DRAFT - DO NOT CITE OR QUOTE
&EPA
EPA/635/R-10/002C
www.epa.gov/iris
TOXICOLOGICAL REVIEW OF
FORMALDEHYDE
INHALATION TOXICITY
(CAS No. 50-00-0)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
VOLUME II of IV
Hazard Characterization
March 17,2010
NOTICE
This document is an Inter-Agency Science Consultation draft. This information is distributed
solely for the purpose of pre-dissemination 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. It is being circulated for review
of its technical accuracy and science policy implications.
1
U.S. Environmental Protection Agency
Washington, DC
-------
1 DISCLAIMER
2
3 This document is a preliminary draft for review purposes only. This information is
4 distributed solely for the purpose of pre-dissemination peer review under applicable information
5 quality guidelines. It has not been formally disseminated by EPA. It does not represent and
6 should not be construed to represent any Agency determination or policy. Mention of trade
7 names or commercial products does not constitute endorsement or recommendation for use.
8
9
This document is a draft for review purposes only and does not constitute Agency policy.
Il-ii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS—TOXICOLOGICAL REVIEW OF FORMALDEHYDE
(CAS No. 50-00-0)
LIST OF TABLES xi
LIST OF FIGURES xx
LIST 01 ABBREVIATIONS AM) ACRONYMS xxv
FOREWORD xxxii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xxxiii
VOLUME I
1. INTRODUCTION 1-1
2. BACKGROUND 2-1
2.1. PHYSICOCHEMICAL PROPERTIES OF FORMALDEHYDE 2-1
2.2. PRODUCTION, USES, AND SOURCES OF FORMALDEHYDE 2-1
2.3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE 2-4
2.3.1. Inhalation 2-5
2.3.2. Ingestion 2-10
2.3.3. Dermal Contact 2-11
3. TOXICOKINETICS 3-1
3.1. CHEMICAL PROPERTIES AND REACTIVITY 3-1
3.1.1. Binding of Formaldehyde to Proteins 3-1
3.1.2. Endogenous Sources of Formaldehyde 3-3
3.1.2.1. Normal Cellular Metabolism (Enzymatic) 3-3
3.1.2.2. Normal Metabolism (Non-Enzymatic) 3-5
3.1.2.3. Exogenous Sources of Formaldehyde Production 3-5
3.1.2.4. FA-GSH Conjugate as a Method of Systemic Distribution 3-6
3.1.2.5. Metabolic Products of FA Metabolism (e.g., Formic Acid) 3-6
3.1.2.6. Levels of Endogenous Formaldehyde in Animal and Human
Tissues 3-6
3.2. ABSORPTION 3-9
3.2.1. Oral 3-9
3.2.2. Dermal 3-9
3.2.3. Inhalation 3-9
3.2.3.1. Formaldehyde Uptake Can be Affected by Effects at the
Portal of Entry 3-10
3.2.3.2. Variability in the Nasal Dosimetry of Formaldehyde in
Adults and Children 3-12
3.3. DISTRIBUTION 3-13
3.3.1. Levels in Blood 3-13
3.3.2. Levels in Various Tissues 3-15
3.3.2.1. Disposition of Formaldehyde: Differentiating covalent
Binding and Metabolic Incorporation 3-16
This document is a draft for review purposes only and does not constitute Agency policy.
II-iii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
3.4. METABOLISM 3-20
3.4.1. In Vitro and In Vivo Characterization of Formaldehyde Metabolism 3-20
3.4.2. Formaldehyde Exposure and Perturbation of Metabolic Pathways 3-23
3.4.3. Evidence for Susceptibility in Formaldehyde Metabolism 3-24
3.5. EXCRETION 3-25
3.5.1. Formaldehyde Excretion in Rodents 3-26
3.5.2. Formaldehyde Excretion in Exhaled Human Breath 3-27
3.5.3. Formaldehyde Excretion in Human Urine 3-31
3.6. MODELING THE TOXICOKINETICS OF FORMALDEHYDE AND DPX 3-32
3.6.1. Motivation 3-32
3.6.2. Species Differences in Anatomy: Consequences for Gas Transport and
Risk 3-34
3.6.3. Modeling Formaldehyde Uptake in Nasal Passages 3-40
3.6.3.1. Flux Bins 3-41
3.6.3.2. Flux Estimates 3-41
3.6.3.3. Mass Balance Errors 3-42
3.6.4. Modeling Formaldehyde Uptake in the Lower Respiratory Tract 3-42
3.6.5. Uncertainties in Formaldehyde Dosimetry Modeling 3-44
3.6.5.1. Verification of Predicted Flow Profiles 3-44
3.6.5.2. Level of Confidence in Formaldehyde Uptake Simulations 3-45
3.6.6. PBPK Modeling of DNA Protein Cross-Links (DPXs) Formed by
Formaldehyde 3-48
3.6.6.1. PBPK Models for DPXs 3-48
3.6.6.2. A PBPK Model for DPXs in the F344 Rat and Rhesus
Monkey that uses Local Tissue Dose of Formaldehyde 3-50
3.6.6.3. Uncertainties in Modeling the Rat and Rhesus DPX Data 3-51
3.6.7. Uncertainty in Prediction of Human DPX Concentrations 3-53
VOLUME II
4. HAZARD CHARACTERIZATION 4-1
4.1. HUMAN STUDIES 4-1
4.1.1. Noncancer Health Effects 4-1
4.1.1.1. Sensory Irritation (Eye, Nose, Throat Irritation) 4-1
4.1.1.2. Pulmonary Function 4-11
4.1.1.3. Asthma 4-19
4.1.1.4. Respiratory Tract Pathology 4-26
4.1.1.5. Immunologic Effects 4-30
4.1.1.6. Neurological/Behavioral 4-42
4.1.1.7. Developmental and Reproductive Toxicity 4-45
4.1.1.8. Oral Exposure Effects on the Gastrointestinal Tract 4-56
4.1.1.9. Summary: Noncarcinogenic Hazard in Humans 4-56
This document is a draft for review purposes only and does not constitute Agency policy.
Il-iv DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
4.1.2. Cancer Health Effects 4-57
4.1.2.1. Respiratory Tract Cancer 4-57
4.1.2.2. Non-Respiratory Tract Cancer 4-84
4.1.2.3. Summary: Carcinogenic Hazard in Humans 4-107
4.2. ANIMAL STUDIES 4-109
4.2.1. Noncancer Health Effects 4-110
4.2.1.1. Reflex Bradypnea 4-110
4.2.1.2. Respiratory Tract Pathology 4-120
4.2.1.3. Gastrointestinal Tract and Systemic Toxicity 4-201
4.2.1.4. Immune Function 4-216
4.2.1.5. Hypersensitivity and Atopic Reactions 4-225
4.2.1.6. Neurological and Neurobehavioral Function 4-250
4.2.1.7. Reproductive and Developmental Toxicity 4-285
4.2.2. Carcinogenic Potential: Animal Bioassays 4-324
4.2.2.1. Respiratory Tract 4-324
4.2.2.2. Gastrointestinal Tract 4-326
4.2.2.3. Lymphohematopoietic Cancer 4-328
4.2.2.4. Summary 4-335
4.3. GENOTOXICITY 4-335
4.3.1. Formaldehyde-DNAReactions 4-335
4.3.1.1. DNA-Protein Cross-Links (DPXs) 4-336
4.3.1.2. DNA Adducts 4-341
4.3.1.3. DNA-DNA Cross-Links (DDXs) 4-343
4.3.1.4. Single Strand Breaks 4-344
4.3.1.5. Other Genetic Effects of Formaldehyde in Mammalian Cells 4-345
4.3.2. In Vitro Clastogenicity 4-345
4.3.3. In Vitro Mutagenicity 4-347
4.3.3.1. Mutagenicity in Bacterial Systems 4-347
4.3.3.2. Mutagenicity in Non-mammalian Cell Systems 4-353
4.3.3.3. Mutagenicity in Mammalian Cell Systems 4-353
4.3.4. In Vivo Mammalian Genotoxicity 4-360
4.3.4.1. Genotoxicity in Laboratory Animals 4-360
4.3.4.2. Genotoxicity in Humans 4-362
4.3.5. Summary of Genotoxicity 4-370
4.4. SYNTHESIS AND MAJOR EVALUATION OF NONCARCINOGENIC
EFFECTS 4-371
4.4.1. Sensory Irritation 4-376
4.4.2. Pulmonary Function 4-379
4.4.3. Hypersensitivity and Atopic Reactions 4-382
4.4.4. Upper Respiratory Tract Histopathology 4-383
4.4.5. Toxicogenomic and Molecular Data that May Inform MO As 4-385
4.4.6. Noncancer Modes of Actions 4-387
4.4.7. Immunotoxicity 4-389
This document is a draft for review purposes only and does not constitute Agency policy.
II-V DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
4.4.8. Effects on the Nervous System 4-390
4.4.8.1. Irritant Threshold Detection 4-390
4.4.8.2. Behavioral Effects 4-391
4.4.8.3. Neurochemistry, Neuropathology, and Mechanistic Studies 4-392
4.4.8.4. Summary 4-392
4.4.8.5. Data Gaps 4-393
4.4.9. Reproductive and Developmental Toxicity 4-393
4.4.9.1. Spontaneous Abortion and Fetal Death 4-393
4.4.9.2. Congenital Malformations 4-396
4.4.9.3. Low Birth Weight and Growth Retardation 4-396
4.4.9.4. Functional Development Outcomes (Developmental
Neurotoxicity) 4-397
4.4.9.5. Male Reproductive Toxicity 4-398
4.4.9.6. Female Reproductive Toxicity 4-399
4.4.9.7. Mode of Action 4-400
4.4.9.8. Data Gaps 4-402
4.5. SYNTHESIS AND EVALUATION OF CARCINOGENICITY 4-402
4.5.1. Cancers of the Respiratory Tract 4-402
4.5.2. Lymphohematopoietic Malignancies 4-408
4.5.2.1. Background 4-408
4.5.2.2. All LHP Malignancies 4-410
4.5.2.3. All Leukemia 4-414
4.5.2.4. Subtype Analysis 4-418
4.5.2.5. Myeloid Leukemia 4-419
4.5.2.6. Solid Tumors of Lymphoid Origin 4-421
4.5.2.7. Supporting Evidence from Animal Bio-Assays for
Formaldehyde-Induced Lymphohematopoietic Malignancies 4-423
4.5.3. Carcinogenic Mode(s) of Action 4-427
4.5.3.1. Mechanistic Data for Formaldehyde 4-428
4.5.3.2. Mode of Action Evaluation for Upper Respiratory Tract
Cancer (Nasopharyngeal Cancer, Sino-Nasal) 4-439
4.5.3.3. Mode(s) of Action for Lymphohematpoietic Malignancies 4-446
4.5.4. Hazard Characterization for Formaldlehyde Carcinogenicity 4-453
4.6. SUSCEPTIBLE POPULATIONS 4-454
4.6.1. Life Stages 4-454
4.6.1.1. Early Life Stages 4-455
4.6.1.2. Later Life Stages 4-459
4.6.1.3. Conclusions on Life-Stage Susceptibility 4-459
4.6.2. Health/Disease Status 4-460
4.6.3. Nutritional Status 4-461
4.6.4. Gender Differences 4-462
4.6.5. Genetic Differences 4-462
This document is a draft for review purposes only and does not constitute Agency policy.
Il-vi DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
4.6.6. Co-Exposures 4-464
4.6.6.1. Cumulative Risk 4-464
4.6.6.2. Aggregate Exposure 4-465
4.6.7. Uncertainties of Database 4-465
4.6.7.1. Uncertainties of Exposure 4-465
4.6.7.2. Uncertainties of Effect 4-466
4.6.8. Summary of Potential Susceptibility 4-467
VOLUME III
5. QUANTITATIVE ASSESSMENT: INHALATION EXPOSURE 5-1
5.1. INHALATION REFERENCE CONCENTRATION (RfC) 5-2
5.1.1. Candidate Critical Effects by Health Effect Category 5-3
5.1.1.1. Sensory Irritation of the Eyes, Nose, and Throat 5-3
5.1.1.2. Upper Respiratory Tract Pathology 5-5
5.1.1.3. Pulmonary Function Effects 5-6
5.1.1.4. Asthma and Allergic Sensitization (Atopy) 5-10
5.1.1.5. Immune Function 5-16
5.1.1.6. Neurological and Behavioral Toxicity 5-17
5.1.1.7. Developmental and Reproductive Toxicity 5-25
5.1.2. Summary of Critical Effects and Candidate RfCs 5-33
5.1.2.1. Selection of Studies for Candidate RfC Derivation 5-33
5.1.2.2. Derivation of Candidate RfCs from Key Studies 5-40
5.1.2.3. Evaluation of the Study-Specific Candidate RfC 5-66
5.1.3. Database Uncertainties in the RfC Derivation 5-69
5.1.4. Uncertainties in the RfC Derivation 5-72
5.1.5. Previous Inhalation Assessment 5-74
5.2. QUANTITATIVE CANCER ASSESSMENT BASED ON THE NATIONAL
CANCER INSTITUTE COHORT STUDY 5-74
5.2.1. Choice of Epidemiology Study 5-75
5.2.2. Nasopharyngeal Cancer 5-76
5.2.2.1. Exposure-Response Modeling of the National Cancer
Institute Cohort 5-76
5.2.2.2. Prediction of Lifetime Extra Risk of Nasopharyngeal Cancer
Mortality 5-79
5.2.2.3. Prediction of Lifetime Extra Risk of Nasopharyngeal Cancer
Incidence 5-81
5.2.2.4. Sources of Uncertainty 5-83
5.2.3. Lymphohematopoietic Cancer 5-88
5.2.3.1. Exposure-Response Modeling of the National Cancer
Institute Cohort 5-88
This document is a draft for review purposes only and does not constitute Agency policy.
II-vii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
5.2.3.2. Prediction of Lifetime Extra Risks for Hodgkin Lymphoma
and Leukemia Mortality 5-91
5.2.3.3. Prediction of Lifetime Extra Risks for Hodgkin Lymphoma
and Leukemia Incidence 5-93
5.2.3.4. Sources of Uncertainty 5-95
5.2.4. Conclusions on Cancer Unit Risk Estimates Based on Human Data 5-99
5.3. DOSE-RESPONSE MODELING OF RISK OF SQUAMOUS CELL
CARCINOMA IN THE RESPIRATORY TRACT USING ANIMAL DATA 5-102
5.3.1. Long-Term Bioassays in Laboratory Animals 5-104
5.3.1.1. Nasal Tumor Incidence Data 5-104
5.3.1.2. Mechanistic Data 5-105
5.3.2. The CUT Biologically Based Dose-Response Modeling 5-106
5.3.2.1. Major Results of the CUT Modeling Effort 5-111
5.3.3. This Assessment's Conclusions from Evaluation of Dose-Response
Models of DPX Cell-Replication and Genomics Data, and of BBDR
Models for Risk Estimation 5-111
5.3.4. Benchmark Dose Approaches to Rat Nasal Tumor Data 5-118
5.3.4.1. Benchmark Dose Derived from BBDR Rat Model and Flux
as Dosimeter 5-118
5.3.4.2. Comparison with Other Benchmark Dose Modeling Efforts 5-125
5.3.4.3. Kaplan-Meier Adjustment 5-128
5.3.4.4. EPA Time-to-Tumor Statistical Modeling 5-129
5.4. CONCLUSIONS FROM THE QUANTITATIVE ASSESSMENT OF
CANCER RISK FROM FORMALDEHYDE EXPOSURE BY INHALATION.. 5-133
5.4.1. Inhalation Unit Risk Estimates Based on Human Data 5-133
5.4.2. Inhalation Unit Risk Estimates Based on Rodent Data 5-133
5.4.3. Summary of Inhalation Unit Risk Estimates 5-135
5.4.4. Application of Age-Dependent Adjustment Factors (ADAFs) 5-136
5.4.5. Conclusions: Cancer Inhalation Unit Risk Estimates 5-137
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE-RESPONSE 6-1
6.1. SUMMARY OF HUMAN HAZARD POTENTIAL 6-1
6.1.1. Exposure 6-1
6.1.2. Absorption, Distribution, Metabolism, and Excretion 6-1
6.1.3. Noncancer Health Effects in Humans and Laboratory Animals 6-4
6.1.3.1. Sensory Irritation 6-4
6.1.3.2. Respiratory Tract Pathology 6-5
6.1.3.3. Effects on Pulmonary Function 6-8
6.1.3.4. Asthmatic Responses and Increased Atopic Symptoms 6-9
6.1.3.5. Effects on the Immune System 6-10
6.1.3.6. Neurological Effects 6-11
6.1.3.7. Reproductive and Developmental Effects 6-12
This document is a draft for review purposes only and does not constitute Agency policy.
II-viii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
6.1.3.8. Effects on General Systemic Toxicity 6-13
6.1.3.9. Summary 6-14
6.1.4. Carcinogenicity in Human and Laboratory Animals 6-14
6.1.4.1. Carcinogenicity in Humans 6-14
6.1.4.2. Carcinogenicity in Laboratory Animals 6-20
6.1.4.3. Carcinogenic Mode(s) of Action 6-21
6.1.5. Cancer Hazard Characterization 6-24
6.2. DOSE-RESPONSE CHARACTERIZATION 6-25
6.2.1. Noncancer Toxicity: Reference Concentration (RfC) 6-25
6.2.1.1. Assessment Approach Employed 6-25
6.2.1.2. Derivation of Candidate Reference Concentrations 6-25
6.2.1.3. Adequacy of Overall Data Base for RfC Derivation 6-26
6.2.1.4. Uncertainties in the Reference Concentration (RfC) 6-29
6.2.1.5. Conclusions 6-32
6.2.2. Cancer Risk Estimates 6-32
6.2.2.1. Choice of Data 6-32
6.2.2.2. Analysis of Epidemiologic Data 6-33
6.2.2.3. Analysis of Laboratory Animal Data 6-36
6.2.2.4. Extrapolation Aporoaches 6-37
6.2.2.5. Inhalation Unit Risk Estimates for Cancer 6-41
6.2.2.6. Early-Life Susceptibility 6-41
6.2.2.7. Uncertainties in the Quantitative Risk Estimates 6-42
6.2.2.8. Conclusions 6-45
6.3. SUMMARY AND CONCLUSIONS 6-45
REFERENCES R-l
VOLUME IV
APPENDIX A: SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITIONS A-l
APPENDIX B: SIMULATIONS OF INTERINDIVIDUAL AND ADULT-TO-CHILD
VARIABILITY IN REACTIVE GAS UPTAKE IN A SMALL SAMPLE
OF PEOPLE (Garcia et aL 2009) B-l
APPENDIX C: LIFETABLE ANALYSIS C-l
APPENDIX D: MODEL STRUCTURE & CALIBRATION IN CONOLLY ET AL.
(2003, 2004) D-l
APPENDIX E: EVALUATION OF BBDR MODELING OF NASAL CANCER IN THE
F344 RAT: CONOLLY ET AL. (2003) AND ALTERNATIVE
IMPLEMENTATIONS E-1
APPENDIX F: SENSITIVITY ANALYSIS OF BBDR MODEL FOR FORMALDEHYDE
INDUCED RESPIRATORY CANCER IN HUMANS F-l
This document is a draft for review purposes only and does not constitute Agency policy.
Il-ix DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
APPENDIX G: EVALUATION OF THE CANCER DOSE-RESPONSE MODELING
OF GENOMIC DATA FOR FORMALDEHYDE RISK ASSESSMENT G-l
APPENDIX H: EXPERT PANEL CONSULTATION ON QUANTITATIVE
EVALUATION OF ANIMAL TOXICOLOGY DATA FOR
ANALYZING CANCER RISK DUE TO INHALED FORMALDEHYDE.. H-l
This document is a draft for review purposes only and does not constitute Agency policy.
II-x DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES
Table 2-1. Physicochemical properties of formaldehyde 2-2
Table 2-2. Ambient air levels by land use category 2-6
Table 2-3. Studies on residential indoor air levels of formaldehyde (non-occupational) 2-8
Table 3-1. Endogenous formaldehyde levels in animal and human tissues and body fluids 3-8
Table 3-2. Formaldehyde kinetics in human and rat tissue samples 3-21
Table 3-3. Allelic frequencies of ADH3 in human populations 3-25
Table 3-4. Percent distribution of airborne [14C]-formaldehyde in F344 rats 3-26
Table 3-5. Apparent formaldehyde levels in exhaled breath of individuals attending a
health fair, adjusted for methanol and ethanol levels which contribute to the
detection of the protonated species with a mass to charge ratio of 31 reported
as formaldehyde (m/z = 31) 3-29
Table 3-6. Measurements of exhaled formaldehyde concentrations in the mouth and nose,
and in the oral cavity after breath holding in three healthy male laboratory
workers 3-30
Table 3-7. Extrapolation of parameters for enzymatic metabolism to the human 3-53
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC 4-59
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer 4-71
Table 4-3. Epidemiologic studies of formaldehyde and pharyngeal cancer (includes
nasopharyngeal cancer) 4-78
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers 4-98
Table 4-5. Respiratory effects of formaldehyde-induced reflex bradypnea in various
strains of mice 4-112
Table 4-6. Respiratory effects of formaldehyde-induced reflex bradypnea in various
strains of rats 4-113
Table 4-7. Inhaled dose of formaldehyde to nasal mucosa of F344 rats and B6C3F1
mice exposed to 15 ppm 4-116
Table 4-8. Exposure regimen for cross-tolerance study 4-117
This document is a draft for review purposes only and does not constitute Agency policy.
Il-xi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-9. Summary of formaldehyde effects on mucociliary function in the upper
respiratory tract 4-127
Table 4-10. Concentration regimens for ultrastructural evaluation of male CDF rat
nasoturbinates 4-129
Table 4-11. Enzymatic activities in nasal respiratory epithelium of male Wistar rats
exposed to formaldehyde, ozone, or both 4-130
Table 4-12. Lipid analysis of lung tissue and lung gavage from male F344 rats exposed
to 0, 15, or 145.6 ppm formaldehyde for 6 hours 4-138
Table 4-13. Formaldehyde effects on biochemical parameters in nasal mucosa and lung
tissue homogenates from male F344 rats exposed to 0, 15, or 145.6 ppm
formaldehyde for 6 hours 4-139
Table 4-14. Mast cell degranulation and neutrophil infiltration in the lung of rats
exposed to formaldehyde via inhalation 4-140
Table 4-15. Summary of respiratory tract pathology from inhalation exposures to
formaldehyde—short term studies 4-143
Table 4-16. Location and incidence of respiratory tract lesions in B6C3F1 mice
exposed to formaldehyde 4-146
Table 4-17. Formaldehyde effects (incidence and severity) on histopathologic changes in
the noses and larynxes of male and female albino SPF Wistar rats exposed to
formaldehyde 6 hours/day for 13 weeks 4-148
Table 4-18. Formaldehyde-induced nonneoplastic histopathologic changes in male
albino SPF Wistar rats exposed to 0, 10, or 20 ppm formaldehyde
and examined at the end of 130 weeks inclusive of exposure 4-149
Table 4-19. Formaldehyde-induced nasal tumors in male albino SPF Wistar rats
exposed to formaldehyde (6 hours/day, 5 days/week for 13 weeks) and
examined at the end of 130 weeks inclusive of exposure 4-150
Table 4-20. Formaldehyde effects on nasal epithelium for various concentration-by-
time products in male albino Wistar rats 4-153
Table 4-21. Rhinitis observed in formaldehyde-treated animals; data pooled for male
and female animals 4-154
This document is a draft for review purposes only and does not constitute Agency policy.
II-xii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-22. Epithelial lesions found in the middle region of nasoturbinates of
formaldehyde-treated and control animals; data pooled for males and
females 4-155
Table 4-23. Cellular and molecular changes in nasal tissues of F344 rats exposed to
formaldehyde 4-156
Table 4-24. Percent body weight gain and concentrations of iron, zinc, and copper in
cerebral cortex of male Wistar rats exposed to formaldehyde via inhalation
for 4 and 13 weeks 4-158
Table 4-25. Zinc, copper, and iron content of lung tissue from formaldehyde-treated
male Wistar rats 4-158
Table 4-26. Total lung cytochrome P450 measurements of control and formaldehyde-
treated male Sprague-Dawley rats 4-159
Table 4-27. Cytochrome P450 levels in formaldehyde-treated rats 4-160
Table 4-28. Summary of respiratory tract pathology from inhalation exposures to
formaldehyde, subchronic studies 4-162
Table 4-29. Histopathologic findings and severity scores in the naso- and
maxilloturbinates of female Sprague-Dawley rats exposed to inhaled
formaldehyde and wood dust for 104 weeks 4-166
Table 4-30. Histopathologic changes (including tumors) in nasal cavities of male
Sprague-Dawley rats exposed to inhaled formaldehyde or HC1 alone and
in combination for a lifetime 4-170
Table 4-31. Summary of neoplastic lesions in the nasal cavity of f344 rats exposed to
inhaled formaldehyde for 2 years 4-173
Table 4-32. Apparent sites of origin for the SCCs in the nasal cavity of F344 rats
exposed to 14.3 ppm of formaldehyde gas in the Kerns et al. (1983)
bioassay 4-174
Table 4-33. Incidence and location of nasal squamous cell carcinoma in male F344
rats exposed to inhaled formaldehyde for 2 years 4-175
Table 4-34. Summary of respiratory tract pathology from chronic inhalation exposures
to formaldehyde 4-183
Table 4-35. Cell proliferation in nasal mucosa, trachea, and free lung cells isolated
from male Wistar rats after inhalation exposures to formaldehyde 4-194
This document is a draft for review purposes only and does not constitute Agency policy.
II-xiii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-36. The effect of repeated formaldehyde inhalation exposures for 3 months on
cell count, basal membrane length, proliferation cells, and two measures of
cell proliferation, LI and ULLI, in male F344 rats 4-196
Table 4-37. Formaldehyde-induced changes in cell proliferation and (ULLI) in the nasal
passages of male F344 rats exposed 6 hours/day 4-198
Table 4-38. Cell population and surface area estimates in untreated male F344 rats and
regional site location of squamous cell carcinomas in formaldehyde
exposed rats for correlation to cell proliferation rates 4-199
Table 4-39. Summary of formaldehyde effects on cell proliferation in the upper
respiratory tract 4-202
Table 4-40. Summary of lesions observed in the gastrointestinal tracts of Wistar rats
after drinking-water exposure to formaldehyde for 4 weeks 4-206
Table 4-41. Incidence of lesions observed in the gastrointestinal tracts of Wistar rats
after drinking-water exposure to formaldehyde for 2 years 4-209
Table 4-42. Effect of formaldehyde on gastroduodenal carcinogenesis initiated by
MNNG and NaCl in male Wistar rats exposed to formaldehyde (0.5%
formalin) in drinking water for 8 weeks 4-212
Table 4-43. Summary of benign and malignant gastrointestinal tract neoplasia
reported in male and female Sprague-Dawley rats exposed to
formaldehyde in drinking water at different ages 4-214
Table 4-44. Incidence of hemolymphoreticular neoplasia reported in male and
female Sprague-Dawley rats exposed to formaldehyde in drinking water
from 7 weeks old for up to 2 years (experiment BT 7001) 4-215
Table 4-45. Battery of immune parameters and functional tests assessed in female
B6C3F1 mice after a 3 week, 15-ppm formaldehyde exposure 4-218
Table 4-46. Summary of the effects of formaldehyde inhalation on the mononuclear
phagocyte system (MPS) in female B6C3F1 mice after a 3-week, 15 ppm
formaldehyde exposure ( 6 hours/day, 5 days/week) 4-219
Table 4-47. Formaldehyde exposure regimens for determining the effects of
formaldehyde exposure on pulmonary S. aureus infection 4-221
Table 4-48. Summary of immune function changes due to inhaled formaldehyde
exposure in experimental animals 4-226
This document is a draft for review purposes only and does not constitute Agency policy.
II-xiv DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-49. Study design for guinea pigs exposed to formaldehyde through different
routes of exposure: inhalation, dermal, and injection 4-232
Table 4-50. Sensitization response of guinea pigs exposed to formaldehyde through
inhalation, topical application, or footpad injection 4-233
Table 4-51. Cytokine and chemokine levels in lung tissue homogenate supernatants in
formaldehyde-exposed male ICR mice with and without Der f sensitization 4-240
Table 4-52. Correlation coefficients among ear swelling responses and skin mRNA
levels in contact hypersensitivity to formaldehyde in mice 4-249
Table 4-53. Summary of sensitization and atopy studies by inhalation or dermal
sensitization due to formaldehyde in experimental animals 4-251
Table 4-54. Fluctuation of behavioral responses when male AB mice inhaled
formaldehyde in a single 2-hour exposure: effects after 2 hours 4-259
Table 4-55. Fluctuation of behavioral responses when male AB mice inhaled
formaldehyde in a single 2-hour exposure: effects after 24 hours 4-259
Table 4-56. Effects of formaldehyde exposure on completion of the labyrinth test by
male and female LEW. IK rats 4-263
Table 4-57. Summary of neurological and neurobehavioral studies in inhaled
formaldehyde in experimental animals 4-279
Table 4-58. Effects of formaldehyde on body and organ weights in rat pups from
dams exposed via inhalation from mating through gestation 4-289
Table 4-59. Formaldehyde effects on Leydig cell quantity and nuclear damage in adult
male Wistar rats 4-298
Table 4-60. Formaldehyde effects on adult male albino Wistar rats 4-299
Table 4-61. Formaldehyde effects on testosterone levels and seminiferous tubule
diameters in Wistar rats following 91 days of exposure 4-300
Table 4-62. Effects of formaldehyde exposure on seminiferous tubule diameter and
epithelial height in Wistar rats following 18 weeks of exposure 4-302
Table 4-63. Incidence of sperm abnormalities and dominant lethal effects in
formaldehyde-treated mice 4-302
This document is a draft for review purposes only and does not constitute Agency policy.
II-xv DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-64. Body weights of pups born to beagles exposed to formaldehyde during
gestation 4-303
Table 4-65. Testicular weights, sperm head counts, and percentage incidence of
abnormal sperm after oral administration of formaldehyde to male
Wistar rats 4-305
Table 4-66. Effect of formaldehyde on spermatogenic parameters in male Wistar rats
exposed intraperitoneally 4-306
Table 4-67. Incidence of sperm head abnormalities in formaldehyde-treated rats 4-307
Table 4-68. Dominant lethal mutations after exposure of male rats to formaldehyde 4-308
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation
exposure studies 4-311
Table 4-70. Summary of reported developmental effects in formaldehyde oral exposure
studies 4-317
Table 4-71. Summary of reported developmental effects in formaldehyde dermal
exposure studies 4-318
Table 4-72. Summary of reported reproductive effects in formaldehyde inhalation
studies 4-319
Table 4-73. Summary of reported reproductive effects in formaldehyde oral studies 4-322
Table 4-74. Summary of reported reproductive effects in formaldehyde intraperitoneal
studies 4-323
Table 4-75. Summary of chronic bioassays which address rodent leukemia and
lymphoma 4-329
Table 4-76. Formaldehyde-DNA reactions (DPX formation) 4-340
Table 4-77. Formaldehyde-DNA reactions (DNA adduct formation) 4-343
Table 4-78. Formaldehyde-DNA interactions (single stranded breaks) 4-344
Table 4-79. Other genetic effects of formaldehyde in mammalian cells 4-346
Table 4-80. In vitro clastogenicity of formaldehyde 4-348
Table 4-81. Summary of mutagenicity of formaldehyde in bacterial systems 4-350
This document is a draft for review purposes only and does not constitute Agency policy.
II-xvi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 4-82. Mutagenicity in mammalian cell systems 4-355
Table 4-83. Genotoxicity in laboratory animals 4-361
Table 4-84. MN frequencies in buccal mucosa cells of volunteers exposed to
formaldehyde 4-364
Table 4-85. MN and SCE formation in mortuary science students exposed to
formaldehyde for 85 days 4-364
Table 4-86. Incidence of MN formation in mortuary students exposed to formaldehyde
for 90 days 4-365
Table 4-87. Multivariate regression models linking genomic instability/occupational
exposures to selected endpoints from the MN assay 4-369
Table 4-88. Genotoxicity measures in pathological anatomy laboratory workers and
controls 4-370
Table 4-89. Summary of human cytogenetic studies 4-372
Table 4-90. Summary of cohort and case-control studies which evaluated the
incidence of all LHP cancers in formaldehyde-exposed populations
(ICD-8 Codes: 200-209) and all leukemias (ICD-8 Codes: 204-207) 4-412
Table 4-91. Secondary analysis of published mortality statistics to explore alternative
disease groupings within the broad category of all lymphohematopoetic
malignancies 4-419
Table 4-92. Summary of studies which provide mortality statistics for myeloid
leukemia sub-types 4-420
Table 4-93. Summary of mortality statistics for Hodgkin's lymphoma, lymphoma and
multiple myeloma from cohort analyses of formaldehyde exposed workers 4-422
Table 4-94. Summary of chronic bioassays which address rodent leukemia and
lymphoma 4-424
Table 4-95. Incidence of lymphoblastic leukemia and lymphosarcoma orally dosed in
Sprague-Dawley rats 4-425
Table 4-96. Available evidence for susceptibility factors of concern for formaldehyde
exposure 4-469
This document is a draft for review purposes only and does not constitute Agency policy.
II-xvii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 5-1. Points of departure (POD) for nervous system toxicity in key human and
animal studies 5-19
Table 5-2. Effects of formaldehyde exposure on completion of the labyrinth test by
male and female LEW. IK rats 5-23
Table 5-3. Developmental and reproductive toxicity PODs including duration
adjustments - key human study 5-31
Table 5-4. Summary of candidate studies for formaldehyde RfC development by
health endpoint category 5-36
Table 5-5. Adjustment for nonoccupational exposures to formaldehyde 5-64
Table 5-6. Summary of reference concentration (RfC) derivation from critical study and
supporting studies 5-68
Table 5-7. Relative risk estimates for mortality from nasopharyngeal malignancies
(ICD-8 code 147) by level of formaldehyde exposure for different
exposure metrics 5-78
Table 5-8. Regression coefficients from NCI log-linear trend test models for NPC
mortality from cumulative exposure to formaldehyde 5-79
Table 5-9. Extra risk estimates for NPC mortality from various levels of continuous
exposure to formaldehyde 5-80
Table 5-10. ECooos, LECqoos, and inhalation unit risk estimates for NPC mortality from
formaldehyde exposure based on the Hauptmann et al. (2004) log-linear
trend analyses for cumulative exposure 5-81
Table 5-11. ECooos, LECooos, and inhalation unit risk estimates for NPC incidence from
formaldehyde exposure based on the Hauptmann et al. (2004) trend
analyses for cumulative exposure 5-82
Table 5-12. Relative risk estimates for mortality from Hodgkin lymphoma
(ICD-8 code 201) and leukemia (ICD-8 codes 204-207) by level of
formaldehyde exposure for different exposure metrics 5-90
Table 5-13. Regression coefficients for Hodgkin lymphoma and leukemia mortality
from NCI trend test models 5-90
Table 5-14. Extra risk estimates for Hodgkin lymphoma mortality from various levels
of continuous exposure to formaldehyde 5-91
This document is a draft for review purposes only and does not constitute Agency policy.
II-xviii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
Table 5-15. Extra risk estimates for leukemia mortality from various levels of
continuous exposure to formaldehyde 5-91
Table 5-16. ECooos, LECooos, and inhalation unit risk estimates for Hodgkin lymphoma
mortality from formaldehyde exposure based on Beane Freeman et al.
(2009) log-linear trend analyses for cumulative exposure 5-93
Table 5-17. ECoos, LECoos, and inhalation unit risk estimates for leukemia mortality
from formaldehyde exposure based on Beane Freeman et al. (2009)
log-linear trend analyses for cumulative exposure 5-93
Table 5-18. ECooos, LECooos, and inhalation unit risk estimates for Hodgkin
lymphoma incidence from formaldehyde exposure, based on Beane
Freeman et al. (2009) log-linear trend analyses for cumulative exposure 5-94
Table 5-19. EC0os, LECqos, and inhalation unit risk estimates for leukemia incidence
from formaldehyde exposure based on Beane Freeman et al. (2009)
log-linear trend analyses for cumulative exposure 5-94
Table 5-20. Calculation of combined cancer mortality unit risk estimate at 0.1 ppm 5-100
Table 5-21. Calculation of combined cancer incidence unit risk estimate at 0.1 ppm 5-100
Table 5-22. Summary of tumor incidence in long-term bioassays on F344 rats 5-105
Table 5-23. BMD modeling of unit risk of SCC in the human respiratory tract 5-125
Table 5-24. Formaldehyde-induced rat tumor incidences 5-128
Table 5-25. Human benchmark extrapolations of nasal tumors in rats using
formaldehyde flux and DPX 5-134
Table 5-26. Summary of inhalation unit risk estimates 5-135
Table 5-27. Total cancer risk from exposure to a constant formaldehyde exposure
level of 1 |ig/m3 from ages 0-70 years 5-137
Table 6-1. Summary of candidate Reference Concentrations (RfC) for co-critical studies.... 6-27
Table 6-2. Effective concentrations (lifetime continuous exposure levels) predicted
for specified extra cancer risk levels for selected formaldehyde-related
cancers 6-36
Table 6-3. Inhalation unit risk estimates based on epidemiological and experimental
animal data 6-42
This document is a draft for review purposes only and does not constitute Agency policy.
II-xix DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
Figure 2-1. Chemical structure of formaldehyde 2-1
Figure 2-2. Locations of hazardous air pollutant monitors 2-5
Figure 2-3. Modeled ambient air concentrations based on 1999 emissions 2-7
Figure 2-4. Range of formaldehyde air concentrations (ppb) in different environments 2-9
Figure 3-1. Formaldehyde-mediated protein modifications 3-2
Figure 3-2. 3H/14C ratios in macromolecular extracts from rat tissues following exposure
to 14C and 3H-labeled formaldehyde (0.3, 2, 6, 10, 15 ppm) 3-18
Figure 3-3. Formaldehyde clearance by ALDH2 (GSH-independent) and ADH3
(GSH-dependent) 3-20
Figure 3-4. Metabolism of formate 3-22
Figure 3-5. Scatter plot of formaldehyde concentrations measured in ppb in direct breath
exhalations (x axis) and exhaled breath condensate headspace (y axis) 3-31
Figure 3-6. Reconstructed nasal passages of F344 rat, rhesus monkey, and human 3-36
Figure 3-7. Illustration of interspecies differences in airflow and verification of CFD
simulations with water-dye studies 3-37
Figure 3-8. Lateral view of nasal wall mass flux of inhaled formaldehyde simulated in
the F344 rat, rhesus monkey, and human 3-38
Figure 3-9. CFD simulations of formaldehyde flux to human nasal lining at different
inspiratory flow rates 3-39
Figure 3-10. Single-path model simulations of surface flux per ppm of formaldehyde
exposure concentration in an adult male human 3-43
Figure 3-11. Pressure drop vs. 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 3-45
Figure 3-12. Formaldehyde-DPX dosimetry in the F344 rat 3-47
Figure 4-1. Delayed asthmatic reaction following the inhalation of formaldehyde after
"painting" 100% formalin for 20 minutes 4-20
This document is a draft for review purposes only and does not constitute Agency policy.
II-xx DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continued)
Figure 4-2. Formaldehyde effects on minute volume in naive and formaldehyde-
pretreated male B6C3F1 mice and F344 rats 4-115
Figure 4-3. Sagittal view of the rat nose (nares oriented to the left) 4-121
Figure 4-4. Main components of the nasal respiratory epithelium 4-122
Figure 4-5. Decreased mucus clearance and ciliary beat in isolated frog palates
exposed to formaldehyde after 3 days in culture 4-126
Figure 4-6. Diagram of nasal passages showing section levels chosen for morphometry
and autoradiography in male rhesus monkeys exposed to formaldehyde 4-135
Figure 4-7. Formaldehyde-induced cell proliferation in male rhesus monkeys exposed to
formaldehyde 4-136
Figure 4-8. Formaldehyde-induced lesions in male rhesus monkeys exposed to formaldehyde
4-137
Figure 4-9. Frequency and location by cross-section level of squamous metaplasia in
the nasal cavity of F344 rats exposed to formaldehyde via inhalation 4-172
Figure 4-10. Effect of formaldehyde exposure on cell proliferation of the respiratory
mucosa of rats and mice 4-190
Figure 4-11. Alveolar MP Fc-mediated phagocytosis from mice exposed to 5 ppm
formaldehyde, 10 mg/m3 carbon black, or both 4-223
Figure 4-12. Compressed air in milliliters as parameter for airway obstruction
following formaldehyde exposure in guinea pigs after OVA sensitization and
OVA challenge 4-235
Figure 4-13. OVA-specific IgGl (IB) in formaldehyde-treated sensitized guinea pigs
prior to OVA challenge 4-235
Figure 4-14. Anti-OVA titers in female Balb/C mice exposed to 6.63 ppm
formaldehyde for 10 consecutive days, or once a week for 7 weeks 4-236
Figure 4-15. Vascular permeability in the tracheae and bronchi of male Wistar rats
after 10 minutes of formaldehyde inhalation 4-238
Figure 4-16. Effect of select receptor antagonists on formaldehyde-induced vascular
permeability in the trachea and bronchi of male Wistar rats 4-239
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continued)
Figure 4-17. The effects of formaldehyde inhalation exposures on eosinophil
infiltration (Panel A) and goblet cell proliferation (Panel B) after Der f
challenge in the nasal mucosa of male ICR mice after sensitization and
challenge 4-241
Figure 4-18. NGF in BAL fluid from formaldehyde-exposed female C3H/He mice
with and without OA sensitization 4-243
Figure 4-19. Plasma Substance P levels in formaldehyde-exposed female C3H/He
mice with and without OVA sensitization 4-244
Figure 4-20. Motor activity in male and female rats 2 hours after exposure to
formaldehyde expressed as mean number of crossed quadrants ± SEM 4-256
Figure 4-21. Habituation of motor activity was observed in control rats during the
second observation period (day 2, 24 hours after formaldehyde exposure) 4-257
Figure 4-22. Motor activity was reduced in male and female LEW. IK rats 2 hours
after termination of 10-minute formaldehyde exposure 4-258
Figure 4-23. The effects of the acute formaldehyde (FA) exposures on the
ambulatory and vertical components of SLMA 4-260
Figure 4-24. Effects of formaldehyde exposure on the error rate of female LEW. IK
rats performing the water labyrinth learning test 4-264
Figure 4-25. Basal and stress-induced trunk blood corticosterone levels in male
LEW. IK rats after formaldehyde inhalation exposures 4-269
Figure 4-26. NGF production in the brains of formaldehyde-exposed mice 4-274
Figure 4-27. Mortality corrected cumulative incidences of nasal carcinomas in the
indicated exposure groups 4-325
Figure 4-28. Leukemia incidence in Sprague-Dawley rats exposed to formaldehyde
in drinking water for 2 years 4-330
Figure 4-29. Unscheduled deaths in female F344 rats exposed to formaldehyde for
24 months 4-332
Figure 4-30. Cumulative leukemia incidence in female F344 rats exposed to
formaldehyde for 24 months 4-333
Figure 4-31. Cumulative incidence or tumor bearing animals for lymphoma in
female mice exposed to formaldehyde for 24 months 4-334
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continued)
Figure 4-32. DNA-protein cross-links (DPX) and thymidine kinase (tk) mutants in
TK6 human lymphoblasts exposed to formaldehyde for 2 hours 4-357
Figure 4-33. Developmental origins for cancers of the lymphohematopoietic system 4-409
Figure 4-34A. Association between peak formaldehyde exposure and the risk of
lymphohematopoietic malignancy 4-415
Figure 4-34B. Association between average intensity of formaldehyde exposure and
the risk of lymphohematopoietic malignancy 4-416
Figure 4-35. Effect of various doses of formaldehyde on cell number in (A) FIT-29
human colon carcinoma cells and in (B) human umbilical vein epithelial cells
(HUVEC) 4-433
Figure 4-36. Integrated MO A scheme for respiratory tract tumors 4-446
Figure 4-37. Location of intra-epithelial lymphocytes along side epithelial cells in
the human adenoid 4-450
Figure 5-1. Change in number of additions made in 10 minutes following formaldehyde
exposure at 32, 170, 390, or 890 ppb 5-21
Figure 5-2. Effects of formaldehyde exposure on the error rate of female LEW. IK rats
performing the water labyrinth learning test 5-24
Figure 5-3. Fecundity density ratio among women exposed to formaldehyde in the high
exposure index category with 8-hour time weighted average formaldehyde
exposure concentration of 219 ppb 5-27
Figure 5-4. Estimated reduction in peak expiratory flow rate (PEFR) in children in
relation to indoor residential formaldehyde concentrations 5-41
Figure 5.5. Odds ratios for physician-diagnosed asthma in children associated with in-
home formaldehyde levels in air 5-45
Figure 5-6. Prevalence of asthma and respiratory symptom scores in children associated
with in-home formaldehyde levels 5-48
Figure 5-7. Prevalence and severity of allergic sensitization in children associated
with in-home formaldehyde levels 5-49
Figure 5-8. Positive exposure-response relationships reported for in-home
formaldehyde exposures and sensory irritation (eye irritation) 5-53
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxiii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continued)
Figure 5-9. Positive exposure-response relationships reported for in-home
formaldehyde exposures and sensory irritation (burning eyes) 5-54
Figure 5-10. Age-specific mortality and incidence rates for myeloid, lymphoid, and
all leukemia 5-98
Figure 5-11. Schematic of integration of pharmacokinetic and pharmacodynamic
components in the CUT model 5-109
Figure 5-12. Fit to the rat tumor incidence data using the model and assumptions in
Conolly etal. (2003) 5-112
Figure 5-13. Spatial distribution of formaldehyde over the nasal lining, as
characterized by partitioning the nasal surface by formaldehyde flux to
the tissue per ppm of exposure concentration, resulting in 20 flux bins 5-120
Figure 5-14. Distribution of cells at risk across flux bins in the F344 rat nasal lining 5-120
Figure 5-15. MLE and upper bound (UB) added risk of SCC in the human nose for
two BBDR models 5-124
Figure 5-16. Replot of log-probit fit of the combined Kerns et al. (1983) and
Monticello et al. (1996) data on tumor incidence showing BMCio and
BMCLio 5-127
Figure 5-17. EPA multistate Weibull modeling: nasal tumor dose response 5-131
Figure 5-18. Multistage Weibull model fit 5-132
Figure 5-19. Multistage Weibull model fit of tumor incidence data compared with
KM estimates of spontaneous tumor incidence 5-132
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxiv DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS
ACGIH
American Conference of Governmental Industrial Hygienists
ADAF
age-dependent adjustment factors
ADH
alcohol dehydrogenase
ADS
anterior dorsal septum
AIC
Akaike Information Criterion
AIE
average intensity of exposure
AIHA
American Industrial Hygiene Association
ALB
albumin
ALDH
aldehyde dehydrogenase
ALL
acute lymphocytic leukemia
ALM
anterior lateral meatus
ALP
alkaline phosphatase
ALS
amyotrophic lateral sclerosis
ALT
alanine aminotransferase
AML
acute myelogenous leukemia
AMM
anterior medial maxilloturbinate
AMPase
adenosine monophosphatase
AMS
anterior medial septum
ANAE
alpha-naphthylacetate esterase
ANOVA
analysis of variance
APA
American Psychiatric Association
ARB
Air Resources Board
AST
aspartate aminotransferase
ATCM
airborne toxic control measure
ATP
adenosine triphosphate
ATPase
adenosine triphosphatase
ATS
American Thoracic Society
AT SDR
Agency for Toxic Substances and Disease Registry
AUC
area under the curve
BAL
bronchoalveolar lavage
BALT
bronchus associated lymphoid tissue
BBDR
biologically based dose response
BC
bronchial construction
BCME
bis(chloromethyl)ether
BDNF
brain-derived neurotrophic factor
BEIR
biologic effects of ionizing radiation
B£R
German Federal Institute for Risk Assessment
BHR
bronchial hyperresponsiveness
BMC
benchmark concentration
BMCL
95% lower bound on the benchmark concentration
BMCR
binuclated micronucleated cell ratefluoresce
BMD
benchmark dose
BMDL
95% lower bound on the benchmark dose
This document is a draft for review purposes only and does not constitute Agency policy.
II-XXV DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
BMR benchmark response
BN Brown-Norway
BrdU bromodeoxyuridine
BUN blood urea nitrogen
BW body weight
CA chromosomal aberrations
CalEPA California Environmental Protection Agency
CAP College of American Pathologists
CASRN Chemical Abstracts Service Registry Number
CAT catalase
CBMA cytokinesis-blocked micronucleus assay
CBMN cytokinesis-blocked micronucleus
CDC U.S. Centers for Disease Control and Prevention
CDHS California Department of Health Services
CFD computational fluid dynamics
CGM clonal growth model
CHO Chinese hamster ovary
CI confidence interval
CUT Chemical Industry Institute of Toxicology
CLL chronic lymphocytic leukemia
CML chronic myelogenous leukemia
CNS central nervous system
C02 carbon dioxide
COEHHA California Office of Environmental Health Hazard Assessment
CREB cyclic AMP responsive element binding proteins
CS conditioned stimulus
C x t concentration times time
DA Daltons
DAF dosimetric adjustment factor
DDX DNA-DNA cross-links
DEI daily exposure index
DEN diethylnitrosamine
Der f common dust mite allergen
DMG dimethylglycine
DMGDH dimethylglycine dehydrogenase
DNA deoxyribonucleic acid
DOPAC 3,4-dihydroxyphenylacetic acid
DPC / DPX DNA-protein cross-links
EBV Epstein-Barr virus
EC effective concentration
ED effective dose
EHC Environmental Health Committee
ELISA enzyme-linked immunosorbent assay
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxvi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
EPA
U.S. Environmental Protection Agency
ERPG
emergency response planning guideline
ET
ethmoid turbinates
FALDH
formaldehyde dehydrogenase
FDA
U.S. Food and Drug Administration
FDR
fecundability density ratio
FEF
forced expiratory flow
FEMA
Federal Emergency Management Agency
FEV1
forced expiratory volume in 1 second
FISH
fluorescent in situ hybridization
FSH
follicle-stimulating hormone
FVC
forced vital capacity
GALT
gut-associated lymphoid tissue
GC-MS
gas chromatography-mass spectrometry
GD
gestation day
GI
gastrointestinal
GO
gene ontology
G6PDH
glucose-6-phosphate dehydrogenase
GPX
glutathione peroxidase
GR
glutathione reductase
GM-CSF
granulocyte macrophage-colony-stimulating factor
GSH
reduced glutathione
GSNO
S-nitrosoglutathione
GST
glutathione S-transferase
HAP
hazardous air pollutant
Hb
hemoglobin
HC1
hydrochloric acid
HCT
hematocrit
HEC
human equivalent concentration
5-HI A A
5-hydroxyindoleacetic acid
hm
hydroxymethyl
HMGSH
S-hydroxymethylglutathione
HPA
hypothalamic-pituitary adrenal
HPG
hypothalamo-pituitary-gonadal
HPLC
high-performance liquid chromatography
HPRT
hypoxanthine-guanine phosphoribosyl transferase
HR
high responders
HSA
human serum albumin
HSDB
Hazardous Substances Data Bank
Hsp
heat shock protein
HWE
healthy worker effect
I cell
initiated cell
IARC
International Agency for Research on Cancer
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxvii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
ICD
International Classification of Diseases
IF
interfacial
IFN
interferon
Ig
immunoglobulin
IL
interleukin
LP.
intraperitoneal
IPCS
International Programme on Chemical Safety
IRIS
Integrated Risk Information System
Km
Michaels-Menton constant
KM
Kaplan-Meier
LD50
median lethal dose
LDH
lactate dehydrogenase
LEC
95% lower bound on the effective concentration
LED
95% lower bound on the effective dose
LHP
lymphohematopoietic
LI
labeling index
LM
Listeria monocytogenes
LMS
linearized multistage
LLNA
local lymph node assay
LOAEL
lowest-observed-adverse-effect level
LPS
lipopolysaccharide
LR
low responders
LRT
lower respiratory tract
MA
methylamine
MALT
mucus-associated lymph tissues
MCH
mean corpuscular hemoglobin
MCHC
mean corpuscular hemoglobin concentration
MCS
multiple chemical sensitivity
MCV
mean corpuscular volume
MDA
malondialdehyde
MEF
maximal expiratory flow
ML
myeloid leukemia
MLE
maximum likelihood estimate
MMS
methyl methane sulfonate
MMT
medial maxilloturbinate
MN
micronucleus, micronuclei
MNNG
N-methyl -N-nitro-N-nitrosoguani dine
MOA
mode of action
MoDC
monocyte-derived dendritic cell
MP
macrophage
MPD
multistage polynomial degree
MPS
mononuclear phagocyte system
MRL
minimum risk level
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxviii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
mRNA messenger ribonucleic acid
MVE-2 Murray Valley encephalitis virus
MVK Moolgavkar, Venzon, and Knudson
N cell normal cell
NaCl sodium choride
NAD+ nicotinamide adenine dinucleotide
NADH reduced nicotinamide adenine dinucleotide
NALT nasally associated lymphoid tissue
NATA National-Scale Air Toxics Assessment
NCEA National Center for Environmental Assessment
NCHS National Center for Health Statistics
NCI National Cancer Institute
NEG Nordic Expert Group
NER nucleotide excision repair
NGF nerve growth factor
NHL non-Hodgkin's lymphoma
NHMRC/ARMCANZ National Health and Medical Research Council/Agriculture and Resource
Management Council of Australia and New Zealand
NNK nitrosamine nitrosamine 4-(methylnitrosamino)-l-(3-pyridyl)-butanone
N6-hmdA N6-hydroxymethyldeoxyadenosine
N4-hmdC N4-hydroxymethylcytidine
N2-hmdG N2-hydroxymethyldeoxyguanosine
NICNAS National Industrial Chemicals Notification and Assessment Scheme
NIOSH National Institute for Occupational Safety and Health
NLM National Library of Medicine
NMDA N-methyl-D-aspartate
NO nitric oxide
NOAEL no-ob served-adverse-effect level
NPC nasopharyngeal cancer
NRBA neutrophil respiratory burst activity
NRC National Research Council
NTP National Toxicology Program
OR odds ratio
OSHA Occupational Safety and Health Administration
OTS Office of Toxic Substances
OVA ovalbumin
PBPK physiologically based pharmacokinetic
PC Philadelphia chromosome
PCA passive cutaneous anaphylaxis
PCMR proportionate cancer mortality ratio
PCNA proliferating cell nuclear antigen
PCR polymerase chain reaction
PCV packed cell volume
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxix DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
PEC AM
platelet endothelial cell adhesion molecule
PEF
peak expiratory flow
PEFR
peak expiratory flow rates
PEL
permissible exposure limit
PFC
plaque-forming cell
PG
peri glomerular
PHA
phytohemagglutinin
PLA2
phospholipase A2
PI
phagocytic index
PLM
posterior lateral meatus
PMA
phorbol 12-myristate 13-acetate
PMR
proportionate mortality ratio
PMS
posterior medial septum
PND
postnatal day
POD
point of departure
POE
portal of entry
PTZ
pentilenetetrazole
PUFA
polyunsaturated fatty acids
PWULLI
population weighted unit length labeling index
RA
reflex apnea
RANTES
regulated upon activation, normal T-cell expressed and secreted
RB
reflex bradypnea
RBC
red blood cells
RD50
exposure concentration that results in a 50% reduction in respiratory rate
REL
recommended exposure limit
RfC
reference concentration
RfD
reference dose
RGD
regional gas dose
RGDR
regional gas dose ratio
RR
relative risk
RT
reverse transcriptase
SAB
Science Advisory Board
see
squamous cell carcinoma
SCE
sister chromatid exchange
SCG
sodium cromoglycate
SD
standard deviation
SDH
succinate dehydrogenase; sarcosine dehydrogenase
SEER
Surveillance, Epidemiology, and End Results
SEM
standard error of the mean
SEN
sensitizer
SH
sulfhydryl
SHE
Syrian hamster embryo
SLMA
spontaneous locomotor activity
This document is a draft for review purposes only and does not constitute Agency policy.
II-XXX DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
SMR
standardized mortality ratio
SNP
single nucleotide polymorphism
SOD
superoxide dismutase
SOMedA
N6-sulfomethyldeoxy adenosine
Spl
specificity protein
SPIR
standardized proportionate incidence ratio
SSAO
semicarbozole-sensitive amine oxidase
SSB
single strand breaks
STEL
short-term exposure limit
TBA
tumor bearing animal
TH
T-lymphocyte helper
THF
tetrahydrofolate
TK
toxicokinetics
TL
tail length
TLV
threshold limit value
TNF
tumor necrosis factor
TP
total protein
TRI
Toxic Release Inventory
TRPV
transient receptor potential vanilloid
TWA
time-weighted average
TZCA
thiazolidine-4-carboxylate
UCL
upper confidence limit
UDS
unscheduled DNA synthesis
UF
uncertainty factor
UFFI
urea formaldehyde foam insulation
ULLI
unit length labeling index
URT
upper respiratory tract
USD A
U.S. Department of Agriculture
VC
vital capacity
VOC
volatile organic compound
WBC
white blood cell
WDS
wet dog shake
WHO
World Health Organization
WHOROE
World Health Organization Regional Office for Europe
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxxi DRAFT—DO NOT CITE OR QUOTE
-------
This page intentionally left blank.
This document is a draft for review purposes only and does not constitute Agency policy.
II-xxxii DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4. HAZARD CHARACTERIZATION
4.1. HUMAN STUDIES
This chapter discusses epidemiologic studies of site-specific cancers and other adverse
health effects that may be caused by exposure to formaldehyde. The primary focus is on the
literature describing inhalation exposure and its potential carcinogenic and noncarcinogenic
health risks. In addition, oral, dermal, and ocular exposures to formaldehyde are discussed.
The noncancer health effects section is organized by endpoint, beginning with sensory
irritation and followed by pulmonary function, asthma, respiratory tract pathology, immunologic
responses, neurological and behavioral responses, and, finally, developmental and reproductive
outcomes.
The carcinogenicity section is divided into two parts, respiratory tract and non-respiratory
tract cancers. The first part discusses site-specific cancers that are chiefly located in the
respiratory tract where direct contact with formaldehyde occurs: nasopharyngeal cancers (NPCs),
nasal and paranasal cancers, other respiratory tract cancers, and lung cancers. The second part
on non-respiratory tract cancer discusses those cancers at other sites with more distant exposure
to formaldehyde than respiratory epithelium—mainly, lymphohematopoietic (LHP) cancer, brain
and central nervous system (CNS) cancer, pancreatic cancer, and cancer at other sites.
4.1.1. Noncancer Health Effects
4.1.1.1. Sensory Irritation (Eye, Nose, Throat Irritation)
As a reactive gas, formaldehyde is a sensory irritant. Sensory irritation of the eyes and
respiratory tract by formaldehyde has been observed consistently in clinical and epidemiologic
studies in residential and occupational populations. Binding to sensory nerves at the portal of
entry (POE) results in direct sensory responses (e.g., detection of odor and tissue irritation) as
well as reflex responses to the sensory irritation and neurogenic sensitization. Reflex responses
result from CNS stimulation by the afferent sensory signals and include lacrimation, coughing,
sneezing, and bronchial constriction (BC). An additional reflex seen in rodents is reflex
bradypnea (RB) (also known as reflex apnea [RA]). Formaldehyde-induced sensory irritation
may be evident after acute exposures as well as in chronically exposed individuals.
Formaldehyde-induced neurogenic sensitization and atopy may result in lifelong health effects
from short-term or transient exposures. For this discussion, sensory irritation will include both
direct sensory response to formaldehyde exposure and reflex responses (lacrimation, coughing,
sneezing, RB, and BC, and sensitization.
This document is a draft for review purposes only and does not constitute Agency policy.
4-1 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Eye, nose, and throat irritation in response to formaldehyde inhalation exposure is well
documented (Doty et al., 2004). Broadly, studies examining these endpoints are either controlled
chamber studies with a defined population (e.g., healthy volunteers or sensitive individuals),
worker/student studies, or population (e.g., residential) studies. Chamber studies, by design, are
acute studies, although some researchers have investigated repeated exposures. Occupational,
student, and residential exposures are generally longer duration, although there is variability in
exposure and duration among subjects. Endpoints include both local effects and reflex effects of
sensory irritation. The endpoints for assessing irritation include self-reporting of adverse
symptoms (e.g., pain, burning, itching) as well as objective measures of irritation (e.g., eye-blink
counts, lacrimation) (Doty et al., 2004). The following review focuses on eye, nose, and throat
irritation but studies have documented other types of irritation, including dermal irritation
eczema and dermatitis.
4.1.1.1.1. Epidemiologic literature. A wide variety of epidemiologic studies have assessed the
potential effects of exposure to formaldehyde on endpoints, indicating sensory irritation of the
eye, nose, and throat. These studies generally include three different types of exposure
populations: (1) Residents and visitors exposed to formaldehyde in homes and mobile buildings,
where formaldehyde is present from various sources, including building components, furniture
and home furnishings, heating and cooking combustion as well as active and passive smoking;
(2) various occupational exposures from industrial processes related to wood products, furniture
making, and formaldehyde-based resins; and (3) anatomy students who are exposed under well-
defined conditions during academic courses where they are examining formaldehyde-preserved
cadavers.
4.1.1.1.1.1. Residential epidemiology. Among the residential epidemiology studies of
formaldehyde effects on sensory irritation, one of the strongest studies based on study design,
execution, analysis, and sample size was the observational study undertaken by Ritchie and
Lehnen (1987). In this cross-sectional study of nearly 2,000 Minnesota residents living in 392
mobile and 494 conventional homes, personal data and formaldehyde samples were collected
from residents that had responded to an offer by the state health department to test homes for
formaldehyde. Technicians administered a symptom questionnaire to participating residents at
the time of formaldehyde sample collection. Residents were asked to close doors and windows
of their homes for 12 hours before testing was conducted, and a standardized collection protocol
was used for both sample collection and analysis. Measurements of formaldehyde exposure
were taken from two rooms of the home, usually the bedroom and living room, and were kept
This document is a draft for review purposes only and does not constitute Agency policy.
4-2 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
refrigerated until analysis. Respondents were not aware of the results of the formaldehyde
analyses in their homes at the time they responded to the symptom questionnaire. The results
from Ritchie and Lehnen (1987) provide a clear dose-response relationship in the percentage of
residential occupants reporting eye, nose, and throat irritation. Specifically, eye irritation
responses increase from 1-2% in homes with formaldehyde concentrations lower than 0.1 ppm
to 12-32% in homes with formaldehyde concentrations ranging from 0.1 to 0.3 ppm with 86-
93% of residents reporting . These effects were found in the same concentration range for
people living in either mobile (n = 851) or conventional (n = 1,156) homes. Similar percentages
were found for nose/throat irritation. Reports of irritation were reported for smokers, passive
smokers, and nonsmokers with higher percentages of irritation among smokers, followed by
passive smokers and then nonsmokers. While the participants in this study were self-selected
and not a random residential sample, a clear concentration response was observed, and, even if
participants sought testing because they suspected that they were being exposed to formaldehyde
in their homes, they could not know the measured concentration of formaldehyde when reporting
their irritation symptoms, so recall bias cannot explain the concentration response. Neither can
confounding be an alternative explanation since the authors reported that formaldehyde was the
most important explanatory variable for all the sensory irritation effects of the eye, nose, and
throat.
The results of an adverse association of sensory irritation with formaldehyde reported by
Ritchie and Lehnen (1987) are corroborated by Hanrahan et al. (1984) who conducted a cross-
sectional survey by using a random sample of mobile homes from mobile home parks in
Wisconsin. Sixty-one teenage and adult residents participated. Health questionnaires were self-
administered by each occupant. Respondents were blinded to the results of their home
formaldehyde vapor measurements, which were sampled from two rooms in the homes following
instruction to close windows, refrain from smoking, and turn off gas appliances for 30 minutes
prior to air sampling. Logistic regression analyses were used to ascertain potential symptom risk
ratio dependency on each respondent's age, smoking status, gender, and formaldehyde
concentration measures in the home. Formaldehyde concentrations ranged from 0.1 ppm to
0.8 ppm with a geometric mean of 0.16 ppm. Across this concentration range, a clear and
statistically significant concentration-response relationship was reported in graphical form,
controlling for age, gender, and smoking status. At 0.1 ppm, the regression model showed less
than 5% predicted prevalence of burning eyes. At 0.2 ppm, the midpoint of the exposure
category in Ritchie and Lehnen (1987) that was reported to be the lowest adverse effect level for
eye irritation with 12-32% reporting eye irritation, the regression model of Hanrahan et al.
(1984) showed approximately 17.5% predicted prevalence of burning eyes. The prevalence of
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
burning eyes rose linearly to approximately 65% prevalence at 0.5 ppm, with some diminishment
in the rate of rise up to approximately 80% prevalence at 0.8 ppm. While only 65 out of 208
randomly selected homes volunteered to complete the health questionnaires, the investigators
were able to complete home formaldehyde vapor measurements on all the homes and reported
nearly an identical distribution of formaldehyde concentrations in participating and
nonparticipating homes. Demographic characteristics of some of the non-respondents were
available and were reported as nearly identical to those of participants. There was no indication
of selection bias. Confounding is unlikely to explain such a strong concentration response.
These findings of associations of sensory irritation with residential exposures to
formaldehyde are further supported by studies that did not examine concentration response but
nonetheless assessed the association of formaldehyde with sensory irritation. Similar findings to
those of Ritchie and Lehnen (1987) and Hanrahan et al. (1984) have been reported in other
residential studies of increased symptoms in association with formaldehyde exposure (Liu et al.,
1991; Thun et al., 1982; Dally et al., 1981). Dally et al. (1981) collected data in 100 "complaint
structures" (65% mobile homes, 27% conventional homes). Of these, 60% were from home
owners contacting the health department and 30% from physician referrals. Twenty percent of
the buildings had concentrations below the limit of detection (0.1 ppm), 20% had levels at or
above 0.81 ppm, and overall the concentrations ranged from below detection to above 3 ppm
with an overall median of 0.35 ppm. The median levels were 0.47 and 0.10 ppm for mobile and
conventional homes, respectively. No other contaminants were measured. Eye, nose, and throat
irritation were reported in a high percentage of occupants (eye irritation 68%, burning eyes 60%,
runny nose 60%, dry or sore throat 57%, cough 51%), but these were not reported as a function
of dose or home type. Thus, there was no control group to which rates of irritation could be
compared. However, symptoms reportedly stopped in 89% of occupants when they left the
"complaint structure." The most recent residential study was performed on over 1,000 mobile
homes with 1,394 participants (Liu et al., 1991). Home formaldehyde concentration ranged from
below 0.01 to 0.46 ppm. Analyses used logistic regression to control for potential confounders.
Eye irritation was positively associated with formaldehyde with a clear concentration response
demonstrated with cumulative exposure. During the summer and winter months, formaldehyde
exposure was associated with burning eyes. In the winter months, formaldehyde exposure was
associated with sore throat. There was no association of formaldehyde exposure with cough or
running nose during either season. Liu et al. (1991) also report a synergistic effect on irritation
by formaldehyde exposure and chronic disease prevalence. Thun et al. (1982) reported increased
symptoms of itchy skin and "wheezing and difficulty breathing" in residents in 395 homes
insulated with urea-formaldehyde foam relative to nearby homes without urea foam
This document is a draft for review purposes only and does not constitute Agency policy.
4-4 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
formaldehyde insulation (UFFI); however, there were no measurements of formaldehyde
concentration taken in this study.
While not strictly a residential epidemiology study, Olsen and Dossing (1982) studied
occupational exposures within mobile and non-mobile daycare centers. They reported the mean
concentration in mobile and non-mobile day care centers were 350 (200-450) ppb and 65 (40-
90) ppb, respectively. Adverse eye, nose, and throat irritation were significantly elevated in the
workers (n = 70) in the mobile units as compared with those in non-mobile units (n = 34). The
authors also state that a high percentage of workers in the mobile day cares reported that the
symptoms disappeared after working hours; however, the authors did not report any such
percentages among those working in non-mobile units.
4.1.1.1.1.2. Occupational epidemiology. Horvath et al. (1988) compared irritation symptoms
between 109 workers at a particleboard manufacturing plant and 264 workers at food plants as a
control group. The mean 8-hour time-weighted average (TWA) formaldehyde concentrations
between these two groups were 0.69 ppm (range 0.17-2.93) and 0.05 ppm (range 0.03-0.12),
respectively. Eye, nose, and throat irritation were more common among the former group
(prevalence of symptoms during a work shift: throat sore or burning—test 22.0%, controls 3.9%;
cough—test 34.9%), controls 18.9%>; burning of nose—test 28.4%>, controls 2.0%>; stuffy nose—
test 33.9%o, controls 14.2%>; itching of nose—test 21.1%, controls 7.9%; eyes burning or
watering—test 39.5%, controls 9.1%; eyes itching—test 19.3%, controls 7.1%).
Similar results were reported for frequency of eye and nasal discomfort in a group of
workers involved in the manufacture of formaldehyde resins. These workers were exposed to a
mean concentration of 0.40 mg/m3. Alexandersson and Hedenstierna (1988) reported that the
frequency of eye, nose, and throat irritation was significantly greater in 38 workers exposed to
formaldehyde and solvents in lacquers (average employment duration 7.8 years) as compared
with 18 controls (nonexposed individuals working at the same factory). The frequency of eye
irritation was 65.8% among those exposed and 16.7% among controls. No controls reported
nose/throat irritation, but about 40% of those exposed did.
A Swedish study conducted at a chemical plant found nasal and eye discomfort were
reported by 64 and 24%, respectively, of workers (n = 70) exposed to formaldehyde (range 0.05-
0.50 mg/m3 with a mean of 0.26 mg/m3) versus 25 and 6%, respectively, of the control group
made up of clerks from the local government (n = 36). In addition, the majority of workers
exposed to formaldehyde reported that their symptoms were relieved during weekends and
vacations (Holmstrom and Wilhelmsson, 1988). Another study by the same authors
(Wilhelmsson and Holmstrom, 1992) reported similar results. In this study irritation prevalent
This document is a draft for review purposes only and does not constitute Agency policy.
4-5 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
among 66 workers from a formaldehyde-producing plant was compared with that seen among
36 community clerks. The workers were exposed to 0.26 mg/m3 of formaldehyde (range 0.05-
0.6 mg/m3). The clerks were exposed to an average of 0.09 mg/m3. Nasal and eye discomfort
were reported at rates of 53 and 24%, respectively, among the workers. Among the community
clerks, 3 and 6%, respectively, reported discomfort.
Holness and Nethercott (1989) reported significant increases in eye irritation (42 versus
21%) and nose irritation (44 versus 16%) among 84 funeral service workers as compared with
38 controls (students and individuals from a service organization). The former group had been
actively embalming for approximately 10 years and had nearly twice the pack-years smoked as
the controls. The exposure concentration in both groups was 0.36 and 0.02 ppm, respectively.
4.1.1.1.1.3. Epidemiology on laboratory students. Several studies have monitored sensory
irritation in medical/physical therapy students exposed to formaldehyde during anatomy courses.
These studies have particular advantages: the student population generally has no former
occupational exposure, and, oftentimes, pre-class survey data serve as the control, providing a
better basis for assessing the effects of formaldehyde exposure.
In a study of 24 formaldehyde-exposed anatomy students (personal breathing zone
samples 0.73 ppm, range 0.49-0.93), the prevalence of eye irritation before the start of a cadaver
dissection class was 16%, while after the class, the prevalence was 59%. The increase in eye
irritation was most pronounced, but increases were also observed in the prevalence of irritation
of the nose (21%) and throat (15%) (Kriebel et al., 1993). The authors also reported a tendency
for this increase in intensity between the beginning and end of class to diminish over the
10-week course, especially for eye irritation. However, although the intensity of the irritation
diminished, eye irritation was still present among the students after 10 weeks of intermittent
exposure. The report of increase in post- versus pre-class irritation symptoms in this study was
no greater for asthmatic students (n = 5) compared with non-asthmatic students.
Takahashi et al. (2007) showed that 143 medical students reported various symptoms
(including eye and throat irritation) and that the percentage of students reporting symptoms
increased between the beginning and end of the course 2 months later. After the first day of
class, approximately 35% of students reported eye soreness and about 15% reported throat
irritation. After the course ended, these rates were close to 70% for eye soreness and slightly
above 40% for throat irritation. The reported average room formaldehyde concentration was
2.12 ppm (range 1.7-2.4), while the gas samplers worn on the students' chests averaged 2.4 ppm
(range 1.8-3.8). Another study of students in an anatomy laboratory class in Japan (Takigawa et
al., 2005) measured formaldehyde concentrations and irritation symptoms before and after the
This document is a draft for review purposes only and does not constitute Agency policy.
4-6 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
installation of a ventilation system. This system reduced the median personal formaldehyde
exposure concentration from 2.7 to 0.72 ppm. Before installation of the ventilation system, the
students complained about exacerbation of all the sensory irritation symptoms on average. The
increase in 8 out of 25 symptoms was significantly reduced after installing general ventilation
(p < 0.05). After installation of the ventilation system, a dose-dependent relationship with
formaldehyde was seen for irritated eyes but not for itchy nose.
Akbar-Khanzadeh et al. (1994) detected mean personal area levels of formaldehyde at
1.24 ppm and a range of 0.1-2.94 ppm from personal air sampling devices. Almost 90% of the
students in this study reported eye irritation, 74% reported nose irritation, and close to 30%
reported throat irritation during or after exposure to formaldehyde during the laboratory period
after having completed at least 6 weeks of laboratory sessions with formaldehyde exposure. In
addition, Uba et al. (1989) demonstrated that symptoms of eye, nose, and throat irritation were
correlated with formaldehyde exposure among medical students by comparing students'
responses on a questionnaire completed after a lab with formaldehyde exposure to a
questionnaire completed after a lab with no formaldehyde exposure. The authors compared
questionnaires completed prior to students' first anatomy lab to a questionnaire completed
7 months later. Reports of cough were more frequent after the 7 months. These students were
exposed to a mean level of 1.9 ppm (range 0.1-5.0) while dissecting (measured using portable
infrared spectrophotometer), and a TWA from all laboratory activities ranged from below limits
of detection to 0.93 (measured using personal sampling devices in the students' breathing zones).
4.1.1.1.2. Acute studies: controlled chamber exposures. Results from controlled human studies
demonstrate eye, nose, and throat irritation in association with formaldehyde exposure (Lang et
al., 2008; Yang et al., 2001; Krakowiak et al., 1998; Kulle, 1993; Green et al., 1989, 1987; Kulle
et al., 1987; Sauder et al., 1987, 1986; Schachter et al., 1987, 1986; Witek et al., 1987; Day et al.,
1984; Bender et al., 1983; Anderson et al., 1983; Weber-Tschopp et al., 1977; Andersen, 1979;
Schuck et al., 1966). A key advantage of chamber studies is the ability to monitor and closely
control formaldehyde concentrations during exposure. However, chamber studies may also be
limited by other aspects of the study design, including small number of participants, use of
healthy volunteers, short exposure durations (a few minutes), and often studies were conducted
with only one exposure group and at relatively high concentrations (>1 ppm). The lack of
multiple exposure levels in many studies limits the understanding of exposure-response
relationships. Additionally, numerous reports that demonstrate multiple symptoms of eye, nose,
and throat irritation at levels at or above 1 ppm did not explore lower levels of exposure and can
This document is a draft for review purposes only and does not constitute Agency policy.
4-7 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
only be used for primary hazard identification (Yang et al., 2001; Green et al., 1989, 1987;
Sauder et al., 1987, 1986; Schachter et al., 1987, 1986; Witek et al., 1987; Day et al., 1984).
The National Aeronautical and Space Administration conducted experiments in closed-
environment living, including environmental monitoring and air quality. James et al. (2002)
quantified air pollutants, including formaldehyde, during 30, 60, and 90-day tests in a closed
chamber study of a Lunar-Mars life support chamber. Unfortunately, the detection methods used
during the 30-day test were not sensitive enough to detect formaldehyde at levels below 2
mg/m3. Thus, badge samples were obtained in the 60-day and 90-day tests and provided greater
detection sensitivity (to 0.02 mg/m3). Measured values of formaldehyde increased over time. In
the 60-day test, formaldehyde levels were well above accepted limits (data not shown). Health
effects data are limited since there were only four crew members. One crew member reported
eye and upper airway irritation at formaldehyde concentrations of 0.25 mg/m3 (308 ppb) on day
15. It should also be noted that astronauts are exceptionally healthy individuals, and these data
should be interpreted carefully when determining expected health effects in the general
population. The experimenters determined that formaldehyde levels increased as temperature
increased. Formaldehyde was also linked to murals lining the chamber and was subsequently
removed before executing the 90-day study. Between days 0 and 60, formaldehyde levels
remained between 0.02 and 0.04 mg/m3, with one sharp peak that occurred at day three to 0.07
mg/m3. Between days 60 and 90, formaldehyde concentrations increased to 0.07 to 0.09 mg/m3.
The increase was attributed to an incomplete oxidation of methanol in a catalytic bed rather than
in excessive off-gassing of formaldehyde. No crew members reported any adverse effects in the
90-day study.
A few studies have been conducted that specifically address sensitive populations
(asthmatics) and/or individuals during exercise, which can exacerbate asthma (further details of
these studies are in Section 4.1.1.3, Effects on Asthmatics). In Sauder et al. (1986), 8-minute
bicycle exercise was completed multiple times during the exposure period (3 hours). However,
irritation symptoms were only reported after 2 hours of exposure and do not address whether
changes occurred during the periods of exercise. Overall, reports of eye, nose, and throat
irritation increased with exposure to formaldehyde (3 ppm) compared with reports of irritation
with no exposure to formaldehyde. Green et al. (1987) report that eye, nose, and throat irritation
symptoms were greater immediately after exercise during exposure to 3 ppm formaldehyde.
Additionally, the response levels were similar between asthmatic (n = 16) and non-asthmatic
(n = 22) subjects. Similar effects of exercise on certain symptoms, such as throat irritation, were
reported in 15 asthmatic subjects exposed to 2 ppm formaldehyde at rest and after exercise
(Witek etal., 1987).
This document is a draft for review purposes only and does not constitute Agency policy.
4-8 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
Kulle (1993) and Kulle et al. (1987) enrolled 19 healthy volunteers and exposed them to a
range of formaldehyde concentrations. At 2 ppm, 53% reported mild or moderate eye irritation
(32% mild, 21% moderate). At 3 ppm, 100% of subjects exposed at this level (n = 9) reported
irritation. The reported increase in irritation was shown to correspond with increasing
formaldehyde concentration in a linear fashion. Mild nose/throat irritation was present among
37%) of those exposed to 2 ppm of formaldehyde. Odor detection was very similar to the
distribution seen for eye irritation. Nineteen subjects performed light to moderate exercise while
exposed to 2 ppm; there was no increase in report of eye irritation, but nose/throat irritation did
increase. The data were reanalyzed (Kulle, 1993), and thresholds for irritation were found to be
0.5-1 ppm for eye irritation and 1 ppm for nose/throat irritation.
Yang et al. (2001) reported that eight individuals exposed to varying levels of
formaldehyde (1.65, 2.99, and 4.31 ppm) had mild to moderate eye irritation during the 5-minute
exposures. The increase in irritation was detected at 30 seconds with exposure to 1.65 ppm of
formaldehyde. The highest severity ratings at this concentration occurred between 60 and
90 seconds. Frequency of eye blinking was also measured. The peak in blinking rate occurred
after about 1 minute of exposure and then decreased almost back to a normal rate after 5 minutes
of exposure. Higher formaldehyde concentrations were associated with increased frequency of
blinking compared with the 1.65 ppm exposure.
Other studies have examined responses across multiple exposure levels. For example,
Weber-Tschopp et al. (1977) used two different methods of studying irritation resulting from
formaldehyde exposure. For one, they exposed subjects (n = 33) to an increasing level of
formaldehyde (maximum exposure was 3.2 ppm). This design precluded evaluation of distinct
effects at different exposure levels. The researchers addressed this by examining another group
of subjects (n = 48) that were exposed to 0, 1, 2, 3, or 4 ppm five times for 90 seconds. Levels of
nasal and throat irritation for this discontinuous exposure were slightly higher than the irritation
levels reported among those with continuous exposure. However, this was reversed for eye
irritation; those with continuous exposure reported higher levels of irritation than those with
discrete exposures. An objective measure, eye-blinking rate, was measured for those with
continuous exposure and was found to have a statistically significant increase at 1.7 ppm.
Bender et al. (1983) conducted a study that enrolled individuals who "responded" to
formaldehyde at 1.3 and 2.2 ppm and did not report irritation to the clean air control. They
found that, among these subjects, exposure to 1 ppm of formaldehyde (n = 27) resulted in the
reporting of eye irritation with a median response time of 78 seconds. Reports of irritation were
given as less than slightly irritating for formaldehyde concentrations of 0.3-0.9 ppm.
This document is a draft for review purposes only and does not constitute Agency policy.
4-9 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
Assessment of sensory irritation for pain and discomfort often relies on self-reporting,
using symptom questionnaires and severity ratings (e.g., mild, moderate, severe). In the case of
formaldehyde, subjective ratings of eye irritation correlate positively with eye-blinking
frequency (Lang et al., 2008). Lang et al. (2008) saw an increase in eye blinking after
195 minutes of exposure to formaldehyde at 0.5 ppm with four peak exposures of 1 ppm. After
this amount of time and formaldehyde exposure, there was also an increase in moderate eye
redness. Weber-Tschopp et al. (1977) reported that, among concentrations ranging from 0.03 to
3.2 ppm, eye-blinking frequency was increased at 1.7 ppm; similarly Yang et al. (2001) reported
increased blinking at >1.5 ppm (the lowest concentration examined). There are studies that
suggest that psychological factors (e.g., anxiety) can impact the perception of irritation—and
perhaps more so at lower concentrations (Lang et al., 2008; Ihrig et al., 2006; Dalton, 2003).
However, when Lang et al. (2008) controlled for mood prior to exposure, subjective symptoms
of eye, nasal, and olfactory irritation were significantly related to exposure (0.5 ppm)
Schuck et al. (1966) performed a study that also examines self-reported eye irritation as
well as blinking rate. Fourteen individuals were exposed to formaldehyde concentrations
ranging from 0 to 1 ppm. Increased irritation was reported with increasing formaldehyde
concentration. One subject, judged to be the least sensitive, was still able to detect formaldehyde
levels as low as 0.01 ppm. In addition, the authors examined the blinking rate of participants,
which they found was related to irritation intensity.
Andersen (1979) and Anderson and Molhave (1983) reported on a controlled experiment
in which 16 individuals were exposed to varying levels of formaldehyde for five hours and rated
their level of discomfort over the exposure period. Discomfort occurred within 1 hour at
formaldehyde exposure levels of 1 and 2 mg/m3 (Andersen and Molhave, 1983) After 2 hours,
increasing discomfort was reported among the groups exposed to 0.3 and 0.5 mg/m3. Subject
reported that discomfort was mainly conjunctival irritation and dryness in the nose and throat.
Subjects complained at all four concentrations of formaldehyde: 0.3, 0.5, 1.0, and 2.0 mg/m3 and
of 16 subjects, 3, 5, 15, and 15 subjects complained at each respective exposure concentration
(Andersen and Molhave, 1983).
Controlled chamber studies have also been conducted on various populations of
previously exposed individuals to determine if formaldehyde exposure potentiates an
individual's response to acute exposures. Schachter et al. (1987) reported on 15 laboratory
workers "frequently exposed to formaldehyde" (no quantification of exposure is given; however,
the workers report being exposed for 1 to 7 days per week from a range of 1 to 21 years). Tests
performed at the start of the study found that these individuals had pulmonary function similar to
that seen in healthy individuals. The workers in this study reported subjective measures of eye,
This document is a draft for review purposes only and does not constitute Agency policy.
4-10 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
nose, and throat irritation after 40 minutes of exposure to 2 ppm of formaldehyde. However, the
2 ppm acute exposure in this study may be sufficiently high to induce significant irritation in
most individuals. Krakowiak et al. (1998) reported that 10 asthmatics with occupational
exposure to formaldehyde (via formaldehyde solutions or pure gaseous formaldehyde) exhibited
similar symptom scores to healthy controls (never exposed to formaldehyde in the workplace)
exposed to 0.4 ppm formaldehyde for 2 hours. The mean symptom scores and standard
deviation (SD), which included information on sneezing, rhinorrhea, mucosal edema, and
itching, were 4.6 ±1.6 (mean ± SD) for asthmatics and 4.3 ±1.2 for healthy subjects
immediately after inhalation. These dropped to 1.8 ± 1.2 and 1.2 ± 1.3, respectively, 4 hours
after the exposure. It is unclear if sensitive individuals may not be represented in either of these
groups, as the workers were tolerating their exposures during the work shift "healthy worker"
effect. However, residents (n = 9) exposed to formaldehyde in their homes, who complained
about adverse effects from the material, but with no occupational exposure reported eye, nose,
and throat irritation at a similar rate as controls (individuals in homes without formaldehyde or
individuals in homes with formaldehyde but not reporting adverse effects [n = 9]) after a
90-minute exposure to 1 ppm (Day et al., 1984). The number of individuals reporting eye
irritation, nasal congestion, and throat irritation were seven, three, and two among sensitive
individuals and eight, four, and three among controls, respectively. These individuals may be
considered a sensitive population since they had "previously complained of various
nonrespiratory effects from the UFFI in their homes" (household concentrations unknown).
4.1.1.2. Pulmonary Function
Workers chronically exposed to formaldehyde have exhibited signs of reduced lung
function consistent with BC, inflammation, or chronic obstructive lung disease. Lung function
deficits have been reported both in pre-shift versus post-shift measurements and as a result of
chronic exposures (Pourmahabadian et al., 2006; Herbert et al., 1994; Malaka and Kodama,
1990; Alexandersson and Hedenstierna, 1989; Alexandersson et al., 1982). Decreases in
spirometric values, including vital capacity (VC), forced expiratory volume (FEV), forced vital
capacity (FVC), and FEV/FVC have been reported. Chronic studies also report increased
respiratory symptoms, such as cough, increased phlegm, asthma, chest tightness, and chest colds,
in exposed workers (Pourmahabadian et al., 2006; Herbert et al., 1994; Malaka and Kodama,
1990; Alexandersson and Hedenstierna, 1989; Alexandersson et al., 1982). Similar findings
have been reported for low-level residential formaldehyde exposure, including decreased peak
expiratory flow (PEF) rates (Krzyzanowski et al., 1990).
This document is a draft for review purposes only and does not constitute Agency policy.
4-11 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Worker exposures that report cross-shift differences in spirometric values are consistent
with formaldehyde-induced sensory irritation. Additionally, concordance has been reported
between subjective irritant response and measured changes in pulmonary function, further
supporting the possibility that cross-shift and short-term evidence of BC may be a reflexive
response to sensory irritation. Absolute values for lung function parameters are likely to vary by
gender, age, height, and smoking status and are best compared when normalized to the expected
lung function based on these variables (Schoenberg et al., 1978). Individual variation can also be
addressed by each subject serving as his/her control with measurements taken before, during, and
after exposure. Analysis of the percent change in various parameters in this context may have
greater sensitivity to detect exposure-related changes in function.
In addition to individual variation in baseline lung function, there is variation in bronchial
responsiveness. Reduced lung function parameters in response to methacholine challenge is a
standard test for BC, and this can be used to define responsive, sensitive, or susceptible
individuals. Since formaldehyde-induced BC is measured with these lung function tests,
variability in bronchial responsiveness may impact interpretation of formaldehyde-induced
changes. Experiments with sensitive individuals can help address this question. However,
results need to be normalized in some way to account for differences in responsiveness before
formaldehyde exposure. Researchers have in some cases excluded hyperresponsive individuals
or presented results as a proportion or percent of the unexposed value for each individual.
However, excluding sensitive individuals may bias results towards the null.
The American Thoracic Society (ATS) published an official statement on what
constitutes an adverse health effect of air pollution (ATS, 2000). According to the ATS
statement, exposure that increases the risk of an adverse effect to the entire population can be
considered adverse, even though it may not increase the risk of any individual to an unacceptable
level. For example, a population of asthmatics could have a distribution of lung function such
that no individual has a level associated with significant impairment. Exposure to an air
pollutant could shift the distribution to lower levels that still do not bring any individual to a
level that is associated with clinically relevant effects. However, this would be considered to be
adverse because individuals within the population would have diminished reserve function and
therefore would be at increased risk if affected by another agent.
4.1.1.2.1. Epidemiologic literature. The potential adverse effects of formaldehyde exposure on
pulmonary function in humans can be examined on several time scales of interest. The
epidemiologic literature supports the assessment of exposures among exposed anatomy medical
students where all participants have well-defined and similar duration of exposure (i.e., a
This document is a draft for review purposes only and does not constitute Agency policy.
4-12 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
semester-long class) (Kriebel et al., 2001, 1993; Akbar-Khanzadeh and Mlynek, 1997; Akbar-
Khanzadeh et al., 1994; Uba et al., 1989; Fleisher, 1987), among individuals living or working in
buildings with formaldehyde exposure (Franklin et al., 2000; Krzyzanowski et al., 1990; Main
and Hogan, 1983), and among workers (industrial, manufacture, mortuary, hospital staff, etc.)
(Ostojic et al., 2006; Herbert et al., 1994; Khamgaonkar and Fulare, 1991; Malaka and Kodama,
1990; Nunn et al., 1990; Alexandersson and Hedenstierna, 1989; Holness and Nethercott, 1989;
Holmstrom and Wilhelmsson, 1988; Horvath et al., 1988; Kilburn et al., 1985; Alexandersson et
al., 1982).
The observed effects in the previously unexposed anatomy students provide additional
information on acute exposures in two naive populations (Kriebel et al., 2001, 1993) as well as
special insight into the intermediate stages of possible sensitization (Kriebel et al., 1993).
Kriebel and colleagues (1993) examined the pre-laboratory and post-laboratory PEF in students
attending anatomy classes once per week. They found the strongest pulmonary response when
examining the average cross-laboratory decrement in PEF in the first 2 weeks of the study when
formaldehyde concentrations collected in the breathing zones had a geometric average
concentration of 0.73 ppm. Overall, the students exhibited a 2% decrement in PEF, while the
students with any history of asthma showed a 7.3% decrement in PEF. These findings of acute
decreases in PEF following students' initial anatomy sessions were corroborated by the Kriebel
et al. (2001) study, which used a similar study design applied to another class of anatomy
students.
The Kriebel et al. (1993) study also shows how the acute effects of formaldehyde
exposure were altered following several weeks of weekly episodic exposure. By the fifth week
of class, the pre- and post-laboratory measurements of PEF were no longer reflecting a clearly
demonstrated acute effect, but, following the seventh week of episodic exposure, both pre- and
post-laboratory PEF continued to drop steadily until the class adjourned after 10 weeks. While
the acute effects of formaldehyde exposure appeared to diminish after several weeks of exposure,
the intermediate effect across 9 weeks was a 24 L/minute drop in PEF that was statistically
significant (p < 0.01 after statistical control for random person effects, asthma, interaction
between time and asthma, and eye as well as nose symptoms of irritation).
Similar studies among medical students have been performed. In one study, 34 exposed
and 12 control students completed pulmonary function tests before and after their work in the
laboratory (approximately 3 hours) (Akbar-Khanzadeh et al., 1994). The time-weighted average
exposure ranged from 0.07-2.94 ppm. More than 94% of the subjects were exposed to >0.3 ppm
and 31.7% were exposed to >0.5 ppm. Comparing pre- and post-exposures among the exposed
students, on average FVC decreased by 1.4%, FEV3 decreased by 1.2%, FEVi/FVC increased by
This document is a draft for review purposes only and does not constitute Agency policy.
4-13 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
1.6%, and FVC25-75% increased 2.5%. These average percent changes in the control group are -
0.3, 1.30, 2.31, and 0.6%, respectively. The researchers also calculated correlation coefficients
by examining the relationship between lung function and formaldehyde concentration, but no
association was found. Akbar-Khanzadeh and Mlynek (1997) performed another study with 50
exposed students and 36 controls and reported a larger increase in lung function among controls
when compared with cases after 1-3 hours of exposure (FVC 3.0 versus 0.9, FEVi 4.1 versus
1.2, FEV3 3.3 versus 0.8, forced expiratory flow during the middle of the FVC [FEF25_75%] 6.1
versus 0.7). These differences between cases and controls remained for FEVi, FEV3, and FEF25
75% after 3 hours.
A study of 103 medical students was performed over a period of 7 months in which the
students were exposed to formaldehyde at a time weighted average of <1 ppm with peaks of 5
ppm during anatomy laboratory sessions (Uba et al., 1989). Twelve students were asthmatics.
Unlike the studies by Kriebel et al. (2001, 1993), these researchers did not find a change in
pulmonary function over the course of 7 months. The mean percent change for pulmonary
function before and after the exposure did change slightly, with measures showing decreases in
function at the end of the laboratory session (measurements taken at the 7-month time point:
FVC -0.79%, FEVi -0.48%, FEF25 75% 0.07%, FEVi/FVC 0.24%).
Finally, Fleisher (1987) gave self-administered questionnaires to medical students after
completing an anatomy laboratory session (formaldehyde exposure measures as <1 ppm) and a
pathology/microbiology laboratory session (no formaldehyde exposure). Over 8% of students
reported experiencing shortness of breath during the laboratory with formaldehyde exposure, but
none of the students reported shortness of breath in the laboratory session with no exposure. No
objective measurements of formaldehyde exposure were used.
Three studies have been performed that examine formaldehyde exposure from the
buildings in which individuals live or work. One study included children 6-13 years of age and
measured the levels of formaldehyde in their homes. There was no association between FVC or
FEV and the indoor concentrations of formaldehyde, although there were signs of lower airway
inflammation as measured by levels of exhaled nitric oxide (NO) in children exposed to average
formaldehyde levels >0.05 ppm (Franklin et al., 2000). Municipal employees with their children
(613 adults and 298 children) were randomly sampled in another study of home exposures
(Krzyzanowski et al., 1990). Residential exposures to formaldehyde were based on repeated
samples from each individual's kitchen, living area, and bedroom. The average formaldehyde
concentration was 26 ppb, with a maximum sample value of 140 ppb. The majority of subjects
(83%>) lived in homes with 2-week average concentrations below 40 ppb. Subjects' peak
expiratory flow rates (PEFRs) were determined four times daily, in the morning, at noon, in the
This document is a draft for review purposes only and does not constitute Agency policy.
4-14 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
early evening, and before bed, for 2 weeks. A statistically significant linear relationship between
increased formaldehyde exposure and decreased PEFR was reported in children but not adults.
All statistical models controlled for socioeconomic status, tobacco smoking (current active or
environmental tobacco smoking), and nitrogen dioxide concentrations. Among adults, there was
a statistically significant nonlinear relationship with decreased morning PEFR for formaldehyde
concentration <40 ppb.
Main and Hogan (1983) reported on a group of individuals (n = 21) working in two
mobile trailers for 34 months and exposed to levels of formaldehyde ranging from 0.12 to
1.6 ppm (mean age 38 ± 9 years, 76% male, 19% nonsmokers). The control population was
comprised of individuals who did not work in the trailers (n = 18; mean age 30 ± 6, 50% male,
22% nonsmokers). There were no differences between the exposure and control groups' percent
predicted FEVi or FVC regardless of smoking status.
Several studies allowed for the examination of potential chronic effects of formaldehyde
exposure. These included an occupational study by Malaka and Kodama (1990) that reported
pre-shift pulmonary function as a percentage of expected among the formaldehyde exposed
compared with comparable people not exposed to formaldehyde. This study found that an
average 8-hour TWA formaldehyde exposure of 1.13 ppm from area samples was associated
with statistically significant decrements in FEVi, FEVi/FVC, and FEF25-75% compared with a
referent population. The strongest response was for FEF25 75o,. which showed a 12% drop in
observed function compared with expected function in the unexposed, but it is unclear how to
interpret the potential chronic adverse effect(s) with just the magnitude of the decrement and the
length of the average occupational tenure at this plywood facility (6.5 years), which was not
reported by exposure status.
A study comparing oriented strand board workers (exposed to formaldehyde) with oil/gas
field plant workers (not exposed to formaldehyde) demonstrated a difference in pulmonary
function between the two groups (Herbert et al., 1994). The groups were similar in regard to
measured FVC and FEVi (controlled for age, height, and smoking), but the workers exposed to
formaldehyde had lower FEVi/FVC. In addition, those exposed to formaldehyde showed a
decrease in FVC and FEVi after their shift, with an average pre- and post-shift difference of
47 mL (p = 0.022) and 39 mL (p = 0.044) for FVC and FEVi, respectively (however change
could not be compared with the controls of this study because no post-shift measurements were
taken). Two other occupational studies found no association between formaldehyde and lung
function (Holness and Nethercott, 1989; Horvath et al., 1988). One of these studies was
conducted among funeral workers and an unexposed control group (Holness and Nethercott,
1989). There was no difference in pulmonary function of the two groups at baseline. After
This document is a draft for review purposes only and does not constitute Agency policy.
4-15 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
exposure, there was no change in lung function for the exposed or unexposed when comparing
lung function tests done immediately before and after an embalming procedure (for controls the
repeat measures were taken approximately 2-3 hours after the first measure) (changes in
percentage predicted FVC and FEVi were 0.88 ± 2.95 and -0.03 ± 2.40 for exposed and 1.13 ±
3.98 and 1.45 ± 4.43 for unexposed). Further analysis showed no association between
formaldehyde levels and changes in lung function. Another study (Horvath et al., 1988) found
no differences in pre-shift pulmonary function between the exposed (workers at a particleboard
and molded products operation, formaldehyde measured using individual monitors ranged from
0.17-2.93 ppm) and controls (workers from nearby food processing facilities, formaldehyde
measured using individual monitors ranged from 0.03-0.12 ppm). However, the authors did find
a post-shift decline in FVC and FEVi among controls and FEVi and FEF25_75% among workers
when using paired comparisons for each group. When assigning all controls a formaldehyde
exposure value of 0.05 ppm, there was a correlation detected in pre-and post-shift pulmonary
function changes and formaldehyde, though no specific details on regression analysis were
provided.
A study performed in India (Khamgaonkar and Fulare, 1991) examined individuals
working in anatomy and histopathology departments and exposed to formaldehyde (mean
1.00 ppm, range 0.036-2.27). Controls (individuals not working in laboratories with
formaldehyde) were exposed to an average of 0.102 ppm formaldehyde (range 0-0.52). Lung
function tests were performed on a Monday morning after days of no exposure in order to
examine chronic effects. The FVC and FEVi% of the exposed group, respectively, were 17.12
and 22.94% reduced compared with the control group (Khamgaonkar and Fulare, 1991);
however, while the pool of cases and controls were frequency-matched on age and gender, there
was no mention by the investigators of normalizing the pulmonary function metrics by gender
and height, which would have made for more appropriate comparisons. Kilburn et al. (1985)
also demonstrated reduced pulmonary function (lower percent predicted FVC, FEVi, and
FEF25-75%) among workers occupationally exposed to formaldehyde when compared with
individuals working at jobs without formaldehyde exposure.
Two occupational studies found no association between formaldehyde exposure and
deficits in pulmonary function (Ostojic et al., 2006; Holmstrom and Wilhelmsson, 1988).
Ostojic et al. (2006) examined nonsmoking male health service professionals working in
pathoanatomic laboratories with 8 hours of formaldehyde exposure per day at an unspecified
concentration for at least 4 years (n = 16). The control group was comprised of sixteen age- and
stature-matched nonsmoking male controls. There was no difference in mean FVC or FEVi
between exposed and controls. The researchers also examined values for diffusing lung capacity
This document is a draft for review purposes only and does not constitute Agency policy.
4-16 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
for carbon monoxide and membrane diffusion capacity, which were similar between the exposed
and control groups. However, blood volume of pulmonary capillaries was found to be higher in
the exposed group. Holmstrom and Wilhelmsson (1988) recruited individuals from a chemical
plant where formaldehyde and formaldehyde products were made (n = 70). Exposure levels
varied from 0.05-0.5 mg/m3. A control group was mostly comprised of clerks for the local
government (n = 36). No difference in FEV% was detected between the groups. Mean FVC was
lower than expected among the exposed group (expected values were based on age, sex, smoking
habits, height, and weight). This study went further and measured changes in pulmonary
function for those employed more than 5 years and reported no signs of increasing restrictivity
after 5 years. No associations were seen and there was no correlation between pulmonary
function and cumulative dose of formaldehyde (Holmstrom and Wilhelmsson, 1988).
There have been only two studies that have reported on the longitudinal follow-up of
workers exposed to formaldehyde (Nunn et al., 1990; Alexandersson and Hedenstierna, 1989).
The Alexandersson and Hedenstierna (1989) investigation not only examined the acute effects of
exposure across shift but was able to do so among some of the same workers that had been
studied 5 years earlier (Alexandersson et al., 1982). Statistically significant decreases in
FEVi/FVC and FEF25-75% were noted over the intervening 5 years in nonsmokers after correction
for normal aging and reference lung function spirometry values. The decrease in FEF25-75% was
0.212 ± 0.066 L/second (mean ± SD) for each year of exposure and was highly significant (p <
0.01). For comparison with the 12% drop in the same pulmonary metric reported by Malaka and
Kodama (1990) over an estimated 6.5 years, the extrapolated percentage decrease in FEF25-75%
was computed for the Alexandersson and Hedenstierna (1989) study by using the reported yearly
decrement applied to the pre-shift values at the time of the initial study period. From the
predicted value of 4.57 L/second, a decrease of 0.168 L/second for each year of exposure
regardless of smoking status was calculated. For 6.5 years of exposure, this would result in a
24% drop in FEF25-75%. Formaldehyde concentrations were estimated at 0.42 ppm in the first
Alexandersson et al. (1982) study and at 0.50 ppm in the Alexandersson and Hedenstierna (1989)
study. The study by Nunn et al. (1990) assessed the decrease in FEVi with age. The researchers
calculated the decrease in FEVi to be 42 mL/year among workers exposed to formaldehyde and
41 mL/year for workers who were not exposed to formaldehyde. Thus, they showed no
association between formaldehyde exposure and decreased FEVi.
There are a few important limitations to consider in these occupational studies of
formaldehyde exposure. First, an often-shared weakness is the absence of data on, and
appropriate statistical control of, potential confounding by occupational co-exposures. Also,
studies that did not report pre-shift pulmonary function as a percentage of expected function
This document is a draft for review purposes only and does not constitute Agency policy.
4-17 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
contribute less to an assessment of potential chronic effects because, post-hoc, it is difficult to
calibrate the multiple pulmonary function data for cross-study comparison without knowledge of
the age, gender, smoking status, height, year of birth, etc., that are important determinants of the
pulmonary function metrics of concern.
4.1.1.2.2. Acute studies: controlled chamber exposures. Pulmonary effects of acute
formaldehyde exposure have been studied in both healthy volunteers and sensitive populations
under controlled conditions. Controlled chamber studies have the advantage of measured
controlled exposures, but other factors can limit the usefulness of the studies, especially when
study populations are small and there is high variability in the measured parameters.
In general, acute formaldehyde exposures (0.5-3 ppm) have not induced significant
pulmonary deficits in healthy, non-exercising volunteers (Kulle et al., 1987; Schachter et al.,
1986; Witek et al., 1986; Day et al., 1984; Andersen and Molhave, 1983). However, it is unclear
whether the data analysis in these reports had the statistical power to substantiate the small
deficits reported in occupational and student studies. All four reports had relatively small study
groups of healthy individuals (n = 19 [Kulle et al., 1987], n = 16 [Andersen and Molhave, 1983],
n = 15 [Schachter et al., 1986], n = 15 [Witek et al., 1986], and n = 9 [Day et al., 1984]), and in
some cases the group was further divided by gender. Two studies report the absolute values of
the lung function parameters without adjustment to individual expected function or the
unexposed baseline for each individual (Kulle et al., 1987; Andersen and Molhave, 1983). As
discussed, this decreases the power of the study to detect formaldehyde-induced changes in
pulmonary function. In contrast, Witek et al. (1986) and Schachter et al. (1986) report lung
function as a percent of baseline (although not normalized for age gender and height). Each
study showed an increase in FEVi in formaldehyde-exposed individuals at rest and increases in
maximal expiratory flow (MEF) at 50% of expired vital capacity (MEF50%) (Witek et al., 1986;
Schachter et al., 1986). However, in both reports the SDs of changes in lung function parameters
are quite large, nearly equaling the reported value and exceeding it in several cases. The absence
of normalized raw data, combined with large individual variation, limit the interpretation of these
studies. A small study (Day et al., 1984) that included nine healthy individuals showed no
changes in FEV or FVC after 90 minutes of exposure to 1 ppm of formaldehyde. A more recent
study (Lang et al., 2008) of 21 healthy volunteers exposed to a range of formaldehyde
concentrations (0.0 to 0.15 or 0.3 ppm) reports no formaldehyde-related pulmonary deficits.
However, data were not shown, and it is unclear whether the authors compared absolute or
relative values of lung function and what variation in lung function was present in the study
population. Additionally, the authors did not provide the criteria used to gauge deficit of lung
This document is a draft for review purposes only and does not constitute Agency policy.
4-18 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
function. If a clinically significant deficit was defined (e.g., 20%), then more subtle changes in
pulmonary function, as supported in other studies, would not have been reported.
Similar to these studies of healthy individuals, acute controlled studies including
asthmatics also report no changes in pulmonary function associated with formaldehyde exposure
(Ezratty et al., 2007; Harving et al., 1990; Green et al., 1987; Sauder et al., 1987; Witek et al.,
1987, 1986), including studies of individuals thought to have formaldehyde-induced bronchial
asthma (Krakowiak et al., 1998); however, the number of asthmatic individuals included in each
of these studies was small. (The details of these studies have been reported elsewhere in this
chapter. Briefly, the number of asthmatics in the study/total number of individuals in the study
are as follows: Ezratty et al. [2007]—12/12; Harving et al. [1990]—15/15; Green et al. [1987]—
16/38; Sauder et al. [1987]—9/9; Witek et al. [1987]—15/15; Witek et al. [1986]—15/30;
Krakowiak et al. [1998]—10/20], The same is true for individuals who are frequently exposed to
formaldehyde either at work (n = 15) (Schachter et al., 1987) or at home (n = 18) (Day et al.,
1984).
Small but statistically significant deficits in pulmonary function due to acute
formaldehyde exposure (2 or 3 ppm) have been reported in healthy volunteers during exercise
(Green et al., 1989, 1987; Sauder et al., 1986; Schachter et al., 1986). Although changes in lung
function parameters averaged over experimental groups were generally small, some individuals
exhibited clinically significant deficits, even after only 2 hours of exposure. Deficits in FEVi
and FEF25-75% in the first 30 minutes of a 2-hour exposure at 3 ppm formaldehyde were 2 and
7%, respectively. Changes in lung function were not statistically significant after 60 and
180 minutes of exposure (Sauder et al., 1986), even when assessed as absolute rather than
relative measurements. Thirteen percent (5 of 38 subjects) demonstrated formaldehyde-induced
clinically significant deficits when exposed at 3 ppm during exercise (defined by Green et al.
(1987) as decrease in FEVi > 10% of control).
4.1.1.3. Asthma
A large number of studies have investigated the potential association between
formaldehyde exposure and a continuum of adverse health effects ranging from decrements in
pulmonary function to asthma. In general, epidemiologic studies of adults have reported varied
results between null findings and positive findings. However, The National Research Council
concluded in its report on Formaldehyde that, "Formaldehyde has been shown to cause bronchial
asthma in humans" (NRC, 1981), citing numerous studies demonstrating the induction of asthma
following exposure to formaldehyde (Hendrick and Lane, 1975, 1977; Laffont and Noceto, 1961;
Nova and Touraine, 1957; Paliard et al., 1949; Popa et al., 1969; Sakula, 1975; Schoenberg and
This document is a draft for review purposes only and does not constitute Agency policy.
4-19 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
Mitchell, 1975; Turiar, 1952; Vaughan, 1939). In a subsequent review article on formaldehyde
and the health effects that have been associated with it, Stenton and Hendrick (1994) reported on
formaldehyde and asthma in occupational settings and starkly describe the .first detailed case
report of formaldehyde asthma confirmed by specific inhalation challenge test occurring in a
nursing sister on a renal dialysis unit. Her symptoms were suggestive of late asthmatic reactions
occurring 4 to 5 hours after heavy exposures. The occurrence of late reactions was confirmed in
a series of challenge tests that involved the painting of formalin onto cardboard pieces within a
confined space" (Stenton and Hendrick, 1994; Hendrick 1997). The results of the challenge tests
are illustrated in Figure 4-1 .
140
eantral taatar)
lormilm 25X lor IS
- formalin 25S »or 15
Figure 4-1. Delayed asthmatic reaction following the inhalation of
formaldehyde after "painting" 100% formalin for 20 minutes. Challenge 2
was premedicated with inhaled betamethasone 200 jig.
Source: (Stenton and Hendrick, 1994)
Five years later, the two nurses were re-challenged with the nurse who had left the
dialysis unit having no response to the subsequent challenge while the nurse who had remained
working in the unit developed mild late asthmatic response with peripheral blood eosinophilia
(Stenton and Hendrick, 1994; Hendrick et al., 1982). Stenton and Hendrick (1994) concluded
that these studies "provide clear evidence of formaldehyde's ability to induce asthma" but no
indication of the exposure concentrations to induce it. In a follow-up study of dialysis unit
This document is a draft for review purposes only and does not constitute Agency policy.
4-20 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
staffers exposed to formaldehyde as a sterilizing agent, 8/28 people reported respiratory
symptoms and a prolonged increase in circadian rhythm of peak expiratory flow rate was seen in
one subject (Hendrick and Lane, 1983) implying an increase in airway responsiveness (Stenton
and Hendrick, 1994). It should be noted, however, that while there did appear to be a clear
response to formalin, it is not known what contribution to the response was attributable to
formaldehyde and what contribution might have been attributable to methanol. Further, while
the evidence of a causal association between formaldehyde and asthma is clear, the above studies
do not offer information on the concentrations at which adverse effects would expected in a
population.
There is at least one clinical study in humans that investigated whether exposure to a low
level of formaldehyde (500 |ag/m3) would enhance inhaled allergen responses (Ezratty et al.,
2007). Twelve subjects with intermittent asthma were exposed to either formaldehyde or
purified air in a double-blind crossover study for 1 hour. Following exposure (8 hours), airway
responsiveness to methacholine challenge was measured. No significant effects on
methacholine-induced bronchial hyperresponsiveness (BHR) were detected due to formaldehyde
exposure.
Numerous epidemiologic studies have investigated adverse effects in populations.
Decreased peak expiratory flow rates (PEFR) are an important component in the diagnosis of
asthma and there is evidence of formaldehyde-induced decrements in PEFR (see Section
4.1.1.2). However, the diagnosis of asthma is both a more serious health condition and
diagnostically more complex than decreased PEFR alone and is evaluated here as a distinct
endpoint. A number of epidemiologic studies have investigated the potential association
between formaldehyde exposure and a continuum of adverse health effects from pulmonary
function to asthma.
The association between formaldehyde and asthma has been studied by examining
occupational exposures (Fransman et al., 2003; Malaka and Kodama, 1990), school-related
exposures (Zhao et al., 2008; Smedje and Norback, 2001; Norback et al., 2000) and residential
exposures (Matsunaga et al., 2008; Tavernier et al., 2006; Gee et al., 2005; Delfino et al., 2003;
Rumchev et al., 2002; Garrett et al., 1999; Palczynski et al., 1999; Norback et al., 1995;
Krzyzanowski et al., 1990). The two occupational studies examined the respiratory health of
plywood workers (Fransman et al., 2003; Malaka and Kodama, 1990). The most recent of these
was conducted in New Zealand by Fransman et al. (2003). Personal samples of formaldehyde
exposure were taken. The mean level of exposure was 0.08 mg/m3 (65 ppb) and the majority of
samples were below the limit of detection which was reported to be 0.03 mg/m3 (24 ppb).
Compared with those with low levels of formaldehyde exposure, workers with high levels of
This document is a draft for review purposes only and does not constitute Agency policy.
4-21 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
exposure were more likely to report having asthma (OR=4.3 [95% CI]: 0.7-27.7]). The
association was not seen when examining formaldehyde exposure and use of asthma medication.
The second study of plywood workers was completed in Indonesia. Background levels of
formaldehyde ranged from 0.003 to 0.07 ppm. The highest concentration of formaldehyde
detected in an air sample was in the particleboard unit (range 1.16 to 3.48 ppm). Asthma, which
was defined as "have you ever had an attack of wheezing that made you feel short of breath?",
was found to be positively associated with formaldehyde exposure (Malaka and Kodama, 1990).
Studies of exposure to formaldehyde at schools have been performed in China (Zhao et
al., 2008) and in Sweden (Smedje and Norback, 2001). In the study from China (Zhao et al.,
2008), mean levels of formaldehyde were reported to be 2.3 |ig/m3 (range 1.0-5.0 |ig/m3)
indoors and 5.8 |ig/m3 (range 5.0-7.0 |ig/m3) outdoors. Cumulative asthma and daytime attacks
of breathlessness were found to be associated with outdoor formaldehyde levels. Neither of
these outcomes was associated with indoor concentrations of formaldehyde; however, indoor
levels were found to be associated with nocturnal attacks of breathlessness. In Sweden (Smedje
and Norback, 2001), the levels of formaldehyde measured indoors were higher (mean 4, range
<5.0-72 |ig/m3). One difference between this Swedish study and the study performed in China is
that the Swedish study examined the incidence of asthma over a 4-year period. This study did
not report an association between formaldehyde exposure and the incidence of asthma (OR 1.2
[95% CI: 0.8-1.7]) among the whole study population. However, when the investigators
stratified on history of atopy, they reported that among children without a history of atopy, a new
diagnosis of asthma was significantly more likely at higher concentrations of formaldehyde (OR
1.7 per 10 |ig /m3 [95% CI: 1.1-2.6]) and at higher total concentrations of mold (OR=4.7 per 10-
fold increased in total molds [95% CI: 1.2-18.4] in the classroom air. The finding for adverse
effects of formaldehyde and mold did not appear to control for the other exposure and no
information on the potential correlation between the two exposures was provided. In order to
evaluate the potential for confounding of the reported formaldehyde association by the reported
mold association, the magnitude of effects must be compared on an appropriate scale since the
magnitude of an odds ratio depends on the magnitude of the change in exposure level that is
expected to produce increased risk. Standardizing the units to the reported geometric mean
standard deviation, the result for formaldehyde (GSM=2.3 |ig /m3) is OR1=1.13 per GSD and the
result for mold is OR2=1.02 for a comparison of risks at the GSM to 10*GSM and OR3=1.06 for
1 OR per GSD=exp[ln(OR per ng/m3)/10 ng/m3 * 2.3 |ig /m31 =cxp11 n( 1.7)/10*2.31 = 1.13
2 OR per GSD=exp[ln(OR per 10-fold increase)/ (9*GSM)*1.6 |ig /m3]=exp[ln(4.7)/162*1.6]=1.02
3 OR per GSD=exp[ln(OR per 10-fold increase)/ (9*Minimum)*1.6 |ig /m3]=exp[ln(4.7)/45*1.6]=1.06
This document is a draft for review purposes only and does not constitute Agency policy.
4-22 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
a comparison of risks at the minimum value of total molds (5*103/m3) to 10*minimum. As it
appears that the magnitude of the formaldehyde effect is substantially stronger than that of the
mold effect (following standardization of exposure increment) it can be concluded that the
reported formaldehyde effect could not have been the spurious result of uncontrolled
confounding by mold.
The results of studies measuring residential exposure to formaldehyde and asthma are
varied, with some demonstrating an association and others finding no relationship. A recent
study (Matsunaga et al., 2008) found no association between 24-hour formaldehyde and
prevalence of asthma when pregnant women with an exposure >47 ppb were compared to those
with exposure to <18 ppb. However, they did report at increased risk of atopic eczema. It
should be noted that this study did not assess the risk of incident asthma. A study utilizing self-
reported asthma prevalence as an outcome also found no association with levels of formaldehyde
(mean 25.9 |ig/m3, range 2.0-66.8 |ig/m3) (Palczynski et al., 1999) although they did note that
the incidence of allergic diseases was highest in the highest formaldehyde exposure group but
that the group was too small for statistical evaluation.
A study performed by Tuthill (1984) measured formaldehyde exposure for children
grades K through 6 by using a combination of proxy variables. Overall, there was no
association, but some individual variables did show an increased risk. For example, the reported
risk ratio for having new construction or remodeling performed in the house in the past 4 months
was 2.5 (95% CI: 1.7-3.9). The risk ratio for having new or upholstered furniture in the house
(brought into the house within the past 4 months) was 2.2 (95% CI: 1.2-3.9).
The study by Delfino et al. (2003) assessed whether ambient formaldehyde concentration
measured at a central monitoring site were associated with asthma symptoms. The study
examined 22 10-15 year olds with at least 1 year of physician-diagnosed asthma and living in a
nonsmoking household. The mean levels of formaldehyde were measured to be 7.21 ppb (range
4.27-14.02 ppb). There was a positive association between asthma symptom scores (comparing
children who report symptoms interfering with their daily activities versus those with no
symptoms or symptoms not great enough to affect their daily activities) and high current levels
of formaldehyde (OR 1.90 [95% CI: 1.13-3.19]).
Three studies (Tavernier et al., 2006; Gee et al., 2005; Garrett et al., 1999) were
performed by matching children with and without asthma and comparing the levels of
formaldehyde in their homes. Gee et al. (2005) reported median formaldehyde levels of 0.03
ppm in living rooms and 0.04 ppm in bedrooms. Analyses were limited to univariate
comparisons of formaldehyde levels for cases of existing asthma and controls without asthma.
The concentrations did not differ in a statistically significant manner. The study by Gee et al.
This document is a draft for review purposes only and does not constitute Agency policy.
4-23 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
(2005) was followed up with a more sophisticated analysis of the same children in the same
homes. Tavernier et al. (2006) reiterated the earlier finding by Gee et al (2005) that
formaldehyde was not found to be associated with existing asthma. Tavernier et al. (2006) did
not report the measured levels of formaldehyde but gave the OR for the highest tertile of
exposure compared with the lowest tertile of exposure as 0.99 (95% CI: 0.39-2.50). The width
of this confidence interval suggests that these findings would still be consistent with two-fold
increase in risk.
Garrett et al. (1999) reported on the risk of allergy and asthma-like respiratory symptoms
due to formaldehyde exposure in a cross-sectional survey of households with children with (n =
53) or without (n = 88) doctor-diagnosed asthma. Formaldehyde exposure was characterized by
4 seasonal in-home sampling events across the year for bedrooms and 4-day passive samples
collected in living rooms, kitchens and outdoors. Statistically significant linear trends for
increased risk of having asthma were seen with increasing formaldehyde levels (p < 0.02);
however, the ORs for the association did not remain statistically significant after controlling for
parental allergy and asthma (exact ORs and 95% CIs not given). Garrett et al (1999) also
evaluated the prevalence and severity of allergic sensitization to 12 common allergens and
reported increased prevalence with increasing formaldehyde concentration in the home. The
respiratory symptom score was also increased and demonstrated a significant effect for
formaldehyde in a multiple regression after adjusting for multiple risk factors and interactions.
For the atopy and respiratory symptom endpoints, severity/incidence was increased in the
medium (20-50 |ig/m3) and high (>50 |ig/m3) exposure groups relative to the low (<20 |ig/m3)
exposure group, based on the highest of four seasonal 4-day formaldehyde measurements in the
home. The associations between formaldehyde concentrations and severity of allergic
sensitization are clearly shown and further substantiated with multivariate regression controlling
for potential confounders. In logistic regressions, both the prevalence and severity of allergic
sensitization to 12 common allergens increased with increasing formaldehyde concentration in
the home. The crude association for atopy with an increase in formaldehyde concentration per
10 |ig/m3 was OR=1.34 which increased when adjusted for parental asthma and gender to and
odds ratio of 1.42 per 10 |ig/m3 (95% CI: 0.99-2.04). Passive smoking, the presence of pets,
indoor nitrogen dioxide concentrations, airborne fungal spores and house-dust-mite allergens did
not influence the effect estimates and were unlikely to be confounders. Additionally, a
calculated respiratory symptom score was increased and demonstrated a significant relationship
to increased formaldehyde concentration in a multiple linear regression after adjusting for
multiple risk factors and interactions. For each of these endpoints, severity/incidence was
increased in the medium (20-50 |ig/m3) and high (>50 |ig/m3) exposure groups relative to the
This document is a draft for review purposes only and does not constitute Agency policy.
4-24 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
low (<20 |ig/m3) exposure group, based on the highest of four seasonal 4-day formaldehyde
measurements in the home.
Residential formaldehyde exposure was associated with an increased risk of asthma in a
population-based case-control study of 192 children aged 6 months to 3 years (Rumchev et al.,
2002). The study, which comprises 88 cases of children discharged from the emergency
department of a children's hospital in Perth, Australia, with a primary diagnosis of asthma and
104 controls, provides a positive exposure-response relationship. Seasonal in-home
formaldehyde measurements taken in the living room and subject's bedroom were used to assess
exposure (8-hour passive sampler). The odds ratios (ORs) for risk of asthma by formaldehyde
exposure level category were adjusted for numerous risk factors both familial and environmental
including, familial history of asthma, age, sex, smoking, presence of pets, and attributes of the
home. Of these, age, allergic sensitization to common allergens, and family history of allergy
were independent risk factors for asthma (ORs of 1.09, 2.57, and 2.66, respectively). Categorical
analysis of the data indicates the ORs for asthma were increased in the two highest formaldehyde
exposure groups, reaching statistical significance for household exposures > 60 |ig/m3 (48 ppb)
(OR of 1.39). Analysis of the data with formaldehyde as a continuous variable provides a
statistically significant increase in the risk of asthma (3 % increase in risk per every 10 ug/m3
increase in formaldehyde level. All analyses controlled for other indoor air pollutants, allergen
levels, relative humidity, and indoor temperature as well as other risk factors.
A study of 202 households (mean formaldehyde level of 26 ppb) found that among
children aged 6-15 years old and exposed to environmental tobacco smoke, the prevalence of
asthma was 45.5% for those with measured levels of formaldehyde in the kitchen >60 ppb. The
prevalence of asthma dropped to 15.1% for levels <40 ppb and 0% for 41-60 ppb. No trend in
asthma prevalence was seen for children who were not exposed to environmental tobacco smoke
(Krzyzanowski et al., 1990).
Finally, a study by Norback et al. (1995) reported mean levels of formaldehyde were 29
|ig/m3 (range <5-110 |ig/m3) in the bedrooms of individuals experiencing nocturnal
breathlessness compared with formaldehyde levels of 17 |ig/m3 (<5-60 |ig/m3) among those
without nocturnal breathlessness. The OR for this association was 12.5 (95% CI: 2.0-77.9) and
the effect was substantially stronger in magnitude than the associations observed for toluene,
terpenes and volatile organic compounds which makes confounding by those co-exposures
unlikely.
Formaldehyde has clearly been shown to be a cause of bronchial asthma and several
epidemiologic studies have identified causal evidence of an adverse effect of exposure on
pulmonary function and the incidence of asthma. While there are studies that did not find
This document is a draft for review purposes only and does not constitute Agency policy.
4-25 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
associations, many of those were limited by their study design, exposure measurement and the
definition of prevalent asthma as the health endpoint.
4.1.1.4. Respiratory Tract Pathology
Formaldehyde-induced respiratory tract pathology includes inflammation, rhinitis, goblet
cell hyperplasia, metaplastic changes, squamous cell hyperplasia, and impaired mucociliary
transport. Formaldehyde may bind to the trigeminal nerve and trigger the release of neurogenic
mediators of inflammation that result in tissue edema, lacrimation, mucus production and
leukocyte infiltration. How much inflammation, hyperplasia, and metaplastic change are due to
sensory irritation-induced inflammatory responses compared with formaldehyde-induced direct
cell damage cannot be distinguished. Increased mucus flow and metaplastic changes may
progress in relation to the concentration and duration of exposure to protect the underlying
tissue. When the exposure exceeds protective and defensive mechanisms, permanent damage
results (Swenberg et al., 1983). Nonetheless, these changes serve as a sensitive indicator of
formaldehyde exposure, since they occur before gross cellular damage and focal lesions
(Monticello et al., 1989), and potentially suggest a point at which the concentration and duration
of exposure exceed the protective nature of local responses (increased mucus flow, goblet cell
hyperplasia, squamous metaplasia, etc.) (Swenberg et al., 1983). A number of human studies
have reported nasal lesions associated with exposure to formaldehyde (Pazdrak et al., 1993;
Ballarin et al., 1992; Boysen et al., 1990; Holmstrom et al., 1989c; Edling et al., 1988), while
other studies have documented changes in mucociliary clearance and activity (Holmstrom and
Wilhelmsson, 1988; Andersen and Molhave, 1983). These studies are summarized below.
4.1.1.4.1. Nasal lesions. Ballarin et al. (1992) did a case-control study of 15 workers from a
plywood factory where urea-formaldehyde glue is used. Mean levels of formaldehyde exposure
(8-hour average) were estimated to be 0.09, 0.1, and 0.39 mg/m3 in three regions of the facility
(sawmill, shearing press, and warehouse, respectively). Nasal respiratory samples were
obtained. Stained cells were scored for histopathology. Cytology examination revealed
increased squamous metaplasia cells in 10 of 15 (67%) factory workers (with an average severity
score of 2.3) compared with one of 15 (6%) controls (with an average histology severity score of
1.6). In addition, one formaldehyde exposed worker (n = 15) exhibited mild dysplasia and had
the highest severity score (3.0). Authors suggest that these results may be due to chronic
irritation of the nasal respiratory mucosa. This small study reported only incidence of lesions
and did not score based on severity of lesions. The lesion incidence was not reported in relation
This document is a draft for review purposes only and does not constitute Agency policy.
4-26 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
to dose, so no dose-response relationship could be determined, precluding the establishment of a
point of departure (POD).
Holmstrom et al. (1989) collected nasal biopsy samples from workers exposed to air or to
formaldehyde at a median concentration of 240 ppm. Nasal biopsy samples were scored on a
0-8 range with normal respiratory epithelium as 0 and carcinoma as 8. Observed histologic
changes included loss of cilia, goblet cell hyperplasia, and cuboidal and goblet cell metaplasia
replacing normal columnar epithelium. The incidence associated with each histologic change
was not reported and cannot be compared between formaldehyde-exposed and control
individuals. Moreover, these biologically relevant changes were not analyzed independently in
the analysis. The mean scores were 1.56 (range, 0-4) for the control group and 2.16 (range, 0-4)
for the formaldehyde-exposed group. Although the range of scores in the controls and
formaldehyde-exposed groups were the same (0-4), the difference in mean scores (1.56 versus
2.16) was statistically significant (p < 0.05); scores were worse in the formaldehyde-exposed
group. The authors reported no correlation between the duration of exposure and histologic
changes and no correlation between smoking habits and biopsy scores. The loss of cilia, goblet
cell hyperplasia, and cuboidal and squamous cell metaplasia replacing the columnar epithelium
were increased in the group exposed to formaldehyde and is a biologically relevant change. This
study provides a lowest-observed-adverse-effect level (LOAEL) of 0.240 ppm for nasal
histopathology.
Edling et al. (1988) collected nasal biopsy samples from workers (n = 75) exposed to
formaldehyde at three plants (workers in two of these plants were also exposed to wood dust)
compared with a referent group (n = 25). Concentrations ranged from 0.1 to 1.1 mg/m3 (TWA)
with peaks of 5 mg/m3. Nasal histology was scored from 0 to 8 by increasing severity, from
normal respiratory epithelium (0) to carcinoma (8). A normal respiratory epithelium was noted
in 3 of 75 workers. A loss of cilia and goblet cell hyperplasia (scores of 2) was reported in eight
workers. Mixed cuboid/squamous epithelium (metaplasia), stratified squamous epithelium, and
keratosis were reported in 58 of 75 workers (those with scores of 3, 4, and 5 were combined).
Dysplasia (score of 6) was reported in 6 of 75 formaldehyde-exposed workers. None of the
workers had lesions that warranted a histologic score higher than 6. Histologic scores did not
correlate with duration of exposure but could not be confirmed due to poor reporting. Data from
the referent group were not included. A POD could not be determined from this study.
Boysen et al. (1990) collected nasal biopsy samples from workers exposed to air (n = 37)
or to formaldehyde (n = 37) and sometimes wood dust. The exposed workers were classified
into two exposure groups, 0.5-2 ppm and >2 ppm. Nasal biopsy samples were assessed by using
a histopathology score range of 0-5, based on the pathology of pseudostratified columnar
This document is a draft for review purposes only and does not constitute Agency policy.
4-27 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
epithelium (0) to dysplasia (5). Mean pathology scores for the control were decreased compared
with the formaldehyde-exposed group (1.4 and 1.9, respectively) but did not reach statistical
significance. Little quantitative pathology data were provided, although qualitative histology
revealed a range of observed effects from deciliated epithelial cells to mixed stratified cuboidal,
squamous epithelium to dysplasia. None of the control samples received histologic severity
scores of 4 or 5, indicating that keratinizing stratified squamous epithelium and dysplasia were
not observed in controls. A wider variety of histopathologic lesions were reported in exposed
workers compared with controls, and a greater number of exposed workers had histologic
changes compared with controls. Incidence data for each type of histopathology were not
reported, but the authors wrote that the degree of metaplastic alterations was more pronounced
among the exposed workers. An upper range for the high concentration group (>2 ppm) was not
reported, and median concentrations were not provided.
Pazdrak et al. (1993) exposed human subjects (six men, three women) to 0.4 ppm
formaldehyde in a chamber for 2 hours. Approximately half of the subjects suffered from skin
hypersensitivity to formaldehyde, while the other subjects were healthy. An evaluation of nasal
lavage pretest and following formaldehyde exposure revealed that the hypersensitive and healthy
groups had similarly elevated eosinophil counts at 0 hours after exposure (from
42 x 103 cells/mL to 72 x 103 cells/mL for healthy subjects \p < 0.05] and from
39 x 103 cells/mL to 69 x 103 cells/mL for hypersensitive subjects \p < 0.05]). Similar
eosinophil levels were also seen in both groups at 3 and 18 hours. Both groups had equivalent
increases in lavage albumin and total protein levels following exposure, but basophil counts were
unchanged. Based on evidence of formaldehyde-induced inflammation, these data provide a
LOAEL of 0.4 ppm for nasal histopathology.
4.1.1.4.2. Mucociliary clearance. In addition to abnormal nasal histopathology, changes in
mucociliary clearance were also observed in some of these studies at similar exposure
concentrations. The mucociliary apparatus is an important barrier to infection and exogenous
agents and, thus, is considered as a potential adverse effect. These effects may be due to direct
interaction of formaldehyde with the mucus itself or to Si-induced inflammation in the nasal
tissue that affects mucus production and creation of an effective mucosal barrier.
Andersen and Molhave (1983) reviewed five controlled human studies, one of which
(Andersen and Lundqvist, 1974) examined mucus flow rate in 16 individuals acutely exposed to
0, 0.3, 0.5, 1, or 2 ppm formaldehyde for 4-5 hours in a chamber. Mucus flow rate was
decreased in the anterior and middle third of the ciliated mucosa at 0.3 ppm, but statistical
significance was not determined. This study included smokers and nonsmokers. The small
This document is a draft for review purposes only and does not constitute Agency policy.
4-28 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
sample size, potential confounder effect from smoking, and lack of dose-response relationship
preclude the establishment of a POD.
Holmstrom and Wilhelmsson (1988) demonstrated reduced mucociliary clearance and
nasal mucosal swelling in 70 workers exposed to a median formaldehyde concentration of
0.21 ppm, compared with a referent group of store clerks (n = 36) and was further averaged over
years of exposure. Mucosal swelling and mucociliary activity was measured in the nasal
turbinates. The authors also reported symptoms not only during the weekdays, but also over
weekends and vacation periods. Formaldehyde-exposed subjects self-reported significantly more
nasal discomfort, eye discomfort, deeper airway discomfort, and frequent headache than the
referent group. Groups exposed to formaldehyde had more pronounced mucosal swelling (10.7
nasal resistance score) compared with the reference group (6.5 nasal resistance score). This
difference persisted when data were normalized for differential nasal congestion in the subjects.
Decreased mucociliary activity was seen in 3% of controls and 20% of formaldehyde-exposed
subjects and reached statistical significance (p < 0.05). It is not clear whether impaired
mucociliary clearance was a consequence of altered cell morphology or increased mucus
viscosity. These data provide a LOAEL of 0.21 ppm based on impaired mucociliary clearance.
Thus, mild nasal epithelial lesions observed in formaldehyde-exposed workers have been
observed consistently at levels of about 0.20 ppm to about 2 ppm (Boysen et al., 1990;
Holmstrom et al., 1989; Edling et al., 1988). Of these, Holmstrom et al. (1989) and Edling et al.
(1988) do not appear to be confounded by exposure to wood dust. Nasal biopsy pathology from
formaldehyde-exposed workers is consistent with irritant and reactive properties of
formaldehyde (Ballarin et al., 1992; Boysen et al., 1990; Holmstrom et al., 1989; Edling et al.,
1988; Berke, 1987). Moreover, these findings are supported by results from animal toxicity and
pharmacokinetic and anatomical airflow studies, indicating that, at concentrations less than
1 ppm, inhaled formaldehyde gas does not reach lower regions of the respiratory tract. Of the
available human studies that evaluated histopathology, Holmstrom and Wilhelmsson (1988)
appears to be the most robust and sensitive. The study was carefully designed and included a
large sample of formaldehyde-exposed subjects who were considered separately from workers
exposed to combinations of exposures (formaldehyde and wood dust). Study subjects had been
exposed to formaldehyde regularly for many years. The authors reported not only weekday
exposures but effects reported on weekends and on vacation. Total exposure was carefully
calculated and averaged. The data were controlled for potential confounders, such as smoking.
The endpoint of reduced mucociliary clearance has been substantiated by Andersen and Molhave
(1983) and Holmstrom et al. (1989). Animal studies have also reported formaldehyde-induced
changes on the nasal mucosa and are highlighted in Section 4.2.1.2.
This document is a draft for review purposes only and does not constitute Agency policy.
4-29 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
4.1.1.5. Immunologic Effects
Numerous studies have examined the immunologic responses of individuals exposed to
formaldehyde. This section will discuss four specific areas related to immunotoxicity after
exposure to formaldehyde: increased upper respiratory tract (URT) infections, systemic immune
dysfunction, sensitization and atopy, and production of formaldehyde-protein complexes. Some
studies report increased incidence of URT infections after exposure to formaldehyde (Lyapina et
al., 2004; Krzyzanowski et al., 1990; Holness and Nethercott, 1989). This effect appears to
occur independently of systemic immune changes and may be due to damage to the mucosal
barrier, thus facilitating pathogen access. A number of studies have investigated the hypothesis
that formaldehyde may induce systemic immunomodulation (Ohtani et al., 2004a, b; Erdei et al.,
2003; Thrasher et al., 1990, 1987; Pross et al., 1987). Some studies have also evaluated immune
system effects by investigating the role of reactive oxygen species (ROS) from respiratory burst
associated with immune cells (Lyapina et al., 2004; Gorski et al., 1992) and by assessing
chromosomal damage in immune cells (Orsiere et al., 2006; Yu et al., 2005). In addition to the
effects of formaldehyde on asthmatics and the potential for formaldehyde exposure to exacerbate
asthmatic responses, reviewed in Section 4.1.1.2, numerous studies have investigated whether
formaldehyde may directly induce sensitization and atopic responses by measuring
immunoglobulin E (IgE) levels associated with formaldehyde exposure (Ohmichi et al., 2006;
Vandenplas et al., 2004; Doi et al., 2003; Baba et al., 2000; Palczynski et al., 1999; Krakowiak et
al., 1998; Wantke et al., 1996a, b; Liden et al., 1993; Salkie, 1991; Grammer et al., 1990;
Kramps et al., 1989). Findings are largely negative and suggest that formaldehyde-induced IgE
production is not likely. Lastly, studies have investigated the production of formaldehyde-
specific antibodies, formaldehyde-albumin complexes, and formaldehyde-heme complexes (Kim
et al., 2001; Carraro et al., 1997; Grammer et al., 1993, 1990; Dykewicz et al., 1991; Thrasher et
al., 1990). Heme complex formation is not a strict immunologic endpoint but may trigger
antibody formation and thus it will be discussed in this section. This section will thus summarize
the human studies that have specifically addressed the increased incidence of URT infections,
immunotoxic endpoints, atopy and sensitization, and formation of formaldehyde-heme and
formaldehyde-albumin complexes.
4.1.1.5.1. Increased UR Tinfections. Diverse studies have investigated the possibility that
formaldehyde exposure leads to increased URT infections (Lyapina et al., 2004; Krzyzanowski
et al., 1990; Holness and Nethercott, 1989). Lyapina et al. (2004) studied 29 workers who were
occupationally exposed occupationally to formaldehyde for an average of 12.7 years through
contact with carbamide-formaldehyde glue. The mean values of the average shift concentrations
This document is a draft for review purposes only and does not constitute Agency policy.
4-30 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
of formaldehyde in the application of carbamide-formaldehyde glue to be 0.71 ppm TWA with a
range of 0.32 to 1.57 ppm. The workers were divided into two subgroups, one (n = 12) that
suffered from either a long history (with clinical findings) of chronic mucous inflammation of
the URT with multiple relapses and a second group (n = 17) whose URT inflammations were
short, acute, and predominantly viral. Twenty-one healthy subjects served as controls. A
statistically significant association of self-reported chronic bronchitis and decreased resistance to
URT infection was reported in all the exposed workers compared with controls (p = 0.02). Of
the workers, 41% had a history of chronic respiratory infection and frequent long-lasting
infectious inflammatory relapses (group la). Another group (group lb) consisted of 17 exposed
workers, 12 of whom had no history of recurrent viral infections of the URT. There was a
statistically significant association of frequency and duration of inflammatory relapses between
groups la and lb. No dates were provided regarding when these measurements were made or
over what period of time they were calculated.
Krzyzanowski et al. (1990) measured formaldehyde levels in homes and recorded, by
way of a questionnaire, health histories from adult and child residents. Formaldehyde levels
were reported from samples taken for two 1-week periods in various rooms of the home (kitchen,
living room, subject's bedroom). The average formaldehyde level was 26 ppb in 202 homes, and
levels were stratified into homes with exposure levels below 40 ppb, between 40 and 60 ppb, and
above 60 ppb. Incidences of doctor-diagnosed chronic bronchitis were more prevalent in
children (under age 15) living in homes with higher formaldehyde (>60 ppb) readings in the
kitchen (p < 0.001). This effect was more pronounced (p < 0.001) in children simultaneously
exposed to environmental tobacco smoke. The prevalence of chronic cough was also increased
in adults living in homes with measurable levels of formaldehyde, but data were not shown.
Holness and Nethercott (1989) assessed chronic bronchitis in 87 funeral workers, where the
average formaldehyde exposure was reported at 0.38 ± 0.19 ppm. Chronic bronchitis was
observed in 20 funeral workers (n = 87) exposed to formaldehyde compared with 3 cases of
chronic bronchitis in nonexposed referent controls (n = 38).
These studies suggest that exposure to formaldehyde may be associated with increased
incidence of chronic bronchitis. The mechanism for this association has not been elucidated.
Pathogens may gain access to the URT via a compromised mucosal barrier, as has been shown in
histopathology studies (Section 4.1.1.4).
4.1.1.5.2. Immune function. A number of studies have evaluated the ability of formaldehyde to
induce systemic immunotoxic effects (Ohtani et al., 2004a, b; Erdei et al., 2003; Thrasher et al.,
1990, 1987; Pross et al., 1987). Some studies have reported altered innate immune responses
This document is a draft for review purposes only and does not constitute Agency policy.
4-31 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
associated with formaldehyde exposure (Erdei et al., 2003), while others have noted adaptive
immune response suppression associated with formaldehyde exposure (Thrasher et al., 1990,
1987) and changes associated with alterations to a predominant T—lymphocyte helper 2 (Th2)
pattern (Ohtani et al., 2004a, b). In contrast, Pross et al. (1987) did not observe formaldehyde-
associated changes in systemic immune function.
Erdei et al. (2003) found that Haemophilus influenzae humoral biomarker (H.in.IgG),
Klebsiella pneumoniae biomarker (K.pn.IgG), and elevated monocyte concentrations were
significantly associated with high formaldehyde concentrations in asthmatic children, compared
with nonsensitive children. Briefly, Erdei et al. (2003) compared the immune system responses
in 9- to 11-year-old Hungarian school children whose respiratory systems were immunologically
compromised (chronic respiratory disease, asthma) and normal children who were exposed to
indoor air pollutants, including formaldehyde. In the homes of the children with the highest
levels of pollutants, 49.3% of formaldehyde measurements exceeded the Hungarian indoor
standard of 0.01 ppm, while 20% exceeded the World Health Organization's (WHO's) suggested
indoor level of 0.09 ppm. The authors excluded from consideration all measurements that
exceeded WHO's air quality guidelines in one unidentified city to prevent a "city-related bias,"
since these measurements occurred entirely in that city. The average formaldehyde
concentration in the 123 homes tested was 14 ppm with a range of 0.5 to 46 ppm. H.in.IgG and
K.pn.IgG were significantly associated with high formaldehyde concentrations (p < 0.013 and
p < 0.049, respectively) in sensitive children compared with nonsensitive children. These
markers were also correlated with high levels of nitrogen dioxide, the number of cigarettes
smoked, and exposure to paint, volatile organic compounds, and solvents. Additionally, indoor
formaldehyde exposure was significantly associated with increased monocyte concentrations
(p < 0.017) that are important to the innate immune response (inflammation) in diseased tissue.
The authors concluded that the elevation of immune biomarkers in sensitive children with
respiratory disease is likely the result of high concentrations of toxic indoor air pollutants,
including formaldehyde.
Thrasher et al. (1987) assessed the effects of formaldehyde exposure on cellular
immunity and antibody formation in eight exposed and eight unexposed individuals. The
exposed group consisted of three males and five females. Seven of the exposed individuals had
resided in mobile homes for periods ranging from 2 to 7 years; the eighth was a laboratory
worker who resided in a newly decorated, energy-efficient apartment. Air monitoring in four of
the homes revealed formaldehyde vapor concentrations ranging from 0.07 to 0.55 ppm. Venous
blood samples were collected from all subjects and T- and B-cells were counted and monitored
for blastogenesis. When IgG and IgE antibodies to formaldehyde were monitored in serum, no
This document is a draft for review purposes only and does not constitute Agency policy.
4-32 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
IgE antibodies to formaldehyde were detected in exposed or control subjects. IgG antibody titers
in exposed subjects ranged from 1:8 to 1:256 but essentially were undetected (1:4) in seven of
the controls. T- and B-cell numbers were significantly lower (p < 0.05) in mobile home residents
(48 and 12.6%, respectively) compared with those of control subjects (65.9 and 14.75%,
respectively). As determined by incorporation of 3H-thymidine into 48-hour unaltered
lymphocytes, phytohemagglutinin-stimulated T- and B-cell blastogenesis was significantly
depressed (p < 0.01) in cells of mobile home residents compared with those of control subjects
(17,882 and 28,576 cpm, respectively). Thrasher et al. (1987) concluded that exposure to
formaldehyde decreases the proportion of peripheral T cells.
In a later study, Thrasher et al. (1990) evaluated five groups of subjects with varying
levels and durations of formaldehyde exposure. The groups consisted of (1) asymptomatic
chiropractic students exposed during anatomy classes (controls with only intermittent exposure
to formaldehyde), (2) mobile home residents, (3) office workers, (4) patients with multiple
symptoms who had been removed from the source of formaldehyde for at least a year, and
(5) occupationally exposed patients. All groups were assessed for immunologic function via
white cell, lymphocyte, and T-cell counts, T-helper/suppressor ratios and B-cell counts. When
compared with controls (chiropractic students), the patient groups had significant elevations in
formaldehyde antibody titers and B-cell titers.
Ohtani et al. (2004a, b) reported effects of exposure to formaldehyde and diesel exhaust
particles on cytokine production by human monocyte-derived dendritic cells (MoDCs) and
T cells in vitro. Dendritic cells were stimulated with CD40 ligand and interferon (IFN)-y, T cells
with anti-CD3/CD28 antibodies. Cytokine proteins and mRNA levels were measured in
supernatants by enzyme-linked immunosorbent assay (ELISA) and polymerase chain reaction
(PCR), respectively. Formaldehyde and diesel exhaust particles significantly increased tumor
necrosis factor (TNF)-a levels and suppressed interleukin (IL)-12p40 protein and mRNA levels
in MoDCs. The same treatment suppressed protein synthesis and mRNA expression of IFN-y
and IL-10 in T cells. The authors concluded that their findings support a role of formaldehyde
and diesel exhaust particles in altering the immune response to a Th2-dominant pattern that
furthers allergic inflammation. Further details, such as exposure concentrations and
experimental protocols, are not available.
In contrast, Pross et al. (1987) concluded that formaldehyde does not induce altered
immune activity. The authors evaluated the immunologic response of asthmatic subjects
exposed to UFFI off-gas products. Subjects consisted of 23 individuals with a history of
asthmatic symptoms attributed to UFFI and 4 individuals with asthma unrelated to UFFI off-gas
products. All subjects were exposed in an environmental chamber according to the following
This document is a draft for review purposes only and does not constitute Agency policy.
4-33 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
sequence: (1) room air (placebo) for 30 minutes; (2) 1 ppm formaldehyde gas for 3 hours; (3)
UFFI particles (4 [j,m diameter, 0.5 particles/mL) for 3 hours, commencing 48 hours after
formaldehyde gas exposure; and (4) UFFI off-gas products for 3 hours, commencing 48 hours
after UFFI particle exposure. There was a significant increase in the percentage and absolute
number of eosinophils and basophils in the subjects who lived in UFFI homes but no differences
between exposure groups with respect to lymphocyte subpopulations either before or after UFFI
exposure. However, when T8 suppressor cells were counted, values in the UFFI-exposed group
pre-exposure and postexposure, a small but statistically significant (p < 0.01) increase in T8 cell
count was observed. The biological significance of this increase in T8 cell count in exposed
asthmatics is not known. Pross et al. (1987) concluded that short-term exposure to formaldehyde
was not immunosuppressive and did not result in systemic immune reactivity.
Respiratory burst from immune cells creates ROS that can incur further cellular damage.
Several studies have evaluated, either directly or indirectly, the potential role of ROS as potential
mediators of formaldehyde-associated effects, particularly those caused by immune cells. Gorski
et al. (1992) measured chemiluminescence resulting from the release of free radicals from
granulocytes of healthy and formaldehyde-sensitive subjects. Thirteen subjects with contact
dermatitis who were occupationally exposed to formaldehyde and five healthy volunteers
participated in the study. All underwent skin-prick tests for common allergens as well as a
histamine inhalation provocation test. Subjects were exposed to 0.5 mg/m3 (0.41 ppm)
formaldehyde for 2 hours, and the PEFR was measured immediately before exposure, after 60
and 120 minutes of exposure, and 6 and 21 hours after completion of exposure. Peripheral blood
granulocyte chemiluminescence was measured in the presence of luminol. Free radical
production was increased significantly within 30 minutes of beginning the exposure in subjects
with allergic dermatitis and remained elevated for 24 hours compared with baseline values.
Gorski et al. (1992) concluded that granulocyte chemiluminescence did not increase in healthy,
formaldehyde-exposed patients but was diagnostic for formaldehyde-sensitive patients. These
results also suggest a putative role for oxidative damage associated with formaldehyde exposure,
particularly in sensitized individuals.
Lyapina et al. (2004) also reported effects of formaldehyde exposure on neutrophil
respiratory burst activity (NRBA), the capacity of polymorphonuclear leukocytes to produce
reactive oxygen radicals in response to chemical or microbial stimuli using flow cytometry.
Briefly, Lyapina et al. (2004) studied 29 workers who were occupationally exposed to
formaldehyde for an average of 12.7 years through contact with carbamide-formaldehyde glue
with a mean value of the average shift concentration of formaldehyde reported as 0.71 ppm
TWA with a range of 0.32 to 1.57 ppm. The workers were divided into two subgroups, one (n =
This document is a draft for review purposes only and does not constitute Agency policy.
4-34 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
12) that suffered from either a long history (with clinical findings) of chronic mucous
inflammation of the URT with multiple relapses, and a second group (n = 17) whose URT
inflammations were short, acute, and predominantly viral. Twenty-one healthy subjects served
as controls. A suite of hematological tests and flow cytometric analysis for respiratory burst
activity were performed. Although no significant difference was observed in the spontaneous
and stimulated NRBA (median percentage of oxidizing cells) between the 29 exposed workers
with URT inflammation and the healthy controls (0.83 versus 1.35, respectively), a separate
comparison of the NRBA of 12 workers with chronic, repeating URT infections and 17 workers
with short, infrequent episodes of URT inflammations was significant (0.45 versus 1.00, p =
0.037). When the NRBA of the group with chronic URT infections (n = 12) was separately
compared with that of the healthy controls (n = 21), the results were also significant (0.45 versus
1.35,/? = 0.012). Individuals with chronic URT infections have reduced NRBA that could be
due to formaldehyde exposure. Neutrophils respond to tissue damage or local invasion of
microorganisms and act to phagocytize foreign cells. If neutrophilic activity is hampered or
altered by formaldehyde exposure, then the ability to fight infection will be diminished, leading
to prolonged infection. However, no dose-response pattern of formaldehyde exposure could be
determined from this study.
Other investigators have reported chromosomal damage in immune cells due to
formaldehyde (Orsiere et al., 2006; Yu et al., 2005). Yu et al. (2005) evaluated chromosomal
damage in lymphocytes from 151 exposed and nonexposed workers from a plywood factory
detected by comet assay. The authors reported that chromosomal damage was statistically
elevated in lymphocytes from formaldehyde-exposed workers compared with controls.
However, no information on exposure duration or levels was provided. Orsiere et al. (2006)
studied DNA damage in lymphocytes from 59 hospital employees with formaldehyde exposures
from pathology and anatomy laboratories in five hospitals. Controls were 37 workers from the
same hospitals, matched on gender, age, and smoking habits, with no known exposure to
genotoxic agents. Study participations were excluded if workers had a history of radio- or
chemotherapy or had used therapeutic medications that were known to be mutagenic.
Occupational exposure was determined through 15-minute and 8-hour personal air sampling
during a typical workday. Mean formaldehyde concentrations were 2 ppm (range: <0.1-
20.4 ppm) for 15-minute sampling and 0.1 ppm (range: <0.1-0.7 ppm) for 8-hour sampling. No
change in DNA damage was found between the beginning and end of the workday among
exposed workers (3.9 ± 0.6 versus 3.6 ± 0.5 relative light units/ng DNA). However, exposed
workers had significant elevations in the binucleated micronucleated cell rate (BMCR) per
1,000 cells compared with controls (16.9 ± 9.3 versus 11.1 ± 6.0%; /? < 0.001), but BMCR did
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
not appear to be correlated with formaldehyde concentration. Linear regression analysis showed
that the effect for exposure remained after adjusting for gender, age, smoking, and drinking
habits. For 18 exposed and 18 control workers who underwent cytokinesis-blocked
micronucleus assay (CBMA) combined with fluorescent in situ hybridization (FISH) with pan-
centromeric DNA probe, results showed that the frequency of micronuclei (MN) containing only
one centromere (Cl+MN) was elevated among the exposed compared with unexposed workers
(11.0 ± 6.2% versus 3.1 ± 2.4%; p < 0.001). The effect of exposure remained significant after
controlling for gender, age, smoking, and drinking habits. Results from Yu et al. (2005) and
Orsiere et al. (2006) suggest that formaldehyde exposure may promote chromosomal damage
leading to micronucleated lymphocytes.
Compromised lymphocyte function may significantly contribute to altered immune
status. The mechanism underlying this effect has not been elucidated.
4.1.1.5.3. Sensitization and atopy. Numerous studies have documented formaldehyde-induced
exacerbation of asthmatic responses (see Section 4.1.1.2). The mechanism of this effect has not
been clarified and has led investigators to assess the potential for formaldehyde to directly induce
formation of formaldehyde-specific antibodies, leading to allergic responsiveness. One case
report showed systemic allergic reactions (e.g., anaphylaxis) to formaldehyde in a patient
undergoing hemodialysis (Maurice et al. [1986] referenced in Thrasher et al. [1990]). Some
studies have evaluated the potential association of formaldehyde-specific IgE in already-
sensitized individuals (Baba et al., 2000; Palczynski et al., 1999). Other studies have
investigated whether formaldehyde can directly induce IgE in nonsensitized individuals. Most of
the studies have not identified presence of formaldehyde-specific IgE (Ohmichi et al., 2006;
Krakowiak et al., 1998; Grammer et al., 1993, 1990; Kramps et al., 1989; Thrasher et al., 1987)
and are summarized below. A few studies (Vandenplas et al., 2004; Doi et al., 2003; Liden et
al., 1993) reported positive IgE against formaldehyde, associated with exposure, but the IgE
titers were either transient (Vandenplas et al., 2004) or were positive in a small subset of
previously sensitized subjects (2 of 15) (Liden et al., 1993). Doi et al. (2003) detected IgE
against formaldehyde in two asthmatic children (out of 122 asthmatic children), but the response
severity did not correlate with exposure level.
Palczynski et al. (1999) evaluated whether exposure to formaldehyde might facilitate
specific sensitization to common allergens. The study population was comprised of residents of
apartments built in 1989-1990. Only households with children from 5-15 years were eligible for
the study. A random sample of 120 apartments was selected in which lived a total of 465
persons aged 5-65 years. Individual demographic characteristics and medical histories were
This document is a draft for review purposes only and does not constitute Agency policy.
4-36 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
determined by questionnaire. Residents were tested, using the skin-prick method, for allergen
response to a variety of materials, such as household dust, pollens, and feathers. Total serum IgE
levels were measured, and the presence of formaldehyde-specific IgE antibodies was determined.
Measured mean levels of formaldehyde were 21.05 ± 8.94 ppb. No significant relationship
between respiratory allergy prevalence and indoor exposure to formaldehyde was detected.
Significant increases in serum IgE levels were found in children exposed to both environmental
tobacco smoke and formaldehyde.
Baba et al. (2000) investigated whether production of formaldehyde-specific IgE could be
detected in adult asthmatics. Formaldehyde exposure levels were not documented.
Formaldehyde-IgE was detected in two asthmatic patients (n = 80), one male and one female, but
the titer of IgE did not parallel the severity of the asthmatic responses and could not be linked to
formaldehyde exposure. Thus, formaldehyde-specific IgE-mediated allergy was rare in adult
chronic asthmatics.
Several studies have examined serum for formaldehyde-specific IgE antibodies in groups
of formaldehyde-exposed humans (Ohmichi et al., 2006; Krakowiak et al., 1998; Wantke et al.,
1996a, b; Salkie, 1991; Grammer et al., 1990; Kramps et al., 1989). While formaldehyde-
specific IgE was reported in one study (Wantke et al., 1996a), results from most other studies
failed to find a consistently strong association between formaldehyde-specific IgE or IgG
antibodies in groups of formaldehyde-exposed humans.
Wantke et al. (1996a) detected elevated levels of formaldehyde-specific IgE in 24 of 62
8-year-old children who were students in three particleboard-paneled classrooms in which the
estimated formaldehyde air concentrations were 0.075, 0.069, and 0.043 ppm. In a health
survey, the children reported headaches (29/62), fatigue (21/62), dry nasal mucosa (9/62), rhinitis
(23/62), cough (15/62), and nosebleeds (14/62). The number of children with symptoms in each
classroom decreased with decreasing formaldehyde concentration (49, 47, and 24, respectively,
for the 0.075, 0.069, and 0.043 ppm classrooms). However, the investigators reported that
elevated levels of specific IgE did not correlate with the number and severity of symptoms.
When the children were evaluated after 3 months in a new school that did not have particleboard
paneling and had lower ambient formaldehyde concentrations (0.029, 0.023, and 0.026 ppm), the
number of children reporting symptoms decreased significantly from earlier figures, and, when
measured in 20 of the children, the mean serum levels of formaldehyde-specific IgE declined
significantly compared with pre-moving mean levels.
In contrast, a study by Krakowiak et al. (1998) measured serum IgE levels in asthmatic
and healthy subjects as part of a larger study to characterize the mechanism of
formaldehyde-induced nasal and bronchial response in asthmatic subjects with suspected
This document is a draft for review purposes only and does not constitute Agency policy.
4-37 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
formaldehyde allergy. Ten subjects reported to have formaldehyde rhinitis and asthma and
10 healthy subjects underwent a 2-hour inhalation challenge in an exposure chamber with
formaldehyde at a concentration of 0.5 mg/m3 (0.41 ppm). Formaldehyde-specific serum IgE
antibodies were measured, and cellular, biochemical, and mediator changes were assessed in
nasal lavage before, immediately after, and at 4 and 24 hours after challenge. Challenges with
formaldehyde caused only transient symptoms of rhinitis in both groups. Furthermore, none of
the subjects thought to have occupational asthma developed clinical symptoms of bronchial
irritation. No specific IgE antibodies to formaldehyde were detected in persons with
occupational exposure to formaldehyde. No differences in the nasal response to formaldehyde
were found between subjects reported to have occupational allergic respiratory diseases and
healthy subjects (p > 0.05). The study showed that inhaled formaldehyde at a level as low as
0.5 mg/m3 did not induce a specific allergic response either in the upper or in the lower part of
the respiratory tract. In addition, it demonstrated that there was no difference in nasal response
to formaldehyde between asthmatic subjects occupationally exposed to formaldehyde and
healthy subjects.
Similarly, formaldehyde-specific IgE antibodies were detected in only 1 serum sample
(out of 86) from four groups of formaldehyde-exposed subjects (Kramps et al., 1989). The
subject with detected formaldehyde-specific IgE displayed allergic symptoms. The groups
included (1) 28 subjects living or working in places with formaldehyde-containing construction
materials (e.g., chipboard) and estimated formaldehyde concentrations ranging from 0.08 to
0.37 ppm, (2) 18 occupationally exposed subjects from an anatomy laboratory and in other
unspecified industries where air concentrations were not measured, (3) 12 hospital attendants
who worked with formaldehyde-sterilized hemodialysis equipment, and (4) 28 hemodialysis
patients coming into contact with equipment sterilized with formaldehyde. Other subjective
symptoms, such as headache, eye irritation, and respiratory complaints, were reported by
24/28 subjects in the construction material group and confirm that formaldehyde is an irritant
(reviewed in Section 4.1.1.1). Durations of exposure or length of employment were not reported
for the subjects in this study.
Grammer et al. (1990) studied the immunologic nature of formaldehyde sensitivity in
37 workers who were examined by a group of physicians in response to complaints of
formaldehyde-related illness. Air sampling of formaldehyde ranged from 0.003 to 0.078 ppm,
but specific levels were not tied to specific workplace areas. Blood samples were collected and
assayed for IgE and IgG activity against formaldehyde. None of the workers had IgG activity
against formaldehyde. No IgE antibodies were detected in the other 32 workers. The authors
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
concluded that there was no evidence of an immunologically mediated response to formaldehyde
in this group of workers.
Formaldehyde-specific IgE was not detected in any of a group of 45 medical students
before or after the students attended a 4-week anatomy dissecting course (Wantke et al., 1996b).
Estimates of ambient air concentrations of formaldehyde ranged from 0.059 to 0.219 ppm (0.124
± 0.05 ppm; mean ± SD). However, the survey revealed frequencies of irritation symptoms that
were consistent with other studies (e.g., itching of the skin in 33/45 students, headache in 15/45,
and burning eyes in 13/45).
Similarly, Ohmichi et al. (2006) were unable to correlate formaldehyde exposure with
specific IgE production among eight students attending a gross anatomy laboratory.
Formaldehyde exposure was estimated to range from 0.33 to 1.47 ppm during the laboratory
sessions. The sample size was small, and IgE levels varied substantially (ranging from <19 to
>5,000 international units/mL). Compared with IgE levels taken 90 minutes prior to the start of
the first session, IgE levels measured shortly after the last session and up to 23 days following
the last session showed no association with exposure.
Salkie (1991) investigated the prevalence of formaldehyde-specific IgE in practicing
pathologists who complained of formaldehyde sensitivity. Exposure levels were not reported.
Serum samples were assayed for total IgE and formaldehyde-specific IgE. Of the 46 subjects,
29 self-reported atopy that was confirmed in 12 subjects by positive IgE. Moreover, 29 subjects
complained of formaldehyde-specific sensitivity. However, zero subjects had formaldehyde-
specific IgE, and there was no evidence that atopic individuals were more sensitive to
formaldehyde than non-atopic individuals. The authors noted that atopic individuals may have
selectively reduced their exposure to formaldehyde.
Vandenplas et al. (2004) evaluated a case study of a 31-year-old male who was
accidentally exposed to formaldehyde for 2 hours. The exposure level was not provided. The
subject had smoked a pack of cigarettes a day for 13 years and was admitted to the emergency
room for asthmatic symptoms. Eight days following exposure, increased levels of
formaldehyde-specific IgE antibodies were detected but could not be detected in subsequent
assessments.
A clinical study by Liden et al. (1993) evaluated IgE-specific antibodies against
formaldehyde in 23 subjects who had previously tested positive for skin sensitization by skin
prick test. Subjects were exposed to formaldehyde by skin patch (1% formaldehyde in water).
Ten of the subjects were classified as atopic Though 15 of 23 of the sensitized subjects were
also sensitive to formaldehyde applied by skin patch, formaldehyde-IgE was positive in 2 of 15
This document is a draft for review purposes only and does not constitute Agency policy.
4-39 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
individuals who were not classified as atopic. No dose-response relationship could be
determined from the study design of this study.
Doi et al. (2003) conducted a clinical study in 155 children of which 122 were
asthmatics. No specific exposure to formaldehyde was documented. IgE against formaldehyde
was determined in blood. Formaldehyde-specific IgE was found in two asthmatic children.
Thus, while several studies have documented formaldehyde-specific IgE, the occurrence is rare
and may be transient. Asthmatic children may be more predisposed to form formaldehyde-
specific IgE than non-atopic individuals or adults. The formation of formaldehyde-specific IgE
is quite rare.
4.1.1.5.4. Formaldehyde-albumin and formaldehyde-heme complexes. Numerous studies have
shown that formaldehyde can bind to blood proteins as formaldehyde-heme and formaldehyde-
human serum albumin (formaldehyde-HSA) complexes (Carraro et al., 1997; Grammer et al.,
1993, 1990; Dykewicz et al., 1991; Thrasher et al., 1990). Kim et al. (2001) failed to identify
IgE against formaldehyde-HSA complexes in one case-control subject following industrial
occupational exposure to formaldehyde. These complexes may serve to traffic formaldehyde
throughout the bloodstream and throughout the body. While formaldehyde may be too small to
engender an immune response, these complexes may be able to trigger formaldehyde-protein-
specific antibodies, leading to an immune response, including sensitization.
Thrasher et al. (1990) evaluated five groups of subjects as follows with varying levels
and durations of formaldehyde exposure: asymptomatic chiropractic students exposed during
anatomy classes (controls with only intermittent exposure to formaldehyde), mobile home
residents, office workers, patients with multiple symptoms who had been removed from the
source of formaldehyde for at least a year, and occupationally exposed patients. All groups were
assessed for production of antibodies against formaldehyde-HSA. The level of autoantibodies
was significantly elevated in patients exposed long-term to formaldehyde. From the data,
Thrasher et al. (1990) concluded that exposure to formaldehyde stimulates IgG antibody
production to formaldehyde-HSA.
Grammer et al. (1990) studied the immunologic nature of formaldehyde sensitivity in
37 workers who were examined by a group of physicians in response to complaints of
formaldehyde-related illness. Air sampling of formaldehyde ranged from 0.003 to 0.078 ppm,
but specific levels were not tied to specific workplace areas. Blood samples were collected and
assayed for IgE and IgG activity against formaldehyde and formaldehyde-HSA. None of the
workers had IgG activity against formaldehyde. Five workers had IgE against both HSA alone
and against formaldehyde-HSA complexes. No IgE antibodies were detected in the other 32
This document is a draft for review purposes only and does not constitute Agency policy.
4-40 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
workers. The authors concluded that there was no evidence of an immunologically mediated
response to formaldehyde in this group of workers.
Grammer et al. (1993) described the evaluation of a worker with bronchospasm
symptoms caused by formaldehyde exposure. The worker was evaluated by means of ELISA,
cutaneous tests, and methacholine and formaldehyde inhalation challenges. The ELISA showed
that the worker had positive IgE and IgG titers to formaldehyde-HSA. The worker also had a
positive cutaneous test for formaldehyde-HSA but a negative methacholine challenge at 25
mg/mL and negative formaldehyde inhalation challenges at exposure concentrations of 0.3, 1, 3,
and 5 ppm for 20 minutes. The worker might have developed a positive response if a higher
concentration of formaldehyde had been used for the challenge, but it is more probable that the
worker's symptoms were not caused by immunologically mediated asthma.
Dykewicz et al. (1991) evaluated whether IgE or IgG antibodies to formaldehyde were
related to formaldehyde exposure or to respiratory symptoms arising from such an exposure.
The authors studied 55 potentially exposed subjects (hospital histology technicians, internal
medicine residents, pathology residents, current smokers, and persons with known workplace
exposure to formaldehyde) and compared them to controls with no history of formaldehyde
exposure. Reported workplace formaldehyde concentrations were 0.2-0.64 ppm for pathology
residents, 0.64 ppm for histology technicians, and 0.6-11 ppm for miscellaneous formaldehyde
exposure scenarios. Workplace air concentrations were not measured for the other occupations.
Occupational exposure to formaldehyde averaged 12.45 years for histology technicians,
0.38 years for medical residents, 3.21 years for pathology residents, and 18.34 years for five
subjects exposed to formaldehyde in miscellaneous workplaces. Blood samples were analyzed
for IgE and IgG reactivity with formaldehyde-HSA complexes. Three subjects had IgE against -
HSA; these three and two others had low levels of anti-formaldehyde-HSA IgG. The presence of
IgG and IgE antibodies to formaldehyde was not clearly related to formaldehyde exposure or
pack-years of smoking. One subject had both IgE and IgG antibodies and also suffered from eye
and respiratory symptoms when exposed to formaldehyde at his workplace. However, the
authors concluded that they could not establish a relationship between IgE and IgG levels and
formaldehyde exposure. This study has several limitations. First, the volunteers (hospital
workers) may not be representative of exposed workers in the general population. One of the
exposure groups comprised cigarette smokers. Although the study focused on formaldehyde
antibodies, which would be unaffected by the other chemicals, respiratory symptoms among
smokers would reflect exposures to the constituents of smoke. Dykewicz et al. (1991) concluded
that immunologically mediated asthma caused by formaldehyde is extremely rare and may not
exist at all.
This document is a draft for review purposes only and does not constitute Agency policy.
4-41 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
Carraro et al. (1997) reported development of a reliable assay to effectively measure
formaldehyde-HSA complexes in smokers, ex-smokers, and nonsmokers. A correlation between
formaldehyde-HSA antibodies and smoking status was detected. This study did not correlate
formaldehyde exposure and formaldehyde-HSA antibodies.
Given that formaldehyde is a sensory irritant that is particularly bothersome to
individuals with compromised lung function or asthma, numerous studies have assessed the
ability of formaldehyde to induce immunotoxic effects. Some studies have documented
increased rates of chronic bronchitis and URT infections associated with exposure to
formaldehyde, which suggests a possible immunomodulatory effect. However, of the numerous
articles that have investigated systemic immunomodulatory effects due to formaldehyde
(Lyapina et al., 2004; Gorski et al., 1992; Thrasher et al., 1990, 1988, 1987; Pross et al., 1987),
few have reported significant immune modulation related to formaldehyde exposure. Significant
decreases in specific adaptive immune cell populations do not appear correlated to formaldehyde
exposure (Erdei et al., 2003; Gorski et al., 1992; Thrasher et al., 1990, 1987; Pross et al., 1987).
Thus, the tendency for increased infection rates associated with formaldehyde may not be related
to altered immune function. Perhaps altered mucociliary clearance and disturbed mucosal barrier
may provide greater access for pathogens and result in greater infection rates. Moreover,
formaldehyde has been associated with exacerbation of asthmatic or atopic responses,
particularly in sensitized individuals. However, this effect does not appear to occur by increased
IgE or formaldehyde-specific IgE levels (Ohmichi et al., 2006; Palczynski et al., 1999;
Krakowiak et al., 1998; Wantke et al., 1996b; Salkie, 1991; Grammer et al., 1990; Kramps et al.,
1989). Thus, formaldehyde-associated enhanced allergic responses does not appear to be due to
direct induction of sensitization and may not occur via an immunologic mechanism. Lastly, the
formation of formaldehyde-heme and formaldehyde-HSA complexes has been well documented
(Grammer et al., 1993, 1990; Dykewicz et al., 1991; Thrasher et al., 1990) and may serve as a
biomarker of exposure (Carraro et al., 1997). Moreover, these complexes may provide a means
by which formaldehyde travels throughout the bloodstream and may drive antibody formation
that may lead to immune activation.
4.1.1.6. Neurological/Behavioral
There is some suggestion of neurological impairment in humans following occupational
exposure to formaldehyde; the data are limited and the results from several studies are potentially
confounded by exposure to other solvents. Two studies of histology technicians with
occupational exposure to formaldehyde and other solvents found neurological deficits and poorer
performance on neurocognitive tests associated with formaldehyde exposure (Kilburn et al.,
This document is a draft for review purposes only and does not constitute Agency policy.
4-42 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
1987, 1985). In another study, Kilburn and Warshaw (1992) found no change from initial
performance, for as long as 4 years, in follow-up evaluations of histology technicians with
continuing exposure to formaldehyde. In a preliminary report from a prospective study,
Weiskopf et al. (2009) found a strong association between duration of formaldehyde exposure
and death from amyotrophic lateral sclerosis (ALS). In a controlled exposure study, Bach et al.
(1990) found that, when workers with chronic formaldehyde exposure were challenged with an
acute formaldehyde exposure, they exhibited poorer performance on some neurocognitive tests
compared with workers without chronic exposure undergoing the same acute challenge
conditions. In another controlled exposure study, Lang et al. (2008) found equivocal changes in
reaction time following an acute exposure.
4.1.1.6.1. Epidemiological studies. Kilburn et al. (1985) reported that a group of 76 female
histology technicians displayed statistically significantly greater frequencies of neurobehavioral
deficits (lack of concentration and loss of memory, disturbed sleep, impaired balance, variations
in mood, and irritability), than did a referent group of 56 unexposed female clerical workers.
The technicians had been employed from 2 to 37 years (mean 12.8 years). Analysis of
workplace air samples indicated the presence of several solvents, ranging from 0.2 to 1.9 ppm for
formaldehyde, 3.2 to 102 ppm for xylene, 2 to 19.1 ppm for chloroform, and 8.9 to 12.6 ppm for
toluene. Subsequently, Kilburn et al. (1987) administered a battery of 10 tests to 305 female
histology technicians to assess various aspects of cognitive and motor function. The researchers
analyzed the results by regression analysis with age, years of smoking, and hours per day of
exposure to formaldehyde and other solvents as explanatory variables. Increased daily hours of
exposure to formaldehyde were significantly correlated with decreased performance in several
tests (including several types of memory, dexterity, and balance), whereas hours of daily
exposure to other solvents were only correlated with decreased performance in a single memory
test. In a later prospective study of performance, 318-494 histology technicians were tested in a
battery of neurobehavioral tests, and testing for a subset of subjects was repeated yearly for up to
4 years. No statistically significant decrement in performance was found when initial test results
were compared with retest results to evaluate effects of continuing occupational exposure to
formaldehyde (or other solvents) or possible effects of aging (Kilburn and Warshaw, 1992).
Kilburn (1994) later reported that three anatomists and one railroad worker, occupationally
exposed to airborne formaldehyde for 14-30 years, each showed impaired performance on
several neurobehavioral tests (e.g., choice reaction time, abnormal balance, digit symbol, and
perceptual motor speed).
This document is a draft for review purposes only and does not constitute Agency policy.
4-43 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
Weisskopf et al. (2009) evaluated the association between chemical exposure and death
from ALS, using the cohort of 987,229 people from the prospective Cancer Prevention Study II
of the American Cancer Society. From 1989-2004, 1,156 deaths from ALS were identified from
mortality records from the National Death Index. Exposure assessment occurred prior to follow-
up and was based on a questionnaire; participants were asked about current exposure to 12 types
of chemicals and whether they had been regularly exposed in the past. After controlling for a
number of potentially confounding factors (including age, sex, smoking status, military service,
education, alcohol intake, occupation, vitamin use, and exposure to other chemicals), it was
found that exposure to formaldehyde for a known duration was statistically significantly
associated with increased risk of death from ALS (p < 0.0001) with a relative risk (RR) of 2.47
(95% CI: 1.58-3.86) based on 22 deaths. Weisskopf et al. (2009) reported that the association
had a strongly significant dose-response relationship, with increased duration of exposure
associated with increased RR of ALS mortality with a reported p value for continuous trend of
0.0004. Multivariate adjusted rate ratios were 1.5 for known formaldehyde exposures less than
4 years, 2.1 for 4-10 years, and 4.1 for >10 years. Although the authors indicated that these
results need independent verification, the results of this study of the nearly one million people
followed for 15 years is unlikely to be biased due to its longitudinal design.
4.1.1.6.2. Controlled exposure studies. Bach et al. (1990) examined whether cognitive and
motor performance of humans responded acutely to formaldehyde exposure and whether
previous chronic exposure to formaldehyde affected the responses observed following acute
exposure. Thirty-two men with at least 5 years of occupational exposure to formaldehyde and
29 matched controls were exposed to formaldehyde at concentrations of 0.04, 0.21, 0.48, or
1.10 ppm for 5.5 hours. During the exposure period, symptoms were assessed by using a
standardized questionnaire, and subjects were evaluated in four tests designed to estimate several
aspects of cognitive function. Testing was performed once prior to exposure and twice during
the exposure period. The authors noted that the typical dose-related symptoms of respiratory
irritation were not seen in this study. Previously unexposed subjects reported more headaches,
"heavy head," and physical tiredness than the exposed workers. In both occupationally exposed
and unexposed subjects, decreased performance in an addition test was significantly correlated
with increasing concentration of formaldehyde. Compared with previously unexposed subjects,
occupationally exposed subjects showed significantly decreased performance in three other tests,
although the effect was not dose related. The study did not adjust for several potential
confounders, including prior exposure to other chemical agents, and the age and health status of
the cases and controls. Authors concluded that their data indicated that acute exposure to
This document is a draft for review purposes only and does not constitute Agency policy.
4-44 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
formaldehyde might cause acute effects on CNS functions but that more investigation was
needed to verify their results.
In a study evaluating chemosensory irritation, Lang et al. (2008) assessed possible
changes in reaction time during an acute (4-hour) exposure to formaldehyde concentrations
between 0-0.5 ppm (some exposure sessions also included short concentration peaks of up to
1 ppm) with or without a masking agent (ethyl acetate). Twenty-one healthy volunteers were
exposed once per day to each of 10 different exposure combinations in random order (for a total
of 10 sessions per subject). Reaction time was tested before and after each exposure session.
Significant increases in reaction time were seen at 0.3 ppm formaldehyde, with or without
masking agent, but not at 0.5 ppm. The significance of these findings is unclear.
Performance of 16 healthy volunteers on addition, multiplication, and card punching
tasks was measured by Andersen and Molhave (1983) before and during a 5-hour exposure to
formaldehyde at concentrations up to 2 mg/m3. The authors reported that formaldehyde
exposure had no effect on performance, but results were not presented.
4.1.1.6.3. Summary. The limited information currently available from human studies does not
permit a definitive conclusion regarding an association between formaldehyde exposure and
human neurotoxicity. There is, however, sufficient information to raise a serious concern for this
type of effect, and additional studies are needed.
4.1.1.7, Developmental and Reproductive Toxicity
Epidemiologic studies suggest a convincing relationship between occupational exposure
to formaldehyde and adverse reproductive outcomes in women. Several of these studies deal
with spontaneous abortion following maternal occupational formaldehyde exposure (Taskinen et
al., 1999, 1994; John et al., 1994; Seitz and Baron, 1990; Hemminki et al., 1985, 1982; Axelsson
et al., 1984), but not all reported a significant association between exposure and spontaneous
abortion. A study of fecundability found an increase in time to pregnancy among female
workers exposed to formaldehyde (Taskinen et al., 1999). Three studies that examined the effect
of occupational exposures on the incidence of congenital malformation produced mixed results
(Dulskiene and Grazuleviciene, 2005; Taskinen et al., 1994; Hemminki et al., 1985). A
population-based, semi-ecologic study found an association between atmospheric formaldehyde
exposure and low birth weight (Grazuleviciene et al., 1998).
4.1.1.7.1. Spontaneous abortion. Several epidemiologic studies report a relationship between
occupational exposure to formaldehyde and increases in risk of spontaneous abortion following
maternal occupational formaldehyde exposure (Taskinen et al., 1999, 1994; John et al., 1994;
This document is a draft for review purposes only and does not constitute Agency policy.
4-45 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
Seitz and Baron, 1990; Axelsson et al., 1984). Increased RRs were in the range of 1.7 to more
than 3.0. However, other studies (Hemminki et al., 1985, 1982) found no association between
occupational formaldehyde exposure and spontaneous abortion. Paternal occupational exposure
to formaldehyde was not related to spontaneous abortion (Lindbohm et al., 1991).
The earliest report of an association between spontaneous abortion and formaldehyde
exposure comes from a Swedish cohort study of female laboratory workers (Axelsson et al.,
1984). Subjects were women born in 1935 or later and worked in a university laboratory during
1968-1979. There were 745 women who responded to a mailed questionnaire (response rate =
95%), 556 of whom reported on 1,180 pregnancies that resulted in 997 births. Exposure to
formaldehyde was estimated based on answers to the questionnaires. Formaldehyde exposure
was reported only in connection with 10 pregnancies, of which 5 went to term, 3 were reported
as miscarriages, and 2 were terminated by induced abortion. Excluding the latter, the
spontaneous abortion rate among women exposed to formaldehyde in the first trimester was 3/8
(37.5%) compared with 14/148 (9.5%) in the population of laboratory workers not exposed to
any solvent in the first trimester.
While not computed by the authors, the OR can be calculated as 5.7 (95% CI: 1.2-26.6).
The exposure assessment on which this result is based was methodologically weak but unlikely
to be a source of bias. Given the exploratory nature of this study, potential confounders were not
controlled for, but no other co-exposure was more strongly related to the increased risk of
miscarriage, so this result is not likely to be explained by confounding. Selection bias is also an
unlikely explanation given the high participation rate. However, although this association of an
increased risk of pregnancy loss with formaldehyde exposure is statistically significant, the CI is
wide and chance may be a possible explanation for this finding.
In a 1988 Health Hazard Evaluation, the National Institute for Occupational Safety and
Health (NIOSH) investigated complaints of adverse reproductive outcomes at a plant where
work pants were cut and sewn with a fabric that was treated with a resin that releases
formaldehyde (Seitz and Baron, 1990). In a NIOSH laboratory, the fabric released 163 to 1,430
[j,g of formaldehyde/gram of cloth. TWA personal breathing space formaldehyde levels ranged
from trace to 0.46 ppm, while workstation values ranged from 0.32 to 0.70 ppm. The
investigators studied the outcomes by using a mailed questionnaire. The response rate for
current employees was 98%. There were 296 pregnancies among a cohort of 188 women. The
investigators found increased rates of spontaneous abortion, premature birth, and congenital
malformations. The crude rate of spontaneous abortion was 21% among women working at the
plant while pregnant (4 of 19 pregnancies), 15% among women employed elsewhere while
pregnant (11 of 71 pregnancies), and 5% among women at home while pregnant (10 of 206
This document is a draft for review purposes only and does not constitute Agency policy.
4-46 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
pregnancies). The investigators did not explain how workers employed elsewhere or at home
during pregnancy were categorized compared with current workers, nor did they calculate RRs.
As calculated from data presented in Table 5 of the monograph, the crude OR (not corrected for
multiple observations per woman) for those pregnant while currently working at the plant
compared with all others was 3.2 (95% CI: 0.8-12). There were also excess congenital
malformations (13 versus 2%) and premature births (13 versus 4%) among the live births (both
based on two births each in the exposed group) from the women who were pregnant while
employed at the textile plant compared with women who stayed at home. After the NIOSH
investigation, changes were made in the plant to improve ventilation.
Because the report provides insufficient details of the methodology and the fact that there
was no personal exposure classification in this study, it is difficult to validate the findings in this
report. The results did not take into account other potential risk factors for spontaneous abortion
or correct for multiple pregnancies per woman. Furthermore, the marked differences between
the "home" and "work" pregnancies were difficult to interpret.
A cohort study of effects of paternal occupational exposures in Finland by Lindbohm et
al. (1991) found that exposure to formaldehyde had little effect on the rate of spontaneous
abortions among 99,186 pregnancies listed in the national hospital discharge register. The
analysis was limited to births/spontaneous abortions in 1976 and from May 1980 to April 1982.
Spontaneous abortion incidence came from the hospital discharge register and data from
outpatient clinics. There were 808 pregnant wives among potentially formaldehyde-exposed
fathers. Exposure to formaldehyde was based on employment information listed in the Finnish
1980 census. Compared with pregnancies among wives of unexposed spouses, the age and
socioeconomic level-adjusted ORs were 1.1 for low paternal exposure to formaldehyde and 1.0
for moderate to high paternal exposure. Paternal occupational exposures to ethylene oxide,
gasoline/benzene, and rubber industry chemicals were associated with spontaneous abortion.
The authors hypothesized that the mode of action (MOA) for spontaneous abortion following
male exposure to chemicals is genetic damage to germ cells.
The indirect exposure assessment was a substantial limitation of this study. Some
confounders in a study of this type could not be controlled for (smoking history, previous
spontaneous abortions, alcohol use), and census data could not provide completely accurate
information, potentially masking associations between paternal formaldehyde exposure and
spontaneous abortion.
A case-control study by Taskinen et al. (1994) of effects of maternal occupational
exposure to chemicals used in laboratories in Finland indicated that exposure to formalin, which
is a 37% aqueous solution of formaldehyde, was related to an increased risk of spontaneous
This document is a draft for review purposes only and does not constitute Agency policy.
4-47 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
abortion. The investigators studied subjects from payrolls of Finnish state-employed laboratory
workers, the laboratory workers' union, and a register of workers occupationally exposed to
carcinogens. These records were cross-referenced with the hospital discharge register. The
investigators selected women who had a single spontaneous abortion during the period 1973—
1986 and two controls who had delivered a baby without malformations. The final sample size
was 208 cases and 329 controls after refusals and other exclusions. The response rate was
82.4%.
Information on occupational exposure, health status, medication, contraception, and
pregnancy history came from mailed questionnaires. Industrial hygienists' construction of an
exposure index was based on the subjects' descriptions of their work assignments, use of
solvents including estimates of quantity used, and use of a fume hood. ORs were adjusted for
employment, smoking, alcohol consumption, parity, previous miscarriage, birth control failure,
febrile disease during pregnancy, and other organic solvents found in laboratory work.
Spontaneous abortion was associated with 3-5 days per week of formalin exposure (OR 3.5
[95% CI: 1.1-11.2]). A contemporaneous study of formaldehyde concentrations in similar
Finnish workplaces (pathology and/or histology laboratories) reported workroom air to range
from 0.01 to 7 ppm with a mean of 0.45 ppm formaldehyde (Heikkila et al., 1991 [as cited in
Taskinen et al., 1994]) and that the highest exposures occurred during emptying of sample
containers, dish washing, and preparation of formaldehyde solution.
Although the results of this study indicate an increased risk between spontaneous
abortion and exposure to formaldehyde/formalin, the women were also exposed to several
chemicals concurrently, of which toluene (OR 4.7 [95% CI: 1.4-15.9]) and xylene (OR 3.1 [95%
CI: 1.3-7.5]) were also significantly associated with the incidence of spontaneous abortion.
However, the investigators reported that the women were more likely to be co-exposed to
formalin and xylene, which would make confounding by toluene less likely, and, since xylene
was not as strongly associated with the outcome as was formaldehyde, it too is unlikely to fully
explain the reported relationship between formaldehyde and increased risk of spontaneous
abortion. While it is possible that exposure misclassification may have occurred because of the
indirect assessment of workplace chemical exposure, an overall conclusion is that, since the
exposure assessment was conducted by industrial hygienists, it is unlikely that this form of bias
will have impacted the results of the study to any great extent.
In a U.S. study (John et al., 1994), the results of a case-control study of cosmetologists
also supported an association between spontaneous abortion and the use of formaldehyde-based
disinfectants. The study population came from the 1988 North Carolina cosmetology license
registry. Women on this list who were 22-36 years of age were screened to find those who were
This document is a draft for review purposes only and does not constitute Agency policy.
4-48 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
recently pregnant. The cases were full-time cosmetologists who experienced a spontaneous
abortion before gestational week 20 during 1983-1988. The most recent spontaneous abortion
was used as the reference case. Controls were full-time cosmetologists who delivered a live
infant during the same time period.
Information was based on mailed questionnaires. Women were not told the purpose of
the study in order to avoid selection and recall bias. Of 8,356 women who received the
screening questionnaires, 72.5% responded. Of those, 1,696 qualified for the study and 73.6%
completed a more detailed questionnaire. Among them, 96 women were "absolutely sure" they
had a spontaneous abortion and qualified as cases. There were 1,058 live births that qualified as
controls. Exposure assessment included identification of disinfectants used as well as types of
chemicals used on hair, use of gloves, hours worked, number of procedures involving chemicals,
and use of manicure products. Presence of formaldehyde in the cosmetology profession in
general was confirmed in two NIOSH hazard reports (Almaguer and Klein, 1991; Almaguer and
Blade, 1990). ORs were adjusted for age, smoking, pregnancy characteristics, other jobs, hours
worked, education (cosmetology school), hours standing per week, number of chemical
procedures per week, hair dyes per week, bleachings per week, permanents per week, use of
gloves, beauty salon characteristics, and use of alcohol or formaldehyde disinfectants.
An elevated OR of 2.1 (95% CI: 1.0-4.3) was reported with the use of formaldehyde-
based disinfectants adjusted for maternal characteristics and other workplace exposures. Other
chemical exposures were also associated with spontaneous abortion, including number of
chemical services per week, hair dyes, bleaches, and alcohol-based disinfectants. Strengths of
this study include adjustment for important confounding risk factors for spontaneous abortion,
detailed collection of interview-based information, a favorable response rate, and the fact that the
index population had a high likelihood of formaldehyde exposure. These data provide overall
support for an association between formaldehyde exposure and spontaneous abortion.
In a retrospective cohort study by Taskinen et al. (1999) of female woodworkers in
Finland, exposure to formaldehyde was associated with delayed conception and spontaneous
abortion. The subjects, recruited from a woodworkers' union and other businesses involving
wood processing, were linked to a national register of births. Women were included if they were
born between 1946 and 1975, had a live birth at age 20-40 years during 1985-1995, had worked
in the wood processing industry for at least 1 month, and had first employment in the wood
processing industry beginning at least 6 months before the index pregnancy. The first pregnancy
that fulfilled the above criteria was the index pregnancy. There were 1,094 women with these
criteria. Information about personal characteristics, pregnancies, and exposures were collected
from mailed questionnaires for which the response rate was 64%. After other exclusions
This document is a draft for review purposes only and does not constitute Agency policy.
4-49 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
(primarily infertility history, unknown time to pregnancy, and contraceptive failure), the final
sample included 602 women.
Estimates of mean daily exposure to formaldehyde were based on measurements taken at
the women's factories of employment during the early 1990s. Where measurements were
unavailable, measurements from equivalent industries were used. An exposure index
representing a TWA exposure was established for every person in the study based on the
concentration of workplace formaldehyde multiplied by the proportion of the workday exposed
to formaldehyde. The investigators categorized TWA formaldehyde exposure into categories of
low (mean of 18 ppb), medium (mean of 76 ppb), and high (mean of 219 ppb) exposure.
Time-to-pregnancy data were analyzed by a discrete proportional hazards regression
procedure with, as the outcome, a fecundability density ratio (FDR), in which a ratio of average
incidence densities of pregnancies for exposed women was compared with that of the employed,
unexposed women. As explained by Taskinen et al. (1999), an FDR significantly below unity
suggests that conception was delayed. The age-, employment-, smoking-, alcohol consumption-,
parity-, and menstrual irregularity-adjusted FDR was 0.64 (95% CI: 0.43-0.92) for women
exposed to high formaldehyde levels compared with the unexposed controls, indicating that there
was a substantial delay in time to conception in this group of women. Among a subset of women
with high exposure who did not use gloves, the FDR was even lower (0.51 [95% CI: 0.28-0.92]),
suggesting that these results might be explained in part through dermal contact with
formaldehyde or might indicate an individual's failure to follow appropriate precautions, which
might have increased inhalation exposures in other ways. Exposure to solvents, wood dust and
other dusts, and phenols was not associated with decreased fecundability.
The investigators further conducted an analysis of the risk of spontaneous abortion after
carefully including only women who had the same workplace during the year of the spontaneous
abortion as they had during the beginning of the time-to-pregnancy period. Spontaneous
abortion was associated with formaldehyde exposure in the low exposure group (OR = 2.4 [95%
CI: 1.2-4.8]), in the medium exposure group (OR =1.8 [95% CI: 0.8-4.0]), and in the high
exposure group (OR = 3.2 [95% CI: 1.2-8.3]). Endometriosis was also associated with the
highest formaldehyde level (OR = 4.5 [95% CI: 1.0-20.0]).
This study by Taskinen et al. (1999) was a well-designed population-based case-control
study that appears to have been well executed and appropriately analyzed. The study population
of Finnish women was well defined and adequately selected so as to allow for meaningful
comparisons of health effects between individuals with different levels of exposure to
formaldehyde. The participation rate was 64%, which is low enough to raise a concern about the
potential for selection bias. However, the authors noted that selection bias has not influenced the
This document is a draft for review purposes only and does not constitute Agency policy.
4-50 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
results of other reproductive epidemiology studies reporting results on smoking, irregular
menstruation, and earlier miscarriages, which are known to lengthen the time to pregnancy
(Bolumar et al., 1996; Sallmen et al., 1995; Baird and Wilcox, 1985). Furthermore, there is no
evidence to support conjecture that an individual's decision to participate in this study would be
differential with respect to their workplace formaldehyde exposures while being non-differential
with respect to the other exposures of interest, including organic solvents, wood dust, and
phenols. Since the women who chose to participate in this study were not likely to be aware of
the specific hypotheses under investigation, nor could they have known the formaldehyde
exposures that were independently estimated by an industrial hygienist, selection bias is not a
likely explanation for the findings of adversity.
Data on pregnancy history, including spontaneous abortions, were collected by
questionnaire. Spontaneous abortion is the most common adverse outcome of pregnancy (Klein
et al., 1989), and retrospective self-report of spontaneous abortion has been found to match well
with prospectively collected reproductive histories (Wilcox and Horney, 1984). Many
spontaneous abortions, however, are missed with self-reporting with the magnitude likely
exceeding 25%, but only rarely do women self-report false positive events (Wilcox and Horney,
1984). The effect of such an undercount is to cause a bias towards the null when the likelihood
of undercounting is unrelated to formaldehyde exposure. The implication is that the observed
association of increased risk of spontaneous abortion associated with occupational exposure to
formaldehyde may be an underestimation of the true risk.
Two studies (Hemminki et al., 1985, 1982) specifically assessed the effects of
formaldehyde exposure and reported no significant increase in the risk of spontaneous abortion.
Hemminki and colleagues (1982) conducted a retrospective cohort study of nurses who were
potentially exposed to chemical sterilizing agents, including formaldehyde, ethylene oxide, and
glutaraldehyde. The risk of having a spontaneous abortion among the women on the sterilizing
staff was compared with that among the control population of nursing auxiliaries whom the
supervisory nurses thought to be unexposed to the chemical sterilizing agents during the previous
three decades. However, no measurements of the chemical sterilizing agents were taken.
Information about exposure to chemical sterilizing agents was obtained from the supervising
nurses. When the women were conducting sterilizing procedures during their pregnancies, the
frequency of spontaneous abortion was 15.1% compared with 4.6% for the nonexposed
pregnancies among the sterilizing staff. The increased frequency of spontaneous abortion
correlated with exposure to ethylene oxide but not with exposure to glutaraldehyde or
formaldehyde. The investigators reported that ethylene oxide concentrations have been
measured in many sterilizing units in Finnish hospitals; 8-hour weighted mean concentrations
This document is a draft for review purposes only and does not constitute Agency policy.
4-51 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
have ranged from 0.1 to 0.5 ppm with peak concentrations up to 250 ppm (measurements by the
Finnish Institute of Occupational Health) (Hemminki et al, 1982). No measurements of
glutaraldehyde concentrations were available. Hemminki et al. (1982) reported that exposure to
formaldehyde in the sterilization units may be minimal, particularly when gas chambers are used.
The range of formaldehyde concentrations measured in sterilizing units has been reported as
0.03-3.5 ppm.
It is not clear that the unexposed women who served as controls were an appropriate
comparison group to the sterilizing staff. The investigators reported that, among the sterilizing
staff, those women who were unexposed during pregnancy experienced a rate of spontaneous
abortion of 4.6% but that, among the comparison population of nursing auxiliaries who were
presumed to be unexposed, the rate of spontaneous abortion was 10.5%. Had the nursing
auxiliaries been an appropriate comparison group, it would be expected that their rate of
spontaneous abortion would be similar to the unexposed sterilizing staff. Given this anomaly in
study design and the unknown concentrations of formaldehyde exposure that were assessed as
positive or negative by supervisory nurses regarding occupational exposures in the previous
30 years, it is concluded that this report of no association between formaldehyde exposure and
the risk of spontaneous abortion does not temper the conclusion that formaldehyde exposure has
been shown to increase the risk of spontaneous abortion.
A second study by the same lead author (Hemminki et al., 1985) used a different study
design to reassess the hypothesis that chemical exposures common in the field of nursing could
be risk factors for spontaneous abortion. This case-control study found no increase in the risk of
spontaneous abortion associated with exposure to formaldehyde. The head nurses at each
hospital were asked by the investigators whether each case or control had been exposed to
formaldehyde during a given 3-month period corresponding to the first trimester of a study
participant's pregnancy during 1973-1979. Formaldehyde exposure was assessed as positive or
negative for either use as a sterilizing agent or use of sterilized instruments. The reported crude
OR for formaldehyde exposure was 0.6; no CIs were provided. From the data reported in
Table 2 in Hemminki et al. (1985), the unadjusted OR and its CI can be computed post hoc as
OR (0.70 [95% CI: 0.28-1.73]). The authors acknowledged that the study failed to distinguish
between sterilizing work and the use of sterilized instruments, where only very small exposures
could be expected. Given the likelihood of extreme exposure misclassification and the
presentation of only crude results without control of potential confounding for formaldehyde,
these results do not appear to be exculpatory of a true causal association between formaldehyde
exposure and the risk of spontaneous abortion.
This document is a draft for review purposes only and does not constitute Agency policy.
4-52 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
A meta-analysis of formaldehyde exposure and spontaneous abortion was conducted by
Collins et al. (2001b). However, the published results should be interpreted with caution. This
meta-analysis included one very large null study of paternal formaldehyde exposure along with
seven studies of maternal exposure. The two null studies by Hemminki et al. (1985, 1982) were
also included without consideration of the potentially extreme exposure misclassification that
may have attenuated any true adverse effect. Nevertheless, the overall reported meta-analytic
RR for parental formaldehyde exposure based on eight maternal and paternal exposure studies
was 1.6 (95% CI: 0.9-2.7). For case-control studies the RR was 1.8 (95% CI: 0.7-4.8), and for
cohort studies the RR was 1.7 (95% CI: 1.2-2.3). Collins et al. (2001b) argued that the method
of exposure evaluation may have influenced the observed results; they stated that several of the
studies whose exposures were based on the investigator's judgment were likely misclassified,
which may have obscured the true relationship, while other studies that assessed exposure based
on self-reporting could have suffered from recall bias. They report that RRs were higher for
studies based on self-reported exposures (RR =1.9 [95% CI: 1.3-2.6]) than those based on
objective exposure assessments (RR =1.5 [95% CI: 0.6-3.7]) and suggested that this difference
might reflect recall bias in the exposure assessment. However, for recall bias to have been
operable in these studies, the women who provided self-reported data on pregnancy history and
occupational exposure would have had to appreciate that the hypothesis of interest was the
specific effect of formaldehyde on the risk of spontaneous abortion. In the specific case of the
study by Taskinen and colleagues (1999), the investigator also looked at the effects of other
exposures, such as organic solvents, dust, and phenols, and did not report adverse effects. It is
therefore unlikely that the women providing exposure data were doing so in a manner indicative
of recall bias. If the supposition of non-differential misclassification error in exposure is indeed
correct, the observed results of the meta-analysis would likely have been biased towards the null.
Therefore, the true RR for maternal formaldehyde could be higher than Collins et al. (2001b)
reported and would likely be statistically significant. Had the study of paternal exposure been set
aside, the meta-analysis almost surely would have shown a statistically significant increase in the
risk of spontaneous abortion associated with maternal formaldehyde exposure. This single study
reported a null finding based on exposure assessment from census records of employment, and,
as the largest of the studies in the meta-analysis, it contributed the greatest weight.
Lastly, Collins and coworkers (2001b) suggested that there were potential confounding
factors in each of the workplaces that might have produced the observed findings of increased
risk of spontaneous abortion associated with formaldehyde. While each of these occupational
studies focused on women who were co-exposed to formaldehyde and other chemicals, the
occupational groups were quite different and had different sets of co-exposures. The
This document is a draft for review purposes only and does not constitute Agency policy.
4-53 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
woodworkers in the Taskinen et al. (1999) study were potentially co-exposed to organic solvents
related to painting and lacquering, dusts, and phenols, none of which was shown to be an
independent predictor of adverse risk. The cosmetologists studied by John et al. (1994) were
co-exposed to hair dyes, bleach, alcohol-based disinfectants, and chemicals specific to services,
such as fingernail sculpturing, but, in analyses that were specifically adjusted for other work
exposures and their potentially confounding effects, the investigators reported an OR of 2.1
(95% CI: 1.0-4.3) for the use of formaldehyde-based disinfectants. The laboratory workers
studied by Axelsson et al. (1984) were potentially co-exposed to a wide range of solvents, but the
miscarriage rate was highest among those exposed to formaldehyde, and, for a potential
confounder to entirely explain an observed effect of another exposure, it must be more strongly
associated with the adverse outcome.
It does not appear that the collective results of formaldehyde exposures associated with
increased risk of spontaneous abortion—often in spite of exposures being crudely measured—
can be explained by information bias or confounding.
The findings by Taskinen et al. (1999) of reduced fertility and increased risk of
spontaneous abortion are internally consistent and coherent with other reports of increased risk
of pregnancy loss associated with exposure to formaldehyde (John et al., 1994; Taskinen et al.,
1994; Seitz and Baron, 1990; Axelsson et al., 1984). Absent evidence of alternative explanation
for these findings, it is concluded that exposure to formaldehyde is associated with pregnancy
loss and diminished fertility.
4.1.1.7.2. Congenital malformations. Only three studies have reported on the epidemiologic
evidence of an association between formaldehyde exposure and the risks of births having
congenital malformations. In the earliest study by Hemminki et al. (1985), the investigators
presented an analysis of 34 congenital malformations from the Finnish Register of Congenital
Malformations and compared them with a group of 95 controls from those used in the larger
study. An association was found between formaldehyde exposure and malformations based on
three exposed cases (OR = 1.8).
The case-control study by Taskinen et al. (1994) of effects of occupational exposure to
chemicals used in laboratories in Finland examined the potential effects of exposure to formalin
on both spontaneous abortions and congenital malformation. The investigators reported on a
study of 36 laboratory workers with a child registered in the Finnish Register of Congenital
Malformations and 105 controls. There was no association between formalin and congenital
malformations.
This document is a draft for review purposes only and does not constitute Agency policy.
4-54 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
A Lithuanian study (Dulskiene and Grazuleviciene, 2005) for which only a brief
summary is available in English investigated the risk of congenital heart malformations as a
result of exposure to 43 different agents. The number of births included in the study was not
given in the English abstract. Exposure to residential ambient formaldehyde concentrations of
>2.42 |ig/m3 (0.002 ppm) was associated with a 24% increase in the risk of congenital heart
malformations (OR = 1.24 [95% CI: 0.81-2.07]). The details of this study are unavailable in
English translation, making it impossible to critically analyze details, such as co-exposure and
other possible confounders.
4.1.1.7.3. Low birth weight. A case-control study by Grazuleviciene et al. (1998) examined the
association of low birth weight (<2,500 grams) and air pollutants, including formaldehyde,
particulates, sulfur dioxide, lead, ozone, and nitrogen dioxide, measured in 12 areas in the city of
Kaunas, Lithuania. This city has conducted environmental pollutant measurements since 1993,
and the investigators classified formaldehyde exposure based on the area of residence of the
study subjects. Formaldehyde levels in the 12 districts of Kaunas in 1994 ranged from 1.36 to
5.28 |ig/m3 (0.0011-0.0043 ppm), with a citywide average of 3.14 |ig/m3 (0.0026 ppm).
Information on infants came from a birth registry. There were 244 cases of low birth weight and
4,089 normal controls born in 1994. Personal data came from record-based prenatal interviews,
and pregnancy data came from hospital records.
The crude RR of low birth weight among women exposed to the highest airborne
formaldehyde level was 1.68 (95% CI: 1.24-2.27). After adjustment for age, occupation,
hazardous work, education, marital status, smoking, hypertension, and other air pollutants, the
OR was still elevated but no longer statistically significant (OR 1.37 [95% CI: 0.90-2.09]).
Although formaldehyde exposure was the only single air pollutant associated with low birth
weight, factors such as smoking, marital status, and pregnancy-related factors had more of an
impact on birth weight. Total suspended particulates (OR 2.58 [95% CI: 1.34-4.99]) and
hazardous work (OR 2.62 [95% CI: 1.12-6.10]), which was not defined by the authors, were also
related to low birth weight.
Aside from studies of birth weight deficits from tobacco smoke and occupational
exposure, the literature on exposure to ambient air pollutants to support the investigators'
hypothesis is limited. The strength of the association between total suspended particulates and
low birth weight supports the idea that incidence of birth weight <2,500 grams may be related to
atmospheric pollution, although this finding may not be specific to formaldehyde. Because of
the large number of variables evaluated in the analysis, large fluctuations in the atmospheric
This document is a draft for review purposes only and does not constitute Agency policy.
4-55 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
formaldehyde measurements, co-exposure to other pollutants, and geographic variability of low
birth weight, it is difficult to estimate the impact of formaldehyde alone on low birth weight.
4.1.1.7.4. Summary. Although all studies on potential developmental toxicity of formaldehyde
have limitations and do not uniformly report positive results, the associations between
spontaneous abortion, delayed conception, or reproductive outcomes and formaldehyde exposure
in multiple studies cannot be dismissed, because several studies report concordant findings
across several populations and study methodologies. The results of most of the studies with
positive findings were adjusted for many potentially confounding factors that may be related to
spontaneous abortion and infertility, including smoking and alcohol use, pregnancy and
reproductive history, and other chemical exposures.
The association between fertility and formaldehyde (Taskinen et al., 1999) stands out
because of its strong quantitative statistical analysis, adequate sample size, and rigorous exposure
assessment. This study was designed to specifically assess the effect of formaldehyde on
reproductive outcomes. Furthermore, it was the only study with an exposure assessment based
on quantitative measurements from the subject's workplace. Moreover, the investigators
conducted a multivariable survival analysis that approximates a longitudinal life table or person-
year analysis while simultaneously adjusting for important confounders. The findings were
strengthened by statistically significant associations between formaldehyde and spontaneous
abortion and endometriosis. The fact that the use of gloves may reduce the reproductive effect of
formaldehyde supports the dose-response relationship in this study, and the lack of an association
between time to pregnancy and any other workplace exposures strengthens the specificity of
formaldehyde effects. The results also support associations reported between formaldehyde and
increased risk of spontaneous abortion because subfertility and spontaneous abortion are
biologically linked (subclinical pregnancy losses are increased among women with fertility
problems) (Gray and Wu, 2000; Hakim et al., 1995), and both subfertility and spontaneous
abortion may be related to sensitivity to environmental agents (Correa et al., 1996).
4.1.1.8, Oral Exposure Effects on the Gastrointestinal Tract
No human epidemiology studies exist to determine an association between oral exposure
of formaldehyde and adverse health effects in the gastrointestinal (GI) tract.
4.1.1.9. Summary: Noncarcinogenic Hazard in Humans
Formaldehyde has clearly and consistently been shown to be a potent sensory irritant
with a variety of adverse health effects. Eye, nose, and throat irritation as a result of
This document is a draft for review purposes only and does not constitute Agency policy.
4-56 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
formaldehyde exposure has been documented in a wide range of epidemiologic studies. Workers
chronically exposed to formaldehyde have exhibited signs of reduced lung function consistent
with BC, inflammation, or chronic obstructive lung disease. A well-conducted residential
epidemiology study has convincingly shown a concentration response for decreased pulmonary
function among children with increased formaldehyde exposures. Several cross-sectional studies
have described associations between increased concentrations of formaldehyde and increased
prevalence of asthma. However, two case-control studies that focused on risk factors for the
initial physician diagnosis of asthma, which is indicative of atopic switching, have also shown
concentration-dependent adverse effects associated with formaldehyde exposure.
Results of research on the effects of formaldehyde on tissue histology suggest that
formaldehyde is also responsible for reduced mucociliary clearance and the induction of
histopathologic lesions in the nose. In addition, there is evidence of neurological impairment in
several studies of formaldehyde-exposed histology technicians, but confounding exposures to
other neurotoxic solvents and inconsistent results prevent drawing definitive conclusions
concerning the neurotoxicity of formaldehyde from these studies.
Finally, there is epidemiologic evidence that formaldehyde is associated with adverse
reproductive outcomes. Four of six occupational studies found an increased risk of spontaneous
abortion among formaldehyde-exposed women. Results of other studies suggested associations
among formaldehyde and congenital malformations, low birth weight, and endometriosis. The
strongest evidence of an association between formaldehyde and an adverse reproductive outcome
came from a well-conducted study of infertility in women employed in the wood processing
industry. This study found a greater than threefold increased risk of spontaneous abortion, a
nearly 50% decrease in a measure of delayed conception indicating reduced fertility, and
increased time to pregnancy associated with average daily formaldehyde exposures of 0.15-
1 ppm.
4.1.2. Cancer Health Effects
4.1.2.1. Respiratory Tract Cancer
4.1.2.1.1. NPC. NPC is a very rare form of cancer. The incidence is less than 1 per 100,000
persons throughout most parts of the world. The most common form of NPC arises from the
epithelial cells lining the nasopharynx. This presentation constitutes between 75 and 100% of all
NPCs. There are two types, squamous cell carcinoma (SCC) and nonkeratinizing carcinoma. In
the U.S., the 5-year survival rate for NPC is about 25% (Burt et al., 1992). Certain risk factors
have been implicated in its etiology, including Epstein-Barr virus (EBV), wood dust and
This document is a draft for review purposes only and does not constitute Agency policy.
4-57 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
particles applied to wood in its treatment, exhaust fumes, occupational smoke, and nitrosamines.
The epidemiologic studies of NPC are summarized in Table 4-1.
4.1.2.1.1.1. Cohort studies. The International Agency for Research on Cancer (IARC) reported
on eleven cohort studies of formaldehyde-exposed industry workers (Marsh et al., 2002, 1996,
1994; Andjelkovich et al., 1995; Gardner et al., 1993; Bertazzi et al., 1989, 1986; Stayner et al.,
1988; Blair et al., 1987, 1986; Edling et al., 1987) and results from eight cohort studies of
professional workers (Hall et al., 1991; Hayes et al., 1990; Stroup et al., 1986; Harrington and
Oakes, 1984; Levine et al., 1984; Walrath and Fraumeni, 1984, 1983; Friedman and Ury, 1983)
(IARC, 2006).
Several of these studies measured exposure to formaldehyde at 10 production facilities
that contributed to a cohort that has been studied by Blair et al. (1987, 1986) and Hauptmann et
al. (2004). The National Cancer Institute (NCI) conducted a mortality study of solid tumors
among a cohort of 25,619 workers who were employed in these 10 plants that produced or used
formaldehyde in the U.S. before 1966 (Blair et al., 1987, 1986). Subjects were followed to
January 1, 1980, accruing approximately 600,000 person-years of follow-up. Hauptmann et al.
(2004) updated the cohort to December 31, 1994 and reported a significant excess risk of NPC in
exposed workers based on U.S. population death rates (standardized mortality ratio [SMR] = 2.1
[95% CI: 1.05-4.21]). In addition to the SMR based on an external comparison population, RRs
were presented based on internal comparisons of similar workers in order to minimize potential
selection bias due to the well-known healthy worker effect (HWE). For NPC, RRs increased
with several different exposure metrics, including average exposure intensity, cumulative
exposure, highest peak exposure, and duration of exposure to formaldehyde (p values for tests
for trends were 0.066, 0.025, <0.001, and 0.147, respectively). These results were based on
primary data analyses of the health and exposure data collected by the NCI, according to their
research protocol and analyzed accordingly. As such, the reported statistical p values may be
appropriately interpreted as showing that these workers were at increased risk of NPC associated
with exposure to formaldehyde. These NCI investigations controlled for potential selection bias
due to the HWE and for several potential confounders, including calendar year, age, sex, race,
and pay category. There was no evidence of any differential measurement error that could have
produced the observation of a spurious association. Any non-differential measurement error
would likely have led to an observed effect of formaldehyde that was less than that which would
otherwise have been observed in the absence of measurement error.
This document is a draft for review purposes only and does not constitute Agency policy.
4-58 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
§•
TO
S
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
Hauptmann et al. (2004)
Retrospective cohort
mortality study of 25,619
workers employed at 10
formaldehyde plants in the
U.S. followed from either
plant start-up or first
employment through 1994.
The 10 plants produced
formaldehyde (3 plants),
molding compounds
(3 plants), photographic
film (2 plants), plywood
(1 plant), and formaldehyde
resins (6 plants).
Exposure estimates3 based
on job titles, tasks, visits
to plants by study
industrial hygienists, and
monitoring data
measurements. Peak
exposure = short-term
excursions >8-hour TWA
formaldehyde intensity
and knowledge of job
tasks. Workers
contributed pre-exposure
person time to nonexposed
category. RRs were from
Poisson regression
models, using a 15-year
lag to account for tumor
latency.
Overall
Nonexposed SMR
Exposed SMR
Peak exposure (ppm)
0 RRD
>0 to <2.0
2.0 to <4.0
4.0 or greater
1.56
2.10
1.00
N/A
N/A
1.83
Average intensity of exposure (ppm)
0 RR
<0.5
0.5 to <1.0
1.0 or greater
Cumulative exposure (ppm-years)
0
>0 to <1.5
1.5 to <5.5
5.5 or more
Duration (years)
0
>0 to <5
5 to <15
15 or more
RR
RR
1.00
N/A
0.38
1.67
2.40
1.00
1.19
4.14
1.77
1.00
0.83
4.18
(95% CI: 0.39-23)
(95% CI: 1.05-21)
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
Trendp < 0.001
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
Trend p = 0.066
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
Trend p = 0.025
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
(95% CI: NS)
Trend p = 0.147
(2)
(8)
(2)
(0)
(0)
(7)
(2)
(0)
(1)
(6)
(2)
(3)
(1)
(3)
(2)
(4)
(1)
(2)
Marsh et al. (2002)
vo
Retrospective cohort
mortality study of 7,328
workers hired up to 1984
Worker-specific exposure3
from job exposure matrix
based on available
Cohort study
Overall
I U.S.
SMR
4.94
(95% CI: 1.99-10)
(7)
-------
K
s
TO
>3
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
S"4
>3*
§•
to
s
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
a, Co'
TO Sj-
§ ^
>S
>S
TO
TO'
*
and followed until 1998 in
one plant from Hauptmann
et al. (2004). Mortality was
compared with death rates
in two Connecticut counties
and the U.S.
sporadic sampling data
from 1965-1987, job
descriptions, and verbal
job descriptions by plant
personnel and industrial
hygienists. Exposures
ranked on a 7-point scale
with exposure range
assigned to each rank.
17% of jobs validated with
company monitoring data;
remaining 83% based on
professional judgment.
Assumed pre-1965
exposure levels same as
post-1965 levels.
| County | SMR
Short-term worker (<1 year)
| | SMR
Long-term worker (1 or more years)
| SMR
Year of hire
1941-1946 SMR
1947-1956 SMR
1957 or later SMR
5.00
5.35
4.59
8.13
2.63
Cumulative exposure (ppm-years) county
Unexposed
SMR
0 to <0.004
SMR
3.97
0.004-0.219
SMR
5.89
0.22+
SMR
7.51
Average exposure (ppm) county
Unexposed
SMR
0 to <0.03
SMR
2.41
0.03-0.159
SMR
15.30
0.16+
SMR
4.13
(95%
(95%
(95%
(95%
(95%
(95%
(95%
(95%
(95%
(95%
(95%
CI: 2.01-10)
CI: 1.46-14)
CI: 0.95-13)
CI: 2.98-18)
CI: 0.07-15)
CI: 0.10-22)
CI: 1.22-17)
CI: 1.55-22)
CI: 0.06-13)
CI: 4.16-39)
CI: 0.50-15)
Duration of exposure to >0.2 ppm (years)
Unexposed
SMR
3.01
(95% CI: 0.36-11)
(2)
0 to <1
SMR
4.81
(95% CI: 0.58-17.4)
(2)
1-9
SMR
4.04
(95% CI: 0.10-22.51)
(1)
10+
SMR
27.60
(95% CI: 3.34-100)
(2)
Duration of exposure to >0.7 ppm (years)
Unexposed
<1
1+
SMR
SMR
SMR
3.64
9.51
11.07
(95%
(95%
(95%
CI: 0.99-9.31)
CI: 1.15-34.4)
CI: 0.28-61.67)
(7)
(4)
(3)
(0)
(6)
(0)
(1)
(3)
(3)
(0)
(1)
(4)
(2)
(4)
(2)
(1)
On
O
-------
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
S"4
>3*
s
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
s ?
a, Co
8- a
to Sj-
§ ^
s ^
§ 3
*o»l.
Marsh et al. (2002)
A nested case-control
analysis of all pharyngeal
cancer cases also conducted
with four controls randomly
selected from cohort and
matched on age, year of
birth, race, and sex.
Conditional logistic model
used for nested case-control
analysis.
Nested case-control analysis0
Duration of exposure to >0.2 ppm (years)
Unexposed
OR
1.00
(8)
0 to <1
OR
1.13
(95% CI: 0.24-5.29)
(6)
1-9
OR
1.38
(95% CI: 0.18-9.03)
(3)
10+
OR
9.49
(95% CI: 0.55-701)
(5)
Duration of exposure to >0.7 ppm (years)
Unexposed
<1
1+
OR
OR
OR
1.00
0.52
1.11
(95% CI: 0.08-2.45)
(95% CI: 0.06-11.3)
(16)
(4)
(2)
Hayes et al. (1990)
Proportionate mortality
cohort study of 4,046 U.S.
male embalmers and funeral
directors who died between
1975 and 1985.
Exposure presumed.
Overall
PMR
2.16
(4)
On
Hansen and Olsen (1995)
Proportionate incidence
study of 2,041 men with
cancer who died between
1970 and 1984, identified
from the Danish Cancer
Registry and matched with
the Danish Supplementary
Pension Fund, whose
longest work experience
occurred at least 10 years
before the cancer diagnosis.
The SPIR measured the
proportion of cases of NPC
in formaldehyde-associated
companies relative to the
proportion of cases of NPC
among all employees in
Denmark.
Linked companies through
tax records to the national
Danish Product Register.
Overall
SPIR
1.3
(95% CI: 0.03-3.2)
(4)
-------
K
s
TO
>3
&*
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
§•
TO
S
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
Olsen et al. (1984)
Case-control study of 314
cases of NPC from Danish
Cancer Registry linked to
the Registry during 1970-
1982. Three controls/case
sampled with cancer of the
colon, rectum, breast, and
prostate by age, sex, and
year of diagnosis of cases.
Employment histories
after 1964 from files
maintained by Danish
Cancer Registry evaluated
by industrial hygienists.
Men
Women
OR
OR
0.7
2.6
(95% CI: 0.3-1.7)
(95% CI: 0.3-22)
West et al. (1993)
Case-control study of 104
non-Chinese incident NPC
cases from the Philippine
General Hospital matched
with 104 hospital and 101
community controls.
Personal interview,
including job history.
Industrial hygienists
blinded to case-control
status reviewed and rated
jobs as likely or unlikely
to be exposed. Analysis
by length of exposure,
length of exposure lagged
10 years, time since first
exposure, and age at first
exposure, based on date of
interview or death.
Length of exposure (years)
<15 RRd 2.7
15 or more 1.2
Length of exposure lagged 10 years (years)
<15
15 or more
Years since first exposure
'<25
25 or more
RR
RR
Age at first exposure (years)
<25
25 or older
RRQ
1.6
2.1
1.3
2.9
2.7
1.2
(95% CI: 1.1-6.6)
(95% CI: 0.5-3.2)
(95% CI: 0.7-3.8)
(95% CI: 0.7-6.2)
(95% CI: 0.6-3.2)
(95% CI: 1.1-7.6)
(95% CI: 1.1-6.6)
(95% CI: 0.5-3.3)
On
to
-------
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
s
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 ^ §
6, to'
to Sj-
§ ^
s ^
§ 3
?»»i.
Roushetal. (1987)
Population-based case-
control study of 173 male
cases from the Connecticut
Tumor Registry who died of
any cause from 1935-1975.
605 male controls randomly
selected from state death
certificates during same
time period.
Four categories: I
probably exposed most of
working life; ]_L probably
exposed most of working
life and probably exposed
20+ years before death;
III, probably exposed most
of working life and
probably to high level in
some year; IV^ probably
exposed most of working
life and probably exposed
to high level 20+ years
before death.
Exposure levels
I
II
III
IV
ORe
1.0
1.3
1.4
2.3
(95% CI: 0.6-1.7)
(95% CI: 0.7-2.4)
(95% CI: 0.6-3.1)
(95% CI: 0.9-6.0)
Vaughan et al. (1986a)
Population-based case-
control study of 27
incidence cases of NPC
(during 1980-1983) from a
13-county area (Washington
State Cancer Surveillance
System) and 552 matched
controls from random digit
dialing in same area, for
occupational exposures.
Interview-based
information on lifetime
occupational exposure to
formaldehyde with cases,
next of kin, and controls.
Exposure from available
hygiene data, NIOSH and
other data, and NCI job
exposure linkage system.
Exposure levels based on
investigator's judgment.
Exposure score A:
weighted sum of no. years
spent per job (weight =
estimated formaldehyde
level). B: weighted sum
of no. years spent per job
with 15-year lag (latency).
Intensity
Low
Medium/high
No. years exposed
1-9
10 or more
Exposure score A: no lag
5-19
20 or more
OR'
OR'
OR'
Exposure score B: 15-year lag
ORf
5-19
20 or more
1.2
1.4
1.2
1.6
0.9
2.1
1.7
2.1
(95% CI: 0.5-3.3)
(95% CI: 0.4-4.7)
(95% CI: 0.5-3.1)
(95% CI: 0.4-5.8)
(95% CI: 0.2-3.2)
(95% CI: 0.6-7.8)
(95% CI: 0.5-5.7)
(95% CI: 0.4-10)
On
LtJ
-------
K
s
TO
>3
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
§•
TO
S
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
On
Vaughan et al. (1986b)
Vaughan et al. (2000)
Population-based case-
control study of 27
incidence cases on NPC
(during 1980-1983) from a
13-county area (Washington
State Cancer Surveillance
System) and 552 matched
controls from random digit
dialing in same area, for
residential exposures.
Population-based case-
control study of 196
incident epithelial NPC
patients identified from 5
U.S. cancer registries from
1987-1993 matched with
244 controls from random
digit dialing in the same
geographic regions.
No direct measurements.
Interview information
from cases/controls or
next of kin: residence in
past 50 years, use of
particleboard or plywood,
and lifetime occupational
and chemical exposure
history.
Interviewed for lifetime
occupational and chemical
exposure. Exposure
estimates by industrial
hygienist without
knowledge of case-
controls status.
Probability of exposure:
definitely not or unlikely
(<10%); possible (>10%
and <50%); probable
(>50% and <90%); and
definite >90%). Jobs with
potential exposure
assigned estimated
concentration levels based
on TWA: low (<10 ppm),
moderate (>10 and
<50 ppm), and high
(>50 ppm).
Years of residence in mobile home
1-9 OR8
10 or more
Years of exposure to particleboard
1-9
10 or more
Exposure source
Occupation only
Mobile home only
Both
OR8
OR8
2.1
5.5
1.4
0.6
1.7
2.8
6.7
(95% CI: 0.7-6.6)
(95% CI: 1.6-19)
(95% CI: 0.5-3.4)
(95% CI: 0.2-2.3)
(95% CI: 0.5-5.7)
(95% CI: 1.0-7.9)
(95% CI: 1.2-39)
Possible, probable, or definite exposure (61 cases, 76 controls)
Ever | OR
Duration (years)
1-5 OR11
6-17
18 or more
Cumulative exposure (ppm-years)
0.05-0.40
0.41-1.10
>1.10
OR11
(95% CI: 0.4-2.0)
(95% CI: 0.8—4.1)
(95% CI: 1.3-6.6)
Trend p = 0.033
Probable or definite exposure (27 cases, 30 controls)
1.6
0.9
1.9
2.7
0.9
1.8
3.0
(95% CI: 1.0-2.8)
(95% CI: 0.4-2.1)
(95% CI: 0.9-1.4)
(95% CI: 1.2-6.0)
Trend p = 0.014
Ever
Duration (years)
1-5
6-17
18 or more
OR11
OR11
2.1
2.0
3.3
1.6
(95% CI: 1.1-1.2)
(95% CI: 0.8-5.0)
(95% CI: 0.9-12)
(95% CI: 0.5-5.6)
Trend p = 0.069
-------
K
s
TO
>3
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
§•
TO
S
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
Cumulative exposure (ppm-years)
0.05-0.40
0.41-1.10
>1.10
OR11
Definite exposure (10 cases, 2 controls)
Ever
OR11
13.3
(95% CI:
(95% CI:
(95% CI:
Trend p =
0.7^1.9)
0.7-9.5)
0.7-7.0)
= 0.13
(95% CI: 2.5-70)
>s
>s
TO
JS*
*
Hildesheim et al. (2001)
Population-based case-
control study of 375
incident cases from two
Taiwanese hospitals
between 7/15/91 and
12/31/94. 325 controls
came from a random sample
of households from a
national household
registration system and
were age, sex, and area-of-
residence matched. Tumors
were histologically
confirmed. All subjects
were tested for the EBV.
Exposure metrics were
stratified by seropositivity.
In-person interviews
collected information on
risk factors and job history
for jobs held >1 year,
including length of time
job held, type of industry,
and tasks, tools, and
materials used on the job.
Industrial hygienist
assigned Standard
Industry Classification/
Standard Occupational
Classification codes to
jobs, assigning each a
probability and intensity
of exposure on a 0-9
scale. Exposure metrics
were duration, average
intensity (intensity scale),
average probability
(probability scale),
cumulative (average
intensity), years since 1st
exposure, and age at 1st
exposure. Analysis of
Ever exposed
Duration (years)
<10
>10
EBV posJ
Ever exposed
<10
>10
RR1
RR1
RR1
RR1
RR1
1.4
(95% CI: 0.9-2.2)
.7
.6
(95% CI:
(95% CI:
Trend p =
(95% CI:
(95% CI:
(95% CI:
Cumulative exposure (average intensity-years)
<25
>25
EBV posJ
<25
>25
Years since 1st exposure
<20
>20
RR1
RR1
RR1
RR1
RR1
RR1
(95% CI:
(95% CI:
Trend p =
(95% CI:
(95% CI:
(95% CI:
(95% CI:
0.69-2.3)
0.91-2.9)
= 0.08
1.2-6.2)
0.8-9.7)
0.9-7.7)
0.7-2.4)
0.9-2.7)
= 0.10
0.9-17)
0.8-5.8)
1.0-5.8)
0.8-2.0)
On
-------
K
s
TO
>3
Table 4-1. Cohort and case-control studies of formaldehyde cancer and NPC
§•
TO
S
Reference
Study design
Exposure assessment
Results; statistical significance (number observed deaths)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
nonkeratinizing or
undifferentiated tumors
yielded similar results as
overall analysis.
EBV pos
<25
>25
Age at 1st exposure
<20
>20
EBV posJ
<25
>25
RR1
RR1
RR1
RR1
RR1
RR1
2.3
2.8
1.3
3.4
2.6
3.1
(95% CI: 0.5-10)
(95% CI: 0.8-2.0)
(95% CI: 0.9-12)
(95% CI: 1.1-6.5)
(95% CI: 0.4-24)
aExposure estimates by Hauptmann et al. (2004) were 10 times higher than those of Marsh et al. (2002).
bAdjusted for calendar year, age, sex, race, and pay category (salaried versus wage).
°Results for cumulative and average intensity of exposure are not included here because condition logistic regression produces unstable estimates for this small
number of cases.
dAdjusted for years since first exposure to dust and exhaust fumes.
eAdjusted for age at death, year at death, and availability of occupational information (Roush et al., 1987).
fAdjusted for cigarette smoking, alcohol consumption, gender, and age.
8Adjusted for ethnic origin and cigarette smoking.
hAdjusted for age, sex, race, SEER site, cigarette usage, proxy status, and education.
'Adjusted for age, sex, education, and ethnicity.
JEBV seropositives defined as positive for one of the following anti-EBV antibodies known to be associated with NPC: viral capsid antigen IgA, EBV nuclear
antigen 1 IgA, early antigen IgA, DNA binding protein IgG, and anti-DNase IgG.
N/A = not applicable, NS = not significant, PMR = proportionate mortality ratio.
On
On
-------
1
2
3
4
5
6
7
8
9
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
Following these reports of increased risk of NPC associated with formaldehyde exposure,
a series of post hoc analyses of similar data were undertaken by Marsh and coworkers (Marsh et
al., 2007a, b, 2002, 1996; Marsh and Youk, 2005). Briefly, these studies focused on the specific
findings from a single plant in the NCI cohort (Wallingford, Connecticut) that generated the
majority of the NPC cases.
In the most recent subsequent report, Marsh et al. (2007a) continues to argue against a
formaldehyde-NPC association. Although earlier reports speculated on anecdotal evidence that
the statistically significant excess risk of NPC observed at the Wallingford, Connecticut, plant
reflected the influence of unmeasured nonoccupational risk factors associated with employment
outside the plant, the new report (Marsh et al., 2007a) suggests that occupational or hobby-
related work in silversmithing may have confounded the observed effect of formaldehyde on the
increased risk of NPC. In this report, Marsh et al. (2007a) show that their subjectively assessed
work in silversmithing is strongly associated with NPC. While the reported ORs are indeed quite
high, the estimates are extremely unstable and it is not clear how many a priori hypotheses were
tested for statistical significance. There are no citations of an association between silversmithing
exposures and NPC in the medical literature. Marsh and coworkers mention that there was
concordance of silver manufacturing history in the Wallingford, Connecticut, area. If
silversmithing exposures are indeed independent risk factors for NPC, it would be expected that
the rates of NPC in the surrounding counties with historical silver-related exposures would be
elevated. However they are not increased, as evidenced by the comparability of the increased
rates of NPC among the plant workers compared with both the national and local county rates
that were very similar (Marsh et al., 2007a). The comparable rates indicate the counties' rates of
NPC were very similar to the national rates and weaken an association between silversmithing
and NPC. Given the many post hoc reexaminations of alternative hypotheses to explain the
original NCI findings, it is more likely that silversmithing is an artifactual confounder.
4.1.2.1.1.2. Professional cohort studies. Two cohort studies of professional groups, such as
anatomists, pathologists, embalmers, and funeral directors, examined the risk of NPC and
formaldehyde exposure. In general, measurements of formaldehyde concentrations were not
available in studies of professionals but are generally below 1 ppm (IARC, 1995; Korczynski,
1994; Stewart et al., 1992; Moore and Ogrodnik, 1986). Hayes et al. (1990) reported an excess
risk of NPC among male professional embalmers and funeral directors, based on 4 deaths with
1.9 expected based on age, gender, and calendar-year-specific proportions of deaths in the U.S.
population. Hansen and Olsen (1995) studied male Danish cancer patients employed in
companies in which formaldehyde was used or produced. Only a slight excess risk of NPC was
This document is a draft for review purposes only and does not constitute Agency policy.
4-67 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
found, based on 4 cases with 3.2 expected (standardized proportionate incidence ratio [SPIR]
1.3). Hansen and Olsen (1995) also reported on a significantly elevated risk of sinonasal cancer.
4.1.2.1.1.3. Case-control studies. Five case-control studies (West et al., 1993; Roush et al.,
1987; Vaughan et al., 1986a, b; Olsen et al., 1984) reviewed by IARC in 1995 provided evidence
of excess risks of NPC due to formaldehyde. Most of these studies report significant and
nonsignificant elevations in risk of NPC in the range of 1.5-3.0, with some higher than 5.0. In
its report, IARC (1995) concluded that, taking the data as a whole, formaldehyde appears to have
a causal role in the induction of NPC, recognizing that the conclusion is based on small numbers
of cancer cases.
Three case-control studies have been conducted since the 1995 IARC report. Armstrong
et al. (2000) found no association between formaldehyde exposure and NPC (adjusted OR = 0.71
[95% CI: 0.34-1.43]), controlling for wood dust and industrial heat. Using data from the
Surveillance Epidemiology and End Results (SEER) program, Vaughan et al. (2000) found an
OR for ever-exposed persons of 3.1 (95% CI: 1.0-9.6) among cases of epithelial NPC,
suggesting differences in the etiology of cancers at this site. There was a trend of increasing risk
of NPC with increasing duration of exposure and cumulative exposure, controlling for wood dust
exposure. Finally, Hildesheim et al. (2001) found that exposure to formaldehyde produced
modest risk elevations for duration of exposure (OR =1.6 for 10 years or less and 1.2 for over
10 years of exposure), for cumulative exposure (ORs were 1.3 for <25 years of exposure and 1.5
for 25+ years of exposure), and for years since first exposure. Among those with EB V, the OR
was 2.7 (95%) CI: 1.2-6.2) for ever-exposed persons. The risk was significantly higher among
exposed persons whose work history was within the last 10 years (OR = 4.7 [95%> CI: 1.1-20.0])
and for those followed 20+ years after exposure (OR = 2.8 [95%> CI: 1.1-7.6]).
4.1.2.1.1.4. Summary of NPC studies. Findings from the large NCI cohort studies of NPC risk
due to formaldehyde exposure clearly show a consistent pattern of increased risk with increased
exposures. Post hoc reanalyses have challenged the interpretation of these findings but have not
been able to dispute the reported excess in NPC mortality (Marsh et al., 2007a, b, 2002, 1996;
Marsh and Youk, 2005). The major questions that have been raised by Marsh and coworkers
highlight the observation that the NPC findings appear to depend on the results of 1 of the 10
plants that made up the NCI cohort. While it is theoretically possible for coexposures at that
plant or among those workers to act as potential confounders or modifiers of the observed effect
of formaldehyde on increased risk of NPC, there is no solid evidence of such a relationship that
would outweigh or supersede the reported adverse effects of formaldehyde exposure. While all
This document is a draft for review purposes only and does not constitute Agency policy.
4-68 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
of the cohort members at the Wallingford, Connecticut, plant were also exposed to particulates,
the NCI investigators did observe a dose-response relationship with formaldehyde among
individuals with high particulate exposures, thereby strengthening the causal interpretation of the
formaldehyde relationship with an increased risk of NPC. The described association of a
potential occupational relationship with silversmithing and NPC has no basis in the medical
literature and is inconsistent with the supposition that this activity is common in the locality of
Wallingford, Connecticut, but has not been associated with increased rates of NPC in
surrounding New Haven County (Connecticut). Marsh and coworkers did report significantly
increased rates of pharyngeal cancer (including NPCs) among workers from the Wallingford
plant compared with both the county and national rates. It is more plausible that the observed
association at the Wallingford plant reflects higher formaldehyde exposures than at other plants.
The exposure levels at Plant 2 were even higher than at the Wallingford plant and were
associated with a fivefold increase in risk associated with NPC, even though this was based on a
single observed case and was not significant.
In addition to the evidence from the NCI cohort studies, modest additional evidence is
found in the professional cohort studies of Hayes et al. (1990) and Hansen and Olsen (1995).
The rarity of the disease and difficulties in obtaining valid and reliable historical exposure
estimates are substantial limitations of these cohort studies. Further evidentiary support comes
from the results of several case-control studies that support an increased risk of NPC from
exposure to formaldehyde. The studies of Vaughan et al. (2000) and Hildesheim et al. (2001)
provide evidence of an association of NPC with exposure to formaldehyde. Vaughan et al.
(2000) found a dose-response relationship of NPC with increasing exposure to formaldehyde, as
did Hildesheim et al. (2001). These studies, in general, are easier to conduct and may provide
more statistical power for a specific level of risk estimate than do cohort studies.
4.1.2.1.2. Nasal and paranasal cancer
4.1.2.1.2.1. Case-control studies. Eight case-control studies were evaluated in the 1995 IARC
monograph regarding the risk of nasal cavity and accessory sinuses from exposure to
formaldehyde (Luce et al., 1993; Roush et al., 1987; Hayes et al., 1986; Olsen and Asnaes, 1986;
Vaughan et al., 1986a, b; Brinton et al., 1984; Olsen et al., 1984). Of three studies that identified
a cell type, two reported a positive finding of sinonasal cancer (Hayes et al., 1986; Olsen and
Asnaes, 1986). One of the positive studies did not report any exposure to the potentially
confounding influence of wood dust, while the other two did report an adjustment for exposure
to wood dust. Of the remaining five studies where a cell type was not identified, only Roush et
al. (1987) and Olsen et al. (1984) found positive results. The remaining studies (Vaughan et al.,
This document is a draft for review purposes only and does not constitute Agency policy.
4-69 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
1986a, b; Brinton et al., 1984) from the 1995 IARC monograph did not find associations between
exposure and sinonasal cancer. Study details of the epidemiologic studies of nasal and paranasal
cancer are summarized in Table 4-2. Vaughan et al. (1986b) matched 53 sinonasal cancer
patients to 552 controls. Potential residential exposure to formaldehyde was estimated by
utilizing residence in a mobile home with or without the presence of UFFI or particleboard or
plywood as a surrogate for exposure. The authors found an OR of 1.5 for sinonasal cancer in
subjects reporting residence of 10 or more years in a mobile home with UFFI before diagnosis.
A higher OR (1.8) was reported for less than 10 years of mobile home residence. However,
because actual formaldehyde levels in the subjects' mobile homes are unknown, the exposure
estimates are, at best, imprecise surrogates that typically have the effect of attenuating any true
risk. In Vaughan et al. (1986a), the same cases and controls were examined for occupational
exposures to formaldehyde, but no increase in risk of nasal or paranasal cancer was reported.
More recently, Luce et al. (2002) pooled data from 12 case-control studies. Combined,
these studies had 195 adenocarcinomas and 432 SCCs of the sinonasal passages compared with
3,136 controls. The authors reported a significant increase in the risk of sinonasal
adenocarcinoma in men (adjusted OR = 3.0 [95% CI: 1.5-5.7]; 91 cases) and in women (adjusted
OR = 6.2 [95% CI: 2.0-19.7]; 5 cases) with a high probability of exposure to formaldehyde. For
SCCs, the ORs were more modest: OR = 1.2 in men and OR = 1.5 in women for a high
probability of exposure to formaldehyde. In an analysis of 11 formaldehyde-exposed cases of
sinonasal adenocarcinomas who were not exposed to wood dust, there was an elevated risk in
men (OR = 1.9; 3 cases) and a significantly increased risk in women (OR =11.1 [95% CI: 3.2-
38.0]; 5 cases) with a high probability of exposure to formaldehyde. Limitations of these studies
were the lack of information about the actual levels or intensity of exposure to formaldehyde,
exposure to multiple occupational carcinogens, and the small number of cases in some
subgroups. In spite of those limitations, which generally obfuscate the observation of a true
underlying effect, these studies identified effects of formaldehyde that were statistically
significant predictors of sinonasal cancers.
This document is a draft for review purposes only and does not constitute Agency policy.
4-70 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer
S"4
>3*
§•
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of cases)
o
5 s
to
o
^ Co'
to &•
§ ^
>S
>S
TO
TO'
*
Olsen and Asnaes
(1986)
Case-control study of
histologically confirmed cases
of squamous cell
carcinoma/lymphoepithelioma
of the sinonasal cavities and
paranasal cancers in 215 men
and adenocarcinomas of the
sinonasal cavities and
paranasal cancers in 39 men
matched with 2,465 controls
with other cancers from the
Danish Cancer Registry,
1970-1982.
Employment histories
after 1964 from files
maintained by Danish
Cancer Registry
estimated by industrial
hygienists.
Squamous cell carcinoma/lymphoepithelioma
Ever vs. never
Formaldehyde only
RR
2.0
(95% CI: 0.7-5.9)
Formaldehyde + wood
dust
RR
1.6
(95% CI: 0.8-3.3)
10
or more vears since first exposure
Formaldehyde only
RR
1.4
(95% CI: 0.3-6.4)
Formaldehyde + wood
dust
RR
1.8
(95% CI: 0.7—4.4)
Adenocarcinoma
Ever vs. Never
Formaldehyde only
RR
7.0
(95% CI: 1.1—44)
Formaldehyde + wood
dust
RR
40.0
(95% CI: 22-71)
10 or more years since first exposure
Formaldehyde only
Formaldehyde + wood
dust
RR
RR
9.5
44.0
(95% CI: 1.6-58)
(95% CI: 22-88)
Hayes et al. (1986)
Case-control study of 91 men
with SCC of the nasal cavity
and paranasal sinuses, from
clinical records of six medical
institutions in the Netherlands.
195 controls from living and
deceased males from
municipal residence registries,
from 1978-1981.
Cases selected from
clinical records of six
institutions in the
Netherlands. 91 male
patients and 195
controls from living and
deceased males from
municipal residence
registries with little or
no exposure to wood
dust. Industrial
hygienists evaluated job
histories according to
probability of exposure
based on job records.
Industrial hygienist A
Any exposure
Moderate exposure
High exposure
Industrial hygienist B
Any exposure
Moderate exposure
High exposure
RR
RR
3.0
2.7
3.1
1.9
1.4
2.4
(90% CI: 1.3-6.4)
(90% CI: 1.0-7.2)
(90% CI: 0.9-10.0)
(90% CI: 1.0-3.6)
(90% CI: 0.5-3.4)
(90% CI: 1.1-5.1)
-------
K
s
TO
>3
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer
§•
TO
S
Reference
Study design
Exposure assessment
Results, statistical significance (number of cases)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
Roushetal. (1987)
Population-based case-control
study of 198 male cases of
sinonasal cancer from the
Connecticut Tumor Registry
who died of any cause in
1935-1975. Controls were
605 males dying in
Connecticut during the same
time period, randomly selected
from state death certificates.
Occupations from city
directories and
evaluation of job by
industrial hygienist who
classed exposure into I,
probably exposed to
some level most of
working life; II,
probably exposed to
some level most of
working life and
probably exposed to
some level 20+ years
before death; III,
probably exposed to
some level most of
working life and
probably exposed to a
high level in some
years; IV, probably
exposed to some level
most of working life
and probably exposed
to a high level 20+
years before death.
Exposure levels
I
II
III
IV
Sinonasal cancer
ORa
0.8
1.0
1.0
1.5
(95% CI: 0.5-1.3)
(95% CI: 0.5-1.8)
(95% CI: 0.5-2.2)
(95% CI: 0.6-3.9)
Luce et al. (1993)
Case-control study of men
with sinonasal cancer
(histologically confirmed), 77
with adenocarcinoma, 59 with
squamous cell carcinomas,
and 25 tumors of other types,
matched with 409 controls
from 27 French hospitals and
Industrial hygienist
estimation based on job
histories from personal
interviews. Subjects
were broken out into no
exposure, possible
exposure, or
probable/definite
to
Adenocarcinoma
Possible exposure
Probable/definite exposure
Average level
<2
>2
Duration (years)
<20
>20
OR0
OR0
OR0
1.28
4.15
5.33
1.03
6.86
(95% CI: 0.16-10)
(95% CI: 0.96-18)
(95% CI: 1.28-22)
(95% CI: 0.18-5.77)
(95% CI: 1.69-28)
-------
K
s
TO
>3
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer
§•
TO
S
Reference
Study design
Exposure assessment
Results, statistical significance (number of cases)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
from lists of names supplied
by patients.
exposure. Those
classed as
probable/definite
exposure further
categorized into three
levels of frequency of
exposure during a
normal workweek: 1 =
<5% of the time; 2 = 5-
30% of the time; and 3
= >30% of the time.
Concentration was
categorized into 3
levels: low (<0.0 ppm);
medium (0.1-1 ppm);
and high (>1 ppm).
The exposure index =
concentration x
frequency. Cumulative
level = sum of exposure
indices. Average level
= cumulative
level/duration and
ranged from 1 to 9.
Nearly all cases had had
wood dust exposure.
Cumulative level (years)
<30
30-60
>60
Age 1st exposed (years)
<15
16-20
>20
Date 1st exposed (years)
After 1954
Before 1954
Other cell type carcinomab
Possible exposure
Probable/definite exposure
Average level
<2
>2
Duration (years)
<20
>20
Cumulative level (years)
<30
>30
Age 1st exposed (years)
<20
>20
Date 1st exposed (years)
After 1954
Before 1954
Job exposure matrix
based on interview data
developed for pooled
OR0
OR0
OR0
OR0
OR0
OR0
OR0
OR0
OR0
1.13
2.66
6.91
9.99
4.12
2.74
6.02
4.26
0.81
1.67
3.04
2.82
1.62
2.18
2.21
2.03
2.36
0.48
3.27
95% CI: 0.19-6.95)
95% CI: 0.38-19)
95% CI: 1.69-28)
95% CI: 1.85-54)
95% CI: 0.95-18)
95% CI: 0.58-13)
95% CI: 1.18-31)
95% CI: 1.06-17)
95% CI: 0.15-4.36)
95% CI: 0.51-5.42)
95% CI: 0.95-9.7)
95% CI: 0.94-8.4)
95% CI: 0.48-5.51)
95% CI: 0.65-7.31)
95% CI: 0.73-6.73)
95% CI: 0.63-6.54)
95% CI: 0.76-7.33)
95% CI: 0.05-4.35)
95% CI: 1.15-9.33)
Luce et al. (2002)
LtJ
Pooled analysis of 195
adenocarcinomas and 432
squamous cell carcinomas of
Adenocarcinoma
High probability of exposure
Men
OR
3.0
(95% CI: 1.5-5.7)
-------
K
s
TO
>3
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer
S"4
>3*
§•
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of cases)
the sinus/nasal cavity matched
with 3,136 controls from 12
case-control studies.
analysis. Industrial
hygiene data used to
develop indices of
exposure. 11
formaldehyde cases
reported no exposure to
wood dust.
Sqi
Women
lamous cell carcinoma
High probability of expos
Men
Women
ORe
ure
ORe
ORe
6.2
1.2
1.5
(95% CI: 2.0-20)
(95% CI: 0.8-1.8)
(95% CI: 0.6-3.8)
Brinton et al. (1984)
Case-control study of 160
patients with cancer of the
nasal cavity and paranasal
sinuses from four North
Carolina and Virginia
hospitals matched with 290
hospital controls with other
conditions, based on
occupational exposures.
Interview data on job
history. Estimation of
exposure based on
industry type. Only
two cases employed in
industry associated with
formaldehyde. There
were no deaths in the
high exposure category.
Ov
Re
Ye
erall male and female
sidence in mobile home
ars of exposure to particleb
1 to 9
10 or more
RR
ORf
oard
ORf
0.35
0.6
1.8
1.5
(95% CI: 0.1-1.8)
(95% CI: 0.2-1.7)
(95% CI: 09-3.8)
(95% CI: 0.7-3.2)
Olsen et al. (1984)
Case-control study of 488
cases of nasal cancer linked to
the Danish Cancer Registry
during 1970-1982. Controls
were individuals with cancer
of the colon, rectum, breast,
and prostate. Three controls
per case were selected for the
same distributions of age, sex,
and year of diagnosis as cases.
Employment histories
after 1964 from files
maintained by Danish
Cancer Registry
estimated by industrial
hygienists.
Men
Formaldehyde onlv
Ever exposed
Exposure to wood dust and
formaldehyde
Ever exposed
1st exposure >10 years or
more before diagnosis
RR
RR
RR
RR
2.8
3.1
3.5
4.1
(95% CI: 1.8—4.3)
(95% CI: 1.8-5.3)
(95% CI: 2.2-5.6)
(95% CI: 0.2.3-7.3)
(33)
(23)
(28)
(20)
Hansen and Olsen
(1995)
Proportionate incidence study
of 2,041 men with sinonasal
cancer who died between 1970
and 1984 identified from the
Danish Cancer Registry
matched with the Danish
Linked companies
through tax records to
national Danish Product
Register, where
companies must report
amount of
Overall
Low formaldehyde
Formaldehyde, no wood
dust
Unknown
SPIR
SPIR
SPIR
SPIR
SPIR
2.3
0.8
3.0
5.0
1.0
(95% CI: 1.3—4.0)
(95% CI: 0.02-4.4)
(95% CI: 1.4-5.7)
(95% CI: 0.5-13)
(95% CI: 0.03-6.1)
(13)
(1)
(9)
(2)
(1)
o
2 »
5 s
to
o
!
>{
JS*
*
--J
4^
-------
is*
Table 4-2. Case-control studies of formaldehyde and nasal and paranasal cancer
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of cases)
o
5 ^ §
6, to'
to Sj-
§ ^
s ^
§ 3
?»»i.
Supplementary Pension Fund
whose longest work
experience occurred at least 10
years before the cancer
diagnosis. The measure of
risk was the SPIR, which
measured the proportion of
cases of sinonasal cancer in
formaldehyde-associated
companies relative to the
proportion of cases of
sinonasal cancer among all
employees in Denmark.
formaldehyde used per
year.
Coggon et al. (2003)
Cohort mortality study of
14,014 men employed in 6
factories of the chemical
industry in Great Britain from
periods during which
formaldehyde was produced.
Cohort followed through
2000. Sinonasal cancer
mortality SMRs based on
English and Welsh age and
calendar-year-specific
mortality rates.
Exposures assessment
based on data
abstracted from
company records. Each
job categorized as
background, low,
moderate, high, or
unknown levels. For
analysis of sinonasal
cancer, no gradient used
because of small
number of observed
cases.
Overall
SMR
0.87
(95% CI: 0.11-3.14)
(2)
aAdjusted for age at death, year at death, and availability of occupational information.
bAll had medium to high exposure to wood dust.
0 Adjusted for age and exposure to glues and adhesives.
dAdjusted for age and study.
"Adjusted for age, study, and cumulative exposure to dust.
fAdjusted for cigarette smoking, alcohol consumption, gender, and age.
-------
1
2
3
4
5
6
7
8
9
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
4.1.2.1.2.2. Cohort studies of nasal and paranasal cancer. I ARC (1995) also reported the
results of several cohort studies of professional and industrial workers for nasal and paranasal
cancer (Andjelkovich et al., 1995; Gardner et al., 1993; Hall et al., 1991; Hayes et al., 1990;
Bertazzi et al., 1989; Edling et al., 1987; Blair et al., 1986; Stroup et al., 1986; Harrington and
Oakes, 1984; Levine et al., 1984; Walrath and Fraumeni, 1984, 1983; Friedman and Ury, 1983).
Only a few studies reported any cases of sinonasal cancer. No cases of this type of cancer were
reported in any of the studies of professional workers examined by the IARC. Only 2 cases (2.2
expected) were reported by Blair et al. (1986), and only 1 case (1.7 expected) was reported by
Gardner et al. (1993). The likelihood of finding this rare tumor type in a long-term cohort study
is low.
Three subsequent cohort studies reported on nasal and paranasal cancer. Hansen and
Olsen (1995), in a proportional incidence study, found a significantly increased risk of sinonasal
cavity cancer (SPIR = 2.3 [95% CI: 1.3-4.0]; 13 observed) in 265 Danish industries, where
2,041 of 91,182 cancer patients had at least 10 years of continuous formaldehyde-related work
experience before diagnosis. Coggon et al. (2003), in a cohort study of 14,014 employees in six
chemical factories in Great Britain, found only 2 deaths from sinonasal cancer (2.3 expected
based on national death rates in Great Britain). Finally, Hauptmann et al. (2004) evaluated the
sinonasal cancer risk in the NCI cohort and found three cases (SMR =1.19 [95% CI: 0.38-3.68])
among those with a 15-year lag period.
4.1.2.1.2.3. Summary of nasal and paranasal cancers. The pooled case-control study of Luce
et al. (2002) provides strong evidence of an association between formaldehyde exposure and
increased risk of sinonasal adenocarcinoma. The cohort studies may not have had sufficient
statistical power to show an association, and studies that did not distinguish cancer type may
have aggregated a truly causal relationship with a noncausal relationship with SCC. In summary,
there appears to be increased risk of sinonasal cancer associated with formaldehyde exposure
with or without exposure to wood dust. The effect appears to be stronger when the risk is
stratified by cancer type with higher risks of adenocarcinoma compared with SCC. Taken
together with the NPC findings in the neighboring tissue, it is concluded that there is evidence of
higher risks of sinonasal cancer associated with exposure to formaldehyde.
4.1.2.1.3, Other respiratory tract cancers. Of six cohort studies of buccal/pharynx cancer in
studies of professionals reviewed by IARC (Hayes et al., 1990; Logue et al., 1986; Stroup et al.,
1986; Levine et al., 1984; Walrath and Fraumeni, 1984, 1983), no evidence of a risk associated
with exposure to formaldehyde was reported (see Table 4-3). In studies of industrial worker
cohorts where buccal/pharynx cancer was examined (Andjelkovich et al., 1995; Stayner et al.,
This document is a draft for review purposes only and does not constitute Agency policy.
4-76 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
1988; Blair et al., 1986), only one (Stayner et al., 1988) reported an excess risk of death from this
tumor (SMR = 3.4, based on four deaths in a cohort of 6,741 white women). Three case-control
studies (Merletti et al., 1991; Vaughan et al., 1986a, b) did not find an association between oral
cavity, oropharyngeal, and hypopharyngeal cancers and formaldehyde exposure. However,
Merletti et al. (1991) found an elevated OR of 1.8 associated with probable or definite exposure
to formaldehyde in a study of 86 patients matched with 373 controls. There was no risk of
laryngeal cancer associated with formaldehyde in a case-control study (Wortley et al., 1992) of
235 patients with laryngeal cancer and 547 controls. The OR in that study was 1.0 (adjusted for
age, smoking, drinking, and level of education). IARC (1995) concluded that there was little
evidence of an increased risk of laryngeal cancer.
4.1.2.1.3.1. Cohort studies of other respiratory tract cancers. Hansen and Olsen (1995)
reported a 10% increase in the risk of cancer of the buccal cavity and pharynx (SPIR = 1.1) in
their proportional incidence study of Danish workers. Marsh et al. (1996) reported no excess
risk of buccal cavity cancer cases (SMR = 1.31) based on U.S. rates and no excess based on state
mortality rates (SMR = 1.0). For oropharyngeal cancer, the SMR was 1.84 (based on two cases),
the SMR for hypopharyngeal cancer was 1.41 (based on one case), and the SMR for laryngeal
cancer was 1.47 (based on six cases). The latter risks were elevated even when SMRs were
derived from Connecticut mortality rates.
The Marsh et al. (2002) update also derived elevated risk estimates for oropharyngeal,
hypopharyngeal, and pharyngeal-unspecified cancers. The SMRs ranged from 2.11 to 2.25
based on U.S. rates and 1.52 to 1.89 based on county rates. When combined with NPC
International Classification of Death (ICD) codes 146-149 to increase statistical power, the total
pharyngeal cancer SMRs were significant based on U.S. death rates (SMR = 2.63, n = 22,p<
0.01) and county death rates (SMR = 2.23, n = 22,p< 0.01) and remained significant for both
short-term (less than 1 year) and long-term exposures. Furthermore, using the exposure
estimates of Marsh et al. (1996), both cumulative and average exposure to formaldehyde resulted
in elevated SMRs, some of which were significant for pharyngeal cancer. Coggon et al. (2003)
identified 14 cases of cancer of the larynx (13.1 expected) in their cohort of formaldehyde-
exposed factory workers. Pinkerton et al. (2004) found an excess risk of mortality from buccal
cavity cancer (SMR = 1.33, four deaths observed) and a deficit of the risk of pharyngeal cancer
(SMR = 0.64, three observed). Since the number of observed deaths was small and the risk
estimates were subject to much variation for both studies, no conclusions about cancer risk can
be drawn.
This document is a draft for review purposes only and does not constitute Agency policy.
4-11 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-3. Epidemiologic studies of formaldehyde and pharyngeal cancer (includes nasopharyngeal cancer)
S"4
>3*
&
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number observed deaths for cohort study)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
Marsh et al. (2002)
Retrospective cohort
mortality study of 7,328
workers hired up to 1984
and followed until 1998 in
one plant from Blair et al.
(1986, 1987) and
Hauptmann et al. (2004).
Mortality was compared
with death rates in two
Connecticut counties and
U.S. A nested case-control
analysis was also conducted
with 4 controls matched on
age, year of birth, race, and
sex randomly selected from
cohort. Conditional logistic
model was used for nested
case-control analysis.
Worker-specific exposures
from job exposure matrix
were based on available
sporadic sampling data from
1965-1987, job descriptions,
and verbal job descriptions
by plant personnel and
industrial hygienists.
Exposures were then ranked
on a 7-point scale. An
exposure range was assigned
to each rank. 17% of jobs
validated with company
monitoring data, remaining
83% based on professional
judgment. Pre-1965 levels of
formaldehyde were assumed
to be the same as post-1965
levels.
oo
Cohort studv
Overall
U.S.
SMR
2.63
(95% CI: 1.65-3.98)
(22)
County
SMR
2.23
(95% CI: 1.40-3.38)
(22)
Short-term worker (<1 year)
SMR
2.35
(95% CI: 1.22-4.11)
(12)
Long-term worker (1 or more years)
SMR
2.10
(95% CI: 1.01-3.86)
(10)
Cumulative exp. (ppm-years) county
Unexposed
SMR
1.24
(95% CI: 0.15-4.49)
(2)
>0 to <0.004
SMR
3.31
(95% CI: 1.22-7.21)
(6)
0.004-0.219
SMR
2.06
(95% CI: 0.83-4.24)
(7)
0.22+
SMR
2.30
(95% CI: 0.92-4.73)
(7)
Average Exposure (ppm) county
Unexposed
SMR
1.24
(95% CI: 0.15-4.49)
(2)
>0 to <0.03
SMR
2.02
(95% CI: 0.74-4.40)
(6)
0.03-0.159
SMR
3.82
(95% CI: 1.54-7.88)
(7)
0.16+
SMR
2.03
(95% CI: 0.82-4.19)
(7)
Exposure to <0.2 ppm
SMR
1.72
(95% CI: 0.74-3.39)
(8)
Exposure to >0.2 ppm
SMR
2.68
(95% CI: 1.46-4.49)
(14)
Exposure to <0.7 ppm
SMR
2.12
(95% CI: 1.21-3.45)
(16)
Nested case-control analysis
Cumulative exp. (ppm-years)
<0.004
OR
1.00
(8)
0.004-0.219
OR
0.71
(95% CI: 0.20-2.43)
(7)
0.22+
OR
0.79
(95% CI: 0.18-3.20)
(7)
Average exposure (ppm)
<0.03
OR
1.00
(8)
0.03-0.159
OR
1.71
(95% CI: 0.47-6.10)
(7)
0.16+
OR
0.99
(95% CI: 0.27-3.55)
(7)
Exposure to >0.2 ppm
OR
1.35
(95% CI: 0.45-4.25)
(14)
Exposure to >0.7 ppm
OR
1.60
(95% CI: 0.15-9.77)
(6)
-------
Table 4-3. Epidemiologic studies of formaldehyde and pharyngeal cancer (includes nasopharyngeal cancer)
S"4
>3*
s
Reference
Study design
Exposure assessment
Results, statistical significance (number observed deaths for cohort study)
Coggon et al. (2003)
Cohort mortality study of
14,014 chemical workers
employed in 6 British
factories.
Based on data abstracted
from company records. Each
job was categorized as having
background, low, moderate,
high, or unknown levels of
formaldehyde.
Overall
High exposure
SMR
SMR
1.55
1.91
(95% CI: 0.87-2.56)
(95% CI: 0.70-4.17)
(15)
(6)
Shangina et al.
(2006)
Multicentered, hospital-
based case-control study in
four European countries;
men only. Cancer cases: 34
hypopharyngeal; 316
laryngeal. Controls: 728
hospital patients with
various conditions.
Exposures determined by
local industrial hygienists,
chemists, and physicians.
Coding was established and
standardized. Categories
were developed for 73
agents; frequency was
estimated as the proportion of
time a worker was exposed.
Linear trends were examined
for duration in years,
weighted duration in hours,
and cumulative exposure.
Larvneeal cancer:
Formaldehyde
Ever vs. never
Highest cumulative
(>22,700 mg/m3-hours)
vs. lowest
Tests of trends:
Years exposed
Cumulative exposure
OR
OR
p = 0.06
p = 0.07
1.68
3.12
(95% CI: 0.85-3.31)
(95% CI: 1.23-7.91)
o
S ?
a, Co
8- a
TO Sj-
§ I»
? f
s >
^ ri
?»»i.
vo
-------
1
2
3
4
5
6
7
8
9
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
Hauptmann et al. (2004) combined URT cancers (cancers of the salivary gland, mouth,
nasopharynx, nasal cavity, and larynx). For average intensity of exposure (AIE), the RR was
1.69 for the medium exposure category (0.5 to <1.0 ppm) and 2.21 (p < 0.05) for the high
exposure category (>1.0 ppm). For peak exposure, the RR was 1.24 for the medium exposure
category (2.0 to <4.0 ppm) and 1.65 for the high exposure category (>4.0 ppm). For cumulative
exposure, the RR was 1.92 for the medium exposure category (1.5 to <5.5 ppm) but 0.86 in the
high exposure category (>5.5 ppm-years). The dose trends for these analyses, while suggestive
for average and peak exposures, were not statistically significant.
4.1.2.1.3.2. Case-control studies of other respiratory cancers. Gustavsson et al. (1998)
conducted a case-control study of 545 cases of SCC of the oral cavity, oropharynx, hypopharynx,
larynx, and esophagus, frequency-matched by age and region with 641 controls. Regression
analyses among 545 male cases showed elevated but nonsignificant risks of SCC of the oral
cavity (OR = 1.28), esophagus (OR = 1.90), and larynx (RR = 1.45) associated with
formaldehyde exposure. However, several of the carcinoma types were statistically significantly
associated with exposure to welding fumes, polyaromatic hydrocarbons, asbestos, and metal
dust.
In a case-control study, Laforest et al. (2000) examined 201 patients with squamous cell
hypopharyngeal cancer and 296 patients with squamous cell laryngeal cancer, who were matched
to 296 controls with cancers of other sites in 15 French hospitals. Adjusting for potential
confounders, the OR of hypopharyngeal cancer in patients with a high probability of exposure to
formaldehyde was 3.78 (95% CI: 1.50-9.49). The ORs were significantly increased with both
exposure durations and high cumulative level of exposure.
Marsh et al. (2002) conducted a nested case-control study of the 22 pharyngeal cancer
deaths in the Wallingford, Connecticut, plant cohort. Each of the pharyngeal cancer deaths was
matched on race, sex, age, and year of birth to four controls from the cohort. Twenty of the
22 cases were exposed to formaldehyde, yielding an OR of 3.04 after adjustment for smoking
and year of hire. There was little or no association of pharyngeal cancer incidence in these
workers with either average or cumulative exposure, based on the exposure estimates in this
study. There was a suggested trend of increasing OR with increasing duration of exposure for
any formaldehyde exposure as well as for formaldehyde exposure >0.2 ppm. The results of this
nested case-control analysis are inconclusive because of its low statistical power and
questionable exposure estimates, which differed substantially from those estimated by the NCI
(see Section 4.1.1.1). In addition, the relatively flat dose-response curve in the nested case-
This document is a draft for review purposes only and does not constitute Agency policy.
4-80 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
35
control study contradicts the positive dose-response curves reported (particularly for NPCs) in
the same study, based on SMRs derived from county and U.S. death rates in the cohort analysis.
Shangina et al. (2006) conducted a multicentered case-control study in Europe and Russia
of 34 cases of hypopharyngeal cancer, 316 cases of laryngeal cancer, and 728 controls. With
regard to formaldehyde exposure, a nonsignificant positive association was found for laryngeal
cancer (OR = 1.68, 95% CI: 0.85-3.31). Trends over increasing exposure were found for
duration of exposure in years (p = 0.06) and for cumulative exposures (p = 0.07). The
investigators reported an OR of 3.12 (95% CI: 1.23-7.91) for the highest cumulative exposure
group (>22,700 mg/m3-hours) compared with the unexposed group.
4.1.2.1.3.3. Summary of other respiratory cancers. The evidence for a compound-specific
effect on the risk of buccal/pharynx, oral cavity, oropharynx, hypopharynx, and laryngeal
cancers as a result of exposure to formaldehyde is minimal. Only the study by Laforest et al.
(2000) and, to a lesser extent, that by Shangina et al. (2006) provided evidence of an association
between formaldehyde and these tumors. However, even the study of Laforest et al. (2000) had
major limitations that made the evidence of an association suggestive at best.
4.1.2.1.4. Lung Cancer. None of the cohort studies of workers in specific professions indicated
excess risks of lung cancer. Of the professional studies reviewed, the RRs range from an
extremely low value (SMR) of 0.2 in Hall et al. (1991), based on nine deaths, to an RR
(proportional mortality ratio) of 1.1, based on 70 lung cancer deaths in Walrath and Fraumeni
(1983).
4.1.2.1.4.1. Industrial worker cohort studies oflune cancer. Evidence of a relationship
between formaldehyde exposure and lung cancer is conflicting, with some studies showing
modest increases while others show significant deficits in risk. There is, at best, only weak
evidence from several studies to suggest that exposure to formaldehyde is associated with lung
cancer.
Several industrial cohort studies (Andjelkovich et al., 1995; Bertazzi et al., 1989, 1986;
Stayner et al., 1988; Edling et al., 1987) reported no significant excess risks of lung cancer from
exposure to formaldehyde. No consistent association between formaldehyde exposure and lung
cancer was found in several reports of the NCI 10-plant cohort originally investigated by Blair et
al. (1987, 1986). Hauptmann et al. (2004) gives the most recent report on this cohort, which has
been studied in part or in its entirety by several others (Marsh et al., 1994, 1992a, b; Sterling and
Weinkam, 1994, 1989a, b, 1988; Robins et al., 1988; Liebling et al., 1984; Fayerweather et al.,
This document is a draft for review purposes only and does not constitute Agency policy.
4-81 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
34
1983; Wong, 1983; Marsh, 1982). Marsh et al. (1996) reported small but significantly increased
risks of respiratory cancer in males at one plant from the cohort. The SMRs were 1.22 based on
U.S. rates, 1.34 based on Connecticut rates, and 1.28 based on county rates.
Similarly, Gardner et al. (1993) and Acheson et al. (1984) found a significant but modest
association between lung cancer and formaldehyde exposure (SMR =1.2 [95% CI: 1.1-1.4]). In
workers hired after 1964, the SMR was 1.1. No trends by level or duration of exposure were
found. Pinkerton et al. (2004) and Stayner et al. (1988) studied a cohort of 11,030 workers in
three garment plants and found an SMR of 1.1 for lung cancer. Some studies have found modest
elevations in risk of lung cancer with formaldehyde exposure, some of which were significant.
Coggon et al. (2003) updated the Gardner et al. (1993) study of industrial workers. By taking
data from all six factories together, results showed a statistically significant excess risk of lung
cancer in the high-exposure category when compared with British national mortality rates (SMR
= 1.58 [95% CI: 1.40-1.78]) and to local mortality rates (SMR= 1.28 [95% CI: 1.13-1.44]).
Callas et al. (1996) reanalyzed the cumulative exposure of 279 lung cancer cases among white
male workers from the NCI study, which comprised 80% of the NCI cohort (Blair et al., 1986).
The analysis revealed modest RRs of 1.46, 1.27, and 1.38 for lung cancer in the cumulative
exposure categories 0.05 to 0.5, 0.51 to 5.5, and greater than 5.5 ppm-years, respectively. None
of these RRs were significant. Finally, Matanoski (1991) reported a significant deficit in the risk
of respiratory cancer (SMR = 0.56 [95% CI: 0.44-0.70]; 77 observed) in pathologists
presumably exposed to formaldehyde based on U.S. mortality rates.
4.1.2.1.4.2. Case-control studies. Several case-control studies of lung cancer (Partanen et al.,
1990; Gerin et al., 1989; Bond et al., 1986; Coggon et al., 1984; Fayerweather et al., 1983;
Anderson et al., 1982) showed no excess lung cancer risk associated with potential exposure to
formaldehyde when analyzed by length of exposure, intensity, and potential exposure 5, 10, or
15 years before death or by combinations of these factors. By contrast, Coggon et al. (1984)
reported a statistically significant increase in risk of lung cancer among male patients with any
potential exposure to formaldehyde based on occupations listed on death certificates (SMR =1.5
[95% CI: 1.2-1.8]).
De Stefani et al. (2005) conducted a case-control study of 338 adenocarcinomas of the
lung in male patients admitted to four Montevideo hospitals from 1994 to 2000. The highest
ORs were for smoking (6.0 [95% CI: 3.3-11]) and for former smokers (4.0 [95% CI: 2.1-7.3]).
In addition, three agents (i.e., asbestos, silica dust, and formaldehyde) indicated significant
excess risks of lung adenocarcinoma after adjusting for smoking history. A significant exposure-
This document is a draft for review purposes only and does not constitute Agency policy.
4-82 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
duration relationship was found for formaldehyde for 21+ years of exposure (OR = 3.0 [95% CI:
1.6-5.8]; p = 0.004 for trend).
4.1.2.1.4.3. Summary of lung cancer. The evidence for an association between formaldehyde
and lung cancer is limited. Only one study has found a statistically significant effect (Coggon et
al., 2003). However, there may be other explanations rather than exposure to formaldehyde for
this association. Except for the findings of De Stefani et al. (2005), other studies of lung cancer
and exposure to formaldehyde have not supported this finding, including several well-done
cohort studies that were specifically designed to evaluate lung cancer. Until the Coggon et al.
(2003) study of British formaldehyde workers is replicated or reevaluated to determine the cause
of the excessive lung cancer risk, evidence from that study alone is insufficient at this time to
support an association between lung cancer and formaldehyde exposure.
4.1.2.1.5. Summary of respiratory tract cancers. Recent studies of NPC continue to support an
association with exposure to formaldehyde even at low levels. In some studies, the association
between formaldehyde and NPC persisted even when adjusted for the effect of potential
confounders (Hauptmann et al., 2004). Data from some reports have suggested a dose-response
relationship (Hauptmann et al., 2004; Marsh et al., 2002; Vaughan et al., 2000).
The risk of NPC was significantly elevated among industrial workers with cumulative
exposure, average exposure, and peak exposure to formaldehyde (Hauptmann et al., 2004). The
studies of the single Wallingford plant by Marsh et al. (2002, 1996, 1994) and Marsh and Youk
(2005) also revealed a dose-response trend, although the absolute exposure level estimates were
much lower. The relatively flat dose-response curve seen in the nested case-control study by
Marsh et al. (2002) of all pharyngeal cancers was inconsistent with the positive dose-response
curves reported in the same paper based on county and U.S. death rates, particularly for NPC.
Also of interest was the finding of a statistically significant increase in the risk of NPC in
formaldehyde-exposed workers who were seropositive for EBV in the Hildesheim et al. (2001)
study.
The pooled analysis by Luce et al. (2002) provides evidence of a relationship of sinonasal
cancer, particularly adenocarcinoma, with formaldehyde. However, as with some of the studies
of NPC, the findings are potentially confounded by concurrent exposure to wood dust. When
wood dust exposure was adjusted for in the analysis, the resulting risks were still positive but
based on small numbers and, as a result, subject to much variability. The more recent studies
continued to reveal small significant and nonsignificant associations among cancers of
This document is a draft for review purposes only and does not constitute Agency policy.
4-83 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
31
32
33
34
35
buccal/pharynx, oral cavity, hypopharynx, and larynx and exposure to formaldehyde. However,
the estimates were always based on small numbers.
A recent study that reported statistically significant lung cancer in association with a high
level of exposure to formaldehyde was conducted by Coggon et al. (2003). The investigators
suggested that unknown lifestyle factors, including smoking, could be responsible for the
finding. Despite the results of their analysis, the authors were unconvinced that formaldehyde
was the agent responsible for the elevation in lung cancer risk. However, De Stefani et al. (2005)
also reported a statistically significant risk of lung adenocarcinoma in formaldehyde-exposed
hospital patients even when smoking was controlled for in their analyses.
In all studies of formaldehyde and lung cancer, smoking remains an important
confounder and possibly an effect modifier. Residual confounding of smoking or other
respiratory exposures (e.g., wood dust or chemical or particular exposures) must always be
considered.
4.1.2.2. Non-Respiratory Tract Cancer
4.1.2.2.1. LHP cancers. Cancers of the hematopoietic system include lymphosarcoma,
reticulosarcoma, Hodgkin's disease, non-Hodgkin's disease, multiple myeloma, and all types of
leukemia, including lymphoid and myeloid. Virtually all of the studies of LHP cancers and
formaldehyde are cohort studies and are divided into two groups: professional and industrial.
Several of the studies of professional groups were reviewed in an IARC (1995) monograph and
are briefly discussed in the next section regarding their findings on cancer of the LHP system.
One case-control study of non-Hodgkin's lymphoma is discussed at the end of this section.
4.1.2.2.1.1. Professional cohort studies. Several cohort studies have been undertaken by
professional groups (i.e., anatomists, pathologists, embalmers, and funeral directors) because
their careers are likely to bring them into contact with formaldehyde. Some studies have
reported an increase in the risk of myelogenous leukemia and other LHP cancers (see Table 4-4).
A few of the increased risks were statistically significant. None of the studies of professionals
have used personal exposure measurements of formaldehyde or other chemicals, making
specificity for any single exposure difficult to determine.
Harrington and Shannon (1975) conducted a cohort mortality study of 2,079 British
pathologists (1955-1973) and 12,944 British medical laboratory technicians (1963-1973). When
compared with death rates for England and Wales, the all-cause SMR for the pathologists was
0.60 versus 0.67 for the laboratory technicians. There was a significant increase in the risk of
lymphatic and hematopoietic neoplasia (SMR 2.0; 8 observed with 3.3 expected;p < 0.01)
This document is a draft for review purposes only and does not constitute Agency policy.
4-84 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
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
among male pathologists. However, the SMR for technicians was only 0.6 (3 observed). The
low SMRs suggest that these professionals have a healthier profile compared with the British
population. No actual exposure estimates are available.
Harrington and Oakes (1984) expanded the above study to include 2,307 male and 413
female pathologists and laboratory technicians. Mortality was only examined from 1973 until
1980; deaths that occurred before 1974 were not included in the update. The SMR for leukemia
was 0.91 in men and 9.26 (based on one case) in women. Although the earlier LHP cancer
deaths were not included in this analysis, the investigators say in their conclusion that their
previous suggestion of an increase in certain lymphatic neoplasia was not confirmed in the
present study because of small numbers. The exceptionally low SMRs suggest that this group of
professionals enjoyed a healthier lifestyle compared with the British population as a whole. Just
as in the earlier studies of these professionals, no exposure estimates are available.
Hall et al. (1991) expanded the above study by including the newest members of the
Royal College of Pathologists. The cohort totaled 4,512 individuals, although only 3,069 males
and 803 females were included in the analysis. The reasons for this discrepancy were not
specified, although the authors mentioned that an unknown number of expected deaths for
Northern Irish and female Scottish pathologists were not calculated, 32 pathologists were lost in
follow-up, and cause of death was unknown for 9 individuals. Follow-up was extended from
1980 to 1986. Mortality was enumerated from 1974 to 1987, a period of time that differed from
both of the earlier studies described above. There were statistically not significant excess risks
for lymphatic and hematopoietic cancer (SMR 1.44; 10 observed) and leukemia (SMR 1.52;
4 observed) for both sexes combined, based on mortality rates in England and Wales.
Separately, there was 1 female death in the lymphatic and hematopoietic cancer category (0.57
expected). The most striking observation in this study is that, despite the low cancer mortality
(SMR 0.45 for all cancer; 53 observed but 118.19 expected), there was still an excess (but not
statistically significant) risk of hematopoietic cancers. This finding of an extremely low risk for
all cancers suggests that population death rates may not be appropriate as a referent group—for
example, the SMRs for lung cancer (0.19) and nonneoplastic respiratory diseases (0.23) were
significantly decreased, suggesting a lower prevalence of smoking among the pathologists
compared with the general population of England and Wales. However, the finding of a possibly
increased risk of LHP cancers should be analyzed further by selecting a more appropriate
reference population (another professional group without exposure to formaldehyde) or by
utilizing internal comparisons.
Walrath and Fraumeni (1983) conducted a proportionate mortality study of all embalmers
and funeral directors licensed in the state of New York between 1902 and 1980 who were known
This document is a draft for review purposes only and does not constitute Agency policy.
4-85 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
to have died between 1925 and 1980. The investigators requested death certificates for 1,678
persons but received only 1,263 (75%). The investigators restricted their analysis to 1,132
males. The distribution of the causes of death was compared with the age-, race-, and calendar-
year-specific proportions of deaths for each cause among the male U.S. population. Duration of
exposure was approximated by time since first license. While the methodology could not be
applied in all calculations because of data gaps, excess risks were found for lymphatic and
hematopoietic cancers, with a proportionate mortality ratio (PMR) of 1.2 (observed 25), and for
leukemia, with a PMR of 1.32 and proportionate cancer mortality ratio (PCMR) of 1.19
(12 observed). The PMRs were not affected when the estimates were stratified by latency
(<35 years or 35 years since first license) or by age at first license. Because the cause of death
could not be determined for nearly 25% of the study group, the risk estimates could be
underestimated. The metrics, PMR, and PCMR are not stochastic processes. An increase in one
cause would produce decreases in all the other causes.
Using the proportionate mortality method, Walrath and Fraumeni (1984) studied 1,007
deceased white male embalmers, members of the California Bureau of Funeral Directing and
Embalming, whose deaths occurred between 1925 and 1980. The decedents had to have been
licensed to practice between 1916 and 1978. For lymphatic and hematopoietic cancer, the PMR
was 1.22 (19 deaths observed). For leukemia alone, the PMR was 1.75 and significant
(12 deaths observed,/* < 0.05). Among embalmers licensed for 20 years or longer, the risk of
leukemia increased and was also significant (PMR 2.21; 8 observed; p < 0.05). But this study,
like the study of New York embalmers, had the same limitations discussed above. The
investigators did not provide information on the number of embalmers for whom no cause of
death could be found.
Levine et al. (1984) conducted a cohort mortality study of 1,477 male Ontario
undertakers first licensed between 1928 and 1957 and followed until the end of 1977. Out of
359 subjects who had died, there were 8 deaths from lymphatic and hematopoietic cancers
compared with 6.5 expected. Additionally, there were 4 deaths from leukemia versus
2.5 expected. Because death rates were not available for Ontario before 1950, person-years and
deaths before 1950 could not be counted. No actual exposure estimates are available for these
undertakers.
Stroup et al. (1986) conducted an historic cohort mortality study of 2,317 men who were
members of the American Association of Anatomists between 1888 and 1969. The investigators
derived SMRs from the U.S. white male population and used members of the American
Psychiatric Association (APA) as a comparison group. Vital status was ascertained between
1925 and 1979. Women were excluded from analysis because of the small numbers. Only
This document is a draft for review purposes only and does not constitute Agency policy.
4-86 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
738 deaths were observed versus 1,133.9 expected, based on U.S. death rates (SMR 0.65),
possibly indicating a sizable HWE. However, a slight increase in the risk of lymphatic and
hematopoietic cancers (SMR 1.2; 18 observed) and the risk of leukemia (SMR 1.5; 10 observed)
was evident. A significant increase in the risk of brain cancer (SMR 2.7; 10 observed; p < 0.05)
was also reported. When the leukemia analysis was restricted to the myeloid type, the SMR
increased to 8.8, based on five deaths (p < 0.05). The analysis using the APA group was
restricted to deaths that occurred between 1900 and 1969. This restriction removed five
leukemia deaths and person-years from the analysis because they likely died after 1969. Because
of this, there were only 3 leukemia deaths versus 3.6 expected, based on APA death rates. The
investigators concluded that the etiological agent had not been definitively identified, mentioning
that a wide range of solvents, stains, and preservatives, including formaldehyde, are used to
prepare biological specimens.
Logue et al. (1986) conducted a cohort study of male radiologists and pathologists
registered with the Radiation Registry of Physicians and the College of American Pathologists
(CAP) between 1962 and 1977. Although the main focus was on determining mortality in
radiologists from exposure to ionizing radiation, mortality was also ascertained for pathologists
alone. To derive SMRs, expected deaths were the sum of the products of person-years times
death rates for both cohorts during the follow-up period in white males only. However, there
were no exposure measurements, and the SMRs were not adjusted for calendar time. Of 5,585
members of the CAP, 496 had died by December 31, 1977. Although the SMR was 0.48 for
pathologists for cancer of the lymphatic and hematopoietic system, for the more specific
category of leukemia and aleukemia the SMR was 1.06 (neither was significant). For
radiologists, the SMRs were 0.78 and 1.55, respectively, also not significant. Cause of death
could not be determined for 8% of the deaths. Although age-adjusted rates for leukemia were
also calculated for each cohort, they were only used for comparison between the two separate
professional groups.
Hayes et al. (1990) conducted a proportionate mortality study of 3,649 deceased white
and 397 deceased nonwhite U.S. male embalmers and funeral directors who had died between
1975 and 1985, using records from local licensing boards, state funeral directors' associations in
32 states and the District of Columbia, the National Funeral Directors' Association, and state
offices of vital statistics (n = 894). Expected deaths by cause were derived from 5-year age- and
calendar-year-specific proportions of deaths among appropriate race groups from the U.S.
population. No measured exposure data were available. A PCMR would be derived by
excluding noncancer causes of death. Statistically significant excesses in hematopoietic and
lymphatic cancers were found in white (PMR 1.31 [95% CI: 1.06-1.59]; 100 observed) and
This document is a draft for review purposes only and does not constitute Agency policy.
4-87 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
nonwhite (PMR2.41 [95% CI: 1.35-3.97]; 15 observed) embalmers and funeral directors. The
combined PMR was 1.39 (95% CI: 1.15-1.63). The excess risk was higher for myeloid
leukemia (ML) (PMR 1.61 [95% CI: 1.02-2.41]; 23 observed) and for other unspecified
leukemias (PMR 2.08 [95% CI: 1.21-3.34]; 17 observed) in white males. The risks were
elevated in nonwhite males based on only a few cases (PMR 1.33 [95% CI: 1.10-1.60];
4 observed).
Matanoski (1991) conducted a study of 6,111 male pathologists for NIOSH. Members of
the cohorts were part of an earlier unpublished study. Twenty-nine thousand psychiatrists were
used as a comparison group. Both samples were selected from the membership rolls of
professional associations. A total of 3,787 pathologists died between 1940 and 1978. Women
were excluded from the analysis. Of the population of psychiatrists, 4,788 died by 1980. U.S.
age- and calendar-time-specific death rates from 1925 were used to develop SMRs. Separate
SMRs were based on psychiatrists' death rates. The risk of hematopoietic cancer (excluding
Hodgkin's disease) was elevated (SMR 1.25; 57 observed) based on U.S. white males. For
leukemia, the SMR was 1.35 (31 observed). The SMR for leukemia among psychiatrists was
0.83 (35 observed). Compared with leukemia in psychiatrists, the SMR for pathologists was
1.68 (95% CI: 1.14-2.38). The SMR for other lymphatic cancers was 1.53 (16 observed) and for
LHP cancer 1.22 (64 observed). Comparing the pathologists' death rates to those of psychiatrists
could be thought to have greater validity than if death rates for the U.S. population as a whole
had been used, because of shared socioeconomic circumstances and access to medical care
between the two professional groups. Differences in access to health care might have been
greater for subjects in the earlier part of the study, because improved diagnosis and medical care
for LHP cancers became more broadly available later in the study period. By using SMRs based
on U.S. death rates, which include those who do not have adequate access to medical care, the
difference between expected and observed deaths would be reduced. This is less likely to occur
when one professional group is compared with another professional group, assuming
psychiatrists and pathologists have equal access to care.
4.1.2.2.1.2. Industry worker cohort studies. This section discusses updated industrial worker
studies that show associations between LHP cancer and formaldehyde. The studies by Marsh et
al. (1994), Blair et al. (1986), and Acheson et al. (1984) and the later update by Gardner et al.
(1993) provide estimates of exposure to formaldehyde. The remaining studies generally rely
either on duration of exposure (number of years in the job) as a surrogate (Pinkerton et al., 2004)
or provide no exposure assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
4-88 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
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Marsh et al. (1994), in an early study of the Wallingford plant, which is also part of the
Hauptmann et al. (2004, 2003) and Blair et al. (1986) studies, found SMRs of 0.89 and 0.91,
based on U.S. and county death rates, respectively (25 observed deaths). The authors did not
further discuss this cancer site until after Hauptmann et al. (2003) was published. Blair et al.
(1986) reported on 4,396 deaths from all causes in the 10 formaldehyde-associated factories that
made up the NCI cohort of 26,561 workers employed before January 1, 1966. There was little
evidence of an association with LHP system cancer (SMR 0.91; 56 observed) in exposed white
men, who dominated the cohort.
Hauptmann et al. (2003) updated the cohort mortality study of Blair et al. (1986) that
consisted of predominantly the same (25,619) workers from 10 plants. The primary focus of this
analysis was cancer of the LHP system, including leukemia. The description and demographics
of the current study are the same as those reported by Blair et al. (1986) and Stewart et al.
(1986). In the current update, follow-up was extended through December 31, 1994. The
additional 15 years of follow-up increased the number of deaths from 4,349 to 8,486. Exposures
were not updated for the 4% of workers still in exposed jobs in 1980, but eliminating exposure
estimates for these workers did not change the results since exposures received after this date
were considered so low as to contribute little to the analysis by the authors.
Peak exposure categories were defined as nonexposed, low (0.1-1.9 ppm), medium (2.0-
3.9 ppm), and high (4.0 ppm or greater). Average intensity categories of exposure were defined
as nonexposed, low (0.1-0.4 ppm), medium (0.5 to <0.9 ppm), and high (>1.0 ppm). Cumulative
exposure was defined as nonexposed, low (0.1-1.4 ppm-years), medium (1.5-5.4 ppm-years),
and high (>5.5 ppm-years). Duration of exposure was defined as 0, 0.1-4.9 years, 5.0-
14.9 years, and >15 years. The median TWA exposure level was 0.45 ppm, range 0.01-
4.25 ppm. Only 2.6% of the workers had average exposure intensities of 2 ppm or higher, and
14.3% had peak exposures of 4 ppm or higher. A total of 3,201 workers had no exposure. The
median duration in formaldehyde-exposed jobs was 2 years. The median TWA intensity for
formaldehyde exposure was 0.5 ppm among exposed workers.
A Poisson regression model was stratified for calendar year, age, sex, race, and pay
category (salary/wage). A minimum latency period of 2 years between exposure and death from
a potentially exposure-related LHP cancer was assumed by the investigators to prevent the
inclusion of exposures not likely to contribute to the development of LHP cancer because of their
timing. Other lag times were evaluated that did not improve the regression models.
There were 2,099 cancer deaths. Hauptmann et al. (2003) reported that mortality from all
causes, all cancers, and LHP malignancies were significantly lower among the unexposed (SMRs
0.77, 0.65, and 0.62, respectively). Among the exposed, the SMRs were 0.95, 0.90, and 0.80,
This document is a draft for review purposes only and does not constitute Agency policy.
4-89 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
respectively, for mortality from all causes, all cancers, and LHP malignancies. These SMRs in
part reflect the HWE caused by using external U.S. population death rates as a referent.
Unexposed workers also may have differed from the exposed workers in other ways. The
remaining analyses used internal comparisons, avoiding the HWE. The referent group in this
analysis was the low exposure group rather than the unexposed group, because nonexposed
workers, who are primarily managers, secretaries, and other non-production personnel, were
considered likely to have different socioeconomic characteristics than workers in the production
areas.
Statistically significant positive associations were found for LHP malignancies and
leukemia, particularly myeloid, in certain higher exposure categories in comparison with
employees in the lowest exposure categories (Table 4-4). In the highest peak exposure level, the
RR for LHP malignancies was 1.87 (95% CI: 1.27-2.75; 64 observed) and for ML was 3.46
(95% CI: 1.27-9.43; 14 observed) compared with employees in the low exposure peak level. For
workers with high peak exposure levels, the RR for LHP malignancies was 1.71 (95% CI:
1.14-2.58; 49 observed) and 2.43 (95% CI: 0.81-7.25) for ML. The trend tests for slope were
highly statistically significant for both LHP malignancies (p < 0.002) and ML (p < 0.009).
Significant results for LHP cancers were also seen with the average intensity exposure
metric. RRs were 1.63 (p < 0.05) and 1.50 (p < 0.05) for the medium and high categories,
respectively. The risk of ML was also significantly increased (RR = 2.49) in the highest
exposure category. In contrast, Hauptmann et al. (2003) did not find statistically significant
associations of formaldehyde with LHP cancer, either by cumulative exposure or years of
duration. However, there were positive associations for leukemia (RR = 1.39) and ML (RR =
1.35) when exposure was 15 years or longer.
The authors concluded that formaldehyde may cause leukemia, particularly ML, in
humans. However, because results from other studies were inconsistent, they suggested caution
in drawing definite conclusions. A biological basis for the significant excess risk of LHP cancer
remains unclear. The authors pointed out several studies that indicate changes that are consistent
with chromosomal changes in formaldehyde-exposed persons, such as increased frequencies of
(MN (He et al., 1998; Kitaeva et al., 1996; Suruda et al., 1993), sister chromatid exchanges
(SCEs) (Shaham et al., 2002, 1997; Yager et al., 1986), chromosomal aberrations (CAs) (He et
al., 1998; Bauchinger and Schmid, 1985), and DNA-protein cross-links (DPXs) (Shaham et al.,
1997, 1996a) in peripheral lymphocytes of humans exposed to formaldehyde.
Hauptmann et al. (2003) identified 11 suspected carcinogens used at the plants:
antioxidants (unspecified), asbestos, carbon black, dyes and pigments, hexamethylenetetramine,
melamine, phenol, plasticizers, urea, wood dust, and benzene. Some workers were employed as
This document is a draft for review purposes only and does not constitute Agency policy.
4-90 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
chemists and laboratory technicians. The investigators did not find substantial changes in the
risk estimates after adjusting for exposure to these substances or for working as a chemist or
laboratory technician. They also eliminated the 586 benzene-exposed persons from their
analysis and found similar results (benzene is a known human leukemogen). Smoking was not
likely to explain an increased risk of leukemia in this cohort, because no increase was seen for
smoking-related diseases, including lung cancer. The cohort consisted predominantly of males
(88%). Strengths include the fact that the cohort was large and there was a long period of
follow-up that was 96.6% successful. Internal analyses eliminated the HWE. One potential
limitation that could lead to an underestimate of risks is the 3.4% or 866 lost to follow-up.
The study by Hauptmann et al. (2003) has been criticized extensively by several experts
representing the formaldehyde industry (Tarone and McLaughlin, 2005; Casanova et al., 2004;
Cole and Axten, 2004; Collins, 2004; Collins and Lineker, 2004). Most of the same criticisms
have been repeated in other critiques by the above-mentioned authors and have been addressed in
the discussions concerning the details of the methodology. However, a few new issues have
arisen from these critiques, as follows. One issue pertains to a concern that person-years at risk
of death may have been assigned wrongly to the highest "peak" category of exposure for the
duration of the study period. For example, there is inconsistency in the fact that only 4% of the
original cohort (Blair et al., 1986) had average exposures equaling or exceeding 2 ppm yet 45%
of the person-years were assigned to the peak exposure category. Average exposures are time-
weighted exposures that can have brief excursions over 4 ppm and still average 2 ppm or less.
Only for the peak exposure surrogate were person-year values assigned to the peak category
following the exposure, because it is a test for the possibility that biological changes could have
been initiated from that brief high exposure that might increase the risk of cancer. If these
genetic changes are irrevocable, then the risk of cancer could be increased and subsequent person
x years should be assigned to that higher risk category.
According to Casanova et al. (2004), the assignment of peak exposures in the Hauptmann
et al. (2003) study was questionable because they were based on professional judgement.
However, there are adequate grounds for hypothesizing that the assignment of peak exposure
was completed before determination of vital status and cause of death. It is always possible that
some subjects may be subject to misclassification. Hauptmann et al. (2003) chose this metric
partly because it more closely resembled the exposure that embalmers and pathologists may have
received from formaldehyde. This same criticism could be said about the Coggon et al. (2003)
study as well.
Hauptmann et al. (2003) have also been criticized because the metric "cumulative
exposure" was not significant and did not show a trend. No adequate explanation has been given
This document is a draft for review purposes only and does not constitute Agency policy.
4-91 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
25
26
27
28
29
30
31
32
33
34
35
by Hauptmann et al. (2003) except that it is possible that this metric is not as sensitive for this
agent. Duration of exposure was only weakly associated with a trend of increasing risk. After
15 years' duration, there appeared to be a slight increase in ML based on 10 cases (RR =1.35
[95% CI: 0.56-3.24]). For all leukemia the risk in workers who were exposed 15 or more years
was somewhat higher (RR = 1.39 [95% CI: 0.78-2.49]) based on 22 cases.
Another of the criticisms from these authors discussed the lack of a biologically plausible
explanation for how leukemia could result from exposure to formaldehyde when there appears to
be no recognizable indication of the presence of formaldehyde in excessive quantities in the
blood of animals or any associated metabolites in experimental research animals. Hauptmann et
al. (2004b) responded that there is evidence that genotoxic effects can be detected in vivo in the
bone marrow of rats and in human peripheral lymphocytes.
Stayner et al. (1988) conducted a cohort study of 11,030 workers (82% female) followed
from 1955 or the beginning date of exposure through 1982 in three garment factories. Personnel
records from three garment manufacturing facilities, one in Pennsylvania and two in Georgia,
were used to assemble a cohort of workers who attained a minimum of 3 months of exposure
after the introduction of formaldehyde into these facilities. Formaldehyde resins were used to
treat fabrics, beginning in 1955 and 1959. Although formaldehyde levels were available on a
subset of the employees from monitoring data available from surveys completed in 1981 and
1984, they were not used in this analysis. Instead, the results were stratified by duration and
latency. SMRs were based on U.S. population mortality rates. Based on six cases, the SMRs for
leukemia were 2.43 and 3.81 among workers with 20 or more years since first exposure or at
least 10 years of exposure, respectively. In their conclusions, the authors suggested that,
although the numbers of deaths from LHP cancers were small, the risks were related to duration
and latency.
Pinkerton et al. (2004) updated the Stayner et al. (1988) study by adding 16 years of
follow-up. No new exposure information was added. The mean TWA exposure in 1981-1984
for the three plants was 0.15 ppm. No additional information regarding earlier industrial hygiene
data was available, although the authors stated that the levels of exposure to formaldehyde were
greater in the years before 1980. Stayner et al. (1988) cited independent studies of exposure
levels in similar garment factories in the 1960s that seemed to indicate that the formaldehyde
levels during that period ranged from 0.9 to 2.7 ppm (Blejer and Miller, 1966) in one garment
manufacturing area. Another report (Shipkovitz, 1966) of 10-minute personal exposure samples
indicated a range from 0.3 to 2.7 ppm in eight garment plants. In another study (Ahmad and
Whitson, 1973), the levels ranged from 2 to 10 ppm. Goldstein (1973) calculated that
concentrations in the cutting rooms of garment plants dropped from 10 ppm in 1968 to less than
This document is a draft for review purposes only and does not constitute Agency policy.
4-92 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
2 ppm in 1973 because of an improvement in the resin treating process. The authors assumed
that exposure ceased in 1981 and 1983. This produced an underestimate of exposure based on
duration of employment for about 11% of the cohort who were still actively employed after those
dates. Stayner et al. (1988) speculated that the risks of cancer of the buccal cavity, leukemia, and
other LHP neoplasia may have been due to exposure to the highest potential formaldehyde levels
in the industry between 1955 and 1962, because the resin used to treat permanent press fabrics
still contained a relatively large amount of formaldehyde.
The SMRs were derived from age-, race-, and calendar-time-adjusted U.S. mortality
rates. The analysis was repeated using Georgia or Pennsylvania mortality rates. In addition to
the primary analysis of the underlying cause of death, the analysis used all causes listed on the
death certificates to evaluate multiple cause mortality. As a referent for this, the analysis relied
on multiple cause death rates available since 1960 from the National Death Index maintained by
the U.S. Centers for Disease Control and Prevention (CDC).
Altogether, 608 cancer deaths were observed. The SMR for all cancer was 0.89 (95% CI:
0.82-0.97). The overall SMR for leukemia was 1.09 (24 deaths) and 1.44 (15 deaths) for ML.
After 10 years of exposure, the risk for ML was 2.19. Exposure prior to 1963 was associated
with a risk of 1.61. Among garment workers followed for 20 or more years from initial
exposure, the SMR was significantly elevated for ML (1.91 ;p< 0 .05; 13 deaths), as was the
SMR for multiple cause leukemia (1.92 [95% CI: 1.08-3.17]; 15 deaths) in the subgroup with
10 or more years of exposure to formaldehyde and who were followed for 20 or more years after
first exposure. The multiple cause mortality for ML for this subgroup of workers was also
significant (SMR 2.55 [95% CI: 1.10-5.03]; 8 deaths).
The study by Stayner et al. (1988) has only limited power to detect excess risks of rare
cancers, such as NPC and nasal cancer (13 and 16%, respectively). Limitations to the
interpretations of the findings include a lack of any monitoring data before 1981, particularly
during the critical time period 1955 to 1962, and lack of personal exposure estimates for any
members of the cohort. The possibility exists that misclassification may still be present because
the intensity of exposure to formaldehyde decreased as improvements were made in the resin
systems used to treat fabrics (e.g., a person who worked 5 years beginning in 1955 might have
been subject to greater exposure than a person who worked 5 years beginning in 1993).
However, workers from the 1950s and 1990s were both placed in the same category of having
worked fewer than 10 years. The median duration of exposure was 3.3 years. Work histories
were not updated in the follow-up study; however, the low or background exposure levels that
probably existed after 1981 were not likely to contribute substantially to the risk of cancer. The
use of mortality data to estimate risk, when the case fatality rate was less than 100% for most
This document is a draft for review purposes only and does not constitute Agency policy.
4-93 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
cancer sites evaluated, could potentially produce an underestimate of the actual risk. Despite
these limitations, this study provides additional evidence of an association between leukemia,
especially ML, and formaldehyde in comparison with the general population.
Gardner et al. (1993) reported that the risk of leukemia was not statistically significant
(SMR 0.9) based on 15 deaths among workers employed before 1965. Only four leukemia
deaths were observed after 1964 through 1989, producing an SMR of 0.9.
When Coggon et al. (2003) updated the above cohort study of 14,014 men first employed
before 1965 in six factories by adding 11 additional years of follow-up (ending December 31,
2000), no increase in the risk of leukemia or related cancers of the hematopoietic system was
reported, either in the entire cohort (SMR 0.91; 31 observed) or in the group with the highest
formaldehyde exposure (>2 ppm) (SMR 0.71; 8 observed). Similar results were obtained for
Hodgkin's disease, non-Hodgkin's lymphoma, and multiple myeloma. No other cancers of the
hematopoietic system were evaluated, and no additional analyses were performed to assess a
possible leukemia risk. However, the main finding from this study was a marked association of
lung cancer with formaldehyde (discussed in the lung cancer section). This study's main focus
was respiratory disease, lung cancer, and stomach cancer, not LHP cancers. For cancers of the
LHP system, there was neither latency evaluation nor internal comparisons. The HWE is also
potentially a problem.
Andjelkovich et al. (1995) studied a cohort of 3,929 male iron foundry workers
potentially exposed to formaldehyde between January 1, 1960, and December 31, 1989, in which
127 cancer deaths had occurred during the observation period. An industrial hygienist, after
reviewing work histories, categorized formaldehyde exposure into four levels corresponding to
the approximate midpoint of the ranges: none, low (0.05 ppm), medium (0.55 ppm), and high
(1.5 ppm) for exposure to formaldehyde. Boundaries of these exposure categories were not
given. The authors warned that the assignment of exposure levels was not perfect because
"subjective judgment had to be applied in many instances." SMRs were based on U.S. male
mortality rates, but actual ranges were not specified. The authors also compared the exposed to
2,032 nonexposed workers from the same company. The population-based SMR for
hematopoietic cancer in the exposed population was 0.59 (based on seven observed deaths). For
leukemia the SMR was 0.43, based on two deaths. There were no other analyses for leukemia or
LHP cancers in this study. Because of the uncertainty about workers' true formaldehyde
exposure, there was no analysis by level of exposure, duration, or latency. There were also very
few LHP cancers in the cohort. Thus, these results neither support nor refute an association of
formaldehyde exposure with LHP cancers. The main focus of this paper was on lung cancer risk.
This document is a draft for review purposes only and does not constitute Agency policy.
4-94 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
Bertazzi et al. (1989, 1986), in a cohort mortality study, followed 1,330 male workers
from 1959 through 1986 at a formaldehyde resin plant in Italy. The workers had to have been
employed for at least 30 days at the plant sometime between 1959 and 1980 to be included in the
study. Their mortality was compared with national and local rates adjusted for age and calendar
time period. No individual exposure estimates were available, but mean levels were estimated to
be between 0.2 and 3.8 mg/m3 (0.16 and 3.1 ppm) during the period 1974-1979. The authors
found an SMR of 2.01 (five deaths observed) for cancer of the lymphatic and hematopoietic
system. The study's limitations included incomplete work histories, small numbers of deaths,
and a follow-up period that may not have been sufficient to allow for a latency period for the
development of LHP cancers. As before, the results neither support nor refute an association of
formaldehyde exposure with LHP cancers.
Edling et al. (1987) reported on the incidence of disease in a cohort of 521 blue collar
Swedish workers in plants where abrasives bound with formaldehyde resins were manufactured.
Formaldehyde levels ranged from 0.1 to 1.0 mg/m3 (0.08-0.8 ppm). The workers in the cohort
were employed between 1955 and 1983, and incidence rates were calculated from 1958 through
1981. There were only 24 total cancer cases (28.5 expected) of which 2 (1.0 expected) were
lymphomas and 2 (0.5 expected) were multiple myelomas. Expected cases were determined
through the Swedish National Cancer Register. No other LHP cancers were observed. This
study lacked the power to detect any significant associations between LHP cancer and exposure
to formaldehyde.
Dell and Teta (1995) conducted a cohort mortality study of 5,932 male employees of a
New Jersey plastics manufacturing, research, and development facility. The workers, who had
been employed during the period 1946-1967, were followed-up for an average of 32 years.
SMRs were based on U.S. and New Jersey mortality rates. Hourly workers (n = 3,853) were
analyzed separately from the 2,079 salaried employees. Although no excess risk was evident for
hematopoietic cancer in hourly workers (SMR 0.93; 28 observed), there was an SMR of 1.69
(95% CI: 1.07-2.53; 23 observed) among salaried workers. This association was further
narrowed to mainly research and development workers (eight leukemia deaths observed with
three expected, for an SMR of 2.67). No common exposure was found when work history
records were examined. The decedents were mostly associated with process development in two
research pilot plants, where chemical engineers, lab technicians, and plant operators executed
small-scale product development. Although notebooks referred to benzene and toluene solvents,
no definite connection was made with formaldehyde or any of the solvents. No ambient air
measurements of formaldehyde were available. The findings cannot be assumed to be due to
formaldehyde exposure because of the presence of other potential leukemogens.
This document is a draft for review purposes only and does not constitute Agency policy.
4-95 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Blair et al. (1993) conducted a study that evaluated the risk of non-Hodgkin's lymphoma
from exposure to formaldehyde. This was a population-based, case-control, interview-based
study of 1,867 white males of whom 622 cases had the disease and 1,245 were controls.
Subjects had lived in Iowa and Minnesota between 1980 and 1983. This study was exploratory
and designed to find associations with any environmental exposures and non-Hodgkin's
lymphoma. Subjects or next of kin were interviewed to determine what exposures the cases and
controls may have received based on agricultural exposures, work histories, medical conditions,
and family history. Extra effort was made to collect information about occupation, industrial
exposures, and other selected exposures. The analysis revealed an OR of 1.2 for exposure to
formaldehyde. Similar associations were found for metals and other substances in the study.
This study, because it did not select cases and controls from a population with possible
formaldehyde exposure, could not detect specific relationships between formaldehyde and
non-Hodgkin's lymphoma.
4.1.2.2.1.3. Summary of non-respiratory tract cancers. The Hauptmann et al. (2003) study
appears to provide the strongest evidence of an association for ML in particular. Statistically
significant positive associations were found for LHP malignancies and leukemia, particularly
ML, in certain higher exposure categories in comparison with employees in the lowest exposure
categories. In the highest peak exposure level, the RR for LHP malignancies was 1.87 (95% CI:
1.27-2.75; 64 observed) and for ML was 3.46 (95% CI: 1.27-9.43; 14 observed) compared with
employees in the low-exposure peak level. For workers with high-peak exposure levels, the RR
for LHP malignancies was 1.71 (95% CI: 1.14-2.58; 49 observed) and 2.43 (95% CI: 0.81-7.25)
for ML. The trend tests for slope were highly statistically significant for both LHP malignancies
(p < 0.002) and ML (p < 0.009). Significant results for LHP cancers were also seen with the
average intensity exposure metric. RRs were 1.63 (p < 0.05) and 1.50 (p < 0.05) for the medium
and high categories, respectively. The risk of ML was also significantly increased (RR = 2.49)
in the highest exposure category. In contrast, results showed no associations of formaldehyde
with LHP cancer, either by cumulative exposure or years of duration.
Additional support linking LHP cancer and formaldehyde comes from a study of garment
workers (Pinkerton et al., 2004) and studies of pathologists and other medical workers exposed
to formaldehyde (Matanoski, 1991; Blair et al., 1990; Hayes et al., 1990; Stroup et al., 1986;
Walrath and Fraumeni, 1984, 1983; Harrington and Shannon, 1975). Hayes et al. (1990) and
Stroup et al. (1986) also reported significant excess risks of ML.
Several reports have challenged the association between LHP cancer and formaldehyde
(Casanova et al., 2004; Cole and Axten, 2004; Collins, 2004; Collins and Lineker, 2004; Coggon
This document is a draft for review purposes only and does not constitute Agency policy.
4-96 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
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
et al., 2003; Casanova and Heck, 1987; Heck et al., 1985). These papers argue that a biological
explanation for the excess risk of LHP cancer or leukemia remains unclear. Absent a plausible
MOA by which formaldehyde could cause these cancers, many investigators have been unable to
accept the reported increased risks identified in the epidemiologic literature. Some researchers
have argued against the biological plausibility of formaldehyde-induced lymphoreticular cancers
based solely on the assumption that formaldehyde as a reactive gas does not penetrate past the
POE. This argument is relevant to diseases for which transformation of stem cells in the bone
marrow is essential. However, cancers that arise from more mature cells present outside of the
bone marrow compartment cannot be dismissed with this argument. Although often grouped for
analysis, the lymphohemoreticular system cancers represent many distinct malignancies that may
arise from discrete cell types in different tissues throughout the body. For example, acute
lymphocytic leukemia (ALL) is believed to arise from the transformation of a lymphoid stem cell
in the bone marrow, resulting in a blood-borne leukemia of immature cells of the lymphoid cell
line. However, if transformation occurs in a mature lymphocyte (e.g., post-germinal center B
cell), a chronic lymphocytic leukemia (CLL) results. Although etiologically different, these
cases would both be lymphocytic leukemia. When considering biological plausibility of an
exogenous agent increasing the incidence of ALL, bone marrow toxicity would be expected.
However, when considering the biological plausibility of CLL, bone marrow toxicity would not
be essential. Mutation or epigenetic changes attained in the mature cell may be passed on to
daughter cells during response to antigen and eventually lead to transformation. So the
etiologies of these two leukemias need not be similar. In contrast, a non-Hodgkin's lymphoma
results from transformation of a mature B or T cell resulting in a solid tumor. The etiology of
this cancer is actually similar to CLL. A recent reclassification of lymphoid malignancies by the
WHO designates adult B-cell leukemias and lymphomas as the same disease, with ALL as a
separate disease. Therefore, mortality analysis by ICD code and the standard groupings of those
codes does not reflect the biology of the cancers.
Considering the whole class of LHP cancers, there is a range of biological plausibility for
an agent whose primary action is at the POE. Acute leukemias (ALL and acute myelogenous
leukemia [AML]), believed to arise from transformation of stem cells in the bone marrow, are
less plausible, although trafficking of stem cells to different tissues would be an alternative
etiology for exogenous compounds acting at the POE. In contrast CLL, lymphomas, multiple
myelomas (from plasma B cells), and unspecified cancers may involve an etiology in peripheral
tissues to include cells, cell aggregates, germinal centers, and lymph nodes. An association of
these cancers to an exogenous agent acting at the POE is biologically plausible.
This document is a draft for review purposes only and does not constitute Agency policy.
4-97 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
§•
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
o
5 s
to
o
^ Co'
to &-
§ ^
>S
>S
TO
TO'
*
Harrington and
Shannon (1975)
Cohort mortality study of 2,079
pathologists and laboratory
technicians from the Royal
College of Pathologists and the
Pathological Society of Great
Britain from 1955-1973. The
comparison population came
from national mortality data.
Presumed exposure to
formaldehyde tissue
fixative.
Pathologists
All cause mortality
LHP cancers
Hodgkin's disease
Leukemia
Technicians
All cause mortality
LHP cancers
Hodgkin's disease
Leukemia
SMR
SMR
SMR
SMR
SMR
SMR
SMR
SMR
0.60
2.0
1.4
0.6
0.67
0.5
0.5
p < 0.01
(8)
(1)
(1)
(3)
(0)
(1)
Harrington and
Oakes (1984)
Cohort mortality study of 2,720
pathologists from the Royal
College of Pathologists and the
Pathological Society of Great
Britain from 1974-1980. Vital
status obtained from the census, a
national health registry, and other
sources. SMRs developed from
the English, Scottish, Irish, and
Welsh populations.
Presumed exposure to
formaldehyde tissue
fixative.
All causes
Men
Women
Leukemia
Men
Women
Other LHP cancers
Men
Women
SMR
SMR
SMR
SMR
SMR
SMR
0.56
0.99
0.91
9.26
0.53
(90% CI: 0.05-4.29)
(90% CI: 0.47-43.9)
(90% CI: 0.03-2.54)
(1)
(1)
(1)
(0)
Halletal. (1991)
Cohort mortality study of 4,512
pathologists from the Royal
College of Pathologists and the
Pathological Society of Great
Britain from 1974-1987. Vital
status obtained from the census, a
national health registry, and other
sources. SMRs developed from
the English and Welsh
populations.
Presumed exposure to
formaldehyde tissue
fixative.
All cause mortality
Men
Women
Hodgkin's disease
All cancers
Leukemia
SMR
SMR
SMR
SMR
SMR
0.43
0.65
1.21
1.44
1.52
(95% CI: 0.03-6.71)
(95% CI: 0.69-2.63)
(95% CI: 0.41-3.89)
(176)
(18)
(1)
(10)
(4)
vo
00
Levine et al. (1984)
Cohort mortality study of 1,477
Presumed exposure to
All LHP cancers
SMR
1.24
(8)
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
male Ontario undertakers first
licensed 1928-1957, followed
from 1950 to 1977. SMRs
developed from Ontario
mortality rates.
formaldehyde tissue
fixative.
Leukemia
SMR
1.60
(4)
Stroup et al. (1986)
Cohort mortality study of 2,317
white male members of the
American Association of
Anatomists from 1888 to 1969
who died 1925-1979. SMRs
developed using U.S. population
mortality rates.
Presumed exposure to
formaldehyde tissue
fixative.
All cause mortality
All LHP cancers
Lymphosarcoma and
reticulosarcoma
Hodgkin's disease
Leukemia
Other lymphatic
SMR
SMR
SMR
SMR
SMR
SMR
0.65
1.2
0.7
1.5
2.0
(95% CI: 0.60-0.70)
(95% CI: 0.7-2.0)
(95% CI: 0.1-2.5)
(95% CI: 0.7-2.7)
(95% CI: 0.7-1.4)
(738)
(18)
(2)
(0)
(10)
(6)
Logue et al. (1986)
Cohort mortality study of 4,485
pathologists who were members
of the College of American
Pathologists, 1962-1972,
followed for mortality through
1977. SMRs developed from
U.S. population mortality rates.
Presumed exposure to
formaldehyde tissue
fixative.
LHP cancer other than
leukemia
Leukemia
SMR
SMR
0.48
1.06
(NR)
(NR)
Matanoski (1991)
Cohort mortality study of 6,111
male pathologists from
membership rolls of the
American Medical Association
1912-1950. Mortality was
followed through 1978. SMRs
developed from U.S. population
white male mortality rates.
Presumed exposure to
formaldehyde tissue
fixative.
All cancer
All LHP cancers
Lymphosarcoma and
reticulosarcoma
Hodgkin's disease
Leukemia
Other lymphatic
SMR
SMR
SMR
SMR
SMR
SMR
0.78
1.25
1.31
0.36
1.35
1.54
(95% CI: 0.71-0.85)
(95% CI: 0.95-1.62)
(95% CI: 0.66-2.35)
(95% CI: 0.04-1.31)
(95% CI: 0.92-1.92)
(95% CI: 0.82-2.63)
(508)
(57)
(11)
(2)
(31)
(13)
Hauptmann et al.
(2003)
Retrospective cohort mortality
study of 25,619 workers
employed at 10 formaldehyde
plants in the U.S. followed from
either the plant start-up or first
employment through 1994.
Exposure estimates
based on job titles,
tasks, visits to plants by
study industrial
hygienists, and
monitoring data through
All LHP cancers
Exposed
Unexposed
Peak exposure (dditi)
o
SMR
SMR
RR
0.80
0.62
1.08
(95% CI: 0.69-0.94)
(95% CI: 0.39-1.00)
(95% CI: 0.60-1.94)
(161)
(17)
(17)
o
2 »
5 s
to
o
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
0.1-1.9
RR
1.00
Reference value
(48)
2.0 to <4.0
RR
1.71
(95% CI: 1.14-2.58)
(49)
4.0 or greater
RR
1.87
(95% CI: 1.27-2.75)
(64)
Trend p = 0.002
Averase exposure (ppm)
0
RR
0.91
(95% CI: 0.52-1.59)
(17)
0.1-0.4
RR
1.00
Reference value
(81)
0.5 to <1.0
RR
1.63
(95% CI: 1.11-2.37)
(42)
1.0 or greater
RR
1.50
(95% CI: 1.01-2.24)
(38)
Trend p = 0.050
Cumulative exposure
DDin-vcars)
0
RR
0.74
(95% CI: 0.42-1.30)
(17)
0.1-1.4
RR
1.00
Reference value
(94)
1.5 to 5.4
RR
0.79
(95% CI: 0.52-1.21)
(29)
5.5 or greater
RR
1.03
(95% CI: 0.70-1.52)
(38)
o
5 s
a, Co'
TO Sj-
§ ^
>S
>S
TO
TO'
*
SMRs calculated using sex-, age-
, race-, and calendar-year-
specific U.S. mortality rates.
RRs estimated using Poisson
regression stratified by calendar
year, age, sex, and race; adjusted
for pay category.
1980. Peak exposure
defined as short-term
excursions exceeding
the 8-hour TWA
formaldehyde intensity
and knowledge of job
tasks. Exposures to 11
other compounds were
identified. Workers
contributed pre-
exposure person-time to
nonexposed category.
Poisson regression
models used a 2-year
lag to account for tumor
latency.
Trendp = 0.157
Leukemia |
Peak exposure (ppin)
0
0.1-1.9
2.0 to <4.0
4.0 or greater
RR
RR
RR
RR
78
00
04
46
(95% CI: 0.25-2.43)
Reference value
(95% CI: 1.04-4.01)
(95% CI: 1.31-4.62)
Trendp = 0.001
Average exposure (ppm)
0 ppm RR
0.1-0.4 RR
0.5 to <1.0 RR
1.0 or greater RR
56
00
52
68
(95% CI: 0.19-1.66)
Reference value
(95% CI: 083-2.79)
(95% CI: 0.91-3.08)
(4)
(16)
(20)
(29)
(4)
(32)
(16)
(17)
o
o
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Results, statistical significance (number of observed deaths for cohort
Reference
Study design
Exposure assessment
study)
Trend p = 0.193
Cumulative exposure
(DDm-vears)
0
RR
0.48
(95% CI: 0.16-1.42)
(4)
0.1-1.4
RR
1.00
Reference value
(35)
1.5-5.4
RR
0.90
(95% CI: 0.47-1.73)
(13)
5.5 or greater
RR
1.14
(95% CI: 0.63-2.07)
(17)
Trend p = 0.183
Hodgkin's disease
Peak exposure (ddiii)
0
RR
0.51
(95% CI: 0.06-4.52)
(1)
0.1-1.9
RR
1.00
Reference value
(5)
2.0 to <4.0
RR
3.45
(95% CI: 0.98-12.2)
(7)
4.0 or greater
RR
3.35
(95% CI: 0.97-11.6)
(8)
Trendp = 0.014
Averase exposure (ddiii)
0
RR
0.46
(95% CI: 0.05-3.93)
(1)
0.1-0.4
RR
1.00
Reference value
(7)
0.5 to <1.0
RR
4.70
(95% CI: 1.61-13.8)
(8)
1.0 or greater
RR
3.12
(95% CI: 0.91-10.7)
(5)
Trendp = 0.022
Cumulative (DDm-vears)
0
RR
0.29
(95% CI: 0.04-2.34)
(1)
0.1-1.4
RR
1.00
Reference value
(12)
1.5-5.4
RR
1.35
(95% CI: 0.45-3.99)
(5)
5.5 or greater
RR
1.17
(95% CI: 0.31-4.46)
(3)
Trendp = 0.037
ML
Peak exposure (ddiii)
0
RR
0.67
(95% CI: 0.12-3.61)
(2)
0.1 to 1.9
RR
1.00
Reference value
(6)
2.0 to <4.0
RR
2.43
(95% CI: 0.81-7.25)
(8)
4.0 or greater
RR
3.46
(95% CI: 1.27-9.43)
(14)
o
2 »
5 s
to
o
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
o
5 s
to
o
^ Co'
& 55
to Sj-
§ ^
>s
>S
TO
TO'
*
Trendp = 0.003
Averase exposure room)
0
RR
0.41
(95% CI: 0.08-1.95)
(2)
0.1 to 0.4
RR
1.00
Reference value
(14)
0.5 to <1.0
RR
1.15
(95% CI: 0.41-3.23)
(5)
1.0 or greater
RR
P =
2.49
= 0.086
(95% CI: 1.03-6.03)
(9)
Cumulative Orom-vears)
0
RR
0.32
(95% CI: 0.07-1.51)
(2)
0.1-1.4
RR
1.00
Reference value
(17)
1.5-5.4
RR
0.57
(95% CI: 0.19-1.73)
(4)
5.5 or greater
RR
1.02
(95% CI: 0.40-2.55)
(7)
Trend p = 0.123
All LHP cancers
SMR
0.97
(95% CI: 0.74-1.26)
(59)
Lymphosarcoma and
SMR
0.85
(95% CI: 0.28-1.99)
(5)
reticulosarcoma
Hodgkin's disease
SMR
0.55
(95% CI: 0.07-1.98)
(2)
Other lymphatic
SMR
0.97
(95% CI: 0.64-1.40)
(28)
Leukemia
SMR
1.09
(95% CI: 0.70-1.62)
(24)
Mortality since 1960
Lymphocytic leukemia
SMR
0.60
(95% CI: 0.12-1.75)
(3)
ML
SMR
1.44
(95% CI: 0.80-2.37)
(15)
10+ years of
SMR
2.19
NS
(8)
exposure
20+ years since 1st
SMR
1.91
p > 0.05
(13)
exposure
Multiple cause leukemia
10+ years of
SMR
1.92
(95% CI: 1.08-3.17)
(15)
exposure and 20+
years since 1st
exposure
Pinkerton et al.
(2004)
Cohort mortality study of 11,098
workers in 3 garment plants
exposed >3 months after
formaldehyde was introduced.
Women comprised 81.7% of the
cohort. Vital status was followed
through 1998. SMRs were
calculated by using sex-, age-,
race-, and calendar-year-specific
U.S. mortality rates. Multiple
cause SMRs were derived from
all contributing causes from
death certificates.
Data for 549 randomly
selected employees in 5
departments in 1981 and
1984 used to estimate
overall exposure levels.
Levels presumed to be
0.09-0.20 ppm.
o
to
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
§•
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
o
5 s
to
o
^ Co'
to &-
§ ^
Multiple cause ML
20+ years since 1st
exposure
10+ years of
exposure and 20+
years since 1st
exposure
SMR
SMR
2.02
2.55
(95% CI: 1.
(95% CI: 1.
13-3.34)
10-5.03)
(15)
(8)
>s
>s
TO
JS*
*
Coggon et al. (2003)
Cohort mortality study of 14,014
men employed in 6 factories of
the chemical industry in Great
Britain from periods during
which formaldehyde was
produced. Cohort mortality
followed from 1941 through
2000. SMRs based on English
and Welsh age- and calendar-
year-specific mortality rates.
Exposure assessment
based on data abstracted
from company records.
Jobs categorized as
background, low,
moderate, high, or
unknown levels.
Non-Hodgkin's lymphoma
Overall
High exposure
Leukemia
Overall
High exposure
Multiple myeloma
Overall
High exposure
SMR
SMR
SMR
SMR
SMR
SMR
0.98
0.89
0.91
0.71
0.86
1.18
(95% CI: 0.
(95% CI: 0.
(95% CI: 0.
(95% CI: 0.
(95% CI: 0.
(95% CI: 0.
67-1.39)
41-1.70)
62-1.29)
31-1.39)
48-1.41)
48-2.44)
(31)
(9)
(31)
(8)
(15)
(7)
Andjelkovich et al.
(1995)
Cohort mortality study of 3,929
automotive industry iron foundry
workers exposed from 1960-
1987 and followed through 1989.
SMRs calculated using sex-, age-
, race-, and calendar-year-
specific U.S. mortality rates.
Exposure assessment
based on review of
work histories by an
industrial hygienist.
All LHP cancers
Leukemia
SMR
SMR
0.59
0.43
(95% CI: 0.23-1.21)
(95% CI: 0.05-1.57)
(7)
(2)
Bertazzi et al. (1986)
Cohort mortality study of 1,330
male workers in an Italian resin
plant. Subjects were employed
any time between 1959 and 1980
for at least 30 days. Vital status
followed through 1986. SMRs
calculated using sex-, age-, race-,
and calendar-year-specific
national and local mortality rates.
Exposure assessment
based on reconstruction
of work history.
All LHP cancers
SMR
2.01
(5)
o
LtJ
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
Edling et al. (1987)
Cohort mortality and incidence
study of 521 Swedish workers in
an abrasive production plant with
at least 5 years of employment
between 1955 and 1983,
followed through 1983.
Exposure level of
1-5 |ig/m3.
Lymphoma
Multiple myeloma
SMR
SMR
2.0
4.0
(95% CI: 0.2-7.2)
(95% CI: 0.5-14)
(2)
(2)
Dell and Teta (1995)
Cohort mortality study of 5,932
male employees of a New Jersey
plastics manufacturing, research
and development facility.
Examination of work
histories to identify jobs
where formaldehyde
was involved.
All
Lei
LHP cancers
Hourly workers
Salaried workers
ikemia
Hourly workers
Salaried workers
SMR
SMR
SMR
SMR
0.93
1.69
0.98
1.98
(28)
(23)
(12)
(11)
Walrath and
Fraumeni (1983)
Proportionate mortality cohort
study of 1,132 white male
embalmers licensed to practice
between 1902 and 1980 in New
York who died between 1925
and 1980 identified from
registration files. Deaths were
compared with age-, race-, and
calendar-year-expected numbers
of deaths from the U.S.
population.
No direct
measurements.
Presumed exposure to
formaldehyde tissue
fixative.
All LHP cancers
| PMR | 1.15 | | (21)
Lymphosarcoma and reticulosarcoma
| PMR | 1.08 | | (4)
Hodgkin's disease
| PMR | 1.0 | | (2)
Other lymphatic lymphoma
| PMR | 1.18 | | (5)
Leukemia
| PMR | 1.32 | | (10)
Walrath and
Fraumeni (1984)
Proportionate mortality cohort
study of 1,007 white male
embalmers from the California
Bureau of Funeral Directing and
Embalming who died between
1925 and 1980. Deaths were
compared with age- and
calendar-year-expected numbers
No direct
measurements.
Presumed exposure to
formaldehyde tissue
fixative.
All LHP cancers
| PMR | 1.22 | | (19)
Lymphosarcoma and reticulosarcoma
| PMR | 0.97 | | (3)
Hodgkin's disease
1 pmr|- I I (0)
Other lymphatic lymphoma
| PMR | 1.33 | | (4)
o
2 »
5 s
a, Co'
§- a
TO Sj-
§ ^
>{
>{
JS*
*
4^
O
4^
-------
K
s
TO
>3
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
&
to
s
Results, statistical significance (number of observed deaths for cohort
Reference
Study design
Exposure assessment
study)
of deaths from the U.S.
Leukemia
population.
Licensed <20 years
PMR
PMR
1.75
1.24
p < 0.05
(12)
(4)
Licensed >20 years
PMR
2.21
p < 0.05
(8)
Hayes et al. (1990)
Proportionate mortality cohort
No direct
All LHP cancers
study of 3,649 deceased white
measurements.
PMR
1.39
(95% CI: 1.15-1.63)
(115)
and 397 deceased nonwhite U.S.
Presumed exposure to
Race
male embalmers and funeral
formaldehyde tissue
directors, derived from licensing
fixative.
White
PMR
1.31
(95% CI: 1.06-1.59)
(100)
boards and funeral director
Nonwhite
PMR
2.41
(95% CI: 1.35-3.97)
(15)
associations in the 32 states and
Occupation on death certificate
the District of Columbia.
Embalmer
PMR
1.23
(95% CI: 0.78-1.85)
(23)
Occupation was confirmed on
Funeral director
PMR
1.56
(95% CI: 1.23-1.94)
(78)
death certificate. Deaths were
Other
PMR
1.30
(95% CI: 0.67-2.28)
(12)
compared with age- and
Ase at death
<60
calendar-year-expected numbers
of deaths from the U.S.
PMR
1.35
(95% CI: 0.88-1.98)
(26)
population.
60-74
PMR
1.72
(95% CI: 1.33-2.19)
(66)
>75
PMR
1.16
(95% CI: 0.74-1.74)
(23)
Hodgkin's disease
PMR
0.72
(95% CI: 0.15-2.10)
(3)
Non-Hodgkin's lymphoma
PMR
1.26
(95% CI: 0.87-1.76)
(34)
Lymphosarcoma and reticulosarcoma
PMR
1.12
(95% CI: 0.58-1.96)
(12)
Multiple myeloma
PMR
1.37
(95% CI: 0.84-2.12)
(20)
Other lymphatic lymphoma
PMR
1.35
(95% CI: 0.85-2.01)
(22)
Lymphatic leukemia
PMR
0.74
(95% CI: 0.29-1.53)
(7)
ML
PMR
1.57
(95% CI: 1.01-2.34)
(24)
Other leukemia
o
2 »
5 s
to
o
-------
Table 4-4. Epidemiologic studies of formaldehyde and lymphohematopoietic cancers
S"4
>3*
s
Reference
Study design
Exposure assessment
Results, statistical significance (number of observed deaths for cohort
study)
PMR
2.28
(95% CI: 1.39-3.52)
(20)
Blair etal. (1993)
Population-based case-control
study of 622 white men with
LHP cancers. Cancers selected
from Iowa and Minnesota cancer
surveillance networks diagnosed
between 10/80 and 9/82. 1,245
matched controls for living cases
selected by random digit dialing
if younger than age 65 and from
Medicare records if 65 or older.
Study focused on agricultural
exposures.
Personal interviews of
subjects or next of kin
included job histories,
agricultural exposures,
and chemical exposures.
Job titles used to create
job exposure matrix.
Industrial hygienist
estimated probability
and intensity of
exposures to large
numbers of substances.
Non-Hodgkin's
lymphoma
(formaldehyde
exposure)
Funeral service worker
O O
1.2
2.1
(95% CI: 0.9-1.7)
(95% CI: 0.5-7.9)
(6)
o
S ?
a, Co
TO Sj-
§ ^
s >
*s ^
^ ri
?»»i.
aAdjusted for age, state, smoking, family history of malignant proliferative disease, agricultural exposure to pesticides, use of dye, and direct/surrogate response
to interview.
o
On
-------
1
2
3
4
5
6
7
8
9
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
4.1.2.2.2. Brain and CNS cancer. Several studies of professional groups discussed earlier
investigated brain and other CNS cancers among those exposed to formaldehyde on the job.
Several of these studies found that exposure increased risk two to three times among exposed
professionals (Hall et al., 1991; Stroup et al., 1986; Walrath and Fraumeni, 1984), while others
found modest or no increase in risk (Hayes et al., 1990; Levine et al., 1984; Walrath and
Fraumeni, 1983).
None of the industrial cohort worker mortality studies of exposure to formaldehyde found
a clear relationship between formaldehyde exposure and risk of brain or CNS cancer (Pinkerton
et al., 2004; Coggon et al., 2003; Andjelkovich et al., 1995; Gardner et al., 1993; Stayner et al.,
1988; Blair et al., 1987, 1986). To date, no case-control studies of brain and CNS cancer have
been completed. In the Hauptmann et al. (2004) study, the authors reported that no clear
association was seen for cancer of the brain and CNS and exposure to formaldehyde.
4.1.2.2.3. Pancreatic and other cancers. Two studies (Kernan et al., 1999; Dell and Teta, 1995)
have found increases in the risk of pancreatic cancer in association with possible exposure to
formaldehyde. Collins et al. (2001a) conducted a meta-analysis of fourteen studies (Kernan et
al., 1999; Andjelkovich et al., 1995; Hansen and Olsen, 1995; Gardner et al., 1993; Hall et al.,
1991; Matanoski, 1991; Hayes et al., 1990; Gerin et al., 1989; Stayner et al., 1988; Blair et al.,
1986; Stroup et al., 1986; Levine et al., 1984; Walrath and Fraumeni, 1984, 1983) and found a
small increase in risk (RR = 1.1 [95% CI: 1.0-1.3]).
Other sites that have been examined are stomach cancer (Coggon et al., 2003) (SMR =
1.47;p < 0.05), intraoccular melanoma (Holly et al., 1996) (OR = 2.9 [95% CI: 1.2-7.0]), and
thyroid cancer among women (Wong et al., 2006) (OR = 8.33 [95% CI: 1.16-60.0]; 2 cases).
However, without further substantiation, it is difficult to infer causation based on these isolated
results alone.
4.1.2.3. Summary: Carcinogenic Hazard in Humans
The weight of the epidemiologic evidence at this time supports a link between
formaldehyde exposure and NPC in humans. This conclusion is based on the longitudinal cohort
study of Hauptmann et al. (2004) as well as the case-control studies of NPC and formaldehyde
exposure completed by Hildesheim et al. (2001), Vaughan et al. (2000), and several additional
case-control studies described in the text. With the exception of Hauptmann et al. (2004), most
of the other cohort studies found little or no increased risk of NPC from exposure to
formaldehyde. However, Hauptmann et al. (2004) employed different exposure metrics and
based their analyses on conservative internal comparisons that limited the potential for the HWE
This document is a draft for review purposes only and does not constitute Agency policy.
4-107 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
to obscure true effects. The case-control studies that provide additional evidence of an
association between NPC and formaldehyde have more power and generally rely on better
diagnoses of NPC. Better ascertainment of histologic types of tumors can sometimes also be
obtained if the cases are taken from cancer registries. The NPC risk is also supported by
experimental evidence in animals in which formaldehyde induces nasal cancers (Section 4.2.2).
Since the physiology of the rat nasal passage is somewhat different from that of humans, it is not
possible to obtain a direct site-specific correspondence between the species. However, in both
species, the tumors are found within the same area of the URT where maximum exposure can be
expected to occur.
Several researchers have challenged the conclusion of a relationship between
formaldehyde and NPC. Those critical of the link argue that, given the wide variability in results
across studies and competing explanations, conclusions about any link from the existing studies
are premature. The difficulty in attaining consensus on whether formaldehyde influences the risk
of NPC in humans arises from several limitations inherent in epidemiologic methods and
exposure assessment, as well as from the characteristics of the disease. The most prominent of
these limitations are the rarity of the cancer and imprecise estimates of exposure. Because NPC
is a very rare cancer with an incidence of less than 1 per 100,000, it is difficult to obtain precise
estimates of risk from cohort studies. Although case-control studies are better suited for
studying rare conditions, they are limited in obtaining valid and precise exposure assessments. A
further problem with exposure assessment is isolating formaldehyde exposure from other
potential chemical or particulate exposures that may influence risk of NPC. Imprecise exposure
assessment and the inability to isolate formaldehyde exposure from other exposures are largely
the bases on which Marsh and coworkers have challenged the NCI cohort study (Marsh et al.,
2007a, b, 2002, 1996; Marsh and Youk, 2005). Marsh and coworkers (Marsh et al., 2007a) show
that subjectively assessed exposure to silversmithing is tentatively associated with NPC. Given
that there were no prior citations of an association between silversmithing exposures and NPC in
the medical literature and given the many post hoc reexaminations of alternative hypotheses to
explain the original NCI findings, it is more likely that silversmithing is an artifactual potential
confounder.
It may be expected that, without new approaches for obtaining more accurate and precise
estimates of exposure, further follow-up of current cohorts and future epidemiologic studies of
formaldehyde and NPC will face the same limitations and criticisms found with existing studies.
These limitations notwithstanding, the epidemiologic studies reviewed here represent what may
be currently discernable about a formaldehyde-NPC link in humans by using rigorous
observational methods. As such, concluding any influence of formaldehyde must be made on the
This document is a draft for review purposes only and does not constitute Agency policy.
4-108 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
weight of all human and animal evidence in the face of known and expected limitations in study
design and exposure assessment.
The results of two well-designed cohort studies found a positive association between
formaldehyde-exposed professionals, such as pathologists, embalmers, and funeral directors, and
LHP cancer, particularly ML. The largest cohort study of formaldehyde has the most extensive
exposure assessment (Blair et al., 1986; Stewart et al., 1986), and the cohort was followed for a
median duration of 35 years (Hauptmann et al., 2003). By using cumulative exposure measures
not previously used and by using internal comparison groups, significant increases in the risk of
cancer of the LHP system, particularly ML, were reported. This study demonstrated that
formaldehyde was a risk factor for LHP cancers, independent of other risk factors, such as
benzene and smoking. Hauptmann et al. (2003) found statistically significant dose trends for
peak exposure and AIE. Pinkerton et al. (2004) also found a significant increase in the risk of
ML in garment workers 20 years after their initial exposure and in workers with 10 or more years
of exposure. Additionally, several studies of pathologists, embalmers, and other medical
workers reported greater numbers of observed deaths from leukemia than expected although
many studies of these groups suffer from a substantial HWE based on comparisons with external
death rates. Two of these studies, Hayes et al. (1990) and Stroup et al. (1986), also report a
significantly excess risk of ML in embalmers, funeral directors, and anatomists.
There is a range of biological plausibility for an agent whose primary action is at the
POE. Acute leukemias (ALL and AML), believed to arise from transformation of stem cells in
the bone marrow, are less plausible. In contrast chronic lymphatic leukemia, lymphomas,
multiple myelomas (from plasma B cells), and unspecified cancers may involve an etiology in
peripheral tissues to include cells, cell aggregates, germinal centers, and lymph nodes. An
association of these cancers to an exogenous agent acting at the POE is biologically plausible.
It is the conclusion of this assessment that the weight of the epidemiologic evidence at
this time supports a link between formaldehyde exposure and carcinogenicity in humans.
4.2. ANIMAL STUDIES
This section discusses the available laboratory animal data on the toxicity of inhalation,
oral, and dermal exposures to formaldehyde. A comprehensive database of laboratory animal
studies is available for formaldehyde, including numerous 2-year bioassays by both the
inhalation and oral exposure routes and dermal application studies. Although a large portion of
the literature reports studies focused on toxic effects at the site of contact or POE, general
systemic effects as well as neurobehavioral effects, reproductive and developmental effects,
immunologic changes, and sensitization are well represented in the literature as well.
This document is a draft for review purposes only and does not constitute Agency policy.
4-109 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
RB is a reduction in ventilation rate, minute volume, and other physiological parameters
experienced by rodents exposed to an irritant/reactive gas. Although humans and nonhuman
primates do not exhibit the same change in respiratory rate, these studies are included in Section
4.2.1.1 in order to better understand the effects on RB in interpreting rodent studies presented in
the balance of the chapter. Additionally, although binding to the trigeminal nerve and
subsequent downstream events do not result in pulmonary changes in humans, the mechanism
itself plays a role in understanding other adverse health effects observed in humans.
The available data for the three exposure routes confirm direct formaldehyde-induced
toxicity in tissues present at the POE. These observations are consistent with the
physicochemical characteristics, reactivity, and metabolic pathways of formaldehyde as
discussed in Chapter 3. Indications of cell damage, cell proliferation, and inflammatory
responses are similar for each route of exposure, therefore effects at the POE for inhalation and
oral exposures are described first (Sections 4.2.1.2 and 4.2.1.3, respectively). Given the well-
established nature of these health effects and the wealth of literature for inhalation exposures,
complete study summaries for respiratory tract effects are provided. Studies are organized by
study duration—acute, subchronic and chronic—where some of the chronic bioassays were
designed to address carcinogenic potential. Section 4.2.2 pulls together the findings of chronic
studies across the routes of exposure to evaluate the carcinogenic potential of formaldehyde
exposure.
Although a majority of the oral and inhalation studies focus on health effects at the
POE—respiratory tract and GI tract—the general systemic toxicity of formaldehyde is addressed
where it was integral to the study. Therefore, body weight and organ weight changes, gross
pathology, organ histopathology outside of the POE, blood and urine chemistry, and other
biochemical measures may be included in these study summaries. An overview of general
systemic findings is provided in Section 4.4 for all routes of exposure.
Studies addressing immune function, neurobehavioral effects, sensitization, and
reproductive and developmental effects are addressed across routes of exposure. The specialized
nature of these studies requires discrete treatment, and inclusion of data across routes of
exposure allows for a synthesis of the available information.
4.2.1. Noncancer Health Effects
4.2.1.1. Reflex Bradypnea
Reflex bradypnea (RB), which is believed to be a protective response, is often observed
in rodents exposed to reactive gases. It is primarily characterized by marked decreases in
activity, respiratory rate, body temperature, and metabolic rate. RB is not seen in humans and
This document is a draft for review purposes only and does not constitute Agency policy.
4-110 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
nonhuman primates. An understanding of the RB is important to the interpretation of many of
the animal bioassays examining formaldehyde-induced health effects. Of chief concern is that
the physiological effects of RB, described below, may interfere with appropriate interpretation of
adverse effects noted with formaldehyde exposure. It is important to distinguish between an
effect directly related to RB versus formaldehyde exposure. Additionally the effects of RB may
mask or alter formaldehyde-induced health effects. Secondly, differential respiratory effects of
RB due to species and strain will result in differential inhaled doses at the same exposure level.
This needs to be considered both when comparing the results of animal studies and in
extrapolation to humans. Finally, although humans do not experience RB, the mechanism of RB
as a reflex response to trigeminal nerve stimulation assists in understanding human health related
to localized and reflex responses due to trigeminal nerve stimulation.
Irritant gases have been shown to decrease body temperature, heart rate, and blood
pressure as well as alter blood chemistry in rodents (Pauluhn, 2003, 1996; Jaeger and Gearhart,
1982). Because of their small size, mice can rapidly lower their body temperatures and thus their
metabolic rate and ventilation rate. The hypothermia that results from RB can directly affect
nearly all biological processes (Gordon et al., 2008). Formaldehyde exposure can dramatically
lower ventilation rate and reduce body temperature in mice by as much as 4°C, and it has been
posited that decreased oxygen supply is likely to have profound effects on organisms with
substantial oxygen demands (Jaeger and Gearhart, 1982). The effects of RB are reversible,
though it may take several minutes to several hours to return to pre-exposure conditions
(Pauluhn, 1996; Jaeger and Gearhart, 1982).
The literature on sensory irritation is broad; many studies have investigated species
differences, dose response relationships, tolerance, and cross-tolerance to other sensory irritants
(see Tables 4-5 and 4-6). This discussion focuses on the changes in respiratory rate and minute
volume during formaldehyde exposure. Sensory irritation is often quantified as the statistically
derived exposure concentration that results in a 50% reduction in respiratory rate (RD50) in
rodents (ASTM, 2000; Kane et al., 1979). Kane and Alarie (1977) evaluated various aspects of
sensory irritation, including establishing the RD50, exploring the reproducibility of response,
investigating the effect of tracheal cannulation, and determining the potential for tolerance with
repeated exposure or pre-exposure in male Swiss-Webster mice, caused by formaldehyde and
acrolein. The RD50 was established by exposing four mice for 10 minutes at each concentration
across a range representing approximately 10 to 80% reduction in respiration and calculated by
using least squares regression. The RD50 and its 95% CI for formaldehyde were calculated to be
3.1 (2.1-4.7) ppm (3.8 [2.58-5.77] mg/m3). The tracheal cannulation experiments demonstrated
that the effect on respiratory rate was caused by URT sensory irritation.
This document is a draft for review purposes only and does not constitute Agency policy.
4-111 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-5. Respiratory effects of formaldehyde-induced reflex bradypnea in
2 various strains of mice
3
Species/strain
No./
group
Treatment"
Respiratory effects
Reference
Male Swiss-
Webster mice
4
Duration: 10 minutes.
Exposure: up to 100 ppm.
RD50a =3.1 ppm
(95% CI: 2.1-4.7).
Kane and Alarie
(1977)
Male Swiss-
Webster mice
8
Duration: 3 hours/day for
3 days. Exposure: 0.52, 0.44,
I.16, 1.83,3.10, 5.35, 5.60, and
II.2 ppm.
RD50 = 3.4 ppm
(95% CI: 2.4-4.7).
Kane and Alarie
(1977)
Male Swiss-
Webster mice
4
Duration: 10 minutes (head
only). Exposure: up to 10 ppm.
RD50 = 3.2 ppm
(95% CI: 2.1—4.7).
Steinhagen and
Barrow (1984)
Male Swiss OF,
mice
6
Single 5-minute exposure to
four unspecified
concentrations.
RD50 = 5.3 ppm.
De Ceaurriz et al.
(1981)
Male B6C3F1 mice
4
Duration: 10 minutes (head
only).
Exposure: Range up to 10 ppm.
RD50 = 4.9 ppm
(95% CI: 3.9-6.4).
Steinhagen and
Barrow (1984)
Male B6C3F1 mice
4
Duration: 10 minutes (head
only). Exposure: up to 15 ppm
Pretreatment: 2, 6, or 15 ppm
6 hours/day for 4 days.
Naive mice: RD50 = 4.4 ppm
(95% CI: 0.9-5.0)
Pretreated mice: RD50 =
4.3 ppm (95% CI: 3.4-5.5).
Chang et al.
(1981); Barrow et
al. (1983)
Male C57BL6/F1
mice
3
Whole-body exposure for up to
2 hours.
After 1.25 hours:
Tidal volume reduced by
33%; 68% reduction in
respiratory frequency; C02
production reduced by
50%; %; body temperature
dropped from 37.8 to
34.7°C.
Jaeger and
Gearhart (1982)
5 aExposure concentration that results in a 50% reduction in respiratory rate.
This document is a draft for review purposes only and does not constitute Agency policy.
4-112 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-6. Respiratory effects of formaldehyde-induced reflex bradypnea in various
2 strains of rats
3
Species/strain
No./
group
Treatment
Respiratory effects
Reference
Male Crl-CD rats
4
RD50 = 13.8 ppm.
Male Wistar rats
4
30-minute nose-only exposure
to a range of formaldehyde
concentrations.
RD50 = 10.0 ppm.
Cassee et al.
(1996)
Male F344 rats
4
Duration: 10 minutes (head
only). Exposure: up to
56 ppm.
Pretreatment:2, 6, or 15 ppm
6 hours/day for 4 days
Naive rats: RD50 = 13.1 ppm
(95% CI: 10.6-17.5)
Pretreated rats: RD50 =10.8 ppm
(95% CI: 7.6-16.9)
Chang et al.
(1981); Barrow
et al. (1983)
Male F344 rats
4
Single 10-minute head-only
exposure to a range of
concentrations.
Pretreatment: 15 or 28 ppm
formaldehyde or 10 ppm
chlorine.
Baseline RD50 = 31.7 ppm.
Pre-exposure to formaldehyde-
induced tolerance at 28 ppm
(RD50 = 20.2 ppm) but not
15 ppm.
Pre-exposure to chlorine-induced
tolerance to formaldehyde
(RD50 ranged from 64.5 to
115 ppm, depending on
exposure duration).
Chang and
Barrow (1984)
Male F344 rats
ND
10 minute exposure to acrolein
or acetaldehyde (head only).
Pre-exposed to formaldehyde
at 15 ppm for 6 hours/day for
9 days.
Pre-exposure to formaldehyde-
induced tolerance:
Acetaldehyde (RD50 = 2,991 ppm
in naive versus 10,601 ppm in
preconditioned animals)
Acrolein (RD50 = 6 ppm in naive
versus 29.6 ppm in preconditioned
animals).
Babiuk et al.
(1985)
Male Charles
Rivers CD rats
3
Whole-body exposure for up
to 2 hours.
After 0.7 hours:
Tidal volume reduced by
22%; 20% reduction in
respiratory frequency; C02
production unaffected.
Jaeger and
Gearhart (1982)
4
5
6 Across the literature there is fairly good agreement on RD50 values for various strains of
7 mice (Table 4-5), ranging from 3.1 ppm in male Swiss-Webster mice to 4.9 ppm in male
8 B6C3F1 mice. Rats are less sensitive, with RD50 values ranging from 10 ppm in male Wistar
9 rats to 31.7 ppm in male F344 rats. No reported RD50 for female rodents exposed to
10 formaldehyde exists.
11 Jaeger and Gearhart (1982) evaluated the effect of formaldehyde on respiratory rate, tidal
12 volume, minute volume, carbon dioxide (C02) production (exhaled to air) as a reflection of total
13 metabolism, and core body temperature in male Charles River CD rats and male C57BL6/F1
This document is a draft for review purposes only and does not constitute Agency policy.
4-113 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
mice. Animals (three/concentration) were exposed to 15 ppm (18.4 mg/m3) formaldehyde for up
to 2 hours. Mice exhibited a greater decrease in respiratory frequency and minute volume
compared with the rats. C02 production and body temperature were also affected to a greater
extent in the mice (Table 4-5). The authors postulated that the decreased body temperature in
mice would likely lead to decreased biologic action of formaldehyde in the tissue.
4.2.1.1.1. Tolerance. Tolerance is defined as an increase in the concentration required to elicit
the same degree of RB response and was evaluated by Kane and Alarie (1977). In the first set of
experiments, mice (four/concentration) were exposed 3 hours/day for 4 days at the concentration
associated with either a 30 or 50% decrease in respiratory frequency (specific concentrations not
given) (Kane and Alarie, 1977). Naive animals served as controls for each day. The maximum
response increased with each additional day of exposure, and the diminution of response that was
typically exhibited after 60 minutes of exposure in naive animals was markedly delayed. In the
second set of experiments, mice were exposed to a formaldehyde concentration at one-tenth the
RD50 (i.e., 0.3 ppm) 3 hours/day for 3 days. On the fourth day the animals underwent a similar
exposure protocol to identify the concentration that resulted in an RD50, following the above
protocol. No change in the RD50 was demonstrated. Both of these experiments indicate no
change in tolerance with either type of pretreatment in Swiss-Webster mice.
Chang and Barrow (1984) tested whether tolerance would develop in male F344
(CDF[F344]Crl/Br) rats exposed to formaldehyde. Exposure to formaldehyde at 15 ppm
(18.4 mg/m3) for 6 hours/day, 5 days/week failed to induce tolerance. However, tolerance was
observed following exposure to 28 ppm (34.4 mg/m3) formaldehyde for 4 days. The
concentration-response curve in these animals was significantly different than that of naive
animals, with an increase in the RD50 estimate for this exposure duration from 31.7 to 70.2 ppm.
4.2.1.1.2. Cross-species differences in inhaled dose. Formaldehyde-induced RB lowers both
respiratory rate and tidal volume and thus reduces the inhaled dose of formaldehyde at a given
exposure concentration. Chang et al. (1983) and Barrow et al. (1983) evaluated the species
differences and the effective inhaled dose between rats and mice, since mice seem to be more
sensitive to formaldehyde-induced RB and do not exhibit tolerance as shown in F344 rats.
Groups (four/concentration) of male F344 rats and male B6C3F1 mice were exposed to
formaldehyde concentration ranges of 6.2-48 ppm (7.6-59 mg/m3) or 0.78-14.0 ppm (0.96-
17.2 mg/m3), respectively, for 10 minutes. Pretreated animals used in the tolerance experiments
were exposed to formaldehyde at 2, 6, or 15 ppm (2.45, 7.36, or 18.4 mg/m3) 6 hours/day for
4 days prior to determination of the RD50 and concentration response across the same ranges.
This document is a draft for review purposes only and does not constitute Agency policy.
4-114 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
A concentration-dependent decrease in respiratory rate was seen in both naive and
pretreated rats during formaldehyde exposure. Tolerance (defined as a decrease in respiratory
rate followed by a subsequent return to control values) occurred after 4 minutes of exposure and
was more pronounced at concentrations above 4 ppm. Concentration-response relationships
were very similar for naive and pretreated rats, and the RD50s were similar for both groups
(naive = 13.1 ppm [95% CI: 10.6-17.5]; pretreated = 10.8 ppm [95% CI: 7.6-16.9 ]). In
contrast, naive or pretreated mice did not develop tolerance during exposures. An examination
of concentration-response relationships for mice showed similar RD50 values (naive = 4.4 ppm
[95% CI: 0.9-5.0 ] and pretreated = 4.3 ppm [95% CI: 3.4-5.5]) compared with rats, although
the slopes of the concentration-response regressions were statistically different (Figure 4-2).
F 344 Rats
• Naive (6 ppm)
° Pretreated (6 ppm)
¦ Naive (15 ppm)
n Pretreated (15 ppm)
360
0 2 4 6 810
120
240
110
100
3
CO
o
Q.
X
0
Q.
M—
O
0s
<1)
E
o
>
£
360
0 2 4 6 810
120
240
Time (min.) during exposure
Figure 4-2. Formaldehyde effects on minute volume in naive and formaldehyde-
pretreated male B6C3F1 mice and F344 rats.
Source: Redrawn from Chang et al. (1983).
Exposure of naive or pretreated rats resulted in an increased (compensatory) tidal
volume. However, the increase in tidal volume did not compensate entirely for the decrease in
ventilation rate and was only concentration dependent in pretreated rats. Comparison of tidal
volume from naive and pretreated mice exposed to formaldehyde showed a slight increase in
naive animals but a decrease in pretreated ones. The effect of formaldehyde exposure on tidal
This document is a draft for review purposes only and does not constitute Agency policy.
4-115 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
volume was concentration dependent in both groups of mice. These results indicate that tidal
volume does not compensate entirely for the decrease in respiratory rate and that the
compensation is slightly greater in rats than in mice.
These studies (Barrow et al., 1983; Chang et al., 1983) showed thatB6C3Fl mice sustain
RB, whereas F344 rats develop tolerance more readily both during exposure and with
pretreatment. Thus, these results suggest that the rat may be the more sensitive species for the
effects of inhaled formaldehyde due in part to the difference in sensitivity between mice and rats
as evidenced by an RD50 of 4.9 versus 31.7 ppm and the ability of rats to develop tolerance while
mice appear to sustain RB. Barrow et al. (1983) used the results of these experiments to estimate
an inhaled dose equivalent to the exposure concentration of 15 ppm for the strains of mice and
rats used in the chronic formaldehyde bioassays by Kerns et al. (1983) and Monticello and
Morgan (1994) described in Section 4.1.2 as follows:
Inhaled dose ([j,g/min-cm2) =
HCHO concentration (us/L) x minute volume (L/min)
Nasal cavity surface area (cm2) (5-1)
As shown in Table 4-7, because mice were observed to be able to decrease their minute
volume by approximately 75% as compared with 45% in rats, a twofold higher inhaled dose
would be expected in rats versus mice. This difference may be relevant to the increased
incidence of SCC in the nasal cavity seen in F344 rats when compared with B6C3F1 mice.
Table 4-7. Inhaled dose of formaldehyde to nasal mucosa of F344 rats and
B6C3F1 mice exposed to 15 ppm
Parameter
F344 rats
B6C3F1 mice
HCHO concentration (|ig/L)
18.4
18.4
Minute volume (L/min)
0.114
0.012
URT surface area (cm2)
13.44
2.89
Inhaled dose (ng/min/cm2)
0.156
0.076
Source: Barrow et al. (1983).
4.2.1.1.3. Cross-tolerance. Cross-tolerance of chemically-induced reflex responses has been
examined in several systems in order to better understand the specificity and nature of the
interaction of reactive chemicals (such as formaldehyde with chlorine) with the trigeminal nerve
involved in the RB. Development of cross-tolerance to formaldehyde following preexposure to
This document is a draft for review purposes only and does not constitute Agency policy.
4-116 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
chlorine or to chlorine following preexposure to formaldehyde was shown to be a function of the
duration of the pretreatment in male F344 rats (Chang and Barrow, 1984) (Table 4-8). A 7-day
recovery period resulted in only a slight loss of cross-tolerance from a 4-day pre-exposure to
either chlorine or formaldehyde (data not shown). The cross-tolerance between formaldehyde
and chlorine demonstrated in the Chang and Barrow (1984) study suggests that these chemicals
may act via a common mechanism and may involve the trigeminal nerve. In rats, cross-tolerance
was induced after chlorine exposure but not after formaldehyde exposure, which suggests that
the trigeminal nerve may have different reactive sites that are differentially activated, depending
on the stimulus.
Table 4-8. Exposure regimen for cross-tolerance study
Pre-exposure
Chlorine RD50
FA-p retreated
Na'ive
Formaldehyde
15 ppm
6 hours/day
1 day
22.6 ppm
10.9 ppm
4 days
16.8 ppm
10 days
64.5 ppm
Pre-exposure
Formaldehyde RDS0
Cl-p retreated
Naive
Chlorine
10 ppm
6 hours/day
1 day
64.5 ppm
31.7
4 days
66 ppm
10 days
115 ppm
Source: Chang and Barrow (1984).
Babiuk et al. (1985) evaluated the potential for formaldehyde pretreatment to cause cross-
tolerance with various other inhaled aldehydes, including acetaldehyde and acrolein. Male F344
rats were pretreated with 15 ppm (18.4 mg/m3) formaldehyde 6 hours/day for 9 days and
challenged on the 10th day with the second aldehyde for 10 minutes at various concentrations
(four rats/concentration) to establish an RD50. Exposure to acetaldehyde and acrolein, the two
smallest molecules in the series of aldehydes tested, resulted in cross-tolerance. The RD50 and
its 95% CI for acetaldehyde were estimated at 2,991 (95% CI: 2,411-3,825) ppm in the naive
rats, and this was increased by approximately 3.5-fold to 10,601 (95% CI: 7,902-15,442) ppm in
the rats pretreated with formaldehyde. With acrolein, the RD50 increased approximately fivefold,
from 6.0 (95% CI: 3.5-18.1) ppm to 29.6 (95% CI: 15.6-93.0) ppm. Cross-tolerance with
formaldehyde has only been demonstrated with acetaldehyde, acrolein, and chlorine (Babiuk et
al., 1985; Chang and Barrow, 1984), suggesting that it is not a generalized phenomenon.
Whether the phenomenon of tolerance involves modulation of specific trigeminal nerve
receptors or whether it results from less specific chemical injury of the nasal mucosa has not
This document is a draft for review purposes only and does not constitute Agency policy.
4-117 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
been determined. For example, different mechanisms lead to stimulation of the trigeminal nerve
and are likely to control the decrease in respiratory rate. In particular, acetaldehyde might
interact with sensory nerves via an amino group (Steinhagen and Barrow, 1984; Schauenstein et
al., 1977), whereas the receptor-binding site for formaldehyde and acrolein is believed to be a
thiol group. Furthermore, different binding sites exist on the trigeminal nerve for different
irritants (Nielsen, 1991). Thus, Bos et al. (1992) concluded that the data on tolerance or
"desensitization" versus "sensitization" (as defined strictly on the basis of the respiratory apneic
response) may be the result of adaptation or reversible/irreversible adverse changes. The
mechanisms underlying sensitization or desensitization are not well characterized.
4.2.1.1.4. Formaldehyde binding and activation of trigeminal nerve afferent activity. Kane
and Alarie (1978) evaluated the effect of 11 combinations of acrolein and formaldehyde on
respiratory rate in outbred specific-pathogen-free male Swiss-Webster mice. Exposure
concentrations ranged from 0.12-8.97 ppm (0.28-21 mg/m3) for acrolein and 0.37-9.73 ppm
(0.45-11.9 mg/m3) for formaldehyde. The data were evaluated using a simple model of
competitive antagonism. Comparing the observed and predicted responses indicated no apparent
differences, and paired t-tests showed no statistical significance. The authors concluded that
acrolein and formaldehyde acted at the same receptor site and acted as competitive antagonists
when exposure occurred simultaneously.
Kulle and Cooper (1975) investigated the effects of formaldehyde on trigeminal nerve
afferent activity in adult male Sprague-Dawley rats. The authors isolated both the ethmoid and
nasopalatine branches of the trigeminal nerve and recorded afferent signaling as electrical
activity while reactive gases (formaldehyde, ozone, and amyl alcohol) were passed through the
nasal passages of the anesthetized animals. The authors reported that both branches of the
trigeminal nerve responded similarly to all three chemicals, and they therefore conducted the
balance of their experiments on the nasopalatine branch of the nerve. Nerve response was
calculated as the difference between exposed and control activity, and the threshold for a positive
response was arbitrarily defined as an increase of 0.1 spikes per second. The sensory threshold
was determined by extrapolation from the measured nerve response to a range of formaldehyde
concentrations (0.5-2.5 ppm) or ozone (5.0-29 ppm) for an exposure duration of 2 minutes.
Amyl alcohol exposure (0.3-10.0 ppm) lasted for 25 seconds. Threshold was arbitrarily defined
as an increase of 0.1 spikes per second. The mean thresholds were 0.25 ppm for formaldehyde,
5.0 ppm for ozone, and 0.30 ppm for amyl alcohol, suggesting that the trigeminal nerve is highly
sensitive to formaldehyde and amyl alcohol compared with ozone exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
4-118 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
In a second set of experiments, Kulle and Cooper (1975) investigated the effects of
prolonged formaldehyde-exposure on the odor response to amyl alcohol. Rats were pre-exposed
to a series of amyl alcohol concentrations (0.3, 0.7, 1.0, 3.3, 6.7, or 10.0 ppm [1.08, 2.52, 3.6,
11.9, 24, or 36 mg/m3]) then a 1-hour continuous formaldehyde exposure (0, 0.5, 1.0, 1.5, or
2.0 ppm [0, 0.61, 1.23, 1.84, or 2.45 mg/m3]). There was a progressive decrease in odor
response to amyl alcohol with increasing stimulus of formaldehyde concentration (p < 0.01,
analysis of variance [ANOVA]). The response to formaldehyde concentration was described by
a power function Y = 0.741 x X1'47, where X is the formaldehyde concentration. The effects of
exposure to 2.0 ppm were similar, regardless of whether it was presented immediately as a
separate exposure or as the final concentration of a progressively increasing series. The response
to amyl alcohol did not fully recover within the 1-hour extended recovery period. Thus, it
appeared that the afferent function depression was not due to receptor adaptation or insufficient
time for formaldehyde diffusion away from receptor sites.
In an attempt to elucidate the basis of the differential effects of various types of
aldehydes on sensory irritation, Tsubone and Kawata (1991) recorded the afferent activity of the
surgically isolated ethmoidal nerve (a branch of the trigeminal nerve) during delivery of 0.32-
4.7 ppm (0.39-5.77 mg/m3) formaldehyde, 0.18-7.2 ppm (0.41-16.5 mg/m3) acrolein, and 134—
2,232 ppm (241-4,021 mg/m3) acetaldehyde into the cannulated URT of male Wistar rats
(six/aldehyde) at a flow rate of 200 mL/minutes for 22 seconds. Only one aldehyde was used in
each animal and each exposure was repeated two to four times at different concentrations. The
activity of the nerve was recorded as the number of electrical discharges for a total period of
100 seconds, including pre-inhalation (30 second), inhalation (22 second), and post-inhalation
(48 second) periods. Nitrogen was used as the control gas and as the vehicle to dilute the
aldehyde gases in order to not interfere with the gas chromatography used to analyze the
exposures. The vapor concentrations associated with a 50% increase in nerve activity over the
level of control gas were calculated as approximately 1.8, 1.2, and 908 ppm for formaldehyde,
acrolein, and acetaldehyde, respectively. These results are consistent with the findings of
Steinhagen and Barrow (1984) and the hypothesis that the differences in RD50 are due to
differences in chemical reactivity in the tissue.
In summary, RB is a phenomenon observed in rodents exposed to reactive gases,
believed to be a protective response to the irritant properties of the gas. In comparative studies,
rats appear more sensitive to irritant gases since they have a more pronounced RB response
compared with mice at a given concentration of formaldehyde and because the dose required to
elicit a bradypneic response is higher in rats than in mice. Interestingly, only rats appear to
develop tolerance to irritant gases, while mice sustain an RB response. When formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
4-119 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
exposure is studied in concert with other reactive gases like chlorine and other aldehydes like
acetaldehyde and acrolein, cross-tolerance developed. However, the mechanism underlying this
response is unknown. It is thought that RB may occur as a result of stimulation of the trigeminal
nerve. Thus, although RB appears to be a phenomenon specific to rodents, the mechanism by
which it occurs, trigeminal nerve stimulation, may be applicable to understanding MO As in other
species, such as primates and humans, particularly in regard to sensitization.
4.2.1.2. Respiratory Tract Pathology
The database for evaluating the POE toxicity in the respiratory tract of inhaled
formaldehyde is robust, with well-designed studies that span a duration range of a few hours to
chronic 2-year bioassays. Toxicity testing has been performed in various species, including
mice, rats, hamsters, guinea pigs, dogs, and nonhuman primates. Although a few studies include
examination of tissues outside of the URT, the majority of studies focus on changes in cell
proliferation and cell pathology in the nasal mucosa. Both mice and rats are well-defined animal
models with standard histologic sections established to evaluate various regions of the nasal
passages, divided into Levels 1 to 5 and illustrated in Figure 4-3. Pathology of the nasal mucosa
will be discussed with reference to these sections, and the region examined will be stipulated
(e.g., nasoturbinates, maxilloturbinates, or ethmoid turbinates [ETs]). Additionally, pathology of
the respiratory epithelium will be distinguished from effects on the olfactory epithelium,
although the nature of the lesions is similar.
4.2.1.2.1. Mucociliary clearance. The mucociliary apparatus of the URT is the first line of
defense against airborne toxicants. Comprising a thick mucus layer (epiphase), hydrophase, and
ciliated epithelium, the mucociliary apparatus may entrain, neutralize, and remove particulates
and airborne chemicals from inspired air (Figure 4-4). The mucus serves to entrain or neutralize
and remove exogenous agents from the nasal epithelium (e.g., particles, reactive chemicals). As
reviewed by Kim et al. (2003), the nasal mucus contains proteins, glycoprotein, and lipids but is
primarily water (95%) and is propelled along by movement of the underlying cilia. Degradation
in the continuity or function of the mucociliary apparatus, which provides protection to the nasal
epithelium, would result in higher levels of gases and particles reaching the nasal epithelium
itself and greater penetration of chemicals into the respiratory tract. Therefore, breakdown and
disruption of mucociliary function are adverse effects, since a key bodily defense to exogenous
agents (including infectious agents) is damaged.
This document is a draft for review purposes only and does not constitute Agency policy.
4-120 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
6
Tffi
Figure 4-3. Sagittal view of the rat nose (nares oriented to the left).
Note: The figure shows the normal distribution of nasal mucosae and the section
levels used in contemporary histopathology (Brenneman et al., 2000; Mery et al.,
1994). Sections 1, 2, 4, and 5 correspond to Levels I, II, III, and IV as proposed
by Young (1981). S = squamous, T/R = transitional/respiratory, O = olfactory
mucosa.
Source: Brenneman et al. (2000).
This document is a draft for review purposes only and does not constitute Agency policy.
4-121 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
OM
ECS,
Gl
8V
m
Figure 4-4. Main components of the nasal respiratory epithelium.
Note: OM = osmiophilic membrane; EP = epiphase; HY = hypophase; CI = cilia;
MV = microvilli; CJ = cell junction; CC = ciliated cell; NCC = non-ciliated cell;
GC = goblet cell; NE = nerve; GL = gland; BV = blood vessel; ECS =
extracellular space; BM = basement membrane.
Source: Morgan et al. (1986d).
Mucus flow slows upon formaldehyde exposure, despite an increase in the ciliary
beat of the underlying epithelial cells, which propel the mucus across the nasal epithelium
(Morgan et al., 1986a, c, d; 1983). These findings are consistent with other studies since
airborne pollutants and reactive gases have been shown to decrease mucus flow rates in
several animal models (Mannix et al., 1983; Iravani, 1974; Carson et al., 1966; Dalhamn,
1956; Cralley, 1942). In addition to slowing flow, the mucus layer has been observed
breaking up as it floats on the epiphase, creating gaps in the epiphase and revealing the
hydrophase below (Morgan et al., 1986c, d). Formaldehyde reacts with glycoproteins in
the mucus of the epiphase, creating cross-links between these large molecules; this is
believed to increase the viscosity of the mucus.
This document is a draft for review purposes only and does not constitute Agency policy.
4-122 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
In their first experiments, Morgan et al. (1983) describe progressive mucostasis (slowing
of mucous flow) and ciliastasis (disruption of ciliary beat) with increasing days of exposure to
formaldehyde in male F344 rats (15 ppm; 6 hours/day for 1, 2, 4, or 9 days). Ciliastasis occurred
with greater frequency and across more regions of the nasoturbinate with subsequent days of
exposure. After 9 days, mucostasis was recorded in all but two regions evaluated. Although the
severity and time course of these changes varied across regions of the nose, the process followed
a similar pattern: decreased flow, increased ciliary action, mucostasis, and ciliastasis. Since the
formaldehyde-induced deficits in mucociliary function increased with days of exposure, activity
did not fully recover between exposures (18 hours) (Morgan et al., 1983). Therefore, the
severity and extent of adverse effects are dependent on both the concentration of exposure and
duration (in this case, days of repeated exposures).
In subsequent studies, Morgan et al. (1986c) examined the exposure-response
relationship of formaldehyde effects on mucociliary function and functional recovery 18 hours
after exposure ceased. Exposure regimens similar to the above experiment included additional
exposure concentrations (0.5, 2, and 6 ppm) and an additional time point of 15 days duration.
Exposure at 2 and 6 ppm resulted in the same progression of effects on mucus flow and ciliary
beat. Considering both severity and extent of effects a clear exposure-response relationship was
demonstrated. Additionally, within each exposure group, effects progressed both in severity and
extent by duration of exposure to formaldehyde (from 1 to 4, 9, and 15 days of exposure)
(Morgan et al., 1986c).
Flow and ciliary beat were not reduced, but rather increased, in epithelium from rats
exposed to 0.5 ppm formaldehyde. Mucus flow in 2 of 10 areas assessed was clearly increased
(275 and 200% of controls) after 4 days of exposure to 0.5 ppm formaldehyde. Two other
epithelial regions showed a similar trend (150% of controls), but this change was not statistically
significant. Interestingly, measurements made in corresponding areas after 9 days of exposure
did not show an increase, and measurements in one region were reduced to 37% of control.
Although it is not known whether the observed increase in mucus flow rate is a subtle indication
of an adaptive response to a low level irritant, the increase appears to be transient. It is not
known if flow rate would continue to decrease below control levels for repeated exposures at
0.5 ppm for longer than 9 days.
The regions affected at 15 ppm generally included the lateral aspects of the nasoturbinate
and both the dorsal and medial aspects of the maxilloturbinate. In general there was an anterior
to posterior effect with increasing concentration and time. Additionally, impaired mucociliary
function was more extensive with greater concentration and length of exposure. Nasal lesions
were seen on the nasal epithelium and correlated with those areas where some inhibition of
This document is a draft for review purposes only and does not constitute Agency policy.
4-123 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
ciliary function was measured. Areas without mucus flow but that still retained ciliary function
did not develop epithelial lesions. Morgan et al. (1986c) reported "coagulated mucus," viewed
as a "continuous membrane" over the epithelium after 6 hours of exposure to 15 ppm
formaldehyde. Minor cell damage and infiltrating neutrophils and monocytes were also seen in
these areas. The coagulated mucus was not seen in similarly exposed rats that were allowed
18 hours of recovery before sacrifice. However, ciliated cells were damaged, and there was a
greater presence of neutrophils and macrophages (MPs) after this recovery period. The authors
noted that, as the exposure continued, these areas exhibited increased signs of inflammation and
epithelial damage, eventually resulting in "severe degenerative changes."
Morgan et al. (1986a) refined their study design to implement a nose-only exposure to
formaldehyde in order to better examine the progression of changes in mucociliary function
during short-term exposure, allowing examination of mucus flow immediately following
exposure. Three F344 rats/group were exposed to 15 ppm (18.4 mg/m3) formaldehyde for 10,
20, 45, or 90 minutes or 6 hours. Two groups of rats were exposed to 2 ppm to determine a no
effect level for 90 minutes or 6 hours. The extent and severity of mucostasis and ciliastasis seen
after a 6-hour 15 ppm (18.4 mg/m3) formaldehyde exposure and a 1-hour recovery period were
similar to the earlier study (Morgan et al., 1986a), indicating that similar exposure conditions
were reached with this nose-only apparatus. Ciliastasis and mucostasis were both less severe and
less extensive in a time-dependent manner and at the earlier time points of 10, 20, 45, and
90 minutes. Significant recovery was seen in mucociliary function by allowing a 1-hour
recovery between exposure and sacrifice. Regions of both the nasal septum and lateral wall,
which exhibited no mucus flow when examined immediately after a 6-hour exposure, had
measurable flow after the 1-hour recovery period. Similar recovery was seen at all durations of
exposure. No decreases in mucociliary function were seen after exposure for either 90 minutes
or 6 hours at 2 ppm formaldehyde. However, given evidence of recovery (Morgan et al., 1986a)
and the time taken to dissect and view the tissues ex vivo may have obscured more subtle effects.
To assess more immediate effects on mucociliary apparatus, Morgan et al. (1984a) have
examined formaldehyde effects on the mucociliary apparatus of isolated frog palates. This
system allowed observation of mucociliary function during exposure. Unexposed frog palates
were covered by a continuous sheet of mucus of variable thickness, which was observed to flow
in streams across the palette, exhibiting a wave-like form in some areas of the epiphase. The
authors reported particle movement in a lower, less viscous layer that was consistent with a less
viscous underlying hydrophase, similar to that described in rat mucosa. The basal mucus flow
rate was 0-4 mm/minute, with localized ciliary activity. Short periods of increased mucus flow
were associated with seemingly spontaneous increases in ciliary beat.
This document is a draft for review purposes only and does not constitute Agency policy.
4-124 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Formaldehyde exposure resulted in an initial increase in ciliary beat and mucus flow rate
in all palates exposed at 1.37, 4.36, and 9.58 ppm formaldehyde (but not 0.23 ppm). With
increasing formaldehyde concentration and time of exposure, mucostasis was evident as mucus
became stiff and eventually rigid. Ciliary beat continued after mucostasis was reached until
palates were exposed to 4.36 and 9.48 ppm formaldehyde, when ciliastasis was reached. The
time course to peak mucus flow rate, mucostasis, and ciliastasis was concentration dependent,
with mucostasis reached in less than 3 minutes at 9.48 ppm. In contrast, increased mucus flow
peaked at 8 minutes in palettes exposed at 1.52 ppm formaldehyde, which, though declining,
remained above basal levels after 25 minutes with no mucostasis or ciliastasis noted at this level.
Flo-Neyret et al. (2001) demonstrated reduced mucociliary clearance and decreased
frequency of ciliary beats by using a similar isolated frog palette mucociliary apparatus.
However the palates were exposed by formaldehyde in the Ringer's solution in which the palates
were placed (0, 1.25, 2.5, or 5 ppm). Also, mucus was removed from the palettes and did not
come into direct contact with the formaldehyde. Despite these differences, formaldehyde caused
mucociliary clearance to decrease in a time- and concentration-dependent manner; mucostasis
occurred after 60 minutes of exposure to 5 ppm formaldehyde (Figure 4-5). Ciliary beat was
decreased in a time-dependent manner at 2.5 and 5 ppm exposure but increased at 1.25 ppm
formaldehyde (Figure 4-5). Reduced mucociliary clearance at 2.5 and 5 ppm was consistent with
the reduced ciliary beat. However, clearance decreased at 1.25 ppm formaldehyde, where there
was an apparent increase in ciliary beat. The authors suggest this may be a result of disrupting
the harmonic movement of the cilia, impairing effective mucociliary clearance. Based on study
results, the authors hypothesize that changes in ciliary beat, including excitation at lower
exposures, are likely to be a direct effect of formaldehyde on epithelial cells or other cellular
components of the mucosa.
In summary, numerous studies have identified impaired mucociliary clearance activity
associated with formaldehyde exposure (Table 4-9). Although low-dose and short-term
exposures first increase ciliary beat, impaired mucus flow, slowed ciliary beat, and eventual
mucostasis and ciliastasis have been demonstrated in both in vivo and in vitro exposure systems.
These effects are both concentration and duration dependent and can be seen in as few as
15 minutes from exposure. Repeated inhalation exposures in rats indicate the effect does not
fully recovery in an 18-hour period between exposures, contributing to greater impairment over
extended periods of exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
4-125 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
-•-5.0 ppm
-H-2.5 ppm
-A-1.25 ppm
-O- Control
-•-5.0 ppm
-H-2.5 ppm
-A-1.25 ppm
-O-Control
£ 125-
~ 125-
¦o 100-
100-
75-
75-
50-
50-
25-
25-
O
T
40
Time (min) Time (min)
Figure 4-5. Decreased mucus clearance and ciliary beat in isolated frog palates
exposed to formaldehyde after 3 days in culture.
Source: Flo-Neyret et al. (2001).
Morgan et al. (1983) suggested that the initial stimulation of ciliary activity may be a
defensive response to the irritant gas, possibly indicating some penetration of formaldehyde to
the underlying epithelial cells. Later effects of mucostasis may be a result of cross-linking of
mucus glycoproteins by formaldehyde, creating a rigid mucus that is not able to flow even with a
rigorous ciliary beat. It is unknown if the eventual cessation of ciliary beat is a result of
compound-related effects on ciliated epithelium as formaldehyde diffuses through the mucus or
an indirect effect associated with mucostasis. However, in vitro experiments by Flo-Neyret et al.
(2001) indicate that formaldehyde in solution, supporting isolated frog palates without mucus,
resulted in the same sequence of effects, including increased ciliary beat at the lowest exposure.
These data suggest a role of formaldehyde beyond its ability to form protein cross-links in
mucociliary proteins.
4.2.1.2.2. Cellular pathology. This section summarizes studies that have investigated cellular
pathology in the URT and in the lung. Below, full study descriptions are provided for both
short-term and subchronic duration studies (including, where appropriate, how cell proliferation
relates to the observed formaldehyde-induced pathology).
This document is a draft for review purposes only and does not constitute Agency policy.
4-126 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-9. Summary of formaldehyde effects on mucociliary function in the upper respiratory tract
S"4
>3*
&
to
s
Species
Na
Treatment
Measure of
mucociliary
function
Summary of results by location
Reference
Male F344
rats
10
15 ppm formaldehyde
6 hours/day for 1, 2, 4,
or 9 days
Mucus flow and
ciliary beat
Mucostasis in regions 2, 3, 4, 5, and 8 for all rats after a single
dose.Mucostasis in all but two regions evaluated by day 9. Ciliastasis
followed mucostasis.
Morgan et al.
(1983)
Male F344
rats
6
0, 0.5, 2, 6, or 15 ppm
formaldehyde
6 ourshours/day for 1,
4, 9, or 15 days
Mucus flow and
ciliary beat and
histopathologic
analysis
Flow or ciliary beat were increased at 0.5 ppm.
AfterlAfter 1 day, slowed or halted mucociliary flow at 15 ppm after 6
hours.
After 9 days, slowed or halted mucociliary flow decreased or completely
stopped in all nasal regions evaluated.
Regions affected included lateral aspect of the nasoturbinate and dorsal and
medial aspects of maxilloturbinate.
Morgan et al.
(1986c)
Male F344
rats
3 per
group
15 ppm formaldehyde
for 10, 20, 45, or
90 minutes or 6 hours
Mucus flow and
ciliary beat
Ciliastasis and mucostasis increased in a time- and concentration-dependent
manner, with maximal response at 6 hours.
Significant recovery was observed when a 1-hour recovery period occurred
between exposure and sacrifice.
Morgan et al.
(1986a)
Isolated frog
palates
Not
stated
0.23, 1.37, 4.36, or
9.58 ppm
formaldehyde
Mucus flow rates
and histopathology
Ciliary beat and mucus flow increased from baseline at 1.37, 4.36, and
9.58 ppm.
Over time, mucus became rigid, and ciliastasis occurred
Morgan et al.
(1984a)
Isolated frog
palates
4
0, 1.25, 2.5, and 5 ppm
formaldehyde every 15
minutes for 60 minutes
Mucociliary
clearance and ciliary
beat
Ciliary beat decreased in a time-dependent manner at 2.5 and 5,0 ppm but
was increased at 1.25 ppm.
Mucostasis occurred after 60 minutes at 5 ppm.
Flo-Neyret et
al. (2001)
o
2 »
5 s
to
o
S
>S
TO
TO'
*
N = number of animals in study.
-j^
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
4.2.1.2.2.1. Nasal pathology short-term studies. Inhalation of formaldehyde for a few hours has
been shown to result in damage of the nasal mucosa, depending on the exposure concentration.
Bhalla et al. (1991) observed changes in cell morphology in male Sprague-Dawley rat nasal
epithelia after a single 4-hour exposure to 10 ppm (12.3 mg/m3) formaldehyde. Three exposed
rats were sacrificed 1 hour and 24 hours after exposure, with two control rats at each time point.
Noses were fixed, decalcified, and sliced along the midsagittal plane through the nasal septum.
The exposed turbinates were examined by scanning electron microscopy. Transverse sections
through the hard palate, at the level of the incisive papillae, were prepared for light microscopy
from similarly exposed rats (n = 10). The authors provided detailed descriptions of cell epithelial
organization in untreated rat turbinates and changes observed in formaldehyde-treated rats, as set
forth below. No statistical analysis was provided.
Scanning electron microscope examination of nasoturbinates showed increased mucus,
erythrocyte infiltration, swelling of microvillus cells, and some cell separation in formaldehyde-
treated rats. Nasoturbinates examined 1 day after exposure showed greater effects, including cell
damage, matted cilia, and blebbing of cell membranes. Damage to microvillus cells of the
maxilloturbinate included deformed cilia, cell swelling and rupture, and lack of typical microvilli
on the cell margins. As in the nasoturbinates, damage was more marked 24 hours after exposure.
The epithelium of the ETs exhibited less cell damage than in the nasal and maxillary regions,
with the slight lesions noted in the upper (ET1) portion and little to no damage noted on the mid
and lower (ET2 and ET3) regions. Examination of transverse tissue sections revealed swollen
goblet cells and stretched epithelial cells that formed an epithelial lining approximately 40%
taller than the lining seen in control rats. There was also a patchy loss of ciliated cells in the
respiratory epithelium, where columnar cells were present.
Buckley et al. (1984) investigated the respiratory tract lesions associated with several
sensory irritants. As part of this investigation, male Swiss-Webster mice were exposed to
3.13 ppm (3.85 mg/m3) formaldehyde 6 hours/day for 5 days. A total of nine chemicals were
tested in parallel. The report indicates there were 24-34 mice in each group, although not
detailed for each chemical. One-half of the treatment group and unexposed controls were
sacrificed immediately after the last exposure. The remaining exposed mice were sacrificed
72 hours later. The head, trachea, and lungs were fixed and heads decalcified. Five sections
were taken of each nose at levels equivalent to standard levels 2-6 (Figure 4-3) and were
examined by light microscopy. Details on lung and trachea sections were not given.
Formaldehyde induced lesions in the respiratory epithelium of exposed mice, including
inflammation, exfoliation, erosion, ulceration, necrosis, and squamous metaplasia. The section
This document is a draft for review purposes only and does not constitute Agency policy.
4-128 DRAFT—DO NOT CITE OR QUOTE
-------
1 level for these effects was not given. No effects were reported in the squamous epithelium,
2 olfactory epithelium, trachea, or lungs of formaldehyde-exposed mice.
3 Monteiro-Riviere and Popp (1986) evaluated damage to the respiratory epithelium due to
4 acute formaldehyde exposures. Male F344 (CDF [F344]/CrlBr) rats (three to five per group)
5 were exposed at 0.5, 2.0, 6.0, or 15 ppm (0.62, 2.5, 7.4, or 18.5 mg/m3) formaldehyde
6 6 hours/day for either 1, 2, or 4 days. Rats were sacrificed either immediately after exposure or
7 18 hours later (Table 4-10). After fixation and decalcification, blocks of tissue were collected
8 from transverse sections of the skull. The first block of tissue, 1 |im thick, was taken just
9 posterior of the incisor teeth. The second block was taken halfway between the first block and
10 the incisive papillae. The dorsal nasal conchae, lateral wall, and ventral nasal conchae were
11 microdissected, postfixed, and viewed by transmission electron microscopy (Monteiro-Riviere
12 and Popp, 1986).
13
14 Table 4-10. Concentration regimens for ultrastructural evaluation of male
15 CDF rat nasoturbinates
16
Formaldehydea'b
Duration
Time of sacrifice
Observations
0.5 ppm (3)
6 hours for 1 day
6 hours for 4 days
18 hours later
No lesions.
Altered ciliary configuration.
2.0 ppm (3)
6 hours for 1 day
6 hours for 4 days
18 hours later
No lesions.
Altered ciliary configuration.
6 ppm
(5 each group)
6 hours for 1 day
6 hours for 1 day
6 hours for 2 days
6 hours for 4 days
Immediately
18 hours later
Focal lesions on dorsal and nasal conchae and
lateral wall.
Severity of lesions increased with exposure
duration.
15 ppm
(5 each group)
6 hours for 1 day
6 hours for 2 days
18 hours later
Focal lesions on dorsal and nasal conchae and
lateral wall.
Severity of lesions increased with exposure
duration.
Severity of lesions increased with concentration.
17
18 aNumber of exposed rats is shown in parentheses.
19 bFive control rats were examined for each experiment.
20 Source: Monteiro-Riviere and Popp (1986).
21
22 No lesions were observed at either 0.5 or 2.0 ppm formaldehyde for either 1 day or
23 4 days, evaluated 18 hours after exposure. However, an unusual altered ciliary configuration,
24 including blebbing of the cell membrane, was observed in almost all formaldehyde-treated rats,
25 whereas it was only "occasionally noted" in control rats. Focal lesions in the dorsal and ventral
26 conchae and lateral wall were seen in rats exposed at 6 and 15 ppm for 1 day and sacrificed
27 immediately after exposure. These lesions included cytoplasmic and autophagic vacuoles, loss
28 of microvilli, and hypertrophy. Lesions increased in severity with both exposure concentration
This document is a draft for review purposes only and does not constitute Agency policy.
4-129 DRAFT—DO NOT CITE OR QUOTE
-------
1 and duration. Neutrophil infiltration and intercellular edema were seen after 1 day at 6 and
2 15 ppm. Nonkeratinized squamous metaplasia was noted after 4 days at 6 and 15 ppm in treated
3 rats. Cell death and sloughing were noted after only 2 days of exposure at 6 ppm formaldehyde.
4 As described above, Cassee and Feron (1994) examined the effects of intermittent
5 exposure to formaldehyde (3.5 ppm [4.3 mg/m3]), ozone (0.44 ppm [0.86 mg/m3]), or a
6 combination of the two on changes to the rat nasal epithelium. Exposure occurred through six
7 consecutive 12-hour cycles in which rats were exposed for 8 hours and then not exposed for a
8 further 4 hours. Rats were weighed before the first and after the last exposure periods and
9 sacrificed immediately after the last exposure. To collect tissue for biochemical analysis, skulls
10 were split sagittally and the respiratory epithelium collected. Tissues from six rats were pooled
11 and homogenized to enable the measurement of glutathione (GSH) and the activities of the
12 following enzymes: glutathione S-transferase (GST), glutathione peroxidase (GPX), glucose-6-
13 phosphate dehydrogenase (G6PDH), glutathione reductase (GR), alcohol dehydrogenase (ADH),
14 and formaldehyde dehydrogenase (FALDH). The remaining heads were fixed, decalcified, and
15 sectioned (standard cross sections [Figure 4-3]).
16 All groups, including controls, lost weight during the course of treatment. Rats exposed
17 to formaldehyde, ozone, or both lost more weight than controls (p < 0.05, p< 0.01, and
18 p< 0.001, respectively). Formaldehyde treatment alone increased GPX from 48.6 to
19 64.0 |imole/minute-mg protein (p < 0.05) (Table 4-11). Formaldehyde exposure, in conjunction
20 with ozone, decreased GST from 490 to 389 |imole/minute-mg protein (p < 0.05). No other
21 enzyme activities or tissue GSH levels were affected by formaldehyde exposure.
22
23 Table 4-11. Enzymatic activities in nasal respiratory epithelium of male
24 Wistar rats exposed to formaldehyde, ozone, or both
25
Enzyme
Controls"
Formaldehyde
(3.5 ppm)
Ozone (0.4 ppm)
Bothb
ADH
2.66 (0.99)
3.53 (0.13)
3.40 (0.33)
2.42 (0.61)
GST
490 (32)
494 (24)
514(4)
389 (28)°
GPX
48.6 (4.3)
64.0 (7.9)°
55.6 (2.0)
54.5 (0.3)
G6PDH
58.9 (7)
60.8 (4.7)
65.8(1.0)
45.5 (6.8)
GR
275 (16)
288.2 (16)
279 (17)
236 (14)
FALDH
0.77 (0.03)
0.68 (0.04)
0.68 (0.07)
0.80 (0.08)
26
27 aValues shown are the means and SDs of three measurements of a pooled sample. Units are
28 |imolc/minutc/mg of cytosolic protein.
29 bRats were exposed intermittently, 12-hour cycles of 8 hours exposed and 4 hours unexposed, for 3 days.
30 Different from control, p < 0.05.
31
32 Source: Cassee and Feron (1994).
33
This document is a draft for review purposes only and does not constitute Agency policy.
4-130 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Formaldehyde-exposed rats exhibited lesions in the nasal epithelium at levels 2 and 3 of
the nose, with effects slightly more severe in level 2. Lesions observed include necrosis,
hyperplasia accompanied by squamous metaplasia, and rhinitis. Exposure to formaldehyde in
the presence of ozone resulted in more severe squamous metaplasia (statistics not given). These
findings are similar to those of Monteiro-Riviere and Popp (1986), indicating that single or
repeated exposures can result in cell damage and death. Cell death and increased cell
proliferation were seen here after 3 days of repeated exposures to 3.5 ppm formaldehyde. While
no increases were seen in olfactory epithelium, frank necrosis, squamous metaplasia, and
hyperplasia of both ciliated and nonciliated epithelium were noted at level 2 and 3.
Javdan and Taher (2000) exposed male and female albino Wistar rats (five/group) at 0, 2,
or 5 ppm (0, 2.5, or 6.2 mg/m3) formaldehyde 8 hours/day for either 3 or 30 days. Transverse
tissue sections at the base of the incisive teeth and the first palatine folds were examined by light
microscopy. Lesions reported after 3 days of exposure to 2 ppm formaldehyde included chorion
congestion, cell disarrangement, squamous hyperplasia, atypical mitosis, and epithelial
hyperplasia. Similar lesions were seen after 30 days but were more severe. Effects at 5 ppm
formaldehyde included goblet cell proliferation, olfactory epithelial hyperplasia, calcified
regions, and an abscess on the chorion. These lesions were more severe after 30 days of
exposure.
Kamata et al. (1996a, b) conducted several high-dose studies by inhalation in rats.
Specifically they exposed male F344 rats to 0, 128.4, or 294.5 ppm (0, 158, or 362 mg/m3)
formaldehyde for 6 hours (Kamata et al., 1996a). In a subsequent study in the same laboratory,
male F344 rats were exposed to either 0, 15, or 145 ppm (0, 18.5, or 178 mg/m3) formaldehyde
nose only for 6 hours (Kamata et al., 1996b). Congestion was noted in the nasal cavities of
formaldehyde-exposed rats and was more severe at 145.6 ppm (Kamata et al., 1996b). Rats
exposed to 15 ppm formaldehyde had lesions in the nasal turbinate and trachea (not detailed)
(Kamata et al., 1996b). A slight hypersecretion of mucus was noted in the tracheal epithelium in
the absence of histopathologic changes. Rats exposed to 145.6 ppm had more dramatic lesions
that penetrated more deeply into the respiratory tract. Hyperkeratosis of the squamous
epithelium was found at level 1 of the nasal cavity. Hypersecretion, desquamation, and irregular
mucosal epithelium were seen in levels 2, 3, 4, and 5 of the nasal cavity, with more severe
changes noted in the nasal septum. Increased secretion and desquamation of mucosal cells
occurred in the trachea, and a slight hyperplasia of the alveolar wall was noted in rats exposed to
145.6 ppm formaldehyde (Kamata et al., 1996b)
Hester et al. (2003) carried out a transcriptional analysis of the nasal epithelium of male
F344 rats 24 hours after nasal instillation of 40 |iL of 400 mM formaldehyde. Immediately after
This document is a draft for review purposes only and does not constitute Agency policy.
4-131 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
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
sacrifice, cells were harvested from the nasal cavity for RNA extraction. The authors found
several phase I and II enzymes, indicative of oxidative stress, to be elevated. They also reported
the greatest increase in inflammatory genes, such as iNOS and neuropeptides. In an effort to
phenotypically link any gene changes to pathology, Hester et al. (2003) noted that this exposure
scenario has been demonstrated to induce regenerative hyperplasia with minimal cytotoxicity. In
this regard, they observed no significant change in nine genes involved in three apoptotic
pathways.
In an expansion of their earlier study, Hester et al. (2005) carried out a transcriptional
analysis of the nasal epithelium of male F344 rats that had been exposed to formaldehyde by
nasal instillation for a single exposure, 5 days of exposures, or 28 days of exposure. In addition,
this study also attempted to characterize the comparative toxicity of glutaraldehyde with
structurally similar formaldehyde (van Birgelen et al., 2000). Thus, four animals per group were
instilled with 40 |iL of deionized water (control group), 40 |iL of 400 mM formaldehyde, or
40 |iL of 20 mM glutaraldehyde. Phenotypically, both aldehydes induced similar
histopathologic changes.
Both aldehydes induced similar changes in DNA repair and apoptotic pathways initially,
but the patterns of gene changes were different after about 5 days of exposure. Eight genes were
differentially expressed between formaldehyde and glutaraldehyde that indicated different
pathways for DNA repair, including recombination, base excision repair, and nucleotide excision
repair. Within this group, replication protein 70 and DNA excision repair ERCC1 showed a
twofold induction by formaldehyde compared with glutaraldehyde. Since both of these genes
and their products function by recognizing and removing damaged DNA bases, Hester et al.
(2005) hypothesized that formaldehyde-exposed cells may remove damaged bases more
efficiently than glutaraldehyde-exposed cells
4.2.1.2.2.2. Lune vatholoev: short-term studies. In addition to nasal pathology, several
researchers specifically investigated formaldehyde-induced effects in the trachea, bronchi, and
pulmonary tissues of the deep respiratory tract in a variety of species (Lino dos Santos Franco et
al., 2006; Kamata et al., 1996a, b; Schreibner et al., 1979; Ionescu et al., 1978).
Ionescu et al. (1978) described progressive damage in pulmonary tissue of adult male
rabbits exposed to an aerosol of 3% formaldehyde solution 3 hours/day for up to 50 days
(method of aerosol generation or particle size were not provided). An equivalent air
concentration was not reported and cannot be derived from the information given. Animals were
sacrificed at several time points (3, 7, 15, 20, 30, and 50 days), and fragments of the caudal lobes
of both lungs were taken to examine bronchi (intrapulmonary and distal) and lung parenchyma.
This document is a draft for review purposes only and does not constitute Agency policy.
4-132 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
Enzymatic activity was characterized in frozen sections for P-galactosidase, adenosine
triphosphatase (ATPase), adenosine monophosphatase (AMPase), lactate dehydrogenase (LDH),
malate dehydrogenase, succinate dehydrogenase (SDH), acid phosphatase, Tween-60 esterase,
naphthol-AS-D-acetate esterase, proline oxidase, hydroxyproline epimerase, leucyl
aminopeptidase, and P-glucuronidase. A portion of the lung was fixed and sectioned and viewed
by light microscopy to determine changes in cell populations and tissue pathology.
In addition, biochemical analysis revealed that enzymatic activity of P-galactosidase,
ATPase, AMPase, LDH, malate dehydrogenase, and SDH were all unchanged by formaldehyde
exposure across the course of treatment (Ionescu et al., 1978). The activities of several enzymes
were increased through the course of exposure, including acid phosphatase, Tween-60 esterase,
naphthol-AS-D-acetate esterase, proline oxidase, and hydroxyproline epimerase. Although no
details were reported, the authors described the changes as progressive, with the increase in
proline oxidase and hydroxyproline epimerase seen only in the second half of the treatment
course. The activities of two enzymes, leucyl aminopeptidase and P-glucuronidase, were
observed to decrease rapidly (time frame not provided) (Ionescu et al., 1978)
Histologic changes in the lung tissue were noted after only 3 days of exposure and were
generally progressive throughout the course of treatment. Early changes in the bronchial
epithelium included increased mucus secretion, hyperplasia, and hypertrophy of epithelial cells.
Lymphocyte infiltration was noted in many areas, and a limited thickening of the alveolar walls
was reported after 3 days of exposure. Epithelial cell lesions, thickening of the alveolar, and
infiltration of lymphocytes increased as exposure continued. Mucus cells increased as much as
40% after 40 days of treatment. After 40 days of treatment, Ionescu et al. (1978) observed
"destructive and fibrotic lesions" and provided a detailed description of progressive lesions.
Schreiber et al. (1979) also examined histologic changes in lung tissue after high
formaldehyde exposures. Syrian golden hamsters (34, sex not stated) were exposed to 250 ppm
(308 mg/m3) formaldehyde 1 hour/day for 1, 2, 5, or 15 days. Five hamsters in each treatment
group were sacrificed 2 days after exposure was ended. Three hamsters in each group were
sacrificed 1, 2, or 6 weeks after exposure ended to determine if formaldehyde-induced changes
regressed over time. Tracheal washing was carried out to collect cytologic samples in each
animal prior to sacrifice. Samples were fixed, stained, and examined by light microscopy.
Lungs and tracheae were removed en bloc and fixed, and 20, 1 |im thick cross sections were
taken (location not detailed). The remaining respiratory tissue was sectioned at 200 |im
intervals. Sections were stained and viewed by light microscopy.
Abnormal epithelial cells were found in tracheal washings from formaldehyde-exposed
hamsters. Schreibner et al. (1979) described cells with lobulated nuclei and a coarse chromatin
This document is a draft for review purposes only and does not constitute Agency policy.
4-133 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
pattern, especially in cells showing signs of degeneration (e.g., vacuolization of nuclei and
cytoplasm) (Schreiber et al., 1979). Cell number and damage were not quantified, and there was
no discussion of the effects of exposure duration on treatment, if any, on these observations.
Tracheal washing was normal 2 and 6 weeks after the end of exposure, indicating that the
cytological changes were reversible (Schreiber et al., 1979).
Formaldehyde exposure caused multifocal lesions in the mucociliary epithelium in the
trachea and larger bronchi. Dysplastic and poorly differentiated squamous metaplastic foci
replaced ciliated epithelium (Schreiber et al., 1979). Abnormal nuclear membranes, tonofibrils
around the nuclei, the appearance of nucleoli, and heterochromatin condensation were distinct in
the formaldehyde-treated hamsters. These changes, observed 2 days after formaldehyde
exposure, were reversible over time and not seen 2 and 6 weeks later.
Because of the similarity of form and physiology of rhesus monkey URTs to the human
respiratory tract, the effects of short-term formaldehyde exposure were evaluated in both nasal
and lung tissue in these monkeys by Monticello et al. (1989). Male rhesus monkeys (nine/group)
were exposed to 6 ppm formaldehyde (7.4 mg/m3) 6 hours/day for 5 days/week for either 1 or
6 weeks. Control animals were exposed to the same regimen of filtered air for 1 week. Monkeys
were weighed during the course of exposure and observed for clinical signs of irritation or
sickness. Monkeys were intravenously injected with [3H]-thymidine 18 hours after the last
formaldehyde treatment to evaluate induced cell proliferation. Sections of the nasal passages,
trachea, larynx, lung carina, and duodenum were processed for histoautoradiography. Tissues
fixed and sectioned for examination by light microscopy included nose, adrenal, sternum (bone
marrow), duodenum, esophagus, eyes, gallbladder, heart, kidney, liver, lymph nodes, pancreas,
stomach, spleen, and tongue. The nose was cut into a series of transverse sections, 3 |im thick,
and sections from five levels were examined (Figure 4-6). Lung lobes were trimmed
midsagittally and sectioned with care to include airway bifurcations. Sections of the nasal
passages, trachea (cross section), larynx (cross section), lung carina (frontal section), and
duodenum were also processed for histoautoradiography.
There were no significant changes in body weight over the course of the experiment.
Oronasal breathing was noted in the first 15 minutes of formaldehyde exposure (Monticello et
al., 1989). Monkeys did experience eye irritation (mild lacrimation and conjunctival hyperemia)
during exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
4-134 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 4-6. Diagram of nasal passages, showing section levels chosen for
morphometry and autoradiography in male rhesus monkeys exposed to
formaldehyde.
Source: Redrawn from Monticello et al. (1989).
Formaldehyde-related lesions were reported in the nasal passages, tracheas, and in the
larynx of treated animals (Figure 4-7) (Monticello et al., 1989). Nasal epithelium from treated
animals exhibited many of the histologic lesions described in rodent studies, including loss of
goblet cells, loss of cilia, epithelial hyperplasia, squamous metaplasia, and neutrophilic
inflammatory response in the respiratory epithelium. The lesions were more severe after
6 weeks of exposure and were present over a greater percentage of the epithelium compared with
the 1-week exposure group (p < 0.05) (Figure 4-7).
This document is a draft for review purposes only and does not constitute Agency policy.
4-135 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
25r
20
15
x
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
formaldehyde exposure (Figure 4-8) (Monticello et al., 1989). Similar trends were seen after
only 1 week but were statistically significant only in the respiratory epithelium. Although
increased proliferation in the trachea and carina was statistically significant after 1 week of
exposure, the greater increases seen after 6 weeks of exposure, compared with controls, were not
statistically significant. A small sample size (n = 3) and high variability may have contributed to
the lack of statistical significance. Monticello et al. (1989) noted that increased cell proliferation
was seen in locations with minimal histologic changes, indicating proliferation may be a more
sensitive predictor of adverse health effects of formaldehyde exposure.
A)
w
LU
oc
<
LU
o
<
LL
ir
3
CO
60
50 ¦
40-
30-
20
10-
0-
1
1
B C D E L/T
LEVEL
ZZ] 1 WK EXP B)
6 WK EXP 60
50
40
30
20
10
0
[±l
B C D E L/T
LEVEL
Figure 4-8. Formaldehyde-induced lesions in male rhesus monkeys exposed
to formaldehyde.
Note: Animals were exposed to 6 ppm formaldehyde 6 hours/day, 5 days/week
for 1 or 6 weeks. Bar graph showing levels B-E of the nasal passages and the
larynx/trachea (L/T), depicting percent surface area with formaldehyde-induced
lesions. Morphometry of level A was excluded due to the similarity of normal
features of transitional epithelium to formaldehyde-induced lesions in the
respiratory epithelium. A: One-week exposure group. B: Six-week exposure
group.
* Statistically different from controls (p < 0.05).
| Statistically different from 1-week exposure group (p < 0.05).
Source: Redrawn from Monticello et al. (1989).
There are two reports in the literature assessing changes in pulmonary tissues after acute
formaldehyde exposures (Kamata et al., 1996a, b). Kamata et al. (1996a) exposed male F344
This document is a draft for review purposes only and does not constitute Agency policy.
4-137 DRAFT—DO NOT CITE OR QUOTE
-------
1 rats to 0, 128.4, or 294.5 ppm (0, 158, or 362 mg/m3) formaldehyde for 6 hours. Lung lavage
2 samples were collected and the fluid analyzed for the lipids, free cholesterol, phosphatidyl
3 ethanolamine, phosphatidyl choline, sphingomyelin, and triglyceride.
4 The bronchoalveolar lavage (BAL) was analyzed for triglycerides, cholesterol, and
5 phosphatidylcholine. As in the first experiment (Kamata et al., 1996a), triglyceride
6 concentration was reduced in the lavage of treated animals, in this case, to 16% of controls in
7 lavage in those rats exposed to 145.6 ppm formaldehyde (Table 4-12). Cholesterol concentration
8 was unchanged and phosphatidyl choline was increased to 220% of that of control rats as a result
9 of exposure to 145.6 ppm formaldehyde. However, BAL lipids were unchanged in 15 ppm
10 exposed rats. Triglycerides were reduced in unwashed lung tissue from formaldehyde-treated
11 rats in a concentration-dependent manner and free fatty acids were reduced in rats exposed to
12 145.6 ppm formaldehyde. Neither triglyceride nor sphingomyelin was detected in lung lavage
13 fluid from the high treatment group.
14
15 Table 4-12. Lipid analysis of lung tissue and lung lavage from male F344
16 rats exposed to 0,15, or 145.6 ppm formaldehyde for 6 hours
17
Control"
15 ppma
145 ppma
Lung tissue
Free fatty acids (mg/g lung)
3.30 (0.7)
3.11 (1.23)
1.41 (0.63)b
Triglyceride (mg/g lung)
1.55 (0.23)
0.74 (0.14)°
0.62 (0.17)°
Cholesterol (mg/g lung)
1.72 (0.10)
1.41 (0.25)
1.16(0.55)
Phosphatidyl ethanolamine (mg/g lung)
7.41 (1.81)
7.46 (2.28)
5.49 (1.78)
Phosphatidyl choline (mg/g lung)
11.0(1.49)
9.65 (3.21)
7.53 (3.52)
Sphingomyelin (mg/g lung)
3.44 (0.75)
3.13 (1.28)
2.51 (0.95)
Lung lavage
Triglyceride (mg/lung)
0.31 (0.10)
0.24 (0.09)
0.05 (0.02)°
Cholesterol (mg/lung)
0.04 (0.01)
0.04 (0.01)
0.04 (0.01)
Phosphatidyl choline (mg/lung)
0.66 (0.23)
0.84 (0.35)
1.45 (0.31)°
18
19 aSD given in parentheses.
20 Significant difference from controls (p < 0.05).
21 Significant difference from controls (p < 0.01).
22
23 Source: Kamata et al. (1996b).
24
25 Concentration-dependent decreases were seen in nonprotein sulfhydryl (SH) groups and
26 lipooxygenase in nasal mucosa homogenate and nonprotein SH groups in lung tissue
27 homogenate (Table 4-13). Increases in both lipooxygenase and LDH activities were found in
28 lung tissue homogenate from formaldehyde-exposed rats.
29
This document is a draft for review purposes only and does not constitute Agency policy.
4-138 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
Table 4-13. Formaldehyde effects on biochemical parameters in nasal
mucosa and lung tissue homogenates from male F344 rats exposed to 0,15,
or 145.6 ppm formaldehyde for 6 hours
Control"
15 ppma
145 ppma
Nasal mucosab
Nonprotein SH groups (|iIVI/g tissue)0
1.64 (0.50)
1.29 (0.28)
0.73 (0.21)'
Lipid peroxides (|iIVI/g tissue)
118(23)
71 (16)'
59 (18)'
Glucose-6-dehydrogenase (U/g tissue)d
1.96 (0.10)
1.87 (0.07)
2.07 (0.13)
Lunge
Nonprotein SH groups (|iIVI/g tissue)0
1.83 (0.18)
1.70 (0.11)
1.29 (0.28)'
Lipid peroxides (|iM/g tissue)
72 (8)
95 (15)'
93 (8)g
Glutathione reductase (U/g tissue)d
0.42 (0.25)
0.25 (0.05)
0.22 (0.05)
Lactate dehydrogenase (U/g tissue)d
77.37 (9.28)
88.69 (7.66)
93.62 (4.99)'
'SD given in parentheses.
b5 or 10% nasal mucosa homogenates.
°nmol malonaldehyde/g tissue.
dUnits per gram tissue.
e20% lung homogenates.
Significant difference from controls (p < 0.01).
Significant difference from controls (p < 0.05).
Source: Kamata et al. (1996b).
Lino dos Santos Franco et al. (2006) studied the effects of inhaled formaldehyde on lung
injury and changes in airway reactivity in rats. The extent of local and systemic inflammation
was assessed by changes in leukocyte counts in BAL fluid, blood, bone marrow, and spleen.
Changes of reactivity of isolated tracheae and intrapulmonary bronchi in response to
methacholine were monitored in response to formaldehyde exposure. The authors exposed male
Wistar rats to formaldehyde generated from a 1% solution of formalin. However, they provided
insufficient information for the exposure concentration to be determined. Groups of six animals
were exposed to formaldehyde for either 0, 30, 60, or 90 minutes on 4 consecutive days. All
experiments were carried out 24 hours after the final exposure.
The authors reported a significantly increased number of leukocytes in the BAL fluid of
animals exposed to formaldehyde via inhalation. The effect reached a maximum for the longer
exposure duration (90 minutes). Compared with controls, rats exposed to formaldehyde
90 minutes/day for 4 days also displayed an increase in the number of total blood leucocytes
(1.4 ± 0.06 x 104 versus 0.8 ± 0.01 x io4 cells/mm3). These values are means ± standard error of
the mean (SEM) for six animals/group. The effect appeared to reflect changes in the
mononuclear cell population (1.1 ± 0.02 x 104 versus 0.6 ± 0.003 x 104 cells/mm3) rather than
peripheral blood neutrophils (0.2 ± 0.003 x 104 cells/mm3 in test animals and controls). There
was also an apparently compound-related increase in the total cell count in the spleen
This document is a draft for review purposes only and does not constitute Agency policy.
4-139 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
(112.7 ± 4.4 x 106 versus 94.2 ± 5.5 x 106 cells). However, a change in the number of cells
eluted from bone marrow did not reach statistical significance (54.6 ±1.3 x 106 versus
45.0 ± 4.3 x io6 cells). Lino dos Santos Franco et al. (2006) provided data on dose-dependent
changes in methacholine-induced contractions in isolated tracheae and bronchi obtained from
formaldehyde-exposed and control rats. Although the maximal contractile response induced by
methacholine in tracheae of formaldehyde-treated rats was unchanged compared with controls,
contractions in isolated bronchi were significantly weaker than those observed in controls.
The authors examined the effect of formaldehyde inhalation on rat lung mast cells.
Degranulation and significant neutrophil infiltration were features of the response to
formaldehyde (Table 4-14).
Table 4-14. Mast cell degranulation and neutrophil infiltration in the lung of
rats exposed to formaldehyde via inhalation
Treatment group
Mast cell degranulation
(cells/mm2)a
Neutrophil infiltration
(cells/mm2)a
Controls
0b
0.3 ±0.2
Formaldehyde-exposed
2.0 ± 0.4°
5.2 ± 1.7°
aValues are means ± SEM; n = 6.
b4.2 ± 0.6 cells/mm2 intact mast cells were found in the lungs of controls.
°No statistical analysis was provided by the authors for these changes.
Source: Lino dos Santos Franco et al. (2006).
Selected pharmacological agents were used to explore the mechanism by which exposure
to formaldehyde might have brought about the observed lung infiltration and bronchial
hyporesponsiveness. Lino dos Santos Franco et al. (2006) provided data showing that separate
pretreatment of the animals with compound 48/80, sodium cromoglycate (SCG), and
indomethacin reduced the formaldehyde effect on neutrophil release into BAL but had no effect
on mononuclear cell counts. Compound 48/80 and SCG also reversed the formaldehyde-induced
reduction in bronchial response to methacholine, but indomethacin had the opposite effect
(causing an additional decrease in bronchial responsiveness). In broad terms, these findings
were thought to implicate mast cells as a possible mediator of the toxicological effects of
formaldehyde. Histologically, a significantly increased number of degranulated mast cells were
evident in the pulmonary tissue of rats that were exposed to formaldehyde.
Lino dos Santos Franco et al. (2006) also examined the regulatory role of NO on
formaldehyde-induced bronchial activity. Nitrites generated by cultured cells of BAL from
formaldehyde-treated rats increased about threefold compared with those from controls.
This document is a draft for review purposes only and does not constitute Agency policy.
4-140 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
31
32
33
34
35
However, pretreatment with the NO synthase inhibitor, N-nitro-L-arginine methyl ester,
prevented the formaldehyde-induced bronchial hyporesponsiveness to methacholine but had no
effect on pulmonary leukocyte recruitment. These data implicate the existence of distinct
mechanisms for the induction of lung inflammation versus bronchial hyporeactivity. Further
support for this concept came from an experiment in which rats were pretreated with capsaicin to
examine the involvement of sensory fibers in lung inflammation and the bronchial
hyporesponsiveness induced by formaldehyde inhalation. Although the treatment did not
influence formaldehyde-induced bronchial hyporesponsiveness to methacholine, the number of
leukocytes recovered in the BAL fluid were reduced compared with those of rats exposed to
formaldehyde alone.
4.2.1.2.2.3. Extrapulmonary effects: short-term studies. Kamata et al. (1996a) exposed male
F344 rats to 0, 128.4, or 294.5 ppm (0, 158, or 362 mg/m3) formaldehyde for 6 hours. In
addition, blood samples were monitored for hematology and clinical chemistry parameters,
including red blood cell (RBC) count, hemoglobin (Hb), packed cell volume (PCV), mean
corpuscular volume (MCV), mean corpuscular hemoglobin concentration (MCHC), white blood
cell (WBC) count, and plasma levels of total protein (TP), albumin (ALB), blood urea nitrogen
(BUN), glucose, phospholipids, triglycerides, total cholesterol, cholinesterase, and LDH. Male
rats exposed to 294.5 ppm formaldehyde had increased RBC count, Hb, hematocrit (HCT),
MCV, and serum glucose (p < 0.05) compared with controls (Kamata et al., 1996a). There were
concentration-related decreases in serum measures of TP, ALB, and phospholipids (p < 0.05).
BUN was decreased in rats exposed to 128.4 ppm but increased in the higher treatment group
(p < 0.05). Phospholipid analysis of the lung surfactant indicated a decrease in the production in
formaldehyde-treated animals (p < 0.05). Total free cholesterol, phosphatidyl ethanolamine, and
phosphatidyl choline were reduced to 60, 55, and 38% of controls for rats treated with 294.5 ppm
formaldehyde (p < 0.05). Sphingomyelin was reduced to 32% of controls in the low treatment
group (p < 0.05).
In a subsequent study in the same laboratory (Kamata et al., 1996b), male F344 rats were
exposed to either 0, 15, or 145 ppm (0, 18.5, or 178 mg/m3) formaldehyde nose only for 6 hours
(Kamata et al., 1996b). Fifteen animals were treated at each level and separated into subgroups
of five animals each for tissue collection and the determination of other endpoints. Blood
samples were collected from one subgroup to determine such hematological and clinical
chemistry parameters as RBC count, Hb, PCV, MCV, MCHC, WBC count, and plasma levels of
TP, ALB, BUN, glucose, phospholipids, triglycerides, total cholesterol, LDH, alkaline
phosphatase (ALP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and
This document is a draft for review purposes only and does not constitute Agency policy.
4-141 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
G6PDH. BAL was collected from five animals of each group and analyzed for phospholipids.
Lung homogenate from five animals in each treatment group was analyzed for nonprotein SH
groups, lipid peroxides, and total lipids. The 20,000 x g supernatant of the lung homogenate was
assayed for the activities of GR, G6PDH, and LDH. Similarly, nonprotein SH groups and lipid
peroxidase were measured in homogenates of excised nasal mucosa. At autopsy, organs (brain,
heart, lung, liver, kidney, spleen, and testis) were weighed and tracheae and nasal turbinates
examined. After fixation and decalcification, five sections across the nose were taken,
corresponding to standard sections 1-5 (Figure 4-3).
Several blood parameters were affected after these acute exposures. The WBC count was
slightly increased, from 4.7 x 103 cells/mm3 in control rats to 5.1 x 103 cells/mm3 and 6.1 x 103
cells/mm3 at 15 and 145.6 ppm formaldehyde, respectively (Kamata et al., 1996b). Serum levels
of AST and LDH decreased in an apparent concentration-dependent manner (AST 68 and 54%
of controls and LDH 48 and 28% of controls, respectively). Serum levels of G6PDH and ALT
were decreased similarly across exposure groups at 45 and 78% of controls, respectively.
A synopsis of respiratory pathology findings following short-term exposure to
formaldehyde is presented in Table 4-15.
4.2.1.2.2.4. Nasal pathology: subchronic studies. In a study by Maronpot et al. (1986), female
and male B6C3F1 mice (10/group) were exposed at 0, 2, 4, 10, 20, or 40 ppm (0, 2.46, 4.92,
12.3, 24.6, or 49.2 mg/m3) formaldehyde 6 hours/day, 5 days/week for 13 weeks. Clinical
observations were made daily, and mice were weighed weekly. At autopsy, tissue sections from
each organ system (approximately 50 tissues per mouse) were fixed, stained, and examined by
light microscopy. Noses were fixed, decalcified, and transversely trimmed at three levels: the
incisor teeth, midway between the incisor teeth and first molar teeth, and the second molar teeth
(corresponding to sections 2, 3, and 4 in Figure 4-3).
Although control mice gained weight, mice exposed to 40 ppm formaldehyde lost weight
during the 13-week exposures. Expressed by the authors as a percent of weight gain in controls,
the weight losses were —235% in males and —168.6% in females. Early mortality for both male
and female mice exposed to 40 ppm was 80%. Although gross and histochemical effects in
excised pieces from each organ system were evaluated, endometrial hypoplasia in mice treated
with 40 ppm was the only effect noted outside the respiratory system. The authors considered
this effect secondary to the observed respiratory tract lesions and frank toxicity at 40 ppm
formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
4-142 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-15. Summary of respiratory tract pathology from inhalation exposures to formaldehyde—short-term studies
S"4
>3*
s
Species/strain
No./
group
Treatment
Respiratory effects
LOAEL/NOAEL
Reference
Nasal pathology
Male Sprague-
Dawley rats
3
Single 4-hour exposure to
10 ppm formaldehyde.
Marked histopathologic changes to the nasoturbinates,
maxilloturbinates, ethmoidal turbinates, and goblet and
microvillus cells.
LOAEL =10 ppm.
Bhalla et al.
(1991)
Male Swiss-
Webster mice
24-34
0 or 3.13 ppm formaldehyde
6 hours/day for 5 days.
Histopathologic lesions to the respiratory epithelium,
including inflammation, exfoliation, erosion,
ulceration, necrosis, and squamous metaplasia.
LOAEL = 3.13 ppm.
Buckley et al.
(1984)
Male F344 rats
3-5
0, 0.5, 2, 6, or 15 ppm
6 hours/day for 1, 2, or 4 days.
Histopathologic lesions to the nasal conchae, lateral
wall, and ventral nasal conchae.
NOAEL = 2 ppm for
focal lesions. Some
changes in ciliary
configuration were
evident at all exposures.
Monteiro-Riviere
andPopp (1986)
Male Wistar rats
20
0 or 3.5 ppm formaldehyde
through six consecutive 12-hour
cycles in which rats were
exposed for 8 hours; 10 were
unexposed for 4 hours.
The activity of GPX was increased in respiratory
epithelium homogenates. The nasal respiratory
epithelium showed frank necrosis.
LOAEL = 3.5 ppm.
Cassee and Feron
(1994)
Male F344 rats
5
0, 6, or 15 ppm
[14C]-formaldehyde 6 hours/day
for a single day (naive group).
A pretreated group was exposed
to 6 or 15 ppm formaldehyde
6 hours/day for 4 days prior to
[14C| -formaldehyde exposure.
Cellular necrosis to the nasal epithelium. 10.05%
cellular proliferation.
LOAEL = 6 ppm.
Chang et al.
(1983)
Male and female
Wistar rats
5/sex
0, 2, or 5 ppm formaldehyde
8 hours/day for 3 or 30 days.
Cell disarrangement, squamous hyperplasia, atypical
mitosis, and epithelial hyperplasia.
NOAEL = 2 ppm.
Javdan and Taher
(2000)
Male F344 rats
15
0, 15, or 145.6 ppm
formaldehyde for a single 6-hour
exposure.
Histopathologic lesions in the nasal turbinates and
trachea
LOAEL =15 ppm.
Kamata et al.
(1996b)
Male rhesus
monkeys
9
0 or 6 ppm formaldehyde
6 hours/day for 1 or 6 weeks.
[3H] -thymidine was injected
prior to sacrifice.
Histopathologic lesions, including loss of goblet cells,
loss of cilia, epithelial hyperplasia, squamous
metaplasia, and neutrophilic inflammation.
LOAEL = 6 ppm.
Monticello et al.
(1989)
o
S ?
a, Co
TO Sj-
§ ^
s ^
§ 3
?»»i. ~-
4^
U>
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-15. Summary of respiratory tract pathology from inhalation exposures to formaldehyde—short-term studies
Species/strain
No./
group
Treatment
Respiratory effects
LOAEL/NOAEL
Reference
Tracheal and lung pathology
Syrian golden
hamsters (sex
unstated)
5
0 or 250 ppm 1 hour/day for 1,
2, 5, or 15 days.
Abnormal cells in tracheal lavage, an effect that was
reversed on cessation of treatment.
LOAEL = 250 ppm.
Schreiber (1979)
Male rabbits
(strain unstated)
ND
Aerosol generated from a 3%
formaldehyde solution
3 hours/day for up to 50 days
(air concentration unknown).
Necrosis of the bronchi and lung parenchyma.
Increased activities of acid phosphatase, Tween-60
esterase, naphthol-AS-D-acetate esterase, proline
oxidase, and hydroxyproline epimerase. Reduced
activities of leucyl aminopeptidase and (3-
glucuronidase. Adverse histopathologic changes.
ND.
Ionescu et al.
(1978)
Male F344 rats
6
0, 128.4, or 294.5 ppm for a
single 6-hour exposure.
Phospholipid content was reduced in lung surfactant,
for example, sphingomyelin to 43% of controls in the
low-concentration group.
LOAEL = 128.4 ppm.
Kamata et al.
(1996a)
Male F344 rats
15
0, 15, or 145.6 ppm
formaldehyde for a single 6-hour
exposure.
Biochemical changes in lung homogenates. Altered
lipid content of BAL in high concentration rats.
LOAEL =15 ppm.
Kamata et al.
(1996b)
Male Wistar rats
6
Aerosol generated from a 1%
formalin solution 0, 30, 60, or
90 minutes/day on 4 consecutive
days (air concentration
unknown).
Increased leukocyte count in bronchoalveolar fluid.
Degranulation of mast cells and increased neutrophil
infiltration.
ND.
Lino dos Santos
Franco et al.
(2006)
Extrapulmonary effects
Male F344 rats
6
0, 128.4, or 294.5 ppm for a
single 6-hour exposure.
LOAEL = 128.4 ppm.
Kamata et al.
(1996a)
Male F344 rats
15
0, 15, or 145.6 ppm
formaldehyde for a single 6-hour
exposure.
LOAEL =15 ppm.
Kamata et al.
(1996b)
o
2 »
5 s
a, Co'
TO Sj-
§ ^
>S
>S
TO
TO'
*
ND = not determined; LOAEL = lowest-observed-adverse-effect level; NOAEL = no-observed-adverse-effect level.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
While no statistical comparison was provided, respiratory tract lesions showed an
increased incidence with concentration, as well as an increased distribution throughout the
respiratory tract (Table 4-16). No lesions were seen in the nasal cavity, larynx, trachea, or lung
of control mice or mice treated with 2 ppm formaldehyde. Minimal squamous metaplasia in the
nasal cavity was noted in 1 of 10 male mice treated with 4 ppm formaldehyde, but none were
observed in the female mice. However, squamous metaplasia was observed in all mice in the
higher treatment groups (10, 20, and 40 ppm). Lesions became more severe and penetrated more
deeply into the respiratory tract as exposure concentration increased. Where lesions were present
in the nasal cavities of all mice exposed to 10 ppm, similar lesions were reported in the larynx
and trachea of some animals exposed to 20 ppm and all animals exposed to 40 ppm
formaldehyde. Mice exposed to 40 ppm formaldehyde exhibited lesions as deep as the lung,
including squamous metaplasia, submucosal fibrosis inflammation, and epithelial hyperplasia.
The findings of Maronpot et al. (1986) indicated a no-observed-adverse-effect level
(NOAEL) of 4 ppm and a LOAEL of 10 ppm in mice, based on squamous metaplasia in the
nasal epithelium. Although a LOAEL of 10 ppm was observed, there was 80% mortality for
both sexes at 40 ppm, indicating a very narrow range between the first observed adverse health
effects and frank effect concentrations in mice for this 13-week treatment.
In a study by Woutersen et al. (1987), male and female albino SPF Wistar rats (10/group)
were exposed to 0, 1, 10, or 20 ppm (0, 1.23, 12.3, or 24.6 mg/m3) formaldehyde 6 hours/day,
5 days/week for 13 weeks. Rats were checked daily and weighed weekly. Three longitudinal
sections of lungs, trachea, and larynx and six standard cross sections of the nose were taken for
microscopic examination. Two rats per exposure group were similarly treated for 3 days and
sacrificed 18 hours later, and nasoturbinates were dissected to measure cell proliferation.
Woutersen et al. (1987) noted that the majority of the dose-dependent increases in cell
proliferation seen at section level 3 after 3 days of repeated 6-hour exposures to 10 and 20 ppm
(12.3 and 24.6 mg/m3) formaldehyde occurred in areas of the epithelium showing "clear
squamous metaplasia and hyperplasia." Cell proliferation rates in metaplastic epithelium of
29.5 and 33.2% were much higher than the 1.4 to 2.8% proliferation in the visibly unaffected
respiratory epithelium from rats exposed at 10 ppm formaldehyde. Although there was a slight
trend towards increased cell proliferation in the visibly unaffected epithelium of exposed animals
compared with unexposed controls, the majority of increased cell proliferation resulting from
exposure to 10 and 20 ppm formaldehyde was attributed to the metaplastic epithelium
(Woutersen et al., 1987).
This document is a draft for review purposes only and does not constitute Agency policy.
4-145 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-16. Location and incidence of respiratory tract lesions in B6C3F1 mice exposed to
formaldehyde
Location of respiratory
tract lesions
Control
2 ppm
4 ppm
10
ppm
20
ppm
40
ppm
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Nasal cavity
Squamous metaplasia
Seropurulent inflammation
a
-
-
-
1/10
-
10/10
4/10
10/10
10/10
10/10
10/10
8/10
10/10
10/10
10/10
10/10
Larynx
Squamous metaplasia
6/9
3/9
10/10
7/8
Trachea
Squamous metaplasia
Epithelial hyperplasia
Seropurulent inflammation
Submucosal fibrosis
-
-
-
-
-
-
-
1/10
3/10
4/10
5/10
2/10
10/10
2/10
8/10
9/10
10/10
5/10
5/10
Lung (Bronchus)
Squamous metaplasia
Inflammation
Submucosal fibrosis
-
-
NDb
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-
-
-
-
4/10
3/10
2/10
3/10
2/10
o
2 »
5 s
to
o
S
>S
TO
TO'
*
aDash indicates no lesions recorded in that treatment group.
bND = no data.
Source: Maronpot et al. (1986).
On
-------
1
2
3
4
5
6
7
8
9
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
Statistically significant increases were seen in focal respiratory epithelial hyperplasia and
keratinization in both male and female rats at the highest treatment level (20 ppm) (Table 4-17).
Male rats also had statistically significant increases in observed respiratory epithelial squamous
metaplasia, focal olfactory epithelial thinning, and rhinitis. Both male and female rats treated
with 10 ppm formaldehyde showed statistically significant increases in squamous metaplasia,
hyperplasia, and keratinization of the respiratory epithelium (Woutersen et al., 1987).
Disarrangement of the respiratory epithelium was only significantly increased in female rats, but
this change was observed at both the 10 and 20 ppm treatment levels. Although some lesions
were observed in animals treated with 1 ppm formaldehyde, their incidences were not
statistically significant and the findings were equivocal.
Feron et al. (1988) examined recovery of formaldehyde-induced nasal lesions after
subchronic exposures. Male albino SPF Wistar rats (50-55/group) were exposed to 0, 10, or
20 ppm (0, 12.3, or 24.6 mg/m3) formaldehyde 6 hours/day, 5 days/week for either 4, 8, or
13 weeks. All groups were observed for a total of 130 weeks, including treatment and recovery.
Rats were weighed weekly for the first 13 weeks and monthly thereafter. Rats (five/group) were
sacrificed immediately after the end of exposure (4, 8, or 13 weeks). The balance of the rats
were sacrificed after 130 weeks, inclusive of exposure time. At sacrifice, noses were fixed and
sectioned by using standard section levels.
Formaldehyde exposure (20 ppm) was associated with reduced body weight throughout
the exposure period (4, 8, or 13 weeks). However, body weight in these groups matched that of
controls after 8, 40, and 100 weeks, respectively. Rats exposed to 10 ppm for 8 or 12 weeks had
slightly decreased body weight (further details not given).
Nonneoplastic lesions were reported in the nasal mucosa of rats exposed to either 10 or
20 ppm formaldehyde and examined immediately after exposure was discontinued (4, 8, or
13 weeks). Lesions increased in severity with both exposure duration and concentration (details
of severity and incidence were not provided). Rhinitis, hyperplasia, and squamous metaplasia of
the respiratory epithelium were seen in rats from both dose groups, but changes in olfactory
epithelia were only seen in rats exposed to 20 ppm, where cell disruption, thinning of the
epithelium, and simple cuboidal or squamous metaplasia were also reported. Changes in the
dorsomedial region, at the junction of the respiratory and olfactory epithelium, were similar to
those seen in the olfactory epithelium of rats exposed to 20 ppm formaldehyde. A similar
concentration- and duration-dependent increase in histopathologic changes in nasal epithelium
was observed after the full 130 weeks, which included 126, 122, or 117 weeks of recovery for
the three duration groups, 4, 8, and 13 weeks, respectively (Table 4-17).
This document is a draft for review purposes only and does not constitute Agency policy.
4-147 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-17. Formaldehyde effects (incidence and severity) on histopathologic
2 changes in the noses and larynxes of male and female albino SPF Wistar rats
3 exposed to formaldehyde 6 hours/day for 13 weeks
4
Concentration of formaldehyde (nnm)
0
1
10
20
0
1
10
20
Respiratory epithelium
Severity
Males
Females
Diffuse squamous
Slight
a
-
-
-
-
-
-
3
metaplasia
Moderate
-
-
-
5b
-
-
-
4
Severe
-
-
-
5b
-
-
-
3
Focal squamous
Very slight
-
1
-
-
-
-
1
-
metaplasia
Slight
-
1
6b
-
-
-
7°
-
Moderate
-
-
4
-
-
-
2
-
Focal hyperplasia
Very slight
-
-
1
1
-
-
2
1
Slight
-
-
6b
T
-
1
6b
6b
Moderate
-
-
1
-
-
-
-
-
Focal disarrangement
Very slight
-
-
1
-
-
-
2
1
Slight
-
-
3
-
-
1
6b
6b
Moderate
-
-
1
-
-
-
-
-
Focal keratinization
Very slight
-
2
6b
1
-
-
6b
6b
Slight
-
-
3
6b
-
-
2
4
Moderate
-
-
-
1
-
-
-
-
Olfactory epithelium
Focal thinning
Slight
-
-
-
2
-
-
-
2
Moderate
-
-
-
1
-
-
-
2
Severe
-
-
-
5b
-
-
-
2
Focal squamous
Slight
-
-
-
4
-
-
-
3
metaplasia
Moderate
-
-
-
4
-
-
-
1
Focal keratinization
Very slight
-
-
-
1
-
-
-
-
Slight
-
-
-
2
-
-
-
-
Rhinitis
-
2
5b
10c
-
-
3
2
Larynx
Squamous metaplasia
Very slight
-
-
-
3
-
NEd
NE
-
Slight
-
-
-
1
-
NE
NE
-
Moderate
-
-
-
1
-
NE
NE
-
Keratinization
Slight
- - - 2
-
NE
NE
-
5
6 aDash indicates no lesions reported.
7 bDifferent from control, p < 0.05.
8 Different from control, p < 0,01,
9 dNE = not evaluated.
10
11 Source: Woutersen et al. (1987).
12
13
14 Feron et al. (1988) did not provide a direct comparison among lesions reported at the
15 interim sacrifice and terminal sacrifice after the extended recovery period. However, similar
16 lesions were reported after the recovery period, including focal hyperplasia and stratified
17 squamous metaplasia of the respiratory epithelium, stratified cuboidal or squamous metaplasia in
18 the dorsomedial area, and replacement of olfactory epithelium. The incidence and severity of
This document is a draft for review purposes only and does not constitute Agency policy.
4-148 DRAFT—DO NOT CITE OR QUOTE
-------
1 these lesions in rats exposed to 20 ppm formaldehyde were statistically different from control
2 animals, regardless of exposure duration (Table 4-18).
3
4 Table 4-18. Formaldehyde-induced nonneoplastic histopathologic changes in
5 male albino SPF Wistar rats exposed to 0,10, or 20 ppm formaldehyde
6 (6 hours/day, 5 days/week) and examined at the end of 130 weeks inclusive of
7 exposure
8
4 Weeks
8 Weeks
13 Weeks
Formaldehyde, ppm
0
10
20
0
10
20
0
10
20
Total noses examined
44
44
45
45
44
43
45
44
44
Respiratory epithelium focal hyperplasia
Very slight
0
0
0
0
1
3
0
5a
2
Slight
0
3
8b
2
2
12b
1
6
14b
Moderate
0
0
1
0
1
0
0
0
4
Respiratory epithelium focal stratified squamous
metaplasia
Very slight
3
6
14b
8
16
17"
2
10a
2
Slight
4
2
19b
2
1
20b
3
18b
26b
Moderate
0
2
3
0
0
2
1
5
14b
Severe
0
0
0
0
0
0
0
0
1
Respiratory/olfactory epithelium stratified
0
0
4
0
0
17b
0
2
23b
cuboidal or squamous metaplasia
Rhinitis
7
7
18a
4
6
22a
8
11
23b
Olfactory epithelium replacement by respiratory
epithelium and regeneration
Very slight
0
0
0
0
0
2
0
0
1
Slight
1
0
6
0
0
14b
0
0
12b
Moderate
0
0
1
0
0
3
0
0
12b
Severe
0
0
0
0
0
1
0
0
1
9
10 aSignificantly different from control, p < 0.05.
11 bSignificantly different from control, p < 0.01.
12
13 Source: Feron et al. (1988).
14
15 Although a slight increase in changes to the olfactory epithelium and dorsomedial area
16 was seen in rats treated with 20 ppm formaldehyde for only 4 weeks, these differences were
17 significant and more severe in the 8- and 13-week treatment groups. Replacement of olfactory
18 epithelium by respiratory epithelium was described as slight after 8 weeks of exposure and slight
19 to moderate after 13 weeks of exposure in the 20 ppm treatment groups. Therefore,
20 formaldehyde-induced lesions were not resolved after a considerable nonexposure recovery
21 period of up to 126 weeks (Feron et al., 1988).
22 Feron et al., (1988) derived a correlation between the development of nonneoplastic
23 changes in nasal epithelium and the development of nasal tumors as a result of these subchronic
24 formaldehyde exposures. Two SCCs were reported in rats exposed to 10 ppm formaldehyde but
This document is a draft for review purposes only and does not constitute Agency policy.
4-149 DRAFT—DO NOT CITE OR QUOTE
-------
1 were not considered to be formaldehyde related because of their locations (nasolacrimal duct,
2 incisor tooth). Six tumors were observed in the 20 ppm, 13-week exposure group (Table 4-19)
3 of which three of the tumors were SCCs similar to those observed as a result of chronic
4 formaldehyde exposure. Two polypoid adenomas also were reported in rats exposed to 20 ppm
5 formaldehyde. Feron et al. (1988) concluded that subchronic exposures to 20 ppm formaldehyde
6 could result in an increase in nasal tumors, an effect that followed observation of cellular
7 proliferation.
8
9 Table 4-19. Formaldehyde-induced nasal tumors in male albino SPF Wistar
10 rats exposed to formaldehyde (6 hours/day, 5 days/week for 13 weeks) and
11 examined at the end of 130 weeks inclusive of exposure
12
Tumor type
0 ppm
10 ppm
20 ppm
No. of rats exposed for 4 weeks
44
44
45
Polypoid adenoma
0
0
la
see
0
0
1
No. of rats exposed for 8 weeks
45
44
43
Polypoid adenoma
0
0
la
sec
2
1
1
No. of rats exposed for 13 weeks
45
44
44
see
0
1
3a
Cystic squamous cell carcinoma
0
0
1
Carcinoma in situ
0
0
la
Ameloblastoma
0
0
1
13
14 aTumor considered to be associated with formaldehyde exposure.
15
16 Source: Feron etal. (1988).
17
18
19 A companion study from the same laboratory examined the effects of lower concentration
20 formaldehyde exposures (Zwart et al., 1988). Male and female albino Wistar rats (50/group)
21 were exposed to 0, 0.3, 1, or 3.0 ppm (0, 0.37, 1.2, or 3.7 mg/m3) formaldehyde 6 hours/day,
22 5 days/week for 13 weeks. Body weight, general condition, and behavior were recorded weekly.
23 No effects of formaldehyde exposure on body weight changes were noted, and growth was
24 considered comparable among different exposure groups and controls. Rats were sacrificed
25 during week 14, and noses were fixed and sectioned (exact time after exposure ended not given).
26 Six standard cross sections were examined for each animal by light microscopy, anterior to
27 posterior. Noses were fixed and decalcified, and six standard cross sections were taken and
28 developed.
29 No formaldehyde-related lesions were reported in the respiratory epithelium at section
30 level 3 after 13 weeks of formaldehyde exposure (0.1 ppm, 1 ppm, or 3 ppm). Signs of
This document is a draft for review purposes only and does not constitute Agency policy.
4-150 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
inflammation (rhinitis, sinusitis, mononuclear cell infiltrates) were observed in
formaldehyde-treated rats, but there was no concentration-response relationship (data not
provided). Formaldehyde-related pathology in the anterior part of level 2 epithelium was
reported in 37/50 males and 21/50 female rats exposed to 3.0 ppm for 13 weeks. Both
keratinized and unkeratinized squamous metaplasia were present, and disarranged cells and
hyperplastic respiratory epithelium were found in the transitional zone between squamous and
pseudostratified epithelium at level 2. Foci of keratinized squamous epithelium, glandularization
of goblet cells, and deciliated epithelium were observed by electron microscopy in anterior
sections of level 2 of rats exposed to 3 ppm formaldehyde. Epithelial cells with irregularly
shaped and strongly indented nuclei were described at level 2 in animals exposed to 0.3 and
1 ppm formaldehyde and were considered to be disarranged as well at 3 ppm formaldehyde
exposures.
Although early cell proliferation at level 3 corresponded to basal cell hyperplasia at
3 days, neither effect persisted for the course of the exposure. The authors speculate that this is
an indication of an adaptive response, perhaps through increased function of the mucociliary
apparatus present at level 3. In contrast, the early changes at section level 2 were less dramatic
but persisted through 13 weeks, including clear formaldehyde-related pathology.
Concentration times time (C x t) issues have been investigated for histopathology as well
as for cellular proliferation, outlined above. Specifically, Wilmer et al. (1989, 1987) compared
the effects of 8-hour continuous and 8-hour intermittent formaldehyde exposure in two studies.
Fifty male albino Wistar rats (10/group) were exposed to different exposure regimens to achieve
similar compound-related C x t products. A C x t product of 40 ppm-hours (49.2 mg/m3-hours)
was attained by an 8-hour exposure to 5 ppm (6.2 mg/m3) or a 4-hour exposure to 10 ppm (12.3
mg/m3) (Wilmer et al., 1987). Similarly, an 80 ppm-hours (98.4 mg/m3-hours) C x t product was
attained from continuous 10 ppm exposure or intermittent 20 ppm (24.6 mg/m3) exposure. Rats
were exposed to one of these regimens 8 hours/day for either 3 days (two/group) or 4 weeks
(eight/group). Eighteen hours after exposure ended, rats were injected with [3H]-thymidine and
sacrificed 2 hours later. Noses were fixed and decalcified, and six standard cross sections were
taken and developed.
Thinning and disarrangement of the respiratory epithelium, squamous metaplasia, basal
cell hyperplasia, and rhinitis were seen in formaldehyde-treated rats. Lesions were most severe
in group 4 (20 ppm intermittent). Groups 2 and 3 had similar lesions (10 ppm intermittent and
continuous). Rats in group 1 had mild lesions. Formaldehyde concentration was the major
determinate in severity of nasal lesions. Formaldehyde effects were less severe in group 1 than
in group 3, even though the C x t product was the same, indicating concentration rather than
This document is a draft for review purposes only and does not constitute Agency policy.
4-151 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
duration or cumulative exposure correlates to severity. Epithelial lesions in group 3 rats were
similar among rats exposed to 10 ppm, regardless of duration (groups 2 and 3).
In a follow-up study, Wilmer et al. (1989) assessed both cellular proliferation and
histologic lesions in Wistar rats exposed to formaldehyde in groups that differed by
concentration and time. Group A served as a control group (0 ppm). Group B was exposed to
1 ppm for 8 hours, group C to 2 ppm for 8 hours, group D to 2 ppm for 4 hours (30 minutes for
8 hours), and group E to 4 ppm for 4 hours (30 minutes for 8 hours). The experimental design
and cellular proliferation results are illustrated in Table 4-20. Intermittent exposures at 2 and 4
ppm resulted in formaldehyde-related histopathologic lesions similar to those reported by Zwart
et al. (1988). Disarrangement and squamous metaplasia in respiratory epithelium were observed
at 4 ppm (Table 4-20). Disarrangement, nest-like infolds, goblet cell hyperplasia, and rhinitis
were observed at 2 ppm. Rats exposed continuously for 8 hours at 2 ppm formaldehyde had
fewer lesions than rats intermittently exposed to 2 ppm and were not statistically different from
controls. Although lesions were noted in rats given the continuous 1 ppm, 8-hour treatment,
their incidence was not significantly different from the controls (Table 4-20). It should be noted
that the control rats in this study were reported to have a higher frequency of lesions than
controls in two previous studies from this laboratory employing the same techniques (Zwart et
al., 1988; Woutersen et al., 1987). For example, lesions noted in the respiratory epithelium of 25
control rats included 13 disarrangements, 13 basal cell hyperplasia, and 5 each of goblet cell
hyperplasia, nest-like infolds, and squamous metaplasia. This is in contrast to the data of
Woutersen et al. (1987), who reported no lesions in the respiratory epithelium of 20 control rats
(male and female). Although Zwart et al. (1988) discussed inflammatory lesions in control rats,
no mention was made of the other scored lesions in control animals. Overall, Wilmer et al.
(1989) reported clear adverse effects at 2 ppm formaldehyde, resulting from intermittent
exposure for 8 hours/day, 5 days/week for 13 weeks. The indication of no effects at 1 ppm and 2
ppm continuous exposure should be considered with some caution, given the unusual incidence
of lesions in the control animals.
The results reported by Wilmer et al. (1989, 1987) indicate a greater influence of
concentration, rather than exposure regimen (continuous versus intermittent) on formaldehyde
toxicity. However, these studies were conducted as repeated 8-hour exposure regimens over a
course of days or weeks. Therefore both regimens allowed for a 16-hour recovery time before
the next reexposure and do not represent a true continuous exposure. This research group has
speculated that defensive adaptation of the nasal mucosa may include the function of the
mucociliary apparatus (Feron et al., 1989). Morgan et al. (1986a) have shown formaldehyde
effects on mucus flow and ciliary beat in F344 rats to result from hourly exposures to 15 ppm
This document is a draft for review purposes only and does not constitute Agency policy.
4-152 DRAFT—DO NOT CITE OR QUOTE
-------
1 formaldehyde. However, effects seen in repeated 8-hour exposures may not correspond to those
2 effects resulting from shorter duration exposures to higher formaldehyde concentrations.
3
4 Table 4-20. Formaldehyde effects on nasal epithelium for various
5 concentration-by-time products in male albino Wistar rats
6
Exposure regimen (number of animals)
A (25)
B (22)
C (24)
D (23)
E (25)
0 ppm
1 ppm
2 ppm
2 ppm
4 ppm
Respiratory epithelium
8-Hour
8-Hour
8-Hour
8-Hour
at crosssection level 2
continuous
continuous
intermittent
intermittent
Disarrangement
Focal
12
4
8
3a
8
Diffuse
1
1
0
15b
llc
Necrosis
Focal
4
3
0
2
3
Diffuse
0
0
0
2
2
Basal cell hyperplasia
Focal
9
4
6
11
10
Diffuse
4
0
0
4
11
Squamous metaplasia
Focal
5
0
1
7
16°
Keratinization
0
0
1
0
3
Nest-like infolds
Focal
5
4
11
14°
7
Diffuse
0
3
1
0
1
Goblet cell hyperplasia
Focal
0
1
1
2
1
Diffuse
5
2
8
13b
10
Rhinitis
3
2
3
16°
8
7
8 "p < 0.05, compared with group A.
9 hp < 0.001. compared with group A.
10 cp <0.01, compared with group A.
11
12 Source: Wilmer et al. (1989).
13
14 Rusch et al. (1983a, b) performed a comparative study of formaldehyde effects on the
15 nasal epithelium in F344 rats, Syrian golden hamsters, and cynomolgus monkeys. Groups of
16 animals were exposed at 0, 0.2, 1, or 3 ppm (0, 0.25, 1.2, or 3.7 mg/m3) formaldehyde
17 22 hours/day, 7 days/week for 26 weeks. Six male monkeys, 10 male and 10 female hamsters,
18 and 20 male and 20 female rats were exposed at each exposure level. The experiment was run in
19 two trials, each with its own control group: trial 1 at 0.2 or 1 ppm and trial 2 at 3 ppm. Animals
20 were weighed weekly and physically assessed (details not given). At sacrifice, organ weights
21 were recorded for the kidney, adrenals, heart, and liver. Tissue sections of the lung (4), trachea,
22 and nasal turbinates (4) of each animal were examined by light microscopy (section locations not
This document is a draft for review purposes only and does not constitute Agency policy.
4-153 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
given). Additionally, sections were examined by electron microscopy for rats in the control and
1 ppm treatment groups (five rats per group).
Body weights of both male and female rats in the 3 ppm treatment group were depressed
by 20% between week 2 and the end of the 26-week exposure. Absolute liver weights were
decreased in these animals as well (26% lower in males and 12% lower in females,/? < 0.05).
This decrease in liver weight remained significant for male rats when normalized for body
weight (a ratio of 2.9 in treated versus 3.16 in controls) but not for female rats. No significant
body weight or organ weight changes were seen in hamsters or monkeys. Increased incidences
of congestion (36/156), hoarseness (32/156), and nasal discharge (62/156) were observed in
monkeys in the 3.0 ppm treatment group versus no hoarseness or congestion and only five
observations of nasal discharge in 156 observations for control monkeys. Increased nasal
congestion was noted in the two lower treatment groups of monkeys: 30/156 and
45/156 observations, respectively, versus 9/156 observations in nasal discharge in the controls.
The authors reported an increase in nasal discharge and lacrimation in treated hamsters but no
increases in symptoms in rats. However, observations of adverse symptoms in the control rats
were greater than 10% on some measures.
Rhinitis increased in rats in the 3 ppm treatment group, and the incidence in controls was
notable (Table 4-21). All groups of monkeys showed some rhinitis, and no treatment effects
were observed in either monkeys or hamsters. Monkeys and rats in the high treatment group
(3 ppm) had a greater incidence of lesions in the nasoturbinate epithelium (Table 4-22). Rusch
et al. (1983a, b) noted that most lesions were mild to moderate but were "somewhat more
severe" in the high treatment group. Hamsters did not exhibit a similar increase, with few
lesions noted in the nasal epithelium. Overall, these studies show a clear increase in adverse
health effects at 3 ppm for rats and monkeys, with no adverse effects seen in hamsters at this
treatment level or rats and monkeys at the lower concentrations (0.2 ppm and 1 ppm).
Table 4-21. Rhinitis observed in formaldehyde-treated animals; data pooled for
male and female animals
Syrian golden
F344 rats
Cynomolgus monkeys
hamsters
Trial 1:
I, Control
17/38
4/6
0/14
II, 0.2 ppm
14/39
4/6
0/4
III, 1 ppm
14/38
5/6
0/11
Trial 2:
IV, Control
12/40
2/6
0/9
V, 3 ppm
25/39
4/6
2/16
Source: Rusch etal. (1983a, b).
This document is a draft for review purposes only and does not constitute Agency policy.
4-154 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Table 4-22. Epithelial lesions found in the middle region of nasoturbinates of
formaldehyde-treated and control animals; data pooled for males and
females
F344 rats
Cynomolgus monkeys
Syrian golden
hamsters
Basal cell
hyperplasia
Squamous
metaplasia/
hyperplasia
Squamous metaplasia/
hyperplasia
Nasal epithelium
Trial 1:
I, Control
0/38
2/38
0/6
No lesions noted
II, 0.2 ppm
0/38
1/38
0/6
III, 1 ppm
0/36
3/36
1/6
Trial 2:
IV, Control
4/39
3/39
0/6
No lesions noted
V, 3 ppm
25/37
23/37
6/6
Source: Rusch et al. (1983a, b).
Andersen et al. (2008) examined the effect of formaldehyde exposure at several
concentrations and durations. This study comprised histopathology and cell proliferation data, as
well as genomic analyses at Level II of the nasal cavity. Toxicogenomics analysis was
performed only at Level II because this was the region where the most severe lesions have been
reported in chronic bioassays (Andersen et al., 2008; Monticello et al., 1991; Kerns et al., 1983).
More specifically, Andersen et al. (2008) stated that the histopathologic and cell proliferation
effects at Levels II and III (with similar tissue structure) (Monticello et al., 1991) provided
phenotypic anchoring for the genetic analysis. Table 4-23 summarizes many of the broad
phenotypic findings.
The primary conclusions of this study with regard to the histopathology and cell
proliferation are as follows:
• The presence of inflammatory cell infiltrates in the nasal epithelial tissue of F344 rats is
highly variable and provides no coherent pattern with dose or duration at levels below
6 ppm.
• Hyperplasia was observed following exposure to >2 ppm.
• Metaplasia was observed at 6 ppm on day 5, but not before or after.
• Cell proliferation (as measured by labeling indices) was significantly elevated in
Levels I—III at 6 ppm on day 5 and Level I on day 15, leading to the conclusion that
significant changes in cell proliferation may not occur at exposures to <2 ppm.
• A significant decrease in cell density was observed at Level I in animals exposed to
6 ppm formaldehyde for 15 days, which was posited to be related to tissue remodeling in
response to this concentration.
This document is a draft for review purposes only and does not constitute Agency policy.
4-155 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-23. Cellular and molecular changes in nasal tissues of F344 rats exposed to formaldehyde
Response^
D1
D1R
D5
D6
D6R
D15
0
0.7
2
6
15
0
0.7
2
6
15
0
0.7
2
6
15
0
0.7
2
6
15
0
0.7
2
6
15
0
0.7
2
6
15
I
0
1
6
8
-
4
2
1
7
-
1
1
5
8
-
5
2
4
7
-
6
1
3
7
-
3
1
0
5
-
H
0
0
0
0
-
0
1
3
8
-
0
0
3
8
-
0
0
1
8
-
0
0
2
8
-
0
0
2
7
-
M
0
0
0
0
-
0
0
0
0
-
0
0
0
7
-
0
0
0
0
-
0
0
0
0
-
0
0
0
0
-
PI
39±9
37±15
65±40
155±89a
79±55
56±37
51±44
119±38a
P2
-
-
-
a
-
-
-
-
P3
-
-
-
a
-
-
-
-
CD
321±30
336±64
377±141
400±61
362±61
340±57
321±37
293±53b
G
-
0
1
42
745
-
0
0
0
-
-
0
15
28
-
-
0
0
9
-
0
0
54
o
2 »
5 s
to
o
S
>S
TO
TO'
*
D = day; R = recovery.
I = infiltrations (number out of 8 total animals); H = hyperplasia (number/8); M = metaplasia (number/8).
P1-P3 = proliferation at levels I—III (ULLI).
CD = cell density (cells/mm) at Level I.
G = genes significantly altered at Level II of nasal epithelial tissue.
"Significantly elevated ULLI and LI at Level I on day 5 or significantly elevated 1LI lat Level I on day 15; indexa without numerical value indicates significant
increases in ULLI in all subregions of Levels II and III at day 5.
Statistically significant difference from control (p < 0.05).
Source: Andersen et al. (2008).
On
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Based on their analysis of the microarray data, Andersen et al. (2008) concluded that no
genes were significantly altered by exposure to 0.7 ppm from 1 to 15 days. Exposure to 2 ppm
primarily resulted in gene changes at 5 days of exposure, but not thereafter. One gene was
significantly increased on day 1, but the authors did not identify that gene. At 6 and 15 ppm,
42 and 745 genes were altered at day 1, respectively. After 5 days, gene changes were only
observed at 6 ppm (15 ppm was not examined after day 1). These findings support conclusions
reached by their laboratory in an earlier analysis. Thus, the primary conclusion in the Andersen
et al. (2008) study is that genomic changes, including those suggestive of mutagenic effects, did
not temporally precede or occur at lower doses than phenotypic changes in the tissue. The
implications of this finding will be examined later in Section 4.5.
4.2.1.2.2.5. Lune pathology: subchrottic studies. Studies have also investigated the ability for
formaldehyde to induce pathology in the trachea, bronchi, and lung tissue. These studies have
reported tracheal tissue changes, lung inflammation, necrosis, changes to the biochemistry of
BAL fluid and lung surfactant in a variety of species. Ozen et al. (2003a) noted changes in zinc
concentration in the lung tissue following exposure for formaldehyde. Dallas et al. (1989) and
Dinsdale et al. (1993) observed changes in P450 enzyme activity in the lung associated with
formaldehyde exposure.
Ozen et al. (2003a) measured zinc, copper, and iron content in lung tissue from
formaldehyde-exposed Wistar rats. Adult male rats were exposed to 0, 5, or 15 ppm (0, 6.2, or
18.5 mg/m3) formaldehyde 8 hours/day, 5 days/week for either 4 or 13 weeks. Rats were
checked daily and weighed weekly. At sacrifice, rats were autopsied and examined for gross
pathological changes. Lung tissue was homogenized and analyzed for zinc, copper, and iron.
Body weight gain was depressed in all treatment groups in a concentration-dependent
manner (p < 0.001) (Table 4-24). Formaldehyde-exposed rats consumed less food and water
than controls and showed unsteady breathing, increased nose cleaning, excessive licking,
frequent sneezing, and nasal mucosa hemorrhages. Significant decreases were seen in the zinc
content of lungs after either 5 or 10 ppm formaldehyde exposure (Table 4-25). Copper content
was unchanged from controls in all treatment regimens, whereas iron content was increased after
4 weeks of 5 ppm exposure and after 13 weeks of either 5 or 10 ppm formaldehyde exposure
(Ozen et al., 2003a).
This document is a draft for review purposes only and does not constitute Agency policy.
4-157 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-24. Percent body weight gain and concentrations of iron, zinc, and
2 copper in cerebral cortex of male Wistar rats exposed to formaldehyde via
3 inhalation for 4 and 13 weeks
4
Exposure (mg/m3)
Weight gain (%)a
Zinc (mg/kg)a
Copper (mg/kg)a
Iron (mg/kg)a
4-week data
0
20.11 ±2.87
120 ±6.03
4.60 ± 0.42
25.07 ±2.83
6.1
7.27 ± 1.49e
130 ± 7.26°
5.60 ± 0.50b
23.00 ±2.32
12.2
5.24 ± 1.52e
185 ± 10.36e
5.80 ± 0.60d
22.14 ± 1.95b
13-week data
0
60.53 ±7.84
123 ±6.22
4.67 ±0.38
24.92 ±2.84
6.1
38.41 ±2.53e
155 ± 7.94e
5.41 ± 0.56°
22.00 ±2.41
12.2
25.87 ± 1.32e
163 ± 6.03e
6.10 ± 0.73e
21.00 ± 1.96b
5
6 "Values are means ± SDs (n = 7).
7
8 Statistical significance of differences versus controls, as calculated by the authors:
9 hp < 0.05. ><0.02. dp< 0.002. ><0.001.
10
11 Source: Ozen et al. (2003b).
12
13
14 Table 4-25. Zinc, copper, and iron content of lung tissue from formaldehyde-
15 treated male Wistar rats
16
Concentration
Duration"
Zincbc
Copperb'c
Ironbc
0 ppm
Control
20.7 (1.6)
0.39 (0.05)
12.5 (0.8)
5 ppm
4 weeks
16.1 (1.3)d
0.32 (0.07)
12.9 (1.0)
10 ppm
4 weeks
13.8 (1.2)e
0.36 (0.04)
17.5 (1.3)e
0 ppm
Control
20.0 (1.6)
0.39 (0.05)
12.7 (0.4)
5 ppm
13 weeks
15.3 (1.4)e
0.37 (0.04)
17.9 (l.l)e
10 ppm
13 weeks
13.0 (l.l)e
0.39 (0.05)
22.4 (1.4)e
17
18 aRats were exposed 8 hours/day, 5 days/week for the number of weeks indicated.
19 bConcentrations are expressed as moles/mg of tissue, wet basis.
20 °Values are means (n = 7); SDs shown in parentheses.
21 dp<0 .005, compared with controls, as calculated by authors.
22 ep < 0.001, compared with controls, as calculated by the authors.
23
24 Source: Ozen et al. (2003a).
25
26 There are two reports of lung cytochrome P450 levels after formaldehyde exposure. The
27 first report by Dallas et al. (1989) describes concentration- and duration-dependent changes in P450
28 levels. Male Sprague-Dawley rats were exposed at 0, 0.5, 3.0, or 15 ppm (0, 0.62, 3.7, or 18.5
29 mg/m3) formaldehyde 6 hours/day, 5 days/week for 1 day, 4 days, 12 weeks, or 24 weeks. There
30 were six rats in each exposure group, but the experiment was run in two parts, with three rats in
31 each subgroup. Rats were sacrificed after 1 day, 4 days, 12 weeks, or 24 weeks of exposure, and
32 liver microsomes were prepared. TP and P450 content were determined on each sample.
This document is a draft for review purposes only and does not constitute Agency policy.
4-158 DRAFT—DO NOT CITE OR QUOTE
-------
1 Average P450 levels in control groups ranged from 17-76 pmol P450/mg protein.
2 However, no P450 was detected in lung from formaldehyde-treated animals after 1 day of
3 exposure, with a method detection limit of approximately 10 pmol P450/mg protein. In contrast,
4 P450 levels were elevated significantly above controls in a concentration-dependent manner after
5 4 days of formaldehyde exposure (Table 4-26). Although P450 levels remained elevated in some
6 experimental groups after 12 and 24 weeks of exposure, results were variable and less dramatic.
7
8 Table 4-26. Total lung cytochrome P450 measurements of control and
9 formaldehyde-treated male Sprague-Dawley rats
10
1 Dayab
4 Days
12 Weeks
24 Weeks
Formaldehyde
Expt. 1
Expt. 2
Expt. 1
Expt. 2
Expt. 1
Expt. 2
Expt. 1
Expt. 2
0 ppm
17(6)
44(13)
39(11)
23 (3)
29 (10)
19 (23)
76 (49)
18(11)
0.5 ppm
ND
ND
103 (52)
137 (14)e
87 (ll)d
35(7)
172 (12)°
38(9)
3.0 ppm
ND
ND
357 (10)e
278 (100)e
91 (10)d
67 (34)
92 (103)
30(15)
15 ppm
ND
ND
362 (38)e
334 (4)e
130 (2)e
56 (6)
151 (9)
48 (7)c
11
12 aRats were exposed 6 hours/day, 5 days/week for the duration shown.
13 bCytochrome P450 expressed as pmol P450/mg of protein. Values are means (SDs) (n = 3).
14 Different from control, p < 0.05.
15 dDifferent from control, p < 0.01.
16 "Different from control, p < 0.001, as calculated by the authors.
17 ND = not detected above the limit of detection, approximately 10 pmol/mg protein.
18
19 Source: Dallas et al. (1989).
20
21
22 A later study by Dinsdale et al. (1993) attempted to confirm the increase in P450 levels
23 reported by Dallas et al. (1989). In their first experiment, Dinsdale et al. (1993) treated male
24 Sprague-Dawley rats at approximately 10 ppm (12.3 mg/m3) formaldehyde 6 hours/day for
25 4 days. The formaldehyde vapor was generated from formalin by a concentric jet atomizer. For
26 the second experiment, Dinsdale et al. (1993) similarly exposed rats to formaldehyde, but the gas
27 was generated by the thermal depolymerization of paraformaldehyde as was done by Dallas et al.
28 (1989). The concentration of P450 and activity of several P450 isozymes were measured in lung
29 microsomes (pentoxyresorufin O-dealkylase, benzyloxyresorufin O-dealkylase, ethoxyresorufin
30 O-dealkylase, and 2-aminofluorene N-hydroxylation). ALP and y-glutamyl transpeptidase
31 activity were measured in BAL fluid collected from each animal. No changes were seen in BAL
32 enzyme activity or the activity of lung microsomes for the P450 substrates tested. Cytochrome
33 P450 levels were unchanged in experiment 1, where formaldehyde was generated from formalin.
34 Cytochrome P450 levels were increased in experiment 2 with formaldehyde generated from
35 paraformaldehyde (Table 4-27).
36
This document is a draft for review purposes only and does not constitute Agency policy.
4-159 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Table 4-27. Cytochrome P450 levels in formaldehyde-treated rats
Group
Experiment 1
(formalin)3
Experiment 2
(paraformaldehyde)3
(nmol/mg protein)
Control
82 ±30
85 ±5
10 ppm formaldehyde
73 ±27
125 ± 23b
"Values are means ± SDs (n = 3-5).
bDifferent from controls,/? < 0.05.
Source: Dinsdale et al. (1993).
4.2.1.2.2.6. Extrapulmonary toxicity: subchronic studies. Several studies have investigated
toxicity in organs other than those associated with the respiratory tract. An earlier cross-species
study examined changes in lung tissue resulting from continuous exposure (Coon et al., 1970).
Animals were exposed to 3.7 ppm (4.6 mg/m3) formaldehyde for 90 days. Five species of
animals were studied: male and female Sprague-Dawley and Long-Evans derived rats (15), male
and female Princeton-derived guinea pigs (15), male New Zealand albino rabbits (3), male
squirrel monkeys (Saimiri sciureus) (3), and purebred male beagle dogs (2). Blood samples were
taken for Hb concentration, HCT, leukocyte counts, and serum levels of BUN, AST, ALT, ALP,
and LDH. Sections of heart, lung, liver, kidney, and spleen were fixed and examined from each
species (details of method not provided). Brain, spinal cord, and adrenal tissue also were
examined in monkeys and dogs as well as thyroid from dogs. Liver and kidney sections were
stained for reduced nicotinamide adenine dinucleotide, lactate, isocitrate, and P-hydroxybutyrate.
Tissue sections of the nasal mucosa were not examined in this study.
Hematological parameters were unaffected by formaldehyde treatment. The lung tissue
of all species exhibited interstitial inflammation after 90 days of formaldehyde exposure
(detailed description not provided). Formaldehyde-treated rats and guinea pigs also had focal
chronic inflammation in heart and kidney tissue sections. However, the authors were uncertain
whether the observed changes to heart and kidney were due to formaldehyde exposure.
As mentioned above, Woutersen et al. (1987) exposed male and female albino SPF
Wistar rats (10/group) to 0, 1, 10, or 20 ppm (0, 1.23, 12.3, or 24.6 mg/m3) formaldehyde
6 hours/day, 5 days/week for 13 weeks. Rats were checked daily and weighed weekly. During
week 13, blood samples were taken for Hb, PCV, RBC count, and a differential count of
leukocytes. Urine samples were also analyzed. At sacrifice, blood samples were analyzed for
ALB, creatinine, glucose, TP, BUN, and the enzyme activities (AST, ALT, and ALP). GSH and
protein content were determined in liver homogenates. Organs were examined and weighed:
adrenals, brain, heart, kidneys, liver, lungs, ovaries, pituitary, spleen, testes, thymus, and thyroid.
This document is a draft for review purposes only and does not constitute Agency policy.
4-160 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
No gross pathological changes were seen upon autopsy, but body weights decreased in
both male and female rats at the 20 ppm treatment level. Of the organs weighed, 6 of 11 had
significantly increased relative rates in male rats exposed to 20 ppm formaldehyde. Relative
brain weight was increased in female rats at the same treatment level (Woutersen et al., 1987).
Clinical chemistry parameters of liver and kidney function and hematological parameters
were also measured after the 13-week treatment by Woutersen et al. (1987). Compared with
those of controls, activities of AST, ALT, and ALP were significantly elevated in plasma from
the 20 ppm treated male rats (by 124, 132, and 126%, respectively; p < 0.05). Total plasma
protein was reduced to 95% of controls in the same animals. Although there was an observed
increase in BUN in male rats treated with 1 ppm, this was not considered a treatment effect.
Furthermore, no statistically significant differences were seen for these parameters in female rats
at any concentration level (Woutersen et al., 1987).
Sul et al. (2007) exposed Sprague-Dawley rats to 0, 5, and 10 ppm formaldehyde for
6 hours/day (5 days/week) for 2 weeks and collected lung samples for tissue damage and
genomic analysis. According to their results, 21 genes were altered in a dose-dependent manner
by microarray analysis; 2 were up regulated and 19 were down regulated in the lung tissue of
animals exposed to formaldehyde. However, six of the nine genes further analyzed by PCR did
not show dose dependency (authors did not comment). Although the authors briefly describe the
functions and potential implications for changes in the expression of some of the altered genes,
there is no discussion of the relationship between these altered genes (i.e., there is no pathway
analysis).
In 2006, Im et al. (2006) published a proteomic analysis using the same exposure
protocols (possibly using the same animals as in the Sul et al. [2007] study, although neither
study makes reference to the other). Im et al. (2006) examined DNA damage in lymphocytes
and liver tissues, as well as protein and lipid oxidation in plasma and liver samples. Similar to
changes reported in the lung (discussed elsewhere), using two-dimensional electrophoresis and
matrix-assisted laser desorption ionization time-of-flight mass spectrometry, the authors also
reported dose-dependent changes in the levels of 32 proteins in plasma (19 up, 13 down). None
of the changes in plasma proteins correspond to the changes in lung reported by Sul et al. (2007).
Again, no pathway analysis was provided. Interestingly, Im and colleagues (2006) also
demonstrated a dose-dependent increase in plasma IL-4 and dose-dependent decrease in IFNy,
perhaps indicative of Th-2-mediated inflammatory response. An overview of formaldehyde
exposure-related pathology in the respiratory system of laboratory animals is presented in
Table 4-28.
This document is a draft for review purposes only and does not constitute Agency policy.
4-161 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-28. Summary of respiratory tract pathology from inhalation exposures to formaldehyde, subchronic studies
S"4
>3*
&
to
s
Species/strain
No./
group
Treatment3
Respiratory effects
LOAEL/NOAEL
Reference
Nasal pathology
B6C3F1 mice
(male and
female)
10
0, 2, 4, 10, 20, or 40 ppm
formaldehyde 6 hours/day,
5 days/week for 13 weeks
Minimal squamous metaplasia in 1 of 10 mice
(4 ppm). Squamous metaplasia observed in all mice
at 10 and 20 ppm.
NOAEL = 4 ppm
Maronpot et al.
(1986)
SPF Wistar Rats
(male and
female)
10
0, 1, 10, or 20 ppm
formaldehyde 6 hours/day, 5
days/week for 13 weeks
Increased respiratory epithelial hyperplasia and
keratinization at 20 ppm; squamous metaplasia at
10 ppm in males and females.
NOAEL = 1 ppm
Woutersen et al.
(1987)
SPR Wister rats
(male)
50-55
0, 10, or 20 ppm formaldehyde
for 6 hours/day, 5 days/week
for 4, 8, or 13 weeks
Rhinitis, hyperplasia, and squamous metaplasia in
respiratory epithelium at all doses (number of weeks
not specified).
Squamous metaplasia of olfactory epithelium at
20 ppm (number of weeks not specified)
NOAEL = 1 ppm
Feronetal. (1988)
Wistar rats (male
and female)
50
0, 0.3, 1, or 3.0 ppm
formaldehyde 6 hours/day,
5 days/week for 13 weeks
Keratinized and non-keratinized squamous
metaplasia in level 2 epithelium in 37/50 male and
21/50 female rats at 3 ppm for 13 weeks.
NOAEL = 1 ppm
Zwart et al. (1988)
Wistar rats (male)
10
40 ppm-hours (8 hours at
5 ppm, 4 hours at 10 ppm) or
80 ppm hours (10 ppm
continuous or 20 ppm
intermittently)
Thinning and disarrangement of respiratory
epithelium, squamous metaplasia, most severe in
20 hours intermittent exposure
NOAEL =10 ppm
Wilmer et al.
(1987)
Wistar rats (male)
10
0, 8, or 16 ppm, given either
continuously or intermittently
Disarrangement and squamous metaplasia at 4 ppm.
Continuous exposure yielded less severe lesions
than intermittent exposure
LOAEL = 8 ppm
Wilmer et al.
(1989)
F344 rats (male
and female),
Syrian golden
hamsters (male
and female),
cynomolgus
monkeys
20 rats,
10
hamsters,
6 monkeys
0.0.2, 1, or 3 ppm
22 hours/day, 7 days/week,
26 weeks
Rats: rhinitis at 3 ppm, increased incidence of nasal
lesions at 3 ppm.
Monkeys: rhinitis at all doses, increased incidence
of nasal lesions at 3 ppm.
Hamsters: no significant nasal lesions.
NOAEL = 1 ppm
Rusch et al.
(1983a, b)
Tracheal and lung pathology
Wistar rats (male)
6
0, 5, 15 ppm for 8 hours/day,
5 days/week, 4 or 13 weeks
Significant decreases in zinc content in lung, copper
unchanged, iron increased in lung.
LOAEL = 5 ppm
Ozen et al. (2003)
o
2 »
5 s
to
o
-------
Table 4-28. Summary of respiratory tract pathology from inhalation exposures to formaldehyde, subchronic studies
S"4
>3*
s
Species/strain
No./
group
Treatment3
Respiratory effects
LOAEL/NOAEL
Reference
Sprague-Dawley
rats (male)
6 but n = 5
in some
trials
0, 0.5, 3.0, 15 ppm6
hours/day, 5 days/week, for 1
day, 4 days, 12 weeks, 24
weeks
Increased P450 levels after 4 days at 3 ppm.
NOAEL = 0.5 ppm
Dallas et al.
(1989)
Sprague-Dawley
rats (male)
5
0 or 10 ppm 6 hours/day,
4 days using both formalin and
paraformaldehyde
P450 levels increased at 10 ppm only in groups
treated with paraformaldehyde.
LOAEL =10 ppm
Dinsdale et al.
(1993)
Extrapulmonary effects
Rats and guinea
pigs
15
3.7 ppm for 90 days
Focal chronic inflammation in heart and kidney
tissue.
LOAEL = 3.7 ppm
Coon et al. (1970)
SPF Wistar rats
(male and
female)
10
0, 1, 10, or 20 ppm
6 hours/day, 5 days/week for
13 weeks
Relative brain weight increased in female rats at
20 ppm; increased AST, ALT, ALP in plasma at
20 ppm.
NOAEL =10 ppm
Woutersen et al.
(1987)
o
S ?
a, Co
TO Sj-
§ ^
s ^
§ 3
*¦>1.
-j^
On
U>
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.2.3. Chronic inhalation bioassays. The respiratory pathology observed in chronic
bioassays is consistent with the subchronic studies. As exposure concentration and duration of
exposure are increased, the pathology becomes more severe and penetrates more deeply into the
respiratory tract. These effects are progressive over time. Tumors are reported in several
bioassays, primarily SCCs. Experimental results regarding both the severity of respiratory tract
pathology as well as the tumor incidence vary by species strain and experimental design. As
discussed above rodents experience RB, and species differences in respiratory and physiological
depression would result in differences in absorbed dose in the respiratory tract, given the same
exposure concentration (Chang and Barrow, 1983). Additionally, differences in the nasal
architecture result in species-dependent variation of formaldehyde absorption (flux) within the
respiratory tract (see Section 3.4). Therefore, chronic studies are discussed by species for greater
clarity.
4.2.1.2.3.1. Mice. Early experiments by Horton et al. (1963) subjected mice (C3H, sex
unspecified) to extreme formaldehyde concentrations (0, 0.05, 0.1, and 0.2 mg/L or 41-163 ppm)
in an attempt to simulate lung pathology reported in humans exposed to cigarette smoke. The
mice were exposed 1 hour/day, 3 days a week for up to 35 weeks. The authors did not note the
effects of RB or provide any information on pathology of the URT. There was a clear increase
in histologic changes in the tracheobronchial epithelium by exposure, including basal-cell
hyperplasia, stratification squamous cell metaplasia and atypical metaplasia. Subsequent
exposures to various combinations of formaldehyde and coal tar did result in squamous cell
tumors. The findings of Horton et al. (1963) suggest a role for formaldehyde in lung cancer
under some conditions. However, the exposure design and early deaths in the treatment groups
severely limit the usefulness of these data in human health risk assessment.
In a comprehensive study conducted by Swenberg et al. (1980) (also reported in Kerns et
al. [1983]) in conjunction with Chemical Industry Institute of Toxicology (CUT) and Battelle
Columbus Laboratories, male and female C57BL/6 x C3H Fi (B6C3Fi) mice (approximately
120/sex/concentration) were exposed to 0, 2.0, 5.6, or 14.3 ppm (0, 2.45, 6.87, or 17.5 mg/m3)
formaldehyde 6 hours/day, 5 days/week for 24 months. This exposure period was followed by
up to 6 months of nonexposure to evaluate recovery. Interim sacrifices were conducted at 6, 12,
18, 24, 27, and 30 months (due to unscheduled deaths, no male mice were sacrificed at 18 or
27 months). Exposure generation was accomplished by sublimation of paraformaldehyde, and
exposures were conducted in whole-body chambers. Detailed sectioning and examination of the
nasal passages were conducted at each interim sacrifice, beginning at 12 months, and for all
unscheduled deaths. Gross organ pathology was noted for all animals and complete
This document is a draft for review purposes only and does not constitute Agency policy.
4-164 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
histopathologic examination was conducted on all animals in the control and high-exposure
groups. There were no differences in survival in any exposure group compared with controls.
Generally, poor survival in all groups of male mice was attributed to fighting and infections of
the urogenital tract associated with group housing; 78, 77, 81, and 82 unscheduled deaths were
recorded before 24 months in the 0, 2.0, 5.6, and 14.3 ppm treatment groups, respectively (
n = 119, 120, 120, and 119 males, respectively). After the interim sacrifices (6 and 12 months)
only 17-22 male mice survived to the 24-month scheduled sacrifice. Female mice had much
greater survival with only 30, 34, 19, and 34 unscheduled deaths prior to the 24-month sacrifice.
The authors did not note the effects of RB in mice, although the RD50 for a 10-minute exposure
for male B6C3F1 mice has been reported at 4.9 ppm and 4.4 ppm (Steinhagen and Barrow, 1984;
Chang et al., 1981).
The first examination of the nasal cavities was conducted at the 12-month interim
sacrifice. Inflammation in the nasal turbinates was evident in mice in the 2 and 6 ppm treatment
groups (14/20 and 18/20, respectively), including adenitis of the nasal lacrimal duct, lacrimal
duct, and vomeronasal gland. Inflammation was not present in mice exposed at 15 ppm,
although serous rhinitis was seen in 4 of 20 animals. At 18 months, mice exposed at 2 and
6 ppm no longer exhibited adenitis in the nasoturbinates. Epithelial dysplasia was evident in 4 of
20 mice at 6 ppm exposure. Mice in the high-exposure group had significantly greater nasal
pathology; epithelial dysplasia and squamous metaplasia were reported in 18/19 and 17/19
female mice, respectively, exposed to 15 ppm. After 24 months, squamous epithelial hyperplasia
of the nasolacrimal duct (29/45) and atrophy of the olfactory epithelium (18/45) were also noted
in animals from the high-exposure group (male and female) (Battelle Columbus Laboratories,
1981). Similar pathology was reported in only a small fraction of mice exposed at 2 and 6 ppm
(5/48 and 11/60, respectively).
Three months after cessation of exposure, only nine female mice were available for
sacrifice, but within this small sample the data suggested recovery of nasal lesions: epithelial
dysplasia (4/9), squamous metaplasia (2/9), atrophy of the olfactory epithelium (1/9), and
squamous epithelial hyperplasia of the nasolacrimal duct (1/9) (Battelle Columbus Laboratories,
1981).
Of the 17 male mice that survived to 24 months in the 14.3 ppm exposure group, 2 had
SCC in the nasal cavity (p < 0.05). Of the two tumor-bearing mice, one exhibited significant
epithelial pathology, including rhinitis, dysplasia, squamous metaplasia, and hyperplasia.
Squamous metaplasia of the nasolacrimal duct was the only related pathology reported for the
second mouse. No SCCs were found in female mice, although 48 mice survived to 24 months.
The authors reported no other formaldehyde-related tumors. However, comparisons were based
This document is a draft for review purposes only and does not constitute Agency policy.
4-165 DRAFT—DO NOT CITE OR QUOTE
-------
1 on summary tables by organ. Although lymphomas were analyzed by organ and site (e.g.,
2 increase in salivary gland lymphoma considered separately from mandibular lymphoma), later
3 reanalysis of lymphoma in female mice, based on tumor-bearing animals (TBAs), does indicate
4 an association with formaldehyde exposure.
5
6 4.2.1.2.3.2. Rats. Holmstrom et al. (1989a) evaluated co-exposure of inhaled formaldehyde
7 with wood dust in 16 female Sprague-Dawley rats/group. Rats were exposed in whole-body
8 chambers for 6 hours/day, 5 days/week for 104 weeks to formaldehyde alone at 12.4 ±1.1 ppm
9 (15.21 ± 1.35 mg/m3), wood dust alone (25 mg/m3), or both wood dust (25 mg/m3) and
10 formaldehyde (12.7 ±1.0 ppm) or to room air as the control. The wood dust was generated from
11 grinding of beech. Microscopic measurements of the wood particles indicated that
12 approximately 70% had a geometric diameter of about 10 [j,m, while 10-20% were about 5 [j,m or
13 less. Animals were sacrificed at 104 weeks and histopathology was performed on five transverse
14 sections of the nasal cavity (Figure 4-3) and the lungs (not otherwise specified).
15 There were no differences in mortality among the groups at any time during the study
16 period. Rats exposed to formaldehyde were reported to have exhibited yellow discoloration of
17 the fur, and many displayed eye irritation. Formaldehyde exposure, with and without wood dust,
18 induced squamous metaplasia, keratinization, and dysplasia of the nasal epithelium (Table 4-29).
19
20 Table 4-29. Histopathologic findings and severity scores in the naso- and
21 maxilloturbinates of female Sprague-Dawley rats exposed to inhaled
22 formaldehyde and wood dust for 104 weeks
23
Treatment
Pronounced
squamous
metaplasia
Pronounced
squamous
metaplasia
with
keratinization
Pronounced
squamous
metaplasia
with
presence of
dysplasia
Sum of rats
with
pronounced
metaplasia
and/or
dysplasia
ccscc
Histologic scores
at the level of
naso- and
maxilloturbinates
(mean ± SD)
Formaldehyde
group (n = 16)
7
2
1
10
1
2.25 ± 1.73a
Formaldehyde-
wood dust group
(n= 15)
7
1
4
12
0
2.6 ± 1.88a
Wood dust group
(n= 15)
0
0
0
0
0
1.86 ± 0.83b
Control group
(n= 15)
0
0
0
0
0
1.07 ±0.70
24
25 a/?<0.01.
26 hp < 0.05.
27
28 Source: Holmstrom etal. (1989a).
This document is a draft for review purposes only and does not constitute Agency policy.
4-166 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Among the five levels of the nasal cavity that were examined, Holmstrom et al. (1989a)
presented findings for the naso- and maxilloturbinates since formaldehyde-induced tumors had
been associated with this level (Morgan et al., 1986a, b). The data also suggested an effect of
wood dust on formaldehyde-induced nasal pathology, with a slightly higher histologic score and
greater incidence of dysplasia than formaldehyde exposure alone. One SCC (1/16) occurred in
the group exposed to formaldehyde only but not in the group exposed to formaldehyde and wood
dust. Microscopic examination of the lungs revealed that emphysema (diagnostic criteria not
specified) was more prevalent in both groups exposed to wood dust compared with the control
group (p < 0.05). There was no significant difference in pulmonary epithelial histopathology
among the groups.
Tobe et al. (1985) also evaluated F344 rats (32/group) exposed to inhaled formaldehyde
for 28 months. Exposures were for 6 hours/day, 5 days/week to formaldehyde concentrations of
0, 0.3, 2, and 14 ppm (0, 0.37, 2.45, and 17.2 mg/m3). Fourteen of 32 rats (44%) in the high
concentration group developed nasal SCCs, compared with none in the other exposed groups and
the control group. Tobe et al. (1985) reported increased rhinitis, hyperplasia, and squamous
metaplasia of the nasal respiratory epithelium, including in the low-exposure group (0.3 ppm.)
However, some level of rhinitis, hyperplasia, and metaplasia were also present in controls.
Without a more complete report, it is unknown whether or not the pathology reported at 0.3 ppm
was a formaldehyde-related effect.
Kamata et al. (1997) evaluated the effects of inhaled formaldehyde in male F344
(F344/DuCrj) rats (32/group) exposed for 28 months. Formaldehyde exposure was generated by
metering 37% formalin (containing 10% methanol) into a sprayer in a glass bottle and diluting
with room air. Concentration in the chamber was monitored twice daily by the acetyl acetone
method. Exposures were for 6 hours/day, 5 days/week at nominal formaldehyde concentrations
of 0, 0.3, 2.0, and 15 ppm (0, 0.37, 2.45, and 18.4 mg/m3). Actual levels were 0, 0.3 ± 0.07, 2.17
± 0.32, and 14.85 ± 2.22 ppm (mean ± SD). Rats in the 0 ppm group were given methanol to
inhale at the same concentration (4.2 ppm) as the 15 ppm group. A room control no-exposure
group was also included in the study. All animals were observed for clinical signs once a day
during the study. Body weights and food consumption were recorded weekly. Five animals per
group, randomly selected at the end of 12, 18, and 24 months, and all surviving animals at
28 months were sacrificed for hematological measurements (Hb, RBCs, PCV, MCV, mean
corpuscular hemoglobin [MCH], MCHC, and WBCs), biochemical determinations (TP, ALB,
BUN, ALP, AST, ALT, glucose, albumin/globulin ratio, phospholipids, triglycerides, and total
cholesterol), and pathological examinations. Wet weights were taken on brain, heart, lungs,
liver, kidneys, spleen, testes, and adrenal gland of each rat. Histopathology was performed on all
This document is a draft for review purposes only and does not constitute Agency policy.
4-167 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
moribund or dead animals and those at specified sacrifices on all gross lesions and the following
tissues: pituitary, thyroid, nasal cavity, trachea, esophagus, stomach, small and large intestines,
prostate gland, urinary bladder, muscle, femur, sciatic nerve, spinal cord, and mesenteric lymph
nodes. Histopathologic sections of the nose were obtained from five anatomical levels, but these
did not correspond to the typical levels taken in other bioassays. Most notably, section level B
was anterior and not posterior to the incisor teeth. The incidence data for nasal histopathology
were not reported with respect to section level location, with the exception that the
nonproliferative lesions and tumors reported were described to occur predominantly at levels B
and C.
Yellow discoloration of the coats occurred in animals exposed at the 2 and 15 ppm levels.
Significant decreases in body weight and food consumption were observed in the high
concentration (15 ppm) group throughout the exposure period, and elevated mortality was noted
at 28 months (88.3 versus 31.8% in controls). The first death occurred after 6 versus 18 months
in the control group. Other effects noted in the 15 ppm exposure group include decreased
triglycerides, reduced liver weight (both relative and absolute), and increased relative adrenal
weights.
Treatment-related macroscopic and histopathologic findings were limited to the nasal
cavity. Squamous cell metaplasia was reported in all treatment groups: 16% (0.3 ppm), 37.5%
(2 ppm), and 91% (15 ppm) of exposed rats. Epithelial hyperplasia was similarly present in 12.5,
22, and 91% of the animals, respectively. Since a no-effect level could not be determined, the
authors reported benchmark doses (BMDs) of 0.25 and 0.24 ppm for squamous cell metaplasia
and epithelial hyperplasia (10% response.). Additional lesions only occurring in the 15 ppm
dose group were papillary hyperplasia (2/32), SCC (13/32), squamous cell papilloma (3/32), and
sarcoma (1/32). The majority of the tumors were located at levels B and C of the nasal cavity.
Albert et al. (1982) and Sellakumar et al. (1985) reported on a set of lifetime studies
performed in male Sprague-Dawley rats to evaluate the effects of inhaled formaldehyde alone
and in combination with hydrochloric acid (HC1). Rats were exposed 6 hours/day, 5 days/week
for life. In the first experiment (Albert et al., 1982), 8-week-old male inbred Sprague-Dawley
rats (n = 99) were exposed to a mixture of 10 ppm (12.3 mg/m3) HC1 and 14 ppm (17.2 mg/m3)
formaldehyde, and there were two control groups: air-sham and untreated (n = 50).
Bis(chloromethyl)ether (BCME), a known animal carcinogen (Albert et al., 1975; Kuschner et
al., 1975; Figueroa et al., 1973; Laskin et al., 1971), is formed when formaldehyde and HC1 are
mixed. BCME concentrations were estimated at about 1 ppb in the formaldehyde-HCl mixed
exposures, based on levels in the mixing chamber. Complete necropsies were conducted when
animals died naturally or were killed when moribund. Histologic sections were taken from the
This document is a draft for review purposes only and does not constitute Agency policy.
4-168 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
nasal cavity, larynx, trachea, pulmonary lobes, liver, bladder, kidney, spleen, and other organs
with gross pathologic alterations.
Exposure to the mixed gases (formaldehyde-HCl-BCME) induced nasal lesions,
including epithelial hyperplasia (71%), squamous metaplasia (64%), squamous papilloma (3%),
and SCC (25%) (Albert et al., 1982). Although a few squamous metaplasia were noted in the
larynx, trachea, and bronchi, these lesions were also noted in controls. Mortality in exposed rats
was significantly increased over controls and was approximately 30% when the first carcinoma
was reported (233 days). Mortality in exposed rats rose quickly to approximately 60% after the
first year of exposure. Therefore, the authors used a life-table method to calculate a mortality-
corrected cumulative incidence, reporting a corrected cumulative incidence of 77% at 720 days
after first exposure.
In the second experiment performed in this laboratory (Sellakumar et al., 1985; Albert et
al., 1982), Sprague-Dawley rats were similarly exposed to HC1 (10 ppm) alone, formaldehyde
alone (15 ppm), or a combination of both. The combination exposure was generated in two
different ways to better understand the influence of BCME formation on study results: premixed
at high concentrations and gases fed separately into the inlet air supply at the target
concentrations. BCME concentration measured by a gas chromatography/mass spectrometry
method in the premixed chamber varied between 0.1 and 0.4 ppb. Cage-side observations and
necropsy procedures were as described in Albert et al. (1982) with the exception of the histologic
preparation of the head. The head was cut transversely into four tissue blocks, and sections were
taken from the face of each.
Animals exposed to formaldehyde alone and formaldehyde-HCl (premixed or non-
premixed) showed a marked decrease in body weight after 16 weeks. After 32 weeks rats
exposed to the premixed formaldehyde-HCl (with BCME) had higher mortality compared with
the other mixed gas exposures (p < 0.05). Nasal pathology was similar among rats exposed to
formaldehyde alone or the mixed gases (Table 4-30). Desquamation of respiratory epithelial
cells was reported in the respiratory epithelium that covers the nasomaxillary turbinates and the
nasal septum (approximately section levels 2 and 3). Olfactory epithelium in the ET frequently
showed an inflammatory reaction with seropurulent exudate filling the lumen. Squamous
metaplasia and hyperplasia were reported in the larynx and trachea in all treatment groups.
Tumors arose primarily from the nasomaxillary turbinates and nasal septum. The SCCs
were predominantly moderate to well differentiated, with excessive amounts of keratin occluding
the lumen, killing the animals by asphyxiation. Statistical comparisons by the log rank test (Peto
test) showed that tumor incidence was increased in the premixed formaldehyde-HCl combined
exposure group over formaldehyde alone or the combined formaldehyde-HCl (not premixed).
This document is a draft for review purposes only and does not constitute Agency policy.
4-169 DRAFT—DO NOT CITE OR QUOTE
-------
1 There were no significant differences in the latency among groups, with the average latency
2 varying from 603 to 645 days. Rats exposed to HC1 exposure alone did not develop tumors.
3 The esthesioneuroepithelioma is a unique tumor type observed with a high incidence in
4 an earlier inhalation study of rats exposed to BCME (Kuschner et al., 1975), suggesting that the
5 higher incidence of nasal tumors observed in the premixed-combined formaldehyde-HCl-
6 exposure group may have been due to BCME (Krimsky, 1986) since this premixed protocol was
7 the one most likely to generate BCME. Sellakumar et al. (1985) refuted this assertion, stating
8 that this singular tumor occurred in the absence of other changes in the ethmoid region or in the
9 lungs where BCME was also demonstrated to cause tumors. Furthermore, exposure was
10 approximately one-tenth the cumulative dose in the Kuschner et al. (1975) study that was
11 associated with a single similar tumor. Sellakumar et al. (1985) attributed the higher incidence
12 in the premixed-combination group to traces of other alkylating agents (not BCME) that could
13 have been formed. The results demonstrate that animals exposed to either a combination of
14 formaldehyde-HCl or to formaldehyde alone develop nasal tumors, principally SCCs, at about
15 the same frequency, indicating that HC1 plays little or no role in the carcinogenicity of inhaled
16 formaldehyde.
17
18 Table 4-30. Histopathologic changes (including tumors) in nasal cavities of
19 male Sprague-Dawley rats exposed to inhaled formaldehyde or HC1 alone
20 and in combination for a lifetime
21
Observation
Premixed
HC1-HCHO
Non-premixed
HC1-HCHO
HCHO
HC1
Air
Colony
Number of animals examined
100
100
100
99
99
99
Rhinitis
74
75
74
81
72
70
Epithelial or squamous hyperplasia
54
53
57
62
51
45
Squamous metaplasia
64
68
60
9
5
6
Polyp or papilloma
13
11
10
0
0
0
see
45
27
38
0
0
0
Adenocarcinoma
1
2
0
0
0
0
Mixed carcinoma
0
0
1
0
0
0
Fibrosarcoma
1
0
1
0
0
0
Esthesioneuroepithelioma
1
0
0
0
0
0
Larynx
Hyperplasia
11
22
21
22
2
2
Squamous metaplasia
10
15
4
0
0
0
Trachea
Hyperplasia
18
32
21
26
6
2
Squamous metaplasia
9
8
7
0
0
0
22
23 Source: Sellakumar et al. (1985).
24
25
This document is a draft for review purposes only and does not constitute Agency policy.
4-170 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
In a companion study to the chronic mouse study described above (Kerns et al., 1983;
Swenberg et al., 1980), groups of F344 rats (approximately 120/sex/concentration) were exposed
to 0, 2.0, 5.6, or 14.3 ppm (0, 2.45, 6.87, or 17.5 mg/m3) formaldehyde 6 hours/day, 5 days/week
for 24 months. This exposure period was followed by up to 6 months of nonexposure to evaluate
recovery. Interim sacrifices were conducted at 6, 12, 18, 24, 27, and 30 months. Study
parameters and methods were as described above.
Formaldehyde exposure increased mortality of both male and female rats in all treatment
groups (p < 0.05 for 6 and 15 ppm groups). Severe treatment-related mortality was seen at the
highest exposure group beginning at 12 months with only 30% surviving to the 24 months.
There were no alterations in clinical chemistry, neurofunctional, or ophthalmological
measurements considered to be related to formaldehyde exposure. A concentration-dependent
increase in yellow discoloration of the hair coat was observed. This discoloration dissipated over
the 3-month postexposure period. Rats in the highest-concentration group were dyspneic
(p < 0.01) and emaciated (p < 0.05) and had many facial swellings that on closer examination
were revealed to be carcinomas protruding through the nasal cavity. Neoplastic lesions in the
URT were first observed clinically at day 358 in females and day 432 in males.
Macroscopically, these lesions originated in the anterior portion of the nasal cavity and, in a few
instances, extended into the ETs.
Figure 4-9 shows the frequency of squamous metaplasia by location in the noses of rats
sacrificed at various time points along the 2-year exposure period. Histopathologic lesions were
confined to the nasal cavity and proximal trachea in concentration-dependent fashion. The
morphologic diagnosis of squamous metaplasia was used to designate zones of altered
epithelium that were characterized by the presence of a well-differentiated germinal layer
(stratum germinativum) and superficial layers of epithelium (stratum spinosum and stratum
corneum). Keratin was produced only in areas of squamous metaplasia. Epithelial dysplasia was
detected earlier than squamous metaplasia and was characterized by a mucosa that had
undergone a transition from nonciliated simple cuboidal to one that was several cells thick and
squamoid with an organization and polarity of the individual cells that had changed from vertical
to horizontal with respect to the basement membrane. Similar histomorphologic changes have
also been called basal cell hyperplasia and epidermoid metaplasia (e.g., Albert et al. [1982]).
Figure 4-9 clearly illustrates that concentration is a dominant determinant of lesion distribution.
At low concentrations the lesions occur only in the most anterior region (cross-section level 1).
At 5.6 ppm, the squamous metaplasia in levels 1, 2, and 3 was also associated with purulent
rhinitis and epithelial dysplasia. At the highest concentration, the lesions progress to the more
distal URT, with lesions evident in level 5 and no difference in the incidence at level 1 or 2
This document is a draft for review purposes only and does not constitute Agency policy.
4-171 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
across the various sacrifice times. Statistically significant (p < 0.05) regression of the lesion was
evident at most locations at the 27-month sacrifice (3 months postexposure) (e.g., level 1 in the
2 ppm group, all levels of the 5.6 ppm group, and levels 4 and 5 of the 14.3 ppm group).
A
100
80
60
40
20
E±3 6 Months
^ 12 Months
H 18 Months
ESI 24 Months
^ 27 Months
B
&
£
0)
3
CT
0)
100
80
20
100
80
60
40
20
Level I
Level I
Level I
Level IV
Level V
Location of squamous metaplasia in nasal cavity
Figure 4-9. Frequency and location by cross-section level of squamous
metaplasia in the nasal cavity of F344 rats exposed to formaldehyde via
inhalation.
Note: Exposure concentrations were 2.0 ppm (A), 5.6 ppm (B), or 14.3 ppm (C).
Nasal cavity levels 2, 3, 4, and 5 were not evaluated at the 6- and 12-month
interim sacrifices in the 14.3 ppm exposure group.
Source: Redrawn from Kerns et al. (1983).
This document is a draft for review purposes only and does not constitute Agency policy.
4-172 DRAFT—DO NOT CITE OR QUOTE
-------
1 Furthermore, progression of lesions distally to the lower respiratory tract (LRT) occurred
2 only in the high concentration group. Tracheal pathology observed at 18 months included
3 multifocal areas of minimal to mild epithelial hyperplasia, epithelial dysplasia, or squamous
4 metaplasia. There were no significant tracheal lesions present in the 0, 2.0, or 5.6 ppm exposure
5 groups, and tracheal lesions were not observed during the postexposure period in the 14.3 ppm
6 exposure group.
7 Table 4-31 provides the summary data of all neoplastic lesions in the nasal cavity of
8 exposed rats. The adjusted cumulative incidence rates of SCC in male and female rats from the
9 14.3 ppm exposure group at 24 months were 67 and 87%, respectively. In this group, the
10 formation of zones of squamous metaplasia with zones of squamous epithelial hyperplasia and
11 increased keratin production appeared to precede areas of squamous papillary hyperplasia with
12 foci of cellular atypia. More advanced lesions included carcinoma in situ and invasive SCC of
13 the nasal turbinates. The neoplasia were extremely osteolytic and were associated with excessive
14 keratin production and mild to severe purulent rhinitis. In many animals from the high-exposure
15 group (with or without carcinoma), the excessive accumulation of keratin and inflammatory
16 exudates within the lumen of the URT caused severe dyspnea and death. Polypoid adenomas
17 were also observed in eight rats (four/sex) from the low-exposure group, six male rats from the
18 intermediate-exposure group, and six rats (five males, one female) from the high-exposure group
19 in level 1, 2, or 3. One control male rat had a similar lesion. When adjusted and unadjusted data
20 were analyzed, no significant differences were observed in pair-wise analyses; however, a
21 significant adjusted trend (p < 0.05) was reported for male rats. There was no evidence of
22 progression from polypoid adenoma to SCC.
23
24 Table 4-31. Summary of neoplastic lesions in the nasal cavity of F344 rats
25 exposed to inhaled formaldehyde for 2 years
26
No. of nasal
Undifferentiated
Formaldehyde
cavities
Nasal
carcinoma or
Carcino-
Polypoid
Osteo-
(ppm)
Sex
evaluated
SCC
carcinoma
sarcoma
sarcoma
adenoma
chondroma
0
M
118
0
0
0
0
1
1
F
114
0
0
0
0
0
0
2.0
M
118
0
0
0
0
4
0
F
118
0
0
0
0
4
0
5.6
M
119
1
0
0
0
6
0
F
116
1
0
0
0
0
0
14.3
M
117
51
la
2a
1
4
0
F
115
52
1
0
0
1
0
27
28 'One rat in this group also had an SCC.
29
30 Source: Kerns et al. (1983).
31
This document is a draft for review purposes only and does not constitute Agency policy.
4-173 DRAFT—DO NOT CITE OR QUOTE
-------
1 Morgan et al. (1986b) performed an additional analysis of the slides and tissues
2 from the Kerns et al. (1983) study to more precisely determine the location of each tumor
3 recorded. Additional sections were cut from the existing tissue blocks if a full slide set
4 (i.e., five sections) was unavailable for each animal. For each animal, the location of
5 each tumor was recorded on diagrams of the cross section of the nose, and an attempt to
6 determine the site of origin was made based on the center of the tumor mass. The results
7 for each case were assigned an accuracy rating that was based on the degree of
8 confidence that the pathologist had in the designated site of origin. Results for SCCs are
9 shown in Table 4-32.
10
11 Table 4-32. Apparent sites of origin for the SCCs in the nasal cavity of F344
12 rats exposed to 14.3 ppm of formaldehyde gas in the Kerns et al. (1983)
13 bioassay
14
Total SCC (%)a
Accuracy
Number of
Unable to
Sex
rating
animals
Area lb
Area 2b
Area 3b
Area 4b
determine
Male
High
36
56
28
14
3
NA
Low
25
56
20
8
0
16
Female
High
45
62
27
7
4
NA
Low
15
47
33
13
0
7
Totals
121
57
26
10
3
4
15
16 aRounded to nearest whole number.
17 bArea 1 = lateral aspect of the nasoturbinate and adjacent lateral wall; Area 2 = midventral septum; Area 3 = dorsal
18 septum and roof of dorsal meatus; Area 4 = dorsal and lateral aspect of the maxilloturbinate.
19 NA = not applicable.
20
21 Source: Morgan et al. (1986b).
22
23
24 In the 14.3 ppm exposure group, 98/103 rat noses had adequate numbers and quality of
25 slides for mapping the SCC distribution. Single neoplasia were present in 80 (40/sex), while
26 multiple neoplasia were present in 9 males (21 neoplasia) and 9 females (20 neoplasia). The
27 results were similar for cases with high or low accuracy. For example, more than half (57%) of
28 the SCCs occurred on the lateral side of the nasoturbinate and adjacent lateral wall at the front of
29 the nose (levels 1 and 2); approximately 25% were located on the midventral nasal septum
30 (levels 2 and 3); and about 10% were on the dorsal septum and roof of the dorsal meatus
31 (levels 1, 2, and 3). A small number (3%) were found on the maxilloturbinate (levels 2 and 3),
32 which only involved the medial aspect. All other regions of the nose where SCC was found were
33 considered to be involved as a result of invasion from one or more of the above sites. There
This document is a draft for review purposes only and does not constitute Agency policy.
4-174 DRAFT—DO NOT CITE OR QUOTE
-------
1 were two tumors in the 5.6 ppm group: one male had a single neoplasm on the ventral nasal
2 septum (level 3) while a female had an SCC from the lateral aspect of the maxilloturbinate to the
3 adjacent lateral wall (level 2).
4 On the basis of the morphology of 19 small neoplasia in this study and in additional work
5 described below (Morgan, 1997; Monticello et al., 1996), it was further concluded that the SCCs
6 arose from the epithelium lining the airway and not from the underlying glandular epithelium.
7 This mapping procedure and that of Monticello et al. (1996) described below were in good
8 concordance and showed a clear site specificity; most of the SCC arose in the anterior lateral
9 meatus (ALM) (57%), which is lined by transitional epithelium, and the midventral nasal septum
10 (26%), which is lined by respiratory epithelium (Morgan, 1997).
11 The CUT performed a second bioassay on inhaled formaldehyde in 9-week-old male
12 F344 (CDF[F344]/CrlBr) rats (Monticello et al., 1996). The rats were exposed 6 hours/day, 5
13 days/week for 24 months to 0, 0.7, 2, 6, 10, and 15 ppm (0, 0.86, 2.45, 7.36, 12.3, and
14 18.4 mg/m3) formaldehyde. Study objectives were to repeat the Kerns et al. (1983) bioassay,
15 better defining the concentration response relationship and to seek a correlation between
16 localized data on tumor sites and concomitant cell proliferation assays. Histopathology was
17 performed on six cross-section levels of the nasal cavity on every animal of an unscheduled
18 death and all those of the terminal sacrifice after 24 months. The distribution of lesions for each
19 individual animal was recorded onto epithelial maps of the nasal cavity at 30 selected levels
20 designed to permit accurate localization (Mery et al., 1994). Cell proliferation was measured in a
21 subset of animals (five per treatment group) at 3, 6, 18, and 24 months of exposure in each of the
22 nasal regions to which tumors were mapped (Table 4-33).
23
24 Table 4-33. Incidence and location of nasal squamous cell carcinoma in male
25 F344 rats exposed to inhaled formaldehyde for 2 years
26
Nasal location
No. of
Anterior
No. of
Formaldehyde
nasal
Anterior
Posterior
Anterior
Posterior
Anterior
medial
animals
concentration
cavities
lateral
lateral
mid-
mid-
dorsal
maxillo-
Maxillary
with
(ppm)
examined
meatus
meatus
septum
septum
septum
turbinate
sinus
SCCa
0
90
0
0
0
0
0
0
0
0
0.7
90
0
0
0
0
0
0
0
0
2
96
0
0
0
0
0
0
0
0
6
90
1
0
0
0
0
0
0
1
10
90
12
2
0
0
0
0
0
20
15
147
17
9
8
1
3
4
0
69
27
28 Total number of animals with SCCs, including those too large to allocate and those located in a site not listed in this
29 table.
30 Source: Monticello et al. (1996).
This document is a draft for review purposes only and does not constitute Agency policy.
4-175 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Yellow discoloration of the fur, a consistent response to formaldehyde in rats, was
observed in the rats exposed to 10 and 15 ppm formaldehyde. There were numerous premature
deaths in the 15 ppm exposure group, resulting in significantly decreased survival relative to
controls (18.8 versus 35.7%;p < 0.001). Survival was higher in the three lowest exposure
groups and statistically comparable to controls in the 10 ppm exposure group (35.7 versus
31.3%, respectively).
Control animals showed no histopathologic evidence of disease in the nasal passages.
Buccal cavity SCC, not associated with the nasal cavity, was present in 2 of 90 control animals.
This was considered an incidental finding and within the spontaneous incidence range reported
for this strain of rat. Buccal SCCs were observed in three animals at 15 ppm and in one animal
at 2 ppm. All other neoplastic responses in the respiratory tract were confined to the nose and
considered to have originated from the epithelium lining the nasal airways. The nasal neoplasia
included SCCs and polypoid (transitional) adenomas and were similar in morphologic
characteristics to those described in the Kerns et al. (1983) chronic bioassay. The incidence of
nasal SCCs by location is summarized in Table 4-33, which demonstrates a clear concentration-
response relationship. No SCCs occurred in the two lowest exposure groups or in the controls.
One nasal rhabdomyosarcoma and two nasal adenocarcinomas were reported in animals in the
highest treatment groups.
Regional analysis indicated that the SCCs arose in nasal regions lined with transitional or
respiratory epithelium and were most common in the lateral meatus and the midseptum (Table
4-33). Within the lateral meatus and mid-septum, there was clear evidence of a higher tumor
incidence rate in the anterior sample site (p = 0.001 and 0.02, respectively). Smaller numbers of
SCCs were observed on the medial aspect of the maxilloturbinate and the dorsal septum and on
the posterior lateral wall and lining of the nasopharyngeal meatus (data not shown). No SCCs
were observed in the maxillary sinus, with the exception of one animal exposed to 15 ppm that
had a small tumor in the wall of the ostium of this sinus. Tumor rates across the seven nasal
epithelial sites are presented in Table 4-33. There was an increasing tumor response between the
10 and 15 ppm exposure groups in all sites, except in the ALM. The SCC rates at 10 and 15 ppm
were virtually identical (13.3 and 11.6%, respectively), which is probably attributable to the
occurrence of many large neoplasia in the lateral meatus site that were not suitable and not
counted in the analysis.
The nonlinear tumor response is mirrored by a highly nonlinear response in cell
proliferation measured after 3, 6, 12, and 18 months of exposure. Significant treatment-induced
responses in cell proliferation indices at these time points were only observed at the two highest
exposure concentrations (10 and 15 ppm). Other treatment-induced lesions, predominantly
This document is a draft for review purposes only and does not constitute Agency policy.
4-176 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
epithelial hypertrophy, hyperplasia, squamous metaplasia, and mixed inflammatory cell
infiltrate, were also most severe at these two exposure concentrations. Significant distortion and
destruction of the nasoturbinate architecture occurred in many animals exposed to 15 ppm.
Nasal turbinate adhesions and olfactory degeneration (usually confined to the walls of the
anterior dorsal medial meatus) also occurred in animals exposed to 10 and 15 ppm. Lesions in
the 6 ppm exposure group were limited to focal squamous metaplasia in the anterior regions.
As discussed briefly above, small numbers of polypoid adenomas were also induced by
formaldehyde exposure and were similar in acinar-like structure and location to those in the
Kerns et al. (1983) bioassay. No polypoid adenomas occurred in the control animals or in the
0.7, 2, or 6 ppm exposure groups. A clear concentration response was observed in the 10 and
15 ppm exposure groups. Five of 90 animals (5.6%) in the 10 ppm exposure group and 14 of
147 animals (9.5%) in the 15 ppm exposure group had a polypoid adenoma. Most of these
polypoid adenomas {19%) were located in or adjacent to the lateral meatus. The significance of
these tumors for risk assessment remains to be determined (Morgan, 1997).
Appelman et al. (1988) studied the effects of bilateral intranasal electrocoagulation
damage on susceptibility to inhaled formaldehyde in male SPF Wistar (Cpb: WU) rats. Rats
were exposed 6 hours/day, 5 days/week for 13 or 52 weeks to 0, 0.1, 1.0, or 10 ppm (0, 0.12,
1.23, or 12.3 mg/m3) formaldehyde. These concentrations were chosen because the various
short-term studies performed in the same laboratory (described in Section 4.2.1.2) showed that
formaldehyde was noncytotoxic to the nasal mucosa at levels of 0.3, 1.0, and 2.0 ppm, slightly
cytotoxic at 3 and 4 ppm, and strongly cytotoxic at 10 and 20 ppm (Zwart et al., 1988; Wilmer et
al., 1987; Woutersen et al., 1987). Furthermore, because nasal tumors have only been found at
exposure concentrations that also induced severe degenerative, hyperplastic, and metaplastic
changes in the nasal epithelium (Griesemer et al., 1985; Squire and Cameron, 1984), Feron et al.
(1984) and the investigators at the TNO-CIVO Toxicology and Nutrition Institute postulated that
formaldehyde at a subcytotoxic concentration was only a very weak initiator without promoting
activity. Appelman et al. (1988) used an electrocoagulation method in this study to evaluate if
damage to the mucosa followed by compensatory cell proliferation might render the epithelium
vulnerable to subcytotoxic levels of formaldehyde. One-half of the rats used in the study
(10/group) were damaged bilaterally and then subjected to the first 6-hour exposure to
formaldehyde approximately 20-26 hours after the electrocoagulation procedure. Ten
undamaged rats/group were also exposed at each concentration for either 13 or 52 weeks.
Histopathologic examination included six standard cross-section levels in the nose; livers of all
rats killed at 14 weeks and of all control and 10 ppm exposed rats killed in week 53; larynges,
tracheas, and lungs of all rats of the control and 10 ppm exposed rats killed in week 53; and
This document is a draft for review purposes only and does not constitute Agency policy.
4-177 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
organs and tissues of control and 10 ppm exposed rats with an undamaged nasal mucosa killed in
week 53.
Yellow discoloration of the fur occurred in all animals of the two highest exposure
groups. Growth retardation was observed in the animals killed with or without damaged noses
after 2 weeks of exposure to 10 ppm formaldehyde. No toxicologically significant findings in
the body weights or organ weights of any animals in the other exposure groups were observed.
No relevant differences between groups were found in any of the hematological or urinary
parameters with the exception of frequent oliguria (p < 0.05) in the top exposure group without
nasal coagulation and killed in week 53. Three-way ANOVA revealed a significant increase in
TP content of the liver in rats with damaged noses as compared with rats with undamaged noses,
and there was a significant negative correlation between the formaldehyde exposure level and TP
in these same rats. Hepatic GSH was positively correlated with both nasal damage and age of
the animals. No treatment-related gross findings were observed in animals sacrificed at either 14
or 53 weeks except for yellow discoloration of the fur in rats exposed at the two highest
concentrations. No changes observed in the larynx, trachea, lungs, liver, or other tissues
evaluated were regarded as related to formaldehyde.
Few nasal lesions were noted in intact rats exposed at 0.1 or 1 ppm for either 13 or 52
weeks (n = 10/group). Focal squamous metaplasia was noted in a single animal exposed at
1 ppm for 13 weeks. Rats exposed at 10 ppm formaldehyde demonstrated clear pathology in the
respiratory epithelium progressing from 13 to 52 weeks, including squamous metaplasia, basal
cell hyperplasia, and focal rhinitis. Additionally, focal nest-like infolds of the epithelium were
present in 4 of 10 rats at 52 weeks, and minor changes to the olfactory epithelium were noted
(thinning/disarrangement and focal basal cell hyperplasia.)
All rats with damaged nasal passages exhibited similar minor pathology of the respiratory
epithelium at 13 and 52 weeks (squamous metaplasia, focal basal cell hyperplasia, and focal
rhinitis). Formaldehyde-related effects were note at 52 weeks, where the squamous metaplasia
of the respiratory epithelium was no longer noted in controls (versus 13 weeks) but was clearly
present in all formaldehyde-treatment groups, including progression from focal to diffuse lesions
(at 1 and 10 ppm) and keratinization (3/10 and 4/10 at 0.1 ppm and 10 ppm, respectively). The
formaldehyde effects on the respiratory epithelium were much more severe in rats with damaged
nasal passages, with all animals demonstrating thinning and disarrangement of the olfactory
epithelium and 8 of 10 rats exhibiting "loosely arranged submucosal tissue." Squamous
metaplasia and focal rhinitis of the olfactory epithelium were seen in less than half of the
formaldehyde-treated rats with damage. No changes in the olfactory epithelium due only to
electrocoagulation were encountered.
This document is a draft for review purposes only and does not constitute Agency policy.
4-178 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
The most notable effects of nasal damage from electrocoagulation were the ones at the
highest formaldehyde exposure (10 ppm) on the olfactory epithelium. Damage to the respiratory
epithelium also occurred more posteriorly in rats with damaged noses. Since electrocoagulation
often induced damage that included partial or complete loss of turbinates and septal perforation,
a likely explanation for the posterior distribution of the damage is an abnormal airflow pattern.
This gross damage to the nasal structure may have also disrupted normal mucous production and
flow. Therefore, formaldehyde-induced pathology appearing deeper in the nasal passages,
including the respiratory epithelium, may be due to formaldehyde penetrating more deeply into
the nasal passages and resulting in greater tissue doses in these areas.
Woutersen et al. (1989) conducted a lifetime study in parallel to the 1-year study
described above for Appelman et al. (1988). Male Wistar rats with nasal damage induced by
electrocoagulation (60/group) or without nasal damage (30/group) were exposed 6 hours/day,
5 days/week to the same concentrations as in the previous study (0, 0.1, 1.0, and 10 ppm) for
28 months or for 3 months followed by a 25-month observation period. The general condition
and behavior of the animals were checked daily. Body weight, organ weight, and gross
pathology were evaluated as described for Appelman et al. (1988). Histopathologic examination
was conducted on all animals at the standard six cross sections (see Figure 4-3).
No remarkable findings on behavior were observed except for yellowing of the fur in
animals at the two highest concentrations. There were no relevant differences in mortality (data
not shown). Growth retardation was observed relative to controls in animals with or without
damaged noses exposed to 10 ppm from day 14 onward. Body weights were generally slightly
lower in formaldehyde-exposed animals with an intact nasal mucosa and slightly higher in
exposed animals with damaged noses than in the corresponding controls.
The effects of formaldehyde exposure on the respiratory and olfactory epithelium after
28 months of exposure were similar to those reported for 52 weeks exposure (Appelman et al.,
1988): rhinitis, squamous metaplasia with some keratinization of the respiratory epithelium, and
thinning/disarrangement and slight squamous metaplasia of the olfactory epithelium at the
10 ppm exposure. Effects attenuated from the anterior to posterior sections (I—II, III, IV, and V-
VI). A low incidence of olfactory epithelium replaced by respiratory epithelium (<10%) and
vacuolation and atrophy of olfactory cells (<10%) was reported, this in part may be due to the
larger study size (30 rats per group versus 10). Squamous metaplasia in levels I—II of the
respiratory epithelium at 10 ppm was the only treatment-related pathology remaining in rats
exposed for 3 months followed by a 25-month recovery period.
Similarly, as reported by Appelman et al. (1988), rats with noses damaged by
electrocoagulation did demonstrate increased pathology of the respiratory epithelium.
This document is a draft for review purposes only and does not constitute Agency policy.
4-179 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Formaldehyde exposure at 10 ppm exacerbated these changes, and effects were noted in more
posterior sections than in rats without nasal damage (levels III, IV, and V). Olfactory pathology
was also greater in formaldehyde-treated rats: basal cell hyperplasia, replacement of olfactory
epithelium by respiratory epithelium (10-20% at level III and <10% at level IV). Although the
incidences are low, there is some evidence that effects on the olfactory epithelium may be
increased at the lower formaldehyde exposures (0.1 and 1 ppm.) Analysis of the number of
animals with olfactory pathology would be helpful to better understand the potential of low-level
formaldehyde effects on these less frequent lesions. Interestingly, the recovery of the olfactory
and respiratory epithelium seen in rats with undamaged nasal cavities after a 25-month recovery
period was not evident in rats with damaged noses. Formaldehyde-exposure effects are only
present at the 10 ppm exposure for the respiratory epithelium (squamous metaplasia, basal cell
hyperplasia), and the formaldehyde-related effects on the olfactory epithelium
(thinning/disarrangement, basal cell hyperplasia, and replacement by respiratory epithelium) are
seen at 0.1 and 1.0 ppm as well.
A single SCC, 1 out of 30 rats, was found in each 28-month formaldehyde-treatment
group (1/26, 1/28, and 1/26, respectively) but not in any control animals (n = 52). SCCs were
also noted in rats with noses damaged by electrocoagulation (1/54, 1/58, 0/56, and 15/58 for
control rats and the formaldehyde-treatment groups, respectively). These data clearly indicate a
synergistic effect of high formaldehyde exposure and nasal damage on the formation of SCCs in
rats. One adenosquamous carcinoma and one adenocarcinoma were also reported as increasing
the frequency to 17/58 for all tumors. Additionally SCC was present in two rats in the 0.1 and
1 ppm 3-month exposure groups with damaged noses only, although only one SCC was reported
in the 10 ppm 3-month groups with and without damaged noses. Rats not surviving to
28 months are included in these results, as well as the histopathology reported above. Since no
mortality data are reported, it should be noted that the incidence of both nasal lesions and tumors
are not controlled for early deaths.
In total, 30 tumors were examined from this study. In general, the tumors (26/30 or 87%)
were SCCs, and 69% (18/26) of these clearly originated from the respiratory epithelium lining
the septum or nasal turbinates. The eight other SCCs, derived from the epithelium lining the
nasolacrimal duct, were seen in connection with severe odontodystrophy and periodontitis or
might have originated from the skin or salivary glands. Four remaining rats bearing a nasal
tumor developed a small polypoid adenoma located on the nasoturbinate, an adenocarcinoma
originating from the olfactory epithelium, an adenosquamous carcinoma of the respiratory
epithelium lining the septum or turbinates, or a carcinoma in situ of epithelium in the
nasolacrimal duct.
This document is a draft for review purposes only and does not constitute Agency policy.
4-180 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.2.3.3. Hamsters. Dalbey (1982) examined the effects of inhaled formaldehyde alone for a
lifetime or combined with diethylnitrosamine (DEN) in an initiation-promotion study design
using male Syrian golden hamsters. For the first experiment, hamsters were exposed at either 0
or 10 ppm (0 or 12.3 mg/m3) formaldehyde in whole body chambers 5 hours/day, 5 days/week
for a lifetime (132 controls, 88 exposed). Histopathologic evaluations were carried out on two
transverse sections of the nasal turbinates (otherwise not specified), longitudinal sections of
larynx and trachea, and all lung lobes cut along the bronchus prior to embedding. In the
formaldehyde-only (10 ppm) experiment, mortality was increased relative to unexposed controls
(p < 0.05). No tumors and little evidence of toxicity to the nasal epithelium were observed.
There was no increase in rhinitis. Epithelial hyperplasia and metaplasia were increased in
formaldehyde-treated animals (5% incidence) versus none observed in controls.
The second set of experiments by Dalbey (1982) examined interaction of formaldehyde
exposure on tumor formation from DEN administered subcutaneously. The five treatment
groups included: (1) controls (n = 50); (2) formaldehyde only (n = 50); (3) DEN 0.5 mg, once
per week for 10 weeks (n = 100); (4) formaldehyde exposure for life with DEN injection for the
first 10 weeks given 48 hours after formaldehyde exposure (n = 27); and (5) DEN injection for
10 weeks, followed by formaldehyde exposure for life (n = 23). In all groups hamsters were
exposed at 30 ppm formaldehyde 5 hours/day, once a week. Histopathologic examinations were
conducted as above.
Although weekly exposures to formaldehyde alone (30 ppm once a week) did not
influence mortality, treatment with DEN alone significantly (p < 0.05) increased mortality above
that of untreated controls, and mortality was further elevated (p < 0.05) in the two groups
exposed to both DEN and formaldehyde compared with DEN alone. No respiratory tract tumors
were observed in untreated animals or those receiving only formaldehyde. DEN treatment alone
resulted in a high incidence (77%) of tumors (nasal, larynx, trachea, and lung). Formaldehyde
pre- or posttreatment did not further increase the number of TBAs-. All tumors observed were
classified as adenomas. Formaldehyde pretreatment nearly doubled the number of tumors per
animal in the trachea (but not lung or larynx) (p < 0.05). This increase in tumors initiated by
DEN given 48 hours after formaldehyde exposure suggests a role of formaldehyde- induced
changes in the respiratory tract in tumor promotion (e.g., cell proliferation and inflammation).
4.2.1.2.3.4. Summary. Chronic rodent studies of inhalation exposure to formaldehyde provide a
consistent picture of the agent's toxicity—especially on the URT—on which most studies focus.
All three species tested—hamsters, mice, and rats—had some degree of hyperplastic and
metaplastic change in the nasal passages. The pathology defined in acute and subchronic
This document is a draft for review purposes only and does not constitute Agency policy.
4-181 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
exposures is similarly described in chronic studies, where progression, severity, and presence in
more posterior sections of the nose increase with both the concentration and duration of
exposure.
Pathology of the respiratory epithelium includes rhinitis, goblet cell hyperplasia,
pseudoepithelial cell hyperplasia, squamous metaplasia, and dysplasia (see Table 4-34). At
higher exposures and longer durations of exposure, similar effects are seen on the olfactory
epithelium, present further into the nasal passages. In addition to hyperplasia and squamous
metaplasia, thinning and disarrangement of the olfactory epithelium noted and, in a few cases,
cell damage and replacement of olfactory epithelium with respiratory epithelium appear,
including loss of sensory cells (Woutersen et al., 1989; Kerns et al., 1983; Battelle Columbus
Laboratories, 1981).
Clear species differences in the severity of lesions are present. Although the bioassays in
mice, hamsters, and rats do represent similar exposure concentrations and duration of exposure,
hamsters exhibit little pathology and rats (three strains tested) exhibit gross toxicity and even
increased mortality. Mice similarly exposed exhibit a range of effects on the respiratory
epithelium but not near the severity seen in rats. Many factors may contribute to these observed
species differences. As Chang and Barrow (1983) reported, the increased RB of mice seems to
be protective of POE damage in comparison to that of rats. The reduced ventilation rate and
minute volume of rodents in the presence of a reactive gas can reduce the effective delivered
dose at the same exposure concentration (Chang and Barrow, 1983). Additionally, as illustrated
in the computational fluid dynamic (CFD) modeling (see Section 3.5), there are species
differences in nasal architecture that influence areas of formaldehyde absorption or flux into the
tissue. Localized differences in mucus flow and production as well as metabolic enzymes have
also been posited as having roles in differential toxicity of formaldehyde on the URT (see
Chapter 3).
Formaldehyde-induced tumors were present in exposed rats and mice and primarily
involved SCCs later in life (Kamata et al., 1997; Tobe et al., 1985; Kerns et al., 1983; Swenberg
et al., 1980). Although exposure of male Syrian hamsters to either 10 or 30 ppm did not result in
formaldehyde-induced nasal tumors, a classic initiation-promotion assay with DEN-induced
tumor formation did indicate that formaldehyde increased the tumor burden per animal, where
DEN induced tumors in 77% of the animals (Dalbey, 1982). This study suggests a role for
promotion in the observed carcinogenicity of formaldehyde. Less clear are the implications of
the synergistic effect of formaldehyde exposures and gross damage to the respiratory epithelium
by electrocoagulation on tumor formation (Woutersen et al., 1989).
This document is a draft for review purposes only and does not constitute Agency policy.
4-182 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-34. Summary of respiratory tract pathology from chronic inhalation exposures to formaldehyde
S"4
>3*
&
to
s
Species/strain
No./
group
Treatment"
Respiratory effects
Noncancer
LOAEL/NOAEL
Reference
Chronic bioassays
Mice
C3H mice (sex
unstated)
60
0, 41, 82, or 163 ppm
1 hour/day, 3 days/week for up
to 35 weeks.
Low- and mid-group mice then
exposed at either 122 or
244 ppm during weeks 35-70.
Pathology: Histologic changes in the tracheobronchial
epithelium by exposure, including basal-cell
hyperplasia, stratification squamous cell metaplasia, and
atypical metaplasia.
Carcinogenicity: No SCC formation was evident in
mice exposed to formaldehyde alone.
LOAEL = 41 ppm
No evidence of
carcinogenicity
Horton et al. (1963)
Male and female
B6C3F1 mice
120/sex
0, 2, 5.6, or 14.3 ppm
6 hours/day, 5 days/week for 24
months.
The protocol featured a 6-month
recovery period. Interim
sacrifices occurred at 6, 12, 18,
24, and 30 months.
Pathology: Rhinitis; hyperplasia, dysplasia, and
squamous metaplasia of the nasal epithelium; atrophy of
the olfactory epithelium; glandular adenitis and
nasolacrimal duct hyperplasia and metaplasia.
Carcinogenicity: Nasal SCC in male mice at 24 months
(2/17). No SCC in female mice.
LOAEL = 2 ppm
Evidence of
carcinogenicity
Swenberg et al.
(1980); Kerns et al.
(1983); CUT
(1982); Battelle
Columbus
Laboratories (1981)
Rats
Female Sprague-
Dawley rats
16
0 or 12.4 ppm
formaldehyde ± wood dust
6 hours/day, 5 days/week for
104 weeks.
Pathology: Squamous metaplasia and dysplasia.
Carcinogenicity: One of 16 rats exposed to
formaldehyde alone developed SCCs.
LOAEL = 12.4 ppm
Support for
carcinogenicity
Holmstrom et al.
(1989a)
Male and female
F344 rats
32/sex
0, 0.3, 2, or 14 ppm
6 hours/day, 5 days/week for
28 months.
Pathology: Increased rhinitis, hyperplasia, and
squamous metaplasia of the nasal respiratory epithelium
Carcinogenicity: Nasal SCCs in high concentration
rats (44%).
LOAEL = 0.3 ppm
Support for
carcinogenicity
Tobe et al. (1985)
Male F344 rats
32
0, 0.3, 2, or 15 ppm
6 hours/day, 5 days/week for
28 months.
Pathology: Squamous cell metaplasia and epithelial
hyperplasia.
Carcinogenicity: SCC (13/32), squamous cell
papilloma (3/32), and sarcoma (1/32).
LOAEL = 0.3 ppm
BMDio = 0.24 ppm
Evidence of
carcinogenicity
Kamata et al. (1997)
o
2 »
5 s
to
o
-------
K
s
TO
>3
Table 4-34. Summary of respiratory tract pathology from chronic inhalation exposures to formaldehyde
S"4
>3*
&
to
s
Species/strain
No./
group
Treatment"
Respiratory effects
Noncancer
LOAEL/NOAEL
Reference
Male Sprague-
Dawley rats
100
0 or 15 ppm 6 hours/day,
5 days/week for life.
Pathology: Squamous metaplasia, epithelial
hyperplasia, and polyps/papillomas.
Carcinogenicity: SCCs formed in the nasomaxillary
turbinates and nasal septum (25%).
LOAEL =15 ppm
Evidence of
carcinogenicity
Albert etal. (1982);
Sellakumar et al.
(1985)
Male and female
F344 rats
120/sex
0, 2, 5.6, or 14.3 ppm
6 hours/day, 5 days/week for
24 months. The protocol
featured a 6-month recovery
period. Interim sacrifices
occurred at 6, 12, 18, 24, and
30 months.
Pathology: Lesions of the nasal cavity were the
primary effects, including squamous metaplasia and
epithelial dysplasia, hyperkeratosis, goblet cell
hyperplasia, and rhinitis.
Salivary gland: atrophy, squamous metaplasia, and
sialadenitis.
Carcinogenicity: SCCs were evident in the nasal
cavity of high concentration rats, plus some polypoid
adenomas.
LOAEL = 2 ppm
Evidence of
carcinogenicity
Swenberg et al.
(1980); Kerns et al.
(1983); CUT (1982);
Battelle Columbus
Laboratories (1981);
Morgan et al.
(1986b)
Male F344 rats
90 and 150
controls
0, 0.7, 2, 6, 10, or 15 ppm
6 hours/day, 5 days/week for
24 months.
Pathology: Olfactory degeneration, squamous
metaplasia, epithelial hypertrophy and hyperplasia, and
mixed inflammatory cell infiltrate.
Carcinogenicity: SCCs and polypoid adenomas in the
nasal cavity
LOAEL = 2 ppm
Evidence of
carcinogenicity
Monticello et al.
(1996)
Male SPF Wistar
rats
10
0, 0.1, 1, or 10 ppm
6 hours/day, 5 days/week for 13
or 52 weeks.
An electrocoagulation method
was applied to damage the
noses of Vi of each study group.
Pathology: Formaldehyde-induced focal changes to the
respiratory and olfactory epithelium, including rhinitis,
hyperplasia, and metaplasia (10 ppm).
In rats with damaged noses: squamous metaplasia of the
respiratory epithelium increased at all formaldehyde
exposures. Pathology of the olfactory epithelium
increased at the 10 ppm exposure.
Carcinogenicity: No tumors noted; 1-year study
LOAEL = 0.1 ppm
in rats with damaged
nasal passages
NOAEL = 1 ppm for
rats with intact noses
Appelman et al.
(1988)
o
2 »
5 s
a, Co'
§- a
TO Sj-
§ ^
JS*
*
4^
00
4^
-------
K
s
TO
>3
Table 4-34. Summary of respiratory tract pathology from chronic inhalation exposures to formaldehyde
S"4
>3*
&
to
s
Species/strain
No./
group
Treatment"
Respiratory effects
Noncancer
LOAEL/NOAEL
Reference
Male Wistar rats
30 (without
nasal
damage),
60 (with
nasal
damage)
0, 0.1, 1, and 10 ppm
6 hours/day, 5 days/week for 28
months or for 3 months with a
25-month observation period.
An electrocoagulation method
was applied to damage the nasal
cavity.
Pathology: Intact noses: squamous metaplasia in the
high concentration group exposed for 28 months and
degeneration of the olfactory epithelium.
Changes were more severe in animals with damaged
noses .
Carcinogenicity: SCCs developed in 15/60 rats with
damaged noses exposed at 10 ppm. In other groups, the
incidence of nasal tumors was low irrespective of the
state of nasal damage.
NOAEL = 1 ppm
Evidence of
carcinogenicity
Woutersen et al.
(1989)
Hamsters
Male Syrian
golden hamsters
88 treated
132
controls.
0 or 10 ppm formaldehyde
5 hours/day, 5 days/week for
life.
Pathology: Increased mortality. Epithelial hyperplasia
and metaplasia increased in formaldehyde-treated
animals (5% incidence)
Carcinogenicity: No tumors reported.
LOAEL =10 ppm
No evidence of
carcinogenicity
Dalbey (1982)
Male Syrian
golden hamsters
50
0 or 30 ppm
5 hours/day, 1 day/week for life
± injections with 0.5 mg DEN.
Pathology: Increased mortality in conjunction with
DEN—above DEN-only treated animals. Respiratory
pathology not reported.
Carcinogenicity: Only hamsters receiving DEN
developed tumors (77%, adenomas). There was an
increase in the number of tumors per TB As in the
trachea of animals exposed to formaldehyde 48 hours
prior to DEN (but no increase in TBAs).
LOAEL = 30 ppm.
Evidence for
formaldehyde as a
promoter
Dalbey (1982)
o
2 »
5 s
to
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
32
33
34
35
4.2.1.2.4. Summary of respiratory yatholosv. The progressive pathology of the nasal passages
from inhalation exposure to formaldehyde is well documented, especially in rodents (rats and
mice) (see Tables 4-9, 4-15, 4-28, 4-34). Although there are species differences in tissue dose
(Section 3.4) due to variations in nasal architecture and breathing patterns, the nature and
progression of the pathology is fairly well conserved across species, including nonhuman
primates. The observed formaldehyde-induced pathology includes disruption of the mucociliary
apparatus, rhinitis (serous and purolent), hyperplasia (cell proliferation), metaplasia (transition of
cell type), dysplasia (disarrangement of cells), nest-like infolds and invaginations of the
epithelium, thinning of the epithelial layer and focal to diffuse lesions, atrophy of the olfactory
epithelium, thickening and keratinization (usually of squamous metaplasia), tumors (adenoma,
sarcoma, carcinoma) (Section 4.2.2).
Progression of lesions can be viewed as progression from the anterior to posterior
sections of the nasal cavity or as a progression in severity of lesions at a particular location (e.g.,
level or region) of the nasal passages. In both cases, progression is evident with increasing
exposure concentration and with increasing duration of exposure (Kamata et al., 1997;
Monticello et al., 1996; Morgan et al., 1986b; Takahashi et al., 1986; Sellakumar et al., 1985;
Kerns et al., 1983; Albert et al., 1982). The data suggest that concentration and duration of
exposure do not act in a simply cumulative manner (e.g., C x t). Additionally the influence of
concentration, duration, and repeated exposure may be different for various effects. For
example, some lesions may be transient (e.g., low-exposure cell proliferation), others may have a
threshold and vary little after that (e.g., rhinitis). Additionally, as the nasal epithelium responds
with both adaptive and adverse epithelial changes, the absorption of formaldehyde into the tissue
at that location may be reduced. As respiratory epithelium transitions to squamous metaplasia,
the effective tissue dose of formaldehyde increases posterior to these lesions. As barriers to
formaldehyde flux into the tissue develop (e.g., squamous metaplasia, keratinization),
formaldehyde penetrates more deeply into the nasal passages (Kimbell et al., 2006). Therefore,
although both concentration and duration of exposure do effect the adverse effect, the
relationship is difficult to define and in fact may be different for various adverse effects.
Respiratory histopathology has been commonly reported in response to exposure to
formaldehyde in rats and mice (Lino dos Santos Franco et al., 2006; Javden and Taher, 2000;
Kamata et al., 1996a, b; Cassee and Feron, 1994; Bhalla et al., 1991; Monteiro-Riviere and Popp,
1986; Buckley et al., 1984; Chang et al., 1983), rabbits (Ionescu et al., 1978), hamsters
(Schreibner et al., 1979), and rhesus monkeys (Monticello et al., 1989). The histopathologic
lesions ranged from inflammation to ulceration, necrosis, and metaplasia that occurred in nasal
turbinates, maxilloturbinates, and goblet and microvillus cells (Bhalla et al., 1991). These effects
This document is a draft for review purposes only and does not constitute Agency policy.
4-186 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
were observed at a variety of doses (e.g., 10 ppm for 4 hours, 3.13 ppm for 6 hours for 1, 2, or
4 days, 6 or 15 ppm). Wilmer et al. (1989, 1987) assessed whether a dose and time-dependent
interaction (C x t) is associated with histopathologic lesions. Results indicated that
concentration, rather than duration or cumulative exposure, correlates best with severity of
lesions (Wilmer et al., 1989, 1987).
Histopathologic lesions and changes to biochemistry have been reported in the lung as
well, though these effects were observed following a high dose of formaldehyde. In addition,
changes in clinical chemistry, P450 expression and activity in lung tissue, and gene expression
that is phenotypically anchored to the observed respiratory pathology have been reported.
Extrapulmonary effects have also been noted, including changes in liver chemistry, relative brain
weight, and focal, chronic inflammation in the heart and kidney. Most of these changes occurred
at exposures of 20 ppm, and those that occurred at lower formaldehyde exposures (3.7 ppm)
could not be strictly correlated with formaldehyde exposure.
Some researchers have reported formaldehyde-induced effects in the pulmonary region in
rats, mice, and rabbits. Kamata et al. (1996a) observed reduced lipid content of pulmonary
surfactant in rats exposed to 128.4 or 294.5 ppm formaldehyde. Kamata et al. (1996b) reported
biochemical changes in lung homogenates and altered lipid content of BAL at 145.6 ppm
formaldehyde. Lino dos Santos Franco et al. (2006) observed increased leukocytes (and
neutrophils) and degranulated mast cells recovered in BAL fluid (concentration of 1% formalin
not provided). In rabbits, Ionescu et al. (1978) observed frank necrosis of lung parenchyma after
aerosol inhalation of 3% formalin for 3 hours/day for 50 days (concentration of formaldehyde
not provided). These pulmonary effects may be due to frank toxicity resulting from the high
dose of formaldehyde used in these studies.
Several recent toxicogenomics studies have assessed gene expression changes in nasal
and lung tissue in animals and in humans by using in vivo and in vitro approaches. Hester et al.
(2005, 2003) documented changes in gene expression associated with DNA repair and apoptosis
in nasal tissue from male rats after a single instillation of formaldehyde. Other gene expression
changes were observed in those genes related to xenobiotic metabolism and in cell cycle and
repair. These preliminary results provide an initial basis for forming a phenotypically anchored
set of gene expression changes associated with exposure to formaldehyde and may assist in
determining the underlying MOA, as will be discussed in Section 4.5. Sul et al. (2007)
investigated gene expression genes in lung tissue from formaldehyde-exposed rats. Yang et al.
(2005) performed a proteomics analysis by using lung tissue extracted from formaldehyde-
exposed rats. Two studies used human tracheal cell lines to investigate formaldehyde-induced
gene expression changes in vitro (Lee et al., 2008, 2007). However, the relevance of these
This document is a draft for review purposes only and does not constitute Agency policy.
4-187 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
findings to actual exposures remains unknown. In total, toxicogenomics studies hold promise,
but they must be interpreted with caution until results can be replicated and phenotypically
linked to observable changes.
Thus, formaldehyde-induced respiratory pathology has been commonly described in the
nasal passages and includes cellular proliferation, mucociliary function, and histopathologic
lesions. Pulmonary effects have been documented as well but at high doses. The nasal
pathology may occur as a result of both concentration and duration components of exposure.
4.2.1.2.5. Cell proliferation. Formaldehyde-induced cell proliferation has been demonstrated
under range of exposure conditions in vivo and in vitro as well (Chapter 3). Formaldehyde-
induced mitogenesis may be a primary effect (as demonstrated in the in vitro work) or secondary
to adaptive responses and tissue remodeling (Swenberg et al., 1983). This section provides a
comprehensive discussion of formaldehyde effects on cell proliferation in the epithelial tissues in
the respiratory tract. The majority of the work discussed investigates cell proliferation with
in vivo labeling of proliferating cells, although additional methods, such as flow-cytometry, have
been employed in some instances.
Swenberg et al. (1986) conducted a series of experiments in rodents to assess cell
proliferation in the nasal mucosa after formaldehyde inhalation. Radiolabeled thymidine
[3H]-thymidine was injected intraperitoneally (I.P.) into male F344 rats and B6C3F1 mice after
formaldehyde exposure to assess the extent of in vivo incorporation into proliferating cells. Two
hours later, animals were sacrificed and the nasal passages were fixed, embedded, and sectioned
to examine the nasal mucosa. Slides were exposed for 12 weeks and developed to identify cells
that incorporated the radiolabeled thymidine. The percentage of labeled cells, as indicated by the
presence of five or more grains over the nucleus, was determined by visual count. A total of
4,000 or 1,500 cells were counted per section for rats and mice, respectively.
The first set of studies reported by Swenberg et al. (1986) compared the dose response of
rats and mice. Animals were exposed to 0, 0.5, 2, 6, or 15 ppm (0, 0.61, 2.45, 7.36, or
18.4 mg/m3) formaldehyde 6 hours/day for 3 days. Tritiated thymidine for cell labeling was
injected 2 hours after the end of exposure. No change in the percentage of labeled cells was seen
after 0.5 or 2 ppm formaldehyde exposure. However, the nasal passages of rats exposed at 6 and
15 ppm showed 10- to 20-fold increases over controls in LI at level 2. A similar cell
proliferation response was seen in mice treated with 15 ppm formaldehyde, although no increase
over control was seen in mice exposed to 6 ppm formaldehyde. These findings are consistent
with other data that indicate rats are more sensitive to formaldehyde exposure than mice. This
may be due to differences in the reflex apneic response between the two species. As discussed in
This document is a draft for review purposes only and does not constitute Agency policy.
4-188 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Section 4.2.1.1, mice maintain decreases in minute volume in response to formaldehyde, which
results in a lower overall effective internal dose to the mice.
Comparing cell proliferation rates after 2 versus 18 hours of exposure, Swenberg et al.
(1986) found that the longer exposure duration gave twice the cell proliferation rates after
repeated exposures. Therefore, these researchers conducted a second dose-response study to
examine cell proliferation 18 hours after exposure instead of the shorter exposure duration. The
dose-response study varied dose as well as duration of treatment. Rats were exposed 6 hours/day
to either 0.5, 2, or 6 ppm (0.61, 2.45, or 7.36 mg/m3) formaldehyde over periods of 1, 3, or 9
days. Formaldehyde exposure at 0.5, 2, or 6 ppm for 1 day increased cell proliferation in the
nasal epithelium. However, these increases were transient, and cell proliferation was not
increased after 3 or 9 days of exposure to 0.5 ppm or 2 ppm formaldehyde. Although still
elevated after a 3-day exposure to 6 ppm formaldehyde, cell proliferation returned to control
values after 9 days of exposure to 6 ppm formaldehyde (Swenberg et al., 1986). Therefore,
although concentration is a major determinant of cell proliferation, duration of exposure also
influenced formaldehyde-induced cell proliferation in the nasal epithelium.
Swenberg et al. (1986) directly tested the effects of cumulative exposure versus
concentration for both mice and rats. Animals were treated with one of three regimens, resulting
in the same C x t product: 3 ppm x 12 hours, 6 ppm x 6 hours, or 12 ppm x 3 hours, each
exposure resulting in 36 ppm-hours. The animals were exposed once a day for either 3 or 9 days.
Tritiated thymidine was injected 18 hours after exposure to label of proliferating cells. Tissue
sections from levels 1 and 2 of the nasal passages were examined in each case, and the
percentage of cells labeled was reported as the percentage of proliferating cells (Figure 4-10).
Cell proliferation at level 1 in the nasal cavity was much greater than at level 2 for all C x
t combinations of formaldehyde exposure in both mice and rats (Figure 4-10). The authors noted
that level 1 is more anterior and lacks significant defense from the mucociliary apparatus, which
may account for the observed greater sensitivity to formaldehyde. At all C x t exposure
products, 3 days of exposure resulted in greater cell proliferation than 9 days of exposure. This
was true for both species and for both examined levels of the nasal cavity. The decrease in cell
proliferation by day 9 is consistent with data on rats labeled 18 hours postexposure (Swenberg et
al., 1986).
This document is a draft for review purposes only and does not constitute Agency policy.
4-189 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
30
10
I
-------
1
2
3
4
5
6
7
8
9
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
the important metric. Therefore, it may be concluded that cell proliferation for level 1 of the
nasal passages, where there is less protection of the epithelium, is influenced by concentration,
time, and duration of exposure. Cell proliferation at level 2 appeared to be more dependent on
concentration than time of exposure (Swenberg et al., 1986).
Cassee and Feron (1994) reported a qualitative increase in histochemical staining for
proliferating cell nuclear antigen (PCNA) in the respiratory epithelium of the nasoturbinates,
maxilloturbinates, septum, and lateral wall at levels 2 and 3 of rat nasal passages after repeated
exposures to 3.5 ppm (4.29 mg/m3) formaldehyde 22 hours/day for 3 days. While no increases
were seen in olfactory epithelium, frank necrosis, squamous metaplasia, and hyperplasia of both
ciliated and nonciliated epithelium were noted at these section levels.
Quantitative cell proliferation studies have been conducted by several researchers in the
same laboratory (Reuzel et al., 1990; Wilmer et al., 1989; Zwart et al., 1988; Wilmer et al., 1987;
Woutersen et al., 1987) (Summary Table 4-39). These studies build off of those of Swenberg et
al. (1986), who labeled proliferating cells with [3H]-thymidine in assessing cell proliferation
within the nasal mucosa. The studies, all performed in male albino Wistar rats and using a
similar experimental design, provide the basis for comparing different exposure levels and dose
regimens across studies. Wilmer et al. (1987) demonstrate a concentration-dependent increase in
cell proliferation after 3 days of repeated 8-hour exposures at 5, 10, or 20 ppm (6.13, 12.3, or
24.6 mg/m3) formaldehyde, regardless of continuous versus interrupted exposure conditions
(2.83, 8.87, and 19.8 versus 0.86% proliferation in controls). Similar trends were seen when the
repeated continuous exposures were extended for 4 weeks, but cell proliferation was not
maintained at the same levels. As observed by Swenberg et al. (1986), these results suggest that
duration of repeated exposures may be an important determinant of cell proliferation rates.
Woutersen et al. (1987) reported that the majority of the dose-dependent increases in cell
proliferation seen at section level 3 after 3 days of repeated 6-hour exposures to 10 and 20 ppm
(12.3 and 24.6 mg/m3) formaldehyde occurred in areas of the epithelium showing "clear
squamous metaplasia and hyperplasia." Cell proliferation rates in metaplastic epithelium of
29.5 and 33.2% were much higher than the 1.4 to 2.8% proliferation in the visibly unaffected
respiratory epithelium from rats exposed at 10 ppm formaldehyde. Although there was a slight
trend towards increased cell proliferation in the visibly unaffected epithelium of exposed animals
compared with unexposed controls, the majority of increased cell proliferation resulting from
exposure to 10 and 20 ppm formaldehyde was attributed to the metaplastic epithelium.
Similarly, dose-dependent increases in cell proliferation seen at level 3 after 3 days of
repeated 6-hour exposures at 0.3, 1, and 3 ppm (0.37, 1.23, and 3.68 mg/m3) formaldehyde
(p < 0.001) corresponded to focal basal cell hyperplasia and loss of cilia (Woutersen et al., 1987).
This document is a draft for review purposes only and does not constitute Agency policy.
4-191 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
No necrosis or focal erosion was noted at these levels of formaldehyde exposure. Cell
proliferation was not sustained at this location, and no lesions were noted after 13 weeks of
repeated 6-hour exposures. The authors hypothesized that defensive mechanisms, such as the
mucociliary apparatus, may have provided greater protection of the mucosa at level 3. Swenberg
et al. (1986) drew a similar conclusion when evaluating extended exposures, suggesting that
more posterior sections had a greater adaptive ability than those anterior sections with little
mucociliary function. Both Woutersen et al. (1987) and Swenberg et al. (1986) reported
sustained cell proliferation and development of lesions in the more anterior cross section.
Repeated exposures to 3 ppm formaldehyde (6 hours/day) resulted in significant increases in cell
proliferation in the epithelial cells at level 2, with accompanying disarrangement, focal
hyperplasia, and squamous metaplasia (Woutersen et al., 1987). Although no cell death was
observed at level 2 when viewed by light microscopy, "strongly indented and disarranged nuclei"
were seen by electron microscopy, which may be consistent with apoptosis (Woutersen et al.,
1987). However, later work in the same laboratory indicated no increased cell proliferation at
levels 2 or 3 in male Wistar rats exposed to formaldehyde at 1 or 2 ppm (1.23 and 2.45 mg/m3)
(8-hour repeated exposures for 3 days or 13 weeks) and only minimal response in rats exposed at
4 ppm formaldehyde (interrupted 8-hour exposures for 3 days or 13 weeks) (Wilmer et al.,
1989).
Reuzel et al. (1990) published the only report in which formaldehyde effects on cell
proliferation were studied for longer daily exposure durations: 22 hours/day versus 6-
8 hours/day. Male Wistar rats were exposed to formaldehyde, ozone, or the combination of the
two 22 hours/day for 3 consecutive days. The concentrations of formaldehyde were 0.3, 1.0, or
3.0 ppm (0.37, 1.23, or 3.68 mg/m3). Rats were injected with [3H]-thymidine 2 hours rather than
18 hours after the last exposure. Cell proliferation was quantified by enumerating the percentage
of labeled cells in fixed and stained tissue sections. Cell proliferation on the nasoturbinates,
maxilloturbinates, lateral wall, and septum at levels 2 and 3 were quantified and reported
separately. Cell proliferation was increased at all locations in level 2 at 3 ppm formaldehyde
exposure (p < 0.05) but not at 0.3 or 1 ppm exposures (Summary Table 4-39). Whereas
proliferation of cells in the nasoturbinate, maxilloturbinate, and septum was nearly undetectable
in control animals, 4, 5, and 3% proliferation was reported after repeated 22-hour exposures to 3
ppm formaldehyde. Basal proliferation in the lateral wall was greater than in other areas,
approximately 1% increasing to 6% after exposure to 3 ppm formaldehyde. Although basal
levels of cell proliferation were slightly higher in all areas of level 3, formaldehyde had no
significant effects on cell proliferation in the level 3 areas evaluated. There was a slight trend for
increases at 3 ppm, but all proliferation rates were below 1%. Exposure to 3 ppm formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
4-192 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
also damaged the respiratory epithelium at levels 2 and 3, where cell disarrangement and
hyperplastic and metaplastic lesions were reported.
Roemer et al. (1993) investigated the effects of formaldehyde exposure on cell
proliferation in the trachea and lung in addition to nasal mucosa. Male Sprague-Dawley rats
were exposed head only to 2, 6, or 20 ppm (2.45, 7.36, or 24.5 mg/m3) formaldehyde 6 hours/day
for either 1 or 3 days. Proliferating cells were labeled with 5-bromodeoxyuridine (BrdU), the
label injected 16-22 hours after formaldehyde exposure ended. Free lung cells were harvested
by tracheal lavage, and the majority of isolated cells were MPs (>97%). Epithelial cells were
isolated from the nasal and tracheal mucosa by dissection, physical disaggregation, and enzyme
treatment to release epithelial cells. All cells were fixed and stained with fluorescent dyes to
detect BrdU and total DNA. Flow cytometry was used to determine the percentage of BrdU-
labeled cells as a measure of cell proliferation. Cells undergoing unscheduled DNA synthesis
(e.g., DNA repair) were excluded by cell cycle analysis.
The proportion of BrdU-labeled cells from the nose and trachea increased two- to
threefold above control values after a single 6-hour exposure to formaldehyde (Table 4-35). The
lowest effective dose for increased cell proliferation was 2 ppm for nose and tracheal cell
proliferation (p < 0.05). However, increased proliferation in the nasal mucosa at the lowest dose
was transient, returning to control levels after a 3-day exposure. Cell proliferation remained
increased in the nasal mucosa after exposure to 6 or 10 ppm (7.36 or 12.3 mg/m3) formaldehyde
for 3 days. In contrast, proliferation of tracheal cells appeared to be reduced as a result of a
3-day exposure to 2 or 6 ppm formaldehyde. A similar trend was seen in free lung cells, but the
differences were not statistically significant.
The flow cytometry employed by Roemer et al. (1993) allowed for subtle changes in
proliferation rates to be measured with good discrimination. However, the method of cell
isolation did not allow examination of proliferation rates in discrete regions of the mucosa,
which may have attenuated the magnitude of the response. Additionally, proliferation rates
represent a mix of cell types that were not separated in this analysis, making the findings difficult
to interpret. This may be especially noteworthy in the free lung cells that were reportedly
primarily MPs.
This document is a draft for review purposes only and does not constitute Agency policy.
4-193 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Table 4-35. Cell proliferation in nasal mucosa, trachea, and free lung cells
isolated from male Wistar rats after inhalation exposures to formaldehyde
1 Day"
Control
2 ppm
6 ppm
20 ppm
Nose
1.3b
2.4°
3.7°
2.7
Trachea
1.2
3.1°
2.1°
2.8
Lungd
1.8
2.6
3.3
3.1
3 Days
Control
2 ppm
6 ppm
20 ppm
Nose
1.3
1.4
2.5°
2.3°
Trachea
1.2
0.3°
o
©
2.5°
Lung
1.8
2.2
2.4
5.1
""Exposures were 6 hours/day.
Proliferation is measured as the percent of BrdU-labeled cells.
Statistically different from controls (p < 0.05).
dThe majority of free lung cells were MPs (97%).
Source: Roemer et al. (1993).
Monticello et al. (1990) investigated whether changes in cell proliferation rate correlated
with areas of cell injury or with areas that developed tumors due to formaldehyde exposures by
using a unique metric of cell proliferation. They hypothesized that treatment-related effects on
cell populations could influence the apparent cell proliferation measured as LI, even though no
proliferative effect had occurred. For example, cell death could give an apparent increased
proliferation as a LI (% cells proliferating) by reducing the total number of cells present. This
would be especially true for a stratified epithelium, where the number of basal cells in active
proliferation may not change but cells above the basal layer might die or slough off, thereby
reducing the overall number of population of cells counted. The unit length labeling index
(ULLI) metric was developed to normalize proliferation rates against length of basal membrane
rather than cell population. However, application of a ULLI to the pseudostratified epithelium of
the nasal mucosa introduced additional complexities. First, undamaged mucosa has a single
layer of epithelial cells that have the capability for cell proliferation. Second, cells only become
layered in response to cell damage as a protective measure. Therefore, the total cells present and
the linear cell density should be considered, as well as the number and density of proliferating
cells, in developing an understanding of the proliferative response of these tissues to toxic insult.
Monticello et al. (1990) directly compared the apparent effects of formaldehyde exposure
on cell proliferation when quantified as an LI or as a ULLI. Male F344 rats were divided into
groups (n = 6) and exposed to 0, 2, 6, or 15 ppm (0, 2.45, 7.36, or 18.4 mg/m3) formaldehyde
6 hours/day, 5 days/week for 12 weeks. Rats were administered [3H]-thymidine continuously for
the last 5 days of exposure by surgically implanted osmotic pumps. After sacrifice, nasal
This document is a draft for review purposes only and does not constitute Agency policy.
4-194 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
passages were fixed, and sections from standard level 3 were prepared for examination. Cell
proliferation was quantified at the midseptum and the lateral meatus at this level. Basement
membrane length, total number of cells present, and number of labeled proliferating cells were
recorded for each location. Each of these areas also was scored for the presence of nasal lesions.
The formaldehyde-related lesions included epithelial hyperplasia, squamous metaplasia,
and acute inflammation. These lesions were most severe in animals exposed to 15 ppm, mild at
6 ppm, but absent at 2 ppm. Cell proliferation, measured either as LI or ULLI, was increased in
the level 3 septum and lateral meatus after 13 weeks of exposure to 15 ppm formaldehyde but
not to 6 or 2 ppm (Table 4-36). There was a slight increase in both cell number and labeled cells
in the lateral meatus of rats exposed to 6 ppm formaldehyde, but both measures of proliferation
were unchanged from controls. The increased proliferation in the lateral meatus at 15 ppm was
entirely due to an increased number of labeled cells. Total cells were unchanged at 15 ppm;
therefore, both Lis demonstrated a similar increase over control. In addition to increased labeled
cells in the septum at 15 ppm, total cells were increased from 470 to 640 (p < 0.05). Where the
total cells and linear cell density were increased, the ULLI was proportionally increased over the
LI. These observations are consistent with the development of squamous metaplasia and
hyperplasia seen at 15 ppm. However, while both LI and ULLI showed an eightfold increase in
cell proliferation in the lateral meatus, they gave different results in the septum where cell
number was increased by formaldehyde treatment. LI increased 19-fold and ULLI 25-fold with
repeated exposures to 15 ppm formaldehyde. Although these data are based on only 5-6
animals/group, and only in an extended study, the results suggest that the ULLI and LI may not
be proportional under all conditions studied. In similar experiments the LI and ULLI provided
different indices of proliferation in the olfactory epithelium after methyl bromide exposure
(Monticello et al., 1990). Methyl bromide exposure decreased cell number/mm of basement
membrane in a time-dependent manner, and the LI and ULLI were not proportional across these
changes. At day 3 there was an increase in labeled cells but a decrease in total cells; therefore,
the LI was increased greater than 20-fold, where the ULLI was only increased eightfold. The
authors endeavored to explain why the ULLI and LI yielded different findings. Where ULLI is a
more time-efficient method of assessing cell proliferation, the authors suggested that
representative areas should be quantified by LI to better understand the nature of increased
ULLI.
This document is a draft for review purposes only and does not constitute Agency policy.
4-195 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-36. The effect of repeated formaldehyde inhalation exposures for
2 3 months on cell count, basal membrane length, proliferation cells, and two
3 measures of cell proliferation, LI and ULLI, in male F344 rats
4
Formaldehyde exposure level
(6 hours/day, 5 days/week for 3 months)
0 ppm
2 ppm
6 ppm
15 ppm
Lateral meatus
Total cells
1,800 ± 100
1,800 ± 150
2,300a± 1700
1,900 ± 160
BM length (mm)b
12.7 ±0.6
11.9 ± 0.5
13.4 ±0.3
11.6 ±0.7
Cells/mm BM
150 ±5
150 ± 10
170 ± 10
150 ±5
Labeled cells
130 ± 10
130 ±20
210 ±30
1,400 ± 130
LI
7.2%°
7.2%
9.1%
73.7%
ULLI
10.2 cells/mmd
10.9 cells/mm
15.7 cells/mm
120.7 cells/mm
Septum
Total cells
470 ± 20
460 ± 30
470 ± 20
640a ± 20
BM length (mm)
2.9 ±0.1
2.7 ±0.1
2.9 ±0.1
2.9 ±0.1
Cells/mm BM
160 ± 10
170 ± 10
160 ±3
220a± 10
Labeled cells
20 ± 1
40 ± 10
10 ±2
250 ± 50
LI
4.3%
8.7%
2.1%
39%
ULLI
6.9 cells/mm
14.8 cells/mm
3.45 cells/mm
86.2 cells/mm
5
6 ""Different from control, p < 0.05.
7 bBM is basal membrane length in mm.
8 Calculated from group averages: LI = (labeled cells)/total cells
9 Calculated from group averages: ULLI = (labeled cells)/BM length
10
11 Source: Monticello et al. (1990).
12
13
14 Monticello et al. (1990) reported similar results in a contemporary abstract; although
15 treatment groups were slightly different than in the above experiments, the findings were similar.
16 Rats were exposed to 0, 0.7, 2, 6, 10, or 15 ppm (0, 0.86, 2.45, 7.36, 12.3, or 18.4 mg/m3)
17 formaldehyde 6 hours/day for 4 days, 6 weeks, or 3 months. ULLIs were determined in the
18 septum and lateral meatus (methods not detailed). It is not stated whether [3H]-thymidine
19 labeling was carried out by injection or continuous infusion. Significant increases in cell
20 proliferation were reported after repeated exposures to 6, 10, and 15 ppm for 4 days and 6 weeks.
21 After 3 months of exposure, cell proliferation was still increased in rats exposed to 10 and
22 15 ppm formaldehyde. The authors noted that, although increased cell proliferation was seen at
23 earlier time points, sustained increased cell proliferation was only seen at 10 and 15 ppm, which
24 they considered the clearly carcinogenic doses.
25 Monticello et al. (1991) applied the ULLI measurements in evaluating formaldehyde
26 effects on cell proliferation after short-term and subchronic repeated exposures. Six male F344
27 rats/group were exposed to 0, 0.7, 2, 6, 10, or 15 ppm (0, 0.86, 2.45, 7.36, 12.3, or 18.4 mg/m3)
This document is a draft for review purposes only and does not constitute Agency policy.
4-196 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
formaldehyde 6 hours/day for 1, 4, or 9 days and for 6 weeks, using a 5 days/week regimen.
Rats were injected with [3H]-thymidine 18 hours postexposure to label proliferating cells. All
animals were sacrificed 2 hours later. Nasal passages were fixed, and sections from levels 2 and
3 were prepared for examination. Cell proliferation was quantified for three locations in level 2
(specifically, the lateral meatus, midseptum, and medial aspect of the maxilloturbinate) and for
two regions of level 3 (the lateral wall and midventral septum). Each of these areas also was
scored for the presence of nasal lesions.
As discussed above, proliferating cells were visually identified by the number of grains
over the nucleus, 10 grains indicating a proliferating cell. Cell proliferation was quantified as the
number of proliferating cells per length of basement membrane (cells/mm) and reported as a
ULLI. The report does not indicate the length of membrane viewed for each section as an
indication of how representative the counts are for each region. Lesions associated with
formaldehyde exposure may change the density of cells/mm of basement membrane (Monticello
et al., 1990). Areas of disarranged cells, erosion, metaplasia, or layering of epithelial cells may
exhibit different cell profiles. These processes would alter cell density, and therefore the ULLI,
independent of differential proliferation rates. As such, it is not expected to be proportional to
cell proliferation rates across conditions that have the potential to change cell density
(Monticello et al., 1990).
No formaldehyde-induced epithelial lesions or increases in the ULLI were seen in rats
exposed to 0.7 or 2.0 ppm formaldehyde, regardless of duration (Table 4-37). Formaldehyde-
induced lesions were present in all regions of the nasal epithelium after exposures to 10 and
15 ppm formaldehyde, regardless of duration (Monticello et al., 1991). Incidence and severity of
the lesions increased with concentration and duration of treatment and were correlated to areas
with increased cell proliferation. Rats exposed to 6 ppm formaldehyde developed lesions in the
level 2 nasal passages, where the ULLI was clearly elevated, but not in the deeper level 3
passages. For example, no formaldehyde-related lesions were seen at the lateral meatus and
septum of level 3 at 1, 4, and 9 days of repeated exposure at 6 ppm, although cell proliferation
was increased. This transient increase in ULLI returned to near-control levels after 6 weeks of
repeated exposure (Table 4-37). Monticello et al. (1991) suggested that cell proliferation is a
more sensitive indicator of cellular response and not necessarily dependent on cellular necrosis.
This document is a draft for review purposes only and does not constitute Agency policy.
4-197 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-37. Formaldehyde-induced changes in cell proliferation (ULLI) in
2 the nasal passages of male F344 rats exposed 6 hours/day
3
Exposure concentration
Location3
0 ppm
0.7 ppm
2 ppm
6 ppmb
10 ppmb
15 ppmb
Level 2: lateral meatus
1 day
2.16
1.31
2.36
16.9b
11.2b
12.7b
4 days
1.46
1.37
1.72
30.5b
20.9b
25.8b
9 days
1.44
1.20
1.73
23.5b
28.6b
24.6b
6 weeks
0.91
0.88
1.36
14.4b
23.9b
28.7b
Level 2: midseptum
1 day
1.08
1.01
1.69
3.85
17.9b
16.7b
4 days
1.03
0.97
0.67
10.0b
26. lb
29. lb
9 days
1.09
0.80
0.97
10.9b
19.6b
29. lb
6 weeks
0.41
0.24
0.68
2.10
21.4b
25.9b
Level 2: medial
maxilloturbinate
1 day
2.49
1.75
2.81
18.15b
5.9
5.3
4 days
1.36
1.54
1.09
25.03b
20.3b
19.4b
9 days
1.38
0.80
1.48
22.54b
21.0b
28.7b
6 weeks
1.02
1.21
1.11
16.32b
26. lb
25.lb
Level 3: lateral meatus
1 day
1.83
1.72
2.46
7.53b"°
14.5b
16.4b
4 days
1.10
1.27
1.09
8.77\c
20.0b
30.8b
9 days
1.36
1.40
1.74
7.35b'c
30.6b
40.4b
6 weeks
0.98
0.91
0.86
2.08
24.2b
34.8b
Level 3: midseptum
1 day
3.02
1.74
2.39
4.20
24.4b
19.3b
4 days
2.81
3.09
1.43
9.22\c
18.7b
34.4b
9 days
1.68
1.06
1.43
9.50\c
28.6b
32.5b
6 weeks
2.18
1.54
2.57
2.58
14.0b
27.5b
4
5 aULLI is expressed as the number of labeled cells/mm of basement membrane.
6 indicates significantly different from control, p < 0.05.
7 Indicates a location where epithelial lesions were not seen by light microscopy.
8
9 Source: Monticello et al. (1991).
10
11
12 The sustained cell proliferation at the lateral meatus and midseptum in rats exposed to 10
13 and 15 ppm formaldehyde, locations where SCCs are known to arise, supports a role for
14 compensatory cell proliferation in tumor development. However, Monticello et al. (1991) noted
15 that regional differences in sustained cell proliferation do not always correspond to the
16 occurrence of nasal tumors, primarily SCCs, in formaldehyde-exposed rats. Where sustained
17 cell proliferation has been demonstrated in the medial maxilloturbinate (MMT) at level 2
18 (Monticello et al., 1991), SCCs have not been found to originate in this area at similar exposures
19 (Monticello et al., 1996; Woutersen et al., 1989). Monticello et al. (1991) suggested that the
20 findings of Bermudez and Allen (1984), indicating that the epithelial cells of the maxilloturbinate
This document is a draft for review purposes only and does not constitute Agency policy.
4-198 DRAFT—DO NOT CITE OR QUOTE
-------
1 are more resistant to the genotoxic effects of DEN, support the possibility that differences in
2 regional tissue susceptibility may contribute to site specificity of formaldehyde-related SCCs.
3 Monticello et al. (1996) further explored the correlation between measures of cell
4 proliferation and tumor site by modifying the ULLI to take into consideration the total number of
5 cells in a region that may be subject to increased cell proliferation. The population weighted
6 ULLI (PWULLI) is the product of the expected number of cells on a three-dimensional surface
7 in the nasal mucosa and the ULLI of a cross section of that surface. For this series of
8 experiments, six male F344 rats/group were exposed to 0, 0.7, 2, 6, 10, or 15 ppm (0, 0.86, 2.45,
9 7.36, 12.3, or 18.4 mg/m3) formaldehyde for up to 24 months with interim sacrifices at 3, 6, 12,
10 and 18 months. Before each interim sacrifice [3H]-thymidine was continuously injected for the
11 last 5 days of exposure through a surgically implanted pump. Nasal passages were prepared, and
12 six standard sections were taken and developed as above for [3H]-thymidine-labeled cells.
13 Stained tissue sections were viewed in order to map all nasal tumors. A ULLI was determined
14 for each region (details not provided). The total cell population of each nasal region was
15 estimated from control animals sacrificed at 3 months (Table 4-38). Cell profiles were counted
16 across 0.5 mm of basement membrane length at two locations for each region (site not specified).
17 Total cells per region were estimated from these counts and the modeled surface area expected in
18 each region (Fluid Dynamics Analysis Package version 7.0). It is unclear if one or more rats
19 were used to quantify cell population. Cell counts and variability were not reported.
20
21 Table 4-38. Cell population and surface area estimates in untreated male
22 F344 rats and regional site location of squamous cell carcinomas in
23 formaldehyde-exposed rats for correlation to cell proliferation rates
24
Nasal region
Total cells
(number)3
Area
(mm2)b
Cell density
(cell/mm2)
SCC incidence0
10 ppm
15 ppm
Anterior lateral meatus
976,000
59.5
16,400
12
17
Anterior midseptum
184,000
10.5
17,500
0
1
Anterior dorsal septum
128,000
3.84
33,300
0
3
Anterior medial
maxilloturbinate
104,000
7.63
13,600
0
4
Posterior lateral meatus
508,000
38.1
13,300
2
9
Posterior midseptum
190,000
10.8
17,600
0
1
Maxillary sinus
884,000
38
23,300
0
0
Region not specified0
--
--
--
6
25
25
26 aTotal cell number determined in unexposed rats as a product of representative cell counts and expected surface area
27 of the region.
28 bModeled surface area of the defined region by FDIP version 7.0.
29 The number of animals bearing a tumor located in the region. Animals were exposed 6 hours/day for 24 months
30 prior to sacrifice.
31 Source: Monticello et al. (1996).
This document is a draft for review purposes only and does not constitute Agency policy.
4-199 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
ULLIs were quantified by region of the nasal passages in order to correlate with regional
localization of tumors. For example, the anterior midseptum included cells from the midseptum
from approximately standard section levels 2 to 3. An anterior to posterior pattern of
formaldehyde effects, especially differences in cell proliferation rates, has been well established.
As such, cell proliferation rates would be expected to vary across the nasal regions used in this
analysis. Areas considered to possibly be preneoplastic were not quantified for this work.
Monticello et al. (1996) reported increased ULLIs in the ALM and the MMT of rats exposed to
10 or 15 ppm at all time points (3, 6, 12, and 18 months) but provided no indication of variability
or a statistical analysis, making it difficult to determine where true differences may exist. Some
caution should be used in interpreting the ULLI counts assigned for each region.
The PWULLI was calculated by multiplying the reported ULLIs by the calculated cell
populations by region. SCC incidence by region had a greater correlation to the calculated
PWULLI than the ULLI, R2 = 0.88 versus R2 = 0.46, respectively. The authors noted that the
relative lack of correlation with the ULLI was influenced by findings at the maxilloturbinate
where cell proliferation was high but SCC incidence was low. Other tumor types were not
included in the analysis (polypoid adenomas, adenocarcinomas, and rhabdomyosarcomas).
Additionally, 54 of the SCC tumors could not be accurately localized and were excluded from
the analysis, resulting in exclusion of 30 and 39% of animals with SCCs in the 10 and 15 ppm
treatment groups, respectively. The authors cautioned that the absence of these data might have
skewed the regional analysis of tumor location. Although the purpose of weighting the ULLIs
by total population of cells available in each region is to better represent the chance of a tumor
arising in each region, the cancer incidence was represented by the number of animals, not the
number of tumors, per region. Based on the exclusion of location data (up to 40% of the
animals), lack of variability and significance reported for the ULLI for cell counts, and SCC
incidence considered by animal rather than by tumor, the significance of a greater correlation by
PWULLI versus ULLI is of questionable value.
Monticello et al. (1989) also assessed formaldehyde-induced cell proliferation and
regional site location of lesions in the respiratory tract of rhesus monkeys (see Section 4.2.1.2.2.2
for a full study description). Lis from the histoautoradiograms indicated increased cell
proliferation in transitory, respiratory, and olfactory epithelial cells after the 6-week
formaldehyde exposure. Similar trends were seen after only 1 week but were statistically
significant only in the respiratory epithelium. Although increased proliferation in the trachea and
carina was statistically significant after 1 week of exposure, the greater increases seen after
6 weeks of exposure were not statistically significant. A small sample size (n = 3) and high
variability may have contributed to the lack of statistical significance. The authors noted that
This document is a draft for review purposes only and does not constitute Agency policy.
4-200 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
increased cell proliferation was seen in locations with minimal histologic changes, indicating
proliferation may be a more sensitive predictor of adverse health effects of formaldehyde
exposure. Table 4-39 provides a summary of formaldehyde-induced cell proliferation data.
4.2.1.3. Gastrointestinal Tract and Systemic Toxicity
As with inhalation, the POE is thought to be the principal target tissue in response to oral
exposure. A concentration-dependent pattern of toxicity longitudinally down the GI tract has
been observed upon oral exposure. Some evidence (Til et al., 1989, 1988) suggests that, with
regard to oral exposure, duration in addition to concentration is important in the development of
toxicity.
Formalin and paraformaldehyde were used to dose animals in oral toxicity studies.
Formalin contains 12-15% methanol as a preservative to inhibit the polymerization of
formaldehyde and subsequent precipitation as paraformaldehyde (Kiernan, 2000). The presence
of methanol in formalin may confound the results of a formaldehyde study. Methanol has been
shown to be a developmental and neurologic toxin (e.g., Degitz et al. [2004a, b]; Rogers et al.
[2004, 2002]; Weiss et al. [1996]; Sharpe et al. [1982]). Oral dosing with paraformaldehyde is
preferred because it allows for the preparation of methanol-free formaldehyde in the laboratory
by dissolving paraformaldehyde in slightly basic water.
4.2.1.3.1. Short-term andsubchronic studies. Til et al. (1988) evaluated the oral toxicity of
formaldehyde and acetaldehyde in a subacute study in Wistar (Cpb:WU; Wistar random) rats.
Groups of rats (10/sex/dose) were exposed to paraformaldehyde dissolved in drinking water at 0,
5, 25, and 125 mg/kg-day for 4 weeks. The control group was comprised of 20 rats of each sex.
To account for potential effects of decreased water consumption in treated animals, an additional
control group of 10 male and 10 female rats was given drinking water in an amount equal to the
amount of liquid consumed by the group given the highest dose. Examination of the GI tract was
performed in all dose groups and included the tongue, esophagus, and stomach. Histopathology
for the other tissues was performed on high-dose and control animals.
This document is a draft for review purposes only and does not constitute Agency policy.
4-201 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-39. Summary of formaldehyde effects on cell proliferation in the upper respiratory tract
S"4
>3*
&
to
s
Species
Na
Treatmentb
Measure of cell
proliferation
Summary of results by location0
Reference
Male F344
rats;
male
B6C3F1
mice
NRd
0.5, 2, 6, or 15 ppm
6 hours/day for
3 days
LI: percent labeled cells
on tissue sections
(3H-thymidine I.P.d
2 hours postexposure)
Level 2: Rats exhibited greater increased cell proliferation than mice.
No increase seen in rats or mice at 0.5 or 2.0 ppm.
No increase seen in mice at 6 ppm, but rats had 20-fold increase in
proliferation.
10- to 20-fold increase seen in both rats and mice at 15 ppm.
Swenberg et
al. (1986)
Male F344
rats
NR
0.5, 2, or 6 ppm
6 hours/day
1, 3, or 9 days
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure)
Level 2: Transient increase in cell proliferation on day 1 at 0.5 and 2.0 ppm.
Increase in cell proliferation on days 1,3, and 9 by 6 ppm.
Swenberg et
al. (1986)
Male F344
rats;
male
B6C3F1
mice
NR
3 ppm for 12 hours,
6 ppm for 6 hours, or
12 ppm for 3 hours
3 or 9 days
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 1:
3 days: Greater increased proliferation in rats than mice. Increases
similar for various concentrations yielding the same C x t product.
9 days: Mice exhibited duration-dependent increases in proliferation,
inverse to concentration for constant C x t.
Level 2
3 days: Concentration-dependent increase in cell proliferation.
9 days: Concentration-dependent increase in cell proliferation in rats; no
increase in mice.
Swenberg et
al. (1986)
Male F344
rats
4-5
15 ppm
6 hours/day
1 or 5 days
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 2: Increase in cell proliferation in respiratory epithelium,
nasoturbinates, maxilloturbinates, and lateral wall.
1 day: 5.5 lf versus 0.43% in controls
5 days: 10.1%f
Chang et al.
(1983)®
Male
BC3F1
mice
4-5
15 ppm
6 hours/day
1 or 5 days
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 2: Increase in cell proliferation in respiratory epithelium,
nasoturbinates, maxilloturbinates, and lateral wall.
1 day: 2.14f versus 0.27% in controls
5 day: 3.42%f
Chang et al.
(1983)e
Male
albino
Wistar rats
5d
3.5 ppm
8 hours, twice a day
for 3 days
Qualitative staining for
PCNA on tissue sections
Levels 2 and 3: Increase in cell proliferation in respiratory epithelium,
nasoturbinates, maxilloturbinates, septum, and lateral wall.
Cassee and
Feron (1994)e
o
2 »
5 s
a, Co'
§- a
TO Sj-
§ ^
>S
>S
TO
TO'
*
to
o
to
-------
Table 4-39. Summary of formaldehyde effects on cell proliferation in the upper respiratory tract
S"4
>3*
s
Species
Na
Treatmentb
Measure of cell
proliferation
Summary of results by location0
Reference
Male
albino
Wistar rats
3
0, 5, or 10 ppm
8 hours/day
continuously for
3 days or 4 weeks, or
0, 10, or 20 ppm
8 hours/day
intermittent8
for 3 days or 4 weeks
LI: percent labeled cells
on tissue sections (3H-
thymidine I.P. 18 hours
postexposure for 2 hours)
Section level not stipulated in report.
3 days: 0.86% in controls 4 weeks: 0.68% in controls
2.83%f at 5 ppm continuous 1.33% at 5 ppm continuous
8.87%f at 10 ppm continuous 8.85%h at 10 ppm continuous
9.80%f at 10 ppm interrupted 3.41%f at 10 ppm interrupted
19.8%f at 20 ppm interrupted 13.9%f at 20 ppm interrupted
Wilmer et al.
(1987)e
Male
albino
Wistar rats
2
0, 1, 10, or 20 ppm
6 hours/day for
3 days
LI: percent labeled cells
(18 hour postexposure ex
vivo 3H-thymidine labeled
excised mucosa)
Level 3
Metaplastic epithelium: increased proliferation
31.4% at 10 ppm, 37.6% at 20 ppm
Visibly unaffected respiratory epithelium
1.6% in controls
2.6% at 10 ppm, 2.8% at 20 ppm
Woutersen et
al. (1987)e
Male
albino
Wistar rats
5
0, 1, or 2 ppm
8 hours/day
continuously for
3 days or 4 weeks,
or 0, 2, or 4 ppm
8 hours/day
intermittent8
for 3 days or 4 weeks
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 2
3 days: No change from controls 4 weeks: no change from controls
0.60% in controls 1.03% in controls
0.34%at 1 ppm continuous 0.81%at 1 ppm continuous
0.61 % at 2 ppm continuous 0.91 % at 2 ppm continuous
0.29% at 2 ppm interrupted 1.16% at 2 ppm interrupted
0.58% at 4 ppm interrupted 2.86% at 4 ppm interrupted
Wilmer et al.
(1989)e
Male and
female
albino
Wistar rats
5
0, 0.3, 1, 3 ppm
6 hours/day,
5 days/week
for 3 days or
13 weeks.
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 2: Increased cell proliferation at days 3 and 13 weeks (p < 0.001).
Level 3: Transient dose-dependent increase at 1 and 3 ppm;
only seen at day 3
(p< 0.001).
Note: Results pooled by sex. Data shown graphically on log-normal scale.
Zwart et al.
(1988)e
Male
Wistar rats
5
0, 0.3, 1, or 3 ppm
22 hours/day for
3 days
LI: percent labeled cells
on tissue sections (2 hours
postexposure ex vivo 3H-
thymidine-labeled excised
mucosa)
Level 2: 3 ppm increased cell proliferation in nasoturbinates,
maxilloturbinates, septum, and lateral wall (p < 0.05).
Level 3: No significant increases in cell proliferation.
Reuzel et al.
(1990)
o
S ?
a, Co
TO Sj-
§ ^
s ^
§ 3
*¦>1.
to
o
LtJ
-------
Table 4-39. Summary of formaldehyde effects on cell proliferation in the upper respiratory tract
Cl
S
Species
Na
Treatmentb
Measure of cell
proliferation
Summary of results by location0
Reference
o
5 ^ §
6, to'
to Sj-
§ ^
s ^
§ 3
?»»i.
Male
Sprague-
Dawley
rats
3-5
0, 2, 6, or 20 ppm
6 hours/day for
1 or 3 days
LI: percent labeled cells
by flow cytometry
(5-bromodeoxyuridine I.P.
18 hours postexposure for
2 hours)
Respiratory and olfactory epithelial cells.
1 day: 1.3 % in controls
2.4% at 2 ppmf
3.7% at 6 ppmf
2.7% at 20 ppmf
Tracheal epithelial cells
1 day: 1.2% in controls
3.1% at 2 ppmf
2.1% at 6 ppm
2.8% at 20 ppmf
Free lung cells (>97% MPs):
no significant change.
3 days:
1.4% at 2 ppm
2.5% at 6 ppmf
2.3% at 20 ppmf
3 days:
0.3% at 2 ppmf
0.6% at 6 ppmf
2.5% at 20 ppmf
Roemer et al.
(1993)
Male F344
rats
0.7, 2, 6, 10, or
15 ppm
6 hours/day,
5 days/week for 1, 4,
or 9 days or 6 weeks
ULLI (unit length LI) ( H-
thymidine LP. 18 hours
postexposure for 2 hours)
Level 3
No increases in cell proliferation at 0.7 or 2 ppm.
Level 4
ULLI increases in locations without lesions at 6 ppm.
Increases in ULLI at all locations at 10 and 15 ppm.
Monticello et
al. (1991)
Male
rhesus
monkeys
6 ppm
6 hours/day for
1 or 6 weeks
LI: percent labeled cells
on tissue sections (3H-
thymidine LP. 18 hours
postexposure for 2 hours)
Nasal passages: Duration-dependent increase in cell proliferation at all levels
(B-E) in transitional, respiratory, and olfactory epithelium.
Increased cell proliferation in areas with minimal lesions.
Larynx: trend for increased proliferation
Trachea: increased cell proliferation
1 week : 1.14 versus 0.55% in controls
6 weeks: 3.73%
Carina of trachea: increased cell proliferation.
1 week: 1.34 versus 0.43% in controls
6 weeks: 3.60%
Respiratory bronchioles: no increase in proliferation.
Monticello et
al. (1989)
aN = number of animals per treatment group.
bTreatment is given as the concentration of formaldehyde, duration of exposure each day, and length of the experiment in days and weeks.
°Standard section levels of the nasal passages as shown in Figure 4-3 are given for experiments in rats or mice.
dNR = not reported; LP. = intraperitoneally.
eStudy is described in full in section 4.2.1.2.2.4. .
different from control, p < 0.05.
to
o intermittent exposures were 30 minutes per hour for 8 hours.
hData from one animal only.
-------
1
2
3
4
5
6
7
8
9
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
The rats appeared to be healthy throughout the study, and no effects on growth occurred
despite significant decreases in food and water intake that occurred at the high dose (125 mg/kg-
day). Yellow discoloration of the fur occurred in the rats on the high dosage from week 3
onward. There were no significant changes in hematology among the exposed groups except for
slight (not statistically different) increases in PCVs in the water-restricted group and in high-dose
males. The high-dose groups of the formaldehyde exposed and in the water-restricted controls
had slightly increased urine density, but again this was not statistically significant. Plasma TP
and ALB levels were decreased in the males of the highest dose group. No changes in organ
weights occurred except for relative kidney weights that were slightly increased in the females of
the high-dose group. Gross pathological findings were restricted to the GI tract and revealed a
thickening of the limiting ridge of the forestomach in all animals exposed at the highest dose that
was accompanied by a yellowish discoloration of the mucosa. These latter changes were not
observed in the acetaldehyde-exposed animals. Treatment-related histopathologic changes were
seen in the GI tract only. Slight (8/20) or moderate (12/20) focal hyperkeratosis of the
forestomach and slight focal atrophic gastritis occurred in animals of the high-dose groups only
(Table 4-40). One female had moderate focal papillomatous hyperplasia. No histopathologic
changes were observed in any animals of the lower-dose groups. The study established a
LOAEL and NOAEL for epithelial changes in the GI tract of male and female Wistar rats
exposed to formaldehyde in drinking water at 125 mg/kg-day and 25 mg/kg-day, respectively.
Johannsen et al. (1986) performed a subchronic study by using rats and dogs exposed to
paraformaldehyde dissolved in drinking water. Groups of albino Sprague-Dawley rats (15/sex)
were administered the equivalent of 0, 50, 100, or 150 mg/kg-day in their drinking water for
91 consecutive days. Pure-bred beagle dogs (four/sex/group) were fed a diet with added aqueous
formaldehyde to approximate 0, 50, 75, or 100 mg/kg-day. Dogs were observed daily and rats at
frequent intervals for behavioral reactions. Body weights and food and water intake were
recorded on a weekly basis in both species. Hematology (HCT, Hb, total and differential
leukocyte counts), clinical chemistry (blood sugar, BUN, ALP, AST and ALT in dogs only), and
urine analyses (color, appearance, pH, specific gravity, sugar, protein, and microscopic elements)
were evaluated in 10 male and 10 female rats selected from each test group and in all dogs.
Organ weights were recorded for the adrenals, gonads, hearts, kidneys, livers, lungs, and thyroids
in each species. Histopathology was performed on a set of over 20 or 30 tissues and organs from
rats or dogs, respectively, in the high-dose and control groups only.
This document is a draft for review purposes only and does not constitute Agency policy.
4-205 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-40. Summary of lesions observed in the gastrointestinal tracts of
2 Wistar rats after drinking-water exposure to formaldehyde for 4 weeks
3
Type of lesion
Formaldehyde (mg/kg-day)
0
5
25
125
Number of male rats examined
20
10
10
10
Focal hyperkeratosis of forestomach
Very slight
3
0
0
0
Slight
1
0
0
4
Moderate
0
0
0
6
Focal gastritis
Slight
0
0
0
2
Moderate
0
0
0
1
Dilated fundic glands (single or a few)
0
0
0
0
Submucosal mononuclear cell infiltrate
0
0
0
1
Type of lesion
Number of female rats examined
20
10
10
10
Focal hyperkeratosis of forestomach
Very slight
6
0
0
2
Slight
0
0
0
2
Moderate
0
0
0
6
Focal gastritis
Very slight
0
0
0
1
Slight
0
0
0
1
Moderate
0
0
0
1
Focal papillomatous hyperplasia
0
0
0
1
Polymorphonuclear leukocytic infiltration
0
0
0
1
4
5 Source: Til et al. (1988).
6
7
8 No deaths or abnormal reactions were observed in either species. Significant reductions
9 in weight gain were observed in dogs of both sexes at 100 mg/kg-day, in rats of both sexes at
10 150 mg/kg-day, and in male rats at 100 mg/kg-day of formaldehyde. There was a dose-related
11 decrease in liquid consumption of both sexes in rats given formaldehyde, but there was no
12 overall difference in mean food intake or feed efficiency, so the reductions in body weight gain
13 were considered to be systemic effects. Dogs administered formaldehyde had reduced food
14 consumption and feed efficiency at all doses tested. No significant effects on hematology,
15 clinical chemistry, or urine analyses were observed in either species. No effects in either species
16 were reported on organ weights. The GI mucosa in both species was reported to appear normal
17 with no indication of irritation. This study suggests a NOAEL of 150 mg/kg-day in Sprague-
18 Dawley rats and of 100 mg/kg-day in beagle dogs for formaldehyde in drinking water.
19 Differences in the results for the rats with those reported in other studies (Til et al., 1989, 1988;
20 Tobe et al., 1989) may be due to strain differences or duration of the exposure. The dog may be
21 a more sensitive species than the rat based on these results and on those of 2-week pilot studies.
This document is a draft for review purposes only and does not constitute Agency policy.
4-206 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.3.2. Chronic bioassays: oral exposure to formaldehyde. The same laboratory that tested
formaldehyde and acetaldehyde in a 4-week study (Til et al., 1988) performed a chronic bioassay
with formaldehyde in drinking water. Til et al. (1989) administered paraformaldehyde dissolved
in drinking water to Wistar rats (Cpb: WU; Wistar random) (70/sex/dose). Interim sacrifices
(10/sex/dose) were performed at 12 and 18 months. Formaldehyde was administered in drinking
water to provide target doses of 0, 5, 25, and 125 mg/kg-day. The mean formaldehyde doses
administered were 0, 1.2, 15, or 82 mg/kg-day for males and 0, 1.8, 21, or 109 mg/kg-day for
females. Concentrations were adjusted weekly for the first 12 weeks based on dose estimates
derived from body weight and liquid consumption data. Such adjustments were made every
4 weeks from weeks 12 to 52 and kept constant. Fresh solutions of the test concentrations were
prepared weekly and stored at 15°C.
Endpoints examined included daily observations for condition and behavior, body weight
at weekly intervals for the first 12 weeks and then every 4 weeks thereafter, liquid intake weekly,
and food intake weekly for the first 12 weeks and then every 2 weeks for the remainder of the
study. Samples of blood were taken for hematological and clinical chemistry analyses on weeks
26 and 103. Analysis of blood glucose and urine pH, density, and volume was performed on
samples at weeks 27, 52, 78, and 104. Pooled urine samples were also evaluated for glucose,
occult blood, ketones, urobilinogen, and bilirubin in samples at weeks 27 and 104. Weights of
all major organs were recorded at interim sacrifices and at term. Gross and histopathologic
examinations were carried out on all major tissues of the rats in the high-dose and control
groups. The livers, lungs, stomach, and noses were examined in all rats. Additionally, the
adrenals, kidneys, spleens, testes, thyroids, ovaries, pituitaries, and mammary glands (for
females) were examined in all sacrificed animals at weeks 53 and 79 and at term.
The general health and behavior of the rats were not affected in any of the formaldehyde-
exposed groups. Slight yellowing of the fur did occur in the animals exposed at the mid and high
doses from week 3 onward. The mean body weights were decreased in the males from week 1
and in the females from week 24 onward. At the high dose, liquid consumption was significantly
decreased in both sexes, and food intake was significantly decreased in the males. There were no
toxicologically significant effects on hematological, urinary, or clinical chemistry parameters.
Decreases in absolute heart, liver, and testis (males) weights were attributed to lower body
weights. Relative kidney weights were increased in females of the high-dose group, and relative
brain weights were increased in both sexes of the high-dose group. Relative testis weight was
increased in males. Treatment-related changes in gross pathology were restricted to the
forestomach. Histopathologic examinations at the two interim sacrifices and final sacrifice
This document is a draft for review purposes only and does not constitute Agency policy.
4-207 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
revealed GI tract changes. Renal changes were observed in the high-dose group at final
sacrifice. There was no indication of treatment-related effects in other tissues.
As shown in Table 4-41, significant histopathology in the GI tract was limited to the
forestomach and stomach of rats in the high-dose groups. Some progression with duration of
exposure may have occurred by week 105 because GI lesions were observed in the lower dose
groups at this time point, whereas none were observed in these groups at interim sacrifices. The
histopathologic changes included papillary epithelial hyperplasia in the forestomach that was
frequently accompanied by hyperkeratosis on the limiting ridge or its vicinity. The mucosa
showed an irregular layer of hyperplastic basal cells, but no atypical nuclei or other subcellular
structures were observed. Chronic atrophic gastritis occurred to varying degrees in the stomachs
of all high-dose rats. In some cases the inflammatory process involved the entire mucosa and
was seen to extend to the whole muscularis mucosae and met the criteria for ulceration.
Histologic examination also showed that the incidence and degree of renal papillary
necrosis was increased in animals of the high-dose groups at the terminal sacrifice. This change
was located at the tip of the papilla and was characterized by patchy necrosis of interstitial cells,
capillaries, and loops of Henle. There was no evidence of a dose-related response in chronic
nephropathy. The incidence of chronic nephropathy was lower in the males of the high-dose
group than in controls. In females, the incidence was slightly higher in the test groups than in
controls but only achieved statistical significance at the lowest dose. It is likely that the decrease
in liquid intake incurred in the high-dose groups contributed to the increased incidence and
degree of renal papillary necrosis observed in the high-dose animals because dehydration has
been shown to enhance its production by various analgesics.
The results of this chronic bioassay indicated that formaldehyde is cytotoxic to the
epithelial mucosa of the nonglandular (forestomach) and glandular stomach with a LOAEL of 82
and 109 mg/kg-day and a NOAEL of 15 and 21 mg/kg-day in males and females, respectively.
The findings provided no evidence of carcinogenicity in either the GI tract or systemic sites for
formaldehyde administered in drinking water to Wistar rats at doses as high as 82 mg/kg-day.
This document is a draft for review purposes only and does not constitute Agency policy.
4-208 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 4-41. Incidence of lesions observed in the gastrointestinal tracts of
Wistar rats after drinking-water exposure to formaldehyde for 2 years
Incidence of lesions with formaldehyde dose
(mg/kg-day)a
Males
Females
0 1.2 15 82
0 1.8 21 109
Week 53
Number of rats examinedb
9
10
10
10
10
10
10
9
Forestomach
Focal papillary epithelial hyperplasia
0
0
0
7
0
0
0
5
Glandular stomach
Chronic atrophic gastritis
0
0
0
10c
0
0
0
9°
Focal ulceration
0
0
0
3
0
0
0
1
Focal mononuclear cell infiltrate
1
0
3
0
2
0
0
0
Atypical glandular hyperplasia
0
0
0
0
0
0
0
1
Week 79
Number of rats examined
10
10
10
10
10
9
10
10
Forestomach
Focal papillary epithelial hyperplasia
2
1
1
8
1
0
1
9
Glandular stomach
Chronic atrophic gastritis
0
0
0
10c
0
0
0
10c
Focal ulceration
0
0
0
2
0
0
0
0
Focal squamous metaplasia
0
0
0
1
0
0
0
0
Submucosal inflammatory cell
infiltrate
1
2
0
0
0
0
0
0
Focal mononuclear cell infiltrate
0
1
0
0
1
1
0
0
Glandular dilation
2
4
4
1
2
2
4
0
Week 105
Number of rats examinedb
47
45
44
47
48
49
47
48
Forestomach
Focal papillary epithelial hyperplasia
1
2
1
45°
1
0
2
45°
Focal hyperkeratosis
2
6
4
24°
3
5
3
33°
Focal ulceration
1
1
1
8
0
0
2
5
Focal acanthosis
1
0
2
1
0
0
0
1
Focal basic cell hyperplasia
0
1
1
0
1
0
0
0
Diverticulum
0
0
1
0
0
0
0
0
Exophytic papilloma
0
1
0
0
1
0
0
0
Glandular stomach
Chronic atrophic gastritis
0
0
0
Os
o
0
0
0
o
00
"3-
Focal ulceration
0
0
0
llc
0
0
0
10c
Glandular hyperplasia
0
1
0
20°
0
0
0
13°
Mineralization
3
2
1
0
0
0
0
0
Focal inflammatory cell infiltrate
5
3
2
0
2
3
1
0
incidence in rats that died or were killed when moribund during the experiment or were killed at
week 53, 79, or 105.
bA few rats were lost because of advanced autolysis.
°The values differ significantly (Fisher's exact test) from the control value (p < 0.001).
Source: Til et al. (1989).
This document is a draft for review purposes only and does not constitute Agency policy.
4-209 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Tobe et al. (1989) performed a chronic toxicity study of Wistar rats (Slc:Wistar) exposed
to paraformaldehyde dissolved in drinking water. Groups of 20 male and 20 female rats were
given formaldehyde solution in their drinking water at concentrations of 0, 0.02, 0.10, and 0.50%
for 24 months. Interim sacrifices of six randomly chosen rats from each group were performed
after 12 and 18 months. Based on the estimated average amount of water intake and body
weight, the actual doses of formaldehyde in either sex were reported to be 0, 10, 50, and
300 mg/kg-day. Fresh test solutions were prepared twice each week. The rats were observed
daily for the entire study. Body weights and water and diet intake were measured once weekly
or biweekly. Hematology (RBC, WBC, and Hb) and serum clinical chemistry (TP, ALB, BUN,
uric acid, total cholesterol, inorganic phosphorous, ALP, AST, and ALT) were made at each
necropsy. Organ weights were measured for the brain, heart, lung, liver, kidney, spleen, adrenal,
testis or ovary, pituitary, and thyroid. These organs and the stomach, small and large intestine,
pancreas, uterus, lymph nodes, and all tumors were examined histopathologically.
The general condition of animals in the high-dose group was poor with significantly
reduced body weight gain as well as intake of water and diet. An increase in mortality was also
observed in this group. Some clinical chemistry parameters were altered in this group. No
significant changes in absolute or relative organ weights were observed. Mortality was 100% in
the high-dose group by 24 months. At the 12-month sacrifice, hyperplasia of the squamous
epithelium with or without hyperkeratosis was observed in the forestomach of all high-dose
animals (12/12). Basal cell hyperplasia with growth into the submucosa was also observed in
most cases (10/12). Erosions and/or ulcers with submucosal inflammatory cell infiltrates were
observed in the glandular stomach of most rats (10/12). Regenerative changes of the glandular
epithelium (glandular hyperplasia) were noticed in most cases (10/12) along the limiting ridge of
the fundic mucosa. No lesions were observed in the glandular stomach at the 50 mg/kg-day
dose, and forestomach hyperplasia was observed in only one of six males and in one of eight
females at 18 and 24 months. No lesions in either the forestomach or glandular stomach were
observed in rats treated at 10 mg/kg-day.
This study corroborates the Til et al. (1989) study and shows that the main targets for
formaldehyde toxicity administered by drinking water to rats are the forestomach and glandular
stomach. Although the lesions observed at the 50 mg/kg-day were minimal in this study, Tobe et
al. (1989) designated the NOAEL at 10 mg/kg-day, further supporting the NOAEL of 15 mg/kg-
day from the Til et al. (1989) study.
Takahashi et al. (1986) studied the effects of formaldehyde in an initiation-promotion
model of stomach carcinogenesis in male outbred Wistar rats (Shizuoka Laboratory Center,
Shizuoka). Rats (n = 17) were given 100 mg/L of N-methyl-N'-nitro-N-nitrosoguanidine
This document is a draft for review purposes only and does not constitute Agency policy.
4-210 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
(MNNG) in drinking water and a diet supplemented with 10% sodium chloride (NaCl) for the
first 8 weeks as an initiation phase. This was followed by 0.5% formalin (which contains 12-
15%) methanol) in drinking water for 32 weeks as the promotion phase of the protocol. A
comparison group (n = 10) was given stock water and diet without any supplementation for the
first 8 weeks followed by 0.5% formalin in drinking water for 32 weeks. Animals were observed
daily and weighed once every 4 weeks. Small pieces of the stomach and other tissues in the
peritoneal cavity were fixed for histopathologic examination.
Body weight gain was reduced by exposure to MNNG with sodium chloride, and
formaldehyde exposure during the promotion phase exacerbated this effect. Histopathologic
investigations were restricted to the GI tract. Formaldehyde was shown to statistically increase
the incidence of lesions in the forestomach and stomach in the animals initiated with MNNG
with NaCl as compared with controls receiving no initiation (Table 4-42). Increases in
papilloma in the forestomach, adenomatous hyperplasia in the fundus, and adenocarcinoma in
the pylorus were observed. Histopathology in the animals receiving formaldehyde alone during
weeks 9 through 32 showed an increase in forestomach papillomas but with no lesions in the
glandular stomach (Table 4-42). The adenomatous hyperplasia were defined as proliferative,
noninvasive mucosal lesions, and the adenocarcinomas were defined as well differentiated and
composed of typical glandular structures, demonstrating a tubular pattern and cellular or
structural atypism without metastasis. No definition of criteria for papilloma diagnosis was
provided. The findings in this study are inconsistent with those of Til et al. (1989), who found
no evidence of carcinogenicity in a 2-year bioassay at comparable concentrations (assuming 37%
formaldehyde in formalin results in 0.19% formaldehyde in this study). As discussed above, the
differences may be due to differences in the strains of rat or in the diagnostic criteria. The lack
of more than one test concentration precludes dose-response analysis of this study and provides
only a stand-alone LOAEL of 0.2% formaldehyde in drinking water. The lack of consumption
data precludes an estimation of dose in mg/kg-day.
This document is a draft for review purposes only and does not constitute Agency policy.
4-211 DRAFT—DO NOT CITE OR QUOTE
-------
is*
Table 4-42. Effect of formaldehyde on gastroduodenal carcinogenesis initiated by MNNG and NaCl in male
Wistar rats exposed to formaldehyde (0.5% formalin) in drinking water for 8 weeks
No MNNG initiation prior to 8-week oral exposure to formaldehyde (0.5% formalin in drinking water)
Gastroduodenal
carcinoma
Forestomach
papillomas
Glandular stomach tumors
Fundus
Pylorus
Duodenum
Adenocarcinoma
Adenomatous
hyperplasia
Adenocarcinoma
Preneoplastic
hyperplasia
Adenocarcinoma
Control
0%
0%
0%
0%
0%
0%
0%
Formaldehyde
0%
80%a
0%
0%
0%
0%
0%
MNNG initiation (100 mg/L in drinking water for 8 weeks) prior to 8-week oral exposure to formaldehyde (0.5% formalin in drinking water)
Gastroduodenal
carcinoma
Forestomach
papillomas
Glandular stomach tumors
Fundus
Pylorus
Duodenum
Adenocarcinoma
Adenomatous
hyperplasia
Adenocarcinoma
Preneoplastic
hyperplasia
Adenocarcinoma
Control
13.3%
0%
0%
0%
3.3%
23.3%
10.0%
Formaldehyde
29.4%
88.2%a
0%
88.2%a
23.5%b
41.2%
5.9%
"Significantly different from control animals with MNNG initiation, p < 0.01,
bSignificantly different from control animals with MNNG initiation, p < 0.05.
Source: Takahashi et al. (1986).
to
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
Soffritti et al. (1989) administered formaldehyde in drinking water to Sprague-Dawley
rats at different ages. Formaldehyde was produced via the Formox process, which yields carbon
monoxide, dimethyl ether, and small quantities of C02 and formic acid. In one experiment
denoted as BT 7001, rats (50/sex/dose) that were 7 weeks old were administered 0, 10, 50, 100,
500, 1,000, or 1,500 mg/L for 104 weeks. As is usual for experiments carried out at the
Ramazzini Foundation, all animals were maintained until natural death at which point they were
necropsied and examined histopathologically. In experiment BT 7005, 25-week-old breeder rats
or offspring (beginning postnatal day [PND] 12) were provided drinking water with 0 or 2,500
mg/L formaldehyde for 104 weeks. Fluid and food consumption were measured weekly for the
first 13 weeks and then every 2 weeks thereafter. Individual body weights were recorded for the
first 13 weeks and then every 2 weeks thereafter. Histopathology was routinely performed on all
major tissues, including oral and nasal cavities, the GI tract (esophagus, stomach, and intestines
[4 levels]), primary and secondary lymph organs (thymus, spleen, subcutaneous lymph nodes,
mesenteric lymph nodes, mediastinal lymph nodes, and femur [bone marrow]), head and face
bones, and other organ systems (liver, kidney, bladder, reproductive, and various glands).
Noncancer health effects were not reported.
No GI neoplasia were observed in any of the control rats (experiments BT 7001 and BT
7005). Historical controls for the BT experimental colony (n = 5,259) indicate an incidence of
approximately 1% for benign neoplasia (papillomas and acanthomas) and an incidence of 0.19%
for malignant stomach neoplasia (adenomas, SCCs, and adenocarcinomas, fibrosarcomas, and
leiomyosarcomas taken together). Therefore, the size of the control groups (18-50 rats/sex)
makes detection of background neoplasia unlikely. Similarly, one or two tumors noted in a
treatment group (n = 18-50) would represent an apparent increase in these relatively rare tumors.
Although stomach and intestinal tumors were found in rats exposed to formaldehyde in drinking
water, the low incidence makes it difficult to discern any dose-response effect for individual
neoplasia. The authors report formaldehyde-induced GI tract neoplasia to include benign tumors
(papillomas and acanthomas of the forestomach and adenomas) and malignant tumors
(adenocarcinomas and leiomyosarcomas). The majority of malignant tumors were present in the
duodenum, jejunum, and ileum. Benign and malignant neoplasia were consistently noted in the
two highest exposure groups: 1500 mg/mL formaldehyde in experiment BT 7001 and
2,500 mg/mL in experiment BT 7005 (Table 4-43). Comparison of overall GI neoplasia in
breeder and offspring rats of experiment BT 7005 suggests that rats exposed beginning on
PND 12 had a greater incidence of malignant tumors. However, it should be noted that, although
there may be an apparent increase in overall tumors, summing across sites and locations is
needed before a response can be seen. Even then, several data points are based on a single
This document is a draft for review purposes only and does not constitute Agency policy.
4-213 DRAFT—DO NOT CITE OR QUOTE
-------
1 observed tumor. However, nonspecific MO As, such as mutagenicity and regenerative
2 proliferation, would be expected to act on all cell types at the POE, and summing tumor types
3 may have some utility. These results constitute a weak positive result for cancer due to oral
4 exposure to formaldehyde.
5
6 Table 4-43. Summary of benign and malignant gastrointestinal tract
7 neoplasia reported in male and female Sprague-Dawley rats exposed to
8 formaldehyde in drinking water at different ages
9
Total benign
Total malignant
Experiment
Dose
Sex
tumors
tumors
All tumors
Historical
Not
M
1.08%
0.31%
1.4%
controls
applicable
F
0.97%
0.41%
1.4%
BT 7001
Control
M
a
-
-
(7 weeks old)
F
-
-
-
1,000 mg/mL
M
-
2%b
2%b
F
2%b
-
2%b
1,500 mg/mL
M
4%
6%
10%
F
6%
-
6%
BT 7005
Control
M
-
-
-
breeders
F
—
—
—
(25 weeks
2,500 mg/mL
M
_
5.6%b
5.6%b
old)
F
5.6%b
-
5.6%b
BT 7005
Control
M
-
-
-
offspring
F
-
-
-
(PND 12)
2,500 mg/mL
M
5.6%
8.3%
13.9%
F
-
21.6%
21.6%
10
11 aDash indicates no tumors reported. An incidence was not reported.
12 Percentage is based on the observation of a single neoplasm.
13
14 Source: Soffritti etal. (1989).
15
16 Oral exposure to formaldehyde resulted in a dose-dependent increase in all
17 hemolymphoreticular neoplasia in both male and female rats in experiment BT 7001 (Table
18 4-44) (Soffritti et al., 1989). The most frequent neoplasia noted were lymphoblastic leukemias
19 and lymphomas. The authors combined lymphoblastic leukemias and lymphosarcomas for
20 analysis and summing across sites. This analysis is appropriate since there is broad consensus
21 that "neoplasms presenting as solid tumors and those presenting with blood and marrow
22 involvement are biologically the same disease but with different clinical presentations," as stated
23 in the recent WHO reclassification of hematological malignancies (Harris et al., 2000b).
24 Inspection of the data tables suggests that the only treatment-related effects occurred at 1,000
This document is a draft for review purposes only and does not constitute Agency policy.
4-214 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
and 1,500 mg/L. The reported incidence of lymphoid neoplasia and of all hemolymphoreticular
neoplasia in rats exposed to the vehicle (methyl alcohol, 15 mg/L) was similar to results reported
in rats exposed to 50 or 100 mg/L formaldehyde in drinking water. Soffritti et al. (1989)
provided no statistical analysis of the data. Although the authors report a similar increase in
experiment BT 7005, the apparent 5% increase is representative of a single animal in a treatment
group of 18 and may not represent a true increase, the study's usefulness for
hemolymphoreticular neoplasia being somewhat limited by study size.
Table 4-44. Incidence of hemolymphoreticular neoplasia reported in male
and female Sprague-Dawley rats exposed to formaldehyde in drinking
water from 7 weeks old for up to 2 years (experiment BT 7001)
Lymphoid
Other leukemias
All leukemias and
Treatment
Sex
neoplasia (%)
(%)
lymphomas (%)
M
4
a
4
Control
1
F
2
3
Vehicle
M
10
-
10
control
F
2
4
6
M
2
2
10 mg/mL
F
4
—
4
50 mg/mL
M
10
8
_
10
8
F
-
M
8
2
10
8
100 mg/mL
F
4
4
500 mg/mL
M
12
4
16
8
F
4
4
M
12
12
1,000 mg/mL
F
10
4
14
M
22
22
1,500 mg/mL
F
10
4
14
aDash indicates no neoplasm was reported.
Source: Adapted from Soffritti et al. (1989).
The study of Soffritti et al. (1989) does provide qualitative corroboration of evidence
from other studies that observed formaldehyde toxicity in the forestomach and stomach.
However, the dosages required to induce such lesions in this study were higher than in other
studies. Soffritti et al. (1989) is the only animal study of oral exposure to formaldehyde that
reports an increase in lymphoblastic leukemia or lymphosarcoma. These occurred at the two
highest doses of 1,000 and 1,500 mg/L used in the study of animals exposed from 7 weeks of age
(experiment BT 7001).
This document is a draft for review purposes only and does not constitute Agency policy.
4-215 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.3.3. Summary of toxicity in the GI tract. Short-term and subchronic exposures to
formaldehyde via drinking water for 4 weeks yielded slight to moderate histopathologic lesions
(focal hyperkeratosis) at 125 mg/kg-day in male and female Wistar rats, as well as slight focal
gastritis and submucosal infiltrate in one to two animals of both sexes (Til et al., 1988). No
histopathologic lesions were noted in albino Sprague-Dawley rats or beagle dogs that received
oral doses of formaldehyde in drinking water for 91 days (Johannsen et al., 1986). In both
studies, decreases in weight gain were noted in exposed animals compared with controls.
In a chronic drinking water study, Til et al. (1989) reported that formaldehyde is
cytotoxic to the epithelial mucosa of the nonglandular (forestomach) and glandular stomach with
a LOAEL of 82 and 109 mg/kg-day and a NOAEL of 15 and 21 mg/kg-day in males and female
Wistar rats, respectively. The findings provided no evidence of carcinogenicity in either the GI
tract or systemic sites for formaldehyde administered in drinking water to Wistar rats at doses as
high as 82 mg/kg-day. The incidence and degree of renal papillary necrosis was increased in
animals of the high-dose groups at the terminal sacrifice (Til et al., 1989). Findings by Tobe et
al. (1989) corroborate the Til et al. (1989) study and show that the main targets for formaldehyde
toxicity administered by drinking water to rats are the forestomach and glandular stomach.
Takahashi et al. (1986) studied the effects of formaldehyde in an initiation-promotion model of
stomach carcinogenesis in male outbred Wistar rats (Shizuoka Laboratory Center, Shizuoka,
Japan). In contrast to Til et al. (1989), Takahashi et al. (1986) found increases in incidence of
papilloma in the forestomach, adenomatous hyperplasia in the fundus, and adenocarcinoma in
the pylorus in a 2-year bioassay at comparable concentrations (assuming 37% formaldehyde in
formalin results in 0.19% formaldehyde in this study). Soffritti et al. (1989) administered
formaldehyde in drinking water to Sprague-Dawley rats at different ages. Rats (50/sex/dose, age
7 weeks) were administered 0, 10, 50, 100, 500, 1,000, or 1,500 mg/L formaldehyde in drinking
water for 104 weeks. The authors reported formaldehyde-induced GI tract neoplasia that
included benign tumors (papillomas and acanthomas of the forestomach and adenomas) and
malignant tumors (adenocarcinomas and leiomyosarcomas), albeit at a relatively low incidence
after summing across sites and locations. Interestingly, oral exposure to formaldehyde resulted
in a dose-dependent increase in all hemolymphoreticular neoplasia in both male and female rats
(Soffritti et al., 1989). The most frequent neoplasia noted were lymphoblastic leukemias and
lymphomas.
4.2.1.4. Immune Function
Leach et al. (1983) documented potential immunomodulatory effects of formaldehyde
inhalation exposure. F344 rats were exposed nose only to formaldehyde 6 hours/day,
This document is a draft for review purposes only and does not constitute Agency policy.
4-216 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
5 days/week for up to 30 days. The target concentrations for exposure were 0, 3, 16, 61, and
99 ppm formaldehyde (0, 3.7, 19.7, 75.0, and 122 mg/m3). Body weight and food consumption
were recorded, and blood samples for standard hematology and immune assays were collected
(details not given). Immune measures referenced include in vitro lymphocyte transformation,
hemagglutination assays, and enumeration of B cells, WBCs, and RBCs. No effects were seen at
3 ppm formaldehyde. However, dose-dependent responses were reported for weight loss,
decreased food consumption, increased WBCs, increased segmented neutrophils and nucleated
RBCs, and decreased ability to produce antibodies to sheep RBCs. The results of the
lymphocyte transformation assay were inconsistent, with a 25-30% reduction in stimulation after
exposure to 99 ppm but an initial stimulation seen after 16 and 61 ppm exposures. Further
details were not available, making it difficult to determine if these reported immunomodulatory
effects may have been, in part or in full, secondary to effects on the URT. Subchronic exposures
at 61 and 99 ppm formaldehyde would be expected to result in frank toxic effects in mice (see
Section 4.2.1). However, these findings suggest possible immunomodulatory effects due to
formaldehyde exposure and require further exploration.
Dean et al. (1984) investigated the effects of formaldehyde exposure on a range of
indicators of immune function. Female B6C3F1 mice were exposed to 15 ppm formaldehyde
(18.4 mg/m3) 6 hours/day, 5 days/week for 3 weeks. Three trials were run with a total of 255
formaldehyde-treated mice. Body and organ weights were recorded at sacrifice for control and
formaldehyde-exposed mice (10 per group). Measures of host susceptibility, cell-mediated
immunity MP function, and antibody reactions were conducted 2 to 6 days after the end of
exposure (Table 4-45). Lymphocyte subsets, spleen cellularity, bone marrow cellularity, and
progenitor cell subsets were enumerated. Host susceptibility and delayed type hypersensitivity
were measured in vivo. Lymphocyte proliferation, natural killer cell activity, phagocytosis,
hydrogen peroxide production, and IgM plaque-forming cells (PFCs) were measured ex vivo
after in vivo stimulation in some cases (Table 4-45).
Body weight, organ weights and cellularity, progenitor cell populations, blood cell
counts, and differentials were unchanged in formaldehyde-treated mice (Dean et al., 1984).
Circulating blood monocytes were decreased in treated mice, which may be a reflection of the
local inflammatory response expected in the nasal epithelium (Dean et al., 1984). However,
there was no corresponding decrease in peritoneal MPs. There was a trend, but no statistical
significance, for decreased spleen weight, cellularity, and B cell precursors (87, 83, and 78% of
controls, respectively). The mean body weight of formaldehyde-treated mice was 21.1 versus
20.9 g in control mice, and thymus and spleen weights were not normalized by body weight.
This document is a draft for review purposes only and does not constitute Agency policy.
4-217 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-45. Battery of immune parameters and functional tests assessed in
2 female B6C3F1 mice after a 3 week, 15-ppm formaldehyde exposure
3 (6 hours/day, 5 days/week)
4
Immune function
Model
Challenge
Metric
Host susceptibility
Tumor resistance
PYB6 sarcoma cells
Subcutaneous injection, followed
by skin palpation to track tumor
development
Tumor resistance
16F10 melanoma cells
Lung tumor burden determined by
|2I | UdR incorporation
Bacterial resistance
Listeria monocytogenes
Survival after challenge
Cell-mediated
immunity
Delayed type
hypersensitivity
Keyhole limpet
hemocyanin
Radiometric index of delayed
hypersensitivity responses
Lymphocyte
proliferation
T-cell mitogen, PHA'1
B-cell mitogen, LPSb
(ex vivo)
Ex vivo proliferation, 3 days,
measured by [3H]-thymidine
incorporation
Lymphocyte subsets
None
Percentage of cells positive for cell
surface markers (Thy-1, Mac-1,
Lyt-1)
Natural killer cell
activity
Yac-1 target cells
(51Cr labeled)
(ex vivo)
% cytotoxicity by 51 Cr release
MP function
(both resident and
MVE-1 elicited
MP)
Phagocytosis
Sheep RBCs
(51Cr labeled)
(ex vivo)
51 Cr incorporation as a measure of
RBCs phagocytized
Hydrogen peroxide
production
Pharmacologic
stimulation (ex vivo)
H202 release in culture
Humoral cell
immunity
Antibody PFC
responses, IgM
PFCs
Sheep RBCs,
TVF-LPS, or TNF-
Ficoll
Plaques formed
Progenitor cells
Bone marrow
cellularity (femur)
None
Cell enumeration by a Coulter
counter
Granulocyte-MP
progenitors
None
Cell enumeration by a Coulter
counter
B-cell precursors
None
Clonogenic assay
5
6 aT-cell mitogen, phy to hemagglutinin (PHA-P).
7 ''B-ccll mitogen, lipopolysaccharide (Escherichia coli).
8
9 Source: Dean etal. (1984).
10
11
12 All indicators of natural killer cell function, cell-mediated immunity, and humoral
13 immunity in formaldehyde-treated mice were unchanged from controls (Dean et al., 1984).
14 Phagocytic capacity of both resident and elicited peritoneal MPs was unchanged by
15 formaldehyde treatment. However, hydrogen peroxide production in elicited peritoneal MPs was
16 significantly increased in formaldehyde-treated mice, 78 versus 42 nmol/mg protein (p < 0.05)
17 (Dean et al., 1984).
18 As shown in Table 4-46, several indicators of host resistance in the female B6C3F1 mice
19 were increased after formaldehyde exposure (Dean et al., 1984). Tumor mass and pulmonary
This document is a draft for review purposes only and does not constitute Agency policy.
4-218 DRAFT—DO NOT CITE OR QUOTE
-------
1 foci after B16F10 melanoma cell challenge were significantly reduced in formaldehyde-treated
2 mice, indicating improved tumor immunity (p < 0.05). However, following PYB6 sarcoma cell
3 challenge, formaldehyde-treated mice had a 7.1% tumor incidence versus 11.1% in controls,
4 which was not statistically different. Mortality due to Listeria monocytogenes (LM) was
5 decreased from 70 to 30% (p < 0.05). Because resistance to LM is primarily MP dependent, the
6 authors speculated that this enhanced resistance might be due in part to increased bactericidal
7 activity as was also suggested by increased hydrogen peroxide production ex vivo in elicited
8 peritoneal MPs from female mice (Dean et al., 1984).
9
10 Table 4-46. Summary of the effects of formaldehyde inhalation on the
11 mononuclear phagocyte system (MPS) in female B6C3F1 mice after a
12 3-week, 15 ppm formaldehyde exposure (6 hours/day, 5 days/week)
13
In vivo indicators of MPS
Metric
Formaldehyde effect
Cellularity
Circulating monocytes
Decreased3
CMF progenitor cells
No change3
Resident peritoneal MP
No change3'13
Elicited peritoneal MP
No change3'15
In vivo test of host resistance
LM
Increased resistance3
B16F10 tumor challenge
Increased resistance3
PYB6 tumor challenge
No significant increase3
Ex vivo indicators of MPS
Cell type/activation
Formaldehyde effect
H202 production
Resident, no PMA°
None detected 3'b
Resident, with PMA
None detected 3'b
Elicited, no PMAd
None detected 3'b
Elicited, with PMA
Increased 3'b
Phagocytosis
Resident
No change3
Elicited
No change3
Assessment of MP
Resident
Decreased13
maturation
Elicited
No changeb
Leucine aminopeptidase content
Resident
No changeb
Elicited
No changeb
Acid phosphatase content
Resident
No changeb
Binding of tumor cells
Elicited
No changeb
Resident
No changeb
Lysing of tumor cells
Elicited
Increased at mid-range target-to-effector
cell ratiob
"Deanetal. (1984).
bAdams et al. (1987).
°Phorbol 12-myristate 13-acetate (PMA).
dPeritoneal MPs were elicited with the pyran copolymer Murray Valley encephalitis virus (MVE-2).
Sources: Adams et al. (1987); Dean et al. (1984).
This document is a draft for review purposes only and does not constitute Agency policy.
4-219 DRAFT—DO NOT CITE OR QUOTE
14
15
16
17
18
19
20
21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Overall, the observations of increased hydrogen peroxide production and increased host
resistance in peritoneal MPs distant from the POE suggest that formaldehyde has an effect on the
mononuclear phagocyte system (MPS). The authors postulated that this effect may be indirect,
due in part to the tissue inflammatory response in the URT or a direct systemic effect on the
MPS by formaldehyde exposure (Dean et al., 1984). Subsequent studies by the same researchers
explored the possibility of systemic effects of formaldehyde exposure on MPS function and
maturation stage (Adams et al., 1987). Female B6C3F1 mice were exposed to 15 ppm
(18.4 mg/m3) formaldehyde 6 hours/day, 5 days/week for 3 weeks, as before (Adams et al.,
1987). Both resident and Murray Valley encephalitis virus (MVE-2)-elicited peritoneal MPs
were examined for hydrogen peroxide production, enzymatic activity, phagocytic ability,
binding, and lysis of tumor cells (Adams et al., 1987).
Similar to the findings of Dean et al. (1984), formaldehyde treatment increased hydrogen
peroxide production almost twofold in MVE-2 elicited peritoneal MPs (Adams et al., 1987). As
summarized in Table 4-46, no treatment differences were seen in phagocytic ability in either
resident or elicited MPs (Adams et al., 1987). Resident peritoneal MPs from formaldehyde-
treated mice were not different in their ability to bind or lyse tumor cells. Although
formaldehyde treatment did not increase the ability of elicited MPs to bind tumor cells, lysis of
the target cells (P815 tumor cells) was increased from 28 to 37% by formaldehyde treatment but
only at the midrange target-to-effector-cell ratio tested in the assay (p < 0.05) (Adams et al.,
1987). Although this is statistically significant, the authors questioned the biological
significance of this result since it was not observed at all three target cell ratios tested. However,
an increase in tumor cell lysis in vitro would be consistent with the in vivo increased tumor
resistance previously reported (Dean et al., 1984). The in vitro lysis response curve suggests that
assay conditions may result in a maximum cytolysis near 40%. If so, any treatment effects on
lysis would be difficult to discern at higher effector cell ratios.
Jakab (1992) investigated the effect of formaldehyde exposure on the alveolar MPs and
resistance to respiratory infections. The first set of experiments assessed bactericidal activity by
directly quantifying the pulmonary bacterial loading after exposure to Staphylococcus aureus.
White female Swiss mice were exposed to formaldehyde after bacterial infection (regimens A
and C), before bacterial infection (regimen B), or before and after infection (regimen D)
(Table 4-47).
This document is a draft for review purposes only and does not constitute Agency policy.
4-220 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-47. Formaldehyde exposure regimens for determining the effects of
2 formaldehyde exposure on pulmonary S. aureus infection
3
Pre-infection treatment
Postinfection treatment
Results
Regimen A
None
4 hours
0, 1, 5, 10, or 15 ppma
15 ppm, increased
bacterial loading
Regimen B
18 hours
0, 0.5, or 1 ppmb
None
No effect
Regimen C
None
4 hours
0, 0.5, or 1 ppm
No effect
Regimen D
18 hours
0, 0.5, or 1 ppm
4 hours
0, 0.5, or 1 ppm
1 ppm, increased
bacterial loading
4
5 % 1.2, 6.2, 12.3, or 18.5 mg/m3 formaldehyde.
6 b0, 0.62, or 1.2 mg/m3 formaldehyde.
7
8 Source: Jakab (1992).
9
10
11 For regimen A, mice were exposed to 0, 1,5, 10, or 15 ppm (0, 1.2,6.2, 12.3, or
12 18.5 mg/m3) formaldehyde. For regimens B-D, mice were exposed to 0, 0.5, or 1 ppm (0, 6.2, or
13 1.2 mg/m3) formaldehyde. A 30-minute exposure to an infectious aerosol of S. aureus deposited
14 2 x 105 staphylococci in the lungs. Bacterial loading was determined in homogenized lung tissue
15 by culturing diluted aliquots for an estimate of bacteria present immediately after loading and 4
16 hours later. Bacterial loading was expressed as a percentage change between control and
17 formaldehyde-exposed animals. Mice exposed to 15 ppm formaldehyde for the 4 hours
18 following bacterial infection (regimen A) had approximately an 8% increase in bacteria,
19 indicating decreased host resistance (p = 0.006) (Jakab, 1992) (Table 4-47). Mice receiving
20 lower concentrations of formaldehyde following bacterial infection did not have increased
21 pulmonary bacterial loading. Pre-infection exposure to 0.5 or 1.0 ppm did not change bacterial
22 loading 4 hours after infection (regimen B). However, combining an 18-hour pre-infection
23 formaldehyde exposure with a 4-hour postinfection 1 ppm formaldehyde exposure increased
24 pulmonary bacterial loading by approximately 6.5% (p < 0.05). This effect was not seen with
25 only a 0.5 ppm pre- and posttreatment regimen. Increased bacterial loading indicates that
26 formaldehyde exposure (regimens A and D) reduced pulmonary bacterial resistance. This is in
27 apparent contradiction to the findings of increased host resistance by Dean et al. (1984).
28 However, there are important differences between the studies. The studies by Jakab (1992) are
29 acute studies examining effects at the respiratory tract where direct effects are possible.
30 Additionally, in some cases, the exposures were concurrent with bacterial infection, and it is
31 difficult to distinguish the potential for formaldehyde effects directly on the mucociliary
32 apparatus as a barrier to infection.
This document is a draft for review purposes only and does not constitute Agency policy.
4-221 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
A second set of experiments in the same report (Jakab, 1992) examined the effects of
co-exposure to formaldehyde and carbon black on pulmonary infection with S. aureus. The
particle size distribution of the carbon black aerosol was less than a 5 [j,m aerodynamic diameter
and, therefore, 98% respirable. Female Swiss mice were exposed nose only to formaldehyde and
carbon black. Experiments were run at two target concentrations: (1) 2.5 ppm (3.1 mg/m3)
formaldehyde and 3.5 mg/m3 carbon black or (2) 5 ppm (6.2 mg/m3) formaldehyde and
10 mg/m3 carbon black. Co-exposure was given either for 4 hours after a 30-minute S. aureus
infection or 4 hours/day for 4 days as a pretreatment prior to S. aureus infection. Bacterial
loading was determined 0 and 4 hours after the S. aureus infection to assess bacterial survival.
Formaldehyde-carbon black co-exposure did not alter bacterial survival either as a pretreatment
or posttreatment to bacterial exposure. However, this exposure regimen was not run for
formaldehyde or carbon black separately, and the 4 hours/day for 4 days pretreatment was not
included in the formaldehyde alone experiments (Table 4-47).
Jakab (1992) also assessed the phagocytic activity of alveolar MPs collected by lavage at
various time points after formaldehyde, carbon black, or co-exposure. Female Swiss mice were
co-exposed to 5 ppm (6.2 mg/m3) formaldehyde and 10 mg/m3 carbon black 4 hours/day for
4 days. Mice were sacrificed and alveolar MPs harvested 1, 3, 5, 25, and 40 days after exposure.
Mice exposed only to formaldehyde or carbon black were sacrificed 3, 10, 25, and 40 days after
exposure. Fc-receptor-mediated phagocytosis was assessed ex vivo by using sensitized sheep
RBCs. The phagocytic index (PI) was reported as the total number of RBCs in 100 MPs.
Neither formaldehyde nor carbon black exposure alone significantly changed the PI (Jakab,
1992). These findings are consistent with the first co-exposure experiment, since no changes in
PI were seen immediately after exposure. However, co-exposure did decrease the PI of alveolar
MPs in a time-dependent manner, with maximal decrease to less than 70% of controls by 25 days
after exposure (Figure 4-11). Decreases in the PI reflect changes in both the percentage of
phagocytic MPs and the number of RBCs phagocytized (Jakab, 1992). The PI recovered to
control levels by 40 days postexposure.
This document is a draft for review purposes only and does not constitute Agency policy.
4-222 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
22
23
24
25
26
27
110
100
O hcho
~ Carbon
¦ Carbon + HCHO
u>
0
10
20
30
40
Days after last exposure
Figure 4-11. Alveolar MP Fc-mediated phagocytosis from mice exposed to
5 ppm formaldehyde, 10 mg/m3 carbon black, or both.
Note: Exposure was 4 hours/day for 4 days. Each value represents the mean ±
SEM of five determinations.
Source: Redrawn from Jakab (1992).
Holmstrom et al. (1989b) evaluated the effects of long-term formaldehyde exposure on
antibody production. Female Sprague-Dawley rats were exposed to 12.6 ppm formaldehyde
(15.5 mg/m3) 6 hours/day, 5 days/week for 22 months. Body weight, tumor incidence, and
pathology were reported elsewhere (Holmstrom et al., 1989b). Rats were given a subcutaneous
injection of pneumococcal polysaccharide antigens or tetanus toxoid 21 to 25 days prior to
sacrifice. The two vaccines chosen represent T-cell-dependent and T-cell-independent antigens,
respectively. Antibody titers (IgG and IgM) were determined prior to vaccination and at
sacrifice. Formaldehyde treatment had no effect on antibody titers either before or after
vaccination (Holmstrom et al., 1989b).
4.2.1.4.1. Summary offormaldehyde effects on immune function. Although there were initial
reports of systemic immunomodulation attributed to formaldehyde exposure (Leach et al., 1983),
formaldehyde effects on measures of humoral and cell-mediated immunity were not confirmed
by Dean et al. (1984). The authors did report increased host resistance to both tumor and
bacterial tumor challenges after a 3-week exposure to 15 ppm formaldehyde. An increased
This document is a draft for review purposes only and does not constitute Agency policy.
4-223 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
resistance to these challenges, presented distal to the site of formaldehyde exposure
(administered subcutaneously or intravenously), suggests a systemic effect of formaldehyde
exposure. In addition, increased host resistance and hydrogen peroxide release from peritoneal
MPs were reported and confirmed (Adams et al., 1987; Dean et al., 1984). Chronic
inflammation and tissue damage to the respiratory mucosa expected with formaldehyde exposure
may result in an up regulation of the MPS and therefore increase host immunity. It is unclear if
this response would be specific to formaldehyde or similar to enhancement of immune function
seen with chronic inflammation.
Jakab (1992) demonstrated decreased pulmonary resistance to bacterial infection where
animals were exposed to 15 ppm formaldehyde immediately after bacterial loading or when they
were given an 18-hour pre-exposure to formaldehyde followed by 1 ppm formaldehyde exposure
after bacterial loading. The authors speculated that formaldehyde may directly act on pulmonary
MPs, reducing their effectiveness. However, Jakab (1992) showed that there was no change in
Fc-mediated phagocytosis of alveolar MPs immediately after formaldehyde exposures.
Degradation of the protective mucus layer and possible epithelial cell damage may contribute to
more effective bacterial infection in the presence of formaldehyde without a direct action on MP
function. As mentioned above, degradation of the mucus layer may result in a more potent
inoculation and therefore higher bacterial loading.
Although neither formaldehyde nor carbon black alone impacted Fc-mediated
phagocytosis of alveolar MPs, Jakab (1992) demonstrated that there was decreased Fc-mediated
phagocytosis after formaldehyde and carbon black co-exposure. Carbon black may have acted as
a carrier for formaldehyde, allowing higher levels of formaldehyde to be delivered more deeply
into the lungs than would be seen with formaldehyde alone.
Formaldehyde is known to break down the mucus layer protecting the respiratory tract,
allowing exposure of the underlying epithelium (Morgan et al., 1986a, c, d). Additionally,
formaldehyde can directly induce tissue inflammation through sensory irritation via substance P
from the trigeminal nerve (Fujimaki et al., 2004a). These actions together could contribute to
some of the observed effects on immune response attributed to formaldehyde exposures.
Degradation of the protective mucus layer would make antigens more available to the immune
system. It has been shown that direct application of an antigen to the nasal associated lymph
tissue, bypassing the mucus layer, is a more effective delivery of antigen (Hou et al., 2002).
Therefore, increased availability of these antigens to the immune system may in part explain
observed increased antibody production seen against ovalbumin (OVA) or common dust mite
allergen (Der f) during formaldehyde exposure (Sadakane et al., 2002; Riedel et al., 1996;
Tarkowski and Gorski, 1995). Neurogenic inflammation may also contribute to more efficient
This document is a draft for review purposes only and does not constitute Agency policy.
4-224 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
antigen processing and presentation by activation of resident MPs. These factors are consistent
with the observation that formaldehyde exposures do not affect antibody production to antigens
administered outside of the respiratory tract, even after chronic exposures (Holmstrom et al.,
1989b).
This effect was initially observed several days after exposure was ended with maximal
suppression seen 25 days after a 4-day formaldehyde exposure. The delayed onset of this
response, however, suggests an effect beyond the POE effects observed at the time of exposure.
Table 4-48 presents a summary overview of the effects of formaldehyde on immune function in
laboratory animals.
4.2.1.5. Hypersensitivity and Atopic Reactions
Adverse reactions in humans exposed to formaldehyde in the workplace and homes have
been reported, which are consistent with an allergic response or a chemical sensitivity (see
Section 4.1.1 for details). Rashes and skin reactions are reported in some individuals after
dermal exposures, and in some cases exacerbation of asthma is reported after inhalation of
formaldehyde. However, the reports of human reactions do not allow a clear determination of
whether this sensitization is immunogenic or neurogenic in origin. Formaldehyde-induced
sensitization may have both neurogenic and immunologic components. Numerous animal
studies have been conducted in order to understand the potential for sensitization to
formaldehyde. Although hypersensitivity and allergic sensitization are often considered solely
immunologic in origin, neurogenic mechanisms may result in bronchial hypersensitivity and
increased immunologic sensitization. Therefore, the animal studies regarding formaldehyde-
induced sensitization are evaluated discretely in order to examine these etiologic possibilities.
Classically, hypersensitivity is characterized as an immune response to an antigen,
resulting in an inflammatory reaction that itself damages the tissues or is otherwise harmful
(Kuby, 1991). These reactions may be localized, as in topical dermatitis, or systematic, as in
anaphylactic shock from an allergen. Hypersensitivity can be mediated by a humoral immune
response or by a cell-mediated immune response. Four classes of hypersensitivity are generally
recognized that differ in their immune system components and functions. Although a single
agent (e.g., penicillin) may induce all four types of hypersensitivity, it is more usual for an agent
to primarily induce one form of hypersensitivity.
This document is a draft for review purposes only and does not constitute Agency policy.
4-225 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-48. Summary of immune function changes due to inhaled formaldehyde exposure in experimental animals
S"4
>3*
s
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
F344 rats
8
0, 3, 16, 61, 99 ppm
6 hours/day,
5 days/week for
4 weeks
No effects at 3 ppm. Mixed results at higher doses that were not
consistent.
NAb
Leach et al.
(1983)
B6C3F1
mice
(female)
10
15 ppm 6 hours/day,
5 days/week for
3 weeks
Increased H202 production, and increased host resistance to tumor
formation, but other immune parameters unchanged.
LOAEL 15 ppm
Dean et al.
(1984)
B6C3F1
mice
(female)
Pooled
MPs from
a number
of mice
15 ppm 6 hours/day,
5 days/week for
3 weeks
Increased H202 production in MVE-2-elicited peritoneal MPs.
LOAEL 15 ppm
Adams et al.
(1987)
White
Swiss mice
(female)
18
0, 1, 5, 10, or
50 ppm for 18 hours
before and/or
4 hours after a
30-minute exposure
to bacterial infection
(S. aureus)
Combining an 18-hour pre-exposure to formaldehyde with 4-hour
postexposure to formaldehyde increased bacterial loading at 1 ppm by
6.5%.
LOAEL 1 ppm
Jakab (1992)_
White
Swiss mice
(female)
18
5 ppm (2.6 mg/m3)
formaldehyde and
10 mg/m3 carbon
black 4 hours/day for
4 days
Phagocytic index was decreased by co-exposure to formaldehyde and
carbon black but not by either insult alone.
NA
Jakab (1992)
Sprague-
Dawley rats
(female)
5
12.6 ppm
6 hours/day,
5 days/week,
22 months
Formaldehyde treatment had no effect on antibody titers either before
or after vaccination with pneumococcal polysaccharide antigen or
tetanus toxoid.
NA
Holmstrom et
al. (1989b)
o
S ?
a, Co
TO Sj-
§ ^
s ^
§ 3
?»»i.
NA = not applicable.
to
to
On
-------
1
2
3
4
5
6
7
8
9
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
Chemical sensitivity generally implies a neurogenically induced sensitization (Meggs,
1995). A chemical may directly interact with sensory nerves, releasing mediators that trigger
inflammation, such as substance P (a tachykinin). Repeated exposure to the same chemical is
hypothesized to potentiate neurogenic inflammation (Meggs, 1995). The resulting signs of tissue
inflammation may be similar to immunogenic inflammation, but there would be no requirement
that the immune system recognize the chemical as an antigen for this type of response.
Therefore, a chemical may induce one or more clinical signs of atopic asthma without a type 1,
IgE-mediated hypersensitivity response. One form of sensitivity, , directly affects sensory nerve
endings, resulting in neurogenic inflammation and is a well-known health effect attributed to
formaldehyde. Neurogenic responses may result from the direct and acute interaction of the
chemical with sensory nerve ending receptors of the trigeminal nerve that may lead to persistent
rhinitis and an asthma-like reactive airway dysfunction syndrome that may develop after short-
term human exposures (Brooks et al., 1985). Thus, there is evidence to suggest that neurogenic
inflammation may contribute to observed increases in formaldehyde-induced airway
hyperresponsiveness and atopic responses. The available animal studies that have investigated
formaldehyde-induced airway hyperresponsiveness and atopic responses are summarized below.
4.2.1.5.1. Inhalation studies in experimental animals. This section summarizes animal studies
informing the role of formaldehyde-induced chemical sensitization. The symptoms of
sensitization (atopy, airway hyperresponsiveness) are frequently associated with immunologic
markers (cytokine production, leukocyte infiltration histamine release, and antibody production)
but may be mediated by neurogenic sensory irritation, principally by activation of the trigeminal
nerve (see Section 4.1.1.1 for a discussion of sensory irritation). The animal studies that
illuminate these neurogenic and immunologic responses are discussed outside of the classic
neurotoxicology and immunotoxicology study summary sections to allow synthesis of these data.
Sensitization to chemical exposure by inhalation often manifests as an allergic or
asthmatic response as characterized by BC or BHR. This sensitization may be a result of
immune involvement, as in the case of hypersensitivity, or a neurogenic sensitization, where a
chemical may directly stimulate inflammation. Asthma is a specific manifestation of IgE-
mediated hypersensitivity, characterized by BHR and airway inflammation, resulting in lower
airway obstruction (Fireman, 2003; Kuby, 1991). In asthma, an allergen capable of cross-linking
membrane-bound IgE on mast cells initiates immunogenic inflammation resulting in an influx of
eosinophils, neutrophils, and lymphocytes. Mediators of BC, including histamine, eicosanoids,
and bradykinin (Kuby, 1991), are released during this process. Prior exposure to the allergen can
This document is a draft for review purposes only and does not constitute Agency policy.
4-227 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
increase allergen-specific IgE, potentiating the allergic reaction; this is immunogenic
sensitization.
Biagini et al. (1989) evaluated the effect of a single pulmonary exposure of formaldehyde
on pulmonary mechanics, including BC. The researchers chose cynomolgus monkeys known to
be hyperreactive to methacholine (acetyl-P-methacholine chloride), which is a direct-acting
stimulant of BC (Cain, 2001). Measures of pulmonary mechanics included pulmonary flow
resistance; dynamic compliance; PEFR; FVC; FEV; FEF25_75%, and 50% of VC; and FEFs
normalized for VC. Nine cynomolgus monkeys were exposed to increasing levels of
methacholine for 1 minute at 10-minute intervals (0, 0.125, 0.5, 2, and 8 mg/mL) as an aerosol
(0.065 mL/minute with a mean aerodynamic diameter of 1.0-1.5 (j,m). Pulmonary mechanics
were measured to establish each monkey's response to methacholine. Methacholine challenge,
as the positive control, increased pulmonary flow resistance at increasing levels of methacholine
(0.125, 0.5, 2, and 8 mg/mL) to 196 ± 16, 285 ± 57, 317 ± 64, and 461 ± 120 % of baseline
levels, respectively. After a 2-week recovery period, each methacholine-sensitized monkey was
exposed to 2.5 ppm formaldehyde (generated from formalin, 15% methanol) for 10 minutes.
Measures of pulmonary function were performed at 2, 5, and 10 minutes after exposure.
Formaldehyde exposure increased pulmonary flow resistance from 11.3 ± 1.4 cm H20
prior to formaldehyde exposure to 16.1 ± 2.1, 16.9 ± 2.8, and 20.0 ± 3.4 cm H20 at 2, 5, and
10 minutes after 2.5 ppm formaldehyde exposure (with 142, 150, and 177 % change,
respectively). All other measures of formaldehyde-induced pulmonary mechanics were not
significantly different from controls. Increased pulmonary flow resistance, a measure of
increased BC, was induced by formaldehyde challenge in previously sensitized mice. However,
the differences between methacholine challenge and formaldehyde challenge were not
statistically significant. Although both formaldehyde challenge and methacholine challenge
increased pulmonary flow resistance, there was no correlation between individual methacholine
responsiveness and the magnitude of effect after formaldehyde exposure (p > 0.1). Therefore,
although formaldehyde exposure stimulated BC similarly to a known direct stimulating agent,
formaldehyde may not work through the same site of action as methacholine.
Swiecichowski et al. (1993) assessed pulmonary resistance and airway reactivity due to
formaldehyde exposure alone and in response to increasing doses of acetylcholine chloride (a
direct-acting BC agent) after formaldehyde exposure in vivo. Male Hartley guinea pigs (eight
per group) were exposed at 0.86, 3.4, 9.4, or 31.1 ppm (1.1, 4.2, 11.6, or 38.3 mg/m3)
formaldehyde for 2 hours or at 0.11, 0.31, 0.59, or 1.05 ppm (0.14, 0.38, 0.73, or 1.29 mg/m3)
formaldehyde for 8 hours. Total pulmonary resistance increased after 2 hours formaldehyde
exposure at 9.4 and 31.1 ppm and reached similar peak resistance at the end of the exposure
This document is a draft for review purposes only and does not constitute Agency policy.
4-228 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
period. This effect was rapidly reversible, with values returning to baseline within 30 minutes
after exposure. Although 2-hour exposures at 3 and 1 ppm did not increase pulmonary
resistance, 8-hour exposures at 0.3 and 1 ppm did increase pulmonary resistance to similar levels
as the 2-hour exposure at 30 ppm. The results indicate that both concentration and exposure time
impacted the measured increase in pulmonary resistance. However, a simple multiplicative
model (e.g., C x t) does not adequately represent the effects observed. It is noted that an 8-hour
exposure at 1 ppm (8 ppm-hours), reached approximately the same pulmonary resistance as 2
hours at 9.4 ppm (19 ppm-hours). This may in part be due to a maximum practical increase in
pulmonary resistance in the animals. Conversely, there was no effect at 3 ppm for 2 hours
(6 ppm-hours), although significant increase in pulmonary resistance was recorded after an
8-hour exposure at 0.3 ppm (2.4 ppm-hours). Formaldehyde does not appear to exert its effects
via a classic C x t paradigm. Exposure concentration, however, did seem to impact recovery
time.
In addition, specific pulmonary resistance and airway reactivity to increasing doses of
intravenous acetylcholine chloride, a direct respiratory stimulant, were measured immediately
after formaldehyde exposure for up to 60 minutes. Formaldehyde-induced airway
hyperreactivity was defined as a decrease in the level of acetylcholine chloride needed to
produce twice the basal specific resistance (effective dose [ED]20o). The dose of acetylcholine
chloride required to double the specific pulmonary resistance (ED2oo) and airway reactivity was
decreased in animals exposed for 2 hours to formaldehyde. When the duration was extended to 8
hours of formaldehyde exposure, the effective dose of formaldehyde required to elicit a doubled
pulmonary resistance (ED2oo) in the presence of acetylcholine chloride was decreased to
1.07 ppm. Lower ED2ooS were recorded in formaldehyde-treated animals. This indicates that
less acetylcholine was needed to produce BC when formaldehyde was present. Thus,
formaldehyde can exacerbate BHR. Additionally the formaldehyde-induced effect increased
with duration of exposure, indicating that time as well as exposure concentration are factors in
the magnitude of the response. Directly induced increases in airway hyperreactivity peaked 1
hour after exposure and persisted 6 hours after exposure.
In a second set of experiments, male Hartley guinea pigs were treated for 8 hours at
3.4 ppm (4.2 mg/m3) in order to measure airway hyperreactivity ex vivo (Swiecichowski et al.,
1993). After formaldehyde exposure, tracheae were excised and mounted in tissue baths, where
tracheal contraction was measured in response to direct application of acetylcholine and then
carbachol. Tracheae from similarly exposed guinea pigs were fixed and sectioned for histologic
examination and were assessed for signs of inflammation. Formaldehyde exposure did not
increase ex vivo tracheal constriction and suggests that changes in airway reactivity were
This document is a draft for review purposes only and does not constitute Agency policy.
4-229 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
produced due to both local humoral changes and neural reflexes. However, no changes in
epithelial cell morphology or influx of inflammatory cells were recorded even up to 4 days after
formaldehyde exposure ended. The authors speculated on possible MO As for BHR, such as the
role of an irritant receptor or altered epithelial cell biochemistry. It may be that the window of
acute inflammation occurred early in the exposure protocol and was resolved by the time of first
measurement, after 8 hours of exposure. The absence of inflammatory markers may argue
against a classic type 1 sensitivity.
The binding of an allergen to receptor bound IgE triggers degranulation of mast cells and
basophils, releasing mediators of type 1 hypersensitivity, including the histamine responsible for
BC. Brown Norway (BN) rats are known for their high capacity for IgE production and airway
hyperresponsiveness in response to allergens or other chemicals; they have often been used as a
model of allergic respiratory disease. Ohtsuka et al. (1997) compared the effects of
formaldehyde exposure on the nasal epithelium of F344 and BN rats. If the formaldehyde-
induced inflammatory response in the nasal epithelium is IgE mediated, BN rats would be
expected to display more severe effects of formaldehyde exposure than F344 rats. Both strains
of age- and sex-matched rats were exposed to formaldehyde aerosol for 3 hours/day, 5
days/week for 2 weeks. The aerosol was generated from a 1% formaldehyde solution by a two-
fluid atomizer, and formaldehyde level was maintained at 2 mg (1% sol.)/L (approximately
16 ppm or 20 mg/m3), by adjusting the flow rate for formaldehyde solution to the atomizer.
During the course of exposure, the following clinical signs were monitored: abnormal
respiration, stridor wheezing, nasal discharge, and sneezing. Rats were weighed weekly. Two
days postexposure, rats were sacrificed and tissues from the head, trachea, and lungs were fixed
and sectioned. Transverse sections were taken at the following palatal landmarks from three
animals: level 1 (lateral edge of incisor teeth), level 2 (between incisive papilla and the first
palatal ridge), and level 3 (on the second upper molar). The nasal septa of the remaining two
animals were revealed for examination by electron microscopy.
Formaldehyde-treated F344 rats showed less body weight gain over the 2-week
treatment, resulting in lower body weight at week 1 and week 2 than F344 controls (p < 0.05 and
0.01). Body weights of formaldehyde-treated BN rats were unchanged from BN controls. The
authors observed fewer clinical signs of respiratory irritation in the formaldehyde-exposed BN
rats compared with formaldehyde-exposed F344 rats, such as abnormal respiration (three versus
five) and nasal discharge (three versus five). Histologic analysis of lung and trachea tissues
revealed no distinct signs of inflammation in either strain. Formaldehyde exposure induced cell
damage in URT tissues. Epithelial cell damage was milder and impacted a smaller portion of the
URT in BN rats compared with F344 rats. Squamous metaplasia were present in the respiratory
This document is a draft for review purposes only and does not constitute Agency policy.
4-230 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
epithelium (levels 1 and 2) in both strains in formaldehyde-treated rats. However, a distinct
keratinized layer was noted in level 1 epithelium of F344 rats, and the extent of lesions in level 2
respiratory epithelium was much greater than that seen in BN rats. Additionally, the olfactory
epithelium (level 2) in formaldehyde-exposed F344 rats exhibited degeneration, necrosis, and
desquamation not seen in BN rats. Mild squamous metaplasia was noted in level 3 of the
respiratory epithelium in the treated F344 rats but not the treated BN rats. No pulmonary
function measurements were taken, and, thus, no direct comparison in BHR or BC between BN
and F344 rats in response to formaldehyde can be made. It appears that BN rats are more
resistant to formaldehyde-induced cell damage than are F344 rats, despite the fact that BN rats
are known to be IgE responders. These results suggest that IgE responsiveness may be
protective of formaldehyde-induced cell damage, or IgE may not play a role at all. The authors
note that their earlier research indicated the BN rats have well-developed submucosal glands and
speculate that greater mucus flow may be partly responsible for the greater resistance of BN rats
to the histologic signs of formaldehyde toxicity.
In a subsequent study in the same laboratory, Ohtsuka et al. (2003) compared histology
and cytokine profiles in the nasal mucosa of formaldehyde-treated F344 and BN rats.
Formaldehyde aerosol was generated as above and rats (nine per group) were exposed
3 hours/day for 5 days to approximately 16 ppm of formaldehyde (20 mg/m3). Clinical signs
were recorded daily, and monitored respiratory parameters included abnormal respiration, stridor
wheezing, nasal discharge, and sneezing. Tissue sections of the nose (five rats per group) were
prepared for light microscopy as above: transverse sections at levels 1, 2, and 3. Th-1 cytokines
(IFN-y, IL-2) and Th2 cytokines (IL-4 and IL-5) were determined from the whole nasal mucosa
in four rats of each treatment group.
As expected, lesions and neutrophilic infiltration were more severe in F344
formaldehyde-exposed rats compared with treated BN rats. In addition, lesions were observed in
all three levels of epithelium examined in F344 rats and impacted both respiratory and olfactory
epithelium. Mucosal lesions in formaldehyde-treated BN rats impacted the respiratory
epithelium of levels 1 and 2 only. Changes in formaldehyde-induced cytokine mRNA
expression were modest in both strains. Th-1-related cytokines (IFN-y, 11-2) in formaldehyde-
treated BN rats were significantly decreased compared with control BN rats. A similar, although
not statistically significant, decrease in Th-2 cytokines (IL-4, IL-5) was observed in
formaldehyde-treated BN rats compared with unexposed BN rats. There were no treatment
differences in either Th-1 or Th-2 cytokine expression in formaldehyde-treated F344 rats
compared with unexposed F344 rats. The modest changes in cytokine profile reported in
formaldehyde-treated BN rats were not consistent with type 1 hypersensitivity since type 1
This document is a draft for review purposes only and does not constitute Agency policy.
4-231 DRAFT—DO NOT CITE OR QUOTE
-------
1 hypersensitivity reactions generally result in increased Th-2 cytokines. The mRNA expression
2 results were not corroborated with protein levels and may not have been captured at their peak
3 expression levels.
4 Lee et al. (1984) evaluated the potential for formaldehyde to act as a sensitizing agent
5 through different routes of exposure in guinea pigs. The inhalation studies will be highlighted
6 here. Dermal exposure and associated contact sensitivity results will be discussed in the dermal
7 exposure section (4.2.1.5.2). Three groups of male English smooth-haired guinea pigs
8 (four/group) were exposed via inhalation to either 6 or 10 ppm (7.4 or 12.3 mg/m3)
9 formaldehyde 6 hours/day for 5 consecutive days. Depending on the group, animals were then
10 subjected to bronchial provocation challenge with 2 or 4 ppm formaldehyde on day 7 or days 7,
11 22, and 29 after exposure (see Table 4-49 for clarification).
12
13 Table 4-49. Study design for guinea pigs exposed to formaldehyde through
14 different routes of exposure: inhalation, dermal, and injection
15
Formaldehyde
exposure
Bronchial provocation
challenge
Skin test
Blood drawn
for antibody
titer
Group I—Inhalation
6 ppm formaldehyde3,
days 1-5
Day 7
2 ppm formaldehyde3 for 1
hour
Day 9
Day 14
Group II—Inhalation
10 ppm formaldehyde,
days 1-5
Day 7
2 ppm formaldehyde for 1 hour
Day 9
Day 14
Group III—Inhalation
10 ppm formaldehyde,
days 1-5
Days 7, 22, and 29
4 ppm formaldehyde for 4
hours
Day 31
Day 14
Group IV—Dermal
100 |iL formalin, days 1
and 3
Day 22
2 ppm formaldehyde for 1 hour
4 ppm formaldehyde for 4
hours
Day 7
Day 14
Group V—Injection
37 mg formaldehyde
with Freund's adjuvant
Day 19
2 ppm formaldehyde
Day 7
Day 14
16
17 Source: Lee etal. (1984).
18
19
20 Dermal and injection groups are shown for comparison. Pulmonary hypersensitivity was
21 assessed by measuring respiratory rate and tidal volume in response to exposure to 2 ppm
22 formaldehyde challenge for 1 hour, 2 days postexposure for all three groups, and additional
23 measurements were taken 22 and 29 days postexposure for group III. Blood was drawn to
24 characterize IgE antibodies to formaldehyde in a passive cutaneous anaphylaxis (PCA) assay.
25 Respiratory rate was measured following initial formaldehyde exposure and again after bronchial
26 challenge with formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
4-232 DRAFT—DO NOT CITE OR QUOTE
-------
1 Respiratory rate (exhibited as a pause during expiration) was depressed by 45% following
2 exposure to 10 ppm formaldehyde during the first hour of exposure. During the first hour of
3 exposure, decreased respiratory rate was accompanied by a pause during expiration that has been
4 categorized as RB and indicated sensory irritation. The decreased respiratory rate is consistent
5 with URT sensory irritation and induction of the trigeminal (neurogenic) reflex (Lee et al.,
6 1984). After the first hour of exposure, decreased respiratory rate was characterized by a pause
7 between breaths, which is similar to the breathing pattern seen in mice exposed to formaldehyde
8 via tracheal cannula (Alarie, 1981). This suggests a separate effect of formaldehyde on the LRT
9 after deep penetration of formaldehyde and suggests pulmonary irritation (Lee et al., 1984).
10 However, subsequent bronchial provocation challenge with either 2 or 4 ppm
11 formaldehyde for either 1 or 4 hours failed to elicit immediate or delayed-onset respiratory
12 sensitization (Table 4-50). Respiratory rates were reported as being within ±20% of pre-
13 challenge levels (data not shown) and did not reflect statistical significance (Lee et al., 1984).
14 Moreover, increased respiratory sensitivity was not observed in animals that had received an
15 emulsification of formaldehyde and Freund's complete adjuvant by injection. Only two to four
16 animals given formaldehyde injections in the presence of Freund's complete adjuvant developed
17 a low titer of antibodies to formaldehyde (Lee et al., 1984).
18
19 Table 4-50. Sensitization response of guinea pigs exposed to formaldehyde
20 through inhalation, topical application, or footpad injection
21
Exposure route
Pulmonary sensitization
Dermal sensitization
Antibody production
Inhalation
6 ppm (Group I)
0/4
0/4
0/4
10 ppm (Group II)
0/4
0/4
0/4
10 ppm (Group III)
0/4
2/4
0/4
Topical
0/8
8/8
0/8
Injection
0/4
4/4
2/4
22
23 Source: Lee etal. (1984).
24
25
26 Thus, inhalation exposure to 6 or 10 ppm formaldehyde (8 hours/day for 5 days) followed
27 by bronchial challenge with 2 or 4 ppm formaldehyde failed to result in respiratory sensitivity
28 defined as greater than 20% change in respiratory rate. Second, for animals that received an
29 injection of formaldehyde with Freud's adjuvant, it was not effective in inducing pulmonary
30 sensitivity. While neither inhaled formaldehyde challenge nor injected formaldehyde and
31 Freud's adjuvant emulsion were effective in producing pulmonary sensitivity, this study relied
32 on increased respiratory rate as an indication of hyperresponsiveness and may not be an accurate
This document is a draft for review purposes only and does not constitute Agency policy.
4-233 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
measure of hyperresponsiveness. Thus, overall, conclusions are uninformative due to study
design flaws.
Riedel et al. (1996) tested the effects of formaldehyde inhalation on the development of
sensitization to a known allergen. Female Perlbright-white Dunkin-Hartley guinea pigs (12 per
group) were exposed to 0, 0.13, or 0.25 ppm formaldehyde 8 hours/day for 5 consecutive days.
On day 5, the animals were sensitized to the common model allergen, OVA, in a 3-minute, head-
only exposure to an aerosol of a 5% OVA solution. A booster sensitization with OVA occurred
on day 19. A compressor nebulizer with an output rate of 0.75 mL/minute generated the aerosol.
Particle size ranged from 0.5 to 5.0 [j,m. On day 26, bronchial provocation testing was conducted
with 1% OVA challenge (aerosol). Blood samples were taken and anti-OVA IgG antibodies
were quantified by ELISA. Significant airway obstruction was defined as an increase in
compressed air in the lung that cannot be expired. Three guinea pigs were exposed to
formaldehyde (0.20 ppm) or clean air for 5 days. Immediately after exposure, lung and tracheal
tissues were fixed for histologic and morphometric evaluation. Wall thickness of bronchial and
alveolar septa was measured systematically with a microscope-digitizing-table set.
Significant airway obstruction as measured by compressed air was seen in 3 of 12
controls, 8 of 12 0.13 ppm-exposed, and 10 of 12 0.25 ppm-treated animals after OVA challenge.
The average airway obstruction was increased after 0.25 ppm (mean = 0.35 mL,/> < 0.01) but not
after 0.13 ppm formaldehyde exposure. However, individual response to OVA sensitization was
highly varied, and animals exhibiting a 10-fold increase in obstruction (measured as compressed
air) were seen in both treatment groups (0.13 and 0.25 ppm). Even at the lower exposure (0.13
ppm), biologically significant responses were seen in individuals (Figure 4-12).
Specific anti-OVA antibodies (IgGl class) were not detected in animals prior to
sensitization or in control-treated animals after sensitization . Measurable anti-OVA antibodies
were elevated in 3 of 12 (at 0.13 ppm) and 6 of 12 (at 0.25 ppm) formaldehyde-treated guinea
pigs after sensitization (Figure 4-13). The average anti-OVA titer for the high-dose group was
significantly higher than for controls (p < 0.05). The individual responses at the 0.13 ppm
exposure level indicate that, although the average group OVA titer may not have reached
statistical significance, there was a measureable biological response in three individuals. These
results indicate that formaldehyde exposure can sensitize previously naive (non-sensitized)
animals to OVA.
This document is a draft for review purposes only and does not constitute Agency policy.
4-234 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
1
0,8
0,6
0,4
0,2
0
Figure 4-12. Compressed air in milliliters as parameter for airway
obstruction following formaldehyde exposure in guinea pigs after OVA
sensitization and OVA challenge.
Note: CA = compressed air; FA = formaldehyde; — = median; ** =p < 0.01.
Source: Redrawn from Riedel et al. (1996).
1.000
100
10
Figure 4-13. OVA-specific IgGl (IB) in formaldehyde-treated sensitized
guinea pigs prior to OVA challenge.
Note: EU = experimental units; FA = formaldehyde; — = median; ** = p < 0.01.
Source: Redrawn from Riedel et al. (1996).
Controls
FA 0.13 ppm FA 0.25 ppm
Anti-OVA-lgG1 [EU]
Controls FA 0.13 ppm FA 0.25 ppm
This document is a draft for review purposes only and does not constitute Agency policy.
4-23 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
The only significant, treatment-related histologic change was bronchial edema, with
thickening of the bronchial wall in formaldehyde-exposed animals compared with non-treated
animals subjected to OVA sensitization and subsequent OVA challenge. Bronchial walls were
measured as 40.9 ± 2.5 versus 28.2 ±1.2 [j,m. No signs of inflammation in the bronchial mucosa
were seen with this edema.
Tarkowski and Gorski (1995) exposed female Balb/C mice to 0 or 6.63 ppm (0 or
2 mg/m3) formaldehyde for either 6 hours/day for 10 days or 6 hours/day once a week for
7 weeks. All mice were sensitized intranasally to OVA for 10 days or once a week for 7 weeks.
IgE anti-OVA titers were determined from sera collected from four mice every 8 days (1 day
after OVA booster) by PCA. A parallel experiment to compare the role of the route of
administration was conducted with I.P. rather than intranasal sensitization (1 jag OVA once every
7 days).
OVA titers increased similarly in control mice and mice exposed to formaldehyde once a
week (Figure 4-14). In contrast, mice exposed to formaldehyde 6 hours/day for 10 consecutive
days at the beginning of the experiment had increased anti-OVA beginning after the fourth OVA
sensitization, which continued to increase through seven doses of OVA to a peak of 70 PCA
units (p < 0.01) (Figure 4-14). Anti-OVA IgE titers were significantly different between
formaldehyde-treated and nonexposed mice.
70
60
50
40
30
20
10
0
3
4
5
6
7
The number of OVA doses
Figure 4-14. Anti-OVA titers in female Balb/C mice exposed to 6.63 ppm
formaldehyde for 10 consecutive days or once a week for 7 weeks.
Note: ¦ = control mice; ~ = formaldehyde once a week x 7; • = formaldehyde
10 days.
Source: Redrawn from Tarkowski and Gorski (1995).
This document is a draft for review purposes only and does not constitute Agency policy.
4-236 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Intraperitoneal sensitization to OVA was much more effective than intranasal
sensitization, resulting in titers as high as 1,000 after 4 weeks. However, there were no
differences between controls and animals treated with formaldehyde via the I.P. route of
exposure. Thus, formaldehyde administered intranasally 6 hours/day for 10 days may facilitate
the sensitization to allergens. These changes were not observed when formaldehyde was
administered intranasally once a week for 10 weeks or via I.P. injection (Tarkowski and Gorski,
1995). The authors speculate that formaldehyde may increase permeability of respiratory
epithelium and destruction of immunologic barriers. Thus the respiratory tract may become
vulnerable to inhaled allergens after formaldehyde exposure (Tarkowski and Gorski, 1995).
Ito et al. (1996) conducted three experiments to examine the effects of acute
formaldehyde exposure on bronchoconstriction and the mediators of vascular permeability.
Male Wistar rats (five to eight per group) were exposed to 0, 2, 5, 15, or 45 ppm (0, 2.5, 6.2,
18.5, or 55.4 mg/m3) formaldehyde for 10 minutes. Baseline pulmonary insufflation and blood
pressure were determined prior to formaldehyde exposure and monitored throughout the
experiment. Vascular leakage was measured by injection of Evans blue dye prior to the
experiment and determining extravasation 5 minutes postexposure. Briefly, lungs were perfused
with 0.9% saline through an aortic cannula. The lower portion of the trachea and main bronchi
were removed, and the Evans blue dye remaining was determined and expressed as ng dye/g
tissue. A second experiment was conducted to determine if dye leakage continued to increase
after exposure. Seven rats were exposed to 15 ppm formaldehyde for 10 minutes, as above.
Evans blue dye was injected 5 minutes postexposure, and tissues were perfused and excised
15 minutes later. The final experiment was conducted to determine the effect of certain receptor
agonists on the formaldehyde-induced microvascular leakage. Ten groups of Wistar rats (four to
seven per group) were exposed to 15 ppm formaldehyde and injected with Evans blue dye, as
before. However, each receptor agonist under test or saline sham was injected 4-5 minutes prior
to the 10-minute formaldehyde exposure. Agonists tested included tachykinin NKi receptor
antagonist (CP-99,994) at 1, 3, or 6 mg/kg; a bradykinin B2 receptor antagonist (HOE 140) at
0.65 mg/kg; and a histamine Hi receptor antagonist (ketotifen) at 1 mg/kg.
Formaldehyde exposure did not change pulmonary insufflation pressure or blood
pressure. Formaldehyde increased vascular permeability in a concentration-dependent manner in
both the trachea and main bronchi for the first 5 minutes after exposure, as measured by Evans
blue dye extravasation (Ito et al., 1996) (Figure 4-15). Vascular permeability was not increased
by formaldehyde exposure from 5 to 15 minutes postexposure (experiment 2). Administration of
a selective NKi receptor antagonist (CP-99,994) inhibited the formaldehyde-induced vascular
This document is a draft for review purposes only and does not constitute Agency policy.
4-237 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
permeability, reducing Evans dye extravasation to control levels at the 3 and 6 mg/kg doses
(Figure 4-16).
Trachea Main Bronchi
S 200t
o 100- ¦
0- L_l — ~
Room 2 5 15 45 2 5 15 45
A'r Formaldehyde Formaldehyde
(ppm) (ppm)
Figure 4-15. Vascular permeability in the trachea and bronchi of male
Wistar rats after 10 minutes of formaldehyde inhalation.
Note: Vascular permeability was tested by an increase in Evans blue dye
extravasation in the tissue. Solid bars: formaldehyde; open bars: room air, n = 7.
Values are the means ± SEM of five to seven animals. *p < 0.05 and < 0.01
versus room-air-exposed group (Williams' test).
Source: Redrawn from Ito et al. (1996).
This document is a draft for review purposes only and does not constitute Agency policy.
4-23 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
200r
M
«
O)
E
O)
c
T 100
>i
¦o
0)
3
<0
C
CO
>
111
Trachea
Main Bronchi
0-L
L
ft
ft
O)
CO
<0
Sham Formaldehyde
L
ft
ft
I
Sham Formaldehyde
Figure 4-16. Effect of select receptor antagonists on formaldehyde-induced
vascular permeability in the trachea and bronchi male of Wistar rats.
Note: Vascular permeability was tested by an increase in Evans blue dye
extravasation. Rats were treated i.v. with 1, 3, or 6 mg/kg CP-99,996 (open bars),
0.65 mg/kg HOE 140 (hatched bars), 1 mg/kg ketotifen (solid bars), or vehicle
(shaded bars) before formaldehyde challenge. Sham: animals were exposed to the
sham gas for 15 ppm formaldehyde (10 minutes) after pretreatment with 0.9%
saline (0.5 mL/kg i.v.). Data are the means ± SEM of six to seven rats/group.
*p < 0.05 versus sham-stimulated group (unpaired Student's t test or Welch's
test). < 0.05. < 0.01 versus 0.9% saline-pretreated, formaldehyde-
exposed control group (Williams' test).
Source: Redrawn from Ito et al. (1996).
Neither the bradykinin B2 nor histamine Hi receptor agonists affected formaldehyde-
induced vascular permeability (Ito et al., 1996). Therefore, the immediate effect of
formaldehyde exposure on vascular permeability is mediated, at least in part, through the NKi
receptor but does not seem to require the B2 or Hi receptors. This implies a role for tachykinins
in formaldehyde-induced vascular permeability. These findings suggest a neurogenic
inflammatory response because the tachykinins are released from sensory nerve endings in the
trachea and bronchi, whereas bradykinin is released from mast cells.
Sadakane et al. (2002) investigated the effects of formaldehyde exposure on airway
inflammation caused by Der f. Two groups of male outbred ICR mice (18/group) were exposed
to an aerosol of 0.5% formaldehyde solution produced by an ultrasonic nebulizer for 15 minutes,
once a week for 4 weeks. Two groups were similarly treated but exposed to saline aerosol only.
Details of the aerosol generation and resulting magnitude of exposure were not given. One
group each of control and formaldehyde-exposed mice was sensitized to Der f by an injection 1
day prior to formaldehyde exposure (1.5 mg/animal). The same groups were challenged with
This document is a draft for review purposes only and does not constitute Agency policy.
4-239 DRAFT—DO NOT CITE OR QUOTE
-------
1 intratracheal instillation of Der f (10 (j,g/animal) after 4 weeks. Three days after allergen
2 challenge, mice were sacrificed and blood plasma and lung tissue were collected. Blood plasma
3 was analyzed for Der f-specific immunoglobulins (IgGl and IgE). Lungs from nine mice in each
4 treatment group were homogenized, and Thl cytokine IL-2, Th2 cytokines IL-4 and IL-5,
5 granulocyte macrophage-colony-stimulating factor (GM-CSF), and the "chemokine regulated
6 upon activation, normal T-cell expressed and secreted" (RANTES) protein levels were quantified
7 in the supernatant via ELISA. Lungs from nine mice in each group were fixed, sectioned, and
8 stained to evaluate eosinophil infiltration, lymphocyte infiltration, goblet cell proliferation, and
9 localization of RANTES in the airway epithelium.
10 Der f-specific IgGl was present in blood plasma of sensitized mice but was unchanged
11 by formaldehyde exposure (Sadakane et al., 2002). IgE was too low to titer. IL-2 and GM-CSF
12 were undetected in lung homogenate supernatant, and 11-4 was unchanged by sensitization or
13 formaldehyde exposure. However, RANTES was increased by both formaldehyde exposure and
14 allergen sensitization and challenge (Table 4-51). These increases were more pronounced but
15 less than additive for formaldehyde-exposed, allergen-sensitized mice. IL-5 was increased by
16 allergen but unaffected by formaldehyde exposure only. However, formaldehyde exposure
17 potentiated the IL-5 increase seen with allergen challenge.
18
19 Table 4-51. Cytokine and chemokine levels in lung tissue homogenate
20 supernatants in formaldehyde-exposed male ICR mice with and without
21 Der f sensitization
22
Group
Formaldehyde
Derf
GM-CSF
IL-2
IL-4
IL-5
RANTES
1
-
-
NDa
ND
68.1 ±.9
4.4 ±0.3
200.1 ± 19.7
2
+
-
ND
ND
59.5 ±4.3
4.1 ±0.2
390.6 ±37.4b
3
-
+
ND
ND
70.7 ±4.9
13.6 ± 1.6 ce
479.6 ± 80.0°
4
+
+
ND
ND
62.3 ±5.8
21.6 ± 2.7 c'e'f
593.3 ±58.2c'd
23
24 aNone detected.
25 hp < 0.05 from control.
26 °p < 0.001 from control.
27 dp < 0.05 from Group 2.
28 ep < 0.001 from Group 2.
29 {p < 0.001 from Group 3.
30
31 Source: Sadakane et al. (2002).
32
33
34 Der f sensitization and challenge increased eosinophil infiltration into the interstitium
35 around the bronchi and bronchioles as well as goblet cell proliferation in the bronchial
36 epithelium (Figure 4-17). Formaldehyde exposure exacerbated the eosinophilic and goblet cell
37 responses to a challenge dose of Der f (p< 0.05) (Sadakane et al., 2002). Formaldehyde-induced
This document is a draft for review purposes only and does not constitute Agency policy.
4-240 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
eosinophilic infiltration in the absence of sensitization and challenge was not different from non-
treated, non-sensitized mice.
These results suggest that formaldehyde exposure may aggravate eosinophilic infiltration and
goblet cell proliferation that accompanies allergic responses. This response is associated with an
increase in IL-5, an eosinophilic attractant, and an increase in RANTES, which recruits
eosinophils by chemotaxis in formaldehyde-exposed and Der f challenged animals, although the
effect was not statistically significantly elevated compared with Der f challenge-induced levels
of IL-5 and RANTES alone.
Panel A
Panel B
control
Derf + FA
control
i r
Derf Derf + FA
Figure 4-17. The effects of formaldehyde inhalation exposures on eosinophil
infiltration (Panel A) and goblet cell proliferation (Panel B) after Der f
challenge in the nasal mucosa of male ICR mice after sensitization and
challenge.
Note: ')) < 0.001 compared with control group; bp < 0.001 compared with
formaldehyde group; cp < 0.05 compared with Der f group.
Source: Redrawn from Sadakane et al. (2002).
This document is a draft for review purposes only and does not constitute Agency policy.
4-241 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Fujimaki et al. (2004a) investigated the long-term effects of low-dose formaldehyde
exposure on immunologic and neurological inflammation. Female C3H/He mice were exposed
to 0, 0.082, 0.393, or 1.87 ppm (0, 0.1, 0.48, or 2.3 mg/m3) formaldehyde 16 hours/day,
5 days/week for 12 weeks. Six mice at each exposure level were given injections of OVA plus
adjuvant before the initial exposure and in weeks 3, 6, 9, and 11 of the experiment. Five mice at
each formaldehyde-exposure level did not receive OVA injections. One day after the last
exposure, mice were weighed and blood, BAL, spleen, and thymus were collected from each
animal. After weighing, spleens were disaggregated and spleen cells harvested for cell culture.
Immunophenotype of the spleen cells was determined by flow cytometry (CD4, CD8, CD3, and
CD 19 positive cells). Lymphocyte proliferation in response to lipopolysaccharide (LPS),
phytohemagglutinin A (PHA), or OVA was determined after 72 hours in culture. Splenocytes
were cultured for 48 hours in the presence of LPS, PHA, and OVA (immunized mice only), and
supernatants were collected for cytokine analysis (IL-4, IL-5, and IFN-y). Splenocytes were
cultured for 24 hours in the presence or absence of OVA to assess chemokine production (MCP-
1 and MlPl-a). Anti-OVA IgE, IgGi, IgG2. and IgG3 were quantified in blood plasma.
Body and thymus weights were unchanged by formaldehyde exposure or OVA injection
(Fujimaki et al., 2004a), while, in non-immunized mice, spleen weights were reduced by
formaldehyde exposure from 152 mg in controls to 128, 118, and 121 mg in mice exposed to
0.08, 0.4, and 1.8 ppm formaldehyde, respectively. Spleen weights tended to increase in groups
exposed to 400 and 2,000 ppb formaldehyde compared with controls in OVA-immunized mice
(control: 117.8 mg compared with 400 ppb: 168.6 mg and control: 121.0 mg compared with
2,000 ppb: 153.2 mg, respectively) but were not statistically significant.
To gain insight on the overall pulmonary inflammatory response of mice exposed to
formaldehyde in both immunized and non-immunized mice, the total number and differential
count of MPs, neutrophils, lymphocytes, and eosinophils in BAL were counted and were found
to be unchanged by formaldehyde in non-immunized mice. By contrast, in immunized mice
exposed to 1.8 ppm formaldehyde, the total number of BAL cells, MPs, and eosinophils were
significantly increased compared with non-immunized controls (9.65 versus 2.84, 7.22 versus
2.74, and 2.0 versus 0.02 xlO4 cells, respectively).
To further assess the pulmonary inflammatory response, protein levels of inflammatory
cytokines were determined by ELISA in BAL fluid. Levels of IL-ip in BAL of immunized mice
were decreased by formaldehyde exposure (p < 0.05 at 1.8 ppm formaldehyde), but IL-1J3 levels
after formaldehyde exposure were not different from controls in non-immunized mice (Fujimaki
et al., 2004a). All other cytokines or chemokines were either unchanged (TNF-a, IL-6, and GM-
CSF) or not detected (eotaxin, MlP-la, and MCP-1).
This document is a draft for review purposes only and does not constitute Agency policy.
4-242 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Various neuropeptides, such as brain-derived neurotrophic factor (BDNF), nerve growth
factor (NGF), and substance P are released from vagal nerve endings and mediate a neurogenic
inflammatory response. Levels of BDNF, NGF, and substance P were assessed in BAL fluid
and/or in plasma. BDNF was not detected in BAL or in plasma. NGF levels in immunized mice
were significantly higher than in non-immunized mice in both BAL fluid and in plasma. NGF
levels in immunized mice were significantly attenuated by 0.08 and 0.4 ppm formaldehyde
exposure (Figure 4-18) in both BAL fluid and in plasma. Plasma level of substance P (a
mediator of neurogenic inflammation) was increased by formaldehyde exposures in non-
immunized mice (Figure 4-19) in both BAL fluid and plasma. This increase appears to be dose-
dependent and reaches statistical significance at 2,000 ppb formaldehyde exposure in non-
immunized mice compared with non-immunized controls. Similar to NGF, levels of substance P
increased in OVA-immunized mice compared with non-immunized mice in both BAL fluid and
plasma. Similar to NGF, levels of substance P in OVA-immunized mice were attenuated by
formaldehyde exposure at 80 ppb.
6-F
5-
o>
O
ir ** i
OVA
FA o 80 400 2000
80 400 2000 (ppb)
Figure 4-18. NGF in BAL fluid from formaldehyde-exposed female C3H/He
mice with and without OVA sensitization.
Note: The day after the final formaldehyde inhalation, BAL fluid was collected
from formaldehyde-exposed, non-immunized and formaldehyde-exposed, OVA-
immunized mice, and the production of NGF was determined by ELISA. Data
are mean ± SEM from five to six animals. **p < 0.01.
Source: Redrawn from Fujimaki et al. (2004a).
This document is a draft for review purposes only and does not constitute Agency policy.
4-243 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
6 150
I 100
OVA -
FA. o 80
8 100
Hh + + +
® 80 400 2000 (PPb)
OVA -
FA 0 80 400 2000 § 80 400 2100
-------
1
2
3
4
5
6
7
8
9
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
Formaldehyde exposure (1.8 ppm) increased IFN-y fourfold in LPS-stimulated cultured spleen
cells from non-immunized mice. No other cytokine or chemokine was changed by formaldehyde
exposure in cultured spleen cells from non-immunized mice. In OVA-immunized mice,
formaldehyde had no significant effect on cytokines from stimulated spleen cells. OVA in vitro
stimulation significantly increased the chemokines MTP-1 and MCP-1 for control and
formaldehyde-treated OVA-immunized mice. The OVA-stimulated release of MCP-1 in vitro
was enhanced by formaldehyde exposure in a concentration-dependent manner, increasing
threefold and fourfold at 0.40 and 1.8 ppm, respectively. Increases in MCP-1 correlate with
reported increases in the associated cytokine, RANTES, which recruits eosinophils by
chemotaxis (Sadakane et al., 2002). These formaldehyde-induced increases in cytokine levels
contribute to pulmonary inflammation. The inflammatory response is not mediated by
lymphocytes, since lymphocyte subsets and in vitro cell proliferation were unchanged by OVA
immunization or formaldehyde treatment (Fujimaki et al., 2004a).
Anti-OVA (IgE and IgG2a) levels in plasma were unchanged by formaldehyde exposure.
Anti-OVA IgGi was reduced in immunized mice exposed to 400 ppb formaldehyde compared
with nonexposed animals. However, this effect did not persist as dose increased. Anti-OVA
IgG3 was depressed in immunized mice exposed to 0.08 and 0.4 ppm formaldehyde (Fujimaki et
al., 2004a). Formaldehyde exposure did not induce an inflammatory response in lung or tracheal
epithelium in sections viewed by light microscopy (Fujimaki et al., 2004a). Although there was
a mild infiltration of mast cells into the epithelium of OVA-immunized mice, there were no
effects of formaldehyde treatment on mast cell infiltration.
A recent study by Lino dos Santos Franco et al. (2009) exposed male Wistar rats for
3 days, 90-minutes/day, to 1% formaldehyde (by weight; exact doses not reported) by inhalation.
Of these, one group was sensitized I.P. to OVA (10 |Lxg), a common allergen, immediately
following formaldehyde exposure, and subsequently challenged with OVA 2 weeks later. Other
rats were sensitized and challenged but were not exposed to formaldehyde. PCA reaction as well
as BAL analysis and whole blood analysis were conducted. Immunohistochemical analysis of
platelet endothelial cell adhesion molecule-1 (PECAM-1) expression, an inflammatory mediator,
in lung tissue was also measured. When formaldehyde exposure was followed by OVA
sensitization and challenge, decreased lung inflammation was reported compared with the group
that was OVA-sensitized but had not been exposed to formaldehyde. Reduced lung mast cell
degranulation was also reported in the formaldehyde/OVA group compared with the nonexposed
OVA group. Total circulating leukocytes, total bone marrow cells, and lung protein expression
levels of PECAM-1 were also significantly decreased in formaldehyde/OVA rats compared with
non-formaldehyde exposed OVA rats. The reduction in inflammatory parameters in response to
This document is a draft for review purposes only and does not constitute Agency policy.
4-245 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
formaldehyde may be attributed to different study designs, since in this study animals were
sensitized after exposure rather than prior to exposure. The results suggest that formaldehyde
may functionally alter the activity of certain cells, like mast cells, that may downgrade an
appropriate immune response to antigen and might serve to threaten lung homeostasis. Due to
the unique experimental design of this study, it cannot be directly compared with Sadakane et al.
(2002) or Fujimaki et al. (2004a). In addition, this study did not intend to measure whether
formaldehyde can exacerbate an asthmatic response but rather set out to identify whether
formaldehyde could affect immune homeostasis.
In summary, studies suggest that formaldehyde exposure may induce a predominantly
neurogenic inflammatory response via release of neuropeptide, such as NGF and substance P
from vagal nerve endings. Formaldehyde does not appear to potentiate a systemic immune
response. However, localized pulmonary inflammation can be potentiated by formaldehyde
exposure, as indicated by the increased presence of eosinophils and certain proinflammatory
cytokines (IFN-y). This response does not appear to be mediated by classic immunogenic
mechanisms since studies have failed to report elevated levels of anti-formaldehyde-specific IgE.
Several studies have shown that exposure to formaldehyde can facilitate allergic sensitization in
previously naive animals, and it is thought that this effect may occur due to formaldehyde's
ability to increase microvascular leakage in the nasal epithelium and by causing damage to the
nasal barrier (Ito et al., 1996). Sadakane et al. (2002) demonstrated that formaldehyde exposure
can also exacerbate allergic responses by enhancing the response to challenge allergen. Thus,
formaldehyde may exacerbate allergic responsiveness by aggravating the sensitization response
in previously naive animals by altering the permeability of the mucosal barrier in nasal
compartments. Neurogenically derived inflammation, including stimulation of the trigeminal
nerve and release of braykinin, suggests that the MO A for sensitization may ultimately have its
roots in neurogenic inflammation rather than an immunogenic response. In addition, using a
different protocol, Lino dos Santos Franco et al. (2009) suggest that formaldehyde exposure can
adversely affect lung homeostasis by reducing the activity of important inflammatory mediators
(mast cells, circulating leukocytes, PECAM-1 expression) when it occurs prior to sensitization,
thus downgrading an appropriate immune response.
4.2.1.5.2. Dermal sensitization. Wahlberg (1993) used Hartley strain guinea pigs as test
animals to determine the skin irritancy of a suite of industrial chemicals, including
formaldehyde. Aqueous solutions of the compound in a 0.1 mL volume were applied to the
shaved flanks of guinea pigs and gently rubbed into the skin with a cotton-tipped applicator.
Sites were left open and the treatments repeated once daily for 10 days. A number of indices of
This document is a draft for review purposes only and does not constitute Agency policy.
4-246 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
acute skin irritation were monitored, including erythema via visual scoring and edema and skin-
fold thickness using Harpenden calipers. Varying concentrations of formaldehyde (up to a 10%
solution) induced a dose-dependent increase in skin-fold thickness. Responses also showed
shorter latencies at the higher concentrations. For example, erythema was first observed on
day 2 when 10% formaldehyde was applied, day 5 (for 3%), and day 6 (for 1%).
Lee et al (1984) investigated the role of different routes of exposure in formaldehyde-
induced allergic sensitization. Two sets of four male English smooth-haired guinea pigs received
topical applications of 100 |j,L 37% w/v formalin distributed over two shaved, depilated dorsal
sites two times over the course of 2 days at different sites. The total dose was calculated as
74 |j,g/animal. In addition, eight animals received a single topical application onto a 15 mm area
of the dorsal surface. The applied dose of 25 |j,L formaldehyde was dissolved in saline. Two
other groups of guinea pigs were exposed to either 6 ppm (6 hours/day for 5 days) or 10 ppm
(6 hours/day for 5 days) formaldehyde by inhalation. A third group of guinea pigs was exposed
to 10 ppm formaldehyde for 8 hours/day for 5 consecutive days by inhalation. All animals were
evaluated for contact sensitivity by topical application of 20 mL formaldehyde diluted with
saline and distributed in a 15 mm area on the backs of the shaved guinea pigs (Lee et al., 1984).
Sites were visually inspected for erythema at 1, 6, 24, and 48 hours following the topical
application, and reactions were scored. No erythema was observed in control animals. None of
the guinea pigs in the 6 hours/day inhalation groups (6 and 10 ppm formaldehyde) developed
skin sensitivity tested on day 9 (4 days after the initial exposure regimen ended). Two of four
guinea pigs exposed to 10 ppm formaldehyde for 8 hours/day for 5 consecutive days developed
mild skin sensitization tested on day 31. Contact sensitivity increased in a dose-dependent
fashion in groups of animals that had been sensitized via the dermal route. Thus, dermal
exposure resulted in contact sensitivity. Inhalation exposure did not consistently produce contact
sensitivity.
Arts et al. (1997) used a local lymph node assay (LLNA) and the induction of IgE to
monitor the sensitization of female Wistar rats (low IgE-responders) and BN rats (high IgE
responders). For the LLNA assay, animals were sensitized by the application of varying
concentrations of formaldehyde in raffinated olive oil on the dorsum of both ears on days 0, 1,
and 2. Control animals were treated with raffinated olive oil alone. Animals received an I.P.
injection of BrdU on day 5 and were subsequently sacrificed. Ear-draining lymph nodes were
collected, fixed, and sectioned, and the mitotic activity was monitored following successive
incubation of the sections in anti-BrdU, biotin-labeled rabbit anti-mouse antibody, peroxidase-
conjugated streptavidin, and 3,3-diaminobenzidine tetrahydrochloride. For serum IgE responses,
150 |iL of different concentrations of formaldehyde were applied to the shaved flanks of rats on
This document is a draft for review purposes only and does not constitute Agency policy.
4-247 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
day 1, then 75 |iL of the same chemical at 50% of the initial concentration were applied to the
dorsum of each ear on day 7. The amount of IgE in the blood was measured using ELISA but
appeared to be little affected by formaldehyde treatment in either species of rat. However, the
ear-draining lymph nodes of both strains of rat showed a comparative increase in weight in
response to formaldehyde, and proliferation (BrdU positive) of paracortical cells was observed in
response to increasing doses of the compound. This response was most notable in BN rats
treated with 10% formaldehyde. Arts et al. (1997) concluded that the irritant and sensitizing
properties of formaldehyde may act through non-IgE-immune mechanisms.
Hilton et al. (1998) used the LLNA assay in female CBA/Ca (H-2k haplotype) mice to
compare the skin sensitizing potencies of formaldehyde and glutaraldehyde. The comparison
was set on a quantitative basis by determining the concentration of each compound necessary to
induce a threefold increase in lymph node cell proliferative activity (effective concentration
[EC3]). While both aldehydes induced a dose-dependent proliferative response, the
incorporation of [3H]-methylthymidine was far greater in animals exposed to glutaraldehyde
versus formaldehyde (with EC3 values of 0.002-0.006 mol/L for glutaraldehyde versus 0.11-
0.18 mol/L for formaldehyde). These data indicate the potential of both chemicals to induce skin
sensitization, although the potency of glutaraldehyde was far greater than that of formaldehyde.
Xu et al. (2002) evaluated the extent to which the expression of some cytokines may
change as a result of cutaneous exposure to formaldehyde in mice. Female Balb/C mice were
skin painted with three topical applications of 100 |iL of 17.5% formaldehyde or distilled water
with a 1-day interval between each application. Spleen and draining lymph nodes were
harvested on days 3, 5, 7, 9, or 12 after the last skin painting. In some animals, contact
hypersensitivity was induced by applying 2% formaldehyde to both sides of mouse ears on day 3
following the last skin painting. For this endpoint, the percent increase in thickness of the ears
was monitored. For the cytokines, mRNA expression levels of IL-2, IL-4, IL-5, IL-10, IL-12,
IL-13, IL-15, IL-18, and INF-y were determined semi quantitatively by measuring the amount of
individual mRNAs following amplification with the reverse transcriptase (RT)-PCR. The
relative amounts of cytokine mRNAs were calculated as the ratio of cytokine mRNA to that of
glyceraldehyde-3-phosphate dehydrogenase, as revealed in specific bands on an agarose gel.
Cutaneous formaldehyde treatment was associated with the long-lasting expression of
IL-4 and IFN-y mRNAs in mouse spleen and draining lymph nodes and with IL-15 mRNA only
in mouse spleen. Only IL-13 mRNAs displayed a transient increase in expression in both spleen
and draining lymph nodes. Levels of IL-2, IL-12, and IL-15 were increased in the mouse spleen
but not the lymph nodes. The mouse ear swelling test gave positive correlations with enhanced
expression of mRNA for IL-4 and IFN-y (Table 4-52).
This document is a draft for review purposes only and does not constitute Agency policy.
4-248 DRAFT—DO NOT CITE OR QUOTE
-------
1
2 Table 4-52. Correlation coefficients among ear swelling responses and skin
3 mRNA levels in contact hypersensitivity to formaldehyde in mice
4
Variables
Correlation coefficients
IL-2
IL-4
IFN-y
Ear swelling
0.50
0.74a
0.67a
IL-2
-
0.39
0.60
IL-4
-
-
0.79a
5
6 Statistically significant (p < 0.05).
7
8 Source: Xu et al. (2002).
9
10
11 4.2.1.5.3. Summary of sensitization studies. Several animal studies report increased airway
12 resistance and BC due to inhalation exposures to formaldehyde (Nielsen et al., 1999;
13 Swiecichowski et al., 1993; Biagini et al., 1989; Amdur, 1960). Changes in pulmonary
14 resistance were observed as early as 10 minutes after exposure (Biagini et al., 1989), and
15 reported effect levels ranged from 0.3-13 ppm. Other pulmonary effects were reported in
16 conjunction with BHR, such as increased tracheal reactivity and decreased pulmonary elasticity
17 (Swiecichowski et al., 1993; Amdur, 1960). Although BHR is a common result of Type I
18 hypersensitivity reaction to an allergen, the observation of BHR alone is not sufficient to
19 demonstrate that an agent induces Type 1 hypersensitivity.
20 BHR may be directly induced both pharmacologically and neurogenically (Joos, 2003;
21 Cain, 2001; Meggs, 1995). There is little evidence that formaldehyde itself is an allergen
22 recognized by the immune system, especially via inhalation (Lee et al., 1984). Although
23 formaldehyde exposure has been reported to alter cytokine levels and immunoglobulins in some
24 experimental systems, these immunomodulatory effects do not support a type 1 hypersensitivity.
25 IgE was unchanged (Fujimaki et al., 2004a; Lee et al., 1984), and cytokine profiles were not
26 consistent with the Th-2 cytokines expected in IgE mediated hypersensitivity (Fujimaki et al.,
27 2004a; Ohtsuka et al., 2003).
28 Formaldehyde-induced dermal sensitization show parallel results. The physical signs of
29 irritation and sensitization are consistently shown (e.g., rashes, edema). Some involvement of
30 the immune response has been demonstrated with positive LLNA assays, indicating proliferation
31 of lymphocytes in lymph nodes draining the affected area (Hilton et al., 1998; Arts et al., 1997).
32 Increased expression of Th-2 cytokines in the lymph nodes of mice given dermal applications of
33 formaldehyde does indicate an immune component to the observed sensitization. However, the
34 response does not seem to be mediated by IgE (Arts et al., 1997; Lee et al., 1984).
This document is a draft for review purposes only and does not constitute Agency policy.
4-249 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ito et al. (1996) reported that a tachykinin NKi receptor, but not the histamine Hi or
bradykinin B2 receptors, is involved in formaldehyde-induced vascular permeability.
Neuropeptides NGF and substance P were affected in BAL and stimulated splenocytes from
formaldehyde-exposed mice, with greater effects seen in OVA-immunized mice. Tachykinins
(e.g., substance P and neurokinin A) are produced by nerve cells and can directly stimulate
bronchoconstriction (Van Schoor et al., 2000). Substance P is also a mediator of neurogenic
inflammation. Therefore, although formaldehyde may induce some of the symptoms of type 1
hypersensitivity, these symptoms are more likely neurogenic than immunogenic in origin.
In contrast, formaldehyde enhances immunogenic hypersensitivity of known allergens
(Sadakane et al., 2002; Riedel et al., 1996; Tarkowski and Gorski, 1995). This potentiation
varied based on sensitization protocols (respiratory tract versus systemic, frequency and timing
of immunization, allergen, etc.) and formaldehyde exposure regimens (concentration, continuous
versus intermittent exposures). Taken as a whole, the results support the finding that
formaldehyde exposure can aggravate a type 1 hypersensitivity response (Table 4-53).
4.2.1.6. Neurological and Neurobehavioral Function
4.2.1.6.1. Inhalation exposure. There are a number of published reports examining the effects
of formaldehyde exposure on nervous system structure and function. The reports evaluating
behavioral effects fall into three main categories: (1) behavioral responses evaluated during or
immediately following formaldehyde exposures, which may include effects due to the potential
irritant properties of the chemical, (2) acute or short-term exposures followed by behavioral
assessments conducted 2-24 hours after termination of formaldehyde exposure, which reflect
sustained effects of chemical exposure independent of its irritant properties, and (3) repeated
exposures to formaldehyde followed by neurological assessments performed throughout the
treatment period or several days to weeks after termination of treatment. In addition to reports
evaluating changes in behavior, there are several reports evaluating neuropathological effects or
changes in brain chemistry.
This document is a draft for review purposes only and does not constitute Agency policy.
4-250 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-53: Summary of sensitization and atopy studies by inhalation or dermal sensitization due to
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Inhalation studies
Cynomolgus
monkeys
9
Methacholine-sensitive monkeys
exposed to 2.5 ppm
formaldehyde for 10 minutes
Formaldehyde increased pulmonary resistance after 2, 5, and
10 minutes.
LOAEL
2.5 ppm
Biagini et al.
(1989)
Hartley
guinea pigs
(male)
8
0.86,3.4, 9.4, 31.1 ppm
formaldehyde for 2 hours or
0.11,0.31,0.59, 1.05 ppm for
8 hours
Total pulmonary resistance increased after 2 hours exposure at
9.4 and 31.1 ppm. Effect was reversible and returned to
baseline within 30 minutes. Total pulmonary resistance was
increased after 8 hours exposure at 0.3 and 1 ppm. Amount of
acetylcholine needed to achieve doubled pulmonary resistance
was decreased in animals after 2 hours exposure.
NA
Swiecichowski
et al. (1993)
Hartley
guinea pigs
(male)
5-7
3.4 ppm for 8 hours
No changes in ex vivo tracheal constriction or inflammation.
NA
Swiecichowski
et al. (1993)
F344 rats
and BN rats
5
16 ppm 3 hours/day, 5 days
Modest changes in inflammatory cytokine expression, but
respiratory and olfactory epithelial lesions were more severe in
F344 rats than in BN rats.
NA
Ohtsuka et al.
(2003)
English
smooth-
haired
guinea pigs
4
6, 10 ppm, 6 hours/day, 5 days,
combined with provocation
challenge (2 or 4 ppm on day 7,
or days 7, 22, and 29)
Inhalation challenge with 6 or 10 ppm followed by bronchial
challenge failed to increase respiratory sensitivity
NA
Lee et al.
(1984)
Perlbright-
white,
Duncan-
Hartley
guinea pigs
(female)
12
0, 0.13, 0.25 ppm 8 hours/day,
5 days. The animals were
sensitized to OVA (3 minutes
exposure to 5% OVA aerosol)
Anti-OVA titer was significantly elevated over controls in
animals exposed to 0.25 ppm formaldehyde and showed that
formaldehyde may sensitize previously naive animals to OVA.
NA
Riedel et al.
(1996)
Balb/C mice
(female)
4
0, 6.63 ppm 6 hours/day for
10 days or 6 hours/day
once/week for 7 weeks. All
mice were sensitized to OVA
Formaldehyde administered intranasally for 6 hours/day for
10 days may facilitate sensitization to allergens since anti-OVA
titers were elevated over control animals. However, the length
and duration of exposure appears to affect development of
sensitization.
NA
Tarkowski and
Gorski (1995)
o
5 s
to
o
^ Co'
to &-
§ ^
JS*
*
to
Lt\
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-53: Summary of sensitization and atopy studies by inhalation or dermal sensitization due to
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Wistar rats
(male)
5-8
0, 2, 5, 15, 45 ppmfor
10 minutes
Pulmonary insufflation or blood pressure were not altered.
Vascular permeability increased in concentration-dependent
manner and could be reduced by adding a NK1 selective
antagonist.
NA
Ito et al.
(1996)
Outbred ICR
mice (male)
18
0.5% formaldehyde for 15
minutes, once/week for 4 weeks.
Both control and exposed groups
were exposed to Der f by I.P.
injection 1 day before
formaldehyde and then
challenged with Der f after
4 weeks.
More pronounced RANTES production in formaldehyde-treated
and sensitized rats than in sensitized rats that had not been
exposed to formaldehyde. Formaldehyde also potentiated IL-5
production associated with sensitization.
NA
Sadakane et al.
(2002)
C3H/HeJ
mice
(female)
6
0,0.082,0.393, 1.87 ppm
16 hours/day, 5 day/week,
12 weeks. Mice also given OVA
plus adjuvant before exposure,
and again 3, 6, 9, 11 weeks after
exposure. Some formaldehyde
mice did not receive any OVA
Substance P and NGF were increased dose dependency in
formaldehyde-treated, non-immunized mice but were attenuated
in formaldehyde-treated immunized mice compared with
nonexposed, immunized controls.
LOAEL
0.082 ppm
Fujimaki et al.
(2004a)
Wistar rats
(male)
NA
1% Formaldehyde by weight for
90 minutes for 3 days. One
group was sensitized to OVA
after to formaldehyde exposure
and then challenged with OVA
afterwards. Others were
sensitized and challenged but not
exposed to formaldehyde.
Total circulating leukocytes, bone marrow cells, and lung
protein PECAM expression were significantly decreased in
formaldehyde/OVA rats compared with OVA rats.
NA
Lino dos
Santos Franco
et al. (2009)
Dermal sensitization
Hartley
guinea pigs
5
Skin painted once/day for
10 days with 0.1 mL of 1, 3, or
10% formaldehyde
Varying concentrations (up to 10%) induced dose-dependent
increase in skin-fold thickness. Erythema seen earlier at higher
doses (2 days at 10% formaldehyde vs 5 days at 3% or 6 days at
1%).
Wahlberg et
al. (1993)
o
5 s
to
o
^ Co'
to &-
§ ^
JS*
*
to
Lt\
to
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-53: Summary of sensitization and atopy studies by inhalation or dermal sensitization due to
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
English
smooth-
haired
guinea pigs
4
Group 1: skin painted, 100 |iL
37% formalin twice over 2 days,
Group 2: single topical
application of 25 |iL
formaldehyde
Group 3: 10 ppm 6 hours/day for
5 days by inhalation
Two of four guinea pigs from group 3 had mild skin
sensitization after day 31. Contact sensitivity developed in a
dose-dependent manner in the dermal groups (group 1 and 2).
Lee et al.
(1984)
Wistar and
BN rats
(female)
4
Application of formaldehyde to
ears on days 0, 1,2, followed by
an I.P. injection of BrdU.
Ear-draining lymph nodes increased in weight in response to
formaldehyde, reflected in increased number of BrdU-stained
cells, most notably in BN rats (high IgE responders) treated with
10% formaldehyde
Arts et al.
(1997)
CBA/Ca
mice
NA
Compared glutaraldehyde to
formaldehyde to induce a local
lymph node assay
Glutaraldehyde and formaldehyde induced a dose-dependent
proliferative response that was greater in glutaraldehyde-treated
animals
Hilton et al.
(1998)
Balb/c mice
(female)
3-5
Skin painted with 100 |iL of
17.5% formaldehyde every other
day for days 3, 5, 7, 9, 12
Cutaneous treatment associated with long-lasting expression of
various cytokines from draining lymph nodes and spleen.
NA
Xu et al.
(2002)
o
5 s
to
o
^ Co'
to &-
§ ^
JS*
*
NA = not applicable.
4^
to
U>
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.6.1.1. Behavioral response
Clinical signs
Several studies that were focused on general toxicity or carcinogenicity of formaldehyde
also assessed clinical signs in exposed animals, which may be related to adverse effects on the
nervous system. Procedural details for the assessments, or specific data regarding findings, were
not provided. Signs recorded included uncoordinated locomotion and climbing of cage walls at
20 ppm formaldehyde in rats (Woutersen et al., 1987); restlessness at 15 ppm formaldehyde in
rats (Morgan et al., 1986a); dyspnea, listlessness, and hunched posture at 20 ppm and ataxia at
40 ppm in mice (Maronpot et al., 1986); and dyspnea in rats at 14.3 ppm formaldehyde (Kerns et
al., 1983). Given the lack of information regarding procedures used for these evaluations and the
limited reporting of results, the utility of these data is limited.
Irritant threshold detection
Wood and Coleman (1995) evaluated the irritant properties of acute formaldehyde
exposure in mice. Adult male Swiss mice (eight/group) were initially trained to terminate a
60-second exposure to an irritant gas (ammonia, 1,000 ppm) by poking their noses into a conical
sensor five times to produce a 60-second facial shower of clean air. Each test session consisted
of 25 exposure trials. Following training, response to formaldehyde was evaluated, using the
same testing scenario. Each day mice had a morning exposure session to ammonia and an
afternoon session to formaldehyde. Formaldehyde concentrations tested were different each day,
in sequence from 0, 1, 1.8, 3, 5.6, and 10 ppm (0, 1.23, 2.21, 3.68, 6.87, and 12.3 mg/m3) and
then stepping back again from 10 to 0 ppm. Half of the animals were tested in an ascending
order of formaldehyde concentrations, the other half in a descending order. The frequency of
terminating exposure, error rate, and the time lapse to termination were recorded. The
concentration at which 50% of the formaldehyde deliveries would be expected to be terminated
was estimated (AC50) by simple linear regression or by analysis of covariance on the logit
transform of percentage terminated as a function of log concentration.
All mice were trained successfully to terminate 100% of ammonia exposures, but varied
responses were observed with formaldehyde exposure. In general, time taken to terminate
formaldehyde exposure decreased significantly with increasing formaldehyde concentration.
Mice terminated more exposures to 1 ppm formaldehyde than to air alone (p < 0.0005), and the
error rate, generally below 40%, did not significantly differ with formaldehyde concentration
tested. Each animal had two test sessions with each formaldehyde concentration (once during
the ascending sequence and again during the descending sequence); both the time to termination
(p < 0.0012) and AC50 were decreased in the second series of tests. One method of estimation by
This document is a draft for review purposes only and does not constitute Agency policy.
4-254 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
the authors yielded an AC50 of 3.63 ppm for the first series of tests versus 1.88 ppm for the
second series. A two-way repeated measures ANOVA with replication and concentration as
within variables was highly significant (p < 0.00005). These studies indicate mice are sensitive
to the irritant properties of formaldehyde at exposure concentrations as low as 1 ppm, and
animals reacted more swiftly and with greater accuracy to terminate formaldehyde exposure as
the concentration increased. However, a wide variety of responses was noted on an individual
animal basis. Two of the eight mice terminated 90% of the trials during 1 ppm exposures and
80-100% of trials at all other tested formaldehyde concentrations. One mouse terminated fewer
than 10%) of the formaldehyde exposure trials (1-10 ppm) during the testing regimen but had a
92% response rate to 20 ppm formaldehyde. The remaining five mice responded with increasing
termination frequency as formaldehyde concentration increased from 1 to 10 ppm, with an AC50
of 2.72 ppm (3.34 mg/m3).
4.2.1.6.1.2. Motor activity and habituation. Malek et al. (2003a) examined open field behavior
of rats after acute formaldehyde exposures. Male and female LEW. IK rats (15/sex/group) were
exposed to 0, 1, 2.5, or 5 ppm (0, 1.23, 3.08, or 6.15 mg/m3) formaldehyde for 2 hours.
Formaldehyde was vaporized from aqueous solutions directly below the exposure chamber.
Formaldehyde levels were checked 16 times throughout the 2-hour exposure periods. Mean
formaldehyde levels of 1.01 ± 0.29 ppm, 2.51 ppm (SD is missing), and 5.0 ± 0.27 ppm were
achieved. Locomotor activity was assessed for 3 minutes in an open field 2 hours after
termination of formaldehyde exposure and again 24 hours later, using an automated device to
count the number of squares crossed. Other behaviors were noted, including grooming (face
cleaning, fur licking, and scratching), rearing, sniffing (air and floor), wall climbing, and
defecation.
The authors reported no signs of irritation or changes in activity or food or water intake
during exposure. In general, sniffing was increased after formaldehyde exposure and movement
was decreased (crossed quadrants and climbing) in both male and female rats (p < 0.05).
Significant reductions in horizontal movements (crossed quadrants) were observed at all dose
levels and were characterized by a U-shaped dose response (Figure 4-20). The lowest dose
tested (1 ppm) demonstrated a higher level of activity suppression than the two higher doses, but
all groups were still suppressed relative to controls. Although female rats displayed a greater
level of activity overall, a similar U-shaped dose-response pattern was also observed.
This document is a draft for review purposes only and does not constitute Agency policy.
4-255 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Activity as crossed quadrants
Number
100
80
60
40
20
0
Control 1 ppm 2.5 ppm 5.0 ppm
Formaldehyde
~ Males
¦ Females
**
L
OH
Figure 4-20. Motor activity in male and female rats 2 hours after exposure to
formaldehyde expressed as mean number of crossed quadrants ± SEM.
Greater reductions were observed in the lowest dose group, a pattern that
was evident in both genders. ** = different from control,/; < 0.005.
Source: Drawn from data reported by Malek et al. (2003a).
Activity in the same apparatus was reassessed 24 hours later. As expected, controls
demonstrated habituation to the test apparatus, exhibiting only 20% of the motor activity
observed on day 1 (Figure 4-21). In contrast, formaldehyde-treated animals failed to
demonstrate the same degree of habituation. Activity levels for males observed on day 2 were
60-80% of the activity levels seen on day 1. Formaldehyde-treated females also failed to
habituate and actually demonstrated increases in activity on day 2 relative to day 1 at all
formaldehyde exposure levels.
This document is a draft for review purposes only and does not constitute Agency policy.
4-256 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Female LEW.1 K Rats
Control 1 ppm 2.5 ppm 5.0 ppm
Formaldehyde
Control 1 ppm 2.5 ppm 5.0 ppm
Formaldehyde
Figure 4-21. Habituation of motor activity was observed in control rats
during the second observation period (day 2, 24 hours after formaldehyde
exposure).
Note: Habituation is shown here as the percent decrease in number of crossings
between sessions from day 1 to day 2. The degree of habituation was reduced in
male rats exposed to formaldehyde (left panel) since their activity was closer to
100% of that seen on day 1. Females (right panel) had increased activity on day 2
(greater than 100% of activity on day 1), which is a sensitization rather than
habituation.
Source: Drawn from data reported by Malek et al. (2003a).
A follow-up study by Malek et al. (2003b) further expanded the dose-response analysis
for acute formaldehyde exposure. As described above, male and female LEW. IK rats (10 per
sex per group) were exposed at 0, 0.1, 0.5, or 5 ppm (0, 0.123, 0.615, or 6.15 mg/m3)
formaldehyde for 2 hours. Formaldehyde levels were checked nine times per hour during the
exposure periods, and mean values were found to be 0.13 ± 0.04, 0.48 ± 0.05, and
5.18 ± 0.66 ppm. Open field behavior was evaluated for each animal 2 hours after formaldehyde
exposure. The number of crossed quadrants for both controls and a 5 ppm group were generally
comparable with those observed in the first study, although female values were somewhat lower.
Horizontal movement was decreased by formaldehyde exposure in a dose-dependent manner
with significant reductions in motor activity as low as 0.1 ppm in males and 0.5 ppm in females
(Figure 4-22). The consistency of the findings across studies and between genders provides
This document is a draft for review purposes only and does not constitute Agency policy.
4-257 DRAFT—DO NOT CITE OR QUOTE
-------
1 greater confidence in the effects of low-level formaldehyde exposure on this standard test of
2 neurotoxicity.
3
0 0.1 0.5 5.0 0 0.1 0.5 5.0
formaldehyde (ppm)
4
5 Figure 4-22. Motor activity was reduced in male and female LEW.1K rats
6 2 hours after termination of 10-minute formaldehyde exposure.
7
8 Note: Values are means ± SDs. * = different from control,/* < 0.05. ** = different
9 from controls,/* <0.01.*** = different from controls,/? <0.001.
10
11 Source: Drawn from data reported in Malek et al. (2003b).
12
13
14 Malek et al. (2004) also assessed the capacity of formaldehyde to induce persistent
15 behavioral deficits in mice. Groups of 20 male AB mice received a single 2-hour exposure to 0,
16 1.1, 2.3, or 5.2 ppm (0, 1.3, 2.8, or 6.4 mg/m3) formaldehyde prior to being tested 2 and 24 hours
17 after exposure for a series of behavioral responses, including ambulation (crossed squares),
18 grooming, sniffing, rearing, wall climbing, and defecation. Even though there were no clinical
19 signs of toxicity in any of the exposed groups, a number of behavioral anomalies were apparent
20 in response to formaldehyde exposure, some of which persisted for at least 24 hours, as indicated
21 in Tables 4-54 and 4-55.
22
23
This document is a draft for review purposes only and does not constitute Agency policy.
4-258 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-54. Fluctuation of behavioral responses when male AB mice inhaled
2 formaldehyde in a single 2-hour exposure: effects after 2 hours
3
Open field parameter
Formaldehyde concentration (ppm)a
0
1.1
2.3
5.2
No. of crossed inner squares
34.10 ±7.51
25.30 ±5.03b
21.20 ±3.41b
16.10 ±5.37b
No. of crossed peripheral squares
56.65 ± 9.68
59.55 ±9.75
49.70 ± 13.24
29.15 ±7.47b
Total no. of crossed squares
90.75 ± 11.08
84.85 ± 9.96
71.10 ± 13.91b
44.20 ± 7.42b
Air sniffing
19.35 ±2.5
21.50 ±4.26
16.35 ±3.84c
8.10 ± 1.77b
Floor sniffing
20.95 ±3.72
26.50 ± 4.64b
21.35 ±4.77
22.80 ± 4.02
Grooming
7.95 ±2.26
7.10 ± 3.19
7.05 ±2.48
6.55 ±2.06
Rearing
17.85 ±2.56
13.90 ± 3.19b
11.30 ±2.30b
9.95 ± 1.6lb
Wall climbing
13.20 ±3.09
14.55 ±2.74
13.95 ±2.31
13.95 ± 1.82
No. of excreted fecal boli
0.65 ±0.81
0.75 ±0.85
0.80 ±0.77
0.90 ± 1.12
4
5 aValues are means ± SDs.
6 Statistical significance of differences from controls (p < 0.005).
7 Statistical significance of differences from controls (p < 0.05).
8
9 Source: Malek et al. (2004).
10
11
12 Table 4-55. Fluctuation of behavioral responses when male AB mice inhaled
13 formaldehyde in a single 2-hour exposure: effects after 24 hours
14
Open field parameter
Formaldehyde concentration (ppm)a
0
1.1
2.3
5.2
No. of crossed inner squares
10.40 ±2.35
9.55 ±1.73
9.10 ± 1.25
9.70 ± 1.13
No. of crossed peripheral squares
42.80 ± 9.27
44.85 ± 14.60
44.95 ± 16.56
41.10 ±9.08
Total no. of crossed squares
53.20 ±8.67
54.40 ± 14.77
54.05 ± 15.81
50.80 ±9.15
Air sniffing
13.65 ±2.81
13.30 ±3.21
12.65 ±2.70
12.30 ±4.14
Floor sniffing
21.55 ±3.47
15.85 ±3.94b
13.25 ±4.17b
17.65 ± 3.13b
Grooming
8.35 ±2.56
13.95 ±2.21b
10.20 ±3.33°
11.90 ± 3.26b
Rearing
18.30 ±4.23
12.40 ±2.23b
12.25 ±2.17b
12.00 ±3.32b
Wall climbing
9.25 ±2.38
8.70 ± 1.98
8.20 ±2.14
9.90 ±2.27
No. of excreted fecal boli
0.80 ±0.83
1.20 ±0.83
1.60 ± 0.94°
1.20 ±0.89
15
16 aValues are means ± SDs.
17 Statistical significance of differences from controls (p < 0.005).
18 Statistical significance of differences from controls (p < 0.05).
19
20 Source: Malek et al. (2004).
21
22
23 Usanmaz et al. (2002) assessed spontaneous locomotor activity (SLMA) in Balb/c mice
24 (4-14 per group, sex unspecified) after both acute and subchronic formaldehyde exposures.
25 Prior to the acute exposure, mice were acclimated to the exposure chamber for 4 days but
26 exposed only to clean air. On the fifth day, mice (six/group, sex unspecified) were exposed for
27 3 hours at 0, 1.8, 3.2, 4.5, 6.4, 9.7, or 14.8 ppm (0, 2.2, 3.9, 5.5, 7.9, 11.9, or 18.2 mg/m3)
28 formaldehyde. Mice were removed from the exposure chamber, and SLMA behavior was
This document is a draft for review purposes only and does not constitute Agency policy.
4-259 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
evaluated by direct observation for 5 minutes. In addition to horizontal and vertical movement,
wet dog shake (WDS) behavior was noted. In a separate trial, Balb/c mice (six/group, sex
unspecified) were exposed to 8.2 ppm formaldehyde for 1 week, 2 ppm formaldehyde for
2 weeks, or 3.3 ppm formaldehyde for 3 weeks (3 hours/day, 5 days/week) compared with
controls exposed only to air. SLMA behavior was observed for 5 minutes after the last exposure.
Mice exposed to 8.2 ppm formaldehyde for 1 week, 3.3 ppm formaldehyde for 2 weeks, and
2 ppm formaldehyde for 3 weeks lost weight over the course of the treatment (p < 0.05). All
other treatment groups had weight gain similar to control mice.
As shown in Figure 4-23, acute 3-hour formaldehyde exposures resulted in a dose-
dependent decrease in SLMA. Decreases in horizontal activity were significant for the three
highest dose groups (6.4, 9.7, and 14.8 ppm), and decreases in vertical activity were significant
for all six formaldehyde treatment groups. SLMA was similarly decreased following subchronic
exposures (data not shown here). Although the experimental protocol included longer exposures
and a slightly longer observation period (5 versus 3 minutes) than in Malek et al. (2003a, b), the
results are consistent, indicating decreased activity in formaldehyde-exposed animals several
hours after exposure was ended.
SLMA after acute exposures
100 -¦
lit
s 80 4
» 60-t
5
3 40 -¦
o
o
20 -¦
0
i
I
~
~
¦
~
~
~
Control
FA 1.8 ppm
FA 3.2 ppm
FA 4.5 ppm
FA 6.4 ppm
FA 9.7 ppm
FA 14.8 ppm
Ambulatory Activity
Vertical Activity
Figure 4-23. The effects of the acute formaldehyde (FA) exposures on the
ambulatory and vertical components of SLMA.
Note: FA = formaldehyde exposure concentration. ** = p < 0.01 from controls.
Source: Usanmaz et al. (2002).
Usanmaz et al. (2002) noted an increase in WDS, after the acute exposures, as a possible
preconvulsive effect. However, the mice were only observed for 5 minutes, and it is unclear how
This document is a draft for review purposes only and does not constitute Agency policy.
4-260 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
the researchers distinguished between a WDS due to an irritating odor and a preconvulsive
movement. No other study has noted convulsive effects from formaldehyde exposure in any
species. A second set of trials was reported in the same paper that sought to evaluate
formaldehyde effects on CNS excitability. Balb/c mice (six/group, sex unspecified) were
exposed to 0, 1.8, 6.4, or 14.8 ppm (0, 2.2, 7.9, or 18.2 mg/m3) formaldehyde for 3 hours.
Subchronic exposures were at 2 ppm (2.5 mg/m3) formaldehyde for 3 weeks or 3.3 ppm
(4.1 mg/m3) formaldehyde for 2 weeks. Seizures were induced by I.P. injection of
pentilenetetrazole (PTZ), and the incidence, severity, and course of induced seizures were
recorded. The PTZ injection induced seizures in 83, 88, and 91% of controls, with 16, 38, and
67% mortality in controls in the three trials. Mortality was highly variable in treatment groups as
well. The authors report that PTZ-induced seizures were decreased in incidence by acute
formaldehyde exposure in a dose-dependent fashion with only 33% of mice exposed to 14.8 ppm
formaldehyde experiencing seizures versus 91% in control mice (p < 0.05 at the highest dose
only). However, the methodology for observing and scoring seizures is unclear. Additionally,
there was high mortality and high variability of results for the three similarly treated control
groups. Therefore, it is difficult to assess data quality and interpret these findings.
Boja et al. (1985) exposed male Sprague-Dawley rats to air or to formaldehyde at 5, 10,
or 20 ppm for 3 hours on 2 consecutive days. On the second day, half the rats received the same
exposure as the previous day, while half the rats were switched (e.g., half those rats receiving air
the first day received formaldehyde the second day, and half those receiving formaldehyde the
first day received air the second day), for a total of four possible exposure combinations. During
the exposure period, activity levels were monitored by observation, once per minute for the first
hour and once every 5 minutes for the second hour. At the end of the second exposure session,
rats were sacrificed and brains removed for neurochemical analysis (see Section 4.2.1.6.1.5).
Behavioral results were described in detail only for control and 5 ppm groups. During
the first exposure session, activity levels of formaldehyde-exposed animals were significantly
decreased (approximately 50% of control levels). On the second day of exposure, those animals
previously exposed to formaldehyde exhibited partial recovery, those experiencing their first
formaldehyde exposure behaved similarly to those initially exposed on the first day, and those
animals exposed to formaldehyde for a second time had a greater decrease in activity than during
the first exposure (to approximately 30% of control levels). The authors stated that a similar
effect was seen in animals exposed at 10 ppm but that results at 20 ppm were not interpretable
(data were not presented). Overall, the decreased activity seen in this study is consistent with
effects seen by other authors.
This document is a draft for review purposes only and does not constitute Agency policy.
4-261 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Senichenkova (1991) exposed pregnant female rats to 0 or 0.5 mg/m3 (0 or 403 ppb)
formaldehyde on gestation days (GDs) 1-19 for 4 hours/day. Reproductive aspects of this study
will be discussed in the reproduction section; however, results from behavioral assessments
conducted on the neonates are discussed here. The author stated that maturation of motor
reflexes (assessed as surface righting and pendular reflex), open field behavior, and maze
learning ability were assessed. Detailed descriptions of procedures and results were not provided
for all assessments, but it was stated that motor reflex development did not differ in treated and
control animals. Open field motor activity assessments in 40-day-old (juvenile) offspring
revealed an increase in squares visited and an increased frequency of rearing on the second and
third days of testing, indicating a lack of habituation in the offspring of formaldehyde-treated
dams; similar levels of activity by both measures were found on the first test day. Counts of
defecation and urination were increased on all 3 days of testing. Increased exploratory behavior,
described as increased impulses, was also noted in a learning task (not otherwise described), but
the author stated that learning rate and ability of the formaldehyde-treated group was not
different from controls (no data were provided).
Mobility and neuromuscular excitability (not otherwise described) in offspring of female
white rats were also evaluated by Sheveleva (1971). Dams were exposed to 0.005 or
0.0005 mg/L (approximately 4,000 or 400 ppb, respectively) formaldehyde on GDs 1-19.
Spontaneous mobility (over 15 minutes) and neuromuscular excitability were evaluated in
offspring at 1 or 2 months of age (other results from this study are discussed under
developmental toxicity, above). At 1 month, spontaneous mobility was reduced at the low dose
in males (52% of control levels; p < 0.01) but not at the high dose, and at both doses in females
(to 64 and 56% of control levels at the mid dose and high dose, respectively; p < 0.02). At two
months, there was a dose-related increase in activity for both sexes, statistically significant (p <
0.001) in high-dose females only (391% of control levels).
4.2.1.6.1.3. Learning and memory. The effects of repeated formaldehyde exposures on learning
were investigated by Malek et al. (2003c), using a labyrinth swim maze. In this task, animals are
required to make a series of consecutive right or left turns to gain access to an escape platform
(Malek et al., 2003c). Adult male and female LEW. IK rats (15/sex/group) were exposed to 0,
0.1, 0.5, or 5.4 ppm (0, 0.123, 0.615, or 6.64 mg/m3) formaldehyde 2 hours/day for
10 consecutive days. Formaldehyde levels were checked eight times throughout the 2-hour
exposure periods. Mean formaldehyde levels of 0.1 ± 0.02, 0.5 ± 0.1, and 5.4 ± 0.65 ppm were
achieved. Body weight was measured on days 1, 5, and 10 of the experiment. Two days prior to
beginning the formaldehyde exposures, all subjects were given an acclimation trial in which they
This document is a draft for review purposes only and does not constitute Agency policy.
4-262 DRAFT—DO NOT CITE OR QUOTE
-------
1 were individually placed into the water-filled basin at the start position and allowed to navigate
2 to the escape platform with manual assistance to learn the correct route. Thereafter, the water
3 labyrinth test was run on each day of formaldehyde treatment, 2 hours after completion of each
4 daily exposure. Time taken to complete the test and errors made were recorded for each rat
5 (Table 4-56). An error was defined as swimming toward the start position or circling in the same
6 position without moving forward toward the escape platform. Rats were sacrificed at the end of
7 the experiment, and tissues were taken from the lung, heart, thymus, kidney, liver, pancreas,
8 skeletal muscle, and spleen. Tissues were fixed and prepared for histologic examination by light
9 microscope. No differences were noted in food consumption or body weight gain for either male
10 or female rats (Malek et al., 2003c). No treatment-related differences in organ pathology were
11 reported (with the possible exception of focal microatelectasis (lung collapse at the microscopic
12 level) seen in two to three animals in each formaldehyde-exposed group but not control animals).
13
14 Table 4-56. Effects of formaldehyde exposure on completion of the labyrinth
15 test by male and female LEW.1K rats
16
Male rats
Swimming time (sec)
Error rate (mean)
Day 1
Day 6
Day 10
Day 1
Day 6
Day 10
Control
105
12.2
6.33
7.4
0.5
0.0
0.1 ppma
100
12.9
6.07
7.7
5.0c
3.2°
0.5 ppm
97
16.7c
7.60b
7.6
4.4c
1.8°
5.4 ppm
105
25.7c
10.9°
7.7
5.0c
2.8°
Female rats
Swimming time (sec)
Error rate (mean)
Day 1
Day 6
Day 10
Day 1
Day 6
Day 10
Control
103
12.5
6.47
7.9
0
0.0
0.1 ppm
96
12.3
7.53
7.1
5.2c
3.0°
0.5 ppm
97
14.6c
7.60b
8.0
4.6c
2.2°
5.4 ppm
98
23.5c
9.73°
7.9
5.2c
2.6°
17
18 "Rats were exposed to formaldehyde for 2 hours/day, for 10 consecutive days.
19 ' Different from control, p < 0.05.
20 Different from control, p < 0.005.
21
22 Source: Malek et al. (2003c).
23
24
25 A clear learning curve was evident in control animals, with rats completing the task in
26 less time and with fewer errors over days (Table 4-56). Although the number of errors decreased
27 with increasing experience in all groups, error rates in formaldehyde-exposed rats at all doses
28 were consistently higher than those observed in controls, starting on day 3 (Figure 4-24). All
29 control animals performed without errors by day 6, whereas all treated animals were still making
30 two to three errors on day 10, the final day of testing. Time required (latency) to complete the
31 maze was also reduced over days. Although this measure of performance was not as sensitive as
This document is a draft for review purposes only and does not constitute Agency policy.
4-263 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
error rate, formaldehyde-induced deficits were still evident in the 0.5 and 5.4 ppm exposure
groups of both sexes.
m
E
w
o
m
1 23456789 10
Day of labyrinth test
Figure 4-24. Effects of formaldehyde exposure on the error rate of female
LEW.1K rats performing the water labyrinth learning test
Source: Drawn from data reported in Malek et al. (2003c).
Impaired performance on formaldehyde-treated subjects cannot be attributed to
alterations in swimming ability, since latencies to complete the maze were identical for 0 and
0.1 ppm groups, yet acquisition of the task was still impaired in the 0.1 ppm group based on
number of errors committed (see Figure 4-24). This study reports an adverse effect level of
0.1 ppm for increased error rate in the labyrinth water test, and all dose groups were equally
impaired across a broad range of exposures, 0.1-5.4 ppm. An independent estimate of
swimming speed was not included, so motor competency could not be directly evaluated.
However, comparable latency scores and error rates at the beginning of testing across all groups
and latency scores that track together over days suggest that impaired swimming ability does
account for the observed differences in latency, which are most likely reflective of the increased
control
This document is a draft for review purposes only and does not constitute Agency policy.
4-264 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
number of errors in treated animals (errors usually increase the distance traveled and thus time
required for completion of the trial).
Pitten et al. (2000) evaluated the effects of very brief formaldehyde exposures
(10 minutes) but prolonged duration (90 days) on previously learned performance in a land
version of the labyrinth maze. Adult male and female Wistar rats (13/group) were trained on the
task for 14 days, two trials/day. Animals were required to make a series of five left or right turns
from the entrance of the maze to retrieve a piece of cheese placed in the goal box at the opposite
end. Animals were guided by the experimenter through the maze during this acclimation phase
until all subjects were able to retrieve the food without aid. After animals were trained (but prior
to formaldehyde exposure), performance was assessed once daily for 11 days, and the latency to
complete the maze as well as the number of errors committed when traversing from the entrance
to the goal box was recorded. Animals were then assigned to one of three dose groups (five to
eight/sex/group) such that task performance was equivalent across groups prior to
commencement of formaldehyde exposures. Animals were exposed to 0 ppm, 2.6 ppm (0.25%
formaldehyde solution to yield 3.06 ± 0.77 mg/m3), or 4.6 ppm (0.70% formaldehyde solution to
yield 5.55 ± 1.27 mg/m3) formaldehyde 10 minutes/day, 7 days/week for 90 days. Animals were
assessed for performance in the maze every seventh day, at least 22 hours after the exposure on
the previous day. At the end of the 90-day exposure period, monitoring of maze performance
continued once every 10 days for an additional 30 days. All rats were sacrificed at the end of the
postexposure trials and tissue sections were prepared for histologic examination by light
microscopy, including liver, trachea, lung, kidney, heart, spleen, pancreas, testicle, and brain.
No treatment-related changes in food or water consumption weight gain or in histologic samples
obtained at the termination of the experiment were observed.
Pitten et al. (2000) reported that no gender differences existed as a function of
formaldehyde treatment; therefore, data were presented by combining sexes. Control rats
showed no change in error rate but a slight decrease in running time through the maze during the
course of the experiment. The formaldehyde-exposed groups began with a similar performance
level and error rate as controls, but their performance degraded over the course of formaldehyde
exposure. By the fourth week of exposure, increased numbers of errors were evident in both
exposed groups relative to controls. This trend reached statistical significance at the 12-week
time point, with a greater than twofold increase in number of errors (p < 0.05). Formaldehyde-
treated rats also tended to have increased run times through the maze (p = 0.04), but no
difference was seen by formaldehyde concentration. By 4 weeks after termination of exposure,
no statistical differences among the three groups were evident, but the tendency for the two
exposed groups to make more errors and have longer latencies remained. Since Pitten et al.
This document is a draft for review purposes only and does not constitute Agency policy.
4-265 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
(2000) tested animals after the task was acquired, these results indicate deficits in the retention of
a previously learned task.
Lu et al. (2008) evaluated the effects of formaldehyde on performance of mice in a
Morris water maze. Kunming mice (five males/group) were exposed to formaldehyde at 0.2, 1,
or 3 mg/m3 6 hours/day for 7 days (measured concentrations: 0.2 ± 0.01. 0.99 ± 0.04. and
3.03 ± 0.16 mg/m3). Mice were trained to locate a hidden platform in a large, circular tank
(106 cm diameter, 31 cm deep). Each animal received four training trials per day, beginning
30 minutes after the end of exposure. During training, latency to locate the platform was
recorded for each trial, with a maximum of 60 seconds, after which the animal was guided to the
platform. After the last day of training, an additional trial was conducted with the platform
removed (the probe trial); time spent in each maze quadrant was measured to determine the time
the animal spent searching for the platform in the correct area of the maze. Performance in the
water maze, measured as mean escape latency across the seven training trials, was significantly
impaired in the 3 mg/m3 group. No significant difference was seen at 1 mg/m3, although there
appeared to be an increased latency during the second day of testing. During the probe trial,
control animals spent significantly more time in the correct quadrant, but neither formaldehyde-
exposed group did so. Results of this study indicate deficits in learning and retention of the
Morris water maze following formaldehyde exposure, with greater effects seen in the higher dose
group.
Apfelbach and Reibenspies (1991) published a brief report of formaldehyde effects on
olfactory learning. Ferrets were exposed to 0.25 ppm (0.31 mg/m3) formaldehyde gas
continuously for 6 months. A Y-shaped maze was used to test odor detection, discrimination
between odors, and odor threshold. Ferrets were conditioned to distinguish ethyl acetate
(0.1 vol%) from clean air. Untreated ferrets achieved 75% success after an average of 110 trials.
However, formaldehyde-treated ferrets required on average 320 trials to reach a 75% success
rate. A 90% success rate was achieved by untreated ferrets after 420 trials. However, this level
of success was not reached in formaldehyde-treated ferrets.
The same researchers also tested olfactory function in formaldehyde-treated ferrets, as
summarized in Section 4.2.1.7 (Apfelbach et al., 1992). A decrease in olfactory discrimination
and a reduction in the percentage of olfactory cells in the olfactory epithelium were reported
after 3-12 months exposure to 0.25 or 0.5 ppm formaldehyde. Decreased olfactory sensitivity in
rats exposed to 0.25 or 0.5 ppm formaldehyde has also been reported by the same researchers
(Weiler and Apfelbach, 1992; Apfelbach and Weiler, 1991), and Weiler and Apfelbach (1992)
reported in an abstract that shifts in olfactory thresholds were greater when exposure was
initiated at PND 30 than at adult ages. Given the documented changes in olfactory thresholds,
This document is a draft for review purposes only and does not constitute Agency policy.
4-266 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
observed changes in olfactory learning would likely be confounded by the potential for
decreased olfactory function by formaldehyde exposures, and definitive conclusions regarding
formaldehyde effects specific to learning cannot be made based on these studies.
4.2.1.6.1.4. Neurosensitization. Sorg et al. (1996) studied the potential for formaldehyde
exposure to induce sensitization in the CNS, possibly through the limbic pathways in the brain.
The authors hypothesized that multiple chemical sensitivity (MCS) has an onset and progression
similar to CNS sensitization and may, therefore, be a similar process. These experiments were
conducted to test this hypothesis and to determine whether formaldehyde exposure could be used
as a model for MCS. Behavioral sensitization can be initiated by psychostimulants (e.g.,
cocaine) and manifest as increased locomotor activity upon subsequent challenge with the
stimulant.
Sorg et al. (1996) evaluated cross-sensitization of cocaine-induced increases in activity
from an initial formaldehyde exposure. Female Sprague-Dawley rats (eight to nine) were
exposed to 0 or 11 ppm (0 or 13.5 mg/m3) formaldehyde 1 hour/day for 7 days. Locomotor
activity was measured (by photocell) after saline injection (1 day postexposure) and after cocaine
injection (2 days postexposure). A similar protocol was conducted on days 36 and 37
postexposure. Motor activity levels following saline injection were similar for controls and
formaldehyde-treated rats. However, formaldehyde exposure initiated sensitization to cocaine as
evidenced by a greater increase in locomotor activity in mice treated with formaldehyde
followed by cocaine (p < 0.05) with an average count of crossed grids greater than 40,000 (2
hours) in treated animals compared with 25,000 (2 hours) in controls. The cross-sensitization
was transient, with no treatment effects on cocaine-induced activity either 29 or 37 days
postexposure. When examining individual data, the authors suggested that the formaldehyde-
treated groups in both cases have a cluster of high responders (HRs), suggesting some animals
may have been more sensitive. A second group of similarly treated female rats was pretested for
locomotor activity and divided into subgroups of HRs or low responders (LRs). They were then
given a panel of neurobehavioral tests: anxiety (elevated plus maze, day 11); memory (passive
avoidance training, day 12; passive avoidance test, day 19); and nociceptive test (day 20). Trunk
blood corticosterone levels were determined during stress on day 35 postexposure. No
significant treatment differences were found in the passive avoidance test, nociception, or
corticosterone levels (basal or stress induced). On the elevated plus maze, a two-way ANOVA
indicated no overall formaldehyde treatment effects, but the HR rats had higher open arm time
ratios (indicating greater anxiety) regardless of treatment. Within the treatment groups, the
difference in behavior between HR and LR subgroups was only significant for the formaldehyde-
This document is a draft for review purposes only and does not constitute Agency policy.
4-267 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
treated rats (p < 0.05). Authors suggested that cross-sensitization to cocaine-induced locomotor
activity was caused by enhanced dopamine transmission within the mesolimbic system (ventral
tegmental area to nucleus accumbens projection) following repeated formaldehyde exposure. A
critical role of the hypothalamic-pituitary-adrenal (HPA) axis has also been implicated in cross-
sensitization.
Sorg et al. (1998) and Sorg and Hochstatter (1999) further explored formaldehyde-
induced behavioral sensitization using the cocaine model. In contrast to the results following
exposure to 11 ppm (Sorg et al., 1996), rats exposed to only 1 ppm for 7 days showed no cross-
sensitization to cocaine injection. However, animals exposed to 1 ppm formaldehyde for
4 weeks exhibited increased cocaine-induced vertical activity (with no difference in horizontal
activity) for 4-6 weeks after cessation of exposure. Activity levels of formaldehyde-exposed rats
were approximately threefold those of control rats 3-4 days postexposure and still 1.5-fold
control levels at 4-6 weeks postexposure (p < 0.05).
Sorg et al. (2001) examined changes in corticosterone levels in rats with and without
formaldehyde treatment. Basal corticosterone levels in trunk blood were established in naive
male Sprague-Dawley rats taken directly from their home cage immediately prior to sacrifice. In
an acute trial, male rats were exposed to 0, 0.7, or 2.4 ppm (0, 0.86, or 2.96 mg/m3)
formaldehyde for either 20 or 60 minutes, and trunk blood was collected for corticosterone
analysis. Therefore, these rats were challenged with a new environment (the exposure chamber)
in the presence or absence of formaldehyde. In a separate trial, basal and challenged
corticosterone levels were measured after repeated exposure (1 hour/day, 5 days/week for 2 or
4 weeks). Basal corticosterone levels were measured in trunk blood immediately after removing
the animal from its home cage. Challenged corticosterone levels were measured after rats were
placed into the exposure chamber for a final 20-minute exposure. Body weight was measured at
the beginning of each week of exposure and was unchanged by formaldehyde treatment.
Corticosterone levels were increased over basal levels when rats were placed in the
exposure chamber for 20 minutes (Figure 4-25, panel a) but returned to basal levels after
60 minutes in the exposure chamber (not shown). This response may reflect the stress of the new
environment and acclimatization after 60 minutes in the chamber. Corticosterone levels were the
same in the presence or absence of formaldehyde, indicating no treatment effect.
This document is a draft for review purposes only and does not constitute Agency policy.
4-268 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Panel a
Panel b
Panel c
Corticosterone levels: with no
pre-exposure to formaldehyde
D Basal
* Challenge
jjJ
Control 0 07 2 4
Formaldehyde (ppm)
Corticosterone levels after 2 weeks of
repeated formaldehyde exposure
~ Basal
-halleng
Control 0 07 2
Formaldehyde (ppm)
Corticosterone levels after 4 weeks
of repeated formaldehyde exposure
~ Basal
Challeng
Control 0 07
Formaldehyde (ppm)
Figure 4-25. Basal and stress-induced trunk blood corticosterone levels in
male LEW.IK rats after formaldehyde inhalation exposures.
Note: Panel a: no pretreatment, corticosterone levels after 20-minute
formaldehyde exposure. Panels b and c show both basal and induced
corticosterone levels after a 2- or 4-week pretreatment to formaldehyde
1 hour/day. Challenge to induce corticosterone was a 20-minute reexposure at the
formaldehyde level tested.
Source: Sorg et al. (2001).
Control animals exhibited an increase in basal corticosterone after 2 weeks, which
returned to naive levels after 4 weeks (Figure 4-25, panels b and c). Formaldehyde-treated rats
demonstrated a comparable increase in basal corticosterone levels at 2 weeks, but these levels
did not return to naive levels at 4 weeks as seen with controls. Control and 0.7 ppm exposed rats
showed a similar response to challenge (the final 20-minute exposure). However, rats exposed to
2.4 ppm were hyperresponsive, with exaggerated corticosterone levels during this final exposure.
Differences in basal corticosterone levels after formaldehyde exposure and the
hyperresponsiveness seen in animals exposed at 2.4 ppm provide evidence of
formaldehyde-induced perturbations of the HPA axis. Authors suggested that elevated
corticosterone levels induced by repeated formaldehyde exposures may contribute to the cross-
sensitization to cocaine-induced motor activity.
Formaldehyde-induced changes in the HPA axis may contribute to behavioral effects of
formaldehyde exposure reported by Sorg et al. (2004) and Sorg and Hochstatter (1999). The
authors also reported an enhanced conditioning to odor in animals previously exposed to
This document is a draft for review purposes only and does not constitute Agency policy.
4-269 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
repeated formaldehyde. Male and female Sprague-Dawley rats (60-80 days of age) were
exposed at 1 ppm (1.23 mg/m3) formaldehyde 1 hour/day, 5 days/week for 4 weeks (Sorg and
Hochstatter, 1999). Two weeks after exposure ended, rats were trained to the conditioned fear
task. Rats were conditioned to a fear response by either odor only or odor associated with
footpad shock. Orange-oil extract was used as the odor conditioned stimulus (CS). One day
after conditioning, rats were reintroduced into the environment without an odor cue, and time
spent motionless in the freezing posture (freezing) was observed. On day 2 after conditioning,
rats were placed in a novel environment, and time spent in the freezing posture was evaluated in
the absence and then the presence of odor. This was repeated on day 12 after conditioning to
measure the loss of the freezing response to the conditioned odor.
Both treated and exposed rats showed similar responses on reintroduction into the
conditioning environment in the absence of an odor cue on day 1 (Sorg and Hochstatter, 1999).
As expected, rats conditioned with a footpad shock demonstrated greater time motionless than
odor-trained only rats, and there was no difference between control and formaldehyde-treated
rats. However, in the presence of odor on days 2 and 12, formaldehyde-exposed rats who were
conditioned with odor associated with foot shock spent significantly more time freezing than
odor-only trained rats (p < 0.05); control animals on those days showed no difference in time
freezing in the presence and absence of odor. The authors concluded that the formaldehyde-
treated rats had more difficulty than controls in extinguishing the fear response to the
conditioned odor, and speculated that an enhancement of the fear-conditioned response by
formaldehyde pretreatment supports the hypothesis that sensitization may include effects through
the limbic system of the brain.
In a second experiment, adult male and female Sprague-Dawley rats were exposed at 0 or
2 ppm (2.45 mg/m3) formaldehyde 1 hour/day, 5 days/week for 4 weeks (Sorg et al., 2004). Two
to 3 weeks after exposure ended, rats were trained to the conditioned fear task. Rats were given
a foot shock either associated with an odor (paired group) or unassociated with an odor (unpaired
group). Orange-oil extract was used as the odor CS. After training, freezing behavior was
assessed (1) in the same context in the absence of odor (1 day), (2) in a new context in the
presence and absence of the CS (5 consecutive days), and (3) in another novel context in the
presence and absence of the CS.
Formaldehyde-exposed male rats demonstrated increased conditioned fear response to an
odor CS (orange oil) paired with foot shock with no change in the degree of conditioning to the
context. For female rats, formaldehyde exposure did not affect the percent of time spent
freezing, either in the conditioning context or the novel context in the absence of the conditioned
odor. In contrast, male rats spent an increased time freezing in a novel context in the presence of
This document is a draft for review purposes only and does not constitute Agency policy.
4-270 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
odor, indicating a greater conditioned fear response to the olfactory cue (p < 0.05). This is in
agreement with the previous study where formaldehyde effects were seen in the presence of the
conditioning odor but not the environment (Sorg and Hochstatter, 1999). However, in this study
female rats did not exhibit a similar enhancement of fear conditioning to the olfactory CS.
The authors suggested that repeated exposure to low levels of formaldehyde acts as a
stressor in much the same way as inescapable foot shock, with resulting sensitized responses
within the olfactory/limbic pathways (Sorg et al., 2004). This interpretation is consistent with
work described above in which augmented basal corticosterone levels following repeated
formaldehyde exposures were demonstrated. However, while the fear conditioning in the present
study and cross-sensitization to cocaine described above (Sorg and Hochstatter, 1999) occurred
3-4 weeks after termination of exposure to formaldehyde, the duration of corticosterone
elevation induced by repeated exposure to formaldehyde has not been determined. It is possible
that augmentation of corticosterone levels following formaldehyde exposure results from direct
action of formaldehyde on the HPA axis. Experiments designed to compare HPA activation
following standard stressors (repeated inescapable foot shock or restraint stress), stress induced
by other irritants (chemicals with strong irritant odors but no CNS action), and repeated
formaldehyde exposures are necessary to dissociate primary from secondary action of
formaldehyde on CNS function in this paradigm. It is also possible that enhanced conditioning
to an odor stimulus results from formaldehyde-induced increases in airway irritation, rendering
the conditioned odor stimulus a more salient cue, producing a conditioned response that is not
extinguished as readily as in air-exposed controls. However, damage of the nasal mucosa and
lesions would be expected to be minimal at 1 ppm formaldehyde exposures and most likely
resolved 2 weeks after exposure was ended (see Section 4.2.1.2). Therefore, a more salient cue
for fear conditioning to odor due to physical irritation is not likely. Alternatively, formaldehyde
may act to up regulate olfactory activity, producing a stronger sense of odor during conditioning.
4.2.1.6.1. 5. Neuro chemistry and neuropathology. Several studies that were focused on general
toxicity or carcinogenicity of formaldehyde also assessed histopathology in exposed animals,
including pathological evaluation of the brain. In all cases, details of the pathological evaluation
were not provided. Reported results stated that no significant lesions were seen on unspecified
tissues (Appelman et al., 1988; Maronpot et al., 1986; Kerns et al., 1983) or that an increase in
relative brain weight (data not provided) was considered of no toxicological significance
(Woutersen et al., 1987). The absence of procedural information, or specific reported results,
limits the utility of this information.
This document is a draft for review purposes only and does not constitute Agency policy.
4-271 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Boja et al. (1985) measured changes in several neurotransmitters (norepinephrine,
dopamine, 5-hydroxytryptamine) and their major metabolites (3,4-dihydroxyphenylacetic acid
[DOPAC] and 5-hydroxyindoleacetic acid [5-HIAA]) following one or two 3-hour exposures to
formaldehyde at 0, 5, 10, or 20 ppm. Animals were sacrificed immediately following the second
exposure, and brains were immediately removed, frozen, and sectioned. Regions of interest were
analyzed by high-pressure liquid chromatography with electrochemical detection. Authors stated
that neurotransmitter concentrations were measured in multiple brain regions, but results were
reported only for the 5 ppm exposure and only for the hypothalamus. No change was seen in
concentrations of norepinephrine or 5-hydroxytryptamine for any exposure paradigm. For those
animals exposed twice to formaldehyde, there was a slight (statistically significant) increase in
dopamine and a larger (approximately fourfold) increase in 5-HIAA. DOPAC was increased
(approximately 30%) in animals receiving formaldehyde during the second exposure only.
Recent work by Hayashi et al. (2004) indicates that formaldehyde exposure increases the
activity of periglomerular (PG) cells in the main olfactory bulb. Tyrosine hydroxylase activity
was measured as a marker for activity of olfactory function. The authors surmised that
expression levels of this enzyme are useful markers since it has been reported that the protein is
up regulated after sensory stimulation and is down regulated by odor deprivation or when the
olfactory epithelium is removed (Cho et al., 1996; Stone et al., 1991; McLean and Shipley, 1988;
Baker et al., 1983). Eight-week-old female C3H/HeN mice were exposed at 0, 0.08, 0.4, or
2 ppm (0, 0.1, 0.49, or 2.45 mg/m3) formaldehyde 16 hours/day for 1 day or 12 weeks
(5 days/week). Formaldehyde exposure did not affect body weight. Mice were sacrificed
24 hours after exposure; the brains were removed, fixed, and prepared for sectioning. One side
of the olfactory bulb was sliced into 40 |iin-thick serial frontal sections and immuno-stained for
tyrosine hydroxylase activity. The number of tyrosine hydroxylase-positive PG cells was
determined by examining digital photomicrographs of three tissue sections, averaging the counts
from 10-15 glomeruli per section.
Neither the size of the olfactory bulb (rostrocaudal, dorsoventral, and mediolateral
lengths) nor the total number of PG cells was changed by formaldehyde exposure. The number
of tyrosine hydroxylase-positive PG cells per glomerulus was unchanged by a single
formaldehyde exposure but increased after 12 weeks of repeated exposures. The increases were
similar among treatment groups: 5.54 ± 0.31 at 0.80 ppm, 5.18 ± 0.60 at 0.4 ppm, and 6.0 ± 0.83
at 2 ppm or 196, 167, and 196% of controls, respectively. As an indicator of activity, it is not
unexpected that the enzyme was up regulated after repeated exposure to an odorous compound.
Hayashi et al. (2004) hypothesize that the increased tyrosine hydroxylase activity is an indication
of increased sensitivity and, therefore, may be a model for MCS. However, it is unknown if the
This document is a draft for review purposes only and does not constitute Agency policy.
4-272 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
increase in enzyme activity after repeated exposures is transient or could result in sensitization.
Tyrosine hydroxylase is the first enzyme in the dopamine synthetic pathway, but the role of
dopamine in PG cells is not known. Further research would be needed to understand the
potential for formaldehyde to act as a sensitizing agent in this model.
In an abstract, Kakeyama et al. (2004) outline the results of experiments to address the
effects of subchronic exposure to low levels of formaldehyde on changes in neurotransmitter-
related mRNA expressions in mice forebrains. An unstated number of female C3H/He mice
were exposed 16 hours/day, 5 days/week to 400 ppb (0.49 mg/m3) formaldehyde for 12 weeks.
The authors used RT-PCR methodologies to quantify mRNA encoding for the glutamate receptor
subunits GluRl and GluR2, the dopamine receptor Dl, and the serotonin receptor 5-HT1A in the
neocortex, hippocampus, amygdala, and hypothalamus. Raised levels of mRNA expression were
observed for GluRl in the neocortex and hippocampus; GluRl, GluR2, and the dopamine
receptor Dl in the amygdala; and the serotonin receptor 5-HT1A in the hypothalamus. Reduced
mRNA expression was observed for GluR2 in the hippocampus and neocortex. When other
mice were subjected to a radiofrequency-induced lesion of the hippocampus then exposed to
formaldehyde for 12 weeks as before, the altered expression of GluRl and GluR2 in the
neocortex was abolished. However, the increment of mRNA expression of 5-HT1A in the
hypothalamus was further enhanced. In demonstrating that formaldehyde affects neocortical
GluRl and GluR2 mRNA expressions through a hippocampal function, Kakeyama et al. (2004)
concluded that subchronic exposure to low concentrations of formaldehyde can affect neural
transmission in the forebrain.
Fujimaki et al. (2004b) examined the effects of formaldehyde on NGF in the brain and
hippocampus. Ten female C3H/HeN mice/group were exposed to 0, 80, 400, or 2,000 ppb (0,
0.1, 0.5, or 2.45 mg/m3) formaldehyde 16 hours/day, 5 days/week for 12 weeks. Some groups of
mice received the same treatment after I.P. injection of 10 |ig of OVA and 2 mg alum prior to the
commencement of formaldehyde exposure. For this subgroup, booster injections of OVA were
administered on days 21, 42, 63, and 77 during the formaldehyde exposure regimen.
Quantitative measures of NGF and BDNF in homogenates of whole brain and hippocampus were
obtained by ELISA and mRNA determination. The amount of NGF protein in whole brains
remained unchanged in the non-immunized mice. However, brain NGF levels were significantly
increased in OVA-immunized mice exposed to 80 and 400 ppb (but not 2,000 ppb)
formaldehyde (Figure 4-26). This result was confirmed by parallel increases in the
concentrations of hippocampal NGF mRNA that were produced in immunized mice exposed to
formaldehyde at the same concentrations. However, there were no comparable increases in the
amounts of brain-derived neurotrophic factor in either immunized or non-immunized mice. In
This document is a draft for review purposes only and does not constitute Agency policy.
4-273 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
discussing the mechanisms potentially associated with their results, Fujimaki et al. (2004b)
considered it likely that low-level exposure to formaldehyde could enhance NGF production
through the stimulation of the HPA axis together with immunization.
~ Non-immunized mice
¦ Ova-immunized mice
35-
30-
-|25-
>20-
§ 15"
z 10-
5 -
"
0
0 80 400 2,000
Formaldehyde (ppb)
Figure 4-26. NGF production in the brains of formaldehyde-exposed mice.
Note: Female C3H mice were exposed to formaldehyde 16 hours/day,
5 days/week for 12 weeks. NGF in homogenates of whole brain and
hippocampus were measured by ELISA. Values are means ± SEM (n = 5-6).
* =p < 0.05 and ** =p < 0.01 versus control mice, as calculated by the authors.
Source: Redrawn from Fujimaki et al. (2004b).
The enhancement of NGF in the brains of immunized mice exposed to formaldehyde
gave rise to the suggestion that NGF may promote the survival of hippocampal neurons when
challenged with formaldehyde. To examine whether or not apoptosis plays a role in this process,
Tsukahara et al. (2006) measured the effects of formaldehyde on apoptotic mechanisms
regulating the survival and death of cells and on N-methyl-D-aspartate (NMDA) receptors.
Female C3H/HeN mice (13/group) were exposed to 0 or 400 (393 ± 34) ppb (0 or 490.8 |ig/m3)
formaldehyde 16 hours/day, 5 days/week for 12 weeks. Seven control and formaldehyde-treated
mice were immunized with 10 |ig OVA plus 2 mg aluminum hydroxide prior to exposure.
Subsequently, these mice received OVA via aerosol as a booster during weeks 3, 6, 9, and 11.
Hippocampi were dissected from all animals 1 day after the final exposure and homogenized in
hypotonic buffer. The 12,000 rpm supernatants were analyzed by Western blotting for the
presence of the proteins BC1-2 (which inhibits apoptosis) and Bax (which opposes BC1-2 action
and promotes apoptosis) and the NMDA receptor subtypes 2A and 2B (NR2A and NR2B).
This document is a draft for review purposes only and does not constitute Agency policy.
4-274 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Immunohistochemical analysis was also carried out for the presence of active caspase-3, an
apoptosis marker.
The levels of NR2A and NR2B were unaffected by exposure to formaldehyde in either
immunized or non-immunized mice. Likewise, the number of caspase-3 immunoreactive cells
did not change as a result of formaldehyde exposure. However, when measured amounts of
Bcl-2 and Bax were normalized to the amount of P-tubulin, the ratio Bcl-2/Bax was significantly
increased in immunized mice exposed to formaldehyde. Non-immunized mice did not show this
apparently compound-related response. Consistent with the concept that the proportions of Bcl-2
and Bax are critical for the regulation of cell survival and death, the authors interpreted their data
as an indication that changes to the ratio of Bcl-2/Bax expressions might be an important
adaptive response to the effects of formaldehyde, such that the antiapoptotic changes might
contribute to the protection of hippocampal neurons from the pernicious effects of formaldehyde
exposure itself.
The same research group used the immunized mouse model to determine whether
formaldehyde exposure affected mRNA expression of genes related to synaptic plasticity
(Ahmed et al., 2007). Ten female C3H/HeN mice were exposed to 0 or 400 ppb formaldehyde
16 hours/day, 5 days/week for 12 weeks. All mice were immunized with 10 |ig OVA plus 2 mg
aluminum hydroxide prior to initial formaldehyde exposure then treated in weeks 3, 6, 9, and 11
with aerosolized OVA as a booster. Five treated and control animals were I.P. injected with 1
mg/kg MK-801, a noncompetitive NMDA receptor agonist before the last formaldehyde
exposure. At term, hippocampi were dissected and frozen at -80°C until processing. At that
point, total mRNA was extracted and first strand cDNA was synthesized by using reverse
transcriptase. Expression levels of various proteins/receptors, including NMDA NR2A and
NR2B receptor subunits, dopamine D1 and D2 receptors, cyclic AMP responsive element-
binding proteins (CREB-1 and CREB-2), and the transcription factors FosB and AFosB were
determined by using the PCR. The expression level of each mRNA species was expressed
relative to the sample's content of 18S rRNA. The total protein lysate was also assayed for
pCREB by Western blotting.
In the first of a sequence of histograms, Ahmed et al. (2007) demonstrated a significant
increase in mRNA expression of NR2A as a result of formaldehyde exposure. However, this
effect was abolished in animals treated with MK-801. A similar trend in the mRNA expression
of NR2B in response to formaldehyde exposure did not achieve statistical significance. MK-801
treatment significantly reduced receptor in mRNA expression in the presence of formaldehyde.
The authors provided data showing an increased expression of dopamine D1 and D2 receptor
mRNA response to formaldehyde, in both cases abolished by treatment with MK-801. The
This document is a draft for review purposes only and does not constitute Agency policy.
4-275 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
expression of CREB-1 mRNA also conformed to the pattern of being increased as a result of
formaldehyde exposure but abolished by MK-801. However, the expression of CREB-2 and
FosB/AFosB was unaffected by formaldehyde. When normalized to the amount of P-tubulin,
there were no significant effects of formaldehyde exposure and MK-801 treatment on the protein
levels of pCREB. Finally, there was no significant difference in the expression of transient
receptor potential vanilloid receptor (TRPV1) between control and formaldehyde-exposed mice,
and MK-801 itself did not significantly alter the mRNA level of TRPV1. In seeking to explain
their results, the authors speculated that low-level exposure of immunized mice to formaldehyde
had an effect on hippocampal synaptic plasticity at the mRNA level, as evidenced by the
enhancement of mRNA for NR2A, the dopamine D1 and D2 receptors, and CREB1, with up
regulation compensating for the sustained levels of enhanced protein expression under low-level
formaldehyde exposure. The interpretation of these changes in NR2A mRNA, in the context of
the results of Tsukahara et al. (2006), showing no change in NR2A and NR2B protein
expression, was not discussed.
4.2.1.6.1.6. Neurosenesis. Two papers have examined the effects of subacute exposure to
formaldehyde on the overall size (volume) of discrete cellular areas of the hippocampus in
neonatal rats. The researchers also used an optical fractionator counting method to derive a
plausible estimate of cell number. Asian et al. (2006) studied the effects of formaldehyde
exposure on the number and volume of granular cells in the hippocampal dentate gyrus.
Sarsilmaz et al. (2007) examined the impact of postnatal formaldehyde exposure on brain
hemisphere volume and on the size and cell number of pyramidal cells in the cornu ammonis
region of the hippocampus. The in-life phase was the same in each study, featuring the exposure
of 10 neonatal male Wistar rats/group to 0, 6, and 12 ppm (0, 7.36, and 14.7 mg/m3)
formaldehyde 6 hours/day, 5 days/week for 30 days. Five rats/group were sacrificed at that point
(PND 30), while the rest were maintained without further treatment until PND 90.
For both pyramidal and granular areas, a much lower number of cells was seen on
PND 90 versus PND 30 (p < 0.001). This response was evident irrespective of the amount of
exposure to formaldehyde and is consistent with normal brain development. Compound-specific
effects of formaldehyde on the volume and number of granular and pyramidal cells varied by
dose and over the two time points. There was a small increase in the volume of the granular cell
layer of the dentate gyrus in rats sacrificed on PND 30 (p< 0.001) in response to increasing
formaldehyde concentration (Asian et al., 2006), with no significant change in neuron number;
the increased volume (now accompanied by an increase in neuron number) was still evident at
the low-exposure level on PND 90 (p< 0.01) but not at the high dose. Brain hemisphere volume
This document is a draft for review purposes only and does not constitute Agency policy.
4-276 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
was decreased at both concentrations on PND 30 (p < 0.01) but was increased at both
concentrations (p < 0.01, with a larger magnitude of effect at 6 ppm) on PND 90 (Sarsilmaz et
al., 2007). In the hippocampal cornu ammonis region, the volume of the pyramidal cell layer on
PND 30 was increased in low-dose animals (p < 0.001) but decreased in high dose animals (p <
0.001) as compared with control values; neither group was significantly different from controls
on PND 90. There was a dose-related decrease in total neuron number in the cornu ammonis on
PND 30 (p < 0.01 for both doses); on PND 90 the decrease in neuron number remained
statistically significant in both treatment groups (p < 0.01), but there was no longer any
difference in the magnitude of the effect between doses (Sarsilmaz et al., 2007).
In a third study from the same laboratory, Songur et al. (2008) used the same exposure
paradigm to evaluate changes in oxidant and antioxidant systems in the cerebellum of perinatally
exposed rats. Exposure was carried out as in Asian et al. (2006) and Sarsilmaz et al. (2007),
described above. On PND 30 or 90, cerebellums from seven male rats per group were evaluated
for levels of malondialdehyde (MDA), NO, superoxide dismutase activity (SOD), and
glutathione peroxidase (GPX) activity. Dose-related increases in NO (approximately 20-80%),
MDA (100-160%), and GPX (25-60%) and dose-related decreases in SOD (20-30%) were seen
on PND 30. In general, the magnitude of change from control levels was maintained on PND 90,
with the exception of MDA levels in 6 ppm animals, which appeared to approach control levels
at 90 days. The authors stated that these findings indicated that formaldehyde exposure may
cause neurotoxicity via the production of oxidative damage in the brain. Persistence of the effect
to the 90-day time point (30 days after cessation of exposure) supports the possibility that
formaldehyde may cause long-lasting or permanent changes in the brain following early life
exposure. These results are consistent with the earlier studies by Asian et al. (2006) and
Sarsilmaz et al. (2007), finding permanent changes in brain structure (although in a different
brain region) following early life exposure.
4.2.1.6.1.7. Summary of formaldehyde effects on neurobehavioral and neuroyatholoeical
measures. following exposure via inhalation. As has been demonstrated in mice (Wood and
Coleman, 1995), it is possible that rats experience respiratory tract irritation during low-level
formaldehyde exposure. Perturbations in nervous system function reported with formaldehyde
exposure include reductions in motor activity, lack of habituation, impairment in acquisition of a
new learning task, deficits in retention of a previously learned task, increases in corticosterone
levels, sensitization to cocaine-induced locomotor activity, and enhanced fear conditioning using
an olfactory CS (Table 4-57). Many of these effects were observed at exposure levels at or
below 1 ppm, and some persisted days to weeks after termination of exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
4-277 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Malek et al. (2004, 2003a, b) detected behavioral changes in rats and mice tested 2 to
24 hours postexposure. The mechanism of these behavioral changes is unknown, and available
data do not allow dissociation of direct effects on the nervous system and behavioral responses to
the irritant effects of formaldehyde (control experiments [e.g., using a different aversive odor
with or without irritant properties] were not included). Given that behavioral changes were
observed hours to days after cessation exposure (i.e., beyond the time required for formaldehyde-
induced irritation to subside), it is unlikely that these behavioral changes were caused by
formaldehyde-induced irritation. Similarly, although it is possible that systemic effects of
formaldehyde exposure might cause reduced motor activity during or immediately following
exposure, it is unlikely that these effects can account for the differences in responses of male rats
24 hours after exposure (Malek et al., 2003a). Furthermore, a follow-up study demonstrated
reduced motor activity in animals 2 hours after a 2-hour exposure to much lower levels of
formaldehyde (0.1 ppm), which fall well below the levels identified by Wood and Coleman
(1995) as the AC50 for formaldehyde in mice (Malek et al., 2003b).
Two studies reported significant reductions in learning or retention following brief
periods of repeated exposure to low levels of formaldehyde (Malek et al., 2003c; Pitten et al.,
2000) (Table 4-57). Malek et al. (2003c) reported an increased number of errors in acquiring a
water maze task; testing took place daily 2 hours after termination of a 2-hour exposure. The
work of Pitten et al. (2000) revealed that brief exposures over many weeks led to increases in
errors performing a previously learned task and that the magnitude of the effect increased over
the course of the exposure period. Testing occurred remote from the time of exposure (22 hours
after the previous exposure), and the deficits appeared to persist for several weeks after exposure
terminated, minimizing the possibility that these effects were related to irritant properties of
formaldehyde. Although the exposure levels were moderately high (2.6-4.6 ppm) and continued
over several months, the duration of a single exposure event was very brief (10 minutes).
Sorg and Hochstatter (1999) and Sorg et al. (2004, 2001) suggest that behavioral
sensitization associated with low-level formaldehyde exposure was linked to alterations in HPA
control of corticosterone. Cross-sensitization to the locomotor activity-enhancing properties of
cocaine and changes in response to a conditioned fear paradigm were observed in animals
exposed several weeks earlier to repeated low-level formaldehyde. Direct activation of
mesolimbic dopamine pathways or activation of conditioned fear response in the amygdala by
formaldehyde could underlie these behavioral effects; these observations were also seen by study
authors as consistent with a formaldehyde-induced stress response.
This document is a draft for review purposes only and does not constitute Agency policy.
4-278 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-57. Summary of neurological and neurobehavioral studies of inhaled
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Irritant detection threshold
Male
Swiss
mice
8
0, 1, 1.8, 3,
5.6, or 10 ppm
60-second
exposure
episode to
determine
irritant
response
Sensitivity of mice to acute formaldehyde levels
determines the median concentration at which
50% of exposures were terminated by the subject
(AC50) decreased upon repeat exposure. AC50 =
3.63 for first series, AC50 = 1.88 ppm for second
series.
Time to exposure termination decreased with
increasing formaldehyde concentration.
Time to termination was decreased in repeat
exposures.
NAb
Wood and
Coleman
(1995)
Motor activity and habituation
Male
and
female
LEW. IK
rats
10/sex
0, 1,2.5, or 5
ppm for 2
hours
Reduced horizontal activity: Number of crossed
quadrants reduced 2 hours after exposure at all
doses for males and females.
Reduced habituation: Exposed rats exhibited
greater activity than controls when reintroduced
into the testing environment 24 hours later
(males and females, all doses).
LOAEL=
1 ppm
2 hours
Malek et al.
(2003a)
Male
and
female
LEW. IK
rats
10/sex
0,0.1,0.5, or 5
ppm for 2
hours
Reduced horizontal activity: Number of crossed
quadrants reduced 2 hours after exposure at all
doses for males and females.
LOAEL=
0.1 ppm
2 hours
in males
Malek et al.
(2003b)
Male
AB mice
5-7/sex
0, 1.1, 2.3, or
5.2 ppm for 2
hours
Reduced horizontal activity: Number of crossed
quadrants reduced 2 hours after exposure at all
doses.
LOAEL=
1.1 ppm
2 hours
Malek et al.
(2004)
Balb/c
mice
6
0, 1.8,3.2,4.5,
6.4, 9.7, or
14.8 ppm for 3
hours
Reduced horizontal and vertical activity: Dose-
dependent decreases in crossed quadrants and
rearing.
Significant for males at 1.8 ppm and greater (p <
0.01).
Significant for females at 6.4 ppm or greater (p <
0.01).
LOAEL=
1.8 ppm
3 hours
in males
Usanmaz et
al. (2002)
Balb/c
mice
6
3.3 ppm for 2
weeks or
2 ppm for 3
weeks
3 hours/day,
5 days/week
Reduced horizontal and vertical activity
decreases in crossed quadrants and rearing.
3.3 ppm (2 weeks) and 2 ppm (1 week) (p <
0.01,/? <0.05).
LOAEL=
2 ppm
3 weeks
Usanmaz et
al. (2002)
Sprague-
Dawley
rats
8
0,5, 10, or 20
ppm;
3 hours/day for
1 or 2 days
Reduced activity levels on both days. Decreases
seen at 5 and 10 ppm; data reported only for 5
ppm group
LOAEL=
5 ppm,
3 hours
Boja et al.
(1985)
This document is a draft for review purposes only and does not constitute Agency policy.
4-279 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-57. Summary of neurological and neurobehavioral studies of inhaled
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Rats
0 or 0.5 mg/m3
(0.4 ppm) on
GDs 1-19, 4
hours/day
Increased motor activity on the 2nd and 3rd days
of testing (reduced habituation) in offspring
exposed in utero. Increased number of squares
entered (p < 0.01) and frequency of rearing (p <
0.05).
LOAEL=
0.4 ppm,
gestational
Senichenkova
(1991)
Rats
15
0, 0.005, or
0.0005 mg/L
(approximately
4 or 0.4 ppm),
GDs 1-19
Changes in motor activity at one and two months
in offspring exposed in utero. Decreased
spontaneous mobility at 1 month in both sexes,
increased activity at 2 months in both sexes.
LOAEL=
0.4 ppm,
gestational
Sheveleva
(1971)
Learning and memory
Adult
male and
female
LEW. IK
rats
15/sex
0, 0.1, 0.5, or
5.4 ppm
2 hours for 10
consecutive
days
Impairment in acquisition of a new task. Male
and female rats at all formaldehyde exposures
had significantly more errors in completing a
water labyrinth (p < 0.01).
Male and female rats had longer times to
completion of the maze at 0.5 and 5.4 ppm (p <
0.05, p < 0.01).
LOAEL=
0.1 ppm
2 hours/
10 days
Malek et al.
(2003c)
Adult
male and
female
Wistar
rats
5-8/
sex
0, 2.6, 4.6 ppm
10
minutes/day
for
90 consecutive
days
Deficit in the retention of a learned task. Male
and female rats committed significantly more
errors (p < 0.05) and took more time to complete
the land maze in across the course of the
experiment (p < 0.04).
LOAEL=
2.6 ppm
10 minutes/
90 days
Pitten et al.
(2002)
Ferrets
0.25 ppm
Impairment in acquisition of a new task.
Exposed ferrets only achieved a 75% success
rate in training to discriminate odors in a Y-maze
versus 90% success rate in controls.
Note: The results are confounded with other
effects on the olfactory epithelium. The same
researchers also reported a decrease in olfactory
sensitivity and a reduction in percentage of
olfactory cells in similarly treated animals.
None
established
Apfelbach
and
Reibenspies
(1991)
(abstract
only)
Male
juvenile
and
adult
rats
5/group
0.25, 0.5 ppm
Decreases in olfactory thresholds, in juvenile
animals but not in adults (p < 0.002).
LOAEL=
250 ppm in
juveniles
Weiler and
Apfelbach
(1992)
(abstract
only)
This document is a draft for review purposes only and does not constitute Agency policy.
4-280 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-57. Summary of neurological and neurobehavioral studies of inhaled
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Neurosensitization endpoints
Female
Sprague-
Dawley
rats
8-9
0 or 11 ppm
1 hour/day, 7
days
Increase in cocaine-induced activity: Increased
quadrants crossed after cocaine injection 1 and 2
days after exposure (p = 0.05 andp < 0.04,
respectively).
No change in corticosterone levels 28 days
postexposure.
No change in nociceptive or passive avoidance
test or plus-maze (21, 20, and 13 days
postexposure, respectively) (21 days).
LOAEL=
11 ppm/
7 days
(unbounded)
Sorg et al.
(1996)
Female
and male
Sprague-
Dawley
rats
Various
up to
24/group
11 ppm, 1
hour/day,
7 days
1 ppm, 1
hour/day,
7 days
1 ppm, 1
hour/day,
5 days/week,
4 weeks
Increase in cocaine-induced activity: Increase in
rearing after
cocaine injection 1 day after exposure but not 4-
6 weeks after exposure; 11 ppm for 7 days or 1
ppm for 4 weeks.
No change in rats exposed at 1 ppm for 1 week.
LOAEL=
1 ppm
4 weeks
NOAEL =
1 ppm
7 days
Sorg and
Hochstatter
(1999)
Female
and male
Sprague-
Dawley
rats
Various
up to
24/group
1 ppm, 2
hours/day,
5 days/week,
4 weeks
Increased conditioned fear response in
formaldehyde-treated, foot-shock-conditioned
rats, twofold (p < 0.05).
LOAEL=
1 ppm
4 weeks
Sorg and
Hochstatter
(1999)
Male
Sprague-
Dawley
rats
4-9 or
16
0,0.7, or 2.4
ppm for 20 or
60 minutes
0,0.7, or 2.4
ppm
1 hour/day,
5 days/week
for 2 or 4
weeks
No change in corticosterone in acute (20- and
60-minute) exposures.
Increase in basal corticosterone: 0.7 ppm for 2
or 4 weeks.
Hyperresponsive corticosterone response to
environment:
2.4 ppm for 2 or 4 weeks.
LOAEL=
0.7 ppm/
2 weeks
NOAEL =
0.7 ppm/
20 minutes
Sorg et al.
(2001)
Female
and male
Sprague-
Dawley
rats
4-9 or
16
0 or 2 ppm
1 hour/day,
5 days/week
for 4 weeks
Increased conditioned fear response to an
olfactory cue in formaldehyde-treated, foot-
shock-conditioned male rats. Measured as
increased time freezing when presented with a
novel environment (p < 0.05).
No e ffect in female rats.
LOAEL=
2 ppm/
4 weeks
NOAEL =
2 ppm/
4 weeks
Sorg et al.
(2004)
This document is a draft for review purposes only and does not constitute Agency policy.
4-281 DRAFT—DO NOT CITE OR QUOTE
-------
Table 4-57. Summary of neurological and neurobehavioral studies of inhaled
formaldehyde in experimental animals
Species
No./
group
Treatment"
Observations
LOAEL/
NOAEL
Reference
Neurochemistty and neuropathology
8-week-
old
female
C3H/
HeN
mice
5
0, 0.08, 0.4, or
1 ppm
16 hours/day,
5 days/week
for 1 day or
12 weeks
No change in size of main olfactory bulb by
several measures.
No change in numbers of PG cells.
No change in tyrosine hydroxylase
immunopositive PG cells after 1 day.
Increase in tyrosine hydroxylase-immunopositive
PG cells after 12 weeks to 196, 167, and 196%
of controls at 0.08, 0.40, and 1 ppm,
respectively.
LOAEL=
0.08 ppm/
12 weeks
Hayashi et al.
(2004)
Adult
male
Sprague-
Dawley
rats
8
0, 5, 10, 20
ppm
3 hours/day, 1
or 2 days
No change in norepinephrine or 5-
hydroxytryptamine in hypothalamus.
Increase in DOPAC in hypothalamus after one
exposure.
Increase in dopamine and 5-HIAA in
hypothalamus after two exposures.
LOAEL=
5 ppm/
3 hours
Boja et al.
(1985)
Neurogenesis
Neonatal
Wistar
rats
5
0, 6, or 12
ppm,
6 hours/day,
5 days/week
for 30 days
Changes in volume of granular cell layer of the
dentate gyrus in the hippocampus at postnatal
days 30 and 90 (p < 0.001)
LOAEL=
6 ppm/
30 days
Asian et al.
(2006)
Neonatal
Wistar
rats
5
0, 6, or 12
ppm,
6 hours/day,
5 days/week
for 30 days
Decreases in brain hemisphere volume at PND
30 (p< 0.01)
Changes in volume and cell numbers in the CA
region of the hippocampus on PND 30 (p <
0.01)
LOAEL=
6 ppm/
30 days
Sarsilmaz et
al. (2007)
Neonatal
Wistar
rats
7
0, 6, or 12
ppm,
6 hours/day,
5 days/week
for 30 days
Changes in oxidant and antioxidant systems in
cerebellum on PNDs 30 and 90 (p = 0.017-
0.001). Increases in MDA, NO, and GSH-Px
and decreases in SOD at both time points.
LOAEL=
6 ppm/
30 days
Songur et al.
(2008)
Treatment is given as the formaldehyde concentration in air (ppm) with the length of exposure each day and the
duration of treatment in days, as available.
bNA = not available.
This document is a draft for review purposes only and does not constitute Agency policy.
4-282 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Limited data regarding possible neurochemical changes in the brains of formaldehyde-
exposed, immunized mice (Ahmed et al., 2007; Fujimaki et al., 2004b; Hayashi et al., 2004;
Kakeyama et al., 2004) and rats (Boja et al., 1985) provided conflicting information, and the
implications of these data regarding possible formaldehyde neurotoxicity are difficult to
determine.
In developmental exposure paradigms, changes in brain structure (Sarsilmaz et al., 2007;
Asian et al., 2006), brain chemistry (Songur et al., 2008), and motor activity (Senichenkova,
1991; Sheveleva, 1971) were seen following neonatal or in utero exposure to formaldehyde. In
addition, Weiler and Apfelbach (1992) found juvenile animals to be more sensitive to
formaldehyde-induced changes in olfactory thresholds when compared with adult animals.
These studies raise concern about possible long-lasting neurological effects of early exposure to
formaldehyde. It is important to note, however, that exposure levels in these studies were higher
(250-6,000 ppb) than those producing the behavioral effects in adults described above.
Overall, available data provide substantial evidence of behavioral effects, including
motor activity changes and changes in learning and retention, following repeated exposure to
relatively low levels of formaldehyde. These effects were seen in multiple laboratories, in
studies conducted by different authors, and using different behavioral paradigms. These
conclusions are also supported by more limited data, indicating possible developmental effects
on the nervous system, including changes in brain structure and in the behavior of offspring; the
developmental findings are less robust since they were seen only in individual laboratories and
occurred following exposure to higher concentrations of formaldehyde. Studies evaluating
developmental neurotoxicity at lower doses, comparable to those used in the adult studies, were
not available. None of the available data provide sufficient information to allow a determination
of the mechanism for these behavioral changes, although it is unlikely that they are attributable
to the irritant properties of formaldehyde. The data regarding behavioral sensitization provide
some support for a stress-related mechanism for those findings, but the applicability of this
mechanism to the behavioral changes seen in the other studies, including the learning deficits,
has not been evaluated.
4.2.1.6.2. Oral exposure. Available data regarding neurotoxic effects of formaldehyde exposure
following oral exposure are very limited. Several chronic or subchronic oral toxicity studies
evaluated changes in brain weight or histopathology in rats or dogs following repeated oral
exposures to formaldehyde at doses as high as 300 mg/kg-day, administered in drinking water
(Til et al., 1989, 1988; Tobe et al., 1989; Johannsen et al., 1986). Although data were not
This document is a draft for review purposes only and does not constitute Agency policy.
4-283 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
presented in the publications, all stated that no changes in brain weight or pathology were seen in
the standard evaluations performed in these studies.
Two studies evaluated changes in behavior following exposure to formaldehyde in
drinking water (Venkatakrishna-Bhatt et al., 1997; Venkatakrishna-Bhatt and Panchal, 1992).
Venkatakrishna-Bhatt and Panchal (1992) evaluated changes in performance on a conditioned
avoidance task in adult male albino rats (five/group). Animals were exposed to formaldehyde in
drinking water (10 mg/mL) or by I.P. injection (10 mg/kg) for 60 days. Although it was stated
that water consumption was recorded, the data were not presented, and thus actual exposure
levels cannot be documented. Prior to the initiation of exposure, rats were trained on the
conditioned avoidance task (climbing a wooden pole in response to a warning buzzer, thus
avoiding electric shock from a floor grid). Rats were trained to a predetermined performance
criterion (not described); animals not achieving the criterion were removed from study. Training
and testing conditions (e.g., retest interval and duration of sessions) were not well described.
Data were presented as percent response in behavioral performance (apparently separately for the
escape or avoidance aspects of the task) or percentage decrease in response. No control data
were presented, and pretreatment performance was not described. Figures presented
performance at 10-day intervals, starting with day 0, with each data point stated to represent the
mean for five experimental sets; again, the interval between experimental sets and the number of
trials per set was not specified. Although the authors concluded that a deficit in performance
was demonstrated, the data as presented were difficult to interpret and the conclusion could not
be verified based on the data as presented.
Venkatakrishna Bhatt and Panchal (1992) examined changes in performance on a
conditioned avoidance response, presumably using a procedure similar to the one described
above. Albino rats (sex not specified, five/group) were exposed to formaldehyde in drinking
water at 0.2 or 0.5 mg/mL for 90 days. As described above, rats were trained in performing the
task prior to the start of exposure. Venkatakrishna-Bhatt Bhatr and Panchal (1992) stated that
there was a dose-related deterioration of performance, but no data were presented to support
these conclusions.
In summary, available data are insufficient to conduct a reliable assessment of neurotoxic
effects of formaldehyde following oral exposure. Limited data suggest a lack of overt
neuropathological changes at doses up to 300 mg/kg-day (Til et al., 1989, 1988; Tobe et al.,
1989; Johannsen et al., 1986), but detailed information regarding the types of neuropathological
evaluations performed in those studies is not available, and thus no firm conclusions can be
drawn regarding the potential for neuropathological effects. The two available studies evaluating
This document is a draft for review purposes only and does not constitute Agency policy.
4-284 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
behavioral changes are not considered to provide useful information, and thus effects on nervous
system function could not be evaluated.
4.2.1.6.3. Summary. Overall, there is strong evidence that formaldehyde exposure via
inhalation may cause adverse effects on nervous system function in experimental animals at
relatively low levels of exposure (LOAELs as low as 100 ppb). Although human data regarding
neurotoxicity following formaldehyde inhalation are limited, available data provide support that
the types of effects seen in humans are similar to those found in animal studies. Evidence from
available human controlled inhalation exposure studies indicates that humans may be affected at
doses similar to those used in animal studies; however, the human data are extremely limited.
There are insufficient data to evaluate the potential for neurotoxicity following oral
exposure to formaldehyde. Limited evaluations of brain weight or histopathology in available
chronic or subchronic oral studies found no evidence of formaldehyde-induced changes (Til et
al., 1989, 1988; Tobe et al., 1989; Johannsen et al., 1986). However, reliable studies examining
nervous system function or focused studies of neuropathology following oral exposure to
formaldehyde are not available.
4.2.1.6.4. Other considerations. Major data gaps were found regarding the evaluation of
changes in nervous system structure or function following formaldehyde exposure by both the
inhalation or oral routes.
With respect to inhalation exposure, none of the available human studies resulted in data
sufficient to conduct a reliable dose-response assessment for changes in nervous system function.
Most of the available animal inhalation studies used short exposure durations (acute or
short-term), precluding a reliable evaluation of neurotoxicity following chronic exposure.
Available data for neurodevelopmental exposures are also quite limited, consisting of evaluation
of neuropathology in only one brain region and functional evaluations focused only on changes
in motor activity.
Major data gaps also exist regarding neurotoxicity following oral exposure, with no
relevant human data and extremely limited animal data. Available oral exposure studies were
insufficient to permit a reliable evaluation of the potential for neurotoxicity following oral
exposure to formaldehyde.
4.2.1.7. Reproductive and Developmental Toxicity
The potential for developmental and reproductive effects after formaldehyde exposure by
the inhalation route has generally been considered low, since formaldehyde, as a reactive gas, is
This document is a draft for review purposes only and does not constitute Agency policy.
4-285 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
not expected to penetrate past the POE (NEG, 2003; IPCS, 2002; Collins et al., 2001).
Nevertheless, a number of animal studies have demonstrated effects of formaldehyde on pre- and
postnatal development and on the reproductive system. For example, developmental toxicity
was observed in two studies that evaluated a standard battery of developmental endpoints
resulting from inhalation exposure on GDs 6-10 (Martin, 1990; Saillenfait et al., 1989).
Similarly, oral exposures resulted in developmental effects when administered during
comparable gestational windows (Marks et al., 1980; Hurni and Ohder, 1973). There have also
been reports that identified developmental effects at lower-level formaldehyde exposures that
were administered throughout gestation (Senichenkova and Chebotar, 1996; Senichenkova,
1991; Kitaev et al., 1984; Sheveleva, 1971; Gofmekler and Bonashevskaya, 1969; Gofmekler,
1968; Pushkina et al., 1968). Postnatal functional consequences of developmental exposures
have also been identified (Sarsilmaz et al., 2007; Asian et al., 2006; Weiler and Apfelbach, 1992;
Senichenkova, 1991; Sheveleva, 1971). Additionally, a number of studies suggest that
formaldehyde adversely affects the male reproductive system after both inhalation and oral
exposures (Xing et al., 2007; Zhou et al., 2006; Ozen et al., 2005, 2002; Sarsilmaz et al., 1999;
Chowdhury et al., 1992; Cassidy et al., 1983; Guseva, 1972). This section reviews the available
published studies assessing reproductive and developmental endpoints of formaldehyde.
4.2.1.7.1. Inhalation studies addressing developmental and reproductive toxicity. Saillenfait et
al. (1989) reported a comprehensive and well-documented developmental study in Sprague-
Dawley rats. Pregnant rats were exposed beginning on GD 6 in order to cover critical stages of
development (e.g., implantation and major organogenesis). Female Sprague-Dawley rats
(25/group) were exposed to 0, 5, 10, 20, or 40 ppm (0, 6.15, 12.3, 24.6, or 49.2 mg/m3)
formaldehyde 6 hours/day on GDs 6-20. The onset of pregnancy was determined by the
presence of sperm in a vaginal smear. Dams were exposed to formaldehyde in a dynamic flow
chamber, and formaldehyde concentrations were determined to be 0, 5.17 ± 0.51, 9.92 ± 0.88,
20.04 ± 0.88, and 38.96 ± 3.70 ppm. Dams were weighed on GDs 0, 6, and 21 and sacrificed on
day 21. Upon examination, uterine weights, fetal weights, sex ratio, number of implantation and
resorption sites, and live and dead fetuses were recorded. Fetuses were examined for external
malformations and cleft palate. One-half of viable fetuses were sectioned to assess soft-tissue
alterations. The other half were fixed, stained with alizarin red S, and examined for skeletal
alterations.
Body weight gain of dams and body weight of male and female fetuses were reduced by
exposure to 40 ppm formaldehyde to 49, 78, and 81% of control values, respectively (p < 0.01)
(Saillenfait et al., 1989). Reduced weight gain in dams remained significantly decreased when
This document is a draft for review purposes only and does not constitute Agency policy.
4-286 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
uterine weight was accounted for (p < 0.01). Mean fetal weight of male pups was reduced at
maternal exposures of 20 and 40 ppm formaldehyde (5.53 and 4.42 g versus 5.61 g in controls).
Decreased fetal body weight in females was only seen at 40 ppm (4.27 g versus 5.24 g in
controls). All other pregnancy endpoints were unchanged by formaldehyde exposure (e.g.,
uterine weight, implantation and resorption sites, live fetuses, dead fetuses, and sex ratios). No
major malformations were noted in fetuses. Some minor soft tissue and skeletal anomalies, such
as dilated ureter, missing sternebrae, extra fourteenth rib, and rudimentary thirteenth rib
(statistics not given), were reported. However, these effects occurred at similar frequencies in
control and treatment groups. The incidence of delayed ossification of the thoracic vertebrae
was 8.7% in fetuses from the 40 ppm exposure group versus 1.8% in controls. However, this
difference was not statistically significant. Overall, from these results formaldehyde was neither
lethal to embryos nor teratogenic, only exhibiting fetotoxic effects at exposures of 20 ppm and
above. These are levels where there was a significant decrease in fetal body weight.
Martin (1990) conducted a similar study exposing pregnant rats during similar stages of
development. Mated female Sprague-Dawley rats (25/group) were exposed to 2, 5, or 10 ppm
(2.46, 6.15, or 12.3 mg/m3) formaldehyde 6 hours/day on GDs 6-15. The study included two
control groups: dams placed in the exposure chambers once a day but exposed only to clean air
and dams fed and housed similarly to the experimental groups but never put into the inhalation
chambers. The method of formaldehyde vapor generation and details of the exposure chamber
were not described. Mean formaldehyde exposure concentrations were reported as 1.88, 4.88,
and 9.45 ppm (variability not given, analytical method not discussed). Food consumption and
body weight were recorded. On GD 20, rats were sacrificed, and the following pregnancy
parameters were recorded: live fetuses, dead fetuses and resorptions, number of corpora lutea,
fetal weights, sex ratios, and preimplantation and postimplantation losses. Fetuses were
examined for major malformations, minor external and visceral anomalies, and minor skeletal
anomalies (details not given). Weight gain and food consumption in dams were said to be
reduced at 10 ppm (p < 0.05). Formaldehyde exposure of the dams at 5 and 10 ppm led to an
increased incidence of reduced ossification of the pubic and ischial bones in fetuses on GD 20
(p < 0.05). Reduced ossification correlated with lower fetal weights, and the author considered
both of these findings a result of larger litter size and, therefore, not related to formaldehyde
exposure. However, no tables presenting the data or statistical analysis were provided. All other
pregnancy parameters and fetal anomalies were described as unaffected by formaldehyde
exposure. However, without data presented for the assessed endpoints, background rates of
malformations, trends in the data, and variability, it is difficult to evaluate the Martin (1990)
This document is a draft for review purposes only and does not constitute Agency policy.
4-287 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
comparisons. However, the author's observations of reduced fetal weight and increased
incidence of reduced ossification are consistent with the results of Saillenfait et al. (1989).
Kilburn and Moro (1985) studied similar endpoints but included formaldehyde exposure
during earlier gestational windows. The study report, only available in abstract form and not
found as a subsequent published article, does not provide many methodological details. Female
rats (number and strain not reported) were exposed to 0 or 30 ppm (0 or 37.2 mg/m3)
formaldehyde 8 hours/day during GDs 3-17, 3-12, 8-12, or 9-11. A second experiment
included pair-fed controls for dams exposed to 30 ppm formaldehyde during GDs 3-17, 3-12,
and 8-12. The authors reported reductions in fetal and maternal weight gain that were greater
than decreases in pair-fed controls. Fetal anomalies were noted after 15 days of gestational
exposure (e.g., altered organ size and undescended testes). Although the report indicates some
maternal toxicity and fetotoxic effects (for example, stunted growth), lack of study details and
clear reporting make this report of negligible utility in human health risk assessment.
There are several early studies that examined developmental effects of formaldehyde
exposure administered throughout gestation (Gofmekler and Bonashevskaya, 1969; Gofmekler,
1968; Pushkina et al., 1968). It is unclear if these reports represent the same or overlapping
experimental groups. They were performed in the same laboratory and are reported with a
similar level of detail. The source of formaldehyde, method of vapor generation, exposure
conditions (dynamic versus static), confirmation of exposure concentrations, study design, and
data presentation details were not provided. Absence of such critical information detracts from
the quality of these studies as a coherent record of experimental information, and, thus, these
findings can only be utilized qualitatively in the formaldehyde risk assessment.
In the Gofmekler (1968) study, female rats (36, strain not specified) were continuously
exposed at 0, low, or high formaldehyde concentrations beginning 10-15 days prior to mating
(target concentrations of 0, 0.01, or 0.81 ppm formaldehyde [0, 0.01, or 1 mg/m3]). The author
reported a 14-15% increase in pregnancy duration and a decrease in litter size (data not shown).
However, males and females were mated 6-10 days, and no information was provided on how
mating and conception were confirmed. No external malformations were attributed to
formaldehyde exposure. Concentration-dependent increases in pup body weight and decreases in
lung and liver weight were attributed to formaldehyde exposure. Pup weights were increased
from 5.6 g in controls to 6.0 and 6.3 g in groups 1 and 2, respectively (p < 0.01 andp < 0.001).
Formaldehyde exposure increased pup adrenal weight in both groups and pup thymus and kidney
weight in group 2 only (Table 4-58).
This document is a draft for review purposes only and does not constitute Agency policy.
4-288 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-58. Effects of formaldehyde on body and organ weights in rat pups
2 from dams exposed via inhalation from mating through gestation
3
Exposure
(ppm)a
Body weight
(g)
Relative organ weights (mg/10 g body weight)
Thymus
Heart
Lung
Liver
Adrenals
Kidney
0
5.6
26
61.4
287.1
587.7
3.2
51.4
0.01
6.0b
25.1
61.5
230.2°
557.7d
4.2°
53.4
0.81
6.3°
31.7°
64.5
223.2C
550.8b
3.8d
55.7b
4
5 aDams were exposed to formaldehyde continuously from 10-15 days prior to mating. Exposure concentrations
6 were not validated.
7 ''Different from controls, p < 0.01.
8 Different from controls,/? < 0.001.
9 dDifferent from controls, p < 0.05.
10
11 Source: Gofmekler (1968).
12
13 In a study by Gofmekler and Bonashevskaya (1969), the researchers evaluated organ
14 histopathology in pups from similarly treated dams. Pregnant female albino rats were
15 continuously exposed at two formaldehyde concentrations (groups 1 and 2, as described above).
16 Adult males were similarly exposed during mating. Offspring were examined for malformations,
17 and the organs were fixed and sectioned for histopathologic examination, including hematoxylin
18 and eosin staining, Brachet stain for RNA, and Feulgen stain for DNA. Liver and kidney
19 sections were also stained with Schiff s reagent (which reacts with aldehydes), with Sudan III for
20 lipids, and Pearl's stain for iron. Placenta, uterus, and ovaries from the dams and testes of the
21 males were sectioned, stained, and evaluated. The authors stated that formaldehyde induced no
22 external anomalies (reported elsewhere, but no reference given). The authors also noted
23 involution of lymphoid tissue and changes in liver, mild hypertrophy of Kupffer cells, and
24 numerous extramedullary myelopoietic centers in pups from dams in group 2. Pups from both
25 treatment groups showed reduced glycogen content in the myocardium and the presence of iron
26 in Kupffer cells. There was a localized increase in positive reaction to Schiff s reagent in the
27 basement membrane and intertubular connective tissue of the kidneys. The authors suggested
28 that this was an indication of functional alterations in the renal tubular apparatus. All other
29 tissues examined and histochemical staining indicated no differences due to formaldehyde
30 exposure.
31 Researchers in the same laboratory (Pushkina et al., 1968) studied the effects of
32 formaldehyde exposure on vitamin C (ascorbic acid, an antioxidant) and nucleic acid levels in
33 dams and fetuses as general measures of toxicity. Female white rats (n = 160) were continuously
34 exposed at two formaldehyde concentrations (groups 1 and 2, as described above) from 20 days
35 prior to mating (6-10 days) and then throughout gestation. Dams were sacrificed and ascorbic
36 acid and nucleic acid levels determined in harvested organs (methods referenced but not
This document is a draft for review purposes only and does not constitute Agency policy.
4-289 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
described). No visible malformations in pups were noted. Formaldehyde exposure increased
fetal body weight and organ weight in both groups (data not given). There was an average of
11.3 fetuses per litter for control dams versus 9.8 and 8.6 for groups 1 and 2, respectively. The
authors reported that formaldehyde exposure decreased DNA levels and increased RNA levels in
organs (further details not provided). Formaldehyde exposure resulted in lower vitamin C levels
in the whole fetus (76 and 75% of controls) and in maternal liver specimens (82 and 88% of
controls) for exposure groups 1 and 2, respectively (p < 0.05). In contrast, vitamin C was higher
in fetal liver (127%) of controls) in group 1 (p < 0.05). The significance of these differences is
unknown. The authors considered the results as general measures of biochemical changes and
therefore toxic.
The reports of Gofmekler and Bonashevskaya (1969), Gofmekler (1968), and Pushkina et
al. (1968) lack key methodological details. As discussed above, exposure conditions and actual
formaldehyde concentrations cannot be validated. Although methods were not thoroughly
detailed, results were reported in data tables with statistics and detailed descriptions of observed
pathological changes. However, without validation of exposure concentrations, these findings
can only be considered qualitatively.
In another study, Sheveleva (1971) exposed female mongrel (i.e., not a homogeneous
genetic strain) white rats to 0, 0.0005, or 0.005 mg/L (0, 0.5, or 5 mg/m3) (0, 0.4, or 4 ppm)
formaldehyde on GDs 1-19 (where GD 1 was defined as the day that spermatozoa were detected
in vaginal smears) for 4 hours/day. In each group, 15 dams were terminated on GD 20 for
evaluation of ovarian corpora lutea, uterine implantation sites, pre- and postimplantation loss,
number of live fetuses, fetal length and weight, and external examination for malformations.
Additionally, in each group, six dams were allowed to deliver their litters. Developmental
landmarks were monitored (i.e., ear and eye opening, incisor eruption, emergence of hair coat),
and the pups were further evaluated at 1 and 2 months of age for body weight, threshold of
neuromuscular excitability, total oxygen consumption in 1 hour per 100 g of weight, and
spontaneous mobility over 10 minutes. Maternal toxicity (recorded on GD 17) included
significantly (p < 0.05) decreased leukocyte counts in both treated groups and a number of
additional findings at 0.005 mg/L (i.e., significant reductions in the threshold of neuromuscular
excitability, rectal temperature, and blood hemoglobin level) as well as an increase in
spontaneous mobility over 15 minutes. (It is noted that a significant reduction in blood
hemoglobin was also observed by Sanotskii et al. [1976] in pregnant rats, following 20 days of
gestational inhalation exposure for 4 hours/day to formaldehyde at 6 mg/m3 [4.83 ppm].) Fetal
examinations on GD 20 identified a 50-70% increase in mean preimplantation loss in both
formaldehyde-exposed groups. When pups were 1 month of age, a reduction in spontaneous
This document is a draft for review purposes only and does not constitute Agency policy.
4-290 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
mobility was noted in both treated groups; in pups at 2 months of age, an increase in mobility
was observed in the 0.005 mg/L group. Also, when pups were 2 months of age, there were
alterations in hemoglobin levels and leukocyte counts in both treated groups. Detailed
descriptions of some study methodologies (particularly in regard to neurological and behavioral
assessments) were not provided in the published paper.
Based on a review of the work by Gofmekler (1968) and various epidemiologic studies
available at the time, Kitaev et al. (1984) hypothesized that formaldehyde may exert toxic effects
in the early days of gestation. To study embryotoxic effects of formaldehyde inhalation
exposures, mature female Wistar rats (five to nine per group) were exposed to 0.41 or 1.22 ppm
(0.5 or 1.5 mg/m3) formaldehyde 4 hours/day, 5 days/week for 4 months (Kitaev et al., 1984).
Rats were exposed in dynamic flow chambers and formaldehyde levels measured gravimetrically
(but not reported). Females were mated on day 120 of exposure and mating confirmed by the
presence of sperm in a vaginal smear. Embryos were harvested on the second or third day of
pregnancy (GD 2 or 3) and examined by both light and phase contrast microscopy for changes in
morphology (i.e., evidence of embryonic degeneration). Additionally, maternal weight gain and
organ weights (ovaries, uterus, and adrenal glands) and blood samples (HCT, Hb, and TP) were
monitored as indicators of general toxicity. These parameters were unchanged by formaldehyde
exposure. Formaldehyde exposure at 1.22 ppm for 4 months resulted in an increased number of
degenerating embryos on GD 3 (14.9 versus 4.4% in controls) and a smaller increase of 10.2%
(versus 5.1% in controls; statistical significance not assessed) on GD 2. Indications of
degeneration included reduced size and changes in appearance (granulation of the ooplasm,
wrinkling and degradation of nuclear material). However, it is unclear if litter effects were
accounted for in the statistical analyses, and it is unknown how the affected embryos were
distributed between litters. For dams exposed to 0.41 ppm formaldehyde, the number of
degenerated embryos was not increased on day 2 (3.8 versus 5.1% in controls) but was increased
on day 3 (9.1 versus 4.4% in controls; again, unknown if statistically significant) after maternal
exposure to 1.22 ppm formaldehyde. This observation may be coincidental since it was seen in
dams sacrificed on GD 2 but not in those sacrificed on GD 3. Kitaev et al. (1984) considered
these findings to indicate that repeated exposure to formaldehyde over a 4-month period can
disturb reproductive function, resulting in adverse effects early in embryonic development.
To further explore the effects of inhalation exposures to formaldehyde on reproductive
function, Kitaev et al. (1984) conducted a second series of experiments on 200 similarly treated
female rats. After 4 months of repeated formaldehyde exposure at 0.41 or 1.22 ppm as described
above, organ weights (ovaries and uterus) and blood levels of gonadotropic hormones and
progesterone were determined. However, the day and time of hormone measurement were not
This document is a draft for review purposes only and does not constitute Agency policy.
4-291 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
given in the report, and normal diurnal variations in these hormones could affect the reported
findings if time of day was not accounted for. The length of the estrous cycle was unchanged
during exposure. Formaldehyde exposure modulated gonadotropin levels and relative ovarian
weight, suggesting low-level effects on the female rat reproductive system prior to mating
(Kitaev et al., 1984). Ovarian weight and blood levels of luteinizing hormone (LH) were both
significantly increased after exposures at 0.41 ppm formaldehyde but remained at control levels
in rats exposed at 1.22 ppm. Blood levels of follicle-stimulating hormone (FSH) were increased
approximately 66% from control after 1.22 ppm formaldehyde exposure (p < 0.05).
Progesterone levels were unchanged by formaldehyde treatment. Kitaev et al. (1984) suggested
a role of the hypothalamus-pituitary system based on increased ovary weight, a greater number
of degenerated embryos, and increased LH in rats exposed at 0.41 ppm. They postulated that
these effects were not seen at 1.22 ppm due to a toxic effect exhibited as embryonic
degeneration, thus the absence of a dose-response did not alter the interpretation of the validity
of the adverse response. The study NOAEL was not determined, and the study LOAEL was 0.4
ppm (0.5 mg/m3), based upon increased early embryo loss and on maternal outcomes (increased
ovarian weight and increased blood LH levels) following 4 months of formaldehyde treatment.
For the finding of increase blood FSH levels, the endpoint NOAEL was 0.4 ppm (0.5 mg/m3)
and the LOAEL was 1.2 ppm (1.5 mg/m3).
Senichenkova and Chebotar (1996) and Senichenkova (1991) examined reproductive and
developmental effects of daily formaldehyde exposure on GDs 1-19 of pregnancy, including the
potential effect of formaldehyde exposure on development early in gestation. Additionally, since
anemia adversely affects fetal development, Senichenkova and Chebotar (1996) also examined
formaldehyde effects in iron-deficient dams to determine whether co-exposure further
compromises fetal development. In both studies, female white rats were exposed to 0 or
0.41 ppm formaldehyde (0 or 0.5 mg/m3), 4 hours/day on GDs 1-19. Formaldehyde
concentrations in the dynamic exposure chambers were measured gravimetrically to confirm the
exposure concentration but were not reported (methods not provided). It is unclear if gravimetric
measurements would be sensitive or accurate enough to validate these low-exposure
concentrations without a better understanding of the methodology. This uncertainty in exposure
conditions should be considered in evaluating the reported results.
Mongrel female white rats were exposed at a target concentration of 0 or 0.41 ppm (0 or
0.5 mg/m3) formaldehyde 4 hours/day on GDs 1-19 (Senichenkova, 1991). On GD 20, a subset
of the dams was sacrificed and examined for number of corpora lutea, implantation and
resorption sites, live/dead fetuses, and fetal weights. Fetuses were examined for gross pathology
of the internal organs and skeleton (details not given). Blood pH, partial pressure of CO2, and
This document is a draft for review purposes only and does not constitute Agency policy.
4-292 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
partial pressure of oxygen were measured in both dams and embryos. The remaining dams were
brought to term to study postnatal effects of formaldehyde exposure. Rat pups were observed on
PNDs 1-25 for viability, physical development, and maturation rate of motor reflexes. Behavior
of juvenile offspring (PND 40) was studied in an open field test, and maze learning was tested at
sexual maturity. In a follow-up report, Senichenkova and Chebotar (1996) present blood
chemistry data, pregnancy outcome, and developmental data for similarly treated dams and their
pups in a chemical model of iron deficiency. Intraperitoneal injections of the iron-chelating
agent bipyridyl were given on GDs 12-15 at the threshold embryotoxic dose (1 mL, 25%
solution). On day 20, the dams were sacrificed and dams and fetuses examined as described
above. In addition to blood pH, partial pressure of carbon dioxide and partial pressure of
oxygen, acid metabolic products (not detailed), and true bicarbonates were reported for maternal
and fetal blood. A review of the data from these reports indicates there may be an overlap of the
study groups. Neither paper presents the entire data set; thus, for tranparency and brevity, the
following text discusses the combined findings from both studies as if they were a single study
Formaldehyde exposure did not affect such indicators of pregnancy outcome as number
of corpora lutea, implantation and resorption sites, and live and dead fetuses, all of which were
unchanged (Senichenkova and Chebotar, 1996; Senichenkova, 1991). Although fetal weight was
slightly increased by formaldehyde exposure, 2.35 versus 2.24 g in controls (p < 0.001), neither
fetal length nor bone length were changed (femur and humerus) (Senichenkova and Chebotar,
1996; Senichenkova, 1991). Often, increased fetal weight is the result of early physical
development, and other signs of development, such as ossification, would be expected to be
enhanced as well. The average number of bone centers per limb was increased by formaldehyde
exposure from 2.45 and 2.66 to 2.78 and 2.91 in controls for metacarpal and metatarsal bone
centers, respectively (p < 0.05) (Senichenkova, 1991); these findings were consistent with
increased growth and weight. In contrast, Senichenkova (1991) reported a decrease in the
number of embryos with ossification centers in the hyoid bone (100% in controls versus 91% for
formaldehyde exposure,/* < 0.05), consistent with the results of Saillenfait et al. (1989) and
Martin (1990). However, litter size, a factor influencing fetal weight, was not provided, and it is
unclear if Senichenkova (1991) took litter size into account in the analysis.
Senichenkova and Chebotar (1996) reported increased blood acidosis and decreased
blood alkaline reserves (bicarbonates and total CO2) in formaldehyde-treated dams and their
embryos (p < 0.05). However, this finding should be considered in light of the fact that chronic
blood acidosis may increase bone remodeling and decrease bone density in adults. It is unknown
if the reported blood acidosis could reduce ossification rates in developing embryos. A better
This document is a draft for review purposes only and does not constitute Agency policy.
4-293 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
understanding of exposure conditions and the acid metabolic products measured is needed to
determine the biological relevance of the reported changes in blood acid balance.
Iron deficiency, induced by injections of bipyridyl (an iron-chelating agent), was found to
be fetotoxic. Iron-deficient dams with no formaldehyde exposure had higher rates of
postimplantation death than controls (12.6 ± 5.5 versus 4.8 ± 1.3%). Formaldehyde exposure in
conjunction with iron deficiency increased postimplantation death to 23.1 ± 5.9%. Fetal weight
and litter size were not reported for the bipyridyl treatment groups, but bipyridyl treatment in
conjunction with formaldehyde resulted in a decreased number of metatarsal bone centers (2.21
± 0.12 versus 2.72 ± 0.08 in controls; p < 0.001). This decrease was also significant when
compared with formaldehyde or bipyridyl alone (p < 0.02). However, all pregnancy outcome
parameters were not reported for the bipyridyl treatment.
Fetal anomalies were reported after formaldehyde exposure and were increased by iron
deficiency. The incidence of litters with internal organ anomalies was increased from 1.4% in
controls to 14.2% in formaldehyde-treated dams (Senichenkova, 1991). Undescended testes
were the predominant anomaly described: 20.8% in litters from formaldehyde-treated dams
versus 1.2% in controls (p < 0.05) (Senichenkova, 1991). Similar findings were reported by
Senichenkova and Chebotar (1996). Bipyridyl treatment in conjunction with formaldehyde
exposure increased the overall incidence of fetal anomalies (13.8 ± 2.1% in controls versus 6.6 ±
1.8%) with iron deficiency alone) (Senichenkova and Chebotar, 1996). However, there are
discrepancies between the two papers in the reporting of the anomalies, and it is unclear whether
the experimental groups overlap between papers, where some parameters are identical (which
would lead to double counting of the same animal, including identical SDs) and others are
different. Additionally, the reporting is unclear with respect to the basis of the incidence rates
reported (for example, overall incidence versus incidence within litter or incidence of litters with
anomalies). Unclear reporting, together with some of the uncertainties regarding exposure
conditions, suggests that the data may be of limited quality to support risk assessment.
In the second phase of the studies, pups were delivered and postnatal development was
assessed (Senichenkova, 1991). Eruption of the upper and lower incisors was delayed in pups
from formaldehyde-treated dams, occurring on PND 14 versus PND 12 in controls (p < 0.01).
All other measures of physical postnatal development were unchanged by formaldehyde. To
evaluate postnatal functional outcomes following in utero exposure to formaldehyde, an open
field test was conducted in juvenile rats on 3 consecutive days (PNDs 40-42). Juvenile rats from
formaldehyde-treated dams exhibited increased mobility (crossed squares), rearings, and
defecations/urinations compared with control rats on the second and third open field tests
This document is a draft for review purposes only and does not constitute Agency policy.
4-294 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
(p < 0.05). There were no differences seen in the maze-learning test assessed in mature offspring
of formaldehyde-treated dams (Senichenkova, 1991).
Additional assessments of formaldehyde exposure on neurological development are
described above in the section on neurological and behavioral toxicity in animal studies (Section
4.2.1.6). In brief, studies conducted by Sarsilmaz et al. (2007) and Asian et al. (2006) exposed
10 neonatal male Wistar rats/group to 0, 6, or 12 ppm (0, 7.36, or 14.7 mg/m3) formaldehyde
6 hours/day, 5 days/week for 30 days. At that time, five rats/group were killed and subjected to
neuropathological examination; the remaining rats were maintained until PND 90, at which time
they were killed and evaluated. Asian et al. (2006) examined the number and volume of granular
cells in the hippocampal dentate gyrus, while Sarsilmaz et al. (2007) examined the size and
number of the pyramidal cells in the cornu ammonis of the hippocampus. In both studies, lower
numbers of cells were observed in both treated groups at PND 90 as compared with PND 30.
Although the effects of treatment on the volume and number of granular and pyramidal cells
were somewhat inconsistent, a significant decrease in the number of neurons in the pyramidal
cell layer of the hippocampal cornu ammonis was observed at both PNDs 30 and 90 (Sarsilmaz
et al., 2007).
One other study reported effects on nervous system function following exposure to
formaldehyde during postnatal development. An abstract by Weiler and Apfelbach (1992)
described a study in which juvenile rats (strain not specified) were exposed to 0.25 ppm
(0.31 mg/m3) formaldehyde from PNDs 30-160 or adult rats were exposed to 0.5 ppm
(0.62 mg/m3) formaldehyde for 90 days. Decreased olfactory sensitivity (i.e., increased olfactory
thresholds) was observed and was greater when the exposure was initiated in the young rats, as
compared with the adults.
Evaluation of offspring following prenatal, perinatal, and/or juvenile inhalation exposures
to formaldehyde have also been reported by Kum et al. (2007), Sandikci et al. (2007), and
Songur et al. (2005). Kum et al. (2007) exposed female Sprague-Dawley rats (six dams/group)
and their offspring to 0 or 6 ppm (0 or 7.4 mg/m3) formaldehyde for 8 hours/day in separate
groups with exposures starting on GD 1, on postparturition day 1, or in offspring at 4 weeks of
age and continuing for 6 weeks. In another group, exposures were initiated in adult rats. Mean
body and liver weights were significantly decreased in the offspring exposed in utero and in
early postnatal life, and mean liver weights were also significantly decreased in rats with juvenile
exposures. However, neither body weight nor liver weight was affected in the group with
exposure initiating at an adult age, suggesting a life-stage-related susceptibility to formaldehyde-
induced hepatic toxicity. Evaluation of biomarkers of oxidative stress revealed significantly
increased catalase (CAT) and MDA in the livers of offspring that were exposed prenatally,
This document is a draft for review purposes only and does not constitute Agency policy.
4-295 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
significantly decreased GSH levels in the livers of offspring that were exposed in early postnatal
life, and significantly decreased SOD levels in the livers of offspring that were exposed starting
at 4 weeks of age. No biomarkers were altered in the livers of rats exposed to formaldehyde only
as adults.
A similar study design was used by Sandikci et al. (2007) to examine the effects of 0 or
6 ppm (0 or 7.4 mg/m3) formaldehyde on bronchus associated lymphoid tissue (BALT)
following pre- and perinatal, juvenile, or adult exposures of 6 weeks duration in Sprague-Dawley
rats (six/group). The presence of the lysosomal enzyme alpha-naphthylacetate esterase (ANAE)
served as a marker of T-lymphocytes in peripheral blood and tissue sections. Significant
increases in ANAE-positive T-lymphocytes were found in BALT in all but the in utero exposed
groups as compared with control; this outcome is consistent with the postnatal development of
BALT in the rat. In peripheral blood, ANAE-positive lymphocyte ratios were significantly (p <
0.001) increased as compared with controls at all life stages tested.
Songur et al. (2005) examined the effect (and reversibility) of formaldehyde exposures
during the early postnatal period on zinc, copper, and iron levels and activity of SOD in the lung
tissue of Wistar rats. Litters (12-14/group) were exposed to 0, 6, or 12 ppm (0, 7.4, or
14.9 mg/m3) formaldehyde 6 hours/day, 5 days/week for 30 days. Trace element and
biochemical analyses were conducted on PND 30 or 90. Decreased SOD activity, decreased
levels of copper and iron levels, and increased zinc levels were observed in the lungs of treated
groups following 30 days of treatment and at 90 days (i.e., 60 days posttreatment). Survival was
not affected in neonatal rats. Clinical observations during treatment included evidence of
respiratory irritation and toxicity. Body weight and food and water consumption were also
nonsignificantly decreased as compared with controls.
There are several reports in the literature regarding formaldehyde effects after inhalation
exposure on the male reproductive system in animals (Galalipour et al., 2007; Zhou et al., 2006;
Ozen et al., 2005, 2002; Sarsilmaz et al., 1999; Woutersen et al., 1987; Maronpot et al., 1986;
Guseva, 1972). The earliest report examined the effect of simultaneous exposures to
formaldehyde from air and water (Guseva, 1972). Male rats (n = 12, strain not specified) were
co-exposed to formaldehyde in air and drinking water 4 hours/day, 5 days/week for 6 months.
There were three exposure levels in the experiment of different air and drinking water
concentrations: (1) 0.41 ppm (0.5 mg/m3) formaldehyde in air and 0.1 mg/L in water; (2) 0.20
ppm (0.25 mg/m3) formaldehyde in air and 0.01 mg/L in water; or (3) 0.10 ppm (0.12 mg/m3)
formaldehyde in air and 0.005 mg/L in water. Reproductive function was assessed by mating
two females per male. The time for the onset of pregnancy and the number of pregnancies per
treatment group were recorded. A subset of dams was sacrificed on GD 20 of pregnancy, and
This document is a draft for review purposes only and does not constitute Agency policy.
4-296 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
the number and weight of fetuses was determined. Postnatal development of the remaining dams
was tracked (e.g., times of eye opening and development of hair coat). Nucleic acid levels were
determined in the testes of formaldehyde-exposed rats. Gonadotropic response was assessed by
injecting an emulsion of pituitaries from exposed male rats into unexposed infantile females and
measuring the weight ratios of the uterus and ovaries. Formaldehyde exposure reduced nucleic
acid levels in testes to 88 and 92% of controls in groups 1 and 2, respectively. No other
formaldehyde-induced differences were found.
Woutersen et al. (1987) and Maronpot et al. (1986) examined tissue sections from testes
and ovaries of exposed animals as part of studies primarily addressing respiratory tract toxicity
(see Section 4.2.1.2.2.4 for complete study details). Maronpot et al. (1986) exposed female and
male B6C3F1 mice to 0, 2, 4, 10, 20, and 40 ppm (0, 2.45, 4.91, 12.3, 24.5, and 49.1 mg/m3)
formaldehyde 6 hours/day, 5 days/week for 13 weeks. Decreased weight gain due to
formaldehyde exposure was seen in both male and female mice. Additionally, there was 80%
mortality at the highest exposure (40 ppm). The authors reported endometrial hypoplasia and
lack of ovarian luteal tissue in females exposed to 40 ppm, but no compound-related changes
were observed in testes sectioned and viewed by light microscopy.
In a study by Appleman et al. (1988), male Wistar rats (40/group) with undamaged or
damaged (via bilateral intranasal electrocoagulation) nasal mucosa were exposed for 13 or
52 weeks to 0, 0.1, 1, or 10 ppm (0, 0.124, 1.24, or 12.4 mg/m3) formaldehyde 6 hours/day,
5 days/week. At study termination, mean body weight was decreased, but relative testes weight
was increased in the 10 ppm group (interpreted as a non-adverse outcome that was associated
with the decreased body weight). No treatment-related histopathologic findings were reported
for male reproductive organs (although it is not clear to what extent they were evaluated since
the primary focus of the study was on the nasal epithelium).
Following up on earlier reports of decreased Ley dig cell quality in rats administered I.P.
injections of formaldehyde (Chowdhury et al. [1992], described in Section 5.2.1.8.3), Sarsilmaz
et al. (1999) studied the effects of formaldehyde inhalation on Leydig cells. Adult male Wistar
rats (30) were exposed to 0, 10, or 20 ppm (0, 12.3, or 24.6 mg/m3) formaldehyde 8 hours/day,
5 days/week for 4 weeks. Animals were observed daily and weighed weekly. Rats were
sacrificed on day 29 and autopsied, and testes were weighed, fixed, and sectioned for histologic
examination. Signs of irritation from formaldehyde exposure were noted (frequent eye blinking,
excessive licking, increased frequency of nose cleaning, interrupted breathing, and sneezing).
Body weight gain was reduced by formaldehyde exposure from 11.1% gain in control rats to
4.66 and 2.63% in rats exposed at 10 and 20 ppm, respectively (p < 0.001). As shown in
Table 4-59, relative testes weights were unaffected (reported as proportions but more likely to be
This document is a draft for review purposes only and does not constitute Agency policy.
4-297 DRAFT—DO NOT CITE OR QUOTE
-------
1 percentages), although trends and numerical differences were similar to those reported by Ozen
2 et al. (2002). Sarsilmaz et al. (1999) found that both Ley dig cell quantity and the percentage of
3 cells with normal nuclei were reduced by formaldehyde treatment. Although the dose-dependent
4 reduction in Leydig cell quantity was statistically significant at both exposure levels, the study
5 authors considered the data to be within the normal range.
6
7 Table 4-59. Formaldehyde effects on Leydig cell quantity and nuclear
8 damage in adult male Wistar rats
9
Inhalation
exposure"
Relative
testes
weightbcd
Leydig cell
quantity0'6''
Appearance of nucleus6'8
Normal
Pyknotic
Karyorectic
Karyolytic
Control
0.93 (0.03)
47.27 (7.8)
98
2
0
0
10 ppm
0.92 (0.06)
45.04 (7.8)h
92
2
4
2
20 ppm
0.89 (0.06)
44.36 (7.5)1
76
9
10
5
10
11 aRats were exposed 8 hours/day, 5 days/week for 4 weeks.
12 bStated to represent the ratio of the last day's testicle weight to the body weight but more likely to be the percent
13 of body weight.
14 °Cells within 100 defined areas.
15 dn= 10.
16 Tor each exposure group, 100 defined locations were assessed.
17 fn= 100.
18 8Values presented as percentage of cells.
19 hDifferent from control (p < 0.05), as reported by the authors.
20 'Different from control (p < 0.01). as reported by the authors.
21
22 Source: Sarsilmaz et al. (1999).
23
24
25 It was hypothesized that decreased Leydig cell quality may have been the result of
26 oxidative stress induced by formaldehyde exposure. Ozen et al. (2002), in the same laboratory,
27 investigated changes in testicular iron, copper, and zinc levels as measures of oxidative stress
28 and damage. Adult male albino Wistar rats (seven/group) were exposed at 0, 10, or 20 ppm (0,
29 12.2, or 24.4 mg/m3) formaldehyde 8 hours/day, 5 days/week for either 4 or 13 weeks. Rats
30 were observed daily and weighed once a week. Rats were sacrificed at the end of the exposure
31 period and autopsied, and the testes were removed and weighed. Zinc, copper, and iron levels
32 were determined in testes tissue by using atomic absorption spectrophotometry. Both weight
33 gain and relative testes weight were decreased in a concentration-dependent and duration-
34 dependent manner (Table 4-60). Both zinc and copper levels in rat testes were reduced in a
35 concentration- and duration-dependent manner by formaldehyde exposure. For example, zinc
36 was reduced from 277 to 107 mg/kg after a 4-week x 20 ppm exposure and from 260 to
37 95 mg/kg after a 12-week x 20 ppm exposure (p < 0.001) (Table 4-60). Iron levels in testes were
This document is a draft for review purposes only and does not constitute Agency policy.
4-298 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
increased from 30 to 39 mg/kg after a 4-week 20 ppm exposure (p < 0.01) and from 33 to
58 mg/kg after 12 weeks at 20 ppm (p < 0.05). The authors suggested that alterations in trace
element levels in the testes were consistent with oxidative damage and may have contributed to
changes in Leydig cell function. These researchers also reported alterations in trace metals in
lung tissue from Wistar rats exposed to formaldehyde 8 hours/day, 5 days/week for 4 or
13 weeks. Iron levels were increased at 5 ppm for 13 weeks and 10 ppm for either 4 or
13 weeks. Zinc levels decreased for all formaldehyde exposures. In both cases, the authors
attributed elevated iron levels to oxidative stress. In addition to citing the role of zinc as a
cofactor of cytoplasmic Cu-Zn-SOD, the authors suggested that zinc may have been consumed
by increased FALDH activity. Although oxidative stress and increased FALDH activity may be
relevant to the POE and therefore impact the lung, it is less clear how these changes would occur
in the testes. Taken together, the reports of Ozen et al. (2002) and Sarsilmaz et al. (1999)
suggested a LOAEL of 10 ppm 8 hours/day for 4 weeks for changes in Leydig cell quantity and
quality, decreased testes weight, and changes in trace metal content (zinc, copper, and iron).
Table 4-60. Formaldehyde effects on adult male albino Wistar rats
Inhalation
exposure"
Weight gain
(%)
Testes weight
(%)
Zinc
(mg/kg)
Copper
(mg/kg)
Iron
(mg/kg)
4 Weeks
Control
10 ppm
20 ppm
19.1 (2.7)
5.8 (2.4)b
2.4 (0.6)b
0.94 (0.03)
0.92 (0.02)°
0.91 (0.01)c
277 (16)
132 (8.9)b
107 (6.9)b
6.4 (0.42)
4.2 (0.33)b
3.3 (0.27)b
30 (2.7)
35 (2.8)d
39 (3.1)d
13 Weeks
Control
10 ppm
20 ppm
55.9(2.3)
34.7 (3.5)b
20.8 (1.4)b
0.91 (0.01)
0.84 (0.03)b
0.82 (0.03)b
269 (15)
112 (8.1)b
95 (6.4)b
6.0 (0.34)
3.6 (0.30)b
1.9 (0.17)b
33 (2.6)
52 (3.5)b
58 (3.0)b
'Formaldehyde exposure was 8 hours/day, 5 days/week for either 4 or 13 weeks. Values are means ± SDs of
seven animals..
'Different from control, p < 0.001. as calculated by the authors.
Different from control, p < 0.05, as calculated by the authors.
dDifferent from control,/) < 0.01, as calculated by the authors.
Source: Ozen et al. (2002).
In another study that assessed testicular toxicity (Ozen et al., 2005), male Wistar rats
(18/group) were exposed by inhalation to 0, 5, or 10 ppm (0, 6.2, or 12.4 mg/m3) formaldehyde
8 hours/day, 5 days/week for 91 days. In-life observations in exposed rats included clinical signs
of respiratory irritation and decreased mean food and water consumption. At study termination,
serum testosterone levels and mean seminiferous tubule diameters were significantly decreased
from control in a dose-responsive manner (Table 4-61). Immunohistochemical staining of testis
This document is a draft for review purposes only and does not constitute Agency policy.
4-299 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
tissues showed increased localization of heat shock protein (Hsp) 70 in the cytoplasm of
spermatogonia, spermatocytes, and spermatids of treated animals compared with controls (not
shown here).
Table 4-61. Formaldehyde effects on testosterone levels and
seminiferous tubule diameters in Wistar rats following 91 days of
exposure
Inhalation
exposure"
Testosterone levels
(ng/dL)
Seminiferous tubule
diameters
(nm)
n= 6
o
o
II
Control
406.54 ± 16.82
259.22 ± 16.18
10 ppm
244.01 ±23.86b
236.17 ± 13.09c
20 ppm
141.30 ±08.56b
233.24 ± 10.13°
'Formaldehyde exposure was 8 hours/day, 5 days/week for 91 weeks. Values are means ±
SEMs.
bDifferent from control, p < 0.0001, by one-way ANOVA, as calculated by the authors.
Different from control,/) < 0.001, by one-way ANOVA, as calculated by the authors.
Source: Ozen et al. (2005).
Zhou et al. (2006) investigated the effect of formaldehyde on the testes and the protective
effect of vitamin E against oxidative damage by formaldehyde in adult male rats. In this study,
adult male Sprague-Dawley rats (10/group) were treated for 2 weeks in the following groups:
(1) control rats were administered physiological saline by oral gavage, (2) rats were administered
physiological saline by gavage and exposed to 10 mg/m3 (8.05 ppm) formaldehyde by inhalation
for 12 hours/day, and (3) rats were administered daily gavage doses of 30 mg/kg vitamin E and
exposed to 10 mg/m3 (8.05 ppm) formaldehyde by inhalation for 12 hours/day. Formaldehyde
treatment resulted in significantly decreased (p < 0.05) mean testis weight. Histopathologic
findings in treated rats included atrophy of seminiferous tubules, decreased spermatogenic cells,
and seminiferous cells that were "disintegrated" and shed into the lumina, which was
azoospermic. Interstitial tissue was edematous with vascular dilatation and hyperemia. In the
formaldehyde-treated group, epididymal sperm count and percentage of motile sperm were
significantly decreased, and the percentage of abnormal sperm was increased (p < 0.05), as
compared with control. Evaluation of biochemical markers in testes tissue showed the activities
of testicular SOD, GPX, and GSH were decreased; MDA levels were significantly increased as
compared with control. All observed effects of formaldehyde treatment (i.e., decreased testes
weight, biochemical alterations, histopathologic effects, and sperm count, motility, and
morphology findings) were attenuated by administration of 30 mg/kg-day vitamin E.
This document is a draft for review purposes only and does not constitute Agency policy.
4-300 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
In a study by Golalipour et al. (2007), testicular changes of increasing severity with
increasing duration were reported. A total of 28 Wistar rats, aged 6-7 weeks old, were divided
into four groups including three FA treatment groups (4 hours of exposure/day, 4 days/week for
18 weeks; 2 hours of exposure/day, 4 days/week for 18 weeks; 4 hours of exposure/day, 2
days/week for 18 weeks) and one untreated control. The three FA-treated groups were exposed
via inhalation to formaldehyde. The mean concentration of FA vapor, based on three
measurements during the study (stated as the beginning, during, and end of the study period),
was reported as 1.5 ppm. At the end of the study period, the rats were sacrificed by ether
anesthesia and subsequent cervical dislocation. The left testis was dissected from each rat and a
specimen was taken from each testis. Tissues were fixed, embedded, sectioned (at 4 (j,m), and
stained with hematoxylin and eosin. Using a light microscope, a histopathological examination
was performed on the testes tissues, including morphometric evaluation of the diameter and
height of 20 seminiferous tubules/testis. Golalipour et al. (2007) reported an FA exposure
frequency (or duration)-dependent increase in testicular germ cells and seminiferous tubule
defects. In the most frequent duration treatment group, a severe decrease in germ cells in >85%
of the seminiferous tubules and arrested spermatogenesis were observed. In the mid-level
frequency of duration treatment group, a decrease in the number of germ cells and an increased
thickness of the basement membrane of 75% of the tubules was observed. In the lowest level
duration treatment group, a disruption in the Sertoli and germinal cell arrangement, and
increased spacing between germ cells was observed. Further, the seminiferous tubule diameter
(STD) and seminiferous epithelial height (SEH) was most decreased among the treatment groups
(exhibiting the greatest decrease in the group with the greatest hours and days of exposure)
compared to the control (Table 4-62). The results of this study are consistent with the findings of
other studies of male reproductive system outcomes with inhalation FA exposure (e.g., Ozen et
al., 2005 and Zhou et al., 2006).
Xing et al. (2007) also studied the effects of formaldehyde on sperm development and
reproductive capacity in adult male mice. In this study, male mice (12/group, strain not
specified) were exposed to 0, 21, 42, or 84 mg/m3 (0, 16.9, 33.8, or 67.6 ppm) formaldehyde via
inhalation for 13 weeks at 2 hours/day, 6 days/week. The males were mated to untreated females
in a dominant lethal protocol, and sperm morphology was assessed at study termination. The
percent abnormal sperm was increased significantly (p < 0.05) in all treated groups, as was the
rate of resorptions (p < 0.01) (Table 4-63). The mean number of live fetuses/litter was decreased
in all treated groups, with statistical significance achieved at 84 mg/m3. Although this study did
not assess the number of corpora lutea per dam, thereby precluding the calculation of
This document is a draft for review purposes only and does not constitute Agency policy.
4-301 DRAFT—DO NOT CITE OR QUOTE
-------
1 preimplantation loss, it is nevertheless indicative of formaldehyde-induced sperm morphology
2 changes and dominant lethal effects in male mice.
3
4 Table 4-62. Effects of formaldehyde exposure on seminiferous tubule
5 diameter and epithelial height in Wistar rats following 18 weeks of
6 exposure
7
Seminiferous tubule
Seminiferous tubule
Inhalation
diameters
height
exposure"
fum)
(jim)
n = 7
n = 7
Control
252.12 ±4.82
82.77 ± 2.00
1.5 ppm, 4 h/d, 4 d/w
204.55 ±3.29b
65.26 ± 1.43b
1.5 ppm, 2 h/d, 4 h/w
232.45 ±2.42b
69.46 ± 1.78b
1.5 ppm, 2 h/d, 2 d/w
238.94 ±4.37b
72.80 ± 2.03b
8
9 a Values are means ± SEMs.
10 b Different from control, p < 0.05. as calculated by the authors.
11
12 Source: Golalipour et al. (2007).
13
14
15 Table 4-63. Incidence of sperm abnormalities and dominant lethal effects in
16 formaldehyde-treated mice
17
Sperm abnormalities
Reproductive capacity
Dose
Total abnormal
Aberration rate
Resorption rate
(mg/m3)
sperm heads
(%)
Mean live fetuses/litter
(%)
0
391
3.53 ±0.98
11.00 ± 1.01
2.273
21
568
5.48 ± 1.45
10.67 ± 1.16
9.380b
42
849
6.15 ± 1.36
9.63 ±2.83
10.390b
84
974
9.24 ± 2.13a
9.04 ± 2.98a
12.440 b
18
19 "Significantly different from controls (p < 0.05), as calculated by the authors.
20 bSignificantly different from controls (p < 0.01), as calculated by the authors.
21
22 Source: Xing et al. (2007).
23
24
25 4.2.1.7.2. Oral exposure studies addressing developmental and reproductive toxicity. No
26 contemporary testing guideline studies, such as a prenatal developmental toxicity study or two-
27 generation reproductive toxicity study, have been performed by the oral route for formaldehyde.
28 However, a number of studies have evaluated developmental toxicity and reproductive
29 parameters in rats, mice, and dogs.
30 Hurni and Ohder (1973) tested the developmental toxicity of formaldehyde administered
31 as a 40% w/v solution containing 11-14% w/w methanol in 9-10 pregnant beagle dogs that
This document is a draft for review purposes only and does not constitute Agency policy.
4-302 DRAFT—DO NOT CITE OR QUOTE
-------
1 received the compound in their diet on GDs 4-56. Commercial grade formaldehyde (as a 40%
2 solution) was sprayed on the pellets prior to feeding. Each animal was allotted a diet of 300 g of
3 chow (reduced to 200 g 1 week prior to term) that was promptly consumed (within 5-
4 10 minutes) before the formaldehyde volatilized appreciably. The concentrations of
5 formaldehyde in the chow were 0, 125, or 375 ppm, equivalent to doses of 0, 3.1, or
6 9.4 mg/kg-day, respectively. Dams were allowed to deliver normally and weight gain, gestation
7 length, number of litters, litter size, number of live pups, number of pups surviving through
8 weaning, and pup weights weekly for the first 8 weeks were monitored as indices of the potential
9 reproductive/developmental toxicity of formaldehyde. There were no formaldehyde-related
10 effects in any of the parameters other than progressive pup weights, which were lower by group
11 in litters of dams exposed to formaldehyde (Table 4-64). A developmental impact of
12 formaldehyde was evident in this strain of dog under the conditions of the experiment. At birth,
13 mean pup body weights were 4 and 8.4% less than control for the low- and high-dose groups,
14 respectively; at 8 weeks of age, the pup weight decrements were 8.3% for the low dose and
15 12.5%) for the high dose, as compared with control, and established a LOAEL of 125 ppm. The
16 contribution of methanol, which is a developmental toxin (Deglitz et al., 2004; Rogers et al.,
17 2004) to these outcomes is not known. No internal or skeletal malformations were observed in
18 any of the 264 live-born and 20 still-born pups.
19
20 Table 4-64. Body weights of pups born to beagles exposed to
21 formaldehyde during gestation
22
Time (weeks)
Formaldehyde concentration in chow (ppm)
0
125
375
Average body weight (g)
0
321
308
294
1
547
526
467
2
818
755
706
3
1,078
987
944
4
1,264
1,247
1,166
5
1,601
1,512
1,429
6
2,020
1,816
1,741
7
2,449
2,263
2,145
8
2,957
2,712
2,587
23
24 Source: Hurni and Ohder (1973).
25
26
27 Marks et al. (1980) conducted a developmental toxicity study of formaldehyde in CD-I
28 mice in which 29-35 pregnant animals were gavaged on GDs 6-15 with aqueous formaldehyde
29 (containing 10-15% methanol) at 74, 148, and 185 mg/kg-day. Seventy-six controls were
30 gavaged with water alone. All dams were sacrificed on GD 18, and the numbers of implantation
This document is a draft for review purposes only and does not constitute Agency policy.
4-303 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
sites in each uterine horn were counted. The high dose of formaldehyde was toxic to the dams,
as indicated by the deaths of 22 of 34 mice before GD 18. Thus, the dose of 148 mg/kg-day was
a NOAEL for maternal toxicity in this study. However, it is unclear to what extent an estimated
concurrent dose of up to 75 mg/kg-day methanol may have contributed to this toxic response.
To assess the developmental toxicity of formaldehyde, live fetuses were weighed individually,
sexed, and examined for external, visceral, and skeletal malformations. Fetuses of surviving
high-dose dams and of those of other groups did not show an increased incidence of
malformations. Therefore, Marks et al. (1980) concluded that formaldehyde did not induce fetal
abnormalities and that the 185 mg/kg-day dose level was a NOAEL for the developmental
toxicity of formaldehyde, nor were fetotoxic effects of methanol apparent under the study
experimental conditions.
Seidenberg and Becker (1987) and Seidenberg et al. (1986) included formaldehyde
(purity not indicated) in a survey of the behavior of potential toxicants in a developmental
toxicity screening assay (Chernoff and Kavlock, 1982). The protocol featured the administration
of a borderline toxic dose to 26-30 pregnant ICR/SIM mice on GDs 8-12. Dams were allowed
to deliver, and the neonates were examined, counted, and weighed on PNDs 1 and 3. The
selected formaldehyde dose of 540 mg/kg-day was fatal for 11/30 dams, but the average weight
gain among surviving dams was little changed compared with controls (3.9 ± 2.3 versus
4.0 ±1.0 g). Similarly, there were no changes in perinatal responses in the neonates of exposed
dams compared with controls. For example, the average values for number of neonates/litter,
percent survival, and fetal weights on PNDs 1 and 3 were closely similar to those of controls.
Evidence of toxicity to the male reproductive system has been observed following oral
administration of formaldehyde in a 40% w/v solution containing 11-14% w/w methanol.
Cassidy et al. (1983) administered single oral doses of 100 or 200 mg/kg to five male Wistar
rats/group. Testes from these animals and 20 controls were excised and examined for
spermatogenic abnormalities 11 days after dosing. A significant (19%) increase in testicular
sperm head counts was observed in rats exposed to 200 mg/kg-day formaldehyde as compared
with controls (Table 4-65). The percentage of abnormal sperm heads was also significantly
increased (5%) in the 200 mg/kg-day dose group compared with controls. These data suggest
that formaldehyde can induce morphologic abnormalities in the germ cells of male experimental
animals at dose levels that did not significantly affect testis weights. The contribution of
methanol to these outcomes is unknown.
This document is a draft for review purposes only and does not constitute Agency policy.
4-304 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-65. Testicular weights, sperm head counts, and percentage incidence
2 of abnormal sperm after oral administration of formaldehyde to male Wistar
3 rats
4
Dose (mg/kg)
Mean testes weight (g)
Mean sperm heads x
106/g testis
Abnormal sperm heads (%)
0
3.30
175
4.76
100
3.27
166
5.22
200
3.16
209a
9.1T
5
6 "Significantly different from controls (p < 0.001), as calculated by the authors.
7
8 Source: Cassidy et al. (1983).
9
10
11 Postmortem evaluation of the reproductive organs was conducted in a number of oral
12 studies that ranged between 4 weeks and 2 years in duration (Til et al., 1989, 1988; Tobe et al.,
13 1989; Johanssen et al., 1986). Johannsen et al. (1986) administered 0, 50, 100, or 150 mg/kg-day
14 formaldehyde in the drinking water to Sprague-Dawley rats (15/sex/group) for 91 days and 0, 50,
15 75, or 100 mg/kg-day formaldehyde in basal diet to beagle dogs (4/sex/group) for 91 days; the
16 study reported no treatment-related effects on absolute or relative gonad weights or
17 histopathology for either species. In a 4-week drinking water study conducted by Til et al.
18 (1988), formaldehyde was administered to Wistar rats (10/sex/treated group) at nominal levels of
19 0, 25, and 125 mg/kg-day; gonad organ weights and histopathology were not affected by
20 treatment. Tobe et al. (1989) conducted a chronic (24-month) study in Wistar rats
21 (20/sex/group), with drinking water concentrations of 0, 0.02, 0.1, or 0.5%. According to the
22 study report, gonad weights were measured and histopathology was conducted, but no treatment-
23 related findings were noted. In a chronic (105-week) study (Til et al., 1989) in Wistar rats
24 (70/sex/group), formaldehyde was administered in the drinking water at mean actual levels of 0,
25 1.2, 15, or 82 mg/kg-day to males and 0, 1.8, 21, or 109 mg/kg-day to females; serial sacrifices
26 were conducted at 53, 79, and 105 weeks of study. At study termination (105 weeks), mean
27 testes weights were 30% increased (p < 0.01) in high-dose males as compared with controls, and
28 histopathology evaluation revealed Leydig cell tumors in treated males (incidences of 0/50, 3/50,
29 3/50, and 2/50 for the control through high-dose groups, respectively; historical control tumor
30 incidence data were not provided). The study authors did not judge these findings to be
31 treatment related. By design, none of the subacute to chronic studies included measures of
32 reproductive function (e.g., estrous cyclicity, sperm measures, or reproductive performance).
33 With the exception of Til et al. (1989), detailed mean organ weight and histopathology incidence
34 data were not provided in the published reports, and Til et al. (1988) only included tumor (not
35 non-tumor) data.
This document is a draft for review purposes only and does not constitute Agency policy.
4-305 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.2.1.7.3. Intraperitoneal studies addressing developmental and reproductive toxicity. Other
studies in which formaldehyde was administered by I.P. injection have confirmed the potential
effects on the male reproductive system.
Chowdhury et al. (1992) administered I.P. injections of 0, 5, 10, or 15 mg/kg-day
formaldehyde to Charles foster adult male rats (10/group) for 30 days. On study day 31, blood
was collected for serum testosterone measurements and the rats were sacrificed. The testes were
removed, weighed, fixed in Bouin's solution, and processed for histopathology. The study
authors reported adverse findings in all treated groups, including significant (p < 0.01) mean
body weight gains, serum testosterone levels, and testes weights. Histopathologic evaluation
revealed normal spermatogenic processes and Leydig cells in control animals. However, in
treated rats, gradual cellular degeneration in seminiferous tubules and in Leydig cells was
observed. Marked nuclear damage was noted in the 10 and 15 mg/kg-day groups, with
significantly (p < 0.001) decreased Leydig cell populations and nuclear diameters observed in all
treated groups. Additionally, a decrease in 3P-A5-hydroxy steroid dehydrogenase was noted in
the Leydig cell region of treated rat testes.
In a 30-day study performed by Majumder and Kumar (1995), 10 mg/kg-day
formaldehyde was administered I.P. to eight male Wistar rats. All animals were sacrificed at
term, and the testes, prostate, seminal vesicles, and epididymides were excised and weighed.
With the use of methodologies that were not described in the report other than by reference to the
Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus
Interaction (WHO, 1987), sperm counts, motility, and viability were compared with those of 10
controls (injected I.P. with water alone). As shown in Table 4-66, striking reductions in sperm
count and motility were noted in formaldehyde-treated rats compared with controls. Sperm
viability was also significantly reduced by formaldehyde treatment, though to a lesser overall
extent than sperm count and motility.
Table 4-66. Effect of formaldehyde on spermatogenic parameters in male
Wistar rats exposed intraperitoneally
Parameters
Control (n = 10)
Treated (n = 8)
Sperm count (106/mL)
46.30 ±5.01
20.40 ±2.01a
Sperm viability (%)
87.10 ±0.83
72.60 ±2.32a
Sperm motility (%)
75.00 ± 10.90
22.00 ± 6.40a
Significantly different from controls (p < 0.0001), as calculated by the authors.
Source: Majumder and Kumar (1995).
This document is a draft for review purposes only and does not constitute Agency policy.
4-306 DRAFT—DO NOT CITE OR QUOTE
-------
1 Majumder and Kumar (1995) also carried out an in vitro experiment in which sperm from
2 normal rats were incubated with different concentrations of aqueous formaldehyde at
3 concentrations ranging from 125 pg/mL to 2.5 |ig/mL. Viability of control sperm remained close
4 to 80% for a period of 1 hour, whereas the presence of formaldehyde dose-dependently reduced
5 viability. Thus, only 50% spermatozoa were viable for 30 minutes in the presence of 5 ng/mL
6 formaldehyde compared with 6 minutes in the presence of 500 ng/mL. Sperm motility also was
7 sensitive to the presence of formaldehyde. Less than 10% of sperm was motile for 10 minutes in
8 the presence of 125 pg/mL formaldehyde. The authors of the study considered their data to be
9 good evidence that functional parameters of spermatozoa, such as viability and motility, can be
10 adversely affected by exposure to formaldehyde. Moreover, they suggested that the cumulative
11 effects of I P. administration of formaldehyde on the male rat reproductive system raise an alert
12 that formaldehyde might impair the reproductive health of males who are occupationally exposed
13 to the compound.
14 Odeigah (1997) conducted two short-term in vivo assays to examine sperm head
15 abnormalities and dominant lethal mutations. In the sperm assessment, five daily I.P. injections
16 of 0, 0.125, 0.25, or 0.5 mg/kg formaldehyde were administered to male albino rats (six/group;
17 strain not specified). The rats were killed 3 weeks after the last injection, and epididymal sperm
18 counts and abnormalities were assessed. A dose-related decrease in sperm count was observed,
19 and significantly increased incidences of sperm head abnormalities were found at all treatment
20 levels (Table 4-67).
21
22 Table 4-67. Incidence of sperm head abnormalities in formaldehyde-
23 treated rats
24
Dose (mg/kg)
Total abnormal
sperm heads
Frequency (%) ± SEM
0
90
1.50 ± 0.11
0.125
184
3.09 ± 0.16 a
0.25
436
7.27 ± 0.30 b
0.5
514
8.57 ±0.33 b
25
26 Significantly different from controls (p < 0.05), as calculated by the authors.
27 Significantly different from controls (p < 0.001), as calculated by the authors.
28
29 Source: Odeigah (1997).
30
31
32 In the dominant lethal assay (Odeigah, 1997), five daily I.P. injections of 0, 0.125, 0.25,
33 or 0.5 mg/kg formaldehyde were administered to male rats (5 control rats, 12/treated group).
34 Subsequently, each male was mated with two untreated virgin females per week for
35 3 consecutive weeks. The females were killed 13 days after the midpoint of the mating period
This document is a draft for review purposes only and does not constitute Agency policy.
4-307 DRAFT—DO NOT CITE OR QUOTE
-------
1 and evaluated for live and dead uterine implants. In general, the number of live embryos was
2 decreased with treatment, and the number of dead implants was increased (Table 4-68).
3 Additionally, there was a reduction in fertile matings in females mated 1-7 days after the males
4 had been treated. This study did not assess the number of corpora lutea and therefore precluded
5 the determination of preimplantation loss. Nevertheless, it is indicative of dominant lethal
6 effects on the male germ cells.
7
8 Table 4-68. Dominant lethal mutations after exposure of male rats to
9 formaldehyde
10
Dose
(mg/kg)
Time of
mating
(days)
Fertile
matings3 (%)
Implants per
female3
(mean ± SE)
Live embryos
per female
(mean ± SE)
Dead implants
per female
(mean ± SE)
Dominant
lethal
mutation
indexb
0
0-21
96.67 (29)
7.86 ± 0.2 (29)
7.43 ±0.3
0.43 ±0.8
0
0.125
1-7
75.0(18)
7.18 ±0.3 (18)
5.95 ±0.2
1.23 ±0.5
19.92
8-14
79.17(19)
7.38 ±0.5 (19)
6.30 ±0.5
1.08 ±0.3
15.21
15-21
91.67 (22)
7.68 ± 0.2 (22)
6.89 ±0.3
0.79 ±0.5
7.27
0.25
1-7
33.33 (8)
5.75 ± 0.3 (8)
2.05 ±0.3
3.70 ±0.4
72.41
8-14
50.0 (12)
6.60 ±0.2 (12)
3.91 ±0.2
2.69 ±0.2
47.38
15-21
87.5 (21)
7.25 ±0.4 (21)
6.63 ±0.3
0.62 ±0.5
10.77
0.5
1-7
25.0 (6)
5.05 ± 0.03 (6)
1.10 ±0.5
3.95 ±0.22
85.20
8-14
29.17(7)
5.27 ±0.01 (7)
1.50 ±0.6
3.77 ±0.28
79.81
15-21
83.33 (20)
7.08 ± 0.04 (20)
5.79 ±0.4
1.29 ±0.17
22.07
11
12 "Number of females with implants presented in parentheses.
13 bDominant lethal mutation index:
14 Index = 1 - (Live implants experiment group per female) x 100
15 (Live implants of control group per female)
16
17 Source: Odeigah (1997).
18
19
20 4.2.1.7.4. Dermal exposure studies addressing developmental and reproductive toxicity. In a
21 study designed to assess the embryotoxic effects of dermal exposure to formaldehyde, Overman
22 (1986) applied 0.5 mL of a 37% formaldehyde solution directly to the dorsal skin of female
23 Syrian golden hamsters (four-six/group) on GDs 8, 9, 10, or 11 for 2 hours. To prevent
24 grooming during the treatment period, the animals were anesthetized with Nembutal. At the end
25 of the 2-hour treatment period, the application site was washed thoroughly to remove any
26 remaining formaldehyde. The dams were terminated on GD 15 (i.e., one day prior to expected
27 delivery, since the typical gestation period for the Syrian golden hamster is 16-18 days). The
28 fetuses were removed and fixed in either Bouin's solution or 95% ethanol for visceral or skeletal
29 evaluation, respectively. The uteri were examined for implantation sites. Fixed fetuses were
This document is a draft for review purposes only and does not constitute Agency policy.
4-308 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
weighed, measured (crown-rump), and examined for external abnormalities; fetuses that had
been placed in Bouin's fixative were evaluated for visceral anomalies by using a free-hand
sectioning technique, and those that were placed in ethanol were macerated, stained, and cleared
for skeletal examination. In this study, the dams exhibited signs of dermal irritation and
irritability, but the author reported no treatment-related effects on maternal body weight gain.
The percent of resorption sites was increased (although not significantly) in treated litters as
compared with control (0, 4.2, 8.1, 4.6, and 3.2% resorbed implantation sites for control and
GDs 8, 9, 10, and 11 treatment groups, respectively). No treatment-related effects on fetal
weight, length, or visceral or skeletal development were observed.
4.2.1.7.5. Summary of reproductive and developmental effects. Formaldehyde exposures up to
40 ppm 6 hours/day from GDs 6-15 or 6-20 did not result in external or internal malformations
(Martin, 1990; Saillenfait et al., 1989). Martin (1990) reported delayed skeletal ossification and
dose-dependent decreases in fetal body weight at 5 ppm. Formaldehyde exposure at 40 ppm to
pregnant female Sprague-Dawley rats reduced fetal body weights in male and female progeny
and in male pups of dams exposed to 20 ppm formaldehyde (Saillenfait et al., 1989). Based on
these studies (Table 4-69), the LOAEL for developmental effects in rats is 5 ppm, with a
NOAEL of 2 ppm for decreased fetal weight and delayed skeletal ossification, based on
inhalation exposures during GDs 6-20.
Developmental studies during earlier gestational windows of inhalation exposure to
formaldehyde have reported additional adverse health effects, including delayed ossification,
changes in relative organ weight, undescended testes, biochemical changes (e.g., ascorbic acid
and nucleic acids), and blood acidosis (Senichenkova and Chebotar, 1996; Senichenkova, 1991;
Kilburn and Moro, 1985; Gofmekler and Bonashevskaya, 1969; Gofmekler, 1968; Pushkina et
al., 1968). Kitaev et al. (1984) hypothesized that formaldehyde may affect reproductive function
by stimulating the hypothalamo-pituitary-gonadal (HPG) axis based on their observations of
increased ovary weight, increased number of ovulating cells, and changes in blood levels of
gonadotropins (LH and FSH). Evidence of preimplantation loss, which may be related to HPG
disruption, was observed in this study and by Sheveleva (1971). Additional studies are needed to
better understand developmental effects of formaldehyde exposure during early gestational
windows.
The prenatal developmental toxicity of oral and dermal exposures to formaldehyde has
not been thoroughly studied. Reductions in postnatal growth in beagle pups was observed by
Hurni and Ohder (1973) following in utero exposure to 125 ppm maternal dietary formaldehyde
during GDs 4-56 in beagle dogs. However, gavage dosing during gestation of mice to overtly
This document is a draft for review purposes only and does not constitute Agency policy.
4-309 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
maternally toxic doses (Seidenberg and Becker, 1987; Marks et al., 1980) (Table 4-70) and
dermal application during gestation to hamsters at a dose that caused dermal irritation and
irritability (Overman, 1986) did not result in any observed fetal toxicity (Table 4-71).
Few studies identified effects on maternal toxicity or female reproductive capacity. As
summarized in Table 4-69, exposure of rat dams at 10-40 ppm formaldehyde during pregnancy
has been shown to result in significantly decreased weight gain (Martin, 1990; Saillenfait et al.,
1989; Kilburn and Moro, 1985). Maronpot et al. (1986) reported endometrial hypoplasia with a
lack of ovarian luteal tissue in female rats exposed at 40 ppm but not at 20 ppm. Changes in LH
and FSH levels were reported in dams exposed to 0.41 ppm formaldehyde by Kitaev et al.
(1984), establishing an unbounded LOAEL for maternal toxicity.
Studies designed to assess male reproductive system endpoints in rats following repeated
inhalation exposures to formaldehyde have shown concentration-dependent decreases in Leydig
cell number and quality, effects on seminiferous tubules, decreases in testes weight, alterations in
sperm measures, decreased testosterone levels, alterations in trace metals in the testes, and/or
dominant lethal effects (Zhou et al., 2006; Ozen et al., 2005, 2002; Zhou et al., 2006Sarsilmaz et
al., 1999) (Table 4-72). Based on available studies, the LOAEL for changes in the male
reproductive system in rats following 5 days/week of inhalation exposures is 5 ppm for 3 months
of daily exposures and 10 ppm for 4 weeks of daily exposures; these dose levels are unbounded.
Abnormal sperm were also noted in mice at an inhalation dose of 16.9 ppm 2 hours/day, 6
days/week for 13 weeks (Xing et al., 2007), but, in contrast, Maronpot et al. (1986) reported no
histologic abnormalities in male mice after formaldehyde exposures at 40 ppm 6 hours/day, 5
days/week for 13 weeks. Varied results among studies may be due to species differences or
differences in methods. Although several oral subchronic and chronic studies with formaldehyde
did not identify effects on the testes (Tobe et al., 1989; Til et al., 1988; Johanssen et al., 1986),
Cassidy et al. (1983) observed spermatogenic abnormalities after a single oral dose of 200 mg/kg
to rats, and a chronic drinking water study in rats (Til et al., 1989) reported low incidences of
Leydig cell tumors in all treated groups, compared with none in control (Table 4-73).
Additionally, studies utilizing I.P. injection of formaldehyde in rats have demonstrated testes and
sperm anomalies (Majumder and Kumar, 1995; Chowdhury et al., 1992) and dominant lethal
effects (Odeigah, 1997) (Table 4-74).
This document is a draft for review purposes only and does not constitute Agency policy.
4-310 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
S"4
>3*
&
to
s
Reported study findingsb
LO AEL/N O AELC
Species,
strain, sex
n/Group
Dose; time of
treatment"
Maternal
Offspring
Maternal
Offspring
Reference
Rat, strain
12
0, 0.01, or 0.81 ppm
At 0.01 and 0.81 ppm:
Fetuses:
L: 0.01 ppm
L: 0.01 ppm
Gofmekler
NR, female
(reported as 0.012,
and 1 mg/m3)d;
continuous dosing
10-15 days prior to
mating and during
gestation
t pregnancy duration
(dose-dependent data
not shown)e
At 0.01 and 0.81 ppm:
i fetuses/dam (dose dep., data
not shown)6
t body wt (dose dep., stat. sig.)
i lung and liver wt (dose dep.,
stat. sig.)
t adrenal wt (dose dep., stat.
sig.)
At 0.81 ppm: f thymus and
kidney wt (stat. sig.)
N: ND
N: ND
(1968)
Rat,
12
0, 0.01, or 0.81 ppm
NE
Age of assessment NR.
NEe
N: 0.01 ppm
Gofmekler and
"albino"
strain NR,
female
(reported as 0.012,
and 1 mg/m3)d;
continuous dosing
10-15 days prior to
mating and during
gestation
At 0.81 ppm:
histologic effects in liver (e.g., t
extramedullary hematopoietic
centers), kidney (e.g., t
polymorphism of renal epithelial
cell nuclei)
and thymus
L: 0.08 ppm
Bonashevskaya
(1969)fh
Rat, strain
4
Inhalation and
No effects
No effects
NDg
ND
Guseva (1972)
NR, male
drinking water
co-exposure: 0;
0.10 ppm plus
0.005 mg/L water;
0.20 ppm plus
0.01 mg/L water; or
0.41 ppm plus
0.1 mg/L water; all
treatments
4 hours/day,
5 days/week for
6 months
o
2 »
5 s
to
o
{
>{
JS*
*
4^
U>
-------
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
S"4
>3*
s
Species,
strain, sex
n/Group
Dose; time of
treatment"
Reported study findingsb
LO AEL/N O AELC
Reference
Maternal
Offspring
Maternal
Offspring
Rat, strain
NR, female
NR
Expt 1:
0 or 30 ppmd
Expt 2: 0, pair-fed
control (15, 10, or
5 days), or 30 ppm;
8 hours/day for
15 days (GDs 3-17),
10 days (GDs 3-12),
5 days (GDs 8-12),
or 3 days (GDs 9-
11)
At 30 ppm
50% mortality (10 and
15 day exp)
i wt gain (duration
dep.; 3, 5, 10, and 15
day exp.)
i wt of liver, kidney,
spleen and thymus
t wt of lung and
adrenal6
Fetuses:
At 30 ppm
i fetal wt and growth (duration
dep., 10 and 15 day exp.)
t dev. defects (undescended
testes, large hearts, small
thymuses, small lungs)6
N: ND
L: 30 ppme
N: ND
L: 30 ppme
Kilburn and
Moro (1985)f Ab
Rat, Wistar,
female
Embryo
dev expt:
5-9/group
(42 adult
animals);
maternal
effects:
NR (200
adult
females
total)1
0, 0.4, or 1.2 ppm
(converted from
reported 0, 0.5 or
1.5 mg/m3);
4 hours/day,
5 days/week for
4 months; exposed
females mated to
unexposed males on
120th day exp.
At 0.4 ppm:
t wt of ovaries (stat.
sige)
t LH level (stat. sig.e)
At 1.2 ppm:
t FSH level in blood
(stat. sig.; nonsig. at
0.4 ppme)
At 0.4 ppm:
t no. of embryos and 2
blastomere stage embryos (stat.
sig. in 2nd day preg.)
At 1.2 ppm:
t no. degenerating embryos
(stat. sig. in 3rd day preg.)
L: 0.4 ppm
N: ND
L: 0.4 ppm
N: ND
Kitaev et al.
(1984)8
Rat,
Sprague-
Dawley,
female and
offspring of
both sexes
25
0 (air control
group), 0 (room
control group), 2, 5,
or 10 ppm;
6 hours/day
GDs 6-15.
Exposed females
mated to unexposed
males
At 10 ppm:
i food consumption
(stat. sig.)
i wt gain (stat. sig.)
At 5 and 10 ppm:
Fetuses:
t incidence of reduced
ossification of pubic and ischial
bones (stat. sig. compared with
air control group)
i fetal wts (nonsig.)
t litter size (nonsig.)
L: 10 ppm
N: 5 ppm
L: 5 ppm
N: 2 ppm
Martin (1990)et
o
S ?
a, Co
8- a
TO Sj-
§ I»
? f
s >•
§ 3
*o»l.
oo
to
-------
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
S"4
>3*
s
Species,
strain, sex
n/Group
Dose; time of
treatment"
Reported study findingsb
LO AEL/N O AELC
Reference
Maternal
Offspring
Maternal
Offspring
Rat, "white"
strain NR,
female and
offspring of
both sexes
5-12 (NR
for
formalde-
hyde only)
0, 0.01, or 0.81 ppm
(reported as 0.012
and 1 mg/m3)d;
continuous
10-15 days prior to
mating through
gestation
At 0.01 and 0.81:
i vit. C level in liver
(stat. sig.)
i vit. C level in
placenta
(nonsig.)
Fetuses:
At 0.01 and 0.81 ppm:
i fetuses/female6
t body wt and organ wt (data not
showne)
i vit. C level in whole fetus (stat.
sig.)
At 0.01 ppm:
i vit. C level in fetal liver (stat.
sig.)
L: 0.01 ppm
N: ND
L: 0.01
N: ND
Pushkina et al.
(1968)f
Rat,
Sprague-
Dawley,
female and
offspring of
both sexes
25
0 (air control), 5, 10,
20, 40 ppm; 6
hours/day,
GDs 6-20.
Exposed females
mated to unexposed
males.
GD 21 dams:
At 5 ppm:
t absolute body wt gain
(5 ppm only)
At 40 ppm:
i body wt gain GDs 6-
21 (stat. sig.)
i absolute body wt gain
(stat. sig., dose-
dependent trend 20 and
40 ppm)
GD 21 fetuses:
At 20 and 40 ppm:
i fetal body wt, male (stat. sig.)
At 40 ppm:
Delayed ossification of thoracic
vertebrae (stat. sig., trend
20 ppm)
t unossified sternebrae (nonsig.
at 40 ppm)
i fetal body wt, female (stat.
sig.)
L: 40 ppm
N: 20 ppm
L: 20 ppm
N: 10 ppm
Saillenfait et al.
(1989)
o
S ?
a, Co
8- a
TO Sj-
§ I»
? f
s >•
§ 3
*o»l. x-
4^
U>
U>
-------
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
Cl
S
Species,
strain, sex
n/Group
Dose; time of
treatment"
Reported study findingsb
Maternal
Offspring
LO AEL/N O AELC
Maternal
Offspring
Reference
o
5 ^ §
6* to'
to Sj-
§ ^
s ^
§ 3
*¦>1.
Rat,
mongrel
white,
female and
offspring of
both sexes
NR1
0 or 0.41 ppm
(reported as 0.5
mg/m3)
formaldehyde
(also a 3rd group of
gasoline exposure,
not described in this
table); 4 hours/day
GDs 1-19
Dams GD 20:
i corpora lutea
(nonsig.), embryos dead
before implantation (not
stat. sig.), and
implanted embryos
(nonsig.)
t blood pC02 (stat.
sig.)
Fetuses (GD 20):
Stat. sig. findings include
t fetal wt
t litters w/internal organ
anomalies
i fetuses w/ossification centers
in hyoid bone
t metacarpal bone centers
t metatarsal bone centers
t developmental defects
t blood pC02 and pOj
Pups: i pup wt
Dev. delays (data not shown)
L: 0.41 ppm
N: ND
L: 0.41 ppm
N: ND
Senichenkova
(1991)
Mouse,
mongrel,
female and
offspring of
both sexes
NR (254
dams)1
0 + ethyl alcohol;
0.41 ppm
formaldehyde;
0.41 ppm
formaldehyde +
bipyridyl;
4 hours/day
GDs 1-19.
Induced maternal
iron deficiency
anemia by I.P.
bipyridyl injections
on GDs 12-15;
controls injected
w/25% ethyl
alcohol.
Dams GD 20:
formaldehyde alone:
t blood pCC>2 (stat.
sig.)
formaldehyde +
bipyridyl:
i blood acid metabolic
products (stat. sig.)
i blood true
bicarbonates and C02
conc. (stat. sig.)
Fetuses (GD 20):
Formaldehyde alone:
t cryptorchidism
Formaldehyde + bipyridyl:
t birth defects (stat. sig.)
i dev. delay (stat. sig.)
i blood acid-base measures of
embryos (stat. sig.)
L: 0.41 ppm
N: ND
L: 0.41 ppm
N: ND
Senichenkova
and Chebotar
(1996)
oo
-------
K
s
TO
>3
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
S"4
>3*
&
to
s
Species,
strain, sex
n/Group
Dose; time of
treatment"
Reported study findingsb
LO AEL/N O AELC
Reference
Maternal
Offspring
Maternal
Offspring
Rat,
mongrel,
white,
female and
offspring of
both sexes
15/group
terminated
GD 20,
6/group
littered
0, 0.0005, or
0.005 mg/L (0, 0.4,
or 4 ppm), GDs 1-
19, 4 hours/day
At 0.4 ppm: J, leukocyte
counts
At 4 ppm: J, leukocyte
counts; reduced
threshold of
neuromuscular
excitability, j rectal
temperature, j blood
hemoglobin; f
spontaneous mobility
At 0.4 ppm: f preimplantation
loss; at 1 mo. of age, j
spontaneous mobility; at 2 mo.
of age, i hemoglobin levels and
leukocyte counts
At 4 ppm: f preimplantation
loss; at 1 and 2 mo. of age, j
spontaneous mobility; at 2 mo.
of age, i hemoglobin levels and
leukocyte counts
L: 0.4 ppm
N: ND
L: 0.4 ppm
N: ND
Sheveleva
(1971)
Rat,
Sprague-
Dawley,
female and
offspring of
both sexes
6 dams
0 or 6 ppm
8 hours/day,
6 weeks, starting at
GD 1, PND 1,
4 weeks of age, or
adult age
NE
In offspring exposed in utero and
during early postnatal life: j
mean BW and liver weight; t
markers of oxidative stress
In offspring exposed as
juveniles: jmean liver weight; t
markers of oxidative stress
In offspring exposed only as
adults: no effect
NE
L: 6 ppm
N: ND
Kum et al.
(2007)
Rat,
Sprague-
Dawley,
female and
offspring of
both sexes
6 dams
0 or 6 ppm
8 hours/day;
6 weeks, starting at
GD 1, PND 1,
4 weeks of age, or
adult age
NE
In offspring exposed in early
postnatal life, as juveniles, or as
adults, t ANAE-positive T-
lymphocytes in BALT
In all exposure initiation groups,
t ANAE-positive lymphocyte
ratios
NE
L: 6 ppm
N: ND
Sandikci et al.
(2007)
o
2 »
5 s
a, Co'
§- a
TO Sj-
§ ^
>!
JS*
*
4^
u>
Lti
-------
K
s
TO
>3
Table 4-69. Summary of reported developmental effects in formaldehyde inhalation exposure studies
S"4
>3*
&
to
s
Species,
strain, sex
n/Group
Dose; time of
treatment"
Reported study findingsb
LO AEL/N O AELC
Reference
Maternal
Offspring
Maternal
Offspring
Rat, Wistar,
female and
offspring of
both sexes
12-14
dams
0, 6, or 12 ppm,
6 hours/day;
5 days/week,
30 days
NE
At 6 and 12 ppm, at postnatal
days 30 and 90: respiratory
irritation and toxicity; deer. BW,
FC, WC; | SOD activity, j
levels of copper and iron levels
in lungs, t zinc levels in lungs
NE
L: 6 ppm
N: ND
Songur et al.
(2005)
o
2 »
5 s
to
o
S
>S
TO
TO'
*
ND: not determined; NE: not evaluated; NR: not reported; Ab: abstract only; wt: weight; stat. sig.: statistically significant; p: pressure; 1: length;
To convert concentrations in air (at 25°C) from mg/m3 to ppm: 1 ppm = 1.23 mg/m3; 1 mg/m3 = 0.813 ppm.
Treatment is given as the formaldehyde concentration in air (ppm) with the length of exposure each day and the duration of treatment in days, as available.
bStudies with negative findings are included.
°L: LOAEL; N: NOAEL.
dExposure concentrations not validated; details of formaldehyde vapor generation not reported; exposure during gestation not well characterized in study report.
eNo statistics provided.
fLack of study details.
8See Table 4-72 for reproductive effects.
hGofmekler and Bonashevskaya (1969) seem to report on different findings from the same study (i.e., same animals) as Gofmekler (1968).
'Number/group not clear from study report.
oo
On
-------
is*
Table 4-70. Summary of reported developmental effects in formaldehyde oral exposure studies
Species,
strain, sex
n/
Group
Dose; time of
treatment
Reported developmental effects a
LO AEL/N O AELb
Reference
Maternal
Offspring
Maternal
Offspring
Dog,
beagle,
female and
pups
9-10
0, 125, or
375 ppm, dietary,
GDs 4-56
No effects
At 125 and 375 ppm:
i birth wt and wt gain
through postnatal week 8
L: ND
N: 375 ppm
L: 125 ppm
N: ND
Hurni and
Ohder (1973)
Mouse,
CD-I,
female
29-35
total
0, 74, 148, or
185 mg/kg-day,
GDs 6-15
(aqueous
formaldehyde
solution
contained 10-
15% methanol)
At 185 mg/kg-day:
Mortality
No effects at GD 18
L: 185 mg/kg-day
N: 148 mg/kg-day
L: ND
N: 185 mg/kg-day
Marks et al.
(1980)
Mouse,
ICR/SIM,
female
26-30
total
0 or 540 mg/kg-
day, GDs 8-12
At 540 mg/kg-day:
Mortality
No effects in pups on PND 1
and 3
L: 540 mg/kg-day
N: ND
L: ND
N: 540 mg/kg-day
Seidenberg and
Becker (1987)
ND: not determined.
"Studies with negative findings are included.
bL: LOAEL; N: NOAEL.
-J
-------
^3
K
^ i?
>3
§•
s
Table 4-71. Summary of reported developmental effects in formaldehyde dermal exposure studies
o
>3
0
1
a,
>3
o
o
s
tr>
S?
r?
Species,
strain, sex
n/
Group
Dose; time of
treatment"
Reported developmental effects
LOAEL/NOAEL'
Reference
Maternal
Offspring
Maternal
Offspring
Hamster,
Syrian
golden,
female
4-6
0 or 37%; 0.5 mL
applied to dorsal skin
(hair clipped) for 2 hours
then washed; GDs 8, 9,
10, or 11
Signs of dermal irritation
and irritability
At all GDs of treatment, t
percent resorptions (not sig.)
L: 37%
N: ND
L: 37%
N: ND
Overman (1986)
58
o
3*
S5
<§>
>{
>{
JS*
*
ND, not determined
aL: LOAEL; N: NOAEL.
4^
u>
00
-------
Table 4-72. Summary of reported reproductive effects in formaldehyde inhalation studies
Species,
strain, sex
n /
Group
Dose; time of treatment"
Reported reproductive effects'3
LOAEL/
NOAELc
Reference
Rat, strain
NR, male
4
Inhalation plus drinking water
co-exposure: 0; 0.10 ppmplus
0.005 mg/L water; 0.20 ppm
plus 0.01 mg/L water; or 0.41
ppm plus 0.1 mg/L water; all
treatments; 4 hours/day,
5 days/week for 6 months.
Exposed males mated to
unexposed females
At 0.20 ppm + 0.01 mg/L water and 0.41 ppm + 0.1 mg/L water:
i nucleic acid in testes (dose dep.; data not shown; stat. sig.)
L: 0.20 ppm
+ 0.01 mg/L
water
N: ND
Guseva
(1972)4e
Rat,
Wistar,
female
NR
(200
female)
0, 0.4, or 1.2 ppm;
4 hours/day, 5 days/week for
4 months.
Exposed females mated to
unexposed males on 120th day
of exposure.
At 0.4 ppm:
t wt of ovaries (stat. sig.e)
t LH level in blood (stat. sig. at 0.4 ppmf)
At 1.2 ppm:
t FSH level in blood (stat. sig., dose dep. trendf)
L: 0.4 ppmh
N: ND h
Kitaev et al.
(1984)e
Mouse,
B6C3F1,
male and
female
10
0, 2, 4, 10, 20, or 40 ppm;
6 hours/day, 5 days/week
for 13 weeks
Males and females:
At 20 ppm: J, wt gain
At 40 ppm: f mortality (13 weeks exp.); J, wt loss
Females:
At 40 ppm: t Uterine endometrial and ovarian hypoplasia
L: 20 ppm
N: ND
Maronpot et
al. (1986)
Rat, albino
Wistar,
male
7
6 groups: 0, 10, or 20 ppm;
8 hours/day, 5 days/week for
4 weeks (subacute) or
13 weeks (subchronic)
At 10 and 20 ppm (both durations):
i wt gain (stat. sig., dose dep.)
i relative testes wt (stat. sig., dose and conc. dep.)
1 zinc and copper in testes (stat. sig., dose and conc. dep.)
t iron in testes (stat. sig., dose and conc. dep.)
No effect on testes wt.
L: 10 ppm
N: ND
Ozen et al.
(2002)f
Rat,
Wistar,
male
18
0, 5, or 10 ppm; 8 hours/day,
5 days/week, 91 days
At 5 and 10 ppm:
clinical signs of respiratory irritation, I BW, FC, WC; | scrum
testosterone; j mean seminiferous tubule diameters; t localization
of heat shock protein 70 in cytoplasm of spermatogonia,
spermatocytes, and spermatids
L: 5 ppm
N: ND
Ozen et al.
(2005)
-------
K
s
TO
>3
Table 4-72. Summary of reported reproductive effects in formaldehyde inhalation studies
S"4
>3*
§•
to
s
Species,
strain, sex
n /
Group
Dose; time of treatment"
Reported reproductive effects'3
LOAEL/
NOAELc
Reference
Rat, albino
Wistar,
male
10
0, 10, or 20 ppm
8 hours/day, 5 days/week for
4 weeks
Dose NR:
Irritation: standing hair, interrupted breathing, f eye blinking,
licking, nose cleaning, and sneezing.
At 10 and 20 ppm:
i Body wt gain (dose dep.; stat. sig.)
i Leydig cell quantity (stat. sig.)
t Nuclear damage of Leydig cells (dose dep.; stat. sig.)
L: 10 ppm
N: ND
Sarsilmaz et
al. (1999)
Mouse,
strain not
specified,
male
12
0,21,42, or 84 mg/m3 (0,
16.9, 33.8, or 67.6 ppm);
2 hours/day, 6 days/week;
13 weeks. Exposed males
mated to unexposed females.
In all treated groups: t percentage abnormal sperm, f resorption
rate, and j live fetuses
L: 16.9 ppm
N: ND
Xing et al.
(2007)
Rat,
Wistar,
male
7
0 or 1.5 ppm, 18 weeks
FA exposures:
(1) 4 hours/day, 4 days/week
(2) 2 hours/day; 4 days/week
(3) 2 hours/day, 4 days/week
In all treated groups: sig. j seminiferous tubular diameter and
epithelial height. Other effects in exposure groups:
(1) sig. i germ cells; arrested spermatogenesis
(2) I cells, increased thickness in basal membrane
(3) | spaces between germ cells; disrupted association between
Sertoli and germinal cells
L: 1.5 ppm
N: ND
Golalipour et
al. (2007)
Rat,
Sprague-
Dawley,
male
10
(1) 0 (gavage saline);
(2) 10 mg/m3 (8.05 ppm),
12 hours/day, 2 weeks; or
(3) 10 mg/m3 (8.05 ppm),
12 hours/day, 2 weeks, plus
30 mg/kg-day vitamin E orally
At 10 mg/m3: sig. j testis weight, atrophy of seminiferous tubules,
i spermatogenic cells, disintegrated and sloughed seminiferous
cells; edematous interstitial tissue with vascular dilatation and
hyperemia; j epididymal sperm count and percentage motile
sperm, f percentage abnormal sperm; I SOD, GSH-Px, GSH and
t MDA in testes; vitamin E attenuated all effects
L: 8.05 ppm
N: ND
Zhou et al.
(2006)
Rat,
Wistar,
male
40
0, 0.1, 1, or 10 ppm;
6 hours/day, 5 days/week, 13
or 52 weeks
No effects: testis weight; histopathologic findings8
L: ND
N: 10 ppm
Appleman et
al. (1986)
o
2 »
5 s
a, Co'
§- a
TO Sj-
§ ^
>S
>S
TO
TO'
*
""Treatment is given as the formaldehyde concentration in air (ppm) with the length of exposure each day and the duration of treatment in days.
bStudies with negative findings are included.
°L: LOAEL; N: NOAEL.
dGuseva (1972) was a drinking water and inhalation study.
^ Developmental effects shown in Table 4-69.
^ Statistics not provided in study report.
to
o
-------
8Focus of study was not the reproductive system; only reproductive system findings are addressed in the table; NOAEL and LOAEL in table are based only on
^3 reproductive system findings.
^ ^ h For increased FSH, the LOAEL was 1.2 ppm (1.5 mg/m3) and the NOAEL was 0.4 ppm (0.5 mg/m3).
o
Co ~
os To convert concentrations in air (at 25°C) from mg/m1 to ppm: 1 ppm = 1.23 mg/irr1: 1 mg/m1 = 0.813 ppm.
5" ND: not determined; NR: not reported.
I § i
S- a
5 ^ Sj.
_ ^
O >!
s >•
§ 3
oo
to
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-73. Summary of reported reproductive effects in formaldehyde oral studies
Species,
strain, sex
n /
Group
Dose; time of treatment
Reported reproductive effects3
LOAEL/
NOAELb
Reference
Rat,
Wistar,
male
5 (20
control)
0, 100, or 200 mg/kg, single
gavage dose
At 200 mg/kg:
t (19%) testicular sperm head counts (stat. sig.)
t (5%) abnormal sperm head (stat. sig.)
L: 200 mg/kg
N: 100 mg/kg
Cassidy et al.
(1983)
Rat,
Sprague-
Dawley,
both sexes
15
0, 50, 100, or 150 mg/kg-day,
drinking water; 91 days
No effects: absolute or relative gonad weights;
histopathologic findings of reproductive organs0
L: ND
N: 150 mg/kg-day
Johanssen et
al. (1986)
Dog,
beagle,
both sexes
4
0, 50, 75, or 100 mg/kg-day,
dietary; 91 days
No effects: absolute or relative gonad weights;
histopathologic findings of reproductive organs0
L: ND
N: 100 mg/kg-day
Rat,
Wistar,
both sexes
10
0, 25, or 120 mg/kg-day,
drinking water, 4 weeks
No effects: gonad weights; histopathologic findings of
reproductive organs0
L: ND
N: 120 mg/kg-day
Til et al.
(1988)
Rat,
Wistar,
both sexes
70
0, 1.2, 15, or 82 mg/kg-day
(males), 0, 1.8, 21, or
109 mg/kg-day (females),d
drinking water; 105 weeks
In all treated groups:
Leydig cell tumors observed at 105 weeks of study0
At 82 mg/kg-day:
t mean testes weights
L: 1.2 mg/kg-day
N: ND
Til et al.
(1989)
Rat,
Wistar,
both sexes
20
0,0.02, 0.1, or 0.5% in
drinking water; 24 months
No effects: gonad weights; histopathologic findings of
reproductive organs0
L: ND
N: 0.5%
Tobe et al.
(1989)
o
2 »
5 s
to
o
!
>{
JS*
*
ND, not determined.
aStudies with negative findings are included.
bL: LOAEL; N: NOAEL.
°Focus of study was not the reproductive system; only reproductive system findings are addressed in the table; NOAEL and LOAEL in table are based only on
reproductive system findings.
dActual concentrations.
oo
to
to
-------
K
s
TO
>3
S"4
>3*
&
S
Table 4-74. Summary of reported reproductive effects in formaldehyde intraperitoneal studies
Species,
strain, sex
n /
Group
Dose; time of treatment"
Reported reproductive effects3
LOAEL/
NOAELb
Reference
Rat,
Charles
foster,
male
10
0 or 5 mg/kg-day; 30 days
i body weight gain
i Leydig cell population and cell nuclear diameter
i serum T levels
i testes weights
cellular degeneration of seminiferous tubules
L: 5 mg/kg-day
N: ND
Chowdhury et
al. (1992)
Rat,
Wistar,
male
8
0 or 10 mg/kg-day; 30 days
i sperm count, motility and sperm viability
L: 10 mg/kg-day
N: ND
Majumder and
Kumar (1995)
Rat,
"albino"
strain NR,
male
6
0,0.125,0.250, or
0.60 mg/kg-day; 5 days
At all treatment levels:
i sperm count and t sperm head abnormalities (3 weeks
after the last injection)
L: 0.125 mg/kg-day
N: ND
Odeigah
(1997)
Rat,
"albino"
strain NR,
male
12
0,0.125,0.250, or 0.60
mg/kg-day; 5 days. Exposed
males mated to unexposed
females.
At all treatment levels (at GD 13):
Delayed time to mating
i mean no. implants and live embryos
t dead implants and dominant lethal index following
mating to untreated females
L: 0.125 mg/kg-day
N: ND
o
2 »
5 s
to
o
!
>{
JS*
*
aStudies with negative findings are included.
bL: LOAEL; N: NOAEL.
ND: not determined; NR: not reported; T: testosterone.
4^
U>
to
U>
-------
1
2
3
4
5
6
7
8
9
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
4.2.2. Carcinogenic Potential: Animal Bioassays
Chronic animal studies (inhalation and oral exposures) chronicle tumor incidence in a
variety of rodent models. Study descriptions are provided above in detail (Section 4.2.1,
Table 4-34). The study results are evaluated here for both routes of exposure in context of how
they inform the carcinogenic potential for the three major affected systems: respiratory tract, GI
tract, and LHP system. Experimental design and implementation must be carefully considered
when interpreting study results. For example, some key factors involved in evaluating cancer
bioassays include study size, organs/tissues examined, and study length (especially for late-in-
life tumors).
4.2.2.1. Respiratory Tract
In the respiratory tract, only nasal tumors are considered formaldehyde induced in rodent
studies. The majority of studies were conducted using rats (F344, Wistar, or Sprague-Dawley),
and all studies of 18 months or greater in mice and rats show evidence of formaldehyde-induced
nasal carcinogenicity. The nasal tumors are primarily SCCs, although papillomas, polypoid
adenoma, adenocarcinoma, fibrosarcoma, and esthesioneuroepithelioma have been reported
(Kamata et al., 1997; Monticello et al., 1996; Morgan et al., 1986a, b; Takahashi et al., 1986;
Sellakumar et al., 1985; Kerns et al., 1983; Albert et al., 1982). Although hyperplasia, dysplasia,
and squamous metaplasia of the respiratory epithelium have been observed beyond the nasal
cavity, other respiratory tract tumors have not been significantly increased by formaldehyde
exposure alone.
Increased tumor incidence and decreased latency are correlated with increasing
formaldehyde exposure concentration. Reviewing data from the only lifelong inhalation study
with multiple exposure groups, SCC is first noted at 8 and 9 months for high exposed (15 ppm)
female and male F344 rats autopsied as "early deaths" prior to the 12 month sacrifice, with an
incidence of 43% over the course of the study (unadjusted for mortality) (Kerns et al., 1983). In
contrast only two SCCs were found in male and female rats sacrificed after 24 months of
exposure (incidence of SCC 2.5% at 24 months) (Kerns et al., 1983). In a follow-up study by
Monticello et al. (1996), the incidence of SCC in rats exposed at 15 ppm was 47% with the first
tumor noted at 12 months. The incidence of SCC in male rats exposed at 10 ppm was 22% with
the first SCC noted at 18 months. Moreover, of 90 rats exposed at 6 ppm for 20 months only one
SCC was noted. No SCCs were detected in rats exposed at 0.7 or 2 ppm formaldehyde. These
incidence rates are not mortality adjusted and include animals from each scheduled sacrifice (3,
6, 12, and 18 months). In a lifelong study of male Sprague-Dawley rats exposed at 10 ppm
formaldehyde, the cumulative nasal tumor incidence was calculated as a function of time of
This document is a draft for review purposes only and does not constitute Agency policy.
4-324 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
exposure (Figure 4-27) (Sellakumar et al., 1985). After 2 years of exposure, the probability of
nasal carcinoma was greater than 60%.
O
~ 10 ppm HCI + I4ppm HCHO {Premix)
14 ppm HCHO
Ar- • -A tO ppm HCI + 14ppm HCHO (Not Premix)
C
o
" too
s
IT'
40
20
0
v/-1—
240
720
840
480 600
Days After First Exposure
360
Figure 4-27. Mortality corrected cumulative incidences of nasal carcinomas
in the indicated exposure groups.
Source: Sellakumar et al. (1985).
There is some evidence that less-than-lifetime exposure to formaldehyde can induce nasal
tumors over an extended observation period. Two studies, both in male Wistar rats, report nasal
tumors in response to less-than-lifetime exposures (Woutersen et al., 1989; Feron et al., 1988).
A 13-week exposure at 20 ppm resulted in four nasal tumors (three SCCs), a cystic SCC of the
nasolacrimal duct, and an epithelial tumor on the mandible, for a total of six tumors observed
over 30 months of observation (Feron et al., 1988). No tumors were noted in 13-week controls.
A limited number of formaldehyde-related tumors were noted due to 4 or 8 weeks of exposure
followed by 30 months of observation. Although the tumor incidence of these less-than-lifetime
exposures is low, this is consistent with the 2-year bioassays in Wistar rats. Wistar rats are more
resilient to formaldehyde-induced nasal toxicity than F344 or SD rats (Section 4.2.1), and only 1
of 26 (4%) Wistar rats exposed at 10 ppm for 28 months developed SCC (Woutersen et al.,
1989) versus 22% in F344 rats (Monticello et al., 1996).
Woutersen et al. (1989) also examined the effect of severe nasal damage from
electrocoagulation on formaldehyde-induced SCC in Wistar rats. Nasal tumors were noted in
formaldehyde-exposed rats without damaged noses (exposed for only 3 months and observed for
This document is a draft for review purposes only and does not constitute Agency policy.
4-325 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
25 months). However, the low incidence of tumors in each treatment group (1/26, 2/60, 2/60,
1/58) indicates these data should be considered suggestive even though no SCCs were noted in
control rats with or without damaged noses (n = 83). The studies by Woutersen et al. (1989) did
demonstrate a synergistic effect of nasal damage from electrocoagulation and 10 ppm
formaldehyde exposure (3 months), where 15/58 rats had SCC versus 1/26 with undamaged
noses. The study was originally designed to examine the effect of formaldehyde on the damaged
tissue on cancer promotion. However, it is unclear if the synergistic effect of formaldehyde
exposure on damaged nasal tissue is an effect of formaldehyde on the damaged cells and joint
effects of a mutagen with regenerative proliferation from the nasal damage. It is also possible
the damaged nasal passages may alter airflow in the nasal passages, resulting in significantly
different flux of formaldehyde into the tissue.
There is a single inhalation study (Dalbey, 1982) that investigates the role of promotion
in formaldehyde-induced cancer. Although hamsters exhibit little to no effects of formaldehyde
on the nasal mucosa or other respiratory tract tissues (Rusch et al., 1983a, b; Dalbey, 1982),
DEN-induced respiratory adenomas were increased with formaldehyde exposure (10 ppm)
48 hours prior to DEN injection (but not by formaldehyde alone or formaldehyde exposure after
DEN injection). The number of tracheal tumors per TBA was doubled by formaldehyde
exposure. The study authors note that adenomas should be considered independent tumors and
that the increase in tracheal tumors is of biological significance even given the incidence of
TBAs (77%, DEN alone), was not further increased by formaldehyde exposure. It is of
particular interest that a promotion study in hamsters is positive, since so little nasal pathology
occurs with formaldehyde exposure. The absence of significant hyperplasia and tissue damage
in these animals suggests that formaldehyde may induce subtle changes in the respiratory tract
mucosa that permit formaldehyde to act as a tumor promoter.
4.2.2.2. Gastrointestinal Tract
As with the respiratory tract, the proximal portion of the GI tract exhibits formaldehyde-
induced lesions in the forestomach and glandular stomach (Soffritti et al., 1989; Til et al., 1989;
Tobe et al., 1989; Takahashi et al., 1986). However, data are mixed regarding the carcinogenic
potential of formaldehyde in the GI tract from oral exposures.
Two independent 2-year cancer bioassays in Wistar rats exposed to formaldehyde in
drinking water were both negative; they reported no tumors found at the 24-month sacrifice (Til
et al., 1989; Tobe et al., 1989). Til et al. (1989) exposed rats to a range of formaldehyde doses
(0, 1.2, 15, or 82 mg/kg-day) and evaluated 44-49 animals per sex per dose group at 24 months
of exposure. No formaldehyde-related tumors were found. A smaller study by Tobe et al.
This document is a draft for review purposes only and does not constitute Agency policy.
4-326 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
(1989) failed to note any tumors after a 2-year exposure at 0, 10, 50 or 300 mg/kg-day (eight rats
per sex per treatment group).
In contrast, two lifelong studies in male and female Sprague-Dawley rats provide support
for formaldehyde-induced GI tract tumors (Soffritti et al., 1989). Both stomach and intestinal
tumors are rare; low background rates are expected in this colony of Sprague-Dawley rats.
These studies demonstrate an increase in tumors (although rare) correlated with exposure to
formaldehyde and significantly increased susceptibility to early-lifetime exposure. The authors
provide a detailed report of the background rates of various stomach and intestinal neoplasia for
male (n = 2,677) and female (n = 2,582) rats (Soffritti et al., 1989). From this background pool,
the total incidence of benign and malignant tumors in the stomach and intestine combined is only
1.4% (combining all sites and locations). The majority of tumors are located in the stomach (1%
benign, 0.2% malignant). Usually, a very large study population is needed to detect increases in
rare tumor types. In this study, the study size of each treatment group was relatively small.
Thus, only a few TBAs are responsible for the observed increases. Additionally, a clear dose-
response relationship is not evident (perhaps due to the low incidence) despite the fact that the
greatest tumor incidence was observed in the highest treatment group. As presented above,
apparent increases in both stomach and intestinal neoplasia are noted in formaldehyde-treated
rats (ranging from 1 to 6% by type). When summed across the GI tract, tumor incidence in the
highest treatment group was 8% versus 1.4% in historical controls. Despite the limitations of
group size and lack of dose response, the findings do support the carcinogenic potential for
formaldehyde administered orally. Moreover, these findings are not inconsistent from Tobe et
al. (1989) and Til et al. (1989) because the study design is significantly different.
The second study reported by Soffritti et al. (1989) demonstrates early lifetime
susceptibility with GI tumor incidence of 21.6% in females (n = 37) and 13.9% in males (n = 36)
after exposure to formaldehyde. Sprague-Dawley rats were exposed to formaldehyde in drinking
water for 2 years (0 or 2,500 mg/L). Exposures began on GD 12 in the offspring. The most
common tumor detected was intestinal leiomyosarcoma (13.5% in female offspring) with a
background rate of 0.04% in female rats in the colony.
Soffritti et al. (1989) stands alone in supporting formaldehyde-induced GI tumors. These
findings are largely attributed to a unique study design that included lifelong observation,
neonatal exposure, examination of individual tumor types as well as combined rare tumor types
for analysis, and summation of tumors across locations. The study design results in a more
sensitive assay for rare tumors. Thus, Soffritti et al. (1989) utilized a more appropriate design
and analysis for detecting rare tumors and should not be compared with the results by Tobe et al.
(1989) and Til etal. (1989).
This document is a draft for review purposes only and does not constitute Agency policy.
4-327 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
There is evidence that formaldehyde may act as a tumor promoter by the oral route as
well as the inhalation route (discussed above). Takahashi et al. (1986) reported an increase in
MNNG-initiated GI cancers with formaldehyde exposure (29.4 versus 13.3% TBA in controls);
the greatest difference in tumor-containing versus non-tumorigenic mice was associated with
adenocarcinoma in the glandular stomach (23.5 versus 3.3% in controls). Additionally,
forestomach papillomas and preneoplastic hyperplasia in the glandular stomach were increased
with formaldehyde exposure alone.
The data indicate carcinogenic potential from formaldehyde ingestion in drinking water.
Formaldehyde may act in part as a tumor-promoting agent and shows clear increased
susceptibility from early lifetime exposures.
4.2.2.3. Lymphohematopoietic Cancer
The majority of chronic animal bioassays do not report either leukemia or lymphoma, but
many of these studies did not have adequate study length or study design to detect these
malignancies. Many studies focused the histopathology on the nasal passages and respiratory
tract (Monticello et al., 1996; Holmstrom et al., 1989a; Woutersen et al., 1989; Appleman et al.,
1988; Dalbey, 1982; Horten et al., 1963). Kamata et al. (1997) did examine additional organs,
but there were only five animals at each sacrifice. Similarly, the oral study by Takahashi et al.
(1989) focused on the stomach and intestines. Tobe et al. (1989) only included 20 Wistar rats
per group with interim sacrifices. Therefore, few studies can inform the carcinogenic potential
of formaldehyde on the LHP system. Table 4-75 lists the chronic bioassays that have the
potential to detect LHP malignancies.
Soffritti et al. (1989) first published an observation of formaldehyde-induced leukemia in
animal studies. These study results have been criticized for their combination of lymphatic
leukemia and lymphoma. However, this classification is consistent with the current WHO
classification of lymphoid malignancies in humans where adult B- and T-cell leukemias and
lymphomas are considered the same disease (Harris, 2000a). Although there may be a slight
vehicle effect, a dose response is still readily apparent among the formaldehyde-treated groups
(Figure 4-28). In contrast, the 2-year bioassay in Wistar male in female rats (Til et al., 1989) was
clearly negative for leukemia and lymphoma with only four TBAs in all treatment groups
sacrificed at 24 months. Moreover, the drinking water levels were similar at the highest dose of
both Til et al. (1989) and Soffritti et al. (1989). However, the two study designs differ in length,
which may have influenced results since leukemia is a late-in-life malignancy in rodents. Two-
year survival in the Soffritti et al. (1989) study varied between 50 and 60%. These animals were
available to develop leukemia after the 2-year window of the Til et al. (1989) study. Any
This document is a draft for review purposes only and does not constitute Agency policy.
4-328 DRAFT—DO NOT CITE OR QUOTE
-------
1 potential role of strain differences is unknown. Overall the results of Soffritti et al. (1989) are
2 strong since they indicate an exposure-response relationship in a lifelong study appropriate for
3 late-in-life malignancies. Unlike the GI tract tumors, increased LHP malignancies were not
4 associated with early-life exposure to formaldehyde in drinking water.
5
6 Table 4-75. Summary of chronic bioassays which address rodent leukemia
7 and lymphoma
8
Study
Histopathology
Endpoint
Results
Comments
Drinking water exposure
Soffritti et
Male and female Svr ague-Daw lev rats
al. (1986)
Complete histopathology
Lymphocytic
leukemia and
lymphosarcoma
Increased, showing
a dose-response
Lifelong study
High exposure of
1,500 mg/L in water
Til et al.
Male and female Wistar Rats
(1989)
Complete histopathology
in control and high-dose
group (15 ppm)
Lymphoma,
leukemia
No increase
(three lymphomas
and one leukemia
found in 200
animals at the 2-year
sacrifice)
2-year bioassay
High exposure of
approximately 1,900 mg/L
(82 mg/kg for males and
109 mg/kg for females)
Inhalation exposures
Sellakumar
Male rats, Sprasue-Dawley
et al. (1985);
Albert et al.
(1982)
Necropsy focused on
respiratory tract: also liver,
spleen, kidney, and testes
and organs demonstrating
gross pathology
Lymphoma
No increase
Lifelong study, high
mortality at 24 months
(>80%)
Battelle,
Male rats, F344
Columbus
Laboratories
(1981)
Complete histopathology
in controls and high-dose
group (15 ppm)
Leukemia, all
No increase
Extended study, high
mortality
Female rats, F344
Complete histopathology
in controls and high-dose
group (15 ppm)
Leukemia, all
Increase in
mortality-adjusted
incidence;
p = 0.00563
Extended study, high
mortality. Apparent
elevation in 2 and 6 ppm
treatment groups as well;
statistical comparison to
controls is problematic.
Female mice, C57BL/6xC3HFl
All organs in controls and
high-dose group (15 ppm)
Lymphoma, all
26% increase in
15 ppm group, 16%
in controls;
p = 0.0617
Extended study.
All mice included in
statistics conducted by
BattelleColumbus
Laboratories.
9
10 aOriginal statistical analysis provided by Battelle Columbus Laboratories. Significance set at p < 0.0167. Analysis
11 of adjusted data where time to lesion and survivorship were considered (Cox [1972] and Tarone [1975]).
12
This document is a draft for review purposes only and does not constitute Agency policy.
4-329 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Male and Female Sprague-Dawley Rats Exposed to
Formaldehyde in Drinking Water
~
~ Fpm;
i
fisnl.
i
Exposure (mg/l formaldehyde)
Figure 4-28. Leukemia incidence in Sprague-Dawley rats exposed to
formaldehyde in drinking water for 2 years.
Note: Animals were observed until natural death. The vehicle control contained
the level of methanol in drinking water for the high-dose group (1,500 mg/L).
Source: Soffritti et al. (1989).
Sellakumar et al. (1985) conducted a lifelong inhalation study in male Sprague-Dawley
rats exposed at 10 ppm formaldehyde. Organs outside of the respiratory tract were routinely
examined (liver, kidneys, and testes), including any organ exhibiting gross pathology, so there
was some ability to detect leukemia and lymphoma. However, spleen, thymus, and lymph nodes
were not routinely examined, limiting detection of smaller lesions. Although Sellakumar et al.
(1985) was a lifelong study, there was a high mortality rate at 2 years (>80% from the figure),
again limiting the power of this study to detect late-in-life malignancies. Nonetheless, this study
did not indicate formaldehyde-induced lymphoma or leukemia.
The largest and most comprehensive study of carcinogenic health effects from
formaldehyde inhalation exposures is the study conducted at the Battelle Columbus Laboratory
(1981) and reported by Kerns et al. (1980) and Swenberg et al. (1983). Although the summary
reports of this study do not discuss leukemia or lymphoma rates, mouse lymphoma and rat
leukemia were selected by the study pathologist and biostatistician for analysis (Battelle
Columbus Laboratories, 1981). Statistical analysis performed by Battelle, which accounted for
This document is a draft for review purposes only and does not constitute Agency policy.
4-330 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
time to lesion and survivorship rates, indicated a statistically significant increase in female rat
leukemia (p = 0.0003) and near significant increase in female mouse lymphoma (p = 0.06). No
trend analysis could be conducted since only gross pathology was conducted on mid-dose mice
and rats (2 and 6 ppm). EPA has further analyzed these data to better understand the significance
of these findings.
As noted in the study description (Section 4.2.1.2.3), both male and female rats at the
highest exposure (15 ppm) exhibited significant early deaths due to nasal lesions (see Figure
4-29). No female rats in the highest exposure group remained after the 24-month sacrifice
(which included only 14 animals). Nine male rats were examined after 24 months. Since the
first leukemia in rats was noted at 21 months, the early deaths prior to that time reduced the
number of animals in which the leukemia could have been observed. Unadjusted data do not
show an increase in female rat leukemia, but, when data are adjusted to account for the early
deaths, dramatically different results are apparent. In the unadjusted data, leukemia incidence is
expressed as the number of cases over total animals examined (including early deaths and
interim sacrifices). By using this methodology, there is a lower incidence of leukemia in high
exposed animals, and a slightly higher incidence of leukemia in the mid-exposed groups
compared with controls (Figure 4-30, panel A). However, when the leukemia incidence rates are
calculated only for female rats that survived to at least 21 months (the first case noted),
formaldehyde-induced increases in leukemia are evident in all treatment groups relative to
controls (Figure 4-30, panel B). These results are consistent with the original statistical analysis
conducted by the researchers at the Battelle Columbus Laboratories (1981). Although elevated
in all treatment groups, no exposure-response relationship is evident. The lack of a dose-
response relationship may in part be due to the fact that no animals in the high-exposure group
survived past 24 months. Additionally, the 6 ppm (mid-exposure) group had significant early
deaths between 21 to 24 months compared with the 2 ppm and control groups. A similar
reanalysis of data from male rats did not reveal any relationship between formaldehyde treatment
and leukemia incidence.
Male and female mice exposed to formaldehyde for 24 months did not experience the
same rate of formaldehyde-related mortality (Kerns et al., 1983). However, significant early
deaths were observed in male mice due to infighting. Therefore, the data for male mice may not
inform incidence of late-in-life tumor, such as lymphoma. As discussed above in the full study
description, full histopathology, including spleen, liver, thymus, and lymph nodes, was only done
on the control and high-exposure group (15 ppm) mice. When comparing unadjusted incidence
lymphoma-bearing mice, there is a clear elevation lymphoma in formaldehyde-exposed female
mice (28%) over controls (22%) (Figure 4-31).
This document is a draft for review purposes only and does not constitute Agency policy.
4-331 DRAFT—DO NOT CITE OR QUOTE
-------
18 mos
2
3 Figure 4-29. Unscheduled deaths in female F344 rats exposed to
4 formaldehyde for 24 months.
5
6 Source: Battelle Columbus Laboratories (1981).
7
8
This document is a draft for review purposes only and does not constitute Agency policy.
4-332 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Panel A: Unadjusted Data
Panel B: Adjusted Data
18 mo 21 Mo 24 mo 27 mo 30 I
-0 ppm —US— 2 ppm 6 ppm
21 mos 24 mos 27 mos 30 mos
- Oppm as 2 ppm A 6 ppm 15 ppm
Figure 4-30. Cumulative leukemia incidence in female F344 rats exposed to
formaldehyde for 24 months.
Note: Panel A shows the unadjusted data where incidence rates include all
scheduled sacrifices and early deaths up to the time point shown. Panel B shows
incidence of leukemia only in rats who survived at least 21 months.
Source: Battelle Columbus Laboratories (1981).
This document is a draft for review purposes only and does not constitute Agency policy.
4-333 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
30%
~ Control
28%
25%
20%
15%
10%
0%
5%
¦
12 Mo 18 Mo 24 Mo 30 Mo
Figure 4-31. Cumulative incidence of tumor bearing animals for lymphoma
in female mice exposed to formaldehyde for 24 months.
Note: Mice from the 6-month interim sacrifice are not included since only nasal
passages were examined (p < 0.05).
Source: Battelle Columbus Laboratories (1981).
Although results are somewhat mixed between studies, there is evidence in the animal
bioassays for formaldehyde-induced LHP malignancies. Differences in study design may
account in part for mixed results. The lifelong drinking water study by Soffritti et al. (1989) may
have allowed for malignancies to develop late in life, whereas the drinking water study by Til et
al. (1989) sacrificed all animals at 24 months. Even though the exposure levels were similar, the
studies are not directly comparable. Similarly it is hard to directly compare results from the two
major inhalation studies in rats. Although Sellakumar et al. (1985) is a lifelong study, the
mortality for rats was greater than 80% at 2 years. Additionally, the pathology examination was
much less rigorous than in the Battelle Columbus Laboratories (1981) study, perhaps missing
smaller lesions. Therefore, the increase in formaldehyde-induced leukemia seen in female F344
rats late-in-life (Battelle Columbus Laboratories [1981]) may reflect a more sensitive study
design. Finally, strain differences may account for different susceptibility as well. The two
positive rat studies, by different routes of exposure, along with a positive result for
This document is a draft for review purposes only and does not constitute Agency policy.
4-334 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
formaldehyde-induced mouse lymphoma, make a substantive case for the potential of
formaldehyde-induced LHP malignancies.
4.2.2.4. Summary
Formaldehyde is toxic at the POE. Similar lesions, including increased cell proliferation,
DPX, and focal lesions, are noted in the GI tract or URT, depending on route of exposure.
Similarly, formaldehyde-induced tumors are noted at the POE for both routes of exposure.
Additionally, data exist for both routes of exposure to indicate that formaldehyde may act in part
as a tumor promoter.
When evaluating the studies with adequate study design to assess LHP malignancies,
results are mixed by strain, sex, route of exposure, and length of study. The three positive
studies (Soffritti et al., 1986; Battelle Columbus Laboratories, 1981) had the best histopathologic
examinations and greater sensitivity for detection of late-in-life tumors. Based on these results,
sufficient evidence is available in animal studies to support formaldehyde-induced LHP
malignancies.
4.3. GENOTOXICITY
Formaldehyde has been extensively studied for its mutagenic and genotoxic activity in a
variety of assay systems. The first reported mutagenic activity of formaldehyde was when
Rapoport (1946) described the induction of sex-linked recessive lethals in drosophila larvae fed
on a medium containing formalin. A variety of genotoxic and mutagenic effects have been
subsequently demonstrated in both in vitro and in vivo test systems, including the formation of
DPXs, point mutations, DNA strand breaks, increased MNs, and CAs (Auerbach et al 1977; Ma
and Harris, 1988; Conaway et al 1996; IARC 1995; 2006).
In this section, reactions of formaldehyde with cellular macromolecules, such as DNA
and proteins, and formaldehyde-induced clastogenicity are described. In addition, the evidence
for formaldehyde-induced mutations is considered in the context of the current EPA cancer
guidelines (U.S. EPA, 2005a). Particular emphasis is given to the genotoxic effects of
formaldehyde in humans, described in Section 4.3.4.2.
4.3.1. Formaldehyde-DNA Reactions
Formaldehyde is a reactive chemical and interacts with DNA in several ways, forming
DPXs, DNA adducts, and DNA-DNA cross-links (DDXs) (Fennell, 1994; Casanova et al., 1989;
Heck and Casanova, 1987; Casanova-Schmitz et al., 1984a, b; Casanova-Schmitz and Heck,
1983; Ohba et al., 1979; Donecke, 1978; Brutlag et al., 1969). Formaldehyde also may facilitate
the formation of adducts between other chemicals (endogenous or xenobiotic) and DNA
This document is a draft for review purposes only and does not constitute Agency policy.
4-335 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
(Fennell, 1994; Koppel et al., 1991; Casanova et al., 1989; Heck and Casanova, 1987; Lam et al.,
1985; Casanova-Schmitz et al., 1984a; Casanova-Schmitz and Heck, 1983; Ohba et al., 1979;
Donecke, 1978; Brutlag et al., 1969).
The high reactivity of formaldehyde results in little specificity in reaction sites, indicating
that a range of adducts and cross-links might be expected. However, the spectrum of
formaldehyde-DNA reaction products is difficult to quantify in vivo as many of these are labile
and difficult to measure directly (Fennell, 1994; Casanova et al., 1989). Additionally,
formaldehyde is metabolically incorporated into nucleic acids, and therefore DNA and RNA
assays incorporating radiolabeled formaldehyde need careful interpretation to distinguish
between covalently bound and metabolically incorporated formaldehyde (Casanova et al., 1989;
Heck and Casanova, 1987; Casanova-Schmitz et al., 1984a, b; Casanova-Schmitz and Heck,
1983). Hence, reports of formaldehyde-DNA reactivity in cell-free system results may not
provide a useful measure of exposure (Fennell, 1994). Besides, the question of biological
relevance must also be considered. On the other hand, methods used to extract and measure
DNA-formaldehyde reaction products after in vivo exposures should be evaluated to ensure that
formaldehyde reaction products are neither created nor removed during sample preparation
(Fennell, 1994; Casanova et al., 1989).
4.3.1.1. DNA-Protein Cross-Links (DPXs)
Evidence from numerous experimental models, ranging from cell-free systems to single
cells and in vivo animal and human exposures, suggests that formaldehyde reacts readily with
DNA forming DPXs (Reviewed in Conaway et al 1996; IARC 2006). As shown in Table 4-76,
cross-links between histones and DNA have been demonstrated in isolated chromatin samples on
exposure to formaldehyde from earlier studies (Ohba et al., 1979; Donecke, 1978; Brutlag et al.,
1969). Several in vitro studies demonstrated induction of DPX by formaldehyde exposure in
bacteria (Wilkins and McLeod 1976), yeast (Magana-Schwencke and Ekert, 1978) and
mammalian cells including animal cells (Chinese hamster ovary cells, Chinese hamster V79 lung
epithelial cells, mouse leukemia L1210 cells, mouse hepatocytes, rat Yoshida lymphsarcoma
cells, rat CI8 tracheal epithelial cells, rat hepatocytes, rat nasal, tracheal epitheial cells and aortic
endothelial cells) and human cells (lung and bronchial epithelial cells, fibrobasts, white blood
cells, peripheral blood lymphocytes, Epstein-Barr Virus-Burkitt's lymphoma cells, Jurkat E6-1
cells, HeLa cells, lymphoblastoid cells, gastric mucosal cells and whole blood) as summarized in
Table 4-76.
Ross and Shipley (1980) showed that formaldehyde induces SSBs and DPXs; SSBs are
formed at concentrations >200 |iM and a reduction of radiation-induced breaks (indirect measure
This document is a draft for review purposes only and does not constitute Agency policy.
4-336 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
of DPXs) at 50 |iM. The authors used a [14C]-thymidine-incorporated mouse L1210 cell line to
monitor formaldehyde-induced DNA strand breaks and DPXs. They exposed cells to varying
concentrations of formaldehyde for 2.5 hours. An alkaline-elution technique in the presence or
absence of proteinase K was used to measure strand breaks. In order to detect DPXs, some cells
were exposed to 300 R of X-rays immediately after formaldehyde treatment. Formaldehyde-
induced DPXs were repaired 24 hours after the compound was removed from the culture (Ross
and Shipley, 1980).
Casanova-Schmitz and Heck (1983) have shown that homogenates of rat nasal mucosa
incubated with formaldehyde in vitro followed by extraction with a strong aqueous-immiscible
organic solvent demonstrated increased DPX formation in DNA obtained after enzymatic
proteolysis from the aqueous-organic interface, termed as "interfacial DNA". In the same study,
they have shown that DNA isolated from the nasal, but not olfactory, mucosa of rats exposed to
formaldehyde (2, 6, 15, and 30 ppm 6 hours/day for 2 days) via inhalation showed significant
increase in DPXs in the interfacial DNA >6 ppm, which was shown to be linear in the exposure
range of 2-30 ppm (2.45-36.8 mg/m3). However, DNA in the aqueous phase did not show DPX
formation. Thus, the cross-linked DNA that could be extracted from the interface after
proteolysis was considered to be supporting evidence of chemically induced DPX formation.
The inability of this study to detect DPXs at lower levels of formaldehyde exposure is likely be
due to the protective mechanism of GSH, which catalyzes the conversion of formaldehyde to
formate.
So, in a later study, Casanova and Heck (1987) reported that GSH depletion caused an
increase in DPX formation in the interfacial DNA in the nasal mucosa of F344 rats when a dual-
isotope (3H/14C) method was used. The dual isotope method helps in making the distinction
between metabolic incorporation and covalent binding of formldehyde. Oxidatoin by removal of
one hydrogen atom is required for metabolic incorporation of formaldehyde into cellular
macromolecules, but not in the formation of DNA adducts or DNA-protein crosslinks. Thus, the
ratio of 3H/14C of DNA containing DPX will be higher than the macromoleucules where
formaldehyde is metabolically incorporated. However, the authors further demonstrated that,
when the double isotope method was used, the 3HCHO is oxidized significantly more slowly
than H14CHO under these conditions, resulting in an overestimate of the concentration of cross-
links due to an isotopic effect on the oxidation of 3HCHO catalyzed by formaldehyde
dehydrogenase (FDH). Besides, this method leaves residual formaldehyde that is likely to form
DNA adducts by reacting with deoxyribonucleosides in the DNA hydrolysates (Heck and
Casanova, 1987).
This document is a draft for review purposes only and does not constitute Agency policy.
4-337 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
To overcome this, Casanova et al. (1989) used an improved method, which is based not
on the analysis of residual formaldehyde bound to deoxyribonucleosides in DNA hydrolysates
but on the determination of the total 14C-formaldehyde bound to DNA. This study showed that
formaldehyde was exclusively bound to interfacial DNA, indicating the formation of DPXs.
Hydrolysis of DPXs in different samples quantitatively released formaldehyde. Besides, DPX
formation was detectable at all concentrations of exposure to formaldehyde (0.3-10 ppm for
6 hours). Overall, these studies clearly 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 URT.
Formaldehyde-induced DPXs were also found in the nasal mucosa and extra-nasal tissues
of rhesus monkeys exposed to 0, 0.71, 2, or 6 ppm (0, 0.86, 2.45, or 7.36 mg/m3) formaldehyde
6 hours/day for 3 days (Casanova et al., 1991). 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 or 6 ppm formaldehyde. These data demonstrate direct effects of formaldehyde on DNA in
tissues that correspond to observed tumor sites in humans (nasal and nasopharynx).
Bermudez and Delehanty (1986) observed the formation of DPXs, scheduled (S) and
unscheduled DNA synthesis (UDS), and synthesis of RNA when cultured F344 rat nasal
epithelial cells from the naso- and maxillary turbinates were incubated with formaldehyde.
Unscheduled and scheduled DNA synthesis was stimulated (0.05-0.1 mM) and then inhibited
(0.1-1 mM), depending on the formaldehyde concentration. Experiments by Cosma et al. (1988)
and Cosma and Marchok (1988) showed the induction of DPXs and DNA SSBs in cultured C18
rat tracheal epithelial cells exposed to 200 |xM formaldehyde for 90 minutes (Cosma et al., 1988;
Cosma and Marchok, 1988).
Several human cells (epithelial cells, fibroblasts, buccal cells) or cell lines
(lymphoblastoid cells) exposed to formaldehyde have been shown to form DPX (Craft et al
1987; Costa et al 1997; Emri et al 2004; Li et al 2004).
Craft et al. (1987) detected DPXs by alkaline elution in TK6 human lymphoblastoid cells
immediately after a 2-hour exposure (zero time) to 0, 15, 50, 75, 100, 150, 300, and 600 [xM
formaldehyde with a significant nonlinear increase in DPXs above 50 [xM concentration, which
correlated with the onset of cytotoxicity, but DPXs were completely removed in cultures held for
24 hours before processing. In the zero-time sample, significant increases in DPXs were first
observed at 50 [xM and increased linearly up to 150 [xM. In cells held for 24 hours, there was no
detectable increase in DPXs.
This document is a draft for review purposes only and does not constitute Agency policy.
4-338 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
However, Costa et al. (1997) detected DPXs with paraformaldehyde (which dissociates to
release formaldehyde) at doses that were cytotoxic (>0.003%) but could not discriminate
between the DPX-inducing and cytotoxic effects of this chemical in EBV, human Burkitt's
lymphoma cells (Costa et al., 1997). Grafstrom et al. (1983) reported that the number of DPXs
induced by 100 |jM formaldehyde in vitro in human epithelial cells and fibroblasts of bronchial
origin was similar and that the frequency of these cross-links was proportional to the
concentration of formaldehyde. Besides the bronchial epithelial cells and fibroblasts, the authors
also noted that formaldehyde exposure resulted in DPXs and DNA SSBs in skin fibroblasts and
DNA excision repair-deficient skin fibroblasts. However, 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 (Grafstrom et
al., 1984).
Emri et al. (2004) detected a significant increase in DPX formation in primary human
skin fibroblasts and keratinocytes at 8 hours of exposure in vitro to formaldehyde at 25 |jM with
an approximately linear increase up to 100 |jM. These cells were exposed to 0, 12.5, 25, 50, and
100 |jM formaldehyde for 8 hours and then exposed to 250 |jM methyl methane sulfonate
(MMS) for 2.5 hours. The induction of DPX formation was measured by the ability of
formaldehyde to reduce DNA migration in the comet assay induced by MMS in this study (Emri
et al., 2004).
Li et al. (2004) measured DNA damage in primary human buccal cells by using the
comet assay. The appearance of SSBs, suggesting compound-induced fragmentation of DNA,
occurred at formaldehyde concentrations of 5 and 7.5 |jM. At higher concentrations, the
response slope decreased, indicating DPXs or DDXs (Li et al., 2004). The same laboratory
reported similar findings in primary human peripheral blood lymphocytes and HeLa cells (Liu et
al., 2006). Peak response for SSBs was seen at 10 jjM in both cells, with higher concentrations
resulting in cross-link formation. SSBs in HeLa cells induced by 10 |jM formaldehyde were
repaired by 60 minutes after cells were washed to remove formaldehyde.
Schmid and Speit (2007) tested formaldehyde for its ability to induce DPXs in blood
cultures. They used an indirect method to monitor DPX formation in which the extent of DNA
migration in the comet assay in response to y radiation was compared in formaldehyde-treated
cultures versus controls. A concentration of 25 |iM was required for DPX formation, and repair
of these lesions was rapid, with DPXs induced by concentrations of formaldehyde up to 100 |iM
and completely removed after 8 hours.
This document is a draft for review purposes only and does not constitute Agency policy.
4-339 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-76. Formaldedyde-DNA reactions (DPX formation)
2
Species/Strain
Cell/Strain
Result
References
DNA Interaction
DPX formation in vitro
In vitro
Nucleohistone
+
Brutlag et al 1969
In vitro
Nucleohistone
+
Doenecke 1978
In vitro
Nucleohistone
+
Ohba et al 1979
Bacteria
+
Wilkins and McLeod 1976
Yeast
Saccharomyces cerevisiae
+
Magana-Schwencke and Ekert 1978
Hamster/Chinese
Ovary cells
+
Marinari et al 1984
Hamster/Chinese
Ovary cells
+
Zhitkovich and Costa 1992
Hamster/Chinese
Ovary cells
+
Olinetal 1996
Hamster/Chinese
Ovary cells
+
Garcia et al 2009
Hamster/Chinese
V79 lung epithelial cells
+
Swenberg et al 1983
Hamster/Chinese
V79 lung epithelial cells
+
Merk and Speit 1998
Hamster/Chinese
V79 lung epithelial cells
+
Merk and Speit 1999
Hamster/Chinese
V79 lung epithelial cells
+
Speit et al 2007
Mouse
Leukemia L1210 cells
+
Ross and Shipley, 1980
Mouse
Leukemia L1210 cells
+
Ross et al 1981
Mouse
Hepatocytes
+
Casanova and Heck 1997
Mouse
Hepatocytes
+
Casanova et al 1997
Rat
Yoshida lymphosarcoma cells
+
O'Connor and Fox 1987
Rat
CI8 tracheal epithelial cell line
+
Cosma and Marchok 1988
Rat/F344
Hepatocytes
+
Casanova and Heck 1997
Rat/F344
Nasal mucosa
+
Casanova-Schmitz and Heck 1983
Rat/F344
Nasal epithelium
+
Bermudez et al., 1986
Rat/F344
Primary tracheal epithelial
cells
+
Cosma etal., 1988
RatsAVistar
Aorta endothelial cells
+
Lin et al 2005
Human
Lung/bronchial epithelial cells
+
Fornace et al 1982
Human
Bronchial cell
+
Grafstrom et al., 1983
Human
Bronchial/ Skin fibroblast
+
Grafstrom et al., 1984
Human
Lung/bronchial epithelial cells
+
Grafstrom et al 1984
Human
Lung/bronchial epithelial cells
+
Saladino et al 1985
Human
Lung/bronchial epithelial cells
+
Grafstrom et al 1986
Human
Foreskin fibroblasts
+
Snyder and Van Houten 1986
Human
Lung/bronchial epithelial cells
+
Grafstrom 1990
Human
Bronchial/Skin fibroblasts
+
Olinetal 1996
Human
White blood cell
+
Shahametal., 1996
Human
EBV-Burkitt's lymphoma cells
+A
Costa et al., 1997
Human
Gastric mucosa cells
+
Blasiak et al 2000
Human
Peripheral lymphocyte
+
Quievryn and Zhitkovich 2000
Human
Fibroblast cells
+
Speit et al 2000
Human
Lymphocyte
+
Andersson et al., 2003
This document is a draft for review purposes only and does not constitute Agency policy.
4-340 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Human
Primary skin fibroblasts and
keratinocytes
+
Emri et al 2004
Human
Buccal cells
+
Li et al 2004
Human
JurkatE6-l cells
+
Saito et al 2005
Human
Peripheral lymphocyte
+
Liu et al 2006
Human
HeLa cells
+
Liu et al 2006
Human
Whole blood
+
Schmid and Speit 2007
Human
Lung/bronchial epithelial cells
+
Speit et al 2008
Human
Lymphoblastoid/TK6
+
Craft etal., 1987
DPX formation in vivo
Rat/F344
Nasal mucosa
+
Casanova-Schmitz and Heck 1983
Rat/F344
Nasal mucosa
+
Casanova-Schmitz et al 1984b
Rat/F344
Nasal mucosa
+
Lam etal 1985
Rat/F344
Nasal mucosa
+
Heck et al 1986
Rat/F344
Nasal mucosa
+
Heck Hd and Casanova 1987
Rat/F344
Tracheal implants
+
Cosmaetal., 1988
Rat/F344
Nasal mucosa
+
Casanova et al 1989
Rat/F344
Nasal mucosa
+
Casanova et al 1994
Rhesus monkeys
Nasal, larynx, trachea, and
carina
+
Casanova et al 1991
Human
White blood cell
+
Shahametal., 1996
Human
Peripheral lymphocyte
+
Shahametal., 1997
Human
Peripheral lymphocyte
+
Shaham et al., 2003
'+' indicates a positive test result
indicates a negative test result
A indicates that DNA-protein cross-links formed at cytotoxic concentrations
4.3.1.2. DNA Adducts
In addition to the formation of DPX, there is evidence that formaldehyde forms
hydroxymethyl (hm) DNA adducts in vitro in a variety of cell-free systems (Zhong and Que Hee,
2004a; Cheng et al., 2003; Kennedy et al., 1996; Fennell, 1994; Beland et al., 1984) and nasal
epithelial cells (Zhong and Que Hee 2004b). In cell-free systems, formaldehyde directly reacts
with DNA forming hmDNA adducts (Cheng et al., 2003; Kennedy et al., 1996; Fennell, 1994;
McGhee and von Hippel, 1977a, b, 1975a, b).
Beland et al. (1984) first reported hmDNA adducts in Chinese hamster ovary (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. Various
forms of hmDNA adducts, including N6-hydroxymethyldeoxyadenosine (N6-hmdA),
N4-hydroxymethyldeoxycytidine (N4-hmdC), and N2-hydroxymethyldeoxyguanosine (N2-
hmdG), were detected by high performance liquid chromatography (HPLC) following in vitro
reaction between formaldehyde and calf thymus DNA or individual deoxynucleotides.
This document is a draft for review purposes only and does not constitute Agency policy.
4-341 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
32P-postlabeling studies allowed for much greater analytical sensitivity but did not
confirm the level of N6-sulfomethyldeoxyadenosine found by HPLC. However, either estimate
of adduct formation is much less than the estimate of DPX formation (120 pmol/mg DNA) in
similarly treated rat nuclei (Heck and Casanova, 1987).
Casanova et al. (1989) demonstrated that detection of hmDNA adduct formation was
sensitive to the methodology used, particularly the buffer used for sample preparation.
Specifically, Tris buffer can prevent hmDNA adduct formation due to the abundance of
formaldehyde-reactive primary amine sites in the buffer. In contrast, the tertiary amine sites that
predominate in Bis-Tris buffer do not react with formaldehyde.
Zhong and Que Hee (2004a) observed hmDNA adducts (N6-hmdA, N2-hmdG, and
N4-hmdC) in placental DNA exposed to 100 ppm formaldehyde in vitro for 20 hours at 37°C
followed by hydrolysis of formaldehyde-reacted DNA using bis-Tris buffer. However,
deoxythymidine did not form hydroxymethyl derivatives in this study (Zhong and Que Hee,
2004a). On the other hand, the same investigators were able to detect N6-hmdA and N2-hmdG
adducts in human nasal epithelial cells cultured in the presence of 0, 10, 25, 50, 100, 250, 400, or
500 |ig/mL formaldehyde and using Tris buffer during hydrolysis of adducted DNA. The
toxicity threshold for <90% viability appeared to be between 100 and 250 |ag/mL initial
formaldehyde culture concentration, and even at 500 |j,g/mL concentration it was not toxic, with
a viability was 70% in this study (Zhong and Que Hee, 2004b).
The only report of formaldehyde-induced hmDNA adducts in vivo is a recent study
(Wang et al., 2007), showing an indirect evidence of formation of formaldehyde-induced N6-
hmdA in hepatic and pulmonary DNA from rats exposed to A'-nitrosodi methyl amine and
4-(methylnitrosamino)-1 -(3 -pyridyl)-1 -butanone.
Since the formaldehyde adducts are labile, Fennel (1994) developed a method by
derivatizing them with sodium bisulfite to their sulfomethyl form, whereby he detected
N6-sulfomethyldeoxyadenosine (SOMedA) and N2-sulfomethyldeoxyguanosine by using HPLC.
However, the levels of SOMedA in DNA isolated following incubation of radiolabeled
formaldehyde with isolated rat hepatic nuclei were similar to those in control nuclei. And in
human TK6 lymphoblastoid cells treated with formaldehyde, detection of SOMedA adducts was
precluded by additional radioactive sports. These observations suggest that N6-
sulfomethyldeoxyadenosine adducts are formed at very low levels in formaldehyde-incubated rat
nuclei and that measurement of hydroxymethyldeoxyadenosine would not provide a useful
measure of formaldehyde exposure (Fennell, 1994).
This document is a draft for review purposes only and does not constitute Agency policy.
4-342 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-77. Formaldedyde-DNA reactions (DNA adduct formation)
2
Species/Strain
Cell/Strain
Result
References
DNA Interaction
DNA Adduct Formation in vitro
Cell-free system
Deoxyribonucleosides
+
Cheng et al 2008
Cell-free system
Guanosine
+
Kennedy et al 1996
Cell-free system
Guanosine
+
Cheng et al 2003
Placental DNA
In vitro
+
Zhong and Que Hee 2004a
Calf thymus
In vitro
+
Beland et al 1984
Calf thymus DNA
In vitro
+
Von Hippel and Wang 1971
Cell-free system
In vitro
+
McGhee and von Hippel 1975a
Cell-free system
In vitro
+
McGhee and von Hippel 1975b
Cell-free system
In vitro
+
McGhee and von Hippel 1977a
Cell-free system
In vitro
+
McGhee and von Hippel 1977b
Cell-free system
In vitro
+
Fennell 1994
Cell-free system
In vitro
+
Cheng et al 2003
Rat
Nuclei
+
Fennell 1994
Rats
Nasal epithelial cells
+
Casanova et al 1989
Hamster
CHO cells
+
Beland et al 1984
Rat
Nuclei
+
Heck Hd and Casanova 1987
Human
Nasal epithelial cells
+
Zhong and Que Hee 2004b
DNA adduct formation in vivo
Drosophila
Larvae
+
Alderson, 1985
Rats
Indirect evidence
+
Wang et al 2007
3
4 '+' indicates a positive test result
5 indicates a negative test result
6
7
8 4.3.1.3. DNA-DNA Cross-Links (DDXs)
9 Formaldehyde, besides forming DPXs and DNA adducts, has also been shown to form
10 DDX in vitro. Li et al. (2004) showed that formaldehyde induces DNA strand breaks at low-
11 exposure concentration and DDXs and DPXs at higher concentrations in buccal cells. The
12 authors also showed that formaldehyde induces DDXs in human peripheral blood lymphocytes
13 exposed in vitro when the concentration was more than 25 |jM. However, the formation of
14 DDXs has not been demonstrated in vivo, and the relevance of these modifications in
15 formaldehyde-induced genotoxicity is not known at the moment.
16 Overall, formaldehyde forms predominantly DPXs that are detected in cell-free systems
17 and single cells in vitro and in animal and human tissues in vivo. In rodents, DPXs are formed in
18 nasal epithelia but not in extra-nasal passages, which are completely removed within a day after
19 formation. The DPXs are detected in nasal and extra-nasal tissues of monkeys, suggestive of
This document is a draft for review purposes only and does not constitute Agency policy.
4-343 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
direct effects of formaldehyde in tissues that correspond to observed tumor sites (nasal and
nasopharynx) in humans. Besides, this is used as a basis for cross-species comparison with
humans. Formaldehyde-DNA adducts are labile, constituting a minor fraction of the DNA-
reaction products. DPXs but not DNA adducts appear to play an important role in the
genotoxicity of formaldehyde.
4.3.1.4. Single Strand Breaks
Formaldehyde has been shown to induce DNA single strand breaks in a number of mammalian
cell systems in vitro as well as in vivo as shown in Table 4-78. Additionally, there is some
evidence that DNA single strand breaks (SSBs) may be induced directly by formaldehyde
reactivity (Grafstrom et al., 1984).
Table 4-78. Formaldehyde-DNA interactions (single strand breaks)
Species/Strain
Cell/Strain
Result
References
DNA single strand breaks (in vitro)
Hamster/Chinese
V79 lung epithelial cells
-
Speit et al 2007a
Mouse
Leukemia L1210 cells
(+)
Ross and Shipley, 1980
Mouse
Leukemia L1210 cells
-
Ross et al 1981
Rat
Hepatocytes
+
Demkowicz-Dobrzanski and
Castonguay 1992
Rat
Yoshida lymphosarcoma cells
+
O'Connor and Fox 1987
Rat/F344 trachea
Epithelial cell/ Primary culture
+
Cosmaetal., 1988
Human
Bronchial cell/Skin fibroblast
+
Grafstrom et al., 1984
Human
Peripheral blood lymphocytes
+
Liu et al 2006
Human
HeLa cells
+
Liu et al 2006
Human
Skin fibroblast
+
Snyder and Van Houten, 1986
Human
Lung/bronchial epithelial cells
+
Saladino et al 1985
Human
Lung/bronchial epithelial cells
+
Grafstrom et al 1984
Human
Lung/bronchial epithelial cells
+
Grafstrom 1990
Human
Lung/bronchial epithelial cells
+
Fornace et al 1982
Human
Lung/bronchial epithelial cells
+
Vock et al 1999
Human
Skin keratinocytes/fibroblasts
-
Emri et al 2004
DNA single strand breaks (in vivo)
Mouse
Liver (maternal)
+
Wang and Liu 2006
Mouse
Liver (fetal)
+
Wang and Liu 2006
Rats/Spraque-Dawley
Lung cells
+
Sul et al 2007
Rat
Lymphocytes
+
Im et al 2006
Rat
Lymphocytes
+
Im et al 2006
'no' indicates test was not done in vivo; '+' indicates a positive test result; 'yes' indicates test was done in vivo
; - indicates a negative test result; (+) indicates a weak positive test result.
This document is a draft for review purposes only and does not constitute Agency policy.
4.344 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.3.1.5. Other Genetic Effects of Formaldehyde in Mammalian Cells
Formaldehyde induces several other genetic and related effects in mammalian cells which
are evaluated by in vitro assays such as unscheduled DNA synthesis (UDS), DNA repair
inhibition and cell transformation as summarized in Table 4-79.
UDS, which represents DNA repair activity has been reported in nasal epithelial cells of
F344 rats (Bermudez and Allen 1984; Bermudez and Delahanty 1986), rat hepatocytes (Williams
et al 1989) and Syrian hamster embryo cells (Hamaguchi and Tsutsui 2000). UDS was observed
in HeLa cells (Martin et al 1978), but not in human bronchial epithelial cells (Doolittle et al
1985) upon formaldehyde exposure. These studies suggest that following formaldehyde-induced
DNA damage was followed by DNA repair.
Studies involving human bronchial epithelial cells and skin fibroblasts or keratinocytes
(Grafstrom et al 1984; Emri et al 2004), DNA repair proficient or -deficient cell lines (e.g.
xeroderma pigmentosum) or cell lines hypersensitive to DNA-DNA crosslinks (e.g. Fanconi's
anemia) (Speit et al 2000) it has been shown that formaldehyde causes DNA repair inhibition at a
concentration range of 0.125 mM to 10 mM). Emri et al (2004) have shown that DNA repair
was inhibited in human keratinocytes and fibroblasts after irradiation with UVB and UVC, but
not UVA follwed by treatment with low concentrations of formaldehyde (10 |jM). They
observed that DNA SSB induced by UVB or UVC irradiation alone were repaired within 3-6
hours of exposure, while cells with UV irradiation followed by formaldehyde exposure still had
the strand breaks at the same timepoints suggesting that formaldehyde is likely to contribute to
UV-induced carcinogenesis.
4.3.2. In Vitro Clastogenicity
Clastogenic effects, including increased MNs, CAs, and SCEs are also reported in a range
of in vitro study systems as shown in Table 4-80.
Miyachi and Tsutsui (2005) measured the induction of sister chromatid exchanges
(SCEs) in Syrian hamster embryo (SHE) cells. Cells were exposed to 0, 3.3, 10, and 33 |jM
formaldehyde for 24 hours. SCE levels after 3.3 |jM formaldehyde were not different from
controls, but significant increases were observed at both 10 and 33 |jM. Toxicity as measured by
reduced cloning efficiency was seen only at 33 |jM (Miyachi and Tsutsui, 2005). The same
laboratory used SHE cells to measure the induction of CAs (Hikiba et al., 2005). Cells were
exposed to 0, 33, 66, and 99 |jM formaldehyde for 24 hours prior to staining for analysis and the
percentages of aberrant metaphases were 0, 6, 6, and 71, respectively. The aberrations were
predominantly chromosome gaps and chromosomal breaks and exchanges. The relative colony-
This document is a draft for review purposes only and does not constitute Agency policy.
4-345 DRAFT—DO NOT CITE OR QUOTE
-------
1 forming efficiency remained high (at least 85%) for the concentrations of formaldehyde used in
2 the experiment (Hikiba et al., 2005).
3
4 Table 4-79. Other genetic effects of formaldehyde in mammalian cells
5
Species/Strain
Cell/Strain
Result
References
Unschedule DNA synthesis (UDS)
Rat/F344
Nasal epithelial cells
+
Bermudez and Allen 1984
Rat/F344
Nasal epithelial cells
+
Bermudez and Delehanty 1986
Rat
Hepatocytes
+
Williams et al 1989
Hamster/Syrian
Embryo cells
+
Hamaguchi and Tsutsui 2000
Human
HeLa cells
+
Martin etal 1978
Human
Bronchial epithelial cells
-
Doolittle et al 1985
DNA repair inhibition
Human
Bronchial epithelial cells/skin
fibroblasts
+
Grafstrom et al 1984
Human
Normal fibroblasts (MRC5CV),
XPA cell line, & FA cell line
+
Speit et al 2000
Human
Skin fibroblasts/keratinocytes
+
Emri et al 2004
6
7 XPA, xeroderma pigmentosum, complementation group A (deficient in NER pathway)
8 FA, Fanconi's anemia (cell line has genetic defect leading to hypersensitivity to DNA-DNA cross links; NER,
9 nucleotide excision repair
10
11
12 Schmid and Speit (2007) observed that SCEs were induced in lymphocytes of whole
13 blood cultures at a formaldehyde concentration of 200 |iM, an effect apparently associated with
14 cytotoxicity. This was indicated by a concomitant reduction in the proliferative index. These
15 authors also observed the formation of MNs in their cultures. This effect was statistically
16 significant at a formaldehyde concentration of 300 |iM and above. However, MN formation was
17 confined to those cultures in which formaldehyde treatment commenced 44 hours after the start
18 of the culture. This prompted the conclusion that the level of DPX formation would need to be
19 high for MN formation and that the cells must be exposed after the first mitosis. In examining
20 MN formation more closely, Schmid and Speit (2007) used the FISH technique, employing a
21 "biotin-labeled pan-centromeric chromosome paint specific for all human centromeres."
22 Indicative that formaldehyde was inducing a clastogenic (rather than aneugenic) effect, 81% of
23 MNs in binucleated cells were centromere-negative.
This document is a draft for review purposes only and does not constitute Agency policy.
4-346 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
In summary, formaldehyde forms MNs, SCEs, and CAs in isolated animal and human
cells following in vitro exposure (Table 4-80).
4.3.3. In Vitro Mutagenicity
Mutations may occur during repair 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 is strengthened by examining the
correlation between these genotoxic and clastogenic events and induction of mutations.
Therefore studies are presented with respect to relevance to one or more of the following lines of
evidence for mutagenicity recommended for consideration in the EPA guidance (U.S. EPA,
2005a): (1) that the chemical is DNA reactive and/or has the ability to bind to DNA, (2) that the
chemical generates positive results in in vitro mutagenic test systems (specifically gene
mutations and CAs), and (3) that the chemical induces indications of genetic damage in in vivo
tests (specifically gene mutations and CAs). Numerous studies have demonstrated
formaldehyde-induced DNA mutations under a variety of experimental conditions (reviewed in
IARC 1995, 2006; Ma and Harris 1988; Auerbach et al. 1977; Conaway et al 1996; NTP 2009).
4.3.3.1. Mutagenicity in Bacterial Systems
A number of research reports describe the mutagenicity of formaldehyde in bacterial test
systems using reverse and forward mutation assays as well as specific strains detecting deletions,
insertions and point mutations. Among the bacterial strains, Salmonella typhimurium TA102 and
the Escherichia coli strains containing an AT base pair at the primary reversion site are often
used to detect oxidative compounds, cross-linking agents and hydrazines. In an early National
Toxicology Program (NTP) collaborative study with three laboratories, formaldehyde
consistently tested positive for mutagenicity in Salmonella typhimurium strain TA100 in the
presence of a rat or hamster liver S9 activating system (Haworth et al., 1983). Formaldehyde
was mutagenic with and without metabolic activation in a number of other studies using in
reverse mutation assays with S. typhimurium strains TA98, TA100, TA102, TA104, TA2638,
and TA2638a and E. coli strains WP2 (pkMlOl), WP2 uvrA (pkMlOl), and hrs/r30R (Ryden et
al., 2000; Dillon et al., 1998; Watanabe et al., 1996; Le Curieux et al., 1993; O'Donovan and
Mee, 1993; Zielenska and Guttenplan, 1988; Schmid et al., 1986; Connor et al., 1983, 1985;
Orstavik and Hongslo, 1985; Takahashi et al., 1985; Fiddler et al., 1984; Frei et al., 1984;
Donovan et al., 1983), while other studies (Muller et al., 1993; Jung et al., 1992; Wilcox et al.,
1990; Marnett et al., 1985) show both positive and negative results. These results are
summarized in Table 4-81 and some of the studies are described in greater detail.
This document is a draft for review purposes only and does not constitute Agency policy.
4-347 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-80. In vitro clastogenicity of formaldehyde.
2
Species
Cell/Tissue origin
Without
activation
With
activation
References
Cytogenetic Assays
Chromosomal aberratoins (CA)
Hamster/Chinese
Ovary cells
(+)
(+)
Galloway et al., 1985
Hamster/Chinese
Ovary cells
-
ND
Dresp and Bauchinger, 1988
Hamster/Chinese
Ovary cells
+
ND
Natarajan et al 1983
Mouse
Lymphoma cells
+
ND
Speit and Merk 2002
Hamster/Syrian
Embryo cells
+
ND
Hikiba et al 2005
Hamster/Syrian
Embryo cells
+
ND
Hagiwara et al 2006
Hamster/Chinese
Ovary cells
+
ND
Garcia et al 2009
Hamster/Chinese
Lung fibroblasts
+
ND
Ishidate Jretal 1981
Human
Lymphocytes
+
+
Schmid et al 1986
Human
Lymphocytes
+
ND
Miretskaya and Shvartsman 1982
Human
Lymphocytes
+
ND
Dresp and Bauchinger, 1988
Human
Fibroblasts
+
ND
Levy et al 1983
Micronucleus (MN)
Hamster/Chinese
V79 lung epthelial cells
+
ND
Speit et al 2007b
Hamster/Chinese
V79 lung epthelial cells
+
ND
Merk and Speit 1998
Human
Whole blood cultures
+
ND
Schmid and Speit 2007
Human
Human MRC5CV (normal)
and XP(Repair-deficient) and
FA (repair-deficient) cell lines
+a
ND
Speit et al 2000
Sister Chromatid Exchan
ge (SCE)
Hamster/Chinese
Ovary cells
(+)
(+)
Galloway et al., 1985
Hamster/Chinese
Ovary cells
+
ND
Natarajan et al 1983
Hamster/Chinese
Ovary cells
+
ND
Garcia et al 2009
Hamster/Chinese
Ovary cells
+
ND
Obe and Beek 1979
Hamster/Syrian
Embryo cells
+
ND
Miyachi and Tsutsui 2005
Hamster/Chinese
V79 lung epithelial cells
+
(+)
Basleretal. 1985
Hamster/Chinese
V79 lung epithelial cells
+
ND
Speit et al 2007b
Hamster/Chinese
V79 lung epithelial cells
+
ND
Merk and Speit 1998, 1999
Hamster/Chinese
V79 lung epithelial cells
+
ND
Neuss and Speit 2008
Human
A549 lung epithelial cells
+
ND
Neuss and Speit 2008
Human
A549 + V79 (co-cultivated)
+°
ND
Neuss and Speit 2008
Human
A549 + V79 (co-cultivated)
d
ND
Neuss and Speit 2008
Human
Lymphocytes
+b
ND
Garry etal., 1981
Human
Lymphocytes
+
ND
Krieger and Garry 1983
Human
Lymphocytes
+
ND
Schmid et al 1986
Human
Lymphocytes
+
ND
Obe and Beek 1979
Human
Whole blood cultures
+
ND
Schmid and Speit 2007
This document is a draft for review purposes only and does not constitute Agency policy.
4-348 DRAFT—DO NOT CITE OR QUOTE
-------
Premature chromosome Condensation (PCC)
Hamster/Chinese
Ovary cells
+
ND
Dresp and Bauchinger, 1988
1
2
3
4
5
6
7
8
9
10
11
12
'+' indicates a positive test result
'ND' indicates test was not done
- indicates a negative test result
(+) indicates a weak positive test result
a MN frequency increased in repair-deficient cell lines compared to normal cell lines
b indicates SCE with significant loss of cell viability
c A549 cells exposed for 1 h with formaldehyde then co-cultivated with V79 cells
d A549 cells exposed for 1 h with formaldehyde, cells washed and then co-cultivated with V79 cells
XP, xeroderma pigmentosum; FA = Fanconi's anemia.
This document is a draft for review purposes only and does not constitute Agency policy.
4-349 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-81. Summary of mutagenicity of formaldehyde in bacterial systems
2
Species
Strain
Metabolic
activation
References
+S9
-S9
Mutagenicity Assays
Reverse Mutation
S. typhimurium
TA98, 100, 1535, 1537, 1538
-
-
De Flora, 1981
S. typhimurium
TA100
ND
(+)
Couch etal., 1982
S. typhimurium
TA100
+
-
Haworth et al., 1983
S. typhimurium
TA1535, 1537
-
-
Haworth et al., 1983
S. typhimurium
TA98
(+)
-
Haworth et al., 1983
S. typhimurium
TA98, TA100
+
+
Connor etal., 1983*
S. typhimurium
UTH8414, UTH8413
-
-
Connor etal., 1983*
S. typhimurium
TA97, 98, 100
+
+
Donovan et al., 1983
S. typhimurium
TA102
+
+
De Flora et al., 1984
S. typhimurium
TA100
+
ND
Frei et al., 1984
S. typhimurium
TA100
ND
+
Fiddler et al., 1984
S. typhimurium
TA100
+
(+)
Connor etal., 1985
S. typhimurium
TA98
(+)
-
Connor etal., 1985
S. typhimurium
UTH8414, UTH8413
-
-
Connor etal., 1985
S. typhimurium
TA100
(+)
-
Ashby et al., 1985**
S. typhimurium
TA97, 98, 1535, 1537, 1538
-
-
Ashby et al., 1985**
S. typhimurium
TA98, 100, 102
ND
(+)
Takahashi et al., 1985
E. coli
WP2, WP2 uvrA
ND
+
Takahashi et al., 1985
E. coli
H/R30R, HS30RuvrA
ND
+
Takahashi et al., 1985
E. coli
NG30rec^4, 016polA
ND
-
Takahashi et al., 1985
S. typhimurium
TA97, 98, 100
ND
-
Marnett et al., 1985
S. typhimurium
TA102, 104
ND
+
Marnett et al., 1985
S. typhimurium
TA98, 100
+
+
Oerstavik and Hongslo, 1985
S. typhimurium
TA100
+
+
Schmidetal., 1986
S. typhimurium
TA104
+
ND
Zielenska and Guttenplan, 1988
S. typhimurium
TA102
ND
-
Wilcox etal., 1990
E. coli
WP2 uvrA /(pKMlOl)
ND
+
Wilcox etal., 1990
E. coli
WP2 (pKMlOl)
ND
-
Wilcox etal., 1990
S. typhimurium
TA102
+
ND
Jung etal., 1992
S. typhimurium
TA102
ND
+
Le Curieux et al., 1993
S. typhimurium
TA102
+
ND
Mulleretal., 1993
S. typhimurium
TA98, 100, 102
ND
+
0 'Donovan and Mee, 1993
S. typhimurium
TA1535, 1537, 1538
ND
-
0 'Donovan and Mee, 1993
E. coli
WP2 (pKMlOl),
ND
+
O'DonovanandMee, 1993
E. coli
WP2uvrA (pKMlOl)
O'DonovanandMee, 1993
E. coli
K12 (AB1157)(WT)
ND
+
Graves et al., 1994
E. coli
K12 (AB1886)/(uvrA),
K12 (AB2480)/(rec A/uvrA)
ND
-
Graves et al., 1994
S. typhimurium
TA102, 2638
ND
+
Watanabe et al., 1996
E. coli
WP2 (pKMlOl), WP2uvrA
(pKMlOl)
ND
+
Watanabe et al., 1996
This document is a draft for review purposes only and does not constitute Agency policy.
4-350 DRAFT—DO NOT CITE OR QUOTE
-------
S. typhimurium
TA1535a
_
_
Sarrif etal., 1997
S. typhimurium
TA1537a
+
+
Sarrif etal., 1997
S. typhimurium
TA98, 100A
+
-
Sarrif etal., 1997
S. typhimurium
TA97a
ND
+
Sarrif etal., 1997
S. typhimurium
TA1535, 1537®
_
_
Sarrif etal., 1997
S. typhimurium
TA98®
+
+
Sarrif etal., 1997
S. typhimurium
TA100b
_
+
Sarrif etal., 1997
S. typhimurium
TA100C
+
+
Sarrif etal., 1997
S. typhimurium
TA100, 104®
+
+
Dillon etal., 1998
E. coli (Lac+ reversion)
WP3101P, WP3106P
+
Ohta et al 1999
S. typhimurium
TA102, 2638®
ND
+
Ryden et al., 2000
Forward Mutation
S. typhimurium
TM677
ND
(+)
Couch etal., 1982
S. typhimurium
TM677
+
+
Donovan et al., 1983
S. typhimurium
TM677
+
+
Temcharoen and Thilly, 1983
E. coli
D494uvrB
+
Bosworth et al 1987
Deletion Mutation
E.coli
GP120, GP120A 7-2, 33694
ND
+D
Crosby et al., 1988
Point Mutation
E.coli
GP120, GP120A 7-2, 33694
ND
+
Crosby et al., 1988
Insertion Mutation
E.coli
GP120, GP120A 7-2, 33694
ND
+
Crosby et al., 1988
1
2 '+' indicates a positive test result
3 'ND' indicates test was not done
4 indicates a negative test result
5 (+) indicates a weak positive test result
6 * indicates the use of formalin in mutagenicity assay
7 ** indicates the use of hexamethylmelamine (HEMLA), a formaldehyde-releasing compound, in mutagenicity assay
8 A indicates use of the Standard Plate Method
9 B indicates use of the Preincubation Plate Method
10 c indicates use of the Suspension Method
11 D indicates loss of DNA
12
13
14 Formaldehyde has been shown to be mutagenic in forward mutation assays using S.
15 typhimurium (Couch et al 1982; Donovan et al 1983; Temcharoen and Thilly 1983) as well as in
16 E. coli (Bosworth et al 1987). Temcharoen and Thilly (1983) examined the toxicity and
17 mutagenicity of S. typhimurium strain TM677, using forward mutation to 8-azaguanine
18 resistance, and have shown that formaldehyde induced both toxicity and mutagenicity at
19 minimum concentrations of 0.17 mM (-S9) and 0.33 mM (+S9). It has also been shown that
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
formaldehyde formed as an intermediate by oxidation at the methyl group of
N-nitrodimethylamine, a biologically active N-nitramine of environmental significance, is
mutagenic to S. typhimurium TA100 strain at low concentrations and toxic above 2 |j,mol/plate
(Frei et al., 1984).
Bosworth et al (1987) developed a forward mutation assay in E. coli D494 uvrB strain
transformed with a multi-copy mutator plasmid pGW1700, in which mutations are scored by an
increase in ampicillin-resistant colonies after exposure of bacterial cells during the logarithmic
growth by the test chemicals. This assay is more sensitive to base-pair substitutions, but less
sensitive to frameshift mutations compared to Salmonella/miceosome-based assays. In this assay,
the authors (Bosworth et al 1987) observed positive curvilinear response to formaldehyde
exposure. Crosby et al (1988) used four E. coli strains GP120, GP120A, 7-2, and 33694
containing the xanthine guanine phosphoribosyl transferase (gpt) gene (which detects point
mutations, deletions and insertions) tested the mutagenicity of formaldehyde by exposing for 1
hour at 4 and 40 mM concentrations. They observed 41% large insertions, 18% large deletions
and 41%) point mutations. However, at 40 mM dose there were 92%> point mutations, a majority
of them (62%) being transition mutations at a single AT base pair in the gpt gene. In the same
study they observed frameshift mutations in E. coli that was transformed with naked pSV2gpt
plamid DNA exposed to 3.3 or 10 mM formaldehyde. Thus, the mutation pattern appear to differ
depending on the concentration of formaldehyde exposure to the bacterial strain as well as the
nature of DNA.
Formaldehyde has also been shown to induce primary DNA damage in E. coli and
mutagenic activity in the Ames fluctuation test in S. typhimurium TA100, TA102, or TA98
strains (Le Curieux et al., 1993).
O'Donovan and Mee (1993) observed clear mutagenicity by the pre-incubation exposure
method in S. typhimurium TA98, TA100, and TA102 strains and both is. coli WP2(pKM101) and
WP2uvrA(pKM101) strains, while the standard plate-incorporation assays showed consistent
mutagenicity only with TA100 and WP2wvrA(pKM101) strains and no evidence of mutagenicity
in TA1535, TA1537, or TA1538 strains using either method of exposure in the absence of
metabolic activation. The S. typhimurium and E. coli strains used in this study are histidine and
tryptophan auxotrophs, with an AT base pair at the critical mutation site within the hisG and trpE
genes, respectively, with an intact excision repair system facilitating the detection of cross-
linking agents and both strains carrying the mutator plasmid, pKMlOl, which enhances error-
prone repair. These salmonella strains detect frameshift (TA98 and TA1537) and base-pair
substitutions (TA100, TA102, and TA1535), while theE. coli strains detect base-pair
substitutions (WP2/nrA), These findings are consistent with the suggestion that formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
induces excision-repairable lesions in bacteria and indicate that the presence of the R-factor
plasmid may be required for the expression of its mutagenicity in excision repair-deficient
salmonella (O'Donovan and Mee, 1993).
Dillon et al. (1998) employed salmonella strains TA100, TA102, and TA104 because of
the latter two strains being more sensitive to oxidative mutagens. Formaldehyde was clearly
mutagenic between 6 and 50 |j,g/plate in all three strains with and without metabolic activation
using Aroclor-induced S9 from male F344 rats or male B6C3F1 mice, except for an equivocal
response in TA102 with mouse S9 (Dillon et al., 1998). Using a set of six tester strains
(WP3101-WP3106) of E. coli, each reversible by a mutation involving a single DNA base pair
substitution, Ohta et al. (1999) determined that formaldehyde preferentially induced GC to TA
transversion mutations. Ryden et al. (2000) demonstrated a statistically significant increase in
the number of revertants in S. typhimurium TA102 (2.5-fold) and TA2638a (3-fold) strains by
formaldehyde at >17 |j,g/plate compared with solvent controls.
In summary, formaldehyde induces mutations in several bacterial strains containing an
AT base pair at the primary reversion site that are used to detect oxidative compounds and cross-
linking agents without metabolic activation by exogenous enzyme-activating systems. This
evidence is strengthened by examining the correlation between genotoxic and clastogenic events
and mutation induction.
4.3.3.2. Mutagenicity in Non-Mammalian Cell Systems
Formaldehyde has been shown to be mutagenic in several non-mammalian systems also.
It has been shown to cause gene conversion, strand breaks, crosslinks, homozygosis and related
damage in yeasts (Saccharomyces cervisiae), forward and reverse mutations in molds
(Neurospora crassa), micronuclei formation in spiderworts (Tradescantiapallida), 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), but failed to show micronuclei formation in newt larvae (Pleurodeles
waltl) (Reviewed in Conaway 1996; IARC 2006).
4.3.3.3. Mutagenicity in Mammalian Cell Systems
Several studies demonstrated the mutagenicity of formaldehyde in mammalian cells. In
its report, the Federal Panel on Formaldehyde underlined the role of formaldehyde as an inducer
of gene mutations and CA in a variety of test systems (Report of the Federal Panel on
Formaldehyde, 1982). Results from several studies are summarized in Table 4-82.
This document is a draft for review purposes only and does not constitute Agency policy.
4-353 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
27
28
29
30
Snyder and Van Houten (1986) demonstrated that formaldehyde increases the levels of
misincorporation of bases into synthetic polynucleotides catalyzed by E. coli DNA polymerase I,
indicating that the mutagenicity of formaldehyde may be due to covalent alteration of DNA
bases. They have also shown that formaldehyde-induced DNA damage in human fibroblasts was
not susceptible to repair by the typical "long patch" excision repair mechanism.
Craft et al. (1987) measured the induction of mutations at the thymidine kinase {tk) locus
or at the ouabain resistance (Oud) locus in TK6 human lymphoblastoid cells. The tk mutations
can result from a variety of mutational events, including base pair substitution, small and large
deletions, and chromosome exchange events, while mutations to Oud require specific base pair
substitutions. Single treatment of formaldehyde (0, 15, 30, 50, 125, and 150 |xM) for 2 hours
resulted in a nonlinear increase in ^mutagenesis with increasing slope >125 |iM (Figure 4-32).
To explore a dose-response effect, cells were also exposed as follows: three treatments of 50 |iM
for 2 hours or five treatments of 30 [xM or 10 treatments of 15 |iM for 2 hours (treatments were
spaced 2-4 days apart) with multiple treatments causing an increase in tk mutations, although
their combined effect was less than a single treatment of equivalent C x t (150 [jM for 2 hours).
Lymphoblasts given four treatments of 150 [xM formaldehyde for 2 hours failed to induce
mutations at the Oud locus. Dose-response increases were seen in all exposure scenarios, with
30 [xM being the level of statistical significance. There was little indication of a dose-response
effect until the cumulative concentration was greater than 100 [xM. Formaldehyde-induced
DPXs were no longer evident after 24 hours of exposure; mutants induced in the TK6
lymphoblast cell line showed a similar dose-response curve to the DPXs measured immediately
after exposure ended (Craft et al., 1987).
The same group also studied mutations induced at the X-linked hypoxanthine-guanine
phosphoribosyl transferase (HPRT) locus by eight repetitive treatments of 150 |xM formaldehyde
in TK6 human lymphoblast cell line by Southern blot analysis, wherein half (14/30) of induced
mutants 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 (Crosby et al., 1988).
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 54 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-82. Mutagenicity in mammalian cell systems.
2
Species/Strain
Cell/Strain
In Vivo
test
Without
activation
References
Mutagenicity Assays
Dominant Lethal Mutation
Rat/Albino
Spermatocyte, Live implants
yes
+
ND
Odeigah, 1997
Mouse
Dominant lethal
yes
-
Epstein and Shafner 1968
Mouse
Dominant lethal
yes
-
Epstein et al 1972
Mouse
Dominant lethal
yes
(+)
Fontignie-Houbrechts 1981
Rat
Dominant lethal
yes
(+)
Kitaeva et al 1990
Deletion Mutation
Hamster/Chinese
V79 cells Hprt locus)
no
-
Merk and Speit 1998
Hamster/Chinese
V79 cells Hprt locus)
no
-
Merk and Speit 1999
Hamster/Chinese
V79/HPRT
no
+
ND
Grafstrom et al., 1993
Hamster/Chinese
Ovary HPRT
no
-
+
Graves et al., 1996
Mouse
Lymphoma L5178Y cells (Tk + "
locus)
no
+
Macerer et al 1996
Mouse
Lymphoma L5178Y cells
no
+
ND
Speit and Merk 2002
Human
Bronchial cell
no
+
Grafstrom et al., 1983
Human
Bronchial fibroblasts/epithelial cells
(HPRT locus)
no
+
Grafstrom et al 1985
Human
Bronchial fibroblasts/epithelial cells
(HPRT locus)
no
+
Grafstrom 1990
Human
Lymphoblast/HPRT
no
+a
ND
Crosby et al., 1988
Human
Lymphoblast/tk
no
+
Craft et al 1987
Human
Peripheral lymphocytes
yes
+
ND
Shaham et al 2003
Human
Lymphoblast (TK6)
no
+
Goldmacher and Thilly 1983
Point Mutation
Hamster/Chinese
Ovary HPRT
no
+
ND
Graves et al., 1996
Mouse
Lymphoma cell/ TK+/-
no
+
+
Blackburn et al., 1991
Mouse
Lymphoma cell/ TK+/-
no
+
ND
Wangeheim and Bolcsfoldi, 1988
Human
Lymphoblast/TK6
no
+
ND
Liber etal., 1989
Insertion Mutation
Hamster/Chinese
Ovary HPRT
no
+
ND
Graves et al., 1996
Heritable Mutation
Mouse
Heritable mutation
yes
+
Liu et al 2009
DNA Repair enzyme activity
Human
Peripheral lymphocyte
yes
-
Hayes et al., 1997
This document is a draft for review purposes only and does not constitute Agency policy.
4-355 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Cell Transformation
Mouse
C3H10T1/2 cells
+b
Ragan and Boreiko 1981
Mouse
Embryo fibroblast/C3H/10Tl/2
no
[+]
ND
Boreiko et al., 1983
Mouse
Embryo fibroblast/C3H/10Tl/2
no
[+]
ND
Frazelle et al., 1983
Hamster
Kidney cell/BHK-21/cI.13
no
+
+
Plesner and Hansen, 1983
p53 mutation and/or p53 protein expression
Rats/F344
Nasal squamous cell carcinomas
yes
+°
Recio et al 1992
Rats/F344
Nasal tumor cell lines
No
+
Bermudez et al 1994
Rats/F344
Nasal squamous cell carcinomas
Yes
+d
Wolfetal 1995
Human
Peripheral blood lymphocytes
yes
+
Shaham et al 2003
'no' indicates test was not done in vivo
'+' indicates a positive test result
'ND' indicates test was not done
'yes' indicates test was done in vivo
- indicates a negative test result
(+) indicates a weak positive test result
[+] indicates positive test result after TPA or N-methyl-N-nitro-N-nitrosoguanidine promoter treatment
" indicates loss of DNA
b Positive only in the presence of 12-O-tetradecanoylphorbol 13-acetate (TPA)
0 p53 mutations
d p53 mutated protein detected by immunohistochemistry
This document is a draft for review purposes only and does not constitute Agency policy.
4-356 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
— DPX -A - MF ¦
0 30 60 90 120 150
Formaldehyde concentration (ijM)
Figure 4-32. DNA-protein cross-links (DPX) and thymidine kinase (tk)
mutants in TK6 human lymphoblasts exposed to formaldehyde for 2 hours.
Note: ~ DPXs immediately after exposure, BDPXs 24 hours after exposure, A tk
mutants. Relative survival was 100% at 0, 15, 30, and 50 |iM, 30% at 125 |iM,
and 20% at 150 |iM.
Source: Adapted from Craft et al. (1987).
Liber et al. (1989) followed up the findings of Crosby et al. (1988) by performing
Southern blot, Northern blot, and DNA sequence analysis on the 16 induced and 10 spontaneous
human lymphoblast mutants not showing deletions. Northern blot analysis showed that the point
mutations fell into four categories: normal size and amount of RNA, normal size but reduced
amounts of RNA, reduced size and amounts of RNA, and no RNA. Sequence analysis of
recombinant DNAs from hprt mRNA in formaldehyde-induced mutants showed a preferential
AT to CG transversion at a specific site, with other changes represented to a lesser degree (Liber
et al., 1989).
Even in CHO cells formaldehyde has been shown to induce hprt mutations involving
mostly single-base pair transversions mostly occurring at AT sequences, including three AT to
TA at position 548 of exon 8 and two AT to CG and one GC to TA transversion at other sites
This document is a draft for review purposes only and does not constitute Agency policy.
4-357 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
(Graves et al., 1996). In another study, formaldehyde-induced forward mutations to
trifluorothymidine resistance 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 FADH was incorporated in the exposure medium (Blackburn
et al., 1991).
Formaldehyde-induced DPXs are removed in part through spontaneous hydrolysis and in
part due to active repair processes (Quievryn and Zhitkovich, 2000). Inhibition of specific
proteosomes in XP-A cells inhibited DPX repair, thereby supporting the role of enzymatic
degradation (Quievryn and Zhitkovich, 2000). The half-life of formaldehyde-induced DPXs in
human cell lines was consistent with the findings of Craft et al. (1987), ranging from 11.6 to
13 hours (Quievryn and Zhitkovich, 2000). In the same report, removal of DPXs from human
peripheral lymphocytes was much slower, with a half-life of 18.1 hours. This difference was
primarily in slower active repair of DPXs, with a tm of 66.6 hours for human lymphocytes
versus 23.3 hours for human cell lines (Quievryn and Zhitkovich, 2000).
Since DPX repair involves proteolytic removal of proteins from the DNA, Speit et al.
(2000) hypothesized that single peptides or small peptide chains cross-linked to the DNA are
critical to formaldehyde-induced mutation. However, these authors did not find significant
difference in the induction and repair of DPXs in normal and DNA repair-deficient cell lines but
observed increased susceptibility of the repair-deficient cell lines to formaldehyde-induced MN
induction. In this study, a normal human cell line (MRC5CV1), a xeroderma pigmentosum cell
line deficient in nucleotide excision repair (NER), and a Fanconi anemia cell line, which has a
genetic defect leading to hypersensitivity towards DDXs, were exposed to 125, 250, and 500 |jM
formaldehyde for 2 hours. The authors suggest that more than one repair pathway is involved in
the repair of cross-links and that the altered NER pathway has more severe consequences to
formation of CAs than disturbed cross-link repair (Speit et al., 2000).
The correlation of early DPX formation and mutation is at first counterintuitive since the
cross-linking of protein to DNA inhibits DNA replication. Without active DNA replication,
formaldehyde-DNA adducts and DPXs would not induce replication error and would be unlikely
to result in a change in DNA sequence or mutation. Recent evidence indicates that residual
peptides and short polypeptides that remain cross-linked to DNA after DPX removal may in fact
be the cause of DPX-associated, formaldehyde-induced mutation (Speit et al., 2000).
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, leading the
authors to speculate that formaldehyde was not mutagenic. Since, the hprt locus in the V79
Chinese hamster cell line is primarily sensitive to point mutations and other studies show the
This document is a draft for review purposes only and does not constitute Agency policy.
4-358 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
formation of deletion mutations by formaldehyde at the same locus in human lymphoblasts
(Crosby et al., 1988), Merk and Speit (1998) concluded that the hprt mutation assay is insensitive
to deletion mutations.
Later, using the mouse lymphoma assay, Speit and Merk (2002) demonstrated that
exposure to formaldehyde for 2 hours was mutagenic in a concentration-dependent manner in the
L5178Y mouse lymphoma cells, which was mainly contributed by a strong increase in small
colony mutants, suggestive of CAs (Speit and Merk, 2002). Detailed analysis of both
spontaneous and formaldehyde-induced lesions indicates that recombination or deletion of DNA
from the tk locus was primarily responsible for the loss of heterogeneity, thereby leading to the
observed mutant phenotype. Therefore, it is believed that formaldehyde is mutagenic in the
L5178Y cell mouse lymphoma system by a clastogenic mechanism rather than through point
mutations. This finding is consistent with that of Craft et al. (1987), who demonstrated
formaldehyde mutagenicity at the tk locus of TK6 cells, and also with the findings of Grafstrom
et al. (1984), who demonstrated increased SSB formation in formaldehyde-exposed cell lines.
Formaldehyde has also been shown to induce cell transformation in mouse embryo
fibroblasts (Ragan and Boreiko 1981; Boreiko et al 1983; Frazelle et al 1983). At low
concentrations of 0.017 mM formaldehyde has shown to cause cell transformation in C3H10T1/2
mouse cells (Ragan and Boreiko 1981) and hamster kidney cells in vitro (Plenser and Hansen
1983).
More recently, Shaham et al. (2003) examined the frequency of DPXs and the incidence
of mutant versus wild type p53 tumor suppressor genes in the peripheral blood lymphocytes of a
cohort of workers exposed to formaldehyde. The adjusted mean levels of DPXs were greater in
the lymphocytes of exposed subjects compared with those of unexposed subjects, and exposure
to formaldehyde increased the likelihood of their having a higher level of pantropic p53
(>150 pg/mL). The authors speculated on a possible causal relationship between DPXs and
mutations in p53. Recio et al (1992) demonstrated point mutations in the p53 tumor suppressor
gene in 45% (5 out of 11) of the primary nasal squamous cell carcinomas (SCCs) obtained from
F344rats that were chronically exposed to 15 ppm formaldehyde for 2 years (Recio et al ., 1992).
In summary, the results of in vitro experiments demonstrate the mutagenicity of
formaldehyde. Mutagenicity is observed below levels of significant cytolethality in mammalian
cell lines. Formaldehyde is clearly a DNA-reactive genotoxicant inducing lesions (DPXs) that
show clastogenicity (SSBs, MNs, etc.). The experiments by Speit and Merk (2002) explore
mechanistic links between DPXs, clastogenicity, and the observed locus-specific mutations in
the mouse lymphoma in vitro testing system.
This document is a draft for review purposes only and does not constitute Agency policy.
4-359 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.3.4. In Vivo Mammalian Genotoxicity
4.3.4.1. Genotoxicity in Laboratory Animals
As discussed above, formaldehyde is clearly reactive at the POE in animal studies,
resulting in increased DPXs in the nasal mucosa. Despite formaldehyde's reactivity and
mutagenicity in isolated mammalian cells, clear evidence of mutagenicity does not emerge from
animal bioassays (Table 4-83).
In a chromosomal analysis study (Fontignie-Houbrechts, 1981), formaldehyde given I.P.
at 50 mg/kg to male Q strain mice and analyzed 8-15 days after treatment did not induce any
chromosomal lesions in spermatocytes. Also, in another study from the same group (Fontignie-
Houbrechts et al., 1982), formaldehyde (30 mg/kg) given along with hydrogen peroxide (90
mg/kg) as a mixture to male Q strain mice failed to produce significant increases in
chromosomal lesions in the spermatogonia.
In a different study Natarajan et al. (1983) failed to detect significant differences in MN
induction in bone-marrow cells or CAs in spleen cells of male and female CBA mice given I.P.
6.25, 12.5, and 25 mg/kg formaldehyde compared with saline-treated controls. However, the
same study showed a positive induction of MNs and CAs in vitro. The authors suggest that the
lack of genotoxicity in vivo may be due to the inability of formaldehyde to reach the target cells
in sufficient quantity to induce biological effects.
Kligerman et al. (1984) also found no difference in the incidence of SCEs or
chromosome breakage in the peripheral lymphocytes of male and female F344 rats exposed to
formaldehyde in air at 0.5, 6, or 15 ppm (0.61, 7.36, or 18.4 mg/m3) 6 hours/day for 5 days.
However, in a different study (Migliore et al., 1989), clastogenic effects, such as increased MNs
and CAs, were reported in GI epithelial cells of male Sprague-Dawley rats after oral exposures to
200 mg/kg formaldehyde. In this study, micronucleated cells and nuclear anomalies were
increased in a time-dependent manner in the stomach, duodenum, ileum, and colon of rats, and
the mitotic index was unchanged for these cells compared with controls at 16, 24, and 30 hours.
These clastogenic effects were seen without regenerative cell proliferation, supporting
formaldehyde-induced mutations as primary effects of formaldehyde rather than secondary to
regenerative cell proliferation.
Kitaeva et al. (1990) observed cytopathological and cytogenetic effects of formaldehyde
chronic inhalation in 0.5 and 1.5 mg/m3 doses in the female rat's germ and marrow cells, where
formaldehyde-induced harmful effects were seen in germ cells at <1.5 mg/m3 doses, while the
reliable clastogenic and cytogenetic effects on the marrow cells were induced even at the
0.5 mg/m3 dose, suggesting differences among effects of small doses of formaldehyde on
different cell systems.
This document is a draft for review purposes only and does not constitute Agency policy.
4-360 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-83. Genotoxicity in laboratory animals.
2
Species/Strain
Cells/Organ/Tumor
Result
References
Cytogenetic Assays
Chromosomal aberratoins (CA)
Mice/Q strain
Spermatocyte
-
Fontignie-Houbrechts et al., 1981
Mice/Q strain
Spermatogonia
-
Fontignie-Houbrechts et al., 1982
Mice/CBA
Polychromatic erythrocytes
-
Natarajan et al., 1983
Mice/CBA
Spleen cells
-
Natarajan et al., 1983
Rats/F344
Lymphocytes
-
Kligerman et al 1984
Rats/Spraque-Dawley
Gastric epithelial cells
+
Migliore et al 1989
RatsAVistar
Bone marrow
+
Kitaeva et al 1990
Rats/Spraque-Dawley
Bone marrow
-
Dallas et al 1992
Rats/Spraque-Dawley
Pulmonary lavage cells
+
Dallas et al 1992
Rats/F344
Peripheral blood cells
-
Speit et al 2009
Micronucleus (MN)
Mouse/NMRI
Bone marrow
_
Gocke et al 1981
Mice/CBA
Femoral polychromatic
erythrocyte and spleen cell
-
Natarajan et al., 1983
Rats/Spraque-Dawley
Gastric epithelial cells
+
Migliore et al 1989
Sister Chromatid Exchang
e (SCE)
Rats/F344
Lymphocyte
-
Kligerman et al 1984
Rats/F344
Peripheral blood cells
-
Speit et al 2009
3
4 '+' indicates a positive test result
5 - indicates a negative test result
6
7
8 Dallas et al. (1992) observed a slight increase (7.6 and 9.2%) in CAs in pulmonary lavage
9 cells from male Sprague-Dawley rats exposed to 15 ppm (18.4 mg/m3) formaldehyde in air
10 6 hours/day, 5 days/week for 1 or 8 weeks by inhalation compared with corresponding controls
11 (3.5 and 4.8%), respectively. However, the small study, limited as it was to five animals/group,
12 showed statistically significant increase at the highest dose tested (15 ppm) but not at lower
13 doses (0.5 and 3 ppm). In the same study, no clastogenic effects were seen in bone marrow,
14 which is consistent with formaldehyde acting primarily at the site of first contact.
15 Speit et al (2009) investigated the genotoxicity of formaldehyde in peripheral blood
16 samples of Fischer-344 rats exposed to 0 to 15 ppm formaldehyde by whole-body inhalation for
17 4 weeks (6 h/day, 5 days/week). In this study, the authors found no significant increase in the
18 genotoxic assays such as comet assay with or without gamma-irradiation of blood samples (DNA
19 migration as determined by tail movement or intensity), sister chromatid exchange (SCE) assay
This document is a draft for review purposes only and does not constitute Agency policy.
4-361 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
and micronucleus test (MNT) compared to controls. However, rats given 50 mg/kg
methylmethane sulfonate (MMS) by gavage for 4 hrs (positive control for Comet and SCE
assays) or 10 mg/kg cyclophosphamide (CP) given twice orally (positive control for MNT)
induced significant increase in genotoxicity in this study. The lack of genotoxicity in this study
is not surprising since earlier studies by Casanova-Schimitz et al (1984a) have shown that
formaldehyde does not cause toxicity to bone marrow possibly due to the inability of this
chemical to reach the bone marrow. Although MMS and CP used in this study were positive in
the genotoxicity assays, the data from positive controls can not be used for validation since the
exposure routes of formaldehyde (inhalation) and the positive controls (oral) were different.
No animal studies have examined clastogenic effects of formaldehyde in nasal or
respiratory epithelial cells. Therefore, it is unknown whether similar changes would occur in
response to exposure to formaldehyde via inhalation. However, the negative finding in bone
marrow cannot be considered definitive evidence on the question of the mutagenic potential of
formaldehyde for cells present at the POE. With weak positive results in pulmonary lavage cells
and clear clastogenicity in GI epithelial cells below exposures that trigger regenerative cell
proliferation, the existing evidence, however incomplete, supports the concept of genotoxic
action of formaldehyde at the POE.
4.3.4.2, Genotoxicity in Humans
The majority of the studies on the effects of formaldehyde in exposed humans have
measured various cytogenetic endpoints, such as MNs, SCEs, or CAs in nasal and oral mucosal
cells (considered to be in direct contact with formaldehyde) as well as peripheral lymphocytes.
Since genotoxicity at the proximal sites (oral, nasal) can be readily linked to the reactive nature
of formaldehyde, these studies are discussed first, noting where researchers also collected blood
lymphocyte samples. A subsequent discussion is focused on results in blood lymphocytes.
Finally, the few studies that measured DPXs in exposed humans are discussed. Table 4-89
provides a summary of human cytogenetic studies of formaldehyde.
4.3.4.2.1. Nasal, buccal, and oral mucosal cells. Epithelial cells of the URT and oral cavity are
potential targets of formaldehyde's DNA reactivity and genotoxicity. Several studies indicate
that formaldehyde exposure results in measurable increases in SCEs, MN formation, and DPXs
in nasal, buccal, and oral mucosal cells; however, these genotoxic effects vary with the type of
exposure. Study quality, sample size, availability of exposure measurements, and assay
methodology may in part contribute to variability in study findings. The studies fall into three
This document is a draft for review purposes only and does not constitute Agency policy.
4-362 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
general categories: workers (industrial or professional), students and staff attending anatomy and
mortuary science courses, and subjects in a controlled clinical trial.
Ballarin et al. (1992) observed significantly higher frequency of micronucleated cells in a
formaldehyde exposed group in a plywood factory compared with controls (0.9 ± 0.47 versus
0.25 ± 0,22, p < 0.01). In this study, the frequency of MNs and cytology of respiratory nasal
mucosal cells was examined in 15 nonsmokers exposed to levels of formaldehyde that ranged
between 0.1 and 0.39 mg/m3 (-0.32 ppm) for an average of 6.8 years. Exposed subjects were
compared with age- and sex-matched controls.
Ye et al. (2005) reported significant increases in MNs per thousand cells in nasal mucosal
cells for 18 nonsmoking workers (2.70 ± 1.50) in a formaldehyde manufacturing plant in the
Hugei province of China as compared with controls (1.25 ± 0.41). In addition, higher
frequencies of SCEs in peripheral lymphocytes of workers were also reported (8.24 ± 0.89 versus
6.38 ± 0.41). In this study, the average age of workers was 29 ± 6.8 years, the average duration
at work was 8.5 years (range 1-15 years), and the reported 8-hour TWA was 0.985 mg/m3
(0.8 ppm). The control group consisted of 23 undergraduate students with an average age of
19 + 2.3 years. The 8-hour TWA in the student dormitories was 0.011 mg/m3 (9 ppb). A group
of 16 waiters with an average exposure duration of only 12 weeks and an 8-hour TWA of
0.107 mg/m3 (90 ppb) was also included in the study. The incidence of MNs and SCEs in the
waiters was the same as that in controls. Overall, results from this study suggest that the
genotoxic potential of high-level formaldehyde exposure may have occupational risks in long-
term exposure.
However, in a different study, Speit et al. (2007b) showed that formaldehyde did not
induce MNs in exfoliated buccal mucosa cells of humans exposed up to a maximum of 1 ppm
and a cumulative exposure of 13.5 ppm-hours over 2 weeks. In this study, volunteers exposed to
formaldehyde in closely controlled conditions (4 hours/day for 10 days) with a complex
exposure schedule, amounting to a cumulative total of 13.5 ppm-hours (16.6 mg/m3-hours), were
used. Samples of the buccal mucosa were taken from subjects 1 week before the start of the
experiment, at the start of the experiment, at the conclusion of the series of exposures, and at 7,
14, and 21 days after the completion of exposure. Thus, the subjects served as their own
controls. Two thousand cells per data point were assessed for the frequency of MNs on slides
that were coded by an independent quality assurance organization. As shown in Table 4-84, the
frequency of MN formation was statistically unchanged from that in controls. The apparent
slight increase in subjects evaluated at the conclusion of exposure was caused by frequencies of
MNs in two subjects (5.0 and 4.5 MNs per 1,000 cells). The data as reported show a high
This document is a draft for review purposes only and does not constitute Agency policy.
4-363 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
variability, where the SD approaches or exceeds the mean for each sample point, suggestive of
data with an asymmetrical distribution.
Table 4-84. MN frequencies in buccal mucosa cells of volunteers exposed to
formaldehyde
Sampling point
Group
MN/1000 cells (± SD)
Control data
1 week before exposure
1
0.95 ±0.67
Immediately before exposure series
2
0.86 ±0.84
Test data
Immediately after exposure series
3
1.33 ± 1.45
7 days after exposure
4
0.94 ±0.73
14 days after exposure
5
0.85 ±0.86
21 days after exposure
6
0.44 ± 0.38a
aStatistically significantly different from control values (p < 0.05), as calculated by the authors.
Source: Speit et al. (2007b).
The best evidence of formaldehyde-induced clastogenic changes in peripheral
lymphocytes is found in studies of anatomy class and mortuary class students. Since genetic
damage accumulates with age, the studies in younger adults, where cells are analyzed before and
after exposure, may have greater sensitivity and fewer confounding factors.
Suruda et al. (1993) showed a 12-fold increase in the MN frequency of epithelial cells
from the buccal area of the mouth in mortuary science students exposed to embalming fluids
containing formaldehyde following an 85-day exposure period (Table 4-85). Overall, students
were exposed to 0.33 ppm (0.4 mg/m3) formaldehyde as an 8-hour TWA on days when
embalming was performed (an average of 6.9 embalmings). Blood, oral, and nasal samples were
collected pre- and postexposure. As shown in Table 4-85, nasal epithelial MNs increased by
22% (frequency of micronucleated lymphocytes increased by 28%). By contrast, SCE frequency
decreased by 7.5% after formaldehyde exposure.
Table 4-85. MN and SCE formation in mortuary science students exposed to
formaldehyde for 85 days
Sampling point
Buccal mucosa
(MN/1,000)
Nasal epithelium
(MN/1,000)
Blood
(MN/1,000)
Blood
(SCEs/cell)
Before course
0.046 ±0.17
0.41 ±0.52
4.95 ± 1.72
7.72 ± 1.26
After course
0.60 ± 1.27a
0.50 ±0.67
6.36 ± 2.03a
7.14 ±0.89
'Statistically significant (p < 0.05), as calculated by the authors.
Source: Suruda et al. (1993).
This document is a draft for review purposes only and does not constitute Agency policy.
4-364 DRAFT—DO NOT CITE OR QUOTE
-------
1 Another group (Titenko-Holland et al., 1996) also reported a significant increase in MN
2 frequency of buccal, but not nasal, epithelial cells from mortuary students exposed to embalming
3 fluid. In this study, 28 out of 35 students were sampled before and after a 90-day embalming
4 class. The mean formaldehyde exposure for the subjects providing data on buccal cell MNs was
5 14.8 ± 7.2 ppm-hours (18.2 ± 8.8 mg/m3-hours) for the entire 90-day period and 16.5 ± 5.8 ppm-
6 hours (20.3 ±7.1 mg/m3-hours) for students providing data on nasal cell MNs. Cells were
7 recorded as having either whole chromosomes with centromeres (MN+) or acentric fragments
8 and no centromeres (MINT). Cells with multiple nuclei were present only in samples taken after
9 exposure to embalming fluid. There was a ninefold increase in the MN frequency in buccal cells
10 (p < 0.5) and only a twofold increase (p > 0.05) in nasal cells. In addition, there was a twofold
11 increase in the MN+ frequency in buccal cells (Table 4-86). The authors suggested that
12 chromosomal breakage appears to be the primary mechanism of MN formation.
13
14 Table 4-86. Incidence of MN formation in mortuary students exposed to
15 formaldehyde for 90 days
16
Sampling
point
Buccal cells (n = 19)
Nasal epithelial cells (n = 13)
Total MN
MN+
MN"
Total MN
MN+
MN"
Pre-exposure
0.6 ±0.5
0.4 ±0.4
0.1 ±0.2
2.0 ± 1.3
1.2 ± 1.3
0.5 ±0.5
Postexposure
2.0 ± 2.0a
1.1 ± 1.3
0.9 ± l.la
2.5 ± 1.3
1.0 ±0.8
1.0 ± 0.6a
Rvalue
0.007
0.08
0.005
0.20
0.31
0.03
17
18 aStatistically significant at the level shown, as calculated by the authors.
19
20 Source: Titenko-Holland et al. (1996).
21
22
23 Ying et al. (1997), however, observed higher frequencies of MNs in the nasal exfoliative
24 cells (3.85 ± 1.48 versus 1.20 ± 0.676, paired t-test,/> < 0.001) and oral exfoliative cells (0.857 ±
25 0.558 versus 0.568 ± 0.317,p< 0.001) after formaldehyde exposure, although there was no
26 significant increase in the frequency of lymphocyte MNs (p > 0.05) in students exposed to
27 formaldehyde in anatomy classes (three classes per week for 3 hours over an 8-week duration).
28 In this study, blood samples and nasal swabs were collected before and after the study. The
29 TWA concentration of formaldehyde in anatomy laboratories and student dormitories was 0.508
30 ± 0.299 mg/m3 and 0.012 ± 0.0025 mg/m3, respectively, suggesting that nasal mucosa cells
31 exposed through respiration are the primary target of formaldehyde-induced genotoxicity.
32 In a different study (Ying et al., 1999), however, the same group showed that exposure to
33 formaldehyde affected the composition of lymphocyte subsets (B cells, total T cells, T helper-
34 inducer cells, T cytotoxic-suppressor cells), but no significant difference was reported between
35 lymphocyte proliferation rate and SCEs at the given levels and durations of formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
4-365 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
exposure. This study involved 23 nonsmoking students exposed to 0.508 ± 0.299 mg/m3
formaldehyde for a period of 8 weeks (3 hours , 3 times per week).
Burgaz et al. (2002) reported significantly (p < 0.05) higher mean MN frequencies in
buccal mucosal cells from shoe workers as well as anatomy and laboratory workers (0.62 ±
0.45% and 0.71 ± 0.56%, respectively) compared with unexposed controls (0.33 ± 0.30%). In
this study, the measured air concentrations of formaldehyde in the breathing zone of the anatomy
and pathology laboratory workers were between 2 and 4 ppm (2.5 and 5 mg/m3). MN count per
3,000 cells was measured in buccal smears from shoe workers and from anatomy and pathology
staff, and eighteen male university staff were used as controls.
In a critical review, Speit and Schmid (2006) examined data from studies that have
reported the formation of MNs in nasal or buccal cells of persons either environmentally or
occupationally exposed to formaldehyde. The authors identified a number of issues relating to
study design, exposure regimen, and confounding factors, including MN levels in nasal and
buccal cells well below established background levels, reports limited by the number of cells
observed, variation in standard techniques, and non-concordance between buccal and nasal
findings. However, the authors concluded that, despite these limitations, the weight of evidence
supports the finding that formaldehyde may be genotoxic in human cells in direct contact with
formaldehyde.
4.3.4.2.2. Peripheral blood lymphocytes. Mature lymphocytes are present at the POE as
intraepithelial lymphocytes and within germinal centers in the mucosa. Because more
lymphocytes may be available in the nasal mucosa than the oral mucosa, mouth versus nose
breathing may contribute to variability in findings. Since some of the lymphocytes traffic around
the body, it is reasonable to find clastogenic effects in these relatively long-lived cells reflected
in peripheral blood lymphocytes. Thus, lymphocytes proliferating in response to antigen would
be more vulnerable to DNA reactivity of formaldehyde and to the clastogenic effects in general.
A cytogenetic evaluation by Fleig et al. (1982) of 15 employees exposed for an average
of 28 years in a formaldehyde manufacturing plant revealed no statistically significant increase
in the frequency of CAs in peripheral blood lymphocytes compared with a matched control
group. Likewise, in a different study (Thomson et al., 1984), no compound-related differences
were evident in the frequency of CAs and MNs in lymphocytes from six pathology workers and
five unexposed controls.
Bauchinger and Schmid (1985) observed an increased incidence of CAs (dicentric and
ring chromosomes) in the peripheral lymphocytes of 20 male paper mill workers and supervisors
exposed to formaldehyde (average exposure of 14.5 years) compared with unexposed workers.
This document is a draft for review purposes only and does not constitute Agency policy.
4-366 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
When workers and supervisors were analyzed separately, significant increases were only seen for
supervisors. The average length of exposure for supervisors (n = 11) and workers (n = 9) was
18.9 years and 7.2 years, respectively. Information regarding formaldehyde concentrations for
the two groups was not provided. However, the incidence of SCEs among workers was actually
slightly lower than among the 20 controls. In contrast, the frequency of SCEs in peripheral
lymphocytes of 18 nonsmoking formaldehyde workers was increased over controls (8.24 ± 0.89
versus 6.38 ± 0.41) (Ye et al., 2005) (described in Section 4.3.4.2.1).
Vargova et al. (1992) observed that the percentage of aberrant cells and number of breaks
per cell in the peripheral blood lymphocytes of formaldehyde-exposed workers was 3.08 and
0.045 versus 3.6 and 0.080 in controls in a pressed board factory, respectively, suggesting both
groups to be at an increased risk. However, normal unexposed population had only 1-2%
aberrant cells. The authors also noted that the mitotic index was significantly decreased in
exposed workers compared with controls.
Kitaeva et al. (1996) evaluated the genotoxic effects of formaldehyde among 15
industrially exposed workers and 8 academic laboratory instructors and observed an increase in
the frequencies of CAs and MNs in the lymphocytes of exposed subjects compared with
6 unexposed controls.
Shaham et al. (1996, 1997) found significantly higher levels of DPXs and SCEs in
peripheral blood lymphocytes of workers occupationally exposed to formaldehyde (physicians
and technicians) compared with unexposed control workers. The authors also observed a linear
relationship between years of exposure to formaldehyde and levels of DPXs and SCEs.
Formaldehyde-induced genotoxicity has also been reported in peripheral blood
lymphocytes of anatomy class students and mortuary workers. Vasudeva and Anand (1996) did
not observe significant differences in the incidences of CAs between the formaldehyde exposed
students and the matched, unexposed controls. In this study, peripheral blood lymphocytes from
30 medical students exposed to formaldehyde in a gross anatomy laboratory for 15 months with
average exposures of less than 1 ppm (1.23 mg/m3) formaldehyde were used.
He et al. (1998) used the cytokinesis-blocked MN (CBMN) assay to detect the frequency
of micronucleated peripheral lymphocytes in 13 students exposed to formaldehyde during a
12-week (10 hours/week) anatomy class. Sampling of breathing zone air showed a mean
concentration of 2.37 ppm (3.17 mg/m3). Ten students from the same school, without exposure
to formaldehyde, were used as controls. CAs and SCEs were observed in both groups, and there
were significant increases (p < 0.01) in the frequencies of micronucleated cells and CAs in the
formaldehyde-exposed group compared with the control group.
This document is a draft for review purposes only and does not constitute Agency policy.
4-367 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
In a study involving 97 plasticware workers (34 males and 63 females) exposed to 0.5 to
0.9 mg/m3 formaldehyde, 4.4 to 6.2 mg/m3 styrene and 0.5 to 0.75 mg/m3 phenol for 2 months
to 25 years, Lazutka et al (1999) observed significantly higher CAs than controls (non-exposed
donors matched by age and similar smoking habits as the exposed workers). Although workers
with short and long exposures showed significant increases in the frequency of CAs, the
cytogenetic damage did not increase with exposure duration.
Sari-Minodier et al. (2001), using the CBMN assay in anatomy/pathology laboratory
workers, reported higher frequency of micronucleated peripheral blood lymphocytes than in
matched controls.
Shaham et al. (2002) observed a mean number of 0.27 SCEs per chromosome in the
peripheral lymphocytes of an exposed cohort compared with 0.19 in controls (p < 0.01). This
study involved 90 individuals employed in 14 hospital pathology laboratories and 52 unexposed
controls.
Yu et al. (2005) reported dose-dependent increase in MNs and comet assay parameters
(olive tail moment and comet tail length) in peripheral lymphocytes in 151 workers from two
plywood factories compared with 112 unexposed controls. The TWA exposure level in the
working environment was 0.1-7.88 mg/m3 (0.08-6.42 ppm) formaldehyde compared with a
background level of <0.01 mg/m3 (<0.008 ppm) formaldehyde applicable to controls. In the
comet assay, the authors observed olive tail moments averaging 0.93 (0.78-1.1), 3.03 (2.49-
3.67), and 3.95 (3.53-4.43) for control, low-, and high-exposure individuals, respectively. For
the same subjects, comet tail lengths were 6.78 (6.05-7.6), 11.25 (10.12-12.5), and 12.59 (11.8-
13.43), respectively. In the CBMN assay, MNs/100 cells were 0.27 ± 0.13, 0.41 ± 0.25, and
0.65 ± 0.36, respectively, for control, low-, and high-exposure individuals.
In a population of 18 workers exposed to formaldehyde at a plant in China, with a mean
employment of 8.5 years (range 1 to 15 years), Ye et al (2005) examined nasal and lymphocytes
for cytogenetic effects. This study also included a second group of 16 waiters who worked in a
newly fitted ball room for 12 weeks with a low level exposure to formaldehyde from building
material, tobacco smoke and furniture and a group of 23 college students as a control group. The
background indoor air conentraton of 0.009 ppm formaldehyde was reported in students' dorms.
Significantly increased frequencies of MNs in the nasal mucosal cells and SCEs in peripheral
blood lymphocytes were reported for the workers, but not the waiters in this study.
Orsiere et al. (2006) reported no apparent effect on the DNA damage in peripheral blood
lymphocytes as assessed by a chemiluminescence microplate assay in pathology and anatomy
laboratory workers (n = 59) before and after a 1-day exposure to formaldehyde. This study had
59 exposed workers and 37 controls. However, with the CBMN assay, the authors reported
This document is a draft for review purposes only and does not constitute Agency policy.
4-368 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
statistically significant differences in the frequency of binucleated micronucleated cells
(1.69 ± 0.93 versus 1.11 ± 0.6%) in exposed versus control subjects. Discrimination between
clastogenic and aneugenic events by using FISH with a pan-centromeric DNA probe resulted in a
higher rate of binucleated micronucleated cells (1.91 ± 1.01 versus 1.19 ± 0.56% in controls) and
showed that the frequency of centromeric nuclei was higher in the exposed group than in
controls, though not significantly. Among the centromeric MNs, the frequency of MNs with
only one centromere (Cl+MN) was significantly greater in pathologists/anatomists than in
controls (1.1 ± 0.62 versus 0.31 ± 0.24%, p < 0.001). The authors interpreted their data on
monocentromeric nuclei in anatomists/pathologists as an indication that formaldehyde exposure
might be associated with aneugenic (rather than clastogenic) events.
Based on pooled analysis of two reports (Iarmarcovai et al., 2006a, b) (Table 4-87), MN
frequency ratios in the peripheral lymphocytes of cancer patients, welders, and
anatomists/pathologists were significantly increased compared with the corresponding controls.
The data were taken from three biomonitoring studies by using CBMN/FISH. The incidence of
MNs was scored and then evaluated further for the presence of centromere-negative MNs
(C-MNs), centromere-positive MNs (C+MNs), and, for the latter case, those containing a single
centromere (Cl+MNs) and those containing two or more centromeres (Cx+MNs). Applying
their findings to considerations of the aneugenic mechanism of action of formaldehyde, the
authors hypothesized that the use of centromeric signals enables the identification of endpoints
representing impaired chromosomal migration (with Cl+MN formation) or centrosome
amplification (with Cx+MN formation).
Table 4-87. Multivariate repression models linking genomic
instability/occupational exposures to selected endpoints from the MN assay
Study populations
Number
MNa
C-MN
C+MN
Cl+MN
Cx+MN
Cancer patients versus
controls
10/10
1.85
(1.18-2.87)
2.05
(1.07-3.94)
1.81
(1.02-3.21)
1.68
(0.80-3.53)
1.28
(0.63-2.59)
Welders versus controls
27/30
1.37
(1.09-1.72)
1.39
(0.99-1.95)
1.37
(1.03-1.83)
1.10
(0.80-1.53)
1.31
(0.99-1.74)
Pathologists/anatomists
versus controls
18/18
1.28
(0.86-1.90)
0.79
(0.46-1.36)
1.65
(1.05-2.59)
3.29
(2.04-5.30)
0.68
(0.38-1.20)
aBolded values indicate statistical significance (p < 0.05).
Source: Iarmarcovai et al. (2006b).
Recently, Costa et al. (2008) observed a significant increase in the genotoxicity of
formaldehyde-exposed pathological anatomy laboratory workers (n = 30) compared with
controls (n = 30) in cytogenetic assays. In this study, the authors evaluated the level of exposure
This document is a draft for review purposes only and does not constitute Agency policy.
4-369 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
to formaldehyde near the breathing zone of workers, and TWA of exposure was calculated for
each subject, giving a mean level of exposure to be 0.44 ± 0.08 ppm (range: 0.04-1.58 ppm). As
compared with control subjects, peripheral blood lymphocyte cultures of formaldehyde exposed
workers showed significant increases in MN frequency (5.47 ± 0.76 versus 3.27 ± 0.69;p =
0.003), SCEs (6.13 ± 0.29 versus 4.49 ± 0.16; p < 0.05), and comet assay as determined by tail
length (TL) (60.00 ± 2.31 versus 41.85 ± 1.97; p < 0.05). In addition, Costa et al. (2008)
observed a positive correlation between formaldehyde exposure levels and MN frequency (r =
0.384;p = 0.001) and TL (r = 0.333;p = 0.005) (Table 4-88). However, polymorphic genes of
xenobiotic metabolizing and DNA repair enzymes did not show any significant effect on the
genotoxic endpoints. This is the lowest level of exposure to formaldehyde in the studies
observed so far, wherein a clear indication of genotoxic effects of formaldehyde was
demonstrated.
Table 4-88. Genotoxicity measures in pathological anatomy laboratory
workers and controls
MN assay
SCEs
Comet assay
Mean MN ± SEM
Mean SCE ± SEM
Mean TL (|aM) ± SEM
(range)
(range)
(range)
Controls
3.27 ±0.69
4.49 ±0.16
41.85 ± 1.97
(n = 30)
(0-17)
(3.10-3.06)
(28.85-66.52)
Exposed
5.47 ±0.76
6.13 ±0.29
60.00 ±2.31
(n = 30)
(1-17)
(3.64-8.80)
(33.76-99.09)
p value
0.003
<0.05
<0.05
Source: Costa et al. (2008).
4.3.5. Summary of Genotoxicity
Formaldehyde's genotoxicity has been demonstrated in a variety of in vitro and in vivo
test systems measuring a variety of genetic endpoints. Formaldehyde forms predominantly
DPXs that are detected in cell-free systems and single cells in vitro. DPXs are formed in nasal
epithelia but not in extra-nasal passages of rodents, which are completely removed within a day
after formation. In vivo data in human and mammalian cells demonstrate that formaldehyde is
genotoxic at the site of first contact, including cells of the mouth or the nose. DPXs are also
detected in nasal and extra-nasal tissues of monkeys, suggestive of direct effects of formaldehyde
in tissues that correspond to observed tumor sites (nasal and nasopharynx) in humans. In
addition, this is used as a basis for cross-species comparison with humans. Formaldehyde-DNA
adducts are labile and constitute a minor fraction of the DNA-reaction products and are less
likely to play an important role in the genotoxicity of formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
4-370 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Formaldehyde clastogenicity has been demonstrated by the induction of SCEs, SSBs,
MNs, and CAs in cultured mammalian cells. Formaldehyde induces mutations in salmonella and
escherichia bacterial strains that contain an AT base pair at the primary reversion site that is used
to detect oxidative compounds and cross-linking agents without metabolic activation by
exogenous enzyme-activating systems. Formaldehyde induces mutations in cultured mammalian
cells at levels that do not cause significant toxicity. Despite formaldehyde's reactivity and
mutagenicity in isolated mammalian cells, clear evidence of mutagenicity does not emerge from
animal bioassays.
Formaldehyde exposure causes differential induction of MNs in human nasal epithelial
and buccal epithelial cells, which is significant in industrial exposure workers and students
working in anatomy or mortuary science, respectively. However, recent data and data from
larger studies support a finding of increased MNs in blood lymphocytes, although the issue
remains controversial because of issues relating to study design, exposure regimen, and
confounding factors, including MN levels in nasal and buccal cells well below established
background levels, reports limited by the number of cells observed, variation in standard
techniques, and non-concordance between buccal and nasal findings (Speit and Schmid, 2006).
Several clastogenic effects, such as induction of MNs, SCEs, and CAs, were seen in human
peripheral blood lymphocytes; however, the data are not very clear. Formaldehyde exposure also
caused p53 mutations in rat nasal carcinomas with the expression of mutant p53 protein.
Overall, induction of DPXs as a predominant lesion in vitro and in vivo, clastogenicity,
and mutagenicity with locus-specific mutations in nonhuman and human cells supports the
concept of genotoxic action of formaldehyde at the POE.
A summary of the genotoxicity of formaldehyde in humans is presented in Table 4-89.
4.4. SYNTHESIS AND MAJOR EVALUATION OF N ON CARCINOGENIC EFFECTS
The adverse health effects due to formaldehyde exposure have been extensively studied
in humans and in animal models. Studies of human exposure include occupational exposures,
environmental exposures, and clinical studies of intentionally exposed subjects (Section 4.1).
Occupational exposures are primarily due to inhalation and dermal contact. Animal studies are
available for a variety of routes of exposure, including inhalation, oral, dermal, and intravenous
and I.P. injections (Section 4.2). Additionally, as discussed in Chapter 3, in vitro studies address
biological activity and the metabolic fate of formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
4-371 DRAFT—DO NOT CITE OR QUOTE
-------
K
s
TO
>3
Table 4-89. Summary of human cytogenetic studes
S"4
>3*
§•
to
s
Study population
N
Exposure time (years)
Formaldehyde
concentration (ppm)
Cytogenetic
observations
Reference
Range
Mean
Range
Mean (TWA)
CAs
SCEs
MNs
Analyses of nasal and/or buccal cells
Plywood workers
Age and sex matched controls
15
15
2-19
6.8
0.32-0.83
(1)
+ nasal
Ballarin et al. (1992)
Male mortuary science students
Female mortuary science students
22
7
Buccal and nasal swabs
taken before and after
first 9 weeks of
embalming course
0.1-4.3
1.4
+ buccal
- nasal
- buccal
- nasal
Suruda et al. (1993)
Mortuary science students3
28
Buccal and nasal swabs
taken before and after
first 9 weeks of
embalming course
0.1-4.3
1.4
+buccalb
- nasal
Titenko-Holland et al. (1996)
Female anatomy faculty
Male anatomy faculty
Controls (Females)
8
5
7
NA
23.6
25.6
NA
NA
+ buccal
- buccal
Kitaeva et al. (1996)
Anatomy students
25
Buccal and nasal swabs
taken before and after
8-week anatomy course
0.06-1.06
(0.508)
+ buccal
+ nasal
Ying et al. (1997)
Anatomy/ pathology staff
Controls (University staff)
28
18
1-13
4.70
2-4
NA
+ buccal
Burgaz et al. (2002)
Workers at a formaldehyde plant
Controls
18
23
1-15
8.5
0.8
+ nasal
Ye et al. (2005)
Volunteers
21
10 days
13.5 ppm-hours
- buccal
Speit et al. (2007b)
Analyses of peripheral lymphocytes
Manufacturing workers
Age and sex matched controls
15
15
23-35
28
<5 1971
<1 later
-
Fleigetal. (1982)
Pathology workers
Controls
6
5
4-11
0.9-5.8
-
Thomson et al. (1984)
o
I § i
N
S, to &-
^ a ^
TO S
s >•
§ 3
U>
-J
to
-------
K
s
TO
>3
Table 4-89. Summary of human cytogenetic studes
S"4
>3*
§•
to
s
Study population
N
Exposure time (years)
Formaldehyde
concentration (ppm)
Cytogenetic
observations
Reference
Range
Mean
Range
Mean (TWA)
CAs
SCEs
MNs
Anatomy students0
8
10-week class
1.08-1.99d
0.08-0.6e
1.2
0.3
+
Yager etal. (1986)
Papermakers
Controls
20
20
2-30
14.4
<3
NA
+f
-
Bauchinger and Schmid (1985)
Wood workers
Controls
25
19
<5 to <16
0.45-8.6
NA
-
Vargova et al. (1992)
Male embalming students
Female embalming students
22
7
Blood sampled before
and after first 9 weeks
of embalming course
0.15—4.3
1.4
—
+
Suruda et al. (1993)
Manufacturing workers
Anatomy faculty
Controls
15
8
6
10
17
Up to 4
NA
NA
+
ND
+
Kitaeva et al. (1996)
Medial students
Controls
30
30
Sampled near end of
15-month term
<1
NA
-
Vasudeva and Anand (1996)
Anatomy students
Controls (students)
13
10
12-week class]
2.31s
+
+
+
He at al. (1998)
Physicians
Technicians
Controls (age matched/unexposed)
6
7
20
2-24 10
2-25 | 15
3.1-2.8
1.6
+
+
Shaham et al 1997
Anatomy students
23-25
Blood samples taken
before and after 8-week
anatomy course
0.06-1.06
(0.508)
-
-
Ying et al. (1999, 1997)
Female anatomy/pathology lab workers
Controls (Women)
10
27
1-16
8.9
1.2-15.1
NA
+
Sari-Minodier et al. (2001)
Hospital pathology workers'1
Controls
90
52
1-39
15.4
0.04-0.71
0.72-5.6
0.4
2.24
+J
+
Shaham et al. (2002)
o
I § i
N
S, to &-
^ a ^
TO S
s >•
§ 3
U>
-J
u>
-------
K
s
TO
>3
Table 4-89. Summary of human cytogenetic studes
S"4
>3*
§•
to
s
Study population
N
Exposure time (years)
Formaldehyde
concentration (ppm)
Cytogenetic
observations
Reference
Range
Mean
Range
Mean (TWA)
CAs
SCEs
MNs
Workers at a formaldehyde plant
Controls
18
23
1-15
8.5
0.8
+
-
Ye et al. (2005)
Workers at two plywood factories
Controls
151
112
ND
0.08-6.42
+
Yu et al. (2005)
Pathology or anatomy workers
Controls
59
37
ND
<0.1-20.4k
2k
+
Orsiere et al. (2006)
Pathologists
Controls
18
18
ND
0.4-7.0k
2.3k
+
Iarmarcovai et al. (2006a, b)
Pathological anatomy lab workers
Controls (21 females and 9 males)
30
30
0.5-27
11
0.04-1.58
0.44
+
+
Costa et al. (2008)
Plasticware workers
Controls (non-exposed donors)
97
90
2 mo to 25
yrs
0.5-0.9
mg/m3
+
Lazutkaetal 1999
Wood workers
Controls
40
22
NR
NR
+
Chebotarev et al 1986
School children (1984)
School children (1985)
School children (1986)
Controls (1984)
Preschool controls (1984)
Preschool children (1984)
20
16
18
17
24
13
0.17-0.3
0.26
0.11
0.03
0
0
+
+
Neri et al 2006
Phenolformaldehyde resin workers
Controls
31
74
0.33-30 yr
0.41
0
+
Suskov and Sazonova 1982
o
2 »
5 s
to
o
s
>S
TO
TO'
*
aSame population in Surada et al. (1993) but different slides used. Nineteen complete slide sets for buccal analysis and 13 complete slide sets for nasal epithelial
cell analysis.
"f* bNot dose related; both low- and high-exposure groups had same SCE increase.
-j °Each student sampled before and after 10-week anatomy class.
dBreathing zone samples.
-------
eRoom air samples.
5 a
^3 increase only in 11 supervisors. See text for details.
^ 5 8Average breathing zone during dissection procedure.
§>;¦»• ~ ... ...
o hExposed and controls from 14 hospitals.
to ~
>3
5" 'Low- and high-exposure groups established but numbers not provided.
o s JNot dose related; both low and high groups had same SCE increase.
^ i kDescribed as "mean concentrations for sampling times of 15 minutes."
<§ a S CAs = chromosomal aberrations; SCEs = sister chromatid exchanges; MNs
^ >• applicable.
& a
a,
^ >! ^
" § 5
s >•
§ 3
4^
U>
-J
micronuclei; TWA = time-weighted average; ND = not determined; NA = not
-------
1
2
3
4
5
6
7
8
9
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
Taken together, the human and animal studies support numerous health effects, not only
at the POE as expected for a reactive gas but also on pulmonary function, neurobehavioral
function, reproduction, development, immunomodulation, and sensitization (atopy, asthma). The
discussion below provides a description of the adverse effects seen in each area, summarizing the
data for both human and animal studies. MOA data are discussed where information regarding
formaldehyde's biological activity may be linked to the observed adverse health effects.
4.4.1. Sensory Irritation
Sensory irritation of the eyes, nose, and throat is reported in humans upon direct contact
with formaldehyde gas during inhalation exposures (Holmstrom and Wilhelmsson, 1988;
Ritchie and Lehman, 1987) and includes irritation resulting from acute exposures (Lang et al.,
2008; Yang et al., 2001; Krakowiak et al., 1998; Kulle, 1993; Green et al., 1989, 1987; Kulle et
al., 1987; Sauder et al., 1987, 1986; Schachter et al., 1987, 1986; Witek et al., 1987; Day et al.,
1984; Bender et al., 1983; Weber-Tschopp et al., 1977). Controlled exposures in inhalation
chambers confirm the specificity of these responses to formaldehyde exposure and allow for
assessment of these symptoms through both subjective and objective measures (Kulle, 1993;
Holness and Nethercott, 1989; Green et al., 1987; Kulle et al., 1987; Sauder et al., 1986; Weber-
Tschopp et al., 1977). Eye irritation may be reported as itching, burning, and general discomfort.
Tearing, redness of the eyes, and increased blink frequency are observed and may be quantified
in exposure under controlled conditions (Lang et al., 2008; Yang et al., 2001; Andersen and
Molhave, 1983; Weber-Tschopp et al., 1977; Schuck et al., 1966). Eye irritation appears to be
the most sensitive endpoint in most individuals and may be observed after short exposures
(195 minutes at 0.5 ppm: Lang et al. [2008]; 30 seconds at 1.65 ppm: Yang et al. [2001]).
Itching, burning, and discomfort of the nose, which may be accompanied by increased
mucous production (runny nose), are reported by individuals exposed via inhalation (Krakowiak
et al., 1998; Kulle, 1993; Green et al., 1987; Kulle et al., 1987; Weber-Tschopp et al., 1977).
Throat irritation may also be described subjectively as itching and burning and is often
accompanied by a cough (Krakowiak et al., 1998). Symptoms of eye and mucous membrane
irritation are also reported in numerous rodent studies and support the health effects reported in
humans (see Section 4.1.1.1). Although dermal contact may result in dermatitis and an apparent
hypersensitivity reaction, symptoms do not present upon contact as sensory irritation. There are
no human or animal data that assess sensory irritation from oral exposures.
The time to onset of sensory irritation symptoms and severity of the sensory irritation are
a function of both the air concentration and duration of exposure. Additionally, nose and throat
irritation becomes more prominent at higher exposures and longer duration of exposure (Kulle,
This document is a draft for review purposes only and does not constitute Agency policy.
4-376 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
1993; Kulle et al., 1987). Controlled human laboratory exposures (Yang et al., 2001; Kulle,
1993; Kulle et al., 1987; Cain et al., 1986; Andersen and Molhave, 1983) provide more direct
exposure-response evidence for sensory irritation. These studies are limited to healthy
nonsmoking individuals. Two studies (Cain et al., 1986; Andersen and Molhave, 1983)
document discomfort and irritation of the eye in response to acute exposures as low as 0.25 ppm.
Dose-response relationships are reported in a number of different ways: as an incidence of the
reported symptom among subjects, as a score for severity of the symptom, or in some cases as a
subjective measure, such as blink frequency for eye irritation.
Symptoms of sensory irritation, including eye irritation (burning watering, increased
blinking), nasal irritation (rhinitis, itching/burning), throat/respiratory tract irritation (wheezing,
coughing, phlegm production), have been reported in numerous worker cohorts. Occupational
exposure environments include hospital and medical settings, students, and industrial workers
(Takahashi et al., 2007; Takigawa et al., 2005; Krakowiak et al., 1998; Akbar-Khanzadeh et al.,
1994; Uba et al., 1989; Horvath et al., 1988; Schachter et al., 1987). Formaldehyde levels often
vary in a work environment and peak as well as average exposures may be used to report
occupational exposures. Although sensitive individuals often remove themselves from an
irritating workplace (the HWE), eye, nose, and throat symptoms are still reported in this
environment. Among workers in a plant where formaldehyde resins were used, those exposed to
an average of 210 ppb formaldehyde reported increased symptoms above those in the control
population (Holmstrom and Wilhelmsson, 1988).
These effects have been noted in students, particularly medical students, who are exposed
to formaldehyde in cadaver labs. In a study of 24 formaldehyde-exposed anatomy students
(personal breathing zone samples 0.73 ppm, range 0.49-0.93) (Kriebel et al., 1993), eye, nose,
and throat irritation was present when comparing rates of irritation from the end or middle of
class to before the start of class. Takahashi et al. (2007) showed that 143 medical students
reported various symptoms (including eye and throat irritation) and that the percentage of
students reporting symptoms increased between the beginning (measured after the first day of
class) and the end of the course (2 months later). After the first day of class, approximately 35%
of students reported eye soreness and about 15% reported throat irritation.
Sensory irritation has also been noted in occupational settings. Horvath et al. (1988)
compared irritation symptoms between 109 workers at a particleboard manufacturing plant and
264 workers at food plants. Eye, nose, and throat irritation were more common among the group
in a particleboard manufacturing facility, exposed to a mean concentration of 0.40 mg/m3.
Similarly, Alexandersson and Hedenstierna (1988) reported that the frequency of eye, nose, and
throat irritation was significantly greater (65.8%) in 38 workers exposed to formaldehyde and
This document is a draft for review purposes only and does not constitute Agency policy.
4-377 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
solvents in lacquers as compared with 18 nonexposed individuals working at the same factory
(16.7%). Holmstrom and Wilhelmsson (1988) conducted a study at a chemical plant and
reported nasal and eye discomfort in 64 and 24%, respectively, of workers (n = 70) exposed to
formaldehyde (range 0.05-0.50 mg/m3 with a mean of 0.26 mg/m3) versus 25 and 6%,
respectively, in nonexposed desk clerks (n = 36). Holness and Nethercott (1989) reported
significant increases in eye irritation (42 versus 21%) and nose irritation (44 versus 16%) among
84 funeral service workers (active embalmers, > 10-year experience) as compared with
38 students and individuals from a service organization. The exposure concentration in both
groups was 0.36 and 0.02 ppm, respectively.
Reports of similar symptoms are correlated to indoor residential exposures, providing
exposure-response relationships for the general population in low-level chronic exposure
scenarios. Ritchie and Lehnen (1987) surveyed residents in 2,000 homes classified as having
formaldehyde concentration <0.1 ppm, 0.1-0.3 ppm, and >0.3 ppm. A LOAEL of 200 ppb was
established from the results of Ritchie and Lehnen (1987). Liu et al. (1991) report irritant effects
associated with formaldehyde exposure in mobile homes, where formaldehyde concentrations
ranged from the 0.01 ppm detection limit to 0.46 ppm. Eye irritation (60%), nose/throat
irritation (10—20%), or headache (< 10%) were reported in residents.
MOA
The mucosae of the URT, oral cavity pharynx, and upper airways are complex tissues,
where epithelial and goblet cells predominate. In addition, the nasal mucosa is highly enervated.
The main nerves include the trigeminal nerve and olfactory sensory cells (olfactory epithelium,
the vomeronasal organ, and the organ of Masera) (Feron et al., 2001). A possible MOA for
sensory irritation includes formaldehyde-induced stimulation of the trigeminal nerve (though
whether formaldehyde acts as a direct agonist is unknown). Trigeminal nerve stimulation in the
nasal passages transmits signals to the CNS, which then sends efferent signals back to the nasal
tissues, causing sensory irritation, and possibly systemically via vagal nerve stimulation,
resulting in more systemic effects.
Animal studies are potentially useful models for understanding mechanisms of toxicity,
especially where sufficient human data do not exist. While experimental animal studies provide
a model of secondary effects, rodents also demonstrate RB, an effect not seen in humans. Thus,
species that exhibit bradypnea (like mice and rats) may not be appropriate for assessing
respiratory endpoints. The mechanism underlying RB includes formaldehyde binding to the
sensory nerve endings of the trigeminal nerve, where signals travel to the CNS. The vagus nerve
transmits the efferent signal to produce smooth muscle contraction. The animals become
This document is a draft for review purposes only and does not constitute Agency policy.
4-378 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
inactive, their core temperatures decrease by several degrees C, and their respiratory rates and
minute volumes decrease. However, this is not to say that trigeminal nerve stimulation is not an
appropriate potential mechanism of action in other species or in humans. Since trigeminal nerve
stimulation has been independently confirmed in species without RB, this mechanism may be a
viable explanation for the observed effects.
4.4.2. Pulmonary Function
Workers chronically exposed to formaldehyde have exhibited signs of reduced lung
function, such as BC, inflammation, and chronic obstructive lung disease. Lung function deficits
have been reported in pre- versus post-shift measurements and as a result of chronic exposures
(Pourmahabadian et al., 2006; Herbert et al., 1994; Malaka and Kodama, 1990; Alexandersson
and Hedenstierna, 1989; Alexandersson et al., 1982). Decreases in spirometric values, including
VC, FEV, FVC, and FEV/FVC, have been reported. Chronic studies (Pourmahabadian et al.,
2006; Herbert et al., 1994; Malaka and Kodama, 1990; Alexandersson and Hedenstierna, 1989;
Alexandersson et al., 1982) also report increased respiratory symptoms, including cough,
increased phlegm, asthma, chest tightness, and chest colds, in exposed workers.
Students have also shown decrements in lung function that are associated with exposure
to formaldehyde in laboratories. Kriebel and colleagues (1993) observed a 2% decrement in PEF
in healthy students attending anatomy classes once per week and a 7.3% decrement in PEF in
students with histories of asthma. The strongest pulmonary response was observed when
examining the average cross-laboratory decrement in PEF in the first 2 weeks of the study
(formaldehyde geometric average concentration of 0.73 ppm). These findings were corroborated
by Kriebel et al. (2001) in which a similar study design was applied to another class of anatomy
students.
Similarly, Akbar-Khanzadeh et al. (1994) compared pre- and postexposure pulmonary
function among students before and after working 3 hours in a laboratory (n = 34). On average,
FVC decreased by 1.4%, FEV3 decreased by 1.2%, FEVi/FVC increased by 1.6%, and FVC-25
75% increased 2.5%. These average percent changes in the control group are —0.3%, 1.30%,
2.3 P/o, and 0.6% but were not statistically significant. In a follow-up study, Akbar-Khanzadeh
and Mlynek (1997) recorded FEV values in 50 exposed students and 36 controls and reported a
larger increase in lung function among controls when compared with cases after 1-3 hours of
exposure that persisted after 3 hours after exposure termination. In a similar study, Fleisher
(1987) reported that approximately 8% of students reported experiencing shortness of breath
during the laboratory with formaldehyde exposure, but none of the students reported shortness of
This document is a draft for review purposes only and does not constitute Agency policy.
4-379 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
breath in the laboratory session with no exposure. However, no objective measurements of
formaldehyde exposure were used.
Unlike the study by Kriebel et al. (1993), Uba et al. (1989) did not find a change in
pulmonary function over the course of the 7 months in a study of 96 anatomy laboratory
students. These negative findings may be attributed to differential cross-shift exposures and to
significant differences in FVC on exposed days.
Deficits in pulmonary function have been reported in occupational or residential exposure
studies (Khamgaonkar and Fulare, 1991; Krzyzanowski et al., 1990; Malaka and Kodama, 1990;
Alexandersson and Hedenstierna, 1989; Kilburn et al., 1985; Alexandersson et al., 1982).
Krzyzanowski et al. (1990) documented a significantly decreased PEFR in children
(298 children) who resided in homes with an average formaldehyde concentration of 26 ppb
(maximum sample value of 140 ppb). Among adults, there was a statistically significant
nonlinear relationship with decreased morning PEFR for formaldehyde concentration <40 ppb
(Krzyzanowski et al., 1990). Similarly, Malaka and Kodama (1990) reported that an average
8-hour TWA formaldehyde exposure of 1.13 ppm from area samples was associated with
statistically significant decrements in FEVi, FEVi/FVC, and FEF25-75% compared with a referent
population. Alexandersson and Hedenstierna (1989) investigated not only the acute effects of
exposure across shift but also measured effects of exposure among some of the same workers
that had been studied 5 years earlier (Alexandersson et al., 1982). Statistically significant
decreases (p < 0.01) in FEVi/FVC and FEF25-75% were noted over the intervening five years in
nonsmokers after correcting for aging. Similar decrements have been documented in laboratory
workers in India (Khamgaonkar and Fulare, 1991) and in factory workers (Kilburn et al., 1985).
Alexandersson et al. (1982) reported only slight deficits in lung function 1 day following
occupational formaldehyde exposure in a carpentry shop in Sweden, where the measured
formaldehyde level was 0.36 ppm (0.47 mg/m3). In this case, subjects were compared with
20 nonexposed workers.
Other studies have found no association between formaldehyde and lung function
(Ostojic et al., 2006; Holness and Nethercott, 1989; Holmstrom and Wilhelmsson, 1988; Horvath
et al., 1988). Ostojic et al. (2006) used an interesting measurement, "diffusing lung capacity"
instead of decrements in FEVI or similar measurements. Similarly, Nunn et al. (1990) assessed
the decrease in FEVi with age and showed no association between formaldehyde exposure and
decreased FEVi. Franklin et al. (2000) did not report an association between FVC or FEV and
the indoor concentrations of formaldehyde in children (ages 6-13), although there were signs of
lower airway inflammation as measured by levels of exhaled NO (Franklin et al., 2000).
Similarly, Main and Hogan (1983) did not observe differences between FEVi or FVC at the end
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
35
of the 34 months between mobile home trailer workers compared with controls who did not work
in trailers. The average exposure was reported as ranging from 0.12 to 1.6 ppm.
Occupational studies share certain limitations, including the potential for confounding by
occupational co-exposures. Also, studies that did not report pre-shift pulmonary function as a
percentage of expected function are less useful to assess potential chronic effects because,
post hoc, it is difficult to calibrate for cross-study comparison due to lack of data on important
pulmonary function determinants, such as age, gender, smoking status, height, and year of birth.
Controlled human studies and studies in nonhuman primates also document changes in
formaldehyde-induced pulmonary dysfunction. Acute exposures of healthy non-asthmatic
volunteers resulted in transient decreases in pulmonary function (e.g., decreased FEVi, FVCi,
FEV3, specific airway conductance) (Green et al., 1987; Sauder et al., 1986). Green et al. (1987)
noted differential responsiveness in formaldehyde-exposed subjects; some were responders while
others were nonresponders. This differential response suggests susceptibility in certain subjects
(Green et al., 1987).
Several animal studies document increased airway resistance and BC following
inhalation exposure to formaldehyde (Nielson et al., 1999; Swiecichowski et al., 1993; Biagini et
al., 1989; Amdur et al., 1960). A study using cynomolgus monkeys (Biagini et al., 1989)
demonstrated that methacholine-induced BC can be similarly induced by acute formaldehyde
exposure (10 minutes at 2.5 ppm). Thus, formaldehyde exposure simulated BC observed after
methacholine challenge, but these effects may not occur by a similar MOA. Similar results were
reported in guinea pigs (Swiecichowski et al., 1993; Amdur et al., 1960), rats (Ohtsuka et al.,
1997), and mice (Nielson et al., 1999).
Deficits in pulmonary function have been documented in occupational as well as
controlled chamber human studies and have been corroborated in animal studies exposed to
formaldehyde. However, some of these deficits are slight or transient. Some studies did not
identify a statistically significant decrease in pulmonary function, and others did not observe any
change at all. Pulmonary function alterations appear to be specifically tied to exposure regimen
and may be reversible but remain, nevertheless, an important symptom often associated with
exposure to formaldehyde.
MOA
Formaldehyde-induced inflammation of the airways may contribute to observed
decreases in measures of pulmonary function. Even short-term inflammatory reactions could
reduce the effective diameter of the conductive airways, resulting in lower respiratory volumes in
a number of functional tests. Formaldehyde-induced trigeminal nerve stimulation contributes to
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
34
airway inflammation, which in turn would reduce airway function. Chronic exposures may
result in increased sensitization or chronic inflammatory responses, which could contribute to the
effects seen in the worker and residential populations.
Formaldehyde-induced pulmonary function deficits may also be in part a result of smooth
muscle contraction in repose to trigeminal nerve stimulation. Trigeminal nerve stimulation
transmits signals to the CNS. The resulting efferent signal from the vagal nerve produces
smooth muscle contraction and may result in decreased pulmonary function. Efferent signaling
has also resulted in release of substance P and other neuromodulator compounds, which may
contribute to BC and sensitization of pulmonary responses (asthma, atopy).
4.4.3. Hypersensitivity and Atopic Reactions
Sensitization to inhalational chemical exposure may manifest as an allergic or asthmatic
response that is characterized by BC or BHR. This sensitization may be a result of immune
involvement, as in the case of hypersensitivity, or a neurogenic sensitization, where a chemical
may directly stimulate inflammation. Asthma is a specific manifestation of IgE-mediated
hypersensitivity, characterized by BHR and airway inflammation, resulting in lower airway
obstruction (Fireman, 2003; Kuby, 1991).
A variety of hypersensitivity reactions have been reported following exposure to
formaldehyde. Rashes and skin reactions have been reported in some individuals after dermal
exposures to formaldehyde. Increased expression of Th-2 cytokines in the lymph nodes of mice
given dermal applications of formaldehyde does indicate the involvement of an immune
component to the observed sensitization (Dearman et al., 2005; Hilton et al., 1998; Arts et al.,
1997). However, the response does not appear to be IgE mediated (Arts et al., 1997; Lee et al.,
1984). Gorski et al. (1992) observed an increase in formaldehyde-mediated neutrophil burst in
dermatitis patients exposed in a controlled chamber study and suggests a putative role of
oxidative stress and reactive oxygen species (ROS).
Inhalation exposure has been associated with increased asthmatic responses in asthmatics
in occupational settings. While few available case reports of bronchial asthma suggest direct
respiratory tract sensitization to formaldehyde gas (Lemiere et al., 1995; Burge et al., 1985;
Hendrick et al., 1982; Hendrick and Lane, 1977, 1975), a greater body of human data provides
evidence of an association between formaldehyde exposure and exacerbation of asthmatic
responses in compromised individuals (Kriebel et al., 1993) and particularly in children
(Rumchev et al., 2002; Garrett et al., 1999; Krzyzanowski et al., 1990). Increased asthma
incidence reported after inhalation exposure to formaldehyde led to a NOAEL of 30 ppb
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
(Rumchev et al., 2002). An increased frequency of respiratory symptoms associated with
asthmatic responses and formaldehyde exposure led to a LOAEL of 30 ppb (Garrett et al., 1999).
Exacerbation of response after formaldehyde exposure has been demonstrated in animal
studies as well. Sadakane et al. (2002) demonstrated that formaldehyde exposure exacerbated
sensitization and challenge with Der f and suggested that formaldehyde exposure may aggravate
eosinophilic infiltration and goblet cell proliferation that accompanies allergic responses.
Several animal studies report increased airway resistance and BC due to inhalation exposures to
formaldehyde (Nielsen et al., 1999; Swiecichowski et al., 1993; Biagini et al., 1989; Amdur,
1960). Changes in pulmonary resistance were observed as early as 10 minutes after exposure
(Biagini et al., 1989), and reported effect levels ranged from 0.3 to 13 ppm. BHR is commonly
associated with allergic Type I hypersensitivity reactions but is not sufficient to demonstrate that
an agent induces Type 1 hypersensitivity.
MOA
The MOA underlying this response has not been elucidated. Formaldehyde-induced IgE
production has been reported in some studies (Vandenplas et al., 2004; Wantke et al., 1996a).
Other studies suggest that this effect does not appear to be immunogenic in nature (Fujimaki et
al., 2004; Lee et al., 1984). Although formaldehyde exposure has been reported to alter cytokine
levels and immunoglobulins in some experimental systems (Fujimaki et al., 2004a; Ohtsuka et
al., 2003), these immunomodulatory effects do not support immunogenically mediated type 1
hypersensitivity.
These decrements may be mediated via neurogenic potentiation (Sadakane et al., 2002;
Riedel et al., 1996; Tarkowski and Gorski, 1995). Tarkowski and Gorski (1995) suggest that
formaldehyde may increase permeability of respiratory epithelium and destruction of
immunologic barriers. Tachykinin NK1 receptor and various neuropeptides (NGF and substance
P) have been implicated in formaldehyde-induced sensitization and lend weight of evidence to a
neurogenic MOA (Van Schoor et al., 2000; Ito et al. 1996).
4.4.4. Upper Respiratory Tract Histopathology
Several studies in occupational workers have reported increased squamous cell
metaplasia and reduced mucociliary clearance in nasal and buccal swabs from humans
occupationally exposed to formaldehyde (Holmstrom et al., 1989; Holmstrom and Wilhelmsson,
1988). Evidence of genotoxic effects include increased MNs and CAs in nasal and buccal
epithelial cells from both workers and students exposed to formaldehyde (Ying et al., 1997;
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
31
32
33
34
35
Titenko-Holland et al., 1996; Suruda et al., 1993) and suggest a potential association between
genotoxicity and altered histopathology.
Numerous animal experimental studies in multiple strains of rats, mice, hamsters, rabbits,
and monkeys describe formaldehyde-induced URT pathology (Flo-Neyret et al., 2001; Roemer
et al., 1993; Reuzel et al., 1990; Monticello et al., 1989; Zwart et al., 1988; Wilmer et al., 1987;
Morgan et al., 1986b, 1983; Swenberg et al., 1986; Buckley et al., 1984). Effects are first
observed in the anterior respiratory mucosa and progress through the nasal passages with
increasing exposure concentration and time. The first observed effect includes damage to the
mucociliary apparatus of the nasal passages in response to formaldehyde. Studies conducted
both in vivo and in vitro demonstrate that formaldehyde disrupts mucus flow and ciliary beat that
are dependent on concentration and duration of exposure. Mucociliary apparatus deficits have
been recorded even after 18 hours of recovery following formaldehyde exposure. The
breakdown of the mucociliary apparatus may allow for increased infection and allow the
underlying epithelium to come into contact with exogenous chemicals.
Formaldehyde is highly reactive and may impact all cells in the nasal mucosa, including
epithelial cells (ciliated, columnar, and cuboidal), goblet cells, sensory neurons, and
intraepithelial lymphocytes. The histologic changes of these processes have been described in all
laboratory animals examined and progress from the anterior nares to the posterior regions of the
nasal passages, including the ETs and olfactory epithelium if the concentration and duration of
exposure are great enough.
Humans and nonhuman primates have significantly less complex nasal passages than
rodents. Formaldehyde has lower peak flux in human nasal tissues compared with rodents,
which are obligate nose breathers, but will penetrate more deeply into the human respiratory tract
than in rodents, since humans lack the autonomic RA response. Additionally, humans may
switch to mouth breathing in the presence of an irritant gas, thus bypassing the sensitive nasal
passages and increasing the tissue dose in the mouth and throat. These differences have been
demonstrated by using nonhuman primates where, at comparable concentrations, tissue
pathology and increased cell proliferation progressed further into the respiratory tract than in
rodents (Monticello et al., 1989). Nonhuman primates share common structural respiratory
components and patterns of breathing and do not have a reflex autonomic apnea response.
Despite the anatomical and physiological differences in breathing patterns and different
exposure parameters between humans and rodents, similar toxic effects are reported in tissues at
the POE in humans and laboratory animals. Several occupational studies have reported
increased squamous cell metaplasia in nasal and buccal samples in response to formaldehyde
exposure (Ballarin et al., 1992; Boysen et al., 1990; Holmstrom et al., 1989), paralleling the
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
histologic effects seen in experimental animal studies. A few human epidemiology studies
suggest increased NPCs (see Section 4.5) as well as oral/buccal tumors in response to
formaldehyde exposure (Shangina et al., 2006; Laforest et al., 2000).
The observed formaldehyde-induced URT toxicity is related to its high reactivity and
solubility. Moreover, additional interspecies differences in the surface area and configuration of
the nasal passages and upper airways will influence the areas of high formaldehyde flux in POE
tissues.
MOA
Formaldehyde-induced damage to the mucociliary apparatus of the nasal passages may
occur because formaldehyde may disrupt mucus flow and ciliary beat that is dependent on
concentration and duration of exposure. Formaldehyde reacts with the mucosal glycoproteins
and thus may contribute directly to the breakdown of the mucus layer. As formaldehyde reaches
the cells of the pseudostratified epithelium in the nasal passages, it exerts a range of effects from
direct damage to cell membrane, intracellular proteins, and DNA to alterations in GSH pools and
increased ROS. Adaptive effects include increased mucus flow and goblet cell proliferation as
well as the transition of respiratory epithelium to more insensitive cuboidal cells. With
continued exposure at sufficient concentration, squamous metaplasia develops, creating a
protective layer of keratinized cells. Gradually, this damage exceeds the cell's ability to
compensate for and repair damage; chronic nasal lesions develop, and the cells die both through
general necrosis as well as programmed cell death, depending on the severity of the cellular
damage (Monticello et al., 1989; Swenberg et al., 1983).
Genotoxic effects have been reported in nasal and buccal lesions taken from both workers
and students exposed to formaldehyde (Ying et al., 1997; Titenko-Holland et al., 1996; Suruda et
al., 1993). MN formation occurs in the more sensitive pseudostratified epithelium of the nasal
passages, nasopharynx, and upper airways, since there is only one layer of epithelial cells that are
constantly regenerating. However, the genotoxicity observed in buccal cells is more difficult to
explain, since buccal basal cells are usually covered by protective keratinized cell layers. Cuts,
sores, or other buccal lesions would increase basal epithelial cells' vulnerability to direct
exposure to formaldehyde.
4.4.5. Toxicogenomic and Molecular Data That May Inform MOAs
Over the past several years, studies have begun to examine the effects of formaldehyde
exposure on gene and protein expression. These include studies on in vivo and in vitro changes
in the global expression of mRNA (transcriptomics) and proteins (proteomics) in the tissues and
This document is a draft for review purposes only and does not constitute Agency policy.
4-385 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
cells of humans and rodents exposed to formaldehyde. Currently (2009), nine "-omics" studies
from five research groups are available. These studies are summarized in Section 5.2 and are
evaluated and discussed in the context of their relevance to informing MO As and the dose-
response characterization briefly here.
In 2002, EPA released the Interim Policy on Genomics (U.S. EPA, 2002c), which
addresses how to use genomic data in regulatory decision making. Although the policy
encourages research in genomics, it places limits on its use, stating that genomic data alone are
not sufficient as a basis for decision making. These data thus cannot currently be utilized as the
critical effect in a chemical risk assessment but can be utilized in a weight-of-evidence approach
on a case-by-case basis. The Science Policy Council developed a white paper entitled Potential
Implications of Genomics for Regulatory and Risk Assessment Applications at EPA (U.S. EPA,
2004). This report described three areas where genomic data might be applied in risk assessment
at EPA: MOA analysis, susceptible population, and mixtures assessments. The genomic data on
formaldehyde thus may be applied to a discussion of MOA.
Toxicogenomics studies have investigated the gene and protein expression changes
resulting from formaldehyde exposure in a variety of respiratory tissues, including nasal tissue
(Andersen et al., 2008; Thomas et al., 2007; Hester et al., 2005, 2003), and, in lung tissue (Lee et
al., 2008, 2007; Sul et al., 2007; Im et al., 2006) used human tracheal cell lines to study genomic
changes after exposure to formaldehyde in vitro. Unfortunately, these studies are not directly
comparable because different gene chip technology platforms were used in different tissues, in
both in vivo and in vitro study designs. In general, the gene and protein expression changes
reflect changes in apoptotic pathway genes, oxidative stress, and tissue remodeling. Andersen et
al. (2008) concluded that there was a threshold level where exposure to formaldehyde (6 ppm)
does not elicit changes in nasal epithelium of F344 rats. Overall, Andersen et al. (2008)
concluded that genomic changes were no more sensitive than tissue responses and that
formaldehyde, being an endogenous chemical, is well handled until some threshold is achieved
when toxicity rapidly ensues with genomic and histologic changes. At about 6 ppm, this largely
involves tissue remodeling (and protection), but regenerative hyperplasia occurs at higher doses.
Andersen et al. (2008) conclude that there is a threshold where exposure to formaldehyde does
not elicit changes in F344 nasal epithelial tissue over the duration examined in this study (i.e.,
15 days). Andersen et al. (2008) argue that this is consistent with bioassays that indicate no
tumor formation in rodents below 6 ppm formaldehyde.
The primary conclusion in the Andersen et al. (2008) paper is that genomic changes,
including those suggestive of mutagenic effects, did not temporally precede or occur at lower
doses than phenotypic changes in the tissue. The authors stated as follows:
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 86 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
"The genomic signatures related to these transitions are for cell membrane and
extracellular components, then inflammation and cellular stress, and eventually
apoptosis. Importantly, these hierarchical models indicate that the tissue
responses at low dose concentrations are qualitatively different from those at high
concentrations and linear extrapolations or extrapolations that specify similar
modes of action at high and low doses would be inappropriate."
4.4.6. Noncancer Modes of Actions
Noncancer health effects of interest span numerous organ systems and include
reproductive and developmental effects, neurological/neurobehavioral effects, and a complex
interaction between inflammation and immune and adverse pulmonary function. To date, no
-omics studies have examined changes in reproductive tissue or altered gene expression in
developing animals. In regard to neurological/behavioral effects, one study (Lu et al. [2008],
described in Section 4.1.1.6) has reported elevations in the mRNA for NMDA receptor subunits
in brain homogenates following exposure to 2.4 ppm. Hester et al. (2003) reported a significant
increase in NMDA receptor subunit transcripts, along with other neuropeptide genes, in nasal
tissue. Together, these changes may relate to formaldehyde-induced sensory irritation and,
perhaps, changes throughout the brain.
In regard to inflammatory, immune, and pulmonary effects, transcript and protein
changes in rodent tracheal tissue and lung tissue indicate that exposure to 3 to 38 ppm
formaldehyde results in genes involved in oxidative stress and cell proliferation and may
additionally increase airway ADH3 levels (Lee et al., 2008; Sul et al., 2007; Yi et al., 2007; Im et
al., 2006; Yang et al., 2005). Together, these data provide evidence for adverse pulmonary
effects that may exacerbate or facilitate asthma.
In lung tissue, Yang et al. (2005) identified three proteins up regulated and one protein
down regulated following 15 days of exposure to about 28 ppm formaldehyde. None of the
proteins corresponded with transcript changes reported by Sul et al. (2007). Interestingly, Sul et
al. (2007) reported that only two transcripts were significantly up regulated in the lung in
response to 5-10 ppm formaldehyde, while 19 were down regulated. In this regard, it is worth
considering that changes in proteins may not relate to their regulation but rather to their overall
percent composition in a cell (relative to other protein changes) before and after exposure. In
addition to transcript changes, Sul and colleagues (2007) reported DNA damage and lipid
peroxidation and noted that the observed down regulation of GR would facilitate oxidative stress,
while the down regulation of phospholipase A2 (PLA2) might represent a mechanism for
mitigating lipid peroxidation. It is worth noting that an increase in either of these genes could
also be argued to support similar conclusions (i.e., that GR is up regulated to increase GSH
levels and that PLA2 up regulation explains lipid peroxidation); this highlights the problem with
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 87 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
interpreting these data. Nevertheless, these studies indicate adverse effects in the rodent lung in
response to 5-30 ppm formaldehyde.
In a hypothesis-driven study by Yi et al. (2007), formaldehyde exposure was shown to
increase lung ADH3 levels. Several studies indicate that allergic responses and hyperreactivity
are uncoupled and may relate to ADH3 expression and activity. Que et al. (2005) has shown
that, in a rodent asthma model, ADH3 knockout mice exhibit similar signs of inflammation but
are protected from bronchoconstriction. Lino dos Santos Franco et al. (2006) provided evidence,
in rodents, that formaldehyde may induce inflammatory responses (e.g., leukocyte infiltration in
the lung) through neurogenic mechanisms but that bronchial tone is mediated by NO. The latter
effect is likely to be mediated by S-nitrosoglutathione GSNO and thus influenced by ADH3.
Interestingly, the single nucleotide polymorphisms (SNPs) that Wu et al. (2007) reported as
linked to other polymorphisms in the promoter region was the one demonstrating protection
against asthma. Hedberg et al. (2001) reported that at least one SNP in the promoter region
reduced ADH3 expression. Together, these data indicate that reduced ADH3 expression might
lower GSNO turnover and bronchial tone, thereby reducing signs of asthma. In this regard, it is
conceivable that wheezing and bronchoconstriction are the symptoms that lead to medical
intervention and not the inflammation per se. Thus, while ADH3 polymorphisms may not cause
asthma, ADH3 polymorphisms may influence hyperresponsivity and the likelihood of asthma
diagnosis. This is discussed further in Section 4.6 on susceptible populations. Formaldehyde
has been shown to accelerate GSNO breakdown (Staab et al., 2008a; Yi et al., 2007); thus,
pulmonary responses to formaldehyde may represent a balance between mechanisms that induce
NO (i.e., inflammation) and those that terminate GSNO (i.e., ADH3).
In regard to -omics changes in blood samples, the apparent limited distribution of
formaldehyde may suggest that these changes are secondary to effects at sites of contact but
could also indicate systemic distribution. As noted elsewhere in this report, bradypnea can
induce changes in dosimetry as well as changes in core body temperature and blood gases
(hypoxia itself induces hypothermia in rodents, and thus the reduction in minute volume and gas
changes may both contribute to hypothermia). These physiological responses (hypothermia and
hypoxia) surely induce changes in gene expression. Observed gene and protein changes in the
blood following formaldehyde exposure could also relate to irritation and inflammation at sites
of contact. In this regard, Im et al. (2006) reported changes in cytokines indicative of Th-2
responses. Altogether, the authors identified 32 proteins altered in the plasma of rats exposed to
formaldehyde. Although no coherent mechanisms are apparent from these changes, the authors
posited that they could serve as biomarkers for formaldehyde exposure. The concordance of
such changes across species remains to be demonstrated.
This document is a draft for review purposes only and does not constitute Agency policy.
4-388 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Li et al. (2007) identified dose-response relationships for six genes in human blood
samples that were putatively associated with formaldehyde exposure. Three of these genes are
reported to inhibit apoptosis and were posited as supporting in vitro data by Tyihak et al. (2001);
however, Li and colleagues (2007) did not report any increase in blood cell count or in Hs 680.Tr
cell counts in vitro (i.e., these changes were not phenotypically linked to changes in cell kinetics
or hematology). However, these findings are not inconsistent with those of Hester et al. (2003)
that indicated no significant increase (or decrease) in nine genes involved in apoptotic pathways.
Finally, serum and glucocorticoid-induced protein kinase 1 (SGK1) was elevated in blood
samples and was posited to relate to possible inflammatory and immune responses.
4.4.7. Immunotoxicity
Results from studies that evaluated the immunotoxicity of formaldehyde are mixed. For
example, most human studies that investigated systemic immune effects by measuring increases
in formaldehyde-specific IgE are negative (Doi et al., 2003; Kim et al., 2001; Palcynski et al.,
1999; Krakowiak et al., 1998; Wantke et al., 1996; Grammer et al., 1990). Vandenplas et al.
(2004) reported a transiently positive increased formaldehyde-specific IgE titer in occupationally
exposed workers. In contrast, Thrasher et al. (1990, 1988, 1987) reported positive
formaldehyde-specific IgE titers in small (six to eight person) case studies of exposed workers,
and Carraro et al. (1997) reported elevated IgE titers in smokers. Grammer et al. (1990) did not
report any differences in albumin IgE in formaldehyde-exposed workers compared with controls.
In a residential study, Pross et al. (1986) found that formaldehyde insulation in homes had no
effect on tested human immunologic parameters.
One study suggests that immune system parameters are perturbed by formaldehyde
exposure. Lyapina et al. (2004) reported decreased immune resistance in all 29 workers exposed
to formaldehyde. This effect was associated with decreased neutrophil respiratory burst activity.
A LOAEL of 700 ppb was established.
Results from animal studies are mixed as to whether formaldehyde causes
immunotoxicity. Leach et al. (1983) reported systemic immunomodulation in F344 rats that was
attributed to formaldehyde exposure, but the formaldehyde effects on measures of humoral and
cell-mediated immunity were not confirmed in B6C3F1 mice (Dean et al., 1984). Jakab et al.
(1992) detected no differences in phagocytic ability of alveolar MPs from mice after
formaldehyde exposure. Formaldehyde-exposed rats that were injected with pneumococcus
antigen or tetanus toxoid produced antibodies in amounts similar to nonexposed, infected control
animals (Holmstrom et al., 1989b).
This document is a draft for review purposes only and does not constitute Agency policy.
4-3 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
However, specific immune parameters appear to be affected by formaldehyde exposure.
For example, increased host resistance and hydrogen peroxide release from peritoneal MPs were
reported and confirmed (Adams et al., 1987; Dean et al., 1984) and suggest a putative role for
ROS. Increased host resistance may be mediated by formaldehyde-induced chronic
inflammation and respiratory mucosal damage that causes an up regulation in MPs and therefore
increases host immunity. Jakab et al. (2002) reported reduced pulmonary bacterial resistance in
mice after exposure to formaldehyde, as determined by increased bacterial loading. This result
contrasts with Dean et al. (1984) and is attributed to differential exposure regimens and
experimental design.
Mode of action
Circulating immune cells present in the mouth and upper airways, such as intraepithelial
lymphocytes, direct a local inflammatory response but may also contribute to systemic responses
through secreted cytokines and soluble factors released into the bloodstream (Togias, 1999).
Altered host resistance and hydrogen peroxide release from peritoneal MPs were reported
and confirmed (Adams et al., 1987; Dean et al., 1984) and suggest a putative role for ROS.
Indeed, increased neutrophilic ROS have been associated with formaldehyde-induced dermatitis
(Gorski et al., 1992), and, conversely, decreased neutrophil respiratory burst activity has been
shown in workers with history of formaldehyde-induced respiratory tract inflammation (Lyapina
et al., 2004). Oxidative stress may occur directly as a result of formaldehyde exposure or as a
secondary consequence to inflammatory responses.
4.4.8. Effects on the Nervous System
There is considerable evidence that formaldehyde exposure causes adverse effects on the
nervous system following inhalation at relatively low exposure levels but little or no information
regarding a possible mechanism of action for these effects. Data regarding adverse effects on the
nervous system following oral exposure are very limited, reflecting a data gap in this area.
Relevant data in animals and humans for several types of neurological endpoints, following
inhalation exposure, are summarized below.
4.4.8.1. Irritant Threshold Detection
Humans are exquisitely sensitive to the irritant properties of formaldehyde, as has been
discussed previously (see Section 4.1.1.1). Animal data confirm the irritant properties of this
compound at very low concentrations (Wood and Coleman, 1995).
This document is a draft for review purposes only and does not constitute Agency policy.
4-390 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.4.8.2. Behavioral Effects
Limited data in humans, as well as more robust data in animals, provide evidence of
behavioral changes following exposure to formaldehyde at levels as low as 0.1 ppm. Studies in
animals have found effects that persist for days to weeks after termination of exposure. In spite
of significant limitations, the available human data are consistent with the animal findings.
Several types of behavior have been evaluated in animals following formaldehyde
exposures. The most consistent findings, demonstrated by multiple laboratories and in multiple
species, have been changes in motor activity, habituation, and learning/memory task
performance. Motor activity and habituation have been evaluated under a variety of exposure
conditions, using both rats and mice. Consistent decreases in activity have been seen in adult
animals (Malek et al., 2004, 2003 a, b; Usanmaz et al., 2002). Senichenkova (1991) and
Sheveleva (1971) also found changes in motor activity in offspring following in utero exposure,
including decreased habituation in juvenile rats exposed in utero. Decreased performance in
learning and/or memory paradigms have been seen in multiple laboratories, also in both rats and
mice (Lu et al., 2008; Malek et al., 2003c; Pitten et al., 2000).
Data from controlled human exposures are very limited, but studies have shown
decreased performance in addition tasks and reaction time tasks following acute exposures to
formaldehyde (Lang et al., 2008; Bach et al., 1990). In contrast, Andersen and Molhave (1983)
indicated they found no change in performance on several types of tasks (including addition,
multiplication, and card punching) following acute exposure to volunteers, but supporting data
were not provided.
Data for humans are also available from epidemiology studies of individuals who were
occupationally exposed to formaldehyde. A variety of neurobehavioral deficits, including lack
of concentration and loss of memory, disturbed sleep, impaired balance and dexterity, and
changes in mood, were identified (Kilburn et al., 1987, 1985). However, most of the individuals
evaluated in these studies were also occupationally exposed to other solvents, raising questions
regarding possible confounding of the results due to multiple exposures. In addition, the
formaldehyde exposure information provided in the studies is not sufficient to permit a reliable
dose-response assessment for the effects identified. The types of effects seen in humans in the
available epidemiology studies are, however, consistent with those seen in available animal
studies.
In general (and noting the differences in exposure paradigms and types of tasks),
behavioral effects in animals and humans appear to occur at similar exposure levels. Animal
studies demonstrated LOAELs as low as 100 ppb following acute or repeated exposures (Malek
et al., 2003b, c); human controlled exposure studies have found effects in that same range, with
This document is a draft for review purposes only and does not constitute Agency policy.
4-391 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
LOAELs of approximately 300 ppb following acute exposures (Lang et al., 2008; Bach et al.,
1990).
4.4.8.3. Neurochemistry, Neuropathology, and Mechanistic Studies
Limited data are available regarding neurochemical and neuropathological sequelae of
formaldehyde exposure. Studies from one laboratory (Sorg et al., 2004, 2001) have suggested
that behavioral sensitization to formaldehyde is linked to alterations in HPA control of
corticosterone and changes in mesolimbic dopamine pathways. Neurochemical changes in
response to formaldehyde exposure have also been documented in other laboratories (Fujimaki et
al, 2004b; Hayashi et al., 2004). Some of these data appear to be conflicting, and there are no
definitive data supporting a specific mechanism for formaldehyde effects on the nervous system
at this time. Neuropathological data are also limited, although data from one laboratory
(Sarsilmaz et al., 2007; Asian et al., 2006) suggest a concern for changes in brain structure in
neonatal rats following exposure at 6 or 12 ppm. No human data are available that address these
endpoints. However, a prospective cohort study of nearly one million people followed for 15
years reported strongly significant dose-response associations between death from ALS and
exposure to formaldehyde of a known duration, with longer exposures associated with greater
risk (Weisskopf et al., 2009). This large, well-designed prospective cohort study strongly
supports the causal association of neuropathological effects in humans following long-term
formaldehyde exposure.
4.4.8.4. Summary
Overall, there is strong evidence that formaldehyde exposure via inhalation may cause
adverse effects on nervous system function in experimental animals at relatively low levels of
exposure (LOAELs as low as 100 ppb). Although human data regarding neurotoxicity following
formaldehyde inhalation are limited, available data provide support that the types of effects seen
in humans are similar to those found in animal studies. Evidence from available human
controlled inhalation exposure studies indicates that humans may be affected at doses similar to
those used in animal studies; however, the human data are extremely limited.
There are insufficient data to evaluate the potential for neurotoxicity following oral
exposure to formaldehyde. Limited evaluations of brain weight or histopathology in available
chronic or subchronic oral studies found no evidence of formaldehyde-induced changes (Til et
al., 1989, 1988; Tobe et al., 1989; Johannsen et al., 1986). However, reliable studies examining
nervous system function or focused studies of neuropathology following oral exposure to
formaldehyde are not available.
This document is a draft for review purposes only and does not constitute Agency policy.
4-392 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.4.8.5. Data Gaps
Major data gaps were found regarding the evaluation of changes in nervous system
structure or function following formaldehyde exposure by both the inhalation or oral routes.
With respect to inhalation exposure, none of the available human studies resulted in data
sufficient to conduct a reliable dose-response assessment for changes in nervous system function.
Most of the available animal inhalation studies used short exposure durations (acute or short-
term), precluding a reliable evaluation of neurotoxicity following chronic exposure. Available
data for neurodevelopmental exposures are also quite limited, consisting of evaluation of
neuropathology in only one brain region and functional evaluations focused only on changes in
motor activity.
Major data gaps also exist regarding neurotoxicity following oral exposure, with no
relevant human data and extremely limited animal data. Available oral exposure studies were
insufficient to permit a reliable evaluation of the potential for neurotoxicity following oral
exposure to formaldehyde.
4.4.9. Reproductive and Developmental Toxicity
A number of studies have been identified that indicate an effect of formaldehyde
exposure on reproductive and developmental outcomes. Human data are described in Section
4.1.1.7, and animal studies are addressed in Section 4.2.1.7 of this document.
4.4.9.1. Spontaneous Abortion and Fetal Death
Several epidemiologic studies reported increases in risk of spontaneous abortion
following maternal occupational formaldehyde exposure (Taskinen et al., 1999, 1994; John et al.,
1994; Seitz and Baron, 1990; Axelsson et al., 1984). While these finding have been questioned
(Collins et al., 2001b), upon careful examination, none of the principal biases in epidemiologic
studies, including information bias, selection bias, and confounding, appear to be more likely
causes of these reported findings than the conclusion that they may reflect an underlying causal
process. While each of these occupational studies focused on women who were co-exposed to
formaldehyde and other chemicals, the occupational groups were quite different and had
different sets of co-exposures. The woodworkers in the Taskinen et al. (1999) study were
potentially co-exposed to organic solvents related to painting and lacquering, dusts, and phenols,
none of which was shown to be an independent predictor of adverse risk. The cosmetologists
studied by John et al. (1994) were co-exposed to hair dyes, bleach, alcohol-based disinfectants,
and chemicals specific to services, such as fingernail sculpturing, but, in analyses that were
specifically adjusted for other work exposures and their potentially confounding effects, the
This document is a draft for review purposes only and does not constitute Agency policy.
4-393 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
investigators reported an increased risk for the use of formaldehyde-based disinfectants. The
laboratory workers studied by Axelsson et al. (1984) were potentially co-exposed to a wide range
of solvents, but the miscarriage rate was highest among those exposed to formaldehyde. For a
potential confounder to entirely explain an observed effect of another exposure, it must be more
strongly associated with the adverse outcome. It does not appear that the collective results of
formaldehyde exposures associated with increased risk of spontaneous abortion—often in spite
of exposures being crudely measured—can be explained by information bias or confounding.
Taken together, these findings are consistent with an adverse effect of formaldehyde
exposure on the risk of pregnancy loss. The single study with the strongest quantitative
assessment of that risk is Taskinen et al. (1999), and the results presented are of sufficient quality
to support quantitative risk assessment by using the LOAEL/NOAEL approach.
This study was a well-designed population-based case-control study that appears to have
been well executed and appropriately analyzed. The study population of Finnish women was
well defined and adequately selected to allow for meaningful comparisons of health effects
between individuals with different levels of exposure to formaldehyde. The participation rate
was 64%, which is low enough to raise a concern about the potential for selection bias.
However, the authors noted that selection bias has not influenced the results of other
reproductive epidemiology studies reporting results on smoking, irregular menstruation, and
earlier miscarriages, which are known to lengthen the time to pregnancy (Bolumar et al., 1996;
Sallmen et al., 1995; Baird and Wilcox, 1985). Furthermore, there is no evidence to support
conjecture that an individual's decision to participate in this study would be differential with
respect to their workplace formaldehyde exposures while being nondifferential with respect to
the other exposures of interest, including organic solvents, wood dust, and phenols. Since the
women who chose to participate in this study were not likely to be aware of the specific
hypotheses under investigation nor could they have known the formaldehyde exposures that were
independently estimated by an industrial hygienist, selection bias is not a likely explanation for
the findings of adversity.
Data on pregnancy history, including spontaneous abortions, were collected by
questionnaire. Spontaneous abortion is the most common adverse outcome of pregnancy (Klein
et al., 1989), and retrospective self-report of spontaneous abortion has been found to match well
with prospectively collected reproductive histories (Wilcox and Horney, 1984). Many
spontaneous abortions, however, are missed with self-reporting, with the magnitude likely
exceeding 25%, but only rarely do women self-report false positive events (Wilcox and Horney,
1984). The effect of such an undercount is to cause a bias towards the null when the likelihood
of undercounting is unrelated to formaldehyde exposure. The implication is that the observed
This document is a draft for review purposes only and does not constitute Agency policy.
4-394 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
association of increased risk of spontaneous abortion associated with occupational exposure to
formaldehyde may be an underestimation of the true risk.
The findings by Taskinen et al. (1999) of reduced fertility and increased risk of
spontaneous abortion are internally consistent and coherent with other reports of increased risk
of pregnancy loss associated with exposure to formaldehyde (John et al., 1994; Taskinen et al.,
1994; Seitz and Baron, 1990; Axelsson et al., 1984). Absent evidence of alternative explanation
for these findings, it is concluded that exposure to formaldehyde is associated with pregnancy
loss and diminished fertility.
In animal studies, Sheveleva (1971) noted an increase in preimplantation loss in rats
exposed to 0.04 and 0.4 ppm formaldehyde by inhalation on GDs 1-19, and Kitaev et al. (1984)
observed evidence of degeneration in harvested embryos on GD 2, following 4 months of
maternal inhalation exposure to 0.41 ppm formaldehyde in rats. In a second series of tests
reported in Kitaev et al. (1984), female rats were exposed to 0.41 and 1.22 ppm formaldehyde for
4 months to test the hypothesis that the embryo degeneration could have been the result of
disrupted reproductive hormone levels in the dams. Ovarian weight and blood levels of LH were
increased at 0.41 ppm (but not at 1.22 ppm), and significantly increased levels of FSH were
observed at 1.22 ppm. Kitaev et al. (1984) proposed that effects at the 0.41 ppm might be related
to disruption of the hypothalamic-pituitary axis and that at the higher exposure level (1.22 ppm)
frank toxic effects to the embryo were observed. The increased FSH levels at 1.22 ppm may also
be indicative of hormonal perturbations induced by formaldehyde exposure that could affect
pregnancy maintenance in humans. The finding of treatment-related increased preimplantation
loss in rats appears to support the evidence of spontaneous abortion in the epidemiologic data. In
addition, a dominant lethal study in rats by Odeigah (1997) identified significant
postimplantation loss following pre-mating I.P. formaldehyde exposures to males, suggesting a
potential MOA involving germ cell toxicity. Nevertheless, a number of developmental toxicity
studies in rats did not report treatment-related embryolethality following gestation exposures to
formaldehyde. These included inhalation studies by Martin (1990), Saillenfait et al. (1989), and
Kilburn and Moro (1985), a series of studies by Gofmekler and Bonashevskaya (1969),
Gofmekler (1968), and Pushkina et al. (1968), and studies by Senichenkova and Chebotar (1996)
and Senichenkova (1991). It is noted, however, that, to the extent that these studies evaluated
embryonic or fetal death, the observations were conducted late in gestation and the studies may
not have been designed to detect the preimplantation losses as observed in Kitaev et al. (1984)
and Sheveleva (1971). Additionally, a number of the reports for these studies did not include
sufficient details to engender a high degree of confidence in the reported results. Fetal death was
This document is a draft for review purposes only and does not constitute Agency policy.
4-395 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
also not observed in oral studies with formaldehyde in beagle dogs (Hurni and Ohder, 1973) and
rats (Seidenberg and Becker, 1987).
4.4.9.2. Congenital Malformations
The effect of occupational exposures to formaldehyde on the incidence of congenital
malformations was examined by Dulskiene and Grazuleviciene (2005), Taskinen et al. (1994),
and Hemminki et al. (1985). Results were mixed.
In animal studies, the most frequently observed structural anomaly noted following
inhalation exposures to formaldehyde during gestation was a delay in fetal (i.e., 1st stage) testes
descent (Senichenkova and Chebotar, 1996; Senichenkova, 1991; Kilburn and Moro, 1985),
although similar findings were not reported by Saillenfait et al. (1989) or Martin (1990) in what
appear to be well-conducted prenatal developmental toxicity studies. No study in the available
database specifically examined the 2nd stage of postnatal testes descent in pups. Thus, there is no
evidence to determine if the observed effect represented a developmental delay or if it was
related to disruptions in male reproductive tract ontogeny, which is dependent on normal levels
of fetal testicular testosterone and on the expression of insulin-like hormone-3 (insl3) in fetal
Leydig cells (Klonisch et al., 2004). Senichenkova (1991) observed an increased incidence of
other organ anomalies following formaldehyde exposure during gestation; however, the
anomalies are not characterized in the report. Alterations on fetal organ weights and/or size were
noted in several studies (Kilburn and Moro, 1985; Gofmekler, 1968), but it is difficult to
ascertain if these findings represented agenesis, hypoplasia, or evidence of systemic organ
toxicity. Histopathologic evaluation of pup tissues following maternal gestational exposures to
0.01 and 0.81 ppm formaldehyde was conducted by Gofmekler and Bonashevskaya (1969),
revealing reduced glycogen content in the myocardium, the presence of iron in hepatic Kupffer
cells, and a positive Schiff reaction in the basement membrane (indicating functional alterations
in the renal tubule) at both exposure levels.
4.4.9.3. Low Birth Weight and Growth Retardation
A population-based study (Grazuleviciene et al., 1998) reported an association between
atmospheric formaldehyde exposure and low birth weight, with an adjusted OR of 1.37 (95% CI:
0.90-2.09).
A number of inhalation studies in rats identified reduced fetal weight as an adverse
outcome of in utero formaldehyde exposure and are supportive of the association noted in
humans. In a study that exposed pregnant rats to formaldehyde during GDs 6-20, Saillenfait et
al. (1989) observed significantly decreased male and female fetal rat weights (78 and 81% of
This document is a draft for review purposes only and does not constitute Agency policy.
4-396 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
control values, respectively) at 40 ppm formaldehyde. In a study that exposed the dams from
GDs 6-15, Martin (1990) found decreased fetal weights at exposure levels of 5 and 10 ppm. In
both studies, observations of reduced or delayed skeletal ossification (i.e., the thoracic vertebrae
in Saillenfait et al. [1989] and the pubes and ischia bones in Martin [1990]) were consistent with
the fetal weight deficits. Kilburn and Moro (1985) also reported fetal body weight decreases in
rats at an inhalation exposure level of 30 ppm. Conversely, increased fetal body weight as
compared with controls (generally considered to be non-adverse) was noted by Gofmekler
(1968) and Pushkina et al. (1968) at maternal formaldehyde exposure levels of 0.1 and 0.81 ppm
administered for approximately 2-3 weeks prior to mating and then throughout gestation.
Increased fetal weight was also noted in rats by Senichenkova (1991) and Senichenkova and
Chebotar (1996) following maternal exposures to 0.41 ppm formaldehyde throughout gestation.
Studies that assessed the effects of oral administration of formaldehyde on development
are quite limited. The only oral study identified that found a treatment-related effect on offspring
growth was a study using beagle dogs (Hurni and Ohder, 1973). In this study, formaldehyde was
administered at doses of 3.1 or 9.4 mg/kg-day in the feed during gestation, and pup weight
decrements at postnatal week 8 were 6.3 and 12% in the low- and high-dose groups, respectively.
4.4.9.4. Functional Developmental Outcomes (Developmental Neurotoxicity)
Indications of effects on the developing nervous system were observed in several rodent
studies, although no similar epidemiologic findings in children were identified. These studies
(Sarsilmaz et al., 2007; Asian et al., 2006; Weiler and Apfelbach, 1992; Senichenkova, 1991;
Sheveleva, 1971) are described in detail in Section 4.2.1.6. In the studies by Asian et al. (2006)
and Sarsilmaz et al. (2007), neonatal rats were exposed to formaldehyde for 30 days at 6,000 or
12,000 ppb. Decreases in the volume of discrete areas of the brain were observed at both
exposure levels in both studies, and, additionally, decreased cell numbers were noted in a region
of the hippocampus in the Sarsilmaz et al. (2007) study. Weiler and Apfelbach (1992) exposed
juvenile rats to 0.25 ppm formaldehyde for 130 days or adult rats to 0.5 ppm formaldehyde for
90 days. Olfactory thresholds measured in this study were significantly higher in the rats that
had been exposed as juveniles than in those that had been exposed only as adults. Sheveleva
(1971) observed alterations in spontaneous mobility in 1- and 2-month-old pups from dams that
had been exposed to formaldehyde at 0.04 or 0.4 ppm throughout gestation. In the Senichenkova
(1991) study, maternal rats were exposed to 400 ppb formaldehyde during GDs 1-19, and
functional observational testing was conducted on the juvenile offspring. Changes in open-field
motor activity, exploratory activity, and habituation were observed in the offspring.
This document is a draft for review purposes only and does not constitute Agency policy.
4-397 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.4.9.5. Male Reproductive Toxicity
A number of laboratory animal studies have reported effects of formaldehyde exposure
on male reproductive system endpoints. These effects include decreased testes weight, changes
in Leydig cell quantity and quality, degeneration of seminiferous tubules, decreased testosterone
levels, alterations in biomarkers of toxicity in the testes, and alterations in sperm measures
(Galilapour et al., 2007; Xing et al., 2007; Zhou et al., 2006; Ozen et al., 2005, 2002; Sarsilmaz
et al., 1999; Odeigah, 1997; Majumder and Kumar, 1995; Chowdhury et al., 1992; Til et al.,
1989, 1988; Tobe et al., 1989; Johanssen et al., 1986; Maronpot et al., 1986; Cassidy et al., 1983;
Appelman et al., 1982; Guseva, 1972). Following concurrent exposures to formaldehyde in air
and drinking water for 6 months, Guseva (1972) found decreases in testicular nucleic acid levels.
In a study conducted by Chowdhury et al. (1992), rats were administered I.P. injections of
formaldehyde for 30 days, and evidence of decreased testes weight, serum testosterone levels,
and Leydig cell quality was observed. Sarsilmaz et al. (1999) followed up on these findings
(exposing male rats to formaldehyde via inhalation at 10 and 20 ppm for 4 weeks) and found
decreases in Leydig cell quantity and the percentage of cells with normal nuclei. Hypothesizing
that the reported decreases in Leydig cell quality may have been the result of oxidative stress and
damage, Ozen et al. (2002) evaluated biomarkers of such changes and found that testicular zinc
and copper levels were decreased and iron levels were increased following exposures of adult
male rats to 10 and 20 ppm formaldehyde for 4 or 13 weeks. Additionally, relative testes weight
was decreased in a dose- and duration-dependent manner, although this effect had not been
observed by Sarsilmaz et al. (1999). Ozen et al. (2005) noted decreased serum testosterone
levels, decreased seminiferous tubule diameters, and increased levels of heat shock protein in
spermatogonia, spermatocytes, and spermatids of rats following 91 days of exposure to 10 ppm
formaldehyde. A study by Golalipour et al. (2007) observed decreased numbers of testicular
germ cells, altered sprmatogenesis, and reduced seminiferous tubular diameter and epithelial
height in rats following 18 weeks of formaldehyde inhalation exposure; the severity of the
seminiferous tubule pathology was positively correlated to the number of hours/week of
exposure. Zhou et al. (2006) found decreased testis weight, atrophy of seminiferous tubules,
edematous interstitial tissue, and alteration of epididymal sperm count, morphology, and motility
in rats after 2 weeks of formaldehyde exposure at 8 ppm. Abnormal sperm were also observed in
mice by Xing et al. (2007) after 13-weeks of inhalation exposure at 16.9 ppm, and Cassidy et al.
(1983) reported sperm abnormalities in rats following a single oral dose of 200 mg/kg.
Additionally, Majumder and Kumar (1995) observed significantly reduced sperm count, motility,
and viability following 30 days of I.P. injection of 10 mg/kg-day formaldehyde to male rats.
Also in this study, the ability of formaldehyde to affect sperm parameters was confirmed with in
This document is a draft for review purposes only and does not constitute Agency policy.
4-398 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
vitro testing. A study conducted by Odeigah (1997) demonstrated epididymal sperm count and
morphology abnormalities following five I.P. doses of >0.125 mg/kg formaldehyde and
additionally identified dominant lethal effects (decreased live embryos and increased dead
implants) following mating of treated male rats with untreated females.
Although Til et al. (1989) reported low incidences of Ley dig cell tumors in
formaldehyde-treated rats in a chronic drinking water study, no alterations in testes weight or
histopathologic lesions of the testes were observed in subchronic inhalation studies conducted by
Appleman et al. (1982) or Maronpot et al. (1986) or in subchronic or chronic oral studies by
Johanssen et al. (1986), Til et al. (1988), or Tobe et al. (1989).
No epidemiologic studies have identified an association between formaldehyde exposure
and alterations in the male reproductive system (e.g., see Ward et al. [1984]).
4.4.9.6. Female Reproductive Toxicity
The available database for the assessment of the effects of formaldehyde exposure on the
female reproductive system was limited. In addition to the findings of spontaneous abortions, as
described above, Taskinen et al. (1999) examined fecundability in a cohort of healthy female
wood-processing industry workers and found that conception was significantly delayed in
women who were occupationally exposed to formaldehyde. The FDR, a ratio of average
incidence densities of pregnancies for exposed female employees compared with unexposed
female employees, was lower in women exposed to mean formaldehyde levels of approximately
0.33 ppm (range: 0.15-1.00 ppm) compared with controls (adjusted FDR = 0.64 [95% CI:
0.28-0.92]). An FDR <1.0 is indicative of delayed conception, which is an indicator of reduced
fertility. The subfertility observed in this study is supportive of the association observed
between formaldehyde exposure and spontaneous abortion, since subclinical pregnancy losses
are increased in women with compromised fertility (Gray and Wu, 2000; Hakim et al., 1995),
and both spontaneous abortion and subfertility can be related to exposure to environmental
toxicants (Correa et al., 1996).
As described above, formaldehyde exposures to female rats resulted in decreased ovarian
weight and altered LH and FSH levels (Kitaev et al., 1984). Maronpot et al. (1986) reported
endometrial hypoplasia and lack of ovarian luteal tissue in female mice exposed to 40 ppm
formaldehyde for 13 weeks. Additionally, it is noted that, in developmental toxicity studies that
included repeated exposures of dams before mating and/or during gestation, reports of adverse
pregnancy outcomes were few. Gofmekler (1968) reported an increase in pregnancy duration
and decrease in litter size; however, this finding was not observed in other studies.
This document is a draft for review purposes only and does not constitute Agency policy.
4-399 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
With the exception of spontaneous abortion and increased time to pregnancy, associations
of formaldehyde exposure with adverse female reproductive system outcomes were not observed
in the available epidemiologic data.
4.4.9.7. Mode of Action
A strong case cannot be made for any one MOA that explains one or more of the
reproductive and developmental outcomes observed in formaldehyde epidemiologic or
toxicology studies. This is due to a number of issues, including the following:
(1) inconsistencies in study findings for the toxicology studies, which may be explained by study
quality issues (see detailed descriptions of studies in Sections 4.1 and 4.2); (2) few studies that
allow for comparisons because no study was performed with the same study design (e.g., stage of
exposure, dose, species, and strain); (3) few mechanistic studies available to test hypothesized
MO As; and (4) a bias that is pervasive in the formaldehyde literature that outcomes observed
beyond the POE (the nose) are not expected from inhalation exposure, which is the route of
exposure for most of the developmental and reproductive studies. This discussion presents
putative MO As that have been hypothesized by study authors and the studies that support the
hypothesized MO As. The four hypothesized MO As are not mutually exclusive. They could be
acting alone for certain endpoints (in which case the others are not operable) or in concert for
certain endpoints.
The focus of this discussion is on analyzing possible MO As for the developmental and
reproductive outcomes that were noted most consistently, across toxicology studies and, in some
cases, across human and animal studies. These outcomes include developmental delays, fetal
loss, and sperm quality and quantity effects.
An endocrine-disrupting MOA is supported by some of the reproductive and
developmental epidemiology and toxicology studies. For example, the decreases in fetal body
weight (Martin, 1990; Saillenfait et al., 1989), delayed ossifications (Senichenkova and
Chebotar, 1996; Senichenkova, 1991; Martin, 1990; Saillenfait et al., 1989), and delayed
eruption of incisors (Senichenkova, 1991) noted in rats after gestational exposure to
formaldehyde are consistent with developmental delays. In turn, developmental delays can result
from effects on the hypothalamic-gonadal-pituitary axis in the dam that cause hormonal level
changes in the pup; however, hormone levels in pups were not measured. Kilburn and Moro
(1985) also observed organ size changes and undescended testes after developmental
formaldehyde exposure. These outcomes can also be explained by an endocrine MOA. There
are three studies that directly tested for changes in hormones after formaldehyde exposure.
Kitaev et al. (1984) observed ovarian weight and serum LH and FSH increases after inhaled
This document is a draft for review purposes only and does not constitute Agency policy.
4-400 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
formaldehyde in adult female rats. Chowdhury et al. (1992) assessed serum testosterone levels
in adult rats after formaldehyde IP exposure and found significant decreased testosterone and
testes weights and a decrease in 3-P,A-5-hydroxy steroid dehydrogenase in Leydig cells,
suggesting that formaldehyde affects steroidogenesis. Ozen et al. (2002) also reported
significant serum testosterone level decreases as well as decreased mean seminiferous tubule
diameters. Furthermore, the steroidogenesis MOA leading to reduced testosterone is also
consistent with the sperm quality and quantity decrements observed in the studies by Ozen et al.
(2002), Sarsilmaz et al. (1999), and Odeigah (1997) studies.
In human studies, delayed time to pregnancy and increased incidence of spontaneous
abortion (Taskinen et al., 1999), consistent with some study findings from the toxicology
literature, could also be explained by an endocrine MOA. Alterations in hormone levels could
lead to pregnancy maintenance problems. Extrapolating the Chowdhury et al. (1992) results of
the steroidogenesis MOA to females, formaldehyde exposure could affect progesterone levels
required for pregnancy. However, progesterone levels were unchanged in the female rat in the
one study that assessed progesterone (Kitaev et al., 1984). Consistent with an endocrine
mediated MOA, Maronpot et al. (1986) observed endometrial hypoplasia and lack of ovarian
luteal tissue in females exposed to formaldehyde.
A second hypothesized MOA for some of the developmental and reproductive outcomes
is genotoxicity of the gametes. Such an MOA could explain pregnancy loss in humans
(Taskinen, et al., 1999) and preimplantation loss in animal studies (Xing et al., 2007; Kitaev et
al., 1984; Sheveleva, 1971) and fetal viability (e.g., litter size decreases) after formaldehyde
exposure. Consistent with male gamete genotoxicity, Odeigah (1997) and Xing et al. (2007)
observed reduced fertile matings and live embryos, and increased dead implants in a dominant
lethal study.
Oxidative stress/damage is another MOA that is consistent with testicular toxicity, sperm
effects, and reduced embryo viability. Ozen et al. (2002) investigated the mechanism of
oxidative stress being responsible for the testes quality effects by assessing testicular iron,
copper, and zinc levels. Zinc and copper levels were reduced in the rat testes, consistent with an
oxidative stress MOA. Ozen et al. (2002) also reported increased iron levels and decreased zinc
levels in the lung, consistent with oxidative stress. Another study (Zhou et al., 2006) that
investigated the oxidative stress MOA in the testes observed significant changes in oxidative
stress biochemical markers (decreases in SOD, GPX, GSH, and an increase in MDA levels).
The authors also assessed the protective effect of coadministration with vitamin E, an
antioxidant, on decreased testes weight, biochemical alterations, histopathologic effects, and on
sperm count, motility, and morphology. The study of Pushkina et al. (1968) found reduced
This document is a draft for review purposes only and does not constitute Agency policy.
4-401 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
levels of Vitamin C, another antioxidant, in the fetus and maternal liver after formaldehyde
exposure.
The MO As proposed are not mutually exclusive and in fact could interact with one
another. For example, an endocrine MO A could lead to oxidative stress, and that oxidative stress
could lead to genotoxicity.
4.4.9.8. Data Gaps
The inhalation developmental toxicity studies conducted on formaldehyde and described
in Section 4.2.1.7 comprise an adequate assessment of prenatal developmental toxicity for
application to inhalation reference concentration determination. The assessments of postnatal
developmental toxicity and of reproductive function following inhalation of formaldehyde are
less complete. It is notable that, although the database contains some studies that assess various
aspects of reproductive organ system toxicity, particularly in males, there is no assessment of
multigenerational reproductive function, such as would be evaluated in a two-generation
reproductive toxicity study, nor is there an assessment of potential reproductive effects of
formaldehyde exposure in human males.
Adequate assessments of developmental and reproductive toxicity following oral
exposures to formaldehyde have not been conducted.
4.5. SYNTHESIS AND EVALUATION OF CARCINOGENICITY
4.5.1. Cancers of the Respiratory Tract
Epidemiologic studies of formaldehyde-exposed workers provide sufficient evidence of a
causal association between formaldehyde exposure and upper respiratory tract (URT) cancers
(e.g., nasopharyngeal cancer (NPC; Section 4.1.2.1.1), nasal and paranasal cancers (Section
4.1.2.1.2), and other upper respiratory tract cancers (Section 4.1.2.1.3)). In addition, the
observational evidence from epidemiologic studies reporting associations between formaldehyde
exposure and increased risk of NPC supports a conclusion of a causal association for this specific
cancer. However, epidemiologic studies of rare outcomes such as NPC, which has an expected
incidence of 1 per 100,000 people per year in the United States, do not typically have great
statistical power to rule out the null hypothesis (i.e. no association). However, the weight of
evidence of the several studies reviewed in Section 4.1.2.1.1 provide an accumulation of
consistent observational evidence sufficient to exclude chance as an explanation for the
association. Additionally, there is insufficient evidence of consistent confounding or other bias
across the studies that were considered; thus, confounding and bias were also ruled out as
explanations for the observed association. The lack of a convincing and consistent alternative
This document is a draft for review purposes only and does not constitute Agency policy.
4-402 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
hypothesis of causation - in spite of repeated examinations - further supports the conclusion that
the association between formaldehyde exposure and increased risk of NPC is causal.
The single strongest cohort study, Hauptmann et al. (2004), shows a statistically
significant exposure-response relationship between formaldehyde exposure and URT cancers.
Hauptmann et al. (2004) demonstrated significant excess risk of NPC in exposed workers based
on U.S. population death rates (standardized mortality ratio [SMR] = 2.1; 95% confidence index
[CI] 1.05-4.21) in a large cohort of 25,619 industrial workers. In addition to the SMR based on
an external comparison population, relative risks (RRs) were presented based on internal
comparisons of workers in order to minimize potential selection bias due to the well known
healthy worker effect. Statistically significant exposure-response relationships between
increased formaldehyde exposure and increased risk of NPC were reported for two different
metrics of exposure (peak and cumulative exposure). Relative risks for NPC were also elevated
for increased duration of exposure to formaldehyde and for the average intensity of exposure.
These analyses controlled for potential confounders including calendar year, age, sex, race, and
pay category. While exposure measurement error is likely to be present in any epidemiologic
study, there was no evidence of any differential measurement error that could have produced the
observation of a spurious association. Any non-differential measurement error would likely have
attenuated the effect of formaldehyde was smaller than that which would otherwise have been
observed in the absence of measurement error.
The case-control studies similarly also report associations between formaldehyde
exposure and cancer mortality for NPC. Although other risk factors for NPC (e.g., Epstein-Barr
Virus) and the predominant NPC histological sub-type (SCC versus undifferentiated) vary
significantly across the world, case-control studies consistently provide evidence of an
association between occupational exposure to formaldehyde and NPC (Vaughn et al., 1986a;
Vaughn et al., 2000; Roush et al., 1987; Hildesheim et al., 2001; West et al., 1993). In their
more recent study, Vaughn et al. (2000) used worker histories to estimate each individual
worker's formaldehyde exposure. Workers with more than 1.10 ppm-years of cumulative
exposure were found to be at significantly higher risk of NPC, with an odds ratio (OR) of 3.0
(95% CI 1.3-6.6) (Vaughn et al., 2000). Two different exposure metrics, duration of exposure
and cumulative exposure, were positively associated with increased risk of NPC, with a
significant test for trend (p = 0.014 and 0.033, respectively). The OR also increased in
magnitude as the probability of "Ever" having occupational exposure increased, from OR =1.6
among those whose exposure was judged to be "Possible, probable or definite" to OR =13.3
among those with "Definite" exposure (p-trend <0.001).
This document is a draft for review purposes only and does not constitute Agency policy.
4-403 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
NPC histological sub-type analysis indicates that these associations held for both SCC
and epithelial NPC, but not for the undifferentiating and nonkeritinizing NPC (Vaughn et al.,
2000). However, formaldehyde exposure is also associated with risk of NPC in Taipei, Taiwan,
where greater than 90% of the cases had nonkeritinizing and undifferentiated carcinomas and
less than 10% of the cases were diagnosed as having SCCs (Hildesheim et al., 2001). These
reported associations were strengthened by considering higher probability of exposure (RR =
2.6; 95% CI 1.1-6.3), greater intensity of exposure (RR = 2.1; 95% CI 1-4.2) and EBV
seropositive cases (RR = 2.7; 95% CI 1.2-5.9) (Hildesheim et al., 2001). Case-control studies
have also linked residential exposure to NPC, specifically for years of residence in mobile homes
(Vaughn et al., 1986b) and the use of mosquito coils in the Philippines (West et al., 1993).
Independent testing of 6 brands of East Asian mosquito coils evaluated the emission rates of
carbonyl compounds in the mosquito smoke and reported that formaldehyde and acetaldehyde
had the highest emission rates. Among the three experiments on each of the six brands, the
range of formaldehyde concentrations was from 0.87 |ig/m3 (1 ppb) to 25 |ig/m3 (31 ppb).
As a group, other URT sites of direct contact with formaldehyde upon inhalation (i.e.,
salivary gland, mouth, nasal cavity and larynx) also showed evidence of a trend in increasing
relative risks with increasing average intensity and peak exposure in the Hauptmann et al. (2004)
cohort study, although these trends did not reach the level of statistical significance. The results
from other cohort studies and case-control studies are mixed (between positive associations and
null findings) for associations between formaldehyde exposure and specific cancers of the URT
(IARC, 2006). For rare cancers, extremely large cohorts would be needed to have the statistical
power to detect an association for tumors defined by individual sites (e.g., mouth, salivary gland,
hypopharnyx). Results vary in the smaller cohort studies, where a single case may result in an
elevated risk but taken together the evidence is considered suggestive (Section 4.1.2). Case-
control studies have been useful to better understand potential associations between
formaldehyde exposure and rare cancers of the URT. Luce et al. (2002) evaluated pooled data
from 12 case-control studies and demonstrated a statistically significant increased risk between
formaldehyde exposure and sinonasal cancer. A case-control study by Gustavsson et al. (1998)
suggested an association between formaldehyde exposure and oral squamous cell carcinoma
(SCC), esophageal, and laryngeal cancers, with odds ratios (ORs) of 1.28, 1.90, and 1.45,
respectively. However, the individual ORs were not statistically significant. Hypopharyngeal
cancer was linked with formaldehyde exposure with an OR of 3.78 (95% CI 1.50-9.49) in
another case-control study (Laforest et al., 2000). While the data on site-specific cancers of the
URT is somewhat sparse, they are consistent with a carcinogenic hypothesis and in their large
cohort study, Hauptmann and colleagues (2004) concluded that in spite of the small numbers of
This document is a draft for review purposes only and does not constitute Agency policy.
4-404 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
deaths from cancers of the URT, the positive associations with average intensity and peak
exposure were consistent with the carcinogenicity of formaldehyde at these sites of first contact.
Supporting animal evidence
Inhalation exposure:
Animal studies, primarily rodent bioassays, strongly support the causal relationship
between of formaldehyde exposure and URT carcinogenicity. Formaldehyde-induced cancers
are primarily seen in the nasal passages (Kerns et al., 1983; Monticello et al., 1996; Tobe at al.,
1985; Kamata et al, 1997; Sellakumar et al., 1985), but it should be noted that rodents, unlike
humans, are obligate nose breathers and have convoluted nasal turbinates. Chronic animal
studies (inhalation and oral exposures) report tumor incidence in a variety of rodent models.
Study descriptions are provided above in detail (Section 4.2.1, Table 4-34). The study results are
evaluated here for both routes of exposure in the context of how they inform the carcinogenic
potential at the portal of entry, specifically the URT.
In rodent studies of the respiratory tract, only nasal tumors are considered to be induced
by formaldehyde. Repeated exposures to 10-15 ppm formaldehyde result in gross nasal lesions
and high incidence of nasal tumors (See Table 4-38, Section 4.2.1). Although increased cell
proliferation, squamous metaplasia, dysplasia and focal necrotic lesions have been noted in the
larynx and trachea in some studies, no tumors in these locations have been reported in the rodent
studies. The majority of studies were conducted using rats (F344, Wistar, or Sprague-Dawley),
and all studies of 18 months or greater in mice and rats show evidence of formaldehyde-induced
nasal carcinogenicity. The nasal tumors are primarily SCCs, although papillomas, polypoid
adenomas, adenocarcinomas, fibrosarcomas, and esthesioneuroepitheliomas have been reported
(Kamata et al., 1997; Monticello et al., 1996; Morgan et al., 1986a, b; Takahashi et al., 1986;
Sellakumar et al., 1985; Kerns et al., 1983; Albert et al., 1982). Although hyperplasia, dysplasia,
and squamous metaplasia of the respiratory epithelium have been observed beyond the nasal
cavity, other respiratory tract tumors have not been reported to be significantly increased by
formaldehyde exposure alone.
Increased tumor incidence and decreased latency are correlated with increasing
formaldehyde exposure concentration. Reviewing data from the only lifelong inhalation study
(i.e., until "natural death") with multiple exposure groups, nasal SCCs occurred much earlier in
the high-exposure animals. For example, tumors are first noted at 8 and 9 months following
exposure for high-exposed (15 ppm) female and male F344 rats versus tumors not arising until
24 months in low-exposed rats (2 ppm) (Kerns et al., 1983). In a follow-up study by Monticello
et al. (1996), the incidence of SCC in rats exposed at 15 ppm was 47%, with the first tumor noted
This document is a draft for review purposes only and does not constitute Agency policy.
4-405 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
at 12 months. The incidence of SCC in male rats exposed at 10 ppm was 22%, with the first
tumors observed at 18 months after exposure. Moreover, of the 90 rats exposed at 6 ppm for 20
months, only one SCC was noted. No SCCs were detected in rats exposed to 0.7 or 2 ppm
formaldehyde. These incidence rates are not mortality-adjusted (see Chapter 5, section 5.3.4)
and include animals from each scheduled sacrifice (3, 6, 12, and 18 months). In a lifelong study
of male Sprague-Dawley rats exposed to 10 ppm formaldehyde, the cumulative nasal tumor
incidence was calculated as a function of time of exposure (Sellakumar et al., 1985). After 2
years of exposure, the adjusted probability of nasal carcinoma was greater than 60%.
There is some evidence that less-than-lifetime exposure to formaldehyde can induce nasal
tumors over an extended observation period. Two studies, both in male Wistar rats, report nasal
tumors in response to less-than-lifetime exposures (Woutersen et al., 1989; Feron et al., 1988).
A 13-week exposure at 20 ppm followed by an observation period of 30 months (inclusive of
exposure) in Wistar rats resulted in six nasal tumors including three nasal SCCs, one cystic SCC
of the nasolacrimal duct, one carcinoma in situ and an ameloblastoma, while no tumors were
noted in the corresponding air-exposed controls (Feron et al., 1988). A limited number of
formaldehyde-related tumors were noted from 4 or 8 weeks of exposure followed by 30 months
of observation. Although the tumor incidence of these less-than-lifetime exposures is low, this is
consistent with the 2-year bioassays in Wistar rats. Wistar rats are more resilient to
formaldehyde-induced nasal toxicity than F344 or SD rats (Section 4.2.1), and only 1 of 26 (4%)
Wistar rats exposed at 10 ppm for 28 months developed SCC (Woutersen et al., 1989) versus
22% in F344 rats (Monticello et al., 1996).
The specificity of formaldehyde-induced tumors in the nasal passages of rodents is
believed, at least in part, to be a function of tissue dose. Computational fluid dynamics (CFD)
modeling used to predict formaldehyde tissue flux during inhalation exposures suggests that at
comparable concentrations, tissue flux in the nasal passages of rodents is more intense than for
non-human primates and humans. Modeling predicts a different pattern of formaldehyde flux
into URT tissues of rodents compared to humans, where formaldehyde penetrates more deeply
into the respiratory tract of primates than rodents even considering nose-only breathing for
primates (See Section 3.4). Humans will generally switch to mouth breathing when sensing an
irritating smell and during physical exertion, resulting in direct exposures to the mouth and
greater tissue flux in tissues beyond the bypassed nasal passages. Therefore, species differences
in tissue dose may contribute to formaldehyde-induced tumors in humans beyond the nasal
passages, which are not evident in rodent bioassays.
This document is a draft for review purposes only and does not constitute Agency policy.
4-406 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Oral Exposure:
Consistent with the observed carcinogenic activity of formaldehyde on tissues at the
portal of entry (POE) from inhalation exposure, there is evidence to support POE effects from
oral exposures as well - further strengthening the overall weight of evidence of formaldehyde's
carcinogenicity. As with the respiratory tract, the proximal portion of the gastrointestinal (GI)
tract exhibits formaldehyde-induced lesions specifically in the forestomach and glandular
stomach (Soffritti et al., 1989; Til et al., 1989; Tobe et al., 1989; Takahashi et al., 1986).
However, data are mixed regarding the carcinogenic potential of formaldehyde in the GI tract
from oral exposures.
Two independent 2-year cancer bioassays in Wistar rats exposed to formaldehyde in
drinking water were both negative (Til et al., 1989; Tobe et al., 1989). Til et al. (1989) exposed
rats to a range of formaldehyde doses (0, 1.2, 15, or 82 mg/kg-day) and evaluated 44-49 animals
per sex per dose group at 24 months of exposure. No formaldehyde-related tumors were found.
A smaller study by Tobe et al. (1989) failed to note any tumors after a 2-year exposure at 0, 10,
50 or 300 mg/kg-day (eight rats per sex per treatment group.)
In contrast, two studies that included lifelong observation in male and female Sprague-
Dawley rats provide support for formaldehyde-induced GI tract tumors - one study where
exposure commenced at 7 weeks of age and a second study conducted with breeder rats (25
weeks of age) and their offspring (Soffritti et al., 1989). These studies demonstrate an increase
in GI tumors (although rare) correlated with exposure to formaldehyde and significantly
increased susceptibility to early-lifetime exposure. The authors provide a detailed report on the
background rates of various stomach and intestinal neoplasia for male (n = 2,677) and female (n
= 2,582) rats within the colony (Soffritti et al., 1989). From this background pool, the total
incidence of benign and malignant tumors in the stomach and intestine combined is only 1.4%
(combining all sites and locations), with the majority of tumors located in the stomach (1%
benign, 0.2% malignant). In comparison to colony-specific background rates, apparent increases
in both stomach and intestinal neoplasia are noted in formaldehyde-treated rats (ranging from 1
to 6% by type in rats exposed beginning week 7). When summed across the GI tract, tumor
incidence in the highest treatment group was 8% versus 1.4% in historical controls. Elevations
of individual tumors or a clear dose-response relationship are difficult to discern for rare cancers
where there are only 50 animals per group. Despite the limitations of group size and lack of a
dose-response relationship, the findings do support the carcinogenic potential for formaldehyde
administered orally.
The second study reported by Soffritti et al. (1989) in Sprague-Dawley rats demonstrates
early lifetime susceptibility, with GI tumor incidences of 21.6% in female (n = 37) and 13.9% in
This document is a draft for review purposes only and does not constitute Agency policy.
4-407 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
33
34
35
male (n = 36) offspring after exposure to formaldehyde beginning in utero, versus 5.6% in the
adult breeders. Rats were exposed to formaldehyde in drinking water for 2 years (0 or 2,500
mg/L). Exposures began on gestational day 12 in the offspring. The most common tumor
detected was intestinal leiomyosarcoma (13.5% in female offspring), which has a background
rate of 0.04% in female rats in the colony. The incidence of GI tumors in the adult breeders (n =
18 per sex) was due to one adenocarcinoma in a male rat and one pallipoma/acanthoma in a
female rat. Although severely limited by study size, their occurrence is consistent with the
observation of formaldehyde-induced tumors, given the low background rates for this colony.
The Soffritti et al. (1989) studies stand alone in observing formaldehyde-induced GI
tumors. These findings are largely attributed to a unique study design that included lifelong
observation (i.e., until "natural death"), neonatal exposure, examination of individual tumor types
as well as combined rare tumor types for analysis, and summation of tumors across locations.
The study design results in a more sensitive assay for rare tumors as well as tumors with a long
latency. Thus, Soffritti et al. (1989) utilized a more appropriate design and analysis for detecting
rare tumors, and these findings are not diminished by the null results of Tobe et al. (1989) and
Til et al. (1989).
There is evidence that formaldehyde may act as a tumor promoter by the oral route as
well as the inhalation route (discussed above). Takahashi et al. (1986) reported an increase in N-
Methyl-N'-Nitro-N-Nitrosoguanidine (MNNG)-initiated GI cancers in mice with formaldehyde
exposure (29.4%, versus 13.3% in controls); the greatest difference in tumor response was
associated with adenocarcinoma of the glandular stomach (23.5%, versus 3.3% in controls).
Additionally, forestomach papillomas and preneoplastic hyperplasia in the glandular stomach
were increased with formaldehyde exposure alone.
4.5.2 Lymphohematopoietic Malignancies
4.5.2.1. Background
Lymphohematopoietic (LHP) cancers include neoplasms of both lymphoid and myeloid
cell origins. Cancers of the immune system are described as leukemia if they primarily involve
cells from peripheral blood and bone marrow at diagnosis and lymphomas if they constitute a
solid tumor (Robbins, 2004). Some forms of leukemia which present as an immature immune
cell phenotype are believed to arise from lymphomyeloid stem cells or progenitor cells normally
found in the bone marrow (e.g., acute lymphoblastic leukemia (ALL) and acute myeloid
leukemia (AML)) (Greaves, 2004). However, multiple myeloma, lymphomas and some
leukemias may arise from mature functional lymphocytes present outside of the bone marrow
(Greaves, 2004; see Figure 4-33). Therefore, the use of the general term 'leukemia' is not
This document is a draft for review purposes only and does not constitute Agency policy.
4-408 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
restricted to cancers from a transformed stem cell or progenitor cell in the bone marrow but also
applies to cancers which arise from differentiated cells (e.g., mature lymphocytes).
Epidemiologic studies have reported that formaldehyde exposure is associated with both
leukemia and lymphomas (Chapter 4.1.2.2.1). Specific neoplasms reported to be associated with
formaldehyde exposure include myeloid leukemia, Hodgkin's lymphoma, and multiple
myeloma.
Figure 4-33: Developmental origins for cancers of the lymphohematopoietic
system (Adapted from Greaves (2004).
When evaluating cancers of the LHP system, epidemiologic analysis often groups many
of these cancers together. In part, this may be done to increase the statistical strength of the
analysis. Additionally, since there is a potential for disease misclassification, grouping these
diseases is preferred by some researchers, especially when analyzing older mortality data.
Historically, misclassification may have been due to factors such as poor histopathology and
diagnosis of late-stage disease, where cell line of origin may have been hard to distinguish.
Without the cell surface markers and molecular tools used today to classify cells, diagnosis was
accomplished primarily based on histology. However, as the cancers progress, the leukemic
stem cells may present as less mature cells. For example, poor health surveillance may allow
chronic myeloid leukemia (CML) to remain undiagnosed until the blast crisis, often seen late in
Developmental origins of blood cell cancers
Germ Line or
Early Embryonic
Rare Acute Leukemia
Acute Lymphoblastic Leukemia (ALL)
Acute Myeloid Leukemia (AML)
Myelodysplastic syndrome
Chronic Myeloid Leukemia (CML)
Haemopoietic
Stem cell
(lymphomyeloid)
Childhood ALL and AML
Lymphoid
Stem Cell
Myeloid
Stem Cell
Cutaneous T Lymphoma
Adult T-cell Leukemia
B Chronic Lymphocytic Leukemia
non-Hodgkin's Lymphoma
B or T Prolymphocytic Leukemia
Hairy Cell Leukemia
Myeloma
Mature
lymphocyte
Subsets
Source: M.F. Greaves, IARC Press, (2004)
This document is a draft for review purposes only and does not constitute Agency policy.
4-409 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
the disease. This blast crisis presents as a leukemia of relatively immature myeloid cells and
may have been mistaken for AML without the more recent techniques available for classifying
disease. Since cancers of the LHP system may present with cell surface markers for multiple cell
lines, classification remains problematic in some cases.
Although often grouped for analysis, the LHP cancers represent many distinct
malignancies which may arise from discrete cell types in different tissues throughout the body
(Greaves, 2004). The World Health Organization (WHO) has developed a classification system
for both lymphoid and myeloid leukemia defined by a combination of morphology,
immunophenotype, genetic features and clinical features (Harris et al., 2000a; Harris et al.,
2000b). The historical nomenclature and International Classification of Disease (ICD) codes
used in epidemiologic studies do not correspond to these new classification systems. For
example, both chronic lymphocytic leukemia (CLL) and B-cell lymphomas arise from similar
cell types in the periphery, yet epidemiologic studies have considered them independently even
though they are currently considered to be the same disease in the new WHO classification
system - which would diminish the statistical power to detect an association. Careful re-analysis
of epidemiologic data addresses evidence for the class as a whole (all LHP cancers) but also
various subclasses (e.g. myeloid versus lymphoid).
Therefore, the following analysis first examines the epidemiologic evidence for all LHP
cancers as a class, then all leukemia, to best take advantage of the majority of publications
available in assessing the weight of evidence for the carcinogenicity of formaldehyde. The
subsequent analysis by sub-type draws upon the available evidence for specific diseases or
groups of diseases (e.g., myeloid leukemia, Hodgkin's lymphoma, non-Hodgkin's lymphoma).
Although fewer data are available for the sub-type analysis, these data help clarify which cancers
may contribute to the consistent observation of an association between formaldehyde exposure
and all LHP cancers. Novel combinations of phenotypic sub-types are presented where they are
etiologically relevant. The sub-type analysis frames the subsequent mode of action (MOA)
analysis.
4.5.2.2. All LHP Malignancies
Epidemiologic studies involving formaldehyde-exposed workers provide sufficient
evidence of a causal association between formaldehyde exposure and all LHP malignancies
(Section 4.1.2.1.5). Positive associations between formaldehyde exposure and LHP cancers have
been reported for chemical workers (Wong et al., 1983; Bertazzi et al., 1986), embalmers
(Walrath and Fraumeni, 1983, Walrath and Fraumeni, 1984; Hayes et al., 1990), anatomists and
pathologists (Harrington and Shannon 1975; Hall et al., 1991; Levine at al 1984; Stroup et al.,
1986; Matanoski et al., 1989) (Table 4-90). However, clear associations (in terms of overall
This document is a draft for review purposes only and does not constitute Agency policy.
4-410 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
standardized mortality ratios [SMRs] or proportional mortality ratios [PMRs]) were not reported
in analyses for garment workers, iron-foundry workers, and a large US industrial cohort
(Pinkerton et al., 2004; Andjelkovich et al., 1995; Beane Freeman et al., 2009; Marsh et al.,
1996), although associations were observed in some of these studies when exposure-response
relationships were considered. Several published meta-analyses are available which more
formally assess the strength of association between formaldehyde exposure and mortality from
all LHP cancers. Pooled SMRs indicate stronger associations for professional workers
(embalmers, anatomists and pathologists) than industry workers (Table 4-91). Bosetti et al.
(2008) found similar relationships, with a pooled SMR of 1.31 (95% CI 1.16-1.47) for
'professionals' (i.e. embalmers, anatomists and pathologists) versus a pooled estimate of 0.85
(95% CI 0.74-0.96) for industrial workers. A recent meta-analysis by Zhang et al. (2009) reports
a summary relative risk (RR) of 1.25 (95% CI 1.09-1.43) for both professional and industry
workers for all LHP cancers (ICD 9 codes 200-209). These researchers identified 19 cohort
study analyses, including cohort study updates. Zhang et al. (2009) used the reported RR from
the highest exposure category to increase statistical power and reduce uncertainty regarding
confounding or other bias. Although study selection was controversial, e.g., the inclusion of
multiple reports from a single cohort and the use of one cohort where only a portion of the
workers were formaldehyde-exposed, this meta-analysis is generally supportive of an association
between formaldehyde and LHP malignancies.
The apparent differences by industry/profession may reflect many influences, including
exposure potential and demographic characteristics. External analysis (use of the general
population for comparison) relies on the assumption that cancer incidence rates are expected to
be similar between the general population and the study population in the absence of exposure.
The 'healthy worker effect' is well known, and there may be differences in the magnitude of this
selection bias by industry or profession. For instance, LHP cancer incidence and mortality have
many risk factors including socioeconomic status. Therefore, the consistent positive findings in
professional workers versus mixed results in industrial workers could be influenced by the
appropriateness of the comparison to the general population - that is, a differential extent of
selection bias. Interestingly, salaried workers, but not the hourly workers, in an Italian plastic
manufacturing plant had elevated SMRs for LHP cancers (1.69 (95% CI 1.07-2.53) and
0.93(95%) CI 0.62-1.35), respectively) (Dell and Teta, 1995). Without knowledge of which
worker group is most similar to the comparison population with respect to LHP cancers
mortality, one cannot discern if this potential effect of demographic variability accentuates
effects in professional/salaried workers or obscures the effects in industrial/hourly-wage
workers.
This document is a draft for review purposes only and does not constitute Agency policy.
4-411 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-90. Summary of cohort and case-control studies which evaluated
2 the incidence of all LHP cancers in formaldehyde-exposed populations
3 (ICD-8 Codes: 200-209) and all leukemias (ICD-8 Codes: 204-207). (See
4 Table 4-6 for complete study details and findings)
5
Sluch noniiliilion
S(iul\ (k'Ciils
All I.IIPCillK-OI'S
l.oiikoiniii
Reference'
SMK Aii;il\sis'
Pathologists and
technicians
(n=2,079)
Years of study 1955-1973
2.0 (p<0.01)
{pathologists}
0.5
{technicians}
0.6
{pathologists}
0.5
{technicians}
Harrington and
Shannon, 1975
Pathologists and
technicians
(n=2,720)
1974-1980
NR
0.91 (0.05-4.29)
men
9.26(0.47-43.9)
women
Harrington and
Oakes, 1984
Pathologists
(n=4,512)
Years of study 1974-1987
1.44
(0.69-2.63)
1.52 (0.41-3.89)
Halletal., 1991
Ontario Undertakers
(n= 1,477)
Mortality from 1950-1977
1.24
1.60
Levine et al.,1984
Male Anatomists
(n=2,327)
Mortality from 1925-1979
1.20
(0.7-2.0)
1.5 (0.7-2.7)
Stroup et al., 1986
Male pathologists
(n=4,485)
Mortality through 1977
NR
1.06
Logue et al, 1986
Male pathologists
(n=6,lll)
Participants from 1912-
1950 membership rolls.
Mortality followed through
1978.
1.25
(0.95-1.62)
1.35 (0.92-1.92)
Matanoski et al.,
1989
Chemical industry
workers, men
(n=14,014)
Mortality from 1941-2000
NR
0.91(0.62-1.39)
Coggon et al.,
2003
Chemical workers
(n=2,026)
1.36
(0.5-2.95)
Wong et al., 1983
Industrial workers
(n=25,619)
Mortality followed through
2004
0.94
(0.84-1.06)
1.02
(0.85-1.59)
Beane-Freemen et
al., 2009
Industrial workers
(n=7,328)
0.89
Marsh etal., 1996
{Subset of NCI
cohort reported by
Hauptmann et al.,
2003}
Garment workers
(n= 11,098)
Mortality followed through
1998
0.97
(0.74-1.26)
1.09 (0.70-1.62)
Pinkerton et al.,
2004
Resin plant workers
(n=l,330)
Employed between 1959-
1980
Mortality through 1986
2.01
NR
Bertazzi et al.,
1986
Plastic
manufacturing
1.69
(salaried workers)
1.98
(salaried workers)
Dell and Teta,
1995
This document is a draft for review purposes only and does not constitute Agency policy.
4-412 DRAFT—DO NOT CITE OR QUOTE
-------
(n=5,932)
0.93
(hourly workers)
0.98
(hourly workers)
PMR Ansiljsis"
Embalmers, New
York
(n=l,132)
Licensed between 1925-
1980
1.15
1.32
Walrath and
Fraumeni, 1983
Embalmers, CA
(n= 1,007)
Licensed between 1925-
1980
1.22
1.75 (p<0.05)
1.24
(<20 years)
2.21 (P<0.05)
(>20 years)
Walrath and
Fraumeni, 1984
Embalmers, U.S.
(n=4,046)
1.39
(1.15-1.63)
White 1.31
(1.06-1.59)
Nonwhite 2.41
(1.35-3.97)
1.52 2
(0.98-2.35)
White 1.44
(p<0.05)
Nonwhite 2.72
(p<0.05)
Hayes et al., 1990
C;ise-( onirol Studios'
American cancer
Society Cancer
Prevention Study II:
(n=362,828 men)
Results for men reporting
formaldehyde exposure,
and occupations related to
formaldehyde exposure
1.22 (0.84-1.77)
(formaldehyde
exposed)
3.44 (1.11-10.68)
{formaldehyde
exposure and
occupation}
0.96 (0.54-1.71)
(formaldehyde
exposed)
5.79 (1.44-23.25)
{formaldehyde
exposure and
occupation}
Stellman et al.,
1998
White men
diagnosed with
leukemia
(Iowa and
Minnesota)
(n=622)
Recruited in 1980-1983
NR
1.0 (0.7-1.4) Low
0.7 (0.2-2.6) High
Blair etal., 1993
1
2 Notes:
3 1. Relative risk estimate (SMR, PMR, or OR) presented with 95% confidence intervals, where available.
4 2. PMR for leukemia for the total group calculated from the published data for lymphatic leukemia (204, myeloid leukemia
5 £205)., and other/unspecified_(206, 207).
6
7
8 The only study which has data to inform the effects of either exposure level or the
9 appropriateness of an external comparison group on the association between formaldehyde
10 exposure and all LHP cancer mortality is the National Cancer Institute (NCI) cohort study of
11 industrial workers (Blair et al., 1986; Beane Freeman et al., 2009), which presents relative rates
12 based on internal comparisons for 3 different exposure metrics. Although SMR analysis with an
13 external comparison group did not indicate increased mortality from all LHP cancers (0.94, CI
14 0.84-1.06, for the exposed workers), internal analysis using the low-exposed workers as the
15 comparison group demonstrates positive exposure-response relationships for increased mortality
This document is a draft for review purposes only and does not constitute Agency policy.
4-413 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
from all LHP malignancies cancers and peak exposure across the study periods (1965-2004)
(Figure 4-34A and 4-34B) (Beane Freeman et al., 2009), with a statistically significant trend (p <
0.05) for every year since 1977. These results, indicating a positive exposure-response
relationship among plant workers, who most likely have similar demographic characteristics, are
noteworthy given the apparent lack of association when SMRs for the same cohort are calculated
against mortality rates for the general population. The lack of an apparent association with
SMRs may be attributable to the healthy worker effect and/or some other difference between the
exposed workers and the general population.
Although the association between formaldehyde exposure and all LHP cancer mortality
in industrial and professional cohorts is mixed, the strength of the internal analysis of the NCI
cohort, in the absence of positive SMRs compared to the general population, suggests that SMR
analyses may not be the most appropriate methodology for assessing LHP cancer mortality.
Given the potential for demographic differences between an industrial workforce and the general
population, the results of the internal analysis of the NCI industrial cohort provide a higher
quality analysis - and therefore should be given significantly more weight than SMR analyses of
industrial workers that could not distinguish their findings from the null. Given the consistency
and strength of the positive associations for all LHP cancers cancer mortality in professional
cohorts (embalmers, anatomists and pathologists) taken together with the strong positive results
of the NCI cohort, human epidemiologic evidence are sufficient to conclude that there is a causal
association between formaldehyde exposure and mortality from all LHP malignancies (as a
group).
4.5.2.3. All Leukemia
Like the analysis of all LHP cancers, an analysis of all leukemia combines diseases which
differ significantly in cell of origin and etiology, including acute and chronic forms of both
myeloid and lymphatic leukemia. This class also includes other leukemia (e.g., erythraemia) and
a general category of 'other and unspecified leukemia' (ICD-8 207). Regardless, there is some
utility in evaluating the all leukemia mortality data because many studies provided results for this
grouping. Also, the diagnosis of leukemia versus solid LHP tumors is fairly distinct thereby
limiting misclassification of the endpoint.
This document is a draft for review purposes only and does not constitute Agency policy.
4-414 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
vii] . ......
:S60 1932 18SO 2i*X» 198C- *&?u ^ r&£ i860 2->Ll£
r.tliH ,i ,ir i f orv? pi" Follows,ip
Figure 4-34A. Association between peak formaldehyde exposure and the risk
of lymphohematopoietic malignancy.
Relative risks for medium-peak (2.0 to <4.0 ppm) and high-peak (>4.0 ppm)
formaldehyde exposure categories compared with the low exposed category (>0
to <2.0 ppm) and P values for trend tests among the exposed person-years for
lymphohematopoietic malignancies are shown by year of end of follow-up, 1965-
2004. Values plotted at 0.1 represent RR = 0 due to no cases in the exposure
category values plotted at 20 represent RR = infinity due to no cases in the
referent category. The small graphs above the relative risk plots represent the
exposure-response trend P values based on two-sided likelihood ratio tests (1 df)
of zero slope for continuous formaldehyde exposure among exposed person-years
only. The points represent the relative risk estimates based on the cumulative
number of cases and person-years accrued from the start of the study to that point
in time and for 2004 are equivalent to the relative risk estimates presented in
Table 2. HLP = lymphohematopoietic malignancies, NHL = non-Hodgkin
lymphoma, HDG = Hodgkin lymphoma, MM = multiple myeloma, LEU =
leukemia, LYL = lymphatic leukemia, MYL = myeloid leukemia, RR = relative
risk.
Source: Beane Freeman, et al. (2009)
This document is a draft for review purposes only and does not constitute Agency policy.
4-415 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
0 -1 J— * «*- •« , » — r I— -* * .»
I960- wn 1980 1990 W *96© W® 1980 1890 2000 1960 WO W *990 5PWJ0
Gmemlar year of «ci of fe»sw-u£
Figure 4-34B. Association between average intensity of formaldehyde
exposure and the risk of lymphohematopoietic malignancy.
Relative risks for medium (0.5 - 0.9 ppm) and high ( > 1.0 ppm) average-intensity
formaldehyde exposure categories compared with the low exposed category (0.1 -
0.4 ppm) and P values for trend tests among the exposed person-years for
lymphohematopoietic malignancies by year of end of follow-up, 1965 - 2004.
Values plotted at 0.1 represent RR •• = due to no cases in the exposure category.
The small graphs above the relative risk plots represent the exposure - response
trend P values based on two-sided likelihood ratio tests (1 df) of zero slope for
continuous formaldehyde exposure among exposed person-years only. The points
represent the relative risk estimates based on the cumulative number of cases and
person-years accrued from the start of the study to that point in time and for 2004
are equivalent to the relative risk estimates presented in Table 3 . HLP =
lymphohematopoietic malignancies, NHL = non-Hodgkin lymphoma, HDG =
Hodgkin lymphoma, MM = multiple myeloma, LEU =
leukemia, LYL = lymphatic leukemia, MYL = myeloid leukemia; RR = relative
risk.
Source: Beane Freeman, et al. (2009)
This document is a draft for review purposes only and does not constitute Agency policy.
4-416 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Although results are mixed across the studies (Table 4-90), an association between
formaldehyde exposure and leukemia mortality is supported by cohort analyses of embalmers,
pathologists and anatomists (Hayes et al., 1990; Walrath and Fraumeni, 1983; Walrath and
Fraumeni 1984; Hall et al., 1991; Levine et al., 1984; Stroup et al., 1986; Matanoski et al., 1989).
Formaldehyde exposure and formaldehyde-related occupation are associated with leukemia
diagnosis in a case-control study (RR = 5.79 (95% CI 1.44-23.25), but not formaldehyde
exposure alone (RR = 0.96; 95% CI 0.54-1.71) (Stellman et al., 1998).
In contrast, SMR analyses of the industrial cohorts do not indicate a similar association
(Coggon et al., 2000; Beane Freeman et al., 2009, Pinkerton et al, 2004). Although the SMR
analysis provided for the NCI cohort does not indicate a positive association for all leukemia
using an external reference group (Beane Freeman et al., 2009), the SMR for exposed versus
unexposed workers within the cohort suggests all leukemia is elevated 2.1-fold with this internal
comparison (95% CI 0.99-4.56)4. A positive exposure-response relationship further strengthens
the association of formaldehyde exposure to leukemia mortality (Beane Freeman et al., 2009).
Where the referent group is defined as 'low exposed' individuals, leukemia is elevated in the
highest peak exposure category (RR = 1.42; 95% CI 0.92-2.18) compared to both the referent
group and the unexposed category (RR = 0.59; 95% CI 0.25-1.36), and there is a statistically
significant trend across all groups (p = 0.02). Categorical analysis for the average intensity and
cumulative exposure metrics suggests greater mortality in the high-exposure groups versus the
'low exposed' individuals (RR = 1.10 [95% CI 0.68-1.78] and 1.11 [0.7-1.74], respectively), but
analysis of individual results across the exposure-response range indicates cumulative exposure
is a better predictor (p = 0.08 for trend across all exposed and unexposed.)
Several meta-analyses have been conducted for formaldehyde exposure and leukemia
which indicate a positive association. Collins et al. (2004) report an overall RR for 18 available
studies of 1.1 (CI 1.0-1.2), suggesting an association of leukemia with formaldehyde exposure.
This association was stronger for both pathologists/anatomists (1.4; CI 1.0-1.9) and embalmers
(RR = 1.6; 1.2-2.0) than for industrial workers (RR = 0.9; 0.8-1.0). Study design also impacted
the apparent strength of association, with stronger associations seen in case—control studies (RR
= 2.4; 0.9-6.5) versus cohort studies (RR = 1.0; 0.9-1.2). Bosetti et al. (2008) reported an
association between formaldehyde exposure and leukemia mortality with a pooled RR of 1.39
(95% CI 1.15-1.68) for 8 groups of professional workers. In the same analysis, the pooled RR
for the 4 industrial cohorts was 0.90 (0.75-1.07). Zhang et al. (2009) reported a pooled RR of
Var
f f
In
SMR
W
LExposed
\ V
SMR
1
1
LNon exp osed J j
()bsl;
Obs,
Exposed Non exp osed
This document is a draft for review purposes only and does not constitute Agency policy.
4-417 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
1.54 (95% CI 1.18-2.00) for all cohorts identified in their meta-analysis, although this pooled RR
should be considered with some caution, as myeloid leukemia alone was included in the analysis
where available (Zhang et al., 2009).
4.5.2.4. Subtype Analysis
Given the associations discussed above between formaldehyde exposure and both all
LHP cancers and all leukemia, further analysis is needed to examine if the observed increase in
all LHP cancers is primarily a reflection of increased leukemia, or if other types of LHP cancers
may be elevated as well. Although analysis of mortality data by sub-type may provide a better
understanding of the specific disease associations, there are potential pitfalls as well. Chief
among these concerns are the potential for disease misclassification (especially in studies with
older mortality data) and lack of statistical power as the number of observed cases is reduced by
considering sub-types. Case control studies by design address specific diseases and are well-
suited for sub-type analysis, but often provide little exposure information. The following
analysis will draw from the available data to examine which forms of LHP malignancies may be
associated with formaldehyde exposure.
There has been speculation that the association between formaldehyde exposure and
increases in all LHP cancers and all leukemia are driven by increased myeloid leukemia (Pyatt et
al., 2008; Heck and Casanova 2004; Golden et al., 2006). If this were the case, then mortality
from LHP cancers other than myeloid should not be elevated, once the excess mortality from
myeloid leukemia is accounted for. Only 2 studies provide the data to evaluate this hypothesis -
both conducted by the NCI (Hayes et al., 1990 and Beane Freeman et al., 2009). From the
published data, crude mortality statistics can be calculated for alternative disease groupings
(Table 4-91). In the NCI embalmer study (Hayes et al., 1990), only myeloid leukemia was
statistically elevated in the subtype analysis. For the NCI industrial cohort (Beane Freeman et
al., 2009), elevations were also seen for Hodgkin lymphoma relative to the referent group. In
both cases, the association between formaldehyde exposure and LHP malignancies remains when
myeloid leukemia is dropped from the analysis. Further, similar associations are found when all
leukemia and myeloproliferative diseases are dropped from the analysis and only solid tumors of
lymphoid origin are included (lymphosarcoma and reticulosarcoma, Hodgkin lymphoma, non-
Hodgkin lymphoma and multiple myeloma). These reanalyses illustrate the need for a more
careful sub-type analysis to assess the potential for associations between formaldehyde exposure
and various forms of LHP cancers.
This document is a draft for review purposes only and does not constitute Agency policy.
4-418 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Table 4-91. Secondary analysis of published mortality statistics to explore
alternative disease groupings within the broad category of all lymphohematopoietic
malignancies
ICD-K ( odes
I .S. cmhalmcrs
(\\ hole cohort)
(1 l;i\cs el ;il . I'WtM
PMR
C)5"'n CI)
I .S. indusln
(peak exposure mclric: >4
ppm \s. >0 lo <2 ppni)
(I3e;ine 1 leeiiuiii el ;il . )
Kclali\c risk
CI)
All Lymphohematopoietic
Malignancies
200-209
1.39a
n 15-1 671 b
1.37c
n n3-i snc
Allei'nali\e Disease Croupings
Lxcluile Mj eloid Leukemia
200-204,
206-209
1.35
(0.99-1.85) b
1.31
(0.97-1.75) d'e
Solid tumors of lymphoid
origin
(Lymphosarcoma and
reticulo sarcoma, Hodgkin
lymphoma, non-Hodgkin
lymphoma and multiple myeloma)
200-203
1.24a
(0.84-1.84) b
1.33d
(0.93-1.90) d'e
PMR--
Obs
Exp
Far (log PMR) = — + —
Obs Exp
See Table 2 of Beane Freeman et al. (2009)
jee lauie z oi iseane riceman
d ^ _ DeathsComparisonGrov/PerSOn ~ TimeComparisonGrov
Deaths referent Gmlv /Person - TimereJerent Gmup
pt r 1 D8c
referent Group 1WO
108-19"
103 -14_
PT
referent Group
PT
Comparison Group
where
PT
1 1 rv
T
Comparison Group
Var([ogRR) =
103c x 1.37c
1
= 0.765
1
DeathsComparison Group Deaths referent Group
4.5.2.5. Myeloid Leukemia
The associations between myeloid leukemia and formaldehyde exposure are strong and
consistent (Table 4-92). Of the four studies which formally assess myeloid leukemia mortality,
all are positive, including cohorts of both professional and industrial workers (Beane Freeman et
al., 2009; Hayes et al., 1990; Pinkerton at al., 2003; Stroup et al., 1986). Although few cases
exist for further subtype analysis, the available data indicate either no differences in SMRs for
acute myeloid leukemia (AML) versus chronic myeloid leukemia (CML) (Hayes et al., 1990;
Pinkerton et al., 2003) or suggest CML is more prominent (Blair et al., 2000; Stroup et al., 1986).
This document is a draft for review purposes only and does not constitute Agency policy.
4-419 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-92. Summary of studies which provide mortality statistics for
2 myeloid leukemia sub-types.
Study population
Uclerencc
Myeloid
Leukemia
Acute
Myeloid
Leukemia
Chronic
Myeloid
Leukemia
S.M K Analysis'
Garment workers
(n= 11,098)
1.44 (0.08-2.37)
1.34
(0.61-2.54)
1.39
(0.38-3.56)
Pinkerton at al.,
2003
Anatomists
(n=2,317)
NR
NR
8.8*
Stroup et al., 1986
Industrial workers
(n=25,619)
0.90 (0.67-1.21)
SMR Ratio
1.38 (0.65-2.97)
(exposed/unexposed)
NR
NR
Beane Freeman et
al., 2009
IWIU Analysis1
1 Embalmers, U.S.
(N=4,046)
1.57 (1.01-2.34)
1.52
(0.85-2.52)
1.84
(0.79-3.62)
Hayes etal., 1990
Case-Control Studies'
White men diagnosed
with leukemia
(Iowa and Minnesota)
(n=622)
NR
Low:
0.9 (0.5-1.6)
High:
NR
Low:
1.3 (0.6-3.1)
High:
2.9 (0.3-24.5)
Blair etal., 1993
4
5 * Leukemia SMR 1.5 (0.7-2.7) {5 of 10 deaths due to myeloid}; Chronic Myeloid Leukemia (CML) SMR of 8.8
6 1. Relative risk estimate (SMR, PMR, or OR) presented with 95% confidence intervals, where available.
7
8
9 Walrath and Fraumeni (1983, 1984) note that AML is prominent in their analyses of New
10 York and California licensed embalmers; however, they do not provide PMR analyses for CML.
11 Walrath and Fraumeni (1983 and 1984) report leukemia cell types - for both studies the majority
12 of myeloid leukemia are acute (5/6 and 4/6, respectively, for New York State and California
13 embalmers). However, PMRs cannot be calculated for AML versus CML in this paper, as
14 comparison rates are not available from the 1920's through the 1960's - the timeframe with the
15 majority of deaths. The authors do contrast the observed rate of AML in the cohort to the
16 background rate for AML in white men in the 1970s - but given the potential misclassification of
17 late stage CML as AML, especially historically, this may not be an appropriate comparison.
18 Additionally, one would expect older data to over-represent AML rather than CML due to
19 diagnosis in the early decades of CML in the blast crisis as AML. Therefore, although these
20 studies support an association between formaldehyde exposure and myeloid leukemia in general
This document is a draft for review purposes only and does not constitute Agency policy.
4-420 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(Walrath and Fraumeni, 1983; 1984), the reported AML and CML subtype information does not
allow a satisfactory sub-type analysis for myeloid leukemia.
4.5.2.6. Solid Tumors of Lymphoid Origin
Multiple myeloma, Hodgkin lymphoma, non-Hodgkin lymphoma, lymphosarcoma,
reticulosarcoma, and other lymphomas may all be derived from immune cells outside of the bone
marrow compartment, in peripheral blood, in the gut and respiratory mucosa and immune tissues
at the POE (e.g., lymph nodes, mucosa-associated lymphoid tissue (MALT), gut-associated
lymphoid tissue (GALT) (Greaves, 2004). The only meta-analysis to specifically address
lymphoid malignancies found evidence for increased lymphoma (Hodgkin lymphoma (pooled
RR = 1.23; 95% CI 0.67-2.29) and multiple myeloma (1.31; 1.02-1.67), but not for non-Hodgkin
lymphoma (1.08; 0.86-1.35) (Zhang et al., 2009). As seen in Table 4-93 below, individual study
results are mixed for these lymphoid cell-line malignancies, as they are for all LHP cancers and
all leukemia above. Although these tumors are from mature lymphocytes, there is still variability
in the etiology, natural history and risk factors for the many sub-types which are included in
these categories.
There is evidence for an exposure response relationship for both Hodgkin lymphoma and
multiple myeloma in the NCI industrial cohort among exposed workers (Beane-Freeman et al.,
2009). Clear exposure response relationships for Hodgkin lymphoma are defined with all three
metrics of exposure, peak average intensity and cumulative exposure (p=0.01, p=0.05 and
p=0.08 respectively for mortality through 2004). These associations have been evident from first
follow-up through the current publication, and statistically significant for the majority of the
follow-up period demonstrating that this is a strong and consistent finding in the NCI cohort
(Figure 4-34 A&B) (Beane-Freeman et al., 2009). Although the overall SMR for multiple
myeloma does not indicate an association, trends across time indicate consistent elevation of
multiple myeloma mortality with both peak and average intensity of exposure, where the
statistical strength of the association with peak exposure increases with follow-up (Figure 4-34
A&B).
This document is a draft for review purposes only and does not constitute Agency policy.
4-421 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 4-93. Summary of mortality statistics for Hodgkin lymphoma,
lymphoma and multiple myeloma from cohort analyses of formaldehyde
exposed workers.
Siudj popiiliiiion
Reference
llod^kin
I.Miinhoniii
Noii-llod^kin
I.Miinhoniii
Mn lliplo
\l\cloill;i
SMR An;il\sis"
Pathologists
(n=2,079)
1.4
2.0 (p<0.05)
NR1
Harrington and
Shannon, 1975
Pathologists
(n=4,512)
1.21 (0.03-6.71)
1.44 (0.69-2.63)
NR
Halletal., 1991
Male Anatomists
(n=2,327)
_ 2
0.7 (.1-2.5) 3
2.0 (0.7-4.4) 4
NR
Stroup et al., 1986
Male pathologists
(n=6,lll)
0.36 (0.04-1.31)
1.31 (0.66-2.35) 3
1.54 (0.82-2.63)4
NR
Matanoski et al.,
1989
Chemical workers
(n=2,026)
2.94 (0.33-
10.63)
NR
NR
Wong et al., 1983
British Chemical plants
(n=14,014)
0.36(0.01-2.01)
0.89 (0.41-1.70)
1.18(0.48-
2.44)
Coggonetal., 2003
Swedish workers- abrasive
production plant (n=911)
NR
2.0 (0.2-7.2)
4.0 (0.5-14)
Edlingetal., 1987a
Industrial workers
(n=25,619)
1.42 (0.96-2.10)
0.86 (0.70-1.05)
0.94 (0.71-
1.25)
Beane Freemen et
al., 2009
PMR An;il\sis"
Embalmers, New York
(n=l,132)
0.87 (p<0.05)
1.083
1.224
NR
Walrath and
Fraumeni, 1983
Embalmers, CA
(n= 1,007)
_ 2
3.103
1.334
NR
Walrath and
Fraumeni, 1984
Embalmers, U.S.
N=4,046
0.72 (0.15-2.10)
1.26 (0.87-1.76)
1.12 (0.58-1.96)3
1.35 (0.84-2.01)4
1.37 (0.84-
2.12
Hayes et al., 1990
( ;iso-( onirol Siudies''
Women mi ( oiiiiccliciil in <>t> 11
\k
1 ^ ( 1 u-1 " )
\k
W ang el al . :ou<>
White men, Iowa and Minnesota
(n=622)
NR
1.2(0.9-1.7)
NR
Blair etal., 1993
ACS Cancer Prevention Study
II (n=128)
NR
NR
1.8 (0.6-5.7)
Boffetta et al., 1999
Men, ACS Cancer Prevention
Study II (n=45,399)
NR
0.92 (0.5-1.68)
2.88 (0.40-10.5)5
0.74 (0.27-
2.02)
Stellman et al., 1998
Danish workers (n=l,098)
NR
NR
Heineman et al.,
1992
Danish women (607)
NR
NR
1.6 (0.4-5.3)
Pottern et al., 1992
Notes :
1: NR is not reported
2. "—" no cases observed.
3. Lymphosarcoma and reticulosarcoma only
4. "other lymphoma"
5. Formaldehyde exposure in a wood-related occupation. RR for wood-related occupation alone was not elevated
0.97 (0.55-1.73)
6. Relative risk estimate (SMR, PMR, or OR) presented with 95% confidence intervals, where available
This document is a draft for review purposes only and does not constitute Agency policy.
4-422 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.5.2.7. Supporting Evidence from Animal Bio-Assays for Formaldehyde-Induced
Lymphohematopoietic Malignancies
Chronic animal studies provide supporting evidence for formaldehyde-induced leukemia
and lymphoma (Soffritti et al., 1989; Battelle Laboratories, 1981). Although the majority of
chronic animal bioassays do not report either leukemia or lymphoma, it should be noted that
many studies focused primarily on respiratory tract and did not provide routine examination of
other organs, limiting the detection of leukemia and lymphoma (Horten et al., 1963; Holmstrom
et al., 1989; Wouterson et al., 1989; Appleman et al., 1988; Monticello et al., 1996; Dalbey,
1982). Kamata et al. (1986) did examine additional organs, but there were only 5 animals at
each sacrifice. Similar issues are seen with some of the drinking water studies, where Takahashi
et al. (1986) focus on the stomach and intestines, and the study in Wistar rats by Tobe et al.
(1989) only included 20 animals per sex per exposure group with interim sacrifices. Therefore,
few studies remain to be explored which are informative about the carcinogenic potential of
formaldehyde on the LHP system. Table 4-94 lists the studies from the chronic bioassays which
have the potential to detect LHP malignancies.
Soffritti et al. (1989) were the first to publish a finding of formaldehyde-induced
leukemia in an animal bio-assay. Sprague-Dawley rats (50 per each sex) were exposed to
formaldehyde in drinking water at 0, 10, 50, 100, 500, 1000, 1500 mg/L given ad libitum.
Lymphoblastic leukemia and lymphosarcomas were the most common lesion reported, and
exhibited an apparent dose-dependent increase in both male and female rats (Table 4-95). All
hemolymphoreticular cancers summed together reflected similar trends, but the other types did
not show similar relationship with exposure (e.g. immunoblastic lymphosarcoma). A subsequent
publication of full study results reports all lesions together as lymphoma and leukemia, providing
less-specific information (Soffritti et al 2002.) Additionally, there is a large discrepancy in
reported lesions between the two study reports where nearly double the incidence is reported in
2002. Dr. Soffritti explains the increase in reported LHP malignancies as a result of tabulating
the full histopathological findings in the 2002 report (personal communication, 23 July 2009).
Although the second report from Soffritti et al. (2002) have been broadly criticized on both of
these points (summing of dissimilar lesions and discrepancies in reporting), these errors do not
iupune the original report. The 1989 report does distinguish between different sub-types
providing positive results for lymphoblastic leukemia/lymphosarcoma and histological
examinations were consistently conducted between treatment groups - even if further results
were published at a later timepoint. Therefore, the findings of Soffritti et al. (1989) are
considered as supportive of the biological plausibility of formaldehyde-induced
lymphohematopoietic malignancies.
This document is a draft for review purposes only and does not constitute Agency policy.
4-423 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-94: Summary of chronic bioassays which address rodent leukemia
2 and lymphoma
3
Stuilv
Histopathology
Enripoint
Results
C oni nients
Drinking Water Exposure
Male and Female Sprague-Dawley Rats
Sofritti et al.
1986
Complete histopathology
Lymphocytic
leukemia and
lymphosarcoma
Increased, showing a
dose-response
Life-long study
High exposure of 1,500 mg/1 in
water
Male and Female Wistar Rats
Til et al, 1989
Complete histopathology
in control and high dose
group (15 ppm)
Lymphoma,
leukemia
No increase
(3 lymphomas and 1
leukemia found in 200
animals at the 2 yr
sacrifice)
2-year bioassay
High exposure of approximately
1900 mg/1
(82mg/kg for males and 109
mg/kg for females)
Inhalation Exposures
Male rats, Sprague-Dawley
Sellakumar et
al., 1985 ;
Albert etal.,
1982
Necropsy focused on
respiratory tract: also
liver, spleen, kidney and
testes and organs
demonstrating gross
pathology
Lymphoma
No increase
Life-long study - high mortality at
24 months (>80%)
Male rats, ¥2
(44
Batelle,
Columbus
Laboratories,
1981
Complete histopathology
in control and high dose
group (15 ppm)
Leukemia, all
No increase
Extended study - high mortality
Female Rats, F344
Batelle,
Columbus
Laboratories,
1981
Complete histopathology
in control and high dose
group (15 ppm)
Leukemia, all
Increased in mortality
adjusted incidence.
P=0.00561
Extended study - high mortality
Apparent elevation in 2 ppm and
6 ppm treatment groups as well
(fig5-xx)- statistical comparison
to controls is problematic
Female mice - C57BL/6xC3HFl
Batelle,
Columbus
Laboratories,
1981
All organs in control and
high dose group (15
ppm)
Lymphoma, all
26% in FA-exposed
(15ppm)
16% in control
P=0.0617
Extended study
All mice included in statistics
conducted by Batelle Lab.
4
5 1. Original statistical analysis provided by Battelle, Columbus Laboratories. Significance set at P<0.0167. Analysis of adjusted
6 data where time to lesion and survivorship were considered, Cox (1972, Tyrone (1975).
7
This document is a draft for review purposes only and does not constitute Agency policy.
4-424 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Table 4-95. Incidence of lymphoblastic leukemia and lymphosarcoma orally
dosed in Sprague-Dawley rats
Lymphoblastic leukemia and
lymphosarcoma
Tumor bearing animals
(%)
Male
Female
control
3
1
vehicle
8
2
10
0
2
50
4
6
100
8
4
500
8
4
1000
12
10
1500
22
10
Source: Soffritti et al. (1989)
The two-year bioassay by Til et al. (1989) in male and female Wistar rats indicates no
increase in leukemia and lymphoma, with only 4 tumor-bearing animals in all treatment groups
sacrificed at 24 months. The drinking water levels were similar at the highest dose of both
studies. The major difference in study design is length, which may have influenced results, as
leukemia is a late-life malignancy in rodents. Two-year survival in the Soffritti et al. (1989)
study varied between 50-60%. These animals would be available to develop leukemia after the
two-year window of the Til et al (1989) study. Any potential role of strain differences is
unknown. Overall, the results of Soffritti et al. (1986) are strong, showing an exposure-response
relationship, in a lifelong study, appropriate for late-life malignancies. Unlike the GI tract
tumors, early-life exposure to formaldehyde in drinking water did not increase LHP
malignancies (Soffritti et al., 1989).
The largest and most comprehensive study of health effects from formaldehyde
inhalation exposures is the study reported by Kerns et al. (1983) and Swenberg et al. (1980)
conducted at the Columbus Laboratory of Battelle Corporation (1981). Although the summary
reports of this study do not discuss leukemia or lymphoma rates, mouse lymphoma and rat
leukemia were selected by the study pathologist and biostatistician for analysis (Battelle
Laboratory, 1981). Statistical analysis performed by Battelle Laboratories which accounted for
time-to-lesion and survivorship rates did indicate a statistically significant increase in female rat
leukemia (P = 0.0003) and a nearly significant increase in female mouse lymphoma (P = 0.06).
No trend analysis could be performed, as only gross pathology was conducted on mid-dose mice
This document is a draft for review purposes only and does not constitute Agency policy.
4-425 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
and rats (2 and 6 ppm, respectively). EPA has further analyzed these data to better understand
the significance of these findings. The percentage of lymphoma—bearing female mice increased
from 18% in control mice to 28% in mice exposed at 15 ppm for 24 months (6 hr/day, 5
days/wk) (P<0.05). Female rat leukemia was similarly elevated to 26%, 22% and 24% by
inhalation exposure to 15 ppm, 6 ppm and 2 ppm, respectively, versus 16% in controls when
early deaths prior to the first observed leukemia are removed from the analysis (21 months). In
contrast, leukemia was not elevated in formaldehyde-exposed male F344 rats within the same
study.
Differences in study design may account in part for mixed results. The lifelong study by
Soffritti et al. (1989 may have allowed for detection of malignancies developing late in life,
whereas the other drinking water study by Til et al. (1989) sacrificed all animals at 24 months.
Even though the exposure levels were similar, the studies are not directly comparable. Likewise,
it is hard to directly compare results from the two major inhalation studies in rats. Although a
life-long study, the mortality for rats in the Sellakumar et al. (1985) study was greater than 80%
at 2 years. Additionally, the pathology examination was much less rigorous than in the Battelle
Laboratory study, perhaps missing smaller lesions. Therefore, the increase in formaldehyde-
induced leukemia seen in female F344 rats late in life (Battelle laboratories, 1981) may be
reflecting a more sensitive study design. Finally, strain differences may account for different
susceptibilities as well. In summary, the available evidence from chronic animal studies
supports the biological plausibility of the formaldehyde-induced LHP malignancies observed in
epidemiologic studies. The two positive rat studies, by different routes of exposure, along with a
positive result for formaldehyde-induced mouse lymphoma make a substantive case for
formaldehyde-induced LHP malignancies.
The epidemiologic studies provide sufficient evidence to conclude that there is a causal
association between formaldehyde exposure and lymphohematopoietic malignancies. When data
are evaluated for all leukemia together, again there is sufficient evidence to establish a causal
association, with consistent positive results in individual studies as well as 3 independent pooled
analyses. Mortality from myeloid leukemia, as well as mortality attributed to "other and
unspecified leukemia" is consistently elevated where reported. In addition, strong evidence for a
causal relationship between formaldehyde exposure and Hodgkin lymphoma is provided by the
consistent associations seen between formaldehyde exposure and Hodgkin lymphoma in the NCI
industrial cohort, with elevations observed across decades of follow-up and significant exposure-
response relationships for all three exposure metrics examined in the most recent follow-up
(Beane Freeman et al., 2009).
This document is a draft for review purposes only and does not constitute Agency policy.
4-426 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
4.5.3 Carcinogenic Mode(s) of Action
The US EPA 2005 Guidelines for Carcinogen risk Assessment recommend a Mode of
Action (MOA) analysis when data are available for evaluation. The purpose of this MOA
analysis is to determine if sufficient data exist to adequately inform the exposure-response
relationship for cancer below the range of observed data in either human or animal studies.
Since the majority of the data supporting the carcinogenicity of formaldehyde comes from
animal bio-assays and epidemiological studies of workers, EPA must extrapolate from the
observed risk of cancer mortality/incidence in those studies to levels considered protective of
human health for lifelong environmental exposures. In this context, the US EPA cancer
guidelines provide a framework to review MOA information for relevant data to establish an
MOA informing appropriate low-dose extrapolation.
The supporting data for the MOA evaluation of formaldehyde are complex, and presented
across multiple sections of a large document; therefore, this section includes a brief summary of
the biological actions of formaldehyde and key mechanistic data which are believed to be
relevant to the MOA evaluation (Section 4.5.3.1). This information is not intended as a stand-
alone description of the evidence for a particular mechanism, but is intended to highlight the
major supporting arguments and direct the reader to text providing more detailed discussion.
The summary of data discussed below combines what is known about the human cancer
of concern (nasopharyngeal cancer, sinonasal cancer, leukemia and other lymphohematopoietic
cancers) with the potential formaldehyde-specific mechanisms of action to postulate
carcinogenic modes of action for each cancer or group of cancers (Section 4.5.3.2 and 4.5.3.2).
The resulting evaluation provides multiple possible MO As for formaldehyde-induced cancers
where some key mechanistic events may be commonly at work in different tissues, and some key
events may be more relevant to a specific tissue/cancer type. Each of these MO As is evaluated
with respect to its relevance to human cancer, and the overall weight of evidence for its
relevance to formaldehyde-related human cancer.
Overall, multiple MO As considered relevant to humans are presented for each cancer
type. Although some MO As may have a greater level of supporting evidence, this reflects in part
how well a particular mechanism or key event may have been studied. For example, there are a
large number of studies across many testing systems, and levels of biological organization to
support the mutagenicity of formaldehyde. In contrast other likely MO As, such as viral
reactivation, have little direct mechanistic evidence, but the available evidence is supportive.
The MO As considered most relevant to upper respiratory tract cancers (e.g. NPC and
sinonasal cancer) are: 1) direct mutagenicity; 2) inhibition of DNA repair mechanisms; 3)
formaldehyde-induced cell proliferation; 4) cytotoxicity-induced cell proliferation; 5) tumor
This document is a draft for review purposes only and does not constitute Agency policy.
4-427 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
promotion activity; and 6) localized immunosuppression/viral reactivation (Section 4.5.3.2). The
majority of these MO As would apply equally to immune cells present at the site of first contact
and may also contribute to those lymphohematopoietic cancers which arise from peripheral
immune cells (e.g. Hodgkins lymphoma, multiple myeloma and some forms of leukemia).
Additional MO As are considered, specifically for formaldehyde-induced leukemia are: 1)
damage of a circulating hematopoietic stem cell or progenitor cell at the site of first contact; and
2) bone marrow toxicity (Section 4.5.3.3).
In summary - no single MOA is singled out as the best explanation for cancer resulting
from formaldehyde exposure. Only one MOA - cytotoxicity induced cell proliferation - suggests
an exposure threshold below which the MOA would not be active. However this MOA is the
least applicable to humans and other MO As are considered operative at exposures below
exposures associated with cytotoxicity-induced cell proliferation. Therefore, multiple MO As are
considered supported by formaldehyde-specific mechanistic information which provide
biological plausibility for the cancers observed in formaldehyde exposed populations.
4.5.3.1 Mechanistic Data for Formaldehyde
4.5.3.1.1 DNA Reactivity/Genotoxicity/Mutagenicity
An agent's genotoxic potential and ability to induce mutations is a key consideration in
assessing a carcinogenic MOA, as cancer results from a series of genetic and epigenetic
alterations affecting genes that control cell growth, division and differentiation (Hanahan and
Weinberg, 2000; Vogelstein et al., 1988; Kinzler and Vogelstein, 2002). The US EPA Cancer
guidelines suggest several lines of evidence which are key to evaluating a mutagenic MOA: 1) Is
the chemical under study DNA-reactive and/or has the ability to bind to DNA; 2) Does the
chemical generate positive results in in vitro mutagenic test systems (specifically gene mutations
and chromosomal aberrations); and 3) Does the chemical induce manifestations of genetic
damage in in vivo tests (specifically gene mutations and chromosomal aberrations) and 4) Does
the chemical have properties and structure-activity relationships (SAR) similar to known
mutagens (US EPA, 2005). As reviewed in Section 4.3 above, there is adequate evidence for
formaldehyde-induced genotoxicity and mutagenicity for consideration of these key events in
formaldehyde's carcinogenic MOA.
Formaldehyde induces a variety of genotoxic and mutagenic events when tested both in
vitro and in vivo systems including DNA-protein crosslinks (DPC or DPX), point mutations,
DNA single strand breaks (SSB) and chromosomal aberrations (CAs) (See Section 4.3).
Formaldehyde, as a reactive chemical, also forms DNA adducts and DNA-DNA crosslinks
(DDC) and may act to form adducts between other chemicals and DNA (Brutlag et al., 1969;
This document is a draft for review purposes only and does not constitute Agency policy.
4-428 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Donecke, 1978; Ohba et al., 1979; Fennel, 1999; Casanova-Schmitz and Heck, 1983; Casanova-
Schmitz and Heck, 1984; Heck and Casanova, 1987 and Casanova et al., 1989). The high
reactivity of formaldehyde results in little specificity indicating that a range of adducts and
crosslinks might be expected.
Numerous studies have shown that formaldehyde induces genotoxic and mutagenic
effects under a variety of experimental conditions (see section 4.3 for a detailed discussion, also
reviewed by IARC 2006; Ma and Harris 1988 and Auerbach et al, 1977). As discussed,
formaldehyde is known to directly react with DNA forming DPC and DNA adducts. Mutations
may occur during repair of formaldehyde-induced DNA damage, or as a result of replication
errors during mitogenesis. Additionally, there is some evidence that DNA single strand breaks
(SSB) may be induced directly by formaldehyde reactivity (Grafstrom et al, 1984). Clastogenic
effects including increased micronuclei (MN), chromosomal aberrations (CAs) and sister
chromatid exchanges (SCEs) are also reported in a range of in vitro study systems.
Formaldehyde caused a concentration-dependent icrease in calstogenicity (e.g. MN) in
human cell lines deficient in either DNA nucleotide excision repair (NER) or DDC repair
systems even though there is no change seen in DPC induction or removal between these cell
lines (Speit et al., 2000). These data suggest that alteration of DNA repair, not DPC removal,
contributes to formaldehyde-induced clastogenicity. Since DPC repair involves proteolytic
removal of proteins from the DNA, the authors hypothesize that single peptides or small peptide
chains cross-linked to the DNA as in the case of DPC are critical to formaldehyde-induced
mutations.
Formaldehyde-induced MN and CAs are associated to concentration-dependent
mutagenic effects in L5178Y mouse lymphoma cells (Speit and Merk, 2002). Detailed analysis
of both spontaneous and formaldehyde-induced lesions indicate that recombination or deletion of
DNA from the thymidine kinase {tk) locus was primarily responsible for the loss of heterogeneity
leading to the observed mutant phenotype. Therefore, it is believed that formaldehyde is
mutagenic by a clastogenic mechanism, rather than through point mutations in the L5178Y
mouse lymphoma cell system. This finding is consistent with Craft et al. (1987) who
demonstrated formaldehyde-induced mutagenicity in the tk locus of TK6 human lymphoblastoid
cells, while Graftsrom et al, (1984) demonstrated increased SSBs in formaldehyde-exposed
human cell lines. The elegant series of experiments by Speit and Merk provide the possible links
between DPC, clastogenicity and locus-specific mutations firmly demonstrating formaldehyde-
induced mutations in the in vitro mouse lymphoma testing system.
Formaldehyde is genotoxic at the portal of entry (POE) in animal studies, resulting in
increased DPC formation in the nasal mucosa as discussed above. However, there are no animal
This document is a draft for review purposes only and does not constitute Agency policy.
4-429 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
studies which directly examine the mutagenicity in nasal or respiratory epithelial cells in the
early stages of exposure. It is likely that the mutations are seen in advanced stage of the tissue
around the transformation stage with formaldehyde exposure. With weak positive results in
pulmonary lavage cells (Dallas et al., 1992) and clastogenicity demonstrated in gastro-intestinal
epithelial cells of rats (Migliore et al., 1989), below exposure levels which trigger regenerative
cell proliferation, the existing evidence, although thin, supports clastogenic effects of
formaldehyde.
Clastogenic effects are consistently reported in humans exposed to formaldehyde in the
industrial workplace or during anatomy or mortuary classes (See section 4.3 for a full
discussion). Increased micronuclei have been reported in nasal epithelial cells from industry
workers (Ballarin et al., 1992; Ye et al., 2005), buccal epithelial cells from anatomy and
mortuary science students and/or staff (Kitaeva et al., 1996; Titenko-Holland et al., 1996; Burgaz
et al., 2001; 2002 compared to corresponding controls). Comparisons of micronuclei in nasal
and buccal cells of anatomy students before and after classes where they are exposed to
formaldehyde indicate an increase in clastogenicity (Ying et al., 1997). An examination of
exfoliated buccal and nasal cells in mortuary students indicates greater increases in centromere-
negative micronuclei, suggesting the effects are due to chromosome breakage or clastogenicity
rather than aneuploidy (Titenko-Holland et al., 1996). Micronuclei were also increased in a
dose-dependent manner in buccal cells as well as peripheral blood lymphocytes (PBLs) in
mortuary students during the course of an embalming class; however, SCEs were reduced in
post-exposure samples (Suruda et al., 1993). Buccal, oral and nasal cells present at the portal of
entry may be directly exposed to formaldehyde and thus reports of clastogenic effects are
consistent with direct interaction of formaldehyde at the POE. There is some supporting
evidence for the mutagenicity of formaldehyde in human populations. Shaham et al. (2003)
reported a increase in mutant p53 protein in the PBLs of individuals with mean formaldehyde
exposure duration of 16 years. Additionally there was is a significant association between
mutantp53 protein and DPC in this study suggesting a relationship between the formaldehyde's
genotoxic effects. More recently, Zhang et al., (2010) have reported aneuploidy in circulating
hematopietic stem cells in formaldehyde exposed workers with increases in both monosomy7
and trisomy 8.
In summary, there are several lines of evidence supporting mutagenic effects of
formaldehyde exposure:
1) Formaldehyde directly interacts with DNA generating DPC,
2) DPC in tissues at the POE exhibit a dose-response relationship to formaldehyde
exposure,
This document is a draft for review purposes only and does not constitute Agency policy.
4-430 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
3) Formaldehyde-induced DPC are associated with formaldehyde-induced MN and CAs,
4) Mutations induced by formaldehyde due to small deletions and rearrangements in
DNA in various experimental systems are consistent with formaldehyde's observed
clastogenic effects (MN and CAs),
5) Formaldehyde-induced mutations and clastogenic effects occur at levels below where
significant cytotoxicity is detected, and
6) Formaldehyde exposure has been correlated to similar increased MN and CAs in
human buccal and oral cells corresponding to sites where formaldehyde-induced tumors
arise.
4.5.3.1.2 Inhibition of DNA repair
Studies indicate that formaldehyde exposure may inhibit DNA repair mechanisms
directly (See Section 4.3.1.5). Graftsrom (1985) first documented formaldehyde effects on DNA
repair mechanisms, reporting that formaldehyde treatment of human bronchial fibroblasts in vitro
inhibited repair of 06-methyl-guuanine adducts induced by N-methyl-Nitrosurea (NMU).
Inhibition of DNA repair in human keratinocytes and fibroblasts cultured at 10 |iM
formaldehyde affected repair of DNA single strand breaks from ultraviolet light but was specific
to UVB and UVC, not impacting repair of single strand breaks from UVA (Emri et al., 2004).
To determine if formaldehyde may have similar effects in exposed humans, Hayes et al.,
(1997) assessed the activity 06-alkylguanine-DNA alkyltranferase (AGT) an enzyme critical in
repairing DNA damage induced by alkylating agents in formaldehyde-exposed mortuary students
previously shown to have increased micronuclei in both buccal cells and peripheral lymphocytes
(Suruda et al., 1993). AGT activity was lower in mortuary students with prior embalming
exposures versus students with no prior exposure (p=0.08). Seventeen of 23 students had lower
AGT activity after the 9 week course (p<0.05) with a larger proportion of naive students
demonstrating decreased activity (7 of 8) versus previously exposed students (10 of 15).
Although detailed exposure measurements were taken for each student, the changes in AGT
activity were not correlated to cumulative exposure (ppm-hrs).
4.5.3.1.3 Protein to protein cross-links
Formaldehyde is a reactive molecule that is likely to interact with both low molecular
weight cellular components (e.g., reduced glutathione [GSH]) as well as high molecular weight
cellular components. Unlike nuclear DNA, which has additional membrane barriers to exposure
(i.e., nucleus), extracellular and intracellular proteins, are obvious primary targets for interacting
with formaldehyde. Formaldehyde is a well-known cross-linking agent that is used in the
This document is a draft for review purposes only and does not constitute Agency policy.
4-431 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
fixation of tissues, inactivation of toxins and viruses (e.g. preparation of vaccines), and study of
protein-protein interactions (Metz et al., 2006). Using several identical synthetic polypeptides
differing on one amino acid, Metz et al (2004) have shown that formaldehyde initially reacts
with the primary amino and thiol groups of amino acids forming unstable methylol adducts,
which later are partially dehydrated forming labile Schiff bases that are capable of forming
crosslinks with other amino acid residues, such as arginine, asparagine, glutamine, histidine,
tryptophan, and tyrosine through methylene bridges, but not between two primary amino groups.
The same group (Metz et al 2006) has also shown that formaldehyde forms seven intramolecular
crosslinks in proteins with defined structure, such as insulin, involving arginine, tyrosine and
lysine and the N-terminus of insulin was converted to a imidazolidinone adducts similar to that
observed with the synthetic peptide (Metz et al 2004). (Figure 3-1 provides a general reaction
scheme for formaldehyde-mediated modifications of amino acids.)
4.5.3.1.4 Break-down of the mucociliary apparatus
The mucociliary apparatus of the upper respiratory tract is the first line of defense against
airborne toxicants. Comprised of a thick mucus layer (epiphase), hydrophase and ciliated
epithelium, the mucociliary apparatus may entrain, neutralize and remove particulates and
airborne chemicals from inspired air (Figure 4-4). Formaldehyde reacts with the components of
the mucous layer (proteins, glycoprotein, and lipids), crosslinking proteins. Formaldehyde
exposure induces slowing of the mucous flow, stiffing and breaking up of the mucous layer and
eventual mucostasis where gaps have been observed exposing the underlying hydrophase and
epithelium. Although ciliary beat first increases in response to formaldehyde exposure, perhaps
to compensate for reduced flow of the epiphase, ciliastasis ensues with both higher levels of
exposure, and increased duration of exposure. Altered ciliary has been noted in as little as 15
minutes of exposure (1.25 ppm) with functional deficits in the mucociliary apparatus at 30
minutes. Altered ciliary beat has been reported at the lowest concentration tested (0.5ppm) for a
single 6 hour exposure. Severity of effects increase with both duration and level of exposure
(see section 4.2.1.2.1).
4.5.3.1.5 Induced cell proliferation
There are several reports apparently demonstrating formaldehyde-induced proliferation in
cells below cytotoxic levels of exposure. This phenomenon has been reported from studies
involving both in vitro and in vivo exposures. Tyihak et al. (2001) demonstrated significantly
increased cell proliferation in both HT-29 human colon carcinoma and human umbilical vein
endothelial cell (HUVEC) lines treated with 0. ImM (the lowest dose) formaldehyde compared
This document is a draft for review purposes only and does not constitute Agency policy.
4-432 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
30
31
32
33
34
35
36
37
38
to untreated controls (P<0.0001). This effect was quantified as both an increase in cell number
over time (Figure 4-35), and an increase in the percentage of cells undergoing mitosis at each
time point. The authors also report a significant (P<0.01) inhibition of apoptosis in
formaldehyde-treated cells as compared to untreated cells (data not shown here). In a novel
system using xenotransplanted human tracheobronchial epithelial cells, formaldehyde was shown
to induce increased cell proliferation at doses below those required for a "massive toxic effect"
(Uraetal., 1989).
20-
18-
16-
14-
12-
10-
8-
6-
4-
2-
0-
A. HT-29 Cells
¦ Control
¦ 0.1 mM
DlmM
~ 10 mM
1
Day 1 Day 2 Day 3
Post-treatment
20
18
16
4- 14
0
r 12
1 10
C 8
u
B. HUVEC Cells
¦ Control
¦ 0.1 mM
~ 1 mM
DIP mM
Day 1 Day 2 Day 3
Post-treatment
Figure 4-35. Effect of various doses of formaldehyde on cell number in (A)
HT-29 human colon carcinoma cells and in (B) human umbilical vein
epithelial cells (HUVEC).
Values are average of three samples + SD; * P < 0.01 and ** P < 0.0001
compared to corresponding controls.
Source: Tyihak et al 2001.
Some animal studies have demonstrated increased cell proliferation after formaldehyde
exposures by both inhalation and ingestion (See section 4.2.1). However, whether sustained
increases in cell proliferation over baseline rates are observed upon exposure to sub-cytotoxic
doses of formaldehyde remains unclear. Several of the inhalation studies demonstrate increased
cell proliferation in the nasal epithelium at formaldehyde exposures levels that were sub-
cytotoxic—i.e. in the absence of significant cell death. Acute formaldehyde exposures (1 to 3
days) induced increased cell proliferation at discrete locations in the nasal mucosa, where cell
This document is a draft for review purposes only and does not constitute Agency policy.
4-433 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
proliferation was measured as a labeling index (percentage of cells pulse-labeled with tritiated-
thymidine. Reuzel et al. (1990) reported increased cell proliferation in the nasal passages
including the nasoturbinates, maxilloturbinates septum and lateral wall in male Wistar rats
exposed at 3 ppm, but not at 0.3 or 1 ppm, formaldehyde for 22 hours/day for 3 days. Zwart et
al. (1988) reported increased cell proliferation after exposure to 1 or 3 ppm formaldehyde, 6
hours/day for 3 days or 13 weeks in male and female albino Wistar rats. These increases were
transient at level 3 but sustained at level 2 of the nose and were not correlated with cytotoxicity.
In contrast, Wilmer et al. (1989), from the same group of investigators and using a similar
exposure regimen, reported no increase in cell proliferation after repeated 8-hour exposures at 1
or 2 ppm formaldehyde for 3 days or 4 weeks. Swenberg et al. (1986) demonstrated a transient
increase in cell proliferation after a single 8-hour exposure to 0.5 or 2 ppm formaldehyde in
male F344 rats but no increases after 3 days or repeated 8-hour exposures. The authors suggest
that adaptive responses of the nasal mucosa contribute to the transient nature of formaldehyde-
induced cell proliferation. After a series of acute studies at various formaldehyde concentrations,
Swenberg and coworkers concluded that, in addition to cell proliferation being concentration-,
dose- and time-dependent, the response varies by species and by location of exposure in the nose
(Swenberg et al., 1983, Swenberg et al, 1986).
Other methods of quantifying cell proliferation in the nasal mucosa have demonstrated
formaldehyde-induced cell proliferation at similar low exposure concentrations. Roemer et al.
(1992) measured cell proliferation by flow-cytometry in epithelial cells harvested from the nose
and trachea of male Sprauge-Dawley rats exposed to 2 ppm formaldehyde for 6 hours/day for 1
or 3 days and found increased cell proliferation after the 1-day exposure. These increases were
transient and were not evident after 3 days of exposure. Cassee and Feron (1994) identified
proliferating cells by staining for the presence of proliferating cell nuclear antigen (PCNA).
Formaldehyde exposure at 3.6 ppm for 6 consecutive periods of 12 hours (8-hour exposures
followed by 4-hour-periods of non-exposure) over three days, qualitatively increased the
expression of PCNA in respiratory epithelium at levels 2 and 3 of the nose in albino male Wistar
rats (Cassee and Feron, 1994). Hyperplasia, squamous metaplasia and frank necrosis were also
reported for these tissues.
Monticello et al. (1990, 1991, 1996) conducted in vivo cell proliferation studies in which
they exposed F344 rats for short durations (1, 4, 9 and 42 days) as well as much longer durations
(13, 26, 52 and 78 weeks) to exposure concentrations of 0, 0.7, 2.0, 6.0, 10.0 and 15.0 ppm.
These data are unique in that they also included low exposure concentrations. The authors
reported statistically significant increases in cell proliferation only at 6.0 ppm and higher
exposure concentrations in the short duration study and only at 10.0 ppm and higher
This document is a draft for review purposes only and does not constitute Agency policy.
4.434 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
concentrations in the longer duration study. These data have undergone considerable statistical
analysis in several papers as well as in this document. Conolly et al. (2002, 2003) and Gaylor and
Conolly (2004) interpreted these data, when combined, as indicating a non-monotonic behavior
at low dose. In other words, formaldehyde was judged to result in a reduction in cell proliferation
at low dose in comparison to baseline rates, with increased proliferation effect kicking in only at
exposures that were cytotoxic. However, as shown in Appendix 5-3 and in Subramaniam et al.
(2008), Crump et al. (2008), analysis of the individual animal data shows considerable
uncertainty and variability, both quantitative and qualitative, in the interpretation of these cell
proliferation data. (For example, even the control data vary over an order of magnitude in some
cases. See Figures 5-22 and 5-23 in Appendix 5-3.) These analyses (which were based on the
replicate animal data used in the above studies) considered regional formaldehyde dose to the
tissue (flux), nasal site and duration of exposure, as well as the number of cells at a given site.
The overall conclusion in Section 5.3.3 (and detailed in Appendix 5-3) is that the cell
proliferation dose-response at low dose could be reasonably described by both monotonic (with
and without a threshold) and non-monotonic curves.
Only one study, by Monticello et al. (1989), quantified cell proliferation in primates after
formaldehyde exposure; this study, reported an 18-fold increase in cell proliferation in the nasal
epithelium (respiratory and transitional), larynx, trachea and carina of male Rhesus monkeys
exposed to 6 ppm formaldehyde compared to controls (See section 4.2.1 for detailed study
description). The authors also noted that increased cell proliferation was seen in locations with
minimal histological changes, indicating proliferation may be a more sensitive predictor of
adverse health effects of formaldehyde exposure.
4.5.3.1.6 Cytolethality and resulting regenerative cell proliferation
The toxic and cytolethal effects of formaldehyde exposure at the POE are well
documented after both inhalation and oral exposures (See Section 4.2.1). The nature and
progression of tissue injury has been best documented in rodent inhalation assays. Early effects
on the nasal mucosa include altered ciliary beat and mucus flow, hyperplasia and metaplasia of
nasal epithelium (Morgan et al., 1986a; Morgan et al., 1986b; Monteiro-Riviere and Popp, 1986;
Maronpot et al 1986; Rusch et al., 1983 and Monticello et al., 1986). These first changes may be
considered adaptive responses. Squamous epithelium may thicken and transitional epithelium
may change to squamous epithelium as evidenced by squamous hyperplasia, squamous
metaplasia and thickening of the epithelium in these anterior portions of the nose. Tissue
damage may be transient at lower formaldehyde exposures as these changes serve to protect
tissue from formaldehyde's reactivity. However, higher formaldehyde concentrations can
This document is a draft for review purposes only and does not constitute Agency policy.
4-43 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
overwhelm these adaptive responses and result in gross tissue damage. Frank necrosis and focal
erosions have been reported in time- and concentration-dependent manner in rodent bioassays.
Both adaptive changes and cytolethality are associated with cell proliferation. However,
where adaptive changes are successful, e.g. prevent continued toxic insult to the tissue, cell
proliferation is transient. Exposure regimens where the adaptive changes are not adequate to
protect the tissue, would result in continued cytoxicity and cell death. Sustained damage to the
epithelium would result in sustained cell proliferation to compensate for cell death. A series of
rodent bioassays present convincing evidence that chronic inhalation exposures 6 hours a day, 5
days a week at 6, 10 and 15 ppm formaldehyde do result in sustained damage to the nasal
epithelium, sustained cell proliferation and tumor development (Kerns et al., 1983, Morgan et al.,
1986; Monticello et al., 1990; Monticello et al., 1991 and Monticello et al., 1996). Work by
Monticello and coworkers demonstrate that chronic repeated inhalation exposures at 6, 10 or 15
ppm formaldehyde result in sustained cell proliferation at the lateral meatus, mid-septum and
maxilloturbinates of rat nasal passages (Monticello et al., 1991 and Monticello et al., 1996).
4.5.3.1.7 Evidence for promotion
There is some evidence, although mixed, that formaldehyde may promote tumor
development by other carcinogens, and known initiating agents by various routes of exposure.
Formaldehyde exposure in drinking water (0.5% formalin) increased glandular stomach
adenocarcinomas in male Wistar rats after initiation with 100 mg/L, N-methyl-N'-
nitrosoguanidine (MNNG), compared to MNNG-only-treated rats (Takahashi et al., 1986). In
white non-inbred rats, inhalation exposures (3, 30 or 300 ug/m3 formaldehyde 7hr/day, 5
days/week for 1 year) increased tumor multiplicity per animal and decreased latency of
benzo[a]pyrene induced tumors in white non-inbred rats (Yanysheva et al., 1998.) Similarly,
formaldehyde skin application decreased tumor latency, in 7,12-dimethylbenz(a)anthacene
(DMBA) initiated hairless Oslo mice (Iversen, 1986). Although formaldehyde exposure also
increased the tumor multiplicity in Syrian golden hamsters where diethylnitrosamine (DEN)
(0.25mg I.P.) was the tumor initiator, positive results were only reported for the exposure
regimen where hamsters were exposed to formaldehyde via inhalation 48 hours prior to DEN
injection, and then one a week thereafter for life. However, formaldehyde did not increase the
number of tumors per tumor bearing animals when only administered beginning one week after
all DEN injections. In contrast, bladder cancer was not enhanced by intravesical instillation of
0.5ml of 0.3% formalin, one week after instillation of N-methyl-N-nitrosourea (MNU) in male
Fisher rats (Homma et al., 1986).
This document is a draft for review purposes only and does not constitute Agency policy.
4-436 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
The observed promotion activity of formaldehyde has been tested in several systems, by
different routes of exposure. By several routes of exposure, formaldehyde enhanced tumor
development at a site where formaldehyde did not induce tumors alone, without the initiating
agent (Takahashi et al., 1986, Yanysheva et al., 1998 and Iversen, 1986). Promotion activity in
these studies was evidenced by increased in tumor bearing animals (oral route), increase in
tumors per animal (inhalation routes) and decreased tumor latency compared to those animals
only exposed to the initiating agent (inhalation route) (Takahashi et al., 1986, Yanysheva et al.,
1998 and Iversen, 1986). Although these experiments do not indicate how formaldehyde acts as
a promoter in these systems, it is possible formaldehyde-induced mutation, increased cell
proliferation or other toxic action could enhance tumor development from another agent.
4.5.3.1.8 Localized Immunosuppression
Formaldehyde exposure has induced localized immune suppression in experimental
animals (Dean et al., 1984) and in exposed workers (Lyapina et al., 2004). Repeated inhalation
exposures in rodents depopulated the URT and pulmonary tissues of resident macrophages,
resulting in a transient decrease in POE host defenses (Admas et al., 1987). After cessation of
exposure, the mononuclear phagocyte (MP) populations were replenished and there was a
subsequent increase in host defense representing both increased MP numbers and increased
bacteriocidal activity of the MPs. These data suggest that peak exposures of formaldehyde may
present localized immunosuppression for components of the mononuclear phagocyte system
(MPS) in tissues at the site of first contact.
A number of studies have evaluated the ability of formaldehyde to induce systemic
immunotoxic effects in humans (Ohtani et al., 2004a, b; Erdei et al., 2003; Thrasher et al., 1990,
1987; Pross et al., 1987). Some studies have reported altered innate immune responses
associated with formaldehyde exposure (Erdei et al., 2003), while others have noted adaptive
immune response suppression associated with formaldehyde exposure (Thrasher et al., 1990,
1987) and changes associated with alterations to a predominant T—lymphocyte helper 2 (Th2)
pattern (Ohtani et al., 2004a, b). In contrast, Pross et al. (1987) did not observe formaldehyde-
associated changes in systemic immune function.
Numerous studies have reported increased respiratory tract infections in formaldehyde
exposed individuals both in occupational and residential environments (Lyapina et al., 2004;
Krzyzanowski et al., 1990; Holness and Nethercott, 1989). Incidences of physician-diagnosed
chronic bronchitis were more prevalent in children (under age 15) living in homes with higher
formaldehyde (>60 ppb) readings in the kitchen (p < 0.001) but this effect was more pronounced
(p < 0.001) in children simultaneously exposed to environmental tobacco smoke (Kryzanowski
This document is a draft for review purposes only and does not constitute Agency policy.
4-437 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
et al., 1990). The prevalence of chronic cough was also increased in adults living in homes with
measurable levels of formaldehyde, but data were not shown. Holness and Nethercott (1989)
assessed chronic bronchitis in 87 funeral workers, where the average formaldehyde exposure was
reported at 0.38 ± 0.19 ppm. Chronic bronchitis was observed in 20 funeral workers (n = 87)
exposed to formaldehyde compared with 3 cases of chronic bronchitis in nonexposed referent
controls (n = 38). A statistically significant association of self-reported chronic bronchitis and
decreased resistance to URT infection was reported in formaldehyde exposed workers compared
with controls (p = 0.02) (Lyapina et al., 2004). Of the workers, 41% had a history of chronic
respiratory infection and frequent long-lasting infectious inflammatory relapses (group la).
Another group (group lb) consisted of 17 exposed workers, 12 of whom had no history of
recurrent viral infections of the URT. There was a statistically significant association of
frequency and duration of inflammatory relapses between groups la and lb.
Lyapina et al. (2004) also reported effects of formaldehyde exposure on neutrophil
respiratory burst activity (NRBA), the capacity of polymorphonuclear leukocytes to produce
reactive oxygen radicals in response to chemical or microbial stimuli using flow cytometry. A
suite of hematological tests and flow cytometric analysis for respiratory burst activity were
performed. Although no significant difference was observed in the spontaneous and stimulated
NRBA (median percentage of oxidizing cells) between the 29 exposed workers with URT
inflammation and the healthy controls (0.83 versus 1.35, respectively), a separate comparison of
the NRBA of 12 workers with chronic, repeating URT infections and 17 workers with short,
infrequent episodes of URT inflammations was significant (0.45 versus 1.00 ,p = 0.037). When
the NRBA of the group with chronic URT infections (n = 12) was separately compared with that
of the healthy controls (n = 21), the results were also significant (0.45 versus 1.35 ,p = 0.012).
Individuals with chronic URT infections have reduced NRBA that could be due to formaldehyde
exposure. Neutrophils respond to tissue damage or local invasion of microorganisms and act to
phagocytize foreign cells. If neutrophilic activity is hampered or altered by formaldehyde
exposure, then the ability to fight infection will be diminished, leading to prolonged infection.
However, no dose-response pattern of formaldehyde exposure could be determined from this
study.
4.5.3.1.9 Potential for systemic transport of formaldehyde
In aqueous solution formaldehyde exists in equilibrium with it's hydrated form
methanediol (CH2OH2) (Kd = 5.5xl0-4). The equilibrium favors methanediol at physiological
temperature and pH (>99.9%) and is readily reversible. In biological systems, as free
formaldehyde is removed from aqueous solution through binding with serum proteins and
This document is a draft for review purposes only and does not constitute Agency policy.
4-43 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
cellular components, the equilibrium is reestablished by dehydration of methanediol to free
formaldehyde. The reversible nature of this hydration reaction describes how a pool of free
formaldehyde may be sustained in biological systems.
There is strong and consistent evidence in biological testing systems in vitro that treating
cells with formaldehyde in an aqueous media results in significant cytoxicity, cell proliferation,
clastogenic effects and clear evidence of mutational events (Section 4.3). Similarly, animal
bioassays where formaldehyde is administered in drinking water report portal of entry toxicity
including hyperplasia, increased cell proliferation, focal lesions and tumors (Section 4.2.1). It
should be noted that URT tissues are covered by an aqueous mucous layer, through which
formaldehyde must pass to react the cellular components of the URT. It has been postulated that
formaldehyde transports through this mucous layer and the underlying tissues as methanediol
(Georgieva et al., 2003).
The dynamic equilibrium between the hydrated and unhydrated forms of formaldehyde in
biological systems is well understood. Since the hydration reaction favors methanediol, it is
expected that exogenous formaldehyde which reaches the blood will primarily exist as
methanediol and is subject to physiological elimination. As free, unhydrated formaldehyde
continues to react with serum proteins and cellular components, the blood levels of methanediol
are expected to reduce as it is dehydrated to maintain equilibrium. Although some attempts to
measure significant changes in free formaldehyde levels in blood after inhalation exposure have
not been successful, the half-life in blood has been measured after i.v. injection at approximately
2 minutes (McMartin et al., 1979). Additionally, the detection of antibodies to formaldehyde-
hemoglobin adducts and formaldehyde-albumin adducts in exposures workers, smokers and
laboratory animals exposed via inhalation provides direct evidence that formaldehyde is able to
react with serum albumin and hemoglobin in biological systems (Thrasher et al., 1990, Grammer
et al., 1990; Grammer et al., 1993; Dykewicz et al., 1991; Varro et al., 1997 and Li et al., 2007).
These data support the hypothesis that exogenous formaldehyde may reach and transport through
the blood. If so, formaldehyde (or methanediol) may reach sites distal to the portal of entry.
4.5.3.2 Mode of Action Evaluation for Upper Respiratory Tract Cancer (Nasopharyngeal
Cancer, Sino-nasal)
From the above discussion, it can be seen that numerous mechanisms of action for
formaldehyde-induced cancer can be reasonably supported based on various known biological
actions of formaldehyde (e.g., mutation, cell proliferation, cytotoxicity, and regenerative cell
proliferation). Additionally, alternative actions, such as immunosuppression or viral
reactivation, are possible, although less data exist to evaluate these MO As. Rather than a single
This document is a draft for review purposes only and does not constitute Agency policy.
4-439 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
22
23
24
25
26
27
28
29
30
31
32
33
34
35
MO A, it is plausible that a combination of these factors contribute to cancer incidence in an
exposed population. Considering multiple factors may help to better understand the biological
and mechanistic basis for the increases in cancer incidence observed in exposed human
populations. Unlike animal bioassays, human epidemiological studies may reflect not only the
effects of the agent of concern but also numerous other risk factors (e.g., viral status, diet,
smoking, etc.). Additionally, human studies may be impacted by biological human variability
across individuals, cancer biology (sub-types), wide variability in exposure regimens in human
populations, etc. Therefore, if the purposes of exploring the carcinogenic MOA of an agent are
to better understand the relevance of a given carcinogen to human populations and to inform the
exposure-response analysis, then discussions of MO As which recognize the interaction of an
agent with human variability and various risk factors is an appropriate analysis.
a) Direct mutagenicity of formaldehyde in cells at the site of first contact: Mutations,
the permanent heritable changes in the genome of the cell, are a primary mechanism for
the activation of oncogenes or the inactivation of tumor suppressor genes. Mutagenicity
is the most widely recognized determinant of chemical-induced carcinogenicity, and it is
difficult to set aside the relevance of direct formaldehyde-induced mutations from its
demonstrated carcinogenicity. Formaldehyde-induced mutation in mucosal cells of the
URT, throat and buccal cavity may serve to initiate cells, or provide subsequent mutageic
events to already initiated cells. Since the mucosal cells have proliferative capacity, and
cell proliferation is a normal tissue function, mutations may be fixed and passed to
daughter cells due to baseline cell proliferation of the tissue.
Relevance to humans: This MOA is relevant to humans. The well-documented DNA
reactivity (e.g. DPC and DNA adducts) and clastogenicity of formaldehyde in the URT of
laboratory animals is a direct effect of formaldehyde on tissues of first contact. As this is
a direct acting agent - no distribution or metabolism is required for the genotoxic action -
there is little expected species variability. As discussed in chapter 3, there are species
differences in flux of formaldehyde into the respiratory mucosal tissues, but this
introduces species differences in dosimetry - not mechanism. Finally, the clastogenic
effects in nasal and buccal epithelial cells in formaldehyde- exposed workers confirms
the direct genotoxic effects of formaldehyde at the first site of contact in humans.
b) Decrease in DNA repair function within cells at the site of first contact: A decrease
in DNA repair capacity in these tissues by formaldehyde may increase total mutations
over time due to either endogenous or exogenous sources of mutation. Although there
This document is a draft for review purposes only and does not constitute Agency policy.
4-440 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
are only a few studies which have explored the potential for formaldehyde to reduce
DNA repair capacity, the evidence is positive, both in vitro testing systems, and in one
study of occupationally exposed humans (Grafstrom 1985; Hayes et al., 1997).
Relevance to humans: This MOA is considered relevant to humans. The general
population is exposed to various carcinogens, many with mutagenic potential, at sites of
first contact including; air pollution, tobacco products, nitrosamines and viruses.
Additionally, there are endogenous sources of DNA damage and mutagenicity in humans
(e.g. lipid peroxidation, oxidative stress). The demonstration of reduced DNA repair
activity (06-alkylguanine-DNA alkyltransferase activity) in formaldehyde-exposed
mortuary students suggests this toxic action of formaldehyde is possible in humans.
c) Formaldehyde-induced cell proliferation: Formaldehyde-induced cell proliferation in
the oral and respiratory mucosa may be considered a key event in conjunction with the
genotoxic effects, and induced mutational events observed with formaldehyde exposure.
This MOA is intended to describe events which occur below exposure levels which
induce cell death and mucosal lesions. Therefore this MOA is comprised of two key
events:
a. Formaldehyde-induced genotoxicity or mutation
b. Formaldehyde-induced cell proliferation
DNA replication during cell proliferation may serve to translate DNA damage or a
formaldehyde-related DNA lesion into a permanent change in the sequence of nucleic
acids during replication of the DNA- e.g. 'fix' a mutation from DNA damage.
Additionally formaldehyde-induced cell proliferation may provide an opportunity for
initiated cells to proliferate, increasing the potential for additional mutation events and
transformation. The increased cell proliferation observed in the mucosal tissues in direct
contact with formaldehyde during inhalation exposures may serve to amplify the risk of
cell transformation from mutation alone. Researchers have noted that increased cell
proliferation may be transient in some locations as adaptive responses compensate
(Swenberg 1983). However, evidence in both monkeys and rodents indicate that
increased cell proliferation in repeated exposures across time do result in sustained cell
proliferation. Data in Rhesus monkeys indicates increased cell proliferation is observed
beyond the nasal cavities to the larynx, trachea and carnia (first tracheal
branching)(Monticello et al., 1989). Additionally, the authors note that cell proliferation
This document is a draft for review purposes only and does not constitute Agency policy.
4.441 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
is a more sensitive indicator of effects on the epithelium, observed even when minimal
histological changes were present.
Human Relevance: Both formaldehyde-induced mutation and cell proliferation are
direct effects on the oral and nasal mucosa, well documented in rodent models with
supporting evidence in human epidemiological studies. Therefore both key events are
relevant to humans. As noted above, there are species differences in localized flux of
formaldehyde into the tissues of the oral and respiratory tract based on structural
differences in the airways, as well as breathing patterns. Although these differences may
effects the dosimetry of the formaldehyde absorption into the tissues, this only influences
the magnitude of response at any given location. Data from exposed Rhesus monkeys
which documents formaldehyde-induced cell proliferation in tissues beyond the nasal
cavity, and tissues with minimal histological changes supports a role for cell-proliferation
in the observed cancers in humans, which occur beyond the nasal cavities, and in tissues
without formaldehyde-related focal lesions.
d) Cytotoxicity-induced cell proliferation (CICP): Cell death followed by compensatory
cell proliferation is a reasonable MOA for agent-induced cancer. It should be noted that
the exposure conditions which result in CICP in rodents is known to result in significant
DNA reactivity and genotoxicity. Therefore, formaldehyde-induced mutations cannot be
excluded from this MOA. The animal bioassays support the carcinogenic potential of
formaldehyde in this context (Kerns et al., 1983; Selkemur et al, 1983 and Monticello et
al 1986). The majority of squamous cell carcinomas (SCCs) seen in formaldehyde-
exposed rats have been localized to the lateral meatus and mid-septum in the nasal
passages (Morgan et al., 1986; Monticello et al., 1996), while polyploid adenomas have
predominantly been reported at the maxilloturbinates (Morgan et al., 1986; Monticello et
al., 1996). Morgan et al. (1986) speculated that the maxilloturbinate was less susceptible
to SCC due to metabolic differences. However, Monticello et al. (1996) later suggested
that the smaller population of cells available at the maxilloturbinate accounted for fewer
SCCs observed at that site. Regardless, for those locations where SCCs do arise in rats
chronically exposed to formaldehyde, a clear temporal relationship can be demonstrated
for dose regimens capable of producing sustained epithelial damage and sustained cell
proliferation to eventual tumor development. Conversely, tumors are not observed in
these rodent models at those sites in the nasal passages without sustained cell
proliferation.
This document is a draft for review purposes only and does not constitute Agency policy.
4-442 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
Relevance to humans: Human exposure to formaldehyde would most likely involve
chronic exposures to indoor levels of formaldehyde, and episodic exposures in the
environment or from an occupational exposure (See review in Chapter 2). An exposure
scenario parallel to that used in chronic rodent bioassays is unlikely (e.g. 2-15 ppm 6-8
hours/day, 5 days/week for 10-30 months). Exposure conditions are difficult to assess
especially in retrospective studies. However, only the most extreme industrial work
conditions would result in human exposures similar to those that produce sustained
compensatory cell proliferation in animal studies (i.e. 6-15 ppm 6 hrs/day, 5 days per
week). Gross tissue lesions as reported in rodents from repeated chronic exposures at 6
and 10 ppm formaldehyde have not been reported from workplace exposure, and only
minor histopathological changes have been noted (Boysen et al., 1990 and Holmstrom
and Wilhelmsson et al., 1989). It is possible that workers were episodically exposed to
formaldehyde levels which resulted in cell death and focal or gross lesions requiring cell
proliferation for tissue remodeling or repair. However, it is unexpected that these
conditions would be relevant to human environmental exposures. Therefore, although
regenerative cell proliferation is retained as a reasonable MOA for formaldehyde
carcinogenicity in experimental animals, it is unclear whether it is relevant to the
extrapolation of health risks to formaldehyde exposures in the general environment.
e) Promotion: Several animal studies indicate that formaldehyde exposure may promote
tumor formation due to other carcinogenic or initiating agents. There are positive data
by several routes of exposure (oral, dermal and inhalation) and promotion has been
reported as an increase in tumor bearing animals, an increase in tumors multiplicity or a
decrease in tumor latency with formaldehyde exposure in conjunction with the initiating
agent compared to tumors from the initiating agent along, or formaldehyde alone. The
specific key events which may explain this promotion effect are unknown but may
include several of the mechanisms discussed as potential MO As for formaldehyde:
mutagenicity, mitogenesis, co-carcinogenicity, immunosuppression. Promotion is
considered here as a separate MOA, since these activities are noted for experimental
conditions and tumor sites where formaldehyde did not induce tumors in the absence of
the initiating agent.
Relevance to humans: Although the human epidemiologic literature doesn't address
issues of tumor promotion, the nature of the cancers of concern indicate that chemical
promotion may be relevant to cancer incidence for these sites. Many of the risk factors
for NPC and other mouth and oral and URT cancers include direct mutagens (e.g.
This document is a draft for review purposes only and does not constitute Agency policy.
4.443 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
smoking, dietary nitrosamines) where a promoting agent would be expected to increase
cancer incidence with these other risk factors. Additionally, the well known viral risk
factors for cancers of the mouth and URT also suggest a role for promoting agents to
human cancer incidence. Although only tangential evidence, this does suggest that the
promoting activity of a chemical agent, would be relevant to the agent's carcinogenicity
at these sites. Therefore, the potential for formaldehyde to act as a promoter with other
initiators - is considered relevant to formaldehyde's carcinogenic MOA.
f) Increased URT infections / viral reactivation: Inhalation exposure to formaldehyde
has been shown to decrease the defenses of the body against infection through two
mechanisms: 1) damage to the protective mucous barrier and function of the mucociliary
apparatus; and 2) localized immunosuppression. These effects have been demonstrated
in both exposed humans and controlled animal experiments. Additionally, increased
respiratory tract infections are associated with formaldehyde exposure in several
populations. Common viral agents (e.g. Epstein barr virus) are known risk factors for
NPC, sinonasal cancers and other URT cancers. Although direct evidence does support
increased URT infections due to formaldehyde exposure, and URT infections are
considered risk factors for URT cancers, direct evidence for formaldehyde-related
infections leading to cancer is lacking. There is however one epidemiological study
which finds the association between formaldehyde and NPC is strengthened in Epstein
barr virus sero-positive cases versus sero-negative cases. These data suggest a possible
role for formaldehyde in infection, viral reactivation, or co-carcinogenicity with a viral
agent.
Relevance to humans: The potential role of increased URT infections and
immunosuppression at the portal of entry is considered to relevant to humans. Data in
humans are available to support both key events in this MOA. Additionally,
epidemiological studies are conducted in human populations where individuals may be
exposed to various viral agents across the study period. Therefore, toxic actions by
formaldehyde which may increase URT infections, or viral-reactivation at the site of first
contact, could be acting in conjunction with viral agents to contribute, in part, to observed
associations between formaldehyde exposure and increased URT cancer.
Summary and integration of key events:
Each of the hypothesized MO As discussed above to better understand the carcinogenic
potential of formaldehyde is supported by formaldehyde-specific evidence, either in animal
This document is a draft for review purposes only and does not constitute Agency policy.
4.444 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
studies, human studies or both. For those key events studied in animal models such as cell
proliferation, genotoxicity, degradation of the mucociliary apparatus and CICP, supporting
evidence is available in more than one species, multiple strains (e.g. rats) and has been reported
by multiple researchers. Therefore the overall database supporting these key events in laboratory
studies, and their corresponding MO As is fairly large. In contrast, some key events relevant to
humans, but less studied in animal models may have a small supporting database (e.g. increased
respiratory tract infections). These alternative MO As are retained as potentially relevant to the
carcinogenic action of formaldehyde as the intent of this discussion is to identify modes of action
will may contribute to the observation of increased upper respiratory tract cancers in exposed
human populations. It is noted that additional study is needed to better understand the range of
effects formaldehyde may have at sites of first contact in humans.
The MO As which include genotoxicity, mutation, decreased DNA repair, increased cell
proliferation and CICP are interrelated. Conditions which provide both a source of cell
proliferation and increased mutation would be expected to increase neoplastic transformation.
Formaldehyde acts on the target tissue, the respiratory epithelium, to induce each of these events.
However, these key events operate across different exposure ranges and present different
exposure response relationships. For example, formaldehyde-induced mutations would be
expected across the exposure range, where any incremental increase in genotoxicity and
formaldehyde-related mutation would contribute to background levels, with the potential to
increase cancer risk incrementally. In contrast, focal and gross lesions to the respiratory mucosa
due to cytolethality are not observed unless exposure concentrations are sufficient to provide
localized tissue doses (flux) required to result in cell death and related compensatory cell
proliferation. Since tissue dose (flux) is dependent on not only exposure concentration but also
duration of exposure and location in the respiratory tract (Section 3.4), and varies by species,
correlation of exposure concentrations to tissue responses directly are complex. Exposure
response relationships for the key events (cell proliferation, genotoxicity, degradation of the
mucociliary apparatus and CICP) are reported by exposure concentration, not tissue flux, which
would be a more biologically relevant measure.
Although the tissue dose-response relationships for formaldehyde induced mutation,
mitogenesis and cytolethality are different, the effects at the tissue level cannot be easily
disaggregated. At any given exposure concentration, target cells in the respiratory tract will
experience different effective tissue concentrations of formaldehyde. Measurement of cell
proliferation, DNA protein crosslinks or genotoxicity may require examining a population of
cells which would have been subject to different flux rates of formaldehyde (See chapter 3).
Similarly, when evaluating the tumor dose response, cells within the target tissue will represent a
This document is a draft for review purposes only and does not constitute Agency policy.
4.445 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
range of target tissue formaldehyde concentrations. Therefore, an integrated MOA scheme is
hypothesized where key events may influence the observed tumor response differentially across
the exposure response range (Figure 4-36). This schematic illustrates the potential for
genotoxicity and formaldehyde-induced mutation to occur where tissue dose (flux of
formaldehyde into the tissue is minimal). Where tissue dose is increased, formaldehyde-induced
cell proliferation is observed in addition to genotoxicity. As tissue dose increases and
formaldehyde effects on the respiratory mucosa are more severe, gross pathology including focal
and gross lesions due to cell death are noted. Therefore, several of the MO As presented above
may be operative and relevant to human exposures at exposure levels resulting in minimal tissue
flux - a) direct formaldehyde genitoxicity and resulting mutation, b) inhibition of DNA repair
and c) formaldehyde induced cell proliferation in conjunction with mutation. CICP, which
involves localized and gross tissue lesions would be operative at higher exposure levels. There is
little data to inform the dose range over which the remaining hypothesized MO As may operate
(promotion and increased respiratory tract infections/viral action).
Integrated Dose Response Curve forFormaldehyde
i i
Tumor
Response
Exposure Concentration
Mutagenesis I I
Basal Cell Proliferation I I
Induced mitogenesis >
Adaptive Cell Proliferation ^
Regenerative Proliferation ^
Figure 4-36: Integrated MOA scheme for respiratory tract tumors
4.5.3.3 Mode(s) of Action for Lymphohematopoietic Malignancies
4.5.3.3.1 MOA evaluation for Leukemia
Leukemia may arise from stem cells and progenitor cells in the bone marrow (e.g. acute
and chronic myeloid leukemia) or from mature lymphocytes (e.g. chronic lymphatic leukemia,
This document is a draft for review purposes only and does not constitute Agency policy.
4-446 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
hairy cell leukemia) (Figure 4-33, Section 4.5.2). Although there is a consistent association
between formaldehyde exposure and forms of leukemia when considered as group of diseases
(Table 4-91, Section 4.5.2), the strongest and most consistent associations are seen specifically
with myeloid leukemia. Little evidence supports an association between formaldehyde exposure
and other specific leukemia subtypes, although two studies support a strong association between
formaldehyde and "other leukemia and unspecified leukemia (ICD-9 code 207). Therefore, this
MOA evaluation will focus on mechanisms which may impact all forms of leukemia (e.g. bone
marrow toxicity) or those specific to myeloid leukemia. The mechanistic data supporting the key
events in this analysis are presented in section 4.5.3.1.
a) Direct effects of formaldehyde on a circulating stem cell or progenitor cell present at
the portal of entry: Hematopoietic stem cells do circulate throughout the body and can
be harvested from peripheral blood. Formaldehyde exhibits a range of toxic effects at the
site of first contact including genotoxic effects believed to be mediated by direct DNA
reactivity (Section 4.3). Formaldehyde is known to directly react with blood components
in formaldehyde exposed humans and animals resulting in both hemoglobin and albumin
adducts (Thrasher et al., 1990, Grammer et al., 1990; Grammer et al., 1993; Dykewicz et
al., 1991; Varro et al., 1997 and Li et al., 2007). Therefore, it has been hypothesized that
formaldehyde could react with DNA in circulating hematopoietic stem cells (Zhang et al.,
2009) resulting in heritable mutations which may contribute to leukemia incidence.
Recently Zhang et al. (2010) have tested the hypothesis that exogenous formaldehyde
may damage circulating stem cells. Clastogenic effects were found in circulating
hematopoietic stem cells cultured from formaldehyde exposed workers. The reported
aneuploidy was demonstrated as significant increases in both monosomy 7 and trisomy 8.
These specific chromosomal changes are consistent with those reported for agent-induced
myeloid leukemia (Zhang et al., 2010).
Relevance to Humans: This hypothesized MOA is considered relevant to humans.
Supporting evidence is found in humans for formaldehyde direct reactivity with blood
proteins (e.g. albumin and hemoglobin) as well as clastogenic effects in circulating
hematopoietic stem cells in formaldehyde exposed workers.
b) Bone marrow toxicity: Direct bone marrow toxicity is the most studied leukemogenic
action for an endogenous agent (e.g. benzene, ionizing radiation). It is believed that an
agent which exerts its toxicity on the bone marrow, resulting in translocations and
heritable mutations in hematopoietic stem cells may cause leukemia. It has been
This document is a draft for review purposes only and does not constitute Agency policy.
4.447 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
hypothesized that formaldehyde may transport to the bone marrow in its hydrated form
(methandiol) and react with cellular proteins, and DNA casuing direct effects on
components of the bone marrow. Pancytopenia (a reduction in blood borne cells formed
in the bone marrow) is a symptom of direct bone marrow toxicity and is observed with
other leukemogenic agents (e.g. benzene, ionizing radiation). A recent review of 8
published studies of formaldehyde exposed workers in China by Tang et al. (2009)
indicates 7 of the studies provide evidence of reduced white blood cell counts, platelet
levels and hemoglobin levels associated with formaldehyde exposure. A study of
occupationally exposed nurses provided a correlation between decreased white blood cell
counts and formaldehyde exposure (Kuo et al., 1997). A recent study by Zhang et al.
(2010) provides the best evidence for bone marrow toxicity, where they report not only a
reduction in white blood cell counts, but reductions in cell counts of all the blood cells, as
well as increased mean cell volume. Although these reductions did not meet the clinical
definition of pancytopenia (when averaged across the study population), reduction of all
blood borne cells formed in the bone marrow is consistent with the bone marrow toxicity
associated with pancytopenia seen with other leukemogens (Zhang et al., 2010).
Relevance to Humas: This hypothesized MOA is considered relevant to humans.
Supporting evidence is found in humans for bone marrow toxicity in formaldehyde
exposed workers.
4.5.3.3.2 MOA evaluation for Lymphomas (e.g. Hodgkin lymphoma, Multiple myeloma)
The general MOA for formaldehyde is based on direct chemical reactivity and toxic
effects at the portal of entry (POE). Formaldehyde is directly and indirectly genotoxic, and
reacts with cellular proteins and DNA in cells which it comes into contact. Additionally,
immunosuppression, viral reactivation and promotion effects are relevant to lymphoma and
related malignancies. Therefore, the key events for the adult cell lymphoid cancers would
include these actions. Lymphoid tumors (e.g. lymphocytic leukemia, B-cell lymphoma, mantle
cell lymphoma [a rare form on non-Hodgkin lymphoma] and myeloma) may arise from cells
present at the portal of entry (POE) (Figure 4-33). The location and function of mature
lymphocytes contribute to their vulnerability to transformation by agents at the POE. Therefore,
a brief summary of the immuno-biology of these cells is provided in order to provide context for
the MOA evaluation:
Location: Lymphocytes are present in the oral and respiratory tract epithelium, as well
as in cell aggregates and tertiary immune structures (e.g. germinal centers) in the mucosal
This document is a draft for review purposes only and does not constitute Agency policy.
4-448 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
tissues (Zuercher and Cebra, 2002; Zuercher et al., 2002; Wu et al 1997 and Kupper et
al., 1990). These mucosa-associated lymphoid tissues (MALT) provide the opportunity
for formaldehyde to directly interact with components of the immune system present at
the POE (Wu et al., 1997, Claeys et al. 1996, Park et al., 2003 and Fujimura 2000).
Intraepithelial lymphocytes are present in the pseudostratified epithelium of the
nasopharyngeal passages and there are aggregates of immune cells and germinal cells
present in these tissues. Crypts containing mature lymphocytes exist at the surface of the
nasal epithelium (Fujimura 2000). Microfold cells or M-cells form the crypts, where the
lymphocytes are covered by a thin membrane (Figure 4-37). Functionally, these
lymphocytes identify and process foreign antigens at the POE (Fujimura 2000).
Therefore the mature lymphocytes within these crypts, exposed to exogenous agents, are
involved in active immune responses to foreign antigens.
Clonal Expansion: Mature lymphocytes (both B and T-cells) clonally expand their
populations in response to an exogenous antigen when a humoral immune response is
stimulated. Therefore cell proliferation is a normal function of these mature lymphocytes
and occurs every time there is an infection. Cell proliferation of mature B and T-cells,
responsive to a particular antigen, occurs in active germinal centers (including those
within the respiratory tract). Cells may be exposed to exogenous agents during the
immune response, or cells responding in the germinal center may have previously been in
the epithelium or M-cell crypt.
Somatic Hypermutation: Normal immune function includes the process of somatic
hypermutation where B-cells undergo DNA rearrangement of the variable region genes to
produce novel antibodies specific to a given antigen. This process is key to adaptive
immunity and demonstrated by the basic principles of immuno-biology which underlie
vaccination theory. Gene sequencing of adult B-cell lymphomas and leukemias indicate
that the chromosomal regions involved in somatic hypermutation correspond to known
oncogenes in these cancers. The vulnerability of these processes is evidenced by the
observation that approximately 90-95% of adult lymphomas and leukemias are of B-cell
origin (Gordon et al., 2003). Formaldehyde-induced protein-protein crosslinking could
disrupt cell processes including somatic hypermutation and cell mitosis, resulting in
agent-induced translations similar to those found in spontaneous B-cell malignancies.
This document is a draft for review purposes only and does not constitute Agency policy.
4.449 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Figure 4-37. Location of intra-epithelial lymphocytes along side epithelial cells in
the human adenoid. (Source Fujimura et al., 2000)
a) Direct or indirect formaldehyde-induced mutation in cells at the site of first contact:
Immune cells including intraepithelial lymphocytes, and cells in mucosal associated
lymph tissue (MALT) are collocated with the epithelial cells from which URT cancers
arise (Figure 4-37). Therefore the direct and indirect mutagenic potential for
formaldehyde is equally applicable to components of the immune system present in these
tissues. Mutations, the permanent heritable changes in the genome of the cell, are a
primary mechanism for the activation of oncogenes or the inactivation of tumor
suppressor genes. Mutagenicity is the most widely recognized determinant of chemical-
induced carcinogenicity, and it is difficult to set aside the relevance of direct
formaldehyde-induced mutations from its demonstrated carcinogenicity. Formaldehyde-
induced mutation in immune cells present at the site of first contact, may initiate cells or
or provide subsequent mutagenic events to already initiated cells. The competence of our
immune system relies on the proliferation of peripheral lymphocytes in response to
immune challenge (infection). Additionally, heritable changes to the variable gene
regions in B-cells generated during somatic hyper-mutation are essential to adaptive
immunity (e.g. immunization) demonstrating that permanent heritable changes in the
DNA of peripheral B-cells are passed to daughter cells and retained in the body for
decades. Any agent-induced mutations would be similarly propagated and retained with
the potential to contribute to the transformation of mature lymphocytes.
This document is a draft for review purposes only and does not constitute Agency policy.
4-450 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Relevance to humans: This MOA is relevant to humans. The well-documented DNA
reactivity (e.g. DPC and DNA adducts) and clastogenicity of formaldehyde at the POE in
laboratory animals is a direct effect of formaldehyde on tissues of first contact, and these
mechanisms are considered relevant to humans. As with epithelial cells, clastogenic
effects in peripheral lymphocytes are documented in formaldehyde-exposed students and
workers, confirming the genotoxic effects of formaldehyde in immune cells, from which
lymphomas and related diseases may arise (Section 4.5.2Figure 4-33).
b) Formaldehyde-induced protein-protein crosslinks may disrupt somatic-
hypermutation: Although not as well studied as DNA-protein crosslinks, formaldehyde
also formes crosslinks between amino acids on proteins (Section 4.5.3.1.3 for details).
Specific oncogenes for malignancies which arise from mature B-cells are linked to errors
in the process of somatic hyper-mutation (Greaves et al. 2004). If formaldehyde creates
protein crosslinks in competent B-cells which effect the process of DNA rearrangement,
formaldehyde may generate translocations and related oncogenes similar to those
observed in observed in spontaneous B-cell malignancies.
Relevance to humans: This hypothesis has not been tested in either exposed human or
animal test systems. However, the link between somatic-hypermnutation and B-cell
oncogenes is well established and perturbation of this process by an exogenous agent is a
reasonable extension of the current understanding of the etiology of B-cell malignancies.
c) Increased URT infections / viral reactivation: Inhalation exposure to formaldehyde
has been shown to decrease the defenses of the body against infection through two
mechanisms: 1) damage to the protective mucous barrier and function of the mucociliary
apparatus; and 2) localized immunosuppression (Section 4.5.3.1). These effects have
been demonstrated in both exposed humans and controlled animal experiments.
Additionally, increased respiratory tract infections are associated with formaldehyde
exposure in several populations. Common viral agents (e.g. Epstein barr virus) are
known risk factors for malignancies which arise from mature lymphocytes. Thus,
increased URT infections or viral reactivation due to formaldehyde exposure may
influence the incidence of these cancers.
Relevance to humans: The potential role of increased URT infections and
immunosuppression at the portal of entry is considered to be relevant to humans. Data in
humans are available to support both key events in this MOA. Additionally, co-exposure
to infectious agents (including viruses) would be expected in participants in an
This document is a draft for review purposes only and does not constitute Agency policy.
4-451 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
epidemiological study, suggesting an MOA which acted in conjunction with infectious
agents may be relevant to agent-induced cancer. Therefore, toxic actions by
formaldehyde which may increase URT infections, or viral-reactivation at the site of first
contact, could be acting in conjunction with viral agents to contribute, in part, to observed
associations between formaldehyde exposure and increased lymphoma and related
diseases.
4.5.3.3.3 Summary and evaluation of hypothesized MOA(s) for Lymphohematopoietic
Malignancies
The well-documented direct toxic action of formaldehyde on cells at the site of first
contact is a general effect based on the reactivity of formaldehyde with cellular components (e.g.
proteins and DNA) (Section 4.5.3.1). As a general effect, it is reasonable that these toxic effects
would be relevant to all cells which come into contact with formaldehyde. The current debate
regarding the biological plausibility of formaldehyde-induced lymphohematopoeitic
malignancies centers around a perspective that the diseases within this general grouping are
systemic cancers arising only out of bone marrow toxicity (Heck et al., 2006, Pyatt et al., 2008)
and that it is implausible for formaldehyde to induce bone marrow toxicity. The above MOA
evaluation expands the current debate by considering the impact of POE toxicity on elements of
the immune system and cancers might arise from these cells (Section 4.5.3.3.2) and by
presenting data which support the observation that formaldehyde is associated with bone marrow
toxicity and damage to circulating stem cells in exposed humans (Section 4.5.3.3.1).
As significant increases in free formaldehyde in peripheral blood from exogenous
exposure has not been detected (Heck et al., 1987), it has been hypothesized that formaldehyde
does not transport and therefore cannot exert toxic effects outside of the tissues at the site of first
contact (Heck et al., 2006, Pyatt et al., 2008). In contrast to this hypothesis, effects are seen in
formaldehyde-exposed humans which indicate systemic effects on the hematopoietic system
including reduced white blood cell counts, clastogenic effects in peripheral blood lymphocytes
and aneuploidy in circulating stem cells (Tang et al., 2008, Zhang et al., 2010 and Section
4.5.3.1). These observed effects in humans are consistent with agent-induced bone marrow
toxicity and are observed with other well-studied exogenous leukemogens (e.g. benzene and
ionizing radiation.) It is unknown if formaldehyde is distributed systemically to exert its effects
directly on cells in the bone marrow or if damage to circulating stem cells or progenitor cells
would be sufficient to result in the observed effects in humans (Zhang et al., 2010). Additional
research is needed to better determine the potential for systemic transport of formaldehyde
considering both detection of its hydrated form (methylene glycol) as well as formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
4-452 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
protein adducts (e.g. FA-GSH and FA-albumin). Similarly the results of Zhang et al. (2010)
need to be extended (analysis for additional chromosomal aberrations) and repeated. Although
further evidence is needed to better understand the hypothesized mechanisms for formaldehyde-
induced effects on hematopoietic stem cells, the observed hematotoxic effects in humans cannot
be set aside. Therefore, however unlikely, the current data support the biological plausibility of
formaldehyde effects on the hematopoietic system.
4.5.4. Hazard Characterization FOR Formaldehyde Carcinogenicity
Formaldehyde is Carcinogenic to Humans by the Inhalation Route of Exposure
Based on the results from a large and well-followed longitudinal cohort study of
industrial workers and several case-control studies, the epidemiologic evidence is sufficient to
characterize the association between formaldehyde exposure and nasopharyngeal cancer (NPC)
as causal in humans (Hauptmann et al., 2004; Hildesheim et al., 2001; Vaughan et al., 2000). As
a group, upper respiratory tracts sites of direct contact with formaldehyde upon inhalation (i.e.,
salivary gland, mouth, nasopharynx, nasal cavity and larynx) also showed sufficient evidence of
a causal association. Case-control studies have demonstrated associations between formaldehyde
exposure and rare cancers of the URT. Luce et al. (2002) evaluated pooled data from 12 case-
control studies and demonstrated a statistically significant increased risk between formaldehyde
exposure and sinonasal cancer. Hypopharyngeal cancer was linked with formaldehyde exposure
with an OR of 3.78 (95% CI 1.50-9.49) in another case-control study (Laforest et al., 2000).
Hauptmann and colleagues (2004) concluded that in spite of the small numbers of deaths from
cancers of the URT, the positive associations with average intensity and peak exposure were
consistent with the carcinogenicity of formaldehyde at these sites of first contact. The finding
that formaldehyde inhalation causes nasal squamous cell carcinoma in animals (Section 4.2.1.2)
further supports the determination of a causal association of formaldehyde exposure and
increased risk of upper respiratory tract cancer in humans. Both humans and animals developed
tumors within the upper respiratory tract, the POE site expected to receive direct exposure to
formaldehyde.
Overall, there is a consistent association between formaldehyde exposure and various
forms of lymphohematopoietic (LHP) cancers, with all leukemias, myeloid leukemia
specifically, Hodgkin lymphoma and multiple myeloma demonstrating the greatest strength and
consistency of results. Where exposure-response data exist, exposure-response trends have been
seen for all LHP malignancies, all leukemia, myeloid leukemia and Hodgkin lymphoma
(Pinkerton et al., 2004; Beane Freeman et al., 2009). Taken together, the data demonstrate a
consistent association, across various worker populations, with the expected temporal association
This document is a draft for review purposes only and does not constitute Agency policy.
4-453 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
to exposure and defined exposure-response relationships in two different worker cohorts. The
strongest associations tend to be with myeloid leukemia and Hodgkin lymphoma. The criterion
of reasonable biological plausibility is easily met for the majority of the diseases which
contribute to an observation of all LHP cancers, specifically the cancers derived from mature
lymphocytes. The potential for formaldehyde-induced LHP cancer is further supported by the
results of animal bioassays, where formaldehyde-induced leukemia and lymphoma had been
demonstrated in 3 independent studies in two species (rats and mice) and both sexes.
4.6. SUSCEPTIBLE POPULATIONS
"Susceptible subpopulations" is used here to refer to factors, such as life stage, genetics,
health status, etc., that may predispose individuals to greater response to an exposure. This
greater response could be achieved either through differences in exposure to the chemical or
differences in underlying toxicokinetic (TK) and toxicodynamic (TD) differences between the
susceptible and other populations. For example, life stages may include the developing
individual before and after birth up to maturity (e.g., preconception, embryo, fetus, young child,
adolescent), adults, or aging individuals. Another susceptibility factor is genetics. Specifically,
susceptible subpopulations may also include people with specific genetic polymorphisms that
render them more vulnerable to a specific agent or people with specific diseases or pre-existing
conditions (e.g., asthmatics). The term may also refer to gender differences, lifestyle choices, or
nutritional state (USEPA, 2002, Section 4.3.2.3).
A discussion of a comprehensive list of all possible susceptibility factors affecting
exposure and response to formaldehyde, or any chemical, is not possible. Therefore, the
discussion of susceptibility factors focuses on 1) factors hypothesized to be of importance to
formaldehyde; and 2) factors for which there are available formaldehyde data. A partial list of
these factors includes gender, genetic polymorphisms, preexisting disease status, nutritional
status, diet, and previous or concurrent exposures to other chemicals. Qualitatively, the presence
of multiple susceptibility factors will increase the variability that is seen in a population response
to formaldehyde toxicity.
4.6.1. Life Stages
Individuals at different life stages are physiologically, anatomically, and biochemically
different. Examples include physiological changes that occur through the lifespan (Selevan et
al., 2000). They may also have distinctive exposure pathways (i.e., transplacental, breast milk
ingestion), and exhibit differences in behavior (U.S. EPA, 2006b; NRC, 1993). Early life stages
(i.e., during development, prior to mature adulthood) and the later life stages (i.e., aging) differ
This document is a draft for review purposes only and does not constitute Agency policy.
4.454 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
greatly from mature adulthood in body composition, organ function, and many other
physiological parameters that can impact the TK and/or TD of chemicals and their metabolites
(ILSI, 2003). This section presents and evaluates the pertinent published literature available to
assess whether and how individuals of differing life stages may respond differently to
formaldehyde.
4.6.1.1. Early Life Stages
4.6.1.1.1. Factors influencing exposure and dosimetry
For all life stages, the primary exposure routes for formaldehyde include inhalation and,
in some cases, ingestion (see Chapter 5). Some exposure scenarios may be child specific. For
example, to the extent that the presence of baby furniture produced with formaldehyde in a
child's house contributes to greater concentrations in a child's room, exposures for very young
children in those circumstances may be increased (Environment California, 2008). As with all
chemicals, placental transfer and breast milk ingestion are exposure pathways that are unique to
early life stages. Studies assessing early life stage exposure pathways to formaldehyde have not
been performed. Presumably, unmetabolized formaldehyde reacts too quickly to be effectively
transported from mother to fetus by placental transfer; in addition, formaldehyde is not lipophilic
and is therefore unlikely to accumulate in breast milk. However, the relevant dose metric for
formaldehyde-related effects may vary depending on the specific target of concern (e.g., direct
toxicity at the portal of entry versus systemic effects); insufficient information is currently
available to determine whether individuals in different life stages are at higher risk for exposure
to specific target tissues.
There are some calculations however which shed light on lifestage differences in the
inhaled tissue dose at the portal of entry. Using respiratory tract surface areas and ventilation
rates reported in the literature and the scheme in USEPA (1994), Ginsberg et al. (2005)
calculated that overall extrathoracic absorption of highly reactive and soluble gases is similar in
adults and children. These results are in agreement with that of Garcia et al. (2009) who used
computational fluid dynamics to study differences in the nasal dosimetry of reactive, water-
soluble gases between 5 adults and 2 children, aged 7 and 8 years old. Overall uptake efficiency,
average flux (rate of gas absorbed per unit surface area of the nasal lining) and maximum flux
levels over the entire nasal lining did not vary substantially between adults and children (1.6-fold
difference in average flux and much less in maximum flux). On the other hand, the local flux of
formaldehyde varies between the two children by a factor of 2 to 4 at various distances along the
septal axis of the nose. The results in Garcia et al. (2009) have been further described and
evaluated in Appendix 3-1. Under normal resting breathing conditions, it is expected that very
This document is a draft for review purposes only and does not constitute Agency policy.
4-455 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
little formaldehyde is delivered to the lung. However, under higher activity as well as mouth
breathing scenarios, both of which appear likely to happen more regularly in children,
formaldehyde dose to the lung will be substantial.5
The toxicokinetic characteristics of formaldehyde are described in Chapter 3, with
absorption and distribution studies discussed in Sections 3.2 and 3.3. Studies to assess
differential absorption or distribution of formaldehyde in early life stages have not been
performed and represent a significant data gap. The metabolism of formaldehyde is described in
section 3.4. Expression of the enzymes that metabolize formaldehyde (ALDH2 and FALDH,
and specifically ADH3) is known to be developmentally regulated and thus may alter the TK of
formaldehyde in early life stages. ADH3 is ubiquitously expressed and is present in the human
fetus, neonate, and 1- to 10-year-old children (Hines and McCarver, 2002; Estonius et al., 1996).
During early development in rodents, when neurulation first begins and forms collections of
somites along the neural tube, ADH3 activities are significantly lower (at 8-10 and 11-13 somite
stages) and suggest a decreased ability to detoxify formaldehyde in the early embryo (Harris et
al., 2003). ADH mRNA expression levels appear to be age related, with decreased expression of
ADH common in premature neonates and infants up to 5 months old. Thereafter, ADH
expression increases and is dependent on body weight (Ginsberg et al., 2004). Benedetti et al.
(2007) reported that decreased ADH expression persisted until age 2 to 5 years. Westerlund et
al. (2005) tracked the ontogeny of ADH3 specifically and reported that ADH3 expression was
ubiquitous in mouse and rat embryos and was the only ADH enzyme to be consistently localized
to brain tissue, suggesting a housekeeping function. Thus, neonates and very young children
may have a decreased ability to metabolize formaldehyde due to differential expression of ADH3
in development compared that of with adults; however, activity levels of this enzyme and
alternate pathways specific to children are not available in the literature.
4.6.1.1.2. Life-stage exposure and adverse health outcomes
In general, exposure to toxic agents during early development (i.e, pre-conception,
prenatal stages, or postnatal development) may affect organ development and may also lead to
increased disease susceptibility later in life. Following early life stage exposure to
formaldehyde, a number of adverse health outcomes have been observed, including alterations in
the respiratory, reproductive, and neurological systems. For example, the developing respiratory
tract may be more vulnerable to insult compared with an adult respiratory tract, and thus,
increase the severity of response. The potential for reproductive and developmental toxicity of
5 For example, in the case of ozone concentrations of 0.1 ppm, a moderately reactive gas, Ginsberg et al. (2008)
predict a 5-fold variation in the dose to the deep lung between quiet and heavy breathing conditions for an 8-year
old child.
This document is a draft for review purposes only and does not constitute Agency policy.
4-456 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
formaldehyde is discussed in detail in Sections 4.1.1.7 (human studies) and 4.2.1.7 (animal
studies), while effects on the nervous system are discussed in Sections 4.1.1.6 and 4.2.1.6
(human and animal studies, respectively). The specific case of formaldehyde exposure and
pulmonary effects is discussed in detail in Sections 4.1.1.1 to 4.1.1.4 and 4.2.1.1 to 4.2.1.4. A
brief summary of identified effects of formaldehyde that may indicate susceptibility during
particular life stages is provided below.
4.6.1.1.2.1. Pre-conceytion.
Exposure prior to conception may damage reproductive organs and/or germ cells that
could affect reproduction and/or damage the genetic makeup of the offspring. Effects on
reproduction are discussed in Sections 4.1.1.7 and 4.2.1.8. In summary, an epidemiological
study (Taskinen et al., 1999) reported significantly delayed conception among female workers
exposed to formaldehyde at average daily ambient formaldehyde levels; these effects could be
consistent with adverse effects on either pre-conceptional and/or gestational exposure. One
animal study (Maronpot et al., 1986) reported endometrial hypoplasia and lack of ovarian luteal
tissue in female mice exposed for 13 weeks to 40 ppm formaldehyde via inhalation, suggesting
the potential for treatment-related alterations to the female reproductive system. Since the
exposure was to the adult, the findings suggest that preconceptional FA exposure caused female
reproductive system effects that in turn could affect pregnancy.
In the rodent study of Kitaev et al. (1984), a three-fold increase in embryo degeneration on
gestational days 2-3 was observed after FA exposure to the dams during premating. Since the
exposure was to the adult in these three studies, the findings suggest that preconceptional FA
exposure caused female reproductive system effects and/or affected the gametes.
4.6.1.1.2.2. Prenatal.
A population-based study (Grazuleviciene et al., 1998) found an association between
atmospheric formaldehyde exposure and low birth weight, yielding an adjusted OR of 1.37 (95%
CI: 0.90-2.09). Three studies (Dulskiene and Grazuleviciene, 2005; Taskinen et al., 1994;
Hemminki et al., 1985) that examined the effect of occupational exposures on the incidence of
congenital malformation produced mixed results.
Results from Taskinen et al. (1999) support associations between formaldehyde exposure,
subfertility, and spontaneous abortion. Subfertility and spontaneous abortion are biologically
linked (subclinical pregnancy losses are increased among women with fertility problems) (Gray
and Wu, 2000; Hakim et al., 1995), and both subfertility and spontaneous abortion may be
related to sensitivity to environmental agents (Correa et al., 1996).
This document is a draft for review purposes only and does not constitute Agency policy.
4-457 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Two experimental animal studies (Martin, 1990; Saillenfait et al., 1989) evaluated a
standard battery of developmental endpoints following inhalation exposure on GDs 6-10, but
effects were minimal. Similarly, Chernoff and Kavlock (1982), Marks et al. (1980), and Hurni
and Ohder (1973) reported minimal reproductive or developmental effects in rodents in studies
in which dams were exposed orally during early gestation. When formaldehyde was
administered via inhalation throughout gestation in female rats, some developmental effects,
including increased pup weight and decreases in lung and liver weight in newborns, were
reported at 0.01 and 0.4 ppm (Senichenkova and Chebotar, 1996; Senichenkova, 1991; Kitaev et
al., 1984; Gofmekler and Bonashevskaya, 1969; Gofmekler, 1968; Pushkina et al., 1968). Two
studies also reported changes in motor activity in offspring of dams exposed via inhalation to 0.4
ppm formaldehyde during gestation (Senichenkova, 1991; Sheveleva, 1971).
4.6.1.1.2.3. Postnatal
Following early life stage exposure to formaldehyde, a number of adverse postnatal
outcomes are possible, including effects on the developing and adult respiratory, reproductive,
and neurological systems. The potential for increased risk of childhood cancer is also discussed
below.
4.6.1.1.2.3.1. Respiratory toxicity.
Formaldehyde is known to induce changes in pulmonary function and cause pulmonary
irritation in human studies (Rumchev et al., 2002; Garrett et al., 1999; Krzyzanowski et al., 1990;
Holmstrom et al., 1989; Holmstrom and Wilhelmsson, 1988; Ritchie and Lehnen, 1987) and
animal studies (Ohtsuka et al., 2003, 1997; Riedel et al., 1996; Swiecichowski et al., 1993; Lee et
al., 1984). Exposure to formaldehyde in early life can cause damage to the lungs and
permanently influence airway function, resulting in increased vulnerability to toxicants later in
life. Thus, young children may demonstrate increased susceptibility to formaldehyde-related
health effects. Krzyzanowski et al. (1990) reported an association between physician-diagnosed
asthma and chronic bronchitis in children who lived in homes with formaldehyde levels that
were higher than 60 ppb, after controlling for socioeconomic status and ethnicity. Rumchev et
al. (2002) reported a statistically significant increased risk of asthma with increased residential
concentrations of formaldehyde. Garrett et al. (1999) found an increased association between
bedroom concentration of formaldehyde and increased risk of atopy in children. These studies
suggest that formaldehyde exposure may exacerbate responses in sensitive airways, particularly
in children. Exacerbation of response has also been noted in asthmatic adults and will be
discussed below.
This document is a draft for review purposes only and does not constitute Agency policy.
4-458 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
Another child-specific concern is that respiratory irritation may have greater impact on
lung function in children due to their smaller lung size; this is true even if the lung development
is normal. Irritation is commonly accompanied by inflammation, which can have a greater
impact on children's airways because they are narrower than adult airways. Thus, less
inflammation is required for significant airway obstruction in children than in adults.
4.6.1.1.2.3.2. Developmental neurotoxicity.
In neonatal exposure paradigms, changes in brain structure (Sarsilmaz et al., 2007; Asian
et al., 2006), and brain chemistry (Songur et al., 2008) were seen in young rats following
inhalation exposures (6000 or 12000 ppb, 5 days per week from postnatal day 0-30). In addition,
Weiler and Apfelbach (1992) found juvenile animals to be more sensitive to formaldehyde-
induced changes in olfactory thresholds when compared with adult animals (shifts in olfactory
thresholds were greater when exposure [at 250 ppb] was initiated at PND 30 than at adult ages).
These studies are consistent with the hypothesis that early life exposure to FA can lead to long-
lasting neurological effects. Exposure levels in these studies (250-6,000 ppb) were in the same
range as those producing the behavioral effects in adults (as low as 100 ppb), but provide limited
information regarding relative sensitivity as no NOAELs were identified, and (with the exception
of Weiler and Apfelbach), similar parameters were not measured in adult animals using the same
exposure paradigms.
4.6.1.2. Later Life Stages
In general, older adults may be at risk for increased susceptibility to exposure to
environmental chemicals by virtue of their slower metabolism and increased incidence of altered
health status (Benedetti et al., 2007; Ginsberg et al., 2005; U.S. EPA, 2005a). Additionally,
adverse effects of earlier exposure to some toxicants may be observed in older adults as a result
of latency in expression of the effect (Olsen et al., 1997; Sweeney et al., 1986). No studies have
examined the differential effects of formaldehyde exposure for elderly adults (>65 years old) as
compared to other age groups.
4.6.1.3. Conclusions on Life-Stage Susceptibility
In summary, timing both of the exposure and of the assessment of health outcomes may
be important for understanding the relative risk of adverse effects from formaldehyde exposure
during different life stages. There are known developmental differences in kinetics across life
stages, including differences in enzymes involved in formaldehyde metabolism, but the
contribution of these differences to formaldehyde-related health effects is unknown. Similarly,
This document is a draft for review purposes only and does not constitute Agency policy.
4-459 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
information regarding life-stage differences in respiratory physiology raises possible concern
regarding increased exposure to children, but studies for formaldehyde are not available.
Available data do support an increased risk for adverse effects on lung function in children. The
overall body of evidence shows some support, although minimal, for susceptibility in
reproductive or developmental endpoints associated with exposure to formaldehyde. Some
studies observed altered development of the nervous system following formaldehyde exposure
during early life. Older adults may be at risk for increased susceptibility to formaldehyde
because of slower metabolism and clearance rates. Elderly adults have an increased probability
of having both altered health status and altered metabolism, which could impact their ability to
process and recover from an adverse effect. The available data are consistent with some life-
stage susceptibility differences for FA at the level of TD or TK differences, the results are
nonetheless inconclusive due to the number of data gaps.
4.6.2. Health/Disease Status
The factor for which we have the greatest evidence is pre-existing disease, and
specifically asthma. Numerous studies have assessed the potential for increased susceptibility to
formaldehyde in asthmatics. Formaldehyde does not induce airway hyperreactivity directly
(Sheppard et al., 1984) and has not been shown to increase airway hyperreactivity in either
asthmatics or non-asthmatics (Pazdrak et al., 1993; Harving et al., 1991; Kulle et al., 1987).
Significantly decreased forced expiratory volume (FEVi) measurements were reported among
asthmatics in two studies (Casset et al., 2006; Green et al., 1987), while others did not find any
significant change in FEVi following formaldehyde exposure (Ezratty et al., 2007; Frigas et al.,
1984).
A few available case reports of bronchial asthma do suggest direct respiratory tract
sensitization to formaldehyde gas (Lemiere et al., 1995; Burge et al., 1985; Hendrick et al., 1982;
Hendrick and Lane, 1977, 1975). All cases displayed marked changes in FEVi or pulmonary
airflow rate in response to acute challenges with formaldehyde gas at exposure levels <3 ppm.
Formaldehyde-induced IgE production has been reported in some studies (Vandenplas et al.,
2004; Wantke et al., 1996a).
There is a large quantity of human data providing evidence of an association between
formaldehyde exposure and increased incidence of asthma or exacerbation of asthmatic
responses in compromised individuals. For example, Krzyzanowski et al. (1990) reported an
association between physician-diagnosed asthma and chronic bronchitis in children who lived in
homes with formaldehyde levels that were higher than 60 ppb, after controlling for
socioeconomic status and ethnicity. Rumchev et al. (2002) reported a statistically significant
This document is a draft for review purposes only and does not constitute Agency policy.
4-460 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
increased risk of asthma with increased residential concentrations of formaldehyde. Garrett et al.
(1999) found an increased association between bedroom concentration of formaldehyde and risk
of atopy in children. These studies suggest that formaldehyde exposure may exacerbate
responses in sensitive airways, particularly in children. Exacerbation of response has also been
noted in adults. Kriebel et al. (1993) reported a greater decrease in peak expiratory flow (PEF)
in asthmatic medical students (7.3% decrement) compared with non-asthmatic medical students
(2% decrement) after 2 weeks exposure to formaldehyde (average concentration 0.73 ppm) in an
anatomy lab. This effect does not appear to be immunogenic in nature (Fujimaki et al., 2004a;
Lee et al., 1984).
Several animal studies document increased airway resistance and bronchial constriction
following inhalation exposure to formaldehyde (Nielson et al., 1999; Swiecichowski et al., 1993;
Biagini et al., 1989; Amdur et al., 1960). Sadakane et al. (2002) demonstrated that formaldehyde
exposure exacerbated sensitization and challenge with a common dust mite allergen (Der f) as
measured by increased eosinophil infiltration into the interstitium around the bronchi and
bronchioles as well as goblet cell proliferation in the bronchial epithelium; they suggested that
formaldehyde exposure may aggravate eosinophilic infiltration and goblet cell proliferation that
accompanies allergic responses. The MOA underlying this response is unknown. These
decrements may occur indirectly in response to formaldehyde and may be mediated via
neurogenic potentiation (Sadakane et al., 2002; Riedel et al., 1996; Tarkowski and Gorski, 1995).
In particular, Tarkowski and Gorski (1995) suggest that formaldehyde may increase permeability
of respiratory epithelium and destruction of immunologic barriers. Thus, the respiratory tract
may become vulnerable to inhaled allergens after formaldehyde exposure (Tarkowski and
Gorski, 1995).
In summary, the data indicate that formaldehyde exposure can aggravate a type I
hypersensitivity response and that this hypersensitivity may in turn increase the susceptibility to
FA exposure in these individuals. Formaldehyde exposure may predetermine an asthmatic
phenotype or may induce new incidences of asthma via indirect mechanisms, though definitive
evidence and a proposed mechanism remain to be determined. Individuals that exhibit
chemically induced sensitivity and are exposed acutely or chronically to formaldehyde in
residential and occupational settings might exhibit adverse responses at lower concentrations of
formaldehyde than the average healthy person.
4.6.3. Nutritional Status
Limited available data indicate that certain types of malnutrition may increase
susceptibility to formaldehyde exposure. Senichenkova and Chebotar (1996) reported increased
This document is a draft for review purposes only and does not constitute Agency policy.
4-461 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
fetal anomalies in fetuses from iron-deficient pregnant mice after formaldehyde exposure
compared with anemic mice that had not been exposed to formaldehyde. Forced iron reduction
(induced by addition of bipyridyl treatment in pregnant mice) in utero increased the overall
incidence of fetal anomalies when paired with formaldehyde exposure (Senichenkova and
Chebotar, 1996). The findings are difficult to evaluate due to poor reporting and have not been
substantiated by other laboratories.
4.6.4. Gender Differences
Males and females can differ greatly in body composition, organ function, and many
other physiological parameters that may influence the toxicokinetics of chemicals and their
metabolites in the body (Gochfeld, 2007; Gandhi et al., 2004).
The human epidemiology data set does not support any specific gender susceptibilities
for noncancer effects due to formaldehyde exposure. In general, data suggest that nonpregnant
women, on a per kg body weight basis, may have slightly lower air intake than men, which
would suggest that women may be less susceptible to inhaled pollutants like formaldehyde than
men, but this has not been investigated in the available formaldehyde literature.
A few isolated reports have investigated potential gender differences in development of
nasal pharyngeal carcinomas following exposure to formaldehyde. One case-control study
identified a higher OR for sinonasal adenocarcinomas in women (OR = 6.2 [95% CI: 2.2-19.7])
compared with the OR observed in men (OR = 3.0 [95% CI: 1.5-5.7]) following exposure to
formaldehyde (Luce et al., 2002). However, the overall body of evidence remains scant.
There are a few reports concerning differential formaldehyde-induced effects on the male
and female reproductive systems. Ozen et al. (2002), Sarsilmaz et al. (1999), and Woutersen et
al. (1987) reported reduced Ley dig cell numbers in adult male rats exposed by inhalation. In
female mice, inhalation exposure to formaldehyde resulted in endometrial hypoplasia and lack of
ovarian luteal tissue (Maronpot et al., 1986). The clinical significance of these effects in humans
is unknown, and due to limited data it is unclear whether the female or male reproductive system
is more susceptible to perturbation by formaldehyde.
4.6.5. Genetic Differences
There are some data for polymorphisms in humans that affect formaldehyde TK. As
discussed in Section 3.4, the primary metabolizing enzymes of formaldehyde are ALDH2 and
ADH3, with the latter enzyme considered more relevant to low exposures. Polymorphisms in
ALDH2 have been shown to have implications in human risk assessment, specifically in regard
to acetaldehyde metabolism (Ginsberg et al., 2002). Teng et al. (2001) demonstrated the
This document is a draft for review purposes only and does not constitute Agency policy.
4-462 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
importance of ALDH2 for formaldehyde metabolism in rat hepatocytes at fairly high
formaldehyde concentrations (2.5 mM and greater). Cheng et al. (2008) investigated the
relationship between occupational exposure to formaldehyde and genetic polymorphisms of
ALDH2 and CYP2E1. There was a positive relationship between the concentration of formic
acid in the urine and ALDH2 genotypes (x2 = 9.241 ,p< 0.05). Urinary formic acid
concentration may be affected by formaldehyde exposure concentration and ALDH2 genotype
(Cheng et al., 2008) for individuals that have high exposure levels. Thus, although ALDH2 may
not be involved in formaldehyde metabolism if exposure levels are low, polymorphisms of this
enzyme may lead to differences in response at higher exposure levels.
Wu et al. (2007) looked for and identified two SNPs in ADH3 among a population of
Mexican asthmatic children 4 to 17 years of age. Carrying one or two copies of the minor allele
for one SNP resulted in a decreased RR of asthma (RR = 0.66-0.77). For the second SNP,
homozygotes for the minor allele had an RR of 1.6 for asthma. The functional characteristics of
these SNPs are unknown. Studies evaluating whether any of the polymorphisms affect the
expression, regulation, stability, or activity of the enzyme in vivo are lacking; therefore, the
relative susceptibility of individuals with different polymorphisms cannot be characterized at this
time.
One study (Hedberg et al., 2001) identified three polymorphisms in human ADH3
involving four base-pair substitutions in the promoter region of which one (C^T) showed
reduced activity (-50-70% of control). Hedberg et al. (2001) reported differences in allele
frequencies among Chinese, Spanish, and Swedish groups, consisting of Asian-Caucasian
differences and ethnic subgroups among Caucasians. Their results suggest that a small
percentage of Caucasians may have decreased ADH3 expression and thus, be more susceptible to
formaldehyde exposure. Additional studies to validate these findings have not been performed.
The relative activity level of these enzymes may also impact the metabolism of
formaldehyde. In pharmacokinetic studies, deletion of ADH3 increased the sensitivity of mice to
formaldehyde (Deltour et al., 1999) and was deleterious to yeast (Achkor et al., 2003). These
results suggest that deficiencies in ADH3 may confer an increased susceptibility to formaldehyde
toxicity (Teng et al., 2001). The importance of properly functioning enzymes also suggests that
genetic differences in ADH3 or ALDH2 may affect the response to formaldehyde exposure.
However, comparable human data are not available.
Race/ethnicity may be a surrogate for genetic differences but racial or ethnic groups may
also reflect socioeconomic, and/or cultural factors that are distinct from genetics. Possible ethnic
differences may be related to genetic polymorphisms of enzymes ALDH2 and ADH3, relevant
for formaldehyde metabolism. ALDH2 variants, present primarily in East Asians, are known to
This document is a draft for review purposes only and does not constitute Agency policy.
4-463 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
have protective effects against alcoholism but were not found in the people of Indo-Trinidadian
descent (Moore et al., 2007) or in American Indians or Alaska natives (Ehlers, 2007). However,
there is no direct evidence to associate these variants to differential susceptibility to
formaldehyde exposure, nor is there direct evidence to associate these ethnic groups specifically
with differential susceptibility to formaldehyde. Further, no studies have specifically assessed
ethnic variability in responses to formaldehyde.
There are complex pathways through which genetic polymorphisms in ADH3 can
potentially affect differential susceptibility to formaldehyde. Firstly, ADH3 is central to the
metabolism of formaldehyde. However, ADH3 itself may indirectly contribute to the adverse
effects of formaldehyde on pulmonary physiology (Thompson et al., 2009; Staab et al., 2008a, b;
Thompson and Grafstrom, 2008). Exposure to formaldehyde is itself thought to alter the activity
of ADH3 resulting in the perturbation of critical metabolic pathways. ADH3 participates in the
oxidation of retinol and long-chain primary alcohols, as well as the reduction of S-
nitrosoglutathione (GSNO). The activity of ADH3 toward some of these substrates has been
shown to be significantly increased in the presence of formaldehyde. ADH3 has recently also
been shown to contribute to NO signaling through its dual role in metabolizing GSNO, an
endogenous bronchodilator and reservoir of NO (Staab et al., 2008a; Hess et al., 2005; Jensen et
al. 1998). Through its regulatory function on GSNO, ADH3 may thus play a central role in
regulating bronchial tone allergen-induced hyperresponsiveness (Gerard, 2005; Que et al., 2005).
As concluded by California Environmental Protection Agency (CalEPA) (2008), "the
dysregulation of NO by formaldehyde [in this manner] helps to explain the variety and
variability in the toxic manifestations following formaldehyde inhalation."
4.6.6. Co-Exposures
4.6.6.1. Cumulative Risk
When considering health risks, it is important to consider the impact of co-exposures to
other agents that may interact with the chemical under evaluation. Co-exposure to other
pollutants, particularly those that produce some of the same metabolites and similar health
effects as formaldehyde, is likely to occur in both occupational and nonoccupational settings.
Due to effects on metabolic enzymes (inducing and/or inhibition) as well as direct effects
on organ system function, co-exposures may alter the way in which formaldehyde is metabolized
and cleared from the body. Inhibition or induction of the enzymes responsible for metabolism of
chemicals may alter susceptibility to toxicity (Lash and Parker, 2001; IARC, 1995; U.S. EPA,
1985a). Smokers may be at increased risk for effects of formaldehyde exposure, because
formaldehyde is one of the components of cigarette smoke and is likely to heighten the point-of-
This document is a draft for review purposes only and does not constitute Agency policy.
4-464 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
entry effect when combined with occupational or residential exposures to inhaled formaldehyde.
However, no evidence is available to evaluate the potential aggregate effects.
4.6.6.2. Aggregate Exposure
In addition, multiple routes of exposure to a single agent may increase the cumulative
risk by increasing the overall body burden of the chemical. A human aggregate exposure model
developed by McKone and Daniels (1991) incorporated likely exposures from air, water, and soil
media through inhalation, ingestion, and dermal contact. The authors hypothesized that the
aggregate exposure could be age-dependent but did not present any data for persons of differing
life stages. The role of multiple exposures on different genders, genetic susceptibility, or altered
health and nutrition status has not been investigated. The available database regarding the
potential for multiple routes of exposure (or aggregate exposure) formaldehyde is limited.
Guseva (1972) specifically assessed the reproductive and developmental effects caused
by co-exposure to formaldehyde via both inhalation (0.25 mg/m3) and ingestion (0.01 mg/L)
routes in male rats. The authors reported reduced nucleic acid levels in testes to 88 and 92% of
controls, which suggests a possible toxic gonadotropic effect. The ability of male rats (receiving
combined exposure to formaldehyde at a low concentration level for a long period of time) to
reproduce was preserved since all the cohabited females were impregnated. The number and
weight of the fetuses and newborn rat pups in the experimental co-exposure groups did not differ
substantially from those figures observed in the control group. No developmental defects or
anomalies were observed in the offspring for up to 1 month postnatally. Thus, at low exposures,
the reproductive effects due to combined ingestion and inhalation exposure are unknown.
4.6.7. Uncertainties of Database
There is a need to better characterize the implications of formaldehyde exposures to
susceptible populations. A number of areas where the database is currently insufficient are
identified below.
4.6.7.1. Uncertainties of Exposure
Although information exists on early life exposure to formaldehyde, a number of
uncertainties regarding children's susceptibility remain. First, inhalation is believed to be of
most concern for formaldehyde, since formaldehyde vapors are released from insulation or from
ambient sources of formaldehyde, including secondary production from other pollutants involved
in photo-oxidant reactions. Any additional pathways of exposure for children have not been
characterized. Since formaldehyde is nearly ubiquitous in the environment, it is difficult to
This document is a draft for review purposes only and does not constitute Agency policy.
4-465 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
quantify the total exposure. Second, children have different respiratory, metabolic, and activity
rates compared with healthy adults, potentially influencing ADME and target tissue exposure to
formaldehyde. However, studies to identify the specific changes in absorption of formaldehyde
and its metabolites across developmental stages and across organs have not been performed. In
addition, exposure prenatally may be altered based on whether formaldehyde or its metabolites
pass through the placenta, but placental transfer data are not available. Third, no quantitative
models have been developed to characterize these differences for formaldehyde. Formaldehyde-
specific PBPK models and their validation will aid in understanding the uncertainties associated
with formaldehyde exposure in children.
Given the large proportion of time that most individuals in the U.S. spend indoors,
exposure scenarios where indoor concentrations to formaldehyde are high (e.g., in homes or in
trailers; see section 2.3.1) may play a significant role and may be of particular concern to the
elderly or health-impaired individuals who spend relatively more time at home. Further
evaluation of the effects of co-exposures and pathways of exposure and aggregate risk is needed.
An estimate of the multiple exposure pathways is needed to know where along the dose-response
curve to place an incremental exposure to formaldehyde.
4.6.7.2. Uncertainties of Effect
Studies specifically designed to evaluate effects after early and later life stage exposure
are needed in order to more fully characterize potential life-stage-related differences in
formaldehyde toxicity, including the defining of critical windows during development. For
example, life-stage-specific neurotoxic and pulmonary effects, particularly in the developing
fetus, need further evaluation. The preconceptional period may be a critical window for FA
exposure and reproductive and developmental effects, based on rodent studies of reproductive,
embryonic and gamete effects. Data specific to the carcinogenic effects of formaldehyde
exposure during early life stages do not exist. The reduction in fertility seen in some studies
(Gray and Wu, 2000; Taskinen et al., 1999; Hakim et al., 1995) is not adequately described and a
well-established MOA has not been identified, but some have been hypothesized including
altered sperm quality (Ozen et al., 2002; Sarsilmaz et al., 1999; Woutersen et al., 1987). Further,
spontaneous abortion/fetal loss occurring early in gestation, prior to maternal knowledge of the
pregnancy, can lead to misclassification of the effect as infertility (see Sections 4.1.1.7 and
4.2.1.7).
More research is needed to clarify the role of genetic polymorphisms in formaldehyde
metabolism. Similarly, data gaps pertaining to gender differences remain. A potential impact of
nutritional status and iron deficiency on formaldehyde toxicity needs further investigation.
This document is a draft for review purposes only and does not constitute Agency policy.
4-466 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
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
A fair body of evidence suggests that asthmatics are more susceptible to formaldehyde exposure
than the general population, however the mechanism of action for this increased susceptibility is
unknown.
In the studies discussed above, there are a number of examples of studies that assessed
multiple susceptibility factors that are worth noting. For example, the Krzyzanowski et al.
(1990) study reported asthma and chronic bronchitis cases for two interacting potential
susceptible groups, in children and those with high exposure (due to living in homes with
formaldehyde levels that were higher than 60 ppb). Similarly, the Garrett et al. (1999) study
assessed the same two interacting potential susceptible groups.
The study of Senichenkova and Chebotar (1996) assessed developmental effects in
mouse fetuses after in utero iron-deficiency and FA exposure. Thus, the study findings must be
considered in light of possible interactions between life stage exposure differences and
nutritional status differences.
Studies to understand the nature of the interactions between the various susceptibility
factors for FA have not been performed.
4.6.8. Summary of Potential Susceptibility
There is some evidence to demonstrate susceptibility for various populations exposed to
formaldehyde. Available data are summarized in Table 4-96 where FA susceptibility factors are
presented by those with data for increased FA susceptibility and those with data for differences
but with an unknown impact on FA susceptibility.
Exposure to FA during early developmental and later life stages may be of concern.
However, human exposure to the developing fetus is unknown since it is not known whether
formaldehyde or one of its metabolites crosses the placenta. However, there is very limited life-
stage-specific information regarding the TK of formaldehyde. Life-stage-specific TK has not
been characterized, and, thus, no PBPK models exist to effectively evaluate the risk to early life
stages. Children may be more susceptible to noncancer health effects as a result of inhalation
exposure to formaldehyde due to increased respiratory rates. There are no studies to evaluate
whether formaldehyde exposure in early life (e.g., pregnancy) is associated with an increased
risk of childhood cancer.
The weight of evidence supports a plausible association between formaldehyde exposure
and aggravated asthmatic responses in humans and this association is corroborated by limited
evidence from animal studies. Formaldehyde does not appear to directly induce airway
hyperreactivity but may sensitize airways to subsequent exposures. One issue in interpreting the
available studies that assessed the relationship between asthma and FA could not distinguish
This document is a draft for review purposes only and does not constitute Agency policy.
4-467 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
between the cases of asthma that were due to earlier FA exposure vs. those without a direct link
to FA exposure.
No direct link exists between formaldehyde exposure and differential susceptibility in
different ethnic groups, although genetic polymorphisms in the enzymes involved with
formaldehyde metabolism, ADH3 and ALDH2, provide some support for differential
susceptibility to alcoholism in a number of ethnic groups. The evidence for differential gender
responses to formaldehyde exposure is equivocal. Co-exposures may result in altered
metabolism and clearance, but there is no evidence that co-exposures are a critical part of
formaldehyde-mediated differential susceptibility.
Thus, given the available data, increased susceptibility to adverse effects of
formaldehyde is most strongly supported for three populations: 1) Preconception and perinatal
exposure based on reproductive and developmental effects; 2) children, whose exposure may be
higher by virtue of their increased activity level and respiratory rate; and 3) asthmatics who may
exhibit exacerbation of response to formaldehyde.
This document is a draft for review purposes only and does not constitute Agency policy.
4-468 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 4-96. Available evidence for susceptibility factors of concern for
2 formaldehyde exposure
3
Factor
Evidence that factors increase
susceptibility to FA
Evidence that factors show differences
but unknown impact on susceptibility
Life Stage
¦ Preconception
¦ Prenatal
¦ Postnatal
Developmental effects reported
suggesting that critical windows of
exposure may be relevant:
¦ Reproductive outcomes (Taskinen et
al., 1999; Maronpot et al., 1986)
¦ Embryo effects (Kitaev et al., 1984)
¦ Structural- and functional
developmental outcomes (Martin,
1990; Saillenfait et al., 1989;
Sheveleva, 1971; Seninchenkova,
1991)
¦ Lung function outcome
(Krzyzanowski et al., 1990; Rumchev
et al., 2002; Garrett et al., 1999)
¦ Developmental neurotoxicity (Weiler
and Apfelbach, 1992)
¦ Possible life stage level differences in
some enzymes involved in FA
metabolism (Harris et al., 2003;
Ginsberg et al., 2004; Westerlund et al.,
2005; Benedetti et al., 2007)
¦ Mixed reports of associations between
prenatal exposure and developmental
outcomes in human studies (positive
association: Grazuleviciene et al., 1998)
¦ Possible life stage level differences in
some enzymes involved in FA
metabolism (e.g., |ADH expression over
first 5 months; Ginsberg et al., 2004)
¦ Developmental neurotoxicity (Sarsilmaz
et al., 2007; Asian et al., 2006; Songur et
al., 2008)
Disease Status
¦ Bronchial asthma (Lemiere et al.,
1995; Burge et al., 1985; Hendrick et
al., 1982; Hendrick and Lane, 1977,
1975)
¦ Increased airway resistance and
bronchial constriction (Nielson et al.,
1999; Swiecichowski et al., 1993;
Biagini et al., 1989; Amdur et al.,
1960)
¦ Mixed results for forced expiratory
volume (FEV1) measures affected by FA
exposure in asthmatics (Casset et al.,
2006; Green et al., 1987; Ezratty et al.,
2007; Frigas et al., 1984)
Nutritional Status/Diet
¦ Iron-deficiency in utero
(Senichenkova and Chebotar, 1996).
Genetics
¦ Polymorphisms
¦ For high FA exposure: Urinary formic
acid levels affected by ALDH2
genotype (Cheng et al., 2008)
¦ In mice, ADH3 increased sensitivity to
FA (Achkor et al., 2003)
¦ Differences among ADH3 alleles and
asthma outcome (Wu et al., 2007)
¦ Differences among ethnic groups in
ADH3 alleles (Hedberg et al., 2001)
Gender
¦ Gender differences in incidence of
nasopharyngeal carcinoma following FA
exposure (Luce et al., 2002)
4
5
e - End of Volume II -
This document is a draft for review purposes only and does not constitute Agency policy.
4-469 DRAFT—DO NOT CITE OR QUOTE
------- |