A
r
EPA
United States
Environmental Protection
Agency
Research and Development
Office of Health and
Environmental Assessment
Washington DC 20460
EPA/600/8-85/004F
September 1985
Final Report
Mutagenicity and
Carcinogenicity
Assessment of
1,3-Butadiene
-------
-------
EPA/600/8-85/004F
September 1985
Final Report
MUTAGENICITY AND CARCINOGENICITY ASSESSMENT
OF
1,3-BUTADIENE
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C.
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade names
or commercial products does not constitute endorsement or recommendation for
use.
-------
CONTENTS
Preface . v
Authors, Contributors, and Reviewers vi
1. SUMMARY AND CONCLUSIONS . .1-1
1.1. Summary 1-1
1.2. Conclusions 1-5
2. INTRODUCTION 2-1
3. TOXICOLOGY 3-1
3.1. Acute Toxicity . . 3-1
3.2. Subchronic Toxicity . . . . :. 3-1
3.3. Chronic Toxicity. . . . 3-3
3.4. Reproductive Toxicity 3-5
4. METABOLISM AND PHARMACOKINETICS 4-1
4.1. Metabolism. 4-1
4.1.1. In vitro Metabolism 4-1
4.1.2. In vivo Metabolism . . 4-3
4.2. Pharmacokinetics . 4-4
5. MUTAGENICITY OF 1,3-BUTADIENE AND ITS REACTIVE METABOLITES. ... 5-1
5.1. Mutagenicity of 1,3-Butadiene . 5-1
5.2. Metabolism of 1,3-Butadiene and Reaction of Metabolites
. with DNA . . ... 5-3
5.3. Mutagenicity of 3,4-Epoxybutene 5-6
5.4. Genotoxicity of l,2:3,4-Diepoxybutane .......... 5-7
5.4.1. Studies in Bacteria ..... 5-10
5.4.2. Studies in Fungi 5-10
5.4.3. Studies in Mammalian Cells 5-13
5.4.4. In vivo Studies 5-16
5.5. Mutagenicity of 4-Vinyl-l-cyclohexene and its
Metabolites 5-22
5.6. Summary of Mutagenicity Studies ....... 5-22
6. CARCINOGENICITY 6-1
6.1. Animal Studies 6-1
\
6.1.1. Chronic Toxicity and Carcinogenicity
Studies in Mice 6-1
6.1.2. Chronic Toxicity Studies in Rats. 6-6
iii
-------
CONTENTS (continued)
6.1.3. Carcinogenicity of Related Compounds 6-10
6.1.4. Discussion of Carcinogenicity Studies ....... 6-12
6.2. Epidemiologic Studies 6-15
6.2.1. McMichael et al. (1974, 1976) 6-16
6.2.2. Andjelkovich et al. (1976, 1977) 6-19
6.2.3. Checkoway and Williams (1982) 6-24
6.2.4. Meinhardt et al. (1982) 6-27
6.2.5. Matanoski et al. (1982) 6-29
6.2.6. Summary of Epidemiologic Studies 6-33
6.3. Quantitative Estimation 6-36
6.3.1. Procedures for Determination of Unit'Risk 6-37
6.3.1.1. Description of the Low-Dose Extrapo-
lation Model 6-38
6.3.1.2. Calculation of Human Equivalent
Dosages from Animal Data 6-40
6.3.1.2.1. Adjustments for Less
Than Lifetime Duration
of Experiment 6-43
6.3.1.3. Interpretation of Quantitative
Estimates 6-44
6.3.1.4. Alternative Models 6-45
6.3.1.5. Internal Dose vs. External Concentration. 6-46
6.3.2. Calculation of Quantitative Estimates 6-47
6.3.2.1. Mouse-to-Human Extrapolation 6-48
6.3.2.2. Rat-to-Human Extrapolation 6-51
6.3.3. Comparison of Human and Animal
Inhalation Studies 6-54
6.3.4. Relative Potency 6-62
6.3.5. Summary of Quantitative Estimation 6-70
REFERENCES 7-1
-------
PREFACE
The Mutagenicity and Carcinogenicity Assessment of 1,3-Butadiene was
prepared to serve as a source document for Agency-wide use. This document was
developed primarily for use by the U.S. Environmental Protection Agency's (EPA)
Office of Air Quality Planning and Standards (OAQPS) to support decision-making
regarding possible designation of 1,3-butadiene as a hazardous air pollutant.
Because OAQPS requested that this assessment focus only on the mutagenicity
and carcinogenicity of 1,3-butadiene, an evaluation of other health hazards
has not been included herein. This document, therefore, is not a comprehensive
health assessment document. The exposure information herein has not been
rigorously reviewed, and is used for illustrative purposes only. An analysis
of the ambient exposure and exposure to populations adjacent to emission
sources will be carried out separately by OAQPS.
In the development of this assessment document, the relevant scientific
literature through July 1, 1985, has been incorporated. Key studies have been
evaluated and the summary and conclusions have been prepared so that the
mutagenicity, carcinogenicity, and related characteristics of 1,3-butadiene are
qualitatively identified. Measures of dose-risk relationships relevant to
ambient exposures are also discussed so that the adverse health responses can
be placed in perspective with possible exposure levels.
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
The Carcinogen Assessment Group and the Reproductive Effects Assessment
Group within the Office of Health and Environmental Assessment were responsible
for preparing this document.
PRINCIPAL AUTHORS
Steven Bayard, Ph.D. Chapters 1 and 6
Robert P. Beliles, Ph.D. Chapters 1, 2, 3, 4, and 6
Arthur Chiu, Ph.D, M.D. Chapters 1, 2, and 6
Herman 0. Gibb, B.S., M.P.H. Chapters 1 and 6
Aparna Koppikar, M.B.B.S.* Chapters 1 and 6
Brian Sadler, Ph.D.t Chapter 4
Sheila L. Rosenthal, Ph.D. Chapter 5
Consultant
tResearch Triangle Institute, Research Triangle Park, NC
REVIEWERS
The following individuals provided external peer review of the mutagenicity
chapter of this document.
George R. Hoffman, Ph.D.
Department of Biology
Holy Cross College
Worcester, MA
Stanley Zimmering, Ph.D.
Division of Biology and Medicine
Brown University
Providence, RI
The following individuals provided peer review of this document and/or
earlier drafts of this document.
Harriet Ammann, Ph.D.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Research Triangle Park, NC
Jerry Blancato, Ph.D.
Exposure Assessment Group
Office of Health and Environmental Assessment
Washington, D.C.
-------
J.A. Bond, Ph.D.
Lovelace Inhalation Toxicology Research Institute
Albuquerque, NM
Chao W. Chen, Ph.D.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, DC
Ila Cote, Ph.D.
Pollutant Assessment Branch
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
Research Triangle Park, NC
Daphne Kamely, Ph.D.
Exposure Assessment Group
Office of Health and Environmental Assessment
Washington, DC
Robert E. McGaughy, Ph.D.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, DC
Debdas Mukerjee, Ph.D.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Cincinnati, OH
David Patrick
Pollutant Assessment Branch
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
Research Triangle Park, NC
Peter W. Preuss, Ph.D.
Deputy Director
Office of Health and Environmental Assessment
Washington, DC
Anita Schmidt
Risk Management Branch
Existing Chemical Assessment Division
Office of Toxic Substances
Washington, DC
Marvin A. Schneiderman, Ph.D.
6503 E. Halbert Road
Bethesda, MD
VI 1
-------
Hugh L. Spitzer, B.A.
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Washington, DC
T.B. Starr, Ph.D.
Chemical Industry Institute of Toxicology
Research Triangle Park, NC
Peter E. Voytek, Ph.D.
Reproductive Effects Assessment Group
Office of Health and Environmental Assessment
Washington, DC
Paul D. White, B.A.
Exposure Assessment Group
Office of Health and Environmental Assessment
Washington, DC
Jeannette Wiltse, Ph.D.
Risk Management Branch
Existing Chemical Assessment Division
Office of Toxic Substances
Washington, DC
EPA Science Advisory Board
The External Review Draft (February 1985) of this document was
dently peer-reviewed in public sessions of the Environmental Health
of EPA's Science Advisory Board.
indepen-
Committee
-------
1. SUMMARY AND CONCLUSIONS
1.1. SUMMARY
1,3-butadiene is a colorless gas with a slight aromatic odor at room
temperature and pressure. It is used mainly by the styrene-butadiene rubber
and polybutadiene rubber industries. No deaths and very few toxic effects have
been reported from acute exposure to the vapor. The symptoms resulting from
acute exposures are lethargy, drowsiness, and irritation to the mucous membranes
of the eyes and the mouth.
The available information on the mutagenicity of 1,3-butadiene is quite
limited in that only three studies have been reported. All three studies, how-
ever, indicate that 1,3-butadiene is a mutagen in Salmonella typhimuriunr. The
mutagenicity is observed only in the presence of a liver S9 metabolic activa-
tion system. No whole animal studies have been reported. These results sug-
gest that 1,3-butadiene is a promutagen in bacteria (i.e., its mutagenicity
depends on metabolic activation).
There is no information on the metabolism of 1,3-butadiene in humans. _In_
vitro data suggest that 1,3-butadiene is metabolized to 3,4-epoxybutene (epoxy-
butene) and then to 1,2:3,4-diepoxybutane (diepoxybutane). Evidence in rats
and mice suggests that 1,3-butadiene is metabolized to 3,4-epoxybutene j_n_
vivo, indicating that the metabolic pathway outlined on the basis of in vitro
data may occur in vivo.
3,4-Epoxybutene is a monofunctional alkylating agent, is a direct-acting
mutagen in bacteria (S_. typhimurium, Klebsiella pneumoniae, and Escherichia
coli), and induces sister chromatid exchange and chromosomal aberrations in
mice. Diepoxybutane is a bifunctional alkylating agent, and as such it can
form cross-links between two strands of DNA. It is mutagenic in bacteria (_K.
1-1
-------
pneumoniae and £. typhimurium), fungi (yeast and Neurospora crassa), and the
germ cells of Drosophila melanogaster. It also induces DNA damage in cultured
hamster cells and in mice, is clastogenic in fungi and cultured rat cells, and
produces chromosome damage/breakage in V_. melanogaster germ cells. Therefore,
the evidence indicates that 3,4-epoxybutene and diepoxybutane are mutagens/clas-
togens in microbes and animals.
Two lifetime inhalation carcinogenicity studies have been carried out in
mice and rats. There was a marked increase in incidences of primary tumors
among the exposed groups of mice in both sexes. These tumors included lymphomas,
hemangiosarcomas, alveolar/bronchiolar adenomas (and carcinomas), acinar cell
carcinomas, granulosa cell tumors or carcinomas, forestomach papillomas and
carcinomas, and hepatocellular adenomas and carcinomas. The study had to be
terminated at 60 to 61 weeks instead of the planned 104 weeks because of
excessive deaths from the neoplasia among the exposed mice.
In female rats (Sprague-Dawley) exposed to 1,3-butadiene, increased inci-
dences of mammary tumors, thyroid follicular cell adenomas, and uterine stromal
sarcomas were observed. In the male rats, increases in tumor incidences were
found in the exposed animals in the form of Leydig cell tumors and exocrine
pancreatic adenomas. Zymbal gland tumors were increased in both sexes of ex-
posed rats. The tumor sites involved were different in the mice and rats among
the exposed groups. The severity of the cancers was also widely different; in
the rats, no increase in mortality secondary to neoplasia was observed, and
there was no early termination of the experiments. In addition to the differ-
ences found in the two sexes, rats were affected less than mice.
The epidemiologic studies evaluated in this review were of workers engaged
in the production of synthetic rubber since synthetic rubber is produced from
styrene and 1,3-butadiene. Three of the studies specifically identified their
1-2
-------
study populations as being exposed to butadiene, as well as styrene. The
concurrent exposure of the workers in these studies to styrene presents a
potential problem since there is some limited evidence that styrene may be a
carcinogen, and in particular, a leukemogen.
Of the three studies on workers specifically identified as being exposed
to 1,3-butadiene, two were cohort studies, and one was a cross-sectional study
designed to look at certain hematologic parameters. One of the cohort studies,
a large study of almost 14,000 styrene-l,3-butadiene rubber (SBR) workers at
eight plants, found none of the standardized mortality ratios (SMRs) for cancer
to be significantly elevated. Some bias may have occurred, however, due to a
possible underascertainment of total deaths and a possible overestimation of
deaths among blacks. A second cohort study of workers, specifically identified
as being exposed to SBR, found that the.SMR for lymphatic and hematopoietic
cancer was of a borderline significance for a subcohort of workers employed at
a plant during the time when a batch production process was in operation.
Solvent exposure may have been a confounding factor, however. In this study,
exposure to 1,3-butadiene as well as styrene was actually measured, but the
measurements were not historic; they were taken at the time of the study. The
third study, in which workers were specifically identified by the authors as
being exposed to 1,3-butadiene, was a cross-sectional investigation designed
to look at certain hematologic parameters. It found no evidence of any hemato-
logic abnormality in the study population. Exposure to 1,3-butadiene, as well
as to styrene, toluene, and benzene was measured in this study.
Two studies found an association between employment in the synthetic rubber
industry and an elevated risk of cancer. One of these studies, a case-control
study of deaths among rubber plant workers from cancer of certain sites, diabe-
tes mellitus, and ischemic heart disease, found workers in the synthetic rubber
1-3
-------
area of the plant to have the highest risk ratio for deaths from lymphatic and
hematopoietic cancer (ICD 200-209). There may have been some confounding due
to organic solvent exposures however. The second study, a cohort mortality
study of rubber plant workers, found that an excess of lung cancer deaths
occurred among workers in the synthetic rubber area of the plant. This finding
was based on only three deaths, however, and there was no control for smoking.
Given the inconsistency of results from different studies, the possible
confounding due to exposure to solvents, styrene, and possibly other chemicals,
and the potential biases in some of the studies, the epidemiologic data would
have to be considered inadequate for evaluating whether a causal association
exists between 1,3-butadiene exposure and cancer in humans.
Based on the linearized multistage model, and estimates of internal dose
from external concentrations, a maximum likelihood estimate of incremental
unit risk was calculated for 1,3-butadiene, using the geometric mean from the
pooled male and pooled female significant tumor responses of the NTP mouse
study. The mean value of q^ = 2.5 x 10"^ (ppm)"* was then used to predict
human cancer responses in several epidemiologic studies, and the predicted and
actual responses were compared. The comparisons were hampered by a scarcity
of information concerning actual exposures, age distributions, and work histo-
ries. In addition, because there was no consistent cancer response across all
of the epidemiologic studies, the most predominant response, cancer of the
lymphatic and hematopoietic tissues, was chosen as being the target for 1,3-
butadiene. While the predicted and observed responses are consistent, in view
of the uncertainties in the epidemiologic data, a fairly wide range of unit
risk values and exposure estimates also predicts human response satisfactorily.
Given the uncertainties, no better estimate of unit risk can be made.
1-4
-------
1.2. CONCLUSIONS
1,3-Butadiene has been shown to be an indirect mutagen in bacteria. Two
of its potential metabolites, 3,4-epoxybutene and diepoxybutane, are genotoxic
in prokaryote as well as eukaryote test systems. Exposure of rodents to 1,3-
butadiene results in ovarian tumors in mice and testicular tumors in rats,
which offers suggestive evidence that 1,3-butadiene (or, more likely, a metab-
olite of 1,3-butadiene) may reach the germ cells. There is also evidence that
the dimer of 1,3-butadiene, 4-vinyl-l-cyclohexene, causes ovarian tumors in mice.
The total body of evidence, including metabolism, mutagenicity, and car-
cinogenicity data, suggests that 1,3-butadiene may present a genetic risk to
humans. However, mutagenicity studies in mammalian test systems, as outlined
in the EPA's Proposed Guidelines for Mutagenicity Risk Assessment (1984a),
should be conducted to further characterize the mutagenic potential of 1,3-
butadiene.
On the basis of sufficient evidence from studies in two species of rodents,
and inadequate epidemiologic data, 1,3-butadiene can be classified, according
to EPA's Proposed Guidelines for Carcinogen Risk Assessment (1984b), as a
"probable" human carcinogen, Group B2. Using the classification scheme of the
International Agency for Research on Cancer, 1,3-butadiene would be classified
as a "probable" human carcinogen, Group 2B.
A carcinogenic potency and related upper-bound estimate of incremental
lifetime cancer risk can be estimated from the animal studies. These risk
estimates are developed for the purpose of evaluating the possible magnitude
of the public health impact if 1,3-butadiene is a human carcinogen.
Using a multistage model which is linear at low doses, a1 95% upper-limit
i
incremental unit risk for 1,3-butadiene is estimated on the basis of the NTP
1 '
(1984) mouse study to be q± = 6.4 x 10"1 (ppm)"1. The upper-bound nature of
1
"1
1-5
-------
this estimate is such that the true risk is not likely to exceed this value
and may be lower. In addition to a 95% upper-limit incremental unit risk, a
measure of carcinogenic potency was determined for 1,3-butadiene. Among the
55 chemicals that the Carcinogen Assessment Group has evaluated as suspect
carcinogens, 1,3-butadiene ranks in the third quartile.
1-6
-------
2. INTRODUCTION
1,3-Butadiene (CAS No. 106-99-00) is a colorless gas produced as an
ethylene co-product, by oxidative dehydrogenation of n-butenes, or by dehydro-
genation of n-butanes. In 1977, between 2.1 and 7.3 billion pounds of 1,3-
butadiene were produced or imported. 1,3-Butadiene ranked 36th in U.S.
domestic chemical production in 1983. It is used as an intermediate in the
production of polymers, elastomers, and other chemicals. The major use of
1,3-butadiene is in the manufacture of styrene-butadiene rubber (synthetic
rubber). In addition, 1,3-butadiene is used as an intermediate to produce a
variety of industrial chemicals, including the fungicides, captan and captofol.
The U.S. Food and Drug Administration (FDA) has approved 1,3-butadiene for use
in the production of adhesives used in certain types of food containers.
Although 1,3-butadiene has been found in U.S. drinking water, it is pri-
marily an air contaminant. It has been detected in cigarette smoke, incinera-
tion products of fossil fuels, gasoline vapor, and automotive exhaust (Miller,
1978). Concentrations ranging from 1 to 9 ppb have been detected in urban air
(Neligan, 1962). Higher concentrations, up to 45 ppm, have been reported in
air samples and factory emissions at petrochemical plants (NTP, 1984).
Approximately 65,000 workers are potentially exposed to 1,3-butadiene,
according to a report prepared by the National Institute for Occupational
Safety and Health (NIOSH, 1984). The greatest occupational exposure is likely
to occur in plants that manufacture 1,3-butadiene or use it to produce polymers
or elastomers. Occupational hazards from exposure to 1,3-butadiene exist from
inhalation of airborne concentrations and, to a lesser extent, by dermal con-
tact. The current permissible exposure limit of 1,000 ppm as an 8-hour time-
weighted average was adopted by the U.S. Occupational Safety and Health Admin-
2-1
-------
istration (OSHA) from the 1968 Threshold Limit Values (TLV) set by the American
Conference of Government Industrial Hygienists (ACGIH).
While 1,3-butadiene was introduced into commerce because of its use in
the manufacture of synthetic rubber in the early 1940s, little information on
the toxicity of this material is available except for scattered reports from
eastern European countries. Indeed, it was not until 1981 that the first
chronic toxicity study was conducted. To date, adequate long-term investiga-
tions in non-rodent experimental animals are not available. Both the Interna-
tional Agency for Research on Cancer (IARC, 1982) and the U.S. Department of
Health and Human Services (DHHS, 1985) have recognized that occupational expo-
sure in the rubber industry leads to an increased risk of cancer. Because many
materials are used in the rubber industry, it may be impossible to identify any
single causative agent, such as 1,3-butadiene, as a carcinogen, especially
since other chemicals that are generally recognized as carcinogens (i.e.,
benzene and acrylonitrile) are also used in this industry segment.
Both IARC (1982) and DHHS (1985) recognize diepoxybutane, a metabolite of
1,3-butadiene, as a carcinogen. Based on reports of the carcinogenic potential
of 1,3-butadiene, the ACGIH published a Notice of Intended Change (ACGIH, 1983)
in their 1983-1984 Threshold Limit Values (TLV). They proposed to classify
1,3-butadiene as an industrial substance suspected of carcinogenic potential
for man and assigned no numerical TLV. More recently, the ACGIH recommended a
TLV of 10 ppm (ACGIH, 1984). Their previous TLV (the basis of OSHA's current
permissible exposure limit) was 1,000 ppm based on mild irritation in man and
limited effects in rats and guinea pigs at higher concentrations. Based on
the same animal carcinogenicity studies, NIOSH issued a Current Intelligence
Bulletin on 1,3-butadiene. They classified the chemical as a potential occu-
pational carcinogen and recommended that worker exposure be reduced to the
2-2
-------
fullest extent possible. In addition, NIOSH recommended that 1,3-butadiene be
regarded as a potential occupational teratogen and possible reproductive hazard
(NIOSH, 1984). The Office of Toxic Substances (OTS) of the U.S. Environmental
Protection Agency assessed the risk to workers exposed during the production
of 1,3-butadiene and during the use of 1,3-butadiene in the production of
synthetic rubbers, plastics, and resins. They determined that 1,3-butadiene
was a "probable" human carcinogen (category B2 according to EPA's Proposed
Guidelines for Carcinogen Risk Assessment, 1984b). OTS concluded, using a one-
stage model with pooled tumors from male mice, that the upper limits of car-
cinogenic risk for year-round exposure at 10, 1, and 0.1 ppm are 10~1, 10~2,
and ID"3, respectively (U.S. EPA, 1985).
The National Toxicology Program (NTP) is currently undertaking a series of
investigations of 1,3-butadiene. These.studies will provide additional infor-
mation on the pharmacokinetics and toxicity of the chemical. In addition, the
carcinogenic response at lower airborne concentrations may be established.
2-3
-------
-------
3. TOXICOLOGY
3.1. ACUTE TOXICITY
Information on the toxicity of 1,3-butadiene resulting from acute exposure
is limited. For rats and mice, the median lethal concentrations of 1,3-buta-
diene for periods of exposure ranging from 2 to 4 hours are above 100,000 ppm.
The oral LDso values for rats and mice are 5.48 g/kg and 3.21 g/kg, respec-
tively. The major acute toxic effects are irritation of the respiratory tract,
mucous membranes, and eyes, and narcosis (NTP, 1984).
3.2. SUBCHRONIC TOXICITY
A 3-month toxicity study in rats preceded the 2-year chronic inhalation
toxicity study conducted at Hazleton Laboratories Europe, Ltd. in England
(1981a), and sponsored by the International Institute of Synthetic Rubber
Producers, Inc. (IISRP). Further details of the chronic investigation as well
as the results with regard to the carcinogenicity of 1,3-butadiene are presented
in the carcinogenicity chapter of this document. The airborne concentrations
of 1,3-butadiene used in the 3-month study were 1,000, 2,000, 4,000, and
8,000 ppm; a group exposed to filtered air (0 ppm) served as controls. The
authors considered that there were no effects attributable to exposure to
the test chemical on growth rate, food consumption, hemograms, blood biochem-
ical investigations, or pathological evaluation. The only effect the inves-
tigators considered to be related to 1,3-butadiene exposure was a moderate
increase in salivation, particularly among female rats during the last 6 to 8
weeks of exposure at the higher airborne concentrations (Crouch et a!., 1979).
These results are consistent with an earlier study by Carpenter et al. (1944)
in which the investigators found only a slight reduction in body weight gain
among rats and guinea pigs exposed for 8 months, 7.5 hours/day, 6 days/week,
3-1
-------
to an airborne concentration of 8,000 ppm. These authors noted no effects
among animals exposed at 600 and 2,300 ppm.
Preliminary inhalation toxicity studies in mice were used as a basis for
dose selection for the chronic studies of B6C3F1 mice, conducted at Battelle
Pacific Northwest Laboratories and sponsored by the NTP (1984). Further details
of the chronic exposure and the results with regard to the carcinogenicity of
1,3-butadiene are presented in the carcinogenicity chapter of this document.
Two range-finding studies, a 15-day and a 14-week study, were conducted at
International Bio-Test Laboratories. In the 15-day study, weight loss at
airborne concentrations of 1,250 ppm was observed. Even the mice exposed to
8,000 ppm, the highest airborne concentration, survived the exposure period.
In the 14-week study, reduced body weight and death were observed among mice
treated at 2,500 ppm or more. Necropsy findings were not reported (NTP, 1984).
Miller (1978) reviewed a series of papers from Russian investigators,
particularly Ripp (1967), and summarized the subchronic toxic effects in rats.
Ripp (1967) exposed rats to airborne 1,3-butadiene concentrations of 1, 3, and
30 mg/m3 (1 mg/m3 = 0.45 ppm). The highest concentration in this study
(equivalent to 13.5 ppm) is only about 1/50 of the lowest concentration (600
ppm) in any of the other studies reported in this section. At 30 mg/m3, blood
cholinesterase was elevated, blood pressure was lowered, and motor activity was
decreased to 60% of the pre-exposure rate. Histopathological evaluation re-
vealed no changes at 1 mg/m3 except for congestion in the spleen and hyperemia
and leukocyte infiltration in the cardiac tissue. The changes in the cardiac
tissues were more marked at the higher levels with hemmorhage and reduced
cellular RNA reported at the highest concentration. At 3 and 30 mg/m3,
atelectasis, interstitial pneumonia, and emphysema were noted in the lung
tissue. These results, showing an adverse response at such low levels, may
3-2
-------
indicate that continuous exposure is more hazardous than intermittent exposure,
the regimen used in all other investigations. An alternative explanation
for these findings at such low levels is that Ripp (1967) examined more sensi-
tive indicators of toxicity than other investigators.
3.3. CHRONIC TOXICITY
A 2-year chronic inhalation toxicity study of the effects of airborne
concentrations (1,000 and 8,000 ppm) of 1,3-butadiene on rats was conducted at
Hazleton Laboratories Europe, Ltd. in England (HLE, 1981a) and was sponsored by
the IISRP. Among rats exposed to 8,000 ppm, clinical signs, consisting of exces-
sive secretion of the eyes and nose plus slight ataxia, were observed between
months 2 and 5 of the study. Variations in mean body weight suggested no
consistent adverse effect. Review of the hemograms, blood chemistry, urine
analysis, and behavioral testing was likewise not indicative of an adverse
effect. In female rats exposed to either 1,000 or 8,000 ppm, subcutaneous '
masses appeared earlier and at higher incidences than in the control. A dose-
related increase in liver weights was observed among rats at the necropsy
performed at 52 weeks and among these killed at the termination of the study.
This could indicate that the chemical induces liver enzymes. Otherwise, no
significant changes were noted at the 52-week kill. Increased alveolar meta-
plasia and nephropathy were observed among males of the 8,000 ppm treatment
groups at the termination of the study. Marked or severe nephropathy occurred
in 27% of the male rats in the high-dose group as compared with 9% to 10% in
the control and the low-dose groups. The authors considered nephropathy to be
the cause of some of the early deaths in this study.
A lifetime chronic inhalation study in B6C3F1 mice at 1,3-butadiene con-
centrations of 625 and 1,250 ppm administered for 6 hours/day, 5 days per week,
was sponsored by the NTP (1984). The exposures were prematurely terminated
3-3
-------
after 61 weeks due to deaths resulting largely from the development of cancer.
No increases in clinical signs could be associated with exposure to the test
chemical. In addition to the neoplastic changes described in the carcinogen-
icity chapter, testicular atrophy (control-0/50, 625 ppm-19/47, 1,250 ppm-11/48)
and ovarian atrophy (2/49, 40/45, 40/48) were elevated among mice at both
doses. Furthermore, among male mice, there was a significant increase in
liver necrosis at both doses. In female mice, liver necrosis was significantly
elevated only at the higher airborne concentration. While neoplastic lesions
of the nasal cavity were not found at any dose, there was an increase in non-
neoplastic changes at the high dose. At 1,250 ppm, chronic inflammation of
the nasal cavity (male, 33/50; female, 2/49), fibrosis (male, 35/50; female,
2/44), cartilaginous metaplasia (male, 16/50; female 1/49), osseous metaplasia
(male, 11/50; female, 2/49), and atrophy of the sensory epithelium (male,
32/50) were observed. No non-neoplastic lesions of the nasal cavity were
found in the controls (NTP, 1984). Huff et al. (1985) have suggested that the
lack of neoplasms in the nasal cavity as compared to the lungs may reflect a
requirement for biotransformation of 1,3-butadiene to a reactive epoxide inter-
mediate. The nasal cavity changes then suggest that the intact molecule may
have some adverse effect at 1,250 ppm. However, as discussed in the metabolism
and carcinogenicity chapters, exposure to 1,3-butadiene did not decrease
minute volume, as occurs with other respiratory irritants.
In summary, since a no-effect dose has not been established for the non-
carcinogenic chronic toxicity, further investigations are warranted. These
investigations should focus on the liver, testes, and ovaries because in mice
these tissues are adversely altered at the lowest concentrations. In addition,
the minimum effect dose for cardiac and respiratory tract changes needs to be
further explored. For a complete understanding of the toxicity, non-rodent
3-4
-------
species might be used.
3.4. REPRODUCTIVE TOXICITY
A teratogenicity study sponsored by the IISRP was conducted at Hazleton
Laboratories Europe, Ltd. in England (HLE, 1981b). Pregnant female Charles
River CD rats (Sprague-Dawley), obtained from Charles River Ltd., were exposed
to airborne concentrations of 1,3-butadiene (200, 1,000, and 8,000 ppm), 6
hours/day, on days 6 to 15 of gestation. Each .treatment group comprised 24
female rats; 40 were assigned to the control group (referred to as 0 ppm
exposure) which was exposure to filtered air. A group of 26 pregnant rats was
administered acetylsalicylic acid at a dose of 250 mg/kg by gavage and served
as a positive control. The rats were observed daily and weighed at intervals
during the study. On day 20 of gestation the females were killed and necrop-
sied, and the uterine contents were inspected. One-third of each litter was
examined for soft tissue abnormalities using a modified Wilson's technique
(free hand sectioning of the heads). The remainder were examined for visceral
and, after preparation, skeletal anomalies.
The positive control group had sufficient adverse reproductive and terato-
genic effects to establish the responsiveness of this strain of rats in the
hands of these investigators. The treated dams were not affected by exposure
to 1,3-butadiene except for reduced weight gain in those exposed to 8,000 ppm.
Selected data for other reproductive toxicity end points are summarized in
Table 3-1. The authors concluded that there was embryonic growth retardation
(fetal weight and length), and slight embryolethality (percentage implantation
loss) in all dose groups as a result of the maternal toxicity and that the
magnitude of the effect was dose related. The relationship between maternal
toxicity and the fetal effects was a subjective judgment and not experimentally
established. The authors further concluded that at the highest doses there was
3-5
-------
TABLE 3-1. EFFECTS OF 1,3-BUTADIENE EXPOSURE ON PREGNANT RATS
Parameter Control
0 200
% preimplantation loss 3.6 6.0
Mean fetal weight (g) 3.3 3.2
Mean crown-rump length 37.8 37.2
(mm)
% of fetuses w/skeletal 86 90
variants
% of fetuses w/skeletal 23 27*
minor defects
% of fetuses w/skeletal 0.6 2.2
major defects
1,000 8,000
4.9 7.3
3.2 3.1a
37.2 35. 9b
86 98^
21 21
3.8a 5.9b
^Different from control (p < 0.05) using the Wilcoxon Test.
Different from control (p < 0.01) using the Wilcoxon Test.
^Different from control (p < 0.05) using the Fisher Exact Test.
3-6
-------
an indication of teratogenicity based on the presence of major fetal defects.
Most of the major skeletal defects (marked or severe wavy ribs) were among
fetuses of the 200 and 1,000 ppm groups. Some investigators might not consider
such changes as major defects. There was a significant increase in "minor"
skeletal defects at the lowest (200 ppm) airborne concentration. Furthermore,
a statistically significant (p < 0.05) increase in minor external and visceral
defects was noted among fetuses in the treatment groups. Subcutaneous hematomas
were the most frequent finding in the lower exposure groups. The frequency of
fetuses with lens opacities was increased at the highest exposure (HLE, 1981b).
Whether these increases represent a qualitative dose-response change in view
of the changes at the higher dose levels or are due to other factors, such as
maternal toxicity, cannot be determined from the information available. Never-
theless, it would appear that the fetuses from dams exposed to concentrations
of 200 to 8,000 ppm were adversely affected in a manner perhaps only related to
growth.
In an earlier but inadequately reported study (Carpenter et al., 1944),
decreased litter size was present among rats exposed to 6,700 and 2,300 ppm
1,3-butadiene, but not at 600 ppm. Carpenter et al. (1944) and the unpublished
study sponsored by IISRP (HLE, 1981b) provide only suggestive evidence that
1,3-butadiene causes adverse reproductive effects in female rats. Further
investigations using other species would be useful in evaluating this possible
hazard. In view of the ovarian and testicular toxicity noted in the NTP (1984)
chronic inhalation study, the scope of these investigations should not be
limited only to exposure during pregnancy.
3-7
-------
-------
4. METABOLISM AND PHARMACOKINETICS
The major hazard with regard to the adverse effects of exposure to 1,3-
butadiene is via inhalation. There is also a slight potential for dermal
exposure. Absorption through the gastrointestinal tract is possible and might
result from contamination of water or leaching of the compound into food.
Useful information concerning the pharmacokinetics of 1,3-butadiene has
only recently become available. The older information is limited by the lack,
of sensitivity of the analytical techniques used, and also because the extremely
high doses used frequently approached lethal doses.
Shugaev (1969) exposed rats for 2 hours to an airborne 1,3-butadiene con-
centration of 130,000 ppm and measured tissue concentrations. He found the
highest concentration of the chemical in the peri renal fat (152 mg %) and lower
concentrations, ranging from 36 to 50 mg %, in the liver, brain, spleen, .and
kidney. The tissue concentrations following exposure to the 1X50 concentration
in the brain tissue of rats and mice (only brain concentration was determined
in mice) were 50.8 and 54.4 mg % for rats and mice, respectively.
4.1. METABOLISM
4.1.1. In vitro Metabolism
In light of the implication that the mutagenic effect of 1,3-butadiene
towards Su typhimurium is due to the formation of reactive metabolites (de
Meester et al., 1978, 1980), the formation of these metabolites in mammalian
species is of vital interest. In addition, the metabolite, diepoxybutane, has
also been demonstrated to be carcinogenic by dermal application and subcutane-
ous injection (IARC, 1982). Malvoisin and Roberfroid (1982) demonstrated that
1,3-butadiene is metabolized to epoxybutene, 3-butene-l,2-diol, diepoxybutane,
and 3,4-epoxy-l,2-butanediol by rat liver microsomes. The pathway proposed
4-1
-------
by these authors for 1,3-butadiene metabolism, based on data obtained from in
vitro experiments, is that epoxybutene is formed from 1,3-butadiene via micro-
somal oxidase (P450) and that it is then further oxidized to diepoxybutane by
the same enzyme system or converted to 3-butene-l,2-diol by epoxide hydro!ase.
The 3-butene-l,2-diol then reacts with molecular oxygen to form the epoxide,
3,4-epoxy-l,2-butanediol, again via the microsomal oxidase system. Figure 5-1
depicts the proposed pathway of 1,3-butadiene metabolism. Liver microsomes
prepared from animals pretreated with phenobarbital metabolized more 1,3-buta-
diene to 3,4-epoxy-l,2-butanediol and diepoxybutane than microsomes from animals
not pretreated. Microsomes prepared from animals pretreated with Metyrapone
and SKF 525A metabolized less 1,3-butadiene to the diepoxide but not the diol.
Styrene oxide and vinylcyclohexene oxide almost completely inhibited the forma-
tion of the diol. Both were observed to have substantially greater affinity
than 1,3-butadiene as substrates for the epoxide hydrolase enzyme. The authors
stated that epoxybutene is a poor substrate for epoxide hydrolase relative to
styrene oxide. They suggested that this may account for the fact that 1,3-buta-
diene is a more potent mutagen than styrene (Malvoisin et al., 1980; Malvoisin
and Roberfroid, 1982).
The in vitro formation of epoxybutene (vinyl oxirane) by rat liver micro-
somal preparations has been confirmed by Bolt et al. (1983). These investiga-
tions also found that microsomes prepared from animals pretreated with pheno-
barbital or 3-methylcholanthrene metabolized more 1,3-butadiene to epoxybutene,
and that microsomes prepared from animals pretreated with SKF 525A or diethyl-
dithiocarbamate metabolized less 1,3-butadiene than microsomes from animals not
pretreated.
Malvoisin and Roberfroid (1982) found that epoxybutene reacts both chemi-
cally and enzymatically with glutathione-S-transferase to form a glutathione
4-2
-------
conjugate. The significance of this observation with respect to the toxicity
of epoxybutene awaits further investigation. Citti et a!. (1984) demonstrated
the formation of an N-7 guanine adduct of epoxybutene after incubation with
either deoxyguanosine or DNA. The authors suggested that the formation of these
adducts may account for the mutagenic effect of 1,3-butadiene and its reactive
epoxide metabolites.
4.1.2. In vivo Metabolism
Bolt et al. (1983) exposed rats (Wistar) to 1,3-butadiene by inhalation
and found epoxybutene in the exhaled air, thus confirming the formation of
this mutagen in in vivo experiments. These results were corroborated by Filser
and Bolt (1984) and Bolt et al. (1984), who also found epoxybutene in expired
air of rats after exposure to 1,3-butadiene. Their results are described in
more detail in section 4.2.
Laib et al. (1985) reported on binding of 14C-l,3-butadiene to liver and
nucleoproteins and DNA. Four rats (Wistar) and 24 B6C3F1 mice were exposed to
labeled (1,4-C14) 1,3-butadiene at an air concentration of about 700 ppm for
6.6 or 4 hours, respectively. More than 98% of the total activity could be
accounted for when the animals were sacrificed 30 minutes after exposure. The
radioactivity expressed as cpm/mg in the liver of early nucleoproteins and DNA
has been tabulated in Table 4-1. The late-eluting nucleoproteins had the same
rat/mouse ratio. These results show that mice incorporate 1,3-butadiene or
its metabolites into the nucleoproteins and DNA at a rate greater than the rats.
All four of the in vivo studies cited above reported that the metabolism of
1,3-butadiene to epoxybutene was at a maximum. In rats this maximum conversion
appears to take place at atmospheric concentrations of 1,3-butadiene greater
than 1,000 ppm (Bolt et al., 1983, 1984; Filser and Bolt, 1984; Laib et al.,
1985). Pretreatment of rats with Aroclor 1254 (polychlorinated biphenyl)
4-3
-------
TABLE 4-1. MEAN RADIOACTIVITY IN LIVER NUCLEOPROTEINS
AND DNA OF MOUSE AND RAT AFTER EXPOSURE TO 14C-1,3-BUTADIENE
cpm/mg
rat mouse
cpm/mg/hour
rat mouse
nucleoprotein
323 613
DNA
720 628
48.9
109
153
157
SOURCE: Laib et al., 1985.
rendered the metabolism of 1,3-butadiene less complete at atmospheric concentra-
tions up to 12,000 ppm (Bolt et al., 1984). The metabolism of 1,3-butadiene
in mice approaches maximum conversion at 1,800 ppm (Laib et al., 1985).
4.2. PHARMACOKINETICS
The pharmacokinetics of 1,3-butadiene have been studied using two different
methods of exposure. The first involved the exposure of rats (Bolt et al., 1983,
1984; Filser and Bolt, 1984; Laib et al., 1985) and mice (Laib et al., 1985)
to various atmospheric concentrations of 1,3-butadiene in a sealed chamber.
1,3-Butadiene was introduced into the chamber at selected concentrations, and the
declines in these concentrations were measured over time. The pharmacokinetics
of this system were described by Filser and Bolt (1981). Filser and Bolt (1981)
assumed that the decline in the concentration of 1,3-butadiene in the chamber
was a result of uptake and metabolism of the compound by the animal. Based on
these assumptions, certain pharmacokinetic parameters may be estimated. By
4-4
-------
inhibiting the metabolism of the compound, the relative contributions of uptake
and metabolism may also be assessed. In addition-, the results from such experi-
ments have been used to estimate the kinetics of the compound under the
assumption that the volume of the chamber is infinite (Filser and Bolt, 1981).
The resulting parameter estimates were reported to apply to environmental
exposure in which the concentration of the compound in the atmosphere is
constant and unaffected by the animal.
Initial experiments in rats (Wistar) by Bolt et al. (1983) indicated
that, under the conditions described above, the decline in the concentration
of 1,3-butadiene in a closed system follows zero-order kinetics at atmospheric
concentration above 1,000 ppm and that below that concentration, the decline
in the concentration of 1,3-butadiene follows first-order kinetics. When the
concentration of 1,3-butadiene in the chamber was above 1,000 ppm, there was
a constant accumulation of the primary metabolite epoxybutene. When the concen-
tration of 1,3-butadiene was 1,000 ppm or less, the concentration of epoxybutene
declined in approximately first-order fashion, which suggests that epoxybutene
was reabsorbed and further metabolized. This observation was confirmed by
Filser and Bolt (1984) in experiments in which rats were administered epoxybutene
by inhalation in a closed system. In these experiments, a first-order decline
of epoxybutene was observed at concentrations as high as ~3,500 ppm. The
same rate of decline was observed when epoxybutene was administered intraperito-
neally and then exhaled.
A kinetic analysis of 1,3-butadiene in rats using the closed-system
technique was conducted and reported by Bolt et al. (1984). As observed in their
previous studies, the decline in the concentration of 1,3-butadiene in the
chamber exhibited zero-order kinetics at atmospheric concentrations above 1,000
ppm and first-order kinetics below this concentration. About 20% of the initial
4-5
-------
content of the chamber remained after 6 hours. Using equations reported in an
earlier article (Filser and Bolt, 1981) the pharmacokinetic parameters shown
in Table 4-2 were calculated. Using these parameters, Filser and Bolt (1984)
found that only 30% of epoxybutene that had been predicted to be formed from
the metabolism of 1,3-butadiene under saturated conditions was expired by the
animals under these conditions. The remaining 70% apparently was either retained
in the body, further metabolized, or excreted via other routes.
Laib et al. (1985) conducted experiments with mice (B6C3F1) that were
similar to the rat (Sprague-Dawley) experiments of Bolt et al. (1984). The
pharmacokinetic parameters obtained for mice are compared to those of rats
in Table 4-1. Of particular interest is that the rate of uptake of 1,3-butadiene
(kigVl)* the static and dynamic equilibrium constants (Keq and Kst, respective-
ly)* the total clearance (Cltot)» tne maximum rate of metabolism (Vmax), and
the atmospheric concentration of 1,3-butadiene resulting in the saturation of
uptake and/or metabolism are greater in mice than in rats. Using the pharmaco-
kinetic parameters calculated for rats (Bolt et al., 1984) and for mice, Laib
et al. (1985) concluded that the rate of metabolism in mice was approximately
twice that in rats at any given atmospheric concentration of 1,3-butadiene.
In investigations in progress for the .National Toxicology Program (NTP)
at Lovelace Inhalation Toxicology Research Institute (LITRI) (NTP, 1985a),
rats (Sprague-Dawley) and mice (B6C3F1) were exposed to 1,3-butadiene via in-
halation using a different exposure paradigm. In the initial studies, animals
were exposed nose-only to constant atmospheric concentrations of l^C-l,3-buta-
diene for 6 hours and then placed in metabolism cages which allowed for the
separate collection of urine, feces, and expired air. There is one important
difference between this method of exposure and the closed-chamber method used
by Bolt and coworkers. In the LITRI studies, animals were exposed to a constant
4-6
-------
TABLE 4-2. PHARMACOKINETIC PARAMETERS FOR 1,3-BUTADIENE
Parameter
k!2Vl
Keq
i/ a
Kst
C1tota
V b
"max
Metabolic
saturation
Bolt et al.
(1984)
rats
5,750
2.3
0.5
4,490
220
1,000
Laib et al .
(1985)
mice Units,
10,280 mL'hr'1
3.2
1.0
6,750 mL'hr'1
400 ymoles*hr~1*kg~1
1,800 ppm
aKst and C1tot calculated for V-^ -
max va^c' f°r saturation range.
SOURCE: Laib et al., 1985.
4-7
-------
atmospheric concentration of 1,3-butadiene for a fixed length of time, and then
exposure was discontinued. In contrast, using the closed-chamber method, ani-
mals are exposed to a constantly decreasing concentration of 1,3-butadiene. The
latter procedure allowed for rebreathing of the parent compound and exhaled
metabolites.
In the preliminary studies conducted at LITRI, rats were exposed for 6
hours to atmospheric concentrations of 14r,-is3-butadiene of 125, 1,700, and
12,800 pg/L. Mice were exposed to concentrations of 13, 145, and 1,900 yg/L
for 6 hours. These concentrations correspond to 70, 930, and 7,100 ppm,
respectively, for rats and 7, 80, and 1,040 ppm, respectively, for mice under
the conditions at LITRI (25°C, 620 mm). Preliminary experiments were also con-
ducted in which rats and mice were exposed to 1,3-butadiene for 6 hours and
their respiratory parameters were measured by plethysmographs.. No significant
concentrated-related differences were reported for minute volume in rats. In
mice, respiratory irritants decreased the minute volume.
The fraction of the inhaled compound that was absorbed and retained was
calculated (see Table 4-3) based on the respiratory parameters obtained from
the plethysmographic studies and the amount of radioactivity recovered in
another study from animals that were similarly exposed and placed in metabolism
cages. Dose-dependent retention was observed in both species. This is in
general agreement with Bolt et al. (1984) and Laib et al. (1985). However,
using the calculated areas under the curves of exposure concentration versus
time from 0 to 6 hours as a measure of relative exposure, the levels at which
dose-dependent uptake occurred are lower in the LITRI report than in the Bolt
studies. This is most likely due to the differences in methods of exposure.
The LITRI study also indicates that at similar concentrations mice received 3
to 4 times the dose (pmol/kg) of that received by rats (Table 4-3).
4-8
-------
z.
o
1 1
1
=C
_J
-
CQ
LU
~S*r
^
LU
O
=C
CO
CO
u_
o
z.
o
h-H
f
0-
o;
CD
oo
CO
oo
1
LU
CQ
^^
H-
^_j
c
0)
o
£_
0)
Q-
O)
c
d) T3
r- O)
T3 C
1C -r-
4J (O
3 4->
CO CD
1 I-
oo
>
01
c:
O)
i- T3
(T3 n~
4-> c\j r-n
<£> "d" CO O OO <7>
CM
i r~-
LfJ O O - r-< * LO
oo o o oo oo
CVJ r-H O «*
A A
co r
r 1
LO O O CO LO O
CVJ O O r-H «d- O
rH I***. 00 r-H O">
«i «\ «i
r-H CSJ i)
1-1
o o o r-~ o o
r oo o oo «*
CTl r-H O
i^ . » i
v> . 0
C£ £
A
fO
oo
CTl
< 1
1
LU
C_5
a;
o
oo
4-9
-------
The proportions of the absorbed radioactivity accounted for in the breath,
urine, feces, and body of rats and mice after exposure to 14C-l,3-butadiene
in the LITRI study are presented in Table 4-4. As can be seen from the table,
the data suggest that the rats displayed a dose-related increase in the propor-
tion of the absorbed radiolabel exhaled during the 65 hours after exposure.
The rat data also suggest that a concomitant dose-related decrease in the propor-
tion of the absorbed radiolabel excreted in urine was also observed. The data
from mice showed no dose-related trend in the pattern of excretion at the
exposure levels used. The proportion of the absorbed radiolabel excreted in
feces and retained in the body after 65 hours was similar over the range of
exposures examined. Although these results are expressed in terms of total
radioactivity and the radioactive species in any of these media were not identi-
fied, dose-dependent metabolism of 1,3-butadiene can be inferred.
In summary, the work of Bolt and coworkers (Bolt et al., 1983, 1984;
Filser and Bolt, 1984; Laib et al., 1985) and LITRI (NTP, 1985a) provide infor-
mation on the metabolism and pharmacokinetics of 1,3-butadiene. The apparent
differences between the results from these two laboratories are probably due
to different methods of exposure. The published papers from Bolt's laboratory
and the preliminary reports from LITRI are in general agreement that the uptake
of 1,3-butadiene in rats and mice is concentration-dependent and that its
dose-dependent metabolism is observed at relatively higher atmospheric concen-
trations in mice than in rats.
4-10
-------
o
"I
1 UJ
«=C OC.
UJ =)
a; co
CQ O
O_
< X
l l UJ
^>
Qi
Q UJ
UJ I
1 U-
UJ
O
UJ
CQ LO
- 2:
1 1 Q t t 1
Q O > i
<=C OQ Q
Cc: «=C
UJ 1
Q 3: =3
UJ 1 CQ
CQ 1
Q; 2: co
O i i «
co «-
CQ CD 1
«c 2: o
O 1-1
1 1 1 y
CD UJ
«C o;
2: Q
uj 2:
O =C
Q£
UJ OO
CL. UJ
O
UJ
u_
1 f"^l
=c
UJ
] «
CQ U)
^C 21
^~ i^
§
T3
cn o
1 > f~ r*i
C -r-
cu c cu
O t- JT
£- fO 4->
CU E
Q. CU C
S- -r-
CO
CU
o
CU
U.
c
1
-o
CU
4-> CU
cu c
t_ -r-
0 t-
X =3
CU
M
C
CU
o
t- s:
CU +J
a. to
cu
4_
CO
£1
O
CU +->
t- (0
3 t-
O C _1
o. cu --»,
x o cn
UJ C i
O *-^'
o
w
CU
r-
O
Q.
00
O O «3" t 1 l£> CM
CM CM i) rt t 1 t I
LO CM CM 1 ,-H CO
i i
CO CM OO O CT> f~-
r-- to 10 CM «* oo
CM *^~ r** "Ci" CM >!*
LO CD O CO LO CD
CM o o < i «=c o
f- ' r~- co r i cri
-"CM r*
r- 1
in CU
-!-> 0
tO -r-
rr* y
tf
0
Q.
a;
Q-
L^.
^^
«-
-o
CU
(O
.S
4^
t/J
CU
W5
cn
p
CU
0
cu
Q_
ro
LO
00
cr>
t 1
D_
1
1
Ul
o
o
co
4-11
-------
-------
5. MUTAGENICITY OF 1,3~BUTADIENE AND ITS REACTIVE METABOLITES
This chapter is concerned with the mutagenicity of 1,3-butadiene, which is
a gas at room temperature, and also includes a discussion of the mutagenicity
of its reactive metabolites (3,4-epoxybutene and l,2:3,4-diepoxybutane). The
available evidence suggests that 1,3-butadiene is mutagenic by virtue of its
metabolism to mutagenic intermediates.
5.1. MUTAGENICITY OF 1,3-BUTADIENE
1,3-Butadiene was tested for its mutagenic potential in the Salmonella
typhimurium histidine reversion assay by de Meester et al. (1978). The sample
of 1,3-butadiene studied was 99.5% pure and was obtained from Matheson Gas Prod-
ucts, Belgium. Salmonella strains TA1530, TA1535, TA1537, TA1538, TA98, and
TA100 were exposed to 1,3-butadiene vapors for 20 hours in closed desiccators.
Mutagenic activity was observed in strains TA1530 and TA1535 both in the pres-
ence and absence of S9 prepared from Aroclor-pretreated rats. The bacteria
were exposed to only one dose of 1,3-butadiene and that dose was not clearly
9
specified. This study suggests that 1,3-butadiene is a direct-acting, base-
pair substitution mutagen in bacteria.
In a subsequent study, de Meester et al. (1980) exposed strain TA1530 to
1,3-butadiene vapors for 24 hours at 0, 4, 8, 16, 24, and 32% (vol/vol) in
closed desiccators. In the absence of S9 mix or in the presence of S9 prepared
from untreated rats, no increase in the revertant frequency was observed.
However, when the bacteria were exposed to 1,3-butadiene in the presence of S9
mix prepared from phenobarbitone- or Aroclor 1254-pretreated rats, mutagenic
activity was observed. The number of histidine revertants increased in a
dose-related fashion from 17 per plate in the absence of 1,3-butadiene up to
255 per plate at 16% (vol/vol) 1,3-butadiene. These results suggest that
5-1
-------
1,3-butadiene itself is not a mutagen, and that it is metabolized to mutagenic
intermediates that cause base-pair substitutions.
Data suggesting that the mutagenic metabolites are volatile were also
reported by de Meester et al. (1980). Plates containing Salmonella but no S9
mix were coincubated in 1,3-butadiene atmospheres (4%-32%) with plates contain-
ing Salmonella plus S9 mix from chemically pretreated animals. The reversion
rates were enhanced in both sets of plates, indicating that 1,3-butadiene was
metabolized in the plates containing the S9 mix to volatile mutagens which were
active in the plates without the S9 mix. The number of mutant colonies was
proportional to the level of 1,3-butadiene in the desiccators up to 16%.
This second study by de Meester and coworkers (de Meester et al., 1980)
contradicts their earlier observation (de Meester et al., 1978) that 1,3-buta-
diene is mutagenic in the absence of S9 mix. If, in the earlier study, plates
containing bacteria plus S9 mix were incubated in the same dessicators with
plates containing bacteria but no S9 mix, volatile mutagenic metabolites of
1,3-butadiene generated in the first set of plates may have been responsible
"i
for the mutagenic effects observed in the second set of plates.
A study by Poncelet et al. (1980) supports the conclusion of de Meester et
al. (1980) on the mutagenic potential of 1,3-butadiene in Salmonella strain
TA1530. Mutagenic effects were observed when the assays were performed in a
16% gaseous atmosphere of 1,3-butadiene in the presence of Aroclor-induced
S9 mix. When the bacteria were exposed to 1,3-butadiene under the conditions
of the plate incorporation method or preincubation in liquid medium, mutageni-
city was not observed. The 1,3-butadiene sample was 99.5% pure and was obtained
from Matheson Gas Products, Belgium.
In summary, the weight of the available evidence suggests that 1,3-butadi-
ene is a promutagen in bacteria; its mutagenicity depends on metabolic activa-
5-2
-------
tion by S9 mix prepared from chemically induced animals. No whole-animal muta-
genicity studies have been reported.
5.2. METABOLISM OF 1,3-BUTAOIENE AND REACTION OF METABOLITES WITH DNA
As described in the previous section, 1,3-butadiene itself does not appear
to be mutagenic. Mutagenic activity is observed only when 1,3-butadiene has
been metabolized to reactive intermediates. A scheme for the metabolism of
1,3-butadiene and the alkylating ability of the probable metabolites are brief-
ly discussed in this section.
Malvoisin et al. (1979) and Malvoisin and Roberfroid (1982) studied the
metabolism of 1,3-butadiene in vitro using rat liver microsomes. They reported
that the metabolism proceeds via a mixed-function oxidase-catalyzed oxidation
to 3,4-epoxybutene, and they suggest that this compound is subsequently metabo-
lized to l,2:3,4-diepoxybutane (diepoxybutane) and 3-butene-l,2-diol (Figure
5-1). Both 3,4-epoxybutene and diepoxybutane are probably reactive intermedi-
ates, whereas 3-butene-l,2-diol and its metabolite 3,4-epoxy-l,2-butanediol are
probably detoxification products. Preliminary evidence in rats suggests that
1,3-butadiene is metabolized to 3,4-epoxybutene in vivo (Bolt et al., 1983),
indicating that the metabolic pathway outlined in Figure 5-1 on the basis of i_n_
vitro data may occur in vivo. Both 3,4-epoxybutene and diepoxybutane are muta-
genic, as described more fully below. Malvoisin and Roberfroid (1982) state
that their unpublished results indicate that 3-butene-l,2-diol and 3,4-epoxy- '
1,2-butanediol are not mutagenic.
The alkylating activity of the two reactive metabolites of 1,3-butadiene
(3,4-epoxybutene and diepoxybutane) has been investigated, 3,4-epoxybutene in
two studies and diepoxybutane in one study. Hemminki et al. (1980) found that
3,4-epoxybutene alkylated 4-(p-nitro-benzyl)-pyridine (NBP) and deoxyguanosine,
which are nucleophiles that were used as models for DNA. The NBP reaction was
\
5-3
-------
CH2=CH-CH=CH2
1,3-Butadiene
NADPH
On microsomes
CH9=CH-CH-CH2
V
3.4-Epoxybutene
Epoxide
hydrolase
CH2= CH-CHOH-CH2OH
3-Butene-1,2-diol
i
NADPH
O2, microsomes
CH2-CH-CHOH-CH2OH
V
3,4-Epoxy-1,2-butanediol
NADPH
02, micfosomes
CH2-CH-CH-CH2
V
.1 ,2:3 ,4-Diepoxy butane
Figure 5-1. A hypothetical scheme for the metabolism of 1,3-butadiene.
SOURCE: Malvoisin and Roberfroid, 1982.
5-4
-------
carried out at 37°C using the test compound at 0.286 yM.. The deoxyguano-
sine reaction was carried out at 37°C using the test compound at 0.1 M. In
both cases, aliquots of the reaction mixture were assayed for alkylation at 0
minutes, 20 minutes, 1 hour, 3 hours, and 5 hours. The reaction rates were
determined from the initial rates, and the results were reported using epichlo-
rohydrin as a reference. 3,4-Epoxybutene alkylated NBP and deoxyguanosine at
rates that were 31% and 14%, respectively, of the alkylation by epichlorohy-
drin. In agreement with Hemminki et al. (1980), Citti et al. (1984) published
data suggesting that 3,4-epoxybutene alkylates deoxyguanosine and DNA in vitro.
The main products formed for both deoxyguanosine and DNA were 7-(2-hydroxy-3-
buten-1-yl)deoxyguanosine and 7-(l-hydroxy-3-buten-2-yl)deoxyguanosine.
Lawley and Brookes (1967) reported that diepoxybutane reacts with DNA in a
manner typical of bifunctional alkylating agents and causes interstrand cross-
linking in DNA. Salmon sperm DNA was dissolved in 0.5 mM sodium citrate at 2
mg/mL (5.4 mM DNA phosphorus) and 25 ml was treated with redistilled diepoxybu-
tane (2.4 mg/mL, 28 mM) at 37°C. Samples were withdrawn after 2, 5, 7, 24, 48,
72, 120, 168, and 193 hours. Ultraviolet spectroscopy was used to measure the
reaction of diepoxybutane with DNA. At 37°C, diepoxybutane reacted with DNA
slowly, as shown by changes in ultraviolet absorption of the reaction mixtures.
The extent of interstrand cross-linking (i.e., covalent linkage of the two
DNA strands by the reaction of diepoxybutane with a nucleotide base in each
DNA strand) was studied by measuring the reversible denaturation (renaturation)
of diepoxybutane-treated DNA. In this experiment, diepoxybutane-treated DNA
was first incubated at 60°C for various time periods and then rapidly cooled.
At 60°C, untreated DNA denatures when it is dissolved in a solution of low
ionic strength (i.e., the two strands separate). When cooled, the DNA rena-
tures (i.e., the two strands rejoin to reform the typical double-stranded DNA
5-5
-------
molecule). Diepoxybutane-treated DNA renatured to a greater extent than did
normal untreated DNA, suggesting that diepoxybutane-treated DNA was cross-linked
by a diepoxybutane bridge covalently joining the two DNA strands.
In summary, the above three in vitro studies indicate that 3,4-epoxybutene
and diepoxybutane can alkylate DNA and that diepoxybutane causes interstrand
cross-links in DNA. In addition, preliminary in vivo data suggest that 1,3-
butadiene alkylates DNA in rats and mice after inhalation exposure (Laib et
a!., 1985), presumably via epoxide intermediates.
5.3. MUTAGENICITY OF 3,4-EPOXYBUTENE
In addition to studying the mutagenicity of 1,3-butadiene, de Meester et
al. (1978) tested its monoepoxide metabolite, 3,4-epoxybutene, by the plate
incorporation method in Salmonella. Strains TA1537, TA1538, and TA98 exhibited
a negative response. Reversion to histidine prototrophy was observed with
strains TA1530, TA1535, and TA100, with maximal numbers of revertants occurring
at a 3,4-epoxybutene concentration of 100 pmol/plate.
The mutagenic potential of 3,4-epoxybutene was studied in the fluctuation
test with Klebsiella pneumoniae as the test organism (Voogd et al., 1981). The
compound was obtained from K and N (ICN, Life Sciences Division, New York), was
analytical grade, and was not further purified. The chemical was dissolved and
diluted in dimethylsulfoxide and subsequently added to broth which was inoculated
with the test organism. The genetic characteristic studied was streptomycin
resistance. The average spontaneous mutation rate for streptomycin resistance
was 0.1676 x 10~9. Triplicate experiments were averaged, and the results were
expressed as the quotient of the observed and spontaneous mutation rates. At 1
and 2 mM 3,4-epoxybutene, the quotients were 1.7 and 2.5, respectively, provid-
ing evidence of a dose-related positive response.
In addition to studying the alkylating activity of 3,4-epoxybutene,
5-6
-------
Hemminki et al. (1980) studied its mutagenicity (reversion to tryptophan proto-
trophy) in !£_. coli strain WP2 uvrA. The concentrations of 3,4-epoxybutene
used were not specified but were based on toxicity determinations. Although
the study suggests that 3,4-epoxybutene is mutagenic in IE_. coli WP2 uvrA, there
was no indication of a dose-related response; the result for only one dose was
reported and that dose was unspecified. Although this study is of limited use
for risk assessment purposes, it supports the study of Voogd et al. (1981) in
suggesting that 3,4-epoxybutene is mutagenic in bacteria.
Unpublished data suggest that 3,4-epoxybutene injected intraperitoneally
induces sister chromatid exchange (SCE) and chromosomal aberrations in mouse
bone marrow cells (Allen, personal communication, 1985). SCE involves the
reciprocal exchange of DMA segments between sister chromatids and is con-
sidered an indication of DNA damage. A. dose-related increase in SCEs was ob-
served at 3,4-epoxybutene doses ranging from 10 mg/kg to 150 mg/kg (Table 5-1).
A dose-related increase in chromosomal aberrations at 25-150 mg/kg was also ob-
served (Table 5-2). The types of aberrations observed included chromatid and
chromosome/isochromatid breaks, chromatid deletions, fragments, acentric frag-
ments, and polyploidy. It is possible that the 3,4-epoxybutene was biotrans-
formed after intraperitoneal injection and that the positive results for SCE and
chromosomal aberrations in bone marrow cells were due to a metabolite of 3,4-
epoxybutene.
In summary, 3,4-epoxybutene is mutagenic in bacteria and induces SCE and
chromosomal aberrations in mice.
5.4. GENOTOXICITY OF 1,2:3,4-DIEPOXYBUTANE
Information on. the genotoxicity of diepoxybutane can be found in a review
on the genotoxicity of several epoxides by Ehrenberg and Hussain (1981). This
review covers much of the early literature on diepoxybutane that is not discussed
5-7
-------
TABLE 5-1. SCE IN MOUSE BONE MARROW CELLS AFTER IN VIVO EXPOSURE
TO 3,4-EPOXYBUTENE
Treatment
Control
Corn oil
Cyclophosphamide
(15 mg/kg)
3,4-Epoxybutene
10 mg/kg
25 mg/kg
50 mg/kg
100 mg/kg
150 mg/kgb
Number
of animals
4
4
4
4
4
4
3
1
SCE/cella
4.65 ± 1.250
4.43 ± 0.499
28.65 ± 3.496
6.68 ± 0.737
17.85 ± 3.793
18.78 ± 4.117
36.83 ± 6.585
47.63 ± 8.42
aMean ± SCE of 3-4 animals/group, except at 150 mg/kg where three out of four
animals died due to toxicity; 30 cells per animal were assayed for SCE.
^Inhibition of cell cycling occurred.
SOURCE: Allen, personal communication, 1985.
5-8
-------
TABLE 5-2. CHROMOSOME ABERRATIONS IN MOUSE BONE MARROW
AFTER IN VIVO EXPOSURE TO 3,4-EXPOXYBUTENE
Number
Dose of
Control
No treatment
BrdU only
BrdU + corn
oil
3,4-Epoxybutene
25 mg/kg
50 mg/kg
100 mg/kg
150 mg/kg
Cycl ophosphami de
.15 mg/kg
animals
4
4
4
4
4b
4
1C
(positive
4
Total
number
of cells
scored
400
400
400
400
400
400
100
control )
400
Gaps3
11
2
1
6
. 9
18
16
9
Total
aberrations
scored
1
3
3
12
20
49
20
14
Aberration
per cell
0.0025
0.0075
0.0075
0.0300
0.0500
0.1225
0.2000
0.0350
aChromatid and chromosome gaps are not included as aberrations,
"One mouse in this dose group was a mosaic (39/40).
cThree animals died before harvest.
SOURCE: Allen, personal communication, 1985.
5-9
-------
below, including testing in plants. This section on the genotoxicity of
diepoxybutane covers more recent information and provides an indication of the
types of genotoxic effects rather than an exhaustive survey of the literature.
5.4.1. Studies in Bacteria
Voogd et al. (1981) investigated the mutagenic potential of diepoxybutane
in IK. pneumoniae in the same paper cited previously for the mutagenicity of
3,4-epoxybutene. The source of the diepoxybutane was Merck (Darmstadt, F.R.6.);
it was of analytical grade and was not further purified. At an equal chemical
concentration (1 mM), diepoxybutane was approximately 16 times as mutagenic as
3,4-epoxybutene. The quotients of observed and spontaneous rates of mutation
to streptomycin resistance for 0.05, 0.1, 0.2, 0.5, and 1 mM diepoxybutane were
1.7, 3.1, 6.2, 15.7, and 27, respectively. These results clearly indicate that
diepoxybutane is mutagenic in K_. pneumoniae and provide strong evidence of a
dose-related response as well.
Diepoxybutane is also mutagenic in the S^. typhimurium histidine reversion
assay (Wade et al., 1979). Plate incorporation assays were performed with
strains TA98 and TA100, and averages of two to five determinations were report-
ed. At 0.02, 0.10, and 0.50 mg of diepoxybutane per plate, there were 196,
325, and 663 revertant colonies per plate with strain TA100 and 32, 22, and 29
revertant colonies per plate with strain TA98. These results suggest that
diepoxybutane is a base-pair substitution mutagen in S^. typhimurium because it
produced a dose-related positive response in strain TA100. Although strain
TA100 is not specific for mutagens that induce base-pair substitutions, it
responds well to such mutagens, and the result in strain TA98, which detects
many frameshift mutagens, was negative.
5.4.2. Studies in Fungi
The mutagenic potential of diepoxybutane in the yeast Saccharomyces
5-10
-------
cerevisiae was studied by Olszewska and Kilbey (1975). They used a diploid
yeast strain that is homozygous for the ilv mutation and therefore requires
isoleucine and valine to grow. Kinetic studies of the induction of revertants
were carried out by treating cells for various times with 0.1 M diepoxybutane at
25°C. About 1.2 x 106 to 1.5 x 106 cells were plated per petri dish. Diepoxy-
butane induced an increase in ilv reversions with increasing time of exposure
up to 20 minutes. At 5, 10, 15, and 20 minutes, the number of ilv+ revertants
per 106 cells was about 3, 10, 20, and 38, respectively. Reversion of the ilv
mutation indicates that diepoxybutane induces point mutations in yeast.
Zaborowska et al. (1983) have shown that diepoxybutane induces mitotic
crossing-over and mitotic gene conversion in the SBTD and D7 strains of _S_.
cerevisiae. Stationary-phase cells were treated with 0.4% (vol/vol) diepoxy-
butane (Merck, purity not specified) at.30°C for 15, 30, and 45 minutes. The
results of these experiments are shown in Table 5-3. The frequency of mitotic
crossing-over in the SBTD strain was dose (exposure time) related. Dose-
dependence of mitotic crossing-over in strain D7 is less clear. Dose-related
increases in mitotic gene conversion were obtained in both strains. Taken
together, these results indicate that diepoxybutane is recombinogenic in yeast,
and recombinogenicity is an indication of DNA damage.
Luker and Kilbey (1982) reported that diepoxybutane causes point mutations
and multigenic deletions in Neurospora crassa. They developed a Neurospora
heterokaryon in which both point mutations and deletions can be detected by
the use of selective techniques. Point mutations were scored by reversion to
adenine independence. Deletions were detected by first assaying for resistance
to £-fluorophenylalanine (pFPA) and then testing for sensitivity to cycloheximide,
These two genes are closely linked on chromosome V.
Information on the source and purity of the sample of diepoxybutane studied
5-11
-------
TABLE 5-3. INDUCTION OF MITOTIC GENE CONVERSION AND MITOTIC CROSSING-OVER IN
SBTD AND 07 STRAINS OF
S. cerevisiae BY DIEPOXYBUTANE
Exposure to 0.4%
diepoxybutane (min)
SB TO strain
0
15
30
45
07 strain
0
15
30
Survival
100
100
93.2
41.3
100
100
78.4
Mi t otic
crossovers (%)
0
1.4
2.8
6.1
0
1.5
1.7
Convertants3
0.7
86.6
167.1
515.2
0
40.4
278.0
aConvertants calculated per 107 survivors in the SBTD strain and per 106
survivors in the D7 strain.
SOURCE: Zaborowska et a!., 1983.
5-12
-------
was not provided. Suspensions of Neurospora conidia were treated for various
times with 0.1 M diepoxybutane in 0.067 M phosphate buffer (pH 7.0) at 27°C.
The treatments were terminated by filtering off the mutagenic solution and
washing the cells with 10% sodium thiosulfate solution. Diepoxybutane induced
dose (exposure time) related increases in both adenine reversions and pFPA
resistance, as shown in Table 5-4. The results shown in Table 5-5 suggest that
about one-fourth of the pFPA-resistant mutants were deletions rather than
point mutations because pFPA resistance was associated with sensitivity to
" f
cycloheximide 26.3% of the time. The evidence therefore shows that diepoxybu-
tane induces both point mutations and multigenie deletions in Neurospora.
5.4.3. Studies in Mammalian Cells
Dean and Hodson-Walker (1979) tested diepoxybutane for the induction of
chromosomal aberrations in cultured rat liver epithelial-like cells. The
sample of diepoxybutane studied was obtained from Fluka A.G., Switzerland.
Its purity was not described. The epithelial-like cell line used, designated
RLi, is neardiploid, having a chromosome number of 44 or 45 (compared to the
normal number of 42 in the rat karyotype). The appropriate concentrations of
diepoxybutane for testing were determined from cytotoxicity studies. Because
diepoxybutane is volatile, sealed flask cultures of the rat liver cells were
used. The cell cultures were exposed to diepoxybutane for 24 hours, and colcemid
at 0.3 pg/mL was added 2 hours before harvesting the cells with 0.25% trypsin.
Distilled water was added to the harvested cells to produce hypotonic conditions.
The cells were fixed in methanol/acetic acid (3:1). Chromosome preparations
were made by air-drying the cells on microscope slides and staining with Giemsa.
The slides were randomly coded, and 100 metaphases from each slide were analyzed
for structural chromosome changes. The results (Table 5-6) suggest that diepoxy-
butane is clastogenic in rat liver cells, producing significant chromatid
5-13
-------
TABLE 5-4. INDUCTION.OF ADENINE,REVERSIONS AND pFPA RESISTANCE IN N. crassa
BY 0.1 M DIEPOXYBUTANE
Treatment Survival
(min) .., - (%) .."
0 100.0
10 78.1
20 60.4
30 "40.8
Heterokaryotic
-> .conidia screened
(x 106) for
pFPA
:ad+i r. resistant
" 15.64 4.17
6.90 1.38
"5.54~ ''"i."l3
2.53 0.51
Number of
mutants
scored
ad-
5
22
34
53
fpr
2
12
21
21
Mutation
frequency
x 10-6
acT fprr
0.32 0.48
3.19 8.70
6.03 18.58
20.95 41.18
TABLE 5-5. ANALYSIS OF pFPAr MUTANTS AS PUTATIVE DELETIONS
Treatment Number of
with 0.1 M ,pFPA-resistant Number acquiring
diepoxybutane mutants sensitivity to
(min) .tested , . ,. ; .cycloheximide
10
15
20
30
45
60
Total
12
7
41
94
25
11
190
5
2
10
21
10
2
50
Putative
deletions
(%)
42
29
24
22
40
18
26.3
SOURCE: Luker and Kilbey, 1982.
5-14
-------
TABLE 5-6. EFFECTS OF DIEPOXYBUTANE ON CHROMOSOMES OF RAT LIVER CELLS
Diepoxybutane
0
0.1
0.5
1.0
Number
of cells
analyzed
269
117
24
29
Chromatid
gaps
2.6
11.1
8.3
0
Chromatid
deletions
0.4
4.3
20.8
0
Chromatid
exchanges
0
8.5
33.3
3.4
Chromosome
aberrations
0.4
0.9
0
0
SOURCE: Dean and Hodson-Walker, 1979.
5-15
-------
lesions at 0.1 ug/mL. At the higher concentrations, fewer metaphases were
apparently available for cytogenetic analysis, perhaps because of cytotoxic
effects at these doses.
Perry and Evans (197b) reported that exposure of cultured Chinese hamster
ovary cells to diepoxybutane resulted in a dose-related induction in SCEs.
The cells were treated with 10 yM bromodeoxyuridene and 0.3, yM or 3 yM
diepoxybutane for two cycles of DNA replication before treatment with colcemid
(2 hours at 0.2 VM), collection by mitotic shakeoff, pretreatment with 75 mM
KC1, and fixation in methanol/acetic acid (3:1). Twenty mitotic cells were
scored for each dose. SCEs/cell at 0, 0.3, and 3 PM diepoxybutane were 12.2,
20.2, and 90.9, respectively.
5.4.4. In vivo Studies
Studies of the mutagenic potential of diepoxybutane in whole animals have
been carried out in mice and Drosophila melanogaster (fruit flies).
Conner et al. (1983) studied the ability of diepoxybutane to induce j_n_
vivo SCE in bone marrow, alveolar macrophages, and regenerating liver cells in
mice. The sample of diepoxybutane studied was 97% pure and was obtained from
Aldrich Chemical Co. It was dissolved in phosphate-buffered saline just prior
to injection. The dose-response studies were performed in three intact and
three partially hepatectomized Swiss Webster mice (mean weight of 27 g).
Diepoxybutane (10-291 ymol/kg) was injected intraperitoneally just prior to
bromodeoxyuridine infusion (10 mg/mL; intravenous flow rate of 3.6 mL/24 hours).
Colchicine (3.3 mg/kg) was then injected intraperitoneally. Bone marrow and
alveolar macrophage cells from intact mice and regenerating liver cells from
hepatectomized mice were harvested 4 hours later and analyzed for SCEs. As
shown in Table 5-7, diepoxybutane produced similar dose-dependent responses for
SCE in bone marrow, alveolar macrophages, and regenerating liver cells. Conner
5-16
-------
TABLE 5-7. SCE FREQUENCIES INDUCED IN BONE MARROW, ALVEOLAR MACROPHAGES,
AND REGENERATING LIVER CELLS OF SWISS WEBSTER MICE
FOLLOWING INJECTION OF DIEPOXYBUTANE
Diepoxybutane
(umol/kg)
Hepatectomi zed
0
10
39
97
193
291
Intact mice
0
10
39
97
193
291
Bone marrow
Aa
mice
4.2 +_-0.6
7.2 +_ 0.9
8.1 +_ 1.4
10.9 +_ 1.5
22.3 +_ 3.5
32.0 +_ 6.6
3.0 _+ 0.8
5.4 +_-0.6
8.8 +_ 0.8
9.7 +_ 1.5
14.6 +_ 2.1
27.3 +_ 3.7
Bb
66
63
72
63
44
22
64
58
63
63
43
46
Alveolar
macrophages
A B
3.5 + 0.6
7.9 +_ 0.4
9.8 +_ 1.0
13.4 +_ 3.4
23.6 + 4.5
30.4 +_ 4.9
3.6 +_ 0.5
6.7 +_ 1.2
10.2 _+ 0.3
12.8 +_ 2.3
17.1 +_ 2.2
28.6 +_ 3.8
77
71
70
75
59
23
63
54
57
62
41
42
Regenerating
liver
A B
3.7 ^0.9 80
7.0 +_ 1.4 73
11.9 +_ 2.6 69
14.9 f 4.0 69
28.9^6.1 43
31.1 +_ 5.9 29
aMean SCE/cell +_ S.D. of three mice at each dose. Individual animal means
were calculated from SCEs scored in 20 cells of each type.
bMean number of second division cells observed in 100 consecutive metaphases,
SOURCE: Conner et al.5 1983.
5-17
-------
et al. (1983) also reported that the DNA lesions did not persist beyond one
cell cycle. These results show that diepoxybutane is very effective in produ-
cing a dose-dependent SCE response, but that the initially induced lesions
disappear in subsequent division cycles.
The sex-linked recessive lethal test in Drosophila, which tests the muta-
genic potential of chemicals in the germ line of an intact animal, was positive
for diepoxybutane in two studies from the same laboratory. In the first study,
Sankaranarayanan (1983) exposed wild-type Berlin-K adult 3- or 4-day-old male
flies to 2 mM diepoxybutane in 5% sucrose by feeding for 48 hours. The diepoxy-
butane sample studied was obtained from Fluka, A.G., Switzerland. Information
on its purity was not reported. The males were mated to Oster females to raise
three successive 2-day broods (A, B, and C), and the f\ female progeny were
used in the tests for lethals. Brood A tests mature spermatozoa, brood B tests
late spermatids, and brood C tests early spermatids. The results are shown in
Table 5-8. A concurrent no-exposure control was not done in this study or in
the subsequent study described below. However, a historical control value for
Drosophila of 0.18% has been established- in the same laboratory from the evalu-
ation of 13,151 chromosomes (Vogel, 1976). The results suggest that diepoxybu-
tane is a strong inducer of sex-linked recessive lethal mutations (6.5%, 4.8%,
and 4.6% compared to the historical control value of 0.18%). The results also
indicate that mature spermatozoa (brood A) respond with higher frequencies of
recessive lethals than late and early spermatids (broods B and C).
Similar positive results (Table 5-8) were obtained in a later study in
Drosophila (Sankaranarayanan et al., 1983). The experimental details in this
study were identical to those described above, except that Canton-S and ebony
males were exposed to diepoxybutane and mated to Muller-5 females. These
results demonstrate that the strong mutagenic response is independent of the
5-18
-------
TABLE 5-8. FREQUENCIES OF SEX-LINKED RECESSIVE LETHALS INDUCED BY 2 mM
DIEPOXYBUTANE IN POSTMEIOTIC MALE GERM CELLS OF D. melanogaster
Experimental
strain
Brood3
Number of
chromosomes
Lethals
Number
Berlin-Kb
Canton-Sc
Experiment 1
Experiment 2
Ebonyc
Experiment 1
Experiment 2
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
800
914
840
934
949
960
938
951
350
932
912
925
876
924
885
42
32
30
88
68
66
64
54
14
88
65
56
57
41
33
6.5
4.8
4.6
9.4
7.2
6.9
6.8
5.7
4.0
9.4
7.1
6.1
6.5
4.4
3.7
aBrood A corresponds to treatment of mature spermatozoa; Brood B corresponds
to late spermatids; and Brood C corresponds to early spermatids.
bjaken from Sankaranarayanan, 1983.
cTaken from Sankaranarayanan et al., 1983.
5-19
-------
strains employed. The strongly positive results in the sex-linked recessive
lethal test provide clear evidence that diepoxybutane reaches the gonads and is
strongly mutagenic in germ cells of Drosophila.
There is evidence that diepoxybutane induces chromosome damage in germ
cells of Drosophila (Zimmering, 1983). Treated males carried an X chromosome
in the form of a closed ring and a Y chromosome carrying dominant markers, one
at the end of long arm of the Y and one at the end of the short arm. The males
were permitted to feed on a solution of 1.25 mM diepoxybutane in 5% sucrose for
24 hours and then mated with repair-proficient (ordinary) females or to repair-
deficient females. There was no evidence of toxicity in the treated males.
The FI offspring were scored for complete loss of the X or Y chromosomes (in
ring-X males, virtually all complete loss is attributable to ring loss) and for
partial loss of the Y chromosome, indicated by the loss of one but not both of
the Y chromosome markers. Complete loss indicates chromosome breakage and/or
sister chromatid exchange. Partial loss of the Y chromosome is a consequence
of breakage. Results shown in Table 5-9 provide evidence of a relatively strong
effect on complete loss (5-6%) and a significant increase in partial loss which
is most apparent from matings with the repair-deficient females (approx. 3%).
In summary, the results of the Drosophila experiments assaying for sex-
linked recessive lethals and chromosome loss provide strong evidence that
diepoxybutane is a mutagen and a chromosome damaging agent in germ cells of
Drosophila. In addition, the results of the SCE assay in mice suggest that
diepoxybutane is a DNA damaging agent in mice.
The diepoxybutane was injected intraperitoneally in the mice and fed to
the Drosophila. It is possible that the diepoxybutane was biotransformed
before reaching mouse bone marrow or Drosophila gonads. Therefore, the muta-
genic, clastogenic, or DNA-damaging effects described in these studies may be
5-20
-------
TABLE 5-9. CHROMOSOME LOSS IN D. melanogaster
FROM MATINGS OF MALES WITH REPAIR-PROFICIENT (RP)
AND REPAIR-DEFICIENT (RD) FEMALES
Series Female
Control RP
Treated
Control RD
Treated
N
8551
7390
3178
1285
Complete
loss
51
515
30
82
Partial
loss
0
8
2
39
Percent induced
Complete Partial
loss loss
6.37 0.11
5.44 2.97
All induced frequencies are statistically significant at or below the 0.01 level
The repair-deficient mutant was mei-9a, which is deficient in excision repair.
SOURCE: Zimmering, 1983.
5-21
-------
due to a metabolite of diepoxybutane.
5.5. MUTAGENICITY OF 4-VINYL-l-CYCLOHEXENE AND ITS METABOLITES
Under certain conditions, such as during rubber curing, 1,3-butadiene can
dimerize to form 4-vinyl-l-cyclohexene (Rappaport and Fraser, 1976). Unpublished
data from the National Toxicology Program indicate that 4-vinyl-l-cyclohexene was
not mutagenic in the Salmonella preincubation assay (strains TA100, TA1535,
TA1537, and TA98) in the presence or absence of liver S9 mix prepared from
chemically induced rats or hamsters (NTP, 1985b).
In contrast with the negative mutagenicity response of 4-vinyl-l-cyclohexene,
its potential mono- and diepoxide metabolites (including 4-vinyl-l,2-epoxycyclo-
hexane, 4-epoxyethyl-l,2-dihydrocyclohexane, and 4-vinyl-l-cyclohexene diepoxide)
are mutagenic or clastogenic in various in vitro prokaryotic and eukaryotic
test systems (Murray and Cummins, 1979; Simmon and Baden, 1980; Turchi et al.,
1981; Voogd et al., 1981). These compounds are base-pair substitution mutagens,
in agreement with the data for 1,3-butadiene and its potential epoxide metabo-
lites.
5.6. SUMMARY OF MUTAGENICITY STUDIES
The available information on the mutagenicity of 1,3-butadiene is quite
limited in that only three studies have been reported. All three studies,
however, indicate that 1,3-butadiene is a mutagen in S_. .typhimurium. The
weight of the available evidence suggests that 1,3-butadiene is mutagenic only
in the presence of a liver S9 metabolic activation system. No whole-animal
studies have been reported. These results suggest that 1,3-butadiene is a pro-
mutagen in bacteria (i.e., its mutagenicity depends on metabolic activation).
In vitro data suggest that 1,3-butadiene is metabolized to 3,4-epoxybu-
tene and then to diepoxybutane. Preliminary evidence in rats suggests that
1,3-butadiene is metabolized to 3,4-epoxybutene in vivo, indicating that the
5-22
-------
metabolic pathway outlined on the basis of in vitro data may occur in vivo.
There is no information on the metabolism of 1,3-butadiene in humans.
However, a scientific basis for the extrapolation from animal metabolism data
to humans for 1,3-butadiene is provided by the similarities in the epoxidation
of the isolated double bond in benzo(a)pyrene by organ and tissue cultures from
animal and human sources (Autrup et al., 1980).
3,4-Epoxybutene is a monofunctional alkylating agent, is a direct-acting
mutagen in bacteria (S_. typhimurium, IK. pneumoniae, and E_. coli), and induces
SCE and chromosomal aberrations in mice. Diepoxybutane is a bifunctional
alkylating agent, and as such it can form cross-links between the two strands
of DNA. It is mutagenic in bacteria (K_. pneumoniae and S_. typhimurium), fungi
(yeast and Neurospora), and the germ cells of Drosophila. It also induces DNA
damage in cultured hamster cells and in mice, is clastogenic in fungi and cul-
tured rat cells, and produces chromosome damage/breakage in Drosophila germ
cells. Therefore, the evidence indicates that 3,4-epoxybutene and diepoxybu-
tane are mutagens/clastogens in microbes and animals.
Exposure of rodents to 1,3-butadiene results in ovarian tumors in mice
(Huff et al., 1985) and testicular tumors in rats (Hazleton Laboratories
Europe, Ltd., 1981a), which offers suggestive (not sufficient) evidence that
1,3-butadiene (or, more likely, a metabolite of 1,3-butadiene) may reach the
germ cells. There is also evidence that the dimer of 1,3-butadiene, 4-vinyl-
1-cyclohexene, causes ovarian tumors in mice (NTP, 1985b).
The total body of evidence from the metabolism, mutagenicity, and carcino-
genicity data suggests that 1,3-butadiene may present a genetic risk to humans.
However, mutagenicity studies in mammalian test systems, as outlined in the U.S.
EPA's Proposed Guidelines for Mutagenicity Risk Assessment (1984a), should be
conducted to further characterize the mutagenic potential of 1,3-butadiene.
5-23
-------
-------
6. CARCINOGENICITY
The purpose of this chapter is to provide an evaluation of the likelihood
that 1,3-butadiene is a human carcinogen and, on the assumption that it is a
human carcinogen, to provide a basis for estimating its public health impact
and evaluating its potency in relation to other carcinogens. The evaluation of
carcinogenicity depends heavily on animal bioassays and epidemiologic evidence.
However, additional factors, including mutagenicity, pharmacokinetics, and
other toxicological characteristics have an important bearing on both the
qualitative and quantitative assessment of carcinogenicity. This section pre-
sents an evaluation of the animal bioassays, the epidemiologic evidence, and
the qualitative and quantitative aspects of risk assessment.
6.1. ANIMAL STUDIES
6.1.1. Chronic Toxicity and Carcinogenicity Studies in Mice
A chronic toxicity and carcinogenicity inhalation study of 1,3-butadiene in
B6C3F1 mice, sponsored by the NTP, was conducted at Battelle Pacific Northwest
Laboratories. Preliminary inhalation toxicity studies in mice were used as a
basis for dose selection for a chronic study. A 15-day study and a 14-week
study were conducted at International Bio-Test Laboratories. In the 15-day
study, weight loss at airborne concentrations of 1,250 ppm was observed. The
mice exposed to 8,000 ppm, the highest airborne concentration, survived the
exposure period. In the 14-week study, reduced body weight and death were
observed among mice treated at 2,500 ppm or more. Necropsy findings were not
reported (NTP, 1984). The changes in the 14-week study at 2,500 ppm indicated
that 1,250 ppm is probably a maximum tolerated dose.
The mice used in the chronic study were obtained from Charles River
Laboratories and were exposed to graded concentrations of 625 and 1,250 ppm
6-1
-------
of 1,3-butadiene for 6 hours/day, 5 days/week. The exposures were conducted in
dynamic negative-pressure exposure chambers, and the chamber concentrations
were generated by mixing the test gas with filtered air. The chamber concen-
trations were measured 7 to 12 times a day for the first 150 days with a photo-
ionization detector, and thereafter by gas chromatography. It was intended
that the dimer (4-vinyl-l-cyclohexene) concentration in the test material be-
fore use should be controlled to less than 100 ppm. However, three cylinders
with slightly more than 100 ppm of dimer were used because replacements were
not available. The mice were 8 to 9 weeks of age when the exposures began, and
were housed individually throughout the study. There were 50 mice per sex per
dose group.
The mice were weighed weekly for the first 12 weeks of the study, and
monthly thereafter. They were examined for subcutaneous masses beginning after
the 12th week. Clinical signs were recorded weekly. Histopathological evalua-
tion (32 tissues) was performed on all mice.
While the original plan was for this to be a 2-year study, all surviving
mice were killed after week 60 to 61 because of excessive deaths among the
treated mice. Many of these deaths were caused primarily by the developing
neoplasia. The survival (mice at risk; corrected for mice that were missing or
were accidentally killed) at this early termination was as follows:
Airborne concentration (ppm)
0 625 1.250
Males
Fema1es
49/49
46/46
11/50
14/47
7/46
30/48
There were no increases in clinical signs that could be associated with
exposure to 1,3-butadiene except those related to tumor development and death.
6-2
-------
The body weights were not affected by inhalation exposure to the test chemical,
There was a marked increase in the overall frequency of mice with primary
tumors, as indicated below:
Airborne concentration (ppm)
0 625 1,250
Males
Females
10/50
6/50
44/50
40/50
40/50
46/49
In addition to a marked increase in the number of animals with primary
tumors, there was also an increase in the number of animals with multiple
primary tumors. Among the tumor-bearing male mice, there were 11,,73, and 61
such tumors in the control, low-, and high-exposure groups, respectively. In
the females, there were 6, 66, and 100 tumors in the tumor-bearing animals of
the control, low-, and high-exposure groups, respectively.
The histopathologic evaluation indicated significant increases in tumors
of various types, as shown in Table 6-1. These tumors began to appear remark-
ably early in the course of the study. Lymphomas were diagnosed in mice dead
at 22 and 20 weeks of exposure for males and females, respectively, of the high-
exposure group. The first tumors of this type were found in low-dose mice at 24
and 29 weeks, respectively. Of the survivors, two males in the low-dose group
and one in the high-dose group had lymphomas. In the females, one lymphoma
was found among the surviving control mice, three in the low-dose group, and
one in the high-dose group. Many of the early deaths were judged to be caused
by this type of tumor. While this tumor type is sometimes associated with
immunosuppression, no evidence of immunosuppression was reported in the histo-
pathologic evaluation of the lymphoid and hematopoietic tissues.
6-3
-------
TABLE 6-1. SUMMARY OF THE STATISTICALLY SIGNIFICANT INCIDENCE OF TUMORS
IN MICE EXPOSED FOR 60-61 WEEKS TO 1,3-BUTADIENE
Tumor type and site
Hemangiosarcomas
(heart)
Malignant lymphomas
(hematopoietic system)
Al veol ar/bronchi ol ar
adenoma
adenoma/carci noma
Acinar cell carcinoma
(mammary)
Granulosa cell tumor
or carcinoma (ovary)
Forestomach
(All papilloma
and carcinoma)
Hepatocellular
adenoma
adenoma/carci noma
Sex
M
F
M
F
M
F
M
F
F
F
M
F
F
F
Airborne
oa
0/50
p=0.032
0/49
p=0.001
0/50
p<0.001
1/50
p<0.006
2/50
p=0.10
3/49
p<0.001
2/50
p<0.001
3/49
p<0.001
0/50
p=0.007
0/49
p<0.001
0/49
p=0.354
0/49
p<0.001
0/50
p=0.025
0/50
p=0.016
concentration
625b
16/49
p<0.001
11/48
p<0.001
23/50
p=0.001
10/49
p=0.003
12/49
p=0.003
9/48
p=0.056
14/49
p<0.001
12/48
p=0.010
2/49
p=0.242
6/45
p=0.010
7/40
p=0.037
5/42
p=0.018
1/47
p=0.485
2/47
p=0.232
(ppm)
l,250b
7/49
p=0.006
18/49
p<0.001
29/50
p=0.001
10/49
p=0.003
11/49
p=0.007
20/49
p<0.001
15/49
p<0.001
23/49
p<0.001
6/49
p=0.012
13/48
p<0.001
1/44
p=0.473
10/49
p<0.001
4/49
p=0.056
5/49
p=0.015
aThe p-values were calculated with the Cochran-Armitage Trend Test.
t>The p-values were calculated with the Fisher Exact Test.
6-4
-------
The heart was the principal organ in which hemangiosarcomas occurred. The
first hemangiosarcomas were diagnosed at.32 and 42 weeks in the low- and high-
dose males and at 41 and 43 weeks among the females. The cardiac hemangiosar-
comas may have caused some of the mice to die early. Atypical cardiac endo-
thelial hyperplasia, a likely preneoplastic lesion, was not observed among the
controls but was present in treated males (625 ppm - 10%, 1,250 ppm - 4%) and
females (625 ppm - 10%, 1,250 ppm - 16%).
Alveolar/bronchiolar adenomas and carcinomas occurred (both separately and
combined) at increased frequency in both male and female mice. In the high-
dose groups the first such lesions appeared at week 42 for males and week 50
for females. Neoplastic changes in the lungs of the controls were not detected
until the termination of the study.
Among the 10 control male mice with primary tumors, eight had hepatocell-
ular adenomas and/or carcinomas. This type of tumor is normally observed among
male mice of this strain in 2-year bioassays. It may be that the inclusion of
mice with this type of tumor in considering the number of tumor-bearing animals
tends to deemphasize the frequency of compound-induced neoplasia. On the other
hand, among the females, the frequency of hepatocellular adenomas and/or carci-
nomas was increased. The occurrence of this type of tumor among females of
this strain is more suggestive of adverse chemical-related effects. Among the
male mice, there was a significant increase in liver necrosis at both doses.
In the female mice, liver necrosis was significantly elevated only at the
higher airborne concentration.
In addition to the neoplastic changes in the ovary and forestomach,
ovarian atrophy and forestomach epithelial hyperplasia were elevated among
the mice at both doses. Since Zymbal gland tumors have been reported in the
chronic rat study to be discussed, it is worth noting that in this study, one
6-5
-------
also occurred in high-dose female mice and two occurred in high-dose male mice.
This tumor is not normally found in control mice, even at the termination of a
2-year study. Testicular atrophy was observed in male mice at both dose levels;
however, the increase in tumors of the testes that had occurred in the rats did
not occur in the mice.
An audit of this chronic study was conducted by the NTP. Potential dis-
crepancies that might have significantly influenced the interpretation of this
study were resolved. It should be rioted that in 1981 some genetic variation
was observed in the male C3H parents of B6C3F1 mice used as test animals. NTP
(1984) has noted that the effect of genetic nonuniformity in the hybrid mice on
the results is unknown, but that the results were valid because of the use of
matched concurrent controls. The NTP considered that this study provided clear
evidence of carcinogenicity, which is the highest classification in their
system of categorizing evidence of carcinogenicity.
6.1.2. Chronic Toxicity Studies in Rats
A 2-year chronic inhalation toxicity study using rats as the experimental
animals was conducted by Hazleton Laboratories Europe, Ltd. (1981a) in England.
The study was sponsored by the International Institute of Synthetic Rubber
Producers, Inc. As previously discussed, the chronic study was preceded by
a 3-month toxicity study. The highest concentration used in that study was
8,000 ppm, which produced only minimal signs of toxicity and moderate
salivation in the female rats. Thus, the highest dose in the chronic rat
inhalation study was not established as a maximum tolerated dose. In addition,
no explanation of the eightfold difference between the low and high dose was
offered.
For the chronic investigation (Hazleton Laboratories Europe, Ltd., 1981a),
Charles River CD rats (Sprague-Dawley rats obtained from Charles River Ltd.)
6-6
-------
were exposed to graded concentrations of 1,000 and 8,000 ppm of 1,3-butadiene.
The exposures (6 hours/day, 5 days/week) for 111 and 105 weeks for males and
females, respectively, were conducted in a dynamic negative-pressure exposure
chamber. The chamber concentrations were generated by mixing the test gas with
filtered air. The concencentrations were measured with an infrared gas analy-
zer. The dimer (4-vinyl-l-cyclohexene) concentrations in the test material
before use were less than 1,000 ppm, but samples in the 700 to 800 ppm range
were used, and the dimer concentration of these averaged 413 ppm. The rats
were 4 1/2 weeks of age when the exposures began, and were housed five to a
cage throughout the study. There were 110 rats per sex per dose group, and a
similar number of rats exposed to filtered air served as a control group.
The rats were weighed weekly and examined for subcutaneous masses and
other clinical signs. Blood chemistries, hemograms, and urine analyses were
evaluated at 3, 6, and 12 months. Neuromuscular function was evaluated period-
ically through week 77 of the study. Ten rats per sex per dose were killed and
necropsied at week 52. Histopathological evaluations were performed on all
rats from the high-dose group and the control group. Tissues from the low-dose
group that were deemed to be of toxicological significance were also examined.
Variations in mean body weight suggested no consistent adverse effect.
Review of the hemograms, blood chemistry, urine analysis, and behavioral test-
ing was likewise not indicative of an adverse effect.
In the females of the treated groups, subcutaneous masses appeared earlier
and at a higher incidence than in the control group. A dose-related increase
in liver weights was observed at the necropsy performed at 52 weeks and at the
termination of the study. This could indicate that the chemical induces liver
enzymes. Otherwise, no significant changes were noted at the 52-week kill.
In the control group, 45% of the males and 46% of the females survived
6-7
-------
until the end of the study (note: corrected for interim kill). In the high-
dose group, survival was 32% and 24%, while in the low-dose group 50% and 32%
survived. The decreased survival in the high-dose group was statistically
significant.
Increased alveolar metaplasia and nephropathy were observed among males of
the 8,000-ppm treatment groups at the termination of the study. Marked or se-
vere nephropathy occurred in 27% of the male rats in the high-dose group, as
compared with 9 to 10% in the control and the low-dose groups. The authors con-
sidered nephropathy to be the cause of some of the early deaths in this study.
The frequency of metaplasia was 5/44 in the surviving male rats (8,000 ppm) as
compared to 5/45 in the controls.
With regard to the carcinogenic potential of 1,3-butadiene, the authors
of this study concluded that exposure of male and female rats under the
conditions of this investigation was associated with significant increases in
both common and uncommon tumors. Furthermore, they stated that the results of
this 2-year inhalation study supported the premise that 1,3-butadiene is a
suspect weak oncogen.
The incidence of selected neoplasms is shown in Table 6-2. In the females
there was an increase in mammary carcinoma tumors (control - 8%, 1,000 ppm - 42%,
8,000 ppm - 38%). Also, in the females follicular thyroid adenomas were encoun-
tered more frequently among the treated females than among the controls (control
- 0%, 1,000 ppm - 2%, 8,000 ppm - 8%).
In the males there was an increase in Leydig cell adenomas (control - 0%,
1,000 ppm - 2%, 8,000 ppm - 7%). A single Leydig cell tumor (unspecified) was
observed in one male of each exposed group. Exocrine pancreatic adenomas were
increased in the male rats of the high-dose group (control - 3%, 1,000 ppm - 1%,
8,000 ppm - 10%). One carcinoma was observed in this tissue in the males of the
6-8
-------
TABLE 6-2. SUMMARY OF THE INCIDENCES OF TUMORS IN RATS
EXPOSED TO 1,3-BUTADIENE (100 RATS PER SEX PER DOSE GROUP)
Tumor type and site
Multiple mammary
gland tumors
Thyroid follicular
(adenoma and carcinoma)
Uterine cervical/
stromal sarcoma
Leydig cell
(adenoma and carcinoma)
Pancreatic exocrine
Carcinoma
Adenoma
Zymbal gland
(carci noma)
Sex
F
F
F
M
M
M
M
F
Airborne
Ob
50
p<0.001
0
p<0.001
1
p=0.115
0
p<0.003
0
3
p=0.019
0
p=0.384
0
p=0.037
concentration (ppm)
1,000C
79
p<0.001
4
p=0.06
4
p=0.184
3
p=0.12
0
1
p=0.879
1
p=0.5
0
8,OOOC
84
p<0.001
11
p<0.001
5
p=0.106
8
p<0.001
1
10
p=0.041
1
p=0.5
4
p=0.061
aComplete information on the number of animals examined was not available.
In calculating incidences, it is assumed that 100 animals per sex per group
survived the first year and were histologically examined.
bThe p-values were calculated using the Cochran-Armitage Trend Test.
cThe p-values were calculated using the Fisher Exact Test.
6-9
-------
high-dose group.
Zymbal gland tumors were increased in the high-dose group when male and
female rats were combined. However, the increase in Zymbal gland tumor inci-
dence is based on the assumption that all glands were examined, and this is not
likely to have been the case. While not statistically significant, there were
four gliomas among the high-level males versus one in the controls. Gliomas
were present among the exposed mice. Thus, the incidence of Zymbal gland
tumors could have been higher than reported. Except for Zymbal gland tumors,
the increase in tumors in this investigation was limited to those developed in
hormonal-dependent tissues.
The results of the Hazleton (1981a) study have not been published, and the
final unpublished report does not include detailed individual histopathological
evaluations. While the U.S. EPA's Carcinogen Assessment Group has requested
these particular data, they have not been made available. Furthermore, because
the report is unpublished, no independent data quality evaluation was performed,
and requested data were not available (July 1, 1985). Qualitatively, this
investigation shows a positive finding of the carcinogenicity of 1,3-butadiene
in the rat. Quantitatively, however, the uncertainty about the numbers of tissues
actually examined severely limits its usefulness in animal-to-human risk extra-
polation (see section 6.3.).
6.1.3. Carcinogenicity of Related Compounds
Recently, a draft report from the NTP (1985b) on the toxicology and carcino-
genicity of 4-vinyl-l-cyclohexene, a dimer of 1,3-butadiene encountered in
the offgassing during tire curing, has become available. The test material
was administered in corn oil by gavage 5 days/week for 103 weeks to groups of
50 F344/N rats and B6C3F1 mice of each sex. The doses were 200 and 400 mg/kg,
and corn oil alone served as a vehicle treatment for control (0 ppm) rats. The
6-10
-------
doses were selected on the basis of 14-day and 13-week studies. The end points
of survival, body weight gains, and histopathologic effects were the basis
for determination of 400 mg/kg as a maximum tolerated dose.
The NTP concluded that there was clear evidence of carcinogenicity of 4-
vinyl-1-cyclohexene in female mice as shown by increased ovarian neoplasms at
both doses. The incidences of mixed benign tumors among the female mice were
0/49, 25/48 (52%), and 11/47 (23%) in the control, 200, and 400 mg/kg groups,
respectively. The incidence of granulosa cell tumor or carcinoma combined
was 1/49 (2%), 10/48 (21%), and 13/47 (13%) for the 0, 200, and 400 mg/kg
treatments, respectively. In addition, the increased incidence (vehicle control,
0/50; 200 mg/kg, 3/49; and 400 mg/kg, 4/48) of adrenal gland adenomas was
judged to be probably related to the test material. With regard to the effects
among the male mice, it was concluded that the evidence for carcinogenicity
was equivocal based on marginal increases of malignant lymphomas (vehicle ,
control, 3/37; 200 mg/kg, 5/39; 400 mg/kg, 4/7) and alveolar/bronchiolar adenomas
combined (vehicle control, 3/37; 200 mg/kg, 5/39; and 400 mg/kg, 3/7). The
sensitivity for detecting the carcinogenic response may have been limited,
however, by the poor survival at the high dose. The response in the male rats
was considered to provide inadequate evidence of carcinogenicity, at least in
part because of excessive mortality. Among the female rats, the increased
incidence of adenomas or squamous cell carcinomas (combined) of the clitoral
gland at the low dose (vehicle control, 1/50; 200 mg/kg, 5/50) despite excessive
mortality at the high dose, was considered to provide equivocal evidence of
carcinogenicity (NTP, 1985b).
Van Duuren et al. (1963) applied epoxybutene (butadiene monoxide), dl-1,2;
3,4-diepoxybutane and meso-l,2;3,4-diepoxybutane three times a week for life to
the skin of 30 male Swiss-Mi 11erton mice. The diepoxybutanes were in 10%
6-11
-------
acetone. The monoxide was applied undiluted. Ninety mice served as solvent
controls, and 207 mice were used as untreated controls. While the combined
tumor incidence (papillomas and squamous cell cancer) was elevated only for
meso-l,2;3,4-diepoxybutane as compared to the controls (20% vs. 8.9%), the
cancer incidence was higher for all three compounds as compared to the controls
(butadiene monoxide, 3.3% vs. 0.5%; dl-l,2;3,4-diepoxybutane, 3.3% vs. 0%;
and meso-l,2;3,4-diepoxybutane, 13.3% vs. 0%).
In another study, these investigators (Van Duuren et al., 1966) injected
dl-l,2;3,4-diepoxybutane in tricaprylin once a week subcutaneously into female
Swiss-Millerton mice at doses of 0.1 and 1.1 mg per mouse. No local sarcomas
were observed in the controls, while 5/50 and 5/30 mice at the low and high
doses, respectively, developed fibrosarcomas. In the same investigation,
female Sprague-Dawley rats were injected intraperitoneally with 1 mg of dl-1,2;
3,4-diepoxybutane in tricaprylin once a week, and 9/50 developed local fibrosar-
comas.
6.1.4. Discussion of Carcinogenicity Studies
The two chronic studies available at this time are compared in Table 6-3.
There is an obvious difference between the mice and the rats with regard to
carcinogenic response. The mice might be expected to respond more than rats
for several reasons. For example, if the carcinogenic response is elicited by
a metabolite, as has been suggested (de Meester et al., 1978, 1980), mice, be-
cause of their higher rate of metabolism, might be expected to yield a greater
response than rats. Furthermore, some of the rats were exposed at the airborne
concentrations expected to produce metabolic saturation. In addition, the
mice could have had less activity because of group housing, thereby reducing
respiration rate.
6-12
-------
TABLE 6-3. SIGNIFICANT'EFFECTS OF EXPOSURE TO 1.3-BUTADIENE ON SPRAGUE-DAWLEY RATS
AND B6C3F1 MICE IN INHALATION STUDIES
Rats3 (Hazleton Laboratories Europe, Ltd., 1981a)
1,000 ppm 8,000 ppm
Neoplasms:
Hales
Females
Leydig cell adenomab
Mammary gland:
f i broadenoma/card nomab
Thyroid: follicular cell
adenoma13
Uterus: stromal sarcomab
Nonneoplastic lesions:
Males
Leydig cell adenomab
Pancreas: exocrine tumors'3
Brain: glioma
Mammary gland:fibroadenoma/
carcinoma'3
Thyro1d:follicular cell
adenoma*3
Uterus: stromal sarcomab
Zymbal gland: carcinoma13
Increased focal alveolar .
epltheliallzation
Nephropathy
Micec (National Toxicology Program, 1984)
625 ppm
1,250 ppm
Neoplasms:
Males
Females
Heart: heraang1osarcomab
Malignant lymphomab
Lung: alveolar/bronchiolar
adenoma and carcinoma13
Forestomach: papillomab
Preputial gland:
squamous cell carcinoma''
Brain: gliomad
Heart: hemangiosarcomab
Malignant lymphomab
Lung: alveolar/bronchiolar
adenoma and carcinoroab
Forestomach: papillomab
Ovary: granulosa cell
tumorb
Nonneoplastic lesions:
Males Forestoroach: epithelial
hyperplasia"3
Liver necrosisb
Testicular atrophyb
Females
Liver necrosis'3
Forestomach: epithelial
hyperplas1ab
Ovary: atrophyb
Uterus: involutionb
Heart: hemangiosarcomab
Malignant lymphoroab
Lung: alveolar/bronchiolar
.adenoma and carcinoma13
Preputial gland:
squamous cell carcinomad
Zymbal gland: carcinomad
Brain: gliomad
Heart: hemangiosarcomab
Malignant lymphomab
Lung: alveolar/bronchiolar
adenoma and carcinoma13
Forestomach: pap1llomab
Mammary gland: acinar cell
carcinoma13
Ovary granulosa cell tumorb
Liver: hepatocellular adenoma
or carcinoma (combined)13
Forestomach: epithelial
hyperplasiab
Liver necrosis'3
Nasal cavity lesions (chronic
inflammation, fibrosls, car-
tilaginous metaplasia, osse-
ous metaplasia, atrophy of
sensory epithelium13
Testicular atrophyb
Forestomach: epithelial
hyperplasiab
Ovary: atrophyb
uterus: involution13
aGroups of 100 male and female Sprague-Dawley rats were exposed to air contain-
ing 0, 1,000 or 8,000 ppm 1,3-butadiene 6 hours/day, 5 days/week for
105 weeks (female), or 111 weeks (male); survival in dosed groups decreased.
bStatistically significant (p < 0.05).
cGroups of 50 male and female B6C3F1 mice were exposed to air containing 0, 625
or 1,250 ppm 1,3-butadiene 6 hours/day, 5 days/week for 60 weeks (male) or 61
weeks (female); survival in dosed groups decreased and was the reason for early
termination.
^Considered uncommon at 60 weeks.
SOURCE: National Toxicology Program, 1984.
6-13
-------
The higher dimer content of the material to which the rats were exposed
might also be expected to contribute to the difference in the effective dose of
1,3-butadiene. In addition, intralaboratory variations in dimer formation in
the test atmospheres might be a contributing factor.
The role that the presence of the dimer at higher concentrations in the
rat study might have played is unknown. The dimer could have inhibited tumor
response by inducing metabolizing enzymes, by competing at active sites, or by
inducing competing tumors. The recent draft NTP report (1985b) indicates that
in mice the incidence of granulosa cell tumors was elevated in the gavage
study with 4-vinyl-l-cyclohexene, as was observed in the mouse inhalation bio-
assay by NTP (1984). The dimer also induced a marginal increase in malignant
lymphomas, and increased alveolar/bronchiolar tumors were present in the mouse
inhalation study. The gavaged dimer induced tumors of the clitoral gland.
While not reported in the inhalation study, this is a hormonally sensitive
tumor, as were most of the tumors in the rat inhalation bioassay.
The tumors in the rats exposed to airborne concentrations of 1,3-butadiene
are largely characterized as occurring in hormonal-dependent tissues. Some
suggestion of this is observed in the mice, but not to so great an extent. The
occurrence of a similar array of tumors in the mice could have been masked
by the early deaths from more rapidly developing neoplasia. It is worth noting
that Zymbal gland tumors developed in both species, but were not as marked in
the mice. Likewise, brain gliomas were found in both test species.
In the NTP mouse inhalation study, hemangiosarcomas of the heart, a very
rare tumor, were markedly elevated in both groups exposed to 1,3-butadiene.
While other carcinogenic agents have been reported to induce hemangiosarcomas,
in most instances these tumors have been found in the spleen or liver.
6-14
-------
The positive response of the monoepoxybutene and diepoxybutane metabolites
after dermal, subcutaneous, and intraperitoneal administration, supplemented by
short-term tests and metabolic information, strongly indicates that 1,3-butadiene
acts as a carcinogen by means of metabolic activation.
In summary, the two positive rodent inhalation bioassays provide suffi-
cient evidence that 1,3-butadiene is an animal carcinogen. In addition, re-
lated compounds, the dimer and the metabolites, are positive in animal carci-
nogenic bioassays. This adds to the weight of evidence with regard to the
classification of 1,3-butadiene as a carcinogen.
6.2. EPIDEMIOLOGIC STUDIES
The manufacture of styrene-butadiene rubber (SBR) involves the use of, and
hence exposures to, several different chemicals. The two major components of
SBR polymers are styrene and butadiene. In a typical recipe for the production
of SBR, butadiene and styrene account for 26% and 9%, respectively, of the
total ingredients. It should be pointed out that water accounts for 63% of the
volume. At room temperature styrene is a clear, colorless liquid, while buta-
diene is a gas.
Two other agents, toluene and benzene, need to be considered, although they
are not used directly in the manufacture of SBR. Toluene exposures result from
its periodic use as a tank-cleaning agent; it may also exist as an impurity of
styrene. Benzene exposures may occur as an impurity of styrene or toluene.
Occupational epidemiologic studies investigating the potential health haz-
ards associated with the production of synthetic rubber have been very limited
in number. Because styrene and butadiene are the two basic materials used in
the manufacture of SBR, with benzene and toluene as byproducts, it is at best
difficult to assess the singular contribution of each. Benzene exposure has
been identified with excessive risk, particularly acute leukemia (Linet, 1985).
6-15
-------
Styrene also may be a leukemogen in humans (Ott et al., 1980; Lilis and Nichol-
son, 1976). Although many studies of rubber production workers have been con-
ducted, only a few of those studies are relevant to butadiene exposure. Those
studies, which are reviewed here, include studies of workers specifically iden-
tified as working in styrene-butadiene production or the manufacture of synthet-
ic rubber. A study was also included if it was a preliminary study to one in
which the workers were identified as SBR or synthetic rubber workers, or if it
added to the interpretation of one of the studies of SBR or synthetic rubber
workers.
6.2.1. McMichael et al. (1974. 1976)
In 1974, McMichael et al. identified, through company records, a historic
prospective cohort of 6,678 hourly paid male workers in a rubber tire manufac-
turing plant in Akron, Ohio. The cohort was composed of all active and retired
male employees aged 40 to 84 years as of January 1, 1964.
During the 9-year follow-up period from 1964 through 1972, 1,783 workers
died. Death certificates were obtained for 99.5% of these workers, and the
causes of death were coded according to the 8th revision of the International
Classification of Diseases (ICD), by a National Center for Health Statistics
nosologist.
SMRs were calculated for all males using the 1968 U.S. male population as
a standard population. SMRs for all causes for the active age range of 40-64
and the full age range of 40-84 were 93 and 99, respectively. In cause-specif-
ic SMRs, statistically significant excesses were observed for stomach cancer
(SMR = 219, observed = 12, expected = 5.5, p < 0.01), lymphosarcoma (SMR = 251,
observed = 6, expected = 2.4, p < 0.05), and leukemia (SMR = 315, observed =
11, expected = 3.5, p < 0.001), in the active age range of 40-64. For the full
age range of 40-84, significant SMR increases were observed for cancers of the
6-16
-------
stomach (SMR = 187, observed = 39, expected = 20.9, p < 0.001), prostate, (SMR
= 142, observed = 49, expected = 34.4, p < 0.05), lymphosarcoma (SMR = 226,
observed = 14, expected = 6.2, p < 0.01), diabetes mellitus (SMR = 143, observed
= 43, expected = 30, p < 0.05), and arteriosclerosis (SMR = 154, observed =
34, expected = 22.1, p < 0.05).
McMichael et al. (1976) attempted to evaluate the relationship of these
mortality excesses to specific jobs within this plant by designing a nested
case-control study. Out of a total of 1,983 deaths observed during the 10-year
follow-up period of 1964 through 1973, 455 individuals who had died from certain
specific causes were selected as cases. The specific causes of death included
stomach, colorectal, respiratory, prostate, and bladder cancers; cancers of the
lymphatic and hematopoietic systems; lymphatic leukemias; ischemic heart dis-
ease; and diabetes mellitus. Out of these 455 cases, 353 deaths were attrib-
uted to cancers and 102 to noncancer causes. From the remainder of the plant
population of male workers, an age-stratified random sample of 1,500 individuals,
with 500 individuals in each age group of 40-54, 55-64 and 65-84 was obtained.
Complete work histories were obtained for 99% (1,482) of this age-stratified
sample drawn as a control group.
As of January 1, 1964, the plant population of male workers had a racial
composition of 86% white and 14% black. Thirty-eight percent, 30%, and 32% were
in the 40-54, 55-64, and 65-84 age ranges, respectively. Forty-eight percent
had begun work in the plant at least 25 years prior to 1964, and 99% had worked
for at least 10 years by 1964.
Work exposure histories were restricted to the period from 1940 through
1960. Cumulative job exposures of less than 2 years were excluded from the
analyses. Because follow-up extended to 1972, the period between first exposure
and death could range from 12 to 32 years, which should allow for the observation
6-17
-------
of occupationally induced cancers in adults.
For each of the cause-specific mortality groups, as well as the group of
controls, rates of exposure for minimum duration of 2 years and 5 years were
calculated for each of 16 occupational title groups (OTGs) in order to ascer-
tain any dose-response relationships. The exposure rates in each case group
were age-adjusted by the direct method of adjustment to the age distribution of
the controls. For the nine cause-specific mortality groups, the ratios of
their age-adjusted job classification exposure rates to the rates within the
sample of controls were calculated in order to provide an approximation for
the more conventional odds ratios.
For all of the causes of death under investigation, there were statisti-
cally significant (p < 0.001) associations with many of the work areas in which
workers had had at least 5 years of exposure. In the synthetic plant area, the
significant (p < 0.001) risk ratios were 6.2 for lymphatic and hematopoietic
cancer, 3.9 for lymphatic leukemia, 3.0 for ischemic heart disease, and 2.2 for
stomach cancer. Among the various work areas, the risk ratios for lymphatic
leukemia and for lymphatic and hematopoietic cancer were the highest in the
synthetic plant. Spirtas (1976) reported, however, that the risk ratio for
lymphatic and hematopoietic cancer dropped from 6.2 to 2.4 when a smaller
matched control group was used; the statistical significance of the 2.4 risk
ratio was not indicated.
McMichael et al. (1976) reported that a case-control study (McMichael et
a!., 1975) had found an association between lymphatic leukemia and solvent expo-
sure in the rubber industry. Many of the lymphatic leukemia deaths were the
same as those reported in the McMichael et al. (1976) study. Further analysis
by Checkoway et al. (1984) of 11 of the lymphatic leukemia cases studied by Mc-
Michael et al. also found an association between lymphatic leukemia and solvent
6-18
-------
exposure. Spirtas (1976) reported that of the six deaths from neoplasms of
lymphatic and hematopoietic tissue of individuals who had worked in the synthetic
rubber area of the plant in the McMichael et al. (1976) study, three were due
to leukemia. Of these three individuals, two had had experience with solvents
other than in the synthetic plant. Thus the role of the transfer of individuals
from one work area into another needs to be investigated. Also, racial factors
could not be accounted for in exposure calculations because data on race were
not available for much of the study population at the time of sampling.
6.2.2. Andjelkovich et al. (1976, 1977)
The mortality experience during the period from January 1, 1964 through
December 31, 1973 of a historic prospective cohort of 8,938 male rubber workers
(known as the "1964 cohort") from a single plant located in Akron, Ohio, was
observed by Andjelkovich et al. in 1976. Any person who was 40 years of age
or more on January 1, 1964, and was an active or living retired hourly worker
from the plant under study, was included in the 1964 cohort.
Data were collected from company records, life insurance death claims,
and bureaus of vital statistics of several states. A trained nosologist coded
the causes of death according to the 8th revision of the ICD. Follow-up was
achieved for 96.7% of the cohort. Out of 8,938 males, 94% (8,418) were white
males and were equally distributed in the age groups 40-54, 55-64, and 65-84.
Although 6% of the cohort consisted of black males, the major analyses were
done on white males. During the 10-year observation period 2,373 (28%) of the
white males died. SMRs were calculated using the age-, race-, and sex-specific
rates of the 1968 U.S. population. SMRs for deaths from all causes in the
40-64, 65-84, and 40-84 age groups were, respectively, 92 (observed =619), 95
(observed = 1,754, p < 0.05) and 94, (observed = 2,373, p < 0.01).
Many cause-specific SMRs showed increases, but only two disease SMRs,
6-19
-------
those for neoplasms of lymphatic and hematopoietic tissue (SMR = 138, observed
«* 40) and chronic rheumatic heart disease (SMR = 170, observed 16) for the
age group 65-84 were statistically significant (p < 0.05). The only statisti-
cally significant (p < 0.05) SMR for the age group 40-64 was for cerebrovascular
disease (SMR = 138, observed = 48). On further detailed examination of neoplasms
of lymphatic and hematopoietic tissue, statistically significant excesses were
found for monocytic leukemia (SMR = 441, observed = 3, p < 0.01) and other neo-
plasms of lymphatic and hematopoietic tissue (SMR = 276, observed = 10, p <
0.001) in the age group 65-84. There were no deaths by either of these causes
in the age group 40-64.
An important finding of this study is that it found a high mortality rate
in workers who had retired between the ages of 40 and 64, the mandatory retire-
ment age being 65. The SMR for all causes for this group was 202, which is
highly statistically significant (observed = 299, p < 0.001). The SMR for
almost every cause analyzed was elevated, and out of 26 categories, 13 of them
were statistically significant. For malignant diseases, the authors found
significant elevations in SMRs for malignant neoplasms of the prostate (SMR =
278, observed = 4, p < 0.05); large intestine (SMR = 231, observed = 6, p <
0.05); trachea, bronchus, and lung (SMR = 241, observed = 28, p < 0.001);
and brain and central nervous system (SMR = 323, observed = 3, p < 0.05). In
non-malignant diseases, SMRs were statistically significant for 1) benign neo-
plasms and neoplasms of unspecified nature (SMR = 541, observed = 2, p < 0.01);
2) endocrine, nutritional, and metabolic diseases (SMR = 396, observed = 11,
p < 0.001); 3) diseases of the nervous system and sense organs (SMR = 577,
observed = 6, p < 0.001); 4) chronic rheumatic heart disease (SMR = 440, ob-
served = 7, p < 0.001); 5) ischemic heart disease (SMR = 180, observed = 112,
p < 0.001); 6) cerebrovascular disease (SMR = 258, observed = 22, p < 0.001);
6-20
-------
7) other respiratory disease (SMR = 309, observed = 5, p < 0.01); 8) diseases
of the digestive system (SMR = 357, observed = 16, p < 0.01); and 9) symptoms
and ill-defined conditions (SMR = 263, observed = 4, p < 0.05).
As opposed to these increases, the SMR for deaths from all causes for ac-
tive workers in the 40-64 age group was 61 (observed = 320), substantially
lower than the SMR of 202 for retired workers in this age group. The overall
SMR for active and retired workers combined was 92 (observed = 619), showing a
dilution effect by the active workers and confirming the "healthy worker"
effect. Some cause-specific SMRs were elevated slightly in this active group,
but none of them were statistically significant.
In an attempt to evaluate the relationship between the mortality increases
and occupational exposures, Andjelkovich et al. (1977) re-analyzed the same.'...
data for 28 work areas within the plant, under study in 1977. The OTG of each
person was decided on the basis of the most representative department (obtained
from personnel folders) in which the individual had worked. The closing date
for the active workers was December 31, 1973, while for the retired or termin-
ated workers the period of study was from the date of hire to the last date
worked.
All causes of death and cause-specific SMRs were calculated by using the
experience of the entire cohort as a reference group. Marginal increases in
SMRs for all causes were observed for many OTGs for all three of the age groups
considered: 40-64, 65-84, and 40-84 years. The only statistically significant
excess observed was for cast film manufacture (SMR = 230, observed = 7, p <
0.05) in age group 65-84. Statistically significant (p < 0.05) SMR deficits
from all causes were observed for OTGs: 1) product fabrication (tire and beads),
2) product fabrication (valves, tubes, and flaps), and 3) bulk chemicals and
metal products, for at least one or more of the three age groups.
6-21
-------
SMR increases were statistically significant for all neoplasms in the fol-
lowing four OTGs: 1) cast film manufacture for age group 40-84, 2) special
products manufacture for age group 40-84, 3) milling for age groups 65-84 and
40-84, and 4) miscellaneous for age groups 40-64 and 40-84. Out of these four
OTGs, the first three departments dealt with the manufacture of industrial
products.
For selected cancers, the SMRs were significantly (p < 0.05) elevated in
age group 40-84 in certain OTGs, namely, cancer of the stomach in compounding
and mixing (SMR = 479, observed = 3), and milling (SMR = 369, observed = 6);
cancer of the large intestine in special products (SMR = 629, observed =4);
cancers of the trachea, bronchus, and lung in synthetic latex (SMR = 434,
observed = 3); cancer of the prostate (SMR = 212, observed = 10); and leukemia
(SMR » 246, observed = 6) in general services. All SMRs, except the one for
the compounding and mixing department, showed statistically significant excesses
of deaths in more than one age group.
For non-malignant diseases, Andjelkovich et al. (1977) calculated signifi-
cantly (p < 0.05) elevated SMRs for diabetes mellitus, acute myocardial infarc-
tion, arteriosclerosis, and suicide in various OTGs for various age groups.
Significant excesses of deaths in the general service department and arterio-
sclerosis in the shipping and receiving department were observed in more than
one age group. Statistically significant deficits were observed for ischemic
heart disease in the industrial chemicals department and acute myocardial
infarction in the stock preparation department, both in the 40-84 age group.
Since the authors were aware that the job transfer patterns of the deceased
workers were not necessarily representative of the job transfer patterns of the
entire cohort, they used the available information on deceased workers to esti-
mate the length of time spent in the most representative department. A simple
6-22
-------
random sample of 50 deceased workers was chosen, and detailed work histories
were reviewed for them. The 50 workers had spent an average of 28.3 years in
the industry. On an average, each worker had spent 50% of his work time in
his most representative OTG. However, the fraction of time spent in the most
representative OTG ranged from less than 10% up to 100% of total employment
duration.
The Andjelkovich et al. studies have a number of limitations. First,
their use of the most representative departments should be questioned in view
of the fact that the people under study could have worked in these departments
from 10% to 100% of their total employment duration. The only elevated SMR
in synthetic latex was for cancer of the trachea, bronchus, and lung, based on
only three deaths, while there was no control for cigarette smoking, which is
a potential confounder. Another limitation was the use of 1968 mortality
data for trachea, bronchus, and lung cancer to calculate the expected number
of deaths. Mortality for trachea, bronchus, and lung cancer was rising sharply
during the period 1964 to 1973, the follow-up period of this study. The use of
1968 data may have underestimated the expected number of deaths and thus over-
estimated the SMR. Although a statistically significant excess of deaths for
cancers of lymphatic and hematopoietic tissue was observed for persons whose
most representative department was general services, this job category does not
necessarily involve contact with SBR production. With regard to the question-
able 'exposure, Taulbee et al. (1976) reported that in an analysis of the work
histories of 37 leukemia cases and four matched controls per case from the 1964
cohort of Andjelkovich et al., none of the cases was found to have worked in
the OTG "synthetic plant." Some of the cases had worked in departments in
which there may have been exposure to the synthetic process, but this associa-
tion was not statistically significant (p < 0.05), nor was the association
6-23
-------
found to increase by duration of exposure.
One positive aspect of the Andjelkovich et al. (1977) study is that the
entire cohort is used as a reference group, which should reduce the "healthy
worker effect" and allow for a more unbiased evaluation. It would have been
interesting, however, to see the comparison of SMRs calculated from the U.S.
population (1968), which was used as a standard population in the 1976 study,
with the SMRs calculated from the internal cohort as a reference group.
Both the McMichael et al. and the Andjelkovich et al. studies are sugges-
tive of some health problems in the synthetic plant, indicating that specific
exposure investigations should be undertaken.
6.2.3. Checkoway and Williams (1982)
Since the study of McMichael et al. (1976) indicated the potential pres-
ence of carcinogens in SBR plants, Checkoway and Williams conducted a combined
industrial hygiene and hematology cross-sectional survey at the same plant
studied by McMichael et al. The objectives of the Checkoway and Williams study
were to quantify workplace exposures to styrene, butadiene, benzene, and toluene,
and to relate exposure levels to hematologic measurements.
During the week of May 15-19, 1979, personal breathing-zone air samples
were collected with both a charcoal tube and a passive diffusion dosimeter
during the day and evening shifts for seven different departments. The depart-
ments were: 1) tank farm, 2) reactor and recovery, 3) latex blending and
solution, 4) shipping and receiving, 5) storeroom, 6) factory service, and 7)
maintenance areas. Sampling periods ranged from 4 to 6 hours. Charcoal tubes
were changed at intervals of 1 to 2 hours during the sampling period to avoid
overloading.
Blood samples of male hourly production workers for the same departments
were obtained on 4 separate days, May 15-18. Of the 163 workers (26-65 years
6-24
-------
of age with a median age of 45 years; 144 whites and 19 blacks) who participat-
ed in the industrial hygiene survey, 154 (135 whites and 19 blacks) also par-
ticipated in the blood survey. Because of work scheduling demands, blood
samples were collected from participants from each of the departments on all 4
days, thereby minimizing any bias due to day effect. The hematological parame-
ters measured included red cell count (RBC), hemoglobin concentration, hemato-
crit, mean corpuscular volume (MCV), mean corpuscular hemoglobin concentration,
reticulocyte count, platelets, total leukocytes (WBC), and differential distri-
butions of neutrophils, neutrophil band forms, eosinophils, basophils, monocytes,
and lymphocytes. Medical histories were obtained by means of questionnaires.
Data from persons who reported positive histories of either malignant disease,
radiation therapy, or current anemia of known etiology were excluded from the
analyses. Only one individual, who reported a history of leukemia, was excluded
from the study.
The mean 8-hour time-weighted averages and ranges show that all four chem-
ical exposures were well below the American Conference of Governmental Indus-
trial Hygienists (AC6IH) Threshold Limit Values (TLVs) recommended at that
time. The TLVs in parts per million (ppm) for butadiene, styrene, benzene, and
toluene are 1,000, 100, 10, and 100, respectively. With the exception of
butadiene and styrene^ for which time-weighted averages of 20.03 ppm and 13.67
ppm, respectively, were observed in the tank farm area, the mean levels for the
four chemicals in all other departments were less than 2 ppm. Even the maximum
concentration of benzene, the most strongly suspected leukemogen of the four
chemicals analyzed, was less than 1 ppm in all plant departments.
With regard to the hematologic survey, there were generally no associations
(p > 0.05) of hematologic values with chemical exposures. Red blood cell count
was negatively associated with butadiene and styrene exposure, while basophil
6-25
-------
count was positively related to the aforementioned chemicals, as measured by
Pearson product moment analysis. The negative association of styrene with
erythrocyte counts and the positive association of the basophil proportions
with butadiene persisted after controlling for age and medical status in step-
wise multiple linear regression analyses. However, there were curious opposing
findings for mean corpuscular hemoglobin concentrationa positive relationship
to butadiene and a negative association with styrene.
Mean hematologic mesurements adjusted for age and medical status were com-
pared for tank farm area workers and all other workers. The tank farm workers
had slightly lower levels of circulating erythrocytes, hemoglobin, platelets,
and neutrophils, in addition to mean corpuscular red cell volumes and neutro-
phil band circuits that were slightly higher than those of the other workers.
This study was undertaken to quantify exposure levels and to find out if
there is any evidence of hematopoietic toxicity in relation to these exposure
levels. With the exception of the tank farm area, the average exposures to the
four chemicals assayed were uniformly less than 2 ppm; even in the tank farm
area, the styrene and butadiene concentrations were considerably lower than the
recommended ACGIH TLVs, although they were considerably higher than in other
workplaces studied.
Because this study is cross-sectional in design, it is very limited with
regard to determining whether styrene-butadiene exposure is carcinogenic.
Individuals in the plant who may have developed cancer probably left the work
force and hence were not available for blood sampling. The industrial hygiene
survey findings cannot be generalized to the past, since the concentrations may
have differed quantitatively as well as qualitatively. It can be concluded
that there was no pronounced evidence of hematologic abnormality in this study
population.
6-26
-------
6.2.4. Melnhardt et al. (1982)
Meinhardt et al. (1982) reported on a retrospective cohort mortality study
conducted by NIOSH at two adjacent SBR facilities in eastern Texas. This study
was motivated by the report of two men who had worked at both plants, and who
had died of leukemia in January 1976.
Personnel employment records documenting the employment of 3,494 workers
from plant A and 2,015 from plant B were available since January 1943 and
January 1950, respectively, to the study cutoff date of March 31, 1976. The
study cohorts from plants A and B consisted of 1,662 and 1,094 white males who
had had at least 6 months of non-management and non-administrative employment,
respectively. The average lengths of employment for the study cohort in plants
A and B were, respectively, 9.48 and 10.78 years.
At the time of the study, environmental samples were obtained at each
plant. At plant A, time-weighted average exposures of styrene, butadiene, and
benzene were 0.94 ppm (0.03-6.46), 1.24 ppm (0.11-4.17) and 0.10 ppm (0.08-0.14)
respectively. For plant B, time-weighted average exposures for styrene and
butadiene were 1.99 ppm (0.05-12.3) and 13.5 ppm (0.34-174.0), respectively
(benzene was not measured). No historical monitoring data were available for
either plant.
The study cohorts from plants A and B accounted for 34,187 and 19,742
person-years at risk of dying. It is also important to note that the survival
status of 54 individuals (3.25%) from cohort A and 34 individuals (3.11%)
from cohort B was unable to be determined. In subsequent analyses, these
individuals were considered to be alive.
The age, race, sex, calendar time, and cause-specific mortality rates of
the U.S. population were applied to the appropriate strata of person-years at
risk in order to obtain the expected number of cause-specific deaths in the
6-27
-------
study populations. Differences in observed and expected numbers of deaths were
evaluated by a test statistic based on the Poisson distribution.
In cohorts from plants A and B, observed and expected numbers of deaths
were compared for the following cause-specific categories: tuberculosis;
malignant neoplams (including cancers of the lymphatic and hematopoietic tis-
sues); all other cancers; diseases of the nervous, circulatory, respiratory,
and digestive systems; accidents; and all other causes. With the exception of
mortalities from cancers of the lymphatic and hematopoietic tissues, there were
deficits (in some instances, striking deficits) in the cause-specific SMRs for
the study cohorts in both the plants.
With regard to the total number of deaths due to all causes, cohorts A and
B had observed numbers of deaths of 252 (SMR = 80) and 80 (SMR = 66), respec-
tively. Although it is possible that the "healthy worker" effect may, in part,
explain these deficits, the relative magnitudes of the deficits, particularly
for plant B, suggest that there may have been an underreporting of deaths or that
selection factors in the choice of the study groups reduced the mortality rate.
Meinhardt et al. observed that all five of the individuals from plant A
whose underlying cause of death was leukemia began employment before the end of
December 1945. This date corresponds to the time when the batch process for
SBR production was converted to a continuous-feed operation. The decision was
made, therefore, to evaluate the mortality experience of those 600 white male
employees who had had at least 6 months of employment in plant A between Janu-
ary 1, 1943 and December 31, 1945. This subgroup was employed for an average
of 11.9 years, and had an accumulation of 17,086 person-years at risk of dying.
The survival status of 34 individuals (5.7%) was unknown. As in the previous
mortality analyses, with the exception of deaths from cancers of the lymphatic
and hematopoietic tissues, there generally were large deficits in the cause-
6-28
-------
specific observed number of deaths. For malignant neoplasms of lymphatic and
hematopoietic tissues, the SMR was 212 (9 observed, 4.25 expected,_.0.05 < p
< 0.1); for lymphosarcoma and reticulosarcoma the SMR was 224 (3 cases, 1.34
expected, p > 0.05); for Hodgkin's disease the SMR was 213 (1 case, 0.47 ex-
pected, p > 0.05); and for leukemia and aleukemia the SMR was 278 (5 cases,
1.80 expected,- 0.05 < p < 0.1). The total number of observed deaths due to all
causes was 201 (SMR = 83, 242.09 expected, p < 0.05). For cohort B, there
were no significant (p > 0.05) excesses of mortality from any cause. Deaths
from all malignant neoplasms (SMR = 53, observed = 11, expected = 20.78, p <
0.05) and "all other causes" (SMR = 54, observed = 9, expected = 16.80, p <
0.05) were significantly decreased, however.
The authors calculated the likelihood of detecting a doubling and a qua-
drupling of the expected occurrence of leukemia for cohort A, subcohort A, and
cohort B. These likelihoods were 26% and 88% for cohort A, 20% and 77% for.
subcohort A, and 13% and 51% for cohort B. Thus, this study suggests that some
component of the SBR manufacturing process may be a leukemogen.
6.2.5. Matanoski et al. (1982)
This study was performed to determine whether there are health risks
associated with the production of synthetic rubber, specifically styrene-buta-
diene rubber (SBR). The populations studied were obtained from seven U.S. and
one Canadian rubber plants. The study population consisted of males who had
worked for more than one year, and whose records contained birth dates and
employment dates. In addition, workers selected had been employed by the
facility from the date of the facility's first SBR production to December 1976.
The total number of people available from the eight plants surveyed was 29,179,
of whom 13,920 (48%) met the criteria for selection.
Data obtained from the personnel records of each facility included employee
6-29
-------
name, Social Security number, job history, date(s) of employment, birth infor-
mation, death information, and limited data on retirees. Individual workers-
jobs were coded according to first job, last job, and job held longest during
the period of employment. For the analyses, jobs were categorized in four
general work areas: production, utilities, maintenance, and other. In three
plants (plants 3, 4, and 5), race classification was unknown for 176 (85%),
329 (50%), and 4,540 (98%) of the cohort populations, respectively. Plants 3
and 4 were expected to have employed black workers; plant 5 had employed few or
no blacks at any time in its history. In all, 7,209 (52%) of the study popula-
tion were unable to be classified racially. If race was not specified, individ-
uals were assumed to be white males.
Follow-up activities to determine vital status of the study population
in the seven U.S. plants included searching by the Social Security Administra-
tion, tracing through motor vehicle administration records, and contacts by
telephone. Through these follow-up activities, 42% of the study population
were traced. For the Canadian plant, follow-up was performed by searching the
company insurance plan records for death benefit information. Determination
of vital status for the study populations revealed 10,899 workers alive, 2,097
known dead, and 924 lost to follow-up. It was determined from a 10% sample in
each plant that about 4% of the study population assumed to be living was actu-
ally dead. Thus, as many as 440 deaths, or 17% of the total possible deaths,
might have been missed. For the populations who were known dead, 90% of the
death certificates requested were received. Death certificates were coded
according to the 8th revision of the ICD.
Most of the statistical analyses were done using the worker records from
the time when record-keeping systems became complete, or any time thereafter
through December 1976. Females, workers employed less than one year, and those
6-30
-------
with unknown birth dates or employment dates were omitted. The total eligible
population numbered 13,6.08. SMRs for the workers as compared to the general
population were calculated by using a modified version of U.S. Death Rates Pro-
9rams (Monson, 1979). SMRs were calculated separately for the white population
and the black population, and the ratios were combined to correct for differen-
ces in age, race, and calendar time. The total number of deaths occurring
among the eligible study population was 1,995, with an SMR of 81. The study
cohort accounted for 250,000 person-years.
Power calculations were performed to test the ability of the data to deter-
mine increases in risk of 0.1, 0.25, 0.5, and 2 times greater than the U.S.
population. The calculations showed a 100% probability of detecting a twofold
increase in all causes of death and in all cancers, and an 89% probability of
detecting a 25% increase in lung cancer.
The average period of follow-up was 19 to 20 years. .The average age at
death was 61 years. The overall SMR for all causes of death was 81, with SMRs
of 98 for blacks and 78 for whites. The low mortality among members of this
population is, in part, a reflection of the "healthy worker" effect. However,
a question must also be raised as to the effect of the 440 possible deaths that
may have been missed. Furthermore, the large difference in SMRs for blacks and
whites may be due to an undercounting of blacks and an overcounting of whites.
This is further exemplified by similar patterns in SMRs for all accidents,
motor vehicle accidents, and suicides. The SMRs for all causes of death do not
appear to increase with duration of employment but do appear to increase with
increasing follow-up period.
The leukemias found in the population included nine acute, five chronic,
and three unspecified on the death certificate, with a median latent period
of 17 years from time of first employment. The types found were not considered
6-31
-------
to have a distribution remarkably different from that found in the general
'population. The study population as a whole did not demonstrate leukemia in
excess.
None of the analyses demonstrated significant increases in SMRs for other
specific causes in the total study population. Black males appear to have a
significantly elevated risk of arteriosclerotic heart disease (p < 0.05), but
this value may be artificially inflated due to undercounting of the blacks,
resulting in a smaller denominator. Vascular lesions of the central nervous
system were also in excess in black males, but they were not statistically
significant.
Increases in mortality according to job classifications are noteworthy,
but only two of them are statistically significant (p < 0.05). The SMR for
testicular cancer among maintenance workers is 294 (observed = 3, p < 0.05).
SMRs for esophagus, stomach and large intestine, and larynx cancers are
elevated in the utilities and maintenance work areas. The SMR for larynx can-
cer in utilities workers is statistically significant (SMR = 476, observed =
4, p < 0.05). Hodgkin's disease is associated with high SMRs in all work
areas. Very few high SMRs are found for the production area.
It should be further pointed out that, with the exceptions of Hodgkin's
disease and stomach cancer, all of the cancers had relatively long latent peri
ods. For all cancers, half of the individuals had had 12 years of employment
or more.
In addition to complete ascertainment of deaths and racial distribution,
further investigation is needed in order to obtain information specific to
various jobs and the associated exposures of individuals in those jobs. It
would also be useful to separate the number of years employed within each job
category, in order to determine the periods of possible exposure in the work-
6-32
-------
place. Moreover, changes in the SBR process, in plant design, and in worker
practices should be given greater attention in the evaluation of mortality for
the four worker categories studied.
Other methodologic limitations of this study include the fact that less
than 50% of the total population of eight plants was studied. This raises
questions about the population that was excluded due to lack of birth dates or
employment dates. This may have been an older population, which probably had
longer exposure and was therefore more likely to suffer from occupational dis-
eases. Out of eight cohorts, only 50% were followed from 1943, whereas in
the rest of the plants, follow-up starting dates ranged from 1953 to 1970. It
is probable that the employees from the latter plants were not followed long
enough for malignancies to develop.
6.2.6. Summary of Epidemiclogic Studies
McMichael et al. (1974) found significant (p < 0.05) excess mortality from
cancers of the stomach, prostate, and lymphatic and hematopoietic system as
well as from diabetes mellitus, arteriosclerosis, and ischemic heart disease in
their historic prospective cohort study. To evaluate these excesses, McMichael
et al. (1976), in a case-control study, investigated exposure rates for various
jobs within the same rubber plant for several cause-specific deaths. The data
indicated that the most probable health risks were of prostate cancer in jani-
toring and trucking; bladder cancer in milling and reclaiming operations; lym-
phatic and hematopoietic cancers in the synthetic plant; and lymphatic leukemia
in the synthetic plant, inspection-finishing-repair, and tread cementing. Non-
malignant mortality excesses included ischemic heart disease in the synthetic
plant and tread cementing and diabetes mellitus in janitoring and trucking and
inspection-finishing-repair.
Andjelkovich et al. (1976) carried out a similar kind of study (historic
6-33
-------
prospective cohort) in white males in a single rubber plant, and observed sig-
nificantly (p < 0.05) increased SMRs for malignant neoplasms of lymphatic and
hematopoietic tissues (monocytic leukemia and other neoplasms of lymphatic and
hematopoietic tissue), chronic rheumatic heart disease, and cerebrovascular
disease. They also observed high SMRs (p < 0.001) for all causes and for most
of the cause-specific deaths for a group of workers who had retired between the
ages of 40 and 64. Andjelkovich et al. (1977) evaluated these excesses in
mortality ratios in relation to various work areas by using the entire cohort
as a reference group, and found that only malignant neoplasms of the trachea,
bronchus, and lung were associated with the synthetic latex department. This
finding was based on only three observed deaths, however, and no smoking data
were taken.
Checkoway and Williams (1982) carried out an industrial hygiene and hema-
tology cross-sectional survey at the same plant in which McMichael et al. had
conducted their case-control study. With the exception of the tank farm area,
in which 8-hour time-weighted averages for butadiene and styrene were observed
to be 20.3 ppm and 13.67 ppm, respectively, all other departments had mean
exposure levels of less than 2.0 ppm. No evidence of hematologic abnormality
was noted. Because of its cross-sectional design, however, this study could
not be expected to identify an excess cancer risk.
Meinhardt et al. (1982), conducted a retrospective cohort mortality study
at two plants in Texas. Male rubber workers in one of the plants (plant A)
were followed from January 1, 1943 through March 31, 1975; employees in the
other plant (plant B) were followed from January 1, 1950 through March 31,
1976. These intervals of observation are important to note because the two
plants changed from a batch process for SBR production to a continuous-feed
operation in 1946. With regard to cancers of the lymphatic and hematopoietic
6-34
-------
system and lymphatic leukemia, plant A. exhibited excess mortalities, although
these were not statistically significant (p > 0.05); plant B did not show any
mortality excesses. When the mortality experience in plant A was analyzed
further for those workers who had had at least 6 months of employment between
January 1, 1943, and December 31, 1945 (the interval for which the batch pro-
cess was used), excess mortalities for the above-mentioned cancers were shown
to be of borderline statistical significance (0.05 < p < 0.01, two-sided).
It is also of interest to note that all of the employees from the total cohort
of plant A whose causes of death were cancers of the lymphatic and hematopoietic
tissues had been employed between 1943 and 1945. Had the analysis in plant A
commenced with the date of first employment in 1946, the SMRs in question would
have been reduced to zero, and the lack of excess mortality would have been
similar to plant B.
Matanoski et al. (1982) also conducted a retrospective cohort mortality
study in which eight SBR plants were involved. As with the cohort in plant B
investigated by Meinhardt et al., there was a general lack of excess mortali-
ties. It should also be noted that half of the cohort was followed from 1943
to 1979. The date of entry for the remaining half of the cohort ranged from
1953 to 1976, with follow-up terminating in 1979.
Although both the McMichael et al. (1976) and the Meinhardt et al. (1982)
studies found some evidence of an association between styrene-butadiene exposure
and lymphatic and hematopoietic cancer, confounding due to exposures to solvents
cannot be ruled out for either study. In addition, the results from the Mein-
hardt et al. study for the subcohort employed during the batch process of pro-
duction were only of borderline (0.05 < p < 0.1, two-sided) significance, and
a possibly serious underascertainment of deaths, and/or selection factors in
the choice of the study group may have biased the results.
6-35
-------
The Andjelkovich et al. (1977) study found an association between employ-
ment in the synthetic part of the plant with mortality from cancer of the
trachea, bronchus, and lung. The association was based on only three deaths,
however, and there was no control for smoking.
The study by Matanoski et al. of almost 14,000 styrene-butadiene produc-
tion workers found no excesses of cancer mortality that were statistically
significant (p < 0.05). Again, however, a possibly serious underascertainment
of deaths may have biased the results. An undercounting of blacks in the study
population may also have resulted in a potential bias.
The epidemiologic evaluation of SBR workers with regard to the carcinoge-
nicity of 1,3-butadiene is particularly difficult because styrene may also be
a carcinogen and, in particular, a leukemogen (Ott et al., 1980; Lilis and
Nicholson, 1976). Because of the inconsistency of the results from different
studies, the possible confounding due to solvent and styrene exposures, and the
potential for bias in some of the studies, the epidemiologic data are considered
inadequate for determining a causal association between 1,3-butadiene exposure
and cancer in humans.
6.3. QUANTITATIVE ESTIMATION
This section deals with the incremental unit risk for 1,3-butadiene in air,
and the potency of 1,3-butadiene relative to other carcinogens that the CAG has
evaluated. The incremental unit risk estimate for an air pollutant is defined
as the excess lifetime cancer risk occurring in a hypothetical population in
which all individuals are exposed continuously from birth throughout their
lifetimes to a concentration of 1 ppm.or 1 Pg/m3 of the agent in the air they
breathe. This calculation estimates in quantitative terms the impact of the
agent as a carcinogen. Unit risk estimates are used for two purposes: 1) to
compare the carcinogenic potency of several agents with each other, and 2) to
6-36
-------
give a crude indication of the population risk that might be associated with
air exposure to these agents if the actual exposures were known. Hereafter,
the term "unit risk" will refer to incremental unit risk.
The development of the quantitative estimation section will be to first
describe the procedures, assumptions, and uncertainties involved in animal-to-
human extrapolation (section 6.3.1.), and then to use these procedures in the
actual calculation of unit-risk estimates, (section 6.3.2.). These estimates
will then be used with the available human data to determine if the extrapolated
animal values actually predict the human response (section 6.3.3.). Finally,
the carcinogenic potency of 1,3-butadiene will be compared with the potencies
of other compounds that the EPA has evaluated as known or suspect human car-
cinogens.
6.3.1. Procedures for Determination of .Unit Risk
The data used for quantitative estimation are taken from one or both of
the following: 1) lifetime animal studies, and 2) human studies where excess
cancer risk has been associated with exposure to the agent. In animal studies
it is assumed, unless evidence exists to the contrary, that if a carcinogenic
response occurs at the dose levels used in the study, then responses will also
occur at all lower doses, with an incidence determined by an extrapolation
model.
There is no solid scientific basis for any mathematical extrapolation
model that relates carcinogen exposure to cancer risks at the extremely low
concentrations that must be dealt with in evaluating environmental hazards.
For practical reasons, such low levels of risk cannot be measured directly
either by animal experiments or by epidemiologic studies.
The linear nonthreshold model has been adopted as the primary basis for
risk extrapolation to low levels of the dose-response relationship (U.S. EPA,
6-37
-------
1984b). The incremental risk estimates made with this model, and the corres-
ponding 95% upper-limit incremental unit risks, should be regarded as conserva-
tive, representing the most plausible upper limits for the risk, i.e., the true
risk is not likely to be higher than the estimates, but it could be lower.
The mathematical formulation chosen to describe the linear nonthreshold
dose-response relationship at low doses is the linearized multistage model.
The multistage model employs enough arbitrary constants to fit almost any
monotonically increasing dose-response data, and it incorporates a procedure
for estimating the largest possible linear slope (in the 95% confidence limit
sense) at low extrapolated doses that is consistent with the data at all dose
levels of the experiment. This procedure effectively linearizes the model at
low doses. Thus, the multistage model, described below, is fitted to the data
in the observational or experimental range. The fit of the curve allows for a
linear term which dominates the risk estimate at low doses. The 95% upper
limit, q£, described below, is technically an upper-limit estimate on the
linear term, but, practically, functions as the upper-limit low dose-response
function.
6.3.1.1. Description of the Low-Dose Extrapolation Model Let P(d) represent
the lifetime risk (probability) of cancer at dose d. The multistage model has
the form
P(d) = 1 - exp C-(q0
. . .+ qkdk)]
where
q-i _> 0, i = 0, 1, 2 k
6-38
-------
Equivalently,
Pt(d) = 1 - exp [-(qjd + q2d2 + ... + qkdk)]
where
Pjd) = P(d) - P(0)
* 1 - P(0)
is the extra risk over background rate at dose d.
The point estimate of the coefficients q-j, i = 0, 1, 2, ..., k, and con-
sequently, the extra risk function, Pt(d), at any given dose d, is calculated
by maximizing the likelihood function of the data.
The point estimate and the 95% upper confidence limit of the extra risk,
Pt(d), are calculated by using the computer program GLOBAL83, Howe (1983). At
low doses, upper 95% confidence limits on the extra risk and lower 95% confi-
dence limits on the dose producing a given risk are determined from a 95%
upper confidence limit, q£, on parameter q-^ Whenever q-^ > 0 at low doses,
the extra risk ?t(d) has approximately the form Pt(d) = qj x d. Therefore,
q1 x d is a 95% upper confidence limit on the extra risk, and R/q£ is
a 95% lower confidence limit on the dose, producing an extra risk of R. Both
the model and the curve-fitting methodology used, including the step-down
procedure for eliminating the highest dose groups, are discussed in detail by
Anderson et al. (1983). For the Hazleton female rat inhalation study, the
multistage model did not fit the full data set and the highest dose group was
dropped.
When all of the higher-order terms in the multistage model are zero except
for the linear term, the multistage model reduces to the one-hit model, which
6-39
-------
is a true low-dose linear nonthreshold model. As will be seen with'the NTP
male mouse data and the Hazleton female rat data, this is the case with 1,3-
butadiene.
For cases of partial lifetime exposure where time-to-tumor or time-to-
tumor death is known, Crump and Howe (1983a) have developed the multistage
model to include a time-dependent dose pattern. The form of this model is one
which is linear in dose and in which time has a power and form determined by
both the number of assumed stages and the stage affected by the carcinogen.
A best fit is determined by the method of maximum likelihood in the ADOLL183
computer program (Crump and Howe, 1983b). Application of this program to the
NTP mouse data was unsuccessful because of lack of convergence when attempting
to extrapolate from the 60-61 week study to the normal 104-week treatment
period.
6.3.1.2. Calculation of Human Equivalent Dosages from Animal DataFollowing
the suggestion of Mantel and Schneiderman (1975), it is assumed that mg/sur-
face area/day is an equivalent dose between species. Since, to a close approx-
imation, surface area is proportional to the two-thirds power of weight, as
would be the case for a perfect sphere, the dose in mg/day per two-thirds power
of the weight is also considered to be equivalent dose. In an animal experi-
ment, this equivalent dose is computed in the following manner:
Let
Le - duration of experiment
le = duration of exposure .
m = average dose per day in mg during administration of the agent (i.e.,
during le), and
W = average weight of the experimental animal
6-40
-------
The lifetime dose is then
d=le_xm
Le x W
2/3
1,3-Butadiene is slightly soluble in water and can be considered a par-
tially soluble vapor. The dose in m - mg/day of partially soluble vapors is
proportional to the 02 consumption, which in turn is proportional to W2/3 and
is also proportional to the solubility of the gas in body fluids, which can be
expressed as an absorption coefficient, r, for the gas. Therefore, expressing
the 02 consumption as 02 = k W2/3, where k is a constant independent of
species and V = mg/m3 of the agent in air, it follows that:
m
= k W2/3 x v x r
or
d =
= kvr
2/3
W
In the absence of experimental information or a sound theoretical argument
to the contrary, the absorption fraction, r, is assumed to be the same for all
species. Therefore, for these substances (including 1,3-butadiene) a certain
concentration in ppm or u9/m3 in experimental animals is assumed equivalent
to the same concentration in humans. This is supported by the observation that
the minimum alveolar concentration necessary to produce a given "stage" of
anesthesia is similar in man and animals (Dripps et a!., 1977). It is further
6-41
-------
supported by the results of the NTP absorption study in mice and rats (1985a)
discussed in Chapter 4.
As shown previously, since both the NTP mouse and Hazleton rat cancer
bioassays used inhalation exposure in terms of 1,3-butadiene concentrations in
the air, the direct ppm-animal-to-ppm-human exposures are appropriate for low
exposure concentrations. However, as shown in section 4.1., the absorption
fraction, r, is not constant for the high external concentrations presented
to the animals in the bioassays, but decreases with increasing concentration.
Thus, the procedure for determining animal-to-equivalent-human dose must be
adjusted to account for the fact that at high concentrations the internal dose
(in mg/kg) is not directly proportional to the external concentration. The
method of adjustment used is to first calculate the function for the incre-
mental cancer risk to the animal based on internal dose (in mg/kg for within-
species standardization), then convert back to risk for low-dose ppm equivalents
in the animal. Since low-dose conversions from animals to humans for 1,3-buta-
diene are on a ppm equivalence in the air, the final animal risk in units of
ppm will also be that for humans. This is done because there are no human
data on external concentration to internal dose.
In order to convert to internal dose in animals, the results of the NTP
absorption study of butadiene in male rats and mice via inhalation (1985a) are
used. These results, presented in Table 4-3 as pmol/kg of butadiene retained,
are first transformed to yg/kg retained for each of the exposure concentra-
tions~70, 930, and 7,100 ppm for the rats and 7, 80, and 1,040 ppm for the
male mice. Responses for the high-dose rat group and middle-dose mouse group
were adjusted upward by the factor 12/11 to standardize to a 6-hour exposure.
Log pg/kg retained vs. log ppm exposure concentration curves were fitted
and these curves were used to estimate the internal (retained) dose of the
6-42
-------
mouse and rat cancer bioassays. The estimated internal doses for the mice are
25.7 mg/kg (474 pmol/kg) at 625 ppm and 38.9 mg/kg (719 umol/kg) at 1,250
ppm exposure concentration for 6 hours. For the rat data the estimated inter-
nal doses are 10.5 mg/kg (195 pmol/kg) at 1,000 ppm and 37.1 mg/kg (685 ymol/
kg) at 8,000 ppm exposure concentration for 6 hours. Thus, the mice exposed to
625 ppm actually received more than twice the internal dose, on a mg/kg basis,
than did the rats exposed to 1,000 ppm. Even the mice exposed to 1,250 ppm
received a larger internal dose than the rats exposed to 8,000 ppm.
6.3.1.2.1. Adjustments for less than lifetime duration of experiment. When
analyzing quanta! data, if the duration of experiment Le is less than the
natural lifespan of the test animal L, the slope q£, or more generally the
exponent g(d), is increased by multiplying by a factor (L/l_e)3. We assume
that if the average dose d had been continued, the age-specific rate of cancer
would have continued to increase as a constant function of the background rate.
The age-specific rates for humans increase at least by the second power of the
age and often by a considerably higher power, as demonstrated by Doll (1971).
Thus, it is expected that the cumulative tumor rate would increase by at least
the third power of age. Using this fact, it is assumed that the slope q£,
or more generally the exponent g(d), would also increase by at least the third
power of age. As a result, if the slope qj [or g(d)] is calculated at age
Le, we expect that if the experiment had been continued for the full lifespan,
L, at the given average exposure, the slope q^ [or g(d)] would have been
increased by at least (L/l_e)3. This correction is used for extrapolation from
the NTP mouse study, which was terminated after 60-61 weeks instead of running
for a full 2 years.
For time-to-tumor data, this adjustment is also conceptually consistent
with the proportional hazard model proposed by Cox (1972) and the time-to-tumor
6-43
-------
model considered by Crump (1979), where the probability of cancer by age t and
at dose d is given by
P(d,t) = 1 - expC-f(t) x g(d)]
It is also consistent with the partial lifetime exposure extension of the
multistage model developed by Crump and Howe (1983a), which as discussed above
is linear in dose, but has a power function of time.
6.3.1.3. Interpretation of Quantitative EstimatesFor several reasons, the
unit risk estimate based on animal bioassays is only an approximate indication
of the absolute risk in populations exposed to known carcinogen concentrations.
First, there are important species differences in uptake, metabolism, and organ
distribution of carcinogens, as well as species differences in target site
susceptibility, immunological responses, hormone function, dietary factors, and
disease. Second, the concept of equivalent doses for humans compared to animals
on a mg/surface area basis is virtually without experimental verification
regarding carcinogenic response. Finally, human populations are variable with
respect to genetic constitution and diet, living environment, activity patterns,
and other cultural factors.
The unit risk estimate can give a rough indication of the relative potency
of a given agent as compared with other carcinogens. The comparative potency
of different agents is more reliable when the comparison is based on studies
in the same test species, strain, and sex, and by the same route of exposure,
preferably inhalation.
The quantitative aspect of carcinogen risk assessment is included here
because it may be of use in the regulatory decision-making process, e.g.,
setting regulatory priorities, evaluating the adequacy of technology-based
6-44
-------
controls, etc. However, it should be recognized that the estimation of cancer
risks to humans at low levels of exposure is uncertain. Because of the limited
data available from animal bioassays, which is at the high dose levels required
for testing, almost nothing is known about the true shape of the dose-response
curve at low environmental levels. At best, the linear extrapolation model
used here provides a rough but plausible estimate of the upper limit of risk;
i.e., it is not likely that the true risk would be much more than the estimated
risk, but it could be considerably lower. The risk estimates presented in
subsequent sections should not be regarded as accurate representations of the
true cancer risk even when the exposures are accurately defined. The estimates
presented may, however, be factored into regulatory decisions to the extent
that the concept of upper risk limits is found to be useful.
6.3.1.4. Alternative Models--The methods used by the CA6 for quantitative
extrapolation from animal to man are generally conservative, i.e., tending
toward high estimates of risk. The most important part of the methodology
contributing to this conservatism is the CAG's use of the linearized multistage
nonthreshold extrapolation model. There are a variety of other extrapolation
models that could be used, most of which would give lower risk estimates.
Among these alternative models, two which are currently popular and which
often tend to give different low-dose extrapolations from the multistage model
are the log-Probit and Weibull models. These models have not been used in
the following analyses because the data do not warrant it. As discussed
below, all models are of limited value for predicting low-dose risks for
1,3-butadiene based on the mouse responses, which were greater than 60% at
the lowest dose tested. With respect to the rat study, even though low-dose
response was less than that of the mouse, the quality of the study, and its
results, have not been peer-reviewed or published. Furthermore, complete
6-45
-------
individual pathology information was not available. Thus, the results of the
extrapolation procedure are limited by the actual data bases.
6.3.1.5. Internal Dose vs. External Concentration--Pharmacokinetic modeling
would be a very useful way to predict the internal concentration of the com-
pounds (parent and metabolites) of interest in this case. To do such modeling,
certain basic in vivo data are necessary. From the present data outlined here,
it is not possible to do such modeling.
Even a simple one-compartment classical model would need data regarding
blood/plasma concentrations of at least the parent compound. At this time
these data are not available. Several other pieces of data are also required
to construct such a model with confidence. For example, elimination of the
unmetabolized parent compound via the kidneys and lungs would need to be
established.
However, it appears after reviewing possible metabolic paths of this
1,3-butadiene that a physiologically based pharmacokinetic model is in order.
Such a model could adequately describe the disposition of the parent compound
and its metabolites in several organs of interest. It would consider many
physiological parameters and would rationally consider scale-up to humans so
that the internal concentrations predicted could be used in risk assessment.
Further, risk assessment based on the internal concentration of dose re-
tained, as is done here, does not consider the possibility that more than one
metabolite is being formed. It is possible that not all of these metabolites
are toxic and that relative concentrations may be changing in a manner differ-
ent from the total labeled materials. Therefore, while-the total retained
labeled materials may be increasing, the amount of toxic metabolite may increase
only to some saturated level. Thus, such a scenario may wrongly estimate the
amount of toxic metabolite(s) within the animal tissues. A physiologically
6-46
-------
based pharmacokinetic model that could consider metabolites and differentiate
mobile metabolites from incorporated metabolites would be extremely useful.
To develop such a physiologically based model would require much more in
vivo data than are currently available. Examples of needed information include
organ concentrations at various times during and after exposure, arterial and
venous blood concentrations, renal and pulmonary clearance rates, and binding
information. Also, the labeled materials would have to be characterized to
discriminate between the parent compound and its metabolites. It might not be
necessary to identify the metabolites, but it would be useful to determine
which portion of metabolites is mobile and which portion is incorporated into
the biomolecules of the various organs and tissues.
The formulation of such a model would improve the data base used for
bettering risk assessment. These types, of models can, with proper verification,
provide reliable estimates of internal concentrations, which may also be re-
solved at organ level.
6.3.2. Calculation of Quantitative Estimates
Human studies have provided inadequate evidence for the carcinogenicity
of 1,3-butadiene. The major weakness is lack of good 1,3-butadiene exposure
information. Additionally, concurrent exposure to several other possible
carcinogens also limits the use of these studies as primary sources for
calculating quantitative risk estimates. For animal-to-human extrapolation,
there are two suitable animal bioassays, the NTP mouse study and the Hazleton
rat study, both showing significant carcinogenic response. As discussed above
and in previous sections, the rat bioassay has deficiencies limiting its use
as the primary data set for animal-to-man extrapolation. Nevertheless, it
will be compared with the results of the mouse bioassay for the purposes of
sensitivity analysis. The mouse study will be considered as the primary study.
6-47
-------
6.3.2.1. Mouse-to-Human ExtrapolationThe NTP (1984) mouse inhalation study
Showed highly statistically significant increases (p < 0.01) both in hemangio-
sarcomas of the heart and malignant lymphomas and in both males and females at
625 ppm and 1,250 ppm. Both of these tumor types are life-threatening and
appeared quite early in the study. As discussed at length in the qualitative
section and shown in Table 6-1, several other tumor sites were also significantly
increased in the study, which was stopped at 60-61 weeks due to high mortality
from tumors in the treated groups.
Using the continuous equivalent dosages, maximum likelihood and 95% upper-
limit incremental, unit risk estimates were calculated from both the male and
female mouse data. For the male mice, the fractions of animals either with
tumors at significantly increased sites, or with tumors considered unusual for
60 weeks (preputial gland squamous cell carcinomas and Zymbal gland carcinomas,
see Tables 6-1 and 6-3) were 2/50, 43/49, and 40/45 for the control, 625-ppm,
and 1,250-ppm groups, respectively. For the females, the fractions of animals
with significantly increased tumors or brain gliomas were 4/48, 31/48, and
45/49. Animals that died before the first tumor was seen (at 20 weeks) were
eliminated. These results are presented in Table 6-4, with internal doses
multiplied by 5/7 to determine an average daily dose. Also shown are the
maximum likelihood estimates (MLE) and the 95% upper-limit incremental unit
risk estimates (q^) based on these data. The initial upper-limit estimates
based on the 60-week (male) and 61-week (female) studies are then adjusted to
project for natural lifetime risk (see section 6.3.1.2.1.). The final estimates
are q£ = 6.1 x 10"1 (mg/kg/ day)"1 for the males and q* = 3.0 x 10"1
(mg/kg/day)-1 for the females.
6-48
-------
TABLE 6-4. MOUSE: CALCULATION OF CANCER RISK BASED ON INTERNAL
DOSES AND CONVERSION TO LOW PPM CONCENTRATIONS.
FRACTIONS (%) OF MICE (NTP, 1984)
WITH AT LEAST ONE OF THE STATISTICALLY SIGNIFICANT INCREASED TUMORS
OR TUMORS CONSIDERED UNUSUAL AT TIME OF TERMINAL SACRIFICE.a
ALSO, MLE AND 95% UPPER-LIMIT INCREMENTAL UNIT RISK ESTIMATES BASED ON
THE LINEARIZED MULTISTAGE MODEL
Males
Nominal
exposure
(ppm)
0
625
1,250
Equivalent
continuous
internal dose*3
(mg/kg/day)
0
18.4
27.8
Number
with tumors/
number
exami nedc
2/50 (4%)
43/49 (88%)
40/45 (89%)
Females
Equivalent
continuous
internal dose
(mg/kg/day)
0
18.4
27.8
Number
with tumors/
number
exami nedc
4/48 (8%)
31/48 (65%)
45/49 (92%)
Tables 6-1 and 6-3 for tumor sites.
blnterpolated daily internal doses x 5/7 since treatment was 5 days per week
for a lifetime.
Examined for either hemangiosarcomas or lymphomas, eliminating animals that
died prior to 20 weeks.
Initial maximum likelihood estimates:
Males
q0 = 0.042
q = 9.35 x
Females
= 0.086
'2 (mg/kg/day)~l
Initial estimates of 95% upper-limit:
qj = 1.17 x lO'1 (mg/kg/day)'1
Adjustment factor for early sacrifice:
(104/60)3 =5.21
Final estimate of 95% upper-limit in (mg/kg/day)-1:
q* = 6.1 x ID'1 (mg/kg/day)-1 q* = 3.0 x 10'1 (mg/kg/day)-1
Final estimate of 95% upper limit in (ppm)-1: 1 ppm = 1.5 mg/kg/day
q-^ = 0
q2 = 3.0 x 10"3 (mg/kg/day)-2
q^J = 6.03 x 10'2 (mg/kg/day)-1
(104/61)3 = 4.96
q£ = 9.2 x ID'1 (ppm)-1
Geometric mean of 95% upper-limit:
q1 = 4.5 x 10"1 (ppm)'
q* = 6.4 x 10"1 ppm"1
6-49
-------
This risk estimate from animal internal dose must be converted back to
units of yg/m3 and ppm. The conversion factor at 25°C and atmospheric
pressure is
M.W. 1,3-butadiene 54.1 3
1 ppm =1.2 x M.W. air= I-2 x 28.8 m9/m
To convert from mg/kg/day internal dose in the mouse to low exposure ppm
in the mouse, the following additional information is needed: (1) the air
volume intake of a mouse per day, and (2) the percentage of 1,3-butadiene
absorbed by the mouse at low exposure levels. For a 35-g mouse, the volume
intake is estimated as
I = 0.0345 (0.035/0.025)2/3 m3/day = 4.3 x 10~2 m3/day
In order to determine the percentage of 1,3-butadiene absorbed at low levels,
the NTP (1985a) absorption study must be used (see Table 4-3). For mice exposed
to 13 yg/L (7 ppm equivalent atmospheric exposure), the absorption rate was
estimated as 54%. Assuming that linear kinetics hold under unsaturated con-
ditions, the mice at lower concentrations will also absorb 54%. In any event,
use of this figure will not cause a large underestimate of the risk.
Putting this information together, the conversion for the mouse for 1 ppm
continuous exposure is: 1 ppm = 2.25 (mg/m3) x 0.54 x 4.3 x 10'2 (m3/day) x
1/0.035 kg = 1.5 mg/kg/day as the internal dose. The final risk conversion for
the male mouse is
6.1 x lO'1 (mg/kg/day)-1 x 1.5 mg/kg/day = 9.1 x 10'1 (ppm)'1
ppm
6-50
-------
and for the female mouse is
qj = 3.0 x 1(T2 (mg/kg/day)-1 x 1.5 mg/kg/day =4.5 x 10"1 (ppm)'1
ppm
Since these male and female mouse data sets are so comparable, the geometric
mean, q-^ = 6.4 x 10"1 (ppm)'1, was chosen as the final 95% upper-limit
incremental unit risk estimate.
6.3.2.2. Rat-to-Human ExtrapolationRats exposed to 1,3-butadiene in the
Hazleton (1981a) inhalation study also developed tumors at multiple sites
(see Table 6-2). Among animals surviving the first year, the fractions of
male rats with at least one of Leydig cell tumors, pancreatic exocrine tumors,
and/or Zymbal gland carcinomas were 4/100, 4/100, and 21/100 for control,
low-, and high-exposure groups. For females with mammary gland carcinomas
only, thyroid foilicular tumors, and/or Zymbal gland carcinomas, the corre-
sponding fractions were 8/100, 46/100, and 50/100. All of the increase in
mammary gland tumors were carcinomas, not adenomas. These figures are shown
in Table 6-5 which presents the corresponding extrapolation results from
the rat to those produced from the mouse.
The internal doses for the rat estimated from the radio!abeled material
collected in the NTP absorption study (1985a) is estimated as 10.5 mg/kg and
37.1 mg/kg for the low and high exposure concentrations, respectively. These
doses were multiplied by 5/7 to determine daily equivalent continuous doses of
7.5 mg/kg/day and 26.5 mg/kg/day. Other calculations are similar to those for
the mouse data. The final 95% upper-limit incremental unit cancer risk estimate
for the male rat is q^ = 7.0 x lO'3 (mg/kg/dayr1, and for the female rat is
ql = 9.4 x 10'2 (mg/kg/day)-1. The 95% upper-limit estimates for the female
mouse and female rat based on internal dose are within a factor of three of
6-51
-------
TABLE 6-5. RAT: CALCULATION OF CANCER RISK BASED ON INTERNAL DOSES
AND CONVERSION TO LOW PPM CONCENTRATIONS.
FRACTION OF RATS (HLE, 1981a)
WITH AT LEAST ONE OF THE SIGNIFICANT TUMORS.9
ALSO, MLE AND 95% UPPER-LIMIT INCREMENTAL UNIT RISK ESTIMATES
BASED ON THE LINEARIZED MULTISTAGE MODEL
Nomi nal
exposure
(ppm)
0
1,000
8,000
Mai
Equivalent
continuous
internal doseb
(mg/kg/day)
0
7.5
26.5
es
Number
with tumors/
number
examined c
4/100
4/100
21/100
Females
Equivalent
continuous
internal dose"
(mg/kg/day)
0
7.5
26.5
Number
with tumors/
number
examined0
8/100
46/100
50/100d
aSee Table 6-2, Table 6-3, and text for tumor sites.
blnterpolated daily internal doses x 5/7 since treatment was 5 days per week for
a lifetime.
cApproximate number surviving first year and examined histologncally.
dThe highest dose group was dropped because of the poor fit of the model.
Maximum likelihood estimates:
Males
qg = 0.035
= o
Females
q0 = 0.083
q = 7.1 x lO'2 (mg/kg/day)-1
q2 = 2.8xlO"3 (mg/kg/day)-2
95% upper-limit estimate in (mg/kg/day)-1:
qt = 7.0 x ID'3 (mg/kg/day)-1
= 9.4 x 10'2 (mg/kg/day)
-1
Final estimate of 95% upper-limit in (ppm)-1: 1 ppm = 0.6 mg/kg/day
q = 4.2 x 10"3 (ppm)"1 q£ = 5.6 x 10'2 (ppm)"1
6-52
-------
each other; the 95% upper-limit estimate based on the male rat is significantly
less than the other estimates.
The final 95% upper-limit estimates, q^, in units of (ppm)""1, are con-
verted similarly to those for the mouse with the exception that the daily
volume for a 700 g rat must be estimated. This is
I = 0.105 (0.700/0.113)2/3 m3/day = 0.354 m3/day.
In order to determine the percentage absorbed at low levels, the same 54% used
for the mouse was used for the rat. Table 4-3 shows that mice and rats exposed
to similar concentrations (in ^g/L) retained very similar percentages of the
inhaled material. Therefore, it seems reasonable to assume that if the lowest
rat exposure was as low as the lowest mouse exposure, the rat would have retained
about the same 54% of the dose as did the mouse.
The final conversion from internal to low exposure external concentration
then becomes
1 ppm = 2.25 (mg/m3) x 0.54 x 0.345 m3/day x 1/0.700 kg = 6.0 x 10'1 mg/kg/day
and the final 95% upper-limit incremental unit risk conversion based on the
male rat is
q1 = 7.0 x ID'3 (mg/kg/day)'
and for the female rat is
x 6.0 x IP"1 mg/kg/day = 4.2 x 10~3 (ppm)'1
ppm
9.4 x 10'2 (mg/kg/day)-1 x 6.0 x IP"1 mg/kg/day = 5.6 x 10~2 (ppm)'1
ppm
6-53
-------
In comparison with the mouse data, Table 6-6 shows estimates based on the
female mouse is about eight times higher than those for the female rat. For
the males, the corresponding estimates were some 200 times higher for the
mouse than for the rat.
TABLE 6-6. COMPARISON OF THE ANIMAL-TO-HUMAN UPPER-LIMIT
INCREMENTAL UNIT CANCER RISK ESTIMATES FOR THE MOUSE AND THE RAT
BASED ON EXTERNAL CONCENTRATIONS
Sex
Upper-limit q-^ (ppm)"1
Mouse3 Ratb
Male
Female
Geometric mean
9.2 x 10-1 (ppm)-1
4.5 x 10-1 (ppm)-1
6.4 x 10-1 (ppm)-1
4.2 x lO-3 (ppm)-1
5.6 x lO-2 (ppm)-1
c
aFinal estimates, Table 6-4.
bpinal estimates, Table 6-5.
C6eometric mean not taken because male and female rat tumor responses were
not similar.
6.3.3. Comparison of Human and Animal Inhalation Studies
The purpose of this section is to evaluate whether or not the animal-to-
man extrapolated estimate of 1,3-butadiene-caused cancer is reasonably borne
out by human data. The section considers the limited human data base and
determines to what extent extrapolation from the positive animal data might
overestimate the human response.
While rat and especially mouse exposures to 1,3-butadiene caused a broad
spectrum of cancers, human response associated with the SBR process was neither
extensive nor consistent across studies. Various cohorts displayed excess
6-54
-------
mortality from cancers of the stomach or intestine, prostate, and/or respiratory
system. The most consistent excesses (and, therefore, the focus of this sec-
tion) appear to be restricted to cancers of the lymphatic and hematopoietic
systems, cancers which include leukemias, Hodgkin's disease, and lymphosarcomas.
It must be emphasized that exposure to 1,3-butadiene alone cannot be iso-
lated from exposure to several other potential carcinogens. Always associated
with the SBR process is concurrent exposure to styrene, a compound for which
there is limited evidence of carcinogenicity in animals and inadequate evidence
in humans (IARC, 1982). The small amount of human evidence associated with
styrene exposure and cancer suggests an association with leukemia and, possibly,
lymphomas. Styrene, like 1,3-butadiene, metabolizes to an epoxide; both epox-
ides are the suspected carcinogens. (Ethylene oxide, also an epoxide, is also
associated with leukemias.) In addition to styrene, the SBR process involves
numerous other exposures concurrent with 1,3-butadiene. These concurrent
exposures will not be dealt with in the following analysis, because if the
animal risk extrapolation based on 1,3-butadiene alone overestimates the human
risk, then the animal risk extrapolation will most likely be too high.
Probably the strongest evidence for human cancer associated with the SBR
process is that of Meinhardt et al. (1982), in which workers exposed to the
high-temperature batch polymerization process from 1943 through 1945 showed a
marginally significant increase in cancers of the lymphatic and hematopoietic
tissues, with an SMR of 212 from 9 deaths out,of 600 study members. For workers
first exposed after the process was changed to continuous feed in 1946, with
correspondingly less exposure, no deaths from lymphopoietic system cancers
occurred among more than 1,000 study members. Unfortunately, no exposure
estimates are available for the pre-1946 cohort. For the cohort exposed after
1946, only 1,3-butadiene measurements taken after 1975 are available. They
6-55
-------
show an 8-hour time-weighted average mean concentration of 1.24 ppm butadiene
(± 1.20 standard deviation [SD]), 0.10 ppm benzene (± 0.035 SD), and 0.94
ppm styrene (± 1.23 SD). (Benzene is not used in SBR manufacturing, but may
be present as an impurity of styrene or toluene.) The Meinhardt et al. (1982)
study also contained an analysis from a second plant whose workers were first
exposed in 1950. Based on a cohort of 1,094, the SMR for cancers of the lym-
phatic and hematopoietic tissues was 78, slightly higher than the overall SMR
of 66, the latter being significantly (p < 0.05) less than that of the compa-
rable general population. Meinhardt et al. reported average 1,3-butadiene
exposure levels in 1977 of 13.5 ppm.
The next strongest evidence for cancer associated with the SBR process is
based on the case-control study of McMichael et al. (1976). These authors
estimated an age-standardized risk ratio of 6.2 for lymphatic and hematopoietic
cancers among workers with at least 5 years of exposure in the synthetic plant,
relative to all other workers as controls. (This ratio decreased to 2.4 when
a matched control analysis was used.) The synthetic plant is where the SBR
process is located. McMichael et al. also found a dose-related risk ratio in
the synthetic plant by number of years worked there.
Estimates of exposures in the McMichael et al. study are based upon a
later paper. Checkoway and Williams (1982) measured 1,3-butadiene, styrene,
benzene, and toluene levels at the same synthetic plant in which McMichael et
al. (1976) found that "leukemia and lymphoma (cases) among hourly paid rubber
workers from one company were 6 times more likely than controls to have worked
at jobs in the SBR plant." Exposure levels of 1,3-butadiene typically averaged
below 1 ppm, but exposure levels in the tank farm area averaged 20 ppm.
The most extensive investigation specifically designed to study the health
effects of the SBR process shows very little association of 1,3-butadiene with
6-56
-------
lymphatic and hematopoietic tissue cancer (ICD 200-207). The Matanoski et al.
(1982) study of one Canadian and seven U.S. synthetic rubber plants showed,
possibly, a trend with more exposure as defined by production, maintenance,
utilities, or.other jobs, but none of the SMRs (Table 6-7) are statistically
significant. Only Hodgkin's disease (ICD 201), shows a consistently high SMR
in all three cohorts, but the numbers of cases were small. No exposure esti-
mates were presented in the Matanoski et al. report, but the plants studied were
of the same type as those studied by Meinhardt et al. (1982). Some workers in
seven of the eight plants might have started as early as 1943, and Matanoski
states (personal communication) that the batch process in several of these
plants was continued into the early 1970s. Based on these observations, the
estimates of 4 ppm for production workers, 3 ppm for maintenance workers, and
1 ppm for utility workers have been used for calculation purposes. It is em-
phasized, however, that none of the estimates are based on contemporary measure-
ments.
A February 1985 review draft of this document presented exposure esti-
mates for the Meinhardt early (1943-45) Plant A cohort and for the Matanoski
study that were five times as high as the present estimates. These exposure
estimates have been reduced based on comments by industry representatives to
EPA's Science Advisory Board that actual human exposures in the epidemiologic
studies were considerably lower, both because of the olfactory threshold and
1,3-butadiene's explosive properties. Accordingly, estimates for these two
study groups were reduced to be more in line with these comments. Estimates
for the other Meinhardt groups remain the same since these were measured concen-
trations in the plants. The uncertainty of all these estimates should be
stressed. Any increase or decrease in exposure estimates causes a porportional
change in predicted cancers.
6-57
-------
£ <
*"* jj
-C O
IP
^aec>
co co ^
coaco:
O O ±3
HEIWOPOIET
S. CONFIDE
IHCREHEHTAL
CO
^^ It t **
ol*
rgSiP
QL. CO a!
>" OC
il!"°
°ujS
*** z S
If 5>c/>
o ic
o 23 5
of S f*
S5J2
5 *? ><
UJ O
ecu
Tpo
10
*> O UI -O
t- 01 t- .e ^
20> re it o
O t- flj B
CU *J CLT3 O
ui aj 5
.C i. t-
4J 3 <<-
re ui re i re
a o TJ *J -o*
xi CL 01 ro oj -a
X 4-> T3 >
U CL
re +>
0
CD
"g-N
B T-
03 UI
CO
f
to
u , a;
/t ^ j_i
01
re
f=
t- b e re -i-
o .c -p ex E P
I CL i ex
VI CO CXVO CL
re «-< »-H
.j=E«o |CM
ui c. en < en
LLj VI rTf V> t 1
VO
u
*
to
en
o
CM
o
i
to
en
o
CM
r-l
CM
to
CM
ex ui o
re o ai CM
.C J-> 3 1
ex re ui o
E E ui o
>, -^-^
C3
0
VO
4J IO
a «*
t- r £ en
x: cvT+j i
C 4J OO CCO
cu ^-i i en
3E J- ^O- 9~t
II
01
x
«S 0) --H
»«
. .
J
1
to
C3
CO
CM
0
«*
to
t-H
0
0)
o
.a
re
UI
re
1
CM
VO
0
*
v-t
r-
*2
O. r-4
-M » *-* 4->
OJ t/1 -Q 4-> «* OJ U)
o> a.
TJ T3 4J W
> "O C- «3 * '
fOt/> -|J*r-J^E-r- S
c- o in aj£='OC34-> °-
OJ CL E
> X CO (Or
,.^ CM re
cu o cu
3 CX^^ CO
en co
VO CO
CM VO
co" co
01
* C VJ
o o c
. ^- -r re
_O CD -4-> C
~ O .C O Ol
CM -r-j 3 <->
CO +J T3 C
en >, ui o T-
»-4 ^ re t- re
i o- !E
UI
(U
CL
ex
o
CM
to
CM
CO
O
00
to
0
CM
VO
O
to
VO
^*
>
O
re
VI
re
re
CO
o
to
to
Itilities
_J
1
o
n M
1 C7I t- CU
EC .- 4J i.
uioxoi exreuicxE ^^
ui -r-c ait oait-E
(05e'xc i '^'X'XH-CL
U) -^ _n-r O O> O CU CU _ ^l
UI c JJ ! 4-» UJ t= t. %>J
re~o o.c cu oicxre ^-^
L_ 4-> C CU CL'r 4^ C. re
C'f-OCUCT->»> OL.
*-< o a.o-*-ti- uiccvvj- re
!
1
1
UI
CU O
^ o .c 4-> re t-
ui *i~ 4-> cr E -*-»
r- +j c re * c
<_ re >>r c o
' C. UI CXCM (- U
u
»->
re o en
0.0
U O CM
4-> 03 O
re EO
.C OJ CM
J3
en
»-H
^
c
o
VI
UI
o
CL.
o
c
0
Ijj
3
X
o
CL--I
ex co
re en
t-H
t- 0
O r
c ui
a)
C- t-
O CO
*-> "O
UI C
ai re
P
c c
o o
ul E
UI 3
i- re
o o>
CL CO
oto
OUJ
5"-"
01
<_ 1
[ft
t- E
CO
c
01 CM
0) 1
£M°
u
i ca
4-> 1
C <-H
re M
u
r- re
r 3 "O
c e ai
C7I L. >
f- O t_
Ul V>_ 0)
U)
p >,^
O .O O
C
T3 C
ui oi re
t- C J=
0) -r- 4->
o e
c e -a
03 01 01
O +J -U
01 U
-O T3 T-
> L. 01
i. 01 t-
01 3 CL
(/I O
J3 CL Ul
O -C
t- LO re to
O O O) O
u- -a
o o
p 0)
I/I II L. V
IV< f\ 1 \ M
w*
-------
Is the unit risk estimate.based on the mouse tumor data a reasonable
extrapolation? To answer this question we must be able to estimate the predic-
ted effect based on actual (or estimated) exposure. (We must also assume the
effect on these cancers from other exposures to be nil.) For illustration, we
choose the Meinhardt et al. (1982) study plant B, with estimated exposure of
13.5 ppm. Although this exposure was measured in 1977, we will consider it to
be representative of exposures from 1950 through 1976. We must also know that
the 1,094 study members converted to 19,742 person-years at risk for an average
of 18 years per person. Since average employment or exposure is given as 10.78
years, we can estimate the expected contribution as follows:
1. The continuous lifetime equivalent exposure based on 10 working years
out of about 50 possible remaining years is:
13.5 ppm x 10.78 years x 240 day x 8 hours = 0.64 ppm
50 365 24~
2. The inital best estimate (MLE) of risk based on the average male and
female mouse data for a 1 ppm continuous exposure for approximately
40/70 of a lifetime is q1 = 2.5 x 10~2 (ppm)"1.* This geometric
mean MLE incremental risk estimate will be used to predict human
excess deaths for the present purpose of deciding whether the number
of deaths among industrial workers is consistent with the expected
deaths derived using the animal extrapolation procedure. The initial
estimate (Table 6-4) is used since it projects the limited animal
1/2
Llf
*From Table 6-4: [(9.35 x 1Q-2) x (3.0 x 1Q-3)] x 1.5 = 2.5 x 10-2 (ppm)-l
This estimate is valid only in the 1 ppm exposure range. (An arithmetic mean
yields an estimate nearly three times as high.)
6-59
-------
observation period (60-61 weeks out of a 2-year lifetime) to the
limited human observation period (average of up to 29 years out of 50
years remaining lifetime).
3. Based on the MLE* unit risk factor qx = 2.5 x 10"2 (ppm)"1, each of the
1,094 workers would be expected to have an additional lifetime risk .of
R = 2.5 x lO-2 (ppm)-1 x 0.64 = 1.6 x 10'2
4. Based on 1,094 workers at risk for 18 of their 50 remaining years,
this converts to the following expected excess number of cancers:
1.6 x ID'2 x 1,094 x 18 = 6.3
50
5. Adding 6.3 to the 2.55 cancer deaths expected based on no exposure
(Table 6-7), we could expect, with the exposure, to observe 8.8
deaths from cancers of the lymphatic and hematopoietic tissues. The
probability of observing two or fewer deaths with 8.8 expected is
2 > x
P (deaths <_ 2|x = 8.8) = e-' x = 0.0073
X=0 X!
or p < 0.01.
The statistical power to detect a predicted SMR of (8.8/2.55) or 3.5 is
given by Beaumont and Breslow (1981) as ZI-B = Za - 2 (SMR0-5 - 1)E°-5. For
6-60
-------
plant B this is
= 1.96 - 2 (3.5°-5 - 1) (2.55)0-5 = -1.75
which corresponds to a power of 0.96, or 96% at a = 0.05. This means that the
study was powerful enough (p = 0.96) to detect the 6.3 predicted deaths at the
p = 0.05 level.
Results based on similar calculations are presented in Table 6-7 for the
other two Meinhardt cohorts and for three Matanoski cohorts. They show incon-
sistent results. For the 1943-1945 plant A Meinhardt cohort, the deaths pre-
dicted from animal extrapolation actually underpredict the observed human
response (p < 0.5). For one other Meinhardt cohort and the Matanoski utilities
cohort, the predicted and observed results are not significantly different,
although the power to detect the predicted difference in these three cases is
low 7% or less. For the two larger Matanoski cohorts, the extrapolated deaths
do significantly (p < 0.05 in both cases) overpredict response. However, if
as suggested in the review of this study, that underascertainment might have
missed approximately 17% of the deaths, then neither of these results from the
two larger Matanoski cohorts would have been statistically significant.
The interpretation of these results, if we dismiss momentarily the large
uncertainties in the exposure estimates, is that the predicted deaths are con-
sistent with the observations. In fact, no predicted results will satisfy the
observed results in all six cohorts. Were we to lower the risk estimates to
try to better accommodate the Meinhardt Plant B cohort, we would further under-
predict the observed results for the Meinhardt early plant A cohort. Based on
the information we have, no single extrapolated value can predict the human
response. Considering the uncertainties in the human exposure data, and the
6-61
-------
ascertainment in the Matanoski study cohort, the estimate based on animal
extrapolation is consistent and the best that can be achieved at present.
Finally, the same analysis as computed for lymphatic and hernatopoietic
cancers in Table 6-7, can be done for all cancers on the theory that 1,3-buta-
diene might be a broad-spectrum carcinogen in humans as it is in mice. This
analysis is presented in Table 6-8. Since we have used the same extrapola-
tion from the mouse data, the same number of excess deaths as predicted in
Table 6-7 will result, but these excess deaths are spread out over all cancers.
It should be noted that only one of the SMRs is statistically significant and
that all, in fact, are less than one. Based on the predicted excess deaths
extrapolated from animal data and estimated exposures, two of the cohorts ex-
perienced significantly fewer deaths than predicted. For one of these cohorts,
the Meinhardt plant B, the deficit in observed versus expected deaths was
significant even if there were no predicted deaths. For the other, the
Matanoski production workers, an underascertainment of 16 deaths (17%) would
more than explain the statistical significance at the p = 0.05 level (one-
sided).
Comparing Tables 6-7 and 6-8, we see fairly similar results: weak, if
any, evidence of a human carcinogenic risk from 1,3-butadiene, but also no
strong evidence that the unit risk extrapolation from animal to human results
is unreasonable, or that it seriously overpredicts a potential risk.
6.3.4. Relative Potency
One of the uses of quantitative estimation is to compare the relative
potencies of different carcinogens. To estimate relative potency, the unit
risk slope factor is multiplied by the molecular weight, and the resulting
number is expressed in terms of (mmol/kg/day)-l. This is called the relative
potency index.
6-62
-------
5
a
in
C£ LU
H 00
OO 3
O
1 = LU
:n o i
t-H u_
3i°
O OO
LU OO i^
h- 1 00
LU
LU CO O
_l 1 X
CO LU LU
1 LU O
C£. LU
00 rH
O
LU
C£
a.
tn
4J
c
E
O
o
4-» "O
O OJ
OJ 4J
4-> o LO ja
I- OJ -r- JC
OJ "a T3 4-> LO
gOJ IO II O
O t- 0) O
Cu 4-> Q.T3 -O
(/) OJ O
J= J- t-
tO (/) (O I (O
-o*
o a. aj 10 oj "D
X 4-> -0 > OJ
VJ OJ O t- 4->
t/t -r- QJ OJ U
O) E "O (/> tn OJ
O O OJ 3 J3 a.
X C. C. O O X
LU If- O. E OJ
OJ {/>
O r
C 10
»* T> t.
LO -r- OJ
cn <*- +-»
c c
O *r-
U
C£
00
OJ
U> LO OJ
t- O O.
OJ CM X
O 1 I*'
c o
O rH
I O OJ
OJ
{/I
0
r- OJ
rcl *u»
Crt
i a. i a.
tn co 0.^0 CL
+J CT^ ^- O1 CM
(O t 1 t 1
4-> o ^ o ir>
t/) I O*> C. O"i
LU tf- *-4 14- * )
R
a-
l-H
i
cn
o
co
l-H
l-H
1
CO
I-H i cn
Z: ( -0. rH
OJ r
I/I IO
tr- 0 E
0 JC 4-> 0.
II
OJ
Ol i_
cn 3
10 tn
O) Q.CM
> X
«C OJrH
^
rH
CM
to
O
rH
LO
rH
1
LO
3-
O
CM
CO
C3
cn
CM
o
l-H
CM
O
rH
1 4->
EtO C
r^S r^
O_ rH Q_
IO
t=
Ol
1
4-> O) O. IP- -r-
0) Crt JD 4-5 ^^ OJ
OJ O. IO T3 4J
, -a «... «
10 wj 4J «i jz; E -
> X CO IO r 4-> OJ wl OJ
4->
cn ro aj
-H z:
a*
to
rH
to"
1
O
0
rH
cn
o
rH
1
CO
P-.
o
cn
CO
o
to
LO
0
i-H
en
cn
to
CM
CO*
-a o
~t^
co 4J -a
cn >, m o
-_- r- Cl-
*-f
(/)
E
a.
a.
CM
to
co"
CO
CM
LO
rH
rH
rH
1
CM
CO
O
LO
cn
C3
CO
to
rH
CO
to
l-H
CO
co
to
A
CO
OJ
Maintenanc
**
OJ
§.
CL
f-H
8
CO
rH
1
CO
CD
g
CM
1
CO
o
LO
en
o
CO
CM
LO
CM
O
LO
LO
Utilities
c
o
u)
o
CL
OJ
O
c:
O
+->
IO
r~
X
o
Q.I-H
O.CO
10 cn
"^
E £
c3°
C tfl
^
I CO.
C: rH
10 h»J
o
r- IO
"*- i
r- 0 1-
tn v- OJ
1^1
a c
v» aj 10
t- C J=
OJ.r- 4->
U E
c: «- -o
IO OJ OJ
OJ O
a TJ -
OJ T3
* OJ *
a> 2 o.
5
.a ex tn
o jr
« 4->
t- LO IO LO
O O OJ O
M- -O
0 0
4-> 0)
-------
Figure 6-1 is a histogram representing the frequency distribution of
potency indices of 55 suspect carcinogens evaluated by the CAG. The actual
data summarized by the histogram are presented in Table 6-9. Where positive
human data are available for a compound, they have been used to calculate the
index. Where no human data are available, animal oral studies and animal
inhalation studies have been used, in that order. In the present case, only
the animal inhalation studies provide sufficient evidence of carcinogenicity
and have sufficient exposure information.
The potency index for 1,3-butadiene based on the NTP mouse inhalation study
(NTP, 1984) is 9.9 x 10-2 (mmol/kg/day)-l. This is derived as follows: the
upper-limit incremental unit risk estimate from the inhalation study is qt =
6.4 x 10-1 (ppm)-1. Transforming this to mg/kg/day, the conversion factor for
1,3-butadiene in humans (assuming 54% absorption at low concentrations) is
1 ppm - 2.25 mg/m3 X 20 m^/day x 0.54 x 1/70 kg = 0.35 mg/kg/day
Then
-1
6.4 x ID'1 (ppm)-1 x 1 ppm/0.35 mg/kg/day = 1.8 x 10° (mg/kg/day)
Multiplying by the molecular weight of 54.1 gives a potency index of 9.9 x 1()1(
Rounding off to the nearest order of magnitude gives a value of 1Q2, which is
the scale presented on the horizontal axis of Figure 6-1. The index of 9.9 x
IQl lies in the third quartile of the 55 chemicals that the CAG has evaluated
as known or suspect human carcinogens. Thus, in terms of potency alone, 1,3-
butadiene would place in the lower half of these carcinogens. However, the
fact that 1,3-butadiene causes so many fatal tumors in animals and sharply
6-64
-------
>-
o
z
UJ
o
UJ
20
18
16
14
12
10
8
6
4 |
2 I
0
4th 3rd
QUARTILE QUARTILE
2nd
QUARTILE
10"1
1 2 34 56
LOG OF POTENCY INDEX
1st
QUARTILE
2 x 10 +3
I
8
Figure 6-1. Histogram representing the frequency distribution of the potency
indices of 55 suspect carcinogens evaluated by the Carcinogen Assessment Group,
6-65
-------
o
OX
O)
o e
ro i t-
e«
tu
oo
x
C (U
O) T3
01
CO
CO
CO
CO
CO
CO
O 3
=> 00
_i as
o
CO
in :E
in
CO -O
z: uj
(O
UJ OO
i-« cC
CJ
r TD
O)
a>
CO
6-66
CO
CO
r- O
-------
f
T3
CU
Z5
C
5
tr
O
o
CTl
to
UJ
l
03
1
CU
It- T3
0 => ^
4-> OX
t_ *r- . 1 gj
CU CT CJ1T3
"O O5 O C
t- re i -i-
O E ~
>>O
O X
cr cu
CU T3
-M C
O !-
Q.
£_
. re
O O1
CU !-
i CU
21
1 1
1
">»
-Q (0
CU "O
o en
CO
en
s
cn c re
CO T-
- o t-
O. "O C£ CU
3 CU
O CO 1 1 -r-
t- re t-
C3 -O O
10
cu re
0 E
i cu c:
co -a <
> -I
cu >
J CU 10
M- re
o £
3
3:
CU
-Q
3
^:
CO
e£
O
to
C
rs
o
CL
0
O
i i O ' ' < 1
+ 44-
i l
O O 4- O
CD O O O
i 1 i li l i l
X X X X
O1 00 00 OO
Cn I * Cn "^J"
co 10 t~^ r*~>
en oo to oo
CM i 1 1
CM CM
CM 1 1
1 0 0
O r-) 1
.-i X X
X CM O OO
cn i i o LO
CO
CM OO CO OO
CO _J _J _1
1 II II II 1
CM i i in in
till
tO CM * O
CD 1 * OO CD
II II
p^ P-~ CT^ Ot
CD to r~- r^.
i i
CO
re
+> CO
cu c
o re
CU i- JC
co c: o +J
cu re co i cu
re +j re o t_
jr cu jc: re o
4-J O +J £_ i
CU £_ CU 4-> -CT
o o cu o
T3 i t_ 1 -r-
CU _E O 1 t-
+J O i CM (
re T c~ *» i
c: Q o CM CM
r^ I re ** *
£- CM X t-< t I
o « cu * «
i t i jr: t i I-H
-C
C_3
, l
4-
i i
0
I 1
X
1 1
<3-
o^
t 1
1 1
CM
1
CD
i 1
X
i l
00
CD
CM
CO
i i
oo
i
i
to
£
O
H-
o
o
-sr
c_>
4-
oo
1 I
X
0
0
r- 1
^
3
>*
T-4
CO
CO
oo
1
1
o
J^^
f 1
s»
E
3
1
o
5_
CM
4-
CM
. CD
r- 1
X
1 1
in
LT>
OO
OO
O
CD
CM
CO
i~~i
OO
en
CM
i
o
LO
I
a
a
oo
4-
CM
4-
0
, t
X
«=t-
< 1
oo
in
CM.
S
i-H
CQ
CM
OO
, ,
t l
CTl
I
cn
cu
M
cu
_Q
o
C-
o
o
o
CM O
4-
CM
4- 0
0 0
^_f ]
X X
r-l ^-<
cn
cn oo
14
CM
' O
^^^f ^_)
^>O X
t-H -3-
F-H ^H
03
OO CM
1 CO
i i i i
^± CM
LO cn
oo o
i |
LO LO
'a? ,*
-a cu
QJ *r T"J
C 5_ -r-
0) 0 C_
i i O
>!-C r
-CO CU _C
cu cu re
o c _sr cu
t- CU 4-J C
o ~o cu cu
!c: >r~ o **^
t- C O 4-J
Q -i i CU
1 1 ^ ' O d-
<-< Q
«* CM
4- 4-
* i l
4- 4-
0 0
^H - i 1
X X
,-1 IO
cn
O CM
oo oo
00 rH
oo
o
oo o
CD CO
CM CM
CO CO
i t CM
1 1
LO r-<
0 r-l
tO CM
t 1
CU
c
cu
o
o
*-*
C T
i C
<- -I
"O f~^
1 t
CD. CM
CM O
CM ,-(
4- 1
0 0
t-H i-H
X X
I-H cn
ur>
O CM
oo cn
i i
CO
o
i i
r-~ x
i cn
CD cn
CO CO-
CM CM
CO CO
r^ oo
1 1
to cn
to oo
1 1
CM to
CM O
Ii , 1
cu
r- C
tM !-
re £_
!- T3
a >,
-C O
l £_
>5 O
C r
cu _c
-C O
Q. T-
r- Q-
CD 1 1 1
CM
4-
CSJ
0
1 1
X
CM
00
T 1
1 1
. 1
CO
CM
CO
^
1
1
r-H
1 1
(-
cu
JZ
cu
i>
+J
cu
0
o
r
-C
o
CM
V)
CO
1
1
1
r~^
[cu
|C7>
rO
CL
o>
c
3
0
"o
c
o
-o
cu
=3
o
o
6-67
-------
r- ^
o
OJ
i
1
o
o
Cft
«S
Ul
i
cu
<*- -o
03 *-»
4-> OX
(_ M- 1-4,O
C O)
OJ TJ
4J C
O'r-
0-
c_
r- .C
3 0>
O'l-
0) «^ 4-*
(/) t 4 !
CD /** O
in
0) -r-
_J (1) to
i+- nJ
O £
3
OJ
|
O
Compounds
vo
+
10
o
X
r-4
« 1
H- 4
o"
o
CO
-
CO
00
oo
CO
1
CM
' «=»
LO
Bis(chloromethyl)ether
..j.
4-
CO
+
O
X
oo
p-
co
r-l
T 4
«=!
'co
CM
CO
i i
1
CO
cn
ID
0
r-4
CO
Ethyl ene di bromide (ED
i4
4~
,_,
0 ,
X
CM
r-4
tt ' '
O
f"^
X
LO,
co
CM
oo
_J
CO
t 4
CM
LO
Ethyl ene oxide
CO
-f-
CO
o
X
r-4
CO
CO
1
CO
CO
CO
CO
CM
00 ,
1 1
CO
1
ID
Heptachlor
co
rf-
CM
-t-
0
X
LO
=1-
00
CM
1
IQ
r-l
CO
CM
CO
. t i
r-l
1
OO
t4
i 1
Hexachl orobenzene
t4
t-
r-H
O
X
CM
< 4
CM
CM
|
O
I 1
x
LO
6
co
I
H-l
co
CO
ID
1
00
Hexachl orobutadiene
co co co co
4- -1- + +
CO CO CM CM
O O CD O
t 1 r 1 r-H t 1
X X X X
r-l CO LO <
cn cn cn cn
0000
cn cn cn cn
CM CM CM CM
CM
LO r-l * CO
r co co
«=1- r-l r-H r-l
CO
CM CO CO
CO _1 _1
1 II II 1
LO [*"* cn
i i i
«^f LO cn
CO CO CO
i i i
cn cn co
t 1 r 1 LO
CO CO
Hexachl orocycl ohexane
technical grade
alpha isomer
beta isomer
gamma isomer
LO
"*"
UD
O
, 1
X
CM
r-l
cn
CO
CO
0
r 1
X
CM
LO
CO
CM
OO
i i
00
LO
LO
LO
CO
c
Hexachl orodi benzodi ox~
CM CO
-r -r
CM CM
O O
i 1 t 1
X X
LO 0
CM LO ,
CM CM
O O
CM CM
*^ * ^S
LO
CD r-4
r-H CM
r-H r-H
OO 00
oo oo
CM"
CM
1
1
LO
CO
1
O
CM
i-H
O
Nickel refinery dust
Nickel subsulfide
CO «3" CO CM *
CO CO CM CM CO
O O O CD O
t It It It It 1
X X X X X
CM «=}* Cn CM ^f
t 1 t 4 CM CM t 1
«* CM co o r-~
t-» O LO O i4
I 4 r-4 t 1 i 1
^ . --^
*r-4Xt-H
cr cr
-O JD
4-> 4->
0 0
c c:
cn LO co co cn
^t- i i
LO CO CM
CM "3r LO CM CO
CO CQ CQ CO CO
CM CM CM CM CM
CO OO OO OO OO
cn LO co CM cn
1 1 1 1 1
LO CO
-------
T3~
CD
Z3
C
r
4-5
C
o
o
.
T
l£>
LU
1
CQ
1 .
CD
0 H
t- £ Sx"
CD C Ol CD
"O cn o "O
£- ra , cr
O E v r-
O X
C CD
CD -a
4-> C
O !-
0_
£_
ra
Z3 -C
o 01
CD T-
i CD
^
1 1
'
-Q ro
CD T3
CL-~~.
O O)
oo -^
Ol
E
Ol c ra
CZ O T-
T- o £_
CL "O C£ CD
ZS CD .< jj
O to i i T-
f- 03 i.
CD JO O
in
ro i
CD ro
0 E
r- CD C
CD -a =c
> ip-.
O> >
_i
O
CO ^t"
CO CO
CM CM
00 00
* 1 h t
in 10
CO CO
cn co
=3- 10
OO OO
rO C
a> T-
£_ E
Z5 CO
'>>'>>
N-nitroso-N-meth
N-nitroso-diphen
CO . r-l
CO
+ 0
0 O
r-l i 1
X X
*
CM cn
CO r-l
CM
0
r-l
X
co cn
"**t" r 1
CO CQ
CM CM
00 OO
r-H | |
CO CM
1 1
CO ldT>
CO O
i
rH
CD
VO
t
I 1
Q
a
o
t
X
0
a
CL
i i
*^*
o
o
X
00
00
in
10
i i
CM
1
0
i I
X
1 1
LO
CO
-
1 '
1
OO
tl
r-.
CM
T 1
CM
trachloroethyleni
m
l
CO
+
CM
0
X
LT>
i i
co
i i
CO
CM
CO
*~~l
CM
1
in
CO
1
i I
O
CO
00
CD
C
CD
-C
CL
ro
O
O
0
0
X
1 1
^J-
t 1
CO
1 1
CM
1
O
1 1
X
1 1
r I
CO
CM
CO
oo
^
r *
t 1
0
cn
ichloroethylene
£_
1
O
0
0
X
< 1
in
CM
1C
f ^
t i
CM
1
0
r-l
X
Lf>
r 1
. 1
OO
OO
1
1 1
o
1
lyl chloride
r
r 0*
ro o Jc: t_
c n: 4J 3 aj
O 4-> . CL
"O O 4-> O CL
CD ro CD Z3 CD CD
"1 f~5i M CD O)
-O - - 1 Z5 > CO 40 CZ -r-
, O> CD tO C T-
~a c- -o ai -r- 4- o to
"3O ~rO4->rac:OO
i CXT3 O O i
ZJXr- r- OJL.r-ctO
uajor jzrcx4->ro
rO i to rp t_l s c-
O ro O) 4-J T- jr 4->
c t- . ra -a i
a| o jz: >,jz: c Q. ro cn
rO 4-> C C 0 -r- tO -^
rO O T- CD JO T~ >~
>i CL C ro to CD i
h~* O ro QJ to 3 c i ^
o a> o o t- ra r
C CD r O 4-> Z5
1 ra i O 4-> ro _Q
^ E c rs >,
O jz: jz: aj -r- 4_> .,_
E -4J£_ -r-4JCDT3
onzf 3 ro _rz CD
01 o -a zs « o ro
rO >. O £- a> 4-> r
4-5 x ^ Tri cu o CD t_ ra rs
tOCCDE 3CLJCO
CDT-OtorO!_OQ.4j,
o +j T- re «« to i ra ra
cz i 4-> JD CL -a a
CD Z5 ra CD CL-C C O
"o E i ~ ro "" CD i "O
NT- EC CDt-CDc
CD-I ! CD to jr> -r .>*: ra
+-> t- i 4-> tO T- ZS T-
ro ro ro i/l CD >, err t i
O^CZ-r^ CLZ3Et_CU *~*
"Or rac£-'raT3t04->ra'
rO "p- > C cr T"J
CCD OCD raot_-^.T3
i i -C i i ex, CD -i-ooic:
4-3 JO CL"O 4-> *4 _^i n
II >,CDraocra ^^O
' o ra 'i/ior^'ai'oE'
"O to JO ro O E O
** ^3 CD in 'I QJ > cr r~ (3
CDCU4_)QJj^;_c-j_^jQj ^»
1 CD -o qt
czrooo Q-j-^-t cjrr
O> JO !- r C CLtO>T-4->
' "O "O to T- Z3 T- CD
t- to C CD t_ CD <*-
> cu -r- cr -a i CD 'tojrr
cDtoorccu-r-T3cujrr 01
4->J= torocD4->Oi<4-.r-
i- ZJ t- - - CL X !- 2 |
IOOCD tOCU4_)= T3C-
J3 rJu
. " CL O X CD JD r ' O to 2 r
CUZ3XCD4-> CU r ra o
O CD OCDJsitO rOE
C»« t-CLO-r- r-
cum «QJ c r CD to CD
"O cn t/j 4_s tj rD jz; cu »p c
> (D -i 2 C -r- O X+J
CUS--O CU>Clr-tOO)>,
40 ^^ c?0 " tO ° '0"fi^
C tO tO T- C 3 >r- CU 4-> -r-r-l
CD CD j-i T- CD to CL t
CLI c o cr --^ zj CD >^.-
o o ra T- c. croo>,
oo ro ro ro jrr cn c »-~o r-> s^
OO eC ro O O -r- 1 JO 1 -
ra JO o
6-69
-------
decreases the latency period increases concern beyond that based simply on
relative potency.
Ranking of the relative potency indices is subject to the uncertainties of
comparing potency estimates for a number of chemicals based on different routes
of exposure in different species, using studies whose quality varies widely.
Furthermore, all of the indices are based on estimates of low-dose risk using
linear extrapolation from the observational range. Thus, these indices are not
valid to compare'potencies in the experimental or observational range if line-
arity does not exist there. The uncertainty of the estimate of low exposure
potency for 1,3-butadiene is increased because the response at 625 ppm for the
mouse is greater than 60%, which is probably not on the linear part of the
dose-response curve.
6.3.5. Summary of Quantitative Estimation
Based on. the linearized multistage model and an external concentration to
internal dose conversion, a 95% upper-limit incremental unit cancer risk of
q* - 6.4 x 1CT1 (ppm)-1 was calculated for 1,3-butadiene using the geometric
mean of the 95% upper-limit incremental risk estimates from, the pooled male
and pooled female significant tumor responses of the NTP mouse study. The
quantitative estimates based on mouse-to-man extrapolation were then used to
predict human responses in several epidemiologic studies, and the predicted
and actual responses were then compared. The comparisons were hampered by a
scarcity of information in the epidemiologic data concerning actual exposures,
age distributions, and work histories. In addition, because there was no
consistent cancer response across all of the studies, the most predominant
response, cancer of the lymphatic and hematopoietic tissues, was chosen as
being the target for 1,3-butadiene. Based on the comparisons between the
predicted and observed human response, the extrapolated value from the mouse
6-70
-------
data was consistent with human response, but in view of all the uncertainties
and apparent inconsistencies in the epidemiologic data, a fairly wide range
of potency estimates and exposure scenarios would also be satisfactory.
In addition to a 95% upper-limit incremental unit risk, a measure of car-
cinogenic potency was determined for 1,3-butadiene. Among the 55 chemicals
that the CAG has evaluated as known or suspect human carcinogens, 1,3-butadiene
ranks in the third quartile. Based on the wide spectrum of cancers and sharply
decreased latency associated with these cancers, however, 1,3-butadiene should
evoke more concern than the potency numbers alone indicate.
6-71
-------
-------
7. REFERENCES
Allen, J.W. 1985. Unpublished data from FDA-EPA Interagency Agreement on
Cytogenetic Effects of Medical Device Chemicals (Monomers). Personal
communication.
American Conference of Governmental Industrial Hygienists (ACGIH). 1983.
Threshold limit values for chemical substances and physical agents in the
work environment, with intended changes for 1983-84. Cincinnati, OH.
American Conference of Governmental Industrial Hygienists (ACGIH). 1984.
Threshold limit values for chemical substances and physical agents in the
work environment, with intended changes for 1984-85. Cincinnati, OH.
Anderson, E.L., and the Carcinogen Assessment Group. 1983. Quantitative
approaches in use to assess cancer risk. Risk Analysis 3:277-295.
Andjelkovich, D., J. Taulbee, and M. Symons. 1976. Mortality experience of a
cohort of rubber workers, 1964-1973. J. Occup. Med. 18:387-394.
Andjelkovich, D., J. Taulbee, M. Symons, and T. Williams. 1977. Mortality of
rubber workers with reference to work experience. J. Occup. Med. 18:387-
Autrup, A., F.C. Wefald, A.M. Jeffrey, H. Tate, R.D. Schwartz, B.F. Trump, and
C.C.Harris. 1980. Metabolism-of benzo(a)pyrene by cultured tracheobron-
chial tissues from mice, rats, hamsters, bovines, and humans. Int J
Cancer 25:293-300. " *
Beaumont, J.I. and N.E. Breslow. 1981. Power consideration in epidemiologic
studies of vinyl chloride workers. Am. J. Epidemic!. 114:725-734.
Bolt, H.M., J.G. Filser, and F. Stoermer. 1984. Inhalation pharmacokinetics
based on gas uptake studies. V. Comparative pharmacokinetics of ethylene
and 1,3-butadiene in rats. Arch. Toxicol. 55:213-218.
Bolt, H.M., G. Schmiedel, J.G. Filser, H.P. Rolzhauser, K. Lieser, D. Wistuba
and V. Schurig. 1983. Biological activation of 1,3-butadiene to vinyl
oxirane by rat liver microsomes and expiration of the reactive metabolites
by exposed rats. J. Cancer Res. Clin. Oncol. 106:112-116.
Carpenter, C., C. Shaffer, C. Weil, and H. Smith. 1944. Studies on the
inhalation of 1,3-butadiene; with a comparison of its narcotic effect
with benzol, toluol, and styrene, and a note on the elimination of
styrene by the human. J. Ind. Hyg. Toxicol. 26:69-78.
Checkoway, H., and T.M. Williams. 1982. A hematology survey of workers at a
styrene-butadiene synthetic rubber manufacturing plant. Am. Ind Hva
Assoc. J. 43:164-169. , * y"
7-1
-------
Checkoway, H., T. Wilcosky, P. Wolf, and H. Tyroler. 1984. An evaluation of
the association of leukemia and rubber industry solvent exposures. Am.
0. Ind. Med. 5:239-249.
Citti, U, P.6. Gervasi, G. Turchi, G. Bellucci, and R. Bianchini. 1984.
The reaction of 3,4-epoxybutene with deoxyguanosine and DNA in vitro:
synthesis and characterization of the main adducts. Carcinogenesis
5:47-52.
Conner, M., J. Luo, and 0. Gutierrez de Gotera. 1983. Induction and rapid
repair of sister-chromatid exchanges in multiple murine tissues in vitro
by diepoxybutane. Mutat. Res. 108:251-263.
Cox, C.R. 1972. Regression model and life tables. J. Roy. Stat. Soc. B
34:187-220.
Crouch, C., D. Pullinger, and I. Gaunt. 1979. Inhalation toxicity studies
with 1,3-butadiene. Part 2: Three-month toxicity study in rats.
Am. Ind. Hyg. Assoc. J. 40:796-802.
Crump, K.S. 1979. Dose-response problems in carcinogenesis. Biometrics
35:157-167.
Crump, K.S., and Watson, W.W. 1979. GLOBAL79. A FORTRAN program to extra-
polate dichotomous animal carcinogenicity data to low dose. Natl. Inst.
of Environ. Health Science Contract No. l-ES-2123.
Crump, K.S., and R.B. Howe. 1983a. The multistage model with a time-dependent
dose pattern: applications to carcinogenic risk assessment. EPA Contract
No. 68-01-3111. Unpublished.
Crump, K.S., and R.B. Howe. 1983b. ADOLL1-83 and ADOLL2-83, FORTRAN programs
for implementing a multistage model with a time-dependent dose pattern.
Unpublished instructions for use of computer programs.
Dean, B.J., and G. Hodson-Walker. 1979. An in vitro chromosome assay using
cultured rat-liver cells. Mutat. Res. 64:329-337.
de Heester, C., F. Poncelet, F. Roberfroid, and M. Mercier. 1978. Muta-
genicity of butadiene and butadiene monoxide. Biochem. Biophys. Res.
Commun. 80:298-305.
de'Meester, C., F. Poncelet, F. Roberfroid, and M. Mercier. 1980. The
mutagenicity of butadiene towards Salmonella typhimurium. Toxicol.
Lett. 6:125-130.
Doll, R. 1971. Weibull distribution of cancer. Implications for models of
carcinogenesis. J. Roy. Statistical Soc. A 13:133-136.
Dripps, R.D., Eckenhoff, J.E., and Vandam, L.D. 1977. Introduction to
anesthesia, the principles of safe practice. 5th ed. Philadelphia,
PA: W.B. Saunders Company, pp. 121-123.
7-2
-------
Ehrenberg, L., and S. Hussain, 1981. Genotoxicity of some important epoxides.
Mutat. Res. 86:1-113.
Filser, J.G., and H.M. Bolt. 1981. Inhalation pharmacokinetics based on gas
uptake studies. I. Improvement of kinetic model. Arch. Toxicol. 47-
279-292.
Filser, J.G., and H.M. Bolt. 1984. Inhalation pharmacokinetics based on gas
uptake studies. VI. Comparative evaluation of ethylene oxide and buta-
diene monoxide as exhaled reactive metabolites of ethylene and 1,3-buta-
diene in rats. Arch. Toxicol. 55:219-223.
Hazleton Laboratories Europe, Ltd. (HLE). 1981a. The toxicity and carcino-
genicity of butadiene gas administered to rats by inhalation for approx-
imately 24 months. Prepared for the International Institute of Synthetic
Rubber Producers, New York, NY. Unpublished.
Hazleton Laboratories Europe, Ltd. (HLE). 1981b. 1,3-Butadiene: Inhalation
teratogenicity study No. 2788 522/3. Prepared for the International
Institute of Synthetic Rubber Producers, New York, NY. Unpublished.
Hemminki, K., K. Falck, and H. Vainio. 1980. Comparison of alkylation rates
and mutagenicity of directly acting industrial and laboratory chemicals
Arch. Toxicol. 46:277-285.
Howe, R.B. 1983. GLOBAL83: an experimental program developed for the U.S.
Environmental Protection Agency as an update to GLOBAL82: a computer
program to extrapolate quanta! animal toxicity data to low doses (May
1982). K.S. Crump and Co., Inc., Ruston, LA. Unpublished.
Huff, J.E., R.L. Melnick, H.A. Solleveld, J.K. Haseman, M. Powers, and R.A.
Miller. 1985. Multiple organ carcinogenicity of 1,3-butadiene in B6C3F1
mice after 60 weeks of inhalation exposure. Science 227:548-549.
International Agency for Research on Cancer (IARC). 1982. IARC monographs
on the evaluation of the carcinogenic risk of chemicals to humans
Supplement 4. Lyon, France.
Laib, R.J., R. Kreiling, and H.M. Bolt. 1985. Species differences in buta-
diene metabolism between mouse and rat as evaluated by inhalation phar-
macokinetics and covalent binding of reactive butadiene metabolites The
Toxicologist: Abstracts of the 1985 Annual Meeting of the Society of
Toxicology 5:131.
Lawley, P.O., and P. Brookes. 1967.
difunctional alkylating agents.
Interstrand cross-linking of DNA by
J. Mol. Biol. 25:143-160.
Lilis, R., and W.J. Nicholson. 1976. Cancer experience among workers in ,
chemical plant producing styrene monomers. Proceedings of Styrene-
Butadiene Briefing, Covington, KY, April 30, 1976. DHEW (NIOSH) Pub.
No. 77-129. Cincinnati, OH, pp. 22-27.
7-3
-------
Linet, M.S. 1985. The leukemias: epidemiologic aspects. Oxford, England:
Oxford University Press.
Luker, M.A., and B.O. Kilbey. 1982. A simplified method for the simulta-
neous detection of intragenic and intergenic mutations (deletions) in
Neurospora crassa. Mutat. Res. 92:63-68.
Halvoisin, E., G. Lhoest, F. Poncelet, M. Roberfroid, and M. Mercier. 1979.
Identification and quantitation of l,2-epoxybutene-3 as the primary
metabolite of 1,3-butadiene. J. Chrom. 178:419-429.
Malvoisin, E., and M. Roberfroid. 1982. Hepatic microsomal metabolism
of 1,3-butadiene. Xenobiotica 12:137-144.
Malvoisin, E., M. Mercier, and M. Roberfroid. 1980. Enzymic hydration of
butadiene monoxide and its importance in the metabolism of butadiene.
Adv. Exp. Med. Biol. 136A-.437-444.
Mantel, N., and Schneiderman, M.A. 1975. Estimating "safe" levels, a
hazardous undertaking. Cancer Res. 35:1379-1386.
Matanoski, 6.M., L. Schwartz, J. Sperrazza, and J. Tonascia. 1982. Mortality
of workers in the styrene-butadiene rubber polymer manufacturing industry.
Johns Hopkins University School of Hygiene and Public Health, Baltimore,
MO. Unpublished.
McMichael, A.J., R. Spirtas, and L.L. Kupper. 1974. An epidemiologic study
of mortality within a cohort of rubber workers, 1964-72. J. Occup. Med.
16:458-464.
McMichael, A.J., R. Spirtas, L.L. Kupper, and J.F. Gamble. 1975. Solvent
exposure and leukemia among rubber workers: an epidemiologic study.
J. Occup. Med. 17:234-239.
McMichael, A.J., R. Spirtas, J.F. Gamble, and P.M. Tousey. 1976. Mortality
among rubber workers: Relationship to specific jobs. J. Occup. Med.
18:178-185.
Meinhardt, T.J., R.A. Lemen, M.S. Crandall, and R.J. Young. 1982. Environ-
mental epidemiologic investigation of the styrene-butadiene rubber indus-
try. Scand. J. Work Environ. Health 8:250-259.
Miller L M. 1978. Investigation of selected potential environmental contam-
inants: butadiene and its oligomers. EPA-560/2-78-008. U.S. Environ-
mental Protection Agency, Office of Toxic Substances, Washington, DC.
Monson, R.R. 1979. U.S. death rates programs, versions USDR 27 and USDR 38.
Harvard University School of Public Health, Boston, MA. Unpublished.
Murray, M.P., and J.E. Cummins. 1979. Mutagenic activity of epoxy embedding
reagents employed in electron microscopy. Environ. Mutagen. 1:307-313.
7-4
-------
National Institute for Occupational Safety, and Health (NIOSH). 1984. Current
Intelligence Bulletin 41: 1,3-butadiene. Washington, DC.
National Toxicology Program (NTP). 1984. Toxicology and carcinogenesis
studies of 1,3-butadiene (CAS 106-99-0) in B6C3F1 mice (inhalation
National Toxicology Program (NTP). 1985a. Quarterly report from Lovelace
Research Institute, January 1 through March 31, 1985. Interaqencv
agreement 22-Y01-ES-0092. (L. Birnbaum, NTP Project Officer).
National Toxicology Program (NTP). 1985b. Draft report on the toxicology and
carcinogenesis studies of 4-vinylcyclohexene in F344/N rats and B6C3F1
mice. NIH Publication No. 85-2559.
on.' Health 5H£|_r^arbons in the Los An9eles atmosphere. Arch.
Olszewska E., and B.J. Kilbey. 1975. The mutagenic activity of diepoxybu-
tane in yeast. Mutat. Res. 33:383-390. uiepoxyou
°U' 19*80 ' V^f-f' H'C' Sc$arnw?ber' E-J- Schneider, and J.R. Venable.
1980. A mortality survey of employees engaged in the development or
manufacture of styrene-based products. J. Occup. Med. 22(7) -445-460.
₯n': andKH'J: Evans; 1975- Cytological detection of mutagen-carcinogen
exposure by sister chromatid exchange. Nature 258:121-125. ai^ni^en
unrfrn . Bogaert, M. Lambotte-Vandepaer,
M. Roberfroid and M. Mercier. 1980. Influence of experimental factors
on the mutagenicity of vinyl ic monomers. Arch. Toxicol. Suppl? 4:63-6e!
Rappaport S.M., and D.A. Fraser. 1976. Gas chromatographic - mass soectro
metric identification of volatiles released from a9 rJbber stock during
simulated vulcanization. Anal. Chem. 48:476-481.
R1PP'pn;K;n th67*, Hyg;enic bas1s for the Permissible concentration of butadi-
mutagemc
Res.
of the ebony (UV and X-ray sensitive and Canton <
"
1n t1ssues
7-5
-------
Simmon, V.F., and J.M. Baden, 1980. .MuUgenic activity of vinyl compounds
and derived epoxides. Mutat. Res. 78:227-231.
Soirtas, R. 1976. Mortality among rubber workers: relationship to jobs with
P styrene-butadiene exposures. Proceedings of NIOSH Styrene-Butadiene
inton KY Aril 30, 1976. DHEW (NIOSH) Pub. No. 77-129.
styrene-butadiene exposres. roceeng
Briefing, Covington, KY, April 30, 1976.
Cincinnati, OH, pp. 9-21.
Taulbee, J.', D. Adjelkovichv T. -Wni^msV1 ,J.f .' Gamble, and P. Wolf. 1976.
A study of possible associations; Between, exposure to SBR processes and
mortality from leukemia and related diseases based on toxicologic,
industrial: hygiene and epidemiol,ogi,c considerations (for workers in
the 1951 and 1964^ cohorts, and:^eathsJ96J-1973 Proceedings of NIOSH
Styrene-Butadiene Briefing, Covington,, KY, April 30, 1976. OHEW (NIOSH)
Pub. No. 77-129. Cincinnati, OH, pp. 113-129.
Turchi, 6., S. Bonatti , U Cftti ,' 'p'J. GeryasiT and T/TO&'dancjol 6'. 1981.
Alkylating properties and genetic activity of 4-vi nyl cycl ohexene meta-
bolites and stfuctu rally related epoxides. Mutat. Res. 83:419-430.
U.S. Department of Health and Human Services (DHHS). 1985. Fourth annual
report on carcinogens,: Washington,* PC.
U.S. Environmental Protect! ^Agency>;;i984a.;;; Proposed guidelines for muta-
genicity risk assessment: request for comments. Federal Register 49.
46314-46321. ; ; -:.: : , .^C ,"
U.S. Environmental Protection Agency.' 1984b. Proposed guidelines for car-
cinogen risk assessment,.; Federal -Register 49(227) :46294-46301.
U.S. Environmental Protection Agency, 1935^ Assessment of cancer risks to
workers exposed to 1,3-butadiene during production of 1,3-butadiene
monomer and production of synthetic rubbers, plastics, and resins.
Office of Toxic Substances, Washington, iDC.
Van Duuren, B.L., N. Nelson,'u"6rris,"E".'DI"Palmes, and F.L. Schmftt. ^1963.
Carcinogenicity of epoxides, lactones, and peroxy compounds. J. Natl.
Cancer';Inst. 31:4-55*.,,;., ,,_;.;, ^';.'i^:-,C ,L-. .' , : '
Van Duuren, B.L., L. Langsteth,!. Orris, G. feebor, N. Nelson, and M. Kuschner,
1966 Carcinogenicity in epoxidesi lactones, and peroxy compounds. IV.
'Tumor response in'epithelial ,and-connective tissue in mice and rats. J.
Natl .'Canter Mtf; 37:825-838.
Vogel, E. 1976, ;Mutagenicity of carcinogens in DrosopjiiLa as a function of
qenotype-Gontrolled metabolism.-In: £F.J, de Serres, J.R. Fouts, J.R.
Bend, and R.M. Phi 1 pot (eds;j^ ir"vitrcTmetabolic activation in muta-
genesis testing. Amsterdam: Else.yier/North-Holland Biomedical Press,
pp. 63-79.
Voogd, C.E., J.J. van de Stel,.and.J.J,0,A.A, Jacobs. 1981.
action of aliphatic epoxides., Mutat. Res, 89:269-282.
The mutagenic
Z-6
-------
Wade, M.J., J.W. Moyer, and C.H. Nine. 1979.
epoxides. Mutat. Res. 66:367-371.
Mutagenic action of a series of
Zaborowska, D., Z. Swietlinska, and J. Zuk. 1983. Induction of mitotic
recombination by UV and diepoxybutane and its enhancement by hydroxyurea
in Saccharomyces cerevisiae. Mutat. Res. 120:21-26.
Zimmering, S. 1983. The mei-9a test for chromosome loss in Drosophila: a
review of assays of 21 chemicals for chromosome breakage"; Environ'
Mutagen. 5:907-921.
7-7
-------
-------
-------
S* 03
8
en
co
Ui
O 5 IT
5 5'5
I 8 3
r* < o
a c&
= 0
3S
o> X
'
> mi
ffi 3
(B < ~|
' "
CD 0)1
-
L}. 3
O 3" 3i
i'i?
SSaJ
ro
o>
oa
------- |