Assessment of Health Risks
to Garment Workers and Certain Home Residents
from Exposure to Formaldehyde
April 1987
Office of Pesticides and Toxic Substances
U.S. Environmental Protection Agency
-------
Table of Contents
Page
List of Tables v
List of Figures viii
Preface x
Authors, Contributors, and Reviewers xi
Executive Summary xiii
1. Risk Characterization 1-1
1.1. Non-Cancer Effects 1-1
1.1.1. Sensory Irritation 1-1
1.1.2. Cellular Changes 1-4
1.2 Carcinogenic Effects 1-6
1.2.1. Studies of Humans 1-7
1.2.2. Studies in Animals 1-12
1.2.3. Additional Supportive Evidence 1-14
1.3. Exposure in Residential and Apparel
Manufacturing Settings 1-16
1.3.1. Residential Exposure 1-17
1.3.2. Exposure in Apparel Manufacturing 1-19
1.4. Quantitative Risk Assessment 1-19
1.4.1. Non-cancer Risk Assessment 1-19
1.4.2. Cancer Dose-Response Assessment 1-23
1.4.3 Numerical Risk Estimates 1-29
2. Background 2-1
3. Physical-Chemical Properties 3-1
4. Hazard of Carcinogenic Effects 4-1
4.1. Long- and Short-Term Animal Tests 4-1
4.2. Data Selection for Quantitative Analysis 4-18
4.2.1. Polypoid Adenomas/Other Tumors Observed.. 4-19
-------
4.3. Short-Term Tests: Genotoxicity and Cell
Transformation 4-22
4.4. Other Effects/Defense Mechanisms 4-35
4.4.1. Introduction 4-35
4.4.2. Sensory Irritation 4-35
4.4.3. Cell Proliferation, Cytotoxicity,
and the Mucous Layer 4-37
4.5. Metabolism and Pharmacokinetics .-• 4-53
4.5.1. Absorption 4-53
4.5.2. Pharmacokinetics 4-53
4.5.3. Summary 4-69
4.6. Structure-Activity Relationships 4-70
4.7. Epidemiologic Studies Reviewed 4-77
4.7.1. Introduction 4-77
4.7.2. Review of Studies - Overview
* and Discussion 4-81
4.7.3. Conclusion 4-104
4.8. Weight-of-Evidence 4-107
4.8.1. Assessment of Human Evidence 4-107
4.8.2. Assessment of Animal Studies 4-111
4.8.3. Categorization of Overall Evidence 4-114
5. Hazard of Noncarcinogenic Effects 5-1
5.1. HCHO-Related Effects of the Eyes and
Respiratory System 5-1
5.1.1. Eye 5-2
5.1.2. Olfactory System 5-4
5.1.3. Upper Airway Irritation 5-4
5.1.4. Lower Airway and Pulmonary Effects 5-4
5.1.5. Asthma 5-7
5.1.6. Summary 5-10
5.2. Irritation/Sensitization--Dermal and Systemic ... 5-11
5.3. Cellular Changes : 5-14
5.4. Central Nervous System Effects 5-19
5.4.1. Neurochemical Changes 5-20
5.4.2. Human Studies 5-21
5.4.3. Conclusion 5-24
ii
-------
Tables Title . Page
6-8 Comparison of Non-UFFI Canadian Homes 6-32
by Average Formaldehyde Concentration
6-9 ORNL/CPSC Mean Formaldehyde Concentrations 6-34
(ppm,) as a Function of Age and Season
(Outdoor Means Are Less Than 25 ppb
Detection Limit)
6-10 Frequency Distribution of Formaldehyde Levels 6-36
in Washington Conventional Non-UFFI Homes
6-11 Pre-1980 Monitoring Data for Garment 6-40
Manufacturing and Closely Related Industries
6-12 Recent Monitorign Data for Formaldehyde 6-41
in the Garment Manufacturing Industry
6-13 NIOSH Monitoring Results - Ranges by 6-44
Deparment
^
6-14 Formaldehyde Concentration Levels 6-45
(ppm) - Garment Manufacturing
7-1 Carcinoma/Adenoma Tumor incidence in Fischer 7-2
344 Rats and Male B6C3F1 Mice
7-2 Estimated Individual and Population Risks 7-101
Based Upon Squamous Cell Carcinoma Data
From CUT Study. Population Risks (number
of excess tumors) Appear in Parentheses
Below Individual Risk Estimates
7-3 Risk Estimates Using Polypoid Adenoma Data 7-13
7-4 Upper Bound Risk Estimates Based on the 7-19
CUT Data for Given exposures to HCHO
7-5 Estimated Lifetime Excess Risks Calculated 7-20
from the Epidemiological Studies
7-6 Risk Estimates Based on the Tobe Study 7-35
8-1 Summary of Selected Cross Sectional Studies 8-5
8-2 Summary of Selected Controlled Human Studies 8-9
8-3 Exposure Ranges for Selected Endpoints 8-20
vii
-------
List of Figures
Figures Title Page
1-1 Relationship Between Concentrations of 1-20
Formaldehyde Causing Irritation and Cellular
Effects and Milestone Exposure Levels
4-1 Frequency of squamous metaplasia in the 4-5
nasal cavity of Fischer 344 rats exposed
to 2.0 ppm (top), 5.6 ppm (middle), or
14.3 ppm (bottom) of formaldehyde gas for
24 months. Nasal cavity Levels I, II, IV,
and V were not evaluated at the 6- and 12-
month interim sacrifices in the 14.3 ppm
exposure group.
4-2 Frequency of squamous metaplasia in the 4-7
nasal cavity of BBCSF^ mice exposed to
14.3 ppm of formaldehyde gas.
*
4-3 Drawing indicating the level of sections 4-38
from the nasal passages of rats and mice.
4-4 Simplified reaction sequence from drug 4-62
N-demethylation (cytochrome-P-450-dependent
monooxygenase) to HCHO, formate, and CO2
production (from Waydhas et al., 1978).
Reactions are: la, HCHO dehydrogenase
(GSH); Ib, aldehyde dehydrogenase;
Ic, catalase (peroxidatic mode);
2a, 10-formyltetrahydro-folate sythetase;
2b, 10-formyltetrahydro-folate dehydrogenase;
2c, catalase (peroxidatic mode).
4-5 Tetrahydrofolic acid pathway and 1-carbon 4-64
transfer for HCHO metabolism.
4-6 Overall metabolism of HCHO. 4-70
6-1 Levels of Mobile Homes Corresponding to 6-23
Year of Manufacture.
6-2 Frequency of Formaldehyde Levels, By 6-25
Home Age, Exceeding 1.0, 0.4, and 0.2
ppm In Clayton and Wisconsin Data
Combined.
6-3 Frequency Distribution of Levels in 6-39
Conventional Homes.
Vlll
-------
List of Tables
Table Title _Page
1-1 Health Effects and Representative Exposure 1-2
Levels
1-2 . Summary of Cancer Risks Associated with 1-31
Formaldehyde Exposure
4-1 Summary of Neoplastic Lesions in Nasal 4-2
Cavity of Fischer 344 Rats Exposed to
Formaldehyde Gas
4-2 Summary of Neoplastic Lesions in the Nasal 4-10
Cavity of Sprague-Dawley Rats
4-3 Incidence of Polypoid Adenoma as Reported 4-20
by PWG
4-4 ' Effect of Formaldehyde Exposure on Cell 4-39
Proliferation in Level B of the Nasal
Passages
4-5 Effect of the Time of 3H-Thymidine Pulse 4-40
on Cell Replication After HCHO Exposure to Rat
4-6 Effect of HCHO Concentration vs. Cumulative 4-40
Exposure on Cell Turnover in Rats (Level B)
4-7 Effect of HCHO Concentration vs. Cumulative 4-41
Exposure on Cell Turnover in Rats (Level A)
4-8 Effect of HCHO Concentration vs. Cumulative 4-42
Exposure on Cell Turnover in Mice (Level A)
4-9 Frequency of Squamous Metaplasia in Level 2 4-45
of the Rat Nasal Cavity
4-10 Incidence of Lesions Other Than Tumors in 4-46
the Larnyx of Rats Exposed to Acetaldehyde
[Numeric]
4-11 Incidence of Epidermoid and Adenoid 4-46
Squamous Carcinomas in Rats Exposed to
Hexamethylphosphoramide
4-12 Nasal and Laryngeal Cancer in Rats Treated 4-76
with Acetaldehyde by Inhalation for 27 Months
4-13 Summary of Studies Relevant to Formaldehyde 4-78
-------
Table Title P&fi
4-14 Formaldehyde Levels to Which Occupational 4-83
Groups Might Be Exposed
4-15 Power Calculations for SMR Studies 4-85
4-16 Conditional Power Calculations for PMR 4-89
Studies
4-17 Power Calculations for Case-Control Studies 4-91
4-18 Fisher's Combined p for SMR and PMR Studies 4-101
4-19 Predicted Human Relative Risks 4-103
for Selected Occupations
5-1 Reported Health Effects of Formaldehyde 5-1
at Various Concentrations
5-2 Delayed Type Hypersensitivity (Human) 5-13
•* Due to Low Levels of Formaldehyde
5-3 Significant Findings in Nasal Turbinates 5-16
in Rats
5-4 Significant Findings in Nasal Turbinates 5-16
in Monkeys
5-5 Total Incidence by Groups of Monkeys 5-16
6-1 Populations at Risk 6-4
6-2 Use of Pressed-Wood Products in 6-7
Home Construction
6-3 Summary of Residential Formaldehyde 6-16
Monitoring
6-4 Summary of Residential Monitoring Data 6-21
from Randomly-Sampled Homes
6-5 Potential Effects of Temperature and - 6-27
Relative Humidity Changes on Formaldehyde
Air Concentrations
6-6 Frequency of Observations Found in 6-28
Concentration Intervals by Clayton
Environmental Consultants
6-7 Frequency of Observations Found in 6-29
Concentration Intervals by Wisconsin
Division of Health
VI
-------
7.
J . J .
5.6.
5.5.1. Animal Studies
5.5.2. Human Data
5.5.3. Conclusion
Effects on Visceral Organs
Exposure Assessment
6. 1.
6.2.
6.3.
6.4.
6.5.
6.6.
Introduction
Estimates of Current Human Exposure
Populations at Risk
6.3.1. Home Residents
6.3.2. Garment Workers
Sources of HCHO in Population Categories
of Concern
6.4.1. Homes Containing Pressed-Wood Products...
6.4.2. Garment Manufacture
HCHO Levels in Homes and Garment
Manufacturing Sites
6.5.1. HCHO Levels in Homes
6.5.2. Manufactured Homes
6.5.3. Conventional Homes
Summary
5-24
5-26
5-28
5-30
6-1
6-1
6-2
6-3
6-3
6-4
6-4
6-5
6-5
6-14
6-15
6-15
6-20
6-30
6-38
6-46
7.1.
7. 2.
7. 3.
7 4
Risk Estimates Based on
Souamous Cell Carcinoma Data •
Risk Estimates Based on Polypoid
Adenoma Data
Pyecont' a i- i nn of Ri
-------
7.4.4. Other Considerations-Squamous
Papillomas 7-33
7.4.5. Conclusion 7-35
7.5 Summary 7-36
8. Estimates of Noncancer Risks 8-1
8.1. 'Introduction 8-1
8.2. Studies Reviewed 8-2
8.3. Limitations of Studies 8-2
8.3.1. Study Design Limitations 8-3
8.3.2. Bias Limitations 8-4
8.4. Results 8-4
8.5. Discussion 8-17
9. References 9-1
Appendix 1 Expert Panel Report on HCHO Pharmacokinetic
Data and CUT Response
Appendix 2
Appendix 3
Appendix 4
Appendix 5
Individual Summaries of Epidemiologic
Studies Reviewed
Estimates of Risk Using Various Extrapolation
Models
Documentation of High to Low Dose Extrapolation
Models Used in Quantitative Risk Assessment-
Concise Description
Sensitivity Analysis of CUT Rat Data Using
the Linearized Multistage Model
IV
-------
Figures Title Page
7-1 Comparison of the Upper Bound Risks Based' 7-21
on the Animal Data to Estimated Lifetime
Excess Risks Based on the Epidemiological
Studies.
8-1 Predicted Irritative Response Over A Range 8-22
of HCHO Levels. Data From Hanrahan et.al.
(1984)
8-2 Predicted Irritative Response Over A Range 8-23
of HCHO Levels. Data from Anderson and
Molhave (1984)
8-3 Predicted Irritative Response Over A Range 8-24
of HCHO Levels. Data from Kulle (1985)
8-4 Eye Irritation Response Over a Range of 8-25
HCHO Levels
IX
-------
PREFACE
The Health Risk Assessment of Formaldehyde (HCHO) was
prepared to serve as source document for Agency-wide use. This
document was developed primarily for use by the U.S.
Environmental Protection Agency's (EPA) Office of Toxic
Substances to support decision-making regarding possible
regulation of HCHO under Section 6 of the Toxic Substances
Control Act. Because this document focuses on inhalation
exposure to HCHO, this document should not be regarded as a
comprehensive assessment of the health effects from oral and
dermal exposure. In addition, only two exposure categories are
extensively reviewed. Assessment of other categories will be
done as needed by other EPA program offices.
In the development of this assessment document, the relevant
scientific literature available through February 1, 1986,. has
been incorporated, except that the epidemiologic section reflects
studies available through March 1987. Key studies have been
evaluated and the summary and conclusions have been prepared so
that the health effects and related characteristics of HCHO are
qualitatively identified. Measures of dose-risk relationships
relevant to inhalation exposure are also discussed so that the
adverse health responses can be placed in perspective with
possible exposure levels.
-------
Authors, Contributors, and Reviewers
The Existing Chemical Assessment Division within the Office
of Toxic Substances was responsible for preparing this document.
Principal Authors
Richard Hefter, M.S.
Cheryl Siegel Scott, M.S.P.H.
Harry Milman, Ph.D.
Elizabeth Margosches, Ph.D.
Gary Grindstaff, M.S.P.H.
Gregory Schweer, M.S.
Jeanette Wiltse, Ph.D.
Contributing Authors
Mary Argus, Ph.D.
Angela Auletta, Ph.D.
Frederick Dicarlo, Ph.D.
William Parian, Ph.D.
David Klauder, Ph.D.
Carl Mazza, Ph.D.
George Semeniuk, Ph.D.
Reviewers
The following individuals provided peer review of this
document and or earlier drafts of this document.
Irwin Baumel, Ph.D.
Office of Regulatory Support
Office of Research and Development
Washington, D.C.
Steven P. Bayard, Ph.D.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, D.C.
David L. Bayliss, Ph.D.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, D.C.
Robert Beliles, Ph.D.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, D.C.
XI
-------
VicXi L. Dellarco, Ph.D.
Reproductive Effects Assessment Group
Office of Research and Development
Washington, D.C.
Karen East, M.S.
Chemical Review and Analysis Branch
Office of Policy, Planning, and Evaluation
Washington, D.C.
A. M. Jarabek, Ph.D.
Environmental Criteria and Assessment Office
Cincinnati, Ohio
Nancy Pate, Ph.D.
Pollutant Assessment Branch
Office of Air Quality Planning and Standards
Research Triangle Park, NC
Peter Preuss, Ph.D.
Office of Health and Environmental Assessment
Washi ngton, D.C.
Van M. Seabaugh, Ph.D.
Hazard Evaluation Division
Office of Pesticide Programs
Washington, D.C.
Paul White, B.A.
Exposure Assessment Group
Office of Health and Environmental Assessment
Washington, D.C.
EPA Science Advisory Board
The May 1985 draft of this document was independently peer-
reviewed in public sessions of the Environmental Health Committee
of EPA's Science Advisory Board.
xii
-------
Executive Summary
EPA has conducted an extensive analysis of the experimental
and exposure information on formaldehyde in order to characterize
the potential risk to humans from exposure to the chemical.
The major non-cancer effects posed by formaldehyde are due to
the irritation (i.e., irritation of the eyes, nose, throat and
lungs) and cellular changes (i.e., effects on the mucociliary
system of the nose). A large number of observations of people in
various settings support a conclusion that the generally observed
range over which more than 95% of people respond is 0.1-3.0 ppm of
formaldehyde. Generally, little risk from non-cancer health
effects from exposure to formaldehyde is attributed to cases where
exposures are one hundred-fold less than a no- or lowest-observed
effect level. Although quantitative estimates of non-cancer risk
are not possible, fewer responses are expected to be associated
with fewer and less intense exposures.
EPA has classified formaldehyde as a "Probable Human
Carcinogen" (Group Bl) under its Guidelines for Carcinogen Risk
Assessment. Based on a review of epidemiologic studies, EPA has
concluded that there is "limited" evidence to indicate that
formaldehyde may be a carcinogen in humans. Nine studies reported
statistically significant associations between site-specific
respiratory neoplasms and exposure to formaldehyde or formaldehyde-
containing products.
XI 11
-------
An examination of studies in animals has indicated that there
is "sufficient" evidence of carcinogenicity of formaldehyde in
animals by the inhalation route. This is based on the induction by
formaldehyde of an increased incidence of a rare type of malignant
tumor (i.e., nasal squamous-cell carcinoma) in both sexes of rats,
in multiple inhalation experiments, and in multiple species (i.e.,
rats and mice). In these long-term laboratory studies, tumors were
not observed beyond the initial site of nasal contact.
Supportive evidence for the carcinogenicity of formaldehyde
was obtained from short-term tests designed to measure effects on
DNA. Formaldehyde is mutagenic in numerous bacterial test systems
and test systems using fungi and insects (Drosophila). It also
tranforms cells in culture and causes DNA cross-linking, sister
chromatid exchange (SCE) and chromosome aberrations. In addition/
formaldehyde has been shown to form adducts with DNA and with
proteins in both in vivo and in vitro test systems. Its ability to
interfere with DNA repair in human cells has also been shown.
Structure-activity correlations support the prediction of
potential carcinogenicity. Formaldehyde is one of several
aldehydes which have been shown to have carcinogenic activity in
experimental animals. Acetaldehyde, the closest structural
analogue of formaldehyde, induces the same type of malignant tumor
in the respiratory and olfactory epithelium of the nose of rats as
does formaldehyde.
Results from studies in rats by the Chemical Industry
Institute of Toxicology were used to estimate the human cancer
xiv
-------
risk. The malignant tumor data were used to extrapolate human
cancer risk because only this response in formaldehyde-exposed rats
was definite and unequivocal in both sexes of rats, was dose-
related, and was confirmed in several rat inhalation studies. In
the absence of compelling biological evidence on the mechanism of
action, as in the case for formaldehyde, EPA's Guidelines for
Carcinogen Risk Assessment specify the selection of the linearized
multistage procedure for estimating human cancer risk. Using this
procedure, the upper bound estimate for excess lifetime risk of
developing cancer is 3xlO~4 (Group Bl) for apparel workers exposed
to formaldehyde at the 0.17 ppm level; 2x10"^ (Group Bl) for
residents of mobile homes who are exposed for 10 years to an
average level of 0.10 ppm; and lxlO~4 (Group Bl) for residents of
some conventional homes who are exposed for 10 years to an average
level of 0.07 ppm. The upper bound estimate for an ambient
exposure of 1 ug/m^ (0.00082 ppm) for 70 years (the unit risk) is
1.3xlO~5 (Group Bl).
Since some of the existing information supports the use of
non-linear risk assessment models to extrapolate cancer risk to
humans, and since considerable uncertainty exists in the risk
estimates, the real risk may be lower than that projected by the
upper bound, linear estimate. The lower bound is always recognized
to be as low as zero. However, the predicted excess lifetime
cancer risk estimates using an upper bound based on the rat nasal
carcinoma data are about equivalent to the excess cancer incidence
observed in the epidemiologic studies.
xv
-------
1. Risk Characterization
This risk-characterization presents the major conclusions of
EPA's risk assessment of formaldehyde. It reviews the underlying
scientific foundation for the findings, describes the strengths
and weaknesses of the supporting data, and discusses the
uncertainties and potential sources of controversy attending
EPA's interpretation of the data and projection of risk. The
risk characterization is divided into three sections which
discuss the qualitative aspects of the risk assessment, the
exposure, and the quantitative risk estimations at current
exposure levels. A summary of the health effects of formaldehyde
and representative exposure levels is presented in Table 1-1.
1.1. Non-cancer Effects
The major non-cancer effects posed by exposure to
formaldehyde are due to the irritating nature of the chemical.
These effects are sensory irritation which is readily perceived
by the exposed individual and cellular changes which are less
evident but still important.
1.1.1. Sensory Irritation
The well documented health effects from acute inhalational
exposures are concentration dependent, with individuals
responding above a threshold concentration. These effects
include irritation of the eyes, nose, throat and lungs, the
intensity of which is dependent upon the extent and duration of
exposure, and may result in extreme discomfort and inability to
function normally at work or in routine daily activities.
1-1
-------
TABLE 1-1
HEALTH EFFECTS AND REPRESENTATIVE EXPOSURE LEVELS
Formaldehyde
Concentration
(ppm).
Health Effects
(Exposure time)
a
Representative
Exposure
Levels
< 0.05
0.1
0.5
1.0
2.0
3.0
5.0
15.0
Human eye irritation
begins in sane people
(minutes-hours)
Human mucociliary inhibition
and squamous metaplasia,
mid-point of range in one study
(0.1-1.1 ppm) (years)
Human nose and throat irritation
begins; most people have eye
irritation (minutes-hours)
Rat squamous metaplasia, and
mucociliary system LOEL (months)
Human (most) experience nose and
throat irritation (minutes)
Monkey squamous metaplasia LOEL
(weeks)
Rat observed 1% cancer incidence
(years)
Human lower airway effects begin
(minutes-hours)
Rat observed 50% cancer incidence
(years)
Mouse observed 1% cancer incidence
(years)
Ambient background
New mobile homes
10-yr average
Current OSHA PEL
(8 hr TWA)
Highest recorded
honec
a
b
c
Duration of exposure causing the effect is indicated in parentheses.
LOEL = lowest observed effect level
Urea-formaldehyde foam insulated home
1-2
-------
Due to varying sensitivities, all individuals do not exhibit
these acute effects at the same formaldehyde concentration.
Thus, the number of persons who respond in a population will
increase with increasing concentrations of formaldehyde. A dose-
response relationship has not been quantitatively characterized
for the general population. However, a large number of
observations of people in various clinical and nonclinical
settings support a conclusion that the generally observed range
over which most people respond (more than 95% response) is 0.1-
3.0 ppm of formaldehyde.
Eye irritation occurs first at the lower end of the range;
the percentage of individuals that respond increases up to a
concentration of formaldehyde of 1.0 ppm, the concentration at
which virtually all persons exhibit some degree of eye
irritation. Irritation of the nose and throat frequently occurs
above 1.0 ppm with most persons responding by 3.0 ppm. Exposures
greater than 3.0 ppm are generally intolerable for more than
short periods. These acute effects are usually reversible.
Tolerance to low levels of formaldehyde can occur in individuals
after 1-2 hours of exposure, but symptoms can return if exposure
is interrupted and then resumed.
In addition to its direct irritant effects on the
respiratory system, formaldehyde has been shown to cause
bronchial asthma-like symptoms in humans. Although asthmatic
attacks may, in some cases, be due specifically to formaldehyde
sensitization or allergy, the evidence for this is
inconclusive. Even so, a small number of reports indicate that
1-3
-------
formaldehyde may be an inhalant sensitizer causing allergic
reactions. There are no sufficiently well controlled studies to
establish the extent of such sensitization in the population, nor
are induction concentrations of formaldehyde known. However, the
concentrations of formaldehyde required to elicit such attacks
are higher than would be expected in most non-occupational
envi ronments.
1.1.2. Cellular Changes
The primary point of contact of formaldehyde upon exposure
by inhalation is the nose. Inhalation of formaldehyde above a
threshold level which varies from person to person causes a
number of cellular effects which can impair the normal
functioning of the nose and are dependent on the concentration
and duration of exposure.
A major function of the nose is to prepare the inhaled air
for the lungs. This includes warming, moistening, and filtering
the inspired air. Dust and many bacteria found in the inspired
air are precipitated in the mucus that bathes the mucous membrane
and are moved outward by the action of the cilia of the nasal
passage. Research indicates that formaldehyde has a number of
effects on the workings of this mucociliary apparatus.
Effects on the mucociliary system of laboratory animals have
been observed in several short-term exposure studies. In one
study, male rats were exposed for 6 hours per day for up to 14
days, to 0.5, 2, 6, or 15 ppm of formaldehyde. At 15 ppm, the
stopping of mucous flow (mucostasis) followed by cessation of
ciliary activity (ciliastasis) was clearly shown. Only slight
1-4
-------
effects were noted in animals being exposed to 6 ppm or 2 ppm.
At 0.5 ppra no effects were observed. In other short-term studies
formaldehyde caused cell proliferation in the nasal epithelium at
doses of 2 ppm and higher. Cell proliferation is a part of the
restorative process to repair cellular damage.
In chronic studies, cellular effects, i.e., rhinitis
(inflammation of the nasal mucosa), epithelial dysplasia
(displacement of one cell type with another one), and squamous
metaplasia (replacement of normal mucosal cells with squamous
cells), developed in the nasal cavities of rats and monkeys after
exposures for 12 months and 26 weeks, respectively, to 2-3 ppm of
formaldehyde. After 24 months of exposure, the incidence of
squamous metaplasia in rats increased to nearly 100 percent. In
both rats and monkeys, a NOEL (no observed effect level) of 1.0
ppm for squamous metaplasia was determined, with a LOEL (lowest
observed effect level) of 2.0 ppm in rats and 3.0 ppm in
monkeys. The potential relationship between squamous metaplasia
and carcinogenesis is presented in section 1.4.2.2.
Evidence of cellular damage in humans is Limited. One study
in which humans were occupationally exposed from four to nine
years (mean = seven years) to formaldehyde in the range of 0.1-
1.1 ppm, time-weighted average (TWA) concentration, showed loss
of ciliary activity and development of squamous metaplasia.
Caution must be used when generalizing from this study because of
the small number of exposed persons examined (20) and the
possibility of confounding exposure. Five individuals in the
formaldehyde-exposed group exhibited nasal cavity changes.
1-5
-------
The mucociliary system is an important defense mechanism in
the removal of foreign particles and bacteria which enter the
upper respiratory system. A reduction in the efficient operation
of this defense mechanism, including formation of squamous
metaplasia, by exposure to formaldehyde may increase the risk of
persons exposed to formaldehyde to develop infections and other
respiratory diseases.
1.2. Carcinogenic Effects
EPA has classified formaldehyde as a "Probable Human
Carcinogen" (Group Bl) under its Guidelines for Carcinogen Risk
Assessment. This classification is based on the following:
*
o limited evidence of carcinogenicity in humans (i.e.,
several epidemiologic studies show positive associations
between respiratory site-specific cancers and exposure to
formaldehyde);
o sufficient evidence of carcinogenicity in animals (i.e.,
formaldehyde induced an increased incidence of rare,
malignant nasal squamous-cell carcinoma in mice and rats,
and in multiple experiments); and
o additional supportive evidence (i.e., studies
demonstrating formaldehyde's mutagenic activity in
numerous test systems using bacteria, fungi, and, insects,
and its ability to transform cells in culture and cause
DNA damage in other in vitro assays for mutagenicity.
Also, structure-activity analysis indicates that
formaldehyde is one of several carcinogenic aldehydes.)
1-6
-------
1.2.1. Studies of Humans
The EPA has examined 28 epidemiologic studies relevant to
formaldehyde. Three of these studies, two cohort* (Blair et al.,
1986; 1987 in press; Stayner et al., 1986) and one case-control2
(Vaughan et al., in press), were well conducted and specifically
designed to detect small to moderate increases in formaldehyde-
associated human risks. Each of these three studies observed
statistically significant associations between respiratory site-
specific cancers and exposure to formaldehyde or formaldehyde-
containing products. These associations are noteworthy since
during inhalation, tissues in the nose, nasal sinuses, buccal
cavity (mouth), pharynx,3 and lungs come into direct contact with
formaldehyde. In each of the above three studies, the
populations studied were also undoubtedly exposed to other
chemicals and these exposures may have contributed to the
observed increases in cancer risk. Only the study by Vaughan
et al. (1986a,b) controlled for smoking and alcohol consumption.
A cohort study follows a group of exposed individuals for a
specified time period and measures the incidence of site-specific
deaths. The observed number of site-specific deaths which
occurred in the time period are compared to the number of site-
specific deaths which would be expected based on mortality rates
of a standard population.
A case-control study identifies cases with the disease of
interest and controls who do not have the disease. The cases and
controls are compared with respect to past exposure.
^ The pharynx is the passage between the nasal cavity and the
larynx. The nasopharynx, hypopharynx, oropharynx, and
laryngopharynx comprise the pharyngeal region.
1-7
-------
The Blair et al. (1986; 1987 in press) cohort study observed
significant excesses in lung and nasopharyngeal cancers among U.S.
workers occupationally exposed to formaldehyde at 10 industrial
sites. Blair et al. (1986), however, argued that the lung cancer
excesses provided little evidence of an association with
formaldehyde exposure since the lung cancer risk did not increase
consistently with either increasing intensity or cumulative
formaldehyde exposure. EPA, after reviewing the data, has
concluded that the significant excesses in total lung cancer
mortality, in analyses either with or without a latency period
equal to or greater than 20 years, and together with nasopharyngeal
cancer mortality among formaldehyde-exposed workers are meaningful
despite the lack of significant trends with exposure.
Misclassification of exposure (or lack of specificity between
exposure categories) and categorization of deaths into four
exposure levels which lowers the power to detect small increases in
risk, may have accounted for the observed lack of a significant
dose-response relationship. The significance of these findings is
reinforced by the fact that the site of the tumors seen in humans
(the nasopharyngeal region) is similar to that seen in animals.
Blair et al. (1987) performed further analyses of the
nasopharyngeal cancers regarding exposure to formaldehyde and
particulates. For those workers with particulate exposure, the
trend between increasing nasopharyngeal risk and increasing
cumulative formaldehyde exposure was not statistically significant,
however, the authors concluded that formaldehyde and particulates
appeared to be. a risk factor for nasopharyngeal cancer.
1-8
-------
The Stayner et al. (1986) cohort study reported statistically
significant excesses in mortality from buccal cavity tumors among
formaldehyde-exposed garment workers. The standardized mortality
ratio (SMR), a ratio of the observed number of deaths to an age-
adjusted number of deaths expected in the group, was highest among
workers with a long duration of employment (exposure) and follow-
up period (latency). A significant excess in deaths from cancer
of the tonsils was also reported, but there were too few deaths to
examine any trends with exposure.
Results from the case-control study by Vaughan et al.
(1986a,b) showed a significant association between nasopharyngeal
cancer and having lived 10 or more years in a "mobile home".
Persons for whom this association was drawn had lived in mobile
homes that were built in the 1950s to 1970s. This study also
reported significant associations between sinonasal cancer and
orohypopharyngeal cancer and exposure to resins, glues, and
adhesives (SAIC, 1986). Mo significant trends were found in
cancer incidence at any of these sites with respect to
occupational formaldehyde exposure; however, the risk estimates
for the highest exposure level and cancers of the orohypo- and
naso-pharynx appeared elevated. As stated earlier, however, this
population, like the two previously discussed, was also
undoubtedly exposed to other chemicals which may have contributed
to the observed increases in cancer risk.
Several residential and occupational characteristics were _a
priori selected as likely surrogates for formaldehyde exposure.
Among these were mobile home residency and occupational resins,
glue, and adhesive exposure.
1-9
-------
EPA previously had reviewed 25 other epidemiologic studies.
These studies had limited ability (lower power) to detect small to
moderate increases in formaldehyde-related risks due to (1) small
sample sizes; (2) small numbers of observed site-specific deaths;
and (3) insufficient follow-up. Even with these potential
limitations, six of the 25 studies (Acheson et al., 1984a; Hardell
et al., 1982; Hayes et al., 1985; Liebling et al., 1984; Olsen et
al., 1984; Stayner et al., 1985) reported significant associations
between excess site-specific respiratory (lung, buccal cavity, and
pharyngeal) cancers and exposure to formaldehyde.
The Olsen et al. (1984), Hayes et al. (1986), and Hardell et
al. (1982) studies reported significant excesses of sinonasal
cancer in individuals who were exposed to both formaldehyde and
wood-dust, or who were employed in particleboard manufacturing
where formaldehyde is a component of the resins used to make
particleboard. Only the Hayes et al. (1986) and Olsen et al.
(1984) studies controlled for wood-dust exposure; the detection
limits in both studies, however, exceeded corresponding expected
excesses in the incidence of sinonasal tumors and, therefore, no
significant excesses were likely to have been observed.
The Acheson et al. (1984a) study conducted in the United
Kingdom supports the results of Blair et al. in that, when
compared to mortality rates of the general population, significant
excesses in mortality from lung cancer were observed in one of six
formaldehyde resin producing plants in England. A trend of
borderline significance with dose was observed for this one
plant. Acheson et al. concluded that the increases in mortality
1-10
-------
from lung cancer were not related to formaldehyde exposure since
the elevation and trend were not statistically significant when
compared with local lung cancer rates. EPA believes that the
risks and trends from analyses using local lung cancer rates as
the comparison risks appeared sufficiently increased for
corroborative use.
The remaining two studies reported significant excesses of
buccal cavity cancer among garment workers in 3 plants (Stayner et
al., 1985) and excesses of buccal cavity and pharyngeal cancer
among formaldehyde resin workers in 1 plant (Liebling et al.,
1984). Portions of the Liebling et al. (1984) and Blair et al.
(1986, 1987) studies overlapped as did portions of the two Stayner
et al. (1985; 1986) studies. However, the non-overlapping
portions and improved design of the more recent studies (i.e.,
Blair et al. 1986, 1987; Stayner et al. 1986) reinforce the
conclusions of the earlier studies.
Analyses of the remaining 19 epidemiologic studies have
indicated the possibility that observed leukemia and neoplasms of
the brain and colon may be associated with formaldehyde
exposure. The biological support for such postulates, however,
has not yet been demonstrated.
1-11
-------
Based on a review of these studies, EPA has concluded that
there is "limited" evidence to indicate that formaldehyde may be a
carcinogen in humans.^ Nine studies reported statistically
significant associations between site-specific respiratory
neoplasms•and exposure to formaldehyde or formaldehyde-containing
products. This is of interest since inhalation is the primary
route of exposure in humans. Although the common exposure in all
of these studies was formaldehyde, the epidemiologic evidence is
categorized as "limited" primarily due to possible exposures to
other agents which may have confounded the findings of excess
cancers.
•
1.2.2. Studies in Animals
The principal evidence indicating that formaldehyde causes
cancer in animals comes from studies conducted by the Chemical
Industry Institute of Toxicology (CUT) (Kerns et al. , 1983) and
those by Albert et al. (1982) and Tobe et al. (1985). The CUT
study was a well conducted, multidose inhalation study in rats and
mice. In this study, a statistically significant increase in
malignant tumors (i.e., squamous cell carcinomas) was seen in the
nasal cavities of male and female rats dosed at 15 ppm. In
addition, a small increased incidence of squamous cell carcinoma,
while not statistically significant, was seen in male mice.
EPA's Guidelines for Carcinogen Risk Assessment define limited
evidence of carcinogenicity in humans as indicating that "...a
causal interpretation is credible, but that alternative
explanations, such as chance, bias, or confounding, could not
adequately be excluded."
1-12
-------
Because this.type of nasal Lesion is rare in mice, these data can
be considered to have biological importance. Benign tumors (i.e.,
polypoid adenomas) were seen in male rats in the CUT study at all
dose levels and in female rats exposed to 2 ppm of formaldehyde.
Notably, the dose-response curve for the benign tumors in this
study was not linear; the tumor, incidence was highest at 2.0 ppm
and decreased at higher doses.
Tobe et al. also observed a statistically significant
increase in the numbers of squamous cell carcinomas in the same
strain of male rats as was used in the CUT study. Albert et al.
reported a statistically significant elevation of the same
*
malignant tumor type in male rats of a different strain. In both
the Tobe et al. and Albert et al. studies benign squamous cell
papillomas were seen. This observation was in contrast to the
CUT study in which polypoid adenomas were the only benign tumors
observed. Hamsters have been tested in long-term inhalation
studies (Dalbey, 1982) but no increased incidence of tumors was
seen in formaldehyde-treated animals. However, deficiencies in
the study design and poor survival limit the interpretation of the
results from these studies.
Additional studies in animals that indicate an association
between exposures to formaldehyde and cancer are those by Dalbey
(1982) in which formaldehyde enhanced the production of tumors
induced by a known animal carcinogen (i.e., diethylnitrosamine);
Mueller et al. (1978) in which formalin (a water solution of
formaldehyde) produced lesions in the oral mucosa of rabbits which
showed histological features of carcinoma in situ; and studies by
1-13
-------
Watanabe et al. (1954; 1955) in which injections of formalin and
hexamethylenetetramine (from which formaldehyde is liberated in
vivo) produced sarcomas (malignant tumors) and one adenoma (benign
tumor) at the site of injection.
Based upon a review of these studies, EPA has concluded that
there is "sufficient" evidence of carcinogenicity of formaldehyde
in animals by the inhalation route. This finding is based on the
induction by formaldehyde of an increased incidence of a rare type
of malignant tumor (i.e., nasal squamous-cell carcinoma) in both
sexes of rats, in multiple inhalation experiments, and in multiple
species (i.e., rats and mice). In these long-term laboratory
studies, tumors were not observed beyond the initial site of nasal
contact nor have other mammalian in vivo tests shown effects at
distant sites.
1.2.3. Additional Supportive Evidence
Other relevant information which is considered in carcinogen
assessments include results from short-term tests designed to
measure effects of a chemical on DNA. Tests for point mutations,
numerical and structural chromosome aberrations, DNA
damage/repair, and in vitro cell transformation provide evidence
for the potential mechanisms of carcinogenicity. A battery of
EPA's Guidelines for Carcinogen Risk Assessment define
sufficient evidence of carcinogenicity from studies in
experimental animals as indicating that "...there is an increased
incidence of malignant and benign tumors: (a) In multiple
species or strains; or (b) in multiple experiments (preferably
with different routes of administration or using different dose
levels); or (c) to an unusual degree with regard to incidence,
site or type of tumor, dose-response effects, as well as
information from short-term tests or on chemical structure."
1-14
-------
tests which measure different endpoints helps to characterize the
chemical's response spectrum. In general, the wider the range and
the greater the intensity of response of a substance in short-term
tests, the more likely it is that the substance may cause cancer.
Formaldehyde is mutagenic in numerous bacterial test systems
and test systems using fungi and insects (Drosophila). It also
transforms cells in culture and causes DNA cross-linking, sister
chromatid exchanges (SCE) and chromosome aberrations. In
addition, formaldehyde has been shown to bind with DNA and with
proteins in both in vivo and in vitro test systems. Its ability
to interfere with DNA repair in human cells has also been shown.
»
Structure-activity correlations support the prediction of
potential carcinogenicity. Formaldehyde is one of several
aldehydes which have been shown to have carcinogenic activity in
experimental animals. Of those tested, acetaldehyde, is the
closest structural analogue of formaldehyde. Like formaldehyde,
acetaldehyde damages the respiratory and olfactory epithelium,
however, formaldehyde appears to be more potent than
acetaldehyde. The main impact of formaldehyde, probably because
of its greater reactivity, occurs more in the anterior portion of
the nose than that of acetaldehyde. Exposure to either aldehyde
leads to the formation of nasal squamous cell carcinoma;
acetaldehyde, however, also induces another type of malignant
nasal tumor, adenocarcinoma. Polypoid adenoma (benign tumor) were
seen following exposure to formaldehyde whereas squamous-cell
papilloma (benign tumor) were found following treatment with
acetaldehyde. The utility of benign tumors in risk assessment is
discussed in sections 1.4., 4.2.1., and 7.4.
1-15
-------
1.3. Exposure in Residential and Apparel Manufacturing Settings
EPA's risk assessment focuses on two large populations
chronically exposed to low levels of formaldehyde. These
populations include: (1) persons who reside in mobile and
conventional homes constructed using "significant amounts" of
urea-formaldehyde (UF) pressed-wood (i.e, homes in which UF
pressed wood is used for floor underlayment and, in some cases,
for wall paneling), and (2) apparel workers who are exposed to
formaldehyde that is emitted from durable press fabrics.
Available air monitoring data, although not collected under
any comprehensive nation-wide survey, indicate that exposure
levels in both settings have declined over the last 5 years. This
is consistent with the increased commercial use of lower-emitting
formaldehyde source material (pressed wood products and durable
press resins).
Measurements of formaldehyde levels are strongly affected by
a number of factors that add to the overall uncertainty of the
data. These factors include the monitoring methods employed, the
amount and age of the formaldehyde source material present at the
site, the extent of ventilation at the site, and the ambient
temperature and humidity. High temperature and humidity, for
example, are known to increase emissions of formaldehyde from
pressed wood products and durable press fabrics. However, many of
the formaldehyde monitoring efforts did not report or document
adequately these variables.
1.3.1. Residential Exposure
Most of the monitoring data collected in residences over the
1-16
-------
last decade have been in older homes, homes in which urea-
formaldehyde foam insulation (UFFI) has been installed, or homes
in which the occupants have expressed health complaints. Perhaps
more representative of current formaldehyde levels in new mobile
arid conventional homes are several recent monitoring studies
conducted in California, Tennessee and Texas. These studies
evaluated homes built after 1980, when builders were using energy-
efficient (tighter) construction and most pressed wood producers
had begun to use low-emitting resin adhesives.
Recent monitoring results indicate that formaldehyde levels
in new (less than one year old) conventional homes generally fall
within the range of 0.05 ppm to 0.2 ppm; few measurements exceeded
0.3 ppm. In new mobile homes, formaldehyde levels monitored
generally fall within the range of 0.2 ppm to 0.3 ppm with the
highest levels measured near 0.4 ppm, the ceiling level targeted
by Department of Housing and Urban Development regulations that
govern mobile home construction. The larger range of values
observed in conventional homes is attributed primarily to the
greater variation in design and use of UF pressed wood products in
their construction. By contrast, mobile homes have less design
variation and, for the most part, generally use pressed wood
products more extensively. EPA has developed computer models to
estimate initial formaldehyde levels in conventional homes built
using significant amounts of pressed wood. Although these models
have not been fully validated, they yield expected values that
fall within the range of 0.1-0.2 ppm.
EPA estimates that every year approximately 631,000 persons
1-17
-------
move into new conventional homes that contain significant amounts
of pressed wood. In the same period, about 780,000 persons move
into new mobile homes.
Under normal conditions, the amount of formaldehyde released
from pressed wood products decreases with time, lowering the
levels in these residences. Although numerous studies have
investigated the decrease in emissions from uncoated pressed wood
in the months immediately after its manufacture, little
quantitative information is available on long-term (10-year)
formaldehyde emissions from pressed wood. In lieu of long-term
emission decay data, EPA derived a decay curve function by
*
combining the results of two large monitoring surveys and
statistically determining the best-fit curve to the data as a
function of home age. Combined, the two surveys reported almost
1,200 measurements in 400 mobile homes that were constructed
during 1970-1980 and ranged .in age from one day to nearly 10
years. EPA has used the exponential function derived from these
data for quantitative cancer risk assessment purposes to calculate
expected 10-year averages for formaldehyde levels in homes built
today. The calculated 10-year averages are 0.07 ppm for
conventional homes built using significant amounts of UF pressed
wood and 0.1 ppm for mobile homes. As these estimates are derived
from historical data, a significant source of uncertainty
associated with these estimates is the unknown long-term emission
characteristics of the UF resins used today to manufacture the
pressed wood products used in these homes.
1-18
-------
1.3.2. Exposure in Apparel Manufacturing
The U.S. apparel industry employed over 1.1 million workers
in 1983. There are approximately 22,600 apparel manfacturing
establishments located in every state of the country, and each
employs an.average of 59 workers.
Monitoring data collected by the National Institute for
Occupational Safety and Health (NIOSH), the Occupational Safety
and Health Administration (OSHA), or otherwise reported in the
literature indicate that formaldehyde levels in these facilities
were generally below 3.0 ppm prior to 1980. In later years, the
levels have generally fallen below the 1.0 ppm level. Recent
industrial hygiene studies by NIOSH of two large manufacturing
sites that produce mens' shirts indicated that the mean exposure
level for both plants was 0.17 ppm.
1.4. Quantitative Risk Assessment
The risk assessment identified two biological effects for
which the data are sufficient to evaluate quantitatively. These
are acute sensory/cellular effects of the upper respiratory tract
and cancer. A combination of results obtained from studies in
animals and humans were used to assess the acute sensory/cellular
effects while the cancer risk estimates were derived from modeling
data obtained from studies in animals.
1.4.1. Non-cancer Risk Assessment
Figure 1-1 illustrates the relationship between the doses
associated with sensory irritation and cellular effects in the
nasal cavity and the exposure levels for a number of population
groups. Instead of using high-to-low dose extrapolation models,
1-19
-------
I
to
o
a)
b)
10-yr awj.
scneoorv.
homes
NIGEH
gatnent
vorters
10-yr avg.
ncbile tares
avg. new
ncbile hones
with <%e, nose, and ttroat aerBoty irritation
taige vhere most persons
NJtL
ciliaty ettects
(rats)
t&L IfJtL
squamae aqiaicus
metaplasia metaplasia
(cats and nonkeys) (cats)
nutaolasia
"oL'UJ'oU'oL'UJUJ'o^UJUJUo ,,a
3.0
FK1RE 1-1, I^lationEhip between ccnoentcations ct
caLeinj asreoty irritation and cellular effects and mikstor^ exposure luvels.
-------
the degree of concern from these effects is approximated by
comparing existing exposures to lowest effect levels. Generally,
little risk is attributed to cases where exposures are one
hundred-fold less than a no- or lowest-observed effect level.
1.4.1.1. Sensory Irritation
The onset of sensory irritation in humans exposed to
formaldehyde occurs over a wide range of formaldehyde
concentrations (i.e., 0.1-3.0 ppm). This range overlaps the
expected human exposures identified in this assessment. This
means that there is no margin between existing exposures and
levels of formaldehyde that are associated with sensory irritation
in some humans. Thus, it would seem that some humans may be
currently experiencing some degree of sensory irritation to
f
existing levels of formaldehyde in new to moderately new mobile
homes and garment manufacturing operations (see Figure l-l(a)).
Due to the large variations in human sensitivity to the irritative
effects of formaldehyde, the prediction of response for a
population would require a characterization of both the frequency
of individual human responses and the severity of effects with
increasing exposure. However, available data do not allow the
development of a well defined dose-response relationship for these
irritation effects. For the exposure conditions presented in
Figure 1-1(a), only a small percentage of persons would respond
and in all likelihood the eye irritation would be very mild and
transitory when an individual enters the home or workplace. The
people at greatest risk of experiencing discomfort due to
1-21
-------
formaldehyde-induced irritation are new homeowners during the
first year of occupancy, particularly under conditions of high
temperature and humidity which are typically associated with
elevated levels of formaldehyde in these homes.
1.4.1.2. • Cellular Effects
Formaldehyde causes cellular changes in the upper respiratory
tract. Studies in animals have shown that formaldehyde can
inhibit mucociliary action after only a few days of exposure, with
a NOEL of 0.5 ppm in rats. Long-term exposure studies have shown
squamous metaplasia in the nasal cavities of rats and monkeys.
The NOEL for this effect in both species is 1.0 ppm, with LOELs
of 2.0 ppm (rats) and 3.0 ppm (monkeys). One study of humans
showed nasal cavity effects in some persons exposed in the range
of 0.1-1.1 ppm (Edling, -et al., 1985).
From these values, it appears that humans and animals may
respond similarly (within a factor of 10) to the cellular effects
of formaldehyde in the nose. Formaldehyde exposures in mobile and
conventional homes and to garment workers fall somewhat below the
NOELs and LOELs for cellular effects as determined from studies in
animals (Figure l-l(b)) Since the anticipated exposures in the
identified populations are close to those associated with effects
in humans and animals, it is expected that home residents and
garment workers may be at some risk of experiencing these non-
cancer effects.
Although quantitative estimates of risk are not possible, the
frequency and severity of response are dose related. Fewer
responses are expected to be associated with less frequent and
1-22
-------
less intense exposure. In addition, the cellular effects are
expected to be reversible once formaldehyde exposure is
eliminated.
1.4.2. Cancer Dose-Response Assessment
In principal, data from studies of humans are preferred for
making numerical risk estimates. However, as is often the case,
the available epidemiologic data on formaldehyde were not suitable
for low dose quantitative cancer risk estimation, mainly because
of a lack of adequate exposure information in the studies.
Accordingly, results from studies in animals were used to estimate
low-dose human cancer risk. In addition, even though the
epidemiolggic studies were not suitable for quantifying a dose-
response curve, those studies with observed statistically elevated
cancer risks provided some support for the animal-based predicted
upper bound risk. This comparison, while yielding valuable
information to the assessment, should be viewed with caution since
exposure levels in these epidemiologic studies were subject to
some variation.
1.4.2.1. Selection of Data
Of the carcinogenicity studies with formaldehyde in animals,
EPA has selected the CUT study in rats as the best study for
cancer risk extrapolation. This study was well designed, well
conducted, included multiple doses, and used a large number of
animals per dose.
Each of the remaining inhalation studies suffered from
various limitations which precluded their use in quantitative risk
assessment. The CUT study in mice showed a limited tumor
1-23
-------
response oaly at. the highest dose of formaldehyde, while the
Albert et al. (1982) study had only a single formaldehyde-exposed
group. Although the Tobe et al. (1985) study contained multiple
dose groups, a tumor response was seen only at the highest dose,
and the number of animals per group was relatively small. Lower
cancer risks than those estimated from the CUT study in rats
would have been predicted had the Agency been able to use the CUT
study in mice for risk extrapolation, while higher cancer risks
would have been estimated had the results from the Tobe et al.
(higher by a factor of ten) or Albert et al. studies been used.
Two types of nasal tumors were observed in the CUT study in
^
rats, squamous cell carcinomas (malignant tumor) and polypoid
adenomas (benign tumor). EPA's risk assessment relied only on the
malignant tumor data of the CUT study to predict human cancer
risks because: (1) the malignant tumor response in formaldehyde-
exposed rats was definite and unequivocal in both males and
females, whereas the frequency of benign tumors reached
statistical significance only when the incidences in males and
females were pooled; (2) the malignant tumor response in the CUT
study in rats showed an increasing dose-related trend, while the
benign tumor response showed a decreasing trend; (3) unlike the
benign tumor response which was not confirmed by the other rat
inhalation studies, similar malignant tumor types were found both
in all rat and mouse inhalation studies with formaldehyde and in a
study of acetaldehyde, a close structural analogue of formaldehyde,
The appearance of benign nasal tumors in rats following
inhalational exposure to formaldehyde in the CUT study
1-24
-------
contributes-to the qualitative weight-of-the-evidence that
formaldehyde may pose a carcinogenic hazard, but because of the
attendant uncertainties they were not included in the
quantitative estimate of human cancer risk. Had the Agency
chosen to use the benign tumor response in the quantitative
estimation of human cancer risk, the predicted values would have
been about ten-fold greater than those reported in Section 1.4.3
using the malignant tumor response alone.
1.4.2.2. Choice of Mathematical Extrapolation Model
Since risks at low exposure levels cannot be measured
directly either by experiments in animals or by epidemiologic
4
studies, a number of mathematical models have been developed to
extrapolate from results at high doses to expected responses at
low doses. The Office of Science and Technology Policy (OSTP)
published principles on model selection which states that:
"No single mathematical procedure is recognized as the most
appropriate for low dose extrapolation in carcinogenesis.
When relevant biological evidence on mechanism of action
exists, the models or procedures employed should be
consistent with the evidence. When data and information are
limited, however, and when much uncertainty exists regarding
the mechanism of carcinogenic action, models or procedures
which incorporate low dose linearity are preferred when
compatible with the limited information."
Data relevant to selecting a model for extrapolation of
cancer risk associated with exposure to formaldehyde were
reviewed; some of the biological information support a direct
1-25
-------
relationship between exposure and carcinogenicity while other
data are consistent with a non-linear response. The Agency,
however, did not conclude that enough information was available
to propose an extrapolation model for formaldehyde that was
different from the one recommended by the OSTP and EPA's
Guidelines for Carcinogen Risk Assessment (i.e., linearized
multistage procedure). The Agency has presented various other
models for comparative purposes.
Biologic evidence on mechanism of action, which can aid in
model selection, largely is inferred from a variety of types of
studies. These are limited and suggestive of several mechanisms
t
for formaldehyde. Mutagenicity studies suggest a direct
relationship (i.e., a linear one) between exposure to
formaldehyde and carcinogenicity. Thus, the ability of
formaldehyde to cause point mutations, chromosome aberrations and
DNA damage is consistent with the chemical's ability to initiate
the carcinogenic reaction.
The steep curvilinearity of the rat nasal carcinoma dose-
response data in the CUT study in rats suggests, however, that
cancer development is greatly accentuated above certain con-
centrations. In keeping with this observation are the results of
experiments on DNA synthesis and cell proliferation following
short-term formaldehyde exposures and the conversion of normal
mucosal cells to squamous cell epithelium (squamous metaplasia)
following longer exposures which indicate that certain toxic
effects are only noted above certain formaldehyde
concentrations. Any relationship between cell proliferation
1-26
-------
following formaldehyde exposures and the carcinogenic process is
currently unknown. Likewise, although squamous metaplasia may
represent a step in the formation of squamous cell carcinoma, its
specific role is uncertain. No lesions that may represent stages
in a continuum between the squamous metaplasia and carcinoma were
identified in the CUT study.
The CUT also conducted molecular dosimetry experiments
attempting to relate ambient exposures to formaldehyde with
tissue-specific levels of formaldehyde-DNA adducts. Use of the
data generated by these experiments in risk extrapolation models
yields lower estimates of risk, sometimes significantly lower
>
than use of the experimental doses. The CUT data have been
reviewed by EPA scientists and a review panel of non-government
scientists to determine whether or not they should be used in the
quantitative risk assessment. Both groups concluded that the
study had several shortcomings which preclude its use in
modifying the doses used in quantitative risk assessment, and
they provided three reasons for their conclusion. First, the
experimental methodologies must be validated to assure that the
experimental assumptions were scientifically sound and that the
formaIdehyde-DNA-protein complexes were identified properly;
second, the single intracellular target used in the study may be
inadequate; and third, and perhaps most important, the use of an
acute exposure model in the CUT study may not be appropriate
because chronic, not acute exposure is most relevant to risk
assessment.
1-27
-------
Different extrapolation models fit the observed data
reasonably well but there are large differences among them in the
risks calculated at low doses. EPA's Guidelines for Carcinogen
Risk Assessment state, however, that goodness of fit to the
observed tumor data by a given model is not an effective means of
discriminating among models. In the absence of compelling
biological evidence on the mechanism of action, as in the case
for formaldehyde, EPA's guidelines specify that the linearized
multistage procedure will be used, with the possible presentation
of various other models for comparative purposes. The analysis
showed that of the models examined, only the one-hit model
*
produced higher risk estimates (about ten fold higher).
Studies show that non-human primates and rats respond
similarly to formaldehyde exposure. Accordingly, an interspecies
scaling factor was not used in the risk extrapolation. This
position was supported by the Consensus Workshop on
Formaldehyde. Consequently, the response of rats and humans was
judged to be the same at equivalent exposure levels and
durations. However, if a conversion factor, such as nasal
surface area, had been used the estimated human cancer risks
would have been about an order of magnitude higher.
1-28
-------
1.4.3. Numerical Risk Estimates
The risk estimates for the linearized multistage procedure,
upper bound (UB) and maximum likelihood estimates (MLE)7 at
various exposure levels are presented in Table 1-2. Risks at any
exposure level range from the upper bound to zero. An
established procedure does not yet exist for making "most likely"
or "best" estimates of risk within the range of uncertainty
defined by the upper bound and zero. The upper bound estimate
for excess lifetime risk of developing cancer is 3 x 10~4
[Group Bl]8 for apparel workers exposed to formaldehyde.at the
0.17 ppm level, 2 x 10 [Group Bl] for residents of mobile homes
who are exposed for 10 years to an average level of 0.10 ppm; and
1 x 10~4 [Group Bl] for residents of some conventional homes who
are exposed for 10 years to an average level of 0.07 ppm. The
upper bound unit risk estimate for an ambient exposure of 1 ug/m^
' The shapes of most models' upper bound estimates tend to
parallel the shapes of the models themselves, unless a procedure
has been devised to provide otherwise. This is the case for the
linearized multistage procedure, which provides a linear upper
bound estimate at low dose. The maximum likelihood estimate
(MLE), which is the estimate given by a fitted model, takes only
the experiment to which the model has been fitted into account.
The upper bound estimate, on the other hand, is intended to
account for experiment to experiment variability as well as
extrapolation uncertainties.
° EPA's Guidelines for Carcinogen Risk Assessment recommend
categorizing chemicals in Group B (Probable Human Carcinogen)
when "the evidence of human carcinogenicity from epidemiologic
studies ranges from almost 'sufficient1 to 'inadequate.1 To
reflect this range, the category is divided into higher and lower
degrees of evidence. Usually, category Bl is reserved for agents
for which there is at least limited evidence of carcinogenicity
to humans from epidemiologic studies."
1-29
-------
(0.00082 ppm) for 70 years is 1.3 x 10~5 [Group Bl]. The fitted
model gives the maximum likelihood estimate curve and, specific
to the CUT study, it has a pronounced S-shape. By contrast, as
the linearized multistage procedure's upper bound estimate is
traced toward lower doses, its linear nature accomodates
increasing variability and extrapolation uncertainty. Both
estimates are shown in Table 1-2 to illustrate how the
perspectives they give on risk differ. Thus at 3 ppm (which is
in the experimental range), the difference between the MLE and
the UB is ten-fold, whereas at about one-tenth of that exposure,
a 100,000 fold difference is generated.
The lower bound on risk is always recognized to be as low as
zero. The upper bound estimate is ordinarily shown to allow for
extrapolation uncertainty. It is for this reason, along with
adherance to EPA's Guidelines for Carcinogen Risk Assessment,
that the upper bound was selected to represent potential human
risk. While some of the existing information on formaldehyde is
consistent with non-linear interpretations, some support for a
linearized upper bound comes from the epidemiologic studies. The
excess cancer incidences observed in the epidemiologic studies
are about the same as the upper bound on lifetime risk based on
the rat nasal carcinoma data.
1-30
-------
TABLE 1-2
SUMMARY OF CANCER RISKS ASSOCIATED WITH FORMALDEHYDE EXPOSURE
Population Segement
(Exposure Level)
Lifetime
Individual Risk
Current OSHA std. (3 ppm)
UBb 6 x 10'3, [B1J
MLEC 6 x 10"4 [81]
Garment Workers
NIOSH
(0.17 ppm)
UB 3 x 10"* [Bl]
MLE 4 x 10"9 [Bl]
Mobile Home
Residents
(0.10 ppm 10-yr average)
UB- 2 x 10~ [Bl]
MLE 2 x 10"10 [Bl]
Conventional Home*
Residents
(0.07 10-year average)
UB 1 x 10"4 [Bl]
MLE 6 x 10"11 [Bl]
Home/Environment
Background Upper Limit
(0.05 ppm)
10 yr.
70 yr,
UB 7.0 x 10"f [Bl]
MLE 1.0 x 10"11 [Bl]
"*
[Bl]
UB 5.0 X 10
MLE 1.0 x 10"10 [Bl]
* For homes containing substantial amounts of urea-formaldehyde
pressed wood (e.g./ floor underlayment and/or paneling)
b upper Bound
° Maximum Likelihood Estimate
d Airborne Unit Risk, 1 ug/m3 - 70 yrs; Lifetime individual risk,
UB = 1.3 x 10~5 [Bl]
1-.11
-------
2. BACKGROUND
In November 1979, EPA received information that the interim
results of a 24-month bioassay in rats conducted by CUT showed
that a number of the rats had developed nasal cancer after
inhalation of HCHO.
In November of 1980, the Federal Panel on Formaldehyde,
formed by several Federal agencies under the aegis of the
National Toxicology Program, published a report finding that
CIIT's bioassay methodology was consistent with accepted testing
standards. Using the data available through the 18-month point
of the CUT study, the Federal Panel concluded that "formaldehyde
should be presumed to pose a risk of cancer to humans." Also- in
November 1980, CUT presented the preliminary results of the full
study. CUT pathologists reported finding statistically
significant increases, as compared with controls, in the
incidence of malignant tumors in rats exposed to HCHO vapor at
the highest of the three levels they tested (14.3 ppm).
In February 1982, based on its evaluation of the toxicity
and exposure data on HCHO then available, EPA decided that,
although HCHO had been found to be carcinogenic under the
conditions of the test, the available information as to HCHO's
cancer risk to humans did not meet the statutory criteria for
priority designation under section 4(f) of TSCA.
To assist its evaluation of HCHO the Agency funded the
National Center for Toxicological Research to sponsor a Consensus
Workshop on Formaldehyde (the Workshop). The Workshop was held
2-1
-------
in Little Rock, Arkansas from October 3 through 6, 1983. Overj D
government, industry,'university, and public interest
organization scientists served on the following eight Panels:
(1) Exposure; (2) Epidemiology; (3) Carcinogenicity/
Histopathology/Genotoxicity; (4) Immunology/Sensitization/
Irritation; (5) Structure Activity/Biochemistry/Metabolism;
(6) Reproduction/Teratology; (7) Behavior/Neurotoxicity/
Psychological Effects; and (8) Risk Estimation. Each Panel
(except the Risk Estimation Panel) was charged with the task of
reviewing the major scientific studies relevant to that Panel's
area. The Panel members were also asked to address a number of
discussion topics and prepare a consensus report addressing those
topics.
When the Panel deliberations were finished, draft reports
were provided to the Risk Estimation Panel. The Risk Estimation
Panel was charged with the task of determining how the data could
be assessed to make reasonable risk estimates for humans exposed
to HCHO at various levels and through different routes.
The decision process of the February 1982 decision under
section 4(f) of TSCA generated considerable controversy and
formed the basis for a lawsuit by the Natural Resources Defense
Council (NRDC) and the American Public Health Association (APHA)
(NRDC v. Ruckelshaus, No. 83-2039, filed in the United States
District Court for the District of Columbia, July 18, 1983).
In view of public controversy concerning the process and
policy issues associated with the Agency's section 4(f) decision
2-2
-------
on the cancer hazard of HCHO, EPA announced in the FEDERAL
REGISTER of November 18, 1983 (48 FR 52507) its decision to
rescind its February 1982 decision, and to ask the public to
submit views, arguments, and data relevant to determining whether
HCHO should be given priority consideration under section 4(f) of
TSCA. Comments were due at EPA by January 17, 1984; EPA
announced that it expected to reach a new decision by May 18,
1984.
On May 23, 1984 EPA announced in the Federal Register (49 FR
21898) that two HCHO exposure categories triggered section 4(f)
of TSCA (possible widespread cancer risk). The exposures which
led to the decision are those associated with manufacture of
apparel from fabrics treated with HCHO-based resins and residence
in conventional and manufactured homes containing construction
materials in which certain HCHO-based resins are used.
In addition to HCHO's potential cancer risks, HCHO's other
effects should be considered in any action to reduce health
effects from HCHO. The assessment of the risks from acute
respiratory effects was prepared to be considered along with the
carcinogenic risk assessment in the overall investigation of
HCHO. The hazard discussion of noncarcinogenic effects in the
risk assessment is based in part on reports from the Consensus
Workshop on Formaldehyde, a report of the Cosmetic Ingredient
Review Expert Panel, a hazard assessment by Ulsamer et al.
(1984), and the National Research Council report titled HCHO and
other Aldehydes prepared under contract to EPA. The risk
2-3
-------
assessment focuses on the possibility of determining a dose-
response for these noncancer effects because while many of the
effects are well documented, the dose-response patterns in the
human population are not. Methods used by HUD and OSHA to relate
the proportion of the human population responding at particular
exposure levels have been analyzed. In addition, EPA has
reviewed selected human studies to determine if dose-response
relationships can be described.
2-4
-------
3. PHYSICAL-CHEMICAL PROPERTIES
HCHO is the simplest member of. the aldehyde chemical
category. It exists in many different forms. Both liquid a-t-i
gaseous HCHO polymerize readily at ordinary temperatures and can
bo kept in pure monomeric state only for a limited time. Pure
monomeric HCHO is a colorless, pungent gas. Agueous HCHO, called
formalin, is a clear/ colorless solution containing about 37
percent by weight of dissolved HCHO in water (room temperature),
usually with 6 to 15 percent methanol added to prevent
polymerization. Solutions containing over 30 percent by weight
become cloudy on standing and precipitate polymer at ordinary
temperatures. Concentrated liquid HCHO-water systems containing
up to around 95 percent HCHO are obtainable, but the temperature
necessary to maintain solution clarity and -irevent separation of
solid polymer increases from around room t-e-iperature to 120°C as
the solution concentration is increased. !"••>.e other forms of HCHO
are polymers, the best known of which are para-HCHO and trioxane
(trioxymethylene). HCHO is sold and transported only in solution
or in the polymerized state.
The molecular weight of HCHO is 30. It has the followinq
structural formula:
0
II
H-C-H
3-1
-------
The chemical name used by Chemical Abstracts Service is
HCHO, and its Chemical Abstract number is 50-00-0. Synonyms*
include HCHO; HCHO qas; HCHO solution; formalin; formalin 40;
formalin 100%; formic aldehyde; methaldehyde; methanal; methyl
aldehyde; methylene glycol; methylene oxide; oxomethane;
oxymethylene; paraform; oara-HCHO; polyoxymethvlene qlycols;
0(-polyoxymethylene; fl -polyoxymethylene; tetraoxymethylene;
ai. -tr ioxane ; trioxane; and c(-tr ioxymethylene .
Dry HCHO qas condenses on chillinq to qive a liquid that
boils at -19°C and freezes to a crystalline solid at -188C.
Vapor pressure is 400 mn at -33°C. HCHO qas is flammable havinq
a heat of combustion of 4.47 kcal per qram. It forms explosive
mixtures with air and oxygen. At atmospheric pressure,
flammability is reported to ranqe from 12." to 80 volume oercent,
HCHO-air mixtures containinq 65 to 70 Derc--,t beinq the most
readily flammable. HCHO is soluble in wat~r, acetone, benzene,
diethyl ether, chloroform and ethanol (IARC, 1982). Solutions
obtained with the nonpolar solvents are somewhat more stable but
also precipate polymer on storage. HCHO solutions have a
definite flash point which is lowered by the oresence of
methanol. The flash point of commercial HCHO 37.5% solution with
14.0% methanol (by weight) is 56°C (132°F). In view of their
unique nature, it is recommended that flash ooint values for HCHO
solutions be regarded as approximations and that the solutions be
regarded as potentially flammable at least 10°F below the
reported figures.
*Includes synonyms for polymeric forms of HCHO.
3-2
-------
The partial pressure of HCHO vanor over commercial solutions
is also increased by methanol. The partial pressure of HCHO over
37 percent solution containing 9 percent methanol is 4.2 mm at
35°C, whereas a 37 percent solution containina 1 percent methanol
has a partial pressure ot 2.7 mm under the same conditions
(Walker, 1975).
3-3
-------
4. HAZARD OP CARCINOGENIC EFFECTS
4.1. Long- and short-term Animal Tests
The first long-term study reporting the carcinogenicity of
HCHO in animals by the inhalation route was one by Kerns et al.
(1983) (CIIT-sponsored study performed at Battelle Columbus
Laboratories) which reported statistically significant increased
levels of squamous cell carcinomas in the nasal cavities of rats
at 14.3 parts per million (ppm). In the study, groups of
approximately 120 male and 120 female Fischer 344 strain *rats and
C57BL/6 X C3HFj strain mice, were exposed by inhalation to 0,
2.0, 5.6, or 14.3 ppm of HCHO gas for 6 hours per day, 5 days per
week, for 24 months. The exposure oeriod was followed by up to
six months of nonexposure. Interim sacrifices were conducted at
6, 12, 18, 24, 27 and 30 months. All major tissues from each
organ system (approximtely 50 tissues/animal) in the control and
high exposure groups were examined histologically. Sguamous cell
carcinomas were observed in the nasal cavities of 103 rats (52
females and 51 males) and 2 male mice exposed to 14.3 ppm of HCHO
and in 2 rats (one male and one female) exposed to 5.6 ppm of
HCHO gas. The first tumor clinically observed in female rats of
the 14.3 ppm group was at 358 days past first exposure and 432
days for males. The adjusted cumulative incidence rate (Kaplan-
Meier life table analysis) of squamous cell carcinomas in -nale
and female rats of the 14.3 ppm exposure group at 24 months was
67 and 87%, respectively. Tumors in male mice were discovered at
the 24-month sacrifice. The incidence of nasal carcinomas in
rats showed a dose-response relationship. See Table 4-1 for a
summary of tumor response in rats.
4-1
-------
Table 4-1.
9MMKT Of NBQPIAST1C UNIONS IN TOE NASAL CAVITY Of FISCHER 344 RATS EXFOSO) TO KXMAUEHYIE GAS*
Formaldehyde
(ppm)
0
2.0
5.6
14.3
Sex
H
F
N
F
N
F
H
f
tto. of nasal
cavities
evaluated
118
114
118
118
119
116
117
115
Squanous cell
carcinoma
0
0
0
0
1
1
51
52
ftiorly
Differentiated
carcinoma
0
0
0
0
0
0
0
1
Adeno-
carcinona
0
0
0
b
0
0
1
0
Undi f f erent iated
carcinoma or
sarcoma
0
0
0
0
0
0
2*
0
Carci no-
sarcoma
0
0
0
0
0
0
1
0
Iblypoid
adenoma
1
0
4
4
5
0
2
0
Cfetao-
chondrau
1
0
0
0
0
0
0
0
I
K)
*Table adapted from Kerns et al. (1983)
a A rat in this group also had a squamous cell carcinoma.
-------
Although the two squamous carcinomas in mice at 14 ppm were
not statistically significant in comparison with the incidence in
control mice in the study, the finding suggests that the effect
is related to HCHO exposure because the natural background rate
for such nasal cancers is very low in this strain of mice, with
only one neuroepithelioma and one angiosarcoma having been
reported by Stewart et al., 1979 (Kerns et al., 1983).
The difference in susceptibility of rats and mice may be
due, in part, to a greater reduction in respiratory minute volume
in mice than in rats during exposure to an irritating agent. In
a study by Chang et al. (1983) changes in minute volume, nasal
cavity disposition, and cell proliferation were examined. It was
found that mice exposed to 15 ppm HCHO for 6 hours experienced an
approximately 50% reduction in minute volume whereas rats
exhibited at 20% decrease. If a "dose" of HCHO is calculated
from adjusting for reduction in minute volume and other data, it
can be seen that for mice the dose received at 14.3 ppm in the
Kerns et al. (1983) study is one-half that received by rats at
14.3 ppm (see also Swenberg et al., 1983). Thus, the tumor
response in mice at 14.3 ppm is comparable to the response in
rats at 5.6 ppm. Interestingly, mice and rats at these exposures
showed nearly identical tumor responses, i.e., two squamous cell
carcinomas out of approximately 240 mice and rats.
In addition to the squamous cell carcinomas, small numbers
of benign tumors characterized as polypoid adenomas were observed
in rats at each dose level. These benign tumors exhibited a
4-3
-------
dose-response relationship with a negative trend. However,
because this type of benign nasal tumor is rare in control rats
it is likely related to HCHO exposure. For a further discussion
of these lesions and their use in this risk assessment see Data
Selection for Quantitative Analysis after this section.
Significant squamous metaplasia was also observed. See
Figure 4-1 for frequency and locations (also see Figure 4-3). In
rats at 2.0 ppm, purulent rhinitis, epithelial dysplasia, and
squamous metaplasia were present in the anterior portion of the
turbinates (Level I) at 12 months. The frequency of metaplasia
increased up to 24 months and then decreased significantly
(p<0.05) at 27 months (three months post exposure). In the 5.6
ppm group, purulent rhinitis, epithelial dysplasia, and squamous
metaplasia were observed in the anterior and middle portions of
the nasal cavity (Levels I, II, and III). Siqnificant (p<0.05)
regression of squamous metaplasia was noted at 27 months (post
exposure). Similar but more severe and extensive lesions were
observed in the 14.3 ppm exposure group in all regions of the
nasal cavity. Significant regression of squamous metaplasia was
only observed in the posterior portion of the nasal cavity
(Levels IV and V). In all exposure groups, epithelial dysplasia
was detected earlier than squamous metaplasia.
4-4
-------
Figure 4-1. Frequency of squamous metaplasia in the nasal
cavity of Fischer 344 rats exposed to 2.0 ppm (top), 5.6
ppm (middle), or 14.3 ppm (bottom), of formaldehyde yas for
24 months. Nasal cavity Levels I, II, IV, and V were not
evaluated at the 6- and 12-month interim sacrifices in the
14.3 ppm exposure group. Figure taken from Kerns et al.
(1983).
4-5
-------
Inflammatory A dysplastic, and squamous metaplastic alterations
of the respiratory epithelium of mice were observed. These
lesions were most severe in the 14.3 ppm exposure group (see
Figure 4-2). .\ few mice in the 5.6 ppm group had dysplastic
changes'with serous rhinitis at 18 months in Level II. By 24
months, a majority (<90%) of mice in the 14.3 ppm group had
dysplastic and metaplastic changes that were associated with
seropurulent rhinitis. At that time period, only a few mice in
the 5.6 ppm exposure group had dysplasia, metaplasis or serous
rhinitis in Level II. Mice in the 2.0 ppm group were generally
free of significant lesions with only a few animals with serous
rhinitis at 24 months.
One complication noticed during the Kerns et al. (1983)
study was a spontaneous outbreak in rats of sialodacryeo-
adenitis. The evidence for this consisted of (a) decreased body
weight in all dosed and control rat groups at about the 52nd week
of the experiment/ followed by prompt recovery of body weight;
and (b) histopathologic demonstration of typical lesions in
lacrimal and salivary glands of dosed and control rats in the
12-month sacrifice groups. Evidence of sialodacryeoadenitis was
not found in rats sacrificed at 6 or 18 months or in those with
unscheduled deaths. Virus isolation, viral antigen
demonstration, and serologic tests for.antibodies were not
attempted in rats or mice.
4-6
-------
Figure 4-2. Frequency of squamous metaplasia in the nasal
cavity of B6C3F] mice exposed to 14.3 ppm of formaldehyde
gas. Figure taken from Kerns et al. (1983).
With regard to HCHO in the exposure chamber in the Kerns
study, a panel of experts reviewed the method of generation of
HCHO and monitoring and agreed that "the Battelle approach to
HCHO vapor generation (heating paraformaldehyde) was a suitable
adaptation of accepted methods and principles and, therefore, was
sound and based upon the best available technology. The same
type of assessment applied to the chamber air monitoring system,
which also combined two well established procedures" (Gralla et
al., 1980).
Other studies support the results of the Kerns (CIIT)
study. In two studies reported by Albert et al. (1982) (comolete
results for one study and preliminary results for the other),
rats were exposed for life by inhalation to HCHO alone, mixtures
of hydrochloric acid (HCL) and HCHO, or HCL alone.
In the first study, 99 male Sprague-Dawley rats were exposed
to a mixture of HC1 and HCHO (premixed at high concentrations
4-7
-------
before introduction into the exposure chamber to maximize the
production of bis(chloromethyl) ether (BCME)). This was done
because the investigators were studying the hazard associated
with the use of HC1 and HCHO in close proximity in the
workplace. A report had suggested the production of BCME from
mixing HCL and HCHO gas. The average concentrations were 10.6
ppm HCL, 14.7 ppm HCHO, and about 1 part per billion (ppb)
BCME. Of the 99 animals exposed to the test mixture, 25
developed squamous cell carcinomas of the nose. The contribution
by the BCME was thought to be minimal because the expected
response to 1 ppb of BCME was estimated to be less than 1.5
percent (based on authors' comparison of 20 exposure days at 100
ppb of BCME with 500 exposure days at 1 ppb of BCME) and there
was a 25 percent incidence of nasal tumors in the study. In
addition, BCME normally produces neurogenic carcinomas (mainly
esthesioneuroepitheliomas), none of which were seen in the
study. The uncertainty of comparing different factors involved
in dose-rate versus total delivered dose in tumore induction is
not resolved, however.
The second Albert et al. (1982) study, in which male
Sprague-Dawley rats (100 per oroup) were exposed to HCL alone
(10.2 ppm), premixed HCL-HCHO mixture (14.3 ppm HCHO/10.0 ppm
HCL), nonpremixed HCHO-HCL mixture (14.1 ppm HCHO/9.5 ppm HCL),
or HCHO alone (14.2 ppm), showed statistically significant
numbers of squamous cell carcinomas of the nasal cavity in the
rats exposed to HCHO alone and the HCL-HCHO mixtures. A control
4-8
-------
group of 100 rats was used. No nasal cancers were seen in the
HCL-only exposed rats or in the controls. Also, it appeared that
the irritant HCL gas did not enhance the carcinogenic response of
HCHO because the frequency of tumors was similar in the HCHO
alone and HCHO-HCL groups. Table 4-2 displays final results of
the study (Sellakumar, 1985). Work by Swenberg et al. (1984) in
which Sprague-Dawley and Fischer 344 rats were exposed to 0, 10,
20 or 50 ppm of HCL gas for 6 hours per day, 5 days per week, for
up to 90 days, indicates that HCL may be considerably less
irritating to the uoper respiratory tract of rats than HCHO. For
those rats exposed for four days and killed 18 hours later, only
the 50 ppm group had significant HCL-induced lesions, consisting
of focal degeneration, epithelial hyperplasia, and early sguamous
metaplasia on the dorsal tip of the maxilloturbinate of the most
anterior section. Since maximum nasal irritation in rats from
HCHO occurs within a few days after exposure begins, the lack of
significant nasal.irritation from HCL in the 10 and 20 ppm groups
indicates that HCL in the Albert study was not sufficiently
irritating to draw conclusions regarding the role of irritation
in HCHO-related carcinogenesis.
In a study reported by Tobe et al. (1985), groups of 32 male
Fischer 344 rats were exposed to HCHO for 6 hours per day, 5 days
per week, for 28 months. The five test groups were as follows:
colony control, room control, 0.3, 2.0, and 15 ppm HCHO. The
significant finding was sguamous cell carcinoma (14 cases) and
papilloma (5 cases) in the 15.0 ppm group. No tumors were
4-9
-------
Table 4-2.
SUMARY CF NEOPLASTie LESIONS IN THE NASAL CAVITY GP SPRAGUE-CAWLEY RATS*
Air controls
(99 rats)
Squamous cell carcinoma 0
Pap i llama or polyps 0
i
o
Fibrosarcona 0
Adenocarcinana 0
Esthesioneuroepitheliama 0
Colony
controls
(99 rats)
0
0
0
0
0
10 ppm HCL 14 ppm HCHO
(99 rats) (100 rats)
0 38
0 10
0 1
0 0
0 0
PremixecJ 14 ppm
HCHO & 10 ppm
HCL (100 rats)
45
13
1
1
1
Nan-premixed 14 p|*n
HCHO & 10 ppm HCL
(100 rats)
27
10
0
2
0
*Fron Sellakurar (1985)
-------
observed in the 0.3 and 2.0 ppm groups. Rhinitis, squamous
metaplasia and hyperplasia of the nasal respiratory epithelium
were observed in all HCHO exposed groups.
It should be noted that while the carcinoma response was
similar between the Kerns, Albert and Tobe studies, the benign
tumor response was markedly different. In the Kerns study only
benign polypoid adenomas were observed, whereas in the Albert and
Tobe studies benign papillomas were observed. The basis for
these differences is difficult to explain. It could represent a
strain difference or some unknown factor. (Tobe used the same
strain of rats as Kerns, Fischer 344, but the small number used
at each dose as compared to Kerns (32 vs. 240) may explain the
failure of polypoid adenomas to be detected.) Consequently,
statements about the significance of' these lesions in discussions
of human risk must be approached with caution.
Two other chronic inhalation studies with HCHO designed to
investigate possible cocarcinogenic effects of this agent in the
upper and lower airways have been reported (Horton et al., 1963,
and Dalbey et al., 1982). Since the nasal tissues were not
systematically examined histologically, the value of these
studies in assessing the carcinogenicity of HCHO is accordingly
limited. In spite of these reservations, the studies have some
bearing on HCHO carcinogenicity.
In the study reported by Horton et al. (1963), C3H mice were
exposed to coal tar aerosol and/or to HCHO at concentrations of
40, 80, 160 ppm. Exposures were carried out for 1 hr/day, 3
4-11
-------
days/week for 35 weeks, except for the 160 ppm group which was
exposed only for 4 weeks because of toxicity. Only 15 mice
survived to 1 year. There is no mention of histopathological
evaluation of nasal tissues, so presumably no grossly visible
tumors were observed. Coal tar aerosol exposure resulted in lung
tumor formation in 5 animals (1 invasive carcinoma), but HCHO
exposure did not. No evidence was found for any cocarcinogenic
effects of HCHO. The major shortcomings of this study for
evaluating the carcinogenicity of HCHO are that too few animals
survived past one year, the individual exposures were short, most
groups were exposed only for 35 weeks, and complete
histopathology of nasal tissues was not reported.
In a study by Dalbey (1982) male Syrian golden hamsters were
used to study the chronic effect of HCHO and diethylnitrosamine
(DEN). In the first part of the study, 88 hamsters were exposed
to 10 ppm HCHO, 5 times/week .for life. There were 132 untreated
controls. The second part of the study examined HCHO's
promotional potential. For the second part of the study the
hamsters were divided into 5 groups: 50 untreated controls; 50
hamsters exposed to 30 ppm HCHO, 5 hrs/day, 1 day/week for life;
and 3 groups receiving DEN injections of DEN (0.5 mg, once per
week for 10 weeks). Of the three groups receiving DEN, one
consisted of 100 hamsters receiving only DEN, a second group of
50 hamsters were exposed to 50 ppm (5 hrs) of HCHO 48 hours prior
to each injection of DEN, and the third group was exposed to 30
ppm HCHO (5 hrs/day, 5 days/week) for life, beginning 2 weeks
after the last DEN injection.
4-12
-------
In the hamsters exposed to 10 ppm HCHO for life there was no
evidence of carcinogenic activity/ but survival time was reduced
(p<0.05) relative to controls. Toxic effects of HCHO in the
nasal cavity were limited to hyperplastic and metaplastic areas
in 5% of the hamsters. No hyperplasia or metaplasia was observed
in the controls. The incidence of rhinitis was similar in both
control and exposed hamsters, and was not considered to be
related to exposure of HCHO.
Caution must be exercised when comparing this part of the
Dalbey (1982) study with the Kerns et al. (1983) study. One
factor that should be considered is that the pathology evaluation
in the Dalbey (1982) study was less rigorous. Only 2 sections of
the nasal turbinates were examined as compared to sections taken
from 5 anatomical levels of the nasal cavity of rats in the Kerns
et al. (1983) study.
Also, the Kerns et al. (1983) study used three HCHO exoosure
levels ( 2.0, 5.6, and 14.3 ppm) whereas only 10 ppm of HCHO was
used in the Dalbey (1982) study. If one compares the ppm-hrs/week
received by rats at 5.6 ppm in the Kerns study and hamsters at 10
ppm, one sees that the ppm-hrs/week for the hamster is equivalent to
a hypothetical dosing regimen of 8.3 ppm for rats (5.6 ppm X 6
hr/days X 5 days = 168 ppm-hrs/week vs. 10 ppm X 5 hr/days X 5
days = 250 ppm-hrs/week; this is equivalent to 250 ppm-hrs/wk * 6
hr/day X 5 d/wk =8.3 ppm). Since only two squamous cell
carcinomas were seen in the Kerns et al. (1983) study out of 240
rats at 5.6 ppm, the likelihood of detecting a tumor in the
4-13
-------
Dalbey (1983) study, which used 88 hamsters, is poor. in fact,
there is a 30% probability that the absence of tumors is due to
chance.
Finally, it is well established (Kane et al., 1979 and
Buckley et al., 1984) that many sensory irritants decrease the
respiratory rate of exposed animals. For rats the RD50
(concentration required to reduce respiratory rate by 50%) for
HCHO is approximately 32 ppm and for mice it is 3.1 ppm. If one
assumes that a HCHO-related sensory irritant response is found in
hamsters (little data have been developed on the hamster in this
regard) (see Alarie, 1985), it is possible that the hamsters in
the Dalbey (1982) study reduced their respiratory rate, which
would further lessen the dose to target tissue.
Although an RD50 value for HCHO has not been reported for
hamsters, a study by Feron et al. (1978) comparing the responses
of hamsters, rats, and rabbits to .acrolein vapor indicates that
hamsters may resemble mice more in their respiratory response to
HCHO than rats. The hamsters were slightly affected (nasal
cavity lesions) at 1.4 ppm and severely affectd at 4.9 ppm by the
acrolein. In contrast/ rats were slightly affected at 0.4 ppm
and were more severely affected at 1.4 and 4.9 ppm. This
response is similar to the difference in response (nonneoplastic
lesions) between rats and mice in the Kerns et al. (1983) study
where rats were affected at all dose levels whereas mice, because
of a lower RD50 value for HCHO, were affected only slightly at
5.6 ppm and more seriously at 14.3 ppm. However, in studies by
4-14
-------
Kendrick et-al. (1976) and Rubin et al. (1978) in which the
effects of tobacco smoke inhalation were investigated for rats
and hamsters, it was found that at least for tobacco smoke, rats
and hamsters had similar reductions in breathing rate and minute
volume and in the amount of particulate matter retained in the
test animals. Consequently, conclusions regarding the comparison
of HCHO-induced respiratory changes in rats and hamsters must
await further research.
In the second part of the Dalbey (1982) study no tumors were
observed in untreated hamsters or those hamsters receiving only
HCHO. However, 77% of the DEN-treated controls had a tumor
(adenomas) at ten or more sites in the respiratory tract.
Although HCHO exposure concurrent with, or after, DEN treatment
did not increase the number of tumor-bearing animals (TBA), the
number of tumors per animal (tracheal tumors) was nearly doubled
over DEN-only controls when HCHO was administered 2 days prior to
each of 10 weekly DEN treatments, whereas post-HCHO treatment had
no measurable effect. Thus, under conditions of the test, HCHO
was a cofactor in chemical carcinogenesis. However, there was a
corresponding decrease in lung tumors in hamsters exposed to both
agents; this suggests that the effect on the trachea may be
within the limits of experimental variability (Consensus Workshop
on Formaldehyde, 1984). In addition, survival in the HCHO-DEN
groups was poor, which further complicates the findings of this
part of the Dalbey study.
4-15
-------
In a study by Rusch et al. (1983), groups of 6 male
Cynomologus monkeys, 20 male and 20 female Fischer 344 rats, and
10 male and 10 female Syrian golden hamsters were exposed to 0,
0.20, 1.0 and 3.0 ppm HCHO for 22 hrs/day, 7 days/week, for 26
weeks. The most significant finding was squamous metaplasia/
hyperplasia in rats and monkeys at 3.0 ppm; little or no response
was seen at the lower exposure levels. Hamsters did not show any
significant responses at any exposure level. The results from
this study indicate that concentration may be more important than
total dose if squamous metaplasia/hyperplasia is the response
measured, when the results are comoared to those of the Kerns et
al. (1983) study. In the Kerns study, squamous metaplasia was
found in rats in the 2.0 ppm exposure group during the course of
the exposure (2.0 ppm 6 hr/day, 7 days/week, for life). However,
in the Rusch et al. (1983) study rats exposed to 1.0 pom HCHO had
no squamous metaplasia, even though they received a total dose
2.5 times that received by the rats at 2.0 ppm in the Kerns et
al. (1983) study. This study design was unlikely to show any
neoplastic response because of its small number of animals and
short duration.
The carcinogenicity of HCHO also has been tested by a
variety of other routes of administration including subcutaneous
injection in rats (Watanabe et al. 1954, 1955), ingestion by mice
and rats (Delia Porta et al. 1968, 1970), and application to the
buccal mucosa in rabbits (Meuller et al., 1978). Because of the
experimental protocols used, none of these studies permits firm
4-16
-------
conclusions regarding HCHO carcinogenicity. Nonetheless, some of
the studies give definite clues that HCHO may be carcinogenic to
a variety of target tissues as well as to a variety of animal
species (and not only to the nasal epithelium of rats).
In one experiment, Meuller et al. (1978), applied a solution
of 3% formalin to the oral mucosa of rabbits, using an "oral
tank." Each exposure lasted for 90 minutes and was repeated 5
times per week for a period of 10 months. As a result, 2 out of
6 rabbits developed grossly visible leukoplakias that, according
to the authors, showed histological features of carcinoma in
situ. Unfortunately, the information given on the
histomorphology of the lesions is very scanty.
Other experiments which suggest that HCHO produces
carcinogenic effects are those by Watanabe et al. (1954, 1955),
who injected rats (strain unknown) subcutaneously with formalin
and with hexamethylenetetramine (HMT, from which HCHO is
liberated in vivo) and produced injection-site sarcomas.
However, several other studies carried out with HMT by
Brendel (1964) who administered HMT by gavage to rats and Delia
Porta et al. (1968, 1970) who administered HMT in drinking water
to mice and rats, resulted in negative findings. The
significance of these finding must be tempered by the fact that
chemicals often give disparate results by different routes of
exposure. For example, hexamethylphosphoramide (HMPA) is a
potent nasal carcinogen by inhalation, but was not carcinogenic
in rats fed HMPA for 2 years (Lee and Trochimowicz, 1984).
4-17
-------
A study by Klenitzky (1940) in which "formol oil" was
applied 50 times to the cervix uteri of mice resulted in no
tumors.
Finally, a study by Spangler and Ward (1983) has been
interpreted as showing weak promoting activity of HCHO:acetone
solutions on Sencar mouse skin. However, in another study by
Krivanek et al. (1983) on CD-I mice, no promotion was observed
when nonirritating applications of HCHOracetone solutions were
used (Consensus Workshop on Formaldehyde, 1984).
4.2. Data Selection for Quantitative Analysis
As discussed above, there are a number of studies available
which indicate the carcinogenic potential of HCHO. For the
purpose of Quantitative Risk Assessment, the Agency generally
chooses a well-designed and conducted study that uses the most
sensitive species of animal (EPA, 1986).
In the case of HCHO, the Kerns et al. (1983) study (CUT
study)fits these criteria. This study has been reviewed by a
number of panels (IRMC, 1984; Consensus Workshop on Formaldehyde,
1984) and has been found to be of sufficient quality for risk
estimation purposes. The selection of the Kerns et al. (1983)
study is consistent with EPA's Carcinogen Risk Assessment
Guidelines (EPA, 1986). Since squamous cell carcinomas were the
only statistically significant malignant tumors observed in the
study, they are the primary end point used for quantitative risk
assessment. A small number of benign tumors, were also
observed. The Guidelines state that benign tumors should be
4-18
-------
combined with malignant tumors for risk estimation unless the
benign tumors are not considered to have the potential to
progress to the associated malignancies. The following discusion
explores this question.
4.2.1. Polypoid Adenomas/Other Tumors Observed
In the Kerns et al. (1983) study, a small number of polypoid
adenomas were reported in the rats: 1, 8, 6, and 5 adenomas in
the 0, 2.0, 5.6, and 14.3 ppm groups, respectively. Because a
number of questions were raised about the accuracy of the
diagnosis of these lesions, they were reexamined by a pathology
working group (PWG) (Boorman, 1984; Consensus Workshop, 1984).
The results of the PWG reexamination are as follows: 1, 8, 5,
and 2 adenomas in the 0, 2.0, 5.6, and 14.3 ppm groups,
respectively. However, two adenomas diagnosed at 2.0 ppm and one
at 5.6 ppm were borderline calls between focal hyperplasia and
small benign tumors. See Table 4-3 for a breakdown by dose and
sex. In addition, two lesions originally diagnosed as nasal
carcinomas were rediagnosed as adenocarcinoma and poorly
differentiated carcinoma which were thought to be morphologically
related. This has relevance to the following discussion of the
potential of polypoid adenomas to progress to a cancer.
The PWG was asked to speculate about the possible
progression of the polypoid adenomas. The consensus of the PWG
was that there was no evidence that polypoid adenomas progressed
to squamous cell carcinomas and that they should not be combined
with squamous cell carcinomas for statistical purposes (Boorman,
1984). This recommendation was accepted by the Risk Estimation
4-19
-------
Table 4-3.
INCIDENCE OP POLYPOID ADENOMA AS
REPORTED BY PWG
DOSE (ppm) Statistically
Sex _0 2.0 5.6 14. 3 Total Significant3
M 1 4* 5** 2 12 No
Nasal cavities
evaluated*** (118) (118) (119) (117)
F 04004 No
Nasal cavities
evaluated (114) (118) (116) (115)
Combined 1 8 5 2 16 Yes at 2.0 ppm
Nasal cavities
evaluated (232) (236) (235) (232)
aOne tailed Fisher exact test. Significance determined for each
dose level.
*Two tumors in this grouo were judged to be borderline
lesions between small benign tumor and focal hyperplasia.
**0ne tumor in this group was judged to be a borderline
lesion between small benign tumor and focal hyperplasia.
***From Kerns et al. 1983.
4-20
-------
Panel of the Consensus Workshop on Formaldehyde (1984). In
addition, an analysis of the localization of the tumors observed
in the Kerns study by Morgan et al. (1985) revealed that the
squamous cell carcinomas generally occurred on the anterior
portion of the lateral aspect of the nasoturbinate and adjacent
lateral wall (57%) or the mid-neutral nasal septum (26%). In
contrast, the polypoid adenomas were confined to a small area of
the anterior nasal cavity and were restricted to the margins of
the naso- and maxilloturbinates and lateral wall adjacent to
these margins. Consequently, it appears unlikely that polypoid
adenomas represent the benign counterpart of squamous cell
carcinomas.
A small number of other cancers were seen in the Kerns et
al. (1983) study. These included one adenocarcinoma, one poorly
differentiated carcinoma, one carcinosarcoma, and two poorly
differentiated carcinoma/sarcoma. The Carcinogenicity/
Histopathology/Genotoxicity Panel of the Consensus Workshop on
Formaldehyde (1984) stated that "[f]he polypoid adenomas can be
evaluated separately and in combination with the nonsquamous
carcinomas that were observed in the 14 ppm rats."
Since an adenocarcinoma and a morphologically similar
carcinoma were seen in the study, the polypoid adenomas may
represent the benign counterpart of these lesions. The PWG
stated that these lesions might arise de novo, originate from
submucosal glands, arise in polypoid adenomas, or a combination
of the above. Also, the PWG stated that "not enough information
4-21
-------
was available about nasal cavity tumors to predict the
possibility of benign tumors progressing to carcinomas."
However, a recent analysis by Swenberq and Boreiko (1985) states
that polypoid adenomas are likely to be the benign counterpart of
adenocarcinomas and may be more common in control animals than
previously thought. In the Kerns study, one polypoid adenoma was
present in the same section as an adenocarcinoma in the 15 ppm
exposure group. In contrast, no adenocarcinomas were found in
the 2 or 6 ppm exposure group, even though more polypoid adenomas
were found at each of these two exposure levels than at 15 ppm.
Even if polypoid adenomas are considered to be the benign
counterpart of adenocarcinomas, the conversion rate is low (a
conversion ratio of 1:15). As for the possibility that the
polypoid adenomas may be the benign counterpart of
carcinosarcomas, this seems unlikely due to different tissue
type. Added to this are the lack of dose-response, diagnostic
uncertainties (3 of the 12 tumors were borderline calls), and the
poor statistical significance of these lesions.
Finally, as discussed earlier, in the Albert et al. (1982)
and the Tobe et al. (1985) studies, papillomas rather than
polypoid adenomas were observed and in the Kerns et al. (1983)
study only polypoid adenomas were observed. This intraspecies
(and intrastrain since Tobe et al. and Kerns et al. used Fischer
344 rats) difference also adds to the uncertainty in using the
polypoid adenoma data for risk estimation purposes. Whether the
difference in benign tumors observed is due to a strain difference^
4-22
-------
is not clear. However, the separate appearance of two distinct
types of benign tumors further calls into question the significance
of these lesions regarding their ability to progress to squamous
cell carcinomas and their relevance in estimating human risk.
Because the nature and progression of benign nasal tumors is
not well understood, studies on other chemicals can be looked to
for elucidation.
Studies by Lee and Trochimowicz (1984), Takano et al (1982),
and Reznik et al. (1980) have examined the morphology of nasal
tumors in rats caused by exposure to hexamethylphosphoramide
(HMPA), 1,4-dinitrospiperazine (DtfP), and 1,2-dibromo-3-
chloropropane (DBCP), respectively. In the Reznik et al. (1980)
study on DBCP, 78% of the tumors in male and 66% in female F-344
rats in the low dose group were benign (adenomas and squamous-
cell papillomas). However, in the high dose group 89% and 76% of
the tumors in males and females, respectively, were malignant
(adenocarcinomas and squamous-cell carcinomas). It does not
appear that the shift from primarily benign tumors at the low
dose to primarily malignant tumors at the high dose means that
the benign tumors were progressing to their malignant
counterparts. Most of the benign tumors were located in the
anterior part of the nasal cavity, while most of the malignant
tumors were located in the region of the ethmoturbinates and the
posterior part of the nasal septum. Adenomas and adenocarcinomas
were often seen in rats at the same time and dose, but in
different parts of the nasal cavity.
4-23
-------
In the'Takano et al . (1982) study with DNP using F-344 rats!
5 different proliferative Lesions were seen: simple hyperplasia,
papillary hyperplasia, nodular hyperplasia, papilloma, and
carcinoma (mostly adenocarcinomas). Papillary hyperplasia and
papilloma were mainly located in the anterior regions of the
nasal cavity. Nodular hyperplasia and adenocarcinoma, on the
other hand, were found in the posterior regions. Also, these
pairs of lesions often coexisted in their respective locations.
The conclusion of the authors was that papillary hyperplasia
progresses to papilloma and that nodular hyperplasia progesses to
adenocarcinoma.
The studies by Lee and Trochimowicz (1982, 1984) using
Sprague-Dawley rats showed that HMPA caused mainly epidermoid
(squamous cell) carcinomas (71%), adenoid squamous carcinoma
(15%) and squamous cell papilloma (8.2%). (A small number of
adenomatous polyps were seen with adenoid squamous carcinomas.)
The squamous cell papillomas were mostly exophytic, which may
indicate that they may not represent the benign counterpart of
the epidermoid carcinomas for two reasons. First, in the Takano
et al. (1982) study/ nodular hyperplasias rather than papillomas
seemed to progress to adenocarcinomas. Second, the papillomas
and nodular hyperplasias are similar to human exophytic and
inverted papillomas, respectively. In humans, squamous
carcinomas apparently arise from inverted papillomas rather than
exophytic papillomas (Takano et al., 1982).
The experience with other chemicals (see Lee and
4-24
-------
Trochimowicz, 1982) and the foregoing illustrate the variability
of the types and locations of the tumors found. Also, except in
limited cases, the progression of preneoplastic and benign
neoplasms to malignant neoplasms is not known with any
assurance. Although some studies of chemicals show a tumor
profile that is predominantly benign at low doses and malignant
at high doses (MTP, 1982a; NTP, 1982b) other studies, such as of
HMPA, show primarily malignant tumors at all dose levels
eliciting a response. This may be the result of a speed-up of
the carcinogenic process at higher doses in the former case or as
in the latter one the chemical may be so 'potent that even at low
doses progression is completed before termination of the study.
Because there are so many uncertainties associated with the
polypoid adenoma data, it is recommended that (1) they not be
combined with squamous cell carcinomas (pooling) for statistical
purposes, and (2) risk estimates should be generated separately
using the polypoid adenoma data for analysis purposes (see
Sections 7.2 and 7.4).
4.3. Short-Tern Testa: Genotoxicity and Cell Transformation
HCHO affects genetic material in a wide range of test
systems (Auerbach et al. (1977); Ragan and Boreiko (1981):
Boreiko et al. (1982); Golmacher and Thilly (1983); Ulsamer et
al. (1984); Consensus Workshop on Formaldehyde (1984); Dooley et
al. (1985); Ma et al. (1985); Scott et al. (1985); Cantoni and
Cattabeni (1985), and Stankowski et al. (1986)). Mutajenic
activity of HCHO has been demonstrated in viruses, Escherichia
4-25
-------
coli, Pseudomonas pluonescens, Salmonella typhimur ium, and
certain strains of yeast, fungi, Pro soph ilia, grasshopper, and
mammalian cells (Ulsamer et al., 1984). HCHO's ability to cause
gene mutations, single strand breaks in DNA, DNA-protein cross-
links, sister chromatid exchanges (SCE), and chromo.some
aberrations has been demonstrated (Consensus Workshop on
Formaldehyde, 1984). In vitro studies have shown HCHO's ability
to transform BALB/c 3T3 mouse cells, BHK 21 hamster cells, and
C2H-10T1/2 mouse cells and to enhance the transformation of
Syrian hamster embryo cells by SA7 adenovirus, and to inhibit DNA
repair (Consensus Workshop on Formaldehyde, 1984). In a study by
Ragan and Boreiko (1981), treatment of C3H/10^/2 cells with HCHO
did not result in significant rates of transformation. However,
if HCHO treatment was followed by continuous treatment with the
tumor promoter 12-0-tetradecanoyl phorbol-13-acetate, significant
transformation occurred. HCHO also causes increases in the
frequencies of observed mutations in the presence of other
mutagens, such as X-rays, ultraviolet radiation, and hydrogen
peroxide. Compared to its effects on strains of E. coli and
Saccharomycea cereviaiae with normal repair mechanisms, HCHO
caused greater lethal and mutagenic effects in excision repair-
deficient strains (Ulsamer et al., 1984).
In reviewing much of the above literature, the Consensus
Workshop on Formaldehyde (1984) "found that the recent work is
more likely to find HCHO a mutagen than earlier studies, and is
also more likely to show a dose-response relationship. These
4-26
-------
results are-most probably attributable to the greater
sophistication in the way the later assays were carried out. It
should be noted that in the above studies, the relationship
between the cytotoxicity induced by HCHO and mutagenicity or
transformation induced by this agent is typical of most mutagens/
carcinogens that are positive in these assays. The data we have
reviewed are consistent with HCHO acting as a weak mutagen (i.e.,
less than a ten-fold increase over background)." In certain
bacterial tests it might be considered weak acting, but in a
recent NTP Drosophila sex-linked recessive lethal test, HCHO
would not be considered a weak acting mutagen (Woodruff et al.,
1985). In fact, in some cases HCHO is used as a test standard.
In vitro Cytogenetic studies have shown HCHO to be an efficient
inducer of sister chromatid exchanges (SCEs) and chromosomal
aberrations (Natarajan et al., 1983).
In a study by Grafstrom et al. (1983) using cultured
bronchial epithelial and fibroblastic cells, HCHO's ability to
cause the formation of cross-links between DNA and proteins,
cause single-strand breaks in DNA, and to inhibit the resealing
of single-strand breaks produced by ionizing radiation has been
shown. HCHO also inhibited the unscheduled DNA synthesis that
occurs after exposure to ultraviolet irradiation or to benzo-
[a]pyrene diolexepoxide, but at doses substantially higher than
those required to inhibit the resealing of X-ray induced single-
strand breaks, suggesting HCHO could exert its effects by both
damaging DNA and inhibiting DNA repair.
4-27
-------
As a follow-up to the above study, Grafstrom et al. (1984)
investigated the repair of DNA damage caused by HCHO in human
bronchial epithelial cells and fibroblasts, skin fibroblasts, and
DNA excision repair-deficient skin fibroblasts from donors with
xeroderma pigmentosum. Exposure of these cell types to HCHO
caused similar levels of DNA-protein cross-links and removal of
such cross-links in all cell types. The half-life for the cross-
links was about 2-3 hours. An examination of the induction and
repair of DNA single-strand breaks showed that the production of
the breaks was .dose dependent, and that there removal occurred at
rates similar to the removal of cross-links. In addition, the
results indicate that exposure to HCHO causes single-strand
breaks without the involvement of excision repair, and that
excision repair of HCHO damage may increase the single-strand
break frequency. HCHO also enhanced cytotoxicity of ionizing
radiation and N-methyl-n-nitrosaourea in normal bronchial
epithelial cells and fibroblasts. The authors speculated that
the inhibition of DNA-repair probably involves the interaction of
HCHO with cellular proteins of importance in DNA repair. They
noted that the repair of DNA lesions caused by ultraviolet
radiation has been shown to be inhibited by alkylating agents.
In a related study, Grafstrom et al. (1985) investigated the
effect of HCHO on the repair of 0^-methylguanine and the ability
of HCHO to potentiate the mutagenicity of N-methyl-N-nitrosourea
(NMU) in normal human fibroblasts. When rate of DNA repair was
measured for NMU-treated cells that were incubated with HCHO, a
4-28
-------
significantly lower rate of DNA repair was observed compared to
NMU-treated cells. The authors proposed that HCHO inhibits DMA
repair by binding to the active site of 0 -alkylguanine DMA
alkyltransferase. Also, although NMU and HCHO are weak mutagens,
addition of HCHO to NMU-treated cells resulted in a significantly
higher mutation frequency than was found with HCHO or NMU
alone. The increase may be due to HCHO's inhibiting O -methyl-
quanine repair (Grafstrom et al., 1985).
However, the results of a study by Snyder and Van Houten
(1986) question the finding of Grafstrom et al. (1983) that HCHO
inhibits UV-induced, unscheduled DMA synthesis. They found that
the inhibition only occurs when thymidine is used as a precursor,
which suggests an uptake artifact. Also, their results indicate
that HCHO has no significnat effect on the rate of repair of
x-ray-induced strand breaks or those by bleomycin. Consequently,
Snyder and Van Houten believe that it is likely that HCHO has no
significant effect on the sealing of most DMA breaks in human
fibroblasts; their work did not support a conclusion that the
ligation step of excision repair is preferentially sensitive to
HCHO.
In a study by Craft and Skapek (1986) using human
lymphoblasts, it was found that when these cells were exposed in
vitro to HCHO using single or multiple treatment regimens, a
difference in the cumulative induced mutant fraction was .
observed. Single treatment exposures (0-150 uM X 2 hr) resulted
in a nonlinear increase in induced mutant fraction. The multiple
4-29
-------
exposure experiment u^ing either three treatments of 50 uM X 2
hr, five treatments of 30 uM X 2 hr, or ten treatments of 15 uM X
2 hr, all treatments administered on alternative days, resulted
in additive (linear) increases in mutant fraction.
The multiple treatment regimen produced a lower rate of
mutations compared to an equivalent single dose; an induced
mutant fraction of 2.2 ± 0.2 X 10"6 for the five 30 uM exposures
vs. 4.8 ± 0.4 X 10~6 for the single 150 uM treatment.
A recent study by Casanova-Schmitz et al. (1984) has
reported the difference between metabolic incorporation and
covalent binding in the labelling of macromolecules in rat nasal
mucosa and bone marrow by inhaled [*4C]- and C^H] HCHO. Rats
were exposed to labelled HCHO at concentrations of 0.3, 2, 6, 10,
or 15 ppm for 6 hrs, one day following a single pre-exposure to
the same concentration of unlabelled HCHO. The principal finding
reported by the authors was the apparent nonlinearity in the
amount of covalent binding of HCHO to DNA of the respiratory
mucosa. The amount of HCHO covalently bound to mucosal DNA at 6
ppm was reported to be 10.5 times higher that at 2 ppm, whereas
covalent binding to protein increased in a linear manner with
increases in HCHO concentration. No covalent binding was noted
in tissues from the olfactory mucosa or bone marrow. The
apparent nonlinear covalent binding of DNA between 2 and 6 ppm
has been used as an input in quantitative estimation of risk from
HCHO exposure (Starr and Buck, 1984). Whether Casanova-Schmitz
et. al. measured covalent binding of HCHO to macromolecules has
4-30
-------
been questioned. See section 4-5.2 for a further discussion of
this issue.
Various studies have been undertaken to determine whether
HCHO has genotoxic effects in vivo. In mice, the dominant lethal
test was' found to be negative (doses up to 40 mg/kg, IP).
However, in a more recent dominant lethal assay using higher
doses (50 mg/kg, IP) and a different mouse strain specify,
marginally positive results were obtained, but only in the first
and third week of the seven weeks studied (Consensus Workshop on
Formaldehyde, 1984). However, the positive response obtained may
not be indicative of a mutagenic change for the following
reasons:
"1. The week to week variation in implantations data are
common in dominant lethal studies. Therefore,
concurrent controls should be included in each weekly
mating. It is not clear from the paper how the control
matings were conducted; the control value is shown as a
mean with no indication of the extend of weekly
variation in control population.
2. The index of implantation deaths should probably be
analyzed on the basis of "per pregnant female" and not
on the total numbers as done in the paper.
3. Preimplantation losses as shown in week 1 and 3 should
be viewed as an index of dominant lethal effect only if
the losses were found to be due to death of preimplants
and not due to failure of oocytes to become fertilized."
(IRMC Report on Systemic Effects, 1984)
Negative results were obtained when the induction of micronuclei
or chromosomal aberrations were used as an endpoint. A small
increase in .sister chromatid exchanges (SCE's) has been reported
in the bone marrow of mice exposed to high (>25 ppm) HCHO
concentrations. Unfortunately, technical problems were
4-31
-------
encountered during the HCHO exposures, and the actual
concentrations required to elicit this effect are not known
(Consensus Workshop on Formaldehyde, 1984). In a recent study by
Ward et al. (1984) measuring changes in sperm morphology, mice
were treated by gavage with five consecutive daily doses of
formalin (100 mg/kg; 10 animals). No increase in abnormal sperm
morphology was observed in the treated mice.
The possibility of genetic effects in humans caused by
inhalation of HCHO has been investigated by a number of persons.
In a study reported by Spear (1982), significant numbers of
SCE's in eight students exposed to HCHO during an anatomy
laboratory class were found. Mean HCHO levels were 1 ppm during
dissections. Mierouskiene' and Lekevicius (1985) have reported a,
statistically significant increase in chromosome aberrations in a
group of 50 workers exposed to phenol, styrene, and HCHO. The
control group consisted of 25 individuals which had no
occupational exposure to chemical substances. The finding of
increased chromosome aberrations was independent of age, exposure
length, and smoking habit. In a study by Bauchinger and Schmid
(1985) using lymphocytes from 20 males exposed to HCHO and
unexposed males employed by a paper factory, a significantly
increased incidence of dicentrics or dicentrics and ring
chromosomes was observed for 11 exposed workers employed as
supervisors. Their total mean exposure time was 2.5 times longer
than for the 9 exposed paper machine operators. SCE values were
not significantly different for smoking and nonsmoking HCHO
workers when compared with the control group.
4-32
-------
No genetic effects in humans were seen in studies by FLeig
et al. (1982), Ward et al. (1984), and Thomson et al. (1984). In
the study by Fleig et al. (1982), 15 employees exposed to HCHO
during HCHO manufacture and processing were studied. The
employees had 23 to 35 years of exposure. Mean HCHO levels did
not exceed 5 ppm before 1971 and 1 ppm after that date, with most
workers exposed to a maximum of 0.25 ppm (post 1971). No
increase in chromosome aberrations was observed as compared to
controls. Similarly, in a study of pathology staff exposed to
HCHO by Thomson et al. (1984), no difference in chromosome
aberrations induction and SCE frequencies was seen between the
exposed and control groups (6 exposed and 5 controls). Time-
weighted average levels of HCHO ranged from 1.14 to 6.93 mg/m^,
with peaks greater than 11.0 tng/m^. The pathology workers were
generally exposed to HCHO for 2-4 hours per day, 2-3 days per
week. In the Ward et al (1984) study, sperm count, morphology,
and fluorescent body frequency in 11 autopsy service workers
exposed to HCHO and 11 controls were evaluated. Time-weighted
average HCHO levels ranged from 0.61 to 1.32 ppm (weekly exposure
range 3-40 pm hours). No significant differences in the
endpoints studied were observed between exposed and control
groups.
Finally, Connor et al. (1985) examined the mutagenicity of
urine from HCHO-exposed autopsy service workers. An exposed
group of 19 and a control group of 20 were matched by sex, age,
and use of tabacco, alcohol, and recreational drugs. Urine
4-33
-------
samples were tested using 5. typhinurium TA100 and TA98 with and
without S9 activation. Except for a sample from a smoker in the
control group and three samples from an individual receiving
metronidazole therapy, most samples produced little'or no
increase in revertants using either strain. However, a
significant number of the samples from the exposed group were
toxic to TA100 and TA98. Similar findings have been reported for
aluminum workers (exposed and control) and in nursing
personnel. The toxicity from the three studies appear to be
identical, but the toxicant has not been identified (Connor et
al., 1985).
As noted above, the literature reports conflicting data
concerning chromosomal effects in humans. However, the weight of,
these data seems to indicate little potential for these effects
in the workplace, but this judgement must be tempered by the
limitations of the studies.
4-34
-------
4.4. Other Effects/Defense Mechanisms
4.4.1. Introduction
The cancer response observed in the Kerns et al. (1983)
study was very nonlinear, 1% of the rats responded at 5.6 ppm
while 50% responded at 14.3 ppm. A number of hypotheses have
been developed which attempt to explain this response and the
different responses seen in rats and mice in that study. These
hypotheses are based on the noncarcinogenic effects of HCHO.
Although these noncarcinogenic effects are not easily separated,
it is possible to discuss the nature of the effects themselves,
and how they may relate to results seen in long-term animal tests
of HCHO, by examining three subject areas: sensory irritation;
cell-proliferation; and the mechanics of the mucous layer
"defense" system.
4.4.2. Sensory Irritation
In the Kerns et al. (1983) study/ the response observed in
mice as compared to rats is markedly different, 2 mice responding
at 14.3 ppm versus 103 rats responding at this concentration.
Also, in studies using hamsters (Dalbey, 1982), no tumor response
was seen. One of the reasons given for the difference between
rats and mice is the observation that mice exposed to 14.3 ppm
reduce their breathing rate in response to the irritant
properties of HCHO. Such an effect may be occuring in hamsters
at the doses tested, but experimental evidence is lacking. How
reduction in breathing rates (which is an effective defense.
mechanism at certain concentrations) is weighed in terms of HCHO
cancer risk assessment is discussed below.
4-35
-------
It is well established that sensory irritants evoke
responses by stimulating the free nerve endings of the afferent
trigeminal nerve located in the corneal, nasal, and oral
mucosa. Besides burning sensations, sensory irritants cause a
number of physiological reflex responses, one of which is a
decrease in respiratory rate. A number of chemicals have been
studied in this regard and have RDSO's established for them. The
RD50 value is the concentration of an irritant that causes a 50%
reduction in respiratory rate. A proposal to use RDSO's to
establish concentration standards for human exposure to sensory
irritants has been made (see Kane et al., 1979; and Buckley et
al., 1984). A number of chemicals have been investigated and
RD50 values established, including HCHO and hydrogen chloride.
Consideration of this effect may be important in interpreting
inhalation bioassays because a doubling of a nominal
concentration to which an animal is exposed may not result in a
doubling of the dose actually received by the animal. For
instance, the RD50 value of HCHO for Swiss-Webster mice is 3.13
ppm. Consequently* results from a study using a dosing regimen
with concentrations above and below this number should be
interpreted in the light of the fact that the dose actually
received by the test animals does not increase in the same
proportion as the nominal concentration. Also, it should be
noted that respiratory rate suppression could change over the
course of a chronic study (Dallas et al. 1985).
4-36
-------
In the case of the Kerns study/ experimental data (Chanq et
al., 1981; 1983) indicate that mice exposed to 14.3 ppm HCHO
reduced their breathing rate to such an extent that an adjusted
exposure concentration would show the mice being dosed with
approximately the same amount of HCHO as rats at 5.6 ppm, where
the same cancer response was observed. If this factor is wholly
responsible for the difference in response between rats and
mouse, then adjusted doses can be used to calculate risks from
mouse data. Thus, it can be postulated that if mice could be
exposed to levels of HCHO that would approximate the amount rats
received at 14.3 ppm, then the response in mice would be similar.
The evidence indicates that mice are more sensitive or
better able to respond to the sensory effects of HCHO than rats,
and it may be this response which accounts for the different
carcinogenic response observed in rats and mice in the Kerns et
al. (1983) study. Adjusting dose levels for this response shows
that mice may be as sensitive as rats to the carcinogenic
potential of HCHO. Hamsters, on the other hand, appear to be
less sensitive to HCHO, although the response of hamsters and
rats to tobacco smoke is similar (as discussed in the section on
animal testa, two factors may account for the absence of an
observed effect in hamster; limited pathology work in the study,
and a low test dose.
4.4.3. Cell Proliferation, Cytotoxicity, and the Mucous Layer
Another important consequence of HCHO's irritant properties
is its effects on cell proliferation and damage it can cause to
4-37
-------
the mucociliary clearance system (resoiratory and olfactory
epithelium) of the nasal cavity. These effects have been cited
(Starr et al., 1984) as important factors in HCHO induced
carcinoaenicty from the standpoint of their impact on the
mucociliary clearance system/ as a orereouisite for HCHO induced
cancer, and in understanding the importance of concentration
versus cumulative exposure. These factors have an important
impact on the model chosen for quantitative risk assessment and
the weighing of noncarcinogenic effects as a cancer risk factor.
4.4.3.1. Cell Proliferation and Cytotoxicity
Studies by Swenberg et al. (1983) and Chang et al. (1983)
have reported the relationship between HCHO concentration and
cumulative exposure on cell turnover in the nasal cavity of rats
and mice. A diagram indicating the codina of the nasal cavities
of rats and mice for the test data discussed below is provided in
Figure 4-3.
Figure 4-3. Drawing indicating the level of sections from
the nasal passages of rats and mice. Figure taken from
Swenberg et al. 1983.
4-38
-------
In one test, rats and mice were exposed to 0, 0.5, 2, 6, and 15
ppm HGHO 6 hrs/day for 3 days, and then to 3H-thymidine 2 hours
after the end of exposure. As illustrated in Table 4-4,
increased cell proliferation as measured by increased
incorporation of labelled thymidine into cells was evident in
rats at 6 and 15 ppm and in mice at 15 ppm.
Table 4-4.
EFFECT OF FORMALDEHYDE EXPOSURE ON CELL PROLIFERATION
IN LEVEL B OP THE NASAL PASSAGES*
Exposure*
Control
0.5 ppm
2 ppm
6 ppm
15 ppm
% of Labelled Respiratory
Rat
0.22 _+ 0.03
0.38 _+ 0.05
0. 33 _f 0.06
5.40 +_ 0.82
2.83 jf 0.81
Epithelial Cells***
Mouse
0.12 _+ 0.02
0.09 _+ 0.04
0.08 _f 0.04
0.15 _+ 0.06
0.97 _+ 0.04
*Table taken from Swenberg et al. (1983).
**A11 animals exposed for 6 hrs/day for 3 days
***Mean + standard error.
When the labelled thymidine is administered 18 hours after
the last exposure, a greater increase in cell turnover is seen as
illustrated in Table 4-5. The increase in cell labelling may be
because 2 hours post exposure may not be the most sensitive time
for DNA synthesis due to initial inhibition by HCHO {Swenberg et
al., 1983).
4-39
-------
Table 4-5.
EFFECT OF THE TIME OP JH-THYMIDINE PULSE ON CELL
REPLICATION AFTER HCHO EXPOSURE TO RAT (LEVEL B)*
Post-Exposure
Time of Pulse
2 hours
18 hours
% Labelled Cells**
0/ppm 6 ppm***
0. 26 jf 0.03 1.22 + 0. 17
0.54 _f 0.06 3.07 +_ 1.09
*Table taken from Swenberg et al. (1983).
**Mean ± standard error.
***6 ppm, 6hr/day for three day.
To determine whether concentration is more important than
cumulative dose, a series of concentration time products were
tested. Each product equaled 36 ppm-hrs of exposure. The
results of this test, which appear in Table 4-6, indicate that,
at least for the effect measured, concentration has a greater
affect in level B of the rat nasal cavity.
Table 4-6.
EFFECT OF HCHO CONCENTRATION vs. CUMULATIVE
EXPOSURE ON CELL TURNOVER IN RATS (Level B)*
% Labelled Cells**
Exposure
Control
3
6
12
ppm
ppm
ppm
X
X
X
12 hrs
6 hrs
3 hrs
3 days +
0
1
3
9
.54 +
.73 _+
.07 _+
.00 _+
18 hrs
0
0
1
0
.03
.63
.09
.88
10 days +
0.
0.
0.
1.
26 _+
49 +.
53 jf
73 _+
18 hrs
0
0
0
0
.02
.19
.20
.65
*Table taken from Swenberg et al. (1983)
**Mean •»• standard error.
4-40
-------
However, the amount of labeling measured in the most
anterior region of the nasal cavity indicates the opposite. The
extent of labeling was essentially the same at each HCHO exposure
level. See Table 4-7 for details.
Table 4-7.
EFFECT OF HCHO CONCENTRATION vs. CUMULATIVE
EXPOSURE ON CELL TURNOVER IN RATS (Level A)*
% Labelled Cells
Exposure After 3 Days Exposure**
Control 3.00 jf 1.56
3 ppm X 12 hrs 16.99 _* 1.50
6 ppm X 6 hrs 15.46 _* 10.01
12 ppm X 3 hrs 16.49 _* 2.07
*Table taken from Swenberg et al. (1983)
**Mean _+ standard error.
Whether this difference in cell proliferation between levels
A and B is due to differences in mucociliary clearance in the
respective regions, to HCHO-laden mucous flowing from posterior
to anterior regions (Swenberg et al., 1983) or simply that the
"capture" capacity of Level A is exceeded which allows pass-by of
HCHO to L«v«l B and beyond, or some other reason, is unknown.
The data developed on mice regarding cell proliferation are
not as clear. In a test to measure differences between
concentration and cumulative exposure there was an inverse
response as illustrated in Table 4-8. Perhaps the ability of
mice to reduce their breathing rate at high HCHO concentrations
played a role.
4-41
-------
Table 4-8.
EFFECT OP HCHO CONCENTRATION vs. CUMULATIVE
EXPOSURE ON CELL TURNOVER IN NICE (Level A)*
% of Labelled Cells
Exposure After 10 Days Exposure**
Control 1.24 _* 0.57
3 ppm X 12 hrs 10. 14 jf 3.20
6 ppm X 6 hrs 4.72 +_ 1.61
12 ppm X 3 hrs 1. 76 _+ 0.49
*Table taken from Swenberg et al. (1983)
**Mean •*• standard error.
The difference between rats and mice has not been adequately
explained, except that there appears to be a significant species
difference regarding cell proliferation.
A study by Rusch et al. (1983) supports the concept that
concentration may be more important than cumulative exposure, at
least for rats. In the study, five qroups of 6 male Cynomoloqus
monkeys, 20 male and 20 female Fischer 344 rats, and 10 male and
10 female Syrian golden hamsters were exposed to 0, 0.2, 1.0, and
3.00 ppm for 22 hrs per day, 7 days per week for 26 weeks.
The most significant finding was squamous metaplasia/
hyperplasla in rats and monkeys at 3.0 ppm. Hamsters were not
affected at any dose level. However, the most significant
finding is that even though rats at 1.0 ppm in the Rusch et al.
(1983) study received a cumulative exposure 2.5 times greater
than rats at 2.0 ppm in the Kerns et al. (1983) study, which
experienced squamous metaplasia, they were largely free of
4-42
-------
squamous metaplasia. This strengthens the conclusion of the
Swenberg et al. (1983) study which indicates that, at least for
rats, concentration is more of a factor than cumulative exposure
for metaplasia.
In- the Kerns et al. (1982) study, significant levels of
noncarcinogenic lesions were noted in rats at all dose levels.
The severity and extent of these lesions were concentration and
time dependent and seem to be correlated with the cancer
response, i.e., these lesions preceded the appearance of squamous
cell carcinomas and their severity increased with increasing
cancer response. This observation, tied with the data showing
increases in cell proliferation due to HCHO exposure and a
threshold for squamous hyperplasia/metaplasia of between 1 and 2
ppm, leads some to the hypothesis that these effects are
important determinants in HCHO induced carcinogenicity and that
they help explain the nonlinearity of the cancer response.
Another factor suggested to contribute to the possibility of a
nonlinear response is the role of the mucous layer in trapping
and removing HCHO. This hypothesis is that when its removal
capacity is exceeded or its flow impeded, HCHO can then impact
the respiratory epithelium, thus causing the noncarcinogenic
effects noted above. A discussion of the role of the mucous
layer follows this section.
As noted previously, there was a 50 fold increase in cancer
response due to a slightly more than a doubling of the dose in
the Kerns study (5.6 to 14.3 ppm). What was the change in
4-43
-------
response of the noncarcinogenic effects? Using data developed
for the incidence of s-quamous metaplasia in rats in the Kern
study, a rough comparison can be made. The incidence of squamous
metaplasia in level 2 (level B in Figure 4-3) of the rat nasal
cavity was chosen because it showed a positive correlation with
concentration rather than cumulative dose, moreover/ it is in the
middle of the anterior part of the nasal cavity where the
squamous cell carcinomas were observed, and it is of the same
cell type as the carcinomas. If one compares the percentage
incidence of squamous metaplasia in the three dose groups at the
sacrifice points in Table 4-9, one sees a clear dose-response,
but not a 50-fold increase between 5.6 and 14.3 ppm; there
appears to be only a 2-fold increase or less. While increased
cell turnovers could lead to greater interaction of HCHO and
single-strand DNA, and thus an enhancement of the cancer
response, incidence of squamous metaplasia alone does not appear
to explain the extreme nonlinearity observed. A major limitation
of this comparison is that it does not account for the severity
and extent of the lesions which presumably increased at higher
concentrations.
t
4-44
-------
Table 4-9.
FREQUENCY OP SQUAMOUS METAPLASIA IN LEVEL 2
OP THE RAT NASAL CAVITY*
Incidence (Percent) of Squamous Metaplasia
Dose (ppm) Month of Sacrifice
6 12 18 24 27
5.6 • 50 45 60 65 30
14.3 75 90 98 100 100
'Estimated from Figure 4-1.
Other chemicals such as acetaldehyde and hexamethyphos-
phoramide (HMPA) are cytotoxic and cause cancer in rats. Data on
these chemicals may provide some insight.
If one examines the incidence and severity of the
noncarcinogenic lesions seen in the Woutersen (1985) acetaldehyde
study and the tumor response, one sees a roughly dose-related
response, i.e., a doubling of dose doubles the response seen
(cancer and noncancer). Although the olfactory epithelium was
severely affected at the highest dose/ the cancer response is
hardly increased over the next lower dose (see the section on
Structure Activity Relationships for a full discussion of the
data on acetaldehyde).
An anatomical region that had a high incidence of noncancer
lesions that was dose-related was the larnyx (mostly squamous
metaplasia). Table 4-10 illustrates this response. However,
only one tumor was observed in the larnyx.
4-45
-------
Table 4-10.
INCIDENCE OP LESIONS OTHER THAN TUMORS
IN THE LARNYX OF RATS EXPOSED TO
ACETALDEHYDE [NUMERIC]*
Dose (ppm)
0 750 1500 3000/1000
Number of Male Rats 50 50 51 47
Squamous metaplasia 3 6 23 41
Hyperkeratosis 1 4 13 32
*Data from Feron (1984)
The relationship between the noncancer and cancer response
seen in rats exposed to HMPA is unremarkable. As Table 4-11
indicates, an increase in dose did not lead to a many fold
increase in the cancer response although HMPA severely damaged
the nasal mucosa of the rats.
Table 4-11.
INDIGENCE OP EPIOERMOID AND ADENOID SQUAMOUS
CARCINOMAS IN RATS EXPOSED TO
HEXAMETHYLPHOSPHORAMIDE*
Dose (ppb) 0 10 50 100 400 4,000
No. of Rats
Examined 396 200 194 200 219 215
Tumor Incidence (%)
Epidermoid carcinoma 0 0 12.4 29.5 62.6 55.8
Adenoid squamous
carcinoma 0 0 2.1 2.5 9.6 19.1
*Data from Lee and Trochimowicz (1982)
4-46
-------
As a general matter, it appears that there is no clear
relationship between cell proliferation/cytoxicity and tumor
response. It is clear that there is much variation in the way
tissues respond to carcinogens, and no firm conclusions can be
drawn. .The appearance of noncancer lesions is not surprising
given the acute toxicity of many carcinogens. However, it is
impossible at this time to clearly link the noncancer effects in
the Kerns study to the appearance of cancers and the nonlinearity
of the response. On the other hand, it is plausible that the
noncancer effects may enhance the cancer response of HCHO and
other carcinogens by providing an increased opportunity for HCHO
to interact with single-strand DNA during cell replication or to
promote an initiated cell. Consequently, prudence would dictate
that situations which cause cell proliferation or lesions shouli
also be avoided. This includes short-term peaks especially if
cell proliferation and cytoxicity contribute to the carcinogenic
process. Also, it must be remembered that there is a natural
background rate of cell turnover in the nasal mucosa which can
provide the opportunity for mutagenic/carcinogenic events to
occur. Although such events may be rare, only one such sequence
of events may need to occur in a population of 10,000 persons
over 70 years to give a cancer risk of 1 X 10" .
Finally when discussing acute responses to a chemical such
as irritating effects, it should be remembered that there can be
a no-effect level in individuals at or below which no response is
observed no matter how many days of exposure occur. However,
4-47
-------
once a minimum effect concentration is reached, the duration of
the exposure may have a major impact on the severity of the
effect. Although the occurrence of squamous metaplasia and other
responses to acute effects may influence the expression of a
carcinogenic response, the absence of these acute responses does
not signify a no-effect level or the absence of a carcinogenic
response. For instance, the well-known carcinogen urethan causes
skin tumors (papillomas and squamous cell carcinomas) in mice,
but not epidermal hyperplasia or inflammatory reactions (Iversen,
1984).
4.4.3.2. Mucous Layer
Besides HCHO's effect on cell proliferation and respiratory
response, it also has a major impact on the mucociliary system o,fi
the nasal cavity through its ability to cause ciliastasis and
cell mortality at elevated concentrations. In addition, it has
been postulated that below certain HCHO concentrations (1-2 ppm)
the mucous layer can trap and remove much inhaled HCHO, thus
preventing it from reaching underlying cells. However, once the
mucous layer is saturated, HCHO can then begin to affect the
underlying cells as described in the section above. When this
occurs, the mucociliary clearance system is seriously compromised
which allows a greater amount of HCHO to reach the respiratory
epithelium. If the mucous layer removed most inhaled HCHO below
1 ppm then it would represent a threshold phenomenon at least for
the nasal cavity. However, the evidence for this is lacking.
The discussion below describes the effects caused by HCHO on the
mucociliary system and its role in protecting the nasal mucosa.
4-48
-------
The nasal cavity is primarily composed of ciliated
respiratory and olfactory epithelium which is covered by a moving
blanket of mucus. Mucus is composed of approximately 95% water,
0.5-li glycoproteins, and other minor constituents. The human
nose has three functions, two of which depend on the presence of
a mucous layer. The first function of the nose is to inform us
of the presence of noxious gases, if these stimulate the
receptors of the olfactory nerves. The second function of the
nose is to drain the secretions of the sinuses and of the
lacrimal (tear) glands. The third function of the nose is to
prepare the inhaled air for the lungs. This includes warming,
moistening, and filtering inspired air. Dust and many bacteria
found in the inspired air are impinged in the mucous that bathes
the mucous membrane and, by the action of the cilia of the nasal
passage, are moved outward (Tuttle et al., 1969).
As research by Morgan et al. (1983, 1984, 1986) indicates,
HCHO has a number of effects on the workings of the mucociliary
apparatus. Using in vitro and in vivo techniques, Morgan et al.
(1983a) examined mucous flow patterns in the rat nasal cavity and
the effect of HCHO on the mucociliary apparatus. Results of the
in vitro analysis indicate that mucus was present as a flowing
continuous coat over the respiratory epithelium except on the
most anteriorcentral extremity of the nasoturbinates and the
anteriomedial extremity of the maxilloturbinates. Mucous flow
rates ranged from 0.28 to 9.02 mm/minute. When rats were exposed
to 15 ppm HCHO, 6 hrs per day for 1, 2, 4, or 9 days, mucostasis
4-49
-------
accompanied by ciliastasis was evident in a number of anatomicaT
regions of the rat nasal cavity. In another study by Morgan et
al. (1986), male rats were exposed for 6 hours per day for 1, 2,
4, 9, or 14 days, to 0.5, 2, 6 or 15 ppm HCHO. There was a clear
dose-dependent affect on mucociliary activity. At 15 ppm there
was significant inhibition of mucociliary activity. Only slight
effects were noted in animals exposed to 2 or 6 ppm. At 0.5 ppm
no effects were observed. Finally, using frog palate, Morgan et
al. (1984) found that mucostasis, and ciliastasis occurred at
4.36 and 9.58 ppm, respectively. At 1.37 ppm an initial increase
in ciliary activity was observed but there was no mucostasis or
ciliastasis, while at 0.23 ppm there was no effect.
The above results indicate that a concentration relationship
exists where mucociliary flow would be impaired at 15 ppm and
less so at 6 and 2 ppm. This range corresponds to the range
where the steep dose-response in carcinogenic!ty of HCHO was seen
in the Kerns et al. (1983) study. Recent work by Bogdanffy et
al. (1985) demonstrated the ability of HCHO to bind with proteins
in human and rat nasal mucus and bovine serum albumin.
Incubation of HCHO in vitro with these materials indicated that
binding i3 rapid and that the main binding constituent in nasal
mucus is albumin. Consequently, some fraction of inhaled HCHO
would be expected to be bound and removed, thus protecting the
underlying epithelium.
Whether the mucous layer has some finite capacity to absorb,
retain, bind, and metabolize HCHO and wash it away to prevent it
4-50
-------
from reaching the underlying cells, or the response seen is
simply the overt signs of gradual cell toxicity, is unknown.
However, a number of factors can be considered when discussing
the protective ability of the mucous layer. First, humans can
detect HCHO at levels below 1 ppm which indicates that, at least
in the olfactory region of the nasal cavity, HCHO is not
completely removed by the mucous layer. The mucous layer is
reported to be immobile or flowing extremely slowly in this
region (CUT, 1984). However, it would seem that if a greatly
reduced removal capacity of the mucous layer in the olfactory
region played a role, this region should have been a target for
effects in the Kerns et al. (1983) study. This was not the
case. The significant neoplastic and nonneoplastic effects were
generally seen in the anterior regions of the rat nasal cavity.
Second, in a study by Casanova-Schmitz et al., 1984), which
measured the difference between metabolic incorporation and
covalent binding of labelled HCHO to macromolecules, it was found
that covalent binding to protein increased in a linear manner
with increases in airborne concentrations (0.3-15 ppm). However,
the finding is complicated by the fact that labelled
extracellular as well as intracellular protein was measured and
the fact that the overall results of this study have been
questioned. Consequently, the relative proportions of these two
constituents may not be able to be compared at each dose level.
A discussion of the formation of DNA-HCHO adducts as studied in
this experiment may be found in section 4.5.2.4.
4-51
-------
Finally, no data exist to show that other than a constant
proportion of HCHO reaches the respiratory epithelium at
concentrations below 2.0 ppm, levels that are generally not
acutely toxic to the underlying cells. At higher concentrations,
above the acute toxicity threshold, it is reasonable to expect
that higher, nonconstant proportions of HCHO reach the underlying
cells because of damage to and eventual destruction of the
mucociliary clearance system.
4.4.3.3. Conclusion
In conclusion, it is consistent with some of the data
described above to assume that HCHO's irritant and cytotoxic
properties may have contributed to the nonlinearity of the
malignant tumor response seen in the Kerns et al. (1983) study.
HCHO's demonstrated ability to increase cell turnover could
provide greater opportunity for HCHO to interact with nuclear
material. As the concentration of HCHO increases, greater cell
proliferation and cell death occur which provide even more
opportunities for HCHO-DNA interactions. To what degree the
mucous layer protects against HCHO's cytotoxic effects is not
clear, but the experimental data do .suggest that it does play a
role. Although data show that HCHO reacts with protein in the
mucous layer/ data have not been developed to show that the ratio
between the airborne concentration and the amount entering tarqet
cells is nonlinear. Regarding the impact of changes in
respiratory response to sensory irritants, it is likely that this
response is responsible for the different response of rats and
4-52
-------
mice in the Kerns et al. (1983) study. Also, other data
presented by Swenberg et al. (1983) indicate that rats only
inhaled twice as much HCHO at 15 ppm as they did at 5.6 ppm,
which indicates that the dose-response curve may be even steeper
when target tissue dose is plotted rather than concentration.
It seems likely that many of the factors outlined above have
contributed to the differences seen among species in their
response to HCHO as well as the steep dose-response seen in the
Kerns et al. (1983) study. However, an examination of the data
described in the sections above (1) does not support the concepts
that the action of the mucous layer presents a barrier to HCHO or
that it causes a nonlinear relationship between air
concentrations and the amount reaching target cells at levels
below overt acute toxicity, (2) that the appearance of and
severity of noncancer lesions can be used to predict the nature
of the cancer response, and (3) that the appearance of
noncancerous lesions is a necessary prerequisite for cancer
induction.
4.5. Metabolisa and Pharmacokinetica
4.5.1. Absorption
HCHO can enter the body as a result of inhalation,
ingestion, or dermal absorption. Absorption of HCHO through the
upper respiratory tract in dogs has been estimated to exceed 95%
of the inhaled dose (Egle, 1972). Nasal deposition in rats in
excess of 98% has been reported (Dallas et al., 1985). Studies
bv Heck et al. (1983) indicate that most of the radiolabel from
4-53
-------
radiolabelled HCHO inhaled by rats was found in tissues from thg-
anterior portion of the nasal cavity (the concentration was 10-
100 fold greater than other tissues). Radiolabel was found in
other tissues, but it is unlikely given HCHO's rapid metabolism
that the radiolabel found was HCHO. In another study by Heck et
al. (1982a), the quantity of HCHO was measured in rat tissue
before and after inhalation of HCHO or chloromethane. The
analytical method used cannot distinguish between free and bound
HCHO. Measured HCHO concentrations were as follows; 0.42 umol/g
for nasal mucosa, 0.097 umol/g for brain, and 0.20 umol/g for
liver. Inhalation of 6 ppm HCHO for 6 hrs/day for 10 days did
not significantly alter the nasal mucosa, brain, and liver HCHO
concentrations. A study by Bogdanffy et al. (1985) demonstrated
that nasal mucus reacts rapidly with HCHO, and suggests that the
main binding constituent in nasal mucus is albumin.
Following oral exposure of dogs to HCHO, formate levels in
the blood increased rapidly, indicating rapid uptake and
metabolism (Malorny et al., 1965). Dermal absorption has been
demonstrated in guinea pigs (Usdin and Arnold, 1979), but does
not appear to be significant in comparison to inhalation or
ingestion. Studies have also demonstrated the dermal absorption
of [ C] HCHO in rats and monkeys, and rabbits (Ulsamer et al.
1984). The chemical form of the radiolabel has not been
determined, but it has been reported by Ulsamer et al. (1984)
that data from in vitro diffusion studies using rabbit skin
indicate that free HCHO cannot be detected enzymatically.
4-54
-------
In a study by Bartnik et al. (1985), the percutaneous
absorption of HCHO was determined by applying [ C] HCHO-
containing cream (0.1% HCHO) to the backs of rats. Over 70% of
the radiolabel was found in the treated skin, 2.3% in urine, 1.8%
in the carcass, 0.7% in feces, and 1.3%. in C02 after 48 hours.
Thus total percutaneous absorption was 6.1% of the applied
dose. However, some fraction of this number may represent
methanol or formic acid which was present in the radiolabelled
sample (2 and 3%, respectively). Similar results were obtained
by Robbins et al. (1984) using rabbits to which 14C-labelled HCHO
was applied as a solution under an occluded patch.
Concentrations up to 37 ug HCHO per patch did not significantly
alter the proportion absorbed. When *4C-labelled DMDHEU
containing cloth was applied to the backs of rabbits under
various occlusion/perspiration conditions, only insignificant
amounts of the radiolabel penetrated the epidermis. Even under
the most severe test condition, only 2.5% of the radiolabel was
transferred from the cloth to the animal of which 80-90% was
found in the skin directly under the patch.
4.5.2. Pharmacokinetics
4.5.2.1. Conversion to formate
HCHO that enters the body appears to be converted rapidly to
formate and C02 (Malorny et al., 1965; McMartin et al., 1979) or
to combine with tissue constituents. The conversion of HCHO to
formate occurs following intravenous (i.v.) infusion,
subcutaneous injection, gastric intubation, or inhalation.
4-55
-------
Studies using i.v. infusion of 0.2M HCHO to dogs have shown
only a small amount of HCHO appears in the plasma during exposure
(Malorny et al., 1965). This becomes undetectable within 1 hour
after cessation of infusion. The peak formate concentration
following HCHO infusion was the same as when formate (0.2M)
itself was infused. The plasma half-life for formate (between 80
and 90 min.) was also similar. In the same study, HCHO could not
be detected after oral administration of 0.2M HCHO, although
formate increased rapidly in the plasma with a half-life of 81.5
minutes.
Similar experiments using Cynomolgus monkeys, in which 0.2M
HCHO was infused i.v., showed no accumulation of HCHO in blood
(McMartin et al., 1979). The blood half-life was estimated to be
1.5 minutes. Similar half-lives for blood HCHO have been
observed in rats, guinea pigs, rabbits, and cats (Rietbrock,
1969). Studies by Heck (1982b) have shown that [14C] formate and
[14C] HCHO have similar distribution patterns in rat blood cells
and plasma following i.v. injection, and follow the same decay
curve. In a somewhat different experiment, McMartin et al.
(1979) administered C14-labelled methanol by gastric
intubation. Again, HCHO could not be detected in the blood
although formate levels increased rapidly. A study in which
humans were exposed to HCHO gas (0.78 mg/m3) for 3 hours also
demonstrated a rapid rise in blood formate levels (Einbrodt et
al., 1976). Gottschling et al. (1984) studied a group of
veterinary medical students exposed to HCHO. An examiniation of
4-56
-------
pre- and post-exposure urine samples did not indicate a formic
acid shift. In a more recent study by Heck et al. (1985), HCHO
concentrations in the blood of rats and humans were
investigated. The rats (8 exposed and 8 controls) were exposed
to approximately 14.4 ppm HCHO for 2 hrs. Blood was collected
immediately after exposure and analyzed by gas chromatography/
mass spectrophotometry. No significant differences were seen
between exposed and control rats. Six human volunteers (4M, 2F)
were exposed to 1.9 ppm of HCHO for 40 minutes. Venous blood was
analyzed for HCHO levels before and after exposure. There was
not a statistically significant effect of exposure on the average
HCHO blood concentrations of the volunteers. However,
significant differences were seen in some of the subjects'
(either decrease or increase) HCHO concentration between blood
taken before and after exposure.
The rapid conversion of. HCHO to formate occurred in many
tissues in the various species examined, including human
erythrocytes (Malorney et al., 1965), liver and brain; sheep
liver; rat brain, kidney, and muscle, rabbit brain; and bovine
brain and adrenals (Uotila and Koivusalo, 1974). The enzymes
involved have been studied by Strittmatter and Ball (1975) as
well by Uotila and Koivusalo (1974). The oxidative process is
initiated by formation of S-formyl glutathione, which is then
oxidized by NAD and finally cleaved by thiol esterase, releasing
formic acid and glutathione. HCHO also has been reported to be
4-57
-------
oxidized to formic acid by a nonspecific aldehydehydrogenase ana
through the tetrahydrofolic acid pathway (Huennekens and Osborn,
1959).
4.5.2.2. Reaction with Glutathione
The intracellular level of glutathione (GSH) affects the
metabolism and toxicity of HCHO. In a study by Ku and Billings
(1984), the relationship between HCHO metabolism and toxicity and
GSH concentration in isolated rat hepatocytes was investigated.
When hepatocytes were pretreated with diethyl maleate (DEM) to
deplete GSH, the initial rate of HCHO disappearance was decreased
approximately 50%. The concentration of HCHO used (5.0 mM) was
not toxic to OEM-treated or untreated cells. Thus cell viability
was not a factor. HCHO was also shown to decrease GSH
intracellular concentration in a dose and time-dependent
manner. DEM treatment followed by HCHO addition caused a similar
reduction in GSH concentrations even though DEM pretreatment
resulted in varying GSH concentrations. In studies measuring the
cell toxicity of HCHO, it was found that DEM pretreatment greatly
decreased cell viability, whereas HCHO treatment alone did not
after 60 minutes. When DEM was added to incubations containing
HCHO, toxicity was increased at all incubation times. Only .at
120 minutes was there substantial toxicity in the HCHO only
treatment group. DEM/HCHO treatment increased lipid neroxidation
at HCHO concentrations which decreased cell viability in other
studies. DEM alone had no affect on lipid peroxidation. To
determine if the enhanced toxicity is due to DEM treatment and
4-58
-------
consequent reductions in GSH concentrations, L-methionine was
added to DEM/HCHO treated cells. L-methionine treatment reduced
reduction in GSH concentration and prevented HCHO stiumulation of
lipid peroxidation and loss of cell viability. Thus, enhanced
toxicity can be traced to the effects of DEM on GSH
concentrations. As a further check on this hypothesis, the
addition of free radical scavengers, ascorbate, BHT, and
-tocopherol protected the cells from HCHO induced toxicity in
DEM-pretreated cells. On the contrary, the addition of
scavengers had no effect on HCHO-induced toxicity in the absence
of DEM treatment, which suggests a non-free-radical mode of
toxicity.
In studies using isolated, perfused rat lungs and livers,
Ayres et al. (1985) reported dose-related reductions in GSH
concentrations. However, the concentrations needed to
significantly reduce GSH levels are many times higher than those
expected in the environment. Studies by Heck et al. (1980) and
Casanova-Schmitz et al.(1984) indicate that carcinogenic
concentrations of HCHO in the rat (15 ppm) did not reduce
nonprotein sulfhydryl levels in rat nasal tissue or produce
plasma HCHO levels approaching the lowest HCHO concentration
causing GSH depletion in isolated lung/ liver. Thus, GSH
depletion does not appear to be a critical factor in HCHO-related
toxicity (Ayres et al., 1985).
The role of GSH depletion on the formation of DMA-protein
cross-links (DPX) has been reported by Casanova-Schmitz and Heck
4-59
-------
(1984, 1985). It was found that when GSH levels were depleted by
phorone/ increases in interfacial (IF) DNA (a measure of DPX) and
the concentration of DPX in isolated DNA were observed. The
yield of DPX was reported to increase nonlinearly with HCHO
concentration for both normal and GSH depleted rats. In
addition, DPX yields were greater at low HCHO concentrations in
GSH depleted rats indicating that the metabolism of HCHO at low
concentrations may be more effective (the significance of the
reported nonlinearity in the formation of DPX is discussed in
section 4.5.2.4). However, a study by Lam et al. (1985) reported
that the concentration of nonprotein sulfhydryls in rat nasal
mucosa was not significantly reduced at 6 or 15 ppm HCHO, which
indicates that HCHO's toxic effects at these levels are not
related to GSH depletion.
4.5.2.3. Conversion to C02 and other metabolites
Additional studies (DuVigneaud et al., 1950) have shown that
following subcutaneous administration of 14C-HCHO to rats,
approximately 81% of the radioactivity was found in choline.
Almost 60% of a subcutaneous dose of 14C-formate appeared as
*4C02» with small amounts of radioactivity in choline. Neely
(1964) administered radiolabelled HCHO intraperitoneally (i.p.)
to rats and found that 82% of the radiolabel was recovered as C02
and 13-14% as urinary methionine, serine, and a cysteine
adduct. At lower doses, only radiolabelled methionine was
formed. The author postulated that C02 was derived from serine
(formed from glycine and N5,N10 methylene tetrahydrofolate) by
4-60
-------
deamination to pyruvate and oxidation in the Krebs cycle. In a
study by Mashford and-Jones (1982), it was found that in rats
administered 4 mg/kg of radiolabelled HCHO, most was exhaled
within 48 hrs as C02; 5.5% was found in the urine. At a dose 10
times higher (40 mg/kg), 78% was exhaled as C02 after 48 hrs,
while 11% was found in the urine. When HCHO was administered to
rats by inhalation, 40% of the radiolabel was found in tissues,
40% was exhaled, and 20% appeared in the urine (Heck, 1982b). It
was found by Heck (1983), chat the greatest amount of radiolabel
in the rat nasal mucosa was found in RNA, with a lesser amount in
protein and a small amount in DNA.
In a study of the disposition of HCHO in mice, Billings et
al. (1984) found that 70-75% of i.p. injections of 6 mg/kg or 100
mg/kg of 14C-HCHO was exhaled as C02 within 4 hours, with an
additional 10% exhaled at the end of 24 hours. When the rates of
C02 excretion between mice dosed with 100 ug/kg HCHO or 100 mq/kg
formate were examined, it was found that the rate of C02
excretion in mice given HCHO was slower than the formate-dosed
mice. Since formate is an intermediate in HCHO oxidation, the
authors speculated that HCHO might accumulate in tissues.
However, subsequent testing did not bear this out. Robbins et
al. (1984), using rabbits found the following distribution of
radioactivity 48 hours after intravenous injection of ^4C-HCHO:
blood 1.58%; skin/muscle/organs 3.26%; urine 4.14%; and 37.03% in
co2
4-61
-------
The formation of methionine from C-HCHO and homocysteine
had previously been demonstrated by Berg (1951). Formation of
methionine would also account for the labelled choline observed
by Duvigneaud et al. (1950) via methylation of phosphatidyl
ethanolamine. More recent work by Pruett et al. (1980) has
demonstrated the incorporation of *4C-HCHO into the nucleic acid
and protein fractions of WI 38 human diploid fibroblasts. Most
of the radiolabel was found in RNA with lesser amounts in DNA and
protein. The purine bases of both DNA and RNA were most heavily
labelled.
In addition to the serine pathway to CO, postulated above
(Neely, 1964), two other pathways have been identified, and are
diagrammed in Figure 4-4.
Figure 4-4. Simplified reaction sequence from drug
N-demethylation (cytochrome-P-450-dependent monooxygenase
to HCHO, formate, and CO, production (from Waydhas et al.
1978). Reactions are: la, HCHO dehydrogenase (GSH);
Ib, aldehyde dehydrogenase; Ic, catalase (peroxidatic
mode)? 2a, 10-formyltetrahydrofolate synthetase; 2b,
10-formyltetrahydrofolate dehydrogenase; 2c, catalase
(peroxidatic mode).
4-62
-------
Waydhas et al. (1978), McMartin et al. (1977), and Palese
and Tephyl (1975) have demonstrated that the catalase reaction
(Figure 4-4) is not of major importance and that the primary
pathway to CC^ from formate occurs via the tetrahydrofolic acid
pathway. This has been demonstrated in rat liver perfusates
(Waydhas et al., 1978) monkeys (McMartin et al., 1977), and rats
(Palese and Tephly, 1975). Since the tetrahydrofolic acid
pathway (Figure 4-5, from Kitchens et al., 1976) can lead to the
transfer of the carbon from formate to a number of other
compounds (including serine), it is not clear that the
10-formyltetrahydrofolate dehydrogenase reaction (Figure 4-4) is
the only reaction of importance for C02 production in this
pathway.
4-63
-------
Glycln*
I
ATP
V
ineslnlc
acid and
pur In«*
ATP T ^
S«rln«
Glyclns
NAOPH HfOf
n5-'°FM
Xl
FM2
ThymidylIc
HemecystAln*
VIT. a
12
M«m I on I n«
and
tcld
«<4 • N>0«for«yir«rrahy4retelle acid
H4 • N**f«rvrir«fr«nr«ro*olle acid
4»N',Mt0-«»th«rtylt«fr«ftydro*olle acid
• N' •fernlalnetctrahydrefelle acid
• w'.N10 -«n«tnyl«n«T«tranydrofoNc acid
Figure 4-5. Tetrahydrofolic acid pathway and 1-carDon
transfer for HCHO raetaoolism.
4-64
-------
4.5.2.4. Reactions with Macromolecules
Besides being converted rapidly to C02 and formate, and
being incorporated into other chemicals, HCHO can alkylate
macromo lecules such as amino acids, proteins, nucleotides, RNA
and DNA- (LJlsamer et al., 1984; Casanova-Schmi tz and Heck, 1983,
1984; Consensus Workshop on Formaldehyde, 1984; Mizenina et al.,
1984; Solomon and Varshavsky, 1985; Schouten, 1985; Foekens,
1985).
A recent study by Casanova-Schmitz et al. (1984) has
reported the difference between metabolic incorporatin and
covalent binding in the labelling of macromolecules in rat nasal
mucosa and bone marrow by inhaled [*4C]- and [^Hj HCHO. Rats
were exposed to labelled HCHO at concentrations of 0.3, 2, 6, 10,
or 15 ppm for 6 hrs, one day following a single pre-exposure to
the same concentration of unlabelled HCHO. The difference
between metabolic incorporation and covalent binding was
determined by the use of a phenol extraction procedure. This
procedure allows the separation of macromolecules into phases
after centrifugation (aqueous (AQ), organic, and interfacial (IF)
phases). DNA can be recovered from both the AQ and IF phases.
The convalently bound DNA is recovered from the IF phase. In.
this way, the authors claim that the relationship between
metabolically labelled DNA and cross-linked DNA can be
determined. The principal finding reported by the authors was
the apparent nonlinearity in the amount of covalent binding of
HCHO to DNA of the respiratory mucosa. The amount of HCHO
4-65
-------
covalently bound to mucosal DNA at 6 opm was reported to be 10.5
times higher that at 2 ppm, whereas covalent binding to protein
(intra- and extracellular) increased, in a linear manner with
increases in HCHO concentration. No covalent binding was noted
in tissues from the olfactory mucosa or bone marrow. The
apparent nonlinear covalent binding of DNA between 2 and 6 ppm
has been used as an input in quantitative estimation of risk from
HCHO exposure (Starr and Buck, 1984).
The Casanova-Schmitz study has been reviewed by Cohn et al.,
1985, EPA, and the Science Advisory Board (SAB) (1985). Its
implications for quantitative cancer risk assessment have also been
addressed. Cohn et al. came to the conclusion that the data were
interesting but preliminary in nature and thus not useable as inpta
to quantitative risk assessment. The SAB is composed of a group of
non-EPA scientists who advise the Administrator of EPA regarding the
scientific adequacy of agency risk assessments, testing and
assessment guidelines, research proposals, etc. The Agency agreed
with the SAB to seek a review of the study by an independent group
of scientists. The group's report to EPA (Report No. TR-835-20,
Expert Review of Pharmacokinetic Data: Formaldehyde) is provided in
Appendix 1 and is summarized by the group as follows:
1. Some doubt still remains as to the validity of the
assumptions which form the basis for distinguishing
metabolically incorporated and crosslinked (or adducted)
CH20, i.e., ^H/14C in DNA.
2. Experimental methods and controls were adequate with
respect to monitoring the C^O administration and
analysis of dual-labeled materials. However, the
chloroform/iso-amylalcohol/phenol extraction for DNA and
DNA crosslinked to proteins was not validated in terms
4-66
-------
of the identities of materials separated nor the overall
efficiency and consistency of extraction. The
occurrence of underlying variability of incorporation
due to kinetic isotope effects on the disposition of
tritiated CH20 can neither be assessed nor discounted.
3. Sufficient documentation is still unavailable to state
unequivocably that all the crosslinked DNA-protein
complexes occur in the IF-DNA fraction.
4. There remains a need for an effective biochemical
dosimeter to measure the dose of CF^O delivered to the
cells of the nasal epithelium. The data provided by
Casanova-Schmitz et al. are not considered a
sufficiently well-validated measure of this parameter.
5. The nonprooortionality of the calculated concentration
of bound i^C (CH20)-DNA as a function of the
administered dose is documented adequately. Whether the
nonproportionality truly reflects crosslink formation or
is due to the small sample size, to a constant loss in
the recovery of IF-DNA, or to artifactual disturbances
in the JH/iqC ratio remains to be elucidated.
6. The increase in concentration of bound C with the
concentration of CH^O is well documented, as is the
increase in the difference in the -3H/-1>4C ratio between
IF-and AQ-DNA. The power of separate comparisons for
the 0.3 and 2 ppm doses is low because of small sample
size relative to the coefficient of variation. This
limits the potential for inferences about no-response
levels and low-dose extrapolations.
7. The study of Casanova-Schmitz et al. is an important
first step toward quantitative assessment of the
intracellular level of Cf^O in the nasal mucosa of the
rat following inhalation exposure. At its present level
of validation, however, it does not provide a basis for
such quantitation. Furthermore, the selection of an
acute study model may not be appropriate to the
assessment of chronic toxicity.
In response to the report, the Chemical Industry Institute
of Toxicology (CUT) submitted detailed comments which strongly
disagree with the expert groups conclusions. CIIT's comments are
provided in Appendix 1. As with many emerging areas of
investigation there are bound to be disagreements among
4-67
-------
scientists. This is one of those cases. Additional work
underway at CUT using primates may resolve areas of
disagreement. Until the issues raised concerning the Casanova-
Schmitz study are resolved/ this study will not be used as a
basis for an alternate measure of HCHO exposure.
4.5.2.5. Endogenous HCHO
Endogenous HCHO is primarily produced from the degradation
of serine with some contribution from the degradation of other
amino acids. Oxidative demethylation of N,N-dimethylglycine
(from choline degradation) also contributes significantly to
endogenous HCHO. HCHO is also produced from a wide range of
xenobiotics (Dahl and Hadley, 1983). Cytochrome P-450-dependent
N-demethylation of drugs can contribute HCHO. Other xenobiotics
including dihalomethanes, methanol, dimethylnitrosamine,
hexamethylphosphoramide (HMPA), bis(chloromethyl) ether (BCME),
dibromoethane, and dimethylsulfoxide lead to the production of
HCHO. HCHO is also formed in vitro in the presence of an amine
acceptor, apparently by nonenzymatic breakdown of
N ,N* -methylene-tetrahydrofolate. This reaction produces
alkaloids from biogenic amines or drugs in vitro and probably in
vivo. The role of HCHO in xenobiotic transformation has also
been studied (Kucharczyk et al., 1984).
Whereas the conversion of HCHO to C02 occurs in a similar
manner in the different species studied, the relative importance
of each reaction differs among species and tissues. Thus, the
rat is able to convert formate to C02 at more than twice the rate
4-68
-------
of monkeys (or humans) and, as a result, has lower blood formate
levels (McMartin et al./ 1977) and does not excrete formate in
the urine (Neely, 1964). Man additionally possesses 50% more
hepatic dehydrogenase than do rats (Goodman and Tephly, 1971).
Den Enge'lse et al. (1975) have shown that mouse (C3Hf/A) and
hamster (Syrian golden) lungs do not convert formate to CC>2 as
efficiently as liver tissue does.
4.5.3. Summary
In summary, free HCHO is not usually found in plasma or
other body tissues in measurable quantities (this may be a
function of the analytic technique and not necessarily the
absence of free HCHO), endogenous HCHO that is produced may be
reasonably presumed to be metabolized rapidly to formate or to
enter the one-carbon pool. When exogenous exposure occurs, HCHO
is likewise rapidly metabolized to formate and excreted,
converted to C02 and/or incorporated into other molecules. The
same pathways seem to occur in all mammalian species examined to
date, but reaction rates differ among various species and
tissues. Neither the ratio of metabolic deactivation to binding
(to tissue or small molecules) nor the effect of route of
exposure on this ratio is known with assurance at this time.
However, Casanova-Schmitz et al. (1984) have made an important
contribution in this area. Eqle's work (1972) suggests that the
respiratory tract tissues would receive the greatest dose.
Although effects at other body sites cannot be ruled out, the
weight of the evidence indicates that effects at sites distant
from the area of exposure would not be expected.
4-69
-------
The overall metabolism of HCHO is summarized in Figure 4-6
!adopted from Kitchens et al., 1976):
?rot«ia« tad Kucltic Acids
Hucitic Acids
Labii* a»thyl group*
and on« carbon oatibolin
Urina 44 Sodiua S*lt
Figures 4-6. Overall metabolism of HCHO (from Kitchens
et al., 1976).
As can be seen from Figure 4-6, HCHO is more chemically
active than any of its direct metabolites and would, therefore,
appear to be the chemical substance of most concern for
carcinogenicity. The possibility exists, that the actual
carcinogenic agent may be an amino HCHO-acid (or other) adduct
(EPA, 1981).
4.6. Structure-Activity Relationships
HCHO is structurally similar to other aldehydes such as
acetaldehyde, malondialdehyde and glycidaldehyde. These
aldehydes have been shown to have oncogenic activity in
laboratory animals. For instance, inhalation of acetaldehyde has
produced tumors of the nose and larynx in hamsters and tumors of
4-70
-------
the nose in rats, and glycidaldehyde has produced skin tumors in
mice in skin painting tests. Since acetaldehyde is the closest
in structure to HCHO, and its effects on animals have been
compared in a previous section, the significant studies related
to its oncogenic potential will be described.
Acetaldehyde was not mutagenic in the standard Ames test
using Salmonella typhimurium (Commoner, 1976) and Escherichia
coli WP2uvrA (Hemminki et al., 1980). It had weak mutagenic
activity in the fruit fly Drosophila melanogaster (Rapoport,
1948). The potential of acetaldehyde to damage chromosomes has
been indicated by the dose-dependent sister chromatid exchanges
in the Chinese hamster ovary cells (Obe and Ristow, 1977) and
human lymphocyte cells (Ristow and Obe, 1978).
The carcinogenic effects of the inhalation of acetaldehyde
vapor were studied in hamsters by Feron (1979). A group of 210
male hamsters, which were further divided into six subgroups of
35 each, were exposed to 1500 ppm acetaldehyde vapor (7 hr/day, 5
days/wk) alone or simultaneously with benzo(a)pyrene (BP) as a
weekly intratracheal injection for 52 weeks. The weekly
concentrations of BP used were 0.0625, 0.125, 0.25, 0.5, and 1.0
mg/animal. The maximum dose of BP administered throughout the
entire experiment was 52 mg/animal. A group of 210 control
animals were exposed to air alone or simultaneously with the same
concentrations of BP. At the end of the treatment period, 5
randomly selected animals from each group were killed and
autopsied. All remaining animals were allowed to recover for 20
weeks and sacrificed by week 72.
4-71
-------
Exposure of hamsters to 1500 ppm acetaldehyde vapor produced
abnormalities in the respiratory tract which were characterized
primarily by reversible hyperplastic/ metaplastic, and
inflammatory changes. Neoplastic alterations attributable to
acetaldehyde exposure alone were not found. Intratracheal
instillation of the highest dose of BP (52 mg, 1 mg/wk for 52
weeks) combined with inhalation of acetaldehyde produced twice as
many tracheal tumors (squamous cell carcinoma and squamous
adenocarcinoma) and a shorter latent period as intratracheal
instillation of BP alone. However, such a synergistic effect of
acetaldehyde was not noticeable at any of the lower BP levels.
No significant differences in the number of tumors in the larynx,
bronchi, bronchioles, or alveoli were found among the different
treatment groups.
In a separate experiment, groups of 35 male and female
hamsters were treated intratracheally with acetaldehyde for a
period of 52 weeks. The intratracheal instillations were given
either weekly or fortnightly with acetaldehyde (2% and 4%) alone
or in the presence of either BP (0.25% and 0.5%) or
diethylnitrosamine (DEN, 0.5%), two proven carcinogens. Interim
sacrifices of 3 animals/sex/group were performed after 13, 26,
and 52 weeks. All remaining animals were sacrificed after 104
weeks.
Intratracheal administration of acetaldehyde at both dose
levels caused severe hyperplastic and inflammatory changes in the
bronchioalveolar region of the respiratory tract; however, only
4-72
-------
one case of pulmonary adenoma was found out of 134 animals
treated with acetaldehyde alone. This is not considered to be an
indication of carcinogenic activity of acetaldehyde. Despite the
high degree and frequency of peribronchiolar adenomatoid lesions
found following intratracheal instillation of acetaldehyde, the
neoplastic response of the bronchioalveolar tissues was clearly
lower in animals treated with BP plus acetaldehyde than in those
given BP alone. Thus, acetaldehyde inhibited the effect of BP.
Similarly, the carcinogenic effect of DEN was also not influenced
by the treatment with acetaldehyde.
In another study, Feron et al. (1982), studied respiratory
tract tumors in male and female hamsters exposed to high
concentrations of acetaldehyde vapor alone or simultaneously with
either benzo(a)pyrene (BP) or diethylnitrosamine (DEN) were
studied. The animals were exposed 7 hrs/day, 5 days/wk for 52
weeks to an average concentration of acetaldehyde of 2500 ppm
during the first 9 weeks; 2250 ppm during weeks 10-20; 2000 ppm
during weeks 21-29; 1800 ppm during weeks 30-44; and 1650 ppm
during weeks 45-52. Animals exposed to air or air plus BP or DEN
served as controls. Following the 52-week treatment period,
there was a 29-week recovery period after which all hamsters were
killed for autopsy. All remaining animals were sacrificed after
31 weeks.
At the end of the exposure period, (i.e., at week 52)
distinct histopathological changes, similar to those of the
previous studies, were found in the nose, trachea, and larynges
of animals exposed to acetaldehyde. No tumors were found in
hamsters killed immediately at the end of the exposure period.
4-73
-------
Acetaldehyde-exposed animals which were found dead or
sacrificed at week 81 exhibited inflammatory, hyperplastic, and
metaplastic changes in the nose and larynx, suggesting a
persistence of those alterations. Tumors were encountered in the
nose (adenoma, adenocarcinoma, anaplastic carcinoma) and the
larynx (papilloma, carcinoma in situ, squamous cell carcinoma,
adeno-squamous carcinoma); animals exposed to acetaldehyde plus
BP or DEN also exhibited tumors of the trachea and the lung. The
neoplastic and nonneoplastic lesions in the larynx were mainly
located either on the true vocal folds or in the most anterior
part of the larynx. None of the animals exposed to air alone
demonstrated nasal or laryngeal tumors nor atypical laryngeal
hyperplasia and metaplasia. The incidence of nasal and laryngeaj
tumors in hamsters exposed to acetaldehyde and treated with
either BP or DEN was similar to that found in hamsters exposed to
acetaldehyde alone. Carcinomas in situ and squamous cell
carcinomas of the larynges were found after combined treatment,
but were not observed after treatment with either BP or DEN
alone. Based upon these findings, the authors concluded that
"acetaldehyde is an irritant as well as a carcinogen to the nose
and larynx with a weak initiating and a strong 'promoting'
(cocarcinogenic) activity" (sic).
Finally, in a study by Woutersen et al. (1985) male and
female Wistar rats were exposed to 0, 750, 1500, and 3000/1000
ppm of acetaldehyde for 6 hrs/day, 5 days/week for 27 months.
There were significant nonneoplastic lesions of the olfactory
4-74
-------
epithelium at each exposure level. In contrast, significant
nonneoplastic lesions were seen in the respiratory epithelium
only at the highest dose. Statistically significant numbers of
adenocarcinomas were observed at each dose level in males and
females. Squamous cell carcinomas were observed at the two
highest dose levels in males and at the highest dose level in
females. Most of the tumors originated from the olfactory
epithelium. Table 4-12 presents summary tumor response of the
nasal cavity for this study.
4-75
-------
Table 4-12.
Nasal and Larynqeal Cancer in Rats Treated with
Acetaldehyde by Inhalation for 27 Months
(Woutersen et al. 1985)
Site and Tumor Type
Incidence a
Males
Acetaldehyde (ppm)
U 750 1500 3000
Females
Acetaldehyde (ppm)
0 750 1500 3000
Nose
Papilloma
Adenoca rc i noma
Carcinoma in situ
Squamous cell
carcinoma
0/49 0/52
0/49 16/50b
0/49 0/49
1/49
Metastasizing 0/49
squamous cell carcinoma
1/49
0/49
0/53 0/49 0/50
30/53b 20/49b 0/50
0/53 1/49 0/50
10/53b l4/49b 0/50
0/53 1/49 0/50
1/48 0/53 0/53
6/48c 26/53b 20/53b
0/48 3/53 5/53
0/48 5/53 17/53b
0/48 0/53 0/b3
Larynx
Carcinoma in situ
0/49 0/49
0/53
0/49
0/50
0/48 1/53
0/53
a Incidence is expressed as the number of animals with tumors over the number
of animals examined.
|0.0
-------
4.7. Epideaiologic Studies Reviewed
4.7.1. Introduction
The EPA has reviewed the available cohort and case-control
studies related to formaldehyde. Many of these studies have been
released within the past three years. Only one study is
currently ongoing which relates to evaluating human risks
associated with formaldehyde exposure a case-control study of
nasal cancers by the Centers for Disease Control. Only cohort
and case-control designed studies were analyzed for this review
since they yield the best quality of information for judging
causality. Table 4-13 identifies these studies. Although these
studies are of optimal designs for evaluative purposes, many
studies suffer from limitations that can potentially influence
their conclusions. Major drawbacks are: (1) the inference of
formaldehyde exposure levels from industrial hygiene data;
(2) the inability to completely separate the contributions of
HCHO from the contributions of other occupational or personal
exposures; (3) small sample sizes for the cohort studies;
(4) small numbers of observed site-specific deaths: and
(5) insufficient follow-up.
One outcome of the design limitations is low statistical
power in each study to detect small relative risks for rare for.r.s
of cancer.* The ability of a well-conducted study to detect -an
increased risk depends upon sample size, years of follow-up,
*The power of a study is the ability to detect true
association of the exposure and disease. If a study is
likely to conclude that the exposure is not associated with
a disease, when in fact an association existed, it has a low
power for detecting that association.
4-77
-------
Table 4-13
auniary of Studies Relevant Tb Ranraldehyte
Type cf Stidy
1. 3*
2.
3.
4.
5. StK
6. SW.
7.
a
9.
10. £MR
11.
12.
13. R-R
14. EM*
15. EM*
16. EM*
17. EM*
Atlrr
i (19G2)
tferringtm and Srenrrn
(1975)
larringtcn and Gates
(1982)
levineet al. (1981)
Strop et al. (1981)
Wrg (1983)
Taberdav Associates
(1982)
Adissmet al. (1984a)
(1983)
Bsrtazzi et al. (1984)
Rlair et al. (1986,
1987)
Stayneretal. (1986)
W&lrath and Fkarcni
(1983)
Walxath and FtciiiaiL
(1984)
(1983)
Ii<±>ling et al. (1984)
Stayer et al. (1935)
Study Qxip
Ehthologists
EfettologLsts
F&ttologists
Cntario nerticians (irale)
Aratcnists
QianicaL vgotters
Gianical vodexs
CharnkaL vortets
ChaniaaL waters
or
U32TS
Qarcert. vcateis
N.Y. ertalners and
fimetal direobccs
O^l i (i »ma
Qiamical voders (rrale)
pxxicticn
chankal vcckers
Gannait voders
Ftefibtmt. Crop
a) U.S. vhite nale age and calendar year-
ncrtality rates; b) maters cf the
Psychiatric Assxiaticn
Ireland, Wales cr Scotland male ag2 and
year iyars ffr- nortality rates
Ireland age-sex and calendar
ncrtality rates
U.S. vhite nale age and calendar
ncrtality rates
ExprBire
/teent
Ptsait
ftsent
a) ESychiatrists; b) U.S. vhite naLe ag&-
qpecifjc nortality rates
U.S. age-sac and CRlmfer >ear-speci£k:
ncrtality rates
U.S. age-sex and nFQprrfo'" y^r-typr** f^r*
ncrtality rates
Mate ncrtality rates cf Qgland/foaLes
U.S. age-raoe-sex and calendar year
specific ncrtality rates
b&ticnal and Vx^d nale specific ncrtality rates
U.S. ae-raoe-sex and calendar yesc
ncrtality rates
Pkaant
Ereaant
ftsent
U.S. age-raoe-setc and caleidar
ncrtality rates
U.S. age-race and calenchr
prcporticr£ of dsaths
U.S. age-race and rqif»rir^f jear-spacifjc
prqxxticrB cf deaths
U.S. age-race-eeK and calendar year-
specific prqacrticns cf deaths
U.S. vhite nsLe -vf and calendar
year-specific prqpcrtians of deaths
U.S. age-race and calendar year-specific
prcporticfB of deaths
Eresent
Eresent
tteent
Asant
-------
Table 4-13 (art.)
4i
VO
Tffieaf Study
18. FM*
19. Gee-Central
20. Case-Central
21. Case-Central
22. Case-Control
23. G»se-Cantral
24. Gse-Gtntrol
25. CaserOntral
26. Case-Control
27.
28. Case-Control
and Qaiffemen
(1983)
F^etveather et ai.
(1982)
Brintan et al. (1984a)
Tblaet al. (1980)
Ibnterg et al. (1983)
(artfell et al. (1982)
Osenet al. (1984)
layeset al. (1986)
Rushet al. (1985)
[fertertnet al. (1985)
et al.
(1986a,b)
Study Qnap
Textile voters
Qiaidcal watkeos
t^bsal ard paranosal
sirus cancEsr casas
in NCandXA
Isasal ard paramsal sioos
CBIXET fftgpft in Finland
tfesal ardsincnasal carioar
nwK in Finland, SUeden,
and Etniaric
isasal and
canoer cases n
isasal and sincnasal mmer
casas in Damadc
tfesal and sincrfisaL
cases in FbLIand
SincnaBal and
cancer in Qmacticut
FJaspLratoy canaer cases
fixin a retroBiMctive
cchxt of rale wooiorters
Sincnasal and ftaryngeal
canaar in ViesMngbcn
Qcckf
Eata
U.S. age and caLendar yaar-e%ea.fic pnqpccticrB
of vhite fiaiale ctetJe
Che male aifdqee natched &ac ags,
ad >Btal aesvios cate, plant
locaticn and pa/ class
TWo hospital centrals cr cne hospital
control ard cne ffanaaofri ccrtioL
ftsent
Bresat:
Bneeait
natched fix: age, sac, race, state,
eocnonic area of USLB! residence, and
>ear of hrpital adnissicn
Che rcn-respiratEay canoer control
natchscl fiar age and sex
Che rr>lmpr«-a1 rRpnar control
natchsJ &x age-ot-diayicBis,
sex and oxntiy
ffeferot gccup identifiei fron
previxus study (lardell et al. , 1981)
Calm, redbun, pxetate cr treast controls
iratxhed fir age-eex-^ear of cUagTcsis
Living and rbceasad populaticn controls
Ctntrols saiplsd frxjn Connecticut
death certificates
Three controls selecfai fron the ochort
and natched £r age
Controls sarpLed iron pcpulatkn
natchad fix age and sex
Resent
Hresent
Present
Breeent
fresent
Eteaent
aBqx)Sure levels are inferraJ fron industrial hygiene studies of similar voters.
levels are infiarai fron rurtoer of ^ears erplx^«d in oocupatiai.
-------
magnitude of the increase, background incidence of the disease,
desired statistical significance, and type of analysis.
Several newly released studies have strengthened our
knowledge regarding the potential carcinogenicity of HCHO. These
new studies have contributed stronger evidence and suggest that
HCHO may be a human carcinogen. In particular 9 studies (Acheson
et al., 1984a; Blair et al., 1986, 1987; Hardell et al., 1982;
Hayes et al., 1986; Stayner et al., 1985; and Vaughan et al. ,
1986b) show among different groups statistically significant
associations between site-specific respiratory cancer and
exposure to HCHO-containing products. Three of these studies
(Blair et al., 1986; Blair et al., 1987, Stayner et al., 1986;
and Vaughan et al., 1986a,b) were specifically designed to detect
moderate elevations in human risk. In addition, the Epidemiology
Panel of the Consensus Workshop on Formaldehyde (1984) and the
EPA (1984b) examined a group of studies and concluded that a
group of professionals (anatomists, pathologists, embalmers, and
undertakers) have a significantly increased mortality from
leukemias and brain neoplasms. These excesses in mortality can
not be attributed to diagnostic bias since these excesses remain
when other professional or like socioeconomic groups are used as
referents.
4-80
-------
4.7.2. Review of Studies Overview and Discussion
Twenty-eight studies (Table 4-13) of populations that may
have been exposed to HCHO have been reviewed. Appendix 2
contains a description of each of the studies. These studies
were of cohort or case-control designs. Results were expressed
as Standardized Mortality Ratios or Proportionate Mortality
Ratios ** or as odds ratios***. Eleven studies were of chemical
or industrial workers and seven studies were of medically-related
professions. For medically-related professions, e.g.,
morticians, embalmers, anatomists, and pathologists, the exposure
was to formalin. This group has diverse chemical exposures, but
formalin is one exposure which is common. Ten other case-control
studies examined occupational etiologies of sinonasal cavity and
pharyngeal cancers. Exposure in these studies was examined
directly by quantitating HCHO levels or indirectly through
**Standardized Mortality Ratios (SMR), from cohort studies, are
measures of the extent to which mortality in the exposed cohort
under study compares to the mortality experience among
unexposed persons. An SMR divided by 100 is called a risk
ratio. The SMR analysis uses death rates of a general
population to derive the expected number of deaths.
Proportionate Mortality Ratios (PMR) are measures in which the
cause-specific proportions of mortality among the exposed
(observed deaths) are compared to the expected proportion of
deaths among the unexposed (general population). In the PMR
study, large excesses of deaths due to one cause can de'flate
the remaining proportions and can, thus, bias comparisons of
the other causes of deaths.
***The odds ratio (OR), from a case-control study, gives the
extent to which exposed individuals are represented among the
affected cases more than among the controls to whom they are
compared. If the disease under study is rare, the odds ratio
is numerically very close to its associated SMR, but the causal
inference is not as direct. In addition, an odds ratio
obtained from a case-control study nested within a cohort
design can be used to support conclusions from the cohort
study.
4-81
-------
particular occupations where HCHO exposure has been known to
occur.
The sparsity of individual exposure data made it difficult
to separate formaldehyde from the other occupational or
residential exposures. Table 4-14 shows that twelve studies have
exposure data for individual members of the study; 6 of these
studies (Acheson et al., 1984a; Blair et al. , 1986; Stayner et
al., 1986; Fayerweather et al., 1982; Partanen et al., 1985; and
Vaughan et al., 1986) have enough information to examine an
exposure-response gradient. Other studies inferred exposure by
citing previous industrial hygiene data of similar occupational
groups. This review used exposure estimates identified in EPA
(1984b) of similar occupations as a surrogate for those
epidemiologic studies where exposure levels were not
identified. It is not known whether the individuals under study
did or did not have formaldehyde exposures at the levels
identified in EPA (1984b). Table 4-14 presents exposure
estimates for occupations identified in this review.
Each of the 28 studies has been evaluated with respect to
bias, confounding, and chance. Excesses, both statistically
significant and not statistically significant, in site-specific
mortality have been emphasized in this review. Deficits were
also noted. Deficits are hard to interpret except when examining
therapeutic treatments. Findings that are not statistically
significant are important in light of the small numbers of site-
specific neoplastic deaths expected or observed in many of the
studies (usually fewer than 5 deaths). Low statistical power is
4-82
-------
Table 4-14
Formaldehyde Levels to Which Occupational Groups Might Be Exposed
Occupation
Average
Formaldehyde
Level (ppm)
Reference
Embalmers
0.3 - 0.9
0.2 - 0.9
0.1 - 5.3
1.37 - 1.70
Levine, 1984
NIOSH, 1980
Kerfoot and Mooney, ]
EPA, 1984b
Anatomists
0.07 - 0.14
Stroup, 1984
Pathologists
0.85a, 3.2*
EPA, 1984b
Resin Manufacturing
Textile Manufacturing
Apparel Manufacturing
Wood Furniture Manufacturing
Particleboard Manufacturing
2.2 - 3.3
0.24a, 1.40b
0.70a, 0.42b
0.64a, 0.23b
0.10b, 1.30b
0.33a, 0.31b
Bertazzi et al., 198!
EPA, 1984b
EPA, 1984b
EPA, 1984b
EPA, 1984b
EPA, 1984b
Personal sample
b Area Sample
4-83
-------
characteristic of several studies. The power of a study is the
ability to detect a certain level of risk. Insufficient follow-
up and small sample sizes in the cohort studies compound to low
power through insufficient person-years and through cancers not
yet having appeared. Thus, elevations in specific-site cancers
in individual cohort studies which are not consistently observed
across all studies should not be totally discounted because they
are not statistically significant. Likewise, the absence of rare
cancers, e.g., nasal, in all cohort studies (except Blair et al.,
1986) may be a reflection of power. A similiar situation may be
observed in the case-control studies. Small numbers of cases for
any given exposure lowers the detection power. Thus,
associations with a specific neoplastic site may not be
consistently observed across like exposures.
The question of the validity of multiple comparisons always
arises in an examination of many studies and sites. Twenty-eight
studies have been reviewed and to account for multiple
comparisons by dividing a commonly accepted p-value by the number
of comparisons yields a stringent rejection value. This
rejection value will not be employed for this review since its
use would have the impact of diminishing the statistical power in
the 28 studies to detect a true positive. As previously
identified, many of these studies already suffer from low
statistical power.
Tables 4-15 through 4-17 present power calculations for the
reviewed studies. Each table summarizes, by study design,
observed and expected numbers of deaths for neoplasms of the
4-84
-------
Ta b I e
Studies
Study
Size
Matanoskl (1982)
pathologist*
1556
pathologist*
1439
oo
ui
Levlne et al. (1934)
•ortIclans
H77
Harrington and Shannon
-------
Harrington and (takes (1984)
•die pathologlsts
Table 4-15 (cont.)
2507
Srroup (1984)
anatomists
2239
.£»
I
00
en
(I98J)
che«lcal workers
2067
Tabersnan Associates (1982)
formaldehyde
exposed
chemical workers
867
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
brain
leukernI a
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
NG
9
NG
2
0
12
20
10
18
10
0
II
1
8
NG
3
6
2
0
2
0
NG
3
NG
3
NG
0
NG
22.0
NG
1.2
3.0
I.I
0.,*
6.8
43.0
18.5
3.7
14.4
6.7
0.4
a
1.9
3.8°
NG
11.6
NG
1.6
4.4
0.6
NG
NG
5.2
NG
0.7
2.0
NG
NG
•41 i i
' 1.6
~" -• — _
•331
67 J.O
"»' 5.1
"'* 2.2
'28 1.4
'<» |.6
•271
'25 ,.a
148 2.2
8.0
•579
•212
OS i f.
"3 1.9
IBM A >
100 4. j
'* 2.6
308 7 ?
' • f.
58 2.4
'« 5.8
'« 4.0
-..
—
i.a
Jr
.5
6. 1
2.4
1.5
1.8
2.0
2.5
10.5
---
2.0
5.0
3.0
9.0
2.7
7.3
4.7
-------
Table 4-15 (cont. )
oo
Achason ut a I. ( I984a>
chemical workers
7716
BIP plant
Berta//! at al. (1985)
formaldehyde
exposed
resin workers
Blair et al. a (1986)
formaldehyde
exposed
manufacturing and production
workers
4462
26561
Kith 20 years latency*
Stayner et al. (1986)
garment workers
I 1010
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
lung
nose
boccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
buccal cavity and pharynx
nasopharynx
lung
colon
brain
lymphopoietic
leukemia •
Hodgkln's disease
nasal
lung
nasopharyngeal
buccal cavity and pharynx
buccal cavity
pharynx
colon
brain
lymphopoletIc
leukemia
connective tissue
nasal
5
205
NG
5
20
9
0
166
0
NG
5
NG
NG
3
NG
0
18
6
201
42
17
56
19
14
2
148
3
6
4
2
NG
5
18
9
4
0
4.3
196.0
NG
12.5
26.3
11.4
1. 1
MI.O
0.7
NG
3.7
NG
NG
1. 1
NG
NG
19
2
192
48
21
62
24
10
2.2
no
i
3.9
1.2
1.8
NG
7.0
19.8
7.9
I.I
0.6
:if ft >•:.
40 '•« l"l
76 '-6 , 7
79 '-8 2.*0
«•' 5.6
"8 13* • ,'
'•' 1.4
6-' 7.7
1 36 •> a
0 2-8 3.2
"* ~ ~
27i 5.0 6.0
96 '-6 , ,
•300 ... ''
»i ..' ,;;»
87 4 , .
•* 1.5
81 .<> ,.a
91 -4 ,.4
80 s
* 17
'42 2 0 ->"-,
•" 2.2
91 3-4 4.0
•135
300 4.3 ;;;
155 ?•<> 3.0
•343
113 --- """
71 '; \
'• 5 2.6
91 '•/ 1.9
'14 2.i 24
*. ^
•364
i > .. ,.
9.0
-------
Table 4-15 (cant.)
• p < 0.05
d These nunbars «are obtained using Molina's tables of Polsson's Exponential Binomial Unit (Molina, 1942).
NG, observed or expected nunbar ot deaths not given In paper.
c Age-spec I Me mortality rates of psychiatrists used as the comparison group
d
Power MAS not calculated lor sites where statistically significant elevations ware observed,
* As described In Levlne et al. (1984) °
Because the expected nu»ber of deaths was large, least detectable relative risks were obtained by the approximation ot Beaunont and Bras low (1981).
g
SMR for the Blair et al. study are for analyses based on white males with time-weighted average exposure ot >0.1 ppm formaldehyde.
h White males with cumulative formaldehyde exposure greater than 0 ppn-years.
Observed and expected numbers ot deaths are tor white male wage workers.
QO
o>
-------
Table 4-16
Con d 111 on a I Hone r Calculations tor FMH Studies'
Study
il/t
oo
MJ
Malrath and Fraumenl (1981)
NY «*balmers
and funeral
directors
1132
Malrath and Frat
C«l Horn I a
embalmers
an I (1984)
1050
Harsh U9bJ)
resin
Manufacturing
workers
2490
lleblIng et al. (1984)
formaldehyde
resin
workers
24
Cancer
Site
buccal cavity and pharyn>
lung
colon
brain
lymphopoietic
Ieukemla
nasal
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
Ieukemla
nasal
buccal cavity and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
buccal cavity and pharynx
lung
colon
brain
Iymphopoletlc
leukemia
nasal
Least detectable
•Ith power6
90*
8
72
29
9
25
12
0
8
41
30
9
19
12
0
0
NG
NG
NG
2
NG
0
2
NG
NG
NG
1
NG
0
7 I
• . i
66.8
20.3
5.8
20.6
a.5.
0.5"
6.1
42.9
16.0
4.7
15.5
6.9
0.6d
0.8
NG
NG
NG
2.3
NG
NG
0.2
NG
NG
NG
0.5
NG
NG
1 13
108
•143°
156
121
140
131
96
•187
•193
123
•175
-__
86
•870
— — —
— _
— — —
217
---
230
130
240
160
210
860
220
140
170
7~2~0~d
530
350
860
250
140
280
170
240
1060
260
151
190
880d
660
—
--.
400
^
1060
-------
Table 4-16 (cont.)
Stayner et al.
garment
•or ker s
Delzell and Grutterman
textl la
•or ker s
256 buccal cavity
lung
colon
brain
lymphatic and
hematopolet Ic
1 eukaml a/a 1 eukaml a
nasal
4462 buccal cavl ty and pharynx
lung
colon
brain
lymphopoietic
leukemia
nasal
3
1 1
NG
1
10
4
0
18
106
115
17
121
45
NG
0.4
12.2
NG
2.1
6.1
2.4
NG
18.0
117.8
115
18.9
64.2
37.5
NG
•750
90
_--
48
163
168
100
90
100
90
188
120
180
- — —
380
220
330
-_.
170
120*
170
130
145
200
440
250
380
180
130*
130*
180
140
150
<=> • p<0.005
4 Conditional on the observed number ot deaths since distribution ot H and H might not have a Polssln distribution (Mlettlnen and Nang, 1981),
These numbers Mere obtained using Molina's tables ot Polsson's Exponential Binomial Limit (Molina. 1942).
c Power was not calculated tor those sites where statistically significant elevations were observed.
d
As published In Levlne et al. (1984).
* Because the expected numbers ot deaths was large, least detectable PMK's were obtained by tha approximation ot Beaumont and tireslow (1982).
-------
Table 4-17
Power Calculations for Case-Control Studiesa'b
Least Relative Odds to
Study
Fayerweather et al. (1982)
cancer deaths
in chemical
workers
Brinton et al. (1983)
481
481
(1:
160
Size
cases/
controls
1) match)
cases/
Cancer Site
lung, bronchus
and trachea
lymphopoietic
prostate
brain
nasal cavity
Exposure Ratio = P
o
forma Idehyde:
males workers (20%)
textile workers:
Odds
Ratio
0.74
0.72
3.20
0.45
detect
80%
2.0
3.5
4.4
1 1.6
with power0
90%
2.2
4. 1
5.5
16.8
V0
nasal and paranasal
sinus cases in NC and VA
nasal and uasopharyngeal
cases in Sweden
Olsen et al. (1984)
nasal cancer
cases in Denmark
290 controls
(1:2 match)
541 controls
839 cases/
2465 controls
and sinuses
nasopha ryngea1
nasal cavity
and sinuses
Hayes et al_. (1984)
nasal and nasal
sinsus cases in
the Netherlands
144 cases/
353 controls
(1:2 match)
nasal cavity
and sinuses
females (17.4%) 1.8 2.5
manufacture: *5.8
males (0.8%)
formaldehyde:
females (0.1%)
males (4.2%)
textile dust:
females (2.5%)
males (1.9%)
formaldehyde: (Classification A)
males with no or low
level wood dust
exposure (6.2%) 2.8 3.4
males with high
level wood dust
exposure (50%) 1.9 6.9
formaldehyde: (Classification A)
males, controlled *1.9
for high wood dust exposure
2. 8
•2.8
2.8
1.3
0.7
15.0
2.8
2.4
18. O
3.2
2.7
3. 7
8.0
-------
Tattle 4-17 (cont. )
Partenen et al. (1985)
nested respiratory
cancer case-control
study
Vaughan et a I. (in press)
sinonasal and
pharyngeal cases in
the U. S.
55 cases/
169 controls
respiratory
system
S3 cases/
552 controls
nasal cavity
and sinuses
yf>
NI
27 cases/
552 controls
nasopharyngea 1
174 cases/
552 controls
oro-hypo-
pharyngeal
formaldehyde:
ever exposed (26.6%) 1.4
level of exposure
0. 1 - 1.0 ppm (16%) 1.5
> 1.0 ppm (7.7%) 1.4
occupational formaldehyde:
cumulative exposure
5-9 years (6.3%) 1.1
10 t years (10.9%) 0.3
occupational:
resins, glues and
adhesives
low exposure (6.5%) 2.0
high exposure (2.3%) *3.8
domestic:
mobile home residence
1 + years (12.0%) 1.7
occupational formaldehyde:
no. of years exposed
1 - 9 (25.0%) 1.2
10 * (10.3) 1.6
occupational:
stains, varnishes,
solvents
low exposure 0. 9
high exposure *4.0
domestic:
mobile home residence
1 - 9 years (12.0%) 2.1
10 + years (3.7%) *5.6
occupational formaldehyde:
no. of years exposed
1 - 9 (25.0%)
10 + (10.3%)
occupa t i ona1:
resins, glues,
adhesives e
low exposure
high exposure
0.6
1.2
1.3
•3.9
2.5
2. 7
3.4
4.0
2. 7
3.2
2.7
3.2
4.5
3.7
1.7
2. 1
2. a
3. 1
3.9
4.8
3. 3
3.7
3. 3
4.0
5.6
4. 7
1.9
2.4
-------
Table 4-17 (cont.)
stains, varnishes,
solvents
low exposure6 1.0
high exposure6 *3.0
domestic:
mobile home residence
1-9 years (12.0%) 1.0 2.0 2.2
10 + years (3.7%) 0.9 2.9 3.3
*p<0.05
a Power calculations for Tola et al. (1980), Rxish et al. (1985) and Hemberg et al. (1983) could not be calculated due
to the unknown exposure ratio (po) among the controls.
Power was not calculate:! for studies where statistically significant elevations were observed.
c Obtained fran the study by Fayerweather et al. (1982) or was calculated using the method in Rothnan and
f Boice (1982) for the studies by Brinton et al. (1983), Olsen et al. (1984), Hayes et al. (1984), and
£ Hardell et al. (1982).
d Matching ratio not identified by Hardell et al. (1982).
e Prevalence of the exposure among the controls not cited by Vaughan et al. (As reported in SAIC, 1986).
-------
hematopoietic site cancer as an .example, Table 4-15 shows that
Levine et al. (1984) could detect, with 80 percent power, a
relative risk of 2.3 or greater and with 90 percent power, a
relative risk of 2.6 or greater.
The'above seven sites were selected for several reasons.
First, because the exposure is generally by inhalation and nasal
tumors were seen in the chronic rat study (Kerns, 1983), the
respiratory system is a reasonable place to look for effects.
Nasal tumors, buccal cavity tumors, and pharyngeal tumors were
included because man, unlike the rat, is not an obligatory nose
breather and inhaled formaldehyde would initially contact these
areas. Last, the Epidemiology Panel, Consensus Workshop on
Formaldehyde (1984) and Levine et al. (1984) report significant
excesses in brain, leukemia, and colon cancer mortality when
results across studies were combined.
Epidemiologists use five criteria for judging whether an
association is causal. These criteria are: 1) strength of the
association, 2) consistency across studies, 3) temporally correct
association (disease occurs after exposure), 4) specificity of
the association, and 5) coherence with existing data. The
reviewed epidemiologic studies are of a cohort or case-control
design, designs which examine health consequences of previous
exposure, thus permitting point 3 to be' satisfied.
The Blair et al. (1986), Blair et al. (1987), Vaughan et al.
(1986a,b), and Stayner et al. (1986) were designed to detect
moderate increases in formaldehyde-related risks. The Blair et
al. (1986) and Vaughan et al. (1986) studies observed significant
4-94
-------
associations with nasopharyngeal cancer and apparent exposure to
formaldehyde, in either the occupational or residential
environment. Vaughan et al. (1986) reported a significant
association between the incidence of nasopharyngeal cancer and
living 10 or more years in a mobile home. This study also
reported statistically significant associations between sinonasal
cancer and oro-hypopharyngeal cancer and exposure to resins,
glues, and adhesives (SAIC, 1986). Mobile home residency and
occupational resins, glues and adhesives exposure were a priori
selected as likely surrogates for formaldehyde exposure. No
statistically significant associations were found between cancer
incidence at any of these sites with respect to occupational
formaldehyde exposure as assessed using an occupational linkage
system. The risk estimates, however, for the highest exposure
level and cancers of the oro-hypopharynx and nasopharynx appeared
elevated. These results for the occupational formaldehyde
exposure most likely were biased towards, the null hypothesis
since a large proportion of the case interviews were with the
next-of-kin, respondents less likely to report or remember all
jobs the case had ever held.
Blair et al. (1986, 1987) observed excesses of lung and
nasopharyngeal cancer mortality among U.S. workers exposed to
formaldehyde in 10 plants. The highest risks were observed for
lung cancer among men with a 20-year latency and for
nasopharyngeal cancer among men with exposure to formaldehyde-
containing particulates. An apparent dose trend was observed
between nasopharyngeal cancer mortality and exposure to
4-95
-------
formaldehde and particulates (Blair et al., 1987); no clear
trends were observed between lung cancer mortality and
formaldehyde level (Blair et al., 1986). Blair et al. (1986)
argued the data provide little evidence that lung cancer
mortality is associated with formaldehyde exposure at levels
experienced by workers in this study, although they concluded
that simultaneous exposure to formaldehyde and particulates
appear to be a risk factor for nasopharyngeal cancer. Blair
et al. (1987) additionally state that further investigation is
needed regarding the dose-dependent association between
nasopharyngeal cancer mortality and exposure to formaldehyde and
particulates. The significant excesses in total lung cancer
mortality, in analysis either with or without a latency period
equal to or greater than 20 years, and in nasopharyngeal cancer
mortality among ever-exposed workers are meaningful. Inhalation
is the primary route of exposure for this cohort. Second,
misclassification of exposure, the lack of specificity between
the narrow exposure categories, may account for the lack of a
statistically significant trend between lung cancer mortality and
formaldehyde level. Blair et al. (1986) relied upon historical
industrial hygiene data, process changes, and human recall to
reconstruct past exposure to formaldehyde. The observed wide
variations in historical industrial hygiene data for any given
job and the reliance upon human memory may have contributed to
misclassification.
Stayner et al. (1986) reported significant mortality
excesses from neoplasms of the buccal cavity, connective tissue,
and tonsils among formaldehyde-exposed garment workers. The risk
4-96
-------
ratio for buccal cavity cancer was significantly elevated among
workers with a long duration of employment (exposure) and follow-
up period (latency). Although it is not presented in Stayner et
al. (1986), EPA calculations showed a statistically significant
trend between buccal cavity cancer mortality and increasing
duration of employment. A significant excess in deaths from
cancer of the tonsils (located in the oropharynx) was also
reported, but there were too few deaths (only 2) to examine any
trends with exposure.
The significant associations between formaldehyde exposure
and excesses in site-specific buccal cavity and respiratory
cancers support observations from EPA's previous review of the
epidemiological literature. These other studies had limited
ability to detect formaldehyde-related risks due to lower
power. Even with this potential limitation, 6 studies (Olsen et
al., 1984; Hardell et al. , 1982; Hayes et al., 1986; Acheson et
al., 1984a; Liebling et al., 1984; and Stayner et al., 1985)
reported significant associations between excess site-specific
respiratory or buccal cavity and pharyngeal cancer and exposure
to formaldehyde.
The Hayes et al. (1986) and Olsen et al. (1984) studies
report significant excesses of sinonasal cancer and exposure to
both formaldehyde and wood-dust. Both studies controlled for
simultaneous wood-dust exposure, and by doing so, the detection
limits exceeded excesses in expected sinonasal neoplastic risk.
Hardell et al. (1982) reported a significant excess in si.nonasal
cancer and employment in particleboard manufacturing.
4-97
-------
The other 3 studies examined mortality among workers
occupationally exposed to formaldehyde-containing products.
Acheson et al. (1984a) observed a significant elevation in lung
cancer among formaldehyde resin workers in 1 plant in the U.K.
Acheson et al. (1984a) observed at this plant a marginally
significant trend with dose. Acheson et al. (1984a) concluded
that the increased lung cancer mortality and positive trend were
not related to formaldehyde exposure since analyses using local
cancer rates as the comparison were not statistically
significant. EPA believes that the risks and trends from
analyses using local lung cancer rates as the comparison appeared
sufficiently increased for corroborative use. EPA notes,
however, it was not known how many of the formaldehyde-exposed
lung cancer deaths were included in the deaths of the local
comparison group. This would lead to a reduction in power since
the same death could be counted in both the numerator and
denominator.
The 2 other studies (Stayner et al., 1985 and Liebling et
al., 1984) reported statistically elevated SMR's for buccal
cavity cancer among garment workers and for buccal cavity and
pharyngeal cancers among formaldehyde resin workers in 1 plant.
Portions of the Liebling et al. (1984) and Blair et al. (1986,
1987) studies overlapped as did portions of the two Stayner et
al. (1985; 1986) studies. The nonoverlapping portions and
improved design of the more recent studies (i.e., Blair et al.,
1986; Blair et al., 1987; Stayner et al., 1986) reinforce the
conclusions of the earlier studies.
4-98
-------
The studies of embalmers, anatomists and pathologists
reported deficits in mortality from site-specific respiratory
cancers. Although individual characteristics are not known, this
observation may be a reflection of decreased smoking habits among
these professional groups in comparison to the general
population. Since expected deaths are based on general
population site-specific mortality, the number of expected deaths
may be biased upwards with the resultant SMR being lower. The
lack of lung cancer excesses in these studies may additionally
reflect the lower statistical power to detect moderate increases
in site-specific respiratory cancer mortality.
Site-specific excesses in lymphopoietic, leukemia, colon, .
and brain neoplasms have been observed in five studies
(Harrington and Shannon, 1975; Harrington and Oakes, 1984;
Stroup, 1984; Walrath and Fraumeni, 1983; '.Valrath and Fraumeni,
1984), but these excesses were not statistically significant
across all studies. This lack of consistency may reflect lack of
a causal relationship or may reflect limited power to observe
excesses at specific sites because of small sample sizes,
insufficient follow-up, different exposure levels, and different
routes of exposure in the individual studies. The Epidemiology
Panel of Consensus Workshop on Formaldehyde (1984) summarized the
observed and expected numbers of site-specific cancers from both
SMR and PMR studies and noted statistically significant excesses
in mortality from brain cancer and leukemia among embalmers,
pathologists, and anatomists. Levine and co-workers (1984) 'jse.i
4-99
-------
this same method and, additionally, noted a statistically
significant excess in mortality from colon cancer.
The same conclusions were reached through the use of another
approach by summarizing all SMR and all PMR site-specific
findings according to Fisher's combined probabilities method
(Sokal and Rolf, 1969). Fisher's combined probabilities can be
used on different sets of data that test the same scientific
hypothesis and where for one reason or another a joint
statistical analysis is not optimal. This methodology does not
assume all studies are equal. Again, Fisher's combined
probabilities may not be an ideal test. Its use on discrete or
count data tends to bias the results toward non-significance
(Gastwirth, 1983). A detailed description of the procedures and
analyses are presented in EPA, 1984b. Table 4-18 presents the
summarized results. Thus, where site-specific mortality has been
reported, it can be concluded from SMR studies that brain cancer
mortality is significantly elevated (p<0.05) among pathologists,
anatomists, and embalmers, and from PMR studies that leukemia and
brain cancer mortalities are significantly elevated (p<0.05) for
these professions. Likewise, from PMR studies, colon and
lymphatic and hematopoietic cancer mortalities are significantly
elevated (p<0.05) for manufacturing workers.
To further examine power, the human data were compared to
estimates of the upper bound risk that were calculated based on
the malignant tumors in Kerns (1983). This comparison assumes
the excess risk calculated from the animal low-dose extrapolation
is the excess above a risk of one for the study population
4-100
-------
Cancer Site
Table 4-18
Fisher's Conbined Probability(p) for SMR and for PMR Studies
Study Design Study Population Fisher's pa
Buccal cavity SMR
and pharynx
PMR
Lung SMR
PMR
Colon SMR
PMR
Brain SMR
PMR
Lymphatic and SMR
hematopoietic
PMR
Leukemia SMR
PMR
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Vorkers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
Embalmers, Anatomists,
Manufacturing Workers
EiTibaLTiers, Anatomists,
Manufacturing Workers
Pa tho legists
Patho legists
Patho legists
Patho legists
Pathologists
Patho loists
Pathologists
Pathologi sts
Pathologists
Pathologists
Pathologists
Pathologists
0.64
0.40
0.36
0.10
0.99
0.35
0.54
0.83
<0.35
0.22
<0.01
0.48
<0.01
0.95
0.02
0.75
0.08
0.56
0.17
<0.01
0.17
0.26
0.04
0.10
Probability of -2 Z__ L-i Pj distributed' as a chi-square with 2k degrees of
freedom (Sokal and Rolf, 1969). Stall values indicate a statistically significant
elevation over 100.
4-101
-------
relative to the U.S. population (Margosches and Springer,
1983). Hence, human site-specific neoplastic relative risks are
calculated by adding the percentage increase in site-specific
tumors to a relative risk of one.
Two major assumptions are made to carry out this
prediction. First, the number of human cancer site-specific
deaths approximates lifetime incidence. Second, site-to-site
concordance between animals and humans does not necessarily hold,
and only one site is examined at a time.
Table 4-19 presents the upper bounds on predicted human
relative risk for seven neoplastic sites. Based on Table 4-19,
we would expect to see a relative risk of around 1.26 for buccal
cavity and pharyngeal neoplasms for funeral service workers
(morticians and embalmers). It must be noted that the predicted
relative risks vary greatly with the mortality. The rarer the
cause of mortality the higher the predicted relative risk will
be. Comparing the occupation-associated predicted human relative
risks with the least detectable risks identified in Table 4-15 to
4-17, we see that very few of the SMR and PMR studies had 90%, or
even 80%, power to detect this upper bound predicted human nasal
sinus and cavity risk. Most of case-control studies had 30%
power to detect such excess nasal cavity and sinus risks, if
average exposure to formaldehyde for the cases was around
1.3 ppm. Note, the reviewed studies had over 90% power to deter",
a predicted upper bound on nasal cavity and sinus relative risk
(RR=101.0) that was based on animals bearing either nasal cavity
and sinus squamous cell carcinomas or polypoid adenomas if
4-102
-------
4-19
Cancer Site
ELneral Service Nbnufectiring
Vtrtem ^pjarel Resin Flznitxre article Bard
(1.70Hn)b (0.64pon)b (1.40ppn)b (1.30ppm)b (0.10ppn)b (0.33 con)
Rraai cavity 1.26c
and pharynx
ling
Gblm
frain
r\jirt'TY"T 1 '11 f& If
Ta.Jtema
ISbsal cavity
and sirus
1.02
1.05
1.24
1.05
1.13
6.45
a This procedure is describee
on an asanpticn that eases
is the acess above a risk
to an inexposed or g=neral
tVt^ h (+ art hi
EfcpoFticns
b Eeranal ex:
1.23=
1.02
1.04
1.22
1.05
1.12
5.83
lb/^fetc
B ri* c
of one |
1 1 1 ulaf'i
nan relative risk = 1
*TT 'tilths C^lOl'^^^* fri
ri^M'-r-jbgd in EE
insure esturates
A, 198fc
fixm EE
(5.65)d
(1.38)
(1.90)
(5.44)
(2.04)
(3.41)
1.53=
1.04
1.10
1.51
1.12
1.28
(101) 12.45
• ili ii i i ii 1 1 f
pares arc S
Rio list-art fr
br ahuraa e
en.
+ P(d)UO^
1.44= 1.03= l.llc
1.04 1.00 1.01
1.09 1.01 1.02
1.42 1.03 1.11
1.10 1.01 1.03
1.23 1.01 1.06
10.65 1.60 2.42
Textile
(0.70 ppn)
1.26c
1.02
1.05
1.24
1.06
1.13
6.50
fcringsr (1983). ft is based
cxi anirral kw-dcse extrapolaticn
ygoBed prpilfrt-.im relative
5-stage nodal
an I960 ncrtality data.
•
^ 1964D. -
c l£per bond predicted huran relative ri^s Vnffrl en a P(d)UCL
P(d) is estinatad fron the nnber of aninals bearing aqiBnaJs'c
d Lfper boxid predicted hjtan relative ri*s basad en P(d)UCL> s^taqe nccel,
P(d) is estinatal fixm the trtal rinber cf aninals baaring aqLarDSceQ"carcinaias or
polypoid adenaras.
4-103
-------
exposure was around 0.64 ppm or greater. None of the studies
could detect the predicted relative risks that were based on
animals bearing only squamous cell carcinomas for brain, or for
leukemia, or lymphopoietic, or colon, or lung neoplasia for the
populations studied.
4.7.3 Conclusion
The EPA has examined 28 studies which related to
formaldehyde. Three of these studies (Vaughan et al., 1986a,b;
Blair et al., 1986, 1987; and Stayner et al., 1986) were designed
to detect moderate elevations in human cancer risks; the
remaining 25 studies had detection limits that exceeded
corresponding expected excesses in site-specific neoplastic
risks. Results from 9 studies (Vaughan et al., 1986a,b; Blair et
al., 1986, 1987; Stayner et al., 1986; Olsen et al., 1984;
Hardell et al., 1982; Hayes et al., 1986; Acheson et al., 1984a;
Liebling et al., 1984; and Stayner et al., 1984) suggest that
lung, nasopharyngeal> sinonasal, and oro-hypo-pharyngeal cancers
are associated with formaldehyde exposure.
In each of the above 9 studies, the populations were also
undoubtedly exposed to other agents and these exposures may have
contributed to the observed increase in cancer risk. Five
studies, however, addressed confounding.' Vaughan et al.
(1986a,b) controlled for sinokirv-j and alcohol consumption in their
analyses. Hayes et al. (1986) and Olsen et al . (1984) controlled
for wood-dust exposure; the detection limits in both studies
exceeded corresponding expected excesses in sinonasal neoplastic
4-104
-------
risk. Stayner et al. (1986) measured possible confounders such
as phenol or particulate dusts in their study and thought the
contribution of the confounders to the observed excesses in
buccal cavity or pharyngeal mortality were minimal to none.
Mote, Stayner et al. could not measure the impact of smoking on
their observed excesses in mortality. Blair et al. (1986), on
the other hand, stated that the lack of a consistent elevation
for tobacco-related causes of deaths suggested that their
cohort's smoking habits did not differ substantially from those
of the general population. Regarding diagnostic bias accounting
for the observed brain cancer excesses, the brain cancer excesses
remained when other like socioeconomic groups were used as the
comparison. Socioeconomic status may be a confounder in the
observed associations with upper respiratory cancers, but no data
currently exist for evaluation. As identified earlier, smoking,
sometimes associated with socio-economic status, either has been
taken into account in analyses or was thought to be appreciably
similar within the individual study comparison groups.
Based on the above human evidence, formaldehyde can be
placed in the "limited evidence of carcinogenicity" category.
This category is defined as "indicates that a causal
interpretation is credible, but that alternative explanations,
such as chance, bias, or confounding, could not adequately be
addressed" (EPA, 1986).
Formaldehyde should not be placed in the categories
"inadequate evidence" or "sufficient evidence of
carcinogenicity." "Inadequate evidence" is defined as
4-105
-------
"indicating that one or two conditions prevailed: (a) there were-
few pertinent data, or (b) the available studies, while showing
evidence of an association, did not exclude chance, bias, or
confounding." Placemen't into the "inadequate evidence" category
would imply that the studies contained insufficient data to
consider adjusting for alternative interpretations. The
aforegoing discussion shows this is not the case. "Sufficient
evidence of carcinogenicity" is defined as "indicates that there
is a causal relationship between the agent and human cancer." A
variety of plausible important exposures could have confounded
these results, but no adjustment could be made for them. In
addition, the association between mobile home residence and
nasopharyngeal cancer was a first report; future epidemiological
studies would be needed to confirm a causal association between
the formaldehyde exposure in a mobile home and the incidence of
nasopharyngeal cancer.
On this basis, EPA has concluded that the epidemiological
evidence is "limited".
4-106
-------
4.8. Weight-of-Evidence
4.8.1. Assessment of Human Evidence
EPA examined 28 epidemiologic studies relevant to HCHO.
Three of these studies, two cohort (Blair et al., 1986; 1987
Stayner et a'l. , 1986) and one case-control (Vaughan et al.,
(1986a,b), were well conducted and specifically designed to
detect small to moderate increases in HCHO-associated human
risks. Each of these three studies observed statistically
significant associations between respiratory site-specific
cancers and exposure to HCHO or HCHO-containing products. In
each of the above three studies, the populations .studied were
also undoubtedly exposed to other chemicals and these exposures
may have contributed to the observed increases in cancer risk.
Only the study by Vaughan et al. (1986a,b) controlled for smoking
and alcohol consumption.
The Blair et al. (1986; 1987) cohort study observed
significant excesses in lung and nasopharyngeal cancers among
U.S. workers occupationally exposed to HCHO at 10 industrial
sites. Blair et al. (1987) conclude that formaldehyde and
particulates appear to be a risk factor for nasopharynyeal
cancer. Blair et al. (1986), however, argued that the lung
cancer excesses provided little evidence of an association with
HCHO exposure since the lung cancer risk did not increase
consistently with either increasing intensity or cumulative HCHO
exposure. EPA, after reviewing these data, has concluded that
the significant excesses in total lung cancer mortality, in
4-107
-------
analyses either with or without a latency period equal to or
greater than 20 years, and together with nasopharyngeal cancer
mortality among HCHO-exposed workers are meaningful despite the
lack of significant trends with exposure.
The Stayner et al. (1986) cohort study reported
statistically significant excesses in mortality from buccal
cavity tumors among HCHO-exposed garment workers. The SMR was
highest among workers with a long duration of employment
(exposure) and follow-up period (latency).
Results from the case-control study by Vaughan et al.
(1986a,b) showed a significant association between nasopharyngeal
cancer and having lived 10 or more years in a "mobile home".
Persons for whom this association was drawn had lived in mobile
homes that were built in the 1950s to 1970s. This study also
reported significant associations between .sinonasal cancer and
orohypopharyngeal cancer and exposure to resins, glues, and
adhesives (SAIC, 1986). No significant trends were found in
cancer incidence at any of these sites with respect to
occupational HCHO exposure; however, the risk estimates for the
highest exposure level and cancers of the orohypo-and naso-
pharynx appeared elevated. However, this population, like the
two previously discussed, was also undoubtedly exposed to other
chemicals which may have contributed to the observed increases in
cancer risk.
EPA reviewed 25 other epidemiologic studies. These studies
had limited ability (lower power) to detect small to moderate
4-108
-------
increases in HCHO-related risks due to (1) small sample sizes;
(2) small numbers of observed site-specific deaths; and (3)
insufficient follow-up. Even with these potential limitations,
six of the 25 studies (Acheson et al., 1984a; Hardell et al.,
1982; Hayes et al., 1985; Liebling et al., 1984; Olsen et al.,
1984; Stayner et al., 1985) reported significant associations
between excess site-specific respiratory (lung, buccal cavity,
and pharyngeal) cancers and exposure to HCHO.
The Olsen et al. (1984), Hayes et al. (1986), and Hardell
et al. (1982) studies reported significant excesses of sinonasal
cancer in individuals who were exposed to both HCHO and wood-
dust, or who were employed in particleboard manufacturing where
HCHO is a component of the resins used to make particleboard.
Only the Hayes et al. (1986) and Olsen et al. (1984) studies
controlled for wood-dust exposure; the detection limits in both
studies, however, exceeded corresponding expected excesses in'the
incidence of sinonasal tumors and, therefore, no significant
excesses were likely to have been observed.
The Acheson et al. (1984a) study conducted in the United
Kingdom supports the results of Blair-et al. in that, when
compared to mortality rates of the general population,
significant excesses in mortality from lung cancer were observed
in one of six HCHO resin producing plants in England. A trend of
borderline significance with dose was observed for this one
plant. Acheson et al. concluded that the increases in mortality
from lung cancer were not related to HCHO exposure since the
4-109
-------
elevation and trend were not statistically significant when
compared with local lung cancer rates. EPA believes that the
risks and trends from analyses using local lung cancer rates as
the comparison risks appeared sufficiently increased for
corroborative use.
The remaining two studies reported significant excesses of
buccal cavity cancer among garment workers in 3 plants (Stayner
et al., 1985) and excesses of buccal cavity and pharyngeal cancer
among HCHO resin workers in 1 plant (Liebling et al., 1984).
Portions of the Liebling et al. (1984) and Blair et al. (1986,
1987) studies overlapped as did portions of the two Stayner
et al. (1985; 1986) studies. However, the non-overlapping
portions and improved design of the more recent studies (i.e.,
Blair et al. 1986, 1987; Stayner et al. 1986) reinforce the
conclusions of the earlier studies.
Analyses of the remaining 19 epidemioLogic studies have
indicated the possibility that observed leukemia and neoplasms of
the brain and colon may be associated with HCHO exposure. The
biological support for such postulates, however, has not yet been
demonstrated.
Based on a review of these studies, EPA has concluded that
under EPA's Guidelines for Carcinogenic Risk Assessment there is
"limited" evidence to indicate that HCHO may be a carcinogen in
humans. Nine studies reported statistically significant
associations between site-specific respiratory neoplasms and
exposure to HCHO or HCHO-containing products. This is of
4-110
-------
interest since inhalation is the primary route of exposure in
humans. Although the common exposure in all of these studies was
HCHO, the epidemiologic evidence is categorized as "limited"
primarily due to possible exposures to other agents which may
have confounded the findings of excess cancers.
4.8.2. Assessment of Animal Studies
The principal evidence indicating that HCHO is able to
elicit a carcinogenic response in animals are the studies by CUT
(Kerns et al., 1983), Albert et al. (1982) and Tobe et al.
(1985). In the CUT study, statistically significant numbers of
squamous cell carcinomas of the nasal cavity of Fischer 344 male
and female rats were seen. The CUT study was a well conducted,
multidose inhalation study. In addition, while not statistically
significant, a small number of squamous cell carcinomas were seen
in male mice. Because this type of nasal lesion is rare in mice,
these data must be considered biologically significant. Benign
tumors (i.e., polypoid adenomas) were seen in male rats in the
CUT study at all dose levels and in female rats exposed to 2 ppm
of HCHO. Notably, the dose-response curve for the benign tumors
in this study did not mirror the carcinoma response; the tumor
incidence was highest at 2.0 ppm and decreased at higher doses.
Tobe et al. also observed a statistically significant
increase in the numbers of squamous cell- carcinomas in the same
strain of male rats as was used in the CUT study. Albert et al.
reported a statistically significant elevation of the same
malignant tumor type in male racs of a different strain. In both
4-111
-------
the Tobe et al. and Albert et al. studies benign squamous cell
papillomas were seen. This observation was in contrast to the
CUT study in which polypoid adenomas were the only benign tumors
observed. Hamsters have been tested in long-term inhalation
studies (Dalbey, 1982) but no increased incidence of tumors was
seen in HCHO-treated animals. However, deficiencies in the study
design and poor survival limit the interpretation of the results
from these studies.
Additional support is provided by studies by Dalbey (1982)
in which HCHO increased the production of tumors caused by a
known animal carcinogen; Meuller et al. (1978) in which a
solution of formalin produced lesions in the oral mucosa of
rabbits which showed histological features of carcinoma in situ;
and studies by Watanabe et al. (1954, 1955) in which injections
of formalin and hexamethyenetetramine produced injection site
sarcomas and one adenoma.
HCHO is mutagenic in numerous test systems, and it is able
to transform a number of cell lines. In addition, HCHO has been
shown to be able to form adducts with DNA in both in vivo and in
vitro tests (Consensus Workshop on Formaldehyde, 1984). Its
ability to interfere with DNA repair mechanisms has also been
demonstrated. However, evidence demonstrating HCHO's mutagenic
potential in in vivo tests is lacking (IRMC Report on Systemic
Effects, 1984b). The literature reports conflicting data
concerning chromosomal effects in humans. However, the weight of
these data seems to indicate little potential for these effects
4-112
-------
in the workplace, but this judgement must be tempered by the
limitations of the studies.
Although HCHO's acute effects do not demonstrate its
carcinogenicity, they do help explain differences in species
response, and the severity of the carcinogenic response in the
animal studies. HCHO's acute effects may be factor in the
promotion of its carcinogenic potential at concentrations greater
than 1 ppm in rats and possibly in humans.
Another factor that bears on the possible carcinogenicity of
HCHO, is the different responses seen in laboratory animals to
HCHO. HCHO has been studied in rats, mice, hamsters, and monkeys
by inhalation. In rats a highly statistically significant
response was obtained in two strains. In mice only males of one
strain showed a marginal response, while hamsters and monkeys
showed no neoplastic response. However, the studies of
respiratory response to sensory irritants indicate that when dose
received is adjusted for reductions in respiratory rate, rats and
mice appear to respond similarly. The cancer test data on
hamsters are negative, but this finding is tempered by poor
survival, limited pathology, and other factors. The study using
monkeys (Rusch et al.) indicates that, at least for nonneoplastic
lesions (squamous metaplasia), rats and monkeys respond
similarly.
HCHO is not the only aldehyde that is carcinogenic in
animals. Acetaldehyde, the closest aldehyde to HCHO in
structure, is carcinogenic in hamsters and rats, causing cancers
4-113
-------
in the nose and trachea of the former, and nasal cancers in the
latter (by inhalation). In addition, other aldehydes such as
glycidaldehyde and malondialdehyde have been shown to be
carcinogenic.
Finally, HCHO's rapid metabolism and pharmacokinetic data,
the protective action of the mucous layer, and respiratory
response to sensory irritants have been discussed in the HCHO
literature as factors that may bear on judgements of the
mganitude of the potential human cancer risk posed by HCHO and
will be discussed in sections 7 and 9.
In conclusion, based upon a review of the above data, EPA
has concluded that there is "sufficient" evidence of
carcinogenicity of HCHO in animals by the inhalation route. This
finding is based on the induction by HCHO of an increased
incidence of a rare type of malignant tumor (i.e., nasal
squamous-cell carcinoma) in both sexes of rats, in multiple
inhalation experiments, and in multiple species (i.e., rats and
mice). In these long-term laboratory studies, tumors were not
observed beyond the initial site of nasal contact nor have other
mammalian in vivo tests shown conclusive effects at distant
sites.
4.8.3. Categorization of Overall Evidence
In conducting risk assessments of suspect carcinogens, EPA
generally evaluates the overall weight-of-evidence including both
primary and secondary evidence of carcinogenicity. As specified
in EPA's Guidelines for Cancer Risk Assessment (EPA, 1986),
4-114
-------
primary evidence derives from long-term animal studies, and
epidemiological data insofar as this is available. Secondary, or
supplemental, evidence includes structure-activity relationships,
the results of short-term tests, pharmacokinetic studies,
comparative metabolism studies, and other toxicological responses
which may be relevant.
In the process of categorizing HCHO, two lines of evidence
were assessed, one of which is an assessment of studies of humans
and the other is the assessment of evidence from studies in
animals. The results from each assessment are then combined to
characterize the overall evidence of carcinogenicity. The EPA
Guidelines also suggest that quantitative risk numbers be coupled
with EPA classifiations of qualitative weight of evidence.
Consequently, based on the determination that there is
sufficient evidence that HCHO is an animal carcinogen and the
determination that there is limited human evidence, HCHO can be
classified under the draft guidelines as a Group Bl-Probable
Human Carcinogen.
4-115
-------
5. HAZARD OF NONCARCINOGENIC EFFECTS
5.1. HCHO-Related Effects of the Eyes and Respiratory System*
Irritation of the eyes and mucous membranes is the principal
effect of Low concentrations of HCHO observed in humans. Human
responses'to airborne HCHO at various concentrations are
summarized in Table 5-1. Table 5-1 shows a wide range in HCHO
concentrations reported to cause specific health and sensory
effects. At concentrations below 0.05 ppm none of the effects
listed have been reported.
Table 5-1.
REPORTED HEALTH EFFECTS OF FORMALDEHYDE
AT VARIOUS CONCENTRATIONS
Approximate HCHO
Health Effects Reported Concentration, ppm*
None reported 0-0.05
Odor threshold 3.05-1.0
Eye irritation 0.01-2.0**
Upper airway irritation 0.10-25
Lower airway and pulmonary 5-30
effects
Pulmonary edema, inflammation, 50-100
pneumonia
Death 100+
*Range of thresholds for effect listed.
**The low concentration (0.01) was observed in the presence of
other pollutants that may have been acting synergisticaily.
*Unless otherwise cited, from NTRC (1981
5-1
-------
Most persons can perceive the odor of HCHO at about 1 ppm,
but some persons can detect it as low as 0.05 ppm. Eye
irritation has been reported at concentrations as low as 0.05
ppm. At concentrations at or above 1 ppm, nose, throat, and
bronchial irritation have been noted. Such irritation was nearly
uniformly reported by persons when the concentration reached 5
ppm. HCHO concentrations exceeding 50 ppm cause severe pulmonary
reactions, including pneumonia, bronchial inflammation, pulmonary
edema, and sometimes result in death.
Table 5-1 shows the variability and overlap of thresholds
for responses among subjects. Tolerance to olfactory, ocular, or
upper respiratory tract irritation occurs in some persons.
Factors such as smoking habits, socioeconomic status, preexisting
disease, and interactions with other poliuiants and aerosols are
expected to modify these responses.
5.1.1. Eye
A common complaint of persons exposed to HCHO vapor is eye
irritation. Some persons can detect HCHO at 0.01 ppm, but it
produces a more definable sensation of eye irritation at 0.05-0.5
ppm. Marked irritation with watering of the eyes occurs at a
concentration of 20 ppm in air. Permanent eye damage from HCHO
vapor at low concentration is thought not to occur bec'ause people
close their eyes to avoid discomfort. Increased blink rates are
noted at concentrations of 0.3-0.5 ppm in persons studied in
chambers. Blink rate, although used as an objective measure of
eye irritation, appears variable for any given subject. In smog-
5-2
-------
chamber testa human subjects tested could readily detect and
react to HCHO at as low as 0.01 ppm. The irritant effects of
HCHO seem to be accentuated when it is mixed with other gases.
Accidental splash exposures of human eyes to aqueous
solutions'of HCHO have resulted in a wide variety of injuries,
depending on concentration and treatment. These range from
discomfort and minor, transient injury to delayed, but permanent,
corneal opacity and loss of vision.
In summary, human eyes and adnexal are very sensitive to
HCHO, detecting atmospheric concentrations of 0.01 ppm in some
cases (when mixed with other pollutants) and producing a
sensation of irritation at 0.05-0.5 ppm. Tolerance to eye
irritation is reported to occur. Lacrimation is produced at
higher levels, but damage is prevented by riosing the eyes in
response to discomfort. Aqueous solutions of HCHO accidentally
splashed into the eyes must be immediately flushed with water to
prevent serious injury, such as lid and conjunctival edema,
corneal opacity, and loss of vision.
5-3
-------
5.1.2. Olfactory System
The odor threshold of HCHO is usually around 1 ppm, but may
be as low as 0.05 ppm for a small percent of the population.
General olfactory fatigue with associated increases in olfactory
thresholds for rosemary, thymol, camphor, and tar has been
reported among plywood and particleboard workers and is thought
to be associated with HCHO exposure.
5.1.3. Upper Airway Irritation (Nose and Throat)
Symptoms of upper airway irritation include the feeling of a
dry or sore throat, tingling sensation of the nose, and are
usually associated with lacrimation and pain in the eyes.
Irritation occurs over a wide range of concentrations, usually
beginning at approximately 0.1 ppm, but is reported more
frequently at 1-11 ppm (see Table 5-1). "olerance to eye and
upper airway irritation may occur after 1-2 hours of exposure.
However, even if tolerance develops, the irritation symptoms can
return after a 1- to 2-hour interruption of exposure.
Finally, examinations of the nose and throat reveal chronic
changes that are more severe in persons occupationally exposed to
higher concentrations HCHO. Exposure to HCHO can cause
alterations in the nasal defense mechanisms that include a
decrease in mucociliary clearance and a loss of olfactory
sensit ivity.
5.1.4. Lower Airway and Pulmonary Effects
Lower airway irritation which is characterized by cough,
chest tightness, and wheezing is reported often in people exposed
to HCHO at 5-30 ppm.
5-4
-------
In a study of workers exposed to phenolic resin fumes by
Schoenberg and Mitchell (1975), there was evidence of chronic
airway obstruction in workers exposed for more than five years.
This was measured by lower FEV 1.0/FVC and MEF 50%/FVC ratios.
However, as opposed to the high percentage of workers reporting
acute respiratory symptoms, only small decreases in pulmonary
function during the workday and workweek were found. In a
similar study, it was found that workers exposed to a phenol-HCHO
type resin, hexamethylenetetramine-resorcinal, experienced
significant acute lung effects (lung function measured before and
after shifts) as measured by decrements in tests measuring "small
airways" effects. However, there was no difference in baseline
lung function tests in the exposed and control populations
(Gamble et al., 1976). Finally, workers exposed to HCHO from the
manufacture of fiberglass batts and the f-:
-------
trailers, morticians, and residents of UFFI homes,
respectively. However, Main and Hogan did find significantly
increased symptoms of eye and throat irritation and headache and
fatigue among the exposed group. The residents of UFFI homes
experienced a high frequency of eye irritation and moderate rates
of nasal congestion and tearing when exposed to 1.0 ppm HCHO for
90 minutes in a chamber (Day et al., 1984). No significant
increase in respiratory disease was found in the morticians
studied by Levine et al.
In reviewing a number of morbidity studies, including the
Schoenberg et al. (1975) and Gamble et al. (1976) studies, the
Epidemiology Panel of the Consensus Workshop (1984) concluded
that:
No important reductions in forced vital capacity were
observed. Reductions in forced expiratory volume in
one second and forced expiratory volu.-.e (expressed as a
percentage of forced vital capacity) vhen observed were
small. These were not detected when exposure to
formaldehyde was solely as a vapor. There was either a
weak or absent association of reduced pulmonary
function tests with exposure in the few studies- where
this factor was analyzed. Workshift (acute) changes in
pulmonary function tests (PFT) have been assessed only
when other dust'was present and/or the formaldehyde
itself was a particuiate or incorporated in
particles. Acute PFT reductions 'were not consistently
present, were small and showed no regular association
with exposure. Although some symptoms were present,
the changes in PFT were clinically insignificant, and
there is no convincing evidence formaldehyde exposure
results in restriction or obstruction at the doses
studied. There is some suggestion that the symptoms
are reversible and of minor import. However, because
of the demonstated irritant potential of formaldehyde,
selection bias may be occurring in the exposed
populations so that these studies are likely to
underestimate adverse effects of formaldehyde exposure.
5-6
-------
Studies cited that were not available to the Panel do not
materially affect their findings.
Pulmonary edema and pneumonitis could result from very high
HCHO concentrations, 50-100 ppm. It is not known what
concentrations/durations are lethal to humans, but concentrations
exceeding 100 ppm would probably be extremely hazardous to most
and might be fatal in sensitive persons.(NRC, 1981).
5.1.5. Asthma
In addition to its iirect irritant effects on the
respiratory system, HCHO has been shown to' cause bronchial
asthma-like symptoms in humans (Hendrick et al., 1982; Surge et
al.,~ 1985; and Nordman et al., 1985). Although asthmatic attacks
may in some cases be due specifically to HCHO sensitization or
allergy, the evidence for this is less than certain (Consensus
Workshop, 1984). HCHO seems to act more :ommonly as a direct
airway irritant in persons -who have bronchial asthma from other
causes (Surge et al . , 1985 and N'ordman et al . , 1985). However,
the HCHO concentrations required to elicit such attacks are
relatively high, higher than would be expected in most
nonoccupational environments. For example, no
bronchoconstriction was observed in seven mild asthmatics who
were exposed to 1 ppm HCHO for 10 minutes at rest and to 1 or 3
ppm during mild exercise (Sheppard et al., 1984). In a study of
21 asthmatics living in UFFI homes, no consistent bronchial
effects were produced from three hour exposures to: Placebo,
0.54 ppm HCHO, UFFI particles 0.5/rnl, and HCHO free UFFI off-
5-7
-------
gases (Lees et al. , 1985). Witek et al. (1985) reports no
effects in healthy and asthmatic individuals exposed to 2.0 ppm
HCHO with and without mild exercise, and in a group of laboratory
workers routinely exposed to HCHO.
In a study of 230 persons who had been exposed to HCHO and
suffered asthma-like symptoms, 218 did not respond when
challenged with 2.0 ppm HCHO, including 71 subjects with
demonstrated bronchial hyperactivity (histamine or methachoiine
challenge test). The 12 individuals that did respond were
diagnosed as "true HCHO asthmatics" and all had been exposed
occupationally (Nordman et al., 1985). Finally, in a study of 15
workers occupationally exposed to HCHO who were exposed to
approximately 4.0 ppm HCHO under controlled conditions, six
workers developed immediate asthmatic reaciions, which were most
likely due to its irritation properties, while three workers
developed what was diagnosed as HCHO-causei hypersensitivity
(Surge et al., 1985).
Persons with bronchial asthma respond to numerous agents,
such as exogenous irritants and allergens, respiratory
infections, cold air, smoke, dust, and stress. The asthmatic
person seems to represent an extreme on the scale of respiratory
sensitivity to inhaled irritants. A paper by Brooks et al.
(1985) reports two cases in which asthma-like symptoms may have
been caused by a single exposure to high levels of an irritating
agent. Symptoms persisted for at least four years and were
accompanied by early inflammatory responses in the lung. Mo
5-8
-------
documented preexisting respiratory illness was found. The
authors have termed the illness observed reactive airways
dysfunction syndrome (RADS). Because many occupations have the
potential for episodic, high level HCHO exposure, RADS should be
considered as a possibility.
Diagnosis of immune sensitization has been based upon
knowledge that individuals were exposed to HCHO before the onset
of symptoms, reported complaints and symptoms, and spirometric
pattern on obstructive changes in respiratory function upon
bronchial provocation by inhalation challenge with HCHO.
Although the production of specific Immunoglobulin Type E (IgE)
antibody has been demonstrated to other chemicals (e.g.,
trimellitic anhydride, and nhthalic anhydride), IgE antibody has
not been shown to' be produced in response -3 HCHO exposure.
However, a study by Patterson et al. (1986; has demonstrated the
presence of IgE antibodies against HCHO-hunan serum albumin
conjugates and human serum albumin (HSA). The authors believe
the immunologic response is HCHO related because of a similar
pattern in dogs immunized with HCHO or HCHO-dog albumin and the
fact that anti-HSA antibodies have not been identified in
patients reactive to other hapten-HSA compounds. Respiratory
sensitization with HCHO has not been demonstrated with animals
(Lee et al., 1984). In some human studies in which patients
complained of respiratory illness, they did not respond
positively to bronchial challenge testing with HCHO gas, but it
does appear from the work of Hendrick et al. (1982), Surge et al
5-9
-------
(1985), and Nordman et al. (1985) that HCHO can induce
hypersensitivity by the inhalation route. However, the data
indicate that this may be a rare event. In addition, no data are
available describing induction concentrations, but it appears
that challenge concentrations as low as 1.0 ppm can elicit a
response (Nordman et al., 1985).
5.1.6. Summary
A number of lower airway.and pulmonary effects may occur
from HCHO exposure. Thresholds have not been established for the
irritant effects of inhaled HCHO. However, within the range of
0.1 to 3 ppm, most people experience irritation of the eyes,
nose, and throat (Consensus Workshop, 1984). In most healthy
persons exposed to HCHO, concentrations greater than 5 ppm will
cause cough and possibly a feeling of ches- tightness. In some
susceptible persons, concentrations below 5 ppm can cause these
symptoms, including wheezing. In persons with bronchial asthma,
the irritation caused by HCHO can precipitate an acute asthmatic
attack, sometimes at concentrations below 5 ppm. Although
conclusive evidence is not available, it appears'that HCHO is
capable of inducing respiratory tract allergy, but data are
lacking on induction concentrations. In concentrations greater
than 50 ppm, severe lower respiratory tract effects can occur,
with involvement not oniy of the airways but also of aiveoiar
tissue. Acute injury of this type includes pneumonia and
pulmonary edema. Finally, a dose-response analysis of the ha.^an
data appears in section 8.
5-10
-------
5.2. Irritation/Sensitization—Dermal and Systemic
In reviewing any analysis of respiratory effects, it is
important to remember that irritation and sensitization are two
distinct physiologic responses. Irritation is a purely local,
immediate response resulting from a chemical reaction between
HCHO and the epithelial lining. The irritant response will
resolve with cessation of exposure. It is scientifically
accepted that there is a threshold for the irritant response.
A chemical sensitization response is a far more complicated
physiologic effect. Some chemical sensitization responses are
mediated by the immunologic system, for others antibodies have
not yet been identified and the mechanism is as yet unknown. The
sensitization response may have one or more components, immediate
and/or delayed. The key distinction between sensitization and
irritation, is the absence of a clear threshold in the former.
Once an individual is sensitized, he/she v^il respond to low
effect-triggering exposures. There is debate in the scientific
community as to whether or not a threshold exists for the initial
chemical sensitizing event(s), but the data are not available to
resolve the issue.
It is established that HCHO is a primary skin sensitizing
agent producing allergic contact dermatitis. It is also probably
a cause of immunoiogic contact urticaria (Consensus Workshop,
1984).
HCHO induces allergic contact dermatitis by a delayed type
(Type IV) hypersensitivity mechanism. Besides contact with HCHO
5-11
-------
itself, allergic contact dermatitis can be caused by contact with
disinfectants and tissue preservatives containing HCHO, HCHO
releasers (resins in clothing, and paper products), and with
preservatives used for cosmetics, detergents, polishes, etc.
Table 5-2 illustrates some induction concentrations which
induce sensitivity and the range of challenge concentrations
which elicit the allergic reaction. The threshold for induction
has not been clearly defined, but it has been estimated as less
than 5 percent formalin in water. The appropriate threshold for
elicitation of allergic contact dermatitis in sensitized subjects
ranges from 30 ppm for patch testing to 60 ppm for actual product
concentrations of HCHO (formalin). However, because of the
limited data base these estimates should be used with caution
(Consensus Workshop, 1984). Data (induction and challenge
concentrations) regarding the ability of HIHO-resin treated
textiles to cause allergic contact dermatitis in garment workers
for instance are lacking.
5-12
-------
Table 5-2.
DELAYED TYPE HYPERSENSITIVITY (HUMAN SKIN) DUE
TO LOW LEVELS OF FORMALDEHYDE*
Induction Challenge Results (No.
Concentration Concentration Reacting Humans)
370 ppm 3,700 ppm 0/45
3,700 ppm 3,700 ppm 4/48 (4.5%)
11,000 ppm 3,700 ppm 5/58 (5.7%)
18,500 ppm 3,700 ppm 4/52 (7.7%)
Unknown 30 ppm 4/8 (50%)
(clinical) 60 ppm 5/8 (63%)
100 ppm 6/8 (75%)
10,000 ppm 8/8 (100%)
Unknown 32 ppm 0/14
55 ppm 2/14 (14%)
144 ppm 7/14 (50%)
*IRMC 1984a
The CIR Expert Panel (1984) stated that "the formulation and
manufacture of a cosmetic product should be such as to ensure use
at the minimal effective concentration of -ormaldehyde, not to
exceed 0.2 percent measured as free formal .iehyde. "
HCHO skin irritation is non-immunologic? how its mechanism
may differ from other forms of dermal irritation is not known.
Induction of contact urticaria by HCHO has been reported and
is presumably a Type 1 allergy (Consensus Workshop, 1984).
However, proof that the immunological reactions are due to an
allergic response must await the demonstration of specific immune
reactions such as the production of IgE or IgG antibody specific
for HCHO (IRMC Subgroup on Sensitization, 1984a). Nonimmunoiogic
contact urticaria which requires multiple applications at the
same site has been reported (Consensus Workshop, 1984).
5-13
-------
Sensitivity caused by the release of HCHO into the blood
from blood dialysis treatment has been reported. Frequent
eosinophilia (increase in eosinophil leukocytes) and some severe
hypersensitivety and asthma-like reactions have been associated
with this occurrence. Antibodies reacting with HCHO conjugated
red blood cells is evidence of Type II auto allergy. The asthma-
like reactions are suggestion of Type I allergy (Consensus
Workshop, 1984). However, commenting on this the IRMC Subgroup
stated that:
The hemodialysis patient population should not be
considered a source of IgE antibody since: (1)
formaldehyde levels present during dialysis have been
markedly reduced; (2) these reactions were due to
systemic exposure and primarily induced an antigenic
change in red blood cell surface markers; (3) only one
possible case has been reported of (anaphylactic)
sensitization by this route; this may resemble some
reactions caused by endotoxins present in dialysis
equipment. In this case patients were exposed to
allergens other than formaldehyde (personal
communication from Ronaid M. Easterl_r.g, M.D.).
However, a study by Patterson et al. (1986) has demonstrated
the presence of IgG, IgM, IgA, and IgE antibodies against HCHO
human serum albumin (HSA), but no correlations could be drawn
between the antibodies against HCHO-HSA and symptoms or
complication in patients using dialysis equipment sterlized with
HCHO.
5.3. Cellular Changes
Inhalation exposure to HCHO causes a number of cellular
effects depending on the concentration and duration of exposure.
In the Kerns et al. (1983) study, rats exposed to 2.0 oprr.
HCHO experienced rhinitis, epithelial dyspiasia, and squamous
5-14
-------
metaplasia after 12 months of exposure. The frequency of
squamous metaplasia increased to nearly 100 percent at the end of
the exposure period at 24 months. Considerable regression was
noted at 27 months (see Figure 4-1). In a more recent study by
Tobe et al. (1985), slight increases in rhinitis and squamous
metaplasia were observed in F-344 rats exposed to 0.3 ppm for 28
months and those found dead. However, the frequency of squamous
metaplasia falls within the 15 percent background rate for this
type of lesion as seen in the Kerns et al. (1983) study.
A study by Rusch et al. (1983), which measured similar
endpoints in monkeys, rats, and hamsters, reported a NOEL for
squamous metaplasia of 1.0 ppm. Table 5-3 clearly shows that a
threshold for this response exists at about 1 ppm (rats in the
Kern study experienced squamous metaplasia at 2.0 ppm). A
similar threshold level is suggested for -rrikeys as Table 5-4
indicates. Although, the authors did not attribute the one case
of squamous metaplasia to HCHO exposure, it is possible that HCHO
is causing effects, other than squamous metaplasia, at or beiow
1.0 ppn due to the increased incidence of nasal discharge in
monkeys at 0.2 and 1.0 ppm as illustrated in Table 5-5. Such a
response may be due to damaged cilia of the respiratory
epithelium. Data submitted by Woutersen et al. (1984b) on a
subchronic (13-week) inhalation toxicity study with HCHO in fats
(10 rats per sex at each dose) showed no squamous metaplasia in
the controls, 3 of 20 at 1 ppm, 20 of 20 at 10 ppm, and 20 of 20
at 20 ppm.
5-15
-------
Table 5-3.
Significant Findings in Nasal Turbinates
in Rats*
Group
I (combined
(controls)
II
III
IV
Level (ppm)
0
0.20
1.00**
3.00
Squamous
Meta/Hyperplasia
5/77
1/38
3/36
23/37
Basai Ceil
Hyperplasia
4/77
0/38
0/36
25/37
*Adapted from Rusch et al. (1983)
**NOEL
Group
I
II
III
IV
V
Table. 5-4.
Significant Findings in Nasal Turbinates
in Monkeys*
Level [ppm]
0
0.2
1.00**
0
3.00
Squamous Meta/Hyperplasia
0/6
0/6
1/6
0/6
6/6
*Adapted from Rusch et al. (1983)
**NOEL
Table 5-5.
Total Incidence By Groups of Monkeys*
Grou
(ppm]
Hoarse
Congestion
Nasal discharge
II
III
IV
V
(0)
0
0
9
(0.2)
0
0
30
(1.0)
0
0
45
(0)
0
0
5
(3.0)
32
36
62
*Adapted from Rusch et al. (1983). Out of a total of 156
observations per group.
5-16
-------
The effect of HCHO on nasal mucociliary function in the rat
has been studied by Morgan et al. (1986) (see Section
4.4.3.2.). Male Fischer 344 rats were exposed for 6 hours per
day for 1, 2, 4, 9 or 14 days, to 0.5, 2, 6 or 15 ppm HCHO.
There was a clear dose-dependent affect on mucociliary
activity. At 15 ppm there was significant inhibition of
mucociliary activity which progressed from anterior to posterior
regions of nasal tissue. Only slight effects were noted in
animals exposed to 2 or 6 ppm. At 0.5 ppm no effects were
observed.
The affects of HCHO on the human nasal system has been
studied by a number of authors. Anderson and Moihave (1983)
reported decreases in nasal mucus flow rates at air
concentrations as low as 0.38 ppm. In a study of five employees
of a sporting goods store in which pressed wood panels were used •
in the basement, Solomons and Cochrane (1934) report finding
nasal turbinate swelling in all five employees that persistent at
least four months beyond the point measures were taken to reduce
exposure to the point that no irritation symptoms remained.
Unfortunately, actual HCHO concentrations were not measured.
However, the lack of eye irritation may indicate that HCHO
concentration had been reduced to below about 1.0 ppm or nay
indicate tolerance to HCHO. Initial concentrations nay have been.
much higher (>3 ppm) because the employees could not stay in the
basement for more than a few minutes due to intolerable eye and
upper respiratory tract irritation, choking, and marked
5-17
-------
dyspnea. Lacroix et al. (1985) report on the clinical assessment
of 76 children who had been exposed to UFFI. Among the many
symptoms observed were abnormal nasal mucosa and nasal
secretions. Finally, in a study of workers processing
particleboard by Edling et al. (1985), it was found that the
group exposed (20 men) had a significantly higher nasal mucosa
histological score (points were assigned to eight factors
describing histological characteristics, e.g. loss of cilia,
keratosis, etc.) than a referent group of 25 men. HCHO exposure
levels were in the range of 0.1-1.1 ppm. Average exposure time
for the men was seven years. Five of the exposed group (25%) had
swollen or dry changes, or both, of the nasal mucosa. This was
characterized histologically as loss of cilia and goblet cells,
squamous metaplasia, and, in some instances, mild dysplasia.
In summary, it is clear that observable cellular changes
begin to occur above 1 ppm HCHO in animals, with the extent and
severity dependent on concentration and duration of exposure.
Based on data developed in rats and monkeys the NOEL for squamous
metaplasia and rhinitis can be placed at 1.0 ppm. The human data
indicate that mucocillary clearance system effects may be
occurring in humans at concentrations as low as 0.1 ppm, but data
in this regard are sparse. Consideration of the animal data
indicates that the rat model is a reasonable predictor of human
effects, even though a rat is obliged to breathe through its
nose, whereas a human is not.
5-18
-------
The practical consequence of the cellular changes noted is a
disturbance of the mucociliary clearance mechanism. Since this
system is an important defense mechanism in the removal of
particulate matter, including microbes, impairment of this
defense mechanism may increase the susceptability of persons to
infections and other respiratory diseases (Proctor, 1982; Comroe,
1974; Widdicombe, 1977). Reports suggesting that inhibition of
the clearance system may predispose certain children to
respiratory infections were reviewed by the Consensus Workshop
(1984). Their conclusion was that better designed studies are
needed to characterize this effect. In a study by Tuthill (1984)
of respiratory disease in children and woodstove use, it was
found that the strongest relationship of all study variables was
that of HCHO exposure. Excess acute respiratory illness was
significantly related to HCHO exposure. However, HCHO
concentration was estimated using parameters such as remodeling,
UFFI in walls, and mobile homes. Thus the results of this study
must be tempered by this and certain design limitations.
5.4. Central Nervous System Effects
Reports in the literature link HCHO with a number of
behavioral and physiologic effects such as thirst, dizziness and
apathy, inability to concentrate, sleep disturbances, etc.
Central nervous system (CNS) responses to HCHO have been tested
in a variety of ways, including determination of optical
chronaxy, electroencephalographicaliy, and by measuring the
sensitivity of the dark-adapted eye to light. Responses are
5-19
-------
reported to begin in some persons at 0.05 ppm and are maximal in
individuals at about 1.5 ppm. HCHO at less than 0.05 ppm
probably has little or no objective adverse effect (NRC, 1981).
However, in general, how HCHO affects the CNS is not clearly
defined (Consensus Workshop, 1984).
5.4.1. Neurochemicai Changes
Studies using radiolabeled HCHO have shown radioactivity in
the brains of rats after inhalation exposures. However, the
chemical identity of the radioactive material has not been
determined. It is unlikely to be HCHO because of its rapid
metabolism. Some kind of condensation product or labeled amino
acid from one-carbon metabolism may be present.
HCHO has been shown to affect the firing rate of
nasopalatine and ethmoidal nerves of the trigeminal nasal sensory
system. Besides being able to effect changes in the respiratory
rate of animals, HCHO also appears to be able to depress
trigeminal nerve response to other irritants, although the data
in this regard are not conclusive because of the testing
protocol.
HCHO (at high concentrations) has been reported to cause
cerebral acid proteinase activity in rats in one study and
decrease in cerebral R>JA concentration, together with decreases
in the succinate dehydrogenase and acid proteinase activities, in
another (Consensus Workshop, 1984). In a study by Boja et ai.
(1985) in which rats were exposed to 5.0 ppm HCHO, for 3 hours on.
2 consecutive days, levels of 5-hydroxyindoleacetic acid,
5-20
-------
3,4-dihydroxyphenylacetic acid, and dopamine were increased in
the hypothalamus. The toxicological significance of these
neurochemical changes is unciear.
Whether HCHO is capable of causing morphological changes in
the CNS is unciear. In two studies reviewed by the Consensus
Workshop (1984), conflicting results were seen. In one study,
structural and cytochemical changes were seen in the cerebral
amygdaloid complex of rats exposed to 1 to 3 mg/m^ of HCHO for 3
months. In contrast, monkeys injected intravenously over several
hours with HCHO for a total dose of 0.9 g/kg showed no
histoiogically detectable effects in the CNS.
5.4.2. Human Studies
Several reports are available which link chronic HCHO
exposure to a number of psychological/behavioral problems
including depression, irritability, memory loss and decreased
attention capacity, and sleep disturbances. Unfortunately, these
studies for the most part have involved field surveys using
subjective self-report symptom inventories. Control data are
often inadequate or completely absent. This is a significant
problem when dealing with HCHO, which in addition to any direct
toxic effects possibly associated with it, produces distinct
olfactory cues which may stimulate a spectrum of secondary
psychological reactions (e.g., expectancies, irritations,
anxieties, fears, etc.). These reactions may in turn exacerbate,
mask, or interfere with the more direct neurologic, biochemical,
and physiological responses to HCHO (Consensus Workshop, 1984).
5-21
-------
Nine studies of human populations were reviewed by the
Consensus Workshop (1984), but most had serious methodologic
problems. For instance in studies by Dally et al., Sardinas et
al., Garry et al., and Woodbury and Zenz, health complaints such
as headaches and difficulty in sleeping were linked to HCHO
exposure. However, these studies do not include control
populations and suffered from selection bias.
Thun and Aitman have pointed out some of the difficulties in
prevalence surveys of symptoms in residents from UFFI homes,
including olfactory cues, respondent and recall biases, and the
objective outcomes measured. Mo significant difference was found
in the occurrence of headaches or insomnia in residents of homes
with UFFI, compared to neighborhood controls.
In contrast, a study by Olson and Dossing found a
significantly greater prevalence of nose and throat irritation,
unnatural tiredness, and headaches in exposed subjects than in
controls. While this study overcomes many of the design problems
previously discussed, responses still may have been based by an
awareness of the subjects of the study goals and hypotheses.
Attempts have been made to evaluate reported symptoms using
formal tests of neuropsychological function. A study by Schenker
et al. found that persons living in UFFI homes who had 'complained
of memory impairment had negative results on formal tests of
memory function, although positive findings were seen for many
regarding attention span. In addition, a study by Anderson found
no effect on performance'tests of 16 healthy volunteers exposed
5-22
-------
to HCHO under controlled laboratory conditions (Consensus
Workshop). More recent studies by Kilburn et al. (1985a, 1985b)
of histology technicians showed disturbances of memory, mood,
equilibrium and sleep that occurred simultaneously with headache
and indigestion in 76 women technicians, while male technicians
were not appreciably different than a male comparison group.
When 25 technicians were evaluated using neurobehavioral tests
(block design, digit symbol, and embedded figure), a few deficits
were seen compared to expected results (IRMC Subgroup on Systemic
Effects, 1984b).
Commenting on the human data the Workshop Panel stated that
the information "suggests that formaldehyde may affect the
psychological functioning of the individual in three ways: (1)
directly, as a result of the immediate tox.c properties of the
substance on the peripheral and central nervous systems; (2)
indirectly, as a result of the individual's monitoring and
awareness of the aforementioned changes and his/her
interpretation and reaction to such changes, which, in turn,
feeds back into the central nervous system; and (3) as a result
of the individual's psychological reaction and concomitant CMS
response to the olfactory properties of the substance. In
practice, these processes are interdependent, yet this simple
analysis of a complex series of responses underlines the need to
control for 'expectancy' effects in formaldehyde research to
permit a differentiation of the direct effects of formaldehyde on
psychological functions from it secondary effects. "
5-23
-------
5.4.3. Conclusion
Based on the body of data available on the effects of HCHO
on the nervous system, the Workshop Panel concluded that:
The effects of formaldehyde .and/or its metabolites
on the biochemistry of the nervous system have not been
clearly defined. Various possibilities exist whereby
such effects might be mediated.
Some evidence exists that exposure to fomic acid
(the principal metabolite of formaldehyde) in vapor
form at high concentrations exercises nervous system
toxicity in intact rats.
The irritant effects of formaldehyde may be
reflected in altered function of sensory nerves such as
the trigeminal nasal sensory system. The presence of
morphological changes in the CMS has been observed in
one study and not in another.
The difficulties inherent in any study of
psychological/behavioral effects of formaldehyde have
not yet been overcome in the course of conducting field
surveys.
Epidemiologic studies evaluating neuro-
psychological symptoms potentially d'ja to occupational
or environmental exposure to formaldehyde have failed
to overcome the problems commonly associated with such
studies. However, some studies merit further
investigation.
5.5. Developmental and Reproductive Effects
5.5.1. Animal Studies
A number of studies have been reported which measured the
potential for teratogenic or reproductive effects of HCHO.
Ulsamer et al. (1984) reviewed four inhalation studies. No
teratogenic effects were reported. However, other effects in
dams and fetuses were reported such as, increased duration of
gestation and body weight of offspring, microscopic changes in
the liver, kidneys, and other organs of fetuses from exposed
5-24
-------
dams, and decreased Levels of nucleic in the testes of exposed
males.
• A dermal study by Overman (1985), reported that applications
of formalin to the backs of pregnant hamsters for 2 hours per day
on day 8, 9, 10 or 11 of gestation increased resorptions but did
not cause birth defects. The author speculated that the
increased incidence of resorptions may have been stress related
because of evidence that rats and mice subject to stress
experience increased resorptions, but no teratogenic effects in
the survivors (see Kimmel et al., 1976).
A study by Marks et al. (1980) was reviewed by the Consensus
Workshop (1984) which concluded that it was the only adequate
s-tudy (at that time) of possible teratogenic effects of HCHO in
mammals.
The Workshop review is as follows:
Marks and colleagues intubated pregnant mice on
days 6 through 15 of gestation with 0, 74, 148 or 185
mg/kg/day. At the highest dose, 22 of the 34 pregnant
mice died. At that dose, there was an increased
incidence of resorptions, but that increase was not
statistically significant. At no dose did the
incidence of resorptions differ between the treated and
control groups. There were also no treatment-related
differences in the mean number of implantations,
stunted fetuses, live fetuses per litter, or average
fetal body weight per litter. At a dose which killed
more than 50 percent of the dams, no adverse
reproductive outcomes were observed except for the
increase in the incidence of resorptions that was not
statistically significant.
To measure the teratogenic potential of HCHO generated in
vivo, a number of investigators exposed animals to hexamethylene-
tetramine by feeding or by drinking water. Studies by Delia
5-25
-------
Porta et al., Hurni and Ohder, and Natvig et al. were variously
reviewed by the CIR Expert Panel (1984), Ulsamer et al. (1984),
MRC (1981), and the Consensus Workshop (1984). No malformations
were noted in any of the studies.
Glycerol formal (GF), a possible slow HCHO-releasing agent,
has been reported to be teratogenic in the rat. However, Asby et
al. (1986) studied the hydrolysis of GF and its effects in a
mouse bone marrow micronucleus assay, which is known to be
sensitive to certain slow HCHO-releasing agents. No hydrolysis
was observed and the micronucleus assay was negative.
Consequently, the teratogenic activity of GF is unlikely to be
due to HCHO (Asby et al., 1986).
The Consensus Workshop (1984) reviewed studies of
reproductive effects. In one study, prolonged diestrus, but no
impairment of reproductive function was reported. Ovarian
involution and endomentrial atrophy were observed in another
study, but only in female mice exposed to 40 ppm HCHO (a
concentration which killed 80 percent of the animals). Other
studies were reviewed but were found to be of little value
because of methodologic problems.
5.5.2. Human Data
No data have been found linking HCHO to teratogenic effects
in humans.
5-26
-------
Review of reports of reproductive effects by the Consensus
Workshop (1984) and the IRMC Systemic Effects Subgroup (1984b)
did not lead to firm conclusions regarding HCHO's potential to
cause reproductive effects for a number of reasons. In a study
by Shumilina, workers exposed to urea-HCHO resins were reported
to have a threefold increase in menstrual disorders and produced
more babies weighing between 2500 and 3000 g than the controls.
The IRMC Systemic Effects Subgroup concluded that because of a
lack of information on the worker environment and the
socioeconomic conditions of the study and control groups, plus
the fact that other conditions such as stress and personal and
nutritional habits are associated with the effects reported, the
role of HCHO in the development of the reported disorders is
uncertain. In a better designed study, reviewed by the IRMC
Subgroup and the Workshop, Olson and Doss-r.g studied a group of
female workers in a mobile home day care center who were exposed
to 0.43 mg/m3 of HCHO. They reported increased incidence of eye
irritation, headache, and use of analgesics in the group. In
addition, 30 percent of the exposed group had a history of
menstrual irregularity. The Consensus Workshop (1984) felt that
these two studies point to the need for further research, but do
not show a causal relationship between exposure to HCHO and
menstrual disorders.
In two other reports reviewed by the IRMC (1984b), an
increased incidence of miscarriages, changes in menstrual cycies,
and an increase in ovarian cysts were reported in one study of
5-27
-------
female hiatotechnicians and a high incidence of sexual
dysfunction among male workers making fiber-reinforced plastic
was reported in another. In both instances, the workers were
exposed to chemicals other than HCHO, especially the male
workers. Consequently, it is uncertain whether the effects
reported are due to HCHO itself, to another chemical agent, or to
the interaction of numerous chemicals.
The Consensus Workshop (1984) reviewed three studies related
to the potential of HCHO to cause germ-cell mutations. A study
by Fonlignie-Houbrechts reported increased pre- and post-
implantation losses in the first week of mating, following
exposure of male mice to 50 mg/kg of HCHO by injection, and an
increase in preimplantation loss in the third week. No evidence
of increased dominant lethal effects were seen in a study by
Epstein et al. where mice were exposed at loses of up to 40
mg/kg, IP. Finally, Cassidy reported increased sperm
abnormalities in rats exposed to a 200 mg/kg, but not in rats
given 100 mg/kg orally. "Thus the data are not consistent and do
not adequately test the possibility that formaldehyde causes
germ-cell mutations" (Consensus Workshop, 1984). However, these
data may not be inconsistent given different routes of
exposure. More work in this area may be needed.
5.5.3. Conclusion
Ulsamer et al. (1984), the Workshop, and the IRMC Subgroup
concluded the following regarding the potential of HCHO to cause
teratogenic or reproductive effects.
5-28
-------
Ulsam«r at al.:
Th« currently available data do not show that the
embryo La unusually sensitive to formaldehyde nor is
there any information to show that formaldehyde is
teratogenic in rodents when administered orally or
applied dermally in nontoxic amounts to the dams.
Also, the in vitro data do not provide any evidence to
support the conclusion that formaldehyde causes terata
at exposure concentrations that are not toxic to the
adult.
Inhalation of formaldehyde has caused fetotoxic
effects but not teratogenic effects. Further studies
of formaldehyde exposure by inhalation are needed to
elucidate the meaning of these changes. Limited
evidence suggests that formaldehyde may affect the
menstrual cycle and perhaps reproduction in women
repeatedly exposed. Additional work is needed to
validate these findings.
IRMC Subgroup:
Reproductive function depends upon a sensitive and
integrated feedback system between the nervous system
and the reproductive organs. Thus, many chemicals that
affect the nervous system have the potential to affect
reproduction. It is possible that formaldehyde, by
affecting the nervous system induces indirect changes
in reproductive behavior and reproduction. Although
mechanisms for such have not been delineated, several
recent reports that show an increase in the incidences
of brain tumors in humans exposed to formaldehyde
provide indirect evidence of the potential of
formaldehyde to significantly affect the CNS.
5-29
-------
Consensus Panel:
In summary, the panel could find no evidence
clearly demonstrating that formaldehyde caused adverse
reproductive outcomes. What it found was a paucity of
information from which to make inferences and data that
suggested hypotheses to be tested in future studies.
This panel feels that formaldehyde poses little, if
any, 'risk as a potential human teratogen. This
judgment is based on the irritation potential of
formaldehyde at extremely low ambient concentrations
(0.05 ppm), existing data from in vivo mammalian
studies, and toxicokinetic and metabolism data
indicating an extremely short half-life (not detected
to 1.5 min) of the parent compound, and relatively
short half-life (80 to 90 min) of the only known
metabolite (formate) in the blood, regardless of the
route of exposure.
5.6. Effects on Visceral Organs
The potential effects of HCHO on visceral organs has
received relatively little attention. One recent review article
by Beall and Ulsamer (1984) summarizes the association between
exposure to HCHO and effects on the liver. Based on the
literature reviewed, it appears that HCHO causes hyperemia or
inflammation in liver and kidney in rats. Microscopically, HCHO
also causes cloudy swelling, cytoplasmic vacuolization, and
necrosis in the liver, and hyperemia, edema, and necrosis in the
kidney. Macroscopic changes in the liver have also been produced
by HCHO. When exposure is repeated over a period of weeks,
changes include a mottled appearance and a decrease in liver
weight. Following a single high exposure, liver size may
increase. Effects on viscera could result from indirect
mechanisms or be secondary to other effects near the site of
first contact. Under some circumstances, GSH detoxification
mechanisms may be involved (IRMC, 1984b).
5-30
-------
Transient'effects on the hematopoietic system occurred in
rats and nice after 6 months of exposure to HCHO by inhalation.
These effects were reflected by statistically significant
decreases in (1) reticulocytes in female mice exposed to 2.1, 5.6
or 14.3 ppm; (2) mean corpuscular hemoglobin in male and female
rats exposed to 14.3 of HCHO; and (3) mean corpuscular hemoglobin
concentration in male rats exposed to 2.1, 5.6 or 143. ppm of
HCHO. Male and female rats had significant (p<0.05) increases in
mean corpuscular hemoglobin, mean corpuscular hemoglobin
concentration, and myeloid to erythroid ratios after 13 weeks of
exposure by inhalation to 12.7 ppm of HCHO. This could indicate
myeloid hyperplasia or erythroid hypoplasia. Thus, it is
possible effects on visceral organs could be partially caused .
through changes in the hematopoietic system as well as through
other mechanisms (IRMC, 1984b).
Gibson (1984), in reply to Beall and Ulsamer (1994), notes
the absence of any hepatic changes attributed to HCHO in the
Kerns et al. (1983) (CUT study) study. Also, other than effects
in the respiratory systems of rats and mice, HCHO has not been
shown to cause toxic effects at other sites. In the Tobe et al.
(1985) study, where rats were exposed to 15 ppm EiCHO for 28
months, no changes other than in the respiratory system could be
attributed to HCHO exposure. A decrease in liver weight was
noted, but this was assumed to be caused by a decrease in food
intake, not by a direct toxic effect.
5-31
-------
Addressing the issue of systemic effects, the Consensus
Workshop (1984) stated that "There is no convincing evidence in
experimental animals that inhalation exposure causes significant
primary toxicologic effects in organs other than the upper
respiratory tract."
5-32
-------
6. EXPOSURE ASSESSMENT
6.1. Introduction
The sources of HCHO can be grouped into two major
cateqories: commercial production and indirect production. The
chemical is not imported in any appreciable quantities.
Commercially, HCHO is produced from the catalytic oxidation
of methanol, using either silver oxide or a mixed-metal oxide as
the catalyst. Processes accounting for the indirect production
of HCHO include the photochemical oxidation of airborne
hydrocarbons released from incomolete combustion processes, the
production of HCHO during incomplete combustion of hydrocarbons
in fossil fuels and refuse, and certain natural processes.
The 1984 commercial production of HCHO amounted to about 6
billion pounds. The major derivatives are urea-HCHO resins,
phenol-HCHO resins, acetal resins, and bu.anediol. The urea- a-.;
phenol-HCHO resins account for about 53 percent of HCHO
production. Adhesives and plastics are the major end uses.
The "consumption" of HCHO can be broken down into three
major categories: nonconsumptive uses, pseudo-consumptive uses,
and consumptive uses. In nonconsumptive uses, the chemical
identity of the HCHO does not change. In pseudo-consumptive
uses, the chemical identity of HCHO does change, but it is not
irreversibly altered. Under appropriate conditions, some or all
of the original HCHO may be regenerated. Consumptive uses, on
the other hand, are those uses in which HCHO serves as a
feedstock for the preparation of other chemicals. The
6-1
-------
derivatives are irreversibly fomed and usually contain only
residual levels of unreacted HCHO. Under extreme conditions,
such as very high temperatures or highly acidic conditions, some
of the derivatives may deqrade and release HCHO.
HCHO's major nonconsumptive uses are (1) disinfectant, (2)
preservative, (3) deodorant, and (4) textile and paper uses.
The major pseudo-consumptive uses are (1) urea-HCHO resins
which are used in fiberboard, particleboard, plywood, laminates,
urea-HCHO foams, molding compounds, and paper, textiles, and
protective coatings; (2) urea-HCHO concentrates which are used to
produce time-release fertilizers, and (3) hexamethylenetetramine
which is used as a special anhydrous form of HCHO to cure resins
and to treat textiles and rubber.
The major consumptive uses are (1) melamine-HCHO resins
which are used for molding compounds, fiberboard, particleboard,
plywood, laminates, paper and textiles, (2) phenol-HCHO resins
which are used in fiberboard, particleboard, plywood moldina
compounds, and insulation; (3) nentaerythritol which is used to
oroduce alkyd resins, (4) 1,4-butanediol which is used to produce
tetrahydrofuran/ (5) acetal resins which are used in the
manufacture of engineering plastics, and (6) trimethyloloropane
which is used in the production of urethanes.
6.2. Estiaates of Current Human Exposure
To obtain estimates of human exposure to HCHO, the Agency
commissioned a contractor study (Versar, 1982). This studv
integrated the existing monitoring data, engineering or modeling
6-2
-------
estimates, use data, population estimates, and assessment of the
likelihood of exposure from HCHO-related activities into an
exposure assessment detailing those activities having a high HCHO
exposure potential. EPA updated some portions of this assessment
to reflect new data received in response to the FEDERAL REGISTER
notice of November 18, L983 and other data gathered by EPA. The
combined data were used as the basis for the Hay 1985 draft risk
assessment.
Subsequent to the draft risk assessment, the Agency
commissioned additional contractor studies to assess garment
worker (PEI, 1985) and residential (Versar, 1986a,b,c) exposure
to HCHO in more depth. The exposure estimates from these reports
were used as the primary basis for this risk assessment. The
conclusions of these contractor reports are summarized in this
document; more detailed information regarding exposure can be
obtained by referring to the contractor reports.
6.3. Populations at Risk
The two populations at.risk examined here are certain home
residents and garment workers.
6.3.1. Home Residents
Based on a projection of manufactured housing starts by
Schweer (1987), it is estimated that 7,800,000 persons may occupy
new manufactured homes during the next ten years. This figure
assumes 295,000 starts per year and 2.64 persons per home.
Similarly, an estimated 214,000 new conventional homes
containing significant quantities of pressed wood products as
construction materials will be started each year for the next ten
6-3
-------
years with an occupancy rate of 2.95 persons for a total of
6,310,000 persons.
6.3.2. Garment Workers
The number of potentially exposed garment workers is
estimated to be 777,000 (Versar, 1982) out of 1,100,000 workers
employed in the U.S. apparel industry (Ward, 1984). This figure
may drop in the future due to increased foreign competition and
the introduction of labor saving equipment.
6.3.3. Summary
Table 6-1 presents population estimates for the two housing
segments. Assuming that the number of potentially exposed
garment workers remains steady at 777,000, then a total of almost
15,000,000 persons over the next ten years may have the potential
to be exposed to elevated levels of HCHO.
Table 6-1.
POPULATIONS AT RISK
Population
Category Estimates
per yr 10 yrs
Manufactured homes 779,000 7,790,000
Conventional homes 631,000 6,310,000
* Schweer (1987)
6-4
-------
6.4. Sources of HCHO in Population Categories of Concern
The principal sources of HCHO in the two population
categories of concern are HCHO-based resins, principally urea-
HCHO (UF) resins. In homes, these resins are used to bond the
wood plys used to make plywood and to bind the wood particle and
fibers used to make particleboard and medium density
fiberboard. For garments, HCHO-based resins are used to impart
permanent press finishes to the garments.
6.4.1. Homes Containing Pressed-Wood Products
6.4.1.1. Pressed-wood product descriptions
Pressed-wood products are used in flooring, interior walls
and doors, cabinetry, and furniture. The three principal types
of products containing UF-resin are particleboard, medium-density
fiberboard (MDF), and hardwood plywood.
Particleboard is composition board comprised of 6 to 10
percent resin (by weight), and small wood particles. UF resin is
used in the majority of particleboard (about 90 percent of total
production capacity). The 1983 production of particleboard was
over 3 billion square feet, of which 70 percent was used in
furniture, fixtures, cabinets, and similar products. The
remaining 30 percent was used for construction, including deckina
in manufactured home manufacture and flooring underlayment in
conventional housing.
Recent data indicate that particleboard is used in home
construction at a rate of 0.16 square feet (ft2) (~ 0.5 m2) ner 1
cubic foot (ft^) of indoor air volume in mobile homes. The
6-5
-------
loading rate (ft2/ft3) in conventional homes is lower on average,
approximately 0.05 ft2/ft3 (~ 0.17 m2/m3) (see Table 6-2).
However/ loading rates in conventional homes may vary
considerably from homes that contain only particleboard as a
cabinet material to homes whose floors are constructed with
particleboard underlayment.
MDF is also a composition board. It is comprised of wood
fibers and 8 to 14 percent UF resin solids by weight.
Approximately 95 percent of MDF production (over 600 million
square feet in 1983) was used to manufacture furniture, doors,
fixtures, and cabinetry. No data are available on the precise
extent of MDF's use in either mobile or conventional homes.
Unlike the two composition boards discussed above, hardwood
plywood is a laminated product; the resin is used as a glue to
hold thin layers of wood and veneers toqether. Of the nearly 4.3
billion sguare feet consumed in 1983, 55 percent was used for
indoor paneling, 30 percent for furniture and cabinets, and 15
percent for doors and laminated flooring.
6-6
-------
Table 6-2 . Use of Pressed-wood Products
in
Construction
Cateoorv
Mew Hoses (U.S.)«'b
Percent units containing
Hardwood plywood paneling
Particleboard underlayment
Average loading rates,6
Hardwood plywood paneling
P«rticltbo*rd und«r Uyntnt
Partidrtotrd shelving
P«rticl«too*rd kitchtn c«t)intts
ToUl p«rticl«too*nj
Homts
Pvrctnt units containing
P«rtidflOO*rd
Avtragt loading rates (nr^/m3)
ToUl particlcboard
Existing Hants (U.S.)*
Percent units containing .
Hardwood plywood paneling
Particleooard
Average loading rate (nrVm3)
HardMod plywood paneling
Particleooard
SFO
7.6
30. 5
0.066
0.118
0.010
0.039
0.167
100
35.5
90.3
0.096
0.058
of
TH
9.3
9.2
O.OS9
0.092
0.016
0.052
0.160
100
0.145 0.100
8.5 most
1.7 most
0.049
0.033
0.020
0.059
0.112
1.0
O.S
100 100
0.079 0.479
most
most
1.0
0.5
Qita reflect only interior uses of UF pressed wood products.
Loading rates are for those hones containing these products.
.•Source: NPA (1964) and H?m (1984) for conventional homes - Based on
interpretation of the results of a survey of 900 home builders (103
responses) regarding the extent of use of particleboard and harduaod
plywood paneling in new nones containing these products (NAHB 1984).
Source:
hones.
Meyer and Hermanns (1984a), NAH8 (1984). MHI (1984) for mobile
(Footnotes continued on next page)
6-7
-------
Table 6-2. Footnotts (continued)
**? of product surf act arta/m* of indoor air volunt.
<%ourct: InttrArt (1983) - bastd on in-hont survtys it 9 SFO, 1 TN. 1
And 1 W. Total loiding includts undtrUywtflt. stw1«in9 and
SFO loadings rangtd from 0.028 to 0.491
'Source: Scnuttt (1981) - Bas«d on in-hoi» survtys at 31 SFO. Avtragt
loadings bastd on tan»s containing thts* products.
f SFO • Singlt family dMtlling
TH • TOHnhOUM
IV • *j1tifamily dMtlling
PW . No&ilt
6-8
-------
6.4.1.2. HCHO release from pressed-wood products
Each of the pressed-wood products described above contain UF
resins which release HCHO over time. The release is attributable
to two basic sources (Podall, 1984):
1. • Free (unreacted) HCHO present as a result of incomplete
crosslinking during resin cure.
2. Decompos.it ion of unstable UF resin or resin-wood
chemical species as a result of their intrinsic
instability and/or due to hydrolysis.
Free HCHO, which is present in cured resin, at low levels (<1
percent) is the most significant source of HCHO release from
pressed-wood products in the initial period after they are
manufactured (Podall, 1984). The specific time period in whicti
free HCHO dominates releases is not known.
The second source, decomposition and hydrolysis, pertains to
the large proportion of HCHO-bearing species like methylene
ureas, urea methylene ethers, and cellulose-crosslinked species
that may release HCHO for a much longer period of time (Podall,
1984). These species differ in their susceptibility to
hydrolytic attack and decomposition, and their relative rates and
durations of release can only be hypothesized at this time.
Release of HCHO from UF-resin containing pressed-wood
products is complex/ with numerous interrelated aspects. The
pressed-wood product manufacturing process, and other factors,
affect the amount of each HCHO-releasing species present in the
finished product. The resin formulation has a direct effect on
6-9
-------
release; resins with a low HCHO:urea ratio have, when cured, a
lower level of free HCHO but may be less stable and more
susceptible to hydrolysis (Myers, 1984). Other additives to the
resin, such as acid catalysts, change the resin chemistry and
influence the release profiles. The conditions under which the
resin is cured affect bond strength, determining to some extent
the stability of the resin components. The character of the wood
itself also affects HCHO release; the more acidic the wood, the
greater the tendency for acid hydrolysis and HCHO release
(Podall, 1984).
Under normal use conditions, the release of HCHO decreases
with time, as discussed previously. Emission reductions linked
to product aging relate to a decrease over time in both the HCHO
present in the board as a residual from manufacturing and the
latent HCHO present in the board in hydrolytically labile resin
and wood components. The emission rate decay curve for a board
is apparently exponential with time; the residual HCHO is emitted
at relatively high rates followed by a slow release of latent
HCHO. Althouoh the short-term emission rate behavior of boards
has been reported in numerous studies, little quantitative
information is available on the long-term emission rates,
particularly for newer products made with low HCHO-urea ratio
resins or treated with scavengers.
6.4.1.3. Other Sources of HCHO
Indoor HCHO concentrations may be attributable to sources
other than pressed-wood products containing UF resin. The other
sources can be characterized as follows:
6-10
-------
o Urea-HCHO foam insulation (UFFI) (existing homes only)
o Products with phenol HCHO resins (PF)
softwood plywood
hardboard
wafarboard
oriented strand board
fibrous glass insulation
fibrous glass ceiling tiles
o Consumer products that may contain HCHO resins
upholstry fabric
drapery fabric
other textiles
o Combustion products
unvented kerosene and gas appliances
smoke from tobacco products
combustion of wood or coal in fireplaces
o Ou td oo r a i r
ventilation system air exchange
Compared to pressed-wood products, with the exception of
UFFI, the other sources are usually minor contributors to HCHO
concentrations in conventional and manufactured homes.
The Consumer Product Safety Commission (CPSC) in 1982
prohibited the installation of UFFI in residential buildings and
schools. Although it was later overturned by a Federal court,
the CPSC ban on UFFI caused the virtual elimination of the UFFI
industry (Formaldehyde Institute, 1984). There is considerable
debate among the regulatory agencies'and the UFFI industry as to
the extent of long-term HCHO emissions from UFFI presently in
place (Hawthorne et al., 1983). UFFI is not discussed in detail
in this section; refer to Versar (1986c) for further information
and references.
Though no residential sources of HCHO have been as well-
studied as urea-HCHO foam insulation and pressed-wood products
made from UF resins, there are fairly complete data on the
6-11
-------
importance of pressed-wood products with PF resins, on fabrics
treated with UF resins for permanent press, on fueled appliances,
and on cigarette smoke as sources of residential levels.
Common applications of PF resin pressed-wood products
include roof and wall sheathing, subflooring, and siding. Small
amounts are used for shelving, cabinets, indoor paneling, and
fixtures (APA, 1934). Pnenol-HCHO resins are inherently nore
stable than are UF resins, and pressed-wood products nad-e of ?F
resin emit HCHO at much lower rites than do products made with UF
resins. The small amount of HCHO that is emitted from the panel
products is the result of residual HCHO that remains in the resin
(APA, 1984).
There are several published studies on HCHO emissions from
PF pressed-wood panel products. Myers and Nagaoka (1981) found
that HCHO levels in chamber tests rarely exceeded 0.1 ppm in the
presence of PF particleboard at 25°C. Matthews et al. (1983,
reports X-XV) tested PF hardboard and softwood plywood and
obtained similar results. Myers (1983) measured higher levels
(0.3 ppm) initially in tests of waferboard and particleboard made
with PF resins, but levels declined rapidly. The American
Plywood Association (APA,. 1984) has submitted data (reviewed by
Versar, 19S6c) indicating that PF-resin pressed wood products
>
emit little HCHO.
Other generic product lines containing PF that are used in
construction applications are fibrous glass insulation and
6-12
-------
ceiling tiles. In 1983, as a result of a study on HCHO release
from consumer products (Pickrell et al., 1982), CPSC decided to
further evaluate HCHO emissions from fibrous glass insulation and
ceilinq .tiles. These products, when compared with other products
tested, 'were among the highest group of emitters tested by
Pickrell et al. (1982). Concern about.these test results arose
because of the high loading rates of these products in homes.
Under normal use conditions (in attics), insulation would be
subjected to temperatures much higher than normal room
temperatures, thereby increasing potential HCHO emissions.
Further evaluation by Matthews et al. (1983) and Matthews
and Westley (1983) (under contract to CPSC) indicated that a
predicted increase of no more than 0.022 ppm in indoor HCHO level
would result from use of new ceiling tile = or new insulation. As
the products age, the HCHO emission rates and resulting indoor
concentrations would be expected to decline significantly.
Available data on treated fabrics (Pickrell et al., 1982,
1984) indicate that, with emission rates only as high as 115
ug/m2/hr, these can be relatively important sources in homes only
with large surface areas of furnishings like draperies (at least
when new). The data on combustion appliances show that HCHO
release is a function of whether the appliance is tuned and
functioning properly. Gas stoves may emit less than 2 to nearly
30 mg HCHO per hour of use; gas heaters can emit less than 5 to
over 60 mg/hr, depending on the efficiency of burning; and new
kerosene heaters emit up to of 3.9 mg/hr of HCHO (Traynor et al.,
6-13
-------
1982; airman et al. , 1983; Fortmann et al., 1984; Traynor et al. ,
1983; Caceres et al, 1983).
The emissions data on sidestrean cigarette snoke range from
20 ug per cigarette (Bardana, 1984) to nearly I.5 mg/cigarette
(reported by Matthews et al., 1984). Several studies, however,
concur on an emission rate of 1.0 to 1.2 mg/cigarette. The
importance of this source is obviously related to use patterns.
Studies where numerous persons chain-smoked in a poorly
ventilated room (Timm and Smith, 1979) did indeed show that HCHO
levels were elevated after a short period of time, but other
studies (Traynor and :;itschke, 1984) in the homes of smokers
indicated that, at a smoking rate of 10 cigarettes per day, HCHO
levels were not elevated over controls with similar loading rates
of other sources.
6.4.2. Garment Manufacture
The principal source of HCHO in the garment manufacturing
workplace, is the release of HCHO from fabric treated with resins
that impart durable or permanent press properties. The textiles
normally treated are blends of cotton, acetate, and rayon. These
fabrics account for 60-80 percent of the textile produced
annually.
The resin of choice is dimethyldihydroxyethylene urea
(DMDHEU) and its alkylated derivatives. It is estimated that
approximately 90 percent of the durable press resin market is
accounted for by DMDHEU. Other resins used are urea-HCHO,
melamine HCHO, and carbamate resins, plus a HCHO/sulfur dioxide
vapor phase process.
6-14
-------
HCHO is released from treated fabric in three phases. In
Phase I, any HCHO loosely held by Van der Waal forces is released
as the fabric is dried. Release of HCHO by this mechanism is
usually complete by the time garment workers receive the
fabric. Surface desorption occurs during Phase II. This
represents the release of HCHO which is not covalently bound to
the fabric, and can last up to 240 hours. The material is
normally stored during this phase, and increased ventilation can
increase the rate at which HCHO is desorbed. Phase III, in which
hemiacetal hydrolysis is the mechanism of release, is thought to
be the phasa of HCHO release which results in worker exposure at
the manufacturing site. Release of HCHO by the hydrolysis
mechanism is independent of air changes, but dependent on
humidity and temperature (Ward, 1984)
6.5. HCHO Levels in Homes and Garment Manufacturing Sites
6.5.1. HCHO Levels in Homes
Table 6-3 briefly summarizes the residential HCHO monitoring
studies reviewed by Versar (1986a, c). .However, because of the
changing nature of pressed-wood products with UF resins and the
constant evolution and improvement in monitoring techniques, this
residential monitoring data base is not the most appropriate for
describing current HCHO exposure in homes. Many data sets are
based on investigation of homes from which complaints of HCHO
symptoms have been filed; these data sets may not be
representative of average exposure because of bias toward high
concentrations. Homes studied before 1980 were built with
6-15
-------
Table 6-3. Sumtary of Residential Formaldehyde Monitoring
Study/sailing dite(s)
(reference)
Nujfcerof hows
(ppn) or
Nu*er of samples range of neans
Ran«e (pp.)
Coments
COMVENIIOMAl HOMES
Flaming A Associates
New York Study
(Iraynor A Nitschke 1964)
Univ. Washington (1982-1963)
(Breysse 1964)
IBL (1979-present)
(Ginun et al. 1963)
30
59
24 (EE)
16 (W)
113
0.040
0.007-0.IS)
<0.I to > 1.0
<0.005 to 0.214
<0.005 to 0.079
Non-complaint hows.
Prinarily complaint hones. Only 4 of II)
•easuranents >0.5 ppn.
Includes energy-efficient (It) and older.
weatherized (W) non-complaint hones around
the U.S.
o\
Geonet (1978)
(Hoschandreas et al. 1978)
17
-714
a\
Canadian UFFI/ICC (1981)
(UFFI/ICC 1981)
ORNl/CPSC 40 lennessee hone
Study (1982)
(Hawthorne et al. 1984)
Private Washington labs (1983)
(Breysse 1984)
UK study (-1980-1982)
(Everett 1983)
378
29
25 specified
conventional
SO
76
0.02 to 0.16 0.25
<0.05 to >0.5
Includes conventional, "experimental", dm
apart«ent hones around U.S. Non conpldini
hones. Assuming 601 of total aldehydes i->
fonMldehyde.
Study of Uff I and non-UFf I hones; wean is
for non-UFFI hones. (Mean M/UFFI = 0 OV1
ppn for 1.897 hones).
Study of UFFI and non 4JTFI homes; nedn is
for non-UFFI hones. (Mean w/UFf 1 = 0 .Oil
ppn for II hones).
None exceeded 1.0 ppn. 45 of 76
0.05 and 0.09 ppn.
Study was of UFFI and non-UTf I hones, mt-jo
is for non-UFFI hones. (Mean w/UFfl , U on
PP»)
-------
Table 6-3. (continued)
Study/sampling date(s) _*_•
(reference)
Dutch study (19/1-1980)
(Van der Ual 1982)
Houston, lexas study (1980)
(Stock and ftendei 1985)
a\
.i Sacrawnto. California (1982)
-J (Uagner 1982)
San Francisco, California (1984)
(Sexton et al. 19856)
Iowa study (1980)
(Schutte et al. 1982)
SAI California survey (1984)
(SA1 1984)
— • • _
ber of hones lumber of sables
— — — __
36
5 _
38 conventional 38
1 energy-efficient 1
conventional
19 aparUnents 19
II condoniniuns ||
3 energy-efficient 3
condoniniiMs
12 i;
SI 5)
31 3/2
6 _
64
.
Mean (pp») or
range of Means
--
—
0.04
0.0?
0.08
0.09
0.18
0.106
0.038
0.063
0.084
0.050
Range (pp»)
_
0.032 to 1.444
(range of
•axiauns)
0.048 to 0.602
<0.008 to 0.14
0.04 to 0. II
0.02 to 0.21
<0.008 to 0.29
0. IS to 0.20
0.0/8 to 0. 163
0.013 to 0.085
0.013 to 0.34
0.046 to 0. 153
0.018 to 0. 120
Coments
Prior to control i^loMUtion. largely
conplaint hones.
After panel coating.
Non-conplaint hones; no UFFI.
All hones less than 5 yrs old. All hcnis
«ith air exhange rates less than 0.5 per
hour.
Only 2 hones less than 6 yrs old.
39 h«es «ore than 10 yrs old.
Non-canplatnt hones.
Non-conplaint hones.
New. non -coqplaint hones.
Older, non conplaint hones.
Indiana Board of Health
study (19f9-1983)
(Konopinski 1983)
120
120
0.09
NO to 1.35
Study of UFfI and non4TF I homes; mean is
for non-UTFl nones; includes some conpUmt
notes. (Mean M/UTFI - 0.05 pan for 119
hones).
-------
Table G-3. (continued)
Study/sanpling date(s)
(reference)
Mean (pp») or
Muter of hows Muter of samples range of Means
(pp»)
Cowents
Godish (1983)
Conn (1961)
Swiss hows (1963)
(Kuhn and Uanner 1984)
Netherlands study (1961-1982)
(Cornet 1983 - Holland study)
Swedish horns (1975-19/7)
(Sundin 1978)
Danish hows (1913)
(Andersen et al. 19)5)
MOBILE HOMES
Geowt (19/8)
(Itoschandreas et al. 19/8)
Univ. Washington (1982-1983)
(Breysse 1984)
NHI (1984)
(Owners 1984)
Clayton (1900 I9UI)
(Sinyh et al. I9fl2a)
29
103
46
52
319
23
430
259
84
822
15
0.05
0.021
0.048 to 0.055
0.58
1.44
0.21
0.34
0.62
(adjusted)
0.03 to 0.0) Study of UfFI and non-UFFI hows; mean i-,
for non-UFFI hows containing no
particleboard flooring, cabinetry or
paneling. (Mean M/UFF1 = 0.01 ppm for ?B
hows).
~0.I to O.I Highest level prior to occupancy
— Hows without particleboard, as measured l>y
the chromotropic acid Method.
O.I to 2.0 Few details available.
0.07 to 1.8) HOWS known to have particleboard
construction Materials.
0.01 to 0.46 Assuming 601 of total aldehydes is
forMaldehyde. Non-ccMplaint hows
<0.l to >I.O 37 of 822 wasurewnts >I.O pp. c«plai..i
hows.
0.24 to 0.46 3-Month old how built specifically (or i,-.i
0.02 to 2.9 Mon complaint, occupied and noiioccupied
(adjusted) Concentration by how age evaded
-------
Table 6-3. (continued)
Study/sailing date(s)
(reference)
ftartwr of hows
Mean (ppm) or
r of samples range of means Range (pp.)
Coments
Wisconsin (I960)
(Anderson et al. 1983)
Minnesota (1960-1961)
(Stone et al. I960
lennessee (1962-1963)
(Hodges 1964)
Kentucky (19/9-1960)
(Conyers 1964)
lexas study (1962-1963)
(Univ. lexas 1963)
SA1 California survey (1964)
(SAI 1964)
California State survey (1964)
(Sexton et al. I985a. I985t»)
131
109
55
103
121
920
663
663
0.36
0.61
0.30
0.23
0.43
0.18
0.114
0.091
0.02 to 2.26 Mon-complaint. occupied h«es.
Concentration by hoje age evaluated
ho"e «9» <* yrs. Complaint hones
0.02 to 1.43 Complaint hones; no age data.
0.02 to 1.92 Complaint homes, see Idble 21 for data by
how age.
0.01 to 1.99 Complaint hows, see Idble 28 for data by
home age.
0.04 to 0.35 Non-co^>laint homes. Excludes results (run
one county (El Paso) where evaporative
coolers Mere in use.
0.068 to 0.144 Passive LBL sampler; one Meek; non
-------
with products made of high HCHOrurea ratio resins that are no
longer on the market; they cannot be considered as baseline
exposures for that reason. The most appropriate data for
describing current exposures in mobile and conventional homes
are, therefore, those generated by random sampling of
noncomplaint homes after 1980, preferably after 1982 (when
manufacturers began using resins with mole ratios of 1.5 (F:U) or
less). These restrictions on the "appropriate" data base still
leave a considerable volume of monitoring data on levels in
homes. Table 6-4 summarizes the noncomplaint (random) data on
HCHO levels in conventional and mobile homes.
6.5.2. Manufactured Homes
HUD has recently promulgated changes in its Manufactured
Home Construction and Safety Standards (24 CFR 3280). The
changes, published in the FEDERAL REGISTER of August 9, 1984 (47
FR 31996), set product emission standards for particleboard (0.3
ppm) and plywood (0.2 ppm). HUD believes that if the product
standards are met and no other major emitters of HCHO are present
(e.g., medium density fiberboard), ambient levels will not exceed
0.4 ppm (EPA estimate of 0.15 ppm as a 10 year average) under
certain temperature, humidity, and ventilation rate conditions.
The HUD regulations, however, were designed to reduce acute
reactions to HCHO and are not based on HCHO's potential
carcinogenicity in humans.
6-20
-------
Table 6-4. Sunwary of"«esidential Monitoring Data from Randomly-Sampled Hunts
Nurb«r
of harms
Convention*)
30
40
17
29
31
6
120
29
103
78
SI
Mobile
2
259
137
121
3
663
«ea/> (pom)
0.040
—
o.os
0.060
0.063
0.064
0.09
0.05
0.027
0.07
0.038
0.21
0.62
0.3S
0.18
0.114
0.091
Range (pom)
0.007 - 0.151
-------
EPA estimates a ten-year average ambient HCHO level of O.LO
ppm for new manufactured homes. EPA has used this estimate and
the estimated 10-year average for new homes that just meets the
HUD target level of 0.4 ppm (0.15 ppm) in the quantitative cancer
risk assessment. Another study has reported average levels of
0.54 ppm for manufactured homes less than three years old and
0.19 ppm for homes older than three years (State of Wisconsin,
1983). The Exposure Panel of the Workshop (1984) reported
studies that showed average ambient levels of 0.38 ppm for
manufactured homes not subject to complaints about HCHO odor by
residents, and averages of 0.38 ppm to 0.90 ppm for complaint
homes. Thus, an unrealistic worst case exposure estimate was not
used to estimate human risk. Also, only 10 years of exposure
were assumed for manufactured homes. Specific exposure data
follow.
The average HCHO level in mobile homes appears to have
declined in recent years due to the use of lower-emittina Dressed
wood products in mobile home construction and to the natural
decay of HCHO emissions from products in existing mobile hones.
Average levels in the existing stock of mobile homes are now
around 0.2 to 0.5 ppm/ with mean levels in individual homes
(including complaint homes) ranging from less than 0.1^ to over
1.0 ppm.
This apparent decline is shown graphically in Firjure 6-1.
The Conyers (1984) study of complaint mobile homes, initiated in
1980, showed mean HCHO levels of 0.85 ppm in new homes. An
6-22
-------
I
00 -
08 -
0 1 -
0 • -
O>
KJ
5 05
s
5
03-
02 -
01 -
060
040
Oil
02*
OM
0 Ik
0 12
0 10
OM
OO6
006
IMO
!•'•
!•'•
10'k
!•»«
1112
till
ItUtNO
- - IMOtl DAIAIClAVION/WISCOMSIMDAIASt I PHtUICItOI
0 iMOai OAIA ICOMVtHS IM4I
A IM2 OAIA IHOOtitS IM4I
• IM3MOAIA IUMIV ItXAS IM4I
* IM4OAIA IMHI IM4I
«) IM4 OAIA ISCNIOM*I«I IMb VCAH INMTHICMSIUOV WAS INIIIAUU
L-J
• •
—f 1 1 1 1 1 r 1 1 1 1 \ 1 r~
1069 1970 1971 1972 1973 1074 I97& I97C 1977 1970 1079 1000 1001 1902 1903
VtAHOt MANUFACTURE
I i'jUK. C-l . LEVELS IN MOBILE HOMES t OH Ht SPONGING TO VEAH OF MANUFACTURE
~T~
I9U4
-------
exponential function describing the relationship between HCHO
level and home age (r2=0.35) for the combined Singh (1982) and
Anderson (1933) data (i.e. the Clayton/Wisconsin data set) (1200
data points) predicts an average level of 0.5 ppm in new 1970 to
1980 vintage mobile homes (noncomplaint). Results of studies
begun in more recent years (University of Texas, 1984; MHE, 1984;
Sexton et al., 1985; Groan et al., 1985) indicate that initial
HCHO levels in new homes on average fall within the range of 0.2
to 0.3 ppm.
Using the exponential function describing the Clayton/
Wisconsin data to estimate decay of HCHO emissions over time, 10
year average concentrations can be estimated. For initial
concentrations in new homes of 0.5 ppm (i.e., Clayton/Wisconsin
data set), 0.4 ppm (i.e., the HUD target level), and 0.25 ppm
(i.e., midpoint of range of recent study of new home levels), the
10-year average concentration estimates are 0.19 ppm, 0.15 ppm,
and 0.10 ppm, respectively.
The fraction of homes with elevated levels of HCHO also
appears to have declined in recent years. Figure 6-2 shows that
the majority of homes less than 215 days old in the Clayton/
Wisconsin data set had HCHO concentrations above 0.4 ppm. More
recent studies indicate that this fraction is decreasing. The
California survey of 663 mobile homes (Sexton et al. , 1985)
reported levels exceeding 0.4 ppm only in two and three-year old
homes. The Texas study (University of Texas, 1984) reported that
6-24
-------
too -
CO -
o, FREQUENCY
I OF
^ EXCEE DANCE
in
20 -
1ECENO
A PfftCENI •' 1.0
• rfMCENI > 04
• PfftCEMI > 0.2
MEDIAN
2/S SSO S2S 1100 13/5 I65O I92S 2200 2476 275O 3025 3300 3S75
HOME AGE. DAYS
.,:IM i (,- „' . FIUOUENCYOF f OMMAI OtHYDE LEVELS. BY HOME AGE. EXCEEDING
1 0. 04. AND 0 ? .MM., IN Cl AY ION AND WISCONSIN DAI A COMOINf D
-------
the highest mean in any group of homes was 0.35 ppm (ten hones in
one county less than one year old); it is likely that one or more
of these had levels above 0.4 pom, but not approaching 1.0 ppm.
Levels -neasured at any one temperature and humidity can,
however,- be misleading. Table 6-S which illustrates the effect
of temperature and humidity changes on a 0.4 ppm reading at 25°C
and 50 percent relative humidity (the HUD target) shows that
under more extreme conditions (30°C/70 percent RH), the predicted
level could rise to 0.92 ppm. Because changes in temperature and
humidity occur over the course of a day and with seasonal weather
fluctuations, homes without constant climate control would
therefore be affected.
These data illustrate clearly that HCHO levels in homes are
the functions of multiple variables; neither age nor temperature
and humidity, nor any other variables can account for all
variations in residential levels (Versar, 1986b).
As the foregoing illustrates, HCHO levels in new
manufactured homes were tending toward 0.4 ppm and in some cases
above, until about 1979. After that date, mean HCHO levels in
new manufactured homes began to fall or level off slightly below
0.4 ppm. Even so, peak levels above 0.4 ppm can be expected at
times du« to adverse temperature and humidity conditions. The
freguency for such peaks is not known with confidence, but based
on the data available (see Tables 6-6 and 6-7, and Figure 6-1)
they could be expected to occur in a substantial fraction of new
manufactured homes.
6-26
-------
Table 6-5. Potential Effects of rawperature and Relative Munidity
Changes on Formaldehyde Air Conctnt rat ions (pa«>*
Relative htnidity
re*p<
S9f
68f
77»r
86-F
rrature
ciro
(20«C)
(25»C)
(30^)
301
0.08
0.1S
0.24
0.40
401
o.n
0.19
0.32
O.S3
SOI
0.14
0.24
0.40
0.66
601
0.17
0.29
0.48
0.79
701
0.19
0.33
O.S6
0.92
*CilcuIat«d using equations in Myers, 1984Mfticn were developed
primrily froi data on reUtiweiy new pressed Mood products and new
hovs. ASSIMH a te^erature coefficient of 8.930 and a nuridtty
coefficient of 0.0195. Assumes a base fomaldeftyde aeasurement of
0.40 pp* at 2S*C and SO percent relative hundity.
6-27
-------
Table 6-6.
FREQUENCY OF OBSERVATIONS FOUND IN CONCENTRATION
INTERVALS BY CLAYTON ENVIRONMENTAL CONSULTANTS
Concentration
Interval (ppn)
0.
•
•
•
•
•
•
•
•
•
1.
2.
0 -
11 -
21 -
31 -
41 -
51 -
61 -
71 -
81 -
91 -
1 -
1 -
Number
.10
.20
.30
.40
.50
.60
.70
.80
.90
1.00
2.00
3.00
of homes
Percent of
<0.5yrs >0.5
3.
7.
6.
7.
5.
6.
5.
5.
6.
12.
24.
7.
6
9
5
2
3
5
8
8
5
2
5
9
139
a.
4.
36.
16.
0.
12.
16.
4.
0.
4.
0.
0.
2
Sanpled Homesa
-1 yr All
0
0
0
0
0
0
0
0
0
0
0
0
5
8.
19.
14.
9.
5.
4.
4.
3.
3.
7.
14.
4.
Homes
1
7
3
3
0
6
6
9
9
7
7
2
259
a 259 "noncomplaint" mobile homes up to eight years old were
sampled in 1980-1981. Three measurements were typically taken
in each single-wide home and four measurements were taken in
each double-wide home. The data in the Table reflect the
average concentration measured in each home.
Source: Versar (1986a) statistical analysis of data supplied by
Singh et al. (1982).
6-28
-------
Table 6-7.
FREQUENCY OF OBSERVATIONS FOUND IN CONCENTRATION
INTERVALS BY WISCONSIN DIVISION OF HEALTH
Concentration
Interval (ppm)
0.
•
•
•
•
•
•
•
•
•
1.
2.
0 -
11 -
21 -
31 -
41 -
51 -
61 -
71 -
81 -
91 -
1 -
1 -
Number
.10
. 20
.30
.40
. 50
.60
.70
.80
.90
1.00
2.00
3.00
of observations
JL°-
2
29
0
10
10
13
10
7
2
2
10
0
Perc
5 yrs
.63
.0
.0
. 5
. 5
.2
. 5
.9
.6
.6
.5
.0
38
rent of
>0.5
3.
13.
21.
14.
11.
12.
8.
5 .
3.
0.
5.
0.
21
Observations'1
-1 yr All
8
6
1
6
3
2
9
6
3
0
2
5
3
14.
20.
18.
14.
9.
8.
5.
3.
2.
0.
3.
0.
Ho ne s
1
4
4
0
2
0
2
6
2
7
3
3
976
a 137 "noncomplaint" mobile homes up to nine years old were
sampled in 1980-1981. Each home was sampled at least six
times at monthly intervals. The data in the table reflect
the results of 976 measurements.
Source: Versar (1986a) statistical analysis of data supplied by
Wisconsin Division of Health (1984).
6-29
-------
6.5.3. Conventional Homes
The average HCHO levels reported in several monitoring
studies of conventional homes range from less than 0.03 to 0.09
ppm (see Table 6-4). Newer homes and energy efficient homes with
low air exchange rates tend to have higher HCHO levels (often
exceeding 0.1 ppm) than older hoines (Versar, 1986c) . Results of
recent studies indicate that initial HCHO levels in new
cofiventional homes generally fall within the range of 0.05 to 0.2
ppm; few measurements exceeded 0.3 ppm (Stock and Mendez, 1985?
Hawthorne et al., 1984; SAI, 1984; Wagner, 1982). Computer
modeling to estimate initial HCHO levels in conventional homes
built using significant amounts of pressed wood (i.e., either
underlayment, paneling or both) yields values ranging from 0.1 to
0.2 ppm (Versar, 1986 ). Using the exponential decay function
described in Section 6.5.2, the 10 year average concentration for
a home with an initial concentration of 0.15 ppm (i.e.,
approximate midpoint of range of new home levels) is estimated to
be 0.07 ppm. Summaries of some of the major HCHO monitoring
studies are presented below.
The Lawrence Berkeley Laboratory (LBL) has summarized HCHO
concentrations in 40 residential indoor environments since 1979
(Girman et al., 1983). They have found that HCHO concentrations
in homes designed to be energy-efficient are somewhat higher than
concentrations in conventional homes. The maximum reported value
is. 0.214 ppm in an energy-efficient home in Mission Viejo,
California. Data are not sufficient to allow calculation of mean
levels.
-------
As part of the development of an indoor air pollution model
based on outdoor pollution and air exchange rates/ Moschandreas
et al. (1978) studied the patterns of indoor aldehyde levels
monitored in 17 houses in the U.S. These data can be useful if
we assume HCHO constitutes 60 percent of total aldehydes, based
on LBL data (Girman et al., 1983). The 17 houses had an average
aldehyde concentration of 0.09 ppm. Applying the 60 percent
factor, the average HCHO concentration for the houses would be
0.05 ppm. The highest mean for any one home was 0.26 ppm; the
range for that home was 0.2 to 0.45 ppm. Another home with a
mean of 0.20 ppm reported a range of 0.07 to 0.5 ppm. For no
other conventional home did levels exceed 0.4 ppm.
A University of Iowa Study (Schutte et al. , 1981), oerfomed
for the Formaldehyde Institute, monitored 31 conventional,
detached homes not containing urea-HCHO foam insulation (UFFI)
for HCHO concentrations in the indoor air. Samples were
evaluated in relation to outdoor HCHO concentrations, age of the
home, and other environmental factors monitored at each of the
sampled homes. The average indoor concentration found in the
homes was 0.063 ppm (standard deviation = 0.064) with a ranrje of
0.013 to 0.34 ppm. In only 5 of the 31 homes were average
concentrations higher than or egual to 0.1 ppm.
The 1981 Canadian study (UFFI/ICC,. 1981) also studied nor-
UFFI homes. Table 6-8 summarizes these data, showing that levels
in none of the 378 homes exceeded 0.2 ppm.
6-31
-------
Table 6-8. Comparison of Non-UFFI Canadian Homes
by Average HCHO Concentration
Av0r40fc
'OHM I uQnyQV
concentration (DOB)
«.OI
.01-. 025
.025-. 040
.040-. 055
.055-. 070
.070-. 085
.085-. 10
.1-.15
.15-. 20
Totals
MJMMr of
48
111
97
67
30
15
—
9
1
378
fltrcenugt
12.7
29.4
25.7
17.7
7.9
4.0
—
2.4
0.3
100.1
CkJRjIjtivt
percentage
12.7
42.1
67.8
85. S
93.4
97.4
-
99.8
100.)
Source: (ffl/lCC (1981).
6-32
-------
A report by Virgil J. Konopinski (1983) of the Indiana State
Board of Health summarizes the results of a series of
investigations conducted from 1979 through 1983 to determine HCHO
levels .in conventional homes in Indiana. The mean HCHO level in
the 120 homes without UFFI was 0.09 ppm (0.05 for homes with
UFFI). That mean could be skewed by the maximum concentration of
1.35 ppm reported in one home. Neither the age of the homes nor
the age of the UFFI installations was reported.
From April to mid-December 1982, Oak Ridge National
Laboratory (ORNL) with the U.S. Consumer Product Safety
Commission (CPSC) studied indoor air quality in 40 east Tenessee
homes. The objective of the study was to increase the data base
of HCHO monitoring in a variety of American homes and further
examine the effect of housing types, inhabitant lifestyles, and
environmental factors on indoor pollutant levels.
Homes to be sampled were selected based on a stratification
to ensure representative home age, insulation types, and heatinq
sources. All were voluntarily enrolled. Twice a month, four
samplers at each location monitored HCHO levels in three rooms
and outside the house. Samplers were exposed to the air for
24-hour periods. No modifications to the residents' life styles
were requested during these measurements.
Table 6-9 summarizes these data by home age and season
(indicative of temperature and humidity). HCHO measurements in
the 40-home east Tennessee study led to the following major
conclusions:
6-33
-------
Table 6-9 . o»t/c«C »an Forbid**, concentration, (pp.,
" " " « of Agt and Season ((
poo Detection Limit)
*9» of house
—^—— — — — , .
all
0-5 years
5-15 yMn
older
0-5 years
5-15 years
older
all
Season
-
all
all
all
all
spring
siamer
fall
spring
suwer
fall
spring
staler
fall
spring
SMer
fall
i*
^~^^^™™™^^— •»•
0.062
0.084
0.042
0.032
0.087
0.111
0.047
0.043
0.049
0.034
0.036
0.029
0.026
0.062
0.083
0.040
s«
0.07?
0.091
0.042
0.042
0.093
0.102
0.055
0.040
0.048
0.03S
0.051
0.037
0.023
0.076
0.091
0.047
•
5903
3210
1211
1482
1210
1069
931
626
326
259
757
341
384
2593
1736
1574
n
— ^~~~a^_i
40
18
11
11
NOU:
> • *Mn concentrations.
s • standard aviation.
• • ntflfctr of Mtsurvwtts.
IneludM nom Kith and without UFFI.
Hawthorn* tt al. (1984)
6-34
-------
(1) The average HCHO levels exceeded 100 ppb (0.1 ppm) in 25
percent of the homes.
(2) HCHO levels were found to be positively related to
temperature in homes. Houses with UFFI were freauently
found to exhibit a temperature-dependent relationship
with measured HCHO levels.
(3)- HCHO levels generally decreased with increasing age of
the house. This is consistent with decreased emission
from materials due to aging.
(4) HCHO levels were found to fluctuate significantly both
during the day and seasonally.
Studies by Breysse (1984) evaluated conventional, non-UFFI
homes. The University of Washington studied 59 such homes;
private laboratories in the state studied an additional 25. The
freouency distribution for measured levels are presented in Table
6-10. A total of 6 of the 189 samples (3.1 percent) were over.
0.5 ppm and 56 samples (26.5 percent) were over 0.1 ppm.
Traynor and Nitschke (1984) monitored indoor air pollutants
in 30 homes with and without suspected combustion (and other)
sources. The average HCHO level observed in all the test homes
was 40 ppb; a high value of 151 ppb was found in one of the
tested residences categorized as containing new furnishings and
new paneling as a suspected pollution source.
6-35
-------
Table 6-10. frequency Distribution of Fonw)d>nydt l»v«ls
in UitAington Conventional Non-lFFI HONK
Fomaldtnydt
concentration
(p.)
> 1.0
> O.S - 0.99
> 0.1 - 0.49
< 0.1
TOTAL OBSERVATIONS
Nufeer of Sarnies
S9 U. ttsn 2S Private
tae* l*t> hoMS
2 0
2 2
41 9
«8 6S
113 7ft
1-189
Frequency
(percent)
1.0
2.1
2ft. S
70.4
Source: SrtysM (1984)
6-36
-------
The results can be summarized as follows:
o The 4 homes with no identified source had a ranqe of
means of 0.007 to 0.034 ppm.
o The 3 homes with new furnishings had a range of means of
0.015 to 0.061 ppm.
o . The 4 homes with cigarette smokers had a range of means
of 0.032 to 0.060 ppm.
o The 18 homes with gas, coal, and wood fueled
appliances/heaters had a range of means of 0.012 to
0.056 ppm.
o The 12 homes with a combination of sources reported a
range of means from 0.013 to 0.064.
Variations in home levels could not be attributed to combustion
sources.
Stock and Mendez (1985) measured HCHO concentrations inside
78 homes in the Houston, Texas area during the summer of 1980.
No mobile homes, UFFI homes, or complaint homes were samoled.
Indoor concentrations ranged from less than 0.008 ppm to 0.29 ppm
with an average value of 0.07 ppm for detectable concentrations
(Number of samples, N»75). Three energy efficient condominiums
had, as a housing category, the highest mean level (0.18 ppm).
Condominiums (N*ll), apartments (N»19), and energy-efficient
houses (N»7) represented the mid-range with mean levels of 0.09,
0.08, and 0.07 ppm, respectively; the mean of 38 conventional
houses was 0.04 ppm.
Wagner (1982) measured HCHO levels in 12 California homes
that fall into a prescribed "worst-case" category of buildinq and
occupancy characteristics (i.e., low infiltration and ventilation
rates, new construction, presence of gas stoves). Weekly averaqe
6-37
-------
concentrations ranged from 0.078 to 0.163 ppm with a mean of
0.106 ppm.
Sexton et al. (1985) measured HCHO levels in 5 1 home
dwellings.- Weekly average concentrations ranged from 0.013 to
0.085 ppm with a geometric mean of 0.035 ppm and an arithmetic
mean of 0.038 ppm. Seventy-six percent of the homes were more
than 10 years old and only two were less than six years old.
A downward tend in HCHO levels in conventional homes is seen
in Figure 6-3. The relative proportion of low HCHO levels in
homes that have been monitored has increased over the past six
years, and the proportion of high levels have decreased. These
data are limited and caution in interpretation is recommended
(Versar, 1986a).
6.5.4. Garment Worker Exposure
HCHO levels in apparel manufacturing facilities were
generally below 3 ppm prior to 1980 (see Table 6-11). OSHA had
established a 3 ppm TWA (time-weighed average) in 1967. However,
OSHA is presently considering establishing a new level (see 50 FR
50412; December 10, 1985). The ACGIH (American Conference of
Government. Industrial Hygienists) recommended level is 1 ppm
TWA. In recent years, HCHO levels observed were generally below
1 ppm (see Table 6-12). The data in Tables 6-11 and 6-12 must be
viewed with caution because in 1983, the National Institute for
Occupational Safety and Health (NIOSH) discovered that the
commercially prepared inpregnated charcoal tubes which had been
used in previous personal monitoring studies were unstable.
6-38
-------
100
cr\
LJ
2
D
O
MOSCHANOHEAS*!*!
SIUCK ANUMCNUtZ I.I9UOOAIAI
SCHUIIE «l4l. (ISaOOAIAt
HAWIHURNt »t* II9A2UAIAI
IHAVNUH ANONIISCHKL IliMiJ 1984 UAI A|
oui
OOiOl
0203
0104
04
I OHMAI IJMIVIll I UNI I NltfAIION (HHMI
'"l'"i-r,-!. | ,-tMiiii.nl y IJisli -iliuliim ,; U-vcIs in Conwenl ional llui..es
-------
Table 6-11. PRE-1980 MONITORING DATA FOR GARMENT MANUFACTURING AND
CLOSELY RELATED INDUSTRIES
i«mr,/fM.m, tm
TflUlUilMU*
farvtut M««'*CtM't
t*»"«»ft titt" ovttf* ct*ttM
(l.f. • «'»»» lft«9t. ClOtH*
»«f •••VAtMtt tlO'tt. fU.)
»4.f«C tm/tf».tW-t
•»'
Witr""-"*1**
M^MllWmui.1^
^^••^•flt frvtl
»t^t^.t »r,fl
»,^,-,M ,r»»t
*t*M*»M If tl
•S
••?
«
MS
M
•S
^^,1^^**
0.1 - 1.4 (1171)'
1 • 11 (INS)'
0.1 • 2.7 (T\a«4rM. IMC)
«0.l . 1.4 (TM. 1171)'
0.15 • O.M (TWA. 1N4)'
O.I • 2.7 (TW*-«rt«. 1M«)
fl.OOC-fl.M4 (TWA>ftf1<«4l .
117})
2.000 • 1.140 (T»gu»»f »»«•).
117!)
<9f (Cri». 1171)
2. no (c«n.
• •• fttrtDt^ MMH9I14JM •Mll44)lf.
4 MM IMCI'ttO <• «•• HfOtXt.
' S««lt tyft «4M NMtflM.
J Nr«|M«l *r •*«» «| iMctf«t4.
••t MtKtM.
IS
6-40
-------
Table 6-12.
RECENT MONITORING DATA FOR FORMALDEHYDE IN THE GARMENT
MANUFACTURING INDUSTRY
SIC
C.**
mi
tin
1111
llrt
CM,
NMcN*l«r MMir4 f»tkiMit.
CMtUl MMlrltt. Olr»ln«j.
M*. ft
•TTM Skirl U.
MMtrlck MMlt« Wilt. l«.
•brrttkwrt. Ml
f»M»llM Skirl Mf. U..
llMkAllAd Skirl to.
AMrlcnt. CA
C. t. MtlkMM, to.
C. r. totktMy to.
lUftlaa Skirt to.k
t*rtm Skirl C«.
•1 !«•(*. U
•rrev Skirl U.
lr« Co«M, Nl« C.
••lltrtft |M« (•
AH*M«. t*
IN t». Iw
ftkrlc ly9*/lrr*lBCn|
•S*
•S
**lyt«Un. p*lipttlir-c*ll««.
kltMt. H|lo« kltMt; ONOMlU*
•S
•S
^ysar^..ls.
•S; f«ra»ltekr*-k«tf|<«-fc»»«< re* In
•Si f«f««l4ckydr-k«it4 r«»l»
yrMnrt4
•Si UravltftkfM-kttctf rttU
•Si OMOMU. ftttvttt, fit-
•S; f*r**l4*ky4r-k«»(4 rttln
ft t- 1*1*4. »r(-c«l
•S
•t
•t
"~:r
0.010-0. 10* |IM*-l«rta^l.vIM4)
1.0 |itr««a. IMll
1.0 (tcrrc*. IMll
1.0 ItcrcM. IMll
0.14-0. II llM*-«rt*. IMI)
0.11-0.44 (Cr«k. IMll
0.00-0.11 inu «r«». IMI)
O.II-O.tl |IIM-«ro. IMll
0.11 I»f4». IMll
IMir
0.00-0.10 IIM-*'**. IMll
O.IO-O.M |IMI-«rt*. IMI)
O.t4 (r*«k. IMll
•O.OM-3.SI (llM-f«r»WMl.lMl)
O.IS-0.41 |IUA.|r««. IMll
1.01-1. II |HM-krctlkl«« l*M.
• o^AAl
0.01-1.0) |IIU.«rc*. IMO)
0.1-1.0 |(r«k. IMO)
O.OM-0.194 rnM-Hri«Ml. IMI)
0.00-O.U (lH«-«rtt. IMll
O.IS-O.MinM-Hr»*««l. IMI)
O.II-O.M IIIM.«rct. IMI)
• .M *.U jllU.r... IMll
o.ii |»ta. IMII
0.440 |t«lllM-*rc«. IMll
0. IM^IMA-pa««l. IMO)
0.110 ('««». IMO)
J 000 (Urt(*. IMO)
0.000 Ikrrcn. *»U «n»"0«i«)
.r^:,.
i
i
i
i
i
U
i
14
10
£•
M
10
14
M
M
11
1
1*
10
10
Ml
II
10
11
U
M
1
1
1
1
1
t-.Hi
Ml 1*4
•t
•S
1
1
1
1
4
4
S
4
1.4
S
s
4
1.4
S
s
4
4
4
V
4
S
4
1.4
i
S
•S
M
•S
•t
•S
-------
Table 6-12. (continued)
en
NJ
JIN
»W
2142
21S1
?MI
IMS
Mil f CM U.
CM». M
MtMM. K
l«fl*r C«.
IMMlrlct
M».. M
terfctr felt !•».
••tic*. M
I to**It. lie.
• itMrct. M
Mttllfc l«i I«C.
. CA
»'|. C*.
•IralagM*. M
rt»U.
ftferlC
M
M
M
M
•S
•.«-•.
•.M |IIM..r««. IN!)
t !•-• I! |IM-*«rt«Ml. .
• !•••.41 IIM «rr*. |MO|
• It I.II jtc.k. IMOI
It.OOO |Scrc*«. IMI)
•.000 (fcrtt*. |«00)
• •M-O.IM (IM
. IMI)
. IMI)
• 00 IHM.HM*Ml. IMII
1.000 ikrre.. |Ml|
N
1
10
II
SIC
In
*•
<«H IM.
NMlflt4 » I CM in.
f I CM 110
CIA IMri SJS *•»
•>Mftr »r**4 l«4lt«l«r lukt*
» I C*K M4.
••! »»t(l'lt4 U Ikt rt>*r«MC.
r«tln.
c*n
for IS
IMl
I* Ifct
»r*«IHU«
U ill c«4t 1111.
t«*»lrt. Ik* >H«
«**!(*
«rr*
Ik*
I.M
M
4
1.4
S
•S
n
•i
•s
•$
•s
n
kr**lhln« I^K.
Ikli «••»*•, can *li« k» cUttl«U« U SIC <•*» lilt. 1111. «•< lilt.
-------
Table 6-13. NFOSH Monitoring Data—Ranges
by Department
(i SAMPLES)
CUTTING
(12)
(29)
COLLAR
(33)
(27)
PARTS
(30)
(46)
ASSEMBLY
r (7)>
(66)
PACKAGING
(45)
(20)
ADMINISTRATION
(30)
(26)
tf fl I
* ° 1
1 h 1
1 ° 1
1 fl 1
1 ° 1
1 0 I
1 ° 1
4 m 1
« 9 <• |
^ ft . - -. „..!
1 o 1
1 ° 1
0 0.1 0.2 0.3 0.4 0.$ 0.
8 HOUR ttfA FORMALDEHYDE CoNceNtRAtloN LEVEL
-------
Thus, the nonitoring data above may be suspect since the loss of
HCHO from the tubes was not consistent. Consequently, the HCHO
levels recorded most likely represent lower levels than actual
conditions. The NIOSH method at that time was also used by OSHA.
NIOSH subsequently developed a stable medium for collecting
the HCHO and did two in-depth industrial hyqiene studies. The
surveys were done at two large manufacturing sites producing
men's dress shirts. HCHO exposure levels were determined for 54
of 72 job titles in two different plants. The number of
individuals within each job title whose exposure levels were
sampled was based on the trotal number of employees in that
category and reflect a 95 percent confidence level that the
highest and lowest exposed individuals were included in the
sampling. A summary of the data are presented in Tables 6-13 and
6-14. These tables show that all levels of exposure were less
than 0.51 ppm TWA. Also, as Table 6-13 illustrates, the combined
range of data was very narrow (0.01-0.39 ppm) for 5 of the 6
departments in the two plants. The range of mean concentrations
of all departments (0.13-0.20 ppm) is very narrow and compares
well within the overall combined mean exposure level of 0.17 onn,
which was used for the quantitative cancer risk assessment. In
addition, the average exposure levels used in EPA's section 4(f)
determination (EPA, 1984), 0.23 ppm (area) and 0.64 ppm
(personal) (Versar, 1982), were also used for this cancer risk
assessment.
6-43
-------
Table 6-14,
U1
FORNU.DEHYOE. CONCENTRATION LEVELS (PPM)
GARMENT MANUFACTURING
ADMINISTRATION
CUTTING
f f*l • A *K
COLLAR
PARTS
ASSEMBLY
P Art/ A/» • u<»
•ACK AGING
•r o«rri_t^
S6
11
60
76
1*9
65
(137)
— Wm__ lEtf1"-
0.01
<0.01
0.02
<0.01
*0.01
*0.01
«0.01
- 0.51
- 0.39
- 0.39
- 0.35
- 0.35
- 0.27
- 0.51)
o,n
V. ±J
0.1*1
0.16
0.20
0.1?
0,14
(0.17)
-------
All of the determinations made in the NIOSH studies were at
one point in time and may not reflect the variation of exposure
over a longer period. Factors that could affect variation in
HCHO levels in these plants include variation in ambient
temperature, humidity, type of fabric or resin system, and volume
of stored materials or completed work.
The exposure range across departments, within plants, as
well as between plants, appears to be narrow. Both these plants
were large manufacturing sites, producing similar products. Both
plants had central ventilation/cooling systems. This type of
plant may potentially represent only 10 percent of the total
number of manufacturing sites (though up to 25-30 percent of the
workforce may work in such plants) (Ward, 1984).
6.6. Summary
The data presented above indicate that HCHO levels in new
manufactured homes are generally below 0.5 ppm, with 10-year
averages for new HUD Standard homes of 0.15 ppm or less.
However, some fraction of new homes experience peak levels that
could exceed 1.0 ppm for periods of time. It would be expected
that as temperature/humidity exceed 75°F/50% RH, HCHO levels
would rise as Table 6-3 illustrates. Thus, depending on heating
and cooling preferences, HCHO levels in new homes may
substantially exceed the reported mean for new homes.
The situation is similar for conventional homes, although
reported mean levels are lower, 0.03 to 0.09 ppm. However,
because conventional housing is much more heterogeneous, peak
6-46
-------
levels in some new homes may substantially exceed reported
means. Although temperature and humidity conditions play a large
role, construction techniques which tend to limit air exchanges,
such as in energy efficient homes, and building product mixes are
also of. major importance. The ten-year average HCHO
concentration for a new home built with significant amounts of
pressed wood is estimated to be 0.07 ppm.
Reported HCHO levels during garment manufacture are below
1.0 ppm and in some plants below 0.5 ppm, and the NIOSH data
indicate rather tight ranges (none exceeding 0.51 ppm). However,
much of the reported monitoring data must be approached with
caution due to the technical fault discussed earlier. Building
design, ventilation, and temperature/humidity changes may be
responsible for daily or seasonal variations.
6-47
-------
7. ESTIMATES OP CANCER RISKS
In principal, data from studies of humans are preferred for
making numerical risk estimates. However, as is often the case,
the available epidemiologic data on HCHO were not suitable for
low dose quantitative cancer risk estimation, mainly because of a
lack of adequate exposure information in the studies.
Accordingly, results from studies in animals were used to
estimate low-dose human cancer risk. This is done by fitting
mathematical models to the observed animal data. In addition,
even though the epidemiologic studies were not suitable for
quantifying a dose-response curve, those studies with observed
statistically elevated cancer risks provided some support for the
animal-based predicted upper bound risk. This comparison, while
yielding valuable information to the assessment, should be viewed
with caution since exposure levels in these epidemiologic studies
were subject to some variation.
7.1. Risk Estimates Baaed on Squamoua Ceil Carcinoma Data
Data from three different studies were considered for their
appropriateness to this risk assessment, studies by Kerns et al.
(1983) (the CUT study>, Albert et al. (1982) (the NYU study),
and Toba *t al. (1985). Dose-response modeling was applied to
the CUT data for Fischer 344 rats using squamous cell carcinomas
of the nasal turbinates as an endpoint. See Table 7-1 for the
statistical significance of the response in the CUT and Tobe
studies. The NYU study provides corroborating evidence of a
similar response in another strain of rats (Sprague-Dawley).
7-1
-------
""'able 7^1
Carcinoma tumor incidence in Fischer 344 rats and male B6C3R1 mice
Fisher Exact Test Results
I
10
Species
Fischer 344
Rats of the
CUT Study
(•ales and
females
combined)
of the CUT
Study (Hales)
Dose (ppm)
Control
0/156 (0)
2.0
5.6
0/159 (0)
2/153 (.01)
pf-0.24
14.3
94/140 (.67)
pf<0.01
B6C3F1 Mice
Control
0/109 (0)
2.0
0/100 (0)
5.6
0/106 (0)
14.3
2/106 (.02)
pf«0.24
Fischer 344
Rats of the
Tobe ftudv
Control
0/32 (0)
0.3
0/32 (0)
2.0
0/32 (0)
15.0
14/32 (.44)
pf<0.01
'- Numbers in parentheses are proportions responding. " ~~~
- Fisher Exact Test p-value. Small values indicate that the response in
dosed animals may be significantly different from the response in the
control animals. This p-value should be compared to 0.017 for
significance at the 0.05 level. This is a multiple comparison, which uses
a critical value of o/k for k (in this case k = 3) comparisons with the same
control group.
-------
Table 7-1
Neoplastic polypoid adenoma incidence in Fischer 344 rats,
Fisher Exact Test Results
Species
Polypoid
Adenomas in
Fischer 344 Rats
of the CUT
Study.
Control
1/156 (1)
Dose (ppm)
2.0
7/159 (.04)
pf=0.04
-j
i
-------
That study, however, was considered less appropriate for risk
estimation since it contains only one nonzero exposure
concentration, and, based on the CUT data, one would expect the
true dose-response curve in the experimental range to be highly
nonlinear'. The Tobe study was not relied on for primary risk
estimation because a tumor response was seen only at the highest
dose group and the number of animals per group was relatively
small (32). However, risk estimates based on the Tobe data are
discussed in sections 7.3 and 7.4. Although not statistically
significant, the squamous cell carcinoma response in two B6C3F1
mice of the CUT study at 15 ppm is suggestive of carcinogenicity
from formaldehyde inhalation in another species due to the rarity
of this tumor. This data set was not considered for dose-
response modeling, however, because of the limited response at
the highest dose level. The CUT study was chosen as the source
of data for several reasons: it was an experiment by inhalation,
which is the primary route of exposure to man; the quality of the
study is considered to be high; and it includes four exposure
levels and responses at those levels for determining the shape of
the dose-response curve (Grindstaff, 1985).
It was decided to estimate the risk of tumor to rats
chronically exposed up to time of death without intervention, or
to a terminal sacrifice at 24 months. With some adjustments for
earlier sacrifice kills discussed below, this was estimated from
the CUT data. The dosing regimen assumed is that of the CUT
study, where exposure was six hours per day for five days per
7-4
-------
week. For estimation purposes, the animals that lived beyond 24
months were included with the animals sacrificed at 24 months.
An adjustment was necessary to correct for animals that died
very early in the CUT study or that were sacrificed prior to 24
months. The rats that died prior to the appearance of the first
squamous cell carcinoma at 11 months were not considered at
risk. Rats sacrificed at 12 and 18 months would be treated as
though they would have responded in the same proportion as the
rats that remained alive at the respective sacrifice times.
From this approach an estimate of the probability of death
with tumor within 24 months and an estimate of its variance was
obtained. The number of animals at risk and the number with
tumors that would give the same estimates of mean and variance
for a 24-month study with no interim kills at 12 and 18 months
was determined, and used as the input data for risk analysis.
The data adjusted for sacrifice kills obtained in this manner are
0/156, 0/159, 2/153, 94/140 (figures rounded), at nominal dose
levels of 0, 2, 5.6, and 14.3 ppm. These numbers were for the
significance tests in Table 7-1.
Another method, which was not used, would simply omit from
the analysis all rats sacrificed prior to 24 months. The data
adjusted for sacrifice kills by this method are 0/156, 0/159,
2/155, and 95/141. The two constructed data sets produce a
negligible difference in estimated risk at very low doses under
the dose-response model discussed below.
7-5
-------
Administered dose expressed as ambient air concentration was
used directly as the measure of dose in this assessment. An
alternative method suggested by Casanova-Schmitz et al. (1984)
and Starr et al. (1984) using data derived from the formation of
HCHO-DNA-adducts was not used because of the uncertainty
associated with this approach (as discussed previously).
However, use of these data reduces the maximum likelihood
estimate of risk approximately by a factor of 50 and reduces the
upper bound estimate of risk by a factor of 3.
Since risk at low exposure levels cannot be measured
directly either by animal experiments or by epidemiologic
studies, a number of mathematical models and procedures have been
developed to extrapolate from high to low doses. Different
extrapolation methods may give a reasonable fit to the observed
data but may lead to large differences in the projected risks at
low doses. In keeping with.EPA's Guidelines for Carcinogen Risk
Assessment and the OSTP Principle Number 26, the choice of low
dose extrapolation method is based on consistency with current
understanding of the mechanisms of carcinogenesis and not solely
on goodness of fit to the observed tumor data. When data and
information are limited, and when uncertainty exists regarding
the mechanisms of carcinogenic action, the OSTP principles
suggest that models or procedures which incorporate low-dose
linearity are preferred when compatible with the limited
information available. EPA's Guidelines recommend that the
linearized multistage procedure be employed in the absence of
7-6
-------
adequate information to the contrary and specify the possible
presentation of various other models for comparative purposes.
This presentation is given in Appendix 3. In addition, see Conn
(1984), Siegel et al. (1983), Brown (1984), Sielken (1983) and
Clement Associates (1982) for discussions concerning quantitative
methods/models for quantifying the potential risks to humans from
HCHO based on the Kerns et al. (1983) study.
The behavior of eleven models used to extrapolate risks was
examined in Appendix 3. These were all dichotomous models
("tumor-no tumor" models). These models along with their
parameter estimates, standard errors, log-likelihoods, and 2
/v
goodness-of-fit test statistics and p-values are presented for
the CUT Fischer 344 rat data on squamous cell carcinomas. Those
interested in the underlying assumptions of these models and
their mathematical form are referred to Appendix 4.
Each of the eleven models listed in Appendix 3 was used to
extrapolate risks from the CUT rat study. They were the
additive and independent forms of the probit, logis.tic
regression, Weibull, and gamma-multihit models and the one, three
and five stage multistage models.
The multistage model without restrictions on the order of
the polynomial in dose is the model of choice. As discussed
above, the Guidelines specify that unless another model can be
justified, the linearized multistage procedure will be employed
(EPA, 1986). In the case of HCHO, we know that it is mutagenic,
can react with nuclear material and processes, is structurally
7-7
-------
related to other carcinogens, is cytotoxic, and is clearly
carcinogenic in the rat. All reasons that taken together justify
use of the linearized multistage procedure.
The formulation of the model for quantal response data was
preferred' to one including time as a variable. Based on
simulation studies conducted under contract to EPA, it was not at
all clear that inclusion of time as a variable would provide
improved estimation, and there would have been some question
about the validity of the results in this case, due to lack of
knowledge of the cause of death of experimental animals, and due
to adjustments made for sacrifice data (Howe et al., 1984). Risk
is summarized as model-derived point estimates and associated
upper bounds in the dose ranges of interest. The latter
corresponds to the number from a linearized multistage model
procedure.
Although arguments have been made that there may be a dose
level below which the added risk of cancer is zero, there is no
consensus within the scientific community on this topic. Through
use of mathematical models of dose-response, there is currently
no way to demonstrate either the existence or nonexistence of a
threshold. In addition, if any thresholds exist, they are likely
to vary among members of the population at risk, and may be
modified by other environmental agents. Therefore, use of a
dose-response model incorporating a single threshold would
provide an estimate of an average population threshold that wouii
have little practical utility. In the absence of clear evidence
7-8
-------
of a threshold and quantifiable supporting data that could be
utilized in risk assessment, it was felt that for dose-response
modeling of HCHO it should be assumed that there is no
threshold. Such a conclusion is supported by the Consensus
Workshop on Formaldehyde (1984). In addition, although much data
have been developed to elucidate the possible mechanism for the
nonlinear carcinogenic response observed in the CUT rat study,
at this time low dose linearity cannot be ruled out.
The likelihood of response was treated as equal in rats and
humans for the same exposure regimen and proportion of lifetime
exposed. Although differences have been apparent in
susceptibility among the species that have been exposed to HCHO,
there are no data suggesting that man may be less susceptible
than rats.
The estimated risk to rats is based on the CUT dosing
regimen for a period of 24 months, which may need to be adjusted
upward to obtain an estimate of risk for lifetime exposure. It
may also be necessary to convert the estimated risk to a shorter
exposure duration in some cases, or to adjust for a different
exposure schedule (i.e., other than six hours per day, five days
per week). However, there is little scientific knowledge that
addresses these problems. Consequently, each estimate of
lifetime risk from the model (assumed to be equivalent for humans
and rats as discussed in the proceeding issue) was multiplied by
the proportion of a human lifetime actually exposed.
Hypothetically, then, at an exposure concentration producing a
7-9
-------
lifetime risk of 1/1000, the risk for exposure of half a lifetime
was estimated as 1/2000. Similarly, if exposure was for 45 hours
per week instead of for 30 as in the CUT study, the risk
estimate would be multiplied by 45/30 = 1.5, giving 3/2000.
Exposure of half a lifetime but at 45 hours per week would give
.5 times 1.5 times 1/1000 = 1/1333. It should be noted that due
to the upward curvature of the dose-response curve the resultant
risk will be less than if the scaling factor were applied to dose
before substitution into the model if the factor is greater than
one, and would be greater if the factor is less than one (in both
instances the difference would be less than 2 fold). If the
response curve were linear there would be no difference between
scaling risk or scaling dose. It is acknowledged that this rule
for adjustment is based on very simplified assumptions.
The unit risk and estimated individual and population risks
to humans for various exposure categories are presented in Table
7-2.
7-10
-------
Table 7-2.
ESTIMATED RISKS BASED ON SQUAMOUS CELL
CARCINOMA DATA PROM CUT STUDY.
POPULATION RISKS (number of excess tumors) APPEAR
IN PARENTHESES BELOW INDIVIDUAL RISK ESTIMATES.
Category
Nbbile Home
Residents
1. Based on
current
monitoring
data
2. Based on HUD
target level
Nbnufacturers
of /£parel •
1. OSHA standard
2. Personal
sample
3. Area
sample
4. NIOSH data
Population
7,800,000*
777,000
Conventional
Home Residents
Unit Risk
6,310,000*
Exposure
0.10 ppn
(112 hrs/wk
for 10 yrs)
0.15 ppn
(112 hrs/wk
for 10 yrs)
3.0 ppn
(36 hrs/wk
for 40 yrs)
0.64 ppn
(36 hrs/wk
for 40 yrs)
0. 23 ppn
(36 hrs/wk
for 40 yrs)
0.17 ppn
(36 hrs/wk
for 40 yrs)
0.07 ppn
(112 hrs/wk
for 10 yrs)
1 ug/m3—
0.00082 ppn
(for 70 yrs)
Maximum Likehihood
Estimate of Risk
2 X 10
'10
[81]
1 X 10~9 [Bl]
6 X lO"4
6 X 10~7 [Bl]
9 X 1CT9 [Bl]
4 X 10~9 [Bl]
6 X 10'
rll
[Bl]
Upper Bound
Estimate of Risk
1.5 X 10"4 [Bl]<
(1,170)
2 X ID"4 [Bl]
(1,560)
6 X 10~3
1 X 10~3 [Bl]
(777)
4 X 10"4 [Bl]
(311)
3 X 10"4 [Bl]
(233)
1 X ID"4 [Bl]
(630)
1.3 X 10'5 [31]
* Population estimates are based on anticipated additions to the housing
stock over the next 10 years as estimated by Schweer (1987).
**Classification under EPA's Guidelines for Carcinogen Risk Assessment—
[Bl]=Probable Human Carcinogen.
7-11
-------
7.2. Risk Estimates Based on Polypoid Adenoma Data
There appears to be little credible evidence that polypoid
adenomas progress to any of the malignant tumors seen in the Kern
et al. (1983) study. However, while the adenomas should not be
combined statistically with the squamous carcinomas for hazard
identification purposes, they represent an endpoint that can be
quantified separately for analysis pruposes.
Because it is beyond the capability of the various
extrapolation models to fit data with a negative slope, an
alternative extrapolation procedure is to drop the two highest
doses and use the data from the 2.0 ppm rat exposure group
(straight line to zero). However, since the true slope of the
dose-response curve is unknown below 2.0 ppm, this approach may
vastly overestimate the true risk if the curve is convex, and
underestimate it if it is concave. The reason the occurrence of
polypoid adenomas has a negative slope probably lies with the
fact that the cell type in the respiratory epithelium from which
these tumors arise is lost sooner and to a greater extent with
increasing dose due to squamous metaplasia. The less respiratory
epithelium available the smaller the chance for adenomas to
develop. Other explanations are also possible as discussed in
section 7.4.1.
Risk estimates using polypoid adenomas appear in Table
7-3. For polypoid adenoma as the endpoint instead of squamous
cell carcinoma there is no difference between the two procedures
described earlier to adjust for animals at risk. The first
7-12
-------
observation of a polypoid adenoma was in a rat sacrificed at 10
months. Eliminating all rats dead of any cause prior to that
time and applying the method used for the carcinoma data leads to
7/159 for the response at 2 ppm with 1/156 at control, the same
as if all rats dead prior to an including the 18 month sacrifice
were excluded.
Table 7-3.
RISK EXTIMATES USING POLYPOID ADENOMA DATA
95 Upper
Maximun Likelihood Confidence
Category Cose Estimate of Risk Limit on Risk
Mobile Hone
Residents '
Based on HUD 0.15 ppn 1 X 10~3 3 X 10~3
Target Level (112 hrs/wk
for 10 yrs)
Manufacturers
of Apparel
1. Personal sample 0.64 ppn 8 X 10~3 2 X 10~2
(36 hrs/wk
for 40 yrs)
2. Area sample 0.25 ppn 3 X 10~3 5 X 10~3
(36 hrs/wk
for 40 yrs)
3. NIOSH data 0.17 ppn 2 X 10~3 5 X 10~3
(36 hrs/wk
for 40 yrs)
Unit Risk 1 ug/m3— 1.7 X KT4
0.00082 ppm
(for 70 yrs)
7-13
-------
7.3. Uncertainty in Risk Estimates
Model-derived risk estimates should be viewed in the proper
context. The upper bound estimate should not be viewed as a
point estimate of risk. As the Guidelines state (EPA, 1986):
"the linearized multistage procedure leads to a plausible upper
limit to the risk that is consistent with some proposed
mechanisms of carcinogenesis. Such an estimate, however, does
not necessarily give a realistic prediction of the risk. The
true value of the risk is unknown, and may be as low as zero."
Other factors are also important.
As Table 7-2 illustrates, there is a wide range between the
MLE and upper bound estimates, approximately 4 or 5 orders of
magnitude. This illustrates the statistical uncertainty of the
estimates generated due to the input data from the study used,
which in this case is highly non-linear. For instance, the
individual risks for apparel workers range from 1 X 10 [81] to
6 X 10"' [81]. In addition, it has been shown that the MLE is
sensitive to small changes in response data when the response is
very nonlinear in the experimental range. For instance, the dose
giving a risk of 1 X 10 (MLE) varies significantly due to small
changes in the response data of the Kerns et al. (1983) study
(Conn, 198Sb). The following illustrates this:
Response at 2 ppm Dose for Risk of
(malignant) 1 X 10~6 (MLE)
1. 0 (actual) 0.67 ppm
2. 1/1,000 0.0022 ppm
3. 1 0.0006 ppm
7-14
-------
Ten perturbations of the squamous cell carcinoma data for
the Fischer 344 rats were selected by slight alteration in one of
the dose-response proportions or the elimination of a dose level
from the study in an attempt to show sensitivity to these
perturbations was examined by modeling. These estimates appear
in Appendix 5. It was found that, in general, slight
perturbations of the data do not significantly disturb the
predictive power of the model for upper bound estimates. This is
not the case for MLEs. Only extreme perturbations significantly
affect upper bound risk estimates. Consequently, when modeling
data that are very non-linear, one should not place great
certainty on MLE estimates. In addition, model choice can lead
to uncertainty. As Appendix 3 illustrates, there is a wide
divergence in risk estimates obtained using the CUT rat data.
Independent background, tolerance distribution models such as,
the probit, logit, and Weibull, produce estimates indicating
virtually zero risk (probit predicts zero risk). The independent
and additive background gamma-multihit models produce similar
results. However, when additive background models are used risk
estimates are much higher, with the multistage model giving the
highest risks. As discussed in section 7.1, the linearized ,
multistage procedure was used for primary risk estimation.
As discussed above, the major contributor to the uncertainty
seen in the risk estimates using the multistage model is the
steep dose-response seen in the Kerns et al (1983) study. There
were no carcinomas at 2 ppm, 2 at 5.6 ppm, and 103 at 14.5 ppm,
7-15
-------
which is a 50-fold increase for only a 2.5 times increase in
dose. If changes in respiratory rate are taken into account (the
rats at 14.3 ppm are receiving the equivalent of a 12 ppm
exposure—use of this data leads to no significant change in
estimated risks at exposures of concern) (Grinstaff, 1985), there
is a 50-fold increase for only a doubling of the dose.
HCHO's ability to cause rapid cell proliferation, cell
killing and subsequent restorative cell proliferation, its
ability to interact with single-strand DNA (during replication),
interfere with DNA repair, its demonstrated mutagenicity, and the
fact that the dose was delivered to a finite area may help
explain the abrupt increase in the response. However, none of
these factors demonstrate the presence of a threshold or minimal
risk at exposures below those that cause significant
nonneoplastic responses such as cell proliferation, restorative
cell growth, etc. For instance, although HCHO causes varying
degrees of cell proliferation in the nasal mucosa of rats due to
HCHO exposure, it must be remembered that there is a natural rate
of cell turnover in this tissue. While it is low in comparison
to HCHO induced increases, it does provide the opportunity for
HCHO to react with single-strand DNA during cell replication
possibly resulting in a mutant ceil which, if proper conditions
are met, could result in a neoplasm. While an event such as this
may be rare, it is not unreasonable when one considers that the
opportunities for this event to occur are great due to the
immense number of cell-turnovers which may lead to defects in
7-16
-------
some cells of the population of the individuals exposed. Even
so, the marked nonlinearity of the response introduces
considerable uncertainty into any discussion of the possible
mechanism of HCHO induced carcinogenicity at exposures below the
experimental range.
The different responses seen in the animals tested also
leads to a degree of uncertainty. Although rats, mice, and
hamsters have been tested in long-term bioassays, only in rats
have statistically significant numbers of neoplasms been
observed. Only two carcinomas were seen in mice at the highest
dose in the CUT study, but the nature of this response is
complicated by the fact that mice are able to reduce their
breathing rate to a greater extent than rats. If this effect is
accounted for, the "dose" mice received at 14.3 ppm is
approximately that which the rats received at 5.6 ppm, where two
carcinomas were observed. Consequently, on a. "dose" received
basis, rats and mice may be equally sensitive to HCHO. Although
no neoplasms were seen in the hamster study, a number of factors
may be responsible. First, there was poor survival. About 40%
of the 88 hamster died before eighty weeks, and only 20 hamsters
survived ninety weeks or more. If a response comparable to that
of the CUT study were expected, 25% or five of the hamsters
surviving ninety weeks or more would have had tumors. However,
the duration of the study may not have permitted them to be
grossly visible. Second, the limited pathology protocol may not
have been able to detect small tumors. And third, the dosing
7-17
-------
regimen and physiologic factors (changes in breathing rate) may
have been factors (see section 4.1).
Although the foregoing helps explain some of the species
differences observed, there remains the possibility that other,
unknown, factors may be important. However, in any event, no
data have been developed to show that humans would respond
differently to HCHO than rats and data exist showing that rats
and monkeys respond similarly to HCHO when nasal irritation and
squamous metaplasia are used an endpoints.
It is often useful to compare lifetime excess risks
estimated from the epidemiologic studies to those risks estimated
from animal data. Tables 7-4 and 7-5 and Figure 7-1 present such
a comparison. Estimated lifetime excess risk can be determined
for either occupational or domestic exposure to HCHO. This
comparison assumes that exposure to HCHO is associated with an
increase in neoplasms at one site only and that the site-specific
excess risk observed in the epidemiological study is the excess
above a risk of one for the study population relative to the U.S.
population (Margosches and Springer, 1983). Hence, lifetime
excess risks based on the epidemiological studies are calculated
by multiplying the excess risk observed in the epidemiologic
study by the site-specific mortality ratio. 1980 mortality data
are used in this calculation.
The estimated lifetime excess risks were based on
significant associations observed in the Blair et al. (1986),
Vaughan et al. (19S6a,b), Hayes et al. (1986), Stroup, Harrington
and Oakes (1982), and Harrington
7-18
-------
Table 7-4.
Upper Bound Risk Estimates Based on
the CUT Data for Given Exposures to HCHO
Animal Based
Exposure -
Resin Worker
Furniture Worker
Patholocists
Mobile Home
Residents
(10 years)
Level (ppm)
0.
1.
0.
1.
3.
0.
24
4
1
3
2
19
Upper
5
3
1
2
6
3
X
X
X
X
X
X
Bound3
A
io'4
io-3
A
10'4
io-3
io-3
10"4
a Based on the linearized multistage model and the rat data
from Kerns et al. (19R3).
7-19
-------
Table 7-5
Estimated Lifetime Excess Risks
Calculated from the Epidemic-logic Studies
Exposure Author
Resins Blair et al.
Resin, Glue Vaughan et al.
HCHO & Wood Hayes et al .
Pathologists Harrington &
Shannon
Harrington &
Oakes
Anatomists Stroup
a Estimated lifetime excess risk
Site
Lung
Nasopharynx
Nasal
Cavity &
Sinuses
Nasal
Cavity &
Sinuses
Leukemia
Brain
Brain
r*
Risk
Ratio
1.32b
2.0C
3.8
1.9C
2.0
3.31
2.7
Estimated Lifetime
Excess Riska
2 X 10~2
8 X 10~4
7 X 1CT4
2 X 10"3
2 X 10~2
1 X 10"2
8 X 10~3
of site-specific deaths
- (KK-!) *|pr0protion
of site specific
leaths -»
Mortality proportion based on 1980 deaths.
k Analysis of white male wage workers with greater than 20 years latency
and HCHO exposure above Oppm-year.
c Analysis of white male wage worker with HCHO exposure greater than Oppm-
year.
7-20
-------
Pathologists
10
10
-4
10"
3.2 ppm
io-2
10
Stroup, Harringtons Harringto
Brain Oakes, sShannon,
Brain Leukemia
Resin Workers
0.24 ppm 1.4 ppm
o-5 i
F
»-' 1 1
p
o-3 i
D-2l ID'
Vaughan
et al.,
SNC*
Blair et al,
Nasopharynx
Blair et al.,
Lung- 20 yr
latency
Furniture Workers
10
-5
Mobile Home
Residents
10
-5
0.1 pom
1.3 ppm
10
-4
10
-3
10"
Hayes et al.
SVC, Controlled
for high wood dust
exposure
0.19 ppm
10"
10
-3
10
-2
T'aughan et al. ,
-lasooharvnx
Figure 7-1. Comparison of the upper bound risks based on the animal
data to estimated lifetime excess risks based on the epidemiological
studies. Animal-based upper bound risks for the identified exposure
level to HCKO are above the line. The estimated excess lifetime
risks based on the observed excesses in site-specific neoplasms are
below the line.
7-21
* Nasal sinus and cavity neoplasms.
10"
10
-1
-------
and Shannon (1975) studies. For example, when one examines lifetHls
risks from exposure to resins, the estimated lifetime excess risk
associated with the 35% increase in lung cancer among white males
with a greater than 20 years latency reported by Blair et al. (1986)
— 2
would be 2 X 10 and the estimated lifetime excess risk associated
with their reported 200% increase in nasopharyngeal cancers would be
8 X 10" . The 280% increase observed by Vaughan et al., (as
reported in SAIC, 1986) for nasal sinus and cavity neoplasms in
conjunction with exposure greater than 10,000 hours to resins,
glues, and adhesives gives an estimated lifetime excess risk of 7 X
10" . The upper bound risk for an exposure of 0.24 ppm HCHO based
on the animal data is 5 X 10" , and for an exposure of 1.4 ppm
HCHO, would be 3 X 10~3.
Comparing the results reported by Hayes et al. (1986) is
more complicated since Hayes et al. do not delineate the exposed
population. However, if one chooses an exposure group, such as
furniture workers who may be exposed to both wood dust and HCHO,
one can make some observations. The reported exposure for
furniture workers ranges from 0.1 ppm to 1.3 ppm HCHO as an
8-hour, time-weighted-average. Upper bound risks based on the
animal data associated with these exposures are 1 X 10"** and
2 X 10" , respectively. Using the 90% increase in nasal cavity
and sinus risk observed in analyses which controlled for high
wood dust exposure, the estimated lifetime excess risk based on
the Hayes et al. study would be 2 X 10~3.
Thus, when individual tumor types are examined, one can see
that the upper bounds are not indicating larger excesses than
7-22
-------
seen in certain studies given uncertainties about exposure.
Although HCHO's potential carcinogenic effects are not expected
to be limited to one site in humans because humans do not
necessarily breathe through their noses as rats do, the analysis
described above provides a check of the risks derived from animal
data and those seen in human studies.
Finally, a factor that can have a major bearing on
population risk estimates is the quality of the available
exposure data. Assumptions made in reporting exposure levels can
have a major impact. For instance, it is not uncommon during a
monitoring exercise to find a number of samples that are below
the detection limit of the analytical technique used. Thus, when
a mean exposure level is calculated it should be realized that if
the nondetectable (ND) samples are counted as 0 the calculated
mean will understate the actual situation. Conversely, if the ND
samples are counted as the limit of detection, the mean win
overstate the true situation. Another factor that can skew
exposure estimates are changes in non-governmental exposure limit
recommendations and the number of years over which the data are
collected. Since a number of years of exposure data are often
used to calculate means, it is possible that the mean will be
weighted by samples taken prior to changes in voluntary exposure
limits. Thus, the reported mean could be substantially
overestimating the true situation. For instance, in the garment
industry, HCHO levels have apparently been falling since the Late
70's and early 80's as a result of increased concern and a
7-23
-------
downward revision of the ACGIH recommendation for HCHO.
Consequently, an industry average calculated from data predating
1980 could cause the reported mean to be overstated. This may
have a significant impact on the estimated population risks. For
the apparel industry there are approximately 800,000 workers
exposed to HCHO. The mean personal exposure level used for the
section 4(f) determination and this assessment is 0.64 ppm which
leads to population risks of <1-777 (MLE-upper bound). If the
mean area exposure level of 0.23 ppm is used, and there is some
evidence that personnel levels may now be approaching this
figure, population risk estimates would range from <1-311, which
is a 60% induction at the high end. However, the exposure data
for apparel workers are poor in its ability to characterize the
industry, and great confidence cannot be placed on an industry
mean as a fair representation of actual exposure levels in the
approximately 20,000 sites where workers are exposed.
The data for mobile homes is qualitatively better in its
ability to characterize this group because of a greater number of
well conducted monitoring surveys. Mobile home exposure studies
have been done by HUD, state and local government agencies, and
academic researchers. Although data is generally only available
to estimate 10 year averages, data from complaint and non-
complaint homes produce 10 year averages which range from 0.19 to
0.25 ppm.
7-24
-------
7.4. Presentation of Risk Estimates
As discussed in "Data Selection for Quantitative Analysis"
above, the recommendation is that risk estimates should be
separately derived from squamous cell carcinoma and polypoid
adenoma data. However, three positions can be taken concerning
the presentation of the risk estimates. One is to calculate risk
estimates separately for squamous cell carinomas and polypoid
adenomas. The second is to add the risk estimates for an overall
estimate of carcinogenic risk, and the third is to assume some
conversion rate for the benign tumors and then add the risk
estimates as in the second position. These are discussed below.
7.4.1. Separate Risk Estimates Derived From Squamous
Cell Carcinoma and Polypoid Adenoma Data
Because two risk estimates can be calculated, the
significance and uncertainties associated with each must be
explained.
The squamous carcinomas observed in the Kerns et al. (1983)
study are frank evidence of carcinogenicity in the rat. The
response at 14.3 ppm HCHO was highly significant in both sexes.
While not significant at 5.6 ppm, the observation of two squamous
cell carcinomas in 240 rats is considered biologically
significant, since the historical incidence of squamous cell
carcinomas in male and female F-344 rats is 1 in 3,000 rats (NTP,
1985). In addition, significant numbers of squamous carcinomas
were observed in rats in two other long-term inhalation studies
(Albert et al., 1982; and Tobe et al., 1985). Consequently,
there is little uncertainty about the carcinoma results.
7-25
-------
There is a positive dose-response relationship for squamous
ceil carcinomas in the Kerns et al. (1983) study. However,
because of the nonlinearity of the dose-response relationship,
there is a wide divergence between the upper bound and maximum
likelihood estimates (MLE) of risk. This introduces a large and
variable level of uncertainty into the risk estimates (see
preceeding section—Uncertainty in Risk Estimates).
The situation for the polypoid adenoma data is not clear.
Although apparently causally related to HCHO exposure, the
statistical significance is poor. The adenomas are not
significant at any dose level for male and female rats
separately. Only when the response is pooled at 2.0 ppm is there
significance. However, even this is questionable since the
response rate in male and female rats is not comparable. At the
5.6 ppm level responses in males and females were significantly
different from one, another. ' Moreover, two of the responses at
2.0 ppm were borderline diagnostic calls between focal
hyperplasia and polypoid adenoma (Boorman, 1984), and if these
two responses are dropped, significance is lost at 2 ppm. Also,
there is a negative dose-response relationship. Several possible
explanations for these observations follow in roughly increasing
order of likelihood (SAB, 1985):
o lack of a causal relationship,
o tumor modulating factors in the rat, which are induced
or enhanced by HCHO exposure,
o chance (random) fluctuations in the data,
o target size decreases with loss of ceil type of origin,
7-26
-------
o differences in time to tumor,
o differences in diagnostic efficiency between the two
kinds of tumors, and
o competition with the simultaneously occurring carcinomas
. at higher doses.
Thus, it is difficult to adequately characterize the relationship
between HCHO exposure and the polypoid adenoma response. Because
of the negative dose-response relationship, it was necessary to
drop the two highest doses (5.6 and 14.3 ppm) and extrapolate
from 2.0 ppm to 0 (a straight line from 2.0 to 0 ppm). Thus, if
the true dose-response relationship is concave between 0 and 2.0
ppm, estimated risks will be too low. If upward convex they will
be too high (this seems more plausible given the benign and
malignant tumor responses in the Tobe et al. (1985) study and the
squamous cell carcinoma response in the Kerns et al. (1983)
study).
As discusssed above, there is a greater level of certainty
in the squamous cell carcinoma response and risk estimates
derived from them. Conversely, because of the nature of the
polypoid adenoma response, its weak statistical significance at
best, and the manner of risk estimation, the confidence in this
response and associated risk estimates is low.
7.4.2. Calculate Risks Separately But Add The Risks
The rationale for this option is that the polypoid adenomas
together with the squamous cell carcinomas observed in the Kerns
et al. (1983) study are an indication of HCHO's potential human
carcinogenicity. Moverover, benign tumors may be expected to
7-27
-------
appear in the human population (not just in the nasai cavity).
It may also be assumed that they have some ability to progress to
cancers as a result of the promoting activity of other agents or
of the initiating agent. Consequently, adding the risk estimates
from the'benign (polypoid adenomas) and malignant (squamous cell
carcinomas) data provides an overall estimate of carcinogenic
risk to humans.
While such a line of reasoning is plausible, a number of
factors must be considered.
First, if the separate risk estimates are added, that
estimated, from the squamous cell carcinoma data is dwarfed by the
estimated adenoma response. For instance, the upper bound
estimate of risk to garment workers exposed to 0.64 ppm of HCHO
is 1 X 10~3 using squamous cell carcinoma data. The risk
estimate based on benign tumors at the same concentration is
2 X 10~2. Adding the two estimates gives 2.1 X 10~2. Following
the Guidelines (EPA, 1986) this would be rounded to one
significant figure, i.e., 2 X 10". Thus, the contribution to
the risk estimate from the frank experimental evidence of
carcinogenic!ty is removed. In addition, the uncertainties
unique to estimates of risk based on the squamous cell carcinoma
and polypoid adenoma data are not carried clearly forward in a
combined estimate of risk.
The second, and major assumption is that there is
equivalence between benign and malignant tumors, i.e. a benign
tumor will progress to a cancer. This is necessary because the
7-28
-------
combined risk estimate is nearly, entirely weighted by a risk
component generated from the benign polypoid adenoma tumor
data. Since the estimate is presented as a cancer risk estimate,
equivalence (progression) must be assumed. However, the basis
for this assumption must be reviewed. Certainly there is much
literature on the progression of benign tumors/ but equivalence
is not automatically assumed, especially when the experimental
study (the Kerns study) suggests otherwise (see section 4.2.1).
It may not be correct to assume that the majority of tumors
estimated for the human population from the Kerns study will
occur in the nasal cavity, since humans are not obliged to
breathe through their nose. Consequently, it may be worthwhile
to look at the nature of benign tumors seen in the nasal cavity
of humans and animals as well as in other epithelial tissues in
humans.
As discussed in section 4.2.1., the nature and progression
of benign tumors in the nasal cavity of rats is poorly
understood. The polypoid adenomas observed in the Kerns et al.
(1983) study do not appear to be the benign counterparts of the
squamous cell carcinomas or other cancers observed. The
situation for humans is similar, although based on clinical
experience some generalizations can be made. However, it must be
remembered that the clinical cases are .the result of diverse
causes and may not share the same course as an HCHO-induced
lesion. The following discussion is presented to highlight the
uncertainty involved in any discussion of cancer induction.
7-29
-------
The common types of benign lesions seen in the nasal cavity
of humans are nasal polyps, squamous papillomas, and transitional
type papillomas.
Nasal polyps are a common clinical condition in humans and
are frequently associated with allergic rhinitis, inflammatory
diseases, and other disorders (Paludetti, 1983; Jacobs, 1983;
Frazer, 1984; Drake-Lee, 1984). These polyps are not considered
to be true neoplasms, but are merely inflammatory hypertrophic
swellings (Robbins, 1974).
On the other hand, squamous and transitional type papillomas
are true neoplasms. Squamous papilloma of the vestibule is the
most common tumor of the nasal cavity, representing approximately
one-third of all benign tumors found. Malignant change is
considered a rare event (Friedmann and Osborn, 1982).
Transitional type papillomas have an incidence that is reported
to vary from 0.4 to 19 percent of all nasal and sinus neoplasms
of the mucosa (Bosley, 1984; Friedmann and Osborn, 1982; Hyams,
1971; Sellars, 1982; Lampertico et al., 1963; Seydell, 1933).
Their clinical appearance may vary from that of firm, bulky,
opaque polypoid lesions with marked vascularity to having the
same appearance as common inflammatory nasal polyps (Bosley,
1984; Perzin et al., 1981), and are variously described as
inverted squamous papilloma, cylindrical or transitional ceil
papilloma, and inverting papilloma (Friedmann and Osborn,
1982). The reported associated frequency of squamous cell
carcinoma with transitional papilloma is between 1.5 to 50
7-30
-------
percent (Bosley, 1984; Friedmann. and Osborn, 1982; Hyams, 1971;
Snyder et al. 1972; Ridolfi et al., 1977; Lasser et al., 1976;
Vrabec, 1975; Osborn, 1970; Yamaguchi et al., 1979; Brown, 1964).
The most common benign mucosal gland tumor is the
microcystic papillary adenoma, which is the human counterpart of
the rat polypoid adenoma (Kerns, 1985). In humans, these tumors
represent 1.6 percent of all tumors of the nose and sinuses and
2.4 percent of all tumors in the nasal cavity region. In
addition, malignant transformation has never been encountered
(Friedmann and Osborn, 1982)
A. number of benign tumors are seen in the oral mucous
membrane of humans. Fibromas, papillomas, hemangiomas,
lymphangiomas, and less commonly myoblastomas and congenital
epulis. However, in contrast approximately 90 percent of oral
malignancies are squamous cell carcinomas (Robbins, 1974).
The two most common benign tumors of the human larynx are
polyps and papillomas, other less common types run the gamut of
every cell type found within the larynx (Robbins, 1974).
Squamous papillomas are the most common type of benign tumors
seen in the larynx and are the most common of all childhood
laryngeal tumors. These are frequently divided into adult and
juvenile groups. However, recent work has contradicted some of
the classical descriptions used to separate adult from juvenile
papillomas (Nikolaidis, 1985).
However, while juvenile papillomas are thought not to, or to
rarely, undergo malignant transformation (Nikolaidsis, 1985; and
7-31
-------
Robbins, 1974) the adult type is.regarded as having the potenital
to progress to a malignancy (squamous cell carcinoma) (Robbins,
1974). In a study of 83 cases (73 juvenile and 10 adult) of
tumors of the larynx, only one adult case was associated with a
squamous cell carcinoma (Nikolaidis, 1985). This suggests a less
than one-to-one relationship between papillomas and carcinomas:
it should be pointed out that there was surgical intervention and
no follow-up.
Although the above does not show that benign tumors caused
by HCHO will not progress to a malignant neoplasm, it does show
the great uncertainty involved in assuming that there is a one-
to-one relationship between risk estimates generated from benign
and malignant data sets.
7.4.3. Calculate Risks Separately But Add the Risk After
Assuming a Conversion Rate for the Benign Tumors
This option is the same as the option described in section
7.4.2 except that the risk estimates generated from the benign
tumor data would be adjusted to reflect the potential to progress
to malignancies. This method may provide a more realistic
estimate of carcinogenic risk, but it still suffers from the
problem of adding estimates derived from different extrapolation
procedures.
However, if one were to assume certain conversion rates
based on the bioassay and human data, overall estimates of cancer
risk can be presented.
7-32
-------
In the Kerns study, 15 polypoid adenomas were observed in
the exposed groups (see Table 4-1) versus one adenocarcinoma
(possible malignant counterpart), which is a 1:15 ratio.' For the
most common human benign tumor of the nasal mucosa, transitional
type papilloma, Friedmann and Osborn (1982) have documented 12
possible conversions related to between 700 and 800 papillonas, a
ratio of 1:50. In contrast, the human counterpart of the rat
polypoid adenoma is the microcystic papillary adenoma (Swenberg
and Boreiko/ 1985) which has never been reported to convert to
malignancy (Friedmann and Osborn, 1982). In addition, multistage
carcinogenesic protocols on mouse skin and in rat liver produce
malignant to benign ratios of 1:20 to 1:100 (Swenberg and
Boreiko, 1985). As discussed above, the conversion rates range
from 0 to about 7 percent. Since there is uncertainty in any of
the ratios discussed, assuming a 1:10 ratio, or a 10 percent
conversion rate, appears reasonable. If the estimates in Tables
7-2 and 7-3 are combined using a 10 percent conversion rate, then
risk estimates would be about a factor or 2.5 higher than those
based on the malignant tumor data.
7.4.4. Other Conaiderations-Squamous Papillomaa
In contrast to the Kerns study, the studies by Tobe and
Albert found significant numbers of squamous papillomas instead
of polypoid adenomas. Because these tumors are of the same cell
type as the squamous cell carcinomas, these lesions are thought
to represent the benign counterpart of the carcinomas (Consensus
Workshop on Formaldehyde, 1984). However, Takano et al. (1982)
7-33
-------
have questioned the importance of papillomas in relation to
carcinoma development. In any event, papillomas were seen in two
HCHO studies suggesting that both types of benign tumors should
be evaluated.
The relationship between the papilloma/carcinoma response in
the Tobe and Albert studies is quite constant. In the Tobe
study, 5 of 19 tumors observed or 26 percent were papillomas. A
similar result was seen in the Albert study were 34 of 110 tumors
or 24 percent were papillomas. The papilloma response is clearly
HCHO related and appeared in two strains of rats (Fischer 344 and
Sprague-Dawley rat). It is not clear why no papillomas were
observed in the Kerns study. However, the behavior of papillomas
seen in the human population should also be evaluated since there
is no reason to assume that the polypoid adenoma response in the
CUT study is more important in determining human risk than the
papilloma response.
To determine the contribution benign tumors would make to
risk estimates derived from the Tobe study, three data sets were
used to derive risk estimates. (These estimates were not used as
primary estimators of risk because there was no response at the
lower dose levels which consequently leads to higher estimates of
risk than those derived from the CUT data. ) One consisted of
the benign tumors (5 per 32 rats), the second was the carcinoma
response (14 per 32 rats), and finally combined malignant/benign
(19 per 32 rats). The results from the exercise appear in Table
7-6.
7-34
-------
Table 7-6.
RISK ESTIMATES BASED ON TOBE STUDY
Response at 15.0 Risk Estimate Added Risk
ppm (32 rats) (upper bound) Estimates
Papilloma 5 2.3 X 10~3 5.1 X 10'3
Carcinoma 14 2.8 X 10~3
Pooled 19 2.8 X 10~3
carcinoma/
papiiloma
As Table 7-6 illustrates, the benign tumors do not have any
impact on the risk estimate derived from the pooled data. If one
were to simply add the separate benign and malignant risk
estimates, the resulting estimate would be about two times higher
than that derived from the pooled data. Discounting the benign
estimate by 90 percent (assumes 10 percent will progress) and
then adding it to the malignant estimate gives a combined number
that is only slightly higher (3.0 X 10~3) than that derived from
the pooled data (2.8 X 10~3). This illustrates some of the
uncertainty one encounters when using different methods of risk
estimation and, in this case, shows that if pooled or a
conversion rate is used, the benign tumors contribute little to
the risk ••timates.
7.4.5. Conclusion
Because of the uncertainties associated with the polypoid
adenoma data set, its statistical significance, the manner of
risk estimation, and the question of progression to malignant
7-35
-------
tumors, it ia recommended that unadjusted risk estimates (for
progression) derived from them not be added to estimates derived
from squamous cell carcinomas. It is recommended that risk
estimates derived from squamous cell carcinomas be used to
estimate human risk because of the frank expression of
carcinogenicity in the rat, evidenced by a statistically
significant, positive dose-response relationship. Little weight
should be accorded risk estimates derived from adding adjusted
benign risks to carcinoma risks because of the uncertainties (1)
inherent in adding risk estimates derived from different
mathematical procedures, (2) the nature of the benign tumor
response, and (3) uncertainties surrounding the rate of benign to
malignant conversion.
7.5. Summary
Although a number of factors that represent more or lesser
degrees of uncertainty have been discussed above in relation to
the quantitative estimates of human cancer risk, no factor alone
or in combination with others indicates that the estimates of
risk are not reasonable as upper bounds. The true risk could
certainly be lower, but no data or modeling procedure is
available to determine the true risk. Consequently, it is
recommended that the risk estimates derived from the CUT rat
squamous carcinoma data be used as the estimates of potential
human cancer risk from exposure to HCHO with due consideration
given to the strengths and weaknesses of the data base.
7-36
-------
8. ESTIMATES OF HONCAHCER RISKS
8.1. Introduction
Although some of HCHO's noncarcinogenic effects are well
characterized, the problem of determining the dpse-reponse
characteristics in populations for these effects remains. This
section focuses on human data to determine if dose-response
relationships can be drawn. Six cross-sectional and three
controlled human studies were selected for review. These studies
were identified by Battelle Columbus (1985) from review articles by
Ulsamer et al. (1984) and the NRC (1982). From this group, six
studies were extensively analyzed for presence of dose-response
relationships. Additional searching identified six studies which
were thought to have shown a possible dose-relationship; these were
also analyzed. For a study to be selected for review, the exposure
level and the prevalence or incidence of a health effect must have
been identified.
Generally, an increase in the prevalence or incidence of eye,
nose, or throat irritant effects with increasing HCHO exposure was
observed across the studies. Since exposures are identified as
ranges in seven of the studies, reductions in the prevalence or
incidence of the irritant endpoints from small changes in HCHO
levels are difficult to quantify.
One study of randomly selected mobile home residents (Hanrahan
et al., 1984) reports a dose-response relationship between the
reporting of eye irritation and HCHO level. Three clinical studies
of volunteers (Kulle, 1985, Andersen and Molhave, 1984, and Bender
et al., 1983) report irritant responses over several exposure
8-1
-------
levels and EPA analyzed these data in a similiar manner as that of
Hanrahan et al. (1984). The predicted response curve of Andersen
and Molhave and Bender et al. are similiar to that of Hanrahan et
al., but the response curve for Kulle predictes lower percentage
response,.
None of the twelve studies provides adequate data to quantify
population risks for the irritant effects of HCHO. The studies, at
best, provide a qualitative estimate of population response over a.
wide exposure range and quantitative estimates of responses for
very selected groups.
8.2. Studies Reviewed
Studies examining ocular or nasal effects are reviewed since
studies of these endpoints comprise the majority of literature
which reports both exposure level and magnitude of the effect.
Studies which examined dermal responses were not selected since
exposure is by either patch testing or dermal injection. Results
from this route of exposure are often difficult to generalize for
dose-response relationships. The reviewed studies are of two
designs: cross-sectional and controlled clinical experiments. A
search of the literature did not reveal any case-control designed
study. The cross-sectional studies were of mobile home residents
and of occupationally exposed workers. In the clinicai studies,
small groups of healthy volunteers, fewer than 30, were exposed to
varying concentrations of HCHO and their responses were recorded.
8.3. Limitations of Studies
Even though response trends are identified for individuals
under study, the studies reviewed have major limitations which
8-2
-------
prevent their use to infer the magnitude of general population
risks.
8.3.1 Study Design Limitations
The data on acute effects come from controlled human studies
or cross-sectional studies. The majority of the studies were
designed as cross-sectional studies, also known as survey
studies, where random or nonrandom sampling frames were
employed.* In addition, two of these studies' designs did not
incorporate a nonexposed or control group (Garry et al. and
Anderson et al.). Without a control group, it is impossible to
determine the attributable magnitude of a reported symptom.
A cross-sectional study measures the study factor level, in
this instance the HCHO level, and disease outcome at the same
time. This type of study does not incorporate a follow-up
period, so that the prevalence of the disease outcome, and not
the incidence, is obtained.** Cross-sectional studies are often
used to generate hypotheses, but they have serious limitations in
making causal inferences.
Controlled human studies test etiologic hypotheses and can
identify acute effects. The three controlled human studies
identified did not utilize a randomization scheme. Study
participants were self-selected and may not be representative of
the general population.
* Use of nonrandom sampling prevents extrapolating the results
beyond the studied population. Random sampling, on the other
hand, allows statistical inference from the studied population
to the general population.
**Prevalence is the number of cases existing with the outcome at
a single point in time. Incidence is the number of new cases
observed over a period of time.
8-3
-------
8.3.2 Bias Limitations
Other limitations regard potential bias in the results. The
cross-sectional studies do not control for confounding exposures,
such as other occupational exposures, and an observed effect may
not solely be due to HCHO exposure. Second, many studies report
subjective or self-reported symptoms. These symptoms have not been
medically verified, and thus results may be biased by over-
reporting or under-reporting. A third problem of the cross-
sectional studies concerns the quality of the environmental
exposure measurement. HCHO levels vary depending on the season of
the year, hour of the day, temperature, and humidity. These
factors must be known to evaluate the intensity of the disease
endpoint.
8.4 Results
The principal acute effects of HCHO, reported by all studies
which were extensively examined, are those of irritations to eyes,
nose, throat, upper respiratory tract, and skin. Tables 8-1 and 8-
2 list, by study design, the prevalence or incidence of each effect
and the associated HCHO level.
Evaluation of the results documented ia the different papers
indicates that these effects exist in varying degrees in people
exposed to HCHO. The exposure levels may range between 0.037-3.0
ppm. However, the intensity of the symptoms differs depending on
the location of the study (mobile homes, industry, anatomy lab), on
ambient air conditions, and on individual characteristics and
personal habits.
8-4
-------
TABLE 8-1 SUMMARY OF SELECTED CHOSS SECTIONAL STUDIES3'b
EXPOSURE
LEVEL (ppn) SYMPTOMS
ESTIMATED
SUBJECf OF STUDY PREVALENCE STUDY COMMENTS
<0.02-0.78
0.037
oo
I
en
Cold Symptons
Headaches
Sneezing
Any nasal
abnormality
Seborrhea
Pressure in chest
Sneezing
Inflamed
mucous membranes
Persistent courjh
& phlegm
Itch
Rash
Shortness of
breath
Chest sputnum
Burning sensation
in heart region
24C
64C
91C
females (>16 yrs) 2.
females (5-15 yrs) 4.
females (>16 yrs) 1,
males & females 1.73^
Oil yrs)
males & females 11.85d
males (>16 yrs) 11.74°
females (5-15 yrs) 2.86C
males & females 4.49d
Oil yrs)
50 non- 10%
hexamethylene-
tetramine workers 23%
re soreino1 17%
workers in a 17%
tire
manufacturing 20%
plant 11%
Texas Indoor
Air Quality
Study
Study ot mobile homes.
Gamble et al
HR exposed workers had si(jnitic\in(
reductions in expiratory flow at In
lung volumes, indicating
resistance in small airways
compared to a rjroup of workers
from the entire plant.
a
Rattelle Columbus (
Only those studies when; l»otli exposure and prevalence are represented.
h iTev.'jIi'iu'i? r~.it jo c.il i -ul.ii »l I nin the il.il d
-------
TABLE 8-1 (Continued)
EXPOSURE
LEVEL (ppm)
SYMPTOMS
SUBJECT OF STUD*
ESTIMATE!)
PREVALENCE
STUD*
CCMMENPS
0.02-0.05
0.04-0.09
0.19-0.44
CO
I
Persistent cough
& phlegm
Itch
Rash
Shortness of breath
Chest sputnum
Burning sensation
in heart region
Eye irritation
Nose, throat
irritation
Eye irritation
Sinusitis
Nose, throat
irritation
52 hexamethylene-
tetramine
resorcinol workers
34 permanent day
care center staff
13%
35%
23%
19%
23%
17%
15%
23%
70 mobile hone day 57%
care center staff 16%
73%
Olsen and
Dossing
Control group and 34
mobile home day care centers.
-------
TABLE 8-1 (Continued)
EXPOSURE
LEVEL (ppm)
<0. 10-2.84
0.1-0.8
03
1
0.1-3.0
SYMPFCMS
Burning eyes
Ifetering eyes
Dry throat
Swollen glands
Diarrhea
Running nose
Sneezing
Phlegm
Inlteezing
tough
Headache
Rash
Runny nose
Dry/Sore throat
toughing
Ear, nose, throat
Cough & wheeze
Respiratory
problan
Diarrhea
Headaches
Nausea & v/oniting
Skin rash
SUBJECT OF STUD*
Residents of
mobile hones
Residents of
Mobile hones
Adults
0-12 yrs
Adults
3-12 yrs
0-2 yrs
Adults
0-12 yrs
Adults
3-12 yrs
0-2 yrs
Adults
3-12 yrs
0-2 yrs
Adults
3-12 yrs
0-2 yrs
Aiults
ESTIMATED
PREVALENCE STUD* CQMMENFS
25% Anderson Study of mobile hoaes.
20% et al.
24%
6%
10%
35%
45%
25%
17%
44%
29%
11%
34% Hanrahan et al. Study of mobile hones.
33%
28%
79% Garry et al. Mobile hones.
60% Measurement of formaldehyde
38% vary with month of measurement.
54%
61%
36%
24%
22%
19%
58%
50%
38%
0%
20%
15%
38%
0*
m
-------
TABLE 8-1 (Continued)
EXPOSURE
LEVEL (ppm) SYMPTOMS SUBJECT OF STUDY
0.40-0.806 Cough Present-line
>5 yrs
1-5 yrs
<1 yr
Previous on line
Never on line
Phlegm Present- line
>5 yrs
i 1-5 yrs
00 <1 yr
Previous on line
Never on line
Dyspnea Present line:
>5 yrs
1-5 yrs
<1 yr
Previous on line
Never on line
ESTIMATED
PREVALENCE STUDY
Schoenberg
33% & Mitchell
30%
40%
12.5%
6.7%
26.7%
20%
26.7%
0.0%
6.7%
6.7*
20.0%
20.7%
12.5%
6.7%
COMMENTS
63 filter manufacturing workers
No significant differences
(p>.05) among any of the groups
in either FVC or FEVj 0. The
group, present line more than
5 years or more, had a lower
FEVj Q/KVC ratio; anU signifi-
cantly lower (p<0.05) MEFcQ/FVC
than the never-on-line group.
u
Kx pi is ure
ure tor pi c.seri t-on- I i ne-woi ket a only
-------
TABLE .8-2 (Continued)
Level of
Bxposure (ppn)
3.0
5.0
Symptom Study Subject
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Odor perception
Conj unct ival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Response Author
30
80
75
15
20
190
200
10
0.24
0.40
0.80
1.60
Conj unct ival
irritation and
dryness in nose,
throat
Conjunctival
irritation and
dryness in nose,
throat
Conjunctival
irritation and
dryness in nose,
throat
Conjunctival
irritation and
dryness in nose,
throat
Healthy students
2C (19%)
5 (31%)
15 (94%)
15 (94%)
Andersen
and •
c Number of conplaints among 16 subjects after a 5-hour exposure to formaldehyde.
3-10
-------
TABLE 8-2 SUMMAPY OF SELECTED CONTROLLED HUMAN STUDIES
Levels of
Exposure (ppn)
Symptom
Study Subject
Response
Author
0.1
0.2
0.5
1.0
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Odor perception
Conjunctival
sensitivity
Nose, Throat
sensitivity
Throat dryness
Anatomy lab students 133
Each group contains 25
six students
20
14
15
21
15
35
35
35
2
30
18
20
4
40
30
40
2
Rader*3
^Response represents the log of a weighted average of the concentration x time-factor.
°Dose response trend was observed for all complaints.
3-9
-------
TABLE 8-2 (Continued)
Level of
Exposure (pan)
Symptom
Study Subject
Response
Author
0.0
0.5
1.0
2.0
3.0
Eye/>tose/Throat
Irritation
Eye Irritation
Nose/Throat
Irritation
Eye/Nose/Throa t
Irritation
Eye Irritation
^se/Throat
Irritation
Eye/Nose/Throa t
Irritation
Eye Irritation
Nose/Throat
Irritation
Eye/Nose/Throa t
Irritation.
Eye Irritation
>fose/Throat
Irritation
Eye/Nose/Throat
Irritation
Eye Irritation
Nose/Throat
Irritation
Healthy Volunteers 3d (14%)
1 (41)
3 (14%)
le (10%)
0 (0%)
1 (10%)
6d (27%)
4 (18%)
2 (9%)
12d (55%)
10 (46%)
7 (32%)
9r (100%)
9 (100%)
2 (22%)
Kulle
dA total of 22 subjects were exposed
? A total of 10 subjects were exposed
A total of 9 subjects were exposed.
9-11
-------
Table 3-2 (continued,'
Level of
Exposure
(ppm)
0
0.35
0. 56
0.7
0.9
1.0
Sy. Tip torn
Eye
Eye
Eye
Eye
Eye
Eye
Study Subject Response Author
irritation Research Staff - Bender at al.
irritation 41.7%gh
irritation 53.8%^-
irritation 57. l%3
irritation 60.0%k
irritation 74. 1%^
9 Subjects with HCHO response time less than clean air response time.
n Total of 12 subjects were exposed.
L Total of 26 subjects were exposed.
J Total of 7 subjects were exposed.
1
Total of 5 subjects were exposed.
Total of 27 subjects were exposed.
8-12
-------
Five of the studies (e.g.. Texas Air Quality, 1983, Anderson
et al., 1983, Garry at al., 1980, Olsen and Dossing, .1982 and
Hanrahan et al. , 1984) were of occupants of mobile homes. Among
these studies/ the symptoms with the highest prevalence across
different age groups were headaches, muscle aches, eye symptoms
(burning eyes, watery eyes, itchy eyes), nose symptoms, and
coughing. Some differences were detected among the different age
groups. The youngest groups (0-2 yrs) experienced a higher rate
of diarrhea while the adults had a relatively high rate of
complaints from headaches (Texas Air Quality, 1983). Three
studies (Texas Air Quality, 1983; Anderson et al., 1983.
Hanrahan, 1984) report significant (p<0.05) dose-response
relationships between certain acute effects and HCHO level. The
Texas Air Quality (1983) study reports significant increases in
the prevalence of certain acute effects such as headaches,
sneezing, and nasal symptoms among occupants exposed to either
1.0 ppm HCHO or greater, or 2.0 ppm HCHO or greater. Anderson et
al. (1983) reported in occupants of a random sample of 100 mobile
homes that burning of the eye was significantly associated with
the level of HCHO in the home. Not only did the prevalence of
burning eyes increase significantly with increasing mean HCHO
level, but also the proportion of individuals who believed their
burning ey«a were related to household conditions went from 50
percent to 87 percent. Hanrahan et al. (1984) reported a
significant dose-response relationship between burning eyes/aye
irritation and HCHO among study volunteers who lived in mobile
homes. This observation had been adjusted for age.
9-13
-------
When studies of mobile home residents are evaluated, it nust
be noted that not only ambient conditions within the hone, but
also seasonal temperature/humidity fluctuations can affect the
rates of off-gassing (Anderson and Lundquist, 1985). Because
most mobile homes are tightly sealed and do not use a continuous
influx of outside air, other gases such as carbon monoxide, which
were not measured, may contribute to the acute effects
experienced by the residents. Hanrahan et al. (1984), however,
stated that these factors did not influence HCHO levels in their
study.
In three studies of workers in occupational settings,
statistically significant increases in the number of complaints
from acute sensory effects were observed among workers exposed to
HCHO. The reported symptoms — itch, rash, breathing better away
from work, cough, chest tightness, burning eyes, running nose ^ni
burning sensation in the heart region — all were significantly
increased in a group of rubber workers exposed to a HCHO-resin
when compared to non-HCHO exposed workers (Gamble et al.,
1976). Assessment of lung function in these workers showed
significant reductions in expiratory flow rates, with the
greatest reductions being shown by smokers. Area sampling of
formald«hyd« showed a mean concentration of 0.06 mg/m3 for those
workers Who were in the exposed group. In another study of
acrylic-wool filter manufacturing workers exposed to phenol-
formaldehyde resin, Schoenberg and Mitchell (1975) observed
significant increases in the prevalence of cough and cough-plus-
phlegm symptoms when compared to never-on-line and previous-line
8-14
-------
workers. Workers exposed for 5 or more years had lower lung
function parameters (FEV^/FVC ratio) than a group of workers who
had smoked more but who had never been consistently exposed to
resin fumes. Breathing zone measurements of HCHO ranged 0.40 -
0.80 ppm in this study, with higher levels (8.48 - 13.04 ppm)
observed when cross-current fans were not operational. Kerfoot
and Mooney (1975) also reported nose and eye irritation in
morticians. These results are qualitative in nature; they
indicate that morticians who spent more time embalming than in
general funeral work more often complained of upper respiratory
irritation. For all three studies, other chemical exposure's were
present and it is not known to what degree the observed effects
were due to possible interactions.
Four studies were of volunteers in controlled clinical
experiments. Findings from these studies are similar to those of
mobile home and occupational populations. Rader (1975), in
testing six student volunteers in an anatomy laboratory, found
that the concentration levels of HCHO in ambient air are affected
by seasonal changes, time of measurements, room temperature, and
humidity level. Dose and response showed correlations and there
was a statistically significant increase in each of the dose
groups ov«r the control group for the total complaint score* of
acute effects. The summed complaint score was for the acute
effects: odor perception, conjunctival sensitivity, nose/throat
irritation, throat dryness, nasal secretions, and tear flow.
*The complaint score was a sum of the number of complaints tines
the severity of the response.
9-15
-------
In a study by Bender at al- (1983), varying size groups of
volunteers preselected by responding positively to HCHO at 1.3
and 2.2 ppm HCHO, were exposed to HCHO or clean air for six
minutes (0, 0.35, 0.56, 0.7, 0.9, or 1.0 ppm). Eye irritation
was measured as response time which was the length of time from
initial exposure of the subject's eyes to the gas until eye
irritation was noticed. Subjects were also asked to rate the
severity of the response using a 0-3 scale (0=none, l=slight,
2=moderate, 3=severe).
Although only the response at 1.0 ppm was statistically
different than clean air, there was a trend toward earlier
response to HCHO with increasing concentration. If the exposure
groups for 0.7 and 0.9 ppm had been larger (5 and 7 subjects,
respectively), the response might have been statistically
different than clean air.
Severity of response was rated slightly to moderately
irritating only at 1.0 ppm. The rating was less than slightly
irritative for 0.35 to 0.9 ppm. In addition, severity was rated
lower at the end of the six minute exposure indicating dimunition
of response. This effect has been noted by Weber-Tschopp et al.
(1977) and Kane et al. (1977) (Bender et al.).
Andersen and Molhave (1984) assessed the human health
effects associated with prolonged exposure to HCHO under
controlled thermal and atmospheric conditions. They observed an
increasing trend in eye and nose irritation between exposure
levels of 0.3 to 2.0 mg/m^ HCHO. Among 16 subjects, human
response increased from 19 percent to 94 percent over this
8-16
-------
exposure range. In addition, mean mucous flow rate decreased at
the higher concentrations of HCHO. Changes in airway resistance
were significant for nasal pressure drop, vital capacity, and
several lung function parameters.
Finally, Kulle (1985), of the University of Maryland,
examined irritant symptomology among volunteers who were exposed
randomly to HCHO concentrations of 0.5, 1.0, 2.0, and 3.0 ppm.
Odor and irritation determinations were made before exposure and
at 180 minutes after exposure completion. Statistically
significant increases in the number of eye and eye/nose/throat
combined symptoms were observed for exposures over 2.0 ppm. The
number of subjects detecting HCHO odor was statistically
increased at HCHO levels of 0.5 ppm and above. Kulle notes that
for subjects exposed to 1.0, 2.0, and 3.0 ppm HCHO, a significant
linear trend with dose was observed for both odor and eye
irritation and for all eye/nose/throat irritation.
8.5. Discussion
Both HUD and OSHA have assessed the acute effects due to
HCHO. HUD's assessment was used to support changes in the
Manufactured Home Construction and Safety Standards, while OSHA's
assessment will be used to support a possible change in OSHA's
permissible exposure level for HCHO.
HUD'» assessment consisted of an evaluation of the cost-
benefit relationships of regulatory alternatives to control -iCHO
levels in mobile homes. A computer model was developed using
data from mobile home residents in Wisconsin, Minnesota, and
Washington to assess the relationship between HCHO levels and
8-17
-------
mobile home age. The cost of ilLnes3 was calculated for a 20-
year exposure period under various assumptions, such as initial
HCHO concentrations in homes, type and cost of resulting health
effect, and number of homes and persons affected. HUD used the
CPSC injury-cost model to estimate an average cost of illness
assuming that exposure to HCHO could cause one of several types
of health problems (for example, dermatitis of the face or
conjunctivitis). The HUD document did not report the incidence
or prevalence of symptoms for persons residing in homes with
varying detectable levels of HCHO. HUD's study method assumed
that 75 percent of the occupants of the mobile home with HCHO
would experience a health problem, but the concentration
producing this effect was not derived or estimated.
There were no data presented in the HUD analysis which
support a dose-response relationship between sensory effects and
HCHO levels in mobile homes. Data presented do support a
qualitative relationship. It is also important to note, however,
that HUD's review does not address the question of concentration
levels of HCHO in the mobile home and the magnitude of the
possible effect on the resident.
OSHA has produced an assessment of both noncancer irritant
and canc«r effects. For the noncancerous effects assessment,
OSHA relics on data submitted by industry (SOCMA, 197,9) and
certain assumptions. SOCMA collected information on nose and eye
irritation from 17 industries where HCHO exposure occurs and
calculated average exposure levels. OSHA only used the endpoint
"nose irritation" in their assessment, which SOCMA defined as the
8-18
-------
ability to detect HCHO odor. One must assume, however, that odor
recognition coincides with eye, nose, and throat irritation. it
may in some individuals, but not in others. For instance, an
individual may have a high odor threshold (1.5 ppm), but a low
eye irritation threshold. Odor perception (strength) is very
subjective.* Consequently, odor recognition and strength should
only be used as qualitative markers of HCHO level and any
corresponding eye, nose, and throat irritation.
In summary HUD's and OSHA's approaches provide some
qualitative measure of acute effects from HCHO exposure. These
techniques, however, can not identify a true dose-response
relationship. The individual reviewed studies can only be used
in the same manner; for qualitative estimates of population-based
risks.
All but two of the reviewed studies estimate the prevalence
of irritant effects for a given exposure level. Table 3-3
presents response data over a range of exposures for three acute
endpoints. As can be seen from the table, reductions in the
prevalence of these endpoints from small changes in HCHO levels,
say from 0.4 to 0.1 ppm, are difficult to quantify.
Only the studies by Bender et al. (1983), Andersen and
Molhave (1984), Hanrahan et al. (1984), and Kulle (1985)
presented response data over a range of doses so as to allow
estimation of irritation prevalence for a particular exposure
*01factory receptors can become saturated when breathing HCHO for
a period of time and, when this occurs, people become refractory
to the odor perception. Also, when saturation occurs, it would
be extremely difficult to link odor perception to the
manifestation of symptoms such as irritation.
8-19
-------
Table 8-3. RXPOSURR RANGES FOR SELECTED ENDPOINTS
Acute Effect
Exposure Level
PrevaJ ence
Author
CD
to
o
Nose Irritation
Fye Irritation
Couqh and
Wheez inn
0.04-0.09
<0. 01-2. 84
0.10-3.00
0.19-0.44
0.04-0.09
0.10-0.09
0.40-0.80
0.02-0.05
<0. 10-2.84
0.10-1.00
0.40-0. 80
23%
45%
79%
73%
15%
57%
25%
13%
44%
54%
31*
Olsen and Dossinq
Anderson et al.
Garry et al.
Olsen and Dossinq
Olsen and Dossinq
Anderson et al.
Olsen and Dossinq
Gamble et a 1 .
Anderson et al .
Garry et al.
Schoenherq and
Mitchel1
-------
level. The data of Andersen and Molhave (1984) and Kulle (1985)
have been analyzed by EPA using logistic regression analyses for
comparability. Hanrahan et al. (1984) presented in their paper
results of logistic regression analyses of their data. The
Hanrahan et al. (1984) results were controlled for age, gender,
and smoking. Figures 8-1 - 8-4 show the percent response
predicted at selected exposures for eye irritation (Hanrahan et
al., 1984) and eye, nose, and throat irritation (Andersen and
Molhave, 1984 and Kulle, 1985). The trends for all three curves
are statistically significant. The predicted response curves for
Hanrahan et al., (1984) who studied randomly sampled mobile home
residents, and for Andersen and Molhave (1984) and Bender et al.
(1983), who clinically studied volunteers, are very similiar.
The response curve for Kulle (1985), on the other hand, predicts
a lower percentage response than the three above studies for a
given exposure level. Likewise, for exposure levels above 0.5
ppm, the upper 95% confidence intervals for predicted response
from Kulle1s data are lower than the 95% confidence bounds of the
Andersen and Molhave and the Hanrahan et al. predicted response
curves.
8-21
-------
oo
I
ro
K)
Figure 8-1. Predicted irritative response oven a
range of HCHO levels
(Data from Hanrahan. et al. 1984).
Response (%)
100
80
60
40
20
*
i
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Exposure Level (ppra)
o Predicted response *95X loner CI *95% upper CI
-------
oo
I
t\J
Figure 8-2. Predicted irritative response oven a
range of HCHO levels
(Data from Andersen and Molhave 1984).
Response (%)
100
80
60
40
20
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Exposure Level (ppm)
o Predicted response *95% lower CI *95% upper CI
-------
00
I
K>
Figure 8-3
Response (X)
100 r
80
60
20
o
t
Predicted irritative response oven a
range of HCHO levels
(Data from Kulle 1985).
0.5 1 1.5 2
Exposure Level (ppm)
oPredicted response *95X lower CI
2.5
*95X upper CI
-------
00
tSJ
in
Figure 8-4. Eye irritation response oven a
range of HCHO levels
(Data from Bender, et al.) .
Response (%)
100 r
80
60
40
20
0.1 0.2
0.3 0.4 0.5 0.6 0.7
Exposure Level (ppm)
o Response
0.8 0.9
-------
Car* put be taken in inferring from the results in Figures
8-1 to 8-4 to the general population. First, three of the
studies are based on study subjects who are volunteers and
selection bias may be present. The one study whose population
was randomly selected is of a cross-sectional design, making
causation difficult to infer. Second, all studies had small
numbers of subjects, 16 in Andersen and Molhave, 28 in Bender et
al., 61 in Hanrahan et al., and 22 in Kulle. Larger studies of
randomly selected subjects are needed to estimate general
population risks.
In conclusion, none of the reviewed studies provide adequate
data to quantify general population risks for the acute effects
of HCHO. At best, the studies provide a qualitative estimate of
population responses over a wide range of exposure and
quantitative estimates of responses for very select populations.
Consequently, for small differences in concentration, say 0.4 to
0.8 ppm, it is not possible to determine the change in response
rates of a given exposed population.
8-26
-------
9.
REFERENCES
Acheson, E.D., Gardner, M.J., Pannett, B., Barnes, H. R. ,
Osmond, C. and Tyalor, C.P. 1984a . The Lancet 1: 6L1.
Acheson, E.D., Barnes, H. R. , Gardner, M.J., Osmond, C. ,
P-annett, B. and Taylor, C.P. 1984b. The Lancet 1: 1066.
Alarie, Y. 1985. Personal communication to Richard
Hefter. N'o studies are known of the respiratory response
of hamsters to sensory irritants. However, since hamsters
have a trigeminal nerve syste:n, a similar response to
sensory irritants as seen in rats and mice would be
e xpected .
Albert, R.E., Sellakumar, A.R., Laskin, 5., Kuschner, M.,
Nelson, N. , and Snyder, D. A. 1982. Gaseous . formaldehyde
and hydrogen chloride induction of nasal cancer in the
rat. J. Natl. Cancer Inst. 63: 597-603.
A?A. 1984. American Plywood Association comments
pertaining to the advance notice of proposed rulemaking on
formaldehyde under section 4(f) of the Toxic Substances
Control Act. Docket OPTS 62033.
Andersen, I., Lundquist, G.R., and Molhave, L. 1975.
Indoor air pollution due to chipboard used as a
construction material. Atmosphere Environment 9:1121-1127.
Anderson, H.A. , Dally, K.A. , and Eckmann, A.D. et al .
1983. The epidemiology of mobile home formaldehyde vapor
concentration and resident's health status. Wisconsin
Division of Health, Wisconsin State Laboratory of Hygiene,
and University of Wisconsin Department of Preventive
Medicine.
Anderson, H.A. , Dally, K.A., Hanrahan, L.P., Eckmann, A.D.,
Kanarek, M.S., and Rankin, J. 1983. The epidemiology of
mobile home formaldehyde vapor concentration and residents'
health status. Wisconsin Division of Health. Report to
U.S. Environmental Protection Agency, EPA-905/1-83-001.
Andersen, I., and Molhave, L.
studies with formaldehyde, pp
(ed.) Formaldehyde Toxic ity.
Publishing Corporation, 1983.
1984. Controlled human
. 154-165. In James E. Gibson
Washington: Hemisphere
Ashby, J. and Ratpan, F. 1986. Evidence to Associate the
Teratogenicity of Glycerol Formal with Ethylene-Glycol
Monomethylether (EGME) Rather than with Formaldehyde.
Environ. Muta. 8: 6.
9-1
-------
Auerbach, C., Moutschen-Dahmen, M., and Moutschen, J.
1977. Genetic and Cytogenetical Effects of Formaldehyde
and Related Compounds. Mutation Res. 39: 317-362.
Ayres, P.H., Marshall, T.C., Sun, J.D., Bond, J.A., and
Hobbs, C.H. 1985. Interaction of Formaldehyde with
Glutathione in the Isolated/Ventilated Perfused Lung and
the Isolated Perfused Liver. J. Toxicol. Environ. Health
15: 655-662.
Bardana, E.J. 1984. Effect of cigarette srnoke on
formaldehyde data. J. Occ. Med. 26(6): 410 (letter).
Bartnik, F.G., Gloxhuber, CHR. , and Zi-nmermann, V. 1985.
Percutaneous Absorption of Formaldehyde in Rats.
Toxicology Letters 25: L67-172.
3attelle Columbus. 1985. Formaldehyde Dose-Response
Estimations for Non-Cancer Outcomes, Draft Report. Task
93, Contract No. 68-01-6721, U.S. Environmental Protection
Agency, OPTS, OTS, EED, DDB.
Battelle Columbus Laboratories. Final Report on a Chronic
Inhalation Toxicology Study in Rats a-vj Mice Exposed to
Formaldehyde to Chemical Industry Ins-.itute of Toxicology,
submitted September 18, 1981; revised December 31, 1981.
Bauchinger, M. and Schmid, E. 1985. Cytogenetic Effects
in Lymphocytes of Formaldehyde Workers of a Paper
Factory. Mutation Res. 158: 195-199.
Beall, J.R. and Ulsamer, A.G. 1984. Formaldehyde and
Hepatotoxicity: A Review. J. Toxicol. Environ. Health
14: 1-21.
Bender, J.R., Mullin, L.S., Graepel, G.J., and Wilson, W.E.
1983. Eye Irritation Response of Humans to Formaldehyde.
Am. Ind. Hyg. Assoc. J. 44: 463-465.
Berg, P. 1951. Synthesis of labile methyl groups by
guinea pig tissue in vitro. J. Biol. Chem. 190: 31-38.
Bertazzi, P.A., Zochetti, C., Pesatori, A., Radice, L. and
Vai, T. 1984. Mortality of workers exposed to formaldehyde
in resin manufacturing. Presented at the XXI Congress on
Occupational Health, Dublin, Ireland, September 1984.
Billings, R.E., Ku, R.H., Brower, M.E., Dallas, C.E., and
Theiss, J.C. 1984. Disposition of Formaldehyde (CH20) in
Mice. Toxicologist 4: 8.
9-2
-------
Blair, A., Stewart, P., O'Berg, M., Gaffey, W., Walrath,
J., Ward, J., Bales, R., Kaplan, S. and Cubit, D. 1986,
JNCI, 76: 1071-1084.
31air, A., Stewart, P.A., Hoover, R.N., craumeni, J.F. Jr.,
Walrath, J., O'Berg, M., and Gaffey, W. In Press. Cancers
of the ^"aso/Harynx and Oropharynx and Formaldehyde
Exposure. J. Natl. Cancer Inst.
Qogdanffy, M.S., Morgan, P.M., Morgan, K.T., and Starr,
T.3. 1985. Binding Kinetics of Fornaldehdye to Rat and
Hunan N'asal Mucus, and Bovine Se run Albumin. Asnet-ACS DM 3
Meeting, Boston, MA.
3oja, J.W., Nielsen, J.A., Foldvary, E., and Truitt, E.3.
Jr. 1985. Acute Low-Level Formaldehyde Behavioral and
Neurochemical Toxicity in the Rat. Prog. N'euro-Psycho-
Pharnacol. and Biol. Psychiat. 9: 671-674.
Boorrnan, G.A. 1984. Letter to Dr. James Swenberg, Chemical
industry Institute of Toxicology, Research Triangle Park,
North Carolina.
Boreiko, C.J., Couch, D.B., and Swenberg, J.A. 1982.
Mutagenic and Carcinogenic Effects of Formaldehyde.
Environ. Sci. Res. 25: 353-367.
Bosley, C.E. and Pruet, C.S. 1984. Inverted Sinonasal
Papillomas. Ear, Nose and Throat Journal 63: 509-513.
Brendel, R. 1964. Untersuchungen an ratten zur
vertraeglichkeit von hexamethylentetramin
examethylenetetramine tolerance in rats].
Arzneimittelforsch 14: 51-53.
Breysse, P.A. 1984. Formaldehyde levels and accompanying
symptoms associated with individuals residing in over 1000
conventional and mobile homes in the State of Washington.
Stockholm: August 1984, Proc. 3rd Int'l. Conf. Indoor Air
Qual. and Climate.
Brinton, L.A., Blot, W.J., Becker, J.A., Winn, D, M.,
Browder, J.P., Farmer, J.C. and Fraumeni, J.F. 1984a. A
case-control study of cancers of the nasal cavity and
paranasal sinuses. American Journal of Epidemiology 119:
896-906.
Brinton, L.A., Blot, W.J. and Fraumeni, J.F. 1984(b).
Nasal cancer in the textile and apparel industries.
Submitted for publication.
9-3
-------
Brooks, S.M., Weiss, M.A., and Bernstein, I.L. 1985.
Reactive Airways Dysuction Syndrome. J. Occup. Med. 27-
473-476.
Brown, 3. 1964. The Papillomatous Tumours of the Mose.
J. Laryngol. Otol. 73: 339-905.
Brown, K.G. L934. Risk Assessment of Laboratory Rats
Exposed to Formaldehyde Vapors. National Institute of
Environmental Health Sciences, Research Triangle Park,
North Carolina.
3uckl
Barrow
Sensory
Appl
ey, L.A., Jiang, Y.Z., Janes, R.A., Morgan, K.T. and
iw, C.S. 1984. Respiratory Tract Lesions Induced by
iry Irritants at the RDrg Concentration. Toxicol.
Pharmacol. 74: 417-429.
Burge, P.S., Harries, M.G., Lam, W.K., O'Brien, I.M. and
Patchett, P.A. 1935. Occupational Asthma Due to
Formaldehyde. Thorax 40: 225-260.
Caceres, T., Soto, H., Lissi, E. and Cisternas, R. 1983.
Indoor House Pollution: Appliance Emissions and Indoor
Ambient Concentrations. Atmospheric Environment 17:
1009-1013.
Cantoni, 0. and Cattabeni, F. 1985. Inhibition of DNA
Repair by Metal Compounds and Formaldehyde.
Casanova-Schmitz, M., David, R.M., and Heck, H. D'A.
1984. Oxidation of formaldehyde and acetaldehyde by NAD+-
dependent dehydrogenases in rat nasal mucosal
homogenates. Biochem. Pharmacol. 33: 1137-1142.
Casanova-Schmitz, M. and Heck, H. d'A. 1983. Effects of
Formaldehyde Exposure on the Extractability of DNA from
Proteins in the Rat Nasal Mucosa. Toxicol. Appl.
Pharmacol. 70: 121-132.
Caaanova-Schmitz, M. and Heck, H.d'A. 1984. Differential
Labelling of Rat Nasal Mucosal DNA Fractions by C^C]- and
C3H]-Pormaldehyde (CH20). Toxicologist 4: 29.
Casanova-Schmitz, M., Starr, T.B. and Heck, H.D'A. 1984.
Differentiation Between Metabolic Incorporation and
Covalent Binding in the Labeling of Macromolecules in the
Rat Nasal Mucosa and Bone Marrow by Inhaled Ci^C]- and [ H]
Formaldehyde. toxicol. Appl. Pharmocol. 76: 26-44.
Casanova-Schmitz, M. and Heck, H. d'A. 1984. Effects of
Glutathione (GSH) Depletion on Formaldehyde-Induced DNA-
Protein Cross-Links in the Rat Respiratory Mucosa.
Toxicologist 4: 29.
9-4
-------
Commoner, B. 1976. "Reliability of Bacterial Mutagenesis
Techniques to Distinguish Carcinogenic and Noncarcinogenic
Chemicals." United States Environmental Protection Agency
Report No. 600/1-76/002. PB 259934. Springfield, VA.
l.'ational Technical Information Service.
Couiroe, Jr., J.H. et al. 1974. Defense Mechanisms of the
Lung's (Chap. 17). in: Physiology of Respiration, po.
220-229.
Conners, J.C. 1984 (March 9). Manufactured Housing
Institute. Results of a single-wide demonstration home.
Report to Shirley Wiseman, General Deputy Assistant
Secretary of Housing, HUD, Washington, D.C.
Connor, T.H., Ward, Jr., J.B., and Legator, M.S. 1985.
Absence of Mutagenicity in the Urine of Autopsy Service
Workers Exposed to Formaldehdye: Factors Influencing
Mutagenicity Testing of Urine. Int. Arch. Occup. Environ.
Health 56: 225-237.
Consensus Workshop on Formaldehyde. 1984. Conclusions of
the Epidemiology Panel. Little Rock, Arkansas.
Consensus Workshop on Formaldehyde. 1984. Final Report:
Deliberations of the Consensus Workshop on Formaldehyde,
October 3-6, 1983, Little rock, Arkansas.
Conyers, E.P. 1984. Letter and enclosures sent to
G. Schweer (USEPA/OTS). Frankfort, KY: Kentucky
Department for Health Services, Radiation and Product
Safety Branch.
Cornet, J.P. 1983. Results and conclusions of the search
for formaldehyde in houses and retirement homes in which no
chipboard is used as building material. Haarlem District,
The Netherlands: Product Analysis Agency, [In Dutch;
English trans.]
Coyne, L.B., Cook, R.E., Mann, J.R., Bonyoucos, S.,
McDonald, O.F. and Baldwin, C.L. 1985. Am. Ind. Hyg.
Asao. J., 46: 609-619.
Craft, T.R. and Skopek, T.R. 1986. Formaldehyde
Mutagenesis in Human Lymphoblasts in vitro; Effect of Dose
Rate on Cumulative Induced Mutant Fraction. Environ. Muta.
8: 19.
Dahl, A.R. and Hadley, W.M. 1983. Formaldehyde Production
Promoted by Rat Nasal Gytochrome P-450-Dependent
Monooxygenases with Nasal Decongestants, Essences,
Solvents, Air Pollutants, Nicotine, and Cocaine as
Substrates. Toxicol. Appl. Pharmacol. 67: 200-205.
9-6
-------
Casanova-Schmitz, M. and Heck, H. d'A. 1985. DMA-Protein
Cross-Linking Induced by Formaldehyde (FA) in the Rat
Respiratory Mucosa: Dependence on FA Concentration in
Vormal Rats and in Rats Depleted of Glutathione (GSH).
Toxi.colegists 5: 128.
Chang, J.C.F., Gross, E.A., Swenberg, J.A. and
Barrow, C.S. 1983. \'asal Cavity deposition
histopathology, and cell proliferation after single or
repeatad formaldehyde exposures in 36C3F1 mice and F-344
rats. Toxicol. Appl. Pharmacol. 68: 161-176.
Chang, J.C.F., Steinhagen, W.H., and Barrow, C.S. 1981.
Effect of single or repeated formaldehyde exposure on
.ninute volume of B6C3F1 mice and F-344 rats. Toxicol.
Appl. Pharmacol. 6: 451-459.
CUT. 1981. Final report on a chronic inhalation
toxicology study in rats and mice exposed to
formaldehyde. Chemical Industry Institute of Toxicology,
December 31, 1981.
CIR Expert Panel. 1984. Final Report on the Safety
Assessment of Formaldehyde. Journal ~>f the American
College of Toxicology 3: 157-184.
Clement Associates. 1982. Formaldehyde Risk Assessment
for Occupationally Exposed Workers. Clements Associates,
Arlington, VA.
Conn, M.S. 1981. Revised carcinogenic risk assesment for
urea formaldehyde foam insulation: estimates of cancer
risk due to inhalation of formaldehyde releasd'ed by UFFI.
Washington, D.C.: U.S. Consumer Product Safety Commission.
Cohn, M.S. 1984. Formaldehyde carcinogenic risk
assessment: exposure in manufactured housing. In:
Comments of the staff of the U.S. Consumer Product Safety
Commission to the Department of Housing and Urban
Development's Proposed Revisions of the Manufactured Home
Construction and Safety Standards (24 CFR Part 3280), April
1984.
Cohn, M.S., DiCarlo, F.J., Turturro, A. and Ulsamer, A.G.
1985. Letter to the Editor: Toxicol. and Appl. Pharmacol.
77: 363-364.
Cohn, M.S. 1985. Personal communication on Sensitivity of
MLE Estimates.
9-5
-------
Dalbey, W.E. 1982. Formaldehyde and Tumors in Hamster
Respiratory Tract. Toxicology 24: 9-14.
Dallas, C.E., Theiss, J.C., Harrist, R.B , and Fairchild,
E.J. 1985. Effects of Subchronic Formaldehyde Inhalation
on Minute Volume and Nasal Deposition in Sprague-Qawley
Rats. J. Toxicol. Environ. Health 16: 553-564.
Day, J.H., Lees, R.E.M., Clark, R.H., and Pattee, P.L.
1984. Respiratory Responses to Formaldehyde and Off-Gas of
Urea Formaldehyde Foan Insulation. Can. Med. Assoc. J.
131: 1061-1065.
Delia Porta, G. and Cabral, J.G. 1970. Studio della
tossicita transolancentare di cancerogenesi in ratti
trattati con esametilentetramina. [Transplacental toxicity
and carcinogenesis studies in rats with hexamethylene-
tetramine]. Tumori 56: 325-334.
Delia Porta, G. and Colnaghi, M.I.G. 1968.
Noncarcinogenicity of hexamethylenetetramine in mice and
rats. Food Cosmet. Toxicol. 6: 707-715.
Delzell, E. and Grufferman. 1983. Cancer and other causes
of death among female textile workers, 1976-78. Journal of
the National Cancer Institute. 71(4): 735-739.
Den Engelse, L. , Gekbbink, M. and Emmelot, P. 1975.
Studies on lung tumors. III. Oxidative metabolism of
demethylnitrosamine by rodent and human lung tissue. Chem.
Biol. Interact. 11: 535-544.
Department of Housing and Urban Development. 1981. An
evaluation of formaldehyde problems in residential mobile
homes. Final Report. Office of Policy Development and
Research, Washington, DC.
Dooley, J.F., Blackburn, G.R., Schreiner, C.A., and
Mackerer, C.R. 1985. Inhibition of the Mutagenicity of
Formaldehyde (HCHO) in the L5178Y TK +/- Mouse Lymphoma
Assay by Formaldehyde Dehydrogenase (FDH): Application to
Characterization of Mutagenic components. EMS Abstracts: 9.
Drake-Lee, A.B., Lowe, D. , Swanston, A. and Grace, A.
1984. Clinical Profile and Recurrence of Nasal Polyps.
Journal of Laryngology and Otology 98: 783-793.
DuVigneaud, V.T., Varby, W.G, and Wilson, J.E. 1950.
Incorporation of the carbon of formaldehyde and formate
into the methyl groups of choline. J. Am. Chem. Soc. 72:
2819-2820.
9-7
-------
Edling, C., Odkvist, L. , and Hellquist, H. 1985.
Formaldehyde and the Nasal Mucosa. Br. J. Ind. Med. 42:
570-571.
Egle, J.L. 1972. Retention of inhaled formaldehyde,
propionaldehyde, and acrolein in the dog. Arch. Environ.
Health 25: 119-124.
Einbrodt, J.J., Prajsnar, D. and Erpenbeck, J. 1976. Der
formaldehyde und ameisensaeurespiegel im blut und urin beim
menschen nach formaldehyd-exposition. [Effect of
formaldehyde exposure on people in school and living
areas.] Zentralbl. Arbeitsmed. Arbeitsschutz Prophyl.
26(8): 154-158.
Environmental Protection Agency. 1981. Priority Review
Level 1: Formaldehyde.
Environmental Protection Agency. 1984a. Formaldehyde;
Determination of Significant Risk; Advance Notice of
Proposed Rulemaking and Notice. 49 FR 21870-21898.
Environmental Protection Agency. 1984b. Determination of
Significant Risk. 49 FR 21880-21884.
Environmental Protection Agency. 1986. Proposed
Guidelines for Carcinogen Risk Assessment. 49 FR 46294.
Everett, L.H. 1983. Urea formaldehyde foam and
formaldehyde emission, UK experience with cavity wall
insulation. A contribution to the formaldehyde workshop,
Little Rock, Arkansas. England: Building Research
Station.
Fayerweather, W.E., Pell, S. and Bender, J.R. 1982.
Case-control study of cancer deaths in DuPont workers with
potential exposure to formaldehyde. In: Formaldehyde:
toxicity, epidemiology, and mechanisms. JJ Clary, JE
Gibson, RS Waritz, eds. New York; Marcel Dekker, Inc.
Feron, V.J. 1979. Effects of Exposure to Acetaldehyde in
Syrian Hamsters Simultaneously Treated with Benzo(a)pyrene
or Diethylnitrosamine. Prog. Exp. Tumor Res. 24: 162
(Karger, Basel).
Feron, V.J. 1984. Summary Tumor Results-Acetaldehy3e.
Proided to Dr. H. Milman, U.S. EPA, Washington, DC.
Feron, V.J., Kruysse, A., Til, H.P. and Immel, H.R.
1978. Repeated Exposure to Acrolein Vapour: Subacute
Studies in Hamsters, Rats, and Rabbits. Toxicology 9:
47-57.
9-8
-------
Feron, V.J., Kruysse, A. and Wouterson, R.A. 1982.
Respiraory Tract Tumors in Hamsters Exposed to Acetaldehyde
Vapour Alone or SimiItaneously to Benzo(a)pyrene or
diethylnitrosamine. Eur. J. Cancer Clin. Oncol. 18: 13.
Fleij, I., Petri, M. , Stocker, W.G. and Thiess, A.M.
1982. Cytogenetic Analyses of 31ood Lymphocytes of Workers
Exposed to FormaIdahydein Formaldehyde Manufacturing and
Processing. J. Occup. Med. 24: 1009-1012.
Foeke.-is, J.\., Rennie, P.S., Cheng, H. and Bruchovsky, \'.
1985. In Situ Cross-Linking of Androgen Receptors to
1,'uclear Acceptor Sites of Rat Prostate with Formaldehyde.
J. Biol. chem. 260: 10093-10098.
Formaldehyde Institute. 1984. Comments pertaining to the
Advance Notice of Proposed Rulemaking on Formaldehyde under
Section 4(f) of the Toxic Substances Control Act. Docket
OPTS 62033.
Fortmann, R.C., Borrazzo, J.E. and Davidson, C.I. 1984.
Characterization of Parameters Influencing Indoor Pollutant
Concentrations. Proceedings of the Third International
Conference on Indoor Air Quality and Climate, Stokholm,
Sweden.
Frazer, J.P. 1984. Allergic Rhinitis and Nasal Polyps.
Ear, Nose, and Throat Journal 63: 172-176.
Friedmann, I. and Osborn, D. A. 1982. Pathology of
Granulomas and Neoplasms of the Nose and Paranasal Sinuses,
Churchill Livingstone, New York.
Gamble, J.F., McMichael, A.J., Williams, T. and
Battigelli, M. 1976. Respiratory function and symptoms:
An environmental-epidemiological study of rubber workers
exposed to a phenol-formaldehyde type resin. Am. Ind. Hyg.
Assoc. J., 37: 499-513.
Garry, V.F., Oatman, L, Pleus, R., and Gray, D. 1980.
Formaldehyde in the home: Some environmental dis-ease
perspectives. Minn. Med., 63: 107-111.
Gastwirth, J.L. 1983. Combined tests of significance in
EEO cases. Presented at the 1983 Annual Meeting of the ASA
in Toronto, Canada.
Gibson, J.E. 1984. Comments on "Formaldehyde and
Hepatotoxicity" by Beall and Ulsamer. J. Toxicol. Environ.
Health 14: 465-467.
9-9
-------
airman, J.R., Geisling, K.L., and Hodgson, A.T. L983.
Sources and concentrations of formaldehyde in indoor
environments. Presented at the 75th Air Pollution Control
Association Annual Meeting, New Orleans, L.A. June 20-25,
1332. Washington, DC: Energy and Environmental Division,
U.S. Department of Energy. Contract No. DE-AC03-76 5F00093.
Godish, T. 1933. Interpretation of one-time formaldehyde
sampling results from measurements of environmental
variables. Chicago, IL: Proc. APCA Specialty Conference -
Measurement and Monitoring of Non-Criteria (Toxic)
Contaminants in Air.
Goldmacher, V.S. and Th i I ly, W.G. 1933. Formaldehyde is
Mutagenic for Cultured Hunan Cells. Mutation Res. 116:
417-422.
Goodman, J.I. and Tephly T.R. 1971. A comparison of rat
and human liver formaldehyde dehydrogenease. Biochem.
Biophys. Acta 252: 439-505.
Gottschling, L.M., Beaulieu, H.J., and Melvin, W.W.
1984. Monitoring of Formic Acid in Urine of Humans Exposed
to Low Levels of Formaldehyde. Am. Ind. Hyg. Assoc. J.
45: 19-23.
Grafstrom, R.C., Curren, R.D., Yang, l.L. and Harris, C.C.
1985. Genotoxicityof Formaldehyde in Cultured Human
Bronchial Fibroblasts. Science 228: 39-90.
Grafstrom, R.C., Fornace, A., Jr. and Harris, C.C. 1984.
Repair of DNA Damage Caused by Formaldehyde in Human Cells.
Cancer Research 44: 4323-4327.
Grafstrom, R.C., Fornace, A.J., Jr., Autrup, H., Lechner,
J.F. and Harris, C.C. 1983. Formaldehyde Damage to DNA
and Inhibition of DNA Repair in Human Bronchial Cells.
Science 220: 216-218.
Gralla E.J., Heck, H d'A., Hrubesh, L.W. and
Meadows/ G.W. 1980. A report of the review of the
formaldehyde exposure made by the CUT ad hoc analytical
chemistry investigative team held at the Battelle Memorial
Columbus (Ohio) Laboratory. Releigh, NC: Chemical
Industry Institute of Toxicology. CUT Docket No. 62620.
Grindstaff, G.F. 1985. Revised Cancer Risks for
Formaldehyde. Memorandum to Richard Hefter.
Groan, W.J., Gramp, G.D., Garrison, S.B., and Walcott,
R.J. 1985. Factors that influence formaldehyde air levels
in mobile homes. Forest Products Journal 35: 11-18.
9-10
-------
Hanrahan, L.P./ Anderson, H.A., Dally, K.A., Eckmann, A.D.
and Kanarek, M.S. 1983. An Investigation of the
Offgassing Decay Function in Aging Mobile Homes with
Climate Corrected Formaldehyde Readings. Wisconsin
Division of Health, Wisconsin State Laboratory of Hygiene,
and j'niversity of Wisconsin, Department of Preventive
Medici ne.
Hanrahan, L.P. Dally, K.A., Anderson, H.A., Kanarek, M.S.
and Rankin, J. 1904. AJPA 74: 1026-1027.
Hardell, L., Johansson, 3. and Axelson, 0. 1982. American
Journal of Industrial Medicine 3: 247-2S7/
Harrington, J. and Oakes, D. 1984. Mortality Among
British Pathologists. Br J Ind Med 41: 138-191.
Harrington, J. and Shannon, H. 1975. Mortality study of
oathologists and medical laboratory technicians. British
Medical Journal 4: 329-332.
Hawthorne, A. R. , Gammage, and Dudney, C.S. et al. 1984.
Oak Ridge National Laboratory. An indoor air quality study
of forty east Tennessee homes. Draft report. Washington,
DC: U.S. Department of Energy *FTP/.-.-001701. ORNL-5965.
Hayes, R.B., Raatgever, J.W. and deBruyn, A. 1984. Tumors
of the nose and nasal sinuses: A case-control study.
Presented at the XXI Congress on Occupational Health,
Dublin, Ireland, September 1984.
Hayes, R.B., Raatgever, J.W. and Gerin, M. 1986. Int. J.
Cancer, 37: 487-492.
Heck, H. d'A. 1982. CUT Activities 2: 3-7.
Heck, H. d'A and Casanova-Schmitz. 1983. Reaction of
Formaldehyde in the Rat Nasal Mucosa. In: Formaldehyde:
Toxicology-Epidemiology-Mechanisms. Clay, J.J., Gibson,
J.E., and Waritz, R.S., Eds. Marcel Dekker, Inc., New
York.
Heck, H. d'A, Casanova-Schmitz, M., Dodd, P.B., schachter,
E.N., Witek, T.J. and Tosun, T. 1985. Formaldehyde (CH20)
Concentrations in the Blood of Humans and Fischer-344 Rats
Exposed to CH20 Under Controlled Conditions. Am. Ind. Hyg.
Assoc. J. 46: 1-3.
Heck, H. d'A, Chin, T.Y., and Schmitz, M.C. 1980.
Distribution of 14-C-formaldehyde in rats after inhalation
exposure. In Formaldehyde Toxicity, ed. J. Gibson, pp.
26-37. Washington: Hemisphere.
9-11
-------
Heck, H. d'A, White, E.I. and Casanova-Schmitze. 1982.
Biomed. Mass Spectrom. 9: 347-353.
'iemminki, X., Falck, K. and Vainio, H. 1980. Comparison
of Alkylation Rates and Mutagenicity of Directly Acting
Industrial and Laboratory Chemicals. Arch. Toxicol. 46:
277. '
Hendrick, D.J., Rando, R.J., Lane, D.J., and Morris, M.J.
1982. Formaldehyde Asthma: Challenge Exposure Levels and
Fate After Five '/ears. J. Occup. Med. 24: 393-397.
Hern'oerg, S. , Westerhol.m, ?., Schultz-Larsen, K. ,
Dogerth, R., Kuosma, E., Englund, A., Hansen, H.S. and
Mutanen, ?. 1983. \rasal and sinunasal cancer. Scand. J.
Work. Environ. Health 9: 315-326.
Hodges, H.E. 1984. Letter with enclosures to USEPA,
Office of Toxic Substances. Mashville, TN: Tennessee
Department of Health and Environment, Air Pollution Control
Divi sion.
Horton, A.W., Tye, R. and Stemmer, K.L. 1963.
Experimental carcinogenesis of the lu-ig. Inhalation of
gaseous formaldehyde or an aerosol of coal tar by C3H
mice. J. Natl. Cancer Inst. 30: 31-43.
Howe, R.B., Crockett, P.W. and Crump, K.S. 1984. Weibull-
A and Probit-A Fortrain Programs Implementing Weibull and
Generalized Probit Models with Additive Background Response
Rates for the Analysis of Animal Bioassay Data. Battelle,
Washington, DC.
HPMA. 1984. Hardwood Plywood Manufacturers'
Association. Comments before the USEPA, May 23, 1984,
Formaldehyde: Determination of significant risk, NAPR, and
notice. Docket No. OPTS 62033. Reston, VA: Hardwood
Plywood Manufacturers' Association.
Huennekens, F.M and Osborn, M.J. 1959. Folic acid
coenzymes and 1-carbon metabolism. Adv. Enzymol. 21:
369-446.
Hyams, V.J. 1971. Papillomas of the Nasal Cavity and
Paranasal Sinuses A Clinico-Pathologi Study of 315 Cases.
Ann. Otol. 80: 192-206.
IARC Monograph. 1982. Some Industrial Chemicals and Dye
Stuffs. IARC 29: 345-389.
IRMC. 1984. Draft Report of the Risk Assessment Subgroup
of the IRMC-Formaldehyde Work Group.
9-12
-------
IRMC. 1984a. Report of the Subcommittee on Formaldehyde
Sensitization.
IRMC. I934b. Report of the Subgroup on Systemic Effects--
For^aliehyda.
Iversen, O.H. 1934. :Jrethan (Ethyl Carbamate) Alone is
Carcinogenic for Mouse Skin. Carcinogenesis 5: 911-915.
Jacobs, R.L., Freda, A.J. and Culver, W.G. 1933. Primary
N'asal Polyoosis Annals of Allergy 51: 500-505.
Kane, L.E., Barrow, C.S. and Alarie, Y. 1979. A Short-
Term Test to Predict Acceptable Levels of Exposure to
Airborne Sensory Irritants. Am. Ind. Hyg. Assoc. J. 40:
207-229.
Kane, L.E. and Alarie, Y. 1977. Sensory Irritation to
Formaldehyde and Acrolein During Single and Repeated
Exposures in Mice. Am. Ind. Hyg. Assoc. J. 38: 509-521.
Kendrick, J., Nettesheim, P., Guerin, M., Caton, J.,
Dalbey, W., Griesemer, R., Rubin, I., and Maddox, W.
1976. Tobacco Smoke Inhalation Studies in Rats. Toxicol.
Appl. Pharmacol. 37: 557-569.
Kerfoot, E.J., and Mooney, T.F. 1975. Formaldehyde and
paraformaldehyde study in funeral homes. Am. Ind. Hyg.
Assoc. J., 36 533-537.
Kerns, W.D., Pavkov, K.L., Donofrio, D.J., Gralla, E.J. and
Swenberg, J.A. 1983. Carcinogenicity of Formaldehyde in
Rats and Mice After Long-Term Inhalation Exposure. Cancer
Research 43: 4382-4392.
Kilburn, K.H., Warshaw, R., Boylen, C.T., Johnson, S.-J.S.,
Seidman, B., Sinclair, R., and Takaro, T. 1985a.
Pulmonary and Neurobehavioral Effects of Formaldehyde
Exposure. Arch. Environ. Health 40: 254-160.
Kilburn, K.H., Seidman, B.C., and Warshaw, R. 1985b.
Neurobehavioral and Respiratory Symptoms of Formaldehyde
and Xylene Exposure in Histology Technicians. Arch.
Environ. Health 40: 229-233.
Kimmel, C.A., Cook, R.O., and Staples, R.E. 1976.
Teratogenic Potential of Noise in Mice and Rats. Toxicol.
Appl. Pharmacol. 36: 239-245.
9-13
-------
Kitchens, J.F, Casner, R.E., Edwards, G.S., Harvard, W.E,
III and Macri, B.J. 1976. Investigation of selected
potential environmental contaminants: formaldehyde.
Washington, DC: U.S. Environmental Protection Agency.
EPA-560/2-76-009.
Klenitzky, J.S. 1940. On experimental cancer of the
uterine cervix. Bull. Biol. Med. Exp. 9: 3-6.
Konopinski, V.J. 1983. Residential formaldehyde and
carbon dioxide. Indiana State Board of Health,
Indianapolis, IN. 329-334.
Krivanek, N.D., Chroney, N.C. and McAlack, J.W. 1983.
Skin Initiation Promotion Study with Formaldehyde in CD-I
Mice. In: Formaldehyde: Toxicology-Epidemiology-
Mechanisrns. J.J. Clary, J.E. Gibson, and R.S. Waritz,
Eds., Marcel Dekker, Mew York.
Ku, R.H. and Billings, R.E. 1984. Relationship Between
Formaldehyde Metabolism and Toxicity and Glutathione
Concentrations in Isolated Rat Hepatocytes. Chem. Biol.
Interactions 51: 25-36.
Kucharczyk, N., Yang, J.T., Wong, K.K., and Sofia, R.D.
1984. The Formaldehyde-Donating Activity of N5, N10-
Methylene Tetrahydrofolic Acid in Xenobiotic
Biotransformation. Xenobiotica 14: 667-676.
Kuhn, M. and Wanner, H: U. 1984. Indoor air pollution by
building materials. August 1984. Stockholm, Sweden:
Proc. Intl. Con. Indoor Air Qual. and Climate.
Kulle, T.J. Letter to C.S. Scott, U.S. EPA. August 28,
1985.
Lacroix, J., Chad, Z., Gauthier, M., LaPointe, N., Haley,
N., Lapierre, J. G., and Masson, P. 1985. Urea-
Formaldehyde Foam Insulation-Clinical Experience in 76
Children. Union Medicale Du Canada. 114: 542-547.
•*
Lara, C.W., Casanova-Schmitz, M., and Heck, H. d'A. 1985.
Acetaldehyde (AA) but not Formaldehyde (FA) Depletes Nasal
Mucosal Glutathione: An Investigation of the Role of
Organic Peroxides. Toxicologist 5: 128.
Lampertico, P., Russell, W.O. and Macomb, W.S. 1963.
Squamous Papilloma of Upper Respiratory Epithelium. Arch.
Path. 75: 293-302.
Lasser, A., Rothfeld, P.R., and Shapiro, R.S. 1976.
Epithelial Papilloma and Squamous Cell Carcinoma of the
Nasal Cavity and Paranasal Sinuses. Cancer 38: 2503-2510.
9-14
-------
Lee, H.K., Alarie, Y., and Karol, M.H. 1984. Induction of
Formaldehyde Sensitivity in Guinea Pigs. Toxicol. AppL.
Pharmacol. 75: 147-155.
Lee, K.P. and Trochinowicz, H.J. 1982. Induction of :«'asal
Turaqrs in Rats Exposed to Hexa;nethylphosphoranide by
Exhalation. J. MatL. Cancer last. 68: 157-164.
Lee, K.P. and Trochinowicz, H.J. 1984. Morphogenesis of
Masai Tumors in Rats Exposed to Hexamethylphosphoranide by
Inhalation. Environmental Research 33: 106-113.
Lees, R.E.M., Clark, R.H., and Day, J.H. 1985.
Respiratory Responses to Formaldehyde, Formaldehyde Free
UFFI Off-Gas and Particles in UFFI Related Asthma. J.
Allergy Clin. Immun. 75: 169.
Levine, R.J., And je Ikovich, D.A. and Shaw, L.K. 1984. The
mortality of Ontario undertakers and a review of the
formaldehyde-related mortality studies. Journal of
Occupational Medicine 26(10): 740-746.
Levine, R.J., DalCorso, R.D., Blunden, P.B., and
Battigelli, M.C. 1984. The Effects of Occupational
Exposure on the Respiratory Health of West Virginia
Morticians. J. Occup. Med. 26: 91-98.
Liebling, F., Rosenman, K., Pastides, H., Griffith, R. and
Lemeshow, S. 1984. American Journal of Industrial
Medicine 5: 423-429.
Ma, T.-H., Harris, M.M., and Lin, G. 1985. Genotoxicity
Studies of Formaldehyde Using Tradescantia-Micronucleus
Test. EMS Abstracts: 42.
Main, D. and Hogan, T.J. 1983. Health Effects of Low-
Level Exposure to Formaldehyde.. J. Occup. Med. 25: 896-
900.
Malorny, G., Rietbrock, N. and Schneider, M. 1965. Die
oxydation des Formaldehyds zu ameisensaeure im Blut, ein
Beitrag zum Stoffwechsel des Formaldehyds. [Oxidation of
formaldehyde to formic acid in blood, a contribution to the
metabolism of formaldehyde.] Naunyn Schmiedebergs Arch.
Exp. Pharmakol. 250: 419-436.
Margosches, E.H. and Springer, J. One Way Animal Data Can
Help with Epidemiology Planning. Presentation at Eastern
North American Region, Biometric Society Meetings, March
1983.
9-15
-------
Marks, T.A., Worthy, W.C., and Staples, R.E. 1980.
Influence of Formaldehyde and Sonacide© (Potentiated Acid
Glutaraldehyde) on Embryo and Fetal Development in Mice.
Teratology 22: 52-58.
Marsh, G. 1933. Proportional mortality among chemical
workers exposed to formaldehyde. British Journal of
Industrial Medicine 39: 313-322.
Marsh, G. 1983. Mortality among workers from a plastic
producing plant: A natched case-control study nestle.1 in a
retrospective cohort study. Journal of Occupational
Medicine 25(3): 219-2030.
Mashford, P.M. and Jones, A. R. 1982. Xenobiotics 12:
119-124.
Matanoski, G. 1980. Data presented at the 49th meeting of
the Interagency Collaborative Group or Environmental
Carcinogenesis, February 6, 1980, Bethesda, Maryland.
Matthews, T.G., Howell, T.C. and Gammage, R.B. 1981(a).
Formaldehyde Release from Plywood, Particleboard,
Fiberboard, and Paneling. Oak Ridge National Laboratory,
Monthly Report I, CPSC-IAG-81-1360.
Matthews, T.G., Allen, R.J. and Gammage, R.B. 1981(b)
Formaldehyde Release from Plywood, Particleboard,
Fiberboard, and Paneling. Oak Ridge National Laboratory,
Monthly Report II, CPSC-IAG-81-1360.
Matthews, T.G. and Westley, R.R. 1983. Determination of
Formaldehyde Emission Levels from Ceiling Tiles and
Fiberglas Insulation Products. Oak Ridge National
Laboratory, Report No. CPSC-IAG-82-1181.
Matthews, T.G., Daffron, C. R. , Hawthorne, A.R., Reed, T.J.,
and Tromberg, B.J. 1984. Formaldehyde emissions from
consumer and construction products: potential impact on
indoor formaldehyde concentrations. August 1984.
Stockholm: Proc. 3rd Intl. Conf. Indoor air Quality and
Climate.
McMartin, K.E., Martin-Amat, G., Makar, A.B. and
Tephly, T.R. 1977. Methanol poisoning. V. Role of
formate metabolism in the monkey. J. Pharmacol. Exp.
Therap. 201: 564-572.
McMartin, K.E, Martin-Amat, G., Noker, P.E. and
Tephly, T. R. 1979. Lack of role for formaldehyde in
methanol poisoning in the monkey. Biochem. Pharmacol. 28:
645-649.
9-16
-------
Meittinen, O.S and Wang, J. 198L. An alternative to the
proportionate mortality ratio. Americal Journal of
Epidemiology 116: 144-148.
Meullsr, R., Raabe, G. and Schumann, D. 1978. Leukoplakia
indaced by repeated deposition of formalin in rabbit oral
rnucosa: Long-term experiments with a new "oral tank."
Exp. Pathol. 16: 36-42.
Meyer, 3., and Hermanns, K. 1984a. Formaldehyde indoor
air problems. Seattle Washington: University of
Washington. ?roc. Air Poll. Control Assoc. Annual Meeting,
1984, San Francisco, CA. APCA paper 34-35.2.
Meysr, 3., and Hermanns, K. 1984b. Diurnal variations of
formaldehyde exposure in mobile homes. Berkeley, CA:
Lawrence Berkeley National Laboratory, Report #13573.
MHI. 1984. Manufactured Home Inst. Comments pertaining
to the advance notice of proposed rulemaking on
formaldehyde under section 4(f) of the Toxic Substances
Control Act.
Mierauskiene, J.R. and Lekevicius, R.K. 1985. Cytogenetic
Studies of Workers Occupationally Exposed to Phenol,
Styrene and Formaldehyde. Mutation Fes. 147: 308-309.
Mizenina, O.A., Kiseleva, N.P., Kaftanova, A.S., and
Dobrov, E.N. 1984. Formaldehyde-Induced Crosslinking of
RNA with Protein in Small Ribonucleoprotein Particles
Reconstituted from Tobacco Mosaic Virus Protein and
Fragments of its RNA.
Morgan, K.T., Patterson, D.L. and Gross, E.A. 1983.
Formaldehyde and the Nasal Mucociliary Apparatus. In:
Formaldehyde: Toxicology-Epidemiology-Mechanisms. Clary,
J.J., Gibson, J.E., and Waritz, R.S., Eds. Marcel Dekker,
Inc., New York.
Morgan, K.T., Patterson, D.L. and Gross, E.A. 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, K.T., Patterson, D.L. and Gross, E.A. 1984. Froc
Palate Mucociliary Apparatus: Structure, Function, and
Response to Formaldehyde Gas. Fundam. Appl. Toxicol. 4:
58-68.
9-17
-------
Morgan, K.T., Jiang, X.-Z., Starr, T.B., and Kerns, W.D.
1985. More Precise Localization of Nasal Tumors Associated
with Chronic Exposure to Formaldehyde Gas. Toxicol. Appl.
Phamacol., Submitted.
Mor-jan, K.T., Patterson, D.L., and Gross, E.A. L986.
Responses of the Nasal Mucociliary Apparatus of F-344 Rats
ti Formaldehyde Gas. Toxicol. Appl. Pharmacol. 32_: L-13.
Moschanireas, D.J., Stark, J.W.C., McFadden, J.-E., and
Morse, 5.S. 1978. Geonet, Inc. Indoor air pollution in
the residential environment, Vol. I, Data Analysis anl
Interpretation. Washington, DC: U.S. Environmental
Protection Agency. E?A-600/7-73-229a.
Myers, G.E. and Nagaoka, M. 1981. Formaldehyde
Emission: Methods of Measurement an
-------
NPA. 1984. National Particleboard Association. 4-F
comment: Comments of the National Particleboard Assoc. on
advance notice of proposed rulemaking on formaldehyde under
section 4(f) of the Toxic Substances Control Act.
Neely, W.B. 1964. The metabolic fate of formaldehyde-14C
iritraperitoneally administered to the rat. Biochem.
Pharmacol. 13: 1137-1142.
Nikolaidis, E.T., Trost, D.C., Buchholz, C.L. and
Wilkinson, E.J. 1985. The Relationship of Histologic and
Clinical Factors in Laryngeal Papillomatosis. Arch.
Pathol. Lab. Med. L09: 24-29.
Nordman, H., Keskinen, H., and Tuppurainen, M. 1985.
Formaldehyde Asthma-Rare or Overlooked? J. Allergy Clin.
Immunol. 75: 91-99.
Obe, G. and Ristow, H. 1977. Acetaldehyde Not Ethanol
Induces Sister Chromatid Exchanges in Chinese Hamster Cells
In Vitro. Mutat. Res. 56: 211.
Occupational Safety and Health Administration. 1984.
Preliminary assessment of the health effects of
Formaldehyde. November 5, 1984.
Olson, J.H., and Dossing, M. 1982. Formaldehyde Induced
Symptoms in Day Care Centers. Am. Ind. Hyg. Assoc. J. 43:
366-370.
Olsen, J.H. and Jensen, O.M. 1984. Case-control study on
sinonasal cancer and formaldehyde exposure based on a
national data linkage system for occupation and cancer.
Presented at the Society for Epidemiologic Research, 17th
Annual Meeting, Houston, Texas, June 13-15, 1984. Am J
Epidemiology 120: 459.
Osborn, D.A. 1970. Nature and Behavior of Transitional
Tumors in the Upper Respiratory Tract. Cancer 25: 50-60.
Overman, D.O. 1985. Absence of Embryonic Effects of
Formaldehyde After Percutaneous Exposure in Hamsters.
Toxicology Letters 24: 107-110.
Palese, M. and Tephly, T.R. 1975. Metabolism of formate
in therat. J. Toxicol. Environ. Health. 1: 13-24.
Paludetti, G., Maurizi, M., Tassoni, A., Tosti, M. and
Altissim, G. 1983. Nasal Polyps: A comparative Study of
Morphologic and Etiopathogenetic Aspects. Rhinology 21:
347-360.
9-19
-------
Partanen, T., Kauppinen, T., Nurminen, M., Nickels, J.,
Hernberg, S. , Hakuliner, T., Pukkala, E. and Savoner, E.
1985. Scand J. Work Environ. Health 11: 409-415.
Patterson, R., Pateras, V., Grammer, L.C., and Harris,
K.E. 1986. Human Antibodies Against Formaldehyde Human
Serum Albumin Conjugates or Human Serum Albumin in
Individuals Exposed to Formaldehyde. Int. Archs. Allergy
Appl. Immun. 79: 53-59.
Perzin, J.H., Lefkowitch, J.H. and Hui, R.M. 1981.
Bilateral Nasal Squamous Carcinoma Arising in
Papillomatosis. Cancer 48: 2375-2382.
Pickrell, J.A., Griffis, L.C. and Hobbs, C.H. 1982..
Release of formaldehyde from various consumer products.
Albuquerque, NM: Lovelace Biomed. andEnviron. Res.
Institute.
Pickrell, J.A., Griffis, L.C., Hobbs, C.H., Kanapilly, G.M.
and Mokler, B.V. 1984. Formaldehyde release from selected
consumer products: influence of chamber loading, multiple
products, relative humidity, and temperature.
Environmental Science Technol. 18: 682-686.
Podall, H. 1984. A review of the state-of-the-art on
urea-formaldehyde resins for wood anJ causes of
formaldehyde release. Draft report. U.S. Environmental
Protection Agency, Office of Toxic Substances, Economics
and Technology Division, Washington, D.C.
Proctor, D.F. 1982. The Mucociliary System (Chap. 10).
In: The Nose: Upper Airway Physiology and the Atmospheric
Environment. Proctor/Anderson (eds.). Elsevier Biomedical
Press, pp. 245-278.
Pruett, J.J., Scheuenstuhl, H. and Michaeli, D.L. 1980.
The incorporation and localization of aldehydes (highly
reactive cigarette smoke components) into cellular
fractions of cultured human lung cells. Arch. Environ.
Health. 35: 15-20.
Rader, J. 1974. Irritative Effects of Formaldehyde in
Laboratory Halls, Analytical and Experimental
Investigations. Dissertation for M.D. Degree. Institute
of Pharmacology and Toxicology, University of Wurzberg,
Federal Republic of Germany.
Ragan, D.L. and Boreiko, C.J. 1981. Initiation of
C3H/10T1/2 Cell Transformation by Formaldehyde. Cancer
Letters 13: 325-331.
9-20
-------
Rapoport, I.A. 1948. Mutation Under the Effect of
Unsaturated Aldehydes. Dokl. Akad. Nauk. S.S.S.R. 61: 713
(in Russian).
Report of the Federal Panel on Formaldehyde. 1982
Environ. Health Perspect: 43: 139-168.
Reznik, G., Reznik-Schuller, H. , Ward, J.M. and Stinson,
S.F. 1980. Morphology of Nasal-Cavity Tumours in Rats
After Chronic Inhalation of 1,2-Dibromo-3-Chloropropane.
Br. J. Cancer 42: 772-781.
Ridolfi, R.L./ Lieberman, P.H., Erlandson, R.A. and Moore,
O.S. 1977. Schneiderian Papillomas: A Clinicopathologic
Study of 30 Cases.. Am. J. Surg. Pathol. 1: 43-53.
Rietbrock, N. 1969. Kinetik and Wege des
Methanolumsatzes. [kinetics and pathways of methanol
metabolism.] Naunyn Schmiedebergs Arch. Pharmakol. Exp.
Pathol. 263: 88-105.
Ristow, H. and Obe, G. 1978. Acetaldehyde Induces Cross-
links in DNA nad Causes Sister-chromatid Exchanges in Human
Cells. Mutat. Res. 58: 115.
Robbins, J.D., Norred, W.P., Bathija, A., and Ulsamer,
A.G. 1984. Bioavailability in Rabbits of Formaldehyde
from Durable-Press Textiles. J. Toxicol. Environ. Health
14: 453-463.
Robbins, S.L. 1974. Pathologic Basis of Disease. W.B.
Saunders Co., Philadelphia, pp. 845-849, 885-892.
Rothman, K.J. and Boice, J.D. 1982. Epidemiologic
analysis with a programmable calculator, 2nd Edition.
Boston: Epidemiology Resources, Inc.
Roush, G., Walrath, J., Stayner, L., Kaplan, S. and
Blair, A. 1985. American Journal of Epidemiology.
Rubin, I.B., Gill, B.E., Guerin, M.R., Kendrick, J., and
Nettesheim, P. 1978. Correlation of Respiratory
Parameters in Hamsters with the Lung Deposition of
Radiolabelled Cigarette Smoke. Environ. Res. 16: 70-76.
Rusch, G.M., Clary, J.J., Rinehart, W.E. and Bolte, H.F.
1983. A 26-Week Inhalation Toxicity Study with
Formaldehyde in the Monkey, Rat, and Hamster. Toxicol.
Appl. Pharmacol. 68: 329-343.
9-21
-------
SAB. 1985. EPA Science Advisory Board, Environmental
Health Committee: Review of Draft Document-Preliminary
Assessment of Health Risks to Garment Workers and Certain
Home Residents from Exposures to Formaldehyde, Draft May
1985. Letter to Lee M. Thomas, October 1, 1985, from
Griesemer, R.A. and Nelson, N.
SAJ. 1984. Science Applications, Inc. Formaldehyde: A
survey of airborne conentrations and sources. Final
report. Sacramento, CA: State of California Air Resources
Board. Contract No. A2-059-32.
SAIC. 1986. Science Applications International
Corporation. Formaldehyde Exposure and Cancers of the
Nose, Sinus, and Pharynx. Final Report, Contract No. 68-
01-6280, U.S. Environmental Protection Agency, OPTS, OTS,
EED, DDB.
Schoenberg, J.B. and Mitchell, C.A. 1975. Airway Disease
Caused by Phenolic (Phenol-Formaldehyde) Resin Exposure.
Arch. Environ. Health 30: 574-577.
Schottenfeld, D. and Fraumeni, J.F., Jr. 1982. Cancer
Epidemiology Prevention, pp 46-47, W.B. Saunders Co.
Schouten, J.P. 1985. Hybridization Selection of Covalent
Nucleic Acid-Protein Complexes. J. Biol. Chem. 260:
9929-9935.
Schutte, W., Frank, C.W., Hoffman, J., Sailer, S., and
Scarpellino, C. 1981. Final report to the Formaldehyde
Institute. Iowa City, Iowa. University of Iowa.
Schweer, G. 1987. Estimates of Populations Exposed to
Formaldehyde in New Residential Construction. Memo dated
March 2, 1987 from G. Schweer (EPA/OTS/EED) to G. Semeniuk
(E)A/OTS/CCD).
Scott, M.J., Ward, J.B., Dallas, C.E., and Theiss, °J.C.
1985. Chromosome Damage Observed in Lung but not Bone
Marrow or Sprague-Dawley Rats Exposed to Formaldehyde by
Inhalation. EMS Abstracts: 53-54.
Sellakumar, A.R., Snyder, C.A., Solomon, J.J., and Albert,
R.E. 1985. Carcinogenicity of Formaldehyde and Hydrogen
Chloride in Rats. Toxicol. Appl. Pharmacol. 81: 401-406.
Sellars, S.L. 1982. The Inverted Nasal Papilloma. J.
Laryngol. Otol. 96: 1109-1112.
9-22
-------
Seydell, E.M. 1933. Fibro-Epithelial Tumors of the Nose
(Papillomata) and Their Relationship to Carcinoma. Ann.
Otol. 42; 1081-1103.
Sexton, K., Petreas, M.X., Lui, J.S. and Kulasingam, G.L.
1985. California Department of Health Services.
Formaldehyde concentrations measured in California mobile
homes. Detroit, MI: Proc. 78th Annual Meeting Air Poll.
Control Assoc.
Sexton, K., Liu, K.S. and Petreas, M.X. 1985. Measuring
indoor air quality by mail. Berkeley, CA: California
Dept. of Health Sevices. In press.
Sheppard, D., Eschenbacher, W.L., and Epstein, J. 1984.
Lack of Bronchomotor Response to up to 3 ppm Formaldehyde
in Subjects with Asthma. Environ. Res. 35: 133-139.
Siegel, D.M., Frankos, V.H., and Schneiderman, M.A. 1983.
Formaldehyde Risk Assessment for Occupationally Exposed
Workers. Reg. Toxicol. Pharmacol. 3: 355-371.
Sielken, R.L. 1983. Incorporating Time into the
Estimation of the Potential Human Cancer Risk from
Formaldehyde Inhalation. Institute of Statistics, Texas
A&M University.
Singh, J., Walcott, R., St. Pierre, C., Coffman, M.A.,
Ferrel, T.W., Opthoff, D., and Montgomery, D. 1982.
Clayton Environmental Consultants Inc. Evaluation of
formaldehyde problem in mobile homes - testing and
evaluation. Washington, DC: Office of Policy Development
and Research, U.S. Department of Housing and Urban
Development. Contract No. HC-5222, Volume 3.
Snyder, R.N. and Perzin, K.H. 1972. Papillomatosis of the
Nasal Cavity and Paranasal Sinuses (Inverted Papilloma
Squamous Papilloma) A Clinicopathologic Study. Cancer 30:
668-690.
Snyder, R.D. and Van Houten, D. 1986. Genotoxicity of
Formaldehyde and an Evaluation of its Effects on the DNA
Repair Process in Human Diploid Fibroblasts. Mutation Res.
165: 21-30.
SOCMA: Booz, Allen, and Hamilton. 1979. Preliminary
study of the costs of increased regulation offormaldehyde
exposure in the U.S. workplace. Prepared for Synthetic
Organic Chemical Manufacturers Association. Scarsdale, New
York.
9-23
-------
Sokal, R.R. and Rohlf, F.J. Biometry, W.H. Freeman and
Company, San Francisco, 1969.
Solomon, M.J., and Varshavsky, A. 1985. Formaldehyde-
mediated DNA-Protein Crosslinking: A Probe for In Vivo
Chromatin Structures. Prod. Natl. Acad. Sci. USA 82:
6470-6474.
Solomons, K. and Cochrane, J.W.C. 1984. Formaldehyde
Toxicity: Part I. Occupational Exposure and a Report of 5
Cases. S. FR. Med. J. 66: 101-102.
Spangler, F. an dWard, J.M. 1983. Skin Initiation-
Promotion Study with Formaldehyde in Sencar Mice. In:
Formaldehyde: Toxicology-Epidemiology-Mechanisms. J.j.
Clary, J.E. Gibson and R.S. Waritz, Eds. Marcel Dekker, New
York.
Spear, R. 1982. Formaldehyde Study Shows Gene Effect in
Anatomy Students. The New Physcian 6: 17.
Stankowski, L.F. Jr., Tuman, W.G., Godek, E.G., and Kasper,
G.J. 1986. Induction of Mammalian Cell Mutations By
Formaldehyde. Environ. Muta. 8: 81.
Starr, T.B. and Buck, R.D. 1984. The Importance of
Delivered Dose in Estimating Low-Dose Cancer Risk From
Inhalation Exposure to Formaldehyde. Fundam. Appl.
Toxicol. 4: 740-753.
Starr, T.B., Gibson, J.E., Barrow, C.S., Boreiko, C.J.,
Heck, H. d'A., Levine, R.J., Morgan, K.T., and Swenberg,
J.A. 1984. Estimating Human Cancer Risk From
Formaldehyde: Criticla Issues. Chemical Industry Institute
of Toxicology, Research Triangle Park, North Carolina.
Stayner, L., Smith, A.B., Reeve, G., Blade, L. , Elliott,
L., Keenlyside, R. and Halperin, W. 1984. Proportionate
mortality study of workers exposed to formaldehyde in the
garment industry. Presented at the Society for
Epidemiologic Research 17th Annual Meeting, Houston, Texas,
June 13-15, 1984. Am J Epidemiology 120: 458-9.
Stewart, H.L., Dunn, T.B., Snell, K.C. and Deringer, M.K.
Tumors of the Respirator Tract. In: V.S. Turosou (ed.),
Pathology of Laboratory Animals, Vol. 2, pp. 251-288.
Lyon, France: IARC, 1979.
Stock, T.H. and Mendez, S.R. 1985. University of Texas,
School of Public Health. A survey of typical exposures to
formaldehyde in Houston area residences. Am. Ind. Hyg.
Assoc. J. 46: 313-317.
9-24
-------
Stone, R., DePaso, D., Shephard, E., To, T., Tucker, B. and
Villarreal, E. 1981. Technology and Economics Inc. An
evaluation of formaldehyde problems in residential mobile
homes. Draft final report. Washington, DC: Dept. of
Housing ad Urban Development, Office of Policy Development
and Research. Contract No. HC-5105.
Stroup, N., Blair, A., and Erickson, G. 1984. Brain
cancer and other causes of death in anatomists. Presented
at the Society for Epidemiologic Research 17th Annual
Meeting, Houston, Texas, June 13-15, 1984. Am J
Epidemiology 120: 500
Strittmatter, P. and Ball, E.G. 1975. Formaldehyde
dehydrogenase, a glutathione-dependent enzyme system.
J. Biol. Chem. 213: 445-461.
Swenberg, J.A., Barrow, C.S., Boreiko, C.J., Hick, H.
I'A., Levine, F.J., Morgan, K.T., and Starr, T.B. 1983.
Nonlinear biological responses to formaldehyde and their
implications for carcinogenic risk assessment.
Carcinogenesis 4: 945-952.
Swenberg, J.A., Gross, E.A., Martin, J., and Popp, J.A.
1983. Mechanisms of formaldehyde toxicity. In: J.E.
Gibson (ed.), Formaldehyde Toxicity. Hemisphere Publishing
Corp., Washington.
Swenberg, J.A. and Boreiko, C.J. 1985. Appropriateness of
Polypoid Adenoma for Quantitative Risk Assessment.
Chemical Industry Institute of Toxicology.
Swenberg, J.A., Heck, H. d'A., Dudek, B.R., and Halliwell,
W.H. 1984. Inhalation Toxicity of HCL Gas to Sprague-
Dawley and F-344 Rats. Toxicologist 4: 27.
Tabershaw Associates. 1982. Historical prospective
mortality study of past and present employees of the
Celanese chemical and plastics plant located in Bishop,
Texas. Tabershaw Associates, Rockville, Maryland.
Takano, T., Shirai, T.S., Ogiso, T., Tsuoa, H. , Baba, S.
and Ito, N. 1982. Sequential Changes in Tumor Development
Induced by 1,4-Dinitrosopiperazine in the Nasal Cavity of
F344 Rats. Cancer Research 42: 4236-4240.
Thomson, E.J., Shackleton, S. and Harrington, J.M. 1984.
Chromosome Aberrations and Sister-Chromatid Exchange
Frequencies in Pathology Staff Occupationally Exposed to
Formaldehyde. Mutation Res. 141: 89-93.
9-25
-------
Timm, W. and Smith, P. 1979. Formaldehyde odor and healtn
problems within residences. Journal of Thermal Insulation,
Vol. 3, p. 104.
Tola, S., Hernberg, S. , Collan, Y., Linderborg, H. and
Korkala, M.L. 1980. A case-control study of the etiology
of nasal cancer in Finland. Int. Arch. Occup. Environ.
Health 46: 79-85.
Tobe, M., Kaneko, T., Uchida, Y., Kamata, E., Ogawa, Y.,
Ikeda, Y. and Saito, M. 1985. Studies of the Inhalation
Toxicity of Formaldehyde. National Sanitary and Medical
Laboratory Service (Japan), pp. 1294.
Traynor, G.W., Anthon, D.W. and Hollowell, C.D. 1982.
Technique for Determining Pollutants from a Gas-Fired
Range. Atmos. Environ. 16: 2979-2988.
Traynor, G.W., Allen, J.R., Apte, M.G., Girman, J.R. and
Hollowell, C.D. 19B3. Pollutant Emissions from Portable
Kerosene-Fired Space Heaters. Environ. Sci. and Technol.
17: 369.
Traynor, G.W. and Nitschke, A. 1984. Field survey of
indoor air pollution in residences with suspected
combustion-related sources. August 1984. Stockholm:
Proc. 3rd Int'l. Conf. Indoor Air Quality and Climate.
Tuthill, R.W. 1984. Woodstoves, Focmaldehyde, and
Respiratory Disease. Am. J. Epid. 120: 952-955.
Tuttle, W.W. and Schottelius, B.A. 1969. Gas Exchange-
Respiration. In: Textbook of Physiology. The C.U. Mosby
Company, St. Louis.
UFFI/ICC. 1981. Urea Formaldehyde Foam Insulation
Information and Coordination Centre. The report on the
national testing survey to the board of review by the
UFFI/ICC. Canada.
Ulsamer, A.G. , Beall, J.R., Rang, H.K., Frazier, J.A.
Overview of Health Effects of Formaldehyde. 1984. In
Saxsena, J. (ed.), Hazard Assessment of Chemicals--Current
Developments. New York, Academic Press, Inc. 3: 337-400.
University of Texas, School of Public Health. 1983. Final
report. Texas indoor air quality study. Houston, TX:
University of Texas.
9-26
-------
Uotila, L. and Koivusalo, M. 1974. Formaldehyde
dehydrogenase from human liver; purification, properties,
and evidence for the formation of glutathione thiol esters
by the enzyme. J. Biol. Chem. 249: 7653-7663.
Usdin, V.R. and Arnold, G.B. 1979. Transfer of
formaldehyde to guinea pig skin. Gillette Research
Institute Report. Contract No. 78-391.
Van der Wai, J.F. 1982. Formaldehyde measurements in
Dutch houses, schools, and offices in the years 1977-
1980. Atmosphere Environment 16: 2471-2478.
Vaughan, T.L., Strader, C., Davis, S. and Daling, J.R.
1986a. Submitted for publication.
Vaughan, T.L., Strader,, C., Davis,,- S. and Daling, J.R.
1986b. Submitted for publication.
Versar. 1982. Final Draft Report—Exposure Assessment for
Formaldehyde. Versar, Inc... Springfield-, .VA.
Versar. 1986a. Maximum Levels of Formaldehyde Exposure in
Residential Settings. EPA Contract No;: 681-02-3968, Task
No. 14. Versar, Inc., Springfield, VA. .
Versar. 1986b. Formaldehyde exposure model-description
and demonstration. Final report. Washington DC: US
Environmental Protection Agency, Office of Toxic
Substances, EPA Contract NQ. 68-02-3968.
Versar. 1986c. Formaldehyde Exposure in Residential
Settings: Sources, Levels, and Effectiveness of Control
Options. EPA Contract No. 68-02-3968, Task No. 14. Versar
Inc., Springfield, VA.
Vital Statistics of the U.S. 1974. . 19,70 Volume II-
Mortality, Part A. United States Department of Health,
Education and Welfare. Health Resources Administration
75-1104.
Vrabec, D.P. 1975. The Inverted Schneiderian Papilloma: A
Clinical and Pathological Study. Laryngoscope 85: 186-220.
Wagner, B.S. 1982. Lawrence Berkeley Laboratory.
Residential indoor air quality/air infiltration study.
Washington, D.C.: U.S. Department of Energy. Contract No.
DE-AC03-76SF00098.
Walrath, J. and Fraumeni, J. 1983. Mortality patterns
among embalmers, International Journal of Cancer 31:
407-411.
9-27
-------
Walrath, J. and Fraumeni/ J.F. 1984. Cancer Research 44:
4638-4641.
Ward, E. 1984. Memorandum: Formaldehyde Interim Exposure
Report. U.S. EPA, Washington, DC.
Ward, Jr., J.B., Hokanson, J.^A;, Smith, E.R., Chang, L.W.,
Pereira, M.A., Whorton, Jr., E.B. and Legator, M.S.
1984. Sperm Count, Morphology and Fluorescent Body
Frequency in Autopsy Service Workers Exposed to
Formaldehyde. Mutation res. 130: 417-424..
Watanabe,*F., Matsunaga, T., Soejimai T. and Iwata, Y.
1954. Study oh aldehyde carcinogenic!ty; Part I.
Experimentally induced rat sarcomas by repeated injection
of formalin. Gann 45? .451-452.
Watanabe, F. and Sugimoto, S. 1955. Studies on aldehyde
carcihogenicity. Part II. Seven cases of transplantable
sarcomas of rats developed in the area of repeated
subcutaneous injections of urotropin (hexamethylene-
tetramine). Gann 46: 365-367.
Waydhas, C., Weigl, K. and Sies, H. 1978. The disposition
of formaldehyde and formate arising from drug N-
demethylations dependent on cytochrome P-450 in hepatocytes
and in perfused rat liver. Eur. J. Biochem. 89(1): 143-150
Weber-Tschopp, A., Fischer, L., and Grandjean, E. 1977.
Irritating Effects of Formaldehyde (HCHO) on Humans. Int.
Arch. Occup. Environ. Health 39: 207-218.
Widdicombe, J.G. 1977. Defense Mechanisms of the
Respiratory System (Chap. 9). In: Respiratory Physiology
II. University Park Press, Baltimore, pp 291-315.
Wisconsin Division of Health. 1984. Magnetic tape of data
collected by Anderson et al. (1983).
Witek, T.J., Schachter, E.N., Brody, D., Tosun, T., Beck,
G.J., and Leaderer, B.P. 1985. A Study of Lung Function
and Irritation From Exposure to Formaldehyde in Routinely
Exposed Laboratory Workers. Chest 88: 65.
Wong, 0. 1983. An epidemiologic mortality study of a
cohort of chemical workers potentially exposed to
formaldehyde. In: Formaldehyde: toxicity. J.J. Clary,
J.E. Gibson, R.S. Waritz, eds. New York: Hemisphere
Publishing Corp.
9-28
-------
Woodruff, R.C., Mason, J.M., Valencia, R., and
Zinunering, S. 1985. Chemical-Mutagenesis testing in
Droaophila. v. Results of 53 Coded Compounds Tested foi* the
National Toxicology Program. Environ. Mutagen.
7: 677-702.
Woutersen, R.A., Appelman, L.M., Wilmer, J.W.G.M., Spit,
B.J., and Falke, H.E. 1984. Subchronic (13-week)
Inhalation Toxicity Study with Formaldehyde in Rats
(Appendices),, pp. 1-88. Submitted to Dr. Harry A. Milman,
U.S. EPA,.Washington, DC.
Woutersen, R.A., Van Garderen-Hoetmer, A., and Appelman,
L.M. 1985. . : Li,fe-Span (27,,months) inhalation
Carcinogeni.cTity.. Study of Acetaldehyde in Rats. Final
Report. ' Report No. V85.145/190I72 Dated June, 1985 from
the Civo Institutes TNO>;Division for Nutrition and Food
Research TNO, the Netherlands.
Yamaguchi, K.T., Shaapshay, S.M., Incze, J.S., Vaughan,
C.W. and Strong, M.S. J. Qtolaryngol. 8: 171-178.
9-29
------- |