A EPA
Toxicological Review of Formaldehyde - Inhalation
EPA/635/R-21/286C
www.epa.gov/iris
ASSESSMENT OVERVIEW
for the
TOXICOLOGICAL REVIEW OF FORMALDEHYDE -
INHALATION
[CASRN 50-00-0]
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
December 2021
NOTICE
This document is an Interagency Science Consultation draft. This information is distributed solely for
the purpose of pre-dissemination peer review under applicable information quality guidelines. It has
not been formally disseminated by EPA. It does not represent and should not be construed to represent
any Agency determination or policy. It is being circulated for review of its technical accuracy and science
policy implications.
Integrated Risk Information System
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
This document is a draft for review purposes only and does not constitute Agency policy.
i DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
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.
ii DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
CONTENTS
AUTHORS xii
1. SUMMARY AND ASSESSMENT METHODS 1
1.1. ASSESSMENT METHODS 5
1.1.1. Literature Search and Screening 6
1.1.2. Study Evaluation 7
1.1.3. Results Display and Evidence Synthesis 11
1.1.4. Evidence Integration 14
1.1.5. Dose-response Analysis 25
2. SUMMARY OF TOXICOKINETICS 26
3. NONCANCER HEALTH EFFECTS 28
3.1. SENSORY IRRITATION 29
3.1.1. Literature Identification 29
3.1.2. Study Evaluation 30
3.1.3. Synthesis of Human Health Effect Studies 30
3.1.4. Mode-of-action Information 32
3.1.5. Overall Evidence Integration Judgment and Susceptibility for Sensory Irritation 33
3.1.6. Dose-response Analysis 34
3.2. PULMONARY FUNCTION 38
3.2.1. Literature Identification 39
3.2.2. Study Evaluation 39
3.2.3. Synthesis of Human Health Effect Studies 40
3.2.4. Mode-of-action Information 43
3.2.5. Overall Evidence Integration Judgment and Susceptibility for Pulmonary Function 44
3.2.6. Dose-response Analysis 45
3.3. IMMUNE-MEDIATED CONDITIONS, FOCUSING ON ALLERGIES AND ASTHMA 47
3.3.1. Literature Identification 48
3.3.2. Study Evaluation 48
3.3.3. Synthesis of Human Health Effect Studies 49
3.3.4. Mode-of-action Information 54
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
3.3.5. Overall Evidence Integration Judgments and Susceptibility for Immune-mediated
Conditions including Allergies and Asthma 56
3.3.6. Dose-response Analysis 57
3.4. RESPIRATORY TRACT PATHOLOGY 64
3.4.1. Literature Identification 64
3.4.2. Study Evaluation 65
3.4.3. Synthesis of the Human Health Effect Studies 66
3.4.4. Synthesis of the Animal Health Effect Studies 66
3.4.5. Mode-of-action Information 70
3.4.6. Overall Evidence Integration Judgment and Susceptibility for Respiratory Tract
Pathology 71
3.4.7. Dose-response Analysis 72
3.5. NERVOUS SYSTEM EFFECTS 75
3.5.1. Literature Identification and Study Evaluation 76
3.5.2. Evidence Synthesis and Overall Evidence Integration Judgement for Nervous
System Effects 76
3.6. REPRODUCTIVE AND DEVELOPMENTAL TOXICITY 77
3.6.1. Literature Identification 77
3.6.2. Study Evaluation 78
3.6.3. Developmental and Female Reproductive Toxicity 78
3.6.4. Male Reproductive Toxicity 81
3.6.5. Dose-response Analysis 84
3.7. REFERENCE CONCENTRATION (Rfc) FOR NONCANCER HEALTH EFFECTS 88
3.7.1. Summary of cRfCs and osRfCs across Noncancer Health Effects 88
3.7.2. Selection of the RfC and Discussion of Confidence 90
3.7.3. Basis and Interpretation of the RfC 92
3.7.4. Previous IRIS Assessment: Reference value 94
4. CARCINOGENICITY 94
4.1. METHODS FOR IDENTIFYING AND EVALUATING STUDIES 95
4.1.1. Literature Identification 95
4.1.2. Study Evaluation 95
4.2. UPPER RESPIRATORY TRACT CANCERS 96
4.2.1. Synthesis of Human Health Effect Studies 97
4.2.2. Synthesis of Animal Health Effect Studies 103
This document is a draft for review purposes only and does not constitute Agency policy.
iv DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
4.2.3. Mode-of-action Information 106
4.2.4. Overall Evidence Integration Judgments and Susceptibility for Upper Respiratory
Tract Cancers 109
4.3. LYMPHOHEMATOPOIETIC (LHP) CANCERS Ill
4.3.1. Synthesis of Human Health Effect Studies Ill
4.3.2. Synthesis of Animal Health Effect Studies 119
4.3.3. Mode-of-action Information 120
4.3.4. Overall Evidence Integration Judgments and Susceptibility for LHP Cancers 123
4.4. WEIGHT-OF-EVIDENCE SUMMARY FOR CARCINOGENICITY 125
4.4.1. Weight-of-evidence Narrative Summary 126
4.5. INHALATION UNIT RISK (IUR) FOR CARCINOGENICITY 128
4.5.1. Derivation of Cancer Unit Risk Estimates for Nasal Cancers 129
4.5.2. Derivation of Cancer Unit Risk Estimates for Myeloid Leukemia 144
4.5.3. Estimates of Cancer Risk based on "Bottom-up" Approach 152
4.5.4. Summary of Unit Risk Estimates and the Preferred Estimate for Inhalation Unit
Risk 153
4.5.5. Previous IRIS Assessment: Inhalation Unit Risk 158
5. REFERENCES 159
This document is a draft for review purposes only and does not constitute Agency policy.
v DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
TABLES
Table 1. Evidence integration judgments for noncancer health effects and the reference
concentration (RfC) 2
Table 2. Cancer evidence integration judgments, carcinogenicity descriptor and inhalation unit
risk (IUR) for cancer incidence 3
Table 3. General approach to literature search strategies 6
Table 4. Information most relevant to describing primary considerations informing causality
during evidence syntheses 12
Table 5. Considerations that inform judgments regarding the strength of the human and animal
evidence 16
Table 6. Framework for strength of evidence judgments (human evidence) 19
Table 7. Framework for strength of evidence judgments (animal evidence) 20
Table 8. Overall evidence integration judgments for characterizing the integrated evidence for
noncancer health effects and cancer outcomes 23
Table 9. Criteria for applying cancer descriptors to evidence integration judgments for cancer
types 24
Table 10. Evidence integration summary for effects on sensory irritation 34
Table 11. Eligible studies for POD derivation and rationale for decisions to not select specific
studies 35
Table 12. Summary of derivation of PODs for sensory irritation 36
Table 13. Derivation of cRfCs for sensory irritation 37
Table 14. Evidence integration summary for effects on pulmonary function 45
Table 15. Eligible studies for POD derivation and rationale for decisions to not select specific
studies 46
Table 16. Summary of derivation of PODs for pulmonary function 46
Table 17. Derivation of the cRfCfor pulmonary function 47
Table 18. Evidence integration summary for effects on immune-mediated conditions, including
allergies and asthma 56
Table 19. Eligible studies for POD derivation and rationale for decisions to not select specific
studies 58
Table 20. Summary of derivation of PODs for allergies and current asthma based on
observational epidemiological studies 60
Table 21. Derivation of the cRfC for allergy-related conditions and asthma 63
Table 22. Evidence Integration Summary for Effects of Formaldehyde Inhalation on Respiratory
Pathology 71
Table 23. Eligible studies for POD derivation and rationale for decisions to not select specific
studies 72
Table 24. Summary of derivation of PODs for squamous metaplasia 74
Table 25. Derivation of cRfCs for respiratory tract pathology 75
Table 26. Evidence integration summary for effects of formaldehyde inhalation on
developmental or female reproductive toxicity in humans 80
Table 27. Evidence integration summary for effects of formaldehyde inhalation on reproductive
toxicity in males 83
Table 28. Eligible studies for POD derivation and rationale for decisions to not select specific
analyses 84
This document is a draft for review purposes only and does not constitute Agency policy.
vi DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Table 29. Summary of derivation of PODs for developmental and reproductive toxicity in
females 85
Table 30. Summary of derivation of PODs for reproductive toxicity in males 86
Table 31. Derivation of cRfCs for Female Reproductive or Developmental Toxicity and Male
Reproductive Toxicity 87
Table 32. Organ/System-specific RfCs (osRfCs) for formaldehyde inhalation 90
Table 33. Proposed RfC for formaldehyde-inhalation 92
Table 34. Summary considerations for upper respiratory tract (URT) carcinogenesis (the primary
support for genotoxicity or mutagenicity is noted; see Toxicological Review for
additional details) 107
Table 35. Evidence integration summary for effects of formaldehyde inhalation on URT cancers 110
Table 36. Incidence of hematopoietic cancers in B6C3F1 mice and F344 rats [source: (Kerns et
al., 1983; Battelle, 1982)] 120
Table 37. Summary conclusions regarding plausible mechanistic events associated with
formaldehyde induction of lymphohematopoietic cancers 121
Table 38. Evidence integration summary for effects of formaldehyde inhalation on LHP cancers 124
Table 39. Relative risk estimates for mortality from NPC (based on ICD code) and regression
coefficients from NCI log-linear trend test models3 by level of cumulative
formaldehyde exposure (ppm x years). Source: Beane Freeman et al. (2013) 131
Table 40. ECooos, LECooos, and unit risk estimates for nasopharyngeal cancer mortality based on
the Beane Freeman et al. (2013) log-linear trend analyses for cumulative
formaldehyde exposure 132
Table 41. ECooos, LECooos, and unit risk estimates for nasopharyngeal cancer incidence based on
the Beane Freeman et al. (2013) log-linear trend analyses for cumulative
formaldehyde exposure 133
Table 42. F344 rat nasal cancer data 135
Table 43. Benchmark concentrations and human equivalents using formaldehyde flux to nasal
tissue as a dose-metric 135
Table 44. BBDR model estimated extra risk of SCC in human respiratory tract compared with
EPA's modeling of extra risk of NPC from the human occupational data 140
Table 45. Unit risk estimates derived from benchmark estimates using animal data and
formaldehyde flux as dose-metric 142
Table 46. Comparison and basis of unit risk estimates for NPC in humans and nasal SCCs in rats 142
Table 47. Strengths and uncertainties in the cancer type-specific unit risk estimate for NPC 144
Table 48. Relative risk estimates for mortality from leukemia (based on ICD codes) and
regression coefficients from NCI log-linear trend test models by level of
cumulative formaldehyde exposure (ppm x years). Source: Beane Freeman et
al. (2009) 146
Table 49. ECoos, LECoos, and unit risk estimates for myeloid plus other/unspecified leukemia
mortality based on log-linear trend analyses of cumulative formaldehyde
exposure data from the Beane Freeman et al. (2009) study 148
Table 50. ECoos, LECoos, and unit risk estimates for myeloid plus other/unspecified leukemia
incidence based on Beane Freeman et al. (2009) log-linear trend analyses for
cumulative formaldehyde exposure 149
Table 51. ECoos and LECoos estimates for mortality and incidence and unit risk estimates for all
leukemia and myeloid leukemia using alternate approaches (all person-years) -
shaded estimate is preferred 149
This document is a draft for review purposes only and does not constitute Agency policy.
vii DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Table 52. Dose-response modeling (all person-years) and (incidence) unit risk estimate
derivation results for different leukemia groupings - shaded estimate is
preferred 150
Table 53. Strengths and uncertainties in the cancer type-specific unit risk estimate for myeloid
leukemia 151
Table 54. Summary of inhalation unit risk estimates from occupational epidemiological studies 153
Table 55. Inhalation unit risk 156
This document is a draft for review purposes only and does not constitute Agency policy.
viii DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
FIGURES
Figure 1. Overview of assessment methods for hazard identification 5
Figure 2. Process for evidence integration 15
Figure 3. Schematic of the rat upper respiratory tract depicting the gradient of formaldehyde
concentration formed following inhalation exposure 27
Figure 4. Prevalence of eye irritation among study groups exposed to formaldehyde in
residential settings and controlled human exposure studies 32
Figure 5. Likely mechanistic association between formaldehyde exposure and sensory irritation 33
Figure 6. Association of PEFR measured at bedtime and in the morning with household mean
formaldehyde concentration among children less than 15 years of age
(Krzyzanowski et al., 1990) 42
Figure 7. Possible mechanistic associations between formaldehyde exposure and decreased
pulmonary function 43
Figure 8. Relative risk estimates for prevalence of allergy-related conditions in children and
adults in relation to formaldehyde in residential and school settings 51
Figure 9. Relative risk estimates for prevalence of asthma in children and adults in relation to
formaldehyde by exposure level in general population and occupational studies 53
Figure 10. Possible mechanistic associations between formaldehyde exposure and immune-
mediated conditions, including allergic conditions and asthma 55
Figure 11. Squamous metaplasia incidence in chronic pathology studies of rats 68
Figure 12. Possible mechanistic associations between formaldehyde exposure and respiratory
tract pathology 71
Figure 13. Candidate RfCs (cRfCs) with corresponding POD and composite UF 89
Figure 14. Organ or system-specific RfC (osRfC) scatterplot 91
Figure 15. Illustration of noncancer toxicity value estimations 93
Figure 16. All epidemiological studies reporting nasopharyngeal cancer risk estimates 98
Figure 17. All epidemiological studies reporting sinonasal cancer risk estimates 101
Figure 18. All epidemiological studies reporting oropharyngeal/hypopharyngeal cancer risk
estimates 103
Figure 19. Incidence of nasal squamous cell carcinomas in rats exposed to formaldehyde for at
least 2 years 105
Figure 20. All epidemiological studies reporting myeloid leukemia risk estimates 114
Figure 21. Epidemiological studies reporting acute myeloid leukemia risk estimates 115
Figure 22. Epidemiological studies reporting paired estimates of acute myeloid leukemia risk
estimates and myeloid leukemia risk estimates 116
Figure 23. All epidemiological studies reporting multiple myeloma risk estimates 118
This document is a draft for review purposes only and does not constitute Agency policy.
ix DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
ABBREVIATIONS
Acronym
Definition
ADAF
age-dependent adjustment factors
ADME
absorption, distribution, metabolism, excretion
ALS
amyotrophic lateral sclerosis
AML
acute myeloid leukemia
ATS
American Thoracic Society
BAL
bronchoalveolar lavage
BBDR
biologically based dose-response
BMC
benchmark concentration
BMR
benchmark response
CASRN
Chemical Abstracts Service Registry Number
CFD
computational fluid dynamics
CFU
colony-forming unit
CML
chronic myeloid leukemia
cRfC
candidate reference concentration
DPX
DNA-protein cross-link(s)
ETS
environmental tobacco smoke
FDR
fecundability density ratio
FEF
forced expiratory flow
FVC
forced vital capacity
GLP
good laboratory practice
GM
granulocyte, monocyte
HEC
human equivalent concentration
HERO
Health and Environmental Research Online
HSPC
hematopoietic stem or progenitor cell
IARC
International Agency for Research on Cancer
ICD
International Classification of Disease
IUR
inhalation unit risk
LEC
lowest effective concentration
LH
luteinizing hormone
LHP
lymphohematopoietic
LOAEL
lowest-observed-adverse-effect level
LOD
limit of detection
LRT
lower respiratory tract
This document is a draft for review purposes only and does not constitute Agency policy.
x DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Acronym
Definition
MDA
malondialdehyde
MDS
myelodysplastic syndrome
MLE
maximum likelihood estimate
MOA
mode of action
MOR
mortality odds ratio
NALT
nasal associated lymphoid tissues
NCHS
National Center for Health Statistics
NCI
National Cancer Institute
NOAEL
no-observed-adverse-effect level
NPC
nasopharyngeal cancer
NRC
National Research Council
NTP
National Toxicology Program
OR
odds ratio
osRfC
organ/system-specific reference concentration
PBL
peripheral blood lymphocyte
PBPK
physiologically based pharmacokinetic
PECO
Populations, Exposures, Comparisons, Outcomes
PEFR
peak expiratory flow rate
PMR
proportionate mortality ratio
POD
point of departure
POE
portal of entry
RfC
reference concentration
ROS
reactive oxygen species
RR
relative risk
see
squamous cell carcinoma
SES
socioeconomic status
SMR
standardized mortality ratio
TSCE
two-stage clonal expansion
TSFE
time since first exposure
TTP
time-to-pregnancy
TWA
time-weighted average
UF
uncertainty factor
URT
upper respiratory tract
WBC
white blood cell
This document is a draft for review purposes only and does not constitute Agency policy.
xi DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
AUTHORS
Assessment Team*
Barbara Glenn (Chemical Manager)
Andrew Kraft (Chemical Manager)
Thomas Bateson
Susan Makris
Deborah Segal
Ravi Subramaniam
Suryanarayana Vulimiri
Glinda Cooper
Jason Fritz
Jennifer Jinot
John Whalan
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
U.S. EPA/ORD/CPHEA
Separated from U.S. EPA
U.S. EPA/Region 8
Retired from U.S. EPA
Retired from U.S. EPA
*Please see the Toxicological Review for a full list of contributors,
production team, contractor support, executive direction, and reviewers.
This document is a draft for review purposes only and does not constitute Agency policy.
xii DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Toxicological Review of Formaldehyde - Inhalation
1. SUMMARY AND ASSESSMENT METHODS
This IRIS health assessment presents a systematic evaluation of the publicly available studies
relevant to inhalation exposure to formaldehyde and potential adverse health outcomes. The purpose
of the review was to identify hazards that may result from formaldehyde inhalation and describe the
level of confidence in the supporting evidence. When there was sufficient confidence in the evidence
supporting a hazard and appropriate studies and data were available, toxicity values were derived using
either analyses of dose-response or selected no-adverse-effect or lowest-adverse-effect levels (NOAEL
or LOAEL). The results of the assessment are summarized in Tables 1 and 2.
The evidence identification, evaluation, and integration framework depicted in Figure 1 was
used to conduct the assessment. Potential health hazards were evaluated, including sensory irritation;
reduced pulmonary function; immune system effects, focusing on allergic conditions and asthma;
respiratory tract pathology; nervous system effects; reproductive and developmental toxicity; and
cancer. Several well-studied cancer sites were specifically evaluated, including cancers of the upper
respiratory tract (i.e., nasopharyngeal cancer, sinonasal cancer, cancers of the oropharynx/hypopharynx,
and laryngeal cancer) and of the lymphohematopoietic system (i.e., Hodgkin lymphoma, multiple
myeloma, myeloid leukemia, and lymphatic leukemia).
This document is a draft for review purposes only and does not constitute Agency policy.
1 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 1. Evidence integration judgments for noncancer health effects and the
2 reference concentration (RfC)
Noncancer Health Effect
Confidence in
Health Effect
POD Basis
Confidence in
POD
UFC
osRfC
(mg/m3)
Decreased pulmonary function
evidence indicates
[likely]
Human
high
3
0.007
Allergic conditions
evidence indicates
[likely]
Human
high
3
0.008
Current asthma symptoms or degree
of asthma control
evidence indicates
[likely]
Human
medium
10:
0.006:
Sensory irritation
evidence
demonstrates
Human
medium
10
0.009
Female reproductive or
developmental toxicity
evidence indicates
[likely]
Human
low
10
0.01
Respiratory tract pathology
evidence
demonstrates
Rat
medium
_Q
o
m
0.003b
Male reproductive toxicity
evidence indicates
[likely]
Rat
low
3000
0.001
Nervous system effects
evidence suggests
Not Derived
-
-
-
Confidence in
Health Effects
PODs Basis
Confidence in
PODs
UFC
Confidence in
Database
RfC
(mg/m3)
Overall
Confidence
RfC :
Medium or High
Human
Medium or High
3 or 10
High
0.007
High
Abbreviations and definitions: RfC, reference concentration: an estimate (with uncertainty spanning perhaps an
order of magnitude) of a continuous inhalation exposure of a chemical to the human population (including
sensitive subpopulations), that is likely to be without risk of deleterious noncancer effects during a lifetime.
osRfC, organ- or system-specific RfC: an RfC based on the evidence for effects on that particular organ or system.
UFC: composite (total) uncertainty factor; POD: point-of-departure.
aBasis for RfC—sensory irritation, decreased pulmonary function, current asthma symptoms or degree of asthma
control, and allergic conditions. The corresponding osRfCs (i.e., based on human studies with medium or high
confidence in the health effects and PODs) are highlighted in gray, which also have the lowest UFC values.
bThese two osRFCs and the RfC are based on multiple studies and candidate values, sometimes with different UFcs
applied. The UFC values shown in this table and Figure 2-2 reflect the candidate values selected to represent each
osRfC [i.e., the UFC applied to the POD from Krzyzanowski et al. (1990) for asthma and from Woutersen et al.,
(1989) for respiratory pathology].
This document is a draft for review purposes only and does not constitute Agency policy.
2 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 2. Cancer evidence integration judgments, carcinogenicity descriptor and
2 inhalation unit risk (IUR) for cancer incidence
Cancer Type Investigated
Evidence
integration
judgment for
Cancer Type
Risk
Unit Risk
Estimate Basis
Unit Risk
Estimate
(perng/m3)
ADAF-adjusted
Unit Risk Estimate
(per |ig/m3)a
Confidence in
the Unit Risk
Estimate
Nasopharyngeal cancer
(NPC) (nasal cancer in
animals)
evidence
demonstrates b
Human
6.4 x 10
1.1 X 10
medium
Animal0
8.9 x 10"6 to
1.8 x 10"5
NAd
medium
Myeloid leukemia
evidence
demonstrates e
Human
3.4 x 10"5
NAf
low
Sinonasal cancer
evidence
indicates [likely]
No usable
data
-
-
Oropharyngeal/
Hypopharyngeal cancer
evidence
indicates [likely]
No usable
data
-
-
Multiple myeloma
evidence
indicates [likely]
No usable
data
-
-
Hodgkin lymphoma
evidence
suggests
Not Derived
-
-
Laryngeal cancer
evidence
inadequate
Not Derived
-
-
Lymphatic leukemia
evidence
inadequate
Not Derived
-
-
Cancer Descriptor:
Carcinogenic to Humans
Total cancer risk (IUR) :
1.1 x 10"5 per |ig/m3; Confidence in the IUR is Medium
Abbreviations and definitions: IUR, inhalation unit risk: the upper-bound excess lifetime cancer risk estimated to
result from continuous exposure to an agent at a concentration of 1 ng/m3 in air. ADAF: age-dependent
adjustment factor.
aADAF adjustments are recommended for cancers for which there is sufficient evidence that formaldehyde has, at
least in part, a mutagenic mode of action (MOA) (see Section 2.2.4).
b The judgment of evidence demonstrates for NPC cancer is based on robust human evidence of increased risk in
groups exposed to occupational formaldehyde levels, and robust animal evidence of nasal cancers in rats and
mice that exhibits steeply increasing incidence at high formaldehyde levels. Strong mechanistic support is
provided across species (primarily rats, but also mice, monkeys, and humans), including genotoxicity, epithelial
damage or remodeling, and cellular proliferation that are consistent with neoplastic development in a regional,
temporal, and dose-related fashion.
This document is a draft for review purposes only and does not constitute Agency policy.
3 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
cWhile the preferred unit risk estimate for NPC is based on a cancer mortality study in humans, several estimates in
general agreement with each other were also derived based on animal nasal tumor incidence. These estimates
used multiple mechanistic and statistical models, including biologically based dose-response (BBDR) modeling
(see Section 2.2.1). In addition, cRfCs for one mechanism contributing to nasal cancer development, specifically
cytotoxicity-induced regenerative cell proliferation, were estimated to be between 0.006 and 0.018 mg/m3 based
on calculations using animal data. Specifically, this narrow range of cRfCs was estimated based on PODs from a
pathology study of hyperplasia, labeling studies of proliferating cells, and BBDR modeling results (see Section
2.2.1).
dNA= not applicable; an ADAF-adjusted value was not calculated for the unit risk estimates based on the animal
data on nasal cancer, as the human unit risk estimate for NPC was the preferred estimate.
e The judgment of evidence demonstrates for myeloid leukemia is based on robust human evidence of increased
risk in groups exposed to occupational formaldehyde levels. Supporting mechanistic evidence consistent with
leukemia development is provided across numerous studies of peripheral blood isolated from exposed workers,
including evidence of mutagenicity and other genotoxic damage in lymphocytes and myeloid progenitors, and
perturbations to immune cell populations. The animal evidence is inadequate and the findings to date suggest
that there may be a lack of concordance across species for leukemia, as leukemia was not increased in two well-
conducted chronic bioassays of rats or mice, and the available animal data provide weak mechanistic support for
LHP cancers. No MOA has been established to explain how formaldehyde inhalation can cause myeloid leukemia
without systemic distribution (inhaled formaldehyde does not appear to be distributed to an appreciable extent
beyond the upper respiratory tract to distal tissues).
fNA = not applicable; no ADAF adjustment is recommended for myeloid leukemia because the MOA is unknown
(see Section 1.3.3).
gThe full lifetime (ADAF-adjusted) IIIR estimate is based on the ADAF-adjusted estimate for nasopharyngeal cancer
(which includes a mutagenic MOA; see Section 1.2.5). Less-than-lifetime exposure scenarios with a very large
fraction of exposure during adulthood may not warrant ADAF adjustment, and one may choose to use the
unadjusted unit risk estimate of 6.4 x 10"6 per ng/m3.
This document is a draft for review purposes only and does not constitute Agency policy.
4 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Literature Searches (Hazard-specific)
1
Reference retrieval
Reference lists
Inclusion/ exclusion criteria
(based on PECO)
Reference screening by hazard domain
o Excluded references grouped by PECO category
o Included references grouped by lines of
evidence (human, animal, mechanistic)
o Literature search diagrams by hazard domain
~
Evaluation of study methods (Outcome/ endpoint specific)
Outcome-specific
evaluation criteria
for health effects
studies in humans
and animals;
informed by ADME
research
Study evaluation tables
o By outcome and study
o Study confidence by outcome
Medium
iT
Syntheses of results
Low
Unlnformative
Interpretation of
results from health
effect studies in
humans and animals
(consistency,
magnitude of effect,
dose-response,
coherence, etc.)
Evaluation and
interpretation of
mechanistic evidence
Judgments on health effects
separately for human or animal
studies
based on health effects and biological
plausibility from mechanistic studies
Overall evidence judgments
regarding potential of chemical to
cause health effects in humans, using
judgments on human and animal
evidence and mechanistic inference3
Evidence demonstrates
Evidence indicates (likely)
Evidence suggests
Evidence inadequate
1 Figure 1. Overview of assessment methods for hazard identification.
This figure illustrates the flow of evidence through the assessment, sequentially focusing on the most
useful information, as well as the decision-making processes for arriving at evidence integration
judgments regarding the potential for noncancer health effects and for developing specific types of
cancer. * Mechanistic inference considered during evidence integration included biological plausibility,
relevance of animal study results to humans, and identification of susceptible groups. Notes: for this
assessment, "compelling evidence of no effect" was not reached for any evidence and, as such, is not
discussed further. Importantly, hazard identification for carcinogenicity includes an additional step of
assigning a descriptor regarding the potential for formaldehyde to cause cancer (this step is not shown,
but is discussed in Section 1.1 below). Abbreviations: HERO, Health and Environmental Research Online;
PECO, Populations, Exposures, Comparisons, Outcomes; ADME, absorption, distribution, metabolism,
excretion; MOA, mode-of-action.
2 1.1. ASSESSMENT METHODS
3 The approaches implemented throughout different stages of this assessment can be grouped
4 into (1) those used to identify and evaluate individual studies; (2) those used to synthesize the evidence,
5 including interpreting the degree of support for particular human health hazards by integrating different
6 lines of evidence (i.e., human, animal, and mechanistic studies) and coming to summary conclusions;
7 and (3) selecting and analyzing studies and data to derive quantitative (dose-response) toxicity values.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
The process involves a successive focusing on the more informative outcomes/endpoints within each
hazard domain and the most methodologically robust studies to judge and integrate the evidence
within, and across, the human, animal, and mechanistic evidence.
1.1.1. Literature Search and Screening
The literature search strategy used to identify primary research was conducted using the
databases and approaches listed in Table 3. A separate, systematic literature search strategy was
developed for each health effect considered in the assessment. These strategies are described in detail
in Appendix Section A.5, with populations, exposures, comparators, and outcomes (PECO) criteria, and
diagrams depicting the search and sorting process. Health effects and search terms were selected after
reviewing the draft Toxicological Review for Formaldehyde (U.S. EPA. 2010) and other relevant health
assessments or reviews of formaldehyde toxicity. A series of comprehensive literature searches was
conducted annually beginning in 2012 through 2016, after which the completed 2017 Step 1 draft IRIS
formaldehyde-inhalation assessment was suspended at the request of senior EPA management. When
the IRIS assessment was unsuspended in March 2021 (http://www.epa.gov/sites/production/files/2021-
03/documents/iris program outlook mar2021.pdf), systematic evidence mapping (SEM) methods were
employed to survey the newer literature and expedite updating the unsuspended draft (see Appendix F
for the methods and results of the formaldehyde SEM update). In these searches, electronic database
queries for published and unpublished studies were supplemented using various approaches to identify
additional papers, including review of reference lists in identified publications and national-level health
assessments.
Table 3. General approach to literature search strategies
Databases
Health Effect-specific Searches3
Web of Science
ToxNet
PubMed
TSCATS2
[formaldehyde, formalin, paraformaldehyde, OR CASN 50-00-0] AND:
Sensory Irritation15
Pulmonary Function15
Immune-Mediated Conditions, focusing on Allergies and Asthma
Respiratory Tract Pathology
Developmental and Reproductive Toxicity
Nervous System Effects
Cancer
Inflammation and Immune Effects (mechanistic information0)
aSpecific parameters and keywords for each hazard-specific database search strategy are included in Appendix A.5.
bA systematic search strategy was not applied to the database of animal studies on this health outcome. Sensory
irritation in animals is a well-described phenomenon. For pulmonary function, there was an extensive set of
research studies in humans, and therefore, the few studies on this endpoint in animals were not reviewed.
This separate, systematic literature search was performed to augment the analyses of mechanisms relevant to
other health effect-specific searches. Details are not included in this Overview (see Appendix A.5.6).
This document is a draft for review purposes only and does not constitute Agency policy.
6 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
The citations for primary health effects studies were screened using inclusion and exclusion
criteria based on health effect-specific Populations, Exposures, Comparisons, Outcomes (PECO)
considerations. In general, although studies of other routes of exposure might inform the mechanistic
understanding of potential health effects, the formaldehyde database is large and the toxicokinetics
following inhalation exposure are expected to differ significantly from those observed after exposure via
other routes (see Section 2 of this Overview); thus, mechanistic descriptions were focused on inhalation
exposure studies (an exception includes studies of genotoxicity). Ambient levels of formaldehyde in
outdoor air are significantly lower than those measured in the indoor air of workplaces or residences,
and the narrow range of exposures (encompassing 0.005 mg/m3 or less) evaluated in the few
epidemiological studies of outdoor exposure limited their sensitivity to find any associations with health
outcomes even if they existed. Therefore, the few studies examining health effects in relation to
outdoor formaldehyde concentrations were excluded. Other exclusions were based on specific
outcome-specific criteria relating to each health hazard, which are summarized in each of the respective
health effect-specific PECO tables (see Sections 3 and 4 of this Overview) and documented in the
Appendices.
In addition to the health effects listed in Table 3, relevant literature on additional topics (e.g.,
formaldehyde exposure, toxicokinetics) was identified. While a thorough effort was made to identify all
relevant studies for each of these topic areas, these discussions do not include specific tracking of the
selection of individual studies (e.g., PECO-based inclusion and exclusion criteria). The references
identified and selected through the literature search process (i.e., all included and excluded studies),
including bibliographic information and abstracts, can be found on the Health and Environmental
Research Online (HERO) web site: http://hero.epa.gov/formaldehyde.
For the literature update from 2016-2021 using SEM approaches (overlapping with the searches
used for the 2017 draft), while the aforementioned description of the search and screening process was
largely identical (see Appendix F) a few differences are important to note. Most notably, after screening
the studies for PECO relevance, only those studies meeting the PECO criteria and judged as likely to have
a potential impact on the conclusions or toxicity values described in the suspended 2017 draft are
synthesized in this assessment. Studies meeting PECO criteria that were judged to have no impact on
those conclusions or toxicity values are summarized in Appendix F, along with explanations for these
decisions. These latter studies are not further discussed or synthesized in the assessment.
1.1.2. Study Evaluation
All human and experimental animal health effect studies identified in the search and included by
the screening process described above, without regard to study results, were considered for use in
assessing the evidence for health effects associated with inhalation exposure to formaldehyde. Study
methods were evaluated to assign a level of confidence in the results of the study with respect to the
health outcome under consideration. These evaluations were performed on a health outcome-specific
This document is a draft for review purposes only and does not constitute Agency policy.
7 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
basis, rather than a study-specific basis; thus, a single study was sometimes evaluated multiple times for
different endpoints, sometimes involving slightly different considerations. The evaluations focused on
potential sources of bias or other limitations (including reduced sensitivity) that can affect the validity or
interpretation of study results. The general procedure involved evaluating specific methodological
features, which differed somewhat between observational epidemiology, animal toxicology, and human
controlled exposure studies. Sets of studies for each hazard-specific outcome were evaluated by a
primary reviewer. The results of the evaluations were then commented on by a second reviewer who
also evaluated each study; the evaluation decisions were discussed, and any differences were resolved.
A study confidence level was drawn and the evaluation, including relevant study characteristics and an
indication of the expected impact of any identified limitations on the results (when possible), was
documented in tables (see Appendices A.5.2 - A.5.5, A.5.7 - A.5.9).
Systematic evaluations of individual mechanistic studies were also conducted in relation to
several important health hazards where a reasonable number of studies were available but the
mechanistic interpretations were not well established. Specifically, this included: biomarkers of
genotoxicity in exposed humans, and mechanistic data related to potential nervous system effects or
potential respiratory health effects. For these studies, the literature identification methods and study
confidence conclusions were similarly documented (see Appendices A.4.7, A.5.6, A.5.7).
The study confidence levels were high, medium, and low confidence, and not informative, and
are presented as italicized text in the various assessment documents. High confidence studies generally
had no significant methodological limitations for an outcome, while medium confidence studies were
considered well conducted but had specific issues that might introduce some uncertainty about
attribution of the results solely to formaldehyde exposure on the health outcome in question.
Methodological limitations of low confidence studies are considered significant, but the outcome-
specific results might still be of limited use (e.g., as support for observations from other studies; to
identify potential data gaps). The results of studies identified as not informative are not discussed in the
Toxicological Review or this Overview.
In some situations, study author(s) were contacted to obtain key study details or results that
were not presented. A decision to contact an author was based on whether the missing information
might result in the reevaluation of methodological features and possibly change the study confidence
level, or if it was useful for dose-response analyses. Any additional study details obtained from the
authors are noted.
Evaluation of observational epidemiological studies
For each type of health outcome examined, the studies were evaluated for each of the
categories of information relevant to internal validity (bias) that could lead to an under- or overestimate
of risk, and to other features that could affect the interpretation of the results or limit the ability to
detect a true association (e.g., narrow exposure range). The categories used for the epidemiological
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
studies included: population selection, exposure (measurement and levels/range), outcome
ascertainment, consideration of confounding, and analytic approach. The potential for selection bias,
information bias (relating to exposure and to outcome), and confounding were evaluated. The expected
direction of bias was explicitly considered and the impact of a potential bias on effect estimates was
incorporated in the determination of overall confidence. Emphasis was placed on discerning a bias that
would be expected to produce a substantive change in the estimated effect estimate, which resulted in
a categorization of low confidence. If a study or individual analysis was judged to have multiple severe
limitations, or if reporting deficiencies precluded the ability to conduct an evaluation for multiple
categories, a study or individual analysis was concluded to be not informative.
Evaluation of controlled exposure studies in humans
A process incorporating aspects of the evaluation approaches used for epidemiological studies
and experimental animal studies (see below) was used to evaluate controlled exposure studies in
humans. The evaluation categories included: exposure generation, outcome classification,
consideration of possible bias (i.e., randomization and blinding), consideration of confounding (i.e.,
adequacy of randomization), and details of analysis and presentation of results. A study was judged to
be low confidence if the exposure generation method resulted in exposure to substances other than
formaldehyde, allocation to the order of exposure categories was not random, or if subjects were not
blinded to their exposure order.
Evaluation of animal toxicological studies
In general, toxicology study evaluations considered related categories to the epidemiological
studies. The categories were based on the design of a toxicology study, and included test animals,
experimental design (e.g., duration of exposure, timing of endpoint evaluations, allocation procedures),
exposure conduct, endpoint evaluation procedures, and data presentation and analysis. Since
experimental studies should attempt to control all variables, any study limitation interpreted as capable
of influencing the data was considered to have negatively affected the quality (e.g., validity, accuracy) of
the results. Thus, these "confounding factors" differ from what would be deemed a potential
"confounder" in epidemiological studies. Observations in low confidence experimental animal studies
were determined to have a high likelihood of being influenced by factors other than formaldehyde
exposure alone, or there were significant concerns that the observations were attributable to non-
specific effects (e.g., toxicity; irritation).
Evaluation of mechanistic studies
For the datasets described previously, evaluations of individual mechanistic studies involving
formaldehyde inhalation in experimental animals or in vitro models of gaseous formaldehyde exposure
considered the same general features evaluated for more apical measures of toxicity (i.e., evaluations of
exposure quality and study design were emphasized). The specific criteria, however, were simplified to
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
accommodate the increased heterogeneity of the available mechanistic studies, as compared to the
data available for apical measures of toxicity. Similarly, study evaluations of individual mechanistic
studies involving exposed humans emphasized consideration of exposure assessment, study design,
outcome ascertainment, and comparison groups for potential sources of bias and their potential impact.
For the mechanistic studies related to potential noncancer respiratory effects, given the large number of
studies identified, individual studies were characterized as high or medium confidence, low confidence,
or not informative. Subsequent to this, groupings of studies or related endpoints were evaluated to
assess the strength of the evidence for different "mechanistic events" as robust, moderate, slight, or
indeterminate. Robust evidence required multiple high or medium confidence studies, while moderate
evidence required at least one high or medium confidence study and some supporting information (see
Appendix A.5.6 for additional details). For studies of genotoxicity biomarkers in exposed humans, a
confidence level of high, medium, low, or not informative was assigned to each study, consistent with
evaluations of human health effect studies.
Exposure-specific considerations in experimental studies
Experimental exposure to formaldehyde by inhalation is typically achieved through volatilization
of formalin or depolymerization of paraformaldehyde. Methanol, present in aqueous formaldehyde
solutions to inhibit polymerization, is a potential confounder of associations between observed effects
and formaldehyde exposure and, because methanol can be converted to formaldehyde endogenously, it
can also introduce quantitative uncertainty. Thus, a critical evaluation of exposure quality (with an
emphasis on the test article used to generate formaldehyde) was applied to experimental studies,
although conclusions about the level of concern varied by health outcome. Specifically, far greater
concern was raised for potential impacts of methanol coexposure on non-respiratory health effects (i.e.,
nervous system effects, developmental and reproductive system effects, LHP cancers), as compared to
respiratory health effects. This disproportionate level of concern was primarily based on two factors: (1)
as compared to formaldehyde, which does not appear to be distributed to systemic sites in appreciable
amounts, inhaled methanol would be readily transported beyond the URT and could elicit direct effects
at distal target tissues; and (2) certain, systemic effects evaluated in this assessment (i.e., nervous
system effects; reproductive and developmental toxicity) are known to be a target of methanol toxicity,
while other health effects, although they are generally less well studied, have not been clearly
associated with methanol exposure. Separately, for some endpoints (e.g., nervous system effects), the
study evaluations also considered the potential impact of the irritant and odorant nature of
formaldehyde gas, and the inescapable nature of these exposures (animals cannot terminate exposure
at irritating levels), which can complicate interpretations of causality. Similarly, uncertainties introduced
by phenomena such as reflex bradypnea, an irritant response to formaldehyde that can occur in rodents
but not humans, are discussed in the evidence syntheses. Thus, during study evaluation, care was taken
This document is a draft for review purposes only and does not constitute Agency policy.
10 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
to consider the exposure protocols in detail, including the duration between exposure and testing and
whether the tested exposure levels were likely to introduce variables such as reflex bradypnea.
1.1.3. Results Display and Evidence Synthesis
For each hazard category, or specific hazard endpoint, and depending on the data available,
separate syntheses were developed for each of the three lines of evidence: namely, human and animal
health effect studies, and mechanistic studies. One notable exception is the mechanistic evidence
related to potential respiratory health effects. Given the abundant data available and the assumed
interdependence of the mechanisms involved across these health effects, the data were identified and
evaluated in a single overarching analysis (see Appendix A.5.6). For brevity, while detailed discussions
and analyses are included in the Appendix, the mechanistic syntheses in the Toxicological Review focus
on the primary conclusions that could be drawn from the analysis and any outstanding issues,
uncertainties, or data gaps that might remain. The evidence syntheses, which incorporate the
evaluations of the strengths and limitations of the available studies, are narrative summaries analyzing
the information provided by each line of evidence regarding the potential for exposure to formaldehyde
via inhalation to result in specific health effects. All informative human and animal health effect studies
(see above), or mechanistic studies (i.e., when individual studies were evaluated) were considered in
assessing the evidence; however, the focus of the synthesis was on the high and medium confidence
studies, when available. Low confidence studies supported the evaluation of consistency, or if no or few
higher confidence studies were available, low confidence studies were considered in greater depth.
Descriptive information about study methods and detailed results were generally presented in tabular
or graphical displays, with supportive text in the Toxicological Review. The evidence syntheses discuss
the nature and breadth of the available literature, highlighting details that contribute to the analysis of
the strength of the evidence within and across the three lines of evidence, as described in the next
section.
The synthesis of the separate lines of evidence, human health effect studies, animal health
effect studies, and mechanistic studies, involved related considerations that differed due to the nature
of the study designs and applicability of the data considered within each line of evidence (Table 4).
Consistency, magnitude of effects, and dose-response gradients were emphasized in the synthesis of
results of epidemiological and controlled human exposure studies. While the precision of effect
estimates could add to the strength of evidence for a health effect, all of the results were summarized,
regardless of precision. Consistency between studies was examined by comparing study results by
confidence level, specific methodological features that contributed to potential bias, exposure setting,
and level of exposure. The primary considerations for synthesizing the results of animal studies were
consistency (e.g., across species and across research groups, with consideration of study confidence),
magnitude and severity of the effects, dose-response, and coherence of findings for related effects. The
information from mechanistic studies in humans or animals relevant to each apical outcome was
This document is a draft for review purposes only and does not constitute Agency policy.
11 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 synthesized, highlighting information that could inform either biological plausibility, susceptibility,
2 relevance to humans, or an improved understanding of dose-response. Given the exposure-related
3 issues specific to formaldehyde, and the abundance of data available, the mechanistic evaluations in this
4 assessment focused almost exclusively on in vivo studies of inhalation exposures, with rare exception
5 (e.g., evaluation of in vitro genotoxicity studies).
6 Table 4. Information most relevant to describing primary considerations informing
7 causality during evidence syntheses
Consideration
Description and Synthesis Methods
Consistency
• Examines the similarity of results (e.g., direction; magnitude) across studies.
When inconsistencies exist, the synthesis considers whether results were "conflicting"
(i.e., unexplained positive and negative results in similarly exposed human populations
or in similar animal models) or "differing" (i.e., mixed results explained by differences
between human populations, animal models, exposure conditions, or study methods)
(U.S. EPA, 2005a) based on analyses of potentially important explanatory factors such
as:
- Confidence in studies' results, including study sensitivity (e.g., some study
results that appear to be inconsistent may be explained by potential biases or
other attributes that affect sensitivity, resulting in variations in the degree of
confidence accorded to the study results)
- Exposure, including route (if applicable), levels, duration, etc.
- Populations or species, including consideration of potential susceptible groups
or differences across lifestage at exposure or endpoint assessment
- Toxicokinetic information as an explanation for any observed differences in
responses across route of exposure, other aspects of exposure, species, or
lifestages
The interpretation of the consistency of the evidence and the magnitude of the
reported effects will emphasize biological significance as more relevant to the
assessment than statistical significance. Statistical significance (as reported by
p-values, etc.) provides no evidence about effect size or biological significance, and a
lack of statistical significance will not be automatically interpreted as evidence of no
effect.
This document is a draft for review purposes only and does not constitute Agency policy.
12 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Consideration
Description and Synthesis Methods
Strength (effect
magnitude) and
precision
Biological
gradient/dose-response
Coherence
• Examines the effect magnitude or relative risk, based on what is known about the
assessed endpoint(s), and considers the precision of the reported results based on
analyses of variability (e.g., confidence intervals; standard error). In some cases,
this may include consideration of the rarity or severity of the findings (in the context
of the health effect being examined).
Syntheses will analyze results both within and across studies and may consider the
utility of combined analyses (e.g., meta-analysis). While larger effect magnitudes and
precision (e.g., p < 0.05) help reduce concerns about chance, bias or other factors as
explanatory, syntheses should also consider the biological or population-level
significance of small effect sizes. Thus, a lack of statistical significance should not be
automatically interpreted as evidence of no effect.
• Examines whether the results (e.g., response magnitude; incidence; severity)
change in a manner consistent with changes in exposure (e.g., level; duration),
including consideration of changes in response after cessation of exposure.
Syntheses will consider relationships both within and across studies, acknowledging
that the dose-response (e.g., shape) can vary depending on other aspects of the
experiment, including the outcome and the toxicokinetics of the chemical. Thus,
when dose-response is lacking or unclear, the synthesis will also consider the potential
influence of such factors on the response pattern.
• Examines the extent to which findings are cohesive across different endpoints that
are known/expected to be related to, or dependent on, one another (e.g., based on
known biology of the organ system or disease, or mechanistic understanding such
as toxicokinetic/dynamic understanding of the chemical or related chemicals). In
some instances, additional analyses of mechanistic evidence from research on the
chemical under review or related chemicals that evaluate linkages between
endpoints or organ-specific effects may be needed to interpret the evidence. These
analyses may require additional literature search strategies.
Syntheses will consider potentially related findings, both within and across studies,
particularly when relationships are observed within a cohort or within a narrowly
defined category (e.g., occupation; strain or sex; lifestage of exposure). Syntheses will
emphasize evidence indicative of a progression of effects, such as temporal- or dose-
dependent increases in the severity of the type of endpoint observed.
This document is a draft for review purposes only and does not constitute Agency policy.
13 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
Consideration
Description and Synthesis Methods
Mechanistic evidence
related to biological
plausibility
Natural experiments
• There are multiple uses for mechanistic information and this consideration
overlaps with 'coherence'. This examines the biological support (or lack thereof)
for findings from the human and animal health effect studies and becomes more
impactful on the hazard conclusions when notable uncertainties in the strength
of those sets of studies exist. These analyses can also improve understanding of
dose- or duration-related development of the health effect. In the absence of
human or animal evidence of apical health endpoints, the synthesis of
mechanistic information will drive evidence integration conclusions (when such
information is available).
Syntheses can evaluate evidence on precursors, biomarkers, or other molecular or
cellular changes related to the health effect(s) of interest to describe the likelihood
that the observed effects result from exposure. This will be an analysis of existing
evidence, and not simply whether a theoretical pathway can be postulated. This
analysis may not be limited to evidence relevant to the PECO, but may also include
evaluations of biological pathways (e.g., for the health effect; established for other,
possibly related, chemicals). The synthesis will consider the sensitivity of the
mechanistic changes and the potential contribution of alternative or previously
unidentified mechanisms of toxicity.
• Specific to epidemiological studies and rarely available, this examines effects in
populations that have experienced well-described, pronounced changes in exposure
to the chemical of interest (e.g., blood lead levels before and after banning lead in
gasoline)
1.1.4. Evidence Integration
For transparency in the sequential decision steps taken to draw overall evidence integration
judgments, a two-step, sequential process was used (Figure 2). First, judgments regarding the strength
of the evidence from the available human and animal studies were made in parallel. These judgments
incorporated mechanistic evidence (or MOA understanding) in exposed humans and animals,
respectively, that informed the biological plausibility and coherence of the available human or animal
health effect studies. Second, the animal and human evidence judgments were combined to draw an
overall conclusion(s) that incorporated inferences drawn based on information on the human relevance
of the animal evidence (i.e., based on default assumptions or empirical evidence), coherence across the
human and animal evidence streams, and susceptibility.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
Human and animal evidence judgments from Step 1 and the overall evidence integration
conclusion from Step 2 were reached using decision frameworks based on considerations originally
described by Austin Bradford Hill (Hill. 1965).
STEP 1: INTEGRATION OF HEALTH EFFECT
AND MECHANISTIC EVIDENCE IN HUMANS
OR ANIMALS
HUMAN EVIDENCE JUDGMENT
The synthesis of evidence about health
effects and mechanisms from human studies
is combined (integrated) to make a
judgment about health effects in human
studies
ANIMAL EVIDENCE JUDGMENT
The synthesis of evidence about health
effects and mechanisms from animal studies
is combined (integrated) to make a
judgment about health effects in animal
studies.
Figure 2. Process for evidence integration.
In the first step, the strength of the human and, separately, the animal evidence for each
noncancer health effect (or groups of related effects) and specific cancer type (or groups of related
cancer types) was summarized using the following terms: robust, moderate, slight, and indeterminate,
which are presented as italicized text in the various assessment documents. Note that the conclusion
regarding the strength of the animal evidence reflects an interpretation about whether formaldehyde
exposure causes the health effect(s) of interest in experimental animals, and the conclusion for the
human evidence reflects an interpretation about the evidence for a causal association between
formaldehyde exposure and the health effect that can be drawn from the available studies in humans or
human cells. The strength of the human and animal evidence was determined starting from the
evidence syntheses that summarized the evidence from the available human and animal health effects
studies, respectively, and then considering coherent effects from mechanistic evidence, which could add
to or detract from the strength of evidence. Note, however, that the lack of mechanistic data explaining
an association did not discount results from human or animal health effect studies. To draw these
judgments, a modified set of considerations was applied to evidence from studies in humans and
animals (Table 5).
STEP 2: OVERALL INTEGRATION OF EVIDENCE
FOR HAZARD ID
EVIDENCE INTEGRATION CONCLUSION
The judgments regarding the human and
animal evidence are integrated in light of
evidence on the human relevance of the
findings in animals, susceptibility, and the
coherence of the findings across evidence
streams to draw a conclusion about the
evidence for health effects in humans.
This document is a draft for review purposes only and does not constitute Agency policy.
15 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 5. Considerations that inform judgments regarding the strength of the human and animal evidence3
Consideration
Increased Evidence Strength
(of the Human or Animal Evidence)
Decreased Evidence Strength
(of the Human or Animal Evidence)
The structured categories and criteria in Tables 6 and 7 will guide the application of strength of evidence judgments for an outcome or health effect. Evidence synthesis
scenarios that do not warrant an increase or decrease in evidence strength will be considered "neutral." These ideas build upon the discussion for assessing causality of
disease in Hill (1965), although there are some differences in the use or interpretations of the terms (see Toxicological Review).
Risk of bias;
sensitivity (across
studies)
• An evidence base of high or medium confidence studies increases
strength.
• An evidence base of mostly low confidence studies decreases strength. An
exception to this is an evidence base of studies where the primary issues
resulting in low confidence are related to insensitivity. This may increase
evidence strength in cases where an association is identified because the
expected impact of study insensitivity is toward the null.
• Decisions to increase strength for other considerations in this table should
generally not be made if there are serious concerns for risk of bias.
Consistency
• Similarity of findings for a given outcome (e.g., of a similar
magnitude, direction) across independent studies or experiments
increases strength, particularly when consistency is observed across
populations (e.g., location) or exposure scenarios in human studies,
and across laboratories, populations (e.g., species), or exposure
scenarios (e.g., duration; route; timing) in animal studies.
• Unexplained inconsistency (conflicting evidence) decreases strength.
Generally, strength should not be decreased if discrepant findings can be
reasonably explained by study confidence conclusions, variation in
population or species, sex, or lifestage, exposure patterns (e.g., intermittent
or continuous), levels (low or high), duration or intensity. However, any
decisions about decreased strength will be determined by the extent to
which residual questions about the evidence may persist.
Strength (effect
magnitude) and
precision
• Evidence of a large magnitude effect (considered either within or
across studies), can increase strength. Effects of a concerning rarity
or severity can also increase strength, even if they are of a small
magnitude.
• Precise results from individual studies or across the set of studies
increases strength, noting that biological significance is prioritized
over statistical significance.
• The presence of small effects is not typically used to decrease confidence in
a body of studies. However, if effect sizes that are small in magnitude are
concluded not to be biologically significant, or if there are only a few
studies with imprecise results, then strength is decreased.
• In animal studies, an example of evidence that can decrease strength
involves an effect for which there is a lesser level of concern under some
conditions (e.g., rapid reversibility after removal of exposure). Note that
many reversible effects are of high concern. Such a decision is informed by
factors such as the toxicokinetics of the chemical and the conditions of
exposure fsee U.S. EPA (1998)1, judgments regarding the potential for
This document is a draft for review purposes only and does not constitute Agency policy.
16 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Consideration
Increased Evidence Strength
(of the Human or Animal Evidence)
Decreased Evidence Strength
(of the Human or Animal Evidence)
delayed or secondary effects, as well as the exposure context focus of the
assessment (e.g., addressing intermittent or short-term exposures).
Biological
gradient/dose-
response
• Evidence of dose-response increases strength. Dose-response may
be demonstrated across studies or within studies and it can be dose-
or duration-dependent. It may also not be a monotonic dose-
response (monotonicity should not necessarily be expected), and the
analysis will consider the extent to which this might be explained by
the available evidence (e.g., different outcomes may be expected at
low versus high doses due to activation of different mechanistic
pathways or induction of systemic toxicity at very high doses).
• Decreases in a response after cessation of exposure (e.g., symptoms
of current asthma) also may increase strength by increasing certainty
in a relationship between exposure and outcome (this is applicable
to human observational studies, but not experimental studies).
• A lack of dose-response when expected based on biological understanding
and having a wide-range of doses/exposures evaluated in the evidence
base can decrease strength.
• In rare cases, and typically only in toxicology studies, the duration of
exposure might reveal an inverse association with effect magnitude (e.g.,
due to tolerance or acclimation). Similar to the discussion of reversibility
above, a decision about whether this decreases strength depends on the
exposure context focus of the assessment and other factors.
• If the data are not adequate to evaluate a dose-response pattern, then
strength is neither increased nor decreased.
Coherence
• Biologically related findings within an organ system, or across
populations (e.g., sex) increase strength, particularly when a
temporal- or dose-dependent progression of related effects is
observed within or across studies, or when related findings of
increasing severity are observed with increasing exposure.
• An observed lack of expected coherent changes (e.g., well-established
biological relationships), particularly when observed for multiple related
endpoints, will typically decrease evidence strength. The decision to
decrease depends on the strength of the expected relationship(s), and
considers factors (e.g., dose and duration of exposure) across studies of
related changes.
Mechanistic
evidence related
to biological
plausibility
• Mechanistic evidence of precursors or health effect biomarkers in
well-conducted studies of exposed humans or animals, in
appropriately exposed human or animal cells, or other relevant
human or animal models (for the human or animal evidence stream,
respectively) increases strength, particularly when this evidence is
observed in the same cohort/population exhibiting the health
outcome.
• Evidence of changes in biological pathways, or providing support for
a proposed mode-of-action (MOA) in models also increases
strength, particularly when support is provided for rate-limiting or
• Mechanistic understanding is not a prerequisite for judging the evidence,
and thus absence of knowledge should not be used a basis for decreasing
strength NTP (2015) NRC (2014a). The human relevance of animal findings
is assumed unless there is sufficient evidence to the contrary [see U.S. EPA
(2005a) IARC (2006)1.
• Mechanistic evidence in well-conducted studies that demonstrates that the
health effect(s) are unlikely to occur, or only likely to occur under certain
scenarios (e.g., above certain exposure levels), can decrease evidence
strength. A decision to decrease depends on an evaluation of the strength
of the mechanistic evidence supporting vs. opposing biological plausibility,
This document is a draft for review purposes only and does not constitute Agency policy.
17 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Consideration
Increased Evidence Strength
(of the Human or Animal Evidence)
Decreased Evidence Strength
(of the Human or Animal Evidence)
key events, or conserved across multiple components of the
pathway or MOA.
as well as the strength of the health effect-specific findings (e.g., stronger
health effect data require more certainty in mechanistic evidence opposing
plausibility).
1
This document is a draft for review purposes only and does not constitute Agency policy.
18 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Decision frameworks, with criteria described in Tables 6 and 7 were used to develop the
2 judgments concerning the strength of evidence for a health effect within each of the human and animal
3 evidence bases, weighing the strengths and weaknesses of both positive and null studies. These
4 frameworks, which add clarity, consistency, and transparency to the evidence evaluations and
5 conclusions, are consistent with generally accepted principles in epidemiology and toxicology and are
6 meant to convey a distribution of confidence in each body of evidence pertaining to a hazard, a process
7 that relies on expert judgment.
8 Table 6. Framework for strength of evidence judgments (human evidence)
Strength of
Evidence Judgment
Description
Robust
... evidence in human
studies
(strong signal of
effect with little
residual uncertainty)
A set of high or medium confidence independent studies reporting an association
between the exposure and the health outcome, with reasonable confidence that
alternative explanations, including chance, bias, and confounding, can be ruled out across
studies. The set of studies is primarily consistent, with reasonable explanations when
results differ; an exposure-response gradient is demonstrated; and the set of studies
includes varied populations. Additional supporting evidence, such as associations with
biologically related endpoints in human studies (coherence) or large estimates of risk or
severity of the response, may increase confidence but are not required.
In exceptional circumstances, a finding in one study may be considered to be robust, even
when other studies are not available (e.g., analogous to the finding of angiosarcoma, an
exceedingly rare liver cancer, in the vinyl chloride industry).
Mechanistic evidence from exposed humans or human cells, if available, may add
support, informing considerations such as exposure-response, temporality, coherence,
and MOA, thus raising the level of certainty to robust for a set of studies that otherwise
would be described as moderate.
Moderate
... evidence in human
studies
(signal of effect with
some uncertainty)
A smaller number of studies (at least one high or medium confidence study with
supporting evidence), or with some heterogeneous results, that do not reach the degree
of confidence required for robust. For multiple studies, there is primarily consistent
evidence of an association, but there may be lingering uncertainty due to potential
chance, bias or confounding.
For a single study, there is a large magnitude or severity of the effect, or a dose-response
gradient, or other supporting evidence, and there are not serious residual methodological
uncertainties. Supporting evidence could include associations with related endpoints,
including mechanistic evidence from exposed humans or human cells, if available, based
on considerations such as exposure-response, temporality, coherence, and MOA, thus
raising the level of certainty to moderate for a set of studies that otherwise would be
described as slight.
Slight
One or more studies reporting an association between exposure and the health outcome,
where considerable uncertainty exists. In general, only low confidence studies may be
available, or considerable heterogeneity across studies may exist. Supporting coherent
This document is a draft for review purposes only and does not constitute Agency policy.
19 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Strength of
Evidence Judgment
Description
... evidence in human
studies
(signal of effect with
large amount of
uncertainty)
evidence is sparse. Strong biological support from mechanistic evidence in exposed
humans or human cells may also be independently interpreted as slight. This also
includes scenarios where there are serious residual uncertainties across studies (these
uncertainties typically relate to exposure characterization or outcome ascertainment,
including temporality) in a set of largely consistent medium or high confidence studies.
This category serves primarily to encourage additional study where evidence does exist
that might provide some support for an association, but for which the evidence does not
reach the degree of confidence required for moderate.
Indeterminate
... evidence in human
studies
(signal cannot be
determined for or
against an effect)
No studies were available in humans or situations when the evidence is inconsistent or
primarily of low confidence.
Compelling evidence
of no effect
... in human studies
(strong signal for
lack of an effect with
little uncertainty)
Several high confidence studies showing null results (for example, an odds ratio of 1.0),
ruling out alternative explanations including chance, bias, and confounding with
reasonable confidence. Each of the studies should have used an optimal outcome and
exposure assessment and adequate sample size (specifically for higher exposure groups
and for susceptible populations). The set as a whole should include the full range of levels
of exposures that human beings are known to encounter, an evaluation of an exposure-
response gradient, and an examination of at-risk populations and lifestages.
1 Table 7. Framework for strength of evidence judgments (animal evidence)
Strength of Evidence
Judgment
Description
Robust
... evidence in animals
(strong signal of effect
with little residual
uncertainty)
The set of high or medium confidence experiments includes consistent findings of
adverse or toxicologically significant effects across multiple laboratories, exposure
routes, experimental designs (e.g., a subchronic study and a two-generation study), or
species, and the experiments can reasonably rule out the potential for nonspecific
effects (e.g., resulting from toxicity) to have resulted in the findings. Any inconsistent
evidence (evidence that cannot be reasonably explained by the respective study design
or differences in animal model) is from a set of experiments of lower confidence. At
least two of the following additional factors in the set of experiments support a causal
association: coherent effects across multiple related endpoints (may include
mechanistic endpoints); an unusual magnitude of effect, rarity, age at onset, or
severity; a strong dose-response relationship; or consistent observations across animal
lifestages, sexes, or strains. Alternatively, mechanistic data in animals or animal cells
that address the above considerations or that provide experimental support for an
MOA that defines a causal relationship with reasonable confidence may raise the level
This document is a draft for review purposes only and does not constitute Agency policy.
20 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Strength of Evidence
Judgment
Description
of certainty to robust for evidence that otherwise would be described as moderate or,
exceptionally, slight or indeterminate.
Moderate
... evidence in animals
(signal of effect with
some uncertainty)
A set of evidence that does not reach the degree of certainty required for robust, but
which includes at least one high or medium confidence study and information
strengthening the likelihood of a causal association. Although the results are largely
consistent, notable uncertainties remain. However, while inconsistent evidence or
evidence indicating nonspecific effects (e.g., toxicity) may exist, it is not sufficient to
reduce or discount the level of concern regarding the positive findings from the
supportive experiments or it is from a set of experiments of lower confidence. The set
of experiments supporting the effect provide additional information supporting a
causal association, such as consistent effects across laboratories or species; coherent
effects across multiple related endpoints (may include mechanistic endpoints); an
unusual magnitude of effect, rarity, age at onset, or severity; a strong dose-response
relationship; or consistent observations across exposure scenarios (e.g., route, timing,
duration), sexes, or animal strains. Mechanistic data in animals or animal cells that
address the above considerations or that provide information supporting an
association between exposure and effect with reasonable confidence may raise the
level of certainty to moderate for evidence that otherwise would be described as
slight.
Slight
... evidence in animals
(signal of effect with
large amount of
uncertainty)
Scenarios in which there is a signal of a possible effect, but the evidence is conflicting
or weak. Most commonly, this includes situations where only low confidence
experiments are available and supporting coherent evidence is sparse. It also applies
when one medium or high confidence experiment is available without additional
information strengthening the likelihood of a causal association (e.g., corroboration
within the same study or from other studies). Lastly, this includes scenarios in which
there is evidence that would typically be characterized as moderate, but inconsistent
evidence (evidence that cannot be reasonably explained by the respective study design
or differences in animal model) from a set of experiments of higher confidence (may
include mechanistic evidence) exists. Strong biological support from mechanistic
studies in exposed animals or animal cells may also be independently interpreted as
slight. Notably, to encourage additional research, it is important to describe situations
for which evidence does exist that might provide some support for an association but
is insufficient for a conclusion of moderate.
Indeterminate
...evidence of the effect
under review in animals
No animal studies were available, or a set of low confidence animal studies exist that
are not reasonably consistent or are not informative to the hazard question under
evaluation.
(signal cannot be
determined for or
against an effect)
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
Strength of Evidence
Judgment
Description
Compelling evidence of
no effect
... in animals
(strong signal for lack of
an effect with little
uncertainty)
A set of high confidence experiments examining a reasonable spectrum of endpoints
relevant to a type of toxicity that demonstrate a lack of biologically significant effects
across multiple species, both sexes, and a broad range of exposure levels. The data are
compelling in that the experiments have examined the range of scenarios across which
health effects in animals could be observed, and an alternative explanation (e.g.,
inadequately controlled features of the studies' experimental designs; inadequate
sample sizes) for the observed lack of effects is not available. The experiments were
designed to specifically test for effects of interest, including suitable exposure timing
and duration, postexposure latency, and endpoint evaluation procedures, and to
address potentially susceptible populations and lifestages.
The second stage of evidence integration combined the animal and human evidence judgments,
while also considering mechanistic information on the human relevance of the animal evidence,
relevance of the mechanistic evidence to humans (especially in cases where animal evidence was
lacking), coherence across lines of evidence, and information on susceptible populations, to arrive at an
overall evidence integration judgment regarding the evidence for causation (Table 8). This evidence
integration framework interprets the guidance and examples provided in the cancer guidelines (U.S.
EPA. 2005a) to allow clarity and consistency in the evaluation of each hazard. The evidence integration
framework illustrates the principle that evidence in humans generally has greater weight for causal
conclusions than evidence in animals. In the absence of sufficiently justifiable MOA information, effects
in animal models are assumed to be relevant to humans. In this assessment, for potential health
hazards where the evidence from animal models strongly influenced the overall evidence integration
judgment, the available mechanistic evidence was considered in light of human relevance.
For each potential health effect evaluated, a narrative summary and evidence integration
judgment regarding the available evidence were developed. The overall evidence integration
judgments: evidence demonstrates, evidence indicates [likely], evidence suggests, and evidence
inadequate (to judge hazard) are presented as bolded text throughout the assessment and are
accompanied by a description of the conditions of expression (e.g., exposure levels, exposure patterns)
in the studies that served as the basis for the judgment. Importantly, for the purposes of this
assessment, the same evidence integration approach was used to draw evidence integration judgments
for both noncancer health effects and specific cancer types. For carcinogenicity, a final step of
categorizing the totality of the evidence using a "descriptor" was performed (U.S. EPA. 2005a), as
described below.
This document is a draft for review purposes only and does not constitute Agency policy.
22 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 8. Overall evidence integration judgments for characterizing the integrated
2 evidence for noncancer health effects and cancer outcomes
Overall
evidence
integration
judgment in
narrative
Explanation and Example Scenarios
Evidence
demonstrates
This signifies a very high level of certainty that formaldehyde exposure caused the health effect.
For this assessment, if the data were amenable, a toxicity value was estimated.
• This category was3 used if there was robust human evidence supporting an effect.
• This category could also be used with moderate human evidence and robust animal evidence
if there was strong mechanistic evidence that MOAs and key precursors identified in animals
were anticipated to occur and progress in humans.
Evidence
indicates
[likely]b
This reflects a reasonable certainty that the relationship between formaldehyde exposure and
the health outcome was causal, although there may be some outstanding questions that remain.
For this assessment, if the data were amenable, a toxicity value was estimated.
• This category was used if there is robust animal evidence supporting an effect and moderate-
to-indeterminate human evidence when strong mechanistic evidence was lacking.
• This category was also used with moderate human or animal evidence supporting an effect
and slight or indeterminate evidence from the opposite evidence stream. In these scenarios,
any uncertainties in the moderate evidence were not sufficient to reduce or discount the
level of concern, or mechanistic evidence in the opposite evidence stream (e.g., precursors)
existed to increase confidence in the moderate evidence.
Evidence
suggests (but is
not sufficient to
infer)c
This conveys some concern that formaldehyde may cause a particular health outcome, but there
were very few studies that contributed to the evaluation, the evidence was very weak or
conflicting, or the methodological conduct of the studies was poor. Given the substantial degree
of uncertainty, additional research would provide valuable information for future evaluations.
Although it may sometimes be possible to develop toxicity values for evidence in this category,
given the particulars of the available data in this assessment, toxicity values were not estimated.
• This category was used if there was slight human or animal evidence.
• This category could also be used with moderate human or animal evidence and sliaht or
indeterminate evidence in the opposite evidence stream. In these scenarios, there were
outstanding issues regarding the moderate evidence that reduced the level of concern or
confidence in the reliability of the findings, or mechanistic evidence from the opposite
evidence stream (e.g., null results in well-conducted evaluations of precursors) existed to
decrease confidence in the moderate evidence.
• Exceptionally, when there is general scientific understanding of mechanistic events that result
in a hazard, this category could also be used if there was strong mechanistic evidence that is
sufficient to identify a cause for concern—in the absence of adequate conventional studies in
humans or in animals (i.e., indeterminate evidence in both).
This document is a draft for review purposes only and does not constitute Agency policy.
23 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Overall
evidence
integration
judgment in
narrative
Explanation and Example Scenarios
Evidence
lnadequated
This conveys either a lack of information or an inability to interpret the available evidence. A
toxicity value was not estimated.
• This category was used if there was indeterminate human and animal evidence.
• This category could also be used with slipht-to-robust animal evidence and indeterminate
human evidence if strong mechanistic information indicated that the animal evidence was
unlikely to be relevant to humans.
A conclusion of inadequate is not a determination that the agent does not cause adverse health
outcomes or is safe. It generally indicates that further research is needed.
Note: This table does not supersede or alter any EPA guidance. It is meant only to provide added transparency for
conclusions drawn regarding the level of evidence from human, animal, and mechanistic studies.
aTerminology of "was" refers to the default option; terminology of "could also be" refers to alternative options.
bFor some applications, such as benefit-cost analysis, to better differentiate the categories of evidence
demonstrates and evidence indicates (likely), the latter category should be interpreted as evidence that supports
an exposure-effect linkage that is likely to be causal.
cHealth effects characterized as having evidence demonstrates and evidence indicates (likely) (and, in some cases,
evidence suggests) are evaluated for use in dose-response assessment. When the database includes at least one
well-conducted study and a judgment of evidence suggests is drawn, quantitative analyses may still be useful for
some purposes (e.g., providing a sense of the magnitude and uncertainty of estimates for health effects of
potential concern, ranking potential hazards, or setting research priorities), but not for others [see related
discussions in (U.S. EPA, 2005a)l. It is critical to transparently convey the extreme uncertainty in any such
estimates.
dSpecific narratives for each of the health effects with an evidence integration judgment of evidence inadequate
may be deemed unnecessary.
1 For carcinogenesis only, the weight of evidence as to whether formaldehyde inhalation
2 exposure is carcinogenic to humans is summarized using descriptors, consistent with EPA guidelines
3 (U.S. EPA. 2005a) (Table 9). For this assessment, the descriptors build upon the overall evidence
4 integration judgments for individual cancer types, as described in Table 8; however, this does not alter
5 or supersede any EPA guidance.
6 Table 9. Criteria for applying cancer descriptors to evidence integration judgments for
7 cancer types
Cancer Descriptor
Criteria
Carcinogenic to humans
• This descriptor was used if the evidence demontrates that, for at least one cancer
type, formaldehyde inhalation exposure caused the increase in cancer incidence or
mortality.
• This descriptor could also be used in rare instances if for the evidence indicates
that formaldehyde inhalation exposure likely causes different cancer types across
evidence bases (e.g., when one type of cancer is based on human evidence and
tumors at another site is supported by animal evidence, consistent with EPA
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
Cancer Descriptor
Criteria
guidance (U.S. EPA, 2005a) that site-concordance is not reauired). Such a decision
would depend on mechanistic understanding (i.e., in this example, the decision
would consider differences in tumor types or ADME across species).
Likely to be carcinogenic to
humans
• This descriptor was used if the evidence indicates that, for at least one cancer
type, formaldehyde inhalation exposure likely caused the increase in cancer
incidence or mortality.
• Similar to the rationale provided above, this descriptor could also be used in rare
instances when the evidence suggests formaldehyde inhalation exposure may
cause multiple tumor types, depending on mechanistic inference.
Suggestive evidence of
carcinogenic potential
This descriptor was used if, for the evidence relating to carcinogenicity, the evidence
was only suggestive that formaldehyde inhalation exposure may cause any of the
observed increases in cancer incidence or mortality for any cancer type. This would
reflect a substantial degree of uncertainty in any potential causal association.
Inadequate evidence to
assess carcinogenic potential
This descriptor was used if the evidence was inadequate to draw a conclusion
regarding cancers of any type with any confidence. This might reflect a lack of
information or highly conflicting information.
Not Likely to be carcinogenic
to humans
This descriptor conveys a high degree of certainty that there is negligible concern for
carcinogenic effects. A substantial amount of evidence would be required to support
this descriptor (see (U.S. EPA, 2005a)).
1.1.5. Dose-response Analysis
The Toxicological Review includes an inhalation reference concentration (RfC). The inhalation
RfC is defined as an estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous
daily exposure of formaldehyde to the human population (including sensitive subgroups) that is likely to
be without an appreciable risk of deleterious effects during a lifetime. A carcinogenicity assessment was
also performed, including derivation of an inhalation unit risk value (IUR), which is an upper-bound
excess cancer risk estimated to result from continuous exposure to an agent at a concentration of
1 ng/m3 in air for a lifetime. In addition, organ/system-specific RfCs (osRfCs) were derived for the
various noncancer health endpoints, when supported by the available evidence, which may be useful
when considering cumulative risk scenarios. Multiple candidate RfCs (cRfCs) were sometimes compared
before choosing a representative osRfC. Where relevant, mechanistic understanding regarding the
development of specific health effects (e.g., temporal progression; potential thresholds in dose-
response), as well as knowledge of susceptibility, was used to inform approaches to derive points of
departure (PODs), uncertainty factors, or confidence levels for the quantitative estimates (e.g., osRfCs;
RfC; IUR). A confidence level of high, medium, or low was assigned to each osRfC and the RfC based on
the reliability of the associated POD and cRfC calculation(s). Confidence in the completeness of the
database for each osRfC and the overall RfC was also assigned. These decisions were used to draw an
overall level of confidence in the RfC. Likewise, an overall level of confidence was assigned to the IUR.
This document is a draft for review purposes only and does not constitute Agency policy.
25 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Toxicological Review of Formaldehyde - Inhalation
Where possible, the assessment attempts to describe the level of response observed across different
exposure levels within the range of the data and transparently discusses the uncertainties and
assumptions when deriving and applying the different toxicity value estimates (e.g., cRfCs, IUR).
2. SUMMARY OF TOXICOKINETICS
Several detoxification and removal processes exist for formaldehyde, including at the site(s) of
first contact (e.g., human nasal passages for inhalation). Much of what is known regarding the uptake
and distribution of formaldehyde is based on experiments using monkeys and rats. Species differences
in the structure of the airways and breathing patterns, as well as the composition of the surface
epithelium at various nasal locations, are important considerations when interpreting results in
experimental animals and extrapolating observations to humans. While the nasal passages in humans
are generally similar to other mammalian species, one key difference is that humans and non-human
primates have nasal passages adapted for both oral and nasal (oronasal) breathing, as opposed to
obligate nasal breathing in rodents. A second key difference regards the shape and complexity of the
nasal turbinates, with relatively simple shapes in humans, and complex, folded patterns in rodents. In
general, these differences provide better protection of the rodent lower respiratory tract (LRT) against
inhaled toxicants than is provided to the human LRT (Harkema et al.. 2006).
Uptake of formaldehyde is based on rough estimates determined from the amount of
formaldehyde removed from the air and indicates that the vast majority of formaldehyde is removed
from inhaled air by the upper respiratory tract (URT) in monkeys (Casanova et al.. 1991; Monticello et
al.. 1989), dogs (Egle, 1972), and rats (Kimbell et al.. 2001b; Chang et al.. 1983; Heck et al.. 1983; Kerns
et al.. 1983). Further, dosimetric modeling studies in humans have shown close agreement with
observations of exposed rodents. Overall, a concentration gradient of inhaled formaldehyde follows an
anterior to posterior distribution, with high concentrations of formaldehyde distributed to nasal
squamous, transitional and respiratory epithelium, and less uptake by olfactory epithelium, and very
little formaldehyde reaching more distal sites such as the larynx or lung. The possibility that more
extensive distribution to the LRT may occur when people are breathing through the mouth during
exercise or when they have an upper respiratory tract infection has not been investigated.
As inhaled formaldehyde enters the URT, it interacts with the mucociliary apparatus, the first
line of defense against inhaled materials in the nose. In nasal mucus, most of the formaldehyde is
rapidly converted to methanediol (~99.9%) and a minor fraction remains as free formaldehyde (~0.1%)
(Bogdanffy et al.. 1986; Fox et al.. 1985). Formaldehyde levels are reduced through interactions with
components of the mucus and through mucociliary clearance; through reactions with cellular materials
at the plasma membrane of the respiratory epithelium; via interactions with glutathione (GSH) and
This document is a draft for review purposes only and does not constitute Agency policy.
26 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Toxicological Review of Formaldehyde - Inhalation
other macromolecules in the intracellular and extracellular space; through localized metabolism and
conjugation reactions; and through reversible interactions with intracellular materials. This results in
the formation of a gradient of formaldehyde across the tissue space, with the greatest formaldehyde
concentration at the apical surface of the mucosa, and the lowest levels of formaldehyde at deeper
components of the tissue, such as the nasal associated lymphoid tissues (NALT) and blood vessels.
Models developed by Schroeter et al. (2014) and Campbell et al. (2020) highlight the fact that at
sufficiently low levels of exogenous formaldehyde, the flux of exogenous formaldehyde is reduced due
to the presence of endogenous formaldehyde, that is, a lesser concentration gradient across the tissues
results in a reduced uptake of inhaled formaldehyde. These results add to the characterization of the
uncertainty in formaldehyde dose-response at low exposure concentrations (discussed in further detail
in Section 1.1.3 and 2.2.1 of the main document).
Several of the key considerations for evaluating the distribution of inhaled formaldehyde to the
portal-of-entry (the rat nose is depicted) and systemic sites are represented in Figure 3.
inspired air and
formaldehyde (red)
_ Mucus
Cilia
Epithelium
[epithelial cells
(EC) and goblet
cells (GC)]
l Basement membrane
Lamina propria
[systemic circulation
(SC) and lymphoid
tissue (NALT)]
VJALT*
Gradient
formaldehyde
concentration
Rat nose [formaldehyde
(in red), squamous
epithelium (se),
respiratory epithelium
(re), nasal turbinates with
olfactory epithelium (NT-
oe) or with re (NT-re),
cribiform plate (CP), and
olfactory bulb (OB)]
Nasopharynx (to lower
respiratory tract)
Figure 3. Schematic of the rat upper respiratory tract depicting the gradient of
formaldehyde concentration formed following inhalation exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
The gradient both from anterior to posterior locations, as well as across the tissue depth is depicted.
Modeling based on observations in rodents predicts a similar pattern of distribution in humans. Drawn
based in part on images by NRC (2011) and Harkema et al. (2006).
In the respiratory tissue, formaldehyde can be metabolized to formate, which can either enter
the one-carbon pool leading to protein and nucleic acid synthesis, or be further metabolized to C02 and
eliminated in expired air or excreted in urine unchanged. Alternatively, upon interactions with cellular
macromolecules, it can form DNA-protein cross-links (DPX), protein adducts (Edrissi et al.. 2013b; Edrissi
et al.. 2013a), or other products, as demonstrated by concentration-dependent increases in DPX
formation in rat and monkey nasal passages. Recently, analytical methods have been developed that
can distinguish between N2-hm-dG adducts from exogenous (inhaled) formaldehyde and N2-hm-dG
adducts from endogenous formaldehyde (Lu et al.. 2012; Lu et al.. 2011; Moeller et al.. 2011; Lu et al..
2010). DNA monoadducts (Yu et al.. 2015; Lu et al.. 2011; Moeller et al.. 2011; Lu et al.. 2010) and DPX
(Lai et al.. 2016) derived from exogenous formaldehyde were detectable in nasal tissues, but not in
distal tissues (including the bone marrow) of experimental animals exposed by inhalation, suggesting
that exogenous formaldehyde is not systemically distributed. Also, toxicokinetic studies demonstrate
that labeled carbon from inhaled formaldehyde measured in bone marrow of rats was the result of
metabolic incorporation from the 1-Carbon (1C) pool, not covalent binding, further supporting the lack
of transport of formaldehyde or metabolites of formaldehyde to the distal tissues (Casanova-Schmitz et
al.. 1984). Finally, inhalation exposure to formaldehyde does not appear to alter blood formaldehyde
levels (approximately 0.1 mM across different species), suggesting that inhaled formaldehyde is not
significantly absorbed into blood (Kleinniienhuis et al.. 2013; Casanova et al.. 1988; Heck et al.. 1985).
The toxicokinetics of formaldehyde may be influenced by certain formaldehyde-related effects,
such as altered mucociliary clearance (e.g., (Morgan. 1983)), reflex bradypnea in rodents (Chang J et al..
1983; Chang et al.. 1981), and dynamic changes in tissue structure (Kamata et al.. 1997), which have the
potential to modulate formaldehyde uptake and clearance. For example, during repeated inhalation
exposure to formaldehyde, mice, and to a lesser extent rats, lower their minute volume thereby
restricting the intake of the gas (Chang J et al.. 1983; Chang et al.. 1981), which may impact dosimetric
adjustment if the dose-response results from rodent studies are extrapolated to humans.
3. NONCANCER HEALTH EFFECTS
Based on the current understanding of the toxicokinetics of formaldehyde inhalation exposure,
several practical working assumptions were applied to this assessment. Although some uncertainties
remain, the organization and analyses in the assessment assume that inhaled formaldehyde is not
This document is a draft for review purposes only and does not constitute Agency policy.
28 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
distributed to an appreciable extent beyond the upper respiratory tract to systemic sites; thus, it is
assumed that inhaled formaldehyde is not directly interacting with tissues distal to the portal-of-entry to
elicit effects. Similarly, it is assumed that formaldehyde does not cause appreciable changes in normal
metabolic processes associated with formaldehyde in distal tissues. Thus, for the purposes of this
assessment, studies examining potential associations between levels of formaldehyde or formaldehyde
byproducts measured in distal tissues and health outcomes are not considered relevant to inhaled
formaldehyde.
Research on several noncancer respiratory health effects was evaluated: sensory irritation;
pulmonary function; immune-mediated conditions, focusing on allergic conditions and asthma; and
respiratory tract pathology. An overarching evaluation of the mechanistic information pertinent to any
or all potential noncancer respiratory system health effects was performed (see Appendix A.5.6), with
the most pertinent results summarized within each section. Evaluations were also performed for
noncancer systemic (i.e., non-respiratory) health effects: nervous system effects; reproductive or
developmental toxicity.
3.1. SENSORY IRRITATION
Individuals exposed to formaldehyde in indoor air reported symptoms of irritation in the eyes,
nose, and throat; eye irritation is the most sensitive effect. Controlled human exposure studies
evaluated frequency and severity of symptoms during brief periods of exposure, and a few studies also
evaluated objective measures, such as conjunctival redness or frequency of eye blinking.
Epidemiological studies of exposure to indoor formaldehyde among residential populations evaluated
symptoms of irritation, including burning and watering eyes, sneezing and rhinitis, sore throat, and
coughing. This review of sensory irritation focused on symptoms and other measures of eye irritation,
which is an immediate response to formaldehyde exposure (Andersen and Molhave. 1983: Andersen.
1979).
3.1.1. Literature Identification
While the review focused on the more informative controlled human exposure studies and
observational studies in residential populations, occupational studies and studies of students exposed to
embalming fluid during dissection labs were also reviewed. The bibliographic databases, search terms,
and specific strategies used to search them are provided in Appendix A.5.2, as are the specific PECO
criteria and the methods for identifying literature from 2016 - 2021 are described in Appendix F.
Mechanistic studies relevant to sensory irritation were separately identified (and evaluated) as part of
the overarching review of mechanistic data informing respiratory effects (see Appendix A.5.6 for
additional details and supporting analyses).
This document is a draft for review purposes only and does not constitute Agency policy.
29 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
3.1.2. Study Evaluation
The controlled human exposure studies were able to evaluate symptoms in a controlled
environment; therefore, the dose-response relationship was more precise and potential confounders
were less of a concern. However, the study groups were selected for age (younger adults) and were
healthy enough to conform to study protocols. These studies evaluated formaldehyde concentrations
above 0.1 mg/m3, while exposure levels in the residential studies ranged between 0.01 (LOD) to
approximately 1 mg/m3, with a large proportion of residences less than 0.1 mg/m3. The studies of
residential formaldehyde exposure included a wider range of ages (adults and children) and potentially
susceptible individuals; some had existing respiratory and other health conditions.
3.1.3. Synthesis of Human Health Effect Studies
The controlled human exposure studies showed that the irritant response to formaldehyde is an
immediate phenomenon apparent at concentrations of 0.1 mg/m3, the lowest concentration evaluated,
and higher, and resolves when exposure is removed (Sauder et al.. 1986; Andersen and Molhave, 1983).
Both prevalence and severity of symptoms were associated with concentration. In addition, a large
variability in sensitivity to the irritant properties of formaldehyde at specific concentrations was
observed (Mueller et al.. 2013; Berglund et al.. 2012). Because of the wide variability in responses, it has
been difficult for experimental studies to characterize the exposure-response relationship in the lower
range of concentrations experienced by the general population.
Only a few studies evaluated whether symptom prevalence or severity changed over the course
of the exposure period. Controlled exposure studies by one research group indicate that irritant
responses to 2.46 mg/m3 do not differ across groups as a result of previous, routine formaldehyde
exposure (Schachter et al.. 1987; Schachter et al.. 1986). Studies that examined change in response
during exposures at relatively high levels (>1 mg/m3) reported higher symptom scores initially with
subsequent declines suggestive of acclimation during exposure (Green et al.. 1987; Schachter et al..
1986; Andersen and Molhave. 1983). However, at lower concentrations (0.3 and 0.5 mg/m3), the
initiation of symptoms was delayed and symptom severity continued to increase during the exposure
period (Andersen and Molhave. 1983). Overall, these few studies suggest that some acclimatization
may occur over a few hours at higher concentrations, however this phenomenon may not be apparent
when concentrations are lower (<1 mg/m3).
Two studies investigating the prevalence of symptoms of irritation in relation to residential
formaldehyde exposure observed a statistically significant relationship between increasing
formaldehyde concentration (from approximately 0.01 to >0.60 mg/m3) and symptoms of irritation
using logistic regression models with adjustment for age, gender, smoking behavior and other potential
confounders (Liu et al.. 1991; Hanrahan et al.. 1984). Data were collected on symptoms occurring since
participants had moved into their homes (Hanrahan et al.. 1984) or those that occurred during the two
weeks prior to the end of the one-week formaldehyde sampling period (Liu et al.. 1991). Although the
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
sampling period used by Hanrahan et al. (1984) was short (one hour), the study ruled out several
exposure sources that might contribute to variability in concentrations. Other emissions released from
the same sources as formaldehyde that also may contribute to eye irritation, including phenols, pinene,
and terpenes, were not adjusted for in the analysis, but were present at lower levels compared to
formaldehyde. A strong dose-response relationship with formaldehyde, as a cumulative measure (ppm-
hour) or a one-hour concentration, was reported by these two medium confidence studies, indicating
that the associations are unlikely to be entirely explained by unmeasured confounding from
coexposures. Although the studies were limited by low participation rates, a potential source of
selection bias if related to formaldehyde concentrations, participants were randomly selected for
recruitment and the investigators noted that the characteristics of respondents and non-respondents,
such as age of housing stock, demographics, and formaldehyde concentrations, were comparable.
Figure 4 graphs prevalence of eye irritation (or burning eyes) by formaldehyde concentration
reported by controlled human exposure studies and residential studies that evaluated concentrations
below 1 mg/m3. These results are complementary for the most part and indicate a consistent pattern in
response to formaldehyde concentrations between 0 and 1 mg/m3. The study by Bender et al. (1983)
used a protocol that involved exposure to the eyes only, although the concentration-response pattern
was similar to the studies that evaluated exposure via inhalation. Two controlled human exposure
studies that also evaluated concentrations below 1 mg/m3 reported results using a different metric, a
subjective symptom score rather than symptom prevalence (Mueller et al.. 2013; Lang et al.. 2008). The
results of the two studies differed. Lang et al. (2008) reported an increase in symptom scores for eye
irritation at 0.3 mg/m3, whereas Mueller et al. (2013) reported no effect related to formaldehyde
exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
31 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Toxicological Review of Formaldehyde - Inhalation
I
0.8 "
cu
0.6 -
i
cu
>
cu
Anderson 1983
0.4 -
K u lie 1987
Bender 1983
Main & Hogan 1983
~ Olsen & Dossing 1982
O Hanrahan 1984
~ Liu 1991
0.0
0.5
1.0
F o rm a Id e h y d e { m g/m )
Figure 4. Prevalence of eye irritation among study groups exposed to formaldehyde in
residential settings and controlled human exposure studies.
Exposure in these studies was either in mobile trailer offices or residential mobile homes. Prevalence at
formaldehyde concentrations measured among comparison groups is graphed if reported (Holness and
Nethercott, 1989; Holmstrom and Wilhelmsson, 1988; Horvath et al., 1988; Olsen and Dossing. 1982).
Error bars are standard error (SE) calculated by EPA. Average weekly concentrations in three categories
for Liu et al. (1991) were estimated from the midpoint of each category of reported weekly cumulative
exposure (ppm-hour) and an assumption that individuals spent 60% of a 24-hour period at home.
3.1.4. Mode-of-action Information
Sensory irritation is understood to occur as a result of direct interactions of formaldehyde with
cellular macromolecules in the nasal mucosa leading directly or indirectly to stimulation of trigeminal
nerve endings located in the respiratory epithelium (see Figure 5; see Appendix A.5.6 for additional
details, related analyses, and discussion). This potential mechanism is most directly supported by
studies demonstrating increases in afferent nerve activity after acute exposure to approximately
0.5 mg/m3 formaldehyde or lower (Tsubone and Kawata, 1991; Kulle and Cooper. 1975), although
several related findings provide additional confirmation. While other mechanistic changes (e.g.,
oxidative stress; airway inflammation; damage or dysfunction of the respiratory epithelium, not shown)
and biological differences (e.g., nasal morphology; underlying allergy, infection, or other respiratory
conditions) are expected to be strong modifiers of this sequence of events, this pathway is interpreted
as likely to be the dominant mechanism by which formaldehyde exposure causes sensory irritation. All
of the mechanistic events in this pathway are supported by robust or moderate evidence, and the
relationships described are largely well-understood biological phenomena or have been demonstrated
This document is a draft for review purposes only and does not constitute Agency policy.
32 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Toxicological Review of Formaldehyde - Inhalation
following formaldehyde exposure. This mechanistic understanding provides strong support for the
biological plausibility of this effect. Although the primary support for an MOA reliant on stimulation of
receptors on nasal trigeminal nerve endings is from studies in experimental animal models, the
mechanistic events presumed to be driving sensory irritation after formaldehyde exposure are expected
to be conserved in humans.
Possible Initial Alterations Secondary Alterations
Effector-Level Changes
Key Hazard Feature
1s oxidative
stress in URI
URTTRPA1
binding
Trigeminal nerve
stimulation in URT
Centrally mediated
sensory irritation
Legend
^ Plausibly an initial
effect of exposure
~ Key feature of sensory
irritation
EVIDENCE
Q) Robust
( ) Moderate
(j Slight
RELATIONSHIP
—> Robust
--> Moderate
Slight
Figure 5. Likely mechanistic association between formaldehyde exposure and sensory
irritation.
An evaluation of the formaldehyde exposure-specific mechanistic evidence informing the potential for
formaldehyde exposure to cause respiratory health effects identified this sequence of mechanistic events
as likely to be the dominant mechanism by which formaldehyde inhalation could cause sensory irritation.
3.1.5. Overall Evidence Integration Judgment and Susceptibility for Sensory Irritation
Studies in humans provide robust evidence of sensory irritation based on the controlled human
exposure studies and observational epidemiological studies, and this effect also is well described and
accepted across a range of experimental animal species (robust), as well as in an established MOA based
on mechanistic evidence in animals (this MOA is interpreted to be operant in humans). Overall, the
evidence demonstrates that inhalation of formaldehyde causes sensory irritation in humans, given
appropriate exposure circumstances (Table 11). The primary basis for this conclusion is based on
residential studies with mean formaldehyde concentrations >0.05 mg/m3 (range 0.01->1.0 mg/m3) and
controlled human exposure studies testing responses to concentrations 0.1 mg/m3 and above.
This document is a draft for review purposes only and does not constitute Agency policy.
33 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 10. Evidence integration summary for effects on sensory irritation
Human Evidence
Animal Evidence
Additional Interpretations
Evidence
Integration
Judgment
Robust, based on:
Human health effect studies:
• 4 high and medium confidence studies
of symptom prevalence among adults
and children in residential settings
(mean >0.05 mg/m3 formaldehyde,
range 0.01->0.1 mg/m3) and
numerous high and medium
confidence studies involving acute
exposure (controlled human exposure
studies)
• Numerous high and medium
confidence studies with longitudinal
designs (occupational, panel studies
of pathology/anatomy lab courses)
• Consistent observations of irritation
symptoms in all studies; clear dose-
response gradients
Biological Plausibility: No directly
relevant human mechanistic studies
were found
Robust, based on:
Animal health effect studies:
Although these were not
formally evaluated,
formaldehyde inhalation-
induced sensory irritation in
rodents is a well-established
phenomenon (most notably,
as reflex bradypnea in mice
and rats)
Biological Plausibility. Robust
and moderate evidence for
mechanistic events from
animal studies identifies
stimulation of the trigeminal
nerve as the dominant MOA
• Relevance to humans:
Assumed, based on similarities
in systems mediating the
presumed MOA across species
• MOA: Established. Trigeminal
nerve stimulation is likely to be
the dominant mechanism
• Potential Susceptibilities:
Potentially large variations in
sensitivity are expected, due
primarily to differences in
nasal health (inflammatory
status; allergy) and physiology
• Other: This effect does not
appear to worsen with longer
exposure durations, although
uncertainties remain
The evidence
demonstrates
that inhalation of
formaldehyde
causes sensory
irritation in
humans given
appropriate
exposure
circumstances
Primarily based
on well-
conducted
residential
studies with
mean
formaldehyde
concentrations
>0.05 mg/m3 and
controlled human
exposure studies
testing >0.1
mg/m3
2 3.1.6. Dose-response Analysis
3 Because of the rapid nature of the irritant response generated by inhalation of formaldehyde,
4 the studies considered to be the most informative for derivation of a cRfC were those where the
5 exposure assessment was concurrent with the outcome assessment. Data from studies in humans
6 involving residential populations with continuous exposure, as well as controlled human exposure
7 studies evaluating acute effects, were determined to be pertinent to the derivation of a cRfC.
8 Study selection
9 The high and medium confidence studies that included information about dose-response
10 relationships for sensory irritation are presented in Table 12, which indicates for each study whether the
11 study was used to develop a POD or the rationale for why the study was not suitable.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
Table 11. Eligible studies for POD derivation and rationale for decisions to not select
specific studies
POD
Rationale for Decision
Reference
Endpoint
Derived?
Not to Advance
(Hanrahan et al., 1984)
Eye irritation: Prevalence
Yes
(Liu et al., 1991)
Eye irritation: Prevalence
No
Incomplete reporting of modeling
results; provided support for
analyses using Hanrahan et al.
(Kulle et al.. 1987)
Eye irritation: Prevalence
Yes
(Andersen and
Eye irritation: Prevalence
Yes
Molhave. 1983)
(Mueller et al.. 2013)
Eye irritation: Tear film break-up time,
symptom score using visual analogue
scale (VAS)
No
An exposure-response trend was not
observed for either endpoint.
Difficult to define an adverse
response level cutoff for these
endpoints
(Lang et al.. 2008)
Eye irritation: Conjunctival redness,
blinking frequency, symptom score
No
Difficult to define an adverse
response level cutoff for these
endpoints and appeared to less
sensitive than symptom score
Derivation of PODs
The PODs and supporting information from the studies are presented in Table 13. Two studies
involving adults and children in a residential exposure setting, for which there was medium confidence,
presented results based on responses at multiple exposure levels, although only the study by Hanrahan
et al. (1984) provided the quantitative results of statistical analyses necessary for dose-response analysis
(Liu et al.. 1991; Hanrahan et al.. 1984). Hanrahan et al. (1984) used one-hour average formaldehyde
measurements taken in two rooms in the mobile homes of a group including teenagers and adults, and
presented the predicted concentration-response for prevalence of "burning eyes" experienced by the
participants since moving into the homes from a logistic regression model that adjusted for age, gender,
and smoking. The mathematical expression for the dose-response pattern and a BMCLio was
determined from a graph of the predicted prevalence and upper and lower 95% confidence bounds for
several concentrations between 100 and 800 ppb (0.12-0.98 mg/m3).1 The concentration
corresponding to a 13% prevalence of "burning eyes" was calculated from the model based on a 10%
increase in irritation as a result of formaldehyde exposure in addition to an assumed background
prevalence of 3%. The background prevalence of 3% was considered to be a reasonable estimate, but
the impact of using alternative estimates (1% and 2%) was evaluated and was found to be minimal.
PODs also were determined using two controlled human exposure studies of formaldehyde for
which there was medium confidence that evaluated multiple levels of exposure (Kulle et al.. 1987;
1EPA estimates that 44% of the average measured concentrations were below 100 ppb (see Appendix B.1.2 for modeling details
and supporting rationale).
This document is a draft for review purposes only and does not constitute Agency policy.
35 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Andersen and Molhave. 1983). Kulle et al. (1987) evaluated results for participants exposed for 3 hours
2 once a week to five concentration levels (including a clean air exposure), while Andersen and Molhave
3 (1983) exposed subjects for 5-hour periods to four concentration levels with a 2-hour clean air exposure
4 prior to each trial. The occurrence of irritation symptoms during the clean air exposure was not
5 reported. Two sets of models were evaluated using the data from (Andersen and Molhave. 1983) and
6 estimates of 0% and 3% for prevalence of irritation during the clean air exposure. The BMC of 0.37
7 mg/m3 derived from the model using a baseline prevalence of 3% was selected.
8 Table 12. Summary of derivation of PODs for sensory irritation
Endpoint and
Reference
Population
Observed Effects by Exposure Level3
POD (mg/m3)
Residential exposure
Symptom prevalence
fHanrahan et al.. 1984)
Teenage and
adult (M and F),
n = 61
Third degree polynomial model fit to In prevalence odds
using presented results of logistic regression analysis: upper
95% confidence bound for predicted prevalence between
<0.123-0.98 mg/m3, BMCi0%: concentration where an
increased prevalence of 10% over a 3% background
prevalence is anticipated.
BMCiob 0.19
BMCL10 0.09c
Controlled human exposure
Symptom prevalence
fKulle et al.. 1987)
Nonsmoking,
healthy, n = 10-
19, Mean age
26.3 yr,
(M and F)
Exposure
mg/m3
nd proportion responding
0 0.62 1.2 2.5 3.7
BMCio 0.85c
BMC/2d 0.42
%
trend, p <
Probit moc
5 0 26 53 100
0.05
el BMC = 0.69 ppm
Symptom prevalence
f Andersen and
Molhave. 1983)
Healthy students,
n = 16, age 30-33
years, 31.2%
smokers (M and
F)
Exposure
end of exp
mg/m3
nd percentage responding (prevalence at the
osure)
0.3 0.5 1.0 2.0
BMCio 0.37c
BMC/2d 0.19
%
Assuming
0% LogLog
3% LogLog
19 31 94 94
jrevalence for clean air dose
stic model BMC = 0.26 mg/m3
stic model BMC = 0.37 mg/m3
Concentrations reported in publication converted to mg/m3
bBMCio benchmark concentration at 10% increase in prevalence over estimated 3% background prevalence. An
increase of 10% was selected consistent with EPA guidance (U.S. EPA, 2012) because the endpoint, burning eyes,
was considered a minimally adverse outcome.
cThe POD was not adjusted for a 24-hour equivalent concentration because the timing of formaldehyde
measurements was concluded to be appropriate to the timeframe of reported symptoms.
dThe BMD models did not account for the correlated measures between concentration levels (each participant was
exposed to each concentration). Therefore the 95% confidence limit for the BMC estimated by the model is too
narrow to use as the POD. A factor of 2 was used to adjust the BMC to identify a lower estimate that
approximates the BMDL
This document is a draft for review purposes only and does not constitute Agency policy.
36 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Toxicological Review of Formaldehyde - Inhalation
Derivation ofcRfCs
Table 14 describes the uncertainty factors used to adjust the POD to derive the cRfC for each of
the three studies. For the cRfC for sensory irritation in adult (and teenage) populations (residential
exposures) in Hanrahan et al. (1984), a UFH of 10 was used. Although the study population in Hanrahan
et al. (1984) comprised randomly selected households in mobile homes with individuals representing a
range of age, gender, health behavior, occupational status, and health status, the identified PODs were
not based on evaluation of differential susceptibility among subgroups with conditions or characteristics
that may contribute to variation in response. For the controlled human exposure studies (Kulle et al..
1987; Andersen and Molhave. 1983). a factor of 10 was applied to account for variation in the broader
human population not represented by participants in controlled human exposure studies. No other
uncertainty factors greater than one were applied.
Table 13. Derivation of cRfCs for sensory irritation
Endpoint (Reference; Population)
POD
POD Basis
UFa
UFh
UFl
UFS
UFd
UFcomposite
cRfC (mg/m3)
SENSORY IRRITATION
Eve irritation svmDtoms (^Hanrahan et al..
1984); adult M+F, n = 61, residential,
prevalence at POD 13%)
0.09
BMCLio
1
10
1
1
1
10
0.009
Eve irritation symptoms (fKulle et al..
1987): adult M+F. n = 10. controlled
exposure)
0.42
BMC/2
1
10
1
1
1
10
0.04
Eve irritation symptoms (^Andersen and
Molhave. 1983): adult M+F. n = 16.
controlled exposure)
0.19
BMC/2
1
10
1
1
1
10
0.02
Organ system-specific RfC (osRfC)
The POD was derived using the dose-response model using prevalence data from the residential
population in Hanrahan et al. (1984) is 0.09 mg/m3. The study by Hanrahan et al. (1984) is pertinent to
the U.S. general population because (1) the population was randomly selected from the general
population in the study area; (2) the exposure levels were concluded to reflect the usual, relatively
constant formaldehyde concentrations in the residences; and (3) exposed individuals included a range of
ages (teenagers and adults), men and women, some with chronic disease. Moreover, a significant
proportion of the study population was estimated to be exposed to average formaldehyde
concentrations below 0.05 mg/m3.
The PODs based on the two controlled human exposure studies were 0.19 and 0.42 mg/m3
(Kulle et al.. 1987; Andersen and Molhave. 1983), less than an order of magnitude greater than the
BMCL estimated from residential exposure. There is less confidence in the PODs based on these studies
because (1) the study participants were young, healthy volunteers, not representative of the age
distribution and health status in the general population; (2) the PODs are based on small sample size,
This document is a draft for review purposes only and does not constitute Agency policy.
37 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
more subject to random variation; and (3) formaldehyde concentrations were high, imposing substantial
uncertainty regarding responses at the low tail of the exposure distribution.
Therefore, the cRfC of 0.009 mg/m3 based on Hanrahan et al. (1984) was chosen as the osRfC for
sensory irritation. Confidence in the POD is medium because of uncertainties in the concentration
measurements relative to the study period for which the symptoms were being assessed. There is
extensive literature on this response to formaldehyde and the completeness of the database is high.
Because sensory irritation is an immediate response to exposure, the osRfC is applicable to short-term
as well as long-term exposure scenarios.
3.2. PULMONARY FUNCTION
Several studies in humans examined the effect of formaldehyde inhalation on pulmonary
function in various populations using different study designs. The systematic review process assigned
controlled human exposure studies of acute exposure involving healthy individuals to the review of
pulmonary function and the studies involving asthmatic volunteers to the review of effects on immune-
mediated conditions and their results are summarized there (see section 1.2.3). However, since all of
these studies involved measurements of pulmonary function, the results of the studies involving
participants with asthma have been integrated with the evidence from studies of acute exposure in
healthy individuals in this section. The few animal studies of analogous endpoints (all acute exposure)
were not included in the hazard evaluation. Changes in pulmonary function measures involving acute or
intermediate-duration exposures have been evaluated using experimental study designs (controlled
human exposure studies), panel studies of medical school anatomy students, and occupationally
exposed populations. In addition, occupational groups with long-term exposures are available, which
compared effects in exposed groups to effects in referents using different exposure metrics, such as
time-weighted average (TWA) or cumulative measures. Population-based studies of adults and children
that analyzed cross-sectional associations with average indoor formaldehyde concentrations also have
been conducted. Generally, groups exposed to formaldehyde at work experienced TWA concentrations
above 0.2 mg/m3 with intermittent peaks above 1 mg/m3. Students meeting once or twice a week in
anatomy labs experienced fluctuating concentrations during dissections averaging between 0.1 and
>1.0 mg/m3. Formaldehyde concentrations in residential or primary school settings are much lower and
less variable (<0.1 mg/m3). EPA included both the higher exposure and the lower exposure studies in its
evaluation of pulmonary function effects.
Poor pulmonary function as well as a decrease in pulmonary function is an important health
endpoint, associated with the development of chronic respiratory disease, coronary heart disease, and
mortality (Clayton et al.. 2014; Menezes et al.. 2014; Young et al.. 2007; Sin et al.. 2005; Schroeder et al..
2003; Schunemann et al.. 2000; Sorlie et al.. 1989). Spirometric measures (the focus of this section) are
commonly used diagnostic criteria. EPA considered a decrease in mean values to suggest a shift toward
a decline in the respiratory health status of the population. Consistent with this, the American Thoracic
This document is a draft for review purposes only and does not constitute Agency policy.
38 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Toxicological Review of Formaldehyde - Inhalation
Society (ATS) evaluated the clinical significance of small average declines in pulmonary function
observed in a population in response to air pollutants and concluded that although the magnitude of the
observed declines may not be clinically relevant to an individual, a shift in the population distribution
toward lower pulmonary function, assuming the association is causal, may have a large impact on public
health (ATS. 2000).
3.2.1. Literature Identification
This review focused on standard quantitative measures of pulmonary function (i.e., spirometry;
peak flow measurements). The bibliographic databases, search terms, and specific strategies used to
search them are provided in Appendix A.5.3, as are the specific PECO criteria and the methods for
identifying literature from 2016 - 2021 are described in Appendix F. Mechanistic studies relevant to
pulmonary function were separately identified (and evaluated) as part of the overarching review of
mechanistic data informing respiratory effects (see Appendix A.5.6 for additional details and supporting
analyses).
3.2.2. Study Evaluation
Forty-two observational epidemiological studies and eleven controlled human exposure studies
were evaluated for sources of bias and sensitivity. Pulmonary function is assessed using spirometry,
which measures the volume and speed of air that is exhaled or inhaled. Several parameters can be
measured during spirometric testing to characterize an individual's respiratory health. The
measurement of pulmonary function outcomes used by the studies in this section was considered to be
adequate if they followed the guidelines published by the American Thoracic Society (Tepper et al..
2012; Miller et al.. 2005a; Miller et al.. 2005b; Pellegrino et al.. 2005). or provided a description of the
protocols and reference equations that were used. In addition to the use of conventional spirometric
equipment to measure forced expiratory volume (FEVi), forced vital capacity (FVC), and forced
expiratory flow (FEF), peak expiratory flow (PEF) has been measured in research settings using portable
flow meters operated by study participants trained in their use. Although it requires careful training and
monitoring, this method has the advantage that it can be used in large epidemiological studies and
multiple measurements can be obtained over time. Studies of residential exposure to formaldehyde
were conducted in this way (Krzyzanowski et al.. 1990).
Lung function varies by race or ethnic origin, gender, age, and height, and is best compared
when normalized to the expected lung function based on these variables (Pellegrino et al.. 2005;
Hankinson et al.. 1999). Analyses were considered to be limited if they did not adjust or otherwise
account for these variables. Smoking status also was considered as a potential confounder. FEVi and
peak expiratory flow rate (PEFR) exhibit diurnal variation and this complicates the interpretation of
changes across a work shift or during a laboratory session if no comparisons were made with an
This document is a draft for review purposes only and does not constitute Agency policy.
39 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
unexposed group (Tepper et al.. 2012; Lebowitz et al.. 1997). Studies with no comparison group were
given less weight in evaluating the results of studies that measured short-term changes.
The healthy worker effect and survivor (lead time) bias was a concern for several cross-sectional
occupational studies, some of which had no other major limitations. Removal of individuals more
sensitive to the irritant effects of formaldehyde from jobs or tasks with formaldehyde exposure likely
occurred in industries with high formaldehyde exposures, and this type of selection bias might result in
an attenuation of risk estimates or a null finding if these individuals also experienced effects on
pulmonary function.
3.2.3. Synthesis of Human Health Effect Studies
While studies involving acute exposure either observed no change or inconsistent responses,
studies of occupational populations exposed over long periods and children exposed in residential
settings reported declines in pulmonary function. The controlled human exposure studies of healthy or
asthmatic volunteers consistently did not observe changes even at high concentrations, although two
studies by one research team observed small decrements (<5%) when longer exercise components (15
minutes) were included. Studies using shorter exercise components (8-10 minutes) reported no
changes. One exception among asthmatic participants was a heightened response to a dust mite
challenge in the formaldehyde inhalation arm compared to the clean air exposure in one study that used
nose clips, although a different study did not observe an increased response in a study with a similar
design but using a pollen challenge and no nose clip. Many of the studies of occupational groups or
anatomy students observed pulmonary function declines over the course of the work day or lab;
however, most did not account for diurnal changes, limiting the interpretation of these results. The few
studies of exposure during dissection labs that included an unexposed comparison group generally
reported that referent groups also experienced a change (increase or decrease) in pulmonary function.
A panel study using repeated peak expiratory flow measures taken by students trained in the procedure
at multiple points during dissection lab sessions found that PEF declined during the labs, and these
declines became attenuated over successive weeks (Kriebel et al.. 2001).
The review of the epidemiological literature provides evidence that long-term formaldehyde
exposure is associated with declines in pulmonary function, including FEVi, FVC, FEF, and PEF. Although
precision was low for most studies, pulmonary function was generally lower in highly exposed
occupational groups employed at exposed jobs for long durations compared to their nonexposed or
lesser-exposed comparison groups. The occupational groups under study were exposed to high average
formaldehyde concentrations (>0.2 mg/m3) in a variety of industries with different formaldehyde
sources. Employees had worked at these jobs for at least 5 years, and in a few studies, for more than 10
years. While a few studies conducted longitudinal analyses, most of the occupational studies were
cross-sectional in design, recruiting only current employees, and likely were limited by lead time bias, a
selection bias that results in attenuated effect estimates.
This document is a draft for review purposes only and does not constitute Agency policy.
40 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
Three studies conducted longitudinal analyses of small groups of workers with continued
exposure over 4-6 years (Lofstedt et al.. 2011; Nunn et al.. 1990; Alexandersson and Hedenstierna,
1989). For FEVi (only one study tested multiple parameters), all the longitudinal studies reported no
change in the full cohorts over the study period; however, one (Nunn et al.. 1990) reported that among
exposed nonsmokers, the annual decline was -45 mL/year (95% CI: -28, -62 mL/year), which is 50%
greater than the expected rate of decline in FEVi in nonsmokers [29 mL/year; (Redlich et al.. 2014; Lee
and Fry, 2010)1. In addition, Alexandersson and Hedenstierna (1989) reported a decline in FEF25-75 at a
TWA concentration of 0.42-0.5 mg/m3, with FEF25-75 percentage declining by -168 ± 46 mL/second
(10.1 L/minute) for each year of exposure over a five-year period (p < 0.001). The annual decrease was
corrected for normal aging and reference pulmonary function spirometry values. Consistent with the
results for FEVi, there was a larger decrease among nonsmokers compared to smokers, likely reflecting
the already reduced FEVi in smokers (-212 mL/sec/yr and -60 mL/sec/yr, respectively).
The longitudinal studies were limited with respect to duration of follow-up and sample size.
Given the large amount of within-person variability in these measures when assessed over time, these
studies would have had limited sensitivity to detect a small longitudinal change. Loss to follow-up of
exposed participants with symptoms also was evident. This type of selection bias also could result in an
attenuated effect estimate. Overall, the longitudinal analyses appear to be inconsistent, but while
hindered by a lack of sensitivity, seem to support a conclusion that occupational exposure may result in
declines in FEVi and FEF overtime.
Most of the occupational studies adjusted for smoking in statistical analyses or otherwise
addressed potential confounding by smoking, and two studies found no correlation between pulmonary
function measures and cigarette smoking indicating that smoking was not a confounder in the cohorts
(Malaka and Kodama, 1990; Holmstrom and Wilhelmsson, 1988). Potential confounding by coexposures
is an uncertainty for this review. However, many independent associations with formaldehyde for one
or more pulmonary function measures were found using statistical models that addressed potential
confounding (e.g., dust), and, since a pattern of reduction in pulmonary function was observed across
several different exposure settings (all involving high formaldehyde exposure), confounding by a
coexposure becomes less compelling as an alternative explanation for the observed associations.
Results among four studies of residential exposure among adults are difficult to compare
because different methods were used to assess pulmonary function and two of the studies did not
report results quantitatively (Norback et al.. 1995; Broder et al.. 1988). A cross-sectional study of
residential formaldehyde exposure in a large, randomly selected sample in Arizona observed an
association with declines in PEFR among adult smokers at formaldehyde concentrations between 0.049
and 0.172 mg/m3, but not among the group as a whole (Krzyzanowski et al.. 1990). Another study
among elderly nursing home residents observed an elevated risk of low pulmonary function defined as
values falling in the lower 20% of the distribution associated with formaldehyde concentrations above
the median level measured in each nursing home (overall median and range: 0.007 mg/m3 and
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
0.001-0.021 mg/m3; (Bentaveb et al.. 2015)). Two additional studies in primarily adult residential
populations exposed to concentrations between 0.009 and 0.279 mg/m3 reported no associations,
although the outcomes evaluated by each study were not equivalent (Norback et al.. 1995; Broder et al..
1988).
There are few studies of residential exposure among children; however, Krzyzanowski et al.
(1990) found a clear dose-response relationship among the children in their Arizona study where most
household concentrations were less than 0.045 mg/m3. A linear relationship between increased
formaldehyde exposure and decreased PEFR among children exposed to average concentrations of
0.032 mg/m3 was reported by this large, population-based, cross-sectional study of residential
formaldehyde exposure. The investigators reported a statistically significant decrease of
-1.28 ± 0.46 L/minute in PEFR per ppb household mean formaldehyde for all children. Figure 6 shows
the incremental decrement in PEFR measured at bedtime versus morning and shows differences in the
morning among asthmatics and nonasthmatics. Asthmatic children (15.8% of the total) showed a
steeper decline in PEFR in the morning at formaldehyde concentrations less than 0.049 mg/m3 (40 ppb).
The analysis of multiple PEFR measurements for each individual resulted in an increased statistical
power to detect an association at the lower formaldehyde levels present in the homes. The statistical
model adjusted for potential confounders including asthma status, smoking status, SES, N02 levels,
episodes of acute respiratory illness, and the time of day. Two other studies among children exposed to
similar levels of formaldehyde, but with limitations that reduced their sensitivity, did not find an
association for either FVC or FEVi (Wallner et al.. 2012; Franklin et al.. 2000).
HCHO (ppb)
Figure 6. Association of PEFR measured at bedtime and in the morning with
household mean formaldehyde concentration among children less than 15 years of
age (Krzyzanowski et al.. 1990).
Reproduced with permission.
This document is a draft for review purposes only and does not constitute Agency policy.
42 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Toxicological Review of Formaldehyde - Inhalation
3.2.4. Mode-of-action Information
There is mechanistic support, primarily from studies in animals, although a definitive MOA(s) has
not been fully defined (see Figure 7; see Appendix A.5.6 for additional details, related analyses, and
discussion). Overall, the most relevant mechanistic events included inflammatory structural alterations
and eosinophil increases in the lower airways that appear to be at least partially related to indirect
activation of sensory nerve endings. Inflammatory changes in the lower airways are supported most
directly by evidence from short-term studies of rodents at 0.3-2.5 rrig/m3 formaldehyde (Fujimaki et al..
2004; Riedel et al.. 1996), although indirect effects (e.g., biomarkers of airway oxidative stress) at lower
levels have been suggested in human studies (Flamant-Hulin et al.. 2010; Franklin et al.. 2000).
However, the initial cellular or tissue modifications that ultimately lead to these later events are not
understood, and it is unclear whether and to what extent certain events would be triggered with
chronic, low-level exposure. Although other important mechanistic events would likely be identified
with additional study, the available data provide reasonable support for the biological plausibility of the
observed associations and identify what is likely to be an incomplete mechanism by which formaldehyde
inhalation could cause decreased pulmonary function. Variation in sensitivity is likely to be affected by
underlying respiratory health status.
o -<:>—_
1s oxidative Sensory nerve ^ LRT neuro- ^ LRT micro •T Eosinophils -f- airway edema/ Decreased pulmonary
stress in LRT stimulation in LRT peptides vascular leakage in LRT* inflammatory function
structural change*
O ->0 -o O^O=S0
URT protein/DNA ^oxidative "t" URT URT mucociliary URT epithelial Mucus membrane Decreased pulmonary
modification stress in URT neuropeptides dysfunction damage change in URT function
Legend
Plausiblyan initial
effect of exposure
Key feature of decreased
pulmonary function
EVIDENCE
Robust
( ) Moderate
{ ; Slight
RELATIONSHIP
—> Robust
*¦> Moderate
Slight
'effects are amplified
with allergen exposure
Figure 7. Possible mechanistic associations between formaldehyde exposure and
decreased pulmonary function.
An evaluation of the formaldehyde exposure-specific mechanistic data informing the potential for
formaldehyde exposure to cause respiratory health effects identified these sequences of mechanistic
events as those most directly relevant to interpreting effects on pulmonary function. Evidence of airway
inflammatory changes, including eosinophil recruitment to both the upper and lower respiratory tract
(URT and LRT; upper pathway), is considered as likely to represent an incomplete mechanism by which
formaldehyde inhalation could cause decreased pulmonary function, although whether certain events
occur at lower exposure levels is unclear, and other unexplored mechanistic events are expected to
This document is a draft for review purposes only and does not constitute Agency policy.
43 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
contribute. URT modifications, primarily structural changes (bottom pathway), may also contribute;
however, this is not interpreted as likely to be a significant contributing mechanism.
1 3.2.5. Overall Evidence Integration Judgment and Susceptibility for Pulmonary Function
2 Overall, based on the moderate human evidence from observational epidemiological studies, as
3 well as slight animal evidence from mechanistic studies supporting biological plausibility, the evidence
4 indicates that long-term inhalation of formaldehyde likely causes decreases in pulmonary function in
5 humans given appropriate exposure circumstances (Table 16). The primary basis for this conclusion
6 includes a study of children and adults in a residential setting (mean, 0.03 mg/m3, maximum 0.17
7 mg/m3) and several studies of workers with long-term exposure to >0.2 mg/m3. The evidence is
8 inadequate to interpret whether acute or intermediate-term (hour to weeks) formaldehyde exposure
9 might cause this effect.
This document is a draft for review purposes only and does not constitute Agency policy.
44 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Table 14. Evidence integration summary for effects on pulmonary function
Human Evidence
Animal Evidence
Additional Interpretations
Evidence
Integration
Judgment
Moderate for Lona-Term Exposure (vrs). based on:
Slight based on:
• Relevance to humans:
The evidence
Human health effect studies:
Animal health effect
For MOA, related
indicates that
• 1 high and 2 medium confidence studies in residential
studies: Do not add
changes are expected
long-term
and school populations indicating that susceptible
support. No studies
to occur in humans,
inhalation of
individuals may experience reduced pulmonary
of exposures >1 day
given similarities
formaldehyde
function at lower average concentrations (mean, 0.03
Biological Plausibility:
across species in the
likely causes
mg/m3, maximum 0.17 mg/m3), and numerous high or
Robust and moderate
systems that appear
decreases in
medium confidence studies showing a pattern of
evidence for several
to be involved, and
pulmonary
lower mean pulmonary function in formaldehyde-
mechanistic events,
some support is based
function in
exposed occupational groups across a variety of
primarily from
on studies in both
humans given
exposure settings and countries
experimental animal
humans and animals
appropriate
• Dose-response trends from 4 high or medium
studies, provides
(e.g., lower airway
exposure
confidence adjusted analyses indicate an independent
support for
oxidative stress).
circumstances
association and argue against confounding
inflammatory
• MOA. Not established,
• Longitudinal declines in 1 occupational population and
changes in the lower
but likely to involve
Primarily based on
a panel study of medical students, but null or
airways, including
airway eosinophil
a study of children
equivocal associations from other studies, all with
eosinophil increases,
increases and
and adults in a
possible differential loss to follow-up resulting in low
which appear to be
stimulation of airway
residential setting
sensitivity
at least partially
sensory nerve
(mean, 0.03
Biological Plausibility: Some indirectly supportive
dependent on
endings.
mg/m3, maximum
mechanistic data from high or medium confidence
indirect stimulation
• Potential
0.17 mg/m3) and
human studies exists related to increased lower airway
of sensory nerve
Susceptibilities:
several studies of
oxidative stress following exposures likely to span
endings. While
Variation in sensitivity
workers with
months to years
evidence exists for
is anticipated to
long-term
some changes in the
depend on age and
exposure to >0.2
Indeterminate for Acute or Intermediate-Term Exposure
range of
respiratory health.
mg/m3
(hrs-wks), based on:
0.3-0.5 mg/m3 with
• Other: None
Human health effect studies:
exposure for several
[Note: The
Small reductions in two controlled human exposure
weeks, some
evidence is
studies of healthy volunteers (1 lab) with longer exercise
potential associations
inadequate to
periods, but no associations with other exposure
in the identified,
draw judgments
protocols in studies involving healthy subjects or
incomplete MOA
regarding acute or
asthmatics; inconsistent results among studies of medical
pathway have only
intermediate-
school dissection labs and cross-shift measurements in
been tested at higher
term exposure
occupational studies
(i.e., >1 mg/m3)
(hrs-wks)]
Biological Plausibility: Increases in lower airway
levels and with
eosinophils were not observed in the few low confidence
shorter term
acute studies in humans available
exposures
3.2.6. Dose-response Analysis
Study selection
The high and medium confidence studies that included information about dose-response
relationships for decreased pulmonary function are presented in Table 17, which indicates for each
study whether the study was used to develop a POD or the rationale for why the study was not suitable.
This document is a draft for review purposes only and does not constitute Agency policy.
45 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 15. Eligible studies for POD derivation and rationale for decisions to not select
2 specific studies
Reference
Endpoint
POD Derived?
Rationale for Decisions to Not Select
Krzyzanowski et al. (1990)
PEFR
Yes
Malaka and Kodama (1990)
FEVi, FEF25-75
No
Incomplete reporting of modeling results
Kriebel et al. (2001)
PEFR
No
Difficult to use modeling results because of
covariance in model coefficients
Wallner et al. (2012)
FEF25-75
No
Incomplete reporting of modeling results
3 Derivation of PODs
4 Declines in PEFR were associated with increases in 2-week average indoor residential
5 formaldehyde concentrations, with greater declines observed in children (5-15 years of age) compared
6 to adults (Krzyzanowski et al.. 1990). This study of effects in a residential population used the most
7 thorough exposure assessment protocol and repeated measurements of PEFR, thus enhancing the
8 ability to detect an association at lower concentrations. Mean formaldehyde levels were 26 ppb
9 (0.032 mg/m3), and more than 84% of the homes had concentrations 40 ppb (0.049 mg/m3) and lower.
10 A BMCio of 0.033 mg/m3 and a BMCLio of 0.021 mg/m3 were determined from the regression coefficient
11 from a random effects model of PEFR among children with and without asthma reported by the study
12 authors. Table 18 summarizes the study and the derivation of the POD for pulmonary function.
13 Table 16. Summary of derivation of PODs for pulmonary function
Endpoint and
Reference
Population
Results by Exposure Level3
BMC and BMCL
(mg/m3)
PODADjb
(mg/m3)
PEFR
^Krzyzanowski
et al.. 1990)
Residential,
prevalence
202 households, 298 children
aged 5-15 years, current asthma
prevalence 15.8%;
613 adults and adolescents >15
yr, 24.4% current smokers,
current asthma prevalence
12.9%
Random effects model; decreased PEFR,
children
-1.28 ± 0.46 L/minute-ppb (95% upper
bound -2.04 L/minute-ppb)
Formaldehyde concentrations: Mean
0.032 mg/m3, maximum 0.172 mg/m3
BMCioc 0.033
BMCLio 0.021
0.021
Concentrations reported in publication converted to mg/m3.
bThe POD was not adjusted for a 24-hour equivalent concentration because formaldehyde is present in all indoor
environments and time-activity information for participants was not reported.
cBMCio benchmark concentration at 10% increase in prevalence over background prevalence. A BMR of 10%
reduction in PEFR was selected as a cut-off point for adversity, based on rationales articulated by the American
Thoracic Society (see Appendix B.1.2 for details on the rationale).
14 Derivation ofcRfCs
15 Table 19 describes the uncertainty factors used to adjust the POD and the resulting cRfC. For
16 the POD for decreased PEFR among children from Krzyzanowski et al. (1990). a UFH of 3 was used with
17 support from the model results reported by the authors. While the BMC was defined as the
This document is a draft for review purposes only and does not constitute Agency policy.
46 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Toxicological Review of Formaldehyde - Inhalation
concentration where a 10% decrease in PEFR among all the children in the study was predicted to occur,
the model results also predicted the degree of response among asthmatic and healthy children.
Multiple observations in the study indicate that a UFH of 3 applied to the endpoint can be expected to be
protective of asthmatic children and other susceptible individuals. EPA used the published regression
coefficients from the random effects model to calculate the predicted decrease in PEFR from the
baseline level (i.e., formaldehyde concentration equal to zero) for each group. At the BMC
corresponding to a 10% decrease overall (0.033 mg/m3), the asthmatic children experienced a
decrement in PEFR that was 1.5-fold greater than that of the nonasthmatic children. Further, at the
BMCL selected as the POD (0.021 mg/m3), the decrease in PEFR among asthmatic children was 10.5%
while that in nonasthmatic children was 7.2%, a 1.5-fold difference. The authors stated that other
characteristics affecting variability, such as acute respiratory illness episodes during the observation
period, environmental tobacco smoke in the home, or socioeconomic status, did not increase sensitivity.
These observations indicate that a UFH of 1 is not appropriate since the asthmatic children experienced a
larger decline in PEFR compared to the healthy children. However, a UFH of 3 can be expected to be
protective of asthmatic children and other susceptible individuals.
Table 17. Derivation of the cRfC for pulmonary function
Endpoint (Reference; Population)
POD
POD Basis
UFa
UFh
UFL
UFS
UFD
UFcomposite
cRfC (mg/m3)
PULMONARY FUNCTION
Peak expiratory flow rate (fKrzvzanowski
et al.. 1990): Children M + F. n = 298.
residential)
0.021
BMCLio
1
3
1
1
1
3
0.007
Selection of osRfCs
The cRfC for pulmonary function of 0.007 mg/m3 (Krzyzanowski etal., 1990) was chosen as the
osRfC. This population-based study used a thorough exposure assessment based on two-week average
measurements in multiple rooms and two different seasons. Hence, confidence in the POD value is
high. The hazard conclusion is based on several studies in diverse exposure settings, and the
completeness of the database is considered high.
3.3. IMMUNE-MEDIATED CONDITIONS, FOCUSING ON ALLERGIES AND ASTHMA
This section examines the evidence pertaining to the effect of formaldehyde exposure on
immune-mediated responses, primarily in the respiratory system, including allergy-related conditions
(e.g., rhinitis; rhinoconjunctivitis) and asthma; dermal sensitization is not a focus of this review.
Epidemiological studies have investigated potential associations between formaldehyde and outcomes
relevant to various exposure durations and time windows in children and in adults using formaldehyde
measurements conducted in occupational, residential, and school-based settings. A few studies
This document is a draft for review purposes only and does not constitute Agency policy.
47 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
described other respiratory conditions in infants and toddlers, but these outcomes were not the focus of
the review. Controlled human exposure studies also are available that evaluated pulmonary function
responses to formaldehyde among subjects with asthma, but their results are most informative to the
pulmonary function outcome and are included in the integration of evidence in that section (see Section
1.2.2). Only two of these studies are relevant to the evaluation of effects on immune-related endpoints;
these studies assessed responses to an allergen challenge during formaldehyde exposures: dust mite in
Casset et al. (2006) and grass pollen in Ezratty et al. (2007). While exposures were high in occupational
settings (>0.1 mg/m3), formaldehyde concentrations measured in schools and homes averaged between
0.03 and <0.1 mg/m3.
Experimental animal studies were ultimately concluded to be unsuitable models (indeterminate)
for evaluating allergy-related conditions and asthma as apical endpoints. However, in the context of the
health effects data available, these findings, as well as a few studies that indirectly suggest that
respiratory immune function could be affected by formaldehyde exposure, are discussed within the
wider context of potential mechanistic changes that might explain respiratory health hazards.
3.3.1. Literature Identification
The focus of this search was on studies with a direct measure of formaldehyde exposure in
relation to measures of allergic respiratory conditions, eczema, or current asthma, reflecting the
question of whether formaldehyde exposure influences the sensitization response to respiratory
allergens. The bibliographic databases, search terms, and specific strategies used to search them are
provided in Appendix A.5.4, as are the specific PECO criteria and the methods for identifying literature
from 2016 - 2021 are described in Appendix F. Additionally, mechanistic studies relevant to immune-
mediated conditions, including potential immunological changes in distal tissues and blood, were
separately identified (and evaluated) as part of the overarching review of mechanistic data informing
respiratory effects (see Appendix A.5.6 for additional details and supporting analyses). Ultimately (see
Section 1.2.3 of the Toxicological Review for details), the animal hypersensitivity studies were included
as part of the overarching review of respiratory system-related mechanistic information (see Appendix
A.5.6) rather than as apical health effect studies; thus, they provided information about potential
mechanisms for the reviewed outcomes in the human studies.
3.3.2. Study Evaluation
The category of allergic sensitization and allergies includes allergic sensitization based on skin
prick tests and history of allergy-related symptoms. Because the time windows for exposure
assessments used in the studies had uncertain relevance to when sensitization may have occurred,
lower confidence was placed in the results of skin prick tests for studies in adults than in children (these
studies are not discussed in this Overview). For symptoms, International Study of Arthritis and Allergies
in Children (ISAAC) questionnaires for rhinitis or rhinoconjunctivitis were considered to provide an
This document is a draft for review purposes only and does not constitute Agency policy.
48 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
adequate basis for case ascertainment in studies in Europe and the United States; in studies in other
areas (i.e., areas that have not been included in ISAAC), specific mention of validation of the
questionnaire was needed to receive a high confidence rating.
Studies that ascertained asthma outcomes using American Thoracic Society (ATS)-based
questionnaires or subsequent variations [ISAAC, European Community Respiratory Health Survey
(ECHRS)] for prevalence of current asthma that include questions on medication use and symptoms
were considered to provide an adequate basis for case ascertainment in studies in Europe and the
United States; in studies in other areas (i.e., areas that have not been included in ISAAC), specific
mention of validation of the questionnaire was needed to receive this level of confidence. Some studies
included results for more than one asthma measure; in this assessment, outcomes that were defined
over a recent period were included (e.g., symptoms in the past 12 months), but outcomes defined over
a lifetime (e.g., ever had asthma) were not, as the formaldehyde measures available do not reflect
cumulative exposures that could be related to cumulative risk. Studies that did not clearly delineate the
period of ascertainment were included, but lower confidence was placed in these studies.
The age of study participants is an important consideration in the interpretation of various
measures. Specificity of symptom questions is reduced in the very young (<5 years) because wheezing
can occur with respiratory infections in infants and young children, and specificity is reduced at older
ages (e.g., >75 years) because of the similarities in symptoms and medication use for chronic obstructive
pulmonary disease and asthma (Abramson et al.. 2014; Taffet et al.. 2014). Rumchev et al. (2002), a
study of emergency room visits for asthma in children ages 6 months to 3 years, and two other studies
that examined wheezing episodes among infants (Roda et al.. 2011; Raaschou-Nielsen et al.. 2010), were
thus classified as not informative with respect to asthma.
The evaluation of controlled exposure studies of responses among asthmatic subjects examined
four primary elements: the type of exposure (paraformaldehyde preferred over formalin or undefined
test articles), use of randomization procedures to allocate exposure, blinding of the participant and of
the assessor to exposure, and the details regarding the analysis and presentation of results.
3.3.3. Synthesis of Human Health Effect Studies
Allergic conditions and sensitization
The general population studies in children and adults provide evidence of an association
between formaldehyde exposure and prevalence of rhinitis or rhinoconjunctivitis. The exposure range
was similar in these studies (0.04-0.06 mg/m3) and estimated RRs were comparable for rhinitis endoints
ranging from 1.14 to 1.21 for comparisons of the higher exposed to the referent groups. (Figure 8).
These studies were conducted in school children in France (Annesi-Maesano et al.. 2012). Romania
(Neamtiu et al.. 2019), and Korea (Yon et al.. 2019), and in adults in France (Billionnet et al.. 2011) and
Japan (Matsunaga et al.. 2008). The classification of rhinoconjunctivitis by Annesi-Maesano et al. (2012)
is the most sensitive and specific of the measures, and the narrower confidence intervals in this study
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
reflect the larger sample size. No other pollutants (e.g., NOx, PM2.5, acetaldehyde, acrolein, and
environmental tobacco smoke) analyzed by this study were associated with rhinoconjunctivitis.
Although the effect size is small, these are relatively common conditions and could result in a large
impact in the population. A stronger association with formaldehyde inhalation (two-fold risk) was seen
in the only study of atopic eczema (Matsunaga et al.. 2008). Eczema, while not indicative of an allergic
respiratory response, is often associated with other allergic disorders, including those affecting the
respiratory system. The five of the six high and medium confidence results for rhinoconjunctivitis,
rhinitis, and eczema (Matsunaga et al. reported results for two outcomes) showed associations with
formaldehyde exposure and residential formaldehyde concentrations in one null study were very low
(0.004 mg/m3). Consistent results are seen in studies in both children and adults in school and
residential settings (Figure 8). Two of the studies had sufficient sample size and range of exposure to
examine dose-response patterns and observed the highest relative risk estimates in the highest
exposure groups. Further, an analysis by categories of rhinitis severity in children observed a statistically
significant increasing trend in risk (Yon et al.. 2019). Two population-based studies evaluated atopy
based on skin prick tests (Garrett et al.. 1999; Palczynski et al.. 1999), but confidence in these analyses
was lower than for the studies of allergy symptoms. It was not certain that the time frame represented
by the exposure measurements were relevant to the development of sensitization as measured by skin
prick tests. Overall, the evidence indicates that formaldehyde exposure at levels seen in the general
population studies can enhance the immune hypersensitivity response to allergens.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
A. Highest Exposure Group per Study
Children
Norback et al. 2017-
(rhinitis) -
Neamtiu et al. 2019-
(eye, nose and skin symptoms) -
Yon et al. 2019-
(rhinitis)-
Annesi-Maesano et al. 2012-
(rhinoconjunctivitis) -
Adults
Billionnet et al. 2001-
(rhinitis)- —
Matsunaga et al. 2008-
(rhinitis) - —
(eczema) -
0.4
No quantitative results reported;
reported no association
Formaldehyde Levels (mg/m3)
Total Approximate Referent RR
N Midpoint
462
280
246
6,683
916
998
998
0.004
0.045
0.027
0.044
0.06
0.07
0.07
Per unit
mg/m3
<0.035
3.23
Per 0.01 1.21
mg/m3
<0.019 1.19
<0.028 1.14
<0.058 1.22
<0.058 2.25
4 5
10
Figure 8. Relative risk estimates for prevalence of allergy-related conditions in
children and adults in relation to formaldehyde in residential and school settings.
High and medium confidence studies are depicted for rhinitis (diamond) and eczema (circle) and symptom
combinations (square). Open symbols are for studies in children; closed symbols are for studies in adults.
Results from the highest exposure group in each study are depicted.
Asthma
The available general population studies also provide evidence of an association between
formaldehyde exposure and prevalence of current asthma, as determined by symptoms or medication
use in the past 12 months in studies with higher exposures (e.g., above 0.05 mg/m3), but associations
are not seen in settings with a lower exposure range (Figure 9). The six medium or high confidence
studies in homes or schools with relatively low exposures (<0.05 mg/m3, most from approximately
0.02-0.04 mg/m3) report relative risks around 1.0. This set of studies includes a variety of designs and
populations; the school-based studies are relatively large (from 1,014 to 6,683 total participants). Six
medium confidence general population studies in children or adults with exposures of 0.05-0.1 mg/m3
were available. Two of these included both children and adults (Zhai et al.. 2013; Krzyzanowski et al..
1990), and each provides evidence of a greater susceptibility in children. A limitation of the
Krzyzanowski et al. (1990) analysis is the relatively small number in the highest exposure group (n = 21).
The summary RR in children calculated for this review combining these two studies was 4.5 (95% CI:
0.76, 27). One other study of children (mean age 10 years) was a hospital-based case-control study that
calculated an OR of 2.74 per quartile increase in formaldehyde concentration (95% CI: 1.098, 5.516)
using a more specific diagnosis for prevalent asthma including symptoms over the previous 3 or more
months, and an FEVi increase of 15% in response to (B-agonist inhalation (Liu et al.. 2018). Of note, a
Canadian intervention study of impacts on symptom exacerbation among asthmatic children from
increasing ventilation rates in homes reported that a 50% reduction in formaldehyde concentrations in
the bedroom was associated with a 14 to 20% decrease in the annual change in some symptoms or
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
medical care in the intervention group (Laioie et al.. 2014). However, other coexposures also were
reduced by the intervention resulting in uncertainty in the independent effect of formaldehyde,
although the reductions were to a lesser extent and separate effects of the other factors were not
analyzed. Two other medium confidence studies with exposures above 0.05 mg/m3 were conducted
only in adults (Billionnet et al.. 2011; Matsunaga et al.. 2008); EPA has lower confidence in the results of
Matsunaga et al. (2008) because of the lower sensitivity and specificity of the asthma ascertainment
(self-report of medication use for asthma). The pattern of results is indicative of an elevated risk, as
none of the point estimates are below 1.0; however, the confidence intervals around each of the
estimates is relatively wide.
Relatively strong associations were seen in three studies examining prevalence of current
asthma in relation to formaldehyde exposure in occupational settings (i.e., >0.10 mg/m3). A greater
than three-fold increased risk of asthma was seen in each of these studies; the summary RR calculated
for this review was 3.79 (95% CI: 1.98, 7.28). One of the wood worker studies addressed potential
confounding by dust exposure by the inclusion of this variable in the analysis (Malaka and Kodama,
1990), and another study specifically noted that the measured dust levels were not related to high
formaldehyde exposure and that the asthma symptoms were not strongly related to other exposures
(Fransman et al.. 2003). The results from these studies may represent underestimates of risk, primarily
because these were prevalent cohorts with 2 or more years of work duration who would have lost
affected individuals prior to the study. In addition, in two of the studies, the comparison group included
workers who may have also been exposed to formaldehyde or other respiratory irritants, resulting in an
underestimate of the relative risks (Fransman et al.. 2003; Herbert et al.. 1994). Overall, given the
strength of the relative risks, the consistency of the associations seen in the three different workplaces
and populations, and the likelihood that the observed associations were underestimates of the true
associations, these studies collectively support a strong association between formaldehyde
concentrations above 0.10 mg/m3 in occupational settings and increased prevalence of current asthma.
This document is a draft for review purposes only and does not constitute Agency policy.
52 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
A. General Population, Low I xposures (< 0.0S0 rnE/m3)
Children
VcwtUl.,2K'3
Aiw 0.050 mg/m ]
Children
Liu et at, 2dB-
Ct a i&ll-
HT^MPOIVSt Ct . 3990 -
Adults
inai e 0.01 mg/m3)
r« , lOTa -
Fransmnn et al. 2003 (all |
(duration > 6.5 ye&r&| -
{>0.080 rrg/m")
Malbkfi ai d Kodaira, 19*90 -
-f
-r4-
0.5 1 2 5 10
Relative Risk
Formaldehyde levels {mg/m')
Total Approximate
N Midpoint Referent
99
112
93
0.20
0.08
0.08
0.16
0.12
¦cO.OS
<0.tM9
¦:o.oa
<0.049
<0.058
RR
S.48
1-S
2.65
1 Figure 9. Relative risk estimates for prevalence of asthma in children and adults in
2 relation to formaldehyde by exposure level in general population and occupational
3 studies.
High and medium confidence studies in the general population with highest exposure categories at
midpoints of <0.05 mg/m3 (Panel A) and >0.05 mg/m3 (Panel B), and occupational populations with
exposures >0.1 mg/m3. Laioie et al. (2014! was not included in the figure because the study assessed
percent change in current asthma symptoms over 12 months, not relative effect.
This document is a draft for review purposes only and does not constitute Agency policy.
53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
Two studies examined symptom frequency and medication use in the past 4 weeks, a measure
of asthma control among children with asthma. This population could represent a group with greater
susceptibility or vulnerability than the general population. Venn et al. (2003) reported a two- to
threefold increased risk of frequent symptoms associated with the highest quartile of exposure (>0.032
mg/m3) compared with <0.016 mg/m3, with some evidence of an increased risk at even lower exposures.
For nighttime symptoms, which may be most relevant with respect to measurements taken in the
bedroom, the relative risk estimate was 3.33 (95% CI: 1.23, 9.02). The case definition of wheezing
during the past year is interpreted as relevant to the definition of current asthma as used in this
assessment, since 88% of the cases also reported using a reliever inhaler in the past year. These results
were not impacted by inclusion of measures of room dampness in the models. In a smaller study of 37
low-income children in Boston, Dannemiller et al. (2013) observed higher formaldehyde levels in homes
of children with poor asthma control compared to those with better asthma control (geometric mean
0.066 and 0.042 mg/m3, p = 0.078).
Most of the acute formaldehyde exposure studies among adults with asthma provide little or no
evidence of short-term effects; no controlled exposure studies have been conducted in children with
asthma. Only two of these studies included an assessment of the response to an allergen challenge,
with effects on FEVi observed in one study (Casset et al.. 2006) but not the other (Ezratty et al.. 2007).
One difference in these studies is that the Casset et al. (2006) protocol used a nose clip, thus resulting in
inhalation solely by mouth.
3.3.4. Mode-of-action Information
The mechanistic information that may inform the potential for formaldehyde to affect allergic
conditions or asthma includes animal models using ovalbumin as an experimental allergen, which can
provide insight into some of the mechanistic changes that are relevant to these human conditions, while
not fully capturing the phenotype of human asthma or allergy-related conditions (see Section 1.2.3 of
the Toxicological Review for details on the decision to use animal hypersensitivity studies as mechanistic
support). The mechanistic evidence that provides the most direct information regarding the potential
role of formaldehyde in respiratory hypersensitivity responses consists of a set of high or medium
confidence studies (Larsen et al.. 2013; Fuiimaki et al.. 2004; Ito et al.. 1996; Riedel et al.. 1996;
Swiecichowski et al.. 1993).2 These studies differed in the conditions under which formaldehyde
affected the relevant endpoints, specifically increased bronchoconstriction and airway
hyperresponsiveness, using short-term and acute exposures in sensitized and nonsensitized animals.
The data do not indicate that formaldehyde is itself immunogenic, but instead suggest that
formaldehyde may augment immune responses to other allergens.
2Note: Swiecichowski et al. (1993) and Leikauf (1992) are considered to involve the same cohort of animals.
This document is a draft for review purposes only and does not constitute Agency policy.
54 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Toxicological Review of Formaldehyde - Inhalation
As shown in Figure 10, the analysis identified several pathways describing potential associations
between the most relevant mechanistic data available (see Appendix A.5.6 for additional details, related
analyses, and discussion). The mechanistic evidence indicates that formaldehyde exposure can induce
bronchoconstriction and lead to the development of hyperresponsive airways,3 particularly with
allergen sensitization. These heightened responses may be due to a combination of potentially
progressive changes, including neurogenic increases in tachykinins and eosinophil recruitment and
activation in the lung; however, there was an absence of reliable data supporting mechanistic changes
that are typically thought to be essential for sensitization (e.g., IgE). The mechanistic studies also
provide consistent evidence that formaldehyde may stimulate a number of immunological and
neurological processes related to asthmatic responses; however, a molecular understanding of how
formaldehyde exposure favors asthmatic Th2 responses has not been experimentally established.
Initial Alterations
Secondary Alterations
Effector-Level Changes
Key Hazard Feature
O
--><
^ airway edema/
inflammatory
structural change*
Airway hyper-
responsiveness*
T1 oxidative Sensory nerve ^ LRT neuro- 1s LRT micro- "t" Eosinophils
stress in LRT stimulation in LRT peptides vascular leakage in LRT*
0 ' C D O O O
f' stress 1s oxidative ^ CD8+T cells IL-4, 4* IFNy Altered B
hormone stress in blood in blood in blood cells
a -o —o o o
URT epithelial 'f URT fairway 1* CD8T T cells & ^Eosinophils Sustained airway Allergic Airway hyper-
damage inflammatory neuropeptides Th2-related in LRT* inflammation* sensitization responsiveness*
cells & factors cytokines in LRT
Altered antibody Allergic Airway hyper-
responses* sensitization responsiveness*
Legend
^ Plausibly an initial
effect of exposure
^_ Key feature of
[[.*1 respiratory immune-
mediated conditions
EVIDENCE
ijQi Robust
( ) Moderate
( ; Slight
RELATIONSHIP
—^ Robust
--> Moderate
Slight
'effects are amplified
with allergen exposure
Figure 10. Possible mechanistic associations between formaldehyde exposure and
immune-mediated conditions, including allergic conditions and asthma.
An evaluation of the formaldehyde exposure-specific mechanistic evidence informing the
potential for formaldehyde exposure to cause respiratory health effects identified these mechanistic
pathways. Similar to effects on pulmonary function, events related to indirect stimulation of lower
3Hyperresponsive airways (or hyperresponsiveness) represents a mechanistic event (supported by robust evidence) and a
potential key feature of respiratory health hazards that is defined to encompass any of a range of relevant airway features,
including hyperreactivity (exaggerated response) and hypersensitivity (lower dose to elicit response).
This document is a draft for review purposes only and does not constitute Agency policy.
55 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Toxicological Review of Formaldehyde - Inhalation
respiratory tract (LRT) sensory nerve endings (top pathway) were considered as likely to represent an
incomplete mechanism by which formaldehyde inhalation could cause airway hyperresponsiveness,
although whether certain events occur with chronic, low-level exposure remains unclear. While the
observed alterations to circulating antibodies (i.e., primarily IgG, not IgE) after formaldehyde exposure
might contribute to the development of both allergic sensitization and airway hyperresponsiveness
(middle pathway), in the absence of additional clarifying data, this was not identified as a likely
mechanism for these effects. Likewise, the slight evidence of altered T cell-related airway responses
and, secondarily, inflammatory eosinophil responses might be useful for explaining allergic sensitization
(bottom pathway) if additional data were available to better explain the pattern and strength of these
associations. Conversely, sustained airway inflammation, at least in animals previously sensitized to an
allergen, was considered likely to be an incomplete explanatory mechanism for airway hyper-
responsiveness. It is expected that there would be overlap between the top and bottom pathways for
airway hyperresponsiveness.
3.3.5. Overall Evidence Integration Judgments and Susceptibility for Immune-mediated Conditions
including Allergies and Asthma
Overall, based primarily on a moderate level of human evidence supporting an association from
the available epidemiological studies, as well as slight animal evidence from mechanistic studies
supporting biological plausibility (including molecular and cellular inflammatory changes and evidence of
hypersensitivity), the evidence indicates that inhalation of formaldehyde likely causes increased risk of
prevalent allergic conditions and prevalent asthma symptoms, as well as decreased control of asthma
symptoms given appropriate exposure circumstances (Table 21). The primary basis for this conclusion
includes studies of occupational settings (>0.1 mg/m3) and population studies where formaldehyde
concentrations measured in schools and homes averaged between 0.03 and <0.1 mg/m3.
Table 18. Evidence integration summary for effects on immune-mediated conditions,
including allergies and asthma
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence
Integration
Judgment
Moderate for Alleraic Conditions, based on:
Human health effect studies:
Small elevated risks in five out of six high or
medium confidence studies of prevalence
of rhinitis, conjunctivitis, and eczema
among adults and children in residential
and school settings with exposures in the
range of 0.04-0.06 mg/m3 formaldehyde.
Very low formaldehyde concentrations
were measured in the one null study.
Sliaht for Immune-Mediated
Respiratory Effects based on:
Animal health effect studies:
Experimental animal models are
generally considered to be
unable to reproduce the overt
manifestations of allergic
conditions and asthma
Biological Plausibility. Robust
evidence for mechanistic events
• Relevance to humans:
for mechanistic data,
while several events
supported by animal
data (e.g., amplified
bronchoconstriction;
eosinophil increases)
have an unclear direct
linkage to complex
human diseases like
asthma, these findings
The evidence
indicates that
inhalation of
formaldehyde
likely increases
the prevalence of
allergic
conditions in
humans given
appropriate
This document is a draft for review purposes only and does not constitute Agency policy.
56 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence
Integration
Judgment
Moderate for Asthma, based on:
Human health effect studies:
• Elevated risks in eight medium
confidence studies of prevalence of
current asthma in adults and children,
change after an intervention to reduce
exposure, or reduced symptom control
in children in residential settings
including homes with >0.05 mg/m3
formaldehyde; greater susceptibility
among children
• No elevated risk of current asthma in 6
high or medium confidence studies with
relatively low exposures (<0.05 mg/m3),
but associations with adequacy of
asthma control were observed in 1 study
at this lower exposure level
• Strongly elevated risks in 3 medium
confidence studies in occupational
settings with exposures from 0.100 to
>0.500 mg/m3
Biological Plausibility (both conditions)\
Studies in humans do not provide robust or
moderate evidence for mechanistic events
that clearly support the development of
asthma, although effects in the blood, such
as cytokine, cell, and antibody changes,
might contribute
exists in relation to
formaldehyde-induced
augmentation of responses to
allergens and airway
bronchoconstrictor effects in
animal models. Although
several events typically
associated with asthma were
not corroborated (i.e., slight or
inadequate evidence exists for
these events), moderate
evidence for mechanistic events
exists for stimulation by
formaldehyde of important
immunological and neurological
processes, including airway
eosinophil increases and other
inflammatory changes, that can
be reasonably associated with
effects on airway hyperreactivity
or other responses relevant to
the development of allergic
conditions and asthma
inform the potential for
exposure to result in
changes to relevant
neurological and
immunological
constituents present in
both human and rodent
airways
• MOA\ Not established,
but several incomplete
MOAs involving airway
inflammatory changes
are considered likely to
be involved
• Potential Susceptibilities:
Variation in sensitivity
anticipated depending
on respiratory health,
physiologic changes
during pregnancy, age,
and exposure to tobacco
smoke
• Other. None
exposure
circumstances
The evidence
indicates that
inhalation of
formaldehyde
likely increases
the prevalence of
asthma
symptoms in
humans, as well
as decreased
control of asthma
symptoms given
appropriate
exposure
circumstances
Both judgments
are primarily
based on studies
of occupational
settings (>0.1
mg/m3) and
population
studies where
formaldehyde
concentrations
measured in
schools and
homes averaged
between 0.03-
<0.1 mg/m3
1 3.3.6. Dose-response Analysis
2 Study selection
3 The high and medium confidence studies that included information about dose-response
4 relationships for allergic conditions and current asthma are presented in Table 22, which indicates for
5 each study whether a POD was developed or the rationale for why the study was not suitable.
This document is a draft for review purposes only and does not constitute Agency policy.
57 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 19. Eligible studies for POD derivation and rationale for decisions to not select
2 specific studies
Respiratory immune-mediated Conditions: Allergic Conditions
Annesi-Maesano et al.
Rhinoconjunctivitis prevalence:
Yes
(2012)
Children
Matsunaga et al. (2008)
Atopic eczema
Yes
(Yon et al.. 2019)
Rhinitis prevalence
No
Minimal details provided on
formaldehyde distribution
(Neamtiu et al.. 2019)
Allergy-like symptoms (eyes, nose
and skin)
No
Provided support for use of Annesi-
Maesano et al. (2012)
Garrett etal. (1999)
Atopy prevalence (skin prick tests):
Children
No
Uncertain window of exposure with
respect to skin prick test results
Palczvnski et al. (1999)
Atopy prevalence (skin prick tests):
Children
No
Uncertain window of exposure with
respect to skin prick test results; too few
individuals in third tertile
Respiratory Immune-mediated Conditions: Current Asthma
Krzvzanowski et al. (1990)
Current asthma prevalence:
Children
Yes
Annesi-Maesano et al.
Current asthma prevalence:
Yes
(2012)
Children
Matsunaga et al. (2008)
Current asthma prevalence: Adults
No
Definition of current asthma was narrow
and resulted in ascertainment of fewer
cases than would be expected
Palczvnski et al. (1999)
Current asthma prevalence:
Children and adults
No
Uncertainty regarding asthma definition
(current, ever?); few cases in third tertile
(n<5)
Kim etal. (2011)
Current asthma prevalence:
Children
No
Provided support for use of Annesi-
Maesano et al. (2012)
Mi etal. (2006)
Current asthma prevalence
No
Provided support for use of Annesi-
Maesano et al. (2012)
Respiratory Immune-related Conditions: Asthma Control
Venn et al. (2003)
Asthma control: Children
Yes
Dannemiller et al. (2013)
Asthma control: Children
Yes
3 Derivation of PODs
4 Allergic conditions and sensitization
5 The selected high confidence studies presented a dose-response analysis using formaldehyde as
6 three (Annesi-Maesano et al.. 2012) or four groups (Matsunaga et al.. 2008). NOAELs and LOAELs were
7 identified in each of these studies based on the pattern of risk seen across the exposure groups; the
8 PODs were based on NOAELs. The study by Annesi-Maesano et al. (2012) used a relatively long
9 exposure period (5 days) and was a very large study in a school-based sample of children in France
10 (n = 6,683) with analysis presented by tertile. Matsunaga et al. (2008) used 24-hour personal samples in
11 a study of 998 pregnant women in Japan. The primary limitation of the Matsunaga et al. (2008) study
12 was that it was conducted only among adults, and so was less able to address the variability in
13 susceptibility that would be anticipated within a population.
This document is a draft for review purposes only and does not constitute Agency policy.
58 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
For allergy-related conditions (rhinoconjunctivitis), EPA selected NOAEL and LOAEL values of
0.024 and 0.040 mg/m3, respectively, in the Annesi-Maesano et al. (2012) study. Higher values
(NOAEL = 0.046, LOAEL = 0.062) were selected based on the study in adults by Matsunaga et al. (2008).
Current asthma
Several residential and school-based exposure studies examined prevalence of current asthma
in relation to formaldehyde exposure in adults and children in relatively low exposure settings. The six
medium or high confidence studies at exposures of <0.050 mg/m3 do not indicate risk at these lower
exposure levels. Several of the relative risk estimates from the individual studies at these exposure
levels were limited by low statistical power. However, the consistency of the results, and the absence of
an increased risk in the study by Annesi-Maesano et al. (2012), a large school-based study (n = 6,683)
that used a 5-day sampling period for formaldehyde measurement, strengthens the basis for
interpreting this set of studies as indicating an absence of risk of current asthma below 0.05 mg/m3.
Based on the study by Annesi-Maesano et al. (2012) and this collection of studies, EPA selected a NOAEL
of 0.042 mg/m3for risk of current asthma.
Krzyzanowski et al. (1990) examined prevalence of current asthma in children (5-15 years of
age) in higher exposure residential settings (>0.05 mg/m3). These results are based on a relatively large
sample size, with a comprehensive exposure assessment protocol. An increased prevalence of current
asthma was seen in the highest exposure group in a categorical analysis. The exposure range in this
group was 0.075-0.172 mg/m3, but the study notes that few values were above 0.11 mg/m3. Based on
this information, EPA selected a LOAEL based on the midpoint of the range estimated as 0.075 to 0.11
mg/m3 (midpoint of 0.092 mg/m3). The middle exposure category was selected as a NOAEL, although
confidence in this NOAEL is less, given the imprecision of the estimate (n with asthma = 1).
EPA identified two studies that examined degree of asthma control in children with asthma in
relation to formaldehyde measures in the home (Dannemiller et al.. 2013; Venn et al.. 2003). The larger
sample size, longer sampling period, and more detailed dose-response analysis makes Venn et al. (2003)
a stronger basis for providing a POD. EPA selected a NOAEL of 0.027 mg/m3 (no or weak relative risks
seen below this value) and a LOAEL of 0.041 mg/m3 (two- to threefold increased risk of symptoms was
seen). The Venn et al. (2003) analysis also evaluated dose-response trends using logistic regression, and
EPA used the reported odds ratio per quartile exposure for frequent nighttime symptoms indicating
poor asthma control and the median exposure values for each quartile to estimate the concentration
associated with a 5% increase in prevalence of symptoms above that observed in the referent group (for
modeling details, see Appendix B.1.2). A BMR of 5% was selected because asthma attacks are overt
effects, generally requiring the use of drugs to control symptoms (i.e., a frank or adverse effect) (U.S.
EPA. 2012).
Table 23 presents the studies with the epidemiology data and sequence of calculations leading
to the derivation of a point of departure for each data set with effects relating to allergies and asthma.
This document is a draft for review purposes only and does not constitute Agency policy.
59 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 20. Summary of derivation of PODs for allergies and current asthma based on
2 observational epidemiological studies
Endpoint and
PODadj
Reference
Population
Observed Effects by Exposure Level
(mg/m3)
Allergic conditions
Rhinoconjunctivitis
Children
Prevalence 12.1%,
NOAEL: 0.024
(prevalence); school-
(M and F)
OR (95% CI) (adjusted)
LOAEL: 0.040
based exposure (5
n = 6,683
<0.0191 mg/m3 1.0 (referent)
days)
>0.0191-0.0284 1.11 (0.94,1.37)
Annesi-Maesano et al.
>0.0284- ~0.055 1.19 (1.03, 1.39)
(2012)
NOAEL selection: 0.024 mg/m3, midpoint of second exposure category
LOAEL selection: 0.040 mg/m3, midpoint of third exposure category
Atopic eczema
Adult women
Atopic eczema
Atopic
(prevalence); personal
(pregnancy
(5.7% prevalence)
eczema
monitor-based
cohort)
NOAEL: 0.046
exposure (24 hours)
n = 998
mg/m3 n OR (95% CI)
LOAEL: 0.062
Matsunaga et al.
<0.022 298 1.0 (referent)
(2008)
0.023-0.033 299 1.03 (0.47,2.29)
0.034-0.057 301 1.11 (0.50,2.42)
0.058-0.161 100 2.36 (0.92,6.09)
(trend p-value) (0.08)
0.058 to 0.161 vs. <0.058 2.25 (1.01,5.01)
per 0.0123 mg/m3 1.16 (0.99,1.35)
[Stronger associations in women with no family history of atopy]
For atopic eczema NOAEL selection: 0.046 mg/m3, midpoint for third
category; LOAEL selection: 0.062 mg/m3, estimated median of fourth
category (based on correspondence with Dr. Matsunaga)
For rhinitis NOAEL selection: 0.062 mg/m3, median of fourth category
Current Asthma/Degree of Asthma Control
Current asthma
Children
Exposure (mg/m3) na OR (95% CI)
NOAEL: 0.042
(prevalence);
(M and F)
<0.0191 2,200 1.0 (referent)
school-based
n = 6,683
exposure (5 days)
>0.0191-0.0284 2,200 1.10 (0.85,1.39)
Annesi-Maesano et al.
>0.0284-~0.055 2,200 0.90 (0.78,1.07)
(2012)
Approximation, based on tertiles, with total n = 6,590
NOAEL selection: 0.042 mg/m3, midpoint of third exposure category
Current asthma
Children
Exposure (mg/m3) N Proportion with asthma
NOAEL: 0.062
(prevalence);
(M and F)
<0.049 248 0.12
LOAEL: 0.092
residence-based
n = 298
0.049-0.074 24 0.04
exposure (two 1-week
0.075-0.172 21 0.24
periods)
(trend p-value) (0.03)
Krzvzanowski et al.
Only a few values were reported to be above 0.11 mg/m3.
(1990)
NOAEL selection: 0.062 mg/m3, midpoint of second category
LOAEL selection: 0.092 mg/m3, estimated midpoint of third category
This document is a draft for review purposes only and does not constitute Agency policy.
60 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Toxicological Review of Formaldehyde - Inhalation
Endpoint and
Reference
Population
Observed Effects by Exposure Level
PODadj
(mg/m3)
Asthma control among
children with asthma,
residence-based
exposure (3 days)
Venn et al. (2003)
Children
(M and F)
n = 194
Exposure (mg/m3) N Proportion
OR
(95% CI)
NOAEL: 0.027
LOAEL: 0.041
From
regression
results:
BMCL5:
0.0133
Frequent nighttime symptoms
<0.016 39 0.41 1.0 (referent)
0.016-0.022 35 0.49 1.40 (0.54,3.62)
0.022-0.032 36 0.53 1.61 (0.62,4.19
0.032-0.083 33 0.67 3.33 (1.23,9.01)
(trend p-value) (0.02)
per quartile increase 1.45 (1.06,1.98)
Frequent daytime symptoms
<0.016 37 0.62 1.0 (referent)
0.020-0.022 34 0.47 0.47 (0.47,1.25)
0.022-0.032 37 0.73 2.00 (0.71,5.65)
0.032-0.083 32 0.73 2.08 (0.71,6.11)
(trend p-value) (0.05)
per quartile increase 1.40 (1.00,1.94)
NOAEL selection: 0.027 mg/m3, median of third category
LOAEL selection: 0.041 mg/m3, median of fourth category (based
on correspondence with Dr. Venn)
Asthma control among
Children
Geometric mean formaldehyde (mg/m3)
NOAEL: 0.042
people with asthma,
(M and F)
Very poor control (score <12, n = 6) 0.066 mg/m3
residence-based
n = 37
All others (score >12, n = 31) 0.042 mg/m3 p =
0.078
exposure
(30 minutes)
Dannemiller et al.
(2013)
Derivation ofcRfCs
Table 24 describes the uncertainty factors used to adjust the PODs and the resulting cRfCs for
allergy-related conditions and asthma. For rhinoconjunctivitis among children from Annesi-Maesano et
al. (2012). a UFH of 3 was used for the POD. Childhood is a susceptible lifestage for asthma and allergy,
and the sample size of 6,600 children was large enough to have characterized an adequate spectrum of
human variability. However, a UFH of 1 was not used because susceptibility among subsets of the study
population was not specifically assessed. For the cRfC for atopic eczema in women by Matsunaga et al.
(2008), a UFh of 3 was used. Matsunaga et al. (2008) was a study of pregnant women, a sensitive
population for eczema prevalence, however no information was available for other sensitive lifestages,
including children, a subgroup with a higher prevalence of eczema compared to adults.
A UFh of 3 was used for the POD for current asthma prevalence among children from Annesi-
Maesano et al. (2012) using the same rationale as described above for rhinoconjunctivitis. For current
asthma prevalence among children with residential exposure (Krzyzanowski et al.. 1990), a UFH of 10
was used because susceptibility among subsets of the population was not specifically assessed, and the
precision of the NOAEL was lower compared to that in Annesi-Maesano et al. (2012). For Venn et al.
(2003), a UFh of 3 was used because the POD was based on the degree of asthma control in children
This document is a draft for review purposes only and does not constitute Agency policy.
61 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Toxicological Review of Formaldehyde - Inhalation
with asthma, a highly sensitive group. (A UFH of 1 was considered but the number of individuals in the
two higher exposure groups was relatively low [n = 31-35], and likely did not characterize a wide range
of human variability).
The PODs for all studies were based on the NOAEL; therefore, a UFL of 1 was applied. Further, a
UFs of 1 was used, based on the following rationale: (1) The definitions of prevalence of
rhinoconjunctivitis, current asthma, or atopic eczema involved symptoms occurring during the past 12
months, while asthma control included symptoms during the past 4 weeks. These time frames are
components of validated definitions for these conditions and are expected to capture the occurrence of
symptoms that tend to be intermittent. (2) The evaluation of children using residential or school-based
exposures is presumed to represent several years of exposure. This reflects a large portion of what is
expected to be a vulnerable lifestage for these effects (particularly for asthma-related measures).
Consistent with the rationale for developmental effects, this would not require the application of a UFS.
The study of the occurrence of atopic eczema during the past 12 months in a group of pregnant women
was an exception where a subchronic UF of 3 was applied to the POD. In Matsunaga et al. (2008), the
exposure assessment corresponded to the time during pregnancy, which is a less-than-lifetime window
of vulnerability. However, this outcome may have been pre-existing in a portion of the study sample
and the window of susceptibility may not have been sufficiently represented by the shorter exposure
period (Cho et al., 2010). Therefore, a UF of 1 was not applied.
This document is a draft for review purposes only and does not constitute Agency policy.
62 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Toxicological Review of Formaldehyde - Inhalation
Table 21. Derivation of the cRfC for allergy-related conditions and asthma
Endpoint (Reference; Population)
POD
POD basis
ufa
UFh
ufl
UFS
ufd
UFcomposite
cRfC
(mg/m3)
ALLERGY-RELATED CONDITIONS
Rhinoconiunctivitis prevalence [Annesi-
Maesano et al. (2012)
children M+F, n = 2200 at POD, school-
based exposure]
0.024
NOAEL
1
3
1
1
1
3
0.008
Atopic eczema prevalence [Matsunaga et
al. (2008) adult F (pregnant) n = 301 at
POD, personal monitor-based exposure]
0.046
NOAEL
1
10
1
1
1
10
0.005
ASTHMA
Current asthma prevalence [Annesi-
Maesano et al. (2012) children M+F. n =
2200 at POD, school-based exposure]
0.042
NOAEL
1
3
1
1
1
3
0.01
Current asthma prevalence [Krzvzanowski
et al. (1990) children M+F. n = 24 at POD.
residential]
0.06
NOAEL
1
10
1
1
1
10
0.006
Degree of asthma control [Venn et al.
(2003) with asthma M+F. n = 35 at POD.
residential]
0.013
BMCLs
1
3
1
1
1
3
0.004
Selection ofosRfC
The osRfCfor allergy-related conditions is based on one study in children (Annesi-Maesano et
al.. 2012) and one study in adults (Matsunaga et al.. 2008). Both PODs were based on NOAELs and are
interpreted with high confidence. In particular, the large study of children (n = 6,683) by Annesi-
Maesano et al. (2012) was better able to address the variability in susceptibility that would be
anticipated within a population. EPA selected an osRfC of 0.008 mg/m3, based on the overall greater
strength of Annesi-Maesano et al. (2012). The completeness of the database relating formaldehyde
exposure to allergic sensitization is considered to be high, based on the variety of endpoints,
populations, and exposure scenarios considered in these studies.
There were three cRfCs developed for asthma based on current asthma and degree of asthma
control (Annesi-Maesano et al.. 2012; Venn et al.. 2003; Krzyzanowski et al.. 1990). The POD based on
Annesi-Maesano et al. (2012) was derived from a NOAEL using a large study with a relatively long
exposure measurement period, supported by a collection of several other smaller studies. Although the
effect estimates derived by Venn et al. (2003) were less precise because of relatively small group sizes,
the POD derived from Venn et al. (2003) reflects the response among a susceptible population,
asthmatic children. To account for the different uncertainties in the PODs from the three studies, the
median of the three PODs, 0.006 mg/m3, was selected for the osRfC. The confidence in the PODs was
medium. As there was a relatively small number of limited studies (e.g., low statistical power,
incomplete reporting of study results and exposure measures) examining asthma risk in relation to
This document is a draft for review purposes only and does not constitute Agency policy.
63 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Toxicological Review of Formaldehyde - Inhalation
exposures between 0.05 and 0.1 mg/m3 and a scarcity of data pertaining to asthma control among
people with asthma, the database for asthma was considered to be medium.
3.4. RESPIRATORY TRACT PATHOLOGY
This section describes research on formaldehyde inhalation and pathology endpoints in the
respiratory system in experimental animal studies and observational studies in humans. Numerous well-
conducted experimental animal studies, while testing relatively high formaldehyde concentrations,
provide consistent support for concentration- and, to a lesser extent, duration-dependent upper
respiratory tract hyperplasia and metaplasia after formaldehyde exposure. These data are supported by
a set of four studies in formaldehyde-exposed workers that demonstrate consistent findings of an
elevated prevalence of nasal lesions such as hyperplasia and metaplasia. The evidence for metaplasia, in
particular, is considered to be the best representation of a potential health hazard.
In the URT, both hyperplasia and metaplasia often reflect adaptive tissue responses. These
cellular responses help reduce the impact of stressors by changing the structure or function of the
locally affected tissue (Harkema et al.. 2013). Hyperplasia, generally a response to cell injury, involves
an increase in the population of resident cells that results in additional cell layers noticeable by
histology, whereas metaplasia, which typically occurs following prolonged or repeated insults, results in
the replacement of one differentiated cell type with another, more resilient cell type (Harkema et al..
2013). Importantly, squamous metaplasia results in a hardened, drier, and non-ciliated skin-like layer
(Tomashefski. 2008). Along with the acquisition of a protective, barrier-type phenotype, this
metaplastic change causes a loss of normal tissue function, including reduced mucous secretion and
ciliary clearance (Harkema et al.. 2013). Thus, this loss of normal function is judged to be an adverse
outcome in and of itself (i.e., independent from its potential role in progression to cancer). As an
interpretation regarding adversity is less clear for hyperplasia, this discussion emphasizes the data on
squamous metaplasia.
3.4.1. Literature Identification
This review focused on histopathological endpoints and signs of pathology in respiratory
(including nasal) tissues. The bibliographic databases, search terms, and specific strategies used to
search them are provided in Appendix A.5.5, as are the specific PECO criteria and the methods for
identifying literature from 2016 - 2021 are described in Appendix F. The mechanistic studies related to
pathology endpoints were considered in the overarching mechanistic evaluation informing all potential
respiratory health effects (see Appendix A.5.6 for additional details and supporting analyses), the most
relevant results of which are summarized herein.
This document is a draft for review purposes only and does not constitute Agency policy.
64 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
3.4.2. Study Evaluation
Hyperplasia can be precipitated by damage to the nasal epithelium, which is evaluated
histologically by measures of, for example, cell loss or necrosis, epithelial degeneration, and erosions.
Relatedly, squamous metaplasia is an adaptive response to continued toxic insult that involves cellular
substitution. Thus, it is useful to consider these cellular damage-related endpoints in the context of
hyperplasia and metaplasia. While evaluations of necrosis- and cytotoxicity-related pathology can be
informative, these endpoints were generally inconsistently measured or poorly reported across the
available studies and are therefore only summarily discussed, whereas the potential development of
hyperplasia and metaplasia was documented in nearly all the long-term histopathological studies.
Studies that evaluated related outcomes, such as mucociliary flow rates, cellular proliferation counts
based on DNA labeling, and mucosal swelling (which generally only investigated acute or short-term
exposure), were included and summarized as part of the respiratory system MOA evaluation.
Given the large number of long-term exposure studies with information on URT pathology and
the focus of the assessment on the effects of lifetime formaldehyde exposure, this section focuses on
animal studies of subchronic or chronic exposure, and on human studies of occupational exposure
where exposed employees were generally employed for longer than five years. Exceptions include
discussion of shorter-term studies that might inform the potential for relationships between lesion types
and studies specifically considering differences in exposure paradigm for lesion induction.
For human studies that evaluated histopathological lesions in nasal biopsies, the evaluation
emphasized either a detailed explanation of how tissues were evaluated and scored, or a citation for a
standard method. Cross-sectional studies among occupational cohorts likely were influenced by the
selection of the workforce toward individuals less responsive to the irritant properties of formaldehyde,
with a reduction in sensitivity. Confidence in these studies was downgraded because of this limitation.
Age, gender, and smoking were considered to be important confounders to evaluate for effects on
pathological endpoints. Confounding by other coexposures in the workplace specific to the
occupational setting also was considered. Higher confidence was placed in studies with the ability to
differentiate between exposed and unexposed, or between low and high formaldehyde exposure.
In addition to general factors considered for all toxicology studies of formaldehyde inhalation
exposure (see Appendix A.5.1), factors specific to the interpretation of respiratory tract pathology were
considered to give greater weight to results from the large database of well-conducted studies. These
factors included: (1) the use of too few test subjects; (2) a failure to report lesion incidence or severity;
(3) the lumping of multiple lesions (e.g., squamous metaplasia and hyperplasia); (4) a failure to report
quantitative incidences or statistical analyses; (5) the use of insensitive sampling procedures (multiple
sections across multiple levels of the respiratory tract were preferred); and (6) use of an exposure
duration or follow-up that is likely insensitive for detecting slow-developing lesions (a duration of >1
year was preferred). Most studies of respiratory pathology used paraformaldehyde or freshly prepared
formalin, which yield high purity formaldehyde gas. In studies that tested commercial formalin,
This document is a draft for review purposes only and does not constitute Agency policy.
65 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
coexposure to methanol was less of a concern for investigations of URT respiratory pathology because
most inhaled methanol bypasses the nose and is readily absorbed in the lungs for systemic distribution.
3.4.3. Synthesis of the Human Health Effect Studies
The epidemiological studies that evaluated pathological endpoints in the nasal epithelium
indicated that formaldehyde exposure is associated with higher scores indicating a higher prevalence of
cells with morphological changes including squamous metaplasia. There was no evidence of a time-
dependent relationship with formaldehyde. Additionally, there was no indication that coexposure with
wood-dust or smoking modifies the pathological effects of formaldehyde.
Cross-sectional studies among occupational cohorts likely were influenced by the selection of
the workforce in favor of individuals less responsive to the irritant properties of formaldehyde, with
resulting bias toward null results. Despite this methodological limitation and subsequent reduction in
sensitivity, most of the studies observed increases in histopathological outcomes among exposed
workers, which increased confidence in the reported exposure-related associations. Nasal biopsies were
taken in four occupational studies, and tissues were subsequently stained and cell structure examined
according to variations of the Torjussen et al. (1979) method. The original Torjussen method scored
morphological characteristics of the nasal epithelium using a whole number between 0 and 8, with 0
indicating normal epithelium, 8 indicating carcinoma, and the midpoint of 4 signifying stratified
squamous epithelium with a horny layer. Despite the variations of this scale, in each study the lowest
numbers (0 or 1) always indicated normal cell structure while increasingly higher numbers indicated
more disruptive cellular changes. Although the focus of this section is nonneoplastic histopathological
lesions, the studies compared the means of the total score between exposed and referent groups.
Although more equivocal in one study (Boysen et al.. 1990), the four studies examining
histopathology found that participants exposed to average formaldehyde levels between 0.05 and 0.6
mg/m3 had a higher average histopathology score than their respective comparison group (Ballarin et al..
1992; Holmstrom et al.. 1989b; Edling et al.. 1988). While the studies were limited by probable survival
bias and, in some cases, other limitations resulting in a bias toward the null, a consistent association
with histopathological endpoints, including squamous metaplasia, was observed. Therefore, the
observational human data provide moderate evidence that inhaled formaldehyde induces
histopathological lesions in the URT, including squamous metaplasia.
3.4.4. Synthesis of the Animal Health Effect Studies
A large database of well-designed studies has characterized formaldehyde-induced respiratory
tract pathology in mice, hamsters, and monkeys, but primarily in rats. The durations of these studies
ranged from a few hours to longer than 2 years, and several studies having included recovery periods
that explored the reversibility of lesions. Because of the abundance of studies of respiratory pathology,
this section focuses on longer duration (i.e., chronic and subchronic) studies interpreted with high or
This document is a draft for review purposes only and does not constitute Agency policy.
66 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
medium confidence, primarily studies in rats (see Section 1.2.4 in the Toxicological Review for an
expanded discussion of pathology in other species). Finally, although other nasal lesions have been
observed to develop after formaldehyde exposure (e.g., necrosis), this summary focuses on the more
reliably evaluated and more consistently reported information on hyperplasia and metaplasia.
Only a few studies evaluated sections of the respiratory tract distal to the nasal cavity, and these
evaluations were generally less rigorous (e.g., examining only a single tissue section). Pathological
findings in the lower respiratory tract were generally not identified in higher confidence studies, and are
not discussed in detail in the assessment. However, the limited evidence for lesions beyond the nasal
cavity in rats suggests that concentration is an important variable in long-term studies. Laryngeal or
tracheal lesions, including hyperplasia and squamous metaplasia, were only observed at high
concentrations, with no evidence for effects across multiple rat strains at levels <12 mg/m3. Findings in
a single study of rhesus monkeys observed changes in URT regions proximal to the nasal cavity (but not
the lungs) at lower concentrations (i.e., exposure for <6 weeks to 7.4 mg/m3 formaldehyde in Monticello
et al. (1989), which might suggest that the monkey nose is less efficient than the rodent nose at
scrubbing formaldehyde from inhaled air.
Hyperplasia and metaplasia have been consistently reported in multiple rodent species/strains,
and in monkeys, with consistent and clear indications of concentration-dependence. These studies also
identify a clear relationship between formaldehyde exposure duration and the development of
squamous metaplasia, with somewhat weaker data indicating a duration-dependency for hyperplasia.
Both squamous metaplasia and hyperplasia appear to be at least partially reversible after exposure
ceases. Due to the high reactivity and water solubility of formaldehyde, nasal metaplasia and
hyperplasia have primarily been assessed (and subsequently observed) in the epithelium lining the
anterior regions of rodent nasal passages (typically levels I, II, and III: level I refers to the area posterior
to the nostrils, with higher levels indicating more posterior sites) following formaldehyde inhalation
exposure, mostly in regions containing respiratory epithelium.
Squamous metaplasia, in particular (which, as previously mentioned, is considered adverse), has
been observed after chronic, subchronic, and short-term exposure to inhaled formaldehyde. Overall,
the most robust responses (i.e., higher incidence or severity at lower formaldehyde concentrations)
occur following chronic exposure, particularly in rats. As compared to rats, other laboratory rodents
appear to require higher levels (i.e., mice) or exhibit a reduced response (i.e., hamsters), suggesting that
there may be differences in species sensitivity to formaldehyde-induced squamous metaplasia. These
differences in sensitivity are likely at least partially due to differences in the magnitude of reflex
bradypnea across species. Multiple chronic rat studies have reported clear increases in squamous
metaplasia following exposures of approximately 2.5-2.7 mg/m3 (Kamata et al.. 1997; Kerns et al.. 1983;
Battelle. 1982) or 11.3-11.6 mg/m3 (Woutersen et al.. 1989; Appelman et al.. 1988). although some data
suggest that slight increases might be present at lower levels (i.e., 0.4-1.2 mg/m3; (Kamata et al.. 1997;
Woutersen et al.. 1989). With subchronic exposure, squamous metaplasia is observed in rat noses at
This document is a draft for review purposes only and does not constitute Agency policy.
67 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Toxicological Review of Formaldehyde - Inhalation
higher concentrations (i.e., >11.3 mg/m3) in high confidence studies by Appelman et al. (1988).
Woutersen et al. (1987), and Feron et al. (1988), the results of which are supported by consistent
observations in two medium confidence studies (Andersen et al.. 2010; Zwart et al.. 1988), although
these latter studies observed increases at lower exposure levels (i.e., 2.5-3.7 mg/m3). The rat data from
medium or high confidence studies of chronic formaldehyde exposure are summarized in Figure 11.
10°1 ~ ~ ¦ ~
90- O
80-
a, 70"
S 60-| O
¦| 50H
^ 40i
30-
20-
10^
O
0 3 6 9 12 15 18
Formaldehyde concentration (mg/m3)
High confidence
Medium confidence
~
Kerns, 1983 (F344; 2yr; n=27-100)
o
Sellakumar, 1985 (SD; life; n=99-100)
¦
Woutersen, 1989 (Wistar; 2.5yr; n=26-28)
o
Kamata, 1997 (F344; 2.5yr; n=32j
~
Appelman, 1998 (Wistar; lyr; n=10)
Figure 11. Squamous metaplasia incidence in chronic pathology studies of rats.
Smaller symbols reflect smaller sample sizes. High confidence studies are outlined in black.
The duration-dependency of these lesions in rat studies represents an important consideration.
For squamous metaplasia, the duration of exposure affects the locations at which lesions develop, as
well as their severity, probably in parallel with increases resulting from increasing formaldehyde
concentration. The association with lesion location is demonstrated by the results of Kerns et al. in the
supporting Battelle report (1983; 1982) who observed that, in anterior nasal regions (i.e., level I and II)
of F344 rats exposed to >2.5 mg/m3, the incidence of squamous metaplasia increased from <20% to
100% with increasing duration (i.e., 6-24 months); however, in posterior nasal regions (i.e., levels 111—V),
a duration-dependent increase in incidence was only observed at 17.6 mg/m3. In some instances, noted
by Kerns et al. (1983; 1982). more posterior lesions were entirely unique to longer exposure durations as
compared to shorter exposures (e.g., level III at 6.9 mg/m3 only with 24 months of exposure). Regarding
This document is a draft for review purposes only and does not constitute Agency policy.
68 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Toxicological Review of Formaldehyde - Inhalation
severity, squamous metaplasia was observed to increase (i.e., from slight focal lesions to metaplasia
with keratinization) with exposure duration increases from 13 to 52 weeks of exposure to 11.6 mg/m3 in
Wistar rats (Appelman et al.. 1988). Similarly, at >11.6 mg/m3 in Wistar rats, an increase in the severity
of squamous metaplasia in respiratory epithelium occurred as exposure duration increased from 4 to 8
to 13 weeks (Feron et al.. 1988). Several studies in rats, which compared longer-term exposure to
shorter-term exposure, confirm the important role for exposure duration in lesion development by
demonstrating that the increases in lesions were not attributable to longer latencies after the
formaldehyde exposures were begun (Woutersen et al.. 1989; Feron et al.. 1988). When animal ages at
evaluation and formaldehyde exposure levels were matched, comparisons of subchronic exposure to
chronic exposure (Woutersen et al.. 1989) and of short-term exposure to subchronic exposure (Feron et
al.. 1988) revealed greater incidences or severity of these lesions with the longer exposure durations.
Comparisons of the formaldehyde concentrations at which significant increases in hyperplasia
are observed across studies of differing exposure duration do not provide as clear a picture regarding
the potential duration-dependence of formaldehyde-exposure induced hyperplasia. However, like the
results for metaplasia, several rat studies comparing exposures of differing exposure duration (e.g.,
chronic versus subchronic) demonstrate that increasing exposure duration results in increases in the
incidence or severity of hyperplasia in the respiratory epithelium when testing the same formaldehyde
concentrations and anatomical levels (Woutersen et al.. 1989; Appelman et al.. 1988; Feron et al.. 1988;
Kerns et al.. 1983; Battelle, 1982). Considering the notable influence of exposure duration on
metaplasia at formaldehyde levels ranging from 2.5 to 2.7 mg/m3 in rat studies (Kamata et al.. 1997;
Kerns et al.. 1983; Battelle. 1982), the easier reversibility of hyperplasia, and the generally more robust
effects of duration on the incidence of metaplasia as compared to hyperplasia across species, exposure
duration appears to be more important to the development of metaplasia in laboratory animals than to
the development of hyperplasia. Rat studies by (Wilmer et al.. 1989, 1987) indicate that formaldehyde,
perhaps similar to mortality responses following acute exposure to some other local irritants (see
below), does not appear to adhere strictly to Haber's rule for the induction of nasal pathology. Although
duration of exposure has a clear and substantial role for the development of these nasal lesions (see
discussion above), the experiments by (Wilmer et al.. 1989, 1987) suggest that a powers equation (Cn x t
= K) where n is >1 may better represent formaldehyde exposure-induced nasal lesions than C x t = K, at
least when interpreting short-term or subchronic exposure (the exposure scenarios examined by Wilmer
et al.). Although a value for n was not identified for formaldehyde, or specifically for exposure-induced
nasal pathology, studies of acute exposure to other local irritants and the concentration-duration
dependence for mortality suggest that the value for n, on average, is approximately 1.8-1.9 (ranging
from 0.5 to 4.0)4. It is difficult to speculate where within this range a value for n might be most
4Values of n for 11 local irritants as estimated by Berge et al. (1986) averaged 1.9 (range: 1.0-3.5), while 21 local irritants relying
on data in rats or mice, as summarized by California EPA (2008). averaged 1.8 (range: 0.5-4.0). Of potential interest to this
assessment, the chemicals included ammonia (n = 2.0) and acrolein (n = 1.2).
This document is a draft for review purposes only and does not constitute Agency policy.
69 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
applicable to formaldehyde, particularly within the context of respiratory pathology and long-term
exposures (i.e., since these n values are for mortality after acute exposure); however, based on the data
discussed in previous sections, it might be reasonable to expect that an n defined for associations with
hyperplasia should be higher than one defined for metaplasia.
Overall, a number of well-conducted studies across multiple species (i.e., rats, mice, and
monkeys) demonstrate a clear association between formaldehyde exposure and the development of
respiratory tract pathology, primarily in the nasal cavity.
3.4.5. Mode-of-action Information
Histopathological lesions in the respiratory tract following formaldehyde exposure appears to
result, at least in part, from a series of increasingly severe effects including altered mucociliary function,
damage to the nasal epithelium (e.g., sustained cytotoxicity), and sustained reparative cell proliferation
culminating in a hyperplastic epithelium, or transitioning to an adaptive, metaplastic tissue (see
Figure 12; see Appendix A.5.6 for additional details, related analyses, and discussion). Consistent with
observations of metaplasia without hyperplasia in some of the rodent health effect studies, this
pathway illustrates that metaplasia can develop following damage (noting that damage does not need
to be overt) to the epithelium in the absence of hyperplasia (i.e., hyperplasia may not be an essential
precursor). All the mechanistic events and relationships between events in the proposed pathway are
based on robust or moderate evidence, indicating that this is likely a mechanism by which formaldehyde
exposure causes squamous metaplasia. Specifically regarding the well-established alterations to mucus
flow and proliferation, mucociliary function appears to be affected at relatively low concentrations (e.g.,
0.25-0.3 mg/m3) in humans (Holmstrom and Wilhelmsson. 1988; Andersen and Molhave. 1983).
whereas multiple high and medium confidence rodent studies do not see notable changes in either
mucociliary function or proliferation below 1.23 mg/m3 (increases generally occur above 2.5 or
sometimes 3.5 mg/m3) [e.g., (Monticello et al.. 1996; Monticello et al.. 1991; Morgan et al.. 1986a;
Morgan et al.. 1986c)1. Overall, consistent with some of the animal health effect studies, these data
suggest that concentration is likely to be more of a driver of these mechanistic effects than duration
(noting that duration still contributes). Because modification of epithelial cell health and function in the
URT can occur via multiple direct and indirect mechanisms following formaldehyde inhalation, which are
expected to vary due to differences in both exposure duration and intensity, there are likely to be other
plausible mechanisms by which formaldehyde exposure could cause this health effect. The current
understanding provides strong biological support for an association between formaldehyde exposure
and respiratory tract pathology. Additionally, as many of the mechanistic events in this pathway have
been observed in both humans (sometimes indirectly) and experimental animals, including effects on
mucociliary function and cell proliferation as well as evidence of elevated oxidative stress, findings from
experimental animals are considered relevant to humans.
This document is a draft for review purposes only and does not constitute Agency policy.
70 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Possible Initial Alterations Secondary Alterations
f1 oxidative URT protein/ DNA URT mucociliary
stress in URT modification dysfunction
Effector-Level Changes Key Hazard Feature
URT epithelial URT epithelial Squamous
damage proliferation metaplasia
Legend evidence relationship
+ Plausibly an initial if ^ Robust -^-Robust
effect of exposure
' ) Moderate --> Moderate
Q Key feature of respiratory
tract pathology Slight Slight
Figure 12. Possible mechanistic associations between formaldehyde exposure and
respiratory tract pathology.
Ari evaluation of the formaldehyde exposure-specific mechanistic evidence informing the potential for
formaldehyde exposure to cause respiratory health effects identified this sequence of mechanistic events
as likely to be a mechanism by which formaldehyde inhalation could cause respiratory tract pathology,
specifically squamous metaplasia, although it is assumed that other plausible pathways explaining this
association have yet to be defined.
1 3.4.6. Overall Evidence Integration Judgment and Susceptibility for Respiratory Tract Pathology
2 Overall, the strength of the evidence for hyperplasia and squamous metaplasia include robust
3 evidence of an effect in animals and moderate human evidence from observational epidemiological
4 studies, supported by more limited findings in mechanistic studies of exposed humans and strong
5 support for a plausible MOA based largely on mechanistic evidence in animals (with coherent findings in
6 human studies). Therefore, the evidence demonstrates that inhalation of formaldehyde causes
7 respiratory tract pathology in humans given appropriate exposure circumstances (Table 26). The
8 primary basis for this conclusion is based on rat bioassays of chronic exposure which consistently
9 observed squamous metaplasia at formaldehyde exposure levels >2.5 mg/m3.
10 Table 22. Evidence Integration Summary for Effects of Formaldehyde Inhalation on
11 Respiratory Pathology
Human Evidence
Animal Evidence
Additional Interpretations
Overall Evidence
Integration
Judgment
Moderate, based on:
Human health effect studies:
Of the 4 occupational studies
interpreted with medium
confidence (less sensitive due
to healthy survival bias), 3
observed a higher prevalence
Robust, based on:
Animal health effect studies:
• Consistent evidence of squamous
metaplasia and hyperplasia in the
nasal respiratory epithelium
across numerous independent
studies interpreted with high or
• Relevance to humans:
Similarities in the function and
properties of the nasal
epithelium across species, as
well as similar mechanistic and
apical effects observed in both
humans and animals, support
the relevance of findings in
The evidence
demonstrates that
inhalation of
formaldehyde
causes respiratory
tract pathology in
humans given
appropriate
This document is a draft for review purposes only and does not constitute Agency policy.
71 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Human Evidence
Animal Evidence
Additional Interpretations
Overall Evidence
Integration
Judgment
of abnormal nasal
histopathology including loss of
ciliated cells, hyperplasia and
squamous metaplasia at
concentrations ranging from
0.1-2 mg/m3, while the
remaining (1) study had more
equivocal findings
Biological Plausibility:
Mechanistic changes in 2
studies (1 interpreted with
medium confidence) in humans
provides evidence of changes in
mucociliary clearance and
mucus flow beginning at
formaldehyde concentrations
of 0.25-0.3 mg/m3
medium confidence, with
generally the most sensitive
effects being metaplasia
observed after chronic exposure
to >2.5 mg/m3 formaldehyde
• Metaplasia and hyperplasia in
monkeys (limited data), rats,
mice, and hamsters; hamsters
and mice were less sensitive
• Multiple studies provided clear
evidence of a concentration-
dependence for lesion
development, as demonstrated
by increases in the incidence,
severity, and anatomical location
of the observed lesions with
increasing exposure
Biological Plausibility: Robust or
moderate evidence for mechanistic
events based predominantly on
experimental animal studies
supports a biological progression of
changes that appears to include
mucocilliary dysfunction, epithelial
damage and, oftentimes, cellular
proliferation, leading to the
eventual development of nasal
lesions, including squamous
metaplasia.
experimental animals to
humans.
• MOA\ Although it may be
incomplete, a MOA involving
effects on mucocilliary
function and epithelial cell
health is well supported and
considered to be a major
contributor the these effects
• Potential Susceptibilities:
Variation in sensitivity may
depend on differences in URT
immunity and nasal structure
or past injury (e.g., studies
support increased sensitivity
of rodents with intentionally
damaged nasal cavities), and
males may be more sensitive
than females
• Other: Animal studies suggest
that lesion development may
be driven more by
concentration than duration,
particularly for hyperplasia.
Estimates for formaldehyde
were not identified; estimates
for other irritants indicate that
concentration is ~1.8-1.9-fold
(on average) more influential
regarding exposure-induced
mortality after acute exposure
exposure
circumstances
Primarily based on
rat bioassays of
chronic exposure
which consistently
observed
squamous
metaplasia at
formaldehyde
exposure levels
>2.5 mg/m3
1 3.4.7. Dose-response Analysis
2 Study selection
3 Of the medium and high confidence studies that exposed rodents to formaldehyde for at least
4 one year, the chronic rat bioassays by Kerns et al., (1983; 1982) and Woutersen et al., (1989) were
5 considered to be the most informative for dose-response analysis (see Table 27 for the rationale
6 supporting this decision). As an interpretation regarding adversity was less clear for hyperplasia, dose-
7 response analysis relied on the data on squamous metaplasia.
8 Table 23. Eligible studies for POD derivation and rationale for decisions to not select
9 specific studies
Respiratory Pathology (Animal Exposure Duration >52 Weeks; Humans All Employed >5 Years)
Kerns etal. (1983):
Battelle (1982)
Squamous metaplasia: nasal
turbinates, Fischer 344 rats
Yes
This document is a draft for review purposes only and does not constitute Agency policy.
72 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Toxicological Review of Formaldehyde - Inhalation
Respiratory Pathology (Animal Exposure Duration >52 Weeks; Humans All Employed >5 Years)
Kerns etal. (1983):
Battelle (1982)
Squamous metaplasia: nasal
turbinates, B6C3F1 mice
No
Mice are far less susceptible for this
endpoint
Woutersen et al. (1989)
Squamous metaplasia: nasal
turbinates, Wistar rats
Yes
Appelman et al. (1988)
Squamous metaplasia: nasal
turbinates, Wistar rats
No
Limited sample size (n = 10/group) and
exposure duration (1 year), as compared
to Kerns et al. (1983: 1982) (n = up to
~100/group; 24 months) and Woutersen
et al. (1989) (n = 30/group; 28 months)
Kamata et al. (1997)
Squamous metaplasia: nose and
trachea, Fisher 344 rats
No
Some quantitative uncertainty
associated with methanol coexposure;
small sample size at 28 months;
metaplasia results pooled across
scheduled sacrifices
Derivation of PODs
There was high confidence in both studies selected for POD derivation, as both studies were
well designed and executed with adequate reporting of data (notably, Kerns et al. (1983; 1982) was
conducted under GLP conditions). Table 28 summarizes the derivation of PODs using data from these
studies. In determining the BMR level for the POD from Kerns et al. (1983; 1982). the average severity
score was in the range of minimal-to-mild at the lowest dose for both the 18-month and 24-month
durations for Level 1. This finding supports a BMR of 0.1 extra risk, representing a minimal level of
adversity. Due to difficulties modeling the 24-month data, the 18-month data, for which incidence rises
more gradually, were chosen even though these data would be less preferred (see Toxicological Review
Section 2.2.1). Interspecies extrapolation of the rat BMCL level to humans was carried out in two steps.
First, average flux values in the Level 1 region of the rat corresponding to the rat BMCL derived from the
incidence of squamous metaplasia were estimated. Next, the exposure concentration at which any
region in the human nose is exposed to this same level of formaldehyde flux at the inspiratory rate of 15
L/min was estimated.
For the POD from Woutersen et al. (1989) the same minimal adversity was assumed and a BMR
of 0.10 extra risk was used; however, a dosimetry model for flux to the nasal lining of the Wistar rat was
not available. U.S. EPA (2012) concluded that internal dose equivalency in the extrathoracic region for
rats and humans is in general achieved through similar external exposure concentrations; that is, even
for highly soluble and reactive gases ppm equivalence is a more appropriate default method for
extrapolation than an approach based on adjustment by the ratio of surface area to minute volume.
Confidence in the POD calculation based on Woutersen et al. (1989) was medium, while
confidence based on Kerns et al. (1983); Battelle (1982) was low. Confidence is lower in the latter due
to extrapolation well below the tested formaldehyde concentrations, a BMCL was based on the 18-
month exposure although the response was greater in magnitude after 24 months, and modeling of the
incidence at Level 1 in the nose, although concentrations in Level 2 were lower.
This document is a draft for review purposes only and does not constitute Agency policy.
73 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Toxicological Review of Formaldehyde - Inhalation
Table 24. Summary of derivation of PODs for squamous metaplasia
Endpoint and
Reference
Model
BMR
Rat BMC
(mg/m3)
Rat BMCLa
(mg/m3)
Flux3
(pmol/mm2-h)
Human
Exposure
(mg/m3)
Human PODbADJ
(mg/m3)
Squamous metaplasia
Kerns etal. (1983):
Battelle (1982)F344 rat.
M & F, 18 months,
Level 1
Log-probit
0.10
0.587
0.456
685
0.484
0.086c
Squamous metaplasia
Woutersen et al.
(1989) Wistar Rats. M,
28 months, Level 1
Log-logistic
0.10b
1.00
0.526
N/A
N/A
0.094d
Approximate average flux over nasal lining at this level corresponding to the BMCL.
bPODADj is the human equivalent of the rat BMCL duration adjusted (6/24) x (5/7) for continuous daily exposure.
cHuman extrapolation was based on modeled estimates of regional formaldehyde tissue flux. If extrapolation is based on ppm
equivalence instead, value increases by 1.14-fold.
dHuman extrapolation was based on ppm equivalence derived from pharmacokinetic principles.
Derivation of cRfC
Table 29 describes the uncertainty factors used to adjust the POD to the resulting cRfCs for each
of the two selected studies. For both PODs, a UFA of 3 was applied to address residual uncertainties in
interspecies extrapolation after dosimetry modeling (Kerns et al.. 1983; Battelle. 1982) or an assumption
of ppm equivalence (Woutersen et al.. 1989) was used to estimate a human equivalent concentration
and account for toxicokinetic differences between animals and humans. A UFH of 10 was applied to
both PODs to address the limited variability in susceptibility factors encompassed by these typical
studies of inbred laboratory animal populations. Finally, a UFS of 3 was applied for the Kerns et al.
(1983); Battelle (1982) study because it was based on 18-month exposure data in lieu of the 24-month
exposure data available in the same study. Specifically, the lesion incidence data were higher with
longer exposure duration (i.e., 24 months versus 18 months), and thus a lower POD would be expected
if the 24-month data could have been modeled. Although the 18-month exposure duration reduced the
uncertainty associated with extrapolating to lifetime exposure compared with a shorter duration, this
reduction was considered incomplete, and a factor of 3 was applied.
This document is a draft for review purposes only and does not constitute Agency policy.
74 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Toxicological Review of Formaldehyde - Inhalation
Table 25. Derivation of cRfCs for respiratory tract pathology
Endpoint (Reference; Population)
POD
POD basis
UFa
UFh
ufl
UFs
UFd
UFcomposite
cRfC (mg/m3)
RESPIRATORY TRACT PATHOLOGY
Sauamous metaolasia: [Kerns et al. (1983):
Battelle (1982) F344 rat. M & F. 18
months, Level 1]
0.088
BMCLio
3
10
1
3
1
100
0.0009
Sauamous metaplasia: [Woutersen et al.
(1989) Wistar Rat. M. 28 months. Level 11
0.094
BMCL10
3
10
1
1
1
30
0.003
Selection of the osRfC
The osRfC for respiratory tract pathology is based on squamous metaplasia observed in anterior
rodent nasal passages in two studies of long-term exposure. EPA could discern no particular basis to
select either the Woutersen et al. (1989) study or Kerns et al. (1983); Battelle (1982) study over the
other on grounds of confidence in the study methods, or known differences in sensitivity between
Wistar and F344 rats. In addition, the PODs were nearly identical and the cRfCs were very similar for the
two datasets (i.e., cRfCs of 0.0009 for Kerns et al. (Kerns et al.. 1983; Battelle. 1982) and 0.003 for
Woutersen et al. (1989) which are comparable given the limited precision of the calculations). However,
there was lower confidence in the derivation of the POD from (Kerns et al.. 1983; Battelle. 1982), which
involved an extrapolation well below the tested formaldehyde concentrations. In addition, the cRfC for
(Kerns et al.. 1983; Battelle. 1982) involved the application of an uncertainty factor for exposure
duration. While exposure duration is important to the development of this lesion, such effects appear
to be more dependent on exposure concentration. Thus, if a factor describing the concentration-
duration relationship5 were available for formaldehyde (and interpretable in the context of metaplasia),
a data-defined UFs could have been applied. Considering these uncertainties and the comparability of
the cRfCs, to represent the results of both studies, the cRfC from Woutersen et al. (1989) was used to
derive an osRfC of 0.003 mg/m3 for the respiratory pathology endpoint. Since the POD basis for this
value is from Woutersen et al. (1989) the confidence in the POD is considered medium. Completeness
of the database for respiratory tract pathology is high, based primarily on numerous well-conducted
long-term studies in experimental animals.
3.5. NERVOUS SYSTEM EFFECTS
Numerous studies reported data suggesting that formaldehyde inhalation might result in
noncancer nervous system effects; however, few studies in humans were available and the animal data
were often compromised by significant methodological limitations. In addition, there was generally
5Studies of other irritants have, on average, identified a factor of ~1.8-1.9 for relationships between acute exposure and
mortality (i.e., the observed mortality is more attributable to concentration, by 1.8- to 1.9-fold, than duration; see Toxicological
Review Section 1.2.4). A value for formaldehyde was not identified, nor were values for long-term exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
75 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
weak consistency in the evidence across well-conducted studies, a potential mode-of-action for nervous
system effects without systemic distribution of inhaled formaldehyde has not been established, and the
database is considered incomplete. Overall, conclusive evidence of a nervous system health hazard in
humans exposed to formaldehyde was not identified (i.e., suggestive evidence). Thus, this Overview
provides only a brief synopsis. However, given the potential for nervous system effects reported across
a variety of study types, and the general lack of comprehensive and rigorous experiments, a clear need
for additional studies, particularly well-conducted studies relevant to childhood exposure, was
identified.
3.5.1. Literature Identification and Study Evaluation
Literature identification and study evaluations were conducted in a manner similar to the other
noncancer health effect sections (see Appendix A.5.7 and Appendix F for details). The study evaluations
emphasized an analysis of potential issues relating to exposure (e.g., for these systemic effects, known
or presumed coexposure to methanol represented a serious study deficiency) and the irritant effects of
formaldehyde.
3.5.2. Evidence Synthesis and Overall Evidence Integration Judgement for Nervous System Effects
Data were available and analyzed relating to the following outcomes:
• Amyotrophic lateral sclerosis (ALS): several medium and high confidence observational
epidemiological studies were available, generally without quantified exposure levels and with
outstanding questions of consistency.
• Developmental neurotoxicity: the evidence primarily consisted of a medium confidence animal
study and some studies reporting potentially relevant mechanistic findings. Given the potential
for effects in children exposed to formaldehyde, this represents a notable data gap.
• Neural sensitization (i.e., an exposure-induced increased responsiveness of the nervous system
to other stimuli): several animal studies were available, with questions of human relevance.
• Motor-related behaviors: numerous human and animal studies of low confidence and some
studies reporting potentially relevant mechanistic findings were available.
• Learning and memory: numerous human and animal studies of low confidence and some studies
reporting potentially relevant mechanistic findings.
Among these outcomes, the studies of ALS are of particular note. An association between
formaldehyde exposure and ALS was suggested across four studies in the United States, Sweden and
Denmark by two separate groups of researchers (Peters et al.. 2017; Seals et al.. 2017; Roberts et a I..
2016; Weisskopf et al.. 2009). Positive associations observed in a large prospective study (Weisskopf et
al.. 2009) were somewhat corroborated by a few (but not most) comparisons in the other studies,
noting that some associations were based on a very small number of cases or secondary analyses.
However, two of the studies had uncertainties in the assignment of individual exposure to formaldehyde
(Roberts et al.. 2016; Fang et al.. 2009), and the third did not observe a dose-response relationship when
the data were stratified by estimated formaldehyde levels (Peters et al.. 2017). The observed
This document is a draft for review purposes only and does not constitute Agency policy.
76 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
association reported by the study in Denmark was not corroborated by a second study that examined
joint effects by multiple health and chemical risk factors (Bellavia et al.. 2021). In addition, the results
were not verified in another study in a different population, which had greater certainty in individual
exposure assessments (Pinkerton et al.. 2013). Thus, the currently available human evidence was not
considered sufficient to identify a clear hazard. However, the unexpected nature of the observed
associations between formaldehyde exposure and this rare and fatal disease across a growing number
of studies (the first association was reported in 2009, with corroborating evidence in 2015 and 2016)
identifies an urgent need for additional research.
Overall, while a number of studies reporting evidence of potential neurotoxic effects were
available, due to limitations identified in the database (e.g., poor methodology, lack of consistency), the
integration of the evidence ultimately resulted in a determination that the evidence suggests but is not
sufficient to infer that formaldehyde inhalation might cause multiple manifestations of nervous system
health effects in humans given relevant exposure circumstances. The data were considered insufficient
for developing quantitative estimates of risk.
3.6. REPRODUCTIVE AND DEVELOPMENTAL TOXICITY
Studies in humans, and a number of animal studies have analyzed effects of inhaled
formaldehyde on pre- and post-natal development and on the female and male reproductive systems.
The health effects studies of human exposure included studies of residential exposure during pregnancy
and fetal and infant growth measures, as well as occupational epidemiological studies conducted in
different industries and countries that evaluated decreased fecundity,6 spontaneous abortion, and
adverse birth outcomes associated with formaldehyde exposure among men and women. A few studies
also analyzed sperm quality parameters. Exposure levels in the occupational settings were high (>0.1
mg/m3) with intermittent peaks depending on specific uses. Animal studies investigated manifestations
of developmental toxicity (i.e., decreased survival, decreased growth, or increased evidence of structural
anomalies), female reproductive toxicity (ovarian and uterine pathology, ovarian weight, and hormonal
changes), and effects on the male reproductive system. However, all of the available medium and high
confidence studies exposed animals to high formaldehyde concentrations (>5 mg/m3), and exposure
protocols for the remaining studies were limited (i.e., the use of formalin, or an uncharacterized test
substance). This review assesses health effects of exposure for females and males separately.
3.6.1. Literature Identification
The literature searches focused on reproductive and developmental outcomes in
epidemiological studies and animal studies, as well as mechanistic studies. The bibliographic databases,
search terms, and specific strategies used to search them are provided in Appendix A.5.8, as are the
6The capacity to conceive and deliver a baby.
This document is a draft for review purposes only and does not constitute Agency policy.
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
Toxicological Review of Formaldehyde - Inhalation
specific PECO criteria and the methods for identifying literature from 2016 - 2021 are described in
Appendix F.
3.6.2. Study Evaluation
The epidemiological analyses that assigned individual-level exposures based on formaldehyde-
specific quantitative information, such as formaldehyde measurements or reported frequency of
product use, were considered to have greater accuracy than studies that defined participants as
exposed or nonexposed. Several studies classified individuals based on work processes, an informed
source, or occupation/industry codes from census data; there was less certainty about whether these
exposure classifications successfully distinguished high exposure from low or no exposure. Exposure
misclassification and the inclusion of individuals with probable low or infrequent exposure as exposed
reduced the sensitivity of analyses; these analyses were considered to be of low confidence.
For studies in experimental animals, a key consideration for the interpretation of developmental
and reproductive outcomes associated with inhalation exposures to formaldehyde was the potential for
coexposure to methanol, a known developmental and reproductive toxicant, when the test article was
an aqueous solution of formaldehyde. Studies that used formalin but did not control for methanol and
studies that did not characterize the formaldehyde source were assigned a low confidence rating and
contributed little to the synthesis of evidence regarding formaldehyde effects on development or the
reproductive system.
3.6.3. Developmental and Female Reproductive Toxicity
Synthesis of human health effect studies
The observational studies of reproductive toxicity or pregnancy outcomes evaluated
associations with exposure during pregnancy in three studies and with occupational exposure among
cosmetologists, woodworkers, laboratory workers, and hospital staff. The evidence regarding
fecundability7 (e.g. time to pregnancy or TTP), spontaneous abortion, pre- and post-natal growth and
other birth outcomes, and male reproductive toxicity was synthesized Time-to-pregnancy is a measure
of fertility and has been characterized in terms of number of menstrual cycles to the recognition of
pregnancy.8 Increased TTP reflects potential effects on gametogenesis, transport, fertilization,
migration, implantation, or survival of the embryo (Baird et al.. 1986). Thus, the measure encompasses
both developmental and reproductive toxicity, reflects an impact on multiple biological processes in
both partners, and is sensitive to the detection of early events before a pregnancy is clinically
recognized. One medium confidence retrospective cohort study evaluated effects on TTP in relation to
maternal occupational exposure to formaldehyde (Taskinen et al.. 1999). The fecundability density ratio
7 A couple's probability of conception in one menstrual cycle.
sTime-to-pregnancy of greater than 12 months of unprotected intercourse is indicative of reduced fertility. Time-to-pregnancy
is not a measure of infertility, as these studies only include women who became pregnant and had a live birth.
This document is a draft for review purposes only and does not constitute Agency policy.
78 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
(FDR) for individuals in the highest formaldehyde exposure category (mean 8-hour TWA exposure of
0.27 mg/m3) compared to nonexposed individuals was 0.57 (95% CI: 0.37, 0.85) in a model that adjusted
for potential confounders and phenol exposure. Other coexposures in the workplace were ruled out as
potential confounders. An ancillary analysis suggested that dermal exposure may have contributed to
risk of increased TTP in this cohort; this is an uncertainty with regard to the TWA concentrations
associated with this outcome.
Two medium confidence studies provided evidence that formaldehyde exposure to female
workers is associated with an increased risk of spontaneous abortion (Taskinen et al.. 1999; John et al..
1994). Of the six studies included in this review, three were determined to be low confidence, primarily
because of concerns about exposure misclassification, with probable decreased study sensitivity (Steele
and Wilkins, 1996; Hemminki et al.. 1985; Hemminki et al.. 1982). A fourth low confidence study
evaluated dose-response patterns, an important consideration for the synthesis of formaldehyde
associations, and, despite potential confounding by another exposure, found associations similar in
magnitude to the medium confidence studies (Taskinen et al.. 1994).
These studies examined diverse occupational groups exposed to different combinations of
chemical exposures and products containing formaldehyde (wood working, cosmetology, research
laboratories). Relatively high odds ratios were associated with formaldehyde exposure; odds ratios
reported by the medium confidence studies were 2.1 (95% CI: 1.0, 4.3) and 3.2 (95% CI: 1.2, 8.3)
(Taskinen et al.. 1999; John et al.. 1994). The studies addressed potential confounders, including other
workplace exposures, and found that formaldehyde was independently associated with spontaneous
abortion. Studies of hospital, nursing, or medical employees generally did not report an association,
although these low confidence studies tended to use less precise exposure assessment methods,
reducing their sensitivity.
The epidemiology literature is limited regarding formaldehyde exposure and birth outcomes.
One medium confidence birth cohort study reported decreases in birth weight and head circumference,
respectively, with each 1 ng/m3 unit increase in formaldehyde concentration measured in the mother's
homes at 34 weeks gestation (Franklin et al.. 2019). Gestational age was not associated with exposure.
The median concentration in the homes was 0.0028 mg/m3 and 23.3% of samples were below the LOD
in this relatively small study. Another medium confidence pregnancy cohort study in South Korea
observed lower birth weights associated with increasing formaldehyde concentration measured at mid
to late pregnancy (mean concentrations were 0.08 mg/m3), although there was evidence of confounding
in the positive direction by volatile organic compounds (Chang et al.. 2017). An elevated association
with congenital malformations and maternal exposure was reported by a set of low confidence studies
among female hospital or laboratory workers (Zhu et al.. 2006; Saurel-Cubizolles et al.. 1994; Stucker et
al.. 1990: Hemminki et al.. 1985: Ericson et al.. 1984). although the precision of the odds ratios was low
(wide confidence intervals overlapping 1.0).
This document is a draft for review purposes only and does not constitute Agency policy.
79 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Toxicological Review of Formaldehyde - Inhalation
Synthesis of animal health effect studies
Several studies in experimental animals evaluated developmental toxicity (survival, growth, and
morphological alterations), and a few evaluated reproductive toxicity in females, however they all were
weak (low confidence) studies with methodological limitations. Notably, for most of the studies, lack of
information about the test substance or the described use of formalin, with known or presumed
methanol coexposures limited interpretation of their results. Effects on fetal survival, pre- or post-natal
growth, or morphological alterations were observed in several studies and sometimes more than one
rodent species, and maternal toxicity did not appear to be a confounding influence. However,
inconsistencies in response were also observed, and clear dose-response relationships were not
discernable. The studies tested concentrations ranging from 0.5 to 49 mg/mg3.
Mode of action information
No experimentally established MOA exists, and any potential mechanisms have not been well-
studied for any effects on development or the female reproductive system. However, evidence of
elevated oxidative stress in the blood of occupationally exposed adults might provide a potential
indirect linkage (Bono et al.. 2010). Evidence of elevated oxidative stress and hormonal alterations in
the blood and other tissues of adult rodents also might provide indirect evidence, as it is recognized that
both oxidative stress and the hypothalamic-pituitary-gonadal (HPG) and hypothalamic-pituitary-adrenal
(HPA) axes have potential roles in developmental toxicity as well as female reproductive function (Sari et
al.. 2004; Sorg et al.. 2001; Kitaev et al.. 1984).
Overall evidence integration judgments and susceptibility for developmental or female reproductive
toxicity
Overall, the evidence indicates that inhalation of formaldehyde likely causes increased risk of
developmental or female reproductive toxicity in humans given appropriate exposure circumstances.
This conclusion is based on moderate evidence in observational studies finding increases in TTP and
spontaneous abortion risk among occupationally exposed women; the evidence in animals is
indeterminate, and a plausible, experimentally verified MOA explaining such effects without systemic
distribution of formaldehyde is lacking (see Table 31). The primary basis for this conclusion is from
studies of women with occupational exposures involving periodic peaks.
Table 26. Evidence integration summary for effects of formaldehyde inhalation on
developmental or female reproductive toxicity in humans
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence
Integration
Judgment
Moderate, based on:
Human health effect studies:
Two medium confidence studies in
two independent populations
(woodworkers, cosmetologists):
Indeterminate for developmental
toxicity, based on:
Animal health effect studies:
Mixed findings for evidence of
decreased fetal survival (pre- or post-
Relevance to humans:
Relevant health
effects observed in
humans are the
The evidence
indicates that
inhalation of
formaldehyde
likely causes
This document is a draft for review purposes only and does not constitute Agency policy.
80 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Toxicological Review of Formaldehyde - Inhalation
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence
Integration
Judgment
decreased fecundability and
increased spontaneous abortion risk.
Supporting evidence of association
with spontaneous abortion from one
low confidence study among
laboratory workers. All studies
evaluated multiple exposure
categories with highest risk at highest
exposure level. Null evidence from 5
low confidence studies with low
sensitivity.
Two medium confidence studies of
pregnancy cohorts indicating
decreased birth weight and head
circumference.
Two low confidence studies of
maternal exposure among health
workers with low precision showing
small increased risk of nonspecific
malformations
Biological Plausibility: No direct
evidence. However, evidence of
elevated oxidative stress in the blood
of exposed adults might provide a
potential indirect linkage (see
explanation at right)
implantation loss) and altered fetal
or post-natal growth across multiple
low confidence studies.
Mixed findings for evidence of
structural anomalies across multiple
low confidence studies
Biological Plausibility: No direct
evidence. However, evidence of
elevated oxidative stress and
hormonal alterations in the blood of
adult rodents might provide a
potential indirect linkage, as it is
recognized that both oxidative stress
and the hypothalamic-pituitary-
gonadal (HPG) axis may play a role in
developmental toxicity
Indeterminate for female
reproductive toxicity, based on:
Animal health effect studies:
Three low confidence studies in rats
and mice: decreased ovarian weight,
ovarian and uterine histopathology,
and hormonal alterations
Biological plausibility: Some evidence
of altered female reproductive
hormones suggests a possible role for
neuro-endocrine mediated pathways
primary basis for the
hazard determination.
MOA: No
experimentally
established MOA
exists, and any
potential mechanisms
have not been well
studied
Potential
susceptibilities: In the
absence of a
mechanistic
understanding,
specific
susceptibilities are
unknown
increased risk of
developmental or
female
reproductive
toxicity in
humans given
appropriate
exposure
circumstances
Primarily based
on studies of
women with
occupational
exposures to
formaldehyde
concentrations as
high as
1.2 mg/m3
3.6.4. Male Reproductive Toxicity
Synthesis of human health effect studies
Two medium confidence studies from one research group reported associations with lower
sperm motility (total and progressive), delayed fertility, and spontaneous abortion (Wang et al.. 2015;
Wang et al.. 2012). A quantitative, individual-level exposure assessment was conducted; average
exposures in the workplace were 0.2-3 mg/m3. Progressive motility and total motility were inversely
associated with the formaldehyde exposure index, a cumulative measure of exposure based on a job
exposure matrix, and a strong association was observed in logistic models of below-normal values of
these motility measures (Wang et al.. 2015). For example, odds ratios of 2.58 (95% CI: 1.11, 5.97) and
3.41 (95% CI: 1.45, 7.92) were found for progressive motility less than 32% in the low and high exposure
groups, respectively, compared to the community-based referent group. TTP and spontaneous abortion
also were associated with paternal exposure to formaldehyde in this cohort (Wang et al.. 2012). Two
low confidence studies with low sensitivity found no association (Lindbohm et al.. 1991; Ward et al..
1984).
This document is a draft for review purposes only and does not constitute Agency policy.
81 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Toxicological Review of Formaldehyde - Inhalation
Synthesis of animal health effect studies
Fourteen studies in rodents assessed effects on the male reproductive system following
inhalation formaldehyde exposure, although 8 of the studies had substantial methodological limitations
and were categorized as low confidence. The six remaining medium or high confidence studies
(examining five cohorts of rats or mice) were conducted by three research teams and only tested high
formaldehyde concentrations (>5mg/m3). In all of these studies paraformaldehyde was administered to
the test animals and study methods provided adequate characterization of the exposure paradigm
(Vosoughi et al.. 2013; Vosoughi et al.. 2012); (Sapmaz et al.. 2018; Ozen et al.. 2005; Ozen et al.. 2002;
Sarsilmaz et al.. 1999). Their studies reported that formaldehyde inhalation resulted in adverse testes
and epididymides histopathological changes in mice (Vosoughi et al.. 2013) and rats (Sapmaz et al..
2018; Ozen et al.. 2005; Sarsilmaz et al.. 1999), and decreased sperm count, motility, and morphology in
mice (Vosoughi et al.. 2013). The decreases in sperm count (44-49%), sperm motility (40-46%) and
abnormal sperm morphology were observed at 35 days posttreatment involving concentrations >12.2
mg/m3 to paraformaldehyde for 10 days (Vosoughi et al.. 2013). The delayed response suggests that the
effects may have resulted from a disruption of spermatogenesis. Decreases in serum testosterone in
mice (32-49% at 24 hours postexposure) and rats (6-9% with 91 days exposure) also were observed
with exposure levels ranging from 6-25 mg/m3 (Vosoughi et al.. 2013; Ozen et al.. 2005). a response that
is biologically consistent with the Leydig cell pathology also associated with these exposure levels
(Vosoughi et al.. 2013; Sarsilmaz et al.. 1999). Results from the low confidence studies were largely
consistent {Han, 2015, 2453275},"(Zhou et al.. 2011a),"(Zhou et al.. 2011b),"(Zhou et al.. 2006); (Golalipour
et al.. 2007);(Appelman et al.. 1988);(Maronpot et al.. 1986);(Xing et al.. 2007). Since the available
studies only tested very high formaldehyde levels (i.e., the lowest levels tested were often >12 mg/m3,
with only a few studies testing 6.15 mg/m3 as the lowest exposure level), significant uncertainties
remain. Taken together, however, these studies provide coherent evidence of toxicity to the male
reproductive system spanning biochemical, cellular, tissue, and functional levels.
Mode of action information
No experimentally established MOA exists, and any potential mechanisms have not been well-
studied for any effects on the male reproductive system. However, mechanistic data provide some
support for indirect effects, including multiple biomarkers of oxidative stress, as well as heat shock
protein induction, that have been observed in the testes or epididymides of exposed rats in well
conducted studies (Sapmaz et al.. 2018; Zhou et al.. 2011b; Ozen et al.. 2008; Zhou et al.. 2006; Ozen et
al.. 2005; Ozen et al.. 2002). Heat shock protein (Hsp) immunoreactivity and oxidative stress resulting in
hypomethylated sperm (no studies were identified that evaluated sperm methylation changes) have
been linked to human male infertility (Werner et al.. 1997).
This document is a draft for review purposes only and does not constitute Agency policy.
82 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
Overall evidence integration judgments and susceptibility for male reproductive toxicity
Overall, the evidence indicates that inhalation of formaldehyde likely causes increased risk of
reproductive toxicity in men given appropriate exposure circumstances, based on robust evidence in
animals that presents a coherent array of adverse effects in two species, and slight evidence from
observational studies of occupational exposure levels, and no plausible, experimentally verified MOA
explaining such effects without systemic distribution of formaldehyde. However, some support for
indirect effects in rodents is provided by relevant mechanistic changes in male reproductive organs
(Table 32). The primary basis for this conclusion is based on bioassays in rodents testing formaldehyde
concentrations >6 mg/mg3.
Table 27. Evidence integration summary for effects of formaldehyde inhalation on
reproductive toxicity in males
Evidence
Additional
Integration
Human Evidence
Animal Evidence
Interpretations
Judgment
Slight, based on:
Robust, based on:
Relevance to humans:
The evidence
Human health effect studies:
Animal health effect studies:
Some uncertainty
indicates that
One medium confidence study of
• Four high or medium confidence
regarding the
inhalation of
exposure among male woodworkers:
studies in mice and rats reporting
relevance of the
formaldehyde
inverse association with sperm
dose-related qualitative or
animal evidence
likely causes
motility measures, increased
quantitative histopathological
exists, as the studies
increased risk of
prevalence of TTP, spontaneous
lesions of the testes or
only tested extremely
reproductive
abortion and birth defects
epididymides; consistent
high concentrations
toxicity in men
Null evidence for effects on sperm
observations from 5 of 6 low
expected to cause
given appropriate
counts and morphology in one low
confidence studies
strong irritant effects
exposure
confidence study with low precision
• One high confidence study in mice
that may not occur in
circumstances
Biological Plausibility: No directly
reporting dose-related effects on
humans; however,
relevant studies were identified
epididymal sperm; support from
given concordant
Primarily based
four low confidence studies in rats
findings in a well-
on bioassays in
• Two high confidence studies in
conducted study of
rats and mice
mice and rats reporting dose-
humans and the
testing
related decreased serum
absence of evidence
formaldehyde
testosterone (and decreased serum
to the contrary, the
concentrations
LH in one study), with support from
relevance to humans
above 6 mg/mg3
one low confidence study in rats
is presumed
(no medium or
• Mixed results for organ weight
MOA: No
high confidence
changes (i.e., testes, epididymis)
experimentally
studies tested
Biological Plausibility. Multiple
established MOA
lower exposure
levels)
biomarkers of oxidative stress, as
exists, and any
well as heat shock protein induction,
potential mechanisms
have been observed in the testes or
have not been well-
epididymides of exposed rats in well
studied; however,
conducted studies. Heat shock
mechanistic data
protein (Hsp) immuno-reactivity and
provide some support
oxidative stress resulting in
for indirect effects
hypomethylated sperm have been
Potential
linked to human male infertility
susceptibilities: No
specific data were
available to inform
potential differences
in susceptibility.
This document is a draft for review purposes only and does not constitute Agency policy.
83 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 3.6.5. Dose-response Analysis
2 Study selection
3 The dose-response analysis for developmental and female reproductive toxicity used data from
4 one medium confidence epidemiological study that assessed dose-response relationships for the
5 outcomes, TTP and spontaneous abortion, although the timing of exposure measurements had
6 uncertain relevance to responses during the pregnancies that ended in spontaneous abortion. For male
7 reproductive toxicity, two studies of rats exposed for 13 weeks, that assessed relatively sensitive
8 endpoints, were considered appropriate for the derivation of toxicity values (see Table 33 for study
9 selection rationales).
10 Table 28. Eligible studies for POD derivation and rationale for decisions to not select
11 specific analyses
Reference
Endpoint
POD
Derived?
Rationale for
Decisions to Not Select
Taskinen et al. (1999)
Time-to-pregnancy
Yes
Taskinen et al. (1999)
Spontaneous abortion
No
Uncertain temporal applicability of
exposure data for evaluating 1st
trimester effects
Franklin etal. (2019)
Birth weight, head circumference
No
Uncertainties in exposure distribution
due to large % < LOD and impact on
quantitative results
Chang et al. (2017)
Birth weight
No
Evidence of confounding by co-exposure;
Log transformed formaldehyde
concentration
Ozen et al. (2002)
Relative testes weight, 13-week
exposure
Yes
Ozen et al. (2005)
Serum testosterone, Wistar rat,
13-week exposure
Yes
Ozen et al. (2005)
Seminiferous tubule diameter,
Wistar rat, 13-week exposure
No
Analysis of pooled tissues;
interpretability to individual rats
uncertain
(2013): Vosoughi et al.
(2012)
Seminiferous tubule diameter,
NMRI mice, 10-day exposure
No
Short exposure duration
(2013): Vosoughi et al.
(2012)
Sperm abnormalities, NMRI mice,
10-day exposure
No
Short exposure duration
(2013): Vosoughi et al.
(2012)
Serum testosterone, NMRI mice,
10-day exposure
No
Short exposure duration
Vosoughi et al. (2013)
Testes weight, NMRI mice, 10-day
exposure
No
Short exposure duration
Sarsilmaz et al. (1999)
Leydig cell quantity or nuclear
damage, Wistar rat, 4-week
exposure
No
Short exposure duration
Sarsilmaz et al. (1999)
Testes weight (relative), Wistar rats,
4-week exposure
No
Short exposure duration; non-preferred
metric (absolute testes weight preferred)
Sapmaz et al. (2018)
Seminiferous tubule measures,
Sprague-Dawley rats, 4- and 13-
week exposure
No
Short exposure duration (for 4-week
experiment); single exposure level
This document is a draft for review purposes only and does not constitute Agency policy.
84 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Toxicological Review of Formaldehyde - Inhalation
Derivation of PODs
Developmental and female reproductive toxicity
Taskinen et al. (1999) presented fecundability density ratios (FDR) for increased time to
pregnancy for index pregnancies of women in three exposure categories for jobs held beginning at least
6 months prior to the index pregnancy. TTP was elevated in the high exposure group relative to the
unexposed group and the middle 8-hour TWA exposure level was selected as a NOAEL (Table 34).
The mean 8-hour TWA concentrations reported for each exposure category were adjusted for
likely background formaldehyde exposures experienced by the employees when they were not
conducting work tasks involving formaldehyde exposure. Normally, exposures from occupational
studies are adjusted to account for the daily breathing volume appropriate to an environmental (versus
occupational) setting and for exposure every day of the year (U.S. EPA. 1993). However, with
formaldehyde, there is potential for exposure outside of work from in-home and environmental sources
of formaldehyde. Therefore, the POD represents exposure during an 8-hour workday.
Table 29. Summary of derivation of PODs for developmental and reproductive toxicity
in females
Endpoint and Reference
Population
Observed Effects by Exposure Level
POD (mg/m3)
Time to pregnancy in females
Occupational prevalence
study Taskinen et al.
(1999)
Adult
women,
n = 602
Time to Pregnancy by Formaldehyde Category;
Fecundability density ratio (FDR)a
Mean 8-hr # FDRb 95% CI
TWA (mg/m3)
NOAEL = 0.106
LOAEL = 0.278
Not exposed 367 1.00
0.042 119 1.09 0.86-1.37
0.106 77 0.96 0.72-1.26
0.278 39 0.64 0.43-0.92
Fecundability density ratio = ratio of average incidence
densities of pregnancies in exposed compared to employed
unexposed women
Discrete proportional hazards regression; adjusted for
employment, smoking, alcohol consumption, irregular
menstrual cycles and # children
Comparison: index pregnancies that occurred when
participants were not employed in exposed workplace
Concentrations converted to mg/m3.
b8-hr TWA reported by authors were recalculated by EPA to account for background formaldehyde exposure while working in
"nonexposed" work areas.
Male reproductive toxicity
Both studies selected for candidate reference value derivation exposed the animals to
paraformaldehyde via inhalation (Ozen et al.. 2002) (Table 35). In Ozen et al. (2002). statistically
significant duration- and dose-dependent decreases in testis weight (relative to body weight) were
This document is a draft for review purposes only and does not constitute Agency policy.
85 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
observed after 4 and 13 weeks of formaldehyde exposure. Although absolute organ weights are
preferred for this measure, because testes weights are generally conserved when body weight is
decreased, mean body weights were also significantly decreased with exposure; thus, this response
pattern suggests that the organ weight decreases were likely due to a direct effect on the testis (note: in
this case, decreased relative testis weight is likely an underestimate of the more appropriate decrease in
absolute testis weight). For the decreased testis weight at week 13 (Ozen et al.. 2002), a LOAEL of 12.3
mg/m3 was adjusted for continuous exposure based upon the experimental paradigm to yield a PODadj
of 2.93 mg/m3 (POD adj — 12.3 mg/m3 x 8 hr exposed per day/24 hours per day x 5 days exposed per
week/7 days per week).
In Ozen et al. (2005), statistically significant dose-dependent decreases in serum testosterone
levels (6 to 9% decreases from control values) were observed following 91 days of inhalation exposure.
At the same exposure levels, significant decreases of 23 to 26% from control were noted in mean
seminiferous tubule diameters, an effect that could have been directly related to testosterone
decreases. A BMCLisd of 0.208 mg/m3 was calculated. U.S. EPA (2012) indicates that for highly soluble
and reactive gases that interact with tissue at the point of entry or for gases with systemic penetration
ppm equivalence is an appropriate default method for extrapolation.
Table 30. Summary of derivation of PODs for reproductive toxicity in males
Endpoint and Reference
Species/
Sex
Model
BMR
(mg/m3)
BMC
(mg/m3)
BMCL
(mg/m3)
PODADJa
(mg/m3)
Ozen et al. (2005)
Decreased relative
testes weight (13 wk)
Rat/M
LOAEL
N/A
N/A
N/A
2.93
Ozen et al. (2005)
Decreased serum
testosterone (13 wk)
Rat/M
Exponential
(M2)
1SD
0.284
0.208
0.050
aPODADj is the human equivalent of the rat BMCL duration adjusted (6/24) x (5/7) for continuous daily exposure.
Derivation ofcRfCs
A UFh of 10 was applied to the developmental toxicity POD based on reduced fecundity in
reproductive age women in an occupational cohort studied by Taskinen et al. (1999) to account for
variation in the broader human population not represented by occupationally exposed groups. No other
adjustments were made to this cRfC (Table 36).
For interspecies uncertainty for results in the animal studies, an assumption of ppm equivalence
(which is derived from pharmacokinetic principles) (Ozen et al.. 2005; Ozen et al.. 2002), male
reproductive toxicity was used to estimate a human equivalent concentration. Then a UFA of 3 was
applied to account for residual uncertainties in interspecies extrapolation from the two cRfCs for
reproductive toxicity in males derived from rat studies. A UFS of 10 was applied to both PODs to
approximate the potential effect of lifetime exposure, as these effects are not necessarily dependent on
This document is a draft for review purposes only and does not constitute Agency policy.
86 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Toxicological Review of Formaldehyde - Inhalation
a specific exposure window and they are expected to worsen with continued exposure. In addition, a
UFl of 10 was applied to the POD for relative testis weight, which was based on a LOAEL (Ozen et al..
2002). Finally, a UFH of 10 was applied to both PODs to account for the limited variability in
susceptibility factors encompassed by these typical studies of inbred laboratory animal populations.
Table 31. Derivation of cRfCs for Female Reproductive or Developmental Toxicity and
Male Reproductive Toxicity
Endpoint (Reference; Population)
POD
POD basis
UFa
UFh
UFl
UFS
UFd
UFcomposite
cRfC (mg/m")
FEMALE REPRODUCTIVE OR DEVELOPMENTAL TOXICITY
Delayed pregnancy Taskinen et al. (1999):
pregnant F, n = 77 at POD)
0.106
NOAEL
1
10
1
1
1
10
0.01
MALE REPRODUCTIVE TOXICITY
Relative testes weight Ozen et al. (2005):
adult rat M, 13-week exposure)
2.93
LOAEL
3
10
10
10
1
3000
0.001
Serum testosterone Ozen et al. (2005):
adult rat M, 13-week exposure)
0.05
BMCLisd
3
10
1
10
1
300
0.0002
Derivation of the osRfC
The cRfC for effects on delayed pregnancy (Taskinen et al.. 1999) was chosen as the osRfC.
Although TTP is a sensitive measure of effects on the reproductive system, confidence in the POD is
judged to be low because the outcome was evaluated in a healthy working population with relatively
high exposure, which raises uncertainty about its applicability to more diverse populations. More
complete assessments of developmental endpoints by epidemiology or toxicology studies were not
available. Thus, the completeness of the database is considered low. As a mechanistic understanding is
lacking, the relevant time period for exposure effects on TTP through unrecognized fetal losses or
factors controlling the ability to conceive could range from the weeks just prior and after conception, to
the entire period of prior exposure during the life of the individual. Thus, the osRfC is 0.01 mg/m3.
The cRfC derived from Ozen et al. (2002) was considered the stronger of the two candidates for
male reproductive toxicity, and thus was chosen to represent the osRfC. The magnitude of the testes
weight response in Ozen et al. (2002) was greater than the testosterone decreases observed in Ozen et
al. (2005). and a number of other rodent studies in the formaldehyde database demonstrated similar
testis (and epididymal) weight deficits, while specific evidence of treatment-related serum testosterone
decreases was quite limited. The osRfC is 0.001 mg/m3 using the cRFC from Ozen et al. (2002). The
confidence in the POD derived from its results is low, given that the lowest formaldehyde concentration
tested in this study was 12 mg/m3. Confidence in the database is also considered low because, while a
number of published studies evaluated reproductive toxicity in males, the interpretation of study results
is complicated by their methodological limitations and exclusive use of formaldehyde concentrations
above 6 mg/m3, and data are lacking regarding functional endpoints.
This document is a draft for review purposes only and does not constitute Agency policy.
87 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 3.7. REFERENCE CONCENTRATION (Rfc) FOR NONCANCER HEALTH EFFECTS
2 3.7.1. Summary of cRfCs and osRfCs across Noncancer Health Effects
3 The RfC was chosen to reflect an estimate of continuous inhalation exposure to the human
4 population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious
5 effects during a lifetime. The RfC was determined from the group of osRfCs, which in turn were selected
6 from the cRfCs in each health effect system. Figure 13 presents the cRfCs derived for each health effect
7 system, the points of departure from each study, and the uncertainty factors that were applied to them.
8 As summarized in Figure 13 and Table 37, the osRfCs for each health effect system were either selected
9 from among the cRfCs or the values were combined. The rationales for osRfC selection were described
10 previously in the hazard evaluations for each health effect system.
This document is a draft for review purposes only and does not constitute Agency policy.
88 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
0 >. =
p CTi
— n £
3 C C
m c
13 .2
ce
£? §
_Of u
<
E
XT
<
E u
ai ~
U_ CC X3
Eye irritation symptoms (humans)
Hanrahan et al., 1984
Eye irritation symptoms (humans)
Kulle et al, 1987
Eye irritation symptoms (humans)
Andersen, 1983
Peak expiratory flow rate (humans)
Krzyzanowski et al,, 1990
Rhinoconjunctivitis prevalence (children)
Annesi-Maesana et al., 2012
Atopic excema prevalence (preganant women)
Matsunaga et al., 2008
Asthma control (children with asthma)
Venn et al., 2003
Current asthma prevalence (children)
Annesi-Maesano et al., 2012
Current asthma prevalence (children)
Krzyzanowski et al., 1990
Squamous metaplasia (male Wistar rats)
Woutersen et al., 1989
Squamous metaplasia (F344 rats of both sexes)
Kerns et al., 1983
Delayed pregnancy (pregnant women)
(Taskinen et al., 1999)
¦0 Relative testes weight (male Wistar rats)
-o £• (Ozen et al., 2002)
2 u
Q. X
cu o
Serum testosterone (male Wistar rats)
"5 (Ozen et al., 2005}
Composite UF
A Candidate RfC
• POD(HEC)
J
~ •
~
A
0.0010 0.0100 0.1000
log formaldehyde concentration (mg/ m3)
Figure 13. Candidate RfCs (cRfCs) with corresponding POD and composite UF.
Note: as PODs reflect exact values, and cRfCs are rounded to 1 significant figure, the extrapolation is not
exact.
This document is a draft for review purposes only and does not constitute Agency policy.
89 DRAFT-DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 32. Organ/System-specific RfCs (osRfCs) for formaldehyde inhalation
Health Effect
Basis
Reference(s) [Species]
UFC
osRfC
(mg/m3)
Integrated
hazard
judgment
Confidence in
POD
Estimate(s)a
Database
Completeness*1
Sensory Irritation
Hanrahan et al. (1984) [humanl
10
0.009
evidence
demontrates
medium
high
Pulmonary Function
Krzyzanowski et al. (1990)
[human]
3
0.007
Evidence
indicates
(likely)
high
high
Allergy-related
Conditions
Annesi-Maesano et al. (2012)
[human]
3
0.008
Evidence
indicates
(likely)
high
high
Asthma (prevalence of
current asthma/degree
of asthma control)
Annesi-Maesano et al. (2012);
Venn et al. (2003); Krzyzanowski
et al. (1990) [humanl
10c
0.006
Evidence
indicates
(likely)
medium
medium
Respiratory Pathology
Kerns et al. (1983); Battelle
(1982); Woutersen et al. (1989)
[rat]
30c
0.003
evidence
demontrates
medium
high
Female Developmental
Toxicity
Taskinen et al. (1999) [humanl
10
0.01
Evidence
indicates
(likely)
low
low
Male Reproductive
Toxicity
Ozen et al. (2002) [rati
3000
0.001
Evidence
indicates
(likely)
low
low
This table presents the osRfCs, the studies and uncertainty factors used to derive them, and the level of confidence in the
evidence integration, the PODs, and the completeness of the database.
aThis reflects a judgment regarding how well the study-specific data are able to estimate a no-effect or minimal-effect level of
response (e.g., a lower level of confidence would be applied to high concentration studies which required extrapolation far
below the lowest tested concentration to estimate a POD). A low confidence level means that the POD derived is expected to
be less accurate.
bAlthough no UFD was applied to any cRfC, it is recognized that the evidence databases for the various health effects are not
equal. This level of confidence was added to emphasize the health areas where additional research could reduce existing
uncertainties. A low confidence level means the degree of certainty regarding the RfC is lower.
cThese two osRFCs are based on multiple studies and candidate values, sometimes with different UFCs applied. The UFC values
shown in this table and Figure 2-2 reflect the candidate values selected to represent each osRfC [i.e., the UFC applied to the
POD from Krzyzanowski et al. (1990) for asthma and from Woutersen et al., (1989) for respiratory pathology].
2 3.7.2. Selection of the RfC and Discussion of Confidence
3 Choice of the RfC involved consideration of both the level of certainty in the estimated osRfCs,
4 as well as the level of certainty in the observed health effect(s). Thus, the collection of studies and
5 results used to characterize the hazard(s) and derive the osRfCs, as well as the cRfC calculations
6 themselves (including derivation of the PODs and the application of UFs), were considered when
7 choosing the RfC. These considerations are illustrated separately in Table 37, and as a composite
8 depiction of certainty in Figure 14. Based on this analysis, an RfC for formaldehyde of 0.007 mg/m3 was
9 selected. This value is within the narrow range (0.006-0.009 mg/m3) of the group of respiratory system-
10 related osRfCs derived from PODs that are the lowest of those identified in human population studies
This document is a draft for review purposes only and does not constitute Agency policy.
90 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Toxicological Review of Formaldehyde - Inhalation
for formaldehyde hazards (i.e., sensory irritation, pulmonary function, allergy-related conditions and
current asthma prevalence or degree of control). These osRfCs are each interpreted with high or
medium confidence in the hazard conclusion and in the POD estimate, and very low composite
uncertainty factors were applied.
An overall confidence level of high, medium, or low is assigned to reflect the level of confidence
in the study(ies) and hazard(s) used to derive the RfC, the completeness of the database, and the RfC
itself, as described in Section 4.3.9.2 of EPA's Methods for Derivation of Inhalation Reference
Concentrations and Application of Inhalation Dosimetry (U.S. EPA. 1994). Overall confidence in the RfC is
high; the RfC is based on a spectrum of adverse effects reported in multiple well-conducted studies of
exposed humans. Most of the study populations were exposed to formaldehyde levels in a residential
or school setting, and some of the studies focused on sensitive individuals. Finally, the hazard
conclusions are supported by an extensive literature database.
osRfCs and RfC (formaldehyde mg/ m3)
i
i ¦
Pulmonary function (Krzyzanowski et al., 1990)
CD
u
~
Allergy-related Conditions (Annesi-Maesano et al., 2012)
CD
"D
•
Sensory Irritation (Hanrahan et al., 1984)
M—
c
o
o
Respiratory Tract Pathology (Kerns et al., 1983; Woutersen et al., 1989)
u
i_
QJ
~
Asthma (Venn et al., 2003; Annesi-Maesano et al., 2012; Kryzanowski et al., 1990)
_£=
CuO
•
Female Reproductive and/ or Developmental Toxicity (Taskinen et al., 1999)
X
o
Male Reproductive Toxicity (Ozen et al., 2002)
Figure 14. Organ or system-specific RfC (osRfC) scatterplot.
This document is a draft for review purposes only and does not constitute Agency policy.
91 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Toxicological Review of Formaldehyde - Inhalation
Organ/system RfCs (osRfCs) that are represented by larger shapes and that are closer to the top of the
graph are interpreted with higher confidence regarding the basis from which the value was derived (see
Table-2-11), and with less uncertainty (i.e., lower UFs were applied). Size of the shape represents
confidence in the study(ies) and health hazard (i.e., hazards with evidence demonstrates judgments are
larger than those with evidence indicates [likely] judgments), POD estimate(s) (for the purposes of this
graphic, confidence in the POD was given slightly greater weight than the others), and completeness of
the available evidence database for each health outcome: larger shapes indicate higher confidence; solid
shapes indicate studies in humans; hollow shapes indicate animal studies. For composite UF, if multiple
studies served as the basis for an osRfC, the composite UF associated with the candidate value selected to
represent the osRfC was used (see Table 2-11). The dashed line represents the proposed overall RfC of
0.007 mg/m3; the circled osRfCs indicate the cluster of effects selected as the basis for this value.
Table 33. Proposed RfC for formaldehyde-inhalation
Health Effect(s) Basis
RfC (mg/m3)
Overall Confidence
Sensory irritation, pulmonary function, allergy-related conditions,
and degree of asthma control/prevalence of current asthma in
humans3
0.007
High
aBased on the following studies: (Annesi-Maesano et al., 2012: Matsunaga et al., 2008: Venn et al., 2003: Krzvzanowski
et al., 1990; Hanrahan et al., 1984)
3.7.3. Basis and Interpretation of the RfC
The RfC is an estimate of exposure that is likely to be without an appreciable risk of adverse
health effects over a lifetime. As illustrated in Figure 15, the selected RfC is at the upper end of the
range of outdoor formaldehyde levels recorded in some locations, and it would be expected that levels
in indoor air would exceed this concentration in most situations. However, it is important to reiterate
that this level is interpreted to be without appreciable risk. It is also important to note that the RfC does
not provide information about the magnitude of the risk of respiratory-related effects that might occur
at different concentrations above the RfC (e.g., at 0.02 or 0.03 mg/m3). As illustrated in Figure 15, nearly
all the study-specific findings of effects (e.g., LOAELs, BMCs) were not observed until formaldehyde
levels were in the upper end of the range of average indoor air concentrations, with effects generally
being observed at or above ~35-40 ng/m3. As an example comparison, a fairly large study of 398 homes
in Los Angeles, CA, Houston, TX, and Elizabeth, NJ, between 1999 and 2001 reported formaldehyde
levels of 22 ± 7.1 ng/m3 (Weisel et al.. 2005). One study that contributed to the RfC derivation involved
an analysis of the degree of asthma control in children with current asthma, and the RfC is expected to
apply to this susceptible subgroup in the population. Although current asthma symptoms and allergic
conditions were not observed in studies of children with exposures less than the range of 0.02-0.05
mg/m3, at 0.021 mg/m3, a 10.5% decrease in peak expiratory flow rate among asthmatic children could
be estimated (the regression model included a term for asthma status), based on a model using results
of Krzyzanowski et al. (1990). Thus, attributes that increase susceptibility in individuals are expected to
play a role in increasing the advent of adverse responses to formaldehyde levels above the RfC (e.g.,
somewhere between 0.007 and 0.04 mg/m3).
This document is a draft for review purposes only and does not constitute Agency policy.
92 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
Although the RfC is designed to apply to exposures over a lifetime, the relevant window of
exposure for some of the effects observed in the contributing studies may be less than lifetime. Sensory
irritation is an immediate response to reactive compounds such as formaldehyde. The relevant window
of exposure for effects on asthma outcomes also is less than lifetime, although the time frame for the
control of asthma symptoms (i.e., a few weeks) is expected to differ from that for the prevalence of
current asthma symptoms or a decrease in pulmonary function (i.e., the past 12 months). In addition,
the relevant window of exposure for the osRfC for female reproductive or developmental outcomes is
from conception to the end of the pregnancy. Thus, while the RfC is a concentration associated with
minimal risk over a lifetime of exposure, a few of the hazards or outcomes supporting the RfC could be
relevant to a shorter exposure time frame. Such interpretations might be informed by the information
presented in Figure 13 (POD to cRfC calculations) and Figure 15 (below).
Outdoor Indoor Air (normal conditions) Indoor Air (atypical; e.g., some sealed mobile homes)
Hanrahan {1984);
Burning eyes
Krzyzanowski (1990);
PEFR measures
Annesi-Maesano {2012);
Rhinoconjunctivitis
Krzyzanowski {1990);
asthma prevalence
Annesi-Maesano (2012);
asthma prevalence
Venn {2003);
asthma control
Woutersen (1989);
squamous metaplasia
Kerns {1983);
squamous metaplasia
Taskinen {1999);
t ime-to-p reg nan cy
Ozen {2002);
testes weight
fih
H [>
a
o
5H
Eh
H-
2-
xh - 4-
- - -m
B-
"Q—G-
- -O
O
¦o
Effects continue
at higher levels
Effects not observed
until higher levels
-O
lb lis 20 40 60 SO 100 200 400 600 800 lobo
Formaldehyde Concentration (|ig/ m3)
2000 4000 6000
; RfC; H cRfC: No appreciable risk O NOAEL, POD, N/C: No adverse change (study) 0 LOAEL or above: adverse effect in study (size = effect magnitude) — - UFs
•0. PODali(: Negligible risk (adjusted study data) ~ BMCL POD: Negligible risk in study ¦ BMC: 5-10% change (study data) —POD adjustment BMCtoBMCL
Figure 15. Illustration of noncancer toxicity value estimations.
This figure provides a representation of the estimates from studies supporting the osRfCs, including a
summary of formaldehyde exposure data. Formaldehyde exposure estimates reflect approximates of the
range (boxes), medians or means (black vertical bars), and more commonly reported estimates
(gradations), based on the data discussed in Appendix A.1.2. Horizontal lines in the figure reflect the
extrapolation process for arriving at points of departure (PODs) and toxicity values (unfilled symbols) in
This document is a draft for review purposes only and does not constitute Agency policy.
93 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Toxicological Review of Formaldehyde - Inhalation
the context of the study-specific evidence for effects (filled symbols; effect magnitude estimated based on
study figures, tables, or reported regressions; see previous sections). Note: The x-axis is intentionally not
on a linear or log scale so as not to convey a false level of precision. Abbreviations: cRfC (candidate RfC);
N/LOAEL (no/lowest-observed-adverse-effect level); UFs (uncertainty factors); BMCL (benchmark
concentration, lower confidence bound).
3.7.4. Previous IRIS Assessment: Reference value
An inhalation RfC for formaldehyde has not previously been derived. In 1990, an oral RfD of
0.2 mg/kg-day was developed. This value was based on reduced weight gain and histopathology
(primarily of the gastrointestinal system) in Wistar rats during a two-year bioassay in which
formaldehyde was administered in the drinking water (Til et al.. 1989). A UFC of 100 was applied to the
NOAEL to account for inter- and intraspecies differences. This RfD was interpreted with medium
confidence, based on high confidence in the principal study and medium confidence in the database.
4. CARCINOGENICITY
Multiple review articles and meta-analyses have examined the epidemiological evidence
informing potential associations between formaldehyde and cancer endpoints (e.g., Checkoway et al..
2012; Bachand et al.. 2010; Zhang et al.. 2009; Bosetti et al.. 2008; Collins and Lineker, 2004; Collins et
al.. 2001; Oiaiarvi et al.. 2000; Collins et al.. 1997; Blair et al.. 1990). The vast majority of studies focused
on cancers of the upper respiratory tract (URT) and lymphohematopoietic (LHP) system. Other cancer
types studied include bladder, brain, colon, lung, pancreas, prostate, and skin. However, aside from
lung and brain cancer, few studies showed evidence of increased risks; a cursory review of the studies of
lung and brain cancer did not provide any indication of an association with formaldehyde exposure (see
Appendix A.5.9). Given the large number of studies available on URT and LHP cancers, other cancer
types were not systematically evaluated.
The occurrences of URT cancers in humans have been described and grouped according to the
International Classification of Disease (ICD) codes. The specific cancers of the URT that are commonly
reported are sinonasal cancers (nose and nasal sinuses), cancers of the pharynx (nasopharynx,
oropharynx, and hypopharynx), and laryngeal cancer. Rarely, cancers of the buccal cavity are reported,
but as this grouping includes lip, tongue, salivary glands, gums, and the floor of the mouth, which
combine cancers of potentially different etiology and cell origin, cancers of the buccal cavity were not
reviewed. Thus, the above groupings were used for literature identification and hazard analyses.
In human studies, the specific LHP cancers that were formally reviewed were Hodgkin
lymphoma, multiple myeloma, myeloid leukemia, and lymphatic leukemia. Non-Hodgkin lymphoma is a
non-specific grouping of dozens of different lymphomas and classification systems for specific subtypes
This document is a draft for review purposes only and does not constitute Agency policy.
94 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
that have changed over time, complicating the evidence synthesis for this cancer type. As a cursory
review of the available studies did not suggest an association between formaldehyde exposure and non-
Hodgkin lymphoma, this endpoint was not formally reviewed.
4.1. METHODS FOR IDENTIFYING AND EVALUATING STUDIES
4.1.1. Literature Identification
The primary focus of this review was whether exposure to inhaled formaldehyde is associated
with specific URT or LHP cancers in humans or, in separate searches (i.e., nasal and LHP cancer studies
were searched separately), in animals. The bibliographic databases, search terms, and specific
strategies used to search them are provided in Appendix A.5.9, as are the specific PECO criteria and the
methods for identifying literature from 2016 - 2021 are described in Appendix F.
4.1.2. Study Evaluation
Human studies
The epidemiological studies generally examined occupational exposure to formaldehyde either
in specific work settings (e.g., cohort studies) or in case-control studies. The overwhelming majority of
information bias in epidemiological studies of formaldehyde stems from the use of occupational records
to gauge exposures with some degree of exposure misclassification or exposure measurement error
considered to be commonplace. Thus, a primary consideration in the evaluation of these studies was
the ability of the exposure assessment to reliably distinguish between levels of exposure within the
study population, or between the study population and the referent population. A large variety of
occupations were included within the studies; some represent work settings with a high likelihood of
exposure to high levels of formaldehyde, and some represent work settings with variable exposures and
in which the proportion of people exposed is quite small. In the latter case, the potential effect of
formaldehyde would be "diluted" within the larger study population, limiting the sensitivity of the study.
EPA categorized the exposure assessment methods of the identified studies into four groups (A through
D), reflecting greater or lesser degree of reliability and sensitivity of the measures. Outcome-specific
associations based on Group A exposures were considered without appreciable information bias due to
exposure measurement error, while other groups were considered increasingly biased towards the null.
Studies with small case counts may have little statistical power to detect divergences from the
null but are not necessarily expected to be biased, and no study was excluded solely on the basis of case
counts as this methodology would exclude any study which saw no effect of exposure. Therefore,
cohort studies with extensive follow-up that reported outcome-specific results on a number of different
cancers, including very rare cancers such as nasopharyngeal cancer (NPC) and sinonasal cancer, were
evaluated even when few or even no cases were observed, if information on the expected number of
This document is a draft for review purposes only and does not constitute Agency policy.
95 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
cases in the study population was provided so that confidence intervals could be presented for the
effect estimate. Studies with five or fewer exposed cases were considered to have low confidence.
Other considerations included an evaluation of limitations in effect estimates that may have
been confounded by exposure to other substances in the workplace that were known risk factors for
URT or LHP cancers and were likely to have been highly correlated with formaldehyde, as well as strong
healthy worker effects, and other selection biases.
Animal studies
Studies of cancer development in experimental animals exposed for at least subchronic duration
(shorter exposure durations were not prioritized for review, given the robust database), and which
performed histopathological evaluations of respiratory tract or hematopoietic tissues, were evaluated
(with preference given to studies that included a reasonable latency for cancers to develop, such as
conducting histopathological evaluations at >1 year of age). As these evaluations consider many of the
same studies previously evaluated for inclusion in the noncancer respiratory tract pathology section,
many parallels exist between both sets of evaluations, although several notable differences exist. For
example, duration of exposure was more important for evaluations of dysplasia and neoplasms, as
compared with evaluations of noncancer respiratory tract lesions. In addition, whereas a substantial
emphasis was placed on the characterization of the severity of the lesion for noncancer respiratory tract
changes, severity was not considered integral to the identification of cancers and dysplasia. Generally,
the study authors did not provide statistical comparisons for reported respiratory tract tumors; given
the rarity of these neoplasms in unexposed animals, any observations of malignant tumors were
considered to be biologically relevant, abnormal changes. Finally, although most studies used
paraformaldehyde or freshly prepared formalin as the test article, some studies tested commercial
formalin. Coexposure to methanol was considered to be a major concern for LHP cancers; it was
considered to be less of a concern when identifying effects of inhaled formaldehyde on respiratory
cancers. A final minor difference involved the preference for microscopic examination of several tissues
applicable to assessing potential LHP cancers, and a preference for blinded assessment of the slides.
4.2. UPPER RESPIRATORY TRACT CANCERS
This section examines the evidence pertaining to the carcinogenic effect of formaldehyde
exposure on the URT of humans and animals. The specific endpoints considered included diagnoses of
nasopharyngeal cancer, sinonasal cancer, cancers of the oropharynx and hypopharynx, and laryngeal
cancer; however, as the studies of laryngeal cancer did not contribute to the hazard conclusion (i.e.,
indeterminate), these data are not discussed in this Overview (see Section 1.2.5 in the Toxicological
Review). This section also describes experimental animal studies examining the potential for cancers of
the nasal cavity and proximal regions of the URT, and mechanistic studies relevant to interpreting
potential carcinogenic effects on the URT.
This document is a draft for review purposes only and does not constitute Agency policy.
96 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
4.2.1. Synthesis of Human Health Effect Studies
Nasopharyngeal Cancer
The evidence for formaldehyde exposure and the risk of nasopharyngeal cancer presents
consistent findings of increased risk in exposed groups across several studies, including results classified
with high, medium, and low confidence. These studies examined different populations, in different
geographical locations, under different exposure settings and employing different study designs.
Fourteen of 17 studies reported increased risks of nasopharyngeal cancer with at least one metric of
formaldehyde exposure—often with both clear statistical significance and dose-response relationships
(see Figure 16). These included the results of a large cohort study of 25,619 U.S. workers (Beane
Freeman et al.. 2013) classified with high confidence, and all four sets of results classified with medium
confidence. Nine studies in eight independent populations reported relative effect estimates greater
than three-fold. The study results exhibited a biologically coherent temporal relationship consistent
with a pattern of exposure to formaldehyde and subsequent death from nasopharyngeal cancer,
allowing time for cancer induction, latency, and mortality.
The reported dose-response relationships showing that multiple measures of increased
exposure to formaldehyde were repeatedly associated with increased risk of mortality from
nasopharyngeal cancer were especially strong among studies primarily focused on squamous cell
carcinomas. Excluding nasopharyngeal cancer cases with undifferentiated or nonkeratinizing histology,
Vaughan et al. (2000) reported a clear dose-response with increased probability of exposure. Among
those subjects considered to be "definitely exposed," there were increasing risks of nasopharyngeal
cancer with increasing duration of formaldehyde exposure (p < 0.001) and with increased cumulative
formaldehyde exposure (p < 0.001). Further evidence of dose-response relationships was reported by
Beane Freeman et al. (2009) for peak formaldehyde exposures (p = 0.005, model including exposed and
unexposed person-years), and, to a lesser degree, for cumulative exposures (p = 0.06, model including
exposed and unexposed person-years) and with average intensity of formaldehyde exposure (p = 0.09,
model including exposed and unexposed person-years).
The evaluation of potential biases resulted in reasonable confidence that alternative
explanations have been ruled out, including chance, bias, and confounding within individual studies or
across studies. There are reasonable explanations for the lack of findings in the three studies with very
low background rates of nasopharyngeal cancer. The NPC results from the Coggon et al. (2014); Meyers
et al. (2013); Siew et al. (2012) studies were all considered to lack sensitivity to detect any true effect
because there was a very low number of expected cases in study populations, which contributed to their
classifications of low confidence.
This document is a draft for review purposes only and does not constitute Agency policy.
97 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
100'
10-
3-
2-
0.1-
oil I O O I fs
I II II II II 1
Hildesheim Yu West Olsen Roush
Asian Populations
European and U.S. Populations
en en en en cn en en " en
enenenenoooenenenenenenso-SSensenenenenQzenenenenenenenenenenenen
oooosssoooooodcococoencoenenenencooooooooooooo
I I I I I I I I I I I I I I
V, ^
A
> q> K9> K9>
1.
<> ~
~ ^
^ A
A A
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
medium and low confidence. Seventeen informative studies evaluated sinonasal cancer among study
subjects with formaldehyde exposure based on occupational history, including 3 sets of results classified
with medium confidence—one of which represents a large, pooled analysis of 12 case-control studies
(see Figure 17). These studies examined different populations, in different locations, under different
exposure settings, and used different study designs.
For sinonasal cancer, it is important to consider the histological subtype or types in each report
(squamous cell carcinoma, adenocarcinoma, or mixed). Sinonasal cancer is exceedingly rare; eight
studies reported zero cases in their study populations. With expected rates for sinonasal cancer as low
as 0.3 cases per 100,000 people each year, these studies lacked the statistical sensitivity to detect an
association with formaldehyde and were classified with low confidence. Of the nine studies that did
observe cases of sinonasal cancer, results from six reported increased risks of sinonasal cancer that
appeared to be associated with exposure to formaldehyde—four of six sets of results had been classified
with medium confidence (Beane Freeman et al.. 2013; Luce et al.. 2002; Roush et al.. 1987; Olsen and
Asnaes, 1986) and two with low confidence (Teschke et al.. 1997; Hansen and Olsen. 1995).
Associations were stronger for adenocarcinomas than for squamous cell carcinomas. However, both
histological cell type groupings, and a mixed type group, yielded results that were consistently
elevated—with a clear demonstration of statistical significance for the adenocarcinomas. Two medium
confidence studies reported at least a three-fold increase in risk for adenocarcinoma. Potential
confounding by wood dust was addressed and ruled out by the authors. Each of the other three sets of
results that did not report some increase in risk associated with formaldehyde exposure had been in the
group classified with low confidence, in part due to their lack of sensitivity to detect a true effect
(Coggon et al.. 2014; Siew et al.. 2012; Pesch et al.. 2008).
A dose-response relationship was observed in a large, pooled analysis of 12 case-control studies.
Luce et al. (2002)9 pooled 196 cases of sinonasal adenocarcinoma and 432 cases of squamous cell
carcinoma and were able to contrast risks in three levels of exposure probability with the risk in the
unexposed. An exposure-response relationship for adenocarcinoma, controlling for coexposure to wood
dust, was observed for both men and women with the highest risks among those with the highest
probability of exposure. The odds ratio (OR) among men with the highest cumulative exposure was 3.0
(95% CI: 1.5, 5.7), while it was 5.8 (95% CI: 1.7, 19.4) among women. No dose-response pattern was
observed for squamous cell carcinoma. Analyses by Luce et al. (2002)allowing for a 20-year induction
period showed only minimal impacts on the magnitude of relative risk; longer latency periods were not
evaluated, which leaves some uncertainty.
The evaluation of chance, bias, and confounding within individual studies or across studies
resulted in the conclusion that these alternative explanations for the observed associations could be
reasonably ruled out. While smoking and alcohol may be independent risk factors for sinonasal cancer
9Note the pooled study by Luce et al. (2002) includes data from 12 publications and thus represents substantially more
information than a single result (see Toxicological Review for additional details).
This document is a draft for review purposes only and does not constitute Agency policy.
99 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
they are unlikely to be related to formaldehyde exposure and therefore unlikely to be across-the-board
confounders. Wood dust, however, is a potential confounder as many wood-related jobs also have
exposures to formaldehyde and the association between wood dust exposure and sinonasal cancer is
extremely strong, with relative risks greater than 30-fold (Olsen and Asnaes, 1986). Wood dust may be
an independent risk factor for sinonasal cancer; however, the majority of investigators presented
analytic results for formaldehyde among workers who were either not exposed to wood dust (Hansen
and Olsen. 1995; Olsen and Asnaes. 1986), or else controlled for the potential confounding of the effects
of wood dust on the risk of sinonasal cancer and did not find wood dust to be a confounder (Luce et al..
2002). Although many of the analyses lacked precision due to the rarity of sinonasal cancer, the
observations of multiple instances of very strong associations in different settings reduces the likelihood
that chance, confounding, or other biases can explain the observed associations.
This document is a draft for review purposes only and does not constitute Agency policy.
100 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
100'
10'
1 -
0.1
Beane
Freeman
r 1
Luce
4-
vvvvvvvv
Zero Cases
Squamous Cell
I C arcin o m a s
Mixed cell type Carcinomas
2 2 2 2 2 2 ^ 2
2 2 2
o o o o
o o o
*<}
^ t c
*° ' o-- S
*vv
^ ^
o- <
0
vi? ^ yN 0- •&
¦ •a-
v o
Figure 17. All epidemiological studies reporting sinonasal cancer risk estimates.
Results are grouped by histological type as squamous cell carcinomas, mixed cell types, or
adenocarcinoma. SMR: standardized mortality ratio. SPIR: Standardized Proportional Incidence Ratio.
RR: relative risk. OR: odds ratio. TSFE: time since first exposure. For each measure of association, the
number of exposed cases is provided in brackets. For studies with multiple metrics of exposure, only the
highest category of each exposure metric is presented. Note that two studies (Luce et al., 2002; Olsen
and Asnaes, 1986) reported separate results for squamous cell carcinoma and adenocarcinoma and
appear twice in the figure. Also note that the pooled analysis by Luce et al. (2002) includes data from 12
publications and thus represents substantially more information than a single set of results.
1 Oropharyngeal/Hypopharyngeal cancer
2 Evidence describing an association between formaldehyde exposure and the risk of
3 oropharyngeal/hypopharyngeal cancer was available from nine reports on six distinct study
This document is a draft for review purposes only and does not constitute Agency policy.
101 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
populations—four reports on three cohort studies (Coggon et al.. 2014; Meyers et al.. 2013; Marsh et
al.. 2007; Marsh et al.. 2002) and five reports on three case-control studies (Laforest et al.. 2000;
Gustavsson et al.. 1998; Vaughan, 1989; Vaughan et al.. 1986a, b).
Increased risks of oropharyngeal/hypopharyngeal cancer were reported by two medium
confidence studies associated with multiple metrics of formaldehyde exposure, but little other evidence
of increases in risk across one other medium and two low confidence was observed (see Figure 18). The
strength of the association was variable with several studies reporting results near the null, and two
medium confidence studies reporting three- to five-fold increases in risk among the highly exposed. One
study observed dose-response relationships using multiple metrics of exposure. The evaluation of bias
and sensitivity resulted in reasonable confidence that alternative explanations have been ruled out,
including chance, bias, and confounding within individual studies or across studies.
This document is a draft for review purposes only and does not constitute Agency policy.
102 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde -
Inhalation
o
>
—
o
QL
1001
10 -
3 -
2 -
1 -
0.1 -
X O CL
LU >, ^
0 = 0
I I I I I O
'vaughan 1986a, b,
198 9
M arsh 2002
< i o i
o n ly
i_i
« C o g g o n
i
Undifferentiated Oro-and Hypopharynx
Hypopharyx only
m o) co
CO CO CO
c CO CD O) t-
— Tf id cd o o)
V) V) to w W
I I I I I I I I
I I I I I I I I I I I
v"5>
,v"?>
v"?>
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Toxicological Review of Formaldehyde - Inhalation
Holmstrom et al.. 1989a; Woutersen et al.. 1989; Sellakumar et al.. 1985; Kerns et al.. 1983; Dalbev.
1982). The most consistent animal evidence of formaldehyde-induced respiratory cancers was the
development of squamous cell carcinomas (SCCs), with the most useful data from studies of exposed
rats (see Figure 19). Following exposure of rats to formaldehyde for two years, an increase in SCCs was
observed in five of six studies interpreted with medium or high confidence. SCCs were not reproducibly
detected below 6 mg/m3 formaldehyde; however, none of the available rat studies tested exposure
between 3 and 6 mg/m3, introducing some uncertainty.
Specifically regarding SCCs, these exposure-induced tumors were restricted to the nasal cavity,
were not observed in other respiratory tract regions, such as the larynx and lung, and generally
developed in animals that were observed for longer than 12 months. The locations of the induced SCCs
were consistent with both the distribution of inhaled formaldehyde and locations of other
formaldehyde-induced nasal pathologies. There were clear species differences in the severity of SCCs,
with hamsters displaying little evidence of toxicity and rats exhibiting amplified responses as compared
to mice (likely attributable to a lower inhaled dose of formaldehyde). While these tumors were
detected in exposed male and female Fischer 344 (F344) and Sprague Dawley (SD) rats, findings in
Wistar rats were less clear. The rat studies are summarized in Figure 19.
This document is a draft for review purposes only and does not constitute Agency policy.
104 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Toxicological Review of Formaldehyde - Inhalation
to
U
U
to
¦O
QJ
"O
fTJ
0)
u
£
u
C
50-
45-
40-
35-
30-
25-
- Z 20-
s.O
\p
O
c
15-
10-
A
~
A
~
~
Kerns et al., 1983
(male F344; high)
Kerns et al., 1983
(female F344; high)
Monticello et al., 1996
(male F344; high)
Woutersen et al., 1989
(male Wistar; high)
Kamata et al., 1997
(male F344; medium)
Sellakumar et al., 1985
(male SD; medium)
Holmstrom et al., 1989
(female SD; medium)
A
A
~
r
0WEA—Mr
~
10
15
20
Formaldehyde concentration (mg/m )
Figure 19. Incidence of nasal squamous cell carcinomas in rats exposed to
formaldehyde for at least 2 years.
% incidence data from the high (black outline and fill) and medium (gray outline and no fill) confidence
studies are arrayed. Different shapes represent different rat strains.
In addition to SCCs, precancerous dysplastic lesions were induced in rats and mice (Holmstrom
et al.. 1989a; Morgan et al.. 1986b; Kerns et al.. 1983), sometimes at lower formaldehyde
concentrations than those at which malignant tumors were observed. The dysplasia and neoplasms
were predominantly localized to anterior regions of the nasal respiratory epithelium, although the
lesions progressed to more posterior locations with increasing duration and concentration of
formaldehyde exposure, with one study reporting that dysplasia can develop in portions of the proximal
trachea in rats.
SCC development depended on the duration of observation and, based on an increasing
incidence and severity of lesions in animals exposed for longer periods of time, the formaldehyde
exposure duration. Most notably, the lesion incidence, as well as the tumor invasiveness and latency,
was reproducibly shown, across two species, to worsen with increasing exposure concentration.
This document is a draft for review purposes only and does not constitute Agency policy.
105 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
4.2.3. Mode-of-action Information
In F344 rats chronically exposed to formaldehyde, there is a clear temporal, dose-responsive
and biological relationship in the appearance of exposure-related genotoxicity, sustained epithelial
damage, cellular proliferation, and eventual development of SCC or polypoid adenoma (PA; a benign
lesion independent from, and not a precursor to, SCC), consistent with similar relationships evident in
analogous URT tissues from both the monkey and human databases. Furthermore, the chronic
formaldehyde exposure concentrations reported to elicit nasal cytotoxic pathology appear to be higher
in the rats and monkeys evaluated experimentally, compared with the results from human
epidemiological cohorts, whereas formaldehyde-associated genotoxicity has been induced in analogous
portal-of-entry tissues from rats, monkeys, and humans exposed to similar formaldehyde concentrations
(see Toxicological Review for details). Together, genotoxicity, cellular proliferation, and cytotoxicity-
induced tissue regenerative proliferation exhibit multiple layers of coherence as a function of species
and anatomy, temporality, concentration, and duration of exposure. When integrated, this evidence
forms a biologically relevant MOA for formaldehyde-induced URT carcinogenesis (U.S. EPA. 2005a). A
summary and evaluation of the mechanistic evidence is presented in Table 40.
Strong, consistent evidence from rodents and monkeys supports the role for both direct (i.e.,
potentially DNA-protein crosslinks, DPX, or hmDNA adduct associated) mutagenicity as well as indirect
genotoxicity, mutagenicity, and regenerative proliferation resulting from respiratory tissue pathology, in
rodent URT carcinogenesis (see Toxicological Review Section 1.2.5 Upper Respiratory Tract Cancer
Mode-of-Action Analysis for details). DNA labeling studies in rodent nasal epithelium suggest that cell
division may also accelerate in response to marginally cytotoxic tissue concentrations resulting from
short-term, lower level, or discontinuous exposure scenarios, although this evidence was neither strong
nor consistent across similar studies and model systems. Observations of mutagenicity, cytotoxic
epithelial pathology, and proliferation correspond histologically, anatomically, temporally, and dose-
responsively with subsequent SCC and PA formation, consistent with contribution of both mutagenesis
and regenerative proliferation to rodent URT carcinogenesis following formaldehyde exposure.
Mutagenicity is presumed to be a relevant component of URT carcinogenesis in humans,
supported by strong evidence of direct genotoxicity in both rodents and monkeys and consistent
observations of direct genotoxicity and mutagenicity from human epidemiological studies. Increased
nasal epithelial cell proliferation (in rats and monkeys) coincides anatomically with dysplastic lesions
found in tissues from similar species as well as with progressive, proliferative lesions in the nasal/buccal
epithelium and nasopharynx of chronically exposed humans. This cross-species concordance, combined
with the observation that cellular proliferation may be induced at lower exposures or following shorter
durations of exposure than those eliciting tissue metaplasia, suggests that cellular proliferation in the
presence of marginal tissue toxicity may also be potentially relevant to human URT carcinogenesis, as
this episodic exposure scenario may be more frequently encountered in human populations than the
continuous, chronic high-level exposures traditionally employed in rodent cancer bioassays. Increasing
This document is a draft for review purposes only and does not constitute Agency policy.
106 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Toxicological Review of Formaldehyde - Inhalation
incidence or severity of nasal dysfunction and progressive pathology is associated with escalating
formaldehyde exposure concentration or duration in humans, monkeys, and rats. While POE tissue
sensitivity to formaldehyde toxicity may quantitatively differ from humans to rats and other rodents,
qualitatively similar nasal dysfunction and pathology consistent with pre-neoplastic stages of cancer
progression are observed across analogous tissues from all affected species, and therefore conclusions
derived from these model systems are presumed relevant to human URT carcinogenesis. Given this
presumed relevance, the potential for an increased susceptibility of specific human populations to
developing URT cancers can be informed by both the human data and relevant mechanistic evidence
from experimental model systems.
Table 34. Summary considerations for upper respiratory tract (URT) carcinogenesis
(the primary support for genotoxicity or mutagenicity is noted; see Toxicological Review
for additional details)
Hypothesized
Mechanistic
Event
Experimental Evidence
Pertinent to Mechanistic Event
Human Relevance
Weight-of-evidence and
Biological Plausibility
Direct, or
presumed direct,
genotoxicity and
mutagenicity
(and indirectly
supporting
information)
• 'T* MN in URT tissue from human students
and workers at average concentrations as
low as 0.1 mg/m3 (subchronic-to-chronic
exposure) {Aglan, 2019, 6196781}: (Ballarin
et al.. 1992): (Burgaz et al.. 2001): (Burgaz et
al.. 2002): (Costa et al.. 2019): (Costa et al..
2008): (Ladeira et al.. 2013): (Peteffi et al..
2015): (Viegas et al.. 2013): (Viegas et al..
2010): (Ye etal.. 2005)
• 'T* DNA monoadducts in nasal tissues of
exposed rats and monkeys using highly
sensitive methods (short-term or subchronic
exposure) (Yu et al.. 2015: Lu et al.. 2011:
Moeller et al.. 2011: Lu et al.. 2010)
• 'T* DPX in URT tissues of monkeys (acute
exposure) and F344 rats (acute-to-
subchronic exposure) (e.g.. (Lai et al.. 2016:
Georgieva et al.. 1999: Casanova et al..
1994)) [note: not observed in 2 short- term
controlled exposure studies]
• No effect on MN incidence nasal tissue (Speit
et al.. 2011) or in BAL cells (Neuss et al..
2010) in single studies in rats (28d exposure)
• While several studies suggest a role for
exposure-induced modifications to the
tumor suppressor, p53, in SCC development
(see Appendix A.4.5), a short-term study in
mice deficient for Trp53 (encodes p53) failed
to observe increases in tumors (Morgan et
al.. 2017)
Yes. Markers of direct
genotoxicity correspond
anatomically and temporally
with subsequent URT
neoplasia in experimental
animal models, are consistent
with increased MN induction
following exposure in
humans and are presumed
relevant to human
carcinogenesis.
Strong and consistent
evidence for formaldehyde-
induced direct genotoxicity
and mutagenicity exists
from both experimental
animal models and human
molecular epidemiology to
support a significant role
for mutagenicity in URT
carcinogenesis.
This document is a draft for review purposes only and does not constitute Agency policy.
107 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Hypothesized
Mechanistic
Event
Experimental Evidence
Pertinent to Mechanistic Event
Human Relevance
Weight-of-evidence and
Biological Plausibility
• Indirect support: strong and consistent
evidence of mutagenicity (increased
incidence of MN, CA, and chromosome
aneuploidies) in PBLs of human workers (see
Section 4.3)
• Indirect support: strong and consistent
evidence of genotoxicity and mutagenicity in
numerous in vitro mammalian and non-
mammalian systems (see Appendix A.4)
Cytotoxicity-
induced
regenerative
proliferation
• \1/ Nasal mucociliary function, 'T* nasal
hyperplasia, keratinization or squamous
metaplasia, URT rhinitis, irritation, and
inflammation in humans (acute-to-chronic
exposure)
• \1/ Nasal cilia content, 'T* hyperplasia and
squamous metaplasia in URT tissues from
monkeys (acute-to-subchronic exposure)
• Associated with 'T* URT cell proliferation in
rhesus monkeys
• \1/ Nasal mucociliary function, 'T* nasal
rhinitis, hyperplasia and squamous
metaplasia dysplasia in various rat strains
and B6C3F1 mice (acute-to-chronic
exposure)
• Associated with 'T* URT proliferation (rats;
mice)
Yes. Increasing incidence or
severity of URT dysfunction
or pathology is positively
associated with
formaldehyde exposure in
humans, monkeys, and rats.
A continuum of similar
epithelial pathology is
observed across affected
species at POE tissues, and
therefore the resulting
increased cellular turnover
observed in experimental
models is presumed relevant
to human carcinogenesis.
Strong and consistent
evidence exists which
associates the nasal
epithelial pathology-driven
proliferation with SCC
abundance following
formaldehyde exposure in
rodent experimental
models to support a
significant role for
regenerative proliferation
in URT carcinogenesis.
Cellular
mitogenesis in
the absence of
cytotoxic tissue
pathology
• Clear evidence of 'T* URT cell proliferation
under conditions also resulting in tissue
pathology in rhesus monkeys
• Exposure to subcytotoxic concentrations not
evaluated
• Clear evidence of 'T* URT cell proliferation
under conditions also resulting in tissue
pathology in Wistar and F344 rats
(>4 mg/m3)
• Suggestive evidence of 'T* URT cell
proliferation under conditions not clearly
causing tissue pathology (<4 mg/m3)
Yes. Cellular proliferation
may be increased at lower
exposures or following
shorter durations of exposure
than that eliciting tissue
pathology, which suggests
that mitogenesis may be
directly stimulated by
formaldehyde exposure.
Proliferation is expected to
accelerate and enhance
carcinogenesis in both
humans and animals and is
presumed relevant to human
carcinogenesis.
Limited and inconsistent
evidence associates cellular
proliferation with
formaldehyde exposures
below those eliciting
cytotoxic pathology in the
rat nasal epithelium, which
precludes a determination
as to the importance of this
phenomenon in URT
carcinogenesis.
Oxidative stress,
immune disease
and dysfunction
in the URT
• 'T* LRT infection frequency, inflammation,
allergic outcomes in children; 'T* leukocyte
activation, allergy symptoms, chronic URT
inflammation and \|/ infection resistance in
adult workers (subchronic-to-chronic
exposure)
Yes. Nasal infection, markers
of persistent inflammation or
immune dysfunction are
positively associated with a
range of formaldehyde
exposure in both humans and
rodents. Oxidative stress and
While evidence exists
supporting oxidative stress,
chronic inflammation and
various immune
dysfunctions following
formaldehyde exposure in
humans and experimental
This document is a draft for review purposes only and does not constitute Agency policy.
108 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Toxicological Review of Formaldehyde - Inhalation
Hypothesized
Mechanistic
Event
Experimental Evidence
Pertinent to Mechanistic Event
Human Relevance
Weight-of-evidence and
Biological Plausibility
• 'T* LRT oxidative stress, markers of
inflammation and leukocyte recruitment in
rats and mice; 'T* airway wall thickening or
remodeling in mice and rats following
ovalbumin sensitization
• 'T* Malignancy and neutrophil involvement of
lung metastases, \|/ lung natural killer (NK)
cell numbers and activity in C57BL/6 mice
chronic inflammatory
diseases
(immunosuppression) are
presumed relevant to human
carcinogenesis. The
relevance of other immune
system dysfunctions to
human carcinogenesis, such
as allergy, is less clear.
animal models, the
evidence supporting
associations between these
effects and URT
carcinogenesis is
insufficient to evaluate the
contribution of these
effects independently in
either humans or
experimental animals.
4.2.4. Overall Evidence Integration Judgments and Susceptibility for Upper Respiratory Tract Cancers
Robust evidence from human epidemiological studies of groups exposed to occupational
formaldehyde levels supports a causal association between inhalation of formaldehyde and
nasopharyngeal cancer, while moderate evidence supports a causal association for sinonasal, and
oropharyngeal/hypopharyngeal cancers (see Table 41). Consistent increases in risk were reported by
numerous high and medium confidence studies of diverse populations in different geographic locations
and exposure settings that accounted for expected temporal relationships for cancer induction and
progression, with several reporting a large magnitude of relative risk (RR > 3). A dose-response gradient
was reported for various measures of exposure, including cumulative exposure, duration of exposure,
and peak exposure. Robust evidence with site concordance for nasal cancers also is provided from
studies in experimental animals (rats and mice). However, the relevance of the rodent nasal cancers
and related mechanistic changes to human cancers not localized in or around the nasal cavity is
questionable and difficult to infer. The incidence of lesions, as well as the tumor invasiveness and
latency, was reproducibly shown to worsen with increasing formaldehyde exposure level. The
distribution of tumors was dependent on duration of exposure as well as formaldehyde concentration.
Mechanistic changes associated with the development of cancer were consistently observed in humans
and experimental systems, including genotoxicity, epithelial damage and proliferation, and eventual
cancer development in relevant URT tissues. The lesions exhibited a temporal and dose-response
relationship coherent with carcinogenesis. The evidence is sufficient to conclude that a mutagenic
mode of action (MOA) of formaldehyde is operative in formaldehyde-induced nasopharyngeal
carcinogenicity. Therefore, the evidence demonstrates that formaldehyde inhalation causes
nasopharyngeal cancer in humans and the evidence indicates that formaldehyde inhalation is likely to
cause sinonasal and oropharyngeal/hypopharyngeal cancer, given appropriate exposure circumstances.
These conclusions were primarily based on studies of groups exposed to occupational formaldehyde
levels and coherent findings in chronic rodent bioassays where tumors were generally only observed at
formaldehyde concentrations above 6 mg/m3.
This document is a draft for review purposes only and does not constitute Agency policy.
109 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Table 35. Evidence integration summary for effects of formaldehyde inhalation on
URT cancers
Human Evidence
Animal Evidence
Mechanistic Interpretation
Evidence Integration
Judgment
Robust for Nasopharvnaeal Cancer
Robust for Nasal Cancers.
• Relevance to humans:
The evidence
(NPC). based on:
based on:
Findings were consistent
demonstrates that
Human health effect studies:
Animal health effect studies:
and coherent across species
formaldehyde
Consistent increases in risk across
• Tumors of the URT
(including humans). While
inhalation causes
numerous high, medium and low
(predominantly SCCs) were
cancer site concordance is
nasopharyngeal
confidence studies, which
consistently observed in mice
not required for hazard
cancer in humans
demonstrated very strong
and in several strains of rats in
determination U.S. EPA
given appropriate
associations (8 studies reported at
numerous high and medium
(2005a), given the known
exposure
least a 3-fold increase in risk),
confidence studies, at
reactivity and distribution of
circumstances
evidence of dose-response
concentrations >6 mg/m3
inhaled formaldehyde, a
relationships across multiple metrics,
• The lesions progressed in
lesser level of confidence in
The evidence
and a temporal relationship
incidence, severity to more
the relevance of the animal
indicates that
consistent with causality
posterior locations with
data is inferred for
formaldehyde
Biological Plausibility:
increasing duration and
Oropharvnaeal
inhalation likely
Mechanistic evidence from human
concentration of formaldehyde
/Hvoooharvnaeal Cancer
causes sinonasal
studies indicates a clear biological
exposure, or duration of
(noting that oronasal vs.
cancer and
relationship with genotoxicity,
observation
nasal breathing in humans
oropharyngeal
epithelial damage and proliferation,
• Most notably, the lesion
adds plausibility for these
/hypopharyngeal
and eventual cancer development in
incidence, as well as the tumor
cancers)
cancer given
relevant URT tissues
invasiveness and latency, was
• /WO/4: Together,
appropriate exposure
reproducibly shown to worsen
genotoxicity, cellular
circumstances
Moderate for Sinonasal Cancer.
with increasing exposure level.
proliferation and
based on:
Biological Plausibility: Changes
cytotoxicity-induced
These conclusions
Human health effect studies:
consistent with cancer
regenerative proliferation
were primarily based
Increases in risk across a set of
development were observed
exhibit multiple layers of
on studies of groups
medium and low confidence studies,
across multiple species, with a
coherence as a function of
exposed to
including 3 studies reporting at least
clear biological relationship
species, anatomy,
occupational
a 3-fold increase in risk for
among the appearance of
temporality, concentration
formaldehyde levels
adenocarcinoma, and the largest
genotoxicity, sustained
and duration of exposure,
and coherent
study demonstrating a clear dose-
epithelial damage, cellular
and when integrated, form a
findings in chronic
response relationship
proliferation, and eventual
biologically-relevant MOA
rodent bioassays
Biological Plausibility: (see NPC)
tumor development in rats.
(U.S. EPA. 2005a).
where tumors were
While most findings were
Furthermore, the chronic
generally only
Moderate for
localized to the nasal cavity,
formaldehyde exposure
observed at
Orooharvnaeal/Hvoooharvnaeal
some evidence indicates that
concentrations reported to
formaldehyde
Cancer, based on:
more distal changes, including
elicit nasal cytotoxic
concentrations above
Human health effect studies:
dysplasia, can occur with very
pathology appear to be
6 mg/m3
• Increased risks in two medium
high formaldehyde exposures
higher in the rats and
confidence studies that evaluated
or different breathing
monkeys evaluated
multiple metrics of exposure and
(oronasal)
experimentally, compared
reported 3- to 5-fold increases in
Biological Plausibility: A
with the results from human
those highly exposed, including 1
spectrum of mechanistic
epidemiological cohorts,
which demonstrated clear dose-
evidence, including evidence of
whereas formaldehyde-
response relationships across several
DNA monoadducts and DPX in
associated genotoxicity has
metrics
the URT of exposed rodents
been induced in analogous
• Little evidence of increases in risk
and monkeys, provides strong
tissues from rats, monkeys
(near the null) across 1 medium and
support for the plausibility of
and humans exposed
2 low confidence results
similarly (<0.9 mg/m3)
This document is a draft for review purposes only and does not constitute Agency policy.
110 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
Human Evidence
Animal Evidence
Mechanistic Interpretation
Evidence Integration
Judgment
Biological Plausibility: Mechanistic
evidence from human studies in
relevant POE tissues (e.g., buccal
cells) demonstrate mechanistic
changes consistent with the
development of cancer, including
genotoxicity
the evidence for nasal cancers
in experimental animals
• Potential Vulnerabilities'.
There is negligible evidence
to evaluate the potential risk
to sensitive populations/
lifestages
4.3. LYMPHOHEMATOPOIETIC (LHP) CANCERS
This section examines the evidence pertaining to the carcinogenic effect of formaldehyde
exposure on lymphohematopoietic (LHP) cancer in humans and animals. The specific endpoints
included: Hodgkin lymphoma, multiple myeloma, myeloid leukemia, and lymphatic leukemia; however,
as the studies of lymphatic leukemia did not contribute to the overall conclusion (i.e., indeterminate),
these data are not discussed in this Overview. This section also discusses experimental animal studies
examining histopathological lesions associated with leukemia or lymphoma, and mechanistic studies
relevant to interpreting potential carcinogenic effects on these tissues.
4.3.1. Synthesis of Human Health Effect Studies
Myeloid Leukemia
Evidence describing the association between formaldehyde exposure and the risk of myeloid
leukemia was available from 13 epidemiological papers reporting on 10 different study populations: 3
case-control studies (Talibov et al.. 2014; Hauptmann et al.. 2009; Blair et al.. 2001) and 10 cohort
studies (Coggon et al.. 2014: Pira et al.. 2014: Meyers et al.. 2013: Saberi Hosniieh et al.. 2013: Beane
Freeman et al.. 2009; Hayes et al.. 1990; Ott et al.. 1989; Stroup et al.. 1986; Walrath and Fraumeni,
1984, 1983). Hauptmann et al. (2009) combined the study populations from Hayes et al. (1990) with
those from Walrath and Fraumeni (Walrath and Fraumeni. 1984. 1983) and reconstructed individual
exposure estimates. Checkoway et al. (2015) reanalyzed Beane Freeman et al. (2009) with different
definition of the exposure categories and presented results for specific sub-types of myeloid leukemia.
For the purposes of this evaluation, cancer cases reported as monocytic leukemia or nonlymphocytic
leukemia were included as myeloid leukemia.
All 13 informative studies reported increased risks of myeloid leukemia associated with
exposure to formaldehyde; these studies examined different populations in different locations and
exposure settings and using different study designs. Consistent reports of elevated risks were provided
by the five studies with population-level exposure assignments (Pira et al.. 2014: Haves et al.. 1990:
Stroup et al.. 1986; Walrath and Fraumeni. 1984, 1983). The results from Walrath and Fraumeni (1984,
1983) and Hayes et al. (1990) were classified with medium confidence, while the results from the other
two studies were classified with low confidence. Although the exposure settings in studies of
This document is a draft for review purposes only and does not constitute Agency policy.
Ill DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Toxicological Review of Formaldehyde - Inhalation
anatomists and embalmers involved coexposure to methanol, whether there is an association of
methanol exposure with leukemia is not known, and elevations in leukemia risk also were observed in
studies involving other exposure settings. Four high and medium confidence studies with individual-
level exposure assignments (Coggon et al.. 2014; Meyers et al.. 2013; Beane Freeman et al.. 2009;
Hauptmann et al.. 2009) also showed elevated risks; three of the studies allowed for the evaluation of
dose-response relationships with increased formaldehyde exposures using multiple metrics of exposure
(see Figure 20). A pattern of increasing dose-response was indicated in analyses of exposure duration
(Meyers et al.. 2013; Hauptmann et al.. 2009), cumulative exposure (Meyers et al.. 2013), and with peak
exposure metrics (Meyers et al.. 2013; Beane Freeman et al.. 2009; Hauptmann et al.. 2009). These
three studies with high confidence results also observed some indication of an increase in mortality risk
at about 15-20 years since the initial exposure consistent with a biologically relevant induction/latency
period; Hauptmann et al. (2009) showed a clear increase in risk at 20+ years since first exposure.
Studies with higher quality exposure data based on individual-level exposure assessment
generally reported stronger associations. The results at the highest levels of formaldehyde showed an
approximately two- to three-fold relative increase in risk of mortality (Meyers et al.. 2013; Beane
Freeman et al.. 2009; Hauptmann et al.. 2009; Blair et al.. 2001). One study's results that were classified
with medium confidence due to exposure measurement error (Coggon et al.. 2014) showed a slightly
elevated risk for those workers with the highest job exposures, but also slight decreased risk for those
with the highest duration of exposure. Alternative explanations for the associations observed by the
high and medium confidence studies can be reasonably ruled out, reinforced across studies by the
consistency in results and dose-response patterns. Four other studies with results classified as low
confidence were less consistent, possibly because these studies were limited by low case numbers and
missing or imprecise exposure information (Talibov et al.. 2014; Saberi Hosnijeh et al.. 2013; Blair et al..
2001; Ott et al.. 1989).
Different measures of exposure reflected different risks both within and among studies,
although all provided some evidence of increased mortality from myeloid leukemia associated with
formaldehyde exposure. One study showed the strongest relationship of myeloid leukemia mortality
with duration of formaldehyde exposure (Hauptmann et al.. 2009). Another showed increased risks for
peak exposure and average exposure but not for cumulative exposure or "any" exposure (Beane
Freeman et al.. 2009). A third study showed increased risk in the study population as a whole that was
stronger among workers with the longest duration of exposure and workers with the greatest length of
time since first exposure to formaldehyde (Meyers et al.. 2013). As the different measures of exposure
are likely to be correlated, it may not be possible to single out one exposure metric as most biologically
meaningful.
The pattern of increased risk of myeloid leukemia reflects the associations seen within two
subtypes, acute myeloid leukemia (AML) and chronic myeloid leukemia (CML). However, among the
studies with separate estimates by subtype, risks were elevated for both AML and CML, with the
This document is a draft for review purposes only and does not constitute Agency policy.
112 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Toxicological Review of Formaldehyde - Inhalation
associations for CML appearing to be as strong as or stronger than the associations with AML
(Checkoway et al.. 2015; Saberi Hosnijeh et al.. 2013; Blair et al.. 2001; Stroup et al.. 1986). Six studies
reported specific results for AML; two were classified with high confidence (Meyers et al.. 2013;
Hauptmann et al.. 2009), and four with low confidence (Checkoway et al.. 2015; Talibov et al.. 2014;
Saberi Hosnijeh et al.. 2013; Blair et al.. 2001). Both of the high confidence results showed non-
significantly elevated risks of AML associated with formaldehyde, as did three out of four of the low
confidence results—although substantially higher risks were reported in the high confidence results.
The precision of these more specific analyses was very low (a total of 0 to 6 exposed cases were
observed in these studies). The Checkoway et al. (2015) reanalysis of Beane Freeman et al. (2009)
reported non-significant increased risks of AML and CML with a redefinition of peak exposure that
shifted nine cases of myeloid leukemia from the highest category of peak exposure in Beane Freeman et
al. (2009) to the lowest category (referent group) in Checkoway et al. (2015).10 Checkoway et al. (2015)
also reported stronger effects of peak exposure with CML compared to AML but the number of cases in
each exposure category was small.
Results specific to AML are plotted in Figure 21. Four of these six studies reported effect
estimates for both ML and AML (Checkoway et al.. 2015; Meyers et al.. 2013; Saberi Hosnijeh et al..
2013; Hauptmann et al.. 2009) on a total of 14 specific metrics of exposure. To assess whether the
results for AML were comparable to those for ML, the pair-wise effect estimates were plotted against
each other in Figure 22. The correlation between the AML results and the ML results was 0.72
(p < 0.0001) and the slope was 0.97, indicating a very strong alignment among these studies and
strongly suggesting that the collective results for the broader group of ML cases may be inferred to
represent AML as well.
10ln Beane Freeman et al. (2009). for peak exposure there were 4 cases of ML who were unexposed, 14 cases with peak
exposure from >0 to <2 ppm, 11 cases with peak exposure from 2 to <4 ppm and 19 cases with peak exposure >4 ppm. In
Checkoway et al. (2015) the new definition of peak exposure and the recategorization results in 27 cases of ML with peak
exposures from 0 to <2 ppm, 11 cases with peak exposure from 2 to <4 ppm, and 10 cases with peak exposure >4 ppm. The
Checkoway et al. (2015) results were classified with low confidence due to information bias and low sensitivity.
This document is a draft for review purposes only and does not constitute Agency policy.
113 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
All Studies Reporting Myeloid Leukemia Risk Estimates
Population-
level exposure
assessment
Individual-level exposure assessment
I I
100
10
0.1
IB §
Q tt o <
i 1 1! 1
-C £1 SZ _C
O) O) CB CD
=
1
I
£
gj
i
o
~L
cm
1
UJ
w
03
c
Lll
g
<5
s
s
I
"S5
cn
CD
i
CT)
x
I
o
I
lis |2
< 6
2 I
& I
S Z
S °
I US H
Q Q I— X < O
Ills i p, .
Blair
Hauptmann
Beane Freeman/
Checkoway
Meyers
+
Coggon
Studies of Anatomists
and Embalmers
J L
Industrial Workers
Garment
Workers
General
Population j
-—>
¦—¦
¦—¦
"Jf
s.
, ,
SO
CSJ
o
o
a>
?iT
CO
CO
CO
CO
II
n
II
II
"Si"
•—¦
T—
T—
CO
II
o
'—1
II
II
ii
TT
II
G>
c
II
c
c
C
II
II
CJ5
II
||
||
c
ii
n
ii
ii
c
II
II
CsJ
II
c
c
C
c
c
ii
•*t
II
ii
o
«—.
O)
&
r-
c
d
c
C
c
c
o
c
.c.
,C-
c.
,c~.
o
,c.
¦C.
c
CO
T
in
o
r--
c
C
c
CO
CO
in
CO
iq
•—<
•—>
¦—¦
¦—>
¦—'
o>
CO
CO
s
CSI
CD
£>
fsj
II
II
ll
II
ll
CO
CO
CO
csi
csi
csi
II
x—
•*—
T—1
CO
II
o
csi
n
II
II
Csi
q
¦
cm
cm
cm
cm
cr
II
II
II
ii
II
II
cm
ll
II
II
ii
II
cm
II
II
II
cm
cm
cm
cm
cm
—
II
II
II
2
S
2
s
cr
cm
cr
cm
cm
cm
2
cm
cm
cm
cm
cm
2
cm
cm
cm
2
2
2
2
2
cm
cm
cm
QC
co
co
CL
Q.
CL
o
o
o
o
o
O
co
cr
cm
cm
X
X
CO
o
o
o
co
CO
CO
CO
CO
o
o
cm
o
^>>>>
//
jf
J?
-------
Toxicological Review of Formaldehyde - Inhalation
All Studies Reporting Acute Myeloid Leukemia Risk Estimates
Higfi confidence resuHs
100-,
10
0.1
Low confidence results
1
£
nil
I I I
Meyers
Haupimann
1
I
3
Em balm era
Garment Workers
Industrial
Workers
tp
&
I*
¦»
if-k.
c
c
c
c
c
C
pi
V>
CD
*3
¦P*
g
S
g
8
g
g
J l_
General
Population |
I
7 ?
£ Si
¥
5
—I 1 1 1 1 I
^ ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Toxicological Review of Formaldehyde - Inhalation
All Paired Effect Estimates of AML vs. ML
Correlation = 0 72
Slope = 0.97
p < 0.0001
o
2 0
0 1 2 3 4 5
AML Relative Effect Estimates
Figure 22. Epidemiological studies reporting paired estimates of acute myeloid
leukemia risk estimates and myeloid leukemia risk estimates.
Based on six paired effect estimates from Hauptmann et al. (2009), five paired estimates from Meyers et
al. (2013), two paired effect estimates from Checkoway et al. (2015) and one pair of effect estimates from
Saberi Hosniieh et al. (2013).
Multiple Myeloma
Evidence describing the association between formaldehyde exposure and the risk of multiple
myeloma was available from 14 epidemiological studies: 5 case-control studies (Hauptmann et al.. 2009;
Heineman et al.. 1992; Pottern et al.. 1992; Boffetta et al.. 1989; Ott et al.. 1989) and 9 cohort studies
(Coggon et al.. 2014; Pira et al.. 2014; Meyers et al.. 2013; Beane Freeman et al.. 2009; Stellman et al..
1998; Band et al.. 1997; Dell and Teta, 1995; Hayes et al.. 1990; Edling et al.. 1987).
Seven of the 14 studies considered to be informative and included in the review reported
increased risk of death from multiple myeloma associated with exposure to formaldehyde; (Hauptmann
et al.. 2009; Band et al.. 1997; Dell and Teta. 1995; Heineman et al.. 1992; Pottern et al.. 1992; Boffetta
et al.. 1989; Edling et al.. 1987). Four reported mixed or null results (Coggon et al.. 2014; Meyers et al..
2013; Beane Freeman et al.. 2009; Ott et al.. 1989), and three studies reported decreased risk of death
from multiple myeloma associated with exposure to formaldehyde (Pira et al.. 2014; Stellman et al..
1998; Band et al.. 1997) (see Figure 23). Among all the studies that used individual-level exposure
assessment, the study with the highest quality exposure assessment methodology was the National
Cancer Institute study (Beane Freeman et al.. 2009) among industrial workers. The most pronounced
effects in this high confidence study showed a two-fold increased risk of mortality from multiple
myeloma associated with the highest level of peak exposure to formaldehyde (RR = 2.04; 95% CI: 1.01,
This document is a draft for review purposes only and does not constitute Agency policy.
116 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Toxicological Review of Formaldehyde - Inhalation
4.12). The evaluation of the study for this review produced reasonable confidence that alternative
explanations were ruled out, including chance, bias, and confounding.
The findings by Beane Freeman et al. (2009) are supported by the results of one medium
confidence study (Hayes et al.. 1990) and two low confidence studies (Dell and Teta, 1995; Edling et al..
1987). all with population-level exposure assessments. The occupational exposures in the three studies
involved very high peaks and were consistent with Beane Freeman et al. (2009) in showing an elevated
risk, although none was able to rule out chance. Hauptmann et al. (2009) and Ott et al. (1989) assessed
individual-level exposure but only presented results specific to formaldehyde exposures for the study
population as a whole. Similarly, the study of garment workers, a medium confidence study, Meyers et
al. (2013) relied on individual measures of the timing of exposure but did not have formaldehyde
concentration data beyond the industrial hygiene data used to plan the study (Stayner et al.. 1988).
Continuous area monitoring showed that formaldehyde levels were relatively constant with no
substantial peak levels over the work shift (Stayner et al.. 1988), a possible explanation for differences in
results. A set of four studies that assessed individual-level exposure gathered minimal information (e.g.,
questionnaire data on "ever" exposure to formaldehyde) on formaldehyde exposure and were
considered to be low confidence (Stellman et al.. 1998; Heineman et al.. 1992; Pottern et al.. 1992;
Boffetta et al.. 1989). The weaknesses of their relatively imprecise exposure assessment may have
precluded their ability to detect an association, thus explaining their generally null results. Overall, the
collection of studies with analyses of multiple myeloma found an association with formaldehyde
exposure limited to groups of people who experienced high peak exposures.
This document is a draft for review purposes only and does not constitute Agency policy.
117 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
All Studies Reporting Multiple Myeloma Risk Estimates
i 1
Population-level
exposure
assess men!
Individual-level Bxposure assessment
100-1
s
I
*¦»
UJ
**
V
o
&
UJ
a:
10-
3-
2-
1-
0.1
I I
i
High Peak
[ Expc»sufe5|
£
1
i
u
1
till £
J r
5 3 & 1
i s
Seane Freeman
4-+
ACS
Danish
Meyers
-| Coggon
:"+-
studies studies
I I I I
Industrial workers
-7 R
^ %
- ~ T - - s s -
S SE 8 8
—¦ — £2
i s «
• c
I R
3 s s
£ £
s e
I 1 £
^ ^ ^ ^ J*
J3 \f >
-------
1
2
3
4
5
6
7
8
9
10
11
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
Toxicological Review of Formaldehyde - Inhalation
Matanoski. 1989; Solet et al.. 1989; Robinson et al.. 1987; Stroup et al.. 1986; Walrath and Fraumeni.
1984. 1983).
The results of the 12 studies considered to be informative and included in the review were not
consistent. The study of the largest cohort of formaldehyde-exposed workers Beane Freeman et al.
(2009) reported an elevated mortality risk from Hodgkin lymphoma for the cohort as a whole
(SMR = 1.42; 95% CI: 0.96, 2.1; 27 cases) and a pronounced increase in risk among those workers with
the highest peak formaldehyde exposures (RR = 3.96; 95% CI: 1.31, 12.02; 11 cases)—results that were
classified with medium confidence. However, the other medium confidence result from Gerin et al.
(1989) was an OR = 0.5 (95% CI: 0.2, 1.2; 8 cases). The results of the other 10 studies (all low
confidence) were largely null, based on small numbers of cases and wide confidence intervals. The high
survival rate for Hodgkin lymphomas (86%) indicates that mortality data may not be a good proxy for
incidence data for this LHP cancer subtype. Given the relatively weak evidence, these data are not
illustrated in this Overview.
4.3.2. Synthesis of Animal Health Effect Studies
This section considers incidence data for histopathological lesions associated with leukemia or
lymphoma; other evidence supportive of the development of these cancers (e.g., hematological
changes) is discussed in the mode-of-action section. Two medium or high confidence animal bioassays
(in addition, two low confidence studies are briefly discussed in the Toxicological Review) evaluated the
carcinogenic potential of inhaled formaldehyde with respect to lymphohematopoietic (LHP)
malignancies (Kamata et al.. 1997; Kerns et al.. 1983; Battelle, 1982). The majority of formaldehyde
exposure studies in animals focused primarily on the respiratory tract and did not provide routine
examination of other tissues, preventing their ability to inform leukemia and lymphoma.
The largest and most comprehensive cancer bioassay evaluating formaldehyde inhalation
exposure in animals is the chronic study in B6C3F1 mice and F344 rats conducted by Kerns et al., with
documentation in the supporting Battelle report (1983; 1982). The cumulative incidence of lymphoma
(in B6C3F1 mice) and leukemia (in F344 rats) as indicated in the summary tables of this report are shown
in Table 42. The p-values reported by the authors were based on a Cox-Tarone test for the comparison
that adjusts for reduced survival (Battelle. 1982). There was a suggestion of a possible slightly increased
incidence in lymphoma (p-value, 0.06) in female mice, and a slightly decreased incidence in leukemia in
female rats (p-value, 0.006) at the high dose. Taken together with the exposure-induced increases in
bone marrow hyperplasia in rats, this represents an area of uncertainty warranting additional study. A
separate study in male F344 rats also did not report any significant intergroup differences in non-nasal
neoplasms using histopathological evaluations that included tissues relevant to leukemia or lymphoma
(Kamata et al.. 1997), although specific incidence data were not provided to compare with the results of
the more comprehensive bioassay. In addition, high mortality at 18.5 mg/m3 (the next lower group was
2.43 mg/m3) limited this study's ability to detect long-term effects (e.g., surviving rats: 0/32 at 28
This document is a draft for review purposes only and does not constitute Agency policy.
119 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
months; ~3/32 at 24 months). Given the findings in the well-reported bioassay by Kerns et al. (1983;
1982), there is a need for additional animal studies specifically designed to target LHP cancers as the
main endpoint.
Table 36. Incidence of hematopoietic cancers in B6C3F1 mice and F344 rats [source:
(Kerns et al., 1983; Battelle, 1982)1
Endpoint, Species
Sex
Incidence or % Incidence
p-valuesa
0 ppm
18.5 mg/m3
Lymphoma, B6C3F1 Mice
Male
0/119 (0%)
0/115 (0%)
Female
19/121 (16%)
27/121 (22%)
0.062
Leukemia, F344 Rats
Male
11/120 (9%)
5/120 (4%)
0.690
Female
11/120 (9%)
7/120 (6%)
0.006
aThe authors' p-values were based on a Cox-Tarone test that adjusts for reduced survival.
While the results of both Kerns et al. (1983; 1982) and Kamata et al. (1997) suggest that LHP
cancers do not appear to develop in F344 rats, given the identified limitations of the available studies
and the few suggestive changes that were reported (i.e., bone marrow hyperplasia in rats and slight but
uncertain increases in lymphomas in mice), it is difficult to draw definitive conclusions (i.e.,
indeterminate evidence) as to whether formaldehyde exposure might be capable of causing leukemia or
lymphoma in animals based on the currently available evidence.
4.3.3. Mode-of-action Information
The mechanistic database pertinent to leukemogenesis was evaluated based upon the
fundamental assumption that exogenous formaldehyde is not distributed appreciably beyond the
portal-of-entry. The available evidence supports some events that could contribute to plausible
mechanistic pathways relating formaldehyde exposure to LHP carcinogenesis (summarized in Table 43).
However, the database was insufficient to support the evaluation or development of any specific MOA.
There is largely consistent and strong evidence linking genotoxicity and mutagenicity in circulating blood
cells with formaldehyde exposure in studies of humans. Both temporal and dose-response relationships
have been demonstrated in these studies, and mechanistic pathways exist that support a biologically
plausible relationship between formaldehyde exposure and cancer, even though the mechanistic
pathways explaining such systemic effects are unclear (NRC. 2014b). In addition, the evidence
supporting noncancer systemic effects following formaldehyde exposure (e.g., reproductive or
developmental toxicity) provides additional plausibility for cancers at systemic sites. It is important to
note that systemic delivery of formaldehyde is not a prerequisite for the observed mechanistic changes,
as some of the reported systemic effects might result from direct interactions with formaldehyde in the
URT, while others could plausibly result indirectly from events such as URT irritation, cytotoxicity,
oxidative stress, and inflammation locally initiated at the POE.
This document is a draft for review purposes only and does not constitute Agency policy.
120 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 37. Summary conclusions regarding plausible mechanistic events associated
2 with formaldehyde induction of lymphohematopoietic cancers
Hypothesized
Mechanistic Event
Experimental Support for Mechanistic Event
Human
Relevance
Evidence Integration
considering Biological
Plausibility
Formaldehyde-
induced DNA
damage to
peripheral blood
leukocytes
• HSPC aneuploidy and structural chromosome damage in
myeloid progenitors (CFU-GMs) from one population of
human workers occupationally-exposed to median levels
of 1.6 mg/m3 (Lan et al.. 2015: Zhang et al.. 2010): ~T
Monosomy and polysomy in multiple chromosomes
(especially monosomy 1, 5, 7) consistent with damage
observed in patients with MDS or AML (Bassig et al..
2016: Lan et al.. 2015): and 'T* breaks, deletions, and
translocations in chromosome #5. Assay methodology
could not distinguish whether formaldehyde exposure is
associated with a potential tendency toward cytotoxicity
in CFU-GM cells either in vivo or during the in vitro cell
culture period. Inconsistencies in assay protocol reported
by Gentry et al. (2013), which were addressed by
Rothman (2017).
• Deficiencies in progenitor cells (CFU-GM and BFU-E) in
exoosed mice (Zhao et al.. 2020). although results mav
be confounded by methanol coexposure (low
confidence)
• 'T* genotoxicity or mutagenicity in circulating PBLs from
exposed humans, including increases in strand breaks,
MN. CA (Costa et al.. 2019: Wans et al.. 2019: Asian and
Mansour, 2018: Zendehdel et al.. 2018: Costa et al..
2015: Peteffi et al.. 2015: Kirsch-Volders et al.. 2014),
NBUDs, or SCE induction at >0.14 mg/m3 (Jiang et al..
2010). and DPX at higher exoosures (Lin et al.. 2013:
Shaham et al.. 2003)
• ^ DPX in PBLs from mice (Ye et al.. 2013), although
results may be confounded by methanol coexposure (low
confidence)
• 'T* MN in human PBLs and buccal cells from exposed
humans, and associations with years of exposure, in
studies evaluating both tissues (Ladeira et al.. 2011:
Viegas et al.. 2010)
Yes. Evidence
comes
primarily
from exposed
humans.
Strong and consistent human
data exist associating
formaldehyde exposure with
various genotoxic outcomes
in myeloid progenitors and
PBLs, and dose-response
relationships demonstrated.
Genotoxicity in circulating
leukocytes shows
concordance with similar
endpoints in POE tissues.
Aneugenic damage observed
in CFU-GMs from
formaldehyde-exposed
human workers is associated
with MDS or AML in humans.
Together this evidence
constitutes the strongest
support for the biological
plausibility for LHP cancer
induction by formaldehyde.
Evidence of
formaldehyde-
induced impacts
other than
genotoxicity on
circulating blood
cell populations,
including
inflammatory
changes or immune
system dysfunction
• \1/ CFU-GM colony formation in human workers
occupationally-exposed to median levels of 1.6 mg/m3
(Zhang et al.. 2010), which mav reflect not only altered
bone marrow progenitor cell viability, but also immune
dysfunction or altered activation
• Numerous published studies reporting divergent changes
in various peripheral blood cell populations from
formaldehyde exposed humans, including: 'T*
pancytopenia in a few studies and reasonably consistent
decreases in total WBCs; \1/ or ^ in some lymphocyte
populations, with decreased CD8 T cells likely at
concentration >0.5 mg/m3; and fluctuations in immune
cell numbers and immune/inflammation markers show a
complex pattern with concentration, with decreases in
blood cell number and decreased cytotoxic response
generally at higher concentrations, some of which are
Yes. Most of
the available
data come
from human
studies.
The evidence supporting
changes in populations or
function of circulating blood
leukocytes following human
exposure to formaldehyde is
strong in terms of a
frequency of alterations, but
different patterns in changes
are reported (e.g., specific
direction of changes in
various lymphocyte
subpopulations, or in blood
levels of soluble signaling
mediators). LHP cancer risk
increases with loss of normal
immune function.
This document is a draft for review purposes only and does not constitute Agency policy.
121 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Hypothesized
Mechanistic Event
Experimental Support for Mechanistic Event
Human
Relevance
Evidence Integration
considering Biological
Plausibility
consistent with observations in AML patients (Kim et al..
2015). Other studies indicate immune cell activation
generally observed at lower concentrations <0.36
mg/m3.
Formaldehyde-
induced systemic
oxidative stress
• 'T* Malondialdehyde-dG adducts in whole blood DNA
from pathologists, compared to workers and students in
other science labs (Bono et al.. 2010), elevated plasma
malondialdehyde (MDA) and plasma p53 associated with
each other and with urinary formate concentrations
(imprecise marker of formaldehyde exposure) among
cosmetics workers (Attia et al.. 2014), and ^ 15-F2t
isoprostane levels in the urine of formaldehyde-exposed
workers (Romanazzi et al.. 2013)
• Inconclusive evidence for and against involvement by
genes that regulate oxidative stress in formaldehyde
associations with DNA damage risk in PBL in humans
• \1/ GSH, 'T* ROS, 'T* MDA in bone marrow, peripheral
blood mononuclear cells, liver, spleen and testes (Ye et
al.. 2013). although markers of oxidative stress were not
correlated with changes in DPX
Yes. Some
human data
available, and
results from
experimental
models are
presumed
relevant to
humans
without
evidence to
the contrary.
Limited human and rodent
evidence supports the
association between
formaldehyde exposure and
induction of oxidative stress
beyond the POE. While
biologically plausible, the
available evidence is
inadequate to determine
what role such oxidative
stress may play in LHP
carcinogenesis.
Formaldehyde-
induced changes in
the bone marrow
niche
• DNA adducts linked to inhaled (exogenous)
formaldehyde were not found in the bone marrow of
monkeys or rats in studies using highly sensitive
detection methods
• 'T* Bone marrow hyperplasia in rats from one study
(Kerns et al.. 1983: Battelle. 1982). unclear if other
results were negative or null (Sellakumar et al.. 1985):
(Kamata et al.. 1997) due to imprecise reporting
• Dose-related 'T* DPX in the bone marrow of formalin-
exposed mice (Ye et al.. 2013), although results may be
confounded by methanol coexposure
• HSPC mobilization and the BM-MSC niche is regulated by
cytokines, hormones and signals, which may be
distributed through circulation as a result of
inflammation; however, these effects have not been
directly evaluated following formaldehyde exposure
Yes.
Available data
are from
experimental
models
presumed
relevant to
humans.
The limited evidence
available is currently
inadequate to evaluate any
effect on bone marrow or
stromal cells following
formaldehyde exposure,
although such an effect
appears consistent with
current understanding of
hematopoiesis.
Evidence of
formaldehyde-
induced changes in
gene expression or
post-transcriptional
regulation in
peripheral blood
leukocytes or bone
marrow
• Limited study reported some statistically significant
differences in mRNA expression in either nasal or whole
blood samples from human volunteers associated with 5
day exposures up to 1 mg/m3 formaldehyde, however
study limitations prevent interpretation that results were
related to formaldehyde exposure (Zeller et al.. 2011) In
F344 rats, significant changes in both miRNA and mRNA
expression were reported in the nasal epithelium and
circulating white blood cells following inhalation
exposure to 2.5 mg/m3 formaldehyde for 1 or 4 weeks;
no changes were observed in miRNA expression in the
bone marrow, and mRNA was not evaluated (Rager et
al.. 2014): "Immune system/inflammation" markers were
enriched in both nasal tissue and WBCs at both time
points; and 'T* WBC miR-326 expression, associated with
Yes.
Available data
are from
experimental
models
presumed
relevant to
humans.
Limited rodent evidence
supports the association
between formaldehyde
exposure and epigenetic
effects in circulating
leukocytes; the available
human evidence is
inadequate. Insufficient
evidence is available to
determine what role
epigenetics may play in LHP
carcinogenesis.
This document is a draft for review purposes only and does not constitute Agency policy.
122 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Toxicological Review of Formaldehyde - Inhalation
Hypothesized
Mechanistic Event
Experimental Support for Mechanistic Event
Human
Relevance
Evidence Integration
considering Biological
Plausibility
bone marrow metastasis in other models (Valencia et
al.. 2013)
4.3.4. Overall Evidence Integration Judgments and Susceptibility for LHP Cancers
In human studies, robust evidence for myeloid leukemia and moderate evidence for multiple
myeloma supports a causal association with inhalation of formaldehyde (see Table 44). The assessment
of LHP cancers was based on epidemiological studies of groups with occupational formaldehyde levels
either in specific work settings (e.g., cohort studies) or in case-control studies. Aneuploidy in
chromosomes 1, 5, and 7 in circulating myeloid progenitor cells, considered a potential primary target
for LHP carcinogenesis, was associated with occupational formaldehyde exposure. The type of
aneuploidies observed in the formaldehyde-exposed asymptomatic human workers are also found in
patients with leukemia, specifically MDS and AML, as well as other worker cohorts at increased risk of
developing leukemias, which provides support for the plausibility of an association between chronic
formaldehyde exposure and leukemogenesis. Moreover, the strong and consistent evidence from a
large set of studies that observed mutagenicity in circulating leukocytes of formaldehyde-exposed
humans, specifically chromosomal aberrations (CA) and micronucleus (MN) formation, provides
additional evidence of biological plausibility for these cancer types. Further support is provided by
studies that observed perturbations to immune cell populations in peripheral blood associated with
formaldehyde exposure. In particular, decreases in RBCs, WBCs, and platelets, along with a 20%
decrease in CFU-GM colony formation in vitro were observed in the same exposed group (Zhang et al..
2010), suggesting both a decrease in the circulating numbers of mature RBCs and WBCs as well as
possible decreases in the replicative capacity of myeloblasts.
Increased LHP cancers have not been observed in a well-reported chronic rodent bioassay
involving inhalation exposure of both rats and mice to formaldehyde, nor in another rat bioassay that
failed to report the incidence of non-nasal neoplastic lesions. Further, positive associations with
leukemia have not been reported in rodent studies, although there are notable uncertainties in the
available data (i.e., increased bone marrow hyperplasia in rats; slight but uncertain increases in
lymphoma in mice; and a general lack of rigorous evaluation of non-respiratory tissues). Thus, there
appears to be a lack of support for the human epidemiological evidence from rodent bioassays, although
concordance across species is not necessarily expected (U.S. EPA. 2005a). The apparent lack of
consistency in results raises uncertainties about the currently available research results on these
diseases, including both how formaldehyde-induced LHP cancers might arise without substantial
distribution to target sites. Notably, the available animal evidence was judged as indeterminate and not
compelling evidence of no effect (see assessment Preface), as there are important uncertainties that
prevent such an interpretation. Thus, the animal evidence does not detract from the strength of the
This document is a draft for review purposes only and does not constitute Agency policy.
123 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Toxicological Review of Formaldehyde - Inhalation
association between formaldehyde exposure and myeloid leukemia (and related mechanistic changes)
in epidemiological studies (NRC, 2014b). Differences in physiology between humans and rodents, as
well as the relative insensitivity of rodent models to reflect the human pathogenesis of myeloid
leukemia, in particular, may together contribute to the potential lack of concordance between the
abundant human epidemiological data and the limited results available from rodent bioassay data.
Taken together, based on the robust and moderate human evidence for these cancers, the
evidence demonstrates that formaldehyde inhalation causes myeloid leukemia in humans given
appropriate exposure circumstances, and the evidence indicates that formaldehyde inhalation likely
causes multiple myeloma in humans given appropriate exposure circumstances. Separately, based on a
limited number of epidemiological studies and potentially relevant mechanistic evidence in exposed
humans, the evidence integration results in a judgment that the evidence suggests but is not sufficient
to infer that formaldehyde inhalation might cause Hodgkin lymphoma, given appropriate exposure
circumstances. While mechanisms for the induction of myeloid leukemia and multiple myeloma are yet
to be elucidated, they do not appear to require direct interactions between formaldehyde and bone
marrow constituents, and either are different in animals or the existing animal models tested thus far do
not characterize the complex process leading to cancers in exposed humans. These conclusions were
primarily based on epidemiological studies of groups with occupational formaldehyde exposure.
Notably, evidence exists to suggest a lack of concordance between chronic rodent bioassays and human
epidemiological evidence.
Table 38. Evidence integration summary for effects of formaldehyde inhalation on
LHP cancers
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence Integration
Judgment
Robust for Mveloid Leukemia, based on:
Human health effect studies:
Consistent increases in risk across a set of
high and medium confidence, independent
studies with varied study designs and
populations, demonstrating strong
associations (1.5- to 3-fold increase in risk),
clear dose-response relationships across
multiple measures of increasing exposure,
and a temporal relationship consistent with
causality (e.g., allowing time for induction,
latency, mortality)
Biological Plausibility (also of potential
relevance to LHP cancer types below):
Evidence from high and medium confidence
studies of exposed humans identifies relevant
mechanistic changes for cancers of the blood
such as myeloid leukemia, including impacts
Indeterminate for Any LHP
Cancer Type, based on:
Animal health effect studies:
Overall, the available data
do not provide evidence
supporting the development
of LHP cancers in a high
confidence chronic bioassay
of rats and mice, a second
medium confidence rat
bioassay, and two other low
confidence, long-term
exposure studies.
Biological Plausibility:
Although some potentially
relevant changes have been
observed in mechanistic
studies of exposed animals
• Relevance to humans:
The evidence for
carcinogenicity is from
studies in humans
• MOA: No MOA exists
to explain how
formaldehyde might
cause LHP cancers
without systemic
distribution (i.e.,
without direct
interactions of inhaled
formaldehyde with
constituents in bone
marrow tissue);
however, given the
mechanistic changes in
exposed humans, it is
reasonable to infer
The evidence
demonstrates that
formaldehyde
inhalation causes
myeloid leukemia in
humans given
appropriate exposure
circumstances
The evidence
indicates that
formaldehyde
inhalation likely
causes multiple
myeloma in humans
given appropriate
exposure
circumstances
This document is a draft for review purposes only and does not constitute Agency policy.
124 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
Human Evidence
Animal Evidence
Additional
Interpretations
Evidence Integration
Judgment
on peripheral immune cell populations (which
seem to be affected in a complex manner),
and elevated levels or severity of DNA or
chromosomal damage in circulating
myeloblasts and mature lymphocyte
populations. The DNA damage exhibits
aneugenic characteristics similar to that found
in humans with, or at increased risk for, AML
Moderate for Multiple Myeloma, based on:
Human health effect studies:
• Increases in risk across high, medium, and
low confidence studies, spanning an
approximate 1.2- to 4-fold increase in risk
with the highest confidence evidence
showing a 2-fold increase, and with very
limited evidence of a dose-response
relationship in 1 high confidence study
• Increases were limited to groups of people
who experienced high peak exposures, and
2 low confidence studies reported inverse
relationships with duration of exposure
Slight for Hodakin Lymphoma, based on:
Human health effect studies:
• Significantly increased risk in the highest
peak exposure group with a dose-response
relationship in 1 medium confidence study
of industrial workers
• An inconsistent pattern of risks across
studies, many with <5 exposed cases
(e.g., inflammatory and
immune changes in systemic
tissues and bone marrow
hyperplasia in rats), the
inability to detect DNA
adducts from exogenous
formaldehyde in bone
marrow, and evidence
related to genotoxicity (i.e.,
in systemic tissues) or other
more directly relevant
changes was weak. Overall,
the mechanistic data do not
suggest a judgment other
than indeterminate for LHP
cancers in animals
that an undefined
MOA is likely to involve
modulatory effects on
circulating immune
cells
• Potential
Vulnerabilities: There is
no evidence to
evaluate the potential
risk to sensitive
populations or
lifestages for lymphatic
leukemia and Hodgkin
lymphoma
• Other, the high survival
rate for lymphatic
leukemia and Hodgkin
lymphoma may
indicate that mortality
data are not a good
proxy for incidence
The evidence
suggests that
formaldehyde
inhalation might
cause Hodgkin
lymphoma, given
appropriate exposure
circumstances
These conclusions
were primarily based
on epidemiological
studies of groups
with occupational
formaldehyde
exposure. While
evidence exists to
suggest a lack of
concordance
between chronic
rodent bioassays and
human
epidemiological
evidence, notable
uncertainties prevent
an animal evidence
judgment of
compelling evidence
of no effect
1 4.4. WEIGHT-OF-EVIDENCE SUMMARY FOR CARCINOGENICITY
2 "Formaldehyde is Carcinogenic to Humans by the Inhalation Route of Exposure"
3 This conclusion is supported by several lines of evidence. Specifically, the hazard descriptor of
4 Carcinogenic to Humans is independently substantiated by two lines of evidence, namely the evidence
5 demonstrates that formaldehyde inhalation causes nasopharyngeal cancer and, separately, myeloid
6 leukemia, in exposed humans given appropriate exposure circumstances. In addition, this conclusion is
7 corroborated by several other lines of evidence, for which the evidence indicates that formaldehyde
8 inhalation likely causes that cancer type in exposed humans, namely sinonasal cancer,
This document is a draft for review purposes only and does not constitute Agency policy.
125 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
oropharyngeal/hypopharyngeal cancer, and multiple myeloma given appropriate exposure
circumstances.
These overall evidence integration judgments, as well as the strength of the human and animal
evidence (i.e., robust, moderate, slight, indeterminate), were based on the currently available evidence
using the approaches presented in the description of methods in the Introduction to this Overview
(Section 1), which included a consideration of mechanistic evidence when drawing each conclusion.
Note that, as the site-specific relationship of the animal data to the specific human cancer types
involved additional considerations, the inference regarding the relevance of the animal data to each
specific human cancer is presented herein as a component of the conclusions regarding the evidence for
an effect in animals.
4.4.1. Weight-of-evidence Narrative Summary
Two separate lines of evidence independently substantiate this conclusion:
Nasopharyngeal Cancer — The evidence demonstrates that formaldehyde inhalation causes
nasopharyngeal cancer in humans given appropriate exposure circumstances, based on robust
epidemiological evidence of an increased risk of the occurrence of nasopharyngeal cancers from studies
of groups with occupational formaldehyde levels in several geographic locations and diverse exposure
settings; robust evidence from long-term bioassays in two animal species providing consistent and
reliable evidence of nasal cancers following exposure; and reliable and consistent mechanistic evidence
in both animals and humans supporting causality. The nasopharynx, although not typically specified in
animal studies, is the region adjacent to the nasal cavity where the animal evidence was predominantly
observed, providing plausible coherence between the animal and human data (and thus, the animal
evidence is reflected as robust). The evidence is sufficient to conclude that a mutagenic MOA of
formaldehyde is operative in formaldehyde-induced nasopharyngeal carcinogenicity.
Myeloid Leukemia — The evidence demonstrates that formaldehyde inhalation causes myeloid
leukemia in humans given appropriate exposure circumstances, based on robust human evidence of an
increased risk of the occurrence of myeloid leukemia in epidemiological studies among different
occupational populations exposed to occupational formaldehyde levels representing diverse exposure
settings. The findings from the occupational cohorts are further supported by other studies of human
occupational exposure providing strong and coherent mechanistic evidence that formaldehyde exposure
is associated with additional endpoints directly relevant to LHP cancers, including an increased
prevalence of multiple markers of genotoxicity in peripheral blood lymphocytes and myeloid
progenitors. Indirect support is also provided by evidence of other systemic health effects (e.g.,
reproductive or developmental toxicity) and mechanistic evidence indicating changes in immune cell
populations and markers of inflammation (e.g., oxidative stress) in the peripheral blood of exposed
humans and animals, although the exact pattern of immune-related changes across studies and species
was difficult to interpret. Notably, leukemia was not increased after formaldehyde exposure in the two
This document is a draft for review purposes only and does not constitute Agency policy.
126 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
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
Toxicological Review of Formaldehyde - Inhalation
high or medium confidence rodent bioassays of chronic exposure, including one testing both sexes of
rats and mice, and the evidence for genotoxicity in the peripheral tissues of exposed rodents is weak,
providing indeterminate evidence of LHP cancers in animals. Taken together, it appears that
mechanisms yet to be elucidated that do not involve direct interactions of formaldehyde in the bone
marrow need to be considered, and that either the mechanistic pathways stimulated by formaldehyde
are different in animals or that the existing animal models tested thus far do not characterize the
disease process in humans for this cancer type.
Additional supporting evidence
Sinonasal Cancer — The evidence indicates that formaldehyde inhalation likely causes sinonasal
cancer in humans given appropriate exposure circumstances, based on moderate epidemiological
evidence from studies of groups exposed to occupational formaldehyde levels in several countries with
diverse exposure settings that found an increased risk of the occurrence of sinonasal cancers; robust
evidence from long-term bioassays in two animal species providing consistent and reliable evidence of
nasal cancers following exposure; and consistent and reliable mechanistic evidence in both humans and
animals supporting causality. Sinonasal cancers, although not typically specified in animal studies,
include cancers of the nasal cavity, where the animal evidence was predominantly observed. The
evidence is sufficient to conclude that a mutagenic MOA of formaldehyde is operative in formaldehyde-
induced sinonasal carcinogenicity.
Oropharyngeal/Hypopharyngeal Cancer — The evidence indicates that formaldehyde inhalation
likely causes oropharyngeal/hypopharyngeal cancer in humans given appropriate exposure
circumstances, based on moderate epidemiological evidence from studies of groups exposed to
occupational formaldehyde levels with diverse exposure settings that found an increased risk of the
occurrence of oropharyngeal/hypopharyngeal cancer, which is further supported by relevant
mechanistic changes (e.g., in buccal cells), and supporting animal evidence that is interpreted as slight to
moderate when incorporating human relevance. Specifically, although cancer site specificity across
species is not required (U.S. EPA. 2005a), and while the available animal bioassay evidence and
mechanistic data indicate that associations are biologically plausible, very few (if any) tumors would be
expected in comparable regions of the rodent respiratory tract with nasal breathing, and only at very
high formaldehyde concentrations. Thus, taking into consideration the toxicokinetics of inhaled
formaldehyde, oronasal breathing in humans (i.e., which would be expected to result in greater
distribution to these regions proximal to the mouth and nasopharynx), and the robust animal evidence
for relevant effects in the nasal cavity, the animal evidence integration judgment is presented as a lesser
amount of evidentiary support (i.e., slight-to-moderate) when integrating the evidence for human
cancers at these particular sites.
Multiple Myeloma — The evidence indicates that formaldehyde inhalation likely causes multiple
myeloma in humans given appropriate exposure circumstances, based on moderate human evidence of
This document is a draft for review purposes only and does not constitute Agency policy.
121 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
an increased risk of the occurrence of multiple myeloma in epidemiological studies of groups exposed to
occupational formaldehyde levels with diverse exposure settings, which is further supported by
mechanistic changes of potential relevance in systemic tissues of exposed humans. The animal evidence
is considered indeterminate, suggesting a need for additional study.
Other information
The remaining evidence relevant to evaluating the potential for formaldehyde inhalation to
cause cancer did not contribute to the carcinogenicity conclusion above, including systematic
evaluations of:
Hodgkin Lymphoma — slight epidemiological evidence suggested the possibility of an increased
risk of Hodgkin lymphoma. The animal evidence was indeterminate, and the mechanistic information
was not interpreted to alter these conclusions. Taken together, the evidence suggests but is not
sufficient to infer that formaldehyde exposure might have caused Hodgkin lymphoma.
Laryngeal Cancer — indeterminate-to-slight animal evidence suggested the possibility of an
increase in tumors at sites relevant to laryngeal cancer, primarily based on supportive mechanistic
changes (e.g., dysplasia at very high levels). Specifically, like the rationale provided for
oropharyngeal/hypopharyngeal cancers, tumors would be unexpected in the rodent larynx, which is
even farther from the portal of entry; thus, given the same considerations above, the coherence of the
robust animal evidence supporting nasal tumors was considered weak (i.e., presented as indeterminate-
to-slight) when integrating the evidence for human laryngeal cancers. The human evidence was
indeterminate. Overall, the evidence was inadequate to draw evidence integration judgments for this
cancer type.
Lymphatic Leukemia — all the evidence related to lymphatic leukemia was indeterminate; thus,
the evidence was inadequate to draw evidence integration judgments for this cancer type.
4.5. INHALATION UNIT RISK (IUR) FOR CARCINOGENICITY
Unit risk estimates for cancer were derived from different data sets available from both
epidemiological and experimental animal studies. In addition, an approach to bound low-dose cancer
risks from formaldehyde exposure using DNA adduct concentrations in nasal epithelium and bone
marrow from animal experiments and U.S. cancer incidence statistics (a "bottom-up" approach) is
summarized to provide some perspective on the uncertainty in extrapolating from high-dose animal
toxicology or human occupational data (Starr and Swenberg, 2016, 2013). Unit risk estimates could be
derived for two cancer types for which the evidence supporting a human health hazard was sufficiently
strong: nasal cancers (i.e., nasopharyngeal cancer in human studies; nasal squamous cell carcinoma in
experimental animal studies) and myeloid leukemia.
Specifically, unit risk estimates were derived based on dose-response modeling of mortality and
cumulative formaldehyde exposure for nasopharyngeal cancer (NPC) and myeloid leukemia in a human
This document is a draft for review purposes only and does not constitute Agency policy.
128 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
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
Toxicological Review of Formaldehyde - Inhalation
occupational cohort. Cumulative exposure, which incorporates both average concentration and the
duration of time over which exposure occurred, is generally the preferred metric for quantitative
estimates of lifetime risk from environmental exposure to carcinogens, and thus cumulative exposure
was chosen as the exposure metric for calculations in this assessment. The "true" exposure metric best
describing the biologically relevant delivered dose of formaldehyde is unknown. Few epidemiological
studies presented dose-response analyses based on cumulative measures of formaldehyde
concentration that could support the derivation of unit risk estimates; estimates were derived only for
NPC and myeloid leukemia.
In experimental animals, multiple approaches, including biologically based dose-response
(BBDR) modeling, and statistical time-to-tumor modeling, were used to derive unit risk estimates based
on data in rats. Results from the different approaches were evaluated and compared. In addition, other
approaches based on mechanistic hypotheses, including derivation of cRfCs based solely on cell
proliferation (one mechanism that contributes to cancer risk) and assessing the potential impacts of
endogenous formaldehyde concentration on dosimetric estimates, were explored quantitatively and
compared.
The unit risk estimates from the well-conducted human occupational study were preferred.
However, while the estimates for nasopharyngeal cancer and myeloid leukemia could be combined to
derive an IUR for formaldehyde, there is considerable scientific uncertainty in the data used to estimate
a unit risk for myeloid leukemia. Therefore, the unit risk estimate for myeloid leukemia is not included
in the IUR calculation in this draft assessment. The following sections outline the best supported
approaches for each cancer subtype based on the data currently available. The strengths and
weaknesses of the statistical approaches, as well as the rationale supporting each estimate, are
presented, including scientific judgments of confidence in the estimates for each cancer type.
4.5.1. Derivation of Cancer Unit Risk Estimates for Nasal Cancers
Derivation of a nasal cancer unit risk estimate based on human data
The quantitative analysis of nasal cancer from epidemiological studies is based on the NPC
results from the latest follow-up of the NCI cohort of industrial workers exposed to formaldehyde
(Beane Freeman et al.. 2013). While the evidence supporting a human health hazard from sinonasal
cancer and oropharyngeal/hypopharyngeal cancers from studies in occupational cohorts and
experimental animals also was sufficiently strong, it was not possible to derive a unit risk estimate for
these cancer types. Out of almost 14,000 deaths observed in the NCI cohort, there were 10 deaths from
NPC, 5 deaths from cancers of the nose and nasal sinus, and oropharyngeal/hypopharyngeal cancer was
not analyzed. Only the data for NPC could be modeled with adequate precision. The NCI cohort study is
the largest of the three independent industrial worker cohort studies [the other two being Meyers et al.
(2013) and Coggon et al. (2014)1 and, more importantly, it is the only one with sufficient individual
exposure data for dose-response modeling. In addition, the NCI study is the only one that used internal
This document is a draft for review purposes only and does not constitute Agency policy.
129 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Toxicological Review of Formaldehyde - Inhalation
comparisons rather than standardized mortality ratios (SMRs), thus minimizing the potential impact of
the healthy worker effect by addressing unmeasured confounding, which can bias effect estimates.
The NCI cohort consists of 25,619 workers (88% male) employed prior to 1966 in any of the 10
plants in the study. The most recent follow-up, based on 998,239 person-years of observation (through
2004) reported a total of 13,951 deaths (Beane Freeman et al.. 2013). Beane Freeman et al. (2013)
analyzed 10 deaths from NPC as well as deaths from other solid tumors. A detailed exposure
assessment was conducted for each worker in the NCI cohort, based on exposure estimates for different
jobs held and tasks performed (Stewart et al.. 1986). Exposure estimates were made using several
different metrics—peak exposure,11 average intensity, cumulative exposure, and duration of exposure.
Respirator use and exposures to formaldehyde-containing particulates and other chemicals were also
considered.
Dose-response modeling of data from the NCI cohort
The results of the internal analyses (i.e., comparing exposed workers to an internal referent
group of other workers in the cohort) of Beane Freeman et al. (2013) for NPC using the cumulative
exposure metric are presented in Table 45 (Tables 2-15 and 2-16 in the Toxicological Review). The
relative risks (RRs) were estimated using log-linear Poisson regression models stratified by calendar year,
age, sex, and race and adjusted for pay category. Beane Freeman et al. (2013) used a 15-year lag
interval in estimating exposures to account for a latency period for the development of solid cancers,
including NPCs. Models with alternative lag intervals (2-20 years) produced similar results. The NCI
investigators used the low-exposure category as the reference category to "minimize the impact of any
unmeasured confounding variables since nonexposed workers may differ from exposed workers with
respect to socioeconomic characteristics" (Hauptmann et al.. 2004). In this review, the nonexposed
person-years were included in the primary cancer risk analyses to be more inclusive of all the dose-
response data. The analyses adjusted for pay category, a measure of socioeconomic status, thus
possible SES differences between exposed and nonexposed were at least partially addressed. Final
results for the exposed person-years only are also presented for comparison.
Cumulative exposure was included as a continuous variable in the log-linear models analyzed by
Beane Freeman et al. (2013) (general model form: RR = e(BX, where (B represents the regression
coefficient and X is exposure). The regression coefficients are presented in Table 45.
nSome of the strongest exposure-response relationships in the NCI cohort (Beane Freeman et al.. 2013) (e.g., for NPC) were
observed for the peak exposure metric. It is not clear how to extrapolate RR estimates based on peak exposure estimates to
meaningful estimates of lifetime extra risk of cancer from continuous exposure to low environmental levels. If a short-term
(<15 minute) excursion above the 8-hour TWA concentration for a job was observed, or expected based on industrial hygiene
expertise, then that job was assigned to a peak exposure category, namely none, >0 to <0.5 ppm, 0.5 to <2.0 ppm, 2.0 to <4.0
ppm, or >4.0 ppm. Individual workers may have experienced these peak concentrations rarely, intermittently, or routinely, and
in jobs they held for a long time or only briefly. At a given time point, a worker's peak exposure estimate is the highest peak
exposure category ever attained by the worker. As such, this exposure metric is not interpretable in terms of a lifetime
exposure risk.
This document is a draft for review purposes only and does not constitute Agency policy.
130 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Toxicological Review of Formaldehyde - Inhalation
Table 39. Relative risk estimates for mortality from NPC (based on ICD code) and
regression coefficients from NCI log-linear trend test models3 by level of cumulative
formaldehyde exposure (ppm x years). Source: Beane Freeman et al. (2013)
Relative Risk Estimates
for Nasopharyngeal
Cancer
Rate Ratio (Number of Deaths)
p-trend, All
Person-years'5
p-trend,
Exposed
Person-years0
0
>0 to <1.5d
1.5 to <5.5
>5.5
1.87 (2)
1.0 (4)
0.86 (1)
2.94 (3)
0.07
0.06
Regression coefficients
for nasopharyngeal
cancer
Person-years
P (per ppm x year)®
Standard error (per ppm x year)®
All
0.04311
0.01865
Exposed only
0.0439
0.01852
aModels stratified by calendar year, age, sex, and race and adjusted for pay category; cumulative exposures
calculated using a 15-year lag interval for NPC and a 2-year lag interval for LHP cancer types.
bLikelihood ratio test (1 degree of freedom) of zero slope for formaldehyde exposure (as a continuous variable)
among all (nonexposed and exposed) person-years.
likelihood ratio test (1 degree of freedom) of zero slope for formaldehyde exposure (as a continuous variable)
among exposed person-years only.
Reference category for all categories.
eSource: Personal communications from Laura Beane Freeman to Jennifer Jinot (February 22, 2013 and February
21, 2014) and to John Whalan (August 26, 2009).
Prediction of lifetime extra risk of nasopharyngeal cancer mortality
To predict the extra risk of NPC mortality from environmental exposure to formaldehyde:
Extra risk = (Rx - Ro) 4- (1 - Ro)
where Rx is the lifetime risk in the exposed population and Ro is the lifetime risk in an
unexposed population (i.e., the background risk). Extra risk estimates were calculated using the p
regression coefficients and a life table program that accounts for competing causes of death.12 U.S. age-
specific 2010 all-cause mortality rates and 2000-201013 NPC mortality rates for all race and sex groups
combined were used to specify the all-cause and cause-specific background mortality rates in the life
table program. Risks were computed up to age 85 years because cause-specific mortality (and
incidence) rates for ages above that are less reliable. Conversions between occupational formaldehyde
exposures and continuous environmental exposures were made to account for differences in the
number of days exposed per year (240 versus 365) and in the amount of air inhaled per day (10 versus
20 m3). An adjustment was also made for the 15-year lag period. The reported standard errors for the
regression coefficients were used to compute the one-sided 95% upper confidence limits (UCLs) for the
extra risks based on a normal approximation.
12
This program is an adaptation of the approach that was previously used in BEIR IV, "Health Risks of Radon and Other Internally
Deposited Alpha Emitters." National Academy Press, Washington, DC, 1988, pp. 131-134.
13Typically, 5-year ranges are used as the basis for population cause-specific disease and mortality rates; a larger range is used
here to get better stability in the rates because NPC is a rare cancer.
This document is a draft for review purposes only and does not constitute Agency policy.
131 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
Consistent with EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the life table
program was used to estimate the exposure level (effective concentration [ECX]) and the associated
(one-sided) 95% lower confidence limit (LECX) corresponding to an extra risk of 0.05% (x = 0.0005).
Although EPA guidelines emphasize the use of exposure levels associated with a 10% extra risk level for
the POD for low-dose extrapolation, for epidemiological studies, this can result in the need to
extrapolate upward to risks well above those that were observed in the study populations. Thus, a 1%
extra risk level is typically used for epidemiological data. However, NPC has a very low background
mortality rate (e.g., lifetime background risk is about 0.00019); therefore, even a 1% extra risk (i.e., 0.01)
would be a large increase relative to the background risk. This is consistent with the fact that, even with
a large cohort followed for a long time, only 10 NPC deaths were observed in the NCI follow-up through
2004.14 Based on the life table program, the 1% level of risk for NPC mortality is associated with an RR
estimate of 53, a level substantially higher than was observed in the epidemiological study. A 0.05%
extra risk level yields an RR estimate of 3.6, which better reflects the RRs in the range of the data. Thus,
0.05% extra risk was selected for determination of the POD, and the LEC value corresponding to that risk
level was used as the POD.
Because formaldehyde is a mutagenic carcinogen and the weight of evidence supports the
conclusion that formaldehyde carcinogenicity for URT cancers can be attributed, at least in part, to a
mutagenic MOA, a linear low-dose extrapolation was performed in accordance with EPA's cancer
guidelines (U.S. EPA. 2005a). The ECooos, LECooos, and inhalation unit risk estimates for NPC mortality are
presented in Table 46.
Table 40. ECooos, LECooos, and unit risk estimates for nasopharyngeal cancer mortality
based on the Beane Freeman et al. (2013) log-linear trend analyses for cumulative
formaldehyde exposure
ECooos
LECooos
Unit Risk3
Unit Risk
Person-years
(ppm)
(ppm)
(per ppm)
(per mg/m3)
All
0.191
0.112
4.5 x 10"3
3.7 x 10"3
Exposed only
0.187
0.111
4.5 x 10"3
3.7 x 10"3
aUnit risk = 0.0005/LECooos.
Prediction of lifetime extra risk of nasopharyngeal cancer incidence
EPA cancer risk estimates are typically derived to represent a plausible upper bound on
increased risk of cancer incidence, as from experimental animal incidence data. Cancer data from
epidemiological studies are more often mortality data, as is the case in the NCI study. For cancers with
low survival rates, mortality-based estimates are reasonable approximations of cancer incidence risk.
14Eleven NPCs were reported on death certificates and included in NCI's SMR analyses, but one of these cases was apparently
misclassified on the death certificate, so only 10 cases were used to estimate the RRs in the internal comparison analyses
(Beane Freeman et al.. 2013).
This document is a draft for review purposes only and does not constitute Agency policy.
132 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Toxicological Review of Formaldehyde - Inhalation
However, for NPC, the survival rate is substantial [51% at 5 years in the 1990s in the United States,
according to Lee and Ko (2005)1, and incidence-based risks are preferred because EPA is concerned with
cancer occurrence, not just cancer mortality.
Therefore, an additional calculation was done using the same regression coefficients provided
by Dr. Beane Freeman but with age-specific NPC incidence rates for 2000-2010 from NCI's Surveillance,
Epidemiology, and End Results (SEER) Program in place of the NPC mortality rates in the life table
program (www.seer.cancer.gov). The incidence-based calculation relies on the reasonable assumptions
that NPC incidence and mortality have the same dose-response relationship for formaldehyde exposure
and that the incidence data are for first occurrences of NPC or that relapses provide a negligible
contribution. The calculation also takes advantage of the fact that NPC incidence rates are negligible
compared with the all-cause mortality rates and thus no special adjustment to the population at risk to
account for live individuals who have been diagnosed with NPC is necessary.
The resulting ECooos, LECooos, and inhalation unit risk estimates for NPC incidence are presented
in Table 47. The unit risk estimate for cancer incidence is two-fold higher than the corresponding
mortality-based estimate, for all person-years, reflecting the high survival rates for NPC.
Table 41. ECooos, LECooos, and unit risk estimates for nasopharyngeal cancer incidence
based on the Beane Freeman et al. (2013) log-linear trend analyses for cumulative
formaldehyde exposure
Person-years
ECooos (ppm)
LECooos (ppm)
Unit Risk3 (per ppm)
Unit Risk (per mg/m3)
All
0.0942
0.0550
9.1 x 10"3
7.4 x 10"3
Exposed only
0.0925
0.0546
9.2 x 10"3
7.5 x 10"3
aUnit risk = 0.0005/LECooos-
The preferred estimate for the inhalation cancer unit risk for NPC is the estimate of 9.1 x 10"3
per ppm derived using incidence rates for the cause-specific background rates, for all person-years. The
results from the exposed person-years are essentially identical.
Because NPC is a rare cancer in the United States, with a relatively low number of cases
occurring per year, a rough calculation was done to ensure that the unit risk estimate derived for NPC
incidence is not implausible in comparison to actual case numbers. For example, assuming an average
constant lifetime formaldehyde exposure level of 5 ppb for the U.S. population (probably at the very low
end of potential lifetime averages) the inhalation unit risk estimate for NPC equates to a lifetime extra
risk estimate of 4.6 x 10"5. Assuming an average lifetime of 75 years (this is not EPA's default average
lifetime of 70 years but rather a value more representative of actual demographic data) and a U.S.
population of 300,000,000, this lifetime extra risk estimate suggests a crude upper-bound estimate of
180 incident cases of NPC attributable to formaldehyde exposure per year. Alternatively, for 20 ppb
(probably toward the upper end of potential lifetime averages), the calculation suggests a crude upper-
This document is a draft for review purposes only and does not constitute Agency policy.
133 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Toxicological Review of Formaldehyde - Inhalation
bound estimate of 730 incident cases of NPC per year. Both upper-bound estimates, using different
assumed lifetime exposure levels, are well below the estimated 2,300 total incident NPC cases per year
calculated from the SEER NPC incidence rate of 0.75/100,000.15,16
Dose-response modeling of nasal SCC tumor incidence in the F344 rat
Dose-response analyses of cancer risk were calculated using conventional multistage-Weibull
time-to-tumor modeling and a biologically based clonal expansion model of cancer, both based on nasal
squamous cell carcinoma (SCC) incidence data from laboratory bioassays using F344 rats. The
biologically based modeling was informed by a large body of mechanistic data on cell replication, DNA
protein cross-link (DPX) and DNA monoadduct formation, and dosimetry modeling of formaldehyde flux
to local tissue, and was therefore considered useful to provide potential information on the shape of the
dose-response curve as well as the interpretation and extrapolation of results from the rat bioassays to
humans.
These models were employed to derive multiple PODs and corresponding human equivalent
concentrations. Unit risks derived by straight line extrapolation from a point of departure as well as a
candidate RfC (cRfC) derived from these human equivalent concentrations were presented, with the
cRfC interpreted as the concentration below which nasal cancers arising from increased cell proliferation
due to cytotoxicity are unlikely to occur (some researchers have argued that protection against this
putative precursor event is sufficient to prevent a cancer response). cRfCs for this mechanism
contributing to cancer were also derived from modeling of data on cell proliferation and basal
hyperplasia in F344 rats and Wistar rats, respectively.
Approaches to modeling the animal nasal tumor incidence
An increased incidence of nasal SCCs was seen in two long-term bioassays using F344 rats
(Monticello et al.. 1996; Kerns et al.. 1983; Battelle, 1982), with similar incidences between the two
studies even though they were conducted 13 years apart (and similar incidences between males and
females in Kerns et al., which tested both sexes). Therefore, for greater power in dose-response
analysis, these data were combined (see Table 48).
15This crude NPC incidence rate is similar to a published NPC incidence rate for the United States of 0.7/100,000 person-years
(Lee and Ko. 2005). The age-adjusted NPC incidence rate from SEER was also 0.75/100,000.
16With the application of age-dependent adjustment factors, the lifetime unit risk estimate for NPC would increase by a factor
of 1.42, and the crude upper-bound estimates of the incident cases per year attributable to formaldehyde exposure would
similarly increase by a factor of 1.42. The resulting adjusted upper-bound estimates of 260 and 1,030 for 5- and 20-ppb
exposure levels, respectively, are still well below the estimated total number of 2,300 incident cases per year in the United
States.
This document is a draft for review purposes only and does not constitute Agency policy.
134 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Toxicological Review of Formaldehyde - Inhalation
Table 42. F344 rat nasal cancer data
Formaldehyde Exposure Levels
Incidence of SCC Tumors
References
0, 0.7, 2.0, 6.01, 9.93, and 14.96 ppm
(0, 0.86, 2.5, 7.4,12.2, and 18.4 mg/mB)
0/341, 0/107, 0/353, 3/343, 22/103,162/386
(Kerns et al.. 1983: Battelle,
1982) & (Monticello et al..
1996) (combined bioassavs)
Several models (described below) were used to calculate BMCs and the corresponding BMCLs
(95% lower confidence bounds on dose) at a benchmark response (BMR) level at the lowest end of the
range of the observed data ((U.S. EPA. 2012); see Table 49). Benchmark concentrations at the 0.005 as
well as 0.01 extra risk levels were determined with the BBDR models. The BMCs and corresponding
BMCLs were then converted to their human equivalent concentrations (HECs) based on formaldehyde
flux to the nasal tissue obtained using computational fluid dynamic (CFD) modeling in the rat and human
(Kimbell et al.. 2001b). The HEC corresponding to a particular benchmark level in the rat was then
calculated by assuming that continuous lifetime exposure to a given steady-state flux of formaldehyde
leads to equivalent risk of nasal cancer across species. This extrapolation included an adjustment to the
laboratory exposure regimen for continuous exposure (multiplication by 6/24 x 5/7).
Table 43. Benchmark concentrations and human equivalents using formaldehyde flux
to nasal tissue as a dose-metric
Models
Rat Benchmark Cone, (ppm)
Human Equivalent Cone. (ppm)a
Extra Riskb
Extra Riskb
0.005
0.01
0.05
0.1
Dose
metric3
0.005
0.01
0.05
0.1
Multistage Weibull time-to-
tumor
EC
LEC
4.28
3.57
5.93
5.52
6.84
6.41
Flux
EC
LEC
0.35
0.30
0.49
0.46
0.57
0.53
Weibull with threshold
Schlosser et al. (2003)
EC
LEC
5.91
5.58
6.12
5.94
6.40
6.22
Flux
EC
LEC
0.75
0.71
0.78
0.76
0.82
0.79
Rat BBDR "model l"c
EC
LEC
4.99d
4.95
5.37d
5.19
Flux
EC
LEC
0.42
0.41
0.45
0.43
Rat BBDR "model 2"c
EC
LEC
5.41
5.25
5.75
5.59
Flux
EC
LEC
0.45
0.44
0.48
0.46
EC=BMC and LEC=BMCL at the specified extra risk; these abbreviations are used here to facilitate comparisons to the modeling
of the human data.
aThe human equivalent benchmark concentrations decrease by a factor of 1.4 if flux estimates based on Schroeter et al. (2014)
are used instead of Kimbell et al. (2001b).
bThe BMR of 0.005 is lower than the value of 0.0085 corresponding to the lowest observed tumor response, corrected for
survival, and was used only with the BBDR modeling because these models incorporate precursor response data related to
cellular proliferation. Because benchmark concentrations at 0.005 and 0.010 extra risk levels were reported, they were not
calculated at the higher levels when BBDR modeling was used.
cSee text for a description of models 1 and 2.
dBenchmark concentrations corresponding to the hockey-stick model in Conolly et al. (2003) as discerned from Figure 5 of their
paper were ECoos = 4.84 ppm and EC0i = 5.48 ppm. LEC levels could not be estimated since confidence bounds were not
reported by these authors.
This document is a draft for review purposes only and does not constitute Agency policy.
135 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Toxicological Review of Formaldehyde - Inhalation
Weibull time-to-tumor modeling
Because higher exposures were associated with both earlier tumor occurrence and increased
mortality in the rats, methods that can reflect the influence of competing risks and intercurrent
mortality on site-specific tumor incidence rates were preferred. For this reason, EPA used the
multistage Weibull time-to-tumor model (Portierand Bailer. 1989; Krewski et al.. 1983), which (a)
modeled the replicate animal data, (b) included the exact time of observation of the tumors and
therefore gave appropriate weight to the amount of time each animal was on study without a tumor,
and (c) acknowledged earlier tumor incidence with increasing dose level.
Weibull modeling of the grouped incidence data assuming a threshold in dose
This assessment also presents results from statistical modeling of the same data by Schlosser et
al. (2003) in Table 49. These authors did not carry out a time-to-tumor analysis of the individual animal
data but applied a Kaplan-Meier survival adjustment of the grouped incidence data. The best fit in
Schlosser et al. (2003) was obtained with the polynomial and Weibull (shown) models for the tumor
incidence data with a nonzero intercept (threshold) on the dose axis.
Biologically based dose-response modeling
A biologically based dose-response (BBDR) time-to-tumor model for the formaldehyde-induced
rat nasal tumors was available (Conolly et al.. 2003; CUT, 1999). This model consisted of interfacing
dosimetry models for formaldehyde and formaldehyde-induced DPX in the rat nasal passages (Kimbell et
al.. 2001a; Kimbell et al.. 2001b; Conolly et al.. 2000) with two-stage clonal expansion (TSCE) models for
predicting the probability of occurrence of nasal SCC (Conolly et al.. 2003). Formaldehyde-induced
changes in cell replication and DPX concentrations were considered a function of local formaldehyde
flux to each region of nasal tissue as predicted by computational fluid dynamics (CFD) modeling on
anatomically accurate representations of the nasal passages of a single F344 rat. DPX tissue
concentrations were calculated in Conolly et al. (2003) using a physiologically based pharmacokinetic
model developed in Conolly et al. (2000). In addition to the data from the two tumor bioassays, these
authors included all historical control data on 7,684 animals obtained from National Toxicology Program
F344 rat inhalation and oral bioassays. Conolly et al. (2003) characterized the dose-response for cell
replication rates as a J-shaped curve, indicating that at low exposure concentrations cell division rates
decreased below that determined for the unexposed case. In addition, these authors used a hockey
stick shaped curve such that the dose-response for cell division rates remained changed from the
baseline only at 6 ppm and higher exposure concentrations. This resulted in more conservative
estimates of risk when used in the clonal expansion model for cancer. The BBDR models for the rat used
here for the purpose of calculating benchmark concentrations were based on Conolly et al. (2003) with
the following modifications.
This document is a draft for review purposes only and does not constitute Agency policy.
136 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Toxicological Review of Formaldehyde - Inhalation
"Model 1" presented in Table 49 was based on the more conservative, "hockey stick", model in
Conolly et al. (2003), with one critical modification. Conolly et al. (2003) added historical control data
from all NTP studies to the concurrent controls, whereas the model used here included historical data
from only the inhalation route of exposure.17
"Model 2" presented in Table 49 made major modifications to Conolly et al. (2003) in regard to
model structure as well as values for input parameters: (1) the shape of the dose-response for the
division rates of normal cells as a function of formaldehyde flux, aN(flux), was monotone increasing
without a threshold in dose, and obtained by fitting the cell replication data for 13-week exposure
duration in Monticello et al. (1996); (2) the dose-response for the division rates of initiated cells was
assumed to be a sigmoid-shaped curve, increasing monotonically with flux; (3) the death rate of an
initiated cell was assumed to be proportional to its division rate at all formaldehyde flux values and
given by Pi(flux) = K-ai(flux), where k is an unknown estimated constant of proportionality; and (4) as in
model 1, only historical controls from NTP inhalation studies were added to the concurrent controls.
Estimated impact of a revised dosimetry model incorporating endogenous formaldehyde
Schroeter et al. (2014) revised the dosimetry model of Kimbell et al., used for the flux estimates
in the table above, to include endogenous formaldehyde production and to explicitly model
formaldehyde pharmacokinetics in the respiratory mucosa. EPA estimated the extent to which the
results in the above table change if flux estimates from Schroeter et al. were used. The average flux
over non-squamous regions of the rat nose was roughly one-third18 of that in the human based on the
dosimetry in Schroeter et al. (2014) in which endogenous formaldehyde was taken into account,
compared to a ratio of roughly one-half based on the dosimetry in Kimbell et al. (2001b). As a result, the
benchmark concentrations calculated in the above table were not appreciably altered (decreasing by
roughly a factor of 1.419) if the revised dosimetry model by Schroeter et al. was applied.
Threshold-based RfC approach for precursor lesion data in the rat: cell proliferation and hyperplasia
The highly curvilinear and steeply increasing dose-responses for DPX formation and cell
proliferation, concomitant with the highly nonlinear observed tumor incidence in the F344 rat, have led
to mechanistic arguments that formaldehyde's nasal carcinogenicity arises only in response to
significant cytotoxicity-induced regenerative cell proliferation (Conolly et al.. 2002; Morgan. 1997). In
particular, Conolly et al. (2003) and Conolly et al. (2004) inferred from BBDR modeling results that the
direct mutagenicity of formaldehyde is not an important contributor compared to the importance of
17ln accordance with generally accepted practice when using historical controls (Haseman. 1995):(Peddada et al.. 2007).
ls0.33 at 0.1 ppm, 0.32 at 1 ppm.
19This is only approximate because the various components of the BBDR modeling were not recalibrated or rerun in light of the
revised flux estimates for both species. Furthermore, the above estimate is for resting inspiration, whereas the human flux
values in this assessment pertain to an equal apportionment of sleeping, sitting, and light activity levels.
This document is a draft for review purposes only and does not constitute Agency policy.
137 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
cytotoxicity-induced cell proliferation in explaining the rat tumor response. Thus, candidate RfCs (cRfCs)
derived from available experimental data relevant to this mechanism were presented and discussed.
The interpretation of these cRfCs was that they may help identify formaldehyde concentrations
below which it is unlikely that hyperplastic lesions develop or that cancers arising from cytotoxicity-
induced regenerative cell proliferation occur. Cytotoxicity-induced regenerative cell proliferation and
the subsequent development of hyperplastic lesions were considered precursor events that, if protected
against, would prevent these mechanisms from contributing to the cancer response. Below these cRfCs,
formaldehyde may still increase the risk of nasal or upper-respiratory cancer through direct
mutagenicity or other mechanisms, but the magnitude of cancer risk may be significantly lower due to
the absence of increased cellular proliferation or hyperplasia.
Significantly increased cell proliferation and hyperplasia (increased cellular proliferation that is
identified to be pathologically "abnormal" in tissues) have been observed in response to exposure to
formaldehyde, and these data were used to estimate benchmark PODs to calculate cRfCs. Schlosser et
al. (2003) modeled the dose-response for cellular proliferation using labeling index data reported by
Monticello et al. (Monticello et al.. 1996) and calculated a value of 0.44 ppm for the HEC corresponding
to the BMCLoi (rat BMC and BMCL of 4.79 and 3.57 ppm, respectively) using dose-response functions
that allowed for a threshold in dose.20
Although Monticello et al. (1996) represented the longest duration cell proliferation study
available that included a range of exposure durations and nasal regions, five other medium or high
confidence cellular proliferation studies testing formaldehyde exposure durations of 12-13 weeks were
also available. Based on the findings from these studies, reasonable alternatives or adjustments to the
Schlosser et al. (2003) estimate were presented in the assessment. The range of results from the
various cell labeling data attempt to represent some key uncertainties; these include the single-day time
frame (the last day of exposure) over which cell labeling was carried out (a methodological constraint
intrinsic to all available cellular labeling studies) and the specific averaging approach employed in
Schlosser et al. (2003), where the labeling index was weighted by exposure durations and averaged over
several locations on the F344 rat nose. Such a time-weighted averaging underweights early exposures
that may have contributed significantly to carcinogenesis (note: the few studies that investigated latent
effects in rats did observe an increased tumor incidence at 1-2+ years following high-level formaldehyde
exposure lasting only ~13 weeks (Woutersen et al.. 1989; Feron et al.. 1988). It was estimated that the
data from these additional studies would result in benchmark levels that were roughly 2- to 3-fold
lower.
Separately, EPA developed a benchmark POD based on modeling the incidence of basal
hyperplasia reported by Woutersen et al. (1989) in a 28-month bioassay using Wistar rats. The BMC and
BMCL at the benchmark response of 0.1 extra risk21 were 1.68 and 1.108 ppm, respectively. The HEC
20They also modeled with functions that were constrained to pass through the origin, but BMCL values are not reported.
21A 10% BMR was chosen to reflect "minimal adversity," consistent with the noncancer respiratory pathology modeling.
This document is a draft for review purposes only and does not constitute Agency policy.
138 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Toxicological Review of Formaldehyde - Inhalation
corresponding to the BMCL was 0.1609 ppm when adjusted for continuous human lifetime exposure,
which was roughly three times lower than the HEC derived from the time-weighted averaged labeling
index by Schlosser et al. (2003). As a point of comparison, this value is roughly similar to the LECooos
derived from EPA's modeling of the NPC risk from the NCI epidemiology data.
Based on these estimates, proliferation-based cRfCs were estimated as follows:
1) The HEC derived from Schlosser et al. (2003) was 0.44 ppm (0.54 mg/m3); the other cell-labeling
studies indicated a 2-fold or 3-fold lower adjustment to this value, i.e., values of 0.27 and 0.18
mg/m3, respectively. Applying a UF = 3 to reflect other uncertainties in extrapolating from
animals to humans and a UF = 10 to account for human variability (total UF = 30) resulted in
cRfCs based on cell proliferation data that ranged from 0.006 mg/m3 to 0.018 mg/m3.
2) The hyperplasia data from Woutersen et al. (1989) resulted in an HEC of 0.1609 ppm (0.1979
mg/m3); applying the UFs described above (total UF=30), the cRfC = 0.007 mg/m3.
As noted earlier, it has been argued that the rat nasal tumors can be quantitatively explained
based solely on formaldehyde's cytotoxic potential. In accordance with this point of view, a cRfC
estimated from benchmark concentrations derived using the two rat BBDR models in Table 49 may be a
reasonable approximation for the dose at which there is no regenerative cell proliferative contribution
to the nasal or upper-respiratory cancer response. A cRfC of 0.017 mg/m3 may be obtained in this
manner corresponding to the average HEC estimated using the two models at a benchmark response of
0.005 extra risk and reduced by an uncertainty factor of 30. This value is encompassed by the range of
0.006-0.018 mg/m3 obtained for the proliferation-based cRfCs above.
However, the direct mutagenicity of formaldehyde plays a key role in its carcinogenicity.
Cytogenetic effects in occupational studies and the formation of DPX in experimental animals have been
reported at exposures well below those considered to be cytotoxic (e.g., approximately 0.7-2 ppm in
rats). In addition, genotoxicity is itself thought to be one of the mechanisms by which formaldehyde
exerts its cytotoxic action, arguing against a demarcation of one MOA over the other along the
concentration axis. Overall, because formaldehyde-induced tumors are not fully explained only by
indirect mutagenicity (i.e., due to regenerative cell proliferation) at any exposure, and since other
modes of action also contribute to the tumor response, the use of an RfC approach was not preferred.
Extrapolation using a human BBDR model and evaluation of uncertainties
Subsequent to their model for predicting the risk of rat nasal cancer, Conolly et al. (2004)
developed a corresponding BBDR model for humans that Conolly et al. used for the purpose of
extrapolating the observed risk in the rat to human exposures. This model was conceptually very similar
to the rat two-stage clonal expansion model in Conolly et al. (2003) but did not incorporate any data on
human responses to formaldehyde exposure. Specifically, it used DPX concentrations and values of local
formaldehyde flux to the tissue obtained from the PBPK and CFD models and incorporated a more
detailed biological hypothesis and mechanistic data than are normally employed in modeling cancer risk.
This document is a draft for review purposes only and does not constitute Agency policy.
139 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Toxicological Review of Formaldehyde - Inhalation
For perspective, Table 50 presents continuous human lifetime extra risk estimates from the
Conolly et al. (2004) model following inhalation exposure to formaldehyde concentrations of 1.0 ppb-
1.0 ppm, in comparison to human risk estimates derived from EPA's modeling of the NPC mortality in
the NCI occupational data. Conolly et al. (2004) developed two models: the "optimal model" in Table 50
refers to derivations using the best fit, a J-shaped curve, to the dose-response for the time-weighted
averaged cell-labeling data in rats such that values at 0.7 ppm and 2.0 ppm were below the control
value; the "conservative model" was derived using a hockey stick-, rather than J-, shape in the low dose
portion (i.e., values at 0.7 and 2.0 ppm were the same as the control). In calculating risk estimates,
Conolly et al. (2004) used a statistical upper bound of the model parameter (kmu)22 (which related DPX
to the probability of mutation per cell generation) and used maximum likelihood (MLE) values for all
other model parameters. Since there is uncertainty inherent to using any statistical model to
extrapolate outside the range of observed data, the relevant question in the context of using the BBDR
modeling for such extrapolation is whether it decreased uncertainty in extrapolating risk (i.e., as
compared to default approaches) or if, by explicitly identifying the sources of uncertainty, the BBDR
modeling pointed to approaches and data needs that may have helped reduce the uncertainty.
Table 44. BBDR model estimated extra risk of SCC in human respiratory tract
compared with EPA's modeling of extra risk of NPC from the human occupational data
Formaldehyde Levels:
0.001 ppm
0.01 ppm
0.10 ppma
1.0 ppm
Conolly et al. (2004) "J-
shape optimal model"
-1.0 x 10"5
-1.0 x 10"4
-9.1 x 10"4
-5.0 x 10"3
Conolly et al. (2004)
"hockey stick
conservative model"
+3.1 x 10"s
+3.2 x 10"7
+3.5 x 10"6
+2.7 x 10"4
EPA analysis of NCI
NPC, MLE (UCL)b
+1.2 x 10"6 (+2.1 x 10"6)
+1.3 x 10"5 (+2.3 x 10"5)
+1.8 x 10"4 (+4.1 x 10"4)
+2.7 xlO1 (+8.7 xlO1)
aThe mortality-based ECooos (LECooos) from the NCI epidemiology data correspond roughly to 0.2 (0.1) ppm.
bMLE = maximum likelihood estimate; UCL=95% upper confidence limit.
Uncertainties and confidence in the BBDR modeling and extrapolation
The assessment included a careful evaluation of the level of confidence and sources of
uncertainties in different components of both the rat and human BBDR models. Of the potential issues
identified, those related to replication rates of normal and initiated cells, and the use of historical
control animals were found to have major impacts on qualitative and quantitative conclusions from the
22The model estimated MLE value for kmu was found to be zero, leading to the inference by the authors that formaldehyde's
direct mutagenic action is not relevant to carcinogenicity in the rat or human, and that the observed tumor response in the rat
can be explained on the basis of regenerative cellular proliferation to cell injury.
This document is a draft for review purposes only and does not constitute Agency policy.
140 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
modeling. In particular, modeling results were unstable in response to slight perturbations in the
assumed values for the division rates for initiated cells, and there are currently no data of any kind even
in rats to inform the effect of formaldehyde on the kinetics of initiated cells. The model was also
extremely sensitive to the inclusion of historical control animals. Because SCC in the nose is a rare
tumor, Conolly et al. (2004. 2003) included in their model control rats from all NTP cancer bioassays.
When the NTP control data were restricted to those animals from NTP inhalation studies, the upper
bound human risk estimate obtained by Conolly et al. (2004) (i.e., with everything else in their modeling
retained unchanged) was increased by 50-fold (Crump et al.. 2008). If only concurrent controls were
used, the model for extrapolation of risk to humans (the human BBDR model) became numerically
unstable (i.e., the MLE and upper-bound estimates of risk became infinite). Subramaniam et al. (2007)
and Crump et al. (2008) provide details. The human extrapolation BBDR model exhibited extreme
uncertainty at all exposure concentrations, above as well as below the human equivalent concentrations
that were calculated in Table 49 [see (2009; Crump et al.. 2008)1.
Unit risk estimates based on animal data, considering confidence in the available models
Overall, use of biologically based modeling allowed utilization of various data, including
mechanistic information, in an integrated manner for modeling the incidence of nasal SCCs in F344 rats
and for deriving benchmark levels for extrapolation. In this way, the rat BBDR modeling improved the
dose-response modeling of the observed nasal cancers in the F344 rat, and multiple BBDR model
implementations provided similar estimates of risk and confidence bounds in the general range of the
observed rat tumor incidence data. Therefore, the rat BBDR models were used to calculate benchmark
concentrations for points of departure (PODs). In addition, given the reasonable confidence in flux
estimates derived from the rat and human CFD models, model-derived formaldehyde flux values were
used in deriving human equivalent concentrations corresponding to these PODs and candidate unit risk
estimates using these values were calculated.
However, it was determined that the human BBDR modeling was extremely uncertain and did
not provide robust measures of human nasal SCC risk at any exposure concentration. Therefore, the
human BBDR modeling was not used to directly calculate risk at human exposure scenarios.
The assessment presents strong arguments in support of a low dose linear extrapolation from
the POD. Given formaldehyde's direct mutagenic potential, following the procedures in EPA's cancer
guidelines (U.S. EPA. 2005a) for when the knowledge of the MOA does not support an alternative
approach, a low dose linear approach was used to predict low-dose formaldehyde cancer risk from the
rat data. Extrapolation was carried out as a straight line drawn to the origin from the HEC corresponding
to the BMDL. Unit risks were calculated using several modeling approaches, including modifications to
the rat BBDR model, as shown in Table 51 below. The unit risks corresponding to BMRs at the 0.005 or
0.01 extra risk levels spanned a remarkably tight range of 0.01-0.03 per ppm.
This document is a draft for review purposes only and does not constitute Agency policy.
141 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Table 45. Unit risk estimates derived from benchmark estimates using animal data
2 and formaldehyde flux as dose-metric
Models
Unit Risk Estimates (1/ppm)
LECoos
LECoi
LECos
LEClo
Weibull with threshold3 Schlosser et al.
(2003)
0.014
0.066
0.127
Multistage Weibull time-to-tumor
0.033
0.109
0.189
Rat BBDR "model 1"
0.012
0.023
Rat BBDR "model 2"
0.011
0.022
Estimates using steady-state DPX as a dose metric were identical.
Note = values were not estimated for vacant cells.
3 Selection of a unit risk estimate for nasal cancers
4 The unit risk estimates derived using the available human and animal data on nasal cancers are
5 similar (see Table 52), with the human estimate being only slightly lower than those values estimated
6 using rat bioassay and mechanistic data.
7 Table 46. Comparison and basis of unit risk estimates for NPC in humans and nasal
8 SCCs in rats
Human NPC Estimate
Animal Nasal Cancer Estimate
Study/Endpoint
Beane-Freeman et al., 2013
(NCI industrial cohort): NPC mortality
Kerns et al., 1983; Monticello et al., 1996:
Incidence of nasal squamous cell carcinoma
in rats
Model features
Estimation of inhalation unit risk using
Poisson regression model and life table
analysis:
• U.S national incidence data
• Regression coefficients from log-linear
models of nasopharyngeal (NPC)
mortality (exposed and unexposed
workers)
• Linear low-dose extrapolation from
LEC
Multiple mechanistic and statistical models,
including BBDR modeling, used for modeling
tumor incidence. Mechanistic information
included:
• Dosimetric (CFD) modelling of
formaldehyde flux to rat, monkey and
human airway lining
• PBPK model for rats incorporating dose-
response data on DNA-protein crosslinks
• Site-specific cell labeling measurements in
nose
Linear low-dose extrapolation was carried
out from human equivalent dose at BMCL
POD
95% lower bound on concentration at
0.05% incidence (approx. 0.05 ppm)
95% lower bound on concentration at 0.5%
incidence (approx. 0.2 ppm)
Unit risk estimate3
7.4 x 10"3 per mg/m3
(9.1 x 10"3 per ppm)
8.9 x 10 3 to 1.8 x 10 2 per mg/m3
(1.1 x 10"2 to 2.2 x 10"2 per ppm)
aNote that these estimates are provided for comparison purposes and do not represent ADAF-adjusted values.
This document is a draft for review purposes only and does not constitute Agency policy.
142 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Toxicological Review of Formaldehyde - Inhalation
Ultimately, it was determined that the human data provided a more appropriate basis for
estimating human nasal cancer risk than did the rodent data, given that a well-conducted
epidemiological study was available with appropriate quantitative analyses. However, candidate unit
risks in Table 52 at 0.005 extra risk were comparable to that derived using the occupational data on
nasopharyngeal cancers (see Section 2.2.1 in the Toxicological Review). Because the unit risk estimates
from the human data were preferred, the rodent-based estimates were not adjusted for the assumed
increased early-life susceptibility arising from the determination of a mutagenic MOA for URT cancers;
however, if the rodent-based estimates were to be used, ADAFs should be applied (U.S. EPA. 2005b).
As previously described, using the human NPC data, a plausible upper bound lifetime extra
cancer mortality unit risk of 4.5 x 10"3 per ppm (3.6 x 10"6 per ng/m3) of continuous formaldehyde
exposure was estimated using a life table program and linear low-dose extrapolation of the excess NPC
mortality and log-linear modeling results (for cumulative exposure) reported in a high confidence
occupational epidemiological study (based on 10 NPC deaths). Applying the same regression coefficient
and life table program to background NPC incidence rates yielded a lifetime extra cancer (incidence) unit
risk estimate of 9.1 x 10"3 per ppm (7.4 x 10"6 per ng/m3).
The weight of evidence supports the conclusion that formaldehyde carcinogenicity for URT
cancers such as NPC can be attributed, at least in part, to a mutagenic MOA. Therefore, because there
were no chemical-specific data to evaluate susceptibility of different lifestages, increased early-life
susceptibility was assumed for NPC and age-dependent adjustment factors (ADAFs) were applied,
consistent with EPA's Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
Carcinogens (U.S. EPA. 2005b).
The application of ADAFs resulted in a lifetime unit risk estimate of 1.1 x 10"5 per ng/m3 (1.3 x
10"2 per ppm) for NPC incidence, adjusted for postulated increased early-life susceptibility, assuming a
70-year lifetime and constant exposure across age groups.
Uncertainties and confidence in the selected unit risk estimate for nasal cancers
The strengths and uncertainties in the unit risk estimate for NPC incidence are summarized in
Table 53. One of the largest sources of uncertainty in the NPC estimate has to do with the rarity of the
cancer and, thus, the small number of exposed cases (n = 8) that informed the dose-response analysis.
It is important to note that, although a unit risk estimate could only be calculated for NPC (for which an
evidence integration judgment of evidence demonstrates was drawn), the systematic evaluation of
evidence on URT cancers also resulted in a judgment that the evidence indicates (based on studies in
occupational cohorts and animals) that inhalation of formaldehyde likely causes sinonasal cancer and
oropharyngeal/hypopharyngeal cancer, given relevant exposure circumstances.
This document is a draft for review purposes only and does not constitute Agency policy.
143 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Toxicological Review of Formaldehyde - Inhalation
Table 47. Strengths and uncertainties in the cancer type-specific unit risk estimate for
NPC
Strengths
Uncertainties
• IIIR estimated from data that
are directly relevant to
humans.
• Based on the results of a
large, high confidence
epidemiological study
involving multiple industries
with detailed, individual
cumulative exposure
estimates and allowance for
cancer latency.
• Low-dose linear extrapolation
is supported by a mutagenic
MOA (i.e., not a default).
• Similar unit risk estimates
derived using rat bioassay
and mechanistic data on
nasal cancers.
• NPC is a very rare cancer. This study followed more than 25,000 workers
for over 40 years and observed a statistically significant increase in relative
risk associated with the highest category of average exposure intensity,
however, only 10 cases occurred. The small number of deaths creates
uncertainties for the dose-response modeling (borderline model fit for
cumulative exposure including exposed and unexposed person-years,
p = 0.07).
• Uncertainty about optimal exposure metric(s). Cumulative exposure is the
standard metric used for unit risk estimates. Use of cumulative exposure
assumes equal importance of concentration and duration on cancer
incidence; yet, associations with peak exposure in epidemiological studies
and the nonlinear shape of the dose-response from animal bioassays
suggest greater influence of concentration.
• Although statistically significant increases in risk for NPC were reported by
multiple studies for several metrics of exposure (duration, cumulative, time
since first exposure, peak), the relationship with cumulative exposure in
the study used for IUR derivation was less precise (p-trend = 0.07 based on
the regression coefficient for the continuous model).
• Some uncertainty in the low-dose extrapolation is introduced based on the
potential for endogenous formaldehyde to reduce the uptake of the inhaled gas at
low doses, as demonstrated in modeling efforts bvSchroeter et al. (2014) and
Campbell et al. (2020).
Based on the attendant strengths and uncertainties outlined above, there is medium confidence
in the unit risk estimate for NPC incidence. The greatest uncertainty was related to the small number of
cases that contributed to the statistical analysis and resulting imprecision in modeling the shape of the
dose-response curve.
4.5.2. Derivation of Cancer Unit Risk Estimates for Myeloid Leukemia
The quantitative analyses of myeloid leukemia are based on results from the latest follow-up of
the NCI cohort of industrial workers exposed to formaldehyde (Beane Freeman et al.. 2009). While the
evidence for a human health hazard from multiple myeloma from studies in occupational cohorts also
was sufficiently strong, a unit risk estimate was not derived for this cancer type because no association
with cumulative exposure was indicated by the NCI study. Although no association was indicated for
cumulative exposure and myeloid leukemia in this study, the combination of myeloid leukemia and
other/unspecified leukemia was marginally associated (p = 0.1) with cumulative formaldehyde exposure.
The evaluation of this combined group is supported by analyses by NCI during the 1980s and 90s that
compared diagnoses on death certificates to original hospital diagnoses and found that as many as 50%
This document is a draft for review purposes only and does not constitute Agency policy.
144 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Toxicological Review of Formaldehyde - Inhalation
of deaths classified as other or unspecified leukemia were originally diagnosed as myeloid leukemia
(Percy et al.. 1990; Percy et al.. 1981).23
Derivation of a myeloid leukemia unit risk estimate based on human data
Choice of epidemiology study
Similar to the unit risk estimate for NPC, the estimate for myeloid leukemia is based on results
from the latest follow-up of the NCI cohort of industrial workers exposed to formaldehyde (Beane
Freeman et al.. 2009), the largest (25,619 workers) of the three independent industrial worker cohort
studies and the only one with sufficient individual exposure data for dose-response modeling. Beane
Freeman et al. (2009) conducted dose-response analyses of 123 deaths attributed to leukemia and
leukemia subtypes, as well as deaths from other LHP malignancies. As previously described, this well-
conducted study is the only one that used internal comparisons rather than standardized mortality
ratios (reducing the impact of potential unmeasured confounding), and it included a detailed exposure
assessment conducted for each worker based on exposure estimates for different jobs held and tasks
performed (Stewart et al.. 1986). and exposure estimates were made using several different metrics-
peak exposure, average intensity, cumulative exposure, and duration of exposure.
Dose-response modeling of data from the NCI cohort
The results of the internal analyses (i.e., comparing exposed workers to an internal referent
group of other workers in the cohort) of Beane Freeman et al. (2009) for LHP cancer types using the
cumulative exposure metric are presented in Table 54 (Tables 2-15 and 2-16 in the Toxicological
Review). The relative risks (RRs) were estimated using log-linear Poisson regression models stratified by
calendar year, age, sex, and race and adjusted for pay category. A two-year lag interval was used to
determine exposures to account for a latency period for LHP cancers. For all cancer types, the NCI
investigators used the low-exposure category as the reference category to "minimize the impact of any
unmeasured confounding variables since nonexposed workers may differ from exposed workers with
respect to socioeconomic characteristics" (Hauptmann et al.. 2004). In this review, the nonexposed
person-years were included in the primary cancer risk analyses to be more inclusive of all the dose-
response data. The analyses adjusted for pay category, a measure of socioeconomic status, thus
possible SES differences between exposed and nonexposed were at least partially addressed. Final
results for the exposed person-years only are also presented for comparison.
23ln the Percy et al. (1990; 1981) studies, only about 10% of leukemia deaths were classified as "other or unspecified" based on
hospital diagnoses (versus 29% from death certificates in the Beane Freeman et al. (2009) study), and 51% (Percy et al.. 1981)
and 53% (Percy et al.. 1990) of leukemia deaths were myeloid leukemias based on hospital diagnoses (versus 39% from death
certificates in the Beane Freeman et al. (2009) study), suggesting that about a third or more of the "other or unspecified"
leukemia deaths in the Beane Freeman et al. (2009) study were probably myeloid leukemias. Percy et al. (1990) reported in
their study that "Of the nearly 600 deaths from leukemia NOS [other or unspecified] nearly 50% were originally diagnosed as
myeloid... Obviously myeloid leukemia is grossly underreported on death certificates."
This document is a draft for review purposes only and does not constitute Agency policy.
145 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 Cumulative exposure was included as a continuous variable in the log-linear models (general
2 model form: RR = e|3X, where (B represents the regression coefficient and X is exposure). The regression
3 coefficients are presented in Table 54.
4 Table 48. Relative risk estimates for mortality from leukemia (based on ICD codes)
5 and regression coefficients from NCI log-linear trend test models3 by level of
6 cumulative formaldehyde exposure (ppm x years). Source: Beane Freeman et al.
7 (2009)
Relative Risk Estimates
Cancer Type
Rate Ratio (Number of Deaths)
p-trend, All
Person-years'3
p-trend, Exposed
Person-yearsc
0
>0 to <1.5d
1.5 to <5.5
>5.5
Leukemia
0.53 (7)
1.0(63)
0.96 (24)
1.11 (29)
0.08
0.12
Myeloid leukemia
0.61(4)
1.0(26)
0.82 (8)
1.02 (10)
0.44
>0.50
Other/unspecified
leukemia
0.77 (2)
1.0(15)
1.65 (10)
1.44 (9)
0.13
0.15
Regression Coefficients
Cancer type
Person-years
p (per ppm x year)'
Standard Error (per ppm x year)'
Leukemia
All
0.01246
0.006421
Exposed only
0.01131
0.00661
Myeloid leukemia
All
0.009908
0.01191
Exposed only
0.008182
0.01249
Myeloid leukemia plus
other/unspecified leukemia6
All
0.01408
0.007706
Exposed only
0.01315
0.007914
aModels stratified by calendar year, age, sex, and race and adjusted for pay category; cumulative exposures
calculated using a 15-year lag interval for NPC and a 2-year lag interval for LHP cancer types.
bLikelihood ratio test (1 degree of freedom) of zero slope for formaldehyde exposure (as a continuous variable)
among all (nonexposed and exposed) person-years.
likelihood ratio test (1 degree of freedom) of zero slope for formaldehyde exposure (as a continuous variable)
among exposed person-years only.
Reference category for all categories.
ep-trend values for the myeloid and other/unspecified leukemia categories combined are 0.10 for all person-years
and 0.13 for exposed person-years only.
'Source: Personal communications from Laura Beane Freeman to Jennifer Jinot (February 22, 2013 and February
21, 2014) and to John Whalan (August 26, 2009).
8 Approaches used for quantitative risk assessment of myeloid leukemia
9 EPA explored several approaches for deriving a unit risk estimate for myeloid leukemia based on
10 cumulative exposure. A standard approach for deriving the unit risk estimate was considered using the
11 regression coefficient for myeloid leukemia and cumulative exposure; however, the p-value (0.44) for
12 that regression coefficient was far from 0.05, indicating a poor model fit. The poor model fit could be
This document is a draft for review purposes only and does not constitute Agency policy.
146 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
due, at least in part, to inadequate statistical power, likely exacerbated by the underreporting of
myeloid leukemia deaths. The regression coefficient for all person-years for myeloid leukemia is only
slightly lower than that for all leukemia (0.0099 and 0.0125 per ppm-years, respectively). The
association with all leukemia cancer had a lower p-value of 0.08 and should include all the myeloid
leukemia deaths, both specified and unspecified. The "other/unspecified" leukemias comprise a
sizeable portion of all leukemia deaths (almost 30%) in the cohort and presumably include a good
proportion of unclassified myeloid leukemias (Percy et al.. 1990; Percy et al.. 1981). To address this
underreporting, two additional approaches for deriving a unit risk estimate for myeloid leukemia were
considered.
One approach involved using the all leukemia grouping.24 Use of the all leukemia background
rates in the life table calculations (described in more detail below) might inflate the unit risk estimate for
myeloid leukemia by increasing the background risk relative to which the formaldehyde-related risks are
calculated. However, the inclusion of any leukemia subtypes not related to formaldehyde exposure
should theoretically dampen the dose-response relationship (lowering the regression coefficient)
relative to that for all the myeloid leukemias alone; thus, this should mitigate at least some of the effect
of using the all leukemia background rates.
The preferred approach involved using a combined grouping of the myeloid leukemia and
other/unspecified leukemias subcategories. The myeloid and other/unspecified leukemias grouping had
a stronger association with cumulative exposure (p-trend = 0.10 for all person-years) in the Beane
Freeman et al. (Beane Freeman et al.. 2009) study than did myeloid leukemia alone and it captures the
unclassified myeloid leukemias with the least inclusion of nonmyeloid leukemias. There is likely more
uncertainty associated with the background rates for the other/unspecified leukemias than for the
specified myeloid and lymphocytic leukemia subtypes; however, the benefits of focusing on the myeloid
plus other/unspecified leukemias rather than the broader "all leukemias" grouping in attempting to be
more inclusive of all the myeloid leukemias were deemed to outweigh any additional uncertainty
associated with the background rates. Although the unit risk estimate based on the preferred approach
of using myeloid plus other/unspecified leukemias inevitably includes some nonmyeloid leukemias, it is
considered the best approach for deriving a unit risk estimate for myeloid leukemia specifically.25
24The all leukemia category includes all 123 leukemias observed in the cohort. Of these, 48 (39.0%) were myeloid, 37 (30.1%)
were lymphoid, and 36 (29.3%) were other/unspecified; the remaining 2 (1.6%) were monocytic leukemias (ICD-8 code 206).
25Although the inclusion of cancer subtypes not necessarily causally associated with the chemical exposure in the grouping of
cancers represented in the regression coefficient and the corresponding background rates for the life table analysis is overt
here, it is not uncommon that, due to data limitations, unit risk estimates based on human data reflect cancer groupings
broader than what might be strictly causally associated with the chemical exposure (e.g., all leukemias or all lung cancers). As
noted in the text, any inclusion of unassociated cancer subtypes in the derivation of the regression coefficient should
theoretically attenuate the coefficient in a manner that would offset the use of the unassociated subtypes in the background
rates in the life table analysis.
This document is a draft for review purposes only and does not constitute Agency policy.
147 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Toxicological Review of Formaldehyde - Inhalation
Prediction of lifetime extra risk of myeloid leukemia mortality and incidence
Unit risk estimates for myeloid leukemia mortality (and incidence) were calculated from the
regression results using the different approaches discussed above and the same general methodology
described for the NPC mortality estimates with the exception of the use of a 2-year lag period, as
selected by Beane Freeman et al. (2009). Mortality (and incidence) rates from the time frame 2006-
2010 were used in the life table program. Although the background mortality rates of leukemia are
higher (lifetime risk of 0.0062 according to the life table analysis) than those of NPC, the 1% extra risk
level typically used as the basis for the POD for epidemiological data still corresponds to an RR estimate
(2.5) that would be above the highest categorical result reported, even after adjusting the RR estimates
upward relative to the 0-exposure group (because our primary analyses include the nonexposed
workers). A 0.5% extra risk level yields an RR estimate of 1.8, which better corresponds to the RRs in the
range of the data. Thus, the LEC value corresponding to 0.5% extra risk (LECoos) was selected for the
POD for all leukemia and for myeloid leukemia and myeloid plus other/unspecified leukemias, which
have lower background rates than all leukemia (lifetime risks of 0.0031 and 0.0046, respectively).
There are insufficient data to establish the MOA for formaldehyde-induced myeloid leukemia;
thus, linear low-dose extrapolation was performed as the default approach, in accordance with EPA's
cancer guidelines (U.S. EPA. 2005a). The ECoos, LECoos, and inhalation unit risk estimates for myeloid plus
other/unspecified leukemia mortality are presented in Table 55.
Table 49. ECoos, LECoos, and unit risk estimates for myeloid plus other/unspecified
leukemia mortality based on log-linear trend analyses of cumulative formaldehyde
exposure data from the Beane Freeman et al. (2009) study
Person-years
ECoos (ppm)
LECoos (ppm)
Unit Riska(per ppm)
Unit Risk(per mg/m3)
All
0.253
0.133
1
o
T—1
X
00
cn
3.1 x 10~2
Exposed only
0.269
0.135
3.7 x 10"2
3.0 x 10"2
aUnit risk = 0.005/LEC0os
All leukemia and myeloid leukemia have substantial survival rates;26 thus, it is preferable to
derive incidence estimates. Unit risk estimates for leukemia incidences were calculated as described
above for the NPC incidence estimates. The incidence-based calculation relies on the assumptions that
incidence and mortality for these leukemia subtype groupings have the same dose-response relationship
for formaldehyde exposure and that the incidence data are for first occurrences of the cancers or that
relapses provide a negligible contribution. The first assumption is more uncertain for all leukemia,
myeloid leukemia, and myeloid plus other/unspecified leukemias than it was for NPC because these are
26Survival rates were 55.0% at 5 years for all leukemia [http://seer.cancer.gov/statfacts/html/leuks.htmll. 23.4% at 5 years for
acute myeloid leukemia [http://seer.cancer.gov/statfacts/html/amvl.htmll. and 59.1% at 5 years for chronic myeloid leukemia
[http://seer.cancer.gov/statfacts/html/cmvl.htmll based on 2002-2009 SEER data.
This document is a draft for review purposes only and does not constitute Agency policy.
148 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 groupings of subtypes with quite different survival rates (e.g., see footnote 26). The incidence-based
2 calculation also takes advantage of the fact that incidence rates for these cancer types are negligible
3 compared with the all-cause mortality rates and thus no special adjustment to the population at risk to
4 account for live individuals who have been diagnosed with these cancers is necessary.
5 The ECoos, LECoos, and inhalation unit risk estimates for myeloid plus other/unspecified leukemia
6 incidence are presented in Table 56. The incidence unit risk estimate is about 10% higher than the
7 mortality estimate. This difference is lower than the ~24% increase that would have been seen for
8 specified myeloid leukemias alone (see the LECoosS in Table 57). This is because the difference between
9 age-specific mortality and incidence rates for the other/unspecified leukemias is not very large, and for
10 some age groups the mortality rates are actually larger than the incidence rates. This irregularity is to
11 be expected for "other/unspecified" classifications because greater attention is given to diagnosing
12 incident leukemia cases than to accounting for causes of death, so one would anticipate less
13 underreporting of myeloid leukemias as incident cases than as causes of death on death certificates.
14 Table 50. ECoos, LECoos, and unit risk estimates for myeloid plus other/unspecified
15 leukemia incidence based on Beane Freeman et al. (2009) log-linear trend analyses for
16 cumulative formaldehyde exposure
Person-years
ECoos (ppm)
LECoos (ppm)
Unit Risk3 (per ppm)
Unit Risk (per mg/m3)
All
0.224
0.118
1
O
T—1
X
pm)
ppm)
Unit Risk Estimate
(per ppm)a
Unit Risk Estimate
(per mg/m3)
Incidence
Mortality
(Incidence)
(Incidence)
Myeloid leukemia
0.378
0.127
0.468
0.157
3.9 x 10"2
3.2 x 10"2
All leukemia
0.156
0.229
5.9 x 10"2
4.8 x 10"2
This document is a draft for review purposes only and does not constitute Agency policy.
149 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
0.0846
0.124
Myeloid + Other/Unspecified
0.224
0.118
0.253
0.133
4.2 x 10"2
3.4 x lO"2
aUnit risk estimate = 0.005/(LECoo5 for incidence).
incidence background rates also include monocytic leukemia, but that contribution is negligible.
1 Thus, the preferred unit risk estimate for myeloid leukemia is the estimate of 4.2 x 10"2 per
2 ppm.27 The unit risk estimate calculated using the exposed person-years only is essentially
3 indistinguishable from the preferred estimate using all person-years (see Table 56). The unit risk
4 estimates from the other approaches considered are fairly close, with the unit risk estimate based on
5 the myeloid leukemia category's being virtually identical to the preferred estimate based on myeloid
6 plus other/unspecified leukemias, and the estimate based on all leukemia being somewhat greater (see
7 Table 58).
8 Table 58 summarizes some of the key information comparing the different approaches
9 considered for the derivation of the unit risk estimate for myeloid leukemia.
10 Table 52. Dose-response modeling (all person-years) and (incidence) unit risk
11 estimate derivation results for different leukemia groupings - shaded estimate is
12 preferred
Deaths
Regression
SE
Unit Risk
Unit Risk
in NCI
Coefficient
(per ppm
Estimate
Estimate
Cancer Grouping
Cohort
(per ppm x year)
x year)
p-value
(per ppm)
(per mg/m3)
Myeloid leukemia
48
0.009908
0.01191
0.44
3.9 x 10"2
3.2 x 10"2
All leukemia
123
0.01246
0.006421
0.08
5.9 x 10"2
4.8 x 10"2
Myeloid +
Other/Unspecified
84a
0.01408
0.007706
0.10
4.2 x 10"2
3.4 x lO"2
leukemias
aThis is the sum of the leukemias classified as myeloid and those classified as "other/unspecified." At least 70-80%
of this number are expected to be myeloid leukemias, assuming that a third to a half of leukemias not otherwise
specified on death certificates are myeloid leukemias, as discussed above.
13 Selection of a unit risk estimate for myeloid leukemia
14 The best estimate that can be derived for myeloid leukemia was calculated using human
15 occupational data from the NCI industrial cohort (Beane Freeman et al.. 2009). As previously described,
16 a plausible upper bound lifetime extra cancer mortality unit risk of 3.8 x 10"2 per ppm (3.1 x 10"5 per
17 Hg/m3) of continuous formaldehyde exposure was estimated using a life table program and linear low-
27Comparable to calculations done for NPC, a rough calculation was done to ensure that the unit risk estimate derived for
myeloid leukemia incidence is not implausible in comparison to actual case numbers. For example, assuming an average
constant lifetime formaldehyde exposure level of 20 ppb for the U.S. population, the inhalation unit risk estimate for myeloid
(and other/unspecified) leukemia equates to a lifetime extra risk estimate of 8.4 x 10~4. Assuming an average lifetime of 75
years (this is not EPA's default average lifetime of 70 years but rather a value more representative of actual demographic data)
and a U.S. population of 300,000,000, this lifetime extra risk estimate suggests a crude upper-bound estimate of 3,400 incident
cases of myeloid leukemia attributable to formaldehyde exposure per year. This upper-bound estimate is well below the
estimated 17,100 total incident myeloid leukemia (not including other/unspecified leukemias) cases per year.
This document is a draft for review purposes only and does not constitute Agency policy.
150 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Toxicological Review of Formaldehyde - Inhalation
dose extrapolation of the excess myeloid plus other/unspecified leukemia mortality and log-linear
modeling results (for cumulative exposure) reported in a well-conducted occupational epidemiological
study (based on 84 deaths). Applying the same regression coefficient and life table program to
background myeloid leukemia incidence rates yielded a lifetime extra cancer (incidence) unit risk
estimate of 4.2 x 10"2 per ppm (3.4 x 10"5 per ng/m3).
Since there is no knowledge as to whether a mutagenic MOA might be operative for
formaldehyde-induced myeloid leukemia, no adjustments for increased early-life susceptibility (i.e.,
application of age-dependent adjustment factors) were made for myeloid leukemia, consistent with
EPA's Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S.
EPA. 2005b).
Uncertainties and confidence in the selected unit risk estimate for myeloid leukemia
The strengths and uncertainties in the unit risk estimate for myeloid leukemia incidence are
summarized in Table 59. The primary uncertainty in this estimate relates to the complexities in the
study-specific data for cumulative formaldehyde exposure and mortality from myeloid leukemia.
Table 53. Strengths and uncertainties in the cancer type-specific unit risk estimate for
myeloid leukemia
Strengths
Uncertainties
• IUR estimated from data that are
directly relevant to humans.
• Based on the results of a large,
high confidence epidemiological
study involving multiple
industries with detailed,
individual cumulative exposure
estimates and allowance for
cancer latency.
• Moderate number of deaths to
model (n = 84).
Uncertainties with a potentially greater impact:
• Although the dose-response relationship with peak exposure was
marginally significant (p = 0.07), and statistically significant associations
were reported for several metrics of exposure in other studies, the
reported relationship with cumulative exposure showed a non-
significant, small increase in risk for myeloid leukemia (based on the
regression coefficient for the continuous model), potentially due in part
to misclassification of myeloid leukemia cases.
• The association with cumulative exposure was stronger for the other/
unspecified grouping of leukemia diagnoses (n = 36) than for myeloid
leukemia alone (n = 48). Although a sizable proportion of this category
is assumed to include myeloid leukemia cases, the stronger association
is surprising given the more heterogeneous set of leukemia cases in this
category, some presumably not associated with formaldehyde
exposure. Hence, the association would be expected to be attenuated.
• Uncertainty about optimal exposure metric(s). Use of cumulative
exposure assumes equal importance of concentration and duration on
cancer incidence. The specific metrics analyzed differed across studies,
and the results of the NCI study were not completely consistent with
those of other studies (associated only with peak exposure).
Uncertainties likely to have a minor impact:
This document is a draft for review purposes only and does not constitute Agency policy.
151 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Toxicological Review of Formaldehyde - Inhalation
Strengths
Uncertainties
• Grouping of myeloid leukemias used for exposure-response modeling
includes non-myeloid leukemias.
• Borderline model fit for myeloid plus other/unspecified leukemias
(p = 0.1) and uncertain shape of exposure-response function.
Based on the attendant strengths and uncertainties outlined above, there is low confidence in
the unit risk estimate for myeloid leukemia incidence. This uncertainty is discussed further in the
summary section below. However, given the judgment that the available evidence demonstrates that
formaldehyde inhalation causes myeloid leukemia in humans given appropriate exposure circumstances,
and the associated public health burden that it poses (e.g., myeloid leukemia is far more prevalent than
NPC), EPA thoroughly considered the complexity in the data and used an innovative approach to derive
and present potential unit risk estimates for myeloid leukemia. A charge question will be provided for
the peer-review panel regarding the development of a unit risk estimate for myeloid leukemia and
asking for advice about how, if at all, the unit risk estimate might inform the quantification of risk for
cancer.
4.5.3. Estimates of Cancer Risk based on "Bottom-up" Approach
Starr and Swenberg (2016) and Swenberg et al. (2011) developed an approach to bound low-
dose human cancer dose-response from formaldehyde exposure in a manner that only uses information
regarding: (1) background incidence of the target tumors (nasopharyngeal cancers, leukemia, and
Hodgkin lymphoma) in the U.S. population, (2) assumptions as to the target tissue for a key event
interaction with formaldehyde for each type of tumor, and (3) measures of internal formaldehyde tissue
dose in laboratory animals derived from either endogenously produced formaldehyde or from
exogenous exposure to formaldehyde. The tissue dose measures are based on highly sensitive
measurements in rats and monkeys of formaldehyde-induced DNA adducts (Yu et al.. 2015; Lu et al..
2011; Moeller et al.. 2011; Lu et al., 2010).
To develop these bounding estimates, the authors attributed the background tumor incidences
to endogenous formaldehyde in the presumed target tissues (as measured by endogenous adducts).
The approach assumed extra cancer incidence to be linearly related to exogenous adduct levels, with a
slope equal to the ratio of background tumor incidence to background tissue dose of endogenous
formaldehyde.
Risk estimates from this approach are claimed by the authors to produce conservative upper
bounds, primarily because (a) the method attributes all the background risks of specific cancers to
endogenous formaldehyde although other environmental factors might also contribute to these
cancers, (b) lower confidence bounds on measured adduct levels are used, and (c) the above linear
relationship is assumed.
This document is a draft for review purposes only and does not constitute Agency policy.
152 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Toxicological Review of Formaldehyde - Inhalation
Swenberg et al. (2011) and Starr and Swenberg (2016) then compared these values with the risk
estimates in EPA's 2010 draft Toxicological Review, which were obtained by linearly extrapolating to
lower doses from a point of departure (a lower bound on the concentration associated with the
benchmark response) derived by dose-response modeling of the epidemiological data. When adduct
data from rats were used, the estimates Swenberg and Starr derived at 1 ppm exposure concentration
were 2.67 x 10"4 for nasal cancer based on Yu et al. (2015) and were at most 12.6 xlO4 for leukemia
(based on the limit of detection, LOD, from Lu et al. (2010), since no exogenous adducts were detected
in bone marrow). In monkeys (Yu et al.. 2015), the Swenberg and Starr bottom-up estimates were
2.69 x 10"4 for NPC and were less than 1.24 x 10 s for leukemia. In comparison, the EPA upper-bound
risk estimates were higher than the adduct-based upper-bound estimates by 40-fold for NPC and at least
45-fold (rat adduct data) or over 45,000-fold (monkey adduct data) for leukemia.
EPA concludes that the bottom-up approach does not necessarily provide an upper bound on
the slope of the dose-response at low exogenous exposures, primarily because the ratio of background
tumor incidence to internal endogenous concentration at the true target tissue may underpredict the
slope of the dose-response above that endogenous concentration. This is further discussed in Crump et
al. (2014). Furthermore, the approach assumes direct interaction of inhaled formaldehyde with a
particular target tissue; if other sites of interaction and mechanisms are involved, the measures of DNA
adducts in a specific tissue could lead to underestimates of the cancer potency when utilizing the
"bottom-up" approach. Nonetheless, the bottom-up approach, which uses cancer incidence in the
general population and is independent of the tumor dose-response data, can potentially provide some
perspective on the likely contribution of a specific mode-of-action and the uncertainty in risk estimates
derived from occupational exposures or derived by extrapolating downward from higher dose animal
data where other phenomena may be occurring.
4.5.4. Summary of Unit Risk Estimates and the Preferred Estimate for Inhalation Unit Risk
As discussed previously, the NPC unit risk estimate based on data from the human occupational
epidemiology study of the NCI updated by Beane-Freeman et al. (2013) was preferred over estimates
based on rodent cancer bioassay data. The best estimate for myeloid leukemia was also derived from
the human occupational epidemiology study of the NCI updated by Beane-Freeman et al. (2009). These
estimates are presented in Table 60.
Table 54. Summary of inhalation unit risk estimates from occupational
epidemiological studies3'13
Cancer Type
Preferred Unit Risk
Estimate
(ppm"1)
ADAF-adjusted
Unit Risk Estimate
(ppm"1)
Preferred Unit Risk
Estimate
((Hg/m3)"1)
ADAF-adjusted Unit
Risk Estimate
((Hg/m3)"1)
Nasopharyngeal
0.0079°
0.013
6.4 x 10^c
1.1 X 10"5
This document is a draft for review purposes only and does not constitute Agency policy.
153 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Toxicological Review of Formaldehyde - Inhalation
Myeloid leukemia
0.042
NAd
3.4 x 10"5
NAd
aThe inhalation unit risk estimate is typically expressed as the (upper-bound) increase in cancer risk estimated for
an exposure increase of 1 ng/m3.
bThe unit risk estimates are all for cancer incidence.
cAdult-based (rescaled) unit risk estimate for NPC intended for the application of ADAFs.
dNA = not applicable; no ADAF adjustment is recommended for myeloid leukemia.
However, the data reported for myeloid leukemia (Beane Freeman et al.. 2009) are complex and
there are reasons for and against the use of these data in the derivation of the inhalation unit risk (IUR).
Given the judgment that the available evidence demonstrates that formaldehyde inhalation causes
myeloid leukemia in humans given appropriate exposure circumstances, and the associated public
health burden that it poses (e.g., myeloid leukemia is far more prevalent than NPC), EPA thoroughly
considered the complexity in the data and used an innovative approach to derive and present potential
unit risk estimates for myeloid leukemia. Some important uncertainties are discussed in greater detail
below.
• Despite the quality of the literature base for the formaldehyde assessment and the judgment that
the available evidence demonstrates that formaldehyde inhalation causes myeloid leukemia in
humans given appropriate exposure circumstances, the only study suitable for dose-response
quantification for myeloid leukemia may be viewed as insufficient for developing a quantitative
estimate of risk with an acceptable level of confidence.
o The Beane-Freeman study failed to observe an association between cumulative formaldehyde
exposure and myeloid leukemia (p = 0.44), despite a reasonable number of cases (n = 48) and
adequate follow up. The peak exposure metric was marginally associated (p = 0.07). This result
raises questions about the relative importance of the intensity of exposure and duration in the
association of myeloid leukemia mortality. On the other hand, myeloid leukemia mortality
increased with time since first exposure, cumulative exposure, and exposure duration in two
other occupational cohorts (garment workers and embalmers).
o The available animal studies do not provide any compelling evidence for an association between
formaldehyde inhalation and myeloid leukemia. Thus, there are no other data that can be used
to support the POD estimate that can be derived from the only suitable human study.
• Analyses from NCI comparing causes of death recorded on death certificates with original diagnoses
in hospital records suggest a misclassification of myeloid leukemia cases (n = 48), with a significant
proportion reported as "other/unspecified" (n = 36).
o In the Percy et al. (1990; 1981) studies, only about 10% of leukemia deaths were classified as
"other or unspecified" based on hospital diagnoses (versus 29% from death certificates in the
Beane Freeman et al. (2009) study), and 51% (Percy et al.. 1981) and 53% (Percy et al.. 1990) of
leukemia deaths were myeloid leukemias based on hospital diagnoses (versus 39% from death
certificates in the Beane Freeman et al. (2009) study), suggesting that about a third or more of
the "other or unspecified" leukemia deaths in the Beane Freeman et al. (2009) study were
probably myeloid leukemias. Percy et al. (1990) reported in their study that "Of the nearly 600
deaths from leukemia NOS [other or unspecified] nearly 50% were originally diagnosed as
myeloid... Obviously myeloid leukemia is grossly underreported on death certificates."
This document is a draft for review purposes only and does not constitute Agency policy.
154 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Toxicological Review of Formaldehyde - Inhalation
o Because it is likely that a proportion of myeloid leukemia cases were reported as
"other/unspecified," a more complete estimate of the association of cumulative formaldehyde
exposure with myeloid leukemia might be obtained using the regression results for a
combination of myeloid leukemia and other/unspecified leukemia,
o Although a unit risk estimate that combines myeloid leukemia and other/unspecified leukemia
overtly includes cancer subtypes not necessarily causally associated with the chemical exposure,
it is sometimes the case that, due to data limitations, unit risk estimates are based on human
data that reflect cancer groupings broader than what might be strictly causally associated with
the chemical exposure (e.g., all leukemias or all lung cancers). The inclusion of unassociated
cancer subtypes in the derivation of the regression coefficient should theoretically attenuate the
association.
o A comparison of the unit risk estimates for all leukemia, myeloid leukemia plus other
unspecified leukemia, and myeloid leukemia (ICD-8/9: 205) indicates that all of the estimates
are within a factor of 1.5. Unit risk estimates were 3.9 x 10 "2, 4.2 x 10"2, and 5.9 x 10"2 for all
leukemia, myeloid leukemia plus other unspecified leukemia, and myeloid leukemia (ICD-8/9:
205), respectively.
• The approach for combining myeloid leukemia and other/unspecified leukemia to estimate risk,
while arguably consistent with the identified misclassification of myeloid leukemia on death
certificates (Percy et al.. 1990; Percy et al.. 1981) is uncommon, and retains significant quantitative
uncertainties, including some inconsistencies in statistical results. To a limited extent, it might also
be viewed as combining cancer types that differ in terms of the cell of origin and other
characteristics of cancer development (e.g., latency; MOA).
o The combination of myeloid leukemia and other/unspecified leukemia in the regression model
yields a p-value of 0.1. While the number of cases is increased by n = 36, cancers in this
category, with the exception of the myeloid leukemia cases, were not identified to be causally
associated with formaldehyde exposure during the hazard evaluation. The inclusion of cancers
not causally associated with formaldehyde exposure would be expected to attenuate the
association, but in contrast to this expectation, there was a stronger association for the
regression model of other/unspecified leukemia alone (p = 0.13) compared to the model of
myeloid leukemia alone (p = 0.44). There is not a clear explanation for why the association
would be stronger for the more heterogeneous leukemia category,
o There is likely more uncertainty associated with the background cancer rates in the U.S.
population for the other/unspecified leukemias than for the specified myeloid and lymphocytic
leukemia subtypes. The survival rates of the other/unspecified cancers had to be estimated by
subtracting myeloid and lymphocytic leukemia rates from the rates for all leukemia,
o As the Beane-Freeman study did evaluate myeloid leukemia, the use of either myeloid leukemia
plus other/unspecified leukemias or the even broader category of all leukemias would represent
deviations from using the most specific diagnoses possible. Depending on the extent to which
the combined cancer types differ (e.g., in terms of cancer development), this could introduce
significant quantitative uncertainties. However, such a decision to group or focus on individual
cancer types must also consider the number and power of the available studies to be capable of
detecting changes with reasonable confidence.
This document is a draft for review purposes only and does not constitute Agency policy.
155 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 • Given the completely unknown MOA for myeloid leukemia, it is possible and perhaps likely that
2 there are dose and duration effects for the development of myeloid leukemia following
3 formaldehyde inhalation that are not fully understood.
4 o Acknowledging the complexity of the different dose metrics available in the observational
5 studies, as well as the lack of an association between cumulative exposure and myeloid
6 leukemia in the Beane-Freeman study, it is possible that the specific, individual exposure metrics
7 in this study failed to fully capture the patterns of exposure with which the development of
8 myeloid leukemia is causally associated. Importantly, this concern is independent of the
9 identified hazard for myeloid leukemia, as myeloid leukemia mortality was increased in
10 association with the peak exposure metric in this study (industrial workers) and others, as well
11 as with duration-dependent metrics including time since first exposure, cumulative exposure,
12 and exposure duration in two other occupational cohorts (garment workers and embalmers).
13 o As information supporting a nonlinear extrapolation from the identified POD is not available for
14 myeloid leukemia, the current approach uses a default linear extrapolation. It is possible that
15 additional study on the development of this cancer after formaldehyde exposure could provide
16 support for the linear extrapolation or, alternatively, support a nonlinear approach.
17 Considering these uncertainties in the myeloid leukemia unit risk estimate, the selected IUR,
18 summarized in Table 61, reflects the estimate for NPC incidence alone. For benefits analyses and certain
19 other situations, "central" estimates of risk per unit dose may be preferred over (upper bound) unit risk
20 estimates. Therefore, the assessment also provides estimates of risk per unit dose resulting from linear
21 extrapolations of risk from the central estimate (here, the EC, or effective concentration associated with
22 the benchmark response level of risk).
23 Table 55. Inhalation unit riska,b
Cancer type
Preferred Unit Risk
Estimate
(ppm-1)
ADAF-adjusted
Unit Risk Estimate
(ppm"1)
Selected Unit Risk
Estimate
((Hg/m3)"1)
ADAF-adjusted Unit
Risk Estimate
((Hg/m3)"1)
Nasopharyngeal
0.0079°
0.013
6.4 x 10"6 c
1.1 X 10"5
aThe inhalation unit risk estimate is typically expressed as the (upper-bound) increase in cancer risk estimated for
an exposure increase of 1 ng/m3.
bThe unit risk estimate is for cancer incidence.
cAdult-based (rescaled) unit risk estimate for NPC intended for the application of ADAFs.
24 Sources of uncertainty associated with the selected inhalation unit risk
25 The availability of suitable human data from which to derive unit risk estimates eliminates one
26 of the major sources of uncertainty inherent in most unit risk estimates—the uncertainty associated
27 with interspecies extrapolation. The NCI study used as the basis for the selected unit risk estimate was
28 considered a high-quality study for the purposes of deriving unit risk estimates. The NCI study is a large
29 longitudinal cohort study that developed individual-worker exposure estimates using detailed
30 employment histories and formaldehyde concentration measurements. In addition to the detailed
This document is a draft for review purposes only and does not constitute Agency policy.
156 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
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
Toxicological Review of Formaldehyde - Inhalation
exposure assessment, the study used internal analyses and gave careful consideration to potential
confounding or modifying variables. Moreover, the NCI study comprises a large cohort that has been
followed for a long time. Nonetheless, uncertainties in the derived unit risk estimates are inevitable.
The primary uncertainty in the selected inhalation unit risk is the lack of inclusion of an estimate
for the prevalent cancer, myeloid leukemia, due to complexities in the quantitative data, the strengths
and limitations of which are outlined above. Other important sources of uncertainty in the selected unit
risk estimate are the retrospective estimation of individual worker exposures, the dose-response
modeling of the NCI data, the exposure metric, and the high-to-low exposure extrapolation. These
factors, particularly the latter two, could have a large impact on the unit risk estimate. The former two
factors could result in either overprediction or underprediction of the true risk, although regarding the
retrospective estimation of exposures, comparisons with the Marsh et al. (1996) exposure estimates
suggest that the NCI exposure estimates might be overestimates, which would tend to underpredict the
true risk. The latter two factors, the use of cumulative exposure as the exposure metric and the use of
linear high-to-low exposure extrapolation, which are related, would tend to overpredict the true risk.
Additional sources of uncertainty include the use of a single study for the derivation of the unit
risk estimate and the derivation of unit risk estimates for the general population from an occupational
study. The first factor could result in either overprediction or underprediction of the true risk. The
second factor would tend to underpredict the true risks.
While the proven genotoxicity and mutagenicity of formaldehyde and the observation of human
cytogenetic effects in human occupational exposures provide strong support for preferring the linear
extrapolation, an uncertainty in the low dose-response comes from the potential for endogenous
formaldehyde levels in respiratory tissue to reduce the uptake of the inhaled gas at low doses, as
demonstrated in modeling efforts by Schroeter et al. (2014) and Campbell et al. (2020). This would be
expected to result in an overprediction of the true risk.
Sources of uncertainty expected to have minimal quantitative impact include the inability to
derive unit risk estimates for potential cancer sites other than NPC and myeloid leukemia, the derivation
of incidence estimates from mortality data, the influence of confounding and modifying factors (with the
possible exception of particulates, where a modifying effect cannot be ruled out; if particulates are
modifying the NPC risk, the NCI data would tend to overestimate the risk from formaldehyde alone,
possibly to a more considerable extent), and the application of the ADAFs used to address assumed
increased early life susceptibility.
Although substantial uncertainty exists with respect to the low-exposure extrapolation, the
estimate is based on human data from a large, high-quality epidemiological study and mutagenic and
cytogenic modes-of-action are well documented. Furthermore, the estimate is similar to estimates
derived from rodent data. Based on these considerations, overall confidence in the selected inhalation
unit risk is medium.
This document is a draft for review purposes only and does not constitute Agency policy.
157 DRAFT—DO NOT CITE OR QUOTE
-------
Toxicological Review of Formaldehyde - Inhalation
1 4.5.5. Previous IRIS Assessment: Inhalation Unit Risk
2 In 1989, an inhalation unit risk of 1.3 x 10~5 per ng/m3 was developed based on nasal squamous
3 cell carcinomas (SCCs) in F344 rats from Kerns et al. (Kerns et al.. 1983). The data were modeled from
4 the estimates of the probability of death with tumor and its variance using a linearized multistage
5 procedure. It was recommended that this unit risk not be used if the air concentration exceeds 8 x 102
6 Hg/m3, as above that concentration the 1989 unit risk may not be appropriate.
This document is a draft for review purposes only and does not constitute Agency policy.
158 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Toxicological Review of Formaldehyde - Inhalation
5. REFERENCES
Abramson, MJ; Perret, JL; Dharmage, SC; McDonald. VM; McDonald. CF. (2014). Distinguishing adult-
onset asthma from COPD: a review and a new approach [Review], The International Journal of
Chronic Obstructive Pulmonary Disease (Online) 9: 945-962.
http://dx.doi.org/10.2147/COPD.S46761
Aglan. MA; Mansour. GN. (2018). Hair straightening products and the risk of occupational formaldehyde
exposure in hairstylists. Drug Chem Toxicol 43: 1-8.
http://dx.doi.org/10.1080/01480545.2018.15Q8215
Alexandersson. R: Hedenstierna. G. (1989). Pulmonary function in wood workers exposed to
formaldehyde: A prospective study. Arch Environ Health 44: 5-11.
http://dx.doi.org/10.1080/00Q39896.1989.9935865
Andersen. I. (1979). Formaldehyde in the indoor environment - health implications and the setting of
standards. In PO Fanger; O Valbjorn (Eds.), Indoor climate: Effects on human comfort,
performance, and health in residential, commercial, and light-industry buildings (pp. 65-87).
Copenhagen, Denmark: Danish Building Research Institute.
Andersen. I; Molhave, L. (1983). Controlled human studies with formaldehyde. In JE Gibson (Ed.),
Formaldehyde toxicity (pp. 154-165). Washington, DC: Hemisphere Publishing.
Andersen. ME; Clewell, HJ; Bermudez, E; Dodd, DE; Willson, GA; Campbell. JL; Thomas. RS. (2010).
Formaldehyde: integrating dosimetry, cytotoxicity, and genomics to understand dose-
dependent transitions for an endogenous compound. Toxicol Sci 118: 716-731.
http://dx.doi.org/10.1093/toxsci/kfq303
Andjelkovich, DA;Janszen, DB; Brown. MH; Richardson. RB; Miller. FJ. (1995). Mortality of iron foundry
workers: IV. Analysis of a subcohort exposed to formaldehyde. J Occup Environ Med 37: 826-
837. http://dx.doi.org/10.1097/00043764-199507000-00Q12
Annesi-Maesano. I; Hulin. M; Lavaud. F; Raherison. C; Kopferschmitt. C; de Blav. F; Charpin. DA; Denis. C.
(2012). Poor air quality in classrooms related to asthma and rhinitis in primary schoolchildren of
the French 6 Cities Study. Thorax 67: 682-688. http://dx.doi.org/10.1136/thoraxinl-2011-200391
Appelman. LM; Woutersen. RA; Zwart. A; Falke. HE; Feron. VJ. (1988). One-year inhalation toxicity study
of formaldehyde in male rats with a damaged or undamaged nasal mucosa. J Appl Toxicol 8: 85-
90. http://dx.doi.org/10.1002/iat.25500802Q4
ATS (American Thoracic Society). (2000). What constitutes an adverse health effect of air pollution? Am J
Respir Crit Care Med 161: 665-673. http://dx.doi.Org/10.1164/airccm.161.2.ats4-00
Attia, D; Mansour. N; Taha, F; El Dein, AS. (2014). Assessment of lipid peroxidation and p53 as a
biomarker of carcinogenesis among workers exposed to formaldehyde in cosmetic industry.
Toxicol Ind Health 32: 1097-1105. http://dx.doi.org/10.1177/0748233714547152
Bachand, AM; Mundt, KA; Mundt, DJ; Montgomery, RR. (2010). Epidemiological studies of formaldehyde
exposure and risk of leukemia and nasopharyngeal cancer: A meta-analysis. Crit Rev Toxicol 40:
85-100. http://dx.doi.org/10.3109/104084409Q3341696
Baird, DP; Wilcox. AJ; Weinberg. CR. (1986). Use of time to pregnancy to study environmental
exposures. Am J Epidemiol 124: 470-480.
Ballarin, C; Sarto, F; Giacomelli, L; Bartolucci, GB; Clonfero, E. (1992). Micronucleated cells in nasal
mucosa of formaldehyde-exposed workers. Mutat Res Genet Toxicol 280: 1-7.
http://dx.doi.org/10.1016/0165-1218(92)90012-0
This document is a draft for review purposes only and does not constitute Agency policy.
159 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Band. PR; Le. ND; Fang. R; Threlfall. WJ; Astrakianakis. G; Anderson. JT; Keefe. A; Krewski. D. (1997).
Cohort mortality study of pulp and paper mill workers in British Columbia, Canada. Am J
Epidemiol 146: 186-194. http://dx.doi.org/10.1093/oxfordiournals.aie.a009250
Bassig, BA; Zhang. L; Vermeulen, R; Tang. X; Li. G; Hu. W. ei; Guo, W; Purdue. MP; Yin. S; Rappaport, SM;
Shen, M. in; Ji. Z; Qiu, C; Ge. Y; Hosgood, HP; Reiss, B; Wu. B; Xie, Y; Li. L; Yue, F. ei; Freeman.
LEB; Blair. A; Hayes, RB; Huang. H; Smith. MT; Rothman, N; Lan, Q. (2016). Comparison of
hematological alterations and markers of B-cell activation in workers exposed to benzene,
formaldehyde and trichloroethylene. Carcinogenesis 37: 692-700.
http://dx.doi.org/10.1093/carcin/bgw053
Battel le (Battelle Columbus Laboratories). (1982). A chronic inhalation toxicology study in rats and mice
exposed to formaldehyde. Research Triangle Park, NC: Chemical Industry Institute of Toxicology.
Beane Freeman. LE; Blair. A; Lubin, JH; Stewart. PA; Hayes, RB; Hoover. RN; Hauptmann, M. (2013).
Mortality from solid tumors among workers in formaldehyde industries: an update of the NCI
cohort. Am J Ind Med 56: 1015-1026. http://dx.doi.org/10.10Q2/aiim.22214
Beane Freeman. LE; Blair. A; Lubin. JH; Stewart. PA; Hayes, RB; Hoover. RN; M. H. (2009). Mortality from
lymphohematopoietic malignancies among workers in formaldehyde industries: The National
Cancer Institute Cohort. J Natl Cancer Inst 101: 751-761. http://dx.doi.org/10.1093/inci/dip096
Bellavia. A; Dickerson. AS; Rotem. RS; Hansen. J; Gredal. O; Weisskopf. MG. (2021). Joint and interactive
effects between health comorbidities and environmental exposures in predicting amyotrophic
lateral sclerosis. Int J Hyg Environ Health 231: 113655.
http://dx.doi.Org/10.1016/i.iiheh.2020.113655
Bender. JR; Mullin. LS; Grapel. GJ; Wilson. WE. (1983). Eye irritation response of humans to
formaldehyde. Am Ind Hyg Assoc J 44: 463-465. http://dx.doi.org/10.1080/15298668391405139
Bentayeb, M; Norback, D; Bednarek, M; Bernard. A; Cai, G; Cerrai, S; Eleftheriou, KK; Gratziou, C; Hoist,
GJ; Lavaud, F; Nasilowski, J; Sestini, P; Sarno, G; Sigsgaard, T; Wieslander, G; Zielinski, J; Viegi, G;
Annesi-Maesano, I; Study, G. (2015). Indoor air quality, ventilation and respiratory health in
elderly residents living in nursing homes in Europe. Eur Respir J 45: 1228-1238.
http://dx.doi.org/10.1183/09031936.00Q82414
Berglund, B; Hoglund, A; Esfandabad, HS. (2012). A bisensory method for odor and irritation detection of
formaldehyde and pyridine. Chemosensory Perception 5: 146-157.
http://dx.doi.org/10.1007/sl2Q78-011-9101-9
Billionnet, C; Gay, E; Kirchner, S; Leynaert, B; Annesi-Maesano. I. (2011). Quantitative assessments of
indoor air pollution and respiratory health in a population-based sample of French dwellings.
Environ Res 111: 425-434. http://dx.doi.Org/10.1016/i.envres.2011.02.008
Blair. A; Saracci, R; Stewart. PA; Hayes, RB; Shy, C. (1990). Epidemiologic evidence on the relationship
between formaldehyde exposure and cancer [Review], Scand J Work Environ Health 16: 381-
393. http://dx.doi.org/10.5271/siweh.1767
Blair. A; Zheng. T; Linos. A; Stewart. PA; Zhang. YW; Cantor. KP. (2001). Occupation and leukemia: A
population-based case-control study in Iowa and Minnesota. Am J Ind Med 40: 3-14.
http://dx.doi.org/10.1002/aiim.1066
Boffetta. P; Stellman. SD; Garfinkel. L. (1989). A case-control study of multiple myeloma nested in the
American Cancer Society prospective study. Int J Cancer 43: 554-559.
http://dx.doi.org/10.1002/iic.29104304Q4
Bogdanffy, MS; Randall. HW; Morgan. KT. (1986). Histochemical localization of aldehyde dehydrogenase
in the respiratory tract of the Fischer-344 rat. Toxicol Appl Pharmacol 82: 560-567.
http://dx.doi.org/10.1016/0041-008X(86)90291-7
Bono. R; Romanazzi, V; Munnia, A; Piro, S; Allione, A; Ricceri, F; Guarrera, S; Pignata, C; Matullo, G;
Wang. P; Giese, RW; Peluso, M. (2010). Malondialdehyde-deoxyguanosine adduct formation in
This document is a draft for review purposes only and does not constitute Agency policy.
160 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Toxicological Review of Formaldehyde - Inhalation
workers of pathology wards: the role of air formaldehyde exposure. Chem Res Toxicol 23: 1342-
1348. http://dx.doi.org/10.1021/txlQ0083x
Bosetti, C; Mclaughlin. JK; Tarone, RE; Pira, E; La Vecchia, C. (2008). Formaldehyde and cancer risk: a
quantitative review of cohort studies through 2006 [Review], Ann Oncol 19: 29-43.
http://dx.doi.org/10.1093/annonc/mdm202
Boysen, M; Zadig, E; Digernes, V; Abeler, V; Reith, A. (1990). Nasal mucosa in workers exposed to
formaldehyde: a pilot study. Occup Environ Med 47: 116-121.
http://dx.doi.Org/10.1136/oem.47.2.116
Broder, I; Corey, P; Cole. P; Lipa, M; Mintz, S; Nethercott, JR. (1988). Comparison of health of occupants
and characteristics of houses among control homes and homes insulated with urea
formaldehyde foam: II initial health and house variables and exposure-response relationships.
Environ Res 45: 156-178. http://dx.doi.org/10.1016/S0013-9351(88)80045-8
Burgaz, S; Cakmak, G; Erdem, O; Yilmaz, M; Karakaya, AE. (2001). Micronuclei frequencies in exfoliated
nasal mucosa cells from pathology and anatomy laboratory workers exposed to formaldehyde.
Neoplasma 48: 144-147.
Burgaz. S; Erdem. O; Cakmak. G; Erdem. N; Karakaya, A; Karakaya, AE. (2002). Cytogenetic analysis of
buccal cells from shoe-workers and pathology and anatomy laboratory workers exposed to n-
hexane, toluene, methyl ethyl ketone and formaldehyde. Biomarkers 7: 151-161.
http://dx.doi.org/10.1080/1354750011Q113242
Campbell Jr. J: Gentry. PR: Clewell III. HJ: Andersen. ME. (2020). A kinetic analysis of DNA-deoxy guanine
adducts in the nasal epithelium produced by inhaled formaldehyde in rats-assessing
contributions to adduct production from both endogenous and exogenous sources of
formaldehyde. Toxicol Sci 177: 325-333. http://dx.doi.org/10.1093/toxsci/kfaal22
Casanova-Schmitz, M; Starr. TB; Heck. HP. (1984). Differentiation between metabolic incorporation and
covalent binding in the labeling of macromolecules in the rat nasal mucosa and bone marrow by
inhaled [14C]- and [3H]formaldehyde. Toxicol Appl Pharmacol 76: 26-44.
http://dx.doi.org/10.1016/0041-008x(84)90026-7
Casanova. M; Heck. H; Everitt, Jl; Harrington. WW. Jr; Popp, JA. (1988). Formaldehyde concentrations in
the blood of rhesus monkeys after inhalation exposure. Food Chem Toxicol 26: 715-716.
http://dx.doi. org/10.1016/0278-6915(88)90071-3
Casanova. M; Morgan. KT; Gross. EA; Moss. OR; Heck. H. (1994). DNA-protein cross-links and cell
replication at specific sites in the nose of F344 rats exposed subchronically to formaldehyde.
Fundam Appl Toxicol 23: 525-536. http://dx.doi.org/10.1006/faat.1994.1137
Casanova. M; Morgan. KT; Steinhagen, WH; Everitt. Jl; Popp. JA; Heck. H. (1991). Covalent binding of
inhaled formaldehyde to DNA in the respiratory tract of rhesus monkeys: pharmacokinetics, rat-
to-monkey interspecies scaling, and extrapolation to man. Toxicol Sci 17: 409-428.
http://dx.doi. org/10.1016/0272-0590(91)90230-2
Casset. A; Marchand. C; Purohit. A; le Calve. S; Uring-Lambert. B; Donnav. C; Meyer. P; de Blav. F. (2006).
Inhaled formaldehyde exposure: effect on bronchial response to mite allergen in sensitized
asthma patients. Allergy 61: 1344-1350. http://dx.doi.Org/10.llll/i.1398-9995.2006.01174.x
Chang J. CF; Gross. EA; Swenberg. JA; Barrow. CS. (1983). Nasal cavity deposition, histopathology, and
cell proliferation after single or repeated formaldehyde exposures in B6C3F1 mice and F-344
rats. Toxicol Appl Pharmacol 68: 161-176.
Chang. JCF; Gross. EA; Swenberg. JA; Barrow. CS. (1983). Nasal cavity deposition, histopathology, and
cell proliferation after single or repeated formaldehyde exposures in B6C3F1 mice and F-344
rats. Toxicol Appl Pharmacol 68: 161-176. http://dx.doi.org/10.1016/0041-008x(83)90001-7
This document is a draft for review purposes only and does not constitute Agency policy.
161 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Chang. JCF; Steinhagen. WH; Barrow. CS. (1981). Effect of single or repeated formaldehyde exposure on
minute volume of B6C3F1 mice and F-344 rats. Toxicol Appl Pharmacol 61: 451-459.
http://dx.doi.org/10.1016/0041-008x(81)90368-9
Chang. M; Park. H; Ha. M; Hong. YC; Lim, YH; Kim. Y; Kim. YJ; Lee. D; Ha. EH. (2017). The effect of
prenatal TVOC exposure on birth and infantile weight: the Mothers and Children's
Environmental Health study. Pediatr Res 82: 423-428. http://dx.doi.org/10.1038/pr.2017.55
Checkoway, H; Boffetta, P; Mundt, DJ; Mundt, KA. (2012). Critical review and synthesis of the
epidemiologic evidence on formaldehyde exposure and risk of leukemia and other
lymphohematopoietic malignancies [Review], Cancer Causes Control 23: 1747-1766.
http://dx.doi.org/10.1007/slQ552-012-0055-2
Checkoway, H; Dell. LP; Boffetta. P; Gallagher. AE; Crawford. L; Lees. PS; Mundt. KA. (2015).
Formaldehyde Exposure and Mortality Risks From Acute Myeloid Leukemia and Other
Lymphohematopoietic Malignancies in the US National Cancer Institute Cohort Study of
Workers in Formaldehyde Industries. J Occup Environ Med 57: 785-794.
http://dx.doi.org/10.1097/JOM.000000000000Q466
CUT (Chemical Industry Institute of Toxicology). (1999). Formaldehyde: hazard characterization and
dose-response assessment for carcinogenicity by the route of inhalation.
Clayton. TC: Meade. TW: Turner. EL: De Stavola. BL. (2014). Peak flow rate and death due to coronary
heart disease: 30-year results from the Northwick Park Heart cohort study. 1: e000164.
http://dx.doi.org/10.1136/openhrt-2014-00Q164
Coggon. D: Harris. EC: Poole. J: Palmer. KT. (2003). Extended follow-up of a cohort of British chemical
workers exposed to formaldehyde. J Natl Cancer Inst 95: 1608-1615.
http://dx.doi.org/10.1093/inci/dig046
Coggon. D; Ntani, G; Harris. EC; Palmer. KT. (2014). Upper Airway Cancer, Myeloid Leukemia, and Other
Cancers in a Cohort of British Chemical Workers Exposed to Formaldehyde. Am J Epidemiol 179:
1301-1311. http://dx.doi.org/10.1093/aie/kwu049
Collins. JJ; Acquavella, JF; Esmen, NA. (1997). An updated meta-analysis of formaldehyde exposure and
upper respiratory tract cancers. J Occup Environ Med 39: 639-651.
http://dx.doi.org/10.1097/00043764-199707000-000Q9
Collins. JJ; Esmen. NA; Hall. TA. (2001). A review and meta-analysis of formaldehyde exposure and
pancreatic cancer [Review], Am J Ind Med 39: 336-345. http://dx.doi.org/10.10Q2/1097-
0274(200103)39:3<336::AID-AJIM1022>3.0.CQ;2-K
Collins. JJ; Lineker, GA. (2004). A review and meta-analysis of formaldehyde exposure and leukemia
[Review], Regul Toxicol Pharmacol 40: 81-91. http://dx.doi.Org/10.1016/i.vrtph.2004.04.006
Conolly, RB; Kimbell, JS; Janszen, D; Schlosser, PM; Kalisak, D; Preston. J; Miller. FJ. (2003). Biologically
motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Toxicol Sci
75: 432-447. http://dx.doi.org/10.1093/toxsci/kfgl82
Conolly. RB; Kimbell. JS; Janszen. D; Schlosser. PM; Kalisak. D; Preston. J; Miller. FJ. (2004). Human
respiratory tract cancer risks of inhaled formaldehyde: dose-response predictions derived from
biologically-motivated computational modeling of a combined rodent and human dataset.
Toxicol Sci 82: 279-296. http://dx.doi.org/10.1093/toxsci/kfh223
Conolly, RB; Kimbell. JS; Janszen. DB; Miller. FJ. (2002). Dose response for formaldehyde-induced
cytotoxicity in the human respiratory tract. Regul Toxicol Pharmacol 35: 32-43.
http://dx.doi.org/10.1006/rtph.2Q01.1515
Conolly, RB; Lilly, PD; Kimbell. JS. (2000). Simulation modelling of the tissue disposition of formaldehyde
to predict nasal DNA-protein cross-links in Fischer 344 rats, rhesus monkeys, and humans.
Environ Health Perspect 108: 919-924. http://dx.doi.org/10.2307/3454325
This document is a draft for review purposes only and does not constitute Agency policy.
162 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Costa. S; Carvalho. S; Costa. C; Coelho. P; Silva. S; Santos. LS; Gaspar. JF; Porto. B; Laffon. B; Teixeira. JP.
(2015). Increased levels of chromosomal aberrations and DNA damage in a group of workers
exposed to formaldehyde. Mutagenesis 30: 463-473. http://dx.doi.org/10.1093/mutage/gev002
Costa. S; Coelho. P; Costa. C; Silva. S; Mayan, O; Santos. LS; Gaspar. J; Teixeira. JP. (2008). Genotoxic
damage in pathology anatomy laboratory workers exposed to formaldehyde. Toxicology 252:
40-48. http://dx.doi.Org/10.1016/i.tox.2008.07.056
Costa. S; Costa. C; Madureira, J; Valdiglesias, V; Teixeira-Gomes, A; Guedes de Pinho, P; Laffon. B;
Teixeira. JP. (2019). Occupational exposure to formaldehyde and early biomarkers of cancer risk,
immunotoxicity and susceptibility. Environ Res 179: 108740.
http://dx.doi.Org/10.1016/i.envres.2019.108740
Crump. KS; Bussard, DA; Chen. C; Jinot, J; Subramaniam, R. (2014). The bottom-up approach does not
necessarily bound low-dose risk [Letter], Regul Toxicol Pharmacol 70: 735-736.
http://dx.doi.Org/10.1016/i.vrtph.2014.10.008
Crump. KS; Chen. C; Fox. JF; Subramaniam. R; van Landingham, C. (2009). Reply to: Letter to the editor.
Formaldehyde risk assessment [Letter], Ann Occup Hyg 53: 181-189.
Crump. KS; Chen. C; Fox. JF; Van Landingham. C; Sumbramaniam, R. (2008). Sensitivity analysis of
biologically motivated model for formaldehyde-induced respiratory cancer in humans. Ann
Occup Hyg 52: 481-495. http://dx.doi.org/10.1093/annhyg/men038
Dalbev. WE. (1982). Formaldehyde and tumors in hamster respiratory tract. Toxicology 24: 9-14.
http://dx.doi. org/10.1016/0300-483X(82)90058-0
Dannemiller. KC; Murphy. JS; Dixon. SL; Pennell. KG; Suuberg. EM; Jacobs. DE; Sandel. M. (2013).
Formaldehyde concentrations in household air of asthma patients determined using
colorimetric detector tubes. Indoor Air 23: 285-294. http://dx.doi.org/10.llll/ina.12024
Dell. L; Teta, MJ. (1995). Mortality among workers at a plastics manufacturing and research and
development facility: 1946-1988. Am J Ind Med 28: 373-384.
http://dx.doi.org/10.1002/aiim.47002803Q7
Edling, C; Hellquist, H; Odkvist, L. (1988). Occupational exposure to formaldehyde and histopathological
changes in the nasal mucosa. Br J Ind Med 45: 761-765.
http://dx.doi.org/10.1136/oem.45.ll.761
Edling. C; Jarvholm, B; Andersson, L; Axelson, O. (1987). Mortality and cancer incidence among workers
in an abrasive manufacturing industry. Br J Ind Med 44: 57-59.
http://dx.doi.Org/10.1136/oem.44.l.57
Edrissi, B; Taghizadeh, K; Dedon, PC. (2013a). Quantitative analysis of histone modifications:
formaldehyde is a source of pathological n(6)-formyllysine that is refractory to histone
deacetylases. PLoS Genet 9: el003328. http://dx.doi.org/10.1371/iournal.pgen.1003328
Edrissi. B; Taghizadeh. K; Moeller, BC; Kracko, D; Doyle-Eisele, M; Swenberg, JA; Dedon. PC. (2013b).
Dosimetry of N6-formyllysine adducts following [13C2H2]-formaldehyde exposures in rats. Chem
Res Toxicol 26: 1421-1423. http://dx.doi.org/10.1021/tx40Q320u
Egle. JL. Jr. (1972). Retention of inhaled formaldehyde, propionaldehyde, and acrolein in the dog. Arch
Environ Health 25: 119-124. http://dx.doi.org/10.1080/00Q39896.1972.10666147
Ericson. A; Eriksson. M; Kallen. B; Westerholm. P; Zetterstrom. R. (1984). Delivery outcome of women
working in laboratories during pregnancy. Arch Environ Health 39: 5-10.
Ezratty, V; Bonay, M; Neukirch, C; Orset-Guillossou, G; Dehoux, M; Koscienlny, S; Cabanes, PA;
Lambrozo, J; Aubier, M. (2007). Effect of formaldehyde on asthmatic response to inhaled
allergen challenge. Environ Health Perspect 115: 210-214. http://dx.doi.org/10.1289/ehp.9414
Fang. F; Quinlan, P; Ye. W; Barber. MK; Umbach. DM; Sandler. DP; Kamel, F. (2009). Workplace
exposures and the risk of amyotrophic lateral sclerosis. Environ Health Perspect 117: 1387-1392.
http://dx.doi.org/10.1289/ehp.090058Q
This document is a draft for review purposes only and does not constitute Agency policy.
163 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Feron. VJ; Bruvnties. JP; Woutersen. RA; Immel. HR; Appelman. LM. (1988). Nasal tumours in rats after
short-term exposure to a cytotoxic concentration of formaldehyde. Cancer Lett 39: 101-111.
http://dx.doi. org/10.1016/0304-3835(88)90045-6
Flamant-Hulin, M; Caillaud, D; Sacco, P; Penard-Morand, C; Annesi-Maesano, I. (2010). Air pollution and
increased levels of fractional exhaled nitric oxide in children with no history of airway damage. J
Toxicol Environ Health A 73: 272-283. http://dx.doi.org/10.1080/152873909032492Q6
Fox. CH; Johnson. FB; Whiting. J; Roller. PP. (1985). Formaldehyde fixation [Review], J Histochem
Cytochem 33: 845-853.
Franklin. P; Dingle. P; Stick. S. (2000). Raised exhaled nitric oxide in healthy children is associated with
domestic formaldehyde levels. Am J Respir Crit Care Med 161: 1575-1759.
http://dx.doi.Org/10.1164/airccm.161.5.9905061
Franklin. P; Tan. M; Herny, N; Hall. GL. (2019). Maternal Exposure to Indoor Air Pollution and Birth
Outcomes. Int J Environ Res Public Health 16. http://dx.doi.org/10.3390/iierphl6081364
Fransman, W; Mclean. D; Douwes, J; Demers, PA; Leung. V; Pearce, N. (2003). Respiratory symptoms
and occupational exposures in New Zealand plywood mill workers. Ann Occup Hyg 47: 287-295.
http://dx.doi.org/10.1093/annhyg/meg046
Fujimaki, H; Kurokawa, Y; Kunugita, N; Kikuchi, M; Sato. F; Arashidani, K. (2004). Differential
immunogenic and neurogenic inflammatory responses in an allergic mouse model exposed to
low levels of formaldehyde [Erratum], Toxicology 197: 1-13.
http://dx.doi.Org/10.1016/i.tox.2005.01.001
Garrett. MH: Hooper. MA: Hooper. BM: Ravment. PR: Abramson. MJ. (1999). Increased risk of allergy in
children due to formaldehyde exposure in homes. Allergy 54: 330-337.
http://dx.doi.Org/10.1034/i.1398-9995.1999.00763.x
Gentry, PR; Rodricks, JV; Turnbull, D; Bachand, A; Van Landingham, C; Shipp, AM; Albertini, RJ; Irons. R.
(2013). Formaldehyde exposure and leukemia: Critical review and reevaluation of the results
from a study that is the focus for evidence of biological plausibility [Review], Crit Rev Toxicol 43:
661-670. http://dx.doi.org/10.3109/10408444.2013.818618
Georgieva, AV; Conolly. RB; Schlosser, PM; Kimbell, JS. (1999). Correlation of inhaled formaldehyde flux
predictions with regional DNA-protein crosslinks (DPX) measurements in rat nasal passages.
Toxicol Sci 48(1-S): 172.
Gerin, M; Siemiatycki, J; Nadon, L; Dewar, R; Krewski, D. (1989). Cancer risks due to occupational
exposure to formaldehyde: Results of a multi-site case-control study in Montreal. Int J Cancer
44: 53-58. http://dx.doi.org/10.1002/iic.291044011Q
Golalipour, MJ; Azarhoush, R; Ghafari, S; Gharravi, AM; Fazeli, SA; Davarian, A. (2007). Formaldehyde
exposure induces histopathological and morphometric changes in the rat testis. Folia Morphol
(Warsz) 66: 167-171.
Green. DJ; Sauder. LR; Kulle. TJ; Bascom. R. (1987). Acute response to 3.0 ppm formaldehyde in
exercising healthy nonsmokers and asthmatics. Am Rev Respir Dis 135: 1261-1266.
http://dx.doi.Org/10.1164/arrd.1987.135.6.1261
Gustavsson. P; Jakobsson. R; Johansson. H; Lewin. F; Norell. S; Rutkvist. LE. (1998). Occupational
exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: A
case-control study in Sweden. Occup Environ Med 55: 393-400.
Hall. A; Harrington. JM; Aw. TC. (1991). Mortality study of British pathologists. Am J Ind Med 20: 83-89.
http://dx.doi.org/10.1002/aiim.47002001Q8
Hankinson, JL; Odencrantz, JR; Fedan, KB. (1999). Spirometric reference values from a sample of the
general US population. Am J Respir Crit Care Med 159: 179-187.
http://dx.doi.Org/10.1164/airccm.159.l.9712108
This document is a draft for review purposes only and does not constitute Agency policy.
164 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Hanrahan. LP; Dally. KA; Anderson. HA; Kanarek. MS; Rankin. J. (1984). Formaldehyde vapor in mobile
homes: A cross sectional survey of concentrations and irritant effects. Am J Public Health 74:
1026-1027. http://dx.doi.Org/10.2105/aiph.74.9.1026
Hansen. J; Olsen, JH. (1995). Formaldehyde and cancer morbidity among male employees in Denmark.
Cancer Causes Control 6: 354-360. http://dx.doi.org/10.1007/BF00Q51411
Harkema, JR; Carey, SA; Wagner. JG. (2006). The nose revisited: A brief overview of the comparative
structure, function, and toxicologic pathology of the nasal epithelium [Review], Toxicol Pathol
34: 252-269. http://dx.doi.org/10.1080/0192623060Q713475
Harkema. JR; Nikula, KJ; Haschek, WM. (2013). Respiratory system. In W Haschek; C Rousseaux; M Wallig
(Eds.), Haschek and Rousseaux's handbook of toxicologic pathology (3rd ed., pp. 1935-2003).
Waltham, MA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-415759-0.00Q51-0
Haseman, JK. (1995). Data analysis: statistical analysis and use of historical control data. Regul Toxicol
Pharmacol 21: 52-59. http://dx.doi.org/10.1006/rtph.1995.1009
Hauptmann, M; Lubin, JH; Stewart. PA; Hayes, RB; Blair. A. (2004). Mortality from solid cancers among
workers in formaldehyde industries. Am J Epidemiol 159: 1117-1130.
http://dx.doi.org/10.1093/aie/kwhl74
Hauptmann. M; Stewart. PA; Lubin. JH; Beane Freeman. LE; Hornung, RW; Herrick, RF; Hoover. RN;
Fraumeni. JF. Jr; Blair. A; Haves. RB. (2009). Mortality from lymphohematopoietic malignancies
and brain cancer among embalmers exposed to formaldehyde. J Natl Cancer Inst 101: 1696-
1708. http://dx.doi.org/10.1093/inci/dip416
Haves. RB; Blair. A; Stewart. PA; Herrick. RF; Mahar. H. (1990). Mortality of U.S. embalmers and funeral
directors. Am J Ind Med 18: 641-652. http://dx.doi.org/10.1002/aiim.47001806Q3
Heck. H; Casanova-Schmitz, M; Dodd, PB; Schachter, EN; Witek, TJ; Tosun, T. (1985). Formaldehyde
(CH20) concentrations in the blood of humans and Fischer-344 rats exposed to CH20 under
controlled conditions. AIHA J 46: 1-3. http://dx.doi.org/10.1080/15298668591394275
Heck. H; Chin. TY; Schmitz, MC. (1983). Distribution of [14C] formaldehyde in rats after inhalation
exposure. In JE Gibson (Ed.), Formaldehyde toxicity (pp. 26-37). Washington, DC: Hemisphere
Publishing.
Heineman, EF; Olsen. JH; Pottern, LM; Gomez. M; Raffn, E; Blair. A. (1992). Occupational risk factors for
multiple myeloma among Danish men. Cancer Causes Control 3: 555-568.
http://dx.doi.org/10.1007/BF00Q52753
Hemminki, K; Kyyronen, P; Lindbohm, ML. (1985). Spontaneous abortions and malformations in the
offspring of nurses exposed to anaesthetic gases, cytostatic drugs, and other potential hazards
in hospitals, based on registered information of outcome. J Epidemiol Community Health 39:
141-147. http://dx.doi.Org/10.1136/iech.39.2.141
Hemminki. K; Mutanen, P; Saloniemi, I; Niemi, ML; Vainio, H. (1982). Spontaneous abortions in hospital
staff engaged in sterilizing instruments with chemical agents. J Occup Environ Med 285: 1461-
1463.
Herbert. FA; Hessel. PA; Melenka. LS;Yoshida. K; Nakaza. M. (1994). Respiratory consequences of
exposure to wood dust and formaldehyde of workers manufacturing oriented strand board.
Arch Environ Health 49: 465-470. http://dx.doi.org/10.1080/00039896.1994.99550Q2
Hill. AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-300.
Holmstrom. M; Wilhelmsson, B. (1988). Respiratory symptoms and pathophysiological effects of
occupational exposure to formaldehyde and wood dust. Scand J Work Environ Health 14: 306-
311.
Holmstrom. M; Wilhelmsson. B; Hellquist, H. (1989a). Histological changes in the nasal mucosa in rats
after long-term exposure to formaldehyde and wood dust. Acta Otolaryngol 108: 274-283.
http://dx.doi.org/10.3109/000164889Q9125528
This document is a draft for review purposes only and does not constitute Agency policy.
165 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Holmstrom. M; Wilhelmsson. B; Hellquist. H; Rosen. G. (1989b). Histological changes in the nasal mucosa
in persons occupationally exposed to formaldehyde alone and in combination with wood dust.
Acta Otolaryngol 107: 120-129. http://dx.doi.org/10.3109/000164889Q9127488
Holness, PL; Nethercott, JR. (1989). Health status of funeral service workers exposed to formaldehyde.
Arch Environ Occup Health 44: 222-228. http://dx.doi.org/10.1080/00Q39896.1989.9935887
Horvath, EP. Jr; Anderson. H. Jr; Pierce. WE; Hanrahan, L; Wendlick, JD. (1988). Effects of formaldehyde
on the mucous membranes and lungs: A study of an industrial population. JAMA 259: 701-707.
http://dx.doi.org/10.1001/iama.1988.0372005003702Q
IARC (International Agency for Research on Cancer). (2006). Formaldehyde, 2-butoxyethanol and 1-tert-
butoxypropan-2-ol [IARC Monograph], Lyon, France. https://publications.iarc.fr/Book-And-
Report-Series/larc-Monographs-On-The-ldentification-Of-Carcinogenic-Hazards-To-
Humans/Formaldehvde-2-Butoxvethanol-And-l-Em-Tert-Em-Butoxypropan-2-ol-2006
Ito, K; Sakamoto. T; Hayashi, Y; Morishita, M; Shibata, E; Sakai, K; Takeuchi, Y; Torii, S. (1996). Role of
tachykinin and bradykinin receptors and mast cells in gaseous formaldehyde-induced airway
microvascular leakage in rats. Eur J Pharmacol 307: 291-298. http://dx.doi.org/10.1016/0Q14-
2999(96)00285-3
Jiang. S; Yu. L; Cheng. J; Leng, S; Dai. Y; Zhang. Y; Niu, Y; Yan, H; Qu. W; Zhang. C; Zhang. K; Yang. R; Zhou.
L: Zheng. Y. (2010). Genomic damages in peripheral blood lymphocytes and association with
polymorphisms of three glutathione S-transferases in workers exposed to formaldehyde. Mutat
Res 695: 9-15. http://dx.doi.Org/10.1016/i.mrgentox.2009.09.011
John. EM: Savitz. DA: Shy. CM. (1994). Spontaneous abortions among cosmetologists. Epidemiology 5:
147-155. http://dx.doi.org/10.1097/00001648-199403000-000Q4
Kamata, E; Nakadate, M; Uchida, O; Ogawa, Y; Suzuki. S; Kaneko, T; Saito, M; Kurokawa, Y. (1997).
Results of a 28-month chronic inhalation toxicity study of formaldehyde in male Fisher-344 rats.
J Toxicol Sci 22: 239-254.
Kerns. WD; Pavkov, KL; Donofrio, DJ; Gralla, EJ; Swenberg, JA. (1983). Carcinogenicity of formaldehyde in
rats and mice after long-term inhalation exposure. Cancer Res 43: 4382-4392.
Kim. JL; Elfman, L; Wieslander, G; Ferm, M; Toren, K; Norback, D. (2011). Respiratory health among
Korean pupils in relation to home, school and outdoor environment. J Korean Med Sci 26: 166-
173. http://dx.doi.Org/10.3346/ikms.2011.26.2.166
Kim. Y; Jekarl, DW; Kim. J; Kwon, A; Choi. H; Lee. S; Kim. YJ; Kim. HJ; Kim. Y; Oh. IH; Kim. M. (2015).
Genetic and epigenetic alterations of bone marrow stromal cells in myelodysplastic syndrome
and acute myeloid leukemia patients. Stem Cell Research 14: 177-184.
http://dx.doi.Org/10.1016/i.scr.2015.01.004
Kimbell, JS; Overton. JH; Subramaniam, RP; Schlosser, PM; Morgan. KT; Conolly, RB; Miller. FJ. (2001a).
Dosimetry modeling of inhaled formaldehyde: Binning nasal flux predictions for quantitative risk
assessment. Toxicol Sci 64: 111-121.
Kimbell. JS; Subramaniam. RP; Gross. EA; Schlosser. PM; Morgan. KT. (2001b). Dosimetry modeling of
inhaled formaldehyde: comparisons of local flux predictions in the rat, monkey, and human
nasal passages. Toxicol Sci 64: 100-110.
Kirsch-Volders. M; Bonassi. S; Knasmueller. S; Holland. N; Bolognesi. C; Fenech. MF. (2014).
Commentary: Critical questions, misconceptions and a road map for improving the use of the
lymphocyte cytokinesis-block micronucleus assay for in vivo biomonitoring of human exposure
to genotoxic chemicals-A HUMN project perspective. Mutat Res Rev Mutat Res 759: 49-58.
http://dx.doi.Org/10.1016/i.mrrev.2013.12.001
Kitaev, EM; Savchenko, ON; Lovchikov, VA; Altukhov, VV; Vishnyakov, YS. (1984). Razvitie zarodyshey i
nekotorye pokazateli reproductivnoy funktsii u krys posle ingalyatsionnogo vozdeystviya
formal'degida do oplodotvoreniya [Akush Ginekol 10: 49-52.
This document is a draft for review purposes only and does not constitute Agency policy.
166 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Kleinniienhuis. AJ; Staal. YC; Duistermaat. E; Engel. R; Woutersen. RA. (2013). The determination of
exogenous formaldehyde in blood of rats during and after inhalation exposure. Food Chem
Toxicol 52: 105-112. http://dx.doi.Org/10.1016/i.fct.2012.ll.008
Krewski, D; Crump. KS; Farmer. J; Gaylor, DW; Howe. R; Portier, C; Salsburg, D; Sielken, RL; Van Ryzin, J.
(1983). A comparison of statistical methods for low dose extrapolation utilizing time-to-tumour
data. Fundam Appl Toxicol 3: 140-160. http://dx.doi.org/10.1016/S0272-0590(83)80075-X
Kriebel, D; Myers, D; Cheng. M; Woskie, S; Cocanour, B. (2001). Short-term effects of formaldehyde on
peak expiratory flow and irritant symptoms. Arch Environ Health 56: 11-18.
http://dx.doi.org/10.1080/00039890109604Q49
Krzyzanowski, M; Quackenboss, JJ; Lebowitz, MP. (1990). Chronic respiratory effects of indoor
formaldehyde exposure. Environ Res 52: 117-125. http://dx.doi.org/10.1016/S0Q13-
9351(05)80247-6
Kulle, TJ; Cooper. GP. (1975). Effects of formaldehyde and ozone on the trigeminal nasal sensory system.
Arch Environ Occup Health 30: 237-243.
Kulle. TJ; Sauder, LR; Hebel, JR; Green. DJ; Chatham. MP. (1987). Formaldehyde dose-response in
healthy nonsmokers. J Air Pollut Control Assoc 37: 919-924.
http://dx.doi.org/10.1080/0894063Q.1987.10466285
Ladeira. C: Viegas. S: Carolino. E: Gomes. MC: Brito. M. (2013). The influence of genetic polymorphisms
in XRCC3 and ADH5 genes on the frequency of genotoxicity biomarkers in workers exposed to
formaldehyde. Environ Mol Mutagen 54: 213-221. http://dx.doi.org/10.1002/em.21755
Ladeira. C: Viegas. S: Carolino. E: Prista. J: Gomes. MC: Brito. M. (2011). Genotoxicity biomarkers in
occupational exposure to formaldehyde-the case of histopathology laboratories. Mutat Res
721: 15-20. http://dx.doi.Org/10.1016/i.mrgentox.2010.ll.015
Laforest, L; Luce. D; Goldberg. P; Begin, D; Gerin, M; Demers, PA; Brugere, J; Leclerc, A. (2000). Laryngeal
and hypopharyngeal cancers and occupational exposure to formaldehyde and various dusts: A
case-control study in France. Occup Environ Med 57: 767-773.
http://dx.doi.org/10.1136/oem.57.ll.767
Lai. Y; Yu. R; Hartwell, HJ; Moeller, BC; Bodnar, WM; Swenberg, JA. (2016). Measurement of Endogenous
versus Exogenous Formaldehyde-Induced DNA-Protein Crosslinks in Animal Tissues by Stable
Isotope Labeling and Ultrasensitive Mass Spectrometry. Cancer Res 76: 2652-2661.
http://dx.doi.org/10.1158/0008-5472.CAN-15-2527
Lajoie, P; Aubin, D; Gingras, V; Daigneault, P; Ducharme, F; Gauvin, FD; Fugler, D; Leclerc. JM; Won. D;
Won. D; Courteau, M; Gingras, S; Heroux, ME; Yang. W; Schleibinger, H. (2014). The IVAIRE
Project - A Randomized Controlled Study of the Impact of Ventilation on Indoor Air Quality and
the Respiratory Symptoms of Asthmatic Children in Single Family Homes. Indoor Air 25: 582-597.
http://dx.doi.org/10.llll/ina.12181
Lan. Q; Smith. MT; Tang. X; Guo. W; Vermeulen. R; Ji. Z; Hu. W; Hubbard. AE; Min. S; Mchale. CM; Qiu. C;
Liu. S; Reiss. B; Beane Freeman. L; Blair. A; Ge. Y; Xiong. J; Li. L; Rappaport. SM; Huang. H;
Rothman. N; Zhang. L. (2015). Chromosome-wide aneuploidy study (CWAS) of cultured
circulating myeloid progenitor cells from workers occupationally exposed to formaldehyde.
Carcinogenesis 36: 160-167. http://dx.doi.org/10.1093/carcin/bgu229
Lang. I; Bruckner. T; Triebig, G. (2008). Formaldehyde and chemosensory irritation in humans: A
controlled human exposure study. Regul Toxicol Pharmacol 50: 23-36.
http://dx.doi.Org/10.1016/i.vrtph.2007.08.012
Larsen, ST; Wolkoff, P; Hammer. M; Kofoed-Sgrensen, V; Clausen. PA; Nielsen. GD. (2013). Acute airway
effects of airborne formaldehyde in sensitized and non-sensitized mice housed in a dry or humid
environment. Toxicol Appl Pharmacol 268: 294-299.
http://dx.doi.Org/10.1016/i.taap.2013.02.006
This document is a draft for review purposes only and does not constitute Agency policy.
167 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Lebowitz. MP; Krzvzanowski. M; Quackenboss. JJ; Orourke. MK. (1997). Diurnal variation of PEF and its
use in epidemiological studies. Eur Respir J 10: S49-S56.
Lee. JT; Ko. CY. (2005). Has survival improved for nasopharyngeal carcinoma in the United States?
Otolaryngol Head Neck Surg 132: 303-308. http://dx.doi.Org/10.1016/i.otohns.2004.09.018
Lee. PN; Fry, JS. (2010). Systematic review of the evidence relating FEV1 decline to giving up smoking
[Review], BMC Med 8: 84. http://dx.doi.org/10.1186/1741-7015-8-84
Leikauf, GD. (1992). Mechanisms of aldehyde-induced bronchial reactivity: role of airway epithelium.
Res Rep Health Eff lnstl-35.
Lin. D; Guo, Y; Yi. J; Kuang, D. an; Li. X; Deng. H; Huang. K. un; Guan, L. ei; He. Y; Zhang. X; Hu. D. ie;
Zhang. Z; Zheng. H; Zhang. X; Mchale, CM; Zhang. L; Wu. T. (2013). Occupational exposure to
formaldehyde and genetic damage in the peripheral blood lymphocytes of plywood workers. J
Occup Health 55: 284-291. http://dx.doi.org/10.1539/ioh.12-0288-QA
Lindbohm, ML; Hemminki, K; Bonhomme, MG; Anttila, A; Rantala, K; Heikkila, P; Rosenberg. MJ. (1991).
Effects of paternal occupational exposure on spontaneous abortions. Am J Public Health 81:
1029-1033. http://dx.doi.Org/10.2105/aiph.81.8.1029
Liu. KS; Huang. FY; Hayward, SB; Wesolowski, J; Sexton. K. (1991). Irritant effects of formaldehyde
exposure in mobile homes. Environ Health Perspect 94: 91-94.
http://dx.doi.org/10.2307/3431298
Liu. QB; Wang. W; Jing. W. (2018). Indoor air pollution aggravates asthma in Chinese children and
induces the changes in serum level of miR-155. Int J Environ Health Res 29: 1-9.
http://dx.doi.org/10.1080/09603123.2018.15Q6569
Lofstedt. H; Westberg. H; Selden. Al; Brvngelsson. IL; Svartengren. M. (2011). Respiratory symptoms and
lung function in foundry workers using the hot box method: A 4-year follow-up. J Occup Environ
Med 53: 1425-1429. http://dx.doi.org/10.1097/JQM.0b013e3182363cl7
Lu. K; Collins. LB; Ru. H; Bermudez, E; Swenberg, JA. (2010). Distribution of DNA adducts caused by
inhaled formaldehyde is consistent with induction of nasal carcinoma but not leukemia. Toxicol
Sci 116: 441-451. http://dx.doi.org/10.1093/toxsci/kfq061
Lu. K; Craft. S; Nakamura, J; Moeller, BC; Swenberg. JA. (2012). Use of LC-MS/MS and stable isotopes to
differentiate hydroxymethyl and methyl DNA adducts from formaldehyde and
nitrosodimethylamine. Chem Res Toxicol 25: 664-675. http://dx.doi.org/10.1021/tx200426b
Lu. K; Moeller. B; Doyle-Eisele, M; Mcdonald. J; Swenberg. JA. (2011). Molecular dosimetry of N2-
hydroxymethyl-dG DNA adducts in rats exposed to formaldehyde. Chem Res Toxicol 24: 159-
161. http://dx.doi.org/10.1021/txl003886
Luce. D; Leclerc, A; Begin, D; Demers, PA; Gerin, M; Orlowski, E; Kogevinas, M; Belli. S; Bugel, I; Bolm-
Audorff, U; Brinton, LA; Comba, P; Hardell, L; Hayes, RB; Magnani, C; Merler, E; Preston-Martin,
S; Vaughan, TL; Zheng. W; Boffetta, P. (2002). Sinonasal cancer and occupational exposures: a
pooled analysis of 12 case-control studies. Cancer Causes Control 13: 147-157.
http://dx.doi.Org/10.1023/A:1014350004255
Malaka. T; Kodama. AM. (1990). Respiratory health of plywood workers occupationally exposed to
formaldehyde. Arch Environ Health 45: 288-294.
http://dx.doi.org/10.1080/00039896.199Q.10118748
Maronpot, RR; Miller. RA; Clarke. WJ; Westerberg, RB; Decker. JR; Moss. OR. (1986). Toxicity of
formaldehyde vapor in B6C3F1 mice exposed for 13 weeks. Toxicology 41: 253-266.
http://dx.doi.org/10.1016/0300-483X(86)90180-0
Marsh. GM; Stone. RA; Esmen, NA; Henderson. VL; Lee. KY. (1996). Mortality among chemical workers in
a factory where formaldehyde was used. Occup Environ Med 53: 613-627.
http://dx.doi.Org/10.1136/oem.53.9.613
This document is a draft for review purposes only and does not constitute Agency policy.
168 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Marsh. GM; Youk. AO; Buchanich. JM; Cassidv. LP; Lucas. U; Esmen. NA; Gathuru. IM. (2002). Pharyngeal
cancer mortality among chemical plant workers exposed to formaldehyde. Toxicol Ind Health
18: 257-268. http://dx.doi.org/10.1191/0748233702thl49oa
Marsh. GM; Youk. AO; Buchanich. JM; Erdal, S; Esmen. NA. (2007). Work in the metal industry and
nasopharyngeal cancer mortality among formaldehyde-exposed workers. Regul Toxicol
Pharmacol 48: 308-319. http://dx.doi.Org/10.1016/i.vrtph.2007.04.006
Matanoski, GM. (1989). Risks of pathologists exposed to formaldehyde (final report). (DHHS Grant No. 5
R01-OH-01511-03). Baltimore, MD: Johns Hopkins University Department of Epidemiology.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml?searchQuery=PB91173682
Matsunaga, I; Miyake, Y; Yoshida, T; Miyamoto, S; Ohya, Y; Sasaki. S; Tanaka, K; Oda, H; Ishiko, O; Hirota,
Y; Group. OMaCHS. (2008). Ambient formaldehyde levels and allergic disorders among Japanese
pregnant women: Baseline data from the Osaka maternal and child health study. Ann Epidemiol
18: 78-84. http://dx.doi.Org/10.1016/i.annepidem.2007.07.095
Menezes, AM; Perez-Padilla, R; Wehrmeister, FC; Lopez-Varela, MV; Muino, A; Valdivia, G; Lisboa, C;
Jardim, J. R.; de Oca. MM; Talamo, C; Bielemann, R; Gazzotti, M; Laurenti, R; Celli, B; Victora, CG;
team. P. (2014). FEV1 is a better predictor of mortality than FVC: the PLATINO cohort study.
PLoS ONE 9: el09732. http://dx.doi.org/10.1371/iournal.pone.0109732
Meyers. AR; Pinkerton. LE; Hein. MJ. (2013). Cohort mortality study of garment industry workers
exposed to formaldehyde: Update and internal comparisons. Am J Ind Med 56: 1027-1039.
http://dx.doi.org/10.1002/aiim.22199
Mi. YH; Norback. D; Tao. J; Mi. YL; Ferm. M. (2006). Current asthma and respiratory symptoms among
pupils in Shanghai, China: Influence of building ventilation, nitrogen dioxide, ozone, and
formaldehyde in classrooms. Indoor Air 16: 454-464. http://dx.doi.Org/10.llll/j. 1600-
0668.2006.00439.x
Miller. MR; Crapo, R; Hankinson, J; Brusasco, V; Burgos. F; Casaburi, R; Coates, A; Enright, P; van Per
Grinten, CP; Gustafsson, P; Jensen. R; Johnson. DC; Macintyre, N; Mckay, R; Navajas, D;
Pedersen, OF; Pellegrino, R; Viegi, G; Wanger, J; Force. AET. (2005a). General considerations for
lung function testing [Review], Eur RespirJ 26: 153-161.
http://dx.doi.org/10.1183/09031936.05.000345Q5
Miller. MR; Hankinson. J; Brusasco. V; Burgos. F; Casaburi. R; Coates. A; Crapo. R; Enright. P; van Per
Grinten. CP; Gustafsson. P; Jensen. R; Johnson. DC; Macintyre, N; Mckay, R; Navajas, D;
Pedersen. OF; Pellegrino. R; Viegi. G; Wanger. J; Force. AET. (2005b). Standardisation of
spirometry. Eur Respir J 26: 319-338. http://dx.doi.org/10.1183/09031936.05.000348Q5
Moeller, BC; Lu. K; Doyle-Eisele, M; Mcdonald. J; Gigliotti, A; Swenberg, JA. (2011). Determination of N2-
hydroxymethyl-dG adducts in the nasal epithelium and bone marrow of nonhuman primates
following 13CD2-formaldehyde inhalation exposure. Chem Res Toxicol 24: 162-164.
http://dx.doi.org/10.1021/txl004166
Monticello. TM; Miller. FJ; Morgan. KT. (1991). Regional increases in rat nasal epithelial cell proliferation
following acute and subchronic inhalation of formaldehyde. Toxicol Appl Pharmacol 111: 409-
421. http://dx.doi.org/10.1016/0041-008X(91)90246-B
Monticello. TM; Morgan. KT; Everitt. Jl; Popp. JA. (1989). Effects of formaldehyde gas on the respiratory
tract of rhesus monkeys: Pathology and cell proliferation. Am J Pathol 134: 515-527.
Monticello. TM; Swenberg. JA; Gross. EA; Leininger, JR; Kimbell, JS; Seilkop, S; Starr. TB; Gibson. JE;
Morgan. KT. (1996). Correlation of regional and nonlinear formaldehyde-induced nasal cancer
with proliferating populations of cells. Cancer Res 56: 1012-1022.
Morgan. PL; Dixon. D; King. DH; Travlos, GS; Herbert. RA; French. JE; Tokar, EJ; Waalkes, MP; Jokinen,
MP. (2017). NTP research report on absence of formaldehyde-induced neoplasia in Trp53
haploinsufficient mice exposed by inhalation. (Research Report 3). Research Triangle Park, NC:
This document is a draft for review purposes only and does not constitute Agency policy.
169 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
National Toxicology Program.
https://ntp.niehs.nih.gov/ntp/results/pubs/rr/reports/rr03 508.pdf
Morgan. KT. (1983). Localization of areas of inhibition of nasal mucociliary function in rats following in
vivo exposure to formaldehyde. Am Rev Respir Dis 127: 166.
Morgan. KT. (1997). A brief review of formaldehyde carcinogenesis in relation to rat nasal pathology and
human health risk assessment [Review], Toxicol Pathol 25: 291-305.
http://dx.doi.org/10.1177/0192623397025003Q7
Morgan. KT; Gross. EA; Patterson. PL. (1986a). Distribution, progression, and recovery of acute
formaldehyde-induced inhibition of nasal mucociliary function in F-344 rats. Toxicol Appl
Pharmacol 86: 448-456. http://dx.doi.org/10.1016/0041-008X(86)90372-8
Morgan. KT; Jiang. XZ; Starr. TB; Kerns. WD. (1986b). More precise localization of nasal tumors
associated with chronic exposure of F-344 rats to formaldehyde gas. Toxicol Appl Pharmacol 82:
264-271. http://dx.doi.org/10.1016/0041-008X(86)90201-2
Morgan. KT; Patterson. PL; Gross. EA. (1986c). Responses of the nasal mucociliary apparatus of F-344
rats to formaldehyde gas. Toxicol Appl Pharmacol 82: 1-13. http://dx.doi.org/10.1016/0Q41-
008X(86)90431-X
Mueller. JU; Bruckner. T; Triebig, G. (2013). Exposure study to examine chemosensory effects of
formaldehyde on hyposensitive and hypersensitive males. Int Arch Occup Environ Health 86:
107-117. http://dx.doi.org/10.1007/s00420-012-Q745-9
Neamtiu. IA; Lin. S; Chen. ML; Roba. C; Csobod. E. va; Gurzau. ES. (2019). Assessment of formaldehyde
levels in relation to respiratory and allergic symptoms in children from Alba County schools,
Romania. Environ Monit Assess 191: 591. http://dx.doi.org/10.1007/slQ661-019-7768-6
Neuss, S; Zeller, J; Ma-Hock, L; Speit, G. (2010). Inhalation of formaldehyde does not induce genotoxic
effects in broncho-alveolar lavage (BAL) cells of rats. Mutat Res Genet Toxicol Environ Mutagen
695: 61-68. http://dx.doi.Org/10.1016/i.mrgentox.2009.12.001
Norback, D; Bjornsson, E; Janson, C; Widstrom, J; Boman, G. (1995). Asthmatic symptoms and volatile
organic compounds, formaldehyde, and carbon dioxide in dwellings. Occup Environ Med 52:
388-395. http://dx.doi.Org/10.1136/oem.52.6.388
NRC (National Research Council). (2011). Review of the Environmental Protection Agency's draft IRIS
assessment of formaldehyde (pp. 1-194). Washington, DC: The National Academies Press.
http://dx.doi.org/10.17226/13142
NRC (National Research Council). (2014a). Review of EPA's Integrated Risk Information System (IRIS)
process. Washington, DC: The National Academies Press, http://dx.doi.org/10.17226/18764
NRC (National Research Council). (2014b). Review of the Formaldehyde Assessment in the National
Toxicology Program 12th Report on Carcinogens. Washington (DC): National Academies Press
(US), http://dx.doi.org/10.17226/18948
NTP (National Toxicology Program). (2015). Handbook for conducting a literature-based health
assessment using OHAT approach for systematic review and evidence integration. U.S. Dept. of
Health and Human Services, National Toxicology Program.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookian2015 508.pdf
Nunn. AJ; Craigen. AA; Darbvshire. JH; Venables. KM; Taylor. AJN. (1990). Six year follow up of lung
function in men occupationally exposed to formaldehyde. Br J Ind Med 47: 747-752.
http://dx.doi.org/10.1136/oem.47.ll.747
OEHHA (California Office of Environmental Health Hazard Assessment). (2008). Air toxics hot spots
program technical support document for the derivation of noncancer reference exposure levels.
Appendix G: Value of the Haber's Law Exponent (n) for various gases and vapors for acute RELs
developed using OEHHA (1999) procedures. Oakland, CA: California Environmental Protection
Agency, https://oehha.ca.gov/media/downloads/crnr/appendixgnexponentfinal.pdf
This document is a draft for review purposes only and does not constitute Agency policy.
170 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Oiaiarvi. IA; Partanen. TJ; Ahlbom. A; Boffetta. P; Hakulinen. T; Jourenkova. N; Kauppinen. TP; Kogevinas.
M; Porta. M; Vainio. HU; Weiderpass. E; Wesseling. CH. (2000). Occupational exposures and
pancreatic cancer: A meta-analysis. Occup Environ Med 57: 316-324.
http://dx.doi.Org/10.1136/oem.57.5.316
Olsen, JH; Asnaes, S. (1986). Formaldehyde and risk of squamous cell carcinoma of the sino nasal
cavities. Occup Environ Med 43: 769-774.
Olsen. JH; Dossing. M. (1982). Formaldehyde induced symptoms in day care centers. AIHA J 43: 366-370.
http://dx.doi.org/10.1080/15298668291409866
Ott, MG; Teta, J; Greenberg, HL. (1989). Lymphatic and hematopoietic tissue cancer in a chemical
manufacturing environment. Am J Ind Med 16: 631-644.
http://dx.doi.org/10.1002/aiim.47001606Q3
Ozen, OA; Akpolat, N; Songur, A; Kus, I; Zararsiz, I; Ozacmak. VH; Sarsilmaz, M. (2005). Effect of
formaldehyde inhalation on Hsp70 in seminiferous tubules of rat testes: An
immunohistochemical study. Toxicol Ind Health 21: 249-254.
http://dx.doi.org/10.1191/0748233705th235oa
Ozen. OA; Kus. MA; Kus. I; Alkoc, OA; Songur. A. (2008). Protective effects of melatonin against
formaldehyde-induced oxidative damage and apoptosis in rat testes: An immunohistochemical
and biochemical study. Sys Biol Reprod Med 54: 169-176.
http://dx.doi.org/10.1080/193963608024224Q2
Ozen. OA; Yaman. M; Sarsilmaz. M; Songur. A; Kus. I. (2002). Testicular zinc, copper and iron
concentrations in male rats exposed to subacute and subchronic formaldehyde gas inhalation. J
Trace Elem Med Biol 16: 119-122. http://dx.doi.org/10.1016/S0946-672X(02)80038-4
Palczynski, C; Krakowiak, A; Hanke, W; Walusiak, J; Gorski, P. (1999). Indoor formaldehyde exposure and
airway allergic diseases. Int Rev Allergol Clin Immunol 5: 65-69.
Peddada, SD; Dinse, GE; Kissling, GE. (2007). Incorporating historical control data when comparing
tumor incidence rates. J Am Stat Assoc 102: 1212-1220.
http://dx.doi.org/10.1198/0162145060000Q1356
Pellegrino, R; Viegi, G; Brusasco, V; Crapo, RO; Burgos. F; Casaburi, R; Coates, A; van Per Grinten, CP;
Gustafsson, P; Hankinson, J; Jensen. R; Johnson. DC; Macintyre, N; Mckay, R; Miller. MR;
Navajas, D; Pedersen, OF; Wanger, J. (2005). Interpretative strategies for lung function tests. Eur
Respir J 26: 948-968. http://dx.doi.org/10.1183/09031936.05.000352Q5
Percy, C; Stanek, E. Ill; Gloeckler, L. (1981). Accuracy of cancer death certificates and its effect on cancer
mortality statistics. Am J Public Health 71: 242-250. http://dx.doi.Org/10.2105/AJPH.71.3.242
Percy, CL; Miller. BA; Gloeckler Ries, LA. (1990). Effect of changes in cancer classification and the
accuracy of cancer death certificates on trends in cancer mortality. Ann N Y Acad Sci 609: 87-99.
http://dx.doi.Org/10.llll/i.1749-6632.1990.tb32059.x
Pesch. B; Pierl. CB; Gebel. M; Gross. I; Becker. D; Johnen. G; Rihs. HP; Donhuiisen. K; Lepentsiotis. V;
Meier. M; Schulze. J; Bruning. T. (2008). Occupational risks for adenocarcinoma of the nasal
cavity and paranasal sinuses in the German wood industry. Occup Environ Med 65: 191-196.
http://dx.doi.org/10.1136/oem.2007.033886
Peteffi. GP; Basso da Silva. L; Antunes. MV; Wilhelm. C; Valandro. ET; Glaeser. J; Kaefer. D; Linden. R.
(2015). Evaluation of genotoxicity in workers exposed to low levels of formaldehyde in a
furniture manufacturing facility. Toxicol Ind Health 32: 1763-1773.
http://dx.doi.org/10.1177/0748233715584250
Peters. TL; Kamel, F; Lundholm, C; Feychting, M; Weibull, CE; Sandler. DP; Wiebert, P; Sparen, P; Ye. W;
Fang. F. (2017). Occupational exposures and the risk of amyotrophic lateral sclerosis. Occup
Environ Med 74: 87-92. http://dx.doi.org/10.1136/oemed-2016-10370Q
This document is a draft for review purposes only and does not constitute Agency policy.
171 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Toxicological Review of Formaldehyde - Inhalation
Pinkerton. LE; Hein. MJ; Meyers. A; Kamel. F. (2013). Assessment of ALS mortality in a cohort of
formaldehyde-exposed garment workers. 14: 353-355.
http://dx.doi.org/10.3109/21678421.2013.778284
Pira, E; Romano. C; Verga, F; La Vecchia, C. (2014). Mortality from lymphohematopoietic neoplasms and
other causes in a cohort of laminated plastic workers exposed to formaldehyde. Cancer Causes
Control 25: 1343-1349. http://dx.doi.org/10.1007/slQ552-014-0440-0
Portier, CJ; Bailer. AJ. (1989). Testing for increased carcinogenicity using a survival-adjusted quantal
response test. Fundam Appl Toxicol 12: 731-737.
Pottern, LM; Heineman, EF; Olsen, JH; Raffn, E; Blair. A. (1992). Multiple myeloma among Danish
women: Employment history and workplace exposures. Cancer Causes Control 3: 427-432.
http://dx.doi.org/10.1007/BF00Q51355
Raaschou-Nielsen, O; Hermansen, MINI; Loland, L; Buchvald, F; Pipper, CB; Sgrensen, M; Loft. S; Bisgaard,
K, (2010). Long-term exposure to indoor air pollution and wheezing symptoms in infants. Indoor
Air 20: 159-167. http://dx.doi.Org/10.llll/i.1600-0668.2009.00635.x
Rager, JE; Moeller, BC; Miller. SK; Kracko, D; Doyle-Eisele, M; Swenberg, JA; Fry, RC. (2014).
Formaldehyde-Associated Changes in microRNAs: Tissue and Temporal Specificity in the Rat
Nose, White Blood Cells, and Bone Marrow. Toxicol Sci 138: 36-46.
http://dx.doi.org/10.1093/toxsci/kft267
Redlich. CA: Tarlo. SM: Hankinson. JL: Townsend. MC: Eschenbacher. WL: Von Essen. SG: Sigsgaard. T:
Weissman. DN: Setting. ATSCoSitO. (2014). Official American Thoracic Society technical
standards: spirometry in the occupational setting : Supplementary materials [Supplemental
Data], Am J Respir Crit Care Med 189.
Riedel, F; Hasenauer, E; Barth, PJ; Koziorowski, A; Rieger, CHL. (1996). Formaldehyde exposure enhances
inhalative allergic sensitization in the guinea pig. Allergy 51: 94-99.
http://dx.doi.Org/10.llll/i.1398-9995.1996.tb00041.x
Roberts. AL; Johnson. NJ; Cudkowicz, ME; Eum, KD; Weisskopf, MG. (2016). Job-related formaldehyde
exposure and ALS mortality in the USA [Letter], J Neurol Neurosurg Psychiatry 87: 786-788.
http://dx.doi.org/10.1136/innp-2015-310750
Robinson. CF; Fowler. D; Brown. DP; Lemen, RA. (1987). Plywood mill workers' mortality patterns 1945
1977 (revised March 1987). (NIOSH/00197140). Cincinnati, OH: NIOSH.
Roda, C; Kousignian, I; Guihenneuc-Jouyaux, C; Dassonville, C; Nicolis, I; Just. J; Momas, I. (2011).
Formaldehyde exposure and lower respiratory infections in infants: findings from the PARIS
cohort study. Environ Health Perspect 119: 1653-1658. http://dx.doi.org/10.1289/ehp.1003222
Romanazzi, V; Pirro, V; Bellisario, V; Mengozzi, G; Peluso, M; Pazzi, M; Bugiani, M; Verlato, G; Bono. R.
(2013). 15-F2t isoprostane as biomarker of oxidative stress induced by tobacco smoke and
occupational exposure to formaldehyde in workers of plastic laminates. Sci Total Environ 442:
20-25. http://dx.doi.Org/10.1016/i.scitotenv.2012.10.057
Rothman. N; Lan. Q; Smith. MT; Vermeulen. R; Zhang. L. (2017). Response to letter to the editor of
Carcinogenesis by Pira et al., 2017 [Letter], Carcinogenesis 38: 1253-1255.
http://dx.doi.org/10.1093/carcin/bgxlll
Roush. GC; Walrath. J; Stavner. LT; Kaplan. SA; Flannerv. JT; Blair. A. (1987), Nasopharyngeal cancer
sinonasal cancer and occupations related to formaldehyde: A case-control study. J Natl Cancer
Inst 79: 1221-1224.
Rumchev, KB; Spickett, JT; Bulsara, MK; Phillips. MR; Stick. SM. (2002). Domestic exposure to
formaldehyde significantly increases the risk of asthma in young children. Eur Respir J 20: 403-
408. http://dx.doi.org/10.1183/09031936.02.002450Q2
Saberi Hosnijeh, F; Christopher. Y; Peeters, P; Romieu, I; Xun, W; Riboli, E; Raaschou-Nielsen. O;
Tjgnneland, A; Becker. N; Nieters, A; Trichopoulou, A; Bamia, C; Orfanos, P; Oddone, E; Lujan-
This document is a draft for review purposes only and does not constitute Agency policy.
172 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Toxicological Review of Formaldehyde - Inhalation
Barroso. L; Dorronsoro. M; Navarro. C; Barricarte. A; Molina-Montes. E; Wareham. N; Vineis. P;
Vermeulen. R. (2013). Occupation and risk of lymphoid and myeloid leukaemia in the European
Prospective Investigation into Cancer and Nutrition (EPIC). Occup Environ Med 70: 464-470.
http://dx.doi.org/10.1136/oemed-2012-101135
Sapmaz, HI; Yildiz, A; Polat, A; Vardi, N; Kose, E; Tanbek, K; Cuglan, S. (2018). Protective efficacy of
Nigella sativa oil against the harmful effects of formaldehyde on rat testicular tissue. 8: 548-
553. http://dx.doi.org/10.4103/2221-1691.245970
Sari. DK; Kuwahara, S; Tsukamoto, Y; Hori, H; Kunugita, N; Arashidani, K; Fujimaki, H; Sasaki. F. (2004).
Effect of prolonged exposure to low concentrations of formaldehyde on the corticotropin
releasing hormone neurons in the hypothalamus and adrenocorticotropic hormone cells in the
pituitary gland in female mice. Brain Res 1013: 107-116.
http://dx.doi.Org/10.1016/i.brainres.2004.03.070
Sarsilmaz, M; Ozen, OA; Akpolat, N; Kus, I; Songur, A. (1999). The histopathologic effects of inhaled
formaldehyde on leydig cells of the rats in subacute period. Firat Univ Med Sci J 13: 37-40.
Sauder, LR; Chatham. MP; Green. DJ; Kulle, TJ. (1986). Acute pulmonary response to formaldehyde
exposure in healthy nonsmokers. J Occup Environ Med 28: 420-424.
http://dx.doi.org/10.1097/00043764-198606000-000Q8
Saurel-Cubizolles. MJ; Hays. M; Estrvn-Behar. M. (1994). Work in operating rooms and pregnancy
outcome among nurses. Int Arch Occup Environ Health 66: 235-241.
http://dx.doi.org/10.1007/bf0Q454361
Schachter. EN; Witek. TJ. Jr; Brodv. DJ; Tosun. T; Beck. GJ; Leaderer. BP. (1987). A study of respiratory
effects from exposure to 2.0 ppm formaldehyde in occupationally exposed workers. Environ Res
44: 188-205. http://dx.doi.org/10.1016/S0013-9351(87)80227-X
Schachter. EN; Witek. TJ. Jr; Tosun. T; Leaderer. BP; Beck. GJ. (1986). A study of respiratory effects from
exposure to 2 ppm formaldehyde in healthy subjects. Arch Environ Occup Health 41: 229-239.
http://dx.doi.org/10.1080/00Q39896.1986.9938338
Schlosser, PM; Lilly, PD; Conolly, RB; Janszen, DB; Kimbell, JS. (2003). Benchmark dose risk assessment
for formaldehyde using airflow modeling and a single-compartment, DNA-protein cross-link
dosimetry model to estimate human equivalent doses. Risk Anal 23: 473-487.
http://dx.doi.org/10.llll/1539-6924.00328
Schroeder, EB; Welch. VL; Couper, D; Nieto, FJ; Liao, DP; Rosamond. WD; Heiss, G. (2003). Lung function
and incident coronary heart disease - The atherosclerosis risk in communities study. Am J
Epidemiol 158: 1171-1181. http://dx.doi.org/10.1093/aie/kwg276
Schroeter, JD; Campbell. J; Kimbell. JS; Conolly, RB; Clewell, HJ; Andersen. ME. (2014). Effects of
endogenous formaldehyde in nasal tissues on inhaled formaldehyde dosimetry predictions in
the rat, monkey, and human nasal passages. Toxicol Sci 138: 412-424.
http://dx.doi.org/10.1093/toxsci/kft333
Schunemann. HJ; Porn. J; Grant. BJB; Winkelstein. W. Jr; Trevisan. M. (2000). Pulmonary function is a
long-term predictor of mortality in the general population: 29-year follow-up of the Buffalo
health study. Chest 118: 656-664. http://dx.doi.Org/10.1378/chest.118.3.656
Seals. RM; Kioumourtzoglou. MA; Gredal. O; Hansen. J; Weisskopf. MG. (2017). Occupational
formaldehyde and amyotrophic lateral sclerosis. Eur J Epidemiol 32: 893-899.
http://dx.doi.org/10.1007/slQ654-017-0249-8
Sellakumar, AR; Snyder, CA; Solomon. JJ; Albert. RE. (1985). Carcinogenicity of formaldehyde and
hydrogen chloride in rats. Toxicol Appl Pharmacol 81: 401-406. http://dx.doi.org/10.1016/0Q41-
008X(85)90411-9
This document is a draft for review purposes only and does not constitute Agency policy.
173 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Shaham. J; Bomstein. Y; Gurvich. R; Rashkovskv. M; Kaufman. Z. (2003). DNA-protein crosslinks and p53
protein expression in relation to occupational exposure to formaldehyde. Occup Environ Med
60: 403-409. http://dx.doi.Org/10.1136/oem.60.6.403
Siew, SS; Kauppinen, T; Kyyronen, P; Heikkila, P; Pukkala, E. (2012). Occupational exposure to wood dust
and formaldehyde and risk of nasal, nasopharyngeal, and lung cancer among Finnish men.
Cancer Management and Research 4: 223-232. http://dx.doi.org/10.2147/CMAR.S30684
Sin. DP; Wu. LL; Man. SFP. (2005). The relationship between reduced lung function and cardiovascular
mortality: a population-based study and a systematic review of the literature [Review], Chest
127: 1952-1959. http://dx.doi.Org/10.1378/chest.127.6.1952
Solet, D; Zoloth, SR; Sullivan. C; Jewett, J; Michaels. DM. (1989). Patterns of mortality in pulp and paper
workers. J Occup Med 31: 627-630.
Sorg, BA; Bailie. TM; Tschirgi, ML; Li. N; Wu. WR. (2001). Exposure to repeated low-level formaldehyde in
rats increases basal corticosterone levels and enhances the corticosterone response to
subsequent formaldehyde. Brain Res 898: 314-320. http://dx.doi.org/10.1016/S00Q6-
8993(01)02208-9
Sorlie, PD; Kannel, WB; O'Connor. G. (1989). Mortality associated with respiratory function and
symptoms in advanced age: the Framingham study. Am J Respir Crit Care Med 140: 379-384.
http://dx.doi.Org/10.1164/airccm/140.2.379
Speit. G: Schutz. P: Weber. I: Ma-Hock. L: Kaufmann. W; Gelbke. HP: Durrer. S. (2011). Analysis of
micronuclei, histopathological changes and cell proliferation in nasal epithelium cells of rats
after exposure to formaldehyde by inhalation. Mutat Res 721: 127-135.
http://dx.doi.Org/10.1016/i.mrgentox.2011.01.008
Starr. TB; Swenberg, JA. (2013). A novel bottom-up approach to bounding low-dose human cancer risks
from chemical exposures. Regul Toxicol Pharmacol 65: 311-315.
http://dx.doi.Org/10.1016/i.vrtph.2013.01.004
Starr. TB; Swenberg. JA. (2016). The bottom-up approach to bounding potential low-dose cancer risks
from formaldehyde: An update. Regul Toxicol Pharmacol 77: 167-174.
http://dx.doi.Org/10.1016/i.vrtph.2016.01.021
Stayner, LT; Elliott. L; Blade. L; Keenlyside, R; Halperin, W. (1988). A retrospective cohort mortality study
of workers exposed to formaldehyde in the garment industry. Am J Ind Med 13: 667-681.
http://dx.doi.org/10.1002/aiim.47001306Q6
Steele. LL; Wilkins, J. R. (1996). Occupational exposures and risks of spontaneous abortion among female
veterinarians. Int J Occup Environ Health 2: 26-36.
Stellman, SD; Demers, PA; Colin. D; Boffetta, P. (1998). Cancer mortality and wood dust exposure among
participants in the American Cancer Society Cancer Prevention Study-ll (CPS-II). Am J Ind Med
34: 229-237. http://dx.doi.org/10.1002/(SICI)1097-0274(199809)34:3<229::AID-AJIM4>3.0.CQ;2-
Q
Stewart. PA; Blair. A; Cubit. DA; Bales. RE; Kaplan. SA; Ward. J; Gaffev. W; O'Berg. MT; Walrath. J. (1986).
Estimating historical exposures to formaldehyde in a retrospective mortality study. Appl Ind Hyg
1: 34-41.
Stroup. NE; Blair. A; Erikson. GE. (1986). Brain cancer and other causes of death in anatomists. J Natl
Cancer Inst 77: 1217-1224.
Stucker, I; Caillard, JF; Collin. R; Gout. M; Poyen, D; Hemon, D. (1990). Risk of spontaneous abortion
among nurses handling antineoplastic drugs. Scand J Work Environ Health 16: 102-107.
http://dx.doi.org/10.5271/siweh.1811
Subramaniam, RP; Crump. KS; Van Landingham, C; White. P; Chen. C; Schlosser, PM. (2007).
Uncertainties in the CUT model for formaldehyde-induced carcinogenicity in the rat: A limited
This document is a draft for review purposes only and does not constitute Agency policy.
174 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
sensitivity analysis-l. Risk Anal 27: 1237-1254. http://dx.doi.Org/10.llll/i. 1539-
6924.2007.00968.x
Swenberg, JA; Lu. K; Moeller, BC; Gao, L; Upton. PB; Nakamura, J; Starr. TB. (2011). Endogenous versus
exogenous DNA adducts: Their role in carcinogenesis, epidemiology, and risk assessment
[Review], Toxicol Sci 120: S130-S145. http://dx.doi.org/10.1093/toxsci/kfq371
Swiecichowski, AL; Long. KJ; Miller. ML; Leikauf, GD. (1993). Formaldehyde-induced airway
hyperreactivity in vivo and ex vivo in guinea pigs. Environ Res 61: 185-199.
http://dx.doi.org/10.1006/enrs.1993.1063
Taffet, GE; Donohue, JF; Altman, PR. (2014). Considerations for managing chronic obstructive pulmonary
disease in the elderly [Review], Clinical Interventions in Aging 9: 23-30.
http://dx.doi.org/10.2147/CIA.S52999
Talibov, M; Lehtinen-Jacks, S; Martinsen, J I; Kjaerheim, K; Lynge, E; Sparen, P; Tryggvadottir, L;
Weiderpass, E; Kauppinen, T; Kyyronen, P; Pukkala, E. (2014). Occupational exposure to solvents
and acute myeloid leukemia: A population-based, case-control study in four Nordic countries.
Scand J Work Environ Health 40: 511-517. http://dx.doi.org/10.5271/siweh.3436
Taskinen, H; Kyyronen. P; Hemminki, K. (1994). Laboratory work and pregnancy outcome. J Occup Med
36: 311-319. http://dx.doi.org/10.1097/00043764-199403000-000Q8
Taskinen. HK: Kyyronen. P: Sallmen. M: Virtanen. SV: Liukkonen. TA: Huida. O: Lindbohm. ML: Anttila. A.
(1999). Reduced fertility among female wood workers exposed to formaldehyde. Am J Ind Med
36: 206-212. http://dx.doi.org/10.1002/(sici)1097-0274(199907)36:l<206::aid-aiim29>3.0.co:2-
d
ten Berge. WF: Zwart. A: Appelman. LM. (1986). Concentration-time mortality response relationship of
irritant and systemically acting vapours and gases. J Hazard Mater 13: 301-309.
http://dx.doi. org/10.1016/0304-3894(86)85003-8
Tepper, RS; Wise. RS; Covar, R; Irvin, CG; Kercsmar, CM; Kraft. M; Liu. MC; O'Connor. GT; Peters. SP;
Sorkness, R; Togias, A. (2012). Asthma outcomes: Pulmonary physiology [Review], J Allergy Clin
Immunol 129: S65-S87. http://dx.doi.Org/10.1016/i.iaci.2011.12.986
Teschke, K; Morgan. MS; Checkoway, H; Franklin. G; Spinelli, JJ; van Belle. G; Weiss. IMS. (1997).
Surveillance of nasal and bladder cancer to locate sources of exposure to occupational
carcinogens. Occup Environ Med 54: 443-451. http://dx.doi.Org/10.1136/oem.54.6.443
Til. HP; Woutersen, RA; Feron, VJ; Hollanders. VHM; Falker, HE; Clary, JJ. (1989). Two-year drinking-
water study of formaldehyde in rats. Food Chem Toxicol 27: 77-87.
http://dx.doi.org/10.1016/0278-6915(89)90001-X
Tomashefski, J. (2008). Dail and Hammar's pulmonary pathology: Neoplastic lung disease. In J
Tomashefski (Ed.), (3rd ed.). New York, NY: Springer.
Torjussen, W; Solberg, LA; Hogetveit, AC. (1979). Histopathological changes of the nasal mucosa in
active and retired nickel workers. Br J Cancer 40: 568-580.
Tsubone. H; Kawata. M. (1991). Stimulation to the trigeminal afferent nerve of the nose by
formaldehyde, acrolein, and acetaldehyde gases. Inhal Toxicol 3: 211-222.
http://dx.doi.org/10.3109/08958379109145285
U.S. EPA (U.S. Environmental Protection Agency). (1993). [Printout of carcinogenicity data for
formaldehyde as verified 2/3/88] [Database],
U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry [EPA Report], (EPA/600/8-90/066F).
Research Triangle Park, NC.
https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=71993&CFID=51174829&CFTOKEN=250
06317
This document is a draft for review purposes only and does not constitute Agency policy.
175 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Toxicological Review of Formaldehyde - Inhalation
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk assessment
[EPA Report] (pp. 1-89). (EPA/630/R-95/001F). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://www.epa.gov/risk/guidelines-neurotoxicity-risk-
assessment
U.S. EPA (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen risk assessment
[EPA Report], (EPA630P03001F). Washington, DC.
https://www.epa.gov/sites/production/files/2013-09/documents/cancer guidelines final 3-25-
05.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for assessing
susceptibility from early-life exposure to carcinogens [EPA Report], (EPA/630/R-03/003F).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
https://www.epa.gov/risk/supplemental-guidance-assessing-susceptibilitv-early-life-exposure-
carcinogens
U.S. EPA (U.S. Environmental Protection Agency). (2010). Toxicological Review of Formaldehyde
(Inhalation) (External Review Draft 2010).
http://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=223614
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance [EPA
Report], (EPA/100/R-12/001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https://www.epa.gov/risk/benchmark-dose-technical-guidance
Valencia. K: Martin-Fernandez. M: Zandueta. C: Ormazabal. C: Martmez-Canarias. S: Bandres. E: de la
Piedra. C: Lecanda. F. (2013). miR-326 associates with biochemical markers of bone turnover in
lung cancer bone metastasis. Bone 52: 532-539. http://dx.doi.Org/10.1016/i.bone.2012.10.033
Vaughan, TL. (1989). Occupation and squamous cell cancers of the pharynx and sinonasal cavity. Am J
Ind Med 16: 493-510. http://dx.doi.org/10.1002/aiim.47001605Q3
Vaughan. TL; Stewart. PA; Teschke, K; Lynch, CF; Swanson, GM; Lyon, JL; Berwick. M. (2000).
Occupational exposure to formaldehyde and wood dust and nasopharyngeal carcinoma. Occup
Environ Med 57: 376-384. http://dx.doi.Org/10.1136/oem.57.6.376
Vaughan. TL; Strader, C; Davis. S; Paling, JR. (1986a). Formaldehyde and cancers of the pharynx, sinus
and nasal cavity: I. Occupational exposures. Int J Cancer 38: 677-683.
http://dx.doi.org/10.1002/iic.291038051Q
Vaughan. TL; Strader. C; Davis. S; Paling, JR. (1986b). Formaldehyde and cancers of the pharynx, sinus
and nasal cavity: II. Residential exposures. Int J Cancer 38: 685-688.
http://dx.doi.org/10.1002/iic.291038Q511
Venn. AJ; Cooper. M; Antoniak, M; Laughlin, C; Britton, J; Lewis. SA. (2003). Effects of volatile organic
compounds, damp, and other environmental exposures in the home on wheezing illness in
children. Thorax 58: 955-960. http://dx.doi.org/10.1136/thorax.58.ll.955
Viegas. S; Ladeira. C; Gomes. M; Nunes. C; Brito. M; Prista. J. (2013). Exposure and genotoxicity
assessment methodologies - the case of formaldehyde occupational exposure. Current
Analytical Chemistry 9: 476-484. http://dx.doi.org/10.2174/1573411011309030Q17
Viegas. S; Ladeira. C; Nunes. C; Malta-Vacas. J; Gomes. M; Brito. M; Mendonca. P; Prista. J. (2010).
Genotoxic effects in occupational exposure to formaldehyde: A study in anatomy and pathology
laboratories and formaldehyde-resins production. J Occup Med Toxicol 5: 25.
http://dx.doi.org/10.1186/1745-6673-5-25
Vosoughi, S; Khavanin, A; Salehnia, M; Asilian Mahabadi, H; Shahverdi, A; Esmaeili, V. (2013). Adverse
effects of formaldehyde vapor on mouse sperm parameters and testicular tissue. Int J Fertility
Sterility 6: 250-257.
This document is a draft for review purposes only and does not constitute Agency policy.
176 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Toxicological Review of Formaldehyde - Inhalation
Vosoughi. S; Khavanin. A; Salehnia. M; Mahabadi. HA; Soleimanian. A. (2012). Effects of simultaneous
exposure to formaldehyde vapor and noise on mouse testicular tissue and sperm parameters.
Health Scope 1: 110-117. http://dx.doi.org/10.17795/ihealthscope-7973
Wallner, P; Kundi, M; Moshammer, H; Piegler, K; Hohenblum, P; Scharf, S; Frohlich, M; Damberger, B;
Tapplere, P; Hutter, HP. (2012). Indoor air in schools and lung function of Austrian school
children. J Environ Monit 14: 1976-1982. http://dx.doi.org/10.1039/c2em30059a
Walrath, J; Fraumeni, JF. Jr. (1983). Mortality patterns among embalmers. Int J Cancer 31: 407-411.
http://dx.doi.org/10.1002/iic.29103104Q3
Walrath. J; Fraumeni. JF. Jr. (1984). Cancer and other causes of death among embalmers. Cancer Res 44:
4638-4641.
Wang. H; Li. H. eC; Lv. M; Zhou. D; Bai, L; Du. L; Xue, X. ia; Lin. P. u; Qiu, S. (2015). Associations between
occupation exposure to Formaldehyde and semen quality, a primary study. Sci Rep 5: 15874.
http://dx.doi.org/10.1038/srepl5874
Wang. HX; Zhou. DX; Zheng. LR; Zhang. J; Huo, YW; Tian, H; Han. SP; Zhang. J; Zhao. WB. (2012). Effects
of paternal occupation exposure to formaldehyde on reproductive outcomes. J Occup Environ
Med 54: 518-524. http://dx.doi.org/10.1097/JQM.0b013e31824e6937
Wang. K; Wang. TW; Xu. J; Zhu, Y; Jian, L; Au. W; Xia, ZL. (2019). Determination of benchmark dose
based on adduct and micronucleus formations in formaldehyde-exposed workers. Int J Hyg
Environ Health 222: 738-743. http://dx.doi.Org/10.1016/i.iiheh.2019.05.008
Ward. JB. Jr: Hokanson. JA: Smith. ER: Chang. LW: Pereira. MA: Whorton. EB. Jr: Legator. MS. (1984).
Sperm count, morphology and fluorescent body frequency in autopsy service workers exposed
to formaldehyde. Mutat Res Environ Mutagen Relat Subj 130: 417-424.
http://dx.doi.org/10.1016/0165-1161(84)90014-1
Weisel, CP; Zhang. J; Turpin, BJ; Morandi, MT; Colome, S; Stock. TH; Spektor, DM; Korn, L; Winer. AM;
Kwon, J; Meng, QY; Zhang. L; Harrington. R; Liu. W; Reff, A; Lee. JH; Alimokhtari, S; Mohan. K;
Shendell, D; Jones. J; Farrar, L; Maberti, S; Fan. T. (2005). Relationships of indoor, outdoor, and
personal air (RIOPA): Part I. Collection methods and descriptive analyses (pp. 1-107; discussion
109-127). (ISSN 1041-5505
EISSN 2688-6855
HEI Research Report 130-1, NUATRC Research Report 7). Boston, MA: Health Effects Institute.
Weisskopf, M; Morozova, N; O'Reilly. EJ; Mccullough, ML; Calle, EE;Thun, MJ; Ascherio, A. (2009).
Prospective study of chemical exposures and amyotrophic lateral sclerosis mortality. J Neurol
Neurosurg Psychiatry 80: 558-561. http://dx.doi.org/10.1136/innp.2008.156976
Werner. A; Meinhardt, A; Seitz, J; Bergmann, M. (1997). Distribution of heat-shock protein 60
immunoreactivity in testes of infertile men. Cell Tissue Res 288: 539-544.
http://dx.doi.org/10.1007/s00441005Q839
Wilmer. JWG. M; Woutersen. RA; Appelman. LM; Leeman. WR; Feron. VJ. (1987). Subacute (4-week)
inhalation toxicity study of formaldehyde in male rats: 8-hour intermittent versus 8-hour
continuous exposures. J Appl Toxicol 7: 15-16. http://dx.doi.org/10.1002/iat.25500701Q4
Wilmer. JWG. M; Woutersen. RA; Appelman. LM; Leeman. WR; Feron. VJ. (1989). Subchronic (13-week)
inhalation toxicity study of formaldehyde in male rats: 8-hour intermittent versus 8-hour
continuous exposures. Toxicol Lett 47: 287-293. http://dx.doi.org/10.1016/Q378-
4274(89)90147-1
Woutersen. RA; Appelman. LM; Wilmer. JWG. M; Falke, HE; Feron. VJ. (1987). Subchronic (13-week)
inhalation toxicity study of formaldehyde in rats. J Appl Toxicol 7: 43-49.
http://dx.doi.org/10.1002/iat.25500701Q8
This document is a draft for review purposes only and does not constitute Agency policy.
Ill DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Toxicological Review of Formaldehyde - Inhalation
Woutersen. RA; van Garderen-Hoetmer. A; Bruiinties. JP; Zwart. A; Feron. VJ. (1989). Nasal tumours in
rats after severe injury to the nasal mucosa and prolonged exposure to 10 ppm formaldehyde. J
Appl Toxicol 9: 39-46. http://dx.doi.org/10.1002/iat.25500901Q8
Xing. SY; Ye. L; Wang. NN. (2007). Toxic effect of formaldehyde on reproduction and heredity in male
mice. Journal of Jilin University - Medicine Edition 33: 716-718.
Ye. X; Ji. Z; Wei. C; Mchale, C; Ding. S; Thomas. R; Yang. X; Zhang. L. (2013). Inhaled formaldehyde
induces DNA-protein crosslinks and oxidative stress in the bone marrow and other distant
organs of exposed mice [Abstract], Environ Mol Mutagen 54: S41.
Ye. X; Yan, W; Xie, H; Zhao. M; Ying, C. (2005). Cytogenetic analysis of nasal mucosa cells and
lymphocytes from high-level long-term formaldehyde exposed workers and low-level short-term
exposed waiters. Mutat Res 588: 22-27. http://dx.doi.Org/10.1016/i.mrgentox.2005.08.005
Yon. DK; Hwang. S; Lee. SW; Jee, HM; Sheen. YH; Kim. JH; Lim, DH; Han. MY. (2019). Indoor Exposure and
Sensitization to Formaldehyde among Inner-City Children with Increased Risk for Asthma and
Rhinitis. Am J Respir Crit Care Med 200: 388-393. http://dx.doi.org/10.1164/rccm.20181Q-
1980LE
Young. RP; Hopkins. R; Eaton. TE. (2007). Forced expiratory volume in one second: not just a lung
function test but a marker of premature death from all causes [Review], Eur Respir J 30: 616-
622. http://dx.doi.org/10.1183/09031936.000217Q7
Yu. R: Lai. Y: Hartwell. HJ: Moeller. BC: Dovle-Eisele. M: Kracko. D: Bodnar. WM: Starr. TB: Swenberg. JA.
(2015). Formation, Accumulation, and Hydrolysis of Endogenous and Exogenous Formaldehyde-
Induced DNA Damage. Toxicol Sci 146: 170-182. http://dx.doi.org/10.1093/toxsci/kfv079
Zeller. J: Neuss. S: Mueller. JU: Kuhner. S: Holzmann. K: Hogel. J: Klingmann. C: Bruckner. T: Triebig. G:
Speit. G. (2011). Assessment of genotoxic effects and changes in gene expression in humans
exposed to formaldehyde by inhalation under controlled conditions. Mutagenesis 26: 555-561.
http://dx.doi.org/10.1093/mutage/ger016
Zendehdel, R; Vahabi, M; Sedghi, R. (2018). Estimation of formaldehyde occupational exposure limit
based on genetic damage in some Iranian exposed workers using benchmark dose method.
Environ Sci Pollut Res Int 25: 31183-31189. http://dx.doi.org/10.1007/sll356-018-3Q77-9
Zhai, L; Zhao. J; Xu. B; Deng. Y; Xu. Z. (2013). Influence of indoor formaldehyde pollution on respiratory
system health in the urban area of Shenyang, China. Afr Health Sci 13: 137-143.
http://dx.doi.org/10.4314/ahs.vl3il.19
Zhang. L; Steinmaus, C; Eastmond, DA; Xin, XK; Smith. MT. (2009). Formaldehyde exposure and
leukemia: A new meta-analysis and potential mechanisms [Review], Mutat Res Rev Mutat Res
681: 150-168. http://dx.doi.Org/10.1016/i.mrrev.2008.07.002
Zhang. L; Tang. X; Rothman, N; Vermeulen, R; Ji. Z; Shen, M; Qiu, C; Guo, W; Liu. S; Reiss, B; Freeman. LB;
Ge. Y; Hubbard. AE; Hua, M; Blair. A; Galvan, N; Ruan, X; Alter. BP; Xin. KX; Li. S; Moore. LE; Kim.
S; Xie. Y; Haves. RB; Azuma. M; Hauptmann. M; Xiong. J; Stewart. P; Li. L; Rappaport. SM; Huang.
H; Fraumeni. JF. Jr; Smith. MT; Lan. Q. (2010). Occupational exposure to formaldehyde,
hematotoxicity, and leukemia-specific chromosome changes in cultured myeloid progenitor
cells. Cancer Epidemiol Biomarkers Prev 19: 80-88. http://dx.doi.org/10.1158/1055-9965.EPI-Q9-
0762
Zhao. Y. un; Magana, LC; Cui, H; Huang. J; Mchale. CM; Yang. X. u; Looney, MR; Li. R. ui; Zhang. L. (2020).
Formaldehyde-induced hematopoietic stem and progenitor cell toxicity in mouse lung and nose.
Arch Toxicol 95: 693-701. http://dx.doi.org/10.1007/s00204-020-Q2932-x
Zhou. D; Zhang. J; Wang. H. (2011a). Assessment of the potential reproductive toxicity of long-term
exposure of adult male rats to low-dose formaldehyde. Toxicol Ind Health 27: 591-598.
http://dx.doi.org/10.1177/0748233710393401
This document is a draft for review purposes only and does not constitute Agency policy.
178 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Toxicological Review of Formaldehyde - Inhalation
Zhou. D; Zhang. J; Wang. H; Xue. Y. (2011b). Effect of formaldehyde exposure on structure and function
of epididymis in adult rats: A histological and biochemical study. Toxicol Environ Chem 93: 134-
144. http://dx.doi.org/10.1080/02772248.2010.5Q1145
Zhou. DX; Qiu, SD; Zhang. J; Tian, H; Wang. HX. (2006). The protective effect of vitamin E against
oxidative damage caused by formaldehyde in the testes of adult rats. Asian J Androl 8: 584-588.
http://dx.doi.Org/10.llll/i.1745-7262.2006.00198.x
Zhu, JL; Knudsen, LE; Andersen. AM; Hjollund, NH; Olsen, J. (2006). Laboratory work and pregnancy
outcomes: a study within the National Birth Cohort in Denmark. Occup Environ Med 63: 53-58.
http://dx.doi.org/10.1136/oem.2005.0212Q4
Zwart, A; Woutersen, RA; Wilmer, JWG, M; Spit. BJ; Feron, VJ. (1988). Cytotoxic and adaptive effects in
rat nasal epithelium after 3-day and 13-week exposure to low concentrations of formaldehyde
vapour. Toxicology 51: 87-99. http://dx.doi.org/10.1016/0300-483X(88)90083-2
This document is a draft for review purposes only and does not constitute Agency policy.
179 DRAFT—DO NOT CITE OR QUOTE
------- |