DRAFT-DO NOT CITE OR QUOTE
mm iliA
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 I of IV
Introduction, Background,
and Toxicokinetics
March 17, 2010
NOTICE
This document is an Inter-Agency Science Consultation review 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.
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document is a preliminary draft for review purposes only. 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. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy.
I-ii DRAFT—DO NOT CITE OR QUOTE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. PI IYSICOCIIHYIICAI. 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.
I-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.
I-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 MOAs 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.
I-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.
I-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.
I-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. Maj or 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.
I-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.
I-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.
I-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.
I-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.
I-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.
I-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.
I-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.
I-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.
I-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.
I-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, LECooos, 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.
I-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. ECoos, LECoos, 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.
I-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.
I-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.
I-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.
I-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) HT-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.
I-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 5-110
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.
I-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
BfR
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.
I-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
CO2 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.
I-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.
I-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
ly mphohematopoi eti c
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.
I-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.
I-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.
I-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.
I-xxxi DRAFT—DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic inhalation exposure to
formaldehyde. It is not intended to be a comprehensive treatise on the chemical or toxicological
nature of formaldehyde.
In Chapter 6, Major Conclusions in the Characterization of Hazard and Dose Response,
EPA has characterized its overall confidence in the qualitative and quantitative aspects of hazard
and dose response by addressing knowledge gaps, uncertainties, quality of data, and scientific
controversies. The discussion is intended to convey the limitations of the assessment and to aid
and guide the risk assessor in the ensuing steps of the risk assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
This document is a draft for review purposes only and does not constitute Agency policy.
I-xxxii DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGERS
John E. Whalan, D.A.B.T.1
EPA-ORD-NCEA
Danielle DeVoney, PhD, D.A.B.T., PE2
EPA-ORD-NCEA
AUTHORS
Thomas Bateson, PhD
EPA-ORD-NCEA
Susan Euling, PhD
EPA-ORD-NCEA
Jennifer Jinot, PhD
Susan Makris, PhD
EPA-ORD-NCEA
Kathleen Raffaele, PhD
EPA-ORD-NCEA
John Schaum, PhD
EPA-ORD-NCEA
Ravi Subramaniam, PhD
EPA-ORD-NCEA
Suryanarayana Vulimiri, PhD
EPA-ORD-NCEA
CONTRIBUTORS
Gillian Backus, PhD3
Stanley Bar one, PhD
EPA-ORD-NCEA
1 Chemical Manager since July 2003.
2 Chemical Manager since June 2009
3 Separated from the Agency prior to final revisions to document.
This document is a draft for review purposes only and does not constitute Agency policy.
I-xxxiii DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
David Bayliss, PhD3
EPA-ORD-NCEA
Ted Berner, PhD
EPA-ORD-NCEA
David Bussard
EPA-ORD-NCEA
David Farrar, PhD
EPA-ORD-NCEA
John Fox, PhD
EPA-ORD-NCEA
EPA-ORD-NCEA
Karen Hogan, PhD
EPA-ORD-NCEA
Rosemarie Hakim, PhD3
EPA-ORD-NCEA
Babashaheb Sonawane, PhD
EPA-ORD-NCEA
Chad Thompson, PhD3
EPA-ORD-NCEA
Larry Valcovic, PhD3
EPA-ORD-NCEA
John J. Vandenberg, PhD
EPA-ORD-NCEA
Lisa Vinikoor, PhD
EPA-ORD-NCEA
Paul White, PhD
EPA-ORD-NCEA
This document is a draft for review purposes only and does not constitute Agency policy.
I-xxxiv DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
CONTRACTOR SUPPORT
The literature search and preliminary drafts of this document as well as support for
editing and formatting were provided by Oak Ridge Institute for Science and Education
(ORISE), Oak Ridge Associated Universities (ORAU), Department of Energy, under
Interagency Agreement (IAG) Project No. 03-18. The ORISE individuals who contributed to
this effort include Sheri Hester, George Holdsworth, Bobette D. Nourse, Wanda Olson, and Lutz
W. Weber.
Assistance with the biologically based dose response model evaluation was provided by
ENVIRON International Corporation of Monroe, Louisiana (subcontractors to ORISE; Project
No. 03-18). The primary scientists involved in the work were Kenny S. Crump and Cynthia Van
Landingham.
INTERNAL EPA REVIEWERS
Daniel Axelrad, PhD
Office of Policy, Economics, and Innovation
Iris Camacho, PhD
Office of Pollution Prevention and Toxics
Christina Cinalli, PhD
Office of Pollution Prevention and Toxics
Rebecca Edelstein, PhD
Office of Pollution Prevention and Toxics
Ernest Falke, PhD
Office of Pollution Prevention and Toxics
Stiven Foster, PhD
Office of Solid Waste and Emergency Response
Greg Fritz
Office of Pollution Prevention and Toxics
Susan Griffin, PhD
EPA Region 8
This document is a draft for review purposes only and does not constitute Agency policy.
I-xxxv DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Timothy Leighton, PhD
Office of Pesticide Programs
Elizabeth Margosches, PhD
Office of Pollution Prevention and Toxics
Timothy McMahon, PhD
Office of Pesticde Programs
Julie Migrin-Sturza, PhD
Office of Policy, Economics, and Innovation
Greg Miller, PhD
Office of Children's Health Protection and Environmental Education
Deirdre Murphy, PhD
Office of Air and Radiation
Marion Olson, PhD
EPA Region 2
Andrea Pfahles-Hutchens, PhD
Office of Pollution Prevention and Toxics
Jennifer Seed, PhD
Office of Pollution Prevention and Toxics
This document is a draft for review purposes only and does not constitute Agency policy.
I-xxxvi 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.
I-xxxvii DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
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
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of
formaldehyde. IRIS Summaries may include oral reference dose (RfD) and inhalation reference
concentration (RfC) values for chronic and other exposure durations, and a carcinogenicity
assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal of entry [POE]) and
for effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per (j,g/m3 air breathed.
Development of these hazard identification and dose-response assessments for
formaldehyde has followed the general guidelines for risk assessment as set forth by the National
Research Council (NRC) (1983). EPA Guidelines and Risk Assessment Forum Technical Panel
Reports that may have been used in the development of this assessment include the following:
Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines
for Mutagenicity Risk Assessment (U.S. EPA, 1986b), Recommendations for and Documentation
This document is a draft for review purposes only and does not constitute Agency policy.
1 -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
of Biological Values for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for
Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and
Limit Concentration Issues in Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of
Inhalation Reference Concentrations and Application of Inhalation Dosimetry (U. S. EPA,
1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995),
Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998), Science Policy Council Handbook. Risk
Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002a), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA, 2005b), Science Policy Council Handbook. Peer Review (U.S. EPA, 2006a), and A
Framework for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA,
2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through April, 2009,
but some criticial literature after this date has been considered in this assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
1 -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
2. BACKGROUND
This chapter provides an overview of the physical and chemical characteristics of
formaldehyde. Also provided in this chapter are a description of the production, uses, and
sources of formaldehyde and information regarding environmental levels and human exposure.
A description of the toxicokinetics and toxicodynamic processes involved in formaldehyde
toxicity for the inhalation, oral, and dermal routes can be found in Chapter 3 (Toxicokinetics).
2.1. PHYSICOCHEMICAL PROPERTIES OF FORMALDEHYDE
Formaldehyde (CASRN 50-00-0) is the first of the series of aliphatic aldehydes and is a
gas at room temperature. Its molecular structure is depicted in Figure 2-1. It is noted for its
reactivity and versatility as a chemical intermediate. It readily undergoes polymerization, is
highly flammable, and can form explosive mixtures with air. It decomposes at temperatures
above 150°C.
Figure 2-1. Chemical structure of formaldehyde.
At room temperature, pure formaldehyde is a colorless gas with a strong, pungent,
suffocating, and highly irritating odor. Formaldehyde is readily soluble in water, alcohols, ether,
and other polar solvents. A synopsis of its physicochemical properties is given in Table 2-1.
2.2. PRODUCTION, USES, AND SOURCES OF FORMALDEHYDE
Formaldehyde has been produced commercially since the early 1900s and, in recent
years, has been ranked in the top 25 highest volume chemicals produced in the U.S. (National
Toxicology Program [NTP], 2002). In 2003, 4.33 million metric tons of formaldehyde were
produced in the U.S. (Global Insight, 2006). In 2000, worldwide formaldehyde production was
estimated to be 21.5 million metric tons, (International Agency for Research on Cancer [IARC],
2006).
H
C=0
H
This document is a draft for review purposes only and does not constitute Agency policy.
2-1 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 2-1. Physicochemical properties of formaldehyde
2
Name
Formaldehyde
International Union for Pure and Applied
Chemistry name
Formaldehyde
Synonyms
Formic aldehyde
Methanal
Methyl aldehyde
Methylene oxide
Oxomethane
Oxymethylene
Chemical Abstracts Service Index name
Formaldehyde
CASRN
50-00-0
Formula
HCHO
Molecular weight
30.03
Density
Gas: 1.067 (air= 1)
Liquid: 0.815 g/mL at-20°C
Vapor pressure
3,883 mm Hg at 25°C
Log Kow
-0.75 to 0.35
Henry's law constant
3.4 x 10 atm-m3/mol at 25°C
2.2 x 10 2 Pa-m3/mol at 25°C
Conversion factors (25°C, 760 mm Hg)
1 ppm = 1.23 mg/m3 (v/v)
1 mg/m3 = 0.81 ppm (v/v)
Boiling point
-19.5°C at 760 mm Hg
Melting point
-92°C
Flash point
60°C; 83°C, closed cup for 37 %, methanol-free
aqueous solution; 50°C closed cup for 37%
aqueous solution with 15% methanol
Explosive limits
73% upper; 7% lower by volume in air
Autoignition temperature
300°C
Solubility
Very soluble in water; soluble in alcohols, ether,
acetone, benzene
Reactivity
Reacts with alkalis, acids and oxidizers
3
4 Sources: American Conference of Governmental Industrial Hygienists (ACGIH) (2002);
5 International Programme on Chemical Safety (IPCS) (2002); Agency for Toxic Substances and
6 Disease Registry (ATSDR) (1999); Gerberich and Seaman (1994); Walker (1975).
7
8
9 Formaldehyde is a chemical intermediate used in the production of plywood adhesives,
10 abrasive materials, insulation, foundry binders, brake linings made from phenolic resins, surface
11 coatings, molding compounds, laminates, wood adhesives made from melamine resins, phenolic
12 thermosetting, resin curing agents, explosives made from hexamethylenetetramine, urethanes,
13 lubricants, alkyd resins, acrylates made from trimethylolpropane, plumbing components from
14 polyacetal resins, and controlled-release fertilizers made from urea formaldehyde concentrates
This document is a draft for review purposes only and does not constitute Agency policy.
2-2 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
(IPCS, 1989). Formaldehyde is used in smaller quantities for the preservation and embalming of
biological specimens. It is also used as a germicide, an insecticide, and a fungicide, as well as an
antimicrobial agent in soaps, shampoos, hair preparations, deodorants, lotions (e.g., suntan lotion
and dry skin lotion), makeup, and mouthwashes, and is present in hand cream, bath products,
mascara and other eye makeup, cuticle softeners, nail creams, vaginal deodorants, and shaving
cream (IPCS, 2002; AT SDR, 1999).
Formaldehyde is commonly produced as an aqueous solution called formalin, which
usually contains about 37% formaldehyde and 12-15% methanol. Methanol is added to formalin
to slow polymerization that leads eventually to precipitation as paraformaldehyde.
Paraformaldehyde has the formula (CH20)n where n is 8 to 100. It is essentially a solid form of
formaldehyde and therefore has some of the same uses as formaldehyde (Kiernan, 2000). When
heated, paraformaldehyde sublimes as formaldehyde gas. This characteristic makes it useful as a
fumigant, disinfectant, and fungicide, such as for the decontamination of laboratories,
agricultural premises, and barbering equipment. Long-chain polymers (e.g., Delrin plastic) are
less inclined to release formaldehyde, but they have a formaldehyde odor and require additives
to prevent decomposition (U.S. EPA, 2008).
The major sources of anthropogenic emissions of formaldehyde are motor vehicle
exhaust, power plants, manufacturing plants that produce or use formaldehyde or substances that
contain formaldehyde (i.e., adhesives), petroleum refineries, coking operations, incineration,
wood burning, and tobacco smoke. Among these anthropogenic sources, the greatest volume
source of formaldehyde is automotive exhaust from engines not fitted with catalytic converters
(NEG, 2003). The Toxic Release Inventory (TRI) data for 2007 show total releases of
21.9 million pounds with about half to the air and half to underground injection (EPA TRI
Explorer, http://www.epa.gov/triexplorer/) (U.S. EPA, 2009a).
Formaldehyde is formed naturally in the lower atmosphere during the oxidation of
hydrocarbons (i.e., methane and terpene), which react with hydroxyl radicals and ozone to form
formaldehyde and other aldehydes, as intermediates in a series of reactions that ultimately lead
to the formation of carbon monoxide and carbon dioxide, hydrogen and water. Formaldehyde
can also be formed in a variety of other natural processes, such as decomposition of plant
residues in the soil, photochemical processes in sea water, and forest fires (National Library of
Medicine [NLM], 2001).
Formaldehyde emitted to the ambient air primarily reacts with photochemically generated
hydroxyl radicals in the troposphere or undergoes direct photolysis (IPCS, 2002). Overall, half-
lives for formaldehyde in air can vary considerably under different conditions. Estimates for
atmospheric residence time in several U.S. cities ranged from 0.3 hours under conditions typical
This document is a draft for review purposes only and does not constitute Agency policy.
2-3 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
of a rainy winter night to 250 hours under conditions typical of a clear summer night (assuming
no reaction with hydroperoxyl radicals). Given the generally short daytime residence times for
formaldehyde, there is limited potential for long-range transport (IPCS, 2002). In cases where
organic precursors are transported long distances, however, secondary formation of
formaldehyde may occur far from the anthropogenic sources of the precursors.
Formaldehyde is released to water from the discharges of both treated and untreated
industrial wastewater from its production and from its use in the manufacture of formaldehyde-
containing resins (ATSDR, 1999). Formaldehyde is also a possible drinking-water disinfection
by-product from the use of ozone and/or hydrogen peroxide. In water, formaldehyde is rapidly
hydrated to form a glycol, and the equilibrium favors the glycol.
2.3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
General population exposure to formaldehyde can occur via inhalation, ingestion and
dermal contact. Each of these pathways and associated media levels are discussed below.
Formaldehyde exposure can also occur occupationally via three main scenarios:
• The production of aqueous solutions of formaldehyde (formalin) and their use in the
chemical industry (e.g., for the synthesis of various resins, as a preservative in medical
laboratories and embalming fluids, and as a disinfectant).
• Release from formaldehyde-based resins in which it is present as a residue and/or
through their hydrolysis and decomposition by heat (e.g., during the manufacture of
wood products, textiles, synthetic vitreous insulation products, and plastics). In general,
the use of phenol-formaldehyde resins results in much lower emissions of formaldehyde
than those of urea- based resins.
• The pyrolysis or combustion of organic matter (e.g., in engine exhaust gases or during
firefighting) (IARC, 2006).
Industries with the greatest potential for exposure include health services, business
services, printing and publishing, manufacture of chemicals and allied products, manufacture of
apparel and allied products, manufacture of paper and allied products, personal services,
machinery (except clerical), transport equipment, and furniture and fixtures (IARC, 1995).
This document is a draft for review purposes only and does not constitute Agency policy.
2-4 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
2.3.1. Inhalation
The most current ambient air monitoring data for formaldehyde come from EPA's air
quality system database (EPA's AirData Web site: http://www.epa.gov/air/data/index.html)
(U.S. EPA, 2009b). These data have been collected from a wide variety of sources, including
state and local environmental agencies, but have not been collected from a statistically based
survey. The most recent data, for the year 2007, come from 188 monitors located in 33 states as
shown in Figure 2-2 (U.S. EPA, 2008). The annual means for these monitors range from 0.7-
45.03 (J,g/m3 (0.56-36.31 ppb) and have an overall average of 3.44 |ig/m3 (2.77 ppb). The annual
means are derived by EPA by averaging all available daily data from each monitor. Table 2-2
shows a breakout of the data by land use category based on the annual means from each monitor
for 2005, 2006, and 2007. The land use is established on the basis of the most prevalent land use
within 0.25 miles of the monitor. The mobile category (land near major highways or interstates
such that it is primarily impacted by mobile sources) has the highest mean levels, and
agricultural lands have the lowest.
Monitor Locator Map
United States
Hazardous Air Pollutants
Shaded states have monitors
HAP Monitoring Site: ~ (175)
Source: US EPA Office of Air and Radiation, AQS Database Tuesday, November 17, 2009
Figure 2-2. Locations of hazardous air pollutant monitors.
Dasgupta et al. (2005) measured formaldehyde levels in 5 U.S. cities during 1999 - 2002.
Samples were collected over approximately a one month period in the spring or summer.
Mean levels were 5.05 ppb in Nashville, TN; 7.96 ppb in Atlanta, GA; 4.49 ppb in
Houston, TX; 3.12 ppb in Philadelphia, PA; and 2.63 in Sydney, FL.
This document is a draft for review purposes only and does not constitute Agency policy.
2-5 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 2-2. Ambient air levels by land use category
2
Formaldehyde Exposure by Category"
Agriculture
Commercial
Forest
Industrial
Mobileb
Residential
Number of data points
17
166
19
61
16
282
Mean ± standard deviation
2.08 ±0.98
3.26 ±2.76
2.79 ±2.17
6.28 ± 14.45
6.84 ±7.28
2.75 ± 1.71
Minimum
0.34
0.20
0.40
0.14
2.02
0.17
Maximum
4.34
20.61
7.33
74.72
23.39
12.35
3
4 "Values are |ig/m\
5 b"Mobile" is ambient air in locations primarily impacted by mobile sources.
6
7 Source: AirData for 2005, 2006, and 2007 (U.S. EPA, 2009b).
8
9
10 Under the National-Scale Air Toxics Assessment (NATA) program, EPA has conducted
11 an emissions inventory for a variety of hazardous air pollutants (HAPs), including formaldehyde
12 (U.S. EPA, 2006c). The NATA uses the emissions inventory data to model nationwide air
13 concentrations/exposures (U.S. EPA, 2006c). The results of the 1999 ambient air concentration
14 modeling for formaldehyde suggest that county median air levels range from 0 to 6.94 |ig/m3 (0
15 - 5.59 ppb) with a national median of 0.56 |ig/m3 (0.45 ppb) (see Figure 2-3). Similar results
16 were found for the year 2002: county concentrations ranged from 0.12 to 9.17 |ig/m3 (0.097 -
17 7.38 ppb) with median of 0.78 |ig/m3 (0.63 ppb). NATA has not provided updated concentration
18 maps for 2002. The 1999 map shows the highest levels in the far west and northeastern regions
19 of the U.S. While these modeling results can be useful, it is important to consider their
20 limitations. Some of the geographical differences result from differences in methods used by
21 states supplying the data. For example, the high levels indicated for Idaho result from the large
22 amount of wood burned during forest fires and the relatively high emission factor that Idaho uses
23 (compared with other states) to estimate formaldehyde emissions from forest fires. A
24 comparison of modeling results from NATA to measured values at the same locations is
25 presented in U.S. EPA (2006c). For 1999, it was found that formaldehyde levels were
26 underestimated at 76% of the sites (n = 68). One possible reason why the NATA results appear
27 low compared to measurements is that the modeling has not accounted for secondary formation
28 of formaldehyde in the atmosphere.
This document is a draft for review purposes only and does not constitute Agency policy.
2-6 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Harp nek
Career
Saeramei
ere on
Distribution of U.S. Ambient Concentrations
Highest In U.S.
95
Percentile
90
75
50
£5
Leweet I n U.S.
6.94-
1.51
1.22
0.34-
0.56
0.37
0.001 5
Count/ Median Ambient Pollutant Concentration
( micrograms / cubic meter)
Source: U.S. EPA / CAQPS
1999 NA1A Hatfonal-Scale A"r Toxics Assessment
1999 Estimated County Median Ambient Concentrations
Formaldehyde — United States Counties
Figure 2-3. Modeled ambient air concentrations based on 1999 emissions.
In general, ambient levels of formaldehyde in outdoor air are significantly lower than
those measured in the indoor air of workplaces or residences (ATSDR, 1999; IARC, 1995).
Indoor sources of formaldehyde in air include volatilization from pressed wood products,
carpets, fabrics, insulation, permanent press clothing, latex paint, and paper bags, along with
emissions from gas burners, kerosene heaters, and cigarettes (NLM, 2001). In general, the major
indoor air sources of formaldehyde can be described in two ways: 1) those sources that have the
highest emissions when the product is new with decreasing emission over time, as with the first
set in the examples above; and 2) those sources that are reoccurring or frequent such as the
second set of examples above. Gilbert et al. (2006) studied 96 homes in Quebec City, Canada
and found elevated levels in homes with new wood or melamine furniture purchased within the
previous 12 months. A summary of indoor data is provided in Table 2-3. Results vary
depending on housing characteristics and date of study.
This document is a draft for re\'iew purposes only and does not constitute Agency policy.
2-7 DRAFT—DO NOT CITE OR QUOTE
-------
1
Table 2-3. Studies on residential indoor air levels of formaldehyde (non-
occupational)
Citation
No. of
Samples
Target Population/House Type
Mean (jig/m3)
Range
Gig/m3)
Gold et al., 1993
Complaint homes
Older conventional homes
<60
24 - 960
Hare et al., 1996
Newly built homes
91
Hare et al., 1996
30 days after installing pressed wood
42 - 540
Gammage and
Hawthorne, 1985
>1200
131
>500
260
Homes with UFFI
Homes without UFFI
Complaint mobile homes
Newer mobile homes
Older mobile homes
60 - 144
30-84
120 - 1080
1032
300
12 - 4080
12 - 204
0 - 5040
Hawthorne et al.,
1986 a,b
18
11
11
40
Conventional homes 0-5 yr
Conventional homes 5-15 yr
Conventional homes >15 yr
Conventional homes overall
96
48
36
72
24 - 480
U.S.EPA, 1987
560
Noncomplaint, conventional,
randomly selected
Noncomplaint, mobile homes,
randomly selected
32-109
109-744
6-576
12-3480
Health Canada and
Environment
Canada, 2001
151
Residential (Canadian) noncomplaint
homes
35
?-148
Zhang etal., 1994
a,b
6
Residential, carpeted, non-smoking
homes
66
42-89
Gilbert et al., 2006
96
Residential (Canadian)
29.5
9.6-90.0
Shah and Singh,
1988
315
Residential & commercial
59
23-89
Stock, 1987
43
Conventional homes
84
96-216
Krzyzanowski et al.,
1990
202
Conventional homes
31
2
3 Note: 1 ppb =1.2 ng/m3
4
5
6 Salthammer et al. (2010) present a thorough review of formaldehyde sources and levels
7 found in the indoor environment. Based on an examination of international studies carried out in
8 2005 or later they conclude that the average exposure of the population to formaldehyde is 20 to
9 40 «g/m3 under normal living conditions. They used the diagram shown in Figure 2-4 to
This document is a draft for review purposes only and does not constitute Agency policy.
2-8 DRAFT—DO NOT CITE OR QUOTE
-------
1 summarize data they found on the range of formaldehyde air concentrations (in ppb) in different
2 environments.
Remote Air
A
Kura
<
al Air
~
Urban/
Jr
Normal Indoor A^r
Polli
ited Indoor Air
b
xtreme Conditions
h
0.1 1 10 100 1000
3
4 Figure 2-4. Range of formaldehyde air concentrations (ppb) in different
5 environments. Source: Salthammer et al. (2010)
6
7
8 Data on formaldehyde levels in outdoor and indoor air were collected under Canada's
9 National Air Pollution Surveillance program (IPCS, 2002; Health Canada and Environment
10 Canada, 2001). The effort included four suburban and four urban sites sampled in the period
11 1990-1998. A Monte Carlo analysis applied to the pooled data (n = 151) was used to estimate
12 the distribution of time-weighted 24-hour air exposures. This study suggested that mean levels
13 in outdoor air were 3.3 |ig/m3 (2.7 ppb) and mean levels in indoor air were 35.9 |ig/m3
14 (29.2 ppb) (Health Canada and Environment Canada, 2001). The simulation analysis also
15 suggested that general population exposures averaged 33-36 |ig/m3 (27-30 ppb).
16 Since the early to mid 1980s, manufacturing processes and construction practices have
17 been changed to reduce levels of indoor formaldehyde emissions (ATSDR, 1999). A 2008 law
18 enacted by the California Air Resource Board (CARB. 2008, Final Regulation Order: Airborne
19 Toxic Control Measure to Reduce Formaldehyde Emissions from Composite Wood Products;
20 http://www.arb.ca.gov/regact/2007/compwood07/fro-final.pdf) has limited the amount of
21 formaldehyde that can be released by specific composite wood products (i.e., hardwood
This document is a draft for re\'iew purposes only and does not constitute Agency policy.
2-9 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
plywood, particle board, and medium density fiberboard) sold, supplied, or manufactured for use
in California. For this reason the mean indoor air levels presented by Health Canada and
Environment Canada (2001) (based on samples collected from 1989-1995) may overestimate
current levels. In addition, the Canadian indoor air data may overestimate formaldehyde levels
in U.S. homes, because many residential homes in Canada use wood burning stoves more
frequently and have tighter construction (due to colder winters), leading to less dilution of indoor
emissions. The outdoor air levels, however, appear to have remained fairly constant over recent
years, and the median outdoor level from the Canadian study (2.8 |ig/m3) (2.3 ppb) is very
similar to the median of the U.S. monitoring data (2.83 |ig/m3) (2.3 ppb) in 1999.
Even though formaldehyde levels in construction materials have declined, indoor
inhalation concerns still persist. For example, recent studies have measured formaldehyde levels
in mobile homes. ATSDR (2007) reported on air sampling in 96 unoccupied trailers provided by
the Federal Emergency Management Agency (FEMA) used as temporary housing for people
displaced by Hurricane Katrina. Formaldehyde levels in closed trailers averaged 1,250 ± 828
|ig/m3 (mean ± standard deviation [SD]) (1.04 ± 0.69 ppm), with a range of 12-4,390 |ig/m3
(0.01-3.66 ppm). The levels decreased to an average of 468 ± 324 |ig/m3 (0.39 ± 0.27 ppm),
with a range of 0.00-1,960 |ig/m3 (0.00-1.63 ppm) when the air conditioning was turned on.
Levels also decreased to an average of 108 ± 96 |ig/m3 (0.09 ± 0.08 ppm), with a range of 12-
588 |ig/m3 (0.01-0.49 ppm) when the windows were opened. ATSDR (2007) found an
association between temperature and formaldehyde levels; higher temperatures were associated
with higher formaldehyde levels in trailers with the windows closed. They also noted that
different commercial brands of trailers yielded different formaldehyde levels.
In December 2007 and January 2008, the Centers for Disease Control and Prevention
(CDC) measured formaldehyde levels in a stratified random sample of 519 FEMA-supplied
occupied travel trailers, park models, and mobile homes ("trailers") (CDC, 2008). At the time of
the study, sampled trailers were in use as temporary shelters for Louisiana and Mississippi
residents displaced by hurricanes Katrina and Rita. The geometric mean level of formaldehyde
in sampled trailers was 95 |ig/m3 (77 ppb), and the range was 3.7-730 |ig/m3 (3-590 ppb).
2.3.2. Ingestion
Limited U.S. data indicate that concentrations in drinking water may range up to
approximately 10 |ig/L in the absence of specific contributions from the formation of
formaldehyde by ozonation during water treatment or from leaching of formaldehyde from
polyacetyl plumbing fixtures (IPCS, 2002). In the absence of other data, one-half this
concentration (5 |ig/L) was judged to be a reasonable estimate of the average formaldehyde in
This document is a draft for review purposes only and does not constitute Agency policy.
2-10 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Canadian drinking water. Concentrations approaching 100 |ig/L were observed in a U.S. study
assessing the leaching of formaldehyde from domestic polyacetal plumbing fixtures, and this
concentration was assumed to be representative of a reasonable worst case (IPCS, 2002).
Formaldehyde is a natural component of a variety of foodstuffs (IARC, 1995; IPCS,
1989). However, foods may be contaminated with formaldehyde as a result of fumigation (e.g.,
grain fumigation), cooking (as a combustion product), and release from formaldehyde resin-
based tableware (IARC, 1995). Also, the compound has been used as a bacteriostatic agent in
some foods, such as cheese (IARC, 1995). There have been no systematic investigations of
levels of formaldehyde in a range of foodstuffs that could serve as a basis for estimation of
population exposure (Health Canada and Environment Canada, 2001). According to the limited
available data, concentrations of formaldehyde in food are highly variable. In the few studies of
the formaldehyde content of foods in Canada, the concentrations were within a range of
<0.03-14 mg/kg (Health Canada and Environment Canada, 2001). Data on formaldehyde levels
in food have been presented by Feron et al. (1991) and IPCS (1989) from a variety of studies,
yielding the following ranges of measured values:
• Fruits and vegetables: 3-60 mg/kg
• Meat and fish: 6-20 mg/kg
• Shellfish: 1-100 mg/kg
• Milk and milk products: 1-3.3 mg/kg
Daily intake of formaldehyde was estimated by IPCS (1989) to be in the range of 1.5-
14 mg for an average adult. Similarly, Fishbein (1992) estimated that the intake of formaldehyde
from food is 1-10 mg/day but discounted this on the belief that it is not available in free form.
Although the bioavailability of formaldehyde from the ingestion of food is not known, it is not
expected to be significant (ATSDR, 1999). Using U.S. Department of Agriculture (USDA)
(1979) consumption rate data for various food groups, Owen et al. (1990) calculated that annual
consumption of dietary formaldehyde results in an intake of about 4,000 mg or approximately
11 mg/day.
2.3.3. Dermal Contact
The general population may have dermal contact with formaldehyde-containing
materials, such as some paper products, fabrics, and cosmetics. For example, nail hardeners
contain formalin, and cosmetics products such as hand creams and suntan lotions contain
formaldehyde-releasing agents (but not formaldehyde) as preservatives. Generally, though,
This document is a draft for review purposes only and does not constitute Agency policy.
2-11 DRAFT—DO NOT CITE OR QUOTE
-------
1 dermal contact is more of a concern in occupations that involve handling concentrated forms of
2 formaldehyde, such as those occurring in embalming and chemical production.
3
This document is a draft for review purposes only and does not constitute Agency policy.
2-12 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
3. TOXICOKINETICS
3.1. CHEMICAL PROPERTIES AND REACTIVITY
Formaldehyde (HCHO) is the smallest aldehyde (30 g/mol) and is a gas at room
temperature. It is highly water soluble and reactive. In water, less than 0.1% of formaldehyde
exists unhydrated, with the majority reported to be in the hydrated form, methylene glycol
(CH2(OH)2) (Priha et al., 1996). Formaldehyde reacts readily with high and low molecular
weight biological constituents.
3.1.1. Binding of Formaldehyde to Proteins
Formaldehyde is a reactive molecule that is likely to react with both low molecular
weight cellular components (e.g., reduced glutathione[GSH]) as well as high molecular weight
components. Unlike deoxyribonucleic acid (DNA), which has some additional barriers to
exposure (i.e., nucleus), extracellular and intracellular proteins are obvious targets for interacting
with formaldehyde. Formaldehyde is a well-known cross-linking agent that is used in the
fixation of tissues, preparation of vaccines, and study of protein-protein interactions (Metz et al.,
2006). However, the exact nature of the protein modifications used for these purposes is not yet
fully characterized (Metz et al., 2006, 2004). Figure 3-1 provides a general reaction scheme for
formaldehyde-mediated modifications of amino acids. In step 1, formaldehyde reacts with
primary N-terminal amines to form a labile methylol adduct. This adduct can undergo
dehydration (step 2) to form an imine, or Schiff base (-N=CH2). Metz et al. (2004) examined
the types of formaldehyde-protein reactions that are likely to occur in vivo by synthesizing
several identical polypeptides with one varying amino acid (X) within the sequence VELXVLL
(V=valine, E=glutamate, L=Leucine, X=varying amino acid). Several peptides with reactive
amino acids did not exhibit modifications, suggesting that the peptide sequence/structure affects
the ability of formaldehyde to react with amino acids. Peptides that were modified indicated
formation of methylol adducts (Figure 3-1, step 1) or a mixture of methylol and imine adducts
(Figure 3-1, step 2).
This document is a draft for review purposes only and does not constitute Agency policy.
3-1 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
[1] Protein—NH2
O
~ A .
H H
Protein—N
H
OH
(Am=+30)
H
/
[2] Protein—N
Protein—N=CH2 + H20
(Am=+30)
(Am=+12)
Protein
Protein
[3] Protein—N=CH2 +¦
Protein^
A
OH
OH
Figure 3-1. Formaldehyde-mediated protein modifications.
Note: Formaldehyde reacts with primary TV-terminal amines to form a methylol
adduct [1], which increases the molecular weight by 30 Da (Am). This labile
adduct can rearrange to form an amine, or Schiff base [2], that results in an
increase in MW of 12 Da. Schiff bases can react with certain amino acids to form
intra- or intermolecular methylene bridges [3], The two amino acids depicted in
step 3 may be within the same protein or possibly from two different proteins.
Source: Metz et al. (2004).
Mucus is composed of water, electrolytes, polysaccharides, and about 0.5% soluble
proteins (Priha et al., 1996; Bogdanffy et al., 1987). Bogdanffy et al. (1987) showed that
although human nasal mucus can bind 70% of 100 mM formaldehyde, irreversible binding of
[14C]-formaldehyde to serum albumin (the major protein in mucus) was shown to be insignificant
after a 1-hour incubation. Irreversible binding (50% or more) did not occur until after about 7
hours of incubation. These data suggest that the protein content of mucus may not provide a
significant formaldehyde irreversible sink. Nonetheless, the solubility of formaldehyde in mucus
along with mucus flow and ingestion likely indicate that much of the inhaled dose is removed—
perhaps as much as 42% in rodents (IARC, 2005; Schlosser, 1999).
This document is a draft for re\'iew purposes only and does not constitute Agency policy.
3-2 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
In general, formaldehyde interacts with proteins. Studies carried out in cell culture media
containing serum and formaldehyde have shown that such mixtures are quite labile. For
example, during a 60-minute incubation of formaldehyde with complete cell media (i.e., with
fetal calf serum) at 38°C, gas chromatography-mass spectrometry (GC-MS) exhibited very
different peak profiles at different points during the incubation (Proctor et al., 1986). In contrast,
GC-MS chromatograms of cell media containing formaldehyde but no serum proteins appeared
relatively unchanged throughout the incubation. Compared to cell culture medium alone,
complete media were considered to provide a more suitable model for the hypothetical
interactions that formaldehyde could undergo in vivo (including perhaps blood).
3.1.2. Endogenous Sources of Formaldehyde
Endogenous formaldehyde is produced through a) normal cellular metabolism through
enzymatic or non-enzymatic reactions, and also as a detoxification product of xenobiotics during
cellular metabolism:
3.1.2.1. Normal Cellular Metabolism (Enzymatic)
Formaldehyde is produced during normal metabolism of methanol, amino acids (e.g.,
glycine, serine, methionine), choline, dimethylglycine, and methylamine and through the folate-
dependent endogenous one-carbon pool etc.
i) One of the endogenous sources for formaldehyde production is methanol,
formed during normal cellular metabolism. However, this fraction may also be derived
through consumption of fruits, vegetables and alcohol (Shelby et al . 2004; IPCS, 1997).
In studies conducted with healthy humans whose diet was devoid of methanol-containing
or methanol-generating foods (such as cereals containing aspartame, a precursor of
methanol) and who abstained from alcohol consumption, the background blood levels of
methanol range from 0.25-4.7 mg/L (Reviewed in Shelby et al 2004 [CERHR]).
Methanol is metabolized to formaldehyde predominantly by hepatic alcohol
dehydrogenase-1 (ADH1) in primates and by ADH1 and catalase (CAT) in rodents,
ADH1 requiring nicotinamide adenine dinucleotide (NAD+) as a cofactor.
ii) Dimethylglycine (DMG), one of the byproducts of choline metabolism
endogenously present in the body, is an indirect source of endogenous formaldehyde.
Two specific dehydrogenases, i) dimethylglycine dehydrogenase (DMGDH) which
converts DMG to sarcosine (methylglycine) and ii) sarcosine dehydrogenase (SDH)
which converts sarcosine to glycine, have been shown to non-covalently bind to the
This document is a draft for review purposes only and does not constitute Agency policy.
3-3 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
folate enzyme, tetrahydrofolate (THF). Further, these dehydrogenases form "active
formaldehyde" by removing the 1-carbon groups from THF (Binzak et al., 2000).
iii) Another source of endogenous formaldehyde is methylamine (MA), an
intermediary component of the metabolism of adrenaline, sarcosine, creatine, lecithin,
and other dietary sources (Yu and Zuo 1996). The enzyme semicarbozole-sensitive
amine oxidase (SSAO), predominantly present in the plasma membrane of endothelial
smooth muscle cells and in circulating blood, converts methylamine to formaldehyde,
hydrogen peroxide and ammonia. The formaldehyde thus released has been shown to
cause endothelial injury eventually leading to atherosclorosis (Kalasz, 2003). Yu et al.
(1997) have shown that adrenaline, released in the body as a response to stress, is known
to be deaminated by the enzyme monoamine oxidase, with further conversion of
methylamine to formaldehyde by SSAO (Yu et al . 1997). Creatine is another precursor
for methylamine which is metabolized by SSAO to form formaldehyde. It has been
shown that short-term, high-dose dietary supplementation of creatine in healthy humans
causes a significant increase in urinary methylamine and formaldehyde levels (Poortmans
et al., 2005).
iv) Endogenous formaldehyde is also a constituent of the one-carbon pool, a
network of interrelated biochemical reactions that involve the transfer of one-carbon
groups from one compound to another (usually the transfer of the hydroxymethyl group
of serine to tetrahydrofolic acid).
Tyihak et al. (1998) have demonstrated that formaldehyde, but not the methyl radical or
methyl cation, is involved in the enzymatic transmethylation and demethylation reactions, and
suggested the presence of a formaldehyde cycle in cells for the production and removal of
formaldehyde utilizing the transfer through methionine -> S-adenosylmethionine S-adenosyl-
homocysteine -> homocysteine (Tyihak et al., 1998). However, these studies did not clearly
show whether the formaldehyde released in this cycle is in free or bound form.
Formaldehyde has been shown to be produced in normal and leukemic leukocytes from
N5-methyl-THF by enzymatic degradation (Thorndike and Beck, 1977). This is a two-step
reaction involving 1) enzymatic conversion of the methyl-THF to formaldehyde followed by 2)
nonenzymatic reaction of formaldehyde with an amine. Thorndike and Beck (1977) showed that
leukocyte (granulocyte and lymphocyte) cell extracts from normal individuals and patients with
chronic lymphocytic leukemia (CLL) or chronic myelocytic leukemia (CML) incubated with
14C-methyl-THF and saturating amounts of tryptamine produced free formaldehyde which is
detected as its corresponding carboline derivative formed with tryptamine. These results
This document is a draft for review purposes only and does not constitute Agency policy.
3-4 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
demonstrate the activity of the enzyme N5, N10-methylene THF reductase which oxidizes N5-
methyltetrahydrofolate to N5, N10 methylene THF. The authors noted that the enzyme levels
were in the order of normal granulocytes < normal lymphocytes < granulocytes from a CML
individual < lymphocytes from a CLL individual (Thorndike and Beck, 1977), suggesting
increased activity of formaldehyde producing enzyme in leukemic cells compared to normal
leukocytes. Overall, formaldehyde might be a byproduct as well as an intermediary product in
several of these reactions.
3.1.2.2. Normal Metabolism (Non-Enzymatic)
i) Formaldehyde can also be formed non-enzymatically by the spontaneous reaction of
methanol with hydroxyl radicals, wherein cellular hydrogen peroxide is the precursor for
hydroxyl radicals generated through Fenton reaction (Cederbaum and Qureshi, 1982).
ii) Another mechanism of nonenzymatic production of formaldehyde is through lipid
peroxidation of polyunsaturated fatty acids (PUFA) (Shibamoto, 2006; Slater, 1984). In this
mechanism, reactive oxygen species (ROS) generated during oxidative stress abstract a hydrogen
atom from a methylene group of polyunsaturated fatty acids (PUFA) in cell membranes causing
autooxidation of lipids with the eventual production of free radicals (e.g., peroxy radical). It is
known that a certain level of oxidative stress and lipid peroxidation does occur in normal
individuals, and these cellular metabolic processes are likely to contribute to endogenous
formaldehyde production.
3.1.2.3. Exogenous Sources of Formaldehyde Production
Microsomal cytochrome P450 enzymes catalyze oxidative demethylation of N-, O- and
S-methyl groups of xenobiotic compounds whereby formaldehyde is produced as a primary
product, which is subsequently incorporated into the one-carbon pool by reacting with
tetrahydrofolic acid or is oxidized to formate (Dahl and Hadley, 1983; Heck et al., 1982). Also,
some special peroxidases, such as peroxide-dependent horseradish peroxidase enzymatically
catalyze xenobiotics to generate formaldehyde in the body. In particular, an ethyl peroxide-
dependent horseradish peroxidase has been shown to act on A'A'-di methyl aniline and produce
equimolar amounts of A-methylaniline and formaldehyde (Kedderis and Hollenberg, 1983).
The tobacco-specific nitrosamine nitrosamine 4-(methylnitrosamino)-l-(3-pyridyl)-butanone
(NNK) is another source of formaldehyde. It has been shown that formaldehyde is also
produced during the methyl hydroxylation of NNK by rat liver microsomes (Castonguay et al.,
1991). Also recent studies have demonstrated the formation of formaldehyde-DNA adducts in
This document is a draft for review purposes only and does not constitute Agency policy.
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
NNK-treated rats using a highly sensitive liquid chromatography-electrospray ionization-tandem
mass spectrometry with selected reaction monitoring (Wang et al., 2007), suggesting formation
of formaldehyde from nitrosamines. Cigarette smoke is also a source of exogenously produced
methylamine which is converted to formaldehyde by SSAO (Yu, 1998).
3.1.2.4. FA-GSH Conjugate as a Method of Systemic Distribution
Formaldehyde is primarily metabolized by alcohol dehydrogenase (ADH3) which uses
the formaldehyde-glutathione hemiacetal adduct as the substrate. Sanghani et al. (2000) have
shown that due to high circulating concentrations (50-fold) of glutathione in human blood, the S-
(hydroxymethyl)glutathione (HMGSH) adduct, the nonenzymatic product of formaldehyde with
glutathione is the major form of formaldehyde seen in vivo (Sanghani et al., 2000). It is likely
that the reversibly bound HMGSH may be transported to different tissues through circulation,
but, specific experimental evidence is lacking.
3.1.2.5. Metabolic Products of FA Metabolism (e.g., Formic Acid)
Formate is converted to carbon dioxide (CO2) in rodents predominantly by a folate-
dependent enzyme pathway (Dikalova et al., 2001). Formate is also oxidized to CO2 and water
by a minor pathway involving catalase located in rat liver peroxisomes (Waydhas et al., 1978;
Oshino et al., 1973). In the folate-dependent pathway, tetrahydrofolate (THF)-mediated
oxidation of formate and the transfer of one-carbon compounds between different derivatives of
THF has been described.
Endogenous levels of formate also will be affected by dietary intake of methanol-
producing or methanol-containing diets since methanol is initially converted to formaldehyde
and eventually metabolized to formate. It has been shown in several studies in human subjects
who were restricted on consuming methanol producing diets, aspartame or alcohol, that the
endogenous blood concentrations of formate ranged from 3.8 to 19.1 mg/L (Shelby et al 2004
[CERHR]). The biological half life of formic acid is 77-90 min (Owen et al., 1990b). The levels
of formate in the urine of unexposed individuals range from 11.7 to 18 mg/L (Boeniger, 1987).
One source of formic acid intake is through diet which ranges from 0.4 to 1.2 mg per day
(Boeniger 1987). The half life for plasma formate is -30 minutes or longer (Boeniger 1987).
3.1.2.6. Levels of Endogenous Formaldehyde in Animal and Human Tissues
Heck et al. (1982) estimated that endogenous levels of formaldehyde (free as well as
bound) in rats ranged from 0.05 to 0.5 |imole/g (1.5-15 jug/g) of wet tissue as analyzed by the
stable isotope dilution with GC-MS method (Heck et al., 1982). Although the levels of free
This document is a draft for review purposes only and does not constitute Agency policy.
3-6 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
formaldehyde cannot be measured due to their high reactivity and short half life, they were
calculated by Heck et al. (1985) using an indirect method. They added a molar excess of GSH or
THF to the test tube containing formaldehyde in aqueous solution enabling complete binding.
When estimated, they observed that the amount of formaldehyde detected was equal to the total
amount added to the reaction suggesting that the formaldehyde measured contained both free and
bound forms. Further, they calculated the free formaldehyde concentration using the
dissociation constant of the HMGSH adduct and cellular concentration of GSH. Human
formaldehyde dehydrogenase has been shown to have a dissociation constant of 1.5 mM for the
formaldehyde-GSH hemithioacetal adduct (Uotila and Koivusalo 1974), while the folate enzyme
product N5,N10-methylene-THF has a dissociation constant of 30 mM (Kallen and Jencks 1966b;
a). This could be evaluated using the Michaels-Menton constant (Km) of formaldehyde
dehydrogenase for the GSH adduct (~4 |iM at 25°C), whereby they calculated the free
formaldehyde level to be around 3-7 |iM or 1-2% of the total formaldehyde as measured by GC-
MS in rat tissues (Heck 1982).
Cascieri and Clary (1992) estimated the total body content of formaldehyde in human
body based on the following assumptions. For an individual with an average body wt of 70 kg
and with body fluids accounting for 70% of body weight, total formaldehyde content is
distributed in -49 kg of body mass or 49 L of body fluids, owing to the water solubility and
uniform distribution of formaldehyde in body fluids. It has been shown that the average blood
concentration (mean ± S.E.) of formaldehyde in unexposed rats and humans was 2.24 ± 0.07 and
2.61 ± 0.14 |ig/g of blood, respectively (Heck et al., 1985), and in unexposed rhesus monkeys it
was 2.42 ± 0.09 |ig/g of blood (Casanova et al., 1988), overall giving an average of
approximately 2.5 ppm (2.5 mg/L) formaldehyde across the species. All these studies used
pentafluorophenyl hydrazine derived formaldehyde using GC-MS analysis (Table 3-1).
Assuming these values, the body content of total formaldehyde is 122.5 mg (49 L x 2.5 mg/L) or
1.75 mg/kg body wt at any given time. Formaldehyde given intravenously to rhesus monkeys
has been shown to have a half life of -1.5 min in blood, wherein formaldehyde in blood was
measured by the dimedone method (McMartin et al., 1979). Using this information Cascieri and
Clary (1992) calculated that the human body generates approximately 40.83 mg/min [(122.5 mg
/2 x 1.5] of formaldehyde. Biotransformation of formaldehyde to carbon dioxide in the liver
alone has been estimated at 22 mg/minute (Owen et al., 1990a).
Free formaldehyde is detected in body fluids and tissues using dimedone (Szarvas et al.,
1986) or 2,4-dinitrohenyhydrazine (DNPH) or pentafluoropheyl hydrazine (PFPH) derivative
(Heck et al., 1985) or as a fluorescent derivative (Luo et al. 2001) as trapping agent and detected
by analytical techniques such as thin-layer chromatography (TLC), high-performance liquid
This document is a draft for review purposes only and does not constitute Agency policy.
3-7 DRAFT—DO NOT CITE OR QUOTE
-------
1 chromatography (HPLC) and gas-chromatography mass spectrometry (GC-MS). Data from
2 several studies is summarized in Table 3-1. Using 14C-labeled dimedone, a chemical which
3 condenses with free formaldehyde forming a product termed "formaldemethone" enabling
4 radiometric detection, Szarvas et al (1986) estimated the levels of endogenous formaldehyde in
5 human blood plasma to be 0.4-0.6 |ig/mL and in human urine to be 2.5-4 |ig/mL (Szarvas et al .
6 1986).
7
8 Table 3-1. Endogenous formaldehyde levels in animal and human tissues
9 and body fluids
10
Tissue
Method
Detected as
Formaldehyde levels
Reference
Not
specified
Not specified
Not specified
0.003-0.012 ppm
(3-12 ng/g)
Hileman 1984
Not
specified
GC-MS with stable
isotope dilution method
As PFPH-
derivative
1.5-15 ppm
(0.05-0.5 (imole/g)
Heck et al 1982a
Blood
GC-MS with select ion
monitoring
As PFPH-
derivative
2.24 ± 0.07 ppm
(2.24 ±0.07 ng/g)
Heck et al 1985
Blood
GC-MS with select ion
monitoring
As PFPH-
derivative
2.61 ± 0.14 ppm
(2.61 ±0.14 ng/g)
Heck et al 1985
Plasma
Reverse phase HPLC-
fluore scent detection
As product of
ampicillin
1.65 ppm
(1.65 (.ig/mL)
Luo et al 2001
Heart
perfusate
HPLC
As DNPH adduct
0.089 - 0.126 ppm
(2.98 - 4.21 nmol/mL)
Shibamoto 2006
Blood
GC-MS with select ion
monitoring
As PFPH-
derivative
2.42 ± 0.09 ppm
(2.42 ± 0.09 ng/g)
Casanova et al
1988
Plasma
Radiometric method
As formalde-
methone adduct
0.4 - 0.6 ppm
(0.4 - 0.6 (ig/mL)
Szarvas et al 1986
Urine
Radiometric method
As formalde-
methone adduct
2.5 - 4.0 ppm
(2.5 - 4.0 (.ig/niL)
Szarvas et al 1986
11
12 Values in the parenthesis, originally cited in the references, are converted to parts per million (ppm) as indicated.
13 PFPH, pentafluorophenyl hydrazone derivative; DNPH, dinitrophenyl hydrazine; GC-MS, gas-chromatography mass
14 spectrometry; HPLC, high performance liquid chromatography.
15
16
17 Hileman (1984) reported that the endogenous levels of metabolically derived
18 formaldehyde will be in the range of 3-12 ng/g of tissue (Hileman 1984). So for an average 70
19 kg individual, the endogenous level of metabolically derived formaldehyde would be 210 jug to
20 840 |ig (3-12 ng/g x 1000 gx 70).
21
This document is a draft for review purposes only and does not constitute Agency policy.
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
35
3.2. ABSORPTION
3.2.1. Oral
Oral absorption of [14C]-formaldehyde (7 mg/kg) in rats resulted in 40% elimination as
14C-carbon dioxide (14C02), with 10% excretion in urine, 1% excretion in feces, and much of the
remaining 49% retained within the carcass, presumably due to metabolic incorporation (IARC,
1995; Buss et al., 1964).
3.2.2. Dermal
Jeffcoat et al. (1983) reported on the disposition of various doses of [14C]-formaldehyde
dermally administered to rats, guinea pigs, and monkeys. Very little (<1% of the applied dose)
of the radiolabel was found in the major organs excised during necropsy. As noted by the
authors, the disposition of formaldehyde when administered via the dermal route was markedly
different to that observed when the compound was administered intravenously or
intraperitoneally. In the latter cases, there was much evidence of metabolic activity, and
substantial portions of the load were expired as C02. The difference appeared to be the result of
a reaction of dermally applied formaldehyde with macromolecules at or near the skin surface or
of its evaporation. In general, portions of the load that succeed in entering the circulation
probably do so bound to macromolecules or by incorporation of the radiolabel via the one-
carbon pool. Likewise, Bartnik et al. (1985) who applied [14C]-formaldehyde to the shaved
backs of rats concluded that the overwhelming majority of the formaldehyde load remained
sequestered in the outer layers of skin at or near the site of application. At the end of the various
measurements, approximately 70% of the dose was found in the treated skin, with a marked
localization of the remaining radioactivity in the uppermost layers. This fraction of the load was
considered to be permanently sequestered, most likely as a result of irreversible binding to
macromolecular components.
3.2.3. Inhalation
Studies indicate that the majority of inhaled formaldehyde is absorbed in the upper
respiratory tract (URT) but that the extent of the scrubbing in this region varies significantly
across species. In dogs, nearly 100% of nasally inhaled formaldehyde is absorbed (Egle, Jr.,
1972). Lower respiratory tract (LRT) studies designed to collect formaldehyde via a tube
inserted into the lower trachea revealed that nearly 95% of formaldehyde was absorbed during
the first pass through the upper respiratory tract (Egle, Jr., 1972), an effect observed with
multiple ventilation rates. The rat nasal passages also scrub nearly all of the inhaled
formaldehyde (on average —97%) (Morgan et al., 1986). In computational dosimetry modeling
This document is a draft for review purposes only and does not constitute Agency policy.
3-9 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
based on anatomically realistic representation of the human nasal airways from a single
individual, approximately 90% of inhaled formaldehyde was predicted to be absorbed in the nose
at resting inspiration. As the inspiratory rate increased, this fraction decreased to about 70% at
light exercise and to 58% at heavy exercise conditions (see Figure 1 in Kimbell et al. [2001b]).
The normal human breathing mode during heavy exercise is oronasal (with -54% of airflow
being oral) (ICRP 66, 1994). Consequently, it is estimated that during heavy exercise breathing
(50 L/min) the flux of formaldehyde into tissue (or rate of mass transported per mm2 of tissue
surface area) in the first six to eight generations of the tracheobronchial airways is comparable to
that in the nasal region (Overton et al., 2001).
It is important to note that the computer simulations mentioned above are based on
anatomical representations of a single individual. Significant anatomical variations occur in
human nasal airways. For example, the nasal volumes of 10 adult nonsmoking subjects between
18 and 50 years of age in a study in the U.S. varied between 15 and 60 mL (Santiago et al.,
2001), and disease states can result in considerable further variation (Singh et al., 1998).
Species differences in kinetic factors have been argued to be the key determinants of
species-specific lesion distributions for formaldehyde and other reactive inhaled gases. Airway
geometry is an important determinant of inhaled-formaldehyde dosimetry in the respiratory tract
and its differences across species. These issues will be discussed in a later section on dosimetry
modeling.
3.2.3.1. Formaldehyde Uptake Can Be Affected by Effects at the Portal of Entry
Certain formaldehyde-related effects have the potential to modulate its uptake and
clearance. The mucociliary apparatus of the upper respiratory tract is the first line of defense
against airborne toxins. Comprising a thick mucus layer (epiphase), hydrophase, and a ciliated
epithelium, the mucociliary apparatus may entrain, neutralize, and remove particulates and
airborne chemicals from inspired air. As reviewed by Wolfe (1986), 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). Degradation in
the continuity or function of this mucociliary apparatus could result in a lower clearance of
inhaled pollutants at the portal of entry.
Morgan et al. (1983) first reported defects in mucociliary function in F344 rats exposed
to 15 ppm formaldehyde 6 hours/day for 1-9 days. Mucostasis occurred in several regions in all
rats after a single 15 ppm exposure. Ciliastasis occurred with greater frequency and across more
regions of the nasoturbinate in subsequent days of exposure. The authors observed that
mucostasis preceded ciliastasis in most cases, and vigorous ciliary activity was noted in areas
This document is a draft for review purposes only and does not constitute Agency policy.
3-10 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
without mucus flow. Morgan et al. (1984a) also studied formaldehyde effects on the mucociliary
apparatus of isolated frog palates in vitro. Mucostasis was evident as mucus became stiff and
eventually rigid with increasing formaldehyde concentration and time of exposure. Ciliary beat
continued even after mucostasis, but ciliastasis ultimately occurred when exposure reached 4 and
9 ppm.
When a rodent is exposed to an irritant, its inhaled dose and pattern of deposition can be
profoundly affected by reflex bradypnea, a protective reflex seen in rodents but not in humans.
Reflex bradypnea can occur when the trigeminal nerve is exposed to a sufficient concentration of
an irritant, such as formaldehyde. It is manifest as markedly decreased activity or prostration,
reduced metabolism, hypothermia (as much as 5°C), significantly reduced respiratory rate and
minute volume, and altered blood and brain chemistry. Because of their small size, rodents are
able to rapidly lower their metabolism and body temperature and therefore their oxygen demand.
The consequence is that their inhaled dose of an irritating chemical is dramatically lowered.
Reflex bradypnea is quantified as the RD50, which is the concentration of a chemical that results
in a 50% decrease in respiratory rate. It can take as much as two hours for rodents to fully
recover from the effects of reflex bradypnea. The clinical manifestations of reflex bradypnea can
easily be misconstrued as toxicity. None of the studies described in this assessment took into
account the fact that reflex bradypnea may have confounded the results. Reflex bradypnea is
discussed in depth in section 4.2.1.1.
Sensory irritation studies suggest that formaldehyde activates the trigeminal nerve by
activating nociceptors through the modification of receptor amino acids, possibly including thiol
groups. Cassee et al. (1996) measured sensory irritation to formaldehyde, acetaldehyde, and
acrolein in male Wistar rats, following a 30-minute nose-only exposure. Formaldehyde and
acrolein elicited similar responses, whereas acetaldehyde was far less irritating. The authors
suggested that the differences in sensitivity to the aldehydes might be explained by differences in
physicochemical properties and by regional differences in activities of detoxifying enzymes for
each chemical. In addition, it has been suggested that acetaldehyde might interact with sensory
nerves via an amino group (Steinhagen and Barrow, 1984), whereas the receptor-binding site for
formaldehyde and acrolein is believed to be a thiol group. Differential binding sites for sensory
irritants in the trigeminal nerve have been reported (Nielsen, 1991).
Sensory irritation effects are discussed in depth in Chapter 4 but are noted here because
stimulation of the trigeminal nerve by formaldehyde can result in significantly lower pulmonary
ventilation, and formaldehyde exposure in rodents at concentrations that approach the RD50.
Barrow et al. (1983) have estimated the "inhaled dose" equivalent to an exposure concentration
of 15 ppm in mice and rats used in the chronic formaldehyde bioassays by Kerns et al. (1983)
This document is a draft for review purposes only and does not constitute Agency policy.
3-11 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
and Monticello and Morgan (1994). Their results indicate that, because mice are observed to
decrease their minute volume by approximately 75% as compared to 45% in rats, a twofold
greater inhaled dose would be expected in rats versus mice. This difference may be relevant to
the increased incidence of squamous cell carcinoma of the nasal cavity in F344 rats as compared
to B6C3F1 mice. Chang et al. (1983) estimated a reduction of 25% in the minute volume of
F344 rats. Yokley et al. (2008) have recently published a model that accounts for physiological
changes in ventilation rate induced by sensory irritation in rats. Thus, the "standard" minute
volumes used for rats and mice need to be adjusted downward when calculating dosimetric
adjustment factors for extrapolation of adverse effects to humans (Thompson et al., 2008). This
question is further discussed in the section on modeling the dosimetry.
Another effect that modulates dosimetry is the dynamic tissue remodeling of nasal
airways that occurs as a consequence of exposure to reactive gases. For example, formaldehyde
dosimetry is influenced by the occurrence of squamous metaplasia, an adaptive tissue conversion
to squamous that occurs in nasal epithelium exposed to toxic levels of formaldehyde. The
metaplasia has been observed to occur in rats at exposure concentrations of 3 ppm and higher
(Kimbell et al., 1997). Squamous epithelium is known to absorb considerably less formaldehyde
than other epithelial types (Kimbell et al., 1997). Overall, the highest flux levels of
formaldehyde in the simulations of the rat nose in Kimbell et al. (2001a) are estimated in the
region just posterior to the nasal vestibule. A consequence of squamous metaplasia would be to
"push" the higher levels of formaldehyde flux toward the more distal regions of the nose
(Kimbell et al., 1997). Subramaniam et al. (2008) discussed this issue further in the context of
uncertainties in the modeling of formaldehyde dosimetry.
3.2.3.2. Variability in the Nasal Dosimetry of Formaldehyde in Adults and Children
Garcia et al. (2009) used computational fluid dynamics to study human variability in the
nasal dosimetry of reactive, water-soluble gases in 5 adults and 2 children, aged 7 and 8 years
old. They considered two model categories of gases, corresponding to maximal and moderate
absorption at the nasal lining. We focus here only on the "maximum uptake" simulations in
Garcia et al. (2009). In this case, the gas was considered so highly reactive and soluble that it
was reasonable to assume an infinitely fast reaction of the absorbed gas with compounds in the
airway lining. Although such a gas could be reasonably considered to represent formaldehyde,
these results cannot be fully utilized to inform quantitative estimates of formaldehyde dosimetry
(and it does not appear to have been the intent of the authors either). This is because the same
boundary condition corresponding to maximal uptake was applied on the vestibular lining of the
nose as well as on the transitional and transitional epithelial lining on the rest of the nose. This is
This document is a draft for review purposes only and does not constitute Agency policy.
3-12 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
not appropriate for formaldehyde as the lining on the nasal vestibule is made of keratinized
epithelium which is considerably less absorbing than the rest of the nose (Kimbell et al. 2001).
The Garcia et al. (2009) study and the results of their analyses have been further
described and evaluated in Appendix 3-1. 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 (1.6-fold difference in average flux and much
less in maximum flux), and the mean values of these quantities were comparable between adults
and children. These results are in agreement with conclusions reached by Ginsberg et al. (2005)
that overall extrathoracic absorption of highly and moderately reactive and soluble gases
(corresponding to category 1 and 2 reactive gases as per the scheme in USEPA [1994]) is similar
in adults and children. On the other hand, Figure 6A of the paper (reproduced as Figure A in
Appendix 3-1), shows significant interhuman variability in flux values at specific points on the
nasal walls. The local flux of formaldehyde varies among individuals by a factor of 3 to 5 at
various distances along the septal axis of the nose.
3.3. DISTRIBUTION
3.3.1. Levels in Blood
Inhalation studies in several species indicate that exposure to formaldehyde does not
result in elevated levels in blood. These studies were carried out over a wide range of exposure
concentrations and durations. Rats exposed to 14 ppm formaldehyde for 2 hours exhibited no
increase in blood formaldehyde levels [2.25 ± 0.07 |ig/(g blood) in treated animals compared
with 2.24 ± 0.07 |ig/(g blood) in control animals] when measured by GC-MS using a stable
isotope dilution technique (Heck et al., 1985, 1982). Similarly, mean formaldehyde blood levels
in humans (n = 6) exposed to 1.9 ppm formaldehyde for 40 minutes in a walk-in chamber (2.77 ±
0.28 |ig/g blood) were not statistically different from measurements in the same population
before exposure (mean of 2.61 ± 0.14 jug/g) (Heck and Casanova-Schmitz, 1984). The
variability in the levels was large. At the individual level, the data showed both increase and
decrease in blood levels relative to pre-exposure levels, which was attributed by the authors as
plausibly due to temporal variations in baseline levels in humans, particularly since the
experiment did not control food intake prior to exposure. Studies in rhesus monkeys have
revealed endogenous formaldehyde levels (2.4 |ig/g blood) comparable to humans and that levels
were also unaltered following exposure to 6 ppm formaldehyde via inhalation 6 hours/day for
4 weeks, measurements being taken at both 7 minutes and 45 hours post final exposure
(Casanova et al., 1988).
This document is a draft for review purposes only and does not constitute Agency policy.
3-13 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
It is important to keep in mind that the GC-MS method is not capable of detecting
irreversibly bound formaldehyde; for example, formaldehyde levels detected by this method,
even in the anterior nasal mucosa of rats exposed to 6 ppm of formaldehyde, were not elevated
over control levels. Furthermore, the GC-MS method does not differentiate between free and
reversibly bound adducts of formaldehyde (Heck et al., 1982). Thus, measured levels represent
total formaldehyde concentration that includes free formaldehyde as well as reversibly bound
adducts. Based on the known Michaelis-Menten constant, Km, for formaldehyde dehydrogenase
with respect to the GSH adduct formation, Heck et al. (1982) estimated under certain
assumptions that free formaldehyde comprised only about 1-2% of the total formaldehyde
measured by their method. Furthermore, as shown by Metz et al. (2006, 2004), formaldehyde
reactions with primary amino and thiol groups can, in a second step, react with many other
amino acids to form stable methylene bridges. Presumably, such reactions would not be
detectable by using the methods employed by Heck et al. (1982).4 Thus, the limited
interpretation of GC-MS measurements of blood levels suggests that formaldehyde does not
appreciably reach the blood, is rapidly metabolized or interacts with macromolecules when it
escapes metabolism, or is otherwise undetected.
Results from an earlier experiment using radiolabeled formaldehyde in rats are consistent
with the conclusion based on the GC-MS measurements of no appreciable increase in blood
levels of formaldehyde. Following a 6-hour exposure of F344 rats to 15 ppm of [14C]-
formaldehyde (Heck et al., 1983), the concentrations of 14C in the nasal mucosa were 28-fold
higher than those in the blood. The observed half-life of the terminal phase of the radioactivity
was long (55 hours); on the other hand, it is known that the half-life of free formaldehyde in the
rat blood is very short. Therefore, the authors concluded that the radioactivity was likely due to
modification of macromolecules or metabolic incorporation rather than slow metabolic clearance
of formaldehyde. The terminal decline of the radioactivity in the packed cell fraction of the
blood was much slower and observed to be consistent with incorporation into erythrocytes.
In the same paper, Heck et al. (1983) report on the similarity in the pharmacokinetics of
radiolabeled formaldehyde and radiolabeled formate in the rat blood, supporting their hypothesis
that oxidation of formaldehyde to formate and subsequent incorporation of this compound
through one-carbon metabolism were major factors in the disposition of formaldehyde. Studies
by Gottschling et al. (1984) have also established that the main product of metabolic clearance of
4 Additionally, note that, although Heck et al. (1982) demonstrated that formaldehyde concentration can be
accurately measured from glutathione and tetrahydrofolate adducts, similar experiments were not performed by using
protein samples or cellular extracts (i.e., in the presence of various amino acids). In addition, standard curves for
predicting formaldehyde concentration in tissues were generated in aqueous solutions rather than biological samples.
This document is a draft for review purposes only and does not constitute Agency policy.
3-14 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
formaldehyde is formate, which is either further metabolized to CO2 and water, incorporated into
the one-carbon pool, and/or eliminated in the urine as a sodium salt at about 13 mg/L urine.
3.3.2. Levels in Various Tissues
The radiolabeling studies indicated high levels of 14C in the rat nasal mucosa (equivalent
concentrations of 14C-formaldehyde in the nasal mucosa of rats naively exposed to 15 ppm
14C-formaldehyde were 2,148 ± 255 nmol/g compared with 76 ± 11 nmol/g in plasma). In
contrast, the GC-MS studies did not detect elevated formaldehyde in this region. This is not to
be interpreted as a discrepancy, because the radiolabeling study did not distinguish among
radiolabeled species and thus the measured radioactivity could potentially be free or bound
formaldehyde, formate, or any [14C] metabolically incorporated into macromolecules.
In concurrent studies, Casanova-Schmitz et al. (1984) resolved the question as to whether
the higher [14C] levels in the nasal mucosa were a consequence of GSH depletion and a
subsequent reduction in GSH-dependent clearance of formaldehyde. An important result in
these studies was that there was no significant difference in labeling in either the nasal mucosa or
in plasma between naive F344 rats and those pre-exposed to unlabeled 15 ppm formaldehyde
6 hours/day for the 9 previous days. These findings indicated little or no apparent effect on the
disposition of formaldehyde following short-term exposure to relatively high levels of
formaldehyde. In contrast, Farooqui et al. (1986) reported decreases in GSH in several tissues
3 hours after a sublethal I.P. injection of formaldehyde but not after 6 and 9 hours. Taken
together, these data suggest that formaldehyde exposure does not result in long-term alterations
in cellular GSH levels and that repeated inhalation exposure does not alter the dosimetry to the
bloodstream or formaldehyde body burden.
Heck et al. (1983) determined the 14C concentrations in different tissues in the F344 rat
body by exposing rats in a head-only chamber to various concentrations (5-24 ppm) of
radiolabeled formaldehyde for 6 hours. (Concentrations of 14C in internal organs and tissues
relative to that in plasma did not appear to vary much as exposure concentrations increased;
therefore only averages over the concentration range were reported.) Except for the esophagus,
levels in the heart, spleen, lung, intestines, liver, and kidney were 1-3 times higher relative to
that in plasma. Labeling in the esophagus was high (fivefold relative to plasma). The authors
attributed this relatively higher dose to mucociliary action in the nose and trachea. The data also
indicate that the brain, testes, and erythrocytes appear to have about threefold lower 14C levels
than plasma. Pre-exposure to formaldehyde (for 9 days) did not alter the measured radioactivity
in the nasal mucosa or plasma. Thus, it was concluded that the single exposure findings may
also be qualitatively extended to chronic exposures.
This document is a draft for review purposes only and does not constitute Agency policy.
3-15 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
The total radiolabel measured in the bone marrow (femur) of F344-rats exposed for 6
hours to 0.3-15 ppm of radiolabeled formaldehyde in the Casanova et al. (1984) experiment was
high (generally within a factor of 0.5 of the total labeling in the nasal respiratory mucosa).
Nearly half of the 14C was contained in the DNA in this tissue presumably on account of the high
rate of cell turnover in the bone marrow, indicating that the carbon derived from 14C-
formaldehyde was utilized for DNA synthesis (Casanova-Schmitz et al., 1984).
Chang et al. (1983) described visceral labeling (via autoradiography) in rats, following
exposure to 15 ppm [14C]-formaldehyde 6 hours/day for 4 days. The authors attributed this
labeling to mucociliary clearance and grooming-related ingestion of formaldehyde.
In summary, following exposure to radiolabeled formaldehyde, the radioactivity was very
high in the nasal mucosa but was also extensively distributed to various tissues. In particular,
levels in the bone marrow were high. On the other hand, formaldehyde levels in the blood
measured by GC-MS were not significantly elevated. Thus, the authors considered it unlikely
that the elevated 14C in various tissues was due to free formaldehyde. Instead, these levels were
thought to arise from either rapid metabolic incorporation or formation of covalent adducts or
incorporation via carboxylation reactions of the 14CC>2 formed during metabolism.
The data presented thus far in this section illustrate that measuring the distribution of the
absorbed formaldehyde based on 14C-radiolabeling and GC-MS studies alone is problematic
because it is difficult to resolve (through these studies) whether it is free, reversibly bound,
irreversibly bound, formate, one-carbon pool, etc. This is of significance with regard to
understanding the availability of the absorbed formaldehyde. More indirect methods had to be
developed to further examine the disposition of formaldehyde; however, as discussed below, the
interpretation of these approaches may also not be straightforward.
3.3.2.1. Disposition of Formaldehyde: Differentiating Covalent Binding and Metabolic
Incorporation
The motivation in presenting this section is twofold, as follows:
1. As concluded above, subsequent studies were necessary to ascertain whether measured
radiolabeling in different experiments was due to formaldehyde adducts or incorporation
of [14C] one-carbon units of formaldehyde into macromolecules via the one-carbon pool.
2. DNA protein cross-links (DPXs) formed by formaldehyde (covalently bound in this case)
have been regarded as a surrogate dose metric for the intracellular concentration of
formaldehyde (Hernandez et al., 1994; Casanova et al., 1991, 1989). This is particularly
relevant because of the nonlinear dose response for DPX formation due to saturation of
This document is a draft for review purposes only and does not constitute Agency policy.
3-16 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
enzymatic defenses at high concentrations (Casanova et al., 1991, 1989). Thus, the
ability to measure DPX is an important development.
An important question is whether the formaldehyde disposed in the form of DPX is
detected in remote tissues. A set of elegant but complex experiments involving dual isotope
labeling (14C and 3H) was carried out to this end by the Heck and Casanova-Schmitz and their
coworkers. Casanova-Schmitz et al. (1984) and Casanova-Schmitz and Heck (1983) used dual
isotope labeling of formaldehyde as a way to partially distinguish between formaldehyde adducts
formation and metabolic incorporation. In separate experiments, F344 rats were exposed to 3H-
and 14C- formaldehyde at different exposure concentrations (0.3-15.0 ppm), and the 3H/14C
ratios of different phases of DNA were measured. Only the highlights of the results and
significant issues are presented here. The overall conclusions from these experiments were as
follows:
• Labeling in the nasal mucosa was due to both covalent binding and metabolic
incorporation.
• DPX was formed at 2 ppm and greater concentrations in the respiratory mucosa.
• In the bone marrow, formaldehyde did not bind covalently to bone marrow
macromolecules at any exposure concentration. The labeling of bone marrow
macromolecules was found to be entirely due to metabolic incorporation and not due to
covalent binding.
Macromolecules such as DNA and protein can be isolated from tissue homogenates by
extraction into three phases: an organic phase consisting of proteins, an aqueous phase consisting
of only double-stranded DNA, and an interfacial phase consisting of both DNA and protein.
Single-stranded (but not double-stranded) DNA was particularly likely to form adducts. DNA
from this interfacial phase can be further purified and has been shown to consist of DPXs
(Casanova-Schmitz and Heck, 1983). Because both [14C]-formaldehyde and [3H]-formaldehyde
can become incorporated into DNA and protein metabolically as well as by cross-linking, the
3H/14C ratio in such cross-linked material should be higher than in material that primarily
contains metabolically incorporated formaldehyde. Figure 3-2 shows the labeling of tissue from
the nasal respiratory mucosa and bone marrow (distal femur) in rats exposed to [14C]-
formaldehyde and [3H]-formaldehyde vapor.
This document is a draft for review purposes only and does not constitute Agency policy.
3-17 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
A Nasal Respiratory Mucosa
B
Bone .Marrow
0
^.
1
o
~:
OX* IIH
a ox* (»oi
.... 0 ?'•>*<•
« ON* UQI
--ox**
[CH20], ppm
[CH20], ppm
Figure 3-2. JH/14C ratios in macromolecular extracts from rat tissues
following
15 ppm).
following exposure to 14C- and 3H-labeled formaldehyde (0.3, 2, 6,10,
Note that the small yield of interfacial (IF) phase from bone marrow tissue precluded
further analysis; this is prima facie evidence for the lack of significant DPXs in this
tissue.
Source: Casanova-Schmitz et al. (1984a).
In the nasal mucosa the interfacial phase has a significantly higher 3H/14C ratio than the
material in the aqueous phase. This suggests that interfacial DNA has significantly more 3H, a
phenomenon likely explained by additional [3H]-formaldehyde molecules present as DPXs prior
to extraction. The amount of interfacial DNA was found to have a clear dose response. These
cross-links were also judged to be due to exogenous formaldehyde. Likewise, the organic phase
of the nasal mucosa showed a similar increase in 3H/14C ratio at higher concentrations, a result
that could be attributed to various inter- and intra-protein adducts (Metz et al., 2004; Trezl et al.,
2003; Skrzydlewska, 1996).
In contrast, analysis of macromolecules at the distal femur location presents a different
pattern (Figure 3-2, part B). First, the interfacial phase was not detected during extraction,
suggesting that there were few or no DPXs to be detected. Second, there was no increase in
3H/14C ratio in the organic (i.e., protein) phase as a function of dose. Therefore, it was
concluded that either radiolabeled formaldehyde or formate reached the distal site and was
subsequently incorporated into macromolecules. According to the mechanistic interpretation of
these studies, the quantity plotted on the ordinate in Figure 3-2 (the ratio of 3H/14C between the
This document is a draft for review purposes only and does not constitute Agency policy.
3-18 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
tissue and the exposure gas) should approach unity as metabolism becomes saturated and more
adduct formation occurs, particularly for protein. Indeed, this is what is observed (see
Figure 3-2, part A). In contrast, there is no dose effect in the femur, suggesting that the labeling
at all doses in that tissue may be due to metabolic incorporation and not due to the parent
formaldehyde.
(Note: These data were originally shown in the absence of an analysis of isotope effects
on covalent binding and metabolism. Subsequent studies determined that [3H]-formaldehyde is
oxidized less rapidly than [14C]-formaldehyde and unlabeled formaldehyde. This suggests that
the 3H/14C ratio, and therefore the amount of formaldehyde covalently bound to tissue, is likely
overestimated because more [3H]-formaldehyde remains unmetabolized, i.e., free to bind [Heck
and Casanova, 1987], The authors hypothesized that this overestimate was relatively greater at
the lower concentrations.)
Similar results were obtained in GSH-depleted rats (Casanova and Heck, 1987). Again,
these authors observed a dose-dependent increase in the 3H/14C ratio in the interfacial DNA and
organic fractions of disrupted cells of the respiratory and olfactory mucosa and no such increases
in bone marrow. Interestingly, at 10 ppm exposure (only), GSH-depleted rats exhibited a higher
3H/14C ratio in the organic phase than did normal rats. Casanova and Heck (1987) posited that
much of the covalent binding at 6 ppm and lower was due to binding to extracellular proteins,
whereas the higher 3H/14C ratio in GSH-depleted rats at 10 ppm was due to more intracellular
binding.
In their first experiment to measure DPX concentrations, Casanova-Schmidt et al. (1984)
and Casanova and Heck (1987) used the dual isotope method (3H/14C) mentioned above. In this
experiment, DPX was observed only at formaldehyde concentrations >2 ppm. Subsequently,
Casanova et al. (1989) developed a more sensitive method using high-performance liquid
chromatography (HPLC) for measuring DPX. In this method, tissue homogenates were digested
with a proteolytic enzyme and extracted with a phenolic solvent. DPX was detected in the nasal
mucosa of rats at formaldehyde concentrations as low as 0.3 ppm. This method was also used to
measure DPX in the nasal region, the larynx, trachea and carina, and major intrapulmonary
airways (airway diameters > 2mm) of rhesus monkeys exposed for 6 hours to 0.7, 2.0 and 6.0
ppm of formaldehyde. DPX was detected in the nose (including the nasopharynx) at all
concentrations and at 2.0 and 6.0 ppm in the larynx, trachea, carina and other lower airways.
However, DPX was not detectable in the bone marrow of these monkeys at any concentration.
Overall, Heck and Casanova-Schmitz and their coworkers interpreted the results of these
various experiments to mean that inhaled formaldehyde could not reach distant sites in the body.
It may be noted in this context that Shaham et al. [1996] reported elevated DPX levels in the
This document is a draft for review purposes only and does not constitute Agency policy.
3-19 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
white blood cells of laboratory workers exposed to formaldehyde. These data are further
reported in Chapter 4.)
3.4. METABOLISM
Formaldehyde is primarily metabolized by glutathione-dependent formaldehyde
dehydrogenase (FALDH) and aldehyde dehydrogenases (ALDHs). Numerous studies now
recognize FALDH as a member of the alcohol dehydrogenase (ADH) family, specifically ADH3
(Thompson et al., 2009; Liu et al., 2004, 2001; Hedberg et al., 2003; Fteog et al., 2003; and the
references in each of these). The remainder of this report will refer to FALDH as ADH3.
3.4.1. In Vitro and In Vivo Characterization of Formaldehyde Metabolism
Formaldehyde is oxidized to formate by two metabolic pathways (Figure 3-3). The first
pathway involves conversion of free formaldehyde to formate by the so-called low-Km
(Km = 400 |iM) mitochondrial aldehyde dehydrogenase-2 (ALDH2). The second pathway
involves a two-enzyme system that converts glutathione-conjugated formaldehyde
(S-hydroxymethylglutathione [HMGSH]) to the intermediate S-formylglutathione, which is
subsequently metabolized to formate and GSH by S-formylglutathione hydrolase.
ADH1
Methanol ______ >»
NAD+ NADH
GSH
NAD+
ALDH2
HMGSH
ADH3
NADH
GSH
NADH
Formate ^
S-fbrmyl GSH
S-formyl GSH hydrolase
Figure 3-3. Formaldehyde clearance by ALDH2 (GSH-independent) and
ADH3 (GSH-dependent).
The Km value for ALDH2 and free formaldehyde is about 400 |iM (Teng et al.,
2001), whereas the Km value for HMGSH and ADH3 is 6.5 |iM (Uotila and
Koivusalo, 1974a,b). The ADH-mediated reactions are reversible in the presence
of excess reduced nicotinamide adenine dinucleotide (NADH).
Source: Adapted from Teng et al. (2001).
This document is a draft for review purposes only and does not constitute Agency policy.
3-20 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
Though ADH3 is rate limiting in this second pathway, the affinity of HMGSH for ADH3
(Km = 6.5 |iM) is about 100-fold higher than that of free formaldehyde for ALDH2. In addition
to the kinetic properties, this member of the ADH gene family (H00g et al., 2003, 2001; Liu et
al., 2001; Jornvall et al., 2000; Estonius et al., 1996) appears to be ubiquitously expressed in
organ tissues (Molotkov et al., 2002; Ang et al., 1996a, b), exhibits cytoplasmic and nuclear
localization (Fernandez et al., 2003), and is the most abundant ADH family member in the liver
and brain (Gaiter et al., 2003).
In vitro studies have examined the clearance of formaldehyde in several human and rat
tissues (Table 3-2). Examination of formaldehyde metabolism in the rat nasal and olfactory
mucosa indicates nearly identical pharmacokinetics in the rat liver on a per mg of cell lysate
basis (Casanova-Schmitz et al., 1984b). Similar results have been obtained in the absence of
GSH, where other ALDH family members oxidize formaldehyde, albeit with significantly lower
affinity (i.e., higher Km). Hedberg et al. (2000) demonstrated that human buccal tissue lysate
kinetics are in close agreement with those reported for purified human liver ADH3 (Uotila and
Koivusalo, 1974a). Additionally, micro-array analysis indicates that these cells express far more
ADH3 and S-formylglutathione hydrolase than ALDH1 or ALDH2 (Hedberg et al., 2001a). The
results of Ovrebo et al. (2002) are not easily compared with the other studies in Table 3-2
because these studies were in intact cell cultures. However, it is apparent that the
pharmacokinetic values in these human cells are comparable to intact rat liver cells.
Table 3-2. Formaldehyde kinetics in human and rat tissue samples
Source
Km (uM)
v
* max
(nmol/mg protein x min)
Reference
Purified human liver ADH3
6.5
2.77 + 0.12
Uotila and Koivusalo (1974a, b)
Rat olfactory mucosa (+ GSH)
2.6 + 0.5
1.77 + 0.12
Casanova-Schmitz et al. (1984b)
Rat olfactory mucosa (- GSH)
647 + 43
4.39 + 0.14
Casanova-Schmitz et al. (1984b)
Rat respiratory mucosa (+ GSH)
2.6+2.6
0.90 + 0.24
Casanova-Schmitz et al. (1984b)
Rat respiratory mucosa (- GSH)
481+88
4.07 + 0.35
Casanova-Schmitz et al. (1984b)
Rat liver (+ GSH)
5.0 + 1.9
2.0 + 0.3
Casanova-Schmitz et al. (1984b)
Human bronchial explants3
5,100
3.3
Ovrebo et al. (2002)
Human bronchial epithelial3
1,400
6.1
Ovrebo et al. (2002)
Rat hepatocytes3
1,250
4.2
Ovrebo et al. (2002)
Human buccal tissue (+ GSH)
11+2
2.9 + 0.6
Hedberg et al. (2000)
Human buccal tissue (- GSH)
360 + 90
1.2 + 0.7
Hedberg et al. (2000)
Human keratinocytes
n.d.b
14.5 + 1.8
Hedberg et al. (2000)
Human fibroblasts
ad.
17.9 + 1.4
Hedberg et al. (2000)
"These studies were carried out in intact cells by measuring the formation of formate. This likely explains the nearly
1,000-fold increase in apparent Km, since much of the formaldehyde was likely to be bound extracellularly. The
remaining studies used either purified enzyme or cell lysates (as indicated) and measured the formation of NADH.
bn.d. = not determined.
This document is a draft for review purposes only and does not constitute Agency policy.
3-21 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
The data in Table 3-3 along with data indicating the ubiquity of ADH3, indicate that
many human tissues and cells, particularly in the respiratory tract, appear to exhibit significant
capacity to metabolize formaldehyde. Molecular biology techniques have demonstrated the
importance of ADH3 in formaldehyde clearance. For example, ADH-knockout studies have
shown that the median lethal dose (LD50) values for formaldehyde in wild type, ADH1
ADH3 , and ADH4 mice strains were 0.200, 0.175, 0.135, and 0.190 g/kg, respectively
(Deltour et al., 1999). Although the statistical significance was not reported, the data indicate
that deletion of ADH3 increases the sensitivity of mice to formaldehyde.
The pharmacokinetics of formate are complex. Formate can undergo adenosine
triphosphate (ATP)-dependent addition to tetrahydrofolate (THF), which can carry either one or
two one-carbon groups. Formate can conjugate with THF to form N10-formyl-THF and its
isomer N5-formyl-THF, both of which can be converted to N5,N10-methenyl-THF and
subsequently to other derivatives that are ultimately incorporated into DNA and proteins via
biosynthetic pathways (Figure 3-4).
C0?
H^FSLATf 10-NC0-H^FOlATe HqFOLATF
t
S¦10-CH-Hqfolatf
t
tyotftTf ^i»ii S,i
Sf« fitt ^ bUHP nTKP
HCv mt
Figure 3-4. Metabolism of formate.
Note: 1, formyl-THF synthetase; 2, formyl-THF dehydrogenase.
Source: Adapted from Black et al. (1985).
Elevated levels of formate in urine have been detected following inhalation of methanol
or formate under certain conditions (Liesivuori and Savolainen, 1987), although the
interpretation of this finding is unclear. There is also evidence that formate generates CO2
This document is a draft for review purposes only and does not constitute Agency policy.
3-22 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
radicals and can be metabolized to CO2 via catalase and via the oxidation of N10-formyl-THF
(Dikalova et al., 2001, and references therein). The significance of formate in formaldehyde
toxicity is unclear. Black et al. (1985) reported that hepatic tetrahydrofolate levels in monkeys
are 60% of those in rats and that primates are far less efficient in clearing formate than are rats
and dogs. Studies involving [14C]-formate suggest that about 80% is exhaled as 14CC>2, 2-7% is
excreted in the urine, and about 10% undergoes metabolic incorporation (Hanzlik et al., 2005,
and references therein). Mice deficient in formyl-THF dehydrogenase exhibit no change in LD50
(via I.P. dose) for methanol or in oxidation of high doses of formate (Cook et al., 2001). It has
been suggested that rodents efficiently clear formate via folate-dependent pathways,
peroxidation by catalase, or an unknown third pathway. Conversely, primates do not appear to
exhibit such capacity and are more sensitive to metabolic acidosis following methanol poisoning
(Cook et al., 2001).
3.4.2. Formaldehyde Exposure and Perturbation of Metabolic Pathways
The enzyme ADH3 has received renewed attention in recent years because of new
functions that have been attributed to it. ADH3 is central to the metabolism of formaldehyde;
however, exposure to formaldehyde in turn alters the activity of ADH3 (in multiple dose-
dependent ways), thereby leading to perturbation of critical metabolic pathways. These are
briefly mentioned below (refer to cited papers for details).
1. Exposure to formaldehyde increases cell replication. These proliferating epithelial and
inflammatory cells are rich in both the messenger ribonucleic acid (mRNA) and protein
of ADH3 (Nilsson et al., 2004; Hedberg et al., 2000). Studies in the rodent lung suggest
that increases in ADH3 in such cells dramatically alter the biology of other important
ADH3 substrates that are involved in protein modification and cell signaling (Que et al.,
2005).
2. ADH3 also participates in the oxidation of retinol and long-chain primary alcohols, as
well as the reduction of S-nitrosoglutathione (GSNO) (Staab et al., 2009; Thompson et
al., 2009; Hedberg et al., 2003; H00g et al., 2003; Molotkov et al., 2002; Liu et al., 2001;
Jornvall et al., 2000; Jensen et al., 1998). The activity of ADH3 toward some of these
substrates has been shown to be significantly increased in the presence of formaldehyde.
Staab et al. (2009) showed that (in cultured cells) GSNO can accelerate ADH3-mediated
formaldehyde oxidation and, likewise, that formaldehyde increases ADH3-mediated
GSNO reduction nearly 25-fold. The following effects may be noted with regard to the
relevance of such perturbations.
This document is a draft for review purposes only and does not constitute Agency policy.
3-23 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
a. GSNO is an endogenous bronchodilator and reservoir of nitric oxide (NO)
activity (Jensen et al., 1998). Details on the ADH3-mediated reduction of GSNO
are shown in Thompson and Grafstrom (2008).
b. ADH3 is implicated in playing a central role in regulating bronchiole tone and
allergen-induced hyperresponsiveness (Gerard, 2005; Que et al., 2005).
c. 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."
3.4.3. Evidence for Susceptibility in Formaldehyde Metabolism
Teng et al. (2001) provided evidence that inhibition of ADH1, ALDH2, and ADH3 has
significant impact on formaldehyde toxicity. The authors speculated that deficiencies in any of
these enzymes would confer an increased susceptibility to formaldehyde toxicity (Teng et al.,
2001). Polymorphism in ALDH2 has been shown to have implications in human risk
assessment, specifically with regard to acetaldehyde metabolism (Ginsberg et al., 2002). It is
worth noting, however, that Teng et al. (2001) only demonstrated the importance of ALDH2 in
rat hepatocytes with formaldehyde concentrations of 2.5 mM and greater. Since this
concentration is fivefold greater than the 0.5 mM Km for free formaldehyde, ALDH2
involvement is not unexpected at such high concentrations. Teng et al. (2001) also demonstrated
the importance of ADH1 in driving the reverse reaction (i.e., formaldehyde to methanol) by
coadministration of NADH-generators. This would have the effect of prolonging the life of
formaldehyde by continuous recycling. This is not surprising, given that many ADH reactions
are reversible. However, levels of nicotinamide adenine dinucleotide (NAD+) are normally
much higher than NADH.
To date, two studies have reported polymorphisms in ADH3, using the new
nomenclature.5 ADH3 transcription appears to be regulated by specificity protein (Spl), with a
minimal promoter located at positions -34 to +61. The reported polymorphisms in ADH3
involve four base-pair substitutions in the promoter region and no polymorphisms in the coding
region (Hedberg et al., 2001b). The three polymorphisms include —197/—196 (GG^AA), -79
(G^A), and +9 (C^T). The genotype frequencies are shown in Table 3-3. Of these alleles, the
+9 (C^T) polymorphism (in the putative Spl minimal promoter region) reduced transcriptional
5 Other epidemiologic studies investigating links between ADH3 and oral cancer use the older nomenclature and
thus refer to Class I ADH (i.e., ADH1) enzymes.
This document is a draft for review purposes only and does not constitute Agency policy.
3-24 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
activity twofold in in vitro reporter gene experiments. According to Hedberg et al. (2001b), no
studies have demonstrated differences in ADH3 enzyme activity in humans. More recently,
single nucleotide polymorphisms in ADH3 have been reported to be associated with childhood
risk of asthma, although the functional relevance of these polymorphisms has not been published
(Wu et al., 2007).
Table 3-3. Allelic frequencies of ADH3 in human populations
Allele frequencies (%)
Population, n
AA_t97/_i96
GG.1Q7/.196
A_79
G.79
T+9
C+9
Chinese, 83
22
78
100
-
—
100
Spanish, 95
41
59
62
38
—
100
Swedish, 96
47
53
67
33
1.5
98.5
Source: Adapted from Hedberg et al. (2001b).
Alterations in THF pathways may also have an impact on formaldehyde toxicity. These
could result from polymorphisms in various enzymes or differences in folate intake and
absorption. Species differences in tetrahydrofolate levels (Black et al., 1985) are thought to play
a role in the differential responses to methanol across species. Cook et al. (2001) speculate that
rats have redundant pathways for formate clearance that may be absent or less efficient in
primates.
3.5. EXCRETION
The main product of metabolic clearance of formaldehyde is formate, which is further
metabolized to CO2 and water, incorporated into the one-carbon pool, or eliminated in the urine.
There is also some evidence that formaldehyde is present in exhaled breath; however, it is
unclear whether this originates from endogenous sources, or is simply a function of ambient
formaldehyde dissolved in fluids lining POEs. The following sections describe first experiments
in laboratory species and then available data in humans. Broadly, these studies address two
important questions that may be of relevance for risk assessment. First, it may be of interest to
know what levels of formaldehyde are exhaled for comparison with inhaled levels, and whether
there is any relationship between external exposure and exhaled levels. Second, there are recent
studies that have attempted to relate genetic polymorphisms and changes in gene transcription
level to levels of putative urinary formaldehyde biomarkers.
This document is a draft for review purposes only and does not constitute Agency policy.
3-25 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
3.5.1. Formaldehyde Excretion in Rodents
Heck et al. (1983) determined the relative contributions of various excretion pathways in
F344 rats following inhalation exposure to formaldehyde. Table 3-4 indicates that the relative
excretion pathways were independent of exposure concentration (at least between 0.63 and
15 ppm). Nearly 40% of inhaled [14C]-formaldehyde appeared to be eliminated via expiration,
probably as CO2 (it should be recalled that nearly 100% of inhaled formaldehyde is absorbed).
Within 70 hours of a 6-hour exposure to formaldehyde, about 17 and 5% were eliminated in the
urine and feces, respectively. Nearly 40% of inhaled [14C]-formaldehyde remained in the
carcass, presumably due to metabolic incorporation.
Table 3-4. Percent distribution of airborne [14C]-formaldehyde in F344 rats
Source
Concentration of formaldehyde (ppm)
0.63
13.1
Distribution (%)a
Expired air
39.4 ± 1.5
41.9 ±0.8
Urine
17.6 ± 1.2
17.3 ±0.6
Feces
4.2 ± 1.5
5.3 ± 1.3
Tissues and carcass
38.9 ± 1.2
35.2 ±0.5
"Values are means ± standard deviations (n = 4).
Source: Heck et al. (1983).
Mashford and Jones (1982) examined elimination pathways of formaldehyde in rats
exposed by I.P. injection. Urine and exhaled gases were collected from rats exposed to 4 or
40 mg/kg [14C]-formaldehyde. At 48 hours postinjection, 82 and 78% of the radiolabel were
exhaled as 14CC>2, whereas exhaled [14C]-formaldehyde was not detected. Mashford and Jones
(1982) also further identified the urinary metabolites. Five hours after injection of the higher
dose, formate was determined to comprise 80% of the urinary metabolites. The authors were
unable to detect cysteine derivatives observed in other studies (see below) in the urine of these
rats prior to or after formaldehyde exposure. The authors stated that if formaldehyde were to be
excreted in urine containing cysteine, then thiazolidine-4-carboxylate (TZCA) would likely be
produced. They speculated that species differences in urinary compounds may produce
formaldehyde conjugates (or artifacts).
Hemminki (1982) reacted formaldehyde and acetaldehyde with cysteine,
N-acetylcysteine, and GSH and found that formaldehyde reacted most rapidly with cysteine to
form TZCA. Similarly, acetaldehyde reacted preferentially with cysteine, albeit slower than
This document is a draft for review purposes only and does not constitute Agency policy.
3-26 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
formaldehyde, to form a thiazolidine derivative. However, when each aldehyde was
administered I.P. (10% formaldehyde, 50% acetaldehyde), thioether concentrations (nmol/mol
creatinine) significantly increased in the 24 and 48 hour urine of acetaldehyde-treated rats but
not formaldehyde-treated rats. These data suggest that formaldehyde is not appreciably excreted
in urine and thus cysteine conjugates are not likely to represent formaldehyde exposure.
Most recently, Shin et al. (2007) attempted to show that formaldehyde inhalation
increased urinary TZCA levels in Sprague-Dawley rats. Treated rats were exposed to 3.1 and
38.1 ppm formaldehyde for 6 hours/day for 2 weeks, and urine was collected for 3 days. The
TZCA level in four control rats was 0.07 ± 0.02 mg/L, whereas levels in the 3 and 38 ppm
groups were 0.18 ± 0.045 and 1.01 ± 0.36, respectively. Notably, the concentrations in the four
highest exposed animals (0.71, 0.70, 1.20, and 1.43 ppm) exhibited a nearly twofold range.
However, these comparisons are confounded if the exposures have any influence on urine
production and urine cysteine levels. The study does not provide any data that might allow one
to examine this issue.
3.5.2. Formaldehyde Excretion in Exhaled Human Breath
Several human and animal studies have attempted to measure the concentration of
formaldehyde in exhaled breath. None of the human studies were designed to distinguish
between exogenous (room air) and endogenous (systemic) formaldehyde in exhaled breath. In
order to discern whether endogenous formaldehyde is excreted into the lungs, test subjects must
breathe formaldehyde-free air. Because subjects were breathing room air, which contained 9-10
ppb formaldehyde in two studies and unspecified concentrations in two other studies, it is
impossible to ascertain whether there was any endogenous formaldehyde in their exhaled breath.
One study demonstrates that the formaldehyde concentration is lower in exhaled breath than in
inhaled breath (Cap et al., 2008). Also, none of the human studies investigated whether there is
any relationship between exhaled formaldehyde levels and food intake, life stage, smoking, or
health status. This assessment identifies a critical research need for further studies on the
measurement of exhaled formaldehyde.
Proton transfer reaction mass spectrometry (PTR-MS) has been applied to measure trace
compounds in exhaled breath including volatile organics and specifically formaldehyde. The
basic method of PTR-MS is based on the transfer of protons from H30+ to gases in exhaled
breath and the in-line monitoring of products where gases are tentatively identified by the mass
to charge ratio (m/z) where a m/z of 31 is consistent with protonated formaldehyde (Hansel et al.,
1995; Lindinger et al., 1998). It is important to note that reaction products from methanol and
ethanol may also produce fragments with an m/z ratio of 31 (Kusch et al., 2008). Up to 1% of
This document is a draft for review purposes only and does not constitute Agency policy.
3-27 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
the mass of ethanol and methanol in exhaled breath may be detected with an m/z ratio of 31 and
thus be identified as formaldehyde, so accurate quantitation of formaldehyde should adjust for
this contribution (Spanel and Smith, 2008). Selected ion flow tube mass spectrometry (SIFT-
MS) is an application of PTR-MS developed for real-time analysis of trace gases in breath
(Smith and Spanel, 2005; Spanel and Smith, 2007).
Moser et al. (2005) measured levels of 179 volatile organic compounds (VOCs) in the
exhaled breath of 344 individuals. This study was not designed to ascertain whether exhaled
formaldehyde is of endogenous origin, but rather to demonstrate that proton transfer reaction-
mass spectrometry can be used as a new method for rapid screening of large collectives for risk
factors (e.g., smoking behavior), potential disease biomarkers, and ambient air characterization.
The study was conducted at a health fair. The test subjects had a mean age of 61.6 years; 63%
were males and 14% were smokers. Samples of room air were collected and evaluated in
parallel with exhaled breath samples. The authors note that formaldehyde was detected in room
air, but did not report the levels; rather they stated that the background concentrations were
negligible. Of the 179 volatile organic compounds measured, data were reported for 14,
including formaldehyde and formic acid. The formaldehyde levels in exhaled breath spanned
from 1.2 to 72.7 ppb with a median of 4.3 ppb and 75th percentile of 6.3 ppb. No explanation
was provided for this wide range in values, and there was no distinction of the data by sex, age,
health, or smoking status.
Because the test subjects were breathing ambient indoor air which contained an
unspecified concentration of formaldehyde, it is impossible to ascertain whether exhaled
formaldehyde was from an endogenous or exogenous source, or both. Moser et al. also note that
significant differences in exhaled breath composition could be found between smokers and non-
smokers for 32 of the 179 chemicals measured, but the 32 chemicals were not named and no
substantiating data were provided. Formaldehyde may have been among these 32 chemicals
since it is a major component of cigarette smoke (NLM, 2001).
The report by Moser et al. does not provide the limit of detection for any of the
compounds measured or details of the analytical method. The minimum formaldehyde level
reported in exhaled breath was 1.23 ppb. The method employed by Moser et al. does not adjust
the detection of apparent formaldehyde {m/z) by accounting for the contribution of methanol and
ethanol. The levels of methanol and ethanol in exhaled breath were reported, however. Based
on 1% of the mass of these chemicals contributing to the m/z = 31 in the PTR-MS method
(Spanel and Smith, 2008), Table 3-5 demonstrates that the highest formaldehyde levels reported
may have been an artifact of high methanol and ethanol in exhaled breath. Because the in-room
formaldehyde concentrations were not reported, it is unknown how much of this formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
3-28 DRAFT—DO NOT CITE OR QUOTE
-------
1 represents formaldehyde levels in inhaled air. Additionally, since these samples were simple
2 exhaled breath samples and not SIFT-MS samples, it is impossible to distinguish between air
3 which reached the pulmonary region versus air which only entered the upper airways.
4
5 Table 3-5. Apparent formaldehyde levels in exhaled breath of individuals
6 attending a health fair, adjusted for methanol and ethanol levels which
7 contribute to the detection of the protonated species with a mass to charge
8 ratio of 31 reported as formaldehyde (m/z = 31)
9
Chemical
Minimum
25th
percentile
Median
75th
percentile
97.5th
percentile
Maximum
Methanol (m/z = 31)
13.367
106.227
161.179
243.185
643.614
1536.499
1% of methanol
predicted as m/z = 31
0.13
1.06
1.61
2.43
6.44
15.36
Ethanol
11.583
23.1
34.664
64.24
549.24
9779.768
1% of ethanol predicted
as m/z = 31
0.12
0.23
0.35
0.64
5.49
97.80
Mass of m/z = 31
attributable to methanol
and ethanol
0.25
1.29
1.96
3.07
11.93
113.16
Mass of m/z = 31
reported as
formaldehyde
1.23
3.1
4.26
6.33
39.8
72.7
10
11 Source: Moser et al. (2005).
12
13
14 Turner et al. (2008) compared levels of volatile compounds in exhaled breath to levels in
15 emissions from the skin in five males (3 in their mid 20s, and the others 42 and 49 years old).
16 The subjects fasted overnight, and measurements were taken before and after ingesting 75 g of
17 glucose. The source of the inhaled air was laboratory air which contained an unreported
18 concentration of formaldehyde. Formaldehyde was not detected in the exhaled breath of any
19 subjects. The limit of detection was 5 ppb or better.
20 Wang et al. (2008) measured the concentrations of formaldehyde and 9 other chemicals
21 in the exhaled breath of three healthy male laboratory workers. The limit of quantification for
22 formaldehyde was not reported. A series of measurements were taken in the nose and mouth,
23 and also in the oral cavity during breath holding. Table 3-6 presents the median formaldehyde
24 levels and geometric standard deviations for the three subjects. The authors noted that
25 formaldehyde in exhaled breath was at a level somewhat lower than the ambient air
26 concentration, but they could not be certain of its origin.
27
28
This document is a draft for review purposes only and does not constitute Agency policy.
3-29 DRAFT—DO NOT CITE OR QUOTE
-------
1 Table 3-6. Measurements of exhaled formaldehyde concentrations in the
2 mouth and nose, and in the oral cavity after breath holding in three healthy
3 male laboratory workers
4
Subject
Formaldehyde
(median ppb / eg)
A
Mouth
5/2.3
Nose
7/2.1
Oral cavity
5/2.3
B
Mouth
7/2.3
Nose
5/2.1
Oral cavity
6/1.9
C
Mouth
4/2.5
Nose
6/1.9
Oral cavity
6/1.9
Laboratory air
9.6 ±1.5
5
6 Source: Wang et al. (2008).
7
8
9 Cap et al. (2008) evaluated relationships between volatile organic compounds measured
10 in exhaled breath and exhaled breath condensate. Exhaled breath condensate consists of
11 aerosolized particles of airway lining fluid evolved from the airway wall by turbulent airflow
12 that serve as seeds for substantial water vapor condensation, which then serves to trap water
13 soluble volatile gases. This study also attempted to ascertain whether the source of each
14 compound was endogenous or exogenous. According to the published article and electronic
15 communication with Dr. Patrik Spanel, a co-author for this study, the limit of quantification was
16 3 ppb or better. Measurements of formaldehyde in the direct exhaled breath of 34 subjects (25 to
17 62 years; 11 males; 2 smokers) varied from 0 to 12 ppb with a mean of 2 ppb and a median of 1
18 ppb. Measurements of formaldehyde in exhaled breath condensate ranged from 0 to 12 ppb with
19 a mean of 2 ppb and a median of 0 ppb. Two smokers exhaled formaldehyde concentrations of 0
20 and 3 ppb. The scatter plot in Figure 3-5 shows that most subjects exhaled formaldehyde
21 concentrations of 0-3 ppb (x axis), which was considerably lower than the ambient air
22 concentration of 9.6 ±1.5 ppb. The authors concluded that the exhaled concentrations were
23 influenced by exogenous levels in room air, which were not controlled. Dr. Spanel offered two
24 plausible explanations for why exhaled formaldehyde levels are lower than inhaled levels. He
25 noted that some subjects were tested within several minutes of entering the laboratory from
26 outdoors, so they may not have had time to acclimate to the higher indoor air concentration.
27 Another explanation is that a substantial portion of inhaled formaldehyde, which is highly
This document is a draft for review purposes only and does not constitute Agency policy.
3-30 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
reactive, was retained in the respiratory tract and thus not exhaled; that is, 9.6 ppb of
formaldehyde was inhaled but only a mean of 2 ppb was exhaled.
In this and other human studies, there was no adjustment for an artifact in the analytical
method that makes it impossible to distinguish between formaldehyde and reaction products for
1% of exhaled methanol and ethanol because they have the same mass to charge ratio (m/z = 31).
In fact, the concentration of methanol and ethanol that is misidentified as formaldehyde exceeds
the reported concentrations of exhaled formaldehyde. Thus, it is highly likely that the actual
exhaled formaldehyde concentration in Cap et al. (2008) was significantly lower than 2 ppb, and
that there was little or no endogenous formaldehyde in the exhaled breath. This would be
consistent with an animal study in which Mashford and Jones (1982) detected no exhaled
formaldehyde in rats injected I.P. with 40 mg/kg [14C]-formaldehyde. Over 48 hours, 78% of the
radioactivity was exhaled as 14CC>2 and 11% was excreted in the urine as formate, N-
(hydroxymethyl)urea, N,N'-bis-(hydroxymethyl)urea, and possibly polymethylene urea. In
summary, there are insufficient data at this time to confidently establish a concentration of
formaldehyde in exhaled breath that can be attributed to endogenous sources.
14
12
10
f 8
CL
U „
CD 6
LU
4
2
0
0 2 4 6 8 10 12 14
direct (ppb)
Figure 3-5. Scatter plot of formaldehyde concentrations measured in ppb in
direct breath exhalations (x axis) and exhaled breath condensate headspace
(y axis)
Source: Cap et al. (2008).
3.5.3. Formaldehyde Excretion in Human Urine
Gottschling et al. (1984) examined urinary formic acid in 35 veterinary students.
Personal monitoring badges were worn and returned after class, and urine samples were taken
This document is a draft for review purposes only and does not constitute Agency policy.
3-31 DRAFT—DO NOT CITE OR QUOTE
P ° O
formaldehyde
O y=-0.167x + 2.3
0O R2 = 0.0136
%
ax£hniD-Oi—o~~cd o 1 r
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
prior to class and within 2 hours after the class. Mean exposure levels were about 100 ppb.
Baseline averages of urinary formic acid (as a sodium salt) were 12.47 mg/L and ranged from
2.43 to 28.38 mg/L among subjects. Postexposure formate levels were slightly elevated but were
not statistically significant. Moreover, formate levels decreased in several individuals relative to
pre-exposure levels. The authors concluded that variability in urinary formate may mask any
changes and that monitoring formate within 2 hours of exposure is not informative. It is worth
noting, however, that interpretation of this finding is confounded due to the fact that diet was not
controlled and because no markers for urinary normalization were employed (Boeniger, 1987).
Boeniger (1987) reviewed previously published data on formate in urine (some of which
were in German). In one occupational study, workers were exposed to an average formaldehyde
exposure of 1.28 mg/m3 over a 6-hour work shift. This implies an average intake of 6 mg;6
Boeniger reported a range of 2.5 to 13 mg. However, the original study reported that post-shift
formate levels were 152 mg/L, whereas the levels were only 24 mg/L 6 days later (no exposure).
Considering that only a small percentage of inhaled formaldehyde would be excreted in urine, it
is unclear how (or whether) formaldehyde exposure, with the highest total dose of 13 mg, could
be responsible for the observed increase.
In the previously described study by Shin et al. (2007), human urine samples were shown
to contain TZCA, although variability was not reported. A subsequent study reported that urine
TZCA levels were higher in individuals living in newer apartments (0.18 ± 0.121 mg/g
creatinine) as compared to older apartments (0.097 ± 0.040 mg/g creatinine) (Li et al., 2007)7.
The authors cited this as evidence that TZCA is a urinary marker for formaldehyde exposure,
even though TZCA levels were not correlated to measured (or estimated) formaldehyde
exposures. The individuals also differed significantly in age (21.5 vs. 28.6, p = 0.053) and
differed in smoking percentage (10 vs. 27%). Clearly these two studies do not establish a
relationship between human formaldehyde exposure and urine TZCA levels.
3.6. MODELING Till TOXICOKINETICS OF FORMALDEHYDE AND DPX
3.6.1. Motivation
Airway geometry is expected to be an important determinant of inhaled formaldehyde
dosimetry in the respiratory tract and its differences across species. The uptake of formaldehyde
in the upper respiratory tract is highly nonhomogenous and spatially localized and exhibits
strong species differences. Species differences in kinetic factors have been argued to be the key
61.28 mg/m3 / l,000L/m3 x 13.8L/minute x 60 minutes/hour x 6 hours.
7 This study is described in greater detail in Chapter 5.
This document is a draft for review purposes only and does not constitute Agency policy.
3-32 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
determinants of species-specific lesion distributions for formaldehyde and other reactive inhaled
gases. In the first subsection here, the advantage gained in the quantitative risk assessment by
modeling these differences in the upper respiratory tract is explained. Such a model was
constructed by using computational fluid dynamic (CFD) methods for modeling airflow and
regional formaldehyde uptake in the rat, rhesus monkey, and human nose by scientists at the
Chemical Industries Institute of Toxicology (CUT) Centers for Health Research (presently the
Hamner Institutes of Health Sciences). While frank effects were seen only in the upper
respiratory tract in rodents, mild lesions were also present in the major bronchi olar region of the
rhesus monkey. Therefore, with regard to extrapolation of cancer risk from animal bioassays to
humans, the upper and lower human respiratory tract should both be considered potentially at
risk of developing formaldehyde-induced squamous cell carcinoma. Therefore, formaldehyde
dose to the entire human respiratory tract human respiratory tract was modeled in order to
develop a dose-response relationship that considered the entire respiratory tract. Accordingly, in
the second subsection results of modeling airflow and regional formaldehyde uptake in the
human lower respiratory tract by Overton et al. (2001) are provided. Unsteady effects were
argued to be insignificant at resting breathing rates, and therefore steady-state inspiratory flow
was assumed. Since these models have been described in various reports and publications, the
technical details are consigned to appendices or the literature is referenced.
The fluid dynamics modeling in the respiratory tract comprises two steps (Kimbell et al.
2001): airflow through the lumen and formaldehyde uptake by the lining of the respiratory tract.
Flow streamlines in the CFD simulations agreed reasonably well with experimentally observed
patterns in casts of the rat, monkey, and human nasal passages as well as with measurements of
velocity taken in hollow molds of the human nose. Pressure drop as a function of volumetric
flow also compared well with measurements made in rats. Unlike the airflow simulations, no
validation of the regional formaldehyde uptake simulations (that is, the spatial distribution of
uptake) was possible. (It was possible to compare simulations of overall uptake with
experimentally observed values.) In this assessment, several indirect qualitative and quantitative
lines of evidence were relied on to provide general confidence in the formaldehyde regional
uptake profile in the F344 rat nasal passages. In addition to the agreement mentioned earlier
with respect to airflow profiles, this evidence includes general agreement between measured and
predicted levels of formaldehyde DPXs (in the "high-tumor" regions) when simulations of
airflow and regional formaldehyde uptake were used as input to a physiologically based
pharmacokinetic (PBPK) model for DPX kinetics (Cohen-Hubal et al. 1997). Such indicators
are not available for the simulation of uptake patterns in the human. With regard to modeling the
lower respiratory tract, calculations in Overton et al. (2001) are based on an idealized
This document is a draft for review purposes only and does not constitute Agency policy.
3-33 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
representation of airways in the human lung and of air flow through them (referred to in the
literature as "single-path" models). Overton et al. (2001) did not attempt to validate their
simulations. However, the following observation may be noted in support of their model. In the
case of the deposition of coarse and fine particulate matter, single-path models have traditionally
provided a reasonably accurate representation of the average deposition in a given generation of
the lung airways (i.e. airways at a given depth in the lung) for a normal human population.
There were mass balance errors in the CFD calculations (Kimbell et al. 2001). Mass
balance errors associated with formaldehyde uptake into tissue ranged from less than 14% for the
rat, monkey, and human at resting minute volume to approximately 27% at the highest
inspiratory flow rates of 31.8 and 37 L/minute in the human.
If DPX formation and cell proliferation are driven by formaldehyde flux, then these
quantities may also be expected to exhibit site and species differences, thus arguing for linking
these quantities to the modeled formaldehyde flux (Conolly et al. 2000).
The third subsection evaluates PBPK models for DPX kinetics in the F344 rat and rhesus
monkey, using CFD simulations of formaldehyde flux to the nasal lining as input (Klein et al.,
2009; Subramaniam et al., 2007; Conolly et al., 2000), and discusses their uncertainties. This
subsection further discusses issues related to the scaling of the animal models to the human.
This assessment uses internal dose metrics computed by using the models described in
this section so as to derive more accurate human equivalent concentrations from the animal
bioassays than would be obtained by averaging over the respiratory surface area. The strengths
and uncertainties associated with the data and the models and their relevance to the hypothesized
mode of action are discussed in some length.
3.6.2. Species Differences in Anatomy: Consequences for Gas Transport and Risk
As discussed earlier, formaldehyde is highly reactive and water soluble (categorized as a
category 1 gas), thus its absorption in the mucus layer and tissue lining of the upper respiratory
tract is known to be significant. The regional inhaled dose of formaldehyde to the respiratory
tract of a given species depends on the amount of formaldehyde delivered by inhaled air, the
absorption characteristics of the nasal lining, and reactions in the tissue. The amount delivered
by inhaled air is a function of the major airflow patterns, air-phase diffusion, and absorption at
the airway-epithelial tissue interface. The dose of formaldehyde to the epithelial tissue, which is
different from the amount delivered, depends on the amount absorbed at the airway-tissue
interface, water solubility, mucus-to-tissue phase diffusion, and chemical reactions, such as
hydrolysis, protein binding, and metabolism. It has been argued strongly that species differences
in these kinetic factors are determinants of species-specific lesion distributions for formaldehyde
This document is a draft for review purposes only and does not constitute Agency policy.
3-34 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
and other inhaled gases (Moulin et al., 2002; Bogdanffy et al., 1999; Ibanes et al., 1996;
Monticello et al., 1996; Monticello and Morgan, 1994; Morgan et al., 1991).
Because of the convoluted nature of the airways in the upper respiratory tract, the
absorption of such gases in the upper respiratory tract is highly nonhomogeneous. There are
large differences across species in the anatomy of the upper respiratory tract (Figure 3-6) and in
airflow patterns (Figure 3-7). Therefore, as shown in the simulations in Figure 3-8, it may be
expected that the uptake patterns and thus risk due to inhaled formaldehyde will also show
strong species dependence. Morgan et al. (1991) concluded that airflow-driven dosimetry plays
a critical role in determining the site specificity of various formaldehyde-induced responses,
including tumors, in the nose of the F344 rat. The convoluted geometry of the airway passages
in the upper respiratory tract, as seen from the cross sections of the nose in Figure 3-6, renders an
idealized representation of fluid flow and uptake profiles almost impossible. For these reasons,
Kimbell et al. (1993), Kepler et al. (1998), and Subramaniam et al. (1998) developed
anatomically realistic finite-element representations of the noses of humans, F344 rats, and
rhesus monkeys. These representations were subsequently used in physical and computational
models (Figure 3-6). This assessment utilizes dosimetry derived from these representations.
An accurate calculation of species differences in formaldehyde dosimetry in the upper
respiratory tract is important to the extrapolation problem for another reason. The upper
respiratory tract in rats is an extremely efficient scrubber of reactive gases (97% uptake)
(Morgan et al., 1986), thereby protecting the lower respiratory tract from gaseous penetration.
On the other hand, there is considerably more fractional penetration of formaldehyde into the
lower respiratory tract of the rhesus monkey than in the rat (see Figure 3-8). Therefore, an
accurate determination of scrubbing in the upper respiratory tract is important to delineate
species differences in the level of risk to the lower respiratory tract. Thus, in the case of the
rhesus monkey, the model by Kepler et al. (1998) included the trachea. Because the human
model also had to address the potential for oronasal breathing, an idealized single-path model of
the lower respiratory tract was attached to a model of the upper respiratory tract (Overton et al.,
2001). It is important to note that the models mentioned above represent nasal passages
reconstructed from a single individual from each species (Kimbell et al., 2001a, b; Conolly et al.,
2000; CUT, 1999; Subramaniam et al., 1998). This is discussed later in the context of
intraspecies variability.
The highly localized nature of uptake patterns shown in Figure 3-8 means that averaging
uptake over the entire nasal surface area would dilute the regional dose over areas where
response was observed and that an extrapolation based on such averaging would clearly not be
accurate.
This document is a draft for review purposes only and does not constitute Agency policy.
3-35 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
<&L
3-
/
•i
¦t' .
Vfj
I '*
F344 Rat
Rhesus Monkey
m
Human
Figure 3-6. Reconstructed nasal passages of F344 rat, rhesus monkey, and
human.
Note: Nostril is to the right, and the nasopharynx is to the left. Right side shows
the finite element mesh. Left-hand side shows tracings of airways obtained from
cross sections of fixed heads (F344 rat and rhesus monkey) and magneti c
resonance image sectional scans (humans). Aligned cross sections were
connected to form a three-dimensional reconstruction and finite-element
computational mesh. Source: Adapted from Kimbell et al. (2001a). Additional
images provided courtesy of Dr. J.S. Kimbell, CUT Hamner Institutes.
This document is a draft for review purposes only and does not constitute Agency policy.
3-36 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
(A) F344 Rat
CFD-Model simulated inspiratory airflow
Streams observed in experimental water-dye studies
(B) Rhesus Monkey
CFD-Model simulated inspiratory airflow
Streams observed in experimental water-dye studies
(C) Human
fl
CFD-Model simulated inspiratory airflow
Black bars = Axial velocities measured in hollow molds
White bars = CFD-simulated axial velocities
£
3.0 Plot 2
<
Plot 3 3.0
¦i n li ii ii -
Plot 4
J (In
Plot 5 3.0
¦ mnii Mi -
Plot 6
ii n r m n
Plot 7 3.0
2.0
n 11 n 1111 n m
Plots
rfi rfi r v
Normalized airway diameter
Figure 3-7. Illustration of interspecies differences in airflow and verification
of CFD simulations with water-dye studies.
Note: Panels A and B show the simulated airflow pattern versus water-dye
streams observed experimentally in casts of the nasal passages of rats and
monkeys, respectively. Panel C shows the simulated inspiration airflow pattern,
and the histogram depicts the simulated axial velocities (white bars) vs.
experimental measurements made in hollow molds of the human nasal passages
Dye stream plots were compiled for the rat and monkey over the physiological
range of inspiration flow rates. Modeled flow rates in humans were 15 L/minute.
Source: Adapted from Kimbell et al. (2001a).
This document is a draft for review purposes only and does not constitute Agency policy.
3-37 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
F344 Rat
Key
pmol/(mm2-hr-ppm)
Rhesus Monkey
Human
2000
1500
1000
h 500
Figure 3-8. Lateral view of nasal wall mass flux of inhaled formaldehyde
simulated in the F344 rat, rhesus monkey, and human.
Note: Nostrils are to the right. Simulations were exercised in each species at
steady-state inspiration flow rates of 0.576 L/minute in the rat, 4.8 L/minute in the
monkey, and 15 L/minute in the human. Flux was contoured over the range from
0-2,000 pmol/(mm2-hour-ppm) in each species.
Source: Kimbell et al. (2001a).
Another factor to consider in the extrapolation is that monkeys and humans are oronasal
breathers while rats are obligate nose-only breathers. Thus, for humans and monkeys, oronasal
or oral breathing implies a significantly higher uptake in the lower respiratory tract. It is known
that a significant fraction of the human population breathes normally through the mouth.
Finally, activity profiles are also determinants of extraction efficiency (see Figure 3-9) and of
breathing route (Niinimaa et al., 1981). Given the fact that formaldehyde-induced lesions were
This document is a draft for review purposes only and does not constitute Agency policy.
3-38 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
observed as far down the respiratory tract as the first bifurcation of the lungs in exposed
monkeys, the entire human respiratory tract should be considered when extrapolating data from
rats.
2082 pmol/(mm2-hr-ppm)
7.4 L/min
15 L/min
18 L/min
25.8 L/min
31.8 L/min
Figure 3-9. CFD simulations of formaldehyde flux to human nasal lining at
different inspiratory flow rates.
Note: Right lateral view. Uptake is shown for the non-squamous portion of the
epithelium. The front portion of the nose (vestibule) is lined with keratinized
squamous epithelium and is expected to absorb relatively much less
formaldehyde.
Source: Kimbell et al. (2001b).
This document is a draft for review purposes only and does not constitute Agency policy.
3-39 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
3.6.3. Modeling Formaldehyde Uptake in Nasal Passages
CUT scientists chose the F344 strain of the rat since it was assumed to be anatomically
representative of its species and because it is widely used experimentally, most notably in
bioassays sponsored by the National Toxicology Program and at CUT. The approximate
locations of squamous, mucus-coated, and nonmucus epithelial cells were mapped onto the
reconstructed nasal geometry of the CFD models. Taken together, these regions of nonmucus
and mucus-coated cells comprise the entire surface area of the nasal passages (see original
papers and CUT [1999] for further details on reconstruction and morphometry). Types of nasal
epithelium overlaid onto the geometry of the models were assumed to be similar in
characteristics across all three species (rat, monkey, and human) except for thickness, surface
area, and location. Species-specific mucosal thickness, surface area, and location were estimated
from the literature, from the documentation of the CFD models, or by direct measurements
(Conolly et al., 2000; CUT, 1999). The nasal passages of all three species were assumed to have
a continuous mucus coating over all surfaces except specific areas in the nasal vestibule. As
discussed at the beginning of this chapter, formaldehyde hydrolyzes in water and reacts readily
with a number of components of nasal mucus. Absorption rates of inhaled formaldehyde by the
nasal lining were therefore assumed to depend on where the epithelial lining is coated by mucus
and where it is not.
To calculate an airflow rate that would be comparable among species, the amount of
inspired air (tidal volume, VT) was divided by the estimated time involved in inhalation (half the
time a breath takes, or (l/2)(l/[breathing frequency, f]). Thus, an inspiratory flow rate was
calculated to be 2VTf, or twice the minute volume. Predicted flux values represent an average of
one nasal cycle. Minute volumes were allometrically scaled to 0.288 L/minute for a 315 g rat
from data given by Mauderly (1986). Simulations were therefore carried out at 0.576 L/minute
for the rat.
The fluid dynamics modeling in the respiratory tract comprises two steps: modeling the
airflow through the lumen (solution of Navier-Stokes equations) and modeling formaldehyde
uptake by the respiratory tract lining (solution of convective-diffusion equations for a given
airflow field). Details of these simulations, including boundary conditions for air flow and mass
transfer, are provided in Kimbell et al. (2001a, b, 1993) and Subramaniam et al. (1998).
Formaldehyde absorption at the airway-to-epithelial tissue interface was assumed to be
proportional to the air-phase formaldehyde concentration adjacent to the nasal lining layer in
monkeys and humans (see the original paper [Kimbell et al., 2001a] for a more detailed
elaboration of the calculations for these coefficients).
This document is a draft for review purposes only and does not constitute Agency policy.
3-40 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
Because formaldehyde is highly water soluble and reactive, Kimbell at al. (2001a)
assumed that absorption occurred only during inspiration. Thus, for each breath, flux into nasal
passage walls (rate of mass transport in the direction perpendicular to the nasal wall per mm2 of
the wall surface) was assumed to be zero during exhalation, with no backpressure to uptake built
up in the tissues. Overton et al. (2001) estimated the error due to this assumption to be small,
roughly an underestimate of 3% in comparison to cyclic breathing. Also, this assumption is the
same as that used in default methods for reference concentration determination and has been
used in other PBPK model applications to describe nasal uptake (Andersen and Jarabek, 2001).
3.6.3.1. Flux Bins
A novel contribution of the CUT biologically motivated dose-response model is that cell
division rates and DPX concentrations are driven by the local concentration of formaldehyde.
These were determined by partitioning the nasal surface by flux, resulting in 20 "flux bins."
Each bin was comprised of elements (not necessarily contiguous) of the nasal surface that
receive a particular interval of formaldehyde flux per ppm of exposure concentration (Kimbell et
al., 2001a). The spatial coordinates of elements comprising a particular flux bin were fixed for
all exposure concentrations, with formaldehyde flux in a bin scaling linearly with exposure
concentration (ppm). Thus, formaldehyde flux was expressed as pmol/(mm2-hour-ppm).
3.6.3.2. Flux Estimates
Formaldehyde flux was estimated for the rat, monkey, and human over the entire nasal
surface and over the portion of the nasal surface that was lined by nonsquamous epithelium.
Formaldehyde flux was also estimated for the rat and monkey over the areas where cell
proliferation measurements were made (Monticello et al., 1991, 1989) and over the anterior
portion of the human nasal passages that is lined by nonsquamous epithelium. Figure 3-8 shows
the mass flux of inhaled formaldehyde to the lateral wall of nasal passages in the F344 rat, rhesus
monkey, and human (Kimbell et al., 2001a).
Maximum flux estimates for the entire upper respiratory tract were located in the mucus-
coated squamous epithelium on the dorsal aspect of the dorsal medial meatus near the boundary
between nonmucus and mucus-coated squamous epithelium in the rat, at the anterior or rostral
margin of the middle turbinate in the monkey, and in the nonsquamous epithelium on the
proximal portion of the mid-septum near the boundary between squamous and nonsquamous
epithelium in the human (see Kimbell et al. [2001b] for tabulations of comparative estimates of
formaldehyde flux across the species).
This document is a draft for review purposes only and does not constitute Agency policy.
3-41 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
The rat-to-monkey ratio of the highest site-specific fluxes in the two species was 0.98. In
the rat, the incidence of formaldehyde-induced squamous cell carcinomas in chronically exposed
animals was high in the anterior lateral meatus (Monticello et al., 1996). Flux predicted per ppm
in this site and flux predicted near the anterior or proximal aspect of the inferior turbinate and
adjacent lateral walls and septum in the human were similar, with a rat-to-human ratio of 0.84.
3.6.3.3. Mass Balance Errors
Overall uptake of formaldehyde was calculated as 100% x (mass entering nostril - mass
exiting outlet)/(mass entering nostril). Mass balance errors for air, 100% x (mass of air entering
nostril - mass exiting outlet)/(mass entering nostril), and inhaled formaldehyde, 100% x (mass
entering nostril - mass absorbed by airway walls - mass exiting outlet)/(mass entering nostril),
were calculated. Mass balance errors associated with simulated formaldehyde uptake from air
into tissue ranged from less than 14% for the rat, monkey, and human at 7.4 and 15 L/minute to
approximately 27% at the highest inspiratory flow rates of 31.8 and 37 L/minute (Kimbell et al.,
2001b). Kimbell et al. (2001b) corrected the simulation results for these errors by evenly
distributing the lost mass over the entire nasal surface.
3.6.4. Modeling Formaldehyde Uptake in the Lower Respiratory Tract
Lesions were observed in the lower respiratory tract of rhesus monkeys exposed to 6 ppm
formaldehyde. Therefore it is appropriate to consider the human lower respiratory tract as
potentially at risk for formaldehyde-induced cancer. Accordingly, fluid flow and formaldehyde
uptake in the lower respiratory tract were also modeled for the human in the CUT approach by
using dosimetry estimates for the human lower respiratory tract.
The single-path idealization of the human lung anatomy captures the geometrical
characteristics of the airways for a given lung depth, and of airflow through these airways, in an
average, homogeneous sense. For particulates, this has provided a reasonable representation of
the average deposition in a given generation of the lung airways for a normal human population.
The one-dimensional model by Weibel (1963) is generally considered adequate unless the fluid
dynamics at airway bifurcations need to be explicitly modeled. However, such an idealization of
the lung geometry has been successfully used in various models for the dosimetry of ozone and
particulate and fibrous matter. Most likely, the lung geometries of the susceptible population,
such as those with chronic obstructive pulmonary disease, would depart significantly from the
geometry described in Weibel (1963). Unlike the accurate representation of the nasal anatomy
used in the CFD modeling, the lung geometry is idealized in the CUT approach as a typical path
Weibel geometry. This captures the lung structure in an average, homogeneous sense for a given
This document is a draft for review purposes only and does not constitute Agency policy.
3-42 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
lung depth. The single-path model used to calculate formaldehyde uptake in the human
respiratory tract (Overton et al., 2001; CUT, 1999) applied a one-dimensional equation of mass
transport to each generation of an adult human symmetric, bifurcating Weibel-type respiratory
tract anatomical model, augmented by an upper respiratory tract. The detailed CFD modeling of
the upper respiratory tract was made consistent with the upper respiratory tract in the single-path
model by requiring that the one-dimensional version of the nasal passages have the same
inspiratory air-flow rate and uptake during inspiration as the CFD simulations for four daily
human activity levels. The reader is referred to Overton et al. (2001) for further details of the
simulations. Results most relevant to this assessment are shown in Figure 3-10.
180O
Minule Volume:
• 7.5 Umin (Nasal B eathing)
—9.0 Umin (Nasal B eathir>g)
—25.0 L/min (Masai 6 eaUung)
—50.0 L/min (Oronas il Breathing)
1600
e
E
1200 -
jr
E
LS-
CL
'J)
O
Mouth > *
1000 -
800 -
o
'CL
X
D
_l
LL
600 -
LU
o
<
400 -
LI-
CC
200
URT
Tracneobrorvcfiiai Region
-200
10
15
20
25
5
0
5
MODEL GENERATION
Figure 3-10. Single-path model simulations of surface flux per ppm of
formaldehyde exposure concentration in an adult male human.
Source: Overton et al. (2001).
The primary predictions of the model, as shown in Figure 3-10, were that more than 95%
of the inhaled formaldehyde would be retained and formaldehyde flux in the lower respiratory
tract would increase for several lung airway generations from that in the posterior-most segment
of the nose and then decrease rapidly, resulting in almost zero flux to the alveolar sacs.
This document is a draft for re\'iew purposes only and does not constitute Agency policy.
3-43 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
Overton et al. (2001) modeled uptake at higher inspiratory rates, including those at
50 L/minute of minute volume (well beyond levels where the oronasal switch occurs in the
normal nasal breathing population). At these rates Figure 3-8 indicates that formaldehyde flux in
the mouth cavity is comparable (but a bit less) to that occurring in the nasal passages. Overton et
al. (2001) did not model uptake in the oral cavity at minute volumes less than 50 L/minute. This
would be of interest because mouth breathers form a large segment of the population.
Furthermore, at concentrations of formaldehyde where either odor or sensory irritation becomes
a significant factor, humans are likely to switch to mouth breathing even at resting inspiration.
At a minute volume of 50 L/minute, Overton et al. (2001) assumed, citing Niinimaa et al. (1981),
that 0.55 of the inspired fraction is through the mouth. Therefore, based on the results in
Figure 3-8, it is not unreasonable to assume that for mouth breathing conditions at resting or
light exercise inspiratory rates, average flux across the human mouth lining would be
comparable to the average flux across the nasal lining computed in Kimbell et al. (2001a, b).
3.6.5. Uncertainties in Formaldehyde Dosimetry Modeling
3.6.5.1. Verification of Predicted Flow Profiles
The simulated streamlines of steady-state inspiration airflow predicted by the CFD model
agreed reasonably well with experimentally observed patterns of water-dye streams made in
casts of the nasal passages for the rat and monkey as shown in panels A and B in Figure 3-7.
The airflow velocity predicted by CFD model simulations of the human also agreed well with
measurements taken in hollow molds of the human nasal passages (panel C, Figure 3-8) (Kepler
et al., 1998; Subramaniam et al., 1998; Kimbell et al., 1997, 1993). However, the accuracy and
relevance of these comparisons are limited. The profiles were verified by video analysis of dye
streak lines in the molds of rats and rhesus monkeys, although this method is reasonable for only
the major airflow streams.
Plots of pressure drop vs. volumetric airflow rate predicted by the CFD simulations
compared well with measurements made in rats in vivo (Gerde et al., 1991) and in acrylic casts
of the rat nasal airways (Cheng et al., 1990) as shown in Figure 3-11. This latter comparison
remains qualitative due to differences among the simulation and experiments as to where the
outlet pressure was measured and because no tubing attachments or other experimental apparatus
were included in the simulation geometry. The simulated pressure drop values were somewhat
lower, possibly due to these differences.
Inspiratory airflow was assumed to be constant in time (steady state). Subramaniam et al.
(1998) considered this to be a reasonable assumption during resting breathing conditions based
on a value of 0.02 obtained for the Strouhal number. Unsteady effects are insignificant when
This document is a draft for review purposes only and does not constitute Agency policy.
3-44 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
this number is much less than one. However, this assumption may not be reasonable for light
and heavy exercise breathing scenarios.
¦ Nasal Mold C1
In Vivo
-A- Nasal Mold C2
CIITCFD
0 100 200 300 400 500 600 700 800 900 1000
Airflow rats (ml/min)
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 (Gerde et al., 1991).
Source: Kimbell et al. (1997).
3.6.5.2. Level of Confidence in Formaldehyde Uptake Simulations
Unlike the airflow simulations, it was not possible to evaluate the formaldehyde uptake
calculations directly. Since the mass transfer boundary conditions were set by fitting overall
uptake to the average experimental data for various exposure concentrations, it was not possible
to independently verify even the overall uptake values with empirical data. This assessment has
relied on several indirect qualitative and quantitative lines of evidence listed below to provide
general confidence in the uptake profile for the F344 rat nasal passages, as modeled in CUT
(1999), when gross averages are considered over certain regions of the nasal lining.
In an earlier simulation, where the nasal walls were set to be infinitely absorbing of
formaldehyde, uptake of inhaled formaldehyde in the upper respiratory tract was predicted to be
90% in the rat for simulations corresponding to the resting minute volume in the F344 rat. This
estimate compared reasonably well with the range of 91-98% observed by Morgan et al.
(1986a).
This document is a draft for review purposes only and does not constitute Agency policy.
3-45 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
Morgan et al (1991) showed general qualitative correspondence between the main routes
of flow and lesion distribution induced by formaldehyde in the rat nose. In their initial work
with a CFD model that represented a highly reactive and soluble gas, Kimbell et al. (1993)
described similarities in computed regional mass flux patterns and lesion distribution due to
formaldehyde. When the results from this work in the coronal section immediately posterior to
the vestibular region were considered, simulated flux levels over regions such as the medial
aspect of the maxilloturbinate and the adjacent septum (where lesions were seen) were an order
of magnitude higher than over other regions, such as the nasoturbinate (where lesions were not
seen).8
The results of a PBPK model by Cohen-Hubal et al. (1997) provide a reasonable level of
confidence in regional uptake simulations for the F344 rat when gross averages over nasal sites
are carried out. Cohen-Hubal et al. (1997) linked the CFD dosimetry model for formaldehyde to
a PBPK model for formaldehyde-DPX concentration in the F344 rat. This PBPK model was
calibrated by optimizing the model to combined_DPX data from high-tumor incidence and low-
tumor incidence regions of the rat nose that were obtained in separate experiments by Casanova
et al. (1991, 1989). These data were obtained at 0.3, 0.7, 2.0, 6.0, and 10 ppm for both regions
and also at 15 ppm for the high-tumor region. Model prediction of DPX concentrations for the
high-tumor region only compared well with the experimental data, including 15 ppm for which
the model had not been calibrated. This is shown in Figure 3-12.
The CFD simulations do not model reflex bradypnea, a protective reflex seen in rodents
but not in humans. As discussed at length in sections 3.2.3.1 and 4.2.1.1, it is reasonable to
expect a range of 25% (Chang et al., 1983) to 45% (Barrow et al., 1983) decrease in minute
volume in F344 rats at the exposure concentration of 15 ppm. Explicit omission of this effect in
the modeling is, however, not likely to be a source of major uncertainty in the modeled results
for uptake of formaldehyde in the rat nose for the following reason. The CFD model for the
F344 rat was calibrated to fit the overall experimental result for formaldehyde uptake in the F344
rat at 15 ppm exposure concentration. This was carried out by adjusting the mass transfer
coefficient used as boundary condition on the absorbing portion of the nasal lining. Thus, the
reflex bradypnea occurring in those experimental animals is phenomenologically factored into
the value used for the boundary condition. Nonetheless, some error in the localized distribution
of uptake patterns may be expected, even if the overall uptake is reproduced correctly.
8 However, this 1993 CFD model differed somewhat from the subsequent model by Kimbell et al. (2001a) used in
this assessment. In the 1993 model, the limiting mass-transfer resistance for the gas was assumed to be in the air
phase; that is, the concentration of formaldehyde was set to zero at the airway lining. Furthermore, this same
boundary condition was used on the nasal vestibule as well, while, in the more recent model, the vestibule was
considered to be non-absorbing. Unfortunately, Kimbell at al. (2001a) did not report on correspondences between
This document is a draft for review purposes only and does not constitute Agency policy.
3-46 DRAFT—DO NOT CITE OR QUOTE
-------
1 Furthermore, since the same value for the mass transfer coefficient was used in human
2 simulations (as obtained from calibration of the rat model), there is additional uncertainty in the
3 modeled human flux estimates. This issue was not addressed by Kimbell et al. (2001a, b),
4 Conolly et al. (2004), or Schlosser et al. (2003), and we are unable to assess the extent of this
5 error more accurately.
HCHO Flux (jiM-mm/min)
2
5 15
x
CL
Q
o
Z
JE
Q
-C
5
5.0
7.5
12.5
0
17.5
2.5
0
HCHO Exposure Concentration (ppm)
HCHO Flux (nM-mm/mln)
34-5
241.5
2.5
2
c:
r"J
o>
'¦-I
QC
L-n
X
2.5
7.5
12.S
17.5
8.0 2.5 5.0 7.5 10.& 12.S 15.0 17.5
HCHO Exposure Concentration (ppm)
6
7 Figure 3-12. Formaldehyde-DPX dosimetry in the F344 rat.
8 Panel A: calibration of the PBPK model using data from high and low tumor
9 incidence sites. Panel B: model prediction compared against data from high
10 tumor incidence site. Dashed line in panel A shows the extrapolation outside the
11 range of the calibrated data.
12 Source: Cohen-Hubal et al. (1997).
13
flux patterns and lesion distribution.
This document is a draft for review purposes only and does not constitute Agency policy.
3-47 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
3.6.6. PBPK Modeling of DNA Protein Cross-Links (DPXs) Formed by Formaldehyde
3.6.6.1. PBPK Models for DPXs
As can be seen from the previous sections, measuring the distribution of the absorbed
formaldehyde and identifying its form have proven difficult. Because of the high reactivity of
formaldehyde, rapid metabolism of formaldehyde, and complexity of formate clearance, dose
surrogates (or biomarkers) of exposure have been used to characterize the extent of absorption
and distribution of formaldehyde. As with other soluble and reactive gases, typical PBPK
models that predict steady-state blood concentrations are not useful for predicting formaldehyde
dosimetry at this time. As noted previously, inhalation exposure to formaldehyde has not been
shown to increase blood formaldehyde levels. Thus, most modeling efforts for formaldehyde
have focused on disposition at the site of contact.
As discussed earlier, the concentration of DPXs formed by formaldehyde has been
treated as a surrogate for the tissue dose of formaldehyde in earlier efforts by Casanova et al.
(1991) and in EPA's efforts to update its health assessment of formaldehyde (Hernandez et al.,
1994). These efforts used data from rats and rhesus monkeys (Casanova et al., 1991, 1989).
Using DPXs in this manner allowed the incorporation of both clearance and metabolism of
formaldehyde and the incorporation of the effect of saturation on detoxification of formaldehyde
at higher doses. Calculation of the average DPX concentration from these data was seen as a
surrogate for the area under the curve (AUC) of the reactive formaldehyde species in the
epithelium. Based on these data, Casanova et al. (1991) developed a PBPK model for predicting
DPXs in these species and for extrapolating to the human.
The Casanova et al. (1991) model consists of three anatomical compartments
representing different parts of the upper respiratory tract of the rhesus monkey. The results
indicated a 10-fold difference in DPX formation between rats and monkeys, due primarily to
species differences in minute volume and differing quantities of DNA in the nasal mucosa.
Casanova et al. (1991) then developed a monkey/rat scaling factor for these parameters by taking
the ratio of nasal mucosa tissue between the two species, a determinant that was proportional to
the total body weight differences between the two species. Using these scaling factors in their
model, the authors' predictions in monkey (based on the rat data) were in close agreement with
observed DPXs in monkey, particularly at higher formaldehyde concentrations. However, the
model overpredicted DPX formation in the monkey at lower formaldehyde concentrations.
Subsequent rat-human and monkey-human scaling results predicted much lower DPX formation
in man. Again, the values obtained at lower concentrations may have been overpredicted, as was
the case for rat-monkey extrapolation.
This document is a draft for review purposes only and does not constitute Agency policy.
3-48 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
Georgieva et al. (2003) developed a model for the uptake and disposition of
formaldehyde in the rat nasal lining. This model was designed to predict the distribution of
formaldehyde in the nasal mucosa. The model indicated that, at 6 ppm exposure, a steady-state
elevation of 15-20 |iM formaldehyde would be achieved within 30 seconds. Furthermore, this
same elevation was predicted when the exposure was 6 ppm formaldehyde for 60 minutes.
Given that human blood formaldehyde levels are predicted to be about 100 ±15 |iM (Heck et al.,
1985) and assuming that blood formaldehyde concentration is roughly equivalent to the
concentration predicted at the basement membrane of the epithelium, this model predicts roughly
a 15-20% increase in blood formaldehyde. However, it should be noted that a 40-minute
inhalation exposure of humans to 1.99 ppm formaldehyde did not lead to a measurable increase
in blood formaldehyde (Heck et al., 1985).
Franks (2005) published a mathematical model for predicting the disposition of
formaldehyde in the human nasal mucosa and blood. The calculated concentrations of
formaldehyde in the mucus, the epithelium, and the blood attained steady-state profiles within a
few seconds of exposure. The increase of the formaldehyde concentration in the blood was
predicted to be insignificant compared with the existing pre-exposure levels in the body: an
increase of 0.00044 mg/L in blood formaldehyde following exposure to 1.9 ppm formaldehyde
for up to 8 hours. The model described formaldehyde concentration gradients across the mucus,
epithelial, and submucosal compartments in the human nose. Transport of formaldehyde was
governed by the following processes: diffusional (in the mucus); a combination of diffusional,
two first order terms representing intrinsic reactivity of formaldehyde and binding to DNA, and
Michaelis-Menten kinetics representing enzymatic metabolism (in the epithelial layer); a first-
order term representing non-enzymatic removal governed by the blood perfusion rate (in the
submucosal compartment). The model used the values for the first order reaction rate constants
and the Michaelis-Menten parameters (Vmax and Km) estimated by Conolly et al. (2000) in their
model for extrapolating the rat and rhesus monkey data to the human. The modeling in Franks
(2005) was not calibrated or validated against experimental data, but the predictions of
negligible penetration of free formaldehyde to the blood are qualitatively in agreement with the
conclusions in Heck et al. (1985).
Following the efforts by Casanova and co-workers, Cohen-Hubal et al. (1997), Conolly et
al. (2000), and Georgieva et al. (2003) developed models that linked local formaldehyde flux
from CFD models to DPX predictions. The focus here will be on the Conolly et al. (2000) effort
for the following two reasons: it explicitly incorporates regional formaldehyde dosimetry in the
nasal lining by using results from CFD modeling of airflow and gas uptake and it brings data
This document is a draft for review purposes only and does not constitute Agency policy.
3-49 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
across species (rat and rhesus monkey) to bear on model calibration, such a situation being
relatively rare in chemical health risk assessments.
3.6.6.2. A PBPK Model for DPXs in the F344 Rat and Rhesus Monkey that Uses Local Tissue
Dose of Formaldehyde
In earlier risk assessment efforts (Hernandez et al., 1994; Casanova et al., 1991; U.S.
EPA, 1991b), the average DPX concentration was considered a surrogate tissue dose metric for
the AUC of the reactive formaldehyde species. Conolly et al. (2003) assigned a more specific
role for DPXs, treating local DPX concentration as a dose surrogate indicative of the
intercellular concentration of formaldehyde, leading to formaldehyde-induced mutations. These
authors indicated that it was not known whether DPXs directly induced mutations (Conolly et
al., 2003; Merk and Speit, 1998). This is discussed in detail in the mode-of-action sections in
this document. The Conolly et al. (2000) model for the disposition of inhaled formaldehyde gas
and DPX in the rat and rhesus nasal lining is relatively simple in terms of model structure
because it consists of a single well-mixed compartment for the nasal lining as follows:
1. Formaldehyde flux to a given region of the nasal lining is provided as input to the
modeling and is obtained in turn as the result of a CFD model. This flux is defined as the
amount of formaldehyde delivered to the nasal lining per unit time per unit area per ppm
of concentration in the air in a direction transverse to the airflow. It is locally defined as
a function of location in the nose and the inspiratory flow rate and is linear with exposure
concentration.
2. The clearance of formaldehyde from the tissue is modeled as follows:
a. a saturable pathway representing enzymatic metabolism of formaldehyde, which
is primarily by formaldehyde dehydrogenase (involving Michaelis-Menten
parameters Vmax and Km)
b. a separate first-order pathway, which is assumed to represent the intrinsic
reactivity of formaldehyde with tissue constituents (rate constant kf)
c. first-order binding to DNA that leads to DPX formation (rate constant kb)
3. The clearance or repair of this DPX is modeled as a first order process (rate constant
kloss)-
DPX data
DPX concentrations were estimated from a study by Casanova et al. (1994) in which rats
were exposed 6 hours/day, 5 days/week, plus 4 days for 11 weeks to filtered air (naive) or to 0.7,
2, 6, or 15 ppm (0.9, 2.5, 7.4, or 18 mg/m3) formaldehyde (pre-exposed). On the 5th day of the
This document is a draft for review purposes only and does not constitute Agency policy.
3-50 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
12th week, the rats were then exposed for 3 hours to 0, 0.7, 2, 6, or 15 ppm 14C-labeled
formaldehyde (with pre-exposed animals exposed to the same concentration as during the
preceding 12 weeks and 4 days). The animals were sacrificed and DPX concentrations
determined at two sites in the nasal mucosa. Conolly et al. (2000) used these naive rat data to
develop a PBPK model that predicted the time-course of DPX concentrations as a function of
formaldehyde flux at these sites.9
3.6.6.3. Uncertainties in Modeling the Rat and Rhesus DPX Data
3.6.6.3.1. Half-life of DPX repair. In the development of the PBPK model for DPXs, Conolly
et al. (2000) assumed a value of 6.5 x 10 3 minute-1 for kioss, the first-order rate constant for the
clearance (repair) of DPXs, such that the DPXs predicted at the end of a 6-hour exposure to
15 ppm were reduced to exactly the detection limit for DPXs in 18 hours (the period between the
end of 1 day's 6-hour exposure and the beginning of the next). This determination of rapid
clearance was based on an observation by Casanova et al. (1994) that the DPX concentrations
observed in the pre-exposed animals were not significantly higher than those in naive animals (in
which there was no significant DPX accumulation). However, in vitro data (Quievryn and
Zhitkovich, 2000) indicate a much slower clearance, with an average kioss of
9.24 x 10~4 minute-1.
Subramaniam et al. (2007) examined the Casanova et al. (1994) data and argued that
there was a significantly decreased (~ 40%) level of DPXs in high tumor regions of pre-exposed
animals vs. naive animals at 6 and 15 ppm and that the weight of the tissues dissected from those
regions increased substantially, indicating a thickening of the tissues. After testing the outcome
of changing the tissue thickness in the PBPK model for DPXs, it was apparent to these authors
that such a change alone could not account for the dramatic reduction in DPX levels after pre-
exposure, even with the higher value of kioss used by Conolly et al. (2000). Therefore, in
addition to the gross increase in tissue weight, these data indicated either an induction in the
activity of enzymes that remove formaldehyde (aldehyde and formaldehyde dehydrogenase) or
other changes in the biochemical properties of the highly exposed tissue that must have occurred.
Given such a change, Subramaniam et al. (2007) concluded that the experimental results in
Casanova et al. (1994) were consistent with the smaller experimental value of kioss indicated by
the Quievryn and Zhitkovich (2000) data. In particular, they argued that if Vmax increased with
exposure (in a tissue region- and dose-specific manner), then it was possible to explain the naive
9 Note that Conolly et al. (2000) stated that they used the pre-exposed data, but this statement appears to be in error
(see Subramaniam et al. [2007]).
This document is a draft for review purposes only and does not constitute Agency policy.
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
vs. pre-exposed data of Casanova et al. (1994), with the value of kioss effectively measured in
vitro by Quievryn and Zhitkovich (2000). Furthermore, this value was measured directly, rather
than obtained by indirect interpretation of measurements made at only two time points where
significant changes in the tissue had occurred. Therefore, Subramaniam et al. (2007) considered
the use of this lower value for kioss to be more appropriate. Consequently, they re-implemented
and re-optimized the Conolly et al. (2000) model with this modification and found that the fit so
obtained to the acute DPX data was excellent. The re-implemented model will be used in this
assessment, and more details can be found in Subramaniam et al. (2007).
It should be noted that this slower DPX repair rate was obtained in an in vitro study by
using human cell lines that were transformed and immortalized. However, it appears that DPX
repair in normal cells would be even slower. When non-transformed freshly purified human
peripheral lymphocytes were used instead, the half-life for DPX repair was about 50% longer
than in the cultured cells (Quievryn and Zhitkovich, 2000).
3.6.6.3.2. Statistical uncertainty in parameter estimates and extrapolation. Klein et al. (2009)
developed methods for deriving statistical inferences of results from PBPK models. They used
the Conolly et al. (2000) model for this purpose, specifically because of the sparse time-course
information in the above DPX data. However, they used the value of kioss deduced from
Quievryn and Zhitkovich (2000) and fitted the model simultaneously to both the rat and rhesus
monkey data, as opposed to the sequential fitting in Conolly et al. (2000). They found that the
predicted DPX concentrations were extremely sensitive to Vmax and tissue thickness as was also
concluded by Georgieva et al. (2003) and Cohen-Hubal et al. (1997). Km was seen to be
substantially different across species, a finding that was attributed plausibly to the involvement
of more than one enzyme (Klein et al., 2009; Georgieva et al., 2003). Klein et al. (2009)
concluded that the two efforts (Conolly et al. [2000] vs. Klein et al. [2009]) resulted in
substantially different predictions outside the range of the observed data over which the models
were calibrated.
The differences between these models occur in spite of the fact that both methods use all the
available DPX data in both species and the same model structures. At the 0.1 ppm exposure
concentration, in general these authors obtained three- to fourfold higher DPX concentrations
averaged over a 24-hour period after exposure. Furthermore, the standard deviations in Klein et
al. (2009) for Vmax and Km were an order of magnitude higher and that for kf was 35-fold lower
than the corresponding standard deviations reported in Conolly et al. (2000). The relatively
larger standard deviation for kf resulted in this parameter becoming negative in Conolly et al.
(2000) at even half the standard deviation below the maximum likelihood estimate (MLE) value.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Note that, at a negative value of kf, formaldehyde would be produced as opposed to being
cleared through its intrinsic reactivity.
Klein et al. (2009) concluded that these "remarkable differences outside the range of the
observed data suggest caution in the use of these models in a predictive sense for extrapolating to
human exposures."
3.6.7. Uncertainty in Prediction of Human DPX Concentrations
Conolly et al. (2000) used both the rat and rhesus monkey data to predict human DPX
concentrations and constructed a PBPK model for the rhesus monkey along similar lines as for
the F344 rat. In the rhesus monkey model, they maintained the same values of kb, kioss, and kf as
in the rat model but optimized the values of Vmax and Km against the rhesus monkey data from
Casanova et al. (1994). The rat and rhesus monkey parameters were then used to construct a
human model (see Conolly et al. [2000] for a more detailed report of implementing the rhesus
monkey model and the extrapolations to the situation in humans).
For the human, the model used the value of Km obtained in the rhesus monkey model and
the epithelial thickness averaged over three regions of the rhesus monkey nose. The maximum
rate of metabolism, Vmax, which was estimated independently for the rat and rhesus monkey by
fitting to the DPX data available for these species, was then extrapolated to the human by
assuming a power law scaling with body weight (BW) (i.e., Vmax = a x BWb), and the coefficient
"a" and exponent "b" were derived from the independently estimated values of (Vmax)RAT and
(Vmax)MONKEY- Table 3-7 gives the values of Vmax and Km in the Conolly et al. (2000)
extrapolation.
Table 3-7. Extrapolation of parameters for enzymatic metabolism to the
human
Parameter
F344 rat
Rhesus monkey
Human
Vmax (pmol/min-mm3)
1,008.0
91.0
15.7
Km (pmol/mm3)
70.8
6.69
6.69
Source: Conolly et al. (2000).
The extent of mechanistic data across species, as available in this case, is rarely seen with
other chemicals, and the above scale-up procedure was an attempt to use both the rodent and
primate DPX data. However, allometric relationships across species are generally based on
regressing data from multiple species and usually multiple sources of data points. Thus, the
empirical strength of a power law derived by using two data points (F344 rat and rhesus
This document is a draft for review purposes only and does not constitute Agency policy.
3-53 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
monkey) is extremely weak for use as an allometric relationship that can then be used to
extrapolate to the human.
The following observations indicate the high level of uncertainty in the values of the
parameters Vmax and Km in the Conolly et al. (2000) models for predicting DPXs. First, Km
varies by an order of magnitude across the rat and monkey models but is then considered
invariant between the monkey and human models (Conolly et al., 2000). Second, the values in
Conolly et al. (2000) for Vmax/Km, the low-dose limit of the rate of enzymatic metabolism, is
roughly similar between the rat and monkey but lower by a factor of six in the human.
Another factor that can substantially influence the above extrapolation of DPXs in the
human is that Conolly et al. (2000) assumed the tissue to be a well-mixed compartment with
regard to formaldehyde interaction with DNA and used the amount of formaldehyde bound to
DNA per unit volume of tissue as the DPX dose metric. Considering formaldehyde's highly
reactive nature, the concentrations of formaldehyde and DPX are likely to have a sharp gradient
with distance into the nasal mucosa (Georgieva et al., 2003). Given the interspecies differences
in tissue thickness, there is consequent uncertainty as to whether DPX per unit volume or DPX
per unit area of nasal lining is the more appropriate dose metric to be used in the extrapolation.
In particular, it may be assumed that the cells at risk for tumor formation are only those in the
epithelium and that measured DPX data (in monkeys and rats) are an average over the entire
tissue thickness. Since the epithelial DPXs in monkeys (and presumably humans) would then be
more greatly "diluted" by lower levels of DPX formation that occur deeper into the tissue than in
rats, it could be predicted that the ratio of epithelial to measured DPXs in monkeys and humans
would be much higher than the ratio in rats.
End of Volume I
This document is a draft for review purposes only and does not constitute Agency policy.
3-54 DRAFT—DO NOT CITE OR QUOTE
------- |