative lesions; EH = epithelial
hyperplasia.
bNumber of animals with the lesion (n = 8).
Cassee and Feron (1994)
Wistar rats; male; 20/group.
Exposure: Rats were exposed in dynamic
nose-only chambers for 3 days (6
consecutive 12-hour periods of 8 hours of
exposure to FA followed by 4 hours of
nonexposure). Rats sacrificed immediately
(i.e., within 30 minutes) after last
exposure.
Test article: Paraformaldehyde.
Actual concentrations were 0 and 4.4 (SE ±
0.1) mg/m3 FA.
Histopathologic evaluation of the
respiratory tract included standard cross
sections of the head (see cross sections in
(Reuzel et al.. 1990).
Controls
FA
Type of lesions
lla
llla
II
III
Disarrangement, flattening and slight basal cell hyperplasia
Minimal
0/5
1/5
0/5
0/5
Slight
0/5
0/5
0/5
0/5
Frank necrosis
0/5
0/5
5/5
5/5
Hyperplasia accompanied by squamous metaplasia
Slight
0/5
0/5
2/5
3/5
Moderate
0/5
0/5
2/5
2/5
Marked
0/5
0/5
1/5
0/5
Rhinitis
Slight"
0/5
0/5
0/5
0/5
Moderate
0/5
0/5
0/5
4/5
Marked
0/5
0/5
5/5
1/5
standard cross section level II and III.
blnflux of neutrophils mainly observed.
Main limitations: hyperplasia
metaplasia were reported together.
and
This study also evaluated the nasal
changes induced by exposures to ozone
alone and FA and ozone. Data presented
here in the Results column are for FA-only
exposures.
(Cassee et al.. 1996b)
Wistar rats; male; number of animals per
group varied but are reported in the
Results column.
Exposure: Rats were exposed to FA in
dynamic nose-only chambers 6 hours/day
for 1 or 3 days. Rats sacrificed immediately
after last exposure.
Test article: Paraformaldehyde.
Actual concentrations were 0,1.2, 3.9, and
7.9 mg/m3.1
Histopathologic evaluation of the
respiratory tract included standard cross
sections at levels II, III, and/or IV of the
nose (see (Reuzel et al.. 1990)
for cross-sectional levels).
1-day exposure: no treatment-related histopathological nasal lesions
observed
Histopathological changes from 3 days of exposure3
FA (mg/m3)
Site, type, degree, and incidence of lesions
0
1.2
3.9
Number of noses examined
19
5
6
Disarrangement, necrosis, thickening, and desquamation of
respiratory/transitional epithelium15
Slight (mainly disarrangement)
0
0
3
Moderate
0
0
2
Severe (extensive)
0
0
0
Basal cell hyperplasia and/or increased number of mitotic figures in
respiratory/transitional epithelium
Slight (focal)
0
0
4
Moderate
0
0
2
Severe (extensive)
0
0
0
This document is a draft for review purposes only and does not constitute Agency policy.
A-430 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and Study Design
Results
Increased incidence of "single-cell necrosis" in olfactory epithelium0
A few necrotic cells
0
0
0
A moderate number of necrotic cells
0
0
0
Many necrotic cells
0
0
0
Atrophy of olfactory epithelium
Slight (mainly disarrangement)
0
0
0
Moderate (focal)
0
0
0
Severe (extensive)
0
0
0
Rhinitis
Slight
2
1
0
Moderate
1
0
0
Severe
0
0
0
Main limitations: small N; failure to report
data for the 7.9 mg/m3 group.
This study also evaluated the combined
effects of FA, acetaldehyde, and acrolein
on nasal epithelium. Data presented here
in the Results column are for FA-only
exposed rats.
aData for 7.9 mg/m3 group NR.
bChanges observed at levels II and III.
cChanges observed at levels III and IV.
Monteiro-Riviere and Popp (1986)
Fischer 344 rats; male; 3-5/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day for either 1,2, or4 days. Interim
sacrifices were performed either
immediately or 18 hours after last
exposure.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.6 (±0.1),
2.7 (±0.4), 7.3 (±0.1), and 18.2 (±0.4)
mg/m3.1
Histopathologic evaluation of the
respiratory tract included transverse
sections of the skull that contained the
dorsal nasal concha, lateral wall, and
ventral nasal concha.
Main limitations: small N; (note: only 3 of
5 rats/ treatment group were evaluated in
"detail"); failed to report lesion incidence
and severity
Cellular occurrence
of ultrastructure
lesionab
Cytoplasmic
vacuoles
Autophagic
vacuoles
Loss of microvilli
Hypertrophy
SER in apical region
Intracytoplasmic
lumen
Mitochondrial
swelling
Neutrophils
Intercellular edema
Ciliated
cells
mucous
7.3
mg/m3c
ALL
BA
CI
7.3 mg/m3
(l-day)d
ALL
BA
CI
CI, GO
NC
7.3 mg/m3
(2-day)
CI
CI, GO
CI
7.3 mg/m3
(4-day)
NC
BA, CU, NC
CI, CU, BR
CI, GO
NC
CI, BR
Nonkeratinized
squamous cells
Abbreviations: BA, basal cells; CI, ciliated cells; CU, cuboidal cells; BR,
brush cells; NC, nonciliated columnar cells; GO, goblet cells; SER, smooth
endoplasmic reticulum; ALL, all cell types; +, indicates presence.
Nucleolar segregation, pyknotic nuclei, and internalized cilia not
observed.
bThese lesions were not observed at 0.6 mg/m3 (1 or 4 days exposure) or
2.7 mg/m3 (1 or 4 days exposure) FA.
cRats in this group were immediately sacrifice after exposure.
dNumber of days of exposure, rats sacrificed 18 hours later.
Cellular occurrence
of
18.2 mg/m3
18.2 mg/m3
ultrastructure lesionab
(l-day)c
(2-day)
Cytoplasmic vacuoles
CU, NC
NC
Autophagic vacuoles
BA, CI, CU, NC
BA, CU, NC
This document is a draft for review purposes only and does not constitute Agency policy.
A-431 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference and Study Design
Results
Loss of microvilli
BA, CI, CU
CI, CU, NC
SER in apical region
NC
NC
Nucleolar segregation
BA, CU
BA, CU
Pyknotic nuclei
CU
CI
Internalized cilia
CI
CI
Neutrophils
+
Intercellular edema
+
Nonkeratinized squamous
+
+
cells
Abbreviations: BA, basal cells; CI, ciliated cells; CU, cuboidal cells; BR,
brush cells; NC, nonciliated columnar cells; GO, goblet cells; SER,
smooth endoplasmic reticulum; ALL, all cell types; +, indicates presence.
Hypertrophy, Intracytoplasmic lumen, mitochondrial swelling, and
ciliated mucous cells not observed.
bThese lesions were not observed at 0.6 mg/m3
(1 or 4 days exposure)
or 2.7 mg/m3 (1 or 4 days exposure) FA.
cNumber of days of exposure, rats sacrificed 18 hours later.
Speit et al. (2011)
No FA-related histological changes observed in levels 1—IV of rats exposed
Fischer 344 rats; males; 6/group.
to 0.63,1.23, 2.48, and 7.53 mg/m3.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
Histopathological analysis of nasal lesions after 4 weeks
hours/day, 5 days/week for 4 weeks.
Incidence and grading of findings3
Test article: Formalin (methanol
FA (mg/m:
)
concentration NR).
Gradeb
0
12.3
18.4
Actual concentrations were 0, 0.63 (±0.6),
Level 1
1.23 (±0.14), 2.48 (±0.18), 7.53 (±0.42),
Metaplasia, squamous
1
0
1
0
12.3 (±0.48), 18.4 (±0.06) mg/m3.1
2
0
5
0
Histopathologic evaluation of the
respiratory tract included 4 levels of the
nasal cavity: 1 (nasal septum, lateral
3
0
0
4
4
0
0
2
Degeneration, (multi) focal
2
0
0
1
3
0
0
3
meatus [wall], maxilloturbinate,
0
0
nasoturbinate), II (nasal septum, lateral
meatus [wall]), and III and IV
(nasopharynx).
Inflammation, (multi) focal
2
0
0
1
3
0
0
4
Level II
Main limitations: Formalin; small N
Metaplasia, squamous
2
0
0
1
3
0
0
5
Degeneration, (multi) focal
1
0
0
1
2
0
0
2
3
0
0
3
Inflammation, (multi) focal
2
0
0
1
Level III
Metaplasia, transitional
1
0
0
4
2
0
0
1
Level IV
Metaplasia, transitional
1
0
0
2
2
0
0
3
aNumber of animal with lesions (6 analyzed per group).
bl = minimal; 2 = slight; 3 = moderate; 4 = severe/marked.
This document is a draft for review purposes only and does not constitute Agency policy.
A-432 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
1Study authors originally reported FA concentrations in ppm. These values were converted based on 1 ppm = 1.23
mg/m3, assuming 25°C and 760 mm Hg.
Abbreviations: FA—Formaldehyde; NA—Not applicable; ND—Not detected; NR—Not reported; SD—Standard
deviation; SE—Standard error of the mean.
A.5.6. Mechanistic Evidence Related to Potential Noncancer Respiratory Health Effects
Note: Large sections of this analysis are redundant to synthesis text, figures, and tables
presented in the Toxicological Review and Assessment Overview. However, the entirety of the
analyses and discussion is included below to contextualize the conclusions described in the
Toxicological Review with the appropriate methodological considerations, supporting analyses, and
other information of potential interest.
Organization and Methods
This evaluation provides an integrated discussion characterizing potential relationships
between the mechanistic changes observed following formaldehyde inhalation in the context of
potential respiratory effects, but it does not attempt to explicitly define a single mode of action.
Literature search strategy
Through 2017, studies were identified through one of two strategies, namely, identification
of studies relevant to mechanisms for potential respiratory effects during systematic searches for
health hazard-specific toxicity information (see Appendix A.5.2-A.5.5), or through an independent
systematic literature search focused on inflammation- and immune-related changes (discussed
here). This latter effort was undertaken to identify mechanistic information related to changes in
the respiratory tract, blood, and lymphoid tissues that might not have been captured by health
effect-specific systematic searches. The comprehensiveness of this strategy was compared against
citations in the recent National Academy of Sciences review of the National Toxicology Program
Report on Carcinogens (NRC. 2014). and some supportive information from that report is noted in
this analysis16 (i.e., hematological findings from four foreign language studies: (Cheng etal.. 2004):
{Tang 2003, }; (Tonget al.. 2007): and (Yang, 2007, }. Given the breadth of this topic, this section
uses a hierarchical approach to screen, sort, and distill information from over 10,000 references
identified across these searches. Thus, additional steps were taken to focus this analysis on the
most influential information. In addition to criteria identifying studies as relevant to assessing
potential respiratory system changes, studies that failed to report a specific estimate of
formaldehyde exposure (e.g., concentration, duration) were not considered. Also, studies of in vitro
exposure to formaldehyde in solution and of exposure routes other than inhalation, which may
inform mechanistic understanding were initially kept for possible further review or qualitative
16
Also identified from the NRC review and considered, but not ultimately included, in this section: {Qian, 1988,
} (an abstract); (Pongsavee. 2011) (ex vivo exposure to nongaseous formaldehyde; did not meet the inclusion
criteria); and (Vargova et al.. 1992) (evaluated and considered "not informative").
This document is a draft for review purposes only and does not constitute Agency policy.
A-433 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
support of POE-related findings. However, given the large number of studies reporting results from
inhalation exposure in vivo or gaseous exposure of airway cells, and considering the uncertainties
associated with the toxicokinetics of noninhalation exposures, these comparably far less influential
mechanistic data were ultimately not included in the final analysis described herein. These
considerations informed the focus of the separate, systematic evidence map, developed to update
the literature from 2017 to 2021 (see Appendix F).
Literature Search
A systematic evaluation of the literature database on studies examining potential
mechanistic events pertaining to noncancer respiratory health effects in relation to formaldehyde
exposure was initially conducted in August 2014, with yearly updates through 2017 (a separate
Systematic Evidence Map updates the literature from 2017-2021 using parallel approaches, see
Appendix F). The search strings used for the pre-2017 literature search were designed to
emphasize identification of mechanistic effects related to inflammation or immune-related changes,
as the expectation was that most other relevant mechanistic effects would be identified through the
health effect-specific literature searches in Appendix A.5.2-A.5.5. However, these strings (see Table
A-62) returned a much wider range of studies than expected. Thus, the primary source of studies
for this section comes from this specific literature search, while a small number of studies not
identified through this search are included based on searches and screening protocols from the
health effect-specific searches. Additional search strategies included:
• Addition of nonoverlapping (many references identified by the search terms in Table A-62
were also identified by health effect-specific literature searches) references describing
mechanistic effects relevant to interpreting respiratory effects, as identified by other health
effect-specific literature searches.
• Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
EPA. 2010a], the ATSDR toxicological profile of formaldehyde fATSDR. 19991. and the
National Toxicology Program (NTP) report on carcinogens background document for
formaldehyde (NTP. 2010). Note: although no specific references were added to the
literature search as a result of this review, several references are footnoted as supportive
information.
After manual review and removal of duplication citations, the articles identified from
database searches were initially screened within an EndNote library for relevance; title and
abstract were considered simultaneously in this process, followed by subsequent review of the full
text The search and screening strategy, including exclusion categories applied and the number of
articles excluded within each exclusion category, is summarized in Figure A-32. Based on this
process, 140 studies were identified and evaluated for consideration in the Toxicological Review.
Given the size of the database of mechanistic studies available for review, some constraints were
placed on the studies considered for inclusion. Studies that failed to include a comparison to
quantified formaldehyde exposure (e.g., levels; duration) were excluded. As noninhalation studies
This document is a draft for review purposes only and does not constitute Agency policy.
A-434 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 poorly replicate the distribution of inhaled formaldehyde, studies of noninhalation exposure and
2 nongaseous in vitro exposure were set aside for possible use (note: these were ultimately not
3 included in the final analysis because EPA concluded that a sufficient number of mechanistic studies
4 employing inhalation exposure were identified). Similarly, a single thesis identified during the
5 literature search was ultimately not included in the final analysis. Given the multitude of
6 potentially relevant studies returned, and because this review focuses on mechanisms most likely
7 to be relevant to respiratory tract effects in humans, nonmammalian models and tissue systems
8 other than those that might be related to formaldehyde-induced respiratory effects (i.e., other than
9 studies of the respiratory tract, or circulatory or immune-related effects) were excluded. The
10 specific inclusion and exclusion criteria used in the screening step are described in Table A-63.
Table A-62. Summary of supplemental literature search terms for mechanistic
studies relevant to potential noncancer respiratory health effects
Database
Search (no date limit thru 8/31/2014)
PubMed
searched 9/4/2014
(*formaldehyde OR formalin) AND ("Adaptive immunity" OR asthma OR "atopic dermatitis"
OR immune OR "innate immunity" OR redox OR allergic OR allergy OR "mucosal immunity"
OR Eosinophil* OR Inflammation OR "Lung function test" OR "Nitric oxide" OR Wheezing
OR rhinosinusitis OR lymphocyte OR bronchiolitis OR glucocorticoid OR IgE OR basophil OR
"histamine-releasing factor" OR "mast cell" OR "reactive nitrogen species" OR "reactive
oxygen species" OR "oxidative stress" OR isoprostane OR "Airway remodeling" OR
phagocytosis OR "toll-like" OR "respiratory immunity" OR autoimmune OR interleukin OR
"immune system" OR "allergic rhinitis" OR "chronic obstructive pulmonary disease" OR
copd OR corticosteroids OR "Chronic bronchitis" OR fibrocyte OR hematopoie* OR
"Epithelial injury" OR "epithelial repair" OR Thl7 OR "Airway hyperresponsiveness" OR
"Airway smooth muscle" OR "airway hyperreactivity" OR "Bronchoalveolar lavage" OR
neutrophil OR cytokine OR Bronchiectasis OR th2 OR th9 OR "t cell" OR leukotriene OR
"Bronchial epithelial cell" OR "Dendritic cell" OR Endothelin OR "growth factor" OR Lipoxins
OR Prostaglandin OR cyclooxygenase OR "matrix metalloproteinase" OR ovalbumin OR
"tumor necrosis factor" OR Phosphodiesterase OR "Bronchopulmonary dysplasia" OR
Adipokine OR Eicosanoid OR bronchoconstriction OR Phospholipase OR Hyperpnoea OR
bronchiectasis OR "corticosteroid responsiveness" OR "Type 2" OR "muscarinic receptor
antagonism" OR "obstructive airway" OR Immunomodulation OR lipocalins OR allergen OR
corticosteroids OR "Vascular endothelial growth factor" OR bronchiectasis OR
immunodeficiency OR "Muscarinic receptor" OR inflammatory OR Complement OR
"Myeloid suppressor cell" OR immunoglobulin OR mucin OR Autophagy OR Leukocyte OR
macrophage OR BALT OR "extracellular lining fluid") NOT (nocicept* OR pain OR "formalin
test" OR "formalin-induced" OR "formaldehyde-fixed" or "formalin-fixed" OR
"paraformaldehyde-fixed" OR "formaldehyde fixation" OR "formalin fixation" OR "10%
formalin" OR "10% buffered formalin" OR "10% neutral buffered formalin" OR vaccin* OR
inactivated OR "formalin-killed" or "formaldehyde-killed" OR dental OR formalinized)
Web of Science
searched 9/5/2014
(TS=("formaldehyde" OR "formalin") AND TS=("Adaptive immunity" OR "asthma" OR
"atopic dermatitis" OR "immune" OR "innate immunity" OR "redox" OR "allergic" OR
"allergy" OR "mucosal immunity" OR Eosinophil* OR "Inflammation" OR "Lung function
test" OR "Nitric oxide" OR "Wheezing" OR "rhinosinusitis" OR "lymphocyte" OR
"bronchiolitis" OR "glucocorticoid" OR "IgE" OR "basophil" OR "histamine-releasing factor"
OR "mast cell" OR "reactive nitrogen species" OR "reactive oxygen species" OR "oxidative
stress" OR "isoprostane" OR "Airway remodeling" OR "phagocytosis" OR "toll-like" OR
This document is a draft for review purposes only and does not constitute Agency policy.
A-435 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Database
Search (no date limit thru 8/31/2014)
"respiratory immunity" OR "autoimmune" OR "interleukin" OR "immune system" OR
"allergic rhinitis" OR "chronic obstructive pulmonary disease" OR "copd" OR
"corticosteroids" OR "Chronic bronchitis" OR "fibrocyte" OR hematopoie* OR "Epithelial
injury" OR "epithelial repair" OR "Thl7" OR "Airway hyperresponsiveness" OR "Airway
smooth muscle" OR "airway hyperreactivity" OR "Bronchoalveolar lavage" OR "neutrophil"
OR "cytokine" OR "Bronchiectasis" OR "th2" OR "th9" OR "t cell" OR "leukotriene" OR
"Bronchial epithelial cell" OR "Dendritic cell" OR "Endothelin" OR "growth factor" OR
"Lipoxins" OR "Prostaglandin" OR "cyclooxygenase" OR "matrix metalloproteinase" OR
"ovalbumin" OR "tumor necrosis factor" OR "Phosphodiesterase" OR "Bronchopulmonary
dysplasia" OR "Adipokine" OR "Eicosanoid" OR "bronchoconstriction" OR "Phospholipase"
OR "Hyperpnoea" OR "bronchiectasis" OR "corticosteroid responsiveness" OR "Type 2" OR
"muscarinic receptor antagonism" OR "obstructive airway" OR "Immunomodulation" OR
"lipocalins" OR "allergen" OR "corticosteroids" OR "Vascular endothelial growth factor" OR
"bronchiectasis" OR "immunodeficiency" OR "Muscarinic receptor" OR inflammatory OR
"Complement" OR "Myeloid suppressor cell" OR "immunoglobulin" OR "mucin" OR
"Autophagy" OR "Leukocyte" OR "macrophage" OR "BALT" OR "extracellular lining fluid"))
NOTTS=(nocicept* OR "pain" OR "formalin test" OR "formalin-induced" OR "formaldehyde-
fixed" OR "formalin-fixed" OR "paraformaldehyde-fixed" OR "formaldehyde fixation" OR
"formalin fixation" OR "10% formalin" OR "10% buffered formalin" OR "10% neutral
buffered formalin" OR vaccin* OR "inactivated" OR "formalin-killed" or "formaldehyde-
killed" OR "dental" OR "formalinized")
lndexes=SCI-EXPANDED, CPCI-S, BKCI-S, BKCI-SSH Timespan=AII years
Toxline
searched 9/3/2014
Part 1
@SYN0+@AND+(@OR+"Adaptive+immunity"+asthma+"atopic+dermatitis"+immune+"inn
ate+immunity"+redox+allergic+allergy+"mucosal+immunity"+Eosinophil*+lnflammation+
"Lung+function+test"+"Nitric+oxide"+Wheezing+rhinosinusitis+lymphocyte+bronchiolitis
+glucocorticoid+lgE+basophil+"histamine-
releasing+factor"+"mast+cell"+"reactive+nitrogen+species"+"oxidative+stress"+isoprosta
ne+"Airway+remodeling"+phagocytosis+"toll-
like"+"respiratory+immunity"+autoimmune+interleukin+"immune+system"+"allergic+rhin
itis"+"chronic+obstructive+pulmonary+disease")+(@OR+formaldehyde+formalin+@term+
@rn+50-00-0)+@NOT+(@OR+nocicept*+pain+"formalin+test"+"formalin-
induced"+"formaldehyde-fixed"+"formalin-fixed"+"paraformaldehyde-
fixed"+"formaldehyde+fixation"+"formalin+fixation"+"buffered+formalin"+"neutral+buffe
red+formalin"+vaccin*+inactivated+"formalin-killed"+"formaldehyde-
killed"+dental+formalinized)+@NOT+@org+pubmed+pubdart+"NIH+reporter"
@SYN0+@AND+(@OR+"Adaptive+immunity"+asthma+"atopic+dermatitis"+immune+"inn
ate+immunity"+redox+allergic+allergy+"mucosal+immunity"+Eosinophil*+lnflammation+
"Lung+function+test"+"Nitric+oxide"+Wheezing+rhinosinusitis+lymphocyte+bronchiolitis
+glucocorticoid+lgE+basophil+"histamine-
releasing+factor"+"mast+cell"+"reactive+nitrogen+species"+"oxidative+stress"+isoprosta
ne+"Airway+remodeling"+phagocytosis+"toll-
like"+"respiratory+immunity"+autoimmune+interleukin+"immune+system"+"allergic+rhin
itis"+"chronic+obstructive+pulmonary+disease")+(@OR+formaldehyde+formalin+@term+
@rn+50-00-0)+@NOT+(@OR+nocicept*+pain+"formalin+test"+"formalin-
induced"+"formaldehyde-fixed"+"formalin-fixed"+"paraformaldehyde-
fixed"+"formaldehyde+fixation"+"formalin+fixation"+"buffered+formalin"+"neutral+buffe
red+formalin"+vaccin*+inactivated+"formalin-killed"+"formaldehyde-
killed"+dental+formalinized)+@AND+@org+"nih+reporter"
This document is a draft for review purposes only and does not constitute Agency policy.
A-436 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Database
Search (no date limit thru 8/31/2014)
Part 2
@SYNO+@AND+(@OR+copd+corticosteroids+"Chronic+bronchitis"+fibrocyte+hematopoi
e*+"Epithelial+injury"+"epithelial+repair"+Thl7+"Airway+hyperresponsiveness"+"Airway
+smooth+muscle"+"airway+hyperreactivity"+"Bronchoalveolar+lavage"+neutrophil+cytok
ine+Bronchiectasis+th2+th9+"t+cell"+leukotriene+"Bronchial+epithelial+ceN"+"Dendritic+
cell"+Endothelin+"growth+factor"+Lipoxins+Prostaglandin+cyclooxygenase+"matrix+meta
lloproteinase"+ovalbumin+"tumor+necrosis+factor"+Phosphodiesterase+"Bronchopulmo
nary+dysplasia"+Adipokine+Eicosanoid+bronchoconstriction+Phospholipase+Hyperpnoea
+bronchiectasis+"corticosteroid+responsiveness"+"Type+2"+"muscarinic+receptor+antag
onism"+"obstructive+airway"+lmmunomodulation+lipocalins+allergen+corticosteroids+"
Vascular+endothelial+growth+factor"+bronchiectasis+immunodeficiency+"Muscarinic+re
ceptor"+inflammatory+Complement+"Myeloid+suppressor+cell"+immunoglobulin+mucin
+Autophagy+Leukocyte+macrophage+BALT+"extracellular+lining+fluid")+(@OR+formalde
hyde+formalin+@term+@rn+50-00-
0)+@NOT+(@OR+nocicept*+pain+"formalin+test"+"formalin-induced"+"formaldehyde-
fixed"+"formalin-fixed"+"paraformaldehyde-
fixed"+"formaldehyde+fixation"+"formalin+fixation"+"buffered+formalin"+"neutral+buffe
red+formalin"+vaccin*+inactivated+"formalin-killed"+"formaldehyde-
killed"+dental+formalinized)+@NOT+@org+pubmed+pubdart+"NIH+reporter"
@SYNO+@AND+(@OR+copd+corticosteroids+"Chronic+bronchitis"+fibrocyte+hematopoi
e*+"Epithelial+injury"+"epithelial+repair"+Thl7+"Airway+hyperresponsiveness"+"Airway
+smooth+muscle"+"airway+hyperreactivity"+"Bronchoalveolar+lavage"+neutrophil+cytok
ine+Bronchiectasis+th2+th9+"t+cell"+leukotriene+"Bronchial+epithelial+ceN"+"Dendritic+
cell"+Endothelin+"growth+factor"+Lipoxins+Prostaglandin+cyclooxygenase+"matrix+meta
lloproteinase"+ovalbumin+"tumor+necrosis+factor"+Phosphodiesterase+"Bronchopulmo
nary+dysplasia"+Adipokine+Eicosanoid+bronchoconstriction+Phospholipase+Hyperpnoea
+bronchiectasis+"corticosteroid+responsiveness"+"Type+2"+"muscarinic+receptor+antag
onism"+"obstructive+airway"+lmmunomodulation+lipocalins+allergen+corticosteroids+"
Vascular+endothelial+growth+factor"+bronchiectasis+immunodeficiency+"Muscarinic+re
ceptor"+inflammatory+Complement+"Myeloid+suppressor+cell"+immunoglobulin+mucin
+Autophagy+Leukocyte+macrophage+BALT+"extracellular+lining+fluid")+(@OR+formalde
hyde+formalin+@term+@rn+50-00-
0)+@NOT+(@OR+nocicept*+pain+"formalin+test"+"formalin-induced"+"formaldehyde-
fixed"+"formalin-fixed"+"paraformaldehyde-
fixed"+"formaldehyde+fixation"+"formalin+fixation"+"buffered+formalin"+"neutral+buffe
red+formalin"+vaccin*+inactivated+"formalin-killed"+"formaldehyde-
killed"+dental+formalinized)+@AND+@org+"nih+reporter"
Abbreviations: Majr = major topic (filter); TS = the requested "topic" is included as a field tag.
Table A-63. Inclusion and exclusion criteria for mechanistic studies relevant
to potential noncancer respiratory health effects
Included
Excluded
Population
0.63 Experimental animals
0.64 Humans
0.65 Irrelevant species or matrix* including
nonanimal species (e.g., bacteria) and studies of
inorganic products
Exposu re
0.66 Quantified (e.g., levels;
duration) exposure to
formaldehyde in indoor
air
0.67 Not specific to formaldehyde* (e.g., other
chemicals)
0.68 No specific comparison to formaldehyde
exposure alone (e.g., formaldehyde levels, duration,
This document is a draft for review purposes only and does not constitute Agency policy.
A-437 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Included
Excluded
or similar in a study of exposure to a mixture)—
NOTE: full text screening only
0.69 Nonrelevant exposure paradigm* (e.g., use as a
pain inducer in nociception studies)
0.70 Outdoor air exposure
Comparison
0.71 Inclusion of a
0.72 Case reports (selected references used for
comparison group (e.g.,
illustration)
pre- or postexposure; no
exposure; lower
formaldehyde exposure
level)
Outcome
0.73 Examining
0.74 Not relevant endpoints for section* including
mechanistic endpoints
carcinogenicity studies and endpoints related to
relevant to interpretions
contact dermatitis
of potential respiratory
0.75 Exposure or dosimetry studies*
health effects
0.76 Use of formaldehyde in methods* (e.g., for
fixation)
0.77 Processes related to endogenous formaldehyde*
0.78 Related to hazard endpoints only* (including
genotoxicity; see those hazard sections)—NOTE: full
text screening only
Other
• Original primary
0.79 Not a unique, primary research article*
research article
including reviews, reports, commentaries, meeting
abstracts, duplicates, or untranslated foreign
language studies (these were determined to be off
topic or unlikely to have a significant impact based
on review of title, abstract, and/or figures).
This document is a draft for review purposes only and does not constitute Agency policy.
A-438 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Toxline, TSCATS,
& DART Web of Science
Not / \ Subject
PubMed / \ Area
Figure A-30. Literature search documentation for sources of primary data
pertaining to inhalation formaldehyde exposure and mechanistic data
associated with potential noncancer effects on the respiratory system (reflects
studies identified in searches conducted through September 2016; see Appendix F
for literature identification from 2016-2021).
This document is a draft for review purposes only and does not constitute Agency policy.
A-439 ' DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
Organizing and judging the evidence for mechanistic events and associations between events
Due to the importance of considering the toxicokinetics of inhaled formaldehyde, the human
and animal experiments interpreted with high or medium confidence and low confidence were
organized according to the tissue compartment and general type of change being examined.
Individual experiments or groups of closely related experiments across studies were divided into
mechanistic events, representing empirically observable biological changes that may inform how
formaldehyde exposure might be associated with a respiratory health effect(s). Mechanistic event is
used in this section as a generic term for types of endpoints, which may or may not be required
for—or even influence—a mode of action; thus, mechanistic events are not necessarily key events,
which are necessary precursor steps (or markers of such) in a mode of action {U.S. EPA, 2005,}.
The level of evidentiary support for each mechanistic event was characterized based on the criteria
presented in Table A-64. These criteria emphasize the confidence and consistency of the data
across studies. Other relevant considerations (e.g., effect magnitude, dose-response, coherence) are
discussed when conclusions across studies could be drawn, but these judgments were often difficult
due to the heterogeneous nature of the available mechanistic studies. This section presents the
broad conclusions drawn from sets of related studies.
Potential associations between mechanistic events were judged based on the
tissue(s)/region(s) assessed and known biological roles within those tissues for the identified
mechanistic events. The basis for each association was not individually documented, but these are
generally discussed in the synthesis sections below and/or the study evaluation tables in the "Study
Evaluations" section below.
This document is a draft for review purposes only and does not constitute Agency policy.
A-440 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-64. Criteria and presentation of strength of the evidence for each
mechanistic event and for potential associations between events relating to
potential respiratory health effects
Evidence
judgment3
Mechanistic events
Associations between mechanistic
events
Criteria for conclusions
Presentation
Criteria for conclusions
Presentation
STRONGEST
Robust
Direct evidence supporting an
effect in multiple, consistent high
or medium confidence studies b
O
Emphasized in
Text
Formaldehyde-specific data
demonstrate a linkage (i.e.,
inhibition of mechanistic
event "A" prevents or
reduces the occurrence of
event "B"; events "A" and
"B" are linked by
concentration, location,
and temporality)
->
WEAKEST
Moderate
Direct or indirect (e.g., genetic
changes) evidence supporting an
effect in at least 1 high or medium
confidence study, with supporting
evidence (e.g., consistent changes
suggesting an effect in low
confidence studies)b
O
Emphasized in
Text
• An association between
events "A" and "B" is
known based on
established (basic)
biology
• An association has been
demonstrated for
similar chemicals and/or
effects
->
Slight
• Evidence supporting an effect
in 1 hypothesis-generating high
or medium confidence study
• Evidence suggesting an effect in
multiple, reasonably consistent
low confidence studies
r.
N. '
Minimal
Discussion in
Text
An association is justifiable,
or even expected, based on
underlying biology, but it
has not been well-
established (note: events
for which an association is
unlikely based on
established understanding
of underlying biology are
not linked)
¦¦¦¦>
Indetermin
ate
• Evidence suggesting an effect in
1 low confidence study
• A set of low confidence studies
with inconsistent results
Not included in
figures; may be
noted in text
N/A
N/A
• Evidence cannot be interpreted
(no data; no pattern in results
within and/or across studies)
• Data suggest no change
Not included in
figures or
synthesis text
N/A
N/A
aFor consistency, the judgments used to describe the within-stream conclusions for apical health effect endpoints
were applied, although the criteria used herein were less rigorous (i.e., when evaluating individual studies and
sets of studies). Unlike within-stream conclusions, these terms are not bolded as they do not reflect evidence
stream conclusions.
'The presence of a comparable or stronger set of studies with directly conflicting evidence results in the identification of the
next weaker evidence descriptor (e.g., robust evidence with conflicting data would be moderate); note that the purpose of
this evaluation was not to identify mechanistic events for which there was robust evidence of no change; however, the
plausibility of the pathways (considering evidence for a lack of changes in expected events) is discussed in later sections.
This document is a draft for review purposes only and does not constitute Agency policy.
A-441 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Display and analysis of the mechanistic evidence
This chapter first describes the data for mechanistic events within each of the assessed
tissue locations, and then analyzes the most informative data (i.e., preference is given to robust
evidence) integrated across tissue compartments, both of which highlight potential effects on
specific tissue components and/or functions. Both analyses include a discussion of the mechanistic
events interpreted as the most likely to be due to (or most closely related to) direct interactions
with inhaled formaldehyde molecules (i.e., "plausible initial effects of exposure"), as well as
important apical toxicity endpoints (i.e., "key features of a potential hazard") discussed in previous
sections (see Sections 1.2 and 1.3). In the first portion of this section, the network-based
presentation serves to evaluate the interconnectivity of mechanistic changes within and across
tissue compartments, and across potential noncancer respiratory system health effects. As an
integrated overview, the analysis focuses primarily on the mechanistic events with robust and
moderate evidence of formaldehyde-induced changes (see Figure A-33), but also includes
consideration of the mechanistic events with slight evidentiary support (see Figure A-34). Where
data clearly suggest a dependence on exposure duration or exposure level to elicit an effect, these
associations are discussed. Note that this illustration is likely not a comprehensive picture of all
potential formaldehyde-induced mechanistic changes or interactions between events, as it is based
exclusively on events for which formaldehyde-specific data are available and which were captured
by the literature search and screening process described above.
In the latter portion of this section, the network of mechanistic changes across tissues is
distilled to the subsets of evidence that best link initial effects of formaldehyde inhalation in a linear
fashion to key features for each of the noncancer respiratory system health effects evaluated in
previous sections (see Figure A-35). In this analysis, for each of the more apical toxicity endpoints,
the sequence of events interpreted to have the most reliable evidence (e.g., mechanistic events and
associations with robust evidence are preferred) from a "plausible intial effect of exposure" are
organized in a linear fashion, regardless of tissue region. This latter analysis attempts to simplify
the data and emphasize the mechanistic events supported by the evidence interpreted with the
highest confidence, but it is not intended to convey the majority of the available information.
Aspects of this latter analysis are similar to components of the adverse outcome pathway (AOP)
approach {Villeneuve, 2013; 2014, }. These analyses only consider mechanistic events identified in
formaldehyde-specific studies. The data supporting each sequence of events depicted in Figure A-
34 are summarized into an interpretation regarding the biological plausibility of that sequence
being a mechanism by which formaldehyde exposure might cause noncancer respiratory health
effects. The synthesis text focuses on generalized summary findings regarding the identified
mechanistic events rather than observations in individual studies. Thus, individual study
references are not frequently cited in the text; these specific supporting references can be found in
the tables at the end of each tissue compartment-specific section (see Tables A-66-A-72).
This document is a draft for review purposes only and does not constitute Agency policy.
A-442 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
Because a large number of relevant articles (mostly experimental studies with multiple,
relevant endpoints) were considered in this analysis, a method was developed to distinguish the
experiments likely to provide the most useful information from those providing less informative
data or a comparably negligible amount of information. Individual mechanistic studies were
evaluated using basic screening-level criteria (see Table A-65) for each relevant endpoint or group
of related endpoints (e.g., hematological parameters) assessed by the study authors; thus, a study
may be evaluated multiple times. Expert judgment of the totality of the potential limitations was
used to determine a final level of confidence in the utility of the study results, with the reasoning
documented. In some instances, notation is included regarding the sensitivity of the methods and
whether they can provide information with direct relevance to interpreting cellular, structural, or
functional changes related to potential respiratory system health effects. Although this information
was not used in study evaluations, it was considered when developing the synthesis.
The study evaluation decision criteria were different for observational epidemiology
studies and experimental studies, although both sets of criteria emphasized exposure-related
considerations. As such, Tables 1-66 to 1-72 are first organized according to mechanistic effect
type, and then within each effect type into observational and controlled exposure studies. The
intent of the criteria applied, and the purpose of this mechanistic evaluation, was to focus on
potential mechanisms associated with constant, chronic inhalation exposure to formaldehyde.
Some studies of other effects that might be related to respiratory health effects have been evaluated
in other sections of the Appendix and support evaluations of potential respiratory hazards; these
evaluations informed the interpretation of overlapping studies presented in this section, as well as
in the MOA analyses presented in the toxicological review. Studies of cellular proliferation,
mucociliary function, and genotoxicity were separately reviewed, with the relevant conclusions
directly incorporated into the MOA analyses described in the Toxicological Review. The application
of the decision criteria presented in Table A-65 to the identified mechanistic studies is presented.
Interpretations of the usefulness of the individual mechanistic studies for evaluating the effect(s) in
question were drawn based on the results of applying the decision criteria. These interpretations
were high or medium confidence—experiments considered very useful for describing potential
formaldehyde inhalation-induced effects (since both medium and high confidence studies were
considered well conducted, additional criteria were not applied to distinguish one from the other).
In contrast, low confidence experiments might provide useful information, but should be considered
in the context of other available data. Not informative studies were interpreted as providing
negligible information regarding the potential for formaldehyde inhalation to cause the effect(s) of
interest and were ultimately not included in the mechanistic analyses, given the identified
limitations and the large number of available studies. Note that studies evaluating tissues
interpreted as unlikely to be contributing to respiratory health effects (e.g., liver) are included in
This document is a draft for review purposes only and does not constitute Agency policy.
A-443 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 the Appendix Tables below, but are not included in the MOA analyses presented in the Toxicological
2 Review or the systematic evidence map; the relative importance and ultimate decision to not
3 include such information in the mechanistic analyses may change if the conclusion regarding their
4 lack of relevance to respiratory health effects were to change with additional, future research.
This document is a draft for review purposes only and does not constitute Agency policy.
A-444 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-65. Decision criteria for the evaluation of mechanistic studies relevant to potential noncancer respiratory
effects
Observational studies preferences
Experimental studies (human or animal, controlled exposure) preferences
Generally, (not strictly scored) studies were considered low
Generally, (not strictly scored) studies were considered low confidence if they
confidence if they had multiple (2) unmet preferences and not
had multiple (2-3) unmet preferences and not informative if the majority of
informative if the majority of preferences were not met:
preferences were not met:
Exposure duration
System
duration >5 days (acute exposures noted)
in vivo with nose-only or whole-body inhalation exposure
daily exposures of several hours
Exposure levels
Test article
inhaled concentration accurately quantified in exposed
explicit use of paraformaldehyde (PFA) or methanol-free preparations of
group
formaldehyde; note: experiments of non-URT tissues/models (including
use of an appropriate referent group
lung) were automatically "low confidence" if this preference was not met)
exposure contrast expected to allow for detection of
differences across groups
Comparability
Exposure paradigm
endpoint result comparisons can discern effects of
duration of >5 days (acute exposures noted)
formaldehyde exposure alone (e.g., controlling for co-
periodicity of >5 hours/day and >5 days/week (if >1 day)
exposures, blinding)
Sample size
Exposure levels
>10 persons/ group to (theoretically) reduce variability
inhaled concentration was quantified (as ppm, mg/L or mg/m3)
at least one tested exposure level of <3 mg/m
(Note: studies only testing above 10 mg/m were considered "excessive")
Reporting
Comparability
clear description of methods
endpoint result comparisons can discern effects of formaldehyde
detailed, quantitative reporting of results
exposure alone (e.g., controlling for other experimental manipulations,
including chamber air exposure).
Sample size
>10 humans or >5 animals/ group to (theoretically) reduce variability
Reporting
clear description of methods
detailed, quantitative reporting of results
*
This document is a draft for review purposes only and does not constitute Agency policy.
A-445 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Evaluation of Individual Mechanistic Studies for Use in Describing Potential MOAs for Respiratory Effects
Important notes on Tables A-66 to A-72: Based on the assumption that most labs used commercially available formalin for convenience, the test
article is assumed to be formalin (and is documented as such) if the test article was not reported; in some cases, multiple endpoints evaluated in
the same row were interpreted as being informative to differing degrees; some specific, more apical endpoints described in the previous hazard
sections are excluded from these tables; N/R= not reported; FA= formaldehyde). Studies on the implications of altered endogenous
formaldehyde levels are not extracted into the tables below, although there may be some contextual discussion (e.g., to inform potential
susceptibility) in the Toxicological Review.
Table A-66. URT-specific structural modification, sensory nerve-related changes, or immune and inflammation-
related changes
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Observational Epidemioloav Studies
(Lvapina et
al.. 2004)
Symptomatic and
nonsymptomatic
human workers
with carbamide-
FAglue (n=29)
Exposed workers: 0.87±
0.39 mg/m3 (n=21
nonexposed); duration
mean: 12.7± 9.6 years
Assessment of
chronic URT
inflammation
Statistically significant increase in
subjective symptoms and objective
clinical findings of chronic, URT
inflammation (e.g.,
hypertrophy/atrophy of mucus
membranes; rhinitis) and decreased
neutrophil function (but N/C in
leukocyte cell counts) in workers;
symptomatic workers exhibited
decreased resistance to infections
(increased frequency, duration)
High or Medium Confidence
[mixture exposure]
(Bono et al.,
2016)
Human plastic
laminate workers
(n=50) and office
personnel
controls (n=45);
males only
Controls (mean±SE and
range): 0.035±0.0034
(0.016-0.11) mg/m3;
Workers: 0.211±0.015
(0.049-0.444); duration
unclear
Nasal epithelial ROS
(MidG adducts; a
marker of oxidative
stress and lipid
peroxidation)
Increased adducts with increasing
formaldehyde exposure (p trend=
0.002), with statistically significant
increases at > 0.066 mg/m3 (i.e.,
<0.025 mg/m3 = 47.6; 0.025-0.066
mg/m3 = 59.2; and >0.066 mg/m3 =
105.5 adducts)
High or Medium Confidence
[unknown duration]
(Holmstrom
and
Wilhelmsso
Two exposed
groups (n= 170
total; =90%
male); 70
formaldehyde
Exposed workers:
chemical plant: 0.05-0.5
mg/m3, mean 0.26 [SD
0.17 mg/m3]. Furniture
factory: 0.2-0.3 mg/m3,
Symptoms of URT
inflammation
Histopathology
scores
Symptoms of nasal obstruction and
nasal watery discharge more frequent
in exposed (p <0.05). When divided
into subgroups based on exposure
time, there were no signs of increasing
Low Confidence [Inclusion of
only current workers and long
duration of employment raises
possibility of healthy worker
survival effect due to irritation
n. 1988)
This document is a draft for review purposes only and does not constitute Agency policy.
A-446 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(note:
mucociliary
function data
below)
production
workers; 100
workers exposed
to wood dust and
formaldehyde at
five furniture
factories;
Referent: (n=36;
=55% male) from
government, with
no history of
formaldehyde or
wood dust
exposure
mean 0.25 [SD 0.05
mg/m3]. Referent mean
0.09 mg/m3 (based on 4
measurements in 4
seasons); duration of
employment >10 years
nasal restrictivity after employment >5
years.
Formaldehyde-only nasal specimens
mean histological score: 2.16 (range
0-4) (p <0.05) compared to referent
group 1.56 (range 0-4); while
formaldehyde-dust group had mean
score 2.07 (range 0-6) (p >0.05).
No correlation observed between
smoking habits and biopsy score, nor
was a correlation found between the
duration of exposure and any
histological changes.
effects; referent group not well
matched (different type of work
activity; undersampled males);
crude measures of effect
(Norback et
al.. 2000)
Primary school
personnel in
Sweden (n=234)
0.003-0.016 (mean=
0.0095) mg/m3; duration
unclear (working at least
20 h/wk; assumed length
months or more)
Assessment of
acoustic rhinometry
and factors in nasal
lavage
Formaldehyde was significantly
associated with multiple measures of
nasal obstruction
Formaldehyde was positively
associated with biomarkers for
eosinophils (eosinophil cationic
protein; lysozyme); N/C in a neutrophil
marker (myeloperoxidase) or albumin
Low Confidence [mixture
exposure (formaldehyde was
independently associated with
these changes, but so were N02
and Aspergillis)-did not
evaluate confounding; some
school measures below the
limit of detection]
(Priha et al..
2004)
Human MDF
board workers
(n=22) versus
wood dust (n=23)
and nonexposed
(n=15)
0.19± 0.11 mg/m3 (MDF
board) versus 0.11± 0.08
mg/m3 (note: VOCs 3-
fold higher in MDF than
wood); pre- and post-8-
hr workshift
Nasal lavage cell and
cytokine counts
N/C in cell counts
Increased postshift total protein vs.
unexposed controls
Increased post- vs. preshift NO (nitrite)
in wood and MDF workers
Decreased post- vs. preshift TNFa in
wood workers
Low Confidence [short duration;
minimal exposure differential;
role of VOCs not accounted for]
NOTE: ACUTE (8 hr; cross-shift)
Controlled-Exposure Studies in Humans or Primary Human Cells
(Pazdrak et
al.. 1993)
Human
occupationally-
exposed (n=10
males and
females) with
Formalin (assumed: test
article NR): 0.5 mg/m3
for 2 hr with follow-up
out to 16-18hr
Nasal lavage cell and
protein counts
Note: changes were
associated with
scoring measures of
Increased number of eosinophils,
albumin, and total protein; N/C
basophils
Increased proportion of eosinophils
and decreased proportion of epithelial
Low Confidence [formalin; short
duration; somewhat small
sample size; lack of investigator
blinding (nonissue for
automated albumin measures)]
This document is a draft for review purposes only and does not constitute Agency policy.
A-447 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
positive reaction
to FA: "allergic";
11 "nonallergic"
control males
nasal symptoms (e.g.,
sneezing; edema)
cells; N/C in proportion of basophils,
neutrophils, or mononuclear cells (i.e.,
lymphocytes and monocytes)
Effects max 10 min after exposure and
declining, but still significant, at
16-18hr; effects observed regardless
of "allergy"
NOTE: ACUTE; authors noted
albumin changes may indicate
increased mucosal permeability:
albumin percentage, also called
the "permeability index," was
elevated at 10 min postexposure
only
(Krakowiak
etal.. 1998)
Human workers
with bronchial
asthma or
healthy subjects
(n=10 each)
Formalin (assumed: test
article NR): 0.5 mg/m3
for 2 hr with follow-up
out to 24 hr
Nasal lavage cell and
protein counts
Note: changes were
associated with
scoring measures of
nasal symptoms (e.g.,
sneezing; edema)
Increased eosinophils, leukocytes,
total cell counts, and permeability
index at 30 minutes after exposure,
but not at 4 hr or 24hr after exposure;
N/C in basophils
(changes were observed regardless of
asthmatic designation)
N/C in mast cell tryptase or eosinophil
cationic protein
Low Confidence [formalin; short
duration; small sample size; lack
of investigator blinding
(nonissue for automated
albumin measures)]
NOTE: ACUTE; albumin
percentage, aka "permeability
index" was used to indicate
mucosal permeability; no effect
on FEVi, etc.
(Falk et al.,
1994)
Human sympto-
matic for nasal
distress (n=7) or
controls (n=6)
Formalin (assumed from
description of test
article)
Symptomatic: 0.021,
0.028,0.073, 0.174
mg/m3; <2 hr
Healthy: 0.023, 0.29,
0.067,0.127 mg/m3; <2
hr
Nasal mucosa
swelling by
rhinostereometry
FA increased mucosal swelling at
>0.073 mg/m3 in symptomatic
persons, but swelling was unchanged
in healthy controls
Low Confidence [formalin; short
duration; small sample size]
NOTE: ACUTE; assay is relevant
to inflammation, but limited in
scope and exposure contrast
(He et al.,
2005)
Human student
volunteers (n=10)
Ocular exposure to
wood-panel generated
formaldehyde gas 0,1, 2,
or 3 mg/m3; 5 min/d for
4d
Nasal lavage
substance P
Substance P was increased
significantly at 3 mg/m3
Low Confidence [exposure
route- unknown relevance of
ocular exposure route to
inhaled exposure level, but
considered to be reasonable
due to similarities in access of
gas to trigeminal nerve endings
for this endpoint; short
duration and periodicity;
somewhat small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-448 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Bardet et
al.. 2014)
In vitro (human
primary nasal
cells); n=5
experiments
(cells: one donor)
Formalin gas: 0.2 mg/m3
for lhr/day for 1, 2, or 3
days
Nasal cell cytokine
secretion
(at 72 hrs for all
exposures)
Slight, statistically significant,
decreased IL-8 with 3 exposures only;
N/Cin IL-6
Not Informative [in vitro;
formalin; short duration; small
sample size; comparable in vivo
inhaled exposure level
unknown]
Controlled-Exposure Studies in Animals. Animal Cells, or Immortalized Human Cells
(Fuiimaki et
al.. 2004b)
Female C3H mice
(n=5-6 per
group)
PFA 0, 0.098, 0.49, or
2.46 mg/m3; 12 wks
Serum cytokines and
neuropeptides (see
explanation at right)
D/D increased Substance P without
OVA (no change + OVA) at 2.46
mg/m3; FA decreased OVA-induced
NGF elevation at 0.098-0.49 mg/m3
(N/C with FA alone)
Body weight decreased at >0.49
mg/m3
High or Medium Confidence
[small sample size]
Note: although serum measure,
discussed in the context of
changes in the URT, so included
here
Sensitization: i.p. lOug OVA prior to FA
exposure; aerosol OVA boost for 6 min on wks
3, 6, 9, and 11
(Monticello
et al.. 1989)
Young adult male
rhesus monkeys
(n=3/group)
PFA 0 or 7.38 mg/m3 for
1 or 6 wk (6 hr/d, 5
d/wk)
Nasal histopathology
Goblet cell loss, hyperplasia and
neutrophil inflammatory response at 1
wk
High or Medium Confidence
[high exposure level]
Note: n=3 monkeys/group
considered a reasonable sample
(Andersen
et al.. 2010)
Male F344/CrlBR
rats (n=7-8)
PFA 0, 0.86,2.46,7.38,
12.3, or 18.5 mg/m3 for
1,4, or 13 wk (6 h/d, 5
d/wk)
Nasal histology
Nasal mRNA analyses
(Note: modeling
results not
considered)
mRNA changes: altered cellular
immune response at 1 wk at 12.3-18.5
mg/m3, with changes in DNA repair
and cell cycle at > 2.46 mg/m3; by 4
wk, immune/injury response is lost; by
13 wk, pervasive changes noted
High or Medium Confidence
Note: unclear, indirect
interpretability of mRNA
profiling
(Andersen
et al., 2008)
Male F344 rats
(n=8 for
histopath; n >5
for genomics)
PFA 0, 0.86, 2.46, or 7.38
mg/m3 for up to 3 wks (6
hr/d, 5 d/wk); also acute
(18.5 mg/m3) and
instillation
Nasal histopathology,
and microarray (high
flux regions)
Inflammatory cell infiltration was
observed at 7.38 mg/m3 at >l-d
exposure; microarray changes at >2.46
mg/m3 at 5d, but only at 7.38 mg/m3
at 15 d (1 gene at 2.46 mg/m3,1 d);
mostly stress-response related
High or Medium Confidence
NOTE: unclear, indirect
interpretability of genomic
endpoints; note: nasal
instillation caused more robust
changes
(Woutersen
et al., 1989)
Male Wistar rats
(n>20/ group)
PFA 0, 0.12,1.23, or 12.3
mg/m3 for 28 months (6
hr/d, 5 d/wk)
Nasal pathology
No treatment-related changes at
0.12-1.23 mg/m3; evidence of
damage, inflammation, proliferation at
12.3 mg/m3
High or Medium Confidence
(Rager et
al.. 2014)
Male Fischer rats
(n=3 biological
replicates/group)
PFA 0 or 2.46 mg/m3 for
7 d, 28 d or 28 d with 7d
recovery (6hr/d)
miRNA microarray of
nasal respiratory
epithelium
Nasal miRNAs were changed after 7 d
or 28 d (84 or 59 transcripts), not with
recovery; associated with
High or Medium Confidence
[very small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-449 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
inflammation and immunity, or tumor
suppression
NOTE: unclear, indirect
interpretability of endpoints
(Tsubone
and
Kawata,
Male Wistar rats
(n=6/group; each
rat received 2-4
exposures of PFA
or control air)
PFA 0.39-5.78 mg/m3
through upper airway for
22 seconds(under
anesthesia)
Ethmoidal nerve
activity (nasal
trigeminal nerve
branch)
Afferent nerve activity was increased
by FA, with a 50% increase in activity
at =2.2 mg/m3 (although FA stimulated
nerve activity at all levels- =20% at
0.62 mg/m3)
High or Medium Confidence
[short duration]
NOTE: ACUTE: surgical
procedures considered internally
controlled (since rats served as
own controls)
1991)
{Kulle, 1975,
39238}
Male SD rats
(n=5)
PFA 0.62,1.23, 1.85, or
2.46 mg/m3 for 1 hr or
0.62-3.08 mg/m3 for 25
sec (with anesthesia)
Nasopalantine nerve
responses (similar to
ethmoidal in
preliminary tests)
Sensory threshold from 25 s exposure:
0.31 mg/m3
Trigeminal response to an odorant
(amyl alcohol) is decreased at >0.62
mg/m3 FA
High or Medium Confidence
[slightly small sample size; short
duration]
NOTE: ACUTE; surgical
procedures internally controlled
(Yonemitsu
et al.. 2013)
TRPA1 knockout
(KO) or wild type
(WT) mice
(n=3-5)
Formalin at up to 123
mg/m3 (varied by
experiment and chamber
location, but all
exposures considered
"excessive"); ACUTE
Responses related to
effects on the
trigeminal nerve
Formalin vapor (3 min) activated
secondary trigeminal system neurons
(according to c-fos activity) in WT but
not KO mice.
Consistent with this, formalin vapor
accelerated wakefulness and induced
avoidance behaviors in WT but not KO
mice; and labeling studies confirmed
TRPA1 expression on trigeminal
afferents innervating the nasal mucosa
High or Medium Confidence
[small sample size; short
duration; formalin; excessive
levels; see below for
explanation]
NOTE: ACUTE; effects of related
chemicals such as acrolein were
similarly blocked in KO mice.
Given the difficult nature of
studying this event, the
consistency of effects across
related chemicals, and the well-
accepted role forTRPAl in
acrolein-induced sensory effects
(based largely on Bautista et al.,
2006), these results are judged
to provide indirect evidence
interpreted with high or medium
confidence and not direct
evidence interpreted with low
confidence.
This document is a draft for review purposes only and does not constitute Agency policy.
A-450 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Rager et
al.. 2013)
Male cynomolgus
macques
(n=2-3/group)
PFA 0,2.46, or 7.38
mg/m3 for 2 days (6
hr/d)
Nasal miRNA screen
and molecular target
verification
3 and 13 miRNAs were dysregulated
by exposure, including associations
with decreased apoptosis signaling (at
2) and increased epithelial
proliferation (at 6)
Low Confidence [short duration;
n=2 primates: small sample
size]
NOTE: Unclear direct relevance
of miRNA changes
(Clement et
al.. 1987)
Female Wistar
Rats (n=10)
PFA 0 or 18.5 mg/m3 for
12 weeks (6 hr/d, 5
d/wk)
URT epithelial
structure and
junctional proteins by
IHC and TEM
Basal lamina degeneration, and goblet
cell hypertrophy of respiratory
epithelium
FA reduced levels of junctional
proteins but did not cause destroy the
junctional complex when assessed by
TEM
Note: body weight significantly
decreased by FA (<5%)
Low Confidence [excessive
exposure levels]
(Cassee et
al.. 1996b)
Male Wistar
albino rats
(>3/group)
PFA 0, 1.23, 3.94, or 7.87
mg/m3 for 1 or 3 d
(6hr/d)
Nasal histopathology
and biochemistry
Evidence of damage and inflammation
at 3 d, >3.94 mg/m3
Increased GPx and NPSH (3 d, >3.94
mg/m3; latter at 1 d, 7.87 mg/m3 too),
not GST, FDH, ADH, or GR in
respiratory epithelium
Low Confidence [short duration;
very small sample size]
NOTE: ACUTE or 3d; NPSH:
nonprotein sulfhydryl groups
(Cassee and
Feron,
Male Wistar rats
(n=20/ group;
n=6+/endpoint)
PFA 4.43 mg/m3 for 3
days (intermittent)
Note: weights decreased
in all groups
Nasal enzyme activity
Nasal GSH
Increased GPx
N/C in ADH, GST, G6PDH, GR, or FDH
N/C in cytosolic GSH (slightly
increased)
Note: rhinitis and necrosis also
reported
Low Confidence [short duration
and unclear periodicity; high
exposure level]
1994b)
(Abreu et
al.. 2016)
C57BL/6 mice
(n=12 M+F/
treatment group
and n=6
M+F/control)
Formalin (assumed) 0,
0.25,1.2, and 3.7 mg/m3
for 8 h (aldehyde mixture
data not included herein;
authors noted some
exposure cross-
contamination)
Nasal epithelial
histology
(morphology only)
(blinded measures
6-8 h postexposure)
N/C in nasal epithelium, except small,
but significant, decreases in cilia at
0.25 mg/m3
Low Confidence
[formalin; short duration and
periodicity; some coexposure to
acetaldehyde possible but
unclear]
Note: ACUTE
(Monteiro-
Riviere and
Male F344 rats
(n=3 examined in
detail)
PFA 0, 0.62,2.46,7.38,
or 18.5 mg/m3 for up to
URT respiratory
epithelium ultra-
structural pathology
Inflammation (neutrophil infiltration;
goblet cell hypertrophy) at >7.38
mg/m3; duration-dependency shown
Low Confidence [short duration;
very small sample size; controls
not air exposed]
Popp, 1986)
This document is a draft for review purposes only and does not constitute Agency policy.
A-451 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
4 days (6 h/d); controls
not air-exposed
NOTE: no statistical comparisons
of structural changes
(McNamara
et al., 2007)
In vitro mouse
and rat dorsal
DRG neurons
(n=300+ neurons)
or HEK293 cells (n
> 5); (note:
relevance is as
URT stimulus)
Formalin or methanol
controls (levels irrelevant
to inhalation exposure);
ACUTE experiments
Activation and
specific inhibition of
"sensory nerve cell"
activity
Formalin, but not methanol,
specifically activated TRPA1 in vitro.
This specific activation was confirmed
using TRPA1 knockout DRG neurons as
well as specific pharmacologic
inhibitors. TRPA1 inhibition also
reduced formalin-induced pain
behaviors in vivo.
Low Confidence [in vitro;
unknown exposure level
relevance; short duration]
Note: ACUTE; methanol
controls; categorized as low
confidence rather than excluding
due to less concern for methanol
effects on receptors in nasal
mucosa
(Tani et al.,
1986)
Male rabbits
(strain
unspecified)
n= unclear
Formalin 12.3 mg/m3
(acute) directly infused
into either the URT
(nasal) and/ or LRT (lung)
Pharmacologic
intervention studies
on respiratory and
cardiac function
(compared to
acrolein and
ammonia)
The effects of formaldehyde on
respiration and heart rate were only
observed with nasal exposure, not
lung. Inhibition of afferent sensory
nerve activity abrogated the
formaldehyde effects.
Low Confidence [formalin; short
duration; unknown sample size]
NOTE: ACUTE; categorized as
low confidence rather than
excluding due to less concern for
methanol effects on receptors in
nasal mucosa
(Kunkler et
al.. 2011)
In vitro trigeminal
root ganglia (rat)
neurons (n=9-15)
Formalin (levels
irrelevant to inhalation
exposure); ACUTE
experiments
Agonist/antagonist
studies of TRP
channel-mediated
CGRP release
Formaldehyde stimulated release of
CGRP from adult trigeminal neurons
(Note: inhibitor studies not tested on
FA, but acrolein was through TRPA1)
Low Confidence [in vitro;
formalin; short duration; high,
unknown exposure level]
NOTE: ACUTE; categorized as
low confidence rather than
excluding due to less concern for
methanol effects on receptors in
nasal mucosa
(Zhao et
al.. 2020)
Male Balb/c
mice (n=3,
pooled into
single sample
for nose and
lung samples);
2 experiments
by different
researchers
Formalin
0, 3 mg/m3 for 2 weeks
(8 h/d, 5 d/wk)
Burst-forming unit-
erythroid (BFU-E),
and colony-forming
unit-granulocyte
macrophage (CFU-
GM) colonies in
nose, lung, spleen,
and bone marrow
Nose (ex vivo) results:
Decreased formation of BFU-E in
both experiment 1 and II
Decreased formation of CFU-GM in
experiment 1; N/C in experiment II
Nose (in vitro treatment):
400 uM formaldehyde significantly
decreased BFU-E not CFU-GM
formation (both nonsignificantly
decreased across doses)
Low Confidence [formalin;
small sample size; in vitro (for
cell treatments)]
This document is a draft for review purposes only and does not constitute Agency policy.
A-452 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Hester et
al.. 2003)
Male F344 rats;
n=3-4
Formalin (assumed,
based on description);
nasal instillation (400mM
in 40|_iL aliquot/nostril)
Respiratory
epithelium gene
expression
24 of 1,185 genes upregulated, and 22
downregulated
Not Informative [formalin;
short duration; very small
sample size; high, unknown
exposure level; exposure route]
NOTE: ACUTE
(Ohtsuka et
al.. 2003)
Male BN and
F344 rats;
n=4/group
Formalin aerosol 1% for
3 hr/d for 5 d vs. water
Nasal mucosa
cytokines and
structure
Degeneration and neutrophil
inflammation (F344> BN)
Decreased IFN-y and IL-2 in BN; N/C in
F344; N/C in IL-4 or IL-5 in BN or F344
Not Informative [formalin;
short periodicity; small sample
size; high, unknown exposure
levels]
(Macpherso
n et al.,
In vitro; n > 7;
transfected cells
(HEK293T cells
neuroendocrine;
immortalized
human kidney)
Formalin (levels
irrelevant to inhalation
exposure); ACUTE
experiments
Activation and
specific inhibition of
"sensory nerve cell"
activity
Formalin activated TRPA1. This
selective activation was confirmed by
inhibition of pain-related behaviors
induced by formalin in vivo.
Not Informative [in vitro;
formalin; short duration; high,
unknown exposure level;
limited reporting]
NOTE: ACUTE
2007)
This document is a draft for review purposes only and does not constitute Agency policy.
A-453 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-67. LRT (e.g., lung, trachea, BAL) markers of structural modification, immune response, inflammation, or
oxidative stress
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
Observational Epidemioloav Studies
(Franklin et
al.. 2000)
Human healthy
children (n= 224;
age =9.5 yr);
FA levels in bedroom and
living room were
dichotomized into > or <
0.062 mg/m3; duration
unknown
exhaled nitric oxide
(eNO); Note:
technique used
excludes NO
originating from the
upper airway
eNO ("reflects airway inflammation")
significantly increased in children of
homes with higher FA levels, after
correcting for multiple other variables
High or Medium Confidence
[limited exposure contrast;
accuracy of single measure
questionable]
Note: authors suggest species
differences in inflammation
locale
(Bentaveb
et al., 2015)
Human elderly
(>65 years)
European nursing
home individuals
(n=600 from 20
homes)
Indoor FA levels in main
common room ranged
from approximately
0.005-0.01 mg/m3
(median =0.006) over 1
week of sampling;
duration unknown
eNO (marker of lower
airway inflammation)
eCO (marker of CO
exhalation and
smoking)
FA was not associated with eNO
FA was associated with increased eCO
Note: FA was associated with
increased reported COPD and FVC, but
not FEV1, asthma diagnosis or
symptoms, or cough
High or Medium Confidence
[limited exposure contrast;
unclear whether adjusted for
co-exposures]
Note: PM co-exposure was not
associated with eNO or eCO;
N02 was associated with
decreased eNO
(Flamant-
Hulin et al.,
Human school
children (34
asthmatics; 70
nonasthmatics);
[Low] yards: 0.0036
(0.0024-0.0044) mg/m3
and rooms: 0.025
(0.013-0.036) mg/m3
[High] yards: 0.0058
(0.0049-0.0068) mg/m3
and rooms: 0.044
(0.038-0.047) mg/m3;
unknown duration
Fractional exhaled
nitric oxide (FeNO)—
"reliable, noninvasive
marker of airway
inflammation" [Note:
"nasal
contamination" was
prevented]
FeNO significantly increased in both
nonasthmatics and asthmatics with
high versus low FA exposure in
classrooms, but not schoolyards; in
nonasthmatics, a stronger association
was found for atopic versus nonatopic
children
High or Medium Confidence
[accuracy of single measure
questionable]
Note: authors hypothesized that
atopic status might modify
airway response to
formaldehyde; called changes
"bronchial inflammation"
2010)
(Roda et al.,
2011)
French infants
(n=2940 with
assessment at
birth and 12
months)
Median 0.020 mg/m3;
IQR 0.014-0.027 mg/m3;
LOD 0.008 mg/m3.
LRT infections (with
or without wheeze)
Note: although URT
infections were
queried, these data
were NR
Significantly increased LRT infection:
32% or 41% increase per 0.0124
mg/m3 increase in formaldehyde
(without and with wheeze,
respectively)
High or Medium Confidence
[specificity and sensitivity of
predictive model not tested on
a separate sample]
(Rumchev
et al., 2002)
Australian
children (ages 6-
Mean 0.030 and 0.028
and maximum 0.224 and
Lower respiratory
tract infection
Increased emergency room visits for
this case definition
Low Confidence [recruitment
process not described;
This document is a draft for review purposes only and does not constitute Agency policy.
A-454 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
36 months); 88
cases, 104
controls
0.190 mg/m3,
respectively, in bedroom
and living room.
involving wheezing
(assuming
misclassification of a
many of the
discharges as asthma
rather than infection)
uncertainty as to how well this
case definition describes LRT
infection and the length of time
between emergency room visit
and subsequent exposure
measure]
Controlled-Exposure Studies in Humans or Primary Human Cells
(Casset et
al.. 2006b)
Human (n=19
with mild asthma
and allergy to
mite allergen)
Formalin 0.1 mg/m3 for
30 minutes; placebo at
=0.03 mg/m3 double-
blind randomized;
restricted to mouth
breathing only
Sputum (lower airway
mucus) eosinophils
and ECP
Authors note a trend, not statistically
significant, towards increased
eosinophil counts (=38±9% vs. 11±3%,
FA vs. air controls), and an increase in
ECP (439± 171 vs. 156± 58 ng/l, FA vs.
air controls)
Low Confidence [formalin; short
duration; not clear that
restriction to mouth breathing
is realistic for typical inhalation]
NOTE: ACUTE; within-subjects
comparison between air and FA
(Ezrattv et
al.. 2007)
Human (n=12
intermittent
asthmatics with
allergy to pollen)
Formalin 0.5 mg/m3 for
60 min; randomized
allocation (no
nonexposed controls)
Sputum (lower airway
mucus) cell counts
and released factors
N/C in sputum Total cell counts, WBC
subtypes, or factors (e.g., ILs, MCP,
TNF)
Low Confidence [formalin; short
duration]
NOTE: all exposed to both air
and FA: internally controlled
Controlled-Exposure Studies in Animals, Animal Cells, or Immortalized Human Cells
(Fuiimaki et
al.. 2004b)
Female C3H mice
(n=5-6 per
group)
PFA 0, 0.098, 0.49,2.46
mg/m3; 12 wks
BAL cell counts
BAL cytokines and
neuropeptides
No significant changes in cell counts
with FA alone; macrophages and
eosinophils increased at 2.46 mg/m3
with OVA+FA; N/C in neutrophils or
lymphocytes
No significant changes in cytokines
with FA alone (NGF was D/D
increased)
FA with OVA D/D decreased IL-ip at
2.46 mg/m3 and NGF at 0.098-0.49
mg/m3; N/C in TNF-a, GM-CSF, or IL-6;
MCP-1, MlP-la, and eotaxin were not
detectable
Body weight decreased at >0.49
mg/m3
High or Medium Confidence
[small sample size for some
groups/end points]
Note: MlP-la, eotaxin, MCP-1,
BDNF, and Substance P levels
insufficient for testing
Sensitization: i.p. lOug OVA prior to FA
exposure; aerosol OVA boost for 6 min on wks
3, 6, 9, and 11
Formaldehyde (bottled
pressurized gas) 0, 0.13,
Airway histology and
morphometry
With FA, lung bronchi had intramural
edema (wall thickening) by
High or Medium Confidence
[small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-455 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Riedel et
al. 1996)
Female Dunkin-
Hartley guinea
pigs (n=3)
0.31 mg/m3 for 5 d (8
hr/d)
morphometry; no evidence of cellular
lower airway inflammation by
histology
Note: histology after FA with
OVA not examined
Sensitization: 0.5% inhaled OVA; OVA boost at
2wk
Challenge: 1% inhaled OVA lwk later
(Ito et al.,
1996)
Male Wistar rats
(n=7)
Formalin (with MeOH
controls) 2.46, 6.15,
18.5, or 55.4 mg/m3 for
10 min
Airway microvascular
leakage (Evans blue)
in trachea and main
bronchi
D/D increased leakage at >6.15
mg/m3, which resolved in <20 minutes
Leakage at 18.5 mg/m3 was inhibited
by NK1 receptor antagonism, but not
by hista-mine HI or bradykinin B2 R
antagonists
55.4 mg/m3 MeOH alone induced
slight leakage in main bronchi, but not
trachea)
High or Medium Confidence
[short duration]
Note: figure comparisons
presented against room air, not
MeOH, controls, but
comparisons made to MeOH
controls in text
(Jakab,
1992)
In vivo and Ex
vivo Female Swiss
mice (n=5+ mice/
determination)
PFA 0,0.62 1.23, 6.15,
12.3, or 18.5 mg/m3 for
4-18 hr or 4 d (4 hr/d); ±
carbon black
Pulmonary
bactericidal activity to
inhaled
Staphylococcus
And ex vivo alveolar
macrophage function
Pulmonary antibacterial activity was
reduced: at 1.23 mg/m3for 18 hr
before and 4 hr postbacterial
challenge (postexposure alone
reduced at 18.5 mg/m3)
N/C in ex vivo alveolar macrophage Fc
receptor-mediated phagocytosis of
RBCs at 6.15 mg/m3 for 4 d (FA +
carbon black, but not FA alone, caused
a robust decrease)
High or Medium Confidence
[short duration]—in vivo
pulmonary bactericidal activity
Note: ACUTE
Low Confidence [ex vivo; short
duration]
(Swiecicho
wski et al.,
Male Hartley
guinea pigs
(n=5-12/group)
PFA at 4.18 mg/m3 for 2
or 8 hours (multiple
experiments)
Airway Histology
(trachea)
No change histological evidence of cell
infiltration or epithelial damage up to
96 hr after exposure to 4.18 mg/m3 for
8 hr
High or Medium Confidence at
1.23 mg/m3 and above [short
duration]
Low Confidence below 1.23
mg/m3 and ex vivo [ex vivo;
sample size of 5 at 1 or more
levels below 1 ppm]
NOTE: ACUTE
1993)
(Ozen et al.,
2003)
Male albino
Wistar rats (n=6)
PFA at 6.15 and 12.3
mg/m3 for 4 or 13 weeks
(8 hr/d)
Lung tissue
homogenate
measures of trace
elements
Zn was dose-dependently decreased
(>6.15 mg/m3 for both exposure
durations;
High or Medium Confidence
[high levels]
NOTE: unclear relevance of
endpoints; authors claim Fe
This document is a draft for review purposes only and does not constitute Agency policy.
A-456 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
Fe was dose-dependently increased
(>6.15 mg/m3 with 13 wk; significant
only at 12.3 mg/m3 after 4 wk); Cu was
unchanged
change linked to oxidative stress
and Zn change linked to
decreased DNA synthesis, but no
direct evidence
(Avdin et
al.. 2014)
Male SD rats
(n=6/group)
Test article unclear, but
appears to be formalin in
this experiment at 0,
6.48 (low), 12.3
(moderate), or 18.7
mg/m3 for 4 wk (8 hr/d,
5 d/wk)
Lung tissue total
antioxidant and total
oxidant levels (TAS
and TOS; kit uses
vitamin E and H202 as
reference,
respectively
Lung tissue oxidative
stress index (OSI:
TOS/TAS) and
apoptotic index
Lung irisin (hormone
may regulate obesity)
Increased TOS and OSI, and decreased
TAS and irisin, at > 12.3 mg/m3
formaldehyde
Increased lung apoptotic index at
>6.48 mg/m3
Note: Carnosine supplementation
reduced changes.
Low Confidence [formalin; high
levels]
(Luo et al.,
2013)
In vitro and ex
vivo (intact
trachea) from SD
rats (sex NR); n=
as low as 4 (some
inhibitor assays),
as high as 28
(trachea)
Formalin (assumed, test
article NR; levels
irrelevant to inhalation
exposure); ACUTE (bath
application) experiments
Isc currents in
trachea and
epithelium from
trachea with various
inhibitors
TRPV channel
expression and
labeling
Formaldehyde caused a dose-
dependent, sustained increase in
currents in isolated trachea and airway
epithelia
TRPV-1 channels were localized to
intraepithelial nerve endings and
inhibition of TRPV-1 or substance P
activity (blocking NK-1R) inhibited
current increases
CI- released in response to
formaldehyde was blocked several CI
channel blockers and involed cAMP
Low Confidence [in vitro and ex
vivo (intact trachea); formalin;
unknown exposure level
relevance]
Note: ACUTE, some inhibition
experiments had n=4, but
magnitude of inhibition was
robust with small variabilty
(Lundberg
and Saria,
Male SD rats
(sample size NR)
Direct injection of
formaldehyde (assumed
to be formalin); 50|_iL
volume unknown
comparison to inhalation
exposure
Tracheal mucosal
reactivity (Evans blue
extravasation)
Formaldehyde injection caused
extravasation which was reduced or
abolished by capsaicin pretreatment
Low Confidence [formalin;
inferred high levels; short
duration; nonspecific reporting]
NOTE: ACUTE
1983)
This document is a draft for review purposes only and does not constitute Agency policy.
A-457 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Larsen et
al.. 2013)
Male BALB/cA
mice (n=10/
group)
PFA 0.49, 2.21, or 4.9-7.0
(dry vs. humid air)
mg/m3; 60 min
BAL counts
FA did not affect BAL "degree of lung
inflammation" (data not shown;
unclear if this reflects comparisons of
total cell counts or comparisons of
individual cell types, as data were
presented for OVA, i.e., neutrophils,
lymphocytes, eosinophils,
macrophages)
Low Confidence [short duration;
for BAL endpoints: poor
reporting: FA alone groups data
NR; OVA without FA and OVA
with FA groups combined]
NOTE: ACUTE
Sensitization: pre-FA i.p. lug OVA, with O.lug
OVA boosts i.p. on days 14 and 21 (note: FA on
day 31)
Challenge: 0.2% OVA aerosol for 20 min on Days
29 and 30
(Wu et al.,
2013)
Male Balb/c mice
(n=8/group)
Formalin 0 or 3 mg/m3
for 4 wk (6 h/d, 5 d/wk)
with or without OVA
aerosol
BALF cell counts
Lung tissue cytokines,
neuropeptides, and
histology/IHC
Total cells, eosinophils, and
lymphocytes were increased in BALF
by FA alone, and all of these cells
(minus lymphocytes but plus
neutrophils) were increased more
robustly by FA+ OVA
Histopathology: increased
inflammation
FA increased lung IL-4, IL-ip,
substance P, and CGRP, but not IFNy;
more robustly by FA+OVA (peptide
changes by IHC also)
TRPA1 and TRPV1 antagonists reduced
FA+OVA-induced eosinophil counts
(anti-TRPAl also decreased
neutrophils), and lung factors (except
IL-1)
Low Confidence [formalin;
pharmacological interventions
did not include effects of FA
alone]
Sensitization: s.c. 80 ng OVA on Days 10,18,
and 25
Challenge: 1% OVA aerosol 30min/d on Days
29-35
(Qiao et al.,
2009)
Male Wistar rats
(n=8/group)
Formalin 0, 0.51 or 3.08
mg/m3 for 3 wk (6 hr/d)
BALF cell counts
Lung histology and
cytokine levels
"slight but insignificant pulmonary
abnormalities" with FA alone; OVA
3.18 mg/m3 changed airway structure
N/C in BAL total cells or eosinophils
with 3.18 mg/m3, but >0.51 mg/m3
dose-dependently increased both in
presence of OVA; 3.18 mg/m3 FA alone
increased IFNy and decreased IL-4;
FA+OVA increased IL-4
Low Confidence [formalin]
Sensitization: i.p. OVA on Days 10 and 18
Challenge: 1% OVA 30 min/d for 7 d
Male Balb/c mice
(n=6/ group)
Formalin 0, 0.5, or 3
mg/m3 for 21 d (6 hr/d)
BALF cell counts
Cell infiltration and airway remodeling
in 3 mg/m3 FA+ OVA
Low Confidence [formalin]
This document is a draft for review purposes only and does not constitute Agency policy.
A-458 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Liu et al.,
2011)
Pulmonary histology
and cytokines
Increased % Eosinophils at > 0.5
mg/m3, which is amplified by OVA; N/C
IFNy
Increased lung IL-4 and IL-6 at 3
mg/m3; with OVA, this is observed at
0.5 mg/m3
Sensitization: i.v. 20 mg OVA on Day 10 and 1
Challenge: 1% OVA aerosol for 30 min/d for 7 d
(Ye et al.,
2013b)
Male Balb/c mice
(n>9/ group/
endpoint)
Formalin 0, 0.5,1, or 3
mg/m3 for 7 d (8 hr/d)
ROS (dichlorohydro-
flourescein and MDA)
and GSH in Lung
Dose-dependent decrease in GSH
levels in lung at >0.5 mg/m3
Dose-dependent increase in DCFH and
MDA in lung at >1 mg/m3
Co-administered GSH attenuated
effects
Low Confidence [formalin]
(Abreu et
al.. 2016)
C57BL/6 mice
(n=12 M+F/
treatment group
and n=6 M+F/
control)
Formalin (assumed) 0,
0.25,1.2, and 3.7 mg/m3
for 8 h (aldehyde mixture
data not included herein;
authors noted some
exposure cross-
contamination)
Lung histology (cells
and morphology)
(blinded measures
6-8 h postexposure)
Lung cytokine,
catalase, and SOD
levels/ activity
FA increased distended alveoli at 3.7
mg/m3; N/C in total mononuclear or
polymorphonuclear cells
N/C in IL-1, IL-6, TNF, CCL2, or MIP-2,
or in antioxidants; increased
keratinocyte chemoattractant at 0.25
mg/m3 only
Note: N/C in lung mechanics except
increased airway inertance (might
indicate an impedence of airflow) at
3.7 mg/m3
Low Confidence
[formalin; short duration and
periodicity; some coexposure to
acetaldehyde possible- unclear]
Note: ACUTE
(Sandikci et
al.. 2007b)
SD rats (n=6/
group) at GDI [1],
PND1 [II], PND28
[III] or adults [IV]
Formalin (assumed: test
article NR): 0 or 7.38
mg/m3 for 6 weeks (8
hr/d, 7 d/wk)
BALTT lymphocyte
CD4+, CD8+ counts
(by IHC)
Increased BALTT lymphocytes (ANAE+
as marker); CD4+ T cell counts and size
of BALT increased in Groups III and IV;
CD8+ T cell counts increased in Group
III
Note: body weight was significantly
decreased in Groups 1 and II
Low Confidence [formalin; high
exposure levels]
Note: limited assays
(Sandikci et
al.. 2007a)
Female SD rats at
GDI [i], PND1 [ii],
PND28 [iii], or
PND90 [iv] (n=6)
Formalin (assumed; test
article NR) 0, 7.38 mg/m3
for 6 weeks (8 hr/d, 7
d/wk)
BALTT lymphocyte
counts; BALT size
Note: body weight
decreased by FA in
groups i and ii
CD4+ cell counts increased in groups iii
and iv; CD8+ cell counts increased in
group iii (group iv N/S increased)
Increased size of BALT in adults (iii &
iv)
Low Confidence [formalin; high
exposure levels]
This document is a draft for review purposes only and does not constitute Agency policy.
A-459 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Jung et al.,
2007)
Female C57BL/6
mice (n=10/
group)
Formalin (assumed; test
article NR) 0,6.15,12.3
mg/m3 for 2 wk (6 hr/d,
5 d/wk)
Lung oxidative stress
(intracellular, by flow)
BAL and lung
homogenate counts,
and histopath.
Cytokine mRNA and
protein
Oxidative stress (DCFH-DA) at >6.15
mg/m3
Total BAL cells increased (2-fold) at
12.3 mg/m3; Slight changes in B220+ B
cells (4,) and CD3+ or CD4+ T cells (1^)
were not interpreted as significant;
CD8+ T cells were only slightly; N/C
in neutrophils
Large increase in eosinophil counts
from BAL, and in flow counts and gene
expression of lung tissue at 12.3
mg/m3, eosinophil infiltration, and
epithelial damage, by histopath at
>6.15 mg/m3
Increased IL-4, IL-5, and IL-ip (not IL-
13) in lung at 6.15 and 12.3 mg/m3
body weights decreased =10%
Low Confidence [formalin; high
exposure levels; statistical
significance of flow data NR]
Note: Th2 cytokines
(Sul et al.,
2007)
Male SD rats
(n=10/group)
Formalin (assumed; test
article NR) 0,6.15,12.3
mg/m3 for 2 weeks
Lung tissue oxidative
stress and mRNA
array
Lipid peroxidation (MDA) and protein
oxidation were increased at 12.3
mg/m3
Changes in 21 genes, including D/D
decrease in 3 immune-related genes:
HSP70ia, complement 4 binding
protein, and Fc receptor IgG low
affinity III
Low Confidence [formalin; high
levels]
NOTE: utility of mRNA results by
themselves unclear
(Lu et al.,
2005)
Male Kun Ming
mice (n=5)
Formalin 0, 0.5,1, or 3
mg/m3 for 10 d (6 h/d)
BALF IL-4
(undetected in
serum)
D/D Increased IL-4 at >1 mg/m3 FA
Blocked by vanilloid (TRPV) receptor
antagonist, CPZ
Low Confidence [formalin; small
sample size]
(Ahn et al.,
2010)
Male SD rats
(n=4/group)
Formalin (assumed; test
article NR) 0, 2.46, or
24.6 mg/m3 for 2 wk (6
h/d)
BAL fluid proteomic
analysis
6 proteins increased (3 inflammatory
serpins, anti-inflammatory annexin, an
erythrocyte protein associated with
trauma or inflammation, and a
metabolic enzyme); 5 proteins were
decreased
Low Confidence [formalin]
NOTE: unclear utility of
measures
This document is a draft for review purposes only and does not constitute Agency policy.
A-460 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Kimura et
al.. 2010)
Male Wistar
(n=5-6)
Formalin 1.23, 6.15,
18.5, or 55.4 mg/m3 for
up to 45 min
Airway microvascular
leakage (lung- main
bronchi and trachea)
BALF counts of
leukocytes
Shed epithelial cells
in BALF
D/D increase leakage by 15 min at >
1.23 mg/m3; not exacerbated with
longer/ repeated exposure
Note: Leakage induced by substance P
was not inhibited by pre-FA exposure,
but preinhalation of the same mg/m3
abolished FA-induced leakage and pre-
FA inhibited capsaicin-induced
leakage; however, 20 hr between
exposures allows for recovery of
tachykinins and leakage by FA
exposure
Inhibition of mast cell activation (HI
receptor antagonist), but not
cyclooxygenase products
(indomethacin), blocked FA leakage at
6.15 mg/m3;
increased shed epithelial cells 20 h,
but not immediately, after 6.15 mg/m3
for 30min
Increased BALF neutrophils with
preinhalation at 6.15 mg/m3, but N/C
eosinophils or mononuclear cells
Low Confidence [formalin; small
sample size; short duration]
Note: Authors hypothesize
preinhalation of FA depletes the
amount of tachykinins available
at the target site (but not
desensitization of NK1
receptors), in part b/c capsaicin
can no longer induce a response;
also, because of recovery, up to
6.15 mg/m3 does not cause
irreversible damage to airway
sensory nerves, but that
prolonged exposure (>7 d) might
exacerbate neurogenic airway
inflammation
(Dallas et
al.. 1987)
Male SD rats
(n=2/timepoint;
unclear
reporting)
PFA 0, 0.62, 3.69, or 18.5
mg/m3 for 1 wk to 24 wk
(6h/d, 5d/wk)
Flow cytometry
DNA/RNA analysis of
alveolar cell
proliferation/ health
Increased RNA index in alveolar cells at
all FA levels at 1 wk; only at > 3.69
mg/m3 at 8 wk; N/C in DNA (e.g., % S
phase)
[Note: same alveolar samples had
chromatid breaks at 18.5 mg/m3]
Low Confidence [small sample
size; unclear reporting]
NOTE: unclear specificity/ utility
of methods
(Kim et al.,
2013a)
Female C57BL/6
mice (n=5
"experiments";
number of mice/
group unclear)
Formalin (assumed; test
article NR) 0, 6.15, or
12.3 mg/m3 for 2-3 wk
(6 hr/d, 5 d/wk)
Lung cell counts
BAL cell counts
Ex vivo cellular
functional assays
N/C in lung tissue total cells, but
number of NK1 cells markedly
decreased (this recovered by 2 wks
postexposure) at 12.3 mg/m3
Lung NK1 cell mRNA and protein
markers (IFNy, perforin, and CD122)
were D/D decreased at > 6.15 mg/m3
Low Confidence [formalin; high
levels; small sample size]
Not Informative: ex vivo
experiments or in vitro FA
treatment of NK precursors
showing reduced differentiation
to mature cells
This document is a draft for review purposes only and does not constitute Agency policy.
A-461 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
BAL total cells increased, but number
of NK cells decreased at 12.3 mg/m3
N/C in other lung or BAL lymphocyte
populations (e.g., % CD4+ or CD8+
cells)
(Sadakane
et al., 2002)
Male ICR mice
(n=9 or 18)
Formalin 0.5% for 4 wk
(15 min/ wk)
Lung IHC cell counts
and cytokine analysis
N/C in lung eosinophil recruitment or
goblet cell proliferation by FA alone,
but Derf-induced eosinophil
recruitment was exacerbated by FA
Increased RANTES in lung by FA alone,
and exacerbated increase to Der f-
changes with FA for IL-5 and RANTES;
N/C in lung IL-2 or IL-4
Low Confidence [formalin;
unquantified high levels; short
periodicity]
Sensitization: i.p. with 3 mg/mL Der f (house
dust mite allergen) prior to FA
Challenge: intratracheal 10 ng Der f 3 hr after
last exposure (note: measures 3 d later)
(Sandikci et
al.. 2007a)
Female SD rats at
PND1, PND28, or
PND90 (n=3)
Formalin (assumed; test
article NR) 0 or 7.38
mg/m3 for 6 wk (8 hr/d,
7 d/wk)
Lung and BALT
histology
N/C in exposed PND1 group
Increased apoptotic cells in lungs and
BALT of PND28 and PND90 groups
Authors: apop. cells likely lymphocytes
Low Confidence [formalin; high
level; small sample size]
(Matsuoka
et al.. 2010)
Male ICR mice (n>
7)
Formalin at 0.12 mg/m3
for up to 24 hr; also, a
single experiment at 3.69
mg/m3 for 24 hr
lung ROS (80HdG)
and NO metabolites
(nitrates/ nitrites); at
3.69 mg/m3: LPS
response
Decreased ROS lung; N/C in NOs or
lung NOs after LPS injection
Low Confidence [formalin; short
duration]
NOTE: ACUTE
(Yan et al.,
2005)
Male Kun Ming
mice (n=6)
Mixture (test article
wood panels) 0, 0.5,1, or
3 mg/m3 for 72 hr (24
hr/d)
Lung NOS activity and
NO measurement
Increased NOS activity at 3 mg/m3 FA
(p = 0.06 at 1 mg/m3)
NO was detected more frequently in
samples from 3 mg/m3 FA group (50%
vs. 17%)
Low Confidence [wood panel
exposure; lack of controls for
co-exposure; short duration]
NOTE: NO detection did not
include statistical comparisons
(Dinsdale et
al.. 1993)
Male SD rats
(n=4,6, or 10)
PFA or Formalin 12.3
mg/m3 for 4 d (6 hr/d)
Lung enzymes (in BAL
or tissue)
Lung histology
Increased cytochrome P450 and
decreased y-glutamyl transpeptidase
with PFA exposure (not with formalin)
No abnormalities (i.e., signs of injury
or repair) by histology
Low Confidence [small sample
size; excessively high levels;
short duration] NOTE: Endpoints
not very informative for
inflammation (injury response,
possibly)
(Rager et
al.. 2011)
In vitro (human
lung cancer cell
PFA 1.23 mg/m3 for 4 hr
or air controls
In vitro epithelial cell
miRNA microarray
and IL-8 secretion
Increased IL-8 release >16-fold with FA
Low Confidence [in vitro; short
duration; exposure level
This document is a draft for review purposes only and does not constitute Agency policy.
A-462 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
line); n=6
replicates
89 miRNAs were downregulated by FA;
the 4 most robust were associated
with inflammatory response pathways
comparability to inhalation
unclear]
(Zhao et
al.. 2020)
Male Balb/c
mice (n=3,
pooled into
single sample
for nose and
lung samples);
2 experiments
by different
researchers
Formalin
0, 3 mg/m3 for 2 weeks
(8 h/d, 5 d/wk)
Burst-forming unit-
erythroid (BFU-E),
and colony-forming
unit-granulocyte
macrophage (CFU-
GM) colonies in
nose, lung, spleen,
and bone marrow
Lung (ex vivo) results:
Decreased formation of BFU-E in
experiment II; N/C in experiment 1
Decreased formation of CFU-GM in
experiment II; N/C in experiment 1
Lung (in vitro treatment):
Up to 400 uM formaldehyde caused
N/C in BFU-E not CFU-GM formation
Low Confidence [formalin;
small sample size; in vitro (for
cell treatments)]
(Maiellaro
et al.. 2014)
Pregnant Wistar
rats (n=5; note:
individual pup
data for n=10
pups did not
appear to
account for
litters)
Formalin 0.92 mg/m3
from GD1-GD21: 1 hr/d,
5 d/wk
BAL cell counts and
factors
Lung factors
N/C in parental BAL total cells,
monocytes, lymphocytes, or
granulocytes
N/C in parental lung IL-4, IL-6 or IL-10;
Decreased birth weight in offspring
24 hr after OVA challenge, offspring
have: decreased BAL total cells,
mononuclear cells, neutrophils, and
eosinophils; Increased BAL IL-10, but
decreased IL-6 and TNFa (N/C in IL-4)
Not Informative [formalin, short
periodicity; small sample size;
offspring comparisons do not
include FA alone; did not
appear to account for litter
effects]
Sensitization: s.c. 10 ng OVA with sc boost after
7d
Challenge: 7 d later, 1% OVA aerosol 15 min/d,
3d
(Maiellaro
et al.. 2016)
Pregnant Wistar
rats (n=5 dams;
note: individual
pup data for n=10
pups did not
appear to
account for
litters)
Formalin 6.13 mg/m3
from GD1-GD21: 1 hr/d,
5 d/wk
BAL cell counts and
factors in pups on
=PND45
Increased (amplified) total BAL
leukocytes
Increased (amplified) BAL
mononuclear cells and neutrophils
Increased (amplified) myeloperoxidase
Decreased (slightly reduced)
eosinophils and eosinophil peroxidase
Not Informative [formalin, short
periodicity; small sample size;
offspring comparisons do not
include FA alone; did not
appear to account for litter
effects]
Sensitization on PND 30: s.c. 10 ng OVA
Challenge: 14d later, 1% OVA aerosol 15 min/d,
3d
(Silva
Ibrahim et
al.. 2015)
Pregnant Wistar
rats (n=5 dams;
10 pups/group
for experiments;
Formalin 0.92 mg/m3
from GDs 1-21: 1 hr/d, 5
d/wk
Cell number, cytokine
and neutrophil
marker (MPO) in BAL
Function of BAL cells
24hr after LPS challenge, offspring
exposed to formaldehyde have
reduced immune responses to LPS (i.e.
decreased BAL cells and granulocytes-
Not Informative [formalin;
short periodicity; offspring
comparisons do not include FA
without LPS; small sample size;
This document is a draft for review purposes only and does not constitute Agency policy.
A-463 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
note: individual
Lung gene and
N/C in lymphocytes or monocytes;
did not appear to account for
pup data for n=10
proteins
decreased MPO and oxidative burst-
litter effects]
pups did not
N/C in phagocytosis; decreased IL-6
appear to
Randomly assigned pups all received 5 mg/kg
and increased IFN and IL-10;
account for
lipopolysacharride (LPS) injections at PND 30
decreased TLR4 and NFkB)
litters)
(Ibrahim et
Pregnant Wistar
Formalin 0.92 mg/m3
Total BALcell number
Increased cell number by LPS was
Not Informative [formalin;
al.. 2016)
rats (n=5 dams;
from GDs 1-21: 1 hr/d, 5
and cytokine gene
reduced in offspring exposed to
short periodicity; offspring
10 pups/ group
for experiments;
d/wk
expression
formaldehyde
Formaldehyde increased IFN
comparisons do not include FA
without LPS; small sample size;
note: individual
Randomly assigned pups all received 5mg/kg
expression, decreased IL-6, TLR4, and
did not appear to account for
pup data for n=10
lipopolysacharride (LPS) injections at PND 30
NF-kB expression, and caused N/C in
litter effects]
pups did not
IL-10, as compared to LPS
appear to
Note: effects rescued by
account for
vitamin C
litters)
(da Silva et
Male Wistar rats
Formalin 1% for 3 days
BAL cell counts
FA increased total BAL cells, activated
Not Informative [formalin;
al.. 2015)
(n=6/ group)
(90 min/ d); rats exposed
Lung vascular
mast cells, and neutrophils (latter
unquantified high levels; static
in static chambers 5 rats/
permeability
based on myeloperoxidase activity)
exposure chamber and group
time
BAL and lung
cytokines
(all measures at 24 h
postexposure except
permeability, which
was immediate)
FA did not change trachea
permeability (Evans blue), but did
increase it in lung parenchyma and
bronchii
FA increased TNF, IL_6, and N/C IL-10
in BAL, and increased IL-10, but not IL-
6 mRNA in lung tissue
Note: while reduced effects were
reported as reduced with laser
therapy, laser therapy-only controls
were not used
exposure; short duration and
periodicity]
(Murta et
Male Fischer rats
Formalin (assumed) 1%,
BAL cell counts
FA increased total leukocyte,
Not Informative [formalin;
al.. 2016)
(n=7)
5%, or 10% for 5 d (3 x
Lung histopathology
macrophages at 10%, and lymphocytes
unquantified high levels; static
20 min/d)
and chemokine levels
at >5%; N/C in neutrophils or
eosinophils; >5% caused lung
parenchyma damage; >1% increased
CCL5 and 10% CCL2 (N/C in CCL3)
exposure chamber; short
periodicity]
This document is a draft for review purposes only and does not constitute Agency policy.
A-464 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Kilburn and
Mckenzie,
Male and female
Syrian golden
hamster (n=6-14)
PFA "low": 3.69 or 7.38
mg/m3 or "high": >246
mg/m3 for 4 hr; alone,
with carbon dust, or
evaporated onto carbon
Lower airway PMN
Leukocyte
recruitment and
cellular changes by
histology
Although cytotoxic effects were
observed at >3.69 mg/m3, FA alone did
not induce PMN leukocyte
recruitment; FA + carbon caused
leukocyte recruitment 2hr
postexposure, which peaked at =20 hr
and resolved by 1 wk; recruitment was
similar at "low" and "high" levels
Not Informative [short
duration, precision of exposure
levels unclear; reporting
difficult to follow, and data NR
for all exposure levels indicated
as tested; nonexposed controls
did not appear to be included]
1978)
(Persoz et
al.. 2010)
In vitro (human
immortalized
lung cells); n=4
experiments
Formalin gas: 0.050
mg/m3 for 30 minutes, ±
TNFa sensitization
Lung cell Cytokine
secretion
(at 24 hr post-FA)
N/C in IL-6, IL-8, or MCP-1 without TNF
a sensitization
Increased IL-8 only with sensitization
Note: air exposure alone increased IL-8
Not Informative [formalin; in
vitro; short duration; unknown
exposure level relevance; small
sample size; controls exhibited
effects from air-only exposure]
(Persoz et
al.. 2011)
In vitro (human
immortalized
lung cells); n=4
experiments
Formalin gas: 0.050
mg/m3 for 30 minutes,
with or without
aspergillus spores (Asp)
Lung cell cytokine
secretion
(at 24 hr post-FA)
N/C in IL-8 or MCP-1 mRNA or protein
Not Informative [formalin; in
vitro; short duration; unknown
exposure level relevance; small
sample size; controls exhibited
effects from air-only exposure]
(Persoz et
al.. 2012)
In vitro (human
immortalized
lung cells); n>3
experiments
Formalin gas: 0.050
mg/m3 for 30 min;
treatment with
sensitizers (i.e., TNFa or
MCM)
Bronchial or alveolar
cytokine secretion
(at 24 hr post-FA)
IL-8 production in alveolar cells
induced by TNFa or macrophage-
conditioned media (MCM) increased
by FA
MCP-1 production in bronchial cells
induced by sensitizers increased by FA
N/C om IL-8 or MCP-1 otherwise
Note: expression affected by air alone
Not Informative [formalin; in
vitro; short duration; unknown
exposure level relevance; small
sample size; controls exhibited
effects from air-only exposure]
(Kastner et
al.. 2013)
In vitro (human
immortalized
lung cells); n=3
experiments
Formalin gas: 0.2 mg/m3
for 30 min, 1 hr, or 2
hr/day once or for 4 d
Lung cell cytokine
secretion and
epithelial barrier
function/ viability
(at 24 hr post-FA)
N/C in IL-6 or IL-8 release, or TEER
(measures disruption to epithelial cell
monolayer) by FA alone
Note: viability affected by air exposure
Not Informative [formalin; in
vitro; short duration; unknown
exposure level relevance; small
sample size; controls exhibited
effects from air-only exposure]
(Lino-Dos-
Santos-
Female Wistar
rats (n=5)
Formalin 1% or methanol
vehicle for 3 days
(90min/d), ±
ovariectomy
BAL counts
Ex vivo lung IL-10
1 d after challenge: FA/OVA versus
OVA alone decreased total cell counts,
including mononuclear cells,
neutrophils, and eosinophils
Not Informative [formalin
(MeOH controls); naive not
chamber exposed; unquantified
This document is a draft for review purposes only and does not constitute Agency policy.
A-465 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
Franco et
Sensitization: After FA, s.c. 10 |og OVA, with s.c.
boost 7 d later
Challenge: After 7 d, 1% OVA aerosol for 15 min
FA/OVA versus OVA alone: Robust IL-
10 increase
high levels; FA alone untested;
small sample size]
al.. 2013a)
(Lino-Dos-
Santos-
Male Wistar rats
(n=5-6)
Formalin 1% for 3 days
(90 min/d)
Lung cellular
oxidative burst (flow)
and tissue
oxidative stress-
peroxynitrite (3-NT)
Increased cellular oxidative burst
(DFFH, ± OVA)
Increased lung nitration (peroxynitrite
formation; without OVA)
Not Informative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity]
Note: vitamin C, E blunted
effects
Franco et
al.. 2010)
Sensitization: immediately post-FA, s.c. 10 |og
OVA; boost 1 wk later with s.c. 10 |og OVA
injection
Challenge: 1 wk later with 1% aerosol OVA (15
min)
(Macedo et
al.. 2016a)
Male Wistar rats
(n=6)
Formalin 1% for 3 days
(90 min/d)
Lung (or lung cells)
oxidative stress
indicators: H202,
nitrites, oxidative
burst, enzyme activity
and gene expression
of redox-related
proteins
Formaldehyde exposure increased
H202 and N02, but not DCFH-DA
(oxidative burst), and exposure
increased expression of cNOS and
iNOS, SOD and catalase, but did not
affect the activity of enzymes
associated with detoxification
processes (e.g., glutathione reductase)
Not Informative [formalin;
unquantified high levels; short
duration and periodicity]
Note: Photobiomodulation
(laser) therapy blunted effects
(Lima et al.,
2015)
Male Fischer rats
(n=7)
Formalin 1, 5, or 10% for
5 days (20 min x 3/d)
Trachea or diaphragm
muscle (DM)
oxidative stress
indicators: carbonyl
protein, lipid
peroxidation, and
catalase activity; and
inflammatory cell
influx
In Trachea: increased lipid
peroxidation at 1 and 5, but not 10%;
N/C in catalase or inflammatory cell
influx; increased mucus deposits at
5%, and increased metaplasia and
ulceration at 10%
In DM: increased lipid peroxidation at
1 and 5, but not 10%; increased
carbonyl protein and increased
inflammatory cell influx at 10%;
decreased catalase at >1%
Not Informative [formalin;
unquantified high levels; short
duration and periodicity;
controls not chamber exposed]
(Lino dos
Santos
Male Wistar rats
(n=5)
Formalin 0,1% for 3 days
(90 min/d)
BAL nitrites
FA increased BAL nitrites, which was
exacerbated with OVA sensitization
Notlnformative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity]
Franco et
Sensitization: immediately post-FA, i.p. 10 ng
OVA; boost 1 wk later with s.c. injection
al.. 2009)
This document is a draft for review purposes only and does not constitute Agency policy.
A-466 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
Challenge: 1 wk later with aerosolized OVA
(Lino-Dos-
Santos-
Franco et
al.. 2013b)
Male Wistar rats
(n=5-8)
Formalin 1% or naive for
3 days (90min/ d), with
or without subsequent
OVA
Lung mRNA
Ex vivo Lung factors
FA increased iNOS and COX-1, but not
COX-2, expression in lung (OVA and FA
seemed to attenuate induction by
other)
FA/OVA vs. OVA increased NO and
LTB4 (both inhibited by inhibition of
NOS or by inhibition of COX), but not
TXB2 or PGE2
Note: suggests mast cell- and NO-
mediated effects
Notlnformative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity; comparisons
reported did not include all
relevant controls (e.g., FA
alone; air alone)]
Sensitization: after FA inhalation, s.c. lOug OVA
with same boost 7 d later
Challenge: after 1 wk, 1% OVA aerosol for 15
min
(Lino-Dos-
Santos-
Franco et
al.. 2011b)
Male Wistar rats
(n=5/ group)
Formalin 1% for 3 days
(90 min/d)
BAL cell counts
Lung ROS
Ex vivo lung cytokines
in explants or
cultured BAL cells
FA increased total BAL cells,
mononuclear cells, and neutrophils
FA decreased SOD, but not catalase,
GPX, GR, or GST activity in lung tissue;
mRNA expression for SOD, catalase,
NOS, and COX was increased
FA increased IL-ip and IL-6 in explants;
increased N02 and H202 in BAL cells
Notlnformative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity; some ex vivo]
(Lino dos
Santos
Franco et
al.. 2006)
Male Wistar (n=5-
6)
Formalin 1% or methanol
vehicle for 4 days (30,
60, or 90 min/d)
BAL cell counts
Lung IHC
Ex vivo BAL nitrites
Increased BALTotal cells (90 min only),
mononuclear cells (60 and 90 min),
and neutrophils (30, 60, or 90 min)
Increased ex vivo cultured BAL cell
release of nitrites
Lung IHC showed mast cell
degranulation and neutrophil
infiltration
Note: number of cells recovered in BAL
was significantly reduced by capsaicin
(depletes neuropeptides from sensory
nerve endings), but bronchial
hyporesponsiveness not altered;
conversely L-NAME (inhibits NO
synthase) did not affect BAL cells, but
did restore bronchial responsiveness;
Not Informative [formalin
((MeOH controls); unquantified
high levels; small sample size;
short duration and periodicity;
comparisons reported to naive
rats rather than MeOH controls;
some ex vivo]
NOTE: if a relevant MOA is
identified from more
informative studies,
pharmacological intervention
endpoints might be
reconsidered
This document is a draft for review purposes only and does not constitute Agency policy.
A-467 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
administration of 48/80 to deplete
mast cells blunted FA-induced effects
on both BAL cell counts and bronchial
response
(Lino-Dos-
Santos-
Franco et
al.. 2011a)
Female Wistar
rats (n=5)
Formalin 1% or naive for
3 days (90 min/d), with
or without ovariectomy
BAL counts and mast
cell degranulation
FA increased total BAL cell counts,
mononuclear cells and neutrophils,
but not eosinophils
Decreased lung mast cell number and
increased degranulation
Not Informative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity; impact of sham
surgery/ FA alone untested;
na'ive not chamber exposed]
(Lino-Dos-
Santos-
Franco et
al.. 2010)
Male Wistar rats
(n=5-6)
Formalin 1% for 3 days
(90 min/d)
Pulmonary vascular
permeability (Evans
blue)
BAL cell counts
Ex vivo cultured BAL
cells
factors/cytokines
Phagocytosis (flow)
Increased BAL mononuclear cells and
neutrophils, but N/C in eosinophils or
in lung ICAM-1
Increased vascular permeability (±
OVA)
FA increased ex vivo LTB4; FA+OVA
increased BAL LTB4, TXB2, IL-lb,ll-
6,VEGF
N/C in phagocytosis;
Not Informative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity; some ex vivo]
Note: vitamin C and E blunted
effects
Sensitization: immediately post-FA, s.c. 10 ng
OVA; boost 1 wk later with s.c. 10 ng OVA
injection
Challenge: 1 wk later with 1% aerosol OVA (15
min)
(Kita et al.,
2003)
Male Hartley
guinea pigs
(n=10+/gi"oup)
Nasal Instillation of
saline or Formalin 0.1 or
1.0%; 3x/wk for 6 wk
BAL cell counts
N/C in BAL fluid cell counts by FA with
passive or active sensitization (not
measured for FA alone)
Not Informative [formalin; high,
unknown levels; short
periodicity; exposure route;
effect of FA alone not
measured]
Sensitization: intradermal anti-OVA serum on
day 38 (passive) or i.p. 2 mg OVA on Day 3
(active) with boost i.p. 10 mg OVA day 24
Challenge: 1 mg/mL nebulized OVA 15 min after
last FA exposure on day 45
(Kita and
Oomichi,
1974)
In/Ex vitro:
trachea from
guinea pigs (n=3)
Formalin gas: 39.4 or
67.7 mg/m3 for <30
minutes
In vitro ciliary beat
frequency
FA decreased CBF 50% in 11.5 minutes
(39.4 mg/m3) or 4.5 minutes (67.7
mg/m3)
Not Informative [formalin;
excessively high levels; short
duration; ex vitro; small sample
size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-468 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and notes*
(Lino dos
Santos
Franco et
al.. 2009)
Male Wistar rats
(n=5)
Formalin 0,1% for 3 days
(90 min/d)
BAL cell counts
Lung mast cell
degranulation
Increased Total BAL cells, mononuclear
cells, and neutrophils (eosinophils
undetected); FA inhibited OVA-
induced increases in all cell counts
FA increased mast cell degranulation;
FA inhibited OVA induced
degranulation
FA induced PECAM expression; FA
inhibited OVA induced increases
Not Informative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity]
Sensitization: immediately post-FA, i.p. lOug OVA;
boost 1 wk later with s.c. injection
Challenge: 1 wk later 1% aerosol OVA for 15 min
Table A-68. Changes in pulmonary function involving provocation (e.g., bronchoconstrictors; allergens; etc.)
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Observational Epidemioloav Studies
(Gorski and
Krakowiak,
1991)
Human textile
and shoemakers
(n=367)
Not exceeding 0.5
mg/m3 (duration at least
1 year (average= -12
years)
Bronchial hyper-
reactivity to
histamine
Bronchial hyperreactivity in 11
nonbronchitic patients (14
bronchitic/2 asthmatic ppl)
Low Confidence [incomplete
and confusing methods and
results; comparisons unclear]
Controlled-Exposure Studies in Humans or Primary Human Cells
(Krakowiak
et al., 1998)
Human workers
with bronchial
asthma or
healthy subjects
(n=10 each)
Formalin (assumed: test
article NR): 0.5 mg/m3
for 2 hr with follow-up
out to 24 hr
Bronchial provocation
responses (histamine)
N/C in Bronchial reactivity to
histamine (Note: scoring measures of
nasal symptoms were elevated)
Low Confidence [formalin; short
duration; small sample size]
NOTE: ACUTE; no effect on FEVi,
etc.
(Casset et
al., 2006a)
Human (n=19
with mild asthma
and allergy to
mite allergen)
Formalin =0.1 mg/m3 for
30 minutes; placebo at
=0.03 mg/m3 double-
blind randomized;
restricted to mouth
breathing only
Airway response to
mite allergen (Note:
large allergen size
chosen to deposit in
large airways)
A lower level of allergen was necessary
to induce bronchoconstriction
following FA exposure and FA
exposure: both immediate and late-
phase responses; note: N/C in
pulmonary function tests with FA
exposure alone prior to allergen
challenge
Low Confidence [formalin; short
duration; not clear that
restriction to mouth breathing
is realistic for typical inhalation]
NOTE: ACUTE; within-subjects
comparison between air and FA
(Ezrattv et
al.. 2007)
Human (n=12
intermittent
Formalin 0.5 mg/m3 for
60 minutes; randomized
Allergen (pollen)-
induced changes in
N/C in pulmonary function by allergen
(a borderline decreased response, p =
0.06, was observed) or to MCh
Low Confidence [formalin; short
duration]
This document is a draft for review purposes only and does not constitute Agency policy.
A-469 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
asthmatics with
allergy to pollen)
to air or FA first (no
nonexposed controls)
airway FEV1 and MCh
responses (note: did
not appear to test
MCh w/o allergen) 8
hr later
responsiveness after allergen
challenge; note: N/C in pulmonary
function by FA
NOTE: ACUTE; within subjects
comparison between air and FA
Controlled-Exposure Studies in Animals. Animal Cells, or Immortalized Human Cells
(Riedel et
al.. 1996)
Female Dunkin-
Hartley guinea
pigs (n=12)
Formaldehyde (bottled
pressurized gas) 0, 0.16,
0.31 mg/m3 for 5 d (8
hr/d)
Airway response to
OVA
Increased OVA challenge-induced
airway obstruction by 0.31 mg/m3 (3,
7, and 10 animals exhibited airway
obstruction across groups)
High or Medium Confidence [no
comparison group with FA
without OVA]
NOTE: guinea pigs have been
shown to be more sensitive to
airway constriction from
toxicants than other animals]
Sensitization: 0.5% inhaled OVA; OVA boost at
2wk
Challenge: 1% inhaled OVA 1 wk later
(Leikauf,
1992)
[considered
same cohort
as
(Swiecicho
wski et al.,
Male Hartley
guinea pigs (n=5-
7)
PFA 0, 0.12,0.37, 1.23,
3.69, 12.3, or 36.9
mg/m3 for up to 8 hr
Bronchial reactivity to
i.v. acetylcholine
Increased specific resistance at >12.3
mg/m3 with 2 hr; Increased at >1.23
mg/m3 with 8hr (i.e., duration >
concentration); with 8 hr,
hyperreactivity persisted >24 hr
postexposure
See Swiechichowski et al., 1993
NOTE: ACUTE
1993)
(Swiecicho
wski et al.,
Male Hartley
guinea pigs
(n=5-7/group)
PFA from 0.12-123
mg/m3, for 2 or 8 hours
(multiple experiments)
Airway reactivity
Ex vivo airway
reactivity (trachea)
Increased pulmonary resistance
(reversible bronchoconstriction) and
airway reactivity to acetylcholine at
>1.23 mg/m3 (not at 0.36 mg/m3) for 8
hr; at > 12.3 mg/m3 (not at <3.6
mg/m3) for 2 hr
Increased ex vivo reactivity (smooth
muscle contraction) at 4.18 mg/m3for
8 hr
High or Medium Confidence at
1.23 mg/m3 and above [short
duration]
Low Confidence below 1.23
mg/m3 and ex vivo [ex vivo;
sample size of 5 at 1 or more
levels below lppm]
NOTE: ACUTE; duration
appeared to be more important
than FA level for pulmonary
resistance
1993)
(Larsen et
Male BALB/cA
mice (n=10)
PFA 0.49, 2.21, or 4.9-7.0
(dry vs. humid air)
mg/m3; 60 min
Airway reactivity
Increased airway reactivity (decreased
expiratory flow rate) in humid air in
OVA-sensitized mice at 7 mg/m3
High or Medium Confidence
[short duration]
al.. 2013)
This document is a draft for review purposes only and does not constitute Agency policy.
A-470 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Sensitization: pre-FA i.p. 1 ng OVA, with 0.1 ng
OVA boosts i.p. on days 14 and 21 (note: FA on
day 31)
Challenge: 0.2% OVA aerosol- 20min on day
29&30
Increased bronchoconstriction in a dry
environment without OVA
sensitization at 4.92-7.0 mg/m3 (with
OVA sensitization reducing the
response to formaldehyde)
NOTE: ACUTE; suggests that
environmental humidity may
affect acute airway reactivity
induced by formaldehyde;
experiments on inflammatory
markers (below) considered less
informative
(Liu et al.,
2011)
Male Balb/c mice
(n=6/ group)
Formalin 0, 0.5, or 3
mg/m3 for 21 d (6 hr/d)
Airway reactivity
Slightly increased responsivity to MCh
compared to saline controls; robust
amplification in 3mg/m3 FA+OVA
group
Low Confidence [formalin]
Sensitization: i.v. 20 mg OVA on Days 10 and 21
Challenge: 1% OVA aerosol for 30 min/d for 7 d
(Qiao et al.,
2009)
Male Wistar rats
(n=8/group)
Formalin 0, 0.51 or 3.08
mg/m3 for 3 wk (6 hr/d)
Airway response to
methylcholine
3.08 mg/m3 FA alone increased
hyperresponsiveness to MCh, which
was amplified with OVA administration
at > 0.51 mg/m3
Low Confidence [formalin]
Sensitization: i.p. OVA on Days 10 and 18
Challenge: 1% OVA 30 min/d for 7 d
(Wu et al.,
2013)
Male Balb/c mice
(n=8/group)
Formalin 0, 3 mg/m3 for
4 wk (6 h/d, 5 d/wk)
Airway responsivity
to Methylcholine
(MCh)
Airway was slightly hyperesponsive to
MCh by FA alone, but severely so in
FA+OVA groups
TRPA1 and TRPV1 antagonists reduced
FA+OVA-induced airway
responsiveness
Low Confidence [formalin;
pharmacological interventions
did not include effects of FA
alone]
Sensitization: s.c. 80 ng OVA on days 10,18, and
25
Challenge: 1% OVA aerosol 30 min/d on Days
29-35
(Biagini et
al.. 1989)
Male cynomolgus
monkeys (n=9)
Formalin 3.08 mg/m3 for
10 min (challenge
experiment)
Bronchoreactivity to
methylcholine (all
with MCh)
Increased bronchoconstriction by FA
challenge at 2, 5, and 10 min
postchallenge
Low Confidence [formalin; short
duration; FA without
methylcholine untested]
(Maiellaro
et al.. 2014)
Pregnant Wistar
rats (n=5)
Formalin 0.92 mg/m3
from GDs 1-21: 1 hr/d, 5
d/wk
Tracheal response to
MCh
24hr after OVA challenge, offspring
have: decreased tracheal response to
MCh
Note: Decreased birth weight in
offspring.
Nonmanipulated group exhibits large,
unexplained differences from vehicle
control (and has reporting limitations)
Not Informative [formalin;
short periodicity; offspring
comparisons do not include FA
alone; unclear comparability for
some groups; small sample size]
Sensitization: s.c. 10 ng OVA with sc boost after
7 d
Challenge: 7 d later, 1% OVA aerosol 15 min/d,
3d
Pregnant Wistar
rats (n=5 dams;
Formalin 0.92mg/m3
from GDs 1-21: 1 hr/d, 5
d/wk
Response to MCh
24 h after LPS challenge, offspring
exposed to formaldehyde have
decreased MCh response
Not Informative [formalin;
short periodicity; offspring
This document is a draft for review purposes only and does not constitute Agency policy.
A-471 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Silva
Ibrahim et
10 pups/ group
for experiments)
Randomly assigned pups all received 5 mg/kg
lipopolysacharride (LPS) injections at PND 30
comparisons do not include FA
without LPS; small sample size]
al.. 2015)
(Kita et al..
2003)
Male Hartley
guinea pigs
(n=5-7/group)
Nasal Instillation of
saline or Formalin 0.1 or
1.0%; 3x/wk for 6 wk
Bronchoconstriction
to methylcholine
N/C in airway response to MCh by FA
or FA with passive sensitization, but
induced by FA with active sensitization
Not Informative [formalin; high,
unknown levels; short
periodicity; exposure route]
Sensitization: intradermal anti-OVA serum on
day 38 (passive) or i.p. 2 mg OVA on day 3
(active) with boost i.p. 10 mg OVA Day 24
Challenge: 1 mg/mL nebulized OVA 15 min after
last FA exposure on day 45
(Lee et al..
1984)
Male English
guinea pigs (n=4)
Formalin: 7.38 or 12.3 mg/m3for 5 days
FA challenge with 2.46 or 4.9 mg/m3 for 1 or
4hr, respectively on Days 7, 22, and 29
Respiratory rate change from prechallenge
baseline
N/C in pulmonary sensitivity (either
immediate or delayed-onset) to
formaldehyde challenge
Note: 2/4 animals exhibited dermal
sensitivity (likely contact-mediated) to
topical FA; 12.3 mg/m3 caused 40-50%
respiratory rate decrease for >5 hr
(later time points NR)
Not Informative [formalin;
small sample size; high
exposure levels; no comparison
to controls with no prior
formaldehyde exposure
(unclear if this, by itself, caused
effects); unclear reporting]
(Lino-Dos-
Santos-
Female Wistar
rats (n=5)
Franco et
Formalin 1% or methanol
vehicle for 3 days (90
min/d), ± ovariectomy
al.. 2013a)
Lung oxidative stress,
microvascular
leakage and mast cell
degranulation; ex
vivo tracheal
reactivity
1 d after OVA challenge: FA/OVA
versus OVA alone: Reduced MPO and
vascular permeability; decreased mast
cell degranulation
Decreased tracheal reactivity
Not Informative [formalin
(MeOH controls), naive not
chamber exposed; high,
unquantified levels, FA alone
untested; small sample size]
Sensitization: After FA, s.c. 10 ng OVA, with s.c.
boost 7 d later
Challenge: After 7 d, 1% OVA aerosol for 15 min
(Lino-Dos-
Santos-
Female Wistar
rats (n=5)
Franco et
Formalin 1% or naive for
3 days (90 min/d), with
or without ovariectomy
al.. 2011a)
Ex vivo trachea
response
N/C in ex vivo tracheal response to
methacholine
Not Informative [formalin,
naive not chamber exposed; ex
vivo; high, unquantified levels,
FA alone untested; small
sample]
(Lino dos
Santos
Male Wistar
(n=5-6)
Formalin 1% or methanol
vehicle for 4 days (30,
60, or 90 min/d)
Ex vivo airway
responsivity
Decreased ex vivo bronchial, but not
tracheal, response to methacholine
Not Informative [formalin
(MeOH controls); naive not
chamber exposed; high,
This document is a draft for review purposes only and does not constitute Agency policy.
A-472 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Franco et
al.. 2006)
Note: number of cells recovered in BAL
was significantly reduced by capsaicin
(depletes neuropeptides from sensory
nerve endings), but bronchial
hyporesponsiveness not altered;
conversely L-NAME (inhibits NO
synthase) did not affect BAL cells, but
did restore bronchial responsiveness;
administration of 48/80 to deplete
mast cells blunted FA-induced effects
on both BAL cell counts and bronchial
response
unquantified levels,
comparisons to naive rats
rather than MeOH controls;
small sample size]
NOTE: if a relevant MOA is
identified from more
informative studies,
pharmacological intervention
endpoints might be
reconsidered
(Lino-Dos-
Santos-
Franco et
al.. 2013b)
Male Wistar rats
(n=5-8)
Formalin 1% or naive for
3 days (90 min/d), with
or without subsequent
OVA
Ex vivo bronchial
response to MCh
Prior FA exposure reduced OVA-
induced ex vivo bronchial
hyperresponsiveness
Note: N/C in respiratory resistance or
elastance with FA alone
Not Informative [formalin;
na'ive not chamber exposed;
high, unquantified levels; short
duration and periodicity;
comparisons did not include all
relevant controls (e.g., FA
alone; air alone); small sample
size]
Sensitization: after FA inhalation, s.c. 10 pg OVA with
same boost 7 d later
Challenge: after 1 wk, 1% OVA aerosol for 15 min
Table A-69. Serum (primarily) antibody responses
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Observational Epidemioloav Studies
(Wantke et
al.. 1996a)
Human children
in schools (n=62)
vs. control (n=19)
Particleboard schools:
0.053,0.085, or 0.092
mg/m3 (n=18, 22, 22);
brick schools: 0.036,
0.028, or 0.032 mg/m3
(n=18, 22, 22); unclear
duration (<2.5 yr)
Serum FA-specific IgE
Before switching schools, 40% of
students had elevated FA-specific IgE,
which significantly decreased 3
months after switch to low-FA schools
(p <0.002)
Note: while symptoms correlated to
FA levels, FA-specific IgE did not
High or Medium Confidence [no
blinding, but not clearly an
issue]
Note: Natural experiment (pre-
and postschool switch) with
limited exposure contrast and
assays
(Kim et al.,
1999)
Human medical
students (n=167)
3.74±3.48 mg/m3 for up
to 4 years of school
(periodicity NR)
Serum FA-specific IgG
and IgE (antibodies to
14 (8.4%) students had FA-specific IgG,
which was not related to duration of
High or Medium Confidence
Note: Limited assays
This document is a draft for review purposes only and does not constitute Agency policy.
A-473 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
and nonexposed
controls (n=67)
FA-human serum
albumin conjugate)
schooling (No relationship to
symptoms)
N/C in FA-specific IgE
(Avdin et
al.. 2013)
Human male
fiberboard
workers
0.25±0.074 mg/m3
(average 7.3 yr
employed; n=46) vs.
nonexposed controls
Serum Antibodies
Decreased IgG and IgM
N/C in IgA
High or Medium Confidence
(Wantke et
al.. 1996b)
Human medical
students (n=45)
0.153± 0.062 mg/m3 for
4 wk (Total: 17 d; 51 hr);
phenol co-exposure
Serum FA-specific IgE
Total IgE
N/C in FA-specific IgE; N/C in total IgE
Low Confidence [37%
participation; phenol co-
exposure; limited periodicity]
Note: limited assays
(Wantke et
al.. 2000)
Human medical
students (n=27);
23 controls
0.265± 0.07 mg/m3 for 5
or 10 weeks
(intermittent—not
specified, but assumed
=3hr/d)
Serum Antibodies and
FA-specific Antibodies
After 5 wk: N/C FA-lgE or Total IgE
After 10 wk: 4/27 students developed
IgE against FA-albumin, but 0/23
developed IgG; N/C in Total IgE
Low Confidence [no reporting of
% participation or population
demographics; limited, unclear
periodicity; phenol co-
exposure]
Note: 1 of 4 positive was a
smoker (4 smokers in study);
limited assays
(Erdei et al.,
2003)
Human (sex NR)
symptomatic
students (9-11 yo
w/ respiratory
issues) (n=176)
0.006-0.057 mg/m3
(average= 0.018 mg/m3);
duration unknown [co-
exposure: NO2, benzene,
toluene, xylene, and dust
mite allergen]
Serum Antibodies
N/C total IgG, IgA, IgM, or IgE (data
NR)
Increased airway pathogen bacteria-
specific IgG (not IgA or IgM) with FA
Low Confidence [comparisons
to "normal" range rather than
to control group; co-exposure;
limited reporting]
Note: symptomatic only; authors
hypothesized increased
bacterial-specific IgG may
represent increased B cell
response (maybe more
infections)
(Zhou et al.,
2005)
Human anatomy
students (n=8)
0.74±0.11 mg/m3 (4-
week course-
intermittent)
Serum FA-specific IgE
antibodies
No students had FA-specific IgE after
exposure
Low Confidence [small sample
(n=8); limited, unclear
periodicity; reporting as yes/no
rather than analytical results,
and no clear comparison to
preexposu re]
This document is a draft for review purposes only and does not constitute Agency policy.
A-474 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Ohmichi et
al.. 2006)
Human anatomy
students (n=8
measured for FA;
n=6 for FA-
specific IgE)
0.41-1.81 mg/m3 (20
laboratory sessions over
10 weeks; laboratory
sessions ranged from
l.l-10hrs, averaging
3hr)
Serum IgE and FA-
specific IgE (threshold
of 0.34 UA/mL)
No significant changes in IgE, and no
positive result for FA-specific IgE (data
presented was highly variable), as
compared to measure 90 min before
1st session of laboratory course
Low Confidence [small sample
(n=6-8); limited and variable
periodicity]
(Thrasher et
al.. 1987)
Human sympto-
matic exposed
subjects, controls
(n=8/ group)
Exposed (mobile home
measures): 0.086-0.68
mg/m3 (residency -6-7
yr); nonexposed: not
measured (authors
assume: <0.037)
Serum FA-specific IgG
and IgE
No detection of FA-specific IgE
Increased FA-specific IgG in all 8
exposed subjects, but only in 1/8
controls (had PD)
Low Confidence [small sample;
symptomatic vs.
nonsymptomatic comparison;
reporting limitations]
(Dvkewicz
et al.. 1991)
Human medical
volunteers (n=55;
31 F, 24M)
Generally, 0.25-0.79
mg/m3 (1 subject up to
13.5 mg/m3); duration
4.53± 1.09 yr
Serum FA-specific IgG
and IgE
N/C in incidence of FA-HSA- specific
IgG or IgE (3 subjects had FA-specific
IgG and IgE, and 2 more had FA-
specific IgG only)
Low Confidence [periodicity
unspecified; unclear exposure
comparison- control levels NR
and variable range in exposed]
(Thrasher et
al.. 1990)
Human various
exposed groups
of patients, and
asymptomatic
controls
"controls"—chiropractic
students (n=28):
assumed > 0.53 mg/m3
for 28 wk (13 h/wk);
mobile home residents
(n=19): 0.05-0.62 mg/m3
for 2-7 yr; office workers
(n=21): assumed
0.012-0.95 mg/m3,
duration N/R;
occupational (n=8):
levels/ duration N/R;
removed from exposure
for >lyr: 0.17-1.0
mg/m3
Serum FA-specific
IgG, IgM, and IgE
Blood autoantibodies
Proportion of pooled titers (IgG, IgM,
and IgE) of FA-specific antibodies (i.e.
% at > 1:8) was greater in all patient
groups than in controls (Note: most
apparent for IgG, but others also
appear elevated; FA-specific IgE was
not found in any of the patients
"removed" from exposure)
Mobile home residents and office
workers had increased autoantibodies
vs. controls (i.e., antismooth muscle or
antiparietal cell)
Low Confidence [controls not
unexposed; patients to
nonpatients comparisons
questionable]
Note: authors argue only real
difference between
asymptomatic control students
and patients is one of duration
of exposure
(Gorski and
Krakowiak,
Human textile
and shoe makers
(n=367)
Not exceeding 0.5
mg/m3 (duration at least
1 year (average =12
years)
Serum FA-specific IgE
Antibodies
No FA-specific IgE in patients tested
(seems to be testing in a small subset
of all subjects)
Low Confidence [incomplete
and confusing methods and
results; comparisons unclear]
1991)
This document is a draft for review purposes only and does not constitute Agency policy.
A-475 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Palczvnski
et al.. 1999)
Human
apartment house
residents (n=465
total, =40%
children)
3 categories of exposure:
<0.025, 0.025-0.05, and
>0.0501 mg/m3; duration
unclear, periodicity
assumed to be constant
Total serum IgE
Note: N=l-2 at high
HCHO levels;
N=27-38 at mid, low
levels
Serum antibodies to
FA
Total IgE was not changed at
0.025-0.5 as compared to <0.025 in
children or adults (n size at >0.05 was
too small to compare); No FA-specific
antibodies were detected (details NR);
note: children exposed to 0.025-0.05
mg/m3 and tobacco smoke had
elevated IgE
Low Confidence: IgE [small
sample size; subsampling for IgE
not reported; minimal exposure
differential; results not
stratified by sex or smoking
status]
Not Informative: FA antibodies
[methods NR; data NR]
(Madison et
al.. 1991)
Human residents,
spill-exposed (n=
41) or unexposed
controls (n=29)
Formaldehyde (PFA):
>2.46 mg/m3 for first 48
hr, then average
dropped to 0.028 mg/m3,
but urea and
methylamines
unmeasured/not
corrected
FA-specific serum
antibodies and
autoantibodies
N/C in FA-specific IgE
Increased FA-specific IgM and IgG
Increased odds ratio of having 1+
autoantibodies (although higher, no
sig. increase in any one auto-antibody)
Not Informative [mixture
exposure; co-exposures not
corrected for; FA in controls
unmeasu red]
(Grammer
et al.. 1990)
Human workers
(Boeing; n=37);
details N/R
0.0037-0.090 mg/m3
(not stratified by
exposure; all exposed;
duration N/R)
Serum FA-specific IgG
and IgE
0/37 had FA-specific IgG
5/37 had elevated IgE (vs. control sera)
that was not specific to FA-HSA or HSA
Not Informative [details on
population N/R; details on
exposure NR; no specific
comparison to FA levels]
Controlled-Exposure Studies in Animals. Animal Cells, or Immortalized Human Cells
(Fuiimaki et
al.. 2004b)
Female C3H mice
(n=5-6 per
group)
PFA 0, 0.098, 0.49, or
2.46 mg/m3; 12 wks (16
hr/d, 5 d/wk)
Serum Antibodies and
Antibodies to Antigen
No change in anti-OVA IgE (variable) or
lgG2a or Total IgE
Decreased anti-OVA IgGi (at 0.49
mg/m3 only) and lgG3 (at 0.098-0.49
mg/m3)
Body weight decreased 20% at 0.49
mg/m3
High or Medium Confidence
[slightly small sample size]
Sensitization: i.p. 10 ng OVA prior to FA
exposure; aerosol OVA boost for 6 min on wks
3, 6, 9, and 11
(Riedel et
al.. 1996)
Female Dunkin-
Hartley guinea
pigs (n=12)
Formaldehyde (bottled
pressurized gas) 0, 0.13,
0.31 mg/m3 for 5 d (8
hr/d);
Serum OVA-specific
IgGl
Increased OVA-specific IgGl by 0.31
mg/m3
High or Medium Confidence [no
comparison group with FA
without OVA]
Sensitization: 0.5% inhaled OVA; OVA boost at
2wk
Challenge: 1% inhaled OVA lwk later
This document is a draft for review purposes only and does not constitute Agency policy.
A-476 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Sapmaz et
al.. 2015)
Male SD rats
(n=5-7)
PFA 0, 6.15,12.3 mg/m3;
4 wks (8 hr/d, 5 d/wk)
Serum Antibodies
Increased IgA, IgM, and complement 3
Decreased IgG
High or Medium Confidence
[slightly small sample size; high
formaldehyde levels]
(Tarkowski
and Gorski,
Female Balb/c
mice (n=4/ group)
Formalin (assumed; test
article N/R) 0 or 2 mg/m3
for 10 d (6 hr/d) or 7 wk
(6 hr/d, 1 d/wk)
Serum OVA-specific
IgE
Increased OVA-specific IgE in mice
exposed for lOd, but not in those
exposed lx/ wk, as compared to
controls
Specific to nasal tissue, as OVA
sensitization via i.p. injection caused
N/C
Low Confidence [formalin; small
sample size]
Note: pinpoints issue of
importance and interpretability
of different sensitization
methods
1995)
Sensitization: intranasal 25 ng OVA lx/wk for 7
wk OR i.p. 1 ng OVA lx/wk for 4 wk
(Wu et al.,
2013)
Male Balb/c mice
(n=8/group)
Formalin 0, 3 mg/m3 for
4 wk (6 h/d, 5 d/wk)
Serum antibodies
FA alone increased total IgE, but not
OVA-IgG or OVA-lgE; FA+OVA
increased IgE compared to OVA alone,
but did not further elevate OVA-IgG or
OVA-lgE (slight, NS increases)
compared to OVA
TRPA1 and TRPV1 antagonists reduced
FA+OVA-induced serum antibodies
Low Confidence [formalin;
pharmacological interventions
did not include effects of FA
alone]
Sensitization: s.c. 80 ng OVA on days 10,18, and
25
Challenge: 1% OVA aerosol 30min/d on day 29-
35
(Kim et al.,
2013b)
Female NC/Nga
(atopic-prone)
mice (n=5-
7/group)
Formalin (assumed; test
article NR) 0, 0.25,1.23
mg/m3 for 4 wk (6 hr/d,
5 d/wk)
Plasma Antibodies
and Antigen-specific
Abs
Plasma IgGl increased by FA alone
(0.25 mg/m3 only), but N/C in total IgE
or lgG2a
FA exacerbates HDM-induced IgE
(>0.25 mg/m3) and lgG2a (0.25 mg/m3
only), but not IgGl
HDM-specific IgE not changed
Low Confidence [formalin; small
sample size]
Note: multiple supplementary
files; HDM-specific IgE data NR
Sensitization: topical house dust mite (HDM;
ear) stimulation (25 mg Df ointment) lx/wk for
4 wk
(Gu et al.,
2008)
Female Balb/c
mice (n=5-6/
group)
Formalin (assumed; test
article NR) 0.12 or 0.98
mg/m3 for 5 wk (24h/d,
5d/wk)
Serum Antibodies and
OVA-specific
Antibodies
N/C in total serum IgG or IgE
Increased OVA-specific IgE in allergen
primed host, only at 5 weeks (not < 4
wk) and only at 0.98 mg/m3; N/C in
OVA-IgG
Low Confidence [formalin; small
sample size]
Sensitization: i.p. lOmg OVA on day 0 and 7 pre-
FA
(Jung et al.,
2007)
Female C57BL/6
mice (n>5/group)
Formalin (assumed; test
article NR) 0,6.15,12.3
mg/m3 for 2 wk (6h/d,
5d/wk)
Serum Antibodies
Increased Total IgGl, lgG3, IgA, and IgE
Decreased Total lgG2a and 2b; N/C
IgM
Note: body wt decreased =10%
Low Confidence [formalin; high
exposure levels; small sample
size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-477 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Holmstrom
et al.,
Female SD rats
(n=8-9 treated
rats; n=6 control)
Formalin (assumed; test
article NR) 15.5±2.3
mg/m3 for 22 months (6
hr/d, 5 d/wk); all rats
vaccinated: anti-tetanus
and Pneumovax
Serum antibody
response to
vaccination
N/C in IgM response to vaccine-related
antigens
Variable increases in IgG against
specific antigens were not statistically
significant
(Note: IgE not measured)
Low Confidence [formalin;
excessively high exposure level;
no unvaccinated comparison
group]
Note: authors indicate B cell
function unchanged
1989a)
(Lee et al.,
1984)
Male English
guinea pigs (n=4)
Formalin: 7.38 or 12.3
mg/m3 for 5 days, with
FA challenge with 2.46 or
4.9 mg/m3for lor4 hr,
respectively
Serum antibody to
formaldehyde
(isotype not
specified) measured 9
or 17 days (i.e., days
14 or 22) after
exposure
N/C antibody response to 2.46 or 4.9
mg/m3 (data NR)
Note: 2/4 animals exhibited dermal
sensitivity (likely contact-mediated) to
topical FA
Low Confidence [formalin;
small sample size; high
exposure levels]
Note: although there was no
comparison to controls with no
prior formaldehyde exposure,
this is not expected to affect this
measure
(Sadakane
et al., 2002)
Male ICR mice
(n=9 or 18)
Formalin 0.5% for 4 wk
(15 min/wk) ±
sensitization of house
dust mite allergen (Der f)
Blood Der f-specific
IgGl and IgE
N/C in Der f-specific IgGl or IgE (latter
appears to have been lower than
detection limit)
Low Confidence [formalin; high,
unknown exposure levels; short
periodicity]
Sensitization: i.p. with 3 mg/mL Der f (house
dust mite allergen) prior to FA
Challenge: intratracheal 10 ng Der f 3 hr after
last exposure (note: measures 3 d later)
(Kita et al.,
2003)
Male Hartley
guinea pigs (n=5-
7/group)
Nasal Instillation of
saline or Formalin 0.1 or
1.0%; 3x/wk for 6 wk
PCA reaction of naive
animals to injected
serum of exposed
animals
Increased anti-OVA IgG at >0.1% FA (at
4hr, but not 7 d after OVA challenge)
in naive animals injected with serum
Not Informative [exposure
route; formalin; high, unknown
exposure levels; short
periodicity; small sample size
(for some endpoints/ groups)]
Sensitization: i.p. anti-OVA serum on after 5 wk
FA (passive) or i.p. 2 mg OVA on day 3 (active)
prior to FA exposure with boost i.p. 10 mg OVA
day 24
Challenge: 1 mg/mL nebulized OVA 15 min after
last FA exposure on day 45
(Lino dos
Santos
Male Wistar rats
(n=5)
Formalin 0,1% for 3 days
(90 min/d)
Skin Antibodies
N/C in skin IgE
Not Informative [formalin;
unquantified high exposure
levels; small sample size; short
duration and periodicity]
Franco et
Sensitization: immediately post-FA, i.p. 10 ng
OVA; boost 1 wk later with s.c. injection
al.. 2009)
This document is a draft for review purposes only and does not constitute Agency policy.
A-478 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Challenge: 1 wk later with aerosolized OVA
Note: unclear endpoint
relevance
(Lino-Dos-
Santos-
Franco et
al.. 2013a)
Female Wistar
rats (n=5)
Formalin 1% or methanol
vehicle for 3 days
(90min/d), ±
ovariectomy
Skin IgE
1 d after OVA challenge: FA/OVA vs.
OVA alone: N/C in cutaneous OVA-
specific IgE
Not Informative [formalin
(MeOH controls); unquantified
high exposure levels; small
sample size; short duration and
periodicity; naive not chamber
exposed]
Note: unclear endpoint
relevance
Sensitization: After FA, s.c. 10 ng OVA, with s.c.
boost 7 d later
Challenge: After 7 d, 1% OVA aerosol for 15 min
Table A-70. Serum markers of immune response (other than antibodies), inflammation, or oxidative stress
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Observational Epidemioloav Studies
(Avdin et
al.. 2013)
Human male
fiberboard
workers
0.25±0.074 mg/m3
(average 7.3 yr
employed; n=46) vs.
nonexposed controls
Serum cell counts,
cytokines and related
factors
N/C in # hematologic cells, WBC, RBC,
Hb, neutrophils, or monocytes; N/C in
helper T, suppressor T, or B
lymphocytes
Increased % of lymphocytes, and
numbers and % of T cell (CD3+) and NK
cell (CD56+)
Increased TNFa, but N/C in
Complement 3 or 4; TNFa increased
more significantly in those not using
protective measures
High or Medium Confidence
Note: annex reviews immune
data
(Bassig et
al.. 2016)
(same cohort
as (Zhang et
al.. 2010)
Human melamine
workers (n=43) or
n=51age- and
sex-matched
unexposed from
different factories
in the same
region of China
1.6 mg/m3 (10% and 90%
= 0.74 and 3.08 mg/m3);
unclear exposure
duration (sampling over
a 3-week period)
Serum cell counts and
soluble markers
Decreased total WBC, Granulocytes,
Monocytes, Platelets, and
Lymphocytes
Decreased CD8+ cells (CD8 effector
memory cells most affected) and NK
cells
N/C in Monocytes, CD4+ cells,
CD4/CD8 ratio, or B cells; N/C in
soluble CD27 or CD30
High or Medium Confidence
This document is a draft for review purposes only and does not constitute Agency policy.
A-479 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Costa et
al.. 2013)
Human pathology
anatomists
(n=35) or
administrative
controls (n=35)
0.44±0.037 mg/m3 (as
high as 0.85 mg/m3 in
peaks); duration of
employment > lyr
Serum lymphocyte
subtypes
Decreased B cells (% CD19+) in
exposed
N/C in T cells or NK cells in exposed
Within the exposed workers: FA
exposure level correlated with
Increased % T cells (CD3+) and % T
helper cells (CD4+), and decreased %
NK cells
High or Medium Confidence
Note: authors suggest
immunosuppression
(Costa et
al.. 2019)
Human anatomy-
pathology lab
workers (n=85) or
administrative
controls (n=87)
8h TWA=0.47±0.037
mg/m3 (range=0.098-
1.71 mg/m3; as high as
3.94 mg/m3 in peaks);
duration of employment
average -12 yr
Serum lymphocyte
subtypes
Increased Cytotoxic (CD8+) T cells and
NK cells; Decreased B cells and
CD4/CD8 ratio; N/C in total T cells or
Helper (CD4+) T cells
High or Medium Confidence
Note: authors suggest
immunostimulation
(Zhang et
al.. 2010)
Human
formaldehyde
melamine
workers
51 Controls: <0.037
mg/m3; 43 Exposed: 1.8
(0.42-6.9) mg/m3;
Duration at least 3
months (41/43 exposed
> 1 year)
Serum immune
markers
22/38 immune/inflammation markers
that were detectable were decreased
Stringent FDR cutoff (10%):
significantly decreased CXCL11 and
CCL17(both -25%)
FDR at 20%: significantly decreased
CRP, TRAIL, SAP, 1L-10, sCD40L, and
Insulin
N/C in TNF-a; other markers below
LOD
High or Medium Confidence
[Note: the strongest correlation
of marker changes was with
monocyte levels (p = 0.05), but
overall the results suggest that
cell counts do not explain the
marker changes]
(Zhang et
al.. 2010)
Human
formaldehyde
melamine
workers
51 Controls: <0.037
mg/m3; 43 Exposed: 1.57
(0.77-6.9) mg/m3;
Duration at least 3
months (41/43 exposed
> 1 year)
Serum cell counts
Proliferation of serum
hematopoietic
progenitor cells
Decreased WBC, lymphocytes,
granulocytes, platelets, and RBC
Increased mean corpuscular volume
N/C in monocytes, hemoglobin
Decreased colony formation in
cultured hematopoietic progenitors
from subjects
High or Medium Confidence
[one ex vivo endpoint: possible
influence of culturing- still
expected to be due to exposure,
but could involve in vitro
amplification of phenomena]
(Jia et al..
2014)
Human plywood
workers (n=118)
and controls
(n=79)
[High] workers: 0.77
(0.44-1.88) mg/m3
(n=70); [Low] workers:
0.18 (0.086-0.23) mg/m3
(n=48); duration >6
Serum lymphocyte
subtypes and
cytokines
Dose-dependent increased % CD19+ B
cells at > 0.18 mg/m3; increased CD56+
NK cells at 0.18 mg/m3 only
N/C in %CD3+, CD4+ or CD8+ T cells
High or Medium Confidence
This document is a draft for review purposes only and does not constitute Agency policy.
A-480 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
months; controls <0.01
mg/m3
Increased IL-10 and decreased IL-8 at >
0.18 mg/m3; Increased IL-4 and
decreased IFNy at 0.77 mg/m3
(Hosgood et
Human
formaldehyde
melamine
workers
51 Controls: 0.032 (0.01-
0.032) mg/m3; 43
Exposed: 1.57
(0.77-3.09) mg/m3;
Duration at least 3
months (41/43 exposed
>1 year)
Serum counts and
analyses of
lymphocyte subsets
Decreased lymphocytes, NK cells, T
cells, and CD8+ T cells
N/C in B cells, or CD4+ T cells (overall;
note: CD4+/FoxP3+ decreased)
T cells subset analyses showed
decreased CD8+ effector T cells and
regulatory T cells
High or Medium Confidence
Note: Authors hypothesized
decreased effector T cells (which
circulate to inflamed tissues)
may reflect decreased response
to antigenic-related
inflammation, and decreased
regulatory cells as decreased
immunosuppression (which may
lead to autoimmunity)
al.. 2013)
Note: Same
cohort as
(Zhang et
al.. 2010)
(Ye et al.,
2005)
Human students
(n=23), waiters
(n= 16), or FA
manufacturers
(n=18)
[High] Manufacturers:
0.985± 0.286 mg/m3 (8.5
yr, 8 hr/d; 1.69
maximum); [Low]
waiters: 0.107± 0.067
mg/m3 (12 wk, 5h/d);
Controls: 0.015 mg/m3
Blood lymphocyte
subset analysis
N/C in waiters exposed to low levels
Increased % B cells and ratio of T
helper to T cytotoxic T cells (CD4/CD8
ratio), and decreased total T cells and
CD8+ T cells in workers exposed to
high levels
High or Medium Confidence
[data not adjusted for age or
gender]
(Bono et al.,
2010)
Human
pathologists
(n=44) and
controls (n=32)
Controls: 0.028±0.0025
mg/m3; Pathologists:
0.032±0.006 or
0.21±0.047 mg/m3 (in
"reduction room");
duration unclear
Serum lymphocyte
ROS (MDA-dG
adducts)
Increased MDA-dG at > 0.066 mg/m3;
N/C in MDA-dG at <0.022 mg/m3 or
0.023-0.066 mg/m3 (significant
association with air-FA levels)
High or Medium Confidence
(unknown duration)
(Romanazzi
et al., 2013)
Human Laminate
workers (males,
yrs employed NR)
0.21±0.10 mg/m3
exposed (n=51);
0.04±0.02 mg/m3
nonexposed (n=54)
15-F2t Isoprostanes in
urine (also measured
cotinine and smoking)
Smoking and air-formaldehyde
exposure were independently
associated with increased IsoP
High or Medium Confidence -
indirect [accuracy of single
measure questionable]
Note: serum and urine
isoprostanes are correlated
[Rodrigo et al., 2007]; thus, this
finding is indirect for serum ROS
(Lvapina et
al.. 2004)
Human workers
with carbamide-
FAglue (n=29)
Exposed workers: 0.87±
0.39 mg/m3 (n=21
Blood neutrophil
oxidative burst
Routine hematology
Significant decreases in neutrophil
function/ oxidative burst were only
detected when comparing the 12
High or Medium Confidence
[mixture exposure]
This document is a draft for review purposes only and does not constitute Agency policy.
A-481 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
nonexposed); duration
mean: 12.7± 9.6 years
Assessment of
chronic URT
inflammation
workers with evidence of URT
inflammation (N/C across full groups)
Decreased erythrocyte count and
hematocrit levels correlated with
duration of exposure (no other
changes)
Note: Authors hypothesized that
decreases in erythrocyte and
hematocrit counts might
indicate FA toxicity on bone
marrow hematopoiesis
(Jakab et
al.. 2010)
Human female
pathologists or
controls (n=37)
0.9 mg/m3 (8hr-TWA
exposure); mean
duration >17 years;
slightly more (not
significant) smokers and
drinkers in exposed
Serum lymphocyte
parameters: CD71 in
fresh cells; apoptosis/
proliferation in cells
cultured with PHA
N/C in T cell activation marker, CD71
Exposure to FA alone increased
apoptosis and 1 out of 3 measures of
cell proliferation in PBLs; N/C % in S
phase
High or Medium Confidence -
CD71 [limited precision of
exposure assessment - sampling
l-3yrs from study]
Low Confidence -other
measures [ex vivo; limited
exposure assess]
(Bellisario
et al.. 2016)
Human nurses
(Italian females,
yrs employed NR)
0.034±0.038 mg/m3
using formalin (n=64);
0.015±0.005 mg/m3 not
using formalin (n=30),
but noting that they did
receive some exposure;
8-hr workshift measures
on 2 separate days
15-F2t Isoprostanes
and malondialdehyde
in urine, normalized
to creatinine (also
measured cotinine)
Smoking and air-formaldehyde
exposure were independently
(positively) associated with increased
oxidative stress biomarkers by
pairwise comparisons and regression
(note: in nurses who used vacuum
sealing techniques, which reduce
formaldehyde exposure, also exhibited
reduced biomarkers).
Low Confidence - indirect
[accuracy of single measure
questionable]; small exposure
differential; formalin test article
Note: serum and urine
isoprostanes are correlated
[Rodrigo et al., 2007]; thus, this
finding is indirect for serum ROS
(Erdei et al.,
2003)
Human (sex NR)
symptomatic
students (9-11 yo
w/ respiratory
issues) (n=176)
0.006-0.057 mg/m3
(average= 0.018);
duration unknown [co-
exposure: NO2, benzene,
toluene, xylene, and dust
mite allergen]
Serum Cell Counts
Increased serum monocyte counts by
linear regression; N/C in RBCs, WBCs,
platelets, lymphocytes, neutrophils
(mostly), or eosinophils (all data NR)
Low Confidence [comparisons
to "normal" range rather than
to control group; co-exposure;
limited reporting]
Note: symptomatic only
(Kuo et al.,
1997)
Human dialysis
nurses (n=51) or
ward nurses
controls (n=71)
Personal sampling
ranged from 0.018-0.11
mg/m3; area sampling
was as high as 3.44
mg/m3 (duration
average= 3 yr; =1/3
employed
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
> 3 yr); control area
levels N/R
(Thrasher et
al.. 1987)
Human sympto-
matic exposed
subjects, controls
(n=8/ group)
Exposed (mobile home
measures): 0.086-0.68
mg/m3 (residency -6-7
yr); nonexposed: not
measured (authors
assume: <0.037)
Serum cell counts
Ex vivo T and B cell
blastogenesis (PHA or
PWM stimulation)
T cell number decreased; B cell counts
were not significantly changed
T cell blastogenesis with PHA (not
PWM: p>0.05, authors call significant)
impaired
Low Confidence [small sample;
symptomatic vs.
nonsymptomatic comparison;
questionable reporting]
(Thrasher et
al.. 1990)
Human various
exposed groups
of patients, and
asymptomatic
controls
"controls"- chiropractic
students (n=28):
assumed > 0.53 mg/m3
for 28wk (13h/wk);
mobile home residents
(n=19): 0.062-0.62
mg/m3 for 2-7 yr; office
workers (n=21): assumed
0.012-0.95 mg/m3,
duration N/R;
occupational (n=8):
levels/ duration N/R;
removed from exposure
for> lyr: 0.17-1.0
mg/m3
Blood cell counts
Decreased WBCs in office workers;
N/C in all T cells, T helper or T
suppressor cells, orT cell H/S ratio
Tal+ lymphocytes (antigenic
stimulation) elevated in all exposed
patient groups
B cells increased in office workers and
removed patients
IL2R+ lymphocytes increased in mobile
home residents and removed patients
Low Confidence [limited
exposure contrast- authors
suggest the only real difference
between asymptomatic control
students and patients is one of
duration of exposure; patients
to nonpatients comparisons
questionable]
(Ying et al.,
1999)
Human anatomy
students (n=23)
0.508± 0.3 mg/m3 for 8
weeks (3 hr/d, 3 d/ wk);
in dormitories: 0.012±
0.003
Serum lymphocyte
subsets
Ex vivo lymphocyte
proliferation (culture
lymphoblast counts)
After exposure compared to before
exposure: Increased % B cells (CD19),
decreased Total T cells (CD3), T helper
(CD4) and T cyto. (CD8) cells; N/C in ex
vivo lymphocyte proliferation rate
Low Confidence [limited
periodicity; some experiments
ex vivo]
Note: internally controlled
(Madison et
al.. 1991)
Human residents,
spill-exposed (n=
41) or unexposed
controls (n=29)
Formaldehyde (PFA):
>2.46 mg/m3 for first 48
hr, then average
dropped to 0.028 mg/m3,
but urea and
methylamines not
measured or corrected
for
Serum cell counts
N/C in WBC, lymphocyte, CD8,
CD8/CD4 ratio, CD19, or CD25 cells
Decreased % CD5+ and % CD4+,
although total counts of these were
unchanged
Increased CD26+ counts and %
Not Informative [mixture
exposure; co-exposures not
corrected for; FA in controls
unmeasu red]
This document is a draft for review purposes only and does not constitute Agency policy.
A-483 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Vargova et
al.. 1992)
Human
Woodworkers
(Czechoslovakia)
Formaldehyde
0.55-10.36 mg/m3 and
other, unquantified
exposures
Serum IgG, IgA, IgM,
IgE
Complement and
other factors
Lymphocyte
proliferation
Increased lymphocyte proliferation to
concanavalin A and decreased
proliferation to phytohaemaglutinin
"no significant differences in natural
cellular and specific humoral
immunity"
Not Informative [mixture
exposure; co-exposures not
corrected for; FA in controls
unmeasured; no description of
recruitment or how referents
were matched- reporting
limited]
(Zhang et
al.. 2010)
Human
formaldehyde
melamine
workers
51 Controls: <0.037
mg/m3; 43 Exposed: 1.57
(0.77-3.09) mg/m3;
Duration at least 3
months (41/43 exposed
>1 year)
In vitro proliferation
of a volunteer's cells
Decreased colony formation in
cultured progenitors with in vitro FA
treatment
Not Informative [formalin
treatment- assumed; single
donor, in vitro; nongaseous
exposure, levels relevance]
Controlled-Exposure Studies in Humans or Primary Human Cells
(Dietrich et
al.. 1996)
In vitro human
leukocytes (single
donor): not
further described
Formalin (assumed; test
article N/R) gas at 0.62
mg/m3 for 1 hour
Heat shock protein 70
levels (Westerns)
FA, but not heat (42°C) stress, caused
a significant increase in HSP70 levels
Not Informative [formalin; in
vitro; short duration; exposure
level relevance unknown;
sample size NR; poor reporting]
Controlled-Exposure Studies in Animals, Animal Cells, or Immortalized Human Cells
(Sorg et al.,
2001a)
Male SD rats
(n=6-9/ group)
PFA (inferred from
citation) 0, 0.86, or 2.95
mg/m3 for 20-60 min, 2
or 4 wk
Serum corticosterone
N/C with acute exposure
Increased CORT at 2.95 mg/m3 at 2 or
4 wk
High or Medium Confidence
Note: unclear utility of endpoint
for respiratory effects
interpretation
(Rager et
al.. 2014)
Male fischer rats
(n=3)
PFA 0 or 2.46 mg/m3 for
7 d, 28 d or 28 d with 7d
recovery (6 hr/d)
miRNA microarray of
blood WBCs
WBCs miRNAs were changed after 7d
or 28d or 28d with recovery (31 or 8 or
3 transcripts); associated primarily
with inflammation and immunity
High or Medium Confidence
[small sample size]
Note: unclear/indirect
interpretation of endpoints
(NTP, 2017)
Male
B6.7?p53tmlBrd
and C3B6.129F1-
Trp53tmlBrd mice
(heterozygote
P53 allele);
n=25/group
PFA 0, 9.23, or 18.45
mg/m3 for 8 weeks (6
hr/d, 5d/wk) with
measures at
approximately 1 year
Whole blood counts
N/C in hematological parameters,
including RBC, WBC, neutropils,
monocytes, eosinophils, platelets,
lymphocytes, reticulocytes,
hemoglobin, hematocrit, MCV, MCH,
or MCHC
High or Medium Confidence
This document is a draft for review purposes only and does not constitute Agency policy.
A-484 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Dean et al.,
1984)
Female B6C3F1
mice (n=10/
group)
PFAOor 18.5 mg/m3 for
3 wk (6 hr/d, 5 d/wk)
Serum cell counts
N/C peripheral blood cell counts,
including WBC differentials, except:
Decreased number of monocytes
(from 43 to 4)
Low Confidence [excessively
high levels: 60-70% RB inferred
at these levels]
Note: monocyte decrease
speculated as peripheral
response to nasal inflammation
and healing
(Avdin et
al.. 2014)
Male SD rats
(n=6/ group)
Test article unclear, but
appears to be formalin in
this experiment at 0,
6.48 (low), 12.3
(moderate), or 18.7
mg/m3 for 4 wk (8 hr/d,
5 d/wk)
Serum total
antioxidant and total
oxidant levels (TAS
and TOS; kit uses
vitamin E and H202 as
reference,
respectively
Serum oxidative
stress index (OSI:
TOS/TAS)
Serum irisin
(hormone- may
regulate obesity)
Increased TOS, and decreased TAS and
irisin, at > 12.3 mg/m3 formaldehyde
Increased OSI at >6.48 mg/m3
Note: serum biochemical parameters
(e.g., cholesterol) are not included
here, but were unchanged. Carnosine
supplementation reduced changes.
Low Confidence [formalin; high
levels]
(Zhang et
al.. 2013)
Male Balb/c mice
(n=9)
Formalin 0, 0.5 or 3
mg/m3 for 2 wk (8 hr/d,
5 d/wk)
Serum cell counts
D/D Decreased serum WBC, RBC, and
lymphocytes, and increased platelets,
at >0.5 FA; decreased intermediate
cells at 0.5 FA; N/C in neutrophils
Low Confidence [formalin]
(Ye et al.,
2013b)
Male Balb/c mice
(n>9/ group/
endpoint)
Formalin 0,0.5,1, or 3
mg/m3 for 7d (8 hr/d)
ROS (dichlorohydro-
flourescein and MDA)
blood mononuclear
cells (PBMC)
Dose-dependent decrease in GSH
levels in PBMC at >1
Dose-dependent increase in DCFH and
MDA in PBMC at 3
Co-administered GSH attenuated
effects
Low Confidence [formalin]
(Im et al.,
2006)
Male SD rats
(n=10)
Formalin (assumed; test
article not specified) 0,
6.15,12.3 mg/m3 for 2
weeks (6 hr/d; 5 d/wk)
Plasma ROS,
cytokines, and
proteomic analysis
Increased MDA & protein carbonyls at
12.3 mg/m3 (note: similar increases in
liver)
D/D Increased IL-4 and decreased IFNy
Other protein changes (e.g, increased
GSTs and ApoE; decreased heme
Low Confidence [formalin; high
levels]
This document is a draft for review purposes only and does not constitute Agency policy.
A-485 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
oxygenase, fibrinogen, ApoAl, SNAP-
25
(Matsuoka
et al.. 2010)
Male ICR mice (n>
7)
Formalin at 0.12 mg/m3
for up to 24 hr; also, a
single experiment at 3.69
mg/m3 for 24 hr
Plasma ROS (80HdG)
and NO (nitrates/
nitrites); NO response
to LPS injection: 3.69
mg/m3
Increased plasma ROS at 0.12 mg/m3
for >8hr and NO at 24hr
Increased plasma SOD activity at 3.69
mg/m3; N/C in plasma IL-6 at 0.12
mg/m3
Decreased N03 with LPS stimulation
Low Confidence [formalin; short
duration]
NOTE: ACUTE
(Sandikci et
al.. 2007b)
SD rats (n=6/
group) at GDI [1],
PND1 [II], PND28
[III] or adults [IV]
Formalin (assumed: test
article NR): 0 or 7.38
mg/m3 for 6 wk (8 hr/d,
7d/wk)
Blood T lymphocyte
counts
Increased blood T lymphocytes
(ANAE+ as marker) in all groups by FA
Low Confidence [formalin; high
exposure levels; use of ANAE as
T lymphocyte marker under all
conditions has been debated]
(Katsnelson
et al.. 2013)
Rat "white"
females (n=12-
15)
Formalin (assumed; test
article NR) 12.8±0.69
mg/m3 for 10 wk (4
hr/d, 5d/wk)
Blood cell counts and
immune markers
(other markers N/C or
not inflammation)
Increased % lymphocytes and albumin;
Decreased % segmented neutrophils,
MDA, GSH, and lymphocyte SDH
activity; some decreased serum amino
acids
Low Confidence [formalin;
excessively high levels; short
periodicity]
(Yu et al.,
2014b)
Male ICR mice
(n=6)
Formalin 20, 40, 80
mg/m3 for 15 d (2 hr/d)
Blood cell counts
Decreased blood WBCs and platelets
at > 40 mg/m3
Low Confidence [formalin;
excessively high levels; short
periodicity]
(Brondeau
et al.. 1990)
Male SD rats
(n=10)
Formalin (assumed; test
article NR) 35.7-75
mg/m3 for 4 hr, with or
without adrenalectomy
Serum cell counts
Decreased WBCs at > 52.9 mg/m3, not
at 35.7 mg/m3; N/C in RBCs
Adrenalectomized rats did not show
decreased WBCs at 60.3 mg/m3
Low Confidence [formalin;
excessively high levels; short
periodicity]
NOTE: ACUTE
(Zhao et
al.. 2020)
Male Balb/c
mice (n=3,
pooled into
single sample
for nose and
lung samples);
2 experiments
by different
researchers
Formalin
0, 3 mg/m3 for 2 weeks
(8 h/d, 5 d/wk)
Burst-forming unit-
erythroid (BFU-E),
and colony-forming
unit-granulocyte
macrophage (CFU-
GM) colonies in
nose, lung, spleen,
and bone marrow
Bone marrow results:
Decreased formation of CFU-GM and
BFU-E in both experiment 1 and II
Low Confidence [formalin;
small sample size]
Not Informative: ex vivo
resu Its
(Wei et al.,
2014)
Male C57BL/6
mice (n=6)
Methanol-free formalin
at 0, 0.5 or 2 mg/kg/day
Serum cytokines for
Thl, Th2, and Thl7
Increased Thl-related cytokines (IFN-y,
TNF, and IL-2), TH2-related cytokines
(IL-4, IL-6, and IL-10), and Thl7-related
Not Informative [levels of
unknown relevance; i.p.
injection]
This document is a draft for review purposes only and does not constitute Agency policy.
A-486 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
for 1 week or 1 month (5
d/wk)
cytokine (IL-17A) at 2 mg/kg/day for 1
or 4 weeks; specific statistically
significant increases only noted for 1
week IL-2 and IL-4 levels (note:
magnitude of change was equal or
greater at 1 month and for all tested
cytokines in all comparisons; in
general, small decreased levels noted
at 0.5 mg/kg)
Note: Kruskal-wallis test
(Ibrahim et
al.. 2016)
Pregnant Wistar
rats (n=5 dams;
10 pups/ group
for experiments;
note: individual
pup data for n=10
pups did not
appear to
account for
litters)
Formalin 0.92 mg/m3
from GDs 1-21: 1 hr/d, 5
d/wk
Blood cell counts and
Myeloperoxidase
activity
Randomly assigned pups all received 5 mg/kg
lipopolysacharride (LPS) injections at PND 30
Increases in total cells and
granulocytes (lymphocytes and
monocytes were unchanged) by LPS
were reduced in offspring exposed to
formaldehyde, as were increases in
myeloperoxidase activity
Not Informative [formalin;
short periodicity; offspring
comparisons do not include FA
without LPS; small sample size;
did not appear to account for
litter effects]
Note: effects rescued by
vitamin C
(Maiellaro
et al.. 2014)
Pregnant Wistar
rats (n=5)
Formalin 0.92 mg/m3
from GDsl-21:1 hr/ d, 5
d/wk
Blood cell counts
Sensitization: s.c. lOug OVA with sc boost after
7d
Challenge: 7 d later, 1% OVA aerosol 15 min/d,
3d
N/C in parental blood total cells,
mono-cytes, lymphocytes, or
granulocytes
Decreased birth weight in offspring
24hr after OVA challenge, offspring
have: decreased blood total cells,
mononuclear cells, neutrophils, and
eosinophils
Not Informative [formalin, short
periodicity, offspring
comparisons do not include FA
alone; small sample size]
(Kum et al..
2007b)
Female SD rats
(n=6)
Formalin (assumed: test
article NR): 0 or 7.38
mg/m3 for 6 weeks (8hr/
d, 7d/wk)
Serum biochemistry
(proteins and factors)
Increased serum urea, but N/C in total
protein, albumin, or creatinine
Note: experiments with FA + xylene
not considered
Not Informative [formalin; high
levels; tests not considered
relevant to inflammation or
respiratory effects]
(Ciftci et al..
2015)
Male Wistar
albino rats (n=10)
Formalin i.p. injection at
9 mg/kg/d every other
day for 2 weeks
Serum markers for
ROS, antioxidants, as
well as beta amyloid
and tumor protein 53
levels
Increased MDA (ROS marker)
Decreased total antioxidants, TP53,
and A-betal-40 (not 1—42)
Not Informative [formalin; high
levels of unknown relevance;
i.p. injection]
This document is a draft for review purposes only and does not constitute Agency policy.
A-487 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Murta et
al.. 2016)
Male Fischer
rats (n=7)
Formalin (assumed)
1%, 5%, or 10% for 5d
(3 x 20min/d)
Blood cell counts,
chemokine levels,
and ROS indicators
FA increased total leukocyte,
lymphocytes at 5%, and decreased
platelets at 10%; N/C in other cell
types; 1% caused increased catalase
and other ROS indicators were
observed; increased CCL2 at 10%,
CCL3 at 1-5%, and CCL5 at 1%
Not Informative [formalin;
unquantified high levels;
static exposure chamber;
short periodicity]
(da Silva
et al.,
2015)
Male Wistar
rats (n=6/
group)
Formalin 1% for 3 days
(90 min/ d); rats
exposed in static
chambers 5 rats/time
Blood cell counts
FA increased total cells, monocytes,
lymphocytes, and neutrophils
Note: while reduced effects were
reported as reduced with laser
therapy, laser therapy-only controls
were not used
Not Informative [formalin;
unquantified high levels;
static exposure chamber and
group exposure; short
duration and periodicity]
(Lino dos
Santos
Franco et
al.. 2006)
Male Wistar rats
(n=5-6)
Formalin 1% or methanol
vehicle for 4 days (30,
60, or 90min/ d)
Serum cell counts
Increased serum leukocytes and
mononuclear cells, but not neutrophils
Not Informative [formalin
(MeOH controls); unquantified
high levels; short periodicity;
small sample size; presented
comparisons to naive rats
rather than MeOH controls]
(Lino-Dos-
Santos-
Franco et
al.. 2011a)
Female Wistar
rats (n=5)
Formalin 1% or naive for
3 days (90min/ d), with
or without ovariectomy
Serum cell counts and
factors
Increased total serum leukocytes
Increased serum corticosterone
Not Informative [formalin;
impact of sham surgery NR;
short periodicity and duration;
unquantified high level; FA
alone untested; naive not
chamber exposed; small sample
size]
(Lino dos
Santos
Male Wistar rats
(n=5)
Formalin 0,1% for 3 days
(90 min/d)
Serum cell counts
Increased Total serum leukocytes and
mononuclear cells, not neutrophils; FA
inhibited OVA-induced increases
Not Informative [formalin;
unquantified high level; small
sample size; short duration and
periodicity]
Franco et
al.. 2009)
Sensitization: immediately post-FA, i.p. 10 ng
OVA; boost 1 wk later with s.c. injection
Challenge: 1 wk later with aerosolized OVA
This document is a draft for review purposes only and does not constitute Agency policy.
A-488 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-71. Effects on other immune system-related tissues (e.g., bone marrow, spleen, thymus, lymph
nodes, etc.)
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
Controlled-Exposure Studies in Animals. Animal Cells, or Immortalized Human Cells
(Fuiimaki et
al.. 2004b)
Female C3H mice
(n=5-6 per group)
PFA 0,0.098, 0.49, or
2.46 mg/m3; 12 wks
Splenic Cell counts
Ex vivo splenic cells
No significant change in counts of
splenic CD3 T cells, CD19 B cells, or
CD4/CD8 ratio
D/D Increased IFNy with LPS
stimulation of cells at 2.46 mg/m3
D/D Increased MCP-1 at > 0.49 mg/m3
in cells of OVA-stimulated mice; N/C in
IFNy, MlP-la or IL-5
Body weight decreased at >0.49
mg/m3
High or Medium Confidence
[small sample size]: cell counts
Low Confidence [small sample
size; ex vivo]: cytokine
measures
Sensitization: i.p. 10 ng OVA prior to FA
exposure; aerosol OVA boost for 6 min on wks
3, 6, 9, and 11
(Rager et
al.. 2014)
Male Fischer rats
(n=3)
PFA 0 or 2.46 mg/m3 for
7 d, 28 d or 28d with 7 d
recovery (6 hr/d)
miRNA microarray of
femur BM cells
N/C in BM miRNAs at any time
High or Medium Confidence
[small sample size]
NOTE: indirect interpretation of
endpoints
(Ma, 2020,
7017056)
Male BALB/c mice
(n=8)
Methanol-free formalin 0
or 2 mg/m3 for 8 weeks
(8h/d, 7d/w)
T cells in the spleen
(mature) and thymus
(immature)
Spleen: Decreased CD8+ and increased
CD4/CD8 ratio; N/C in organ weight
and CD4+ cells
Thymus: Increased CD4/CD8 ratio ;
Decreased organ weight and CD8SP
cells; N/C in CD4SP cells
High or Medium Confidence:
counts
NOTE: experiments in directly
treated cells considered Not
informative for these endpoints
(not extracted)
(Park et al.,
2020)
Female BALB/c
mice (n=10)
Fresh formaldehyde
solution (methanol-free)
0,1.38, 5.36 mg/m3 for 2
weeks (4 h/d, 5 d/wk)
Splenic cytokines, T
cell populations and
Thl/Th2 balance,
differentiation
markers
Spleen: N/C in CD4+ T helper cells, D/D
increased T reg cells
(CD4+CD25+Foxp3+) subset of CD4+
cells; Increased calcinurin and NFAT1
(regulatory and inhibitory functions),
N/C in NFAT2
Spleen (ex vivo production): D/D
decreased IL-4, IL-5, IL-13, IFN-g, IL-
17A, and IL-22 with similar changes in
mRNA for same; [also, N/C in relative
spleen wt. and increased rel. lung wt.
at 5.36 mg/m3]
High or Medium Confidence
[small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-489 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Dean et al.,
1984)
Female B6C3F1
mice (n=6-10/
group/ endpoint,
except n=5 for
splenocyte
assays)
PFA 18.5 mg/m3 for 21d
(6 h/d, 5 d/wk)
Lymphoid organ
weights/ cellularity
Host immunity
response
N/C in thymus or spleen weight; N/C in
BM cells/ femur or spleen cell counts;
N/C in CFU in spleen or BM; N/C in
splenic lymphocyte proliferation or
splenic B cell IgM production
N/C in cell-mediated immunity
(response of spleen lymphocytes to
mitogens, splenocyte cell surface
markers, NKcell cytotoxicity) or
humoral immunity (number of IgM Ab-
producing B cells for 3 separate
antigens)
Decreased host susceptibility to
bacteria challenge, but not tumor
challenge; N/C in hypersensitivity or
NK cytotoxicity
Low Confidence [excessively
high levels small sample size;
some experiments ex vivo]
NOTE: 60-70% RB inferred
(Liu et al.,
2017)
Male ICR mice (n=
10/group)
Unspecified test article
0,1,10 mg/m3 for 20 wk
(2 h/d)
Bone marrow (BM)
polychromatic
erythrocytes
(PCE)/normochromati
c erythrocyte (NCE)
ratio
Dose-dependent decrease in BM
PCE/NCE ratio (markers of
immature/mature RBCs), significant at
>1 mg/m3
Low Confidence [presumed
formalin]
(Ye et al.,
2013b)
Male Balb/c mice
(n>9/ group/
endpoint)
Formalin 0, 0.5,1, or 3
mg/m3 for 7 d (8 hr/d)
ROS (dichlorohydro-
flourescein and MDA)
and GSH in BM and
Spleen
Dose-dependent decrease in GSH
levels in BM and spleen at >1
Dose-dependent increase in DCFH and
MDA in BM and spleen at >1
Co-administered GSH attenuated
effects on GSH, DCFH and MDA in all
tissues
Low Confidence [formalin]
(Zhang et
al.. 2013)
Male Balb/c mice
(n=9)
Formalin 0, 0.5, or 3
mg/m3 for 2 wk (8 hr/d,
5 d/wk)
BM ROS and
cytokines/factors
BM increased megakaryocytes at >0.5
FA
BM ROS (DCFH-DA) D/D increased at
>0.5 FA; GSH decreased, and caspase-3
increased, at 3 FA; BM NFkB, TNFa,
and IL-ip increased at 3 FA
Low Confidence [formalin]
This document is a draft for review purposes only and does not constitute Agency policy.
A-490 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Zhao et al.,
2020)
Male Balb/c mice
(n=3, pooled into
single sample for
nose and lung
samples); 2
experiments by
different
researchers
Formalin
0, 3 mg/m3 for 2 weeks (8
h/d, 5 d/wk)
Burst-forming unit-
erythroid (BFU-E),
and colony-forming
unit-granulocyte
macrophage(CFU-
GM) colonies in nose,
lung, spleen, and
bone marrow
Spleen results:
Decreased formation of CFU-GM in
both experiment 1 and II
Decreased formation of BFU-E in
experiment II; N/C in experiment 1
Low Confidence [formalin; small
sample size]
Not Informative: ex vivo results
(Gu et al.,
2008)
Female Balb/c
and C3H/He mice
(n=10 for in vivo;
n=3 ex vivo
experiments)
Formalin (assumed; test
article NR) 0.098 mg/m3
for 5 wk (in vivo); 0.12,
or 0.98 mg/m3 for 5 wk
(ex vivo); both 24 h/d, 5
d/wk
Splenic cell
phenotypes
Ex vivo cytotoxicity
N/C in T cell or B cell subtypes at 0.08
Increased NK1 cells (NK1.1 expression)
at 0.098 mg/m3
Increased ex vivo NK1 cell cytotoxicity
at >0.12 mg/m3
Low Confidence [formalin]
Not Informative [small sample
size; ex vivo; unclear reporting:
ex vivo cytotoxicity
(Dallas et
al.. 1987)
Male SD rats
(n=2/time point;
unclear
reporting)
PFA 0, 0.62, 3.69, or 18.5
mg/m3 for 1 wk to 24 wk
(6 h/d, 5 d/wk)
Flow cytometry
DNA/RNA analysis of
BM cell proliferation/
health
N/C in RNA or DNA measures (e.g., % S
phase) in BM cells
Low Confidence [small sample
size; unclear reporting]
NOTE: indirect utility for
evaluating respiratory effects or
inflammation
(Kim et al.,
2013b)
Female NC/Nga
(atopic-prone)
mice (n=5-
6/group)
Formalin (assumed; test
article NR) 0, 0.25,1.23
mg/m3 for 4 wk (6hr/d,
5d/wk)
Cytokine mRNA for
spleen
Spleen mRNA: FA D/D increase IL-13
only
With HDM, FA exacerbated IL-4 (0.2),
IL-5 (1.23 mg/m3), IL-13 and IL-17A
(>0.25 mg/m3), but caused D/D
decreased IFNy (>0.25 mg/m3)
Low Confidence [small sample
size; unclear reporting]
NOTE: indirect utility for
evaluating respiratory effects or
inflammation
Sensitization: topical house dust mite (HDM;
ear) stimulation (25 mg Df ointment) lx/wkfor
4 wk
(Kim et al.,
2013a)
Female C57BL/6
mice (n=5
"experiments";
number of mice/
group unclear)
Formalin (assumed; test
article NR) 0, 6.15, or
12.3 mg/m3 2-3 wk (6
hr/d, 5 d/wk)
Spleen and bone
marrow cell counts
Ex vivo cellular
functional assays
N/C in absolute cell number or T cell or
B cells subtypes in spleen or BM; No
significant changes in %CD8 or % B
cells in spleen
Decreased NK1 cells in spleen,
including reduced function, which was
inhibited at 12.3 mg/m3: duration-
dependent
Low Confidence [formalin;
unclear, low sample size; high
levels]
Not Informative: ex vivo
function
This document is a draft for review purposes only and does not constitute Agency policy.
A-491 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Yu et al.,
2014b)
Male ICR mice
(n=6)
Formalin 20, 40, 80
mg/m3 for 15 d (2 hr/d)
BM histology, cell
counts and ROS
Decreased BM cells observed by
pathology and GSH-Px activity at >40
FA
Increased MPO activity and protein
and decreased Prx2 protein at >20 FA
Decreased BM cells (karyocytes) and
CFUs and MMP levels at 80 mg/m3
D/D increased BM oxidative stress
(MDA increased and SOD decreased)
>20 FA
Increased BM apoptosis markers >40
FA
Low Confidence [formalin;
excessively high levels; short
periodicity]
(Yu et al.,
2015a)
Male mice (strain
NR; n=6/group)
Formalin 0, 20, 40, 80
mg/m3 for 15 days (2
h/d)
BM H202 production,
caspase and
antioxidant enzyme
levels/ activity, and
apoptosis
Increased ex vivo caspase-3 activity,
peroxiredoxin levels and H202
production at >20 mg/m3
Increased apoptosis at >40 mg/m3
Low Confidence [formalin-
excessively high levels; short
periodicity]
(De Jong et
al.. 2009)
Male Balb/c mice
(n=6)
Formalin 3.6 mg/m3
nose-only (up to 360
min/d for 3 d)
Ex vivo cytokine
production from
isolated lymph nodes
No cell proliferation in LNs
N/C in IL-4, IL-10, or IFNy production
from isolated cells by FA alone, but FA
with sensitization results in increased
IL-4 and IL-10 (and slight increase in IL-
12), but N/C in IFNg
Low Confidence [formalin; short
duration and periodicity; ex
vivo]
(Zhang et
al.. 2014a)
Balb/c mice
(n=3/sex/group)
Formalin 0, 4, 8 mg/m3
for 7 days (6 h/ d)
Spleen and thymus
weights
Ex vivo spleen cell
lymphocyte
proliferation and ROS
Urine metabolomics
Decreased relative spleen and thymus
weights (only statistically significant
for thymus at 8 mg/m3)
Decreased ex vivo lymphocyte
proliferation and SOD activity at >4
mg/m3 and increased ex vivo ROS at 8
mg/m3
Low Confidence [formalin; ex
vivo; no chamber control
exposure; lowest tested
exposure of 4mg/m3]
Note: some ex vivo assays after
in vivo exposure; n=6 (pooled
sexes assumed- not explicit in
reporting)
(Fuiii et al.,
2005)
Female Balb/c
mice (n=6-10)
Formalin (assumed; test
article NR) 0, 0.25
mg/m3; exposed during
elicitation (reporting
unclear) or sensitization
Ex vivo lymph node
cells all w/
epicutaneous
trinitrochlorobenzene
TNCB
During elicitation: FA increased CD4+ T
cells (IL-4+: Th2, not IFNy+: Thl), not
CD8+, in draining lymph node (LN)
Not Informative [formalin; ex
vivo; reporting for some
experiments unclear; No FA-
only controls; short duration]
This document is a draft for review purposes only and does not constitute Agency policy.
A-492 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(4 wk) or w/ chronic
hypersensitivity
During sensitization (and in CH model):
FA increased LN CD8+ T cells (N/C
CD4+; CD4+CD25+/CD4+ decrease)
(da Silva et
al.. 2015)
Male Wistar rats
(n=6/ group)
Formalin 1% for 3 days
(90min/ d); rats exposed
in static chambers 5 rats/
time
Bone marrow cell
counts
FA caused N/C in total bone marrow
cells
Note: while reduced effects were
reported as reduced with laser
therapy, laser therapy-only controls
were not used
Not Informative [formalin;
unquantified high levels; static
exposure chamber and group
exposure; short duration and
periodicity]
(Ibrahim et
al.. 2016)
Pregnant Wistar
rats (n=5 dams;
10 pups/group
for experiments;
design did not
appear to
account for
potential litter
effects)
Formalin 0.92 mg/m3
from GDs 1-21: 1 hr/d, 5
d/wk
Total cells in femur
lavage
Decreases in total cells by LPS were
further reduced in offspring exposed
to formaldehyde
Randomly assigned pups all received 5mg/kg
lipopolysacharride (LPS) injections at PND 30
Not Informative [formalin;
short periodicity; offspring
comparisons do not include FA
without LPS; small sample size;
did not appear to account for
litter effects]
Note: effects rescued by
vitamin C; effects on dam
uterine tissue not included in
these tables
(Lino dos
Santos
Franco et
al.. 2009)
Male Wistar rats
(n=5)
Formalin 0,1% for 3 days
(90 min/d)
BM cell counts
N/C in total BM cells; FA inhibited
OVA-induced increases)
Sensitization: immediately post-FA, i.p. 10 ng
OVA; boost 1 wk later with s.c. injection
Challenge: 1 wk later with aerosolized OVA
Not Informative [formalin;
unquantified high levels; small
sample size; short duration and
periodicity]
(Lino-Dos-
Santos-
Franco et
al.. 2011a)
Female Wistar
rats (n=5)
Formalin 1% or naive for
3 days (90 min/ d), with
or without ovariectomy
Bone marrow cell
counts
Decreased total bone marrow cells
Not Informative [formalin;
impact of sham surgery;
unquantified high levels; FA
alone untested; naive not
chamber exposed; small sample
size; short duration &
periodicity]
(Lino dos
Santos
Franco et
al.. 2006)
Male Wistar (n=5-
6)
Formalin 1% or methanol
vehicle for 4 days (30,
60, or 90 min/d)
Splenic and bone
marrow cell counts
Increased total splenic cells, but total
bone marrow cells unchanged
Not Informative [formalin
(MeOH controls); unquantified
high levels; small sample size;
short duration and periodicity;
This document is a draft for review purposes only and does not constitute Agency policy.
A-493 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
comparisons to naive rats
rather than MeOH controls]
(Golalipour
et al., 2008)
Wistar albino rats
(n=7; sex N/R)
Mixture (dissection room
vapor of undocumented
composition) =1.85
mg/m3 for 18 wk: 2 hr/d
for 2 d/wk, 4 d/wk, or 4
hr/d for 4 d/wk
Spleen morphometry
Frequency-dependent increases in
white pulp diameter and marginal
zone diameter
Not Informative [mixture
exposure; short periodicity;
poor reporting; controls do not
account for co-exposures;
quantitative comparisons for
results NR]
Table A-72. Effects on other tissues (data extracted for possible future consideration, but not included in the
current analyses)
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Fuiimaki et
al.. 1992)
In vitro Male SD
Rat peritoneal
mast cells (n=3+
experiments)
PFA 0, 1.23,6.15, 12.3,
61.5 mg/m3 for 30 min;
stimuli: substance P,
A23187 (increases
cellular Ca2+ and NO
production), and ant-rat
IgE (in sensitized cells)
Peritoneal mast cell
Histamine release
Enhanced histamine release
stimulated by A23187 and anti-lgE at >
6.15 mg/m3; enhanced release by
substance P at 61.5 mg/m3 (note:
release was inhibited by PLA2
inhibition, but not by antioxidant or
dexamethasone)
Excluded (not tissues of interest)
[In vitro; questionable relevance
of peritoneal cells and exposure
levels]
(Fujii et al.,
2005)
Female Balb/c
mice (n=6-10)
Formalin (assumed; test
article NR) 0,0.25
mg/m3; exposed during
elicitation (reporting
unclear) or sensitization
(4 wk) or w/ chronic
hypersensitivity (CH)—all
w/ epicutaneous
trinitrochlorobenzene
Ear swelling, skin
histopathology
During elicitation: FA suppressed
contact hypersensitivity (i.e.,
decreased ear swelling and edema)
During sensitization (and in CH model):
FA increased swelling, edema, and
mast cell infiltration
Excluded (not tissues of interest)
[Formalin; reporting for some
experiments unclear; No FA-only
controls; endpoint relevance
unclear]
(Dean et al.,
1984)
Female B6C3F1
mice (n=5-10/
group/endpoint)
PFA 18.5 mg/m3 for 21 d
(6 h/d, 5 d/wk)
Peritoneal
macrophage function
N/C in peritoneal macrophage
function, except: FA-increased H202
production by macrophages isolated
after injection with MVE-2 and
stimulation with PMA
Excluded (not tissues of interest)
[Excessively high exposure
levels; small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-494 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Adams et
al.. 1987)
Female B6C3F1
mice (n=10)
PFA 18.5 mg/m3 for 3 wk
(6 h/d, 5d /wk)
Peritoneal
macrophage counts
and function (some in
ex vivo cultures)
N/C in macrophage number or
phagocytosis of antibody-covered
erythrocytes; FA decreased leucine
aminopeptidase expression
FA increased release of ROS in
response to external challenge (MVE-2
priming and PMA stimulus); N/C w/o
challenge
Excluded (not tissues of interest)
[Excessively high levels]
(Kim et al.,
2013b)
Female NC/Nga
(atopic-prone)
mice
Formalin (assumed; test
article NR) 0, 0.25,1.23
mg/m3 for 4 wk (6 hr/d,
5 d/wk)
Atopic dermatitis/
Ear skin Inflammation
Cytokine mRNA for
ear skin
FA increased AD-like clinical skin
inflammation by HDM, but not FA
alone
Mast cell infiltration in dermis by FA
alone, exacerbates HDM eosinophil &
mast cell
Skin mRNA: 0.25 mg/m3 increased IL-
13,IL-17A, COX-2; with HDM, FA
exacerbated these and IFNy, IL-4, and
TSLP; N/C IL-5
Excluded (not tissues of interest)
[Formalin; small sample size]
Note: unclear utility for
evaluating respiratory effects or
inflammation;
multiple supplementary files;
eosinophil data not reported
Sensitization: topical house dust mite (HDM;
ear) stimulation (25 mg Df ointment) lx/wkfor
4 wk
(Maiellaro
et al.. 2014)
Pregnant Wistar
rats (n=5)
Formalin 0.92 mg/m3
from GD1-GD21: lhr/d,
5d/wk
Uterine factors
Decreased uterine IL-10, SOD2, and
cNOS, and increased COX-1, at birth
(N/C in IL-6, IL-4, IFNy, COX-2, iNOS,
SOD1, or catalase)
Decreased birth weight in offspring
Excluded (not tissues of interest)
[Formalin, short duration,
offspring comparisons do not
include FA alone]
Sensitization: s.c. lOug OVA with sc boost after
7 d
Challenge: 7 d later, 1% OVA aerosol 15 min/d,
3d
(Avdin et
al.. 2014)
Male SD rats
(n=6/group)
Test article unclear, but
appears to be formalin
in this experiment at 0,
6.48 (low), 12.3
(moderate), or 18.7
mg/m3 for 4 wk (8 hr/d,
5 d/wk)
Liver tissue total
antioxidant and total
oxidant levels (TAS and
TOS; kit uses vitamin E
and H202 as reference,
respectively
Liver tissue apoptotic
index and oxidative
stress index (OSI:
TOS/TAS)
Increased TOS and decreased TAS, at >
12.3 mg/m3 formaldehyde
Decreased irisin and increased OSI at
>6.48 mg/m3
Note: Carnosine supplementation
reduced changes.
Excluded (not tissues of interest)
[Formalin; high levels]
This document is a draft for review purposes only and does not constitute Agency policy.
A-495 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Bakar et
al.. 2015)
Male Wistar
albino rats (n=7)
i.p. Formalin every other
day at lmg/kg/day for 14
days
Kidney biochemistry,
immunoreactivity for
Bcl-2 and Bax, ROS
and antioxidant
markers, and
electron microscopy
Increased Bcl-2 and Bax
immunostaining, and increased ROS
markers and altered antioxidant
enzyme activities; kidney damage and
inflammation noted
Excluded (not tissues of interest)
[Formalin; high levels of
unknown comparability to
inhaled levels; i.p. injection]
(Matsuoka
et al.. 2010)
Male ICR mice (n>
7)
Formalin at 0.12 mg/m3
for up to 24 hr; also, a
single experiment at 3.69
mg/m3 for 24 hr with LPS
Urine, liver, brain
ROS (80HdG) and NO
metabolites
(nitrates/ nitrites)
Decreased ROS in urine and liver; N/C
in brain; decreased NO in urine, liver
and brain at 0.12 mg/m3 at 24 hr
Increased urinary SOD activity:3.69
mg/m3
Excluded (not tissues of interest)
[Formalin; short duration]
(Kum et al.,
2007b)
Female SD rats
(n=6/group) at
GDI [1], PND1 [II],
PND28 [III] or
adults [IV]
Formalin (assumed: test
article NR): 0 or 7.38
mg/m3 for 6 wk (8 hr/d, 7
d/wk)
Liver oxidative stress
(i.e., SOT, CAT, GSH,
MDA)
CAT activity and MDA levels increased
[1]
GSH decreased in [II]
SOD activity decreased [III]
N/C in adult [IV] oxidative stress
markers
Note: body and liver weight decreased
in 1 and II; liver weight increased in III
Excluded (not tissues of interest)
[Formalin, high levels; limited
assays]
(Kum et al.,
2007b)
Female SD rats
(n=6/ group)
Formalin (assumed: test
article NR): 0 or 7.38
mg/m3 for 6 wk (8 hr/d, 7
d/wk);
Renal oxidative stress
N/C in renal SOD, CAT, GSH-Px, GSH, or
MDA by FA alone
Excluded (not tissues of interest)
[Formalin, high levels; limited
assays]
(Ciftci et al.,
2015)
Male Wistar
albino rats (n=10)
Formalin i.p. injection at
9 mg/kg/d every other
day for 2 weeks
Brain and urine
oxidative DNA
damage
Beta amyloid in brain
Increased A-betai-42 in brain
Increased brain DNA8-Ohdg damage;
slightly increased (nonsignificant-
assumed) DNA damage in urine
Excluded (not tissues of interest)
[high levels of unkown
relevance; i.p. injection;
formalin]
(Ye et al.,
2013b)
Male Balb/c mice
(n>9/ group/
endpoint)
Formalin 0, 0.5,1, or 3
mg/m3 for 7 d (8 hr/d)
ROS (dichlorohydro-
flourescein and
MDA) and GSH in
Liver and Testes
D/D decrease in GSH levels in liver at
>0.5 mg/m3; decreased in testes at 3
mg/m3
D/D increase in DCFH and MDA in liver
at >0.5 mg/m3; in testes at >1 mg/m3;
co-administered GSH attenuated
effects on GSH, DCFH and MDA in all
tissues
Excluded (not tissues of interest)
[Formalin]
This document is a draft for review purposes only and does not constitute Agency policy.
A-496 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Study
System
Exposure
Endpoint(s)
Results
Utility and Notes*
(Jiang et al.,
2015)
In vitro PC12
(immortalized
neuronal) cells
(n=3 technical
replicates)
Formalin (assumed; test
article NR)—in vitro
levels of unknown
relevance
Viability,
neurotrophic factor,
and ROS markers
Decreased BDNF and viability
Increased MDA and other ROS markers
Excluded (not tissues of interest)
[Formalin, high levels of
unknown relevance; in vitro;
small sample size]
(Kim et al.,
2013a)
Female C57BL/6
mice (n=5
"experiments";
number of mice/
group unclear)
Formalin (assumed; test
article NR) 0, 6.15, or
12.3 mg/m3 2-3 wk (6
hr/d, 5 d/wk)
liver cell counts
Ex vivo cellular
functional assays
N/C in absolute cell number or T cell or
B cells subtypes in liver
Excluded (not tissues of interest)
[Formalin; unclear sample size]
(Giilec et
al.. 2006)
Wistar albino rats
(n=10; sex NR)
PFA 0, 12.3 or 24.6
mg/m3 (8 h/d, 5 d/wk) for
4 or 13 wk
Heart oxidative stress
(i.e., SOD, CAT,
TBARS, NO)
Increased SOD at > 12.3 mg/m3 (4 or
13 wk); Decreased CAT at > 12.3
mg/m3 at 4 wks, but not 13 wk; N/C in
TBARS or NO
Excluded (not tissues of interest)
[excessively high levels; limited
assays]
(Xin et al.,
2015)
HepG2 (liver)
cells; n=3
technical
replicates
Formalin; in vitro
(unknown relevance)
Heat shock protein
reporter assays
Increased promotion of HSPA1,
correlated with oxidative stress and
cellular damage
Excluded (not tissues of interest)
[in vitro; high levels; formalin;
small sample size]
This document is a draft for review purposes only and does not constitute Agency policy.
A-497 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde—Inhalation
Synthesis of the identified mechanistic evidence by tissue compartment
The most likely initial effects of formaldehyde exposure include evidence of direct
interactions of formaldehyde with biological macromolecules (e.g., DNA; receptors; redox proteins)
in the upper respiratory tract (URT). These direct interactions would typically not be expected to
occur in other tissue compartments given the lack of substantial distribution of inhaled
formaldehyde to distal sites (see Appendix A.2). While stress hormone increases likely involve
prior modification of the hypothalamic-pituitary-adrenal (HPA) axis, slight evidence of this change
is indicated as a plausible initial effect of exposure due to a general lack of knowledge of the specific
type of stressor(s) (e.g., direct responses due to subtle changes in fear or anxiety; indirect effects
via sustained inflammation) and the nature of the interactions with the HPA axis that might result
from formaldehyde inhalation. The slight evidence of indirect evidence for sensory nerve
stimulation in the LRT is not indicated as a plausible initial effect of exposure because inhaled
formaldehyde is unlikely to reach the LRT in appreciable amounts and it is expected that LRT
sensory nerve activation would be reliant on a secondary response to TRP channel-activating
stimuli such as increased LRT oxidative stress or inflammatory mediators; although, certain
exposure scenarios (e.g., after exposure to high levels of formaldehyde or mouth breathing during
exercise, perhaps only in susceptible individuals) might, in rare scenarios, result in distribution of
minimal amounts of formaldehyde to upper regions of the LRT (see Appendix A.2) that may be
sufficient to induce such receptor-mediated events. Although it is difficult to disentangle the
multiple mechanistic events manifested soon after formaldehyde inhalation, it appears that
formaldehyde can initiate overlapping events in the URT, including effects at the level of the
respiratory epithelial cells and overlying mucociliary layer, as well as at trigeminal nerve endings.
While uncertainties remain17, the effects in the lower respiratory tract (LRT), blood, and other
organs are likely secondary to the changes observed in the URT. Figures A-33 and A-34 illustrate
the potential relationships between the mechanistic events reported from formaldehyde exposure,
based on the more reliable evidence (see Figure A-33) or including evidence that should be
interpreted with greater caution (see Figure A-34). These figures are based on evidence
summarized in Tables 1-66-1-72, and they are discussed according to tissue compartment in the
sections below.
Figures A-33 and A-34 (on the following pages) present network summaries of mechanistic
data related to potential noncancer respiratory health effects of formaldehyde. These figures
present an integrated picture of the mechanistic events identified from studies of formaldehyde
exposure. The figures are organized by tissue type or region (i.e., upper respiratory tract, "URT";
17 Controlled human exposure studies observed pulmonary function deficits when a longer exercise
component (15 minutes) was included. These deficits were not observed by other studies with shorter
periods or no exercise (Green et al„ 1989: Green et al.. 1987). and another study observed airway
hyperresponsiveness with an exposure protocol using nose clips requiring mouth-only breathing (Cassett et
al., 2006).
This document is a draft for review purposes only and does not constitute Agency policy.
A-498 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Toxicological Review of Formaldehyde—Inhalation
lower respiratory tract, "LRT"; "blood"; and other tissues related to immunological responses,
"other"), the data for which are summarized in the following subsections. Figure A-33 presents
events interpreted with greater confidence (i.e., robust or moderate evidence), while Figure A-34
includes events based on slight evidence. In both figures, the mechanistic events and the
relationships between events are characterized as defined in Table 1. Lines with arrows on both
ends indicate events for which the association appears to be bidirectional. The figures also identify
events that are "plausibly an initial effect of exposure," and each event is related to one or more
"key features of a potential hazard" (see explanations above). Note: Some events and relationships
are not shown for clarity, but nearly all mechanistic events from Tables 1-66-1-72 for which at
least slight evidentiary support was concluded are presented. Note that "decreased pulmonary
function" encompasses a range of possible contributing effects including, but not limited to,
increased bronchoconstriction, flow limitation, and decreased bronchodilation.
This document is a draft for review purposes only and does not constitute Agency policy.
A-499 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
URT Changes
4, Mucociliary Mucus membrane
function/555^ change
Epithelial
proliferation
^ Granulocytes
Squamous
metaplasia-
Epithelial
damage ,
TRPA1
stimulation
Protein/DNA^*^
modification
1 *' *
— t Oxidative? /
stress v .J
f LRT
ynfection
Centrally mediatedTr'6em'na'nerve
sensory irritation stimulation
i i
-ex.
-T1 Neuro"
Legend
Plausibly an initial
effect of exposure
Q Robust
( ) Moderate
( ; Slight
Robust
--> Moderate
Slight
I Key feature of a
I potential hazard
Other
peptides
LRT Changes and Airway Function
4/ Pulmonary Function
iP*--.
'T1 Oxidative n
J—^ > I v
( \ stress /;
W
* ft
s
^ Eosinophils \
(& total cells)
't5 Edenja/
inflammatory
structural change
^ Microvascular1'. )
leakage
^Airway hyper
responsiveness
Amplified
response (e.g.,
with allergen)
"-v' .Sustained
x y inflammatioi)''
/ in airway /
Allergic
sensitization
o-
4, RBCs
a
Altered
B Cells
D
O
O
4/ Total
WBCs
'f' Oxidative
stress
o
Altered antibody
responses
(primarily IgG)
4- CD8+T cells
CD4/CD8)
o
f IL-4; 4- IFNy
Blood Changes
Figure A-31. Mechanistic events for respiratory effects of formaldehyde based on robust or moderate evidence.
This document is a draft for review purposes only and does not constitute Agency policy,
A-500" DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
URT Changes t urt infection
4, Mucociliary Mucus membrane /
function^?51^ change
Epithelial
proliferation
Granulocytes
Squamous
metaplasia/
Epithelial
damage ,
Protein/DNA\^ *
modification
TRPA1
stimulation
' | 1 f-
'T> Oxidativei^'W'' /
stressV^/
Centrally mediatedTr'Sem'na' nerve
sensory irritation
stimulation
Legend
Plausiblyan initial
effect of exposure
1Q1 Robust
( ) Moderate
( ; Slight
—> Robust
--> Moderate
Slight
I Key feature of a
I potential hazard
t LRT
^ infection
LRT Changes and Airway Function
4- Pulmonary Function
, _..iuauve
,Q
epithelial
damage!,
Th2-related
cytokines
/
>1 f' Eosinophils
Oxidative
stress
A
(&total cells)
eosinophil
sensory nerve,^
stimulgtitih ! recruitment
& factors'
-¦ t CDS" 7
cells
; /f> Eden^a/
inflammatory
structural change
1" neutrophils &
monocytes (with
antigen only)
Airway hyper
responsiveness
't' NeuroTpeptides
(and NKRI activation)
^ Microvascular
leakage
"Amplified
response (e.g.,
with allergen)
r-
"jJ \ Sustained
V_ J inflammation'
/ in airway /'
m
altered NK cells
'T bone marrow &/" \
spleen oxidative stress ^ _J
¦j" splenic & lymph
cytokine responser - -.
Other
! ,<" T
4, neutrophils
4- chemo-
attractants
o
4, RBCs
' ;
* *
4, platelets
Altered -
B Cells
Allergic
sensitization
o
/t1 Oxidative
\ stress
Ck \
4 Total
i * ;
.--O
.--'''4 CDS"! cells
(f CD4/CD8)
Altered antibody
responses
(primarily IgG)
WBCs
f' stress
hormone
o
f IL-4; 4, IFNy
Blood Changes
Figure A-32. Mechanistic events for respiratory effects of formaldehyde based on robust, moderate, or slight
evidence.
This document is a draft for review purposes only and does not constitute Agency policy,
A-501.' DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
Changes in the URT
Data on formaldehyde-induced mechanistic changes in the URT are largely based on studies
in experimental animals or acutely exposed human volunteers, as most of these endpoints are
difficult to examine in long-term observational epidemiology studies. The specific studies and
summary findings supporting the synthesis below are described in Table 1-73. While the structure
and function of the URT across species is considered similar, interpretation of compensatory or
adaptive changes within the human URT following long-term exposure based on findings in
experimental animals is difficult to infer.
The majority of the events which are potential initial or direct effects of formaldehyde
(see asterisks in Figure A-33) occur at the level of the respiratory epithelium, including evidence
supporting the involvement of formaldehyde in reactions with cellular macromolecules such as
proteins (e.g., detoxifying enzymes) and DNA, effects on the local redox system, and interactions
with sensory nerve endings within the respiratory epithelium. While these events are interrelated,
they could be caused by formaldehyde independently and simultaneously. Although some studies
have reported changes in these initial mechanistic events at formaldehyde concentrations as low as
0.035 mg/m3 following acute or short-term exposure, notable uncertainties remain. For example,
tissue alterations that might increase vulnerability to these changes with continued exposure is
expected (e.g., decreases in mucociliary clearance). Conversely, gradual tissue changes following
exposure might also lead to resilience (e.g., increases in epithelial cell barrier function). More
detailed mechanistic studies characterizing the initial molecular interactions of formaldehyde in the
URT following long-term exposure would help to clarify potential progressive changes in the ability
of formaldehyde inhalation to elicit these intial changes.
Effects on the mucociliary system are likely secondary to the production of reactive
byproducts or covalent modification to mucosal structural components following physical
interactions of formaldehyde with proteins in the mucus. The effects of formaldehyde on mucus
flow patterns appear to include both a concentration and exposure-duration dependency (as well
as variability due to humidity), although a mechanism reliant on direct modification of
macromolecules alone would be expected to be driven largely by concentration. The impact of this
is difficult to define and integrate into the overall mechanistic picture. Persistent changes to the
normally protective mucociliary apparatus or tissue redox capacity are likely to eventually lead to
epithelial damage (which has been shown to correlate with inhibited mucociliary function following
formaldehyde exposure). To repopulate damaged tissue and cells, and to protect against further
insult, damage often leads to cell proliferation or hyperplasia (i.e., an increase in the amount of
tissue due to proliferation of normal cells), and/or the damage can eventually lead to epithelial
lesions such as squamous metaplasia, where cells transition to a different phenotype. This
proliferation, hyperplasia, and/or metaplasia can be adaptive (e.g., response to tissue stress) or
maladaptive, and could lead to subsequent effects on pulmonary function through thickening or
keratinization of the respiratory epithelium, or thickening of mucus, all of which can restrict
This document is a draft for review purposes only and does not constitute Agency policy.
A-502 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
airflow. Formaldehyde exposure-induced damage to the URT epithelial cells could also result in an
altered release of cytokines or other soluble mediators, which, were they to reach the LRT, could
contribute to decreased pulmonary function through airway hyperreactivity and/or
hypersensitivity to challenges such as allergen exposure (Hulsmann and De Jongste, 1996). In
general, the plausible initial mechanistic events and changes in mucus flow patterns observed after
formaldehyde exposure occur at lower formaldehyde levels than those eliciting URT epithelial
lesions (i.e., at <0.3 mg/m3in exposedhumans and >0.6mg/m3 in animals).
Inhaled formaldehyde also appears to directly stimulate trigeminal nerve endings in the
nasal mucosa. Activation of these chemosensory afferents, likely C fibers, is known to initiate
afferent signals that result in the burning sensation characteristic of sensory irritation. This
chemosensory activation is enhanced in the anterior third of the nasal cavity and is typically less
sensitive than olfaction [Hummel and Livermore, 2002], These characteristics are consistent with
the known distribution of inhaled formaldehyde (see Appendix A.2) and with observations that
formaldehyde exposure typically causes chemosensory-related irritation at higher concentrations
than those necessary for olfactory detection in naive individuals (e.g.. as demonstrated by 2012).
The rapid detection of these sensations in exposed individuals suggests a receptor-mediated event
that is dependent on formaldehyde penetration to the nerve endings, which may not have an
exposure duration threshold. Based on mechanistic studies in vitro and ex vivo, activation of the
trigeminal nerve by formaldehyde is likely mediated, at least in large part, through Transient
Receptor Potential A1 (TRPA1) cation channels. To a lesser extent, this activation may also involve
TRPV1 channels, which can be coexpressed and coactivated alongside TRPA1 in certain situations
(Salas et al., 2009). Overall, very little is known about changes in chemosensitivity to inhaled
formaldehyde with repeated exposure over time, as mechanistic studies of long-term exposure
were not identified. With acute, controlled exposure in human volunteers, the initial irritation
response to formaldehyde, which is highly variable across individuals, has been shown to plateau
(e.g., (Green etal.. 1987)) or even decline somewhat (e.g., Bender et al., 1983) when exposure is
continued for several minutes to hours; however, this pattern may depend upon concentration
(Anderson and Molhave, 1983), and changes to this response pattern in humans over time,
particularly with exposure longer than 1 day, remain unclear. Studies of reflex bradypnea in
rodents (see Appendix A.3), which is dependent on the activation of the trigeminal nerve, show that
repeated exposure for up to a month elicits a similar level of activation of this pathway. However,
uncertainties with these data include a nonconstant exposure (i.e., short-term rodent studies
employed work hour-like exposure periodicity) and testing only at reflex bradypnea-inducing
levels (e.g., >1 mg/m3). It is unclear how this informs long-term responses to constant oronasal
exposure in humans (who do not exhibit this reflex) at lower formaldehyde levels. Enhanced
irritation with prolonged exposure could occur directly as a result of sensitization of the receptors
(e.g., TRPA1) to formaldehyde or indirectly by increased access of formaldehyde to trigeminal
nerve endings following damage to juxtaposed epithelial cells. Electrophilic oxidative stress
This document is a draft for review purposes only and does not constitute Agency policy.
A-503 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
products such as hydrogen peroxide and 4-hydroxynonenal are also known to be capable of
stimulating sensory nerve receptors such as TRPA1 (Andersson et al., 2008; Taylor-Clark etal.,
2008), and moderate evidence exists to support the presence of oxidative stress in both the upper
and lower airways. In addition, airway inflammation has been shown to reduce the threshold for
activation of afferent fibers, through an unknown mechanism [Carr and Undem, 2001], Conversely,
however, as this action is mediated predominantly by access of formaldehyde to chemoreceptors,
changes such as the conversion of normal epithelium to squamous epithelium or damage and
destruction of nerve afferents would be expected to reduce or desensitize subsequent irritant
responses. Taken together, this suggests a complex sequence of interactions that might impact
trigeminal nerve chemosensation over time.
Together with the centrally mediated physiological response, stimulation of airway sensory
nerves, including the trigeminal nerve, can also cause a more immediate localized release of
neuropeptides like substance P and calcitonin gene-related protein (CGRP). These released
neuropeptides, particularly substance P, can affect local immune responses by increasing vascular
permeability and leukocyte recruitment, among other things (Sarin et al., 2006, J allergy clinical
immunology), as has been demonstrated with substance P-dependent eosinophil accumulation in
the human nasal mucosa after allergen exposure (Fajac et al., 1995, Allergy 50:970). Observations
of neuropeptide changes, including increased substance P, have been reported at slightly higher
formaldehyde levels than those shown to activate the trigeminal nerve, generally >1 mg/m3. While
URT neuropeptide levels have not been examined in great detail following formaldehyde exposure,
given that the URT represents the primary region of formaldehyde flux, formaldehyde exposure-
induced increases in neuropeptides in model systems and related tissue regions, including the LRT,
are inferred to provide support for the few URT-specific studies that observed elevated
neuropeptide levels. The formaldehyde-specific data further indicate that the neuropeptides are
released from neuronal rather than nonneuronal sources, at least following short-term exposure,
and this release appears to be at least partially dependent on TRPA1 activation. The formaldehyde-
specific URT studies have not examined many of the potential consequences of these changes,
particularly after long-term exposure. Elevated URT neuropeptides might result in local
inflammatory changes ranging from increased histamine and mucus secretion to edema and nasal
obstruction during normal or exaggerated attempts to minimize nasal irritation (Barnes et al.,
1991- neuropeptides in the respiratory tract (2 parts)).
The immune response in the URT following formaldehyde exposure has not been
thoroughly studied, particularly in exposed humans; however, the available evidence does provide
moderate support for granulocyte (e.g., eosinophils; neutrophils) involvement The available data
generally indicate that eosinophils are increased in the URT with acute or short-term exposure at
=0.5 mg/m3, although one study suggests the possible increases at much lower levels in exposed
humans with longer exposure Norback etal.. 20001. Although the role for eosinophils in the upper
airways of exposed individuals remains unclear, airway eosinophils are known to be tightly
This document is a draft for review purposes only and does not constitute Agency policy.
A-504 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
regulated and uncommon in normal airways. In addition to their traditional role as immune
"effectors" (i.e., releasing toxic molecules to destroy invading pathogens), activation of eosinophils
can also cause them to release a number of chemical mediators which damage epithelial cells,
stimulate mucus secretion, induce airway hyperresponsiveness, and perpetuate further
recruitment of inflammatory mediators into the airway (Cohn et al., 2004). Eosinophils, which are
relatively rare (=1%) blood leukocytes, are a hallmark of allergic asthma [Howarth etal., 2000];
however, no formaldehyde-specific studies meeting the inclusion criteria evaluated the URT for
changes in other commonly observed inflammatory markers of allergic individuals such as
activated mast cells or histamine. In addition, the data are unable to inform whether this
inflammatory change persists in the human URT with long-term exposure. It should be recognized
that acute inflammation is a protective response of the tissue to stress or damage; inflammation is
more concerning when it becomes dysregulated, recurrent, and/or persistent
At much higher concentrations (>5 mg/m3), neutrophils also appear to increase within the
upper airways, presumably via migration from the blood. Neutrophils, which are the most common
(>50%) blood leukocyte, are also relatively uncommon (<2%) in healthy airways. These phagocytic
cells, along with eosinophils, are one of the first cells recruited to inflamed tissues shortly after
infection. Both eosinophils and neutrophils can release toxic mediators, including lipid-active
factors and reactive oxygen species (ROS), for which moderate evidence exists to support increased
levels in the URT following formaldehyde exposure, and can damage bystander epithelial cells.
However, in contrast to eosinophils, neutrophils are not thought to be associated with allergic
responses or asthma, although they can be increased in individuals with pulmonary disease
(O'Donnell et al., 2006). Changes in other cells in the URT, including basophils, macrophages, and
lymphocytes, were not observed in the few short-term studies examining them.
Exactly how or why eosinophils and neutrophils migrate to the upper airways following
formaldehyde exposure remains unclear. One possibility is that this response is related to the slight
evidence of increased frequency and duration of URT infections in chronically exposed humans.
However, while this effect might be caused by loss of barrier function (e.g., from epithelial cell
damage or inhibited mucociliary function) leading to increased colonization of the epithelium by
bacteria, this is not temporally plausible for the eosinophil increases observed following acute
exposure. Evidence of specific changes in chemoattractants known to stimulate recruitment of
these cells to the URT (e.g., eotaxin; IL-5; or, indirectly, TNFa or IL-ip, which can stimulate eotaxin
in epithelial cells) was not identified, and thus, the biological explanation for the recruitment of
these cells to the upper airways is unknown. Although not examined, it is also possible that
formaldehyde could directly or indirectly (e.g., through tissue damage) interact with and modify
epithelial components, including pattern recognition receptors, that can trigger release of ROS and
lead to immunological responses (Lambrecht and Hammad, 2012; Holtzman etal., 2014). Overall,
although moderate evidence indicates that inflammatory cells including eosinophils and
neutrophils are increased in the URT following formaldehyde exposure, the data are limited in their
This document is a draft for review purposes only and does not constitute Agency policy.
A-505 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
ability to define the concentration and duration requirements for the effects of formaldehyde
exposure on URT immunological processes, which might inform how these changes are initiated.
This document is a draft for review purposes only and does not constitute Agency policy.
A-506 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
Structural Modification of the Upper Airways
Modification of
biological
macromolecules
Human2: None (note: binding of formaldehyde to albumin and other soluble proteins in
human mucus has been demonstrated in vitro; e.g., Bogdanffy, 1987); hemoglobin adducts
at =0.2 mg/m3, Bono, 2012
Consistent with its known chemistry,
formaldehyde can modify cellular
biological macromolecules, including
Robust
[see Appendix
A.2 and A.4 for
additional
detail]
High or Medium
Animal3: Multiple animal studies demonstrate that inhaled formaldehyde can bind and
modify biological macromolecules, which is consistent with the known biological reactivity
of formaldehyde; evidence includes increased DNA-protein crosslinks (DPXs),
hydroxymethyl (hm) DNA adducts, and reactions with glutathione; (e.g., increased DPXs
are observed at >0.37 mg/m3, Casanova et al., 1989; hmDNA adducts and protein adducts
at >0.86 mg/m3, (Lu et al.. 2010), 2011; Edrissi, 2013)
DNA, and interacts with soluble
factors such as albumin and
glutathione, shortly after exposure
to low concentrations (e.g., <0.5
mg/m3) across a wide range of
exposure durations
Human: N/A (see summary)
Sufficient information for 'Robust'
5
o
1
Animal: N/A (see summary)
from high or medium confidence
studies
Impaired
Mucociliary
Function
£
Human2: decreased mucus flow at >0.3 mg/m3 after acute exposure and pathological
changes in mucociliary clearance in workers at mean exposed levels of 0.25-0.26 mg/m3
after chronic exposure (Andersen and Molhave, 1983; (Holmstrom and Wilhelmsson,
1988).
Decreased mucus flow and ciliary
beat, and impaired clearance, in
humans and rats at >0.25 and >2.5
mg/m3, respectively (observed
across exposure durations),
eventually leading to cilia loss
Robust
¦a
(D
&_
o
.n
CuO
±
Animal3: mucociliary function was generally unaffected at 0.57 mg/m3 after short-term
exposure—minor changes were notable at 2.46 mg/m3; robust changes were observed at
the next highest concentrations tested, >7.27 mg/m3; a general lack of recovery with
longer exposure duration
Human: Increases in ciliary activity at 1.23 mg/m3 in dissociated human nasal epithelial
cells (Wang et al., 2014), with decreased cilia beating frequency in human epithelial cells at
>3.46 mg/m3 (Wang et al., 2014; Schafer et al., 1999): in vitro acute
Suggestive of decreased ciliary beat
and ciliastasis at >5 mg/m3 in
humans and rats with acute
o
1
Animal: Ciliastasis and mucostasis: (Morgan, 1986b) acute 14.76 mg/m3 (not <2.46 mg/m3;
recovery); (Morgan, 1984): acute in vitro (frog palates) >5.36 mg/m3 (authors noted early
exposure, and cilia damage at
>0.5 mg/m3 with short-term
exposure; usually preceded by initial
This document is a draft for review purposes only and does not constitute Agency policy.
A-507 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
activity increase, even at 1.69 mg/m3); structural cilia changes: (Monteiro-Riviere, 1986)
short-term_>0.5 mg/m3, (de Abreu et al., 2016) acute at 0.25, but not 1.2-3.7 mg/m3
effects including slight increases in
activity
This document is a draft for review purposes only and does not constitute Agency policy.
A-508 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure (continued)
Endpoint
Study-Specific Findings from "High or Medium" or "Low" Confidence Experiments
Summary of Evidence (exoosure
duration)
Conclusion
Structural
Change in URT
Mucus
Membrane or
Nasal
Obstruction
High or
Medium
Human: Membrane hypertrophy, atrophy, rhinitis: (Lyapina, 2004) chronic (yrs) 0.87
mg/m3
Mucus membrane damage and
swelling in humans at 0.87 mg/m3
with chronic exposure
Moderate
particularly
in persons
with nasal
damage
Animal: None
5
o
1
Human: Data suggest increased mucosal swelling, nasal obstruction, and/or rhinitis in
workers ((Holmstrom and Wilhelmsson, 1988)) chronic at 0.26 mg/m3 and Norback
et al., 2000): short-term at <0.016 mg/m3, which did not increase in severity with longer
exposure; increase in mucosal swelling in symptomatic nasal distress patients, but not
healthy controls: (Falk, 1994) acute (2 hr) >0.073 mg/m3
Observations at <0.26 mg/m3 in
humans or at >3.5 mg/m3 in rats
support data from the chronic-
duration study and suggest increased
acute vulnerability of people with a
prior nasal condition
Animal: Rhinitis and necrosis in rats after acute or short term (1-3 d) at >3.94 or 4.43
mg/m3
URT Epithelial
Damage or
Dysfunction
[see
Toxicological
Review Section
1.2.4 for
additional data
and discussion]
High or Medium
Human: Indirect data indicating epithelial damage, including loss of ciliated cells, in
occupational studies at 0.1->2 mg/m3 ((Holmstrom and Wilhelmsson, 1988), 1989;
Edling et al, 1987,1988; Ballarin,1992, 3307), with one with more equivocal findings
(Boysen et al., 1990); however, these histopathological symptom scores included
hyperplasia and metaplasia, which complicate interpretation
Duration-dependent epithelial
damage, typically at >2.5 mg/m3 in
subchronic or chronic rat studies,
and with supportive indirect findings
from human studies at
0.1-0.2 mg/m3, generally correlates
with inhibited mucociliary activity
Robust
Animal: Increased epithelial damage and related nasal lesions: duration-dependent,
typically >2.46 mg/m3 in subchronic and chronic studies (e.g., Andersen, 2010; lower in
some longer-term studies) and generally correlating with inhibited mucociliary activity;
goblet cell loss in monkeys (Monticello et al., 1989) short term (1 wk) at 7.38 mg/m3
5
o
1
Human: None
Studies suggest that nasal epithelial
damage is increased, even in
short-term studies, at >2.5 mg/m3
Animal: Goblet cell damage and decreased junctional proteins between epithelial cells in
rats (Arican, 2009): subchronic (12 weeks) at 18.5 mg/m3; mRNA and/or miRNA changes
associated with apoptosis (Rager, 2014): short term (2 d in macques or 28 d in rats) or DNA
repair Andersen et al. (2010): short term (1 wk, but not at 4-13 week durations) at
>2.46 mg/m3; Rhinitis and necrosis in rats after acute or short term (1-3 d) at >3.94 or
4.43 mg/m3
This document is a draft for review purposes only and does not constitute Agency policy.
A-509 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure (continued)
Endpoint
Study-Specific Findings from "High or Medium" or "Low" Confidence Experiments
Summary of Evidence (exoosure
duration)
Conclusion
URT Cellular
(Epithelial)
Proliferation
[see
Toxicological
Review Section
1.2.4 for
additional data
and discussion]
High or Medium
Human: None: indirect data from humans indicating an increase in histopathological
scores that sometimes included hyperplasia were not specific enough to independently
evaluate proliferation
Increased cell proliferation in rats at
all tested durations. Proliferation
increases were typically observed in
the anterior nasal cavity at tested
levels >=3.5-4 mg/m3, and were
generally not observed at <1.23
mg/m3. Sites of proliferation
correlated with the development of
hyperplasia and metaplasia,
although the temporal and exposure
levels specifics of this association are
unclear. Indirect data from
observations of hyperplasia in
exposed animals and humans are
consistent with these data.
Robust -t
Animal: Acute dose-dependent increases in cell proliferation in rats, measured primarily by
DNA labeling during the final days of exposure, were consistently observed following
acute, short-term, and subchronic exposure, and generally with a similar magnitude of
responses across durations. Proliferation was typically highest in anterior regions (e.g.,
"level 2"), with little evidence of proliferation at <1.23 mg/m3, mixed findings between
1.24 and 3.5 mg/m3, and studies generally reporting increases with exposure at higher
levels, particularly with longer exposure duration. These data are supported by consistent
observations of formaldehyde exposure-induced increases in hyperplasia in pathology
studies, some of which provided information showing a correlation between acute
proliferation and hyperplasia and metaplasia. The only rat study that measured exposure
longer than 13 weeks suggests that increases in acute proliferation may begin to decrease
in magnitude with chronic exposure at >6 mg/m3 (Monticello et al., 1996). A few
studies suggest that mice may exhibit less robust responses than rats, while monkeys may
exhibit proliferation in more posterior nasal regaions at >7 mg/m3.
5
o
1
Human: N/A (see summary)
Sufficient information for 'Robust'
from high or medium confidence
studies
Animal: N/A (see summary)
Sensory Nerve-Related Changes
Trigeminal
Nerve
Stimulation
High or
Medium
Human: None
Increased activity of trigeminal nerve
afferents at <0.5 mg/m3 following
acute exposure in animals
Robust -t
Animal: Increased afferent nerve activity: (Tsubone, 1991) acute =20% at 0.62 mg/m3 and
=50% at 2.21 mg/m3; (Kulle, 1975) acute (threshold detection at 25 seconds) at 0.31
mg/m3
5 £
Human: None
This document is a draft for review purposes only and does not constitute Agency policy.
A-510 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure (continued)
Endpoint
Study-Specific Findings from "High or Medium" or "Low" Confidence Experiments
Summary of Evidence (exoosure
duration)
Conclusion
Animal: Indirect evidence: with acute exposure, dose-dependent increase in nerve
currents and CI—release in intact rat trachea (Luo et al. 2013), and stimulation using in
vitro neuronal preparations (McNamara et al., 2007; kunkler et al., 2011)
Supportive indirect evidence from ex
vivo and in vitro experiments
TRPAl and/or
TRPV1
Stimulation
High or Medium
Human: None
Indirect data identify TRPAl as a
molecular target for formaldehyde
exposure-induced sensory effects
Moderate
(TRPAl);
Minimal
(TRPV1: not
shown in
figures)
Animal: Formaldehyde and related chemicals such as acrolein activate the trigeminal
system in wild-type mice, but not TRPAl knockout mice following acute exposure, at least
at high exposure levels (Yonemitsu et al., 2013); taken together with the established role
for TRPAl in acrolein-induced sensory effects (e.g., Bautista et al., 2006); these data
indirectly support a role for TRPAl in sensory nerve-related changes following
formaldehyde exposure
5
o
1
Human: None
Indirect data identify TRPAl and/or
TRPV1, as molecular target(s) of
formaldehyde exposure with acute
or short-term exposure; inhibitor
studies demonstrate that
downstream effects of sensory nerve
stimulation depend on TRPAl or
TRPV1 stimulation.
Animal: Formaldehyde activates the transient receptor potential cation channels, TRPAl
and TRPV1, in in vitro and ex vivo models relevant to acute inhalation exposure of the URT
and upper LRT: (McNamara, 2007; Luo, 2013), and in vivo using formalin as a pain stimulus
(not shown); Inhibition of TRPAl and TRPV1 channels localized to sensory nerve endings
reduce FA exposure-induced nerve currents in rat trachea (Luo et al., 2013) and
immune-related responses in mice (Wu, 2013; Lu, 2005): 1 or 3 mg/m3 for 2 or 4 wk
Neuropeptide
Release
High or
Medium
Human: None
Indirect evidence that Substance P
was increased with subchronic
exposure in a single mouse study at
2.46 mg/m3
Moderate T*
(relevant to
both URT
and LRT;
note:
evidence for
NK Receptor
involvement
is Slight)
Animal: in plasma: Increased substance P in mice with subchronic exposure (Fujimaki,
2004): subchronic at 2.46 mg/m3
5
o
1
Human: in URT: Substance P in nasal lavage is increased in human volunteers with ocular
exposure (He, 2005): 4 d (5 min/d) at 3 mg/m3, but not at 1 mg/m3
Data suggest formaldehyde activates
TRP channels on sensory neurons,
leading to release of CGRP and
substance P, with acute or
short-term exposure at >1 mg/m3
Animal: in URT: Formaldehyde stimulates release of calcitonin gene related-protein (CGRP)
in in vitro models relevant to inhalation exposure of the URT (Kunkler, 2011); Experiments
using the related chemical, acrolein, suggest this is TRPAl-mediated (Kunkler, 2011).
in LRT: Inhibition of substance P receptor (NK1) inhibited formaldehyde-induced currents
in isolated rat trachea (Luo et al., 2013); increased substance P and CGRP in mouse BAL,
This document is a draft for review purposes only and does not constitute Agency policy.
A-511 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure (continued)
Endpoint
Study-Specific Findings from "High or Medium" or "Low" Confidence Experiments
Summary of Evidence (exoosure
duration)
Conclusion
both amplified with ovalbumin (OVA) sensitization, and both involved TRP activation (Wu,
2013): short term at 3 mg/m3
Immune and Inflammation-Related Changes
URT Oxidative
Stress
High or Medium
Human: Increased nasal epithelial MldG adducts (marker for oxidative stress and lipid
peroxidation (Bono et al., 2016): unknown duration (but likely years) at >0.066 mg/m3
Direct and indirect evidence of
elevated reactive oxygen species
(ROS), possibly at very low
concentrations (e.g., at
>0.066 mg/m3, with a maximum of
0.444 mg/m3) with prolonged human
exposure
Moderate T*
Animal: mRNA changes indicating increased stress-response proteins: (Andersen, 2008)
short-term >2.46 mg/m3
Low
Human: Increased nasal lavage nitrites (Priha, 2004): acute (8 hr shift) 0.19 mg/m3
Data suggest elevated oxidative
stress at very low formaldehyde
concentrations with acute and
short-term exposure.
Animal: Increased glutathione peroxidase and/or nonprotein sulfhydryl groups (Cassee,
1996) and (Cassee, 1994): short-term (3 d) 3.94 and 4.43 mg/m3, respectively
Nasal Cellular
Inflammatory
Response
High or Medium
Human: None
Cellular infiltration observed by
histology, primarily neutrophils, but
indirectly supporting other immune
cell infiltration, in short-term animal
studies at 7.38 mg/m3. Indirect
evidence of increases in granulocytes
(and possibly lymphocytes) at 2.46
mg/m3 with short term exposure.
Moderate
granulocytes
(neutrophils,
eosinophils);
Note: data
on
lymphocytes
considered
Indetermina
te
Animal: Increased inflammatory response, mostly neutrophils but also mention of
lymphocytes and other inflammatory cells (e.g., assumed monocytes, basophils and
eosinophils): (Monticello, 1989) short-term (1 or 6 wk) 7.38 mg/m3; "inflammatory cell"
infiltration: (Andersen, 2008) acute or short-term (1 d-3 wk) 7.38 mg/m3; mRNA and
miRNA changes associated with inflammation in rats and nonhuman primates: (Rager,
2014; 2013) short-term (1 or 4 wk, with some miRNA changes reversible with 1 week
recovery) at 2.46 mg/m3: 35 formaldehyde-responsive transcripts altered in the nose
known to be related to immune cells indirectly indicated increases in granulocytes (i.e.,
eosinophil and neutrophil markers) and lymphocyte changes, and Andersen et al.
(2010): short-term (1 wk, but not >4 wk) at >12.3 mg/m3
This document is a draft for review purposes only and does not constitute Agency policy.
A-512 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-73. Summary of changes in the upper respiratory tract (URT) resulting from formaldehyde exposure (continued)
Endpoint
Study-Specific Findings from "High or Medium" or "Low" Confidence Experiments
Summary of Evidence (exoosure
duration)
Conclusion
5
o
1
Human: N/C in nasal lavage cell counts, but increased total protein: (Priha, 2004)
occupationally exposed (8-hr shift) 0.19 mg/m3; Allergy-independent increased
eosinophils, permeability (albumin index) and total protein in lavage: (Pazdrak, 1993)
acute (2 hr) 0.5 mg/m3; increased eosinophils, leukocytes, and permeability (albumin
index) in lavage: (Krakowiak, 1998) acute (2 hr) 0.5 mg/m3 (reversible); indirect evidence of
eosinophil infiltration (increased markers: lysozyme and eosinophil cationic protein), but
not neutrophils, at very low levels (Norback, 2000): <0.02 mg/m3; unknown duration (likely
months or more) in schools
Suaaestive of cellular inflammation,
particularly eosinophils, at 0.5
mg/m3 and indirect markers of
eosinophil recruitment at lower
levels in humans, following acute
exposure; neutrophil inflammation
observed at >6 mg/m3 in rats with
short-term exposure
Animal: Neutrophil inflammation: (Monteiro-Riviere, 1986) short-term_>6 mg/m3
Altered URT
Immunity
(inferred from
URT infections)
High or Medium
Human: Increased frequency and duration of URT infections in symptomatic workers;
increased chronic URT inflammation (and decreased function of blood neutrophils, but N/C
in leukocyte counts) in exposed workers (Lyapina, 2004): chronic (yrs) 0.87 mg/m3 [Note:
recent URT infection was often an exclusion criterion in observational studies focusing on
pulmonary function; see Section A.5.3)
Indirect evidence of decreased
immune capacity in a single study of
chronic human exposure at 0.87
mg/m3 (note: while altered immunity
was observed in an mRNA study,
these changes were not necessarily
indicative of decreased immune
response)
Slight
1^URT
infection
Animal: mRNA chanaes Suaaestive of altered immune response (Andersen, 2010): >12.3
mg/m3 short-term (>1 wk)
5
o
1
Human: None
No evidence to evaluate
Animal: None
This document is a draft for review purposes only and does not constitute Agency policy.
A-513 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
Specific Evaluation and Summary of URT mucociliary function and cellular proliferation
Studies examining the potential effects of formaldehyde exposure on mucociliary function
and cell proliferation were considered for use in identifying potential hazards associated with
respiratory tract pathology effects, but were ultimately determined to be most useful as
mechanistic evidence describing the potential progression of effects on structures within the URT
that might lead to more apical effects (e.g., squamous metaplasia). In contrast to the other
mechanistic studies described in this section, these observational human studies and experimental
animal studies were individually evaluated according to the criteria laid out for human and animal
apical endpoint (i.e., hazard) studies described in Appendix A.5.5, noting that the decisions for the
specific endpoints considered in this section can differ when interpretations of the reliability of the
methods differed from those of the more apical endpoints . Thus, studies were judged as high,
medium, or low confidence, or as "not informative" (i.e., not discussed).
Mucociliary function
Mucociliary function studies in animals, which primarily focused on quantifying mucus flow
rate and qualitative descriptions of ciliary beat frequency and viscosity, were limited to a set of
studies from one research group examining dissected nasal passages. Studies of exposed humans
were similarly limited, with relevant endpoints being evaluated in a prevalence study and an acute,
controlled exposure study. Data are sparse, but in general, mucus flow and/or ciliary beat were
inhibited by formaldehyde exposure as a function of concentration and, at least in rats, exposure
duration. Effects were most pronounced in the anterior nasal regions, with effects progressing
towards posterior regions after extended exposure durations in rats (see Tables A-74 to A-75).
These functional observations are consistent with histological changes observed in experimental
animals, including decreased cilia content in rhesus monkeys after 1 or 6 weeks of exposure to 7.38
mg/m3 (Monticello etal.. 1989) and blebbing of ciliary membranes at formaldehyde concentrations
as low as 0.62 mg/m3, with more overt signs of damage at >7.38 mg/m3, in rats exposed for 1 or 4
days (Montieiro-Riviere and Popp, 1986).
In well-conducted experiments in F344 rats, mucociliary function was generally unaffected
after exposure to 0.57 mg/m3 formaldehyde for <1 to 14 days (Morgan et al., 1986 a, c). Although
sporadic, minor changes were notable at 2.46 mg/m3, including slight increases in mucus flow rate,
inhibition of ciliary beat and mucus flow became clearly apparent at the next highest
concentrations tested, >7.27 mg/m3. Initial increases in mucociliary activity at somewhat lower
level formaldehyde concentrations were also apparent immediately after in vitro exposure,
including increases in ciliary activity at 1.49 mg/m3 in ex vivo frog palates and at 1.23 mg/m3 in
dissociated human nasal epithelial cells (Morgan et al., 1984; Wang et al., 2014), with observations
of mucostasis and ciliastasis at >5.36 mg/m3 in frog palates and decreased cilia beating frequency
in human epithelial cells at >3.46 mg/m3 (Morgan et al., 1984; Wang et al., 2014; Schafer et al.,
1999); however, these in vitro studies are interpreted with low confidence. Two studies in humans
reported consistent effects, with decreased mucus flow at >0.3 mg/m3 after exposure for several
This document is a draft for review purposes only and does not constitute Agency policy.
A-514 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Toxicological Review of Formaldehyde—Inhalation
hours, and pathological changes in mucociliary clearance in workers exposed to mean
formaldehyde levels of 0.25-0.26 mg/m3 for several years (Andersen and Molhave, 1983;
(Holmstrom and Wilhelmsson. 1988).
In rats, impaired function was most frequent in the dorsal and medial maxilloturbinate, the
lateral wall, and portions of the nasoturbinate (Morgan et al., 1986a,c). This is consistent with the
locations of epithelial lesions, which correlate with areas of inhibited ciliary function (Morgan et al.,
1986, Toxicol Appl Pharmacol. 82:1). Similarly, mucus flow was inhibited in the anterior nose of
exposed human volunteers (Andersen and Molhave, 1970). However, whereas mucociliary
function was affected with increasing severity with increasing exposure duration over several days
in rats (Morgan et al., 1986, c), effects on mucus flow rate did not vary with exposure durations of
up to several hours in human volunteers (Andersen and Molhave, 1983). Seemingly consistent with
this finding, mucociliary function in rat nasal passages was reported to recover considerably within
1 hour after 90 minutes of exposure to 18.5 mg/m3 (Morgan etal., 1986a); however, less recovery
occurred after exposure for 6 hours (Morgan et al., 1986a), and little or no recovery was observable
18 hours after exposure for multiple days at similar concentrations (Morgan et al., 1986c). These
data suggest that the initial changes observed in response to exposure may vary somewhat from
the functional changes induced by sustained formaldehyde exposure.
Overall, mucociliary function is affected in a concentration-dependent manner shortly after
formaldehyde inhalation, and this impaired function can be persistent, at least when exposure
exceeds several hours, as indicated by studies in F344 rats and exposed workers. In rats, impaired
function worsens with increasing exposure duration, although durations longer than 2 weeks have
not been tested.
Table A-74. Mucociliary function studies in experimental animals
Reference and study design
Results
Rats
High confidence
Morgan et al. (1986a)
Fischer 344 rats; male; 3-8/exposed groups and
9/control group.
Exposure: Rats were exposed to FA in dynamic
head-only chambers for 10, 20, 45, or 90
minutes or 6 hours with or without a 1-hour
recovery period.
Test article: Paraformaldehyde.
Actual concentrations were within 5% of
nominal concentrations of 0, 2.5, or 18.5
mg/m3.1
Mucociliary function (i.e., mucus flow pattern,
mucus flow rate, and ciliary activity) evaluated
by using dissected nasal mucosa that included
the nasal septum and lateral wall.
Changes in mucociliary function
Group
Observations
Controls
Mean mucus flow rates for nasal septum were
slower (0.91-1.2 mm/min) compared to rates on
the lateral wall (3.61-8.15 mm/min); lateral wall
mucus flow by region (slowest to fastest): anterior,
midregions, posterior
18.5 mg/m3
(no recovery
period)
Ciliastasis and mucostasis observed in specific
regions of nose with discernible differences
between recovery and nonrecovery groups;
ciliastasis increased progressively with duration of
exposure and was observed on anterior and ventral
septum, anterio-medial and dorsal
maxilloturbinate, and lateral wall and lateral
nasoturbinate; distribution of mucostasis exhibited
greater variation within exposure groups compared
This document is a draft for review purposes only and does not constitute Agency policy.
A-515 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Figure 2 from Morgan et al. (1986a) depicting
areas of rat nasal passages used to determine
flow rate on nasal septum and lateral wall.
Main limitations: No major limitations
to ciliastasis; mucostasis exhibited similar site
specificity as ciliastasis but with greater coverage
than ciliastasis (<1 to several mm posterior to
regions of ciliastasis); mucus flow observed over
areas of ciliastasis in anterio-medial and anterio-
dorsal maxilloturbinate, anterior lateral wall, and
anterior septum; mean mucus flow rates reduced in
areas of nasal septum and lateral wall with intact
mucociliary function
18.5 mg/m3
(90-min or 6-hr
exposure with
1-hour recovery
period)
90-min group: recovery characterized to be almost
complete, ciliastasis confined to small regions of
anterio-ventral septum, anterio-medial
maxilloturbinate, anterio-lateral nasoturbinate, and
adjacent lateral wall; extent of ciliastasis similar to
18.5 mg/m3, 20-min group
6-hour group: recovery characterized as
considerable but incomplete, especially in posterior
regions of nose; reduced mucus flow rates
compared to equivalent regions in control rats
2.5 mg/m3
No evidence of impaired mucociliary function
Morgan et al. (1986c)
Fischer 344 rats; male; 6 exposed and 12
controls (n=6 morning, n=6 afternoon)/group.
Exposure: Rats were exposed to FA in dynamic
whole-body chambers 6 hours/day, 5
days/week for 1, 2, 4, 9, or 14 days. Exposure
was followed by an 18-hour recovery period
for some groups.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.57 (0.5-0.6;
range), 2.46 (2.4-2.7), 7.27 (7.0-7.5), and 17.7
(15.0-18.5) mg/m3.1
Mucociliary function and mucus flow rate
evaluated by using dissected nasal mucosa
within 20 minutes after death.
Histopathologic evaluation of the respiratory
tract included transverse sections of the nasal
mucosa tissues used in the evaluation of
mucociliary function.
Figure 1 from Morgan et al. (1986c) depicting
rat nasal passages opened near the midline.
Septum was removed to reveal turbinates.
Arrows indicate direction of mucus flow, and
numbers represent areas assessed for mucus
flow rate. Inset represents lateral aspect of
nasoturbinate showing lateral scroll.
Main limitations: No major limitations
Changes in mucociliary function
Group
Observations (truncated from original article)
Controls
Mucociliary apparatus functioned for 20-60 minutes
after death; minimal inter-animal variation in mucus
flow rate
General
observations for
exposed groups
Concentration- and duration-related defects
included cessation or severe slowing of mucus flow
(mucostasis), loss of ciliary function (ciliastasis), or
alterations in mucus flow patterns; minimal inter-
animal variation; mucostasis observed to generally
be more extensive than ciliastasis, mucus was found
flowing over areas of inactivated cilia
17.7 mg/m3
Duration-dependent mucostasis most frequently
observed on dorsal and medial aspects of
maxilloturbinate, lateral aspect of nasoturbinate
(especially lateral scroll), lateral ridge, and lateral
wall; little or no recovery 18 hours after exposure
7.27 mg/m3
Changes were much less extensive as those in 17.7
mg/m3 group
2.46 mg/m3
Changes were characterized as minimal or absent;
localized inhibition of ciliary activity for few animals
was observed on ventral margin of nasoturbinates
with 9 days of exposure
0.57 mg/m3
No inhibition of mucociliary function observed
Changes in mucus flow rate
Group
Observations
Controls
No significant differences observed between
morning and afternoon groups, combined for
statistical analysis with exposed groups
General
observations
Mucus flow rates found to be characteristic of
specific regions of the nose and observed to be:
slowest on anteromedial naso-and maxilloturbinates
and anterior margin of ethmoid turbinate, fastest on
lateral wall, and intermediate on other regions
This document is a draft for review purposes only and does not constitute Agency policy.
A-516 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
17.7 mg/m3
Reduction of mean mucus flow rate without
histologic changes observed on ventromedial
surface of nasoturbinate (area 1) after 1 day of
exposure, with more pronounced and statistically
significant reductions after 9 days of exposure even
with 18 hours of recovery
7.27 mg/m3
No consistent changes in mucus flow rate observed
except in areas with mucostasis
2.46 mg/m3
No reduction in mucus flow rate observed;
nonstatistically significant increases in mean mucus
flow rates observed on posteromedial aspect of
nasoturbinate (area 10)
0.57 mg/m3
No reductions in mucus flow rate observed;
statistically significant increases in mean mucus flow
rate observed in areas 6 and 9 after 4 days of
exposure but not after 9 days of exposure
Frogs
Low confidence
Morgan et al. (1984)
Leopard frogs; male; 6/group.
Exposure: Frog palates were exposed to FA in
an ex vivo chamber for up to 30 minutes after a
5-minute equilibration period.
Test article: Paraformaldehyde.
Actual concentrations were within 20% of
nominal values and are reported for each
endpoint in the Results column.1
Mucociliary function (i.e., mucus flow and
ciliary activity) evaluated by using dissected
frog palates.
Main Limitations: ex vivo, acute exposure;
nonmamalian model
Group (± SE)
Initial response0
to exposureb
Mucus stasisb
(min ± SE)
Ciliastasisb
(min ±SE)
11.8 (±0.37) mg/m3
6/6
6/6 (1.93±0.13)
6/6 (3.47±0.44)
5.36 (±0.36) mg/m3
6/6
4/6 (8.14±3.27)c
4/6 (13.6±5.18)c
1.69 (±0.10) mg/m3
6/6d
0/6
0/6
0.28 (±0.04) mg/m3
0/6
0/6
0/6
a Response was increased ciliary activity in the presence or absence of increased
mucus flow rate.
b Number of cases in which change was observed/number of cases examined.
c Values in parentheses indicate time to induce the effect for the four positive
cases.
d
The response was variable and generally very slight in this group.
Group
mg/m3
(- SE)
Observations for mucociliary function (truncated from original
article)
11.8
(±0.37)
Increased ciliary activity and mucus flow rate; peak mucus flow
rate followed by rapid decline, cessation of flow, beating cilia,
and changes to mucus flow; ciliastasis preceded by reduced beat
frequency and amplitude
5.36
(±0.36)
Considerable inter-animal variation observed
1.69
(±0.10)
Inter-animal variation observed; initial response involved
variable increase in mucus flow rate and increased ciliary activity
or more frequent surges of increased activity
0.28
(±0.04)
No apparent effect after 30-min exposure
0
Very few ciliated cells observed to be actively beating; any ciliary
beating occurred in individual or small groups of cells; basal
mucus flow rate determined to be 0-4 mm/min
As = anterior septum.
1Study authors originally reported FA concentrations in ppm. These values were converted based on 1 ppm = 1.23
mg/m3, assuming 25°C and 760 mm Hg.
This document is a draft for review purposes only and does not constitute Agency policy.
A-517 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-75. Mucociliary function studies in humans
Study and design
Exposure
Results
Medium Confidence
Andersen and Molhave, 1983
Denmark
Controlled Human Exposure Study
Participants: 16 healthy students,
5 females and 11 males. Mean
age: 23 years; range 20-33years.
31% smokers with one heavy
smoker having >20 cigarettes per
day. None had past formaldehyde
exposure and all had healthy upper
airways. All were habitually nasal
breathers with no history of
chronic or recent acute respiratory
disease.
Methods: Three identical sets of
subject measurements taken each
day, first during control period,
second after 2-3 hours of
exposure and third after 4-5 hours
of exposure. Nasal mucociliary
flow measurements in slits 1-2 are
most anterior and slits 5-6 are
most posterior part of the ciliated
nose.
ANOVA significance at 5%.
Main limitations: short exposure
duration; note: internal control
A 5-hour exposure study. Subjects
assigned to four groups, each group
undergoing four different
exposures over four consecutive
days. Levels were 0.3, 0.5,1.0 and
2.0 mg/m3 formaldehyde with order
decided by latin square design.
Each day began with 2 hour control
period using clean air at 23± 0.5° C,
50+/- 5 % humidity, air velocity
10±3cm/s and air supply rate of 500
m3/h. Control air comprised of
outdoor air filtered through
absolute and charcoal filters.
Following control period,
formaldehyde was added to air,
reaching steady state concentration
after one hour. Formaldehyde
generated by passing air through an
80°C oven containing
paraformaldehyde. Variation
monitored, ranging within ±20%
from the target values.
A statistically significant decrease
in mucus flow rate occurred in the
anterior two-thirds portion of the
ciliated nose (slits 1-4). Mucus
flow rate shown to decrease with
increasing formaldehyde
concentrations starting at 0.3
mg/m3 and then leveling off after
0.5 mg/m3. Flow rate decreases
did not fluctuate with time of
exposure.
Low Confidence
(Holmstrom and
Wilhelmsson, 1988J
Sweden
Prevalence Study
Population: Two exposed groups
170 total; 70 formaldehyde
production workers, Mean age
36.9 years, 87% male, mean
duration employment 10.4 yr. 100
workers exposed to wood dust and
formaldehyde at five furniture
factories. Mean age 40.5 years,
93% male, mean duration
employment 16.6 yr. Referent: 36
persons from local government in
the same village as the furniture
workers, with no history of
occupational exposure to
formaldehyde or wood dust.
Personal sampling in breathing zone
for 1-2 hours in 1985. Total dust
and respirable dust also measured.
Previous measurements 1979-1984
in chemical company combined
with 1985 values to estimate
average annual values for each
participant. Only 1985 values
available for wood factories.
Formaldehyde concentration:
Chemical plant: 0.05-0.5 mg/m3,
mean 0.26 [SD 0.17 mg/m3].
Furniture factory: 0.2-0.3 mg/m3,
mean 0.25 [SD 0.05 mg/m3].
Referent mean 0.09 mg/m3 (based
on 4 measurements in 4 seasons).
Mucociliary clearance is defined to
be pathological if transit time is >
20 minutes for one or both spots.
In formaldehyde only group, 20%
of subjects (14/69, p <0.05
compared to referent) had
clearance times > 20 minutes
compared to 15% of the
formaldehyde-dust group (14/95)
and 3% of the referent group
(1/36).
Formaldehyde-only nasal
specimens had higher mean score
of 2.16 (range 0-4) (p <0.05) while
formaldehyde-dust group had
mean score 2.07 (range 0-6) (p
>0.05). Referent group score was
1.56 (range 0-4). Combining
This document is a draft for review purposes only and does not constitute Agency policy.
A-518 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Toxicological Review of Formaldehyde—Inhalation
Study and design
Exposure
Results
Mean age 39.8 years, 56% male,
mean duration employment 11.4
yr.
Methods: Pretesting
questionnaire, Mucociliary activity
tested using green dye spotted on
both inferior turbinates 1 cm
posterior to the anterior border of
the turbinate. Measured transit
time of spot to rhinopharynx.
Chi-square tests or 2-tailed t-test
for group comparisons.
Main limitations: poor matching of
referent group (i.e., different
occupation type; lower proportion
of males); inclusion of only current
workers and long duration of
employment raises possibility of
healthy worker effect due to
irritation effects; crude measure.
formaldehyde-only and
formaldehyde-dust group mean
score 2.11 (p <0.05). No
correlation observed between
smoking habits and biopsy score,
nor was a correlation found
between the duration of exposure
and any histological changes
Cellular proliferation
A number of quantitative cellular proliferation studies have been carried out in
experimental animals, primarily in rats. While these experiments provide more robust
quantification of changes in cell number compared to histological determinations of tissue
hyperplasia, the data provided by these approaches are limited to active proliferation and do not
directly inform cumulative proliferative responses. For example, the most common approaches
involve in vivo administration of either bromodeoxyuridine (BrdU, a thymidine analog) or tritiated
thymidine ([3H]-thymidine), both of which label newly-synthesized DNA in dividing cells. When
either of these are administered during the last 1-3 days of an exposure (nearly all of the studies
followed a similar protocol), these experiments would only be able to measure the proliferation
actively occurring during the 1-3 days at the end of the exposure; they would provide no
information on proliferation induced earlier during the exposure period, or on adaptive changes to
proliferative responses that might have resulted from those initial exposure effects. Despite this
limitation, these studies still provide useful information on the magnitude of acute proliferation
induced at different concentrations and following different durations of formaldehyde exposure. In
addition, in some studies, histopathology was assessed along with cell proliferation, which may
inform potential correlations between cellular proliferation and apical tissue pathology endpoints.
The studies generally assessed cell proliferation in the anterior part of the nasal cavity, focusing on
discrete regions (i.e., cross section levels) of the epithelium, with a few studies extending their
investigation beyond the nasal cavity to include the trachea, larynx, and carina. There were notable
differences in methodology across studies, including the use of different DNA synthesis-labeling
This document is a draft for review purposes only and does not constitute Agency policy.
A-519 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
agents (i.e., BrdU, [3H] thymidine, 14C), different durations of labeling (i.e., 2 h to 3 d), and different
measures of proliferation (i.e., cell turnover; 14C incorporation; labeling index [LI]: the ratio of
labeled cells to total counted cells; unit length labeling index [ULLI]: the ratio of labeled cells per
mm of basement membrane). While these methodological differences complicate direct
comparisons across studies, increases in cell proliferation were in general consistently observed
across several rat strains, with supportive findings in smaller databases of mice and monkey
studies. Proliferation responses, at least in the anterior nasal cavity of exposed rats, were
concentration-dependent, while in most studies the response magnitude remained relatively
constant across exposure duration (i.e., acute proliferation responses were not notably larger after
longer exposure at similar concentrations; see Figure A-35); the only study to test proliferation
beyond 13 weeks of exposure suggested that response magnitude may actually begin to decrease in
most nasal regions after chronic exposure fMonticello etal.. 19961.
As illustrated in Figure A-35, after <1 week, 1-6 weeks, or >12 weeks of exposure,
proliferation in the nasal epithelium was increased in a concentration-dependent manner in F344
rats, and from a more limited set of studies, in Wistar rats. Proliferation was also shown to increase
in single studies of rhesus monkeys (after exposure for either 1 or 6 weeks to 7.38 mg/m3
formaldehyde; fMonticello etal.. 19891 and B6C3F1 mice (after exposure for 1 to 5 days at
approximately 18.45 mg/m3 formaldehyde; f Chang etal.. 19831: Swenberg et al., 1983).
Interestingly, as with other respiratory tract effects, mice might be less sensitive to changes in
cellular proliferation, although the data relevant to this interpretation are sparse. Specifically,
proliferation in the epithelium lining nasal associated lymphoid tissue (NALT) was observed in
F344 rats, but not in B6C3F1 mice, even at concentrations as high as 18.4 mg/m3 (Kuper etal..
20111. This potential difference could reflect the differential sensitivity to reflex bradypnea across
species (see Section A.3). In rats, although the data were variable across studies, particularly in
Wistar rats exposed for < 1 week {Cassee etal., 1996; Cassee and Feron 1994; fReuzel et al.. 19901:
Wilmer et al., 1989, 3576; Zwart et al., 1988; Woutersen et al., 1987}, the levels of cell proliferation
in regions such as the anterior lateral meatus were typically 1.5- to 25-fold greater than control
levels after exposure to > =12 mg/m3 formaldehyde, regardless of exposure duration. While levels
were similarly increased at =6-7.5 mg/m3 after exposure durations < 13 weeks, the only study to
evaluate longer exposures observed less robust increases in proliferation after chronic exposure, as
compared to proliferation levels after 3 months of exposure (Monticello et al., 1996). The results
across studies were less consistent at formaldehyde concentrations below 4 mg/m3, with several
studies at 2.5-3.67 mg/m3 indicating that proliferation tended to increase in some nasal regions
after >12 weeks (Zwart et al., 1988; Andersen et al., 2010; Meng et al., 2010) and others suggesting
elevations in proliferation at concentrations ranging from 1.24-3.69 mg/m3 with exposure < 1
week (Zwart et al., 1988; (Reuzel et al.. 19901: Roemer, 1993, 7807}, although not all comparisons
in all regions evaluated were statistically significant. Changes at these concentrations were not
observed in several other studies of similar exposure duration, or in any studies examining 1-6
This document is a draft for review purposes only and does not constitute Agency policy.
A-520 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
weeks of exposure. Increases in proliferation were typically not observed at formaldehyde
concentrations below 1.23 mg/m3, although some weak induction was noted in a few studies.
Proliferation generally exhibited a decreasing anterior to posterior gradient and correlated
with sites of respiratory tract pathology. For example, after adjusting for the number of animals
with accurate tumor localization and including target cell population size in the comparison,
increased cell proliferation was correlated (R2 = 0.88) with the incidence of squamous cell
carcinoma; however cell proliferation alone (i.e., without considering target cell population size)
was not as well correlated (R2=0.46; (Monticello etal.. 1996). suggesting that some minimal cell
population size may be important for tumor formation. Cell proliferation has also been shown to be
correlated with hyperplasia and squamous metaplasia; nasal lesions indicative of cytotoxicity such
as cell degeneration, necrosis, or erosion and/or inflammation (Speit et al., 2011; Andersen et al.,
2010; 2008; Monticello etal., 1991). Although most studies demonstrated proliferation in anterior
regions of the nasal cavity, primarily examining sections at cross level 2 (variably including anterior
and/or medial portions of structures such as the lateral meatus, maxilloturbinate, and
nasoturbinate), some studies demonstrated formaldehyde-induced changes in more posterior
regions, including regions outside of the URT. For example, exposure of groups (n=3) of rhesus
monkeys to 7.36 mg/m3 for 1 or 6 weeks resulted in increased proliferation along with slight
histological changes (e.g., inflammation, hyperplasia, and metaplasia) in both the nasal cavity and
extranasal regions including the larynx, trachea, and carina, but not the bronchioles (Monticello et
al.. 1989). In F344 rats, increased proliferation was observed in the nasopharynx at >12.3 mg/m3
(with slight increases at 2.48 mg/m3) after 4 weeks of exposure (Speit et al., 2011). Increased
proliferation in the trachea and lung was observed in SD rats following 1 or 3 days of exposure to
24.6 mg/m3, with mixed findings at lower concentrations, including increased proliferation in the
trachea at 2.5mg/m3 after 1 day of exposure, but decreased proliferation in the trachea with 3 days
of exposure at 2.5-7.4 mg/m3 fRoemer et al.. 19931.
These latter data highlight the complicated nature of the association between formaldehyde
exposure duration and cellular proliferation. While, generally, proliferation appears to be sustained
at similar levels across exposure durations ranging from 1 day to 13 weeks (see Figure A-35), some
studies reported differences in the magnitude of effects in specific regions of the respiratory tract
tissue after different exposure durations. In studies of F344 and Wistar rats exposed to a wide
range of formaldehyde concentrations (0.37-18.5 mg/m3), proliferation induced by formaldehyde
exposure was typically not increased with longer exposure duration (in some instances, it was
slightly decreased, but statistical comparisons were not performed) in various anterior nasal
sections (approximately levels I-III), including comparisons of 3 days to 10 days (Chang etal.. 1983:
Swenberg et al., 1983), 5 days to 15 days (Andersen et al., 2008), and 4 days to 6 weeks (Monticello
et al., 1991) in F344 rats (note: response magnitude increased from 1 to 4 days in the latter study)
and comparisons of 3 days to 4 weeks fWilmer et al.. 19871 and 3 days to 13 weeks in Wistar rats
(Zwart et al., 1988). In several of these studies, the data suggest that formaldehyde concentration
This document is a draft for review purposes only and does not constitute Agency policy.
A-521 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
had a much greater impact on proliferation than exposure duration, although the relative
contributions of concentration versus duration could not be accurately defined (Wilmer et al., 1989;
1987; Chang etal.. 1983: Swenberg et al., 1983). Somewhat complicating this, an increasing
magnitude of proliferation at the same formaldehyde concentration was observed in anterior nasal
regions of F344 rats exposed to 7.4-18.5 mg/m3 for 13 weeks, as compared to 1 or 4 weeks
(Andersen et al., 2010), or for 5 days, as compared to 1 day (Chang etal.. 1983). although an
increase was not observed in B6C3F1 mice in the latter study. Similarly, in a study of rhesus
monkeys, there was a noted exposure duration-dependent increase in proliferation in more
posterior regions (approximately nasal section levels III-V as well as regions posterior to the nasal
cavity) at 7.4 mg/m3 from 1 to 6 weeks of exposure fMonticello etal.. 19891. Interestingly, while
duration-dependent increases in proliferation were observed in anterior nasal regions of F344 rats
exposed to 0.86-18.5 mg/m3 for 1-13 weeks, cell proliferation was greatest at 4 weeks, as
compared to 1 or 13 weeks, when examining central and posterior portions (levels 2-3) of the nasal
cavity (Meng et al., 2010). Finally, as previously mentioned and of particular interest, are the
results of Monticello etal. (1996) in F344 rats exposed to 0.85-18.4 mg/m3 formaldehyde. The
authors observed decreases in proliferation when comparing 3 months of exposure with longer
durations up to 18 months within most of the nasal regions examined, including the lateral meatus,
the anterior and posterior mid-septum, and medial maxilloturbinate; however, the opposite finding
(i.e., duration-dependent increases in proliferation) was observed in the anterior dorsal septum
(Monticello etal.. 1996). Unfortunately, this is the only study that examined proliferation after
chronic exposure and the authors did not report variability or statistical comparisons, which limits
the ability to draw reliable conclusions about a possible drop off in proliferation after 13 weeks of
exposure. Overall, the pattern across studies is mixed but suggestive of possible region-specific
differences in the impact of exposure duration on proliferation, and additional studies would be
needed to clarify the discrepancies.
A large number of well-conducted studies have evaluated acute cellular proliferation after
exposure to a wide range of formaldehyde concentrations for durations ranging from 1 day to 18
months. The data were variable across studies. This variability is assumed to result, at least in part,
from methodological factors that include the selection and preparation of tissue for analysis, the
composition and administration protocol of the labeling agent used to indicate proliferation, when
the proliferation counts were made (e.g., age of the animal), and the units used to express
proliferation data (e.g., LI versus ULLI) (Monticello and Morgan 1997; Goldsworthy etal., 1993;
Monticello et al., 1993; Goldsworthy et al., 1991). Despite this methodological variability, cell
proliferation was consistently increased in response to formaldehyde exposure in anterior portions
of the rat, mouse, and monkey nasal cavity, with studies in rats demonstrating a prominent role for
formaldehyde concentration. While some studies in rats and monkeys demonstrated a role for
exposure duration in cell proliferation within specific regions of the respiratory tract, acute
proliferation in most nasal regions generally remained constant regardless of exposure duration.
This document is a draft for review purposes only and does not constitute Agency policy.
A-522 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Study High/Med Strain Exposure Nasal region depicted Labeling Metric
Roemer et a I1993 H SD Id (3d is less) nose I 22h BrdU 1 LI |
0 Andersen eta I .,2010 H F344 lwk ALM (L2) 3d BrdU ULLI
-B-i Andersen eta I., 200S~ I H [ F344 | 5d High flux 3d BrdU 1 U
O Chang eta I.,1983* H F344 5d (Id is less) NT/MT (Level B) 18h thym. LI
O Swenberg etal., 1983* H F344 3d Level B (note: no statistics) 2hthym. LI
-EF Monticelloetal., 1991 H F344 4d (Id is less) ALM/T (L2) 18h thym. ULLI
Wilmeretal., 1989 H Wistar 3d NT/MT 18hthym. LI
® Wilmeretal., 1987 H Wistar 3d NT/MT 18h thym. LI
-©- Zwartetal., 1988 H Wistar 3d NT/MT/ALM/sept. (L2+L3 ave) 18h thym. turnover
-O- Reuzel etal., 1990 H Wistar 3d NT/MT/ALM/sept. (L2 ave) 2hthym. LI
Study High/Med Strain Exposure Nasal region shown Labeling Metric
-0-
Andersen et al., 2010
H
F344
4wk
ALM (L2)
3d BrdU
ULLI
Andersen et al., 2008
H
F34-4
ISd
High flux (LI)
3d BrdU
U
-B-
Monticello etal., 1991
H
F344
9d
ALM/T (L2)
18h thym.
ULLI
-e-
Monticelloetal., 1991
H
F344
6wk
ALM/T (L2)
18h thym.
ULLI
-B-
Speit et al., 2011
M
F344
4wk
LM (LI; NT similar)
3d BrdU
ULLI
Figure A-33. Nasal cell proliferation in rats exposed to formaldehyde. Summary of rat studies of nasal cell
proliferation (as % change relative to controls) following different durations of formaldehyde exposure, specifically < 1
week (left panel), 1-6 weeks (center panel), or > 12 weeks (right panel). The tables below each panel summarize the
studies, study confidence determinations (only high and medium confidence studies are shown), exposure durations, nasal
regions depicted, cell labeling methods used, and the method of data reporting for each corresponding panel. Note: solid
symbols indicate statistical significance, as identified by the study authors. High confidence studies are indicated by
bolder symbols and with solid, rather than dashed, connecting lines. Data at different tirnepoints from the same study are
indicated by use of the same line colors and general symbol shapes. See Tables A-71 and A-72 for additional details.
Study
High/Med Strain
Exposure
Nasal region shown
Labeling
Metric
¦&
Andersen etal., 2010
H
F344
13wk
ALM (L2)
3d BrdU
ULLI
El
Meng et al,, 2010
H
F344
13wk
ALM (note: p<0.01)
3d BrdU
U
Wilmer etal., 1989
H
Wistar
13wk
NT/MT
18 h thym.
u
e
Zwartetal., 1988
H
Wistar
13wk
NT/MT/ALM (12; NC in 13)
18h thym.
turnover
Monticello etal., 1996
M
F344
12wk
ALM (note: no statistics)
18 h thym.
ULLI
jr|
Monticello etal., 1996
M
F344
6mos
ALM (note: no statistics)
18h thym.
ULLI
•8-
Monticello etal., 1996
M
F344
lyr
ALM (note: no statistics)
18 h thym.
ULLI
Monticello etal., 1996
M
F344
18mos
ALM (note: no statistics)
18 h thym.
ULLI
-A-
Casanova etal., 1994
M
F344
12wk
LM (less in M/PM)
3h "C
11C incorp.
This document is a draft for review purposes only and does not constitute Agency policy,
A-523 " DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-76. Subchronic or chronic exposure cell proliferation studies in
experimental animals
Reference and study design
Results
Rats
High confidence
Andersen et al. (2010)
Fisher 344; male; 8/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for 1, 4, or 13
weeks. Rats sacrificed immediately after
last exposure.
Test article: Paraformaldehyde.
Actual concentrations reported in the
Results column. Target concentrations
were 0, 0.8, 2.5, 7.4,12.3, and 18.5
mg/m3.1
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index at levels 1
(highest FA flux near nose tip), II
(anterior lateral meatus, anterior mid-
septum, medial aspect of
maxilloturbinate), and III (posterior
lateral meatus, posterior mid-septum).
Cell proliferation at each site reported as
number of labeled cells per total cells
(i.e., LI) and as the number of labeled
cells per length (i.e., mm) of basement
membrane (i.e., ULLI).
Supplemental 4A from Andersen et al.
(2010) depicting a schematic illustration
of the nasal cavity levels used for cell
proliferation studies.
Nasal Epithelium ULLI
Formaldehyde (mg/m3)
Site
0
0.8
2.5
High-flux region (HFR)
1 week
12.8±3.5a (7)b
15.0±12.5 (8)
13.8±7.0 (8)
4 weeks
20.3±4.1 (7)
17.8±3.8 (8)
18.5±4.6 (5)
13 weeks
21.9±20.3 (3)
12.2±10.3 (3)
29.1±32.7 (6)
Anterior lateral meatus (ALM)
1 week
31.9±26.3 (8)
32.6±30.2 (8)
25.1±26.1 (8)
4 weeks
26.6±17.1 (8)
34.3±21.3 (8)
26.7±7.9 (8)
13 weeks
aMean ULLI
Nasal Epitl
21.7±15.1 (8)
±SD; bNumberc
7elium ULLI (con
29.7±24.6 (8)
)f animals exam
tinued)
56.3±33.3 (8)
ined.
Formaldehyde (mg/m3)
Site
0
7.4
12.3
18.5
High flux region (HFR)
1 week
12.8±3.5a (7)b
25.2±13.3 (8)
36.1±14.3C (8)
25.3±17.5 (7)
4 weeks
20.3±4.1 (7)
40.9±24.9 (5)
69.2±17.7C (6)
63.6±26.1c (8)
13 weeks
21.9±20.3 (3)
17.4(1)
58.3±27.8 (5)
110.2±46.0C
(7)
Anterior lateral meatus (ALM)
1 week
31.9±26.3 (8)
62.9±50.3 (8)
75.7±31.1d (8)
45.1±25.7 (8)
4 weeks
26.6±17.1 (8)
63.1±21.6C (8)
90.7±17.6C (8)
67.0±10.5C (8)
13 weeks
aMean ULLI
21.7±15.1 (8)
±SD; bNumberc
56.4±17.2 (8)
)f animals exam
83.3±33.3C (8)
ined; dp<0.01;e
91.8±33.1c (8)
d<0.05.
Meng et al. (2010)
Fischer 344; males; 8/group.
Exposure: Rats were exposed to FA in
dynamic chambers (not otherwise
specified) 6 hours/day, 5 days/week for
1, 4, or 13 weeks.
Test article: Paraformaldehyde.
Actual concentrations were not
reported. Target concentrations were 0,
0.86, 2.46, 7.38, 12.3, and 18.5 mg/m3.
Dose-dependent increases in cell proliferation of nasal epithelium at 1, 4,
and 13 weeks of exposure.
Cell proliferation had a decreasing anterior to posterior gradient.
Duration-dependent increases in cell proliferation at the anterior portion
of nasal cavity.
Cell proliferation greatest in the central and posterior regions of the nose
following 4 weeks of exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
A-524 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index in the
anterior lateral meatus (lateral wall) for
both sides of the nose. Cell proliferation
data reported as percentage of Brdll-
labeled cells among the total number of
labeled and unlabeled cells.
FA
(mg/m3)
% BrdU-labeled cells after 13 wk
0
18
0.86
22
2.46
35
7.38
38
12.3
51a
18.5
64a
ap <0.01, compared to control group
Wilmer et al. (1989)
Wistar rats; male; 25/group.
Exposure: Rats were exposed to FA in
dynamic horizontally placed glass
cylinders (with sampling ports at the
inlet and outlet) either continuously for 8
hours/day, 5 days/week for 13 weeks or
intermittently 8 hours/day (successive
periods of 0.5 hour of exposure and 0.5
hour of nonexposure), 5 days/week for
13 weeks.
Test article: Paraformaldehyde.
Actual concentrations were not
determined. Target concentrations
were 0,1.2, or 2.5 mg/m3 for continuous
exposures and 0, 2.5, or 4.9 mg/m3 for
intermittent exposures.1
Cell proliferation studies carried out
after 3 days or 13 weeks of FA exposure
with [3H]thymidine labeling (ip injection
18 hours postexposure) and scoring of
the cells lining the nasal (n=1000) and
maxillary (n=1000) turbinates and the
septum (n=3000).
Percentage of [3H]thymidine labeled cells in nasal epithelium
% labeled cells
Exposure
Exposure x time
After 3 days
After 13 wk
0 mg/m3
0 mg/m3h/day
0.60 (0.37)a
1.03 (0.26)
1.2 mg/m3
(continuous)
9.6
mg/m3h/day
0.34 (0.10)
0.81 (0.54)
2.5 mg/m3
(continuous)
20 mg/m3h/day
0.61 (0.28)
0.91 (0.59)
2.5 mg/m3
(intermittent)
10 mg/m3h/day
0.29 (0.20)
1.16 (0.59)
4.9 mg/m3
(intermittent)
19.6
mg/m3h/day
0.58 (0.32)
2.86 (1.80)
!SDs shown in parentheses.
Zwart et al. (1988)
Wistar rats; male and female;
50/group/sex.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for 13 weeks.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.37
(±0.02), 1.24 (±0.10), and 3.67 (±0.27)
mg/m3.1
Cell proliferation studies carried out
after 3 days or 13 weeks of FA exposure
with [3H]thymidine labeling (i.p. injection
18 hours postexposure) and scoring of
Cell proliferation (based on 5 rats/group/sex)
3 days:
Section III - Exposure-related increase in cell turnover for combined
data (males and female, p <0.001), with statistically significant
differences between males and females (p <0.02).
Section II - Cell turnover statistically significant (p <0.001) in 3.67
mg/m3 group, no difference in 0.37 and 1.24 mg/m3 groups compared
to controls.
13 weeks:
Section III - Statistically nonsignificant decrease in mean cell turnover
for all groups.
Section II - Cell turnover statistically significant (p <0.001) in 3.67
mg/m3 group, no difference in 0.37 and 1.24 mg/m3 groups compared
to controls.
This document is a draft for review purposes only and does not constitute Agency policy.
A-525 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
the cells lining the nasal and maxillary
turbinates (n=1500), septum (n=2000),
and lateral wall (n=1500) at Section III.
Only cells lining the nasal septum were
scored at Section II.
Compared to Section II, cell turnover roughly 10 times greater at Section
Data extracted using Grablt software (mean+SEM converted from log
scale):
mg/m3
Level III (3 d)
Level III (13
wk)
Level II (3 d)
Level II (13 wk)
0
0.517
(0.043)
0.165 (0.029)
0.022
(0.005)
0.041 (0.014)
0.37
0.541
(0.045)
0.133 (0.021)
0.040
(0.008)
0.038 (0.010)
1.24
0.872
(0.104)*
0.141(0.027)
0.034
(0.009)
0.038 (0.005)
3.67
3.71 (0.442)*
0.101(0.027)
0.435
(0.147)*
0.214 (0.050)*
Medium confidence
Casanova et al. (1994)
Fischer 344; male; 8/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for 11 weeks
plus 4 days. On day 5 of week 12, rats
were exposed to labeled FA (i.e.,
H14CHO) in nose-only chambers for 3
hours.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.86
(±0.02), 2.52 (±0.05), 7.23 (±0.16), 12.35
(±0.23), 17.86 (±0.37) mg/m3 for whole
body exposures and 0, 0.86 (±0.02), 2.53
(±0.04), 7.39 (±0.15), and 19.4 (±0.4)
mg/m3 for nose-only exposures.1
Cell proliferation studies carried out by
determining H14CHO incorporation into
DNA (i.e., de novo DNA synthesis) via
liquid scintillation counting.
Cell proliferation lateral meatus (LM) versus medial and posterior
meatuses (M:PM)a
FA (mg/m3)b
0
0.86
2.53
7.39
19.4
Observation
NA
No difference between LM and M:PM
No difference between LM and M:PM
Preexposed (PE) rats: significantly greater (p<0.02)
proliferation in LM than M:PM
Naive (N) rats: greater proliferation in M:PM than
LM
PE rats: significantly greater (p<0.02) proliferation in
LM than M:PM
N rats: greater proliferation in M:PM than LM
aFor whole body exposures to unlabeled FA, rats exposed to 0 mg/m3 were
considered N, whereas rats in the other exposure groups were considered
PE; Concentrations represent those used for nose-only exposures with
H14CHO.
Cell proliferation preexposed versus naive ratsa
FA (mg/m3)b
0
0.86
2.53
7.39
Observation
NA
No difference between PE and N
No difference between PE and N
PE rats: greater (p <0.01) proliferation in LM than in N rats
19.4 PE rats: greater (p <0.01) proliferation in LM and M:PM
than N rats
aFor whole body exposures to unlabeled FA, rats exposed to 0 mg/m3 were
considered N, whereas rats in the other exposure groups were considered
PE; Concentrations represent those used for nose-only exposures with
H14CHO.
This document is a draft for review purposes only and does not constitute Agency policy.
A-526 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Lateral meatus = L; medial and posterior meatuses = M:PM.
Data extracted using Grablt software (mean+SEM):
mg/m3
Lateral
Meatus (3h)
Lateral
Meatus (12
wk)
Med/Posterior
Meatus (3d)
Med/Posterior
Meatus (12
wk)
0.861
69.16
(0.0001)
74.93 (5.76)
57.63 (5.76)
63.40 (5.76)
2.46
80.69 (5.76)
92.22 (5.76)
97.98
(0.0001)
109.5 (5.76)
7.38
115.3 (5.76)
749.3
(161.4)*
201.7 (23.05)
276.7 (23.05)
18.45
149.86
(11.53)
1591
(132.5)*
334.3 (23.05)
1002 (103.7)*
*p<0.05 for 12 wk vs 3h exposure
(Monticello et al.. 1996)
F344 rats; male; 6/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers to FA 6
hours/day, 5 days/week for up to 24
months with interim sacrifices at 3, 6,12,
and 18 months.
Test article: Paraformaldehyde.
Actual FA concentrations were 0 (±0.0),
0.85 (±0.06), 2.52 (±0.18), 7.39 (±0.41),
12.2 (±0.54), or 18.4 (±0.98) mg/m3.1
Cell proliferation studies (6 rats/group)
conducted with surgical implantation of
[methyl-3H]thymidine-containing pumps
(5 days prior to interim sacrifice) and
determining labeling index at 7 locations
in the nasal passages: anterior lateral
meatus, posterior lateral meatus,
anterior mid-septum, posterior mid-
septum, anterior dorsal septum, medial
maxilloturbinate, and maxillary sinus
(excluding ostium). Cell proliferation
data reported as the number of labeled
cell profiles per mm of basement
membrane (i.e., ULLI).
mg/m3
Exposure
(months)
Anterior
lateral
meatus
Posterior
lateral
meatus
Anterior
mid-
septum
Posterior
mid-
septum
Anterior
dorsal
septum
0
3
10.11a
7.69
6.58a
11.94
2.14
6
11.14
11.92
5.73
27.31
3.61
12
8.28
7.67
3.25
31.31
8.63
18
5.74
8.99
4.80
19.86
3.80
0.85
3
10.53
7.82
8.04
13.28
1.08
6
10.09
8.15
3.71
17.04
2.20
12
6.39
5.11
1.72
13.28
1.08
18
6.89
6.40
4.54
18.31
4.95
2.52
3
9.83
11.24b
12.74
13.11b
3.38
6
7.14
9.15
4.78
12.07
2.06
12
6.35
6.19
2.14
10.35
0.92
18
3.66
5.24
3.02
7.20
1.93
7.39
3
15.78
9.65
4.15
10.52
3.55
6
7.98
6.74
3.52
7.76
1.52
12
6.24
5.42
3.06
8.76
2.01
18
3.51
6.47
3.96
12.30
1.96
12.2
3
76.79
15.29
39.01
21.43
5.28
6
53.57
17.97
28.22
15.81
2.64
12
32.42
5.60
10.29
6.79
2.20
18
36.28
19.45
11.92
24.44
3.22
18.4
3
93.22
59.52
75.71
51.79
5.96
6
65.89
44.63
75.32
61.52
26.18
12
74.99
44.73
51.62
60.56
37.52
18
34.62
22.34
30.29
37.06
52.98
an=5 or 6; bn=4
This document is a draft for review purposes only and does not constitute Agency policy.
A-527 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Exposure
(months)
mg/m3
medial
maxilla
turbinate
maxillary
sinus
mg/m3
medial
maxilla
turbinate
maxillary
sinus
3
0
7.84a
8.10
7.39
9.23
ND
6
17.95
ND
10.18
ND
12
7.85
6.31
6.22
12.04
18
5.58
5.95
5.03
9.51
3
0.85
10.33
ND
12.2
89.20
ND
6
9.34
ND
57.83
ND
12
6.79
7.80
43.27
9.15
18
5.08
6.99
42.74
12.12
3
2.52
10.84
3.12
18.4
115.19
10.77b
6
10.41
ND
101.97
13.13
12
5.98
7.73
66.64
17.06
18
3.42
8.52
63.11
13.16
an=5 or 6; bn=3
Table A-77. Short-term exposure cell proliferation studies in experimental
animals
Reference and study design
Results
Rats
High Confidence
Andersen et al. (2008)
Fischer 344 rats; male; 8/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for up to
3weeks. Rats sacrificed at end of single
6-hour exposure (Day 1), 18 hours after
single 6-hour exposure (Day 1 recovery),
at end of 5 days of exposure (Day 5), at
end of 6 days of exposure (Day 6), 18
hours after 6 days of exposure (Day 6
recovery), and at end of 15 days of
exposure (day 15).
Test article: Paraformaldehyde.
Actual concentrations were determined
on a daily basis and reported in the
Results column. Target concentrations
were 0, 0.9, 2.5, 7.4, and 18.5 mg/m3.1
This study also evaluated the effects of a
single FA instillation (40 nL, 400 mM per
nostril). Data presented here in the
Target concentration
(mg/m3)
Actual FA Concentrationsa
Day 1
(mg/m3)
Day 5
(mg/m3)
Day 6
(mg/m3)
Day 15
(mg/m3)
0
0±0
0±0
0±0
0±0
0.9
0.74±0.23
0.79±0.15
0.75±0.16
0.7±0.11
2.5
2.08±0.46
2.14±0.43
2.26±0.49
2.2±0.31
7.4
5.83±1.73
6.43±0.76
6.00±1.25
6.14±0.97
18.5
17.7±5.7
NA
NA
NA
aDaily means ± SD.
Cell proliferation in nasal epitheliuma
Formaldehyde (mg/m3)
Day
Level
Site
Control
0.9
2.5
7.4
1
NA
38.6±8.5b
(13.2±4.6)
36.8±14.7
(10.2±2.8)
65.0±39.8
(16.6±6.0)
155.0±88.9C
(35.5±14.8)c
5
Aim
6.0±2.5
7.5±1.1
7.3±1.7
29.0±21.9C
II
As
5.6±3.0
6.Oil.6
6.6±3.5
14.2±10.3C
Mam
6.5±2.1
6.8±3.1
9.7±3.8
35.1±22.0C
III
Plm
6.4±3.0
8.1±2.4
10.0±4.0
16.1±6.4C
This document is a draft for review purposes only and does not constitute Agency policy.
A-528 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Results column are for inhalation
exposures.
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index at levels I
(front of nose), II (anterior lateral
meatus, anterior septum, medial aspect
maxilloturbinate), and III (posterior
lateral meatus, posterior septum). Cell
proliferation determined only for days 5
and 15 and reported as the number of
labeled cell profiles per mm of basement
membrane (i.e., ULLI).
Ps
8.9±3.0
7.5±3.5
8.0±5.2
15.0±11.9C
1
NA
78.9±54.7
(22.6±17.2)
55.8±37.3
(15.6±10.5)
50.8±44.2
(15.6±13.1)
119.1±38.0
(40.6±ll)c
Aim
12.4±12.4
18.2±11.4
12.1±7.0
19.1±8.7
15
II
As
12.0±9.7
17.6±11.0
10.0±4.6
14.1±8.7
Mam
22.7±23.0
27.2±18.6
20.9±20.6
21.9±16.8
Plm
11.8±10.0
12.6±6.3
11.7±7.6
13.6±7.2
III
Ps
15.9±15.2
13.0±5.9
12.5±6.3
18.3±12.1
aReported as mean±SD; bData represent ULLI. Data in parenthesis
represent LI: (labeled cells/total cells) x 100; cp<0.05.
(Cassee et al.. 1996b)
Wistar rats; male; 5 to 6/group.
Exposure: Rats were exposed to FA in
dynamic nose-only chambers 6
hours/day for 1 or 3 days. Rats
sacrificed immediately after last
exposure.
Test article: Paraformaldehyde.
Actual concentrations were 0,1.2, 3.9,
and 7.9 mg/m3.1
Cell proliferation studies carried out
using deparaffinized standard cross
sections of the nose and semi-
quantitative proliferating cell nuclear
antigen (PCNA) immunostaining. Cell
proliferation studies were also
conducted with surgical implantation of
Brdll-containing pumps (20 hours prior
to sacrifice). Labeling index determined
for the entire epithelium of both sides of
anterior nasal cavity lining the
nasoturbinate, maxilloturbinate, lateral
wall, and septum. Cell proliferation at
each site reported as number of
positive-stained cells per length (i.e.,
mm) of basement membrane (i.e., ULLI).
1 day exposure: no treatment-related changes in cell proliferation
FA (mg/m3)
Cell proliferation measured by PCNA after
3 daysa
1.2
Levels II and III: no increases in ULLIs
3.9
Level II: significant increase in ULLIs at
maxilloturbinate (p <0.05) and nasal
turbinate and lateral wall (p <0.01),
compared to controls
Level III: no increases in ULLIs
7.9
NR
aBased on data from 3 to 5 rats per exposure group and 10 to 12 control
rats.
FA (mg/m3)
Cell proliferation measured by BrdU after 3
daysa
1.2
Levels II and III: no increases in ULLIs
3.9
Levels II and III: no increases in ULLIs
7.9
NR
aBased on data from 3 to 5 rats per exposure group and 10 to 12 control
rats.
This study also evaluated the combined effects of FA, acetaldehyde, and
acrolein on nasal epithelium. Data presented here are for formaldehyde-
only exposed rats
This document is a draft for review purposes only and does not constitute Agency policy.
A-529 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Figure 1 from (Cassee et al.. 1996b)
depicting cross levels of the rat nose
evaluated for cell proliferation.
Chang et al.. 1983; [additional data
from related Swenberg et al. (1983)
reportl
Fischer 344 rats; males; 4-5/exposure
group, 9/control group.
Exposure: Rats were exposed to FA in
head-only chambers 6 hours/day for 1,
3, 5, or 10 days.
Test article: Paraformaldehyde.
Actual concentrations were 0 and 18.5
(±0.1) mg/m3.1 Target concentrations
were 0, 0.62, 2.46, 3.69, 7.38, 14.76, or
18.45 mg/m3 in Swenberg et al. (1983)
report.
Cell proliferation studies carried out
after FA exposure with [3H]thymidine
labeling (i.p. injection 2 or 18 hours
postexposure) and scoring of cells
(n=9000) lining the respiratory
epithelium from the nasal and maxillary
turbinates and lateral wall.
Levels A (with minimal mucociliary
clearance) and B (with extensive
mucociliary clearance) reported in
Swenberg et al. (1983)
Group (18.5 mg/m3)
Labeling index (%) in Level B
Control
0.43±0.05 (9)a
1 day
5.51±0.35 (4)b
5 days
10.05±0.27 (5)b c
aNumber in parentheses represents number of animals studies;
Significantly different from control, p<0.05; Significantly different from 1-
day exposed rats, p<0.05.
% labeled respiratory epithelial cells in Level B (thymidine at 2 h
postexposure)
Formaldehyde Concentration (mg/m3)
Duration
0
0.62
2.46
7.38
18.45
3 days
0.22
(0.03)
0.38 (0.05)
0.33 (0.06)
5.4 (0.82)
2.83 (0.81)
% labeled respiratory epithelial cells (thymidine at 18 h postexposure)
Control
3.69 mg/m3 x 12 h/ day
7.38 mg/m3 x 6 h/ day
14.76 mg/m3 x 3 h/ day
3 days (Level
B)
0.54 (0.03)
1.73 (0.63)
3.07 (1.09)
10 days (Level
B)
0.26 (0.02)
0.49 (0.19)
0.53 (0.2)
1.73 (0.65)
3 days (Level
A)
3.0(1.56)
16.99 (1.5)
15.46 (10.01)
16.49 (2.07)
0, 9.0(0.88) , ,
Mean (SEM); Group sizes and statistical comparisons not reported
Swenberg et al., 1983
in
Note: Pulse labeling with thymidine 18 hours compared to 2 hours
postexposure resulted in =2-fold and =3-fold increase in labeling in control
rats and at 7.38 mg/m3, respectively (Swenberg et al., 1983).
This document is a draft for review purposes only and does not constitute Agency policy.
A-530 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
(Kuper et al.. 2011)
Fischer 344 rats; male; 8/group.
Exposure: Mice were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 day/week for 4 weeks.
Test article: Formalin (10.21% FA).
Actual concentrations were 0, 0.63
(±0.06), 1.23 (±0.14), 2.48 (±0.18), 7.53
(±0.42), 12.3 (±0.48), and 18.4 (±0.06)
mg/m3.1
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index of 2 sections
of NALT and 1 section of a upper-
respiratory tract-draining lymph node
(i.e., posterior and superficial cervical
lymph nodes). Cell proliferation data
reported as Brdll-positive cells per
length (i.e., mm) of epithelium.
Lymph nodes: No FA-related effects on the number of Brdll-positive cells
reported in the follicle and paracortex compartments and medulla
BrdU counts in section 1 of NALT
FA (mg/m3)
Interfollicular
area
Interfollicular
epithelium
Follicular
area
Follicular
epithelium
0
61.9±18.8a
6.5±3.2
73.0±39.1
12.6±17.5
0.63
57.3±17.4
4.9±2.2
53.5±19.4
4.9±3.8
1.23
55.7±17.7
5.9±3.4
52.2±27.9
6.4±6.5
2.48
53.5±12.9
4.3±2.7
49.8±22.1
4.7±3.2
7.53
51.1±14.9
3.3±2.4
47.6±13.9
5.8±5.3
12.3
55.5±15.3
5.5±3.5
51.2±16.2
5.7±2.9
18.4
54.4±11.6
28.2±ll.lb
41.4±14.2
23.6±13.6C
aMean number of Brdll-positive cells±SD; bp <0.001; cp <0.05.
BrdU counts in section 2 of NALT
FA (mg/m3)
Interfollicular
area
Interfollicular
epithelium
Follicular
area
Follicular
epithelium
0
48.3±17.7a
6.3±2.2
62.3±24.1
6.8±1.5
0.63
51.0±16.3
4.4±2.7
58.0±30.5
5.8±5.6
1.23
53.9±12.2
4.1±2.9
47.0±15.3
6.9±3.8
2.48
53.4±14.2
5.1±2.4
52.2±15.1
5.6±4.0
7.53
48.2±12.3
3.5±2.3
47.2±15.0
5.9±2.8
12.3
56.0±16.3
6.4±2.3
56.8±17.4
6.2±4.7
18.4
49.9±9.1
24.5±12.6b
40.1±11.8
22.9±10.5b
aMean number of Brdll-positive cells±SD; bp<0.001.
Monticello et al. (1991)
Fischer 344 rats; males; 4-6/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for 1, 4, or 9
days or 6 weeks.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.85
(±0.01), 2.48 (±0.02), 7.63 (±0.12), 12.2
(±0.11), and 18.2 (±0.28) mg/m3.1
Cell proliferation studies carried out
after FA exposure with [3H]thymidine
labeling (ip injection 18 hours
postexposure) and profiling nasal
epithelial cells in serial sections of Levels
II and III of the nose. Level II included
the lateral meatus with the lateral
aspect of the nasoturbinate, lateral wall,
and lateral aspect of maxilloturbinate
(Site 1); midseptum (Site 2); and medial
aspect of maxilloturbinate (Site 3). Level
Mean until length labeling indices0
Exposure time
mg/m3
Level
Site
1 day
4 days
9 days
6 weeks
0
II
1
2.16b
1.46
1.44
0.91
2
1.08
1.03
1.09
0.41
3
2.49
1.36
1.38
1.02
III
1
1.83
1.10
1.36°
0.98
2
3.02
2.81
1.68°
2.18
0.85
II
1
1.31°'e
1.37
1.20
0.88°
2
1.01°
0.97
0.80
0.24°
3
1.75°
1.54
0.80
1.21°
III
1
1.72°
1.27
1.40
0.91°
2
1.74°
3.09
1.06
1.54°
2.48
II
1
2.36°
1.72
1.73
1.36
2
1.69°
0.67
0.97
0.68
3
2.81°
1.09
1.48
1.11
III
1
2.46°
1.09°
1.74
0.86
2
2.39°
1.43°
1.43
2.57
This document is a draft for review purposes only and does not constitute Agency policy.
A-531 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
III included the lateral wall (Site 1) and
midventral septum (Site 2).
I li ill IV V
KEY:
|^|Siic I
ESS5"*2
1
LEVEL II LEVEL III
Figure 1 from Monticello et al. (1991).
(A) Lateral view of the rat nose with
Levels l-V of the nasal passage. (B) Level
II and (C) Level III represent sites for cell
proliferation studies.
7.63
II
1
16.86cfg
30.51fg
23.5 lfg
14.41f'g
2
3.85°
10.00f
10.85f
2.10
3
18.15cf
25.03f
22.54f
16.32f
III
1
7.53f
8.77cf
7.35f
2.08
2
4.20
9.22cf
9.50f
2.58
12.2
II
1
11.17cf
20.9 lf
28.59f
23.87cf
2
17.90cf
26.12fg
19.62f
21.44 c,f,g
3
5.87°
20.26f
20.95f
26.07cf
III
1
14.48f
20.01cf
30.59f
24.21f
2
24.44f
18.70cf
28.60f
13.98f
18.2
II
1
12.68f
25.78f
24.57cf
28.74cf
2
16.72f
29.10f
29.09cf
25.95cf
3
5.31
19.39f
28.71cf
25.10cf
III
1
16.35d,f
30.80cf
40.36f
34.78cf
2
19.26df
34.43 c'f
32.53f
27.47 c'f
aUnit length labeling index defined as the number of labeled cell
profiles/mm basement membrane; bn=6, unless otherwise indicated; cn=5;
dn=4; eUnless noted, not statistically different from control; f p <0.05
compared to control;g p <0.05 compared to level III.
(Reuzel et al.. 1990)
Wistar rats; male; 5/group.
Exposure: Rats were exposed in
dynamic whole-body chambers 22
hours/day for 3 days to FA.
Test article: Paraformaldehyde.
Actual concentrations were 0, 0.37
(±0.01), 1.4 (±0.0), and 3.8 (±0.1) mg/m3
FA.1
Cell proliferation studies carried out
after FA exposure with [3H]thymidine
labeling (ip injection 2 hours
postexposure) and scoring of the cells
lining the nasal (n=1000) and maxillary
(n=1000) turbinates, lateral wall
(n=1000), and the septum (n=2000).
See diagram from (Cassee et al..
1996b)
(above) for cross levels of the rat nose
evaluated for cell proliferation.
Data extracted using Grablt software (mean from level 2, Figure 3, HCHO
only):
mg/m3
Maxilloturb.
Nasal Turb.
Lateral wall
septum
0
0.351855128
0.291340043
1.19765084
0.172349
0.369
0.287744031
0.842204054
1.04583032
0.221581
1.23
0.221580704
0.337503123
0.54215496
0.221581
3.69
4.456151692*
5.273729396*
5.8261316*
4.627466*
Note: data were also presented for Level 3 (same regions). While slight
increases became noticeable at 3.69 mg/m3, none reached statistical
significance.
This study also evaluated the combined effects of FA and ozone mixtures
on nasal epithelium. Ozone co-exposure resulted in an increase in
proliferation compared to formaldehyde exposure alone. Data are only
presented herein for formaldehyde-only exposures.
Roemer et al. (1993)
Sprague Dawley rats; male; 3 or
5/exposure group, 6 or 10/control
group.
Proportion of BrdU-labeled cells (%) after exposure
Cell origin and
exposure
frequency
Formaldehyde (mg/m3)
Number of
rats per
groupa
2.5
7.4
24.6
This document is a draft for review purposes only and does not constitute Agency policy.
A-532 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Exposure: Rats were exposed to FA in
dynamic head-only chambers 6
hours/day for 1 or 3 days.
Test article: Paraformaldehyde.
Actual concentrations were within 10%
of nominal concentrations of 0, 2.5, 7.4,
or 24.6 mg/m3.1
Cell proliferation studies carried out
after FA exposure with Brdll labeling (i.p.
injection 16-22 hours postexposure) and
flow cytometry analysis of 10,000 cells
per measurement.
Nose
1 exposure
5
1.3 (0.1)b
2.4(0.6)°
3.7 (0.5)°
2.7 (0.8)°
3 exposures
5
NR
1.4 (0.3)
2.5 (0.2)°
2.3 (0.2)°
Trachea
1 exposure
5
1.2 (0.1)
3.1(0.6)°
2.1(0.8)
2.8(0.4)°
3 exposures
5
NR
0.3 (0.1)°
0.6 (0.1)°
2.5 (0.2)°
Lung
1 exposure
3
1.8(0.3)
2.6 (0.6)
3.3 (0.4)
3.1(0.7)
3 exposures
3
NR
2.2 (0.0)
2.4(0.7)
5.1(1.5)
aTwice the number of rats in control groups; Standard error in
parentheses;
statistically significant at p <0.05, compared with controls.
Wilmer et al. (1987)
Wistar rats; male; 10/group.
Exposure: Rats were exposed to FA
(chamber type not reported) either
continuously for 8 hours/day, 5
days/week for 4 weeks or intermittently
8 hours/day (successive periods of 0.5
hour of exposure and 0.5 hour of
nonexposure), 5 days/week for 3 days
and 4 weeks.
Test article: Paraformaldehyde.
Actual concentrations were not
determined. Target concentrations
were 0, 6.2, or 12.3 mg/m3 for
continuous exposures and 0,12.3, or
24.6 mg/m3 for intermittent exposures.1
Cell proliferation studies carried out
after 3 days or 4 weeks of FA exposure
with [3H]thymidine labeling (ip injection
18 hours postexposure) and scoring of
the cells (n=5000) lining the nasal and
maxillary turbinates, the septum, and
the lateral wall.
Percentage of [3H]thymidine labeled cells in nasal epithelium
% labeled cells
Exposure
Exposure x time
After 3 days of
exposure
(n=3)
After 4 weeks
of exposure
(n=3)
0 mg/m3
0 mg/m3h/day
0.86 (0.14)a
0.68 (0.12)
6.2 mg/m3
(continuous)
49.6
mg/m3h/day
2.82 (0.47)b
1.33 (0.75)
12.3 mg/m3
(continuous)
98.4
mg/m3h/day
8.87 (1.51)b
8.85°
12.3 mg/m3
(intermittent)
49.2
mg/m3h/day
9.80 (1.54)d
3.41 (1.25)e
24.6 mg/m3
(intermittent)
98.4
mg/m3h/day
19.77 (2.39)d
13.87 (0.64)d
aSDs shown in parentheses; bp<0.01, compared to controls; cData from one
rat; dp<0.001, compared to controls; ep<0.05, compared to controls.
Medium Confidence
Cassee and Feron (1994)
Wistar rats; male; 20/group.
Exposure: Rats were exposed in
dynamic nose-only chambers for 3 day (6
consecutive 12-hour periods of 8 hours
of exposure to FA followed by 4 hours of
nonexposure). Rats sacrificed
immediately (i.e., within 30 minutes)
after last exposure.
Test article: Paraformaldehyde.
Controls
FA olonea
Site
llb
lllb
II
III
Nasoturbinates
+c
+
+++
+++
Maxilloturbinates
+
+
+++
+++
Septum
+
+
+++
+++
Lateral wall
+
+
+++
+++
aOnly nonnecrotic areas at cross level II showed severe PCNA expression;
Standard cross level II and III through the nose; cPCNA-expression scores:
+, some nuclei stained; ++, a moderate number of nuclei stained; +++,
many nuclei stained.
This document is a draft for review purposes only and does not constitute Agency policy.
A-533 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Actual concentrations were 0 and 4.4 (SE
±0.1) mg/m3 FA alone.1
Cell proliferation studies carried out
using deparaffinized standard cross
sections of the nose and
semiquantitative proliferating cell
nuclear antigen (PCNA) immunostaining.
See diagram from (Cassee et al..
1996b)
(above) for cross sections of a rat nose
examined for PCNA staining by Cassee
and Feron (1994).
In animals exposed to FA alone, no increased PCNA staining observed in
olfactory epithelium.
This study also evaluated the combined effects of FA and ozone mixtures
on nasal epithelium. Ozone co-exposure resulted in an increase in
proliferation compared to formaldehyde exposure alone. Data are only
presented herein for formaldehyde-only exposures.
Speit et al. (2011)
Fischer 344 rats; males; 6/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers 6
hours/day, 5 days/week for 4 weeks.
Test article: Formalin (methanol
concentration NR).
Actual concentrations were 0, 0.63
(±0.6), 1.23 (±0.14), 2.48 (±0.18), 7.53
(±0.42), 12.3 (±0.48), 18.4 (±0.06)
mg/m3.1
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index of 3 levels of
the nasal cavity: I (nasal septum, lateral
meatus [wall], maxilloturbinate,
nasoturbinate), II (nasal septum, lateral
meatus [wall]), and IV (nasopharynx).
Cell proliferation data reported as Brdll-
labeled nuclei per mm of basal lamina
(i.e., ULU).
ULLI for level III not assessed due to author's expectation that this level
was not a sensitive target tissue.
ULLI for nasal level I
mg/m3
Nasal
septum
Lateral
meatus
Maxillo-
turbinate
Naso-
turbinate
0
6.64±1.30a
8.44±3.37
10.21±5.90
14.15±2.93
0.63
8.02±2.57
10.80±1.58b
9.49±3.07
17.13±6.97
1.23
6.04±2.20
9.56±3.68
10.43±5.52
22.60±5.86c
2.48
6.14±3.15
11.56±4.73
9.08±2.65
14.29±5.59
7.53
4.80±3.14
14.85±2.40c
12.95±3.94
20.48±8.12b
12.3
3.83±2.13
52.53±16.30c
52.42±16.88c
74.63±28.90c
18.4
70.86±14.30c
74.21±16.37c
81.96±2.90c
67.50±12.76c
aGroup mean value±SD; bp<0.05; cp<0.01.
ULLI for nasal level II
ULLI for nasal level
IV
mg/m3
Nasal septum
Lateral meatus
Naso-pharynx
0
14.59±6.37a
9.33±4.22
17.81±2.18a
0.63
19.93±7.66
7.58±2.32
21.23±5.19
1.23
22.36±7.04b
8.04±2.92
21.56±3.17
2.48
21.79±5.28b
9.47±3.31
21.33±3.55b
7.53
19.07±6.43
9.28±3.54
20.93±4.13
12.3
26.66±11.31
37.13±5.22c
29.23±4.25c
18.4
62.36±12.30c
55.21±10.99c
73.29±15.87c
aGroup mean value±SD; °p <0.05; cp <0.01.
Relative change (% control) in ULLI in metaplastic/ degenerative (M)
and nonmetaplastic (O) epithelia
Lateral
Maxillo-
Naso-
Nasal septum
meatus
turbinate
turbinate
mg/m3
M
O
M
O
M
O
M
O
Level I
This document is a draft for review purposes only and does not constitute Agency policy.
A-534 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
12.3
58
61
622ab
1195a
513a'c
262a
527ac
139
18.4
1066a
1386a
879ac
1399a
802a
735a
477a'b
280d
Level II
12.3
183
161
398ac
110
NA
NA
NA
NA
18.4
428ac
1188a
592ac
195a
NA
NA
NA
NA
ap <0.01, compared to corresponding untreated control; bp <0.05,
comparison between metaplastic and nonmetaplastic tissues; cp<0.01,
comparison between metaplastic and nonmetaplastic tissues; dp <0.05,
compared to corresponding untreated control.
Woutersen et al. (1987)
Wistar rats; male and female;
10/sex/group.
Exposure: Rats were exposed to FA in
dynamic whole-body chambers for 6
hours/day, 5 days/week for 3 days.
Test article: Paraformaldehyde.
Actual concentrations were 0,1.2
(±0.00), 11.9 (±0.15), and 24.4 (±0.09)
mg/m3.1
Cell proliferation studies carried out
after 3 days of FA exposure with
[3H]thymidine labeling of dissected
nasoturbinates (18 hours postexposure)
and scoring of the cells (n=1000) of the
respiratory epithelium.
Percentage of [3H]thymidine labeled cells in nasal epithelium (males,
n=2/group)
% labeled cells
mg/m3
Visibly unaffected
epithelium
Metaplastic epithelium
0
1.6 (1.2-2.0)3
NR
1.2
1.2 (0.8-1.5)
NR
11.9
2.6 (1.4-3.8)
31.4 (29.5-33.2)
24.4
2.8b
37.6 (32.6-42.5)
aRange in parentheses; bValue based on one rat since most respiratory
epithelium was metaplastic.
Mice
High Confidence
Chang et al., 1983: [additional data from
related Swenberg et al. (1983) reportl
B6C3F1 mice; males; 4-5/exposure
group, 10/control group.
Exposure: Mice were exposed to FA in
head-only chambers 6 h/day for either 1,
3, 5 or 10 days.
Test article: Paraformaldehyde.
Actual concentrations were 0 and 18.5
(±0.1) mg/m3.1 Target concentrations
were 0, 0.62, 2.46, 3.69, 7.38, 14.76 or
18.45 mg/m3 in Swenberg et al. (1983)
report.
Cell proliferation studies carried out
after FA exposure with [3H]thymidine
labeling (ip injection 2 or 18 hours
postexposure) and scoring of cells
(n=4000) lining the respiratory
epithelium from the nasal and maxillary
turbinates and lateral wall.
Group (18.5 mg/m3)
Labeling index (%) in Level B
Control
0.27±0.04 (10)a
1 day
2.14±0.56 (5)b
5 days
3.42±0.84 (4)b
aNumber in parentheses represents number of animals studies.
Significantly different from control, p <0.05.
% labeled respiratory epithelial cells in Level B (thymidine at 2 h
postexposure)
Formaldehyde Concentration (mg/m3)
0
0.62
2.46
7.38
18.45
3 days
0.12
(0.02)
0.09 (0.04)
0.08 (0.04)
0.15 (0.06)
0.97 (0.04)
% labeled respiratory epithelial cells in Level A (thymidine at 18 h
postexposure)
Control
3.69 mg/m3 x 12 h/ day for 10 days
1.24 (0.57)
10.14 (3.20)
This document is a draft for review purposes only and does not constitute Agency policy.
A-535 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
See diagram from Swenberg et al. (1983)
for rats (above) for locations of Levels A
(with minimal mucociliary clearance) and
B (with extensive mucociliary clearance)
7.38 mg/m3 x 6 h/ day for 10 days
4.72 (1.61)
14.76 mg/m3 x 3 h/ day for 10 days
1.76 (0.49)
Mean (SEM); Group sizes and statistical comparisons not reported in
Swenberg et al., 1983
(Kuper et al., 2011)
B6C3F1 mice; females; 6/group.
Exposure: Mice were exposed to FA in
dynamic whole-body chambers 6 h/day,
5 day/wk for 4 wk.
Test article: Formalin (10.21% FA).
Actual concentrations were 0, 0.63
(±0.06), 1.23 (±0.14), 2.48 (±0.18), 7.53
(±0.42), 12.3 (±0.48), and 18.4 (±0.06)
mg/m3.1
Cell proliferation studies conducted with
surgical implantation of Brdll-containing
pumps (3 days prior to sacrifice) and
determining labeling index of 2 sections
of NALT and 1 section of a upper-
respiratory tract-draining lymph node
(i.e., posterior and superficial cervical
lymph nodes). Cell proliferation data
reported as Brdll-positive cells per
length (i.e., mm) of epithelium.
NALT: No FA-related effects on the number of Brdll-positive cells reported
in the follicular and interfollicular compartments and epithelium
Lymph nodes: No FA-related effects on the number of Brdll-positive cells
reported in the follicle and paracortex compartments and medulla
Monkeys
Medium Confidence
(Monticello et al., 1989)
Exposure
Observations between nasal passage epithelia
Rhesus monkeys; male; 3/group.
Exposure: Monkeys were exposed to FA
in dynamic whole-body chambers 6
hours/day, 5 days/week for 1 or 6
weeks.
Test article: Paraformaldehyde.
Actual concentrations were not
determined. Target concentration was
7.4 mg/m3. Controls were sham
exposed to biologically filtered air for 6
weeks.1
Cell proliferation studies carried out
after FA exposure with [3H]thymidine
labeling (iv injection 18 hours
postexposure) and scoring of respiratory
epithelial cells. For nasal passages
(transitional, respiratory, and olfactory
epithelia), larynx, trachea, and carina, Lis
Controls
(6 wk)
Highest Lis in transitional epithelium compared to
respiratory and olfactory epithelia
7.4 mg/m3
(1 wk)
Transitional and respiratory epithelia elevated compared
to controls (p <0.05)
7.4 mg/m3
(6 wk)
Exposure
Transitional epithelium Lis slightly elevated over controls
and had decreased from 1-week group; olfactory
epithelium Lis had mild increase over controls (p <0.05);
respiratory epithelium Lis elevated compared to controls
(p <0.05)
Observations between levels of nasal passages
Controls
(6 wk)
Lis for Levels B-E significantly increased over controls (p
<0.05), anterio-posterior gradient (i.e., greatest to lowest)
in cell proliferation rates
7.4 mg/m3
(1 wk)
Lis for Levels B-E significantly increased over controls
(p<0.05)
7.4 mg/m3
(6 wk)
Levels C-E significantly elevated over 1-week group (p
<0.05)
This document is a draft for review purposes only and does not constitute Agency policy.
A-536 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
Group
Observations within levels of nasal passages
Level A
NR
Level B
Lis for 1- and 6-week groups elevated over controls (p
<0.05) for septum, inferior meatus, inferior turbinate,
lateral wall, and middle turbinate
Level C
Lis for 1- and 6-week groups elevated over controls (p
<0.05) for septum, inferior meatus, inferior turbinate,
lateral wall, and middle turbinate; no increase in Lis for 1-
and 6-week groups over controls for maxillary sinuses
Level D
Lis for 1-week group elevated over controls (p <0.05) for
septum, inferior meatus, inferior turbinate, and lateral
wall; Lis for 6-week group elevated over controls (p <0.05)
for inferior meatus and inferior turbinate
Level E
Lis for 1-week group elevated over controls (p <0.05) for
floor and lateral and dorsal walls; Lis for 6-week group
elevated over controls (p<0.05) for septum, floor, and
lateral and dorsal walls
Group
Observations for nonnasal tissues
Larynx
Lis for 1- and 6-week groups elevated over controls; Lis
increased with duration of exposure
Trachea
Significant elevation in Lis for 1-week (p <0.05) but not 6-
week group over controls; Lis increased with duration of
exposure
Carina
Significant elevation in Lis for 1-week (p <0.05) but not 6-
week group over controls; Lis increased with duration of
exposure
defined as the number of labeled cells
per mm of basal lamina.
A B C D E
Figure 4 from (Monticello et al..
1989) depicting the nasal passage levels
selected for cell proliferation studies. A,
nasal atrium; B, anterior aspect of the
middle and ventral turbinates; C, mid-
region of the maxillary sinuses; D,
posterior nasal cavity; and E,
nasopharynx.
Interanimal variation in Us for trachea and carina
Exposure
Animal #
Trachea U
Carina U
Controls (6 wk)
1
0.29
0.42
2
0.46
0.37
3
0.91
0.50
ave
0.55±0.19a
0.43±0.04a
7Amg/m3 (lwk)
4
1.34
1.09
5
0.90
1.95
6
1.19
0.99
ave
1.14±0.13a
1.34±0.31a
lAmg/m3 (6 wk)
7
8.00
3.86
8
2.30
6.49
9
0.88
0.45
ave
3.73±2.18a
3.60±1.75a
Represents MeaniSEM.
Exposure
Controls (6 wk)
LI in respiratory bronchioles0
0.01+0.001
This document is a draft for review purposes only and does not constitute Agency policy.
A-537 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Toxicological Review of Formaldehyde—Inhalation
Reference and study design
Results
7Amg/m3 (lwk)
0.01±0.003
7Amg/m3 (6 wk)
0.01+0.001
aUs expressed as
percent labeled cells per total cell count from >500
respiratory bronchiolar nucleated epithelial cells per animal.
Changes in the LRT
Although the URT and the LRT are physically and functionally connected, this analysis
delineates findings across these two tissue compartments. This was done due to the distribution of
the overwhelming majority of inhaled formaldehyde to the URT (noting that some data suggest that
oronasal breathing in humans, as compared to nose-only breathing in rodents, might result in slight
differences in the distribution of inhaled formaldehyde, including a possible increase in the portion
reaching proximal regions of the LRT such as the trachea; see Appendix A.2). Thus, evidence
related to studies of BAL (bronchoalveolar lavage) fluid and airway function, both of which may
involve some contribution from URT-related changes but are largely driven by effects on the lung,
are described in this section. The specific studies and summary findings supporting the synthesis
below are described in Table A-78. In general, compared to effects on the URT, the methodological
approaches for evaluating LRT changes are more commonly applied to studies of exposed humans,
so this section considers a wider range of evidence. A greater level of concern exists for the
erroneous attribution of changes in the LRT (and other, non-URT, compartments in subsequent
sections) to inhaled formaldehyde when studies used methanol-containing formalin; thus, findings
from some studies using exposure paradigms similar to those described in the previous section are
interpreted with comparably less confidence.
As previously mentioned, formaldehyde-induced stimulation of TRPA1 receptors on
trigeminal nerve endings distributed within the epithelial cell layer in the URT appears to cause a
localized release of neuropeptides, including substance P, which can cause local inflammatory
changes. Consistent with this, ex vivo models of LRT tissues and low confidence studies of in vivo
exposure suggest that indirect activation of sensory nerve endings in the LRT, presumably of the
vagus nerve, occurs after formaldehyde inhalation exposure. In the URT, this activation is expected
to occur via direct interaction of formaldehyde with receptors. However, while these direct
interactions might occurn in upper portions of the LRT during certain, very rare human exposure
scenarios (e.g., in the trachea at high exposure levels), they would be unexpected in the lungs or
during typical exposure scenarios; thus, this is not considered a plausible initial effect of typical
exposure. Notwithstanding this assumption, the available evidence indicates that formaldehyde
exposure likely causes downstream sequelae in the lung that could be attributed to sensory nerve
activation in the LRT, predominantly related to substance P-related pathways (see below).
However, the mechanistic event(s) critical to understanding this potential relationship remain
unknown: namely, how sensory nerve endings in the LRT would be stimulated without distribution
This document is a draft for review purposes only and does not constitute Agency policy.
A-538 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde—Inhalation
of inhaled formaldehyde to the LRT. The most likely explanations involve a secondary response to
TRP channel-activating stimuli increased via other mechanisms, such as increased LRT oxidative
stress and/or inflammatory mediators released from activated immune cells or damaged epithelial
cells in the LRT. It could also be explained by a central trigeminal-to-vagal neural reflex response to
irritation of the URT (i.e., a "nasobronchial" reflex18); however, the existence of this reflex in
humans is debated and a clear scientific consensus does not exist (Sahin-Yilmaz and Naclerio, 2011;
Togias, 1999: 54 and 2004: 113; giavina-bianchi etal., 2016: 9). No studies specifically designed to
assess any of these potential linkages after formaldehyde exposure were identified.
Studies in several species provide moderate evidence that formaldehyde exposure results in
increased LRT neuropeptides, including substance P (see "Changes in the URT" Section above), as
well as a rapid activation of the primary receptor for substance P, the neurokinin receptor (NKiR),
typically at formaldehyde concentrations >2.5 mg/m3. Further, the activation of this pathway has
been experimentally linked to both formaldehyde-induced leakage of the LRT microvasculature
(which has been observed in rodents at >1.23 mg/m3) as well as airway hyperresponsiveness
(which has been observed in animals and humans at <0.5 mg/m3). In addition to facilitating the
recruitment of inflammatory cells, NKiR activation can promote immune cell survival and
activation through the release of cytokines and chemokines [Tulec et al., 2009], The substance
P-NKiR pathway has been implicated in mast cell degranulation, which can lead to
bronchoconstriction (Van der Kleij and Bienenstock, 2005: 65-80); however, while inhibiting mast
cell activation prevented microvascular leakage in a low confidence rat study after acute exposure
to high levels of formaldehyde (Kimura et al., 2010), an acute medium or high confidence study of a
cohort of guinea pigs failed to observe any changes in mast cells (Swiecichowski et al., 1993;
Leikauf et al., 1992). Importantly, an understanding of potential changes to substance P and NKiR-
dependent effects (e.g., due to desensitization) with long-term formaldehyde exposure remains
unclear. While a transient depletion of neuropeptides from sensory nerve terminals after acute
exposure seems plausible (see Kimura et al., 2010), substance P is still elevated, at least in the
blood, after subchronic exposure (Fujimaki et al., 2003). Overall, the activation characteristics of
this pathway in the LRT across various formaldehyde exposure scenarios have not been
established.
Microvascular leakage can lead to inflammatory structural changes observable by histology,
which are supported by moderate evidence in formaldehyde-exposed rodents, particularly those
sensitized with the allergen, ovalbumin (OVA). The available studies indicate changes including
airway edema (swelling) or thickening of airway walls, with general support for inflammatory
changes in airway bronchi, but not necessarily alveoli. In addition, the pattern of structural changes
varied across studies, with a study in guinea pigs observing airway swelling without signs of
18 Note: neural reflexes involving afferent and efferent activity of the vagus nerve (e.g., across different LRT
regions), some of which may involve C fibers and TRP channels, are better established (Mazzone and Undem,
2016: 96).
This document is a draft for review purposes only and does not constitute Agency policy.
A-539 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
inflammation at low formaldehyde (<0.5 mg/m3) levels (Reidel etal., 1996), while studies in rats
and mice generally observed mild inflammatory-related structural changes at higher levels (i.e.,
>3.0 mg/m3) that only became pronounced with allergen sensitization. It is important to note that
animal models vary in their ability to mimic some features of human airways. Airway responses in
guinea pigs often differ from those in rats and mice, and while no animal model fully recapitulates
human airway function, in many ways the sensitivity of guinea pig airways may be more relevant
than other small mammals (e.g., similar structure of the lung to humans; responsiveness to stimuli
that induce sensitivity in humans) [Shin et al., 2009; Ricciardolo et al., 2008], Alongside airway
inflammation and structural changes, including edema, which could narrow or obstruct airways, an
increased permeability to bronchoconstrictors such as histamine would be expected to influence
airway function, possibly linking these changes to observations of hyperresponsiveness or
decreased pulmonary function.
A moderate association between formaldehyde exposure and increases in LRT eosinophils
was identified, including amplification of the response of these cells in rodents previously exposed
to allergens (see Table A-79). Taken together with similar findings in the URT, a general increase in
airway eosinophils as a result of formaldehyde exposure is supported by robust evidence. As in the
URT, this finding has been reported in the LRT following exposure for several weeks at effective
concentrations above 0.5 mg/m3. The only study of longer-term exposure available (Fujimaki,
2004) indicated that formaldehyde exposure at 2.46 mg/m3, but not «0.5 mg/m3, for three months
caused increased eosinophils in mice sensitized to OVA, but not in unsensitized mice. While the
data are not conclusive, it appears that eosinophil recruitment does not occur immediately after
acute exposure, as this increase was not observed in the available studies of acute exposure (see
Table A-79). Although it has not been mechanistically demonstrated based on increased
eosinophils and other immune cells after acute tachykinin release [Barnes etal., 1998], repeated
release of neuropeptides could plausibly lead to sustained airway inflammation and, depending on
the phenotype of the recruited cells, this could result in airway hyperresponsiveness. In both the
URT and LRT, recruitment of eosinophils might also be related to changes in markers of oxidative
stress observed across formaldehyde exposure paradigms. However, whereas oxidative stress in
the URT may be related to damage to the local epithelial cells, most studies indicate that
formaldehyde exposure does not result in overt damage to the LRT airway epithelium (slight
evidence, at relatively high formaldehyde levels: >5 mg/m3), making this potential linkage less
plausible. It is considered more likely that increases in oxidative stress are the result of changes in
inflammatory factors and immune cells in the LRT, rather than LRT epithelial damage.
The evidence for LRT immunological changes other than those seen in eosinophils is mixed
and generally only suggestive of potential effects. As shown in Figure A-34, slight evidence exists to
suggest that formaldehyde exposure amplifies recruitment of innate immune cells such as
neutrophils and monocytes to the LRT; notably, this finding has only been observed when animals
exposed to >2 mg/m3 were previously sensitized to an allergen. Importantly, few studies examined
This document is a draft for review purposes only and does not constitute Agency policy.
A-540 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
lymphocyte subsets, and no studies reported on the response of lymphocytes in animals sensitized
to allergens or at exposure levels below 5 mg/m3, highlighting important gaps in the literature.
Two studies suggest that CD8+, but not CD4+, T cells may be increased with formaldehyde exposure
above 7 mg/m3 [Sandikci et al., 2007b; Jung et al., 2008], The only study meeting the inclusion
criteria that evaluated lymphocyte changes in both immature and adult animals only observed
changes in animals exposed as adults [Sandikci et al., 2007b], which could suggest that a
functionally mature immune system is necessary for these alterations (the immune system is not
considered to be fully mature in rodents until around six weeks of age [Burns-Nass et al., 2008]).
While these findings should be interpreted with substantial caution, there may be a role for CD8+ T
cells in promoting the recruitment and survival of airway eosinophils, as well as a requirement of
these cells for the development of airway hyperresponsiveness (e.g., to allergen or infection)
[Hamelmann et al., 1996; Schwarze et al., 1999], CD8+ T cells make up a heterogeneous population
of lymphocytes which migrate by recruitment to sites of inflammation, proliferate in response to
antigen stimulation, and help to mediate long-term cellular immunity against foreign pathogens,
particularly viruses. The conventional role for IFNy-producing CD8+ T cells is to inhibit eosinophil
function; however, some emerging evidence suggests that certain CD8+ T cell subpopulations may
induce eosinophil recruitment [Huber and Lohoff, 2015], No data are available to evaluate the
potential for effects of formaldehyde exposure on different subpopulations of LRT CD8+ T cells.
Studies of markers of immune cell activation in the LRT after formaldehyde exposure
generally provide mixed results, making it difficultto draw inferences (see Table A-79). Most
cytokine-related changes reported in the LRT occur at high formaldehyde levels (>5 mg/m3) after
short-term exposure and include slight evidence to support an increase in eosinophil chemotactic
factors, and a decrease in markers and counts of natural killer (NK) cells. NK cells respond rapidly
to infection and appear to have a role in regulating chronic inflammation and infection of the
airways [Cully, 2009], Thus, this change, were it to be experimentally verified, could be associated
with the moderate evidence of an increased propensity for LRT infections, similar to the slight
evidence of altered URT immune responses (see previous section); however, definitive studies
relevant to long-term exposure have not been identified and additional data are necessary to
interpret these alterations in respiratory immune responses as consistent with immune
suppression. A number of consistent studies in exposed rodents do suggest an increase in T helper
type 2 (Th2)-related cytokines, most notably IL-4, with short term exposure at >0.5 mg/m3 and
particularly in animals sensitized to an allergen. The slight evidence supporting increased IL-5, a
Th2 cytokine that can be both synthesized by and act upon airway mast cells and eosinophils and
which is believed to be integral to the development of airway eosinophilia and airway
hyperresponsiveness [Schwarze et al., 1999; Greenfeder et al., 2001], is considered to be
inconclusive (i.e., two low confidence studies testing exposure levels >5 mg/m3). Along with IL-5
and IL-13, IL-4 is recognized for its established role in chronic respiratory disorders [Maes et al.,
2012], and this change may be relevant to other LRT-specific changes. IL-4, which can stimulate T
This document is a draft for review purposes only and does not constitute Agency policy.
A-541 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
cell receptors on CD4+ and CD8+ T cells [Serre etal., 2010], can influence the activation and
development of antigen-specific CD8+ T cell immunity by shifting the phenotype of these cells from
IFN-y production to IL-4 production [Erb and Le Gross, 1996],
The cytokine changes could be related to the moderate evidence for increased LRT
infections and the slight evidence suggesting reduced NK cell numbers (see Tables A-79 and A-73),
as Th2 cytokines have been shown to reduce pulmonary bacterial immunity [Beisswenger et al.,
2006] and NK cells have a role in regulating chronic inflammation and infection of the airways
[Cully, 2009], A key limitation of the data is that the few formaldehyde-specific studies have not
demonstrated consistent increases in CD4+ Th2 cells in the airways of exposed individuals.
Similarly, interactions between airway innate and adaptive immune responses, and between CD4+
and CD8+ T cells, topics of current interest [Gasteiger and Rudensky, 2014; Koya et al., 2007], have
not been well studied following formaldehyde exposure. Experiments focused on these types of
endpoints would help to integrate the currently available data.
The consistent evidence of amplified airway responses to immunogenic stimuli (e.g., to
allergens such as OVA) following formaldehyde exposure is of particular interest As described
above, multiple LRT parameters are affected or exacerbated by the combination of formaldehyde
exposure and sensitization to allergenic materials. At concentrations ranging from 0.31-3 mg/m3
over durations of several days to several weeks, formaldehyde exposure in combination with
allergen sensitization exacerbates immune-related changes, such as: recruitment of eosinophils and
possible increases in IL-4; airway structural changes, including edema; and airway functional
changes, including exaggerated responses to muscarinic receptor agonists. These observations may
be relevant to the associations between human formaldehyde exposure at much lower
concentrations (<0.05 mg/m3) and conditions that may reflect an enhanced response to allergens
(e.g., rhinoconjunctivitis; asthma).
The formaldehyde exposure-induced effects associated with allergen sensitization varied
depending on the specific mechanistic effect and the experimental animal model. This variability
may reflect a lack of consistency in the methods used for sensitization and challenge, or other
experimental design differences across studies. Alternatively, these differences might reflect
variability in susceptibility to these types of effects across different populations or groups of
individuals (e.g., animals of different species, strains, sex, or age). This variable sensitivity of
subsets of the population to formaldehyde-induced effects would be consistent with observations
of substantial inter individual human variability for several potential health effects. Further, these
data suggest that vulnerability to some formaldehyde-induced health effects might be influenced by
the exposure history of the individuals, including exposure to known allergens. The mechanism for
this amplified response to allergens (and, possibly, nonallergenic antigens) due to formaldehyde
exposure, including what airway component(s) formaldehyde may interact with to initiate this
particular alteration, remains unknown. Possible explanations include formaldehyde acting as an
antigen (capable of directly eliciting an antibody response) or as a hapten (capable of eliciting an
This document is a draft for review purposes only and does not constitute Agency policy.
A-542 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde—Inhalation
antibody response when bound to a larger molecule such as a protein), or formaldehyde-induced
chronic inflammation acting as an adjuvant (enhancing immune responses to antigens); however,
these speculations have not been examined by directed testing following inhalation exposure.
While changes in airway responsiveness could be dependent on stimulation of sensory nerve
endings, observations in isolated tracheae by Swiecichowski et al. (1993; Leikauf et al., 1992)
suggest that the amplified response to stimuli is at least partly mediated by interactions with local
immuno-modulatory factors. As airway hyperreactivity and other indicators of immunologic
sensitization are known to be related to markers (e.g., antibodies) in the blood, some evidence
related to these responses are discussed in the subsequent section. Overall, the essential airway
immunologic target(s) of inhaled formaldehyde has not yet been identified and verified, thereby
presenting a key uncertainty.
This document is a draft for review purposes only and does not constitute Agency policy.
A-543 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-78. Summary of changes in the lower respiratory tract (LRT) as a result of formaldehyde exposure
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
Structural Modification of the Lower Airways
Microvascular
Leakage
High or
Medium
Human: None
Demonstrated increased leakage from acute
exposure >6.15 mg/m3 in 1 study, which
might be mediated by substance P
Moderate 'f
Animal: Increased in rats (Ito, 1996): acute at >6.15 mg/m3; note: inhibited at 18.45 mg/m3 by NK1
receptor antagonist (note: substance P binds NKiR), but not histamine or bradykinin antagonists
5
o
1
Human: None
One study suggests acute exposure as low as
1.23 mg/m3 induces microvascular leakage,
although continued exposure appeared to (at
least in the near-term) result in less leakage
Animal: Transiently increased in rats (Kimura, 2010): acute at >1.23 mg/m3 (duration-independent);
Note: leakage blocked by inhibiting mast cells, but not blocking cyclooxygenases; potential additional
mechanistic understanding by injection of formalin into the trachea causing leakage that appeared to
be dependent on substance P release after stimulation of C-fiber afferents (Lundberg and Saria, 1983)
Airway Edema
and/or Other
Inflammatory
Structural
Change
High or
Medium
Human: None
Bronchial edema in 1 short-term study at
0.31 mg/m3
Moderate 'f
may require
higher
exposure levels
and/or allergen
sensitization for
pronounced
changes
Animal: Increased edema in lung bronchi, but not alveoli, without signs of inflammation in lower
airways in guinea pigs (Riedel, 1996): 5 d at 0.31 mg/m3, not 0.16 mg/m3
5
o
1
Human: None
Airway structural changes with allergen
sensitization in 2 species (and, to a lesser
extent, without sensitization) with short-term
exposure at >3 mg/m3
Animal: Airway structural changes consistent with inflammation (e.g., wall thickening; cell infiltration)
in mice (Jung, 2007; Wu, 2013—slight; Liu, 2011—slight) and in mice and rats sensitized with OVA (Wu,
2013; Liu, 2011; Qiao, 2009), but not in nonsensitized rats (Qiao, 2009): all 2-3 wk at >3 mg/m3 [Note:
most studies indicated assessment of bronchial airways]
Airway/Airway
Epithelial Cell
Damage
High or
Medium
Human: None
N/C in a single mouse subchronic study with
i.p. sensitization and up to 2.46 mg/m3
exposure, nor in a guinea pig study at 4.18
mg/m3
Slight
at higher
formaldehyde
levels
Animal: N/C (histology for mouse epithelial cell damage) (Fujimaki, 2004): 12 wk at up to 2.46 mg/m3
N/C in histology in guinea pigs (Swiecichowski, 1993; Leikauf, 1992): acute at 4.18 mg/m3
5
o
1
Human: None
A single short-term study in mice and another
in rats, and indirect evidence from several
studies in rats, suggests damage at higher
formaldehyde levels (e.g., around 4 mg/m3);
however, another similar study did not
observe effects at 12.3 mg/m3
Animal: Increased in mice (Jung, 2007): 2 wk at >6.15 mg/m3 and in rats (Aydin et al., 2015): 4 wk at
>6.15 mg/m3; indirect evidence of damage in rats (Kimura, 2010 and Dallas, 1987 and Sandikci, 2009):
20 h after acute at 6.15 mg/m3 and 1 wk at >0. 62 mg/m3 (effect magnitude decreased with longer
exposures) and 6 wk at 7.38 mg/m3 (in adults, not young), and in mice (de Abreau): 6-8 h after acute
at 3.7 mg/m3, but N/C in rats in another study (Dinsdale, 1993): 4 d at 12.3 mg/m3
This document is a draft for review purposes only and does not constitute Agency policy.
A-544 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-78. Summary of changes in the lower respiratory tract (LRT) as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
LRT Sensory
Nerve
Activation
High or
Medium
Human: None
No evidence to evaluate
Slight
levels required
for potential
activation
unknown (note:
may involve
TRPA1 binding)
Animal: None
5
o
1
Human: None
A single acute rat study and indirect
evidence from potentially related
exposures suggest that lower airway
sensory nerve afferents may be
activated, but the inhaled formaldehyde
levels required for such potential
activation have not been experimentally
demonstrated
Animal: With acute exposure, dose-dependent increase in nerve currents and CI" release
in intact rat trachea (Luo et al. 2013), with supporting evidence of substance P and NK
Receptor involvement. Indirectly, increased substance P and CGRP were observed in
mouse lung tissue, both were amplified with OVA, and both were dependent on TRP
activation (Wu, 2013): short term at 3 mg/m3. Note: the potential involvement of
tracheobronchial reflexes in the pulmonary effects of cigarette smoke constituents, such
as nicotine and formalin, may add indirect support
Immune and Inflammation-Related Changes
[[See Table 1-30 for Cellular and Cytokine Response in BAL and LRT tissues]]
Oxidative
Stress
High or Medium
Human: Increased exhaled nitric oxide, a noninvasive marker of lower airway
inflammation and oxidative stress, in healthy or asthmatic children (Franklin et al.,
2000); Flamant-Hulin et al., 2010): unknown duration (likelv months to vears: classrooms
or homes) at 0.04-0.06 mg/m3, but not in elderly nursing home patients at lower levels
(Bentaveb et al., 2015): unknown duration (likelv months to vears) at 0.005-0.01
mg/m3
Increased biomarkers (indirect
evidence) of oxidative stress in children
at >0.04 mg/m3, but not in elderly
individuals at <0.01 mg/m3 with
prolonged (months-vears) exposure
Moderate T*
in children at
=0.04 mg/m3
Animal: None
5 £
Human: None
This document is a draft for review purposes only and does not constitute Agency policy.
A-545 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-78. Summary of changes in the lower respiratory tract (LRT) as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
Animal: in mice: NO and NOS activity increased with 3 d at 3 mg/m3 (Yan, 2005), GSH
levels decreased with 3 wk at >0.5 mg/m3 (Ye, 2013), and increased ROS and/or lipid
peroxidation markers with 3 wk at >1 mg/m3 (Ye, 2013) or 2 wk at >6.15 mg/m3 (Jung,
2007), but decreased with acute exposure in 1 study (Matsuoka et al., 2010): 24 h at
0.12 mg/m3
in rats: at >12.3 mg/m3 increased total oxidant levels and decreased total antioxidant
level (Aydin et al., 2015): 4 wk, increased lipid peroxidation markers and protein oxidation
markers (Sul et al., 2007): 2 wk, and decreased gamma-glutamyl transpeptidase-
indirect evidence (Dinsdale, 1993): 4 d
Multiple studies in two species suggest
elevated oxidative stress at >1 mg/m3
with short-term exposure
Sustained
Inflammation
High or Medium
Human: Increased exhaled nitric oxide, a noninvasive marker of lower airway
inflammation and oxidative stress, in healthy or asthmatic children (Franklin et al.,
2000); Flamant-Hulin et al., 2010): unknown duration (likelv months to vears: classrooms
or homes) at 0.04-0.06 mg/m3
Immune cell counts are continually
elevated in a subchronic mouse studv
with allergen stimulation at 2.46 mg/m3;
increased biomarkers (indirect
evidence) of lower airway inflammation
are observed in children with prolonged
exposure.
Moderate
may require
allergen
sensitization in
some cases
Animal: Eosinophils and monocyte counts remain elevated with continued exposure for
subchronic duration with allergen (OVA) sensitization (Fujimaki, 2004): 12 weeks at 2.46
mg/m3
5
o
1
Human: None
BAL cell counts and histologic evidence
suggest that inflammation persists for
several weeks with short-term
exposure, and these effects are
amplified by allergen
Animal: Immune cell counts were increased with short term exposure in several studies at
>0.5 mg/m3 (see Table 1-30); histological evidence of inflammation without epithelial
damage was noted in short-term studies, typically at higher concentrations, which were
amplified by allergen (e.g., >3 mg/m3; Wu, 2013; Kimura, 2010)
Immune
Function
(inferred from
LRT infections)
High or Medium
Human: Increased LRT infections in infants (Roda et al., 2011): 32-41% increase in
incidence per 0.0124 mg/m3 increase in formaldehyde (LOD: 0.008 mg/m3); =1-year
exposure at 0.020 mg/m3 (median)
Indirect evidence in a single study of
infants exposed to a median of 0.020
mg/m3 that observed an association
between exposure and increased
infections. One acute mouse studv also
provided indirect support for an
increased likelihood of respiratory
infections.
Moderate
supports an
increased
propensity for
LRT infections,
particularly
during
development
Animal: Decreased antibacterial activity in mice (Jakab, 1992): acute at 1.23 mg/m3,
noting that this finding appeared to be particularly sensitive to the pattern of
formaldehyde exposure
This document is a draft for review purposes only and does not constitute Agency policy.
A-546 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-78. Summary of changes in the lower respiratory tract (LRT) as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence (exposure
duration)
Conclusion
Low
Human: Increased emergency room visits for episodes including LRT infections (Rumchev
et al., 2002): children aged 6-36 months with mean levels of 0.028-0.030 (maximum
0.12-0.22) mg/m3
Direct and indirect evidence of impaired
LRT immune function in children and in
a short-term rat study, respectively
Animal: Decreased expression of immune-related genes in rat lung (Sul et al., 2007),
specifically HSP701a (may be involved in antigen presentation), complement 4 binding
protein (may bind necrotic or apoptotic cells for cleanup), and Fc portion of IgGiii (may be
involved in leukocyte activation): 2 wk at >6.15 mg/m3
Changes in
pulmonary
function with
challenge (e.g.,
with broncho-
constrictors
and/or
allergens)
(Note:
unprovoked
responses are
not included)
High or Medium
Human: None
Acute and short-term studies in two
animal species demonstrate that
formaldehyde increases responsiveness
to allergens and bronchoconstrictors,
particularly with prior sensitization, at
levels as low as 0.31 mg/m3
Robust -t
Hyperresponsive
airways3
(1" effects with
allergen)
Animal: [allergen challenge]: With ovalbumin [OVA] sensitization, increased airway
obstruction in guinea pigs (Riedel, 1996): short-term at 0.31 mg/m3 and increased
reactivity in mice (Larsen, 2013): acute at =5-7 mg/m3 in humid or dry environments;
[acetylcholine challenge]: Increased airway resistance and reactivity in guinea pigs
(Swiecichowski, 1993; Leikauf, 1992): acute at 1.23 mg/m3
Low
Human: [histamine challenge]: Hyperreactive airways with prolonged exposure (Gorski,
1991): >1 year at <0.5 mg/m3, but N/C after acute exposure (Krakowiak, 1998): 2 hr at 0.5
mg/m3; [allergen challenge]: hypersensitivity with acute exposure when exposure was
restricted to mouth breathing in allergic asthmatics with a large allergen (mite) (Casset;
2006): <1 hr at 0.1 mg/m3, but N/C after acute oronasal (normal) exposure in allergic
asthmatics using a different allergen (pollen), including a test of methacholine (MCh)
responsiveness 8 hr after allergen exposure (Ezratty, 2007): 1 hr at 0.5 mg/m3
Suaaestive evidence of increases with
prolonged exposure, and possiblv acute
mouth-breathing exposure when
challenged with specific allergens, but
not acute exposure alone, to <0.5
mg/m3 in human adults; also, increased
at >3 mg/m3 in short-term or acute
studies across three species, particularly
with prior sensitization
Animal: [MCh challenge]: Hyperresponsive airways (increased reactivity and sensitivity)
noted with FA alone in mice and rats (Qiao, 2009; Lui, 2011; Wu, 2013): short-term at >3
mg/m3, and in monkeys (Biagini, 1989): acute at 3.1 mg/m3; in mice and rats, this response
was amplified with OVA sensitization; Note: TRP antagonists reduced the
hyperresponsiveness in mice (Wu et al., 2013)
aAs the challenge stimuli used in the formaldehyde studies included allergens as well as nonimmunological stimuli, and because most experiments did not
attempt to delineate the specifics of the functional changes, "airway hyperresponsiveness" or "hyperresponsive airways" encompasses any of a range of
This document is a draft for review purposes only and does not constitute Agency policy.
A-547 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
possible airway features: hyperreactivity (exaggerated response), hypersensitivity (lower dose to elicit response), altered ventilatory parameters (e.g.,
maximal response; resistance), recovery (longevity of response), or others.
This document is a draft for review purposes only and does not constitute Agency policy.
A-548 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-79. Summary of changes in LRT cell counts and immune factors as a result of formaldehyde exposure
Endpoint(s)
No changes observed
{high or medium confidence experiments are
bolded)
Significant3 increases or decreases
{high or medium confidence experiments are
bolded)
Summary
conclusion
Clarifying notes
and exposure
duration
Duration Concentration(s) [allergen stimulusl
(species) (studv)
Duration Concentration(s) [allergen
(species) stimulusl (studv)
White blood cells (WBCs)
Total WBCs (or Total
Inflammatory Cells)
Acute (g pigs) 0.13-0.31 mg/m3 [-OVA] (Riedel, 1996)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (mice) 0.5-6.2 mg/m3 [-OVA] (Larsen, 2013)
Acute (mice) 0.25-3.7 mg/m3 [-OVA] (de Abreu, 2016)
Subchronic (mice) ¦f- 2.5 mg/m3 [+ OVA] (Fujimaki, 2004)
Short term (mice) 'T* 12.3 mg/m3 [-OVA] (Kim, 2013b); total
Short term (mice) BAL cells
Short term (mice) 'T* 12.3 mg/m3 [-OVA] (Jung, 2007)
Short term (rats) -f 3 mg/m3 [± OVA] (Wu, 2013)
^ 0.5-3.1 mg/m3 [+ OVA] (Qiao, 2009)
Moderate ^
short-term >0.5
mg/m3; amplifies
allergen effect
Granulocytes
Neutrophils
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)
Acute (g pigs) 4.2 mg/m3 [-OVA] (Swiecichowski, 1993)
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Jung, 2007)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) 'T* 3 mg/m3 [+ OVA] (Wu, 2013)
Acute (rats) 'T* 6.2 mg/m3 [-OVA] (Kimura, 2010)
Slight
amplifies allergen
response at >3 mg/m3
(short-term)
Eosinophils
Acute (humans) (trend ^) 0.1 mg/m3[+ Derb f] (Casset, 2007)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (rats) 6.2 mg/m3 [-OVA] (Kimura, 2010)
Subchronic (mice) ¦f- 2.5 mg/m3 [+ OVA] (Fujimaki, 2004)
Short term (mice) 'T* 12.3 mg/m3 [-OVA] (Jung, 2007)
Short term (mice) ^ 0.5-3 mg/m3 [± OVA] (Liu, 2011)
Short term (mice) 'T* 3 mg/m3 [± OVA] (Wu, 2013)
Short term (mice) ^ infer1 >12.3 mg/m3 [+ Der f] (Sadakane,
Short term (rats) 2002)
^ 0.5-3.1 mg/m3 [+ OVA] (Qiao, 2009)
Moderate ^
short-term >0.5
mg/m3; amplifies
allergen effect
Lymphocytes
All
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Kim, 2013b)
Short term (mice) 12.3 mg/m3 [-OVA] (Jung, 2007)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) 'T* 3 [-OVA] mg/m3 (Wu, 2013)
Indeterminate
suggests total number
unchanged
B Cells
Acute (g pigs) 4.2 mg/m3 [-OVA] (Swiecichowski, 1993)
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Kim, 2013b)
Short term (mice (trend \|/) 6.2-12.3 mg/m3 [-OVA] (Jung,
2007)
Indeterminate
allergen stimulus
unstudied
T Cells (CD4+)
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Kim, 2013b)
Short term (mice) (trend ^) 6.2-12.3 mg/m3 [-OVA] (Jung,
2007)
Short term (rats) 'T* (adults) 7.4 mg/m3 [-OVA] (Sandikci,
2007b)
Indeterminate
allergen stimulus
unstudied
This document is a draft for review purposes only and does not constitute Agency policy.
A-549 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
No changes observed
{high or medium confidence experiments are
bolded)
Significant3 increases or decreases
{high or medium confidence experiments are
bolded)
Summary
conclusion
Clarifying notes
and exposure
duration
Endpoint(s)
Duration Concentration(s) [allergen stimulusl
(species) (studv)
Duration Concentration(s) [allergen
(species) stimulusl (studv)
T Cells (CD8+)
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Kim, 2013b)
Short term (rats) 'T* (adults) 7.4 mg/m3 [-OVA] (Sandikci,
Short term (mice) 2007b)
'T* (slight) 12.3 mg/m3 [-OVA] (Jung, 2007)
Slight^
short-term >7 ms/m3.
allergen stimulus
unstudied
NK Cells
Short term (mice) \|/ 12.3 mg/m3 [-OVA] (Kim, 2013b)
Indeterminate
Monocytes
Acute (g pigs) 4.2 mg/m3 [-OVA] (Swiecichowski, 1993)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (rats) 6.2 mg/m3 [-OVA] (Kimura, 2010)
Subchronic (mice) ¦f- 2.5 mg/m3 [+ OVA] (Fujimaki, 2004)
Slight^
long-term >2.5 mg/m3
amplifies allergen
effect
Mast Cells
Acute (g pigs) 4.2 mg/m3 [-OVA] (Swiecichowski, 1993)
Indeterminate
Secreted factors and immune markers
Primarily Thl-related
TNF-a and GM-CSF
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (mice) 0.25-3.7 mg/m3 [-OVA] (de Abreu, 2016)
Indeterminate
suggests unchanged or
highly variable
IFN-v
Short term (mice) 0.5-3 mg/m3 [± OVA] (Liu, 2011)
Short term (mice) 3 mg/m3 [± OVA] (Wu, 2013)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) \|/ 6.2-12.3 mg/m3 [-OVA] (Kim, 2013b)
Short term (rats) 'T* 3.1 mg/m3 [-OVA] (Qiao, 2009)
IL-1
(IL-ip in animals)
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (mice) 0.25-3.7 mg/m3 [-OVA] (de Abreu, 2016)
Subchronic (mice) 4- 2.5 mg/m3 [+ OVA] (Fujimaki, 2004)
Short term (mice) 'T* 3 mg/m3 [-OVA] (Wu, 2013)
Short term (mice) 'T* 6.2-12.3 mg/m3 [-OVA] (Jung, 2007)
Primarily Th2-related
IL-4
Short term (mice) infer3 >12.3 mg/m3 [± Der f] (Sadakane,
Acute (humans) 2002)
0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) 'T* 1-3 mg/m3 [-OVA] (Lu, 2005)
Short term (mice) 'T* 6.2-12.3 mg/m3 [-OVA] (Jung, 2007)
Short term (mice) 'T* 0.5-3 [+ OVA] or 3 [-OVA] mg/m3 (Liu,
Short term (mice) 2011)
Short term (rats) -f 3 mg/m3 [± OVA] (Wu, 2013)
'f 0.5-3.1 mg/m3 [+ OVA]; \|/ 3.1 mg/m3
[-OVA] (Qiao, 2009)
Slight^
IL-4 at >0.5 mg/m3 and
IL-5 at >6.15 mg/m3,
short-term and likely
amplifying allergen
effects
IL-5
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) 'T* 6.2-12.3 mg/m3 [-OVA] (Jung, 2007)
Short term (mice) ^ infer3 >12.3 mg/m3 [+ Der f] (Sadakane,
2002)
IL-10
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Indeterminate
IL-6
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)
Acute (mice) 0.25-3.7 mg/m3 [-OVA] (de Abreu, 2016)
Short term (mice) 'T* 0.5-3 [+ OVA] or 3 [-OVA] mg/m3 (Liu,
2011)
This document is a draft for review purposes only and does not constitute Agency policy.
A-550 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
No changes observed
{high or medium confidence experiments are
bolded)
Significant3 increases or decreases
{high or medium confidence experiments are
bolded)
Summary
conclusion
Clarifying notes
and exposure
duration
Endpoint(s)
Duration Concentration(s) [allergen stimulusl
(species) (studv)
Duration Concentration(s) [allergen
(species) stimulusl (studv)
IL-13
Short term (mice) 6.2-12.3 mg/m3 [-OVA] (Jung, 2007)
NKcell
factors
IL-2R
Short term (mice) 6.2-12.3 mg/m3 (Kim, 2013b)
Indeterminate
Perforin
Eosinophil attractant and
adhesion factors
RANTES
Short term (mice) ^ infer3 >12.3 mg/m3 [± Der f] (Sadakane,
2002)
Slight^
chemoattractants
relevant to eosinophil
recruitment with
ICAM and CCR3
Short term (mice) ^ (indirect15) 12.3 mg/m3 [-OVA] (Jung,
2007)
Eotaxin
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)3
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Short term (mice) ^ (indirect15) 12.3 mg/m3 [-OVA] (Jung,
2007)
ECP
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (humans) ^ 0.1 mg/m3 [+ Der f] (Casset, 2007)
MlP-la
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)3
Other
IL-8
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Acute (in vitro) ^ 1.23 mg/m3 (Rager, 2011)
Indeterminate
MCP-1
Subchronic (mice) 0.1-2.5 mg/m3 [± OVA] (Fujimaki, 2004)3
Acute (humans) 0.5 mg/m3 [+ pollen] (Ezratty, 2007)
Indeterminate
Derf: Dermatophagoides farina (house dust mite); OVA: ovalbumin (major protein of chicken egg whites); both are immunogenic materials used to stimulate
an allergic response.
Gray box = no data meeting the inclusion criteria were available.
Notes: Two studies with evidence that may inform the potential for formaldehyde exposure-induced inflammatory changes in the LRT are not captured in
these tables, specifically a proteomics analysis of the BAL fluid after short-term exposure at >2.46 mg/m3 (Ahn et al., 2010) and an miRNA microarray study of
gaseous paraformaldehyde exposure in a human lung cancer cell line with acute exposure to 1.23 mg/m3 (Rager et al., 2011). Swiecichowski, 1993 may
include information from an earlier study interpreted to have been conducted in the same cohort of guinea pigs (Leikauf et al., 1992).
Primarily, this reflects reporting of a statistically significant change; in rare instances where a p value was not given, changes are indicated if the authors
discussed the change as a significant effect.
bReported as 0.5% formaldehyde solution; concentration assumed to be >12.3 mg/m3 (Sadakane, 2002).
cGene expression levels.
dThese factors were not present at detectable levels regardless of treatment.
This document is a draft for review purposes only and does not constitute Agency policy.
A-551 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
Changes in the blood and lymphoid organs
Although this mechanistic evaluation is focused on mechanisms underlying respiratory
health effects, these effects can be influenced by changes in nonrespiratory tissue compartments,
most notably the blood and lymphoid organs. The direction, magnitude and type of immune
responses observed in the blood should not be assumed to represent immunological changes
occurring in the airways, as responses can differ. The nonrespiratory changes most likely relevant
to respiratory system health effects are immune-related changes because these could induce
extrapulmonary signals (e.g., cellular; secreted factors) to travel through the blood to perfused
regions of the respiratory tract. This section emphasizes changes in exposed humans, unlike the
emphasis on experimental animal studies in the URT and LRT sections, because blood sampling in
humans is more convenient than sampling from respiratory tissue compartments; thus, more
human data are available for changes in the blood.
A number of studies, across different human and animal populations, spanning an array of
formaldehyde exposure scenarios, have reported changes in blood cell counts. Although some of
the specific changes vary across studies, taken together, the data provide robust evidence of an
association between formaldehyde exposure and hematological effects. Although additional studies
clarifying inconsistencies across the studies would be informative, several tentative patterns could
be discerned. Interestingly, looking atthe picture as a whole (see Figures A-33-A-34), the direction
of some changes noted in the blood of individuals exposed to formaldehyde are contrary to the
cellular changes noted in the respiratory tract. For example, data suggest (slight) or support
[moderate) that total cells, neutrophils, and CD8+ T cells are increased in the respiratory tract by
formaldehyde exposure, while these same cells appear to be decreased in the blood
(see Figure A-34). One potential explanation for this difference could involve recruitment of
particular subsets of immuno-responsive cells from the circulation to the irritated and inflamed
respiratory tract, as is observed with viral infections of the respiratory system [Levandowski,
1986]; however, none of the identified human studies report data from both tissue compartments,
and the animal data do not address such a hypothesis. It is plausible that this pattern could reflect
species differences (i.e., LRT data are mostly from animal studies), but this possibility is considered
unlikely given the blood data. As with investigations of the airways, very few studies tested
mechanistic hypotheses for how formaldehyde exposure could affect blood immune cell counts.
Despite this lack of information and variability in responses, the available data support a conclusion
that formaldehyde exposure can modify immune system function in the blood across a range of
concentrations and exposure durations.
One of the most consistent cellular changes observed across studies was a decrease in the
total number of white blood cells (WBCs). This is a nonspecific finding, as WBCs encompass a
spectrum of functional phenotypes, and this change may be driven by decreases in only one or
several subpopulations. When looking more specifically atthe WBCs, moderate evidence of CD8+ T
cell decreases following formaldehyde exposure is provided by several studies, together with a
This document is a draft for review purposes only and does not constitute Agency policy.
A-552 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde—Inhalation
corresponding increase in the ratio of CD4+/CD8+ T cells (see Table 1-80). As mentioned
previously, CD8+ T cells comprise a heterogeneous cell population, which complicates
interpretations regarding the potential impact of decreased numbers in peripheral blood.
Depending on the specific stimuli, stimulated CD8+ T cells can produce interferon-y (IFN-y) and
inhibit production of IL-4 and immunoglobulin (i.e., IgE) responses [Holmes etal., 1997], or their
phenotype can be driven towards production of excess IL-4, a situation hypothesized to be
associated with atopic asthma (Lourenco et al., 2016: 7:638). Moderate evidence provides support
for increases in blood IL-4 (which was similarly increased in the LRT) and decreases in IFN-y after
formaldehyde exposure. A more complete understanding of the phenotype of the depleted CD8+ T
cells would be informative to ascertain whether these changes are related to the profile of secreted
factors observed in the blood after formaldehyde exposure (see Figure A-33).19
Moderate evidence also indicates that formaldehyde exposure alters the number or
percentage of B cells in the circulation. These cells produce antibodies upon stimulation with
antigen (e.g., allergens) and contribute to airway hyperresponsiveness [Hamelman et al., 1997],
While this finding, along with slight evidence of increased antigenic markers, suggests potential for
alteration of the adaptive immune response as a result of formaldehyde exposure, this observation
alone is insufficient to indicate functional changes such as exposure-induced differences in clonal
expansion and differentiation to antibody-producing cells, evidence of which would support a more
convincing biological relationship.
Slight evidence suggests that neutrophils are also decreased in the blood by formaldehyde
exposure. This could plausibly be explained by the suggestive (slight) findings of decreased
lymphocyte and neutrophil chemoattractants in the blood and increased levels in the airways
(possibly attracting blood neutrophils), suggesting that a gradient of these factors across tissue
compartments may be induced and maintained as a result of formaldehyde exposure and, perhaps,
sustained inflammation.
Finally, although variable across studies, several lines of evidence suggest a pattern of
immune cell effects related to formaldehyde concentration, with stimulation at lower formaldehyde
exposure levels and decreases at higher levels. This included changes in total T cells, NK cells, and
IL-10 (and, perhaps, TNF-a). A complex relationship exists between IL-10, NK cells, and subsets of
CD4+ T cells (e.g., Thl and Th2 cells), which direct the type of antibody responses; however, the
specifics of this suggestive (slight) association with formaldehyde exposure remain to be
elucidated. Many of these observations would benefit from additional, more specific studies
on WBCs.
19 Several studies examining the lineage and maturity of immune and non-immune cells in the bone marrow and other systemic
tissues (e.g., blood; spleen) are not discussed in this section. Although it is possible that differences in the maturation
phenotype of cells could indirectly contribute to the immune changes of interest to this section, such alterations would be
expected to cause functional or other detectable changes in more apical mechanistic events relevant to immune responses in
the respiratory system. Thus, this discussion focuses on those mechanistic events considered more directly relevant to these
POE outcomes. Please see Section 1.3.3 of the Toxicological Review for a discussion of these cell lineage and maturation
markers in the context of lymphohematopoietic cancer MOA.
This document is a draft for review purposes only and does not constitute Agency policy.
A-553 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Toxicological Review of Formaldehyde—Inhalation
Red blood cell (RBC) counts were decreased in both human and animal studies (moderate
evidence), generally at formaldehyde concentrations above 0.5 mg/m3. Slight data exist to suggest
that platelets may also be decreased, which could plausibly be related to the single, low confidence
animal study that reported increased megakaryocytes (cells that produce platelets) in the bone
marrow [Zhang et al., 2013], The relevance of these changes to respiratory system health effects is
unknown. It is plausible that sustained increases in oxidative stress (which has been observed in
the blood and, to a lesser extent, other lymphoid tissues) and/or other soluble factors in the blood
resulting from airway inflammation could affect the viability of circulating erythrocytes and
immune cells or the circulating precursors for these cells; however, no evidence exists to
substantiate this hypothesis. An increased level of the circulating stress hormone, corticosterone
(the major animal glucocorticoid; in humans, it is Cortisol), with short-term, but not acute,
formaldehyde exposure is also suggested by slight data. Persistent increases in circulating
glucocorticoids can also negatively impact the function and health of circulating immune cells,
causing immunosuppression of most cell types [O'Connor et al., 2000], However, these potential
linkages have not been examined.
As with findings for WBC changes, antibody, or immunoglobulin (Ig), responses resulting
from formaldehyde exposure are consistently altered, although the specific changes observed
across studies provide a mixed picture. Much of the moderate evidence is based on animal
sensitization models using the protein allergen ovalbumin, although the human data also indicate
changes after exposure. In general, the variable evidence of formaldehyde-induced modification of
humoral immunity in humans demonstrates different patterns of results depending on the
population (e.g., children vs. adults), the duration of exposure, and the specific Ig measure (e.g., Ig
isotype) across studies. The animal studies consistently report amplified responses with allergen
stimulation and/or sensitization, although the pattern and magnitude of these effects appears to
vary depending on the type of allergen and the sensitization protocol used. The Igs most relevant to
the blood and respiratory tract are IgA (IgAl and IgA2), IgE, IgM, and IgG (IgGl, IgG2, and IgG3;
also, IgG4 in humans). No changes of note in IgA or IgM were identified across the available studies.
Slight data suggest that formaldehyde exposure may cause elevated levels of IgE antibodies in
certain exposure scenarios, including in exposed children; however, this finding should be
interpreted with caution, as comparable studies did not observe effects, and explanations for this
inconsistency are not available. IgEs are implicated in allergic hypersensitivity responses of the
airways [Hamelman et al., 1999], although they may not be essential for all hypersensitivity-related
responses (e.g., intrinsic [nonallergic] asthma occurs in one-third of all adult patients; Knudsen et
al., 2009). Despite the variability in models, several of the available studies consistently identified
changes in antibodies of the IgG class (moderate evidence), including increases in IgGs specific to
formaldehyde or antigens (e.g., allergens) to which the subjects had previously been exposed. IgGs
are the most prevalent Ig in the serum of humans, and they are the only Ig that can be transferred to
neonatal/infant circulation (i.e., by crossing the placenta; through breast milk in animals) to
This document is a draft for review purposes only and does not constitute Agency policy.
A-554 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde—Inhalation
influence immunity in offspring [Van de Perre, 2003], None of the included studies examined
antibody titers or transferred immunity with developmental formaldehyde exposure (note: not
informative studies from one lab: Maiellero et al., 2014; Ibrahim et al., 2015, reported immune-
related effects of gestational formaldehyde exposure). While IgEs are most commonly associated
with sensitization-related airway hyperresponsiveness to allergens, subclasses of IgGs also
contribute to allergic responses; however, their exact role in the pathophysiology of airway
disorders remains unclear [Hofmaier etal., 2014; Williams etal., 2012; Bogaert etal., 2009],
Overall, although a body of evidence indicates changes in antibody-mediated responses after
formaldehyde exposure, particularly in regard to IgGs, an explanation for the variable pattern of
changes in Igs (e.g., to formaldehyde alone or with coexposure to different types of antigens by
specific Ig subclasses) does not exist, and the likely consequences of these changes are unknown.
This document is a draft for review purposes only and does not constitute Agency policy.
A-555 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
Formaldehyde-Induced Antibody Response in the Blood
- E
Human: None
No changes in a subchronic mouse
study at <2.46 mg/m3
Moderate
.3
!§> "S
1 5
Animal: No evidence suggesting changes (Fujimaki, 2004): subchronic <2.46 mg/m3
Altered antibody
responses (basis
below)
Total IgE
Low
Human: No evidence suggesting changes (Wantke, 1996b, 2000; Erdei, 2003; Ohmichi, 2006;
(Palczvnski et al., 1999)): short-term <1.8 mg/m3 (duration in Erdei unknown)
Suggestive evidence of increased IgE
in 2 short-term formalin studies in
Total
Moderates1/:
Animal: Evidence of increases in mice, which were increased further by OVA (Wu, 2013; Jung, 2007):
short-term >3 mg/m3; evidence of no changes in mice by FA alone (Kim, 2013; Gu, 2008), although FA
exacerbated HDM-induced IgE (Kim, 2013): short-term 0.12-1.2 mg/m3
mice at >3 mg/m3, but no evidence
for changes in mice or humans at <2
mg/m3
IgG [naive
subjects]
Slight 1s: IgE [3
mg/m3]
High or
Medium
Human: Elevated in one study of children (Wantke, 1996a): years (assumed) at =0.06 compared to
=0.03 mg/m3 (unrelated to symptoms);
N/C in adults (Kim, 1999): 4 years at 3.74 mg/m3
Increased in a single long-term study
of children at <0.1 mg/m3; N/C in a
single long-term study of adults at
IgA [6 mg/m3]
Indeterminate:
IgM [mixed]
Animal: None
3.74 mg/m3
FA-specific
Moderate IgG
[long-term] Slight
IgE [children;
long-term]
Indeterminate:
IgM or IgA
Formaldehyde
(FA)-Specific
Human: No evidence of changes across multiple studies in adults (Wantke, 1996b; Zhou, 2005;
Ohmichi, 2006; Thrasher, 1987; Kim, 1999; Gorski, 1991): short-term (weeks) or long-term (years) at
=0.1-3.74 mg/m3; unclear in 2 long-term adult studies in which a small proportion of subjects did have
FA-lgE (Dykewicz, 1991 and Thrasher, 1990);
one study noted slight increases with longer exposure (Wantke, 2000): 10 wk, not 5 wk, at 0.265 mg/m3
IgE
Low
Animal: Isotype unspecified- no change in guinea pigs with acute challenge (Lee et al., 1984) at 2.5 or
4.9 mg/m3 after short term exposure to 7.4 or 12.3 mg/m3 (note: no measures without formaldehyde)
No clear evidence of changes across
multiple short-term and long-term
studies in adults at <3.74 mg/m3; no
studies in children
Antigen-specific
Moderate IgG
[inhaled antigen]
Slight IgE
[certain
scenarios]
Indeterminate:
IgM or IgA
This document is a draft for review purposes only and does not constitute Agency policy.
A-556 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
Antigen-
Specific IgE
(does not include
FA-specific Ig)
High or
Medium
Human: None
N/C in a single subchronic study with
i.p. sensitization
Animal: N/C in OVA-lgE (Fujimaki, 2004): 12 wks at 0.1-2.46 mg/m3 (OVA i.p.)
5
o
1
Human: None
Two mouse studies suggest
formaldehyde can increase IgE
specific to antigen at =>1 mg/m3, but
this appears to be highly situational
(e.g.. dependent on duration and
periodicity of formaldehyde
exposure, and antigen type and
administration route)
Animal: Increased OVA-specific IgE in mice in 2 studies—(Tarkowski, 1995 and Gu, 2008): 10 d at 2
mg/m3 (but not 1 d/wk for 7 wk, or when OVA sensitization i.p.) and 5 wk at 0.98 mg/m3 with i.p. OVA
(but not <4 wk), respectively; however, N/C in mice in 3 studies: (Wu, 2013): 4 wk at 3 mg/m3 (s.c. OVA
sensitization), (Kim, 2013): 0.2-1.23 mg/m3 for 4 wk (dermal house dust mite, HDM, sensitization), and
(Sadakane, 2002): 4 wk at 0.5% (i.p. Der f sensitization)
Total IgG
High or
Medium
Human: Decreased in a single study of exposed workers (Avdin et al., 2013): 7 vr at 0.264 mg/m3
A single study in adult workers and
another in male rats showed
decreased IgG at 0.264 or >6.15
mg/m3 with long-term or short-term
exposure, but subclass not examined
Animal: Decreased total IgG in rats (Sapmaz, 2015): short-term at >6.15 mg/m3
5
o
1
Human: N/C in children at =0.007-0.07 mg/m3 (Erdei, 2003): unknown duration (likely months-years)
Suggestive evidence based on
increased IgGl in 2 short-term mouse
studies, but a third mouse study and
a human study did not observe
effects at <1 mg/m3
Animal: IgGl (N/C in lgG2a) increased by FA alone, whereas FA exacerbated lgG2a (N/C in IgGl) in
atopic-prone mice (Kim, 2013): short-term 0.25, not 1.2 mg/m3; increased IgGl and lgG3, but
decreased lgG2a and 2b, in C57 mice (Jung, 2007): short-term >6.15 mg/m3;
N/C in IgG Balb/c mice (Gu, 2008): short-term <1 mg/m3
FA-Specific IgG
High or
Medium
Human: Slight (<10%) increase in a single study of adults (Kim, 1999): years at 3.74 mg/m3
Slightly increased in a single
long-term study of adults at 3.74
mg/m3; no studies in children
Animal: None
5
o
1
Human: Increased in two studies (Thrasher, 1987; 1990) and unclear in 1 study in which 5/55 subjects
did have FA-IgG (Dykewicz, 1991): [all 3 studies] years at <0.1-<1.0 mg/m3;
N/C in one study (Wantke, 2000): short-term at 0.265 mg/m3
Suggestive of slight increases in
adults with long-term exposure at <1
mg/m3, but not with short-term
exposure at higher levels; no studies
in children
Animal: Isotype unspecified—no change in guinea pigs with acute challenge (Lee et al., 1984) at 2.5 or
4.9 mg/m3 after short term exposure to 7.4 or 12.3 mg/m3 (note: no measures without formaldehyde)
This document is a draft for review purposes only and does not constitute Agency policy.
A-557 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
Human: None
Increased OVA-lgGl in 1 short-term
Antigen-
High or
Mediurr
Animal: Increased OVA-specific IgGl in guinea pigs (Riedel, 1996): 5 d at 0.31 mg/m3 (inhaled OVA);
questionable decrease (no dose-response) in OVA-lgGl and OVA-lgG3 in mice (Fujimaki, 2004): 12 wks
at 0.49, but not 2.46 mg/m3 (OVA i.p.; N/C in OVA-lgG2)
study in guinea pigs at 0.31 mg/m3
with inhaled allergen, but not a
longer mouse study using injected
allergen
Specific IgG
(does not include
FA-specific Ig)
Human: Increased IgG against 2 bacterial pathogens by linear regression in 3rd grade children with
respiratory complaints (Erdei, 2003): <0.1 mg/m3, unknown duration (likely years, home measures)
1 long-term study suggests increased
IgG sensitization to an airway antigen
5
o
1
Animal: N/C in OVA-IgG or Der f-lgGl in mice (Gu, 2008; Wu, 2013; Sadakane, 2002): up to 5 wk at
0.123-3 mg/m3 or higher; N/C in IgG specific to vaccine antigens in rats (Holmstrom, 1989): 22 months
at 15.5 mg/m3. In all cases, s.c. or i.p. sensitization
by FA in children; multiple studies in
mice and rats suggest that IgG
sensitization does not occur when
antigen sensitization occurs by
injection
° 1
4—'
-£= ~0
bO aj
Human: Decreased IgM, N/C in IgA, in a study of exposed workers (Avdin et al., 2013): 7 yr at
0.26 mg/m3
IgM, but not IgA, decreased in a
single study in adult workers at 0.26
Total IgM or
IgA
X 5
Animal: Increased total IgM and IgA in rats (Sapmaz, 2015): short-term at >6.15 mg/m3
mg/m3 with long-term exposure
3
Human: No evidence of IgA or IgM changes (Erdei, 2003): duration unknown <0.1 mg/m3
IgA increased in 1 short-term study at
o
i
Animal: Increased IgA and N/C in IgM in C57 mice (Jung, 2007): short-term >6.15 mg/m3
>6 mg/m3; N/C in IgM in 2 studies
o 1
Human: None
!§> "S
X 2
Animal: None
No evidence to evaluate
FA-Specific
IgM or IgA
3
Human: Unclear evidence in 1 long-term study in which a small proportion of subjects appear to have
elevated FA-specific IgM (Thrasher, 1990): months-years at =0.1-1 mg/m3
Evidence could not be interpreted
i
Animal: Isotype unspecified- no change in guinea pigs with acute challenge (Lee et al., 1984) at 2.5 or
4.9 mg/m3 after short term exposure to 7.4 or 12.3 mg/m3 (note: no measures without formaldehyde)
Antigen-
o |
Human: None
Specific IgM or
IgA
!§> "S
1 5
Animal: None
No evidence to evaluate
This document is a draft for review purposes only and does not constitute Agency policy.
A-558 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
(does not include
FA-specific Ig)
s
Human: N/C in airway pathogen bacteria-specific IgM or IgA in one study in children (Erdei, 2003):
unknown duration (likely months to years) at <0.1 mg/m3
The minimal data available suggest
that formaldehyde does not alter
these parameters
O
i
Animal: N/C in IgM specific to vaccine antigens in rats (Holmstrom, 1989): 22 months at 15.5 mg/m3
(s.c. injection)
Immune and Inflammation-Related Changes in the Blood
[[See Appendix Table 1-32 for Cellular and Cytokine Response in Blood]]
igh or Medium
Human: Increased marker of lipid peroxidation in adult serum lymphocytes (Bono, 2010): likely months
to years (assumed) at >0.066 mg/m3; Increased F2-lsoprostanes (suggested as the best in vivo
biomarker of lipid peroxidation) in urine (Romanazzi, 2013): 0.21 mg/m3 chronic occupational
(indirect), although smoking and formaldehyde were not additive, both were independently associated
with ROS—Note: serum and urine IsoP measures are correlated (Rodrigo et al., 2007), suggesting that
urine levels may reflect similar serum changes
Two studies in adults indicate
elevated oxidative stress markers in
blood at >0.066 mg/m3 with long-
term exposure. Given the
uncertainty with concluding urine
levels exhibit the same pattern of
Oxidative
Stress
X
Animal: None
association as blood, one study
contributes as indirect evidence
Human: Increased oxidative stress biomarkers (F2-lsoprostanes; malondialdehyde) in urine (Bellisario,
2016): =0.034 mg/m3 work shift occupational (indirect; responses likely reflect short-term exposure)
Several studies in three species
Moderate 'f
Low
Animal: Increased oxidative stress markers in mice (Ye, 2013; (Matsuoka et al., 2010): acute or
short-term as low as 0.12 mg/m3; increased markers and protein indicators in rats (Im, 2006; Aydin,
2015): short term at 6.48-12.3 mg/m3, although 1 study with longer exposure observed a decrease in
MDA, but decreased SDH in lymphocytes (Katsnelson, 2013): 10 wk at 12.8 mg/m3; other indicators
including decreased GSH (Ye, 2013; Katsnelson, 2013) and increased NO and SOD (Matsuoka et
al., 2010) at >1 mg/m3
suggest increases in markers of
oxidative stress with acute or short-
term exposure, even at
formaldehyde levels <1 mg/m3; it is
not clear whether and to what extent
this persists with long-term exposure
Human: None
Increased stress hormone at 3 mg/m3
formaldehyde in a single rodent
study with short-term, but not acute,
exposure
Circulating
Stress
Hormones
Animal: Increased corticosterone in rats with short-term, but not acute, exposure (Sorg, 2001): =3
mg/m3
Slight 'T*
5 3
Human: None
No evidence to evaluate
This document is a draft for review purposes only and does not constitute Agency policy.
A-559 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
Animal: None
Altered
Immune
Function
High or
Medium
Human: None
No evidence to evaluate
Indeterminate
Animal: None
5
o
1
Human: Increased autoantibodies in adults (Thrasher, 1990): long-term at 0.06-0.95 mg/m3
1 study in adults suggests that
autoantibodies are elevated with low
level, long-term exposure: somewhat
in contrast, 1 mouse study suggests
short-term high level exposure
improves host response to bacteria
Animal: Improved cell-mediated immune response to bacteria challenge, but N/C against tumor
challenge or delayed-type hypersensitivity response in mice (Dean, 1984): 3 wk at 18.5 mg/m3;
however, N/C in vitro measures of immune cell function.
Changes in Other Immune-related tissues
Cell counts in
immune
tissues (not
including bone
marrow)
High or Medium
Human: None
Suppression of CD8+ T cells in
immune tissues (e.g., spleen) is
indicated in one 8-wk mouse study,
with indirect support from a second
short-term mouse study, at around 2
mg/m3; effects on CD4+/CD8+ ratio
were mixed across 2 subchronic
mouse studies
Moderate
(for si CD8+ T cell
response in
spleen and
thymus)
Slight
NK cells (in
spleen: 'T* at low
level; \|/ at high
level)
Indeterminate for
other cell counts
Animal: Decreased CD8+ T cells and increased CD4+/CD8+ ratio in both thymus (immature immune
cells) and spleen (mature immune cells) in male mice (Ma et al., 2020): Eight weeks of exposure
at 2 mg/m3; No change in splenic CD4+/CD8+ ratio in female mice (Fujimaki et al., 2004b): 12 wk
at up to 2.46 mg/m3; Increased splenic regulatory T cells (subset of CD4+) and indirect markers for
suppression of effector T cell (CD8+) activity in female mice (Park et al., 2020): short-term
exposure at >1.38 mg/m3
5
o
1
Human: None
Multiple short-term mouse studies
suggest that overall splenic cell T and
B cells are unchanged; however, 2
studies suggest that NK cells may be
affected (1 study showed NK cells
were stimulated at low formaldehyde
levels, and another that high levels
are inhibitory/toxic)
Animal: N/C in tissue weight, total cellularity or T or B cell counts in mice (Kim, 2013b, Gu, 2008; Dean,
1984); altered NK cell number and function was noted in mice, with one study showing decreases (Kim,
2013b): 2-3 wk at 12.3 mg/m3, and another showing increases (Gu, 2008): 5 wk at up to 0.12 mg/m3,
and a third showing N/C in lymphocyte proliferation, functional parameters, IgM production, or NK
cytotoxicity (Dean, 1984): 3 wk at 18.5 mg/m3
Splenic and
Lymph
High or
Medium
Human: None
No evidence to evaluate
Slight oxidative
stress and
cytokine
Animal: None
This document is a draft for review purposes only and does not constitute Agency policy.
A-560 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-80. Summary of changes in the blood and lymphoid organs as a result of formaldehyde exposure (continued)
Endpoint
Study-specific findings from "high or medium" or "low" confidence experiments
Summary of evidence
(exposure duration)
Conclusion
Cytokines and
other Markers
Low
Human: None
1 short-term mouse study suggests
increased oxidative stress at >1
mg/m3, and another \|/ IL-13 at
0.25-1.23 mg/m3, and 3 others
suggest that the response (splenic or
lymph) to antigen stimulation (and 1
study without stimulation), most
notably increased IL-4, is exacerbated
at >0.25 mg/m3 formaldehyde
production,
especially in
response to
antigen
Animal: Spleen: 'T* oxidative stress markers in mice (Ye, 2013): 7 d at >1 mg/m3); exaggerated IFNy
response (at 2.46 mg/m3) of lymphocytes to LPS and 'T* MCP-1 response to OVA in mice (Fujimaki,
2004): 12 wk at >0.49 mg/m3; \|/ IL-13 (Kim, 2013a): short-term at 0.25-1.23 mg/m3; with allergen
(HDM), exacerbated 'f in IL-4, IL-5, IL-13, and IL-17a, but \|/ IFNy (Kim, 2013a): short-term at 0.25 or
1.23 mg/m3;
Lymph Nodes: 'T* IL-4 and IL-10 (and IL-12, slightly), but N/C in IFNy in mice with sensitization (De Jong,
2009): 4 wk at 3.6 mg/m3; thymus: 'T* IL-4 and IL-1B in mice (Jung, 2007): short-term (2 wk) at
>0.5 mg/m3
Bone Marrow
Cell Counts
and Function
High or
Medium
Human: None
No evidence to evaluate
Indeterminate
Animal: 'T bone marrow hyperplasia in rats (Kerns et al., 1983): 24 months at 17.6 mg/m3
Low
Human: None
1 mouse study suggests BM
megakaryocytes may be increased
with short-term exposure at >0.5
mg/m3; Total cell counts are
unchanged with short-term exposure
at <20 mg/m3 in 2 mouse studies,
while excessive levels appear to
cause toxicity
Animal: In mice: N/C in cell counts or functional properties in mice (Dean, 1984): 3 wk at 18.5 mg/m3
[Note: thymus measures also unchanged]; Bone marrow toxicity, impaired function, and decreased cell
counts at excessive levels (Yu, 2014a, 2014b, 2015): short-term at >40 mg/m3; increased
megakaryocytes (Zhang, 2013): short-term at >0.5 mg/m3
Bone Marrow
Cytokines and
other Markers
High or
Medium
Human: None
Indirect evidence suggests no
changes at <2.46 mg/m3
Slight oxidative
stress and
inflammation
Animal: N/C in BM mRNAs or miRNAs in rats (Rager, 2014): short term at 2.46 mg/m3
Low
Human:
3 mouse studies suggest that
oxidative stress is increased with
short-term exposure, even at 0.5
mg/m3.1 short-term mouse study
suggests the BM is damaged and
inflamed, while 1 longer-term rat
study suggests there is no damage
Animal: 'T* indicators of oxidative stress in mice (Ye, 2013; Zhang, 2013; Yu, 2014a, 2014b, 2015):
short-term at >0.5 mg/m3; increased markers of cell death (caspase-3) and inflammation (^ NFkB,
TNFa, IL-ip) in mice (Zhang, 2013 and Yu, 2015): short-term at 3 and 20 mg/m3, respectively; N/C in
DNA or RNA measures of proliferation and health in rats (Dallas, 1987): subchronic at 0.62-18.5 mg/m3
This document is a draft for review purposes only and does not constitute Agency policy.
A-561 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Table A-81. Summary of changes in blood cell counts and immune factors as a result of formaldehyde exposure
Endpoint(s)
No changes observed
(high or medium confidence experiments are bolded)
Significant3 increases or decreases
(high or medium confidence experiments are bolded)
Summary
conclusion
Clarifying notes
Durationb
(species) Concentration(s) [notes] (study)
Duration
(species)b Concentration(s) [notes] (study)
White blood cells (WBCs)
Total WBCs
Years (humans) 0.87 mg/m3 (Lyapina, 2004)
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Years (humans) 4- 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Short term (rats) Bassig, 2016)
Years (humans) >2.46 mg/m3 (Rager, 2014); [indirect]
Unclear0(humans) \|/ <0.29 mg/m3 [mean levels] (Kuo, 1997)
Short term (mice) N/Ah (<1 mg/m3) [yrs, not months] (Thrasher,
1990)
\1/ 0.5-3 mg/m3 (Zhang, 2013)
Moderate 4 4
Possibly concentration-
and/or
duration-dependent,
but this dependence is
unclear
Granulocytes
All
Short term (mice) 18.5 mg/m3 [WBC differentials'1] (Dean,
1984)
Years (humans) 4- 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013,
Bassig, 2016)
Slight 4
most likely neutrophils
at higher
concentrations with
short-term or longer
exposure
Neutrophils
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Short term >9.23 mg/m3 (NTP, 2017)
(mice)Years =0.02 mg/m3 [yr assumed] (Erdei, 2013)
(children) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Years (humans) 0.5-3 mg/m3 (Zhang, 2013)
Short term (mice)
Years (humans) 4- 0.87 mg/m3 [note: function, not counts, in
workers with URT dysfunction] (Lyapina, 2004)
Short term (rats) \|/ 13 mg/m3 (Katnelson, 2013)
Eosinophils
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Years (humans) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Basophils
Years (humans) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Lymphocytes
All
Months (humans) 0.2-0.8 mg/m3 (Jia, 2014)
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Years (humans) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Weeks (humans) 0.51 mg/m3 (Ying, 1999)
Unclear (humans) N/Ah (<1 mg/m3) [yrs vs. months] (Thrasher,
1990)
Short term (mice) 18.5 mg/m3 [WBC differentials0] (Dean,1984)
Years (humans) 4 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Years (humans) Bassig, 2016)
Short term (mice) ^ 0.25 mg/m3 (Aydin, 2013)
Short term (rats) \|/ 0.5-3 mg/m3 (Zhang, 2013)
'T* 13 mg/m3 (Katnelson, 2013)
Indeterminate
multiple changes
noted, but pattern is
indiscernible
This document is a draft for review purposes only and does not constitute Agency policy.
A-562 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Endpoint(s)
No changes observed
(high or medium confidence experiments are bolded)
Significant3 increases or decreases
(high or medium confidence experiments are bolded)
Summary
conclusion
Clarifying notes
Durationb
(species) Concentration(s) [notes] (study)
Duration
(species)b Concentration(s) [notes] (study)
B Cells
Years (humans) 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013,
Bassig, 2016)
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Years (humans) 0.09-0.68 mg/m3 (Thrasher, 1987)
Years (humans) 4- 0.36 [up to 0.69 peaks] mg/m3 (Costa, 2013)
Months (humans) ^ 0.99 [up to 1.69 peaks] mg/m3 (Ye, 2005)
Months (humans) ^ 0.2 and 0.8 mg/m3 (Jia, 2014)
Years (humans) 4- 0.47 [up to 3.94 peaks] mg/m3 (Costa, 2019)
Unclear (humans) 'T* N/Ah (<1 mg/m3) [yrs, not months] (Thrasher,
Weeks (humans) 1990)
¦f 0.51 mg/m3 (Ying, 1999)
Moderate
For altered number of
B cells (direction of
change may differ by
exposure levels or
duration)
T Cells
(Total)
Months (humans) 0.2-0.8 mg/m3 (Jia, 2014)
Unclear (humans) N/Ah (<1 mg/m3) [yrs vs. months] (Thrasher,
1990)
Years (humans) 4- 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Bassig, 2016)
Months (humans 4 0.99 [up to 1.69 peaks] mg/m3 (Ye, 2005)
Years (humans) ¦f- 0.36 [up to 0.69 peaks] mg/m3 (Costa, 2013)
Years (humans) ^ 0.25 mg/m3 (Aydin, 2013)
Years (humans) 0.09-0.68 mg/m3 (Thrasher, 1987)
Years (humans) 0.9 mg/m3 [indirect: apoptosis] (Jakab, 2010)
Weeks (humans) \|/ 0.51 mg/m3 (Ying, 1999)
Short term (rats) ^ 7.4 mg/m3 (Sandicki, 2007a, b)
Slight
mixed results suggests
concentration-
dependence, with \1/ at
higher levels (possibly
'T* at low levels) with
months-years
exposure
T Cells
(CD4+)
Years (humans) 1.6 mg/m3 [>1/ Treg] (Zhang, 2010; Hosgood,
2013; Bassig, 2016)
Months (humans) 0.99 [up to 1.69 peaks] mg/m3 (Ye, 2005)
Years (humans) 0.47 [up to 3.94 peaks] mg/m3 (Costa, 2019)
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Months (humans) 0.2-0.8 mg/m3 (Jia, 2014)
Years (humans) ¦f- 0.36 [up to 0.69 peaks] mg/m3 (Costa, 2013)
Weeks (humans) \|/ 0.51 mg/m3 (Ying, 1999)
Indeterminate
data suggest N/C, but
variable, considering
also studies of spleen
above, suggests effects
may exist at CD4 subset
level
T Cells
(CD8+)
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Years (humans) 0.36 [up to 0.69 peaks] mg/m3 (Costa, 2013)
Months (humans) 0.2-0.8 mg/m3 (Jia, 2014)
[N/C CD4/CD8 ratio in 3 studies and
Thrasher, 1990]
Years (humans) 4 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013)
Months (humans) 4 0.99 [up to 1.69 peaks] mg/m3 (Ye, 2005)
Years (humans) ^ 0.47 [up to 3.94 peaks] mg/m3 (Costa, 2019)
Weeks (humans) >]/ 0.51 mg/m3 (Ying, 1999)[^ CD4/CD8 ratio in
all but one of these studies]
Moderate 4 CD8 and
•f CD4/CD8 ratio
likely related to
concentration
NK Cells
Years (humans) 4 0.36 [up to 0.69 peaks] mg/m3 (Costa, 2013)
Years (humans) 4 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Bassig, 2016)
Years (humans) ^ 0.25 mg/m3 (Aydin, 2013)
Months (humans) ¦f- 0.2, but not at 0.8 mg/m3 (Jia, 2014)
Slight
mixed results suggest
role of concentration
similar to total T cell
findings
This document is a draft for review purposes only and does not constitute Agency policy.
A-563 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Endpoint(s)
No changes observed
(high or medium confidence experiments are bolded)
Significant3 increases or decreases
(high or medium confidence experiments are bolded)
Summary
conclusion
Clarifying notes
Durationb
(species) Concentration(s) [notes] (study)
Duration
(species)b Concentration(s) [notes] (study)
Monocytes
Years (humans) 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Bassig, 2016)
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) ^ =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Short term (mice) \|/ 0.5, but not 3, mg/m3 (Zhang, 2013)
Short term (mice) \|/ 18.5 mg/m3 (Dean, 1984)
Indeterminate
data suggest N/C, at
least in human adults
Red Blood Cells
Years (humans) 0.25 mg/m3 (Aydin, 2013)
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Years (humans) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Years (humans) 4- 0.87 mg/m3 [note: duration] (Lyapina, 2004)
Years (humans) 4- 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013)
Short term (mice) \|/ 0.5-3 mg/m3 (Zhang, 2013)
Moderate 46
suggests combined role
of concentration and
duration
Platelets
Years (humans) 0.87 mg/m3 (Lyapina, 2004)
Short term (mice) >9.23 mg/m3 (NTP, 2017)
Years (children) =0.02 mg/m3 [yr assumed] (Erdei, 2013)
Years (humans) <0.29 mg/m3 [mean levels] (Kuo, 1997)
Years (humans) 4- 1.6 mg/m3 (Zhang, 2010; Hosgood, 2013;
Bassig, 2016)
Short term (mice) ^ 0.5-3 mg/m3 (Zhang, 2013)
Slight 4 7
possible concentration
dependence similar to
above
Secreted factors and immune markers
Primarily Thl-related
TNF-a
Years (humans) 1-g fup to 6 9 pea|
-------
Toxicological Review of Formaldehyde—Inhalation
Endpoint(s)
No changes observed
(high or medium confidence experiments are bolded)
Significant3 increases or decreases
(high or medium confidence experiments are bolded)
Summary
conclusion
Clarifying notes
Durationb
(species) Concentration(s) [notes] (study)
Duration
(species)b Concentration(s) [notes] (study)
IL-8
(neutrophils
)
Months (humans) 0.2-0.8 mg/m3 (Jia, 2014)
Other
Tal
IL-2R
Unclear 3 'T* N/Ah (<1 mg/m3) [yrs, not months, change in
(humans) antigen reactivity markers] (Thrasher, 1990)
Indeterminate
(data suggest N/C in B
cell activation markers)
CD27and
CD30
Years (humans) 1.6 mg/m3 (Bassig, 2016)
Derf: Dermatophagoides farina (house dust mite); OVA: ovalbumin (major protein of chicken egg whites); both are immunogenic materials used to stimulate
an allergy-like response
Gray box = no data meeting the inclusion criteria were available.
Note: one study observing increased substance P and related changes in the serum (Fujimaki et al., 2004) is primarily discussed in the context of changes in the
URTand LRT.
Primarily, this reflects reporting of a statistically significant change; in rare instances where a p value was not given, changes are indicated if the authors
discussed the change as a significant effect.
bHuman study exposure durations are indicated as "years," "months," "weeks," or "acute" and defined based on the anticipated exposure duration for the
majority of the exposed population(s); these durations are interpreted to approximate animal study exposure durations of chronic (>1 year), subchronic
(several months), short term (<30 days), and acute (1 day or less).
cThe comparison presented by Thrasher et al. (1990) reflects differences in exposure duration (years compared to weeks or months), but there appeared to be
minimal difference in concentration.
dThis finding (decreased total WBCs) is supported by 3 studies in humans evaluated by the NRC (2014) [Tang and Zhang, 2003; Tong et al., 2007; Cheng et al.,
2004], but not evaluated in this analysis; additionally, this finding is supported by a study in mice (Yu et al., 2014b) and a study in rats (Brondeau et al., 1990),
which are not included as they only tested formaldehyde levels >20 mg/m3.
eAuthors indicated no changes in "WBC differentials" other than decreased monocytes, but further details NR (Dean, 2013). This test was assumed to include
basic granulocyte and lymphocyte counts.
This finding (decreased erythrocytes) is supported by 1 study in humans evaluated by the NRC (2014) [Yang et al., 2007], but not evaluated in this analysis.
gThis finding (decreased platelets) is supported by 2 studies in humans evaluated by the NRC (2014) [Yang et al., 2007; Tong et al., 2007], but not evaluated in
this analysis, and a mouse study testing excessive formaldehyde levels (Yu et al., 2014b).
hThe exposure level is, in general, considered not applicable (N/A), as the comparison presented by Thrasher et al. (1990) reflected differences in exposure
duration (i.e., years of exposure [Yr], as compared to weeks or months [Mo] of exposure), but there appeared to be minimal differences in concentration from
the controls.
This document is a draft for review purposes only and does not constitute Agency policy.
A-565 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Toxicological Review of Formaldehyde—Inhalation
Consideration of mechanistic changes across tissue compartments
Several interesting relationships across tissue compartments are suggested:
• Evidence of increased oxidative stress, in particular, appears to be conserved across each of
the evaluated tissue compartments. As soluble inflammatory signals can be transmitted
across tissue boundaries with relative ease, it is plausible that these indications of an
increased body burden of free radicals may be an indirect consequence of inflammatory
changes that could be relatively restricted to the airways.
• Observations of increased eosinophils, and to a somewhat lesser extent, neutrophils, in both
the URT and LRT, suggest that the inflammation of the airways caused by formaldehyde
exposure is not restricted to the URT sites directly contacted by the majority of inhaled
formaldehyde.
• Although some more subtle changes appear to occur in the LRT (e.g., inflammation; altered
airway permeability), the data suggest that overt damage to the airway epithelium by
formaldehyde exposure is limited primarily to the URT.
• Key features of several potential health hazards appear to involve mechanistic changes
occurring within multiple tissue compartments, including decreased pulmonary function
and allergic sensitization.
• Although many uncertainties remain, the instances of opposing immune-related responses
in the airways compared to those in the blood suggest immunological communication and
possible recruitment of cells from one compartment to another. One exception to this
pattern was the consistent observation of increased IL-4 in both the LRT and blood. IL-4 is
associated with driving CD4+ T cells towards a Th2 response [Kopf et al., 1993], The
evidence specific to changes in CD4+ T cell populations in either compartment were
inadequate, limiting interpretations of the significance of this finding.
• While many immune-cell-related changes were observed, some only occurred in specific
exposure contexts. For example, neutrophil and monocyte increases in the LRT were
observed only with allergen sensitization, while eosinophil increases were not observed in
studies of exposure less than several weeks; changes in NK cells and other lymphocytes
subsets appeared to vary depending on concentration, and some antibody responses
depended on the antigen (e.g., allergen) type and administration methods. In addition,
immune system studies after developmental exposure represent a significant data gap.
• In general, the evidence becomes less convincing with increasing removal from the
point-of-first-contact for inhaled formaldehyde, with the highest confidence for effects in
the URT, slightly less confidence for effects in the LRT and blood, and a general inability to
draw conclusions regarding the potential for effects in lymphoid organs.
Plausibility of potential associations between mechanistic changes and respiratory system
health effects
Figure A-36 illustrates one or more potential sequences of events from formaldehyde
inhalation to apical outcomes (i.e., key hazard features) described in each of the respiratory system
This document is a draft for review purposes only and does not constitute Agency policy.
A-566 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Toxicological Review of Formaldehyde—Inhalation
health effects sections in the Toxicological Review. Each of these sequences was developed based
on the most reliable mechanistic evidence (i.e., robust or moderate evidence was preferred) that can
plausibly link an initial effect of inhaled formaldehyde to each of these key hazard features, and
which have been demonstrated in formaldehyde-specific studies. Thus, these sequences do not
represent all possible scenarios for which data exist (see Figures A-33 and A-34 for more
comprehensive illustrations), and data not considered in this analysis (e.g., studies of chemicals
closely related to formaldehyde) could identify additional initial alterations and mechanistic events,
as well as more interim changes or relationships between many of the depicted mechanistic events.
As such, this figure may not illustrate the most biologically pertinent sequence of events, but it does
illustrate biologically plausible pathways of effects based on data specific to formaldehyde
exposure. Thus, this is a pragmatic attempt to link early mechanistic events with apical endpoints,
similar to the AOP conceptual framework [Ankley et al., 2010; Villaneuve et al., 2014 a, b]. For each
sequence, an interpretation regarding the likelihood of the presented sequence of events being a
mechanism by which formaldehyde inhalation could cause respiratory system health effects is
provided in the section below. As these interpretations are based on the robustness of the available
evidence, they are primarily based on confidence in the individual studies and the consistency and
coherence of observations across species and experimental paradigms. Other considerations
outlined by Sir Bradford Hill (1965, 71664), including the magnitude and dose-dependency of the
individual study findings, are discussed where the data are available, but these considerations
generally had less of an impact on interpretations. This section references evidence conclusions
from previous sections, as well as studies supporting biological understanding, but individual
formaldehyde-specific studies are generally not referenced.
This document is a draft for review purposes only and does not constitute Agency policy.
A-567 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde—Inhalation
Initial Alterations
Secondary Alterations
Effector-Level Changes Key Hazard Feature
^oxidative URTprotein/DNA URTmucociliary
stress in URT modification dysfunction
URT epithelial URT epithelial Squamous
damage proliferation metaplasia
0 0 <»)
t oxidative URTTRPA1
stress in URT binding
Trigeminal nerve
stimulation in URT
Centrally mediated
sensory irritation
© 0 Robust
--> Moderate
Slight
Key feature of a
potential hazard
© o-
•effects are
amplified with
allergen exposure
-f- oxidative Sensory nerve 't LRT neuro- 'f LRT micro-
stress in LRT stimulation in LRT peptides vascular leakage
^Eosinophils airway edema/ Airway hyper- Decreased
in LRT* inflammatory responsiveness* pulmonary
structural change* function
©
XT -O O O
o
stress ^ oxidative ^ CD8+T cells ^ IL-4, ^ IFNy
hormone stress in blood in blood in blood
Altered B Altered antibody Allergic Airway hyper-
cells responses* sensitization responsiveness*
©
0--SO»D ""O
O
URT epithelial f URT fairway 1s CD8+ Tcells &
damage inflammatory neuropeptides Th2-related
cells & factors cytokines in LRT
^Eosinophils Sustained airway Allergic Airway hyper-
in LRT* inflammation* sensitization responsiveness*
Figure A-34. Possible sequences of mechanistic events identified based on the most reliable evidence available.
This figures presents plausible mechanistic pathways illustrating the most reliable formaldehyde exposure-specific data (i.e., robust or moderate evidence was
preferred) based on currently available information. The figure is organized by respiratory system health effect represented by key features of each hazard
evaluated in the Toxicological Review. The pathways interpreted to most plausibly link possible initial effects of formaldehyde exposure to these apical events
is presented, based on both the confidence in the relationships between events and confidence in the evidence for each of the linked mechanistic events.
These pathways20 are organized in a linear fashion from initial event(s) to key hazard feature(s), and each pathway is numbered, corresponding to the
synthesis that follows. The mechanistic events are grouped into "initial events" and "secondary events" for endpoints that would be expected to occur earlier
and later, respectively, along a sequential mechanistic progression. Generally, for the "initial" events, a preceding or precursor event other than a direct
interaction with formaldehyde is unknown or has not been studied following formaldehyde exposure, or they have been described in previous pathways (e.g.,
see #6). "Effector-level changes" are those events that are most likely to be directly associated with the apical endpoint(s) of interest. The symbols,
descriptors, and arrows are the same as those depicted in Figures A-33-A-34.
20 This approach draws some parallels to the AOP conceptual framework approach (Villenueve et al., 2013; 2014). As such, for those familiar with AOP
terminology, it may be useful to think of the terms used herein according to related AOP terms (e.g., "plausible initial effects of exposure" and "initial
alterations" relate to "molecular initiating events"; "mechanistic events" relate to "key events"; and "key hazard features" relate to "adverse outcomes"].
This document is a draft for review purposes only and does not constitute Agency policy,
A-568" DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
1) Respiratory tract pathology (squamous metaplasia) through epithelial cell damage
Interpretation: This is likely to be a major mechanism by which formaldehyde inhalation
could cause squamous metaplasia.
Consistent with its known chemistry and reactivity, formaldehyde has been shown to react
with DNA and other biological macromolecules at the point of first contact in the URT, where it also
affects tissue redox capacity, presumably either through direct interactions with cellular
macromolecules (e.g., lipids) or indirectly by impacting local tissue detoxification processes. These
initial reactions have been shown to occur following acute and short-term exposure at
concentrations <0.5 mg/m3, and generally, the magnitude of these effects is expected to be driven
largely by formaldehyde concentration and distribution. Distribution of formaldehyde-induced
nasal lesions progresses to more posterior locations with chronic exposure; presumably, this
represents changes in formaldehyde deposition, although this has not been tested. Additionally,
studies have not been performed to address whether long-term exposure may overcome the body's
capacity to regulate or restrict the magnitude of these changes. Elevated oxidative stress could
directly lead to cytotoxic or subcytotoxic epithelial cell damage and/or dysfunction through the
modification of cellular proteins and DNA. Because similar endogenous defense mechanisms (e.g.,
glutathione) are responsible for the detoxification of some free radicals and formaldehyde,
persistent oxidative stress may make these cells more prone to damage directly resulting from
formaldehyde and other inhaled agents. DNA-protein crosslinks (DPXs), which have been observed
at formaldehyde concentrations >0.3 mg/m3 (rats) or >0.9 mg/m3 (rhesus monkeys) and durations
>3 hours (see Appendix A.4), can lead to cellular damage if they are not repaired. Formaldehyde
can modify the structure and function of the mucociliary apparatus, potentially as a result of
covalent modification of soluble factors in the mucus (Morgan et al., 1984) or ciliary proteins
(Hastie et al., 1990). Studies of the mucociliary apparatus following acute exposure provide
evidence for a concentration threshold for functional effects, again highlighting the importance of
formaldehyde concentration and distribution. In rats, DPXs and regions of mucociliary dysfunction
have both been demonstrated to correlate with locations of subsequent respiratory tract pathology
and cell proliferation in the anterior portions of the nasal mucosa following formaldehyde
exposure. The resultant, potentially adaptive, effects on cellular proliferation (i.e., hyperplasia) are
typically dose- and duration-dependent and localized to regions of mucociliary dysfunction and
epithelial damage. Cellular proliferation may be initiated, at least in part, in response to
formaldehyde exposures not associated with histopathological evidence of epithelial cell damage,
since some studies report effects on proliferation at «1 mg/m3. Direct and overt epithelial cell
damage or death associated with squamous metaplasia is not typically observed until formaldehyde
concentrations are above 2 mg/m3. Squamous metaplasia is also localized initially to these
high-flux, anterior regions, but these lesions increase in severity and advance to more posterior
locations with longer exposure. Thus, although some early mechanistic events in this pathway are
This document is a draft for review purposes only and does not constitute Agency policy.
A-569 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
expected to be highly dependent on formaldehyde concentration, the data supports a role for both
exposure duration and concentration in the development of long-term lesions such as squamous
metaplasia.
All of the events in this mechanism are based on robust or moderate evidence, with robust
or moderate evidence for interactions between events, indicating that this mechanism is likely a
major mechanism by which formaldehyde inhalation can cause squamous metaplasia. However,
because modification of epithelial cell health and function in the URT can occur via multiple direct
and indirect mechanisms following formaldehyde inhalation, which are expected to vary due to
differences in both exposure duration and intensity, there are likely to be other important
mechanisms by which formaldehyde exposure could cause respiratory tract pathology.
2) Sensory irritation through trigeminal nerve stimulation
Interpretation: This is likely to be the dominant mechanism by which formaldehyde
inhalation could cause sensory irritation.
With distribution throughout the nasal mucosa, trigeminal nerve endings are well
positioned for direct interactions with inhaled formaldehyde. Trigeminal nerve activation at
unmyelinated C fibers occurs following acute formaldehyde exposure and the resultant
physiological sensation of burning is known to be caused by afferent signaling to the CNS
[Mackenzie et al., 1975], This afferent nerve activity has been demonstrated following
formaldehyde inhalation. Based primarily on indirect evidence (e.g., ex vivo models), activation of
the trigeminal nerve is probably at least partly dependent on direct activation of TRPA1 channels
by formaldehyde (e.g., via binding). Further support for an "irritant receptor" response to
formaldehyde exposure is provided by evidence of competitive inhibition of irritation caused by
chlorine and acetaldehyde [Chang and Barrow, 1984; Babiuk et al., 1985], However, other direct
actions of formaldehyde at trigeminal nerve endings (e.g., binding to other receptors; modification
of ion balance; protein modification) are possible and some other potential pathway scenarios are
suggested. In addition, oxidative stress, such as that elicited in the URT by formaldehyde exposure,
is known to activate TRP channels [Bessac and Jordt, 2008], providing another plausible indirect
mechanism. Based on the proposed sequence of events, sensory irritation would be expected to be
highly variable across individuals due to differences in TRPA1 channel sensitivity or access of
formaldehyde to TRPA1 channels (e.g., due to differences in airway structure, mucus production, or
TRPA1 channel density). Studies of related chemicals suggest that human sensitivity may also be
dependent on demographic factors such as age, sex (women appear to be more sensitive), and
allergy status [Shusterman, 2007; Hummel and Livermore, 2002],
The threshold for activation of exposed rodent nerve endings has been reported at
0.31 mg/m3 formaldehyde. The levels necessary for in vivo activation following acute exposure
may be somewhat higher. Although trigeminal nerve activation may worsen with constant,
repeated exposure to low levels of formaldehyde, as has been demonstrated for other chemicals
This document is a draft for review purposes only and does not constitute Agency policy.
A-570 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
(Brand and Jacquot, 2002), constant exposure or high concentrations could conversely desensitize
this response by excessively stimulatingf the (presumed) irritant receptors. The potential for
sensory irritation to attenuate over time due to processes such as desensitization (e.g., via
internalization of TRPA1 receptors) is unclear, particularly with long-term exposure. Indirect
evidence suggesting either the presence of extremely sensitive individuals in the population or a
role for the duration of exposure in eliciting this effect is provided from residential studies
identifying symptoms associated with sensory irritation at levels as low as 0.1 mg/m3 (e.g., Zhai et
al., 2013; (Liu etal.. 1991): Hanrahan, 1984, 22300). Structural changes to the URT tissue (e.g.,
formaldehyde-induced modification of the epithelial cell layer altering accessibility of sensory
nerve endings) and to the URT response of local immune cells (i.e., inflammatory cells may release
mediators which can stimulate proliferation and/or sensitization of sensory nerve fibers [Carr and
Undem, 2001]) would be expected to be strong modifiers of this effect, introducing an exposure
duration component to the concentration-dependence of receptor binding that is assumed for
activation of TRPA1.
A strong biological understanding exists to identify the physiological sensation of sensory
irritation as being related to stimulated sensory fibers of the trigeminal nerve. While the specific
concentration and duration dependency of activation remain incomplete, based on the robust and
moderate formaldehyde-specific evidence available to support activation of trigeminal nerve fibers
and stimulation of TRPA1 receptors, respectively, along with a general lack of alternative
explanations for chemical-induced sensory irritation, this mechanism is likely the dominant
mechanism by which formaldehyde exposure can cause sensory irritation.
3) Decreased pulmonary function through URT epithelial damage
Interpretation: This is a possible mechanism by which formaldehyde inhalation could
contribute to decreases in pulmonary function, but this is not a major pathway explaining this
potential effect, and other changes are expected to be the primary drivers of any substantial
functional changes.
Airway epithelial cells not only serve as a physical barrier to inhaled pathogens and
antigens, they also participate in the regulation of airway inflammatory responses [Holgate et al.,
1999], The demonstrated modification of the respiratory epithelium in the upper airways by
formaldehyde exposure may affect pulmonary function through both physical, and humoral
mechanisms, although definitive studies for the latter have not been conducted and such factors are
generally tightly controlled and locally acting (e.g., Mayer and Dalpke, 2007 55:353). Modification
to the URT epithelium by formaldehyde, particularly the observed effects on mucociliary function,
is also likely to modify URT barrier and clearance processes, which could increase the impact of
other inhaled antigens on pulmonary function; however, this possibility has not been well-studied.
Physically, swelling of the mucus membrane has been observed in exposed humans at <1 mg/m3
formaldehyde, and this is expected to be highly influenced by the underlying respiratory status of
This document is a draft for review purposes only and does not constitute Agency policy.
A-571 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
the exposed individuals (e.g., allergy status; previous and/or current respiratory infections; etc.).
This swelling can plausibly be linked to narrowing of the airways and impaired pulmonary function,
although this linkage has not been explicitly demonstrated by corresponding effects in the LRT
following formaldehyde exposure and it is unclear to what extent URT swelling would need to
progress before effects on lung function were experienced. Morphological changes in the mucous
membrane can be related to changes in mucus secretion and, possibly, epithelial cell proliferation
[Reader et al., 2003], both of which are observed following formaldehyde exposure. Dysfunction of
airway epithelial cells can also modify their release of humoral factors, which help to regulate
airway smooth muscle contraction and immune cell responses. For example, epithelial cells can
release neutral endopeptidase, which is the major metabolizing enzyme for tachykinins such as
substance P and neurokinin A [Barnes, 1992], and they are known to produce situation-specific
signals that can either promote or inhibit the activity of local immune cells, including dendritic cells,
which contribute to airway remodeling [Lambrecht and Hammad, 2012], In these ways,
modification of the function of URT epithelial cells by formaldehyde exposure might result, in an
indirect manner, in changes in humoral factors that could reach the lower airways and lungs in
minimal amounts. However, direct formaldehyde-specific examinations of such potential
associations, including the requisite exposure parameters (e.g., levels), were not identified.
This sequence of events can plausibly link structural damage and dysfunction of the
epithelium in the URT to potential decrements in pulmonary function. However, a large amount of
missing information, particularly regarding LRT changes, is assumed, and evidence linking these
formaldehyde-induced mechanistic events in the URT to changes in pulmonary function has not
been reliably demonstrated. While these events might contribute to some minimal level of
decrease in pulmonary function, the data are insufficient to identify this sequence of events as a
major mechanism.
4) Airway hyperresponsiveness and/or decreased pulmonary function through LRT
inflammatory changes resulting from sensory nerve activation
Interpretation: This is likely to be an incomplete mechanism by which formaldehyde
inhalation could cause airway hyperresponsiveness and decreased pulmonary function, although
whether certain events occur at low exposure levels is unclear.
Activation of airway sensory nerve endings is known to cause the release of neuropeptides,
including substance P. Short-term formaldehyde exposure appears to cause increases in substance
P, and perhaps other neuropeptides, in the lower airways. In addition, several lines of evidence
identify potential substance P-related changes in the LRT that are at least partially dependent on
TRP channel activation. As discussed previously, while certain, very rare human exposure
scenarios might result in weak activation of the vagus nerve in proximal regions of the LRT (e.g., the
trachea) due to direct interactions with formaldehyde, it is expected that the predominant
explanation (and that most relevant to interpretations) for activation remains unidentified and
This document is a draft for review purposes only and does not constitute Agency policy.
A-572 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
involves indirect pathway(s). One possible explanation involves indirect activation of LRT sensory
nerve endings in association with the formaldehyde exposure-induced increases in LRT oxidative
stress and/or inflammation, as certain electrophilic oxidative byproducts and inflammatory factors
can stimulate TRPA1 channels (Taylor-Clark etal., 2008; Andersson et al., 2008). Alternatively,
substance P could also be directly released from certain subsets of activated immune cells,
including eosinophils [Joos et al., 2000: 55], which are increased in the LRT, although this
hypothesis has not been examined and may be somewhat less plausible, given the apparent
discrepancy in the exposure duration required for substance P increases versus LRT eosinophil
increases in the available studies. Regardless, any indirect pathway(s) would require prior
modification of the LRT microenvironment after formaldehyde exposure through a separate,
undefined mechanism.
Locally, substance P can cause vasodilation and leakage or constriction of airway smooth
muscle, the latter of which appears to be enhanced in asthmatics (who also exhibit elevated
substance P-immunoreactivity in airway nerves; Ollerenshaw et al., 1991: 673), all of which can
contribute to airway narrowing or obstruction (Joos etal., 1994: 1161; Joos etal., 1995: 329). It
should be noted that airway obstruction typically requires much higher doses of agonist than does
leakage (e.g., Yiamouyiannis etal., 1995). Formaldehyde-induced increases in substance P
contribute to microvascular leakage in the LRT (i.e., trachea and main bronchi) following acute
formaldehyde exposure, which has been observed at >1 mg/m3. Specifically, although the effects of
prolonged exposure were not examined, at higher formaldehyde levels (i.e., >10 mg/m3) and with
acute exposure, microvascular leakage was blocked by inhibition of the neurokinin 1 (NKi)
receptor, and perhaps also by inhibiting mast cell activation, but not by inhibition of histamine,
cyclooxygenases, or bradykinin. Substance P is the preferred substrate for NKi receptors. Although
activation of NKi receptors can contribute to structural changes in human airways, these receptors
are more commonly associated with increases in airway inflammation (Schuiling et al., 1999: 423).
As introduced above, NKi receptors are also implicated in establishing the successful recruitment
and adhesion of eosinophils and neutrophils to inflamed airways (Baluk, 1995), at which point
these cells can release bronchoconstrictors. Thus, the increase in LRT eosinophils observed
following formaldehyde exposure (and the slight evidence for increased neutrophils with allergen
sensitization) could be related to elevated substance P. In addition, substance P itself can increase
the responsiveness of the airways to bronchoconstrictors (Cheung et al., 1994: 77,1325). Thus,
either directly, or indirectly, the release of neuropeptides, presumably from stimulated sensory
nerve endings, could result in airway hyperresponsivness. Perhaps relatedly, possible
consequences of increased microvascular leakage and inflammation include airway edema and
related structural changes, which have been reported following short-term formaldehyde
exposures ranging from >0.3 to >3 mg/m3 across studies, although these events have not been
experimentally linked to sensory nerve stimulation or substance P signaling. Taken together, it is
This document is a draft for review purposes only and does not constitute Agency policy.
A-573 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
plausible that substance P-mediated inflammatory alterations to the lower airways, were they of
sufficient severity, could also lead to decreases in pulmonary function.
Several notable uncertainties exist for this plausible mechanistic pathway. As discussed
above, an understanding of the sequence of events preceding the observed changes in the LRT
remains largely incomplete. In addition, and perhaps most importantly, while most of the evidence
is moderate, the data are based almost exclusively on acute or short-term experiments. Similarly,
while evidence for some events at low formaldehyde levels (e.g., <1 mg/m3) exists, some of the
more convincing associations, including the requirement of NKi receptor activation for
microvascular leakage, have only been tested at very high formaldehyde concentrations (e.g.,
>10 mg/m3). Taken together, these limitations raise uncertainties for the relevance of this specific
pathway to chronic, low-level exposure scenarios. Further, several important events related to this
pathway have not been well studied. For example, the available studies have not examined the
potential for sensory nerve activation to modify smooth muscle tone (e.g., regulation of contractile
responses through the electrical activity; release of factors with direct action on smooth muscle
cells, such as acetylcholine), and information does not exist to ascertain whether NK2 receptor
activation by neurokinin A, which can be a more potent bronchoconstrictor than substance P
[Kraneveld et al., 2002], might be involved. Also, while substance P can stimulate mast cell
degranulation and release of bronchoconstrictors such as histamine (Lilly etal., 1995: 1234; Suzuki
et al., 1995: 1447), in vivo evidence of changes in mast cells was not identified. However, given the
recruitment of other immune cells to the airways after formaldehyde exposure, an event that can be
mediated by mast cells [Dawicki, 2007], data on mast cells may represent critical information that
is missing from the present analysis. Overall, based on the consistent moderate evidence for
changes in the LRT that are commonly associated with changes in pulmonary function and airway
responsiveness, this incomplete sequence of events is likely one of the mechanisms by which
formaldehyde exposure could cause airway hyperresponsiveness and decreased pulmonary
function. However, the pertinence of some or all of the components in this pathway with long-term,
low-level formaldehyde exposure is unknown, and it is considered likely that other important
mechanistic events would be identified with additional studies, particularly those testing longer
exposure durations. It remains unclear how directly translatable this pathway, based largely on
animal data, might be to interpreting complex human diseases such as asthma, and notable events
thought to be important to the development or progression of asthma have not been observed.
5) Allergic sensitization and airway hyperreactivity through altered antibody-related
responses in the blood
Interpretation: It is unclear whether this is a possible mechanism by which formaldehyde
inhalation could cause these effects, as an understanding of the potential mechanistic relationships
is incomplete.
This document is a draft for review purposes only and does not constitute Agency policy.
A-574 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Many reactive oxygen and nitrogen species (ROS, RNS) can be essential immunomodulatory
signaling molecules. However, prolonged or excessive exposure to these factors can modify the
structural and functional integrity of a wide range of cell and tissue types. Elevated indicators of
oxidative stress have been identified in nearly all tissues examined following formaldehyde
exposure, including the blood. In the blood of exposed humans, formaldehyde concentrations as
low as 0.1 mg/m3 have been shown to cause lipid peroxidation in peripheral immune cells, typically
with prolonged exposure. The data are not available to demonstrate what might be causing this
increase in free radicals, although factors released into the circulation as a result of pronounced or
sustained airway inflammation would be expected to be capable of causing such an effect.
Specifically regarding the elevated corticosterone levels, which have been reported in rats exposed
for several weeks to much higher formaldehyde levels (3 mg/m3), an excess of glucocorticoids is
typically associated with the inhibition of T cell cytokine secretion and function, although they may
more specifically enhance the Th2 lineage and suppress the Thl lineage (Ashwell et al., 2000: 309;
Elenkov et al., 2004: 1024). However, the varied roles for stress hormones (and free radicals) in the
regulation of immune responses are complex [Glaser et al., 2005: 243], Formaldehyde-specific
studies examining the dynamics of this potential interplay were not identified.
Immunomodulatory effects of circulating stress hormones (and free radicals), could
plausibly be associated with changes in circulating immune cells. As previously mentioned,
although formaldehyde-induced changes in circulating immune cells were consistently observed,
they varied in magnitude and direction across studies, suggesting a complex regulatory
mechanism(s) for these effects. For example, decreases in CD8+ T cells were primarily observed in
the blood of individuals exposed to higher levels of formaldehyde (>0.5 mg/m3), but not in studies
testing lower exposure levels for comparable durations. CD8+ T cells are composed of five
subpopulations with numerous roles for both cell-mediated immunity and Th2-mediated allergies
[Mittrucker et al., 2014], However, the majority of formaldehyde-specific studies evaluating T cell
responses did not distinguish subpopulations of CD4+ or CD8+ T cells, since a number of these
subpopulations have only recently been discovered, and some studies only assessed total T cells
(see Table 1-31). This complicates interpretations of these responses and raises the possibility that
more consistency in changes across studies may exist for specific T cell subpopulations. Perhaps
more importantly, the evidence for changes in CD4+ T cells, which would be highly informative to
this analysis as they are viewed as critical to the development of hypersensitivity [Cohn et al.,
2004], was mixed and uninterpretable. Stimulated CD8+ T cells produce IFN-y, providing a
plausible linkage between the decreases in CD8+ T cells and the decrease in IFN-y at >0.75 mg/m3
formaldehyde in several studies. The observed increase in IL-4 at similar formaldehyde levels is
more complicated, as its regulation is tightly controlled and likely to be mediated by multiple
mechanisms. B cell proliferation and production of IgE and certain IgG subtypes is dependent on
IL-4 and inhibited by IFN-y [Paul, 1987], providing support for a relationship between these
cytokine changes and altered IgG-related responses. The evidence of alterations in the number of B
This document is a draft for review purposes only and does not constitute Agency policy.
A-575 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
cells, as well as the potential relationship between B cell levels and Ig levels, would benefit from
additional study.
Understanding the regulation and function of IgE and IgG responses continues to evolve.
IgE has a clear role in the development of allergic diseases that affect the airways, including allergic
asthma, although IgE may not always be essential (e.g., in other types of asthma; in other allergic
disorders). In contrast, IgG responses are poorly understood. While IgG may help to exacerbate IgE
responses (e.g., patients with increases in both IgE and IgG are at greatest risk for developing
allergic responses) and IgGs alone might induce allergic reactions to certain antigens [Wu and
Zarrin, 2014; Williams etal., 2012; Finkelman, 2007], an excess of IgG antibodies can prevent IgE -
mediated hypersensitivity and persons with increases in IgG alone are not typically at increased
risk for allergic-related responses [Strait et al., 2006; Pandey, 2013; Williams etal., 2012], The
evidence from formaldehyde-specific studies is insufficient to clarify whether IgE-mediated
responses are involved (i.e., the evidence was considered slight, and was generally mixed and
inconclusive), nor is it clear that changes in IgG are related to the development of sensitization or
airway hyperresponsiveness. Further clarification of the observed IgG changes is also necessary, as
some of the changes noted in response to formaldehyde exposure may depend on the duration of
exposure or the specific IgG subtype examined. The antibody-related responses discussed herein
have only been measured in the blood, as compared to samples that might be more directly
informative to immune responses in the airways (e.g., nasal lavage or BAL). This is a notable data
gap, given the somewhat disparate findings regarding immune cell counts in the airways and the
blood. Overall, there are still critical uncertainties in the formaldehyde-specific antibody data.
In typical allergic disorders, changes in CD4+ Th2 cells are present and are thought to play a
prominent role, whereas CD8+ T cell responses are generally lacking. Similarly, although IgG might
contribute to allergic sensitization, the prototypical antibody response in allergy is thought to be
largely driven by IgE. While it is possible that formaldehyde exposure may cause
sensitization-related responses through a predominant IgG response rather than through IgE, the
data demonstrating or proving such a linkage are not currently available. Overall, the available
formaldehyde-specific studies do not provide information sufficient to disentangle the complex
interplay between CD4+ and CD8+ T cells and B cells, regulatory cytokines such as IL-4, and the IgG
and IgE responses that might underly the potential for formaldehyde to induce the interrelated
immune effects of allergic sensitization and airway hyperresponsiveness.
Overall, the potential sequence(s) of events that may underly the observed changes in
circulating antibodies remains poorly defined. Further, although a linkage between IgG responses
and hypersensitivity is plausible, additional clarification is needed regarding the potential role for
these types of changes in the pathogenesis of airway disease. Thus, based largely on an incomplete
understanding of the necessity and ability of changes in IgG to induce these responses, and a lack of
convincing formaldehyde-specific evidence demonstrating changes in IgE, it is unclear whether this
is a possible mechanism by which formaldehyde exposure might cause these immune effects.
This document is a draft for review purposes only and does not constitute Agency policy.
A-576 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
6) Airway hyperresponsiveness and allergic sensitization through airway eosinophilia and/or
sustained airway inflammation
Interpretation: This is a likely a mechanism by which formaldehyde inhalation could cause
airway hyperresponsiveness in those sensitized to allergens, although additional unidentified
events are expected to contribute. It is also a possible mechanism by which formaldehyde
inhalation could cause airway hyperresponsiveness in nonsensitized individuals. Whether this
mechanism is useful for explaining the development of allergic sensitization is unclear.
A number of studies demonstrate that short-term formaldehyde exposure, and possibly
longer-term exposure (the data are sparse), can cause an increase in eosinophils in both the upper
and lower airways, particularly in animals sensitized to allergens. As previously mentioned, an
understanding of how this recruitment occurs remains unclear. Although specific events proving a
linkage have not been demonstrated, other formaldehyde-specific observations may be associated
with this change. For example, airway epithelial cells, which are modified as a result of
formaldehyde exposure, can release immuno-stimulatory factors, including the Th2 cytokines, IL-4
and IL-13, when exposed to allergens [Li et al., 1999], While changes in IL-4 have been noted in the
LRT and could plausibly be related to altered epithelial cells mediating recruitment of eosinophils,
the more important, and thus more convincing, evidence of such a linkage would involve increases
in IL-3, IL-5, IL-13, GM-CSF, and/or eotaxin [Jacobsen etal., 2014; Trivedi and Lloyd, 2007; Wang et
al., 2007]; however, the formaldehyde-specific evidence related to these latter factors is limited and
generally inconsistent Alternatively, eosinophil recruitment could be related to increased
neuropeptide release from stimulated sensory nerve endings, as previously discussed.
Bidirectional communication exists between sensory nerve endings and immune cells of the
airways, and neuropeptide release can be enhanced by various cytokines and neurotrophins,
including nerve growth factor (NGF) [Nokher and Renz, 2005], NGF, which can also induce mast
cell degranulation and shift T cells towards a Th2 response [Mostafa, 2009; de vries et al., 2001]
and drive antigen-induced and tachykinin-mediated increases in inflammatory cells such as
eosinophils [Quarcoo et al., 2004], may also be modified in the airways following formaldehyde
exposure [Fujimaki, 2004] (not shown in Figures A-33-A-35). Specifically regarding eosinophils,
released neuropeptides such as substance P have been shown to prime eosinophils for chemotaxis
by other factors such as leukotrienes or IL-5, and these neuropeptides can induce accumulated
eosinophils to release factors associated with cellular activation, such as eosinophil cationic protein
[Kraneveld and Nijkamp, 2001], Similar to the lack of evidence supporting a linkage with altered
epithelial cell function, formaldehyde-specific data are not available to inform such potential
linkages. Indirectly, neuropeptide release could also be associated with facilitating the recruitment
of eosinophils to the airway by increasing the permeability of the microvasculature, although this
evidence still fails to identify the immuno-attractant stimuli. Given the gaps in these linkages, it is
likely that this sequence of events is incomplete. Of specific note, evidence of changes in CD4+ Th2
cells in the LRT would be expected for each of these potential scenarios leading to eosinophil
This document is a draft for review purposes only and does not constitute Agency policy.
A-577 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for Formaldehyde—Inhalation
recruitment, as these cells release factors such as IL-5 and are known to aid eosinophil recruitment
in multiple experimental scenarios [Trivedi and Lloyd, 2007; Hogan et al., 1998],
Regardless of the mechanism of recruitment, the evidence indicates that airway eosinophils
are increased by formaldehyde exposure, and activated eosinophils are known to affect airway
contractile responses. Thus, even a short-lived increase in eosinophils could increase
bronchoconstriction (e.g., through the release of mediators such as leukotrienes, major basic
protein and M2 receptor antagonists, and through the activation of other immune cells such as mast
cells and basophils, all of which can act on smooth muscle). However, the relationship of increased
eosinophils to airway hyperresponsiveness or allergic sensitization to nonspecific stimuli is more
complicated and depends on a combination of factors, many of which the formaldehyde-specific
data do not address. For example, the longevity of this eosinophilic response following
formaldehyde exposure, particularly in healthy individuals, remains unclear. Short-term eosinophil
effects on pulmonary function with subsequent clearance of these cells from the airways would be
unlikely to lead to prolonged hypersensitivity of the airways, which would be expected to involve
persistent activation of these cells and continued production of pro-inflammatory mediators. A
single animal study suggests that eosinophils persist with subchronic formaldehyde exposure at
2.3 mg/m3 (but not at <0.5 mg/m3) in animals sensitized to allergen [Fujimaki, 2004], and other
indirect evidence indicates that inflammation of the airways persists with long term formaldehyde
exposure, particularly in those sensitized to allergens (see Table 1-80). However, it remains
unknown whether these latter findings reflect the involvement of the populations of immune cells
and secreted factors believed to be critical to the development of airway hyperresponsiveness. As
previously described, the evidence examining the involvement of other important
immunomodulatory events expected to affect airway responsiveness and allergic sensitization,
including activation of basophils and mast cells, recruitment and/or development of a Th2
phenotype in CD4+ T cells, evidence of remodeling21 in the bronchi and/or alveoli, and changes in
secreted factors known to affect smooth muscle reactivity, is generally slight or inadequate. These
represent important data gaps.
Some experimental animal studies also report data suggesting increases in CD8+ T cells in
the LRT at very high levels of formaldehyde (>5 mg/m3) with short term exposure. Similar to the
observed LRT increases in eosinophils, the mechanism(s) mediating this recruitment to the airways
is unknown, but likely to be downstream of formaldehyde-induced changes to epithelial cells
and/or sensory nerve fibers. The observation of this change alongside the moderate evidence of
decreases in CD8+ T cells in the blood, generally suggesting a threshold for this effect around
0.5 mg/m3, is of interest (note: similar trends in changes in other cells populations, including NK
21 "Airway remodeling" has a specific meaning in human airway disease (see Bergeron and Boulet, 2006).
Several formaldehyde-specific animal studies defined the observed airway structural changes as remodeling
(e.g., Liu et al., 2011; Qiao et al., 2009; Wu et al., 2013). Although the studies' data may relate to some aspects
of airway remodeling, they are more generally described herein as inflammatory histologic changes to avoid
misinterpretation.
This document is a draft for review purposes only and does not constitute Agency policy.
A-578 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
cells, were also observed). Recruitment of lymphocytes to inflamed airways from the blood in
response to acute insults is assumed for multiple respiratory disorders [Medoff et al., 2005] and has
been demonstrated with different pathogenic stimuli, including exacerbation of asthma or COPD by
rhinovirus infection [Message et al., 2008; Mallia et al., 2014], In these models, rhinovirus challenge
generally causes an increase in BAL cells, including eosinophils and CD8+ lymphocytes (and
possibly neutrophils), while cell counts in the blood, including CD4+ and CD8+ T cells (and possibly
NK cells) are decreased. In these types of studies, the specific relationship and magnitude of these
changes appears to depend on the "dose" (e.g., viral load), as well as the sequence of pathology (e.g.,
viral challenge in symptomatic individuals). While the exact mechanisms underlying these
complementary changes are unclear, hypotheses include modifications to epithelial cell function
that leads to exaggerated immune responses in the absence of cytotoxicity [Proud, 2011; Gavala et
al., 2013], Thus, some of the observed airway inflammatory responses could be mediated through a
sequence of events resulting from recruitment of certain immune cell populations from the blood to
the airways, which may be directly relevant to changes observed in acutely challenged humans with
airway disorders.
Overall, the evidence for persistent increases in airway immune cells and other
immunomodulatory factors following formaldehyde exposure in individuals with prior allergen
sensitization is interpreted as likely to represent an incomplete mechanism that could lead to
airway hyperresponsiveness, as relevant observations have been reported after long-term
exposure. However, the currently available data are insufficient to indicate this sequence of events
as a likely mechanism for airway hyperresponsiveness in nonsensitized individuals. Owing to the
lack of reliable formaldehyde-specific evidence demonstrating changes in IgE and other
immunomodulatory factors assumed to be essential to the development of allergic responses, it is
unclear whether this is a possible mechanism by which formaldehyde might cause allergic
sensitization. Similarly, it remains unclear how useful this pathway might be to interpreting
complex human diseases such as asthma. Additional studies are needed, particularly those
employing long-term, low-level formaldehyde exposure.
Consideration of mechanistic pathways that may be associated with each potential respiratory
system health effect
Several conclusions are suggested by the analyses of potential mechanistic pathways that
might be associated with individual respiratory health effects, based on the most reliable
formaldehyde-specific data:
• The confidence in the suggested mechanistic associations varies across the respiratory
system health effects. While some uncertainties remain, important mechanistic events
associated with sensory irritation, squamous metaplasia, and to a lesser extent, decreased
pulmonary function, are supported by robust or moderate formaldehyde-specific data, and
the relationships described are largely well-understood biological phenomena or have been
demonstrated following formaldehyde exposure. Comparatively, the understanding of
This document is a draft for review purposes only and does not constitute Agency policy.
A-579 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
mechanisms for potential immune effects is less complete. While moderate evidence exists
for several mechanistic events that are likely to be involved in the development of airway
hyperresponsiveness, the effect(s) at the point of contact that leads to these events is
unclear. The mechanistic evidence describing the potential development of allergic
sensitization is the most limited, as it includes slight evidence for several events, and the
majority of the potential mechanistic relationships have not been experimentally validated
and a clear scientific consensus regarding the relationships does not exist
• The primary mechanism for sensory irritation is considered well understood, although it is
based largely on acute or short-term exposures, and sensitivity is expected to vary between
individuals. While studies clarifying the effects of tissue modification with longer term
exposure in humans would be useful, it is likely that rodents exposed to «0.2 mg/m3
formaldehyde under normal conditions would exhibit this effect However, as exposure to
formaldehyde appears to cause airway inflammation, which can increase the sensitivity and
response magnitude of sensory nerve fibers, inflammation is viewed as a likely modifier of
sensory irritation.
• At least one of the mechanisms by which formaldehyde exposure could cause squamous
metaplasia is considered well understood, and it appears to depend on both exposure level
and duration. Based on the pathway presented, these events are likely to occur at similar or
slightly higher formaldehyde levels than those causing sensory irritation, and while
cumulative tissue modifications with longer exposure or differences in human anatomy may
increase sensitivity, the available experimental animal evidence suggests that pronounced
effects leading to metaplasia are unlikely below 0.5 mg/m3.
• Several contributing mechanistic pathways appear to impact pulmonary function, and the
complex interactions within and across these pathways are expected to involve additional,
unidentified factors. While some important mechanistic changes occur at low formaldehyde
exposure levels (e.g., <0.2 mg/m3 in rodents), data are not available to quantitatively relate
these changes to decrements in pulmonary function. In addition, sensitivity is expected to
be influenced by the respiratory health of exposed individuals. As with the mechanistic
evidence supporting other health effects, much of the data is based on short term exposure.
As exposure duration increases, and in the absence of potential compensatory mechanisms
(which remains largely unexamined), amplification of these mechanistic events is expected.
• Given the lack of clear explanatory mechanisms for allergic sensitization, in particular, and
uncertainties in data that may help to explain airway hyperresponsiveness, as well as an
expectation of a large amount of important information that has not yet been identified in
formaldehyde-specific studies, it is difficult to speculate on the exposure level- and
duration-dependence of these potential pathways. However, some of the important events
that may be involved (e.g., eosinophil increases) suggest a duration-dependence for the
development of persistent changes in the sensitivity of the airways (note: transient
hyperresponsiveness may be possible with short-term exposure), while other important
data suggest that a concentration threshold likely exists in regard to critical changes in the
cellular immune responses. Individual variability, including underlying respiratory health,
is expected to be a significant modifier of these effects.
This document is a draft for review purposes only and does not constitute Agency policy.
A-580 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
A.5.7. Nervous System Effects
Literature Search
A systematic evaluation of the literature database on studies examining the potential for
noncancer nervous system effects in humans or animals in relation to formaldehyde exposure was
initially conducted in 2012, with regular updates as described elsewhere (including a separate
Systematic Evidence Map thatupdates the literature from 2017-2021 using parallel approaches;
see Appendix F). . The search strings used in specific databases are shown in Table A-82.
Additional search strategies included:
• Review of reference lists in the articles identified through the full screening process.
• Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
EPA. 2010b). the ATSDR toxicological profile of formaldehyde (ASTSDR, 1999), and the NTP
report on carcinogens background document for formaldehyde (NTP, 2010).
• "Snowball": review of references in review articles relating to formaldehyde and
neurological effects (based on title and abstract screening), published in English, identified
in the initial database search. For these articles, references were retrieved through Web of
Science and added to the database via electronic export; manual review of references were
conducted for the three reviews that were not found in Web of Science. Review articles that
contained primary data were retained after full text screening.
This broad literature search was designed to identify studies in humans or animals that
examined objective, apical effects on the nervous system, including structural, behavioral, chemical,
and electrophysiological changes, as well as mechanistic studies informing potential biological
associations between formaldehyde exposure and nervous system effects. Given the general lack of
distribution of inhaled formaldehyde to the nervous system, likely in contrast to other routes of
exposure and which complicates interpretations of direct interactions of formaldehyde with
nervous system cells in tissue culture models, this search focused on inhalation exposure studies.
Inclusion and exclusion criteria used in the screening steps are described in Table A-83.
The search and screening strategy, including exclusion categories applied and the number
of articles excluded within each exclusion category, is summarized in Figure A-37. Although these
noninhalation studies were considered for use, possibly to describe (in)consistent findings across
exposure routes or as qualitative support for toxicological or mechanistic findings from inhalation
studies, given the toxicokinetic uncertainties (e.g., possible differential distribution to the CNS),
they ultimately were not included in the synthesis and were not considered further.
This document is a draft for review purposes only and does not constitute Agency policy.
A-581 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-82. Summary of search terms for neurological effects
Database,
Search
Parameters
Terms
PubMed
No date
restriction
(formaldehyde [majr] OR paraformaldehyde) AND (neuron OR neurons OR neurono* OR
neurolo* OR neuronal OR neurotox* OR neurophys* OR neurochem* OR neurotrans* OR
neuropsych* OR neuropath* OR neuromusc* OR nerve OR nerves OR nervous OR
electrophys* OR "evoked potential" OR *encephalog* OR encephalop* OR *sensory OR
sensori* OR "central nervous system" OR CNS OR brain OR spine OR spinal OR spino* OR
*axon* OR *synapt* OR *synaps* OR *myelin* OR dendrite* OR *behavior* OR learn* OR
memory OR *motor OR *motion OR operant OR habituat* OR *coordination OR weakness OR
righting OR reflex OR psychologic* OR mood OR sleep* OR visual OR audit* OR touch OR taste
OR sound OR smell OR "pain sensitivity" OR nociception OR olfact* OR *glia* OR oligoden* OR
astrocyte* OR balance OR sensation OR sensitization OR tremor* OR convuls* OR seizure* OR
grip OR gait OR paralysis OR posture OR mobility OR rearing OR splay OR stereotypy OR
conditioning OR avoidance OR approach OR neuropath* OR attenti* OR aggressi* OR arous*)
NOT ("formalin test" OR "formaldehyde fixation" OR "formalin fixation" OR "formalin fixed"
OR "formaldehyde fixed" OR "formalin-induced" OR "formalin-evoked")
[Note: for quality control, =10% (50) of the 451 excluded article titles were scanned in
PubMed: none were relevant]
Web of Science
No date
restriction
Lemmatization
"off
SU= ("Anatomy & Morphology" OR "Behavioral Sciences" OR "Biochemistry & Molecular
Biology" OR "Cell Biology" OR "Developmental Biology" OR "Life Sciences Biomedicine Other
Topics" OR "Neurosciences & Neurology" OR Pathology OR Pediatrics OR Physiology OR
"Public, Environmental & Occupational Health" OR "Reproductive Biology" OR "Research &
Experimental Medicine" OR Toxicology OR "Veterinary Sciences" OR Psychology) AND TS=
(formaldehyde OR paraformaldehyde OR formalin) AND TS= (neuron OR neurons OR
neurono* OR neurolo* OR neuronal OR neurotox* OR neurophys* OR neurochem* OR
neurotrans* OR neuropsych* OR neuropath* OR neuromusc* OR nerve OR nerves OR nervous
OR electrophys* OR "evoked potential" OR *encephalog* OR encephalop* OR *sensory OR
sensori* OR "central nervous system" OR CNS OR brain OR spine OR spinal OR spino* OR
*axon* OR *synapt* OR *synaps* OR *myelin* OR dendrite* OR *behavior* OR learn* OR
memory OR *motor OR *motion OR operant OR habituat* OR *coordination OR weakness OR
righting OR reflex OR psychologic* OR mood OR sleep* OR visual OR audit* OR touch OR taste
OR sound OR smell OR "pain sensitivity" OR nociception OR olfact* OR *glia* OR oligoden* OR
astrocyte* OR balance OR sensation OR sensitization OR tremor* OR convuls* OR seizure* OR
grip OR gait OR paralysis OR posture OR mobility OR rearing OR splay OR stereotypy OR
conditioning OR avoidance OR approach OR neuropath* OR attenti* OR aggressi* OR arous*)
NOT TS= ("formalin test" OR "formaldehyde fixation" OR "formalin fixation" OR "formalin
fixed" OR "formaldehyde fixed" OR "formalin-induced" OR "formalin-evoked")
[Note: for quality control, -2% (80) of the 3,825 excluded article titles were scanned in WoS:
none were relevant].
ToxNet (Toxline
and DART)
No date
restriction
formaldehyde AND (neurol* OR neurotox*)
(including synonyms and CAS numbers, but excluding PubMed records)
TCATS2
Restricted to
01/01/2010 and
newer
"formaldehyde" OR CAS Number: "50-00-0"
This document is a draft for review purposes only and does not constitute Agency policy.
A-582 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-83. Inclusion and exclusion criteria for studies of nervous
system effects
Included
Excluded
Population
•
Experimental animals
•
Irrelevant species or matrix*, including nonanimal
•
Humans
species (e.g., bacteria) and studies of inorganic
products
Exposu re
•
Quantified (e.g., levels;
duration) exposure to
•
Not specific to formaldehyde* (e.g., other
chemicals)
inhaled formaldehyde
•
No specific comparison to formaldehyde exposure
in indoor air
•
•
(e.g., formaldehyde levels, duration, or similar in a
study of exposure to a mixture)—NOTE: full text
screening only
Outdoor air formaldehyde exposure—NOTE: full
text screening only
Nonrelevant exposure paradigm* (e.g., use as a
pain inducer in nociception studies)
Comparison
•
Inclusion of a
comparison group (e.g.,
pre- or postexposure;
no exposure; lower
formaldehyde exposure
level)
•
Case reports (selected references used for
illustration)
Outcome
•
Nervous system effects
that could indicate a
•
Subjective symptoms, including headache, fatigue,
etc.
hazard (e.g.,
•
Effects other than noncancer nervous system
behavioral, chemical,
effects*, including carcinogenicity studies
structural, or
•
Exposure or dosimetry studies*
physiological)
•
Use of formaldehyde in methods* (e.g., for
•
Mechanistic studies
fixation)
examining aspects of
•
Processes related to endogenous formaldehyde*
nervous system
function
Other
•
Original primary
research article
•
•
Not a unique, primary research article*, including
reviews, reports, commentaries, meeting abstracts,
duplicates, or nonessential untranslated foreign
language studies (these were determined to be off
topic or unlikely to have a significant impact based
on review of title, abstract, or figures).
Related to policy or current practice* (e.g., risk
assessment/management approaches or models)
* Indicates criterion tags used in HERO for excluded studies
This document is a draft for review purposes only and does not constitute Agency policy.
A-583 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
PubMed
Toxline, TSCATS,
& DART
Web of Science
A3
-121
^
+344
+308
+286
^
*
>
+257
+1318
Electronic Duplicate Removal
6529 Articles
-6305
Title & Abstract Screen
Excluded, did not meet criteria for:
Population-122
Exposure-1270
Outcome- 4348
Other- 565
224 Articles
-79
Full Text Screen
Excluded, did not meet criteria for:
Exposure- 30
Outcome-19
Other- 30
+ 2
147 Articles
147 articles "considered": ail human (40) & animal inhalation (42 toxicology and 18 MOA-only) studies
were evaluated (see Appendix BBB); in vitro & non-inhalation studies (47) did not inform Hazard ID
Figure A-35. Literature search documentation for sources of primary data
pertaining to formaldehyde exposure and nervous system effects (reflects
studies identified in searches conducted through September 2016).
This document is a draft for review purposes only and does not constitute Agency policy,
A-584' DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
The studies identified in the literature search and screening process were evaluated using a
systematic approach to identify strengths and limitations, and to rate the confidence in the results.
EPA evaluated observational epidemiology studies of neurobehavioral effects and of risk of
amyotrophic lateral sclerosis (ALS), controlled human exposure studies of neurobehavioral effects,
and experimental animal inhalation exposure studies examining a variety of endpoints (e.g.,
learning and memory; motor activity, habituation, and anxiety; neuropathology). For controlled
inhalation exposure studies (all chamber studies, including mechanistic studies), a separate
evaluation was conducted examining details of the exposure protocol (formaldehyde
administration and measurement (see Appendix A.5.1) that involved controlled formaldehyde
inhalation was evaluated. The accompanying tables in this section document the evaluation.
Studies are arranged alphabetically by first author within each table. The specific criteria for
evaluation are described below.
Human Observational Epidemiology Studies
Amyotrophic lateral sclerosis is a rare neurodegenerative disorder of the motor neurons
with an incidence in Western countries of 1-2 per 100,000 person-years (Ingre et al., 2015). Three
of the studies of ALS evaluated ALS mortality which was not considered to be a limitation. Because
the 5-year survival rate is low, mortality studies of ALS provide a good estimate for incidence of this
disease. Because the disease is rare, the precision of risk estimates reported by these studies is a
major limitation; the number of exposed cases for the case-control studies or total cases
ascertained for the cohort studies generally was small. Established risk factors that should be
considered as potential confounders are age, and sex. Smoking also has been associated with ALS in
multiple studies. Family history also is a risk factor but would not likely be associated with
formaldehyde exposure; therefore controlling for family history was not considered essential.
While potential misclassification of exposure was another limitation for all of the studies, this was a
particular concern for the general population studies, which collected exposure information using
questionnaires (Fang et al., 2009; Weiskopf et al., 2009) or job-exposure matrices based on industry
or occupation Roberts et al. (2015); fPeters etal.. 2017: Seals etal.. 20171. Fang et al. (2009) used a
more detailed evaluation of exposure level and duration based on a structured occupational
questionnaire and classification by industrial hygienists. Peters etal. f20171 and Seals etal. f20171
assigned individuals to exposure categories using the Nordic Occupational Cancer Study
job exposure matrix which contained formaldehyde concentration data specific to either Sweden or
Denmark; data on occupations over time were obtained from national censuses in Sweden (Peters
etal.. 2017) or the National Pension Fund in Demark (Seals etal.. 2017). Roberts et al. (2015) used
data from the National Longitudinal Study in the United States, which obtained information via a
survey on the most recent occupation at the time subjects were enrolled; information on later
occupations during follow-up was not captured.
This document is a draft for review purposes only and does not constitute Agency policy.
A-585 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 In addition to the general considerations for study evaluation, the observational and
2 controlled human exposure studies that assessed a battery of neurobehavioral tests were evaluated
3 with respect to the completeness and appropriateness of the battery of tests used, and the timing of
4 their administration with respect to exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
A-586 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-84. Evaluation of observational epidemiology studies of formaldehyde—neurological effects
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Amyotrophic Lc
teral Sclerosis (ALS)
Bellavia et al.
(2021).
(Denmark)
Population-
based nested
case-control
Cancer cases, 1982-2009,
from Seals et al. (2017)
with data for several
health factors and
environmental risk factors
previously linked with ALS.
Controls, 100 per case
matched on being alive on
index date for case
diagnosis, same birth year
and sex. Excluded
individuals with less than
5 years work experience.
Occupational histories
obtained from Danish
Pension Fund
databases. Used
NOCCA (Nordic
Occupational Cancer
Study)- Danish JEM
for periods 1960-74,
1975-84, and 1985
and after.
Formaldehyde
exposure metric was
ever/never exposed.
Anticipate exposure
misclassification and
large variation in
prevalence and
intensity of exposure
across individuals. In
men, correlations
between
formaldehyde, diesel
exhaust and solvents
were 0.22 and 0.41,
respectively (Phi
coefficients)
Danish National
Patient
Register,
discharge
diagnosis ICD-8
348.0 OR icd-10
G12.2. Case
definition was
1st diagnoses
on or after
1/1/1982 -
12/31/2009.
Evaluated
diabetes,
obesity,
physical/ stress
trauma, CVD
(1977-2009) and
lead, diesel
exhaust and
solvents
Selected joint
predictors and
interactions
using boosted
regression trees
and Logic
regression,
which were
included in a
logistic
regression
model adjusting
for age, SES, and
geography.
Model used a 3
year lag.
1086
incident
cancer
cases, 677
exposed;
111,507
controls
Amyotrophic lateral
sclerosis (incidence)
SB
IB
Cf Oth
Overall
Confidence
Medium
Uncertainty regarding
exposure assessment.
Adequacy of 3 year lag is
unknown.
Seals et al.
(2017)
(Denmark)
Registry-based case
identification using the
Danish National Patient
Occupational histories
obtained from Danish
Pension Fund
Danish National
Patient
Register,
Controls were
matched to
cases by age, sex
Conditional
logistic
regression
3650
incident
cases,
Amyotrophic lateral
sclerosis (incidence)
This document is a draft for review purposes only and does not constitute Agency policy.
A-587 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Population-
based case-
control
Register, 1982-2009 (3650
incident cases). Controls, 4
per case matched on sex,
age, and no ALS diagnosis
in Hospital Register as of
index date obtained from
Central Person Registry
(All Denmark residents
since 1968).
databases. Used
NOCCA (Nordic
Occupational Cancer
Study)- Danish JEM
for periods 1960-74,
1975-84, and 1985
and after. Inputs year
and industry code and
outputs prevalence of
exposure for each job
along with expected
exposure level (ppm)
in exposed. The JEM
has not been
validated to estimate
levels. Cumulative
expected exposure
calculated
(prevalence
multiplied by
expected level)
summed over jobs
and time (3 & 5 year
lags). Exposure
misclassifi cation
expected.
discharge
diagnosis ICD-8
348.0 OR icd-10
G12.2. Case
definition was
1st diagnoses
on or after
1/1/1982 -
12/31/2009.
and calendar
date. Assessed
SES (highest
attained, 5
groups based on
job title), marital
status and
residence. Other
covariates were
relative to 4th
year before
index year:
whether worked
on that year,
years worked
prior, hospital
admission, #
times admitted
prior, #
admissions, prior
diagnoses used
to construct
Charlson
Comorbidity
Index. No
information on
smoking status
adjusted for age,
sex, index date,
SES, marital
status and
residence. In
secondary
analyses
included other
work variables, #
hospital
diagnoses, plus
Charlson
Comorbidity
Index. Exposure
metrics were
dichotomous
(ever exposed
lagged 3 years),
quantiles, and
continuous
1068
exposed;
14,600
controls
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Uncertainty regarding
exposure assessment.
Adequacy of 3 year lag is
unknown.
(Fang et al.,
2009) (United
States)
General
population
(case-control)
Sequential ALS cases
recruited, 1993-1996,
from 2 major referral
centers in New England;
eligibility criteria cases &
controls: lived in New
England at least 50% of
Occupational history
by structured
questionnaire;
industry, occupation,
frequency and
duration; jobs held
before ALS diagnosis
Diagnoses by
board-certified
specialists in
motor neuron
disease using
World
Federation of
Adjusted for age,
sex, area of
residence,
smoking
(ever/never), &
education; no
additional
Unconditional
logistic
regression
models; linear
trend with
lifetime
exposure days,
109 ALS
cases
(n=20
exposed)
253
controls
Amyotrophic lateral
sclerosis (incidence)
SB
IB
Cf Oth
Overall
Confidence
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-588 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
year, mentally competent,
English speakers; 71% of
eligible cases participated;
controls by random
telephone screening,
frequency matched on
sex, age (3 groups), &
region; 76% of eligible
(256 of 270 completed
questionnaire).
or 2 years before
interview (controls);
formaldehyde-
exposed occupations
identified a priori by
industrial hygienist;
calculated life-time
hours of exposure
weighted by
probability in specific
jobs
Neurology El
Escorial criteria
workplace
exposures
associated with
ALS
probability, &
weighted
exposure
duration (4
categories);
effect
modification by
smoking;
missing
occupational
data for 2/111
cases & 3/256
controls
Uncertainty regarding
exposure assessment; small
number of exposed cases
Peters et al.
(2017)
(Sweden)
Nested case-
control study
All Swedish births (1901-
1970) and included in
1990 Swedish Population
and Household census,
N=5,763,437. Controls
randomly selected (5 per
case) from population
alive on date of diagnosis,
matched on birth year and
sex. 25,100 controls.
Occupational history
obtained from 1970,
1980, and 1990
census; included
occupations listed >
10 years prior to
index date;
occupational
exposures assessed
using Swedish version
of JEM (Nordic
Occupational Cancer
Study), prevalence
and level of exposure
at specific calendar
time. Exposure
metric for dose
response, prevalence
multiplied by annual
mean level for each
occupation at time of
Linkages to
National
Patient
Register,
primary or
secondary
diagnosis, ICD-9
335C or ICD-10
G12.2
(inpatient visits
1991-2010 and
outpatient
visits 2001-
2010); follow-
up to date of
first visit,
migration,
death, or
12/31/2010.
5,010 cases
Addressed age
and sex via
matching,
adjusted for
education and
evaluated 12 of
>20 agents
possibly
associated with
ALS. No
adjustment for
smoking status
although
restriction to
blue collar
workers and
farmers may
have partially
addressed
potential
confounding
Conditional
logistic
regression, OR
and 95% CI,
adjusted for
education and
other 11
chemicals;
restricted
analyses to
cases and
controls with at
least one
occupation
listed in any
census and to
blue-collar
workers or
farmers;
sensitivity
analysis
2,647
cases
(n=323
exposed),
13,378
controls
Amyotrophic lateral
sclerosis (incidence)
SB
IB
Cf Oth
Overall
Confidence
Medium
Uncertainty regarding
exposure assessment
This document is a draft for review purposes only and does not constitute Agency policy.
A-589 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
census (mg/m3),
averaged across all
censuses;
dichotomized at
median in controls
index
restricting to <
65 years at i
date, age of
retirement
(Pinkerton et
al.. 2013)
(United States)
Garment
workers
(cohort)
Cohort of garment
workers (N=ll,098)
exposed for > 3 months at
3 facilities (late 1950s to
early 1980s).
Monitoring in 1980s,
geometric mean 0.15
ppm (GSD 1.9 ppm),
constant levels across
departments and
facilities, year of first
exposure (42% before
1963), time since 1st
exposure (median
39.4 years) and
exposure duration
(median 3.3 years)
Vital status
ascertained
through 2008,
ICD-10 G12.2,
ICD-9 335.2,
ICD-8 348.0,
and ICD-7
356.1; ALS
mortality is a
good surrogate
for ALS
incidence
Adjusted for age,
calendar time,
sex, race; no
information on
smoking.
Mortality for
COPD and lung
cancer in cohort
was similar or
greater than
national rates
suggesting
possible bias
away from null.
Life table
analysis,
excluded missing
birth date (n-
55), deaths
(n=8), loss to
follow-up prior
to rate file begin
date (n=13);
SMRs and 95%
CI
N = 11,
022,
414,313
person-
years at
risk; 8 ALS
deaths
Amyotrophic lateral
sclerosis (mortality)
SB IB Cf Oth
Overall
Confidence
High
Small number of cases.
Confounding away from
null not of concern because
effect estimates were null.
(Roberts et al.,
2015) (United
States)
National
Longitudinal
Mortality
Study.
Occupational
(cohort)
Note: same
laboratory and
data handling
procedures as
794,541 men and 674,694
women (recruitment date
unclear, but study from
1973-2011) aged 25+ at
recruitment (national).
Follow-up time provided
by participants.
Internal comparison,
participation unlikely to
be influenced by
knowledge of exposure
and disease.
Self-reported at
enrollment based on
survey regarding last
or most recent job.
Exposure matrix
constructed by
industrial hygienists
at the National
Cancer Institute
based on methods in
Wang et al. (2009).
Metrics included
intensity and
probability of
ALS Mortality
(National Death
Index from
1979-2011) as
underlying
cause; ICD-9
code 335.3
(specific for
ALS) or ICD-10
code G12.2 (for
all motor
neuron
diseases, of
which ALS
Adjusted for
education,
race/ethnicity,
and income
(participants
tended to be
poorer, less
educated, and
less frequently
non-Hispanic
white. One
sensitivity
analysis among
high probability,
Data handling
and analysis as
in Weisskopf et
al. (2009)
HRs provided for
each exposure
intensity and
probability for
men and women
separately.
Additional
sensitivity
analyses to
evaluate validity
472 deaths
in men
(100
exposed);
285 deaths
in women
(61
exposed)
Amyotrophic lateral
sclerosis (mortality)
SB IB Cf Oth
Overall
Confidence
Medium
Uncertainty regarding
exposure assessment,
including the influence of
duration, particularly in
light of the use of a one-
time survey at enrollment;
very small number of
exposed cases (n=2 in jobs
This document is a draft for review purposes only and does not constitute Agency policy.
A-590 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Weisskopf et
al. (2009)
exposure.
Information on other
exposures not
collected/reported.
comprises the
overwhelming
majority)
high exposure
group (all funeral
directors)
included
adjustment for
smoking and
military service.
of exposure and
outcome
assignments and
selection bias,
included follow
up restricted to
75 years or
excluding first 5
years, age
restricted to 35-
75 or 50-75
years at
enrollment, or
restricted to
those employed
at enrollment.
Did not provide
or incorporate
any data on
duration.
with high probability and
intensity of formaldehyde
exposure)
Weisskopf et
al. (2009)
(United States)
American
Cancer Society
Cancer
Prevention
Study II.
General
population
(cohort)
987,229 (414,493 men,
572,736 women) enrolled
in 1982. National
recruitment; no major
illness at baseline, not
missing age or sex data.
Follow-up from 1989
through 2004.
Internal comparison,
participation unlikely to
be influenced by
knowledge of exposure
and disease.
Self-reported, mailed
questionnaire in
1982. Current or past
regular exposure to
formaldehyde and
duration (years) (not
specified, but likely in
occupational
settings). Data on 10
other types of
chemicals and X-ray
exposure also
collected.
Mortality
(National Death
Index),
underlying or
contributing
cause; ICD-9
(1989-1998)
code 335.3 or
ICD-10 (1999-
2004) G12.2.
(ALS represents
> 98% of these
categories)
Adjusted for age,
sex, smoking,
military service,
education,
alcohol,
occupation
(farmer, lab
technician,
machine
assembler,
programmer),
vitamin E use,
and the other
chemical (and
Cox proportional
hazards
modeling,
analyzed with
and without
approximately
1/3 who
reported
exposure but did
not provide
duration data
(i.e., less likely
to be truly
exposed).
1,156 ALS
deaths (36
exposed)
Amyotrophic lateral
sclerosis (mortality)
Overall
SB IB Cf
Oth
Confidence
H
Medium
Uncertainty regarding
exposure assessment; small
number of exposed cases
This document is a draft for review purposes only and does not constitute Agency policy.
A-591 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
X-rays)
exposures
assessed at
baseline.
Neurobehavioral tests and olfactory detection
Broder et al.,
Homes with UFFI
2-day samples in
Sense of smell
Detailed
Prevalence by
1,726 from
1988b
insulation, within 60 miles
homes, 5 hr/day
threshold for
demographic
group and Chi-
UFFI
(Canada;
of Toronto. 4,400 of 8,200
Median ppm
pyridine; three
data collected
square test.
homes,
Toronto)
agreed to be contacted;
Control 0.031
control bottles
720 from
Residences
95% participated.
UFFI 0.038
(mineral oil
control
(household
Control homes randomly
only) plus 3
homes
survey)
selected from streets
bottles with
Additional
adjacent to UFFI homes,
0.00005, 0.008,
reference:
20% participated.
and 0.012%
Broder et al.,
Some demographic and
pyridine.
1988a
symptom data allowed
comparison with
nonparticipants; similar
neighborhood,
demographics.
Replicate tests
conducted.
Variability and
stability of test
kits assessed.
Participant
blinded.
(Kilburn et al..
Recruited from attendees
Self-reported hours
Neuro-
Adjusted for age,
Multiple
305
1989; Kilburn
(female) at annual
per day (based on
behavioral test
number of cover
regression.
et al., 1987)
histology technician
detection of odor)
battery
slipped slides
Coefficients and
(United States)
conferences, 1982 and
(memory,
(for other
designation if p
Workers:
1983. Participation rate
cognition,
solvent
< 0.05 (no
histology
not reported.
spatial relation
exposure),
standard errors)
technicians
integration,
duration of
(survey)
dexterity,
conceptual
motor speed,
balance,
smoking
Sense of smell
SB IB Cf Oth
Overall
Confidence
Not
informative
No appreciable difference
in median exposure
between groups
Neurobehavioral tests
SB IB
Cf Oth
Overall
Confidence
¦
Low
Potential selection bias
(could be influenced by
perceived exposure and
effects), limited detail
presented in results
This document is a draft for review purposes only and does not constitute Agency policy.
A-592 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
reaction time);
1 hour
(Kilburn and
Warshaw,
1992) (United
States)
Workers:
histology
technicians
(survey,
multiple time
points)
Recruited from attendees
(female) at annual
histology technician
conferences, 1982,1983,
1985, 1986, 1989.
Participation rates not
reported.
No information on
intensity or frequency
of exposure
Neuro-
behavioral test
battery
(memory,
cognition,
pattern
recognition,
dexterity,
decision
making, motor
speed,
balance); 2-3
hours
Considered age,
sex, number of
cover slipped
slides (for other
solvent
exposure), years
of exposure
For analysis of
single (first) test
per subject
(n=350),
reported as "not
statistically
significant." For
longitudinal
analysis (n=19),
no decline in
performance
noted
(formaldehyde
exposure not
explicitly
analyzed).
19 with 4
tests, 299
with 2 or 3
tests, 350
with one
test
Neurobehavioral tests
SB
IB Cf Oth
Overall
Confidence
1
¦
LOW
1
Potential selection bias,
limited detail presented in
results. Longitudinal
analysis limited by sample
size and did not specifically
address formaldehyde
exposure
Kilburn, 2000
(United States,
6 states).
Home or office
exposure
(survey)
Exposed (e.g., new mobile
homes or renovated
offices), experienced
"adverse effects almost
daily"; referent group
randomly selected from
voter registration rolls in 4
cities (location and
participation rate not
reported).
No exposure
measures.
Neuro-
behavioral test
battery
Frequency
matched by age
and education
Mean ± SD
percent
prediction
20
exposed,
202
referents
Neurobehavioral tests
SB IB
a
Oth
Overall
Confidence
Not
informative
¦
Likely selection of exposed
based on symptoms; no
exposure measures, limited
covariate data.
Schenker et
al., 1982
(United States)
People self-referred to
occupational and
environmental health
clinic regarding health
Measured in 4 homes
(protocol not
described), ranged
Neurobehavior
al battery
Not addressed
Prevalence
18 adults,
6 children
(from 6
homes)
Neurobehavioral tests
This document is a draft for review purposes only and does not constitute Agency policy.
A-593 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration of
participant selection
and comparability
Exposure measure
and range
Outcome
measure
Consideration
of likely
confounding
Analysis and
completeness
of results
Size
Confidence
Residences
(survey)
effects of formaldehyde
insulation. No comparison
group.
from 0.03 to 0.23
ppm
SB IB Cf Oth
Overall
Confidence
Not
informative
IH
Likely selection of exposed
based on symptoms;
limited exposure measures,
no comparison group
1 Controlled Exposure Studies in Humans
2 Controlled human exposure studies were evaluated using a combination of criteria relevant to experimental animal studies
3 (below) and criteria specific to studies in observational epidemiology studies.
This document is a draft for review purposes only and does not constitute Agency policy.
A-594 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-85. Evaluation of human controlled exposure studies of formaldehyde - nervous system effects
Reference,
setting, and
design
Exposure assessment (quality
descriptor and exposures)
Outcome
classification
Consideration of possible bias
(randomized exposure order,
blinding to exposure) and
confounding
Analysis and
completeness
of results
Size
Confidence
(Andersen
and
Molhave,
Chamber type and analytical
concentrations not provided; testing
during exposure (distractibility likely
contributes)
4d of exposure
Endpoints limited:
sparse methods on
conduct of partial
neurobehavioral
test battery
Exposure order by Latin square
design; blinding not indicated
Comparisons
appear to
represent
pooled sexes;
results data NR
ID
1
II
C
Low
1983)
(Bach et al.,
1990)
Test article not defined (inferred from
Andersen & Molhave, 1983)°;
testing during exposure
(distractibility likely contributes);
acute (5.5h) exposure
Endpoints limited:
sparse methods on
conduct of partial
neurobehavioral tes
battery
Occupation exposure group and
controls from population registry
(attempted matching by age,
education, smoking prevalence but
workers had higher smoking and
lower education; details not
reported); Exposure order by
balanced Latin square design;
blinding not indicated
Results
reporting
incomplete &
difficult to
decipher
n=61
males only
Low
(Lang et al.,
2008)
Analytical concentrations achieved
measured but not reported; testing
immediately after exposure; study
focus on irritation; no indication of
acclimation; recovery not examined
(reaction time); lOd of exposure
Endpoints limited:
decision reaction
time
Exposure order randomly assigned
double blinded
Data= combined
sexes; high
variability in
reaction time
data
n=21
-20%
attrition
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-595 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Studies in Animals: Toxicological Studies
Hazard ID evaluations of chamber studies only encompass studies reporting results
following in vivo inhalation exposures. Noninhalation exposures are expected to involve significant
distribution of formaldehyde beyond the portal of entry (which is not observed to an appreciable
extent following inhalation exposure).
Evaluation of experimental studies
As described in Appendix A.5.I., experimental animal studies were assigned the following
confidence ratings: high, medium, or low confidence, and not informative based on expert judgement
of each study's experimental details related to predefined criteria within five study feature
categories. Not informative studies were designated based on the interpretation that the observed
effect(s) are expected to have been driven by factors other than exposure to inhaled formaldehyde,
or that the study did not provide a sufficient level of detail to evaluate the key methodological
features or the nervous system-specific results. Due to the issues identified, the not informative
experiments are not discussed in the Toxicological Review.
In addition to the general criteria discussed in Appendix A.5.I., considerations specific to
the evaluation of potential nervous system effects were also evaluated. Due to the known
neurotoxicity hazard of methanol, studies failing to use an appropriate test article were
automatically assigned low confidence and, in an effort to avoid confusion with methanol's effects, if
they evaluated high exposure levels (defined here as relying only on exposures >10 mg/m3) they
were deemed to be not informative. Additional criteria included: consideration of the potential
influence of irritation or changes in olfaction on behavioral measures (e.g., exposure during
behavioral training was considered a limitation; a preference was given to behavioral studies with a
period of latency between exposure and endpoint testing of 24 hours, or 2 hours at a minimum);
blinding of the outcome assessors was preferred for subjective measures (e.g., slide evaluation;
behavioral observations; etc.), although this was not necessarily considered a limitation for
automated measures; a sample size of n=10/group was preferred (n=4 at a minimum); methods
include a description of and a preference for endpoint evaluation procedures that are sensitive and
specific for the detection of potential nervous system effects (see Table A-86 for additional details).
Although studies with a longer exposure duration were considered to be most relevant to
interpreting the lifetime neurotoxicity hazard of inhaled formaldehyde, nervous system effects
studies of short term or even acute duration were not automatically considered to be less
informative (i.e., exposure duration < 28 days was indicated as a minor limitation). This is
somewhat in contrast to the interpretation of animal studies in other sections (e.g., respiratory tract
pathology), and this reflects an understanding that neurotoxic effects from very brief exposures can
oftentimes represent important health concerns. Additional considerations that might influence
the interpretation of the usefulness of the studies during the hazard synthesis are noted, including
limitations such as a short exposure duration or the use of only one test concentration or
This document is a draft for review purposes only and does not constitute Agency policy.
A-596 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information for Formaldehyde—Inhalation
concentration that are all too high or too low to provide a spectrum of the possible effects, as well
as study strengths such as very large sample sizes or particularly robust endpoint protocols;
however, this information typically did not affect the study evaluation decisions.
If the conduct of the experimental feature is considered to pose a substantial limitation that
is likely to influence the study results, the cell is shaded gray; a "+" is used if potential issues were
identified, but these are not expected to have a substantial influence on the interpretation of the
experimental results; and a "++" denotes experimental features without limitations that are
expected to influence the study results. Specific study details (or lack thereof] which highlight a
limitation or uncertainty in answering each of the experimental feature criteria are noted in the
cells. For those experimental features identified as having a substantial limitation likely to
influence the study results, the relevant study details leading to this decision are bolded. Studies
are organized according to the type of endpoint(s) evaluated, and then listed alphabetically.
This document is a draft for review purposes only and does not constitute Agency policy.
A-597 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-86. Evaluation of controlled inhalation exposure studies examining nervous system in animals
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
Exposure Qualitv
Test Subjects
Studv Design
Endooint Evaluation
Data Considerations
& Statistical
Analvses
Overall Confidence
Rating Regarding the
Use for Hazard ID
[Main limitations]
Expert judgement
based on conclusions
from evaluation of
the 5 experimental
feature categories
Criteria relevant
to evaluating the
experimental
details within
each
experimental
feature category
Exposure quality
evaluations (see
Appendix A.5.1) are
summarized below;
robust;
adequate; and
shaded box: poor;
relevance of the
tested exposure
levels is discussed in
the hazard synthesis
The species, sex,
strain, and age are
appropriate for the
endpoint(s); sample
size provides
reasonable power to
assess the
endpoint(s); overt
systemic toxicity is
absent or not
expected, or it is
accounted for; group
allocations can be
inferred as
appropriate
A study focus was nervous
system effects; the
exposure regimen is
informative for the tested
endpoint; latency from
exposure to testing
reduces the potential for
irritation-driven responses
Note: No guideline or GLP
studies were identified3
The protocols used to
assess the nervous
system effects are
sensitive for detecting
an effect, complete,
discriminating (i.e.,
specific for the
response in question),
and biologically sound;
experimenter and
sampling bias
minimized
Statistical methods,
group comparisons,
and data
presentation
(including variability)
are complete,
appropriate, and
discerning; selective
reporting bias
avoided
Odorant or Irritant Detection/Effects
(Apfelbach
and Weiler,
1991)
+
Chamber type not
specified
+
N = 5 (exposed) or 10
(controls); males only
Testing during exposure;
controls not air-exposed in
exposure chamber;
possible continuous
exposure
Note: 130d exposure
Training started 30d
after exposures began
(not clear if training
ability prior to
endpoint testing was
affected)
++
Not informative
[Tested during
exposure; missing
controls; training
during exposure]
(Wood and
Coleman,
1995)
++
+
N=8; males only
Testing during exposure;
each animal served as its
own control (multiple
exposures/animal); acute
++
Note: endpoint is not
adverse (irritant
detection)
++
Note: statistical
comparisons not
possible
N/A*
Olfactory
detection/irritation
response
This document is a draft for review purposes only and does not constitute Agency policy.
A-598 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
exposure (60 seconds
on/off for ~lhr)
[Tested during acute
exposure]
Cursory Examinations in Long-Term Toxicity & Carcinogenicity Studies
(Appelman et
al.. 1988)
++
+
N > 10; males only
Behaviors tested during
exposure; study focus not
nervous system-specific;
lyr study
+
Endpoints limited:
cursory cage-side
observations, gross
pathology, & weight
Results data NR;
behavioral effects
not quantified
**
[Tested during
exposure; study
focus not CNS; data
NR]
(Coon et al.,
1970)
+
Multiple species
exposed
simultaneously
+
N=2 (i.e., dogs) to 15
(e.g., rats); age & sex
ratio/group not given
Note: multiple species
tested
Behaviors tested during
exposure; study focus on
overt toxicity and
inflammation; 90d study
Endpoints limited:
cursory cage-side
observations & brain
sections "retained"
(not clear if examined)
Results data NR;
behavioral effects
not quantified; one
death noted, but no
cause provided
Not informative
[Tested during
exposure; limited
endpoints; data NR]
(DHGC. 2010)
Formalin (high
concentration:
methanol may drive
responses)
N = 3-6
Behaviors tested during
exposure; acute exposure
Endpoints limited:
cursory observations
of behavior during
exposure
Effects not
quantified
Not informative
[High formalin levels;
etc.]
(Kerns et al.,
1983)b
++
++
N= 10
Behaviors appear to have
been tested immediately
after exposure; study
focus on carcinogenicity
Note: based on a 2yr GLP-
compliant study (CUT,
1982); this was not noted
in article
+
Endpoints limited:
simple neurofunctional
observations & gross
pathology; methods
provided in original CUT
(1982) study indicate
lack of observer
blinding
Results data NR in
published article;
latency NR; data in
original CUT (1982)
study is qualitative
(normal vs.
abnormal) & is
pooled across test
battery endpoints
**
[Tested immediately
after exposure; study
focus not CNS; data
NR]
(Maronpot et
al.. 1986)
Formalin
++
N= 10
Behaviors tested during
exposure; study design not
nervous system-specific;
13wk study
+
Endpoints limited:
cursory cage-side
observations & gross
pathology
Results data NR;
behavioral effects
not quantified
Not informative
[Formalin; tested
during exposure;
This document is a draft for review purposes only and does not constitute Agency policy.
A-599 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
study focus not CNS;
etc.]
(Morgan et al.,
1986a)
+
Analytical
concentrations not
provided
N = 3-6; males only
Behaviors tested during
exposure; study design not
nervous system-specific;
acute exposure
Endpoints limited:
cursory observations
of distress during
exposure
No quantified
neurological effects
Not informative
[Formalin; small
sample size; tested
during exposure;
etc.]
{Tobe, 1985,
3574}
Formalin (Note:
methanol control
group included in the
chronic study)
+
N = 3-20 (depending
on the experiment,
endpoint & exposure
group); males only
Behaviors tested during
exposure; study design not
nervous system-specific
Note: studies of variable
duration (up to 28 mo)
+
Endpoints limited:
cursory cage-side
observations; gross
pathology, brain wt.
weight also performed
in 28-month study
Results details NR
for many
experiments &
animals; behavioral
effects not
quantified; multiple
dead animals could
not be examined for
comparisons due to
decomposition
**
[Formalin: controlled
for some endpoints;
tested during
exposure; data NR]
(Woutersen et
al.. 1987)
+
Animals were housed
in the inhalation
chambers
++
N=40
Behaviors tested during
exposure; study design not
nervous system-specific
Note: 13wk study
+
Endpoints limited:
cursory cage-side
observations, brain wt.
Results data NR;
behavioral effects
not quantified
**
[Tested during
exposure; data NR]
Neuropathology
(Asian et al.,
2006)
++
N= 3 litters (5 pups);
males only; dam
health during
lactation & pup
health not presented
Note: possible subset
of Songur (Songur et
al., 2003) study;
same animals as
+
Unclear if potential litter
bias was corrected
(although randomized
treatment groups); dams
seemed to be co-exposed
with pups from PND 1-14
Note: 30d of exposure
++
Note: regional or
hemisphere volume
changes not verified by
immunostaining,
leaving interpretations
unclear; sensitive
stereology methods;
random sampling
indicated
As presented, data
do not account for
potential litter
effects (pup means
presented)
Medium
[Small sample size;
potential for litter
effects]
This document is a draft for review purposes only and does not constitute Agency policy.
A-600 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
Sarsilmaz et al. (2007)
study6
(Bian et al.,
2012)
Formalin (high
concentration:
methanol may drive
responses)
N= 3/endpoint/time
point; males only;
mild toxicity:
decreased food
intake (effect not
quantified)
Controls not air-exposed in
exposure chamber; all
groups had anesthesia &
antibiotic injections;
exposures = 1 hr/day
Note: 90d exposure; single
exposure level
+
Number of
slides/animal not
provided; relatively
insensitive method for
cell count
quantification
Note: blinding & other
methods appropriate
++
Not informative
[High formalin levels;
etc.]
(Liu et al..
2010)
Formalin (high
concentration:
methanol may drive
effects)/static
chamber
+
Group size for staining
not clear; males only;
groups determined by
preexposure probe
trial performance
+
Exposures only 30 min
twice daily; 28d
Potential sampling
bias: details on
blinding,
slides/animal, etc. not
provided; imaging
specifics not provided
and qualitative only
+
Hippocampal Nissl
staining not
quantified
Not informative
[High formalin levels;
etc.]
(Mei et al.,
2016)
Formalin
+
N = 8; males only
+
No comparisons to
chamber or air exposure
alone; 8hr/d for 7
consecutive days
Potential sampling
bias: details on
blinding,
slides/animal, etc. not
provided; qualitative
only
No quantitative
results (e.g., counts;
severity scores; etc.)
Not informative
[formalin; potential
sampling bias; no
results
quantification]
(Pitten et al.,
2000)
Formalin/static
chamber
+
N = 5-8
Note: no changes in
body weight were
observed
+
Exposures only 10 min/d
for 90 days
Potential sampling
bias: details on
blinding,
slides/animal, etc. not
provided; qualitative
only
Results data NR
**
[Formalin; potential
sampling bias; data
NR]
(Sarsilmaz et
al.. 2007)
++
N= 3 litters (5 pups);
dam health during
lactation & pup
+
Unclear if potential litter
bias was corrected
++
Note: regional or
hemisphere volume
As presented, data
do not account for
potential litter
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-601 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
health not presented;
males only0
Note: possible subset
of Songur (Songur et
al., 2003) study;
same animals as Asian
et al. (2006) study0
(although randomized
treatment groups); dams
seemed to be co-exposed
with pups from PND 1-14;
30d of exposure
changes not verified by
immunostaining,
leaving interpretations
unclear; sensitive
stereology methods;
random sampling
indicated
effects (pup means
presented)
[Small sample size;
potential for litter
effects]
(Songur et al.,
2003)
+
Analytical
concentrations not
provided
N= 6 pups (likely 3
litters); mild toxicity
(body weight changes
at 30 & 60d, but not
90dd); males only
+
Unclear if potential litter
bias corrected (& not
indicated as randomized);
30d of exposure
Cell counting methods
do not detail how
many slides/animal
were examined (may
be a single slide)
as presented, data
do not account for
potential litter
effects (pup means
presented)
Low
[Small sample size;
potential for
sampling bias and
litter effects]
(Wang et al.,
2014)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
++
N =12 males/group
Note: no changes in
body weight were
observed
++
2hr/d exposure for
subchronic (90 days)
Relative, but not
absolute (preferred),
brain weights were
reported; number of
H&E samples NR
Note: both insensitive
++
Not Informative
[Mixture exposure
only; etc.]
Neural Sensitization-Related Responses
(Sheveleva,
1971)
(translation)
Test article not
defined (assumed to
be formalin)
+
Use of mongrel white
rats; N= 7 dams or 6
offspring/sex
evaluated from 6
litters, so assumed 1
pup/sex/litter
examined, but not
specified; unclear why
7 dams vs. 6 offspring
+
Latency between dam
exposure and testing not
provided: unclear if reflex
bradypnea can influence
these measures (e.g.,
reduced respiration leading
to transiently reduced 02
content in muscle tissue,
causing reduced
excitability); 4hr/d
exposures from GDI-19
"Neuromuscular
excitability" protocol
specifics not provided
(e.g., blinding; how
assessed)
+
Statistical methods
used were not
specified; data
appear to account
for possible litter
effects, but not
clearly described
Low
[Formalin; endpoint
methods NR]
This document is a draft for review purposes only and does not constitute Agency policy.
A-602 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
(Sorg et al.,
1996)
Formalin (high
concentration:
methanol may drive
responses)
+
N > 4; females only
Potential high
concentration irritation-
related responses (that
may affect odor
discrimination in tasks
involving exploration)
were not measured;
exposure lhr/d for 7d;
Note: single exposure level
+
Overall plus maze
activity not provided;
Note: questionable
human relevance of
rodent sensitization
responses
+
Groups divided into
high & low
respondersfor
presentation of most
endpoints &
statistical
comparisons;
statistical
comparisons NR for
1-month recovery
data
Not informative
[High formalin levels;
etc.]
(Sorg et al.,
1998)
+
Chamber type not
provided; declining
HCHO exposures
across days
+
N= 15-24; females
only
+
Imprecise timing of
assessment; unclear effect
of prior cocaine
exposure/handling on
nociception (assumed to be
minimal)
Note: 1 or 4wk exposure;
single exposure level
Experimenter blinding
not indicated; methods
for measuring vertical
activity NR in cited
reference
Note: questionable
human relevance
++
Medium
[Blinding NR; limited
methods description]
Note: relevance of
inescapable stress
unclear
(Sorg and
Hochstatter,
1999)
+
Chamber type and
analytical
concentrations not
provided
+
N = 4; females only
(conditioned fear) OR
N= 8; males only
(approach/avoidance)
Possible effects on
olfactory detection of
conditioned odors by
HCHO nasal effects;
Approach/avoidance
tested during exposure to
formalin vapors
Note: 4 wk exposure; single
exposure level
++
Note: questionable
human relevance of
rodent sensitization
responses
Effects without
cocaine NR: (unclear
influence of prior
cocaine exposure in
conditioned fear
responses)
Low
[Unclear influence of
changes in olfactory
detection or prior
cocaine exposure]
(Sorg et al.,
2001b)
+
Chamber type and
analytical
00
1
t s
Testing during exposure;
exposures < 4wk
Note: single exposure level
+
Methods for measuring
vertical activity NR in
++
Low
This document is a draft for review purposes only and does not constitute Agency policy.
A-603 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
concentrations not
provided
cited reference (but
automated using
photocell counts)
[Tested during
exposure; limited
methods reporting]
(Sorg et al.,
2002)
Formalin (likely high
concentration- not
quantified: methanol
may drive
responses); HCHO
levels NR
+
N = 6-12
Formalin used as an
aversive stimulus- results
more specific to cocaine;
behaviors evaluated
coincident with exposures;
acute exposure
Tests involve odor
detection & irritation-
specific responses:
could confound results
Note: questionable
human relevance
Specific effects of
formaldehyde alone
on behaviors NR;
some data presented
with groups divided
into high & low
responders for
statistical
comparisons
Not informative
[High formalin levels;
etc.]
(Sorg et al.,
2004)
+
Chamber type not
specified
00
1
t s
Possible effect on
olfactory detection of
conditioned odor by HCHO
nasal effects; context
testing prior to
conditioned fear tests may
cause order effects
Note: single exposure level;
4wk exposure
+
Possible contribution of
change in footshock
sensitivity not
examined
Note: questionable
human relevance of
rodent sensitization
responses
++
Low
[Unclear influence of
changes in olfactory
detection]
(Usanmaz et
al.. 2002)
++
+
N = 6; unexplained
overt toxicity (body
weight decrease) with
multiple exposures
Observations immediately
after exposure; acute (3hr)
or short-term (l-3wk)
exposure
Observations not
blinded; 5 min test
duration; peripheral vs.
central square
crossings not
measured, limiting
interpretability
++
Low
[Tested immediately
after exposure; no
blinding]
Motor Activity, Habituation, and Anxiety (& aggression)
(Boja et al.,
1985)®
+
Analytical
concentrations not
provided
+
N = 8; males only
Behaviors tested during
exposure; acute exposure
(3hr/d for l-2d); timing of
exposures (9-12pm vs. 12-
Appropriateness of
protocol for adult
animals is
questionable (methods
+
Statistical
comparisons to air-
only exposure groups
Low
[Tested immediately
after acute exposure;
This document is a draft for review purposes only and does not constitute Agency policy.
A-604 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
3pm) may not have been
same across groups
Note: single exposure level
designed for pups);
"active" vs.
"nonactive" endpoint
readout is nonspecific
NRfor all treatment
groups; higher
exposure groups
data NR and text
suggests results are
somewhat
inconsistent
endpoint methods
questionable]
(Katsnelson et
al.. 2013)
Test article not
defined (assumed to
be formalin; high
concentration:
methanol may drive
effects)
++
N= 12-15
females/group
Testing indicated as
immediately after
exposure;
Note: subchronic (10 wk)
exposure
Protocols not specified,
although hole board
test methods assumed
to be conducted in a
standard manner;
blinding not indicated
++
Not informative
[High levels of test
article assumed to be
formalin; irritation
effects likely]
(Li et al.. 2016)
Formalin; static
chambers
+
N = 15 (inferred);
males only
+
Testing began ~2 h
postexposure
Note: exposure 2 h/day for
Id
Blinding not indicated
for all tests except
forced swim: of
particular concern for
nonautomated novel
object testing; unclear
impact of multiple
tests in same animals
(chosen test order may
reduce impact); % open
time in EPM does not
include % closed time;
note: slight body
weight loss 2.46 mg/m3
++
Low
[Formalin; endpoint
evaluations fail to
control for several
important
variables]
(Liu et al..
2009a)
Formalin (high
concentration:
methanol may drive
effects)/static
chamber
+
N = 8; males only
+
14 d exposure
Note: tested >24hr after
exposure;
Spontaneous
locomotor activity was
assessed subsequent
to aggression tests,
which may influence
anxiety-related
++
Not informative
[High formalin levels;
etc.]
This document is a draft for review purposes only and does not constitute Agency policy.
A-605 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
responses; blinding
not indicated
(Malek et al.,
2003a)
Formalin
++
N= 15/sex
+
2 and 26 hr postexposure;
acute: 2hr
+
3 min test duration;
manual scoring
(blinded); peripheral vs.
central square
crossings not
quantified, limiting
interpretability
+
Assuming data is SE,
some statistical
significance calls are
questionable;
variability unclear:
SE reported is higher
than SD for same
parameters in 2003b
Low
[Formalin]
(Malek et al.,
2003b)
Formalin
++
N= 10/sex
+
2hr postexposure; acute:
2hr
+
3 min test duration;
manual scoring
(blinded); peripheral vs.
central square
crossings not
quantified, limiting
interpretability
++
Low
[Formalin]
(Malek et al.,
2004)
Formalin
+
N = 20; m ales only
+
2 and 26 hr postexposure;
acute; 2hr
+
3 min test duration;
manual scoring
(blinded)
++
Low
[Formalin]
(Senichenkova
. 1991a)
(translation)
Test article not
defined (assumed to
be formalin)
Sex, N, & strain NR;
could not be
evaluated due to lack
of reporting
+
Unclear if litter bias
corrected
Note: 4hr/d exposures
from GDI-19; single
exposure level
Open field protocol
specifics not provided
(e.g., blinding; manual
vs. automated
assessment of activity)
+
Statistical methods
NR
Not informative
[Test article assumed
to be formalin; test
animal and endpoint
protocol details
NR]
(Sheveleva,
1971)
(translation)
Test article not
defined (assumed to
be formalin)
+
Mongrel white rats;
N=6 offspring/sex
evaluated from 6
++
4hr/d exposures from GD1-
19
"Spontaneous
mobility" protocol
specifics not provided
(e.g., blinding; manual
+
Statistical methods
NR
Low
[Test article assumed
to be formalin;
This document is a draft for review purposes only and does not constitute Agency policy.
A-606 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
litters, so assumed 1
pup/sex/litter
examined, but this
was NR
vs. automated
assessment of activity)
missing endpoint
protocol details]
(Sorg et al.,
1998)
+
Chamber type not
provided; declining
HCHO exposures
across days
+
N= 15-24; females
only
+
Imprecise timing of
assessment & unclear
effect of prior exposure to
cocaine/handling on plus
maze endpoints (assumed
to be significant)
Note: 1 or 4wk exposures;
single exposure level
Experimenter blinding
not indicated (note:
activity measures
automated); overall
plus maze activity not
provided; unclear
impact of saline
injection, handling;
methods for measuring
vertical activity NR in
cited reference
++
Activity: Medium
[Blinding NR; limited
methods description;
unclear impact of
prior manipulations]
Plus maze: Low
[Blinding NR; limited
methods description;
overall activity NR;
likely impact of prior
testing]
(Sorg et al.,
2001b)
+
Chamber type and
analytical
concentrations not
provided
+
N = 6; males only
+
No EEG/EMG sham
controls and influence of
37% formalin irritation
responses NR; exposures <
4wk
Note: single exposure level
No preformaldehyde
sleep measures; sleep
pattern methods NR
Note: questionable
adversity of endpoints
++
Low
[limited methods
reporting;
preformaldehyde
comparisons NR]
Note: questionable
adversity
(Usanmaz et
al.. 2002)
++
+
Unexplained overt
toxicity (body weight
decrease) with
multiple exposures; N
= 6
Observations immediately
after exposure; acute (3hr)
or short-term (l-3wk)
exposures
Observations not
blinded; 5 min test
duration; peripheral vs.
central square
crossings not
measured, limiting
interpretability
++
Low
[Tested immediately
after exposure; lack
of blinding]
Learning and Memory
(Chonglei et
al.. 2012)
Mixture (formalin,
benzene, toluene
+
N= 5 males/group
+
Path length or similar
NR (contribution of
motor effects not
++
Not informative
This document is a draft for review purposes only and does not constitute Agency policy.
A-607 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
and xylene)/static
chamber
Testing 30 minutes after
exposure; 2hr/d exposure
for short term (10 days)
tested); visual cues
NR; no blinding
indicated
[Mixture exposure;
endpoint protocol
deficiencies]
(Liao et al.,
2010)
(translation)
Formalin/static
chamber
N=8: pooled sexes
(/V=4/sex); overt
toxicity during
exposure (e.g.,
listlessness; up to
=30% decreased body
weight gain), most
likely from poor
exposure quality, as
only 0.5mg/m3 HCHO
Latency not provided
(assumed that
observations made
immediately after
exposure); no indication of
correction for possible
litter bias
Note: exposures 2hr/d for
28d
Path length or similar
NR (contribution of
motor effects not
tested); pool
temperature, pool
diameter, & platform
size NR; recovery time
between escape
latency trials not
indicated; no blinding
indicated
+
Data= combined
sexes (test often
displays sex
differences)
Not informative
[Formalin; overt
toxicity; endpoint
protocol deficiencies;
etc.]
(Liu et al..
2010)
Formalin (high
concentration:
methanol may drive
effects)/static
chamber
+
Males only; treatment
groups determined by
performance in
preexposure probe
trials, but unclear
exactly how groups
were matched; Note:
W=8-ll
+
Latency for all assessed
time points unclear, but
appears that most had
>24h habituation period
between exposure and
training/testing; exposures
only 30 min twice daily;
28d exposure
++
Note: probe trials
preexposure were
comparable; cued trials
conducted to rule out
HCHO effects on vision
++
Not informative
[High formalin levels;
etc.]
(Lu et al.,
2008b)
Unspecified wood
(possible co-
exposures not
tested)
+
N = 5; males only
Training behaviors
assessed 30 min
postexposure and possible
indirect effects of irritation
on training may influence
performance in the probe
trial test; 7d exposure
+
Path length or similar
NR (contribution of
motor effects not
tested); no blinding
indicated
+
Comparisons across
treatment groups NR
for probe trial test
Low
[Likely mixture
exposure; possible
impact of irritation]
(Mei et al.,
2016)
Formalin
+
N = 8; males only
+
No comparisons to
chamber or air exposure
Path length or similar
NR (contribution of
motor effects not
++
Low
[formalin; endpoint
protocol reporting
This document is a draft for review purposes only and does not constitute Agency policy.
A-608 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
alone; testing 3 hr after
exposure during training;
Note: 8hr/d for 7
consecutive days
tested); pool
temperature, pool
diameter, start
positions & platform
size NR; no blinding
indicated (of concern,
as not automated;
note: cited references
did not contain these
details)
deficiencies; lack of
blinding]
(Malek et al.,
2003c)
Formalin/static
chamber
++
N= 15/sex/group; no
changes in body
weight were observed
+
Latency 2hr postexposure;
exposures for 2hr/dfor 10
days
Motor effects appear
to drive some
responses & were not
tested (path length or
similar NR); possible
influence of changes in
olfaction and/or vision
not tested; blinding
not indicated
+
No A NOVA or trend
tests performed
across the 4 groups
(only pair-wise tests)
Low
[Formalin; endpoint
protocol deficiencies;
no blinding]
(Pitten et al.,
2000)
Formalin/static
chamber
+
N = 5-8
Note: no changes in
body weight were
observed
+
22 hr postexposure;
exposures only 10 min/d
Note: 90 d exposure
+
Possible influence of
changes in olfaction
and/or vision not
tested; path length or
similar NR
+
Data= combined
sexes (test often
displays sex
differences)
Low
[Formalin]
(Wang et al.,
2014)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
+
N = 6 males/group
Note: no changes in
body weight were
observed
+
Testing 30 minutes after
exposure; Note: 2hr/d
exposure for 49-90 days
Path length or similar
NR (contribution of
motor effects not
tested); visual cues
NR; no blinding
indicated
++
Not informative
[Mixture exposure;
endpoint protocol
deficiencies]
This document is a draft for review purposes only and does not constitute Agency policy.
A-609 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
Nociception
(Sorg et al.,
1998)
+
Chamber type NR;
declining HCHO
exposures across
days
+
N= 15-24; females
only
Imprecise timing of
assessment following
exposure; unclear if
cocaine or saline
challenged
Note: single exposure level;
1 or 4wk exposures
+
Experimenter blinding
not indicated
++
Medium
[Unclear exposure to
testing latency]
Functional Observational Battery or Grip Strength
(Chonglei et
al.. 2012)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
+
N= 5 males/group
+
Unclear exposure to testing
latency; 2hr/d exposure for
short term (10 days)
No description of grip
strength protocol
provided
++
Not informative
[Mixture exposure;
endpoint protocol
NR]
(Tepper et al.,
1995)
Carpet emission
exposu res:
formaldehyde not
primary exposure
(BHT, toluene, etc.)
N= 2(nonexposed
controls) or 4; males
only
Behaviors tested
immediately after
exposure
++
Quantitative data
NR for the majority
of measures; some
measures presented
as compared to
preexposure or
summarized
qualitatively
Not informative
[Mixture exposure;
small sample; etc.]
(Wang et al.,
2014)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
+
N = 6 males/group
Note: no changes in
body weight were
observed
+
Unclear exposure to testing
latency; Note: 2hr/d for 49-
90 d
+
No blinding indicated;
Note: 5s inter-trial
delay and 3 trials/d
++
Not informative
[Mixture exposure]
Electrophysiology (for Hazard; see below for MOA)
(Bokina et al.,
1976)
Details of exposure
were not provided
Details on test
subjects were not
provided
Details of study design
were not provided
Details of endpoint
measures were not
provided
No quantitative
comparisons to
controls were
performed
Not informative
[Experimental details
NR]
This document is a draft for review purposes only and does not constitute Agency policy.
A-610 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
Katsnelson,
2013,1987924}
Test article not
defined (assumed to
be formalin; high
concentration:
methanol may drive
effects)
+
N= 12-15/group;
females only
+
Testing indicated as
immediately after
exposure: unclear if RB-
related effects could affect
these impulses
Note: subchronic (10 wk)
exposure
++
Note: Citation for
temporal summation
of impulses protocol
was provided
++
Not informative
[High levels of test
article assumed to be
formalin]
Autonomic Effects (for Hazard; see below for usefulness for MOA)
(Nalivaiko et
al.. 2003)
Unregulated
exposure without
reporting of levels;
no chamber
Note:
paraformaldehyde
+
N = 6-13; males only
No nonexposed groups
indicated (internal
comparisons); acute
exposure; All animals
implanted with electrodes
(duration before tests NR)
+
ECG implantation
procedures NR
Note: endpoint not
considered adverse
++
Not informative
[Exposure levels NR
and unregulated;
etc.]
(Tani et al.,
1986)
Formalin (high
concentration:
methanol may drive
responses)
+
N = 4-5; males only
No nonexposed groups
indicated (internal
comparisons); acute
exposure; all animals
received anesthesia,
surgery, and anticoagulants
(no recovery before
exposure)
Blocker experiments
may be influenced by
prior exposure to
formaldehyde
Note: endpoint not
considered adverse
+
Effects of blocker
experiments without
prior HCHO exposure
NR
Not informative
[High formalin levels;
etc.]
(Yu and
Blessing, 1997)
Formalin (likely high
concentration- not
quantified: methanol
may drive
responses); HCHO
concentrations NR
+
N = 5-16; males only
No nonexposed groups
indicated (internal
comparisons); acute
exposure; all animals
received surgery,
anesthesia, and
catheterization lwk prior
to exposure
++
Note: Endpoint not
adverse
+
Data were pooled
across groups for
some measures
Note: all
comparisons to
preexposure
measures
Not informative
[Formalin levels NR;
etc.]
This document is a draft for review purposes only and does not constitute Agency policy.
A-611 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
(Yu and
Blessing, 1999)
Test article not
defined (assumed to
be formalin); levels
not quantified (likely
high: methanol may
drive responses)
+
N = 4; males only
No nonexposed groups
indicated (internal
comparisons); other
alerting & noxious stimuli
administered pre-HCHO; 2
surgeries- only Id recovery
after cannulation before
exposure; acute exposure
++
Note: Endpoint not
adverse
+
Justification for
selection of resting
periods used for
comparison unclear;
data qualitative only
Not informative
[Test article assumed
to be formalin;
exposure levels NR;
etc.]
NR = not reported; N/A = not applicable;
* Three studies examined an endpoint that is not adverse and has no MOA relevance. These are briefly mentioned in the assessment, as they inform the
irritant/odorant threshold of rodents, but these studies were not used to characterize the potential neurotoxicity hazard.
** Five animal studies sufficient for hazard characterization were not categorized using confidence ratings, and they are not included in the exposure-response
array, as they represent cursory observations with none or minimal data reporting; however, these studies were used to help describe the potential
neurotoxicity hazard.
a See the draft Methanol Toxicological Review (http://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=233771), which proposes an RfC of =2mg/m3.
Assuming methanol is present in the breathing zone somewhere in the range of 1/10-1/3 the levels of formaldehyde when stabilized formalin solutions are
used as the test article (determination of the exact ratio of exposure is not currently available), exposures > 10mg/m3 are assumed to have at least some
methanol-driven effects.
b Kerns is a report of a GLP study by CUT (Battelle, 1982), which was not identified in the literature search [Note: use of GLP or guideline study protocols is
provided to identify the most stringent studies, but did not factor into the confidence ratings or sufficiency evaluations for this particular database],
c Communication with the study author detailed that male rats (2 per litter from 3 separate dams per dose group) were used in the Sarsilmaz et al. (2007)
study. A review from this same laboratory (Songur et al., 2010) indicated that the stereological studies of the hippocampus were conducted to confirm
previous observations (Songur et al., 2003); thus, the separate reports of stereological changes in the CA and DG regions of the hippocampus (Sarsilimaz et al.,
2007 and Asian et al., 2006, respectively) are assumed to represent the same cohort of animals (note: it is possible that these two stereological studies report
effects on a subset of the same animals used in the Songur et al. (Songur et al., 2003) study, but this inference is less clear and is not assumed).
d Note: although pup body weight changes would be of concern as potential confounders for behavioral analyses, endpoints such as neuropathology and brain
weight are unlikely to be secondary to these changes: at least for brain weight, the current literature does not support a consistent causal relationship. In
Songur et al. (Songur et al., 2003), body weight decreases were =10% and 20% at 30d (low and high formaldehyde concentrations, respectively) & =10% at
60d (high concentration only).
e Because data for exposure groups other than 6.15 mg/m3 were not reported by Boja et al. (Boja et al., 1985), the higher exposure groups were not included in
the study quality analysis or the Toxicological Review hazard ID synthesis.
This document is a draft for review purposes only and does not constitute Agency policy.
A-612 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
Studies Specific to Mechanistic Considerations Only
Studies examining mechanistic events related to nervous system effects were systematically
evaluated in order to inform biological plausibility. The evaluations included herein only
encompass animal studies reporting mechanistic results following in vivo inhalation exposures
(including exposures to animals under anesthesia or after surgery). Noninhalation (e.g., oral, i.p.)
animal exposures are expected to involve a different distribution of formaldehyde to systemic sites
such as the nervous system, as compared to inhalation exposure, and thus are likely to involve
mechanisms unrelated to those observed following inhalation. Similarly, in vitro examinations
were also not considered to be informative enough to warrant study quality evaluations, as
appreciable amounts of formaldehyde are unlikely to reach the target cells in the nervous system
following inhalation exposure. Notably, the aqueous formaldehyde solutions used in both in vitro
and noninhalation in vivo studies typically contained methanol as a stabilizer, introducing
additional uncertainties.
Although parallel criteria to those used to evaluate studies describing potential
neurotoxicity hazards (see above) were used to judge the mechanistic studies, the stringency of
some criteria were adapted to accommodate this type of information and additional leniency was
applied for certain parameters (e.g., acute exposure was not considered a limitation). Studies are
organized alphabetically.
This document is a draft for review purposes only and does not constitute Agency policy.
A-613 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-87. Evaluation of studies pertaining to mechanistic events associated with nervous system effects
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
Exposure Quality
Test Subjects
Study Design
Endpoint Evaluation
Data Considerations &
Statistical Analyses
Overall Confidence
Rating Regarding
the Use for MOA
[Main limitations]
Expert judgement
based on conclusions
from evaluation of the
5 experimental feature
categories
Criteria relevant
to evaluating the
experimental
details within each
experimental
feature category3
Exposure quality
evaluations (see B.4.1.2)
are summarized below;
robust;
adequate; and shaded
box: poor; relevance of
the tested exposure
levels is discussed in the
hazard synthesis
The species, sex, strain,
and age are appropriate
for the endpoint(s);
sample size provides
reasonable power to
assess the endpoint(s);
overt systemic toxicity is
absent or not expected,
or it is accounted for;
selection bias minimized
A study focus was nervous
system effects; the exposure
regimen is informative for the
tested endpoint(s); acute
exposure not necessarily a
limitation; manipulations
other than formaldehyde
exposure are adequately
controlled
Endpoint evaluates a
mechanism relevant to
humansb; protocols are
complete, sensitive,
discriminating, &
biologically sound;
experimenter bias
minimized
Statistical methods, group
comparisons, and data
presentation (including
variability) are complete,
appropriate, and
discerning; selective
reporting bias avoided
(Ahmed et
al.. 2007)
++
+
N = 4-5; females only
Lack of OVA-free controls:
inability to separate
effects of OVA &
formaldehyde; possible
altered
distribution/effectiveness
of aerosolized OVA given
after formaldehyde; Note:
12wk exposure; single
exposure level
++
++
Medium
[Control group
deficiencies]
(Bian et al.,
2012)
Formalin (high
concentration:
methanol may drive
effects)
N =
3/endpoint/timepoint
(males); mild toxicity:
decreased food intake
(effect not quantified)
Controls not air-exposed
in exposure chamber; all
groups had anesthesia &
antibiotic injections
Note: exposure 1 hr/d for
90d; single exposure level
++
++
Not informative
[High formalin levels;
etc.]
(Boia et al..
1985)
+
Analytical
concentrations NR
+
N = 8; males only; data
from experiments with
N=1 (air-HCHO NE &
DA levels) not included
in the assessment
+
Timing of exposures (9-
12pm vs. 12-3pm) may
have varied across groups
Note: single exposure
level; acute exposure:
3hr/d for l-2d
+
Molecular verification
of regional "punches"
not performed
+
Higher exposure
groups data NR;
inability to evaluate
findings for exposures
indicated as tested but
NR
Medium
[Selective reporting;
some methods detail
NR]
This document is a draft for review purposes only and does not constitute Agency policy.
A-614 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
(Bokina et
al.. 1976)
Details of exposure
were not provided
Details on test
subjects were not
provided
Details of study design
were not provided
Note: continuous exposure
for 45d
Details of endpoint
measures were not
provided
No quantitative
comparisons to
controls were
performed
Not informative
[Experimental details
NR]
(Fuiimaki et
al.. 2004b)
+
Analytical
concentrations NR
+
N = 5-6; females only;
unclear influence of
splenic effects (e.g.,
decreased weight)
+
For OVA groups: unclear if
prior formaldehyde
exposure had nasal effects
influencing inhaled OVA
booster
distribution/effects; Note:
12 wk exposure
+
Methods for EUSA of
plasma NR: assumed
to be same as BAL
fluid ELISA
++
Medium
[Control group
deficiencies; some
methods detail NR]
(Fuiimaki et
al.. 2004a)
+
Analytical
concentrations NR
+
EUSA data: N=5; males
only
RT-PCR data: A/=3;
(considered major
limitation)
+
for OVA groups: unclear if
prior formaldehyde
exposure had nasal effects
influencing inhaled OVA
booster
distribution/effects; 12 wk
exposure
Methods for brain
dissection &
homogenization, as
well as gel
quantification NR;
ELISA and booster
challenge methods
NR
++
ELISA: Medium
RT-PCR: Low
[Control group
deficiencies; small
sample size; some
methods detail NR]
(Gieroba et
al.. 1994)
Formalin (likely high
concentration- not
quantified: methanol
may drive response)
N= 2 or 6
Unclear contribution of
apnea & bradycardia;
results may be specific to
exposure combined with
restraint & anesthesia;
strong irritation induced
+
Number of sections
analyzed/animal NR
Immunostaining
results were not
quantified across
groups; results are
qualitative only; TFT
cell counts alone NR
Not informative
[High formalin levels;
etc.]
(Havashi et
al.. 2004)
++
+
N = 5; females only
++
Exposures up to 12 wk
+
Possible mild
sampling bias (3
sections, but
selection methods
NR); blinding
indicated
++
High
(Kimura et
al.. 2010)
Formalin
N = 5-6; males only;
systemic toxicity not
evaluated (HCHO
+
Irritation-related effects
probable, as tested near-
simultaneous with
+
Blinding not indicated
for cell type counts
++
Low
[Formalin; possible
overt toxicity]
This document is a draft for review purposes only and does not constitute Agency policy.
A-615 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
tested up to "55
mg/m3)
exposures; acute
exposure; unclear if
anesthesia/dye injection
influenced sensory nerve
responses
(Kulie and
Cooper,
1975)
+
Analytical
concentrations NR
N=3; males only; no
air-only controls
+
All animals underwent
surgery prior to exposure
(no recovery prior to
exposure); some exposures
were complicated by amyl
alcohol co-exposure; acute
exposure
++
Note: unclear
relevance of these
surgical preparations
to human nerve
responses
No quantitative
comparisons to
controls performed
(extrapolated
threshold only)
Low
[small sample size;
comparison group
deficiencies]
(Chonglei et
al.. 2012)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
+
N= 5 males/ group
++
2hr/d exposure for short
term (10 days)
No description of
hippocampal MDA
and GSH protocols
provided
++
Not informative
[Mixture exposure;
etc.]
(Li et al.,
2016)
Formalin; static
chambers
+
N = 7 (inferred); males
only
++
2hr/d exposure for short
term (7 days)
+
Some sampling bias
possible: 3 sections
Note: although not
corrected for neuron
number, location
determined from
atlas; slides were
randomized and
coded for blinded
evaluation
++
Low
[Formalin]
(Liao et al.,
2010)
(translation)
Formalin/static
chamber
N=8: pooled sexes
(A/=4/sex); overt
toxicity during
exposure (e.g.,
listlessness; up to
=30% decreased body
weight gain), most
likely from poor
+
No indication of correction
for possible litter bias;
Note: 2hr/d for28d
Potential sampling
bias: N=5 fields
(assumed to be per
animal), but number
of slides not
indicated (DAB
amplification used) &
no correction made
to account for the
+
Data= combined sexes;
CA3 cell number or
viability measures NR
Not informative
[Formalin; endpoint
protocol deficiencies;
overt toxicity]
This document is a draft for review purposes only and does not constitute Agency policy.
A-616 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
exposure quality, as
only 0.5mg/m3 HCHO
number of neurons
visible/field
(Liu et al.,
2009a)
Formalin (high
concentration:
methanol may drive
effects)/static
chamber
+
N = 5; males only
++
28 d exposures
++
++
Not informative
[High formalin levels;
etc.]
(Liu et al..
2010)
Formalin (high
concentration:
methanol may drive
effects)/static
chamber
+
N=5; males only;
treatment groups
determined by
preexposure probe
trial performance, but
method for matching
groups NR
++
28 d exposures
Methods for
quantification of
western blots NR
++
Not informative
[High formalin levels;
etc.]
(Lu et al.,
2008b)
Unspecified wood
+
Sample sizes for MOA-
related endpoints
were NR, but assumed
to be N=5; males only
++
7 d exposures
Regional brain
dissections were
nonspecific &
methods
incompletely
described; RT-PCR
analyses were semi-
quantitative only
++
Low
[Possible mixture
exposure; endpoint
protocol description
insufficient]
(Matsuoka
et al., 2010)
Formalin
+
N=7-9; males only
+
Did not appear that
controls were air-exposed
in chambers
("noninhalation controls");
acute exposure
Methods for brain
dissection/regions
analyzed NR;
assumed brain
region-specific
analyses were not
conducted
+
High variability in
measures, possibly due
to lack of regional
specificity
Low
[Formalin; endpoint
protocol description
insufficient]
(Mei et al.,
2016)
Formalin
+
N = 8; males only
+
No comparisons to
chamber or air exposure
alone; 8hr/d for 7
consecutive days
No blinding for
biochemical
measures; no
regional specificity
(homogenates)
++
Low
[formalin; some
endpoint protocol
limitations]
This document is a draft for review purposes only and does not constitute Agency policy.
A-617 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
(Nalivaiko et
al.. 2003)
Unregulated
exposure without
reporting of levels;
no chamber
Note:
paraformaldehyde
+
N = 6-13; males only
+
No nonexposed groups
indicated (internal
comparisons); all animals
were implanted with
electrodes, but duration
prior to testing not
provided; acute exposure
+
ECG implantation
procedures NR
++
Not informative
[Exposure levels NR
and unregulated; etc.]
(Ozen et al.,
2003)
+
Analytical
concentrations NR
Unclear contribution
of unexplained overt
toxicity (robust effects
on body weight);
males only; N = 1
++
4 wk or 13 wk exposures
Methods for analyses
of brain tissue were
not clearly described,
even in cited
reference
++
Not informative
[Overt toxicity;
endpoint protocol
description insufficient]
(Sari et al.,
2004)
++
+
N=5/e nd po i nt; females
only
++
12 wk exposure
Cell counts were not
reported as observer
blinded, but were
from serial sections;
RT-PCR analyses
were semi-
quantitative only
++
Medium
[possible experimenter
bias- no blinding]
(Sari et al.,
2005)
++
+
N = 5; females only
Nasal instillation of
toluene may affect
formaldehyde distribution
Cell counts were not
reported as observer
blinded, but were
from serial sections
Data for exposures
without toluene NR
Note: 2004 paper data
cited was not
considered
Not informative
[Data on formaldehyde
exposure alone NR;
etc.]
(Songur et
al.. 2003)
+
Analytical
concentrations NR
N = 6 (assumed 3
litters); mild toxicity
(body weight &
food/water intake
changes): HSP
activation may be
indirectly related to
health/nutrition
+
Litter assignments NR;
unclear if litter bias
corrected; 30d of exposure
Potential sampling
bias: details on
blinding,
slides/animal, etc.
not provided;
nonblinded intensity
ratings subject to
observer bias
+
No statistical
comparisons for HSP
staining
Low
[small sample size;
possible litter and/or
sampling bias]
(Songur et
al.. 2008)
++
Dam health during
lactation & pup health
not presented; sex
and litters/group
+
Unclear if litter bias
corrected (& not indicated
as randomized); dams
++
++
Medium
[Small sample size;
possibly litter effects]
This document is a draft for review purposes only and does not constitute Agency policy.
A-618 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
unknown (likely males
& 3 litters); body
weights were indicated
as measured, but NR;
N= 7 pups
exposed from PND1-14;
30d of exposure
(Sorg et al.,
2001a)
++
+
N = 6-10; males only
+
Possible difference in
harvest day (20 vs 21)
across groups may
contribute to high
variability noted in results;
exposures <4 wk
+
Volume of trunk
blood/animal and
some other details
(e.g., serum isolation)
NR
Note: chamber
exposure itself
(tested) had a large
influence, so critical
to rapidly remove rats
after exposure (as
indicated)
++
Note: sensitive
endpoint, so high level
of variability is as
expected
High
(Sorg et al.,
2002)
Formalin (likely high
concentration; not
quantified: methanol
may drive response)
+
N = 6-12
Formalin used as an
aversive stimulus- results
more specific to cocaine;
acute exposure to
concentrated vapors
Tests involve odor
detection &
irritation-specific
responses could be
confounding results
+
Specific effects of
formaldehyde alone
not tested or NR
Not informative
[Formalin (assumed
high level) levels NR]
(Tani et al.,
1986)
Formalin (high
concentration:
methanol may drive
responses)
+
N = 4-5; males only
+
No nonexposed groups
indicated (internal
comparisons); animals
received anesthesia,
surgery, and drugs with no
recovery before exposure;
acute exposure
+
Blocker experiments
may be influenced by
prior exposure to
formaldehyde (not
tested)
++
Not informative
[High formalin levels]
(Tsukahara
et al., 2006)
++
+
Females only; Western
Blot data: N> 6;
Caspase data: A/=3;
(considered major
limitation)
+
For OVA groups: unclear if
prior formaldehyde
exposure had nasal effects
influencing inhaled OVA
booster
++
(for Western Blot
data)
Caspase data: likely
sampling bias:
number of
++
Western blot: High
Caspase: Low
[Caspase data: small
sample size; likely
sampling bias]
This document is a draft for review purposes only and does not constitute Agency policy.
A-619 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature limitation is
indicated
distribution/effects; 60d
exposure
slides/animal &
neurons visible/field
NR; counts were not
reported as observer
blinded
(Wang et al.,
2014)
Mixture (formalin,
benzene, toluene
and xylene)/static
chamber
+
N = 6-12; males only
Note: no changes in
body weight were
observed
++
2hr/d exposure for
subchronic (90 days);
tested Id postexposure
No description of grip
strength protocol
provided
++
Not informative
[Mixture exposure;
endpoint protocol NR]
(Yu and
Blessing,
1997)
Formalin (likely high
concentration; not
quantified: methanol
may drive responses)
+
N = 5-16; males only
Animals received surgery,
anesthesia, &
catheterization 1 wk prior
to exposures; no
nonexposed groups
indicated (internal
comparisons); acute
exposure
++
+
Data was pooled
across groups for some
measures
Note: all comparisons
to preexposure
measures
Not informative
[Formalin (assumed
high level) levels NR;
etc.]
(Yu and
Blessing,
1999)
Test article not
defined (assumed to
be formalin); levels
not quantified (likely
high: methanol may
drive responses)
+
N = 4; males only
No nonexposed groups
indicated (internal
comparisons); other
alerting & noxious stimuli
administered pre-HCHO; 2
surgeries; only Id
recovery after cannulation
before exposure; acute
exposure
++
+
Justification for
selection of resting
periods used for
comparison unclear;
data qualitative only
Not informative
[Unknown test article
(assumed to be
formalin) levels NR
(assumed high level);
etc.]
(Zitting et
al.. 1982)
Test article results in
co-exposures to
formic acid, acrolein,
& possibly other
chemicals
+
N = 4-5; males only
Formaldehyde levels »
100mg/m3 are overtly
toxic (rats gasped for air
for hours after exposure);
6hr or 3d exposure
+
Evaluations are not
brain-region-specific
+
Details on statistics NR
(e.g., "Student's t test")
Not informative
[Unknown test article
(assumed to be
formalin) at high level;
overt toxicity]
a Mode-of-action study quality evaluations were conducted in a similar fashion as those described above for hazard identification, with minor adjustments to the
types of experimental details considered for meeting sufficiency criteria (e.g., adversity of the endpoint was not considered).
b A mechanism or mode of action is considered relevant to humans unless there is convincing evidence to the contrary.
This document is a draft for review purposes only and does not constitute Agency policy.
A-620 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information for Formaldehyde—Inhalation
A.5.8. Developmental and Reproductive Toxicity
Literature Search
A systematic evaluation of the literature database on studies examining the potential for
noncancer developmental and/or reproductive effects in humans or animals in relation to
formaldehyde exposure was initially conducted in October 2012, with yearly updates (see A.1.1).
The search strings used in specific databases are shown in Table A-88. Additional search strategies
included:
• Review of reference lists in the articles identified through the full screening process.
• Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (US EPA,
2010), the ATSDR toxicological profile of formaldehyde (ASTSDR, 1999), and the NTP
report on carcinogens background document for formaldehyde (NTP, 2010).
• Review of references in 41 review articles relating to formaldehyde and reproductive or
developmental effects, published in English, identified in the initial database search.
References were retrieved through Web of Science and added to the database.
This review focused on reproductive effects in women and men, fetal loss (e.g., spontaneous
abortion), and birth outcomes. Effects in animals included alterations in pre- and postnatal
development (survival, growth, structural alterations) and in the integrity of the male and female
reproductive system (cells/tissues/organs, outcomes, and function). Inclusion and exclusion
criteria used in the screening step are described in Table A-89 and Table A-90, respectively, for
human and animal studies.
After manual review and removal of duplication citations, the 9,854 articles identified from
database and additional searches were initially screened within an EndNote library for relevance;
title was considered first, and then abstract in this process. Full text review was conducted on 261
identified articles. The search and screening strategy, including exclusion categories applied and
the number of articles excluded within each exclusion category, is summarized in Figure C.5.9.-1.
Based on this process, 55 studies were identified and evaluated for consideration in the
Toxicological Review.
This document is a draft for review purposes only and does not constitute Agency policy.
A-621 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-88. Summary of search terms for developmental or reproductive
toxicity
Database,
search date
Terms
PubMed
No date
restriction
(formaldehyde [majr] OR paraformaldehyde OR formalin) AND ("reproductive toxicity" OR
"reproductive toxicology" OR reproductive OR "developmental toxicity" OR "developmental
toxicology" OR development OR developmental OR ontogen* OR "embryo toxicity" OR embryo
OR embryon* OR embryog* OR embryot* OR "fetal loss" OR fetal OR fetus OR fetuses OR
fetotoxi* OR miscarriage or miscarry OR "spontaneous abortion" OR "preimplantation loss" OR
preimplantation OR "postimplantation loss" OR postimplantation OR implantation OR
conception OR resorption OR fertility OR fertile OR infertility OR infertile OR pregnancy OR
gestation OR neonatal OR neonate OR prenatal OR postnatal OR "menstrual cycle" OR
"premature birth" OR "preterm birth" OR "low birth weight" OR "in utero" OR "fetal body
weight" OR "fetal weight" OR pup OR "pup body weight" OR "pup weight" OR ovary OR ovaries
OR ovu* OR sperm OR gamete OR "germ cells" OR "Sertoli cells" OR testes OR testis OR testic*
OR uterus OR uteri* OR epididy* OR prostate OR "seminal vesicles" OR semen OR testosterone
OR "luteinizing hormone" OR LH OR "follicle stimulating hormone" OR FSH OR estrogen OR
estradiol OR "time to pregnancy" OR "time-to-pregnancy" OR TTP OR fecund*)
NOT (fixative OR "formaldehyde fixation" OR "paraformaldehyde fixation" OR "formalin
fixation" OR "formaldehyde fixed" or "paraformaldehyde fixed" OR "formalin fixed" OR
"formaldehyde-fixed" or "paraformaldehyde-fixed" OR "formalin-fixed" OR formocresol OR
dental OR dentistry OR immunogen OR vaccine OR vaccination OR metabolite)
[Note: for quality control, -1% (75) of the 7,589 excluded article titles were scanned in
PubMed: 2 potentially relevant government reports were found and 4 duplicates were
excluded, resulting in 2,810 in the final database.
Web of Science
No date
restriction
Lemmatization
"off
SU=(Toxicology OR "Pharmacology &Pharmacy" OR "Public, Environmental & Occupational
Health" OR "Cell Biology" OR "Reproductive Biology" OR "Biochemistry & Molecular Biology"
OR Pathology OR "Obstetrics & Gynecology" OR "Environmental Sciences" OR "Anatomy &
Morphology" OR Andrology OR "Veterinary Sciences" OR Physiology OR "Developmental
Biology" OR "Research & Experimental Medicine" OR "Life Sciences Biomedicine Other Topics"
OR "Veterinary Sciences") AND TS=(formaldehyde OR paraformaldehyde OR formalin) AND
TS=(formaldehyde OR paraformaldehyde OR formalin) AND TS=(formaldehyde OR
paraformaldehyde OR formalin) AND TS=("reproductive toxicity" OR "reproductive toxicology"
OR reproductive OR "developmental toxicity" OR "developmental toxicology" OR development
OR developmental OR ontogen* OR "embryo toxicity" OR embryo OR embryon* OR embryog*
OR embryot* OR "fetal loss" OR fetal OR fetus OR fetuses OR fetotoxi* OR miscarriage or
miscarry OR "spontaneous abortion" OR "preimplantation loss" OR preimplantation OR
"postimplantation loss" OR postimplantation OR implantation OR conception OR resorption OR
fertility OR fertile OR infertility OR infertile OR pregnancy OR gestation OR neonatal OR
neonate OR prenatal OR postnatal OR "menstrual cycle" OR "premature birth" OR "preterm
birth" OR "low birth weight" OR "in utero" OR "fetal body weight" OR "fetal weight" OR pup
OR "pup body weight" OR "pup weight" OR ovary OR ovaries OR ovu* OR sperm OR gamete OR
"germ cells" OR "Sertoli cells" OR testes OR testis OR testic* OR uterus OR uteri* OR epididy*
OR prostate OR "seminal vesicles" OR semen OR testosterone OR "luteinizing hormone" OR LH
OR "follicle stimulating hormone" OR FSH OR estrogen OR estradiol OR "time to pregnancy" OR
"time-to-pregnancy" OR TTP OR fecund*)
NOT (fixative OR "formaldehyde fixation" OR "paraformaldehyde fixation" OR "formalin
fixation" OR "formaldehyde fixed" or "paraformaldehyde fixed" OR "formalin fixed" OR
"formaldehyde-fixed" or "paraformaldehyde-fixed" OR "formalin-fixed" OR formocresol OR
dental OR dentistry OR immunogen OR vaccine OR vaccination OR metabolite)
This document is a draft for review purposes only and does not constitute Agency policy.
A-622 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Database,
search date
Terms
[Note: for quality control, -2% (40) of the 2,309 excluded article titles were scanned in Web of
Science: none were relevant].
ToxNet (Toxline
and DART)
No date
restriction
(formaldehyde OR paraformaldehyde OR formalin) AND ("reproductive toxicity" OR
"reproductive toxicology" OR reproductive OR "developmental toxicity" OR "developmental
toxicology" OR developmental)
(including synonyms and CAS numbers, but excluding PubMed records); 525 identified; 11
discarded upon importation into EndNote because they were duplicates
This document is a draft for review purposes only and does not constitute Agency policy.
A-623 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-89. Inclusion and exclusion criteria for studies of reproductive and
developmental effects in humans
Included
Excluded
Population
Human
Animals
Exposu re
• Indoor exposure via inhalation to
formaldehyde
• Measurements of formaldehyde
concentration in air
• Formaldehyde-specific assessments in
exposed occupations (wood workers, nurses,
pathologists, cosmetologists)
• Not formaldehyde
• Outdoor formaldehyde
exposure
• Mixtures or industry/job title
analyses
• Not inhalation
Comparison
•
• Case reports
Outcome
• Reproductive toxicity (sperm measures)
• Time-to-pregnancy (fecundity)
• Spontaneous abortion
• Pregnancy
• Birth outcomes
• Exposure studies/no
outcomes evaluated
• Other health outcomes not
related to reproduction or
development
Other
•
• Reviews, reports, meeting
abstract, methodology paper,
laboratory techniques using
formalin, mechanistic studies,
foreign language
Table A-90. Inclusion and exclusion criteria for studies of reproductive and
developmental effects in animals
Included
Excluded
Population
• Experimental animals
• Nonmammalian test species or test
paradigms that are relevant for
evaluation or developmental or
reproductive hazard
• Humans
• Irrelevant species or test paradigms
Exposu re
• Inhalation route, formaldehyde
• Not formaldehyde
• Noninhalation routes of exposure
• Mixture studies
• Ecological studies
Comparison
• Inclusion of a comparison group
(e.g., pre- or postexposure, no
exposure, vehicle exposure, lower
formaldehyde exposure level)
• No comparison group
Outcome
• Pre- and postnatal offspring
biomarkers of:
• No health outcomes evaluated
This document is a draft for review purposes only and does not constitute Agency policy.
A-624 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Included
Excluded
o Survival (e.g., resorptions, death)
• Health outcomes not related to
developmental or reproductive
o Growth (e.g., body weight)
toxicity
o Structural anomalies (e.g., external,
• Mechanistic data irrelevant to
skeletal, or soft tissue malformations
developmental or reproductive
or variations)
outcomes
o Functional deficits
•
• Adult biomarkers of reproductive
toxicity, including:
o Gonadotropic hormone measures
o Reproductive organ weight
o Reproductive organ macro- and
microscopic pathology
o Sperm measures (count, motility,
morphology)
o Reproductive function (e.g., mating,
fertility, parturition, gestation,
lactation)
o Mechanistic data relevant to
developmental or reproductive
outcomes
Other
• Original primary research
• Not original primary research, e.g.,
reviews, reports, commentaries,
meeting abstracts, policy papers
• Duplicates, or untranslated foreign
language studies (judged to be
irrelevant or unlikely to have a
significant impact, based on review of
title, abstract, and/or tables/figures)
• Methodology papers, or studies
describing laboratory techniques
using formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
A-625 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reproductive and Developmental Toxicity (Human and Animal) Literature Search
PubMed
Toxline
-c
k.
o
I
I
4331 citations
636 citations
Q.
E
o
u
n ^
Ji »
« a?
Is
.¦s &
"O
"O
<
a.
3
ts
ra
i
a>
(2
1474 from references in 41 reviews;
50 additional citations from review of
citations in other references
2013 Literature Search Update
2014 Literature Search Update
2015 Literature Search Update
2016 Literature Search Update
6929
+ 1524
8453
-152
8301
+ 595
s
+ 448
-J
1 X
+ 504
1
+ 300
10,154
Web of Science
3638 citations
-9889
265
-202
(after duplicate removal from merged
dataset)
323 Reviews
Electronic duplicate
removal of from the newly
added citations
Title - Abstract Screen
Excluded because did not meet criteria for:
Population 5541
Exposure 1666
Comparison
Outcome 2177
Other 505
(abstract, review, duplicate, foreign language -
not on topic)
Full Text Screen
Excluded because did not meet criteria for:
Human Studies Animal Studies
Exposure
74
Comparison
2
Outcome
11
Other
115 mechanistic
55 articles included in study evaluation including 20 studies in
humans and 35 animal toxicology studies (8 background studies)
Figure A-36. Literature search documentation for sources of primary data
pertaining to formaldehyde exposure and developmental and reproductive
toxicity.
This document is a draft for review purposes only and does not constitute Agency policy,
A-626 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
Human Studies
Participant Selection
Occupational studies of spontaneous abortion may be influenced by selection bias if
participants are recruited from current employees. This is because women with a history of
spontaneous abortion are more prevalent in the working population fAxelsson. 19841.
Time-to-pregnancy also may be increased among current workers because early spontaneous
abortion contributes to this effect (Slama etal.. 2014: Baird etal.. 1986). Four of the reviewed
studies reduced the potential for selection bias by recruiting from union rosters, registers of
licensed practitioners, or graduate school enrollment lists fTaskinen etal.. 1999: Steele and
Wilkins. 1996: Tohn etal.. 1994: Taskinen et al.. 19941. Another case-control study identified
spontaneous abortion events from a nationwide hospital discharge register fLindbohm etal.. 19911.
Thus, selection into the study was not conditional on being currently employed in the industry at
the time of the study. Regardless of the method used to identify the study population, most of the
studies used an appropriate comparison—other employed individuals. Generally, participation
rates reported by study authors were above 70%; thus, participants likely were representative of
the population under study.
Another potential bias may result from which pregnancy (first, pregnancy during defined
time period, most recent) is selected as the index pregnancy in studies of spontaneous abortion.
Studies that focus on the most recent pregnancy may be less sensitive due to time-lapse bias. The
time between a pregnancy ending in spontaneous abortion and a subsequent pregnancy ending in a
live birth is often shorter than two pregnancies, both ending in live births. This can result in a bias
toward identifying live births as the most recent pregnancy (Wilcox. 2010).
Outcome ascertainment
The validity of retrospectively collected self-completed questionnaire data on time-to-
pregnancy has been evaluated by some authors and was found to closely reproduce the
distributions of TTP in the group using a different data source (e.g., data collected during annual
follow-up of a family planning cohort) (Joffe et al., 1995). This finding suggests that data from the
questionnaires can be used to differentiate differences between groups. The comparability of the
distributions based on the two data sources persisted even among individuals for whom the
duration of recall was greater than 14 years. In addition, subfertility, defined as a TTP greater than
12 months using the questionnaires, was identified with high sensitivity (79.9%) and specificity
(94.9%) (Joffe et al., 1993). However, individuals recalled the number of months before conception
with greater error, and these errors increased as the duration of time-to-pregnancy increased.
Longer TTP was both over- and under-estimated (Coonev etal.. 2346932: Toffe etal.. 1995).
Therefore, while individual estimates of TTP may be less precise, the comparison of group means
with respect to levels of formaldehyde exposure is likely to be informative. The studies of TTP and
This document is a draft for review purposes only and does not constitute Agency policy.
A-627 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
formaldehyde exposure collected information about these variables in the same questionnaire;
thus, making it difficult to exclude the possibility that recall of TTP may have been differential with
respect to exposure status.
Validity studies indicate that recall of previous spontaneous abortions is relatively
complete, particularly for losses that occurred after the 8th week of gestation (> 80% of recorded
spontaneous abortions were recalled) (Wilcox and Hornev. 1984). Completeness varies by
occupation; completeness of recall among nurses was better than that among industrial workers
fLindbohm and Hemminki. 1988: Axelsson and Rvlander. 1982). Although elapsed time since the
event occurred may also influence the completeness of recall, this also varied by occupation in a
similar way (not important among nurses) and was not important within the first 10 years after the
event fLindbohm and Hemminki. 1988: Wilcox and Hornev. 19841. It is difficult to evaluate the
validity of self-reports of spontaneous abortion occurring during the 1st trimester using medical
records because these early events often are not recognized or do not require medical intervention;
medical records may not necessarily be an accurate reference (Slama etal.. 2014: Lindbohm and
Hemminki. 1988).
The degree to which the ability to recall a spontaneous abortion or a decision to participate
in the study may be associated with exposure status will affect the potential for bias with either
overestimation or underestimation of effect estimates fSlama etal.. 20141. Several of the studies
identified both cases and referents from the same occupational database or source population, thus
reducing the likelihood that recall was associated with formaldehyde exposure (Taskinen etal..
1999: Steele and Wilkins. 1996: Tohn etal.. 1994: Taskinen etal.. 1994). However, selection bias
has been documented in studies of spontaneous abortion within an occupational group. A study of
exposure to anesthetics among current and previous health personnel at a hospital in Sweden
reported a higher response rate among exposed cases fAxelsson and Rvlander. 19821. While the
rate of response to the mailed questionnaire was relatively high and comparable between the
exposed (85%) and unexposed (84%) female hospital personnel, an additional 20 spontaneous
abortions were found in hospital records for unexposed nonrespondents, whereas no additional
cases were found among exposed nonrespondents. It is difficult to predict the magnitude of the
impact of this potential selection bias on the findings of the reviewed studies, although it may vary
depending on the occupation.
Evaluation of Possible Confounding
Variables associated with time-to-pregnancy include age, gravidity (any previous
pregnancies), educational level, use of oral contraceptives, frequency of intercourse, recent
pregnancy or breastfeeding specific medical conditions, cigarette smoking, alcohol consumption,
and radiation exposure fBaird. 1988: Baird etal.. 1986: Baird and Wilcox. 19851. These individual
characteristics are possible confounders of the relation between formaldehyde exposure and time-
to-pregnancy if they are associated with formaldehyde exposure in the study population.
Spontaneous abortions during the first trimester most commonly result from chromosomal
This document is a draft for review purposes only and does not constitute Agency policy.
A-628 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
abnormalities, and risk factors include maternal and paternal age. Other factors associated with
increased risk include previous pregnancy loss, cigarette smoking alcohol consumption, and
maternal health conditions fWilcox. 2010. p. 153-1571. Almost all of the studies addressed these
potential confounding factors through adjusted analyses or by matching on characteristics
associated with spontaneous abortion risk. Adjusting for previous pregnancy loss or gravidity can
be problematic and potentially result in biased effect estimates because past pregnancy history also
may be related to exposure in ways that are part of the causal pathway. Therefore, adjustment for
these parameters was considered a limitation.
Exposure Assessmen t
A variety of different approaches to the assessment of exposure were used in this set of
studies. These ranged from more specific, robust measures such as estimates of time-weighted
average concentrations (based on job-specific formaldehyde measurements and the proportion of
time spent at the job reported by participants) fWang etal.. 2012: Taskinen etal.. 1999: Seitz and
Baron. 19901 to measures subject to greater misclassification error, such as the self-reported use of
specific products or chemicals, or assignment to exposures by supervisors. In the absence of
formaldehyde measurements, studies assigned exposure based on self-report (Steele and Wilkins.
1996: Tohn etal.. 1994: Saurel-Cubizolles etal.. 1994: Taskinen etal.. 1994: Axelsson etal.. 19841.
an informed source (Hemminki et al.. 1985: Hemminki et al.. 19821 or occupation/industry codes
from census data combined with expert knowledge of industry wide concentrations fLindbohm et
al.. 19911. Studies that used an open-ended question about what chemical exposures a participant
experienced were determined to be not informative and were excluded. The assignment to
exposure categories by third parties (supervisors of the participants or industrial hygienists) likely
resulted in an exposed group with large variation in exposure intensity and frequency with a
reduction in sensitivity. Exposure misclassification and the classification of individuals with
probable low or infrequent exposure as exposed was a major limitation in these and other studies
designated as low confidence fZhu etal.. 2006. 2005: Lindbohm etal.. 1991: Hemminki etal.. 1985:
Hemminki et al.. 19821.
Exposure assignments based on responses to questionnaires are likely to be affected by the
ability to recall exposures, resulting in misclassification. However, unless responses were
influenced by the respondent's pregnancy outcome, the misclassification would more often result in
an attenuation of the risk estimates. A study of women who worked in laboratories at a Swedish
university provides some evidence that differential recall bias may be an important issue. Women
who reported miscarriages that could not be verified in a national birth register, also reported a
higher rate of exposure to solvents f Axelsson and Rvlander. 19821. However, a few validity studies
of questionnaire responses about exposure among women with adverse reproductive and
pregnancy outcomes did not find evidence for differential recall bias. An investigation of the
repeatability of reported exposures among women who experienced a miscarriage did not find an
increase in reported occupational and residential exposures after the event (Farrow et al.. 19961.
This document is a draft for review purposes only and does not constitute Agency policy.
A-629 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information for Formaldehyde—Inhalation
Other studies of questionnaire validity reported that sensitivity and specificity of responses to
specific questions about chemical exposure were similar between individuals reporting a history of
subfertility or adverse pregnancy outcomes, and individuals in the comparison groups floffe etal..
1993: Ahlborg. 19901. Notably specificity was high for questions about specific chemicals,
indicating that false positives for exposure were less likely. Further, other studies have found that
under-reporting rather than over-reporting of exposures is more common (Toffe etal.. 1993:
Ahlborg. 1990: Hemminki et al.. 19851. Therefore, while differential reporting of exposure by
outcome status was evaluated for the studies of formaldehyde, it was not assumed to have
occurred.
The criteria that were important in the evaluation of the studies for these endpoints are
included in Table A-91 below. Information from the published studies pertinent to each of the
evaluation categories was evaluated and conclusions are documented in the table that follows (see
Table A-92). Studies are arranged alphabetically within each table.
Table A-91. Criteria for categorizing study confidence in epidemiology studies
of reproductive and developmental effects
Study
Confidence
Exposure
Study Design and Analysis
High
Work settings: Ability to differentiate between
exposed and unexposed, or between low and high
exposure. Exposure assessment specific to
formaldehyde exposures and based on concentration
data; includes assessment of intensity and frequency.
Exposures characterized for etiologically relevant
time window (e.g., period prior to or during
pregnancy attempt for time-to-pregnancy or first
trimester for spontaneous abortion).
Pregnancy outcomes compared between employed
exposed and employed referent groups.
Spontaneous abortion defined. Analytic approach
evaluating dose-response relationship using analytic
procedures that are suitable for the type of data,
and quantitative results provided. Confounding
considered and addressed in design or analysis; co-
exposures (risk factors for endpoint) relevant to
occupational setting addressed in analyses. Large
sample size (n cases).
Medium
Work settings: Exposure assessment may not include
formaldehyde concentration measurements, but
other information used to differentiate between
exposed and unexposed, or between low and high
exposure levels. Incorporation of information on
intensity and frequency. Referent group may be
exposed to formaldehyde or to other exposures
affecting reproductive or developmental outcomes
(potentially leading to attenuated risk estimates).
One or a few limitations noted but otherwise study
used a strong methodological and analytical design.
While potential confounders may have been
evaluated, co-exposures (risk factors for endpoint)
relevant to occupational setting may not be.
Low
High likelihood of exposure misclassification and no
information on frequency or intensity of exposure;
imprecise assignment of exposure period to relevant
time window for endpoint under study.
Evidence of confounding by other co-exposures in
workplace and only single pollutant analyses
presented; may be small number of exposed cases;
not all important potential confounders addressed.
Not
Informative
Use of an open question regarding occupational
exposures.
Insufficient reporting detail; insufficient number of
exposed cases ascertained; important potential
confounders not addressed (age, gravidity,
smoking).
This document is a draft for review purposes only and does not constitute Agency policy.
A-630 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-92. Evaluation of observational epidemiology studies of formaldehyde - reproductive and developmental
outcomes
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
Residential Studies
Franklin et al.
(2019) (Australia)
Birth cohort
Pregnant women,
all nonsmokers,
recruited prior to
18 weeks
gestation.
Recruited 373
women, 305
(81.7%)
participated; 4
excluded because
of smoking. Birth
data available for
262 live births.
Air monitoring in
homes at 34 weeks
gestation, 7-day
sampling duration
using validated
passive samplers in
bedroom and living
room. LOD 2.4
Hg/m3; used LOD/2
for values < LOD.
House average
Median (range) 2.81
(LOD-17.33)
Hg/m3; 23.3% < LOD.
Uncertainties in
exposure
distribution due to
large % < LOD.
Gestational age,
birth weight, birth
length and head
circumference
from birth records.
Confounders were
selected based on
previous literature.
Adjusted for maternal
age, parity, maternal
asthma, diabetes and
blood pressure,
season of birth.
Distance from main
road and ETS
exposure were
evaluated as
potential
confounders in
models. Adjusted and
unadjusted results
presented.
Gestational age
was normally
distributed.
Birth weight,
birth length and
head
circumference
were
transformed to
z-scores
(accounting for
sex and
gestational
age). General
linear models.
N = 262
Gestational age, birth
weight, birth length, head
circumference
SB
IB
Cf Oth
Overall
Confidence
Medium
¦
Medium
Uncertainties in exposure
distribution due to large %
< LOD, small sample size,
uncertain relationship
between outcomes and
window of exposure (3rd
trimester)
Amiri and
Turner-Henson
(Southeastern
United States)
Cross sectional
study
Pregnant women
in 2nd trimester
(convenience
sample, n = 140)
recruited from
obstetrics and
gynecology
clinics with no
history of chronic
Participants wore
vapor monitor
badges, 24-hour
period, detection
limit 0.003 ppm.
Mean (SD) 0.04
(0.06) ppm = 0.049
(0.074) mg/m3. This
is a measure of total
Ultrasonographic
biometry during
2nd trimester for
head
circumference,
abdominal
circumference,
femur length,
biparietal
Urine cotinine
adjusted for urinary
creatinine (spot
sample, methods and
timing of collection
were not described).
Models adjusted for
maternal
demographics,
Multiple linear
regression for
formaldehyde
as dichotomous
variable (cutoff
at 0.03 ppm)
adjusted for
maternal age,
fetal sex and
N = 88
Ultrasonographic biometry
measurements
SB IB Cf Oth
m
Overall
Confidence
Low
Low participation rate with
no comparisons raises
This document is a draft for review purposes only and does not constitute Agency policy.
A-631 DRAFT-
-D0 NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
disease or high-
risk pregnancy,
19 - 40 years old,
Participation 63%
(n = 88). No
comparison of
those who did
and did not
return the
formaldehyde
badges which
raises a concern
for selection bias.
exposure from
indoors and ambient
air.
diameter,
estimated fetal
weight, and ratio
of abdominal
circumference to
femur length.
Measurements in
mm converted to
percentiles using
gestational age
and the Hadlock
formulas.
Sensitivity and
specificity for IUGR
are 67% and 93%
for BPD, 42% and
100% for HC, 94%
and 100% for AC
and 46% and 90%
for AC/FL ratio.
Hadlock formulas
are based on a
sample of White
women in the US
with uncertain
accuracy for other
races. Over 50% of
the participants
were not White.
obstetric history, and
cotinine. Biometry
measurements were
not correlated with
maternal age,
education, marital
status, yearly family
income or
employment status.
No correlation with
gravida, maternal
smoking or
pregnancy intervals.
BPD was lower
among whites
compared to African-
Americans or other
category. BPD and FL
varied by sex.
race.
Mediation of
tobacco smoke
(urinary
cotinine) on
associations
examined.
concern for selection bias.
Small sample size with
reduction in sensitivity.
Reference population for
BPD measure was not
appropriate for >50% of
participa nts.
(Chang et al.,
2017) (Birth
Pregnant women
were selected
from cohort (n =
383), originally
Personal
formaldehyde
measurements
during mid- or late
Age-specific
weights by gender
using growth
Prenatal variables
from questionnaire
and medical records;
postnatal via
Analyzed birth
weight adjusted
for maternal
age, pre-
N = 360
singleton
newborn
s
Birth weight; mean
difference in weight at 6,
12, 24, and 36 months
This document is a draft for review purposes only and does not constitute Agency policy.
A-632 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
cohort) South
Korea
Mother and
Childrens
Environmental
Health Study
recruited from
hospital;
information on
demographics
and housing
characteristics
via
questionnaire.
Infants followed
at 6 (n=262), 12
(n=234), 24
(n=199), and 36
months (n=92).
pregnancy, 3 days.
Categorized into
two groups below
and above the 75th
percentile and also
continuous with log
transformation.
Mean (SD) 0.082
(0.052) mg/m3,
geometric mean
0.067, 75th
percentile 0.106
mg/m3. Correlation
between TVOCs and
formaldehyde 0.22,
p<0.01.
standard for
Korean children.
questionnaire and
interview. ETS slightly
higher in low
formaldehyde group
but was not
associated with
weight.
pregnancy body
mass index,
education level,
parity, gender,
gestational age
at birth and
residential
factors.
Analyzed
postnatal
weight at each
visit using
multiple linear
mixed models
adjusted for
gender, birth
order,
breastfeeding
and education.
Overall
SB IB
Cf
Oth
Confidence
Medium
Hospital-based cohort with
potential selection bias,
notable attrition over time
Occupational Studies
Axelsson et al.,
1984 (case-cohort)
laboratory work
University
laboratory
workers
identified via
payroll (born
1935 and after,
worked in lab
1968-79); 95%
response; birth
register records
compared for
Self-report (Y/N)
during 1st trimester,
open question; likely
exposure
misclassification, no
information on
intensity or
frequency of
exposure
Spontaneous
abortion & birth
defects; self-report
& birth registry,
1968-1979.
Spontaneous
abortion verified
using hospital
records or via
recall.
Miscarriage rate not
associated with
smoking before or
during pregnancy
(raises uncertainty
about data quality);
inverse association of
solvent exposure with
pregnancy number,
age, and work shift
Unadjusted
analyses for
formaldehyde
Only 10
exposed
pregnane
ies;
potential
ly
unstable
risk
estimate
s
Spontaneous abortion
Birth defects
SB IB Cf Oth
Overall
Confidence
Not
informative
Open-ended question
unreliable for exposure
classification; uncertainty
regarding data quality:
miscarriage rate higher in
nonresponders and not
associated with smoking
This document is a draft for review purposes only and does not constitute Agency policy.
A-633 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
respondents and
nonrespondents.
Ericson et al., 1984
(nested case
control)
Controls (2 per
case) selected
from other
infants in registry
born in 1976 of
laboratory
worker; 50% of
cases and 20% of
controls
responded about
exposure
Lab work identified
by occupational
code in 1975 census;
self-report on work
during pregnancy &
exposure to agents
(open question);
potential
misclassification; no
information on
intensity or
frequency of
exposure
Perinatal deaths (<
7 days) & birth
defects; National
Birth Register,
1976
Controls selected
randomly within
same age (5-yr
categories) and parity
stratum as case. No
information on
smoking or other risk
factors.
Unadjusted
analyses for
formaldehyde
3
exposed
cases
Perinatal deaths
Birth defects
SB IB Cf Oth
Overall
Confidence
Not
informative
Open-ended question
unreliable for exposure
classification; low response
regarding exposure; very
few exposed cases
Hemminki et al.,
1982 (cohort)
hospital staff
Recruited from
nursing staff
working in
sterilizing units
(exposed to
sterilizing agents,
x-rays, or
anesthetic gases)
or auxiliary units
(referent) in all
general hospitals;
Response > 90%
for both exposed
and referent;
recall likely not
related to
exposure
Exposure (Y/N) at
beginning of
pregnancy to
specific agents
assigned by
supervising nurse,
blind to case status,
possible exposure
misclassification,
particularly for
earlier years. No
information on
intensity and
frequency.
Spontaneous
abortion: self
report on
pregnancies, 1951
-1981;
questionnaire &
hospital discharge
register
Regression adjusted
for several risk
factors, and
presented risk
estimates for other
sterilants (ethylene
oxide,
glutaraldehyde).
Formaldehyde results
not adjusted for
other sterilants.
Binary logistic
regression for
exposure
(yes/no)
adjusted for
age, parity,
decade of
pregnancy,
smoking habits,
alcohol, and
coffee
consumption
50
exposed
pregnane
ies (6
spontane
ous
abortion
s); 1,100
unexpos
ed
pregnane
ies (121
spontane
ous
abortion
s)
Spontaneous abortion
SB IB Cf Oth
Overall
Confidence
Low
Assumed sterilant use was
same throughout period; no
information on intensity
and frequency of
formaldehyde exposure
(exposure
misclassification —
decreased sensitivity); no
adjustment for other
sterilants; adjustment for
parity may introduce bias;
This document is a draft for review purposes only and does not constitute Agency policy.
A-634 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
Hemminki et al.,
1985 (case control)
nursing staff
Birth outcomes
from national
discharge register
linked to national
occupational
register.
Occupation
identified for >
87% of exposed
and referent.
Selected hospital
nurses.
Occupation during
1st trimester
identified by head
nurses at all general
hospitals in Finland
plus exposure (Y/N)
to listed substances
(used sterilizing
agent or sterilized
instruments;
formaldehyde
included in list);
potential exposure
misclassification; no
information on
intensity or
frequency.
Spontaneous
abortion & birth
defects, 1973-
1979; hospital
discharge register
linked to personnel
register
Referent with healthy
birth selected from
same hospital as
cases; matched on
age; not adjusted for
other risk factors or
other workplace
exposures
Conditional
logistic
regression.
Unadjusted OR
presented for
FA; no
statistical tests
6
exposed
cases for
spontane
ous
abortion
3
exposed
cases for
birth
defects
Spontaneous abortion and
birth defects
Overall
Confidence
Low
No information on intensity
or frequency (exposure
misclassification —
decreased sensitivity); very
small number of exposed
cases
John et al., 1994
(case control)
cosmetologists
Recruited from
license registry
(currently and
formerly
employed), 74%
with eligible
pregnancy, data
obtained for
71.5% of cases,
74% live births;
restricted
analysis to full-
time workers
Self-report;
response to closed
list (Y/N & frequency
of use), no ambient
measurements;
relevant exposure
period: 1st
trimester;
pregnancies while
full-time
cosmetologist
Spontaneous
abortion, 1983 -
1988, most recent
pregnancy
(decreased
sensitivity because
of time-lapse bias).
Self-report verified
by positive
pregnancy test or
medical care
Regression adjusting
for several risk
factors plus other
work exposures
among full-time
cosmetologists
Adjusted OR,
95% CI,
unconditional
logistic
regression
adjusting for
previous
pregnancy loss,
mother's age at
conception, &
mother's
cigarette
smoking during
1st trimester
67 cases,
351
controls
Spontaneous abortion
SB IB Cf Oth
Overall
Confidence
Medium
Selection of most recent
eligible pregnancy
(decreased sensitivity); no
ambient measurements;
adjustment for previous
pregnancy loss may
introduce bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-635 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
during 1st
trimester.
Lindbohm et al.,
1991 (registry
linkage)
paternal
occupation
Identified all
pregnancies
between 1/1/76
-12/31/77 and
5/1/80 -4/30/82,
excluded
maternal age <
12 and > 50 yr
and missing data
on occupation,
industry or SES
Industry/occupation
code based on
national census;
assignments by
industrial hygienist
(IH) using database
on chemical
exposures and
concentrations;
potential
misclassification into
low and mod/hi, and
exposure window
during
spermatogenesis for
paternal exposure
Spontaneous
abortion identified
in hospital
discharge register
that occurred
during a 2-yr
period close to
census
Adjusted for age, SES,
& maternal exposure
Linear logistic
7,772
regression
unexpos
adjusted for
ed SA,
age, SES, and
820
maternal
potential
exposure to
low, 139
reproductive
moderat
hazards; risk
e/high
odds ratio
comparing
exposed to
unexposed
Spontaneous abortion
Overall
Confidence
Low
Industry/occupation coding
has low specificity;
potential exposure
misclassification and
imprecise assignment of
exposure period to period
of spermatogenesis
relevant to identified
pregnancy
Saurel-Cubizolles
Recruited
et al., 1993
operating room
(cohort,
nurses at 18
retrospective)
hospitals
operating room
(exposed) and
nurses
randomly from
nurses in other
departments
from same
hospital
(unexposed);
data collection in
both groups
Self-reported
exposure (Y/N) to
anesthetics, formol,
& ionizing radiation
during 1st trimester.
No information on
intensity and
frequency.
Ectopic pregnancy:
self-report by
interview.
Interviewed 1987-
1988
Exposed and referent
matched for age,
duration of service,
sex, occupation, and
hospital. Formol
exposure associated
with exposure to
anesthetics. No info
on pelvic
inflammatory disease
but association with
formaldehyde not
likely.
Chi-square
15
analysis for
ectopic
formol; no
pregnane
multivariate
ies of
analyses
734
pregnane
ies; 1
exposed
case
Ectopic pregnancy
SB IB
a
Oth
Overall
Confidence
Low
¦
Small sample size and
unadjusted analyses. No
information on intensity
and frequency of
formaldehyde exposure
(exposure
misclassification —
decreased sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-636 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
conducted the
same
Saurel-Cubizolles
et al., 1994
(cohort,
retrospective)
operating room
nurses
Recruited
operating room
nurses at 18
hospitals
(exposed) and
randomly from
nurses in other
departments
from same
hospital
(unexposed);
data collection in
both groups
conducted the
same
Self-reported
exposure (Y/N) to
anesthetics, formol,
& ionizing radiation
during 1st trimester.
No information on
intensity and
frequency.
Spontaneous
abortion and birth
defects
(malformations
ICD-9): self-report
by questionnaire.
First pregnancy in
or after 1970;
interviewed 1987-
1988
Exposed and referent
matched for age,
duration of service,
sex, occupation, and
hospital.
Formol exposure
associated with
exposure to
anesthetics
Chi-square
analysis for
formol; no
multivariate
analyses
72
spontane
ous
abortion
s (9.4%);
22
pregnane
ies with
birth
defects
(3.4%);
14 major
malform
ations
(2.2%)
Spontaneous abortion and
birth defects
SB IB Cf Oth
Overall
Con fi ctence
Not
informative
No information on intensity
and frequency of
formaldehyde exposure
(exposure
misclassification —
decreased sensitivity).
Possible confounding by
other exposures and no
adjustment (stronger
associations observed for
spontaneous abortion and
anesthetics and ionizing
radiation, but not all birth
defects); no consideration
of impact of gravidity on
risk
Shumilina, 1975
(cross sectional)
cotton textile
workers
Unable to assess;
selection &
response rate not
reported
Range reported;
sampling protocol
not described;
analyzed categories
of textile finishers
and sorted
compared to
saleswomen
Reproductive &
pregnancy history
including LBW.
Gynecological
exam and self-
report; methods
NR
Job demands among
textile workers and
referent (sales
women) were
different; shift work
with standing and
elevated ambient
Prevalence &
SD; incomplete
Reproductive disorders,
and complications of
pregnancy, low birth
weight
SB
IB
a
Oth
Overall
Confidence
Not
informative
1 1 1
This document is a draft for review purposes only and does not constitute Agency policy.
A-637 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
temperature for
exposed
Not informative; reporting
deficiencies; potential
confounding by conditions
in the workplace
Steele & Wilkin,
1996 (cohort,
retrospective)
veterinarians
Recruited from
graduation rolls;
85% of eligible
graduates.
Graduated 1970-
1980; survey
1987
Self-reported
exposure (Y/N) to
specific agents for
specific jobs,
defined exposed
pregnancy if
estimated time of
conception occurred
during years of job
where exposure also
was reported. 81%
reported exposure
to formaldehyde; no
information on
intensity or
frequency of
exposure.
Spontaneous
abortion occurring
for pregnancy
started after
graduation from
veterinary college,
< 20-week
gestation, self-
reported
Compared exposed
pregnancies to
employed women
who reported no
exposure to
formaldehyde or not
employed during
pregnancy. Adjusted
for other risk factors,
but not other
workplace exposures
Unconditional
logistic
regression
adjusting for
maternal age,
gravidity,
previous SA,
alcohol, and
smoking. Also
evaluated
height, previous
stillbirth, and
previous
induced
abortions.
1,757
exposed
pregnane
ies, 482
not
exposed
Spontaneous abortion
SB IE CF Oth
Overall
Confidence
Low
No information on intensity
and frequency of
formaldehyde exposure
which would likely be
variable among
veterinarians (exposure
misclassification —
decreased sensitivity).
Adjustment for gravidity
and previous spontaneous
abortion may introduce
bias.
Seitz & Baron,
1990 NIOSH Health
Hazard
Investigation
(retrospective
cohort)
clothing
manufacturer
Response: 98% of
current
employees, 18%
of former
employees
employed 1984
or after. Possible
survivor bias.
Potential for
Air sampling 1987,
full shift personal
breathing zone for 5
task areas, 14 area
samples full shift in
several locations;
perhaps not
representative of
earlier years;
Self-report,
questionnaire,
pregnancy while
working at plant
compared to
employment at
other locations or
at home;
miscarriage (not
Authors stated no
differences among
groups for other risk
factors including
smoking, alcohol, use
of medications, and
presence of diseases
(diabetes)
Compared
miscarriage and
pregnancy
outcomes by
employment
status when
pregnancy
occurred
(employed at
Pregnane
ies
among
current:
19 at
Rockcastl
e, 71
other,
Miscarriage
SB
IB
Cf Oth
{Vera II
Confidence
Not
informative
¦
No comparison group
(compared pregnancy
This document is a draft for review purposes only and does not constitute Agency policy.
A-638 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
selection bias for
comparisons with
pregnancy
outcomes while
at home (away
from null); not a
concern for
comparisons with
employment at
other locations
during
pregnancy.
exposure range:
TWA 0.17-0.57
mg/m3; job status
when pregnancy
occurred.
defined), birth
outcomes, self-
report
(questionnaire).
Former workers
sent questionnaire
in 1984.
Rockcastle or
other) or at
home. RR (95%
CI), Fisher's
exact test
206
home
history during and not
during job but could not
account for gravidity in that
kind of analysis). Limited
exposure assessment for
earlier years.
Stucker et al., 1993
(birth weight)
Stucker et al., 1990
(spontaneous
abortion) (cohort,
retrospective)
nursing staff
Recruited all
female daytime
nursing staff, <
45 yr old and
currently working
in selected units.
87% response
among all
daytime nursing
staff
Current and
previous jobs; self-
report by interview;
dates of each prior
pregnancy and dates
of occupational
exposure to
cytostatic drugs,
anesthetic agents,
and formaldehyde.
Exposure based on
exposure during or
before the
pregnancy. No
information on
intensity or
frequency of
exposure.
Self-report by
interview
(spontaneous
abortion, birth
weight, small for
gestation age).
Interview 1985-
1986. Mean time
since exposed and
referent
pregnancies,
respectively, was 5
and 10 years
(potential for
differential recall
and
misclassifi cation?)
Exposed and referent
were all female day
time nursing staff
No analyses
were presented
for
spontaneous
abortion.
Linear
regression for
birth weight &
formaldehyde
association,
adjusted for
gestational age;
not adjusted for
other work
exposures;
other work
exposures
(quantitative
results not
reported, just
reported as
348
births
among
formalde
hyde
exposed
pregnane
ies; # of
spontane
ous
abortion
s not
reported
Birth weight
spontaneous abortion
SB IB a Oth
Overall
Confidence
Not
informative
Inclusion of exposure
before pregnancy of
uncertain relevance for
birth weight. No
information on intensity
and frequency of
formaldehyde exposure
(exposure
misclassification —
decreased sensitivity).
Quantitative results not
presented for formaldehyde
for birth weight analysis; no
This document is a draft for review purposes only and does not constitute Agency policy.
A-639 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
"not
significant")
results presented for
spontaneous abortion
analysis
Taskinen et al.,
1994 (case-control)
laboratory workers
Recruited from
payrolls & union
rolls, 82.4%
response,
reduced
likelihood of
selection bias; 2
referents per
case with a live
birth and no
registered SA, 4
referents per
congenital
malformation
case, study
population
restricted to age
20-34 yr,
referents
matched to case
for age (24 mo)
at conception
and year at end
of pregnancy
Self-report, focus on
1st trimester;
exposed &
frequency, reviewed
by industrial
hygienist; calculated
exposure index
based on reported
quantity used,
frequency (#
hours/day and #
days/week), and use
of fume hood
Spontaneous
abortion: hospital
discharge register,
1973-1986
Smoking, alcohol and
employment status
considered a priori,
plus other factors
(parity, previous
miscarriages, febrile
diseases during
pregnancy and used
contraception) with
OR > 1.5 or p value <
0.05; no other work
exposures; possible
confounding by
xylene exposure,
majority of formalin
exposed also exposed
to xylene (OR 3.1)
Conditional
logistic
regression
adjusted for
factors listed in
confounding
column
206 SA
cases,
329
referents
; 36
malform
ation
cases,
105
referents
Spontaneous abortion
SB IB Of Oth
Overall
Confidence
Low
*
Adjustment for parity and
previous miscarriage may
introduce bias; lack of
adjustment for xylene, an
exposure associated with
the spontaneous abortion
and formalin exposure;
evaluation of increasing
frequency of use a strength.
Taskinen et al.,
1999 (cohort,
retrospective)
woodworkers
Recruited from
woodworker's
union (not only
current workers)
reducing
TWA assigned using
measurements and
reported time at
task, sampling
protocol not
Pregnancies
identified from
national birth
register 1985-
1996; live birth.
FDR: Regression
adjusting for several
risk factors plus
phenols, FDR for
dusts & wood dusts
TTP: Discrete
proportional
hazards
regression and
likelihood ratio
Not
exposed
N=367
Low
N= 119
Time-to-pregnancy
SB. IB a Oth
Overall
Confidence
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-640 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
test, FDR (95%
Medium
CI), adjusted for
N=77
employment,
High
smoking and
N=39
alcohol
consumption,
irregular
52
menstrual
spontane
cycles, and # of
ous
children.
abortion
Spontaneous
cases (in
abortion:
women
Unconditional
with
logistic
same
regression,
workplac
odds ratios,
e as
adjusted for
time-to-p
age,
regnancy
employment,
analysis)
smoking and
alcohol, #
exposed cases
not reported
Confidence
likelihood of
survivor bias,
64% returned
questionnaire;
evaluated
exposure
response trend;
period of recall
1-11 years. Not
an optimal design
for spontaneous
abortion: women
with no live
births but at risk
for spontaneous
abortion were
not included.
reported, JEM is a
more robust
exposure
assessment; focused
on 6 months prior to
pregnancy forTTP
relevant exposure
window; evaluated
risk by glove use in
high exposure
group; Exposure
range: 0.01-1.23
mg/m3. Applied
formaldehyde
concentrations from
a comparable
workplace when
data was missing
(missing data was
differential by
exposure level; high
31%, moderate 61%,
and low 46%)
Analysis limited to
first pregnancy
filling criteria; TTP
(FDR): self-report
(question: did
woman get
pregnant during
first menstrual
cycle when not
using
contraception?
Second? Or how
many
months/years?)
Left censoring:
excluded 38
pregnancies as a
result of
contraception
failure & 28 whose
TTP started before
the first job in the
branch.
any previous SA:
self-report
were > 1 in low
exposure category &
equal to 1 (1.02 &
0.93) in middle &
high categories; SA:
reported that other
exposures were not
associated
Expect some error in
individual exposure
assignments
Spontaneous abortion
Overall
Confidence
Medium
Exposures during critical
exposure period(s) for
spontaneous abortion were
not estimated.; excluded
women with no live birth
(missing spontaneous
abortions to women with
no live births)
This document is a draft for review purposes only and does not constitute Agency policy.
A-641 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
Wang et al., 2012
(cohort,
retrospective)
wood processing
100% of eligible
recruited couples
participated; did
not describe
recruitment or
sampling frame;
included if
married males,
Chinese Han
ethnicity, had
formaldehyde
exposure for at
least 24 months;
excluded couples
with possible
nonwork
exposure to
formaldehyde
(i.e., newly
remodeled
homes), or wives
with other
exposures to
reproductive
toxicants &
pregnancies prior
to formaldehyde
exposure
Measurements in
factories,
monitoring on 3
occasions during
different periods;
self-report of
workplace, work
tasks & hours/day
exposed to
formaldehyde; daily
mean exposure =
mean concentration
multiplied by % of
time exposed to
formaldehyde
(referenced
Taskinen et al.,
1999). JEM is a
more robust
exposure
assessment. Did not
report
formaldehyde
estimates; relevant
exposure period:
gametogenesis
Prolonged
time-to-pregnancy
(> 12 months),
spontaneous
abortion, birth
outcomes
(preterm birth,
LBW, sex ratio,
birth defects);
semi-structured
interview using
questionnaire;
data analysis for
most recent
pregnancy;
potential under-
ascertainment
because
interviewed male
partners.
Left censoring: 106
excluded because
wife's pregnancy
began before
exposed
employment
Exposed and referent
matched on age,
married men & from
same area (salesmen
and clerks); exposed
and referent were of
similar age, BMI,
educational level,
income, smoking,
alcohol, frequency of
intercourse.
Confounding
considered: age, BMI,
education, income,
smoking, alcohol, and
frequency of
intercourse.
Adjusted for other
risk factors but not
for other work
exposures (e.g., dust,
phenols)
Logistic
regression,
paternal
exposure risk;
adjusted OR,
95% CI;
compared low
versus high
formaldehyde
exposed.
Comparison of
means
(referent, low,
and high)
exposure,
ANOVA; crude
and adjusted
regression
coefficients and
95% CI; OR and
95% CI for
abnormal
sperm
parameters;
reported results
of all analyses
Did not
report #
exposed
and
referent
cases
Time-to-pregnancy
SB IB Cf Oth
Overall
Confidence
Medium
Exposure levels not
reported (but robust
assessment method).
Dichotomized
time-to-pregnancy in
analysis (low sensitivity).
Spontaneous abortion
birth defects
SB IB Cf Oth
Overall
Confidence
Medium
Exposure levels not
reported (but robust
assessment method).
Other workplace exposures
in woodworking industry
(solvents) have been
associated with the
spontaneous abortion but
not accounted for; analysis
of most recent pregnancy:
possible selection for live
births (time-lapse bias) and
possible impact of gravidity
This document is a draft for review purposes only and does not constitute Agency policy.
A-642 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
on spontaneous abortion
risk
(Wang et al..
2015)
(cohort,
retrospective)
wood processing, 7
industrial sites
Recruited men
aged 23-40 yrs of
age, Chinese Han
ethnicity, and
formaldehyde
exposed at least
24 months.
Excluded men
who lived in
newly built or
recently
decorated house,
men with genital
malformations or
other chronic
diseases.
Comparison: age-
matched male
Han population
volunteers living
in same area
(salesmen and
clerks) not
Referenced Wang et
al. (2012); sampling:
25-minute samples
at 3 times on one
workday, same day
as investigation .
Exposure
information based
on workplace, work
tasks, work duration
and time. Exposure
index based on
formaldehyde
concentration
(mean of 3 samples)
times exposed work
time during work
day times exposure
duration (years).
Two categories with
cutpoint at median.
Concentrations:
Exposed 0.22 - 2.91
Semi-structured
interview using
questionnaire; no
change in lifestyle
or environments 6
mo prior to semen
collection; genital
examination.
Semen sample
within 2 weeks of
exposure
sampling, after a 2-
7 day abstinence.
Semen analysis
within 60 min by
two technicians
using same
apparatus
(computer assisted
semen analysis),
blinded.
Parameters:
semen volume,
Addressed via design,
sex, SES, education,
age. Variables
included in models:
age, body mass index,
education, income,
smoking, drinking,
and abstinence
duration. No
evaluation of other
organic solvents such
as phenol or wood
preservatives.
Multiple linear
regression of In-
transformed
semen
parameters and
logistic
regression of
abnormal
semen
parameters;
reported results
for all
parameters
analyzed
124
(62.3%)
recruited
, eligible
and
agreed
to
participa
te. 75 of
199
eligible
refused
to
provide
sample.
No data
for 10,
W=114
N=81
referents
(40.5% of
eligible),
no
SB IB Cf Oth
Overall
Confidence
Medium
Other workplace exposures
in woodworking industry
(solvents) have been
associated with sperm
motility but not accounted
for; however otherwise
strong design and analysis,
including evaluation of
increasing exposure-
response relationship.
This document is a draft for review purposes only and does not constitute Agency policy.
A-643 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
exposed to
formaldehyde or
other
reproductive
toxicants.
mg/m3, exposure
index 4.54-195.08,
median 56.55;
referent 0 - 0.02
mg/m3.
sperm
concentration,
total sperm count,
sperm progressive
motility and total
sperm motility;
kinematic
parameters (WHO
laboratory manual,
2010), velocity,
linearity,
displacement
measures.
semen
data for
5, N=76
Ward et al., 1984
(cross-sectional)
autopsy service
Groups similar:
exposed and
referent all from
university
(exposed =
autopsy service;
referent = other
medical
branches)
Reported ranges for
TWA and
concentration; area
and personal
breathing zone.
Exposure range:
TWA 0.75-1.62
mg/m3
Sperm
abnormalities
assessed every 2-3
months (3 samples
collected for
standard sperm
parameters); hand
scoring of
morphology (no
QC data)
Matched on sex, age,
tobacco, alcohol, and
recreational drug use
No statistical
analyses; EPA
could compare
prevalence
11 men
per
exposure
group
Sperm parameters
SB IB a Oth
Overall
Confidence
Low
Small sample size;
uncertainty regarding
reliability of morphology
scoring
Zhu et al., 2005
(pregnancy cohort)
laboratory work
Danish National
Birth Cohort, 30-
40% of all
pregnancies, first
pregnancy and
laboratory
technician
(hospital,
university,
Self-report at
gestational weeks
12 - 25 (median 17
weeks), laboratory
work processes
during pregnancy
and 3 months
before conception;
JEM exposure index:
Self-report of TTP
(0-2 months, 3-5
months, 6-12
months, >12
months);
fecundability ratio
Demographic
characteristics of
laboratory
technicians and
teachers comparable
(maternal age,
gravidity, history of
spontaneous
abortion, smoking,
Fecundability
ratios analyzed
within the
exposed group
(exposure index
1-5 vs >=6)
using discrete-
time survival
analysis;
Exposed
N=829,
referent
N=6,250
Time-to-pregnancy
SB IB Cf Oth
Overall
Confidence
Low
Categorized
time-to-pregnancy
(decreased precision),
This document is a draft for review purposes only and does not constitute Agency policy.
A-644 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Consideration
of participant
selection and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely confounding
Analysis and
completeness
of results
Size
Confidence
medical industry,
food industry or
public services),
77.5% initial
cohort; referent
teachers, 73.9%
initial cohort;
entered cohort at
weeks 12-25
(median 17)
exposure level (low
or medium assigned
to work process by
study investigators)
times frequency of
contact.
Formaldehyde: Low:
processed human
blood or tissues,
worked with
experimental
animals or
microorganisms;
Medium: prepared
slides for
microscopy.
Exposure index did
not include use of
protective measures
(40 - 64% used
exhaust/flow
bench). Exposure
tool was not
validated for
formaldehyde
alcohol, BMI, paternal
job). Possible
confounding by other
exposures in lab
adjusted for
covariates listed
in confounding
column
missed pregnancies that
ended before 1st interview.
Variation in probability or
intensity of formaldehyde
exposure possible for work
processes across different
types of labs, did not
account for large
proportion of participants
who used protective
measures to prevent
inhalation exposure. JEM
was not validated for
formaldehyde.
Zhu et al., 2006
(cohort study)
laboratory work
Members of the
Danish National
Birth Cohort, 30-
40% of all
pregnancies, first
pregnancy and
laboratory
technician
Self-report at
gestational weeks
12 - 25 (median 17
weeks), laboratory
work processes
during pregnancy
and 3 months
before conception;
Birth outcomes:
preterm birth,
small for
gestational age,
major
malformations
Demographic
characteristics of
laboratory
technicians and
teachers comparable
(maternal age,
gravidity, history SA,
smoking, alcohol,
Cox regression
within the
exposed group
(exposure index
1-5 vs >6),
hazard ratios
for fetal loss
and
Late fetal
loss:
exposed
9/
unexpos
ed 106;
preterm
birth:
Preterm birth
small for gestational age
major malformations
Overall
Confidence
Low
This document is a draft for review purposes only and does not constitute Agency policy.
A-645 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Consideration
Reference,
of participant
Analysis and
setting, and
selection and
Exposure measure
Outcome
Consideration of
completeness
design
comparability
and range
measure
likely confounding
of results
Size
Confidence
(hospital,
JEM exposure index:
BMI, paternal job).
malformations;
exposed
Variation in probability or
university,
see Zhu et al. (2005)
Possible confounding
logistic
41,
intensity of formaldehyde
medical industry,
above
by other exposures in
regression,
unexpos
exposure possible for work
food industry or
lab
odds ratios for
ed 317;
processes across different
public services),
other
SGA:
types of labs, did not
95% of eligible;
outcomes;
exposed
account for large
referent
adjusted for
80,
proportion of participants
teachers, 95% of
covariates listed
unexpos
who used protective
eligible
in confounding
column
ed 700;
major
malform
ations:
exposed
56,
unexpos
ed 379
measures to prevent
inhalation exposure. JEM
was not validated for
formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
A-646 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
Animal Studies
Only in vivo inhalation exposure studies are used for hazard identification and dose-
response assessment These studies were conducted in inhalation chambers under controlled
experimental conditions. Studies that exposed animals to formaldehyde via other routes were not
included because they are expected to result in significant distribution of formaldehyde past the
portal of entry, which does not occur to a great extent with inhalation exposures.
Evaluation of experimental studies
The experimental animal studies were each assigned confidence ratings of: High, Medium,
or Low Confidence, and "Not Informative" based on an evaluation of the experimental details for
each study and an expert judgement related to predefined criteria for 1) exposure quality, 2) test
animals, 3) study dosing, 4) endpoint evaluation, and 5) data considerations and statistical analysis
(described in Appendix A. 1.1.). The studies designated as "Not informative" included those with
documented chemical co-exposure (in addition to inhaled formaldehyde) that might have
compromised the developmental or reproductive outcomes evaluated, or those that did not present
sufficient information to fully assess the study methods or test results for assessments critical to
study interpretation. The studies judged to be "Not informative" are not discussed in the
Toxicological Review.
Due to the known developmental hazard of methanol, studies failing to use an appropriate
test article (see Appendix A. 1.2) or that did not provide a full characterization of the test substance
were automatically assigned a rating of "Low Confidence", and may be deemed "Not Informative" if
additional study limitations are identified.
In addition to the general criteria discussed in Appendix A.I.I., considerations specific to
the evaluation of potential developmental or reproductive system effects were also evaluated:
• The potential contribution of species and strain-related differences in reproductive
schedules and outcome sensitivity were considered. The age of the animals, life stage, and
critical windows of exposure and assessment were evaluated for potential influence on
study results.
• The power of the study (group size, and sample size for specific endpoints) was considered.
Typical standards for guideline developmental and reproductive toxicity studies (i.e.,
preferably at least 20 dams/group) may not always be relevant to the endpoint-targeted
studies published in the literature. Negative studies with less than 10 test subjects per
group were considered to be "Low confidence."
• Random assignment of animals to exposure groups or to a specific assessment subgroup,
"blinding" to study group, or other procedures that were applied with the intent of
mitigating potential bias was preferred.
This document is a draft for review purposes only and does not constitute Agency policy.
A-647 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for Formaldehyde—Inhalation
• Studies were examined for evidence of severe overt toxicity in parental animals or
offspring, and the potential influence of maternal toxicity on fetal or postnatal offspring
outcomes was considered.
• In general principle, methodologies used to assess specific endpoints were evaluated in
comparison to published standards, guidance, and/or guidelines, although developmental
and reproductive toxicity database contained no guideline studies conducted under strict
Good Laboratory Practice regulations.
• The intent and focus of the study was considered when evaluating limitations in study
design because it is recognized that not all available studies are designed to screen for a
wide array of developmental or reproductive outcomes. Sometimes only part of the data
from a study might be deemed adequate.
• Presentation of detailed methodological information was necessary, given the complexity of
studies that assess developmental and reproductive outcomes, and the potential for small
variation in study design to have an impact on study outcome.
• Inclusion of adequately characterized quantitative and/or qualitative data to support study
conclusions was considered critical to the evaluation of study quality. The report was
examined to determine if the litter was considered the primary unit of analysis for offspring
data.
Additional considerations that might influence the interpretation of the usefulness of the
studies during the hazard synthesis are noted, including limitations such as a short exposure
duration or the use of only one test concentration or concentrations that are all too high or too low
to provide a spectrum of the possible effects, as well as study strengths such as very large sample
sizes or particularly robust endpoint protocols; however, this information typically did not affect
the study evaluation decisions.
If the conduct of the experimental feature was considered to pose a substantial limitation
that is likely to influence the study results, the cell is shaded gray; a "+" is used if potential issues
were identified, but these are not expected to have a substantial influence on the interpretation of
the experimental results; and a "++" denotes experimental features without limitations that are
expected to influence the study results. Specific study details (or lack thereof] which highlight a
limitation or uncertainty in answering each of the experimental feature criteria are noted in the
cells. For those experimental features identified as having a substantial limitation likely to
influence the study results, the relevant study details leading to this decision are bolded. Studies
are organized according to the general outcomes evaluated (i.e., gestation exposures and
developmental outcomes and reproductive outcomes) and then listed alphabetically.
This document is a draft for review purposes only and does not constitute Agency policy.
A-648 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-93. Study quality evaluation of developmental and reproductive toxicity animal studies
Experimental Feature Categories
The study detail(s) leading to identification of a major (bolded) or minor (italicized) experimental feature
limitation is indicated
Exposure Quality
Test Subjects
Study Design
Endpoint Evaluation
Data Considerations &
Statistical Analyses
Overall
Confidence
Rating
Regarding the
Use for MOA
(Main
limitations)
Expert
judgement
based on
conclusions from
evaluation of the
5 experimental
feature
categories
Criteria relevant
to evaluating the
experimental
details within
each
experimental
feature category
Exposure quality
evaluations (see
B.4.1.2) are
summarized
below;
robust;
adequate; and
shaded box: poor;
relevance of the
tested exposure
levels is discussed
in the hazard
synthesis
The species, sex,
strain, and age are
appropriate for the
endpoint(s); sample
size provides
reasonable power to
assess the endpoint(s);
overt systemic toxicity
is absent or not
expected, or it is
accounted for; group
allocations can be
inferred as
appropriate
A study focus was
developmental or
reproductive system
effects; the exposure
regimen is informative
for the tested
endpoint(s);
manipulations other
than formaldehyde
exposure are
adequately controlled'
Endpoint evaluates a
mechanism relevant to
humans"; protocols are
complete, sensitive,
discriminating, and
biologically sound;
experimenter bias
minimized
Statistical methods,
group comparisons, and
data presentation
(including variability) are
complete, appropriate,
and discerning; selective
reporting bias avoided
Gestation Exposures and Developmental Outcomes'"
Al-Saraj (2009)
Test article =
formalin; co-
exposure with
ivermectin
(anhelmintic)
+
7 control does and 26
FA-exposed does;
strain NR
Gestation day not
standardized via
cesarean section;
detailed offspring
evaluation methods not
provided
Only external
examination; no visceral
or skeletal evaluation of
newborn kits
Exposure during
gestation not well-
characterized; dose-
dependent data in dams
and offspring not
shown.
Litter incidences of
external findings not
provided; major
confounding factor: co-
exposure with
ivermectin, a known
Not informative
(Co-exposure to
ivermectin)
This document is a draft for review purposes only and does not constitute Agency policy.
A-649 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
developmental toxicant
in rabbits
Gofmekler
(1968)
Test article NC;
generation
method, analytical
method and
concentrations,
chamber type NR;
exposure regimen
poorly
characterized
+
N = 3 males and 12
females/group; source
and strain NR
+
Limited study design
focused on offspring
growth (body weight
and organ weight)
+
Methods were poorly
described but appeared
appropriate for the
evaluation of offspring
growth
Mean body and organ
weight data reported,
but no variance
provided; statistical
methods not described
although statistical
analysis was conducted.
Age at assessment of
offspring NR;
reproductive (maternal
and litter) data not
provided; overall
limited data reporting.
Low
(Test article NC,
exposure
generation,
animal
strain/source
NR; limited
description of
methods; limited
reporting)
Gofmekler and
Bonashevskaya
(1969)
Test article NC;
generation
method, analytical
method and
concentrations,
chamber type,
exposure regimen
NR
+
N = 12/group; source
and strain NR
+
Limited study design
focused on
developmental
anomalies, offspring
reproductive organ
weights, and
histopathology
+
Methods were poorly
described but appeared
appropriate for the
evaluation of.
Report contained only
verbal summary of
findings. No
quantitative data were
included in the paper
Not informative
(Test article NC,
exposure
generation,
animal
strain/source
NR; limited
description of
methods; limited
reporting)
Guseva (1972)iv
Test article NC;
generation
method, analytical
concentrations
NR; chamber type
NC; co-exposure
with formalin in
drinking water
N = 4/group; source
and strain NR
+
Limited study design
focused on reproductive
function,
developmental
anomalies and postnatal
maturation;
gonadotropic response
to pituitary emulsions,
and testicular nucleic
acids
+
Methods were poorly
described but some
appeared appropriate for
the evaluation of
reproductive function,
developmental anomalies
and postnatal maturation;
gonadotropic response
assay was not a standard
testing paradigm
Only nucleic acid
quantitative data (mean
and variance) were
reported; all other
results were described
verbally; statistical
methods not described
although statistical
analysis was conducted
Not informative
(Test article NC;
oral co-exposure
with formalin;
low N; some
experimental
methods and
data NR)
This document is a draft for review purposes only and does not constitute Agency policy.
A-650 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Kitaev et al.
(1984)d
Test article NC;
generation
method, analytical
concentrations
NR; chamber type
NC
+
N = 5-9/group; source
NR
+
Limited study design
focused on early
embryonic
development, organ
weights, and hormone
measures; time of day
the hormone measures
were taken NR
+
Methods were poorly
described but appeared
appropriate for the
evaluation of early
embryonic development,
organ weights, and
hormone measures
+
Group mean data and
variance presented for
embryos/rats; variance
shown in graphics for
organ weights and
hormone measures;
statistical methods not
described although
statistical analytical
results were described in
text. Statistical
significance NR for some
embryonic outcomes;
relative organ weight
and hormone measure
graphs appeared to be
hand-drawn
Low
(Test article NC;
limited
description of
methods)
Kum et al. (2007)
Test article =
formalin;
generation
method, analytical
concentrations NR
+
N = 6/group; source NR
+
Limited study design
focused on embryonic
and early postnatal
body and liver weights
and MOA (redox
enzymes)
+
Methods were poorly
described but appeared
appropriate for the
evaluation of embryonic
and early postnatal body
and liver weights
+
Group mean data and
variance presented;
maternal toxicity not
reported
Low
(Formalin;
limited
description of
methods;
maternal tox NR)
Martin (1990)
++
Test article =
paraformaldehyde
/
well characterized
exposure methods
+
N = 25 dams/group;
source NR
+
Study design described
as a "teratology study"
although few details
were provided
Methods were not
described; endpoints listed
in the statistical methods
section appeared
appropriate for a screening
level evaluation of
developmental toxicity
+
Inadequate reporting of
methods and
quantitative results. No
group mean data were
presented
Low
(Inadequate
reporting of
methods and
quantitative
results)
Monfared (2012
Test article NC;
generation
method, analytical
methods and
concentrations NR
++
N = 10 dams/group;
strain and source were
reported
+
Limited study design
focused on placental
weight, histopathology,
++
Methods were appropriate
for the evaluation of
placental weight,
+
Group mean placental
weight data and
variance presented;
photomicrographs
Low
(Test article NC;
maternal tox:
NR)
This document is a draft for review purposes only and does not constitute Agency policy.
A-651 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
and ultrastructural
pathology
histopathology, and
ultrastructural pathology
provided; maternal
toxicity not reported
Pushkina et al.
(1968)
Test article NC;
generation
method, analytical
method and
concentrations,
chamber type,
exposure regimen
NR
+
N = 10 females/group;
strain NR
+
Limited study design
focused on ascorbic acid
levels in dams, fetuses,
and placentas
Limited methodological
information provided
+
Group mean ascorbic
acid levels and variance
presented; statistical
methods not described
although statistical
analytical results were
noted in table
Not informative
(Experimental
methods NR)
Saillenfait et al.
(1989)
Test article =
formalin with 10%
methanol; well-
characterized
exposure methods
++
N = 25 dams/group;
strain and source
provided
++
Study design was
equivalent to a
guideline prenatal
developmental toxicity
study
++
Methods well described
and appropriate for a
screening level evaluation
of developmental toxicity.
++
Group incidence and
mean/variance data
presented
Low
(Formalin)
Sanotskii et al.
(1976)
Test article NC;
generation
method, analytical
method and
concentrations
NR; chamber type
NC
N = 334 total females
(females/group NR);
strain and source NR
Limited study design
only evaluated
pregnant vs.
nonpregnant dams (did
not evaluate
reproductive or fetal
parameters)
Limited methodological
information provided
Inadequate reporting of
methods and results (no
primary or mean data
presented); statistical
methods not described
although statistical
analytical results were
noted in text
Not informative
(Experimental
methods and
data NR)
Senichenkova
(1991)
Test article NC;
generation
method, analytical
method and
concentrations
NR; chamber type
NC
N = 137 total dams
(dams/group NR);
strain and source NR
+
Study design focused on
in utero developmental
outcomes (mortality,
growth, visceral,
skeletal outcomes),
select open field
neurotoxicity
measurements in
juveniles, and blood
acid-base status
+
Limited methodological
information provided for
tests conducted; apparent
methods appropriate for
the evaluation of in utero
developmental outcomes.
+
Group mean and
variance data presented;
maternal toxicity not
reported; statistical
methods not described
although statistical
analytical results were
noted in tables
Low
(Test article NC;
exposure
generation,
animal
strain/source, #
dams/group,
maternal tox NR;
limited
description of
methods)
This document is a draft for review purposes only and does not constitute Agency policy.
A-652 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Senichenkova
and Chetobar
(1996)
Test article NC;
generation
method, analytical
method and
concentrations
NR; chamber type
NC
N = 254 total dams
(dams/group NR);
strain and source NR
+
Control group co-
exposure to ethanol;
limited study design
focused on in utero
developmental
outcomes (external
anomalies and skeletal
delays) and blood acid-
base status
+
Limited methodological
information provided for
tests conducted; apparent
methods appropriate for
the evaluation of in utero
developmental outcomes.
+
Group mean and
variance data presented;
statistical methods not
described although
statistical analytical
results were noted in
tables; maternal toxicity
not reported
Low
(Test article NC;
exposure
generation,
animal
strain/source, #
dams/group,
maternal tox NR;
limited
description of
methods)
Sheveleva (1971)
Test article NC;
generation
method, analytical
method NR
+
N = 15 dams/group for
C-section, 6
dams/group for
delivery; strain and
source NR
+
Limited study design
focused on
developmental
parameters, body
weight spontaneous
mobility, temperature,
and hematology
parameters
+
Limited methodological
information provided for
tests conducted; apparent
methods appropriate for
the evaluation of
developmental
parameters.
+
Group mean and
variance data presented;
statistical methods not
described
Low
(Test article NC;
exposure
generation,
animal
strain/source
NR; limited
description of
methods)
Reproductive Outcomes
(Appelman et
al.. 1988)
++
Test article =
paraformaldehyde
; well
characterized
exposure methods
++
N = 40 males/group;
test animals
adequately
characterized
++
Study design focused on
comparison of
subchronic or chronic
exposures to rats with
undamaged or clinically
damaged nasal mucosa;
extensive tissue
evaluation
No indication if
histopathology was
performed on male
reproductive organs
Quantitative testes
weight data were not
presented in the study
results. No
histopathology findings
for male reproductive
organs were reported
Low
(No indication if
histopathology
performed on
male repro
organs;
quantitative
testes weights
not presented)
Golalipour et al.
(2007)
Test article NC;
generation
method NR; open
air exposures (i.e.,
not a controlled
chamber study)
N = 4 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testis
toxicity
++
Methods were appropriate
for the evaluation of testis
toxicity.
++
Group mean data and
variance presented
Low
(Test article NC;
open air
exposures; N =
4/group)
This document is a draft for review purposes only and does not constitute Agency policy.
A-653 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Han et al. (2013)
Test article NC;
generation
method, analytical
method and
concentrations
NR, static chamber
type
++
N = 10 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testis
toxicity and MOA
+
Methods were appropriate
for the evaluation of testis
toxicity.
+
Group mean testis
weight and seminiferous
tubule diameter data
reported but variance
not presented;
quantitative microscopy
findings not presented
Low
(Test article NC;
exposure
generation NR;
static chamber
used; limited
reporting of
study results and
group data)
Maronpot et al.
(1986)
Test article =
formalin; well-
characterized
exposure methods
++
N = 10/sex/group; test
animals adequately
characterized
++
Subchronic study with
limited in-life
observations and
extensive postmortem
evaluation
++
Methods were appropriate
for a screening level
evaluation of general
toxicity following
subchronic exposure; no
special emphasis on
reproductive organs
+
Selected incidence data
presented (survival,
histopathology); mean
body weight data did not
include variance; no
indication of statistical
data analysis
Low
(Formalin;
limited reporting
of methods and
results)
Ozen et al.
(2002)
++
Test article =
paraformaldehyde
; well
characterized
exposure methods
++
N = 7 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testis
toxicity and MOA
++
Methods were appropriate
for the evaluation of testis
toxicity
++
Group mean data and
variance presented
High
(None)
Ozen et al.
(2005)
+
Test article =
paraformaldehyde
; analytical
concentrations NR
++
N = 6 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testis
toxicity
++
Methods were appropriate
for the evaluation of testis
toxicity
++
Group mean data and
variance presented
High
(None)
Sapmaz et al.
(2018)
++
Test article =
paraformaldehyde
; well
characterized
exposure methods
+
N =7 adult males;
strain provided; source
not identified
+
Limited study design
focused on testis
toxicity and biomarkers
of oxidative stress; only
one paraformldehyde
test group
++
Methods were appropriate
for the evaluation of testis
toxicity
++
Group mean data and
variance presented
Medium
(Inadequate
information for
quantitative
analysis of
histopathology
data)
Sarsilmaz et al.
(1999)
+
++
+
++
+
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-654 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Test article =
paraformaldehyde
; analytical
concentrations NR
N = 10 males/group;
test animals
adequately
characterized
Limited study design
focused on testis
toxicity
Methods were appropriate
for the evaluation of testis
toxicity
Group mean data and
variance presented;
unable to determine
what the reported SD
represents for Leydig cell
numbers
(Inadequate
information for
quantitative
analysis of
histopathology
data)
Vosoughi et al.
(2012, 2013)
++
Test article =
paraformaldehyde
; well
characterized
exposure
methods;
analytical
concentrations
reported
++
N =12 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testis
toxicity, sperm
measures, and hormone
(LH, FSH, T) levels
++
Methods were appropriate
for the evaluation of testis
toxicity, sperm measures,
and hormone levels (LH,
FSH, T)
++
Group mean data and
variance presented
High
(None)
Wang et al.
(2013)
Test article NC;
generation
method, analytical
method and
concentrations
NR, static chamber
type
++
N = 10 females/group;
test animals
adequately
characterized
+
Limited study design
focused on ovarian
toxicity, estradiol (E2)
levels, and MOA
++
Methods were appropriate
for the evaluation of
ovarian toxicity and E
levels
++
Group mean data and
variance presented
(graphically) for E2 levels
and ovarian weights
Low
(Test article NC)
(Woutersen et
al.. 1987)
++
Test article =
paraformaldehyde
, generation
method, analytical
methods and
concentrations
reported, dynamic
whole-body
chamber
++
N = 40/sex/group; test
animals adequately
characterized
++
13-week subchronic
study
Report indicates that
testes and ovaries were
weighed at necropsy; no
indication if
histopathology was
performed on male or
female reproductive
organs
Quantitative
reproductive organ
weight data were not
presented in the study
results. No
histopathology findings
for reproductive organs
were reported
Low
(Limited
methods; no
data presented)
Xing et al. (2007)
Test article NC;
generation
method, analytical
method and
++
N =12 males and 24
females/group; test
+
Limited study design
focused on sperm
morphology,
++
Methods were appropriate
for the evaluation of sperm
++
Adequate reporting of
reproductive outcome
results (group incidence
Low
(Test article NC;
exposure
generation,
This document is a draft for review purposes only and does not constitute Agency policy.
A-655 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
concentrations,
chamber type NR
animals adequately
characterized
reproductive success,
and micronucleus assay
morphology and
reproductive outcome.
and mean data with
variance). Micronucleus
data not presented.
strain NR; high
exposure levels)
Zhou et al.
(2006a)
Test article NC;
generation
method and
analytical
concentrations
NR, static chamber
type
++
N = 10 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testes
weight and
histopathology, sperm
measures, and MOA;
co-exposure of one FA-
treated group with
vitamin E to assess
mediation effects
++
Methods were appropriate
for the evaluation of testes
weight and histopathology,
and sperm measures
++
Group mean data and
variance presented
(graphically for testes
weights); appropriate
statistical analysis of
data. Vitamin E co-
exposure group not
included in dose-
response assessment for
FA outcomes
Low
(Test article NC,
exposure
generation NR;
static chamber
used)
Zhou et al.
(2011a)
Test article NC;
generation
method, analytical
method and
concentrations
NR; static chamber
type, exposure
regimen poorly
described
++
N = 10 males/group;
test animals
adequately
characterized
+
Limited study design
focused on testes and
epididymal weight and
histopathology, sperm
measures, testosterone
(T) levels, and MOA
++
Methods were appropriate
for the evaluation of testes
and epididymal weight and
histopathology, sperm
measures, and T levels
++
Group mean data and
variance presented
(graphically for T levels)
Low
(Test article NC;
exposure
generation NR;
static chamber
used)
Zhou et al.
(2011b)
Test article NC;
generation
method, analytical
method and
concentrations
NR; static chamber
type, exposure
regimen poorly
described
++
N =12 males/group;
test animals
adequately
characterized
+
Limited study design
focused on epididymal
weight and
histopathology, sperm
measures, and MOA
++
Methods were appropriate
for the evaluation of
epididymal weight,
histopathology, and sperm
measures
++
Group mean data and
variance presented
(graphically)
Low
(Test article NC;
exposure
generation NR;
static chamber
used)
NR = Not Reported; NC = Not Characterized
Gradations of sufficiency based upon described criteria: ++ = meets sufficiency criteria; + = meets some sufficiency criteria
This document is a draft for review purposes only and does not constitute Agency policy.
A-656 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
A.5.9. Carcinogenicity: Respiratory Tract, Lymphohematopoietic, or Other Cancers
Systematic identification and evaluation of the literature database on studies examining the
potential for carcinogenicity following formaldehyde exposure was performed separately for the
following: (1) human studies of respiratory tract, lymphohematopoietic, or other cancers; (2)
experimental animal studies of respiratory tract (nasal) cancers; and (3) experimental animal
studies of LHP cancers. This section is organized accordingly.
Literature Search
Studies in Humans
A systematic evaluation of the literature database on studies examining the potential for
cancer in humans in relation to formaldehyde exposure was initially conducted in October 2012,
with yearly updates (see A.1.1 for searches through 2016; see Appendix E for details on a separate
Systematic Evidence Map thatupdates the literature from 2017-2021 using parallel approaches).
The search strings used in specific databases are shown in Table A94. Additional search strategies
included:
• Review of reference lists in the articles identified through the full screening process.
• Review of reference lists in the 2010 draft Toxicological Review for Formaldehyde (U.S.
EPA. 2010bl. the ATSDR toxicological profile of formaldehyde (ASTSDR, 1999), and the NTP
report on carcinogens background document for formaldehyde (NTP, 2010).
• Review of references in 11 review articles relating to formaldehyde and cancer, published
in English, identified in the initial database search.
Relevant studies were separated into upper respiratory tract (URT) cancers,
lymphohematopoietic (LHP) cancers, and other cancers (including brain, lung, pancreatic, etc.).
Inclusion and exclusion criteria used in the screening step are described in Table A-95.
Multiple review articles and meta-analyses have examined the epidemiologic evidence
informing potential associations between formaldehyde and cancer endpoints (e.g., Bachand et al.,
2010; Zhang et al., 2009; Bossetti et al., 2008; Collins and Lineker, 2004; Collins et al., 2001; Ojajarvi
et al., 2000; Collins et al., 1997; Blair etal., 1990). The vast majority of studies focused on cancers of
the URT and LHP system. Other cancers endpoints reported in the literature include bladder, brain,
colon, lung, pancreas, prostate, and skin. However, aside from cancer of the brain and lung few
studies showed any evidence of increased risks. Given the large number of studies available on
URT and LHP cancers, the other endpoints were not included in the hazard evaluation. As
numerous studies reported data on cancers of the brain or lung, a summary of the available studies
for each of these endpoints is provided in Appendix XX for information; however, a cursory review
of the available studies did not suggest any consistent association with formaldehyde exposure and,
as such, these endpoints were also not formally reviewed.
This document is a draft for review purposes only and does not constitute Agency policy.
A-657 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
For the hazard evaluation, the URT cancer endpoints were restricted to specific cancers (i.e.,
nasopharyngeal cancer, sinonasal cancer, cancers of the oro- and hypopharynx, and laryngeal
cancer). The specific LHP cancers that were formally reviewed were Hodgkin lymphoma, multiple
myeloma, myeloid leukemia, lymphatic leukemia. Non-Hodgkin lymphoma is a nonspecific
grouping of dozens of different lymphomas and classification systems for specific subtypes have
changed over time, complicating the synthesis of study results for this cancer type. If formaldehyde
is associated with particular non-Hodgkin lymphoma subtypes, then these studies might be not
sensitive enough to detect an association. As review articles and a cursory review of the available
did not suggest an association between formaldehyde exposure and non-Hodgkin lymphoma and,
as such, this endpoint was not formally reviewed.
After manual review and removal of duplication citations, the 624 articles identified from
database searches were initially screened within an EndNote library for relevance; title was
considered first, and then abstract in this process. Full text review was conducted on 271identified
articles. The search and screening strategy, including exclusion categories applied and the number
of articles excluded within each exclusion category, is summarized in Figure B4-8.
Based on this process, 59 studies were identified and evaluated for consideration in the
Toxicological Review.
Table A-94. Summary of search terms for carcinogenicity in humans
Database,
search date
Terms
PubMed
No date
restriction
"formaldehyde"! Majr] AND ("neoplasms"[AII Fields] OR "cancer"[AII Fields] OR
"leukaemia"[AII Fields] OR "leukemia"[AII Fields] OR "multiple myeloma"[ All Fields] OR
("multiple"[AII Fields] AND "myeloma"[AII Fields]) OR "multiple myeloma"[AII Fields] OR
"myeloma"[AII Fields] OR "lymphoma"[AII Fields] OR "nasopharyngeal neoplasms"! All Fields]
OR ("nasopharyngeal"[AII Fields] AND "neoplasms"[AII Fields]) OR "nasopharyngeal
neoplasms"[AII Fields] OR ("nasopharyngeal"[AII Fields] AND "cancer"[AII Fields]) OR
"nasopharyngeal cancer"[AII Fields] OR ("sinonasal" [All Fields] AND "neoplasms"[AII Fields])
OR "neoplasms"[AII Fields] OR "cancer"[AII Fields] OR "oropharyngeal neoplasms"! All Fields]
OR ("oropharyngeal"[AII Fields] AND "neoplasms"[AII Fields]) OR "oropharyngeal
neoplasms"[AII Fields] OR ("oropharyngeal"[AII Fields] AND "neoplasms"[AII Fields]) OR
"laryngeal neoplasms"! All Fields] OR ("laryngeal"[AII Fields] AND "neoplasms"[AII Fields]) OR
"laryngeal neoplasms"[AII Fields] OR ("laryngeal"[AII Fields] AND "cancer"[AII Fields]) OR
"laryngeal cancer"[AII Fields]) AND (Epidemiol*[AII Fields] OR "Case-control studies"[AII Fields]
OR "Cohort studies"[AII Fields] OR "Follow-up studies"[AII Fields] OR "Risk factors"[AII Fields])
Web of Science
No date
restriction
Lemmatization
"off
TS=formaldehyde AND (TS=neoplasms OR TS=cancer OR TS=leukaemia OR TS=leukemia OR
TS=multiple myeloma OR (TS=multiple AND TS=myeloma) OR TS=multiple myeloma OR
TS=myeloma ORTS=lymphoma ORTS=nasopharyngeal neoplasms OR (TS=nasopharyngeal
AND TS=neoplasms) OR TS=nasopharyngeal neoplasms OR (TS=nasopharyngeal AND
TS=cancer) OR TS=nasopharyngeal cancer OR (TS=sinonasal AND TS=neoplasms) OR
TS=oropharyngeal neoplasms OR (TS=oropharyngeal AND TS=neoplasms) OR
TS=oropharyngeal neoplasms OR (TS=oropharyngeal AND TS=neoplasms) OR TS=laryngeal
neoplasms OR (TS=laryngeal AND TS=neoplasms) OR TS=laryngeal neoplasms OR (TS=laryngeal
This document is a draft for review purposes only and does not constitute Agency policy.
A-658 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Database,
search date
Terms
AND TS=cancer) OR TS=laryngeal cancer) AND (TS=Epidemiol* OR TS=Case-control studies OR
TS=Cohort studies OR TS=Follow-up studies OR TS=Risk factors)
ToxNet (Toxline
and DART)
No date
restriction
English, not
including PubMed
Formaldehyde AND (neoplasms OR neoplasms OR cancer OR leukaemia OR leukemia OR
"multiple myeloma" OR (multiple AND myeloma) OR myeloma OR lymphoma OR
"nasopharyngeal neoplasms" OR (nasopharyngeal AND neoplasms) OR "nasopharyngeal
neoplasms" OR (nasopharyngeal AND cancer) OR "nasopharyngeal cancer" OR (sinonasal AND
neoplasms) OR "oropharyngeal neoplasms" OR (oropharyngeal AND neoplasms) OR
"oropharyngeal neoplasms" OR (oropharyngeal AND neoplasms) OR "laryngeal neoplasms" OR
(laryngeal AND neoplasms) OR "laryngeal neoplasms" OR (laryngeal AND cancer) OR
"laryngeal cancer") AND (Epidemiol* OR "Case-control studies" OR "Cohort studies" OR
"Follow-up studies" OR "Risk factors"))
Table A-95. Inclusion and exclusion criteria for evaluation of studies of cancer
in humans
Included
Excluded
Population
• Human
• Animals
Exposu re
• Exposure assessment
for formaldehyde
• Industries or
occupations known to
involve exposure to
formaldehyde
• Not formaldehyde
• Outdoor formaldehyde exposure
•
Comparison
•
• Case reports
Outcome
• Nasopharyngeal
cancer
• Sinonasal cancer
• Cancers of the oro-
and hypopharynx
• Laryngeal
• Specific
lymphohematopoietic
cancers (i.e., Hodgkin
lymphoma, multiple
myeloma, myeloid
leukemia, lymphatic
leukemia
• Bladder, colon, pancreas, prostate, and skin
• Brain and lung cancer studies were initially included
but were subsequently excluded from the systematic
review
• Non-Hodgkin lymphoma
Other
•
• Reviews, reports, letters, commentaries, meeting
abstracts, methodology papers
• Systematic evaluation of study quality
This document is a draft for review purposes only and does not constitute Agency policy.
A-659 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Cancer (Human) Literature Search
TJ
¦_
O
§
>•
3
a
E
o
u
O U OD
rn ^
rc
T3
QJ 03
"O «/> £
<
to
i/>
fO
-a
1
D
c
t{
+j
2
i-
-q
<
X
c
&
0)
I—
0)
1—1
o
3
Li.
to
PubMed
138 Articles
Other primary epidemiologic
studies of specific cancer
endpoints identified from
review articles, government
documents and public
comments.
2014 Literature Search Update
2015 Literature Search Update
2012 Title / abstract screen
Excluded because did not meet
criteria:
Notepi studies 248
Other 203
(language, methodology,
secondary literature)
Toxline, TSCATS,
& DART Web of Science
353 Articles
281 Articles
624 Articles
+ 15
(after duplicate removal from
merged dataset)
639 Articles
2013 Literature Search Update + 21
+ 32
+ 18
722 Articles
2016 Literature Search Update + 12
271 Articles
Full Text Screen
Excluded because did not meet
criteria: 212
59 Articles
59 articles "considered": the informativeness of each study was evaluated
Ultimately, 47 studies were included in the specific cancer endpoint tables.
Figure A-37. Literature search documentation for sources of primary data
pertaining to inhalation formaldehyde exposure and upper respiratory or
lymphohematopoietic cancers in humans through 2016 (see Appendix F for
details on the systematic evidence map updating the literature through 2021).
This document is a draft for review purposes only and does not constitute Agency policy,
A-660 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Supplemental Information for Formaldehyde—Inhalation
Studies in Animals
Based on the available evidence, separate systematic literature evaluations were conducted
as follows: 1) literature related to respiratory tract cancers and 2) literature related to LHP cancers.
These searches were initially conducted in October 2012, with yearly updates (see A.1.1 for
searches through 2016; see Appendix E for details on a separate Systematic Evidence Map that
updates the literature from 2017-2021 using parallel approaches). Similar to the evidence in
humans described above, the animal evidence for cancers other than those of the respiratory tract
and the LHP system were not systematically identified and reviewed; rather, these observations (as
identified through other, health effect-specific searches) were summarily described. For the
respiratory tract, the strategies are summarized in figure format (see Figures B-16); the search
strings used in specific databases are shown in table format (see Tables A-96), with additional
details of the process described below. For LHP cancer searches, the strategies are summarized in
figure format (see Figures B-17); the search strings used in specific databases are shown in table
format (see Tables A-98), with additional details of the process described below.
Respiratory tract (i.e., nasal) cancers in animals
A systematic evaluation of the literature database on studies examining the potential for
respiratory tract cancers following formaldehyde exposure was conducted through September
2016. This search strategy is summarized in Figure B-16; the search strings used in specific
databases are shown in Table A-96 with additional details of the process described below, and the
criteria used for inclusion and exclusion of studies during screening described in Table A-97.
Table A-96. Summary of search terms for respiratory tract cancers in animals
Database,
search date
Terms
PubMed
04/15/2013
No date
restriction
Formaldehyde [majr] AND (animal OR rodent OR rat OR mouse OR hamster) AND (nasal OR
nose OR buccal OR larynx OR lung OR mouth OR pharynx OR sinus OR trachea) AND (cancer
OR dysplasia OR neoplasia OR tumor OR carcinoma OR polyp OR cytotoxicity OR neoplastic OR
promoter OR pathology OR toxicity) NOT (formalin test OR formaldehyde fixation OR formalin
fixed OR formaldehyde fixed OR formalin-induced OR formaldehyde-induced)
Web of Science
03/08/2013
No date
restriction
Lemmatization
"off
Formaldehyde (title) AND (animal OR rodent OR rat OR mouse OR hamster) AND (nasal OR
nose OR buccal OR larynx OR lung OR mouth OR pharynx OR sinus OR trachea) AND (cancer
OR dysplasia OR neoplasia OR tumor OR carcinoma OR polyp OR cytotoxicity OR neoplastic OR
promoter OR pathology OR toxicity) NOT (formalin test OR formaldehyde fixation OR formalin
fixed OR formaldehyde fixed OR formalin-induced OR formaldehyde-induced)
This document is a draft for review purposes only and does not constitute Agency policy.
A-661 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
Table A-97. Inclusion and exclusion criteria for studies of nasal cancers in
animals
Included
Excluded
Population
• Experimental animals
• Not animal studies*
Exposu re
• Exposure to
formaldehyde for an
exposure duration
longer than short term
• Not related to formaldehyde* (e.g., other
chemicals)
• Mixture studies*
• Short study duration*
Comparison
• Inclusion of a
comparison group (e.g.,
pre- or postexposure;
no exposure; lower
formaldehyde exposure
level)
•
Outcome
• Endpoint evaluation
included nasal cancers
• Exposure or dosimetry studies*
• Related to formaldehyde use in methodology*
• Endpoint not nasal cancer*
Other
• Original primary
research article
• Not a unique, primary research article,* including
reviews, reports, commentaries, meeting abstracts,
duplicates, or untranslated foreign language
studies (these were determined to be off topic or
unlikely to have a significant impact based on
review of title, abstract, and/or figures).
• Related to policy or current practice (e.g., risk
assessment/management approaches or modeling
studies)
* Indicates criterion tags used in HERO for excluded studies (insert website link for chemical page)
Identification of additional articles
The reference lists of the review articles identified through the process described above
were manually screened (based on the criteria used for full text screening presented in
Figure B-16) for relevant articles (aka "snowball searching"). These were then compared against
the 229 articles identified from the computerized searches. No additional (0) relevant articles were
identified.
Manual screening for relevance: Title/Abstract/Full Text
The primary research articles identified were screened within an EndNote library for
relevance; title, abstract, and full text were assessed simultaneously. The number of articles
excluded within each category described in Table A-97 is shown in Figure B-15.
Overall, 19 articles were identified as relevant and are cited in the animal nasal cancer
section of the Formaldehyde Toxicological Review (see Appendix B.4 for individual study
evaluations).
This document is a draft for review purposes only and does not constitute Agency policy.
A-662 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
PubMed Web of Science
Filters \ "Tr /' \ s"biect ,/
\ Topic J \ Area /
Figure A-38. Literature search documentation for sources of primary data
pertaining to inhalation formaldehyde exposure and upper respiratory tract
(nasal) cancers in animals.
This document is a draft for review purposes only and does not constitute Agency policy,
A-663 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Lymphohematopoietic cancers (leukemia/lymphoma) in animals
2 A systematic evaluation of the literature database on studies examining the potential for
3 lymphohematopoietic cancers following formaldehyde exposure was conducted through
4 September 2016. This search strategy is summarized in Figure B-17; the search strings used in
5 specific databases are shown in Table A-98 with additional details of the process described below,
6 and the criteria used for inclusion and exclusion of studies during screening described in Table A-
7 99.
Table A-98. Summary of search terms for lymphohematopoietic cancers in
animals
Database,
search date
Terms
PubMed
04/15/2013
No date restriction
Formaldehyde [majr] AND (leukemia OR lymphoma OR hemolymphoreticular) AND (animal
OR rodent OR monkey) NOT (formalin test OR formaldehyde fixation OR formalin fixed OR
formaldehyde fixed OR formalin-induced OR formaldehyde-induced)
Web of Science
03/08/2013
No date restriction
Lemmatization "off"
Formaldehyde (title) AND (leukemia OR lymphoma OR hemolymphoreticular) AND (animal
OR rodent OR monkey) NOT (formalin test OR formaldehyde fixation OR formalin fixed OR
formaldehyde fixed OR formalin-induced OR formaldehyde-induced) (topic)
This document is a draft for review purposes only and does not constitute Agency policy.
A-664 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
Table A-99. Inclusion and exclusion criteria for studies of LHP cancers in
animals
Included
Excluded
Population
• Experimental animals
• Not animal studies*
Exposu re
• Exposure to
formaldehyde
• Not related to formaldehyde* (e.g., other
chemicals)
Comparison
• Inclusion of a
comparison group (e.g.,
pre- or postexposure;
no exposure; lower
formaldehyde exposure
level)
•
Outcome
• Endpoint evaluation
included LHP cancers
• Exposure or dosimetry studies*
• Related to formaldehyde use in methodology*
• Endpoint unrelated to LHP cancer*
Other
• Original primary
research article
• Not a unique, primary research article*, including
reviews, reports, commentaries, meeting abstracts,
duplicates, or untranslated foreign language
studies (these were determined to be off topic or
unlikely to have a significant impact based on
review of title, abstract, and/or figures).
* Indicates criterion tags used in HERO for excluded studies
Identification of additional articles
The reference lists of the review articles identified through the process described above
were manually screened (based on the criteria used for full text screening presented in
Figure B-17) for relevant articles (aka "snowball searching"). These were then compared against
the articles identified from the computerized searches to identify additional relevant articles.
Manual screening for relevance: title/abstract/full text
The primary research articles identified from database searches and evaluation of reference
lists in reviews, were screened within an Endnote library for relevance; given the relatively small
size of the database, title, abstract, and full text were assessed simultaneously. The number of
articles excluded within each category described in Table A-99 is shown in Figure B-17.
Overall, 4 articles were identified as relevant and are cited in the animal
lymphohematopoietic cancer section of the Formaldehyde Toxicological Review
(see Appendix B.4 for individual study evaluation
This document is a draft for review purposes only and does not constitute Agency policy.
A-665 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
ni
u
CO
Filters
PubMed
Major
\ Topic /
Web of Science
\ Subject /
\ Area /
>
CD
"D
0)
M
CL
E
o
u
•— ni
4-» =
If
« s
11/1
to
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Study Evaluations
Studies in Humans
The studies identified for inclusion in the review were evaluated using a systematic
approach to identify strengths and limitations, and to rate the overall confidence in the results. The
accompanying tables in this section document the evaluation of these studies (cohort studies, and
nested case-control studies within occupational cohorts, in Table A-105, and case-control studies in
Table A-106). Studies are arranged alphabetically by author within each table.
The focus of EPA's examination is on several specific types of upper respiratory tract (URT)
and lymphohematopoietic (LHP) cancer. The evaluation of LHP cancers includes four different
subtypes: myeloid leukemia (including monocytic leukemia), lymphatic leukemia, multiple
myeloma, and Hodgkin lymphoma. Among upper respiratory cancers, four different types are
reviewed: sinonasal (SNC), nasopharyngeal cancer (NPC), oro/hypopharyngeal cancer (OHPC), and
laryngeal cancer.Evaluation of Observational Epidemiology Studies of Cancer
The epidemiology studies examined occupational exposure to formaldehyde either in
specific work settings (e.g., cohort studies) or in case-control studies. The considerations with
respect to design, exposure assessment, outcome assessment, confounding and analysis differ for
these different types of studies, and are discussed in more detail below.
Each study identified by the literature search as potentially relevant to inform the causal
evaluation of whether formaldehyde exposure causes cancer was then evaluated and classified for
the study's ability to inform a hazard conclusion for a particular cancer outcome. Study evaluation
encompasses interpretations regarding a variety of methodological features (e.g., study design,
exposure measurement details, study execution, data analysis). Developing an outcome-specific
study evaluation for each cancer outcome encompasses two concepts: minimization or control of
bias (internal validity), and sensitivity/appropriateness (the ability of the study to detect a true
effect). The purpose of this step is not to eliminate studies, but rather to evaluate studies with
respect to potential methodological considerations that could affect the interpretation of or
confidence in the results.
1) Consideration of participant selection and comparability
• Whether there is evidence of selection into or out of the study (or analysis sample) that was
jointly related to exposure and to outcome.
For cohort studies, EPA considered the extent of follow-up, and the likelihood that
completeness of follow-up was related to exposure level. Most of the cohort studies
examining mortality data reported high rates of follow-up with respect to ascertainment
of vital status and ascertainment of cause of death (90-95% or higher); in some cases,
the latter figure (i.e., percentage of decedents with death certificates) was not provided
by the study authors. Two studies were able to obtain only 79% (Hayes et al., 1990) or
75% (Walrath and Fraumeni, 1983b) of the identified death certificates but as both
This document is a draft for review purposes only and does not constitute Agency policy.
A-667 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Supplemental Information for Formaldehyde—Inhalation
studies were of embalmers who were all considered to have been exposed to
formaldehyde, the absence of data (missingness) was considered to have been random.
For case-control studies, controls are optimally selected to represent the population from
which the cases were drawn (e.g., similar geographic area, socioeconomic status, and
time period). A variety of methods were used in the identified studies, including
random digit dialing and use of population registries. The interest and motivation to
participate is generally higher for cases than for controls, particularly in population-
based settings. A low participation rate of either or both groups does not in itself
indicate the occurrence of selection bias; a biased risk estimate is produced if exposure
and disease are jointly related to participation rates, but not if either is independent of
participation rates. For example, a bias is not necessarily produced if cases are more
likely to participate than controls; a bias can be produced, however, if cases with high
exposure are more likely to participate than cases with low exposure. Most of the case-
control studies were conducted using incident (or recently diagnosed) cases, with
participation rates ranging from approximately 75% to 99%. Participation among
population-based controls generally ranged from 75% to 85%, with higher rates seen in
some studies using with hospital-based. Differences in participation rates between case
and controls potentially related to exposure were considered to be more prone to be
biased (Armstrong et al., 2000). Certain studies used cases' next of kin to ascertain the
cases' occupational history from which the individual's exposure to formaldehyde was
derived. The difference in methods for ascertaining exposure histories thus differs
between deceased cases and the controls and creates a potential for selection bias (e.g.,
Vaughan et al., 1986a,b; Vaughan 1989; Yangetal., 2005).
• An uncommon issue related to potential selection bias was the "healthy worker effect" in
cohort studies where a working population compared to that of the general public—a bias
which can result in underestimates of any adverse effect of exposure. While this
phenomenon is generally considered to be a stronger influence in evaluation of
cardiovascular health endpoints, there is evidence that there can be a strong healthy
workers effect in studies of cancer endpoints (Sont et al., 2001). In cohort studies, the
potential for selection bias due to the healthy worker effect was assessed by examination of
the all cause cancer effect estimates; studies with estimates <90% of expected were judged
to be potentially biased towards lower overall cancer occurrence and lower levels of cases
detection resulting underestimates of any true effect Severe underestimates of <80% of
expected cases were noted as well (e.g., Hall etal., 1991; Harrington and Oaks, 1984; Levine
et al., 1984; Matanoski et al., 1989; Robinson et al., 1987; Stroup et al., 1986; Wesseling et
al., 1996).
• For some cancers, the reliance of cohort studies on death certificates to detect cancers with
relatively high survival may have underestimated the actual incidence of those cancers,
especially when the follow-up time may have been insufficient to capture all cancers that
may have been related to exposure. The potential for bias may depend upon the specific
survival rates for each cancer. Five-year survival rates vary among the selected cancers
(see Table A-100), from 86% for Hodgkin lymphoma (HL) to less than 50% for multiple
myeloma (MM), myeloid leukemia (ML), and oro/hypopharyngeal cancer. EPA considered
the likelihood of underreporting of incident cases to be higher for mortality-based studies of
HL and LL which may result in undercounting of incident cases and underestimates of effect
estimates compared to general populations (e.g., Hansen et al., 1984; Hansen and Olsen,
1995; Hayes etal., 1990; Mayr etal., 2010; Soletetal., 1989).
This document is a draft for review purposes only and does not constitute Agency policy.
A-668 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-100. Lymphohematopoietic and upper respiratory cancers: age-
Adjusted SEER incidence and U.S. death rates and 5-year relative survival by
primary cancer sitea
Cancer Site
Incidence Rate
(per 100,000)
2008-2012
Expected
Cases'5
2014
Mortality Rate
(per 100,000)c
2008-2012
Expected
Deaths'5
2014
5-Year
Survival (%)
2005-2011
Lymphohematopoietic Cancers
Hodgkin lymphoma (HL)
2.7
8,336
0.4
1,235
85.9
Multiple myeloma (MM)
6.3
19,451
3.3
10,189
46.6
Lymphatic Leukemia (LL)
6.6
20,377
1.9
5,866
77.6
Acute lymphatic leukemia (ALL)
1.7
5,249
0.4
1,235
67.5
Chronic lymphatic leukemia (CLL)
4.5
13,894
1.4
4,322
81.7
Other
0.4
1,235
0.1
309
80.6
Myeloid & monocytic leukemia (ML)
6.1
18,833
3.4
10,497
37.5
Acute myeloid leukemia (AML)
4.0
12,350
2.8
8,645
25.9
Chronic myeloid leukemia (CML)
1.7
5,249
0.3
926
63.2
Acute monocytic
0.2
617
0.0
0
23.5
Other
0.2
617
0.2
617
33.2
Upper Respiratory Tract Cancers
Nose, nasal, & middle ear6
0.7
2,161
0.1
309
55.3
Nasopharynx
0.6
1,852
0.2
617
59.6
Oropharynx
0.4
1,235
0.2
617
41.7
Hypopharynx
0.6
1,852
0.1
309
32.2
Larynx
3.2
9,880
1.1
3,396
60.6
incidence rates and 5-year survival from Surveillance, Epidemiology, and End Results (SEER), 18 areas. Results.
[http://seer.cancer.gov/csr/1975_2012/results_merged/topic_survival.pdf], last accessed August 14, 2015.
bEPA calculated the expected number of cases based on incidence rates applied to U.S. census population estimate
for 2014 of 308,745,538 (http://www.census.gov/search-
results.html?q=2014+population&page=l&stateGeo=none&searchtype=web)
CU.S. Mortality Files, National Center for Health Statistics, Centers for Disease Control and Prevention
dSEER 18 areas. Based on follow-up of patients into 2012.
eSEER does not publish specific data on sinonasal cancer which would be included in the published category
labeled "Nose, nasal & middle ear."
1 2) The reliance of case-control studies on prevalent cases rather than incident cases.
2 In order to accrue a sufficiently large population of rare cancer cases, some studies may
3 include cases which have been detected over a long period of time and thus include many prevalent
4 cases at the time of analysis. Restriction to only living cases may lead to over-representation of
5 cancer survivors or, if next of kin are used to provide proxy information on cases, the quality of that
6 data may then differ between cases and controls which can be a concern if differences may be
This document is a draft for review purposes only and does not constitute Agency policy.
A-669 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
related to exposure. Hence, EPA considers that there is some risk of selection bias in studies
examining prevalent cases (e.g., Armstrong et al., 2000; Mayr etal., 2010, Pesch et al., 2008; Yang et
al., 2005; Vaughan et al., 1986a, b; Vaughan 1989).
3) Evaluation of exposure assessment
At a minimum, exposed to formaldehyde may be inferred based on the specific occupations
(e.g., carpenter, embalmer, pathologist) or industry (e.g., production or use of formaldehyde resins,
wood-products, paper, textiles, foundries). Independent testing of various workplaces may provide
approximate exposure measurements and ranges for inferred exposures. Details in each study may
reveal the extent of exposure within occupational groups or at the individual-level based on job
histories. Some studies may have documented formaldehyde exposures using exposure monitors
or quantified the absolute or relative exposure for different tasks, which may be matched to
individual occupational patterns using 'job exposure matrices" or JEMs. The quality of the exposure
measure is evaluated with respect to the accuracy of the measures and their related potential for
exposure measurement error which can lead to "information bias." The overwhelming majority of
information bias in epidemiologic studies of formaldehyde stems from the use of occupational
records to gauge exposures with some degree of exposure misclassification or exposure
measurement error considered to be commonplace.
A primary consideration in the evaluation of these studies is the ability of the exposure
assessment to reliability distinguish among levels of exposure within the study population, or
between the study population and the referent population. A large variety of occupations are
included within the studies; some represent work settings with a high likelihood of exposure to
high levels of formaldehyde, and some represent work settings with variable exposures and in
which the proportion of people exposed is quite small. In the latter case, the potential effect of
formaldehyde would be "diluted" within the larger study population, limiting the sensitivity or
informative nature of the study. EPA categorized the exposure assessment methods of the
identified studies into four groups (A through D), reflecting greater or lesser degree of reliability
and sensitivity of the measures (see Table A-101). Outcome-specific association based on Group A
exposures were consider without appreciable information bias due to exposure measurement error
while those based on Groups B-D were considered to be somewhat biased towards the null.
Table A-101. Categorization of exposure assessmentmethods by study design.
Group
Cohort (and nested
case-control within cohort) studies
Case-control and cancer
registry-based studies
A
• Industrial settings with extensive industrial
hygiene data used to determine levels of
exposure (and variability within a worksite);
job exposure matrix takes into account
variability by time and job/task.
• Detailed lifetime job history, more
extensive than industry and
occupation codes, including
information about specific tasks and
setting, combined with job exposure
This document is a draft for review purposes only and does not constitute Agency policy.
A-670 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Group
Cohort (and nested
case-control within cohort) studies
Case-control and cancer
registry-based studies
• (Beane Freeman et al., 2013; Beane Freeman
etal., 2009)
• Highly exposed professions (embalmers) with
comparison to general population, or with
measures capturing variability within the
cohort
• (Hauptmann et al., 2009)
• (Haves et al., 1990)
• {Levine, 1984,}
• (Mevers et al., 2013)
• (Strouo et al., 1986)
• (Walrath and Fraumeni, 1983)
• (Walrath and Fraumeni, 1984)
matrix that takes into account variability
by time, setting, and job/task. Also
includes some kind of validation study or
congruence of ratings based on different
exposure ascertainment measures to be
equivalent to Group A cohort studies with
extensive industrial hygiene data.
•
• (none identified)
B
• Industrial settings with more limited industrial
hygiene data
• (Andielkovich et al., 1995)
• (Coggon et al., 2014; Coggon et al., 2003)
• {Edling, 1987,}
• (Frvzek et al., 2005)
• (Marsh et al., 2007; Marsh et al., 2002)
• {Ott, 1984,}
• Exposed professions (e.g., pathologists) with
comparison to general population, but that do
not have measures capturing variability within
the cohort
• {Bertazzi, 1989,}
• (Hall etal., 1991)
• (Harrington and Oakes, 1984)
• (Li etal., 2006)
• (Matanoski, 1989)
• Detailed lifetime job history, more
extensive than industry and
occupation codes, including
information about specific tasks and
setting, combined with job exposure
matrix that takes into account
variability by time, setting, and job/task.
• (Armstrong et al., 2000)
• (d'Errico et al., 2009)
• (Gerin et al., 1989)
• (Gustavsson et al., 1998)
• (Hildesheim et al., 2001)
• (Pesch et al., 2008)
• (Vaughan et al., 2000)
C
• {Band, 1987,}
• (Dell and Teta, 1995)
• Self-report of exposure
• (Boffetta et al., 1989)
• (Saberi Hosniieh et al., 2013)
• (Stellman et al., 1998)
• Lifetime job history coding based only
on industry and occupation; more
detailed information about specific
tasks and setting not included in
assessment of exposure potential (or,
information on what was collected
was not provided)
• (Blair et al., 2001)
• (Hansen and Olsen, 1995)
• (Laforest et al., 2000)
• (Luce et al., 2002)
• {Olsen, 1984,}
This document is a draft for review purposes only and does not constitute Agency policy.
A-671 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Cohort (and nested
Case-control and cancer
Group
case-control within cohort) studies
registry-based studies
• (Olsen and Asnaes, 1986)
• (Roush et al., 1987)
• (Shangina et al., 2006)
• (West etal., 1993)
• (Wortlev et al., 1992)
• (Yu etal., 2004)
• Self-report of exposure
• (Mavr et al., 2010)
• Lifetime job history, including
tasks/exposure information, but
analysis conducted only for job
categories rather than for an exposure
category
• (Teschke et al., 1997)
• Industrial settings that do not include data to
• Job history limited to information on a
distinguish variability in exposure (e.g., wood
single job (e.g., based on tax record,
workers, with no information on which
death certificate, medical record,
workers were exposed to formaldehyde;
census data)
textile workers with no formaldehyde
• (Heineman et al., 1992)
exposure measures), or that include few
• (Pottern et al., 1992)
people classified as exposed
• (Talibov et al., 2014)
• (Hansen et al., 1994) pharmaceuticals
• High proportion (> 40%) of next-of-kin
• (Hansen and Olsen, 1995) plant used
interviews
lkg/person/yr
• {Vaughan, 1989, 2823477;Vaughan,
• (Jakobsson et al., 1997) grinding stainless
1986a, ;Vaughan, 1986b,}
steel
• (Yang et al., 2005)
D
• (Malker et al., 1990) fiberboard plants
Methods of exposure assessment
• (Siew et al., 2012) any occupational
rated as higher quality but
exposure
downgraded due to validation by
• (Solet et al., 1989) pulp and paper
study authors.
mills
• (Berrino et al., 2003)
• (Robinson et al., 1987) plvwood mill
workers
• Wesseling, 1996,1986612} banana plant
workers
• Methods of exposure assessment rated as
higher quality but downgraded due to
methods used by study authors which were
likely to induce bias.
• (Checkoway et al., 2015)
1 Additional exposure measurement error may arise in circumstances when the time period
2 of exposure assessment is not well aligned with the time period when formaldehyde exposure
This document is a draft for review purposes only and does not constitute Agency policy.
A-672 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
could induce carcinogenesis that develops to a detectable stage (incident cancer) or result in death
from a specific caner. Epidemiology studies regularly explore the analytic impact of different
lengths of'latency periods' which may exclude from the analyses the formaldehyde exposure most
proximal to each individual's cancer incidence or cancer mortality. For analyses of the exposure-
related risks of solid tumors, it is commonplace evaluate latency periods of 10,15, or 20 years by
present results stratified by time since first exposure or to exclude (or in the parlance of
epidemiology, to "lag") exposures in the 10,15, or 20 years immediately prior to death from the
analyses so as to more accurately (potentially) describe what may be the more biologically relevant
window of exposure in time that could have caused carcinogenesis (sometimes called the
etiologically relevant time period). Analyses which do not evaluate latency, may be inducing
exposure measurement error by including irrelevant exposure and were considered to be
somewhat biased towards the null.
An understanding of the effects of exposure measurement error on the results from
epidemiologic analyses is important as it enables the reviewer to place these possible exposure
measurement errors in context. The effect of exposure measurement error on estimates of the risk
of cancer mortality potentially attributable to formaldehyde exposure depends upon the degree to
which that error itself may be related to the likelihood of the outcome of interest. Exposure
measurement error that is similar among both workers who died of a specific cancer, and those
who did not die of that cancer, is termed nondifferential exposure measurement error. Exposure
measurement error that is associated with the outcome (error that is differential with respect to
disease status) can cause bias in an effect estimate towards or away from the null, while
nondifferential exposure error typically results in bias towards the null (Rothman and Greenland.
1998).
4) Outcome measure
The diagnosis of cancers in epidemiologic studies has historically been ascertained from
death certificates according to the version of the International Classification of Diseases (ICD) in
effect at the time of study subjects'deaths [i.e., ICD-8 and ICD-9: (WHO, 1967; 1977)]. The most
specific classification of diagnoses that is commonly reported across the epidemiologic literature
has been based on the first three digits of the ICD code (i.e., Myeloid Leukemia ICD-8/9: 205)
without further differentiation (i.e., Acute Myeloid Leukemia ICD-8/9: 205.0)—although some
studies have reported results at finer levels. In the evaluation of the epidemiologic evidence for
upper respiratory cancers, four different types are reviewed: sinonasal cancer, nasopharyngeal
cancer, oro/hypopharyngeal cancer, and laryngeal cancer. In the evaluation of the epidemiologic
evidence for LHP cancers, four different subtypes are reviewed: myeloid leukemia (including
monocytic leukemia), lymphatic leukemia, multiple myeloma, and Hodgkin lymphoma. In
restricting the causal evaluation of LHP cancers to these four specific subtypes, another category of
This document is a draft for review purposes only and does not constitute Agency policy.
A-673 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
LHP cancer originating from white blood cells, which includes all lymphoma not classified as
Hodgkin was not evaluated.
In the review of study quality for cancer studies, the outcome measure was generally
considered to be accurate as the source of this information was typically from death certificates,
cancer registries, or hospitals. Some studies did provide additional information on histological
typing but the majority did not. Histological type can be informative in understanding the
epidemiologic evidence but the lack of such information was not judged as a major study limitation.
While it is true that death certificates and other administrative records can occasionally contain
errors, the impact of misclassification of outcome on epidemiologic results is to reduce precisions in
effect estimates and not to induce bias.
5) Consideration of likely confounding
EPA evaluated the potential for confounding based on exposures to identified risk factors
for specific, or related, cancers, whether those exposures were found to be risk factors in the
specific study and whether there was a known or likely correlation between those exposures and
formaldehyde. Information on the presence of potential confounders in a particular study was
gleaned from the study itself or from information from outside the study (e.g., information on
exposure levels from other sources).
Risk factors for LHP cancers include pharmaceuticals (chemotherapeutic drugs), biological
agents (e.g., viruses), radiation, and chemical exposures (Cogliano J Natl Cancer Inst 2011;103:1-
13). The primary agents of interest that were considered in the study quality review are the
potential occupational and environmental co-exposures that may be associated with formaldehyde
exposure as well as LHP cancers. Chemotherapeutic drug exposures were not expected to be
correlated with formaldehyde exposures during the etiologically relevant time period for
potentially formaldehyde-related carcinogenesis and were not considered as potential confounders.
Similarly, viral exposures and radiation exposures also were not expected to be correlated with
formaldehyde exposures except, possibly, among embalmers and pathologists who may be co-
exposed by deceased persons who had viral infections or had implanted radiation devices used in
chemotherapy. Each of the chemical and occupational exposures that were reported to be
associated with risks of LHP cancers (i.e., benzene, 1,3-butadiene, 2,3,7,8-tetrachlorodibenzo-para-
dioxin, ethylene oxide, magnetic fields, paint, petroleum refining polychlorophenols, radioisotopes
and fission decay products, styrene, tetrachloroethylene, tobacco smoking trichloroethylene;
Cogliano et al., 2011) was examined in the study quality review and evaluated as a potential
confounder of any association between formaldehyde and specific LHP cancers.
Risk factor for URT cancers include biological agents (e.g., viruses), radiation, and chemical
exposures (Cogliano J Natl Cancer Inst 2011;103:1-13). Viral exposures and radiation exposures
also were not expected to be correlated with formaldehyde exposures except, possibly, among
embalmers and pathologists who may be co-exposed by deceased persons who had viral infections
This document is a draft for review purposes only and does not constitute Agency policy.
A-674 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
or had implanted radiation devices used in chemotherapy. Each of the chemical and occupational
exposures which were reported to be associated with risks of URT cancers (i.e., acid mists, asbestos,
chromium VI, isopropyl alcohol production, leather dust, nickel compounds, radioisotopes and
fission decay products, rubber production, textile manufacturing, tobacco smoking, wood dust;
Cogliano et al., 2011) was examined in the study quality review and evaluated as a potential
confounder of any association between formaldehyde and specific URT cancers.
The specific chemical and occupational exposures, listed above, which were reported to be
associated with LHP or URT cancers are bolded in the lists of co-exposures in each study in the
Exposure Measure column of the study quality tables. This identifies any important co-exposures
which are then evaluated for their potential correlation with formaldehyde exposure to identify
potential confounders.
6) Analysis and results (estimate and variability)
Analyses should be appropriate with respect to study design. When analytic methods are
not matched to the study design, the expected impact on the results was evaluated. For cancer
endpoints, results that examined the effects of including various latency periods using lagged
exposure of strata of time since first exposure allow for the focus of results on different etiological
windows of time that may be more biologically relevant Studies that did not report results looking
at different latencies may be vulnerable to additional exposure measurement error as they evaluate
the effects of formaldehyde exposures during times that may not have any causal effects such as in
the years immediately preceding death.
7) Study sensitivity
Studies with small cases counts may have little statistical power to detect divergences from
the null but are not necessarily expected to be biased and no study is excluded solely on the basis of
cases counts as this methodology would excluded any study which saw no effect of exposure.
Therefore, cohort studies with extensive follow-up which reported outcome-specific results on a
number of different cancers, including very rare cancers such as NPC and SNC, are evaluated even
when few or even no cases were observed, if information on the expected number of cases in the
study population was provided so that confidence intervals could be presented to show the
statistical uncertainty in the associated effect estimated. For example, Coggon et al. (2014)
followed the mortality of 14,008 workers and yet expected only 1.7 deaths from nasopharyngeal
cancer in the exposed workers and observed just one resulting in an unstable estimated RR=0.38
(95% CI: 0.02-1.90). Meyers etal. (2013) followed the mortality of 11,043 workers and expected
only 1.33 deaths from nasopharyngeal cancer and did not observe any deaths, resulting in a SMR=0
(95% CI: 0-2.77). In general, cohort studies should have a sufficiently long follow-up period for any
exposure-related cancer cases to develop and be detected and ideally, allow for analyses of
potential cancer latency. Outcome-specific effect estimates from cohort studies with short follow-
This document is a draft for review purposes only and does not constitute Agency policy.
A-675 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
up could be uninformative depending on the size of the study population and the baseline
frequency of the cancer.
Outcome-specific evaluation of confidence in the precise effect estimate of an association
An outcome-specific evaluations classified with High confidence in the precise effect
estimate is expected to be without appreciable bias and thus represents an accurate estimate of any
reported association between formaldehyde exposures and the risks of cancer. These evaluations
are expected to have methodological features sufficiently sensitive to provide an adequate basis for
interpreting null or weak results as evidence of no or weak risk of cancer. Table A-102 identifies
the outcome-specific evaluations were classified with High confidence.
Table A-102. Outcome-specific effect estimates classified with High
confidence
Reference
Outcome-specific effect estimates
Confidence classification
Beane Freeman et al., 2009
Hodgkin Lymphoma
High
Beane Freeman et al., 2009
Larygeal cancer
High
Beane Freeman et al., 2013
Lymphocitic leukemia
High
Beane Freeman et al., 2009
Multiple myeloma
High
Beane Freeman et al., 2009
Myeloid leukemia
High
Beane Freeman et al., 2013
Nasopharyngeal cancer
High
Hauptmann et al., 2009
Multiple myeloma
High
Hauptmann et al., 2009
Myeloid leukemia
High
Meyers et al., 2013
Multiple myeloma
High
Meyers et al., 2013
Myeloid leukemia
High
An outcome-specific evaluation classified with Medium confidence in the precise effect
estimate may have some potential for residual bias, but the direction of the observed effect is
unaffected and the magnitude of any expected biases are limited. Thus, the observed effect
estimates represent a reasonable estimate of the association between formaldehyde exposures and
the risk of cancer, and are expected to be sufficiently sensitive to provide an adequate basis for
interpreting null or weak results as evidence of no or weak risk of cancer. Table A-103 identifies
the outcome-specific evaluations were classified with Medium confidence.
Table A-103. Outcome-specific effect estimates classified with Medium
confidence
Reference
Outcome-specific effect estimates
Confidence classification
Beane Freeman et al., 2009
Hodgkin lymphoma
Medium
Beane Freeman et al., 2009
Lymphocytic leukemia
Medium
Beane Freeman et al., 2013
Sinonasal cancer
Medium
Coggon et al., 2014
Myeloid leukemia
Medium
Coggon et al., 2014
Laryngeal cancer
Medium
This document is a draft for review purposes only and does not constitute Agency policy.
A-676 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for Formaldehyde—Inhalation
Reference
Outcome-specific effect estimates
Confidence classification
Coggon et al., 2014
Oro/hypopharyngeal cancer
Medium
Gerin et al., 1989
Hodgkin lymphoma
Medium
Hayes et al., 1990
Multiple myeloma
Medium
Hayes et al., 1990
Myeloid leukemia
Medium
Hauptmann et al., 2009
Lymphatic leukemia
Medium
Meyers et al., 2013
Oro/hypopharyngeal cancer
Medium
Walrath and Fraumeni
1983b
Myeloid leukemia
Medium
Walrath and Fraumeni 1984
Myeloid leukemia
Medium
Laforest et al., 2000
Oro/hypopharyngeal cancer
Medium
Luce et al. 2002
Sinonasal cancer
Medium
Olsen and Asnaes, 1986
Sinonasal cancer
Medium
Olsen et al., 1984
Nasopharyngeal cancer
Medium
Roush et al., 1987
Nasopharyngeal cancer
Medium
Roush et al., 1987
Sinonasal cancer
Medium
Vaughan et al., 2000
Nasopharyngeal cancer
Medium
West et al., 1993
Nasopharyngeal cancer
Medium
An outcome-specific evaluation classified with Low confidence in the precise effect estimate
is likely to have some residual bias, or may lack sensitivity to provide an adequate basis for
interpreting null or weak results as evidence of no or weak risk of cancer. For example, an
outcome-specific effect estimate based on fewer than five observed or expected cases of a particular
cancer would be classified with Low confidence based on a lack of sensitivity, even if there were no
appreciable biases. Another study classified with Low confidence might have relied on exposure
assessment methodologies that were unbiased, but nonspecific in nature so as to yield effect
estimates that were likely biased towards the null, and thus, underestimated any true effect.
Similarly, the lack of consideration of latency is a limitation as it may cause measurement error in
improperly including exposure of little biological relevance to cancer occurrence. Concern about
the potential for confounding is a limitation when a co-exposure is a known cause of a particular
cancer endpoint and may be correlated with formaldehyde exposure is a study. Selection bias may
be a limitation when survival rates are long as incidence cases may not be readily detected using
mortality statistics. In general, outcome-specific effect estimates that underestimate any true effect
may still inform a hazard conclusion. However, outcome-specific effect estimates that overestimate
any true effect cannot inform a hazard conclusion and are considered to be uninformative as are
outcome-specific effect estimates, which suffer from strong bias or a complex mixture of biases.
Tables A-105 and A-106 identify the outcome-specific evaluations that were classified with Low
confidence.
Exclusion of studies based judged to be uninformative for the evaluation of causation
In rare circumstances, studies initially judged to be potentially informative were further
evaluated and found to be uninformative. For example, studies of specific LHP subtypes, which
This document is a draft for review purposes only and does not constitute Agency policy.
A-677 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
mention formaldehyde or study the health of workers in an industry expected to be exposed to
formaldehyde but details of the study reveal only extremely limited exposure (Armstrong et al.,
2000) or virtually none at all (Li et al., 2006). Two outcome-specific associations were judged to be
uninformative due, in part, to potential manifestations of the healthy worker effect with
standardized mortality ratio for "all cancers" more than 30% below expected values (SMR<0.7: Hall
et al., 1991; Harrington et al., 1984). Another reason was that a study had co-exposures that are
likely to have been highly correlated with formaldehyde and were known risk factors for LHP
cancers and the independent effect of formaldehyde cannot be inferred (e.g., d'Errico et al., 2009;
Fryzek et al., 2005). Studies with co-exposures to known risk factors for LHP cancers that are not
likely to be highly correlated for formaldehyde or were not risk factor for the specific LHP subtype
in question are included and the potential for confounding is noted for evaluation in the causal
synthesis. Table A-104 identifies the outcome-specific evaluations were classified as
uninformative.
This document is a draft for review purposes only and does not constitute Agency policy.
A-678 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-104. Outcome-specific effect estimates classified as uninformative
Reference
Outcome-specific
effect estimates
Confidence
classification
Critical limitation(s)
Fryzek et al., 2005
Hodgkin lymphoma
Not informative
Confounding
Fryzek et al., 2005
Multiple myeloma
Not informative
Confounding
Hall et al., 1991
Hodgkin lymphoma
Not informative
Selection bias (healthy worker
effect)
Hansen et al., 1994
Hodgkin lymphoma
Not informative
Information bias (minimal exposure)
Hansen et al., 1994
Laryngeal cancer
Not informative
Information bias (minimal exposure)
Hansen et al., 1994
Multiple myeloma
Not informative
Information bias (minimal exposure)
Harrington and Oakes,
1984
Sinonasal cancer
Not informative
Selection bias (healthy worker
effect)
Li et al., 2006
Sinonasal cancer
Not informative
Sensitivity (minimal exposure)
Matanoski et al., 1989
Hodgkin lymphoma
Not informative
Selection bias and Information bias
Solet et al., 1989
Hodgkin lymphoma
Not informative
Multiple
Wesseling et al., 1996
Hodgkin lymphoma
Not informative
Multiple
Wesseling et al., 1996
Multiple myeloma
Not informative
Multiple
Armstrong et al., 2000
Nasopharyngeal
cancer
Not informative
Multiple
Berrino et al., 2003
Laryngeal cancer
Not informative
Confounding
d'Errico et al., 2009
Sinonasal cancer
Not informative
Confounding
Mayr et al., 2010
Sinonasal cancer
Not informative
Confounding
This document is a draft for review purposes only and does not constitute Agency policy.
A-679 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-105. Evaluation of occupational cohort studies of formaldehyde and cancers of the URT (NPC, SN, OHPC)
and LHP (HL, MM, LL, ML)
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
(Andielkovich
etal.. 1995)
United States
Cohort study
of iron
foundry
workers
working
during 1960-
1987 with
follow-up
through
1989.
3,929 male
workers
exposed to
formaldehyde
> 6 months.
Loss to follow-
up 1.3% (1.5%
of
2,032
unexposed
workers).
Median
follow-up =15
years.
Average
follow-up
=20.77 years.
All cancer
SMR = 0.99.
Individual-level exposure
(Yes/No), questionnaire
based on industrial hygienist
review of detailed work
histories; assignments based
on job title and industrial
hygiene data and information
on tasks and plants. Exposure
assessment blinded to
outcome.
Co-exposed to silica. Possibly
co-exposed to polycyclic
aromatic hydrocarbons,
nickel, and chromium.
Mortality:
underlying
cause of death
based on ICD-8
(Social Security
Administration
Pension Benefit
Information,
and National
Death Index).
HL: ICD201.
Higher survival
rates for HL
could
undercount
incident cases,
but median
follow-up is
more than 15
years.
Controlled for sex,
age, race, and
calendar-year
specific mortality
rates.
Nickel and
chromium are
associated with URT
cancers and would
likely be positively
correlated with
formaldehyde
exposure.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
Other co-exposures
are not known risk
factors for these
outcomes.
Exposed vs.
unexposed.
SMRs (95% CI).
Latency not
evaluated.
HL: 1
Larynx: 3
NPC: 0
SNC: 0
SB IB Cf Oth
Overall
•I
Exposure: Group B;
lack of latency
analysis
Confounding
possible for URT
cancers
Low power (few
cases)
SUMMARY:
HL, Larynx, NPC,
SNC: LOW \U
(Low sensitivity
Potential biases)
Band et al.,
1997
Canada
28,200 male
workers
employed at
least one year
Hire and termination dates
and type of chemical process
of pulping (sulfate vs. sulfite).
Individual exposure measures
Mortality:
underlying
cause of death
obtained from
All comparisons
adjusted for age and
sex.
SMRs (95% CI).
HL: 7
Larynx: 12
MM: 12
SB IB Cf Oth
This document is a draft for review purposes only and does not constitute Agency policy.
A-680 DRAFT-
-D0 NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Cohort study
of pulp and
paper
workers,
working
before 1950
with follow-
up through
1982.
by January
1950.
Loss to follow-
up < 6.5% for
workers
exposed to
the sulfate
process (67%
of original
cohort of
30,157 were
exposed to
the sulfate
process) and
loss to follow-
up < 20% for
workers
exposed to
the sulfite
process.
Average
follow-up
=19.42 years.
All cancer
SMP = 1.03.
not derived. As a profession,
workers were likely exposed
to formaldehyde.
Formaldehyde is known to be
an exposure for pulp and
paper mill workers: job-
specific exposures range from
0.2 to 1.1 ppm with peaks as
high as 50 ppm (Korhonen et
al., 2004).
Co-exposed to arsenic,
chlorophenols, sulfuric acid
mists, and chloroform.
Co-exposures to dioxin or
perchloroethylene are also
possible (Kauppinen et al.,
1997 IAOEH;70:119-127).
the National
Mortality
Database based
on ICD version
in effect at time
of death and
standardize to
ICD-9 version
HL: ICD 201
MM: ICD 203.
Higher survival
rates for HL
could
undercount
incident cases,
but average
follow-up is
more than 15
years.
Confounding not
evaluated.
Potential
confounders for
these outcomes
include
chlorophenols, acid
mists, dioxin, and
perchloroethylene
and would likely be
positively correlated
with formaldehyde
exposure.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
Other co-exposures
are not known risk
factors for these
outcomes.
Duration of
exposure
evaluated.
Latency
evaluated as
time since first
exposure.
Exposure: Group C
Confounding
possible for LHP
and URT cancers
SUMMARY:
HL, Larynx, MM:
LOW sU
(Potential biases)
(Beane
Freeman et
al., 2009));
Beane
25,619 workers
(12% female)
followed from
plant start-up or
first employment.
Individual-level exposure
estimates based on job
titles, tasks, visits to
plants by study industrial
hygienists who took
Mortality:
underlying
cause from
death
All comparisons
adjusted for
calendar year, age,
sex, and race.
Internal:
Poisson
regression; RR
(95% CI) by
exposure
HL: 27
MM: 59
LL: 37
ML: 48
IB Cf Oth
Overa I
This document is a draft for review purposes only and does not constitute Agency policy.
A-681 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Freeman,
2013
United States
Cohort study
of workers in
10 plants
using or
producing
formaldehyd
e, follow-up
through
2004.
Related
studies:
Initial 10
plant cohort
follow-up
through 1980
(Blair et al.,
1986,1987).
Second set of
10 plant
follow-ups
through 1994
(Hauptmann
et al., 2003,
2004).
Reanalysis of
1 plant
Deaths were
identified from the
National Death
Index with
remainder
assumed to be
living. Vital status
was obtained for
97.4%.
Median follow-up
42 years.
Average follow-up
=38.96 years.
All cancer SMR =
0.93.
2,000 air samples from
representative jobs, and
plant monitoring data
from 1960 through 1980.
Blinded to outcome.
Median cumulative
exposure was 0.6 ppm-
years (range = 0.0 -
107.4 ppm-years).
Co-exposed to
antioxidants, benzene,
carbon black, dyes and
pigments, melamine,
hexamethylenetetramin
e, phenols, plasticizers,
urea, wood dust.
Beane Freeman et al.
(2013) sampled cohort
members and found no
association between
smoking and
formaldehyde. Blair et
al., 1986 noted that
smoking habits among
this cohort did not differ
substantially from those
of the general
population.
certificates,
ICD-8.
HL: ICD201
MM: ICD203
LL: ICD 204
ML: ICD 205.
Larynx: ICD 161
NPC: ICD 147
SNC: ICD 160.
Higher survival
rates for HL and
LL could
undercount
incident cases,
but median
follow-up is
more than 42
years.
Checkowav
(2015)
AML: 205.0
CML: 205.1
Internal analysis
adjusted for pay
category.
For HL, MM, LL, ML:
Benzene is a
potential
confounder but was
controlled for.
For NPC, SN: Wood
dust is a potential
confounder but was
controlled for.
Eleven co-exposures
examined as
potential
confounders, but
none were found to
be confounders.
categories (4
levels), for
peak, average,
cumulative
exposures.
Latency was
evaluated.
External: SMRs
(95% CI).
Checkowav
(2015)
Cox PH
regression; HR
(95% CI) by
exposure
categories (4
levels collapsed
to 3 by
widening the
ref. cat. due to
small
numbers).
Latency was
evaluated.
Larynx: 48
NPC: 11
SNC: 5
Checkoway
(2015)
AML: 34
CML: 13
Exposure: Group
A
Low power for
SNC
SUMMARY:
SNC: MEDIUM
(Low sensitivity)
HL, Larynx, LL,
ML, MM, NPC:
HIGH
Checkowav
(2015)
SB IB Cf Oth D»era
Exposure Group
A (from Beane
Freeman et al.,
2009)
downgraded to
Group D based
on authors'
decision to
reclassify all
peak exposures <
2 ppm as
unexposed and
to reclassify peak
This document is a draft for review purposes only and does not constitute Agency policy.
A-682 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
(Marsh et al.,
2002, 2007).
Reanalysis of
Beane
Freeman et
al. (2009)
(Checkoway
et al., 2015).
Checkoway et al. (2015)
redefined peak
exposures in the
referent category to
include any exposures <2
ppm of hourly, daily,
weekly or monthly
frequency as well as
exposures > 2 ppm if
they occurred hourly or
monthly.
exposures > 2
ppm as
unexposed if
they were either
very rare or very
common.
SUMMARY:
AML, CML: LOW
(Potential bias
(Beane
Freeman et
al.. 2009));
Beane
Freeman,
2013
United States
Cohort study
of workers in
10 plants
using or
producing
formaldehyd
e, follow-up
through
2004.
Related
studies:
25,619
workers (12%
female)
followed from
plant start-up
or first
employment.
Deaths were
identified
from the
National
Death Index
with
remainder
assumed to
be living. Vital
status was
obtained for
97.4%.
Individual-level exposure
estimates based on job titles,
tasks, visits to plants by study
industrial hygienists who took
2000 air samples from
representative jobs, and plant
monitoring data from 1960
through 1980.
Blinded to outcome.
Median cumulative exposure
was 0.6 ppm-years (range =
0.0 -107.4 ppm-years).
Co-exposed to antioxidants,
benzene, carbon black, dyes
and pigments, melamine,
hexamethylenetetramine,
phenols, plasticizers, urea,
wood dust.
Mortality:
underlying
cause from
death
certificates,
ICD-8.
HL: ICD201
MM: ICD203
LL: ICD 204
ML: ICD 205.
Larynx: ICD 161
NPC: ICD 147
SNC: ICD 160.
Higher survival
rates for HL and
LL could
undercount
incident cases,
but median
All comparisons
adjusted for
calendar year, age,
sex, and race.
Internal analysis
adjusted for pay
category.
For HL, MM, LL, ML:
Benzene is a
potential
confounder but was
controlled for.
For NPC, SN: Wood
dust is a potential
confounder but was
controlled for.
Internal:
Poisson
regression; RR
(95% CI) by
exposure
categories (4
levels), for
peak, average,
cumulative
exposures.
Latency was
evaluated.
External: SMRs
(95% CI).
HL: 27
MM: 59
LL: 37
ML: 48
Larynx: 48
NPC: 11
SNC: 5
SB IB Cf Oth
Overall
Exposure: Group A
Low power for SNC
SUMMARY:
SNC: MEDIUM
(Low sensitivity)
HL, Larynx, LL, ML,
MM, NPC: HIGH
This document is a draft for review purposes only and does not constitute Agency policy.
A-683 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Initial 10
plant cohort
follow-up
through 1980
(Blair et al.,
1986,1987).
Second set of
10 plant
follow-ups
through 1994
(Hauptmann
et al., 2003,
2004).
Reanalysis of
1 plant
(Marsh et al.,
2002, 2007).
Median
follow-up 42
years.
Average
follow-up
=38.96 years.
All cancer
SMR = 0.93.
No information on smoking;
however, according to (Blair
et al., 1986), 'The lack of a
consistent elevation for
tobacco-related causes of
death, however, suggests that
the smoking habits among
this cohort did not differ
substantially from those of
the general population."
Beane Freeman et al. (2013)
report that among a sample
of 379 cohort members, they
"found no differences in
prevalence of smoking by
level of formaldehyde
exposure."
follow-up is
more than 42
years.
Eleven co-exposures
examined as
potential
confounders, but
none were found to
be confounders.
Bertazzi et al.,
1986. Italy
Cohort study of
Italian chemical
workers in
plant
producing
formaldehyde
resins.
1,332 male
workers
ever
employed
in the plant
between
1959 and
1980.
Deaths
were
identified
from vital
statistics
offices.
Individual-level exposure
estimates based on occupational
histories from the personnel
office with supplement
information from 350 employed
workers alive at the end of
follow-up in 1980.
5,731/20,366 (28%) person
years were considered to be
exposed to formaldehyde.
Other exposures included
styrene, xylene, toluene, and
methyl isobutyl ketone.
Death
certificates used
to determine
cause of deaths
from nasal
cancer (ICD-8).
Controlled for age,
sex and calendar
time.
Styrene is associated
with LHP cancers but
not URT cancers.
Other co-exposures
are not known risk
factors for this
outcome.
SMRs (95% CI).
Latency
evaluated.
SNC: 0 cases
SB IB Cf Oth
Exposure Group B
Low power
SUMMARY:
SNC: LOW vU
(Low sensitivity
Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-684 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Vital status
was 98.6%
complete.
Average
follow-up
=15.26
years.
All cancer
SMR =
1.54.
(Boffetta et
al., 1989).
United States
Nested
matched case
control of
MM within
general
population
cohort.
Baseline
enrollment in
1982 with bi-
annual
follow-up in
1984 and
1986.
508,637
men and
676,613
women
(57%) in
American
Cancer
Society's
Cancer
Prevention
Study II,
with
sufficient
data on
occupation.
Loss to
follow-up
1.5%.
Death
certificates
for 84% of
Self-report from baseline
questionnaire occupational
history, based on specific question
about exposure to formaldehyde
(Ever/Never).
Other exposures included
asbestos, chemicals, acids,
solvents, coal or stone dusts, coal
tar, pitch, asphalt, diesel and
gasoline exhausts, dyes, pesticides,
herbicides, textile fibers/dusts,
wood dust, X-rays, and radioactive
material.
Mortality:
underlying or
contributing
cause from
death
certificates
MM: ICD-9:
203.
Analysis
limited to
"incident"
cases (i.e.,
had not
indicated a
history of
cancer in
baseline
questionnaire
Matching controlled
for sex, age, ethnic
group, residence,
smoking, education,
diabetes, X-ray
treatment, farming,
pesticide, and
herbicide exposure.
Other co-exposures
were not associated
with LHP cancers.
Mantel-
Haenszel
matched OR
(95% CI).
Latency not
evaluated.
MM: 128(4
exposed)
SB IB Cf Oth
Exposure Group C
Lack of latency
analysis
Low power (few
exposed cases)
SUMMARY: LOW
(Low sensitivity
Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-685 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
deceased
subjects.
Four
controls
per case
were
matched
for age,
sex, ethnic
group, and
residence.
Coggon 2014
(with Coggon
2003)
Great Britain
Cohort study
of British
chemical
workers in
factories
using or
producing
formaldehyd
e, working
before 1940
with follow-
up through
2012.
Related
studies:
14,008 men in
six chemical
facilities.
Cohort
mortality
followed from
1941 until
December
2012.
Vital status
was 92%
complete.
Cause of
deaths was
known for
99% of 5,185
deaths
through 2000.
This figure
was not
provided on
Individual level categorical
exposure assessment based
on employment records
evaluated occupational
hygienist who classified job
titles according to their
exposure to formaldehyde
based on measurement made
after 1970 and workers' recall
of irritant symptoms prior to
1970. Background exposure
corresponded to <0.1 parts
per million (ppm), low
exposure to 0.1-0.5 ppm,
moderate exposure to 0.6-
2.0 ppm, and high exposure
to >2.0 ppm.
Blinded to outcome.
Each worker assigned the
highest level of exposure ever
Mortality:
underlying
cause from
death
certificates,
ICD-9.
HL: ICD201
ML: ICD205
MM: ICD203.
Larynx: ICD 161
MM: ICD 203
NPC: ICD 147
OHPC: ICD 146-
149 minus 147
SNC: ICD 160.
Note than HL
follow-up was
through 2000
Adjusted for
calendar year, age.
Styrene is associated
with LHP cancers but
not URT cancers.
Asbestos is
associated with URT
cancers, including
laryngeal cancer.
Authors stated that
the extent of co-
exposures was
expected to be low.
Potential for
confounding may be
mitigated by low co-
exposures.
SMRs (95% CI)
by
low/moderate
and high
exposure
categories.
Latency not
evaluated.
NPC: 1
SNC: 2
OHPC: 16
Larynx: 22
HL: 15
MM: 28
ML: 36
Note that
HL results is
from 2003.
SB IB Cf Oth
EH
Exposure: Group B
Lack of latency
analysis
Low power for NPC
and SN
SUMMARY:
NPC, SNC: LOW ^
(Low sensitivity
Potential bias 4/)
HL, Larynx, ML,
MM, OHPC:
MEDIUM vU
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-686 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
Participants
setting, and
and
design
selection
Initial follow-
7,378 deaths
up through
through 2012.
1981
(Acheson et
All cancer
al., 1984).
SMR= 1.10.
Second
follow-up
through 1989
(Gardner et
al., 1993).
Third follow-
up through
2000:
(Coggon et
al., 2003).
Coggon 2014
Internal
Great Britain
comparison
using nested
Nested case-
case-control
control study.
study within
cohort with
Related
10 controls
studies:
per case
Initial follow-
individually
up through
matched by
1981
facility,
(Acheson et
mortality
al., 1984).
status and age
within 2
Second
years.
follow-up
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
experienced (i.e., "ever highly
exposed"). Subjects' assigned
exposure grade may exceed
average workplace exposure.
Potential low-level exposure
to styrene, ethylene oxide,
epichlorohydrin, solvents,
asbestos, chromium salts, and
cadmium.
(Coggon et al.,
2003).
Higher survival
rates for HL and
LL could
undercount
incident cases,
but follow-up is
more than 50
years.
Individual level categorical
exposure assessment based
on employment records
evaluated occupational
hygienist who classified job
titles according to their
exposure to formaldehyde
based on measurement made
after 1970 and workers' recall
of irritant symptoms prior to
1970. Background exposure
corresponded to <0.1 parts
per million (ppm), low
exposure to 0.1-0.5 ppm,
moderate exposure to 0.6-
Incidence or
morality: cancer
registries and
death
certificates,
ICD-code in
effect at time of
diagnosis or
death. Cases
were either
incident
diagnoses,
underlying
cause of death,
or contributing
cause of death.
Matched analysis
controlled for facility
and age.
Styrene is associated
with LHP cancers but
not URT cancers.
Authors stated that
the extent of co-
exposures was
expected to be low.
Potential for
confounding may be
mitigated by low
ORs (95% CI) by
low, moderate,
high exposure
for less than
one year, and
high exposure
for one year or
more.
Latency
evaluated by
exposure
duration and
category at 5
years prior to
diagnosis or
Larynx: 53
Pharynx: 28
OHPC: 27
ML: 45
MM: 28
SB IB Cf Oth
a
Exposure Group B
Latency evaluation
likely to be under-
powered to detect
any effects beyond
a 5-year period.
SUMMARY:
Larynx, ML, MM,
OHPC: MEDIUM ^
This document is a draft for review purposes only and does not constitute Agency policy.
A-687 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
through 1989
(Gardner et
a I1993).
Third follow-
up through
2000 (Coggon
et al., 2003).
2.0 ppm, and high exposure
to >2.0 ppm.
Blinded to outcome.
Each worker assigned the
highest level of exposure ever
experienced (i.e., "ever highly
exposed"). Subjects' assigned
exposure grade may exceed
average workplace exposure.
Potential co-exposure to
styrene and solvents.
Larynx: 161
MM: ICD203
NPC: ICD147
OHPC: ICD 146-
149 minus NPC
SN: ICD 160.
extent of co-
exposures.
death for each
matched set.
(Dell and
Teta. 1995)
United States
Cohort study
of workers in
a plastics
manufacturin
g and
research and
development
facility which
made
phenol-
formaldehyd
e resins,
working
1946-1967
with follow-
5,932 white
men
employed for
at least seven
months.
Vital status
was 94%
complete.
Death
certificates
obtained for
98%.
Average
follow-up 32
years.
All cancer
SMR= 1.02.
Individual exposure measures
not evaluated. Only 111 men
(2%) had work assignments
involving formaldehyde.
However, as the plant
manufactured and used
formaldehyde since 1931, a
larger percentage may have
actually been exposed.
Variation in presumed
exposure by department and
pay status.
Co-exposures: acrylonitrile,
asbestos, benzene, carbon
black, epichlorohydrin, PVC
(vinyl chloride), styrene, and
toluene.
Mortality:
underlying
cause from
death
certificates, ICD
version in effect
at time of
death.
MM: ICD 203.
Adjusted for sex,
race, age, and
calendar-year.
Asbestos is not
associated with LHP
cancers.
Benzene and styrene
were not evaluated
as potential
confounders and
would likely be
positively correlated
with formaldehyde
exposure.
Potential for
confounding is
SMRs (95% CI)
by major
department.
Latency
evaluated with
exposure lag
times of 10 and
15 years.
MM: 8
SB IB a Oth
Overall
Exposure: Group C
Confounding
possible
Low power due to
rarity of exposure
SUMMARY: LOW
(Low sensitivity
Potential biases)
This document is a draft for review purposes only and does not constitute Agency policy.
A-688 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
up through
1988.
unknown but could
have inflated the
observed effect.
(Edling et al.,
1987b)
Sweden
Cohort study
of workers in
a production
plant making
abrasives
bound with
formaldehyd
e resins,
working 1955
to 1981 with
follow-up
through
1983.
521 male
workers
employed at
least 5 years.
Vital status
was 97%
complete.
All cancer
SMR = 0.93.
Whole cohort assumed to be
exposed with some
individual's exposed to high
peak exposures.
Manufacture of grinding
wheels bound by
formaldehyde resins exposed
company workers to 0.1-1
mg/m3 formaldehyde.
59 workers (11%) had
intermittent heavy exposures
to formaldehyde with peaks
up to 20-30 mg/m3.
Co-exposed to aluminum
oxide and silicon carbide.
Incidence (ICD-
8), from
National Cancer
Registry.
MM: ICD-203.
Controlled for sex,
age, and calendar-
year-specific
mortality rates.
Co-exposures are
not known risk
factors for this
outcomes.
SIRs (95%
CI).
Latency
not
evaluated
MM: 2
SB
IB Of Oth
Overall
Exposure: Group B
Latency not
evaluated
Low power
SUMMARY:
MM: LOW vU
(Low sensitivity
potential bias 4/)
Fryzek et al.,
(2005)
United States
Cohort
mortality
study of
workers in
motion
picture film
processing,
working 1960
to 2000, with
2,646 workers
(11% female)
employed at
least 3
months.
178 workers
(7%) excluded
for missing
work histories
or work
outside the
study period.
Individual-level occupational
histories were used to classify
workers in job families
matched to past industrial
hygiene surveys conducted in
house and by state program.
Formaldehyde used in "film
developing" and possibly in
'maintenance'. Personal and
area sample averaged 0.28-
0.29 ppm with range 0.06-
0.52.
Mortality:
underlying
cause from
death
certificates.
HL: ICD-9 201
MM: ICD-9 203.
Higher survival
rates for HL
could
undercount
Controlled for age,
sex, race, and time
period.
Perchloroethylene
may be a risk factor
for multiple
myeloma as may
hydroquinone which
is a metabolite of
benzene, a known
cause of LHP
cancers.
SMRs
(95% CI).
Decade of
exposure,
duration
of
exposure
and time
since first
exposure
were
HL: 0
MM: 2
SB IB Cf Oth
Overall
0
Exposure: Group B
Confounding likely
Low power
SUMMARY: NOT
INFORMATIVE
This document is a draft for review purposes only and does not constitute Agency policy.
A-689 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
follow-up
through
2000.
Vital status
obtained for
99.7%; cause
of death data
for 655 of 666
decedents
(98.3%).
Average
length of
follow-up
=20.58 years.
All cancer
SMR= 1.1.
Co-exposures included
methanol, methyl chloroform,
perchloroethylene, and
hydroquinone.
incident cases,
but average
follow-up is
more than 20
years.
Potential for
confounding is
unknown but could
have substantially
inflated the
observed effect due
to the high
correlation of these
exposures with
formaldehyde.
evaluated
Latency
was
evaluated
as time
since first
exposure.
Critical limitation:
Confounding
(Hall et al.,
1991)
Great Britain
Cohort study
of British
pathologists.
Related
studies:
Initial follow-
up through
1973
(Harrington
and Shannon,
1975);
4,512
pathologists
from the
Royal College
of
Pathologists
and the
Pathological
Society of
Great Britain
from
1974-1987.
Deaths among
those >85
years were
censored.
Vital status
As a profession, pathologists
were highly exposed to
formaldehyde as a main
ingredient in tissue fixative.
NIOSH (Industry Selection for
Determination of Extent of
Exposure, 1979) has reported
mean formaldehyde
concentrations of 4.35 ppm
with range (2.2-7.9).
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc, and
ionizing radiation.
Mortality: cause
of death =
Hodgkin
lymphoma, ICD
8: code 201.
Higher survival
rates for HL
could
undercount
incident cases,
but maximum
follow-up is 13
years with 5%
mortality during
follow-up.
Controlled for age,
sex, and calendar
year.
Chemical co-
exposures are not
known risk factors
for this outcome.
Radiation
exposure likely to
be poorly
correlated with
formaldehyde.
SMRs (95%
CI)
developed
from the
English and
Welsh
populations
Latency not
evaluated.
HL: 1
Low power due to
the rarity of cases.
SB IB Cf Oth
0
Selection:
Extremely healthy
population with
overall cancer SMR
of 0.44
Exposure: Group B
Lack of latency
analysis
Low power
SUMMARY: NOT
INFORMATIVE
This document is a draft for review purposes only and does not constitute Agency policy.
A-690 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Second
follow-up
through 1980
(Harrington
and Oakes,
1984).
was obtained
from the
census, a
national
health
registry, and
other sources
(100%). Cause
of death data
for 222 of 231
individuals
(96.5%).
All cancer
SMR = 0.44.
Critical limitation:
Selection bias
Hansen et al.
(1994)
Denmark
Cohort study
of workers at
a Danish
pharmaceutic
al plant.
10,889 employees (51%
women) ever employed
1964-1988 at a
pharmaceutical plant.
Cases were extracted
from the Danish Cancer
Registry.
All cancer SIR
(men)=0.95
All cancer SIR (women)
= 1.16.
No individual-level
exposures
estimated: whole
cohort assumed to
be exposed.
Formaldehyde was
one of many
exposures in this
industry but not a
main ingredient or
product.
Co-exposures may
have included
asbestos,
antibiotics,
chloroform,
dichloromethane,
Incidence: cases
from Danish
Cancer Registry
classified
according to
ICD-7.
HL: ICD201
MM: ICD203.
Higher survival
rates for HL
could
undercount
incident cases,
although
average follow-
up is 13 years.
Controlled for age,
sex, and calendar
year.
Asbestos is
associated with URT
cancers. Ethylene
oxide is associated
with LHP cancers.
Neither were
evaluated as
potential
confounders.
Potential for
confounding is
mitigated by low
formaldehyde
exposure and likely
SIRs (95% CI).
Latency not
evaluated.
HL: 4
Larynx: 5
MM: 0
Low power
due to the
rarity of
cases and
low
confidence in
formaldehyd
e exposure.
BEE
0
Potential selection:
mortalityfor HL
Exposure Group D
Latency not
evaluated
Low power
SUMMARY: NOT
INFORMATIVE
Critical limitation:
Information bias
(minimal exposure)
This document is a draft for review purposes only and does not constitute Agency policy.
A-691 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
enzymes, ethylene
oxide, glucagon
heparin, insulin,
silica, sex
hormones, sodium
saccharin, and
synthetic agents.
low correlation with
asbestos and
ethylene oxide.
(Hansen
and
Olsen,
1995).
Denmar
k
Cohort
study of
Danish
men,
URT
cancers
diagnos
ed 1970-
1984.
2,041 men with incident
cancer whose longest work
experience occurred at least
10 years before cancer
diagnosis.
Cases matched with
employment records from
pension fund (72%) with
remainder being self-
employed, pensioners, and
unemployed.
External comparison with
general population.
Average follow-up =13 years.
Individual
occupational
histories including
industry and job
title established
through company
tax records.
Considered
exposed if worked
in plant with more
than 1 kg
formaldehyde used
per employee per
year.
Very crude
exposure
assessment.
No information on
co-exposures
except for wood
dust.
Incident cases
identified in
Danish Cancer
Registry (ICD-7).
NPC: 146
SNC: 160
Larynx: 161
HL: 201.
Higher survival
rates for HL
could
undercount
incident cases,
although
average follow-
up is
approximately
13 years.
Controlled for age,
sex, and calendar
time.
Sinonasal cancer risk
was evaluated
controlling for wood
dust.
While other co-
exposures were not
evaluated, the
overall correlation
between co-
exposures in
multiple
occupational
industries is likely to
be low.
SPIRs (95% CI)
(Standardized
proportionate
incidence ratio)
- proportion of
cases for a
given cancer in
formaldehyde-
associated
companies
relative to the
proportion of
cases for the
same cancer
among all
employees in
Denmark.
Latency
addressed by
inclusion
criteria.
NPC: 4
SNC: 13
Larynx: 32
HL: 12
SB IB Cf Oth
Potential selection:
mortality
for HL
Exposure Group D
Low power for NPC
SUMMARY:
HL, Larynx, NPC,
SNC: LOW vU
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-692 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Harrington
and Oakes,
1984.
Great Britain
Second
cohort study
of British
pathologists.
Related
studies:
Initial follow-
up through
1973
(Harrington
and Shannon,
1975);
Third follow-
up through
1987 (Hall,
1991,
626476).
2,720
pathologists
from the
Royal College
of
Pathologists
and the
Pathological
Society of
Great Britain
from
1974-1980.
Deaths among
those >85
years were
censored.
Vital status
was obtained
from the
census, a
national
health
registry, and
other sources
(100%). 96%
of death
certificates
were obtained
with 91
reporting a
cause of
death.
As a profession, pathologists
were highly exposed to
formaldehyde as a main
ingredient in tissue fixative.
NIOSH (Industry Selection for
Determination of Extent of
Exposure, 1979) has reported
mean formaldehyde
concentrations of 4.35 ppm
with range (2.2-7.9).
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc, and
ionizing radiation.
Mortality: cause
of death
sinonasal
cancer.
Controlled for age,
sex, and calendar
year.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
SMRs (95% CI)
developed
from the
English and
Welsh
populations.
Latency not
evaluated.
SNC: 0
Low power
due to the
rarity of
cases.
SB IB Cf Oth Overall
Selection:
Extremely healthy
population with
overall cancer SMR
of 0.61
Exposure: Group B
Lack of latency
analysis
Low power
SUMMARY: NOT
INFORMATIVE
Critical limitation:
Selection bias
This document is a draft for review purposes only and does not constitute Agency policy.
A-693 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
All cancer
SMR = 0.61.
(Hauptmann et
al., 2009).
United States
Nested case-
control study
within
extension of
embalmers
cohorts
described in
Hayes et al.,
1990; Walrath
and Fraumeni,
1983b; 1984.
Embalmers
(8% women)
from national
and state
funeral
directors
associations
and licensing
boards. Died
1960 -1986.
Participation
rate of case
interviews was
220/228 (96%)
and 265/282
eligible
controls
(94%).
Controls
randomly
selected from
individuals in
the funeral
industry
whose deaths
were
attributed to
other causes.
Controls
stratified to be
similar to data
source, sex,
Individual level, based on
lifetime work practices and
exposures to formaldehyde
obtained by interview with
next of kin or co-workers
(96% of cases and controls)
with information on
occupational exposure
resulting from embalming.
Interviewers blinded to
outcome.
Exposure levels assigned
based on laboratory
reconstruction of exposures
for specific work practices.
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc, and
ionizing radiation.
Mortality:
underlying
cause from
death
certificates,
ICD-8.
MM: ICD203
LL: ICD 204
ML: ICD 205.
Higher survival
rates for HL
could
undercount
incident cases,
but average
follow-up is
more than 39
years (485 cases
and
controls/19,104
person-years).
Controlled for date
of birth, age at
death, sex, data
source, and smoking.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
Logistic
ML: 34
regression, OR
(17
(95% CI) by
acute)
exposure
MM: n
categories (4
cases
levels) for
not
duration, number
reported
of embalmings,
but must
cumulative
be
exposure,
greater
average intensity,
than 5
time-weighted
due to
average, and
size of
peak exposure
se(ln(OR
measures.
))¦
LL: 99
Analyses of
duration of
exposure for MM
is proxy for
latency.
SB IB Cf Oth
Overall
¦I
Exposure: Group A
Latency not evaluated
for LLor MM
SUMMARY:
ML: HIGH
LL, MM: MEDIUM ^
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-694 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
and dates of
birth and
death (5-year
intervals).
Hayes et al.
(1990)
United States
Cohort study
of
embalmers.
Related
study:
Hauptmann
et al. (2009)
4,046
deceased
male
embalmers
and funeral
directors,
derived from
state licensing
boards and
funeral
director who
died during
1975-1985
and a death
certificate
could be
obtained.
Death
certificates
obtained for
79% of
potential
study
subjects.
The 21%
missing death
certificates
Individual exposure measures
not derived. Occupation
confirmed from death
certificates.
Separate study estimated
personal formaldehyde
exposures from 0.98 ppm
(high ventilation) to 3.99 ppm
(low ventilation), with peaks
up to 20 ppm.
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc, and
ionizing radiation.
Mortality:
underlying
cause of death
from death
certificates,
ICD-8;
ICD 201 = HL
ICD 203 = MM
ICD 204 = LL
ICD 205 = ML.
Higher survival
rates for HL and
LL could
undercount
incident cases,
and median
follow-up is
unknown.
Controlled for
calendar year, age,
sex, and race.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
PMR (95% CI).
Latency not
evaluated.
HL: 3
Larynx: 7
LL: 7
ML: 24
MM: 20
NPC: 4
SNC: 0
Possible
undercountin
g of cases
due to
abbreviated
death
certificate
search.
SB IB Cf Oth
Overall
4-
Exposure: Group A
Latency not
evaluated
Low power for HL,
NPC, SNC
SUMMARY:
Larynx, LL, ML,
MM: MEDIUM ^
(Potential bias 4/)
HL, NPC, SNC: LOW
(Potential bias 4'
low sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-695 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
considered to
missing at
random
because all
embalmers
were
considered to
be exposed to
formaldehyde
All cancer
PMR (white) =
1.07
(nonwhite) =
1.08.
Jakobsson et
727 male
al. (1997)
employees of
Sweden
2 plants
producing
Cohort study
stainless steel
of workers
sinks and
grinding
sauce pans
stainless
employed at
steel.
least one year
during 1927-
1981 with
minimum
15-year
follow-up.
Of 823
original
workers, 23
No individual exposure
measures.
Presumed exposure was to
phenol-formaldehyde resins
on ribbons or plates in
grinding workers.
Co-exposures may have
included chromium, nickel,
and abrasive dusts including
silicon carbide, aluminum
oxide, silicon dioxide, and
clay.
No wood dust exposures.
Incidence:
cases from
Swedish Tumor
Registry
SN ICD-7 160.
Adjusted for sex,
age, and calendar
year.
Nickel and
chromium are
associated with URT
cancers and would
likely be positively
correlated with
formaldehyde
exposure.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
SIRs (95% CIs).
Latency
addressed by
enforcing a 15-
year waiting
period to begin
observation.
Larynx:l
SNC: 0
Low power
due to the
rarity of
cases.
SB IB Cf Oth
EH
Exposure Group D
Confounding
possible for
laryngeal cancer
Low power
SUMMARY:
Larynx, SNC: LOW
(Potential bias 4,
low sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-696 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
(3%) could not
be identified,
12 died or
emigrated
before 1952
(1%), and 61
did not
exceed the 15
year waiting
period. No
further losses
to follow-up.
All cancer SIR
= 0.9.
Other co-exposures
are not known risk
factors for these
outcomes.
Levine et al.
(1984)
Canada
Cohort study
of
undertakers.
1,477 male
undertakers
first licensed
during 1928-
1977 with
mortality
follow-up
from 1950-
1977.
Vital status
was 96%
complete with
cause of
death
available for
94%.
As a profession,
undertakers/embalmers were
highly exposed to
formaldehyde as a main
ingredient in tissue fixative.
Kerfoot and Mooney (1975)
reported mean formaldehyde
concentrations for
embalmers in funeral homes
of 0.74 ppm with range (0.09-
5.26).
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc, and
ionizing radiation.
Mortality:
underlying
cause from
death
certificates
(ICD-8).
Nose, middle
ear, sinuses:
160
Larynx: 161.
Controlled for
calendar year, age,
and sex.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
SMR, 95% CI.
Latency was
not evaluated
for these
endpoints.
SNC: 0
Larynx: 1
Low power
due to the
rarity of
cases.
Potential selection:
Healthy worker
effect possible
Exposure Group A
Latency was not
evaluated
Low power
SUMMARY:
Larynx, SNC: LOW
(Potential bias 4,
low sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-697 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Average
follow-up 25
years.
All cancer
SMR = 0.87.
Li et a I., 2006
China
Nested case-
cohort study
within a cohort
study of textile
workers.
67 women
diagnosed
during 1989-
1998 with
nasopharynge
al cancers
were
identified in a
cohort of
267,400
female textile
workers born
during 1925-
1958.
Nine
additional
cases (12% of
total) were
excluded due
to lack of
occupational
histories.
3,188 controls
randomly
selected from
the cohort
Individual level, based on
job exposure matrix
developed for this
industry/setting (unclear
extent of industrial hygiene
specifically for
formaldehyde).
No historical measurements
of exposures. No cases
were classified as exposed
and only 10/3,188 controls
(0.3%) were classified as
exposed.
EPA considered the
potential for formaldehyde
exposure to be exceedingly
low.
Co-exposed to cotton dust.
Incidence or
mortality.
Diagnosis of
nasopharyngeal
cancer or
sinonasal
cancer reported
to a cancer and
death registry
operated by the
Shanghai Textile
Industry
Bureau.
NPC: ICD-9 147
SN: ICD-9 160.
Controlled for age
and sex.
Dusts could be a
potential
confounder but due
to the rarity of
formaldehyde
exposure the
correlation would be
minimal.
Cox proportional
hazards modeling
adapted for case
cohort design.
Hazard ratios
(95% CI).
Duration and
latency were not
evaluated.
NPC: 10
No cases
exposed.
Very low
power
due to
the
rarity of
exposur
e.
SB IB a Oth
Overall
Exposure Group B
Very low power due
to the rarity of
exposure
SUMMARY: NOT
INFORMATIVE
(Very low sensitivity
potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-698 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
frequency
matched by
age.
Malker et al.
(1990)
Sweden
Cancer
registry-
based study,
NPC
diagnosed
1961-1979.
471 employed
men with
incident NPC
cancer.
No individual exposure
measures.
Occupations and
industries with potential
exposure to
formaldehyde:
bookbinders, fiberboard
makers, textile workers,
furniture makers,
chemical workers,
physicians, foundry
workers, biologists,
tanners, and skin
processors, worker
employed in veneer and
plywood plants and in
sugar processing plants.
Co-exposure information
not provided.
Incident cases
identified in
Swedish
Cancer-
Environment
Registry.
Microscopic
confirmation
obtained for
99.6% of NPC
cases. 48%
squamous cell
carcinomas,
37% unspecified
carcinomas, 5%
transitional cell
carcinomas, and
3%
adenocarcinom
as.
Controlled for age
and region.
Variation in
exposure was not
evaluated.
Co-exposures were
also not evaluated.
Fiberboard workers
are also exposed to
wood dust.
Wood dust is
associated with URT
cancers and would
likely be positively
correlated with
formaldehyde
exposure.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
SIRs (95%
CI).
Latency
not
evaluated
NPC: 12
SB
IB
¦ oth
Overall
Exposure Group D
Latency not
evaluated
Confounding
possible
Low power for any
one occupation
which may be
potentially exposed
SUMMARY:
NPC: Low \U
(Potential bias 4,
low sensitivity)
Marsh et al.
(2002/2007)
United States
7,328 workers
employed at a
formaldehyde using
plant in Connecticut
Worker-specific
exposure measures
from job exposure
matrix based on
Mortality:
oropharyngeal
code ICD-9:
146.
Controlled for age,
race, sex, and time
period.
SMR (95%CI)
Secondary
analysis for
NPC.
Oro: 5
Hypo:3
SB IB Cf Oth
4-
This document is a draft for review purposes only and does not constitute Agency policy.
A-699 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
Participants
Evaluation of
setting, and
and
Exposure measure and
Outcome
Consideration of
Analysis and
Study
major bias
design
selection
range
measure
likely confounding
results
sensitivity
categories
Nested case-
followed from 1945
available sporadic
Hypopharyngeal
Comparison was with
Low power
Exposure Group B
control study
through 1998.
plant monitoring data
code ICD-9:
U.S. death rates and
EPA derived
due to the
Latency not
within a
Vital status was
from 1965-1987, job
148.
with death rates in 2
SMRs for the
rarity of
evaluated
cohort study
identified from the
descriptions, and
Nasopharyngeal
counties.
combination
cases.
of workers in
National Death Index,
verbal job descriptions
code ICD-9:
of
Low power
one plant
private businesses, or
by plant personnel
147.
Benzene is not
oropharyngea
NPC: cases
using
state and local
and industrial
Pharyngeal ICD-
associated with URT
1,
included in
SUMMARY:
formaldehyd
agencies, and was
hygienists.
9: 146-149.
cancers. Potential
hypopharyng
Beane
Oro- alone & Hypo-
e.
98.4% complete;
confounders were
eal and
Freeman et
alone: LOW
cause of death data
Exposure assessment
Death
evaluated but only
unspecified
al. (2013).
(Potential bias 4,
Related
for 95% of 2,872
did not include the
certificates used
smoking was found to
pharyngeal
low sensitivity)
studies:
deaths.
same industrial
to determine
be a potential
cancer by NPC
Initial 10
hygiene sampling
underlying
confounder and was
cases from all
OHPC together:
plant cohort
Average follow
up
conducted by Stewart
cause of death
controlled for.
pharyngeal
MEDIUM (Potential
follow-up
=32.89 years.
et al. (1986) used in
according to the
cancers.
bias 4/)
through 1980
the Beane Freeman
ICD codes at
Co-exposures to
(Blair et al.,
All cancer SMR
=
(2009, 2013) analyses
time of death.
pigments and particles
Latency not
1986,1987).
1.08.
which included this
plant.
Histological
typing not
were evaluated and
were found not to be
evaluated.
Second set of
reported.
confounding. Marsh
10 plant
Exposure estimates
et al. (2002)
follow-ups
were on average 10
attempted to evaluate
through 1994
times lower than
smoking but data
(Hauptmann
those of other studies
were incomplete. No
et al., 2003,
in this plant (Blair et
other potential
2004).
al., 1987,1986; Beane
Freeman et al., 2009,
confounders were
evaluated.
Third set of
2013).
10 plant
Beane Freeman et al.
follow-ups
From Beane Freeman
(2009, 2013)
through 2004
et al. (2009, 2013): Co-
evaluated 11 potential
(Beane
exposed to
confounders among a
Freeman et
antioxidants, benzene,
carbon black, dyes and
set of 10 plants that
included this one and
This document is a draft for review purposes only and does not constitute Agency policy.
A-700 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
Participants
setting, and
and
Exposure measure and
Outcome
Consideration of
design
selection
range
measure
likely confounding
al., 2009,
pigments, melamine,
did not find any
2013).
hexamethylenetetram
confounding.
ine, phenols,
plasticizers, urea,
wood dust.
Matanoski
3,644
As a profession, pathologists
Mortality: death
Controlled for sex,
(1989)
deceased
were highly exposed to
certificates and
race, age, and
United States
male
formaldehyde as a main
obituary notices
calendar-year-
pathologists,
ingredient in tissue fixative.
used to
expected deaths
Prospective
derived from
determine
from the U.S.
mortality
membership
NIOSH (Industry Selection for
cause of death
population and
cohort study
rolls of
Determination of Extent of
from Hodgkin
psychiatrists.
with two
multiple
Exposure, 1979) has reported
lymphoma (ICD-
external
professional
mean formaldehyde
8: 201).
Variation in
comparison
societies.
concentrations of 4.35 ppm
exposure was not
groups.
with range (2.2-7.9).
Higher survival
evaluated.
Mortality
rates for HL
followed
Co-exposures may have
could
Radiation exposure
through 1978.
included: phenol, methyl
undercount
likely to be poorly
Death
alcohol, glutaraldehyde,
incident cases,
correlated with
certificates
mercury, arsenic, zinc, and
although
formaldehyde.
obtained for
ionizing radiation.
median follow-
94% of
up is probably
Chemical co-
potential
more than 15
exposures are not
study
years since
known risk factors
subjects, 3%
follow-up was
for this outcome.
from obituary
from the early
notices and
20th century
3% presumed
through 1978.
dead.
All cancer
SMR = 0.78.
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
SMRs (95% CI).
Latency not
evaluated.
HL: 2 cases
total
Low power
due to the
rarity of
cases.
0
Selection: Healthy
worker effect
probable with
overall cancer SMR
of 0.78.
Exposure: Group B
Latency not
evaluated
Low power
SUMMARY: NOT
INFORMATIVE
Selection and
information biases
This document is a draft for review purposes only and does not constitute Agency policy.
A-701 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Meyers et al.
(2013)
United States
Prospective
cohort
mortality
study.
Related
studies:
Initial cohort
follow-up
(Stayner et
al., 1988)
Second
follow-up
(Pinkerton et
al., 2004)
Workers in 3
U.S. garment
plants
(n=11,043) in
Georgia and
Pennsylvania
exposed for at
least 3
months (82%
female).
Vital status
was followed
through 2008
with 99%
completion.
Causes of
death were
obtained for
3,904 (99.7%)
of the 3,915
identified
deaths.
Average
follow-up
=37.52 years.
All cancer
SMR = 0.96.
Individual-level exposure
estimates for 549 randomly
selected workers during 1981
and 1984 with 12-73 within
each department.
Formaldehyde levels across
all departments and facilities
were similar.
Exposures ranged from 0.09-
0.20 ppm. Overall geometric
mean concentration of
formaldehyde was 0.15 ppm,
(GSD 1.90 ppm). Area
measures showed constant
levels without peaks.
No other chemical exposures
were identified by the
industrial hygiene surveys.
There was no information on
smoking in this analysis,
however, according to
(Stayner et al., 1988), "the
overall prevalence of
cigarette smokers was 29.4%.
In plant 1 the prevalence was
26.6%, in plant 2 it was
33.5%, and in plant 3 it was
29.4%. These figures are
similar to those reported in a
1980 survey of adult
Americans, in which 29.2% of
Mortality: death
certificates used
to determine
the underlying
cause of death
(ICD-10):
NPC: Cll
OHPC: C09-C10,
C12-C14
SN: C30-31
Larynx: C32.
HL:C81
LL: C91.0-91.3,
C91.5-91.9
ML: C92
MM: C88.7,
88.9, 90.
Higher survival
rates for HL
could
undercount
incident cases,
but average
follow-up is
more than 37
years
Histological
typing not
reported.
Adjusted for sex,
age, race, and
calendar-year
specific US mortality
rates.
No other chemical
exposures were
identified by the
industrial hygiene
surveys that could
influence the
findings.
SMRs (95% CI),
by exposure
categories (3
levels) for
duration, time
since first
exposure
measures.
SRRs (95% CI)
(internal
comparison),
by 3 categories
of duration of
exposure.
Latency effects
were examined
for leukemia.
NPC: 0
OHPC: 6
SNC: 0
Larynx: 4
ML; 21 (14
acute; 5
chronic)
LL: 6
HL: 4
MM: 23
SB IB Cf Oth
4-
Exposure Group A
Latency for
leukemia only
Low power for NPC,
SNC, Larynx, HL
SUMMARY:
Larynx, NPC, SN:
LOW •
(Potential bias 4,
low sensitivity)
HL, MM, OHPC:
MEDIUMS
(Potential bias 4/)
LL, ML: HIGH
This document is a draft for review purposes only and does not constitute Agency policy.
A-702 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
females and 38.3% of males
over the age of 20 were
current cigarette smokers
[NCHS, 1985]."
Ott et al.
(1989)
United States
(West Virginia)
Nested case-
control study
within two
chemical
manufacturing
plants.
29,139 male
workers
followed from
1940-1978.
Loss to follow-
up 3.6%.
95.4% of
death
certificates
obtained.
Frequency
matching of
controls (5:1)
from the total
employee
cohort
according to a
group-
matched
incidence
density
sampling
design.
Individual-level exposure
classification based on
company records of work
assignments linked to
records on department
usage of formaldehyde.
Exposures during 1940 to
1978.
21 different chemicals were
evaluated including
benzene with much cross
exposure.
Mortality:
underlying
cause from
death
certificates, ICD
version in effect
at time of
death.
Higher survival
rates for LL
could
undercount
incident cases,
but average
follow-up is
likely more than
15 years as
follow up was
initiated in 1940
and ceased in
1978.
Unconditional
logistic regression.
Controlled for sex
and age.
Controlling for age
did not change
results.
Benzene was not
evaluated as a
potential
confounder and may
be positively
correlated with
formaldehyde
exposure.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
Potential for
confounding may be
mitigated by rarity of
co-exposures among
cases.
OR (95% CI).
Analyses
conducted with a
5-year exposure
lag. Limited
adjustment for
latency.
MM: 20
ML: 39
LL: 18
<2
exposed
cases for
each
endpoint
Low
power
due to
the
rarity of
exposur
e.
SB IB Cf Oth
a
¦i
Exposure Group B
Latency evaluation
likely to be under-
powered to detect
any effects beyond a
5-year period.
Confounding possible
Low power due to
rarity of exposure
SUMMARY:
LL, ML, MM: LOW <4,
(Low sensitivity
potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-703 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Robinson et
al. (1987)
United States
Prospective
cohort
mortality
study.
Plywood mill
workers
(n=2,283)
employed at
least one year
during 1945-
1955 followed
for mortality
until 1977
with vital
status for 98%
and death
certificates for
97% of
deceased.
Average
follow-up
=25.22 years.
All cancer
SMR = 0.7.
Individual exposure measures
not derived.
Presumed exposure to
formaldehyde-based glues
used to manufacture and
patch plywood.
Co-exposure to carbon
disulfide, pentachlorophenol,
wood dust.
Mortality:
underlying
cause from
death
certificates
(ICD-7)
HL: 201
MM: 203.
Higher survival
rates for HL
could
undercount
incident cases,
but average
follow-up is
more than 25
years.
Adjusted for sex,
age, race, and
calendar-year-
specific U.S.
mortality rates.
Some exposed
workers also
exposed to
pentachlorophenol
for more than 1
year.
EPA concluded that
pentachlorophenol is
likely to be
carcinogenic based
on strong evidence
from epidemiologic
studies of increased
risk of MM.
Potential for
confounding is
unknown but could
have inflated the
observed effect for
MM but not for HL
SMRs (90% CI).
Latency not
evaluated.
MM: 3 cases
HL: 2 cases (2
cases, whole
cohort of mill
workers; 2
cases,
subcohort of
exposed
workers)
SB IB Cf OtH
0.-1
Selection: Healthy
worker effect
probable with
overall cancer SMR
of 0.7.
Exposure Group D
Latency not
evaluated
MM likely
confounded by
pentachlorophenol
Low power
SUMMARY:
MM: Not
informative,
(Low sensitivity,
likely confounding)
HL: LOW sU
(Low sensitivity
potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-704 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
Saberi
Hosnijeh et al.
(2013)
Europe
Prospective
cohort study.
241,465
men and
women
recruited
from 10
European
countries
during 1992-
2000.
Participants
were
predominan
tly ages 35-
70 at
recruitment
and were
followed up
through
2010.
Occupational histories obtained
by questionnaire about ever
working in any of 52
occupations considered to be at
high risk of developing cancer.
Occupational exposures
estimated as "high," "low," and
no exposure by linking to a JEM.
Incident
primary
leukemias
identified
from
cancer
registries,
health
insurance
records,
pathology
registries
and
contact
with
subjects of
their next
of kin.
Controlled for age, sex,
smoking, alcohol,
physical activity,
education, BMI, family
history of cancer,
country, other
occupational
exposures, and
radiation.
Proportional
hazards
regression; HRs
(95% CI).
Latency was
not evaluated.
LL: 67/225
exposed
ML: 49/179
exposed
SB IB a Oth
H
Exposure
Group C
Latency was
not evaluated
SUMMARY:
LL, ML: LOW <4,
(Potential bias
Siew et al.
(2012)
Finland
National
cohort study.
All Finnish
men born
during 1906-
1945 who
participated
in census
and were
employed in
1970 (n=1.2
million).
Cancer cases
identified by
national
registry
Occupational history from
census records were linked to
the national JEM to code each
cohort member with "any"
exposure to formaldehyde or
"none." Only some use of
"industry" information.
3% of NPC cases exposed
5% of SNC cases exposed
Co-exposure wood dust was
collected.
Diagnosis
of cancer
reported to
the Finnish
Cancer
Registry.
Controlled for age, sex,
socioeconomic status,
smoking, and wood
dust.
SIRs (95% CI).
A 20-year
latency period
was assumed.
NPC: 149
SNC: 167.
Baseline
incidence of
NPC in this
population is
the lowest in
the world.
SB IB a Oth
Exposure
Group D
Low power due
to rarity of
exposure
SUMMARY:
This document is a draft for review purposes only and does not constitute Agency policy.
A-705 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
during 1971-
1995.
NPC, SNC:
LOW sU
(Potential bias
Solet et al.
(1989)
United States
Proportionat
e mortality
study of pulp
and paper
workers.
201 white
male pulp and
paper
producing
workers who
died during
1970-1984
and had at
least 10 years
of experience
in the
industry.
All cancer
PMR= 1.31.
Occupational history from
union records identified
workers in the pulp and paper
producing jobs.
Formaldehyde is known to be
an exposure for pulp and
paper mill workers: job-
specific exposures range from
0.2 to 1.1 ppm with peaks as
high as 50 ppm (Korhonen et
al., 2004).
From Band et al. (1997), co-
exposed to arsenic,
chlorophenols, sulfuric acid
mists, and chloroform.
According to a review
(Kauppinen et al., 1997
IAOEH; 70:119-127), co-
exposures to dioxin or
perchloroethylene are also
possible.
Mortality:
underlying
cause from
death
certificate
submitted to
the Union
Pension Fund.
HL: ICD-8 201.
Higher survival
rates for HL
could
undercount
incident cases,
but average
follow-up is
probably more
than 15 years
because
workers had to
have at least 10
years of
experience in
the industry.
Controlled for age,
sex, race, age at
death, and calendar
time.
Confounding not
evaluated.
Potential
confounders for
these outcomes
include
chlorophenols, acids
mists, dioxin, and
perchloroethylene,
which are likely to
have been positively
correlated with
formaldehyde
exposure.
Other co-exposures
are not known risk
factors for these
outcomes.
Potential for
confounding is
unknown but could
PMRs (95% CI).
Latency not
evaluated.
HL: 1 case
Low power
due to the
rarity of
cases.
SB
IB
Cf Oth
Overall
0
-
Potential selection:
mortality for HL
Exposure Group D
Latency not
evaluated
Confounding
possible
Low power
SUMMARY: NOT
INFORMATIVE
Critical limitation:
(multiple potential
biases and
uncertainties)
This document is a draft for review purposes only and does not constitute Agency policy.
A-706 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
have inflated the
observed effect.
Stellman et
al. (1998)
United
States
General
population
cohort.
Baseline
enrollment
in 1982;
follow-up
through
1988.
317,424 men
enrolled in
the American
Cancer
Society's
Cancer
Prevention
Study II in
1982. Follow-
up was 98%
complete.
Median
follow-up 6
years.
Average
follow-up
=5.79 years.
Individual level, based on
questionnaire response
(Yes/No) on formaldehyde
exposure. Excludes wood-
related occupations.
Specific co-exposures
included asbestos and wood
dust.
Mortality:
death
certificates,
MM: ICD-9
203.
Controlled for age,
sex, and smoking.
Co-exposures are
not associated with
LHP cancers.
Poisson
regression,
(internal
comparison)
RRs (95% CI).
Latency not
evaluated.
MM: 14
(4 exposed)
Low power dues to
the rarity of
exposure.
Exposure Group C
Latency not
evaluated
Low power
SUMMARY: LOW
(Low sensitivity
potential bias 4/)
Stroup et al.
(1986)
United
States
Retrospectiv
e cohort
mortality
study.
2,239 deceased
white male
anatomists
identified from
professional
societies who died
during 1925-
1979.
91% of death
certificates of
As a profession,
anatomists were highly
exposed to
formaldehyde as a main
ingredient in tissue
fixative.
Akbar-Khanzadeh and
Mlynek (Akbar-
Khanzadeh, 1997,
626546@@author})
Mortality:
underlying
cause from
death
certificates
(ICD-8),
HL: 201
Larynx: 161
ML: 205
SNC: 160.
Controlled for
calendar year, age,
sex, race compared
with U.S. population.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
SMR (95% CI).
Latency not
evaluated.
HL: 0
Larynx: 1
ML: 5 (1 acute,
3 chronic, 1
unspecified)
SNC: 0
Low power due
to the rarity of
cases.
SB
IB Cf Oth
Overall
1
_
Selection: Healthy
worker effect
probable with
overall cancer SMR
of 0.64.
Exposure Group A
This document is a draft for review purposes only and does not constitute Agency policy.
A-707 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
known deceased
obtained.
Average follow-up
=22.52 years.
All cancer SMR =
0.64.
reported mean
formaldehyde
concentrations in
anatomy laboratories of
1.9 ppm with range (0.3-
4.5).
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc,
and ionizing radiation.
Anatomists may also be
co-exposed to stains,
benzene, toluene,
xylene, chlorinated
hydrocarbons, dioxane,
and osmium tetroxide.
Higher
survival rates
for HL could
undercount
incident
cases, but
average
follow-up is
more than 22
years.
Benzene not
evaluated as
potential
confounder but is a
risk factor for ML.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
Latency not
evaluated
Confounding
possible for ML
Low power
SUMMARY:
HL, Larynx, ML,
SNC: LOW vU
(Low sensitivity
potential bias 4/)
Walrath and
Fraumeni
(1983b)
United
States
Cohort
mortality
study.
Related
study:
Hauptmann
et al. (2009)
1,132 deceased
white male
embalmers
identified from NY
state license
board. Died 1925
-1980.
Death certificates
obtained for 75%.
The 25% missing
death certificates
considered to
missing at random
As a profession,
embalmers were highly
exposed to
formaldehyde as a main
ingredient in tissue
fixative.
Kerfoot and Mooney
(1975) reported mean
formaldehyde
concentrations for
embalmers in funeral
homes of 0.74 ppm with
range (0.09-5.26).
Mortality:
underlying
cause from
death
certificates
(ICD-8)
HL: 201
LL: 204
ML: 205.
Higher
survival rates
for HL and LL
could
undercount
Controlled for
calendar year, age,
sex, and race.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
PMR, 95% CI.
Latency was not
evaluated for
these endpoints.
HL: 7
Larynx: 2
LL: 4
ML: 7
SNC: 0
Low power for
LL due to the
rarity of cases.
SB IB Cf Oth
Overall
4.
Exposure Group A
Latency was not
evaluated.
Low power for
larynx, LL, SNC
SUMMARY:
Larynx, LL, SNC:
LOW sU
(Low sensitivity
This document is a draft for review purposes only and does not constitute Agency policy.
A-708 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
because all
embalmers were
considered to be
exposed to
formaldehyde.
All cancer PMR =
1.11.
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc,
and ionizing radiation.
incident
cases, but
average
follow-up is
likely more
than 15 years
as follow up
was initiated
in 1925 and
ceased in
1980.
potential bias 4/)
HL, ML: MEDIUM
(Potential bias 4/)
Walrath and
Fraumeni
(1984)
United
States
Cohort
mortality
study.
Related
study:
Hauptmann
et al. (2009)
1,007 deceased
white male
embalmers
identified from CA
state license
board. Died 1925
-1980.
Death certificates
obtained for 100%.
All cancer PMR =
1.04.
As a profession,
embalmers were highly
exposed to
formaldehyde as a main
ingredient in tissue
fixative.
Kerfoot and Mooney
(1975) reported mean
formaldehyde
concentrations for
embalmers in funeral
homes of 0.74 ppm with
range (0.09-5.26).
Co-exposures may have
included: phenol, methyl
alcohol, glutaraldehyde,
mercury, arsenic, zinc,
and ionizing radiation.
Mortality:
underlying
cause from
death
certificates
(ICD-8)
HL: 201
LL: 204
ML: 205.
Higher
survival rates
for HL and LL
could
undercount
incident
cases, but
average
follow-up is
likely more
than 15 years
as follow up
was initiated
Controlled for
calendar year, age,
sex, and race.
Radiation exposure
likely to be poorly
correlated with
formaldehyde.
Chemical co-
exposures are not
known risk factors
for this outcome.
PMR, 95% CI.
Latency was not
evaluated for
these endpoints.
ML: 8
Larynx: 2
LL: 4
HL: 0
SNC: 0
Low power due
to the rarity of
cases.
SB IB Cf Oth
Exposure Group A
Latency was not
evaluated.
Low power for HL,
Larynx, LL, SNC
SUMMARY:
HL, Larynx, LL, SNC:
LOW sU
(Low sensitivity
potential bias 4/)
ML: Medium \U
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-709 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants
and
selection
Exposure measure and
range
Outcome
measure
Consideration of
likely confounding
Analysis and
results
Study
sensitivity
Evaluation of
major bias
categories
in 1925 and
ceased in
1980.
Wesseling et
al. (1996)
Costa Rica
Cohort study
of banana
plantation
workers.
26,565 male
workers on the
payrolls of banana
companies as
reported to the
Social Security
Administration
between 1972 and
1979. Cohort
follow-up in the
cancer registry
from 1981 to
1992.
Losses to follow-
up and poor
record keeping
resulted in
difficulty in
assessing
participation rates.
Very low
confidence in data
quality.
Average follow-up
=11.83 years.
All cancer SIR =
0.76 (men).
A list of names of
workers sterilized by
dibromochloropropane
was used to identify
banana plantations
whose workers may have
been exposed to
formaldehyde.
Co-exposed to maneb,
dibromochloropropane,
mancozeb, benomyl,
chlorothalonil.
Incidence:
National
Tumor
Registry.
HL: ICD-9
965-966
MM: ICD-9
973.
Higher
survival rates
for HL and LL
could
undercount
incident
cases, but
average
follow-up is
12 years.
Controlled for age
and sex.
Banana plantation
workers are co-
exposed to several
potential
carcinogens such as
dibromochloropropa
ne, maneb,
mancozeb, benomyl,
and chlorothalonil.
While these
chemical co-
exposures are not
known risk factors
for these outcomes
the fact that co-
exposures were so
high as to cause
sterility in workers
strongly suggests a
large potential for
confounding.
SIR (95% CIs).
Latency was not
evaluated for
these endpoints.
Males:
SB IB
Cf
Oth
Overall
HL: 9 cases
1 1 1
MM: 6 cases
¦ 1 1
0
Selection: Selection
issues (loss to
follow-up, record
keeping). Healthy
worker effect
probable with
overall cancer SIR
of 0.76.
Exposure Group D
Possible
confounding
Very low
confidence in data
quality
SUMMARY: NOT
INFORMATIVE
Critical limitation:
(multiple potential
biases and
uncertainties)
This document is a draft for review purposes only and does not constitute Agency policy.
A-710 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-106. Evaluation of case-control studies of formaldehyde and cancers of the URT (NPC, SN, OHPC) and LHP
(HL, MM, LL, ML)
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
(Armstrong et
al.. 2000)
Malaysia
Population-
based case-
control study
of NPC.
Prevalent and incident
NPC cases (31% female)
during 1987-1992
identified through
treatment or diagnosis
records from 4
radiotherapy centers.
Participation of cases
was 53% due to death
and illness and difficulty
in locating them.
Participation of living
cases who could be
located was 89% (n=282)
and 90% for eligible
controls (n=282).
Selection bias possible.
Cases and controls were
matched on age, sex,
Chinese ethnicity, and
neighborhood.
Participation rate was
somewhat lower in
more affluent
neighborhoods (80% vs.
90%).
Individual-level
exposure status based
on occupational
history obtained by
interview including job
description, worked
performed, calendar
time, machines, tools,
substances used, and
exposures to dusts,
smoke, gases, and
chemicals.
Exposure assessment
blinded to outcome.
Prevalent and
incident cases.
Diagnosis of
NPC: confirmed
by histological
review.
All cases were
squamous cell
carcinomas.
Design controlled
for age, sex,
Chinese ethnicity,
and neighborhood.
Analysis adjusted
for social class,
diet, smoking, and
wood dust.
Other exposures
evaluated were
wood dust,
industrial heat,
textile dusts,
metals, acids,
bases, solvents,
detergents, and
soaps.
Wood dust is a
potential
confounder but
was controlled for.
Conditional logistic
regression; ORs
(95% CI) for each of
22 separate
occupational
exposures.
Latency was
evaluated
(exposures < 1, 5,
10,15, and 20 years
prior to diagnosis).
8/564 subjects
(1.4%) had more
than 10 years of
potential exposure
outside of a 10-year
latency period. This
suggests additional
information bias.
NPC: 282
The power
to evaluate
formaldehy
de as a
hazard is
diminished
as fewer
than 10%
of cases
had any
exposure
to
formaldehy
de.
SB
IB Cf Oth
Overall
0
¦
Selection
issue with substantial
difference in participation
rates.
Exposure Group B
Lack of latency data.
Very low power to detect
any effects beyond a 10-
year period.
SUMMARY: NOT
INFORMATIVE
(multiple potential biases
4/ and uncertainties)
This document is a draft for review purposes only and does not constitute Agency policy.
A-711 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
(Berrino et al.,
2003)
Europe
Population-
based case-
control study
of larynx and
hypopharynx
cancer.
Male residential
populations of six cancer
registries in four
European countries
during 1979-1982.
All patients with newly
diagnosed cancer were
identified with
participation rates of
70% to 92% by center.
Controls participated at
an average rate of 74%.
Controls were selected
from age and sex
stratified random
samples of the local
general population.
Individual-level
exposure status based
on lifetime
occupational history
for all jobs held for
more than one year
obtained from
questionnaire
including job title,
specific tasks, and
calendar time.
Multiple exposure
metrics including
peak, average, and
cumulative exposure
developed by job
exposure matrix.
However, the quality
of the exposure
assessment is further
degraded by the
authors' statements.
Namely, the authors
regarded the "JEM
performance as poor
for formaldehyde
where 14% of jobs
classified as category 1
(unexposed) by the
matrix were judged as
definitely exposed by
the experts." Co-
linearity among crude
exposures (e.g.,
solvents and
formaldehyde had
Incident cases.
Diagnosis of
cancer of the
larynx or
hypopharynx
confirmed by
pathology
review.
Cancer of the
larynx divided
into epilarynx
and
endolarynx.
Analyses of
hypopharynx
grouped
together with
epilarynx while
endolarynx
analyzed
separately.
No separate
analysis of
hypopharynx
without
epilarynx.
Controlled for age
and sex by
selecting controls
from stratified
population
samples.
Analysis controlled
for study center,
age, tobacco
smoking,
socioeconomic
status, alcohol, and
diet.
Exposures to other
compounds were
identified and
evaluated as risk
factors including
asbestos, arsenic,
solvents, and dusts
(wood and other).
Note that solvents
were a stronger
risk factor for
laryngeal cancer
than formaldehyde
(OR=2.21 vs. 1.7).
Co-exposures were
controlled for but
poorly measured
covariates cannot
be well controlled
for.
Unconditional
logistic regression;
OR (95% CI).
Lagged exposures
were evaluated to
account for cancer
latency in selected
analyses.
Larynx
(endolaryn
x): 213
total cases
37 cases
exposed at
least 10
years and
more than
20 years
since first
exposure.
SB IB
Cf Oth
Overall
0
Exposure
Group B downgraded to
Group D based on poor
performance of JEM.
Confounding likely due to
collinearity of exposures to
other risk factors and
potentially poor quality
exposure data which
minimized ability to
control.
SUMMARY: NOT
INFORMATIVE
Critical limitation:
Confounding
This document is a draft for review purposes only and does not constitute Agency policy.
A-712 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Spearman correlation
of 0.4).
Blair et al.
(2001)
United States
Population-
based case
control of
leukemia.
White men, ages > 30
years. Cases (r?=513)
identified 1980-1983
(cancer registry and
hospital network).
Controls (n=l,087)
selected by random digit
dialing (under age 65)
otherwise from lists
provided by the HCFA
and state death files.
Controls were
frequency-matched by
5-year age groups, vital
status at interview, and
state of residence.
Cases participation rate
was 86%. Control
participation rate was
77-79%.
Individual-level
exposure status based
on lifetime farm and
nonfarm occupational
history for all jobs held
for more than one
year obtained from
interview including job
title, industry, and
calendar time.
Other exposures
evaluated included
benzene, other
organic solvents,
petroleum-based oils
& greases, cooking
oils, ionizing radiation,
paper dusts, gasoline
and exhaust vapors,
paints, metals, wood
dust, asbestos,
asphalt, cattle, meat,
solder fumes.
Incident cases.
Diagnosis of
myeloid
leukemia and
lymphatic
leukemia
confirmed by
pathology
review.
Analysis controlled
for age, state,
direct or surrogate
interview, and
smoking.
Other co-exposures
were not evaluated
as potential
confounders.
Logistic regression;
ORs (95% CI) by
exposure categories
(3 levels) for
intensity,
probability,
duration, and time
since first exposure
measures.
Latency not
evaluated.
ML: 22/59
exposed
(14 acute;
8 chronic)
LL: 30/190
exposed
Exposure Group C
Lack of latency analysis
Possible confounding
although relationship
between formaldehyde
and co-exposures is
unknown.
SUMMARY:
LM: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-713 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
d'Errico et al.
(2009)
Italy
Hospital-based
case-control
study of SNC in
the Piedmont
region of Italy.
154 sinonasal cases
during 1996-2000
identified through
treatment or diagnosis
records from all
Piedmont hospital
departments. 5 cases
excluded (3 prevalent
cases, 2 <30 years old).
Participation of incident
cases using full
questionnaire was 76%
(113/149). Participation
of eligible hospital
controls (n=336) was
95%.
Controls frequency
matched for age, sex,
and province of
residence.
Lifetime job history (all
jobs); company, job
title, tasks, size of
work environment,
and other details.
Probability of
exposure was
determined by blinded
expert staff for jobs
lasting 6 or more
months.
Other exposures
evaluated were
arsenic, wood dust,
leather dust, nickel,
chromium, PAHs,
welding fumes, oil
mists, flour dust,
cocoa powder, silica,
coal dust, textile dusts,
acid mists, paint mists,
organic solvents.
Incident cases
by cell type
were taken
from the
regional
Sinonasal
Cancer Registry
reported to
them by
hospitals in the
region.
Analysis controlled
for age, sex,
province of
residence, smoking
and co-exposures.
Wood dust is a
considered an
extremely strong
risk factor for SNC
and a potential
confounder and
was controlled for
but adjusted
results not
presented; just
"loss of statistical
significance."
Unconditional
logistic models; ORs
(95% CI).
Latency was
evaluated with a
10-year latency
period.
SNC: 7/113
exposed
The power
to evaluate
formaldehy
de as a
hazard is
diminished
as fewer
than 10% of
cases had
any
exposure to
formaldehy
de.
IB Cf Oth
Overall
Exposure Group B
Wood dust is a likely
confounder and no effect
estimate adjusted for
wood dust was presented.
Low power
SUMMARY: NOT
INFORMATIVE
Critical limitation:
Confounding
This document is a draft for review purposes only and does not constitute Agency policy.
A-714 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
(Gerin et al.,
1989)
Canada
Population-
based case-
control study.
Related study:
Siemiatycki et
al. (1987)
3,726 male cases, 1979-
1985, from 14 major
area hospitals, which
report to the Quebec
Tumor Registry (97% of
all cancers reported).
533 population controls
participated out of 740
selected (72%).
Interviews and
questionnaires
completed for 82% of
eligible cases of which
18% of interviews were
completed by next of
kin.
Internal and external
comparison.
Controls were patients
with cancer at other
sites with all lung
cancers excluded.
External comparison
with general population.
Lifetime job history
included company
activities, raw
materials and final
product, machines,
tasks involving
machine maintenance,
type of room.
A team of chemists
and hygienists (likely
blinded to outcome)
translated each job
into a list of potential
formaldehyde
exposures based on
their confidence level,
the frequency, and the
duration of exposure.
Incident cases
histologically
confirmed
diagnosis of
Hodgkin
lymphoma
(ICD: 201).
Controlled for age,
ethnic group,
socio-economic
status, smoking,
and dirtiness of
jobs held (white vs.
blue collar).
Additional control
for any of 300 of
the most common
occupational
exposures if the
inclusion changed
the formaldehyde
OR by more than
10%.
Logistic regression;
OR (95% CI).
Latency not
evaluated.
HL: 8/53
exposed.
SB IB Cf Oth
Overall
4-
Exposure Group B
Lack of latency analysis.
SUMMARY:
HL: MEDIUM 4,
(Potential bias 4/).
This document is a draft for review purposes only and does not constitute Agency policy.
A-715 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Gustavsson et
al. (1998).
Sweden.
Population-
based case-
control study
of squamous
cell carcinoma
of oro-/hypo-
pharynx.
138 men between 40-79
years old residing in
Sweden identified by
hospitals reports or
regional cancer
registries with squamous
cell carcinoma of oro-
/hypo-pharynx during
1988-1990.
Interviews completed
for 90% of cases and
85% of controls.
Controls were randomly
selected from
population registries and
frequency-matched by
sex, age, and region.
Detailed occupational
history obtained
through unblinded
interview yielding
information on all
jobs held >1 year,
starting and stopping
times, job title, tasks,
and company.
Histories reviewed by
a blinded industrial
hygienist who coded
jobs based on
intensity and
probability of
exposure to 17
occupational factors.
Incident cases of
squamous cell
carcinomas of
the oro-/hypo-
pharynx.
Diagnosis of
cancer based on
ICD-9 codes 146
(oropharynx) and
148
(hypopharynx)
but not including
code 147
(nasopharynx) on
weekly reports
from
departments of
otorhinolaryngol
ogy, oncology,
and surgery and
from regional
cancer registries.
Controlled for sex,
age, region,
drinking, and
smoking.
Other exposures
evaluated
included:
polycyclic
aromatic
hydrocarbons,
asbestos, quartz,
dusts (general,
leather, wood,
metal, paper,
textile), oil & acid
mists, phenoxy
acids, welding
fumes, manmade
mineral fibers,
nickel, hexavalent
chromium.
Leather dust was a
risk factor for
OHPC but only 5
cases were
exposed.
Asbestos and
metal dust were
risk factors for
laryngeal cancers
with 34 and 41
cases respectively.
Since asbestos and
metal dust were
stronger risk
Unconditional
logistic regression;
RRs (95% CI).
Latency not
evaluated.
Larynx:
23/157
exposed.
OHPC:
13/138
exposed.
SB IB
Cf Oth
Overall
Exposure
Group B
Lack of latency analysis.
For OHPC,
Confounding possible for
the power
laryngeal cancer.
to evaluate
formaldehy
Low power.
de as a
hazard is
SUMMARY:
diminished
Larynx: LOW 4,
as fewer
(Potential bias 4/)
than 10% of
OHPC: LOW 4,
cases had
(Low sensitivity Potential
exposure to
bias 4,).
formaldehy
de.
This document is a draft for review purposes only and does not constitute Agency policy.
A-716 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
factors for
laryngeal cancer
than
formaldehyde and
more common
exposures, there is
a potential for
confounding with
this cancer.
This document is a draft for review purposes only and does not constitute Agency policy.
A-717 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Heineman et al.
(1992).
Denmark.
Cancer registry-
based case-
control study,
MM diagnosed
1970-1984.
2,098 men registered in
both the national
cancer registry and
pension fund. All men
with a specific
occupational history
were included.
Controls frequency
matched on age, sex,
and year of diagnosis.
Individual-level
exposure estimated
by industrial
hygienists based on
occupation listed on
most recent tax
documents.
Incident cases
identified in
Danish Cancer
Registry.
92% of cases
were
histologically
confirmed.
Controlled for age
and gender.
Other compounds
were identified
and evaluated as
independent risk
factors including:
gasoline, oil
products, engine
exhausts, benzene,
dyes, phthalates,
vinyl chloride,
asbestos, and
pesticides.
Asbestos is not a
risk factor for LHP.
'Possible' benzene
exposure was
associated with
MM but not
'probable' Benzene
exposure, so
confounding is
considered to be
unlikely.
Logistic regression,
ORs (95% CI) by
likelihood of
exposure in 3
categories.
Latency not
evaluated.
MM: 835
(185
exposed).
sa ib cf oth
Exposure Group D
Latency not evaluated.
Confounding unlikely.
SUMMARY:
MM: LOW 4,
(Potential bias 4/).
This document is a draft for review purposes only and does not constitute Agency policy.
A-718 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Hildesheim et
al. (2001).
Taiwan.
Population-
based case-
control study.
Related studies:
Yang et al.
(2005).,
Hildesheim et
al. (1997),
Cheng et al.
(1999)
375 men and women
with NPC and 375
controls. Ages <75 years,
July 1991 and January
1995, from two
hospitals.
Participation of eligible
cases was 99% and 87%
for controls.
Controls individually
matched 1:1 on age, sex,
and district/township of
residence.
Lifetime job history
(jobs held for at least
one year since age 16);
job title, typical
activities/duties, type
of industry, and tools
and/or materials used.
Industrial hygienist
assigned scaling to
subjects based upon
intensity and
probability of
exposure on a scale
from 0-9.
Incident cases.
Diagnosis of
nasopharyngea
I was
confirmed by
histological
review with
>90%
diagnosed with
nonkeratinizing
and
undifferentiate
d carcinomas
and 9% with
squamous cell
carcinoma.
Adjusted for age,
sex, education,
ethnicity, and HLA.
Did not adjust for
residence.
Other exposures
identified included:
wood dust,
solvents, and
smoking. All
subjects were
tested for EBV.
The observed
associations were
not materially
affected when
controlling for
wood dust and
solvent exposure.
Smoking was a risk
factor for NPC and
was not controlled
for in the analysis.
Logistic regression;
ORs (95% CI) by
exposure intensity,
exposure probability,
cumulative exposure
and an induction
period of 10 ten years
used to account for
latency.
Conditional logistic
regression was not
used; however,
logistic regression did
control for age and
sex. Area of residence
was expected to be
related to referral
patterns and may not
be related to
exposure
independent of
occupational history.
NPC: 375
cases (74
ever
exposed)
SB IB Cf Oth
Overall
4-
Exposure Group B
Confounding possible
The impact of not
controlling for all matching
factors is unclear but
considered most likely to
bias towards the null and
inflate confidence
intervals.
SUMMARY:
NPC: LOW 4,
(Low sensitivity potential
bias 4,)
This document is a draft for review purposes only and does not constitute Agency policy.
A-719 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Laforest et al.
(2000)
France
Hospital-based
case-control
study of
hypopharynge
al and
laryngeal
cancer.
Male cases (201 primary
hypopharyngeal
squamous cell cancer,
296 laryngeal cancer),
diagnosed during 1989-
1991, from 15 French
hospitals.
Interviews completed
for 79.5% of eligible
cases and 86% of eligible
controls.
Controls frequency
matched on sex, age,
and the same or similar
nearby hospital.
Occupational histories
from questionnaires;
industry and
occupation coding
used with job
exposure matrix for
formaldehyde (and
other exposures).
Exposure assessment
based on job-exposure
matrix that included
level and probability of
exposure to
formaldehyde as well
as duration and
cumulative exposure
to formaldehyde.
Incident cases.
Diagnosis of
hypopharyngea
I and laryngeal
cancers was
histologically
confirmed.
Controlled for sex, age,
alcohol, and smoking.
Induction periods of 5,
10, and 15 years was
also used to account for
latency in evaluating
risk.
Other exposures
evaluated included:
coal dust, leather dust,
wood dust, flour dust,
silica, and textile dust.
Of these, only coal dust
significantly increased
the risk of
hypopharyngeal cancer
in this study but coal
dust was controlled for
in the OHPC analysis.
Unconditional
logistic
regression; OR
(95% CI).
Latency was
evaluated.
OHPC:
201
Larynx:
296
SB IB Cf Oth
Exposure Group C
SUMMARY:
OHPC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-720 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Luce et al.
(2002)
China, France,
Germany, Italy,
Sweden,
United States
Pooled analysis
of 12 case-
control
studies:
Zheng et al.
(1992), Luce et
al. (1992,
1993), Leclerc
et al. (1994),
Bolm-Audorff
et al. (1990),
Comba et al.
(1992a,b),
Magnani et al.
(1993), Merler
et al. (1986),
Hayes et al.
(1986a,b),
Hardell et al.
(1982),
Vaughan et al.
(1986a,b;
1989), Mack
and Preston-
Martin
(Unpub. data),
Brinton et al.
(1984,1985)
Pooled analysis of 12
case-control studies.
Men and women. All
from 7 different
countries diagnosed
with sinonasal cancer
during 1968-1990.
Each individual study
selected controls
intended to be
comparable to the cases
in that study.
Occupational histories
from interview or
questionnaires;
industry and
occupation coding
used with job
exposure matrix for
formaldehyde (and
other exposures).
Diagnoses
originally
assessed in 12
studies. 195
cases were
adenocarcinom
as (169 men
and 26 women)
and 432 were
squamous cell
carcinomas
(330 men and
102 women).
Adenocarcinoma
results in men
controlled for age,
study, and
cumulative
exposure to wood
and leather dust.
All other results
adjusted for age
and study.
Co-exposures were
evaluated as
potential
confounders.
Other occupational
exposures
potentially
affecting risk
estimates were
controlled for
including dusts
(wood, leather,
coal, flour, textile),
silica, asbestos,
and man-made
vitreous fibers.
Unconditional
logistic regression;
OR (95% CI).
Latency evaluated.
SNC: 627
cases(135
adenocarci
nomas
exposed.
132
squamous
cell
carcinomas
exposed)
Exposure Group C
SUMMARY:
SNC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-721 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Mayr et al.
(2010)
Germany
Hospital-based
case-control
study.
Hospital patients
diagnosed at the
University of Erlangen-
Nuremburg, Germany
during 1973-2007.
31 of 58 patients with
identified
adenocarcinoma (53%)
were followed up with a
standardized
questionnaire. 85 of 110
patients with cancer of
the oral cavity (77%)
included as controls.
Controls were other
hospital patients
diagnosed with oral
cancer during the same
time period as cases and
in the same hospital.
Oral cancer could be
related to formaldehyde
exposure but this would
bias towards the null.
Structured interview
with specific questions
about exposure to
formaldehyde (and
other exposures).
Both cases and
controls were blinded
to case status and
study hypotheses, and
were not aware of
their "case" status.
Prevalent
cases.
Diagnosis of
sinonasal
adenocarcinom
a in the
Department of
Otolaryngology
, Head and
Neck Surgery.
Controlled for age
and sex.
Other exposures:
Wood dust,
preservatives,
stains, varnishes,
solvents, and
pickling solutions.
Wood dust is a
considered an
extremely strong
risk factor for SNC
was not controlled
for so there is a
strong possibility of
confounding.
Crude ORs (95% CI).
Methods unstated
for OR
determinations.
Latency not
evaluated.
SNC: 2/31
exposed
Low power
due to the
rarity of
cases.
SB IB Cf Oth
Potential selection issue
(prevalent cases)
Exposure Group C
Latency not evaluated
Wood dust is a likely
confounder.
SUMMARY: NOT
INFORMATIVE
Critical limitation:
Confounding
This document is a draft for review purposes only and does not constitute Agency policy.
A-722 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Olsen and
Asnaes (1986)
Denmark
Cancer registry-
based case-
control study,
SNC diagnosed
1970-1982.
Related study:
Olsen et al.
(1984)
310 men with incident
SN cancer. 215 (69%)
squamous cell &
lymphoepithelioma. 39
(13%) adenocarcinoma.
2,465 controls, selected
among people with
colon, rectum, prostate,
and breast cancer
diagnosed during the
same time period as
cases. Controls were
selected to be similar
with regard to age, sex,
and year of diagnosis.
Employment
histories from 1964
based on linkage to
population registry
data; includes
industry and job
title. Occupational
exposure to
formaldehyde
estimated by
industrial hygienists
based on industry or
occupations.
Incident cases
identified in
Danish Cancer
Registry.
Cancer of the
nasal cavity (ICD-
7 160.0) or
sinuses (ICD-7
160.2-160.9) was
histologically
confirmed. Of all
male cases for
cancer of the
nasal cavity and
paranasal sinuses.
82% were
squamous cell,
lymphoepithelio
ma 18% were
other types.
Matched for age,
sex, and year of
diagnosis. Mantel-
Haenszel summary
estimates of the
relative risk were
used to account for
possible
confounding
because the
subjects were
stratified according
to several
variables.
Wood dust is a
considered an
extremely strong
risk factor for SNC
so exposure to
wood dust was
evaluated as a
potential
confounder and as
an effect modifier.
OR (95% CI)
calculated using the
method of Rothman
and Boice (1979).
Latency was
evaluated.
SNC: 215
squamous
cell and
lymphoepi
theliomas
(13
exposed)
and 39
adenocarci
nomas
(17
exposed)
SB IB Cf Oth
Exposure Group C
SUMMARY:
SNC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-723 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Olsen et al.
(1984)
Denmark
Cancer registry-
based case-
control study,
NPC diagnosed
1970-1982.
Related study:
Olsen and
Asnaes (1986)
266 incident NPC and
488 incident SN cases;
matched approximately
3 controls per case.
Controls matched on
age, sex, and year of
diagnosis from the
Registry.
Employment
histories from 1964
based on linkage to
population registry
data; includes
industry and job
title. Occupational
exposure to
formaldehyde
estimated by
industrial hygienists
based on industry or
occupations.
Authors reported
that 4.2% of control
males and 0.1% of
females were
exposed to
formaldehyde.
Incident cases
identified in
Danish Cancer
Registry.
NPC: ICD 146
SN: ICD 160.0
and 160.2-160.9
9% of NPC and
SNC cases were
sarcomas and
91% were
carcinomas.
Sarcomas were
excluded but
gender-specific
case counts were
not provided for
carcinomas.
Controlled for age,
sex, and year of
diagnosis from the
registry.
Other exposure
evaluated
included: wood
dust, paint,
lacquer, and glue.
Wood dust is
associated with
SNC and was
evaluated as a
potential
confounder of NPC
but was not a risk
factor.
OR (95% CI)
calculated using
programs developed
by Rothman and
Boice (1979).
Latency was
evaluated.
NPC: 266
cases
(number
exposed is
not stated)
SNC: cases
included in
Olsen and
Asnaes
(1986).
SB IB a Oth
Overall
4-
Exposure Group C
SUMMARY:
NPC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-724 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Pesch et al.
(2008)
Germany
Insurance-
based case-
control study.
Male workers insured by
a liability insurance
association for the
German wood-working
industries with an
occupational disease
during 1994-2003.
86/129 cases (67%)
participated (including
29 next of kin). 204/272
controls (75%)
participated (including
69 next of kin).
Controls were selected
from the same insurance
database of workers
with registered
accidents. Controls were
crudely frequency
matched on age with a
cut-off at 60 years.
Median ages were both
69 years with cases
ranging from 41-84
years and controls
ranging from 37-85
years).
Lifetime job history,
with focus on tasks
and exposures in
wood industries.
Because next-of-kin
information on
exposure to wood
additives was
considered poor, the
probability of
exposure to
formaldehyde was
rated by an expert
team as none, low,
medium, or high.
Prevalent cases.
Cases were ever
employed in
German wood
industries and
diagnosed with
histopathologica
My confirmed
sinonasal
adenocarcinoma
Because cases
and controls
were stratified
by age less than
60 years and
greater or equal
to 60 years, the
older cases may
have been
selected for
survival. If so,
this may have
resulted in a
downward bias.
Controlled for age,
smoking, region,
interviewee, and
average wood dust
exposure.
Co-exposure to
wood
preservatives,
varnishes, and
pigment stains
likely.
Wood dust is a
considered an
extremely strong
risk factor for SNC
but was controlled
for.
Logistic regression.
OR (95% CI).
A 5-year latency
period was applied.
SNC: 47/86
cases
exposed
SB IB Of Oth (Vera
Potential selection issue
(prevalent cases) may have
resulted in a downward
bias.
Exposure Group B
Latency evaluation likely to
be under-powered to
detect any effects beyond
a 5-year period.
SUMMARY:
SNC: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-725 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Pottern et al.
(1992)
Denmark
Cancer registry-
based study,
MM diagnosed
during 1970-
1994.
363 female incident
cases; included if found
in pension fund registry.
1,517 age and sex
matched controls alive at
time of case diagnosis.
All women with a specific
occupational history
other than
"Homemaker" were
included.
Individual-level
exposure estimated
by industrial
hygienists based on
occupation listed on
most recent annual
income tax
documents and the
industry associated
with that occupation.
Incident cases
identified in
Danish Cancer
Registry.
ICD code at
time of
diagnosis.
Controlled for age,
sex, and vital status.
Other exposures
evaluated included
19 categories
grouping 47
substances.
Co-exposures were
not evaluated for
confounding but
exposure to organic
solvents (including
benzene)and
radiation were not
risk factors for MM.
Logistic regression,
ORs (95% CI) by
likelihood of
exposure in 3
categories.
Latency not
evaluated.
MM:
60/363
exposed
SB IB Cf Oth
Exposure Group D
Latency not evaluated
SUMMARY:
MM: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-726 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Consideration of
Outcome
likely
measure
confounding
Incident cases
Controlled for age
(from state
at death, year at
tumor
death, and
registries) who
availability of
had died.
occupational
Diagnosis of
information.
nasopharyngea
1 cancer and
Exposure to wood
sinonasal
dust was not found
cancer based
to be a risk factor
on case
for all nasal cancers
registration by
(NPC+SNC). This
the
suggests a lower
Connecticut
potential for
Tumor
confounding by
Registry.
wood dust.
Clinical records
reviewed for
>75% of cases.
Histological
typing not
reported.
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Roush et al.
(1987)
United States
Population-
based case-
control study.
173 male cases of NPC,
198 male cases of
sinonasal cancer
identified from the
Connecticut Tumor
Registry who died during
1935-1975;and 605
male controls dying
during the same time
period and randomly
selected from state
death certificates.
Controls were matched
on sex, date of death,
and state of residence.
Job history obtained
by city directories and
death certificates,
which yielded
information on job,
industry, employer,
and year of
employment. Job data
sought for 1,10, 20,
25, 30, 40, and 50
years prior to death.
An industrial hygienist,
blinded to case status,
classified likely
exposure to
formaldehyde on basis
of job title.
Logistic regression;
ORs (95% CI).
Intensity of the
likelihood of
exposure and
latency evaluated.
NPC:
21/173
exposed
SNC:
21/198
exposed
SB IB a Oth
Overall
4,
Exposure Group C
SUMMARY:
NPC, SNC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-727 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Shangina et al.
(2006)
Europe
Multicenter
case-control
study.
316 male cases of
laryngeal cancer between
the ages of 15-79 years
residing in four European
countries that were
diagnosed during 1999-
2002 and identified by
study centers in Romania,
Poland, Russia, and
Slovakia. 728 male
hospital controls selected
within six months of case
recruitment from
diagnoses excluding
disease related to alcohol
or tobacco. Controls
frequency matched by age
+/- 3 years.
Occupational histories
obtained by interview
and yielded information
on all jobs held >1 year.
A general questionnaire
obtained information of
job titles, tasks,
industries, starting and
stopping times, full-
time/ part-time status,
working environments,
and specific exposures.
A specific questionnaire
was completed for
employment in defined
jobs or industries.
Diagnosis of
laryngeal cancer
was
histologically or
cytologically
confirmed and
included
topographic
subcategories
from ICD-O
code C32
(glottis,
supraglottis,
subglottis,
laryngeal
cartilage,
overlapping
lesion of the
larynx, and
larynx,
unspecified).
Controlled for age,
country, smoking,
and alcohol.
Other exposures
that were found to
be risk factors
included dusts of
"hard alloys" (16
cases) and
chlorinated
solvents (15 cases).
As formaldehyde,
hard alloy dust and
chlorinated
solvents were each
found in fewer
than 6% of cases,
the correlation
between them is
considered to be
small enough to
make confounding
unlikely.
Logistic regression;
ORs (95% CI).
Latency was
evaluated.
Larynx:
18/316
exposed
The power
to evaluate
formaldehy
de as a
hazard is
diminished
as fewer
than 10%
of cases
had any
exposure
to
formaldehy
de.
SB IB Cf Oth
4-
Exposure Group C
Low power due to rarity of
exposure
SUMMARY:
Larynx: MEDIUM 4,
(Potential bias 4,
low sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-728 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Talibov et al.
(2014)
Europe
Mul ticoun try
case-control
study.
Individuals from Finland,
Iceland, Norway, and
Sweden who were
recorded in various
censuses from 1960 -
1990. Acute myeloid
leukemia cases
identified by national
registries up until 2003-
2005 depending on the
country.
Occupational history
from census records
were linked to the
Nordic Occupational
Cancer Study (NOCCA)
JEM to code each
cohort member as
exposed to
formaldehyde.
Exposures were
quantified based on
the proportion of
people in each
occupation considered
to be exposed and the
mean level of
exposure during
specific time periods.
8% of AML cases and
controls were
exposed.
Co-exposures to
solvents was
evaluated.
Diagnosis of
incident cancer
reported to the
National
Cancer
Registries.
Controlled for age
(<50, 50+), sex, and
solvents.
Solvents included:
aliphatic and
alicyclic
hydrocarbons,
aromatic
hydrocarbons,
benzene, toluene,
trichloroethylene,
111-
trichloroethane,
methylene
chloride,
perchloroethylene,
other organic
solvents, and
ionizing radiation.
HRs (95% CI).
A 10-year latency
period was assumed.
AML:
1201/15,33
2 exposed
SB IB a Oth
Overall
4-
Exposure Group D
SUMMARY: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-729 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Teschke et al.
(1997)
Canada
Population-
based case-
control study
of nasal
cancer.
48 incident cases of
nasal cancers (31%
female) older than 19
years, 1990-1992.
Controls were randomly
selected from age and
sex strata of voter list of
the same time period.
6 of 54 cases (11%) were
excluded for lack of
interview as were 36 of
195 controls (18%).
Controls matched on
age and sex.
Standardized
questionnaire
including
occupational,
residential, smoking,
and medical histories
aimed at identifying
exposures considered
to be probably
carcinogenic by IARC.
Occupational data
reviewed by an
industrial hygienist
blinded to case-status.
EPA considered that
workers in the textile
and pulp and paper
mill industries may
have been exposed to
formaldehyde but the
exposure
questionnaire did not
identify them as
exposed.
Pulp and paper mill
workers may also be
co-exposures to dioxin
or perchloroethylene
(Kauppinen et al.,
1997 IAOEH;70:119-
127).
Incident cases
from British
Columbia
Cancer Agency
registry.
Histologically
confirmed
primary
malignant
tumors of the
nasal cavity.
SNC: ICD-0
160.
Controlled for age
and sex.
More than 40
specific
occupational
groups were
evaluated without
control of
confounding.
Confounding not
evaluated.
Potential
confounders for
these outcomes
include
chlorophenols, acid
mists, dioxin, and
perchloroethylene
and would likely be
positively
correlated with
formaldehyde
exposure.
However, on acids
mists are
associated with
URT cancers.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
ORs (95% CIs).
Latency was
evaluated.
SNC: 48
3 cases
exposed to
pulp and
paper mills.
SB IB Cf Oth
Overall
Exposure Group C
Potential confounding for
pulp and paper mill
workers
Low power due to rarity of
exposure
SUMMARY:
SNC: LOW 4,
(Potential bias 4,
low sensitivity)
This document is a draft for review purposes only and does not constitute Agency policy.
A-730 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Vaughan et al.
(2000)
United States
Population-
based case-
control study
of
nasopharyngea
I cancer.
196 cases (32% female)
ages 18-74 diagnosed
during 1987-1993
identified from five
population based cancer
registries.
Interviews completed
for 82% of cases and
76% of the 244 controls.
19% of case interviews
completed by next of
kin.
Controls selected by
random digit dialing in
the same geographical
region frequency
matched by age, sex,
and cancer registry.
Individual-level
exposure based on
industrial hygienist
review of detailed
occupational histories
including industry, job
title, duties and dates
used to estimate
probability, intensity,
and cumulative
exposure.
Incident cases.
Diagnosis of
nasopharyngea
I (any
histological
type) based on
clinical records.
Histological
typing
reported.
Controlled for age,
sex, race, registry,
smoking, proxy
status, and
education.
Wood dust
evaluated as an
independent risk
factor for NPC
controlling for
formaldehyde and
it was not a risk
factor in this data
set. Therefore,
wood dust should
not be a
confounder in this
data set.
Logistic regression;
ORs (95% CI) by
probability of
exposure, duration,
and cumulative
exposure.
Separate analyses by
histological type.
Latency evaluated.
NPC: 79
exposed
cases.
IB a Oth
Exposure Group B
SUMMARY:
NPC: MEDIUM 4,
This document is a draft for review purposes only and does not constitute Agency policy.
A-731 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
Participants,
Consideration of
Analysis and
setting, and
selection, and
Exposure measure
Outcome
likely
results (estimate
Study
design
comparability
and range
measure
confounding
and variability)
sensitivity
Vaughan
231 cases (32% female)
Individual-level
Incident cases.
Controlled for age,
Logistic regression;
NPC: 3/21
(1989)
ages 20-74 years
exposure based on job
Diagnosis of
sex, smoking, and
ORs (95%CI).
exposed
United States
residing in the area
exposure matrix by
squamous cell
alcohol.
OHPC:
covered by Washington
occupation and
cancers of the
Duration of
11/183
Population-
State Cancer
industry for each
pharynx and
NPC analyses
employment and
exposed
based, case
Surveillance System
individual job used to
sinonasal cavity
controlled for race.
occupation are
SNC: cases
control study
during 1980-1983.
estimate probability
based on
surrogates for
included in
of squamous
and intensity of
review of
Wood dust is
intensity of
Luce et al.
cell cancers of
Participation for all cases
exposure.
hospital
associated with
exposure.
(2002).
the pharynx
was 69% (See Vaughan
medical
URT cancers and
and sinonasal
et al., 1986a) and 80.0%
Formaldehyde
records,
would likely be
Latency was
Low power
cavity.
for controls (n=552).
exposure from
surveillance of
positively
evaluated.
for NPC
available industrial
radiotherapy
correlated with
and SN.
Related
=50% of cases interviews
hygiene data, NIOSH
and pathology
formaldehyde
studies:
completed by next of
and other data, and
practices, and
exposure, but
Vaughan et al.
kin.
NCI job exposure
state death
strongest
(1986a, b);
Controls selected by
linkage system.
certificates.
association is with
Included in
random digit dialing in
SNC.
Luce et al.
same residential area as
Occupation as a
(2002)
cases and were
frequency matched on
age and sex with at 2
controls per cases in
each 5-year age and sex
category. May result in
poorer quality exposure
data and a bias towards
the null.
carpenter or
employment in the
"lumber and wood
product
manufacturing"
industry presumed to
be exposed to
formaldehyde.
Potential for
confounding is
unknown but could
have inflated the
observed effect.
Evaluation of major bias
categories
SB IB Cf Oth
Overall
Potential selection issue
(>40% cases represented
by next of kin)
Exposure Group D
Confounding possible
Low power for NPC
SUMMARY:
NPC: LOW 4,
(Low sensitivity
potential bias 4/)
OHPC: LOW
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-732 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Vaughan et al.
(1986a)
United States
Population-
based, case
control study
of cancers (all
types) of the
pharynx and
sinonasal
cavity.
Related
studies:
Vaughan et al.
(1989; 1986b);
SNC cases
included in
Luce et al.
(2002) but not
here.
285 cases (35% female)
ages 20-74 years
residing in the area
covered by Washington
State Cancer
Surveillance System
during 1980-1983.
Participation for all cases
was 69% and 80% for
controls (n=552).
=50% of cases interviews
completed by next of
kin.
Controls selected by
random digit dialing in
same residential area as
cases and were
frequency matched on
age and sex with at 2
controls per cases in
each 5-year age and sex
category.
Individual-level
exposure based on job
exposure matrix by
occupation and
industry for each
individual job used to
estimate probability
and intensity of
exposure.
Formaldehyde
exposure from
available industrial
hygiene data, NIOSH,
and other data, and
NCI job exposure
linkage system.
Incident cases.
Diagnosis of
squamous cell
cancers of the
pharynx and
sinonasal cavity
based on
medical
records,
surveillance of
radiotherapy
and pathology
practices, and
state death
certificates.
2% of cases
were
nonsquamous
cell cancers
(Vaughan,
1989).
Controlled for age,
sex, smoking, and
alcohol.
NPC analyses
controlled for race.
Wood dust is
associated with risk
of URT cancer and
was not evaluated
as a confounder.
However, as this is
a case-control
study the
correlation
between
formaldehyde and
wood dust is
expected to be
small and thus
wood dust would
not be expected to
be a confounder.
Logistic regression;
ORs (95%CI).
Latency was
evaluated.
NPC: 11/27
occupation
ally
exposed.
OHPC:
58/205
occupation
ally
exposed.
SNC: cases
included in
Luce et al.
(2002).
SB IB Cf Oth
4-
Potential selection issue
(>40% cases represented
by next of kin)
Exposure Group B
downgraded to D due to
additional measurement
error from next-of-kin
interviews.
Confounding possible for
SNC but less so for NPC
and OHPC
SUMMARY:
OHPC, NPC: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-733 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Vaughan et al.
(1986b)
United States
Population-
based, case
control study
of cancers (all
types) of the
pharynx and
sinonasal
cavity.
Related
studies:
Vaughan et al.
(1989; 1986a);
SNC cases
included in
Luce et al.
(2002) but not
here.
285 cases (35% female)
ages 20-74 years
residing in the area
covered by Washington
State Cancer
Surveillance System
during 1980-1983.
Participation for all cases
was 69% (See Vaughan
et al., 1986a) and 80%
for controls (n=552).
=50% of cases interviews
completed by next of
kin.
Controls selected by
random digit dialing in
same residential area as
cases and were
frequency matched on
age and sex with at 2
controls per cases in
each 5-year age and sex
category.
Presumed exposure to
formaldehyde based
on structured
telephone interview
information on
occupational and
residential history.
Interview-based
information on
lifetime residential
history from cases,
next of kin, and
controls.
Incident cases.
Diagnosis of
squamous cell
cancers of the
pharynx and
sinonasal cavity
based on
medical
records,
surveillance of
radiotherapy
and pathology
practices, and
state death
certificates.
2% of cases
were
nonsquamous
cell cancers
(Vaughan,
1989).
Controlled for age,
sex, smoking, and
alcohol.
NPC analyses
controlled for race.
Wood dust is
associated with risk
of sinonasal cancer
and was not
evaluated as a
confounder.
However, as this is
a case-control
study the
correlation
between
formaldehyde and
wood dust is
expected to be
small and thus
wood dust would
not be expected to
be a confounder.
Logistic regression;
ORs (95% CI).
Latency was
evaluated.
NPC: 8/27
lived in
mobile
home.
10/27
exposed to
particleboa
rd.
OHPC:
28/205
lived in
mobile
home.
68/205
exposed to
particleboa
rd.
SNC: cases
included in
Luce et al.
(2002).
SB IB Cf Oth
4-
Potential selection issue
(>40% cases represented
by next of kin)
Exposure Group B
downgraded to D due to
additional measurement
error from next-of-kin
interviews.
Confounding possible for
SNC but less so for NPC
and OHPC
SUMMARY:
OHPC, NPC: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-734 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
West et al.
(1993)
Philippines
Hospital-based
case-control
study.
Related study:
Hildesheim et
al. (1992)
104 cases (27% female),
11-83 years old,
predominantly non-
Chinese, from the
Philippine General
Hospital diagnosed
before 1992.
100% of cases
participated. All 104
hospital controls
participated while only
77% of 101 community
controls participated
(Hildesheim et al., 1992).
Hospital controls were
matched on age, sex,
and hospital ward type
(private/public).
Community controls
were matched on age,
sex, and neighborhood
of residence.
Lifetime job history;
details not provided.
Occupational exposure
to formaldehyde
classified by blinded
industrial hygienist as
likely or unlikely to be
exposed; appendix
provides
formaldehyde
exposure rating for
each job category.
Incident cases.
Diagnosis of
NPC
pathologically
confirmed by
histological
review for all
cases.
Histological
typing not
reported.
Controlled for age,
sex, hospital ward
type (or
neighborhood), for
education, years
since first exposure
to dust and
exhaust fumes, diet
including
processed meats,
fresh fish, smoking,
anti-mosquito
coils, and herbal
medicines.
Note that anti-
mosquito coils emit
formaldehyde
0.87- 25 ng/m3 (Liu
et al., 2003).
Controlling for
mosquito coils may
have underestimated
to effect of
formaldehyde.
Conditional logistic
regression; ORs
(95% CI).
Latency was
evaluated.
NPC:
27/104
exposed
SB IB Cf Oth
Exposure Group C
Controlling for exposure to
mosquito coils which emit
formaldehyde may
underestimate the effect
of other formaldehyde
exposures in the
regression analysis.
SUMMARY:
NPC: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-735 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Wortley et al.
(1992)
United States
Population-
based, case
control study
of cancers (all
types) of the
larynx.
235 cases (21% female)
ages 20-74 years
residing in the area
covered by Washington
State Cancer
Surveillance System
during 1983-1987.
Participation for all cases
was 81% and 80% for
controls (r?=547).
7% of cases interviews
completed by next of
kin.
Controls selected by
random digit dialing in
same residential area as
cases and were
frequency matched on
age and sex with at 2
controls per cases in
each 5-year age and sex
category.
Individual-level
exposure based on job
exposure matrix by
occupation and
industry for each
individual job used to
estimate duration and
intensity of exposure.
Formaldehyde
exposure from
available industrial
hygiene data, NIOSH,
and other data, and
NCI job exposure
linkage system.
Incident cases.
Diagnosis of
cancer of the
larynx based on
medical
records,
surveillance of
radiotherapy
and pathology
practices, and
state death
certificates.
94.5% of cases
were
squamous cell
cancers.
Controlled for age,
smoking, and
alcohol. Further
adjustment for sex
did not change
results.
Other exposures:
asbestos, chromium,
nickel, cutting oils,
and diesel fumes.
High risk
occupations (e.g.,
mechanics,
carpenters, painters,
textile machine
operators) likely had
co-exposures to
unidentified
substances.
However, as this is
a case-control
study the
correlation
between
formaldehyde and
those potential
confounders is
expected to be
small and thus
wood dust would
not be expected to
be a confounder.
Logistic regression;
ORs (95%CI).
Latency was
evaluated.
Larynx:
58/235
occupation
ally
exposed
SB IB Cf Oth
Overall
-l
Exposure Group C
SUMMARY:
Larynx: MEDIUM 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-736 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Yang et al.
(2005)
Taiwan
Family-based
case-control
study.
Related
studies:
Hildesheim et
al. (1997;
2001),
Cheng et al.
(1999)
502 cases recruited from
265 families with 2 or
more NPC cases
identified from earlier
study (Hildesheim et al.,
2001). Additional cases
obtained from hospitals
that treat NPC.
Occupational data
available for 65% of
cases and 57% of
controls.
203 cases represented
by next of kin (>40%).
Cases were matched
with 2 groups: First with
1,944 familial controls;
and second with 327
population controls.
Lifetime job history
(jobs held for at least
one year since age 16);
job title, typical
activities/duties, type
of industry, and tools
and/or materials used.
Exposures coded by
industrial hygienist.
Exposures in 10 year
preceding diagnosis of
interview were
excluded.
Collected information
on cigarette smoking,
betel nut
consumption, wood
and formaldehyde
exposure, and
Guangdong and other
salted fish
consumption during
childhood.
Original case
series were
incident cases.
Unclear if
supplemental
cases were
incident or
prevalent.
Diagnosis NPC
confirmed by
histological
review on 502
cases from
national tumor
registry.
Three analyses
(check each and be
specific).
Family control
analysis controlled
for family, age, sex,
education, and
ethnicity.
This analysis did
not control for
partial matching on
education,
ethnicity, or area of
residence. Nor did
it control for
smoking, betel nut
consumption, or
wood.
In this study,
smoking was
inversely
associated with
NPC. Because
smoking is
positively
associated with
formaldehyde,
there may be
negative
confounding by
smoking in this
study.
Unconditional
logistic regression
(95%CI) controlling
for age and sex.
Lagged exposure
partially address
latency.
Controls used here
were originally
matched to an
earlier set of cases,
some of whom were
included here.
NPC: 502
Yu et al. (2004)
Hong Kong
Men and women.
Restaurant workers
(r?=l,225) who died
Occupational history
obtained from union
records. 415 deceased
Mortality:
Underlying
cause of
MOR with Internal
control group
adjusted for age at
Logistic regression.
Mortality odds ratios
(MORs) calculated
NPC: 21
SB IB
Cf Oth
Overall
_
4,
Potential selection issue
(>40% cases represented
by next of kin)
Exposure Group D
Negative confounding
possible
The impact of not
controlling for all matching
factors is unclear but
considered most likely to
bias towards the null and
inflate confidence
intervals.
SUMMARY:
NPC: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-737 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference,
setting, and
design
Participants,
selection, and
comparability
Exposure measure
and range
Outcome
measure
Consideration of
likely
confounding
Analysis and
results (estimate
and variability)
Study
sensitivity
Evaluation of major bias
categories
Mortality odds
ratio.
Related studies:
Ho et al. (2006),
EPD (1999)
during 1986-1995 and
were registered as union
members by 4 major
Chinese-style restaurant
workers' unions in Hong
Kong. Cause of death
available for more than
80% of restaurant
workers.
waiters and 140
deceased waitresses
and kitchen workers
likely exposed to
formaldehyde based on
independent studies of
air quality in service
areas of restaurants.
Authors discuss sources
of exposure.
Co-exposures include
Epstein-Barr virus
(EBV), smoking, salted
and preserved foods,
and other combustion
by-products.
death from
Hong Kong
Census and
Statistics
Department.
NPC: ICD-9
147
Histological
typing not
reported.
death, sex, year of
death, and place of
origin. Adjusted for
age at death, sex,
and year of death
for external control
group.
Most adults (90+ %)
are seropositive for
EBV and thus it
cannot be a
confounder.
Smoking was
evaluated as a
potential
confounder because
49% of staff smoked
compared to 27% of
population, but it
was insufficient to
explain the
observed effects.
Authors stated that
with free fresh food
available to
workers, the
availability of
preserved or salted
food was unlikely to
explain the
observed effect.
for waiters and
waitresses by
internal and external
controls and for
waiters, length of
union membership (a
surrogate for
duration of
exposure).
Latency was not
evaluated.
SB IB Cf Oth
Overall
4-
Exposure Group C Latency
not evaluated
Possible confounding by
smoking
SUMMARY:
NPC: LOW 4,
(Potential bias 4/)
This document is a draft for review purposes only and does not constitute Agency policy.
A-738 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Studies in Animals
Respiratory tract cancer
Similar to other sections, studies were evaluated and assigned the following confidence
ratings: High, Medium, or Low Confidence, and "Not Informative" based on expert judgement of each
study's methodological details related to predefined criteria within five study feature categories (see
Appendix A.1.1). In addition to the general considerations outlined in Appendix A.I.I., criteria
specific to evaluating respiratory tract cancer were evaluated (see Table A-107 for specific details).
With one exception (noted below), studies of experimental animals exposed for at least subchronic
duration (shorter exposure durations were not considered informative to this endpoint, given the
robust database), and which performed histopathological evaluations of respiratory tract tissues,
were evaluated. As these evaluations consider many of the same studies previously evaluated for
inclusion in the noncancer respiratory tract pathology section (see Appendix A.l.6), many parallels
exist between both sets of evaluations. While the important considerations across the two sections
are generally similar, several notable differences exist. For example, duration of exposure was seen
as more important for evaluations of dysplasia and neoplasms, as compared with evaluations of
noncancer respiratory tract lesions. Conversely, whereas a substantial emphasis was placed on the
characterization of the severity of the lesion for noncancer respiratory tract changes, severity was
not considered integral to the identification of cancers and dysplasia. Finally, although most studies
of respiratory pathology used paraformaldehyde or freshly prepared formalin as the test article,
some studies tested commercial formalin. While co-exposure to methanol is a major confounding
factor for systemic endpoints, it is considered to be less of a concern when identifying effects of
inhaled formaldehyde on respiratory pathology, (see Appendix A.l.6 for discussion) Because of the
abundance of animal respiratory pathology studies, only those ranked as having Robust or Adequate
exposure quality, and several ranked as having Poor exposure quality studies solely because they
tested formalin (see evaluations in Appendix A.1.2), were evaluated for their use in describing the
potential for formaldehyde inhalation exposure to cause respiratory tract cancers. Additional
considerations that might influence the interpretation of the usefulness of the studies during the
hazard synthesis are noted, including limitations such as the use of only one test concentration or
concentration that are all too high or too low to provide a spectrum of the possible effects, as well as
study strengths such as very large sample sizes or use of good laboratory practices (GLP); however,
this information typically did not affect the study evaluation decisions.
Studies are grouped according to exposure duration, and then organized alphabetically by
first author. If the conduct of the experimental feature is considered to pose a substantial limitation
that is likely to influence the study results, the cell is shaded gray; a "+" is used if potential issues
were identified, but these are not expected to have a substantial influence on the interpretation of
the experimental results; and a "++" denotes experimental features without limitations that are
expected to influence the study results. Specific study details (or lack thereof) which highlight a
This document is a draft for review purposes only and does not constitute Agency policy.
A-739 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 limitation or uncertainty in answering each of the experimental feature criteria are noted in the
2 cells. For those experimental features identified as having a substantial limitation likely to influence
3 the study results, the relevant study details leading to this decision are bolded.
This document is a draft for review purposes only and does not constitute Agency policy.
A-740 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table A-107. Evaluation of controlled inhalation exposure studies examining respiratory tract cancer or
dysplasia in animals
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Quality
Test Subjects3
Study Design13
Endpoint Evaluation
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
Criteria
relevant to
evaluating the
experimental
details within
each
experimental
feature
category
Exposure quality
evaluations (see B.4.1.2)
are summarized (++ =
"robust"; + = "adequate";
gray box = poor);
relevance of the tested
exposure levels is
discussed in the hazard
synthesis- studies without
tested exposure <15
mg/m3are highlighted
Sample size provides
reasonable power to
assess endpoint(s) in
question (e.g.,
>20/group desired);
species, strain, sex,
& age relevant to
endpoint; no overt
systemic toxicity
noted or expected
The study design is
appropriate and informative
for evaluating respiratory
tract cancer or dysplasia,
including a sufficient
exposure duration and/or
appropriate timing of
endpoint evaluations to
allow for cancer to develop,
and a lack of additional
modifying variables
introduced over the course
of the study. GLP-compliant
studies are highlighted
The protocols used to
assess respiratory
tract cancer or
dysplasia are sensitive
and complete (e.g.,
multiple tissues and
sections examined),
discriminating
(specific), &
biologically sound
(reliable);
experimenter bias
minimized (e.g., slides
blinded to evaluator")
Statistical methods,
group comparisons,
& data/variability
presentation are
appropriate &
discerning; mortality
data are described
Expert judgement
based on
conclusions from
evaluation of the 5
experimental
feature categories
Respiratory Tract Cancers—Chronic
(Appelman
et al.. 1988)
Rat
++
+
Small N (N=10);
Note: randomized
1-year duration short to
allow for cancer
development
+
Blinding of slides for
evaluation NR
++
Medium
[1 year duration]
This document is a draft for review purposes only and does not constitute Agency policy.
A-741 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
(Dalbev.
1982)
++
Note: 5h/d exposure;
days and timing of
exposure NR
++
++
Note: single concentration
(12.3 mg/m3) lifetime study
Blinding of slides for
evaluation NR; only 2
nasal sections;
limited reporting of
histopathology
methods; unclear if
dysplasia considered
+
Locations and
specific incidence of
lesions and other
minor details NR
Medium
[Limited sampling,
evaluation, and
reporting]
Hamster
Holmstrom et
al. (1989a)
Rat
++
Note: high concentration
exposure (15.3 mg/m3);
exposed nocturnally, in
contrast to other studies
+
Small N (N=15-16);
some cannibalism;
non-U RT tumors
-50% across groups
+
2/16 animals in
formaldehyde group
developed emphysema
Note: single concentration
(15.3 mg/m3) 2 yr study
++
Note: slides blinded
+
Locations of lesions
and other minor
details NR
Medium
[Some health
issues noted;
limited reporting]
(Kamata et
al.. 1997)
+
Formalin exposure, with a
methanol control
(assumed to be based on
levels in formalin)
Note: methanol
considered unlikely to
affect endpoint
+
Small N for interim
sacrifices (N=2-5)
Note: mortality rate
doubled at 18.3
mg/m3; exposure
begun at =PND35
++
Note: 2 yr study
+
Blinding of slides for
evaluation NR
++
Medium
[Formalin (with
methanol control)]
Rat
This document is a draft for review purposes only and does not constitute Agency policy.
A-742 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
(Kerns et al.,
1983)
Mouse
See also
(Battelle,
1982) and
(Swenberg
et al.,
1980b)
++
+
Survival to 18
months was <33% in
all groups (N is >25)
Note: randomized
++
Note: data from this study
based on a 2 yr GLP study
(CUT 1982)
+
Only three nasal
sections evaluated;
blinding of slides for
evaluation NR
+
Limited reporting of
dysplasia fin dings
High
[Note: somewhat
limited sampling
and high mortality]
(Kerns et al.,
1983)
Rat
See also
(Battelle,
1982) and
(Swenberg
et al.,
1980b)
++
+
Viral infection at
weeks 52-53
Note: considered
unlikely to influence
these outcomes;
randomized
++
Note: data from this study
based on a 2 yr GLP study
(CUT 1982)
+
Blinding of slides for
evaluation NR
Note: routine analysis
of nasal tissues only
+
Limited reporting of
dysplasia fin dings
High
[Note: transient
viral infection]
(Monticello
et al.. 1996)
Rat
++
++
Note: randomized
++
Note: 2 yr study
+
Blinding of slides for
evaluation NR
Note: routine analysis
of nasal tissues only
++
High
This document is a draft for review purposes only and does not constitute Agency policy.
A-743 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
(Sellakumar
et al., 1985)
+
Air controls direct into
chamber, not through
apparatus
Note: PFA in paraffin oil
(commonly used in
bubbler-type units); high
concentration exposure
(18.2 mg/m3)
++
++
Note: single concentration
(18.2 mg/m3) lifetime study
+
Blinding of slides for
evaluation not
specified
++
High
Rat
see also
(Albert et
al.. 1982)
(Woutersen
et al.. 1989)
++
++
Note: randomized
++
Note: 2 yr study
+
Blinding of slides for
evaluation NR; Note:
routine analysis of
nasal tissues only
++
High
Rat
Respiratory Tract Cancers—Subchronic (note: includes 1 study with only 8 weeks of exposure in genetically modified mice)
Andersen et
al. (2010)
Rat
+
Analytic concentrations
NR
Small N[N=8)
Note: randomized
13 wk duration with no
follow up to allow for cancer
+
Blinding NR; limited
reporting of slide
selection, analysis
methods, and number
of slides evaluated
+
Low
[Short duration;
small sample]
Arican et al.
(2009)
Rat
Analytical method and
concentrations NR
+
Small N (N=10)
Note: randomized
12 wk duration with no
follow up to allow for cancer
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[short duration;
exposure and
outcome methods
lacking]
This document is a draft for review purposes only and does not constitute Agency policy.
A-744 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
Casanova et
al. (1994)
Rat
++
Small N[N=3)
Note: randomized
12 wk duration with no
follow up to allow for cancer
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[short duration;
small N; outcome
methods lacking]
Coon et al.
1970
Dogs
++
Small N[N=2);
limited reporting
(e.g., age, weight,
health status, etc.)
Multiple species housed and
exposed simultaneously;
continuous exposure (>22
h/d); 90d study does not
allow for cancer to develop
Notes: single concentration
(4.6mg/m3) study
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[outcome
methods lacking;
short duration;
group housed for
exposure]
Coon et al.
1970
Guinea pig
++
NR age or number
of male vs female
guinea pigs; small N
[N= 15); limited
reporting (e.g., age,
weight, health
status, etc.)
Multiple species housed and
exposed simultaneously;
continuous exposure (>22
h/d); 90d study does not
allow for cancer to develop
Notes: single concentration
(4.6mg/m3) study
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[outcome
methods lacking;
short duration;
group housed for
exposure]
Coon et al.
1970
Monkey
++
Small N[N=3);
limited reporting
(e.g., age, weight,
health status, etc.)
Multiple species housed and
exposed simultaneously;
continuous exposure (>22
h/d); 90d study does not
allow for cancer to develop
Notes: single concentration
(4.6mg/m3) study
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[outcome
methods lacking;
short duration;
group housed for
exposure]
This document is a draft for review purposes only and does not constitute Agency policy.
A-745 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
Coon et al.
1970
Rabbit
++
Small N[N=2);
limited reporting
(e.g., age, weight,
health status, etc.)
Multiple species housed and
exposed simultaneously;
continuous exposure (>22
h/d); 90d study does not
allow for cancer to develop
Notes: single concentration
(4.6mg/m3) study
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[outcome
methods lacking;
short duration;
group housed for
exposure]
Coon et al.
1970
Rat
++
NR number of male
vs female nor how
many of each strain
exposed; limited
reporting (e.g., age,
weight, health
status, etc.)
Multiple species housed and
exposed simultaneously;
continuous exposure (>22
h/d); 90d study does not
allow for cancer to develop
Notes: single concentration
(4.6mg/m3) study
Blinding NR; slide
selection, analysis
methods, and
number of slides or
regions evaluated NR
+
Qualitative
descriptions only
Not informative
[outcome
methods lacking;
short duration;
group housed for
exposure]
Feron et al.
(1988)
Rat
++
Note: high concentration
exposure (> 12 mg/m3)
++
+
13 wk duration, but long-
term follow up to allow for
cancer to develop
+
Blinding NR; limited
reporting of analysis
methods
+
Limited information
(deaths only) to
inform timing of
tumor development
Medium
[Short duration of
exposure; limited
reporting]
Horton et al.
(1963)
Mouse
+
Analytic concentrations
NR
Note: excessive exposure
level (=200 mg/m3)
+
Limited reporting
(e.g., age, weight,
health status, etc.);
high mortality
35 wk duration with no
follow up to allow for
cancer; exposure paradigm
of lhr/wk considered less
informative
Nasal tissue not
examined; blinding
NR; limited reporting
+
Not informative
[Primary target
tissue not
examined; study
design limited]
Maronpot et
al. (1986)
Mouse
Formalin, methanol
concentrations NR, and
no controls
+
Small N (N=10)
Note: randomized
13 wk duration with no
follow up to allow for cancer
+
Blinding NR; limited
reporting of analysis
methods
++
Low
[Formalin; small
sample]
This document is a draft for review purposes only and does not constitute Agency policy.
A-746 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
National
Toxicology
Program
(2017)
Mouse
+
Analytic concentrations
NR
++
Note: "randomly
assigned"; Males
only; =25 mice/
group; genetically
modified (p53+/-)
8 wk exposure duration;
follow up for 32 wk
Note: although unclear if
exposure or follow up
duration was adequate, the
study employed maximally
tolerated cumulative dose
+
Blinding NR;
examined 3 nasal
cavity sections (and 1
larynx)
Note: 4 additional
pathologists reviewed
all tumor slides
++
Low
[very short (8
week) exposure
duration and
limited follow up
(32 wk) for cancer
development]
Rusch et al.
(1983)
Rat
++
Note: test article was not
stabilized (negligible
methanol) formaldehyde;
concentration <3.6 mg/m3
++
26 wk duration with no
follow up to allow for cancer
+
Blinding NR; limited
reporting of analysis
methods
++
Low
[Short duration of
exposure with no
follow up]
Rusch et al.
(1983)
Monkey
++
Note: test article was not
stabilized (negligible
methanol) formaldehyde;
concentration <3.6 mg/m3
++
26 wk duration with no
follow up to allow for cancer
+
Blinding NR; limited
reporting of analysis
methods
++
Low
[Short duration of
exposure with no
follow up]
Rusch et al.
(1983) -
Hamster
++
Note: test article was not
stabilized (negligible
methanol) formaldehyde;
concentration <3.6 mg/m3
++
26 wk duration with no
follow up to allow for cancer
+
Blinding NR; limited
reporting of analysis
methods
++
Low
[Short duration of
exposure with no
follow up]
Wilmer et al.
(1989)
Rat
+
Analytic concentrations
NR
Note: concentration
tested <5 mg/m3
++
Note: randomized
13 wk duration with no
follow up to allow for cancer
+
Blinding NR
++
Low
[Short duration of
exposure with no
follow up]
This document is a draft for review purposes only and does not constitute Agency policy.
A-747 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature limitations
are indicated.
Exposure Qualitv
Test Subjects3
Studv Design13
Endooint Evaluation0
Data Considerations
& Statistical
Analvsisd
Overall
Confidence Rating
Regarding the Use
for Hazard IDe
(Woutersen
etal.. 1987)
++
+
Small N (N=10)
Note: randomized
13 wk duration with no
follow up to allow for cancer
+
Blinding NR
++
Low
[Short duration of
exposure with no
follow up]
Rat
Zwart et al.
(1988)
Rat
++
Note: concentration <3.6
mg/m3
++
13 wk duration with no
follow up to allow for cancer
+
Blinding NR
+
Qualitative
descriptions only
Low
[Short duration of
exposure with no
follow up]
NR = not reported; N/A = not applicable
x Although blinding of slides for evaluation is considered important, it is identified as only a minor limitation for these endpoints, as the pathology is expected
to be overt and not reliant on subtle quantitative (e.g., cell counting) or qualitative (e.g., slightly increased proliferation) decisions that would be highly
impacted by potential evaluator biases.
aGray = inadequate N (N= 1 or 2) or multiple less essential study details (e.g., sex, strain) NR; + = inadequate N (e.g., N= >2 to <10) or individual less essential
study details NR; ++ = adequate N (using guidance from OECD TG 452 and TG 413: chronic: >20 animals/sex/group; subchronic: 10 animals/sex/group,
respectively).
bGray = test protocols for assessing endpoints could not be evaluated or had critical flaws, timing of exposures expected to compromise the integrity of the
protocols, protocols completely irrelevant to human exposure; + = informative components of the protocol were NR/insufficiently assessed, limited human
relevance or single concentration study; ++ = protocol considered relevant to human exposure.
cGray = uncontrolled variables are expected to confound the results or lack of reporting for lesion incidence and severity; + = limited information provided for
observed lesions (i.e., incidence and/or severity) uncontrolled variables may significantly influence results; ++ = adequate reporting of data, no potential
confounding identified.
dGray = failure to report a sufficient amount of data to verify results; + = failure to report statistical analyses; ++ = adequate reporting of data,
designation for the Use for Hazard ID based on EPA judgment and the following criteria: gray = the presence of generally >2 gray boxes in the study feature
categories; low = failure in 2 categories; medium = failure in 1 category; high = no category failures; the presence of multiple +'s may demote tier level.
1 Lymphohematopoietic cancers
2 Studies examining LHP cancers were evaluated using nearly identical approaches and criteria as those for respiratory cancers
3 (above). One notable difference involved a consideration of the test article as a key component of the review, as co-exposure to methanol
This document is a draft for review purposes only and does not constitute Agency policy.
A-748 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 in studies using formalin could have a substantial impact on the interpretation of potential LHP cancers (see exposure quality evaluation
2 in Appendix A.1.2). A minor difference involved the preference for microscopic examination of several tissues applicable to assessing
3 potential LHP cancers, and a preference for blinded assessment of the slides.
Table A-108. Evaluation of controlled inhalation exposure studies examining lymphohematopoietic cancers in
animals
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects
Studv Design
Endooint Evaluation0
Data
Considerations &
Statistical
Analvsis
Overall
Confidence
Rating Regarding
the Use for
Hazard ID
Criteria relevant
to evaluating the
experimental
details within
each
experimental
feature category
Exposure quality
evaluations (see
B.4.1.2) are
summarized (++ =
"robust"; + =
"adequate"; gray box =
poor); relevance of the
tested exposure levels
is discussed in the
hazard synthesis-
studies without tested
exposure <15 mg/m3
are highlighted
Sample size
provides
reasonable power
to assess
endpoint(s) in
question (e.g.,
>20/group
desired); species,
strain, sex, & age
relevant to
endpoint; no overt
systemic toxicity
noted or expected
The study design is
appropriate and
informative for evaluating
LHP cancer or dysplasia,
including a sufficient
exposure duration and/or
appropriate timing of
endpoint evaluations to
allow for cancer to
develop, and a lack of
additional modifying
variables introduced over
the course of the study.
GLP-compliant studies are
highlighted
The protocols used to
assess LHP cancer or
dysplasia are sensitive
and complete (e.g.,
multiple tissues and
sections examined),
discriminating (specific), &
biologically sound
(reliable); experimenter
bias minimized (e.g.,
slides blinded to
evaluator*)
Statistical methods,
group comparisons,
& data/variability
presentation are
appropriate &
discerning; mortality
data are described
Expert judgement
based on
conclusions from
evaluation of the 5
experimental
feature categories
This document is a draft for review purposes only and does not constitute Agency policy.
A-749 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects
Studv Design
Endooint Evaluation0
Data
Considerations &
Statistical
Analvsis
Overall
Confidence
Rating Regarding
the Use for
Hazard ID
(Kamata et
al.. 1997)
Rat
+
Formalin exposure,
with a methanol
control
+
Small Nfor
interim sacrifices
(N=2-5); Note:
mortality rate
doubled at 18.3
mg/m3; exposure
begun at
=PND35
++
Note: 2 yr study
+
Blinding of slides for
evaluation NR; specific,
routine histopathology
of several tissues
relevant to LHP cancer
(e.g., femur)
++
Medium
[Formalin (with
methanol
control)]
(Kerns et al.,
1983)
Mouse
See also
(Battelle,
1982)and
(Swenberg et
al.. 1980b)
++
+
Survival to 18
months was
<33% in all
groups (N is >25)
Note:
randomized
++
Note: relevant data
from the 2-yr GLP study
report (CUT 1982;
Batelle-Columbus,
1982)
+
Blinding of slides for
evaluation NR;
reported gross lesions
only
+
Limited reporting
High
[Note: somewhat
limited sampling
for potential LHP
cancers and high
mortality]
(Kerns et al.,
1983)
Rat
See also
(Battelle,
1982)and
(Swenberg et
al.. 1980b)
++
+
Viral infection at
weeks 52-53
Note: considered
unlikely to
influence these
outcomes;
randomized
++
Note: relevant data
from the 2-yr GLP study
report (CUT 1982;
Batelle-Columbus,
1982)
+
Blinding of slides for
evaluation NR;
reported gross lesions
only
+
Limited reporting
High
[Note: transient
viral infection;
limited sampling
for potential LHP
cancers]
This document is a draft for review purposes only and does not constitute Agency policy.
A-750 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Experimental Feature Categories
The study details leading to identification of major (bolded) or minor (italicized) experimental feature
limitations are indicated.
Exposure Qualitv
Test Subjects
Studv Design
Endooint Evaluation0
Data
Considerations &
Statistical
Analvsis
Overall
Confidence
Rating Regarding
the Use for
Hazard ID
National
Toxicology
Program (2017)
Mouse
+
Analytic
concentrations NR
++
Note: "randomly
assigned"; Males
only; =25 mice/
group;
genetically
modified (p53+/-
)
8 wk exposure
duration;
follow up for 32 wk
Note: although unclear
if exposure or follow up
duration was adequate,
the study employed
maximally tolerated
cumulative dose;
however, no increase
in any tumors noted
(even nasal SCCs, which
were the focus of the
study hypothesis)
+
Blinding NR; slide
evaluation details NR,
but assessed multiple
relevant tissues
Note: 4 additional
pathologists reviewed
all tumor slides
++
Low
[very short (8
week) exposure
duration and
limited follow up
(32 wk) for cancer
development]
(Sellakumar
et al., 1985)
Rat
see also
(Albert et al.,
1982)
+
Air controls direct
into chamber, not
through apparatus
Note: PFA in paraffin
oil (commonly used
in bubbler-type
units); high
concentration
exposure (18.2
mg/m3)
++
++
Note: single
concentration (18.2
mg/m3) lifetime study
Does not appear to be
an explicit, routine
examination of tissues
relevant to LHP
cancers, or an
evaluation of bone
marrow, in particular
("histologic sections
were prepared from...
other organs where
gross pathology was
present"); Blinding of
slides for evaluation
not specified
++
Low
[no routine
examination of
tissues relevant to
LHP cancers, and
lack of evaluation
of bone marrow
specfically,
severely limits
detection ability]
This document is a draft for review purposes only and does not constitute Agency policy.
A-751 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
Supporting Material for Carcinogenicity
Cancer sites for which data were reported that were not formally reviewed in this
assessment included lung, non-Hodgkin lymphoma, brain, bladder, colon, pancreas, prostate, and
skin cancers. A summary of the studies available on lung, non-Hodgkin lymphoma, and brain are
provided below for information. The data on bladder, colon, pancreas, prostate, and skin cancers
were sparse and, as such, these studies are not summarized.
Lung Cancer
Evidence describing an association between formaldehyde exposure and the risk of dying
from lung cancer is available from 28 epidemiologic studies (Checkaway et al., 2011; De Stefani et
al, 2005; Beane Freeman etal., 2013; Meyers etal., 2013; Coggon etal., 2014; Stern, 2003; Marsh et
al., 2001; Stellman etal., 1998; Band etal., 1997; Chiazze etal., 1997; Jakobsson etal., 1997;
Andjelcovich etal., 1995; Dell and Teta, 1995; Hansen and Olsen, 1995; Hayes etal., 1990; Partanen
et al., 1990; Gerin et al., 1989; Solet et al., 1989; Edling et al., 1987; Robinson et al., 1987; Logue et
al., 1986; Stroup etal., 1986; Bertazzi etal., 1986; Bond etal., 1986; Leibling et al., 1984; Levine et
al., 1984; Walrath and Fraumeni, 1984,1983b). Currently, these are the only primary studies that
provide informative evidence of the effect of formaldehyde exposure on the risk of dying from lung
cancer. A few studies are interpreted as unlikely to be informative (i.e., Fryzek et al. 2005; Hall et
al., 1991; Hansen etal., 1994; Harrington and Oaks, 1984; Wesselingetal., 1996), based on
considerations used to evaluate observational studies in the toxicological review.
Non-Hodgkin Lymphoma
The most specific level of non-Hodgkin lymphoma diagnosis that is commonly reported
across the epidemiologic literature has been based on the first three digits of the Eighth or Ninth
Revision of the ICD code [i.e., non-Hodgkin lymphoma ICD-8 and ICD-9: Codes 200 and 202 (WHO,
1967; 1977); however, early studies reported results for lymphosarcoma/reticulosarcoma alone
(ICD-8/9: Code 200)]. Evidence describing the association between formaldehyde exposure and
the specific risk of non-Hodgkin lymphoma was available from 19 epidemiologic studies—four
case-control studies (Tranah etal., 2009; Wang etal., 2009; Blair etal., 1993; Gerin etal., 1989) and
15 cohort studies (Coggon etal., 2014; Meyers etal., 2013; Beane Freeman etal., 2009; Band etal.,
1997; Andjelkovich et al., 1995; Hansen and Olsen, 1995; Dell and Teta 1995; Hayes et al., 1990;
Matanoski, 1989; Stellman et al. 1988; Robinson et al., 1987; Edling etal., 1987; Stroup etal., 1986;
Walrath and Fraumeni, 1984; 1983b). One study was interpreted as unlikely to be informative (i.e.,
Matanoski, 1989).
Brain Cancer
Evidence describing an association between formaldehyde exposure and the risk of dying
from brain cancer is available from 16 epidemiologic studies (Hauptmann et al., 2009; Beane
Freeman etal., 2013; Meyers etal., 2013; Coggon etal., 2003; Stellman etal., 1998; Band etal.,
This document is a draft for review purposes only and does not constitute Agency policy.
A-752 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
1997; Robinson etal., 1987; Dell and Teta, 1995; Hansen and Olsen, 1995; Andjelcovich etal., 1995;
Matanoski, 1989; Hayes etal., 1990; Stroup etal., 1986; Levine etal., 1984; Walrath and Fraumeni,
1984,1983b). Currently, these are the only primary studies that provide evidence of the effect of
formaldehyde exposure on the risk of dying from brain cancer. A few studies were interpreted as
unlikely to be informative (i.e., Hall et al., 1991; Hansen et al., 1994; Harrington and Oaks, 1984;
Wesseling et al., 1996).
Approaches for Cancer Mode of Action
Formal systematic approaches to identifying and evaluating the literature databases of
studies examining mechanistic data relevant to interpreting the potential for formaldehyde to cause
upper respiratory tract (URT) or lymphohematopoietic (LHP) cancers were not performed. Rather,
these sections consider studies identified through other health effect-specific literature searches,
and evaluate those studies in the context of the specific cancer etiology being considered.
Supplemental literature relevant to interpreting the biological relevance of some mechanistic data
was also identified from review articles and other national-level health assessments. These
sections rely heavily on searches and evaluations performed in the following sections: genotoxicity,
respiratory tract pathology, and integrated noncancer portal of entry mode of action (see
Appendices A.4, A.5.5, and A.5.6).
This document is a draft for review purposes only and does not constitute Agency policy.
A-753 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
APPENDIX B. INFORMATION IN SUPPORT OF THE
DERIVATION OF REFERENCE VALUES AND
CANCER RISK ESTIMATES
B.l. DOSE-RESPONSE ANALYSES FOR NONCANCER HEALTH EFFECTS
A thorough understanding of the exposure-response functions for any association between
exposure and health outcomes supports both the derivation of the traditional toxicity values (e.g.,
RfC) as well as potentially allowing for the estimation of risk above and below those values, and
thus provides a more comprehensive understanding of the effects of formaldehyde exposure on
various health outcomes. The following details on the estimation of points of departure for the
derivation of candidate reference concentrations (cRfCs) are provided to support the derivation of
toxicity values as well as to directly inform the potential computation of benefits analyses which
require detailed information describing the shape of the exposure-response function across a range
of exposures. Such benefits analyses may be used to support a variety of rulemakings.
The technical detail on dose-response evaluation and determination of points of departure
(POD) for relevant toxicological endpoints are provided in this Section. Some of the endpoints were
modeled using the U.S. EPA's Benchmark Dose Software (BMDS, version 2.2). The common
practices used in evaluating the model fit and selecting the appropriate model for determining the
POD, as outlined in the Benchmark Dose Technical Guidance Document (U.S. EPA, 2012) were used.
For some data, alternative methods were used, and these are noted as necessary in the summary of
the modeling results.
B.l.l. Evaluation of Model Fit Using BMDS models
For each dichotomous endpoint, BMDS dichotomous models were fitted to the data using
the maximum likelihood method. Each model was tested for goodness-of-fit using a chi-square
goodness-of-fit test (x2 p-value < 0.10 indicates lack of fit). Other factors were also used to assess
model fit, such as scaled residuals, visual fit, and adequacy of fit in the low-dose region and in the
vicinity of the BMR.
For each continuous endpoint, BMDS continuous models were fitted to the data using the
maximum likelihood method. Model fit was assessed by a series of tests as follows. For each model,
first the homogeneity of the variances was tested using a likelihood ratio test (BMDS Test 2). If Test
2 was not rejected (x2 p-value > 0.10), the model was fitted to the data assuming constant variance.
If Test 2 was rejected (x2 p-value < 0.10), the variance was modeled as a power function of the
mean, and the variance model was tested for adequacy of fit using a likelihood ratio test (BMDS
Test 3). For fitting models using either constant variance or modeled variance, models for the mean
This document is a draft for review purposes only and does not constitute Agency policy.
B-l DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Supplemental Information for Formaldehyde—Inhalation
response were tested for adequacy of fit using a likelihood ratio test (BMDS Test 4, withx2 p-value <
0.10 indicating inadequate fit). Other factors were also used to assess the model fit, such as scaled
residuals, visual fit, and adequacy of fit in the low-dose region and in the vicinity of the BMR.
B.1.2. Noncancer Estimates from Observational Epidemiology Studies
Derivation of BMC and BMCLfor Burning Eyes (Hanrahan et al., 1984)
Hanrahan et al. (1984) conducted a cross-sectional study and reported a concentration-
response relationship for the prevalence of ocular discomfort (i.e., burning eyes/eye irritation) in a
study of 61 teenage and adult residents of mobile homes in Wisconsin during July of 1979. In-home
formaldehyde measurements were obtained for all participants, and measured formaldehyde levels
(average of two approximately 1-hour air samples—one from the kitchen or living room and one
from a bedroom) were used to characterize average in-home exposures.
Hanrahan et al. (1984) reported that prevalent symptoms22 of burning eyes and eye
irritation were significantly associated with in-home formaldehyde exposures, and the authors
provided a graphical representation of the best-fitting logistic regression model results of predicted
prevalence of "burning eyes" for exposures at 100 ppb increments from 100 to 800 ppb. From
inspection of this graph, EPA determined the prevalence of burning eyes predicted at 100 ppb is
approximately 4%. While the published exposure-response results were shown truncated at 100
ppb, Hanrahan etal. (1984) reported that exposures ranged from <100 ppb to 800 ppb, and the
indoor median formaldehyde concentration was 160 ppb. Based on this information, it is
reasonable to assume that there were residential exposures below 100 ppb, and thus the
extrapolation of the published results below 100 ppb is considered to be based on measured
concentrations within the study's observed exposure range. Thus, it is possible to approximate the
functional form of the concentration-response relationship below 100 ppb from the graphical
results because what the investigators presented was the model predicted functional form for all
measured exposures. The reconstruction of that underlying functional form can show the results of
the same Hanrahan et al. (1984) model where they were omitted from the graphic below 100 ppb.
22Hanrahan et al. (1984) reported on the "prevalence" of symptoms; however, it is not clear if this was the "point
prevalence" of symptoms on the day of the formaldehyde sampling, or whether this was the "period prevalence" of
symptoms during the study period (July, 1979).
This document is a draft for review purposes only and does not constitute Agency policy.
B-2 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
FORMALDEHYDE CONCENTRRTI ON IN PPM
Figure B-l. Regression of prevalence of "burning eyes" versus indoor
formaldehyde concentration (ppm) in mobile homes (approximately 1-hour
air samples). Dashed lines show upper and lower 95th percentile confidence
intervals on model results.
1 In Figure 1, the dependent variable is displayed as a predicted percentage prevalence of
2 burning eyes. However, the general epidemiologic method used to model prevalence data is logistic
3 regression, which predicts the log odds of prevalence, which the authors then transformed to
4 prevalence for graphing. In order to describe the underlying functional form of the results
5 displayed, EPA converted the prevalence data back to prevalence odds. Table 1 shows the
6 prevalence values which EPA visually estimated from the plot, as well as the associated prevalence
7 odds, which EPA calculated as estimated prevalence divided by the complement of estimated
8 prevalence, that is p/(l-p). Figure 2 plots the estimated prevalence odds against the residential
9 concentration of formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
B-3 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-l. Concentration-response information for the central estimate of the
effect extracted from Hanrahan et al. (1984).
Residential formaldehyde
concentration (ppm)
Prevalence
(p)
Prevalence odds
(P/[1-P])
0.1
0.0375
0.039
0.2
0.175
0.212
0.3
0.35
0.538
0.4
0.52
1.08
0.5
0.66
1.86
0.6
0.725
2.64
0.7
0.8
4
0.8
0.85
5.67
Prevalence Odds of Burning Eyes by
Formaldehyde Concentration
y = 6
,1949x3 +
3.7689X2 + 0.0309X + 0.03
R2 - n QQQ1
•
••
I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Formaldehyde Concentration (ppm)
Figure B-2. Plot of the prevalence odds by residential concentration-response
information from Table 1.
1 In order to describe the underlying functional form of the model-predicted results from
2 Hanrahan et al. (1984), EPA fit polynomial trend lines from linear up to cubic functions with the
3 intercept fixed at a background prevalence of burning eyes of 3% 23 (using Microsoft Excel) to the
4 discrete prevalence odds data in Figure 2 and found that a third degree polynomial function fit with
5 an R2 value of 0.9991. This indicates nearly a perfect fit to the published model results. Such a high
23Setting the intercept to other value such as 0.01, 0.02, 0.03 made little difference (e.g., at 0.03, the R2 had the
same value of 0.9991, and the model was y=6.1949X3 + 3.7689x2 + 0.0309x + 0.03.
This document is a draft for review purposes only and does not constitute Agency policy.
B-4 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1
2
3
4
5
6
7
8
9
10
Table B-2. Concentration-response information for the upper bound on the
central estimate of the effect extracted from Hanrahan et al. (1984)
Residential formaldehyde
concentration (ppm)
Prevalence
(P)
Prevalence odds
(P/[1-P])
0.1
0.18
0.22
0.2
0.35
0.54
0.3
0.55
1.22
0.4
0.74
2.85
0.5
0.84
5.25
0.6
0.91
10.11
0.7
0.94
15.67
0.8
0.96
24.00
This document is a draft for review purposes only and does not constitute Agency policy.
B-5 DRAFT—DO NOT CITE OR QUOTE
value of R2 would not have been achieved from analysis of the raw data (unavailable), but the
objective here was to recreate the functional form of the modeled data presented by Hanrahan et al.
(1984). The following describes the functional form for the prevalence odds:
P
- = 6.1949 * (exposure)3 + 3.7689 * (exposure)2 + 0.0309 * (exposure) + 0.03
1 — p
(B-l)
Table 2 shows the prevalence values for the upper bound of the published concentration-
response function, which EPA visually estimated from the plot, as well as the associated prevalence
odds, which EPA calculated as estimated prevalence divided by the complement of estimated
prevalence, that is p/(l-p). Figure 3 plots the estimated prevalence odds against the residential
concentration of formaldehyde.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for Formaldehyde—Inhalation
Prevalence Odds of Burning Eyes by
Formaldehyde Concentration
30.00
y = 56.551X3 - 10.388x2 + 2.0796x + 0.03
9- 25.00 - R2 = 0.9995
— 20.00 -
~o
§ 15.00 -
o
C 10.00 JT'#
ro
£ 5.00 —
£ ••••¦'
o.oo •
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Formaldehyde Concentration (ppm)
y = 56.551x3 -
10.388x2 + 2.0796x + 0.03
r2 - n qqqi;
•
«*'
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.
Figure B-3. Plot of the upper bound on prevalence odds by residential
concentration-response information from Table 2.
In order to describe the underlying functional form of the model-predicted results from
Hanrahan et al. (1984), EPA fit polynomial trend lines from linear up to cubic functions with the
intercept fixed at zero (using Microsoft Excel) to the discrete prevalence odds data in Figure 3 and
found that a third degree polynomial function fit with an R2 value of 0.9995. This indicates nearly a
perfect fit to the published model results. The following describes the functional form for the
prevalence odds:
= 56.551 * (exposure)3 — 10.388 * (exposure)2 + 2.0796 * (exposure) + 0.03
1 — p
(B-2)
Selecting a benchmark response (BMR) for the derivation of a reference concentration (RfC)
involves making judgments about the statistical and biological characteristics of the data set A
BMR representing an extra risk of 10% is generally recommended as a standard reporting level for
quantal data. Biological considerations may warrant the use of a BMR of 5% or lower for some
types of effects (e.g., frank effects), or a BMR greater than 10% (e.g., for early precursor effects) as
the basis of the point of departure (POD) for a reference value (U.S. EPA, 2012).
EPA calculated the concentration at which a 10% extra risk of "burning eyes" would have
been observed in these data using the polynomial functions for the main effect to estimate the BMC
and for the upper-bound to estimate the BMCL. In this derivation, 10% extra risk is the benchmark
response (BMR) and the BMC and BMCL for a 10% BMR are noted as the BMCio and BMCLio. Note
that in Hanrahan et al. (1984), the prevalence of "burning eyes" was similar to that of "eye
This document is a draft for review purposes only and does not constitute Agency policy.
B-6 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 irritation." As there is little information available in the literature to estimate the background
2 prevalence of "burning eyes," the background prevalence of "burning eyes" was estimated at 3% (in
3 the absence of formaldehyde exposure) based on the prevalence of "eye irritation." A background
4 prevalence of 3% was considered to be a reasonable estimate. Sensitivity analyses using a
5 background prevalence of 1% and 2% were also evaluated and yielded BMC and BMCL estimates.24
6 Because the extra risk is a function of the prevalence in the exposed (PExposed) and the
7 prevalence in the unexposed (Punexposed) was estimated at 3%, EPA derived PExPosed for 10% extra
8 risk above background.
9 Extra Risk = 0.10 = [PExposed " Punexposed]/[1 " Punexposed] and Punexposed = 0.03, then PExposed = 0.127.
10 (B-3)
11 Because the exposure-response function from Hanrahan et al. (1984) is in terms of the
12 prevalence odds, that value is derived based on PExposed = 0.127. Thus, the prevalence odds =
13 [PExposed]/[l-PExposed] = 0.145. To derive the BMC, solve for the exposure value, which yields
14 prevalence odds of 0.145:
15 0.145 = 6.1949 * (exposure)3 + 3.7689 * (exposure)2 + 0.0309 * (exposure) + 0.03
16 (B-4)
17 Of the three roots, only one is within the exposure range of the data.
18 Exposure = 0.153 ppm formaldehyde = 0.188 mg/m3 formaldehyde [seefootnote25)
19 To derive the interim BMCL, solve for:
20 0.145 = 56.551 * (exposure)3 — 10.388 * (exposure)2 + 2.0796 * (exposure) + 0.03
21 (B-5)
22 Of the three roots, only one is within the exposure range of the data.
23 Exposure = 0.0706 ppm formaldehyde = 0.0868 mg/m3 formaldehyde
24 The BMCio is 0.188 mg/m3. The BMCLio is 0.0868 mg/m3.
24Using a 1% background prevalence to estimate the exposure-response function and the BMC, yields an estimate
of 0.154 ppm = 0.190 mg/m3 formaldehyde, and a BMCL estimate of 0.0768 = 0.0945 mg/m3; using a 2%
background prevalence to estimate the exposure-response function and the BMC, yields an estimate of 0.154 ppm
= 0.189 mg/m3 formaldehyde, and a BMCL estimate of 0.0739 = 0.0909 mg/m3.
^Concentration (mg/m3) = Concentration (ppm) * (Molecular mass/Molar volume) = 0.155 ppm * [(30.03
g/mol)/(24.45 L)] = 0.191 mg/m3 at 25°C.
This document is a draft for review purposes only and does not constitute Agency policy.
B-7 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Supplemental Information for Formaldehyde—Inhalation
Eye Irritation Data from Two Controlled Human Exposure Studies (Andersen, 1979; Andersen
and Molhave, 1983; Kulle et al., 1987; Kulle et al., 1993)
Modeling results are presented that support the derivation of PODs for sensory irritation
based on two controlled human exposure studies. Kulle et al. (1993) reanalyzed results of a study
of eye, nose, and throat irritation among participants exposed to 0, 0.5,1.0, 2.0, and 3.0 ppm for 3
hours once a week with exposure order randomly assigned. Another experimental study exposed a
group of 16 subjects to 0.3, 0.5,1.0, and 2.0 mg/m3 formaldehyde for 5-hour periods with a 2-hour
clean air exposure prior to each trial (Andersen, 1979; Andersen and Molhave, 1983). The order of
exposure concentrations was randomized. The occurrence of irritation symptoms during the clean
air exposure was not reported. Two sets of models were evaluated using the data from Andersen
(1979) and estimates of 0% and 3% for prevalence of irritation during the clean air exposure.
Table B-3. Benchmark dose modeling of sensory irritation using a BMR
of 10%
Model
BMD
BMDL
AIC
p-value
Best
Model
Notes
Andersen and Molhave, 1983 (Assumed response among controls = 0
Gamma
0.209
0.091
58.847
0.0488
Logistic
0.256
0.182
62.408
0.0665
Log Logistic
0.257
0.157
57.33
0.1429
X
Lowest AIC
Log Probit
0.249
0.153
57.965
0.1109
Multistage
0.137
0.068
60.321
0.0161
Multistage
0.137
0.068
60.321
0.0161
Probit
0.239
0.175
65.167
0.0469
Weibull
0.169
-0.077
59.527
0.0404
Quantal-
Linear
0.080
0.060
60.262
0.0247
Andersen and Molhave, 1983 (Assumed response among controls = 3%)
Gamma
0.304
0.142
77.217
0.1946
Logistic
0.201
0.148
76.388
0.0001
Log Logistic
0.369
0.219
74.821
0.4013
X
Lowest AIC
Log Probit
0.350
0.208
75.8
0.3202
Multistage
0.262
0.091
79.039
0.1145
Multistage
0.262
0.091
79.039
0.1145
Probit
0.196
0.149
77.859
0.0005
Weibull
0.233
0.108
78.456
0.1696
Quantal-
Linear
0.091
0.065
80.471
0.152
Kulle etal., 1993
Gamma
0.853
0.497
66.839
0.1819
Logistic
0.760
0.546
64.737
0.3644
Log Logistic
0.852
0.510
67.596
0.1465
Log Probit
0.850
0.541
67.254
0.1594
Multistage
0.676
0.395
65.090
0.3726
Multistage
0.863
0.369
66.134
0.226
Probit
0.694
0.502
64.645
0.3686
X
Lowest AIC
This document is a draft for review purposes only and does not constitute Agency policy.
B-8 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Model
BMD
BMDL
AIC
p-value
Best
Model
Notes
Weibull
0.886
0.501
66.225
0.2108
Quantal-
Linear
0.270
0.191
71.876
0.0629
Log-Logistic Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
dose
15:33 09/16 2015
Figure B-4. Log-logistic model with BMC of 10% extra risk over an assumed
background of 3% and lower confidence limit for the BMCL for prevalence of
conjunctival redness and/or nose or throat dryness; data from Andersen and
Molhave (1983).
This document is a draft for review purposes only and does not constitute Agency policy.
B-9 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-4. Parameter estimates for log-logistic model with BMC of 10% extra
risk over an assumed background of 3% and lower confidence limit for the
BMCL for prevalence of conjunctival redness and/or nose or throat dryness;
data from Andersen and Molhave (1983)
Variable
Estimate
Std. Err.
Lower conf. limit
Upper conf. limit
Background
0.1604
0.0715851
0.0200953
0.300704
Intercept
1.46207
0.609559
0.267359
2.65679
Slope
3.66848
1.12878
1.45611
5.88085
Table B-5. Observed and estimated values and scaled residuals for log-logistic
model with BMC of 10% extra risk over an assumed background of 3% and
lower confidence limit for the BMCL for prevalence of conjunctival redness
and/or nose or throat dryness; data from Andersen and Molhave (1983)
Dose
Est. Prob.
Expected
Observed
Size
Residual
0
0.1604
2.566
3
16
0.295
0.3
0.202
3.232
3
16
-0.144
0.5
0.3731
5.97
5
16
-0.501
1
0.842
13.472
15
16
1.047
2
0.985
15.76
15
16
-1.561
This document is a draft for review purposes only and does not constitute Agency policy.
B-10 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Probit Model, with BMR of 10% Extra Risk for the BMD and 0.95 Lower Confidence Limit for the BMDL
dose
15:25 09/16 2015
Figure B-5. Probit model with BMC of 10% extra risk and 95% lower
confidence limit for the BMCL for prevalence of eye irritation; data from Kulle
et al. (1987).
Table B-6. Parameter estimates for probit model with BMC of 10% extra risk
and 95% lower confidence limit for the BMCL for prevalence of eye irritation;
data from Kulle et al. (1987)
Variable
Estimate
Std. Err.
Lower conf. limit
Upper conf. limit
Intercept
-1.9161
0.36123
-2.6241
-1.20811
Slope
1.10331
0.222381
0.667453
1.53917
Table B-7. Observed and estimated values and scaled residuals for probit
model with BMC of 10% extra risk and 95% lower confidence limit for the
BMCL for prevalence of eye irritation; data from Kulle et al. (1987)
Dose
Est. Prob.
Expected
Observed
Size
Residual
0
0.0277
0.526
l
19
0.663
0.5
0.0862
0.862
0
10
-0.971
1
0.2082
3.955
5
19
0.59
2
0.6143
11.672
10
19
-0.788
3
0.9183
8.265
9
9
0.895
This document is a draft for review purposes only and does not constitute Agency policy.
B-ll DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information for Formaldehyde—Inhalation
Derivation of BMC and BMCL for PEFR in Children fKrzvzanowski etal. 1990)
A cross-sectional study of residential formaldehyde exposure in a large population-based
sample observed a linear relationship between increased formaldehyde exposure and decreased
peak expiratory flow rate (PEFR) among children exposed to average concentrations of 0.032
mg/m3 (26 ppb) (Krzvzanowski et al.. 1990). This study of effects in a residential population used a
thorough exposure assessment protocol and repeated measurements of PEFR, thus, enhancing the
ability to detect an association at the lower concentrations found in the homes. Declines in peak
expiratory flow rate (PEFR) were associated with increases in 2-week average indoor residential
formaldehyde concentrations, with greater declines observed in children (5-15 years of age, n =
208 in analytical data set) compared to adults fKrzvzanowski etal.. 19901. Mean formaldehyde
levels were 26 ppb (0.032 mg/m3), and more than 84% of the homes had concentrations 40 ppb
(0.049 mg/m3) and lower.
EPA calculated the concentration at which a 10% decrement in pulmonary function would
be expected. In this derivation, 10% decrement in a continuous response is considered to be the
benchmark response (BMR). A BMCioo/0 and BMCLioo/0 were determined from the regression
coefficient from a random effects model of PEFR among children reported by the study authors.
Statistical models which adjusted for important covariates (including smoking status, SES, NO2
levels, episodes of acute respiratory illness, and the time of day) did not identify any potential
confounders and those covariates were not included in the final model.
y = 349.6 — 1.28 * (household formaldehyde) — 6.1 * (morning) + 0.09
* (bedroom formaldehyde) * (morning) + 0.0031 * (bedroom formaldehyde)2
* (morning) + 4.59 * (morning) * (asthma) — 1.45 * (bedroom formaldehyde)
* (morning) * (asthma) + 0.031 (bedroom formaldehyde)2 * (morning)
* (asthma)
(B-6)
wherey = PEFR (L/min); household formaldehyde = 2-week household mean concentration;
morning = time of PEFR measurement (0,1); 2-week bedroom mean concentration; current asthma
= doctor's diagnosis and current status (0,1).
This document is a draft for review purposes only and does not constitute Agency policy.
B-12 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
For the purpose of deriving a point of departure for indoor formaldehyde, the primary
estimate of the point of departure was computed for household formaldehyde with morning = 0 and
asthma =0. The regression coefficient ((3) for household formaldehyde was -1.28 ± 0.46 L/minute-
ppb and the 95% one-sided upper bound on the regression coefficient was -2.04 L/minute-ppb;
/? — (critical value for one — tailed a of 0.05 * s. e. of /?) = —1.28 — (1.645 * 0.46) =
-2.04
(B-7)
Based on the background PEFR of 349.6 L/minute, a 10% decrement is 35 L. Dividing 35 L
by the regression coefficient for household formaldehyde of-1.28 L/minute-ppb (i.e., -1.28
L/(minute*ppb)), the change in formaldehyde concentration resulting in a 10% decrement in PEFR
is 27 ppb which is equivalent to 0.033 mg/m3. The BMCL resulting in a 10% decrease from a
background of 349.6 L/minute is 17 ppb (35 L/minute divided by -2.04 L/minute-ppb), which is
equivalent to 0.021 mg/m3.
In order to estimate how much more sensitive asthmatic children were to formaldehyde,
household and bedroom formaldehyde concentrations were assumed to be the same and morning =
1 and asthma = 1. Solving the final regression model for these realizations of household
formaldehyde, bedroom formaldehyde, morning, and asthma yield the following:
—35 L/min = —1.28 * (household formaldehyde) — 6.1 * (1) + 0.09
* (household formaldehyde) * (1) + 0.0031 * (household formaldehyde)2 * (1)
+ 4.59 * (1) * (1) — 1.45 * (household formaldehyde) * (1) * (1)
+ 0.031 (household formaldehyde)2 * (1) * (1)
(B-8)
which simplifies to:
—35—— = 0.0341 * (household formaldehyde)2 — 2.64 * (household formaldehyde) — 1.51
(B-9)
Solving for household formaldehyde yields a BMCi0o/„ (asthmatics) resulting in a 10%
decrease from a background PEFR of 349.6 L/minute of 16 ppb given that asthmatic children were
more sensitive to the respiratory effects of formaldehyde exposure than were children in general
who had BMCioo/0 of 27 ppb.
This document is a draft for review purposes only and does not constitute Agency policy.
B-13 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
Derivation of a BMC and BMCLfor Asthma Exacerbation in Children with Asthma (Venn et al.,
2003)
Venn et al. (2003) studied how indoor formaldehyde exposures affected the proportion of
childhood asthma cases who reported symptoms of asthma attacks (asthma exacerbation). During
an asthma attack, the muscles of the airways constrict thereby limiting air flow and the cells in the
airway produce mucus which further restricts the passage of air. Symptoms included any of the
following: wheezing, chest tightness, breathlessness, or cough (Venn et al., 2003). According to the
Centers for Disease Control and Prevention (Moorman et al., 2012), more than 50% of children with
asthma experienced at least one asthma attack in the previous 12 months yielding an annual rate of
asthma attacks in the general population of children of more than 5%. Approximately 10% of
children with asthma suffer an asthma attack resulting in a visit to the emergency room each year.
The annual mortality rate from asthma among children is 2-3 per million (Moorman et al., 2012).
Venn et al. (2003, see Table 4) divided the children's bedroom formaldehyde exposures into
quartiles and reported a statistically significant exposure-response trend of increasing risk of
symptoms of an asthma attack with increasing quartiles of formaldehyde concentrations (p=0.03)
and then fit a regression model to estimate the "per quartile" increase in risk. Venn et al. (2003)
identified similar exposure-response functions for night-time and daytime symptoms of an asthma
attack (asthma exacerbation) in children with asthma26: for night-time symptoms, the odds ratio
(OR) per exposure quartile increase in formaldehyde concentration was 1.45 (95% CI: 1.06 - 1.98);
for daytime symptoms, the OR per exposure quartile was 1.40 (95% CI: 1.00 - 1.94)27. Results were
adjusted for age, sex, and socioeconomic status. Dampness was also reported to be a risk factor for
symptoms of an asthma attack; however, further adjustment of the formaldehyde results for
dampness made little difference (Venn et al., 2003). No effect of other volatile organic compounds
or nitrogen dioxide on the risk of asthma attacks was found.
As the formaldehyde measures were taken in the children's bedrooms, the RfC derivation is
based on the exposure-response function for night-time symptoms of an asthma attack. The
following table summarizes the results from Venn et al. (2003) specific to the exposure-response
relationship for night-time symptoms of asthma attacks in children with asthma. Note that, by
definition, the OR reported for each exposure level is relative to the odds of being a case in the
reference category, which is the lowest quartile of exposure. In Venn et al. (2003), the reference
category is defined as exposures within the range 0-16 ng/m3. The median concentration within
this range was 12.24 |J.g/m3 (Venn, 2012). In order to estimate the OR per unit increase in
formaldehyde concentration from the reported effect per unit increase in quartile of formaldehyde
26Cases were defined as those whose doctors had prescribed asthma drug treatment at the time of the study
(including the preceding year) (Venn et al., 2003).
"Exposure measurements, pulmonary function measurements, and symptoms of asthma attacks were measured
over a 4-week period.
This document is a draft for review purposes only and does not constitute Agency policy.
B-14 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information for Formaldehyde—Inhalation
exposure, the difference in each quartile's median formaldehyde concentration was computed by
subtracting 12.24 |ig/m:i from each quartile median.
Table B-8. Modeled effect estimates for night-time symptoms of an asthma
attack; data from Venn et al., 2003
Quartile
Median >
Lower
Upper
Lower
Upper
Exposure
Quartile
Reference
Bound
Bound OR
Bound
Bound
Quartile3
Medianb
Quartile
OR by
OR by
by
Modeled
Modeled
Modeled
(pg/m3)
(pg/m3)
(pg/m3)
Quartile3
Quartile
Quartile
ORc
ORc
ORc
0-16
12.24
0
1
l
16.1-22
19.23
6.99
1.4
0.54
3.62
1.45
1.06
1.98
22.1-32
26.55
14.31
1.61
0.62
4.19
2.10
1.12
3.92
32+
41.02
28.78
3.33
1.23
9.01
3.05
1.19
7.73
a Venn et al. (2003);b Venn (2012);c Venn et al. (2003) OR per increasing quartile = 1.45 (95% CI: 1.06 - 1.98).
EPA considered multiple methodologies for identifying a point of departure for this health
endpoint. If the information provided by Venn etal. (2003) had been limited to just the quartile-
specific results, then the one method might have used the results from Table 4 of Venn etal. (2003)
which show the first statistically significant effect occurring in the highest exposure group with a
quartile mean of 41.02 ng/m3 which could represent the LOAEL and thus the corresponding NOAEL
could be the quartile mean of the third exposure group at 26.55 ng/m3. However, because Venn et
al. (2003) also reported a statistically significant exposure-response function (p-trend = 0.02) with
OR=1.45 per exposure quartile (95% CI: 1.06 - 1.98), it is not reasonable to assume there is no
effect at the median of the third quartile because the reported OR for this quartile was 1.61 (95%
CI: 0.62 - 4.19) and the reported exposure-response function corresponds to a modeled OR=2.10
(95% CI: 1.12 - 3.92). Likewise, for the second quartile with a quartile-specific result of 0R=1.4
(95% CI: 0.54 - 3.62), rather than evidence of "no effect," the reported exposure-response function
indicates a modeled OR = 1.45 (95% CI: 1.06 - 1.98), which is consistent with the second quartile-
specific results of OR = 1.4 but has narrower confidence intervals due to the use of data from all the
quartiles rather than just a comparison of the second quartile to the first
The reported exposure-response function from Venn et al. (2003) appears to be a more
precise estimate of the exposure-response relationship for night-time symptoms of poor asthma
control in children with asthma. In order to estimate a point of departure, the units of 'per quartile'
need to be defined in terms of "per ng/m3." As the magnitude of the increase in exposure from the
median of the first quartile to the median of the second quartile is 6.99 |J.g/m3, an estimate of the
effect of exposure per |J.g/m3 can be obtained by scaling the ln(OR) and its standard error by the
difference in quartile medians. The OR = 1.45 per quartile (95% CI: 1.06 - 1.98) is first converted to
the natural log scale as ln(OR) = 0.37156 per quartile (95%: 0.05827 - 0.68310), and then each
term is multiplied by unity as expressed by [(1 quartile)/(6.99 ng/m3)] to yield an effect of ln(OR) =
0.053156 (95% CI: 0.008336 - 0.09773), which when exponentiated back to the OR scale is
This document is a draft for review purposes only and does not constitute Agency policy.
B-15 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
equivalent to an OR = 1.05 per ng/m3 (95% CI: 1.01 - 1.10). This equivalent exposure-response
function in terms of "per ng/m3" retains the same p-trend value of 0.02 because the scaling cancels
out.
According to Table 4 in Venn et al. (2003), the prevalence of night-time asthma symptoms
among the cases in the reference group is 0.41. Because the symptoms of an asthma attack among
children with asthma is considered to be a frank effect (an overt of clinically apparent effect), a
BMR of 5% was used to derive the POD for the derivation of the RfC (U.S. EPA, 2012). Using a
BMR=5% extra risk for symptoms of an asthma attack, the prevalence of symptoms among the
exposed at 5% extra risk compared to the prevalence of symptoms at zero exposure is:
Extra Risk = 0.05 = [PExposed - Punexposed] ~ |1 Punexposed] and Punexposed = 0.41, then PExposed = 0.4395.
(B-10)
Find OR = [PExposed/(l - PExposed)]/[Punexposed/(l " Punexposed)]
= [0.4395/(1 - 0.4395)]/[0.41/(1 - 0.41)] = 1.13
(B-11)
For the derivation of the point of departure, here the benchmark concentration or BMC,
note that the exposure-response function is defined relative to the reference group (those exposed
to the first quartile of formaldehyde exposures) which experienced a median formaldehyde
concentration of 12.24 |J.g/m3 (Venn, 2012 personal communication). So in deriving the BMC, the
first step is to estimate the magnitude of the concentration above the reference concentration of
12.24 ng/m3, which corresponds to a 5% extra risk. For clarity, that value will be called the
"interim BMC05." The second step is to add that interim BMC5 to the median formaldehyde
concentration in the reference group. While it is possible that there are adverse effects of
formaldehyde below the median formaldehyde concentration in the reference group, it should be
understood that the methodology used in this derivation restricts the BMC to be greater than the
median formaldehyde concentration in the reference group. The alternative would be to
extrapolate the exposure-response function down from 12.24 |ig/m3to either the background
ambient formaldehyde concentration, or down to a concentration of zero.
To derive the interim BMC using the linear concentration-response function, solve for:
OR corresponding to a 5% extra risk = 1.13 = (1.05 per |ig/m3)*(Interim BMC5)
Interim BMC5 = 1.08 |J.g/m3
This document is a draft for review purposes only and does not constitute Agency policy.
B-16 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 To derive the interim BMCL using the linear concentration-response function, the one-sided
2 95% upper bound is needed (rather than the upper bound of the two-sided 95% CI around the OR).
3 Using the one-sided 95% upper bound, which is 1.09 (calculation below)28, solve for:
4 OR corresponding to a 5% extra risk = 1.13 = (1.09 per |ig/m3)*(Interim BMCLs)
5 Interim BMCLs = 1.04 ng/m3
6 Adding back the median formaldehyde concentration in the reference category (12.24
7 ng/m3), the BMCLs value is 13.28 |J.g/m3 and this value is selected as the point of departure for the
8 cRfC.
9 Reference: Moorman JE, Akinbami LJ, Bailey CM, et al. National Surveillance of Asthma:
10 United States, 2001-2010. National Center for Health Statistics. Vital Health Stat 3(35). 2012.
11 B.1.3. Noncancer Estimates from Animal Toxicology Studies
12 Analysis of Respiratory Pathology Data from F344 and Wistar Rats
13 This appendix provides support to the decisions and details of modeling the respiratory
14 pathology data in rats and mice in Section 2.1 for deriving candidate human inhalation RfCs based
15 on these endpoints. These involve the following endpoints and studies: squamous metaplasia in
16 F344 rats (Kerns et al. 1983), basal hyperplasia in Wistar rats (Woutersen etal.. 1989). and
17 squamous metaplasia in Wistar rats (Woutersen et al.. 1989).
28To calculate the standard error of the In(OR): [(ln(1.10)-ln(1.01)]/3.92=0.02178. Therefore, the 95% one-sided
upper bound of the In(OR) is [ln(OR)+l.645(0.02178)]=0.08461 and the 95% one-sided upper bound of the OR is
1.09.
This document is a draft for review purposes only and does not constitute Agency policy.
B-17 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Supplemental Information for Formaldehyde—Inhalation
Figure B-6. Midsaggital section of rat nose showing section levels (Kerns et al.
1983) (nostril is to the left).
Formaldehyde flux to the nasal lining was used in analyzing the dose-response data from
Kerns etal. (1983) at the Level 1 cross section (as shown in Figure B-6) of the F344 rat nose, which
is located in the front portion of the rat nose behind the nasal vestibule (Young 1981). Kimbell et al.
modeled formaldehyde flux to the nasal lining; their flux estimates are shown in Figure B-7 as a
contour plot of flux per ppm of exposure (note: only the lateral view of the three-dimensional
surface is presented). These figures indicate that formaldehyde flux per ppm of exposure to the
surface of the Level 1 section would correspond to the upper range (greater than approximately
1,750 pmol/mm2-h-ppm) of flux estimates per ppm exposure. Kimbell et al. divided their total flux
(per ppm of exposure) range in the rat into 20 flux bins with the mean flux in bin 14 equal to 1,764
pmol/mm2-h-ppm of exposure (see Table 1, Kimbell et al., 2001). Therefore, we use flux estimates
from flux bins 14-20 of their paper; the surface-area-weighted average flux per ppm of exposure in
these flux intervals is 1,879.66 pmol/mm2-h per ppm (i.e., 1,528.18 pmol/mm2-h per mg/m3) of
exposure. Therefore, average flux in the Level 1 region corresponding to the BMCLio of 0.448
mg/m3 is estimated to be 1,528.18 x 0.448 ~ 685 pmol/mm2-hr.
In order to extrapolate the above BMCL to the human, one is interested in knowing the
human exposure concentration at which some region in the human nose (see Figure B-7) is exposed
to a formaldehyde flux of 685 pmol/mm2-hr. This is estimated from Table 3 in Kimbell et al.
(2001), which tabulates formaldehyde flux to the human nasal lining at various inspiratory rates.
At any given exposure, the anterior regions of the nose are subject to the highest concentrations of
formaldehyde; therefore, we averaged the data from flux bins 17-20 in their tabulation, which
receive the highest levels of flux. The average flux per ppm of exposure concentration in bins 17-20
in the human is 1741 pmol/mm2-h per ppm of exposure. Thus, the exposure concentration at
which these regions would receive a flux of 685 pmol/mm2-hr is 0.484 mg/m3. This is the human
BMCL corresponding to 0.10 extra risk, which was selected because the observed squamous
This document is a draft for review purposes only and does not constitute Agency policy.
B-18 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 metaplasia was determined to be of minimal-to-mild adversity. This is further adjusted in Tables
2 2.1-4 and 2.1-6 for continuous exposure, (6/24) x (5/7).
3 As shown in Table 2.1-5, squamous metaplasia occurred in several sagittal cross sections
4 (Level 1-5, depicted in Figure B-6) of the F344 rat nose in the Kerns et al. (1983) study. However,
5 accurate estimates of formaldehyde flux over the nasal lining other than Level 1 were not available
6 to EPA, and flux estimates provided in Kimbell et al. (2001) cannot be reliably used for the other
7 cross-sections because of a lack of correspondence with the nasal regions in their paper. Therefore,
8 only the squamous metaplasia data reported for Level 1 was carried forward in calculating a
9 candidate RfC. Details of benchmark dose modeling for data on squamous metaplasia in F344 rat
10 and squamous metaplasia and basal hyperplasia in Wistar rat.
Table B-9. Benchmark dose modeling of rat respiratory histopathological
effects
Model
BMR
AIC
BMD
BMDL
Model
fit
Best
model
Notes
Squamous metaplasia in F344 rat (Level 1)
Mstage
k=2
0.10
97.779
0.351
0.281
Fig. 3
Log-
logistic
0.10
97.322
0.492
0.119
Fig. 3
BMD/BMDL > 4
Log-
Probit
0.10
95.619
0.576
0.448
Fig. 4
V
Lowest AIC
Basal hyperplasia in Wistar rat (anterior, Levels 1 & 2)
Mstage
k=2
0.10
65.842
1.767
1.109
Mstage
k=l
0.10
63.846
1.676
1.108
Fig. 7
V
Lowest AIC
Log-
logistic
0.10
65.975
1.633
0.711
Squamous metaplasia in Wistar rat (anterior, Levels 1 & 2)
Log-
logistic
0.10
71.810
1.003
0.526
Fig. 8
V
Lowest AIC
Mstage
k=2
0.10
72.157
0.917
0.376
Fig. 8
This document is a draft for review purposes only and does not constitute Agency policy.
B-19 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
dose
09:57 05/23 2013
Figure B-7. Lateral view of contour plot of formaldehyde flux to the rat (on the
top) and human nasal lining (on the bottom) using CFD modeling (Kimbell et
al. 2001) (nostril is to the right). The actual surface is three-dimensional. Flux at
a site is linear with exposure concentration and is shown here in terms of per ppm.
Therefore, values shown here need to be multiplied by exposure concentration.
Rectangular boxes on the rat mesh roughly estimate location of section Levels 1 & 2
in Kerns et al. (corresponding to Figure B-6).
F344 Rat
Figure B-8. Midsaggital section of rat nose showing section levels (Kerns et al.
1983) (nostril is to the left}.
This document is a draft for review purposes only and does not constitute Agency policy,
B-20 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
dose
10:03 05/23 2013
Figure B-9. Multistage (top panel) and log-logistic (bottom panel) model fit for
Level 1 squamous metaplasia.
dose
11:09 05/23 2013
Figure B-10. Log-probit model fit for Level 1 squamous metaplasia.
This document is a draft for review purposes only and does not constitute Agency policy.
B-21 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
dose
11:59 05/23 2013
Figure B-ll. Basal hyperplasia in Wistar rat (Woutersen et al.,1989):
multistage model (fc=l) fit.
This document is a draft for review purposes only and does not constitute Agency policy.
B-22 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
12:06 05/23 2013
12:09 05/23 2013
Figure B-12. Squamous metaplasia in Wistar rat (Woutersen etal., 1989): log-
logistic (top panel) and multistage (bottom panel) model fit
1 Reproductive Toxicity in Males
2 Two studies reporting effects on the male reproductive system in rats were considered to
3 be of sufficient quality for candidate reference value derivation (Ozen et al., 2002, 2005). For each
4 endpoint, the BMDL estimate (95% lower confidence limit on the BMD, as estimated by the
5 profile-likelihood method) and AIC value were used to select a best-fit model from among the
6 models exhibiting adequate fit If the BMDL estimates were "sufficiently close," that is, differed by
7 at most xx-fold, the model selected was the one that yielded the lowest AIC value. If the BMDL
8 estimates were not sufficiently close, the lowest BMDL was selected as the POD.
This document is a draft for review purposes only and does not constitute Agency policy.
B-23 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-10. Endpoints selected for dose-response modeling for reproductive
and developmental toxicity in animals
Species (strain)/Sex Endpoint
Concentrations and Effect Data
Ozen et al. (2005), Table 1
Rat (Wistar)/adult males,
13-week exposure
Concentration
(mg/m3)a
0
1.462
2.924
Serum testosterone (ng/L)
No. of animals
Mean ± SD
6
4-06.5 ± 41.20
6
244.0 ± 58.44
6
141.3 ± 20.97
Ozen et al. (2002), Table 2
Rat (Wistar)/adult males,
13-week exposure
Concentration
(mg/m3)b
0
2.905
5.810
Testis weight as percent of body
weight
No. of animals
Mean ± SD
7
0.91 ±0.01
7
0.84 ± 0.03
7
0.82± 0.03
Ozen et al. (2002), Table 2
Rat (Wistar)/adult males,
4-week exposure
Concentration
(mg/m3)a
0
2.905
5.810
Testis weight as percent of body
weight
No. of animals
Mean ± SD
7
0.94 ± 0.03
7
0.92 ±0.02
7
0.91± 0.01
a Reported as 0, 5, and 10 ppm. Conversion: ppm*(30.02598/24.45)*(8 hours/24 hours)*(5 days/7days)
b Reported as 0,12.2, and 24.4 mg/m3. Conversion: (mg/m3)*(8 hours/24 hours)*(5 days/7days)
1 Modeling Results
2 Below are tables summarizing the modeling results for the noncancer endpoints modeled.
3 The following parameter restrictions were applied, unless otherwise noted:
4 • Dichotomous models: For the log-logistic and dichotomous Hill models, restrict slope > 1;
5 for the gamma and Weibull models, restrict power > 1; for the multistage models, restrict
6 betas > 0.
7 • Continuous models: For the polynomial models, restrict the coefficients bl and higher to be
8 nonnegative or nonpositive if the direction of the adverse effect is upward or downward,
9 respectively; for the Hill, power and exponential models restrict power > 1.
10 Serum testosterone fOzen et al.. 20051
11 For the BMD modeling of serum testosterone in male Wistar rats exposed to formaldehyde
12 by inhalation for 13 weeks (Ozen et al., 2005), model fit to the mean responses was good. Fit of the
13 models for variance was marginal because the reported sample estimates of standard deviations
14 (SD) did not change monotonically with concentrations. Nevertheless, it is reasonable to accept the
15 best fitting model because the estimated SD of 41.7 is closer to that reported for the control (41.2),
16 meaning that the 1-SD BMR is estimated reasonably well. As both the means and the control SD are
17 well estimated, the BMD is also estimated reasonably well.
This document is a draft for review purposes only and does not constitute Agency policy.
B-24 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-ll. Summary of BMD modeling results for serum testosterone in male
Wistar rats exposed to formaldehyde by inhalation for 13 weeks (Ozen et al.,
2005); BMR = 1 SD change from the control mean
Model3
Goodness of fit
BMDisd
(mg/kg-d)
BMDLisd
(mg/kg-d)
Basis for model selection
p-value
AIC
Exponential (M2)a
0.84
156.2
0.284
0.208
Exponential Models 2 and 4 provided
the best fit with identical AIC to 4
decimals (156.1811).
Fit of Variance Models (Test 3) was
marginal at p = 0.065 with constant
variance and did not improve when
variance was modeled as a power of
means (P=0.050).
Exponential (M3)
NAC
158.1
0.314
0.209
Exponential (M4)b
0.84
156.2
0.284
0.189
Exponential (M5)c
NA
Hill0
NA
Polynomial l°d
Polynomial 2°
Power
0.14
158.3
0.460
0.348
aConstant variance models are presented (BMDS Test 3 p-value = 0.065), with the selected model in bold. Scaled
residuals for selected model for concentrations 0,1.462, and 2.924 mg/m3 were -0.046, 0.15, and -0.13,
respectively.
bFor exponential model M4, the estimate of d, 1.0498, was close to a boundary (1) and parameter estimates were
close to those for M2. The lower BMDL is a result of having one more free parameter (d) than M2.
c These models could not be fitted (more parameters than dose groups).
dFor the power model, the power parameter estimate was 1 (boundary of parameter space). For the Polynomial 2
model, the b2 coefficient estimate was 0 (boundary of parameter space). Consequently, the models in this row
reduced to the Polynomial 1° model.
Exponential Model 2 with 0.95 Confidence Level
dose
14:46 01/16 2013
Figure B-13. Plot of mean response (serum testosterone, Ozen et al., 2005) by
concentration, with the fitted curve for Exponential Model 2 with constant
variance. BMR = 1 SD change from the control mean. Concentrations are in mg/m3.
This document is a draft for review purposes only and does not constitute Agency policy.
B-25 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Relative Testis Weight at 4 weeks (Ozen et al. 2002)
2 Models were fitted successfully to data for the 4-week exposure duration. Fit of the models
3 for variance was marginal (P=0.026 with constant variance, P=0.047 with modeled variance). It
4 may be reasonable to accept the best fitting model, because the estimated SDs and means are fairly
5 close to the observed values. The customary BMR for body and organ weights is "10% relative
6 deviation," (i.e., a 10% difference from the control mean). However, the change in means across the
7 experimental doses was much less than 10% so the BMDs for 10% relative deviation (16-17 mg/kg-
8 g) fall well above the highest dose (5.8 mg/kg-g), leading to unacceptable extrapolation. The table
9 below reports only the BMDs for the 1-SD BMR.
Table B-12. Summary of BMD modeling results for relative testis weight in
male Wistar rats exposed to formaldehyde by inhalation for 4 weeks (Ozen et
al., 2002); BMR = 1-SD change from the control mean
Model3
Goodness of fit
BMD1SD
(mg/kg-d)
BMDL1SD
(mg/kg-d)
Basis for model selection
p-value
AIC
Exponential (M2)a
NA
-138.2
3.81
2.60
The Polynomial 1° model fits the means
adequately, but the fit of the variance
model is marginal at P-0.047.
Exponential (M3)
NA
-126.4
1,944
1.87
Exponential (M4)b
NA
-126.4
NA
NA
Exponential (M5)c
NAc
NA
NA
NA
Hillc
NA
NA
NA
NA
Polynomial ld
Polynomial 2°
0.529
-138.2
3.841
2.636
Power d
<0.0001
-140.2
3.841
2.636
a Variances were modeled as a power of the means (BMDS Test 3 p-value = 0.047), with the selected model in bold.
Note that the power coefficient in the variance model was 18, which is a boundary artificially imposed by BMDS.
Scaled residuals for selected model for concentrations 0, 2.905, and 5.81 mg/m3.
bFor exponential model M4, the estimate of d, 1.0498, was close to a boundary (1) and parameter estimates were
close to those for M2. The lower BMDL is a result of having one more free parameter (d) than M2.
c These models could not be fitted (more parameters than dose groups).
dFor the power model, the power parameter estimate was 1 (boundary of parameter space). For the Polynomial 2
model, the b2 coefficient estimate was 0 (boundary of parameter space). Consequently, the models in this row
reduced to the Polynomial 1° model.
This document is a draft for review purposes only and does not constitute Agency policy.
B-26 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
Polynomial Model with 0.95 Confidence Level
dose
15:19 01/16 2013
Figure B-14. Plot of mean response (relative testis weight, Ozen et al., 2002)
by concentration, with the fitted curve for a linear model with modeled
variance. BMR = 1 SD change from the control mean. Concentrations are in mg/m3.
Relative Testis Weight at 13 weeks fOzen etal. 20021
Most BMDS models could not be fitted successfully to data for testis weight as a percentage
of body weight (Ozen et al. 2002) at the 13-week exposure duration because they reduce to linear
models that had large scaled residuals (poor fit). The Exponential Model 4 did achieve an
acceptable fit, but the likelihood ratio goodness-of-fit test had zero degrees of freedom. Therefore,
Exponential Model 4 was selected. The target BMR, 10% relative change from the control mean, fell
outside the range of observed responses: the control mean was 0.91 and the response at the high
concentration was 0.84 (8% below the control mean). The BMD was 9.99 while the highest
concentration was 5.81.
An alternative POD is the LOAEL. EPA calculations indicate that if the data are normally
distributed (unverified, but plausible for relative weights), the response at the first concentration
represents a decrease of 7.7% below control (95% confidence interval 4.6% to 11%), and the
response at the second concentration represents a decrease of 11% (95% confidence interval 7.9%
to 14%). The response at the second concentration is closest to the target BMR for organ weights
(10% decrease), so the second concentration (5.81 mg/m3) would be used as the biologically
relevant POD.
This document is a draft for review purposes only and does not constitute Agency policy.
B-27 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-13. Model predictions for relative testis weight (Ozen et al. 2002)
Model3
Goodness of Fit
BMDisd
(mg/m3)
BMDLisd
(mg/m3)
BMDiord
(mg/m3)
BMDLiord
(mg/m3)
Basis for Model
Selection
p-value
AIC
Exponential (M2)
Exponential (M3)b
0.011
-129.70
0.574
0.326
4.68
3.74
Smallest AIC
Exponential (M4)
N/Ac
-134.46
0.204
5.02 x 10"04d
9.99
3.24
Power
0.00705
-128.90
0.621
0.348
4.70
3.75
Polynomial 2e
Linear
0.00598
-128.90
0.621
0.348
4.70
3.75
aModeled variance case presented (BMDS Test 2 p-value = 0.0183), selected model in bold; scaled residuals for
selected model for concentrations 0, 2.905, and 5.81 mg/m3 were -0.01397, 0.2209, and -0.2285, respectively.
bFor the Exponential (M3) model, the estimate of d was 1 (boundary). The models in this row reduced to the
Exponential (M2) model.
cNo available degrees of freedom to calculate a goodness-of-fit value.
dModel curvature becomes extreme near the origin, resulting in a very small BMDL for the 1-SD BMR. Model 4 is
the only one with curvature; the other models are linear and do not fit as well.
eFor the Polynomial 2° model, the b2 coefficient estimate was 0 (boundary of parameters space). The models in
this row reduced to the Linear model.
16:24 07/09 2013
Figure B-15. Plot of mean response by concentration, with fitted curve for
selected model; concentration shown in mg/m3.
1 BMDS Modeling Output
2 Exponential Model. (Version: 1.9; Date: 01/29/2013)
3 The form of the response function is: Y[dose] = a * [c-(c-l) * exp(-b * dose)]
4 Parameter d is defined d=1; it is, therefore, not estimated (it is estimated for M5).
5 A modeled variance is fit
This document is a draft for review purposes only and does not constitute Agency policy.
B-28 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1
2
3
4
Table B-15. Table of data and estimated values of interest
Dose
N
Obs mean
Est mean
Obs std dev
Est std dev
Scaled resid
0
7
0.91
0.91
0.01
0.009464
-0.01397
2.905
7
0.84
0.8379
0.03
0.02504
0.2209
5.81
7
0.82
0.8227
0.03
0.03108
-0.2285
Table B-16. Likelihoods of interest
Model
Log(likelihood)
# Pa rams
AIC
A1
68.44598
4
-128.892
A2
72.44658
6
-132.8932
A3
72.0827
5
-134.1654
R
54.58803
2
-105.1761
4
72.22982
5
-134.4596
Table B-17. Tests of interest
Test
-2 Log(likelihood ratio)
Test df
p-value
Test 1
35.72
4
<0.0001
Test 2
8.001
2
0.0183
Test 3
0.7278
1
0.3936
Test 6a
-0.2942
0
N/A
Benchmark Dose Computation.
BMR = 10% relative deviation
BMD = 9.99109
BMDL at the 95% confidence level = 3.24373
Table B-14. Parameter estimates
Variable
Estimate
Default initial parameter values
Inalpha
-11.5414
-11.2791
rho
-23.5629
-22.6938
a
0.91005
0.9555
b
0.535554
0.280827
c
0.899523
0.817323
d
1
1
This document is a draft for review purposes only and does not constitute Agency policy.
B-29 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 B.2. DOSE-RESPONSE ANALYSIS FOR CANCER
2 B.2.1. Cancer Estimates from Observational Epidemiology Studies
3 Illustration of Life-table Analysis for NPC Risk in Humans Based on Data in Beane Freeman et
4 al. (2013)
5 A spreadsheet illustrating the calculation for the derivation of the lower 95% bound on the
6 effective concentration associated with a 0.05% extra risk (LECooos) for nasopharyngeal carcinoma
7 (NPC) incidence is presented in Table B-18.
This document is a draft for review purposes only and does not constitute Agency policy.
B-30 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table B-18. Extra risk calculation3 for environmental exposure to 0.0550 ppm formaldehyde (the LECooos for NPC
incidence)b using a log-linear exposure-response model based on the cumulative exposure trend results of Beane
Freeman et al. (2013), as described in Section 2.2.1
A
B
C
D
E
F
G
H
1
J
K
L
M
N
O
P
Interval
number
(0
Age
interval
All-
cause
mortality
(xl05/yr)
NPC
incidence
(xl05/yr)
All
cause
hazard
rate
(h*)
Prob of
surviving
interval
(q)
Prob of
surviving
up to
interval
(S)
NPC
cancer
hazard
rate (h)
Cond
prob of
NPC
incidence
in interval
(Ro)
Exp
duration
mid
interval
(xtime)
Cum
exp mid
interval
(xdose)
Exposed
NPC
hazard
rate
(hx)
Exposed
all
cause
hazard
rate
(h*x)
Exposed
prob of
surviving
interval
(qx)
Exposed
prob of
surviving
up to
interval
(Sx)
Exposed
cond prob
of NPC in
interval
(Rx)
1
<1
623.4
0.02
0.0062
0.9938
1.0000
0.00000
0.000000
0
0.0000
0.0000
0.0062
0.9938
1.0000
0.00000
2
1-4
26.5
0.05
0.0011
0.9989
0.9938
0.00000
0.000002
0
0.0000
0.0000
0.0011
0.9989
0.9938
0.00000
3
5-9
11.5
0.06
0.0006
0.9994
0.9927
0.00000
0.000003
0
0.0000
0.0000
0.0006
0.9994
0.9927
0.00000
4
10-14
14.3
0.11
0.0007
0.9993
0.9922
0.00001
0.000005
0
0.0000
0.0000
0.0007
0.9993
0.9922
0.00001
5
15-19
49.4
0.15
0.0025
0.9975
0.9915
0.00001
0.000007
2.5
0.4182
0.0000
0.0025
0.9975
0.9915
0.00001
6
20-24
86.5
0.17
0.0043
0.9957
0.9890
0.00001
0.000008
7.5
1.2547
0.0000
0.0043
0.9957
0.9890
0.00001
7
25-29
96.0
0.18
0.0048
0.9952
0.9847
0.00001
0.000009
12.5
2.0911
0.0000
0.0048
0.9952
0.9847
0.00001
8
30-34
110.2
0.30
0.0055
0.9945
0.9800
0.00002
0.000015
17.5
2.9276
0.0000
0.0055
0.9945
0.9800
0.00002
9
35-39
138.8
0.54
0.0069
0.9931
0.9746
0.00003
0.000026
22.5
3.7641
0.0000
0.0069
0.9931
0.9746
0.00003
10
40-44
201.1
0.80
0.0101
0.9900
0.9679
0.00004
0.000039
27.5
4.6005
0.0001
0.0101
0.9900
0.9679
0.00005
11
45-49
324.0
1.07
0.0162
0.9839
0.9582
0.00005
0.000051
32.5
5.4370
0.0001
0.0162
0.9839
0.9582
0.00008
12
50-54
491.7
1.48
0.0246
0.9757
0.9428
0.00007
0.000069
37.5
6.2734
0.0001
0.0246
0.9757
0.9428
0.00011
13
55-59
711.7
1.70
0.0356
0.9650
0.9199
0.00009
0.000077
42.5
7.1099
0.0001
0.0356
0.9650
0.9198
0.00013
14
60-64
1,015.8
1.85
0.0508
0.9505
0.8878
0.00009
0.000080
47.5
7.9464
0.0002
0.0509
0.9504
0.8876
0.00014
15
65-69
1,527.6
2.19
0.0764
0.9265
0.8438
0.00011
0.000089
52.5
8.7828
0.0002
0.0765
0.9264
0.8436
0.00017
16
70-74
2,340.9
2.08
0.1170
0.8895
0.7817
0.00010
0.000077
57.5
9.6193
0.0002
0.1172
0.8894
0.7815
0.00016
17
75-59
3,735.4
1.85
0.1868
0.8296
0.6954
0.00009
0.000059
62.5
10.4557
0.0002
0.1869
0.8295
0.6951
0.00013
18
80-84
6,134.1
1.86
0.3067
0.7359
0.5769
0.00009
0.000046
67.5
11.2922
0.0002
0.3068
0.7358
0.5766
0.00011
Ro =
0.000662
Rx =
0.001163
Extra Risk = (Rx-Ro)/(l-Ro) = 0.0005
This document is a draft for review purposes only and does not constitute Agency policy.
B-31 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Column
A:
Column
B:
Column
C:
Column
D:
Column
E:
Column
F:
Column
G:
Column
H:
Column 1
Interval index number (i).
5-year age interval (except <1 and 1-4) up to age 85.
All-cause mortality rate for interval i (x 105/year) (2010 data from NCHS).
NPC incidence rate for interval i (x 105/year) (2000-2010 SEER data).
All-cause hazard rate for interval i (h*i) (= all-cause mortality rate x number of years in age interval).0
Probability of surviving interval i without being diagnosed with NPC (qi) (= exp(-h*i)).
Probability of surviving up to interval i without having been diagnosed with NPC (Si) (SI = 1; Si = Si—1 x qi—1, for i>l).
NPC incidence hazard rate for interval i (hi) (= NPC incidence rate x number of years in interval).
Conditional probability of being diagnosed with NPC in interval i (= (hi/h*i) x Si x (1—qi)), i.e., conditional upon surviving up to interval i without having been
diagnosed with NPC [Ro, the background lifetime probability of being diagnosed with NPC, is the sum of the conditional probabilities across the intervals].
Column J: Exposure duration (in years) at mid-interval (xtime).
Column K: Cumulative exposure mid-interval (xdose) (= exposure level (i.e., 0.0550 ppm) x 365/240 x 20/10 x xtime) [365/240 x 20/10 converts continuous
environmental exposures to corresponding occupational exposures].
Column L: NPC incidence hazard rate in exposed people for interval i (hxi) (= hi x (1 + p x xdose), where p = 0.04311 + (1.645 x 0.01865) = 0.07379 per ppm x
year) [0.04311 per ppm x year is the regression coefficient obtained, along with its SE of 0.01865, from Dr. Beane Freeman (see Section 2.2.1). To estimate the LEC0oos (i.e., the
95% lower bound on the continuous exposure giving an extra risk of 0.05%), the 95% upper bound on the regression coefficient is used (i.e., MLE + 1.645 x SE)].
Column M: All-cause hazard rate in exposed people for interval i (h*xi) (= h*i + (hxi - hi)).
Column N: Probability of surviving interval i without being diagnosed with NPC for exposed people (qxi) (= exp(-h*xi)).
Column O: Probability of surviving up to interval i without having been diagnosed with NPC for exposed people (Sxi) (Sxl = 1; Sxi = Sxi-1 x qxi-1, for i>l).
Column P: Conditional probability of being diagnosed with NPC in interval i for exposed people (= (hxi/h*xi) x Sxi x (1—qxi)) [Rx, the lifetime probability of being
diagnosed with NPC for exposed people = the sum of the conditional probabilities across the intervals].
aUsing the methodology of BEIR IV (1988).
bThe estimated 95% lower bound on the continuous exposure level of formaldehyde that gives a 0.05% extra lifetime risk of NPC.
cFor the cancer incidence calculation, the all-cause hazard rate for interval i should technically be the rate of either dying of any cause or being diagnosed with the specific cancer
during the interval [i.e., (the all-cause mortality rate for the interval + the cancer-specific incidence rate for the interval - the cancer-specific mortality rate for the interval [so
that a cancer case isn't counted twice, i.e., upon diagnosis and upon death]) x number of years in interval]. This adjustment was ignored here because the NPC incidence rates
are small compared to the all-cause mortality rates.
MLE = maximum likelihood estimate; SE = standard error
This document is a draft for review purposes only and does not constitute Agency policy.
B-32 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
B.2.2. Cancer Estimates from Animal Toxicology Studies Using Biologically Based Dose
Response (BBDR) Modeling
Biologically based dose-response models were developed in a series of papers and in a
health assessment report by scientists at the Chemical Industry Institutes of Toxicology (CUT)
(Conolly etal., 2004, 2003, 2000; Conolly, 2002; Kimbell et al., 2001a, b; Overton etal., 2001; CUT,
1999) to interpret the tumor incidence observed in F344 rats in two long-term bioassays (Kerns et
al. 1983; (Monticello etal.. 1996) and extrapolate risk from rats to humans. The CUT modeling and
available data, and alternatives based on their original model were evaluated extensively for the
purpose of this assessment and used in calculating the cancer potency. This section of the appendix
separately addresses the BBDR models developed for the F344 rat and the human, and in each case:
first provides clarifying details regarding the model, then summarizes all the issues evaluated, and
finally provides detailed evaluations of key issues.
BBDR Modeling: Model Structure and Calibration in Conolly et al (2003; 2004)
In Conolly et al. (2003), tumor incidence data in the above long-term bioassays were
modeled by using an approximation of the two-stage clonal growth model (Moolgavkar et al., 1988)
and allowing formaldehyde to have directly mutagenic action. Conolly et al. (2003) combined these
data with historical control data on 7,684 animals obtained from National Toxicology Program
(NTP) bioassays. These models are based on the Moolgavkar, Venzon, and Knudson (MVK)
stochastic two-stage model of cancer (Moolgavkar et al., 1988; Moolgavkar and Knudson, 1981;
Moolgavkar and Venzon, 1979), which accounts for growth of a pool of normal cells, mutation of
normal cells to initiated cells, clonal expansion and death of initiated cells, and mutation of initiated
cells to fully malignant cells. The following notations are used in the rest of this appendix:
• N cell, normal cell
• I cell, initiated cell
• LI, labeling index (number of labeled cells/(number labeled + unlabeled cells))
• ULLI, unit length labeling index (number labeled cells/length of basement membrane)
• N, number of normal cells that are eligible for progression to malignancy
• an, division rate of normal cells (hours-1)
• [In, rate at which an initiated cell is formed by mutation of a normal cell (per cell division of
normal cells)
• ai, division rate of an initiated cell (hours-1)
• Pi, death rate of an initiated cell (hours-1)
This document is a draft for review purposes only and does not constitute Agency policy.
B-33 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
• hi, rate at which a malignant cell is formed by mutation of an initiated cell (per cell division
of initiated cells)
Cell replication rates and DPX concentrations are driven by local dose, which is
formaldehyde flux to each region of nasal tissue expressed as pmol/mm2-hour, and predicted by
computational fluid dynamics (CFD) modeling using anatomically accurate representations of the
nasal passages (see Chapter 3). In the CUT model, cell division and mutation is treated as a function
of local flux. The spatial distribution of formaldehyde over the nasal lining was characterized by
partitioning the nasal surface by formaldehyde flux to the tissue (rate of gas absorbed per unit
surface area of the nasal lining), resulting in 20 "flux bins" (see Figure 5-13, Chapter 5). Each bin is
comprised of elements (not necessarily contiguous) of the nasal surface that receive a particular
interval of formaldehyde flux per ppm of exposure concentration (Kimbell et al., 2001a). The
spatial coordinates of elements comprising a particular flux bin are fixed for all exposure
concentrations, with formaldehyde flux in a bin scaling linearly with exposure concentration (ppm).
The number of cells at risk varies across the bins, as shown in Figure 5-14, Chapter 5.
Inputs to the model: The inputs to the two-stage cancer modeling consisted of results from
other model predictions as well as empirical data. These included: regional uptake of formaldehyde
in the respiratory tract predicted by using CFD modeling in the F344 rat and human (Kimbell et al.,
2001a, b; Overton et al., 2001; Subramaniam et al., 1998), discussed in the toxicokinetics section of
the appendix; concentrations of DPXs predicted by a PBPK model (Conolly et al., 2000) calibrated to
fit the DPX data in F344 rat and rhesus monkey (Casanova et al., 1994,1991) and subsequently
scaled up to humans; and cell division rates for normal cells (aN) inferred from labeling index data
on rats exposed to formaldehyde (Monticello etal., 1996,1991,1990).
Calibration: The rat model in Conolly et al. (2003) involved six unknown statistical
parameters that were estimated by fitting the model to the rat formaldehyde bioassay data shown
in Table 5-24 in Chapter 5 fMonticello etal.. 19961: Kerns et al., 1983) plus historical data from
several thousand control animals from all the rat bioassays conducted by the NTP. These NTP
bioassays were conducted from 1976 through 1999 and included 7,684 animals with an incidence
of 13 SCCs (i.e., 0.17% incidence). The resulting model predicts the probability of a nasal SCC in the
F344 rat as a function of age and exposure to formaldehyde. The fit of the Conolly et al. (2003)
model to the tumor incidence data is shown in Figure xxx of the main document.
Modeling formaldehyde's mutational action: Formaldehyde interacts with DNAto form DPXs.
In Conolly et al. (2003), DPX formation is considered proportional to the intracellular dose of
formaldehyde related to its directly mutagenic action. Casanova et al. (1994,1989) carried out two
studies of DPX measurements in F344 rats. In the first study, rats were exposed to concentrations
of 0.3, 0.7, 2, 6, and 10 ppm for 6 hours and DPX measurements were made over the whole
respiratory mucosa of the rat, while in the second study, the exposure was to 0.7, 2, 6, or 15 ppm
formaldehyde for 3 hours and measurements were made at "high" and "low" tumor sites. Conolly et
This document is a draft for review purposes only and does not constitute Agency policy.
B-34 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
al. (2000) used data from the second study to develop a PBPK model that predicted the time course
of DPX concentrations as a function of regional formaldehyde flux (estimated in the CFD modeling
and expressed as pmol/mm2-hour). In the two-stage clonal expansion model the mutation rate of
normal and initiated cells were defined as the same linear function of DPX concentration as follows:
[In - Hi - M-Nbasai + KMU x DPX (B-12)
The unknown constants |1n basai and KMU were estimated by fitting model predictions to the
tumor bioassay data.
Use of labeling data: Cell replication rates in Conolly et al. (2003) were obtained by pooling
labeling data from two phases of a labeling study in which male F344 rats were exposed to
formaldehyde gas at similar concentrations (0, 0.7, 2.0, 6.0,10.0, or 15.0 ppm). The firstphase
employed injection labeling with a 2-hour pulse labeling time, and animals were exposed to
formaldehyde for early exposure periods of 1, 4, and 9 days and 6 weeks (Monticello et al., 1991).
The second phase used osmotic minipumps for labeling with a 120-hour labeling time to quantify
labeling in animals exposed for 13, 26, 52, and 78 weeks fMonticello etal.. 19961. The combined
pulse and continuous labeling data were expressed as one exposure time-weighted average (TWA)
over all sites for each exposure concentration, an was calculated from these labeling data by using
an approximation from Moolgavkar and Luebeck (1992). A dose-response curve for normal cell
replication rates (i.e., otN as a function of formaldehyde flux) was then calculated as shown in
Figure D-l.
Upward extrapolation of normal cell division rates: The extensive labeling data collected by
Monticello et al. (1996,1991) present an opportunity to use precursor data in assessing cancer risk.
However, these empirical data could be used to determine afflux) only for the lower flux range, 0-
9,340 pmol/mm2-hour (see Subramaniam et al. [2008] for the reasons), as shown by the solid line
in Figure D-l, whereas the highest computed flux at 15.0 ppm exposure was 39,300
pmol/mm2-hour. Therefore, Conolly etal. (2003) introduced an adjustable parameter, amax, that
represented the value of afflux) at the maximum flux of 39,300 pmol/mm2-hour. amax was
estimated by maximizing the likelihood of the two-stage model fit to the tumor incidence data. For
9,340 < flux < 39,300 pmol/mm2-hour, afflux) was determined by linear interpolation from
cxn(9,340) to amax, as shown by the dashed line in Figure D-l.
This document is a draft for review purposes only and does not constitute Agency policy.
B-35 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
0.05
^ 0.04
a
£ 0.03
CO
Dd
c
§ 0.02
>
b
ID 0.01
o
0.00
Flux (pmole/mm2/h)
Figure B-16. Dose response of normal (aN) and initiated (ai) cell division rate
in Conolly et al. (2003).
Note: Empirically derived values of aN (TWA over six sites) from Table 1 in Conolly et al. (2003) and
optimized parameter values from their Table 4 were used. The main panel is for the J-shaped dose
response. Insets show J-shaped and hockey-stick shaped representations at the low end of the flux range.
The long arrow denotes the upper end of the flux range for which the empirical unit-length labeling data
are available for use in the clonal growth model. amax is the value of aN at the maximum
formaldehyde flux delivered at 15 ppm exposure and estimated by optimizing against the tumor incidence
data. oti < aN for flux greater than the value indicated by the small vertical arrow. Conolly et al. (2004,
2003) assumed oti = aN at all flux values.
Source: Subramaniam et al. (2008).
Division and death rates of initiated cells: The pool of cells used for obtaining the LI data in
Monticello et al. (1996,1991) consists of largely normal cells with perhaps increasing numbers of
initiated cells at higher exposure concentrations. Because the division rates of initiated cells in the
nasal epithelium, ai, either background or formaldehyde exposed, could not be inferred from the
available empirical data, Conolly et al. (2003) assumed a two-parameter function to link ai to an
ai = aN xjmultb - multc x max[aN - aN(basai> 0]} (B-13)
where otN = aN(flux), aN(basai) is the estimated average cell division rate in unexposed normal cells,
and multb and multc are unknown parameters estimated by likelihood optimization against the
0.002
0.001
0.000
0.002
0.001
0.000
Empirical aN (from ULLI data)
Estimated aN
Estimated otj
a
max
/
/
/
2000 4000 6000
/
i
10000
20000
30000
40000
This document is a draft for review purposes only and does not constitute Agency policy.
B-36 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
tumor data.29 The value of ctNfbasai) was equal to 3.39 x 10~4 hours-1 as determined by Conolly et al.
(2003) from the raw averaged unit length labeling index data. The ratio a.\\aN decreases with flux
approximately from 1.07 to 0.96 over the flux range used in the modeling (see Figure 6 in
Subramaniam et al. 2008).
Death rates of Initiated cells ((3i) are assumed to equal the division rates of normal cells (aN)
for all formaldehyde flux values, that is
Pi (flux) = an (flux) (B-14)
No biological justification for these assumed relationships was provided by the authors. Conolly et
al. (2003) stated that this formulation for ai and (3r provided the best fit of the model to the tumor
data.
Structure of the CIIT human model: Subsequent to the BBDR model for modeling rat cancer,
Conolly et al. (2004) developed a corresponding model for humans for the purpose of extrapolating
the nasal cancer risk estimated by the rat model to humans. Also, rather than considering only
nasal tumors (as in the rat model), the human model was used to predict the risk of all human
respiratory tumors. The human model is conceptually very similar to the rat model, and is based on
an anatomically realistic representation of the human nasal passages in a single individual and an
idealized representation of the LRT. Local formaldehyde flux to the tissue is estimated by a CFD
model for humans (Subramaniam et al., 1998; Kimbell et al., 2001a; Overton et al., 2001). However,
the model does not incorporate any data on human responses to formaldehyde exposure.
Rates of cell division and cell death are, with a minor modification, assumed to be the same
in humans as in rats. The concentration of formaldehyde-induced DPXs in humans is estimated by
scaling up from values obtained from experiments in the F344 rat and rhesus monkey.
The statistical parameters for the human model are either estimated by fitting the model to
the human background data, assumed to have the same value as obtained in the rat model, or, in
one case, fixed at a value suggested by the epidemiologic literature. The delay, D, is fixed at 3.5
years, based on a fit to the incidence of lung cancer in a cohort of British doctors (Doll and Peto,
1978). The two other parameters in the rat model that affect the background rate of cancer (multb
and (ibasai) are estimated by fitting to U.S. cancer incidence or mortality data. These parameters
affect the baseline values for the human ai, |iN, and |i|. Because amax, multfc, and KMU do not affect
the background cancer rate, they cannot be estimated from the (baseline) U.S. cancer incidence
rates. Therefore, in Conolly et al. (2004, 2003), amax and multfc are assumed to have the same
values in humans as in rats, and the human value for KMU is obtained by assuming that the ratio
KMU:[ibasaiis invariant across species. Thus,
29Multb and multc were equal to 1.072 and 2.583, respectively (J-shaped aN), and 1.070 and 2.515, respectively
(hockey-stick shaped aN).
This document is a draft for review purposes only and does not constitute Agency policy.
B-37 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
KMUllw„, = KMU[a] x ft"""—' (B-15)
f^Nbasal(rat)
BBDR Modeling: Evaluation ofConolly et al. (2003) Modeling of Nasal Cancer in the F344 Rat
and Alternative Implementations
Table -7 in the dose-response section of the main document listed various issues that were
evaluated by EPA pertaining to the BBDR modeling. This section of the appendix provides the
relevant details of that evaluation. Following an overview of all the issues only the following four
major issues are further elaborated: physiologically based pharmacokinetic modeling of DPXs, use
of historical controls, the uncertainty and variability in the dose response for normal cell-
replication rates, and sensitivity of model results to uncertainty in the kinetics of initiated cells.
Summary of Issues Evaluated in the Rat BBDR Modeling
Table E-l summarizes model uncertainties and their impact as evaluated by EPA. The key
uncertainties are discussed in considerably more detail in additional sections in this appendix and
in published manuscripts (Crump etal., 2008; Subramaniam etal., 2008, 2007). The results in
Subramaniam et al. (2007) and Crump et al. (2008) have been debated further in the literature
(Conolly etal., 2009; Crump etal., 2009). Other alternatives to the CUT biological modeling (but
based on that original model) are also further explored and evaluated in this appendix.
Table B-19. Evaluation of assumptions and uncertainties in the CUT model for
nasal tumors in the F344 rat
Assumptions, approach,
and characterization of
input data in model
Rationale for
assumption/ap
proach
EPA evaluation
Further
elaboration
of
evaluation3
1
Steady-state flux estimates are
not affected by airway and
tissue reconfiguration due to
long-term dosing.
Histopathologic
changes not
likely to be rate-
limiting factors
in dosimetry.
1) Thickening of epithelium and
squamous metaplasia occurring at later
times for the higher dose (Kimbell et al.
1997) will reduce tissue flux. Not
incorporated in model.
2) These effects will push regions of
higher flux to more posterior regions of
respiratory tract. Likely to affect
calibration of rat model. Uncertainty not
evaluated quantitatively.
3) Calibration of PBPK model for DPXs was
seen to be highly sensitive to tissue
thickness.
Subramaniam
et al. (2008);
Cohen-Hubal
etal. (1997);
Klein et al.
(2010).
This document is a draft for review purposes only and does not constitute Agency policy.
B-38 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Assumptions, approach,
and characterization of
input data in model
Rationale for
assumption/ap
proach
EPA evaluation
Further
elaboration
of
evaluation3
2
DPX is dose surrogate for
formaldehyde's mutagenic
potential. DPX clearance is
rapid and complete in 18
hours.
Casanova et al.
(1994).
Half-life for DPX clearance in in vitro
experiments on transformed cell lines
was 7 times longer than estimated by
Conolly et al. (2004, 2003) and perhaps 14
times longer with normal
(nontransformed) human cells. Some
DPX accumulation is therefore likely.
However, model calibration and dose
response in rat was insensitive to this
uncertainty. See Section E.3 for effect on
scale-up of model to humans.
Quievryn and
Zhitkovich,
(2000);
Subramaniam
et al. (2007);
Section 3.6.6.
3
3
Formaldehyde's mutagenic
action takes place only while
DPX's are in place.
DNA lesions may remain after DPX repair
and incomplete repair of DPX can lead to
mutations (Barker et al. 2005). There is
some potential for formaldehyde-induced
mutation after DPX clearance. Thus, it is
possible that formaldehyde mutagenicity
may be underrepresented in model.
Could not quantitatively evaluate
uncertainty (no data on clearance of
secondary lesions).
Subramaniam
et al. (2008);
Section 4.3.3.
3
4
Hoogenveen et al. (1999)
solution method, which is valid
only for time-independent
parameters, is accurate
enough.
Errors due to
this assumption
thought to be
significant only
at high
concentration
and not at
human
exposures.
EPA implemented a solution method valid
for time-dependent parameters. Results
did not differ significantly from those
obtained assuming Hoogenveen et
al.(1999) solutions. However, impact was
not evaluated for the case where cell
replication rates vary in time.
Crump et al.
(2005);
Subramaniam
et al. (2007)
5
All observed SCC tumors are
rapidly fatal; none are
incidental tumors.
Death is
expected to
occur typically
within 1-2
weeks of
observed tumor
(personal
communication
with R. Conolly).
1) Overall, assumption does not impact
model calibration or prediction.
2) However, because 57 animals were
observed to have tumors at interim
sacrifice times, EPA implementation
distinguished between incidental and
fatal tumors. Time lag between
observable tumor and time of death was
significant compared to time lag between
first malignant cell and observable tumor.
Subramaniam
et al. (2007)
This document is a draft for review purposes only and does not constitute Agency policy.
B-39 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Assumptions, approach,
and characterization of
input data in model
Rationale for
assumption/ap
proach
EPA evaluation
Further
elaboration
of
evaluation3
6
Historical controls from entire
NTP database were lumped
with concurrent controls in
studies.
Large number of
control animals
(7,684).
Intercurrent
mortality was
not expected to
be substantial.
1) Tumor incidence in "all NTP" 10-fold
higher than in "all inhalation NTP"
controls. Including all NTP controls is
considered inappropriate.
2) Low-dose-response curve is very
sensitive to use of historical controls.
3) Model inference regarding relevance of
formaldehyde's mutagenic potential to its
carcinogenicity varies from "insignificant"
to "highly significant," depending on
controls used. (See Appendix F for impact
on human risk.)
Table E-2;
Subramaniam
etal. (2007);
Sec E.3.1
7a
LI was derived from
experimentally measured ULLI.
Derived from
correlating ULLI
to LI measured
in same
experiment.
Significant variation in number of cells per
unit length of basement membrane.
Spread in ULLI/LI =25%. Impact on risk
not evaluated.
Subramaniam
etal. (2008);
7b
Pulse and continuous labeling
data were combined in
deriving aN from LI.
All continuous LI
values were
normalized by
mean ratio of
pulse to
continuous LI for
controls.
Formula used for deriving aN from LI is
not applicable for pulse labeling data.
Pulse labeling is measure of number of
cells in S-phase, not of their recruitment
rate into S-phase; not enough information
to derive aN from pulse data. Impact on
risk predictions could not be evaluated.
Subramaniam
etal. (2008);
Section E.3.2.
2
7c
To construct dose response for
aN, labeling data were
weighted by exposure time (t)
and averaged over all nasal
sites (TWA). At an exposure
concentration, flux was
averaged over all nasal sites.
Site-to-site
variation in LI
was large and
did not vary
consistently with
flux. No
reasonable
approach was
available for
extrapolating
observed time
variation in
labeling in rats
to humans.
1) TWA assigns low weight to early time LI
values, but aN for early time (t) is very
important to the cancer process. Because
pulse ULLI was used for t < 13 weeks,
impact of these ULLIs on risk could not be
evaluated.
2) Time dependence in aN derived from
continuous ULLI does not significantly
impact model predictions.
3) Site-to-site variation of aN is at least
10-fold and has major impact on model
calibration. Variation in tumor incidence
data across sites is 10-fold.
4) Large differences in number of cells
across nasal sites (see Table E-3), so
averaging over sites is problematic.
5) TWA is also problematic because
histologic changes, thickening of
epithelium and metaplasia occur at later
times for the higher dose and would
affect replication rate.
Figures E-l,
E-2, E-3;
Subramaniam
etal. (2008);
Section E.3.2.
3
This document is a draft for review purposes only and does not constitute Agency policy.
B-40 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Assumptions, approach,
and characterization of
input data in model
Rationale for
assumption/ap
proach
EPA evaluation
Further
elaboration
of
evaluation3
7d
TWA aN(flux) rises above
baseline levels only at
cytolethal dose. Above such
dose, aN(flux) rises sharply due
to regenerative proliferation.
Variability in
aN(flux) is partly
represented by
also considering
hockey-stick
(threshold in
dose) when TWA
indicates
J-shaped
(inhibition of cell
division)
description of
aN(flux).
1) Uncertainty and variability in aN were
quantitatively evaluated to be large. In
addition, there are several qualitative
uncertainties in characterization of
aN(flux) from LI.
2) Several dose-response shapes,
including a monotonic increasing curve
without a threshold, were considered in
order to adequately describe highly
dispersed cell replication data. This has
substantial impact on low dose risk.
Figures E-l,
E-2, E-3, E-4,
E-5;
Subramaniam
et al. (2008);
Section E.3.2
8a
Dose response for ai was
obtained from aN, assuming
ratio (oti:otN) to be a two-
parameter function of flux (see
Figures 5-7, 5-9). Parameters
were estimated by optimizing
model predictions against
tumor incidence data.
(oti:otN) was >1.0
in line with the
notion of 1 cells
possessing a
growth
advantage over
N cells.
Satisfies
Occam's razor
principle
(Conolly et al.,
2009).
1) ai:aN in CUT modeling is <1.0 (growth
disadvantage) for higher flux values and is
>1.0 only at lower end of flux range in
model (see Figure 5-9).
2) Because there are no data to inform ai,
sensitivity of risk estimates to various
functional forms was evaluated. Risk
estimates for the rat were extremely
sensitive to alternate biologically
plausible assumptions for oti(flux) and
varied by many orders of magnitude at <1
ppm, including values lower than baseline
risk. All these models described tumor
incidence data and cell replication and
DPX data equally well.
Figures D-2,
E-5, E-6;
Subramaniam
et al. (2008);
Crump et al.
(2009, 2008);
Section E.3.3
8b
Death rate of 1 cells is assumed
equal to division rate of N cells
i.e., Pi(flux) = aN(flux).
Based on
homeostasis (aN
= pN) and
assumption that
formaldehyde is
equally cytotoxic
to N cells and 1
cells. Satisfies
Occam's razor
principle
(Conolly et al.,
2009).
1) In general, data indicate 1 cells are
more resistant to cytolethality and that
ADH3 clearance capacity is greater in
transformed cells. Therefore, plausibility
of model assumption, that Pi = aN, is
tenuous.
2) Alternate assumption, Pi proportional
to ai, was examined. Risk estimates were
extremely sensitive to assumptions on Pi
(see Figure 5-12).
Subramaniam
et al. (2008);
Crump et al.
(2009, 2008);
Section E.3.3
References stated here are in addition to Conolly et al. (2004, 2003).
Note: Risk estimates discussed in this table are for the F344 rat.
1 Given the scope of issues to examine, the evaluation of the BBDR modeling as presented in
2 Conolly et al. (2003), and in alternative approaches considered by EPA, proceeded in stages. First,
3 the dosimetric models for formaldehyde flux and DPXs were evaluated. Confidence in the CFD
This document is a draft for review purposes only and does not constitute Agency policy.
B-41 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
modeling of formaldehyde flux has been assessed in the toxicokinetic modeling section earlier, and
is not repeated here. The evaluation of PBPK models for predicting DPXs is presented in this
section of the appendix.
Second, the Hoogenveen et al. (1999) solution was replaced by one that is valid for a model
with time-varying parameters (Crump et al., 2005; see first entry in Table E-l), and tumors found at
scheduled sacrifices were assumed to be incidental rather than fatal (see second entry in
Table E-l). Third, weekly averaged solutions for DPX concentration levels were used instead of
hourly varying solutions (predicted by a PBPK model). The log-likelihood values and tumor
probabilities remained essentially unchanged. Upon quantitative evaluation, these factors,
although important from a methodological point of view, were not found to be major determinants
of either calibration or prediction of the model for the F344 rat data (Subramaniam et al., 2007).
EPA evaluation first attempted to reproduce the Conolly et al. (2003) results under similar
conditions and assumptions, including the assumption that tumors were rapidly fatal. Figure 5-12
in Chapter 5 shows the results from Conolly et al. (2003) and the predicted probabilities from
Subramaniam et al. (2007) (source code made available by Dr. Conolly). These are compared with
the best-fitting model and plotted against the Kaplan-Meier (KM) probabilities. Although the
results are largely similar, there are some residual differences, and these are detailed in
Subramaniam etal. (2007).
Following Georgieva et al. (2003), Subramaniam et al. (2007) used the DPX clearance rate
constant obtained from in vitro data instead of the assumption in Conolly et al. (2003) that all DPXs
cleared within 18 hours (Subramaniam et al., 2007). With this revision, weekly average DPX
concentrations were larger than those in Conolly et al. (2003) by essentially a constant ratio equal
to 4.21 (range of 4.12-4.36) when averaged over flux bin and exposure concentrations.
Accordingly, cancer model fits to the rat tumor incidence data using the two sets of DPX
concentrations (everything else remaining the same) provided very similar parameter estimates,
except that the parameter KMUrat in eq D-l (and eq D-4) (see Appendix D) was 4.23 times larger
with the Conolly et al. (2003) DPX concentrations. In other words, the product KMU x DPX
remained substantially unchanged. However, it is important to note that the different clearance
rate does significantly impact the scale-up of the two-stage clonal growth model to the human
because the parameter KMUhuman is not estimated separately but related to KMUrat (see eq D-4).
After making the above modifications, the impact of the other uncertainties in Table E-1
were examined. Of the issues in Table E-l, only three uncertainties had large impacts on the
modeling of the F344 rat data. These uncertainties and the evaluation of the PBPK modeling of DPX
will be discussed in more detail below:
1) evaluation and model selection of PBPK models for DPX,
2) use of historical controls,
This document is a draft for review purposes only and does not constitute Agency policy.
B-42 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Supplemental Information for Formaldehyde—Inhalation
3) uncertainty and variability in characterizing cell replication rates from the labeling data,
and
4) uncertainty in model specification of initiated cell kinetics.
Physiologically based pharmacokinetic models for DPX formation: evaluation and model
selection
The CFD modeling discussed in the toxicokinetics section models the transport of
formaldehyde through the air phase to the tissue lining on the respiratory tract. While the above
calculations involved the specification of boundary conditions that appropriately characterize the
air-tissue interface, the internal dose of formaldehyde and its reaction with tissue constituents was
not explicitly modeled. Several physiologically based pharmacokinetic (PBPK) models have been
developed to describe the disposition of formaldehyde in the tissue accounting for formaldehyde
reaction via saturable and first order pathways that include the formation and, in some models
clearance, of DNA protein cross links (DPX) formed by formaldehyde. These models relied wholly
or partly on various experimental measurements of DPX in the upper respiratory tract of the F344
rat and rhesus monkey and in the lower respiratory tract of the rhesus monkey (Casanova et al.
1989,1991,1994), which were discussed earlier in Section xxx {ADME section}. The
measurements, and subsequently the models that were based upon these data, allowed the use of
formaldehyde-DPX as an internal dosimeter of inhaled formaldehyde, in particular, as a surrogate
for the molecular dose associated with formaldehyde's mutagenic potential. These models are
tabulated below in Table XXXX.
Table B-20. PBPK models for formaldehyde-DPX
Model
Dpx data
Animal
species
Human
extrapolation
model
Compartments and pathways
Includes air-phase
formaldehyde flux?
Casanova et
al. (1991)
Casanova et
al. (1989);
6-hr exp; 0.3,
0.7, 2.0, 6.0,
10 ppm
Casanova et
al. (1991);
6-hr exp; 0.7,
2.0, 6.0 ppm
F344 rat
Rhesus
monkey
No
Single well-stirred compartment. Saturable & 1st
order metabolism, 1st order DPX formation but not
clearance.
No
Heck &
Casanova
(1994)
Casanova et
al. (1994);
0.7, 2, 6, 15
ppm
preexposed +
naive groups
F344 rat
No
Similar to Casanova et al. (1991). Included effects of
preexposure, induction of hyperplasia at cone > 6
ppm.
No
Cohen Hubal
et al. (1997)
Casanova et
al. (1989)
above +
Casanova
F344 rat
No
Casanova (1991) model+air-phase transports- 1st
order DPX clearance. Predicted DPX in a more
localized region based on model calibrated over
whole nose
Yes (Kimbell et al. 1997)
This document is a draft for review purposes only and does not constitute Agency policy.
B-43 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Model
Dpx data
Animal
species
Human
extrapolation
model
Compartments and pathways
Includes air-phase
formaldehyde flux?
(1994); 3-hr
exp; 0.7, 2.0,
6.0, 15 ppm
Conolly et al.
(2000)
Casanova et
al. (1989)
above +
Casanova
(1994); 3-hr
exp, 0.7, 2.0,
6.0,15 ppm
Casanova et
al. (1991);
6-hr exp; 0.7,
2.0, 6.0 ppm
F344 rat
Rhesus
monkey
Yes
Similar to Cohen Hubal et al. (1997). Derived
allometric rule based on rat and rhesus model to
develop human extrapolation model
Yes (Kimbell et al. 2001a)
Georgieva et
al. (2003)
Casanova et
al. (1989)
above +
Casanova
(1994) 3hr
exp, 0.7, 2.0,
6.0,15 ppm
F344 rat
No
Multilayer tissue compartment, epithelia of varying
thickness. Saturable & 1st order metabolism, 1st
order DPX formation & clearance, clearance rate
derived from in vitro data
Yes, Kimbell et al. 2001a
Franks et al.
(2005)
Did not use
data on DPX
or
formaldehyde
levels for
calibration.
Parameter
values from
other models
were used.
Model developed for
humans
Continuous distribution of formaldehyde across
mucous, epithelial & blood perfused submucosal
layers; diffusional transport of formaldehyde through
mucous layer; Saturable & 1st order metabolism, 1st
order DPX formation but not clearance. Model
evaluated systemic transport of formaldehyde.
No
Subramaniam
et al. (2007)
Casanova et
al. (1989)
above +
Casanova
(1994) 3hr
exp, 0.7, 2.0,
6.0,15 ppm.
F344 rat
No
Saturable & 1st order metabolism, 1st order DPX
formation & clearance, clearance rate derived from
in vitro data
Yes, Kimbell et al. 2001a
1 Of these, clearance of DPX by repair processes was not considered in the models by
2 Casanova et al. (1991), Heck and Casanova (1994) and Franks et al. (2005), and only Conolly et al.
3 (2000) extended their animal PBPK model to develop a corresponding model for the human. The
4 Conolly et al. (2000) modeling presents other useful features that may be particularly important in
5 the context of modeling formaldehyde dose response. Their PBPK modeling of DPX kinetics
6 explicitly incorporates regional formaldehyde dosimetry in the nasal lining by using results from
7 CFD modeling of airflow and gas uptake. Furthermore, results from their models were used as
8 input to biologically based cancer dose-response (BBDR) modeling developed by the same authors.
9 Because of these reasons, EPA focused on the Conolly et al. (2000) PBPK effort and evaluated it in
This document is a draft for review purposes only and does not constitute Agency policy.
B-44 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
detail. Based upon the evaluation, the model was modified by Subramaniam et al. (2007) and used
in EPA's dose-response assessment. The Conolly etal. (2000) model is first described below.
In earlier risk assessment efforts by Hernandez et al. (1994) and Casanova et al. (1991), the
average DPX concentration was considered a surrogate tissue dose metric for the area-under-the-
curve (AUC) of the reactive formaldehyde species. Conolly et al. (2003) assigned a more specific
role for DPXs, treating local DPX concentration as a dose surrogate indicative of the intercellular
concentration of formaldehyde leading to formaldehyde-induced mutations. These authors
indicated that it was not known whether DPXs directly induced mutations (Conolly et al., 2003;
Merk and Speit, 1998). The Conolly et al. (2000) model for the disposition of inhaled formaldehyde
gas and DPX in the rat and rhesus nasal lining is relatively simple in terms of model structure
because it consists of a single well-mixed compartment for the nasal lining as follows:
1) Formaldehyde flux to a given region of the nasal lining is provided as input to the modeling
and is obtained in turn as the result of a CFD model. This flux is defined as the amount of
formaldehyde delivered to the nasal lining per unit time per unit area per ppm of
concentration in the air in a direction transverse to the airflow. It is locally defined as a
function of location in the nose and the inspiratory flow rate and is linear with exposure
concentration.
2) The clearance of formaldehyde from the tissue is modeled as follows:
a saturable pathway representing enzymatic metabolism of formaldehyde, which is
primarily by formaldehyde dehydrogenase (involving Michaelis-Menten parameters
Vmax and Km);
a separate first-order pathway, which is assumed to represent the intrinsic reactivity of
formaldehyde with tissue constituents (rate constant kf); and
first-order binding to DNA that leads to DPC formation (rate constant kb).
3) The clearance or repair of DPC is modeled as a first order process (rate constant kioss).
DPC concentrations were estimated from a study by Casanova et al. (1994) in which rats
were exposed 6 hours/day, 5 days/week, plus 4 days for 11 weeks to filtered air (naive) or to 0.7, 2,
6, or 15 ppm (0.9, 2.5, 7.4, or 18 mg/m3) formaldehyde (preexposed). On the 5th day of the 12th
week, the rats were then exposed for 3 hours to 0, 0.7, 2, 6, or 15 ppm 14C-labeled formaldehyde
(with preexposed animals exposed to the same concentration as during the preceding 12 weeks
and 4 days). The animals were sacrificed and DPC concentrations determined at two sites in the
nasal mucosa. Conolly et al. (2000) used these naive rat data to develop a PBPK model that
predicted the time-course of DPC concentrations as a function of formaldehyde flux at these sites.30
30Subramaniam et al. (2007) who also used the same data verified that they were on naive rats; however, Conolly
et al. (2000) state that they used data on preexposed rats.
This document is a draft for review purposes only and does not constitute Agency policy.
B-45 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Casanova et al. (1994) observed that the DPC concentrations measured in the preexposed
animals (exposed for 11.5 weeks) were not significantly higher than those in naive (air-exposed
control) animals in which there was no significant DPC accumulation. This was interpreted to mean
that DPC repair is rapid enough to completely eliminate the DPC formed in a single 6-hour exposure
by the beginning of the next day. Based on this observation, Conolly et al. (2000) assumed a value
of 6.5 x 10"3 minute-1 for kloss, the first-order rate constant for the clearance (repair) of DPCs, such
that the DPCs predicted at the end of a 6-hour exposure to 15 ppm were reduced to exactly the
detection limit for DPCs in 18 hours.
Uncertainties in PBPK Modeling of the Rat and Rhesus DPC Data
The above assumption of rapid DPC repair in Conolly et al. (2000) appears to be
questionable on three grounds. First, in vitro data from three human cell lines indicated a much
slower clearance, with an average kloss of 9.24 x 10-4 minute-1 (Quievryn and Zhitkovich, 2000).
While the in vitro data can be uncertain because these cells were transformed and immortalized, it
appears that DPC repair in normal cells would be even slower. When nontransformed freshly
purified human peripheral lymphocytes were used instead, the half-life for DPC repair was about
50% longer than in the cultured cells (Quievryn and Zhitkovich, 2000).
Second, Subramaniam et al. (2007) reexamined the Casanova etal. f19941 data for their
PBPK modeling and concluded that the experimental results in Casanova et al. (1994) were
consistent with the smaller experimental value of kloss indicated by the Quievryn and Zhitkovich
(2000) data. Subramaniam et al. (2007) found a significantly decreased (« 40%) level of DPCs in
the high tumor regions of preexposed animals relative to naive animals at 6 and 15 ppm. This was
accompanied by a substantial increase in weight of the tissues dissected from those regions
indicating a thickening of the tissues as is to be expected from metaplastic transformation of
normal tissue to the squamous type due to formaldehyde toxicity. However, after testing the
outcome of changing the tissue thickness in the PBPK model for DPCs, it was apparent to these
authors that such a change alone could not account for the dramatic reduction in DPC levels after
preexposure, even with the higher value of kloss used by Conolly etal. (2000). Because Vmax was
found to be very sensitive to tissue thickness (as also noted by others; Conolly et al. 2000,
Georgieva et al. 2003, Klein et al. 2010), Subramaniam et al. (2007) increased the value of Vmax
with exposure (in a tissue region- and dose-specific manner) and found that it was possible to
explain the naive versus preexposed data of Casanova et al. (1994) with the 7-fold lower value of
kloss. This was consistent with the hypothesis of either an induction in the activity of enzymes that
remove formaldehyde (aldehyde- and formaldehyde dehydrogenase) or other changes in the
biochemical properties of highly exposed tissue.
Third, the value for kloss used by Conolly et al. (2000) was not obtained from time course
measurements but inferred indirectly from measurements made at only two time points where
significant changes in the tissue had occurred. On account of these reasons, Subramaniam et al.
(2007) considered the use of the lower value for kloss from in vitro observations to be more
This document is a draft for review purposes only and does not constitute Agency policy.
B-46 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information for Formaldehyde—Inhalation
appropriate. The same lower value of kloss was also used by Georgieva et al. (2003). Consequently,
Subramaniam et al. (2007) reimplemented and reoptimized the Conolly et al. (2000) model with
this modification and obtained a good fit to the acute DPC data. The reimplemented model is used
in this assessment
Both models provide good similar fits to the DPC data gathered from different regions of the
nose immediately after single 3.0-hour and 6.0-hour acute exposures. However, they differed
significantly in predictions for weekly averaged DPC values; generally 4-fold higher in
Subramaniam et al. (2007) We return to discussing the impact of these differences in C.1.5 and
C.1.6. in the context of using the PBPK model predictions in the two-stage clonal expansion models
for extrapolating the respiratory cancer risk from rats to humans.
The standard error in the parameter estimates reported by Conolly et al. (2000) are
provided in Table YYYY. These are compared with estimates obtained by Klein et al. (2010)31 who
used the same model structure and data as Conolly et al. (2000) except that they used the value
deduced from Quievryn and Zhitkovich (2000) for the parameter kloss (corresponding to slower
clearance) and fitted the model simultaneously to both the rat and rhesus monkey data instead of
the sequential fitting in Conolly et al. (2000). As in Conolly et al. (2000) these authors also
optimized their model by fitting to the grouped mean data.
The mean values of the estimated parameters are similar in Conolly et al. (2000) and Klein
et al. (2012) for the rat and comparable, within a factor of 2.5, for the monkey. However the
standard errors reported by the two authors are very different. The standard error for the
Michaelis-Menten parameters Vmax and Km are generally much higher, while that for Kf, the first-
order clearance rate constant for formaldehyde, is substantially lower (35-fold) in Klein et al.
(2012). These authors used four different methods for their error estimation, asymptotic,
parametric and nonparametric, and Bayesian, all giving very similar standard errors; therefore,
only those for the asymptotic method are reported. Klein et al. (2012) found Vmax to be highly
correlated with Km in both species. Km was seen by both authors to be substantially different across
species, a finding that was attributed plausibly to the involvement of more than one enzyme (Klein
et al., 2010; Georgieva et al., 2003). The standard error reported for kf by Conolly et al. (2000) is
unusually large. These statistical inferences are particularly relevant in identifying uncertainties
when scaling up the animal models for developing a formaldehyde-DPC PBPK model for humans
which is discussed in the section that follows.
Table B-21. Parameter estimates for PBPK modeling
Parameter
Estimate
Standard error
Conolly et al
Klein et al (asymptotic)
Conolly et al
Klein et al (asymptotic)
Vmax-rat
1,008
1,091
9.5
81.0
31The purpose of this effort was to demonstrate different methods that can be used for deriving statistical
inferences of results from PBPK models.
This document is a draft for review purposes only and does not constitute Agency policy.
B-47 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information for Formaldehyde—Inhalation
Parameter
Estimate
Standard error
Vmax-monkey
91.0
223
1.5
18.8
Km-rat
70.8
59.2
7.4
13.8
Km-monkey
6.69
12.6
1.3
4.4
kf
1.08
1.64
2.1
0.06
Sensitivity to use of historical controls
Use of historical controls: Conolly et al. (2003) combined the historical controls arising from
the entire NTP database of bioassays. Tumor and survival rates in control groups from different
NTP studies are known to vary due to genetic drift in animals over time and differences in
laboratory procedures, such as diet, housing, and pathological procedures (Haseman, 1995; Rao et
al., 1987). In order to minimize extra variability when historical control data are used, the current
NTP practice is to limit the historical control data, as far as possible, to studies involving the same
route of exposure and to use historical control data from the most recent studies {Peddada, 2006
#26}.
Bickis and Krewski (1989) analyzed 49 NTP long-term rodent cancer bioassays and found a
large difference in determinations of carcinogenicity, depending on the use of historical controls
with concurrent control animals. The historical controls used in the CUT modeling controls came
from different rat colonies and from experiments conducted in different laboratories over a wide
span of years, so it is clearly problematic to assume that background rates in these historical
control animals are the same as those in the concurrent control group. There are considerable
differences among the background tumor rates of SCCs in all NTP controls (13/7,684 = 0.0017),
NTP inhalation controls (1/4,551 = 0.0002), and concurrent controls (0/341 = 0.0). The rate in all
NTP controls is significantly higher than that in NTP inhalation controls (p = 0.01, Fisher's exact
test). Given these differences, the inclusion of any type of historical controls is problematic and is
thought to have limited value if these factors are not controlled for (Haseman, 1995).
Influence of historical controls on model calibration and on human model: To investigate
the effect of including historical controls in the CUT model, the analyses in Subramaniam et al.
(2007) were conducted by using the following sets of data for controls (the fraction of animals with
SCCs is denoted in parentheses): a) only concurrent controls (0/341), b) concurrent controls plus
all the NTP historical control data used by Conolly et al. (2003) (13/8,031), c) concurrent controls
plus data from historical controls obtained from NTP inhalation studies (1/4,949) (NTP, 2005).32
The results of the evaluation are shown in Table E-2. For these analyses, the same normal
cell replication rates and the same relationship (see eq D-2 in Appendix D) between initiated cell
and normal cell replication rates as used in Conolly et al. (2003) were used. In all cases, weekly
32Three animals in the inhalation historical controls were diagnosed with nasal SCC. Of these, two of the tumors
were determined to have originated in tissues other than the nasal cavity upon further review (Dr. Kevin Morgan
and Ms. Betsy Gross Bermudez, personal communication). These two tumors, therefore, were not included on the
advice of Dr. Morgan. See Subramaniam et al. (2007) for more details.
This document is a draft for review purposes only and does not constitute Agency policy.
B-48 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Supplemental Information for Formaldehyde—Inhalation
averaged values of DPC concentrations were used. Model fits to the tumor incidence data were
similar in all cases to that shown in Figure 5-12 (see Subramaniam et al. [2007] for a more complete
discussion). The biggest influence of the control data was seen to be on the estimated basal
mutation rate in rats, /iNbasalfrat), which, in turn, influences the estimated mutation effect in
humans through eq D-4 (see Appendix D). amax was also seen to be a sensitive parameter and is
discussed later. See Subramaniam et al. (2007) for other parameters in the calibration.
Table B-22. Influence of control data in modeling formaldehyde-induced
cancer in the F344 rat
Case
A
D
B
E
C
F
Control animals
(combined with
concurrent
controls)
All NTP
historical3
All NTP
historical3
NTP
inhalation
historical3
NTP
inhalation
historical3
Concurrent
only3
Concurrent
only3
Cell replication
dose response
J shape
Hockey stick
J shape
Hockey stick
J shape
Hockey stick
Log-likelihood
-1,692.65
-1,693.68
-1,493.21
-1,493.35
-1,474.29
-1,474.29
uNbasal
1.87 x 10-6
2.12 x 10-6
7.32 x 10-7
9.32 x 10-7
0.0
0.0
KMU
1.12 x 10-7
0.0
6.84 x 10-7
6.18 x 10-7
1.20 x 10-6
1.20 x 10-6
KMU:nNbasal
0.06
(0.0, 0.40)
0.0
(0.0, 0.25)
0.94
(0.26, 6.20)
0.66
(0.2, 5.20)
CO
(0.42, ~)
CO
(0.41, ~)
amax
0.045
(0.029, 0.045)
0.045
(0.029, 0.045)
0.045
(0.026, 0.045)
0.045
(0.027, 0.045)
0.045
(0.027, 0.045)
0.045
(0.027, 0.045)
aValues in parentheses denote lower and upper 90% confidence bounds.
Source: Adapted from Subramaniam et al. (2007).
The ratio KMU: [iNbasai is of particular interest because extrapolation to human in Conolly et
al. (2004) assumed its invariance as given by eq D-4 (see Appendix D). Now, [iNbasai in the human is
estimated independently by fitting a scaled-up version of the two-stage model to human baseline
rates of tumor incidence. Thus, a decrease in the value of [iNbasai estimated in the rat modeling
increases the formaldehyde-induced mutational effect in the human.
The MLE of KMUrat:|iNbasai(raq is zero in (Conollv et al.. 2003). However, in the various cases
examined in Subramaniam et al. (2007) it takes a range of values from 0 to 0.9 mm3/pmol and
undefined (or infinite, when [iNbasai = 0). The 95% upper confidence bound on this ratio ranges from
0.25-6.2 (these values would be four times larger had the Conolly et al. [2003] DPC concentrations
been used) to infinite. Thus, the extrapolation to human risk by using the approach in Conolly et al.
(2004) becomes particularly problematic when only concurrent controls are used, because then the
mutational contribution to formaldehyde-induced risk in humans becomes unbounded. This issue
will be discussed again toward the end of the discussion on historical controls.
It may be noted, however, that absence of tumors in the limited number of concurrent
animals does not imply that the calculation will necessarily predict a zero background probability
This document is a draft for review purposes only and does not constitute Agency policy.
B-49 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
of tumor (i.e., a parameter estimate of |iNhasai = 0). Subramaniam et al. (2007) observed such a
counterexample estimate for |iNhasai in simulations involving the alternate dose-response curves for
otN and ai that are discussed in Section E.3.4. Nonetheless, when [iNbasai = 0, an upper bound for
UNbasai using the concurrent controls could be inferred. Accordingly, the 90% statistical lower
confidence bound on the ratio KMU:[iNbasaiis also reported in Table E-2. Such a value would of
course provide a lower bound on risk by using this model and, therefore, would not be
conservative.
Conolly et al. (2003) estimated KMU to be zero for both their hockey-stick and J-shaped
dose-response models for cell replication. However, the estimate for the coefficient KMU [obtained
using the solution of Crump et al. (2010)] is zero only for the case of the model with the hockey-
stick curve for cell replication and with control data as used by Conolly et al. (2003). It is positive in
all other cases and statistically significantly so in all cases in which either NTP inhalation control
data or concurrent controls were used. With concurrent controls only and the J-shaped cell
replication model, the MLE estimate for KMU (1.2 x 10-6) is larger than the statistical upper bound
obtained by Conolly et al. (2003) (8.2 x 10-7). The estimate would be about 4.2 times larger had the
Conolly et al. (2003) DPC model been used.
Influence of historical controls on dose-response curve: Subramaniam etal. (2007) showed
that inclusion of historical controls had a strong impact on the tumor probability curve below the
range of exposures over which tumors were observed in the formaldehyde bioassays. As shown
there, the MLE probabilities for occurrence of a fatal tumor at exposure concentrations below 6
ppm were roughly an order of magnitude higher when all the NTP historical controls were used,
compared with MLE probabilities predicted when historical controls were drawn only from
inhalation bioassays, and many orders of magnitude higher than MLE probabilities predicted when
only concurrent controls were used in the analysis. (Note that this comparison should not be
inferred to apply to upper bound risk estimates because there were many fewer concurrent than
historical controls, so error bounds could be much larger in the case where concurrent controls
were used.)
However, as shown by these authors, model fits to the tumor data in the 6-15 ppm
exposure concentration range were qualitatively indifferent to which of these control data sets was
used. This observation emphasizes the statistical aspect of the CUT modeling—that significant
interplay among the various adjustable parameters allows the model to achieve a good fit to the
tumor incidence data independent of the control data used. On the other hand, the results in
Subramaniam etal. (2007) show that changes in the control data affect parameter KMU, resulting in
significantly different tumor predictions at lower exposure concentrations. Therefore, the strong
influence of using all the NTP historical controls on the low-dose region of the time-to-tumor curves
presented in Subramaniam et al. (2007) suggests that large uncertainties may arise in extrapolating
to both human and rat (in the low-dose region) from such considerations alone.
This document is a draft for review purposes only and does not constitute Agency policy.
B-50 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
A crucial point needs to be noted with regard to the use of inhalation NTP historical controls
(i.e., cases B and E) in the two-stage clonal growth modeling. The single relevant tumor in the NTP
inhalation studies came from the very first NTP inhalation study, dated 1976, and the animals in
this study were from Hazelton Laboratories, whereas the concurrent animals were all from Charles
River Laboratories. Similar problems arise with inclusion of several other NTP inhalation studies.
As mentioned before, genetic and other time-related variation can lead to different tumor and
survival rates, and in general it is recommended that use of historical controls be restricted to the
same kind of bioassays and to studies within a 5-7 year span of the concurrent animals (Peddada et
al., 2007). Thus, it is problematic to assume that the tumor in the 1976 NTP study is representative
of the risk of SCCs in the formaldehyde bioassays. Even if it were appropriate to consider the 1976
study, this leads to the unstable situation in which, despite all of the "upstream" mechanistic
information used to construct the BBDR model, the only piece of data that might keep the model
predictions of human risk bounded is a single tumor found among several thousand rats from NTP
bioassays (Crump et al., 2008). In summary, although it can be argued that the rate of SCCs among
the controls in the rat bioassay is probably not zero, it is also problematic to assume that this rate
can be adequately represented by the background rate in NTP historical controls or even in NTP
inhalation historical controls.
Effect of historical controls on MOA inferences: Subramaniam etal. (2007) also examined
the contribution of the DPC component (which represents the directly mutagenic potential of
formaldehyde in the model) to the calculated tumor probability, choosing for their case study the
optimized models that use the NTP inhalation control data. In the range of exposures where
tumors were observed (6.0-15.0 ppm), the DPC term was found to be responsible for 58-74% of
the added tumor probability. Below 6.0 ppm the estimated DPC contribution was extremely
sensitive to whether the hockey-stick shape or J-shaped was used to characterize the dose response
for cell replication, and varied between 2% and 80%.
The CUT BBDR cancer modeling has contributed to the weight-of-evidence process in
various formaldehyde risk assessment efforts and papers by lending weight to the argument that
the direct mutations induced by formaldehyde are relatively irrelevant compared to the importance
of cytotoxicity-induced cell proliferation in explaining the observed tumorigenicity in rodent
bioassays and in projecting those observations to human exposures {Bogdanffy, 1999 #34;, 2001
#35;Slikker, 2004 #39;Conolly, 2004 #11}. The reanalyses in Subramaniam etal. (2007) (in
particular, the results in the above paragraph) indicate that, if the CUT mathematical modeling
were used to inform this debate, it would in fact indicate the contrary—that a large contribution
from formaldehyde's mutagenic potential may be needed to explain formaldehyde carcinogenicity.
This discussion is resumed in the context of uncertainties in model specification for initiated cells.
This document is a draft for review purposes only and does not constitute Agency policy.
B-51 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Characterization of uncertainty-variability in cell replication rates
Dose-response for normal cell division rate as used in model
Monticello et al. (1996,1991) used unit length labeling index (ULLI) to quantify cell
replication within the respiratory epithelium. ULLI is a ratio between a count of labeled cells and
the corresponding length (in millimeters) of basal membrane examined, whereas the per-cell
labeling index (LI) is the ratio of labeled cells to all epithelial cells, in this case, along some length of
basal membrane and its associated layer of epithelial cells. Monticello et al. (1996,1991) published
ULLI values averaged over replicate animals for each combination of exposure concentration,
exposure time, and nasal site. These values are plotted in Figure E-l.
To use the ULLI data in clonal growth modeling, ULLI needed to be related to LI, and
thereby to cell replication rate (aN) of normal cells. Conolly et al. (2003) adopted the following
procedure in using these values (Subramaniam et al., 2008):
1) The injection labeled ULLI data were first normalized by the ratio of the average minipump
ULLI for controls to the average injection labeled ULLI for controls.
2) Next, these ULLI average values were weighted by the exposure times in Monticello et al.
(1996,1991) and averaged over the nasal sites. Thus, the data were combined into one
TWA for each exposure concentration.
3) LI was linearly related to the measured ULLI by using data from a different experiment
(Monticello et al., 1990) where both quantities had been measured for two sites in the
nose.
4) Cell replication rates of normal cells (an) were then calculated as an = (-0.5/t)log(l - LI)
(Moolgavkar and Luebeck, 1992), where LI is the labeling index and t is the period of
labeling.
5) This was repeated for each exposure concentration of formaldehyde, resulting in one value
of an for each exposure concentration.
6) Correspondingly, for a given exposure concentration, the steady-state formaldehyde flux
into tissue, computed by CFD modeling, was averaged over all nasal sites. Thus, the
afflux) constructed by Conolly et al. (2003) consisted of a single an and a single average
flux for each of six exposures.
This yielded a J-shaped dose-response curve for cell replication (when viewed on a
nontransformed scale for an), as shown in Figure D-l (see Appendix D) for the full range of flux
values used in their modeling. The authors also considered a hockey-stick threshold representation
of their J-shaped curve for aN in order to make a health-protective choice, and the differences
between the two can be seen from the insets in Figure D-l. In these curves, the cell replication rate
is less than or the same as the baseline cell replication rate at low formaldehyde flux values. The
shape of the dose-response curve for cell replication as characterized in Conolly et al. (2003) is seen
as representing regenerative cell proliferation secondary to the cytotoxicity of formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
B-52 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 (Conolly, 2002). Considerable uncertainty and variability, both quantitative and qualitative, exist in
2 the use and interpretation of these labeling data for characterizing a dose response for cell
3 replication rates. The primary issues are discussed here. Unlike the preceding sections, these have
4 largely not been published elsewhere, so more details are provided.
This document is a draft for review purposes only and does not constitute Agency policy.
B-53 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Exposure < 12 wk (pulse labeling)
S ¦ — ¦ --- -m '
ALM
0
5 10 15
^¦y
s'
AMS
Exposure > 12 wk (continuous labeling)
O '
10 15
o
o
o
o
-' •> "—
U)
in
yy ''
\ 'yf
o
-JF
o
-—
in
jr.
r
in
m
in
o
MMT
o
MMT
10
15
10
15
o
o
o
m
/¦ — —
o
m
j-r"-"**1¦
=/'^'
o
y
o
m
.
.
""
m
m
m
o
PLM
o
PLM
10
15
10
15
o
o
o
in
o
o
in
o
in
in
m
in
o
PMS
o
PMS
5 10 15 0 5 10
Formaldehyde exposure concentration, ppm
15
Figure B-17. ULLI data for pulse and continuous labeling studies.
Note: Data are from pulse labeling study, left-hand side, at 1-42 days of exposure and from the
continuous-labeling study, right-hand side, at 13-78 weeks of exposure for five nasal sites ALM, AMS,
This document is a draft for review purposes only and does not constitute Agency policy.
B-54 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
MMT, PLM, and posterior mid septum [PMS]). Within each graph, lines with more breaks correspond to
shorter exposure times. Data source: Monticello et al. (1996,1991).
Time variability in labeling data
Short-time exposure effects on cell replication: Figure E-l shows the site and time variation
in the raw unit-length labeling index (ULLI) data for 1 day to 78 weeks of exposure duration. The
temporal variation in ULLI is quite different between the "early time" (left panel) and "later time"
(right panel) and these early time effects may be quite important to the cancer modeling. At the
earliest times in the left panel, the data show an increased trend in labeling at 2 ppm for the sites
anterior lateral meatus (ALM), anterior medial septum (AMS), posterior lateral meatus (PLM), and
medial maxilloturbinate (MMT) relative to control. Such an increase is generally indicated for low
flux values also for the 13-week exposure time. This can be seen in the dose response plotted as a
function of flux in Figure E-4.
The early times would be important if, say, repeated episodic exposures were considered,
where adequate time has not elapsed for adaptive effects to take place. Such an exposure scenario
may be the norm in the human context. In the CUT cancer modeling, the LI was weighted by
exposure time. As a consequence, the contribution of the early time labeling data is minimized in
their modeling.
Uncertainty due to combining pulse and continuous labeled data: The formula used for
obtaining aN from LI in Conolly et al. (2003) was due to Moolgavkar and Luebeck (1992) who
derived this formula for continuous LI, cautioning that it is not applicable for pulse labeled data.
However, Conolly et al. (2003) applied this formula to the injection (pulse) labeled data also. Such
an application is problematic because 2-hour pulse labeled data represent the pool of cells in
S-phase rather than the rate at which cells are recruited to the pool, and because the baseline values
of an obtained in this manner from both data sets differ considerably. As such, we are not aware of
any reasonable manner to derive cell replication rates from these pulse data without acquisition of
data at additional time points. Because of these problems in incorporating the pulse-labeled data,
further quantitative analysis of cell replication rates is restricted in this document to the continuous
labeled data fMonticello etal.. 19961. which do not include measurements made before 13 weeks of
exposure. It is unfortunate that the continuous labeled data do not include any early
measurements.
Site and time variability in derived cell replication rate
In the remainder of this section, the factors that are considered in order to represent the
uncertainty and variability in the cell replication data when developing alternate dose-response
curves for afflux) will be elaborated.
The ULLI data for individual animals were provided by CUT, which were transformed to LI
values using the linear relationship from step 3 in Section E.3.2.1. For these replicate data, cell
replication rates of normal cells (an) were then calculated as an = (-0.5/t)log(l - LI) as in Step 4.
Figure E-2 (adapted from Subramaniam et al., 2008) shows the variability in aN due to replicated
This document is a draft for review purposes only and does not constitute Agency policy.
B-55 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for Formaldehyde—Inhalation
animals, exposure times, and nasal sites in the continuous labeled data obtained by Monticello et al.
(1996). In this figure, log otN versus site-specific flux are plotted for six sites and four exposure
times for four to six replicate animals in each case. (The mean ULLI over these replicates were
shown in Figure E-l for each site and time as a function of exposure concentration.) It needs to be
noted that these nasal sites differ considerably in the number of cells estimated at these locations as
shown in Table E-3. Each point in Figure E-2 represents data from a single site for a single animal
at a given time. For comparison, the afflux) in Conolly et al. (2003) is also plotted in this figure at
their averaged flux values (filled circles). For flux >9,340 pmol/mm2-hour, Conolly et al. (2003)
extrapolated this empirically derived afflux) by using a scheme discussed in Appendix D
(see Section D.5) on the upward extrapolation of cell replication rate. The curves shown connecting
the filled circles in the figure represent their linear interpolation (long dashes) among the six
points. Their linear extrapolation for flux value >9,340 pmol/mm2-hour is also shown (short
dashes). Note that the linear interpolation and extrapolation are shown transformed to a
logarithmic scale in this plot
As discussed, the raw labeling data plotted in Figure E-l indicates considerable temporal
variability. In Figures E-3, fitted dose-response curves showing logio(cr«) versus flux with
simultaneous confidence limits separately for each time point for two of the largest sites in
Table E-3 (ALM and PLM) are plotted for the continuous labeled data. Note that flux levels are
different at each site. Simple polynomial models in flux (as a continuous predictor), with time
included as a factor (i.e., a class or indicator variable, ii representing the effect of the zth time) were
used as follows:
log(ajv) = a + b x flux + c x flux2 + d x flux3 + ii (B-16)
This document is a draft for review purposes only and does not constitute Agency policy.
B-56 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Figure B-18. Logarithm of normal cell replication rate aN versus
formaldehyde flux (in units of pmol/mm2-hour) for the F344 rat nasal
epithelium.
Note: Values were derived from continuous unit length labeled data obtained by Monticello et al. (1996)
for four to six individual animals at all six nasal sites (legend, sites as denoted in original paper) and four
exposure durations (13, 26, 52, 78 weeks). Each point represents a measurement for one rat, at one nasal
site, and at a given exposure time. Filled red circles: aN(flux) used in Conolly et al. (2003) plotted at their
averaged flux values (see text for details). Long dashed lines: their linear interpolation among points.
Short dashed line: their linear extrapolation for flux value >9,340 pmol/mm2-hour (see Figure D-l for full
range of extrapolation). Linear interpolation/extrapolation is shown with y-axis transformed to
logarithmic scale.
Source: Subramaniam et al. (2008).
Table B-23. Variation in number of cells across nasal sites in the F344 rat
Nasal site
No. of cells
Anterior lateral meatus
976,000
Posterior lateral meatus
508,000
Anterior mid septum
184,000
Posterior mid septum
190,000
Anterior dorsal septum
128,000
Anterior medial maxilloturbinate
104,000
Note: Mean number of cells in each side of the nose of control animals.
Source: Monticello et al. (1996).
This document is a draft for review purposes only and does not constitute Agency policy.
B-57 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
13 weeks 26 weeks
CNI
CJI
_ 1
CO
8 '
o
8
o
CO
p
O) -5J-
O 1
cn
o
¦ * • •''
If)
>o
0 4000 8000 12000 0 4000 8000 12000
Flux Flux
co
8 '
O
U)
O «
52 weeks
4000 8000 12000
Flux
78 weeks
4000 8000 12000
Flux
Figure B-19. Logarithm of normal cell replication rate versus formaldehyde
flux with simultaneous confidence limits for the ALM.
Source: Subramaniam et al. (2008).
This document is a draft for review purposes only and does not constitute Agency policy.
B-58 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information for Formaldehyde—Inhalation
13 weeks
0 2000 6000
Flux
26 weeks
0 2000 6000
Flux
52 weeks
0 2000 6000
Flux
78 weeks
t 1 r~
0 2000 6000
Flux
Figure B-20. Logarithm of normal cell replication rate versus formaldehyde
flux with simultaneous confidence limits for the PLM.
Source: Subramaniam et al. (2008).
The variability considered is that among animals and any measurement error as well as any
other design-related components of error. Simultaneous 95% confidence limits for log(aN) were
produced using Scheffe's method (Snedecor and Cochran, 1980). These 95% confidence limits span
a range of 0.96 in loglO(aN), or nearly a 10-fold range in median an. There is additional dispersion
in these data that does not appear in Figures E-2 and E-3 for an, derived using the mean value of
ULLI:LI; due to variation in the number of cells per mm basement membrane, the ratio of ULLI:LI
had a spread of approximately ±25% (0.45 to 0.71, mean 0.60) among the eight observations
considered in Monticello et al. (1990). Thus:
1) As suggested by Table E-3, and Figures E-2 and E-3, the shape of aN(flux) in Conolly et al.
(2003) is likely to be very sensitive to how an is weighted and averaged over site and time.
2) Averaging of sites could significantly affect model calibration because of substantial
nonlinearity in model dependence on otN at the 10 and 15 ppm doses associated with high
cancer incidence.
This document is a draft for review purposes only and does not constitute Agency policy.
B-59 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
3) Monticello et al. (1996) found a high correlation between tumor rate and the ULLI
weighted by the number of cells at a site. Therefore, considering these factors while
regressing otN against tissue dose would be important in the context of site differences in
tumor response.
4) A further complexity arises because of histologic changes and thickening that occurs in the
nasal epithelium over time in the higher dose groups (Morgan, 1997), factors that are
likely to affect estimates of local formaldehyde flux, uptake, and replication rates
(Subramaniam et al., 2008).
It is clear from Figures E-l and E-3 that the time dependence in cell replication is
significant It would also be useful to examine if this time dependence affects the results of the
time-to-tumor modeling and if early temporal changes in replication rate are important to consider
because of the generally cumulative nature of cancer risk. The time window over which
formaldehyde-induced cancer risk is most influenced is not known, but the time weighting used by
Conolly et al. (2003) assigns a relatively low weight to labeling observed at early times compared
with those observed at later time points. Finally, initiated cells are likely to be replicating at higher
rates than normal cells as evidenced in several studies on premalignant lesions (Coste et al., 1996;
Dragan et al., 1995; Rotstein et al., 1986). Therefore, LI data as an estimator of normal cell
replication rate would be most reliable at early times when the mix of cells sampled include fewer
preneoplastic or neoplastic cells.
The more relevant question, therefore, is whether the afflux) derived in Conolly et al.
(2003) by a TWA over all sites has an effect on low-dose risk estimates. Given the above
uncertainties and variability not characterized in CUT (1999) or in Conolly et al. (2003), it is
important to examine whether additional dose-response curves that fit the cell replication data
reasonably well have an impact on estimated risk. Such sensitivity analyses are carried out in the
sections that follow.
Alternate dose-response curves for cell replication
Clearly, a large number of alternative afflux) can be developed. In conjunction with the
other uncertainties, mainly the use of control data and alternative model structures for initiated cell
kinetics, the number of plausible clonal growth models to be exercised soon require a prohibitively
large investment of time. Therefore, detailed analyses were restricted to a select set of biologically
plausible choices of curves for afflux), which would allow the identification of a range of plausible
risk estimates (MLEs and statistical bounds). This discussion is further informed by recently
published dose-response data for cell replication (Meng et al., 2010), detailed in Section F.2.3.
Six alternative equations for an were developed by regression analysis of the Monticello et
al. (1996) ULLI data. The replicate data corresponding to the summary data presented in this paper
were kindly provided to EPA by CUT for further analyses. In each of these equations, an is
expressed as a function of formaldehyde flux to nasal tissue (pmol/mm2-hour) and, in one equation
(see eq E-ll) that explored time-dependence, the duration of exposure to formaldehyde in weeks.
All the graphs use flux/10,000 for the x-axis, and thej/-axis expresses logio an.
This document is a draft for review purposes only and does not constitute Agency policy.
B-60 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
One source of uncertainty in the cell replication dose response in Conolly et al. (2003) is the
large value of amax (the cell replication rate corresponding to the upper end of the flux range at 15
ppm exposure) in the upward extrapolation from the empirically determined afflux) (see
Figure D-l and surrounding text in Section D.5). The optimal value of amax was found by Conolly et
al. (2003) to be 0.0435 hour-1. As noted by the authors, an argument in support of this value is that
it corresponds to the inverse of the fastest cell cycle times found in the literature. Because the
model treats the induced replication rates as being time invariant, this means that cells in the high-
flux region(s) divide at the highest cell turnover rate ever observed throughout most of an animal's
life. This does not seem to be biologically plausible (Subramaniam et al., 2008).
Our analysis found that a 20% increase or decrease in the estimated value for amax degraded
the fit to the tumor incidence data considerably. Because of the interplay among the parameters
estimated by optimization, this sensitivity of the model to amax indicates that it is necessary to
examine if other plausible values of amax are also indicated by the data and to what extent low dose
estimates of risk are influenced by the uncertainty in its value. The need for such an analysis is also
indicated by Figure E-2. The value of amax (logiootmax = -1.37) in Conolly etal. (2003) is roughly an
order of magnitude greater than the values of afflux) at the highest flux levels in this figure. If the
data pooled over all sites and times are to be used for afflux), then, based solely on the trend in
aN(flux) in Figure E-2, it appears unlikely that afflux) could increase up to this value of amax.
Visually, these empirically derived data collectively suggest that aN versus flux could be leveling off
rather than increasing 10-fold. Therefore, as an alternative to the approach taken in Conolly et al.
(2003) of estimating amax via likelihood optimization against the tumor data, regressions of the
empirical cell replication data in Figure E-2 were used to extrapolate aN(flux) outside the range of
observation (recognizing the uncertainty and model dependence that still results from
extrapolating well outside the range of observed data).
In fitting dose-response curves to the cell replication data, a functional form was used that
was flexible to allow a variety of monotonic and nonmonotonic shapes, with a parameter that
determined the asymptotic behavior of the dose-response function. This allowed the extrapolation
of ajv(flux) to higher flux levels by only relying on the empirical cell replication data. Then, there is
no need for an adjustable parameter to be estimated by fitting to the tumor data. However, the
plausible asymptotes obtained in this manner spanned a large range. In one case below, the
asymptote suggested by the fit to the empirical cell replication data was judged to be abnormally
high. In this case, the aN versus flux curve was followed until the biological maximum of amax (as
given in Conolly et al. [2003]) was reached.
In three of the six regression models below, the data were restricted to the earliest
exposure time (13 weeks) in Monticello etal. (1996) for which the cell proliferation rate (aw) could
be calculated. The interest in using only the 13-week exposure time arises from observations
(Monticello etal., 1996,1991) that at later times there were more frequent and severe histologic
changes, which may have altered formaldehyde uptake and cell proliferation response.
This document is a draft for review purposes only and does not constitute Agency policy.
B-61 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
Consequently, given that the data in Monticello etal. (1991) for times earlier than 13 weeks could
not be used as explained in Section E.3.2.3, the 13-week responses might better represent
proliferation rates for use in a two-stage model of the cancer process than the rest of the Monticello
etal.(1996) data.
Second, the LI data showed considerable variation among nasal sites, which may be related
to the variation in tumor response among sites. Because the cell replication dose-response curves
used in the cancer model represent all of the sites, it was attempted to include this variation by
weighting the regression by the relative cell populations at risk at each of the sites. This was
carried out for some of the models as stated below.
Finally, in one of the regression models, derived from fitting to all of the Monticello et al.
(1996) ULLI data, time-dependence of aN was considered by using weeks of exposure as a
covariate. In this model, time was a regression (continuous) predictor, not a class variable, and its
coefficient represents the change in logio an per week of exposure.
The following regression models for a« versus flux, denoted in the equations below as Nl-
N6 and shown in Figure E-4, as well as the hockey-stick and J-shaped curves used by Conolly et al.
(2003), shown in Figure D-l, Appendix D, were next used as inputs to the clonal growth model for
cancer:
0.0 0.2 0.4 0.6 0.8 1.0
Figure B-21. Various dose-response models of normal cell replication rate;
Nl.
Note: See text for definitions of N1-N6. Nl: Quadratic; monotone increasing in flux, derived from fit to all of the
Monticello et al. (1996) ULLI data.
This document is a draft for review purposes only and does not constitute Agency policy.
B-62 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
log10(alpha) = -2.565 -0.987 * exp{+2.188*X -(2.162*X)A2 }
0.0
0.2
0.4 0.6
weighted mean flux/10,000
0.8
1.0
Figure B-22. Various dose-response models of normal cell replication rate;
N2.
Note: See text for definitions of N1-N6. N2: Linear-quadratic; decreasing in flux for small values of flux,
derived from fit to all of the Monticello etal. (1996) ULLI data.
Time = 13 weeks
0.4 0.6
weighted mean flux/10,000
Figure B-23. Various dose-response models of normal cell replication rate;
N3.
Note: See text for definitions of N1-N6. N3: Linear-quadratic; decreasing in flux for small values of flux,
derived from fit to the 13-week Monticello et al. (1996) ULLI data, using average flux over all sites for a
given ppm exposure and weighting regression by estimates of the numbers of cells at each of five sites.
This document is a draft for review purposes only and does not constitute Agency policy.
B-63 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Time = 13 weeks
flux/10,000
Figure B-24. Various dose-response models of normal cell replication rate;
N4.
Note: See text for definitions of N1-N6. N4: Quadratic; monotone increasing in flux, derived from
unweighted fit to 13-week Monticello et al. (1996) ULLI data.
Time = 13 weeks
0.0 0.2 0.4 0.6 0.8 1.0
weighted mean flux/10,000
Figure B-25. Various dose-response models of normal cell replication rate;
N5.
Note: See text for definitions of N1-N6. N5: Linear-quadratic-cubic; initially increasing slightly with
increasing flux, then decreasing slightly, and finally increasing, derived from fit to 13-week Monticello et
This document is a draft for review purposes only and does not constitute Agency policy.
B-64 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
al. (1996) ULLI data, using average flux over all sites for a given ppm exposure and weighting regression by
estimates of the numbers of cells at each of five sites.
All Sites, ~ Time + 2nd order in Flux
Time = 13
Time = 26
0.0 0.2 0.4 0.6 0.8 1.0 1.2
x
0.0 0.2 0.4 0.6 0.8 1.0 1.2
x
Time = 52
Time = 78
0.0 0.2
0.4
0.6
x
0.8 1.0 1.2
0.0 0.2
0.4
0.6
x
0.8 1.0
1.2
Figure B-26. Various dose-response models of normal cell replication rate;
N6.
Note: See text for definitions of N1-N6. N6: Linear-quadratic-cubic; initially increasing slightly with
increasing flux, then decreasing slightly, and finally increasing, derived from fit to all Monticello et al.
(1996) ULLI data, using weeks of exposure as a covariate. In this model, time was a regression
(continuous) predictor, not a class variable, and its coefficient represents the decrease in logio aN per
week of exposure time.
1 Nl: Quadratic; monotone increasing in flux, derived from fit to all of the Monticello et al. (1996)
2 ULLI data.
3 aN = Exp{-2.015 - 6.513 x Exp[- (6.735xl0"4 x flux)2]}
(B-17)
4 N2: Linear-quadratic; decreasing in flux for small values of flux, derived from fit to all of the
5 Monticello et al. (1996) ULLI data.
This document is a draft for review purposes only and does not constitute Agency policy.
B-65 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 aN = Exp{-5.906 - 2.272 x Exp[2.188xl0"4 x flux - (2.162x10-4 x flux )2]} (B-18)
2 N3: Linear-quadratic; decreasing in flux for small values of flux, derived from fit to the 13-week
3 Monticello et al. (1996) ULLI data, using average flux over all sites for a given ppm exposure and
4 weighting regression by estimates of the numbers of cells at each of five sites.
5 aN = Exp{-5.274 - 2.792 x Exp[1.407xl0"4 x flux - (1.986xl0"4 x flux)2]} (B-19)
6 N4: Quadratic; monotone increasing in flux, derived from unweighted fit to 13-week Monticello et
7 al. (1996) ULLI data.
8 aN = Exp{-3.858-4.809 xExp[-(9.293xl0-5 x flux)2]} (B-20)
9 N5: Linear-quadratic-cubic; initially increasing slightly with increasing flux, then decreasing
10 slightly, and finally increasing, derived from fit to 13-week Monticello et al. (1996) ULLI data, using
11 average flux over all sites for a given ppm exposure and weighting regression by estimates of the
12 numbers of cells at each of five sites.
13 aN = Exp{-5.488-2.755 xExp[-7.808xl0-5x flux + (2.349x10-4 x flux)2 (B-21)
14 - (2.166xl0~4 xflux)3]}
15 N6: Linear-quadratic-cubic; initially increasing slightly with increasing flux, then decreasing
16 slightly, and finally increasing, derived from fit to all Monticello et al. (1996) ULLI data, using weeks
17 of exposure as a covariate. In this model, time was a regression (continuous) predictor, not a class
18 variable, and its coefficient represents the decrease in logio aN per week of exposure time.
19 aN = Exp{7.785xl0-3 x (weeks) - 5.722 - 2.501 x Exp[1.103xl0"4 x flux (B-22)
20 - (7.223x10-s x flux]2 - (1.575x10~4 x flux)3]}
21 Uncertainty in model specification of kinetics of initiated cells
22 Biological implications of assumptions in Conollv et al. (2003)
This document is a draft for review purposes only and does not constitute Agency policy.
B-66 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information for Formaldehyde—Inhalation
The results of a two-stage MVK model are extremely sensitive to the values for initiated cell
division (oti) and death (Pi) rates, particularly in the case of a sharply rising dose-response curve as
observed of formaldehyde. The pool of cells used for obtaining the available LI data (Monticello et
al., 1996,1991) consists of largely normal cells with perhaps increasing numbers of initiated cells at
higher exposure concentrations. As such there is no way of inferring the division rates of initiated
cells in the nasal epithelium, either spontaneous (baseline) or induced by exposure to
formaldehyde, from the available empirical data. Conolly et al. (2003) considered ai(flux) as a
function of afflux) as given by eq D-2 in Appendix D. As shown in Figure D-l (see Appendix D), oti
is estimated in Conolly et al. (2003) to be very similar to aN. That is, with eq D-2 assumed to relate
ai(flux) to afflux), a J- or hockey-shaped dose-response curve for afflux) necessarily results in a J
or hockey shape for ai(flux).
The J shape for the TWA afflux) in Conolly et al. (2003) could plausibly be explained, as
suggested by the examples in Conolly and Lutz (2004), by a mathematical superposition of dose-
response curves describing the effects of the inhibition of cell replication by the formation of DPCs
(Heck and Casanova, 1999) and cytotoxicity-induced regenerative replication (Conolly, 2002).
However, as explained earlier, there is considerable uncertainty and variability, both qualitative
and quantitative, in the interpretation of the LI data and in the derivation of normal cell replication
rates from the ULLI data. While the TWA values of ULLI indicate a J-shaped dose response for some
sites, as also concluded by Gaylor et al. (2004), this is not consistently the case for all exposure
times and sites as discussed earlier. Notwithstanding this uncertainty and variability, and in the
absence of data, the following essential questions have a significant impact on risk predictions and
need resolution if the model structure in eq D-2 is to be used in a biologically based (or motivated)
sense:
• Should mechanisms that might explain a J-shaped dose response for normal cell replication
be expected to prevail also for initiated cells? An identical question can be posed for the
hockey-stick-shaped curve, which indicates a cytotoxicity-driven threshold in dose
response.
• Would the formaldehyde flux at which the cell replication dose-response curve rises above
its baseline be similar in value for both normal and initiated cells as inferred by the CUT
model in Figure D-l?
The next critical assumption in Conolly et al. (2003) was that made for Pi (the death rate of
initiated cells), namely, Pi(flux) = afflux) (see eq D-3). The rationale for this assumption is
explained by assuming formaldehyde to be equally cytotoxic to initiated and normal cells because
the mechanism is presumed to be via its general chemical reactivity (Subramaniam et al. 2008). In
essence, this assumption brings the cytotoxic action of formaldehyde to bear strongly on the
parameterization of the CUT model.
This document is a draft for review purposes only and does not constitute Agency policy.
B-67 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
There are no data to evaluate the strength of these assumptions, so Subramaniam et al.
(2008) studied the plausibility of various inferences that arise as a result of these assumptions.
These inferences are only briefly listed here (see the paper for further discussion).
• For flux <27,975 pmol/mm2-hour, ai > aN (see Figures D-l and D-2 of Appendix D).
Qualitatively, this concept of a growth advantage is in line with data on epithelial and other
tissue types with or without exposure to specific chemicals.
• For higher flux levels, however, the model indicates ai < an (see Figure D-2). There are no
data to shed further light on this inference.
• At these higher flux levels, initiated cells in the model die at a faster rate than they divide,
indicating the extinction of initiated cell clones in regions subject to these flux levels. There
are no data indicating formaldehyde to have this effect
In evaluating these inferences, Subramaniam et al. (2008) point to various data that indicate
that initiated cells represent distinctly different cell populations from that of normal cells with
regard to proliferation response (Ceder et al., 2007; Bull, 2000; Schulte-Hermann et al., 1997; Coste
et al., 1996; Dragan et al., 1995), have excess capacity to clear formaldehyde and, in general, are
considerably more resistant to cytotoxicity, and may already have altered cell cycle control. The
resistance to toxicity is manifested variably as decreased ability of the toxicant to induce cell death
or to inhibit cell proliferation compared to corresponding effects in normal cells. Therefore, the
influence of formaldehyde on apoptosis likely differs between normal and initiated cells.
As concluded in Subramaniam et al. (2008), taken together, there is much data to suggest
that inferring ai < an at cytotoxic formaldehyde flux levels is problematic and that death rates of
initiated cells are likely to be very different from those of normal cells.
In the absence of data to indicate that eq D-2 and eq D-3 (in Appendix D) are biologically
reasonable approaches to link the kinetics of initiated cells with those of normal cells, alternate
model structures other than those represented by these relationships considered by Conolly et al.
(2003) need to be explored, given that the two-stage model is extremely sensitive to ai and Pi. Such
an evaluation needs to primarily explore if the assumptions in eq D-2 and eq D-3 significantly
impact the intended use of the model, namely extrapolation to low-dose human cancer risk and the
calculation of an upper bound on human risk. Any such alternate model structure needs to provide
a good fit to the time-to-tumor data.
Plausible alternative assumptions for al and pi
Therefore, in the additional sensitivity analysis presented here:
a) initiated cell kinetics are considered to be independent of normal cells, and
b) initiated cell replication dose response cannot take a J shape; this is motivated by the
consideration that lower-than-baseline turnover rate represents an increased amount of
This document is a draft for review purposes only and does not constitute Agency policy.
B-68 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
DNA repair taking place, which may not be consistent with impaired DNA repair in initiated
cells.
Thus, two alternatives were considered to eq D-2 for cr/(flux):
II: ai = yi x [1 + exp(y2/y3)]/{l + exp[-(flux- Y2VY3]} (B-23)
12: ai = max[ai(Il), aNBasai] (B-24)
Here yi, J2, and 73 are parameters estimated by fitting the cancer model to the rat bioassay
data. In eq E-12, a\ increases monotonically with flux from a background level of yi asymptotically
up to a maximum value of yix [1 + Expfy-j/Y-O]- The choice of this functional form in eq E-12 and eq
E-13 was considered in order to be parsimonious while at the same time allowing for a flexible
shape to the dose-response curve. The sigmoidal curve allows for the possibility of a slow rise in
the curve at low dose and an asymptote.
Equation E-13 is a modification of eq E-12 that restricts the rate of division of initiated cells
to be at least as large as the spontaneous division rate of unexposed normal cells. There is evidence
to suggest (e.g., in the case of liver foci) that initiated cells have a growth advantage over normal
cells, with or without exposure to specific chemicals (Ceder et al., 2007; Grasl-Kraupp et al., 2000;
Schulte-Hermann et al., 1999; Coste etal., 1996; Dragan etal., 1995).
In addition, in most runs, an upper bound [amgh) is selected for both crw and ai. This value is
assumed to represent the largest biologically plausible rate of cell division. Following Conolly et al.
(2003), in most cases amgh is set equal to 0.045 hours-1. If a value of ai or crw computed using one of
the above formulas exceeded cthtgh, the value of 07,w/, was used in the computation rather than the
value obtained by using the formula.
As noted above, Conolly et al. (2003) set the rate of death for intermediate cells, /?/, equal to
the division rate of normal cells, /?/ =
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information for Formaldehyde—Inhalation
Results of sensitivity analyses on aN, al, and pi
Further constraints
The number of models that might be constructed if all the possibilities listed above for aN,
ah and ft are to be tried in a systematic manner clearly become exponential and daunting.
(Optimally, it would have been desirable to elucidate the role of a specific modification while
keeping others unchanged to determine risk.) Therefore, in order to carry out a viable sensitivity
analysis while at the same time examining the plausible range of risks resulting from variations in
parameters and model structures, various uncertainties were combined in any given simulation. By
using the constraints described above (see eqs E-6 through E-13 and associated text) for ai, Pi, and
an, 19 models were obtained that provided similarly good fits to the time-to-tumor data (which in
some cases contained only five dose groups).
However, for many of these models, the optimal ai(flux) displayed a threshold in flux even
when the model used for afflux) was a monotonic increasing curve without a threshold (i.e., model
N4 for aN in Figure E-4). Indeed, if a thresholded dose-response curve was plausible for oti based on
arguments of cytotoxicity, then a threshold is all the more plausible for an, and such models are
removed from consideration.
Secondly, the basal value of ai was required to be at least as large as the basal value of an.
Another constraint was placed on the baseline initiated cell replication rate. In the absence of
formaldehyde exposure, ai was not allowed to be greater than two or four times an, even if such
models described the tumor data, including the control data, very well. There are some data that
suggest that baseline initiated cells have a small growth advantage over normal cells, so a huge
advantage was thought to be biologically less plausible.
Finally, because most of the SCCs in the rat bioassays occurred in rats exposed to the
highest formaldehyde concentration (15 ppm), the data from this exposure level have a big impact
on the estimated model parameters. In most runs that incorporated the 15 ppm data, the model
appeared, based on inspection of the KM plots, to fit the 15 ppm data quite well but to fit the lower
exposure data less well. Because of the high level of necrosis occurring at 15 ppm, it is possible that
the data at this exposure may not be particularly relevant to modeling the sharp upward rise in the
dose response at 6 ppm. Furthermore, the principal interest is in the predictions of the model at
lower levels to which human populations may be exposed. Consequently, in order to improve the
fit of the model at lower exposures, some of the alternative models were constructed with the 15
ppm data omitted.
Sensitivity of risk estimates for the F344 rat
Figure E-5 contains plots of the MLE of additional risk computed for the F344 rat at
formaldehyde exposures of 0.001, 0.01, 0.1, and 1 ppm for eight models. Two log-log plots are
provided. For those models for which the estimates of additional risk are all positive, the additional
risks are plotted (panel A), and, for those for which estimates of additional risk are negative, the
negatives of additional risks are plotted (panel B). Only five dose groups were considered (i.e., 15
This document is a draft for review purposes only and does not constitute Agency policy.
B-70 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
ppm data omitted) for models 8, 5,15, and 16. Figure E-6 shows the dose-response curves for otN
and ai for these eight cases (panels A and B corresponding to those in Figure E-5). The specification
and estimated values of the parameters for these models are provided in Tables E-4 and E-5. The
primary results are as follows:
1) Among the models considered, negative values for additional risk can arise only in models
in which the dose response for normal cells is J shaped. Thus, all of the models with
negative dose responses for risk have J-shaped dose responses for normal cells. However,
the converse is not necessarily true as may be noted from model 8. This model has both a
positive dose response for risk and a J-shaped dose response for normal cells. In this case,
the strong positive increase in response of initiated cells at low dose was sufficient to
counteract the negative response of normal cells.
2) For doses below which no tumors were observed, the risk estimates predicted by the
different models span a very large range. This result points to large uncertainties in model
specification (how to relate the kinetics of normal and initiated cells) as well as in
parameter values. As mentioned above, the analysis does not attempt to separate the
influence of the different sources of uncertainty, so this range also incorporates the
uncertainty arising from the use of different control data and that due to amax.
This document is a draft for review purposes only and does not constitute Agency policy.
B-71 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1.E-04 -I
1.E-05 -
1.E-06 -
1.E-07 -
1.E-08C]
<>
1.E-09 -
A
1.E-10 -
1.E-11 -
1.E-12 -
1.E-13 -
1.E-14X
1.E-15 —
0.001 0.01 0.1 1
formaldehyde exposure cone (ppm)
Figure B-27. BBDR models for the rat—models with positive added risk.
Note: All four models provide "similar" fits to tumor data (see text)
O
A
X
/\\\ 8 roo1
ide\s
\n
&6Pr0
,N/\d®
s\f0\\ar
f\ts to
wroof
X
data
O model 8
~ model 15
A model 16
model 17
X
As in Conolly et al. (2003), Hockey
stick aN and a, but using concurrent
controls
This document is a draft for review purposes only and does not constitute Agency policy.
B-72 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
0.001
-1.E-08 H
-1.E-07
J*
tn
¦c
"S-1.E-06
T3
T3
<
LU
>
<-1E-05
CD
LU
-1.E-04
-1.E-03
Exposure cone (ppm)
0.01 0.1
A//
e
%
0
-------
Supplemental Information for Formaldehyde—Inhalation
Model 8
0 10000 20000 30000 40000
Flux (pmole/mm2/hr)
0 10000 20000 30000 40000
Flux (pmole/mm2/hr)
All 8 models in A & B provide
similar fits to tumor data
£=
o
m 0.014
«N
---a,
Hockey stick aN and a, with concurrent
controls
Model 16
Model 17
0 10000 20000 30000 40000
Flux (pmole/mm2/hr)
Flux (pmole/mm2/hr)
Figure B-29. Models resulting in positive added rat risk: Dose response for
normal and initiated cell replication.
This document is a draft for review purposes only and does not constitute Agency policy.
B-74 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Model 3
10000 20000 30000 40000
Flux (pmole/mm2/hr)
«N
---a,
Model 4
0 10000 20000 30000 40000
Flux (pmole/mm2/hr)
All 8 models in A & B provide
similar fits to tumor data
J-shape o^, a, in Conolly (2003) with
inhalation NTP + concurrent controls
Model 5
10000 20000 30000 40000
Flux (pmole/mm/hr)
10000 20000 30000
Flux (pmole/mm2/hr)
Figure B-30. Models resulting in negative added rat risk: Dose response for
normal and initiated cell replication.
Table B-24. Parameter specifications and estimates for clonal growth models
of nasal SCC in the F344 rat using alternative characterization of cell
replication and death rates
Parameters
Model 3
Model 4
Model 5
Model 8
Model 15
Model 16
Historical controls
added to concurrent
Inhalation NTP
Inhalation NTP
Inhalation NTP
Inhalation NTP
Inhalation NTP
Inhalation NTP
Number of dose
groups
6
6
5
5
5
5
DPC concentration
Subramaniam
et al. (2007)
Subramaniam
et al. (2007)
Subramaniam
et al. (2007)
Subramaniam
et al. (2007)
Subramaniam
et al. (2007)
Subramaniam
et al. (2007)
aN definition
N3
N6
N3
N6
N4
N4
ot/ definition
12
12
12
11
11
11
Clhigh
--
0.045
-
0.045
0.045
0.045
6i definition
6i = Kg x a,
6i = Kg x a,
6i = Kg x a,
6i = Kg x a,
6i = Kg x a,
6i = Kg x a,
Vl < 4 (XNBasal
Vl <2 (XNBasal
Log-likelihood
-1,495.34
-1,495.61
-184.02
-184.22
-182.75
-186.37
l^NBasal
7.518E-7
1.664E-6
8.684E-7
9.230E-7
1.037E-6
1.662E-7
This document is a draft for review purposes only and does not constitute Agency policy.
B-75 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Parameters
Model 3
Model 4
Model 5
Model 8
Model 15
Model 16
KMU
3.884E-7
3.471E-7
0.0
0.0
(0.0, 2.093E-6)
4.582E-6
(1.8E-6,1.86E-5)
0.0
KMX (KMU/Hmasal)
0.5166
0.2086
0.0
0.0
(0.0, 4.696)
4.420
(1.53,17.67)
0.0
D0§
214.3
199.7
261.8
254.2
423.2
245.1
Dof§
75.26
79.81
119.7
101.1
100.8
98.83
Vi
1.164E-5
1.006E-5
3.168E-5
2.967E-4
6.888E-4
3.441E-4
V2
1427
1,591
1,825
3,223
4,652
2,818
V3
11,944
13,017
14,207
15,989
54,334
37,896
Ke
0.9893
0.9848
0.9804
0.9504
1.006
0.9660
§See Subramaniam et al. (2007) for an explanation of the time delay constants D0and D0f
Table B-25. Parameter specifications and estimates for clonal growth models
of nasal SCC in the F344 rat using cell replication and death rates as
characterized in Conolly etal. (2003)
Parameters
Model 13
Model 17
Historical controls added to
concurrent
All NTP
NO historical controls
Number of dose groups
6
6
DPC concentration
Conolly et al. (2000)
Subramaniam et al. (2007)
aN definition
J shape
(TWA, Conolly et al. 2003)
Hockey
(TWA, Conolly et al., 2003)
ot/ definition
eq. D-l
eq. D-l
Clhigh
--
-
6i definition
6, = aN
6, = aN
Log-likelihood
-1,692.68
-1,474.29
l^NBasal
1.731E-6
0.0
KMU
0.0
1.203E-6
(1.0E-6,1.427E-6)
KMX (KMU:iiNBasai)
0.0
Infinite
(0.4097, infinite)
D0§
239.5
243.13
DoF5
66.31
68.83
multib
1.047
1.078E+0
multic
1.510
3.347
&max
5.153E-2
0.045
§See Subramaniam et al. (2007) for an explanation of the time delay constants D0 and D0f
This document is a draft for review purposes only and does not constitute Agency policy.
B-76 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
1) At the 10 ppb (0.01 ppm) concentration, MLE risks range from -4.0 xlO"6 to +1.3 xlO"7. At
this dose, models that gave only positive risks resulted in a five orders of magnitude risk
range from 1.2 xlO"12 to 1.3 xlO"7, while narrowing to a four orders of magnitude risk
range from 1.2 xlO"10 to 1.3 xlO"6 at the 0.1 ppm level. This narrowing continues as
exposure concentration increases, and the curves coalesce to substantially similar values at
6 ppm and above (not shown). For all these 8 models, the rat added risk at 6.0 ppm ranged
from 1.8 xlO-2 to 2.1 xlO"2.
2) There does not seem to be any systematic effect on additional risk that depends on
whether the 15 ppm data are included in the analysis.
3) For all of the models except Models 13 and 17 in Figures E-5 and E-6, the additional risk
varies substantially linearly with exposure at low exposures between 0.001 and 1.0 ppm
(departing only to a small extent from linearity between 0.1 and 1.0 ppm). Models 13 and
17 show a quadratic dependence; these models employ the TWA J-shaped and hockey stick
dose-response curves for aN used in Conolly et al. (2003) and the same equations used by
those authors to relate ai and Pi to otN (see eqs D-2 and D-3, Section D-6). However, the
control data in Model 17 was different from those used by Conolly et al.; while all NTP
controls were added to the concurrent controls in Model 13, only concurrent controls were
used in Model 17.
The various model choices presented in Figure E-5 all provided equally good fits to the
time-to-tumor data although within the context of a significant qualification. It was not possible to
simply use the maximized log-likelihood values as a means of comparing the goodness of fit to the
tumor incidence data across all these model choices. This is because many of the model choices
differed in the number of doses or in the number of control animals that were used, so the fits were
compared across such models only visually.
Wherever results from the BBDR modeling are discussed, values of added risk, as opposed
to extra risk, are reported. This is purely for convenience in interpretation. Because of the low
background incidence, these values are only negligibly different from the corresponding extra risk
estimate. The final risk (or unit risk) estimates provided in this document are based on extra risk
estimates.
MOA inferences revisited
The ratio KMU:/iNbasai represents the added fractional probability of mutation per cell
generation (/in - MNbasai):MNbasai due to unit concentration of DPCs. As discussed in Sections E.3.1.2
and E.3.1.5 (see Appendix E), this parameter has a critical impact on the extrapolation as well as on
inferring whether the mutagenic action of formaldehyde is relevant to explaining the observed
tumor incidence or its carcinogenicity at lower concentrations. In that prior discussion, this ratio
was found to be extremely sensitive to the choice of historical control data. The analysis indicates
that, for a given set of control data that is used, uncertainties associated with otN and oti also have a
large impact on this ratio.
As discussed in E.3.1.2, this ratio was infinite when concurrent controls were used because
the MLE value for /iNbasai was found to be zero. The use of these concurrent controls, however, does
This document is a draft for review purposes only and does not constitute Agency policy.
B-77 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
not necessarily imply that /iNbasai will be determined to be zero. In one of the scenarios examined in
the sensitivity analysis, where concurrent controls were used along with the combination of dose-
response curves eq D-9 for otN (see Figure E-4) and eq E-13 for ai, the optimal value of the ratio
KMU:/iNbasai was equal to 0.25. For the models in Figure 5-13A, this ratio was 0 for all except model
17 for which it was infinite. For the models in Figure 5-13B with negative added risk, the ratio
ranged from 0-4.5. For some of those models where KMU:/iNbasai was finite, the upper confidence
bound on this ratio was found to increase by an order of magnitude from the MLE value.
Thus, we conclude that the modeling does not help resolve the debate as to the relevance of
formaldehyde's mutagenic potential to its carcinogenicity.
Confidence bounds: model uncertainty versus statistical uncertainty
For Models 15 and 17 in Figures E-5A and E-6A, 90% CIs for additional risk were calculated
by using the profile-likelihood method. Table E-6 compares the lower and upper confidence
bounds for these models for 0.001 ppm, 0.1 ppm (doses well below the range where tumors were
observed), and 6 ppm (the lowest dose where tumors were observed) with the MLE risk estimates
at these doses. In both cases, these intervals were quite narrow compared with the differences in
risk predicted by different models in Figure E-5. This suggests that model uncertainty is of more
consequence in the formaldehyde animal model than is statistical uncertainty. We also estimated
confidence bounds using the bootstrap method for select models, and determined that these
estimates were in agreement with the bounds calculated using the profile-likelihood method.
These results are not presented here. We return to the calculation of confidence limits when
determining points of departure (PODs).
Table B-26. Comparison of statistical confidence bounds on added risk for two
models
Dose (ppm)
Model
Lower bound
MLE
Upper bound
0.001
Model 15
4.4 x 10"9
1.3 x 10"8
1.6 x 10"8
Model 17
1.2 x 10"14
1.2 x 10"14
1.3 x 10"14
0.1
Model 15
4.5 x 10"7
1.3 x 10"6
1.7 x 10^
Model 17
1.2 x 10"10
1.2 x 10"10
1.3 x 10"10
6
Model 15
1.8 x 10"2
2.1 x 10"2
2.3 x 10"2
Model 17
1.3 x 10"2
1.8 x 10"2
3.0 x 10"2
In conclusion, it is demonstrated that the different formaldehyde clonal growth models can
fit the data about equally well and still produce considerable variation in additional risk and
biological inferences at low exposures. However, even with these large variations, the highest MLE
added risk for the F344 rat is only of the order of 10-6 at 0.1 ppm. Thus, with regard to calculating a
reasonable upper bound that includes model and statistical uncertainty, the relevant question is
whether the range arising out of uncertainties in the rat model amplifies when extrapolated to the
human. Thus, in Appendix F, the human model in Conolly et al. (2004) will be examined.
This document is a draft for review purposes only and does not constitute Agency policy.
B-78 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
Statistical Methods Used in Evaluation
Parameters of the alternate models shown here were estimated by maximizing the
likelihood function defined by the data (Cox and Hinkley, 1974). Such estimates are referred to as
maximum likelihood estimates (MLEs). Statistical confidence bounds were computed by using the
profile-likelihood method (Crump, 2002; Cox and Oakes, 1984; Cox and Hinkley, 1974). In this
approach, an asymptotic 100(1 - a)% upper (lower) statistical confidence bound for a parameter, p,
in the animal cancer model is calculated as the largest (smallest) value of /3 that satisfies
2[Lmax - L*(P)] = Xi-2a (B-26)
where L indicates the likelihood of the rat bioassay data, Lmax is its maximum value, L *(p) is, for a
fixed value of p, the maximum value of the log-likelihood with respect to all of the remaining
parameters, and Xi-2U is the 100(l-2cr) percentage point of the chi-square distribution with one
degree of freedom. The required bound for a parameter, p, was determined via a numerical search
for a value of p that satisfies this equation.
The additional risk is defined as the probability of an animal dying from an SCC by the age of
790 days, in the absence of other competing risks of death, while exposed throughout life to a
prescribed constant air concentration of formaldehyde, minus the corresponding probability in an
animal not exposed to formaldehyde. The MLE of additional risk is the additional risk computed
using MLEs of the model parameters.
The method described above for computing profile-likelihood confidence bounds cannot be
used with additional risk because additional risk is not a parameter in the cancer model. Instead,
an asymptotic 100(1 - a)% upper (lower) statistical confidence bound for additional risk was
computed by finding the parameter values that presented the largest (smallest) value of additional
risk, subject to the inequality
2 [Lmax - L] < Xi-2a (B-27)
being satisfied, with the resulting value of additional risk being the required bound. This procedure
was implemented through use of penalty functions (Smith and Coit, 1995). For example, the profile
upper bound on additional risk was computed by maximizing the "penalized added risk," defined as
(additional risk - penalty), where
penalty = W x {[(Lmax - L) - xi-2a/2]+}2 (B-28)
and []+ equals the quantity in the brackets whenever it is positive and zero otherwise. The
multiplicative weight, W, was selected by trial and error so that the final solution satisfied the
following equation sufficiently well.
2 (Lmax — L) — Xi-2a (B-29)
This document is a draft for review purposes only and does not constitute Agency policy.
B-79 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Supplemental Information for Formaldehyde—Inhalation
The computer code was written in Microsoft Excel 2002 SP3 Visual Basic. Either the regular
Excel Solver or the Frontline Systems Premium Solver was used to make the required function
optimizations. Computation of confidence bounds was highly computationally intensive, and,
consequently, confidence bounds were computed only for selected parameters in selected runs.
For select cases, the bootstrap method was also used to calculate confidence bounds in order to
confirm their accuracy. Values so calculated were found to be in agreement with those calculated
by using the likelihood method.
BBDR Modeling: Sensitivity Analysis ofBBDR Model for Formaldehyde-Induced Respiratory
Cancer in Humans
Maior Uncertainties in the Formaldehyde Human BBDR Model
Subsequent to the BBDR model for modeling rat cancer, Conolly et al. (2004) developed a
corresponding model for humans for the purpose of extrapolating the risk estimated by the rat
model to humans. Also, rather than considering only nasal tumors, it is used to predict the risk of
all human respiratory tumors. The human model for formaldehyde carcinogenicity (Conolly et al.,
2004) is conceptually very similar to the rat model and follows the schematic in Figure 5-11 in
Chapter 5. The model structure, notations, and calibration are described in Appendix D. Unlike the
sensitivity analysis of the rat modeling where a number of issues were examined, a much more
restricted analysis will be presented here for the sake of brevity. A more extensive analysis was
carried out initially that carried forward several of the rat models from Appendix E to the human,
and the lessons learned from those exercises are in agreement with the more restricted
presentation that follows. Table F-l lists the major uncertainties and assumptions in the human
extrapolation model in Conolly et al. (2004).
Table B-27. Summary of evaluation of major assumptions and results in CUT
human BBDR model
Assumptionsa
Rationale in Conolly et al.
(2003) or CUT (1999)
EPA evaluation
Further
elaboration
Cell division rates derived
from rat labeling data were
assumed applicable to
human (except for
assuming different fraction
of cells with replicative
potential).
There are no equivalent LI
data for human or guidance
for extrapolating cell division
rate across species.
Enzymatic metabolism plays a
role in mitosis. Therefore, we
expect interspecies difference in
cell division rate. Basal cell
division rates in humans are
expected to be much more
variable than in laboratory
animals.
Subramaniam
et al. (2008)
This document is a draft for review purposes only and does not constitute Agency policy.
B-80 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Assumptionsa
Rationale in Conolly et al.
(2003) or CUT (1999)
EPA evaluation
Further
elaboration
Parameters for enzymatic
metabolism of
formaldehyde in human
PBPK model for DPC
concentrations: Km varies by
order of magnitude
between rat and monkey
but is same for monkey and
human. Vmax:Km is similar
for rat and monkey but 6-
fold lower for human.
See text (Section 3.6.6.2)
See text (Section 3.6.6.2)
Section 3.6.6.2;
Conolly et al.
(2000);
Subramaniam
et al. (2008);
Klein et al.
(2010)
Anatomically realistic
representation of nasal
passages.
Reduces uncertainty (over
default calculation carried
out by averaging dose over
entire nasal surface).
Computer representation
pertains to that of one individual
(white male adult). There is
considerable interindividual
variability in nasal anatomy.
Susceptible individuals are even
more variable.
Kimbell et al.
(2001a, b);
Subramaniam
et al. (2008,
1998)
KMU:nNbasaiis species
invariant (used to estimate
human).
Human cells are more
difficult to transform than
rodent, both spontaneously
and by exposure to
formaldehyde.
Hnbasai is 0 when concurrent
controls or inhalation NTP
controls in time frame of
concurrent bioassays are used.
Leads to infinitely large KMU for
human.
Subramaniam
et al. (2007);
Crump et al.
(2009, 2008)
Conservative assumptions
were made. Results are
conservative in the face of
model uncertainties.
1) Hockey-stick dose
response for aN was included
even though TWA indicated J
shape.
2) Overall respiratory tract
cancer incidence data for
human baseline rates were
used.
3) Risk was evaluated at
statistical upper bound of the
proportionality parameter
relating DPCs to the
probability of mutation.
Results in Conolly et al. (2004) are
not conservative in the face of
model uncertainties: (a) human
risk estimates are very sensitive
to use of historical controls in the
analysis of the animal bioassay,
(b) human risk estimates are
unboundedly large when
concurrent controls are used in
rat model, and (c) minor
perturbations in model
assumptions regarding division
and death rates of initiated cells
lead to upper bound risks that
were more than 1,000-fold
greater than the highest
estimates in Conolly et al. (2004).
Conolly et al.
(2004);
Subramaniam
et al. (2007);
Crump et al.
(2009, 2008)
Assumptions in this table are in addition to those listed for the BBDR model for the F344 rat.
1 Uncertainty in PBPK Model for Prediction of Human DPC Concentrations
2 Conolly et al. (2000) constructed a PBPK model for the rhesus monkey along similar lines as
3 for the F344 rat, and used the rat and rhesus monkey parameter estimates to develop a model for
This document is a draft for review purposes only and does not constitute Agency policy.
B-81 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
human DPC concentrations. In the rhesus monkey model, they maintained the same values of kb,
kioss, and kf as in the rat model but optimized the values of Vmax and Km against the rhesus monkey
data from Casanova et al. (1994). The resulting human PBPK model used formaldehyde flux
estimates predicted by an anatomically realistic CFD modeling of the nasal passages; except for the
anatomic reconstruction, there were no other human data used to inform the PBPK model.
For the human, the model used the value of Km estimated by the rhesus monkey model and
the epithelial thickness averaged over three regions of the rhesus monkey nose. The maximum rate
of metabolism, Vmax, which was estimated independently for the rat and rhesus monkey by fitting
to the DPC data available for these species, was then extrapolated to the human by assuming a
power law scaling with body weight (BW) (i.e., Vmax = a x BWb), and the coefficient "a" and
exponent "b" were derived from the independently estimated values of (Vmax)rat and
(Vmax) monkey- Table C-l gives the values of Vmax and Km in the Conolly et al. (2000) extrapolation.
Table B-28. Extrapolation of parameters for enzymatic metabolism to the
human in Conolly et al. (2000)
Parameter
F344 rat
Rhesus monkey
Human
Vmax (pmol/min-mm3)
1,008.0
91.0
15.7
Km (pmol/mm3)
70.8
6.69
6.69
Source: Conolly et al. (2000).
In general, laws for allometric scaling across species, such as how enzymatic metabolic rates
vary across organisms, are derived as empirical regression relationships based on data from
multiple species and usually multiple sources of data points. For example, West and Brown (2005)
demonstrate that metabolic rates scale with mass3/4 using data from organisms ranging over 2 7
orders of magnitude in mass (intracellular up to the largest organisms). In Conolly et al. (2000), the
power-law relationship is derived using two data points (F344 rat and rhesus monkey for a single
chemical) with log BW as x-axis and Vmax ony-axis. Because such a regression does not have the
power to delineate the curvature in the scaling function, the empirical strength of the allometric
relationship derived in Conolly et al. (2000) is extremely weak for use in extrapolating from the rat
to the human on the basis of body-weight. Furthermore, as noted earlier, Vmax is highly correlated
to Km, the value of Km appears to vary substantially between the rat and monkey, and as indicated
by the standard error in Klein etal. (2011). its estimation is fairly uncertain. These observations
make the scaling relationship in Conolly et al. (2000) more problematic.
The following observations point to the uncertainty in the values of the parameters Vmax
and Km in the Conolly et al. (2000) models for predicting DPCs. First, Km varies by an order of
magnitude across the rat and monkey models but is then considered invariant between the monkey
and human models (Conolly et al., 2000). Second, the values in Conolly et al. (2000) for Vmax/Km,
the low-dose limit of the rate of enzymatic metabolism, is roughly similar between the rat and
monkey but lower by a factor of six in the human.
This document is a draft for review purposes only and does not constitute Agency policy.
B-82 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
Another factor that can substantially influence the above extrapolation of DPCs in the
human is that Conolly et al. (2000) assumed the tissue to be a well-mixed compartment with regard
to formaldehyde interaction with DNA and used the amount of formaldehyde bound to DNA per
unit volume of tissue as the DPC dose metric. Considering formaldehyde's highly reactive nature,
the concentrations of formaldehyde and DPC are likely to have a sharp gradient with distance into
the nasal mucosa (Georgieva et al., 2003). Given the interspecies differences in tissue thickness,
there is uncertainty as to whether DPC per unit volume or DPC per unit area of nasal lining is the
more appropriate dose metric to be used in the extrapolation. In particular, it may be assumed that
the cells at risk for tumor formation are only those in the epithelium and that measured DPC data
(in monkeys and rats) are an average over the entire tissue thickness. Because the epithelial DPCs
in monkeys (and presumably humans) would then be more greatly "diluted" by lower levels of DPC
formation that occur deeper into the tissue than in rats, it could be predicted that the ratio of
epithelial to measured DPCs in monkeys and humans would be much higher than the ratio in rats.
On the whole, these observations suggest that human extrapolations of DPC concentrations
from the rat or monkey using the human PBPK model in Conolly et al. (2000) may be highly
uncertain.
Sensitivity Analysis of Human BBDR Modeling
Crump et al. (2008) carried out a limited sensitivity analysis of the Conolly et al. (2004)
human model. This analysis was limited to evaluating the effect on the human model of the
following. These evaluations have been the subject of some debate in the literature and at various
conferences (Conolly, 2009; Conolly et al., 2009, 2008; Crump etal. 2009).
1) The use of the alternative sets of control data for the rat bioassay data that were
considered in the sensitivity analysis of the rat model in Appendix E.
2) Minor perturbations in model assumptions regarding the effect of formaldehyde on the
division and death rates of initiated cells (ai, Pi).
As mentioned in Section D.7. one (of the two) adjustable parameter in the expression for the
human ai in Conolly et al. (2004) was determined from the model fit to the rat tumor
incidence data while the second parameter was determined from background rates of
cancer incidence in the human. Therefore, variations considered in ai were constrained
to only those that (a) did not meaningfully degrade the fit of the model to the rat tumor
incidence data and (b) were in concordance with background rates in the human.
Crump et al. (2008) also evaluated these variations with respect to their biological
plausibility. The sensitivity analysis on assumed initiated cell kinetics was thought to be
particularly important because there were no data to even crudely inform the kinetics
of initiated cells for use in the models, even in rats, and the two-stage clonal expansion
model is very sensitive to initiated cell kinetics (Gaylor and Zheng, 1996; Crump,
1994a, b).
This document is a draft for review purposes only and does not constitute Agency policy.
B-83 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Crump et al. (2008) note that, because the purpose of their analysis was to carry out a
2 sensitivity analysis, in order to illustrate certain points, only risks to the general U.S. population
3 from constant lifetime exposure to various levels of formaldehyde under the Conolly et al. (2004)
4 environmental scenario (8 hours/day sleeping, 8 hours/day sitting, and 8 hours/day engaged in
5 light activity) are considered. Fits based on the hockey-stick and J-shaped models were identical,
6 and, of the three estimated parameters ([ibasal, multb, and D), only the estimate of [ibasal differed
7 between the two models.
8 Effect of background rates of nasal tumors in rats on human risk estimates
9 Crump et al. (2008) quantitatively evaluated the impact of different control groups on
10 estimates of additional human risk as follows:
11 1) Concurrent controls plus all NTP controls:, the same as used by Conolly et al. (2004);
12 2) Concurrent controls plus controls from NTP inhalation studies;
13 3) Only concurrent controls;
14 4) Each set of control data was applied with both the J shape and hockey-stick models in
15 Conolly et al. (2004) for afflux) and ai(flux) for a total of six analyses.
16 5) Uncertainties associated with an or ai are not addressed. Parameters amax, multfc, and
17 KMU were estimated in exactly the same manner as in Conolly et al. (2004).
18 Crump etal. (2008) present the following dose-response predictions of additional risk in
19 humans from constant lifetime exposure to various levels of formaldehyde arising from exercising
20 the above six cases. Their plots are reproduced in Figure F-l, where the corresponding curves
21 based on Conolly et al. (2004) are also shown for comparison.
This document is a draft for review purposes only and does not constitute Agency policy.
B-84 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
1E-3
i i i i i |
0.01
0.1
Formaldehyde Exposure (ppm)
Figure B-31. Effect of choice of NTP bioassays for historical controls on human
risk.
Note: Estimates of additional human risk of respiratory cancer by age 80 from lifetime exposure to formaldehyde
are obtained by using different control groups of rats.
Source: Crump et al. (2008).
The lowest dotted curve in Figure F-l represents the highest estimates of human risk
developed by Conolly et al. (2004). This resulted from use of the hockey-stick model for cell
division rates in conjunction with the statistical upper bound for the parameter KMU. As indicated
by the downward block arrows in the figure, their corresponding estimates based on the J-shaped
model were all negative for exposures below 1 ppm.
Consider next the solid curves in the figure, which show predicted MLE added risks that
were positive and less than 0.5. Crump et al. (2008) next examined the added risk obtained when
the MLE estimate of (KMUi^tasai) in these cases is replaced by the 95% upper bound of this
parameter ratio. The upper bound risk estimates in Conolly et al. (2004) were calculated in a
similar manner (but using all NTP historical controls). Except for minor differences, risk estimates
corresponding to such an upper bound and using all NTP controls were very similar in the two
efforts (Crump etal., 2008; Conolly et al., 2004).
Figure F-l shows that the choice of controls to include in the rat model can make an
enormous difference in estimates of additional human risk. For the J-shaped model for cell
replication rate both estimates based on the MLE and those based on the 95% upper bound on
KMU.-fibasai are negative for formaldehyde exposures below 1 ppm. However, when only concurrent
This document is a draft for review purposes only and does not constitute Agency policy.
B-85 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
controls are used in the model in Crump et al. (2008), the MLE from the J-shaped model is positive
and is more than three orders of magnitude higher than the highest estimates obtained by Conolly
et al. (2004). Using only concurrent controls, estimates based on the 95% upper bound on
KMU.-fibasai are unboundedly large (block arrows at the top of the figure). For the hockey-stick
shaped model for cell replication rate, when all NTP controls are used, the estimates based on the
MLEs are zero for exposures less than about 0.5 ppm. If only inhalation controls are added, the
MLEs are about seven times larger than the Conolly et al. (2004) upper bound estimates, and the
estimates based on the 95% upper bound on KMU:fibasai are about 50 times larger than the Conolly
et al. (2004) estimates. If only concurrent controls are used, both the MLE estimates and those
based on the 95% upper bound on KMU:^basai are unboundedly large.
Alternative assumptions regarding the rate of replication of initiated cells
For the human model, Conolly et al. (2004) made the same assumptions for relating ai(flux)
and Pi(flux) to afflux) as in their rat model (Conolly et al., 2003). That is, these quantities were
related by using eqs D-2 and D-3 (see Appendix D). As discussed in the context of the rat modeling,
by extending the shape of these curves to humans, the authors' model brings the cytotoxic action of
formaldehyde to bear strongly on the parameterization of the human model as well.
In the sensitivity analyses of the rat modeling in Appendix E, it was concluded that other
biologically plausible assumptions for ai and Pi resulted in several orders of magnitude variations
in the low dose risk relative to those obtained by models based on the assumptions in Conolly et al.
(2003) but that the highest risks were nonetheless of the order of 10-6 at the 10-ppb level. This
section examines how these uncertainties in the rat model propagate to the human model.
Crump et al. (2008) made minor modifications to the assumed division rates of initiated
cells in Conolly et al. (2004), while all other aspects of the model and input data were kept
unchanged. Two alternatives were considered for each of the J-shaped and hockey-stick models.
Figure F-2 shows the hockey-stick model for initiated cells in rats. In the first modification to the
hockey-stick model (hockey-stick Mod 1), rather than having a threshold at a flux of
1,240 pmol/m2-hour, the division rate increases linearly with increasing flux until the graph
intersects the original curve at 4,500 pmol/m2-hour, where it then assumes the same value as in the
original curve for larger values of flux. The second modification (hockey-stick Mod 2) is similar,
except the modified curve intersects the original curve at a flux of 3,000 pmol/m2-hour.
This document is a draft for review purposes only and does not constitute Agency policy.
B-86 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Formaldehyde Flux [pmol/(m'-h)]
Figure B-32. Conolly et al. (2003) hockey-stick model for division rates of
initiated cells in rats and two modified models.
Source: Crump et al. (2008).
1 Figure F-3 shows the rat J-shaped model for initiated cells. In the first modification to this
2 dose response (J-shaped Mod 1), rather than having a J shape, the division rate of initiated cells
3 remains constant at the basal value until the original curve rises above the basal value and has the
4 same value as the original curve for larger values of flux. In the second modification (J-shaped
5 Mod 2), the J shape is retained but somewhat mitigated. In this modification, the division rate
6 initially decreases in a linear manner similar to that of the original model but with a less negative
7 slope until it intersects the original curve at a flux of 1,240 [im/m2-hour, where it then follows the
8 original curve for higher values of flux.
This document is a draft for review purposes only and does not constitute Agency policy.
B-87 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information for Formaldehyde—Inhalation
Formaldehyde Flux [pmol/(m?-h)]
Figure B-33. Conolly et al. (2003) J-shaped model for division rates of initiated
cells in rats and two modified models.
Source: Crump et al. (2008).
Because the first constraint on the variation in ai was in concordance with the rat time-to-
tumor incidence data, Crump et al. (2008) applied each of the modified models in Figures F-2 and
F-3 to the version of the formaldehyde models in Subramaniam et al. (2007) that employed all NTP
controls and the hockey-stick curve for an. These authors restricted their analysis to this case
because their stated purpose was only a sensitivity analysis as opposed to developing alternate
credible risk estimates. Figure F-4 reproduces (from Crump et al. [2008]) curves of the cumulative
probability of a rat dying from a nasal SCC by a given age for bioassay exposure groups of 6,10, and
15 ppm. For comparison purposes, the corresponding KM (nonparametric) estimates of the
probability of death from a nasal tumor are also shown. Three sets of probabilities are graphed: the
original unmodified one and the ones obtained by using hockey-stick Mod 1 and Mod 2. Crump et
al. (2008) state that the changes in the tumor probability resulting from these modifications are so
slight that the three models cannot be readily distinguished in this graph.33 Thus, the modifications
considered to the models for the division rates of initiated cells caused an inconsequential change
in the fit of the model-predicted tumor incidence to the animal tumor data.
33The largest change in the tumor probability resulting from this modification for any dose group and any age up
through 900 days was found to be less than 0.002, a change so small that it would be impossible to detect, even in
the largest bioassays ever conducted. The changes in tumor probability resulting from the other modifications
described earlier were found to be even smaller. These comparisons were made in Crump et al. (2008) without
reoptimizing the likelihood. The authors note that reoptimization of the model subsequent to the variations would
have made the fit of modified models even better.
This document is a draft for review purposes only and does not constitute Agency policy.
B-88 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information for Formaldehyde—Inhalation
Time (days)
Figure B-34. Very similar model estimates of probability of fatal tumor in rats
for three models in Figure F-2.
Note: The differences are visually indistinguishable. Models were derived from the implementation of Conolly et
al. (2003) with the hockey-stick curves for al(flux) and aN(flux) and variants derived from modifications (Mod 1
and Mod 2, Figure F-2) to al(flux). Model probabilities are compared to Km estimates. The three sets of model
estimates are so similar that they cannot be distinguished on this graph.
Source: Crump et al. (2008).
The above modifications did not affect the basal rate of cell division in the model and
likewise had no effect on the fit to the human background data (Crump et al., 2008).
Crump et al. (2008) noted that, although the threshold model for initiated cells in Conolly et
al. (2003) was replaced with a model that had a small positive slope at the origin, the resulting
curves, hockey-stick Mod 1 and hockey-stick Mod 2, could have been shifted slightly to the right
along the flux axis in order to introduce a threshold for ai without materially affecting the risk
estimates resulting from these modified curves. Thus, "the assumption of a linear no-threshold
response is not an essential feature of the modifications to the hockey-stick model; clearly
threshold models exist that would produce essentially the same effect" (Crump et al. 2008).
Biological plausibility of alternate assumptions
These very small variations made to the ai in Conolly et al. (2003) are
• consistent with the tumor-incidence data (see Figure F-4);
This document is a draft for review purposes only and does not constitute Agency policy.
B-89 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
• small compared with the variability and uncertainty in the cell replication rates
characterized from the available empirical data (at the formaldehyde flux where ai was
varied);
• supported (qualitatively) by limited data, suggesting increased cell proliferation at doses
below cytotoxic;
• perturbations to be expected on any dose response derived from laboratory animal data
because of human population variability in cell replication; and
• biologically plausible because cell cycle control in initiated cells is likely to be disrupted.
The averaged cell replication rate constants as tabulated in Table 1 of Conolly et al. (2003)
and shown by the red curve in Figure E-2 of Appendix E (for various exposure concentrations and
corresponding average formaldehyde flux values in the F344 rat nose) demonstrate an increase
over baseline values only at exposure concentrations of 6 ppm and higher. Increased cell
proliferation at these concentrations of formaldehyde, whether transient or sustained, have been
associated in the literature with epithelial response to the cytotoxic properties of formaldehyde
(Conolly, 2002; Monticello and Morgan, 1997; Monticello et al., 1996,1991). The labeling data are
considered to show a lack of cytotoxicity and regenerative cell proliferation in the F344 rat at
exposures of 2 ppm and below (Conolly, 2002). In the Conolly et al. (2003) modeling it is further
assumed that the formaldehyde flux levels at which cell replication exceeds baseline rates remain
essentially unchanged when extrapolated to the human and for initiated cells for the rat as well as
the human. These assumptions need to be first viewed in the context of the uncertainty and
variability in the data on normal cells discussed in Appendix E.
Arguments for a hockey-stick or J shape over the background have been made in the
literature for sustained and chronic cell replication rates. However, the analyses of the cell
replication data show that the data are not consistently (over each site and time) indicative of a
hockey-stick or J shape as the best representation of the data (see Appendix E). This uncertainty is
particularly prominent when examining the cell replication data at the 13-week exposure time and
the pooled data from the PLM nasal site from Monticello et al. (1996) (see Figures E-l [dotted
curve], E-3B, and E-4 of Appendix E). The earliest exposure time in this experiment was at 13
weeks, and the 13-week cell replication data appear to be more representative of a monotonic
increasing dose response without a threshold; it is possible that early times are of more relevance
to the carcinogenesis as well as for considering typical (frequent short duration) human exposures.
Recently, Meng et al. (2010) measured cell replication in the anterior lateral meatus of the
F344 rat using continuous labeling on rats exposed to all the concentration levels in the Monticello
et al. (1996) experiment Labeling index (i.e., LI, as opposed to ULLI in the Monticello experiment)
was measured as the percentage of BrdU-labeled cells among the total number of cells counted at
the nasal site. Their data are reproduced below in Figure F-5, where the asterisk denotes the
observation of a statistically significant difference from the control group (Dunnett's test, p < 0.01).
This document is a draft for review purposes only and does not constitute Agency policy.
B-90 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplemental Information for Formaldehyde—Inhalation
These data appear to be consistent with a monotonically increasing dose-response shape for cell
replication. Linear regression provided good fits to all of the data (R2 = 0.97) as well as to the
subset of the data obtained by deleting the higher dose data at 10 and 15 ppm exposures (R2 =
0.84). We cite these data in support of considering the modifications carried out in Figure F-2.
For initiated cells, there are no data on which to evaluate the modifications made in
Section F.2.2 to these rates. However, some perspective can be gained by comparing them to the
variability in the division rates obtained from the data on normal cells used to construct the
formaldehyde model. As shown in Figure E-2 and discussed further in Subramaniam et al. (2008),
these data show roughly an order of magnitude variation in the cell replication rate at a given flux.
As part of a statistical evaluation of these data, a standard deviation of 0.32 was calculated for the
log-transforms of individual measurements of division rates of normal cells (Crump etal., 2008).
By comparison, the maximum change in the log-transform division rate of initiated cells resulting
from hockey-stick Mod 2 was only 0.20, and the average change would be considerably smaller.
Thus, although there are no data for initiated cells, it can be said that the modifications introduced
in Crump et al. (2008) for initiated cells are extremely small in comparison to the dispersion in the
data for normal cells.
Formaldehyde (ppm)
Figure B-35. Cell proliferation data from Meng et al. (2010).
The y-axis shows the percentage of Brdll-labeled cells among the total number of cells counted in the ALM section
of the rat nose.
Reproduced with permission from Meng et al. (2010).
Subramaniam etal. (2008) also point to some additional, albeit limited, data, suggesting
that exposure to formaldehyde could result in increased cell replication at doses far below those
This document is a draft for review purposes only and does not constitute Agency policy.
B-91 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Supplemental Information for Formaldehyde—Inhalation
that are considered to be cytotoxic. Tyihak et al. (2001) treated different human cell lines in culture
to various doses (0.1-10 mM) of formaldehyde and found that the mitotic index increased at the
lowest dose of 0.1 mM. These findings considered along with human population variability and
susceptibility (for example, polymorphisms in ADH3 [Hedberg et al., 2001]) indicate that it is
necessary to consider the possibility of small increases in the human ai over baseline levels at
exposures well below those at which cytotoxicity-driven proliferative response is thought to occur.
Heck and Casanova (1999) have provided arguments to explain that the formation of DPCs
by formaldehyde leads to inhibition of cell replication (i.e., if this effect alone is considered, normal
cell replication rate of the exposed cells would be less than the baseline rate). However, this
hypothesis was posed for normal cells. Subramaniam et al. (2008) argue that if an initiated cell is
created by a specific mutation that impairs cell cycle control, the effect would be to mitigate the
DPC-induced inhibition in cell replication, either partially or fully, depending on the extent to which
the cell cycle control has been disrupted. In the absence of data on initiated cells, the above
argument provided biological motivation to the modification applied to the J-shaped model for cell
division (Crump etal. 2008).
Thus, the previous paragraphs suggest that the changes made in the analysis in Crump et al.
(2008) to the assumption by Conolly et al. (2003) regarding the dose response for the division rate
of initiated cells are plausible.
Effect of alternate assumptions for initiated cell kinetics on human risk estimates
Figure F-6 contains graphs of the additional human risks estimated (in Crump et al. [2008])
by applying these modified models for ai and using all NTP controls, compared with those obtained
by using the original Conolly et al. (2004) model. Each of the four modified models presents a very
different picture from that of Conolly etal. (2004). At low exposures, these risks are three to four
orders of magnitude larger than the largest estimates obtained by Conolly et al. (2004).
This document is a draft for review purposes only and does not constitute Agency policy.
B-92 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information for Formaldehyde—Inhalation
Formaldehyde Exposure (ppm)
Figure B-36. Graphs of the additional human risks estimated by applying
these modified models for ab using all NTP controls, compared to those
obtained using the original Conolly et al. (2004) model.
Source: Crump et al. (2008).
These results have been criticized by Conolly et al. (2009) as being unrealistically large and
above the realm of any epidemiologic estimate for formaldehyde SCC. Thus, they argue that the
parameter adjustments made in Crump et al. (2008) are inappropriate. Crump et al. (2009)
rebutted these points by arguing that the purpose of their work was not to provide a more reliable
or plausible model but to carry out a sensitivity analysis. They argued that the changes made to the
model (in their analyses) were reasonable because they did not violate any biological constraints or
the available data. Further, they pointed out that "by appropriately mitigating the small
modifications [they] made to the division rates of initiated cells, the model [would] provide any
desired risk ranging from that estimated by the original model up to risks 1,000-fold larger than the
conservative estimate in Conolly etal. (2004)."
Crump et al. (2008) also evaluated the assumption in equation D-3 of the CUT modeling
pertaining to initiated cell death rates (/?/) by making small changes to ff. They report that they
obtained similarly large values for estimates of additional human risk at low exposures. Obtaining
reliable data on cell death rates in the nasal epithelium appears to be an unusually difficult
proposition (Hester et al., 2003; Monticello and Morgan, 1997), and, even if data are obtained, they
are likely to be extremely variable.
This document is a draft for review purposes only and does not constitute Agency policy.
B-93 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Supplemental Information for Formaldehyde—Inhalation
B.2.3. Estimates of Cancer Risk Using DNA Adduct Data from Animal Toxicology Studies and
Background Incidence
DNA Adduct-Based Approach
Recently, Lu et al. (2010) developed a highly sensitive MS method using [13CD2]-
formaldehyde that reportedly distinguishes whether formaldehyde-induced hydroxymethyl-DNA
monoadducts, in particular, the /V2-hydroxymethyl-dG (/V2-hmdG) adduct, originate from
endogenous or exogenous sources of formaldehyde (Lu et al., 2010; Lu et al., 2011; Moeller et al.,
2011;Yu et al., 2015). They quantified these mono adducts formed from both sources in various
tissues of rats and monkeys: nasal cavity, bone marrow, mononuclear white blood cells, spleen,
thymus, tracheal bronchial lymph nodes, mediastinal lymph nodes, trachea, lung, kidney, liver, and
brain. Swenberg et al. (2011) and Starr etal. (2016) used these adduct measurements and data on
the background incidences of nasopharyngeal cancer, Hodgkin lymphoma, and leukemia in the U.S.
population to develop cancer risk estimates by attributing the background incidences to
endogenous formaldehyde, using the measured endogenous iV2-hmdG adducts formed by
formaldehyde in specific tissues as a biomarker of exposure. Their risk model assumes a linear
relation between cancer incidence and iV2-hmdG adduct levels over the concentration range of
endogenous adducts as well as in the low-exposure range for exogenous adducts.
The authors stated that the approach has the following distinct advantages over traditional
approaches:
• risk estimates are assumed to conservatively bound the added lifetime risk at low
environmental exposures;
• use of the N2-hmdG adduct as an intracellular metric of formaldehyde dose to the DNA has
distinct advantages over the exposure estimates used in analyzing epidemiologic data;
• the method does not rely upon bioassay data from a limited number of animals; and
• the approach overcomes the uncertainty associated with extrapolating downward from
higher doses to typically environmentally relevant doses.
Specifically, their approach for risk estimation used the following steps:
1) DNA mono-adducts were used in the risk model as a marker of exposure (i.e., repairable)
as opposed to a marker of effect (i.e., heritable mutations). iV2-hmdG and iV6-hmdA mono-
adducts of formaldehyde were expressed in units of relevant adducts per 107 dG and 107
dA, respectively. While both adducts were reportedly formed by endogenous
formaldehyde, only /V2-hmdG adducts were detectable from exogenous formaldehyde.
2) Adducts formed endogenously were distinguished from those formed due to exogenous
sources using 13CD2-formaldehyde coupled with MS methods.
This document is a draft for review purposes only and does not constitute Agency policy.
B-94 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
3) Endogenously and exogenously formed mono-adducts were measured in various tissues:
nasal cavity, bone marrow, spleen, thymus, and mononuclear white blood cells (rats); nasal
cavity, bone marrow (monkeys).
4) Adducts were measured in rats after one 6-hour exposure to 0.7, 2.0, 5.8, 9.1, and 15.2 ppm
formaldehyde and five 6-hour exposures to 10 ppm, and in monkeys (cynomolgus
macaques) after two 6-hour exposures to 2 and 6 ppm. There were no measurements
carried out in unexposed animals.
5) No exogenous adducts were detected in any of the distant tissues (bone marrow, spleen,
thymus, white blood cells); therefore, for these tissues the adduct levels were estimated by
considering the limit of detection (LOD) of the method as an upper-bound estimate. This
LOD was converted to the equivalent level of iV2-hmdG adducts per 107 dG. The LOD
values were 0.0177 and 0.001034 iV2-hmdG adducts per 107 dG for the rat and monkey
data, respectively (Swenberg et al. 2011, Table 3).
6) Time-course data were collected in rats at the 10 ppm exposure concentration only. These
data were used to derive the half-life (ti/2) for repair of the iV2-hmdG adduct, and the same
value was assumed for all exposure concentrations.
7) Unit risks for nasopharyngeal cancer (NPC), Hodgkin lymphoma (HL) and leukemia were
calculated as follows:
Determine lower confidence limits on the endogenous iV2-hmdG adduct levels measured in
Step 3.
Assume the endogenous adduct level measured in rats to be the same in humans.
Convert exogenous iV2-hmdG adduct levels from 6-hour exposure values to adduct levels to
be expected under steady-state continuous exposure using the estimated ti/2.
Assume adduct levels are a linear function of exposure (adduct) concentration, passing
through the origin. Calculate the adduct per ppm ratio. Then, from c) above, calculate
the continuous adduct level corresponding to 1 ppm.
Convert the continuous adduct level corresponding to 1 ppm exposure from rat to human
by assuming that adduct levels scale in proportion to formaldehyde flux to the nasal
tissue in each species. For the monkey, assume that humans receive the same levels of
formaldehyde flux.
Consider endogenous and exogenous iV2-hmdG adducts formed by formaldehyde to be
biochemically indistinguishable (both were similarly related to low-dose formaldehyde
carcinogenicity).
Use the U.S. population background lifetime incidence probabilities of NPC (7.25 x 10 4), HL
(2.3 x 10"3), and leukemia (1.3 x 10 2). Swenberg et al. (2011) consider values provided
in the EPA draft assessment (for NPC) and the SEER Cancer Statistics Review (for HL
and leukemia). Attribute these lifetime risks to the endogenous formaldehyde levels
indicated by the adduct levels in step a (i.e., to the lower confidence limit on endogenous
formaldehyde iV2-hmdG adducts in the nose, bone marrow, or mononuclear white blood
This document is a draft for review purposes only and does not constitute Agency policy.
B-95 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 cells). Thus, calculate unit risk estimates for these specific cancers, expressed in units of
2 risk per W2-hmdG adduct per 107 dG.
3 Using the unit risk estimates determined in Step g, calculate upper confidence limit on
4 cancer risks for the continuous steady-state exogenous adduct level calculated in Step e,
5 which corresponds to 1 ppm inhaled formaldehyde exposure concentration.
6 Results
1 The mean A?2-hmdG adduct levels averaged over the animals in any exposure group and the
8 standard deviations in these values are reproduced below in Table 1 from Swenberg et al. (2011),
9 Table B-29. Mean formaldehyde-induced N2-hmdG adducts in rats and
10 monkeys (Swenberg et al., 2011)
Formaldehyde-Induced N" Hydroxymethyl-dG Adducts in Rats
Exposed to 10 ppm Formaldehyde for 1 Day or 5 Days, Rats
Exposed to Different Concentrations of Formaldehyde for 1
Day, and Monkeys Exposed to 2 and 6 ppm Formaldehyde for 2
Days
N~-HOCH2-dG
Exposure (adducts/107 dG ± S.D.)
and
Species
period
Tissues
Endogenous
Exogenous
Rat
l()ppm/l day
Nose
2.63 ± 0.73
1.28 ± 0.49
Bone marrow
1.05 ±0.14
n.d.
Blood
1.28 ± 0.38
n.d.
10 ppm/5 day
Nose
2.84 ± 1.13
2.43 ± 0.78
Bone marrow
1.17 ±0.35
n.d.
Blood
1.10 ±0.28
n.d.
Rat
0.7 ppm/1 day
Nose
3.62 ± 1.33
0.039 ± 0.011
2 ppm/1 day
Nose
6.09 ± 3.03
0.19 ± 0.08
5.8 ppm/1 day
Nose
5.51 ± 1.06
1.04 ± 0.24
9.1 ppm/1 day
Nose
3.41 ± 0.46
2.03 ± 0.43
15.2 ppm/1 day
Nose
4.24 ± 0.92
11.15 ± 3.01
Bone marrow
18.2 ± 0.47
n.d.
Monkey
2 ppm/2 days
Nose
2.49 ± 0.40
0.25 ± 0.04
Bone marrow
17.5 ± 2.6
n.d.
6 ppm/2 days
Nose
2.05 ± 0.54
0.41 ± 0.05
Bone marrow
12.4 ± 3.6
n.d.
Note. n.d., not detected.
11 Source: Swenberg et al. (2011), Table 1
12 Swenberg etal. (2011) calculated what they characterized as "upper-bound" risk estimates
13 at 1 ppm from these aggregate measurements based on the steps outlined in #7 above. These
14 values were then compared with the risk estimates in EPA's 2010 draft toxicological review, which
15 were obtained by linearly extrapolating from a point of departure derived by dose-response
16 modeling of the epidemiological data. When adduct data from rats were used, the risk estimates at
17 1 ppm exposure concentration ranged from 0,9xl0~3 to 7.5x10 3 for NPC and were at most 20.9x10-5
18 for HL and 12.6x10-4 for leukemia using adduct data from the nose, bone marrow and mononuclear
19 white blood cells, respectively. When the corresponding monkey adduct data were used, the risk
This document is a draft for review purposes only and does not constitute Agency policy.
B-96 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 estimates were 0.39x10-3 and 0.54x10-3 for NPC, and were atmost 5.5xl0-6for leukemia. In
2 contrast, the EPA upper-bound risk estimates at lppm were 1.1x10-2 for NPC, 1.7x10-3 for HL, and
3 5.7x10-2 for leukemia, and are higher than the adduct-based upper-bound estimates: 1.5 to 29-fold
4 for NPC, at least 81-fold for HL, and at least 45-fold (rat adduct data) or 10,000-fold (monkey
5 adduct data) for leukemia (Table 3, Swenberg et al. 2011).
6 Basis for Upper-Bound Claim:
7 Swenberg et al. (2011) state thattheir risk estimates are conservative upper bounds, and
8 cite the following reasons as support:
9 1) The background risks of specific cancers are fully attributed to the internal dose
10 represented by the endogenous iV2-hmdG adducts measured in the corresponding tissue.
11 2) Only iV2-hmdG adducts are included (the unit risk would be lower if other higher
12 endogenous adducts are included).
13 3) A linear risk model is assumed.
14 4) Exogenous adduct levels are assumed to be a linear function of exposure concentration,
15 passing through the origin. The slope of this line is based on the mean adduct
16 concentration at 10 ppm exposure which is an overestimate at low exposures because the
17 actual relationship of adduct levels versus ppm is highly nonlinear (upwardly concave).
18 This leads to a more conservative estimate for the cancer risk from step h of #7 above.
19 5) The 95% lower confidence bound on mean adduct level is used, which can be assumed to
20 correspond to the upper confidence bound on estimated risk.
21 6) Monkeys appear to have lower exogenous iV2-hmdG adduct levels than rats; therefore, risk
22 estimates based on scaling rat adduct levels to humans in proportion to formaldehyde flux
23 to nasal tissue would likely err on the side of being an over-estimate for humans.
24 7) Assumptions made in the derivation of ti/2 for iV2-hmdG adduct repair are conservative.
25 The time-course of adduct levels appears biphasic. However, Swenberg et al. (2011)
26 considered only the later slower part of the time course in deriving ti/2, attributing the
27 initial decay to cell-death at the high 10 ppm exposure concentration. Using a longer
28 estimate for ti/2 leads to higher estimates of steady-state adduct levels calculated in step 7e
29 above, thus, overestimating risk due to formaldehyde exposure (personal communication,
30 Dr. T. Starr to R. Subramaniam, 12-12-12).
31 Details on EPA Evaluation of Quantitative Issues
32 The main document pointed to several major issues that bear on the interpretation of the
33 measurements and their analysis. In this appendix, we provide further quantitative details to
34 illustrate our concerns.
35 1) Additivity of endogenous and exogenous adducts: Endogenous /V2-hmdG and /V6-hmdA
36 adducts were both measured in rat and monkey nasal tissues; on the other hand,
37 inhalation of formaldehyde resulted in a concentration-related pattern for exogenous N2-
This document is a draft for review purposes only and does not constitute Agency policy.
B-97 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
hmdG adducts only, and no detectable exogenous iV6-hmdA adducts. If these differences in
regards the observation of /V6-hmdA versus /V2-hmdG adducts are attributable to
differences in the effects of endogenous versus exogenous formaldehyde in inducing DNA
adducts, does this fit with the concept of additivity (step 7f) for endogenous and exogenous
formaldehyde?
2) Potential for interaction between exogenous formaldehyde exposure and endogenous adduct
levels? The Lu et al. (2010) and Moeller et al. (2011) studies used each exposed animal as
its own control rather than using a separate unexposed control group. This is problematic
if there is an exposure-related effect on the endogenous adduct levels, and two
observations point to the possibility of such an effect First, in a similar experiment in the
same laboratory, Lu et al. (2012) exposed rats orally to isotope-labeled methanol but
included a separate unexposed control group. In this case, they found that endogenous N2-
hmdG adducts showed exposure-dependent increases in many tissues compared with
control values. Second, EPA's analysis of the replicate animal data for adduct levels in the
nasal tissues (data kindly provided by Dr. Swenberg) indicates that at low exposures the
exogenous and endogenous adduct levels within a pooled34 group are correlated (see
Appendix II). In view of these observations, it is important to: a) consider the total
(endogenous plus exogenous) iV2-hmdG adduct level measured in an animal, and b)
include measurements from unexposed controls. We return to this point in issue #7
below.
3) Does the use of a linear risk model in Swenberg et al. (2011) necessarily yield an upper bound
on the low-dose risk? Swenberg et al. employ a linear model for modeling cancer risk due to
the endogenous dose and assume that using this model for upward extrapolation from
endogenous levels results in overestimating risk at exposures that are not high enough to
cause cytotoxicity. This assumption is examined first in general below and then
specifically for the formaldehyde adduct data from Swenberg etal. (2011).
By virtue of the additivity assumption (#7f), the effective dose to the DNA is represented by
the total iV2-hmdG adduct (endogenous plus exogenous) level. That is, the bottom-up approach
allows the traditional dose-response curve (extra risk versus externally derived dose) to be
rescaled so that the dose measure associated with zero external dose is now considered a positive
dose equal to the levels found in tissues not exposed to an external source, and the line of zero extra
risk is at a positive risk designated as the background risk. Furthermore, it is reasonable to assume
that the shape of the true dose-response curve is differentiable at the endogenous adduct level, and
is concave upward at dose levels used in rodent bioassays (i.e., following typically used dose-
response functions used in modeling the probability of tumor incidence, the slopes get steeper as
dose increases and the second derivative is positive). Then it is clear from Figure 1 that the
bottom-up approach can never overestimate the relevant low-dose slope; any straight line between
two points on the concave upward curve will underestimate the slope of the curve at the higher of
the two doses. The thick solid curve in the Figure XX represents risk as a function of the lower
34ln order to get adequate DNA for the chromatogram, Lu et al. (2011) pooled (i.e., combined) the DNA from
individual rats into groups of four for the 0.7 ppm exposure data and into groups of two (except for a single sample
in one group) for the 2.0 ppm. There was no pooling of samples at the 6, 9, and 15 ppm exposures.
This document is a draft for review purposes only and does not constitute Agency policy.
B-98 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
confidence bound on total iV2-hmdG adduct The adduct-based unit risk is the slope of the straight
line calculated based on the background risk of developing a specific cancer and the lower
confidence bound on the mean endogenous iV2-hmdG adduct level (indicated by the arrows in the
figure). The dashed line shows the upward extrapolation of this risk. It is possible, nonetheless,
that the extent of underestimation discussed above can be off-set by the conservatism in attributing
all cancers of the specified type to the endogenous dose. However, this is difficult to assess, as
discussed further in Appendix II.
Furthermore, the slope of increased risk with increasing adduct levels may not be linear
even over the range of the endogenous adducts; the slope may be concave upward as endogenous
defensive mechanisms become less effective in dealing with endogenous adduct levels as adduct
levels increase over the endogenous range. This seems a plausible scenario, as organisms would
have evolved some level of defensive mechanisms to deal with endogenous levels of adducts, yet
there is an energy cost associated with over-capacity; thus, these defensive capabilities are not fully
effective over the entire endogenous range, and this is consistent with the observance of
"background" rates of cancer. Under this plausible scenario, the actual slope of the adduct-based
unit risk estimate at the lower confidence bound on the mean endogenous iV2-hmdG adduct level
may be substantially higher than that suggested by a linear relationship over the endogenous range
and, thus, the slope obtained from the linear assumption does not necessarily provide an upper
bound on risk.
It may be noted that the bottom up approach is not consistent with the concept of additivity
to background disease processes on the basis of which local linearity in the proximity of zero
exogenous dose is thought to be reasonable. The bottom up approach requires a linear dose
response below zero exogenous dose which is not required to assume additivity to background.
In Appendix B.2.3, we demonstrate that the bottom up approach should be expected to
substantially underestimate the low-dose risk for formaldehyde by roughly 19-fold when the
principle is applied to the nasal cancers in the F344 rat
1) Use of adduct data from high-dose exposures where cell-killing may occur:
Lu et al. (2011) determined adduct half-lives based on time-course measurements at high
exposure concentrations. This is problematic because it is well known that cytotoxicity
has a strong influence on DNA repair rates (Rajewsky et al. 2000). Another
complication is that cell proliferation will result in diluting the adduct concentration
and needs to be accounted for in extrapolating data from the short-duration high
exposures to continuous steady-state levels.
To exclude adduct loss due to cytolethality at high exposures, Lu et al. (2011) deleted the
initial data in the time-course measurements used to calculate half-life associated with
adduct loss due to repair; see item vii above. However, loss of adducts due to cell-killing
was not considered as a factor by Swenberg et al. (2011) when the calculation in Step 7c
(page S135 of their paper) for the derivation of continuous steady-state levels was
applied to the adduct data generated in Lu et al. (2011) at 6,10, and 15 ppm. Omission
This document is a draft for review purposes only and does not constitute Agency policy.
B-99 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information for Formaldehyde—Inhalation
of this factor will lead to underestimating the steady state adduct levels from
measurements following 6-hour exposures at these concentrations. This contributes to
an underestimation of those human risk estimates in Table 3 of Swenberg et al. (2011)
that are based upon measured exogenous adducts at 15,10, 9, and 6.0 ppm.
2) Basis for assuming similar endogenous adduct levels in rats and humans: The basis for this
claim needs further clarification. The data as reproduced in Table 1 indicate endogenous
iV2-hmdG adduct levels in nasal tissues in monkeys to be lower than those in rats;
therefore, assuming human endogenous levels to be the same as the mean levels measured
in the rat may lead to an underestimate of human NPC risk. (Swenberg et al. 2011 and Lu et
al. 2011 state that endogenous levels are higher in monkeys than in rats but the data they
present in their Tables point to the contrary.)
3) Implication of large variability in endogenous levels: Exogenous /V2-hmdG levels as a
function of exposure concentration are nonlinear (see Figure A2, Appendix), and inter-
individual variability in endogenous iV2-hmdG levels is large (see Figure Al, Appendix). If
total adduct levels are considered (see Figure A3, Appendix), Figures A1-A3 indicate that a
sizable fraction of the animal population will be in the nonlinear region of the iV2-hmdG
adduct vs exposure concentration curve simply by virtue of the higher endogenous levels.
It is reasonable to think that endogenous levels in the human population will have an even
more variable distribution than a particular strain of laboratory animal.35 Therefore, and
considering the issue discussed in #3 above, if endogenous and exogenous formaldehyde
are equipotent as assumed in Step 7f, then total (endogenous + exogenous) adduct levels
should be used when developing a dose-response model based on iV2-hmdG adduct as a
metric of formaldehyde dose to the DNA. If the data permit, endogenous and exogenous
levels in the same animal should be paired.
Swenberg et al. (2011) inferred from their analyses that the higher risk estimates derived
by EPA from the NCI data for NPC are not credible. Taken together, these seven issues suggest that
the conclusion in Swenberg et al is premature and that it is not possible to characterize the results
using this approach as providing a conservative upper bound on cancer risk. Notwithstanding this
limitation, the bottom-up approach in Swenberg et al. (2011) is particularly attractive when other
phenomena such as significant cytotoxicity and subsequent impact on DNA repair prior to
mutations are occurring at higher doses. Because the approach does not use the higher-dose data
(other than to identify the type of tumors of concern for analysis), it provides a unique perspective
on risk estimates derived from these data.
35ln addition to the intrinsic variability of endogenous adducts, the dose dependence of adduct repair rate will
contribute significantly to the variability in the total adduct level.
This document is a draft for review purposes only and does not constitute Agency policy.
B-100 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
total adduct
Figure B-37. Schematic of typical dose-response curves with axes shifted to
include background dose and risk, (endog l.c.b.= endogenous lower
confidence bound; bg= background). Thin solid and dashed lines= unit risk and
extrapolation in bottom up approach.
This document is a draft for review purposes only and does not constitute Agency policy.
B-101 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 Additional Details Regarding Issues Raised during EPA Evaluation of the Results Regarding
2 Potential Cancer Risk Using DNA Adduct Data Presented in Swenberg et al. (2011)
3 This appendix provides further details to support the issues highlighted (in the main text)
4 following EPA's evaluation of the approach and results in Swenberg et al. (2011).
5 The individual animal data for /V2-hmdG adduct levels in the nasal tissues were kindly
6 provided to EPA by Dr. Swenberg and are reproduced in Figures Al (endogenous), A2 (exogenous),
7 and A3 (endogenous plus exogenous); the original paper, Lu etal. (2011), provides only the
8 summary data. In order to obtain adequate DNA for analysis, these authors pooled the tissues from
9 animals exposed to 0.7 and 2.0 ppm concentrations of formaldehyde into groups of 4 and 2 animals
10 per sample, respectively. Data at the higher exposure concentrations were not pooled.
11 The issues are numbered as per their occurrence in the main text.
12 • #3. The replicate exogenous and endogenous iV2-hmdG adduct levels were plotted against
13 each other for each exposure concentration in order to explore if these observations were
14 correlated. These scatter plots are shown in Figures A4 and A5 and indicate that the
15 exogenous and endogenous levels are correlated for the 0.7 and 2.0 ppm but not for the
16 higher exposure concentrations.
17 • #4. In Section 2.2.X we discussed thatthe bottom-up approach in Swenberg etal. (2011)
18 would underestimate the relevant low-dose slope in the context of typically used dose-
19 response functions used in modeling the probability of tumor incidence. This is
20 demonstrated here using the mean adduct data from Lu et al. (2011). Because the adduct
21 data are for rats, and the background tumor incidence in rats can be estimated from
22 historical control data, it is most appropriate to base the discussion on the rat tumor
23 incidence dose-response curve. Table Al reproduces the mean adduct levels as reported by
24 Swenberg et al. (2011) and the summary tumor incidence rate. The data point at 15 ppm is
25 not included because there is considerable toxicity at this level and the purpose of this
26 exercise is illustrative.
27 Table B-30. N2-hmdG adduct levels (Lu et al., 2011) and rat tumor data
28 (Monticello et al., 1996; Subramaniam et al., 2007)
Exposure ppm
Mean exogenous N2-
hmdG (adducts/107 dG)
Total /V2-hmdG
(endogenous3 + exogenous)
(adducts/107 dG)
Tumor incidence
0
0
0
0
0
0
4.70
1/3,602 = 0.00028
0.7
0.04
4.74
0/107
2.0
0.19
4.89
0/353
5.8
1.04
5.74
3/343=0.009
9.9b
2.26
6.96
22/103=0.214
a The mean endogenous level of 4.7 adducts/107 dG as reported by Swenberg et al. (2011, Fig. 2) was used.
b 9.9 ppm was the concentration in the tumor bioassay. However, the adduct levels in Lu et al. (2011) were
measured at 9.1 ppm; therefore, the exogenous adduct level was corrected with a linear extrapolation (value in
Lu etal., 2011, is 2.02).
This document is a draft for review purposes only and does not constitute Agency policy.
B-102 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Supplemental Information for Formaldehyde—Inhalation
These data are fit in Figure A6 with a multistage model with total adduct level for dose and
constrained to include a linear term; P(d) = l-exp(-a-d-b-dc) where d=total iV2-hmdG adduct level.
The value of the slope from the bottom up approach is 5.9-10 5 whereas the slope of the multistage
model fit to the tumor incidence at the background dose (mean endogenous adduct) is 1.1-10 3
which is 19-fold higher. This was analyzed on the basis of mean adduct levels and MLE dose
response but the conclusions would be conceptually similar if looking at upper bound estimates of
risk.
Does the conservative assumption that all the cancers of the specified type are attributable
to the dose off-set the degree of underestimation from a "bottom up" linear fit to a dose-response
curve? If one focuses only on the specified type of tumor, the assumption on its own appears to be
conservative. It is not, however, easy to ascertain whether that degree of conservatism would be
greater than the under-estimation illustrated above with the formaldehyde data. In addition, the
selection of the type of cancer is informed by, and thus dependent on, higher dose data. To the
extent the higher dose data did not detect other types of cancer, the attribution of all observed
cases of the selected tumor may not capture all the relevant cases.
individual animal data but partly pooled for 0.7 & 2 ppm
4 6 8 10
CH20 exposure cone (ppm)
12
14
16
Figure B-38. Endogenous N2-hmdG adducts as a function of formaldehyde
exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
B-103 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
o
CM
LO -
o
o
o
o
o
2 4 6 8 10 12 14
Exposure Cone (ppm)
Figure B-39. Endogenous N2-hmdG adducts as a function of formaldehyde
exposure concentration.
individual animal data
but partly pooled for 0.7 & 2 ppm
15
10
0.5 3.0 5.5 8.0 10.5
CH20 exposure cone (ppm)
13.0
15.5
Figure B-40. Total (endogenous plus exogenous) N2-hmdG adducts as a
function of formaldehyde exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
B-104 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
o
3.5 4.0
Endogenous
Endogenous
Figure B-41. Endogenous and exogenous adduct levels (adducts per 107 dG)
appear to be correlated for data from animals exposed to 0.7 (left) and 2.0
ppm (right) formaldehyde. The individual animals were pooled into several
groups (see text).
This document is a draft for review purposes only and does not constitute Agency policy.
B-105 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
6.0
Endogenous
3.6
Endogenous
4.0
Endogenous
Figure B-42. Endogenous and exogenous adduct levels from individual
animals appear to be uncorrelated for exposures of 6 (left), 9 (right), and 15
ppm (bottom) formaldehyde. Symbols are individual animal data.
This document is a draft for review purposes only and does not constitute Agency policy.
B-106 DRAFT—DO NOT CITE OR QUOTE
-------
0.25 -i
0.20
CD
O
c
CD
TJ
O °-15
C
o
E
cc
0.10
0.05 -
Supplemental Information for Formaldehyde—Inhalation
multistage model fit
f=1-exp(-a*x-b*xc); a, b>0
0.00*
0
Total Adduct (per 10 dG)
multistage model fit
f=1-exp(-a*x-b*xc); a, b>0
CD
o
c
CD
"O
O
c
o
E
0.0020
0.0015
0.0010
0.0005
b.g. incidence
bottom up sl°Pe
0.0000 x - —• 1 1 1
b.g. do§e
Total Adduct (per 107 dG)
Figure B-43. Underestimation of slope of dose response using bottom up
approach. Bottom up slope (dashed line); multistage model fit to tumor incidence
data, highest dose deleted (solid line). Multistage model parameters: a= 1.7847-10-
8, b=l.1421-10-15, c=16.9983. Bottom panel: Axes truncated so that difference
between curves at crossing point is visible.
This document is a draft for review purposes only and does not constitute Agency policy.
B-107 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 APPENDIX C. ASSESSMENTS BY OTHER NATIONAL
2 AND INTERNATIONAL HEALTH AGENCIES
Table C-l. Hazard conclusions and toxicity values developed by other national
and international health agencies
Organization
Conclusions and toxicity values
Agency for Toxic Substances and
Disease Registry (ATSDR, 1999)
Chronic inhalation minimal risk levels (MRL) = 0.008 ppm using a composite
uncertainty factor (UF) of 30, based on clinical symptoms of irritation of eyes and
upper respiratory tract and mild damage to the nasal epithelium in chronically
exposed workers (Holmstrom et al., 1989); Intermediate MRL = 0.03 ppm using
composite UF of 30 based on nasopharyngeal irritation in Cynomolgus monkeys (Rusch
et al., 1983); Acute MRL = 0.04 ppm using UF = 9 based on nasal and eye irritation in
human volunteers (Pazdrak et al., 1993).
Interim Acute Exposure Guideline
Levels (AEGLs) for Formaldehyde,
National Advisory Committee for
AEGLs for Hazardous Substances
(NRC, 2008)
AEGL-1 (nondisabling)—0.90 ppm (1.1 mg/m3) for exposures ranging from 10 min to 8
hr to protect against mild irritation, based on mild irritation in human subjects.
AEGL-2 (disabling)—14 ppm (17 mg/m3) for exposures ranging from 10 min to 8 hr to
protect against mild lacrimation with adaptation in humans.
AEGL-3 (lethal)—100 ppm (123 mg/m3) for a 10-min exposure to 35 ppm (43 mg/m3)
for an 8-hr exposure, the highest nonlethal values in the rat.
National Toxicology Program (NTP,
2011)
Known to be a human carcinogen based on sufficient evidence of
carcinogenicity from studies in humans (consistent findings for nasopharyngeal,
sinonasal, and myeloid leukemia) and supporting data on mechanisms of
carcinogenesis (Twelfth Report on Carcinogens, 2011).
National Institute of Occupational
Safety and Health (NIOSH, 2011)
Potential occupational carcinogen. Recommended exposure limit (REL)—0.016 ppm
(0.04 mg/m3) TWA for up to a 10-hr workday and a 40-hr work wk.
Occupational Safety and Health
Standard 1910.1048
Permissible exposure limit (PEL) for general industry—0.75 ppm (0.92 mg/m3) TWA for
an 8-hr workday; Short-term exposure limit: 2 ppm (2.5 mg/m3), 15-minute duration.
International Agency for Research on
Cancer, Monograph Vol. 88 (IARC,
2006); Monograph Vol. 100F (IARC,
2012)
Sufficient evidence in humans for the carcinogenicity of formaldehyde based on
nasopharyngeal cancer and leukemia (Group 1). Sufficient evidence in experimental
animals for the carcinogenicity of formaldehyde.
European Union, European
Commission, Scientific Committee on
Occupational Exposure Limits (SCOEL,
2016)
Carcinogen group C: genotoxic carcinogen with a mode-of-action-based threshold.
Occupational exposure limit (OEL)—8h-TWA of 0.3 ppm (0.369 mg/m3); STEL 15 min
of 0.6 ppm (0.738mg/m3) based on cytotoxic irritation in studies of human volunteers.
Health Canada (2005)
Residential Indoor Air Quality
Guideline
Short-term exposure: 123 ng/m3 (1-hr average) based on eye, nose, and throat
irritation (Kulle, 1993); long-term exposure: 50 ng/m3 (8-hr average) based on
respiratory symptoms in children with asthma (Rumchev et al., 2002).
(Health Canada, 2001)
Priority Substances List Assessment
Report
The inhalation of formaldehyde under conditions that induce cytotoxicity and
sustained regenerative proliferation is considered to present a carcinogenic
hazard to humans.
This document is a draft for review purposes only and does not constitute Agency policy.
C-l DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 APPENDIX D. SUMMARY OF EXTERNAL PEER
2 REVIEW AND PUBLIC COMMENTS AND EPA'S
3 DISPOSITION
4 This section itemizes the comments and recommendations regarding the June 2010 draft
5 toxicological review of formaldehyde that was released for public review and was also reviewed by
6 a committee of the National Research Council. The report by the NRC committee was sent to the
7 EPA in 2011. In light of the substantive recommendations to adopt a more systematic approach to
8 the assessment, the revision of the assessment involved a fresh start, and now includes explicit
9 rationales and criteria, and thorough documentation of all steps in the process from the literature
10 search through the development of toxicity values. Thus, this is a completely different document
11 Although the comments from the NRC Committee and the public may not be directly applicable to
12 the current assessment draft, many of the issues that were raised remain pertinent, and responses
13 were developed to address all of the comments that were received.
14 D.l. NRC FORMALDEHYDE PANEL SUMMARY RECOMMENDATIONS
15 SPECIFIC TO FORMALDEHYDE AND EPA RESPONSES
16 • General Recommendations From Executive Summary And Chapter 7
17 • Rigorous editing is needed to reduce the volume of the text substantially and address the
18 redundancies and inconsistencies; reducing the text could greatly enhance the clarity of the
19 document.
20 • Response: EPA has taken several steps to reduce the amount of text and to display relevant
21 information more clearly and succinctly in tables and graphs. Section 4 in the draft
22 reviewed by the NRC (Hazard Evaluation) was reorganized to describe the human and
23 animal evidence together by health hazard. An integrated weight of evidence for each
24 hazard is now included to enhance clarity. Repetition is minimized and all summaries and
25 conclusions have been carefully reviewed and edited to prevent inconsistency.
26 • Chapter 1 of the draft assessment needs to discuss more fully the methods of the
27 assessment, including a description of search strategies used to identify studies with the
28 exclusion and inclusion criteria clearly articulated and a better description of the outcomes
29 of the searches (a model for displaying the results of literature searches is provided later in
30 this chapter) and clear descriptions of the weight-of evidence approaches used for the
31 various noncancer outcomes. The committee is recommending not the addition of long
32 descriptions of EPA guidelines but rather clear concise statements of criteria used to
33 exclude, include, and advance studies for derivation of the RfCs and unit risk estimates.
34 • Response: The new Preface to the toxicological review describes the approaches used to
35 identify relevant studies and the process through which specific studies were reviewed for
This document is a draft for review purposes only and does not constitute Agency policy.
D-l DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
hazard identification and for use in derivation of toxicity values. Because literature
searches were conducted for each health hazard independently, the databases, search
strings, inclusion and exclusion criteria and diagrams displaying results are presented by
health hazard in the supplemental materials with a summary included for each health
hazard in Chapter 1. A framework developed for evaluating weight of evidence for
noncancer effects is also transparently described in the new introductory materials.
• Standardized evidence tables that provide the methods and results of each study are needed
for all health outcomes; if appropriate tables were used, long descriptions of the studies
could be moved to an appendix or deleted.
• Response: EPA has developed tables to summarize the studies in humans and animals that
were used to synthesize the evidence for specific endpoints and reduced the amount of text
that simply describes studies.
• All critical studies need to be thoroughly evaluated with standardized approaches that are
clearly formulated and based on the type of research, for example, observational
epidemiologic or animal bioassays. The findings of the reviews might be presented in tables
to ensure transparency.
• Response: EPA implemented these suggestions and applied a framework for systematic
review for the review of epidemiology and toxicology studies of formaldehyde inhalation
relevant to each considered hazard. The studies identified as potentially relevant to the
assessment of hazard during the literature searches were evaluated for their ability to
inform the hazard reviews using standardized approaches and were categorized by a level
of confidence (high, medium, low, and not informative). The issues pertinent to evaluating
the strengths and limitations of individual studies with respect to specific health endpoints
are discussed, and each study evaluation is documented in tables found in the supplemental
material for each health hazard. The results of the study evaluations (e.g., confidence) are
included in the evidence tables that summarize the studies found in each hazard section.
Studies identified as not informative are not included in the evidence tables and do not
contribute to hazard identification or dose-response decisions; these excluded studies are
identified (e.g., in the discussion of methods in each section; in the study evaluation tables
in the supplemental material). A simplified evaluation process was applied to mechanistic
studies informing potential mode of action for respiratory effects and genotoxic endpoints
(epidemiology studies for genotoxicity) and tables documenting the evaluations are found
in the supplemental materials.
• The rationales for selection of studies that are used to calculate RfCs and unit risks need to
be articulated clearly. All candidate RfCs should be evaluated together with the aid of
graphic displays that incorporate selected information on attributes relevant to the
database.
• Response: The rationale for selecting studies for RfCs derivation are presented in the
Preface to the assessment and in Chapter 2 of this toxicological review (see Sections X.X and
2.X). An array of the studies and the candidate values, including key uncertainties, was
developed and is found in Section 2.X to clearly present the information used by EPA in
developing the RfC.
This document is a draft for review purposes only and does not constitute Agency policy.
D-2 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
• The weight-of-evidence descriptions need to indicate the various determinants of "weight."
The reader needs to be able to understand what elements (such as consistency) were
emphasized in synthesizing the evidence.
• Response: EPA has clarified the considerations used in the synthesis of the available
studies pertaining to specific health effects (see Preface to the assessment). The syntheses
of studies in humans or animals for each health effect discuss how well the available data
address each of the criteria detailed in the preface; these are based on the considerations
described by Hill (e.g., consistency, response magnitude, etc.). For noncancer effects, a
framework was developed for synthesizing evidence from studies in humans and animals,
and integrating across all the evidence. This is described in the Preface to the toxicological
review and the results are presented for each health hazard or hazard subcategory. Within
an evidence stream, the evidence was characterized as robust, moderate, slight or
inadequate evidence for a hazard, or compelling evidence that no hazard exists. The lines of
evidence in humans and animals were then considered together, along with biological
plausibility and relevance to humans, when appropriate, to arrive at final causal conclusions
for a particular hazard. Documentation of this evaluation is included for each potential
health hazard, in Chapter 1. The evaluation of weight-of-evidence for carcinogenicity used
the same criteria and framework within epidemiology and animal evidence streams, and
then, based on the 2005 Carcinogenicity Guidelines and supplemental guidance for early life
exposures (EPA, 2005a, b) arrived at a conclusion with regard to causality.
• "In general, the committee found that the draft was not prepared in a consistent fashion; it
lacks clear links to an underlying conceptual framework; and it does not contain sufficient
documentation on methods and criteria for identifying evidence from epidemiologic and
experimental studies, for critically evaluating individual studies, for assessing the weight of
evidence, and for selecting studies for derivation of the RfCs and unit risk estimates" (pp. 3-
4).
• Response: As described above for comments 1.1-1.6, the toxicological review follows a
unifying conceptual framework, which is followed and documented throughout for
identifying the evidence, evaluating individual studies, synthesizing the evidence within and
across studies in humans and animals, and for deriving organ- or system-specific RfCs, the
overall RfC, and unit risk estimates.
• T oxicokinetics
• The committee agrees with EPA's conclusion that "certain formaldehyde-related effects
have the potential to modulate its uptake and clearance" (EPA 2010, pp. 3-5). Some of the
effects, such as changes in mucociliary function and altered nasal epithelium, could occur in
humans. However, reflex bradypnea and related modulating effects seen in rodents do not
occur in phylogenetically higher animals (nonhuman primates) or humans. Thus,
formaldehyde exposures at concentrations relevant for an RfC or unit risk are unlikely to
alter its toxicokinetics.
• Response: Consistent with the comment by the committee, the current draft assessment
does not argue that these effects on toxicokinetics occur at formaldehyde concentrations
relevant for an RfC or unit risk. The study results on changes in mucociliary clearance are
discussed in the supplemental materials and changes in nasal epithelium are discussed in
respiratory pathology hazard section. These discussions examine the concentration and
This document is a draft for review purposes only and does not constitute Agency policy.
D-3 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
duration relationships observed for formaldehyde. Reflex bradypnea in experimental
animals is discussed if relevant to the interpretation of the results of toxicology studies.
• Formaldehyde has also been measured in exhaled breath, but the interpretation of some
measurements made with mass spectrometry has been questioned (Spanel and Smith,
2008; Schripp etal., 2010). Spanel and Smith (2008) showed that a trace contaminant (up
to 1%) of the reagent gas used in real-time mass-spectrometric methods—specifically
proton-transfer reaction mass spectrometry (PTRMS) and selected ion flow tube mass
spectrometry (SIFT-MS)—reacts with endogenous methanol and ethanol that is normally
found in exhaled breath to produce the same main ion (mass-to-charge ratio of 31) as is
used to measure formaldehyde. Thus, they concluded that up to 5 ppb of the formaldehyde
concentration determined in the exhaled breath of humans reported in earlier studies that
did not account for this confounding may be due to methanol or ethanol and not
formaldehyde; that is, 1% of total background concentrations of methanol or ethanol of
about 500 ppb would be misclassified as formaldehyde. The committee concurs with EPA's
concerns as to whether some published exhaled breath measurements of formaldehyde are
analytically valid. The committee also notes that this methodologic problem is
inconsistently addressed by EPA in its reanalysis of the exhaled-breath experiments. The
committee concludes, however, that regardless of the methodologic issue related to breath
analysis, formaldehyde is normally present at a few parts per billion in exhaled breath after
the measurement error associated with a trace contaminant in the reagent gas used in
previous mass spectrometric methods is taken into account
• Response: It is difficult to say what range of formaldehyde concentration may be found in
exhaled breath, although levels are likely to be very low. Subjects in several of the cited
studies were inhaling formaldehyde at concentrations of about 10 ppb, so the inhaled air
contributed to the measurements of formaldehyde in exhaled air.
• A more recent study by Riess et al. (2010), published shortly after the NAS review
commenced, was not hindered by the limitations of previous studies. All subjects in this
study inhaled formaldehyde-free air. No formaldehyde could be detected in exhaled breath
of any subjects, including smokers, using a method with a limit of detection of <0.5 ppb.
EPA has reviewed its new text to discuss the issue consistently.
• Regardless of the technical limitations in the studies, the toxicity values derived in the
toxicological review are intended to protect the population from the extra risk imposed by
inhalation of formaldehyde in the air.
• The committee concludes that formaldehyde is an endogenous compound and that this
finding complicates assessments of the risk posed by inhalation of formaldehyde. The
committee emphasizes that the natural presence of various concentrations of formaldehyde
in target tissues remains an important uncertainty with regard to assessment of the
additional dose received by inhalation.
• Response: The natural presence of formaldehyde in target tissues can complicate assessing
risk on the basis of internal tissue concentrations. For many endpoints, however, including
many for which there is human epidemiology data, there are studies relating inhaled
formaldehyde concentrations directly to observed endpoints, and the target tissue
concentration is not an explicit part of the estimated dose-response relationship. These
studies allow EPA to estimate the extra risk of those endpoints as a result of inhaled
This document is a draft for review purposes only and does not constitute Agency policy.
D-4 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
formaldehyde that adds to naturally occurring formaldehyde concentrations in target
tissues, the extra risk associated with the extra exposure.
• Schroeter et al. developed a dosimetry model that incorporated published values of
endogenous formaldehyde levels. EPA has addressed the approach and results of this paper
in the revised document, and determined that the modeled inhaled flux of formaldehyde
adds linearly to background endogenous levels with inhaled exposure concentration.
• The draft IRIS assessment of formaldehyde provides an exhaustive discussion of
formaldehyde toxicokinetics, carcinogenic modes of action, and various models. Although
the committee agrees with much of the narrative, several issues need to be addressed in the
revision of the draft assessment First, there is broad agreement that formaldehyde is
normally present in all tissues, cells, and bodily fluids and that natural occurrence
complicates any formaldehyde risk assessment Thus, an improved understanding of when
exogenous formaldehyde exposure appreciably alters normal endogenous formaldehyde
concentrations is needed (pp. x and 44 [modes of action]).
• Response: The current draft assessment discusses the studies that evaluated formaldehyde
concentrations in upper respiratory tract tissues and blood after formaldehyde inhalation in
rodents (ref). The studies concluded that DPC in bone marrow associated with inhaled
formaldehyde were the result of metabolic incorporation of the inhaled formaldehyde in the
nasal tissues, not from distribution and direct interactions with the aldehyde in bone
marrow (Casanova-Schmitz and Heck 1983; Casanova-Schmitz et al. 1984). In addition, the
assessment discussed the research using sophisticated measurements of hydroxymethyl
DNA adducts differentiating between inhaled and endogenous formaldehyde in the upper
respiratory tract, blood and other organs (Lu etal., 2010, 2011; Moeller etal. 2011;
Swenberg etal. 2011, 2013; Yu etal., 2015). These studies did not find evidence that
inhaled formaldehyde is distributed substantially beyond the respiratory tract tissues.
Although there are remaining uncertainties regarding the extent that inhaled formaldehyde
is distributed, the lack of systemic distribution is an assumption used in the assessment to
provide a framework for presenting and interpreting the evidence concerning the potential
hazards of formaldehyde inhalation.
• One approach that EPA could use would be to complete an analysis of variability and
uncertainty in measuring and predicting target-tissue formaldehyde concentrations among
species. Only with such an analysis can one begin to identify and address openly and
transparently the question of how much added risk for an endogenous compound is
acceptable.
• Response: This assessment does not make judgments as to whether any specific added risk
is acceptable. The conclusions about potential health impacts are derived from evaluating
the relationships in available studies between different inhaled concentrations of
formaldehyde and observed health effects. As mentioned earlier, results in Shroeter et al.
are consistent with the assumption that inhaled formaldehyde at relevant concentrations
adds to mean endogenous concentrations in nasal tissue. We agree that more data on the
variability of endogenous formaldehyde concentrations among individuals would be useful
to the discussion of when (and in which individuals) tissue levels of exogenous
formaldehyde are significantly greater. For example, when the individual animal data on
DNA adducts formed by formaldehyde in Swenberg et al. (2013) were analyzed, a number
of animals had very high endogenous levels of these adducts. In these animals, even at a
This document is a draft for review purposes only and does not constitute Agency policy.
D-5 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
low inhaled exposure concentration of 2 ppm, the total (endogenous plus exogenous)
internal dose (as measured by the level of DNA adducts) was comparable to the mean total
internal dose measured in the group of animals exposed at 10 ppm (a dose at which
considerable carcinogenicity was observed in animal bioassays). Heck and co-workers
found the variability in endogenous levels to be greater than the difference between mean
endogenous and exogenous levels in nasal tissues of multiple species at the lowest exposure
levels in their studies. However, these data are from a small sample, and data from other
studies (Swenberg et al. 2013) suggest that the population variability in endogenous levels,
and the variation in endogenous levels across tissues, is likely to be large.
• A series of studies using dual-labeled (14C/3H) formaldehyde in rats has been performed to
address the analytic concern (Casanova-Schmitz and Heck 1983; Casanova-Schmitz etal.
1984). The draft IRIS assessment accurately summarizes the main conclusions reached
from those experiments, namely that "labeling in the nasal mucosa was due to both covalent
binding and metabolic incorporation," that "DPC [were] formed at 2 ppm or greater in the
respiratory mucosa," and that "formaldehyde did not bind covalently to bone marrow
macromolecules at any exposure concentration" (up to 15 ppm) (EPA 2010, pp. 3-12). The
labeling of bone marrow macromolecules was found by the investigators to be due entirely
to metabolic incorporation of the radiolabels, not to direct covalent binding of intact
formaldehyde. The committee views those findings as supporting the hypothesis that
inhaled formaldehyde is not delivered systemically under the exposure conditions used in
the studies (0.3-15.0 ppm, 6 hr) (EPA, 2010).
• Response: The current draft assessment concludes that, although uncertainties remain
regarding the extent that inhaled formaldehyde is distributed, the lack of systemic
distribution is an assumption used in the assessment to provide a framework for presenting
and interpreting the evidence concerning the potential hazards of formaldehyde inhalation.
• The committee also found that the more contemporary work performed by Lu et al. (2010)
that examined formaldehyde-induced DNA adducts and DDX cross links provided no direct
evidence of systemic availability of inhaled formaldehyde. The Lu et al. (2010) study used
13CD2-labeled formaldehyde and showed that 13CD2-formaldehyde-DNA adducts and DDX
were confined to the nasal cavity of exposed F344 rats, even though they examined much
more DNA isolated from bone marrow, lymphocytes, and other tissues at distant sites for
the adducts. The male Fischer 344 rats were exposed to [13CD2]-formaldehyde at 10 ppm
for 1 or 5 days (6 hr/day) with a single nose-only unit
• Response: Lu et al. (2010) is discussed in the current draft assessment draft, along with
more recent studies confirming and expanding these observations (Lu etal., 2011; Yu etal.,
2015). EPA agrees that this study shows that the formaldehyde monoadducts and DNA-
DNA cross links are detectable in nasal cavity, but not in bone marrow, of exposed rats. EPA
agrees that this study does not provide evidence that formaldehyde is transported to bone
marrow.
• The strongest data cited by EPA in support of systemic delivery of inhaled formaldehyde
come from several studies in which antibodies to formaldehyde-hemoglobin and
formaldehyde-albumin adducts were detected in blood from exposed workers, smokers,
and laboratory animals. The studies did not definitively demonstrate, however, whether
adduct formation occurs at a site distant from the portal of entry. For example, it is not
known whether the adducts could be formed in the airway submucosal capillary beds or
This document is a draft for review purposes only and does not constitute Agency policy.
D-6 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
reflect systemic delivery of formaldehyde. Moreover, the draft IRIS assessment does not
evaluate the antibody work as critically as the direct chemical-analysis approaches. The
committee found that the draft does not offer a sufficient basis for EPA's reliance on the
antibody data to support the hypothesis that formaldehyde (or its hydrated form,
methanediol) may reach sites distal to the portal of entry and produce effects at those sites.
• Response: Whether the antibodies detected in the blood indicated adducts formed in
airway submucosal capillary beds or in the blood is an uncertainty that is acknowledged in
the current draft assessment. All discussions in the toxicological review follow from the
premise that the evidence base does not support the hypothesis that the observed effects of
inhaled formaldehyde are due to its delivery (in any intact form, including its hydrated
form, methanediol) to systemic organs. These studies are discussed in the section on
possible modes of action for lymphohematopoietic cancers (Section l.X).
• Questions have arisen regarding the possibility that formaldehyde reaches distal sites as
methanediol. However, although equilibrium dynamics indicate that methanediol would
constitute more than 99.9% of the total free and hydrated formaldehyde, the experimental
data described above provide compelling evidence that hydration of formaldehyde to
methanediol does not enhance delivery of formaldehyde beyond the portal of entry to distal
tissues. Furthermore, Georgieva et al. (2003) used a pharmacokinetic modeling approach
that explicitly accounted for the competing processes of hydration, dehydration, diffusion,
reactivity with macromolecules, and metabolism and demonstrated that hydration-
dehydration reaction rates determined from equilibrium studies in water are not applicable
in biologic tissues, given that their use in the model resulted in simulations that were
inconsistent with the available data. For example, the calculated dehydration rate from
equilibrium dynamics studies in water was so small relative to other competing rates that
too little formaldehyde would be available to account for the measured DPC rates. Thus, the
data provide a strong indication that the hydration-dehydration reaction should not be rate-
limiting and can thus be ignored in modeling the disposition of inhaled formaldehyde in
nasal tissues.
• Response: EPA agrees that the hydration-dehydration reaction is not likely to play a
significant role in the disposition of formaldehyde following absorption into nasal tissue.
• EPA also suggested that systemic delivery of formaldehyde-glutathione adducts and latter
release of free formaldehyde may result in delivery of formaldehyde to sites distal to the
respiratory tract However, experimental data supporting that hypothesis are lacking, as
acknowledged by the draft IRIS assessment In fact, additional data based on even more
sensitive analytic methods published since the draft assessment was released casts further
doubt on the hypothesis that formaldehyde reaches the systemic distribution in a form that
can react with macromolecules in tissues remote from the portal of entry (Lu et al. 2011;
Moeller etal. 2011; Swenberg et al. 2011).
• Response: We agree with NAS that the hypothesis of GSH-mediated delivery of
formaldehyde lacks experimental support The current draft assessment includes the
studies by Lu etal. (2011), Moeller etal. (2011), Swenberg etal. (2011), and the more
recent report by Yu et al. (2015).
• The committee also found two divergent statements regarding systemic delivery of
formaldehyde in the draft IRIS assessment. Some parts of the draft assume that the high
This document is a draft for review purposes only and does not constitute Agency policy.
D-7 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
reactivity and extensive nasal absorption of formaldehyde restrict the systemic delivery of
inhaled formaldehyde to the upper respiratory tract (for example, EPA 2010, pp. 4-371).
Under that assumption, systemic responses—including neurotoxicity, reproductive toxicity,
and leukemia—are unlikely to arise from the direct delivery of formaldehyde (or
methanediol) to a distant site in the body, such as the brain, the reproductive tract, and the
bone marrow. Other portions of the document presume systemic delivery of formaldehyde
(or its conjugates) and use this presumption to account in part for the systemic effects (see,
for example, p. 4-1, lines 16-19; p. 4-472, line 18; Section 4.5.3.1.8; and p. 6-23, line 31). The
committee found the inconsistency to be troubling, and the divergent assumptions are not
justified.
• Response: All discussions in this draft toxicological review follow from the premise that the
evidence base does not support the hypothesis that the observed effects of inhaled
formaldehyde are due to its delivery (in any intact form, including its hydrated form,
methanediol) to systemic organs.
• The committee concludes that the issue of whether inhaled formaldehyde can reach the
systemic circulation is extremely important in assessing any risk of adverse outcomes at
nonrespiratory sites associated with inhalation of formaldehyde. Moreover, the committee
concludes that the weight of evidence suggests that it is unlikely for formaldehyde to appear
in the blood as an intact molecule, except perhaps after exposures at doses that are high
enough to overwhelm the metabolic capability of the tissue at the site of entry. Thus,
although many sensitive and selective investigative approaches have been used, systemic
concentrations from inhaled formaldehyde are indistinguishable from endogenous
background concentrations. The committee, however, notes the importance of
differentiating between systemic delivery of formaldehyde and systemic effects. The
possibility remains that systemic delivery of formaldehyde is not a prerequisite for some of
the reported systemic effects seen after formaldehyde exposure. Those effects may result
from indirect modes of action associated with local effects, especially irritation,
inflammation, and stress.
• Response: We agree with NAS that systemic delivery is not a prerequisite for systemic
effects. We also agree with NAS that the systemic effects could be due to indirect or
unknown modes of action. EPA conducted a systematic evaluation of the evidence pertinent
to possible mechanistic events responsible for the observed respiratory effects identified in
the toxicological review. Some of these events related to irritation, inflammation, and
oxidative stress may also be relevant to effects observed at distal sites, and this evidence is
included in the MOA discussions for nervous system effects, reproductive and
developmental toxicity, and myeloid leukemia (see Sections X.X.X).
• Inhaled formaldehyde, a highly reactive chemical, is absorbed primarily in the upper
airways and remains predominantly in the respiratory epithelium. The weight of evidence
indicates that formaldehyde probably does not appear in the blood as an intact molecule
except at doses high enough to overwhelm the metabolic capability of the exposed tissue.
The draft IRIS assessment presents divergent opinions regarding the systemic delivery of
formaldehyde that need to be resolved (pp. x and 44 [mode of action]).
• Response: The revised assessment presents a consistent view on the evidence regarding
the distribution of formaldehyde. All discussions in this draft toxicological review follow
This document is a draft for review purposes only and does not constitute Agency policy.
D-8 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
from the premise that the evidence base does not support the hypothesis that the observed
effects of inhaled formaldehyde are due to its delivery to systemic organs.
• In rewriting the sections of the draft IRIS assessment that pertain to the topics reviewed in
this chapter, EPA should consider the implications of the most recent work. References to
older studies on DNA-adduct measurements may need to be reanalyzed in light of the most
recent analytic techniques that achieved superior sensitivity (for example, Lu et al. 2010).
In particular, the committee finds the recent study of Lu et al. (2010) to be highly
informative and the first one to distinguish clearly between exogenous and endogenous
formaldehyde-induced DNA adducts. Although the study does not challenge the notion that
DNA adducts play only a minor, if any, role in formaldehyde genotoxicity and
carcinogenicity, compared with DNA-protein cross links, it adds to the evidence of the
inability of formaldehyde to reach distant sites. Likewise, the positive study by Wang et al.
(2009) is not adequately described in the draft IRIS assessment, nor is it clear to the
committee why so much emphasis is placed on the study by Craft et al. (1987) (pp. x and 45
[mode of action]).
• Response: The studies by Lu et al. (2010), Wangetal. (2009), and Craft et al. (1987) are
described and evaluated in the current draft (see Section 3.X) and strengths and limitations
are clearly presented. EPA updated the literature annually and all relevant studies are
included in this draft
• Dosimetry modeling of formaldehyde
• The CFD models were fairly evaluated and that the sources of uncertainty in dose metrics
used in dose-response assessments were appropriately treated, [pp 31]
• The committee disagrees with EPA's findings that CFD models are not useful for low-dose
extrapolations. In fact, flux results from the CFD models can easily be scaled from an
exposure of 1 ppm—as given by Kimbell et al. (2001a,b) and Overton etal. (2001)—to
lower concentrations because of the linear flux-concentration relationship that was used by
the authors. Therefore, the committee recommends that the CFD-based approach also be
used to extrapolate to low concentrations, that the results be included in the overall
evaluation, and that EPA explain clearly its use of CFD modeling approaches (p. 31).
• Response: EPA agrees with the committee that "flux results from the CFD models can easily
be scaled from an exposure of 1.0 ppm to lower concentrations because of the linear flux-
concentration relationship that was used by the authors of the model," and has used this
approach in the assessment As explained further in response to questions on EPA's use of
BBDR modeling, the assessment presents rat and human risk estimates based on the BBDR
modeling. This modeling used CFD model calculations as input. Because these BBDR-
predicted values differ from each other by many orders of magnitude, EPA's calculation of
unit risk is based on straight line extrapolation from points of departure, derived using
different implementations of the BBDR model in the rat. Extrapolation to the human is then
based on CFD model-derived wall-mass flux estimates in the rat and human nose.
• The committee notes that the CFD models of Kimbell etal. (2001a,b) do not account for
potential effects of sensory irritation on ventilation inasmuch as only two mass-transfer
coefficients, one for mucus-coated and one for non-mucus-coated epithelial regions of the
nose, were used in all simulations to derive uptake into nasal tissues. However, later
This document is a draft for review purposes only and does not constitute Agency policy.
D-9 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
models that account for DPC cross links and cytotoxicity (Conolly et al. 2000, 2002, 2003,
2004; Georgieva etal. 2003) relied on animal data that were obtained at concentrations that
potentially caused irritation to derive parameters associated with metabolism and
reactivity; thus, the potential effect of altered ventilation was indirectly compensated for in
those model simulations.
• Response: EPA agrees with the committee. The statement on uncertainty in model (BBDR
and DPC) structure associated with effects of sensory irritation on ventilation has been
deleted from the current draft assessment
• The draft IRIS assessment raises the criticism that the nasal CFD models are based on a
single geometry for each species. Thus, the models do not address variability that arises
from differences in airway anatomy. A recent paper by Garcia et al. (2009) evaluated the
effect of individual differences in airway geometry on airflow and uptake of reactive gases,
such as formaldehyde. Although the sample was small (five adults and two children), the
individual differences in airway geometry alone caused the potential flux rates to vary by a
factor of only 1.6 over the entire nose and by a factor of 3-5 at various distances along the
septal axis of the nose. The committee agrees with EPA that although the sample was small,
the estimates of individual variability are consistent with default uncertainty factors applied
to internal dose metrics that account for human variability.
• Response: For noncancer effects, EPA has used an uncertainty factor to address human
variability. For cancer effects, EPA does not apply uncertainty factors for intrahuman
variability but recognizes that there is uncertainty in estimates of unit risk.
• Biology-based dose-response (BBDR) modeling of rat nasal tumors
• 4.1 The committee agrees that [EPA's] sensitivity analysis added value to the
interpretation of the Conolly et al. models (p. 36). The committee also acknowledges that
the draft IRIS assessment provides a thorough review of the BBDR models, the major
assumptions underpinning the extrapolation to humans, and EPA's own series of papers
that evaluated the sensitivity of the BBDR models to these assumptions even though the
committee may not agree with the validity of all the resulting manipulations (p. 42).
0.80 EPA's reanalysis was consistent with its cancer guidelines that specify that the uncertainties and
variability in model parameters must be understood and articulated so that predictions of adverse
responses and extrapolations to human exposures can be appropriately characterized from the
standpoint of human health protection (p. 36).
• Response: The revised draft assessment includes such sensitivity analyses which yield risk
estimates in the range of observed data that is consistent with the observed data but
illustrate the wide variation in potential estimates at low doses if the model is extended far
below the observable data.
• 4.1 The committee questions the degree to which manipulations of the range of model
parameter values can and should be performed to reflect potentially divergent outcomes (p.
36). The committee is concerned about the possibility that those adjustments of the Conolly
et al. models may not be scientifically defensible (p. 43).
This document is a draft for review purposes only and does not constitute Agency policy.
D-10 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Supplemental Information for Formaldehyde—Inhalation
• EPA, on the basis of extreme alternative model scenarios, chose not to use the BBDR models
developed by Conolly et al. (2003, 2004); however, the committee questions the validity of
some of these scenarios (p. 44).
• The NAS committee raises the concern that "because Crump et al. (2008) argue that there
are no data to refute these assumed and arbitrary adjustments of the Conolly et al. models,
they state that the onus is on others to show that such small changes cannot occur (that is,
prove a negative before the authors would accept the contention that the Conolly et al.
models are at all conservative as Conolly et al. suggested). That standard cannot be met" (p.
40).
• Response: In a sensitivity analysis, one makes small changes to the inputs or assumptions
in a model and observes the changes in the output. The purpose of such an analysis, as
recommended by the cancer guidelines, is to establish that predictions from the BBDR
model are robust These changes should be small enough to be consistent with the data
used to develop the model and biological constraints imposed on the model inputs and
assumptions. EPA's sensitivity analyses presented in this assessment draft adhere
rigorously to this requirement. In particular, in the context of model treatment of initiated
cells (the focus of the above NAS comment) EPA's sensitivity analyses are based on
extremely small variations to the initiated cell division rates assumed in the original model.
These variations, as presented in the revised draft assessment, are smaller by an order of
magnitude than those carried out in Crump et al. (2008). The calculations were constrained
to satisfy the conditions (as in Conolly et al., 2004) that model predictions provide good fits
to: a) the formaldehyde combined bioassay tumor incidence data (Kerns et al., 1983;
Monticello et al. 1996) and b) the background rates of respiratory cancers in humans
obtained from the SEER database. Furthermore, it was ascertained that the ratio of initiated
cell division rate to initiated cell death rate was very close to the value of one for any
variations in parameter values in the sensitivity analyses.
• There are no empirical data on division rates for these initiated cells; thus these values were
assumed in the original model. Therefore, in order to provide perspective on the variations
in the division rates of initiated cells that were used for the purpose of the sensitivity
analysis, the revised draft assessment compares them with the empirical variability in
normal cell division rates. For example, the maximum change in log-transformed value for
the analysis labeled mod4 in Figure 2-9 was about 1 /35 th of the standard deviation in the
log-transformed values of the empirically determined normal cell division rate. These
issues are addressed in item vi of the subsection on "uncertainties in BBDR modeling
components" in "BBDR modeling for extrapolation of SCC risk. EPA believes the sensitivity
analysis variations in this revised assessment are consistent with the available data and
biological constraints.
• 4.1 In particular, adjustments of parameter values associated with mutation, birth, and
death rates of initiated cells used in EPA's analysis of alternative models that yielded the
most extreme deviations from the Conolly et al. (2004) low-dose extrapolations also
produced unrealistically high added risks for humans at concentrations that have been
observed in the environment of occupationally exposed workers (100% incidence at
concentrations as low as about 0.1-1 ppm). Thus, the committee recommends that
manipulations of model parameters that yield results that are biologically implausible or
inconsistent with the available data be discarded and not used as a basis for rejecting the
overall model (p. 42).
This document is a draft for review purposes only and does not constitute Agency policy.
D-ll DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Supplemental Information for Formaldehyde—Inhalation
• Response: EPA's revised assessment provides more refined sensitivity analyses. Fig. xxx is
added in response to the above NAS question regarding values resulting from these
analyses. This figure compares values for lifetime human MLE risk estimates between the
values resulting from: 1) EPA's analysis of epidemiological data on nasopharyngeal cancers
(NPC) from the National Cancer Institute (NCI) cohort study of workers occupationally
exposed to formaldehyde, 2) the original Conolly etal. (2004) model for squamous cell
carcinoma in humans as extrapolated from the F344 ratbioassays, and 3) EPA's sensitivity
analyses of that model. In order to do so, the figure highlights values corresponding to an
exposure concentration of 0.15 ppm, the level corresponding to the ECooos derived from the
occupational epidemiology data. At 0.15 ppm, the estimated lifetime added risk estimates
are: +5x10-4 (from epidemiology data), -1x10-3 (Conolly 2004), +9x10-4 (from one model
variant), as well as values ranging from -2x10-3 to +3x10-3 (from other variants). Note that
some of these values are negative because the model estimated risk estimates that were
below baseline. Thus, the sensitivity analyses in the assessment shows that the original
model and its variants, arising from extremely small variations in values of the unknown
initiated cell replication rates used in the original model, result in values that range from
being many orders of magnitude different from, to substantially in agreement with, the
lifetime risks projected from the epidemiology data. These model variations all adhere to
the same biological constraints and provide similar fits to the tumor incidence data when
used in the rat SCC model.
• 4.1 In contrast, Conolly et al. (2003) focused their model parameter estimates to
represent "best-fit," using maximum likelihood estimates, whereas Subramaniam et al. and
Crump et al. pushed parameter assumptions in a single direction to show that different
assumptions that fit the experimental data can yield different results of low-dose
extrapolation (p. 43).
• Conolly and co-workers felt that they made several conservative assumptions in their
models—use of hockey-stick rather than J-shaped models for cell proliferation, use of
overall respiratory tract cancer incidence in humans to calculate basal mutation rates, and
use of an upper bound on the proportionality parameter relating DPC to mutation. EPA
pushed that concept further by making even more conservative assumptions within the
models that cumulatively resulted in radical departures from the results of the Conolly et al.
models with regard to low-dose extrapolation of tumor incidence. The committee notes
that EPA forced changes in the model parameter values in a direction that yielded more
conservative results rather than one that yielded a best fit to the data (p. 43).
• Response: EPA considered central estimates of input parameters. As the NAS supported in
Comment 4.1 above, the current draft assessment also appropriately examines
uncertainties in the inputs and the sensitivity of modelling results to assumptions. For
some modeling assumptions, there is no specific data from which to select a central
estimate or maximum likelihood and EPA evaluates whether the model is sensitive to the
choice of assumptions. EPA's analysis evaluates a continuous range of minor perturbations
to the original formaldehyde model that are all equally consistent with the data used in
developing the model. Resulting risk estimates are both above and below (i.e., vary in both
directions from) that obtained in Conolly et al. (2004). The risk estimates from some of the
model implementations in the new Figure 2-9 in the revised draft are obtained even
without making conservative assumptions or calculating an upper bound; all these models
retained the J shape for the dose response for normal and initiated cell replication. EPA's
sensitivity analysis does not necessarily yield conservative results; risk estimates
This document is a draft for review purposes only and does not constitute Agency policy.
D-12 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
substantially below background levels of human risk are obtained from some variations in
the division rates for initiated cells that are used in the sensitivity analyses. Thus, the
analyses are not constrained to push the model output in a single direction.
• 4.1 The committee was also struck by the relative lack of transparency in the draft IRIS
assessment's description of the decision to use the peer-reviewed BBDR models minimally
(p. 43).
• As a result of the agency's reanalysis of the models, EPA chose not to use the full rat and
human BBDR models to estimate unit risks. Instead, in a benchmark-dose approach, EPA
used the CFD-derived determinations of formaldehyde flux to the entire surface of mucus-
coated epithelium to derive a point of departure based on nasal cancers in rats. It then
extrapolated to zero dose by using a default linearized multistage approach. The committee
is concerned about that approach for low-dose extrapolation. The committee found that the
evaluations of the original models and EPA's reanalysis conflicted with respect to the intent
or purpose of using the formaldehyde BBDR models in human health assessments (p. 43).
• The primary purposes of a BBDR model are to predict as accurately as possible a response
to a given exposure, to provide a rational framework for extrapolations outside the range of
experimental data (that is, across doses, species, and exposure routes), and to assess the
effect of variability and uncertainty on model parameters (p. 5).
• Response: EPA's revised draft has improved transparency in regards to its use of the BBDR
model and its evaluation for low-dose extrapolation. Because the BBDR modeling
integrates various mechanistic information (DNA-protein cross links, cell labeling index
measurements, computational fluid dynamic modeling of formaldehyde flux to the nasal
lining) and time-to-tumor data from individual animals in the tumor bioassay, EPA's revised
draft uses it for multiple purposes through the assessment Firstly, the model is used to
predict risk in the range of the observed rat data (in fact, slightly below the range, allowing
for a benchmark response of 0.005, so that the point of departure is just below the dose
where a change in the curvature of the dose response occurs). Secondly, the BBDR model
provides some perspective on the shape of the dose response used for low-dose
extrapolation. Dose-response curves (shown in the assessment) from the Conolly et al.
model and from the variants constructed for the sensitivity analyses, all exhibit linearity
below roughly 0.05 ppm, and the value of the low dose slope of one such curve is consistent
with that derived from EPA's analysis of epidemiological data on nasopharyngeal cancers
(NPC) from the National Cancer Institute (NCI) cohort study of workers occupationally
exposed to formaldehyde. However no particular value from these BBDR-derived curves
can be selected because of the large variability in values. Thirdly, the BBDR modeling shows
that formaldehyde's mutagenic action could potentially play a significant role in explaining
the observed tumor data as well as its predicted low-dose carcinogenicity, lending support
to using a linear low-dose extrapolation below the observed data. Fourthly, computational
fluid dynamic modeling of formaldehyde flux to the nasal lining, an element in the BBDR
modeling, is used in deriving a candidate reference dose for squamous metaplasia observed
in F344 rats.
• 4.1 Given that the BBDR model for formaldehyde is one of the best-developed BBDR
models to date, the positive attributes of BBDR models generally, and the limitations of the
human data, the committee recommends that EPA use the BBDR model for formaldehyde in
This document is a draft for review purposes only and does not constitute Agency policy.
D-13 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
its cancer assessment, compare the results with those described in the draft assessment,
and discuss the strengths and weaknesses of each approach (p. 5).
• A biologically based dose-response (BBDR) model that has been developed for
formaldehyde could be used in the derivation of the unit risk estimates. EPA explored the
uncertainties associated with the model and sensitivities of various model components to
changes in key parameters and assumptions and, on the basis of those extrapolations,
decided not to use the BBDR model in its assessment (p. 5).
• Response: EPA's revised draft assessment does use two formulations of the BBDR model to
estimate points of departure from the animal nasal cancer data, and to illustrate the
uncertainties that arise in using these and other models for low-dose risk estimation. EPA
clearly explains why it chose to use linear low-dose extrapolation to derive estimates of
reasonable upper-bound on risk at lower doses. The revised assessment also explains why
its preferred estimates of human nasal cancer risks from formaldehyde are derived from
the human epidemiology data rather than from extrapolations of the animal study data. As
explained in response to 4.5, EPA's revised draft uses the BBDR model for multiple
purposes, qualitative and quantitative.
• Comparison of human risk estimates: As recommended by the NAS, it is useful to contrast
lifetime human risk estimates for cancer in the human respiratory tract from the
formaldehyde BBDR model with other estimates. This is shown in Figure 2-9. In this figure,
the epidemiology-based ECooos and LECooos are the maximum likelihood estimate (MLE) and
95% lower confidence bound, respectively, for the continuous exposure level of
formaldehyde that would correspond to a lifetime extra risk of NPC of 0.0005, and the curve
labeled "Lin. Extrap. LEC0005" is the straight line extrapolation drawn from the LECooos-
The dose-response curve obtained from fitting a time-to-tumor model (the multistage-
Weibull) to the cancer bioassay data where extrapolation was based on average
formaldehyde flux to the nasal tissue as dose-metric is also shown.
• Robustness of models: As discussed in the response to Comment 4.5, models used to
estimate human risk must be robust. EPA evaluation shows the human BBDR model to be
numerically unstable on two accounts. EPA has considered very small perturbations of the
dose response for the division rate of initiated cells that was assumed in the original model.
Risk estimates corresponding to a continuous range of perturbations to the original
formaldehyde model, all equally consistent with the data used in developing the model,
span a large continuous range. This range includes values that may be consistent with
human epidemiology as well as very large values and values that are substantially below
background levels of human risk (see Figure 2-9 and surrounding text). For example, the
small perturbation represented by the curve labeled mod4 in Figure 2-9 increased the
estimate of extra risk at 0.15 ppm from -0.001 (the MLE value obtained by Conolly et al.
2004) to roughly + 0.001.
• The second source of numerical instability was the input used for cancer rates in control
animals. The Conolly et al. (2003, 2004) analysis included 7,684 historical control animals
drawn from all National Toxicology Program (NTP) bioassays in addition to the concurrent
control animals. Crump et al. (2008) explored the impact of uncertainties in this usage of
historical control data. When the BBDR model for the rat was run using incidence data from
control animals in only NTP inhalation bioassays added to the incidence data from the
concurrent control animals, human risk estimates from the corresponding human BBDR
This document is a draft for review purposes only and does not constitute Agency policy.
D-14 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
model used for extrapolation were 50-fold higher; when only concurrent control animals
were used (without any historical controls added), human risk estimates could not be
bounded (Crump etal., 2008; Subramaniam etal., 2007). Even in the former case (where
the historical control data was restricted to inhalation bioassays), the human model was
prevented from becoming unstable by a positive tumor incidence in just one animal (Crump
etal., 2008).
• To sum, EPA finds that the Conolly etal. human BBDR model is not robust and therefore
cannot be used to constrain human risk at any exposure concentration. However, EPA did
use BBDR modeling for human extrapolation of nasal cancer risk observed in the rat as
explained below.
• Use of BBDR modeling of nasal cancer risk in the rat: EPA has evaluated the impact of the
uncertainty and variability in the data and assumptions used in the BBDR model developed
for modeling nasal cancer risk in the F344 rat, and has used the evaluations quantitatively
in its dose-response assessment. Given the data, multiple implementations of the model,
including the modeling in Conolly et al. (2003), can be judged to be just as biologically
plausible as the other. Each of the models describe the rat tumor incidence equally well, is
based on different characterizations of the same empirical cell kinetic data, and is based on
the same empirical data on DPC measurements. However, when extrapolated below the
range of observable data, these BBDR models result in risk estimates that vary by many
orders of magnitude. For example, at the 10 ppb (0.01 ppm) concentration, MLE risks range
from -4.0 xl0~6 to +1.3 x 10"7. Atthis dose, models that gave only positive risks result in a
five orders of magnitude risk range from 1.2 xlO"12 to 1.3 xlO"7. Furthermore, EPA finds
model uncertainty to be substantially higher than the statistical uncertainty arising out of a
given model specification (see Appendix F). Thus, BBDR modeling could not be used to
reasonably constrain nasal cancer risk estimates for the F344 rat when extrapolated below
the range of observable data.
• Use of rat BBDR model to estimate PODs: EPA has used the BBDR modeling to calculate
points of departure (PODs) for quantifying cancer risk. Because model uncertainty is
significant, two different implementations of the rat BBDR model are used to reflect
uncertainty in calculating the POD. These PODs are based on formaldehyde flux to the
tissue as an internal dose-metric calculated from fluid dynamic modeling of airflow and
formaldehyde uptake in anatomically realistic representations of the upper respiratory
tract. Extrapolation of these values to the human is also based on formaldehyde flux to the
tissue using fluid dynamic modeling, but in this case for both the upper and lower
respiratory tract The use of BBDR modeling provides greater support for using a POD at
the 0.5% response level (0.005 extra risk). Typically, the BMD is calculated at the 5% or
10% response level. In the case of data combined from the Kerns et al. (1983) and
Monticello et al. (1996) bioassays, the lowest observed tumor incidence of SCC is below the
1% level (at 0.85%). Additionally, the BBDR modeling incorporates a precursor response in
the form of labeling index data. Therefore, it is appropriate to evaluate the POD at the 0.5%
level while still staying in the neighborhood of the experimentally observed response.
• 4.1 The committee is also concerned that EPA directed substantial effort toward
refuting many of the assumptions and conclusions of the Conolly et al. (2003, 2004) models
rather than trying to fill the data gaps that were clearly articulated by the models. Conolly
and co-workers were clear on that point and expressed the need for new data that could
This document is a draft for review purposes only and does not constitute Agency policy.
D-15 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
anchor many of the parameter values that had to be optimized from rather sparse data sets
(p. 44).
• Response: EPA agrees that the formaldehyde BBDR model has helped identify data gaps. A
large data gap identified by EPA is information on division rates of initiated cells in the
respiratory tract As suggested by the NAS such information can be used to anchor
uncertain parameter values. Similar efforts have been directed in the area of modeling liver
cancers to inform the health risk assessments for dioxin and other chemicals. In those
cases, data on foci or nodules36 have been used to estimate rates of initiation and
proliferation, under the assumption that they are preneoplastic lesions. However, such foci
or nodules have not been identified in the case of nasal cancer. As acknowledged by the
NAS, assuming that initiated cells related to tumors in the respiratory tract can be
identified, measurement of their division rates would be extremely difficult Even if this
difficulty were to be surmounted, it is reasonable to suppose that these rates would be at
least as variable as division rates of normal cells. Based on the normal variation in such
rates observed in normal cells (see Figure A-3), and the extreme sensitivity of the
formaldehyde model to small differences in assumed division rates of initiated cells, EPA
concludes that it would be impossible to measure these accurately enough to restrict the
range of risks consistent with the model sufficiently to be useful for practical risk
assessment needs. In the case of preneoplastic foci in the liver, it has not been possible to
confidently decide which cells in foci or nodules represent initiated cells or even whether
the model formulation is correct for those foci (Kopp-Schneider et al., 1998). Quantitative
estimates of risk can be very sensitive to these choices.
• 4.1 EPA's rationale for use of a low-dose linear extrapolation (through zero dose) is the
observed linear relationship between DPC and exposure. The committee evaluated the
strength of this rationale on the basis of [differences in] model predictions in Conolly et al.
(2003) and Subramaniam et al. (2007) for the value of the constant of proportionality
relating DPC to the probability of mutation in the BBDR modeling. However, the committee
had low confidence in deciding which of these approaches was the most scientifically
defensible because too few parameters were experimentally fixed and too many optimized
against one data set [in either case],
• The current parameter estimates that Conolly et al. (2003) optimized from the data, using a
maximum likelihood function, suggest that the proportionality constant for DPC adding to
the mutation rate of a normal (or intermediate) cell should be zero or close to zero. That
suggests that DPC is not directly related to the key events leading to mutation and
carcinogenicity per se. Because this [i.e., mutagenic potential being proportional to DPC
burden] is the only low-dose linear relationship between exposure and a biomarker of
response, EPA contends that the low-dose extrapolations should be linear through zero
dose. For example, Subramaniam et al. (2007) examined alternative choices to parameters
associated with DPC clearance and suggested that in the exposures at which tumors were
seen, the mutagenic mode of action could contribute up to 74% of the added tumor
probability. Because too few parameters were experimentally fixed and too many
optimized against one data set, confidence in deciding whether the Conolly et al. or the
Subramaniam et al. approach is the most scientifically defensible is not high (p. 39).
6To our knowledge, no such preneoplastic foci have been seen for squamous cell carcinomas.
This document is a draft for review purposes only and does not constitute Agency policy.
D-16 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
• Response: EPA is assuming that the NAS comment on low-dose extrapolation refers to
extrapolating the risk of nasal tumors from the rat to human. We agree with the
committee's conclusion that neither the Subramaniam et al. (2007) nor the Conolly et al.
(2004) analyses should be used as the basis for making a mode of action determination.
EPA's decision to use a linear extrapolation to the origin from a point of departure was
based only on the following two considerations: 1) that the BBDR models did not constrain
estimates of human respiratory cancer risk at any exposure concentration, and did not
constrain estimates of rat nasal cancer risk at exposure concentrations below the observed
data in the rat (see response to Comment 4.6) and 2) EPA's determination, based on
multiple sources of data in humans and animals, of a mutagenic contribution to
formaldehyde's carcinogenic potential in the upper respiratory tract of exposed humans
(see Section XX of the document).
• Subramaniam et al. (2007) did not attempt to determine the most appropriate low-dose
relationship. Rather, their analysis, and the use of their results in EPA's draft assessment,
expresses the uncertainty in the assertion in Conolly et al. that formaldehyde's
mutagenicity, as per their model conclusions, did not play a role in its carcinogenicity. The
current draft assessment further clarifies this point of view.
• 4.1 The reanalysis by Subramaniam et al. is used to support the mutagenic mode of
action of formaldehyde and to reduce support for using the BBDR models on the basis of the
uncertainties in parameter estimation and assumptions in the models (p. 43).
• Response: The determination that formaldehyde's direct mutagenic action contributes to
its carcinogenicity in humans was based on multiple sources of data in humans and
laboratory animals. These are detailed in Section 1.X.X (URT cancer MOA) of the
assessment. The analyses in Subramaniam et al. and in other BBDR model implementations
pursued in the draft assessment were partly used to evaluate the uncertainty in an
inference on mode of action made by Conolly et al. Based on BBDR modeling results, these
authors inferred that formaldehyde's mutagenicity did not play a role in its carcinogenicity.
EPA's uncertainty analyses of the BBDR modeling determined that such an inference was
extremely uncertain. To be clear, in some alternate model implementations EPA estimated
parameter values that were consistent with a significant role for formaldehyde's putative
mutagenic action in explaining its tumorigenicity, but these results were not the basis upon
which EPA concluded that there was sufficient weight of evidence for a mutagenic MOA for
upper respiratory tract cancers. The current draft assessment makes this very clear.
• 4.1 Because multiple modes of action may be operational, the committee recommends
that EPA provide additional calculations that factor in regenerative cellular proliferation as
a mode of action, compare the results with those presented in the draft assessment, and
assess the strengths and weaknesses of each approach, (pp. 5) Although the draft IRIS
assessment discusses that [regenerative cell proliferation associated with cytotoxicity]
mode of action, it relies on the mutagenic mode of action to justify low-dose extrapolations.
The committee recommends that EPA provide alternative calculations that factor in
nonlinearities associated with the cytotoxicity compensatory cell proliferation mode of
action and assess the strengths and weaknesses of each approach (p.44).
• Response: Because multiple modes of action are operational, EPA's assessment uses BBDR
modeling that factors in the empirical regenerative cellular proliferation data, thus,
inherently including the nonlinearity to which the above comment points, as well as the
This document is a draft for review purposes only and does not constitute Agency policy.
D-17 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Supplemental Information for Formaldehyde—Inhalation
DNA protein cross-link data representing formaldehyde's directly mutagenic potential. The
cancer slope factors derived in the assessment from the animal nasal cancer data are
consistent with the predictions of the BBDR modeling. The revised assessment also
compares with the BMDLoi derived exclusively from regenerative cell proliferation by
Schlosser et al. (2003). These authors fitted a curve with a threshold in dose to the
exposure time-weighted average (over the entire nose) of the unit length labeling index
data from Monticello etal. (1991,1996). While these points of departure are in agreement
with each other, the BBDR modeling points to significant risk below the presumed
threshold in Schlosser et al.
• The revised assessment also notes that, because the BBDR modeling estimates the constant
of proportionality relating DPC levels to formaldehyde-induced mutation by fitting to the
steeply rising tumor incidence data, EPA's uncertainty analysis of results derived from the
modeling reflects [model] uncertainty associated with a putative mutagenic mode of action
(as an explanation for formaldehyde tumorigenicity).
• 4.1 The committee agrees with EPA that existing data are insufficient to establish the
potential biologic variability in model parameters associated with the mutagenic mode of
action adequately. However, because the mutagenic mode of action is the major reason for
adopting the default low-dose linear extrapolation methods over application of the BBDR
models in the draft assessment, the committee recommends that the manipulations that
lead to such high contributions of mutagenicity to the mode of action for nasal tumors be
reconciled with the observations that formaldehyde is endogenous, that nasal tumors are
very rare in both rats and humans, and that no increases in tumor frequency have been
observed in animal studies at formaldehyde exposure concentrations that do not also cause
cytotoxicity (p. 42).
• Response: EPA agrees with the NAS that there are no data to directly establish the
variability or uncertainty in key unknown model parameters. The EPA cancer guidelines
note that unless there is an established mode of action known to be inconsistent with a
linear estimate of upper-bound risk at low doses, it is EPA's practice to use a linear
approach to estimating an upper-bound on the low-dose risk. That cancers may be due to a
mutagenic mode of action is one rationale for that policy. But, dose-response functions for a
human population may also be approximately linear at low doses due to other factors
including the effect of variation in human responses, as was noted in the NAS report on
Science and Decisions [cite]. The assessment notes that the assessment is evaluating the
extra risk associated with inhaled formaldehyde adding to endogenous concentrations in
nasal tissues and is not estimating the risk associated with the endogenous formaldehyde
concentration. The revised assessment draft concludes that the background rates of nasal
cancers and the background cellular concentration of endogenous formaldehyde are not
inconsistent with the draft assessments estimates of the extra risk associated with
difference inhaled doses of formaldehyde.
• EPA has examined the range of risk estimates obtained when using the BBDR modeling
approach in Conolly et al. for extrapolation in a manner that reflects uncertainty and
variability. This approach is not constrained to assuming a mutagenic mode of action, and
incorporates data related to formaldehyde mutagenicity as well as formaldehyde's effect on
cell proliferation. This course of action follows NAS advice presented as Comment 4.1. As
explained earlier, the range in risk estimates resulting from the BBDR modeling is so large
that low-dose risk cannot be constrained in either the rat or the human. Thus, given the
This document is a draft for review purposes only and does not constitute Agency policy.
D-18 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
uncertainty, it appears reasonable to use a linear extrapolation from a point of departure
estimated using the BBDR modeling (and more than one point of departure was determined
to reflect model uncertainty). EPA also verified (as seen from Figure 2-9), that linear
extrapolation is not inconsistent with the large range of risk estimates predicted if the
BBDR modeling were to be used below the POD.
• It is important to note that the model predicts extra risk (over baseline levels) due to
inhaled exogenous concentrations of formaldehyde. EPA's uncertainty analyses with the rat
formaldehyde BBDR model include the observation of tumors in historical control animals
from NTP inhalation bioassays. Therefore these model implementations were calibrated to
predict the observed levels of spontaneous tumor incidence. Thus, these predictions are
presumably consistent with contributions to baseline risk [if any] arising from endogenous
levels of formaldehyde. The rarity of squamous cell carcinoma in rats is appropriately
accounted for by the inclusion of historical control animals from inhalation bioassays. The
alternate model implementations and the perturbations considered in initiated cell
replication rates were all constrained to reproduce the tumor incidence data. Specifically,
model fits to the time-to-tumor data in all cases were equivalent In other words, all these
results were consistent with no increases in observed tumor frequency in animal studies at
subcytotoxic formaldehyde exposure concentrations.
• 4.1 Crump et al. (2008) made an arbitrary change in the DPX-based effect on initiated
cell replication by theorizing that if an initiated cell is created by a specific mutation that
impairs cell-cycle control, there may be a mitigation of cell replication that is observed in
the low-dose cell proliferation of normal cells (that is, in the negative vs baseline replication
portion of the J-shaped dose-response curve) and hence a shift of the cell division of an
initiated cell in the model toward greater rates at low doses (p. 40).
• The change disconnects the birth and death rates of initiated cells from constraints used by
Conolly et al. based on normal cells. The committee concludes that this change is contrary
to the explanation provided by Monticello et al. (1996), who suggested that it is not a
mutation in cell-cycle check points that results in lower cell-division rates than control at
low exposures but rather an increase in the time that it takes for DNA-repair processes to
eliminate the DPX before the cell can resume the process of cell division that leads to lower
than basal cell-division rates at low exposures. These are two fundamentally different
mechanisms with different connotations for risk—the mutagenic one chosen by EPA and
the DNA-repair mode of action supported by several other publications on DPX cited by
Conolly etal. (2003, 2004) and Monticello etal. (1996) (p. 40).
• Response: The revised assessment does not rely upon the mechanistic hypothesis put
forward in Crump et al. (2008) for what might cause cell-division rates to be lower than
control at low exposures. (EPA has removed speculation as to how minor differences
between initiated and other cells could arise.)
• The revised assessment explains that small potential differences in the division rates of
initiated cells examined in the sensitivity analysis are illustrative that, as the NAS comment
[#] notes, the biological data are not available to directly determine whether initiated cells
have the same or different division rates as uninitiated cells. The perturbations considered
in the sensitivity analyses in the current draft EPA assessment are substantially smaller
than in Crump et al., and are only applied to the J-shaped dose response for cell replication
in the original model. Any mechanistic arguments that one might associate with a J-shaped
This document is a draft for review purposes only and does not constitute Agency policy.
D-19 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
curve for a dose-response relationship for cell replication should apply with equal force to
the J-shaped curves in Figure $$$.
• 4.1 There were zero squamous cell carcinomas in control rats in the two bioassays used
to define the basal mutation rates of normal and intermediate cells in the two-stage, MVK
dose-response model. Conolly et al. (2004) used results from the full National Toxicology
Program historical control database. That is a point of contention by EPA, which believes
that only historical controls from inhalation bioassays (and those in the same laboratory as
the formaldehyde study) can be used in a relevant comparison. Squamous cell carcinomas
are so rare that some leeway in approximating basal rates may have to be accepted, even
though EPA's point is technically correct (p. 40).
• Response: EPA agrees. The rarity of squamous cell carcinoma in rats is appropriately
accounted for by the inclusion of historical control animals from inhalation bioassays in
EPA's uncertainty analyses. Given the reactivity of formaldehyde, to allow for a reasonable
comparison it is considered essential that studies used the same route of exposure; as such,
noninhalation studies were not included in the current analyses.
• 4.1 Estimating parameters for basal mutation rates for a normal to intermediate and
intermediate to malignant transformation in humans is subject to even more uncertainty
than in the rat
• Response: EPA agrees, and has included this in additional uncertainties associated with the
formaldehyde human model.
• 4.1 The first-order clearance of DPX could be slower than that used by Conolly et al.
(2003, 2004). Over time, epithelial tissue in targeted regions of the nose thickens. The
thickening could conceivably dilute DPX concentrations in the measured tissues to such an
extent that residual concentrations 18 hr after exposure are not different from those in
naive animals, and this would affect the determination of DPX clearance rates (pp. 41).
• Response: The revised assessment discusses the uncertainty in clearance rates of DPC and
its impact on model calibration.
• Health endpoints
• Overall, the committee found that the noted outcomes were appropriate to evaluate. EPA
identified relevant studies for its assessment, and on the basis of the committee's familiarity
with the scientific literature, it does not appear to have overlooked any important study.
For a few outcomes, however, as noted below, EPA did not discuss or evaluate literature on
mode of action that could have supported its conclusions. Although EPA adequately
described the studies, critical evaluations of the strengths and weaknesses of the studies
were generally deficient, and clear rationales for many conclusions were not provided. In
several cases, the committee would not have advanced a particular study or would have
advanced other studies to calculate the candidate RfCs (p. 6).
• Irritation
• The committee notes that EPA did not (but should) review research findings on transient-
receptor-potential ion channels and evaluate the use of this evidence for improving
This document is a draft for review purposes only and does not constitute Agency policy.
D-20 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
understanding of the mode of action for sensory irritation and respiratory effects attributed
to formaldehyde exposure (p. 6; and list at end of Chapter P 52).
• Response: EPA agrees with this recommendation and discusses involvement of transient-
receptor-potential ion channels in a more comprehensive MOA discussion for noncancer
respiratory tract-related effects, including sensory irritation (see Section l.X).
• Although the chamber studies are of acute duration, they are complementary with the
residential studies and provide controlled measures of exposure and response. Therefore,
the committee recommends that EPA present the concentration response data from the
occupational, chamber, and residential studies on the same graph and include the point
estimate and measures of variability in the exposure concentrations and responses (p. 6;
also in list at end of the chapter, pp. 52-53).
• Response: EPA agrees with this recommendation and presents the dose-response results
from the literature in graphical form in Section l.X. The prevalence of eye irritation (and
standard errors) reported by the studies of residential populations and controlled human
exposure studies are plotted on the same graph in the range of formaldehyde
concentrations that are common to both (0-1 mg/m3). Because the controlled human
exposure studies examined symptoms at higher concentrations as well, an additional graph
that includes all of the data also is included. The results of the occupational studies on
irritation symptoms are complementary, but the variation in exposure levels in the exposed
groups in these settings was large, and only the mean response in relation to the mean
concentration in the entire exposed group was presented and compared to a referent group.
These data were less informative compared to the exposure-response information from the
residential or controlled human exposure studies.
• The committee found that EPA dismissed the results of the exposure chamber and other
nonresidential studies too readily. Although the exposure durations for the chamber
studies are short relative to the chronic duration of the RfC, the studies provide
complimentary information that could be used for deriving a candidate RfC (also in list at
end of the chapter on p. 52).
• Response: EPA agrees that the controlled human exposure studies provide complimentary
information and relies on these studies in concert with the occupational and residential
studies to establish formaldehyde as an irritant In accordance with the criteria for
selecting studies for the derivation of candidate RfCs (see Table l.X), EPA uses the dose-
response information from epidemiology studies of residential exposure because studies of
good quality are available (Hanrahan et al., 1984; Liu et al., 1991) and compares these to
cRfCs derived from medium confidence controlled human exposure studies (Kulle et al.,
1983; Andersen, 1983).
• 1.1.1.1 The committee agrees with EPA's selection of eye irritation as a critical sensory-
irritation effect caused by formaldehyde exposure because residential, occupational, and
chamber studies have demonstrated that the eyes are more sensitive to irritation from
formaldehyde than the nose and throat.
• Response: EPA agrees that irritant effects on the eye are a sensitive response to
formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
D-21 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
• The committee supports EPA's advancement of the residential studies by Liu et al. (1991)
and Hanrahan et al. (1984) for derivation of candidate RfCs as adequately conducted studies
of a randomly selected general population and agrees with the points of departure
identified by EPA from these studies:
• LOAEL = 95 ppb (Liu et al. 1991)
• BMCL10 = 70 ppb (Hanrahan et al. 1984)
• Response: These two studies are included among those for which candidate RfCs were
considered. Although the results from Liu et al. (1991) were not used to derive a cRfC, the
data can be used to check the estimated POD based on Hanrahan et al. (1984).
• Chapter 4: The committee recommends that EPA address the following in the revision of the
formaldehyde draft IRIS assessment
• Strengthen its critical evaluation of the studies.
• Response: In the current draft assessment, the studies are described in tables or
graphically categorized according to confidence in the study results determined by
systematic evaluation of risk of bias and sensitivity. The contribution of the studies to the
hazard assessment and the strengths and limitations of the studies are documented in
supplemental material (see Section X.X.X).
• Not advance the Ritchie and Lehnen (Ritchie and Lehnen. 1987) study for calculation of a
candidate RfC.
• Response: EPA agrees with this recommendation and does not advance Ritchie and Lehnen
fRitchie and Lehnen. 19871 to derive a candidate RfC.
• Decreased pulmonary function.
• The committee agrees with EPA that formaldehyde exposure may cause a decrease in
pulmonary function, but EPA should provide a clear rationale to support that conclusion (p.
6).
• Response: In the revised draft assessment, the studies of pulmonary function were
evaluated and synthesized using a common framework applied to all hazard categories and
outcomes in the formaldehyde toxicological review. The studies are described in tables
categorized according to confidence in the study results determined by systematic
evaluation of risk of bias and sensitivity. The study evaluations, with the strengths and
limitations of the studies, are documented in supplemental material (see Section X.X.X). A
WOE discussion provides the rationale supporting the conclusion.
• Furthermore, although the committee supports the use of the study by Kryzanowski et al.
(1990) to calculate a candidate RfC, EPA should provide a clear description of how the study
was used to estimate a point of departure and should also consider the studies conducted
by fKriebel etal.. 19931. fKriebel etal.. 20011 and the chamber studies for possible
derivation of candidate RfCs (p. 6; also at end of the chapter).
This document is a draft for review purposes only and does not constitute Agency policy.
D-22 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
• Response: The description of how the POD for Krzyzanowski et al. was derived is found in
Section 2.X.X. EPA evaluated study results from (Kriebel etal.. 1993): (Kriebel etal.. 2001)
to develop a candidate RfC. Kriebel et al. (Kriebel etal.. 20011 evaluated the effect of
formaldehyde exposure during a weekly 2.5-hour laboratory session over a 12-week
anatomy course using a random effects model. For each week, two measures of
formaldehyde exposure were calculated for each student, the average concentration during
that week's laboratory session and the average of all the previous weekly laboratory
sessions. These two measures of formaldehyde exposure were included simultaneously in
the random effects model. Both exposure estimates were associated with peak expiratory
flow rate (PEFR) among the laboratory students. Estimation of a cRfC using these data is
not straightforward due to the simultaneous modeling of the two exposure estimates and
the complication of potential covariance between these effects. Therefore, a POD could not
be determined from these data. The controlled human exposure studies of pulmonary
function were not included in the evaluation of the hazards of subchronic or chronic
exposures because these studies exposed subjects only for minutes or hours while the
review focused on effects related to exposure over a prolonged period.
The committee recommends that EPA address the following in the revision of the
formaldehyde draft IRIS assessment:
• Prepare plots of the findings of the chamber studies to assess the use of pooling their
results.
• Response: The controlled human exposure studies of pulmonary function were not
included in the evaluation of hazard because these studies exposed subjects only for
minutes or hours to high concentrations while the review focused on effects related to
exposure over a prolonged period. Several studies more relevant to the long-term exposure
setting that was the focus of this review were available.
• Provide further justification for its choice of the study by Krzyzanowski et al. fKrzyzanowski
etal.. 19901 for estimating the point of departure.
• Response: The current draft assessment contains a detailed discussion and rationale for
why the study by Krzyzanowski et al. fKrzyzanowski etal.. 19901 was selected for the
development of a candidate RfC (see Section X).
• Respiratory tract pathology
• Animal studies in mice, rats, and nonhuman primates clearly show that inhaled
formaldehyde at 2 ppm or greater causes cytotoxicity that increases epithelial-cell
proliferation and that after prolonged inhalation can lead to nasal tumors. Although the
committee agrees with EPA that the human studies that assessed upper respiratory tract
pathology were insufficient to derive a candidate RfC, it disagrees with EPA's decision not to
use the animal data (pp. 6-7).
• Response: EPA agrees with this point and has evaluated the toxicology studies reporting
respiratory tract pathology to identify a POD and derive a candidate RfC based on incidence
of squamous metaplasia [Woutersen et al., 1989; Kerns et al., 1983] (see Section l.X).
This document is a draft for review purposes only and does not constitute Agency policy.
D-23 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
• The committee concludes that a candidate RfC should be calculated for noncancer pathology
of the respiratory tract (that is, in the nasal epithelium).
• Response: EPA agrees with this point and has evaluated the toxicology studies reporting
respiratory tract pathology to identify a POD and derive a candidate RfC based on incidence
of squamous metaplasia [Woutersen et al., 1989; Kerns et al., 1983] (see Section l.X).
• Do not calculate a candidate RfC for mucociliary clearance.
• Response: EPA has not calculated a candidate RfC for mucociliary clearance.
• Asthma
• In infants and children, wheezing illnesses that are the result of lower respiratory tract
infections are often labeled as asthma, and in adults, the symptoms can overlap with those
of other chronic diseases, such as chronic obstructive pulmonary disease. Thus, a critical
review of the literature is essential to ensure that what is being evaluated is asthma. The
committee notes that this issue is not adequately addressed in the draft IRIS assessment
and that EPA advanced a study (Rumchev et al., 2002) that most likely suffers from
misclassification of infection-associated wheezing in young children as asthma (pp. 7 and
61).
• Response: EPA agrees that the condition experienced by the children in the Rumchev et al.
(2002) study is unlikely to representthe asthma phenotype that characterizes the majority
of research in childhood asthma (with onset typically in grade school). EPA developed
criteria to evaluate the definitions for the measures of allergy, asthma and other respiratory
outcomes reported in the epidemiology studies. This process included consultations with
two groups of clinical and epidemiology experts in allergy and asthma regarding the
reliability, validity, and interpretation of various types of outcome measures used in the
identified observational epidemiology studies. Based on these criteria, the study by
Rumchev et al. (2002) is not included in the set of studies examining asthma.
• The draft IRIS assessment also provides little discussion of the current understanding of the
mechanisms of asthma causation and exacerbation. Given the abundant research available,
the committee recommends that EPA strengthen its discussion of asthma to reflect current
understanding of this complex disease and its pathogenesis (pp. 7).
• Response: See next comment
• Asthma is a complex phenotype on whose pathogenesis substantial research has been
conducted. The discussion of asthma needs to be strengthened to reflect the extensive
literature better. The discussion of mode of action needs to be greatly strengthened and
grounded in current understanding of pathogenesis. The current speculative discussion is
not satisfactory (p. 61).
• RESPONSE: EPA agrees with these two suggestions (5.1.4.2 and 5.1.4.3). The pathogenesis
of asthma, as currently understood, and the potential mode(s) of action through which
formaldehyde may act in the exacerbation of this condition, are discussed in a more
comprehensive MOA discussion for portal of entry noncancer effects, including asthma and
immune-related endpoints (see Section l.X).
This document is a draft for review purposes only and does not constitute Agency policy.
D-24 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
• Although the committee agrees that the study by Garrett et al. (1999) should be used to
calculate a candidate RfC, the approach taken to identifying the point of departure needs
further justification (p. 7).
• RESPONSE: In the current draft assessment, the Garrett et al. (1999) study was considered
for the derivation of a candidate RfC for allergic sensitization, but was not advanced because
of uncertainty with respect to the timing of the exposure measure and its relation to skin
prick test results.
• The committee recommends that EPA address the following in the revision of the
formaldehyde draft IRIS assessment: Strengthen the discussion of asthma to reflect current
understanding of this complex phenotype and its pathogenesis better. There should be
greater clarity regarding the outcomes considered: incident asthma (the occurrence of new
cases), prevalent asthma (the presence of asthma at the time of study), or exacerbation of
established asthma (p. 61).
• Response: As indicated in response to previous comments, EPA agrees with this suggestion.
Based on EPA's consultation with clinical and epidemiology asthma experts, EPA has
divided the studies relating to asthma into studies of incident asthma, studies of prevalence
of current asthma (typically ascertained based on frequency of symptoms or medication use
over the past 12 months), and studies of asthma severity or asthma control (frequency of
symptoms or medication use over a short period of time, e.g., 2-4 weeks). Asthma
exacerbation is a term typically used in clinical trials, and considers the need for use of
systemic corticosteroids. EPA did not identify any studies of formaldehyde exposure that
examined this type of outcome. The current draft Toxicological Review presents the
collection of asthma studies based on type of outcome, population and exposure setting
(e.g., residential, school-based, or occupational exposure; adults or children). This revised
presentation, including both tabular and graphical summaries of the studies, provides
greater clarity regarding the observed results, and how variation in specific features of the
studies (most notably exposure levels) contributes to the variation in the observations.
• Respiratory tract cancer
• However, the draft IRIS assessment does not present a clear framework for causal
determinations and presents several conflicting statements that need to be resolved
regarding the evidence of a causal association between formaldehyde and respiratory tract
cancers. On the basis of EPA cancer guidelines, the committee agrees that there is sufficient
evidence (that is, the combined weight of epidemiologic findings, results of animal studies,
and mechanistic data) of a causal association between formaldehyde and cancers of the
nose, nasal cavity, and nasopharnyx. It disagrees that the evidence regarding other sites in
the respiratory tract is sufficient (pp. 9 and 87).
• Response: EPA thanks the NAS for this comment and has revised the document to describe
the evidence, the elements that contributed to the weight of evidence, and our conclusion
concerning formaldehyde and respiratory tract cancer. The discussion and conclusions in
the document are consistently presented and we reach (epidemiologic) conclusions about
cancers of the nose and nasal cavity (sinonasal cancer), the nasopharynx, the
oro/hypopharynx, and the larynx.
This document is a draft for review purposes only and does not constitute Agency policy.
D-25 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
• EPA's review of the literature on formaldehyde and respiratory cancer was thorough and
appropriate. It would be useful if, in the future, EPA could explicitly state its criteria for
evaluation of the evidence of causality based on its own cancer guidelines. Several sections
of the draft IRIS assessment contain conflicting statements on the evidence of causality that
clearly need to be rectified. The committee finds that, on the basis of EPA's guidelines, there
is sufficient evidence of a causal association between formaldehyde and cancers of the nose
and nasal cavity (ICD8 160) and nasopharynx (ICD8 147) but not other sites of respiratory
tract cancer (p. 87).
• Response: EPA thanks the NAS for this comment and has revised the document to describe
the evidence, the elements that contributed to the weight of evidence, and our conclusion
concerning formaldehyde and respiratory tract cancer. The discussion and conclusions in
the document are consistently presented and we reach (epidemiologic) conclusions about
cancers of the nose and nasal cavity (sinonasal cancer), the nasopharynx, the
oro/hypopharynx, and the larynx.
• The committee agrees that the study by Hauptmann et al. (2004) is an appropriate choice
for the derivation of a point of departure and unit risk. Although it is a high-quality study, it
is important to recognize some of its deficiencies, such as the apparent inconsistency
between the findings in different plants in the study and the weakness of the exposure-
response relationship in connection with cumulative exposure. Furthermore, the study was
found to be missing deaths in a later update of the cohort for lymphatic and hematopoietic
cancers. NCI is updating its cohort for respiratory cancer and other solid tumors. The
update not only will include the missing deaths but will extend the follow-up, and this will
result in nearly twice the amount of deaths (pp. 9 and 88).
• Response: Consistent with the evaluation of all relevant studies considered in the
toxicological review using standardized approaches, the cohort followed by the Hauptmann
et al. (2004) study was evaluated for risk of bias and sensitivity, and this evaluation is
documented in the supplemental material (see Section XX.X) and in the evaluation of hazard
(see Section XX). EPA has incorporated the updated follow-up of this cohort (Beane
Freeman et al., 2013) in its synthesis of the epidemiological studies and used these data in
the derivation of the unit risk value.
• Immunotoxicity
• The draft IRIS assessment presents numerous studies suggesting that formaldehyde has the
ability to affect immune functions. However, EPA should conduct a more rigorous
evaluation of the strengths and weaknesses of the studies; more integration of the human
and animal data would lend support to the conclusions made. The committee agrees with
EPA's decision not to calculate a candidate RfC on the basis of immunotoxicity studies (p.
10).
• Response: The current draft includes a discussion of the quality of the studies of immune
function using a framework developed for evaluating all epidemiology studies in the
assessment. Animal evidence for immunotoxicity was incorporated throughout the
document, integrated with the human data, and used to bolster mode-of-action analysis for
several endpoints (e.g., lymphohematopoietic cancer). Regarding animal studies relevant to
allergy and respiratory hypersensitivity, advice from allergy experts was incorporated
concerning the interpretation of the allergy outcome measures evaluated in epidemiology
This document is a draft for review purposes only and does not constitute Agency policy.
D-26 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
studies. The hypersensitivity-relevant experimental studies provide mechanistic support
and were integrated with the epidemiology studies in evaluating the weight of evidence for
immune system hazard. Although the toxicology studies were not used to derive a
candidate RfC, results from several epidemiology studies contributed to the development of
candidate RfCs for allergy-related conditions and asthma.
• The committee agrees with EPA's decision not to calculate a candidate RfC for
immunotoxicity at this time. The committee recommends, however, that EPA address the
following in the revision of the formaldehyde draft IRIS assessment:
• Provide a more careful evaluation of the relative strengths and weaknesses of the key
studies.
• Response: Each of the key studies was evaluated using several categories relevant to
internal validity (bias) that could lead to an under- or over-estimate of risk, and other
features that can affect the interpretation of the results. The details of this process are
provided in Supplemental Material Section C.5., and the summaries of the results are
included in the tabular displays and discussion of studies in the toxicological review.
• Consider giving additional weight to animal studies in which exposure assessment was
more rigorously controlled (p. 97).
• Response: Details of the exposure protocol, including level of control and source of
formaldehyde, were explicitly considered in the evaluation of controlled exposure studies.
• Neurotoxicity
• The committee found that EPA overstated the evidence in concluding that formaldehyde is
neurotoxic; the human data are insufficient, and the candidate animal studies deviate
substantially from neurotoxicity-testing guidelines and common practice. Furthermore, the
committee does not support EPA's conclusion that the behavioral changes observed in
animals exposed to formaldehyde are not likely to be caused by the irritant properties of
formaldehyde. Data indicate that those changes could occur as a result of nasal irritation or
other local responses; stress, also an important confounder that can affect the nervous
system, was not considered by EPA. The draft IRIS assessment provides conflicting
statements that need to be resolved about whether formaldehyde is a direct neurotoxicant
(p. 10).
• Response: EPA has updated and reconsidered the existing body of evidence for
neurotoxicity. The section has been revised to clearly present the strengths and limitations
of each study, as well as the relative contribution each study made to the overall
conclusions related to potential nervous system effects of formaldehyde exposure.
• Regarding the human data, the NRC indicated that the causal association between
formaldehyde exposure and ALS in one study (Weisskopfetal., 2009) was overstated.
Accordingly, a more detailed discussion of this study and its conclusions, as well as related
studies that have been published since the NRC review, have been added to the current text
A candidate RfC is no longer derived. As in the previous draft, the co-exposure limitations of
the Kilburn et al. studies are acknowledged and discussed. In the revised version, the data
from controlled human exposure studies are now evaluated in greater detail.
This document is a draft for review purposes only and does not constitute Agency policy.
D-27 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Supplemental Information for Formaldehyde—Inhalation
• In the current draft, endpoints in animal studies are critically evaluated alongside the
human data. The candidate animal studies relying on open field testing endpoints are no
longer considered for developing candidate values. In addition, the discussion of
nonguideline test paradigms, including the specific behavioral correlates they may be
capable of distinguishing, has been expanded in the text. The rodent-specific irritant
response, reflex bradypnea, is now explicitly considered for each study relevant to
interpreting the potential neurotoxicity hazard (see Appendix B.4.6). In addition,
discussion of behaviors evaluated at formaldehyde levels at which irritant-related
processes in rodents are expected has been added, and endpoints which are clearly reliant
on olfaction-related behaviors (e.g., odor-cued conditioning in (Sorg and Hochstatter.
199911. in particular, are considered likely to be influenced by irritation and are studies that
also examined the potential for nasal damage were preferred. The current draft includes a
more rigorous examination of the formaldehyde inhalation exposure methods used across
studies, which is now a critical consideration for evaluating how well individual studies
inform the potential for formaldehyde-induced neurotoxicity. An important confounder
identified in the previous draft is now attributed greater weight in the revised draft
Specifically, contamination of formaldehyde solutions with methanol, a known reproductive
and nervous system toxicant, was present in many of the studies. When contamination with
methanol was identified, or when the test article was not reported, the studies are now
attributed much less weight in the overall database and multiple discussions of possible
confounding by methanol-induced toxicity have been added to the current text.
• Potential, stress-induced changes by formaldehyde are considered to be highly relevant
effects of exposure. This has been more fully discussed in the revised text, including a
section on mechanistic information supporting potential indirect, neurotoxic effects. The
current draft now considers the potential for contributions from stress or other
uncontrolled variables to the observed responses. Unfortunately, the design of many of the
identified studies does not permit a separate evaluation of immediate, stress-induced
behaviors and possible direct effects of formaldehyde on neurobehavior. Stress-related
changes that persist after exposures are terminated (e.g., neural sensitization; altered
habituation) are now interpreted with greater concern.
• EPA agrees that the limited systemic availability of formaldehyde and its metabolites makes
it highly unlikely that inhaled formaldehyde is a direct neurotoxicant This viewpoint is
now presented throughout the document (it is now an underlying assumption), and only
potential mechanisms for indirect actions of inhaled formaldehyde are now discussed. As
stated in the U.S. EPA Guidelines for Neurotoxicity Risk Assessment (1998), indirect effects of
exposure are still considered to provide evidence of neurotoxicity.
• Evidence of neurotoxicity at exposure levels comparable to respiratory system effects has
not been conclusively shown for any neurotoxicity endpoint; this is clearly presented in the
current draft. EPA agrees that nearly all of the controlled exposure studies, including the
animal neuroanatomical changes, have significant limitations that reduce their ability to
inform the hazard assessment The limitations of these studies (including lack of clear
exposure-response relationships, study design deficiencies, possible confounders, and a
lack of database corroboration for specific endpoints) has been more transparently
described in the text (see Section 1.2.5) and appendix (see Appendix B.4.6).
• The committee concludes that the draft IRIS assessment overstates the evidence that
formaldehyde is neurotoxic. The selected studies are not sufficiently robust in design to be
This document is a draft for review purposes only and does not constitute Agency policy.
D-28 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
considered well executed for the purpose of neurotoxicity-hazard identification. One study
of rats by Malek et al. (Malek et al.. 2003a) was advanced by EPA for consideration. It was
considered to offer information on an outcome relevant to humans at an appropriate
concentration. Appropriately the study was not used to calculate a candidate RfC, partly
because of uncertainty in extrapolating from the exposure conditions in the study to a
chronic-exposure scenario (pp. 101-102).
• Response: EPA has updated and reconsidered the existing body of evidence for
neurotoxicity.
• As mentioned above, the current draft more clearly delineates the shortcomings of the
database; it is now concluded that the evidence for neurotoxicity is suggestive, due
primarily to limitations in the methodology of the available studies. Although the database
is limited, this is seen as an area of concern deserving further research.
• To specifically address questions related to the design and conduct of the neurotoxicity
studies, detailed discussions of study limitations have been added to the document text,
based on thorough evaluations of the testing methodology and validity for each assessed
endpoint (see Appendix B.4.6). The considerations used to interpret study quality for each
study/endpoint, including possible significant confounders and methodological limitations,
were applied in a comprehensive, transparent, and systematic manner.
• The study by Malek et al. (Malek et al.. 2003a) is not advanced for consideration in the
current draft.
• The committee agrees with EPA's decision not to calculate a candidate RfC on the basis of
the neurotoxicity studies (p. 10).
• Response: EPA agrees with the committee's recommendation and, in the current draft, EPA
does not calculate a candidate RfC on the basis of the neurotoxicity studies.
• The committee recommends that EPA address the following in the revision of the
formaldehyde draft IRIS assessment:
• Reevaluate its conclusions that behavioral changes are unlikely to be related to irritant
properties of formaldehyde (p. 102).
• Response: EPA agrees that irritation-related behaviors can have a significant influence on
many of the neurobehavioral changes observed following formaldehyde inhalation. A more
detailed consideration of the latency between exposure and testing as well as the
formaldehyde concentrations assessed, is now included in evaluations of study quality (see
Appendix B.4.6) and in the synthesis text These considerations are now included as
discussion points related to confounding (or as reasons for exclusion of studies as
noninformative) for select studies examining open field behaviors, neural sensitization, and
learning/memory processes. However, although it has not been sufficiently tested, an
additional discussion has been added regarding the potential for repeated formaldehyde-
induced irritation to elicit indirect, persistent neurological effects.
• Resolve inconsistencies regarding the concentration at which systemic effects of
formaldehyde exposure are expected. The draft IRIS assessment indicates that there is
This document is a draft for review purposes only and does not constitute Agency policy.
D-29 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
some question as to whether formaldehyde should be considered a direct neurotoxicant,
and some portions of the assessment suggest that systemic effects are unexpected at
formaldehyde concentrations less than 20 ppm. That statement is inconsistently made in
other parts of the document (p. 102).
• Responses: EPA agrees that the previous draft contained inconsistent statements regarding
direct or indirect neurological effects of formaldehyde. The revised draft does not include
any text identifying formaldehyde as a direct neurotoxicant. The available neurotoxicity
studies are insufficient to draw conclusions as to what formaldehyde concentrations might
be expected to elicit systemic, nervous system effects. In the animal studies, the suggestive
evidence of indirect neurotoxicity, defined in accordance with the neurotoxicity guidelines,
is generally reported at formaldehyde concentrations well above observations of direct
toxicity in portal-of-entry systems. Potential mechanisms for indirect neurotoxicity are
now succinctly stated in the hazard synthesis, with an emphasis on their highly speculative
nature.
• Reproductive and developmental toxicity
• The draft IRIS assessment states that epidemiologic studies provide evidence of a
"convincing relationship between occupational exposure to formaldehyde and adverse
reproductive outcomes in women." The committee disagrees and concludes that a small
number of studies indicate a suggestive pattern of association rather than a "convincing
relationship" (p. 10).
• Response: EPA agrees that the results of epidemiology studies suggest a pattern of
association between formaldehyde exposure and adverse reproductive outcomes in women.
The epidemiological and toxicological studies of reproductive effects in males and females,
and developmental effects were evaluated for risk of bias and sensitivity using approaches
and criteria described in the supplemental material (see Section X.X), and were categorized
according to the level of confidence (high, medium, and low) in the study results to inform
the hazard assessment. The study results were synthesized and a framework was used to
draw conclusions concerning male and female reproductive hazards and developmental
hazard. Using this framework, EPA concluded there was reasonable evidence for male
reproductive toxicity, inadequate evidence for female reproductive toxicity, and reasonable
evidence for developmental toxicity associated with inhaled formaldehyde exposures.
• The review of the reproductive and developmental outcomes in the draft IRIS assessment
includes relevant outcomes and literature. It does not consistently provide a critical
evaluation of the quality of publications and data presented or note strengths and
weaknesses of each study. That is especially the case with the animal studies (p. 108).
• Response: In the current draft assessment, the epidemiology studies are described in tables
categorized according to the extent they meet evaluation criteria that are provided in the
supplemental material (see Section X.X). The contribution of the studies to the hazard
assessment and the strengths and limitations of the studies are clarified. Likewise,
elements that were used to evaluate the quality of the animal studies are presented in
Section l.x.x., and the ability of the studies to inform the weight of evidence for reproductive
and developmental toxicity in humans is discussed.
This document is a draft for review purposes only and does not constitute Agency policy.
D-30 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
• Animal data also suggest an effect, but EPA should weigh the negative and positive results
rigorously inasmuch as negative results outnumbered positive ones for some end points,
should evaluate study quality critically because some studies of questionable quality were
used to support conclusions, and should consider carefully potential confounders, such as
maternal toxicity, effects of stress, exposure concentrations above the odor threshold, and
potential for oral exposures through licking (p. 10).
• Response: The text and tables in Section 1.X.X describe the criteria used to evaluate the
animal studies and the level of information provided by each study to the assessment of
hazard, in light of strengths and limitations. Evaluation of the toxicology literature included
criteria related to study quality, test subjects, study design, endpoint evaluation, data
considerations/statistical analyses and reporting. Considerations included maternal
toxicity, effects of stress, exposure concentrations above the odor threshold and potential
for oral exposures through licking. A key consideration for the interpretation of
developmental and reproductive outcomes associated with inhalation exposures to
formaldehyde was the potential for co-exposure to methanol, a known developmental and
reproductive toxicant, when the test article was an aqueous solution of formaldehyde.
Studies that used formalin but did not control for methanol, and studies that did not
characterize the formaldehyde source, are identified throughout Such studies were
assigned a "low" confidence rating. The consistency of study results with regard to specific
outcomes was an important consideration in the synthesis of evidence.
• The rationale for the assessment of the body of the epidemiologic evidence as convincing is
not well articulated. Issues regarding the potential portal of entry and mode of action in
relation to reproductive and developmental outcomes are not integrated into the weight-of-
evidence discussion (p. 108).
• Response: The evaluation of hazard for reproductive and developmental outcomes in the
current draft assessment was conducted using a framework for study evaluation and
evidence integration developed for the entire assessment. The current hazard descriptors
are consistent with the overall framework and their selection is described in Section XX.X.
The mode of action for the observed effects on reproduction and development is not known.
The mode-of-action discussion follows from the assumption that observed effects were not
due to systemic distribution of formaldehyde (see Section l.x.x.).
• Although the epidemiologic studies provide only a suggestive pattern of association, EPA
followed its guidelines and chose the best available study to calculate a candidate RfC (p.
10). The point of departure is appropriately selected (p. 108).
• Response: EPA agrees with this comment
• Lymphohematopoietic cancers
• EPA evaluated the evidence of a causal relationship between formaldehyde exposure and
several groupings of LHP cancers—"all LHP cancers," "all leukemias," and "myeloid
leukemias." The committee does not support the grouping of "all LHP cancers" because it
combines many diverse cancers that are not closely related in etiology and cells of origin.
The committee recommends that EPA focus on the most specific diagnoses available in the
epidemiologic data, such as acute myeloblastic leukemia, chronic lymphocytic leukemia, and
specific lymphomas (pp. 11 and 113).
This document is a draft for review purposes only and does not constitute Agency policy.
D-31 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
• Response: EPA agrees with this comment and recommendation. The current draft hazard
assessment focuses on the specific diagnoses of myeloid leukemia, lymphatic leukemia,
multiple myeloma, and Hodgkin lymphoma, and does not draw causal conclusions for the
broad categories of "all leukemias," grouping of nonspecific lymphomas, or "all LHP
cancers."
• As with the respiratory tract cancers, the draft IRIS assessment does not provide a clear
framework for causal determinations. As a result, the conclusions appear to be based on a
subjective view of the overall data, and the absence of a causal framework for these cancers
is particularly problematic given the inconsistencies in the epidemiologic data, the weak
animal data, and the lack of mechanistic data. Although EPA provided an exhaustive
description of the studies and speculated extensively on possible modes of action, the causal
determinations are not supported by the narrative provided in the draft IRIS assessment
Accordingly, the committee recommends that EPA revisit arguments that support
determinations of causality for specific LHP cancers and in so doing include detailed
descriptions of the criteria that were used to weigh evidence and assess causality (pp. 11
and 113).
• Response: The sets of epidemiologic studies related to each outcome were evaluated using
a common framework for determinations of causality. The following considerations were
evaluated: consistency, strength of the observed associations, exposure-response
relationships, and the potential impact of selection bias, information bias, confounding bias,
and chance. When information was available from the published epidemiologic studies, the
influence of time since first exposure or years of follow-up on the relative risk estimates
was evaluated. For example, for myeloid leukemia the following evidence contributed to
the causal determination.
• The causal evaluation for formaldehyde exposure and the risk of developing or dying from
myeloid leukemia placed the greatest weight on four particular considerations: 1) the
consistency of the observed increases in risk across a set of High and Medium confidence
independent results with varied study designs and populations; 2) the strength of the
association showing a 1.5 to 3-fold increase in risk; 3) the reported exposure-response
relationships showing that two measures of increased exposure to formaldehyde were
associated with increased risk of dying from myeloid leukemia; and 4) a biologically
coherent temporal relationship consistent with a pattern of exposure to formaldehyde and
subsequent death from myeloid leukemia allowing time for cancer induction, latency, and
mortality.
• Clarify how EPA determined weight and strength of evidence. The draft assessment should
be revised to discuss the benefits, limitations, and justifications of using one exposure
metric to determine causality and another to calculate cancer unit risk. Because the draft
assessment relies solely on epidemiologic studies to determine causality, further discussion
of the specific strengths, weaknesses, and inconsistencies in several key studies is needed.
As stated in EPA's cancer guidelines, EPA's approach to weight of evidence should include "a
single integrative step after assessing all of the individual lines of evidence" (EPA 2005a,
Section 1.3.3, p. 1-11). Although a synthesis and summary are provided, the process that
EPA used to weigh different lines of evidence and how that evidence was integrated into a
final conclusion are not apparent in the draft assessment and should be made clear in the
final version.
This document is a draft for review purposes only and does not constitute Agency policy.
D-32 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
• Response: As described in the response to 5.1.8.2, the sets of studies related to each
outcome were evaluated using a common framework for determinations of causality for
each cancer outcome and the rationales are described in Section l.XX. The determination of
causality was based on multiple epidemiologic studies that found associations with
different exposure metrics, and which were supported by mechanistic studies in exposed
humans that provided biological support for genotoxic and immunologic changes in
peripheral blood cells. The epidemiological evidence was synthesized using the common
framework developed for the formaldehyde assessment, and then the synthesis conclusion
was integrated with mechanistic evidence. This process is consistent with EPA's cancer
guidelines. The rationale for EPA's selection of the exposure metric used to derive the IUR
is provided in Section 2.XX). The IUR was derived using the regression coefficients for
myeloid leukemia in combination with other/unspecified leukemias and cumulative
exposure to account for likely inaccuracies in the underlying cause of death for myeloid
leukemia as documented by Percy et al. (1981; 1991). EPA selected the exposure-response
results based on cumulative exposure because this exposure metric is most relevant for
estimating life-time risk. The use of this metric also was supported by other studies that
observed associations with similar measures, such as duration of exposure or years since
first exposure.
• Revisit arguments that support determinations of causality of specific LHP cancers and in so
doing include detailed descriptions of the criteria that were used to weigh evidence and
assess causality. That will add needed transparency and validity to the conclusions.
• Response: The synthesis of the epidemiological evidence for specific LHP cancers uses a
common framework for determinations of causality that was developed for the assessment
• If EPA decides to rely on meta-analysis as a tool to assess causation, it should perform its
own meta-analysis with particular attention to specific diagnoses and to variables selected
and combined for analysis. The contrasting conclusions of the published meta-analyses
make it difficult to rely on conclusions from any one analysis (see, for example, Zhang et al.
2009; Bachand etal. 2010; Schwilketal. 2010) (p. 113).
• Response: EPA agrees that the contrasting conclusions in the published meta-analyses
make it difficult to rely on conclusions from any one analysis. EPA does not rely on the
conclusions of published meta-analyses.
• Quantitative assessment
• The committee supports EPA's selection of effects on which it based candidate RfCs but
does not support the advancement of two studies selected by EPA: Ritchie and Lehnen
fRitchie and Lehnen. 19871 and Rumchev et al. (2002). Furthermore, the lack of clear
selection criteria, inadequate discussion of some modes of action, little synthesis of
responses in animal and human studies, and lack of clear rationales for many conclusions
weaken EPA's arguments as presented in the draft IRIS assessment
• Response: The current draft assessment is based on a defined structure with criteria for
systematic review and the integration of evidence to determine causality for formaldehyde
effects. The dose-response assessment (see Section 2) also is based on a defined structure
with criteria for selecting studies for the derivation of candidate RfCs and organ-specific
This document is a draft for review purposes only and does not constitute Agency policy.
D-33 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
RfCs. The studies by Ritchie and Lehnen (Ritchie and Lehnen. 19871 and Rumchev et al.
(2002) were not used to derive RfCs for reasons described in the hazard assessment
• The committee disagrees with EPA's decision not to calculate a candidate RfC for upper
respiratory tract pathology. Many well-documented studies have reported the occurrence
of upper respiratory tract pathology in laboratory animals, including nonhuman primates,
after inhalation exposure to formaldehyde, and the committee recommends that EPA use
the animal data to calculate a candidate RfC for this end point
• Response: As stated in response 6.1.3.1, EPA agrees with this point and has evaluated the
toxicology studies reporting respiratory tract pathology to identify a POD and derive a
candidate RfC based on incidence of squamous metaplasia (Woutersen etal., 1989; Kerns et
al., 1983) (see Section l.X).
• The committee found that EPA dismissed the results of the exposure chamber and other
nonresidential studies too readily. Although the exposure durations for the chamber
studies are short relative to the chronic duration of the RfC, the studies provide
complementary information that could be used for deriving a candidate RfC.
• Response: EPA agrees that the controlled human exposure studies provide complimentary
information and relied on these studies in concert with the occupational and residential
studies to establish formaldehyde as a sensory irritant. The data indicate that this response
may be a more immediate phenomenon. In accordance with the criteria for selecting
studies for the derivation of candidate RfCs (see Table l.X), EPA used the dose-response
information for sensory irritation from epidemiology studies of residential exposure
because these studies evaluated populations including a range of ages, males and females,
and with health conditions perhaps conferring susceptibility (Hanrahan et al., 1984; Liu et
al., 1991) and compared these to cRfCs derived from medium confidence controlled human
exposure studies (Kulle etal., 1983; Andersen, 1983). For other effects, controlled human
exposure studies of acute effects after exposures of minutes or hours, did not contribute to
the evaluation of dose response and development of RfCs. However, evidence from
controlled human exposure studies was synthesized in the hazard assessments for asthma
and nervous system effects.
• Regarding the uncertainty factor that accounts for variability in response of the human
population, the committee suggests application of a value of 3 to calculate the candidate
RfCs on the basis of the work of Garrett et al. (1999), Hanrahan et al. (1984), and Liu et al.
(1991). Those studies included potentially susceptible populations, so the default value of
10 is not necessary. However, uncertainties remain regarding susceptible populations and
factors that affect susceptibility, so a value of 1 is not recommended.
• Response: Notably, the format and approach towards deriving candidate RfCs has been
substantially altered by EPA since the release of the previous draft. Currently, RfCs
corresponding to each health outcome with credible evidence of hazard (e.g., sensory
irritation; pulmonary function) are being separately derived, in addition to an overall RfC.
In part, this is because these organ or system-specific RfCs are more flexible for many risk
management situations. Using the new approach, the application of UFs is somewhat
different, in that the specific UF values (e.g., 3 or 10 for human variability) can differ across
the various health outcomes, even if (theoretically) they are based on the same study.
This document is a draft for review purposes only and does not constitute Agency policy.
D-34 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
• Specifically regarding the UFh factor, EPA guidance states that an uncertainty factor <10 for
human variability can be used if the POD is based on results in a susceptible group. A UFh of
10 was used for the POD for sensory irritation in teenage and adult populations (residential
exposures) in Hanrahan et al. (1984). Although the study population in Hanrahan et al.
(1984) was comprised of randomly selected households in mobile homes with individuals
representing a range of age, gender, health behavior, occupational status, and health status,
the identified PODs were not based specifically on evaluation of more susceptible
subgroups with conditions or characteristics that may contribute to variation in response.
Candidate RfCs were not derived using the Liu et al. (1991) study for sensory irritation or
the Garrett et al. (1999) study for increased asthma symptoms. However, a lower UFH (i.e.,
101/2 = 3) was selected based on study-specific data for some outcomes. These included a
study by Venn et al. (2003) of the degree of asthma control in children with asthma (the
study population consisted of this highly sensitive group), a study by Krzyzanowski et al,
(Krzvzanowski et al.. 1990) of pulmonary function decrements which included model
results comparing increased sensitivity among asthmatics, and a large study by Annesi-
Maesano et al. (2012) of associations with rhinoconjunctivitis and asthma prevalence
among children.
• Regarding the uncertainty factor that accounts for database completeness, the committee
suggests that EPA apply its first option as described in the draft IRIS assessment; that is,
apply a value of 1 with the qualification that further research on reproductive,
developmental, neurotoxic, and immunotoxic effects would be valuable.
• Response: EPA will use an uncertainty factor of 1 with the qualification that further
research is needed for several health endpoints.
• Although there are some gaps in the data on reproductive, developmental, immunologic,
and neurotoxic effects, the likelihood that new effects will be observed at concentrations
below those at which respiratory effects have been observed is low. Thus, the committee
supports the use of a UFD of 1 with the caveat that research of the types noted should be
pursued (p. 9).
• Response: Thank you for the recommendation. EPA will use an uncertainty factor of 1 with
the qualification that further research is needed for several health endpoints, particularly
because the available literature database does not sufficiently address the potential for
developmental toxicity (e.g., developmental neurotoxicity or immunotoxicity) at lower
exposure levels (i.e., those comparable to levels at which respiratory effects are observed).
• Overall, the committee found little synthesis of the relationships among the identified
noncancer health effects; it appeared that EPA was driven by the need to identify the best
study for each health effect rather than trying to integrate all the information. The
committee strongly recommends the use of appropriate graphic aids that better display the
range of concentrations evaluated in each published study selected for quantitative
assessment; the figures may help to identify how findings of studies cluster and especially
identify low or high reference values that may be inconsistent with the body of literature.
Ultimately, such graphics will improve the ability of the assessment and make a compelling
case for the RfC ultimately put forward.
• Response: The current draft presents the candidate RfCs together, including the relevant
PODs and the uncertainty factors applied. In addition, the rationale for selecting the overall
This document is a draft for review purposes only and does not constitute Agency policy.
D-35 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
RfC from the organ/system-specific RfCs includes a scatterplot of the organ/system-specific
RfCs in relation to the average composite UFs applied to derive each one, with the highest
uncertainty factors at the bottom of the graph. The size of the symbols for each
organ/system RfC represents confidence in the study(ies), POD(s) and database: small=
low; medium= medium; large = high. Therefore, the larger RfCs grouped closer to the top of
the graph are associated with higher certainty.
• Regarding calculation of unit risks, the committee agrees that the NCI studies and the
findings of the two follow-ups are a reasonable choice because they are the only ones with
sufficient exposure and dose-response data for risk estimation. However, the studies are
not without their weaknesses, and these need to be clearly articulated in the revised IRIS
assessment.
• Response: The current draft assessment includes a structured presentation of the
limitations and strengths of the epidemiology studies of cancer found in the supplemental
material (see Section X.X) and discussed as appropriate in the synthesis of the evidence in
Section X.X).
• The committee agrees that EPA's choice of NPC, Hodgkin lymphoma, and leukemia data
from the NCI studies to estimate a unit risk is appropriate given that the analysis of Hodgkin
lymphoma and leukemia primarily supports the assessment of uncertainty and the
magnitude of potential cancer risk. However, the mode of action for formaldehyde-induced
Hodgkin lymphoma and leukemia has not been clearly established. Moreover, the highly
limited systemic delivery of formaldehyde draws into question the biologic feasibility of
causality between formaldehyde exposure and the two cancers. Thus, substantial
uncertainties in using Hodgkin lymphoma and leukemia for consensus cancer risk
estimation remain.
• Response: The integration of evidence from the epidemiology studies provided the
rationale for EPA's finding there is sufficient epidemiologic evidence of a causal association
between formaldehyde exposure and increased risks of NPC, sinonasal cancer, and myeloid
leukemia and that there is suggestive epidemiologic evidence of a causal association
between formaldehyde exposure and increased risks of oro/hypopharyngeal cancer and
multiple myeloma. The MOA discussion for myeloid leukemia and multiple myeloma
concluded that the mechanisms for these cancers is not known, although evidence was
discussed that supported the biological plausibility for the conclusion. The cancer hazard
section discusses in depth the uncertainties associated with the causality conclusions, and
the dose-response section (see Section 2) discusses the uncertainties associated with the
derived unit risk estimate.
• Overall, the committee finds EPA's approach to calculating the unit risks reasonable.
However, EPA should validate the Poisson dose-response models for NPC, leukemia, and
Hodgkin lymphoma mortality with respect to adequacy of model fit, including goodness of
fit in the low-dose range, (log) linearity, and absence of interactions of covariates with
formaldehyde exposure. Furthermore, EPA is strongly encouraged to conduct alternative
dose-response modeling by using Cox regression or alternative nonlinear function forms.
• Response: See response to comment 6.13.
This document is a draft for review purposes only and does not constitute Agency policy.
D-36 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Supplemental Information for Formaldehyde—Inhalation
• The draft IRIS assessment does not provide adequate narratives regarding selection of
studies and end points for derivation of unit risks. The committee strongly recommends
that EPA develop, state, and systematically apply a set of selection criteria for studies and
cancer end points. The committee recognizes that uncertainty and variability remain
critical issues as EPA continues to promote quantitative assessment to improve
environmental regulation. There are still technical gaps in developing and applying
quantitative analysis of uncertainty and variability, especially to incorporate from all
sources and at all stages into an overall summary. The NRC Committee to Review EPA's
Toxicological Assessment of Tetrachloroethylene (NRC 2010) made several
recommendations for advancing methodology and promoting applications. Further
research is needed to study various approaches. Small (2008) discussed a probabilistic
framework. Given a set of options related to a key assumption (such as mode of action) or a
key choice (such as cancer end point), a preference score (or prior probability) may be
assigned to each option. The final risk estimate thus also has a weight or probability
attached that combines the preference on all options over each assumption or choice. The
overarching weight is the result of propagation of uncertainty in each assumption or choice
and aggregation of all assumptions over the risk assessment process tree. The collection of
final risk estimates for all permissible combinations of assumption and choice forms an
empirical distribution. That distribution quantifies the full range of variation and
uncertainty in the risk estimate. With the full range of variation of risk estimates and other
information on preference of key assumptions and choices, regulatory policy can depend
less on a single principal study, a single principal dataset, or a principal end point. The risk-
management process may use the distributional properties of the risk estimate to choose a
final risk estimate in the context of all feasible assumptions and choices. The committee
concludes that further development of systematic approaches to quantifying uncertainty
and variation will enable EPA to conduct IRIS assessments in a more transparent and
objective fashion (pp. 107-108).
• Response: Thank you for the description of possible approaches to quantifying uncertainty
and variation in deriving unit risk estimates. The Agency is working on developing methods
to better quantify uncertainty (ref) although no validated approaches have been offered to
date. The current draft presents a number of sensitivity analyses that examine a range of
unit risk estimates associated with different assumptions.
• Derivation of RfC: Overall, the committee is troubled by the presentation and derivation of
the proposed RfC values and strongly recommends the approach illustrated and described
in Figure S-l. A similar approach was recommended by the NRC Committee to Review
EPA's Toxicological Assessment of Tetrachloroethylene and used in recent EPA assessments
of tetrachloroethylene and trichloroethylene. Appropriate graphic aids that enable the
visualization of the concentration ranges of the candidate RfCs may identify a central value,
isolate especially low or high RfC values that might not be consistent with the body of
literature, and ultimately improve the ability of the assessment to make a compelling case
that the RfC proposed is appropriate for the most sensitive end point and protective with
regard to other potential health effects (p. 13).
• Response: The current revised assessment follows a process as outlined in Figure S-l of the
NAS review (p. 13). This is the systematic review process developed for the formaldehyde
assessment and described in the Preface to the toxicological review. The criteria and
rationale for identifying studies with appropriate data for deriving a cRfC are found in
Chapter 2 of the assessment and a figure is included that summarizes the cRfCs for each
This document is a draft for review purposes only and does not constitute Agency policy.
D-37 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
hazard with the range of concentrations that span the POD to the cRfC. Although not
specifically recommended by the NAS, the revised assessment selected organ-specific RfCs
(providing rationale for their derivation), and the assessment includes a scatterplot of the
organ/system-specific RfCs, which aids in providing the rationale for selection of the overall
RfC.
• Regarding calculation of unit [cancer] risks
• The committee agrees that the NCI studies are a reasonable choice because they are the
only ones with exposure and dose-response data sufficient for calculation of the unit risks;
however, the studies are not without their weaknesses, which should be clearly discussed
and addressed in the revised IRIS assessment Although there are uncertainties as
discussed above regarding the causal relationship of formaldehyde exposure and the three
kinds of cancer, EPA's decision to calculate unit risk values for them appears to be
defensible on the basis of the Agency's cancer guidelines. However, EPA should provide a
clear description of the criteria that it used to select the specific cancers and demonstrate a
systematic application of the criteria (p. 10).
• Response: EPA has clarified its discussion of the NCI studies strengths and limitations (see
Section 1.x). The evaluation of cancer types also is expanded, as is the rationale for
selection of cancer types for evaluation of dose-response relationships.
• The calculation of the unit risk values is a complex process, involves many sources of
uncertainty and variability, and is influenced by the low-dose extrapolation used (for
example, linear vs threshold). The committee therefore recommends that EPA conduct an
independent analysis of the dose-response models to confirm the degree to which the
models fit the data appropriately. EPA is encouraged to consider the use of alternative
extrapolation models for the analysis of the cancer data; this is especially important given
the use of a single study, the inconsistencies in the exposure measures, and the
uncertainties associated with the selected cancers (p. 10).
• Response:
• Independent analysis of the dose-response models to confirm model fit to data
• Analytical results quantifying exposure-response relationships were available from the
occupational cohort study conducted by NCI. The published studies provided information
about the Poisson dose-response models used to evaluate cancer mortality, including which
exposure metrics were evaluated, the p-values for exposure-response trend, and the
additional covariates and interaction terms included in the models (Hauptmann et al., 2004;
Beane Freeman et al., 2009; Beane Freeman et al., 2013).
• Additional information describing the model covariates and the impact of different model
forms (e.g., different lag periods, inclusion of terms for coexposures) on the magnitude or
statistical significance of the association of the exposure terms with mortality has been
added to the description of the studies in the assessment.
• NCI described in the published papers their approach to model evaluation, which included
evaluating the models in the entire cohort (nonexposed and exposed) and only among the
exposed workers, evaluating multiple possible lag periods, and adding quadratic terms to
This document is a draft for review purposes only and does not constitute Agency policy.
D-38 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
explore whether such terms indicated significant deviation from a log-linear relationship.
EPA concluded that the approach and level of reporting detail in the papers was acceptable
and obtained from the NCI the regression coefficients for the trend models reported in the
papers. EPA has obtained additional information from NCI about the Institute's review
process for this study (June 27, 2012 email to Barbara Glenn from Laura Beane Freeman).
NCI informed EPA that after publication of the 2003 and 2004 papers, independent
investigators obtained the cohort data and were able to recreate the results using these
models. In addition, for the most recent follow-up of the cohort, with deaths through 2004,
the NCI convened a group of extramural scientists to provide advice on the protocol for how
to conduct the follow-up. At that meeting, the NCI proposed to use the same methodologies
for analysis as in the prior publications. For the 2009 publication, regression models using
the same covariates as the 2003 and 2004 publications were built. In addition, two
researchers independently ran all analyses to confirm that no errors had inadvertently been
introduced. NCI's extensive internal review processes serve as additional layers of
verification and validation above and beyond peer review.
• The following detail on the covariates included in the Poisson regression models was added
to the assessment. The Poisson regression models stratified the cohort by calendar year (5-
year categories), age (5-year categories), sex, and race (white or other) and adjusted for pay
category (salary, ever wage, or unknown) (Hauptmann et al., 2004; Beane Freeman et al.,
2009; Beane Freeman et al., 2013). Multiple lag lengths in exposure were assessed and the
goodness of fit did not differ substantially for the different lag lengths; a 15-year lag was
selected by NCI for solid tumors and a 2-year lag for the lymphohematopoietic cancers.
Eleven potential confounding exposures (including benzene) in the plants were evaluated
by NCI and found not to alter the RR estimates appreciably in any of the models.37
Additionally, to specifically rule out an effect of benzene on the lymphohematopoietic
cancer results, individuals with possible exposure to benzene were excluded from the
analysis, and this did not change the RR estimates. As a final check on the potential for
confounding, Hauptmann et al. (2004) noted that evidence suggests that smoking is not a
confounder because there was no consistent excess or deficit for other tobacco-related
diseases, for example, bladder cancer, emphysema, and ischemic heart disease. The careful
work by NCI to evaluate the potential for confounding is considered sufficient to confirm
that the models fit the data appropriately.
• The NAS comment and recommendation above refers to the evaluation of model fit, and our
response assumes that the NAS panel is concerned specifically with whether the exposure
term in the model adequately fits the data. For the log-linear model, the p-value for a trend
test for the exposure metric in the model indicates the degree to which the log of relative
risk rises (or falls) with increases in the exposure metric.
• The p-values for the tests for trend for each exposure metric were reported in the published
papers. From the 2004 follow-up, the p-values using the cumulative exposure term (ppm-
years) indicated that an increasing trend in cancer relative risk was observed for NPC (p =
0.07), leukemia (p = 0.08), and Hodgkin lymphoma (p = 0.06). The p-values for average
intensity (ppm) indicated a rising trend in relative risk only for Hodgkin lymphoma (p =
0.03). Finally, the p-values for peak exposure (4 categories [ppm]) indicated a rising trend
37The one exception was a model for NPC that included melamine- note that melamine can be combined with
formaldehyde to form a resin and controlling for melamine in an analysis of formaldehyde may essentially be
controlling for formaldehyde, therein resulting in an alteration of the RR.
This document is a draft for review purposes only and does not constitute Agency policy.
D-39 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
in relative risk for leukemia (p = 0.02), myeloid leukemia (p = 0.07) and Hodgkin lymphoma
(p = 0.004).
• Whether or not the association of mortality with formaldehyde exposure varies according
to certain characteristics such as age, gender, race/ethnicity, or other individual attributes
is an important question in assessing risk. Effect modification by the above factors was
evaluated by NCI. According to Beane Freeman etal. (2009), page 755, "We found no
evidence of heterogeneity of relative risks by race (white or other), sex, or pay category
(salaried or hourly)." The evaluation of effect modification (evaluated statistically using a
cross-product term in the model) was conducted for the lymphohematopoietic cancer types
under study, including myeloid leukemia and multiple myeloma, and for all exposure
metrics. Likewise, Hauptmann et al. (2004) tested heterogeneity for the solid cancers and
did not report any significant heterogeneity (see Table 7). Therefore, it was not necessary
to account for variation in risk by these individual characteristics in the estimation of the
unit risk. This information has been added to the description of the studies in response to
the following NAS comment, "One may also wonder whether there were any covariates
(such as sex) that interacted with formaldehyde exposure. The presence of any interactions
that indicate effect modification will make the extrarisk formula (Rx - Ro/(l - Ro) depend
on the covariates involved rather than independent, as assumed in the draft IRIS
assessment" (pp. 137-139).
• Alternative extrapolation models for the analysis of the cancer data
• The NAS commented further in their review saying, "EPA is encouraged to consider the use
of alternative extrapolation models, including Cox regression models and nonlinear model
forms. The details of such modeling activities should be included in an appendix to the IRIS
assessment in sufficient detail that the results can be reproduced." "The authors [Callas et
al., 1998] suggested that Cox regression be used when confounding cannot be well
controlled or when age at cancer death does not follow an exponential distribution" (p.
138).
• EPA agrees that the Cox proportional hazards model is an alternative to the Poisson model;
however, because age was carefully controlled in the analyses, the Poisson regression
results should be essentially the same as those that would be obtained from a Cox analysis.
Callas et al. (1996,1998) have reported, based on analyses of an earlier follow-up of the NCI
formaldehyde cohort, that these two models yield nearly identical RR estimates and CIs
except in situations in which age cannot be closely controlled in the Poisson analysis. The
NCI analyses had a very fine level of control for age by using 5-year age groups, a
nonparametric approach that controls for potential confounding by age even when the risk
function for age may be strongly nonlinear.
• The log-linear Poisson model assumed a linear relationship between log RR and
formaldehyde exposure. One of the published papers described NCI's approach to
evaluating whether the relation of exposure with mortality was log-linear, or whether
nonlinear terms would provide a better fit. This was done by including a quadratic term in
the Poisson analysis to investigate whether there was a departure from the log-linear
model. The authors concluded that there was no evidence of a departure from log-linearity
for NPC (personal communication from Michael Hauptmann, June 11, 2013) and all
leukemia (Beane Freeman.etal., 2009).
This document is a draft for review purposes only and does not constitute Agency policy.
D-40 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
D.2. RESPONSE TO PUBLIC COMMENTS
• Study selection: EPA's Guidelines for Carcinogen Risk Assessment for weight-of-evidence
evaluation were not upheld in the reliance on statistical findings from Zhang (2009) and
disregard of those from Bachand et al. (2010). One commenter questions the reliance on
Hauptmann et al. (2003) and disregard of Beane-Freeman et al. (2009), which offers
critique of the Hauptmann study (missed 1,000 deaths), as well as EPA's failure to include a
more recent meta-analysis carried out by Bachand et al. (2010), which assessed all cohort,
case-control, and proportional mortality ratio studies (including Beane-Freeman and the
corrected Hauptmann data). One commenter identified one of the major flaws of the draft
assessment to be the omission of key studies, causing it to fail to meet the standards of the
IQA and EPA's own guidelines.
• Response: EPA conducted a systematic literature search to identify all relevant primary
publications reporting epidemiology studies of cancer risk among formaldehyde-exposed
populations. Reviews and meta-analyses were used to identify any literature that may have
been missed by the literature search process, but the results of these reviews were not
included in the synthesis of evidence on cancer risk. EPA agrees with the comments by NAS
on the use of meta-analyses (see Comment 6.1.7.2.5). Consistent with the NAS
recommendation, EPA conducted an independent synthesis and integration of the primary
literature and did not rely on the conclusions of published meta-analyses. EPA included the
NCI studies with updated analyses that included the 1,000 deaths that were missing in the
earlier publications.
• Oral exposure and formalin: One commenter noted that the draft review creates
confusion in its handling of issues related to formaldehyde gas by using studies that involve
the ingestion of formalin. The commenter further comments that while the draft recognizes
that two separate chemicals with different characteristics that are commonly referred to as
"formaldehyde," the document fails to adequately distinguish between them and causes
further confusion by using the two chemical names interchangeably. The commenter
further states that this document perpetuates the false assumption that anhydrous
formaldehyde gas is readily soluble in water, rather than providing clarity on this issue to
guide medical researchers about the true composition of formalin to eliminate such errors.
• Studies using formalin: The commenter noted that it would be an error for EPA to issue a
report with the objective of providing information about the chronic inhalation of
formaldehyde gas while relying on studies involving ingestion of formalin (methylene
glycol). The commenter notes that the error of confusing formalin with formaldehyde gas
becomes egregious when it is considered that the studies in question ignore the fact that
formalin contains significant quantities of methanol, and that the amount of formaldehyde
contained in formalin is dramatically overestimated.
• Response: The current draft assessment recognizes that the health effects from oral
exposure to formalin may not be relevant to a hazard assessment of formaldehyde exposure
through inhalation because of differences in distribution and metabolism via these two
routes. Methanol is systemically distributed and is metabolized to formaldehyde in organs
that are not directly exposed to formaldehyde when it is inhaled. Because of these
considerations and because there is a large set of studies using inhalation exposures,
This document is a draft for review purposes only and does not constitute Agency policy.
D-41 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
toxicity studies that used an oral exposure to formalin have not been relied on for the
determination of hazard.
• Oral vs inhalation dose (Til et al., 1989): The commenter stated that any increases in the
resulting blood concentrations by the oral versus inhalation would be best performed using
a kinetic model and that the highest dose administered in the Til et al. (1989) study (109
mg/kg/day in female rats) would be equivalent to inhalation concentrations of
approximately 105 ppm formaldehyde, assuming a body weight of approximately 0.35 kg
and an inhalation rate of 0.29 m3/day. The commenter notes that this concentration cannot
be achieved in an inhalation study because of respiratory irritation and nasal carcinomas in
animals after exposure to > 10 ppm formaldehyde for chronic durations.
• Response: The Til el al. (1989) study is not relied on in the current draft assessment
• Errors
• Data transcription: One commenter notes an apparent misreading of the Batelle report
(Volume A Table 10), in which a reported p-value of 0.0056 from the adjusted Cox/Tarone
pair-wise comparison of the control to 15 ppm for all leukemias, but the report lists the p-
value of 0.0003, which is erroneously taken from the pair-wise analysis of control to 15
ppm for endpoints of uterus, endometrial stromal.
• Response: This information is corrected in the current draft assessment.
• Toxicokinetics—systemic distribution
• Contribution to blood levels from other sources (drugs): One commenter noted that
formaldehyde is also directly released into the blood by FDA approved "prodrugs" through
bioconversion, as described by Dhareshwar and Stella (2008). The commenter references
the study's conclusion that exogenous sources (such as from the bioconversion of prodrugs)
compared to endogenous sources, contribute a very small fraction. The commenter notes
that Dhareshwar and Stella (2008) indicate that formaldehyde is metabolized quickly
enough that accumulation and systemic toxicity is unlikely after casual exposure.
• Response: Although noninhalation studies (such as the study referenced above) are
generally not used in the current assessment, the metabolism and distribution of inhaled
formaldehyde, in the context of endogenous levels of formaldehyde, are thoroughly
discussed in the AD ME appendix (see Section XX), with discussion as appropriate
throughout the main body of the document.
• Leukemia animal studies: One commenter raised concern that the statement in the draft
IRIS profile stating that the study conducted by Battelle Columbus Laboratories (1981)
provides the only evidence of formaldehyde-induced leukemia or lymphoma was incorrect
The draft further states that although there were significant early deaths in some of the
exposure groups, formaldehyde exposure slightly increased leukemia incidence in female
but not male rats. The commenter feels that the term "slightly increased" is imprecise and
fails to reflect that no statistically significant increase was found. The commenter
summarizes the Battelle study and its statistical tests. In addition, the commenter conducts
further statistical tests to the female rat leukemia data and conclude that 1) statistical tests
applied to the female leukemia data that adjust for survival do not indicate a statistically
This document is a draft for review purposes only and does not constitute Agency policy.
D-42 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
significant increase in leukemia incidence and 2) the test applied in the Battelle (1981)
report may inflate the likelihood that the incidence of leukemia would increase significantly
if more of the animals had survived.
• The commenter also noted that the draft review's suggestion that the "adjusted" incidence
of lymphoma in female mice was significantly increased is incorrect The commenter stated
that statistical significance in the methods used in the Battelle (1981) study is achieved with
a p-value of 0.05 divided by the number of dose groups, or p<0.0167.
• The commenter also notes a possible misreading of the Battelle report In the Battelle
Report Volume A Table 10 Analysis of Effects of Formaldehyde in Female Rats, the authors
reported a p-value of 0.0056 from the Adjusted Cox/Tarone pair-wise comparison of the
control to 15 ppm for leukemia, all. The next row in that table with an endpoint of uterus,
endometrial stromal polyp is the one that reports a p-value of 0.0003 for the pair-wise
analysis of control to 15 ppm.
• Response: The animal bioassays that evaluated lymphomas and leukemias were evaluated
systematically for the current draft assessment The summary reports of the studies in the
published literature did not discuss leukemia or lymphoma rates (Kerns etal.. 1983:
Swenberg etal.. 1980b). However, tissue slides were examined histopathologically in all
animals from the control and 15 ppm dose groups at each interim and terminal necropsy;
the lesions examined included lymphoma and leukemia. At the intermediate dose groups of
2 and 6 ppm exposure concentrations, only the proximal target (i.e., the nasal passages)
tissues were examined unless unusual tissue masses or gross lesions were noted, or if the
animals died spontaneously, as indicated by experimental findings fBattelle. 19821. EPA
used the histopathology data of individual animals reported in Table H of Battelle (1982) to
evaluate the incidence of LHP cancers in these bioassays.
• Because the individual animal data and time of death were available from Battelle (1982).
EPA was able to adjust for differential mortality patterns among exposure groups taking
into account individual animal survival times. The results of this analysis are provided in
Section 1.2.3.2 and the details are provided in Appendix B.9. There was no statistically
significant increase in incidence in any of the treatment groups compared to controls; the
maximum increase was seen for lymphoma in female mice at 15 ppm (p-value = 0.29).
• Although chronic formaldehyde inhalation studies in animals do not show an increase in
LHP cancers (these data are summarized in Table 1-59), the detection of
leukemia/lymphoma in these studies may be limited by study design (limited statistical
power; all tissues potentially related to LHP cancers not measured in every study; focus of
histopathological evaluation on nasal tissue; animal deaths first from toxicities other than
LHPs at 15ppm) and generally poor predictive use of rodent models for the detection of
chemical-induced leukemia or lymphoma [refs]. There is a need for studies specifically
designed to target these cancers as the main endpoint. Overall, EPA agrees with the
commenter and confirmed that the available data do not provide evidence supporting the
development of LHP cancers in chronic rodent bioassays; however, given the design of the
available experiments, the studies are generally considered limited in their ability to detect
these types of effects.
• Mechanistic evidence—URT cancers
This document is a draft for review purposes only and does not constitute Agency policy.
D-43 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Supplemental Information for Formaldehyde—Inhalation
• Commenters recommended a review of the 13-week toxicogenomic study by Andersen et al.
(2010) that was recently accepted for publication in Toxicological Sciences (pending
revision).
• One commenter summarized findings by Meng et al. (2010), Hester et al. (2003, 2005), and
Andersen et al. (2008). The commenter feels like EPA fails to properly weigh the best
available science, including major findings from Meng et al. (2010) such as p53 findings.
According to the commenter, this study suggests that formaldehyde induced mutagenicity is
unlikely to play a role in tumorigenesis and supports a threshold concentration for
formaldehyde-induced nasal tumors.
• Draft IRIS Assessment places more weight on a single dose toxicogenomic study with an
inappropriate route of administration (nasal instillation) than a multiple dose inhalation
toxicogenomic study (Datson 2008). One commenter noted that the draft IRIS assessment
is incomplete because it does not discuss the implications of Andersen et al. (2008), which
replicated the dosing protocol used by Hester et al. (2003, 2005) to determine how this
dose and method of delivery compared with inhalation exposure and subsequent
toxicogenomic responses following administration of graded doses. According to the
commenter, the findings by Andersen et al. (2008) challenge the relevance of the effects
(e.g., changes in DNA repair genes) reported by Hester et al. (2003, 2005) and support the
notion that there is a clear, dose dependent transition in the effects of formaldehyde at the
level of transcription, and the cellular and histopathological levels.
• Response: The findings reported by Andersen et al. (2010, 2008), Mengetal. (2010), and
Hester et al. (2003, 2005), along with those of other authors, are discussed in the section on
Respiratory Tract Pathology Hazard (see Section l.XX) and in the Mode-of-Action Analysis
for Upper Respiratory Cancers in the current draft assessment (see Section l.XX). The
evidence on histopathological lesions and cellular proliferation (see Appendix XX for
detailed study data) is presented and compared in relation to time, location and dose. The
relative roles of mutagenicity and cell proliferation in the development of URT cancer are
discussed in the context of temporal and dose patterns, and formaldehyde effects are
compared to the carcinogenic properties of other similar chemicals that cause mutagenicity
or cell proliferation, or both. The findings of Mengetal. (2010) is included in the
mode-of-action analysis. P53 mutations were detected in 5 of 11 SCCs isolated from the
nasal passages of F344 rats following 2-yrs of exposure to 18 mg/m3 formaldehyde (Recio
et al., 1992; Wolf et al., 1995), but were not found in hyperplastic nasal tissue samples
following 90 days of exposure to similar concentrations (Meng et al., 2010). The lack of p53
mutations, or mutations in any other single gene, in hyperplastic nasal tissues after a 90-day
exposure is not considered a priori to be evidence against a role of mutagenicity in
formaldehyde-caused cancer. At 18 mg/m3, nasal squamous metaplasia preceding or
concomitant with hyperplasia is significantly elevated early after first exposure (within 7
days; Section XYZ Respiratory Pathology), prior to the emergence of dysplasia at 365 days,
in the nasal regions of F344 rats, which eventually harbor SCC after 330-548 days (Kamata
etal., 1997; Monticello etal., 1996; Kerns etal., 1983b). The absence ofp53 mutations in
reactive nasal mucosa after subchronic exposure is consistent with p53 mutations as a
selective or permissive factor acting later in formaldehyde-initiated carcinogenesis,
facilitating increased genetic instability and the progression of nascent neoplasms to
respiratory carcinomas, which appear months later (Hanahan and Weinberg, 2011;
Hanahan, 2000,188413).
This document is a draft for review purposes only and does not constitute Agency policy.
D-44 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
• The current draft assessment also discusses the literature evaluating epigenetic activity by
inhaled formaldehyde and effects on gene expression involved in regulation of cell cycle,
apoptosis, DNA repair and growth signaling pathways in nasal tissue from F344 rats after
acute or repeated exposure fAndersen et al.. 20081fRager etal., 2014; Hester etal., 2005;
Andersen etal., 2010); and similar results in nonhuman primates, including changes in
regulators of cellular proliferation, apoptosis, and inflammatory signaling (Rager etal..
20131. The interpretation of this evidence and implications for the role of formaldehyde
inhalation in URT carcinogenesis is discussed.
• Interpretation of Til et al. (1989)
• The commenter notes that there are two observations that can be made from the Til et al.
(1989) study that were not considered in the draft review: 1) development of nasal and
respiratory tumors is not the result of redistribution to these tissues following absorption,
but instead results from a direct contact, portal of entry effect and 2) the incidence of
leukemia in treated rats was not increased following chronic exposure to formaldehyde.
• Response: The Til et al. (1989) study, which exposed animals to formaldehyde in drinking
water, is not discussed in with regard to cancer development in the current draft
assessment.
• Mechanistic evidence—LHP cancers
• Genotoxic markers in peripheral blood: One commenter that inconsistencies are seen
from study to study in the types of effects reported following formaldehyde exposure and
that these studies and their limitations and inconsistencies are summarized in Table 2.
• The commenter further noted that the methods used in the Pala etal. (2008) study do not
differentiate between formaldehyde of endogenous and exogenous origin. Also, the
commenter noted that the presence and/or frequency of chromosomal aberrations in the
peripheral blood are not a validated marker of specific types of cancer.
• Response: Pala et al. (2008) evaluated chromosomal aberrations and sister chromatid
exchange in DNA of peripheral lymphocytes of laboratory workers. The study group was
divided into a group exposed to an 8-hour average concentration of < 0.026 mg/m3 and a
group exposed to > 0.026 mg/m3 formaldehyde measured using personal monitors, and
DNA damage was compared. No differences were observed, albeit both groups were
exposed to low levels of formaldehyde. The analysis evaluated whether a measure of
formaldehyde in the breathing zone of the workers was associated with DNA damage.
Although DNA adducts from endogenous formaldehyde have been measured in peripheral
blood lymphocytes, measures of inhaled formaldehyde are not associated with measures of
endogenous formaldehyde in blood. Therefore it was appropriate to evaluate a measure of
inhaled formaldehyde given the hypothesis tested in this study. EPA agrees that the
frequency of chromosomal aberrations in peripheral lymphocytes is a validated marker of
increased overall cancer risk, which is discussed in the section on mode of action for
lymphohematopoietic cancer (see Section l.XX).
• Zhang et al. (2010) critiques: One commenter conducted additional analysis of data on
Chinese workers included in the Zhang et al. (2010) analysis felt like the protocol for
evaluation of the cells for monosomy 7 and trisomy 8 was not followed. In addition the
This document is a draft for review purposes only and does not constitute Agency policy.
D-45 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
commenter raised concern that the validity of the model results reported by Zhang et al.
(2010) are in question and conclusions should not be drawn from the results of the
application of this model. The commenter further noted that while Zhang et al. (2010)
suggests that monosomy 7 and trisomy 8 are the most frequent cytogenetic changes
observed myeloid leukemia and myelodysplastic syndromes, they are not observed in the
majority of individuals with these diseases.
• Another commenter raised concerns about the Zhang et al. (2010) study. The commenter
noted that In a letter to the editor of Cancer Epidemiology, Biomarkers and Prevention
several scientists (Speitetal., 2010) posed questions about the Zhang (2009) study setup
and concluded that, "because this study is too preliminary and has too many shortcomings,
it is not suited to demonstrate a systemic (geno-)toxic mode of action of inhaled
formaldehyde." The commenter also noted that in response to a Freedom of Information Act
(FOIA) request, the National Cancer Institute (NCI) provided the individual hematology
(CBC) and cytogenetic (monosomy 7 and trisomy 8) data (all of which were pooled in the
original study). Based upon a very preliminary review of these data, scientists including a
hematologisthave identified a number of issues that raise added questions to those in Speit
etal.
• Response: EPA is aware of the published criticisms of Zhang etal. (2010) (Speitetal.,
2010). The current draft also reviews a subsequent study by the same group of
investigators that evaluated chromosomal aberrations in a larger subset of the cohort and
for the entire chromosome (Lan et al., 2015). While these analyses addressed cells more
relevant to hematopoietic outcomes, HSPCs, the findings from these studies are reviewed in
concert with those that compared chromosomal aberrations in peripheral blood
lymphocyte in exposed and unexposed groups. However, EPA agrees that there is not
enough evidence to determine that a direct genotoxic mode of action mediates inhaled
formaldehyde associated lymphohematopoietic cancers.
• DNA adducts (Lu et al., 2010): Another commenter noted that new information regarding
specific DNA adducts that arise from interaction of DNA with formaldehyde of either
endogenous or exogenous (inhaled) origin is available (Lu et al., 2010) and that this data is
far superior to the necessarily coarse exposure characterizations that have been developed
as part of retrospective cohort mortality investigations. The commenter goes on to
summarize the data reported by Lu et al. (2010). One commenter raised concerns about the
data reported by Lu et al., (2010) and noted that overall, the distribution of adducts caused
by inhaled formaldehyde could be consistent with induction of nasal carcinoma, but is not
consistent with induction of leukemia.
• Genotoxicity in bone marrow: The commenter references several studies (Casanova-
Schmitz etal., 1984; Heck and Casanova, 2004; Lu etal., 2010), indicating that formaldehyde
does not form DNA: protein cross links or DNA adducts and notes that the weight-of-
evidence conclusion from these studies is that exogenous formaldehyde is not a direct
genotoxic agent at sites distant from the point of exposure, in particular the bone marrow.
• Response: EPA discusses the findings of Lu etal. (2010) in the current draft assessment
and its contribution to our understanding of the distribution of exogenous and endogenous
formaldehyde (see Section l.XX; Supplemental Material ADME Section 2.XX). The
contribution of this study's findings to our understanding of the mode of action for
carcinogenicity is discussed in the MOA sections for URT cancer and LHP cancers,
This document is a draft for review purposes only and does not constitute Agency policy.
D-46 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
respectively. Along with other evidence, the measurement of formaldehyde-DNA adducts of
exogenous origin in the URT supports a mode of action that includes genotoxicity as a
mechanism for the observed associations with nasopharyngeal and sinonasal cancers. EPA
concluded that the mode of action for myeloid leukemia and multiple myeloma is not
known. Although there is strong evidence of genotoxic effects in peripheral lymphocytes, as
well as hematopoietic stem and progenitor cells from peripheral blood samples, the Lu et al.
(2010) study, as well as others, do not support the hypothesis that these effects are caused
directly by formaldehyde in bone marrow. The section on the MOA for LHP cancers
concludes that it is biologically plausible that formaldehyde-related myeloid leukemia and
multiple myeloma may occur as a result of events in the URT.
• Leukemia biological plausibility: One commenter noted that the presence and/or
frequency of chromosomal aberrations in the peripheral blood are not validated markers of
specific types of cancer and that there is no evidence that circulating hematopoietic stem
cells return to bone marrow during homeostasis (McKinney-Freeman and Goodell (2004).
• Response: EPA agrees that the frequency of chromosomal aberrations in peripheral
lymphocytes is a validated marker of increased overall cancer risk, which is discussed in the
section on mode of action for lymphohematopoietic cancer (see Section l.XX). The current
draft assessment discusses what is known about the hematopoietic stem and progenitor cell
physiology as part of the rationale for biological plausibility (see Section l.XX). As
discussed in the current draft, "As part of their physiological function, HSPCs migrate via the
vasculature to extramedullary tissues such as the liver, lung, small intestine, skin and
kidneys, and return via lymphatics to the bone marrow, by a process termed 'homing,'
which is mediated by cytokines, growth factors and hormones (Granick et al., 2012; Schulz
et al., 2009; Massberg et al., 2007). Although their numbers in the peripheral blood at any
one time constitute a small fraction of the total circulating leukocyte population in both
mice (Massberg et al., 2007) and humans (Zhang et al., 2010; de Kruijf et al., 2014), these
cells can completely replenish bone marrow stem cell populations (Massberg et al., 2007)."
• MOA for LHP cancers: One commenter discussed the two MOAs proposed by the draft for
leukemia and lymphoma and noted that if formaldehyde is a cause of myeloid leukemia or
Hodgkin lymphoma in humans, as posited in the draft IRIS assessment, there should be
cases of a nasal chloroma or nasal lymphoma in exposed workers, but none have been
reported. The commenter noted that no empirical data are cited, or appear to exist to
support the two MOAs proposed by the draft for leukemia and lymphoma.
• Response: A hypothesized scenario where HSPCs damaged in the URT tissues do not return
to the bone marrow, and form local foci of neoplastic leukocytes is discussed in the current
draft assessment (mode of action for lymphohematopoietic cancer [see Section l.XX]),
stating that there is no evidence supporting this hypothesis. The current draft assessment
states what evidence is available to support the possible mechanisms that are discussed.
• Mechanistic evidence—MOA URT cancers
• MOA for URT cancers: The commenter noted that the draft asserts that early mutations
play a key role in formaldehyde-induced nasal tumors in rodents. The commenter noted
that while various in vitro studies indicate that formaldehyde is mutagenic in a number of
test systems (ATSDR. 1999) (IARC 2006), none of these has ever been associated with
formaldehyde-induced nasal tumorigenesis.
This document is a draft for review purposes only and does not constitute Agency policy.
D-47 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
• The commenter noted that while the draft summarized the cell proliferation data from
Meng et al., (2010) [Draft IRIS Assessment, Fig. F-5], it did not mention the p53 findings.
According to ACC, the data suggest that p53 mutation is a late event not involved in the
carcinogenic MOA in formaldehyde-induced carcinogenesis and occurs only after other key
events have occurred (e.g., DNA-protein cross links, cytotoxicity, and cell proliferation). In
addition, the commenter noted that the data from Meng et al. (2010) supports a threshold
concentration for formaldehyde-induced nasal tumors and, therefore, do not support EPA's
conclusion that no threshold exists for formaldehyde-induced nasal tumors, and EPA did
not consider or address these conflicting data.
• Response:
• Weight of evidence for cancer—EPA guidelines
• One commenter expressed concern that the draft assessment does not meet EPA's own
guidelines in terms of objectivity, integrity, and even-handedness. Another commenter
stresses the importance of meeting the requirements of the IQA. One commenter noted that
the draft review fails to provide an objective view of the best available science as mandated
by the IQA and the EPA's Guidelines for Carcinogen Risk Assessment.
• One commenter pointed out that the draft assessment incompletely or incorrectly reports
the findings of studies causing the conclusions drawn to be biased, also noting that this does
not satisfy the standards developed in the IQA or those developed in the EPA guidelines to
justify the application of the descriptor "Carcinogenic to Humans." A commenter specifically
questions the conclusions regarding leukemia and Hodgkin lymphoma where convincing
epidemiological evidence is lacking, and mode-of-action data is both insufficient and
contradicted by negative animal data.
• Response: EPA followed its cancer guidelines in the integration of the evidence concerning
carcinogenicity for the current draft assessment The current draft explains the approach
and provides the criteria used to weigh evidence associated with varying degrees of
confidence both within and across evidence streams (human, animal, and mechanistic),
used to integrate evidence for a final conclusion (see Section s l.xx, l.xx and Supplemental
Material X.XX).
• NCI's peak exposure metric: One commenter noted that the draft IRIS assessment lacks a
discussion of the NCI's reliance on peak exposures and selection of atypical metric versus
more typical exposures used in epidemiology studies. The commenter further noted that
Hauptmann et al., (2003, 2004) did not explain the rationale for the use and development of
the peak exposure metric and that when Beane-Freeman etal., (2009) used a different
metric of potential exposure there was no evidence of increased risks.
• Response: The studies by NCI of the industrial cohort evaluated associations with several
exposure metrics, including duration of exposure, time since first exposure, average
exposure, cumulative exposure and peak exposure. EPA agrees that the peak exposure
metric is difficult to interpret because it is a categorical measure that includes the
experience of 15-minute intervals above an average formaldehyde concentration typical for
the task that may have occurred from 1 to several times over a career for an individual
assigned a particular peak exposure level. However, the evaluation of more than one
exposure metric by cohort studies is considered to be a strength, particularly because
This document is a draft for review purposes only and does not constitute Agency policy.
D-48 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Supplemental Information for Formaldehyde—Inhalation
several cancer endpoints are evaluated in mortality studies and for most of these, the
biologic mechanisms for cancer are not known.
• The study of LHP cancers by Beane-Freeman etal. (2009) did not observe an increasing
trend in myeloid leukemia deaths with cumulative exposure, although two other studies of
different cohorts found an association with increasing duration (Hauptmann et al., 2009;
Meyers et al., 2013) (see Section l.XX). Healthy worker selection bias and imprecision in
the cumulative exposure measure resulting in errors in estimating long-term formaldehyde
exposures across multiple jobs and tasks may have contributed to attenuating the observed
value of relative risk estimated for this study. Further, lack of specification of the leukemia
subtype in some death certificates likely resulted in the incomplete ascertainment of deaths
from myeloid leukemia, and a reduction in statistical power. For example, inconsistency
between the diagnosis recorded on a patient's medical record and the underlying cause of
death on the death certificate can lead to misclassification of the death in occupational
cohort studies, and a potential for bias toward underestimates of exposure-related cancer
risk. Indeed, studies that conducted this type of comparison for cancer deaths during 1970-
1971 and 1985-1986 found inconsistencies for lymphocytic and myeloid leukemia, which
were under-reported, and for "other and unspecified" leukemias, which were over-reported
(Percy et al., 1981; Percy et al., 1991). The overall leukemia classification was consistent
between the death certificate and the hospital medical record (Percy et al., 1981). The
follow-up periods for most of the cohort studies reviewed in this assessment encompassed
these years. EPA included an analysis by NCI that modeled the combined risk of myeloid
leukemia and other/unspecified leukemia deaths in relation to cumulative formaldehyde
exposure, which resulted in improved precision for the regression coefficient
• Plausibility of risk estimated from IUR: One commenter noted that the draft EPA unit
risk factor, when applied to levels at the high end of the daily ambient range or at high
ambient single exposures, results in a cancer risk ranging between 3 and 9 in 1000. The
commenter states that most Americans will be exposed at these levels, which makes the
draft risk factor unreasonable and implausible. The commenter disagrees with EPA's
conclusion that indoor exposure to formaldehyde is responsible for 16 percent of all cases
of Hodgkin lymphoma and 42 percent of all cases of nasal pharyngeal cancer.
• Response: The current draft assessment does not find a hazard for Hodgkin lymphoma.
Nonetheless, the comments suggest that the IUR can be used to estimate the current disease
burden attributable to formaldehyde. The IUR is an upper-bound estimate of the extra risk
per unit lifetime formaldehyde exposure, and it cannot be used for such attributable
fraction estimates. The comment misinterprets some "rough calculations" conducted by
EPA to derive "crude upper-bound estimates" for some lifetime exposure scenarios to
assure that the unit risk estimates for the rare cancers were not implausible (as upper
bounds) in comparison to actual case numbers; they were not
• POD for IUR estimation: One commenter noted that EPA has deviated from the usual 10%
or 1% rate of extra risk in selecting the POD for Hodgkin lymphoma and leukemia and has
instead chosen 0.05% for Hodgkin lymphoma and 0.5% for leukemia. The commenter
pointed out that the same issues exist for the risk rate for nasal pharyngeal cancer (NPC),
which was set at 0.05%. The commenter argued that selection of these rates is arbitrary
and results in a low POD which is then used in linear extrapolation to derive even lower
response levels. The commenter noted that the unit risk factors for Hodgkin lymphoma and
leukemia that result from this analysis are "ultra conservative" and do not agree with the
This document is a draft for review purposes only and does not constitute Agency policy.
D-49 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
underlying epidemiological evidence for these cancers. The commenter noted that the
response levels for NPC are derived from linear extrapolation, although the available data
support a nonlinear mode of action. The commenter stated that the NPC estimate is
unnecessarily conservative and that the Hodgkin lymphoma and leukemia estimates
drastically over estimate cancer risk at low formaldehyde concentrations. The commenter
recommended that EPA use the approach by Schlosser et al. (2003) for the NPC risk factor
and that EPA use the approach of Sielken et al. (2007) for the Hodgkin lymphoma and
leukemia risk factors.
• Response: The current draft assessment selected a 0.5% extra risk level for the derivation
of the POD for myeloid leukemia and a 0.05% extra risk level for the derivation of the POD
for NPC. These extra risk levels are appropriate for the background risks for these cancer
types and the RR estimates observed in the NCI study, as discussed in the draft assessment
(see Section 2.XX (pgXX) for myeloid leukemia and in Section 2.XX (pgXX) for NPC), and are
consistent with EPA guidance (U.S. EPA, 2012). Also, the comment appears to suggest that
the use of these extra risk levels results in overly low PODs and overly conservative unit
risk estimates. In fact, because the exposure-response model used is sublinear (i.e., the risk
increases more than linearly with increasing exposure), the unit risk estimates are lower
than would be obtained with a higher extra risk level (a straight line drawn from higher up
the sublinear curve is steeper than one drawn from lower down the curve). The use of
linear low-exposure extrapolation for both NPC and myeloid leukemia is consistent with the
mode-of-action conclusions in the assessment (see Section XX) and EPA guidance (U.S. EPA,
2005).
• The approach of Schlosser et al. (2003) is for rodent data and is not suitable for the human
NPC data; to the extent that it derives a POD and does not preclude linear low-exposure
extrapolation, it is consistent with the approach used in the draft assessment For reasons
documented in detail in EPA's draft ethylene oxide carcinogenicity assessment (U.S. EPA,
2014), EPA does not agree with the approach typically used by Sielken etal. (2007).
• IUR and low-dose extrapolation: One commenter noted that the draft IUR of 7.7 ppt is
4,000-fold lower than the mean indoor level and 37,000-fold lower than peak indoor levels
when using the figures in Stenton et al. The commenter stated that EPA's cancer risk
assessment is based on the assumption that a single peak formaldehyde exposure at any
time during a lifetime can cause leukemia, and that the predicted extra cancer risks
attributed to peak and average indoor levels of formaldehyde are 4 in 100 and 4 in 1,000,
respectively. The commenter argues that these levels are not realistic or plausible.
• Response: While the (lower bound) lifetime concentration associated with a 10 6 lifetime
cancer risk in the current draft assessment is different from the value stated in the above
comment, the current concentrations associated with lifetime extra risks between 1:10,000
and 1:1,000,000 are still low. Extrapolating risk levels using data from occupational cohort
studies where formaldehyde concentrations were higher than those experienced in
residential communities is associated with uncertainty about the exposure-response
relationship in the lower concentration range where data are not available. EPA concluded
that because current research does not allow the selection of a specific dose-response
model, the selection of a linear model for extrapolation is reasonable and consistent with
EPA cancer guidelines.
This document is a draft for review purposes only and does not constitute Agency policy.
D-50 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
• IUR and selection of dose-response data: One commenter noted that there are problems
with combining nasal tumors, Hodgkin lymphoma, and leukemia for the unit risk estimate,
as well as problems with each individually. Two commenters noted that there were no
significant trends for leukemia or myeloid leukemia by any exposure metric in the study
chosen for use in the dose-response analysis from Beane-Freeman et al. (2009). One
commenter noted that it is not appropriate to include negative data in dose-response
modeling. Two commenters stated that it is inappropriate to use the cumulative exposure
dose metric for Hodgkin lymphoma when the only significant increase in Hodgkin
lymphoma was seen when using the peak exposure dose metric. One commenter noted that
there is no apparent dose-response relationship between cumulative exposure and risk
which could provide adequate data for development of a unit risk factor.
• Response: The combined unit risk estimate for cancer is consistent with EPA's cancer
guidelines. The IUR is intended to address overall cancer risk, not risk associated with any
particular cancer type. Because data were not available to derive a combined IUR that
included cancer types other than nasopharyngeal or myeloid leukemia for which there was
sufficient or suggestive evidence of a causal association, the current IUR may not be
completely protective, although it is intended to be an upper bound.
• The evaluation of hazard for myeloid leukemia concluded that there is convincing evidence
from epidemiology studies to support a causal conclusion. A statistically significant
association is not required in order to select a dataset for the derivation of the IUR once a
hazard determination has been made. The observed association may have been attenuated
as a result of imprecision in the cumulative exposure metric and incomplete ascertainment
of myeloid leukemias from the death certificates. Although the risks are possibly
underestimated, these data are the best available for the derivation of the IUR for cancer.
• Combined IUR: One commenter noted that the combined tumor unit risk factor (URF) has
no precedence in final IRIS toxicological reviews and would result in difficulties in
interpretation. The comment states that EPA's assumption that the estimates of the URF are
normally distributed around the maximum likelihood estimate (MLE) with the 95% upper
confidence limit (UCL) for the URF equal to the MLE plus 1.645 times the standard error is
incorrect. The commenter argued that the assumption is incorrect because:
• The estimation procedure produces the EC or the LEC, but this does not imply that the URF
is the 95% UCL for the ratio of the extra risk level to the EC.
• There is no basis for concluding that the URF (a fixed value divided by a parameter) would
be normally distributed around a mean value, especially because the ratio must be positive
because the EC and LEC by definition must be positive.
• Stating that the approach is statistically based is wrong because the underlying assumption
that the URF estimates are normally distributed has not been shown to be true.
• Response: Although the method used is an approximation, it yields a reasonable estimate
that is strictly bounded and thus cannot be very different from the true combined risk.
0.81 IUR life tables: One commenter noted that all of the life tables for the analyses reported are not
provided in Chapter 5 of the draft review. The commenter stated that in the only life table provided, a
footnote states that the adjustment for the incidence calculation was not performed due to the small
This document is a draft for review purposes only and does not constitute Agency policy.
D-51 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
incidence rates. The commenter requests that EPA show the version of the life table that makes that
adjustment for comparison.
• Response: The result would be identical within the precision of unit risk estimates (1 or 2
significant figures).
• Study selection for selection of POD for NPC: One commenter noted that EPA uses a weak
and scientifically and methodologically challenged epidemiology study to support its point
of departure (POD) for nasal pharyngeal cancers (NPC). One commenter notes that in
general, no single study should drive the overall weight-of-evidence judgment, but in this
analysis EPA has focused on one study, Hauptmann et al. (2004), as a basis for the POD for
NPC. The commenter notes that in the Hauptmann et al. (2004) study, the majority of NPC
cases were at 1 plant of the 10 in the study, and in the other two large occupational cohorts,
there was only a single case of NPC. The commenter points out that in the Hauptmann et al.
(2004) study, workers were assigned to exposure categories based on the highest single
peak exposure experienced during their work history, which may have caused individuals
with very different exposures to be grouped into similar categories. The commenter argues
that while the risk ratios (RR) derived in the Hauptmann et al. (2004) study appear
substantial, an independent analysis of the same plant by Marsh et al. (2002) concluded that
the NPC and other pharyngeal cancers were not associated with formaldehyde. The
commenter notes that a later study by the same author associated employment in the
metal-working industry with the observed cancers (Marsh et al., 2007). The commenter
provides that EPA did not find evidence in Marsh et al. (2007) to support the same claims.
• Response: The NAS commented that EPA's selection of the data from Hauptmann et al.
(2004) to identify a POD for the derivation of an IUR was reasonable (see Comment 7.13.1).
The strengths and limitations of this study, and the resultant impact on the interpretation of
the results are discussed in Section 2.XX of the current draft assessment. The POD was
selected using the cumulative exposure models, not the peak exposure models.
• IUR calculation: Two commenters noted that EPA's use of the (3 estimates and standard
errors of the values from the NCI study to estimate both the risk of mortality and incidence
of NPC, Hodgkin lymphoma, and leukemia was inappropriate because the values from the
NCI study were based on death from the three causes (NPC, Hodgkin lymphoma, and
leukemia). One commenter noted that because the (3 estimates and standard errors are not
appropriate (and cannot be confirmed because they are not reported in the NCI
publications), the URFs estimated from these factors are suspect One commenter noted
that the survival rate from NPC is significant, but that no justification is provided for EPA's
assumption that NPC cancer mortality and incidence share the same dose-response
relationship.
• One commenter noted that one of the issues with the combined endpoint URF derivation is
footnote c from the Appendix C life table which describes ignoring the adjustment for the
all-cause hazard rate for interval'/.'
• Response: EPA used the same assumption for the derivation of the IUR for cancer based on
myeloid leukemia and NPC in the current draft assessment because incidence data were not
collected on the cohort. Because survival rates for NPC are high, use of mortality rates in
the life table analysis would have underestimated the number of cancer cases due to
lifetime formaldehyde exposure over background incidence. Therefore, EPA concluded that
This document is a draft for review purposes only and does not constitute Agency policy.
D-52 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Supplemental Information for Formaldehyde—Inhalation
the assumption that the exposure-response relationship was similar for NPC mortality and
incidence was reasonable. The rationale and discussion of uncertainty related to the
assumption are discussed in Section 2.XX.
• ADAF: One commenter noted that EPA indicates the data have met the EPA requirements
for applying age-dependent adjustment factors. The commenter noted that EPA does not
discuss the criteria as they apply to formaldehyde and the scientific defensibility of applying
ADAFs derived from data for mutagenic carcinogens to formaldehyde, which has a mixed
mode of action for which mutagenicity is only a part One commenter noted that EPA's
application of ADAFs makes the final unit risk factor more conservative and the risk factor
is based on the exposures of embalmers, which do not represent younger individuals by
their exposure patterns.
• Response: The application of the ADAFs is consistent with EPA guidance (U.S. EPA, 2005).
The ADAF adjustment is a default procedure to account for presumed increased early life
susceptibility when a mutagenic mode of action is operable and chemical-specific data are
not available. As a necessarily imprecise default approach, it is not intended to be parsed
across putative modes of action that might be operational to different and unknown extents
at different exposure levels.
• IUR: One commenter noted that the 1 in 100,000 excess risk air concentration of 0.08 ppb
based on the draft URF would not be met anywhere in the world, including locations
ranging from Alert, Nunavut, Canada to the remote South Pacific Island of Eniwetok Atoll, to
physiological concentrations in human breath. The commenter notes that outdoor
concentrations of formaldehyde would be expected to be lower than those in indoor air,
which would likely exceed the excess risk air concentration level more readily. The
commenter noted that the draft unit risk estimate would imply that neither air outdoors or
indoors is safe from a regulatory perspective.
• Response: The IRIS toxicological reviews are health assessments and the IURs are based on
the best available data to inform risk management decisions. The uncertainties associated
with the IURs are discussed in the assessment and also inform risk managers. Risk
management decisions incorporate other considerations as mandated by the statutory
authority relevant to the regulatory programs.
This document is a draft for review purposes only and does not constitute Agency policy.
D-53 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
0.0010
0.0008
0.0006
0.0004
0.0500
0.0400
/
0.0300
/
0.0200
/
O.OIOO
/
o.oooo
1
20000 40000
Flux (priwle/mm'/h)
0.0002 -
° ALM
• AMS
• MMT
• PLM
¦ PMS
— mod-1
modi
mod2
mod3
— mod4
mod5
mod 6
mod7
— mods
0.0000
~i 1 r
3000
1000 2000
Formaldehydeflux(pmol/mm2/h)
4000
Figure D-l. Various assumed dose-response curves for initiated cell division
rates (as function of formaldehyde flux to tissue). Curves differ from each
other only in the flux range 0-1,200 pmol/mm2/h. Inset shows these curves for
the full flux range needed to model bioassay data. Symbols (in gray) represent
empirically derived division rate for normal cells (see Fig. 1); no empirical data
exists on initiated cell rates.
This document is a draft for review purposes only and does not constitute Agency policy.
D-54 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
c\j _
*
ADS
A
ALM
A
AMS
V
MMT
0
PLM
~
PMS
oo _
CJ
o
D)
O
_
i/)
I
I ~:
V
A
* M
A
.-f
i»-f r
i
i
A
n 1 9
1
if £
~
i
it «
A
b °
A
A
A
A
*
"T"
0
2000 4000 6000 8000 10000
Formaldehyde Flux (pmole/mrn2/h)
12000
Figure D-2. Logarithm of replication rate for normal cells («N) versus
formaldehyde flux (in units of pmol/mm2/h) for the F344 rat nasal
epithelium. Values were derived from continuous unit length labeled data
obtained by Monticello et al. (1996), for 4-6 individual animals at all 6 nasal sites
(as shown in legend; sites are as denoted in original article) and 4 exposure
durations (13, 26, 52, 78 weeks). Each point represents a measurement for one rat,
at one nasal site, and at a given exposure time. Filled red circles: ctN values as used
in Conolly et al. (2003) after time-weighting and averaging over sites. Dashed lines:
their linear interpolation between points (short dash); their linear extrapolation for
flux value >9,340 pmol/mm2/h (long dash).
This document is a draft for review purposes only and does not constitute Agency policy,
D-55 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
** Cumulative exposure corresponding to constant lifetime exposure at or less than this p jg 3
level is attributed to roughly 93% of the person-years in the NCI cohort
Figure D-3. Estimates of extra human risk of respiratory cancer from lifetime
exposure to formaldehyde. Estimates from BBDR model runs corresponding to
eight dose response curves for initiated cell division rates (see Figure 2) compared
with various other estimates a) EPA modeling of nasopharyngeal cancer (NPC) risk
from NCI epidemiology data (Hauptmann et al. 2004); the MLE benchmark extra
risk of 0.0005 occurred at 0.15 ppm exposure concentration (EC0005); b)
multistage-Weibull statistical time-to-tumor modeling of the F344 ratbioassay data;
c) conservative upper-bound risk estimate in Conolly et al. (2004) based upon using
the hockey stick dose-response relationship for normal and initiated cells and
calculating a statistical upper bound.
This document is a draft for review purposes only and does not constitute Agency policy.
D-56 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Statistical time-to-
tumor modeling of rat
nasal tumor data;
determine POD
Use BBDR modeling to
integrate various data,
determine POD, effect
of model uncertainty
I
Default extrapol. to human,
including nose and lung based
on tissue flux
Y
Default extrapol. to human
nose lung based on tissue flux
Figure D-4. Decision tree for the use of mechanistic data and BBDR modeling
[Abbreviations: extrapol. = extrapolation; u-v = uncertainty-variability;
POD=point of departure].
This document is a draft for review purposes only and does not constitute Agency policy.
D-57 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 APPENDIX E. SUMMARY OF PUBLIC COMMENTS
2 AND EPA'S DISPOSITION.
3 E.l. INSERT APPENDIX E HERE
This document is a draft for review purposes only and does not constitute Agency policy.
E-l DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information for Formaldehyde—Inhalation
APPENDIX F. SYSTEMATIC EVIDENCE MAP
UPDATING THE LITERATURE FROM 2016-2021
F.l. INTRODUCTION
This systematic evidence map (SEM) updates the literature that was assessed to develop the
2017 Step 1 draft IRIS formaldehyde-inhalation assessment. The completed draft 2017 IRIS
assessment was suspended by EPA f https://www.epa.gov/sites/default/files/2 019-
04/documents/iris program outlook apr2019.pdf) and shared with EPA's OCSPP-OPPT program
for use in developing a risk evaluation under TSCA. However, in 2021, development of the IRIS
assessment was unsuspended (https://www.epa.gov/sites/default/files/2021-
03 /documents/iris program outlook mar2021.pdf). This SEM was developed to identify the
relevant literature published since the suspension of the 2017 draft, in particular studies that may
alter hazard or toxicity value conclusions presented in the 2017 draft Studies identified in this SEM
as possibly impactful to the 2017 draft conclusions have been incorporated into the updated 2021
draft IRIS Toxicological Review.
F.2. METHODS
This SEM identifies and documents the literature relevant to assessing the potential human
health hazards of formaldehyde inhalation from January 2016-May 2021. The search terms and
screening strategies are nearly identical (exceptions noted later in this document) to those used to
develop the 2017 Step 1 draft, and the detailed methods can be found in the Supplemental
Information to the Toxicological Review of Formaldehyde - Inhalation (see Appendix A.5). In
Appendix A.5. supporting materials for each health effect include tables listing the search terms for
each bibliographic database searched, and tables listing the inclusion and exclusion criteria used to
search and screen the identified citations fPECOl.
F.2.1. Specific Aims
The following specific aims were identified for the SEM.
• Identify epidemiological (i.e., human), toxicological (i.e., experimental animal), and
mechanistic literature using an identical literature search approach as was used to develop
the 2017 Step 1 draft IRIS formaldehyde-inhalation assessment reporting effects of
exposure to formaldehyde as outlined in the health effect-specific PECOs found in Appendix
A.5 of the Supplemental Information to the Toxicological Review of Formaldehyde -
Inhalation.
• Tag secondary (not primary research) studies.
This document is a draft for review purposes only and does not constitute Agency policy.
F-l DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 • Create a literature inventory of PECO-relevant studies. The literature inventory summarizes
2 basic features of study design, health system(s), and endpoints assessed.
3 • Assess PECO-relevant studies, within each health effect category, to determine if they are
4 possibly impactful to the 2017 draft assessment decisions on hazard and dose response and
5 document the reasons in a literature inventory.
6 F.2.2. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria and
7 Supplemental Material Tagging
8 A PECO is used to focus the research question(s), search terms, and inclusion/exclusion
9 criteria used in a SEM or systematic review. For this SEM, health effect-specific PECOs were used
10 for the literature search and screening process and can be found in Appendix A. 5 of the
11 Supplemental Information to the Toxicological Review of Formaldehyde - Inhalation. For each
12 health effect, the PECOs list the different populations and endpoints of interest In addition, PECOs
13 tailored to mechanistic studies were used—these also are found in Appendix A.5 of the
14 Supplemental Information to the Toxicological Review of Formaldehyde - Inhalation. The PECO for
15 lymphohematopoietic (LHP) cancer in animal studies is provided below as an example (Table 1).
16 In addition to identifying studies that met the PECO criteria and studies that were excluded,
17 tags were added to nonprimary research studies (i.e., reviews, commentaries, letters, etc).
Table F-l. Example of outcome-specific PECO: LHP cancer in animals
PECO element
Description
Populations
Animal: Nonhuman mammalian animal species (whole organism) of anv lifestage (including
preconception, in utero, lactation, peripubertal, and adult stages).
In-vitro assays and non-experimental animal studies are excluded.
Exposures
Relevant forms:
Formaldehyde (generated from formalin, paraformaldehyde, or other sources)
•
• Animal: Anv exposure to formaldehyde via inhalation routefsl of >1 dav duration, or anv
duration assessing exposure during reproduction or development.
•
• Non-inhalation dosing regimens are excluded for systemic effects (in this SEM).
Comparators
Animal: A concurrent control group exposed to vehicle-only treatment and/or untreated control
(control could be a baseline measurement).
Outcomes
LHP cancers.
This document is a draft for review purposes only and does not constitute Agency policy.
F-2 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
F.2.3. Literature Search and Screening Strategies
Database Searches
To identify relevant studies published since the 2017 draft was developed, separate
searches were conducted for the health effect categories listed in Table 2 encompassing January
2016 to May 2021 (overlapping with the search dates of the 2017 draft). Separate searches across
two databases were conducted for different health outcomes (e.g., sensory irritation, cancer). In
addition to the health effects listed in Table 2, specific search strategies were used to identify
literature on additional topics (e.g., toxicokinetics and mechanistic information related to
respiratory tract cancers and LHP cancers). While the searches for cancer mechanisms primarily
focused on genotoxicity endpoints, the searches for mechanistic research on inflammation and
immune effects and respiratory pathology retrieved studies also relevant to cancer. While earlier
literature updates included a search strategy on exposure to formaldehyde, this research category
was not updated for this search as exposure is not a review topic for the assessment
The search strategies are identical to those used to develop the 2017 Step 1 draft, which used
PubMed, Web of Science andToxNet, although this update did not include ToxNet, which has not been
available since December 2019. Details on the database searches can be found in the Appendix A.5 of
the Supplemental Information to the Toxicological Review of Formaldehyde - Inhalation.
Table F-2. Literature search strategy
Databasesa
Health hazard searchesb
Web of Science
PubMed
(formaldehyde, formalin, paraformaldehyde, OR CASN 50-00-0) AND:
Sensory Irritation0
• Pulmonary Function0
• Immune-Mediated Conditions, focusing on Allergies and Asthma
• Respiratory Tract Pathology in Humans
• Respiratory Tract Pathology in Animals
• Site-specific cancer in Humans
• Upper Respiratory Tract Cancer in Animals
• Lymphohematopoietic Cancer in Animals
• Mechanistic Studies of Upper Respiratory Tract Cancer, focusing on genotoxicity
• Mechanistic Studies of Lymphohematopoietic Cancer, focusing on genotoxicity
• Inflammation and Immune Effects (mechanistic information)01
• Developmental and Reproductive Toxicity
• Nervous System Effects
aPubMed: http://www.ncbi.nlm.nih.gov/pubmed/, Web of Science:
http://apps.webofknowledge.com/WOS GeneralSearch input.do?product=WOS&search mode=.
Specific parameters and keywords for each hazard-specific database search strategy are included in Appendix A.5
of the Supplemental Information to the Toxicological Review of Formaldehyde - Inhalation.
CA systematic search strategy was not applied to the database of animal studies on this health outcome. Sensory
irritation in animals is a well-described phenomenon. For pulmonary function, there was an extensive set of
research studies on humans, and therefore, the few studies on this endpoint in animals were not reviewed.
This document is a draft for review purposes only and does not constitute Agency policy.
F-3 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information for Formaldehyde—Inhalation
dThis separate, systematic literature search was performed to augment the analyses of mechanisms relevant to
other health effect-specific searches.
Screening Process
Studies identified from the database searches were imported into DistillerSR software
(https://www.evidencepartners.com/products/distillersr-systematic-review-software/) for
screening. Both title/abstract (TIAB) and full-text screening were conducted by two independent
reviewers and any screening conflicts were resolved by discussion between the primary screeners
with consultation by a third reviewer if needed. Conflicts between screeners in applying the
supplemental tags were resolved similarly, erring on the side of over-tagging. For citations with no
abstract, articles were initially screened based on all or some of the following: title relevance (title
should indicate clear relevance), and page numbers (articles two pages in length or less are
assumed to be conference reports, editorials, or letters). Eligibility status of non-English studies
was assessed using the same approach with online translation tools or engagement with a native
speaker used to facilitate screening. Full-text records were sought through the EPA's HERO
database for studies screened as meeting PECO criteria or "unclear" based on the TIAB screening. In
addition, references that had potential relevance to other health-outcome specific projects were
identified and then screened within those projects. Access to the example screening form
DistillerSR is available upon request for users who have DistillerSR access.
Although some uncertainties remain, the organization and analyses in the assessment
assume that inhaled formaldehyde is not distributed to an appreciable extent beyond the upper
respiratory tract to distal tissues; thus, it is assumed that inhaled formaldehyde is not directly
interacting with tissues distal to the portal of entry (POE) to elicit systemic effects. Therefore, as a
deviation from the literature screening approach applied to develop the 2017 draft, studies of
exposure routes not involving inhalation, including in vitro studies involving cells from distal
tissues, were not considered to be PECO relevant for this literature update and were excluded.
Similarly, it is assumed that formaldehyde does not cause appreciable changes in normal metabolic
processes associated with formaldehyde in distal tissues. Thus, studies examining potential
associations between levels of formaldehyde (i.e., endogenous formaldehyde) or formaldehyde
metabolites in tissues distal to the POE (e.g., formate in blood or urine, brain formaldehyde levels)
were excluded for most health outcomes, particularly effects on systemic tissues such as the
nervous system and reproductive and developmental effects. However, studies of endogenous
formaldehyde and mechanisms with its potential relevance to circulating hematopoietic precursor
cells and lymphohematopoietic cancers were considered.
F.2.4. Literature Inventory
Human, animal, and mechanistic studies that met PECO criteria after full-text review were
briefly summarized in DistillerSR using a structured data extraction form. Studies were extracted
by one team member and the extracted data were quality checked by at least one other team
This document is a draft for review purposes only and does not constitute Agency policy.
F-4 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information for Formaldehyde—Inhalation
member. The extraction fields in the forms are available in MS Excel format upon request See
fhttps://www.epa.gov/iris/forms/contact-us-about-iris! for requestors who have DistillerSR
access. The literature inventories were exported from DistillerSR in MS Excel format.
For animal studies, the following information was captured: formaldehyde source, study
type (e.g., acute, chronic, developmental), duration of treatment, route, species, strain, sex, exposure
levels tested, exposure units, and endpoints assessed.
For epidemiological studies, the following information was summarized: population type
(e.g., residential/school based, occupational, other), study design (e.g., cross-sectional, cohort, case-
control, ecological, case-report, controlled trial, meta-analysis), study location, lifestage (adults,
children/infants), exposure measurement (air sampling occupational history, other), and
endpoints assessed.
For mechanistic studies, the information gathered was dependent on the study type: human
in vivo, animal in vivo, in vitro/ex vivo, or dosimetry/pharmacokinetic modeling. For
dosimetry/pharmacokinetic modeling references, a summary from the paper's abstract was
excerpted. For all types of mechanistic studies, study details and exposure metrics were
summarized along with the endpoints assessed.
The inventory also includes a decision and explanation as to whether each relevant study is
considered "possibly impactful" (i.e., to the 2017 draft assessment conclusions) or "not impactful,"
as described below.
Considerations for identifying "possibly impactful" studies
Studies that met the PECO criteria after full text screening were further examined to
determine if they could potentially be impactful to the assessment with respect to changing hazard
conclusions or toxicity values presented in the 2017 draft This process relied on information
gathered from the literature inventory and expert judgment by two reviewers. General
considerations for designating studies as possibly impactful are included below, with the specific
rationales documented in the SEM study summary tables:
• Studies with chronic or subchronic exposure durations or including exposure during
reproduction or development, are generally more impactful than studies with acute or
shorter-term exposure durations (e.g., <4 weeks in rodent studies).
• Animal studies with multiple dose groups covering a broad range of dose levels, and
specifically including lower exposure levels, are generally more impactful than single-dose
studies.
• Animal studies employing exposure to formaldehyde without methanol co-exposure (e.g.,
generated from paraformaldehyde) and with adequate inhalation exposure administration
methods were considered more impactful. Methanol, present in aqueous formaldehyde
solutions to inhibit polymerization, is a potential confounder of associations between
observed health outcomes and formaldehyde exposure via formalin. The test article used to
This document is a draft for review purposes only and does not constitute Agency policy.
F-5 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information for Formaldehyde—Inhalation
generate the formaldehyde atmosphere and controls in experimental studies was an
important consideration, particularly for non-respiratory health effects.
• More apical endpoints and those most directly related to the mechanistic uncertainties
identified in the 2017 draft as most relevant to drawing hazard or dose-response judgments
were considered more impactful. The specifics of this consideration vary depending on the
health outcome(s) of interest In some cases, this relevance determination relates to the
potential human relevance of the endpoints, while in others this relates to an ability to infer
adversity.
• For human studies, prioritization considerations depended on the health effect category,
formaldehyde exposure levels, and the extent of the evidence base supporting the hazard
conclusions in the 2017 draft. Studies of noncancer respiratory outcomes identified in the
PECOs among residential populations or school-aged children were prioritized over
occupational studies, which typically involve higher formaldehyde concentrations. Any
study of reproductive or developmental outcomes that conducted an exposure assessment
(qualitative or quantitative) for formaldehyde was considered possibly impactful. In
addition, with some exceptions documented in the inventory tables, studies of ALS,
genotoxicity endpoints, or PECO identified cancer outcomes that conducted an exposure
assessment (qualitative or quantitative) for formaldehyde were generally considered
possibly impactful.
This document is a draft for review purposes only and does not constitute Agency policy.
F-6 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3. RESULTS
2 F.3.1. Sensory Irritation Effects in Human Studies
®
Met PECO
©
Sensory irritation Excluded
©
Not primary research
(supplemental)
Figure F-l. Sensory irritation literature tree (interactive version here).
3 A total of 121 citations were retrieved for the assessment of sensory irritation in humans
4 and five studies were PECO-relevant (Table 3). None of these were deemed to be possibly impactful.
5 Saowakon etal. (2015) already had been included in the 2017 draft.
This document is a draft for review purposes only and does not constitute Agency policy.
F-7 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-3. Studies of sensory irritation effects in humans
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Humans
Aung et al.
(2021)
Occupational
Myanmar
cross-sectional
Air sampling, adults,
medical students and
instructors in anatomy
dissection rooms
Unpleasant odor, eye irritation,
nasal irritation symptoms
Not
impactful
High exposure levels, adults,
health effects well supported
in assessment
Deng et al.
(2020)
only abstract
available (full
text Chinese)
Occupational
China
cross-sectional
Air sampling, adults,
medical students in
anatomy dissection
rooms
Subjective symptoms (e.g., itchy
eyes, nasal congestion, runny
nose)
Not
impactful
High exposure levels, adults,
health effects well supported
in assessment
Sakellaris et al.
(2020)
Occupational
Europe (Portugal, Spain,
Italy, Greece, France,
Hungary, the
Netherlands, Finland)
cross-sectional
Air sampling, adults,
office building occupants
Eye irritation (dry eyes, watering
or itchy eyes, burning or
irritated eyes), respiratory
symptoms (blocked or stuffy
nose, runny nose, dry/irritated
throat, cough
Not
impactful
Adults, health effects well
supported in assessment
Saowakon et al.
(2015)
Not extracted
Not
impactful
Already identified in 2017
draft
Thetkathuek et
al. (2016)
Occupational,
Chacheongsao Province,
Thailand
cross-sectional
Air sampling, adults,
medium-density
fiberboard furniture
workers
Respiratory irritation symptoms
Not
impactful
High exposure levels, adults,
health effects well supported
in assessment
Rows for studies judged as "not impactful" are shaded grey.
This document is a draft for review purposes only and does not constitute Agency policy.
F-8 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.2. Pulmonary Function Effects in Human Studies
6 » ©
MetT^CO Possibly impactful
Pulmonary function
0
Not primary research
(supplemental)
Figure F-2. Pulmonary function effects in humans literature tree (interactive
version here).
2 A total of 30 citations were retrieved for the assessment of pulmonary function effects in
3 humans and six studies were PECO-relevant (Table 4). Of these, one study, Saowakon etal. (2015).
4 was deemed to be possibly impactful but already had been included in the 2017 draft.
This document is a draft for review purposes only and does not constitute Agency policy.
F-9 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-4. Studies of pulmonary function effects in humans
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Human
Saowakon et al.
(2015)
Not extracted
Possibly
impactful
Already identified in 2017
draft
Fsadni et al. (2018)
Schools-based
Malta
cross-sectional
Air sampling, children,
school children
Pulmonary function tests (not
specified)
Not impactful
Important details were not
provided
Asgedom et al. (2019)
Occupational
Ethiopia
cross-sectional
Air sampling, adults,
particleboard workers
Lung function (FVC, FEVI, FEF 25-
75%)
Not impactful
High exposure levels,
adults, health effects well
supported in assessment
Deng et al. (2020)
only abstract available
(full text Chinese)
Occupational
China
cross-sectional
Air sampling, adults,
medical students in
anatomy dissection
rooms
FEVI, FEVI/FVC, PEF, FEF 25%-75%,
MEF25%, FEF50%-75%
Not impactful
High exposure levels,
adults, health effects well
supported in assessment
Neghab et al. (2017)
Occupational
Shiraz, Iran
cross-sectional
Air sampling, adults,
kitchen workers
exposed to cooking
fumes
VC, FVC, FEVI, PEF, FEV1/FVC,
FEV1/VC
Not impactful
High exposure levels,
adults, health effects well
supported in assessment
Zarei et al. (2017)
Occupational
Tehran, Iran
cross-sectional
Air sampling, adults,
foundry coremakers
FVC, FEVI, FEV1/FVC, peak
expiratory flow (PEF), mid forced
expiratory flow (FEF25-75%)
Not impactful
High exposure levels,
adults, health effects well
supported in assessment
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
FEF25 -75% ¦ mid forced expiratory flow, FEFso-75% ~ forced expiratory flow 50-75%/ FEVi- Forced expiratory volume in one second, FVC — forced vital capacity, PEF -
peak expiratory flow, MEF25% - mean flow at 25%, VC -vital capacity.
This document is a draft for review purposes only and does not constitute Agency policy.
F-10 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F.3.3. Immune-Mediated Conditions in Humans, Focusing on Allergies and Asthma
16 @
Met PECO Possibly impactful
©
Not primary research
(supplemental)
Figure F-3. Asthma and immune effects in humans literature tree (interactive
version here).
A total of 1,597 citations were retrieved for the assessment of asthma and immune effects in
humans and 16 studies were PECO-relevant (Table 5). Of these, 11 studies were deemed to be
possibly impactful.
This document is a draft for review purposes only and does not constitute Agency policy.
F-ll DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-5. Studies of immune-mediated conditions in humans, focusing on allergies and asthma
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Human
Branco et al.
(2020)
School-based
Northern Portugal
cross-sectional
Air sampling, children,
preschoolers/primary
school students
Asthma (reported, diagnosed),
wheezing (active)
Possibly
impactful
School-based - children; indoor
formaldehyde concentrations between
10-80 ng/m3
Huang et al.
(2017)
Population-based
Shanghai, China
case-control
Air sampling in residence,
children
Current rhinitis
Possibly
impactful
Population-based - children; indoor
formaldehyde concentrations between
10-80 ng/m3
Isa et al.
(2020a)
School-based
Selangor, Malaysia
cross-sectional
Air sampling in classroom,
children
Rhinitis (past 12 months), skin
allergy (past 12 months)
Possibly
impactful
School-based - children; mean indoor
formaldehyde concentrations between
10-80 ng/m3
Laioie et al.
(2014)
Population-based
Quebec, Canada
intervention study
Air sampling, children,
ventilation intervention
study
Change in prevalence of
asthma symptoms and
medical care
Possibly
impactful
Population-based - children; mean
indoor formaldehyde concentrations
between 10-80 ng/m3
Li et al. (2019)
Population-based
Hong Kong
cohort
Air sampling, birth to 18
mo
Wheeze (new onset)
Possibly
impactful
Population-based - children; mean
indoor formaldehyde concentrations
between 10-80 ng/m3
Liu et al.
(2018a)
Population-based
Changchun, China
case-control
Air sampling in residence,
children
Asthma diagnosis
Possibly
impactful
Population-based - children; indoor
formaldehyde concentrations between
10-80 ng/m3
Madureira et
al. (2016)
Population-based
Porto, Portugal
case-control
Air sampling in residence,
children
Current asthma
Possibly
impactful
Population-based - children; indoor
formaldehyde concentrations between
10-80 ng/m3
Neamtiu et al.
(2019)
School-based
Alba County,
Romania
cross-sectional
Air sampling in classroom,
children
Asthma-like symptoms
(difficult breathing, dry cough,
wheezing in past week),
allergy-like symptoms (skin
conditions such as rash, itch,
eczema; eye disorders such as
red, dry, swollen, itching,
burning, or sensation of "sand
in eyes"; rhinitis such as
itching nose, sneezes, stuffy or
blocked nose)
Possibly
impactful
School-based - children; mean indoor
formaldehyde concentrations between
10-80 ng/m3
This document is a draft for review purposes only and does not constitute Agency policy.
F-12 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Norback et al.
School-based
Air Sampling, children
Rhinitis
Possibly
School-based - children; indoor
(2017)
Johor Bahru,
Malaysia cross-
sectional
impactful
formaldehyde concentrations between
10-80 ng/m3
Yon et al.
School-based
Air sampling in classroom,
Current asthma, rhinitis,
Possibly
School-based - children; mean indoor
(2019)
Seongnam City,
Korea cohort
children
rhinitis severity
impactful
formaldehyde concentrations between
10-80 ng/m3
Yu et al. (2017)
Population-based
Air sampling in residence,
Wheeze (new onset)
Possibly
Population-based - children; mean
Hong Kong
birth to 18 mo
impactful
indoor formaldehyde concentrations
cohort
between 10-80 ng/m3
Asgedom et al.
Occupational
Air sampling, adults,
Respiratory symptoms (cough,
Not
Occupational exposure - adults, health
(2019)
Ethiopia
cross-sectional
particleboard workers
cough with sputum
production, phlegm,
wheezing, shortness of
breath)
impactful
effects well supported in assessment
Dumas et al.
Occupational
Occupational history and
Self-reported incident
Not
Occupational exposure - adults, health
(2020)
United States
cohort
job-task-exposure-matrix,
adults, health workers
(female nurses)
physician-diagnosed asthma
impactful
effects well supported in assessment
El-Fekv et al.
Occupational
Industry/ production
Chronic bronchitis, respiratory
Not
Occupational exposure - adults, health
(2020)
Egypt
cross-sectional
type, adults, factory
workers
symptoms and signs,
respiratory rate, nasal
symptoms, eye symptoms,
skin manifestations
impactful
effects well supported in assessment
Fsadni et al.
School-based
Air sampling in classroom,
Wheezing, rhinitis, eczema,
Not
Only qualitative results reported, e.g.,
(2018)
Malta
cross-sectional
children
acoustic rhinometry, nasal
lavage
impactful
whether statistically significant and
directional arrow
Thetkathuek et
Occupational
Air sampling, adults,
Difficulty breathing, chest
Not
Occupational exposure - adults, health
al. (2016)
Chacheongsao
Province, Thailand
cross-sectional
medium density
fiberboard workers
discomfort, wheeze
impactful
effects well supported in assessment
Rows for studies judged as "not impactfu
1" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
F-13 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.4. Respiratory Tract Pathology in Human Studies
©
Met PECO
Respiratory tract pathology Excluded
human
©
Not primary research
(supplemental)
Figure F-4. Human respiratory tract pathology literature tree (interactive
version here).
2 A total of 5 79 citations were retrieved for the assessment of respiratory tract pathology in
3 humans and one study was PECO-relevant (Table 6). This study was not deemed to be possibly
4 impactful.
This document is a draft for review purposes only and does not constitute Agency policy.
F-14 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-6. Studies of respiratory tract pathology in humans
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Human
Bruno et al.
(2018)
Occupational
Rome, Italy
cross-sectional
Air sampling, adults,
Laboratory pathology
workers
Nasal cytology (muciparous
metaplasia)
Not impactful
Adults, health effects well
supported in assessment
Rows for studies judged as "not impactful" are shaded grey.
This document is a draft for review purposes only and does not constitute Agency policy.
F-15 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.5. Animal Studies of Respiratory Tract Pathology
10
Met PECO
O
Possibly impactful
Respiratory tract pathology
animal
Excluded
©
Not primary research
(supplemental)
Figure F-5. Animal respiratory tract pathology literature tree (interactive
version here).
2 A total of 352 citations were retrieved for the assessment of respiratory tract pathology in
3 animals and ten studies were PECO-relevant (Table 7). Of these, one (NTP. 20171 was deemed to be
4 possibly impactful. Although NTP (20171 was identified in the literature search update and
5 included in the inventory, it already had been included in the 2017 draft Toxicological Review of
6 Formaldehyde-Inhalation.
This document is a draft for review purposes only and does not constitute Agency policy.
F-16 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-7. Animal studies of respiratory tract pathology
Reference
Study design
Exposurea
Endpoints
Impact
Rationale
Animal Studies
NTP (2017)
Mouse (Trp53
haploinsufficient),
Male
Subchronic (8 wk; 6
h/d, 5 d/wk), then held
for 32 wk
Paraformaldehyde
0, 7.5 or 15 ppm (0, 9.2,
18.5 mg/m3)
Inhalation
All major tissues and gross lesions
were collected for histopathology
(including squamous metaplasia in
respiratory tissues)
Possibly
impactful
Already included in 2017
draft
Avdemir et
al. (2017)
Rat (Wistar), both
sexes
Subchronic (6 wk;
8h/d, 5d/wk)
Formalin
0, 6 ppm (0, 7.38 mg/m3)
Inhalation
Lung hematoxylin and eosin
staining for qualitative review of
inflammation and tissue
morphology
Not impactful
Formalin
Cheng et al.
(2016)
Mouse (Kunming),
male
Short-term (up to 7 d;
continuous)
Formalin
0,0.08, 0.8 mg/m3
Inhalation
Hematoxylin and eosin staining for
inflammation and edema
Not impactful
Formalin; not key
endpoints
Abreu et al.
(2016)
Mouse (C57BL/6), both
sexes
Acute (8 h)
Unspecified test article
0, 0.2, 1.0, 3.0 ppm (0,
0.25, 1.23, 3.69 mg/m3)
Inhalation
Lung morphology and nasal
ciliation; histological inflammatory
cell counts in lung and scoring in
nose
Not impactful
Unknown test article;
acute
Lima et al.
(2015)
Rat (Fischer), male
Short-term (5 d; 20-
min x3/d)
Unspecified test article
0,1, 5,10%
Inhalation
Trachea histology and
morphometric analyses, including
mucus production
Not impactful
Unknown test article; high
levels; brief exposures
Liu et al.
(2018b)
Rat (Sprague Dawley),
male
Short-term (4 wk; 8
h/d)
Formalin
0,0.5, 5,10 mg/m3
Inhalation
Lung histopathological architecture
measurements
Not impactful
Formalin; not key
endpoints
Pavani et al.
(2019)
Rat (Wistar), male
Short-term (21 d; 1
h/d)
Unspecified test article
0,40%
Inhalation (vapor)
Pulmonary histopathology
Not impactful
Unknown test article; high
levels; brief exposures
This document is a draft for review purposes only and does not constitute Agency policy.
F-17 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposurea
Endpoints
Impact
Rationale
Saomaz et al.
(2017)
Rat (Sprague Dawley),
male
Short-term (4 wk; 8
h/d) or Subchronic (13
wk; 8 h/d)
Paraformaldehyde 0, 5,
10 ppm (0, 6.2, 12.3
mg/m3)
Inhalation
Hematoxylin and eosin staining
(airway inflammation; morphology;
scored injury); trachea thickness
Not impactful
Not key endpoints
Sholapuri et
al. (2020)
Rat (Wistar), male
Short-term (21 d; 1
h/d)
Formalin
0,40%
Inhalation
Lung histopathology
Not impactful
Formalin; high levels; brief
exposures
Song et al.
(2017)
Mouse (Balb/c), male
Short-term (18 d; 3h/d)
Formalin
0, 2.44 ppm (0, 3.00
mg/m3)
Inhalation
Airway inflammation histology
Not impactful
Formalin; No
formaldehyde-only control
(without ovalbumin [OVA])
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
a Use of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
This document is a draft for review purposes only and does not constitute Agency policy.
F-18 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.6. Site-specific Cancer in Human Studies
© 0
Met PECO Possibly impactful
Human cancer Excluded
Not primary research
(supplemental)
Figure F-6. Human cancer literature tree (interactive version here).
2 A total of 1,555 citations were retrieved for the assessment of cancer in humans and six
3 studies were PECO-relevant (Table 8). Of these, half (three studies) were deemed to be possibly
4 impactful. Checkowav et al. (2015) and Piraetal. (2014) had been included in the 2017 draft.
This document is a draft for review purposes only and does not constitute Agency policy.
F-19 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-8. Studies of site-specific cancer in humans
Reference
Study Design
Exposure
Endpoints
Impact
Rationale
Human
Checkowav et al.
(2015)
Occupational
United States
cohort
Air sampling, occupational
history, and job-exposure matrix,
adults, NCI cohort reanalysis
Cause-specific mortality [non-Hodgkin
lymphoma mortality, chronic
lymphocytic leukemia mortality, Hodgkin
lymphoma mortality, multiple myeloma
mortality, myeloid leukemia mortality,
acute myeloid leukemia mortality,
chronic myeloid leukemia mortality, all
leukemia mortality,
lymphohematopoietic cancer mortality]
Possibly
impactful
Already identified in 2017
draft
Marsh et al.
Occupational
United States
cohort
Air sampling, occupational
history, and job-exposure matrix,
adults, NCI cohort NPC reanalysis
Nasopharyngeal cancer mortality
Possibly
impactful
Additional analyses of
important studies in the 2017
draft
(2016)
Mohner et al.
Occupational
United States
cohort
Occupation-based, adults, NCI
cohort analysis
Mortality from nasopharyngeal cancer
[oropharynx, nasopharynx,
hypopharynx, pharynx, pharynx
(unspecified)]
Possibly
impactful
Additional analyses of
important studies in the 2017
draft
(2019)
Pira et al. (2014)
Occupational
Piedmont, Italy
cohort
Occupational history, adults,
laminated plastics workers
Cause-specific mortality [lymphoma,
myeloma, leukemia, all lymphatic and
hematopoietic tissue neoplasms]
Not impactful
Already identified in 2017
draft
Sernia et al.
Occupational
Italy
cohort
Current occupation, adults,
university laboratory workers
NPC, leukemia/lymphoma
Not impactful
Inadequate exposure
assessment and study results
do not add novel findings to a
health effect that is well
supported in the assessment
(2016)
Xie et al. (2017)
General
population
Hong Kong
case-control
Occupational history and
industrial code, self-report, adults
Nasopharyngeal carcinoma incidence
Not impactful
Inadequate exposure
assessment and study results
do not add novel findings to a
health effect that is well
supported in the assessment
Rows for studies judged as "not i4mpactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
F-20 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.7. Animal Studies of Respiratory Tract Cancer
© Q
Met PECO Possibly impactful
E
Respiratory tract cancer animal Excluded
(21
Not primary research
(supplemental)
Figure F-7. Animal respiratory tract cancer literature tree (interactive version
here).
2 A total of 705 citations were retrieved for the assessment of respiratory tract cancers in
3 animals and two studies were PECO-relevant (Table 9). Of these, one was deemed possibly
4 impactful. This study NTP (20171 was identified in the literature search update and included in the
5 inventory although ithad been included in the 2017 draft Toxicological Review of Formaldehyde-
6 Inhalation.
This document is a draft for review purposes only and does not constitute Agency policy.
F-21 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-9. Animal studies of respiratory tract cancers
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Animal Studies
NTP (2017)
Mouse (Trp53
haploinsufficient),
Male
Subchronic (8 wk; 6
h/d, 5 d/wk), then
held for 32 wk
Paraformaldehyde
0, 7.5 or 15 ppm (0, 9.2,
18.5 mg/m3)
Inhalation
Blood was collected for
hematology, and major tissues
and gross lesions were
collected for histopathology
(nasal and LHP cancer, and
respiratory lesions)
Possibly impactful
Already included in
2017 draft
Soffritti et al.
(2016)
Rat (SD), both sexes
Chronic (continuous
exposure from 6 -
104 weeks of age)
Unspecified test article
0, 50 ppm
Oral (drinking water)
Carcinogenicity study
(presumed to include
evaluation of nasal/URT
tumors)
Not impactful
Oral exposure; high
levels; formalin (note:
would be screened as
excluded, but
inventoried due to
rarity of chronic
exposure duration
studies of cancer)
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
F-22 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F.3.8. Animal Studies of Lymphohematopoietic Cancers
© ©
Met PECO Possibly impactful
:
Lymphohematopoietic cancer Excluded
animal
©
Not primary research
(supplemental)
Figure F-8. Animal lymphohematopoietic cancer literature tree (interactive
version here!
A total of 66 citations were retrieved for lymphohematopoietic cancers in animals and two
studies were PECO-relevant (Table 10). Of these, one was deemed possibly impactful. NTP (20171
was identified in the literature search update and included in the inventory although it had been
included in the 2017 draft Toxicological Review of Formaldehyde-Inhalation.
This document is a draft for review purposes only and does not constitute Agency policy.
F-23 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-10. Animal studies of lymphohematopoietic cancer
Reference
Study design
Exposure
Endpoints
Impact
Rationale
Animal Studies
NTP (2017)
Mouse (Trp53
haploinsufficient),
Male
Subchronic (8 wk; 6
h/d, 5 d/wk), then
held for 32 wk
Paraformaldehyde
0, 7.5 or 15 ppm (0, 9.2,
18.5 mg/m3)
Inhalation
All major tissues and gross
lesions were collected for
histopathology (including LHP
tissues)
Possibly impactful
Already included in
2017 draft
Soffritti et al.
(2016)
Rat (SD), both sexes
Chronic (continuous
exposure from 6 -
104 weeks of age)
Unspecified test article
0, 50 ppm
Oral (drinking water)
Carcinogenicity study
(presumed to include
evaluation of nasal/URT
tumors)
Not impactful
Oral exposure; high
levels; formalin (note:
would be screened as
excluded, but
inventoried due to
rarity of chronic
exposure duration
studies of cancer)
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
F-24 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.9. Mechanistic Studies of Inflammation and Immune-Related Responses
© ®
Met PECO Possibly impactful
Mechanistic inflammation Excluded
Q
Not primary research
(supplemental)
Figure F-9. Mechanistic inflammation and immune effects literature tree
(interactive version here).
2 A total of 1,411 citations were retrieved for the assessment of mechanistic information on
3 inflammation and immune responses (in the respiratory system or at systemic sites) and 56 studies
4 were PECO-relevant (Table 11). Of these, eight were deemed to be possibly impactful (note: one
5 possibly impactful study is repeated under both the animal and in vitro/ex vivo sections). NTP
6 f20171 was identified in the literature search update and included in the inventory table although it
7 had been included in the 2017 draft Toxicological Review of Formaldehyde-Inhalation. In Vitro/Ex
8 Vivo designs and a study of endogenous formaldehyde biology also were included.
This document is a draft for review purposes only and does not constitute Agency policy.
F-25 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-ll. Mechanistic studies relating to respiratory or systemic inflammatory and immune responses
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Human Studies
Bassig et al.
(2016)
Occupational
Guangdong, China
Cross-sectional
Air sampling
Adult formaldehyde
factory workers
WBC counts in blood, with subtype analyses
of cells of both myeloid and lymphoid
lineage (include CD4 T cell subtyping and
cell activation markers)
Possibly
impactful
PBL sub-population
analyses and
lineage studies are
important
endpoints
Costa et al.
(2019)
Occupational
Portugal
Cross-sectional
Air sampling
Adult anatomy-
pathology laboratory
workers
Lymphocyte counts, subpopulations
analyses
Possibly
impactful
PBL sub-population
analyses and
lineage studies are
important
endpoints
Augenreich et al.
(2020)
Occupational
Boone, North
Carolina, USA
Cohort
Air sampling
Adult medical students
in anatomy dissection
rooms
Circulating markers of oxidative stress and
inflammation; brachial artery dilation (arm),
reactive hyperemia (leg), blood
pressure/pulse/heart rate
Not
impactful
ROS measures are
not key endpoints
Bellisario et al.
(2016)
Occupational Torino,
Italy
cross-sectional
Air sampling, adults,
Female surgical nurses
Biomarkers of oxidative stress (urinary
malondialdehyde and 15-F2t-isoprostane)
Not
impactful
ROS markers are
not key endpoints
Bruno et al.
(2018)
Occupational
Rome, Italy
Cross-sectional
Air sampling
Adult pathology
laboratory workers
Counts of neutrophils, eosinophils,
lymphocytes, macrophages, ratio of
mucous-secreting cells and ciliated cells in
the middle portion of the inferior turbinate
Not
impactful
Cell counts
(without sub-
analyses) are not
key endpoints
Ghelli et al.
(2020)
Occupational
Turin, Italy
Cohort
Air sampling
Adult (female) hospital
workers
ROS measures in urine and inflammatory
markers and cytokines in blood. Genotyped
for CYP1A1, GSTT1, GSTM1, TNFa, and IL-6
polymorphisms
Not
impactful
ROS and cytokine-
related measures
are not key
endpoints
Isa et al. (2020a)
School-based
Selangor, Malaysia
Cross-sectional
Air sampling
School children
Fractional exhaled nitric oxide (FeNO, an
airway ROS/inflammation marker)
Not
impactful
ROS markers are
not key endpoints
Isa et al. (2020b)
School-based
Hulu Langat,
Selangor, Malaysia
Air sampling, children,
Suburban and urban
school children
Inflammatory cytokine markers in sputum;
exhaled FeNO
Not
impactful
ROS and cytokine-
related measures
This document is a draft for review purposes only and does not constitute Agency policy.
F-26 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Cross-sectional
are not key
endpoints
Yon et al. (2019)
School-based
Seongnam City,
Korea
Cohort
Air sampling
School children
Serum formaldehyde-specific IgE; airway
function; and exhaled FeNO
Not
impactful
ROS and antibody-
related measures
are not key
endpoints
Animal Studies
Liu et al. (2017)
Mouse (ICR), male
Subchronic (20 wk; 2
h/d)
Unspecified test article
0,1,10 mg/m3
Inhalation
Bone marrow cell MN; polychromatic
erythrocytes (PCE)/normochromatic
erythrocyte (NCE)ratio (immature/mature
RBCs)
Possibly
Impactful
Endpoints noted as
important in draft;
longer duration
study is rare (note:
presumed use of
formalin limits
interpretation)
Ma et al. (2020)
Mouse (Balb/c), male
Subchronic (8 wk; 8
h/d, 7 d/wk)
Formaldehyde in water
(methanol free)
0, 2 mg/m3
Inhalation
DNA damage (comet assay) in peripheral
tissues (e.g., spleen; thymus); % of CD4+ T
cells, CD8+T cells, ratio ofCD4+/CD8+ cells,
and CD4 and CD8 cell phenotyping spleen
weights, percentage of the DN (double
negative), DP (double positive), CD4SP
(single positive) and CD8SP cell populations
in the isolated thymocytes, cytotoxicity in
CD4SP and CD8SP cells, Runx (Runx 1,2,3,
C), Runxl, Runx3, and ThPOK expression in
the DP cells, ROS
Possibly
impactful
Informative
endpoints of
immune cell health
and function
NTP (2017)
Mouse (Trp53
haploinsufficient),
Male
Subchronic (8 wk; 6
h/d, 5 d/wk), then
held for 32 wk
Paraformaldehyde
0, 7.5 or 15 ppm (0,
9.23,18.5 mg/m3)
Inhalation
Hematology
Possibly
impactful
Already included in
2017 draft
Park et al. (2020)
Mouse (BALB/c),
female
Short-term (2 wk; 4
h/d, 5 d/wk)
Fresh formaldehyde
solution (methanol-free)
0, 1.38, 5.36 mg/m3
Inhalation
Splenic cytokines, T cell populations and
Thl/Th2 balance, differentiation markers
Possibly
impactful
T cell
subpopulation
analyses are
This document is a draft for review purposes only and does not constitute Agency policy.
F-27 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
considered
important
Zhao et al. (2020)
Mouse (Balb/c),
Formalin
Burst-forming unit-erythroid (BFU-E), and
Possibly
Important
male
0, 3 mg/m3
colony-forming unit-granulocyte
impactful
endpoints (note:
Short-term (2 wk; 8
Inhalation
macrophage (CFU-GM) colonies in nose,
(POE
formalin; in vitro
h/d, 5 d/wk)
lung, spleen, and bone marrow
tissues);
Not
impactful
(systemic
tissues)
are of less concern
for POE tissues)
Avdemir et al.
Rat (Wistar albino),
Formalin
Blood DNA damage (comet assay) and ROS
Not
Formalin; high
(2017)
both sexes
Subchronic (6 wk; 8
h/d, 5 d/wk)
0, 6 ppm (0, 7.4 mg/m3)
Inhalation (note: i.p. not
PECO relevant)
markers
impactful
level
Avdin et al.
Rat (Sprague-
Formalin
Serum and lung total antioxidant and
Not
ROS and serum
(2014)
Dawley), male
0, 5.27, 10.02 ppm (0,
oxidant status, and oxidative stress index;
impactful
lipid-related
Short-term (4 wk)
6.48,12.3 mg/m3)
Inhalation
serum glucose, protein, albumin, lipids,
cholesterol, HDL, LDL, triglyceride, T
protein; lung irisin levels and
immunostaining
measures are not
key endpoints
Bernardini et al.
Mouse (Swiss), male
Unspecified test article
Lung histopathology; BALcell counts and
Not
Unknown test
(2020)
Short-term (4 wk; 4
0, 0.5,1, 5,10 ppm (0,
inflammatory and ROS markers; global
impactful
article; not key
h/d, 5 d/wk)
0.62, 1.23,6.15, 12.3
mg/m3)
Inhalation
methylation in blood and bone marrow
endpoints
Cheng et al.
Mouse (Kunming),
Formalin
Serum CD4+, CD8+, and CD4/CD8T cell
Not
Formalin
(2016)
male
Short-term (3 or 7 d;
continuous)
0, 0.08, 0.8 mg/m3
Inhalation
counts
impactful
Abreu et al.
Mouse (C57BL/6),
Unspecified test article
Lung mechanics and morphology,
Not
Unknown test
(2016)
female
Acute (single
exposure, assessed 8
h later)
0, 0.2,1, 3 ppm (0, 0.25,
1.23, 3.69 mg/m3)
Inhalation
inflammatory cell counts and cytokines, and
ROS markers
impactful
article; acute
da Silva et al.
Rat (Wistar), male
Unspecified test article
BALcell counts (WBCs, Mono., Lympho.,
Not
Unknown test
(2015)
0,1 %
Neutro., Eosin.), cytokines, and
impactful
article; high levels
This document is a draft for review purposes only and does not constitute Agency policy.
F-28 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Short-term (3 d; 90-
Inhalation
myeloperoxidase activity (inflammation);
min/d)
lung morphometries, microvascular
permeability, and mRNA levels
Duan et al. (2018)
Mouse (BALB/c),
Formalin
Pulmonary eosinophil cationic protein
Not
Formalin; no saline
male
0,1 mg/m3
(histopathology), ROS markers, nuclear
impactful
plus formaldehyde
Short-term (18 d; 5
Inhalation
factor kappa B activation, and cytokine and
control group
h/d)
growth factor levels
Duan et al. (2020)
Mouse (Balb/c), male
Formalin
Airway IgE, cytokines and inflammatory
Not
Formalin; not key
Short-term (21 d; 6
0, 0.5 mg/m3
factors, Thl/Th2 balance, mucus secretion,
impactful
endpoints
h/d)
Inhalation
histopathology, and lung function
Ge et al. (2020a)
Mouse (Balb/c), male
Formalin
CBC; Myeloid progenitor cell (BFU-E and
Not
Formalin
Short-term (2 wk; 8
0,0.5, 3 mg/m3
CFU-GM) colony counts and cytokines;
impactful
h/d, 5 d/wk)
Inhalation
circulating ROS and cytokine markers; bone
marrow histology, ROS, and gene
expression of cell cycle and DNA damage
markers
Han et al. (2016)
Rat (Sprague-
Paraformaldehyde
Serum IgE, thymus Thl and Th2 cytokines,
Not
Nonspecific
Dawley), male
0, 0.83, 1.16 ppm (0,
body weight
impactful
antibodies and
Subchronic (6 wk; 2
1.02,1.43 mg/m3)
cytokines are not
h/d, 5 d/wk
Inhalation
key endpoints
beginning at PND3
Jin etal. (2021)
Mouse (C57BL/6J),
Unspecified test article
Respiratory parameters (e.g., rate) during
Not
Unknown test
both sexes
0, 5 ppm (0, 6.15
exposure; serum lipids; serum cell counts
impactful
article; not key
Short-term (4 d; 6
mg/m3)
and soluble factors (CBC)
endpoints
h/d)
Inhalation
Kane et al. (2018)
Mouse (BALB/c),
Formalin
Serum IgE, IgG; airway hyperreactivity, ROS
Not
Formalin; not key
male
0,1 mg/m3
markers, nuclear factor kappa B and MAPK
impactful
endpoints
Short-term (18 d;
Inhalation
activation; cytokine levels, and mast cell
5h/d)
degranulation
Leal et al. (2018)
Mouse (C57BL6),
Unspecified test article
Lung cytokines and elasticity measures
Not
Unknown test
male
0, 0.92 mg/m3
impactful
article; not key
Short-term (2 wk; 1
Inhalation
endpoints
h/d, 5 d/wk)
Li et al. (2017)
Mouse (Balb/c or
Formalin
Bronchial responsiveness (to methacholine),
Not
Formalin; not key
C57BL/6), male
0, 0.5, 3 mg/m3
Inhalation
BAL cytokines and cell counts (total, eosin.,
impactful
endpoints
This document is a draft for review purposes only and does not constitute Agency policy.
F-29 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Short-term (25 d;
lympho., neutro.); Serum OVA-specific IgE,
6h/d)
IgGl, and lgG2a
Lima et al. (2015)
Rat (Fischer), male
Unspecified test article
Trachea histology and morphometric
Not
Unknown test
Short-term (5d; 20-
0,1, 5,10 %
analyses, including mucus production,
impactful
article; high levels
min x3/d)
Inhalation
glycogen, ROS markers, and inflammatory
cell counts.
Liu et al. (2018b)
Rat (Sprague
Formalin
Lung autophagy, histopathology and BAL
Not
Formalin; not key
Dawley), male
0, 0.5, 5,10 mg/m3
cytokines
impactful
endpoints
Short-term (4 wk; 8
Inhalation
h/d)
Macedo et al.
Rat (Wistar), male
Formalin
BAL ROS markers and cellular oxidative
Not
Formalin; high
(2016b)
Short-term (3 d; 90-
min/d)
0,1 %
Inhalation
burst; lung tissue antioxidant enzyme
measures
impactful
levels
Murta et al.
Rat (Fischer), male
unspecified 0,1, 5,10 %,
BALF cell counts (WBCs, macrophages,
Not
Unknown test
(2016)
Short-term (5d; 20-
min x 3/d)
inhalation
lymphocytes, neutrophils, eosinophils),
inflammatory and ROS markers, and
neutrophil ROS production
Lung tissue inflammatory markers, H&E
staining and morphometry
impactful
article; high levels
Pavani et al.
Rat (Wistar, albino),
Unspecified test article
Lung ROS markers
Not
Unknown test
(2019)
male
Short-term (21 d; 1
h/d)
0, 40 %
Inhalation
impactful
article; high levels
Sapmaz et al.
Rat (Sprague-
Paraformaldehyde
Serum total IgA, IgM, IgG, complement C3
Not
Nonspecific
(2015)
Dawley), male
Short-term (4 wk; 8
h/d)
0, 5, 10 ppm (0, 6.15,
12.3 mg/m3)
Inhalation
impactful
antibody-related
measures are not
key endpoints
Sholapuri et al.
Rat (Wistar), male
Formalin
Hematology parameters (CBC); BAL
Not
Formalin; high
(2020)
Short-term (21 d; 1
h/d)
0, 40 %
Inhalation
histamine; lung histology
impactful
levels
Song et al. (2017)
Mouse (Balb/c), male
Formalin
Serum levels of cytokines, neuropeptides,
Not
Formalin; No
Short-term (25 d)
0, 2.44 ppm (0, 3
mg/m3)
Inhalation
ROS, and IgE; leukocyte counts and cellular
antioxidant levels.
impactful
formaldehyde-only
control (without
OVA);
Wei et al. (2017b)
Mouse (BALB/c),
Formalin
Complete blood cell count; bone marrow-
Not
Formalin; short-
male
0, 3 mg/m3
myeloid progenitor formation assay, ROS
impactful
term (otherwise
This document is a draft for review purposes only and does not constitute Agency policy.
F-30 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Short-term (2 wk; 8
h/d, 5 d/wk)
Inhalation
assay, IL-3 and GM-CSF ELISA, systemic
toxicity, bone marrow cellularity, apoptosis
assay
important
endpoints)
Wei et al. (2017a)
Mouse (BALB/c),
male
Short-term (2 wk; 5
d/wk), followed by 7
d recovery
Formalin
0, 3 mg/m3
Inhalation
Complete blood cell count, bone marrow
histopathology, myeloid progenitor colony-
forming cell assay, ROS and cytokine
measures, and DNA-protein crosslinks
Not
impactful
Formalin; short-
term (otherwise
important
endpoints)
Wen et al. (2016)
Mouse (Balb/c), male
Short-term (2 wk; 8
h/d, 5 d/wk)
Formalin
0, 3 mg/m3
Inhalation
Cell counts (WBCs, lymphocytes,
monocytes, neutrophils, RBCs, platelets);
serum antibody (total) level; ROS markers;
PBL proliferation; serum hemagglutination
titer and delayed-type hypersensitivity
(both after sheep RBC injection)
Not
impactful
Formalin (limits
interpretability of
systemic effects)
Wu et al. (2020)
Mouse (Balb/C), male
Short-term (21 d; 5
h/d)
Formalin
0, 0.8 mg/m3
Inhalation
Pulmonary function; lung histopathology;
airway hyperresponsiveness; lung IgE and
cytokine (including Thl/Th2) levels
Not
impactful
Formalin; not key
endpoints
Zhang et al.
(2018b)
Mouse (Balb/c), male
Short-term (7,14, or
28 d, 2 4h/d for
constant and 12 h/d
for intermittent)
Unspecified test article
0, 0.8 (intermittent) orO,
0.4 (constant) ppm (0,
0.49, or 0.98 mg/m3)
Inhalation
BAL cell counts (total, eosin., neutro.,
lympho.); lung tissue ROS markers,
histology, and cytokine and inflammatory
marker immunohistochemistry
Not
impactful
Unknown test
article; not key
endpoints
In Vitro/Ex Vivo Studies
Zhao et al. (2020)
Mouse (Balb/c), male
Ex vivo primary lung
and nose cells
(systemic cells not
PECO-relevant)
Acute (1 h)
Formalin
0, 50, 100, 200, 400 nM
In media
Burst-forming unit-erythroid (BFU-E), and
colony-forming unit-granulocyte
macrophage (CFU-GM) colonies
Possibly
impactful
Important
endpoints (note:
formalin; in vitro
are of less concern
for POE tissues)
An et al. (2019)
Human immortalized
bronchial epithelial
cells (in vitro
experiments in LHP-
relevant cells were
excluded)
Unspecified test article
0, 20, 40, 60, 80, 100,
120 nM
In media
Cell proliferation, ROS production, and
markers of cell division/proliferation and
ROS
Not
impactful
Unknown test
article; in vitro;
acute
This document is a draft for review purposes only and does not constitute Agency policy.
F-31 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Acute (2h)
Arslan-Acaroz
Human immortalized
Unspecified test article
Cell viability and ROS markers
Not
Unknown test
and Bavsu-
lung epithelial cells
0,100 nM,
impactful
article; in vitro;
Sozbilir (2020)
Acute (4 h)
In media
acute
Bonder et al.
Human immortalized
Unspecified test article
Cell viability and mitochondrial membrane
Not
Unknown test
(2019)
lung epithelial cells
(other in vitro
experiments in this
study excluded as
not PECO relevant)
Acute (24 h)
0, 63, 126, 378, 504, 630
Hmol/L
In media
potential
impactful
article; in vitro;
acute
Cuietal. (2016)
Human immortalized
Unspecified test article
Cell signaling and gene expression, ROS, and
Not
Unknown test
lung cells or Mouse
0, 200 nM
cellular currents
impactful
article; acute
(Balb/c) nasal
In media or instilled
instillation
Acute up to 48 h
Gostner et al.
Human
Unspecified test article
Cell viability; gene expression
Not
Unknown test
(2016)
immortalized, lung
epithelial cells
Short-term (3 d)
0, 0.1, 0.5 ppm (0, 0.12,
0.62 mg/m3)
Gaseous exposure at the
air:liquid interface
impactful
article; not key
endpoints
Jude et al. (2016)
Human primary
FormalinO, 0.2,0.8, 2
Agonist-induced calcium mobilization,
Not
Formalin; in vitro;
airway smooth
ppm (0, 0.25, 0.98, 2.46
cytotoxicity, ROS markers and cytokines in
impactful
acute
muscle (HASM) cells
mg/m3)
co-cultures; cabachol-induced airway
Acute (1 hr, assessed
Vapor delivered to cells
narrowing in slices
at 24 h)
Kim et al. (2018)
Human immortalized
Unspecified test article
ROS production, protein expression of
Not
Unknown test
endometrial adeno-
1011 to 10"3 M
markers associated with cell transformation
impactful
article; in vitro
carcinoma cells
In media
and proliferation
Short-term (6 d)
[Note: study included
due to use of this cell
line to examine
mechanisms
associated with
This document is a draft for review purposes only and does not constitute Agency policy.
F-32 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
epithelial cell-cell
interactions]
Li et al. (2008)
Human immortalized
tracheal epithelial
cells Acute (4 or 24 h)
Unspecified test article
0, 20, 50, 100, 200 nM
In media
Cell viability and expression of MAPK-
responsive genes
Not
impactful
Unknown test
article; in vitro;
acute
Liu et al. (2019)
Human immortalized
bronchial epithelial
cells
Acute (24 h)
Unspecified test article
0, 40, 80, 160 nmol/L
In media
Apoptosis, PI3K-Akt pathway signaling
markers
Not
impactful
Unknown test
article; in vitro;
acute
Mietal. (2019)
Human pulmonary
alveolar epithelial
cells in artificial
airway
Acute (2, 4, or 6 h)
Unspecified test article
0.025 and 40 nM (0.025
HM = ~0.3 ppm)
Nitrogen carrier-
mediated delivery
directly into cells
ROS and cytokine markers
Not
impactful
Unknown test
article; acute
Nazarparvar-
Noshadi et al.
(2020)
Human immortalized
lung epithelial cells
Acute/short-term
(24, 48, and 72 h)
Unspecified test article
0, 25, 50, 100, 150, 200,
300 nM
In media
Cellular viability and DNA damage markers
Not
impactful
Unknown test
article; in vitro
Vitoux et al.
(2018)
Human immortalized
conjunctival
epithelial cells
Acute (15-30 min,
assess at 1 or 24 h)
Formalin
0,100,1,200 ng/m3
Airflow over cells
Expression of inflammatory cytokines
Not
impactful
Formalin; in vitro;
acute
Zhang et al.
(2019)
Human immortalized
lung bronchial cells
Acute (3, 6,12, or 24
h)
Formalin
0, 5,10, 20, 40, 80
Hmol/L
In media
ROS and cytotoxicity markers m
Not
impactful
Formalin; in vitro;
acute
Zhang et al.
(2020b)
Human Immortalized
bronchial epithelial
cells
Formalin
0,10, 40, 80 nM
24 h
DNA damage - comet assay; apoptosis;
mitochondria-mediated apoptosis; reactive
oxygen species levels
Not
impactful
Formalin; in vitro;
non-critical
endpoints
Models, Endogenous Formaldehyde, or Other Studies
Dingier et al.
(2020)
Mouse (C57BL/6
background), ALDH2
and ALDH5 WT,
No formaldehyde
inhalation exposures
(note: included since it
Genotoxicity in peripheral blood cells and
bone marrow (MN assay, SCE); bone
marrow stem cell and progenitor cell
Possibly
impactful
Serves as included
reference study for
discussion of
This document is a draft for review purposes only and does not constitute Agency policy.
F-33 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
single, and double
KO, both sexes (note:
also includes primary
cultures of human
PBLs, fibroblasts, and
buccal cells not
deemed PECO-
relevant)
evaluates essentiality of
formaldehyde
detoxification processes
in normal function)
quantification, lineage characterization, and
B cell maturation; thymic development and
cell lineage characterization; complete
blood cell count, cell cycle profiling
potential sources
of susceptibility
relating to
formaldehyde
detoxification;
hematopoietic
health and cell
production from
bone marrow is
important
endpoint
Abbreviations: WBC = white blood cell; ROS = reactive oxygen species; BAL = bronchoalveolar lavage (F = fluid); RBC = red blood cell; PBL = peripheral blood
leukocyte; CBC = complete blood cell (count).
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
a Use of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
This document is a draft for review purposes only and does not constitute Agency policy.
F-34 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.10. Mechanistic Studies of Respiratory Tract Cancer, Focusing on Genotoxicity
Met PECO Possibly impactful
Respiratory tract cancer Excluded
mechanistic
©
Not primary research
[supplemental)
Figure F-10. Mechanistic respiratory tract cancer literature tree (interactive
version here).
2 A total of 3 62 citations were retrieved for the assessment of mechanistic information
3 informing respiratory tract cancers, focusing on genotoxicity, and 27 studies were PECO-relevant
4 Of these, 8 studies were deemed to be possibly impactful (note: one possibly impactful study is
5 repeated under both the animal and in vitro/ex vivo sections). Table 12 summarizes studies of
6 formaldehyde exposure in humans and animals, as well as in vitro or ex vivo experiments. Several
7 studies relevant to endogenous formaldehyde, pharmacokinetic modeling and dosimetry also were
8 included.
This document is a draft for review purposes only and does not constitute Agency policy.
F-35 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-12. Mechanistic studies relating to respiratory tract cancers, focusing on genotoxicity
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Human Studies
Aglan and
Mansour
(2018)
Occupational
Cairo, Egypt
Cross-sectional
Air sampling
Adult hairstylists
Buccal cell MN frequency
Possibly
impactful
Specific markers; exposures
similar to important studies
in draft
Costa et al.
(2019)
Occupational Portugal
Cross-sectional
Air sampling
Adult anatomy-pathology
laboratory workers
Buccal cell MN and nuclear
budding, genotype analysis of
selected polymorphisms
Possibly
impactful
Specific markers; exposures
similar to important studies
in draft
Peteffi et al.
(2015)
Occupational
Rio Grande do Sul, Brazil
Cross-sectional
Air sampling
Adult furniture workers
Micronucleus (MN) assay in
buccal cells: nuclear buds,
binucleated cells, Karyorrhexis
Possibly
impactful
Specific markers; exposures
similar to important studies
in draft
Bono et al.
(2016)
Occupational Piedmont
region, Italy
Cross-sectional
Air sampling
Adult plastic laminate
workers
Malondialdehyde DNA adducts
in swabbed nasal epithelial cells
Not
impactful
Adducts may or may not lead
to more robust markers
Bruno et al.
(2018)
Occupational
Rome, Italy
Cross-sectional
Air sampling
Adult pathology
laboratory workers
Counts of multinucleated
ciliated cells, Karyorrhexis,
Hyperchromatic SNS from
middle portion of the inferior
turbinate
Not
impactful
Nuclear abnormalities are
non- specific markers
Animal Studies
Leng et al.
(2019)
Rat (Fischer 344), male
Short-term (28 d; 6 h/d)
Deuterated formaldehyde
(no methanol)
0,1, 30, 300 ppb (1.23,
36.9, 369 mg/m3) [13CD2]-
HCHO
Inhalation
DNA adducts in nose, lung (and
other tissues)
Possibly
impactful
Endpoints important to
dosimetry; low exposure
levels
Zhao et al.
(2020)
Mouse (BALB/c), male
Short-term (2 wk; 8 h/d, 5
d/wk)
Formalin
0, 3 mg/m3
Inhalation
Burst-forming unit-erythroid
(BFU-E), and colony-forming
unit-granulocyte macrophage
(CFU-GM) colonies from nose
and lung
Possibly
impactful
Impactful endpoints (Note:
formalin, but less of a
concern in POE)
This document is a draft for review purposes only and does not constitute Agency policy.
F-36 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Bernardini et
al. (2020)
Mouse (Swiss), male
Short-term (4 wk; 4 h/d, 5
d/wk)
unspecified test article
0,0.5,1, 5,10 ppm (0,
0.62, 1.23, 6.15,12.3
mg/m3)
Inhalation
MN, comet assay, and global
methylation in lung
Not
impactful
Unknown test article; no
specific URT measures
Edrissi et al.
(2017)
Rat (F344), male
Short-term (7,14, 21, or 28
d;6 h/d)
[13C]-labeled
formaldehyde
0, 2 ppm (0, 2.46 mg/m3)
Inhalation
FA-lysine adducts in nasal
epithelium, lung, and trachea
Not
impactful
Adducts may or may not lead
to more robust markers
In vitro/Ex vivo Studies
Zhao et al.
(2020)
Mouse (BALB/c), male
Ex vivo primary lung and
nose cells
Acute (1 h)
Formalin
0, 50,100, 200, 400 nM
In media
Burst-forming unit-erythroid
(BFU-E), and colony-forming
unit-granulocyte macrophage
(CFU-GM) colonies
Possibly
impactful
Important endpoints (note:
formalin; in vitro)
Anandaraian
et al. (2020)
Yeast (Schizosaccharomyces
pombe), deletion strains
Short-term (3-5 d)
Formalin
0,0.2, 0.5, 1.25, 1.5, 1.75
mM
(Note: included due to
conserved DNA repair
pathways between yeast
and humans, and potential
relevance to human
susceptibility)
Toxicogenomic profiling of
pathways relating to
formaldehyde detoxification
and DNA repair-including
homologous recombination and
nucleotide excision repair
Not
impactful
Yeast; formalin; high dose
Chen et al.
(2017)
Human immortalized
bronchial epithelial cells
Acute (up to 6 h)
Unspecified test article
0,0.5 mM
In media
Inhibition of chromatin
assembly, formaldehyde-
histone adducts, gene
expression
Not
impactful
Unknown test article; in
vitro; non-critical endpoints
Gonzalez-
Rivera et al.
(2020)
Human immortalized
bronchial epithelial cells
Acute (2h)
Paraformaldehyde
0,1 ppm (0,1.23 mg/m3)
In vitro gaseous exposure
Cell phenotypic alterations; DNA
damage
Not
impactful
In vitro; non-critical
endpoints
Juarez et al.
(2018)
Human immortalized,
osteosarcoma, fibroblast, or
epithelial colorectal
adenocarcinoma cells
Unspecified test article
0, 20, 40, 60, 80, 100 nM
In media
genomic analysis
(Note: included due to analyses
across multiple cell lines which
might reflect genomic
Not
impactful
In vitro; indirect measure; no
cell lines specific to URT
This document is a draft for review purposes only and does not constitute Agency policy.
F-37 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Short-term (5 d; continuous)
signatures relevant to exposure
of URT cells)
Kang et al.
(2016)
Yeast (Saccharomyces
cerevisiae), deletion strains
5 or 15 generations of
exposure
Unspecified test article
0,0.15, 0.3, 0.6 mM
(Note: included due to
conserved DNA repair
pathways between yeast
and humans, and potential
relevance to human
susceptibility)
Toxicogenomic profiling of
pathways relating to RNA
stability and DNA repair-
including homologous
recombination, single strand
annealing, and post-replication
repair
Not
impactful
Yeast; Unknown test article;
high dose
Nazaroarvar-
Noshadi et al.
(2020)
Human immortalized lung
epithelial cells
Acute (24 h; note:
cytotoxicity up to 72 h)
Unspecified test article
0, 25, 50, 100, 150, 200,
and 300 nM
In media
DNA damage (DNA ladder) and
cytotoxicity/ apoptosis
Not
impactful
Unknown test article; in
vitro; non-critical endpoints
Zhang et al.
(2018a)
Human immortalized alveolar
basal epithelial cells
Acute (24 h)
Freshly prepared
formaldehyde solution
25 to 1,500 nM
In media
DNA damage; chromosome
damage; micronucleus
frequency; cytotoxicity
Not
impactful
In vitro (many in vivo studies
exist)
Zhang et al.
(2020a)
Human immortalized
bronchial epithelial cells
Acute (3, 6, 12, 24 h)
Formalin
0, 5,10, 20, 40, 80 nM
In media
DNA strand breaks;
chromosome damage; DNA
repair, ROS, and cell cycle
markers
Not
impactful
Formalin; in vitro; non-
critical endpoints
Zhang et al.
(2020b)
Human Immortalized
bronchial epithelial cells
Acute (24 h)
Formalin
0,10,40, 80 nM
In media
DNA damage - comet assay;
apoptosis; mitochondria-
mediated apoptosis; reactive
oxygen species levels
Not
impactful
Formalin; in vitro; non-
critical endpoints
Modeling, Endogenous Formaldehyde, and Other Studies
Campbell Jr et
al. (2020)
Updated pharmacokinetic model developed here for formaldehyde dG adducts based on the
previously developed models for formaldehyde DPX (Andersen et al., 2010; Conolly et al.,
2000).
Possibly
impactful
Model potentially important
to modeling dosimetry
(Note: discussed with regard
to toxicokinetics, Section
1.1.3, and cancer dose-
response, Section 2.2.1, not
MOA analysis, Section 1.2.5)
This document is a draft for review purposes only and does not constitute Agency policy.
F-38 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Corlev et al.
Excerpt from abstract: extended airway computational fluid dynamic (CFD) models of the rat
Possibly
Model potentially important
(2015)
and human were coupled with airway region-specific physiologically based pharmacokinetic
(PBPK) tissue models to describe the kinetics of formaldehyde. Simulations of aldehyde no-
observed-adverse-effect levels for nasal toxicity in the rat were conducted until breath-by-
breath tissue concentration profiles reached steady state. Human oral breathing simulations
were conducted using representative aldehyde yields from cigarette smoke.
impactful
to modeling dosimetry
(Note: discussed with regard
to toxicokinetics, Section
1.1.3, and cancer dose-
response, Section 2.2.1, not
MOA analysis, Section 1.2.5)
Miller et al.
BBDR: Previously a computational fluid dynamics model was combined with a 2-stage clonal
Possibly
Model potentially important
(2017)
growth model to develop a MOA-based DR model. This paper reports changes that reflect a
better understanding of populations of cells at risk of carcinogenic transformation in the
pharynx, larynx and respiratory bronchiolar portions of the human respiratory tract and
impactful
to modeling dosimetry
(Note: discussed with regard
to cancer dose-response,
inclusion of basal cells in the pool of cells at risk.
Section 2.2.1, not MOA
analysis, Section 1.2.5)
Burgos-
Mouse (C57BL/6 x 129SV
No formaldehyde
Genotoxicity (DNA damage
Not
Included as reference study
Barragan et al.
hybrid background), WT or
inhalation exposures
response markers) in vitro and
impactful
for discussion of potential
(2017)
KO in ALDH2, FANCD2, or
(note: included since it
in vivo (various tissues) when
sources of susceptibility
both (note: also included in
evaluates essentiality of
formaldehyde detoxification
relating to formaldehyde
vitro evaluations in human,
formaldehyde
pathways are disrupted
detoxification
chicken, and mouse cells)
detoxification processes in
normal function)
Starr and
Not
Included due to discussion in
Swenberg
impactful
2017 draft, but non-primary
(2016)
Update to prior non-primary research perspectives on how to calculate cancer risk
research
Yang et al.
Excerpt from abstract: the deposition rates from the linear regressions of CH20, CH5N,
Not
Not impactful to dosimetry
(2020)
C2H60, C2H402, C3H80, C6H6, C7H8, C8H8, and C8H10 of 120 healthy volunteers were
obtained with significantly different from the respective calculated deposition rates... In order
to explore the effects of the breathing models and sampling time on the deposition rates of
VOCs, volunteers were first asked to breathe successively with nasal-in-nasal-out, oral-in-
nasal-out, and oral-in-oral-out breathing models before and after three meals for three
days...In order to further validate the results, the deposition rates of the selected VOCs were
calculated in 120 healthy volunteers using nasal-in-oral-out breathing model for unlimited
impactful
modeling in the assessment
(note: briefly discussed in
the assessment as consistent
with prior observations)
time after the conventional lung function examination.
Yoo and Ito
BBDR: PBPK-computational fluid dynamics hybrid analysis was integrated into the computer
Not
Not impactful to dosimetry
(2018a)
simulated person-based numerical simulation to estimate inhalation exposure and respiratory
impactful
modeling in the assessment
tissue dosimetry with the unsteady breathing cycle model.
(see below)
This document is a draft for review purposes only and does not constitute Agency policy.
F-39 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Yoo and Ito
(2018b)
Excerpt from abstract: In this study, a CSP integrated with a virtual airway was developed and
used to estimate inhalation exposure in an indoor environment. The virtual airway is a
numerical respiratory tract model for CFD simulation that reproduces detailed geometry from
the nasal/oral cavity to the bronchial tubes by way of the trachea. Physiologically based
pharmacokinetic (PBPK)-CFD hybrid analysis is also integrated into the CSP. Through the
coupled simulation of PBPK-CFD-CSP analysis, inhalation exposure under steady state
conditions where formaldehyde was emitted from floor material was analyzed and respiratory
tissue doses and their distributions of inhaled contaminants are discussed quantitatively.
Not
impactful
Not impactful to dosimetry
modeling in the assessment
[these studies by Yoo and Ito
(2018a, b), extended the
Corley et al. (2015) modeling
by superposing on it the
dynamics of formaldehyde
exterior to the respiratory
tract (i.e. within the room
and surrounding the nose
and mouth). As such they do
not provide additional
information of relevance to
the assessment beyond that
discussed in the context of
Corley et al. (2015)]
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
Abbreviations: MN = micronucleus (assay); ROS = reactive oxygen species; BBDR = biologically based dose-response (model).
a Use of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
This document is a draft for review purposes only and does not constitute Agency policy.
F-40 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.11. Mechanistic Studies of Lymphohematopoietic Cancer, Focusing on
2 Genotoxicity
25 (l4)
Met PECO Possibly impactful
LHP cancer mechanistic Excluded
Not primary research
(supplemental)
Figure F-ll. Mechanistic lymphohematopoietic cancer literature tree
(interactive version here).
3 A total of 2,356 citations were retrieved for the assessment of mechanistic information
4 informing lymphohematopoietic cancers, focusing on genotoxicity, and 25 studies were PECO-
5 relevant (Table 13). Of these, 14 studies were deemed to be possibly impactful. Studies relevant to
6 pharmacokinetic modeling or dosimetry also were included. Mundtetal. f20171 was identified in
7 the literature search update and included in the inventory table although it already had been
8 included in the 2017 draft Toxicological Review of Formaldehyde-Inhalation.
This document is a draft for review purposes only and does not constitute Agency policy.
F-41 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-13. Mechanistic studies relating to lymphohematopoietic cancers, focusing on genotoxicity
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Human Studies
Asian and
Mansour (2018)
Occupational
Cairo, Egypt
Cross-sectional
Air sampling
Adult hairstylists
PBLMN
Possibly impactful
Specific markers;
exposures similar to
important studies in draft
Bassig et al. (2016)
Occupational Guangdong,
China
Cross-sectional,
Air sampling
Adult formaldehyde factory
workers
Frequency of monosomy 7 in
isolated CFU-GM cells
Possibly impactful
Specific markers;
exposures similar to
important studies in draft
Costa et al. (2015)
Occupational
Northern and Central
Portugal
Cross-sectional
Air sampling
Adult pathology workers
Chromosomal aberrations, comet
assay, genotype analysis in blood
cells
Possibly impactful
Specific markers;
exposures similar to
important studies in draft
Costa et al. (2019)
Occupational
Portugal
Cross-sectional
Air sampling
Adult anatomy-pathology
laboratory workers
PBLMN and
sister chromatid exchange;
T-cell receptor mutations;
genotype analysis of select
polymorphisms
Possibly impactful
Specific markers;
exposures similar to
important studies in draft
Mundt et al.
(2017)
Occupational
China
Cross-sectional
Additional analysis of Zhang
(2010) results
Adult factory workers
Monosomy of chromosome 7 and
8, complete blood count
Possibly impactful
Already identified in 2017
draft
Peteffi et al.
(2015)
Occupational
Rio Grande do Sul, Brazil
Cross-sectional
Air sampling
Adult furniture workers
Comet assay in PBLs [cell
migration, frequency of damaged
cells, damage index]
Possibly impactful
Markers of DNA damage;
exposures similar to
important studies in draft
Wang et al. (2019)
Occupational
Shanghai, China
Cross-sectional
Air sampling
Adult factory workers
Cytokinesis-blocked MN assay in
PBLs
Possibly impactful
Specific markers;
exposures similar to
important studies in draft
Zendehdel et al.
(2017)
Occupational
Tehran City, Iran
Cross-sectional
Air sampling
Adult melamine workers
Comet assay [tail moment, Olive
moment in PBLs]
Possibly impactful
Markers of DNA damage;
exposures similar to
important studies in draft
This document is a draft for review purposes only and does not constitute Agency policy.
F-42 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Barbosa et al.
Occupational
Porto Alegre, Brazil
Cross-sectional
Air sampling
Adult beauty salon workers
Global DNA methylation (%) in
PBLs
Not impactful
Not specific to
genotoxicity, so less
important endpoint
(2019)
Zendehdel et al.
Occupational
Tehran, Iran
Cross-sectional
Air sampling
Adult melamine workers
DNA damage (comet assay) in
PBLs
Not impactful
Related to Zendehdel et
(2018)
al. (2017), no additional
results.
Animal Studies
Leng et al. (2019)
Rat (Fischer 344), male
Short-term (28 d; 6 h/d)
Deuterated formaldehyde
(no methanol)
0,1, 30, 300 ppb (0, 1.23,
36.9, 369 ng/m3)
Inhalation
DNA adducts in blood, bone
marrow (and other tissues)
Possibly impactful
Endpoints important to
dosimetry; low exposure
levels
Liu et al. (2017)
Mouse (ICR), male
20 wk (2 h/d)
Unspecified test article
0,1,10 mg/m3
Inhalation
Bone marrow cell MN;
polychromatic erythrocytes
(PCE)/normochromatic
erythrocyte (NCE) ratio
(immature/mature RBCs)
Possibly Impactful
Endpoints noted as
important in draft; longer
duration study (note:
presumed use of formalin
limits interpretation)
Ma et al. (2020)
Mouse (Balb/c), male
Subchronic (8 wk; 8 h/d, 7
d/wk)
Formaldehyde in water
(methanol free)
0, 2 mg/m3
Inhalation
DNA damage (comet assay) in
peripheral tissues (e.g., spleen;
thymus); % of CD4+ T cells, CD8+ T
cells, ratio of CD4+/CD8+ cells,
and CD4 and CD8 cell phenotyping
Possibly impactful
Informative endpoints of
immune cell health and
function
Avdemir et al.
Rat (Wistar albino), both
sexes
Subchronic (6 wk; 8 h/d, 5
d/wk)
Formalin
0, 6 ppm (0, 7.38 mg/m3)
Inhalation (note: i.p.
deemed not PECO relevant)
DNA damage (comet assay) and
ROS markers in peripheral blood
Not impactful
Formalin; high level
(2017)
Bernardini et al.
Mouse (Swiss), male
Short-term (4 wk; 4 h/d, 5
d/wk)
unspecified test article
0, 0.5, 1, 5, 10 ppm (0, 0.62,
1.23,6.15, 12.3 mg/m3)
Inhalation
MN, comet assay, and global
methylation in blood and bone
marrow
Not impactful
Unknown test article
(2020)
Edrissi et al.
Rat (F344), male Short-
term (7, 14, 21, or 28 d; 6
h/d)
[13C]-labeled formaldehyde
0, 2 ppm
Inhalation
FA-lysine adducts in bone marrow
and WBCs
Not impactful
Adducts may or may not
lead to more robust
markers
(2017)
This document is a draft for review purposes only and does not constitute Agency policy.
F-43 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
Ge et al. (2020a)
Mouse (Balb/c), male
Short-term (2 wk; 8 h/d, 5
d/wk)
Formalin
0,0.5, 3 mg/m3
Inhalation
Myeloid progenitor cell (BFU-E
and CFU-GM) colony counts and
cytokines; bone marrow histology,
ROS, and gene expression of cell
cycle and DNA damage markers
Not impactful
Formalin; short-term
(otherwise important
endpoints)
Wei et al. (2017b)
Mouse (BALB/c), male
Short-term (2 wk; 8 h/d, 5
d/wk)
Formalin
0, 3 mg/m3
Inhalation
Bone marrow - myeloid
progenitor formation assay, bone
marrow cellularity
Not impactful
Formalin; short-term
(otherwise important
endpoints)
Wei et al. (2017a)
Mouse (BALB/c), male,
Short-term (2 wk; 5 d/wk),
followed by 7 d recovery
Formalin
0, 3 mg/m3
Inhalation
Complete blood count, bone
marrow histopathology, myeloid
progenitor colony-forming cell
assay, ROS and inflammatory
markers, DNA-protein crosslinks
Not impactful
Formalin; short-term
(otherwise important
endpoints)
Zhao et al. (2020)
Mouse (Balb/c), male
Short-term (2 wk; 8 h/d, 5
d/wk)
(note: ex vivo systemic
tissues not PECO relevant)
Formalin
0, 3 mg/m3
Formation of burst-forming unit-
erythroid (BFU-E), and colony-
forming unit-granulocyte
macrophage (CFU-GM) cellular
colonies in bone marrow and
spleen
Not impactful
Formalin; short-term
(otherwise important
endpoints)
Modeling, Endogenous Formaldehyde, and Other Studies
Burgos-Barragan
Mouse (C57BL/6 x 129SV
hybrid background), WT or
KO in ALDH2, FANCD2, or
both (note: also includes
in vitro evaluations in
human, chicken, and
mouse cells)
No formaldehyde
inhalation exposures (note:
included since it evaluates
essentiality of
formaldehyde
detoxification in normal
processes)
Colony Forming Units (CFU) from
bone marrow stem cells and
progenitor cells
Possibly impactful
Serves as included
reference study for
discussion of potential
sources of susceptibility
relating to formaldehyde
detoxification; cell
production from bone
marrow is an important
endpoint
et al. (2017)
Dingier et al.
Mouse (C57BL/6
background), ALDH2 and
ALDH5 WT, single, and
double KO, both sexes
(note: also includes
No formaldehyde
inhalation exposures (note:
included since it evaluates
essentiality of
formaldehyde
Genotoxicity in peripheral blood
cells and bone marrow (MN assay,
SCE); bone marrow stem cell and
progenitor cell quantification,
lineage characterization, and B
Possibly impactful
Serves as included
reference study for
discussion of potential
sources of susceptibility
relating to formaldehyde
(2020)
This document is a draft for review purposes only and does not constitute Agency policy.
F-44 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Mechanistic endpoints
Impact
Rationale
primary cultures of human
PBLs, fibroblasts, and
buccal cells not deemed
PECO-relevant)
detoxification processes in
normal function)
cell maturation; thymic
development and cell lineage
characterization; complete blood
cell count, cell cycle profiling
detoxification;
hematopoietic health and
cell production from bone
marrow is important
endpoint
Garcfa-Calderon et
al. (2018)
Mouse (C57BL/6
background), WT or KO in
ALDH5 or FANCD2 (note:
also includes in vitro
evaluations not deemed
PECO-relevant)
No formaldehyde
inhalation exposures (note:
included since it evaluates
essentiality of
formaldehyde
detoxification in normal
processes)
Bone marrow HSPC lineage,
function, and genotoxicity;
complete blood cell count
Possibly impactful
Serves as included
reference study for
discussion of potential
sources of susceptibility
relating to formaldehyde
detoxification;
hematopoietic health and
cell production from bone
marrow are important
endpoints
Nakamura et al.
(2020)
Mouse (C57BL/6
background), ALDH2 and
ALDH5 WT, single, and
double KO, both sexes
Observed GDO to PND25
No formaldehyde
inhalation exposures (note:
included since it evaluates
essentiality of
formaldehyde
detoxification processes in
normal function)
Postnatal survival and gross organ
observations (e.g., spleen, liver,
lung thymus)
Not impactful
Serves as included
reference study for
discussion of potential
sources of susceptibility
relating to formaldehyde
detoxification
Starr and
Swenberg (2016)
Update to prior non-primary research perspectives on how to calculate cancer risk
Not impactful
Included here because
commented on in existing
draft, but non-primary
research
Abbreviations: PBL = peripheral blood leukocytes; MN = micronucleus; WBC = white blood cell.
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
a Use of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
This document is a draft for review purposes only and does not constitute Agency policy.
F-45 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
F.3.12. Nervous System Effects
0 0
Met PECO Possibly impactful
Nervous system Excluded
42)
Not primary research
(supplemental)
Figure F-12. Nervous system effects literature tree (interactive version here).
2 A total of 2,617 citations were retrieved for the assessment of nervous system effects and
3 14 studies were PECO-relevant (Table 14). Of these, two human studies were deemed to be possibly
4 impactful. Peters etal. (2017) was identified in the literature search update and included in the
5 inventory table although it already had been included in the 2017 draft Toxicological Review of
6 Formaldehyde-Inhalation. None of the identified animal or mechanistic studies were deemed
7 possibly impactful.
This document is a draft for review purposes only and does not constitute Agency policy.
F-46 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-14. Studies of nervous system effects
Reference
Study design
Exposure3
Endpoints
Impact
Rationale
Human Studies
Bellavia et
al. (2021)b
General population
Denmark
case-control
Occupational history and
job-exposure matrix, adults
Amyotrophic lateral sclerosis
(ALS)
Possibly
impactful
Additional study on
health effect for which
there are few studies
Peters et al.
(2017)
General population
Sweden
case-control
Occupational history and
job-exposure matrix, adults
Amyotrophic lateral sclerosis
(ALS) incidence
Possibly
impactful
Already identified in
2017 draft
Animal Studiesc
Askarand
Halloull
(2018)
Rat (Albino, strain not
specified), male
Subchronic (12 wk; 6 h/d,
5 d/wk)
Paraformaldehyde
0, 20 ppm (0, 24.6 mg/m3)
Inhalation
Cerebellar histopathology, cell
counts, and cell morphology;
evaluations of ROS and
inflammatory markers
Not impactful
High levels
Cheng et al.
(2016)
Mouse (Kunming), male
Short-term (Up to 7 d;
continuous)
Formalin
0,0.08,0.8 mg/m3
Inhalation
Morris water maze
Not impactful
Formalin
Duan et al.
(2018)
Mouse (Balb/c), male
Short-term (18 d; 5h/d)
Formalin
0,1 mg/m3
Inhalation
Prefrontal cortex histology; brain
ROS and inflammation markers,
cytokines
Not impactful
Formalin; no saline plus
formaldehyde control
group
Ge et al.
(2019)
Mouse (Kunming), male
Short-term (21 d;
continuous)
Formalin
0,1 mg/m3
Inhalation
Morris water maze, hippocampal
morphology, brain ROS and cell
signaling markers
Not impactful
Formalin
Huang et al.
(2019)
Mouse (Kunming), male
Short-term (14 d; 8 h/d)
Formalin
0, 3 mg/m3
Inhalation
Morris water maze; brain ROS
and inflammatory markers;
hippocampal histopathology and
cell morphology
Not impactful
Formalin
Li et al.
(2016)
Mouse (Kunming), male
Short-term (7 d; 2 h/d)
Formalin
0,1, 2 ppm (0, 1.23, 2.46
mg/m3)
Inhalation
Open field activity; elevated plus
maze test; forced swimming test;
novel object recognition; counts
of TH- and GR-immunoreactive
neurons
Not impactful
Formalin; brief
exposures
This document is a draft for review purposes only and does not constitute Agency policy.
F-47 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Endpoints
Impact
Rationale
Li et al.
(2020)
Mouse (Kunming), male
Short-term (14 d; 8 h/d)
Formalin
0,0.5, 3 mg/m3
Inhalation
Morris water maze; brain ROS
and inflammatory markers;
hippocampal histopathology and
cell morphology
Not impactful
Formalin
Mouse (Balb/c), male
Short-term (7 d; 8 h/d)
Brain neurotransmitters; ROS and
inflammatory markers in
hippocampus, brain stem, and
cerebral cortex
Mei et al.
(2016)
Mouse (Balb/c), male
Short-term (7 d; 8 h/d) (in
vitro experiments not
PECO-relevant)
Unspecified test article
0, 3 mg/m3
Inhalation
Morris water maze; qualitative
hippocampal neuron staining;
brain ROS and GSH
Not impactful
Formalin
Zhang et al.
(2014b)
Rat (Sprague Dawley),
male
Short-term (14 d; 30-min,
2x/d)
Unspecified test article
0,13.5 ppm (0,16.6 mg/m3)
Inhalation
Buried food pellet behavioral
testing; olfactory bulb
synaptosomal and neuronal
markers; olfactory sensory
neuron maturation
Not impactful
Unknown test article;
high levels; brief
exposures
Mechanistic Studies
Cao et al.
(2015)
Mouse (Balb/c), male
Short-term (7 d; 8 h/d)
Unspecified test article
0,0.5, 3 mg/m3
Inhalation
Hippocampus, cortex, and
brainstem ROS and inflammatory
markers
Not impactful
Unknown test article
Eom et al.
(2017)
Drosophila melanogaster
(mutant strains: WT, p53
and p38b)
Acute (6 or 24 h)
Unspecified test article
0,10,100 |jg/m3
Inhalation
Behavioral (movement-based)
quantification; microarray
analyses (note survival test study
design not extracted)
Not impactful
Non-mammalian;
unknown test article
Li et al.
(2015)
mouse (ICR), male, Acute
or short-term (1 or 7 d; 6
h/d)
Unspecified test article
0, 3 ppm (0, 3.69 mg/m3)
Inhalation
miRNA screening of olfactory
bulb
Not impactful
Unknown test article
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
aUse of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
bAn additional study, Seals et al.(2017), was identified from the reference list of Bellavia et al. (2021). As this study was determined to be possibly impactful to
the 2017 draft conclusions on nervous system effects, it was incorporated into the Toxicological Review.
cAnimal studies may include evaluation of mechanistic endpoints.
This document is a draft for review purposes only and does not constitute Agency policy.
F-48 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
1 F.3.13. Reproductive and Developmental Effects
® 0
Met PECO Possibly impactful
1520
Reproduction and development Excluded
©
Not primary research
(supplemental)
Figure F-13. Reproductive and developmental effects literature tree
(interactive version here).
2 A total of 1,544 citations were retrieved for the assessment of reproductive and
3 developmental effects and nine studies were PECO-relevant (Table 15). Of these, five were deemed
4 to be possibly impactful. There were four from the human literature and one from the animal
5 literature. Neither of the identified mechanistic studies were deemed possibly impactful. Wang etal.
6 T20151 was identified in the literature search update and included in the inventory table although it
7 already had been included in the 2017 draft Toxicological Review of Formaldehyde-Inhalation.
This document is a draft for review purposes only and does not constitute Agency policy.
F-49 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Table F-15. Studies of reproductive and developmental effects
Reference
Study design
Exposure3
Endpoints
Impact
Rationale
Human Studies
Amiri and Turner-
Henson (2017)
General population
southeastern U.S.
cross-sectional
Air sampling, prenatal,
exposure during
pregnancy
Biparietal diameter, head circumference,
abdominal circumference, femur length, ratio
of abdominal circumference to femur length
(AC/FL), estimated fetal weight
Possibly
impactful
Health effect for
which there are few
studies
Chang et al.
(2017)
General population
Seoul, South Korea
birth cohort
Air sampling, prenatal,
exposure during
pregnancy
Birthweight, postnatal weight at 6,12, 24, and
36 months
Possibly
impactful
Health effect for
which there are few
studies
Franklin et al.
(2019)
General population
Australia
birth cohort
Air sampling, prenatal,
exposure during
pregnancy
Gestational age, birth length, birth weight,
head circumference
Possibly
impactful
Health effect for
which there are few
studies
Wang et al.
(2015)
Occupational China
cross-sectional
Air sampling and
occupational history,
adults, male plywood
production workers
Semen volume, sperm concentration,
total sperm count, sperm progressive motility
and total sperm motility, curvilinear velocity,
straight line velocity, linearity, time-average
velocity, straightness, mean angular
displacement, amplitude of lateral head
displacement
Possibly
impactful
Already identified in
2017 draft
Animal Studies'3
Sapmaz et al.
(2018)
Rat (Sprague
Dawley), male
Short-term (4 wk) or
Subchronic (13 wk), 8
h/d, 5 d/wk
Paraformaldehyde
0, 5 ppm (0, 6.15 mg/m3)
Inhalation
Testicular tubular atrophy, germinative
epithelium height, seminiferous tubule
diameter; markers of ROS in testicular tissue
Possibly
impactful
Longer duration
study; informative
morphological
endpoints
Ge et al. (2020b)
Rat (Sprague
Dawley), male
Subchronic (8 wk)
Formalin
0, 0.5, 2.46, 5 mg/m3
Inhalation
Testicular seminiferous tubule histopathology
and morphometry, SPOll protein in testicular
tissue
Not
impactful
Formalin
Zang et al. (2017)
Mouse (C57BL/6),
male
Formalin
0, 0.5, 5,10 mg/m3
Sexual behavior (mount latency, intromission
latency, ejaculation latency, mount frequency,
Not
impactful
Formalin
This document is a draft for review purposes only and does not constitute Agency policy.
F-50 DRAFT—DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
Reference
Study design
Exposure3
Endpoints
Impact
Rationale
Subchronic (60 d; 4
h/d)
Inhalation
intromission frequency, copulatory efficacy);
hormone measures (serum T and LH; testicular
T); sperm number and motility; reproductive
organ weights and histopathology
Mechanistic Studies
Fang et al. (2015)
Rat (Sprague
Dawley), male
Short-term (4 wk;
8h/d)
Unspecified test article
0, 0.5, 5,10 mg/m3
Inhalation
mTOR (mammalian target of rapamycin, a
regulator of various cellular processes) mRNA
expression, protein levels, and
immunostaining in testes
Not
impactful
Unspecified test
article
Ibrahim et al.
(2016)
Rat (Wistar), female
(dam)
Gestational (GD1-21;
lh/d, 5d/wk)
Unspecified test article
0, 0.92 mg/m3
Inhalation
Markers of ROS and inflammation in dam
uterus at parturition; inflammation and
immune parameters in offspring after PND30:
BALcell count and myeloperoxidase activity,
lung cytokines and inflammatory markers;
blood and bone marrow cell counts
Not
impactful
Unspecified test
article
Rows for studies judged as "not impactful" are shaded grey; unshaded rows highlight studies incorporated into the updated draft assessment.
aUse of methanol-stabilized formalin was inferred in some studies based on study-specific description (e.g., 37% stock solution).
bAnimal studies may include evaluation of mechanistic endpoints.
This document is a draft for review purposes only and does not constitute Agency policy.
F-51 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Supplemental Information for Formaldehyde—Inhalation
APPENDIX G. QUALITY ASSURANCE FOR THE IRIS
TOXICOLOGICAL REVIEW OF FORMALDEHYDE
This assessment is prepared under the auspices of the U.S. Environmental Protection
Agency's (EPA's) Integrated Risk Information System (IRIS) Program. The IRIS Program is housed
within the Office of Research and Development (ORD) in the Center for Public Health and
Environmental Assessment (CPHEA). EPA has an agency-wide quality assurance (QA) policy that is
outlined in the EPA Quality Manual for Environmental Programs (see CIO 2105-P-01.11 and follows
the specifications outlined in EPA Order CIO 2105.1.
As required by CIO 2105.1, ORD maintains a Quality Management Program, which is
documented in an internal Quality Management Plan (QMP). The latest version was developed in
2013 using Guidance for Developing Quality Systems for Environmental Programs (OA/G-1). An
NCEA/CPHEA-specific QMP was also developed in 2013 as an appendix to the ORD QMP. Quality
assurance for products developed within CPHEA is managed under the ORD QMP and applicable
appendices.
The IRIS Toxicological Review of Forrmaldehyde is designated as Highly Influential
Scientific Information (HISA)/Influential Scientific Information (ISI) and is classified as QA Category
A. Category A designations require reporting of all critical QA activities, including audits. The
development of IRIS assessments is done through a seven-step process. Documentation of this
process is available on the IRIS website: https://www.epa.gov/iris/basic-information-about-
i integrated-risk-information-svstem#process.
Specific management of quality assurance within the IRIS Program is documented in a
Programmatic Quality Assurance Project Plan (PQAPP). A PQAPP is developed using the EPA
Guidance for Quality Assurance Project Plans fOA/G-51. and the latest approved version is dated
March 2020. All IRIS assessments follow the IRIS PQAPP, and all assessment leads and team
members are required to receive QA training on the IRIS PQAPP. During assessment development,
additional QAPPs may be applied for quality assurance management They include
Title
Document number
Date
Program Quality Assurance Project
Plan (PQAPP) for the Integrated Risk
Information System (IRIS) Program
L-CPAD-0030729-QP-1-4
April 2021
An Umbrella Quality Assurance
Project Plan (QAPP) for Dosimetry
and Mechanism-Based Models
(PBPK)
L-CPAD-0032188-QP-1-2
December 2020
Quality Assurance Project Plan
(QAPP) for Enhancements to
Benchmark Dose Software (BMDS)
L- H E EAD-0032189-QP-1-2
October 2020
This document is a draft for review purposes only and does not constitute Agency policy.
G-l DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information for Formaldehyde—Inhalation
During assessment development, this project undergoes one quality auditduring
assessment development including:
Date
Type of audit
Major findings
Actions taken
July 27, 2021
Technical system audit
None
None
During Step 3 and Step 6 of the IRIS process, the IRIS toxicological review is subjected to
external reviews by other federal agency partners, including the Executive Offices of the White
House. Comments during these IRIS process steps are available in the docket [insert chemical docket
number—make sure the comments are in the docket] on http://www, regulations,gov.
During Step 4 [include this section AFTER Step 4] of assessment development, the IRIS
Toxicological Review of Formaldehyde undergoes public comment from [insert date of public
comment]. Following this comment period, the toxicological review undergoes external peer review
by [SAB/NAS/contractor peer-review panel] on [insert date ofERD], The peer-review report is
available on the [NAS/SAB website—include the URL], All public and peer-review comments are
available in the docket [insert chemical docket number—make sure that the ERD public comments are
available in the docket as well],
[Include this section AFTER Step 6] Prior to release (Step 7 of the IRIS process), the final
toxicological review is submitted to management and QA clearance. During this step the CPHEA QA
Director and QA Managers review the project QA documentation and ensure that EPA QA
requirements are met
This document is a draft for review purposes only and does not constitute Agency policy.
G-2 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Supplemental Information for Formaldehyde—Inhalation
REFERENCES
[Multiple references published in the same year by the same author(s) have been assigned a
letter (e.g., 1986a, 1986b) based on order of appearance in the text of the document Those same
letters have been retained for the appendices.]
Abrams. WR: Kallen. RG. (1976). Equilibria and kinetics of N-hydroxymethylamine formation from
aromatic exocyclic amines and formaldehyde. Effects of nucleophilicity and catalyst
strength upon mechanisms of catalysis of carbinolamine formation1. J Am Chem Soc 98:
7777-7789. http://dx.doi.org/10.1021/ia00440a052
Abramson. Ml: Perret. TL: Dharmage. SC: McDonald. VM: McDonald. CF. (2014). Distinguishing
adult-onset asthma from COPD: a review and a new approach [Review], The International
Journal of Chronic Obstructive Pulmonary Disease (Online) 9: 945-962.
http://dx.doi.Org/l 0.2147 /C0PD.S46761
Abreu. M. d: Neto. AC: Carvalho. G: Casquillo. NY: Carvalho. N: Okuro. R: Ribeiro. GC: Machado. M:
Cardozo. A: Silva. AS: Barboza. T: Vasconcellos. LR: Rodrigues. DA: Camilo. L: Carneiro. L:
Tandre. F: Pino. AY: Giannella-Neto. A: Zin. WA: Correa. LH: Souza. MN: Carvalho. AR. (2016).
Does acute exposure to aldehydes impair pulmonary function and structure? Respir Physiol
Neurobiol 229: 34-42. http://dx.doi.Org/10.1016/i.resp.2016.04.002
Adams. DO: Hamilton. TA: Lauer. LP: Dean. TH. (1987). The effect of formaldehyde exposure upon
the mononuclear phagocyte system of mice. Toxicol Appl Pharmacol 88: 165-174.
http://dx.doi. org/10.1016/0041 -008xC87190002-0
Aglan. MA: Mansour. GN. (2018). Hair straightening products and the risk of occupational
formaldehyde exposure in hairstylists. Drug Chem Toxicol 43: 1-8.
http://dx.doi.Org/10.1080/01480545.2018.1508215
Ahlborg. G. Tr. (1990). Pregnancy outcome among women working in laundries and dry-cleaning
shops using tetrachloroethylene. Am J Ind Med 17: 567-575.
http://dx.doi.Org/10.1002/ajim.4700170503
Ahmed. S: Tsukahara. S: Tin-Tin-Win-Shwe: Yamamoto. S: Kunugita. N: Arashidani. K: Fuiimaki. H.
(2007). Effects of low-level formaldehyde exposure on synaptic plasticity-related gene
expression in the hippocampus of immunized mice. J Neuroimmunol 186: 104-111.
http://dx.doi.Org/10.1016/i.ineuroim.2007.03.010
Ahn. KH: Kim. SK: Lee. TM: Teon. HI: Lee. DH: Kim. DK. (2010). Proteomic analysis of bronchoalveolar
lavage fluid obtained from rats exposed to formaldehyde. J Health Sci 56: 287-295.
http://dx.doi.org/10.1248/jhs.56.287
Akbar-Khanzadeh. F: Vaquerano. MU: Akbar-Khanzadeh. M: Bisesi. MS. (1994). Formaldehyde
exposure, acute pulmonary response, and exposure control options in a gross anatomy
laboratory. Am J Ind Med 26: 61-75. http://dx.doi.org/10.1002/aiim.4700260106
Alarie. Y. (1981). Toxicological evaluation of airborne chemical irritants and allergens using
respiratory reflex reactions. In BKJ Leong (Ed.), Inhalation toxicology and technology (pp.
207-231). Ann Arbor, MI: Ann Arbor Science Publishers, Inc.
This document is a draft for review purposes only and does not constitute Agency policy.
R-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Albert. RE: Sellakumar. AR: Laskin. S: Kuschner. M: Nelson. N: Snyder. CA. (1982). Gaseous
formaldehyde and hydrogen chloride induction of nasal cancer in the rat J Natl Cancer Inst
68: 597-603.
Alderson. T. (1967). Induction of genetically recombinant chromosomes in the absence of induced
mutation. Nature 215: 1281-1283.
Alexandersson. R. (1988). Decreased lung function and exposure to formaldehyde in the wood
working industry. A five-year follow-up. Arh Hig Rada Toksikol 39: 421-424.
Alexandersson. R: Hedenstierna. G. (1988). Respiratory hazards associated with exposure to
formaldehyde and solvents in acid-curing paints. Arch Environ Health 43: 222-227.
http://dx.doi.org/10.1080/00039896.1988.9934937
Alexandersson. R: Hedenstierna. G. (1989). Pulmonary function in wood workers exposed to
formaldehyde: A prospective study. Arch Environ Health 44: 5-11.
http://dx.doi.Org/10.1080/00039896.1989.9935865
Alexandersson. R: Hedenstierna. G: Kolmodin-Hedman. B. (1982). Exposure to formaldehyde:
effects on pulmonary function. Arch Environ Health 37: 279-284.
http://dx.doi.Org/10.1080/00039896.1982.10667579
Amiri. A: Turner-Henson. A. (2017). The roles of formaldehyde exposure and oxidative stress in
fetal growth in the second trimester. J Obstet Gynecol Neonatal Nurs 46: 51-62.
http://dx.doi.Org/10.1016/j.jogn.2016.08.007
An. 1: Li. F: Oin. Y: Zhang. H: Ding. S. (2019). Low concentrations of FA exhibits the Hormesis effect
by affecting cell division and the Warburg effect. Ecotoxicol Environ Saf 183: 109576.
http://dx.doi.Org/10.1016/i.ecoenv.2019.109576
Anandaraian. V: Noguchi. C: Oleksak. 1: Grothusen. G: Terleckv. D: Noguchi. E. (2020). Genetic
investigation of formaldehyde-induced DNA damage response in Schizosaccharomyces
pombe. Curr Genet 66: 593-605. http://dx.doi.Org/10.1007/s00294-020-01057-z
Andersen. I. (1979). Formaldehyde in the indoor environment - health implications and the setting
of standards. In PO Fanger; 0 Valbjorn (Eds.), Indoor climate: Effects on human comfort,
performance, and health in residential, commercial, and light-industry buildings (pp. 65-
87). Copenhagen, Denmark: Danish Building Research Institute.
Andersen. I: Molhave. L. (1983). Controlled human studies with formaldehyde. In JE Gibson (Ed.),
Formaldehyde toxicity (pp. 154-165). Washington, DC: Hemisphere Publishing.
Andersen. ME: Clewell. HI: Bermudez. E: Dodd. DE: Willson. GA: Campbell. TL: Thomas. RS. (2010).
Formaldehyde: integrating dosimetry, cytotoxicity, and genomics to understand dose-
dependent transitions for an endogenous compound. Toxicol Sci 118: 716-731.
http: / /dx. doi. or g/10.109 3 /toxsci /kfq 3 0 3
Andersen. ME: III. CH: Bermudez. E: Willson. GA: Thomas. RS. (2008). Genomic signatures and dose-
dependent transitions in nasal epithelial responses to inhaled formaldehyde in the rat
Toxicol Sci 105: 368-383. http:/ /dx.doi.org/10.1093/toxsci /kfnO97
Andersson. B: Eriksson. B: Falsen. E: Fogh. A: Hanson. LA: Nvlen. 0: Peterson. H: Svanborg Eden. C.
(1981). Adhesion of Streptococcus pneumoniae to human pharyngeal epithelial cells in
vitro: differences in adhesive capacity among strains isolated from subjects with otitis
media, septicemia, or meningitis or from healthy carriers. Infect Immun 32: 311-317.
This document is a draft for review purposes only and does not constitute Agency policy.
R-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Andersson. M: Agurell. E: Vaghef. H: Bolcsfoldi. G: Hellman. B. (2003). Extended-term cultures of
human T-lymphocytes and the comet assay: a useful combination when testing for
genotoxicity in vitro? Mutat Res 540: 43-55. http://dx.doi.org/10.1016/S1383-
5718f03100169-4
Andielkovich. DA: lanszen. DB: Brown. MH: Richardson. RB: Miller. FT. (1995). Mortality of iron
foundry workers: IV. Analysis of a subcohort exposed to formaldehyde. J Occup Environ
Med 37: 826-837. http://dx.doi.org/10.1097/00043764-199507000-00012
Apfelbach. R: Weiler. E. (1991). Sensitivity to odors in wistar rats is reduced after low-level
formaldehyde-gas exposure. Naturwissenschaften 78: 221-223.
http ://dx. doi. or g/10.10 0 7 /bfO 113 60 8 5
Appelman. LM: Woutersen. RA: Zwart. A: Falke. HE: Feron. VI. (1988). One-year inhalation toxicity
study of formaldehyde in male rats with a damaged or undamaged nasal mucosa. J Appl
Toxicol 8: 85-90. http://dx.doi.org/10.1002/iat25500802Q4
Arbes. ST: Gergen. PI: Elliott. L: Zeldin. DC. f2005I Prevalences of positive skin test responses to 10
common allergens in the US population: Results from the Third National Health and
Nutrition Examination Survey. J Allergy Clin Immunol 116: 377-383.
http://dx.doi.Org/10.1016/i.iaci.2005.05.017
Arican. RY: Sahin. Z: Ustunel. I: Sarikcioglu. L: Ozdem. S: Oguz. N. (2009). Effects of formaldehyde
inhalation on the junctional proteins of nasal respiratory mucosa of rats. Exp Toxicol Pathol
61: 297-305. http://dx.doi.Org/10.1016/i.etp.2008.09.005
Armstrong. RW: Imrev. PB: Lye. MS: Armstrong. MI: Yu. MC: Sani. S. (2000). Nasopharyngeal
carcinoma in Malaysian Chinese: occupational exposures to particles, formaldehyde and
heat Int J Epidemiol 29: 991-998. http://dx.doi.Org/10.1093/ije/29.6.991
Arslan-Acaroz. D: Bavsu-Sozbilir. N. (2020). Ameliorative effect of boric acid against formaldehyde-
induced oxidative stress in A549 cell lines. Environ Sci PollutRes Int 27: 4067-4074.
http: //dx.doi.org/10.1007/sl 1356-019-06986-v
Asgedom. AA: Bratveit. M: Moen. BE. (2019). High Prevalence of Respiratory Symptoms among
Particleboard Workers in Ethiopia: A Cross-Sectional Study. Int J Environ Res Public Health
16. http: //dx.doi.org/10.3390 /iierph 16122158
Asher. MI: Keil. U: Anderson. HR: Beaslev. R: Crane. I: Martinez. F: Mitchell. EA: Pearce. N: Sibbald. B:
Stewart. AW. (1995). International Study of Asthma and Allergies in Childhood (ISAAC):
rationale and methods. Eur Respir J 8: 483-491.
http://dx.doi.Org/10.1183/09031936.95.08030483
Askar. EM: Halloull. NM. (2018). Formaldehyde-induced neurotoxicity in rat cerebellar cortex and
possible protective effects of fatty acids from omega 3 and wheat germ oil supplement: a
histopathological and biochemical study. J Histotechnol 41: 79-87.
http: //dx.doi.org/10.1080 /01478885.2018.1458176
Asian. H: Songur. A: Tunc. AT: Ozen. OA: Bas. 0: Yagmurca. M: Turgut. M: Sarsilmaz. M: Kaplan. S.
(2006). Effects of formaldehyde exposure on granule cell number and volume of dentate
gyrus: a histopathological and stereological study. Brain Res 1122: 191-200.
http://dx.doi.Org/10.1016/j.brainres.2006.09.005
ATSDR (Agency for Toxic Substances and Disease Registry). (1999). Toxicological profile for
formaldehyde [ATSDR Tox Profile], Atlanta, GA: U.S. Department of Health and Human
Services, Public Health Service, http://www.atsdr.cdc.gov/toxprofiles/tplll.pdf
This document is a draft for review purposes only and does not constitute Agency policy.
R-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
ATSDR. (2008). Draft for toxicological profile for formaldehyde [ATSDR Tox Profile], Atlanta, GA:
U.S. Department of Health and Human Services, Public Health Service.
Attia. D: Mansour. N: Taha. F: El Dein. AS. (2014). Assessment of lipid peroxidation and p53 as a
biomarker of carcinogenesis among workers exposed to formaldehyde in cosmetic industry.
Toxicol Ind Health 32: 1097-1105. http://dx.doi.org/10.1177/0748233714547152
Auerbach. C: Moser. H. (1953a). An analysis of the mutagenic action of formaldehyde-food. I.
Sensitivity of Drosophila germ cells. MGG Mol gen genet 85: 479-504.
http://dx.doi.org/10.1007/RF00308298
Auerbach. C: Moser. H. (1953b). Analysis of the mutagenic action of formaldehyde food. II. The
mutagenic potentialities of the treatment MGG Mol gen genet 85: 547-563.
http://dx.doi.Org/10.1007/RF00308300
Auerbach. C: Moutschen-Dahmen. M: Moutschen. 1. (1977). Genetic and cytogenetical effects of
formaldehyde and related compounds [Review], DNA Repair 39: 317-361.
http://dx.doi. org/10.1016/0165-1110177190011 -2
Augenreich. A: Stickford. 1: Stute. N: Koontz. L: Cope. I: Rennett. C: Ratchford. SM. (2020). Vascular
dysfunction and oxidative stress caused by acute formaldehyde exposure in female adults.
Am J Physiol Heart Circ Physiol 319: H1369-H1379.
http://dx.doi.Org/10.1152/aipheart00605.2020
Aung. W: Sakamoto. H: Sato. A: Yi. E: Thein. Z: Nwe. M: Shein. N: Linn. H: Uchivama. S: Kunugita. N:
Win-Shwe. T: Mar. 0. hn. (2021). Indoor Formaldehyde Concentration, Personal
Formaldehyde Exposure and Clinical Symptoms during Anatomy Dissection Sessions,
University of Medicine 1, Yangon. Int J Environ Res Public Health 18.
http: / /dx. doi. or g/10.3 3 9 0 /ij erphl 8 0 2 0 712
Axelsson. G. (1984). Selection bias in studies of spontaneous abortion among occupational groups. J
Occup Med 26: 525-528.
Axelsson. G: Liitz. C: Rvlander. R. (1984). Exposure to solvents and outcome of pregnancy in
university laboratory employees. Rr J Ind Med 41: 305-312.
Axelsson. G: Rvlander. R. (1982). Exposure to anesthetic gases and spontaneous-abortion: Response
bias in a postal questionnaire study. Int J Epidemiol 11: 250-256.
http://dx.doi.Org/10.1093/ije/ll.3.250
Avdemir. S: Akgun. SG: Receren. A: Yuksel. M: Kumas. M: Erdogan. N: Sardas. S: Omurtag. GZ. (2017).
Melatonin ameliorates oxidative DNA damage and protects against formaldehyde-induced
oxidative stress in rats. International Journal of Clinical and Experimental Medicine 10:
6250-6261.
Avdin. S: Canpinar. H: Undeger. U: Gtic. D: Colakoglu. M: Kars. A: Rasaran. N. (2013). Assessment of
immunotoxicity and genotoxicity in workers exposed to low concentrations of
formaldehyde. Arch Toxicol 87: 145-153. http://dx.doi.org/10.1007/s00204-012-Q961-9
Avdin. S: Ogeturk. M: Kuloglu. T: Kavakli. A: Avdin. S. (2014). Effect of carnosine supplementation
on apoptosis and irisin, total oxidant and antioxidants levels in the serum, liver and lung
tissues in rats exposed to formaldehyde inhalation. Peptides 64C: 14-23.
http://dx.doi.Org/10.1016/j.peptides.2014.ll.008
Rabiuk. C: Steinhagen. WH: Rarrow. CS. (1985). Sensory irritation response to inhaled aldehydes
after formaldehyde pretreatment Toxicol Appl Pharmacol 79: 143-149.
http: //dx.doi.org/10.1016/0041 -008xC85190376-x
This document is a draft for review purposes only and does not constitute Agency policy.
R-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Bach. B: Pedersen. OF: M0lhave. L. (1990). Human performance during experimental formaldehyde
exposure. Environ Int 16: 105-113. http://dx.doi.org/10.1016/0160-4120r90190150-5
Baird. DP. (1988). Using time-to-pregnancy data to study occupational exposures: methodology
[Review], Reprod Toxicol 2: 205-207. http://dx.doi.org/l 0.1016/0890-6238(88190023-8
Baird. DP: Wilcox. AT. (1985). Cigarette smoking associated with delayed conception. JAMA 253:
2979-2983. http://dx.doi.org/10.1001/iama.1985.03350440057031
Baird. DP: Wilcox. AT: Weinberg. CR. (1986). Use of time to pregnancy to study environmental
exposures. Am J Epidemiol 124: 470-480.
Bakar. E: Ulucam. E: Cerkezkavabekir. A. (2015). Protective effects of proanthocyanidin and vitamin
E against toxic effects of formaldehyde in kidney tissue. Biotech Histochem 90: 69-78.
http: //dx.d0i.0rg/l 0.3109!\ 0520295.2014.954620
Ballarin. C: Sarto. F: Giacomelli. L: Bartolucci. GB: Clonfero. E. (1992). Micronucleated cells in nasal
mucosa of formaldehyde-exposed workers. MutatRes Genet Toxicol 280: 1-7.
http://dx.d0i.0rg/l 0.1016/0165-1218C92190012-0
Barbosa. E: Pos Santos. ALA: Peteffi. GP: Schneider. A: Miiller. P: Rovaris. P: Bau. CHP: Linden. R:
Antunes. MY: Charao. MF. (2019). Increase of global PNA methylation patterns in beauty
salon workers exposed to low levels of formaldehyde. Environ Sci Pollut Res Int 26: 1304-
1314. http://dx.d0i.0rg/l0.1007/si 1356-018-3674-7
Bardet. G: Achard. S: Loret. T: Pesauziers. V: Momas. I: Seta. N. (2014). A model of human nasal
epithelial cells adapted for direct and repeated exposure to airborne pollutants. Toxicol Lett
229: 144-149. http://dx.doi.Org/10.1016/i.toxlet2014.05.023
Barrow. CS. (1983). Respiratory and metabolic response of rats and mice to formalin vapor [Letter],
Toxicology 28: 357-359. http://dx.d0i.0rg/l 0.1016/0300-483Xf83190009-4
Barrow. CS: Steinhagen. WH: Chang. TCF. (1983). Formaldehyde sensory irritation. In JE Gibson
(Ed.), Chemical Industry Institute of Toxicology series (pp. 16-25). Washington, PC:
Hemisphere Publishing.
Basler. A: Wvd. H: Scheutwinkel-Reich. M. (1985). Formaldehyde-induced sister chromatid
exchanges in vitro and the influence of the exogenous metabolizing systems S9 mix and
primary rat hepatocytes. Arch Toxicol 58: 10-13. http://dx.doi.org/10.1007/bf00292609
Bassig. BA: Zhang. L: Vermeulen. R: Tang. X: Li. G: Hu. W. ei: Guo. W: Purdue. MP: Yin. S: Rappaport.
SM: Shen. M. in: Ti. Z: Oiu. C: Ge. Y: Hosgood. HP: Reiss. B: Wu. B: Xie. Y: Li. L: Yue. F. ei:
Freeman. LEB: Blair. A: Hayes. RB: Huang. H: Smith. MT: Rothman. N: Lan. 0. (2016).
Comparison of hematological alterations and markers of B-cell activation in workers
exposed to benzene, formaldehyde and trichloroethylene. Carcinogenesis 37: 692-700.
http ://dx. doi. or g/10.10 9 3 /carcin/bgwO 5 3
Batalha. TRF: Guimaraes. ET: Lobo. PTA: Lichtenfels. ATF. C: Peur. T: Carvalho. HA: Alves. ES:
Pomingos. M: Rodrigues. GS: Saldiva. PHN. (1999). Exploring the clastogenic effects of air
pollutants in Sao Paulo (Brazil) using the Tradescantia micronuclei assay. PNA Repair 426:
229-232.
Battelle. (1981). Final report on a chronic inhalation toxicology study in rats and mice exposed to
formaldehyde to Chemical Industry Institute of Toxicology: Volume 1. Research Triangle
Park, NC: Chemical Industry Institute of Toxicology.
This document is a draft for review purposes only and does not constitute Agency policy.
R-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Battelle. (1982). A chronic inhalation toxicology study in rats and mice exposed to formaldehyde.
Research Triangle Park, NC: Chemical Industry Institute of Toxicology.
Bauchinger. M: Schmid. E. (1985). Cytogenetic effects in lymphocytes of formaldehyde workers of a
paper factory. MutatRes Genet Toxicol 158: 195-199. http://dx.doi.org/10.1016/Q165-
1218C85190085-0
Beane Freeman. LE: Blair. A: Lubin. TH: Stewart. PA: Hayes. RB: Hoover. RN: Hauptmann. M. (2013).
Mortality from solid tumors among workers in formaldehyde industries: an update of the
NCI cohort Am J Ind Med 56: 1015-1026. http: / /dx. doi. o r g/10.10 0 2 /aj im. 2 2 214
Beane Freeman. LE: Blair. A: Lubin. TH: Stewart. PA: Hayes. RB: Hoover. RN: M. H. (2009). Mortality
from lymphohematopoietic malignancies among workers in formaldehyde industries: The
National Cancer Institute Cohort. J Natl Cancer Inst 101: 751-761.
http: / / dx. do i. o r g /10.10 9 3 /i nci/di p 0 9 6
Beland. FA: Fullerton. NF: Heflich. RH. (1984). Rapid isolation, hydrolysis and chromatography of
formaldehyde-modified DNA. J Chromatogr A 308: 121-131.
http://dx.doi.Org/l 0.1016/0378-4347r84180202-9
Bellavia. A: Dickerson. AS: Rotem. RS: Hansen. I: Gredal. 0: Weisskopf. MG. (2021). Joint and
interactive effects between health comorbidities and environmental exposures in predicting
amyotrophic lateral sclerosis. Int J Hyg Environ Health 231: 113655.
http://dx.doi.Org/10.1016/j.ijheh.2020.113655
Bellisario. V: Mengozzi. G: Grignani. E: Bugiani. M: Sapino. A: Bussolati. G: Bono. R. (2016). Towards
a formalin-free hospital. Levels of 15-F2t-isoprostane and malondialdehyde to monitor
exposure to formaldehyde in nurses from operating theatres. Toxicology Research 5: 1122-
1129. http://dx.doi.Org/10.1039/c6tx00068a
Bender. TR: Mullin. LS: Grapel. GT: Wilson. WE. (1983). Eye irritation response of humans to
formaldehyde. Am Ind Hyg Assoc J 44: 463-465.
http://dx.doi.org/10.1080/15298668391405139
Bentaveb. M: Norback. D: Bednarek. M: Bernard. A: Cai. G: Cerrai. S: Eleftheriou. KK: Gratziou. C:
Hoist. GT: Lavaud. F: Nasilowski. I: Sestini. P: Sarno. G: Sigsgaard. T: Wieslander. G: Zielinski.
1: Viegi. G: Annesi-Maesano. I: Study. G. (2015). Indoor air quality, ventilation and
respiratory health in elderly residents living in nursing homes in Europe. Eur Respir J 45:
1228-1238. http://dx.doi.Org/10.1183/09031936.00082414
Berglund. B: Hoglund. A: Esfandabad. HS. (2012). A bisensory method for odor and irritation
detection of formaldehyde and pyridine. Chemosensory Perception 5: 146-157.
http: //dx.doi.oi-g/10.1007/sl 2078-011-9101 -9
Berglund. B: Nordin. S. (1992). Detectability and perceived intensity for formaldehyde in smokers
and non-smokers. Chem Senses 17: 291-306. http://dx.doi. or g/10.109 3 /chemse/17.3.291
Bermudez. E: Chen. Z: Gross. EA: Walker. CL: Recio. L: Pluta. L: Morgan. KT. (1994). Characterization
of cell lines derived from formaldehyde-induced nasal tumors in rats. Mol Carcinog 9: 193-
199. http://dx.doi.org/10.1002/mc.2940090403
Bernardini. L: Barbosa. E: Charao. MF: Goethel. G: Muller. D: Bau. C: Steffens. NA: Stein. CS: Moresco.
RN: Garcia. SC: Vencato. MS: Brucker. N. (2020). Oxidative damage, inflammation, genotoxic
effect, and global DNA methylation caused by inhalation of formaldehyde and the purpose
of melatonin. Toxicology Research 9: 778-789. http: //dx.doi.org/10.1093 /toxres/tfaa079
This document is a draft for review purposes only and does not constitute Agency policy.
R-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Berrino. F: Richiardi. L: Boffetta. P: Esteve. T: Belletti. I: Raymond. L: Troschel. L: Pisani. P: Zubiri. L:
Ascunce. N: Guberan. E: Tuvns. A: Terracini. B: Merletti. F: Group. MTW. (2003). Occupation
and larynx and hypopharynx cancer: A job-exposure matrix approach in an international
case-control study in France, Italy, Spain and Switzerland. Cancer Causes Control 14: 213-
223. http://dx.doi.Org/10.1023/a:1023661206177
Bhalla. DK: Mahavni. V: Nguyen. T: McClure. T. (1991). Effects of acute exposure to formaldehyde on
surface morphology of nasal epithelia in rats. J Toxicol Environ Health 33: 171-188.
http://dx.doi.Org/l 0.1080/15287399109531516
Biagini. RE: Moorman. WT: Knecht. EA: Clark. TC: Bernstein. IL. (1989). Acute airway narrowing in
monkeys from challenge with 2.5 ppm formaldehyde generated from formalin. Arch
Environ Health 44: 12-17. http: //dx.doi.org/10.1080/00039896.1989.9935866
Bian. RX: Han. TY: Kim. IK: Choi. IS: Lee. SG: Park. IS: Tung. YD. (2012). The effect of chronic
formaldehyde exposure on the hippocampus in chronic cerebral hypoperfusion rat model.
Toxicol Environ Chem 94: 1211-1224. http: //dx.doi.org/10.1080/02772248.2012.691505
Billionnet. C: Gay. E: Kirchner. S: Levnaert. B: Annesi-Maesano. I. (2011). Quantitative assessments
of indoor air pollution and respiratory health in a population-based sample of French
dwellings. Environ Res 111: 425-434. http: / /dx.doi. or g/10.1016/i.envres.2011.02.008
Binawara. BK: Raninee. CS: Mathur. KC: Sharma. H: Goval. K. (2010). Acute effect of formalin on
pulmonary function tests in medical students. Pak J Physiol 6: 8-10.
Binzak. BA: Vocklev. TG: Tenkins. RB: Vocklev. I. (2000). Structure and analysis of the human
dimethylglycine dehydrogenase gene. Mol Genet Metab 69: 181-187.
http://dx.doi.Org/10.1006/mgme.2000.2980
Blackburn. GR: Doolev. TFI: Schreiner. CA: Mackerer. C. (1991). Specific identification of
formaldehyde-mediated mutagenicity using the mouse lymphoma L5178Y TK +/- assay
supplemented with formaldehyde dehydrogenase. In Vitro Toxicol 4: 121-132.
Blair. A: Zheng. T: Linos. A: Stewart. PA: Zhang. YW: Cantor. KP. (2001). Occupation and leukemia: A
population-based case-control study in Iowa and Minnesota. Am J Ind Med 40: 3-14.
http ://dx. doi. or g/10.10 0 2 /ai im. 10 66
Blasiak. I: Trzeciak. A: Malecka-Panas. E: Drzewoski. I: Woiewodzka. M. (2000). In vitro genotoxicity
of ethanol and acetaldehyde in human lymphocytes and the gastrointestinal tract mucosa
cells. Toxicol In Vitro 14: 287-295. http://dx.doi.org/10.1016/S0887-2333(00)00022-9
Boffetta. P: Stellman. SD: Garfinkel. L. (1989). A case-control study of multiple myeloma nested in
the American Cancer Society prospective study. Int J Cancer 43: 554-559.
http://dx.doi.Org/10.1002/ijc.2910430404
Bogdanffv. MS: Morgan. PH: Starr. TB: Morgan. KT. (1987). Binding of formaldehyde to human and
rat nasal mucus and bovine serum albumin. Toxicol Lett 38: 145-154.
http: //dx.doi.org/10.1016/0378-4274C87190122-6
Bogdanffv. MS: Randall. HW: Morgan. KT. (1986). Histochemical localization of aldehyde
dehydrogenase in the respiratory tract of the Fischer-344 rat Toxicol Appl Pharmacol 82:
560-567. http://dx.doi.org/10.1016/0041 -008Xf86190291 -7
Bogdanffv. MS: Sarangapani. R: Plowchalk. PR: Tarabek. AM: Andersen. ME. (1999). A biologically
based risk assessment for vinyl acetate-induced cancer and noncancer inhalation toxicity.
Toxicol Sci 51: 19-35. http://dx.doi.org/10.1093/toxsci/51.1.19
This document is a draft for review purposes only and does not constitute Agency policy.
R-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Supplemental Information for Formaldehyde—Inhalation
Boia. TW: Nielsen. TA: Foldvarv. E: Truitt. EB. Tr. (1985). Acute low-level formaldehyde behavioural
and neurochemical toxicity in the rat. Prog Neuropsychopharmacol Biol Psychiatry 9: 671-
674. http://dx.doi.Org/l 0.1016/0278-5846C85190038-7
Bokina. AI: Eksler. ND: Semenenko. AD: Merkur'veva. RV. (1976). Investigation of the mechanism of
action of atmospheric pollutants on the central nervous system and comparative evaluation
of methods of study. Environ Health Perspect 13: 37-42.
http://dx.doi.org/10.2307/3428235
Bonassi. S: El-Zein. R: Bolognesi. C: Fenech. M. (2011). Micronuclei frequency in peripheral blood
lymphocytes and cancer risk: evidence from human studies [Review], Mutagenesis 26: 93-
100. http://dx.doi.org/10.1093/mutage/geq075
Bonassi. S: Lando. C: Ceppi. M: Landi. S: Rossi. AM: Barale. R. (2004a). No association between
increased levels of high-frequency sister chromatid exchange cells (HFCs) and the risk of
cancer in healthy individuals. Environ Mol Mutagen 43: 134-136.
http: / /dx. doi. or g/10.10 0 2 /em. 2 0 0 0 6
Bonassi. S: Norppa. H: Ceppi. M: Stromberg. U: Vermeulen. R: Znaor. A: Cebulska-Wasilewska. A:
Fabianova. E: Fucic. A: Gundy. S: Hansteen. IL: Knudsen. LE: Lazutka. 1: Rossner. P: Sram. RT:
Boffetta. P. (2008). Chromosomal aberration frequency in lymphocytes predicts the risk of
cancer: results from a pooled cohort study of 22 358 subjects in 11 countries.
Carcinogenesis 29: 1178-1183. http://dx.doi.org/10.1093/carcin/bgn075
Bonassi. S: Znaor. A: Ceppi. M: Lando. C: Chang. WP: Holland. N: Kirsch-Volders. M: Zeiger. E: Ban. S:
Barale. R: Bigatti. MP: Bolognesi. C: Cebulska-Wasilewska. A: Fabianova. E: Fucic. A: Hagmar.
L: Toksic. G: Martelli. A: Migliore. L: Mirkova. E: Scarfi. MR: Ziino. A: Norppa. H: Fenech. M.
(2007). An increased micronucleus frequency in peripheral blood lymphocytes predicts the
risk of cancer in humans. Carcinogenesis 28: 625-631.
http: / /dx. doi. or g/10.109 3 /carcin/bgl 177
Bonassi. S: Znaor. A: Norppa. H: Hagmar. L. (2004b). Chromosomal aberrations and risk of cancer in
humans: an epidemiologic perspective [Review], Cytogenet Genome Res 104: 376-382.
http: //dx.doi.org/10.1159 /000077519
Bonder. M: Lukasiak. M: Dastvch. 1: Golanski. 1: Watala. C. (2019). Differentiated mitochondrial
function in mouse 3T3 fibroblasts and human epithelial or endothelial cells in response to
chemical exposure. Basic & Clinical Pharmacology & Toxicology Online Pharmacology
Online 124: 199-210. http: / /dx. doi. or g/10.1111 /bcpt. 13117
Bono. R: Munnia. A: Romanazzi. V: Bellisario. V: Cellai. F: Peluso. MEM. (2016). Formaldehyde-
induced toxicity in the nasal epithelia of workers of a plastic laminate plant. Toxicology
Research 5: 752-760. http://dx.doi.org/10.1039/c5tx00478k
Bono. R: Romanazzi. V: Munnia. A: Piro. S: Allione. A: Ricceri. F: Guarrera. S: Pignata. C: Matullo. G:
Wang. P: Giese. RW: Peluso. M. (2010). Malondialdehyde-deoxyguanosine adduct formation
in workers of pathology wards: the role of air formaldehyde exposure. Chem Res Toxicol 23:
1342-1348. http://dx.doi.org/10.1021 /txl 00083x
Bono. R: Vincenti. M: Schiliro'. T: Scursatone. E: Pignata. C: Gilli. G. (2006). N-Methylenvaline in a
group of subjects occupationally exposed to formaldehyde. Toxicol Lett 161: 10-17.
http://dx.doi.Org/10.1016/i.toxlet.2005.07.016
Boreiko. CI: Ragan. PL. (1983). Formaldehyde effects in the C3H/10Ty2 cell transformation assay. In
JE Gibson (Ed.), Formaldehyde toxicity (pp. 63-71). Washington, DC: Hemisphere Publishing
Corporation.
This document is a draft for review purposes only and does not constitute Agency policy.
R-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Bos. PMT: Busschers. M: Arts. THE. (2002). Evaluation of the sensory irritation test (Alarie test) for
the assessment of respiratory tract irritation. J Occup Environ Med 44: 968-976.
http://dx.doi. org/10.1097/01 .iom.0000034342.94005.ee
Bosworth. D: Crofton-Sleigh. C: Venitt. S. (1987). A forward mutation assay using ampicillin-
resistance in Escherichia coli designed for investigating the mutagenicity of biological
samples. Mutagenesis 2: 455-467.
Bouraoui. S: Mougou. S: Brahem. A: Tabka. F: Ben Khelifa. H: Harrabi. I: Mrizek. N: Elghezal. H: Saad.
A. (2013). A combination of micronucleus assay and fluorescence in situ hybridization
analysis to evaluate the genotoxicity of formaldehyde. Arch Environ Contam Toxicol 64:
337-344. http://dx.doi.org/10.1007/s00244-012-9828-6
Bovsen. M: Zadig. E: Digernes. V: Abeler. V: Reith. A. (1990). Nasal mucosa in workers exposed to
formaldehyde: a pilot study. Occup Environ Med 47: 116-121.
http://dx.doi.Org/10.1136/oem.47.2.116
Bracken. Ml: Leasa. DT: Morgan. WKC. (1985). Exposure to formaldehyde: Relationship to
respiratory symptoms and function. Can J Public Health 76: 312-316.
Branco. PT: Nunes. RA: Alvim-Ferraz. MC: Martins. FG: Sousa. SI. (2015). Children's exposure to
indoor air in urban nurseries - Part II: Gaseous pollutants' assessment. Environ Res 142:
662-670. http://dx.doi.Org/10.1016/i.envres.2015.08.026
Branco. PTB. S: Alvim-Ferraz. MCM: Martins. FG: Ferraz. C: Vaz. LG: Sousa. SIV. (2020). Impact of
indoor air pollution in nursery and primary schools on childhood asthma. Sci Total Environ
745: 140982. http://dx.doi.Org/10.1016/i.scitotenv.2020.140982
Braun-Fahrlander. C: Wiithrich. B: Gassner. M: Grize. L: Sennhauser. FH: Varonier. HS: Vuille. TC.
(1997). Validation of a rhinitis symptom questionnaire (ISAAC core questions) in a
population of Swiss school children visiting the school health services. Pediatric Allergy and
Immunology 8: 75-82. http://dx.doi.Org/10.llll/i.1399-3038.1997.tb00147.x
Broder. I: Corey. P: Brasher. P: Lipa. M: Cole. P. (1988a). Comparison of health of occupants and
characteristics of houses among control homes and homes insulated with urea
formaldehyde foam: III. Health and house variables following remedial work. Environ Res
45: 179-203. http://dx.doi.org/l0.1016/S0013-9351C88180046-X
Broder. I: Corey. P: Cole. P: Lipa. M: Mintz. S: Nethercott. TR. (1988b). Comparison of health of
occupants and characteristics of houses among control homes and homes insulated with
urea formaldehyde foam: I Methodology. Environ Res 45: 141-155.
http://dx.doi. org/10.1016/S0013-9351C88180044-6
Broder. I: Corey. P: Cole. P: Lipa. M: Mintz. S: Nethercott. TR. (1988c). Comparison of health of
occupants and characteristics of houses among control homes and homes insulated with
urea formaldehyde foam: II initial health and house variables and exposure-response
relationships. Environ Res 45: 156-178. http://dx.doi.org/10.1016/S0013-9351f88180045-
8
Brondeau. MT: Bonnet. P: Guenier. TP: Simon. P: de Ceaurriz. 1. (1990). Adrenal-dependent
leucopenia after short-term exposure to various airborne irritants in rats. J Appl Toxicol 10:
83-86. http://dx.doi.org/10.1002/iat2550100204
Bruno. E: Somma. G: Russo. C: Porozai. D: Pietroiusti. A: Alessandrini. M: Magrini. A. (2018). Nasal
cytology as a screening tool in formaldehyde-exposed workers. Occup Med (Lond) 68: 307-
313. http://dx.doi.org/10.1093/occmed/kqy052
This document is a draft for review purposes only and does not constitute Agency policy.
R-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Brusick. DT. (1983). Genetic and transforming activity of formaldehyde. In JE Gibson (Ed.),
Formaldehyde toxicity (pp. 72-84). Washington, DC: Hemisphere Publishing.
Buckley. LA: liang. XZ: Tames. RA: Morgan. KT: Barrow. CS. (1984). Respiratory tract lesions induced
by sensory irritants at the RD50 concentration. Toxicol Appl Pharmacol 74: 417-429.
http://dx.doi.org/10.1016/0041-008Xr84190295-3
Buckton. KE: Evans. HI. (1973). Methods for the Analysis of Human Chromosome Aberrations.
Geneva: WHO.
Burgaz. S: Cakmak. G: Erdem. 0: Yilmaz. M: Karakava. AE. (2001). Micronuclei frequencies in
exfoliated nasal mucosa cells from pathology and anatomy laboratory workers exposed to
formaldehyde. Neoplasma 48: 144-147.
Burgaz. S: Erdem. 0: Cakmak. G: Erdem. N: Karakava. A: Karakava. AE. (2002). Cytogenetic analysis
of buccal cells from shoe-workers and pathology and anatomy laboratory workers exposed
to n-hexane, toluene, methyl ethyl ketone and formaldehyde. Biomarkers 7: 151-161.
http://dx.doi.Org/l 0.1080!\ 3547500110113242
Burgos-Barragan. G: Wit. N: Meiser. 1: Dingier. FA: Pietzke. M: Mulderrig. L: Pontel. LB: Rosado. IV:
Brewer. TF: Cordell. RL: Monks. PS: Chang. CI: Vazquez. A: Patel. KT. (2017). Mammals divert
endogenous genotoxic formaldehyde into one-carbon metabolism. Nature 548: 549-554.
http: / /dx. doi. or g/10.10 3 8 /natur e2 3 481
Burnev. PG: Laitinen. LA: Perdrizet. S: Huckauf. H: Tattersfield. AE: Chinn. S: Poisson. N: Heeren. A:
Britton. TR: Tones. T. (1989). Validity and repeatability of the IUATLD (1984) Bronchial
Symptoms Questionnaire: an international comparison. Eur Respir J 2: 940-945.
Campbell Tr. 1: Gentry. PR: Clewell III. HI: Andersen. ME. (2020). Akinetic analysis of DNA-deoxy
guanine adducts in the nasal epithelium produced by inhaled formaldehyde in rats-
assessing contributions to adduct production from both endogenous and exogenous sources
of formaldehyde. Toxicol Sci 177: 325-333. http: //dx.doi.org/10.1093 /toxsci/kfaal22
Cao. FH: Cai. 1: Liu. ZM: Li. H: You. HH: Mei. YF: Yang. X: Ding. SM. (2015). [Toxic effect of
formaldehyde on mouse different brain regions], Sheng Li Xue Bao 67: 497-504.
Cap. P: Drvahina. K: Pehal. F: Spanel. P. (2008). Selected ion flow tube mass spectrometry of exhaled
breath condensate headspace. Rapid Commun Mass Spectrom 22: 2844-2850.
http: / /dx. doi. or g/10.10 0 2 /rem. 3685
Caria. H: Chaveca. T: Laires. A: Rueff. 1. (1995). Genotoxicity of quercetin in the micronucleus assay
in mouse bone marrow erythrocytes, human lymphocytes, V79 cell line and identification of
kinetochore-containing (CREST staining) micronuclei in human lymphocytes. Mutat Res
343: 85-94.
Casanova-Schmitz. M: David. RM: Heck. H. (1984a). Oxidation of formaldehyde and acetaldehyde by
NAD+-dependent dehydrogenases in rat nasal mucosal homogenates. Biochem Pharmacol
33: 1137-1142. http://dx.doi.org/l0.1016/0006-2952C84190526-4
Casanova-Schmitz. M: Heck. H. (1983). Effects of formaldehyde exposure on the extractability of
DNA from proteins in the rat nasal mucosa. Toxicol Appl Pharmacol 70: 121-132.
Casanova-Schmitz. M: Starr. TB: Heck. HP. (1984b). Differentiation between metabolic
incorporation and covalent binding in the labeling of macromolecules in the rat nasal
mucosa and bone marrow by inhaled [14C]- and [3H]formaldehyde. Toxicol Appl Pharmacol
76: 26-44. http: //dx.doi.org/10.1016/0041 -008xr84190026-7
This document is a draft for review purposes only and does not constitute Agency policy.
R-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Casanova. M: Bell. DA: Heck. H. (1997). Dichloromethane metabolism to formaldehyde and reaction
of formaldehyde with nucleic acids in hepatocytes of rodents and humans with and without
glutathione S-transferase T1 and Ml genes. Fundam Appl Toxicol 37: 168-180.
http: / /dx. doi. or g/10.109 3 /toxsci/3 7.2.168
Casanova. M: Devo. DF: Heck. H. (1989). Covalent binding of inhaled formaldehyde to DNA in the
nasal mucosa of Fischer 344 rats: Analysis of formaldehyde and DNA by high-performance
liquid chromatography and provisional pharmacokinetic interpretation. Fundam Appl
Toxicol 12: 397-417. http://dx.doi.Org/l 0.1016/0272-0590(89190015-8
Casanova. M: Heck. H. (1987). Further studies of the metabolic incorporation and covalent binding
of inhaled [3H]- and [14C]formaldehyde in Fischer-344 rats: Effects of glutathione
depletion. Toxicol Appl Pharmacol 89: 105-121. http: / /dx. doi. or g /10.1016 /0 0 41 -
008XC87190181-5
Casanova. M: Heck. H. (1997). Lack of evidence for the involvement of formaldehyde in the
hepatocarcinogenicity of methyl tertiary-butyl ether in CD-I mice. Chem Biol Interact 105:
131-143.
Casanova. M: Heck. H: Everitt. II: Harrington. WW. Tr: Popp. TA. (1988). Formaldehyde
concentrations in the blood of rhesus monkeys after inhalation exposure. Food Chem
Toxicol 26: 715-716. http: //dx.doi.org/10.1016/0278-6915C88190071 -3
Casanova. M: Morgan. KT: Gross. EA: Moss. OR: Heck. H. (1994). DNA-protein cross-links and cell
replication at specific sites in the nose of F344 rats exposed subchronically to
formaldehyde. Fundam Appl Toxicol 23: 525-536.
http: / /dx. doi.org/10.10 0 6 /faat. 199 4.1137
Casanova. M: Morgan. KT: Steinhagen. WH: Everitt. II: Popp. TA: Heck. H. (1991). Covalent binding of
inhaled formaldehyde to DNA in the respiratory tract of rhesus monkeys: pharmacokinetics,
rat-to-monkey interspecies scaling, and extrapolation to man. Toxicol Sci 17: 409-428.
http://dx.doi.Org/10.1016/0272-0590r91190230-2
Cassee. FR: Arts. THE: Groten. TP: Feron. VI. (1996a). Sensory irritation to mixtures of formaldehyde,
acrolein, and acetaldehyde in rats. Arch Toxicol 70: 329-337.
http://dx.doi.Org/10.1007/s002040050282
Cassee. FR: Feron. VI. (1994a). Biochemical and histopathological changes in nasal epithelium of
rats after 3-day intermittent exposure to formaldehyde and ozone alone or in combination.
Toxicol Lett 72: 257-268. http://dx.doi.org/10.1016/0378-4274r94190037-x
Cassee. FR: Feron. VI. (1994b). Histopathological and biochemical changes in nasal epithelium of
rats after 3-day intermittent exposure to a mixture of ozone and formaldehyde. 195: 142.
Cassee. FR: Groten. TP: Feron. VI. (1996b). Changes in the nasal epithelium of rats exposed by
inhalation to mixtures of formaldehyde, acetaldehyde, and acrolein. Toxicol Sci 29: 208-218.
http: / /dx. doi.org/10.10 0 6 /faat. 1996.0024
Casset. A: Marchand. C: Purohit. A: Le Calve. S: Donnav. C: Meyer. P: Pauli. G: De Blav. F. (2006a).
Low exposure to inhaled formaldehyde: Effect on allergen bronchial response in asthmatics
sensitized to mite [Abstract], J Allergy Clin Immunol 117: S23-S23.
Casset. A: Marchand. C: Purohit. A: le Calve. S: Uring-Lambert. B: Donnav. C: Meyer. P: de Blav. F.
(2006b). Inhaled formaldehyde exposure: effect on bronchial response to mite allergen in
sensitized asthma patients. Allergy 61: 1344-1350. http://dx.doi.org/10.1 111/i.1398-
9995.2006.01174.x
This document is a draft for review purposes only and does not constitute Agency policy.
R-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Cederbaum. AI: Oureshi. A. (1982). Role of catalase and hydroxy! radicals in the oxidation of
methanol by rat liver microsomes. Biochem Pharmacol 31: 329-335.
http://dx.doi.Org/l 0.1016/0006-2952r82190179-4
Chan. CF: Sun. WZ: Lin. IK: Lin-Shiau. SY. (2000). Activation of transcription factors of nuclear factor
kappa B, activator protein-1 and octamer factors in hyperalgesia. Eur J Pharmacol 402: 61-
68. http://dx.doi.org/10.1016/S0014-2999r00100431-3
Chang. TCF: Barrow. CS. (1984). Sensory irritation tolerance and cross-tolerance in F-344 rats
exposed to chlorine or formaldehyde gas. Toxicol Appl Pharmacol 76: 319-327.
http://dx.doi.org/10.1016/0041-008Xr84190013-9
Chang. TCF: Gross. EA: Swenberg. TA: Barrow. CS. (1983). Nasal cavity deposition, histopathology,
and cell proliferation after single or repeated formaldehyde exposures in B6C3F1 mice and
F-344 rats. Toxicol Appl Pharmacol 68: 161-176. http: //dx.doi.org/10.1016/0Q41-
008xr83190001-7
Chang. TCF: Steinhagen. WH: Barrow. CS. (1981). Effect of single or repeated formaldehyde
exposure on minute volume of B6C3F1 mice and F-344 rats. Toxicol Appl Pharmacol 61:
451-459. http://dx.doi.Org/10.1016/0041-008xr81190368-9
Chang. M: Park. H: Ha. M: Hong. YC: Lim. YH: Kim. Y: Kim. YT: Lee. D: Ha. EH. (2017). The effect of
prenatal TVOC exposure on birth and infantile weight: the Mothers and Children's
Environmental Health study. Pediatr Res 82: 423-428.
http: / /dx. doi. or g/10.10 3 8 /pr. 2 017.5 5
Chatzidiakou. L: Mumovic. D: Summerfield. AT: Hong. SM: Altamirano-Medina. H. (2014). A Victorian
school and a low carbon designed school: Comparison of indoor air quality, energy
performance, and student health. Indoor Built Environ 23: 417-432.
http://dx.doi.org/10.1177/1420326X14532388
Chaw. YF: Crane. LE: Lange. P: Shapiro. R. (1980). Isolation and identification of cross-links from
formaldehyde-treated nucleic acids. Biochemistry 19: 5525-5531.
Checkowav. H: Dell. LP: Boffetta. P: Gallagher. AE: Crawford. L: Lees. PS: Mundt. KA. (2015).
Formaldehyde Exposure and Mortality Risks From Acute Myeloid Leukemia and Other
Lymphohematopoietic Malignancies in the US National Cancer Institute Cohort Study of
Workers in Formaldehyde Industries. J Occup Environ Med 57: 785-794.
http://dx.doi.Org/10.1097/TOM.0000000000000466
Chen. D: Fang. L: Mei. S: Li. H: Xu. X: Pes Marais. TL: Lu. K: Liu. XS: Tin. C. (2017). Regulation of
Chromatin Assembly and Cell Transformation by Formaldehyde Exposure in Human Cells.
Environ Health Perspect 125: 097019. http://dx.doi.org/10.1289/EHP1275
Cheng. G: Shi. Y: Sturla. ST: Talas. I. R.: Mcintee. El: Villalta. PW: Wang. M: Hecht. SS. (2003). Reactions
of formaldehyde plus acetaldehyde with deoxy guano sine and DNA: formation of cyclic
deoxyguanosine adducts and formaldehyde cross-links. Chem Res Toxicol 16: 145-152.
http: / /dx. doi. or g/10.10 21 /txO 2 5 614r
Cheng. G: Wang. M: Upadhyava. P: Villalta. PW: Hecht. SS. (2008). Formation of formaldehyde
adducts in the reactions of DNA and deoxyribonucleosides with alpha-acetates of 4-
(methylnitrosamino)-l-(3-pyridyl)-l-butanone (NNK), 4-(methylnitrosamino)-l-(3-
pyridyl)-l-butanol (NNAL), and N-nitrosodimethylamine (NDMA). Chem Res Toxicol 21:
746-751. http://dx.doi.org/10.1021/tx7003823
This document is a draft for review purposes only and does not constitute Agency policy.
R-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Cheng. T: Zhang. L: Tang. Y: Li. Z. (2016). The toxicity of continuous long-term low-dose
formaldehyde inhalation in mice. Immunopharmacol Immunotoxicol 38: 495-501.
http://dx.doi. org/10.1080/08923973.2016.1248844
Cheng. Z: Li. Y: Liang. B: Wang. C. (2004). [Investigation of formaldehyde level and health of
personnel in clinical pathology], 29: 266-267.
Chia. SE: Ong. CN: Foo. SC: Lee. HP. (1992). Medical students' exposure to formaldehyde in a gross
anatomy dissection laboratory. J Am Coll Health 41: 115-119.
http://dx.doi.org/10.1080/07448481.1992.9936310
Choi. DW: Moon. KW: Bveon. SH: Lee. EI: Sul. DG: Lee. TH: Oh. EH: Kim. YH. (2009). Indoor volatile
organic compounds in atopy patients' houses in South Korea. Indoor Built Environ 18: 144-
154. http://dx.doi.org/10.1177/1420326X08101945
Chonglei. L: Fan. W: Wei. L: Yihe. 1. (2012). Effects of Exposure to VOCs on Spatial Learning and
Memory Capacity and the Expression of NMDA Receptor in Mice. Journal of Animal and
Veterinary Advances 11: 3355-3364.
Ciftci. G: Aksov. A: Cenesiz. S: Sogut. MU: Yarim. GF: Nisbet. C: Guvenc. D: Ertekin. A. (2015).
Therapeutic role of curcumin in oxidative DNA damage caused by formaldehyde. Microsc
Res Tech 78: 391-395. http://dx.doi.org/10.1002/iemt.22485
CUT. (1999). Formaldehyde: Hazard characterization and dose-response assessment for
carcinogenicity by the route of inhalation (revised edition). Research Triangle Park, NC.
Clarisse. B: Laurent. AM: Seta. N: Le Moullec. Y: El Hasnaoui. A: Momas. I. (2003). Indoor aldehydes:
measurement of contamination levels and identification of their determinants in Paris
dwellings. Environ Res 92: 245-253. http://dx.doi.org/10.1016/S0013-9351C03100039-2
Clement. PA: Stoop. AP: Kaufman. L. (1987). The influence of formaldehyde on the nasal mucosa.
Rhinology 25: 29-34.
Coggon. D: Harris. EC: Poole. I: Palmer. KT. (2003). Extended follow-up of a cohort of British
chemical workers exposed to formaldehyde. J Natl Cancer Inst 95: 1608-1615.
http: / /dx. doi. or g/10.10 9 3/i nci/di g046
Coggon. D: Ntani. G: Harris. EC: Palmer. KT. (2014). Upper Airway Cancer, Myeloid Leukemia, and
Other Cancers in a Cohort of British Chemical Workers Exposed to Formaldehyde. Am J
Epidemiol 179: 1301-1311. http://dx.doi.org/10.1093/aje/kwu049
Cohen-Hubal. EA: Schlosser. PM: Conollv. RB: Kimbell. IS. (1997). Comparison of inhaled
formaldehyde dosimetry predictions with DNA-protein cross-link measurements in the rat
nasal passages. Toxicol Appl Pharmacol 143: 47-55.
Conawav. CC: Whvsner. 1: Verna. LK: Williams. GM. (1996). Formaldehyde mechanistic data and risk
assessment: Endogenous protection from DNA adduct formation [Review], Pharmacol Ther
71: 29-55. http://dx.doi.org/10.1016/0163-7258r96100061-7
Connor. TH: Barrie. MP: Theiss. TC: Matnev. TS: Ward. IB. (1983). Mutagenicity of formalin in the
Ames assay. MutatRes Lett 119: 145-149. http: / /dx. doi. o r g/10.1016 / 016 5 -
7992C83190122-7
Connor. TH: Theiss. TC: Hanna. HA: Monteith. DK: Matnev. TS. (1985a). Genotoxicity of organic
chemicals frequently found in the air of mobile homes. Toxicol Lett 25: 33-40.
http://dx.doi.org/10.1016/0378-4274r85190097-9
This document is a draft for review purposes only and does not constitute Agency policy.
R-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Connor. TH: Ward. TB: Legator. MS. (1985b). Absence of mutagenicity in the urine of autopsy service
workers exposed to formaldehyde: Factors influencing mutagenicity testing of urine. Int
Arch Occup Environ Health 56: 225-237. http://dx.doi.org/10.1007/BF00396600
Conollv. RB: Kimbell. IS: lanszen. D: Schlosser. PM: Kalisak. D: Preston. 1: Miller. FT. (2003).
Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344
rat. Toxicol Sci 75: 432-447. http://dx.doi.org/10.1093/toxsci/kfgl82
Conollv. RB: Lilly. PD: Kimbell. IS. (2000). Simulation modelling of the tissue disposition of
formaldehyde to predict nasal DNA-protein cross-links in Fischer 344 rats, rhesus monkeys,
and humans. Environ Health Perspect 108: 919-924. http: //dx.doi.org/10.23Q7/3454325
Coon. RA: Tones. RA: Tenkins. LI. Tr: Siegel. 1. (1970). Animal inhalation studies on ammonia, ethylene
glycol, formaldehyde, dimethylamine, and ethanol. Toxicol Appl Pharmacol 16: 646-655.
http://dx.doi.Org/10.1016/0041-008Xf70190069-4
Coonev. MA: Buck Louis. GM: Sundaram. R: McGuiness. BM: Lynch. CD. (2009). Validity of self-
reported time to pregnancy. Epidemiology 20: 56-59.
http://dx.doi.Org/l 0.1097/EDE.0b013e31818ef47e
Corlev. RA: Kabilan. S: Kuprat. AP: Carson. TP: Tacob. RE: Minard. KR: Teeguarden. TG: Timchalk. C:
Pipavath. S: Glennv. R: Einstein. DR. (2015). Comparative risks of aldehyde constituents in
cigarette smoke using transient computational fluid dynamics/physiologically based
pharmacokinetic models of the rat and human respiratory tracts. Toxicol Sci 146: 65-88.
http: / /dx. doi. or g/10.109 3 /toxsci /kfvO 71
Cosma. GN: Marchok. AC. (1988). Benzo[a]pyrene- and formaldehyde-induced DNA damage and
repair in rat tracheal epithelial cells. Toxicology 51: 309-320.
http://dx.d0i.0rg/l 0.1016/0300-483Xf88190159-X
Cosma. GN: Wilhite. AS: Marchok. AC. (1988). The detection of DNA-protein cross-links in rat
tracheal implants exposed in vivo to benzo[a]pyrene and formaldehyde. Cancer Lett 42: 13-
21. http://dx.doi.org/10.1016/0304-3835r88190233-9
Costa. M: Zhitkovich. A: Harris. M: Paustenbach. D: Gargas. M. (1997). DNA-protein cross-links
produced by various chemicals in cultured human lymphoma cells. J Toxicol Environ Health
50: 433-449. http://dx.d0i.0rg/l0.1080/00984109708984000
Costa. S: Carvalho. S: Costa. C: Coelho. P: Silva. S: Santos. LS: Gaspar. IF: Porto. B: Laffon. B: Teixeira.
TP. (2015). Increased levels of chromosomal aberrations and DNA damage in a group of
workers exposed to formaldehyde. Mutagenesis 30: 463-473.
http: / /dx. doi. or g /10.109 3 /mutage / ge vO 0 2
Costa. S: Coelho. P: Costa. C: Silva. S: Mayan. 0: Santos. LS: Gaspar. 1: Teixeira. TP. (2008). Genotoxic
damage in pathology anatomy laboratory workers exposed to formaldehyde. Toxicology
252: 40-48. http://dx.doi.Org/10.1016/i.tox.2008.07.056
Costa. S: Costa. C: Madureira. 1: Valdiglesias. V: Teixeira-Gomes. A: Guedes de Pinho. P: Laffon. B:
Teixeira. TP. (2019). Occupational exposure to formaldehyde and early biomarkers of cancer
risk, immunotoxicity and susceptibility. Environ Res 179: 108740.
http: / /dx. doi. or g/10.1016/i. envres. 2019.108740
Costa. S: Garcia-Leston. 1: Coelho. M: Coelho. P: Costa. C: Silva. S: Porto. B: Laffon. B: Teixeira. TP.
(2013). Cytogenetic and immunological effects associated with occupational formaldehyde
exposure. J Toxicol Environ Health A 76: 217-229.
http://dx.d0i.0rg/l 0.1080/15287394.2013.757212
This document is a draft for review purposes only and does not constitute Agency policy.
R-14 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
Costa. S: Pina. C: Coelho. P: Costa. C: Silva. S: Porto. B: Laffon. B: Teixeira. TP. (2011). Occupational
exposure to formaldehyde: ge no toxic risk evaluation by comet assay and micronucleus test
using human peripheral lymphocytes. J Toxicol Environ Health A 74: 1040-1051.
http://dx.doi.org/10.1080/15287394.20n.582293
Craft. TR: Bermudez. E: Skopek. TR. (1987). Formaldehyde mutagenesis and formation of DNA-
protein crosslinks in human lymphoblasts in vitro. Mutat Res 176: 147-155.
http://dx.doi. org/10.1016/0027-5107C87190262-4
Crosby. RM: Richardson. KK: Craft. TR: Benforado. KB: Liber. HL: Skopek. TR. (1988). Molecular
analysis of formaldehyde-induced mutations in human lymphoblasts and E. coli. Environ
Mol Mutagen 12: 155-166. http://dx.doi.org/10.1002/em.2860120202
Crump. KS: Chen. C: Chiu. WA: Louis. TA: Portier. CI: Subramaniam. RP: White. PP. (2010). What
role for biologically based dose-response models in estimating low-dose risk? [Review],
Environ Health Perspect 118: 585-588. http://dx.doi.org/10.1289/ehp.0901249
Cui. Y: Li. H: Wu. S: Zhao. R: Du. D: Ding. Y: Nie. H: Ti. HL. (2016). Formaldehyde impairs
transepithelial sodium transport. Sci Rep 6: 35857. http: / /dx. doi. o r g/10.10 3 8 /sr ep3 5 85 7
d'Errico. A: Pasian. S: Baratti. A: Zanelli. R: Alfonzo. S: Gilardi. L: Beatrice. F: Bena. A: Costa. G.
(2009). A case-control study on occupational risk factors for sino-nasal cancer. Occup
Environ Med 66: 448-455. http://dx.doi.org/10.1136/oem.2008.041277
da Silva. CM: Leal. MP: Brochetti. RA: Braga. T: Vitoretti. LB: Saraiva Camara. NO: Damazo. AS:
Ligeiro-De-Oliveira. A: Chavantes. MC: Lino-Dos-Santos-Franco. A. (2015). Low Level Laser
Therapy Reduces the Development of Lung Inflammation Induced by Formaldehyde
Exposure. PLoS ONE 10: e0142816. http://dx.doi.org/10.1371/iournal.pone.0142816
Dalbev. WE. (1982). Formaldehyde and tumors in hamster respiratory tract Toxicology 24: 9-14.
http: //dx.doi.org/10.1016/0300-483Xr82190058-0
Dallas. CE: Badeaux. P: Theiss. TC: Fairchild. El. (1989). The influence of inhaled formaldehyde on rat
lung cytochrome P450. Environ Res 49: 50-59. http://dx.doi.org/10.1016/S0Q13-
9351C89180021-0
Dallas. CE: Mellard. DN: Theiss. TC: Pentecost. AR: Fairchild El. II. (1987). Distribution of DNA and
RNA content in the bone marrow and alveolar macrophages of rats after subchronic
inhalation of formaldehyde. Environ Res 43: 191-202.
Dallas. CE: Scott. MI: Ward. IB. Tr: Theiss. TC. (1992). Cytogenetic analysis of pulmonary lavage and
bone marrow cells of rats after repeated formaldehyde inhalation. J Appl Toxicol 12: 199-
203. http://dx.doi.org/10.1002/iat.2550120309
Dally. KA: Hanrahan. LP: Woodbury. MA: Kanarek. MS. (1981). Formaldehyde exposure in
nonoccupational environments. Arch Environ Occup Health 36: 277-284.
http://dx.doi.Org/10.1080/00039896.1981.10667638
Dannemiller. KC: Murphy. IS: Dixon. SL: Pennell. KG: Suuberg. EM: Tacobs. DE: Sandel. M. (2013).
Formaldehyde concentrations in household air of asthma patients determined using
colorimetric detector tubes. Indoor Air 23: 285-294. http: / / dx. do i. o r g /10.1111 /ina. 12 0 2 4
Day. TH: Lees. RE: Clark. RH: Pattee. PL. (1984). Respiratory response to formaldehyde and off-gas
of urea formaldehyde foam insulation. Can Med Assoc J 131: 1061-1065.
This document is a draft for review purposes only and does not constitute Agency policy.
R-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
de Ceaurriz. T: Micillino. TC: Bonnet. P: Guenier. TP. (1981). [Prediction of the irritant effects of
chemicals on the human respiratory tract: Advantages of an animal model], Cah Notes Doc
102: 55-61.
De Flora. S. (1981). Study of 106 organic and inorganic compounds in the Salmonella/microsome
test. Carcinogenesis 2: 283-298. http://dx.do i. or g /10.109 3 /car cin / 2.4.2 8 3
De Flora. S: Zanacchi. P: Camoirano. A: Bennicelli. C: Badolati. GS. (1984). Genotoxic activity and
potency of 135 compounds in the Ames reversion test and in a bacterial DNA-repair test
[Review], MutatRes 133: 161-198. http://dx.doi.Org/l 0.1016/0165-1110f84190016-2
De long. WH: Arts. THE: De Klerk. A: Schiif. MA: Ezendam. 1: Kuper. CF: Van Loveren. H. (2009).
Contact and respiratory sensitizers can be identified by cytokine profiles following
inhalation exposure. Toxicology 261: 103-111. http://dx.doi.Org/10.1016/j.tox.2009.04.057
de Serres. FT: Brockman. HE. (1999). Comparison of the spectra of genetic damage in formaldehyde-
induced ad-3 mutations between DNA repair-proficient and -deficient heterokaryons of
Neurosporacrassa. MutatRes 437: 151-163. http://dx.doi.org/10.1016/S1383-
5742f99100081-2
Dean. TH: Lauer. LP: House. RV: Murray. MT: Stillman. WS: Irons. RD: Steinhagen. WH: Phelps. MC:
Adams. DO. (1984). Studies of immune fuction and host resistance in B6C3F1 mice exposed
to formaldehyde. Toxicol Appl Pharmacol 72: 519-529. http://dx.doi.org/10.1016/0Q41-
008Xf84190129-7
Dell. L: Teta. MT. (1995). Mortality among workers at a plastics manufacturing and research and
development facility: 1946-1988. AmJInd Med 28: 373-384.
http://dx.doi.Org/10.1002/aiim.4700280307
Demkowicz-Dobrzanski. K: Castonguav. A. (1992). Modulation by glutathione of DNA strand breaks
induced by 4-(methylnitrosamino)-l-(3-pyridyl)-l-butanone and its aldehyde metabolites
in rat hepatocytes. Carcinogenesis 13: 1447-1454.
Deng. 1: Zhang. X: Liu. 1: Wei. X: Liu. X: Ma. L: Zhao. Y: Li. Z. (2020). [Effects of gaseous formaldehyde
fluctuating exposure on medical students' subjective symptoms and pulmonary function],
Weisheng Yanjiu 49: 921-926.
http://dx.doi.Org/10.19813/i.cnki.weishengvaniiu.2020.06.008
Dhareshwar. SS: Stella. VI. (2008). Your prodrug releases formaldehyde: should you be concerned?
No! J Pharm Sci 97: 4184-4193. http://dx.doi.org/10.1002/ips.21319
DHGC. (2010). [Two Acute/ Subacute Inhalation toxicity studies].
Dickey. FH: Cleland. GH: Lotz. C. (1949). The role of organic peroxides in the induction of mutations.
Proc Natl Acad Sci USA 35: 581-586.
Dietrich. CI: Richards. IS: Bernard. TE: Hammad. YY. (1996). Human stress protein response to
formaldehyde exposure. Exp Toxicol Pathol 48: 518-519. http://dx.doi.org/10.1016/S094Q-
2993C96180071-6
Dillon. D: Combes. R: Zeiger. E. (1998). The effectiveness of Salmonella strains TA100, TA102 and
TA104 for detecting mutagenicity of some aldehydes and peroxides. Mutagenesis 13: 19-26.
Dingier. FA: Wang. M: Mu. A: Millington. CL: Oberbeck. N: Watcham. S: Pontel. LB: Kamimae-
Lanning. AN: Langevin. F: Nadler. C: Cordell. RL: Monks. PS: Yu. R: Wilson. NK: Hira. A:
Yoshida. K: Mori. M: Okamoto. Y: Okuno. Y: Muramatsu. H. ideki: Shiraishi. Y: Kobavashi. M:
Moriguchi. T: Osumi. T: Kato. M: Mivano. S: Ito. E: Koiima. S: Yabe. H: Yabe. M: Matsuo. K:
This document is a draft for review purposes only and does not constitute Agency policy.
R-16 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Ogawa. S: Gottgens. B: Hodskinson. MRG: Takata. M: Patel. KT. (2020). Two aldehyde
clearance systems are essential to prevent lethal formaldehyde accumulation in mice and
humans. Mol Cell 80: 996-1012.el019. http://dx.doi.Org/10.1016/i.molcel.2020.10.012
Dinsdale. D: Riley. RA: Verschovle. RD. (1993). Pulmonary cytochrome P450 in rats exposed to
formaldehyde vapor. Environ Res 62: 19-27. http://dx.doi.org/10.1006/enrs.1993.1085
Doolittle. DT: Furlong. TW: Butterworth. BE. (1985). Assessment of chemically induced DNA repair in
primary cultures of human bronchial epithelial cells. Toxicol Appl Pharmacol 79: 28-38.
http://dx.doi.org/10.1016/0041-008Xr85190365-5
Dotterud. LK: Kvammen. B: Lund. E: Falk. ES. (1995). An evaluation of atopic diseases in relation to
immediate skin test reactions among schoolchildren in the Sor-Varanger community. J Eur
Acad Dermatol Venereol 5: 240-249. http://dx.doi.org/10.Ill 1 /j. 1468-
3083.1995.tb00112.x
Douglas. MP: Rogers. SO. (1998). DNA damage caused by common cytological fixatives. Mutat Res
401: 77-88. http: / /dx.doi.org/10.1016/S0027-5107C97100314-X
Dresp. 1: Bauchinger. M. (1988). Direct analysis of the clastogenic effect of formaldehyde in
unstimulated human lymphocytes by means of the premature chromosome condensation
technique. Mutat Res 204: 349-352. http: //dx.doi.org/10.1016/0165-1218C88190110-3
Duan. 1: Kang. 1: Oin. W: Deng. T: Liu. H: Li. B: Yu. W: Gong. S: Yang. X: Chen. M. (2018). Exposure to
formaldehyde and diisononyl phthalate exacerbate neuroinflammation through NF-kB
activation in a mouse asthma model. Ecotoxicol Environ Saf 163: 356-364.
http://dx.doi.Org/10.1016/i.ecoenv.2018.07.089
Duan. 1: Xie. 1: Deng. T: Xie. X: Liu. H: Li. B: Chen. M. (2020). Exposure to both formaldehyde and high
relative humidity exacerbates allergic asthma by activating the TRPV4-p38 MAPK pathway
in Balb/c mice. Environ Pollut 256: 113375.
http://dx.doi.Org/10.1016/i.envpol.2019.113375
Duan. YY. (2011). [Effects of overexpression of heat shock protein 70 on the damage induced by
formaldehyde in vitro], Zhonghua Laodong Weisheng Zhiyebing Zazhi 29: 349-352.
Dumas. 0: Boggs. KM: Ouinot. C: Varraso. R: Zock. TP: Henneberger. PK: Speizer. FE: Le Moual. N:
Camargo. CA. (2020). Occupational exposure to disinfectants and asthma incidence in U.S.
nurses: A prospective cohort study. Am J Ind Med 63: 44-50.
http ://dx. doi. or g/10.10 0 2 /ai im. 2 3 0 67
Dvkewicz. MS: Patterson. R: Cugell. DW: Harris. KE: Wu. AF. (1991). Serum IgE and IgG to
formaldehyde-human serum albumin: Lack of relation to gaseous formaldehyde exposure
and symptoms. J Allergy Clin Immunol 87: 48-57. http: //dx. doi. or g /10.1016/0 091 -
6749C91190212-7
ECETOC (European Centre for Ecotoxicology and Toxicology of Chemicals). (1995). Formaldehyde
and human cancer risk. (CIS/08/01402)
Edling. C: Hellquist. H: Odkvist. L. (1987a). Occupational formaldehyde exposure and the nasal
mucosa. Rhinology 25: 181-187.
Edling. C: Hellquist. H: Odkvist. L. (1988). Occupational exposure to formaldehyde and
histopathological changes in the nasal mucosa. Br J Ind Med 45: 761-765.
http://dx.doi.org/10.1136/oem.45.ll.761
This document is a draft for review purposes only and does not constitute Agency policy.
R-17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Edling. C: Tarvholm. B: Andersson. L: Axelson. 0. (1987b). Mortality and cancer incidence among
workers in an abrasive manufacturing industry. Br J Ind Med 44: 57-59.
http ://dx. doi. or g/10.113 6 /oem.44.1.5 7
Edrissi. B: Taghizadeh. K: Dedon. PC. (2013). Quantitative analysis of histone modifications:
formaldehyde is a source of pathological n(6)-formyllysine that is refractory to histone
deacetylases. PLoS Genet 9: el003328. http://dx.doi.org/10.1371/iournal.pgen.1003328
Edrissi. B: Taghizadeh. K: Moeller. BC: Yu. R: Kracko. D: Dovle-Eisele. M: Swenberg. TA: Dedon. PC.
(2017). N6-Formyllysine as a Biomarker of Formaldehyde Exposure: Formation and Loss of
N6-Formyllysine in Nasal Epithelium in Long-Term, Low-Dose Inhalation Studies in Rats.
Chem Res Toxicol 30: 1572-1576. http://dx.doi.org/10.1021/acs.chemrestox.7b00075
Egle. TL. Jr. (1972). Retention of inhaled formaldehyde, propionaldehyde, and acrolein in the dog.
Arch Environ Health 25: 119-124. http://dx.doi.org/10.1080/00039896.1972.10666147
El-Fekv. AA: Kabbash. IA: Zavet. HH: El-Sallamv. RM. (2020). Health disorders and safety measures
among workers in Tanta Flax and Oil Company, Egypt. Environ Sci Pollut Res Int
http://dx.doi.Org/l 0.1007/sl 1356-020-11588-0
Emri. G: Schaefer. D: Held. B: Herbst. C: Zieger. W: Horkav. I: Baverl. C. (2004). Low concentrations
of formaldehyde induce DNA damage and delay DNA repair after UV irradiation in human
skin cells. Exp Dermatol 13: 305-315. http://dx.doi.Org/10.llll/i.0906-6705.2004.00157.x
Environment Canada. (2000). Priority substances list assessment report Formaldehyde. Ottawa,
Canada: Health Canada. http://publications.gc.ca/collections/Collection/En40-215-50E.pdf
Eom. HI: Liu. YD: Kwak. GS: Heo. M: Song. KS: Chung. YD: Chon. TS: Choi. 1. (2017). Inhalation
toxicity of indoor air pollutants in Drosophila melanogaster using integrated
transcriptomics and computational behavior analyses. Sci Rep 7: 46473.
http://dx.doi.org/10.1038/srep46473
Epstein. SS: Arnold. E: Andrea. 1: Bass. W: Bishop. Y. (1972). Detection of chemical mutagens by the
dominant lethal assay in the mouse. Toxicol Appl Pharmacol 23: 288-325.
http://dx.doi.org/10.1016/0041-008Xr72190192-5
Epstein. SS: Shafner. H. (1968). Chemical mutagens in the human environment. Nature 219: 385-
387. http: / /dx.doi.org/10.1038/219385a0
Er. TK: Lin. CW: Liu. TC: Chen. CC: Wang. LH: Hsieh. LL: Tsai. WC. (2015). Increase EGFR Mutations
Detection Rate in Lung Adenocarcinoma by Real-Time PCR Screening Followed by Direct
Sequencing. Appl Immunohistochem Mol Morphol 23: 343-348.
http: //dx.doi.org/10.1097/PDM.0000000000000037
Erdei. E: Bobvos. 1: Brozik. M: Paldv. A: Farkas. I: Vaskovi. E: Rudnai. P. (2003). Indoor air pollutants
and immune biomarkers among Hungarian asthmatic children. Arch Environ Occup Health
58: 337-347.
Esterbauer. H: Cheeseman. KH: Dianzani. MU: Poli. G: Slater. TF. (1982). Separation and
characterization of the aldehydic products of lipid peroxidation stimulated by ADP-Fe2+ in
rat liver microsomes. Biochem J 208: 129-140.
Ezrattv. V: Bonav. M: Neukirch. C: Orset-Guillossou. G: Dehoux. M: Koscienlnv. S: Cabanes. PA:
Lambrozo. I: Aubier. M. (2007). Effect of formaldehyde on asthmatic response to inhaled
allergen challenge. Environ Health Perspect 115: 210-214.
http://dx.doi.org/10.1289/ehp.9414
This document is a draft for review purposes only and does not constitute Agency policy.
R-18 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Falk. TE: Tuto. TE: Stridh. G: Bvlin. G. (1994). Dose-response study of formaldehyde on nasal mucosa
swelling. A study on residents with nasal distress at home. Am J Rhinol Allergy 8: 143-146.
http://dx.doi.Org/l 0.2500/105065894781874412
Fang. F: Ouinlan. P: Ye. W: Barber. MK: Umbach. DM: Sandler. DP: Kamel. F. (2009). Workplace
exposures and the risk of amyotrophic lateral sclerosis. Environ Health Perspect 117: 1387-
1392. http://dx.doi.Org/10.1289/ehp.0900580
Fang. 1: Li. DH: Yu. XO: Lv. MO: Bai. LZ: Du. LZ: Zhou. DX. (2015). Formaldehyde exposure inhibits
the expression of mammalian target of rapamycin in rattestis. Toxicol Ind Health 32: 1882-
1890. http://dx.doi.org/10.1177/0748233715592992
Farrow. A: Farrow. SC: Little. R: Golding. 1. (1996). The repeatability of self-reported exposure after
miscarriage. ALSPAC Study Team. Avon Longitudinal Study of Pregnancy and Childhood. Int
J Epidemiol 25: 797-806.
Federal Panel of Formaldehyde. (1982). Report of the federal panel on formaldehyde. Environ
Health Perspect 43: 139-168. http://dx.doi.org/10.1289/ehp.43-1568898
Fenech. M. (1993). The cytokinesis-block micronucleus technique: a detailed description of the
method and its application to genotoxicity studies in human populations [Review], Mutat
Res 285: 35-44.
Fenech. M. (2000). The in vitro micronucleus technique [Review], Mutat Res 455: 81-95.
Fenech. M. (2007). Cytokinesis-block micronucleus cytome assay. NatProtoc 2: 1084-1104.
http ://dx. doi. or g/10.10 3 8/nprot 2007.77
Fenech. M: Chang. WP: Kirsch-Volders. M: Holland. N: Bonassi. S: Zeiger. E: project. HM. (2003).
HUMN project: detailed description of the scoring criteria for the cytokinesis-block
micronucleus assay using isolated human lymphocyte cultures [Comment], Mutat Res 534:
65-75.
Fenech. M: Holland. N: Zeiger. E: Chang. WP: Burgaz. S: Thomas. P: Bolognesi. C: Knasmueller. S:
Kirsch-Volders. M: Bonassi. S. (2011). The HUMN and HUMNxL international collaboration
projects on human micronucleus assays in lymphocytes and buccal cells-past, present and
future [Review], Mutagenesis 26: 239-245. http ://dx. doi. or g/10.109 3 /mutage /geqO 51
Fenech. M: Morlev. AA. (1985). Measurement of micronuclei in lymphocytes. Mutat Res 147: 29-36.
Fennell. TR. (1994a). Development of methods for measuring biological markers of formaldehyde
exposure. Res Rep Health Eff Instl-20; discussion 21-26.
Fennell. TR. (1994b). Development of methods for measuring biological markers of formaldehyde
exposure (pp. 1-20; discussion 21-26). Cambridge, MA: Health Effects Institute.
Feron. VI: Bruvntjes. TP: Woutersen. RA: Immel. HR: Appelman. LM. (1988). Nasal tumours in rats
after short-term exposure to a cytotoxic concentration of formaldehyde. Cancer Lett 39:
101 -111. http://dx.doi.Org/10.1016/0304-3835r88190045-6
Feron. VI: Immel. HR: Wilmer. TWG. M: Woutersen. RA: Zwart. A. (1987). Nasal Tumours in Rats
after Severe Injury to the Nasal Mucosa and Exposure to Formaldehyde Vapour: Preliminary
Results (pp. 8-12). (NIOSH/00176121). Feron, VJ; Immel, HR; Wilmer, JWGM; Woutersen,
RA; Zwart, A.
Feron. VI: Til. HP: de Vrijer. F: Woutersen. RA: Cassee. FR: van Bladeren. PI. (1991). Aldehydes:
Occurrence, carcinogenic potential, mechanism of action and risk assessment [Review],
Mutat Res 259: 363-385. http: //dx.doi.org/10.1016/0165-1218C91190128-9
This document is a draft for review purposes only and does not constitute Agency policy.
R-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Ferris. BG. (1978). Epidemiology standardization project (American Thoracic Society). Am Rev
Respir Dis 118: 1-120.
Fiddler. W: Miller. AT: Pensabene. TW: Doerr. RC. (1984). Investigation on the mutagenicity of N-
nitrosothiazolidine using the Ames Salmonella test. In IK O'Neill; RC von Borstel; CT Miller; J
Long; H Bartsch (Eds.), IARC Scientific Publication No 57 (pp. 95-100). Lyon, France:
International Agency for Research on Cancer. http: //publications.iarc.fr/Book-And-Report-
Series/Iarc-Scientific-Publications/N-Nitroso-Compounds-Occurrence-Biological-Effects-
And-Relevance-To-Human-Cancer-1984
Fishbein. L. (1992). Exposure from occupational versus other sources [Review], Scand J Work
Environ Health 18: 5-16.
Flamant-Hulin. M: Caillaud. D: Sacco. P: Penard-Morand. C: Annesi-Maesano. I. (2010). Air pollution
and increased levels of fractional exhaled nitric oxide in children with no history of airway
damage. J Toxicol Environ Health A 73: 272-283.
http://dx.doi.Org/l 0.1080/15287390903249206
Fleig. I: Petri. N: Stacker. WG: Thiess. AM. (1982). Cytogenetic analyses of blood lymphocytes of
workers exposed to formaldehyde in formaldehyde manufacturing and processing. J Occup
Med 24: 1009-1012.
Fleisher. TM. (1987). Medical students' exposure to formaldehyde in gross anatomy laboratories. N
Y State J Med 87: 385-388.
Fontignie-Houbrechts. N. (1981). Genetic effects of formaldehyde in the mouse. Mutat Res 88: 109-
114. http://dx.doi.org/10.1016/0165-1218C81190095-1
Fontignie-Houbrechts. N: Moutschen-Dahmen. M: Degraeve. N: Gloor. H. (1982). Genetic effects in
the mouse of formaldehyde in combination with adenosine and hydrogen peroxide. Mutat
Res 104: 371-376. http://dx.doi.org/l 0.1016/0165-7992C82190172-5
Fornace. AT. Tr. (1982). Detection of DNA single-strand breaks produced during the repair of
damage by DNA-protein cross-linking agents. Cancer Res 42: 145-149.
Fornace. AT: Lechner. IF: Grafstrom. RC: Harris. CC. (1982). DNA repair in human bronchial
epithelial cells. Carcinogenesis 3: 1373-1377. http: / /dx. doi. or g/10.10 9 3 /car cin / 3.12.13 7 3
Fox. CH: Tohnson. FB: Whiting. 1: Roller. PP. (1985). Formaldehyde fixation [Review], J Histochem
Cytochem 33: 845-853.
Franklin. P: Dingle. P: Stick. S. (2000). Raised exhaled nitric oxide in healthy children is associated
with domestic formaldehyde levels. Am J Respir Crit Care Med 161: 1575-1759.
http: / /dx. doi. or g/10.1164/ai rccm. 161.5.9905061
Franklin. P: Tan. M: Hemv. N: Hall. GL. (2019). Maternal Exposure to Indoor Air Pollution and Birth
Outcomes. Int J Environ Res Public Health 16. http: //dx.doi.org/10.3390/iierphl6081364
Fransman. W: Mclean. D: Douwes. I: Demers. PA: Leung. V: Pearce. N. (2003). Respiratory symptoms
and occupational exposures in New Zealand plywood mill workers. Ann Occup Hyg 47: 287-
295. http://dx.doi.org/10.1093/annhyg/meg046
Frazelle. TH: Abernethv. DT: Boreiko. CT. (1983). Weak promotion of C3H/10T1 /2 cell
transformation by repeated treatments with formaldehyde. Cancer Res 43: 3236-3239.
Frvzek. TP: Chadda. BK: Cohen. SS: Marano. D: White. K: Steinwandel. M: McLaughlin. IK. (2005).
Retrospective cohort mortality study of workers engaged in motion picture film processing.
This document is a draft for review purposes only and does not constitute Agency policy.
R-20 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
J Occup Environ Med 47: 278-286.
http://dx.doi.Org/10.1097/01.iom.0000155712.22617.42
Fsadni. P: Bezzina. F: Fsadni. C: Montefort. S. (2018). Impact of School Air Quality on Children's
Respiratory Health. Indian J Occup Environ Med 22: 156-162.
http://dx.doi.org/10.4103/iioem.ITOEM 95 18
Fuiii. K: Tsuii. K: Matsuura. H: Okazaki. F: Takahashi. S: Arata. 1: Iwatsuki. K. (2005). Effect of
formaldehyde gas exposure in a murine allergic contact hypersensitivity model.
Immunopharmacol Immunotoxicol 27: 163-175. http://dx.doi.org/10.1081/IPH-51768
Fuiimaki. H: Kurokawa. Y: Kakevama. M: Kunugita. N: Fueta. Y: Fukuda. T: Hori. H: Arashidani. K.
(2004a). Inhalation of low-level formaldehyde enhances nerve growth factor production in
the hippocampus of mice. Neuroimmunomodulation 11: 373-375.
http://dx.doi.Org/10.1159/000080147
Fuiimaki. H: Kurokawa. Y: Kunugita. N: Kikuchi. M: Sato. F: Arashidani. K. (2004b). Differential
immunogenic and neurogenic inflammatory responses in an allergic mouse model exposed
to low levels of formaldehyde [Erratum], Toxicology 197: 1-13.
http://dx.doi.Org/10.1016/i.tox.2005.01.001
Fuiimaki. H: Shiraishi. F: Katavama. N. (1992). Enhancement of histamine release from rat
peritoneal mast cells exposed to formaldehyde. Inhal Toxicol 4: 125-136.
http://dx.doi.org/10.3109/08958379209145309
Galloway. SM: Bloom. AD: Resnick. M: Margolin. BH: Nakamura. F: Archer. P: Zeiger. E. (1985).
Development of a standard protocol for in vitro cytogenetic testing with Chinese hamster
ovary cells: Comparison of results for 22 compounds in two laboratories. Environ Mutagen
7: 1 -51. http: //dx.doi.org/10.1002 /em.2860070102
Gammage. RB: Hawthorne. AR. (1985). Current status of measurement techniques and
concentrations of formaldehyde in residences. In V Turoski (Ed.), Formaldehyde: Analytical
Chemistry and Toxicology (pp. 117-130). Washington, DC: ACS Publications.
http://dx.doi.Org/10.1021/ba-1985-0210.ch009
Garcia-Calderon. CB: Beiarano-Garcia. TA: Tinoco-Gago. I: Castro. MI: Moreno-Gordillo. P: Piruat. II:
Caballero-Velazquez. T: Perez-Simon. TA: Rosado. IV. (2018). Genotoxicity of tetrahydrofolic
acid to hematopoietic stem and progenitor cells. Cell Death Differ 25: 1967-1979.
http://dx.doi.Org/10.1038/s41418-018-0089-4
Garcia. CL: Mechilli. M: Proietti De Santis. L: Schinoppi. A: Kobos. K: Palitti. F. (2009). Relationship
between DNA lesions, DNA repair and chromosomal damage induced by acetaldehyde.
Mutat Res 662: 3-9. http://dx.doi.org/10.1016/j.mrfmmm.2008.11.008
Gardner. RT: Burgess. BA: Kennedy. GL. Tr. (1985). Sensory irritation potential of selected nasal
tumorigens in the rat Food Chem Toxicol 23: 87-92. http://dx.doi.org/10.1016/Q278-
6915C85190225-X
Garrett. MH: Hooper. MA: Hooper. BM: Ravment. PR: Abramson. Ml. (1999a). Increased risk of
allergy in children due to formaldehyde exposure in homes. Allergy 54: 330-337.
http: //dx.doi.org/10.1034/i.l 398-9995.1999.00763.x
Garrett. MH: Hooper. MA: Hooper. BM: Ravment. PR: Abramson. MI. (1999b). Increased risk of
allergy in children due to formaldehyde exposure in homes. Errata [Erratum], Allergy 54:
1327.
This document is a draft for review purposes only and does not constitute Agency policy.
R-21 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Ge. T: Yang. H: Lu. X: Wang. S: Zhao. Y: Huang. T: Xi. Z: Zhang. L: Li. R. (2020a). Combined exposure to
formaldehyde and PM2.5: Hematopoietic toxicity and molecular mechanism in mice.
Environ Int 144: 106050. http: / /dx. doi. o r g/10.1016/i.envint2020.106050
Ge. P: Zhang. X: Yang. YO: Lv. MO: Zhou. DX. (2020b). Long-term exposure to formaldehyde induced
down-regulation of SPOll in rats. Inhal Toxicoll-10.
http://dx.doi.Org/10.1080/08958378.2020.1859652
Ge. S: Yan. B: Huang. 1: Chen. Y: Chen. M: Yang. X: Wu. Y: Shen. D: Ma. P. (2019). Diisodecyl phthalate
aggravates the formaldehyde-exposure-induced learning and memory impairment in mice.
Food Chem Toxicol 126: 152-161. http://dx.doi.Org/10.1016/i.fct2019.02.024
Gee. IL: Watson. AFR: Tavernier. G: Stewart. LI: Fletcher. G: Niven. RM. (2005). Indoor air quality,
environmental tobacco smoke and asthma: A case control study of asthma in a community
population. Indoor Built Environ 14: 215-219.
http://dx.doi.org/10.1177/1420326X05054288
Gentry. PR: Rodricks. TV: Turnbull. D: Bachand. A: Van Landingham. C: Shipp. AM: Albertini. RT:
Irons. R. (2013). Formaldehyde exposure and leukemia: Critical review and reevaluation of
the results from a study that is the focus for evidence of biological plausibility [Review], Crit
Rev Toxicol 43: 661-670. http://dx.doi.org/10.3109/10408444.2013.818618
Georgieva. AV: Kimbell. IS: Schlosser. PM. (2003). A distributed-parameter model for formaldehyde
uptake and disposition in the rat nasal lining. Inhal Toxicol 15: 1435-1463.
http://dx.doi.Org/10.1080/08958370390249085
Gerberich. HR: Seaman. GC. (2013). Formaldehyde. In JI Kroshwitz; M Howe-Grant (Eds.), Kirk-
Othmer encyclopedia of chemical technology (4th ed., pp. 929-951). New York, NY: John
Wiley & Sons.
Gerde. P: Cheng. YS: Medinskv. MA. (1991). In vivo deposition of ultrafine aerosols in the nasal
airway of the rat Toxicol Sci 16: 330-336. http: / /dx. doi. o r g/10.1016 /0 2 72 -
0590(91190117-m
Gerin. M: Siemiatycki. I: Nadon. L: Dewar. R: Krewski. D. (1989). Cancer risks due to occupational
exposure to formaldehyde: Results of a multi-site case-control study in Montreal. Int J
Cancer 44: 53-58. http://dx.doi.org/10.1002/iic.2910440110
Ghelli. F: Buglisi. M: Bellisario. V: Santovito. A: Bono. R. (2020). Formaldehyde in hospitals can still
represent a risk factor. Oxidative stress and GSTT1 polymorphism [Abstract], Eur J Public
Health 30. http://dx.doi.org/10.1093/eurpub/ckaal66.340
Gieroba. ZT: Yu. YH: Blessing. WW. (1994). Vasoconstriction induced by inhalation of irritant vapour
is associated with appearance of Fos protein in CI catecholamine neurons in rabbit medulla
oblongata. Brain Res 636: 157-161. http://dx.doi.Org/10.l016/0006-8993(94)90192-9
Gilbert. NL: Gauvin. D: Guav. M: Heroux. ME: Dupuis. G: Legris. M: Chan. CC: Dietz. RN: Levesque. B.
(2006). Housing characteristics and indoor concentrations of nitrogen dioxide and
formaldehyde in Quebec City, Canada. Environ Res 102: 1-8.
http: / /dx. doi. or g/10.1016/i. envres. 2006.02.007
Gilbert. NL: Guav. M: David Miller. I: Tudek. S: Chan. CC: Dales. RE. (2005). Levels and determinants
of formaldehyde, acetaldehyde, and acrolein in residential indoor air in Prince Edward
Island, Canada. Environ Res 99: 11-17. http://dx.doi.Org/10.1016/i.envres.2004.09.009
This document is a draft for review purposes only and does not constitute Agency policy.
R-22 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Ginsberg. GL: Foos. BP: Firestone. MP. (2005). Review and analysis of inhalation dosimetry methods
for application to children's risk assessment [Review], J Toxicol Environ Health A 68: 573-
615. http://dx.doi.Org/l 0.1080/15287390590921793
Gocke. E: King. MT: Eckhardt. K: Wild. D. (1981). [Mutagenicity of cosmetics ingredients licensed by
the European communities], MutatRes 90: 91-109. http: / /dx. doi. o r g/10.1016 / 016 5 -
1218r81190072-0
Gofmekler. VA. (1968). Effect on embryonic development of benzene and formaldehyde in
inhalation experiments. Hyg Sanit 33: 327-332.
Gofmekler. VA: Bonashevskava. TI. (1969). Experimental studies of teratogenic properties of
formaldehyde, based on pathological investigations. Hyg Sanit 34: 266-268.
Golalipour. Ml: Azarhoush. R: Ghafari. S: Gharravi. AM: Fazeli. SA: Davarian. A. (2007).
Formaldehyde exposure induces histopathological and morphometric changes in the rat
testis. Folia Morphol (Warsz) 66: 167-171.
Golalipour. MT: Kord. H: Ghafari. S: Gharravi. AM: Davarian. A: Fazeli. SA: Azarhoush. R. (2008).
Morphometric alterations of the rat spleen following formaldehyde exposure. Folia Morphol
(Warsz) 67: 19-23.
Goldmacher. VS: Thillv. WG. (1983). Formaldehyde is mutagenic for cultured human cells. Mutat
Res 116: 417-422. http://dx.doi.org/l0.1016/0165-1218C83190080-0
Gomaa. M: Elmesallamv. GE: Sameer. MM. (2012). Evaluation of genotoxic effects of formaldehyde
in adult albino rats and its implication in case of human exposure. Life Science Journal 9:
3085-3093.
Gonzalez-Rivera. TC: Sherman. MW: Wang. PS: Chuvalo-Abraham. TCL: Hildebrandt Ruiz. L:
Contreras. LM. (2020). RNA oxidation in chromatin modification and DNA-damage response
following exposure to formaldehyde. Sci Rep 10: 16545.
http: //dx.doi.org/10.1038/s41598-020-73376-7
Gordon. CI: Spencer. PI: Hotchkiss. 1: Miller. DB: Hinderliter. PM: Pauluhn. 1. (2008).
Thermoregulation and its influence on toxicity assessment. Toxicology 244: 87-97.
http://dx.doi.Org/10.1016/i.tox.2007.10.030
Gorski. P: Krakowiak. A. (1991). Formaldehyde-induced bronchial asthma-does it really exist? Pol
J Occup Med Environ Health 4: 317-320.
Gostner. TM: Zeisler. 1: Alam. MT: Gruber. P: Fuchs. D: Becker. K: Neubert. K: Kleinhappl. M: Martini.
S: Uberall. F. (2016). Cellular reactions to long-term volatile organic compound (VOC)
exposures. Sci Rep 6: 37842. http: / /dx.doi. or g/10.10 38/srep37842
Gottschling. LM: Beaulieu. HI: Melvin. WW. (1984). Monitoring of formic acid in urine of humans
exposed to low levels of formaldehyde. Am Ind Hyg Assoc J 45: 19-23.
http://dx.doi.org/10.1080/15298668491399299
Graf. RA: Kater. SB: Gordon. H. (1999). Prolonged Cytosolic Calcium Elevations in Growth Cones
Contacting Muscle. Dev Neurosci 21: 409-416. http: //dx.doi.org/10.1159/000017408
Grafstrom. RC. (1990). In vitro studies of aldehyde effects related to human respiratory
carcinogenesis. MutatRes 238: 175-184.
Grafstrom. RC: Fornace. A. Tr: Harris. CC. (1984). Repair of DNA damage caused by formaldehyde in
human cells. Cancer Res 44: 4323-4327.
This document is a draft for review purposes only and does not constitute Agency policy.
R-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Grafstrom. RC: Fornace. AT. Tr: Autrup. H: Lechner. TF: Harris. CC. (1983). Formaldehyde damage to
DNAand inhibition of DNA repair in human bronchial cells. Science 220: 216-218.
http://dx.doi.org/10.1126/science.6828890
Grafstrom. RC: Hsu. IC: Harris. CC. (1993). Mutagenicity of formaldehyde in Chinese hamster lung
fibroblasts: Synergy with ionizing radiation and N-nitroso-N-methylurea. Chem Biol Interact
86: 41-49. http://dx.doi.org/10.1016/0009-2797r93190110-K
Grafstrom. RC: Sundqvist. K: Dvpbukt. TM: Hvbbinette. SA: Harris. CC. (1986). CYTOTOXIC AND
GENOTOXIC EFFECTS OF ALDEHYDES IN CULTURED HUMAN BRONCHIAL CELLS (pp.
BASEL). (BIOSIS/87/05969). Grafstrom, RC; Sundqvist, K; Dypbukt, JM; Hybbinette, SA;
Harris, CC.
Grafstrom. RC: Willev. TC: Sundqvist. K: Harris. CC. (1986). Pathobiological effects of tobacco smoke-
related aldehydes in cultured human bronchial epithelial cells. In D Hoffman; CC Harris
(Eds.), Mechanisms in Tobacco Carcinogenesis, Banbury Report 23 (pp. 273-285). Cold
Springs Harbor, NY: CSH Press.
Grammer. LC: Harris. KE: Shaughnessv. MA: Sparks. P: Avars. GH: Altman. LC: Patterson. R. (1990).
Clinical and immunologic evaluation of 37 workers exposed to gaseous formaldehyde. J
Allergy Clin Immunol 86: 177-181. http://dx.doi.org/10.1016/S0091-6749r05180063-6
Graves. RT: Callander. RD: Green. T. (1994). The role of formaldehyde and S-
chloromethylglutathione in the bacterial mutagenicity of methylene chloride. Mutat Res
320: 235-243. http://dx.doi.org/10.1016/0165-1218r94190050-7
Graves. RT: Trueman. P: Tones. S: Green. T. (1996). DNA sequence analysis of methylene chloride-
induced HPRT mutations in Chinese hamster ovary cells: Comparison with the mutation
spectrum obtained for 1,2-dibromoethane and formaldehyde. Mutagenesis 11: 229-233.
http: / /dx. doi. or g/10.109 3 /mutage /11.3.229
Green. DT: Bascom. R: Healev. EM: Hebel. TR: Sauder. LR: Kulle. TT. (1989). Acute pulmonary
response in healthy, nonsmoking adults to inhalation of formaldehyde and carbon. J Toxicol
Environ Health 28: 261-275. http://dx.doi.org/10.1080/15287398909531347
Green. DT: Sauder. LR: Kulle. TT: Bascom. R. (1987). Acute response to 3.0 ppm formaldehyde in
exercising healthy nonsmokers and asthmatics. Am Rev Respir Dis 135: 1261-1266.
http: / /dx. doi. or g/10.1164 /arrd. 1987.135.6.1261
Gross. EA: Swenberg. TA: Fields. S: Popp. TA. (1982). Comparative morphometry of the nasal cavity
in rats and mice. J Anat 135: 83-88.
Groten. TP: Schoen. ED: van Bladeren. PT: Kuper. CF: van Zorge. TA: Feron. VT. (1997). Subacute
toxicity of a mixture of nine chemicals in rats: detecting interactive effects with a
fractionated two-level factorial design. Fundam Appl Toxicol 36: 15-29.
http: / /dx. doi. or g/10.10 0 6 /faat. 1996.2281
Gu. Y: Fuiimiva. Y: Kunugita. N. (2008). Long-term exposure to gaseous formaldehyde promotes
allergen-specific IgE-mediated immune responses in a murine model. Hum Exp Toxicol 27:
37-43. http://dx.doi.Org/l0.1177/0960327108088973
Giilec. M: Songur. A: Sahin. S: Ozen. OA: Sarsilmaz. M: Akvol. 0. (2006). Antioxidant enzyme
activities and lipid peroxidation products in heart tissue of subacute and subchronic
formaldehyde-exposed rats: a preliminary study. Toxicol Ind Health 22: 117-124.
http://dx.doi.org/10.1191/0748233706tli248oa
This document is a draft for review purposes only and does not constitute Agency policy.
R-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Guseva. VA. (1973). [Study of the gonadotropic effect in male rats under the effect of formaldehyde
administered simultaneously with air and water]. Gig Sanit 37: 102-103.
Gustafson. P: Barregard. L: Lindahl. R: Sallsten. G. (2005). Formaldehyde levels in Sweden: Personal
exposure, indoor, and outdoor concentrations. J Expo Anal Environ Epidemiol 15: 252-260.
http: / /dx. doi. or g/10.10 3 8 /si .i ea. 75 0 0 3 9 9
Gustavsson. P: Takobsson. R: Tohansson. H: Lewin. F: Norell. S: Rutkvist. LE. (1998). Occupational
exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus:
A case-control study in Sweden. Occup Environ Med 55: 393-400.
Guvton. AC. (1991). Textbook of medical physiology (8th ed.). Philadelphia, PA: W.B. Saunders Co.
Hagiwara. M: Watanabe. E: Barrett. TC: Tsutsui. T. (2006). Assessment of genotoxicity of 14 chemical
agents used in dental practice: Ability to induce chromosome aberrations in Syrian hamster
embryo cells. MutatRes 603: 111-120. http://dx.doi.Org/10.1016/i.mrgentox.2005.08.011
Hall. A: Harrington. TM: Aw. TC. (1991). Mortality study of British pathologists. Am J Ind Med 20: 83-
89. http://dx.doi.Org/10.1002/aiim.4700200108
Hamaguchi. F: Tsutui. T. (2000). Assessment of genotoxicity of dental antiseptics: Ability of phenol,
guaiacol, p-phenolsulfonic acid, sodium hypochlorite, p-chlorophenol, m-cresol or
formaldehyde to induce unscheduled DNA synthesis in cultured Syrian hamster embryo
cells. Jpn J Pharmacol 83: 273-276. http://dx.doi.org/10.1254/iip.83.273
Han. RT: Back. SK: Lee. H: Lee. 1: Kim. HY: Kim. HI: Na. HS. (2016). Formaldehyde-Induced
Aggravation of Pruritus and Dermatitis Is Associated with the Elevated Expression of Thl
Cytokines in a Rat Model of Atopic Dermatitis. PLoS ONE 11: e0168466.
http://dx.doi.org/10.1371/iournal.pone.0168466
Han. SP: Zhou. DX: Lin. P: Oin. Z: An. L: Zheng. LR: Lei. L. (2013). Formaldehyde exposure induces
autophagy in testicular tissues of adult male rats. Environ Toxicol 30: 323-331.
http: / /dx. do i. or g /10.10 0 2 /tox. 21910
Hankinson. TL: Odencrantz. TR: Fedan. KB. (1999). Spirometric reference values from a sample of the
general US population. Am J Respir Crit Care Med 159: 179-187.
http: / /dx. doi. or g/10.1164/ai rccm. 159.1.9712108
Hanrahan. LP: Dally. KA: Anderson. HA: Kanarek. MS: Rankin. 1. (1984). Formaldehyde vapor in
mobile homes: A cross sectional survey of concentrations and irritant effects. Am J Public
Health 74: 1026-1027. http://dx.doi.Org/10.2105/aiph.74.9.1026
Hansen. 1: Olsen. TH. (1995). Formaldehyde and cancer morbidity among male employees in
Denmark. Cancer Causes Control 6: 354-360. http://dx.doi.org/10.1007/BF00051411
Hansen. I: Olsen. TH: Larsen. AI. (1994). Cancer morbidity among employees in a Danish
pharmaceutical plant. Int J Epidemiol 23: 891-898. http: / /dx.doi. or g/10.10 9 3 /ii e /2 3.5.891
Hare. DA: Margosian. RL: Groah. WT: 3rd. AS: Schweer. LG: Koontz. MP. (1996). Evaluating the
contribution of UF-bonded building materials to indoor formaldehyde levels in a newly
constructed house.
Harkema. TR: Carey. SA: Wagner. TG. (2006). The nose revisited: A brief overview of the comparative
structure, function, and toxicologic pathology of the nasal epithelium [Review], Toxicol
Pathol 34: 252-269. http://dx.doi.org/10.1080/01926230600713475
Harrington. TM: Oakes. D. (1984). Mortality study of British pathologists 1974-80. Occup Environ
Med 41: 188-191. http://dx.doi.Org/10.1136/oem.41.2.188
This document is a draft for review purposes only and does not constitute Agency policy.
R-25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Harving. H: Korsgaard. T: Dahl. R: Pedersen. OF: Molhave. L. (1986). Low concentrations of
formaldehyde in bronchial asthma: a study of exposure under controlled conditions. Br Med
J 293: 310.
Harving. H: Korsgaard. 1: Pedersen. OF: Malhave. L: Dahl. R. (1990). Pulmonary function and
bronchial reactivity in asthmatics during low-level formaldehyde exposure. Lung 168: 15-
21. http://dx.doi.org/10.1007/BF02719669
Hauptmann. M: Stewart. PA: Lubin. TH: Beane Freeman. LE: Hornung. RW: Herrick. RF: Hoover. RN:
Fraumeni. IF. Tr: Blair. A: Hayes. RB. (2009). Mortality from lymphohematopoietic
malignancies and brain cancer among embalmers exposed to formaldehyde. J Natl Cancer
Inst 101: 1696-1708. http://dx.doi.org/10.1093/inci/dip416
Haworth. S: Lawlor. T: Mortelmans. K: Speck. W: Zeiger. E. (1983). Salmonella mutagenicity test
results for 250 chemicals. Environ Mutagen 5: 3-142.
http://dx.doi.Org/10.1002/em.2860050703
Havashi. H: Kunugita. N: Arashidani. K: Fuiimaki. H: Ichikawa. M. (2004). Long-term exposure to
low levels of formaldehyde increases the number of tyrosine hydroxylase-immunopositive
periglomerular cells in mouse main olfactory bulb. Brain Res 1007: 192-197.
http ://dx. doi. or g/10.1016/i .brainres. 2003.12.052
Hayes. RB: Blair. A: Stewart. PA: Herrick. RF: Mahar. H. (1990). Mortality of U.S. embalmers and
funeral directors. Am J Ind Med 18: 641-652. http://dx.doi.org/10.1002/ajim.4700180603
Hayes. RB: Klein. S: Suruda. A: Schulte. P: Boeniger. M: Stewart. P: Livingston. GK: Oesch. F. (1997).
06-alkylguanine DNA alkyltransferase activity in student embalmers. Am J Ind Med 31: 361-
365. http://dx.doi.Org/l 0.1002/CSTCTll 097-0274d99703131:3<361 ::ATD-
AJIM13>3.0.CO;2-Z
He. HI: Liu. HL: Wu. I: Lu. ZS: Yan. Y: Yang. X: Li. CM. (2005). A study on the acute irritation
responses and molecular mechanism of gaseous formaldehyde. In X Yang; B Zhao; R Zhao
(Eds.), Indoor Air 2005: Proceedings of the 10th International Conference on Indoor Air
Quality and Climate, vol 5 (pp. 3691-3695). Beijing, China: Tsinghua University Press.
https://www.isiaq.org/docs/PDFs/3691.pdf
He. TL: Tin. LF: Tin. HY. (1998). Detection of cytogenetic effects in peripheral lymphocytes of students
exposed to formaldehyde with cytokinesis-blocked micronucleus assay. Biomed Environ Sci
11: 87-92.
Health Canada. (2001). Priority substances list assessment report Formaldehyde. Hull, Quebec,
Canada: Environment Canada and Health Canada.
Heck. H: Casanova-Schmitz. M: Dodd. PB: Schachter. EN: Witek. TT: Tosun. T. (1985). Formaldehyde
(CH20) concentrations in the blood of humans and Fischer-344 rats exposed to CH20 under
controlled conditions. AIHA J 46: 1-3. http://dx.doi.org/10.1080/15298668591394275
Heck. H: Casanova. M. (1987). Isotope effects and their implications for the covalent binding of
inhaled [3H]- and [14C]formaldehyde in the rat nasal mucosa. Toxicol Appl Pharmacol 89:
122-134. http: //dx.doi.org/10.1016/0041-008XC87190182-7
Heck. H: Chin. TY: Schmitz. MC. (1983). Distribution of [14C] formaldehyde in rats after inhalation
exposure. In JE Gibson (Ed.), Formaldehyde toxicity (pp. 26-37). Washington, DC:
Hemisphere Publishing.
Heck. H: Keller. DA. (1988). Toxicology of formaldehyde. ISI Atlas of Science: Pharmacology 2: 5-9.
This document is a draft for review purposes only and does not constitute Agency policy.
R-26 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Heck. H: White. EL: Casanova-Schmitz. M. (1982). Determination of formaldehyde in biological
tissues by gas chromatography/mass spectrometry. Biomed Mass Spectrom 9: 347-353.
http://dx.doi.Org/10.1002/bms.1200090808
Hedberg. IT: Hoog. 10: Nilsson. TA: Xi. Z: Elfwing. A: Grafstrom. RC. (2000). Expression of alcohol
dehydrogenase 3 in tissue and cultured cells from human oral mucosa. Am J Pathol 157:
1745-1755. http://dx.doi.Org/10.1016/S0002-9440n0164811-0
Heineman. EF: Olsen. TH: Pottern. LM: Gomez. M: Raffn. E: Blair. A. (1992). Occupational risk factors
for multiple myeloma among Danish men. Cancer Causes Control 3: 555-568.
http://dx.doi.Org/10.1007/BF00052753
Hemminki. K: Kyvronen. P: Lindbohm. ML. (1985). Spontaneous abortions and malformations in the
offspring of nurses exposed to anaesthetic gases, cytostatic drugs, and other potential
hazards in hospitals, based on registered information of outcome. J Epidemiol Community
Health 39: 141-147. http://dx.doi.Org/10.1136/iech.39.2.141
Hemminki. K: Mutanen. P: Saloniemi. I: Niemi. ML: Vainio. H. (1982). Spontaneous abortions in
hospital staff engaged in sterilizing instruments with chemical agents. J Occup Environ Med
285: 1461-1463.
Herbert. FA: Hessel. PA: Melenka. LS: Yoshida. K: Nakaza. M. (1994). Respiratory consequences of
exposure to wood dust and formaldehyde of workers manufacturing oriented strand board.
Arch Environ Health 49: 465-470. http: //dx.doi.org/10.1080/00039896.1994.9955002
Hester. SD: Benavides. GB: Yoon. L: Morgan. KT: Zou. F: Barry. W: Wolf. DC. (2003). Formaldehyde-
induced gene expression in F344 rat nasal respiratory epithelium. Toxicology 187: 13-24.
http://dx.doi. org/10.1016/S0300-483Xr03100008-8
Hikiba. H: Watanabe. E: Barrett. TC: Tsutsui. T. (2005). Ability of fourteen chemical agents used in
dental practice to induce chromosome aberrations in Syrian hamster embryo cells. J
Pharmacol Sci 97: 146-152. http://dx.doi.org/10.1254/iphs.fpi04044x
Hildesheim. A: Dosemeci. M: Chan. CC: Chen. CI: Cheng. YT: Hsu. MM: Chen. IH: Mittl. BF: Sun. B:
Levine. PH: Chen. TY: Brinton. LA: Yang. CS. (2001). Occupational exposure to wood,
formaldehyde, and solvents and risk of nasopharyngeal carcinoma. Cancer Epidemiol
Biomarkers Prev 10: 1145-1153.
Holmstrom. M: Rosen. G: Wilhelmsson. B. (1991). Symptoms, airway physiology and histology of
workers exposed to medium-density fiber board. Scand J Work Environ Health 17: 409-413.
http: //dx.doi.org/10.5 2 71 /si weh. 1685
Holmstrom. M: Rynnel-Dagoo. B: Wilhelmsson. B. (1989a). Antibody production in rats after long-
term exposure to formaldehyde. Toxicol Appl Pharmacol 100: 328-333.
http://dx.doi.org/10.1016/0041-008Xr89190318-9
Holmstrom. M: Wilhelmsson. B. (1988). Respiratory symptoms and pathophysiological effects of
occupational exposure to formaldehyde and wood dust Scand J Work Environ Health 14:
306-311.
Holmstrom. M: Wilhelmsson. B: Hellquist. H. (1989b). Histological changes in the nasal mucosa in
rats after long-term exposure to formaldehyde and wood dust. Acta Otolaryngol 108: 274-
283. http://dx.doi.org/10.3109/00016488909125528
Holmstrom. M: Wilhelmsson. B: Hellquist. H: Rosen. G. (1989c). Histological changes in the nasal
mucosa in persons occupationally exposed to formaldehyde alone and in combination with
This document is a draft for review purposes only and does not constitute Agency policy.
R-27 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
wood dust Acta Otolaryngol 107: 120-129.
http://dx.doi.Org/10.3109/00016488909127488
Holness. PL: Nethercott. 1R. (1989). Health status of funeral service workers exposed to
formaldehyde. Arch Environ Occup Health 44: 222-228.
http://dx.doi.Org/10.1080/00039896.1989.9935887
Horton. AW: Tve. R: Stemmer. KL. (1963a). Experimental Carcinogenesis of the Lung. Inhalation of
Gaseous Formaldehyde or an Aerosol of Coal Tar by CEH Mice. J Natl Cancer Inst 30: 31-40.
Horton. AW: Tve. R: Stemmer. KL. (1963b). Experimental carcinogenesis of the lung: Inhalation of
gaseous formaldehyde or an aerosol of coal tar by C3H mice. J Natl Cancer Inst 30: 31-40.
http: / /dx. doi.org/10.1093 /i nci /3 0.1.31
Horvath. EP. Tr: Anderson. H. Tr: Pierce. WE: Hanrahan. L: Wendlick. TP. (1988). Effects of
formaldehyde on the mucous membranes and lungs: A study of an industrial population.
JAMA 259: 701-707. http://dx.doi.Org/10.1001/iama.1988.03720050037020
Hosgood. HP. Ill: Zhang. L: Tang. X: Vermeulen. R: Hao. Z: Shen. M. in: Oiu. C: Ge. Y: Hua. M: Ti. Z: Li.
S: Xiong. 1. un: Reiss. B: Liu. S: Xin. KX: Azuma. M: Xie. Y: Freeman. LB: Ruan. X: Guo. W:
Galvan. N. oe: Blair. A: Li. L: Huang. H: Smith. MT: Rothman. N: Lan. 0. (2013). Occupational
exposure to formaldehyde and alterations in lymphocyte subsets. Am J Ind Med 56: 252-
257. http://dx.doi.org/10.1002/aiim.22088
Hsu. NY: Lee. CC: Wang. TY: Li. YC: Chang. HW: Chen. CY: Bornehag. CG: Wu. PC: Sundell. I: Su. HI.
(2012). Predicted risk of childhood allergy, asthma and reported symptoms using measured
phthalate exposure in dust and urine. Indoor Air 22: 186-199.
http: //dx.doi.org/10.1111 /i.l 600-0668.2011,00753.x
Huang. C: Liu. W: Cai. I: Wang. X: Zou. Z: Sun. CI. (2017). Household formaldehyde exposure and its
associations with dwelling characteristics, lifestyle behaviours, and childhood health
outcomes in Shanghai, China. Build Environ 125: 143-152.
http://dx.doi.Org/10.1016/j.buildenv.2017.08.042
Huang. H: Hopkins. PB. (1993). PNA interstrand cross-linking by formaldehyde - nucleotide-
sequence preference and covalent structure of the predominant cross-link formed in
synthetic oligonucleotides. J Am Chem Soc 115: 9402-9408.
http: / /dx. doi. or g/10.10 21 /) aO 0 0 74a0 0 5
Huang. HF: Solomon. MS: Hopkins. PB. (1992). Formaldehyde preferentially interstrand cross-links
duplex PNA through deoxyadenosine residues at the sequence 5'-d(AT). J Am Chem Soc
114: 9240-9241. http://dx.doi.org/10.1021 /ia00049a097
Huang. I: Lu. Y: Zhang. B: Yang. S: Zhang. 0: Cui. H: Lu. X: Zhao. Y: Yang. X: Li. R. (2019). Antagonistic
effect of epigallocatechin-3-gallate on neurotoxicity induced by formaldehyde. Toxicology
412: 29-36. http://dx.doi.org/10.1016/i.tox.2018.10.022
Hulin. M: Caillaud. P: Annesi-Maesano. I. (2010). Indoor air pollution and childhood asthma:
variations between urban and rural areas. Indoor Air 20: 502-514.
http: //dx.doi.org/10.1111 /i.l 600-0668.2010.00673.x
Hwang. G: Yoon. C: Choi. 1. (2011). A Case-Control Study: Exposure Assessment of VOCs and
Formaldehyde for Asthma in Children. Aerosol Air Qual Res 11: 908-914.
http://dx.doi.Org/10.4209/aaqr.2011.05.0072
I ARC (International Agency for Research on Cancer). (1995). Wood dust and formaldehyde. Lyon,
France. http: //monographs.iarc.fr/ENG/Monographs/vol62/index.php
This document is a draft for review purposes only and does not constitute Agency policy.
R-28 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
IARC (International Agency for Research on Cancer). (2006a). Formaldehyde, 2-butoxyethanol and
l-tert-butoxypropan-2-ol [IARC Monograph], Lyon, France.
https://publications.iarc.fr/Book-And-Report-Series/larc-Monographs-On-The-
Identification-0f-Carcinogenic-Hazards-To-Humans/Formaldehyde-2-Butoxyethanol-And-
l-Em-Tert-Em-Butoxypropan-2-ol-2006
IARC (International Agency for Research on Cancer). (2006b). Preamble to the IARC Monographs
(amended January 2006). http://monographs.iarc.fr/ENG/Preamble/index.php
IARC (International Agency for Research on Cancer). (2012). Formaldehyde [IARC Monograph], In
A review of human carcinogens: Chemical agents and related occupations (pp. 401-435).
Lyon, France. http://monographs.iarc.fr/ENG/Monographs/vollOOF/index.php
Ibrahim. BS: Barioni. ED: Heluanv. C: Braga. TT: Drewes. CC: Costa. SG: Saraiva Camara. NO: Poliselli
Farskv. SH: Lino-Dos-Santos-Franco. A. (2016). Beneficial effects of vitamin C treatment on
pregnant rats exposed to formaldehyde: Reversal of immunosuppression in the offspring.
Toxicol Appl Pharmacol 300: 77-81. http://dx.doi.Org/10.1016/i.taap.2016.03.010
ICRP (International Commission on Radiological Protection). (1994). Human respiratory tract
model for radiological protection. Ann ICRP 24.
Tm. H: Oh. E: Mun. I: Khim. TY: Lee. E: Kang. HS: Kim. E: Kim. H: Won. NH: Kim. YH: Tung. WW: Sul. D.
(2006). Evaluation of toxicological monitoring markers using proteomic analysis in rats
exposed to formaldehyde. J Proteome Res 5: 1354-1366.
http ://dx. doi. or g/10.10 21 /pr0 5 0 43 7b
Inoue. K: Nishimukai. H: Yamasawa. K. (1979). Purification and partial characterization of aldehyde
dehydrogenase from human erythrocytes. Biochim Biophys Acta 569: 117-123.
http://dx.doi.Org/l 0.1016/0005-2744(79190046-9
Ionescu. I: Marinescu. D: Tapu. V: Eskenasv. A. (1978). Experimental chronic obstructive lung
disease: I. Bronchopulmonary changes induced in rabbits by prolonged exposure to
formaldehyde. Morphol Embryol (Bucur) 24: 233-242.
Isa. KNM: Hashim. Z: Talaludin. 1: Norback. D. an: Tabbar. MA: Hashim. TH. (2020a). The Impact of
Exposure to Indoor Pollutants on Allergy and Lung Inflammation among School Children in
Selangor, Malaysia: An Evaluation Using Factor Analysis. Aerosol Air Qual Res 20: 2371-
2383. http://dx.doi.Org/10.4209/aaqr.2020.03.0128
Isa. KNM: Hashim. Z: Talaludin. 1: Than. LTL: Hashim. TH. (2020b). The effects of indoor pollutants
exposure on allergy and lung inflammation: An activation state of neutrophils and
eosinophils in sputum. Int J Environ Res Public Health 17: 5413.
http://dx.doi.org/10.3390/ijerphl7155413
Ishidate. M. Tr: Sofuni. T: Yoshikawa. K. (1981). Chromosomal aberration tests in vitro as a primary
screening tool for environmental mutagens and/or carcinogens. In N Inui; T Kuroki; MA
Yamada; C Heidelberger (Eds.), Mutation, promotion and transformation in vitro (pp. 95-
108). Tokyo, Japan: Japan Scientific Societies Press.
Ito. K: Sakamoto. T: Havashi. Y: Morishita. M: Shibata. E: Sakai. K: Takeuchi. Y: Torii. S. (1996). Role
of tachykinin and bradykinin receptors and mast cells in gaseous formaldehyde-induced
airway microvascular leakage in rats. Eur J Pharmacol 307: 291-298.
http://dx.doi.Org/10.1016/0014-2999f96100285-3
Taeger. RT: Gearhart. TM. (1982). Respiratory and metabolic response of rats and mice to formalin
vapor. Toxicology 25: 299-309. http://dx.doi.org/l 0.1016/0300-483XC82190108-1
This document is a draft for review purposes only and does not constitute Agency policy.
R-29 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Takab. GT. (1992). Relationship between carbon black particulate-bound formaldehyde, pulmonary
antibacterial defenses, and alveolar macrophage phagocytosis. Inhal Toxicol 4: 325-342.
http://dx.doi.Org/l 0.3109 /08958379209145312
Takab. MG: Klupp. T: Besenvei. K: Biro. A: Major. 1: Tompa. A. (2010). Formaldehyde-induced
chromosomal aberrations and apoptosis in peripheral blood lymphocytes of personnel
working in pathology departments. MutatRes 698: 11-17.
http://dx.doi.Org/10.1016/i.mrgentox.2010.02.015
Takobsson. K: Mikoczv. Z: Skerfving. S. (1997). Deaths and tumours among workers grinding
stainless steel: a follow up. Occup Environ Med 54: 825-829.
http://dx.doi.org/10.1136/oem.54.ll.825
Tenkins. MA: Clarke. TR: Carlin. IB: Robertson. CF: Hopper. TL: Dalton. MF: Hoist. DP: Choi. K: Giles.
GG. (1996). Validation of questionnaire and bronchial hyperresponsiveness against
respiratory physician assessment in the diagnosis of asthma. Int J Epidemiol 25: 609-616.
http://dx.doi.Org/10.1093/iie/25.3.609
Tensen. DE: Belka. GK: Du Bois. GC. (1998). S-Nitrosoglutathione is a substrate for rat alcohol
dehydrogenase class III isoenzyme. Biochem J 331: 659-668.
http: //dx.doi.org/10.1042 /bi 3 310 65 9
Tensen. KA: Kirk. I: Kalmark. G: Westergaard. M. (1951). Chemically induced mutations in
Neurospora. Cold Spring Harb Symp Quant Biol 16: 245-261.
http://dx.doi.Org/10.1101/SOB.1951.016.01.020
Ti. Z: Li. X: Fromowitz. M: Mutter-Rottmaver. E: Tung. 1: Smith. MT: Zhang. L. (2014). Formaldehyde
induces micronuclei in mouse erythropoietic cells and suppresses the expansion of human
erythroid progenitor cells. Toxicol Lett 224: 233-239.
http://dx.doi.Org/10.1016/i.toxlet.2013.10.028
Tia. X: Tia. 0: Zhang. Z: Gao. W: Zhang. X: Niu. Y: Meng. T: Feng. B: Duan. H: Ye. M: Dai. Y: Tia. Z: Zheng.
Y, (2014). Effects of formaldehyde on lymphocyte subsets and cytokines in the peripheral
blood of exposed workers. PLoS ONE 9: el04069.
http://dx.doi.org/10.1371/iournal.pone.0104069
Tiang. 1: Zhou. CF: Gao. S: Tian. Y: Wang. C: Wang. L. i: Gu. HF: Tang. XO. (2015). BDNF-TrkB Pathway
Mediates Neuroprotection of Hydrogen Sulfide against Formaldehyde-Induced Toxicity to
PC12 Cells. PLoS ONE 10: e0119478. http://dx.doi.org/10.1371/iournal.pone.0119478
Tiang. S: Yu. L: Cheng. 1: Leng. S: Dai. Y: Zhang. Y: Niu. Y: Yan. H: Ou. W: Zhang. C: Zhang. K: Yang. R:
Zhou. L: Zheng. Y. (2010). Genomic damages in peripheral blood lymphocytes and
association with polymorphisms of three glutathione S-transferases in workers exposed to
formaldehyde. MutatRes 695: 9-15. http://dx.doi.Org/10.1016/i.mrgentox.2009.09.011
Tin. L: Lynch. I: Richardson. A: Lorkiewicz. P: Srivastava. S: Theis. W: Shirk. G: Hand. A: Bhatnagar. A:
Srivastava. S: Conklin. DT. (2021). Electronic Cigarette Solvents, Pulmonary Irritation and
Endothelial Dysfunction:Role of Acetaldehyde and Formaldehyde. Am J Physiol Heart Circ
Physiol. http://dx.doi.org/10.1152/aipheart.00878.202Q
Toffe. M: Vilard. L: Li. Z: Powman. R: Vessev. M. (1993). Long-term recall of time-to-pregnancy. Fertil
Steril 60: 99-104. http://dx.doi.org/10.1016/s0015-0282ri6156044-0
Toffe. M: Villard. L: Li. ZM: Plowman. R: Vessev. M. (1995). A time to pregnancy questionnaire
designed for long-term recall - Validity in Oxford, England. J Epidemiol Community Health
49: 314-319. http://dx.doi.org/10.1136/iech.49.3.314
This document is a draft for review purposes only and does not constitute Agency policy.
R-30 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Tohn. EM: Savitz. DA: Shy. CM. (1994). Spontaneous abortions among cosmetologists. Epidemiology
5: 147-155. http://dx.doi.Org/10.1097/00001648-199403000-00004
lohnsen. RC: Baillie. PL. (1988). Formaldehyde mutagenesis of the eTl balanced region in
Caenorhabditis elegans: dose-response curve and the analysis of mutational events. Mutat
Res 201: 137-147. http://dx.doi.org/10.1016/0027-5107r88190120-0
Tuarez. E: Chambwe. N: Tang. W: Mitchell. AD: Owen. N: Kumari. A: Monnat. RT: Mccullough. AK.
(2018). An RNAi screen in human cell lines reveals conserved DNA damage repair pathways
that mitigate formaldehyde sensitivity. DNA Repair 72: 1-9.
http://dx.doi.Org/10.1016/i.dnarep.2018.10.002
Tude. 1: Koziol-White. C: Scala. 1: Yoo. E: Tester. W: Maute. C: Dalton. P: Panettieri. R. (2016).
Formaldehyde Induces Rho-associated Kinase Activity to Evoke Airway
Hyperresponsiveness. Am J Respir Cell Mol Biol 55: 542-553.
http://dx.doi.org/10.1165/rcmb.2015-0254QC
Tung. R: Engelhart. G: Herbolt. B: Tackh. R: Miiller. W. (1992). Collaborative study of mutagenicity
with Salmonella typhimurium TA102. Mutat Res 278: 265-270.
http://dx.doi.org/10.1016/S0165-1218ri0180006-0
Tung. W: Kim. E: Lee. E: Yun. H: Tu. H: Teong. M: Hwang. K: Sul. D: Kang. H. (2007). Formaldehyde
exposure induces airway inflammation by increasing eosinophil infiltrations through the
regulation of reactive oxygen species production. Environ Toxicol Pharmacol 24: 174-182.
http ://dx. doi. or g/10.1016/i. etap. 2007.05.001
Kamata. E: Nakadate. M: Uchida. 0: Ogawa. Y: Kaneko. T: Kurokawa. Y. (1996). Effects of
formaldehyde vapor on the nasal cavity and lungs of F-344 rats. J Environ Pathol Toxicol
Oncol 15: 1-8.
Kamata. E: Nakadate. M: Uchida. 0: Ogawa. Y: Suzuki. S: Kaneko. T: Saito. M: Kurokawa. Y. (1997).
Results of a 28-month chronic inhalation toxicity study of formaldehyde in male Fisher-344
rats. J Toxicol Sci 22: 239-254.
Kane. LE: Alarie. Y. (1977). Sensory irritation to formaldehyde and acrolein during single and
repeated exposures in mice. Am Ind Hyg Assoc J 38: 509-522.
http: //dx.doi.org/10.1080/0002889778507665
Kane. LE: Barrow. CS: Alarie. Y. (1979). A short-term test to predict acceptable levels of exposure to
airborne sensory irritants. Am Ind Hyg Assoc J 40: 207-229.
http: //dx.doi.org/10.1080/15298667991429516
Kang. I: Duan. T: Song. I: Luo. C: Liu. H: Li. B: Yang. X: Yu. W: Chen. M. (2018). Exposure to a
combination of formaldehyde and DINP aggravated asthma-like pathology through
oxidative stress and NF-kB activation. Toxicology 404-405: 49-58.
http://dx.doi.Org/10.1016/i.tox.2018.05.006
Kang. S. ukYun: North. 1. in: Gavtan. I: Romero. W. oo: De La Rosa. VY: Loguinov. A: Smith. MT:
Zhang. L: Vulpe. CD. (2016). Functional Toxicogenomic Profiling Expands Insight into
Modulators of Formaldehyde Toxicity in Yeast Front Genet 7: 200.
http://dx.doi.Org/10.3389/fgene.2016.00200
Kaplan. WD. (1948). Formaldehyde as a mutagen in Drosophila. Science 108: 43.
http://dx.doi.org/10.1126/science.108.2793.43
This document is a draft for review purposes only and does not constitute Agency policy.
R-31 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Kastner. PE: Le Calve. S: Zheng. W: Casset. A: Pons. F. (2013). A dynamic system for single and
repeated exposure of airway epithelial cells to gaseous pollutants. Toxicol In Vitro 27:632-
640. http://dx.doi.0rg/lO.lOl6/i.tiv.2Oi2.ll.Oll
Katsnelson. BA: Degtyareva. TP: Privalova. LI: Minigaliveva. IA: Slvshkina. TV: Rvzhov. W:
Beresneva. OY. u. (2013). Attenuation of subchronic formaldehyde inhalation toxicity with
oral administration of glutamate, glycine and methionine. Toxicol Lett 220: 181-186.
http: / /dx. doi. or g/10.1016 /i. toxlet. 2013.04.024
Keller. DA: Heck. H: Randall. HW: Morgan. KT. (1990). Histochemical localization of formaldehyde
dehydrogenase in the rat Toxicol Appl Pharmacol 106: 311-326.
http://dx.d0i.0rg/l 0.1016/0041 -008xC901902S0-x
Kelly. TT: Smith. PL: Satola. 1. (1999). Emission rates of formaldehyde from materials and consumer
products found in California homes. Environ Sci Technol 33: 81-88.
http://dx.doi.org/10.1021/es980592%2B
Kennedy. G: Slaich. PK: Golding. BT: Watson. WP. (1996). Structure and mechanism of formation of
a new adductfrom formaldehyde and guanosine. Chem Biol Interact 102: 93-100.
http://dx.doi.org/10.1016/S0009-2797r96103737-4
Kepler. GM: Richardson. RB: Morgan. KT: Kimbell. IS. (1998). Computer simulation of inspiratory
nasal airflow and inhaled gas uptake in a rhesus monkey. Toxicol Appl Pharmacol 150: 1-11.
http: / /dx. doi. or g/10.10 0 6 /taap. 1997.8350
Kerns. WD: Pavkov. KL: Donofrio. DT: Gralla. EI: Swenberg. TA. (1983). Carcinogenicity of
formaldehyde in rats and mice after long-term inhalation exposure. Cancer Res 43: 4382-
4392.
Khan. AH. (1967). The induction of crossing over in the absence of mutation. Sind University
Science Research Journal 3: 103-106.
Kiernan. TA. (2000). Formaldehyde, formalin, paraformaldehyde and glutaraldehyde: What they are
and what they do. Microsc Today 00-1: 8-12.
Kilburn. KH: Mckenzie. WN. (1978). Leukocyte recruitment to airways by aldehyde-carbon
combinations that mimic cigarette smoke. Lab Invest 38: 134-142.
Kilburn. KH: Moro. A. (1985). REPRODUCTIVE AND MATERNAL EFFECTS OF FORMALDEHYDE IN
RATS (pp. 21-26). (BIOSIS/85/04191). Kilburn, KH; Moro, A.
Kilburn. KH: Seidman. BC: Warshaw. R. (1985). Neurobehavioral and respiratory symptoms of
formaldehyde and xylene exposure in histology technicians. Arch Environ Occup Health 40:
229-233. http://dx.doi.Org/10.1080/00039896.1985.10545924
Kilburn. KH: Warshaw. R: Thornton. TC. (1987). Formaldehyde impairs memory, equilibrium, and
dexterity in histology technicians: Effects which persist for days after exposure. Arch
Environ Occup Health 42: 117-120. http://dx.doi.org/10.1080/00039896.1987.9935806
Kilburn. KH: Warshaw. R: Thornton. TC: Husmark. I. (1989). An examination of factors that could
affect choice reaction time in histology technicians. Am J Ind Med 15: 679-686.
http://dx.doi.Org/10.1002/aiim.4700150607
Kilburn. KH: Warshaw. RH. (1992). Neurobehavioral effects of formaldehyde and solvents on
histology technicians: Repeated testing across time. Environ Res 58: 134-146.
http://dx.doi.org/10.1016/S0013-935ir05180210-5
This document is a draft for review purposes only and does not constitute Agency policy.
R-32 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Kim. EM: Lee. HY: Lee. EH: Lee. KM: Park. M: Ti. KY: Tang. TH: Teong. YH: Lee. KH: Yoon. IT: Kim. SM:
Teong. MT: Kim. KD: Kang. HS. (2013a). Formaldehyde exposure impairs the function and
differentiation of NK cells. Toxicol Lett 223: 154-161.
http://dx.doi.Org/10.1016/j.toxlet.2013.09.008
Kim. H: Kim. YD: Cho. SH. (1999). Formaldehyde exposure levels and serum antibodies to
formaldehyde-human serum albumin of Korean medical students. Arch Environ Health 54:
115-118. http://dx.doi.Org/l 0.1080/00039899909602245
Kim. H: Levin. L: Lemasters. GK: Villareal. M: Evans. S: Lockev. IE: Hershev. GKK: Bernstein. PI.
(2012). Validating childhood symptoms with physician-diagnosed allergic rhinitis. Ann
Allergy Asthma Immunol 108: 228-231. http: //dx.doi.Org/10.1016/i.anai.2012.02.004
Kim. TL: Elfman. L: Mi. Y: Wieslander. G: Smedje. G: Norback. D. (2007). Indoor molds, bacteria,
microbial volatile organic compounds and plasticizers in schools - associations with asthma
and respiratory symptoms in pupils. Indoor Air 17: 153-163.
http://dx.doi. org/10.1111 /i.l 600-0668.2006.00466.X
Kim. TL: Elfman. L: Wieslander. G: Ferm. M: Toren. K: Norback. D. (2011). Respiratory health among
Korean pupils in relation to home, school and outdoor environment J Korean Med Sci 26:
166-173. http://dx.doi.Org/10.3346/ikms.2011.26.2.166
Kim. TY: Teong. MS: Park. KY: Seo. ST. (2013b). Aggravation of atopic dermatitis-like symptoms by
consecutive low concentration of formaldehyde exposure in NC/Nga mice [Letter], Exp
Dermatol 22: 219-221. http://dx.doi.org/10.llll/exd.12Q92
Kim. SM: Hwang. KA: Choi. DW: Choi. KC. (2018). The cigarette smoke components induced the cell
proliferation and epithelial to mesenchymal transition via production of reactive oxygen
species in endometrial adenocarcinoma cells. Food Chem Toxicol 121: 657-665.
http://dx.doi.Org/10.1016/i.fct2018.09.023
Kim. Y: Tekarl. DW: Kim. T: Kwon. A: Choi. H: Lee. S: Kim. YT: Kim. HI: Kim. Y: Oh. TH: Kim. M. (2015).
Genetic and epigenetic alterations of bone marrow stromal cells in myelodysplastic
syndrome and acute myeloid leukemia patients. Stem Cell Research 14: 177-184.
http://dx.doi.Org/10.1016/i.scr.2015.01.004
Kimbell. IS: Gross. EA: Tovner. PR: Godo. MN: Morgan. KT. (1993). Application of computational fluid
dynamics to regional dosimetry of inhaled chemicals in the upper respiratory tract of the
rat. Toxicol Appl Pharmacol 121: 253-263. http://dx.doi.org/10.1006/taap.1993.1152
Kimbell. IS: Gross. EA: Richardson. RB: Conollv. RB: Morgan. KT. (1997). Correlation of regional
formaldehyde flux predictions with the distribution of formaldehyde-induced squamous
metaplasia in F344 rat nasal passages. MutatRes 380: 143-154.
http://dx.doi.org/10.1016/S0027-5107r97100132-2
Kimbell. IS: Overton. TH: Subramaniam. RP: Schlosser. PM: Morgan. KT: Conollv. RB: Miller. FT.
(2001a). Dosimetry modeling of inhaled formaldehyde: Binning nasal flux predictions for
quantitative risk assessment. Toxicol Sci 64: 111-121.
Kimbell. IS: Subramaniam. RP. (2001). Use of computational fluid dynamics models for dosimetry of
inhaled gases in the nasal passages [Review], Inhal Toxicol 13: 325-334.
http://dx.doi.Org/10.1080/08958370120442
Kimbell. IS: Subramaniam. RP: Gross. EA: Schlosser. PM: Morgan. KT. (2001b). Dosimetry modeling
of inhaled formaldehyde: comparisons of local flux predictions in the rat, monkey, and
human nasal passages. Toxicol Sci 64: 100-110.
This document is a draft for review purposes only and does not constitute Agency policy.
R-33 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Kimura. R: Kimoto. I: Takeda. M: Mivake. M: Sakamoto. T. (2010). Alteration in airway
microvascular leakage induced by sensorineural stimulation in rats exposed to inhaled
formaldehyde. Toxicol Lett 199: 254-260. http://dx.doi.Org/10.1016/i.toxlet.2010.09.007
Kita. H: Oomichi. S. (1974). Effect of air pollutants on ciliary activity of respiratory tract. Bull Tokyo
Med DentUniv 21: 327-343.
Kita. T: Fuiimura. M: Mvou. S: Ishiura. Y: Abo. M: Katavama. N: Nishitsuii. M: Yoshimi. Y: Nomura. S:
Oribe. Y: Nakao. S. (2003). Potentiation of allergic bronchoconstriction by repeated
exposure to formaldehyde in guinea-pigs in vivo. Clin Exp Allergy 33: 1747-1753.
http://dx.doi.Org/10.llll/i.1365-2222.2003.01826.x
Kitaev. EM: Savchenko. ON: Lovchikov. VA: Altukhov. W: Vishnvakov. YS. (1984). Razvitie
zarodyshey i nekotorye pokazateli reproductivnoy funktsii u krys posle ingalyatsionnogo
vozdeystviya formal'degida do oplodotvoreniya [Akush Ginekol 10: 49-52.
Kitaeva. L: Kitaev. E: Pimenova. M. (1990). Cytopathic and cytogenetic effects of chronic inhalation
of formaldehyde on the female rat's germ and marrow cells. Tsitologiia 32: 1212-1216.
Kitaeva. LV: Mikheeva. EA: Shelomova. LF: Shvartsman. PI. a. (1996). [Genotoxic effect of
formaldehyde in somatic human cells in vivo], Genetika 32: 1287-1290.
Klein. MP: Sinha. BK: Subramaniam. RP. (2011). Statistical inferences from formaldehyde DNA-
protein cross-link data: improving methods for characterization of uncertainty. J Biopharm
Stat 21: 42-55. http://dx.doi.Org/l 0.1080/10543400903531601
Kleinniienhuis. AT: Staal. YC: Duistermaat. E: Engel. R: Woutersen. RA. (2013). The determination of
exogenous formaldehyde in blood of rats during and after inhalation exposure. Food Chem
Toxicol 52: 105-112. http: //dx.doi.org/10.1016/i.fct.2012.11.008
Kligerman. AD: Phelps. MC: Erexson. GL. (1984). Cytogenetic analysis of lymphocytes from rats
following formaldehyde inhalation. Toxicol Lett 21: 241-246.
http: //dx.doi.org/10.1016/0378-4274r84190079-l
Kolmark. G: Westergaard. M. (1953). Further studies on chemically induced reversions at the
adenine locus of Neurospora. Hereditas 39: 209-224.
Krakowiak. A: Gorski. P: Pazdrak. K: Ruta. U. (1998). Airway response to formaldehyde inhalation in
asthmatic subjects with suspected respiratory formaldehyde sensitization. Am J Ind Med 33:
274-281. http://dx.doi.org/10.1002 /fSTCni 097-0274fl 99803133:3<274::A1D-
AJIM9>3.0.CO;2-W
Kreiger. RA: Garry. VF. (1983). Formaldehyde-induced cytotoxicity and sister-chromatid exchanges
inhuman lymphocyte cultures. MutatRes 120: 51-55. http://dx.doi.org/10.1016/Q165-
7992f83190073-8
Kriebel. D: Myers. D: Cheng. M: Woskie. S: Cocanour. B. (2001). Short-term effects of formaldehyde
on peak expiratory flow and irritant symptoms. Arch Environ Health 56: 11-18.
http: //dx.doi.org/10.1080/00039890109604049
Kriebel. D: Sama. SR: Cocanour. B. (1993). Reversible pulmonary responses to formaldehyde. A
study of clinical anatomy students. Am Rev Respir Dis 148: 1509-1515.
http://dx.doi.Org/10.1164/airccm/148.6 Pt 1.1509
Krzyzanowski. M: Ouackenboss. IT: Lebowitz. MP. (1990). Chronic respiratory effects of indoor
formaldehyde exposure. Environ Res 52: 117-125. http://dx.doi.org/10.1016/S0Q13-
9351C05180247-6
This document is a draft for review purposes only and does not constitute Agency policy.
R-34 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Kuehner. S: Holzmann. K: Speit. G. (2013). Characterization of formaldehyde's genotoxic mode of
action by gene expression analysis in TK6 cells. Arch Toxicol 87: 1999-2012.
http://dx.doi.Org/l 0.1007/s00204-013-1060-2
Kuehner. S: Schlaier. M: Schwarz. K: Speit. G. (2012). Analysis of leukemia-specific aneuploidies in
cultured myeloid progenitor cells in the absence and presence of formaldehyde exposure.
Toxicol Sci 128: 72-78. http: / /dx. doi. o r g/10.10 9 3/toxsci/kfs 126
Kulle. TT. (1993). Acute odor and irritation response in healthy nonsmokers with formaldehyde
exposure. Inhal Toxicol 5: 323-332. http://dx.doi.org/10.3109/08958379308998389
Kulle. TT: Cooper. GP. (1975). Effects of formaldehyde and ozone on the trigeminal nasal sensory
system. Arch Environ Occup Health 30: 237-243.
Kulle. TT: Sauder. LR: Hebel. TR: Green. DT: Chatham. MP. (1987a). Formaldehyde dose-response in
healthy nonsmokers. J Air Waste Manag Assoc 37: 919-924.
http://dx.doi.org/10.1080/08940630.1987.10466285
Kulle. TT: Sauder. LR: Hebel. TR: Green. DT: Chatham. MP. (1987b). Formaldehyde dose-response in
healthy nonsmokers. J Air Pollut Control Assoc 37: 919-924.
http://dx.doi.Org/10.1080/08940630.1987.10466285
Kum. C: Kiral. F: Sekkin. S: Sevrek. K: Bovacioglu. M. (2007a). Effects of xylene and formaldehyde
inhalations on oxidative stress in adult and developing rats livers. Exp Anim 56: 35-42.
http://dx.doi.org/10.1538/expanim.56.35
Kum. C: Sekkin. S: Kiral. F: Akar. F. (2007b). Effects of xylene and formaldehyde inhalations on renal
oxidative stress and some serum biochemical parameters in rats. Toxicol Ind Health 23:
115-120. http://dx.doi.org/10.1177/0748233707078218
Kumari. A: Lim. YX: Newell. AH: Olson. SB: Mccullough. AK. (2012). Formaldehyde-induced genome
instability is suppressed by an XPF-dependent pathway. DNA Repair 11: 236-246.
http://dx.doi.0rg/lO.lOl6/i.dnarep.2Oll.ll.OOl
Kunkler. PE: Ballard. CI: Oxford. GS: Hurley. TH. (2011). TRPA1 receptors mediate environmental
irritant-induced meningeal vasodilatation. Pain 152: 38-44.
http://dx.doi.0rg/lO.lOl6/i.pain.2OlO.O8.O2i
Kuo. HW: Tian. GT: Chen. CL: Liu. CS: Lai. IS. (1997). White blood cell count as an indicator of
formaldehyde exposure. Bull Environ Contam Toxicol 59: 261-267.
http://dx.doi.Org/10.1007/s001289900473
Kuper. CF: van Oostrum. L: Ma-Hock. L: Durrer. S: Woutersen. RA. (2011). Hyperplasia of the
lymphoepithelium of NALT in rats but not in mice upon 28-day exposure to 15 ppm
formaldehyde vapor. Exp Toxicol Pathol 63: 25-32.
http://dx.doi.Org/10.1016/i.etp.2009.09.004
Kurttio. P: Norppa. H: Tarventaus. H: Sorsa. M: Kalliokoski. P. (1993). Chromosome aberrations in
peripheral lymphocytes of workers employed in the plywood industry. Scand J Work
Environ Health 19: 132-134. http://dx.doi. or g/10.5 2 71 /sjweh. 1495
Kushch. I: Schwarz. K: Schwentner. L: Baumann. B: Dzien. A: Schmid. A: Unterkofler. K: Gastl. G:
Spanel. P: Smith. D: Amann. A. (2008). Compounds enhanced in a mass spectrometric profile
of smokers' exhaled breath versus non-smokers as determined in a pilot study using PTR-
MS. J Breath Res 2: 026002. http://dx.doi.Org/10.1088/1752-7155/2/2/026002
This document is a draft for review purposes only and does not constitute Agency policy.
R-35 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Kuvkendall. TR: Bogdanffv. MS. (1992). Efficiency of DNA-histone crosslinking induced by saturated
and unsaturated aldehydes in vitro. DNA Repair 283: 131-136.
http://dx.doi. org/10.1016/0165-7992C92190145-8
Ladeira. C: Viegas. S: Carolino. E: Gomes. MC: Brito. M. (2013). The influence of genetic
polymorphisms in XRCC3 and ADH5 genes on the frequency of genotoxicity biomarkers in
workers exposed to formaldehyde. Environ Mol Mutagen 54: 213-221.
http: / /dx. doi. or g/10.10 0 2 /em. 2175 5
Ladeira. C: Viegas. S: Carolino. E: Prista. 1: Gomes. MC: Brito. M. (2011). Genotoxicity biomarkers in
occupational exposure to formaldehyde-the case of histopathology laboratories. Mutat Res
721: 15-20. http://dx.doi.Org/10.1016/i.mrgentox.2010.ll.015
Laforest. L: Luce. D: Goldberg. P: Begin. D: Gerin. M: Demers. PA: Brugere. 1: Leclerc. A. (2000).
Laryngeal and hypopharyngeal cancers and occupational exposure to formaldehyde and
various dusts: A case-control study in France. Occup Environ Med 57: 767-773.
http://dx.doi.org/10.1136/oem.57.ll.767
Lai. Y: Yu. R: Hartwell. HI: Moeller. BC: Bodnar. WM: Swenberg. TA. (2016). Measurement of
Endogenous versus Exogenous Formaldehyde-Induced DNA-Protein Crosslinks in Animal
Tissues by Stable Isotope Labeling and Ultrasensitive Mass Spectrometry. Cancer Res 76:
2652-2661. http://dx.doi.org/10.1158/0008-5472.CAN-15-2527
Lajoie. P: Aubin. D: Gingras. V: Daigneault. P: Ducharme. F: Gauvin. FD: Fugler. D: Leclerc. TM: Won.
D: Won. D: Courteau. M: Gingras. S: Heroux. ME: Yang. W: Schleibinger. H. (2014). The
IVAIRE Project - A Randomized Controlled Study of the Impact of Ventilation on Indoor Air
Quality and the Respiratory Symptoms of Asthmatic Children in Single Family Homes.
Indoor Air 25: 582-597. http://dx.doi.org/10.llll/ina.12181
Lakwiik. N: Van Strien. RT: Doekes. G: Brunekreef. B: Gerritsen. 1. (1998). Validation of a screening
questionnaire for atopy with serum IgE tests in a population of pregnant Dutch women. Clin
Exp Allergy 28: 454-458. http://dx.doi.org/10.1046/i.l 365-2222.1998.00254.x
Lam. CW: Casanova. M: Heck. H. (1985). Depletion of nasal mucosal glutathione by acrolein and
enhancement of formaldehyde-induced DNA-protein cross-linking by simultaneous
exposure to acrolein. Arch Toxicol 58: 67-71. http://dx.doi.org/10.1007/bf00348311
Lan. 0: Smith. MT: Tang. X: Guo. W: Vermeulen. R: Ti. Z: Hu. W: Hubbard. AE: Min. S: Mchale. CM: Oiu.
C: Liu. S: Reiss. B: Beane Freeman. L: Blair. A: Ge. Y: Xiong. I: Li. L: Rappaport. SM: Huang. H:
Rothman. N: Zhang. L. (2015). Chromosome-wide aneuploidy study (CWAS) of cultured
circulating myeloid progenitor cells from workers occupationally exposed to formaldehyde.
Carcinogenesis 36: 160-167. http://dx.doi.org/10.1093/carcin/bgu229
Lang. I: Bruckner. T: Triebig. G. (2008). Formaldehyde and chemosensory irritation in humans: A
controlled human exposure study. Regul Toxicol Pharmacol 50: 23-36.
http://dx.doi.Org/10.1016/i.yrtph.2007.08.012
Larsen. ST: Wolkoff. P: Hammer. M: Kofoed-S0rensen. V: Clausen. PA: Nielsen. GD. (2013). Acute
airway effects of airborne formaldehyde in sensitized and non-sensitized mice housed in a
dry or humid environment Toxicol Appl Pharmacol 268: 294-299.
http://dx.doi.Org/10.1016/j.taap.2013.02.006
Lazenbv. V: Hinwood. A: Callan. A: Franklin. P. (2012). Formaldehyde personal exposure
measurements and time weighted exposure estimates in children. Chemosphere 88: 966-
973. http://dx.doi.Org/10.1016/i.chemosphere.2012.03.029
This document is a draft for review purposes only and does not constitute Agency policy.
R-36 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Lazutka. TR: Lekevicius. R: Dedonvte. V: Maciuleviciute-Gervers. L: Mierauskiene. T: Rudaitiene. S:
Slapsvte. G. (1999). Chromosomal aberrations and sister-chromatid exchanges in
Lithuanian populations: Effects of occupational and environmental exposures. MutatRes
Genet Toxicol Environ Mutagen 445: 225-239. http: //dx.doi.org/10.1016/S1383-
5718C99100128-X
Le Curieux. F: Marzin. D: Erb. F. (1993). Comparison of three short-term assays: Results on seven
chemicals. Potential contribution to the control of water genotoxicity. Mutat Res 319: 223-
236. http://dx.doi.Org/l 0.1016/0165-1218(93190082-0
Leal. MP: Brochetti. RA: Ignacio. A: Camara. NOS: da Palma. RK: de Oliveira. LVF: de Fatima Teixeira
da Silva. D: Lino-Dos-Santos-Franco. A. (2018). Effects of formaldehyde exposure on the
development of pulmonary fibrosis induced by bleomycin in mice. Toxicology Reports 5:
512-520. http ://dx.doi. or g/10.1016/i.toxrep.2018.03.016
Lebowitz. MP: Krzvzanowski. M: Ouackenboss. 11: Orourke. MK. (1997). Diurnal variation of PEF
and its use in epidemiological studies. Eur Respir J10: S49-S56.
Lee. HK: Alarie. Y: Karol. MH. (1984). Induction of formaldehyde sensitivity in guinea pigs. Toxicol
Appl Pharmacol 75: 147-155. http://dx.doi.org/10.1016/0041-008Xr84190085-l
Leikauf. GD. (1992). Mechanisms of aldehyde-induced bronchial reactivity: role of airway
epithelium. Res Rep Health Eff Instl-35.
Leng. 1: Liu. CW: Hartwell. HI: Yu. R: Lai. Y: Bodnar. WM: Lu. K: Swenberg. TA. (2019). Evaluation of
inhaled low-dose formaldehyde-induced DNA adducts and DNA-protein cross-links by
liquid chromatography-tandem mass spectrometry. Arch Toxicol 93: 763-773.
http: //dx.doi.org/10.1007/s00204-019-02393-x
Levy. S: Nocentini. S: Billardon. C. (1983). Induction of cytogenetic effects in human fibroblast
cultures after exposure to formaldehyde or X-rays. Mutat Res 119: 309-317.
http: //dx.doi.org/10.1016/0165-7992C83190179-3
Li. AM: Fung. CK: Yu. IT: Goggins. WB: Chan. GY: Chan. CK: Lau. AP: Leung. TO. (2019). Associations
of wheeze during the first 18 months of life with indoor nitrogen dioxide, formaldehyde,
and family history of asthma: a prospective cohort study. Hong Kong Med J 25 Suppl 3:20-
23.
Li. F: Oin. Y: Gong. S: Zhang. H: Ding. S. (2020). Learning and memory impairment of mice caused by
gaseous formaldehyde. Environ Res 184: 109318.
http ://dx. doi. or g/10.1016/i. envres. 2020.109318
Li. G: Yang. I: Ling. S. (2015). Formaldehyde exposure alters miRNA expression profiles in the
olfactory bulb. Inhal Toxicol 27: 1-7. http://dx.doi.org/10.3109/08958378.2015.1062580
Li. GY: Lee. HY: Choi. YT: Lee. MO: Shin. HS: Kim. HY: Lee. SB: Lee. BH. (2008). Changes in the
Expression of Ras-family Genes in Rats Exposed to Formaldehyde by Inhalation.
Toxicological Research 24: 201-206. http://dx.doi.Org/10.5487/TR.2008.24.3.201
Li. L: Hua. L: He. Y: Bao. Y. (2017). Differential effects of formaldehyde exposure on airway
inflammation and bronchial hyperresponsiveness in BALB/c and C57BL/6 mice. PLoS ONE
12: eO 179231. http://dx.doi.org/10.1371/iournal.pone.0179231
Li. R: Lu. ZS: Qiao. Y: Yao. HC: Yu. FF: Yang. X. (2004). Study on the formaldehyde-induced DNA
damage with comet assay. Shi Yan Sheng WuXue Bao 37: 262-268.
This document is a draft for review purposes only and does not constitute Agency policy.
R-37 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Li. W: Ray. RM: Gao. PL: Fitzgibbons. ED: Seixas. NS: Camp. TE: Wernli. KT: Astrakianakis. G: Feng. Z:
Thomas. DB: Checkowav. H. (2006). Occupational risk factors for nasopharyngeal cancer
among female textile workers in Shanghai, China. Occup Environ Med 63: 39-44.
http://dx.doi.org/10.1136/oem.2005.021709
Li. Y: Song. Z: Ding. Y: Xin. Y. e: Wu. T: Su. T. ao: He. R: Tai. F: Lian. Z. (2016). Effects of formaldehyde
exposure on anxiety-like and depression-like behavior, cognition, central levels of
glucocorticoid receptor and tyrosine hydroxylase in mice. Chemosphere 144: 2004-2012.
http://dx.doi.Org/10.1016/j.chemosphere.2015.10.102
Liao. S: Tiang. L: Zhang. X. (2010). [Effects of inhaled formaldehyde on learning and memory and
expression of CaMK II in hippocampus of Wistar rats of different ages], 35: 342-344.
Liber. HL: Benforado. K: Crosby. RM: Simpson. D: Skopek. TR. (1989). Formaldehyde-induced and
spontaneous alterations in human hprt DNA sequence and mRNA expression. Mutat Res
226: 31-37. http://dx.doi.org/10.1016/0165-7992r89190089-4
Lima. LF: Murta. GL: Bandeira. AC: Nardeli. CR: Lima. WG: Bezerra. FS. (2015). Short-term exposure
to formaldehyde promotes oxidative damage and inflammation in the trachea and
diaphragm muscle of adult rats. Ann Anat 202: 45-51.
http://dx.doi.Org/10.1016/i.aanat.2015.08.003
Lin. D: Guo. Y: Yi. 1: Kuang. D. an: Li. X: Deng. H: Huang. K. un: Guan. L. ei: He. Y: Zhang. X: Hu. D. ie:
Zhang. Z: Zheng. H: Zhang. X: Mchale. CM: Zhang. L: Wu. T. (2013). Occupational exposure to
formaldehyde and genetic damage in the peripheral blood lymphocytes of plywood
workers. J Occup Health 55: 284-291. http://dx.doi.org/10.1539/ioh.12-0288-OA
Lin. Z: Luo. W: Li. H: Zhang. Y. (2005). The effect of endogenous formaldehyde on the rat aorta
endothelial cells. Toxicol Lett 159: 134-143. http://dx.doi.Org/10.1016/j.toxlet2005.05.003
Lindbohm. ML: Hemminki. K. (1988). Nationwide data base on medically diagnosed spontaneous
abortions in Finland. Int J Epidemiol 17: 568-573. http://dx.doi.Org/10.1093/iie/17.3.568
Lindbohm. ML: Hemminki. K: Bonhomme. MG: Anttila. A: Rantala. K: Heikkila. P: Rosenberg. MI.
(1991). Effects of paternal occupational exposure on spontaneous abortions. Am J Public
Health 81: 1029-1033. http: //dx.doi.org/10.2105 /aiph.81.8.1029
Lino-Dos-Santos-Franco. A: Amemiva. RM: de Oliveira. AP: Damazo. AS: Breithaupt-Faloppa. AC:
Vitoretti. LB: Acceturi. BG: Tavares-De-Lima. W. (2013a). The putative role of ovary removal
and progesterone when considering the effect of formaldehyde exposure on lung
inflammation induced by ovalbumin. Clinics 68: 1528-1536.
http://dx.doi.org/10.6061/clinics/2013ri2109
Lino-Dos-Santos-Franco. A: Amemiva. RM: Ligeiro de Oliveira. AP: Breithaupt-Faloppa. AC: Damazo.
AS: Oliveira-Filho. RM: Tavares-De-Lima. W. (2011a). Differential effects of female sex
hormones on cellular recruitment and tracheal reactivity after formaldehyde exposure.
Toxicol Lett 205: 327-335. http://dx.doi.org/10.1016/i.toxlet.2011.06.023
Lino-Dos-Santos-Franco. A: Correa-Costa. M: Durao. AC: de Oliveira. AP: Breithaupt-Faloppa. AC:
Bertoni. 1: Oliveira-Filho. RM: Camara. NO: Marcourakis. T: Tavares-De-Lima. W. (2011b).
Formaldehyde induces lung inflammation by an oxidant and antioxidant enzymes mediated
mechanism in the lung tissue. Toxicol Lett 207: 278-285.
http://dx.doi.Org/10.1016/i.toxlet.2011.09.026
Lino-Dos-Santos-Franco. A: Domingos. HV: de Oliveira. AP: Breithaupt-Faloppa. AC: Peron. TP:
Bolonheis. S: Muscara. MN: Oliveira-Filho. RM: Vargaftig. BB: Tavares-De-Lima. W. (2010).
This document is a draft for review purposes only and does not constitute Agency policy.
R-38 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Differential effects of formaldehyde exposure on the cell influx and vascular permeability in
a rat model of allergic lung inflammation. Toxicol Lett 197: 211-218.
Lino-Dos-Santos-Franco. A: Gimenes-Tunior. TA: Ligeiro-De-Oliveira. AP: Breithaupt-Faloppa. AC:
Acceturi. BG: Vitoretti. LB: Machado. ID: Oliveira-Filho. RM: Farskv. SHP: Moriva. HT:
Tavares-De-Lima. W. (2013b). Formaldehyde inhalation reduces respiratory mechanics in a
rat model with allergic lung inflammation by altering the nitric oxide/cyclooxygenase-
derived products relationship. Food Chem Toxicol 59: 731-738.
http://dx.doi.Org/10.1016/j.fct2013.07.027
Lino dos Santos Franco. A: Damazo. AS: Beraldo de Souza. HR: Domingos. HV: Oliveira-Filho. RM:
Oliani. SM: Costa. SK: Tavares de Lima. W. (2006). Pulmonary neutrophil recruitment and
bronchial reactivity in formaldehyde-exposed rats are modulated by mast cells and
differentially by neuropeptides and nitric oxide. Toxicol Appl Pharmacol 214: 35-42.
http://dx.doi.Org/10.1016/i.taap.2005.ll.014
Lino dos Santos Franco. A: Domingos. HV: Damazo. AS: Breithaupt-Faloppa. AC: de Oliveira. AP:
Costa. SK: Oliani. SM: Oliveira-Filho. RM: Vargaftig. BB: Tavares-De-Lima. W. (2009).
Reduced allergic lung inflammation in rats following formaldehyde exposure: Long-term
effects on multiple effector systems. Toxicology 256: 157-163.
http ://dx. doi. or g/10.1016/i. tox. 2 0 0 8.11.011
Liteplo. RG: Meek. ME. (2003). Inhaled formaldehyde: Exposure estimation, hazard
characterization, and exposure-response analysis [Review], J Toxicol Environ Health B Crit
Rev 6: 85-114. http://dx.doi.Org/l 0.1080/10937400306480
Liu. D: Zheng. Y: Li. B: Yao. H: Li. R: Zhang. Y: Yang. X. (2011). Adjuvant effects of gaseous
formaldehyde on the hyper-responsiveness and inflammation in a mouse asthma model
immunized by ovalbumin. J Immunotoxicol 8: 305-314.
http: //dx.doi.org/10.3109/1547691 X.2011.600738
Liu. KS: Huang. FY: Havward. SB: Wesolowski. I: Sexton. K. (1991). Irritant effects of formaldehyde
exposure in mobile homes. Environ Health Perspect 94: 91-94.
http: //dx.doi.org/10.2307/3431298
Liu. L: Huang. Y: Feng. X: Chen. I: Duan. Y. (2019). Overexpressed Hsp70 alleviated formaldehyde-
induced apoptosis partly via PI3K/Akt signaling pathway in human bronchial epithelial
cells. Environ Toxicol 34: 495-504. http: / /dx. doi. o r g/10.10 0 2 /tox. 2 2 7 0 3
Liu. OB: Wang. W: Ting. W. (2018a). Indoor air pollution aggravates asthma in Chinese children and
induces the changes in serum level of miR-155. Int J Environ Health Res 29: 1-9.
http://dx.doi.Org/10.1080/09603123.2018.1506569
Liu. OP: Zhou. DX: Lv. MO: Ge. P: Li. YX: Wang. ST. (2018b). Formaldehyde inhalation triggers
autophagy in rat lung tissues. Toxicol Ind Health748233718796347.
http: //dx.doi.org/10.1177/0748233718796347
Liu. Y. i: Ye. Z: Luo. H: Sun. M: Li. M. i: Fan. D: Chui. D. (2009a). Inhalative formaldehyde exposure
enhances aggressive behavior and disturbs monoamines in frontal cortex synaptosome of
male rats. Neurosci Lett 464: 113-116. http://dx.doi.Org/10.1016/i.neulet.2009.06.037
Liu. Y: Li. CM: Lu. Z: Ding. S: Yang. X: Mo. 1. (2006). Studies on formation and repair of
formaldehyde-damaged DNAby detection of DNA-protein crosslinks and DNA breaks. Front
Biosci 11: 991-997. http://dx.doi.org/10.2741/1856
This document is a draft for review purposes only and does not constitute Agency policy.
R-39 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Liu. Y: Ye. Z: Yang. H: Zhou. L: Fan. D: He. S: Chui. D. (2010). Disturbances of soluble N-
ethylmaleimide-sensitive factor attachment proteins in hippocampal synaptosomes
contribute to cognitive impairment after repetitive formaldehyde inhalation in male rats.
Neuroscience 169: 1248-1254. http://dx.doi.Org/10.1016/j.neuroscience.2010.05.061
Liu. Y: Yu. D: Xiao. S. (2017). Effects of chronic exposure to Formaldehyde on micronucleus rate of
bone marrow cells in male mice. J Pak Med Assoc 67: 933-935.
Liu. YR: Zhou. Y: Oiu. W: Zeng. TY: Shen. LL: Li. AP: Zhou. TW. (2009b). Exposure to formaldehyde
induces heritable DNA mutations in mice. J Toxicol Environ Health A 72: 767-773.
http://dx.doi.org/10.1080/15287390902841615
Lofstedt. H: Westberg. H: Selden. AI: Lundholm. C: Svartengren. M. (2009). Respiratory symptoms
and lung function in foundry workers exposed to low molecular weight isocyanates. Am J
Ind Med 52: 455-463. http: / / dx. do i. o r g /10.10 0 2 / aiim. 2 0 69 3
Lofstedt. H: Westberg. H: Selden. AI: Rudblad. S: Brvngelsson. IL: Ngo. Y: Svartengren. M. (2011).
Nasal and ocular effects in foundry workers using the hot box method. J Occup Environ Med
53: 43-48. http://dx.d0i.0rg/l 0.1097/TOM.QbOl3e318181 ff05cc
Lovreglio. P: Carrus. A: Iavicoli. S: Drago. I: Persechino. B: Soleo. L. (2009). Indoor formaldehyde
and acetaldehyde levels in the province of Bari, South Italy and estimated health risk. J
Environ Monit 11: 955-961. http://dx.doi.org/10.1039/b819843h
Lu. K. (2009) Molecular binding of formaldehyde to dna and proteins. (Doctoral Dissertation).
University of North Carolina at Chapel Hill, Chapel Hill, NC.
Lu. K: Bovsen. G: Gao. L: Collins. LB: Swenberg. TA. (2008a). Formaldehyde-induced histone
modifications in vitro. Chem Res Toxicol 21: 1586-1593.
http: / /dx. doi. or g/10.10 21 /tx80 0 0576
Lu. K: Collins. LB: Ru. H: Bermudez. E: Swenberg. TA. (2010). Distribution of DNA adducts caused by
inhaled formaldehyde is consistent with induction of nasal carcinoma but not leukemia.
Toxicol Sci 116: 441-451. http://dx.doi.org/10.1093/toxsci/kfq061
Lu. K: Craft. S: Nakamura. I: Moeller. BC: Swenberg. TA. (2012a). Use of LC-MS/MS and stable
isotopes to differentiate hydroxymethyl and methyl DNA adducts from formaldehyde and
nitrosodimethylamine. Chem Res Toxicol 25: 664-675.
http: / /dx. doi. or g/10.10 21 /tx2 0 042 6b
Lu. K: Gul. H: Upton. PB: Moeller. BC: Swenberg. TA. (2012b). Formation of hydroxymethyl DNA
adducts in rats orally exposed to stable isotope labeled methanol. Toxicol Sci 126: 28-38.
http: / /dx. doi. or g/10.109 3 /toxsci /kfr 3 2 8
Lu. K: Moeller. B: Dovle-Eisele. M: Mcdonald. I: Swenberg. TA. (2011). Molecular dosimetry of N2-
hydroxymethyl-dG DNA adducts in rats exposed to formaldehyde. Chem Res Toxicol 24:
159-161. http://dx.d0i.0rg/l0.1021 /txl003886
Lu. Z: Li. CM: Qiao. Y: Liu. Y: Yan. Y: Yang. X. (2005). Type II vanilloid receptor signaling system: One
of the possible mechanisms for the rise in asthma cases. Front Biosci 10: 2527-2533.
http://dx.doi.org/10.2741/1717
Lu. Z: Li. CM: Qiao. Y: Yan. Y: Yang. X. (2008b). Effect of inhaled formaldehyde on learning and
memory of mice. Indoor Air 18: 77-83. http://dx.doi.org/10.ll 11/j.1600-
0668.2008.00524.x
This document is a draft for review purposes only and does not constitute Agency policy.
R-40 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Luce. D: Leclerc. A: Begin. D: Demers. PA: Gerin. M: Orlowski. E: Kogevinas. M: Belli. S: Bugel. I:
Bolm-Audorff. U: Brinton. LA: Comba. P: Hardell. L: Hayes. RB: Magnani. C: Merler. E:
Preston-Martin. S: Vaughan. TL: Zheng. W: Boffetta. P. (2002). Sinonasal cancer and
occupational exposures: a pooled analysis of 12 case-control studies. Cancer Causes Control
13: 147-157. http://dx.doi.Org/10.1023/A:1014350004255
Lundberg. TM: Saria. A. (1983). Capsaicin-induced desensitization of airway mucosa to cigarette
smoke, mechanical and chemical irritants. Nature 302: 251-253.
Luo. YL: Guo. HM: Zhang. YL: Chen. PX: Zhu. YX: Huang. TH: Zhou. WL. (2013). Cellular mechanism
underlying formaldehyde-stimulated CI- secretion in rat airway epithelium. PLoS ONE 8:
e54494. http://dx.doi.org/10.1371/iournal.pone.0054494
Lvapina. M: Zhelezova. G: Petrova. E: Boev. M. (2004). Flow cytometric determination of neutrophil
respiratory burst activity in workers exposed to formaldehyde. Int Arch Occup Environ
Health 77: 335-340. http://dx.doi.Org/10.1007/s00420-004-0516-3
Ma. H: Song. X: Zhang. W: Ling. X: Yang. X: Wu. W: Lou. K: Xu. H. (2020). Formaldehyde inhibits
development of T lymphocytes in mice. Toxicol Environ Chem 102: 473-489.
http://dx.doi.Org/10.1080/02772248.2020.1815202
Ma. TH: Harris. MM. (1988). Review of the genotoxicity of formaldehyde [Review], Mutat Res Rev
Genet Toxicol 196: 37-59. http://dx.doi.org/l 0.1016/0165-1110C88190027-9
Macedo. R: Gomes. F: Leal. M: Barioni. E: Braga. T: Camara. N: Farskv. S: Franco. ALP. (2016a). Low
level laser treatment reduces oxidative stress induced by formaldehyde exposure by the
modulation of gene expression of oxidant and antioxidant enzymes in the lung tissue
[Abstract], Lasers Surg Med 48: 58. http://dx.doi.org/10.1002/lsm.22485
Macedo. RS: Leal. MP: Braga. TT: Barioni. ED: Duro. S: Ratto Tempestini Horliana. AC: Saraiva
Camara. NO: Marcourakis. T: Poliselli Farskv. SH: Lino-Dos-Santos-Franco. A. (2016b).
Photobiomodulation Therapy Decreases Oxidative Stress in the Lung Tissue after
Formaldehyde Exposure: Role of Oxidant/Antioxidant Enzymes. Mediators Inflamm 2016:
9303126. http://dx.doi.org/10.1155/2016/9303126
Mackerer. CR: Angelosanto. FA: Blackburn. GR: Schreiner. CA. (1996). Identification of
formaldehyde as the metabolite responsible for the mutagenicity of methyl tertiary-butyl
ether in the activated mouse lymphoma assay. Proc Soc Exp Biol Med 212: 338-341.
http://dx.doi.org/10.3181/00379727-212-44023
Macpherson. LI: Xiao. B: Kwan. KY: Petrus. Ml: Dubin. AE: Hwang. S: Cravatt. B: Corey. DP:
Patapoutian. A. (2007). An ion channel essential for sensing chemical damage. J Neurosci 27:
11412-11415. http://dx.d0i.0rg/l 0.1523/TNEIJROSCT.3600-07.2007
Madison. RE: Broughton. A: Thrasher. ID. (1991). Immunologic biomarkers associated with an acute
exposure to exothermic byproducts of a ureaformaldehyde spill. Environ Health Perspect
94: 219-223. http://dx.d0i.0rg/l 0.2307/3431314
Madureira. I: Paciencia. I: Cavaleiro-Rufo. 1: de Oliveira Fernandes. E. (2016). Indoor pollutant
exposure among children with and without asthma in Porto, Portugal, during the cold
season. Environ Sci PollutRes Int 23: 20539-20552. http://dx.doi.org/10.1007/sll356-
016-7269-x
Magana-Schwencke. N: Ekert. B. (1978). Biochemical analysis of damage induced in yeast by
formaldehyde. II. Induction of cross-links between DNA and protein. Mutat Res 51: 11-19.
http://dx.d0i.0rg/l 0.1016/0027-5107178190003-9
This document is a draft for review purposes only and does not constitute Agency policy.
R-41 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Magana-Schwencke. N: Ekert. B: Moustacchi. E. (1978). Biochemical analysis of damage induced in
yeast by formaldehyde. I. Induction of single-strand breaks in DNA and their repair. Mutat
Res 50: 181-193. http://dx.doi.Org/l0.1016/0027-5107C78190023-4
Magana-Schwencke. N: Moustacchi. E. (1980). Biochemical analysis of damage induced in yeast by
formaldehyde III. Repair of induced cross-links between DNA and proteins in the wild-type
and in excision-deficient strains. Mutat Res 70: 29-35. http: / /dx. doi. or g /10.1016 /0 0 2 7-
5107C80190055-X
Maiellaro. M: Correa-Costa. M: Vitoretti. LB: Gimenes Tunior. TA: Saraiva Camara. NO: Tavares-De-
Lima. W: Poliselli Farskv. SH: Lino-Dos-Santos-Franco. A. (2014). Exposure to low doses of
formaldehyde during pregnancy suppresses the development of allergic lung inflammation
in offspring. Toxicol Appl Pharmacol 278: 266-274.
http://dx.doi.Org/10.1016/i.taap.2014.05.003
Maiellaro. M: Macedo. RS: Mendes. E: Tavares-De-Lima. W: Ferreira. CM: Lino-Dos-Santos-Franco. A.
(2016). High dose of formaldehyde exposure during pregnancy increases neutrophils lung
influx evoked by ovalbumin in the offspring. Inflamm Res 65: 179-181.
http://dx.doi.Org/10.1007/s00011-015-0901-2
Main. DM: Hogan. TT. (1983). Health effects of low level exposure to formaldehyde. J Occup Environ
Med 25: 896-900. http://dx.doi.org/1 0.1097/00043764-198312000-00013
Malaka. T: Kodama. AM. (1990). Respiratory health of plywood workers occupationally exposed to
formaldehyde. Arch Environ Health 45: 288-294.
http://dx.doi.Org/10.1080/00039896.1990.10118748
Malek. FA: Moritz. KU: Fanghanel. 1. (2003a). Formaldehyde inhalation & open field behaviour in
rats. Indian J Med Res 118: 90-96.
Malek. FA: Moritz. KU: Fanghanel. I. (2003b). A study on specific behavioral effects of formaldehyde
in the rat J Exp Anim Sci 42: 160-170. http://dx.doi.org/1 0.1016/S0939-8600r03180009-3
Malek. FA: Moritz. KU: Fanghanel. I. (2003c). A study on the effect of inhalative formaldehyde
exposure on water labyrinth test performance in rats. Ann Anat 185: 277-285.
http://dx.doi.Org/10.1016/S0940-9602r03180040-7
Malek. FA: Moritz. KU: Fanghanel. 1. (2004). Effects of a single inhalative exposure to formaldehyde
on the open field behavior of mice. Int J Hyg Environ Health 207: 151-158.
http://dx.doi.org/10.1078/1438-4639-00268
Malker. HSR: Mclaughlin. IK: Weiner. TA: Silverman. DT: Blot. WT: TLE. E: Fraumeni. 1. r. I. F. (1990).
Occupational risk factors for nasopharyngeal cancer in Sweden. Br J Ind Med 47: 213-214.
http://dx.doi.Org/10.1136/oem.47.3.213
Marinari. UM: Ferro. M: Sciaba. L: Finollo. R: Bassi. AM: Brambilla. G. (1984). DNA-damaging activity
of biotic and xenobiotic aldehydes in Chinese hamster ovary cells. Cell Biochem Funct 2:
243-248. http: //dx.doi.org/10.1002 /cbf.290020411
Marks. GB: Ezz. W: Aust. N: Toelle. BG: Xuan. W: Belousova. E: Cosgrove. C: lalaludin. B: Smith. WT.
(2010). Respiratory health effects of exposure to low-NOx unflued gas heaters in the
classroom: A double-blind, cluster-randomized, crossover study. Environ Health Perspect
118: 1476-1482. http://dx.doi.org/10.1289/ehp.l 002186
Marnett. LI: Hurd. HK: Hollstein. MC: Levin. DE: Esterbauer. H: Ames. BN. (1985). Naturally
occurring carbonyl compounds are mutagens in Salmonella tester strain TA104. Mutat Res
Fundam Mol Mech Mutagen 148: 25-34. http://dx.doi.org/10.1016/0027-5107C85190204-0
This document is a draft for review purposes only and does not constitute Agency policy.
R-42 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Maronpot. RR: Miller. RA: Clarke. WT: Westerberg. RB: Decker. TR: Moss. OR. (1986). Toxicity of
formaldehyde vapor in B6C3F1 mice exposed for 13 weeks. Toxicology 41: 253-266.
http://dx.doi.Org/l 0.1016/0300-483XC86190180-0
Marsh. GM: Morfeld. P: Zimmerman. SD: Liu. Y: Balmert. LC. (2016). An updated re-analysis of the
mortality risk from nasopharyngeal cancer in the National Cancer Institute formaldehyde
worker cohort study. J Occup Med Toxicol 11:8. http://dx.doi.org/10.1186/sl2995-016-
0097-6
Marsh. GM: Youk. AO: Buchanich. TM: Cassidv. LP: Lucas. LI: Esmen. NA: Gathuru. IM. (2002).
Pharyngeal cancer mortality among chemical plant workers exposed to formaldehyde.
Toxicol Ind Health 18: 257-268. http: //dx.doi.org/10.1191 /0748233702th! 49oa
Marsh. GM: Youk. AO: Buchanich. TM: Erdal. S: Esmen. NA. (2007). Work in the metal industry and
nasopharyngeal cancer mortality among formaldehyde-exposed workers. Regul Toxicol
Pharmacol 48: 308-319. http: / /dx.doi. or g/10.1016/i.vrtph.2007.04.006
Martin. CN: Mcdermid. AC: Garner. RC. (1978). Testing of known carcinogens and noncarcinogens
for their ability to induce unscheduled DNA synthesis in HeLa cells. Cancer Res 38: 2621-
2627.
Matanoski. GM. (1989). Risks of pathologists exposed to formaldehyde (final report). (DHHS Grant
No. 5 R01-OH-01511-03). Baltimore, MD: Johns Hopkins University Department of
Epidemiology.
https://ntrl.ntis. gov/NTRL/dashboard/searchResults.xhtml?search0uery=PB91173682
Matsunaga. I: Mivake. Y: Yoshida. T: Miyamoto. S: Ohva. Y: Sasaki. S: Tanaka. K: Oda. H: Ishiko. 0:
Hirota. Y: Group. OMaCHS. (2008). Ambient formaldehyde levels and allergic disorders
among Japanese pregnant women: Baseline data from the Osaka maternal and child health
study. Ann Epidemiol 18: 78-84. http://dx.doi.Org/10.1016/i.annepidem.2007.07.095
Matsuoka. T: Takaki. A: Ohtaki. H: Shioda. S. (2010). Early changes to oxidative stress levels
following exposure to formaldehyde in ICR mice. J Toxicol Sci 35: 721-730.
http: / /dx. doi. or g/10.2131 /its. 3 5.721
Mavr. SI: Hafizovic. K: Waldfahrer. F: Iro. H: Kiitting. B. (2010). Characterization of initial clinical
symptoms and risk factors for sinonasal adenocarcinomas: results of a case-control study.
Int Arch Occup Environ Health 83: 631-638. http://dx.doi.Org/10.1007/s00420-009-0479-
5
Mcghee. ID: von Hippel. PH. (1975a). Formaldehyde as a probe of DNA structure. I. Reaction with
exocyclic amino groups of DNA bases. Biochemistry 14: 1281-1296.
http://dx.doi.Org/10.1021/bi00677a029
Mcghee. ID: von Hippel. PH. (1975b). Formaldehyde as a probe of DNA structure. II. Reaction with
endocyclic imino groups of DNA bases. Biochemistry 14: 1297-1303.
http: / /dx. doi. or g/10.10 21 /biO 0 677a0 3 0
McNamara. CR: Mandel-Brehm. 1: Bautista. DM: Siemens. I. an: Deranian. KL: Zhao. M: Havward. NT:
Chong. TA: Tulius. D: Moran. MM: Fanger. CM. (2007). TRPA1 mediates formalin-induced
pain. Proc Natl Acad Sci USA 104: 13525-13530.
http://dx.doi.Org/10.1073/pnas.0705924104
Mei. YF: Duan. CL: Li. XX: Zhao. Y: Cao. FH: Shang. S: Ding. SM: Yue. XP: Gao. G: Yang. H: Shen. LX:
Feng. XY: Tia. TP: Tong. ZO: Yang. X. (2016). Reduction of Endogenous Melatonin Accelerates
Cognitive Decline in Mice in a Simulated Occupational Formaldehyde Exposure
This document is a draft for review purposes only and does not constitute Agency policy.
R-43 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Environment Int J Environ Res Public Health 13.
http ://dx. doi. or g/10.3 3 9 0 /ii erphl 3 0 3 0 2 5 8
Meister. A: Anderson. ME. (1983). Glutathione [Review], Annu Rev Biochem 52: 711-760.
http://dx.doi.Org/10.1146/annurev.bi.52.070183.003431
Meng. F: Bermudez. E: Mckinzie. PB: Andersen. ME: III. CH: Parsons. BL. (2010). Measurement of
tumor-associated mutations in the nasal mucosa of rats exposed to varying doses of
formaldehyde. Regul Toxicol Pharmacol 57: 274-283.
http://dx.doi.Org/10.1016/j.yrtph.2010.03.007
Merk. 0: Speit. G. (1998). Significance of formaldehyde-induced DNA-protein crosslinks for
mutagenesis. Environ Mol Mutagen 32: 260-268. http://dx.doi.org/10.1002/fSICniQ98-
22 80fl 998132:3<260::ATD-EM9>3.0.CQ:2-M
Merk. 0: Speit. G. (1999). Detection of crosslinks with the comet assay in relationship to
genotoxicity and cytotoxicity. Environ Mol Mutagen 33: 167-172.
http://dx.doi.Org/l 0.1002/rsicin098-2280n999133:2<167::aid-em9>3.0.co:2-d
Meyers. AR: Pinkerton. LE: Hein. Ml. (2013). Cohort mortality study of garment industry workers
exposed to formaldehyde: Update and internal comparisons. Am J Ind Med 56: 1027-1039.
http: / /dx. doi. or g/10.10 0 2 /ai im. 2 2199
Mi. L: Sui. 1: Wu. Y: Liang. G: Zhang. Y: Pu. Y: Tian. Y: Liu. S: Tiang. L. (2019). Bioinspired in vitro lung
airway model for inflammatory analysis via hydrophobic nanochannel membrane with joint
three-phase interface. Anal Chem 91: 15804-15810.
http://dx.doi.org/10.1021/acs.analchem.9b04114
Mi. YH: Norback. D: Tao. 1: Mi. YL: Ferm. M. (2006). Current asthma and respiratory symptoms
among pupils in Shanghai, China: Influence of building ventilation, nitrogen dioxide, ozone,
and formaldehyde in classrooms. Indoor Air 16: 454-464.
http: //dx.doi.org/10.1111 /i.l 600-0668.2006.00439.X
Migliore. L: Ventura. L: Barale. R: Loprieno. N: Castellino. S: Pulci. R. (1989). Micronuclei and
nuclear anomalies induced in the gastro-intestinal epithelium of rats treated with
formaldehyde. Mutagenesis 4: 327-334. http ://dx. doi. or g/10.109 3 /mutage/4.5.327
Miller. FT: Conollv. RB: Kimbell. IS. (2017). An updated analysis of respiratory tract cells at risk for
formaldehyde carcinogenesis. Inhal Toxicol 29: 586-597.
http://dx.doi.Org/10.1080/08958378.2018.1430191
Mivachi. T: Tsutsui. T. (2005). Ability of 13 chemical agents used in dental practice to induce sister-
chromatid exchanges in Syrian hamster embryo cells. Odontology 93: 24-29.
http: //dx.doi.org/10.1007/sl 0266-005-0055-8
Mivake. Y: Tanaka. K: Arakawa. M. (2011). Sibling number and prevalence of allergic disorders in
pregnant Japanese women: baseline data from the Kyushu Okinawa Maternal and Child
Health Study. BMC Public Health 11: 561. http://dx.doi.org/10.1186/1471-2458-ll-561
Moeller. BC: Lu. K: Dovle-Eisele. M: Mcdonald. I: Gigliotti. A: Swenberg. TA. (2011). Determination of
N2-hydroxymethyl-dG adducts in the nasal epithelium and bone marrow of nonhuman
primates following 13CD2-formaldehyde inhalation exposure. Chem Res Toxicol 24: 162-
164. http://dx.doi.org/10.1021/txl004166
Mohner. M: Liu. Y: Marsh. GM. (2019). New insights into the mortality risk from nasopharyngeal
cancer in the national cancer institute formaldehyde worker cohort study. J Occup Med
Toxicol 14: 4. http://dx.doi.org/10.1186/sl 2995-019-0224-2
This document is a draft for review purposes only and does not constitute Agency policy.
R-44 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Monfared. AL. (2012). Histomorphological and ultrastructural changes of the placenta in mice
exposed to formaldehyde. Toxicol Ind Health 30: 174-181.
http://dx.doi.Org/l 0.1177/0748233712452603
Monteiro-Riviere. NA: Popp. TA. (1986). Ultrastructural evaluation of acute nasal toxicity in the rat
respiratory epithelium in response to formaldehyde gas. Fundam Appl Toxicol 6: 251-262.
http://dx.doi.org/10.1016/0272-0590r86190238-l
Monticello. TM: Miller. FT: Morgan. KT. (1991). Regional increases in rat nasal epithelial cell
proliferation following acute and subchronic inhalation of formaldehyde. Toxicol Appl
Pharmacol 111: 409-421. http://dx.doi.org/10.1016/0041-008Xr91190246-B
Monticello. TM: Morgan. KT. (1994). Cell proliferation and formaldehyde-induced respiratory
carcinogenesis [Review], Risk Anal 14: 313-319. http: //dx. doi. o r g /10.1111 / j. 15 3 9-
6924.1994.tb00246.x
Monticello. TM: Morgan. KT. (1997). Chemically-induced nasal carcinogenesis and epithelial cell
proliferation: A brief review [Review], Mutat Res Fundam Mol Mech Mutagen 380: 33-41.
http: //dx.doi.org/10.1016/S0027-5107(97100125-5
Monticello. TM: Morgan. KT: Everitt. II: Popp. TA. (1989). Effects of formaldehyde gas on the
respiratory tract of rhesus monkeys: Pathology and cell proliferation. Am J Pathol 134: 515-
527.
Monticello. TM: Swenberg. TA: Gross. EA: Leininger. TR: Kimbell. IS: Seilkop. S: Starr. TB: Gibson. TE:
Morgan. KT. (1996). Correlation of regional and nonlinear formaldehyde-induced nasal
cancer with proliferating populations of cells. Cancer Res 56: 1012-1022.
Morgan. KT. (1983). Localization of areas of inhibition of nasal mucociliary function in rats
following in vivo exposure to formaldehyde. Am Rev Respir Dis 127: 166.
Morgan. KT: Gross. EA: Patterson. PL. (1986a). Distribution, progression, and recovery of acute
formaldehyde-induced inhibition of nasal mucociliary function in F-344 rats. Toxicol Appl
Pharmacol 86: 448-456. http://dx.doi.org/10.1016/0041 - 008Xf86190372-8
Morgan. KT: Tiang. XZ: Starr. TB: Kerns. WD. (1986b). More precise localization of nasal tumors
associated with chronic exposure of F-344 rats to formaldehyde gas. Toxicol Appl
Pharmacol 82: 264-271. http:/ /dx.doi.org/10.1016/0041 - 008Xr86190201 -2
Morgan. KT: Kimbell. IS: Monticello. TM: Patra. AL: Fleishman. A. (1991). Studies of inspiratory
airflow patterns in the nasal passages of the F344 rat and rhesus monkey using nasal molds:
Relevance to formaldehyde toxicity. Toxicol Appl Pharmacol 110: 223-240.
http://dx.doi.Org/10.1016/S0041-008Xr05180005-5
Morgan. KT: Patterson. PL: Gross. EA. (1984). Frog palate mucociliary apparatus: Structure,
function, and response to formaldehyde gas. Fundam Appl Toxicol 4: 58-68.
http: //dx.doi.org/10.1016/0272-0590r84190219-7
Morgan. KT: Patterson. PL: Gross. EA. (1986c). Responses of the nasal mucociliary apparatus of F-
344 rats to formaldehyde gas. Toxicol Appl Pharmacol 82: 1-13.
http: //dx.doi.org/10.1016/0041 -008Xr86190431 -X
Morgan. KT: Patterson. PL: Gross. EA. (1983). Formaldehyde and the nasal mucociliary apparatus.
In JJ Clary; JE Gibson; RS Waritz (Eds.), Formaldehyde: toxicology, epidemiology,
mechanisms (pp. 193-209). New York, NY: Marcel Pekker, Inc.
This document is a draft for review purposes only and does not constitute Agency policy.
R-45 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Mori. M: Matsumoto. Y: Kushino. N: Morimatsu. Y: Hoshiko. M: Saga. T: Yamaki. K: Ishitake. T.
(2016). Comparison of subjective symptoms associated with exposure to low levels of
formaldehyde between students enrolled and not enrolled in a gross anatomy course.
Environ Health Prev Med 21: 34-41. http: //dx.doi.org/10.1007/sl2199-015-0497-8
Morita. T: Asano. N: Awogi. T: Sasaki. YF: Sato. S: Shimada. H: Sutou. S: Suzuki. T: Wakata. A: Sofuni.
T: Havashi. M. (1997). Evaluation of the rodent micronucleus assay in the screening of IARC
carcinogens (groups 1, 2A and 2B) the summary report of the 6th collaborative study by
CSGMT/JEMS MMS [Review], Mutat Res 389: 3-122. http://dx.doi.org/10.1016/S1383-
5718C96100070-8
Moser. B: Bodrogi. F: Eibl. G: Lechner. M: Rieder. 1: Lirk. P. (2005). Mass spectrometry profile of
exhaled breath - field study by PTR-MS. Respir Physiol Neurobiol 145: 295-300.
http ://dx. doi. or g/10.1016 /i. resp. 2 0 0 4.0 2.0 0 2
Mueller. B: Schulz. G: Mehlin. A: Herzen. 1: Lang. S: Holme. M: Zanette. I: Hieber. S: Devhle. H:
Beckmann. F: Pfeiffer. F: Weitkamp. T. (2012). Grating-based Tomography of Human
Tissues. AIP Conference Proceedings 1466: 107-112. http://dx.doi.Org/10.1063/l.4742277
Mueller. TU: Bruckner. T: Triebig. G. (2013). Exposure study to examine chemosensory effects of
formaldehyde on hyposensitive and hypersensitive males. Int Arch Occup Environ Health
86: 107-117. http://dx.doi.Org/l 0.1007/s00420-012-0745-9
Mullen. NA: Li. 1: Russell. ML: Spears. M: Less. BP: Singer. BC. (2015). Results of the California
Healthy Homes Indoor Air Quality Study of 2011-2013: impact of natural gas appliances on
air pollutant concentrations. Indoor Air 26: 231-245. http://dx.doi.org/10.Ill 1 /ina. 12190
Miiller. W: Engelhart. G: Herbold. B: Tackh. R: Tung. R. (1993). Evaluation of mutagenicity testing
with Salmonella typhimurium TA102 in three different laboratories. Environ Health
Perspect 101: 33-36. http://dx.doi.org/10.1289/ehp.101-1521147
Mundt. KA: Gallagher. AE: Dell. LP: Natelson. EA: Boffetta. P: Gentry. PR. (2017). Does occupational
exposure to formaldehyde cause hematotoxicity and leukemia-specific chromosome
changes in cultured myeloid progenitor cells? [Review], Crit Rev Toxicol 47: 1-11.
http://dx.doi.Org/10.1080/10408444.2017.1301878
Murta. GL: Duarte Campos. KK: Balthar Bandeira. A: Diniz. MF: Costa. G: Costa. DC: Talvani. A: Lima.
WG: Bezerra. FS. (2016). Oxidative effects on lung inflammatory response in rats exposed to
different concentrations of formaldehyde. Environ Pollut 211: 206-213.
http: / /dx. doi. or g/10.1016/i. envpol. 2 015.12.054
Musak. L: Smerhovskv. Z: Halasova. E: Osina. 0: Letkova. L: Vodickova. L: Polakova. V: Buchancova.
1: Hemminki. K: Vodicka. P. (2013). Chromosomal damage among medical staff
occupationally exposed to volatile anesthetics, antineoplastic drugs, and formaldehyde.
Scand J Work Environ Health 39: 618-630. http://dx.doi.org/10.5271/siweh.3358
Nakamura. I: Hollev. DW: Kawamoto. T: Bultman. ST. (2020). The failure of two major formaldehyde
catabolism enzymes (ADH5 and ALDH2) leads to partial synthetic lethality in C57BL/6
mice. Genes and Environ 42: 21. http: //dx.doi.org/10.1186/s41021-020-00160-4
Nalivaiko. E: De Pasquale. CG: Blessing. WW. (2003). Electrocardiographic changes associated with
the nasopharyngeal reflex in conscious rabbits: Vago-sympathetic co-activation. Auton
Neurosci 105: 101-104. http://dx.doi.org/10.1016/sl566-0702(03)00048-1
This document is a draft for review purposes only and does not constitute Agency policy.
R-46 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Nataraian. AT: Darroudi. F: Bussman. CTM: van Kesteren-Van Leeuwen. AC. (1983). Evaluation of the
mutagenicity of formaldehyde in mammalian cytogenetic assays in vivo and vitro. Mutat Res
122: 355-360. http://dx.doi.Org/l 0.1016/0165-7992C83190019-2
Nazarparvar-Noshadi. M: Dolatabadi. TEN: Rasoulzadeh. Y: Mohammadian. Y: Shanehbandi. D.
(2020). Apoptosis and DNA damage induced by silica nanoparticles and formaldehyde in
human lung epithelial cells. Environ Sci Pollut Res Int 27: 18592-18601.
http://dx.doi. org/10.1007/sl 1356-020-08191-8
Neamtiu. IA: Lin. S: Chen. ML: Roba. C: Csobod. E. va: Gurzau. ES. (2019). Assessment of
formaldehyde levels in relation to respiratory and allergic symptoms in children from Alba
County schools, Romania. Environ Monit Assess 191: 591.
http: //dx.doi.oi-g/10.1007/sl 0661 -019-7768-6
NEG. (2003). Formaldehyde. Stockholm, Sweden: National Institute for Working Life.
http://ebib.arbetslivsinstitutetse/ah/2003/ah2003 ll.pdf
Neghab. M: Delikhoon. M: Norouzian Baghani. A: Hassanzadeh. 1. (2017). Exposure to Cooking
Fumes and Acute Reversible Decrement in Lung Functional Capacity. Int J Occup Environ
Med 8: 207-216. http: / /dx. doi. or g/10.15171 /ii oem. 2 017.110 0
Neghab. M: Soltanzadeh. A: Choobineh. A. (2011). Respiratory morbidity induced by occupational
inhalation exposure to formaldehyde. Ind Health 49: 89-94.
http://dx.doi.org/10.2486/indhealth.MS1197
Neuss. S: Moepps. B: Speit. G. (2010a). Exposure of human nasal epithelial cells to formaldehyde
does not lead to DNA damage in lymphocytes after co-cultivation. Mutagenesis 25: 359-364.
http://dx.doi.org/10.1093/mutage/geq013
Neuss. S: Speit. G. (2008). Further characterization of the genotoxicity of formaldehyde in vitro by
the sister chromatid exchange test and co-cultivation experiments. Mutagenesis 23:355-
357. http: / /dx. doi. o r g/10.10 9 3/mutage/genO25
Neuss. S: Zeller. 1: Ma-Hock. L: Speit. G. (2010b). Inhalation of formaldehyde does not induce
ge no toxic effects in broncho-alveolar lavage (BAL) cells of rats. Mutat Res Genet Toxicol
Environ Mutagen 695: 61-68. http://dx.doi.Org/10.1016/i.mrgentox.2009.12.001
Nielsen. GD. (1991). Mechanisms of activation of the sensory irritant receptor by airborne
chemicals [Review], Crit Rev Toxicol 21: 183-208.
http://dx.doi.Org/10.3109/10408449109089879
Nielsen. GD: Hougaard. KS: Larsen. ST: Hammer. M: Wolkoff. P: Clausen. PA: Wilkins. CK: Alarie. Y.
(1999). Acute airway effects of formaldehyde and ozone in BALB/c mice. Hum Exp Toxicol
18: 400-409. http://dx.doi.org/10.1191/096032799678840246
Niinimaa. V: Cole. P: Mintz. S: Shephard. RT. (1981). Oronasal distribution of respiratory airflow.
Respir Physiol 43: 69-75. http://dx.doi.org/10.1016/0034-5687C81190089-X
Nilsson. TA: Hedberg. IT: Vondracek. M: Staab. CA: Hansson. A: Hoog. 10: Grafstrom. RC. (2004).
Alcohol dehydrogenase 3 transcription associates with proliferation of human oral
keratinocytes. Cell Mol Life Sci 61: 610-617. http://dx.doi.org/10.1007/s00018-003-3433-9
Norback. D: Biornsson. E: Tanson. C: Widstrom. 1: Boman. G. (1995). Asthmatic symptoms and
volatile organic compounds, formaldehyde, and carbon dioxide in dwellings. Occup Environ
Med 52: 388-395. http://dx.doi.Org/10.1136/oem.52.6.388
This document is a draft for review purposes only and does not constitute Agency policy.
R-47 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Norback. D: Hashim. TH: Hashim. Z: Ali. F. (2017). Volatile organic compounds (VOC), formaldehyde
and nitrogen dioxide (N02) in schools in Johor Bahru, Malaysia: Associations with rhinitis,
ocular, throat and dermal symptoms, headache and fatigue. Sci Total Environ 592: 153-160.
http: / /dx. doi. or g/10.1016/j. scitotenv. 2017.02.215
Norback. D: Walinder. R: Wieslander. G: Smedie. G: Erwall. C: Venge. P. (2000). Indoor air pollutants
in schools: nasal patency and biomarkers in nasal lavage. Allergy 55: 163-170.
http://dx.doi.Org/10.1034/i.1398-9995.2000.00353.x
Norsted. SW: Kozinetz. CA: Annegers. IF. (1985). Formaldehyde complaint investigations in mobile
homes by the Texas Department of Health. Environ Res 37: 93-100.
http: //dx.doi.org/10.1016/0013-9351C85190052-0
NRC (National Research Council). (2011). Review ofthe Environmental Protection Agency's draft
IRIS assessment of formaldehyde (pp. 1-194). Washington, DC: The National Academies
Press, http://dx.doi.org/10.17226/13142
NRC (National Research Council). (2014). Review of the Formaldehyde Assessment in the National
Toxicology Program 12th Report on Carcinogens. Washington (DC): National Academies
Press (US), http://dx.doi.org/10.17226/18948
NTP (National Toxicology Program). (2010). Final report on carcinogens. Background document for
formaldehyde [NTP] (pp. i-512).
NTP (National Toxicology Program). (2017). NTP research report on absence of formaldehyde-
induced neoplasia in Trp53 haploinsufficient mice exposed by inhalation. (Research Report
3). Research Triangle Park, NC: National Toxicology Program.
https://ntp.niehs.nih.gov/ntp/results/pubs/rr/reports/rr03 508.pdf
O'Connor. PM: Fox. BW. (1987). Comparative studies of DNA cross-linking reactions following
methylene dimethanesulphonate and its hydrolytic product, formaldehyde. Cancer
Chemother Pharmacol 19: 11-15. http://dx.doi.org/10.1007/BF0Q296247
O'Donovan. MR: Mee. CD. (1993). Formaldehyde is a bacterial mutagen in a range of Salmonella and
Escherichia indicator strains. Mutagenesis 8: 577-581.
http://dx.doi.Org/10.1093/mutage/8.6.577
Obe. G: Beek. B. (1979). Mutagenic activity of aldehydes. Drug Alcohol Depend 4: 91-94.
http://dx.d0i.0rg/l 0.1016/0376-8716(79190044-9
Odeigah. PGC. (1997). Sperm head abnormalities and dominant lethal effects of formaldehyde in
albino rats. MutatRes Genet Toxicol Environ Mutagen 389: 141-148.
http://dx.d0i.0rg/l 0.1016/SI 383-5718C96100136-2
Odkvist. LM: Edling. C: Hellquist. H. (1985). Influence of vapours on the nasal mucosa among
industry workers. Rhinology 23: 121-127.
Ohmichi. K: Komivama. M: Matsuno. Y: Sawabe. Y: Mivaso. H: Fukata. H: Ohmichi. M: Kadota. T:
Nomura. F: Moria. C. (2006). Relationship between exposure to formaldehyde and
immunoglobulin E (IgE) production during the gross anatomy laboratory. J Health Sci 52:
642-647. http://dx.d0i.0rg/l 0.1248/ihs.52.642
Ohta. T: Watanabe-Akanuma. M: Tokishita. S: Yamagata. H. (1999). Mutation spectra of chemical
mutagens determined by Lac+ reversion assay with Escherichia coli WP3101P-WP3106P
tester strains. MutatRes Genet Toxicol Environ Mutagen 440: 59-74.
This document is a draft for review purposes only and does not constitute Agency policy.
R-48 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Ohtsuka. R: Shuto. Y: Fuiie. H: Takeda. M: Harada. T: Itagaki. S. (1997). Response of respiratory
epithelium of BN and F344 rats to formaldehyde inhalation. Exp Anim 46: 279-286.
http://dx.doi.org/10.1538/expanim.46.279
Ohtsuka. R: Shutoh. Y: Fujie. H: Yamaguchi. S: Takeda. M: Harada. T: Doi. K. (2003). Rat strain
difference in histology and expression of Thl- and Th2-related cytokines in nasal mucosa
after short-term formaldehyde inhalation. Exp Toxicol Pathol 54: 287-291.
http://dx.doi.Org/l 0.1078/0940-2993-00266
Olin. KL: Cherr. GN: Rifkin. E: Keen. CL. (1996). The effects of some redox-active metals and reactive
aldehydes on DNA-protein cross-links in vitro. Toxicology 110: 1-8.
http://dx.doi. org/10.1016/0300-483XC96103318-5
Olsen. TH: Asnaes. S. (1986). Formaldehyde and the risk of squamous cell carcinoma of the sinonasal
cavities. Br J Ind Med 43: 769-744. http://dx.doi.org/10.1136/oem.43.11.769
Olsen. TH: Dossing. M. (1982). Formaldehyde induced symptoms in day care centers. AIHA J 43:
366-370. http://dx.doi.org/10.1080/15298668291409866
Orsiere. T: Sari-Minodier. I: Iarmarcovai. G: Botta. A. (2006). Genotoxic risk assessment of pathology
and anatomy laboratory workers exposed to formaldehyde by use of personal air sampling
and analysis of DNA damage in peripheral lymphocytes. Mutat Res Genet Toxicol Environ
Mutagen 605: 30-41. http://dx.doi.Org/10.1016/i.mrgentox.2006.01.006
Orstavik. D: Hongslo. IK. (1985). Mutagenicity of endodontic sealers. Biomaterials 6: 129-132.
http://dx.doi.org/10.1016/0142-9612r85190076-6
Overton. TH: Kimbell. IS: Miller. FT. (2001). Dosimetry modeling of inhaled formaldehyde: The
human respiratory tract. Toxicol Sci 64: 122-134.
http: / /dx. doi. or g/10.109 3 /toxsci /64.1.122
Owen. BA: Dudnev. CS: Tan. EL: Easterly. CE. (1990). Formaldehyde in drinking water: Comparative
hazard evaluation and an approach to regulation [Review], Regul Toxicol Pharmacol 11:
220-236. http://dx.doi.Org/10.1016/0273-2300r90190023-5
Ozen. OA: Akpolat. N: Songur. A: Kus. I: Zararsiz. I: Ozacmak. VH: Sarsilmaz. M. (2005). Effect of
formaldehyde inhalation on Hsp70 in seminiferous tubules of rat testes: An
immunohistochemical study. Toxicol Ind Health 21: 249-254.
http://dx.doi.org/10.1191/0748233705th235oa
Ozen. OA: Kus. MA: Kus. I: Alkoc. OA: Songur. A. (2008). Protective effects of melatonin against
formaldehyde-induced oxidative damage and apoptosis in rat testes: An
immunohistochemical and biochemical study. Sys Biol Reprod Med 54: 169-176.
http://dx.doi.Org/10.1080/19396360802422402
Ozen. OA: Songue. A: Sars. M: Yaman. M: Kus. I. (2003). Changes of zinc, copper, and iron levels in
the lung of male rats after subacute (4-week) and subchronic (13-week) exposure to
formaldehyde. J Trace Elem Exp Med 16: 67-74. http://dx.doi.org/10.1002/itra. 10026
Ozen. OA: Yaman. M: Sarsilmaz. M: Songur. A: Kus. I. (2002). Testicular zinc, copper and iron
concentrations in male rats exposed to subacute and subchronic formaldehyde gas
inhalation. J Trace Elem Med Biol 16: 119-122. http://dx.doi.org/10.1016/SQ946-
672Xr02180038-4
Pala. M: Ugolini. D: Ceppi. M: Rizzo. F: Maiorana. L: Bolognesi. C: Schiliro. T: Gilli. G: Bigatti. P: Bono.
R: Vecchio. D. (2008). Occupational exposure to formaldehyde and biological monitoring of
This document is a draft for review purposes only and does not constitute Agency policy.
R-49 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Research Institute workers. Cancer Detect Prev 32: 2008.
http://dx.doi.Org/10.1016/i.cdp.2008.05.003
Palczvnski. C: Krakowiak. A: Hanke. W: Walusiak. T: Gorski. P. (1999). Indoor formaldehyde
exposure and airway allergic diseases. Int Rev Allergol Clin Immunol 5: 65-69.
Park. 1: Yang. H: Song. MK: Kim. D: Lee. K. (2020). Formaldehyde exposure induces regulatory T cell-
mediated immunosuppression via calcineurin-NFAT signalling pathway. SciRep 10: 17023.
http://dx.doi.org/10.1038/s41598-020-72502-9
Parthasarathv. S: Maddalena. RL: Russell. ML: Apte. MG. (2011). Effect of Temperature and
Humidity on Formaldehyde Emissions in Temporary Housing Units. J Air Waste Manag
Assoc 61: 689-695. http: / /dx.doi.org/10.3155/1047-3289.61.6.689
Pauluhn. 1. (1998). Hazard identification and risk assessment of pyrethroids in the indoor
environment Appl Occup Environ Hyg 13: 469-478.
Pavani. S: Mamatha. C: Chandraprakash. C: Bhaskar. M. (2019). Protective role of (Bronco-T) against
formaldehyde induced antioxidant, oxidative and histopathological changes in lung of male
Wistar rats. Toxicology Reports 6: 718-726.
http ://dx. doi. or g/10.1016/i. toxrep. 2019.07.002
Pazdrak. K: Gorski. P: Krakowiak. A: Ruta. U. (1993). Changes in nasal lavage fluid due to
formaldehyde inhalation. Int Arch Occup Environ Health 64: 515-519.
http: //dx.doi.org/10.1007/BF00381101
Pellegrino. R: Viegi. G: Brusasco. V: Crapo. RO: Burgos. F: Casaburi. R: Coates. A: van Per Grinten. CP:
Gustafsson. P: Hankinson. 1: Tensen. R: Tohnson. DC: Macintvre. N: Mckav. R: Miller. MR:
Navaias. D: Pedersen. OF: Wanger. 1. (2005). Interpretative strategies for lung function tests.
Eur Respir J 26: 948-968. http://dx.doi.org/10.1183/09031936.05.000352Q5
Peng. G: Yang. X: Zhao. W: Sun. 1: Cao. Y: Xu. 0: Yuan. I: Ding. S. (2006). Gaseous formaldehyde-
induced DNA-protein crosslinks in liver, kidney and testicle of Kunming mice. Life Science
Journal 3: 82-87.
Perry. P: Wolff. S. (1974). New Giemsa method for the differential staining of sister chromatids.
Nature 251: 156-158.
Persoz. C: Achard. S: Leleu. C: Momas. I: Seta. N. (2010). An in vitro model to evaluate the
inflammatory response after gaseous formaldehyde exposure of lung epithelial cells. Toxicol
Lett 195: 99-105. http ://dx. doi. or g/10.1016/i .to xlet 2 010.0 3.0 0 3
Persoz. C: Achard. S: Momas. I: Seta. N. (2012). Inflammatory response modulation of airway
epithelial cells exposed to formaldehyde. Toxicol Lett 211: 159-163.
http://dx.doi.Org/10.1016/j.toxlet.2012.03.799
Persoz. C: Leleu. C: Achard. S: Fasseu. M: Menotti. I: Meneceur. P: Derouin. F: Seta. N. (2011). In vitro
repeated co-exposure to formaldehyde and Aspergillus fumigatus of human respiratory
cells. Toxicol Lett 205: S171-S171. http ://dx. doi. or g/10.1016/i .to xlet. 2011.05.595
Pesch. B: Pierl. CB: Gebel. M: Gross. I: Becker. D: Tohnen. G: Rihs. HP: Donhuiisen. K: Lepentsiotis. V:
Meier. M: Schulze. I: Briining. T. (2008). Occupational risks for adenocarcinoma of the nasal
cavity and paranasal sinuses in the German wood industry. Occup Environ Med 65: 191-
196. http://dx.doi.org/10.1136/oem.2007.033886
Peteffi. GP: Basso da Silva. L: Antunes. MY: Wilhelm. C: Valandro. ET: Glaeser. I: Kaefer. D: Linden. R.
(2015). Evaluation of genotoxicity in workers exposed to low levels of formaldehyde in a
This document is a draft for review purposes only and does not constitute Agency policy.
R-50 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
furniture manufacturing facility. Toxicol Ind Health 32: 1763-1773.
http://dx.doi.org/10.1177/0748233715584250
Peters. TL: Kamel. F: Lundholm. C: Fevchting. M: Weibull. CE: Sandler. DP: Wiebert. P: Sparen. P: Ye.
W: Fang. F. (2017). Occupational exposures and the risk of amyotrophic lateral sclerosis.
Occup Environ Med 74: 87-92. http://dx.doi.org/10.1136/oemed-2016-103700
Pickrell. TA: Griffis. LC: Mokler. BY: Kanapillv. GM: Hobbs. CH. (1984). Formaldehyde release from
selected consumer products: influence of chamber loading, multiple products, relative
humidity, and temperature. Environ Sci Technol 18: 682-686.
http://dx.doi.Org/10.1021/es00127a009
Pickrell. TA: Mokler. BY: Griffis. LC: Hobbs. CH. (1983). Formaldehyde release rate coefficients from
selected consumer products. Environ Sci Technol 17: 753-757.
http://dx.doi.org/10.1021/es00118a012
Pierce. IS: Abelmann. A: Spicer. LI: Adams. RE: Glynn. ME: Neier. K: Finlev. BL: Gaffnev. SH. (2011).
Characterization of formaldehyde exposure resulting from the use of four professional hair
straightening products. J Occup Environ Hyg 8: 686-699.
http://dx.doi.org/10.1080/15459624.2011.626259
Pinkerton. LE: Hein. Ml: Meyers. A: Kamel. F. (2013). Assessment of ALS mortality in a cohort of
formaldehyde-exposed garment workers. 14: 353-355.
http://dx.doi.org/10.3109/21678421.2013.778284
Pira. E: Romano. C: Verga. F: La Vecchia. C. (2014). Mortality from lymphohematopoietic neoplasms
and other causes in a cohort of laminated plastic workers exposed to formaldehyde. Cancer
Causes Control 25: 1343-1349. http://dx.doi.org/10.1007/sl0552-014-0440-0
Pitten. FA: Kramer. A: Herrmann. K: Bremer. 1: Koch. S. (2000). Formaldehyde neurotoxicity in
animal experiments. Pathol Res Pract 196: 193-198. http://dx.doi.org/10.1016/SQ344-
0338f00180100-4
Plesner. BH: Hansen. K. (1983). Formaldehyde and hexamethylenetetramine in Styles' cell
transformation assay. Carcinogenesis 4: 457-459.
http://dx.doi.Org/10.1093/carcin/4.4.457
Pongsavee. M. (2011). In vitro study of lymphocyte antiproliferation and cytogenetic effect by
occupational formaldehyde exposure. Toxicol Ind Health 27: 719-723.
http://dx.doi.org/10.1177/0748233710395991
Pottern. LM: Heineman. EF: Olsen. TH: Raffn. E: Blair. A. (1992). Multiple myeloma among Danish
women: Employment history and workplace exposures. Cancer Causes Control 3: 427-432.
http://dx.doi.Org/10.1007/BF00051355
Priha. E: Liesivuori. 1: Santa. H: Laatikainen. R. (1996). Reactions of hydrated formaldehyde in nasal
mucus. Chemosphere 32: 1077-1082. http://dx.doi.org/10.1016/0045-6535f96100Q15-X
Priha. E: Pennanen. S: Rantio. T: Uitti. 1: Liesivuori. I. (2004). Exposure to and acute effects of
medium-density fiber board dust J Occup Environ Hyg 1: 738-744.
http://dx.doi.Org/10.1080/15459620490520774
Pross. HF: Day. TH: Clark. RH: Lees. RE. (1987). Immunologic studies of subjects with asthma
exposed to formaldehyde and urea-formaldehyde foam insulation (UFFI) off products. J
Allergy Clin Immunol 79: 797-810. http://dx.doi.org/10.1016/0091-6749f87190213-2
This document is a draft for review purposes only and does not constitute Agency policy.
R-51 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
Pushkina. NN: Gofmekler. VA: Klevtsova. GN. (1968). [Changes in ascorbic and nucleic acid
concentration under the influence of benzene and formaldehyde], Biull Eksp Biol Med 66:
51-53.
Qiao. Y: Li. B: Yang. G: Yao. H: Yang. 1: Liu. D: Yan. Y: Sigsgaard. T: Yang. X. (2009). Irritant and
adjuvant effects of gaseous formaldehyde on the ovalbumin-induced hyperresponsiveness
and inflammation in a rat model. Inhal Toxicol 21: 1200-1207.
http://dx.doi.org/10.3109/08958370902806159
Que. LG: Liu. L: Yan. Y: Whitehead. GS: Gavett. SH: Schwartz. DA: Stamler. IS. (2005). Protection from
experimental asthma by an endogenous bronchodilator. Science 308: 1618-1621.
http: / /dx. doi. or g/10.112 6 /science. 110 82 2 8
Ouievryn. G: Zhitkovich. A. (2000). Loss of DNA-protein crosslinks from formaldehyde-exposed
cells occurs through spontaneous hydrolysis and an active repair process linked to
proteosome function. Carcinogenesis 21: 1573-1580.
http: / /dx. doi. or g/10.109 3 /carcin /21.8.1573
Raaschou-Nielsen. 0: Hermansen. MN: Loland. L: Buchvald. F: Pipper. CB: Sarensen. M: Loft. S:
Bisgaard. H. (2010). Long-term exposure to indoor air pollution and wheezing symptoms in
infants. Indoor Air 20: 159-167. http://dx.doi.Org/10.llll/i.1600-0668.2009.00635.x
Ragan. PL: Boreiko. CI. (1981). Initiation of C3H/10T1/2 cell transformation by formaldehyde.
Cancer Lett 13: 325-331. http://dx.doi.org/10.1016/0304-3835r81190061-6
Rager. IE: Moeller. BC: Dovle-Eisele. M: Kracko. D: Swenberg. TA: Fry. RC. (2013). Formaldehyde and
epigenetic alterations: microRNA changes in the nasal epithelium of nonhuman primates.
Environ Health Perspect 121: 339-344. http://dx.doi.org/10.1289/ehp.1205582
Rager. IE: Moeller. BC: Miller. SK: Kracko. D: Dovle-Eisele. M: Swenberg. TA: Fry. RC. (2014).
Formaldehyde-Associated Changes in microRNAs: Tissue and Temporal Specificity in the
Rat Nose, White Blood Cells, and Bone Marrow. Toxicol Sci 138: 36-46.
http: / /dx. doi. or g/10.109 3 /toxsci /kft2 67
Rager. IE: Smeester. L: Taspers. I: Sexton. KG: Fry. RC. (2011). Epigenetic changes induced by air
toxics: formaldehyde exposure alters miRNA expression profiles in human lung cells.
Environ Health Perspect 119: 494-500. http: //dx.doi.org/10.1289/ehp.1002614
Ratnavake. WE. (1968). Tests for an effect of the Y-chromosome on the mutagenic action of
formaldehyde and x-rays in Drosophila melanogaster. Genet Res 12: 65-69.
http: //dx.doi.org/10.1017/S0016672300011629
Ratnavake. WE. (1970). Studies on the relationship between induced crossing-over and mutation in
Drosophila melanogaster. Mutat Res 9: 71-83. http://dx.doi.org/10.1016/0027-
5107f70190071-0
Ravault. C: Kauffmann. F. (2001). Validity of the IUATLD (1986) questionnaire in the EGEA study.
IntJTuberc LungDis5: 191-196.
Recio. L: Sisk. S: Pluta. L: Bermudez. E: Gross. EA: Chen. Z: Morgan. K: Walker. C. (1992). p53
mutations in formaldehyde-induced nasal squamous cell carcinomas in rats. Cancer Res 52:
6113-6116.
Reiss. R: Ryan. PB: Koutrakis. P: Tibbetts. ST. (1995). Ozone reactive chemistry on interior latex
paint Environ SciTechnol 29: 1906-1912. http://dx.doi.org/10.1021/es00008a007
This document is a draft for review purposes only and does not constitute Agency policy.
R-52 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Ren. X: Ti. Z: Mchale. CM: Yuh. T: Bersonda. T: Tang. M: Smith. MT: Zhang. L. (2013). The impact of
FANCD2 deficiency on formaldehyde-induced toxicity in human lymphoblastoid cell lines.
Arch Toxicol 87: 189-196. http: / /dx.doi.org/10.1007/s00204-012-0911 -6
Reuzel. PGT: Wilmer. TWG. M: Woutersen. RA: Zwart. A: Rombout. PTA: Feron. VI. (1990). Interactive
effects of ozone and formaldehyde on the nasal respiratory lining epithelium in rats. J
Toxicol Environ Health 29: 279-292. http://dx.doi.org/10.1080/15287399009531391
Riedel. F: Hasenauer. E: Barth. PI: Koziorowski. A: Rieger. CHL. (1996). Formaldehyde exposure
enhances inhalative allergic sensitization in the guinea pig. Allergy 51: 94-99.
http://dx.doi.Org/10.llll/i.1398-9995.1996.tb00041.x
Riess. U: Tegtbur. U: Fauck. C: Fuhrmann. F: Markewitz. D: Salthammer. T. (2010). Experimental
setup and analytical methods for the non-invasive determination of volatile organic
compounds, formaldehyde and NOx in exhaled human breath. Anal Chim Acta 669: 53-62.
http://dx.doi.Org/10.1016/i.aca.2010.04.049
Ritchie. IM: Lehnen. RG. (1985). An analysis of formaldehyde concentrations in mobile and
conventional homes. J Environ Health 47: 300-305.
Ritchie. IM: Lehnen. RG. (1987). Formaldehyde-related health complaints of residents living in
mobile and conventional homes. Am J Public Health 77: 323-328.
http ://dx. doi. or g/10.210 5 /ai ph. 77.3.3 2 3
Roberts. DM: Yates. C: Megarbane. B: Winchester. IF: Maclaren. R: Gosselin. S: Nolin. TP: Lavergne.
V: Hoffman. RS: Ghannoum. M: Group. EW. (2015). Recommendations for the role of
extracorporeal treatments in the management of acute methanol poisoning: a systematic
review and consensus statement [Review], Crit Care Med 43: 461-472.
http: //dx.doi.org/10.1097/CCM.0000000000000708
Robinson. CF: Fowler. D: Brown. DP: Lemen. RA. (1987). Plywood mill workers' mortality patterns
1945 1977 (revised March 1987). (NIOSH/00197140). Cincinnati, OH: NIOSH.
Roda. C: Kousignian. I: Guihenneuc-Touyaux. C: Dassonville. C: Nicolis. I: lust. 1: Momas. I. (2011).
Formaldehyde exposure and lower respiratory infections in infants: findings from the
PARIS cohort study. Environ Health Perspect 119: 1653-1658.
http://dx.doi.org/10.1289/ehp.1003222
Roemer. E: Anton. HI: Kindt. R. (1993). Cell proliferation in the respiratory tract of the rat after
acute inhalation of formaldehyde or acrolein. J Appl Toxicol 13: 103-107.
http://dx.doi.Org/10.1002/iat.2550130206
Romanazzi. V: Pirro. V: Bellisario. V: Mengozzi. G: Peluso. M: Pazzi. M: Bugiani. M: Verlato. G: Bono.
R (2013). 15-F2t isoprostane as biomarker of oxidative stress induced by tobacco smoke
and occupational exposure to formaldehyde in workers of plastic laminates. Sci Total
Environ 442: 20-25. http://dx.doi.Org/10.1016/i.scitotenv.2012.10.057
Ross. WE: Mcmillan. PR: Ross. CF. (1981). Comparison of DNA damage by methylmelamines and
formaldehyde. J Natl Cancer Inst 67: 217-221.
Ross. WE: Shipley. N. (1980). Relationship between DNA damage and survival in formaldehyde-
treated mouse cells. MutatRes 79: 277-283. http: / /dx. doi. o r g/10.1016 / 016 5 -
1218(80190075-0
Rothman. KT: Greenland. S. (1998). Modern epidemiology. In Modern Epidemiology (2 ed.).
Philadelphia, PA: Lippincott-Raven.
This document is a draft for review purposes only and does not constitute Agency policy.
R-53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Rothman. N: Lan. 0: Smith. MT: Vermeulen. R: Zhang. L. (2017). Response to letter to the editor of
Carcinogenesis byPiraetal., 2017 [Letter], Carcinogenesis 38: 1253-1255.
http ://dx. doi.org/10.1093 /carcin/bgxl 11
Roush. GC: Walrath. T: Stavner. LT: Kaplan. SA: Flannerv. IT: Blair. A. (1987). Nasopharyngeal
cancer, sinonasal cancer, and occupations related to formaldehyde: A case-control study. J
Natl Cancer Inst 79: 1221-1224.
Rumchev. K: Spickett. 1: Bulsara. M: Phillips. M: Stick. S. (2004). Association of domestic exposure to
volatile organic compounds with asthma in young children. Thorax 59: 746-751.
http ://dx. doi. or g/10.113 6 /thx. 2003.013680
Rumchev. KB: Spickett. TT: Bulsara. MK: Phillips. MR: Stick. SM. (2002). Domestic exposure to
formaldehyde significantly increases the risk of asthma in young children. Eur Respir J 20:
403-408. http://dx.doi.oi-g/10.1183/09031936.02.00245002
Rusch. GM: Clary. 11: Rinehart. WE: Bolte. HF. (1983). A 26-week inhalation toxicity study with
formaldehyde in the monkey, rat, and hamster. Toxicol Appl Pharmacol 68: 329-343.
http: //dx.doi.org/10.1016/0041 - 008Xf83190276-4
Rvden. E: Ekstrom. C: Hellmer. L: Bolcsfoldi. G. (2000). Comparison of the sensitivities of Salmonella
typhimurium strains TA102 and TA2638Ato 16 mutagens. Mutagenesis 15: 495-502.
Saberi Hosniieh. F: Christopher. Y: Peeters. P: Romieu. I: Xun. W: Riboli. E: Raaschou-Nielsen. 0:
Tjanneland. A: Becker. N: Nieters. A: Trichopoulou. A: Bamia. C: Orfanos. P: Oddone. E:
Luian-Barroso. L: Dorronsoro. M: Navarro. C: Barricarte. A: Molina-Montes. E: Wareham. N:
Vineis. P: Vermeulen. R. (2013). Occupation and risk of lymphoid and myeloid leukaemia in
the European Prospective Investigation into Cancer and Nutrition (EPIC). Occup Environ
Med 70: 464-470. http://dx.doi.org/10.1136/oemed-2012-101135
Sadakane. K: Takano. H: Ichinose. T: Yanagisawa. R: Shibamoto. T. (2002). Formaldehyde enhances
mite allergen-induced eosinophilic inflammation in the murine airway. J Environ Pathol
Toxicol Oncol 21: 267-276.
Saillenfait. AM: Bonnet. P: de Ceaurriz. I. (1989). The effects of maternally inhaled formaldehyde on
embryonal and foetal development in rats. Food Chem Toxicol 27: 545-548.
http: //dx.doi.org/10.1016/0278-6915C89190051 -3
Saito. Y: Nishio. K: Yoshida. Y: Niki. E. (2005). Cytotoxic effect of formaldehyde with free radicals via
increment of cellular reactive oxygen species. Toxicology 210: 235-245.
http://dx.doi.Org/10.1016/i.tox.2005.02.006
Sakellaris. I: Saraga. D: Mandin. C: de Kluizenaar. Y: Fossati. S: Spinazze. A: Cattaneo. A: Mihucz. V:
Szigeti. T: de Oliveira Fernandes. E: Kalimeri. K: Mabilia. R: Carrer. P: Bartzis. I. (2020).
Association of subjective health symptoms with indoor air quality in European office
buildings: The OFFICAIR project Indoor Air 31: 426-439.
http: / /dx. doi. or g/10.1111 /ina. 12749
Saladino. AT: Willev. TC: Lechner. IF: Grafstrom. RC: Laveck. M: Harris. CC. (1985). Effects of
formaldehyde, acetaldehyde, benzoyl peroxide, and hydrogen peroxide on cultured normal
human bronchial epithelial cells. Cancer Res 45: 2522-2526.
Salonen. H: Pasanen. AL: Lappalainen. S: Riuttala. H: Tuomi. T: Pasanen. P: Back. B: Reijula. K.
(2009). Volatile organic compounds and formaldehyde as explaining factors for sensory
irritation in office environments. J Occup Environ Hyg 6: 239-247.
http://dx.doi.org/10.1080/15459620902735892
This document is a draft for review purposes only and does not constitute Agency policy.
R-54 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Salthammer. T: Mentese. S: Marutzkv. R. (2010). Formaldehyde in the indoor environment Chem
Rev 110: 2536-2572. http://dx.doi.org/10.1021/cr800399g
Sandel. M: Murphy. TS: Dixon. SL: Adgate. TL: Chew. GL: Dorevitch. S: Tacobs. DE. (2014). A side-by-
side comparison of three allergen sampling methods in settled house dust. J Expo Sci
Environ Epidemiol 24: 650-656. http: //dx.doi.org/10.1038/ies.2014.30
Sandikci. M: Eren. U: Kum. S. (2007a). Effects of formaldehyde and xylene on alpha-naphthyl acetate
esterase positive T-lymphocytes in bronchus associated lymphoid tissue and peripheral
blood in rats. Rev Med Vet 158: 297-301.
Sandikci. M: Eren. U: Kum. S. (2007b). Effects of formaldehyde and xylene on CD4- and CD8-positive
T cells in bronchus-associated lymphoid tissue in rats. Toxicol Ind Health 23: 471-477.
http://dx.doi.org/10.n77/0748233708089025
Sandikci. M: Sevrek. K: Aksit. H: Kose. H. (2009). Inhalation of formaldehyde and xylene induces
apoptotic cell death in the lung tissue. Toxicol Ind Health 25: 455-461.
http://dx.doi.Org/l 0.1177/0748233709106824
Sanghani. PC: Stone. CL: Ray. BP: Pindel. EV: Hurley. TP: Bosron. WF. (2000). Kinetic mechanism of
human glutathione-dependent formaldehyde dehydrogenase. Biochemistry 39: 10720-
10729. http://dx.doi.org/10.1021/bi9929711
Santovito. A: Cervella. P: Pelpero. M. (2014). Chromosomal damage in peripheral blood
lymphocytes from nurses occupationally exposed to chemicals. Hum Exp Toxicol 33: 897-
903. http://dx.doi.org/10.1177/0960327113512338
Santovito. A: Schiliro. T: Castellano. S: Cervella. P: Bigatti. MP: Gilli. G: Bono. R: Pelpero. M. (2011).
Combined analysis of chromosomal aberrations and glutathione S-transferase Ml and T1
polymorphisms in pathologists occupationally exposed to formaldehyde. Arch Toxicol 85:
1295-1302. http://dx.doi.org/10.1007/s00204-011 -0668-3
Saowakon. N: Ngernsoungnern. P: Watcharavitoon. P: Ngernsoungnern. A: Kosanlavit. R. (2015).
Formaldehyde exposure in gross anatomy laboratory of Suranaree University of
Technology: a comparison of area and personal sampling. Environ Sci PollutRes Int 22:
19002-19012. http://dx.doi.org/10.1007/sl 1356-015-5078-2
Sapmaz. E: Sapmaz. HI: Vardi. N: Tas. U: Sarsilmaz. M: Toplu. Y: Arid. A: Uvsal. M. (2017). Harmful
effects of formaldehyde and possible protective effect of Nigella sativa on the trachea of rats.
Niger J Clin Pract 20: 523-529. http://dx.doi.org/10.4103 /1119-3Q77.183253
Sapmaz. HI: Sarsilmaz. M: Godekmerdan. A: Ogetiirk. M: Tas. U: Kose. E. (2015). Effects of
formaldehyde inhalation on humoral immunity and protective effect of Nigella sativa oil: An
experimental study. Toxicol Ind Health 32: 1564-1569.
http://dx.doi.org/10.1177/0748233714566294
Sapmaz. HI: Yildiz. A: Polat. A: Vardi. N: Kose. E: Tanbek. K: Cuglan. S. (2018). Protective efficacy of
Nigella sativa oil against the harmful effects of formaldehyde on rat testicular tissue. 8: 548-
553. http://dx.doi.org/10.4103/2221-1691.24597Q
Sari-Minodier. I: Orsiere. T: Bellon. L: Pompili. 1: Sapin. C: Botta. A. (2002). Cytogenetic monitoring
of industrial radiographers using the micronucleus assay. MutatRes 521: 37-46.
Sari. PK: Kuwahara. S: Furuva. M: Tsukamoto. Y: Hori. H: Kunugita. N: Arashidani. K: Fujimaki. H:
Sasaki. F. (2005). Hypothalamo-pituitary-adrenal gland axis in mice inhaling toluene prior
to low-level long-term exposure to formaldehyde. J Vet Med Sci 67: 303-309.
http: / /dx. doi. or g/10.12 9 2 /ivms. 67.303
This document is a draft for review purposes only and does not constitute Agency policy.
R-55 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Sari. DK: Kuwahara. S: Tsukamoto. Y: Hori. H: Kunugita. N: Arashidani. K: Fuiimaki. H: Sasaki. F.
(2004). Effect of prolonged exposure to low concentrations of formaldehyde on the
corticotropin releasing hormone neurons in the hypothalamus and adrenocorticotropic
hormone cells in the pituitary gland in female mice. Brain Res 1013: 107-116.
http ://dx. doi. or g/10.1016 /i .brainres. 200 4.03.070
Sarigiannis. DA: Karakitsios. SP: Gotti. A: Liakos. IL: Katsoviannis. A. (2011). Exposure to major
volatile organic compounds and carbonyls in European indoor environments and associated
health risk [Review], Environ Int 37: 743-765.
http://dx.doi.Org/10.1016/i.envint.2011.01.005
Sarrif. AM: Krahn. DF: Donovan. SM: O'Neil. RM. (1997). Evaluation of hexamethylphosphoramide
for gene mutations in Salmonella typhimurium using plate incorporation, preincubation,
and suspension assays. MutatRes 380: 167-177. http://dx.doi.org/10.1016/S0Q27-
5107C97100134-6
Sarsilmaz. M: Kaplan. S: Songur. A: Colakoglu. S: Asian. H: Tunc. AT: Ozen. OA: Turgut. M: Bas. 0.
(2007). Effects of postnatal formaldehyde exposure on pyramidal cell number, volume of
cell layer in hippocampus and hemisphere in the rat: A stereological study. Brain Res 11:
157-167. http ://dx. doi. or g/10.1016/i .brainres. 2007.01.139
Sarsilmaz. M: Ozen. OA: Akpolat. N: Kus. I: Songur. A. (1999). The histopathologic effects of inhaled
formaldehyde on leydig cells of the rats in subacute period. Firat Univ Med Sci J 13: 37-40.
Sarto. F: Finotto. S: Giacomelli. L: Mazzotti. D: Tomanin. R: Levis. AG. (1987). The micronucleus
assay in exfoliated cells of the human buccal mucosa. Mutagenesis 2: 11-17.
http ://dx. doi. or g/10.109 3 /mutage/2.1.11
Sauder. LR: Chatham. MP: Green. DT: Kulle. TT. (1986). Acute pulmonary response to formaldehyde
exposure in healthy nonsmokers. J Occup Environ Med 28: 420-424.
http: / /dx.doi.org/10.1097 /00043764-198606000-00008
Sauder. LR: Green. DT: Chatham. MP: Kulle. TT. (1987). Acute pulmonary response of asthmatics to
3.0 ppm formaldehyde. Toxicol Ind Health 3: 569-578.
http://dx.doi.Org/10.1177/074823378700300408
Saurel-Cubizolles. Ml: Hays. M: Estrvn-Behar. M. (1994). Work in operating rooms and pregnancy
outcome among nurses. Int Arch Occup Environ Health 66: 235-241.
http://dx.doi.org/10.1007/bf0Q454361
Sax. SN: Bennett. DH: Chillrud. SN: Kinney. PL: Spengler. ID. (2004). Differences in source emission
rates of volatile organic compounds in inner-city residences of New York City and Los
Angeles. J Expo Anal Environ Epidemiol 14: S95-S109.
http ://dx. doi. or g/10.10 3 8 /si .i ea. 75 0 0 3 64
Schachter. EN: Witek T T. T. R: Tosun. T: Leaderer. BP: Beck. GT. (1986). A STUDY OF RESPIRATORY
EFFECTS FROM EXPOSURE TO 2 PARTS-PER-MILLION FORMALDEHYDE IN HEALTHY
SUBJECTS. Arch Environ Health 41: 229-239.
Schachter. EN: Witek. TT. Tr: Brodv. DT: Tosun. T: Beck. GT: Leaderer. BP. (1987). A study of
respiratory effects from exposure to 2.0 ppm formaldehyde in occupationally exposed
workers. Environ Res 44: 188-205. http://dx.d0i.0rg/l 0.1016/S0013-9351T87180227-X
Schlink. K: Tanssen. K: Nitzsche. S: Gebhard. S: Hengstler. TG: Klein. S: Oesch. F. (1999). Activity of
06-methylguanine DNA methyltransferase in mononuclear blood cells of formaldehyde-
This document is a draft for review purposes only and does not constitute Agency policy.
R-56 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
exposed medical students. Arch Toxicol 73: 15-21.
http://dx.doi.Org/10.1007/s002040050581
Schlosser. PM. (1999). Relative roles of convection and chemical reaction for the disposition of
formaldehyde and ozone in nasal mucus. Inhal Toxicol 11: 967-980.
http://dx.doi.org/10.1080/089583799196736
Schmid. E: Goggelmann. W: Bauchinger. M. (1986). Formaldehyde-induced cytotoxic, genotoxic and
mutagenic response in human lymphocytes and Salmonella typhimurium. Mutagenesis 1:
427-431. http: / /dx.doi. or g/10.109 3/mutage /1.6.427
Schmid. 0: Speit. G. (2007). Genotoxic effects induced by formaldehyde in human blood and
implications for the interpretation of biomonitoring studies. Mutagenesis 22: 69-74.
http: / /dx. doi. or g/10.109 3 /mutage /gelO 5 3
Schreiber. H: Bibbo. M: Wied. GL: Saccomanno. G: Nettesheim. P. (1979). Bronchial metaplasia as a
benign or premalignant lesion. I. Cytologic and ultrastructural discrimination between acute
carcinogen effects and toxin-induced changes. Acta Cytol 23: 496-503.
Schreider. TP. (1986). Comparative anatomy and function of the nasal passages. In CS Barrow (Ed.),
Toxicology of the nasal passages (pp. 1-25). Washington, DC: Hemisphere Publishing
Corporation.
Schroeter. ID: Campbell. 1: Kimbell. IS: Conollv. RB: Clewell. HI: Andersen. ME. (2014). Effects of
endogenous formaldehyde in nasal tissues on inhaled formaldehyde dosimetry predictions
in the rat, monkey, and human nasal passages. Toxicol Sci 138: 412-424.
http: / /dx. doi. or g/10.109 3 /toxsci /kft3 3 3
Schuck. EA: Stephens. ER: Middleton. IT. (1966). Eye irritation response at low concentrations of
irritants. Arch Environ Health 13: 570-575.
Seals. RM: Kioumourtzoglou. MA: Gredal. 0: Hansen. 1: Weisskopf. MG. (2017). Occupational
formaldehyde and amyotrophic lateral sclerosis. Eur J Epidemiol 32: 893-899.
http://dx.doi.Org/l 0.1007/sl 0654-017-0249-8
Seitz. T: Baron. S. (1990). Health hazard evaluation report No. HETA-87-349-2022, Rockcastle
Manufacturing, Mount Vernon, Kentucky (pp. 87-349). (HETA-87-349-2022). Cincinnati,
OH: National Institute of Occupational Safety and Health.
Sellakumar. AR: Snyder. CA: Solomon. IT: Albert. RE. (1985). Carcinogenicity of formaldehyde and
hydrogen chloride in rats. Toxicol Appl Pharmacol 81: 401-406.
http: //dx.doi.org/10.1016/0041-008XC85190411 -9
Senichenkova. II. (1991a). Embryotoxic effects of industrial environment pollutants: Formaldehyde
and gasoline. Gig Sanit-: 35-38.
Senichenkova. II. (1991b). [Embryotoxic effects of industrial environment pollutants: formaldehyde
and gasoline]. Gig Sanit-: 35-38.
Senichenkova. IN: Chebotar. NA. (1996). Effects of gasoline and formaldehyde on prenatal
development of rats with induced iron micronutrient disorder (iron deficiency). Ontogenez
27: 108-113.
Seow. WT: Zhang. L: Vermeulen. R: Tang. X: Hu. W: Bassig. BA: Ti. Z: Shiels. MS: Kemp. TT: Shen. M:
Oiu. C: Reiss. B: Beane Freeman. LE: Blair. A: Kim. C: Guo. W: Wen. C: Li. L: Pinto. LA: Huang.
H: Smith. MT: Hildesheim. A: Rothman. N: Lan. 0. (2015). Circulating immune/inflammation
This document is a draft for review purposes only and does not constitute Agency policy.
R-57 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
markers in Chinese workers occupationally exposed to formaldehyde. Carcinogenesis 36:
852-857. http ://dx.doi.org/10.1093/carcin/bgvO55
Sernia. S: Pi Folco. F: Altrudo. P: Sbriccoli. B: Sestili. C: Colamesta. V: Del Buono. S: Michetti. A: Ortis.
M: Dawodu. A: Villari. P: La Torre. G. (2016). [Risk of nasopharyngeal cancer, Leukemia and
other tumors in a cohort of employees and students potentially exposed to (FA)
formaldehyde in University laboratories], ClinTer 167: 43-47.
http://dx.doi.Org/l 0.7417/CT.2016.1925
Sexton. K: Liu. KS: Petreas. MX. (1986). Formaldehyde concentrations inside private residences: A
mail-out approach to indoor air monitoring. J Air Pollut Control Assoc 36: 698-704.
http://dx.doi.org/10.1080/00022470.1986.10466104
Sexton. K: Petreas. MX: Liu. KS. (1989). Formaldehyde exposures inside mobile homes. Environ Sci
Technol 23: 985-988. http://dx.doi.org/10.1021/es00066a009
Shaham. 1: Bomstein. Y: Gurvich. R: Rashkovskv. M: Kaufman. Z. (2003). DNA-protein crosslinks and
p53 protein expression in relation to occupational exposure to formaldehyde. Occup
Environ Med 60: 403-409. http://dx.doi.Org/10.1136/oem.60.6.403
Shaham. 1: Bomstein. Y: Meltzer. A: Kaufman. Z: Palma. E: Ribak. 1. (1996). DNA-protein crosslinks, a
biomarker of exposure to formaldehyde - in vitro and in vivo studies. Carcinogenesis 17:
121-126. http ://dx. doi. or g/10.109 3 /carcin/17.1.121
Shaham. 1: Bomstein. Y: Melzer. A: Ribak. 1. (1997). DNA-protein crosslinks and sister chromatid
exchanges as biomarkers of exposure to formaldehyde. Int J Occup Environ Health 3: 95-
104. http: / /dx.doi.org/10.1179/107735297800407695
Shaham. 1: Gurvich. R: Kaufman. Z. (2002). Sister chromatid exchange in pathology staff
occupationally exposed to formaldehyde. Mutat Res Genet Toxicol Environ Mutagen 514:
115-123. http://dx.doi.org/10.1016/SI 383-5718f01100334-5
Shangina. 0: Brennan. P: Szeszenia-Dabrowska. N: Mates. D: Fabianova. E: Fletcher. T: T'Mannetie.
A: Boffetta. P: Zaridze. D. (2006). Occupational exposure and laryngeal and hypopharyngeal
cancer risk in central and eastern Europe. Am J Epidemiol 164: 367-375.
http ://dx. doi. or g/10.10 9 3/ai e /kwi 208
She. Y: Li. Y: Liu. Y: Asai. G: Sun. S: He. 1: Pan. Z: Cui. Y. (2013). Formaldehyde induces toxic effects
and regulates the expression of damage response genes in BM-MSCs. Acta Biochim Biophys
Sin45: 1011-1020. http://dx.doi.org/10.1093/abbs/gmtl05
Sheppard. D: Eschenbacher. W: Epstein. 1. (1984). Lack of bronchomotor response to up to 3 ppm
formaldehyde in subjects with asthma. Environ Res 35: 133-139.
http: //dx.doi.org/10.1016/0013-9351 (84190120-8
Sheveleva. G. (1971). Study of the specific effect of formaldehyde on the embryogenesis and
progeny of white rats. Toksikol Nov Prom Khim Veshchestv 12: 78-86.
Shi. L: Lebrun. P: Camu. F: Zizi. M. (2004). Intrathecal catheterization and solvents interfere with
cortical somatosensory evoked potentials used in assessing nociception in awake rats.
Anesth Analg99: 159-165. http:/ /dx.doi.org/10.1213/01 .ane.OOOOl 14552.20268.7f
Shibamoto. T. (2006). Analytical methods for trace levels of reactive carbonyl compounds formed in
lipid peroxidation systems [Review], J Pharm Biomed Anal 41: 12-25.
http://dx.doi.Org/10.1016/i.ipba.2006.01.047
This document is a draft for review purposes only and does not constitute Agency policy.
R-58 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
Sholapuri. P: Chintha. V: Matcha. B: Pradeepkiran. T. (2020). Beneficial effects ofpolyherbal
formulation (Bronco-T) on formaldehyde-induced lung toxicity in male Wistar rats.
Toxicology Research 9: 798-807. http ://dx.doi.org/10.1093 /toxres/tfaaO83
Siboulet. R: Grinfeld. S: Deparis. P: Tavlet. A. (1984). Micronuclei in red blood cells of the newt
Pleurodeles waltl Michah: induction with X-rays and chemicals. Mutat Res 125: 275-281.
http://dx.doi.org/10.1016/0027-5107r84190077-0
Siew. C: Deitrich. RA: Erwin. VG. (1976). Localization and characteristics of rat liver mitochondrial
aldehyde dehydrogenases. Arch Biochem Biophys 176: 638-649.
http://dx.doi.org/10.1016/0003-986ir76190208-3
Siew. SS: Kauppinen. T: Kyvronen. P: Heikkila. P: Pukkala. E. (2012). Occupational exposure to wood
dust and formaldehyde and risk of nasal, nasopharyngeal, and lung cancer among Finnish
men. Cancer Management and Research 4: 223-232.
http://dx.doi.org/10.2147/CMAR.S30684
Silva Ibrahim. B: Miranda da Silva. C: Barioni. ED: Correa-Costa. M: Drewes. CC: Saraiva Camara. NO:
Tavares-De-Lima. W: Poliselli Farskv. SH: Lino-Dos-Santos-Franco. A. (2015). Formaldehyde
inhalation during pregnancy abolishes the development of acute innate inflammation in
offspring. Toxicol Lett 235: 147-154. http://dx.doi.Org/10.1016/i.toxlet2015.04.001
Skrzvdlewska. E. (2003). Toxicological and metabolic consequences of methanol poisoning. Toxicol
Mech Meth 13: 277-293. http://dx.doi.org/l 0.1080/713857189
Slama. R: Ballester. F: Casas. M: Cordier. S: Eggesbo. M: Iniguez. C: Nieuwenhuiisen. M: Philippat. C:
Rev. S: Vandentorren. S: Vriiheid. M. (2014). Epidemiologic tools to study the influence of
environmental factors on fecundity and pregnancy-related outcomes [Review], Epidemiol
Rev 36: 148-164. http://dx.doi.org/10.1093/epirev/mxtO 11
Slater. TF. (1984). Free-radical mechanisms in tissue injury [Review], Biochem J 222: 1-15.
Smedie. G: Norback. D. (2001). Incidence of asthma diagnosis and self-reported allergy in relation to
the school environment: A four-year follow-up study in schoolchildren. Int J Tuberc Lung
Dis 5: 1059-1066.
Smedie. G: Norback. D: Edling. C. (1997). Asthma among secondary schoolchildren in relation to the
school environment Clin Exp Allergy 27: 1270-1278. http://dx.doi.org/10.1046/i. 1365-
2222.1997.1780977.x
Snyder. RD: van Houten. B. (1986). Genotoxicity of formaldehyde and an evaluation of its effects on
the DNA repair process inhuman diploid fibroblasts. Mutat Res DNA Repair 165: 21-30.
http: //dx.doi.org/10.1016/0167-8817r86190005-2
Sobels. FH: van Steenis. H. (1957). Chemical induction of crossing-over in Drosophila males. Nature
179: 29-31. http://dx.doi.org/10.1038/179029a0
Soffritti. M: Tibaldi. E. va: Padovani. M: Hoel. DG: Giuliani. L: Bua. L: Lauriola. M: Falcioni. L:
Manservigi. M: Manservisi. F: Belpoggi. F. (2016). Synergism between sinusoidal-50Hz
magnetic field and formaldehyde in triggering carcinogenic effects in male Sprague-Dawley
rats. Am J Ind Med 59: 509-521. http://dx.doi.org/10.10Q2/aiim.22598
Solet. D: Zoloth. SR: Sullivan. C: Tewett. I: Michaels. DM. (1989). Patterns of mortality in pulp and
paper workers. J Occup Med 31: 627-630.
This document is a draft for review purposes only and does not constitute Agency policy.
R-59 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Song. T: Kang. T: Lin. B: Li. T: Zhu. Y: Du. T: Yang. X: Xi. Z: Li. R. (2017). Mediating role of TRPV1 ion
channels in the co-exposure to PM2.5 and formaldehyde of balb/c mice asthma model. Sci
Rep 7: 11926. http://dx.doi.Org/l0.1038/s41598-017-11833-6
Songur. A: Akpolat. N: Kus. I: Ozen. OA: Zararsiz. I: Sarsilmaz. M. (2003). The effects of the inhaled
formaldehyde during the early postnatal period in the hippocampus of rats: A
morphological and immunohistochemical study. Neurosci Res Commun 33: 168-178.
http: / /dx. doi. or g/10.10 0 2 /nrc. 10093
Songur. A: Ozen. OA: Sarsilmaz. M. (2010). The toxic effects of formaldehyde on the nervous system
[Review], Rev Environ Contam Toxicol 203: 105-118. http://dx.doi.org/10.10Q7/978-l-
4419-1352-4 3
Songur. A: Sarsilmaz. M: Ozen. OA. (2008). The effects of inhaled formaldehyde on oxidant and
antioxidant systems of rat cerebellum during the postnatal development process. Toxicol
Mech Meth 18: 569-574. http://dx.doi.org/10.1080/15376510701555288
Sorg. BA: Bailie. TM: Tschirgi. ML: Li. N: Wu. WR. (2001a). Exposure to repeated low-level
formaldehyde in rats increases basal corticosterone levels and enhances the corticosterone
response to subsequent formaldehyde. Brain Res 898: 314-320.
http://dx.doi.org/10.1016/S0006-8993r01102208-9
Sorg. BA: Davidson. PL: Hochstatter. T: Sylvester. PW. (2002). Repeated cocaine decreases the
avoidance response to a novel aversive stimulus in rats. Psychopharmacology 163: 9-19.
http://dx.doi.Org/10.1007/s00213-002-1133-z
Sorg. BA: Hochstatter. T. (1999). Behavioral sensitization after repeated formaldehyde exposure in
rats. Toxicol Ind Health 15: 346-355. http://dx.doi.org/l 0.1177/074823379901500309
Sorg. BA: Swindell. S: Tschirgi. ML. (2004). Repeated low level formaldehyde exposure produces
enhanced fear conditioning to odor in male, but not female, rats. Brain Res 1008: 11-19.
http: / /dx. doi. or g/10.1016 /i .brainres. 200 4.02.015
Sorg. BA: Tschirgi. ML: Swindell. S: Chen. L: Fang. 1. (2001b). Repeated formaldehyde effects in an
animal model for multiple chemical sensitivity [Review], Ann N Y Acad Sci 933: 57-67.
http://dx.doi.Org/10.llll/i.1749-6632.2001.tb05814.x
Sorg. BA: Willis. TR: Nowatka. TC: Ulibarri. C: See. RE: Westberg. HH. (1996). Proposed animal
neurosensitization model for multiple chemical sensitivity in studies with formalin.
Toxicology 111: 135-145. http://dx.doi.0rg/lO.l016/0300-483x(96)03371-9
Sorg. BA: Willis. TR: See. RE: Hopkins. B: Westberg. HH. (1998). Repeated low-level formaldehyde
exposure produces cross-sensitization to cocaine: Possible relevance to chemical sensitivity
in humans. Neuropsychopharmacology 18: 385-394.
http: / /dx. doi. or g/10.10 3 8 /si. npp. 13 9 516 0
Sorokin. AB: Khenkin. AM: Shilov. AE. (1988). HIGH VALUE OF KINETIC ISOTOPIC EFFECT DURING
HYDROXYLATION OF CYCLOHEXANE CATALYZED BY IRON PORPHYRINS. Kinet Catal 29:
886-886.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/7956713C
3 - 3598
Souza. AD: Devi. R. (2014). Cytokinesis blocked micronucleus assay of peripheral lymphocytes
revealing the genotoxic effect of formaldehyde exposure. Clin Anat27: 308-312.
http: / /dx. doi. or g/10.10 0 2 /ca. 22291
This document is a draft for review purposes only and does not constitute Agency policy.
R-60 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Speit. G: Kiihner. S: Linsenmever. R: Schiitz. P. (2011a). Does formaldehyde induce aneuploidy.
Mutagenesis 26: 805-811. http://dx.doi.org/10.1093/mutage/ger05Q
Speit. G: Ladeira. C: Linsenmever. R: Schiitz. P: Hogel. T. (2012). Re-evaluation of a reported
increased micronucleus frequency in lymphocytes of workers occupationally exposed to
formaldehyde. MutatRes 744: 161-166. http://dx.doi. or g/10.1016/i.mrgentox.2012.02.009
Speit. G: Merk. 0. (2002). Evaluation of mutagenic effects of formaldehyde in vitro: Detection of
crosslinks and mutations in mouse lymphoma cells. Mutagenesis 17: 183-187.
http: / /dx. doi. or g/10.109 3 /mutage /17.3.183
Speit. G: Neuss. S: Schuetz. P: Froehler-Keller. M: Schmid. 0. (2008a). The genotoxic potential of
glutaraldehyde in mammalian cells in vitro in comparison with formaldehyde. Mutat Res
Genet Toxicol Environ Mutagen 649: 146-154.
http: / /dx. doi. or g/10.1016/i. mr gentox. 2007.08.010
Speit. G: Schmid. 0: Frohler-Keller. M: Lang. I: G. T. (2007a). Assessment of local genotoxic effects of
formaldehyde in humans measured by the micronucleus test with exfoliated buccal mucosa
cells. MutatRes GenetToxicol Environ Mutagen 627: 129-135.
http: / /dx. doi. or g/10.1016/i. mr gentox. 2006.10.013
Speit. G: Schmid. 0: Neuss. S: Schiitz. P. (2008b). Genotoxic effects of formaldehyde in the human
lung cell line A549 and in primary human nasal epithelial cells. Environ Mol Mutagen 49:
300-307. http://dx.doi.Org/l 0.1002 /em.20386
Speit. G: Schutz. P: Hogel. 1: Schmid. 0. (2007b). Characterization of the genotoxic potential of
formaldehyde in V79 cells. Mutagenesis 22: 387-394.
http://dx.doi.org/10.1093/mutage/gem031
Speit. G: Schutz. P: Merk. 0. (2000). Induction and repair of formaldehyde-induced DNA-protein
crosslinks in repair-deficient human cell lines. Mutagenesis 15: 85-90.
http: / /dx. doi. or g/10.109 3 /mutage /15.1.85
Speit. G: Schutz. P: Weber. I: Ma-Hock. L: Kaufmann. W: Gelbke. HP: Durrer. S. (2011b). Analysis of
micronuclei, histopathological changes and cell proliferation in nasal epithelium cells of rats
after exposure to formaldehyde by inhalation. MutatRes 721: 127-135.
http: / /dx. doi. or g/10.1016/i. mr gentox. 2011.01.008
Speit. G: Zeller. 1: Schmid. 0: Elhajouji. A: Ma-Hock. L: Neuss. S. (2009). Inhalation of formaldehyde
does not induce systemic genotoxic effects in rats. Mutat Res Genet Toxicol Environ
Mutagen 677: 76-85. http://dx.doi.Org/10.1016/i.mrgentox.2009.05.020
Sram. RT. (1970). The effect of storage on the frequency of dominant lethals in Drosophila
melanogaster. MGG Mol gen genet 106: 286-288.
Starr. TB: Swenberg. TA. (2016). The bottom-up approach to bounding potential low-dose cancer
risks from formaldehyde: An update. Regul Toxicol Pharmacol 77: 167-174.
http: / /dx. doi. or g/10.1016 /i .vrtph. 2016.01.021
Steele. LL: Wilkins. 1. R. (1996). Occupational exposures and risks of spontaneous abortion among
female veterinarians. Int J Occup Environ Health 2: 26-36.
Stellman. SD: Demers. PA: Colin. D: Boffetta. P. (1998). Cancer mortality and wood dust exposure
among participants in the American Cancer Society Cancer Prevention Study-II (CPS-II). Am
J Ind Med 34: 229-237. http://dx.doi.org/10.1002/fSICniQ97-
0274C199809134:3<229::ATD-ATTM4>3.0.CO:2-O
This document is a draft for review purposes only and does not constitute Agency policy.
R-61 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Stich. HF: Curtis. TR: Parida. BB. (1982). Application of the micronucleus test to exfoliated cells of
high cancer risk groups: Tobacco chewers. Int J Cancer 30: 553-559.
http://dx.doi.Org/10.1002/iic.2910300504
Stock. TH. (1987). Formaldehyde concentrations inside conventional housing. J Air Waste Manag
Assoc 37: 913-918. http://dx.doi.org/10.1080/08940630.1987.10466284
Stroup. NE: Blair. A: Erikson. GE. (1986). Brain cancer and other causes of death in anatomists. J
Natl Cancer Inst 77: 1217-1224.
Stumm-Tegethoff. BFA. (1969). Formaldehyde-induced mutations in Drosophila melanogaster in
dependence of the presence of acids. Theor Appl Genet 39: 330-334.
http://dx.doi.Org/T0.1007/BF00281915
Subramaniam. RP: Chen. C: Crump. KS: Devonev. D: Fox. IF: Portier. CI: Schlosser. PM: Thompson.
CM: White. P. (2008). Uncertainties in biologically-based modeling of formaldehyde-induced
respiratory cancer risk: Identification of key issues. Risk Anal 28: 907-923.
http://dx.doi. org/10.1111 /i.l 539-6924.2008.01083.x
Subramaniam. RP: Richardson. RB: Morgan. KT: Kimbell. IS: Guilmette. RA. (1998). Computational
fluid dynamics simulations of inspiratory airflow in the human nose and nasopharynx. Inhal
Toxicol 10: 91-120. http://dx.doi.org/10.1080/089583798197772
Sul. D: Kim. H: Oh. E: Phark. S: Cho. E: Choi. S: Kang. HS: Kim. EM: Hwang. KW: Tung. WW. (2007).
Gene expression profiling in lung tissues from rats exposed to formaldehyde. Arch Toxicol
81: 589-597. http://dx.doi.Org/10.1007/s00204-007-0182-9
Suruda. A: Schulte. P: Boeniger. M: Hayes. RB: Livingston. GK: Steenland. K: Stewart. P: Herrick. R:
Douthit. D: Fingerhut. MA. (1993). Cytogenetic effects of formaldehyde exposure in students
of mortuary science. Cancer Epidemiol Biomarkers Prev 2: 453-460.
Suskov. II: Sazonova. LA. (1982). Cytogenetic effects of epoxy, phenolformaldehyde and
polyvinylchloride resins in man. MutatRes 104: 137-140. httD://dx.doi.org/10.1016/0165-
7992T82190134-8
Sutton. HC: Downes. TM. (1972). Rate of hydration of formaldehyde in aqueous solution. J Chem Soc
Chem Communl. http://dx.doi.org/10.1039/C39720000Q01
Svensson. S: Some. M: Lundsio. A: Helander. A: Cronholm. T: Hoog. 10. (1999). Activities of human
alcohol dehydrogenases in the metabolic pathways of ethanol and serotonin. Eur J Biochem
262: 324-329. http://dx.doi.Org/10.1046/i.1432-1327.1999.00351.x
Swenberg. 1: Kerns. W: Pavkov. K: Mitchell. R: Gralla. EI. (1980a). Carcinogenicity of formaldehyde
vapor: interim findings in a long-term bioassay of rats and mice. Dev Toxicol Environ Sci 8:
283-286.
Swenberg. TA: Gross. EA: Martin. 1: Popp. TA. (1983a). Mechanisms of formaldehyde toxicity. In JE
Gibson (Ed.), Formaldehyde toxicity (pp. 132-147). Washington, DC: Hemisphere
Publishing.
Swenberg. TA: Gross. EA: Randall. HW. (1986). Localization and quantitation of cell proliferation
following exposure to nasal irritants. In CS Barrow (Ed.), Toxicology of the nasal passages
(pp. 291-300). New York, NY: Hemisphere Publishing Corp.
Swenberg. TA: Gross. EA: Randall. HW: Barrow. CS. (1983b). The effect of formaldehyde exposure on
cytotoxicity and cell proliferation. In JJ Clary; JE Gibson; RS Waritz (Eds.), Formaldehyde,
toxicology, epidemiology, mechanisms (pp. 225-236). New York, NY: Marcel Dekker.
This document is a draft for review purposes only and does not constitute Agency policy.
R-62 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Swenberg. TA: Kerns. WD: Mitchell. RI: Gralla. ET: Pavkov. KL. (1980b). Induction of squamous cell
carcinomas of the rat nasal cavity by inhalation exposure to formaldehyde vapor. Cancer
Res 40: 3398-3402.
Swenberg. TA: Lu. K: Moeller. BC: Gao. L: Upton. PB: Nakamura. 1: Starr. TB. (2011). Endogenous
versus exogenous DNA adducts: Their role in carcinogenesis, epidemiology, and risk
assessment [Review], Toxicol Sci 120: S130-S145. http://dx.doi.org/10.1093/toxsci/kfq371
Swiecichowski. AL: Long. KT: Miller. ML: Leikauf. GD. (1993). Formaldehyde-induced airway
hyperreactivity in vivo and ex vivo in guinea pigs. Environ Res 61: 185-199.
http://dx.doi.org/10.1006/enrs.1993.1063
Taffet. GE: Donohue. IF: Altman. PR. (2014). Considerations for managing chronic obstructive
pulmonary disease in the elderly [Review], Clinical Interventions in Aging 9: 23-30.
http://dx.doi.org/10.2147/CIA.S52999
Takahashi. K: Morita. T: Kawazoe. Y. (1985). Mutagenic characteristics of formaldehyde on bacterial
systems. MutatRes 156: 153-161. http://dx.doi.org/10.1016/0165-1218C85190058-8
Takahashi. S: Tsuji. K: Fujii. K: Okazaki. F: Takigawa. T: Ohtsuka. A: Iwatsuki. K. (2007). Prospective
study of clinical symptoms and skin test reactions in medical students exposed to
formaldehyde gas. J Dermatol 34: 283-289. http://dx.doi.org/10.1111 /i. 1346-
8138.2007.00274.x
Takigawa. T: Usami. M: Yamasaki. Y: Wang. B: Sakano. N: Horike. T: Kataoka. H: Ohtsuka. A: Kira. S.
(2005). Reduction of indoor formaldehyde concentrations and subjective symptoms in a
gross anatomy laboratory. Bull Environ Contam Toxicol 74: 1027-1033.
http://dx.doi.Org/10.1007/s00128-005-0683-2
Talibov. M: Lehtinen-lacks. S: Martinsen. II: Kiaerheim. K: Lvnge. E: Sparen. P: Trvggvadottir. L:
Weiderpass. E: Kauppinen. T: Kyvronen. P: Pukkala. E. (2014). Occupational exposure to
solvents and acute myeloid leukemia: A population-based, case-control study in four Nordic
countries. Scand J Work Environ Health 40: 511-517.
http: / /dx. doi. or g/10.5 2 71 /si weh. 3436
Tani. T: Kogi. K: Horiguchi. Y. (1986). Inhibitory effects of formaldehyde inhalation on the
cardiovascular and respiratory systems in unanesthetized rabbits. Jpn J Pharmacol 40: 551-
559. http://dx.doi.org/10.1254/jjp.40.551
Tarkowski. M: Gorski. P. (1995). Increased IgE antiovalbumin level in mice exposed to
formaldehyde. Int Arch Allergy Immunol 106: 422-424.
http: //dx.doi.org/10.1159/000236876
Taskinen. H: Kyvronen. P: Hemminki. K. (1994). Laboratory work and pregnancy outcome. J Occup
Med 36: 311-319. http://dx.doi.Org/10.1097/00043764-199403000-00008
Taskinen. HK: Kyvronen. P: Sallmen. M: Virtanen. SV: Liukkonen. TA: Huida. 0: Lindbohm. ML:
Anttila. A. (1999). Reduced fertility among female wood workers exposed to formaldehyde.
Am J Ind Med 36: 206-212. http://dx.doi.org/10.10Q2/fsici) 1097-
0274C199907136:1 <206: :aid-aiim29>3.0.co:2-d
Tavernier. G: Fletcher. G: Gee. I: Watson. A: Blacklock. G: Francis. H: Fletcher. A: Frank. T: Frank. P:
Pickering. CA: Niven. R. (2006). IPEADAM study: Indoor endotoxin exposure, family status,
and some housing characteristics in English children. J Allergy Clin Immunol 117: 656-662.
http://dx.doi.Org/10.1016/i.iaci.2005.12.1311
This document is a draft for review purposes only and does not constitute Agency policy.
R-63 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Teixeira. TP: Gaspar. T: Silva. S: Torres. T: Silva. SN: Azevedo. MC: Neves. P: Laffon. B: Mendez. T:
Goncalves. C: Mayan. 0: Farmer. PB: Rueff. T. (2004). Occupational exposure to styrene:
modulation of cytogenetic damage and levels of urinary metabolites of styrene by
polymorphisms in genes CYP2E1, EPHX1, GSTM1, GSTT1 and GSTP1. Toxicology 195: 231-
242. http://dx.doi.Org/10.1016/i.tox.2003.10.010
Temcharoen. P: Thillv. WG. (1983). Toxic and mutagenic effects of formaldehyde in Salmonella
typhimurium. MutatRes 119: 89-93. http://dx.doi.org/10.1016/0165-7992r83190115-X
Teng. S: Beard. K: Pourahmad. 1: Moridani. M: Easson. E: Poon. R: O'Brien. PI. (2001). The
formaldehyde metabolic detoxification enzyme systems and molecular cytotoxic
mechanism in isolated rat hepatocytes. Chem Biol Interact 130-132: 285-296.
http://dx.doi.org/10.1016/S0009-2797r00100272-6
Tepper. IS: Moser. VC: Costa. PL: Mason. MA: Roache. N: Guo. Z: Dyer. RS. (1995). Toxicological and
chemical evaluation of emissions from carpet samples. Am Ind Hyg Assoc J 56: 158-170.
http://dx.doi. org/10.1080/15428119591017196
Teschke. K: Morgan. MS: Checkowav. H: Franklin. G: Spinelli. IT: van Belle. G: Weiss. NS. (1997).
Surveillance of nasal and bladder cancer to locate sources of exposure to occupational
carcinogens. Occup Environ Med 54: 443-451. http://dx.doi.Org/10.1136/oem.54.6.443
Thetkathuek. A: Yingratanasuk. T: Ekburanawat. W. (2016). Respiratory Symptoms due to
Occupational Exposure to Formaldehyde and MDF Dust in a MDF Furniture Factory in
Eastern Thailand. 2016: 3705824. http://dx.doi.org/10.1155/2016/3705824
Thompson. CM: Ceder. R: Grafstrom. RC. (2010). Formaldehyde dehydrogenase: beyond phase I
metabolism. Toxicol Lett 193: 1-3. http://dx.doi.Org/10.1016/i.toxlet2009.ll.023
Thompson. CM: Sonawane. B: Grafstrom. RC. (2009). The ontogeny, distribution, and regulation of
alcohol dehydrogenase 3: Implications for pulmonary physiology [Review], Drug Metab
Dispos 37: 1565-1571. http://dx.doi.org/10.1124/dmd.l 09.027904
Thomson. El: Shackleton. S: Harrington. TM. (1984). Chromosome aberrations and sister-chromatid
exchange frequencies in pathology staff occupationally exposed to formaldehyde. Mutat Res
141: 89-93. http://dx.doi.org/10.1016/0165-7992f84190016-2
Thrasher. ID: Broughton. A: Madison. R. (1990). Immune activation and autoantibodies in humans
with long-term inhalation exposure to formaldehyde. Arch Environ Health 45: 217-223.
http://dx.doi.Org/10.1080/00039896.1990.9940805
Thrasher. ID: Woidani. A: Cheung. G: Heuser. G. (1987). Evidence for formaldehyde antibodies and
altered cellular immunity in subjects exposed to formaldehyde in mobile homes. Arch
Environ Health 42: 347-350. http: //dx.doi.org/10.1080 /00039896.1987.9934357
Tibbetts. AS: Appling. DR. (2010). Compartmentalization of Mammalian folate-mediated one-carbon
metabolism [Review], Annu Rev Nutr 30: 57-81.
http://dx.doi.org/10.1146/annurev.nutr.012809.10481Q
Titenko-Holland. N: Levine. AT: Smith. MT: Ouintana. PI: Boeniger. M: Hayes. R: Suruda. A: Schulte. P.
(1996). Quantification of epithelial cell micronuclei by fluorescence in situ hybridization
(FISH) in mortuary science students exposed to formaldehyde. MutatRes 371: 237-248.
http: //dx.doi.org/10.1016/S0165-1218(96190112-3
Tolbert. PE: Shy. CM: Allen. TW. (1992). Micronuclei and other nuclear anomalies in buccal smears:
methods development. MutatRes 271: 69-77.
This document is a draft for review purposes only and does not constitute Agency policy.
R-64 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Tong. ZM: Zhu. SX: Shi. T. (2007). [Effect of formaldehyde on blood component and blood
biochemistry of exposed workers], 20: 409-410.
Tsubone. H: Kawata. M. (1991). Stimulation to the trigeminal afferent nerve of the nose by
formaldehyde, acrolein, and acetaldehyde gases. Inhal Toxicol 3: 211-222.
http://dx.doi.org/10.3109/08958379109145285
Tsukahara. S: Yamamoto. S: Shwe. TTW: Ahmed. S: Kunugita. N: Arashidani. K: Fuiimaki. H. (2006).
Inhalation of low-level formaldehyde increases the Bcl-2/Bax expression ratio in the
hippocampus of immunologically sensitized mice. Neuroimmunomodulation 13: 63-68.
http://dx.doi.Org/10.1159/000094829
Turner. C: Parekh. B: Walton. C: Spanel. P: Smith. D: Evans. M. (2008). An exploratory comparative
study of volatile compounds in exhaled breath and emitted by skin using selected ion flow
tube mass spectrometry. Rapid Commun Mass Spectrom 22: 526-532.
http ://dx. doi. or g/10.10 0 2 /rem. 3402
U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry [EPA Report],
(EPA/600/8-90/066F). Research Triangle Park, NC.
https://cfpub.epa. gov/ncea/risk/recordisplay.cfm?deid=71993&CFID=51174829&CFTOKE
N=25006317
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report], (EPA/630/P-03/001F). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA. (2010a). Toxicological Review of Formaldehyde - Inhalation Assessment (CAS No. 50-00-
0). In Support of Summary Information on the Integrated Risk Information System (IRIS).
External Review Draft
U.S. EPA (U.S. Environmental Protection Agency). (2010b). Toxicological Review of Formaldehyde
(Inhalation) (External Review Draft 2010).
http://cfpub.epa.gov/ncea/iris drafts/recordisplav.cfm?deid=223614
Uba. G: Pachorek. D: Bernstein. 1: Garabrant. DH: Balmes. TR: Wright. WE: Amar. RB. (1989).
Prospective study of respiratory effects of formaldehyde among healthy and asthmatic
medical students. Am J Ind Med 15: 91-101. http://dx.doi.org/10.1002/aiim.4700150110
Ulsamer. AG: Gupta. KC: Cohn. MS: Preuss. PW. (1982). Formaldehyde in indoor air: Toxicity and
risk. In Proceedings of the 75th Annual Meeting of the Air Pollution Control Association.
Uotila. L: Koivusalo. M. (1974). Formaldehyde dehydrogenase from human liver: Purification,
properties, and evidence for the formation of glutathione thiol esters by the enzyme. J Biol
Chem 249: 7653-7663.
Uotila. L: Koivusalo. M. (1987). Multiple forms of formaldehyde dehydrogenase from human red
blood cells. Hum Hered 37: 102-106. http://dx.doi.org/10.1159/000153684
Uotila. L: Koivusalo. M. (1989). Glutathione-dependentoxidoreductases: Formaldehyde
dehydrogenase. In D Dolphin; R Poulson; 0 Avramovic (Eds.), Glutathione: Chemical,
biochemical, and medical aspects (pp. 517-551). New York, NY: Wiley-Interscience.
Usanmaz. SE: Akarsu. ES: Vural. N. (2002). Neurotoxic effects of acute and subacute formaldehyde
exposures in mice. Environ Toxicol Pharmacol 11: 93-100.
http: //dx.doi.org/10.1016/SI 382-6689C01100109-0
This document is a draft for review purposes only and does not constitute Agency policy.
R-65 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Vargova. M: Tanota. S: Karelova. T: Barancokova. M: Sulcova. M. (1992). Analysis of the health risk of
occupational exposure to formaldehyde using biological markers. Analusis 20: 451-454.
Vasudeva. N: Anand. C. (1996). Cytogenetic evaluation of medical students exposed to
formaldehyde vapor in the gross anatomy dissection laboratory. J Am Coll Health 44: 177-
179. http://dx.doi.org/10.1080/07448481.1996.9937526
Vaughan. TL: Stewart. PA: Teschke. K: Lynch. CF: Swanson. GM: Lyon. TL: Berwick. M. (2000).
Occupational exposure to formaldehyde and wood dust and nasopharyngeal carcinoma.
Occup Environ Med 57: 376-384. http://dx.doi.Org/10.1136/oem.57.6.376
Venn. A: Lewis. S: Cooper. M: Hubbard. R: Hill. I: Boddv. R: Bell. M: Britton. 1. (2000). Local road
traffic activity and the prevalence, severity and persistence of wheeze in school children:
combined cross sectional and longitudinal study. Occup Environ Med 57: 152-158.
http://dx.doi.Org/10.1136/oem.57.3.152
Venn. AT: Cooper. M: Antoniak. M: Laughlin. C: Britton. 1: Lewis. SA. (2003). Effects of volatile
organic compounds, damp, and other environmental exposures in the home on wheezing
illness in children. Thorax 58: 955-960. http://dx.doi. or g/10.1136/thorax.58.11.955
Viegas. S: Ladeira. C: Gomes. M: Nunes. C: Brito. M: Prista. 1. (2013). Exposure and genotoxicity
assessment methodologies - the case of formaldehyde occupational exposure. Current
Analytical Chemistry 9: 476-484. http://dx.d0i.0rg/l0.2174/1573411011309030017
Viegas. S: Ladeira. C: Nunes. C: Malta-Vacas. 1: Gomes. M: Brito. M: Mendonca. P: Prista. 1. (2010).
Genotoxic effects in occupational exposure to formaldehyde: A study in anatomy and
pathology laboratories and formaldehyde-resins production. J Occup Med Toxicol 5: 25.
http://dx.d0i.0rg/l 0.1186/1745-6673-5-25
Vitoux. MA: Kessal. K: Baudouin. C: Laprevote. 0: Melik Parsadaniantz. S: Achard. S: Brignole-
Baudouin. F. (2018). Formaldehyde gas exposure increases inflammation in an in vitro
model of dry eye. Toxicol Sci 165: 108-117. http: //dx.doi.org/10.1093 /toxsci/kfvl25
Vock. EH: Lutz. WK: Ilinskava. 0: Vamvakas. S. (1999). Discrimination between genotoxicity and
cytotoxicity for the induction of DNA double-strand breaks in cells treated with aldehydes
and diepoxides. MutatRes 441: 85-93. http://dx.doi.org/10.1016/S1383-5718f99100038-8
Von Hippel. PH: Wong. KY. (1971). Dynamic aspects of native DNA structure: Kinetics of the
formaldehyde reaction with calf thymus DNA. J Mol Biol 61: 587-613.
http://dx.doi.Org/10.1016/0022-2836f71190066-0
von Kobvletzki. LB: Berner. A: Carlstedt. F: Hasselgren. M: Bornehag. CG: Svensson. A. (2013).
Validation of a parental questionnaire to identify atopic dermatitis in a population-based
sample of children up to 2 years of age. Dermatology 226: 222-226.
http://dx.doi.org/10.1159/000349983
Vosoughi. S: Khavanin. A: Salehnia. M: Asilian Mahabadi. H: Shahverdi. A: Esmaeili. V. (2013).
Adverse effects of formaldehyde vapor on mouse sperm parameters and testicular tissue.
Int J Fertility Sterility 6: 250-257.
Walker. IF. (1975). Formaldehyde (3rd ed.). RE. Krieger Publishing Company: Huntington, NY.
Walrath. 1: Fraumeni. IF. Tr. (1983). Mortality patterns among embalmers. Int J Cancer 31: 407-411.
http://dx.doi.Org/10.1002/ijc.2910310403
Walrath. 1: Fraumeni. IF. Tr. (1984). Cancer and other causes of death among embalmers. Cancer Res
44: 4638-4641.
This document is a draft for review purposes only and does not constitute Agency policy.
R-66 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Wang. B: Liu. DP. (2006). [Detection of formaldehyde induced developmental toxicity assessed with
single cell gel electrophoresis]. Fen Zi Xi Bao Sheng Wu Xue Bao 39: 462-466.
Wang. F. an: Li. C: Liu. W. ei: Tin. Y. (2014). Potential mechanisms of neurobehavioral disturbances
in mice caused by sub-chronic exposure to low-dose VOCs. Inhal Toxicol 26: 250-258.
http://dx.doi.org/10.3109/08958378.2014.882447
Wang. H: Li. H. eC: Lv. M: Zhou. D: Bai. L: Du. L: Xue. X. ia: Lin. P. u: Oiu. S. (2015). Associations
between occupation exposure to Formaldehyde and semen quality, a primary study. Sci Rep
5: 15874. http: / /dx.doi.org/10.1038/srep 15874
Wang. HX: Zhou. DX: Zheng. LR: Zhang. 1: Huo. YW: Tian. H: Han. SP: Zhang. 1: Zhao. WB. (2012).
Effects of paternal occupation exposure to formaldehyde on reproductive outcomes. J Occup
Environ Med 54: 518-524. http://dx.d0i.0rg/l 0.1097/TOM.QbOl 3e31824e6937
Wang. K: Wang. TW: Xu. 1: Zhu. Y: Tian. L: Au. W: Xia. ZL. (2019). Determination of benchmark dose
based on adductand micronucleus formations in formaldehyde-exposed workers. IntJ Hyg
Environ Health 222: 738-743. http://dx.doi.Org/10.1016/i.iiheh.2019.05.008
Wang. M: Cheng. G: Balbo. S: Carmella. SG: Villalta. PW: Hecht. SS. (2009). Clear differences in levels
of a formaldehyde-DNA adduct in leukocytes of smokers and nonsmokers. Cancer Res 69:
7170-7174. http://dx.d0i.0rg/l0.1158/0008-5472.CAN-09-1571
Wang. M: Cheng. G: Villalta. PW: SS. H. (2007). Development of liquid chromatography electrospray
ionization tandem mass spectrometry methods for analysis of DNA adducts of
formaldehyde and their application to rats treated with N-nitrosodimethylamine or 4-
(methylnitrosamino)-l-(3-pyridyl)-l-butanone. Chem Res Toxicol 20: 1141-1148.
http://dx.doi.org/10.1021/tx700189c
Wang. T: Pvsanenko. A: Drvahina. K: Spanel. P: Smith. D. (2008). Analysis of breath, exhaled via the
mouth and nose, and the air in the oral cavity. J Breath Res 2: 1-13.
http://dx.d0i.0rg/l 0.1088/1752-7155/2/3/037013
Wangenheim. 1: Bolcsfoldi. G. (1988). Mouse lymphoma L5178Y thymidine kinase locus assay of 50
compounds. Mutagenesis 3: 193-205. http: //dx.doi.org/10.1093 /mutage/3.3.193
Wantke. F: Demmer. CM: Tappler. P: Gotz. M: Tarisch. R. (1996a). Exposure to gaseous formaldehyde
induces IgE-mediated sensitization to formaldehyde in school-children. Clin Exp Allergy 26:
276-280. http://dx.doi.Org/10.llll /i.1365-2222.1996.tb00092.x
Wantke. F: Focke. M: Hemmer. W: Bracun. R: Wolf-Abdolvahab. S: Gotz. M: Tarisch. R: Gotz. M:
Tschabitscher. M: Gann. M: Tappler. P. (2000). Exposure to formaldehyde and phenol during
an anatomy dissecting course: Sensitizing potency of formaldehyde in medical students.
Allergy 55: 84-87. http://dx.doi.Org/10.1034/i.1398-9995.2000.00307.x
Wantke. F: Focke. M: Hemmer. W: Tschabitscher. M: Gann. M: Tappler. P: Gotz. M: Tarisch. R.
(1996b). Formaldehyde and phenol exposure during an anatomy dissection course: A
possible source of IgE-mediated sensitization. Allergy 51: 837-841.
http://dx.d0i.0rg/l 0.1111 /\.\398-9995.1996.tb00031 ,x
Ward. BT. Tr: Legator. MS: Pereira. MA: Chang. LW. (1983). Evaluation in man and animals of tests
for the detection of population exposures to genotoxic chemicals. In MD Waters; SS Sandhu;
J Lewtas (Eds.), Short-term bioassays in the analysis of complex environmental mixtures
(pp. 461-484). New York, NY: Plenum Press.
Watanabe. K: Sakamoto. K: Sasaki. T. (1996). Comparisons on chemically-induced mutagenicity
among four bacterial strains, Salmonella typhimurium TA102 and TA2638, and Escherichia
This document is a draft for review purposes only and does not constitute Agency policy.
R-67 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Supplemental Information for Formaldehyde—Inhalation
coli WP2/pKM101 and WP2 uvrA/pKMlOl: Collaborative study I. MutatRes 361: 143-155.
http://dx.doi.org/10.1016/s0165-116ir96190249-6
Weber-Tschopp. A: Fischer. T: Grandiean. E. (1977). Reizwirkungen des Formaldehyds (HCHO) auf
den Menschen [Irritating effects of formaldehyde on men], Int Arch Occup Environ Health
39: 207-218. http://dx.doi.org/10.1007/bf00409367
Wei. C: Chen. M: You. H: Oiu. F: Wen. H: Yuan. 1: Xiang. S: Yang. X. (2017a). Formaldehyde and co-
exposure with benzene induce compensation of bone marrow and hematopoietic
stem/progenitor cells in BALB/c mice during post-exposure period. Toxicol Appl Pharmacol
324: 36-44. http://dx.doi.Org/10.1016/i.taap.2017.03.024
Wei. C: Wen. H: Yuan. L: Mchale. CM: Li. H: Wang. K: Yuan. 1: Yang. X: Zhang. L (2017b).
Formaldehyde induces toxicity in mouse bone marrow and hematopoietic stem/progenitor
cells and enhances benzene-induced adverse effects. Arch Toxicol 91: 921-933.
http://dx.doi.Org/10.1007/s00204-016-1760-5
Wei. CN: Harada. K: Ohmori. S: Wei. 01: Minamoto. K: Ueda. A. (2007). Subjective symptoms of
medical students exposed to formaldehyde during a gross anatomy dissection course. Int J
Immunopathol Pharmacol 20: 23-25. http://dx.doi.org/10.1177/03946320070200S2Q5
Wei. H: Tan. K: Sun. R: Yin. L: Zhang. I: Pu. Y. (2014). Aberrant production of Thl/Th2/Thl7-related
cytokines in serum of C57BL/6 mice after short-term formaldehyde exposure. Int J Environ
Res Public Health 11: 10036-10050. http://dx.doi.org/10.3390/ijerphlll010036
Weibel. ER. (1963). Morphometry of the human lung. Berlin, Germany: Springer-Verlag.
http://dx.doi.org/10.1007/978-3-642-87553-3
Weisskopf. M: Morozova. N: O'Reilly. El: Mccullough. ML: Calle. EE: Thun. Ml: Ascherio. A. (2009).
Prospective study of chemical exposures and amyotrophic lateral sclerosis mortality. J
Neurol Neurosurg Psychiatry 80: 558-561. http://dx.doi.org/10.1136/innp.2008.156976
Wen. H: Yuan. L: Wei. C: Zhao. Y: Oian. Y: Ma. P: Ding. S: Yang. X: Wang. X. (2016). Effects of
combined exposure to formaldehyde and benzene on immune cells in the blood and spleen
in Balb/c mice. Environ Toxicol Pharmacol 45: 265-273.
http: / /dx. doi. or g/10.1016/i. etap. 2016.05.007
West. S: Hildesheim. A: Dosemeci. M. (1993). Non-viral risk factors for nasopharyngeal carcinoma in
the Philippines: Results from a case-control study. Int J Cancer 55: 722-727.
http://dx.doi.Org/10.1002/iic.2910550504
WHO (World Health Organization). (1989). Environmental health criteria 89: Formaldehyde.
(RISKLINE/1990090019). http://www.inchem.org/documents/ehc/ehc/ehc89.htm
WHO. (2002). Concise international chemical assessment document 40. Geneva, Switzerland.
https://www.who.int/ipcs/publications/cicad/en/cicad40.pdf
Wilcox. AT. (2010). Fertility and pregnancy: An epidemiologic perspective. In Fertility and
pregnancy: An epidemiologic perspective. New York, NY: Oxford University Press.
Wilcox. AT: Hornev. LF. (1984). Accuracy of spontaneous abortion recall. Am J Epidemiol 120: 727-
733. http://dx.doi.org/10.1093/oxfordiournals.aie.all3940
Wilcox. P: Naidoo. A: Wedd. PI: Gatehouse. DG. (1990). Comparison of Salmonella typhimurium
TA102 with Escherichia coli WP2 tester strains. Mutagenesis 5: 285-291.
http: / / dx. do i. o r g /10.109 3 /mutage / 5.3.2 8 5
This document is a draft for review purposes only and does not constitute Agency policy.
R-68 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Supplemental Information for Formaldehyde—Inhalation
Wilhelmsson. B: Holmstrom. M. (1992). Possible mechanisms of formaldehyde-induced discomfort
in the upper airways. Scand J Work Environ Health 18: 403-407.
http: //dx.doi.org/10.5 2 71 /si weh. 1556
Wilkins. RT: Macleod. HP. (1976). Formaldehyde induced DNA-protein crosslinks in Escherichia
Coli. Mutat Res 36: 11-16. http://dx.doi.org/10.1016/0027-5107r76190016-6
Williams. GM: Mori. H: Mcqueen. CA. (1989a). Structure-activity relationships in the rat hepatocyte
DNA-repair test for 300 chemicals [Review], Mutat Res 221: 263-286.
http://dx.doi.Org/l 0.1016/0165-1110r89190039-0
Williams. HC: Burnev. PG: Pembroke. AC: Hay. RT. (1996). Validation of the U.K. diagnostic criteria
for atopic dermatitis in a population setting. Br J Dermatol 135: 12-17.
http: //dx.doi.org/10.1046 /i.l 365-2133.1996.d01 -925.x
Williams. RL: Lipari. F: Potter. RA. (1989b). Formaldehyde, methanol, and hydrocarbon emissions
from methanol-fueled cars.
Wilmer. TWG. M: Woutersen. RA: Appelman. LM: Leeman. WR: Feron. VI. (1987). Subacute (4-week)
inhalation toxicity study of formaldehyde in male rats: 8-hour intermittent versus 8-hour
continuous exposures. J Appl Toxicol 7: 15-16. http://dx.doi.org/10.1002 /iat25500701Q4
Wilmer. TWG. M: Woutersen. RA: Appelman. LM: Leeman. WR: Feron. VI. (1989). Subchronic (13-
week) inhalation toxicity study of formaldehyde in male rats: 8-hour intermittent versus 8-
hour continuous exposures. Toxicol Lett 47: 287-293. http: / /dx.doi. or g/10.1016/0 3 78-
4274C89190147-1
Witek. TT. Tr: Schachter. EN: Tosun. T: Beck. GT: Leaderer. BP. (1987). An evaluation of respiratory
effects following exposure to 2.0 ppm formaldehyde in asthmatics: Lung function,
symptoms, and airway reactivity. Arch Environ Health 42: 230-237.
Witek. TT. Tr: Schachter. EN: Tosun. T: Leaderer. BP: Beck. GT. (1986). Controlled human studies on
the pulmonary effects of indoor air pollution: Experiences with sulfur dioxide and
formaldehyde. Environ Int 12: 129-135. http://dx.doi.org/10.1016/0160-4120(86)90023-1
Wolf. DC: Gross. EA: Lvght. 0: Bermudez. E: Recio. L: Morgan. KT. (1995). Immunohistochemical
localization of p53, PCNA, and TGF-alpha proteins in formaldehyde-induced rat nasal
squamous cell carcinomas. Toxicol Appl Pharmacol 132: 27-35.
http: / /dx. doi. or g/10.10 0 6 /taap. 1995.1083
Wolff. RK. (1986). Effects of airborne pollutants on mucociliary clearance [Review], Environ Health
Perspect66: 223-237.
Wong. MA: Isaza. R: Cuthbert. IK: Brooks. DE: Samuelson. D. (2012). PERIOCULAR ANTERIOR
ADNEXAL ANATOMY AND CLINICAL ADNEXAL EXAMINATION OF THE ADULT ASIAN
ELEPHANT (ELEPHAS MAXIMUS). J Zoo Wildl Med 43: 793-801.
http: //dx.doi.org/10.1638/2011 -0173R2.1
Wood. RW: Coleman. IB. (1995). Behavioral evaluation of the irritant properties of formaldehyde.
Toxicol Appl Pharmacol 130: 67-72. http: //dx.doi.org/10.1006/taap.1995.1009
Wortlev. P: Vaughan. TL: Davis. S: Morgan. MS: Thomas. DB. (1992). A case-control study of
occupational risk factors for laryngeal cancer. Br J Ind Med 49: 837-844.
http://dx.doi.org/10.1136/oem.49.12.837
This document is a draft for review purposes only and does not constitute Agency policy.
R-69 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Woutersen. RA: Appelman. LM: Wilmer. TWG. M: Falke. HE: Feron. VT. (1987). Subchronic (13-week)
inhalation toxicity study of formaldehyde in rats. J Appl Toxicol 7: 43-49.
http://dx.doi.org/10.1002/iat.2SS0070108
Woutersen. RA: van Garderen-Hoetmer. A: Bruijntjes. TP: Zwart. A: Feron. VI. (1989). Nasal tumours
in rats after severe injury to the nasal mucosa and prolonged exposure to 10 ppm
formaldehyde. J Appl Toxicol 9: 39-46. http: //dx.doi.org/10.1002 /iat.2550090108
Wu. Y: Duan. 1: Li. B: Liu. H: Chen. M. (2020). Exposure to formaldehyde at low temperatures
aggravates allergic asthma involved in transient receptor potential ion channel. Environ
Toxicol Pharmacol 80: 103469. http://dx.doi.Org/10.1016/i.etap.2020.103469
Wu. Y: You. H: Ma. P: Li. L: Yuan. Y: Li. 1: Liu. X: Yao. H: Chen. R: Lai. K: Yang. X. (2013). Role of
transient receptor potential ion channels and evoked levels of neuropeptides in a
formaldehyde-induced model of asthma in Balb/c mice. PLoS ONE 8: e62827.
http://dx.doi.org/10.1371/iournal.pone.0062827
Xie. SH: Yu. IT: Tse. LA: Au. IS: Lau. IS. (2017). Occupational risk factors for nasopharyngeal
carcinoma in Hong Kong Chinese: a case-referent study. Int Arch Occup Environ Health 90:
443-449. http://dx.doi.Org/10.1007/s00420-017-1212-4
Xin. L: Wang. 1: Fan. G: Wu. Y: Guo. S. (2015). Activation of HSPA1A promoter by environmental
pollutants: An early and rapid response to cellular damage. Environ Toxicol Pharmacol 39:
1027-1033. http://dx.doi.org/10.1016/i.etap.2015.03.011
Xing. C: Zhang. SY: Deng. 1: Wang. S. (2007). Urea-formaldehyde-resin gel time as affected by the pH
value, solid content, and catalyst. J Appl Polymer Sci 103: 1566-1569.
http://dx.doi.org/10.1002/app.25343
Yager. TW: Cohn. KL: Spear. RC: Fisher. TM: Morse. L. (1986). Sister-chromatid exchanges in
lymphocytes of anatomy students exposed to formaldehyde-embalming solution. Mutat Res
174: 135-139. http://dx.doi.org/10.1016/0165-7992T86190104-1
Yan. Y: Ye. Z: Lu. ZS: Qiao. Y: Yang. X: Li. CM. (2005). Nitric oxide level associated with gaseous
formaldehyde exposure in lungs of mice. In X Yang; B Zhao; R Zhao (Eds.), Indoor Air 2005:
Proceedings of the 10th International Conference on Indoor Air Quality and Climate, vol 5
(pp. 3851-3854). Beijing China: Tsinghua University Press.
https://www.isiaq.org/docs/PDFs/3851.pdf
Yang. X: Zhang. YP: Chen. D: Chen. WG: Wang. R. (2001). Eye irritation caused by formaldehyde as
an indoor air pollution--a controlled human exposure experiment Biomed Environ Sci 14:
229-236.
Yang. XR: Diehl. S: Pfeiffer. R: Chen. CI: Hsu. WL: Dosemeci. M: Cheng. YT: Sun. B: Goldstein. AM:
Hildesheim. A: Team. CaAGEoNS. (2005). Evaluation of risk factors for nasopharyngeal
carcinoma in high-risk nasopharyngeal carcinoma families in Taiwan. Cancer Epidemiol
Biomarkers Prev 14: 900-905. http://dx.doi.org/10.1158/1055-9965.EPI-04-068Q
Yang. Y: Luo. H: Liu. R: Li. G: Yu. Y: An. T. (2020). The exposure risk of typical VOCs to the human
beings via inhalation based on the respiratory deposition rates by proton transfer reaction-
time of flight-mass spectrometer. Ecotoxicol Environ Saf 197: 110615.
http://dx.doi.Org/10.1016/j.ecoenv.2020.110615
Ye. X: Ti. Z: Wei. C: Mchale. C: Ding. S: Thomas. R: Yang. X: Zhang. L. (2013a). Inhaled formaldehyde
induces DNA-protein crosslinks and oxidative stress in the bone marrow and other distant
organs of exposed mice [Abstract], Environ Mol Mutagen 54: S41.
This document is a draft for review purposes only and does not constitute Agency policy.
R-70 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Supplemental Information for Formaldehyde—Inhalation
Ye. X: Ti. Z: Wei. C: Mchale. CM: Ding. S: Thomas. R: Yang. X: Zhang. L. (2013b). Inhaled formaldehyde
induces DNA-protein crosslinks and oxidative stress in bone marrow and other distant
organs of exposed mice. Environ Mol Mutagen 54: 705-718.
http: //dx. doi. or g/10.10 0 2 /em. 21821
Ye. X: Yan. W: Xie. H: Zhao. M: Ying. C. (2005). Cytogenetic analysis of nasal mucosa cells and
lymphocytes from high-level long-term formaldehyde exposed workers and low-level short-
term exposed waiters. MutatRes 588: 22-27.
http://dx.doi.Org/10.1016/j.mrgentox.2005.08.005
Yeatts. KB: El-Sadig. M: Leith. D: Kalsbeek. W: Al-Maskari. F: Couper. D: Funk. WE: Zoubeidi. T: Chan.
RL: Trent. CB: Davidson. CA: Boundv. MG: Kassab. MM: Hasan. MY: Rusvn. I: Gibson. TM:
Olshan. AF. (2012). Indoor air pollutants and health in the United Arab Emirates. Environ
Health Perspect 120: 687-694. http://dx.doi.org/10.1289/ehp.110409Q
Ying. CI: Yan. WS: Zhao. MY: Ye. XL: Xie. H: Yin. SY: Zhu. XS. (1997). Micronuclei in nasal mucosa,
oral mucosa and lymphocytes in students exposed to formaldehyde vapor in anatomy class.
Biomed Environ Sci 10: 451-455.
Ying. CI: Ye. XL: Xie. H: Yan. WS: Zhao. MY: Xia. T: Yin. SY. (1999). Lymphocyte subsets and sister-
chromatid exchanges in the students exposed to formaldehyde vapor. Biomed Environ Sci
12: 88-94.
Yoklev. KA. (2012). Sensory irritation response in rats II: Recovery and dose-dependence. Bull Math
Biol 74: 1673-1690. http://dx.doi.org/10.1007/sll538-012-9730-4
Yon. DK: Hwang. S: Lee. SW: Tee. HM: Sheen. YH: Kim. TH: Lim. DH: Han. MY. (2019). Indoor
Exposure and Sensitization to Formaldehyde among Inner-City Children with Increased
Risk for Asthma and Rhinitis. Am J Respir Crit Care Med 200: 388-393.
http://dx.doi.org/10.1164/rccm.201810-1980LE
Yonemitsu. T: Kuroki. C: Takahashi. N: Mori. Y: Kanmura. Y: Kashiwadani. H: Ootsuka. Y: Kuwaki. T.
(2013). TRPA1 detects environmental chemicals and induces avoidance behavior and
arousal from sleep. Sci Rep 3: 3100. http://dx.doi.org/10.1038/srepO3100
Yoo. ST: Ito. K. (2018a). Assessment of transient inhalation exposure using in silico human model
integrated with PBPK-CFD hybrid analysis. Sustain Cities Soc 40: 317-325.
http://dx.doi.Org/10.1016/j.scs.2018.04.023
Yoo. ST. un: Ito. K. (2018b). Numerical prediction of tissue dosimetry in respiratory tract using
computer simulated person integrated with physiologically based pharmacokinetic-
computational fluid dynamics hybrid analysis. Indoor Built Environ 27: 877-889.
http: //dx.doi.org/10.1177/1420326X17694475
Yorgancilar. E: Deveci. E: Deveci. S. (2012). Effects of formaldehyde on respiratory mucosa in rats.
International Journal of Morphology 30: 521-523. http://dx.doi.org/10.4067/SQ717-
95022012000200026
Yu. C: Crump. D. (1998). A review of the emission of VOCs from polymeric materials used in
buildings [Review], Build Environ 33: 357-374. http: / /dx. doi. or g /10.1016 / S0 3 6 0-
1323C97100055-3
Yu. G: Chen. 0: Liu. X: Guo. C: Du. H: Sun. Z. (2014a). Formaldehyde induces bone marrow toxicity in
mice by inhibiting peroxiredoxin 2 expression. Mol Med Rep 10: 1915-1920.
http: / /dx. doi. or g/10.3 89 2 /mmr. 2014.2473
This document is a draft for review purposes only and does not constitute Agency policy.
R-71 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Yu. GY: Song. XF: Liu. Y: Sun. ZW. (2014b). Inhaled Formaldehyde Induces Bone Marrow Toxicity via
Oxidative Stress in Exposed Mice. Asian Pac J Cancer Prev 15: 5253-5257.
http://dx.doi.org/10.7314/APTCP.2014.15.13.5253
Yu. GY: Song. XF: Zhao. SH: Liu. Y: Sun. ZW. (2015a). Formaldehyde induces the bone marrow
toxicity in mice by regulating the expression of Prx3 protein. J Huazhong Univ Sci Technolog
Med Sci 35: 82-86. http://dx.doi.org/10.1007/sll596-015-1393-6
Yu. IT: Li. AM: Goggins. W: Leung. TO: Chan. GY: Fung. CK: Chan. CK: Lau. AP. (2017). Association of
wheeze during the first 18 months of life with indoor nitrogen dioxide, formaldehyde, and
family history of asthma: a prospective cohort study. Hong Kong Med J 23 Suppl 2: 19-23.
Yu. ITS: Chin. YL: Wong. TW: Tang. TL. (2004). Deaths from nasopharyngeal cancer among waiters
and waitresses in Chinese restaurants. Int Arch Occup Environ Health 77: 499-504.
http://dx.doi.Org/10.1007/s00420-004-0543-0
Yu. LP: Tiang. SF: Leng. SG: He. FS: Zheng. YX. (2005). Early genetic effects on workers
occupationally exposed to formaldehyde. Zhonghua Yufang Yixue Zazhi 39: 392-395.
Yu. PH: Lai. CT: Zuo. DM. (1997). Formation of formaldehyde from adrenaline in vivo; a potential
risk factor for stress-related angiopathy. Neurochem Res 22: 615-620.
http://dx.doi.Org/10.1023/a:1022478221421
Yu. PH: Zuo. DM. (1996). Formaldehyde produced endogenously via deamination of methylamine. A
potential risk factor for initiation of endothelial injury. Atherosclerosis 120: 189-197.
http://dx.doi.org/10.1016/0021-9150r95105701-3
Yu. R: Lai. Y: Hartwell. HI: Moeller. BC: Dovle-Eisele. M: Kracko. D: Bodnar. WM: Starr. TB:
Swenberg. TA. (2015b). Formation, Accumulation, and Hydrolysis of Endogenous and
Exogenous Formaldehyde-Induced DNA Damage. Toxicol Sci 146: 170-182.
http: / /dx. doi. or g/10.109 3 /toxsci /kfvO 7 9
Yu. YH: Blessing. WW. (1997). Cerebral blood flow in rabbits during the nasopharyngeal reflex
elicited by inhalation of noxious vapor. J Auton Nerv Syst 66: 149-153.
http://dx.doi.org/10.1016/s0165-1838r97100080-5
Yu. YH: Blessing. WW. (1999). Amygdala co-ordinates sudden falls in ear pinna blood flow in
response to unconditioned salient stimuli in conscious rabbits. Neuroscience 93: 135-141.
http://dx.doi.Org/10.1016/s0306-4522r99100097-4
Zang. ZT: Fang. YO: Ti. SY: Gao. Y: Zhu. YO: Xia. TT: Tiang. MH: Zhang. YN. (2017). Formaldehyde
Inhibits Sexual Behavior and Expression of Steroidogenic Enzymes in the Testes of Mice. J
Sex Med 14: 1297-1306. http: //dx.doi.org/10.1016/i.isxm.2017.09.001
Zararsiz. I: Kus. I: Akpolat. N: Songur. A: Ogeturk. M: Sarsilmaz. M. (2006). Protective effects of
omega-3 essential fatty acids against formaldehyde-induced neuronal damage in prefrontal
cortex of rats. Cell Biochem Funct 24: 237-244. http: //dx.doi.org/10.1002 /cbf.1204
Zarei. F: Rezazadeh Azari. M: Salehpour. S: Khodakarim. S: Omidi. L: Tavakol. E. (2017). Respiratory
effects of simultaneous exposure to respirable crystalline silica dust, formaldehyde, and
triethylamine of a group of foundry workers. Journal of Research in Health Sciences 17: El-
E6.
Zeller. 1: Neuss. S: Mueller. TU: Kiihner. S: Holzmann. K: Hogel. 1: Klingmann. C: Bruckner. T: Triebig.
G: Speit. G. (2011). Assessment of genotoxic effects and changes in gene expression in
humans exposed to formaldehyde by inhalation under controlled conditions. Mutagenesis
26: 555-561. http: / /dx.doi. or g/10.1093 /mutage /ger 016
This document is a draft for review purposes only and does not constitute Agency policy.
R-72 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Supplemental Information for Formaldehyde—Inhalation
Zendehdel. R: Touni. FT: Haiipour. B: Paniali. Z: Kheiri. H: Vahabi. M. (2017). DNA damage in workers
exposed to formaldehyde at concentrations below occupational exposure limits. Toxicol
Environ Chem 99: 1409-1417. http://dx.doi.Org/l 0.1080/02772248.2017.1343335
Zendehdel. R: Vahabi. M: Sedghi. R. (2018). Estimation of formaldehyde occupational exposure limit
based on genetic damage in some Iranian exposed workers using benchmark dose method.
Environ Sci PollutRes Int25: 31183-31189. http://dx.doi.org/10.1007/sll356-018-3077-
9
Zhai. L: Zhao. 1: Xu. B: Deng. Y: Xu. Z. (2013). Influence of indoor formaldehyde pollution on
respiratory system health in the urban area of Shenyang China. Afr Health Sci 13: 137-143.
http ://dx. doi. or g/10.4314 /ahs. vl 3 il. 19
Zhang. 1: Sun. R: Chen. Y: Tan. K: Wei. H: Yin. L: Pu. Y. (2014a). Small molecule metabolite biomarker
candidates in urine from mice exposed to formaldehyde. International Journal of Molecular
Sciences 15: 16458-16468. http://dx.doi.org/10.3390/iimsl50916458
Zhang. L: Tang. X: Rothman. N: Vermeulen. R: Ti. Z: Shen. M: Oiu. C: Guo. W: Liu. S: Reiss. B: Freeman.
LB: Ge. Y: Hubbard. AE: Hua. M: Blair. A: Galvan. N: Ruan. X: Alter. BP: Xin. KX: Li. S: Moore.
LE: Kim. S: Xie. Y: Hayes. RB: Azuma. M: Hauptmann. M: Xiong. 1: Stewart. P: Li. L: Rappaport.
SM: Huang. H: Fraumeni. IF. Tr: Smith. MT: Lan. 0. (2010). Occupational exposure to
formaldehyde, hematotoxicity, and leukemia-specific chromosome changes in cultured
myeloid progenitor cells. Cancer Epidemiol Biomarkers Prev 19: 80-88.
http://dx.doi.org/10.1158/1055-9965.EPI-09-0762
Zhang. 0: Yan. W: Bai. Y: Zhu. Y: Ma. 1. (2014b). Repeated formaldehyde inhalation impaired
olfactory function and changed SNAP25 proteins in olfactory bulb. Int J Occup Environ
Health 20: 308-312. http://dx.doi.org/10.1179/2049396714Y.0000000079
Zhang. S: Chen. H: Wang. A: Liu. Y: Hou. H: Hu. 0. (2018a). Combined effects of co-exposure to
formaldehyde and acrolein mixtures on cytotoxicity and genotoxicity in vitro. Environ Sci
PollutRes Int 25: 25306-25314. http://dx.doi.org/10.1007/sl 1356-018-2584-z
Zhang. S: Chen. H: Zhang. 1: Li. I. un: Hou. H: Hu. 0. (2020a). The multiplex interactions and
molecular mechanism on genotoxicity induced by formaldehyde and acrolein mixtures on
human bronchial epithelial BEAS-2B cells. Environ Int 143: 105943.
http://dx.doi.Org/10.1016/j.envint2020.105943
Zhang. S: Zhang. I: Chen. H: Wang. A: Liu. Y: Hou. H: Hu. 0. (2019). Combined cytotoxicity of co-
exposure to aldehyde mixtures on human bronchial epithelial BEAS-2B cells. Environ Pollut
250: 650-661. http://dx.doi.org/10.1016/i.envpol.2019.03.118
Zhang. S: Zhang. 1: Cheng. W: Chen. H: Wang. A: Liu. Y: Hou. H: Hu. 0. (2020b). Combined cell death
of co-exposure to aldehyde mixtures on human bronchial epithelial BEAS-2B cells:
Molecular insights into the joint action. Chemosphere 244: 125482.
http://dx.doi.Org/10.1016/i.chemosphere.2019.125482
Zhang. X: Zhao. Y: Song. 1: Yang. X: Zhang. I: Zhang. Y: Li. R. (2018b). Differential health effects of
constant versus intermittent exposure to formaldehyde in mice: Implications for building
ventilation strategies. Environ Sci Technol 52: 1551-1560.
http://dx.doi.org/10.1021/acs.est7b05015
Zhang. Y: Liu. X: Mchale. C: Li. R: Zhang. L: Wu. Y: Ye. X: Yang. X: Ding. S. (2013). Bone marrow injury
induced via oxidative stress in mice by inhalation exposure to formaldehyde. PLoS ONE 8:
e74974. http://dx.doi.org/10.1371/iournal.pone.0074974
This document is a draft for review purposes only and does not constitute Agency policy.
R-73 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information for Formaldehyde—Inhalation
Zhao. W: Guangvin. P: Yang. X. (2009). DNA-protein crosslinks induced by formaldehyde and its
repair process. Int J Environ Pollut 37: 299-308.
Zhao. Y. un: Magana. LC: Cui. H: Huang. 1: Mchale. CM: Yang. X. u: Loonev. MR: Li. R. ui: Zhang. L.
(2020). Formaldehyde-induced hematopoietic stem and progenitor cell toxicity in mouse
lung and nose. Arch Toxicol 95: 693-701. http://dx.doi.Org/10.1007/s00204-020-02932-x
Zhitkovich. A: Costa. M. (1992). A simple, sensitive assay to detect DNA-protein crosslinks in intact
cells and in vivo. Carcinogenesis 13: 1485-1489.
http: //dx. doi. or g/10.109 3 /carcin /13.8.1485
Zhong. W: Hee. SO. (2004). Quantitation of normal and formaldehyde-modified deoxynucleosides
by high-performance liquid chromatography/UV detection. Biomed Chromatogr 18: 462-
469. http: //dx. doi. o r g/10.10 0 2 /bmc. 3 3 7
Zhong. W: Que Hee. SS. (2004). Formaldehyde-induced DNA adducts as biomarkers of in vitro
human nasal epithelial cell exposure to formaldehyde. Mutat Res 563: 13-24.
http://dx.doi.Org/10.1016/i.mrgentox.2004.05.012
Zhou. DX: Oiu. SD: Zhang. 1: Tian. H: Wang. HX. (2006). The protective effect of vitamin E against
oxidative damage caused by formaldehyde in the testes of adult rats. Asian J Androl 8: 584-
588. http:/ /dx.doi.org/10.1111 /i.l 745-7262.2006.00198.x
Zhou. ES: Kane. YY: Gao. XX: Wu. LF: Lu. ZS: Yan. Y: Qiao. Y: Yang. X. (2005). A pilot investigation on
human serum formaldehyde-specific IgE. Paper presented at 10th International Conference
on Indoor Air Quality and Climate, September 4-9, 2005, Beijing, China.
Zhu. TL: Knudsen. LE: Andersen. AM: Hiollund. NH: Olsen. 1. (2005). Time to pregnancy among
Danish laboratory technicians who were a part of the National Birth Cohort Scand J Work
Environ Health 31: 108-114.
Zhu. TL: Knudsen. LE: Andersen. AM: Hiollund. NH: Olsen. 1. (2006). Laboratory work and pregnancy
outcomes: a study within the National Birth Cohort in Denmark. Occup Environ Med 63: 53-
58. http://dx.doi.Org/10.1136/oem.2005.021204
Zimmermann. FK: Mohr. A. (1992). Formaldehyde, glyoxal, urethane, methyl carbamate, 2,3-
butanedione, 2,3-hexanedione, ethyl acrylate, dibromoacetonitrile and 2-
hydroxypropionitrile induce chromosome loss in Saccharomyces cerevisiae. Mutat Res 270:
151-166. http://dx.d0i.0rg/l 0.1016/0027-5107(92190126-M
Zitting. A. (1982). IV. Biochemical Effects (pp. 43-60). (NIOSH/00126377). Zitting, A.
Zitting. A: Savolainen. H: Nickels. I. (1982). Biochemical and toxicological effects of single and
repeated exposures to polyacetal thermodegradation products. Environ Res 29: 287-296.
http://dx.d0i.0rg/l 0.1016/0013-9351 (82190031 -7
Zwart. A: Woutersen. RA: Wilmer. TWG. M: Spit. BT: Feron. VI. (1988). Cytotoxic and adaptive effects
in rat nasal epithelium after 3-day and 13-week exposure to low concentrations of
formaldehyde vapour. Toxicology 51: 87-99. http://dx.doi.org/10.1016/030Q-
483Xf88190083-2
This document is a draft for review purposes only and does not constitute Agency policy.
R-74 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information for Formaldehyde—Inhalation
This document is a draft for review purposes only and does not constitute Agency policy.
R-75 DRAFT-DO NOT CITE OR QUOTE
------- |