vvEPA
TOXICOLOGICAL REVIEW
of
CHLORDANE (TECHNICAL)
(CAS No. 12789-03-6)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
December 1997
U.S. Environmental Protection Agency
Washington, DC
-------
TABLE OF CONTENTS
Authors and Reviewers iv
Foreword vii
1.0 Introduction 1
2.0 Toxicokinetics Relevant to Assessments 2
3.0 Absorption, Metabolism, Distribution, Excretion, and Toxicokinetics 3
4.0 Hazard Identification 5
4.1 Studies in Humans 5
4.1.1 Noncancer Effects in Humans 5
4.1.2 Cancer Effects in Humans 9
4.1.2.1 Case-control studies 9
4.1.2.2 Occupational cohort studies 11
4.1.2.3 Case reports 13
4.2 Subchronic and Chronic Studies and Cancer Bioassays in Animals—Oral
and Inhalation 15
4.3 Reproductive/Developmental Studies—Oral and Inhalation 26
4.4 Other lexicologically Relevant Studies 30
4.4.1 Mechanistic Studies 30
4.4.2 Genotoxicity 30
4.5 Synthesis and Evaluation of Major Noncancer Effects and Mode of Action—
Oral and Inhalation 31
4.6 Weight of Evidence Evaluation and Cancer Classification 32
4.7 Other Hazard Identification Issues 34
4.7.1 Possible Childhood Susceptibility 34
4.7.2 Possible Gender Differences 34
5.0 Dose-Response Assessments 35
5.1 Oral Reference Dose 35
5.1.1 Choice of Principal Study and Critical Effect—with Rationale and
Justification 35
5.1.2 Methods of Analysis—Benchmark Dose and No-Observed-Adverse-Effect
Level/Lowest-Observed-Adverse-Effect Level 36
5.1.3 Oral Reference Dose Derivation—Including Application of Uncertainty
Factors and Modifying Factors 36
5.2 Inhalation Reference Concentration 37
5.2.1 Choice of Principal Study and Critical Effect—with Rationale and
Justification 37
5.2.2 Methods of Analysis—No-Observed-Adverse-Effect Level/Lowest-
Observed- Adverse-Effect Level and Benchmark Concentration 38
-------
TABLE OF CONTENTS
(cont'd)
5.2.3 Inhalation Reference Concentration Derivation—Including Application
of Uncertainty Factors and Modifying Factors 38
5.3 Cancer Assessment 39
5.3.1 Choice of Study/Data with Rationale and Justification 40
5.3.2 Dose-Response Data 40
5.3.3 Dose Conversion 40
5.3.4 Extrapolation Methods 41
5.3.5 Oral Slope Factor and Inhalation Unit Risk 41
6.0 Major Conclusions in Characterization of Hazard and Dose-Response 42
6.1 Hazard Identification 42
6.2 Dose-Response Assessment 43
7.0 References 45
8.0 Appendixes 54
Appendix A: Benchmark Dose Analysis of Data and Cancer Risk Estimates 54
Appendix B: Summary of and Response to External Reviews and Comments 63
in
-------
AUTHORS AND REVIEWERS
Chemical Manager/Author
Robert E. McGaughy, Ph.D
Health Scientist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
Contributing Authors
Gary L. Foureman, Ph.D. (RfD and RfC Assessments)
Health Scientist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
Peter McClure, Ph.D. (RfD and Cancer Assessments)
Senior Scientist
Syracuse Research Corporation
Syracuse, NY
Reviewers
This document and summary information on IRIS have received peer review by both U.S.
Environmental Protection Agency (EPA) scientists and independent scientists external to EPA
(U.S. EPA, 1994c). Subsequent to external review and incorporation of comments, this
assessment has undergone an Agency-wide review process whereby the IRIS Program Manager
has achieved a consensus approval among the Office of Air and Radiation; Office of Policy,
Planning and Evaluation; Office of Prevention, Pesticides, and Toxic Substances; Office of
Research and Development; Office of Solid Waste and Emergency Response; Office of Water;
and the Regional Offices.
Internal EPA Reviewers
David L. Bayliss, Ph.D
Epidemiologist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
Dharm Singh, Ph.D.
iv
-------
Toxicologist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
Thomas M. Crisp, Ph.D
Biologist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
Henry Spencer, Ph.D.
Pharmacologist
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
Washington, DC
Vanessa Vu, Ph.D.
Director, Risk Assessment Division
Office of Pesticides, Prevention and Toxic Substances
U.S. Environmental Protection Agency
Washington, DC
External Peer Reviewers
John B.Barnett, Ph.D.
Chair, Department of Microbiology and Immunology
West Virginia University School of Medicine
Morgantown, WV
Aaron Blair, Ph.D.
Epidemiology and Biostatistics Program
National Cancer Institute
Bethesda, MD
Benjamin A. Jackson, Ph.D.
Private Consultant, Regulatory Toxicology
Silver Spring, MD
-------
Kaye H. Kilburn, M.D.
Professor of Medicine
University of Southern California School of Medicine
Los Angeles, CA
Summaries of the external peer reviewers' comments and the disposition of their
recommendations are in Appendix B.
VI
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale for
the hazard identification and dose-response information in the Integrated Risk Information System
pertaining to chronic exposure to chlordane. It is not intended to be a comprehensive treatise on
the chemical or toxicological nature of chlordane.
In Section 6, EPA has characterized its overall confidence in the quantitative and qualitative
aspects of hazard and dose-response (U.S. EPA, 1995a). Matters considered in this
characterization include knowledge gaps, uncertainties, quality of data, and scientific
controversies. This characterization is presented in an effort to make apparent the limitations of
the individual assessments and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process. For other general information about this assessment or other questions
relating to the Integrated Risk Information System, the reader is referred to EPA's Risk
Information Hotline at (202) 566-1676.
vn
-------
1.0 INTRODUCTION
Chlordane was first produced in 1947 and was used as an insecticide for agricultural crops
and livestock, for lawns and gardens, and also for underground treatment around the foundation
of homes. In 1978, because of concern over cancer risk, evidence of human exposure and danger
to wildlife, EPA canceled its use on food crops and phased out its other above-ground uses.
From 1983 to 1988 its only approved use was as a termiticide around home foundations, and all
uses were canceled after 1988. However, residues still exist in soils and sediments and chlordane
bioaccumulates in fatty tissue offish and humans; this bioaccumulation is a source of current
concern.
This Toxicological Review is the background document on chlordane from which the three
1997 summary assessments on the Integrated Risk Information System (IRIS), the oral reference
dose (RfD), inhalation reference concentration (RfC), and cancer assessment, are derived. These
RfD and cancer summaries replace the information entered on the IRIS in 1993, and the RfC
summary is new.
The RfD and RfC provide information on long-term toxic effects other than carcinogenicity.
The RfD is based on the assumption that thresholds exist for certain toxic effects such as cellular
necrosis, but may not exist for other toxic effects such as some carcinogenic responses. It is
expressed in units of milligrams per kilogram per day. In general, the RfD is an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population
(including sensitive subgroups) that is likely to be without an appreciable risk of deleterious
effects during a lifetime. The inhalation RfC is analogous to the oral RfD. The inhalation RfC
considers toxic effects for both the respiratory system (portal-of-entry) and for effects peripheral
to the respiratory system (extrarespiratory). It is expressed in units of milligrams per cubic meter.
The carcinogenicity assessment provides information on three aspects of the carcinogenic
risk assessment for the agent in question: (1) the EPA classification (2) quantitative estimates of
risk from oral exposure and (3) inhalation exposure. The classification reflects
a weight-of-evidence judgement of the likelihood that the agent is a human carcinogen and the
conditions under which the carcinogenic effects may be expressed. Quantitative risk estimates are
presented in three ways. The slope factor is the result of the application of a low-dose
extrapolation procedure and is presented as the risk per milligrams per kilogram per day. The unit
risk is the quantitative estimate in terms of either risk per |ig/L drinking water or risk per |ig/m3 of
air breathed. The third form in which risk is presented is a drinking water or air concentration
providing cancer risks of 1 in 10,000, 1 in 100,000 or 1 in 1,000,000.
Development of these hazard identifications and dose-response assessments for chlordane
has followed the general guidelines for risk assessments as set forth by the National Research
Council (1983). Other EPA guidelines that were used in the development of this assessment
include the following: Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1986a), (new)
Proposed Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1996a), Guidelines for
Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Guidelines for Reproductive
Toxicity Risk Assessment (U.S. EPA, 1996b), Interim Policy for Particle Size and Limit
1
-------
Concentration Issues in Inhalation Toxicity (U.S. EPA, 1994a), (Proposed) Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1995b), Methods for Derivation of Inhalation
Reference Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b),
Recommendations for and Documentation of Biological Values for Use in Risk Assessment (U.S.
EPA, 1988a) and Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA,
1995c).
The literature search strategy employed for this compound was based on the Chemical
Abstract Service Registry Number (CASRN) and at least one common name. As a minimum, the
following databases were searched: RTECS, HSDB, TSCATS, CCRIS, GENETOX, EMIC,
EMICBACK, DART, ETICBACK, TOXLINE, CANCERLINE, and MEDLINE and MEDLINE
backfiles.
Information submitted by the public as a consequence of the Federal Register notification on
the development of the IRIS file for this chemical (U.S. EPA, 1996b) was also utilized and
considered in the development of this document.
2.0 CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS
The U.S. EPA (1979) considers technical chlordane (CAS No. 12789-03-6) to be composed
of 60% octachloro-4,7-methanotetrahydroindane (the cis and trans isomers) and 40% related
compounds.
The term chlordane in association with CAS No.57-74-9 refers to a mixture of chlordane
isomers, other chlorinated hydrocarbons and numerous other components. For example, the
mixture used by the National Cancer Institue (NCI) in its 1977 bioassay was described as
94.8% chlordane (cis [or alpha]-chlordane, 71.7%; trans [or gamma]-chlordane, 23.1%) with
heptachlor, 0.3%; trans-nonachlor, 1.1%; c/'s-nonachlor, 0.6%; chlordene isomers, 0.25%;
3% other compounds, and hexachlorocyclopentadiene, 0.25% (NCI, 1977).
Technical chlordane, CAS No. 12789-03-6, is a mixture of chlordane and chlordane related
compounds having a lower percentage of the cis and trans isomers and a larger percentage of
other compounds relative to mixtures with the above CAS number. Dearth and Kites (1991)
identified 147 different compounds in a preparation of technical chlordane. The identity and
percent of total for the 12 most abundant compounds were: c/'s-chlordane, 15%; rram--chlordane,
15%; trans-nonachlor, 9.7%; octachlordane, 3.9%; heptachlor, 3.8%; c/'s-nonachlor, 2.7%;
"compound K," 2.6%; dihydrochlordene, 2.2%; nonachlor III, 2%; and three stereoisomeric
dihydroheptachlors totaling 10.2%. These 12 compounds comprise 67% of the mixture, and the
remaining 33% of the mixture consists of 135 other compounds. Infante et al. (1978) reported
another production sample of technical chlordane to have a composition of 38 to 48% cis- and
rram--chlordane, 3 to 7 or 7 to 13% heptachlor, 5 to 11% nonachlor, 17 to 25% other chlordane
isomers, and a small amount of other compounds. Unless otherwise indicated all studies
described in this document were carried out with technical grade chlordane.
-------
Other physical characteristics of chlordane relevant to this assessment include its solubility
and vapor pressure. The log octanol/water partition coefficient has been estimated at 5.54
(ATSBR, 1994), indicating that it is insoluble in water and has the propensity to accumulate in
fatty tissues. The vapor pressure for pure chlordane has been reported as 1E-5 and for technical
grade chlordane 4.6E-4 mm Hg at 25 °C (IARC, 1991). Based on these values (and a molecular
weight of 409), a saturated vapor concentration for pure chlordane would be around 0.2 mg/m3,
and for technical chlordane (same assumed molecular weight) around 10 mg/m3. In inhalation
studies carried out at or above this concentration the chlordane would be in both vapor and
aerosolized particulate form, and at concentrations near or less than this value the chlordane
would be predominately in vapor form. This is relevant as the form in which the compound is
delivered (i.e., particle or vapor) affects the manner of calculating dose (U.S. EPA, 1994b).
At 25 °C and assumed molecular weight of 409, conversion factors are ppm (v/v) x 16.75 =
mg/m3 and mg/m3 x 0.0597 = ppm (v/v).
3.0 TOXICOKINETICS RELEVANT TO ASSESSMENTS
The primary toxicokinetic determinant of chlordane and its metabolites is their partitioning
into and persistence in fat tissue. Animal studies show rapid and efficient uptake of chlordane and
relatively complete elimination of individual doses. Oxychlordane, an epoxide metabolite of
chlordane, appears to be the most persistent metabolic chlordane residue in both animal and
human tissues. Oxychlordane and heptachlor epoxide, the latter from heptachlor which is present
in nearly all mixtures of chlordane (Section 2.0), are the metabolites of chlordane that are
considered of primary toxicological significance.
Animal studies indicate that single doses of chlordane administered orally are well absorbed,
extensively metabolized, distributed throughout the body, almost completely eliminated in 7 days,
and leave residues in several tissues, predominately fat. Male and female Sprague-Dawley rats
were orally administered single radiolabeled doses of a 3:1 mixture of cis- and rram--chlordane
isomers and the individual isomers and their tissue concentrations and elimination monitored
(Barnett and Borough, 1974). The chemical nature of the excreted radiolabel was determined
chromatographically. Single doses of the isomers were almost completely (> 90%) eliminated
after 7 days, mostly via feces, with only 2 to 6% of this total eliminated in the urine.
Approximately 15% of the radiolabel in the feces was as the administered compound, the
remainder being mostly dechlorination products. Of the residues remaining in the tissues, the
epoxide metabolite (oxychlordane) was the principal residue in all tissues examined, accounting
for 32 to 35% of the total residue content present in the kidney and 53 to 63% in fat. Nye and
Borough (1976) examined the fate of endotracheally administered radiolabeled chlordane in
anesthetized female Sprague-Bawley rats. Administration of the insecticide was as an aerosol
(isomeric composition not given) and blood levels, tissue levels after 1 h, and excretion of
radiolabel were studied. The chemical nature of the excreted radiolabel was not determined. No
radiolabel was detected in exhaled air. Chlordane residues appeared rapidly in the blood, reaching
maximum concentrations 2 to 5 min after administration. Blood levels began to dissipate
immediately, but then slowly subsequent to the maximum. No elimination constants were
calculated. One hour after administration, the lungs contained 24%, the liver 20%, the kidney
-------
0.3%, and the bladder 0.1% of the total dose. Elimination was primarily via the feces with about
52% of the dose eliminated after 6 days, whereas the urine accounted for another 12% (for a total
of 64% as compared to > 90% total in oral studies in 7 days).
Lipid partitioning of chlordane and its metabolites has been documented in both humans and
animals. Concentrations of chlordanes (c/'s-chlordane, rram--chlordane, oxychlordane, and
trans-nonachlor) detected in human cadaver liver samples were 17-fold higher when expressed on
a fat rather than a wet weight basis (average concentration of 0.05 mg/kg fat versus 0.003 mg/kg
wet weight) (Mussalo-Rauhamaa, 1991). This value is in concordance with the results of
Khasawinah (1989) who reported this ratio to be 200 to 300/1 in rats and monkeys after 90 days
of inhalation exposure to technical chlordane. Taguchi and Yakushiji (1988) found elevated
chlordane residues, including oxychlordane and heptachlor epoxide, in the milk of women
exposed to chlordane as a consequence of home termite treatment. After 56 days of feeding
isomeric mixtures of chlordane (3:1, cis-:tmns-) to rats, Barnett and Borough (1974) observed
concentrations of residues in the fat to be approximately 3-fold more than the concentration of
chlordane in the feed, whereas concentrations in other organs were far less than the dietary
concentrations levels (only 1/8 as high as liver).
The lipid partitioning nature of chlordane may also facilitate dermal absorption. Hirai and
Tomokuni (1993) noted higher skin surface/blood ratios of chlordane and its metabolites among
hospital outpatients (n = 196) whose homes had been treated compared with those whose homes
had not been treated with chlordane.
Evidence from humans and animals indicates that oxychlordane is the most persistent
chlordane metabolite. Oxychlordane and heptachlor epoxide have been reported in a large
majority of adipose tissue samples (90 to 92%) taken at surgery from humans (Nomeir and Hajjar,
1987). In four pest control workers, plasma levels of components or metabolites of chlordane
(cis- and rram--nonachlor, oxychlordane, heptachlor epoxide) were monitored over a period of
eight mo after discontinuance of chlordane exposure subsequent to 6 to 8 years of employment
(Takamiya, 1990). In all four workers concentration was highest for tmns-nonachlor with
oxychlordane, c/'s-nonachlor and heptachlor epoxide being present at about the same
concentration. In three of the four workers, the plasma levels of these metabolites did not drop
during the 8 mo exposure-free monitoring period. Among 51 pest control operators,
rram--nonachlor was detected in 37% with a mean concentration range of 0.55 ppb, oxychlordane
in 22% with a mean of 0.29 ppb, and heptachlor epoxide in 20% with a mean of 0.29 ppb
(Saito et al., 1986). In male and female Sprague-Dawley rats that were fed isomeric mixtures
(3:1, cis-:trans-) of chlordane at 1, 5, and 25 ppm for up to 56 days, fat concentrations of
oxychlordane were essentially unchanged 56 days after withdrawal of the chlordane feeding,
whereas other metabolites underwent nearly complete depuration in this period. Because of long
retention time in adipose tissue, oxychlordane is believed to be more toxic than its parent isomers,
which are eliminated relatively rapidly from the body (Satoh and Kikawa 1992); therefore,
oxychlordane may be a major contributor to chlordane toxicity.
Oxychlordane was a minor residue (14 to 19% of the total) in tissues (adipose, plasma, liver,
and red blood cells) of monkeys exposed to aerosolized chlordane (10 mg/m3, 8 h/day,
5 days/week) for 90 days, whereas it was a major component (36 to 51% of the total) in these
-------
same tissues from rats similarly exposed (Khasawinah, 1989). Thus, rat tissues appear to have a
much greater level of metabolic activity towards the precursors of oxychlordane (i.e., chlordane)
than do monkey tissues. In a subgroup of rats examined 90 days after this exposure, the figure for
oxychlordane had risen to 85 to 92% of the total as a result of its selective retention over other
residues. Thms'-nonachlor, a component of technical chlordane, was the major residue in the
sampled monkey tissues comprising 38 to 62% of the total; for rats the corresponding figure was
2 to 13%. Minor quantities of other unidentified chlordane metabolites are excreted in the urine
and feces (Barnett and Borough, 1974) and also may be stored in the body fat for long periods of
time. Alpha (cis)- and gamma (Yram')-chlordane tend to concentrate in muscle of mice following
oral administration (Satoh and Kikawa, 1992).
Adeshina and Todd (1991) developed a methodology to estimate daily human doses of
chlordane from all routes of exposure. Although their method implies that the fat concentration
after steady state conditions have been reached is directly proportional to the daily intake, human
data on dieldrin, which they cite as a chlordane analogue, indicates a nonlinear relationship. This
discrepancy can not be resolved until the long time (9 to 12 mo or longer in humans) required to
attain asymptotic levels (or steady state) in fat is taken into account.
The major metabolites of cis- and trans-chlordane following incubation of the compounds
with human liver microsomal preparations are oxychlordane, chlordene chlorhydrin, and
monohydroxylated dihydrochlordene (Tashiro and Matsumura, 1978). The metabolite pattern
was similar for both isomeric forms.
4.0 HAZARD IDENTIFICATION
4.1 Studies in Humans
4.1.1 Noncancer Effects in Humans
Kilburn, K.H. and J.C. Thornton. 1995. Protracted neurotoxicity from chlordane sprayed to kill
termites. Environ. Health Perspect. 103(7-8): 690-694.
Batteries of neurophysiological and neuropsychological tests were performed on 216 adult
occupants (109 women and 97 men) of an apartment complex, the exterior of which had been
sprayed with an unknown concentration of chlordane 7 years earlier. Chlordane levels were
assayed for 3 to 4 years after the initial spraying and showed indoor concentrations of chlordane
as high as 13.6 jig/ 929 cm2 (wipe samples) and indoor air levels reported as above 0.5 jig/ m3 for
8-h samples. The duration of the individuals' residence at the complex was not noted. Blood and
fat samples from eight of these residents (number actually tested not given) showed measurable
and elevated levels of heptachlor, oxychlordane, and trans-nonachlor. Testing and questionnaire
administration were done in either Spanish or English. The exposed group was compared with
174 unexposed referents matched on age and educational level. Neurophysiological results
showed significant (p < 0.002) slowed reaction times in both English tested (414 ms occupants
versus 309 ms referents, simple; 639 ms occupants versus 564 ms referents, choice) and Spanish
tested (456 ms occupants versus 334 ms referents, simple; 694 ms occupants versus 609 ms
-------
referents, choice) subjects. Balance sway speed was significantly increased (p < 0.01) in both
English tested (0.95 cm/s occupants versus 0.82 cm/s referents with eyes open; and 1.48 cm/s
occupants versus 1.26 cm/s referents with eyes closed) and Spanish tested (0.96 cm/s occupants
versus 0.81 cm/s referents with eyes open; 1.40 cm/s occupants versus 1.17 cm/s referents with
eyes closed) subjects. Neuropsychological tests also showed deficits among exposed occupants
relative to the referent population; both immediate and delayed verbal recall were significantly
lower (p < 0.04) in both English tested (scores of 8.7 occupants versus 10.5 referents for
immediate; 6.7 occupants versus 8.3 referents for delayed) and Spanish tested (scores of 8.4
occupants versus 10.4, immediate; 6.4 versus 8.5, delayed) occupants. Tests of cognitive
function showed significant deficits between exposed and referent populations in vocabulary
scores in both English tested (16.0 occupants versus 18.5 referents, p = 0.0254) and Spanish
tested (11.6 occupants versus 17.5 referents, p = 0.0001) subjects. Tests also showed significant
deficits between exposed and referent populations on digit symbol scores in English tested (48.5
occupants versus 57.5 referents, p = 0.00005) and Spanish tested (38.3 occupants versus 46.9
referents, p = 0.0046) subjects. Profiles of mood scores including those for tension, depression,
anger, vigor, fatigue and confusion, were all significantly higher (p < 0.0005) for subjects of both
languages than the corresponding referent populations. Confounding factors which could affect
the results, such as alcohol consumption, illicit drug use, and neurological or psychiatric diseases,
were addressed and dismissed as not having any influence on the results mostly because these
factors were not different from those in the referent population. However, no dose-response
information, no definitive exposure information (either levels or duration of exposure), or
information about co-exposure to other neurotoxicants, such as lead, could be gleaned from this
report and no effect levels were assigned.
These results show significant impairment of both neurophysiological and psychological
functions in a nonoccupational population exposed to uncertain levels of chlordane, predominately
through the inhalation route via indoor air. The most notable changes were slowing of reaction
time, balance dysfunction, reductions in cognitive function, and deficits of immediate and delayed
recall. These results indicate that neurological effects are a relevant endpoint in humans
chronically exposed to airborne chlordane concentrations around 0.0005 mg/m3. However, no
dose-response information and no definitive exposure information (either levels or duration of
exposure) could be gleaned from this report, and no effect levels can be assigned.
Alvarez, W.C. and S. Hyman. 1953. Absence of toxic manifestations in workers exposed to
chlordane. Arch. Ind. Hyg. Occup. Med. 8:480-483.
Twenty-four male workers at a chlordane manufacturing plant underwent examinations that
included chest X rays, liver function tests (sulfobromophthalein dye excretion, thymol turbidity
tests, total bilirubin in blood), hemoglobin concentration, and urinalysis. Unspecified neurological
examinations were also performed. The workers had been employed from periods of 2 mo to
5 years and had been exposed to chlordane through both inhalation and dermal absorption,
although no exposure levels are given. Chest X-ray films revealed minimal shadows in the lungs
of 7 workers, although these were not considered to be from chlordane exposure. No evidence of
injury to the liver, kidneys, or blood-forming organs was indicated by any of the tests. No
indications of any neurological disturbances were noted.
-------
Fishbein, W.I., J.V. White, and HJ. Isaacs. 1964. Survey of workers exposed to chlordane.
Ind. Med. Surg. 33: 726-727.
Fifteen male workers at a chlordane manufacturing plant underwent liver function tests
(sulfobromophthalein dye excretion), EKGs, urinalysis, and had a chest X ray taken. A routine
physical examination and medical history were also carried out. The duration of employment for
14/15 of the workers was from 9 to 16 years, the other being employed for only 2 years. Four air
samples for chlordane were taken that ranged from 1.2 to 1.7 |ig/m3. All organ function tests and
results from physical examinations were normal. No gastrointestinal or nervous system
disturbances were noted, although it is not apparent to what extent they were examined for these
symptoms.
Menconi, S., J.M. Clark, P. Langenberg, and D. Hryhorczuk. 1988. A preliminary study of
potential human health effects in private residences following chlordane applications for termite
control. Arch. Environ. Health 43(5): 349-352.
A cross-sectional study of 85 chlordane-treated houses (self-selected, totaling 261 people)
was performed via a questionnaire. Indoor air sampling from 107 homes (including those 85
answering the questionnaire) allowed for three categories of exposure levels: < 1 |ig/m3, 1 to
5 |ig/m3, and > 5 |ig/m3. These levels were reported by the authors as unrelated to time after
application, as chlordane levels decline very slowly. The range of time between completing the
questionnaire and the last termite control application, 1 to 24 years (mean 3.8 years.), may be
viewed as duration of exposure. Incidence rates of acute and chronic conditions from the
questionnaires were compared to rates from the 1979 U.S. Health Interview Survey. Incidence
rates for a number of conditions including anemia, migraine, bronchitis, sinusitis, and dermatitis
were found to be significantly elevated (p < 0.05) over those noted in the Health Interview
Survey. When smoking, age, and sex were controlled for the increased rates for sinusitis,
bronchitis, and migraine remained significant (p < 0.05) for the upper two exposure categories.
However, this study population was self-selected and not random and any results are subject to
bias. No effect levels are assigned to this study.
Aldrich, F.D. and J.H. Holmes. 1969. Acute chlordane intoxication in a child. Case report with
toxicological data. Arch. Environ. Health 19(1): 129-132.
Garrettson, L.K., P.S. Guzelian, and R.V. Blanke. 1985. Subacute chlordane poisoning.
J. Toxicol. Clin. Toxicol. 22(6): 565-571.
Olanoff, L.S., WJ. Bristow, J. Colcolough, Jr., and J.R. Reigart. 1983. Acute chlordane
intoxication. J. Toxicol. Clin. Toxicol. 20(4): 291-306.
These three studies report central nervous system effects including headaches, irritability,
excitability, confusion, incoordination, muscle tremors, seizures, convulsions, and coma following
acute oral exposure to insecticidal formulations of chlordane. In one incident (Aldrich and
Holmes, 1969), intermittent clonic convulsions, incoordination, hyporeflexia, and increased
excitability were exhibited by a 4-year-old girl who ingested a 45% chlordane emulsifiable
concentrate at an estimated dose of 0.15 mg/kg.
-------
Fleming, L.E. and W. Timmeny. 1993. Aplastic anemia and pesticides: An etiologic association?
J. Occup. Med. 35: 1106-1116.
This review of 30 years of medical literature identified 280 cases of aplastic anemia
associated with pesticide exposure including numerous organochlorines (including
heptachlor/chlordane), carbamates, and organophosphates. For 90 of these cases, information
about age, sex, and latency was available and, of these, four involved chlordane/heptachlor with
the average age of 50 years and average latency 2 mo. These case studies indicate that
insecticides (including chlordane/heptachlor) are associated with myelotoxicity but nearly always
under conditions of acute high doses. Few cases, if any, appear to be associated with long-term
low-level exposures. Only qualitative and anecdotal information about exposure levels is
available.
McConnachie, P.R. and A.C. Zahalsky. 1992. Immune alterations in humans exposed to the
termiticide technical chlordane. Arch. Environ. Health 47(4): 295-301.
Twenty-seven individuals who had been acutely exposed to technical chlordane either at
home (23) or at work (4) underwent an immunological evaluation. Information given indicated
that durations of exposure ranged from 3 days to 15 mo, and the length of time from exposure
(beginning or end of exposure not indicated) to testing ranged from 4 mo to 10 years. Analysis of
fat biopsies in 14 of the 27 individuals revealed detectable levels of trans-nonachlor and/or
chlordane. All of these individuals complained of health problems (not described) that they
attributed to chlordane. The exposed individuals differed significantly (p > 0.01) from a referent
nonexposed population (n = 118) in distribution of lymphocytes and in their mean response to
decreased function (decreased response to several foreign antigens). Test for autoantibodies were
positive in 11 of 12 individuals tested although the rate in the referent nonexposed population is
not given. These results indicate an association of immunologic dysregulation and acute exposure
to technical chlordane.
4.1.2 Cancer Effects in Humans
4.1.2.1 Case-control studies
Cantor, K.P., A. Blair, G. Everett, R. Gibson, L.F. Burmeister, L.M. Brown, L. Schuman, and
F.R. Dick. 1992. Pesticides and other agricultural risk factors for non-Hodgkin's lymphoma
among men in Iowa and Minnesota. Cancer Res. 52: 2447-2455.
Cantor et al. (1992) compared 622 white men recently diagnosed with non-Hodgkin's
lymphoma (NHL) in Iowa and Minnesota with 1,245 population-based controls matched with
respect to age, vital status at the time of interview, and state of residence. Detailed interviews
-------
over a period from 1981 through 1984 elicited information about their pesticide handling
practices. Chlordane was one of 12 chemicals evaluated as animal (livestock) insecticides; the
others were coumaphos, DDT, dichlorvos, famphur, flyspray (NOS), lindane, malathion,
methoxychlor, nicotine, rotenone, and toxaphene. It was also one of 15 chemicals evaluated as
crop insecticides; the others being aldrin, carbofuran, carboxyl, copper acetoarsenate, DDT,
diazinon, dieldrin, fonofos, heptachlor, lindane, malathion, phorate, turbofos, and toxaphene.
They found that among NHL cases the odds of chlordane use specifically as an animal insecticide
were significantly greater than among controls (OR= 1.7, 95% CI = 1.0 to 2.9; 31 cases and 38
controls). The odds of chlordane use more generally as either an animal or a crop insecticide was
also significant (OR = 1.8, CI = 1.1 to 3.1; 30 cases). The pattern of odds ratios for other
chemicals studied were similar to chlordane, in that livestock use had greater odds ratios than
crop use; this consistency indicated to the authors that the increased odds ratios were due to
greater exposure with livestock than crops rather than some random error. In discussing the
limitations of their study, the authors point out the difficulty of attributing the results to individual
pesticides when the exposures were multiple, and they also point out the possibility that important
associations may have been missed due to nondifferential exposure misclassification because of
the difficulties in accurate recall of past pesticide exposures. They concluded that chlordane, as
well as some other specifically mentioned pesticides, has an important role in the etiology of NHL
among farmers.
Brown, L.M., A. Blair, R. Gibson, G.D. Everett, K.P. Cantor, M. Schuman, L.F. Burmeister,
S.F. Van Lier, and F. Dick. 1990. Pesticide exposures and other agricultural risk factors for
leukemia among men in Iowa and Minnesota. Cancer Res. 50:6585-6591.
In a study of similar design to that of Cantor et al. (1992), the authors conducted a
population-based case-control study of leukemia among 578 white men with leukemia and
1,245 controls living in Iowa and Minnesota. No significantly elevated odds ratios for leukemia
were associated with chlordane exposures. Chlordane was one of 17 crop insecticide exposures
and one of 16 animal insecticide exposures evaluated.
Woods, J.S., L. Polissar, R.K. Severson, L.S. Heuser, and E.G. Kulander. 1987. Soft tissue
sarcoma and non-Hodgkin's lymphoma in relation to phenoxyherbicide and chlorinated phenol
exposure in western Washington. J. Nat. Cancer Inst. 78(5): 899-910.
Woods et al. (1987) published a population-based case-control study of 128 soft-tissue
sarcoma cases and 576 NHL cases among men in western Washington state, where
phenoxyherbicides and chlorinated phenols are widely used in the agricultural, forestry and wood
products industries. There were 694 controls in the study and the case diagnoses were from 1981
through 1984. They investigated other chemical handling practices, including the use of
chlordane, that potentially could modify or independently affect the risk of these diseases.
Although their main interest was herbicides and phenols, one of their findings was an elevated, but
not statistically significant, odds ratio of NHL among farmers who worked with chlordane (OR =
1.46, CI = 0.8 to 2.8). The pooled odds ratios for all occupations studied was 1.6. (0.7 to 3.8).
This study is mentioned here in connection with chlordane use by farmers, not because it is strong
evidence of chlordane risk, but because it supports the findings of Cantor et al. (1992), discussed
above.
-------
Zahm, S.H., D.D. Weisenburger, P.A. Babbitt, R.C. Saal, K.P. Cantor, and A. Blair. 1988.
A case-control study of non-Hodgkin's lymphoma and agricultural factors in eastern Nebraska.
Am. J. Epidemiol. 128: 901.
This abstract reported the results of a population-based case-control study of NHL in
relation to agricultural factors in eastern Nebraska. Among the 385 cases and 1,432 controls in
the study, the use of chlordane was one specific factor significantly associated with NHL (OR =
2.2, lower 95% confidence bounds is above 1.0). The use of three other insecticides was also
significantly associated with NHL, and the authors also mentioned the use of three herbicides as
other factors. As with the other studies in farmers, multiple exposures of individuals to chemicals
in addition to the specific chemical identified is possible.
Zahm, S.H., D.D. Weisenburger, P.A. Babbitt, R.C. Saal, J.B. Vaught, K.P. Cantor, and A. Blair.
1990. A case-control study of non-Hodgkin's lymphoma and the herbicide
2,4-dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiology 1(1): 349-356.
This subsequent publication of the results for eastern Nebraska showed a 50% excess risk of
NHL associated with using the herbicide 2,4-D, and that the risk was not altered by considering
exposure to chlorinated hydrocarbons as a class. Chlordane was not mentioned specifically.
Brown, L.M., L.F. Burmeister, G.D. Everett, and A. Blair. 1993. Pesticide exposures and
multiple myeloma in Iowa men. Cancer Causes Control 4:153-156.
Brown et al. (1993) published a population-based case-control study of multiple myeloma
among 173 white men in Iowa compared with 650 controls in order to evaluate possible
contribution of agricultural risk factors and exposure to individual pesticides. In 173 men with
multiple myeloma (and 650 controls), no statistically significant elevated risks for multiple
myeloma were found for farmers or individuals who handled either classes of pesticides or specific
pesticides. Chlordane was one of seven animal insecticides evaluated, but no information about
potential exposure levels was available.
Pesatori, A.C., J.M. Sonntag, J.H. Lubin, D. Consonni, and A. Blair. 1994. Cohort mortality and
nested case-control study of lung cancer among structural pest control workers in Florida (United
States). Cancer Causes Control 5:310-318.
Pesatori et al. (1994) updated an earlier study of licensed structural pest control workers in
Florida (Blair et al., 1983). The original cohort of 3,827 men was started in 1965 and was
followed until 1977, and the 1994 study updated the follow-up period to 1982. In view of the
elevated lung cancer mortality found in the original study, they conducted a nested case-control
study of 65 lung cancer cases; 294 controls were interviewed. The interviews gathered
information on tobacco use, occupation, work practices, and dietary habits. They found in the
case-control study that tobacco use, diet, and other occupations had little effect on the excess
lung cancer associated with occupational pesticide exposure. This excess was larger among those
licensed before the age of 40 and those licensed for more than 20 years. Because of the small
sample size, the use of individual pesticides could not be evaluated, but carbamates and
phenoxyacetic acids were significantly elevated and use of inorganics, organobromides,
10
-------
organochlorines, and organophosphates were nonsignificantly elevated in these lung cancer cases.
Use of chlordane specifically was not associated with lung cancer.
4.1.2.2 Occupational cohort studies
Wang, H.H. and B. MacMahon. 1979a. Mortality of workers employed in the manufacture of
chlordane and heptachlor. J. Occup. Med. 21: 741-744.
Wang and MacMahon (1979a) studied the mortality patterns of male workers in two plants,
one in Illinois that manufactured chlordane and another in Tennessee which manufactured
heptachlor and endrin. The authors did not have a measure of exposure to chlordane. The
employees working from the start-up of the plants (1946 for the chlordane plant and 1952 for the
other plant) until 1976 were eligible for the study. In the 113 deaths at the two plants, there was
no statistically significant excess of deaths from cancer, even in those followed for more than 20
years after the beginning of employment.
Brown, D.P. 1992. Mortality of workers employed at organochlorine pesticide manufacturing
plants - an update. Scand. J. Work Environ. Health 18: 115-161.
Ditraglia, D., D.P. Brown, N. Namekata, and N. Iverson. 1981. Mortality study of workers
employed at organochlorine manufacturing plants. Scand. J. Work Environ. Health
7(suppl.): 140-146.
Brown (1992) reported follow-up data through 1987 of a cohort mortality study of white
male workers employed at four organochlorine pesticide manufacturing plants, two of which were
the same plants studied by Wang and MacMahon (1979a). The original study (Ditraglia et al.,
1981) included all employees who worked at least 6 mo prior to 1964 and vital statistics were
followed until 1976. There was no excess cancer mortality in either the original study or in
Brown's follow-up study. The small number of deaths in the chlordane plant workers (159)
available for analysis precludes the ability to make definitive conclusions about the effects of
chlordane exposure.
Shindell, S. and S. Ulrich. 1986. Mortality of workers employed in the manufacture of
chlordane: an update. J. Occup. Med. 28:497-501.
Cancer mortality was studied in a cohort of 800 male employees who worked at a U.S.
chlordane manufacturing plant for at least 3 mo during the years 1946 through 1985. Death
certificates were available for 161 of the 181 deaths that occurred within the study period, and
"reliable causes of death" were reported for an additional 11 deaths. Compared with U.S. death
rates, no elevated risks for death from all causes or cancer were found in the cohort or in
subcohorts comprised of either production workers or nonproduction workers. Average blood
concentrations of oxychlordane and heptachlor epoxide measured in 1976 and 1979 among
workers (n = 42 and 70) showed average levels were higher by 34 to 113% in production versus
nonproduction workers. Although there was no elevated risk for death from heart disease in the
cohort (76 observed versus 84.2 expected), elevated risk was found for death from stroke
(20 observed versus 11.7 expected; p < 0.01).
11
-------
Infante, P.P. and C. Freeman. 1987. Cancer mortality among workers exposed to chlordane.
J. Occup. Med. 29(11): 908-909.
In a letter to the editor, Infante and Freeman (1987) challenged the results of Shindell based
on calculation of expected rates according to published studies of this cohort, asserting there was
an increasing trend of the standard mortality rate (SMR) with increasing time of employment.
Shindell responded in the same 1987 publication with a table showing the expected rates and
stating that there was no trend. These data, along with the results of a linear regression
performed by the authors, is shown in Table 1.
Wang, H.H. and B. MacMahon. 1979b. Mortality of pesticide applicators. J. Occup. Med.
21; 741-744.
Wang and MacMahon (1979b) studied the prospective mortality of male pesticide
applicators employed by three nationwide companies for at least 3 mo between 1967 and 1976.
Of the 16,126 men in the study, there were 311 deaths with no statistically significant excess
cancer mortality. A follow-up of this cohort to 1984 was published by MacMahon et al. (1988).
A total of 1,082 deaths were ascertained and 994 death certificates were obtained. Expected
numbers of death by cause were calculated based on U.S. national rates and compared with
observed deaths in the cohort. Only one category of death, lung cancer, showed a significantly
elevated SMR in the total cohort (SMR =135; 90% CI = 114 to 158), and this excess was
restricted to those employed for less than five years. Attributing that finding to chlordane
exposure is unlikely, since there was no increase in lung cancer among termite control operators,
the group most likely to be exposed.
Table 1. Observed and expected deaths (based on U.S. rates) from respiratory or other
cancers in white male workers in a U.S. chlordane manufacturing plant, 1946-1985.
Years Employed
(5 -year intervals)
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35+
Respiratory Cancer
Observed
3
1
2
2
2
1
1
0
Expected
6.1
1.3
1.5
1.2
1.6
1.7
1.0
_
SMR
49
77
133
167
125
59
100
_
Other Cancer
Observed
7
O
4
3
2
2
3
1
Expected
11.1
2.6
3.2
2.5
2.8
2.7
1.0
0.3
SMR
63
115
125
120
71
74
300
333
12
-------
Total 12 14.4 83 25 26.3 95
Note: Linear regression analysis of employment duration class (up to 35 years) versus SMRs for respiratory cancer
or other cancers found no statistically significant trend for increasing SMR for respiratory cancer (p = 0.67) or
other cancers (p = 0.20) with increasing duration of employment.
Source: Shindell and Ulrich (1986).
4.1.2.3 Case reports
Epstein, S.S. and D. Ozonoff. 1987. Leukemias and blood dyscrasias following exposure to
chlordane and heptachlor. Teratogen. Carcinogen. Mutagen. 7(6): 527-540.
Epstein and Ozonoff (1987) presented 25 cases of blood dyscrasias (thrombocytopenic
purpura, aplastic anemia, and leukemia) following acute exposure to chlordane/heptachlor. Four
of these were leukemia cases. In 16 of the 25 cases, no exposure in addition to chlordane or
heptachlor was reported and over 75% of these involved homeowners following termite
treatment, and garden and lawn applications. No other exposure information is given. These
results were included in the review of Fleming and Timmeny (1993), discussed above.
Infante, P., S.S. Epstein, and W.A. Newton, Jr. 1978. Blood dyscrasias and childhood tumors
and exposure to chlordane and heptachlor. Scand. J. Work Environ. Health 4: 137-150.
The authors presented clinical case histories of five children under the age of seven with
neuroblastoma, all of whom had been exposed to chlordane in household situations, either in
utero (three cases) or post-natally. They also presented three cases of aplastic anemia, with
evidence of exposures to other pesticides in addition to chlordane in two of them. Three cases
of leukemia were presented also, two of which had no known exposure to other agents besides
chlordane. (Also discussed in the review by Fleming and Timmeny [1993], above.)
Chadduck, W.M., S.M. Gollin, B.A. Gray, J.S. Norris, C.A. Araoz, and A.F. Tyrka. 1987.
Gliosarcoma with chromosome abnormalities in a neonate exposed to heptachlor. Neurosurgery
21(4): 557-559.
Chadduck et al. (1987) discussed the case of an infant boy with gliosarcoma, an uncommon
pediatric malignant brain tumor. The mother had consumed large amounts (over 2 quarts/day) of
milk from a local dairy throughout pregnancy which was found to be contaminated with
heptachlor. The authors stated that causal attribution to heptachlor was not possible from this
case, but that further study was needed.
Caldwell, G.G., S.B. Cannon, C.B. Pratt, and R.D. Arthur. 1981. Serum pesticide levels in
patients with childhood colorectal carcinoma. Cancer 48: 774-778.
The authors obtained chemical exposure histories of nine adolescent colon cancer patients
treated at a hospital in Tennessee during the years 1974 through 1976. Eight of them had lived on
farms and had been exposed to pesticides. Analysis of the serum of patients and their families for
13
-------
DDT, dieldrin, beta-hexachlorocyclohexane, and heptachlor epoxide showed no overall
differences from 24 adolescent controls living in the same region, and both patients and controls
had levels similar to Mississippi state averages. Therefore, the data provide no evidence that
these pesticides contribute to this rare form of cancer.
Teufel, M., K.H. Niessen, J. Sartoris, W. Brands, H. Lochbiihler, K. Waag, P. Schweizer, and
G.V. Oelsnitz. 1990. Chlorinated hydrocarbons in fat tissue: Analyses of residues in healthy
children, tumor patients and malformed children. Arch. Environ. Contam. Toxicol. 19: 646-652.
This study found no differences in fatty tissue concentrations of chlorinated hydrocarbons
(including heptachlor and heptachlor epoxide) among three groups of children: 183 healthy
children, 33 children with congenital malformations or benign tumors, and 46 children with
malignant tumors. This study does not provide data to support an association between chlordane
or heptachlor and cancer.
Falck, F., A. Ricci, M.S. Wolff, J. Godbold, and P. Deckers. 1992. Pesticides and
polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch.
Environ. Health 47: 143-146.
This study found no statistically significant differences in levels of chlordane residues
(heptachlor epoxide, oxychlordane, trans-nonachlof) in breast fat from 20 women with malignant
breast disease compared with 20 women with benign breast disease (predominately
nonproliferative fibrocystic changes). They did find higher levels of DDE and PCBs in cancer
patients than in women with benign disease. The study fails to make any association between
chlordane residue levels and breast cancer incidence.
4.2 Subchronic and Chronic Studies and Cancer Bioassays in Animals—Oral
and Inhalation
Khasawinah, A.M., C. Hardy, and G. Clark. 1989a. Comparative inhalation toxicity of technical
chlordane in rats and monkeys. J. Toxicol. Environ. Health 28: 327-347.
(The 90-day rat study.)
Wistar rats (35 to 47/sex/group) were exposed to 0, 0.1, 1.0, or 10 mg/m3 technical
chlordane, 8 h/day, 5 days/week, for 13 weeks (adjusted to continuous 24 h exposures to 0.024,
0.24, or 2.4 mg/m3), followed by a 13-week recovery period. The control group was exposed to
an air-only chamber. Monkeys were exposed concurrently with the rats. The test material was
produced by stainless steel concentric jet atomizers. The polydisperse aerosol of chlordane was
introduced into a elutriating glass column in which nonrespirable droplets of chlordane were
removed by sedimentation and impaction. Evaporation of chlordane also occurred in an
elutriating column, enhancing the proportion of chlordane as vapor. The aerosol at the top of the
column was then introduced into the exposure chambers. The concentration of chlordane in the
chamber was determined at least three times during each 8-h exposure. Chlordane concentration
was determined by summation of six peaks termed as characteristic of technical chlordane from
gas chromatography analysis, which included both chlordane isomers, frvms-nonachlor, and
heptachlor. Particle size distribution based on mass deposition was determined at least once per
14
-------
week during the study and reported as 91.9 and 82.9% less than 5.5 jim in aerodynamic diameter
for the high- and intermediate exposure groups, respectively. Ancillary information on particle
sizing and characterization was included in an unpublished version of this study (Velsicol
Chemical Corporation, 1984). A three-stage liquid impinger was used to test particle size
distribution twice each week for the first week and once each week after that for a total of
17 determinations. These data showed that, at 1 mg/m3 chlordane, 51.4% of the particle mass
was <2 jim, 31.5% was between 2 and 5.5 |im, and 17.2% was >5.5 jim. This information was
used to obtain an approximate mass median aerodynamic diameter (MMAD) of 1.8 jim and a
corresponding particle distribution estimate (sigma g) of 3.1, which allow estimation of a regional
deposited dose ratio (RDDR) (Section 5.2.3; U.S. EPA, 1994b). Optical counting and sizing of
particles from air samples taken from both the low-dose and control chambers showed that nearly
all particles were less than 0.5 jam in diameter and that counts were highly variable. This finding
gives little evidence of particulate chlordane exposure at the lowest dose level. As the vapor state
would predominate at this concentration (vapor pressure supporting >0.2 mg/m3, see Section 2),
and as 0.5 |im is in the range of diffusion, the concentration of chlordane at this lowest
concentration is considered to be totally vapor.
Rats were sacrificed at 9 (5/sex/group) and 13 weeks (15/sex/group) of exposure, and
13 weeks postexposure (9/sex/group). Blood chemistry and urinalysis were performed before and
at 5 and 13 weeks of exposure. There were no exposure-related effects on mortality, clinical
signs (except sensitivity to touch at the highest concentration), food consumption, body weight,
ophthalmoscopy, rectal temperature, or urinalysis. White blood cell (WBC) parameters were
affected in female rats at Week 13; elevated total WBC and lymphocyte counts occurred in the
1 mg/m3 (WBC, 10.8 versus 7.1 x 103/mm3 controls; lymphocyte, 8.23 versus 5.45 x 103/mm3
controls) and 10 mg/m3 (WBC, 11.5 versus 7.1 x 103/mm3 controls; lymphocyte, 8.60 versus
5.45 x 103/mm3 controls) females, and increased neutrophil count (2.61 versus 1.42 x 103/mm3
controls) occurred in the 10-mg/m3 females. Hematological changes were noted predominately in
females and included decreased mean corpuscular hemoglobin concentration at all exposure levels
(8%) that was not concentration-related and reduced hemoglobin at 1 (8%) and 10 mg/m3 (9%).
None of these effects are considered adverse as they are all minor changes, do not follow any
clear concentration-response, and are all within normal ranges for these parameters: WBC 3 to
15 x 103/mm3, lymphocyte 4 to 10 x 103/mm3, neutrophil s 1.1 to 4 x 103/mm3 (Mitruka and
Rawnsley, 1981). Thrombocytopenia was also noted in females exposed to the two highest
exposure concentrations where platelet counts were decreased 25 and 35% (p < 0.01)
respectively, from the control count. This decreased count was not accompanied by any
significant increase in clotting time even at the highest concentration tested and the difference was
no longer apparent 13 weeks postexposure. At 13 weeks of exposure there were several blood
chemistry indications of hepatic functional alteration restricted to the highest level of chlordane.
These included a significant decrease in glucose (83% of control for females), increased globulins
(114% of control for males), increased total protein (104% of control for males), decreased
albumin (95% of control for males), an increase in cholesterol (158% of control for females), and
an altered albumin/globulin ratio.
Histopathology was performed on all tissues, including nasal passages (number of levels not
noted), pharynx, larynx, trachea, lungs, and bronchi, in the control and high concentration group.
No effects were observed on the respiratory system. Because effects (apparent increased
15
-------
incidence of liver and thyroid enlargement at the highest dose) were observed in the liver and
thyroid during the 9-week interim sacrifice, these organs were examined in all groups at 13 weeks
of exposure and 13 weeks postexposure. At the end of the exposure period, there was
a statistically significant increase (p < 0.05) in the incidence of rats with centrilobular hepatocyte
enlargement (or hypertrophy), which was observed in all rats of the 10-mg/m3 group (minimal
severity) and in 5/15 males and 5/15 females of the 1 mg/m3 group (slight severity). No animals
were affected in the control or 0.1 mg/m3 groups. A polynomial model was fit to these data to
estimate the air concentration corresponding to a 10% incidence of hypertrophy; the result is
0.64 mg/m3. This information is used in Section 5.3.3 to compare with the oral dose giving the
same toxic effect in the Khasawinah and Grutsch (1989b) study. At 27 weeks (13 weeks
postexposure) only one female rat in the 10 mg/m3 group had slight hepatocyte enlargement.
Increased liver weights (p < 0.01) were observed for male and female rats exposed to 10 mg/m3 at
weeks 9 and 14, but weights were similar to those of the controls by week 27. Exposed rats also
exhibited a concentration-related increase in both hepatic cytochrome P-450 concentration and
microsomal protein. This response of induction involving reversible liver cell hypertrophy and
liver enzymes is considered to be an adaptive response indicative of systemic exposure to
chlordane. Kidney weights were significantly elevated (p < 0.05) in the 1 and 10 mg/m3 males at
9 weeks, and in females and males exposed to 10 mg/m3 at week 14 (p < 0.01), but no
histopathological lesions were observed. Males also exhibited slightly increased height of the
follicular epithelium of the thyroid in 1/15 animals exposed to 1 mg/m3, and in 11/15 animals
exposed to 10 mg/m3, although no abnormalities were noted in animals examined 13 weeks
postexposure. The lowest-observed-adverse-effect level (LOAEL) of 10 mg/m3 chlordane was
determined based on the wide range of biochemical indicators indicating hepatic functional
alteration at this concentration. The no-observed-adverse-effect level (NOAEL) was 1 mg/m3.
Khasawinah, A.M., CJ. Hardy, and G.C. Clark. 1989b. Comparative inhalation toxicity of
technical chlordane in rats and monkeys. J. Toxicol. Environ. Health 28: 327-347.
(The 28-day rat study.)
In a 28-day range-finding study, Wistar (Crl:Cobs WI Br strain) rats (10/sex/group) inhaled
0, 5.8, 28.2, 154, or 413 mg/m3 technical chlordane, 8 h/day, 5 days/week. Body and organ
weights and food and water consumption were measured and hematology, urinalysis, and
microscopic examination were performed. Because of deaths (incidence not reported) occurring
early in the study (females dying earlier than the males), animals in the 154 and 413 mg/m3 groups
were sacrificed after exposures 11 and 3, respectively. At 154 mg/m3, abnormal respiratory
movements and excessive salivation occurred in some animals, occasionally followed by
convulsions and death. Reduced weight gain and food consumption were also observed in the
groups exposed to the two highest concentrations. No deaths or significant weight loss were
observed in the 5.8 and 28.2 mg/m3 groups. There were no exposure-related hematological
alterations. There were, however, several blood chemistry indications of hepatic functional
alteration in rats exposed to 28.2 mg/m3 including a significant decrease in glucose (73% of
control for males and 78% of control for females), increased globulins (116% of control for males
and 119% of control for females), increased total protein (116% of control for females), increased
albumin (117% of control for females), an increase in cholesterol (133% of control for females),
and an altered albumin/globulin ratio. An increase in liver weight was noted in the 5.8 mg/m3
males and the 28.2 mg/m3 males and females, as well as in surviving rats of the 154 and 413
16
-------
mg/m3 groups (no specifics given). Increased kidney and thyroid weight and decreased thymus
weight were also exhibited in the 28.2 mg/m3 group. Livers were enlarged in the 154 mg/m3
males and discolored in the 154 and 413 mg/m3 males (no incidence given). As in the livers of
rats exposed in the 90-day study, centrilobular hepatocyte enlargement (hypertrophy) was
observed in the livers of an unspecified number of animals exposed to 28 mg/m3. Centrilobular
hepatocyte enlargement, with or without vacuolation, was observed among rats exposed to 154
mg/m3. In addition to hepatocellular hypertrophy and vacuolation, necrosis was also noted among
rats exposed to 413 mg/m3. Effects observed in other organs (no incidence given) included
increased follicular epithelial cell height in thyroid tissue and increased incidence of cervical lymph
node enlargement in the 154 mg/m3 males, and epithelial degeneration in bronchi/bronchioles and
cell debris in bronchi, bronchioles, and alveoli of the 413 mg/m3 animals. A LOAEL of 28.2
mg/m3 chlordane was determined based on the wide range of biochemical indicators indicating
alteration of hepatic function at this concentration. The NOAEL was 5.8 mg/m3.
Khasawinah, A.M., CJ. Hardy, and G.C. Clark. 1989c. Comparative inhalation toxicity of
technical chlordane in rats and monkeys. J. Toxicol. Environ. Health 28: 327-347.
(The 90-day monkey study.)
Cynomolgus monkeys (6/sex/group) were exposed to 0, 0.1, 1.0, or 10 mg/m3 technical
chlordane (adjusted to continuous 24 h exposures to 0.024, 0.24, or 2.4 mg/m3, respectively) for
90 days, 8 h/day, 5 days/week. Similar parameters were examined in the monkeys as those
reported for rats. The respiratory system (and pulmonary function as indicated by blood gas
analysis, lung mechanics, and pulmonary ventilation) of the monkeys was unaffected by inhalation
exposure to chlordane. Minor blood effects were noted between controls and some of the dose
groups, including increases in mean corpuscular volume, decreases in WBC, alteration in
prothrombin times and blood urea-nitrogen, and decreased platelet counts, although all values
were within normal limits. The only effects were increased mean liver and thyroid weights for the
10-mg/m3 group compared to controls; however, the weight changes were not statistically
significant and did not correspond to any histopathological changes. Histopathological
examination was the same as for rats save that only control and high dose groups were examined.
There were no exposure-related histopathological findings observed in any animals. No adverse
changes were noted and the NOAEL was 10 mg/m3.
IRDC (International Research and Development Corporation). 1973. Eighteen-month oral
carcinogenic study of chlordane in mice. Unpublished report to Velsicol Chemical Corporation.
MRID No. 00067568. Available from U.S. Environmental Protection Agency.
U.S. EPA. 1986b. Carcinogenicity Assessment of Chlordane and Heptachlor/Heptachlor
Epoxide. Washington, DC: Office of Health and Environmental Assessment. NTIS,
PB87-208757.
Groups of 100 male and 100 female CD-I mice were fed diets containing analytical grade
chlordane at concentrations of 0, 5, 25 or 50 ppm (approximately 0, 0.71, 3.57, or 7.14 mg/kg-
day) for 18 mo. Mice were 6 weeks of age when exposure began. A 6 mo interim sacrifice of
10 mice/sex/group did not reveal compound-related lesions. The terminal sacrifice was at
19.5 mo. No exposure-related changes in body weight gain or food consumption were observed.
17
-------
Many mice were lost due to autolysis so that only about half of the mice were examined
histologically. The number of mice for which tissue was available for histopathological analysis
was less in the high dose group than in the other treated groups but comparable to that in
controls. U.S. Environmental Protection Agency-sponsored examination of the liver histological
slides found statistically significant increased incidence of hepatic carcinomas in males and females
in the 25- and 50-ppm groups (see Table 2). No effects were observed other than cancer, even at
the highest dose tested (50 ppm).
NCI (National Cancer Institute). 1977. Bioassay of chlordane for possible carcinogenicity.
Technical Report Series No. 8. U.S. Department of Health, Education and Welfare; National
Institutes of Health. PB 271 977.
Rat and mouse bioassays were conducted with analytical grade chlordane administered in
the diet (71.7% c/'s-chlordane, 23.1% rram--chlordane, 0.3% heptachlor, 0.6% nonachlor, 1.1%
hexachlorocyclopentadiene, and 0.25% chlordane isomers).
Groups of 50 male and 50 female Osborne-Mendel rats were fed low- or high-dose diets for
80 weeks, and then observed for 29 weeks. Chlordane dietary concentrations were reduced twice
during the study because of obvious toxic effects. Time-weighted average concentrations were
203.5 and 407 ppm for males and 120.8 and 241.5 ppm for females. These concentrations
correspond to approximate average doses of 10.2 and 20.4 mg/kg-day for male, and 6.0 and
12.1 mg/kg-day for female rats. Matched controls consisted of 10 male and 10 female rats.
Table 2. Incidence of liver lesions in CD-I mice exposed to analytical chlordane
in the diet for 18 mo.
Males
(ppm in diet)
Lesion
Hepatocyte hyperplasia
Hepatic nodules
Hepatic small carcinomas
Hepatic large carcinomas
0
20/33
1/33
0/33
3/33
5
34/55
6/55
3/55
2/55
25
7/52
3/52
9/52a
32/52a
50
7/39
0/39
4/39
28/3 9a
0
26/45
0/45
0/45
0/45
Females
(ppm in diet)
5
32/61
1/61
0/61
0/61
25
13/50
4/50
7/50a
25/50a
50
11/37
0/37
2/37
24/37a
(> 5mm)
Total hepatic carcinomas
3/33
5/55 41/52a
32/39a
0/45
0/61
32/50a
26/37a
aSignificantly increased relative to control incidence, one-tailed Fisher Exact test (p < 0.01); p > 0.05 for
comparison between other elevated incidence and respective controls.
Source: U.S. EPA (1986b).
A pooled control group consisted of the matched-control rats and 50 male and 50 female
untreated rats from similar bioassays conducted during periods that overlapped with the chlordane
bioassay by at least a year. Surviving rats were sacrificed at 109 weeks. Major organs and tissues
from sacrificed rats and from rats found dead were examined grossly and microscopically.
18
-------
The average body weights of high-dose male and female rats were consistently lower (by
about 10%) than that of control rats, throughout the study. Obvious clinical signs of toxicity
(rough and discolored hair coats, palpable masses) occurred "frequently" in treated rats during the
first year and increased in frequency during the second year. No statistically significant
differences in survival were found for exposed male rats compared with control rats, but there
was a statistically significant dose-related trend for decreased survival in exposed female rats. No
information about the cause of death was given. No statistically significant increases of liver
tumors or liver nonneoplastic lesions were found in exposed rats compared with incidence in
control rats. Statistically significant increases of proliferative lesions of follicular cells of the
thyroid and of malignant fibrous histiocytoma were found in exposed male rats compared with
controls. The study authors concluded, however, that these differences were not biologically
significant because the incidences were within the range of incidence for control rat populations.
No other exposure-related neoplastic or nonneoplastic lesions were found in rats in this study.
The NOAEL for noncancer effects is 12.1 mg/kg-day based on female rats.
Groups of 50 male and 50 female B6C3F1 mice were fed diets with low and high
concentrations of analytical chlordane for 80 weeks, and then observed for 10 weeks. The low-
and high-dose groups were tested at different calendar times, but each exposed group was tested
with a concurrent control. Chlordane dietary concentrations were changed during the study.
Time-weighted average concentrations were 29.9 and 56.2 ppm for male mice and 30.1 and
63.8 ppm for female mice. These concentrations correspond to approximate average doses of
4.3 and 8.0 mg/kg-day for male, and 4.3 and 9.1 mg/kg-day for female mice. Matched controls
consisted of 10 male and 10 female untreated mice. A pooled control group consisted of the
matched-control mice and 70 male and 80 female untreated mice from similar bioassays
conducted during periods that overlapped with the chlordane bioassay by at least a year.
Surviving mice were sacrificed at 90 to 91 weeks. Major organs and tissues from all mice were
examined grossly and microscopically.
No significant exposure-related changes in average body weight were found in male or
female mice. Survival of exposed female mice was essentially the same as control female mice,
but survival of male mice in both exposed groups was significantly decreased relative to control
males. Neurotoxicity (tremors) was observed in high-dose males and females after 20 weeks.
Exposure-related neoplastic or nonneoplastic lesions in mice were restricted to the liver.
Hepatocytomegaly, nodular hyperplasia, and diffuse hyperplasia were found in exposed male and
female mice, but statistically these incidences were not significantly different from controls
(see Table 3). Statistically significant elevated incidences for hepatocellular carcinomas, relative
to controls, were found in low- and high-dose male mice and in high-dose female mice (Table 3).
Hepatocellular carcinomas were described as displaying a spectrum of histological features.
At one end of the spectrum, nodules of hepatocytes with only moderate changes (from normal
hepatocytes) in staining characteristics, size and shape of cells, and lobular architecture occurred.
At the other end of the spectrum, tumors with clearly anaplastic cytologic characteristics and no
resemblance to normal lobular architecture occurred. Metastasis to the lung occurred in
two high-dose males and three high-dose females. No effects were observed other than cancer,
even at the highest dose tested (50 ppm).
19
-------
Table 3. Incidence of liver lesions in B6C3F1 mice exposed to analytical chlordane
in the diet for 80 weeks
Males
(ppm in diet)
Lesion
Hepatocytomegaly
Diffuse hyperplasia
Nodular hyperplasia
Hepatocellular
carcinoma
Pooled
Control
0/92
3/92
3/92
17/92
Matched
Control
0/18
0/18
2/18
2/18
29.9
1/48
3/48
7/48
16/48a
56.2
56.2
0/49
0/49
43/49b
Pooled
Control
1/78
1/78
2/78
3/78
Females
(ppm in diet)
Matched
Control
0/19
0/19
1/19
0/19
30.1
0/47
3/47
3/47
3/47
63.8
0/49
3/49
3/49
34/49b
"Different from pooled (p = 0.04) and matched (p = 0.06) controls; one-tailed Fisher Exact test.
bDifferent from pooled and matched controls (p < 0.001); one-tailed Fisher Exact test.
Source: NCI (1977).
Khasawinah, A.M. and J.F. Grutsch. 1989b. Chlordane: thirty-month tumorigenicity and
chronic toxicity test in rats. Regul. Toxicol. Pharmacol. 10: 95-109.
Velsicol Chemical Corporation. 1983a. Thirty-month chronic toxicity and tumorigenicity test in
rats by chlordane technical. Unpublished study by Research Institute for Animal Science in
Biochemistry and Toxicology, Japan. MRID No. 00138591, 00144313. Available from U.S.
Environmental Protection Agency.
ICF-Clement. 1987. Pathology peer review of chlordane in F344 rats. Pathology review
participants: Goodman, D.G., A.W. Mackin, R.R. Maronpot, J.A. Popp, R.A. Squire,
J.M. Ward, and M.R. Anver.
U.S. EPA. 1988b. Review of Pathology Working Group Slide Reevaluation of Livers of Rats
in a 30-Month Oral Exposure to Chlordane. Office of Pesticides and Toxic Substances.
Memorandum from Henry Spencer to George LaRocca, March 14.
In a 30-mo oral study, Fischer 344 rats (80/sex/group) were exposed to 0, 1, 5, or 25 ppm
technical chlordane in the diet. Based on food consumption and body weight data, the study
authors calculated the doses to be 0, 0.045, 0.229, and 1.175 mg/kg-day for males and 0.055,
0.273, and 1.409 mg/kg-day for females. Doses were based on a 4-week pilot study in which
hepatic effects (increased liver cell volume and fatty degeneration) were seen at 50 ppm and
higher. Hematology, biochemistry, urinalysis, organ weights, and pathology (of major tissues and
organs) were assessed at weeks 26 and 52 (eight males and nine females/group), and on all
survivors at week 130. Because the original 1983 pathology report was completed before the
publication of the National Toxicology Program's diagnostic criteria for proliferative liver lesions,
and because diagnostic discrepancies were found in the original pathology report, the sponsor of
the study (Velsicol Chemical Corporation) convened a Pathology Working Group (PWG) of six
U.S. pathologists to review the liver histopathological slides from this study. These slides were
20
-------
evaluated on a coded "blind" basis without knowledge of dose group, age or sex. The PWG's
liver pathology findings are reported herein. The assessment of chlordane in the U.S. EPA
(1986b) report was based on the original pathology report (Velsicol Chemical Corporation,
1983a) and is superceded by the data presented here.
No treatment-related clinical signs, deaths, or biochemical and hematological parameters
were observed. The original pathologist noted that exposure-related lesions were found only in
the liver. Absolute liver weight was significantly increased (p < 0.05) in 5- and 25-ppm males
(13.33 and 13.30 gm versus 10.44 gm in controls) at 130 weeks and in 25-ppm females at
26 weeks (5.02 gm versus 4.62 gm in controls) and 52 weeks (5.77 gm versus 5.27 gm in
controls). These increases in liver weight among the 25-ppm but not the 5-ppm animals were
accompanied by severe liver pathology. A high incidence of large granular lymphocyte leukemia
was found in all groups of female and male rats, including controls (Table 4). Elevated incidence
of leukemia in control F344 rats living beyond 104 weeks, relative to rats terminated at
104 weeks, has been reported as a common age-related occurrence that is accompanied by
nonneoplastic liver lesions including necrosis, cytomegaly (i.e., swelling or hypertrophy) and fatty
degeneration (Solleveld et al., 1984; Stromberg and Vogtsberger, 1983). Because age-related
leukemia confounded exposure-related liver lesions, the PWG examined incidence for
nonneoplastic liver lesions in rats without leukemia. Nonleukemic female rats in the 5 and 25 ppm
groups had statistically significant increases in hepatocellular hypertrophy relative
21
-------
Table 4. Incidence of nonneoplastic and neoplastic lesions in F344 ratsa exposed to
technical chlordane in the diet for 30 mo.
Males
(ppm in diet)
Lesion
Large granular lymphocyte
leukemia
Hepatocellular adenoma
Regional hepatocellular
0
43/64
2/64
0/20
1
43/64
4/64
1/23
5
40/64
2/64
0/20
25
46/64
7/64b
0/16
0
25/64
0/64
2/37
Females
(ppm in diet)
1
18/64
1/64
4/43
5
21/64
0/64
9/40c
25
19/64
0/64
15/44C
hypertrophy in nonleukemic
livers
Foci of cellular alteration in
nonleukemic livers
8/20
12/23 14/20
8/20
28/37
55/43
32/40
32/44
"Excludes the 16 interim-kill animals per group.
bOne-tailed Fisher Exact test, p = 0.08.
°p < 0.05 for one-tailed Fisher Exact test; Cochrane-Armitage trend test for females was p = 0.0003.
Source: Khasawinah and Grutsch (1989b).
to controls, whereas nonleukemic females in the 1 ppm group, and nonleukemic males in the 1, 5,
and 25 ppm groups, had no significant increases (Table 4). Overall incidence for hepatocellular
adenomas in 25-ppm male rats (not including interim-kill rats) was increased by a marginal
statistical significance relative to controls (7/64 versus 2/64; Table 4). Incidences for hepatic
neoplasms in other exposed groups were not increased. According to the authors, no increased
incidence of foci of cellular alteration in nonleukemic livers (the first stage of rat liver
carcinogenesis) was found in exposed groups relative to controls.
Incidences for hepatocellular adenomas in the concurrent F344 rat controls (3.1% in males
and 0% in females) were below historical values for control F344 rats in United States (range 0 to
17%) and Japanese laboratories (2.5% for males dying before 104 weeks and 9.6% for males in
studies of longer duration). In consideration of this, and the marginal significance of the observed
increased incidence of hepatocellular adenomas in 25-ppm males compared with controls
(p = 0.08), the PWG concluded that chlordane treatment in this study did not produce
a tumorigenic response in male or female F344 rats. The Toxicology Branch of EPA's Office of
Pesticides and Toxic Substances concurred with this conclusion (U.S. EPA, 1988b). The PWG
also concluded that regional hepatocellular hypertrophy was associated with administration of
5- or 25-ppm chlordane in the diet to female rats for 30 mo. Interpolation of the data on regional
hypertrophy in female rats (Table 4) shows that at a dietary dose of 0.15 mg/kg-day (assuming
100% absorption of chlordane in the diet) the incidence is 10%. This information is used in
Section 5.3.3 to compare with the inhalation dose giving the same toxic effect in the Khasawinah
et al. (1989a) study. Thus, in female rats, 1 ppm (0.055 mg/kg-day) was the NOAEL and 5 ppm
(0.273 mg/kg-day) the LOAEL for chlordane-induced nonneoplastic liver lesions.
Khasawinah, A.M. and J.F. Grutsch. 1989a. Chlordane: 24-month tumorigenicity and chronic
toxicity test in mice. Regul. Toxicol. Pharmacol. 10: 244-254.
22
-------
Velsicol Chemical Corporation. 1983b. Twenty-four month chronic toxicity and turn origeni city
test in mice by chlordane technical. Unpublished study by Research Institute for Animal Science
in Biochemistry and Toxicology, Japan. MRID No. 00144312, 00132566. Available from U.S.
Environmental Protection Agency.
ICR mice (80/sex/group) were given 0, 1, 5, or 12.5 ppm technical chlordane in the diet for
104 weeks. The study authors reported that average doses corresponded to 0, 0.15, 0.75, and
1.875 mg/kg-day, based on an assumed food consumption of 15% of the body weight per day.
Dosages were chosen based on a 4-week pilot study in which a LOAEL of 25 ppm was
determined for hepatic toxicity (increased hepatic cell volume). Hematology, biochemistry,
urinalysis, organ weights, and pathology (of major tissues and organs) were assessed at week 52
(eight males and eight females/group) on all animals that died during the study, and on all
survivors at week 104.
When male mice were analyzed without regard to the presence or absence of tumors, no
consistent, exposure-related changes were found in clinical signs of toxicity, body weights,
mortality, urinalysis parameters, or hematological parameters, except that 12.5-ppm male mice
showed a statistically and probably biologically significant elevation in serum alanine
aminotransferase activity (308 IU/L versus 102 IU/L in controls) at week 104. Statistically
significant increased absolute and relative liver weights were seen in high-dose males (200 and
203%) and females (144 and 129%) compared to control values at the end of the study. Absolute
(but not relative) liver weight was also significantly elevated (p < 0.05) in female mice exposed to
1 ppm (22%) and 5 ppm (14%). No consistent changes in other organ weights were found.
When animals with adenocarcinomas were compared with those without adenomas or
adenocarcinomas, however, the animals with malignant tumors had statistically significant
elevations of all blood biochemistry measurements made (SGOT, SGPT, glucose, urea nitrogen,
cholesterol, and protein). The authors concluded that the prevalence of the malignant tumors, and
not the exposure to chlordane, altered the blood biochemistry parameters.
Exposure-related neoplastic and nonneoplastic lesions were restricted to the liver.
Statistically significant increased incidences, relative to controls, were found for the following
nonneoplastic lesions: hepatocellular swelling (hypertrophy) in male and female mice in the
5- and 12.5-ppm groups, hepatic fatty degeneration in 12.5-ppm males, and hepatic necrosis in
5- and 12.5-ppm male mice. The incidence for hepatic necrosis among male mice was 7/80 for
the controls, 8/80 for the 1-ppm group, 25/80 for the 5-ppm group, and 27/80 for the 12.5-ppm
group. These data are modeled in Appendix A. Male mice in the 12.5-ppm group showed
statistically significant elevated incidence of hepatocellular adenomas (also called hepatocellular
nodules by the study authors) and hepatic hemangiomas compared with controls (Table 5). The
hepatic hemangiomas were described as benign tumors of capillary cells associated with the
hepatocellular adenomas. The 1989 report mentions the occurrence of liver adenocarcinomas but
does not present incidence data. However, the earlier report of this assay, Velsicol Chemical
Corporation (1983a), cited by U.S. EPA (1986b), shows that the hepatocellular carcinoma
23
-------
Table 5. Incidence of hepatocellular adenomas and hemangiomas in male ICR mice
exposed to dietary chlordane for 24 mo.
Dietary Concentration Incidence of Hepatocellular Adenomas Incidence of Hemangiomas
0 12/79 4/79
1 14/79 1/79
5 14/80 8/80
12.5 27/80a 14/80b
aP<0.01.
bP < 0.05.
Source: Khasawinah and Grutsch (1989a).
incidence in males in the 0-, 1-, 5-, and 12.5-ppm groups is 3/71, 3/71, 7/72, and 9/72,
respectively. This data is used for the quantitative cancer risk estimates (Table 6d). One ppm
chlordane (0.15 mg/kg-day) was the NOAEL, and 5 ppm chlordane (0.75 mg/kg-day) was the
LOAEL for nonneoplastic liver lesions (hepatic necrosis) in male ICR mice.
Barrass, N., M. Stewart, S. Warburton, J. Aitchison, D. Jackson, P. Wadsworth, A. Marsden, and
T. Orton. 1993. Cell proliferation in the liver and thyroid of C57B1/10J mice after dietary
administration of chlordane. Environ. Health Perspect. 101(suppl. 5): 219-224.
Barrass et al. (1993) carried out two assays in which chlordane with a concentration of
50 ppm was fed to C57B1/10J mice. The first was a 24-mo bioassay of chlordane in a group of
100 male mice, with histopathological analysis of liver sections. They found that in the animals
surviving after 2 years of treatment the incidence of hepatocellular adenomas and carcinomas was
15/39 and 5/39, respectively. In the animals with unscheduled deaths during the assay, the
incidence of adenomas and carcinomas was 1/40 and 2/40, respectively. The incidence of total
liver tumors in a in concurrent colony control group was 8/400. Therefore, the incidence of
adenomas and carcinomas combined was 23/79, which is statistically significant relative to
controls. However, because the incidence of carcinomas alone in the control group was not
given, the statistical significance of the carcinomas alone can not be calculated nor can the risk of
carcinoma alone be determined.
The second assay was a 6-mo cell proliferation study with nine serial sacrifices in which
replicating cells were labeled by infusing BrdU via subcutaneous mini-osmotic pumps 3 days
before sacrifice. Thyroid and liver cell replication and histopathology were measured. In this
assay, there were no morphological thyroid abnormalities; the thyroid labeling index reached
a peak after 5 days and reached control levels after 99 days. Centrilobular hepatocellular
hypertrophy was seen starting early and increasing throughout the experiment. The liver labeling
index showed an initial peak phase lasting 4 weeks and a sustained elevated phase lasting for the
entire 200-day period but gradually declining by day 200.
24
-------
Malarkey, D.E., T.R. Devereux, G.E. Dinse, P.C. Mann, and R.R. Maronpot. 1995.
Hepatocarcinogenicity of chlordane in B6C3F1 and B6D2F1 male mice: evidence for regression
in B6C3F1 mice and carcinogenesis independent of ras proto-oncogene activation.
Carcinogenesis 16: 2,617-2,625.
These authors reported the age-specific prevalence of hepatocellular adenomas, carcinomas
and two measures of liver toxicity: centrilobular hypertrophy and "toxic change", which consists
of cell necrosis, cell vacuolization and inflammation. These observations were made in males of
two mouse strains, B6C3F1 (210 animals) and B6D2F1 (160 animals), fed diets containing
55 ppm of technical chlordane starting from 63 days of age and continuing until 568 days of age
(72 weeks of treatment) and 668 days (86 weeks of treatment), respectively. This is the same
dietary concentration used in the high-dose groups of the NCI (1977) study in B6C3F1 mice. The
protocol included one group of B6C3F1 mice in which the chlordane was discontinued after
491 days of age (61 weeks of treatment). Control mice were not sacrificed until age 759 days
(99 weeks of treatment) to obtain a sufficient number of spontaneous tumors for analysis of ras
mutations. Mutations in K-ras and H-ras oncogenes were measured in spontaneous tumors from
controls and chlordane treated animals.
They found that in both strains the prevalence of adenomas and carcinomas increased with
time and that the B6C3F1 mice developed tumors about 100 days before the B6D2F1 mice. The
prevalence was nearly 100% in both strains at terminal sacrifice. By week 86, the time of terminal
sacrifice of the B6D2F1 mice, the prevalence of adenomas and carcinomas was 91%. After 72
weeks of treatment (age 568 days), which was about 8 weeks less than the duration of the NCI
(1977) test, the prevalence of carcinomas in B6C3F1 mice was about 88%, as determined from
the prevalence versus time graphs in the Malarkey et al. (1995) paper. This is the same as the
incidence of tumors in the NCI bioassay, which was 43/49 (88%) after 80 weeks of treatment.
Therefore, the incidence of carcinomas in the current study closely matches the experiment done
under similar conditions 18 years earlier in mice of the same strain. Malarkey and co-workers
also found that for the B6C3F1 mice the multiplicity of carcinomas was constant over time at an
average of 1.5 tumors per tumor-bearing animal, whereas the multiplicity of adenomas increased
from about two to four adenomas per tumor-bearing animal from 400 to 568 days. The
multiplicity of both adenomas and carcinomas at the terminal sacrifice was higher in B6C3F1 (5.4
tumors per tumor-bearing animal) than in B6D2F1 mice (3.2 tumors per tumor-bearing animal).
They found that the onset of chronic liver hyperplasia preceded the onset of tumors by over
100 days but that "toxic change" closely paralleled the prevalence of tumors. This indicates that
liver toxicity is not a precursor of tumor development. The severity, but not the prevalence, of
"toxic change" decreased after the prevalence of liver tumors reached 100% in the two strains.
In one group of B6C3F1 mice in which chlordane was withdrawn from the diet after 14 mo,
the multiplicity of carcinomas alone remained at 1.5, but that of adenomas went down 22% from
3.2 to 2.5, and that of combined adenomas and carcinomas went down 30% (from 4.4 to 3.1). In
these experiments, the prevalence of animals with carcinomas went down from 80 to 54%, and
the prevalence of adenomas went down from 100 to 93%, whereas the prevalence of combined
adenomas and carcinomas remained at 100%.
25
-------
There were no mutations of H-ras or K-ras oncogenes in tumors of treated animals of either
strain, whereas H-ras mutations occurred in 4/10 spontaneous tumors in control animals. They
also observed that the proliferative lesions preceding tumor formation in treated animals consisted
of large hepatocytes with acidophilic cytoplasm, whereas in control animals they were of normal
size with basophilic staining. The authors concluded from the oncogene measurements and the
histopathology of the liver lesions that the mechanism of chlordane-induced liver tumors was
different than the mechanism of spontaneous tumors. This evidence, along with the evidence of
regression of some chlordane-induced tumors, indicated to the authors that the tumorigenic
pathways of chlordane-induced tumors are different than those associated with mouse genotoxic
murine hepatocarcinogens. They cited a paper (Fox et al., 1990) that concluded that promoters
or nongenotoxic carcinogens are not believed to directly induce point mutations in the ras gene.
The authors postulate a progression of tumors from a dependent stage, where growth is
dependent on the presence of chlordane, to an autonomous stage, which does not regress after its
withdrawal.
4.3 Reproductive/Developmental Studies—Oral and Inhalation
No multigenerational reproductive studies, by any route, exist for technical chlordane.
Several items within the current chlordane database suggest that reproductive effects could be
a relevant endpoint for chlordane. The study of Cassidy et al. (1994) indicates alterations in
reproductive related behaviors in male rats as a consequence of chlordane exposure that could
affect actual reproductive performance. Data on tissue distribution of chlordane and its
metabolites also indicate the potential for reproductive consequences. Taguchi and Yakushiji
(1988) measured chlordane residues in the milk of 15 mothers who had lived in chlordane-treated
residences for an average period of 1.8 years, all of which were elevated relative to those in the
milk of unexposed mothers. The following residues, in order of their mean concentration, were
detected: frvms-nonachlor, heptachlor epoxide, oxychlordane, c/'s-nonachlor, gamma-chlordene,
trans-chlordane, and c/'s-chlordane. In a subgroup of these women who had been exposed for
approximately 2 years, the overall chlordane residues in milk (0.254 mg/kg milk fat) were similar
to those of PCBs (0.389 mg/kg milk fat). Rani et al. (1992) reported accumulation of heptachlor
(a component of technical chlordane) in the ovary, uterus, and adrenals in nonpregnant rats within
30 min after an oral dose of 120 mg/kg heptachlor. In pregnant rats, levels were markedly
elevated in the uterus compared to nonpregnant rats; the higher accumulation is believed to be
a result of a slower metabolic turnover of heptachlor. These results indicate that chlordane and its
metabolites may have the potential to adversely affect reproductive processes through
bioconcentration in reproductive organs and postnatal development through elimination via milk.
In contrast to the situation for reproductive effects, a number of studies have examined the
potential of chlordane or its metabolites to affect developmental and various postnatal processes
covering a variety of endpoints.
Fischer 344 rats (19 to 23/group) exposed to 0, 21, or 28 mg/kg-day chlordane on
Gestational Days 6 through 19 had significant decreases in maternal weight gain (dose-related,
food intake data not given), but no other clinical signs of maternal toxicity (Narotsky and
Kavlock, 1995). A significant increase in percent loss of pups per litter was seen at both doses;
possible explanations for these findings were functional deficits from prenatal exposure, direct
26
-------
exposure to chemical through maternal milk, or inability of dam to lactate or care for the pups.
The lowest dose, 21 mg/kg-day, is a LOAEL.
In a developmental screening toxicity study, CD-I mice (25 females/group) were exposed to
0 or 50 mg/kg-day chlordane in corn oil via gavage during gestational days 8 through 12 and
monitored through Postnatal Day 3 (Chernoff and Kavlock, 1982). No effects on viability and
postnatal growth were observed in the offspring. No other information was presented.
Pregnant Sprague-Dawley rats were dosed orally with 0, 0.1, 0.5, or 5 mg/kg-day technical
grade chlordane from Day 4 of gestation to Day 21 of lactation, and offspring were dosed from
Postnatal Day 22 to 80 (Cassidy et al., 1994). The hypothesis tested in this study is that
chlordane or its isomers/metabolites act to mimic sex steroids and/or change their concentrations
to alter (in this case to masculinize) functions and behavior. To test this hypothesis, they made
eight behavioral and other functional measurements on male and female offspring. Exposure
during lactation was evidenced by a high concentration of frvms-chlordane, oxychlordane, and
heptachlor epoxide, as well as a smaller amount of heptachlor, in the milk of dams sampled on
Lactation Day 8 which the authors claim is within the time period during which sexual dimorphic
organization occurs. The plasma levels of these compounds in the offspring increased from
Day 40 to Day 80 showing that the animals were exposed during this time. The measurements
and results are listed below.
(1) Plasma chlordane and metabolites: females had greater levels than males at 60 days but there
was no significant difference at 80 days.
(2) Plasma testosterone at 80 days: dose-dependent decrease in females, not significant for
0.1 mg/kg-day but significant at the two highest doses.
(3) Body weights from 25 through 80 days: 5 to 10% increase in weight of females compared to
controls with a non-monotonic dose-response relationship.
(4) Performance in escaping and navigating a water maze in repeated trials on days 76 through
79: treated females had better performance than controls with no consistent pattern as a
function of dose.
(Males did not differ from controls in any of the above tests.)
(5) Open field activity levels at 81 weeks: neither treated males or females differed significantly
from controls.
(6) Auditory startle response: both treated males and females of the low-dose group had
increased response in one of three measures of auditory startle response, but no significant
effect occurred in the other measures or at other doses.
(7) Male mating behavior: low- and mid-dose males performed better than controls but the
improvement of high-dose group males was not significant.
(8) Radiolabeled chloride ion uptake in brain of males: high dose males had reduced uptake of
chloride ion but the results were not stated for the other two dose groups.
Although these results could indicate that chlordane treatment mimics the sex steroids,
masculinizing both females and males, there are several difficulties with this interpretation. For
example, the effects in the high dose group are generally less than the low- and mid-dose groups,
indicating either no effect or a complex dose-response mechanism. Also, the measurements of
testosterone levels show no change in males (even though mating behavior in males is improved),
and smaller and non-significant reductions in testosterone in low-dose females (where the largest
27
-------
effects occur) than in mid- and high-dose females where several effects are less or non-existent.
These observations show that if testosterone or its receptors are somehow involved in the effects
of chlordane, the dose-response model (or mechanism) for the effects must be extremely complex,
and in need of further clarification.
Several investigators (Barnett et al., 1985a,b; Spyker-Cranmer et al., 1982) have assessed
the effects of chlordane on the immunological system of offspring that were exposed during
gestation, and found that chlordane may affect cell-mediated immunity. Pregnant BALB/c mice
(6/group) were fed 0, 0.16, or 8 mg/kg-day chlordane during the 19-day gestation period
(Spyker-Cranmer et al., 1982). At birth, pups were weighed and examined for viability;
alterations in these parameters were not indicated in the study. Pups were randomized within
treatment groups (45 pups/group) and fed normal diets. Offspring were weaned at 28 days, and
immunological tests were performed at 101 days of age. A significant decrease in ear thickness
(0.175 to 0.17 versus 0.065 to 0.05 mm) in response to a challenge with a sensitizing agent
(p < 0.01) was exhibited in a cell-mediated immune assay in the 8 mg/kg-day group in both sexes.
T-cell-dependent humoral immune response to sheep erythrocytes was not significantly different
among groups. The lower dose of 0.16 mg/kg-day is a NOAEL for this endpoint.
Pregnant albino mice (6/group) were exposed orally by gavage to 0, 1, or 2.5 mg/kg-day
technical grade chlordane dissolved in olive oil for 7 consecutive days during late gestation
(i.e., gestational days 12 through 19 (Al-Hachim and Al-Baker, 1973). Each mouse received
5 to 7 doses. It was not indicated whether animals were nursed or foster reared, and thus whether
exposure also occurred through the milk. Groups of 10 randomly selected offspring were tested
for conditioned avoidance response (<37 days of age), electroshock threshold (38 days of age),
and response in open-field tests (6 weeks of age). Exposed animals exhibited depressed
acquisition of avoidance response (mean responses 13.1, 9.7, and 9.7 in controls, low- and
high-dose), increased seizure threshold (mean responses 90.1, 108.6, and 134.9 in controls,
low- and high-dose), and increased exploratory activity (mean responses 93.9, 88.4, and 137.7
in controls, low- and high-dose). A three-way ANOVA was performed and indicated significant
dose-related differences between treatment groups (p < 0.001). The lowest dose, 1 mg/kg-day,
is a LOAEL.
In an effort to understand the effects of prenatal chlordane exposure on adult bone marrow
expansion potential, Barnett et al. (1990a) exposed female mice orally to chlordane at either 0, 4,
or 8 mg/kg-day for 18 days during pregnancy. Offspring were nursed, which would provide some
postnatal chlordane exposure. Bone marrow hematopoietic activity, as measured by the ability of
bone marrow cells to undergo clonal expansion in response to stimulating factors, and spleen
colony forming units (after irradiation) were both evaluated in offspring of these mice at 100 and
200 days of age. Results showed a significant dose-related depression (p < 0.05) of both
measures at 100 and 200 days of age. The bone marrow of offspring exposed to 4 or 8 mg/kg
chlordane had 63 to 75% of control numbers of granulocyte-macrophage colony forming units
(CPU) at 100 days of age; at 200 days postexposure these values were still significantly decreased
at 50 to 77% of control. The formation of splenic-CFU was similarly affected. Bone marrow of
male and female offspring exposed to 4 mg/kg chlordane had only 78 to 87% of control splenic-
CFU at 100 days of age and those exposed to 8 mg/kg had only 67 to 64%. At 200 days of age
female mice exposed to 8 mg/kg still had only 72% of control splenic-CFU and male mice
28
-------
exposed to 4 mg/kg only 84% of control splenic-CFUs. In a subsequent study in which these
same measures were evaluated at 18 days of gestational age (and without chlordane exposure via
milk) these results were confirmed indicating that damage to stem cells occurred during the fetal
period (Barnett et al., 1990b). A LOAEL of 4 mg/kg-day is indicated by these results.
Barnett et al. (1985a) found that in utero exposure to 8 or 16 mg/kg-day chlordane (in
peanut butter) in pregnant BALB/c mice during gestational days 1 through 18 both resulted in
a significant decrease (p < 0.05) in virus-specific delayed type hypersensitivity response
postinjection of an influenza type A virus (degree of footpad swelling) in the offspring, but only at
48 h after virus injection. At 48 h post-injection, the control infected animals had a change in
swelling of 1.17 mm, the 8 mg/kg group 0.54 mm, and the 16 mg/kg group 0.46 mm. There was,
however, no deficit in the chlordane treated animals in ability to resolve the infection. In another
mouse study by Barnett et al. (1985b), delayed type hypersensitivity response in offspring, as well
as depressed mixed lymphocyte reactivity of spleen cells in male offspring, occurred after in utero
exposure to 4 or 16 mg/kg-day during gestational days 1 through 19. In the first study 8
mg/kg-day is an effect level and in the second study 4 mg/kg-day is an effect level. For both of
these studies, offspring were also exposed to chlordane via milk.
Cranmer et al. (1984) administered chlordane (in peanut homogenate) at doses of 0, 0.16, or
8 mg/kg-day to dihybrid female mice throughout gestation to assess plasma corticosterone levels
in adult offspring (45 pups/group). At birth, pups were weighed and examined for birth defects;
malformed, moribund, or small newborns were removed from further study. Offspring were
sacrificed at 101, 400, or 800 days of age. Due to the high number of postnatal deaths in the
high-dose offspring (55% died within the first week), there were no male offspring left for
sacrifice at 800 days. At 101 days of age, plasma concentration of corticosterone was elevated
(p < 0.01) in 0.16 mg/kg-day males at 6.5 jig/dl plasma as compared to 2.6 to 3.2 jig/dl plasma in
controls. At 400 days of age, corticosterone levels remained significantly elevated (p < 0.05) in
0.16- (6.8 versus 4.5 jig/dl plasma) and 8-mg/kg-day males (8.1 versus 4.5 jig/dl plasma) and
0.16-mg/kg-day females (10.8 versus 6.8 |ig/dl plasma). The investigators hypothesized that
elevated plasma corticosterone levels may reflect the diminished ability of the liver to
metabolically reduce corticosterone, or that the elevated levels may possibly be due to
perturbations in corticosteroid production by the hypothalamus or pituitary. Chlordane may be
impairing normal development of the neuro-endocrinological feedback system, leading to transient
or permanent malformations of the system in the adult animal. The investigators also suggested
that the more marked effect on plasma corticosterone levels in male mice compared to female
mice may be a result of chlordane's interference with sex hormone control. As the toxicological
significance of these effects are unclear, no effect levels are assigned.
4.4 Other lexicologically Relevant Studies
4.4.1 Mechanistic Studies
29
-------
The findings of Malarkey et al. (1995) give indications of the mechanism of tumor
formation. The authors exposed B6C3F1 and B6D2F1 mice in a discontinuous protocol (see
Section 4.2), evaluating tumor characteristics and progression with continuous administration of
chlordane and following the withdrawal of dietary administration. They found no H-ras or K-ras
mutations in hepatocellular tumors from chlordane-exposed B6C3F1 or B6D2F1 mice, whereas
H-ras mutations were found in 4/10 spontaneous hepatocellular tumors examined from control
B6C3F1 and in 6/15 liver tumors from control B6D2F1 mice. This observation of lower H-ras
mutation frequency in tumors from exposed mice than in spontaneous tumors from control mice
has been observed for other nongenotoxic agents that cause liver tumors in mice, including
dieldrin, Arochlor, phenobarbital, and chloroform. The majority of neoplastic hepatocytes and
altered hepatocellular foci in exposed mice had acidophilic cytoplasm, whereas neoplastic
hepatocytes and altered hepatocellular foci in controls had basophilic cytoplasm. These two
findings suggest that chlordane-induced tumors arise from different cell phenotypes than
spontaneous tumors. Also, the multiplicity of combined adenomas and carcinomas as well as that
of adenomas decreased after discontinuation of treatment. These results provide evidence for a
progression of neoplasms from a dependent to an autonomous stage of development such that
only the latter persist after the cessation of exposure.
4.4.2 Genotoxicity
The U.S. EPA (1986b) previously reviewed and evaluated published genotoxicity studies on
chlordane. Approximately two-thirds of 25 published studies were judged unacceptable or
inconclusive due to study deficiencies or reporting deficiencies. Several conclusions, listed below,
were drawn based on the remaining nine studies.
(1) Chlordane was not mutagenic in bacterial assays, with or without mammalian activation
systems.
(2) Data obtained for chlordane-induced gene mutations in in vitro tests with mammalian cells
were equivocal.
(3) Chlordane produced negative results in mouse dominant lethal assays, in an in vitro
mammalian cell cytogenetics assay, and in vitro DNA repair assays using mouse, rat, and
hamster hepatocytes.
(4) Chlordane induced sister chromatid exchanges in fish and human cells in culture, and gene
conversion in yeast.
The U.S. EPA (1986b) concluded that the available genotoxicity data did not "provide a fully
comprehensive data set upon which to definitively assess the mutagenic potential of chlordane."
In a recent review of chlordane genotoxicity studies, Jackson et al. (1993) concluded that
three studies provided sufficient evidence to classify chlordane as having "limited evidence of
mutagenicity." These studies found weakly positive results for gene mutation in V-79 cells
(Ahmed et al., 1977), positive results in the mouse lymphoma assay (McGregor et al., 1988),
and positive results for the induction of sister chromatid exchanges in cultured human
lymphocytes (Sobti et al., 1983).
4.5 Synthesis and Evaluation of Major Noncancer Effects and Mode of Action—Oral
and Inhalation
30
-------
The overall hazard profile evaluation for chlordane is one of moderate toxicity. Chronic
exposure, both oral and inhalation, to technical chlordane has been documented to produce
hepatic effects in animals, neurotoxic effects in humans, and possibly to modify sex
steroid-mediated behaviors in animals. Only the hepatic effects have both reliable dose-response
information and a theoretical mode-of-action for which there is supportive data.
Subchronic and chronic studies in animals consistently demonstrate that exposure to
technical chlordane elicits both adaptive and adverse hepatic effects. Technical chlordane and its
metabolites are lipophilic and have been demonstrated to accumulate within fatty tissues in both
animals and humans. A principal metabolite of technical chlordane, generated by enzyme systems
localized largely in hepatic tissues, is the epoxide, oxychlordane. Although the tissue reactivity of
this epoxide has not been thoroughly studied, it (like other epoxides) has the potential to
covalently interact with cellular constituents and macromolecules, which may lead to more
obvious and expressed toxic effects. The long residence of this epoxide in fatty tissues makes this
species suspect in manifesting these effects. This mechanism (meaning the molecular sequence of
events leading to toxic effects) could reasonably be explanatory of the observed liver toxicity in
animals exposed chronically to chlordane. That this mechanism could apply to humans is
supported by the detection of this epoxide in human tissues. However, the human studies
available, although limited in their scope of information, give no further evidence that this
mechanism is operative as no liver toxicity is apparent. Several of these studies looked
specifically for this adversity.
In humans as well as animals neurotoxicity has been identified as the most likely endpoint in
acute exposures to chlordane (Grutsch and Khasawinah, 1991) and for other chlorinated
cyclodienes such as aldrin and dieldrin (Murphy, 1986). Some evidence also exists for blood
(Fleming and Timmeny, 1993) and immunological (McConnachie and Zahalsky, 1992) effects as
being relevant in acute exposures. Although the mechanistic bases for these effects are not
known, biochemical perturbations have been implicated (Murphy, 1986). Effects in humans
resulting from chronic exposure to chlordane are not clear. Occupational studies available on
highly-exposed workers report no neurological symptomatologies although this conclusion was
reached from limited examinations or questionnaires. In contrast to these reports, studies of
individuals exposed from treatment of their residences document neurophysiological and
neuropsychological deficits (Kilburn and Thornton, 1995) from exposure via dermal and
inhalation routes. No dose-response information and little reliable exposure information is
available to confirm and quantify these chronic effects in humans. The RfC proposed in this
assessment (8E-5 mg/m3, Section 5 below) is below that concentration (> 0.5 |ig/m3 or
5E-4 mg/m3) associated with neurological effects in the Kilburn and Thornton (1995) study.
The preliminary results of Cassidy et al. (1994) indicate that chlordane may masculinize
certain behaviors in female rats suggesting that chlordane may chemically mimic androgenic
steroids. If confirmed, these findings would indicate the potential for reproductive effects.
Alterations of some reproductive behaviors were noted among males in this study but no parallel
tests were carried out with females.
From these observations, it is apparent that a multigenerational reproductive animal study
could be considered a data gap in the chlordane database, as well as both subchronic and chronic
31
-------
animal studies in which neurotoxicity is carefully evaluated. Follow-up studies on
neuropsychological and neurophysiological characteristics of chlordane plant workers with known
exposures would also serve to provide confirmatory and possibly quantitative dose-response
information on human neurotoxicity from chronic exposure to technical chlordane. This would be
possible only if former plant workers with known exposure levels and employment records could
be found.
4.6 Weight of Evidence Evaluation and Cancer Classification
The overall weight of evidence for the carcinogenicity of chlordane to humans consists of
information from humans supplemented by information from animal and in vitro systems designed
to model particular aspects of human biology.
There are two main lines of evidence that chlordane is likely to be a human carcinogen.
First, liver tumors occurred in five strains of mice given chlordane in the diet. They occurred in
both males and females, and the incidence was closely replicated in an experiment repeated
18 years after a 1977 NCI bioassay. Tumors were not observed in liver of rats or other sites in
mice, although rats had liver enlargement and cell swelling, which can be signs of preneoplastic
toxicity. Five compounds structurally related to chlordane (aldrin, dieldrin, heptachlor, heptachlor
epoxide, and chlorendic acid) have induced liver tumors in mice. Chlorendic acid has also
produced liver tumors in rats. Although there has been controversy about the relevance of the
chemically-induced mouse liver tumor as a model of human carcinogenicity, this evidence firmly
establishes chlordane as a mouse liver tumor carcinogen.
The second line of evidence is in humans exposed to chlordane by direct dermal contact
and/or inhalation at high concentrations over a prolonged time. This evidence is too weak to
conclude that chlordane is causally related to human cancer, but it is suggestive that such
exposure might be associated with excess risk to a broad range of hematopoietic diseases. This
evidence consists of: (1) excess of NHL in Iowa and Minnesota farmers using chlordane as an
insecticide for livestock and crops, and an elevated but not significant excess odds ratio for NHL
in agricultural and forestry workers in Washington State; (2) several human case reports of blood
dyscrasias (aplastic anemia and leukemia) in people exposed either by handling chlordane at home
or by pesticide applicators; (3) a report of impairment of lymphocyte proliferation and suggestive
autoimmunity in people exposed to chlordane as a home termiticide; and (4) data showing that
people exposed at home have chlordane metabolites in blood. There is no evidence that exposure
to chlordane in food or drinking water is associated with cancer.
There is limited evidence of genotoxicity of chlordane, consisting of positive mutagenicity
findings in V-79 cells, a positive mouse lymphoma assay, and sister chromatid exchange induction
in human peripheral lymphocytes, but no evidence in bacterial assays or in any in vivo assays.
Whereas the genotoxic potential of chlordane is not clearly understood, several
toxicological properties have been described which may play roles in the expression of chlordane
carcinogenicity in rodents, including: chlordane induction of hair follicle nuclear aberrations in
CD-I mice (Schop et al., 1990); irreversible binding of chlordane metabolites to intracellular
macromolecules, including DNA and RNA (Brimfield and Street, 1981); chlordane inhibition of
32
-------
intercellular communication (Telang et al., 1982; Ruch et al., 1990; Bessi et al., 1995); chlordane
stimulation of protein kinase C activity (Moser and Smart, 1989); chlordane induction of in vitro
hepatic lipid peroxidation and DNA single-strand breaks (Hassoun et al., 1993); and chlordane
suppression of in vitro immune responses (Johnson et al., 1987; Chuang et al., 1992).
A search for H-ras and K-ras mutations in liver tumors induced by chlordane in two mouse
strains showed no mutations, whereas spontaneous tumors in the same two strains did show
mutations; this result indicates that chlordane induces tumors by different mechanisms than those
responsible for spontaneous tumors. The absence of ras mutations has been observed in tumors
induced by other non-genotoxic carcinogens. One recent experiment showed that a fraction of
liver tumors induced in mice by chlordane regress (their multiplicity decreases) after withdrawal of
chlordane from the diet; other tests at lower doses are needed to further explore the mechanism of
this regression.
The mode of action for chlordane as a carcinogen is probably not solely as a direct
genotoxin, although DNA damage has not been ruled out. Since it is persistent in body tissues as
an epoxide (ocychlordane), it could affect carcinogenesis at many stages. There is evidence that it
induces malignant liver tumors by a process independent of the induction of spontaneous tumors,
and a suggestion that it might induce two populations of liver tumors, one of which is persistent
after the withdrawal of continuous dosing and the other population regressing after cessation of
continuous dosing. Since these experiments were done only at doses high enough to induce
tumors in 100% of the animals, the low-dose mode of action is not known.
There have been no populations identified as unusually sensitive to the carcinogenic effects
of chlordane. The people at most risk are the ones most highly exposed; these are farmers who
have handled, mixed and applied chlordane to livestock and crops, and people exposed in their
homes to improperly applied chlordane as a termiticide or in garden uses of chlordane not yet
disposed since sales were discontinued in 1988. No effects have been observed in humans from
the small concentrations of chlordane that commonly occur in food.
The main uncertainty about the current risk assessment is whether the effects observed in
highly exposed people will occur from the low concentrations found in food or in water and air
near waste sites where chlordane is found, and if it does occur, what is the probability of cancer
due to that exposure. The estimate of this probability is derived from data on liver cancer in mice,
although the relationship of mice liver tumors to cancers of an unknown site in humans is not
known. There are no experimental data which provide information about the probability that
cancer is induced by low chlordane concentrations; in the absence of such information the linear,
no-threshold dose extrapolation model is the appropriate model for low-dose extrapolation. The
persistence of chlordane in soils and fish and in fat and bone marrow after it has been ingested
serves as a rationale for protective assumptions.
The mechanistic relationship between liver tumors in mice and hematopoietic toxicity and
lymphoma in humans has not been established. However, in the absence of definite information, it
is prudent to regard the liver cancer in mice as an indication that chlordane may be carcinogenic in
humans.
33
-------
This evidence matches closely the criteria of the 1986 Guidelines for Carcinogen Risk
Assessment for a B2 carcinogen (probable) because of inadequate evidence of carcinogenicity in
human studies, sufficient evidence of hepatocellular carcinomas in multiple mice strains, and the
evidence of several toxicological responses believed to play a role in carcinogenesis. Using the
1996 Proposed Guidelines for Carcinogen Risk Assessment, chlordane would be characterized as
a likely carcinogen in humans, and the above paragraphs would constitute the weight-of-evidence
narrative.
4.7 Other Hazard Identification Issues
4.7.1 Possible Childhood Susceptibility
A number of factors may differentially affect the response of children to toxicants such as
chlordane. These factors include diet and physical environment as well as maturation of
physiological and biochemical processes (Roberts, 1992). Although some work has been
attempted in this area for exposure to organochlorines (McConnell, 1992; Fenske, 1992), there is
too little information to make any statements about how these factors may specifically affect the
toxicological responses of chlordane in children, be they cancer or noncancer.
4.7.2 Possible Gender Differences
The extent to which men differ from women in susceptibility to chlordane toxicity is not
known. Although gender differences in the effects of chlordane have been observed in the studies
reviewed in this document (listed below), they are rare and show no consistent pattern.
(1) In the 90-day Wistar rat inhalation study (Khasawinah et al., 1989a), females but not males
had transient changes in white cell populations that were not considered adverse or
dose-related and disappeared 13 weeks after exposure.
(2) In the 30-mo Fischer rat study (Khasawinah and Grutsch, 1989b), hepatocellular hypertrophy
was more prevalent in females than males.
(3) In the 24-mo ICR mice study (Khasawinah and Grutsch, 1989a), the background incidence
and the dose-related trend of liver adenoma and hemangioma incidence was larger in males
than females.
(4) In measurements of plasma corticosterone following in utero exposure, Cranmer et al. (1984)
found greater elevation in males than females, but the significance of the finding is not clear.
Gender differences associated with other agents which cause the same toxic effects seen
here with chlordane, namely liver damage and liver cancer, might offer some insights about gender
differences expected with chlordane. The six observations summarized below (from Calabrese,
1996) may have some bearing on gender differences expected with chlordane, but the situation is
not clear enough to make predictions in the absence of data on chlordane itself.
(1) Liver cancer incidence from aflatoxin Bl exposure in African populations is higher in men
than women.
(2) In rats of several strains the liver metabolism of aflatoxin Bl to the active forms Ml and Ql is
greater in males than females, and the DNA binding of these metabolites is greater in males.
(3) In the United States and Europe liver cancer is about two-fold higher in men than women;
whether this is due to differential exposure to liver carcinogens is not known.
34
-------
(4) Levels of hepatic aryl hydrocarbon hydroxylase, specifically P4501A1, are higher in male than
female rats. These are enzymes that can both activate and detoxify carcinogenic agents.
(5) Female rats are more susceptible than males to carbon tetrachloride-induced liver damage.
(6) Benzene-induced toxicity to the hematopoietic system is greater in female rats than males,
presumably because benzene is distributed more readily into body fat and is retained in tissues
of females for a longer time.
Listed below are three other findings that may have some relevance to gender differences.
(1) Chlordane acts synergistically with other pesticides in binding to estrogen receptors
incorporated into a yeast system (Arnold et al., 1996), although synergism apparently does
not occur with weakly estrogenic chemicals when tested in cells with a high concentration of
estrogen receptors (Ramamoorthy et al., 1997). The Arnold et al. (1996) article was formally
withdrawn by the authors in a 1997 letter to Science (McLachlan, 1997) because they could
not reproduce their findings. Therefore this point is not valid.
(2) Male B6C3F1 mice have higher and more variable spontaneous liver tumor incidence than
females (U.S. EPA, 1986a).
(3) There are pronounced gender differences in the clinical action of psychotropic (Yonkers
et al., 1992) and other drugs (Harris et al., 1995). These differences are of interest because of
the neurotoxic action of chlordane noted above.
5.0 DOSE RESPONSE ASSESSMENTS
5.1 Oral Reference Dose
5.1.1 Choice of Principal Study and Critical Effect—with Rationale and Justification
Chlordane has a particularly wide base of chronic studies which consistently demonstrate
that the liver is the target organ. Most of these studies, however, have limitations. The IRDC
(1973) study in mice was conducted for only 18 mo at high doses (up to 7.1 mg/kg-day) that
resulted in excessive mortality. The NCI (1977) study was also carried out at high doses for
around 20 mo, with the dose levels being adjusted during the exposures to avoid excessive
toxicity. The Barrass et al. (1993) and Malarky et al. (1995) studies were carried out at a single
dose. In contrast, the Kasawinah and Grutsch chronic rat study (1989b) was carried out over
a wide range of exposure values which included dosing for an extended period (30 mo),
an extensive and thorough examination of the target organ (liver), and considering a major
confounding factor, leukemia. However, no clearly adverse effects, hepatic or otherwise, were
noted. This situation is in contrast to the 24-mo mouse study (Khasawinah and Grutsch, 1989a)
wherein hepatic necrosis is observed at a dose level free of any hepatic neoplasia. The occurrence
of such an adverse lesion in mice but not in rats under approximately equal dose-rates indicates
that mouse livers may be more sensitive to the effects of chlordane or its metabolites than are rat
livers. The dose-related occurrence of an adverse effect in the livers of mice justifies the choice of
the Khasawinah and Grutsch (1989a) chronic study in mice as the principal study.
The extensive animal data base consistently demonstrates that the liver is the target organ
for chlordane. Hepatic effects also were noted in inhalation studies in rats (Khasawinah et al.,
35
-------
1989a) indicating that the liver is a common target organ regardless of route of chlordane
administration. As they show progression with dose, from nonadverse to clearly adverse, and as
they are consistent throughout the data base for chlordane, hepatic effects are appropriate as the
critical effect in this assessment. Hepatic necrosis is judged to be the most clearly adverse
noncancerous lesion of those reported, which included fatty degeneration and increased liver cell
volume (hypertrophy). The latter effect is commonly associated with ultrastructural cellular
changes involving increased metabolic capacity due to the presence of the toxicant (Sipes and
Gandolfi, 1991) and are more appropriately considered as nonadverse, adaptive changes.
5.1.2 Methods of Analysis—Benchmark Dose and No-Observed-Adverse-Effect Level/Lowest-
Observed-Adverse-Effect Level
The NOAEL for hepatic necrosis in male mice in the principal study (Khasawinah and
Grutsch, 1989a) is 0.15 mg/kg-day (1 ppm), the LOAEL is 0.75 mg/kg-day (5 ppm). This study
was described in Section 4.2, page 24.
Benchmark dose (BMD) analyses were performed on the incidence of hepatic necroses in
male mice. Specifics on models, data input and calculations are included in Appendix A as are
results documenting that mouse liver necroses were not modeled acceptably by either benchmark
model employed. The NOAEL/LOAEL approach is used for all subsequent quantitation in the
RfD dose-response assessment.
5.1.3 Oral Reference Dose Derivation—Including Application of Uncertainty Factors and
Modifying Factors
Uncertainty factors (UF) are applied to account for recognized uncertainties in extrapolation
from experimental conditions to the assumed human scenario (i.e., chronic exposure over a
lifetime). Historically, UFs are applied as values of 10 in a multiplicative fashion (Dourson and
Stara, 1983). Recent EPA practice, however, also includes use of a partial UF of 101/2 (3.333)
(U.S. EPA, 1994b) on the assumption that the actual values for the UFs are lognormally
distributed. Application of these factors in the assessments is that, when a single partial UF is
applied, the factor is rounded to 3, such that the total factor for a UF of 3 and 10, for example,
would be 30 (3 x 10). When two partial UFs are evoked, however, they are not rounded, such
that a UF of 3, 3, and 10 would result in a total uncertainty of 100 (actually 101/2 x 101/2 x 1Q1).
The NOAEL for male mice chronically exposed to chlordane is 0.15 mg/kg-day. The
following full UFs are applied to the NOAEL: 10 for extrapolation for human variability
(intraspecies differences in response), 10 for consideration of animal to human (interspecies)
extrapolation. A partial (101/2) UF is also applied for lack of any full scale multigenerational
reproductive study. The total UF = 10 x 10 x 1Q1/2 = 300. No modifying factor (MF) is
considered necessary.
RfD = 0. 0.15 mg/kg-day - 300 = 5E-4 mg/kg-day.
5.2 Inhalation Reference Concentration
36
-------
5.2.1 Choice of Principal Study and Critical Effect—with Rationale and Justification
The subchronic study of Khasawinah et al. (1989a,b,c) is the only long-term repeated
inhalation studies currently available in the database. The subchronic study included both
monkeys and rats with rats being used for further analysis. Although monkeys could be more
comparable to humans for exposure purposes, it is not at all clear if monkeys would be more or
less sensitive than humans to the effects of chlordane and its isomers and metabolites. It is clear
from this study that effects were shown in rats and not monkeys when both species were exposed
simultaneously in the same chambers to the same concentrations of chlordane. To base the RfC
on the monkey data would require that the highest concentration, 10 mg/m3, be used as a
"free-standing" NOAEL, a procedure that has not historically been a practice in the IRIS process.
The incentive to use the monkey data as a direct surrogate for humans is ameliorated further due
to the RfC methodology (U.S. EPA, 1994b) allowing estimation of a human equivalent
concentration (HEC) from the rat data. It is, however, interesting to note that the HEC
dosimetric adjustment (done below) implies that in inhaling the same external concentration,
as was the case in this experiment, humans and presumably monkeys would receive less dose than
would rodents. The anticipated result from this situation, that rats would respond to lower
external concentrations than monkeys, is precisely what was seen in this study.
The animal data base consistently demonstrates that liver is the target organ for chlordane.
Although some mild thyroid histopathology also was reported in the subchronic study, hepatic
effects also were noted in chronic oral studies in rats (Khasawinah and Grutsch, 1989b) and mice
(Khasawinah and Grutsch, 1989a) indicating that liver is a common target organ regardless of
route of chlordane administration. As hepatic effects show progression with dose, from
nonadverse to clearly adverse, and as they are consistent throughout the data base for chlordane,
hepatic effects are appropriate as the critical effect in this assessment.
Hepatic-related alterations in blood chemistry are designated as the critical effect. These
are alterations in cholesterol, albumin, globulins and total proteins. In the 28-day inhalation pilot
study in rats (Khasawinah et al., 1989b) a progression of hepatic effects was seen. At the highest
concentrations adverse hepatocellular pathologies (i.e., vacuolation and necrosis) were noted.
At lower concentrations minor hepatic effects (centrilobular hepatocellular hypertrophy) and an
entire pattern of hepatic-related alterations in blood chemistry were noted, these latter alterations
being the critical effect and basis for the LOAEL in this short-term study. Although the
concentration-range in the subchronic study (Khasawinah et al., 1989a) was apparently not high
enough to elicit hepatic vacuolation or necrosis, a pattern of hepatic-related alterations in blood
chemistry nearly identical to those observed in the 28-day study was noted at the highest
concentration used in the subchronic study, 10 mg/m3. The pattern of alterations in blood
chemistry observed in the subchronic and 28-day pilot studies (Khasawinah et al., 1989a,b) such
as cholesterol, albumin, globulins and total proteins are indicative of dysfunction of the liver, the
major site of synthesis of plasma proteins (Kaneko, 1989). The other effect described in the
subchronic study, reversible centrilobular hepatocyte hypertrophy, is commonly associated with
ultrastructural cellular changes involving increased metabolic capacity due to the presence of the
toxicant (Sipes and Gandolfi, 1991) and are more appropriately considered as nonadverse,
adaptive changes.
37
-------
5.2.2 Methods of Analysis—No-Observed-Adverse-Effect Level/Lowest-Observed-Adverse-
Effect Level and Benchmark Concentration
The LOAEL for the rat portion of the principal study (Khasawinah et al., 1989a) was based
on blood chemistry indicators of hepatic dysfunction that were restricted to rats exposed to the
highest concentration of chlordane at 10 mg/m3. The corresponding NOAEL was 1 mg/m3.
Analyses for benchmark concentrations (BMC) were not attempted as the critical effect
(biochemical indicators) was a collection of biochemical alterations that occurred only at a single
concentration and therefore was not amenable to BMD analysis. The critical effect comprises
hepatic-related alterations in blood chemistry including a decrease in glucose and increases in
globulins, albumin, cholesterol, and total protein all of which occurred at the highest
concentration (10 mg/m3) only. Higher chlordane concentrations used in the 28-day pilot study
demonstrate further progressive changes in hepatic function and histopathology and may be
considered as evidence for a concentration-response relationship between chlordane and hepatic
effects. The differences in the duration of these studies precludes combining them for a formal
analysis such as a benchmark dose. Therefore, the NOAEL/LOAEL approach will be used for all
further analyses.
The NOAEL of 1 mg/m3 is duration adjusted to a continuous exposure = NOAEL(ADJ):
1 mg/m3 x (8 h/24 h) x (5/7 days) = 0.24 mg/m3.
5.2.3 Inhalation Reference Concentration Derivation—Including Application of Uncertainty
Factors and Modifying Factors
The NOAEL for hepatic effects in rats in the principal study of Khasawinah et al. (1989a) is
1 mg/m3. The NOAEL(ADJ) would be calculated based on an 8 h daily exposure for 5 of 7 days
per week = 1 mg/m3 x 8/24 x 5/7 = 0.24 mg/m3.
The methodology for development of reference concentrations (U.S. EPA, 1994b) allows
for dosimetric adjustment of particulate aerosols between animals and man to arrive at an HEC.
For this particulate exposure of chlordane, specific information on particulate size and distribution
is requisite to this adjustment. Particle sizing information included in an unpublished version of
this study (Velsicol Chemical Corporation, 1984) was used to estimate a MMAD of 1.8 jim and a
corresponding particle distribution estimate (sigma g) of 3.1. These parameters were obtained by
plotting the given size cut-offs against the cumulative mass; 51.4% of the mass at 2 jim and
82.9% of the mass at 5.5 jim. Although the data available to estimate these particle parameters
were limited to 2 data points, it should be noted that one point (51.4% of total mass at <|im) was,
by definition, inclusive of the MMAD. These estimates of the particle characteristics were thus
judged reasonable estimates as the RDDR for an extrarespiratory effect (ER), the RDDR(ER), of
2.7 obtained from these particle parameters (U.S. EPA, 1994b).
NOAEL(HEC) = NOAEL(ADJ) x RDDR
NOAEL(HEC) = 0.24 mg/m3 x 2.7 = 0.65 mg/m3
38
-------
Uncertainty factors are applied to account for recognized uncertainties in extrapolation from
experimental conditions to the assumed human scenario (i.e., chronic exposure over a lifetime).
Historically, UFs are applied as values of 10 in a multiplicative fashion (Dourson and Stara,
1983). Recent EPA practice, however, also includes use of a partial UF of 101/2 (3.333) (U.S.
EPA, 1994b) on the assumption that the actual values for the UFs are lognormally distributed.
Application of these factors in the assessments is that, when a single partial UF is applied, the
factor is rounded to 3, such that the total factor for a UF of 3 and 10, for example, would be 30
(3 x 10). When two partial UFs are evoked, however, they are not rounded, such that a UF of 3,
3, and 10 would result in a total uncertainty of 100 (actually 101/2 x 101/2 x 101).
The following full UFs are applied to this effect level: 10 for subchronic to chronic
extrapolation, 10 for consideration of intraspecies variation (human variability), a partial UF
is applied for interspecies extrapolation as an HEC has been calculated, and another partial UF
is applied for lack of any reproductive studies. The total UF = 10 x 10 x io'/2 x 1Q1/2 = 1000.
No MF is proposed for this assessment.
RfC = NOAEL(HEC) / UF = 0.65 mg/m3 / 1000 = 7E-4 mg/m3
5.3 Cancer Assessment
No reliable human dose information is available in the case reports of blood dyscrasias or in
the case-control studies of farmers using chlordane. Therefore, quantitative estimates based on
human data are inappropriate.
For animals there are incidence data for hepatocellular carcinoma in mice in three studies
constituting five data sets (males and females in two of the studies). In one of these studies
adenomas were reported in addition to carcinomas; these were not counted in order to be
consistent with the other data sets and because carcinomas by themselves constitute a potential
human health hazard. Because of this omission of adenomas, the risk of chlordane may be slightly
underestimated in comparison to other mouse liver carcinogens where typically adenomas are
combined with carcinomas to obtain a risk estimate.
5.3.1 Choice of Study/Data with Rationale and Justification
The slope estimates from the five dietary mouse liver tumor data sets (IRDC [1973] males
and females, NCI [1977] males and females, and Khasawinah and Grutsch [1989a]) were
combined by taking the geometric mean. These slope estimates are derived in Appendix A.
The justification for the mean is that there is no a priori reason that any one data set is more
appropriate than the others for human risk estimation. The two data sets from the IRDC 1973
were more difficult to fit to a multistage model than the other three, but omitting them from the
average produces little change.
5.3.2 Dose-Response Data
39
-------
The dose-response data are given in Tables 2, 3, and 5 of this document and the details on
the multistage model fit to the data are given in Appendix A.
5.3.3 Dose Conversion
The two steps in estimating the human equivalent dose from the animal dose are:
(1) converting the dietary concentrations to milligrams per kilograms per day in the animals;
and (2) converting the animal dose to the equivalent human dose. The first step had been done
in the IRDC and Khasawanah and Grutsch studies and the assumption of a default value of 0.13 x
concentration was made for the NCI data. The second step used the "3/4 power" assumption
where the human equivalent dose, in units of milligrams per kilograms body weight per day,
is equal to the animal dose x (w/70)A(l/4), and the animal weights, w, were taken from the
NCI report for that data set and the default assumption of 0.030 kg was used for the other two
data sets in absence of information.
In estimating the inhalation unit risks, the assumption is made that effects in animals depend
on the inhaled dose (milligrams per kilograms per day) and that there is 100% absorption. This
assumption was checked by comparing the dose at which a 10% incidence of liver cell
hypertrophy was observed in the 13-week rat inhalation study of Khasawinah et al. (1989a) with
the dose at which the same effect was observed in the 30-mo rat dietary study of Khasawinah and
Grutsch (1989b). In section 4.2 it was shown that these values were 0.64 mg/m3 for the
inhalation air concentration and 0.15 mg/kg-day for the dietary study. The dose absorbed by rats
breathing an air particulate concentration of 0.64 mg/m3 with a MMAD of 1.8 jim and a sigma g
of 3.1 is the product of concentration, minute volume, and fraction of parti culates deposited
divided by the body weight. The minute volume (I) was estimated from the formula I = 0.8 x
(W)A(0.821) = 0.227 m3/day for a 0.216 kg rat (U.S. EPA, 1988a). The fraction of particles
deposited as calculated from the standard Agency model (U.S. EPA, 1994b) is 0.45. Therefore,
the deposited dose is
(0.64 mg/m3 x 0.227 m3/day x 8/24 x 5/7 x 0.45 x 1)/(0.216 kg) = 0.073 mg/kg-day.
This is about one-half the dose at which rats in the chronic dietary study had liver cell
hypertrophy. Although one would expect that a slightly smaller dose would be sufficient to cause
equivalent effects in a chronic study, the similarity of the two dose estimates (0.15 mg/kg-day for
the oral study and 0.073 mg/kg-day from the inhalation study) is noteworthy and warrants the
conclusion that liver cell hypertrophy occurs at approximately the same absorbed dose regardless
of the route of administration.
Based on this analysis, the inhalation unit risk for humans is derived from the oral slope
factor by estimating the intake in humans assuming 100% absorption and a breathing rate of
20 m3/day in the normal fashion.
5.3.4 Extrapolation Methods
The linearized multistage model, Global 86, was used to calculate the upper limit on Q(l)
and the lower confidence limit on the dose corresponding to an incidence of 0.1. The parameters
40
-------
used were the extra risk option and the degree of polynomial determined by the algorithm of
Global 86. Details are presented in Appendix A.
5.3.5 Oral Slope Factor and Inhalation Unit Risk
The oral slope factor, 3.5E-1 per mg/kg-day, is the geometric mean of the following five
slope factors from the individual data sets: CD-I males, 0.858; CD-I females, 0.217; B6C3F1
males, 0.345; B6C3F1 females, 0.114; and ICR males, 0.710. If the calculation were to be done
according to the proposed cancer guidelines, which uses a straight-line extrapolation from the
LED10 point, the corresponding geometric mean would be 5.7E-1 per mg/(kg-day). This slope is
slightly higher than the extrapolation to small doses, reflecting the positive curvature
("sub-linearity") of the dose-response data. Both estimates are included here pending the
fmalization of the currently-proposed guidelines, but the former is used here because the
1986 guidelines have not been replaced.
The inhalation unit risk is 1E-4 per (|ig/m3), derived from the oral slope factor and assuming
a human air intake of 20 |ig/m3 per day, 100% absorption of inhaled chlordane, and a human body
weight of 70 kg. The justification for the conversion from oral to inhalation dose is given above
in Section 5.3.3.
41
-------
6.0 MAJOR CONCLUSIONS IN CHARACTERIZATION OF HAZARD AND
DOSE-RESPONSE
6.1 Hazard Identification
Technical chlordane is a mixture of over 140 different compounds with a carbon skeleton
consisting of joined 5- and 7-carbon rings and having varying degrees of chlorination. It was first
produced in 1947 as an insecticide for agricultural crops and livestock, for lawns and gardens, and
also for underground treatment of building foundations. The U.S. EPA canceled above-ground
uses in 1978 because of concern about cancer risks and canceled all uses after 1988.
These compounds are insoluble in water, soluble in fat, and they bioconcentrate in fat tissue.
The most persistent compounds in biological tissue are oxychlordane (an epoxide) and heptachlor
epoxide, which are of primary toxicologic significance. It persists in rats for at least 56 days after
withdrawal of chlordane from the diet. There is some evidence that steady-state levels in human
fat are not reached before 3 to 12 mo of exposure. This information is discussed in Section 3.0.
People have not been harmed by working in factories where chlordane is manufactured, but
there are several reports of adverse effects in people having direct contact with chlordane, such as
farmers, pesticide applicators, and people excessively exposed in improperly treated buildings.
These effects include NHL in farmers using chlordane and other farm chemicals, and neurological
and immunological impairments, blood abnormalities, and other chronic health problems in people
living in treated residences.
In animal experiments chlordane has induced liver tumors in all five strains of mice that have
been tested and liver and thyroid toxicity in mice and rats. A recent study of cancer mechanisms
in mice concluded that chlordane-induced liver tumors arise from different cell types than
spontaneous tumors, and that some, but not all, of the induced tumors regress after chlordane
dosing is stopped.
Administration of chlordane to rodents during fetal development has resulted in early
mortality after birth, decreases in immunological and bone marrow functioning, and possibly
impairment of normal sex hormone function in young adult animals.
Because of the repeated findings of liver carcinogenicity in mice, chlordane is probably a
carcinogen in humans. Under the 1986 Carcinogen Risk Assessment Guidelines, chlordane is
classed as a probable human carcinogen (B2), and under the proposed 1996 guidelines it would be
characterized as a likely carcinogen.
An area of scientific uncertainty in this assessment concerns the roles of neurotoxic,
immunotoxic, and hematotoxic effects in the chronic toxicity of chlordane in humans. In acute
chlordane exposures, neurotoxicity has been established as a common endpoint with similar
symptomatologies in both humans and animals (Grutsch and Khasawinah, 1991). Chronic studies
with chlordane in animals have considered and found the liver to be the primary target organ.
However, occupational studies of workers with chronic chlordane exposure have established no
liver toxicity, although hematotoxicity has been strongly associated with the high levels of
occupational exposure (Fleming and Timmeny, 1993). Recent epidemiology results from a
42
-------
non-occupationally exposed cohort, however, do show strong evidence of neurotoxicity (Kilburn
and Thornton, 1995) and some indications of immunotoxicity (McConnachie and Zahalsky,
1992). These findings could mean that in rats the liver is more sensitive to chlordane effects than
is the nervous or immune system. Some data do exist for rats (transient tremors in rats observed
in NCI [1977]) indicating neurotoxicity occurring in chronic studies at levels higher than those
eliciting liver effects. This possibility cannot be confirmed, however, as no long-term,
repeated-dose animal study has been conducted in which neurotoxicity has been specifically
evaluated. These findings could also be a result of study bias as occupational studies were
designed to elucidate hepatic effects; both Alvarez and Hyman (1953) and Fishbein et al. (1964)
employed a number of liver function tests and only addressed neurological symptoms through
a general physical exam and did not address immunological effects at all. The possibility exists
that this assessment is based on studies flawed in design and endpoint assessment. Due to this
circumstance, any future alteration of the RfD (or RfC) based on an improved database (such as
a chronic study) would have to be considered in light of what is known about the neurotoxic and
immunotoxic potential of chlordane in both animals and humans.
Another area of scientific uncertainly in this assessment concerns the toxicological
significance of endocrine mimicry effects of chlordane. Toxicity data for this chemical includes a
study demonstrating biochemical and behavioral alterations consistent with technical chlordane
(or its metabolites) mimicking male sex-steroids (Cassidy et al., 1994). That these effects could
include reproductive behaviors is suggested in this study. Tissue distribution data also offer
evidence that developmental/reproductive toxicity could be a potential endpoint as technical
chlordane (or its metabolites) appears to have an increased affinity toward reproductive organs
during pregnancy. Evidence that these altered behaviors and circumstances would have functional
consequences that could alter reproduction is lacking, however, because no multigenerational
reproductive studies exist.
Because chlordane shares the qualities of persistence and bioaccumulation in soils and
tissues with other pesticides still in use, it is likely to be present with these other agents in breast
milk, soils, fish residues, and animal feeds, and is likely to contribute to the etiology of diseases or
conditions (such as NHL) that are associated with exposure to these sources. Therefore, when
the total exposure to chlorinated substances is being assessed, contributions from chlordane and
other pesticides no longer in use should not be neglected.
6.2 Dose-Response Assessment
Quantitative estimates of human risk as a result of low-level chronic chlordane exposure are
based on animal experiments since no adequate human exposure data are available in studies
where adverse effects have been observed in people. Liver toxicity in rats is the effect observed at
the lowest tested doses in both feeding and inhalation studies.
The human chronic dose of ingested chlordane considered to be without adverse effect (the
RfD) is 5E-4 mg/kg-day. This value was based on the occurrence of hepatic necrosis observed in
a lifetime mouse feeding study by first determining the NOAEL (0.15 mg/kg-day) and then
dividing this dose by a composite UF. An uncertainly factor of 300 was used: 10 for
extrapolation from rats to humans; 10 for variations between humans; and 3 because the entire
43
-------
data base is lacking a multigeneration reproductive study, which is considered necessary for a full
characterizati on of toxi city.
The UF for the RfD is 300 and the overall confidence is medium. The components of this
overall rating consist of the confidence in the principal study, rated medium, and confidence in the
entire data base, which is medium. The principal study is a mouse chronic oral study performed
with relatively large group sizes, where histopathological analyses on the known animal target
tissue, the liver, was thoroughly performed. The critical effect of hepatic toxi city was consistent
across routes of exposure as the inhalation RfC is also based on hepatotoxicity. However, recent
evidence indicates that neurotoxicity, a known human endpoint in acute exposures, may be a
relevant endpoint in chronic human exposures, and no chronic animals studies have examined
neurotoxicity. Studies on pre- and postnatal animals indicating chlordane mimicry of sex-steroids
raise reproductive concerns and no multigenerational reproductive studies, by any route, currently
exist. Therefore, there is some concern that the appropriate endpoints have not been adequately
examined in the existing data base. Despite this concern, the wide array of special studies (which
may raise similar controversy for other chemicals having less data than does chlordane) and
numerous chronic studies having consistent results are judged adequate information to rate the
overall confidence in the database as medium.
The human chronic air concentration (RfC) considered to be without adverse effect is
8E-5 mg/m3. This value was based on the occurrence of blood-indicators of hepatic dysfunction
derived from a 3-mo rat inhalation study by first designating a NOAEL (10 mg/m3) and then
dividing this concentration by a composite UF. An uncertainty factor of 3000 was used: 10 for
extrapolation from the 3-mo study to a lifetime study; 10 for extrapolation from rats to humans;
10 for variability between humans; and 3 for data base deficiencies, principally the lack of any
multigeneration reproductive study. It is assumed that these RfD and RfC values are low enough
to protect humans from all forms of toxi city.
The overall confidence in this RfC value is low. The components of this overall rating
consist of the confidence in the principal study, rated medium, and the confidence in the entire
data base, which is low. The principal study is a rat subchronic inhalation study performed with
relatively large group sizes, in which histopathological analyses on the known animal target tissue,
the liver, was thoroughly performed. The confidence in the database for technical chlordane is
rated low. In animal studies, the critical effect of hepatic toxicity was consistent across routes of
exposure as oral studies also indicate hepatotoxicity as the critical endpoint of toxicity
(Khasawinah and Grutsch, 1989a). The same limitations, uncertainties and controversies
surrounding the oral RfD assessment also apply to this assessment (above) and the possibility
exists that this assessment, too, is based on studies flawed in design and endpoint assessment.
Due to this circumstance, any future alteration of the RfC based on improved database (such as a
chronic study) would have to be considered in light of what is made known about the neurotoxic
and immunotoxic potential of chlordane in both animals and humans.
To estimate the human cancer risk to chlordane, the Agency has assumed that, in general,
the risk is proportional to the dose and the quantitative risk of cancer which might appear in
humans is the same as the risk of liver cancer in mice. The liver cancer risk in five different mouse
data sets (males and females in three mouse strains), when converted to an equivalent human
44
-------
dose, averages 0.35 per mg/(kg-day) and the total range is a factor of 8. The slope factor, which
is the lifetime risk from a lifetime average ingestion exposure of 1 |ig/(kg-day), is 3.5E-4.
If chlordane exposure came only from drinking water, the risk for a drinking water concentration
of 1 ng/L would be about 1E-5. A chlordane concentration of 1 |ig/m3 would induce a risk of
about 1E-4.
The relation between the quantitative cancer and noncancer risk estimates given above can
be seen if we were to calculate the cancer risk to people hypothetically exposed continuously via
ingestion to the RfD dose or via inhalation to the RfC concentration. These lifetime cancer risks
would be approximately 1.75E-4 for ingestion and 7E-5 for inhalation, which fall into the range of
E-6 to E-4.
7.0 REFERENCES
Adeshina, F. and E.L. Todd. 1991. Application of biological data in cancer risk estimations of
chlordane and heptachlor. Regul. Toxicol. Pharmacol. 14: 59-77'.
Ahmed, F.E., NJ. Lewis, and R.W. Hart. 1977. Pesticide induced ouabain resistant mutants in
Chinese hamster V79 cells. Chem. Biol. Interact. 19: 369-374.
Aldrich, F.D. and J.H. Holmes. 1969. Acute chlordane intoxication in a child. Case report with
toxicological data. Arch. Environ. Health 19(1): 129-132.
Al-Hachim, G.M. and A. Al-Baker. 1973. Effects of chlordane on conditioned avoidance
response, brain seizure threshold and open-field performance of prenatally-treated mice.
Br. J. Pharmacol. 49:480-483.
Alvarez, W.C. and S. Hyman. 1953. Absence of toxic manifestations in workers exposed to
chlordane. Arch. Ind. Hyg. Occup. Med. 8:480-483.
Arnold, S.F., D.M. Klotz, B.M. Collins, P.M. Vonier, L.G. Guillette, and J.A. Mclachlan. 1996.
Synergistic activation of estrogen receptor with combinations of environmental chemicals.
Science 272: 1,489-1,492.
ATSDR (Agency for Toxic Substances and Disease Registry). 1994. Toxicological Profile for
Chlordane. TP-93/03. U.S. Department of health and Human Services, Public Health Service.
Barrass, N., M. Stewart, S. Warburton, J. Aitchison, D. Jackson, P. Wadsworth, A. Marsden, and
T. Orton. 1993. Cell proliferation in the liver and thyroid of C57B1/10J mice after dietary
administration of chlordane. Environ. Health Perspect. 101(suppl. 5): 219-224.
Barnett, J.R. and H.W. Dorough. 1974. Metabolism of chlordane in rats. J. Agric. Food Chem.
22(4): 612-619.
45
-------
Barnett, J.B., D. Holcomb, J.H. Menna, and L.S.F. Soderberg. 1985a. The effect of prenatal
chlordane exposure on specific anti-influenza cell-mediated immunity. Toxicol. Lett.
25(3): 229-238.
Barnett, J.B., L.S.F. Soderberg, and J.H. Menna. 1985b. The effect of prenatal chlordane
exposure on the delayed hypersensitivity response of BALB/C mice. Toxicol. Lett.
25(2): 173-183.
Barnett, J.B., B.L. Blaylock, J. Gandy, J.H. Menna, R. Denton, and L.S. Soderberg. 1990a.
Long-term alteration of adult bone marrow colony formation by prenatal chlordane exposure.
Fundam. Appl. Toxicol. 14(4): 688-695.
Barnett, J.B., B.L. Blaylock, J. Gandy, J.H. Menna, R. Denton, and L.S. Soderberg. 1990b.
Alteration of fetal liver colony formation by prenatal chlordane exposure. Fundam. Appl.
Toxicol. 15(4): 820-822.
Bessi, H., C. Rast, B. Rether, G. Nguyen-Ba, and P. Vasseur. 1995. Synergistic effects of
chlordane and TPA in multistage morphological transformation of SHE cells. Carcinogenesis
16: 237-244.
Blair, A., D.J. Grauman, J.H. Lubin, and J.F. Fraumeni. 1983. Lung cancer and other causes of
death among licensed pesticide applicators. J. Natl. Cancer Inst. 71: 31-37.
Brimfield, A. A. and J.C. Street. 1981. Microsomal activation of chlordane isomers to derivatives
that irreversibly interact with cellular macromolecules. J. Toxicol. Environ. Health 7: 193-206.
Brown, D.P. 1992. Mortality of workers employed at organochlorine pesticide manufacturing
plants - an update. Scand. J. Work Environ. Health 18: 155-161.
Brown, L.M., A. Blair, R. Gibson, G.D. Everett, K.P. Cantor, M. Schuman, L.F. Burmeister,
S.F. Van Lier, and F. Dick. 1990. Pesticide exposures and other agricultural risk factors for
leukemia among men in Iowa and Minnesota. Cancer Res. 50: 6,585-6,591.
Brown, L.M., L.F. Burmeister, G.D. Everett, and A. Blair. 1993. Pesticide exposures and
multiple myeloma in Iowa men. Cancer Causes Control 4:153-156.
Caldwell, G.G., S.B. Cannon, C.B. Pratt, and R.D. Arthur. 1981. Serum pesticide levels in
patients with childhood colorectal carcinoma. Cancer 48: 774-778.
Calabrese, E.J. 1996. Gender differences to response to toxic substances. Unpublished report to
the U.S. Environmental Protection Agency, May.
Cantor, K.P., A. Blair, G. Everett, R. Gibson, L.F. Burmeister, L.M. Brown, L. Schuman, and
F.R. Dick. 1992. Pesticides and other agricultural risk factors for Non-Hodgkin's lymphoma
among men in Iowa and Minnesota. Cancer Res. 52: 2,447-2,455.
46
-------
Cassidy, R.A., C.V. Vorhees, DJ. Minnema, and L. Hastings. 1994. The effects of chlordane
exposure during pre- and postnatal periods at environmentally relevant levels on sex
steroid-mediated behaviors and functions in the rat. Toxicol. Appl. Pharmacol. 126(2): 326-337.
Chadduck, W.M., S.M. Gollin, B.A. Gray, IS. Norris, C.A. Araoz, and A.F. Tyrka. 1987.
Gliosarcoma with chromosome abnormalities in a neonate exposed to heptachlor. Neurosurgery
21(4): 557-559.
Chernoff, N. and RJ. Kavlock. 1982. An in vivo teratology screen utilizing pregnant mice.
J. Toxicol. Environ. Health 10: 541-550.
Chuang, L.F., Y. Liu, K. Killiam, and R.Y. Chuang. 1992. Modulation by insecticides heptachlor
and chlordane of the cell-mediated immune proliferative responses of rhesus monkeys. In Vivo
6: 29-32.
Cranmer, J.M., M.F. Cranmer, and P.T. Goad. 1984. Prenatal chlordane exposure: effects of
plasma corticosterone concentrations over the lifespan of mice. Environ. Res. 35: 204-210.
Dearth, M.A. and R.A. Kites. 1991. Complete analysis of technical chlordane using negative
ionization mass spectrometry. Environ. Sci. Technol. 25(2): 245-254.
Ditraglia, D., D.P. Brown, N. Namekata, and N. Iverson. 1981. Mortality study of workers
employed at organochlorine manufacturing plants. Scand. J. Work Environ. Health
7(suppl.): 140-146.
Dourson, M.L. and J.F. Stara. 1983. Regulatory history and experimental support of uncertainty
(safety) factors. Regul. Toxicol. Pharmacol. 3: 224-238.
Epstein, S.S. and Ozonoff, D. 1987. Leukemias and blood dyscrasias following exposure to
chlordane and heptachlor. Teratog. Carcinogen. Mutagen. 7(6): 527-540.
Falck, F., A. Ricci, M.S. Wolff, J. Godbold, and P. Deckers. 1992. Pesticides and
polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch.
Environ. Health 47: 143-146.
Fenske, R.A. 1992. Differences in exposure potential for adults and children following
residential insecticide applications. In: Similarities and Differences between Children and Adults:
Implications for Risk Assessment (Guzelian, P.S., CJ. Henry, and S.S. Olin, eds.). Washington,
DC: International Life Sciences Institute Press; pp. 214-225.
Fleming, L.E. and W. Timmeny. 1993. Aplastic anemia and pesticides: An etiologic association?
J. Occup. Med. 35:1,106-1,116.
Fishbein, W.I., J.V. White, and HJ. Isaacs. 1964. Survey of workers exposed to chlordane.
Ind. Med. Surg. 33: 726-727.
47
-------
Fox, T.R., A.M. Schumann, P.O. Watanabe, B.L. Yano, V.M. Maher, and JJ. McCormick.
1990. Mutational analysis of the H-ras oncogene in spontaneous C57BL6 x C3H/He mouse liver
tumors and tumors induced with genotoxic and nongenotoxic hepatocarcinogens. Cancer Res.
50: 4,014-4,018.
Garrettson, L.K., P.S. Guzelian, and R.V. Blanke. 1985. Subacute chlordane poisoning.
J. Toxicol. Clin. Toxicol. 22(6): 565-571.
Grutsch, J.F. and A. Khasawinah. 1991. Signs and mechanisms of chlordane intoxication.
Biomed. Environ. Sci. 4(3): 317-326.
Hirai, Y. and K. Tomokuni. 1993. Relationship between termiticide treatment and human
pollution by chlordane, oxychlordane, and nonachlor. Bull. Environ. Contam. Toxicol.
51: 814-819.
Hassoun, E., M. Bagchi, D. Bagchi, and SJ. Stohs. 1993. Comparative studies on lipid
peroxidation and DNA-single strand breaks induced by lindane, DDT, chlordane and endrin in
rats. Comp. Biochem. Physiol. 104C: 427-431.
Harris, R.Z., L.Z. Benet, and J.B. Schwartz. 1995. Gender differences in pharmacokinetics and
pharmacodynamics. Drugs 50(2): 222-239.
IARC (International Agency for Research on Cancer). 1991. Monographs, Volume 53.
pp. 115-175.
ICF-Clement. 1987. Pathology peer review of chlordane in F344 rats. Pathology review
participants: Goodman, D.G., A.W. Mackin, R.R. Maronpot, J.A. Popp, R.A. Squire, J.M.
Ward, and M.R. Anver.
ICF Kaiser, Inc. 1990a. THRESH: A computer program to compute a reference dose from
discontinuous animal toxicity data using the benchmark dose method. K.S. Crump Division,
Ruston, LA.
ICF Kaiser, Inc. 1990b. THRESHW: A computer program to compute a reference dose from
discontinuous animal toxicity data using the benchmark dose method. K.S. Crump Division,
Ruston, LA.
Infante, P.F. and C. Freeman. 1987. Cancer mortality among workers exposed to chlordane.
J. Occup. Med. 29(11): 908-909.
Infante, P., S.S. Epstein, and W.A. Newton, Jr. 1978. Blood dyscrasias and childhood tumors
and exposure to chlordane and heptachlor. Scand. J. Work Environ. Health 4: 137-150.
IRDC (International Research and Development Corporation). 1973. Eighteen-month oral
carcinogenic study of chlordane in mice. Unpublished report to Velsicol Chemical Corporation.
MRID No. 00067568. Available from U.S. Environmental Protection Agency.
48
-------
Jackson, M.A., H.F. Stack, and M.D. Waters. 1993. The genetic toxicology of putative
nongenotoxic carcinogens. Mutat. Res. 296: 241-277.
Johnson, K.W., N.E. Kaminski, and A. Munson. 1987. Direct suppression of cultured spleen cell
responses by chlordane and the basis for differential effects on in vitro immunocompetence. J.
Toxicol. Environ. Health 22: 497-515.
Kaneko, J.J. 1989. Serum proteins and the dysproteinmias. In: Clinical Biochemistry of
Domestic Animals, 4th Ed. (Kaneko, J.J., ed.). New York, NY: Academic Press; pp. 142-165.
Khasawinah, A. 1989. Chlordane residues in rat and monkey tissues following subchronic
inhalation exposure to technical chlordane. Bull. Environ. Contam. Toxicol. 43: 459-466.
Khasawinah, A.M. and J.F. Grutsch. 1989a. Chlordane: 24-month tumorigenicity and chronic
toxicity test in mice. Regul. Toxicol. Pharmacol. 10: 244-254.
Khasawinah, A.M. and J. Grutsch. 1989b. Chlordane: thirty-month tumorigenicity and chronic
toxicity test in rats. Regul. Toxicol. Pharmacol. 10(2): 95-109.
Khasawinah, A., C. Hardy, and G Clark. 1989a. Comparative inhalation toxicity of technical
chlordane in rats and monkeys. J. Toxicol. Environ. Health 28(3): 327-347. (The 90-day rat
study.)
Khasawinah, A., C. Hardy, and G. Clark. 1989b. Comparative inhalation toxicity of technical
chlordane in rats and monkeys. J. Toxicol. Environ. Health 28(3): 327-347. (The 28-day rat
study.)
Khasawinah, A., C. Hardy, and G. Clark. 1989c. Comparative inhalation toxicity of technical
chlordane in rats and monkeys. J. Toxicol. Environ. Health 28(3): 327-347. (The 90-day
monkey study.)
Kilburn, K.H. and J.C. Thornton. 1995. Protracted neurotoxicity from chlordane sprayed to kill
termites. Environ. Health Perspect. 103(7-8): 690-694.
MacMahon, B., R.R. Monson, H.H. Wang, and T. Zheng. 1988. A second follow-up of
mortality in a cohort of pesticide applicators. J. Occup. Med. 30: 429-432.
Malarkey, D.E., T.R. Devereux, G.E. Dinse, P.C. Mann, and R.R. Maronpot. 1995.
Hepatocarcinogenicity of chlordane in B6C3F1 and B6D2F1 male mice: evidence for regression
in B6C3F1 mice and carcinogenesis independent of ras proto-oncogene activation.
Carcinogenesis 16: 2,617-2,625.
McConnachie, P. R. and A.C. Zahalsky. 1992. Immune alterations in humans exposed to the
termiticide technical chlordane. Arch. Environ. Health 47(4): 295-301.
49
-------
McConnell, E.E. 1992. Ethylenethiourea: comparative response in rodent carcinogenesis studies
as a function of age at first exposure. In: Similarities and Differences between Children and
Adults: Implications for Risk Assessment (Guzelian, P.S., CJ. Henry, and S.S. Olin, eds.).
Washington, DC: International Life Sciences Institute Press; pp. 201-203.
McGregor, D.B., A. Brown, P. Cattanach, I. Edwards, D. McBride, C. Riach, and WJ. Caspary.
1988. Responses of the L5178Ytk+/tk- mouse lymphoma cell forward mutation assay, III.
72 coded chemicals. Environ. Mol. Mutagen. 12: 85-154.
McLachlan, J.A. 1997. Synergistic effect of environmental estrogens: report withdrawn.
Science 277: 462-463.
Menconi, S., J.M. Clark, P. Langenberg, and D. Hryhorczuk. 1988. A preliminary study of
potential human health effects in private residences following chlordane applications for termite
control. Arch. Environ. Health 43(5): 349-52.
Mitruka, B.M., and H.M. Rawnsley. 1981. Clinical Biochemical and Hematological Reference
Values in Normal Experimental Animals and Normal Humans. New York, NY: Masson
Publishing Co.
Moser, GJ. and R.C. Smart. 1989. Hepatic tumor-promoting chlorinated hydrocarbons
stimulate protein kinase C activity. Carcinogenesis 10:851-856.
Murphy, S.D. 1986. Toxic effects of pesticides. In: Casarett and Doull's Toxicology, 3rd Ed.
(Klassen, C.D., M.O. Amdur, and J. Doull, eds.). MacMillan Publishing Co.; pp. 519-581.
Mussalo-Rauhamaa, H. 1991. Partitioning and levels of neutral organochlorine compounds in
human serum, blood cells, and adipose and liver tissue. Sci. Total Environ. 103: 159-175.
Narotsky, M.G. and R.J. Kavlock. 1995. A multidisciplinary approach to toxicological
screening: II. Developmental toxicity. J. Toxicol. Environ. Health 45(2): 145-171.
NCI (National Cancer Institute). 1977. Bioassay of chlordane for possible carcinogenicity.
Technical Report Series No. 8. U.S. Department of Health, Education and Welfare; National
Institutes of Health. PB 271 977.
National Research Council. 1983. Risk Assessment in the Federal Government: Managing the
Process. National Academy Press.
Nomeir, A.A. and N.P. Hajjar. 1987. Metabolism of chlordane in mammals. Rev. Environ.
Contam. Toxicol. 100: 1-22.
Nye, D.E. and H.W. Dorough. 1976. Fate of insecticides administered endotracheally to rats.
Bull. Environ. Contam. Toxicol. 15: 291-296.
50
-------
Olanoff, L.S., WJ. Bristow, J. Colcolough, Jr., and J.R. Reigart. 1983. Acute chlordane
intoxication. J. Toxicol. Clin. Toxicol. 20(4): 291-306.
Pesatori, A.C., J.M. Sonntag, J.H. Lubin, D. Consonni, and A. Blair. 1994. Cohort mortality and
nested case-control study of lung cancer among structural pest control workers in Florida (United
States). Cancer Causes Control 5:310-318.
Ramamoorthy, K., F. Wang, I.C. Chen, S. Safe, J.D. Norris, D.P. McDonnell, K.W. Gaido,
W.P. Bocchinfuso, and K.S. Korach. 1997. Potency of Combined Estrogenic Pesticides.
Science 275: 405-406.
Rani, B.E., N.G. Karanth, and M.K. Krishnakumari. 1992. Accumulation and embryotoxicity of
the insecticide heptachlor in the albino rat. J. Environ. Biol. 13(2): 95-100.
Roberts, RJ. 1992. Overview of similarities and differences between children and adults:
implications for risk assessment. In: Similarities and Differences between Children and Adults:
Implications for Risk Assessment (Guzelian, P.S., CJ. Henry, and S.S. Olin, eds.). Washington,
DC: International Life Sciences Institute Press; pp. 11-15.
Ruch, R.J., R. Fransson, S. Flodstrom, L. Warngard, and I.E. Klaunig. 1990. Inhibition of
hepatocyte gap junctional intercellular communication by endosulfan, chlordane and heptachlor.
Carcinogenesis 11: 1,097-1,101.
Saito, L, N. Kawamura, K. Uno, N. Hisanaga, Y. Takeuchi, Y. Ono, M. Iwata, M. Gotoh,
H. Okutani, T. Matsumoto, Y. Fukaya, S. Yoshitomi, and Y. Ohno. 1986. Relationship between
chlordane and its metabolites in blood of pest control operators and spraying conditions.
Int. Arch. Occup. Environ. Health 58: 91-97.
Satoh, A. and H. Kikawa. 1992. Metabolic fate of cis- and trans-chlordane in mice. Nippon
Eiseigaku Zasshi 47(4): 818-825.
Schop, R.N., M.H. Hardy, and M.T. Goldberg. 1990. Comparison of the activity of topically
applied pesticides and the herbicide 2,4-D in two short-term in vivo assays of genotoxicity in the
mouse. Fundam. Appl. Toxicol. 15: 666-675.
Shindell, S. and S. Ulrich. 1986. Mortality of workers employed in the manufacture of
chlordane: an update. J. Occup. Med. 28:497-501.
Sipes, I.G. and AJ. Gandolfi. 1991. Biotransformati on of toxicants. In: Casarett and Doull's
Toxicology, 4th Ed. (Klassen, C.D., M.O. Amdur, and J. Doull, eds.). McGraw-Hill;
pp. 88-126.
Sobti, R.C., A. Krishan, and J. Davies. 1983. Cytokinetic and cytogenetic effect of agricultural
chemicals on human lymphoid cells in vitro. Arch. Toxicol. 52: 221-231.
51
-------
Solleveld, S.P., J.K. Haseman, and E.E. McConnell. 1984. Natural history of body weight gain,
survival and neoplasia in the F344 rat. J. Natl. Cancer Inst. 72: 929-940.
Spyker-Cranmer, J.M., J.B. Barnett, D.L. Avery, andM.F. Cranmer. 1982. Immunoteratology
of chlordane: Cell-mediated and humoral immune responses in adult mice exposed in utero.
Toxicol. Appl. Pharmacol. 62(3): 402-408.
Stromberg, P.C. and L.M. Vogtsberger. 1983. Pathology of the mononuclear cell leukemia of
Fischer rats. I. Morphologic studies. Vet. Pathol. 20: 698-708.
Taguchi, S. and T. Yakushiji. 1988. Influence of termite treatment in the home on the chlordane
concentration in human milk. Arch. Environ. Contam. Toxicol. 17: 65-72.
Takamiya, K. 1990. Interruption of chronic chlordane exposure and plasma residue levels in
occupational workers. Bull. Environ. Contam. Toxicol. 44(6): 905-909.
Tashiro, S. and F. Matsumura. 1978. Metabolism of trans-nonachlor and related chlordane
components in rat and man. Arch. Environ. Contam. Toxicol. 7: 113-127.
Telang, S., C. Tong, and G.M. Williams. 1982. Epigenetic membrane effects of a possible tumor
promoting type on cultured liver cells by the non-genotoxic organochlorine pesticides chlordane
and heptachlor. Carcinogenesis 3: 1,175-1,178.
Teufel, M., K.H. Niessen, J. Sartoris, W. Brands, H. Lochbiihler, K. Waag, P. Schweizer, and
G.V. Oelsnitz. 1990. Chlorinated hydrocarbons in fat tissue: Analyses of residues in healthy
children, tumor patients and malformed children. Arch. Environ. Contam. Toxicol. 19: 646-652.
U.S. EPA. 1979. Acceptable Common Names and Chemical Names for the Ingredient Statement
on Pesticide Labels. EPA 540/9-77-017. Washington, DC: Office of Pesticide Programs.
U.S. EPA. 1986a. Guidelines for Carcinogen Risk Assessment. F. R. 51:33,992-34,003.
U.S. EPA. 1986b. Carcinogenicity Assessment of Chlordane and Heptachlor/Heptachlor
Epoxide. EPA-600/6-87-004. Washington, DC: Office of Health and Environmental Assessment.
NTIS,
PB87-208757.
U.S. EPA. 1988a. Recommendations for and Documentation of Biological Values for use in
Risk Assessment. EPA 600/6-87/008. Cincinnati, OH: Office of Health and Environmental
Assessment. NTIS, PB88-179874/AS.
U.S. EPA. 1988b. Review of Pathology Working Group Slide Reevaluation of Livers of Rats in
a 30-Month Oral Exposure to Chlordane. Office of Pesticides and Toxic Substances.
Memorandum from Henry Spencer to George LaRocca, March 14.
52
-------
U.S. EPA. 1991. Guidelines for Developmental Toxicity Risk Assessment. F. R. (Decembers)
56(234): 63,798-63,826.
U.S. EPA. 1994a. Interim Policy for Particle Size and Limit Concentration Issues in Inhalation
Toxicity: Notice of Availability. F. R. (October 26) 59: 53,799.
U.S. EPA. 1994b. Methods for Derivation of Inhalation Reference Concentrations and
Application of Inhalation Dosimetry. EPA/600/8-90/066F, October.
U.S. EPA. 1994c. Peer Review and Peer Involvement at the U. S. Environmental Protection
Agency. Administrator's memorandum, June 7.
U.S. EPA. 1995a. Guidance on Risk Characterization. Memorandum from the Administrator,
Carol Browner, March 21.
U.S. EPA. 1995b. (proposed) Guidelines for Neurotoxicity Risk Assessment. F. R. (October 4)
60(192): 52,032-52,056.
U.S. EPA. 1995c. Use of the Benchmark Dose Approach in Health Risk Assessment.
EPA/630/R-94/007, February.
U.S. EPA. 1996a. (new proposed) Guidelines for Carcinogen Risk Assessment, 1996. (These
guidelines are currently only available as a draft.)
U.S. EPA. 1996b. Guidelines for Reproductive Toxicity Risk Assessment. F. R. (October 31)
61(212): 56,274-56,322.
Velsicol Chemical Corporation. 1983a. Thirty-month chronic toxicity and tumorigenicity test in
rats by chlordane technical. Unpublished study by Research Institute for Animal Science in
Biochemistry and Toxicology, Japan. MRTD No. 00138591, 00144313. Available from U.S.
Environmental Protection Agency.
Velsicol Chemical Corporation. 1983b. Twenty-four month chronic toxicity and tumorigenicity
test in mice by chlordane technical. Unpublished study by Research Institute for Animal Science
in Biochemistry and Toxicology, Japan. MRID No. 00144312, 00132566. Available from U.S.
Environmental Protection Agency.
Velsicol Chemical Corporation. 1984. Chlordane: A 90-day inhalation toxicity study in the rat
and monkey. VCL 28/83958/2. Chicago, IL.
Wang, H.H. and B. MacMahon. 1979a. Mortality of workers employed in the manufacture of
chlordane and heptachlor. J. Occup. Med. 21: 745-748.
Wang, H.H. and B. MacMahon. 1979b. Mortality of pesticide applicators. J. Occup. Med.
53
-------
21: 741-744.
Woods, J.S., L. Polissar, R.K. Severson, L.S. Heuser, and E.G. Kulander. 1987. Soft tissue
sarcoma and non-Hodgkin's lymphoma in relation to phenoxyherbicide and chlorinated phenol
exposure in western Washington. J. Natl. Cancer Inst. 78(5): 899-910.
Yonkers, K.A., J.C. Kando, D. Pharm, J.O. Cole, and S. Blumenthal. 1992. Gender differences
in pharmacokinetics and pharmacodynamics of psychotropic medication. Am. J. Psychiatry
149(5): 587-595.
Zahm, S.H., D.D. Weisenburger, P.A. Babbitt, R.C. Saal, K.P. Cantor, and A. Blair. 1988.
A case-control study of non-Hodgkin's lymphoma and agricultural factors in eastern Nebraska.
Am. J. Epidemiol. 128: 901.
Zahm, S.H., D.D. Weisenburger, P.A. Babbitt, R.C. Saal, J.B. Vaught, K.P. Cantor, and A. Blair.
1990. A case-control study of non-Hodgkin's lymphoma and the herbicide
2,4-dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiology 1(1): 349-356.
8.0 APPENDIXES
Appendix A: Benchmark Dose Analyses Cancer Risk Estimates.
(1) Computational Models—Discontinuous (Quanta!) Data
The polynomial mean response regression model (THRESH) (ICF Kaiser, 1990a) and the
Weibull power mean response regression model (THRESHW) (ICF Kaiser, 1990b) were used to
fit data by the maximum likelihood method.
THRESH P(d) = q0 + (1- q0) * (l-exp(-qi (d-d0) - ... -qk (d-d0)k)
THRESHW P(d) = 1 - e -a-P(d-d0^
where
d = dose
P(d) = probability of a response (health effect) at dose d
q0 , qx , q2 , a, P, y, k = estimated parameters
For data input to THRESH, the degree of the polynomial was set to the number of dose
groups minus one, the response type was extra [P(d) - P(0) /I- P(0)]. The risk estimated was
10%. For THRESHW, data inputs were the same and the lower limit of y was set at 1. For both
THRESH and THRESHW calculations were performed with and without the threshold parameter
(do).
(2) Data
54
-------
Incidence of hepatic necrosis in male mice from the study of Khasawinah and Grutsch
(1989a) was modeled.
(3) Model Fit
Model fit was judged by the p-values generated with the /2 goodness-of-fit generated by
THRESH or THRESHW.
(4) Results
Organ Data Modeled
Male mice, hepatic necrosis
BMC 10 (MLE) for
THRESH and THRESHW
Dose
(ppm)
0
1
5
12.5
2.4(3.3)
Observed Incidence
7/80
8/80
25/80
27/80
Predicted Incidence
THRESH and
THRESHW (P= 0.1)
7.5/80
9.7/80
18/80
31/80
Table 1. Benchmark concentration modeling at the 10% response level (BMC10) of hepatic
necrosis observed in male mice (Khasawinah and Grutsch, 1989a). The actual
data from the study (Observed Incidence) is compared against the results
obtained from applying both the THRESH and THRESHW models (Predicted
Incidence). For THRESH and THRESHW, P = 0.1. Both the maximum
likelihood estimate (MLE) and the BMC at 10% response level (BMC10) are
shown.
(5) Discussion
Both models (the polynomial THRESH and Weibull THRESHW) without a threshold
parameter were able to marginally fit the observed data (i.e., the p value for the goodness-of-fit
statistic was > 0.05 at 0.1). With a threshold term estimated, the model fits were p < 0.05,
indicating no fit. With a background term estimated for both models, the model outputs
underpredicted observed incidence at 5 ppm (around 0.3, 25/80) by about 8% and overpredicted
observed incidence at 12.5 ppm by about 5%. The 8% margin of underprediction around the
frequency of 0.3 is nearly equal to the designated response level of 10%. This degree of error in
the critical area where the 10% response level would be derived is deemed unacceptable. At the
response level chosen, 10%, the BMC determined by the models, although not reliable for the
reasons given above, was intermediate between the NOAEL of 1 ppm and the LOAEL of 5 ppm
at 2.4 ppm.
Due to this unacceptable degree of error in the modeled output, the NOAEL/LOAEL
method will be used rather than results from this benchmark analysis.
55
-------
Derivation of Cancer slope factors from bioassay data:
The following tables summarize the derivation of the slope factors for the five data sets.
Table 6a. Dose-response modeling of liver tumor data for mice chronically
exposed to chlordane in the diet.
Dietary
Concentration
(ppm)
Estimated Mouse Dose
(mg/kg-day)
Human Equivalent Dose1
(mg/kg-day)
Tumor Incidence
No. responding/No, examined
Source: IRDC (1973): Male CD-I mice, analytical chlordane in diet for 18 mo, hepatic carcinomas.
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0. 143882
5
25
5
0
0
0.71
3.57
7.14
0
0.1022
0.5137
1.0273
3/33
5/55
41/52
32/39
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2 -Q3xD3).
Results:
All Data: MLEs of dose coefficients: QO = 5.502E-2; Ql = 1.837; Q2 = 0.152; Q3 = 0.
Chi-square fit statistic = 11.843, p =.579E-3. 95% upper limit on low-dose slope = 4.542E-2 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.0571, 95% CLL = LED10 = 0.0439;
1E-4 extra risk = 5.443E-5, 95%CLL = 4.163E-5;
1E-5 extra risk = 5.443E-6, 95%CLL = 4.163E-6;
1E-6 extra risk = 5.443E-7, 95%CLL = 4.163E-7.
High-dose data dropped due to poor fit: P(D) = l-exp(-QO - QlxD - Q2xD2); MLEs of dose coefficients:
QO = 6.860E-2; Ql = 0; Q2 = 5.493.
Chi-square fit statistic = 0.738, p = 0.390. 95% upper limit on low-dose slope = 0.857 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.139, 95% CLL = LED10 = 0.088;
1E-4 extra risk = 4.267E-3, 95%CLL = 1.165E-4;
1E-5 extra risk = 1.349E-3, 95%CLL = 1.166E-5;
1E-6 extra risk = 4.267E-4, 95%CLL = 1.165E-6.
56
-------
Table 6b. Dose-response modeling of liver tumor data for mice chronically
exposed to chlordane in the diet.
Dietary
Concentration
(ppm)
Estimated Mouse Dose
(mg/kg-day)
Human Equivalent Dose1
(mg/kg-day)
Tumor Incidence
No. responding/No, examined
Source: IRDC (1973): Female CD-I mice, analytical chlordane in diet for 18 mo, hepatic carcinomas.
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0.143882.
0
5
25
50
0
0.71
3.57
7.14
0
0.1022
0.5137
1.027
0/45
0/61
32/50
26/37
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2 -Q3xD3).
Results:
All Data: MLEs of dose coefficients: QO = 0; Ql = 0.6814; Q2 = 0.9067; Q3 = 0.
Chi-square fit statistic = 15.41, p = 0.865E-4. 95% upper limit on low-dose slope = 1.515 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.132, 95% CLL = LED10 = 0.0696;
1E-4 extra risk = 1.467E-4, 95%CLL = 6.602E-5;
1E-5 extra risk = 1.468E-5, 95%CLL = 6.602E-6;
1E-6 extra risk = 1.468E-6, 95%CLL = 6.602E-7.
High-dose data dropped due to poor fit: P(D) = l-exp(-QO -QlxD - Q2xD2); MLEs of dose coefficients:
QO = 0;Q1=0;Q2 = 3.573.
Chi-square fit statistic = 2.502, p = 0.114. 95% upper limit on low-dose slope = 0.270 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.172, 95% CLL = LED 10 = 0.143;
1E-4 extra risk = 5.290E-3, 95%CLL = 3.677E-4;
1E-5 extra risk = 1.673E-3, 95%CLL = 3.690E-5;
1E-6 extra risk = 5.290E-4, 95%CLL = 3.691E-6.
57
-------
Table 6c. Dose-response modeling of liver tumor data for mice chronically
exposed to chlordane in the diet.
Dietary
Concentration
(ppm)
Estimated Mouse Dose
(mg/kg-day)
Human Equivalent Dose1
(mg/kg-day)
Tumor Incidence
No. responding/No, examined
Source: NCI (1977): Male B6C3F1 mice, analytical chlordane in diet for 80 weeks (+ 10 weeks
observation), hepatic carcinomas.
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0.143882.
0 (pooled control)
0 (matched control)
29.9
56.2
0
0
4.3
8.0
0
0
0.6187
1.1511
17/92
2/18
16/48
43/49
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2).
Results (using pooled control data):
All Data: MLEs of dose coefficients: QO = 0.183; Ql = 0; Q2 = 1.131.
Chi-square fit statistic = 4.61, p = 3.180E-2. 95% upper limit on low-dose slope = 0.324 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.305, 95% CLL = LED10 = 0.211;
1E-4 extra risk = 9.402E-2, 95%CLL = 3.083E-4
1E-5 extra risk = 2.973E-3, 95%CLL = 3.085E-5;
1E-6 extra risk = 9.402E-4, 95%CLL = 3.086E-6.
Source: NCI (1977): Female B6C3F1 mice, analytical chlordane in diet for 80 weeks (+ 10 weeks
observation), hepatic carcinomas.
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0.143882.
0 (pooled control)
0 (matched control)
29.9
56.2
0
0
4.3
8.1
0
0
0.6187
1.3093
3/78
0/19
3/47
34/49
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2).
Results (using pooled control data):
All Data: MLEs of dose coefficients: QO = 2.95E-5; Ql = 0; Q2 = 0.5242.
Chi-square fit statistic = 7.663, p = 5.638E-3. 95% upper limit on low-dose slope = 0.125 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.448, 95% CLL = LED10 = 0.365;
1E-4 extra risk = 1.381E-2, 95%CLL = 8.026E-4;
1E-5 extra risk = 4.368E-3, 95%CLL = 8.045E-5;
1E-6 extra risk = 1.381E-3, 95%CLL = 8.047E-6.
58
-------
Table 6d. Dose-response modeling of liver tumor data for mice chronically exposed
to chlordane in the diet.
Dietary
Concentration
(ppm)
Estimated Mouse Dose
(mg/kg-day)
Human Equivalent Dose1
(mg/kg-day)
Tumor Incidence
No. responding/No, examined
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0.143882.
0
1
5
12.5
0
0.15
0.75
1.875
0
0.0216
0.1079
0.2698
3/71
3/71
7/72
9/72
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2).
Results:
All Data: MLEs of dose coefficients: QO = 4.28E-2; Ql = 0.378; Q2 = 0.00
Chi-square fit statistic = 0.43, p = 0.73. 95% upper limit on low-dose slope = 0.710 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 2.79, 95% CLL = LED 10 = 0..15;
1E-4 extra risk = 2.64E-4, 95%CLL = 1.41E-4;
1E-5 extra risk = 2.64E-5, 95%CLL = 1.41E-5;
1E-6 extra risk = 2.64E-6, 95%CLL = 1.41E-6.
Source: Khasawinah and Grutsch (1989a); U.S. EPA (1986b): Male ICR mice, technical chlordane in diet for
24 mo, hepatocellular adenocarcinomas alone.
59
-------
Table 6e. Dose-response modeling of liver tumor data for mice chronically exposed
to chlordane in the diet.
Dietary
Concentration
(ppm)
Estimated Mouse Dose
(mg/kg-day)
Human Equivalent Dose1
(mg/kg-day)
Tumor Incidence
No. responding/No, examined
'Human Equivalent Dose = Mouse Dose x (mouse BW/human BW)1/4 = mouse dose x 0.143882.
0
1
5
12.5
0
0.15
0.75
1.875
0
0.0216
0.1079
0.2698
16/71
16/71
22/72
37/72
Model: GLOBAL86 computer algorithm, linearized multistage procedure, extra risk;
Polynomial model form: P(D) = l-exp(-QO -QlxD -Q2xD2).
Results:
All Data: MLEs of dose coefficients: QO = 2.50E-1; Ql = 5.53E-1; Q2 = 4.45
Chi-square fit statistic = 2.67E-2, p = 0.945, 95% upper limit on low-dose slope = 2.34 per mg/kg-day.
MLE of dose (mg/kg-day) at:
0.1 extra risk = 0.104, 95% CLL = LED10 = 4.50E-2;
1E-4 extra risk = 1.81E-4, 95%CLL = 4.27E-5;
1E-5 extra risk = 1.81E-5, 95%CLL = 4.27E-6;
1E-6 extra risk = 1.81E-6, 95%CLL = 4.27E-7.
Source: Khasawinah and Grutsch, 1989a; U.S. EPA (1986b): Male ICR mice, technical chlordane in diet for
24 mo, hepatocellular adenocarcinomas alone.
60
-------
The final quantitative estimate is based on the geometric mean of the five data sets, all
hepatocellular carcinomas in mice. The slope factors (units of risk per milligrams per kilograms
per day) for the individual data sets are: CD-I males, 0.858; CD-I females, 0.217; B6C3F1
males, 0.345; B6C3F1 females, 0.114; and ICR males, 0.710. The geometric mean is 0.349 per
mg/kg-day. In the calculation of human equivalent dose from dietary concentrations, data in the
publications were used (where given) to arrive at the animal dose in milligrams per kilograms per
day units and the human equivalent dose was calculated from the animal dose using the equation:
human dose = animal dose x (w/70)El/4, where w is the adult body weight of the mice. The
animal weights were assumed to be 30 grams in the absence of data in the publications; however,
the NCI study stated that males weighed 35 grams and females weighed 30 grams, and the actual
weights were used rather than the 30 gram default value. In the NCI study, the animal dose was
assumed to be 0.13 x the concentration in the feed as the default assumption in the absence of
animal dose values in the report. Also in the NCI study, the pooled control tumor incidence data
were used rather than the matched control (not given above) data. None of the studies mentioned
above except Khasawinah and Grutsch reported hepatocellular adenomas, and the above data are
for carcinomas alone. The slope factor for combined adenomas and carcinomas from the latter
study is 2.34 per mg/kg-day, which is about three times the slope factor for the carcinomas alone
in that study. Analysis (not shown here) of the Barrass et al. (1993) study, which had data at a
single dose for only combined adenomas and carcinomas, yielded a slope value of 0.47 per mg/kg-
day, which is comparable to the above data for carcinomas alone. Taken together, the data for
combined benign and malignant tumors is not considered extensive or reliable enough for risk
estimation. Neither the male nor the female IRDC data could be fit to the multistage model, and
the procedure of deleting the high dose group was used to get an adequate fit. In the case of the
females, the best fit was obtained by a polynomial with all terms zero except Q(6), the sixth
degree in powers of dose. Since the low-dose slope determination for the IRDC data was judged
to be difficult, the effect of deleting the IRDC data from the geometric mean was evaluated. The
resulting geometric mean was 0.303 per mg/kg-day, which was close enough to the geometric
mean of all five data sets (0.35 per mg/kg-day). Therefore, this difficult data affected the result to
only an insignificant extent. The earlier IRIS summary of the IRDC data sets used a version of
the multistage model where the degree of the polynomial was fixed at 3; since that model did not
adequately fit the data, it was not used here.
The 1996 proposed cancer guidelines recommend an alternate method of low-dose
extrapolation in which a linear extrapolation to small doses is done based on the lower 95%
confidence limit of the dose giving a 10% incidence of tumors. Using this method, the five data
sets would have the following slope values: CD-I males, 1.136; CD-I females, 0.699;
B6C3F1 males, 0.502; B6C3F1 females, 0.218; and ICR males, 0.676. The geometric mean of
these values is 0.567, compared with 0.349, the mean of the slopes from the linearized multistage
model. The higher slope values at the 10% incidence point of the dose-response curve compared
with the low-dose slope estimates is a reflection of the positive curvature (concave upwards) of
all the animal dose-response curves except the ICR mice data set.
61
-------
Appendix B: Summary of and Response to External Reviews and Comments
The Toxicological Review for Chlordane (except for Section 6, which was written after the
document was sent to the external peer reviewers) and the summaries for RfD, RfC, and cancer
have undergone both internal peer review performed by scientists within EPA and a more formal
external peer review performed by scientists chosen by EPA in accordance with the Agency's
Policy for Peer Review and Peer Involvement (U.S. EPA, 1994c). Comments made by the
internal reviewers were addressed prior to submitting the documents for external peer review and
are not part of this appendix. Public comments also were read and considered. Three external
peer reviewers provided written answers to general questions on the overall assessment and on
chemical-specific questions in areas of scientific controversy or uncertainty. A fourth peer
reviewer from the National Cancer Institute was asked to give his comments, without a specified
format, on just the cancer summary document. A summary of comments made by the external
reviewers and EPA's response to these comments follows. All three external peer reviewers
(see Contributors and Reviewers) recommended that this document and the accompanying
assessments were acceptable with minor revisions. The brief comments of the fourth reviewer
dealt with details of the cancer epidemiology studies and his comments were incorporated into the
document.
(1) General Questions
The three external reviewers offered editorial comments and many valuable, but minor,
suggestions, all of which have been incorporated into the text to the extent feasible. Specific
questions to the reviewers, comments by the peer reviewers and responses by EPA are addressed
below.
Question 1: Are you aware of relevant data/studies that were not discussed in the document?
Comments: A total of 13 additional references were mentioned by two of the four
reviewers.
Response to Comments: The readily available and appropriate references were obtained
and evaluated for relevance to the file. The additional studies cited by one of the reviewers on
prenatal chlordane exposure represent additional information on myeloid cell development and are
therefore incorporated into the file. None of the additional articles materially changed the
conclusions of the document.
Question 2: For the RfD and RfC, were the appropriate critical effects chosen? For the cancer
assessment were the tumors observed biologically significant?
Comments: One reviewer states that the critical effect of the RfD should be other than
hepatocellular hypertrophy from Khasawinah and Grutsch (1989b). This reviewer gives the
specific recommendation to substitute the liver effects observed in the chronic mouse study
(Khasawinah and Grutsch, 1989a). Reasons cited for making this comment include the
questionable adverse nature of hepatocellular hypertrophy and the association of this effect with
age-dependent leukemia, a major confounder in this study.
62
-------
One reviewer wanted to know the significance of the finding that H-ras and K-ras
oncogenes were absent in tumors induced by chlordane.
Response to Comments: We agree with the reviewer that the hepatocellular hypertrophy
reported in rats in Khasawinah and Grutsch (1989b) is probably of no toxicological significance.
The hypertrophy evidenced no pathological progression in this study even at the 130 week time
point and was correlated with increased levels of metabolizing enzymes (cytochromes P-450);
it is, therefore, judged as adaptive. The recommendation of the reviewer to substitute the mouse
liver study and effects was instituted.
As a consequence of this action, the critical effect of the RfC (also hepatocellular
hypertrophy) was considered to be invalidated. The toxicology of the inhalation data base was
reconsidered by viewing concurrently the hepatic effects that occurred in the 28-day pilot
inhalation study along with those few effects observed in the subchronic study. This approach
allowed an overall progression of hepatic effects to be elucidated, from necrosis to blood
chemistry indicators of hepatic dysfunction. As these latter effects clearly relate to organ
dysfunction, as they were in the progression of adverse hepatic alterations that took place at
higher concentrations, and as they were seen in both the 28-day pilot study and in the subchronic
study (at the highest concentration only), these indicators were collectively called out as the new
critical effect.
A statement was inserted that promoters and nongenotoxic carcinogens are not believed to
directly induce ras genes.
Question 3: Were the most appropriate studies used for the various assessments?
Comments: One reviewer reiterated the comment about regional hepatocellular
hypertrophy made in item 2 above. The other two reviewers thought that better information
might be available: one would like epidemiology studies looking for infectious diseases and
cancer in people exposed in utero or as children, the other would like neurotoxicology studies
to be done.
Response to Comments: We agree that the data base is not adequate in these respects and
we have already flagged neurotoxicity as a data gap in the Risk Characterization section of the
Toxicological Review and in the Confidence sections of the RfD and RfC. For the RfD, the
chronic mouse study of Khasawinah and Grutsch (1989a) is now being used. For the RfC, the
28-day pilot study (Khasawinah et al., 1989b) is now being used in conjunction with the
subchronic study (Khasawinah et al., 1989a) to provide some form of a concentration-response
continuum.
Question 4: Should other studies be added under the "supporting/additional" category?
Comments: One reviewer thought that information about the dose of chlordane used in the
genotoxic assays and about the response of these assays to structural analogs of chlordane would
63
-------
improve the description of genotoxicity. The two other reviewers had no suggestions for the
improvement of the supporting studies section.
Response to Comments: In this document no attempt was made to reanalyze the
genotoxicity information. Instead, the results of the genotoxicity review done in the EPA
Carcinogenicity Assessment of Chlordane document (U.S. EPA, 1986b) were incorporated
unchanged. No additional genotoxicity studies appeared in the literature search and none were
cited by the reviewers. Since there is little evidence for genotoxicity and whatever evidence there
is based on assays no longer used, we felt that trying to refine the picture by citing structure-
activity relationships and doses relative to standard mutagen would be counterproductive.
Question 5: For noncancer assessments are there other data that should be considered in
developing UFs?
Comments: One reviewer cited two additional papers as possibly being relevant to the
UF for the RfC.
A second reviewer brought up the controversial issue of synergism between weakly
estrogenic pesticides by citing a 1997 letter to Science in which the work of Arnold, et al. (1996)
could not be repeated.
Response to Comments: One of the papers, Nye and Borough (1976), was actually cited in
the document but it deals with the distribution of chlordane in the tissues of rats after intratracheal
administration, and therefore has little bearing on the uncertainty factor for the RfC. A comment
was inserted in section 4.7.2 of the Toxicological Review document discussing the controversy.
This paper was formerly withdrawn by the original authors in a letter to Science (McLachlan,
1997).
Question 6: Are the confidence and weight-of-evidence paragraphs clear and accurate?
Comments: One reviewer thought they were clear and comprehensive.
A second reviewer wanted the document to mention that reversal of neurotoxicity in
humans after withdrawing chlordane exposure had not been studied.
A third reviewer suggested that the confidence statements for both the RfD an RfC should
be recast to present a more balanced view on comments made on the significance of possible
effects that may occur in humans such as neurotoxic, hematologic, and reproductive effects. This
reviewer suggested mentioning that occupational studies revealed no evidence of hematotoxicity
despite the fact that hematologic examinations are a commonly performed health surveillance
procedure; that despite nearly 50 years of potential exposure to chlordane there are no indications
of reproductive effects in humans and that in the existing animal studies there were no
histopathological changes in either the nervous system, blood-forming organs, or reproductive
systems that could be attributed to chlordane and suggested adding to the RfC confidence
statement information about the relevance of the monkey study. This same reviewer made
specific mention to not claim that there is a common mechanism between liver cancer in mice and
64
-------
cancers of the hematopoietic system in humans. This reviewer also mentioned the possibility of
investigating and adding discussions on trends in incidence of liver cancer and NHL in humans
after the introduction of chlordane in 1947.
Response to Comments: Concerning the third reviewer's comments, a statement was not
added about reproductive effects as another reviewer pointed out that this is a significant data
gap. Also, the document already states that there is no evidence of adverse effects in people
engaged in manufacturing operations, but that there are case reports of adverse hematopoietic
effects in pest applicators and evidence of NHL in farmers. Therefore such a statement here
would be misleading. We agree that the link is not established by studies that can be quoted. The
document stated only that there was a possible link, and the wording was changed to indicate that
this link is speculative. Time trends were not added to the discussion because they are inherently
ambiguous when attempting to trace cause and effect. There are too many factors influencing
time trends to be able to conclude anything about chlordane as a cause.
(2) Comments on Chemical-Specific Questions
Question 1: Are case studies of hematopoietic disease in people exposed at home and exposed
as pesticide applicators supportive of the NHL observed in farmers in terms of confirming the
carcinogenicity of chlordane in humans and therefore do they furnish a rationale for upgrading the
classification of human evidence from inadequate to limited?
Comments: Two of the three reviewers said the case studies were not supportive of human
carcinogenicity and the third reviewer said they were, maintaining that the Agency should err on
the side of human safety.
Response to Comments: No change was made in the classification of human evidence of
carcinogenicity. It is still classified as inadequate because scientific judgment is a primary
consideration in EPA assessments and there is no new evidence submitted by the reviewers that
would change the rationale.
Question 2: Is there an underlying mechanism that would explain both the effects observed in
animals (liver toxicity and carcinogenicity in adult exposures and multiple effects after in utero
exposures) and effects in humans (toxicity in the nervous and immune systems)?
Comments: Two of the reviewers stated there was no common mechanism and the third
speculated that both liver and brain could be producing epoxies that could explain the effects.
Response to Comments: No change will be made in the discussion, since the EPA authors
did not find evidence of such mechanisms.
Question 3: Is the Agency correct in its judgment that neurotoxicity is a more appropriate
endpoint for human chronic exposure than liver toxicity?
Comments: One reviewer said that neurotoxicity may be more important, but there is no
dose-response information on which a firm conclusion can be based. A second reviewer said that
65
-------
other end points (besides the liver) are important, but that the end points are likely to be different
for adult exposures than for in utero exposures. A third reviewer said that both neurotoxicity and
liver toxicity are important considerations.
Response to Comments: No changes are necessary in view of these comments.
Additional comment 1: One reviewer concluded his remarks with a list of nine
recommendations, most of which were covered in the above questions. However, two additional
recommendations were: (1) include information, if available, on the activity of known compounds
in the various studies cited in the RfD document in the Additional Studies section as examples of
the ability of chlordane to affect developmental processes; and (2) expand comments on the UF
for calculating RfC to clarify how calculation of HEC addresses interspecies variation.
Response: Both of these requests will be accomplished to the extent feasible, although the
revised RfC does not calculate HEC for reasons elaborated upon in the RfC section.
Additional comment 2: One reviewer suggested a number of editorial clarifications in the
Toxicological Review document and the three summary documents.
Response: Most of these suggestions have been incorporated, but the few that were not
incorporated were rejected for one of three reasons: (1) they asked for details in experimental
protocols that, in our judgment were irrelevant to the outcome; (2) they wanted basic facts about
epidemiologic methods covered elsewhere by textbooks; and (3) they suggested alternative
wording that was equivalent to the existing text.
Additional comment 3: The reviewer from the National Cancer Institute agreed with the
classification of the human evidence of carcinogen!city as less than sufficient, and that the proper
classification is B2, probable human carcinogen. With respect to the Cantor et al. (1992) study
conclusions, he advised against saying that "[the multiple comparisons made could have] biased
the results upwards" because they actually could also increase the chance that relative risk
estimates too low will occur. He pointed out that in the Woods et al. (1987) study the odds ratio
for chlordane exposure was 1.61, not 1.46 as stated in the EPA report.
He also pointed out a paper by Brown et al. (1990) on leukemia in farmers exposed to
pesticides and one by Pesatori et al. (1994) about pest control operators in Florida.
Response: The wording of the Cantor et al. (1992) conclusions was changed to more
closely agree with that in the paper.
In the Woods et al. (1987) paper the 1.61 odds ratio was the overall odds ratio for men of
all occupations exposed to chlordane, whereas the 1.46 ratio quoted was for farmers only. The
former value was added to the text.
The Brown et al. (1990) and the Pesatori et al. (1994) papers were added to the text.
66
------- |