EPA/635/R-08/001F
www.epa.gov/iris
?/EPA
TOXICOLOGICAL REVIEW
OF
THALLIUM AND COMPOUNDS
(CAS No. 7440-28-0)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
September 2009
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
11
-------
CONTENTS—TOXICOLOGICAL REVIEW OF THALLIUM AND COMPOUNDS
(CAS No. 7440-28-0)
LIST OF TABLES v
LIST OF FIGURES vii
LIST OF ABBREVIATIONS AND ACRONYMS viii
FOREWORD ix
AUTHORS, CONTRIBUTORS, AND REVIEWERS x
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION 3
3. TOXICOKINETICS 5
3.1. ABSORPTION 5
3.2. DISTRIBUTION 5
3.3. METABOLISM 6
3.4. ELIMINATION 6
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 8
4. HAZARD IDENTIFICATION 9
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS 9
4.1.1. Incident/Case Reports 9
4.1.2. Population Surveys 16
4.1.3. Occupational Exposure 17
4.2. LESS-THAN-LIFETIME AND CHRONIC STUDIES AND CANCER
BIO AS SAYS IN ANIMALS—ORAL AND INHALATION 18
4.2.1. Oral Exposure 18
4.2.2. Inhalation Exposure 26
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES 26
4.3.1. Reproductive Toxicity 26
4.3.2. Developmental Toxicity 31
4.4. OTHER ENDPOINT-SPECIFIC STUDIES 32
4.4.1. Liver and Kidney Toxicity 33
4.4.2. Cardiotoxicity 40
4.4.3. Neurotoxicity 40
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
MODE OF ACTION 45
4.5.1. Interference with Potassium Transport 45
4.5.2. Disturbance of Mitochondrial Function and Energy Generation 46
4.5.3. Induction of Oxidative Stress 46
4.5.4. Reaction with Thiol Groups 47
4.5.5. Other Endpoint-specific Mechanistic Data 47
4.5.6. Genotoxicity 50
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS 51
4.6.1. Oral 51
4.6.2. Inhalation 54
4.6.3. Mode-of-Action Information 54
iii
-------
4.7. EVALUATION OF CARCINOGENICITY 55
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 56
4.8.1. Possible Childhood Susceptibility 56
4.8.2. Possible Gender Differences 56
5. DOSE-RESPONSE ASSESSMENTS 57
5.1. ORAL REFERENCE DOSE (RfD) 57
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and
Justification 57
5.1.2. Methods of Analysis 60
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs) 63
5.1.4. Candidate RfD Comparison Information 67
5.1.5. Previous RfD Assessment 75
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 75
5.3. CANCER ASSESSMENT 75
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD_AND DOSE
RESPONSE 76
6.1. HUMAN HAZARD POTENTIAL 76
6.2. DOSE RESPONSE 78
7. REFERENCES 79
APPENDIX A. Summary of External Peer Review and Public Comments and Disposition A-l
APPENDIX B. Documentation of Benchmark Dose Modeling B-l
IV
-------
LIST OF TABLES
Table 2-1. Chemical and physical properties of thallium and selected thallium
compounds 4
Table 3-1. Urine concentrations of thallium for the U.S. population from NHANES,
1999-2002 7
Table 4-1. Thallium toxicity in humans following oral exposure 10
Table 4-2. Selected clinical observations in Sprague-Dawley rats treated with thallium
sulfate for 90 days 20
Table 4-3. Incidence of alopecia in rats 21
Table 4-4. Selected blood chemistry values 22
Table 4-5. Thallium toxicity in animals following oral exposure 29
Table 4-6. Thallium toxicity in animals via injection 366
Table 5-1. Incidence data and BMD modeling results for selected clinical observations in
Sprague-Dawley rats treated with thallium sulfate for 90 days 61
Table B-l. A summary of BMDS (version 1.4.1) modeling results based on incidence of
rough coat in male Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-l
Table B-2. A summary of BMDS (version 1.4.1) modeling results based on incidence of
piloerection in male Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-3
Table B-3. A summary of BMDS (version 1.4.1) modeling results based on incidence of
shedding in male Sprague-Dawley rats exposed to thallium sulfate via gavage
for 90 days B-5
Table B-4. A summary of BMDS (version 1.4.1) modeling results based on incidence of
alopecia in male Sprague-Dawley rats exposed to thallium sulfate via gavage
for 90 days B-6
Table B-5. A summary of BMDS (version 1.4.1) modeling results based on incidence of
lacrimation in male Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 day B-6
Table B-6. A summary of BMDS (version 1.4.1) modeling results based on incidence of
exophthalmos in male Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-8
-------
Table B-7. A summary of BMDS (version 1.4.1) modeling results based on incidence of
miosis in male Sprague-Dawley rats exposed to thallium sulfate via gavage
for 90 days B-10
Table B-8. A summary of BMDS (version 1.4.1) modeling results based on incidence of
behavioral findings in male Sprague-Dawley rats exposed to thallium sulfate
via gavage for 90 days B-12
Table B-9. A summary of BMDS (version 1.4.1) modeling results based on incidence of
rough coat in female Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-13
Table B-10. A summary of BMDS (version 1.4.1) modeling results based on incidence
of piloerection in female Sprague-Dawley rats exposed to thallium sulfate
via gavage for 90 days B-17
Table B-l 1. A summary of BMDS (version 1.4.1) modeling results based on incidence
of shedding in female Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-22
Table B-12. A summary of BMDS (version 1.4.1) modeling results based on incidence
of alopecia in female Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-24
Table B-13. A summary of BMDS (version 1.4.1) modeling results based on incidence
of lacrimation in female Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-28
Table B-14. A summary of BMDS (version 1.4.1) modeling results based on incidence
of exophthalmos in female Sprague-Dawley rats exposed to thallium sulfate
via gavage for 90 days B-29
Table B-15. A summary of BMDS (version 1.4.1) modeling results based on incidence
of miosis in female Sprague-Dawley rats exposed to thallium sulfate via
gavage for 90 days B-31
Table B-16. A summary of BMDS (version 1.4.1) modeling results based on incidence
of behavioral findings in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days B-31
VI
-------
LIST OF FIGURES
Figure 5-1. POD (mg/kg-day) with corresponding derived candidate reference value that
would result if histopathologic changes of the skin (hair follicle atrophy)
were used as the critical effect 70
Figure 5-2. PODs (mg/kg-day) with corresponding derived candidate reference values
that would result if clinical observations from MRI (1988) were used as the
critical effect 71
Figure 5-3. POD (mg/kg-day) with corresponding derived candidate reference value that
would result if clinical chemistry changes (suggesting the liver or kidney as a
target) were used as the critical effect 72
Figure 5-4. PODs (mg/kg-day) with corresponding derived candidate reference values
that would result if reproductive toxicity endpoints were used as the critical
effect 73
Figure 5-5. PODs (mg/kg-day) with corresponding derived candidate reference values that
would result if alternative endpoints were used as the critical effect 74
vn
-------
LIST OF ABBREVIATIONS AND ACRONYMS
AchE acetyl cholinesterase
AIC Akaike Information Criterion
ALA aminolevulinic acid
ALT alanine aminotransferase
AST aspartate aminotransferase
ATSDR Agency for Toxic Substance and Disease Registry
BMD benchmark dose
BMDio benchmark dose corresponding to a 10% extra risk
BMDLio 95% lower bound on the benchmark dose corresponding to a 10% extra risk
BMDS benchmark dose software
BMR benchmark response
BUN blood urea nitrogen
CASRN Chemical Abstracts Service Registry Number
ChAT choline acetyltransferase
CHO Chinese hamster ovary
ECso effective concentration necessary to produce a 50% response
EPA Environmental Protection Agency
GI gastrointestinal
GLP good laboratory practice
GSH reduced glutathione
5-HT 5-hydroxytryptamine
i.p. intraperitoneal
IPCS International Programme on Chemical Safety
i.v. intravenous
IRIS Integrated Risk Information System
LD50 median lethal dose
LDH lactate dehydrogenase
LOAEL lowest-observed-adverse-effect level
MAO monoamine oxidase
MDA malondialdehyde
MEPP miniature endplate potential
MRI Midwest Research Institute
NA nucleus accumbens
NHANES National Health and Nutrition Examination Survey
NLM National Library of Medicine
NOAEL no-observed-adverse-effect level
NOEL no-observed-effect level
OSHA Occupational Safety and Health Administration
PAD peripheral arterial disease
POD point of departure
PBPK physiologically based pharmacokinetic
RfC inhalation reference concentration
RfD oral reference dose
s.c. subcutaneous
SCE sister chromatid exchange
Tl thallium
UF uncertainty factor
viii
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to thallium
and compounds. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of thallium and compounds.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration, and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of the data and related uncertainties. The discussion is intended to convey the
limitations of the assessment and to aid and guide the risk assessor in the ensuing steps of the
risk assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
IX
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER/AUTHOR
Octavia Conerly, M.S.P.H.
Office of Water
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
Susan Rieth, M.P.H.
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
CONTRACTOR SUPPORT
Robyn B. Blain, Ph.D.
ICF Consulting
Fairfax, VA
OFFICE OF RESEARCH AND DEVELOPMENT CO-LEAD
Susan Rieth, M.P.H.
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
REVIEWERS
This document has been reviewed by EPA scientists, interagency reviewers from other
federal agencies and White House offices, and the public, and peer reviewed by independent
scientists external to EPA. A summary and EPA's disposition of the comments received from
independent external peer reviewers and from the public is included in Appendix A.
INTERNAL EPA REVIEWERS
Joyce Donohue, Ph.D.
Office of Water
Elizabeth Doyle, Ph.D.
Office of Water
Steven Kueberuwa, Ph.D.
Office of Water
-------
Mike Hughes, Ph.D.
Office of Research and Development
Amal Mahfouz, Ph.D.
Office of Water
David Thomas, Ph.D.
Office of Research and Development
EXTERNAL PEER REVIEWERS
Ronald Baynes, DVM, Ph.D.
North Carolina State University
George Cherian, Ph.D.
University of Western Ontario
Lucio G. Costa, Ph.D. (chair)
University of Washington
George Daston, Ph.D.
The Procter & Gamble Company
Robert Hoffman, M.D.
New York University
Deborah Rice, Ph.D.
Maine Center for Disease Control
XI
-------
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of thallium
and compounds. IRIS Summaries may include oral reference dose (RfD) and inhalation
reference concentration (RfC) values for chronic and other exposure durations, and a
carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal of entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per |ig/m3 air breathed.
Development of these hazard identification and dose-response assessments for thallium
and compounds has followed the general guidelines for risk assessment as set forth by the
National Research Council (1983). EPA guidelines and Risk Assessment Forum Technical
Panel Reports that may have been used in the development of this assessment include the
following: Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a),
Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1986b), Recommendations for and
Documentation of Biological Values for Use in Risk Assessment (U.S. EPA, 1988), Guidelines
for Developmental Toxicity Risk Assessment (U.S. EPA, 199 la), Interim Policy for Particle Size
1
-------
and Limit Concentration Issues in Inhalation Toxicity (U.S. EPA, 1994a), Methods for
Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA, 1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA,
1995), Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk
Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA., 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A
Framework for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA,
2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through May 2009.
-------
2. CHEMICAL AND PHYSICAL INFORMATION
Metallic thallium (Tl) is bluish white or grey, very soft, malleable, and insoluble in water.
Thallium is a Group IIIA metal, one whose salts do not hydrolyze at pH > 7 to form insoluble
hydroxides. According to Mulkey and Oehme (1993), this is a physical property that contributes
to thallium's marked toxicity. Thallium exists in monovalent (thallous, thallium (I), Tl+1) and
trivalent (thallic, thallium (III), Tl+3) states. Monovalent thallium is favored in the standard
potential of Tl+3/Tl+1, coupling with a redox potential of+1.25V (Tl+3 + 2e~ becomes Tl+1).
According to Pearson (1963), monovalent thallium is a Lewis acid (electron pair receiver) that
prefers to interact with inorganic and organic sulfur, carbon, phosphorous, and arsenic moieties
as the electron pair donor (Lewis base). Monovalent thallium ions also are more stable in
aqueous solution, but trivalent thallium (Tl+3) can be stabilized by complexing agents (Sabbioni
et al., 1980a). Trivalent thallium forms more stable organic compounds than monovalent
thallium.
Monovalent thallium is similar to potassium (K+) in ionic radius and electrical charge,
which contribute to its toxic nature. Many of the thallium salts are soluble in water with the
exception of thallium (III) oxide, which is insoluble. Thallium compounds and their chemical
and physical properties are listed in Table 2-1.
Thallium occurs naturally in the earth's crust, with a crustal abundance of approximately
1 mg/kg. In soil, thallium concentrations are on the order of 0.1 to 1 mg/kg; higher
concentrations occur in the vicinity of metallic ore deposits. Measureable concentrations of
thallium are also found in marine water, freshwater, and air. Thallium is taken up by vegetation,
with the extent of uptake determined by soil acidity and plant species (Kazantzis, 2007).
According to the International Programme on Chemical Safety (IPCS) (1996), thallium is
used only in small amounts by industry, and thus worldwide production of pure thallium is low.
Sources for the production of thallium are zinc, lead, and sometimes copper or iron smelters and
sulfuric acid plants as well as a by-product of cadmium production. In 1981 the production of
thallium in the U.S. was discontinued. Thallium is released to the environment through the
combustion of fossil fuels (in particular from coal-fired power-generating plants), refinement of
oil fractions, smelting of ferrous and non-ferrous ores (including lead, copper, and zinc), and by
some other industrial processes such as cement production and brickworks (Kazantzis, 2007;
IPCS, 1996).
Due to its ability to remove hair, thallium (I) sulfate was used in the past as a depilatory
agent. Thallium (I) sulfate was once used in medicine to treat infections, such as venereal
diseases, ringworm of the scalp, typhus, tuberculosis, and malaria. It was also used in the past as
a pesticide for various rodents and insects but has been banned for this use in the U.S. since
1972. Currently, thallium is used in the semiconductor industry and the manufacture of optic
-------
lenses. When thallium is alloyed with mercury, it is used on switches and closures, which can
operate at subzero temperatures. Thallium compounds are also used to manufacture low-melting
glass, low-temperature thermometers, alloys, electronic devices, mercury lamps, fireworks, and
imitation gems. Thallium radioisotopes are used in medicine for scintigraphy of certain tissues
and the diagnosis of melanoma (Ibrahim et al., 2006; National Library of Medicine [NLM],
1998; IPCS, 1996; Agency for Toxic Substance and Disease Registry [ATSDR], 1992;
U.S. EPA, 1991b).
Table 2-1. Chemical and physical properties of thallium and selected
thallium compounds
Name
Metallic thallium
Thallium (I) acetate
Thallium (I) carbonate
Thallium (I) chloride
Thallium (I) nitrate
Thallium (I) oxide
Thallium (III) oxide
Thallium (I) selenite
Thallium (I) sulfate
CASRN
7440-28-0
563-68-8
6533-73-9
7791-12-0
10102-45-1
1314-12-1
1314-32-5
12039-52-0
7446-18-6
Chemical
formula
Tl
T1C2H3O2
T12CO3
T1C1
T1NO3
T12O
T1203
Tl2SeO3
T12SO4
Molecular
weight
204.38
263.43
468.78
239.84
266.39
424.77
456.76
535.72
504.82
Melting point
(°C)
303.5
131
273
430
206
596
717
No data
632
Boiling point
(°C)
1,457
No data
No data
720
430
No data
875
No data
Decomposes
Solubility in
water (g/L)
Insoluble
Very soluble
40.3 (15.5°C)
Very soluble
(20°C)
95.5 (20°C)
Soluble
(as T1OH)
Insoluble
No data
48.7 (20°C)
Sources: IPCS (1996); Downs (1993); ATSDR (1992).
-------
3. TOXICOKINETICS
3.1. ABSORPTION
Studies in humans and animals indicate that thallium compounds are readily absorbed
through various routes of exposure, but few studies provide quantitative measures of absorption.
Mulkey and Oehme (1993) reported that water soluble salts are rapidly and completely absorbed
from the respiratory tract, gastrointestinal (GI) tract, or skin but did not provide data or cite
references to support this conclusion. Thallium ions have been detected in the urine of exposed
humans (Ludolph et al., 1986; Davis et al., 1981; Schaller et al., 1980; Cavanagh et al., 1974;
Gefel et al., 1970) and animals (Thomas and McKeever, 1993; Waters et al., 1992; Leloux et al.,
1987; Talas and Wellhoner, 1983), which implies absorption from environmental sources.
Shaw (1933) determined that 61.6% of an oral dose of thallium (I) sulfate (25 mg/kg Tl)
was absorbed by a dog. Lie et al. (1960) determined that thallium was completely absorbed via
the GI tract, following oral administration of 767 ug/kg 204T1 as thallium (I) nitrate. This was
based on observations in male Wistar-derived rats where the body burden decreased
exponentially and extrapolated to 100% absorption. The same results were obtained when
thallium (as thallium nitrate) was administered by other routes of exposure (intravenous [i.v.],
38 ug/kg; intramuscular, 96 ug/kg; subcutaneous [s.c.], 96 ug/kg; intratracheal, 123 ug/kg; and
intraperitoneal [i.p.], 146 ug/kg). Eighty percent of a single dose of 10 nmol of thallium, as
thallium (I) sulfate, was absorbed within 1 hour from tied-off jejunal segments in anesthetized
rats (Forth and Rummel, 1975; Leopold et al., 1968).
No information was found regarding the absorption of thallium salts via inhalation.
There are a few case reports (Hirata et al., 1998; Ludolph et al., 1986) in which occupational
exposure has been associated with toxicity, but it could not be determined if exposure occurred
via inhalation or another route (e.g., oral or dermal).
The use of thallium salts in the past as depilatory agents, treatment for ringworm of the
scalp, and treatment for night sweats associated with tuberculosis suggests dermal absorption
(Leonard and Gerber, 1997; Reed et al., 1963; Lie et al., 1960).
3.2. DISTRIBUTION
Thallium ions are rapidly distributed (as early as 1 hour after exposure) throughout the
body in both experimental animals (Careaga-Olivares and Gonzalez-Ramirez, 1995; Galvan-
Arzate and Rios, 1994; Aoyama, 1989; Rios et al., 1989; Talas and Wellhoner, 1983; Sabbioni et
al., 1980a, b; Lameijer and van Zwieten, 1977; Andre et al., 1960; Downs et al., 1960; Lie et al.,
1960; Lund, 1956) and humans (Talas et al., 1983; Davis et al., 1981; Cavanagh et al., 1974;
Barclay et al., 1953), regardless of the route of exposure, dose, and length of exposure (Sabbioni
et al., 1980a, b; Lameijer and van Zwieten, 1977). The highest thallium concentrations have
-------
typically been found in the kidney and the lowest concentrations in the brain, with none being
detected in fat tissue. Thallium also has been demonstrated to cross the placenta in humans
(Hoffman, 2000) and experimental animals (Gibson and Becker, 1970).
The distribution of thallium in newborn Wistar rats differed from that in adult Wistar rats.
Newborns administered an i.p. dose of 16 mg/kg thallium (I) acetate (12.4 mg/kg Tl) had the
highest levels of thallium in the testis, heart, and kidneys, in that order, 24 hours after
administration (Galvan-Arzate and Rios, 1994). Levels in the liver and brain were
approximately three- to fourfold lower. In adult rats, the level of thallium in the kidney 24 hours
after an i.p. dose of 16 mg/kg thallium (I) sulfate was approximately twofold higher than the
level present in the testis (Rios et al., 1989). Galvan-Arzate and Rios (1994) also demonstrated
age-related differences in the regional distribution of thallium in the brain. Twenty-four hours
after i.p. injection of 16 mg/kg thallium (I) acetate, the thallium content among all regions of the
brain of newborn rats was homogeneous, whereas the thallium content in the brains of 5- to
20-day-old rats showed a region-dependent distribution, with thallium levels in the cortex
significantly lower than levels in the hypothalamus.
3.3. METABOLISM
Because thallium is an element, it is not metabolized. It is not known if thallium is
transformed from one valence state to another in vivo.
3.4. ELIMINATION
Thallium salts are eliminated mainly via urine and feces, but the amount excreted via
each route varies depending on the species. Thallium also has been found to be excreted in
breast milk, sweat, saliva, and tears. Thallium deposition into hair and nails also is considered an
important route of elimination (Kazantzis, 2007; IPCS, 1996).
A study of a human cancer patient orally administered thallium (I) sulfate and
radiolabeled thallium (I) nitrate (204T1NO3) demonstrated that thallium was mainly excreted in
the urine; 15.3% of the thallium salts were recovered in the urine over 5.5 days with 0.4%
recovered in the feces over 3 days (Barclay et al., 1953).
In a survey of 776 members of the general population (>40 years of age) that participated
in the 1999-2000 National Health and Nutrition Examination Survey (NHANES), the geometric
mean level of thallium in the urine was 0.16 ug/L, with a maximum of 0.86 ug/L (Navas-Acien
et al., 2005).
The Third National Report on Human Exposure to Environmental Chemicals (Centers for
Disease Control and Prevention [CDC], 2005) provides ongoing biomonitoring data for the U.S.
population for environmental chemicals over the periods 1999-2000 and 2001-2002 collected
from NHANES participants. Selected urinary thallium level data from this survey are provided
in Table 3-1. For the U.S. population (ages 6 and older), the geometric mean urinary thallium
-------
-th
concentration for survey years 2001-2002 was 0.165 ug/L, and the 95 percentile concentration
was 0.440 ug/L.
Table 3-1. Urine concentrations of thallium for the U.S. population from
NHANES, 1999-2002
Total, ages 6
and older
Survey
years
99-00
01-02
Sample
size
2413
2653
Urinary concentration of thallium in jig/L urine
[jig/g creatinine]3
Selected percentiles of the U.S. population
Geometric mean
0.176(0.162-0.192)
[0.166(0.159-0.173)]
0.165(0.154-0.177)
[0.156(0.151-0.162)]
50th
0.200(0.180-0.210)
[0.168(0.162-0.176)]
0.180(0.170-0.200)
[0.156(0.148-0.164)]
95th
0.450 (0.420-0.470)
[0.366 (0.338-0.387)]
0.440 (0.410-0.470)
[0.348 (0.337-0.365)]
Age group
6-11 years
12-19 years
20 years &
older
99-00
01-02
99-00
01-02
99-00
01-02
336
362
697
746
1380
1545
0.201 (0.167-0.243)
[0.221 (0.197-0.248)]
0.172(0.147-0.202)
[0.211(0.198-0.226)]
0.202(0.181-0.225)
[0.153 (0.146-0.160)]
0.200(0.182-0.220)
[0.143(0.137-0.150)]
0.170(0.157-0.183)
[0.162(0.153-0.171)]
0.159(0.147-0.173)
[0.153 (0.147-0.159)]
0.200(0.150-0.260)
[0.221 (0.196-0.236)]
0.200(0.160-0.220)
[0.207(0.198-0.221)]
0.210 (0.200-0.240)
[0.154(0.146-0.162)]
0.210(0.190-0.240)
[0.145(0.135-0.152)]
0.180(0.170-0.200)
[0.167(0.154-0.176)]
0.190(0.170-0.200)
[0.152(0.144-0.161)]
0.440 (0.350-0.590)
[0.424 (0.356-0.600)]
0.380 (0.360-0.420)
[0.411(0.389-0.456)]
0.460 (0.430-0.510)
[0.321 (0.265-0.364)]
0.460 (0.400-0.500)
[0.307 (0.299-0.333)]
0.450 (0.420-0.470)
[0.364 (0.325-0.389)]
0.440 (0.400-0.490)
[(0.342(0.313-0.362)]
Gender
Males
Females
99-00
01-02
99-00
01-02
1200
1313
1213
1340
0.197(0.179-0.217)
[0.154(0.147-0.161)]
0.184(0.173-0.196)
[0.146(0.140-0.153)]
0.159(0.145-0.175)
[0.178(0.167-0.189)]
0.149(0.137-0.163)
[0.167(0.158-0.176)]
0.220(0.190-0.240)
[0.156(0.149-0.164)]
0.200(0.190-0.220)
[0.148(0.141-0.156)]
0.180(0.150-0.200)
[(0.182(0.169-0.196)]
0.150(0.150-0.170)
[0.167(0.153-0.179)]
0.440 (0.420-0.480)
[0.338 (0.300-0.364)]
0.420 (0.390-0.460)
[0.307 (0.291-0.342)]
0.450 (0.410-0.490)
[0.380 (0.333-0.462)]
0.430 (0.400-0.500)
[(0.375 (0.348-0.402)]
a95 percentile confidence interval in parentheses.
Source: CDC (2005).
As noted above, thallium elimination is not limited to renal excretion. IPCS (1996)
estimated that in humans renal excretion accounts for approximately 70% of total daily excretion
of thallium. This estimate is based on limited human data.
In contrast to humans, thallium is excreted to a greater extent in the feces than in the
urine of rats and rabbits. IPCS (1996) estimated that in rats about 2/3 of the intake of thallium
-------
was excreted via the GI tract and about 1/3 via the kidney. Lund (1956) determined that, after
26 days, 51.4% of an i.p. dose of 10 mg/kg thallium (I) sulfate in the rat was eliminated via the
feces, while 26.4% was excreted in the urine. Talas and Wellhoner (1983) demonstrated that
thallium (I) acetate administered to rabbits via i.v. injection (as a radioactive tracer) was excreted
mainly in the feces. Both studies found that, although the feces was the major route of excretion
in the rat and rabbit, neither species had high levels in the bile, suggesting that excretion via the
liver was relatively low. Lund (1956) determined that thallium was mainly excreted in the feces
through gastric and intestinal secretions, which is likely associated with potassium excretion.
Lund (1956) demonstrated that rabbits excreted thallium through the kidneys by glomerular
filtration, but approximately one-half the dose filtered was reabsorbed in the tubuli.
In Syrian golden hamsters, thallium (I) sulfate was mainly excreted in the feces after i.p.
administration but was excreted at an equal rate in the feces and urine after an oral dose
(Aoyama, 1989). Shaw (1933) demonstrated that 32 and 61.6% of a single oral dose of
25 mg/kg Tl as thallium (I) sulfate administered to a dog was excreted in the urine at 3 and
36 days after dosing, respectively.
Sabbioni et al. (1980b) determined that thallium (I) sulfate administered at doses of
0.00004-2,000 ug/rat was persistent in the kidneys for 8 days (192 hours) after dosing with 2.5%
of the dose still present at that time (suggesting a half-life of approximately 1.5 days). Lehmann
and Favari (1985) and Lie et al. (1960) estimated the biological half-life of thallium in rats to
range from 3-8 days. The biological half-life in humans has been estimated to be approximately
10 days, with values up to 30 days reported (IPCS, 1996).
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
There are no physiologically based toxicokinetic models for thallium compounds.
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS
Studies of thallium toxicity in humans are comprised of clinical reports, case studies, and
medical surveys. Because case reports largely involved accidental ingestion, intentional
poisoning, or suicide attempts, they do not provide useful information on thallium toxicity
associated with chronic exposure. Available epidemiology studies involving long-term exposure
to thallium are limited by small study populations and insufficient characterization of long-term
exposure. Health effects information was based on self-reporting (via questionnaire) or medical
histories/physical examinations of uncertain scope.
In adults, the average lethal oral dose has been estimated to range from 10 to 15 mg/kg
(Gosselin et al., 1984; Schoer, 1984). Without treatment, death typically follows in about 10-
12 days, but death as soon as 8-10 hours also has been documented (IPCS, 1996).
4.1.1. Incident/Case Reports
As indicated by case reports, the acute toxicity of thallium is characterized by alopecia
(hair loss), severe pain in the extremities, lethargy, ataxia, abdominal pain or vomiting, back
pain, abnormal reflexes, neuropathy, muscle weakness, coma, convulsion, other neurological
symptoms (i.e., mental abnormalities, tremors, abnormal movements, abnormal vision, and
headache), and death (Lu et al., 2007; Tsai et al., 2006; Saha et al., 2004: Sharma et al., 2004;
Rusyniak et al., 2002; Atsmon et al., 2000; Hirata et al., 1998; Feldman and Levisohn, 1993;
Yokoyama et al., 1990; Heyl and Barlow, 1989; Roby et al., 1984; Limos et al., 1982; Davis et
al., 1981; Cavanagh et al., 1974; Gefel et al., 1970; Reed et al., 1963). Symptoms were
observable within 14 hours after a high dose (i.e., 5-10 g of thallium (I) nitrate), with death
occurring 8 days later (Davis et al., 1981). The lowest known single dose of thallium associated
with adverse effects was reported to be 0.31 g of thallium (I) acetate (3.4 mg/kg Tl, assuming a
70 kg body weight) (Cavanagh et al., 1974). This dose caused paresthesia, pain, weakness,
vomiting, and alopecia in a 26-year-old male. Approximately 1 month after the onset of
symptoms, complete recovery occurred following treatment. In adults, doses ranging from 6 to
40 mg/kg have been reported to be lethal (IPCS, 1996). Table 4-1 summarizes the individual
case reports.
-------
Table 4-1. Thallium toxicity in humans following oral exposure
Reference
Sex
Age
Dose
Symptoms"
Final outcome
Males — adult
Gefeletal. (1970)
Cavanagh et al. (1974)
Cavanaghetal. (1974)
Cavanagh et al. (1974)
Davis etal. (1981)
Limosetal. (1982)
Limos etal. (1982)
Roby etal. (1984)
Male
Male
Male
Male
Male
Male
Male
Male
41 years
60 years
56 years
26 years
19 years
56 years
26 years
45 years
Unknown but
chronic;
urine thallium
0.15mg/100mL
0.93 g thallium (I)
acetate in 2 divided
doses
0.93 g thallium (I)
acetate in 3 divided
doses
0.3 Ig thallium (I)
acetate
5-10 g thallium (I)
nitrate
Unknown
Unknown
Unknown;
urine thallium:
2,000 ug/L
High blood pressure; lower back pain; vomiting;
severe pain in the feet; weakness of the calf muscle;
alopecia; slurred speech; atrophic lower limbs;
limited vision
Diarrhea; vomiting; dizziness; back pain; paresthesia
of the feet and lower legs; high blood pressure; facial
weakness; dysphagia; difficulty breathing
Abdominal pain; diarrhea; vomiting; paresthesia;
photophobia, nystagmus, visual impairment; facial
weakness; bilateral ptosis
Paresthesia in both feet; chest pain; tenderness over
the sternum; vomiting, weakness, pain in the knees
and ankles that inhibited walking; alopecia
Nausea; vomiting; slurred speech; paresthesia of
hands and feet; respiratory weakness
Visual disturbances; alopecia; elevated AST and
ALT; high blood glucose and creatine kinase;
decreased myelinated fibers; denervated Schwann
cell clusters
Visual disturbances; alopecia; elevated AST and
ALT; high blood glucose and creatine kinase;
decreased myelinated fibers; denervated Schwann
cell clusters
Burning pain in feet; inability to walk; alopecia;
acute fibrillation
Death
Death within a week of
symptoms
Death within 3 weeks of
symptoms
Recovery
Death
Bedridden; could not
speak
Residual tremors of the
extremities and muscle
weakness of the lower
limbs
Continued neurological
dysfunction
10
-------
Table 4-1. Thallium toxicity in humans following oral exposure
Reference
Heyl and Barlow (1989)
Yokoyamaetal. (1990)
Hantsonetal. (1997)
Hirataetal. (1998)
Atsmon et al. (2000)
Sharma et al. (2004)
Sex
Male
Male
Male
Male
Male
Male
Age
"Five young
men"
3 1 years
48 years
29 years
40 years
48 years
Dose
Unknown
Unknown;
urine thallium:
3.5 mg/L
200 mg thallium (I)
sulfate
Unknown; hair
thallium: 20 ng/g
(32 months after
possible exposure)
Unknown;
urine thallium: 7 mg
Unknown;
serum thallium:
870 ug/100 mL
urine thallium:
5,000 ug/mL
Symptoms"
Follicular plugging of the skin (nose, cheeks, and
nasolabial folds) by keratinous material; crusted
eczematous lesions and acneiform eruptions on the
face; dry scaling on palms and soles; and alopecia
(scalp, eyelashes, lateral eyebrows, arms, and legs).
Skin biopsies (scalp and cheek): disintegrating hair
shafts, gross follicular plugging, and eosinophilic
keratohyalin granules in the epidermis; necrotic
sebaceous glands; pustular lesions on the face:
folliculitis and necrosis of the follicles; (feet) marked
hyperkeratosis and hypergranulosis.
Nausea, vomiting; leg pain; alopecia; abnormal
behavior; decreased conduction velocity of fast
nerve fibers
No overt symptoms within 24 hours; increase in
binucleated cells with micronuclei 15 days after
exposure
Alopecia; abdominal pain; diarrhea; tingling in
extremities; neuropathy
Weakness of the limbs; vomiting; severe
neurological symptoms; alopecia; high blood
pressure; increased ALT and AST; Mees lines;
decreased visual acuity; bilateral foot drop
Painful peripheral neuropathy, decreased
consciousness
Final outcome
4/5 recovered; 1/5
experienced permanent
neurological damage
Recovery
Recovery
Recovery
Recovery
Death
11
-------
Table 4-1. Thallium toxicity in humans following oral exposure
Reference
Sex
Age
Dose
Symptoms"
Final outcome
Females — adult
Robyetal. (1984)
Robyetal. (1984)
Robyetal. (1984)
Hoffman (2000)
Saha et al. (2004)
Female
Female
Female
Female
Female
5 1 years
61 years
80 years
Pregnant; ages
not specified
26 years
Unknown;
serum thallium:
50 ug/100 mL;
urine thallium:
5,000 ug/L
Unknown;
serum thallium:
740 ug/100 mL
Unknown;
serum thallium:
422 ug/100 mL;
urine thallium:
21,600 ug/L
150-1,350 mg
thallium (I) sulfate
Unknown;
serum thallium:
12 ug/100 mL
Numbness and weakness of the legs and hands;
alopecia; fluctuating pulse and blood pressure;
bradycardia; hypotension
Burning chest pain; paresthesia; difficulty speaking
and swallowing; inability to walk; hypotension;
acute respiratory distress syndrome (ARDS)
ARDS
Paresthesia; abdominal pain; muscle weakness;
lethargy; alopecia; Mees lines
Headache, lethargy, abdominal pain, muscle cramps,
joint pain, backache, numbness of fingers, alopecia,
erosion of nails
Persistent ventricular
ectopy and neurological
dysfunction,
necessitating placement
at a nursing home
Death
Death
None specified
Not specified
Both sexes — adult
Brockhaus et al. (1981)
Schoer (1984); Gosselin
etal. (1984)
Rusyniak et al. (2002)
Both
Both
Both
Not reported
Adult
Various
Unknown
10-15 mg/kg
thallium
Unknown; various
levels were detected
in urine
Sleep disorders; tiredness; weakness; nervousness;
headache; other psychic alterations; neurological and
muscular symptoms
None specified
Myalgia; arthralgia; paresthesia; dysesthesia; joint
stiffness; insomnia; alopecia; abdominal pain
Not reported
Death (average lethal
dose)
Recovery in 7 adults; 5
had ongoing psychiatric
problems
12
-------
Table 4-1. Thallium toxicity in humans following oral exposure
Reference
Tsai et al. (2006)
Lu et al. (2007); Kuo et
al. (2005)
Sex
Both
Both
Age
48-year old
female; 52-year
old male
48 and 52 years
Dose
1.5-2 .4 g
1.5 and 2. 3 g/person
(estimated);
Serum thallium:
950-2,056 ug/L
Urine thallium:
1 1,325-14,520 ug/L
Symptoms"
Confusion, disorientation, hallucination, anxiety,
depression, memory impairment, peripheral
neuropathy, erythematous skin rashes, diarrhea,
tachycardia, alopecia
Nausea, vomiting; general aching muscle pain;
numbness of tongue and mouth within a few hours;
severe paresthesia and dysesthesia in hands and feet
(one day postexposure); erythematous rash; diarrhea;
urine retention; hyporeflexia; muscle weakness;
hypoesthesia; acneiform eruptions; alopecia (1-3
weeks); Mees lines (2-3 months).
Skin biopsy: parakeratosis; dilated hair follicles
filled with keratin and necrotic sebaceous materials;
mild epidermal atrophy; vacuolar degeneration of the
basal layer.
Cutaneous nerve biopsy: axonal degeneration; loss
of epidermal nerves indicating involvement of the
small sensory nerves (2 months).
Final outcome
Impairment of memory
and verbal fluency
remained at 6 months;
neuropsychological
impairment persisted at
9 months
At 1 -year follow-up,
persistent paresthesia,
dysesthesia, and
impairment of small
sensory nerve fibers in
skin
Children
Reed etal. (1963)
Feldman and Levisohn
(1993)
Hoffman (2000)
Both
Male
Both
1-11 years
10 years
Transplacental
Unknown
Unknown;
serum thallium:
296 ug/L;
urine thallium:
322 ug/24 hours
Unknown
Alopecia; lethargy; ataxia; abdominal pain;
vomiting; abnormal reflexes; neuropathy; muscle
weakness; coma; convulsion
Alopecia; leg paresthesia; abdominal pain; seizures
Premature birth; low birth weight; alopecia
Neurological
abnormalities;
retardation; psychosis;
death
Recovery
None specified
13
-------
Table 4-1. Thallium toxicity in humans following oral exposure
Reference
Ammendola et al. (2007)
Sex
Male
Age
16 years
Dose
1.3 g thallium
sulfate;
urine thallium:
3,400 ug/L
Symptoms"
Acute stage: GI disturbances, alopecia, and clinical
and electrodiagnostic signs of severe
polyneuropathy.
Final outcome
3 years post-poisoning:
neurological
symptoms making
progress;
electrophysiological
signs of peripheral
neuropathy mainly
confined to lower limbs.
6 years post-poisoning:
persistent weakness and
sensory disturbances of
distal lower extremities;
neurological and
electrodiagnostic
abnormalities affecting
mainly the feet.
aALT = alanine aminotransferase; AST = aspartate aminotransferase.
14
-------
High blood pressure or fluctuating blood pressure was noted upon hospital admission in several
cases (Roby et al., 1984; Cavanagh et al., 1974; Gefel et al., 1970). Elevated serum aspartate
aminotransferase (AST) (formerly referred to as serum glutamic oxaloacetic transferase) and
serum alanine aminotransferase (ALT) (formerly referred to as serum glutamate pyruvate
transaminase), high blood glucose, and creatine kinase values also have been noted in case
reports of thallium exposure (Atsmon et al., 2000; Limos et al., 1982). The same symptoms
were noted across age and sex groupings. Retardation and psychosis were the most common
findings in children (1-11 years old) after nonlethal thallium exposure. Several cases were so
severe that institutionalization was necessary (Reed et al., 1963). Thallium significantly
decreased the conduction velocities of faster nerve fibers in a 31-year-old male, who ingested a
thallium-containing rodenticide, compared with baseline levels recorded following recovery.
In most case-study reports, thallium was detectable in the urine or tissues. In some cases,
thallium could not be definitively associated with the symptoms because other heavy metals
were also found in the blood or urine of the subject.
Hantson et al. (1997) evaluated cytogenetic changes in blood from a 48-year-old man
who accidentally ingested 200 mg of thallium (I) sulfate intended for rodenticide use. Despite
the lack of overt symptoms 24 hours after ingesting the thallium, the man was admitted to the
emergency room and Prussian blue treatments were commenced. Blood samples were obtained
on days 1 and 15 for cytogenetic analysis. Slight increases in mean sister chromatid exchange
(SCE) numbers on days 1 and 15 were not considered related to thallium exposure. A 3.5-fold
increase in binucleated cells with micronuclei (35% versus 10% in the historical controls) was
noted on day 15. The thallium level was determined to be 14.4 ug/dL in blood at the time of
hospital admission, and the concentration in urine was 3,804 ug/g Tl creatinine (reference value,
<1 ug/g Tl).
Fifty-one case histories of women treated for thallium poisoning following external
application of a 3 to 8% thallium (I) acetate ointment were reviewed for signs of possible
thallium intoxication (Munch, 1934). Neurological and GI symptoms were observed in 29 cases
after an unspecified number of applications with 2-24 ounces of the ointment. This was
approximately equivalent to a dose of 53-636 mg/kg Tl per application using a 5.5% ointment on
a 50 kg woman. Alopecia followed several weeks after beginning treatment.
Hoffman (2000) provided case reports and a comprehensive literature review of thallium
poisoning that occurred during pregnancy. Exposures were primarily oral, but some of the cases
involved dermal exposure. The majority of the doses were unreported, but those doses that were
documented ranged from 150-1,350 mg thallium (I) sulfate. Of the 18 cases that met Hoffman's
criteria (cases were excluded if maternal or fetal outcomes were not provided), 5 women were
exposed during the first trimester and 5 during the second trimester; the remaining 8 were
exposed during the third trimester. The ages of the women were not reported. While the
mothers developed the classic symptoms of thallium poisoning, including paresthesia, abdominal
15
-------
pain, muscle weakness, lethargy, alopecia, and Mees lines (single transverse white bands
occurring on the nails), the only consistent finding in their offspring was a trend toward
prematurity and low birth weight. Several of the children had alopecia, particularly those
exposed during the third trimester.
4.1.2. Population Surveys
Several published studies have surveyed populations living near a cement plant in
Lengerich, a small city in northwest Germany. These populations were studied because of their
potential to experience exposure to thallium as a result of its presence as an impurity in pyrite
and its release during the roasting of pyrite for use in making some types of cement. Thallium
was discharged to outdoor air, deposited in soils, and taken up by local crops and indigenous
plants. People who lived near the plant and consumed home-grown foods thus were exposed to
thallium through their diets. Prior to 1979, the concentration of thallium in the pyrite was
400 ppm. After 1979, a pyrite with lower levels of thallium (2 ppm) was used.
Brockhaus et al. (1981) conducted an epidemiologic study of a group of 1,200 people
living near the cement plant in Lengerich. Urinary thallium data were also collected from two
reference populations without increased thallium intake—one group consisting of 31 persons
living in a small (rural) city in northwest Germany and a second group consisting of 10 persons
living in an urban area in Dusseldorf, Germany. The study investigators did not perform specific
tests for toxicity but surveyed for the presence of certain symptoms by using questionnaires.
Thallium exposure was assessed by measurements in urine and hair. The thallium body burden
of the study population was increased over the reference populations, as indicated by a mean
urinary thallium level of 5.2 ± 8.3 ug/L (range: <0.1-76.5 ug/L) in the study population
compared to the reference population means of 0.4 ± 0.2 ug/L (rural) and 0.3 ± 0.2 ug/L (urban)
(range: 0.1-1.2 ug/L). The predominant contributing factor to the thallium burden was
consumption of homegrown fruits and vegetables. When the consumption of homegrown foods
was restricted, thallium exposure was reduced, as indicated by decreased thallium in the urine.
No correlation between dermal or GI symptoms and thallium level was observed. There was a
negative correlation between urinary thallium and hair loss (13.6% with urine levels <2 ug/L,
6.6% with urine levels 2-20 ug/L, and 5.9% with urine levels >20 ug/L; 10.7% with hair levels
<10 ng/g, 9.6% with hair levels 10-50 ng/g, and 2.3% with hair levels >50 ng/g). These data
appear to conflict with other reports that indicate hair loss increases with increasing thallium
exposure. A positive association was observed among thallium levels in urine or hair and the
following self-reported symptoms: sleep disorders, tiredness, weakness, nervousness, headache,
and psychological alterations as well as neurological and muscular symptoms (Brockhaus et al.,
1981).
Dolgner et al. (1983) examined the potential developmental effects of thallium in this
same German population. Of 300 births registered in Lengerich between January 1, 1978, and
16
-------
August 31, 1979, questionnaires on health status and maternal risk factors were completed by the
mothers of 297 infants. One hundred fifty-four urine and 164 hair samples were analyzed for
thallium content. All children with suspected congenital malformations or other abnormalities
were examined physically, and medical histories of mothers were taken. Eleven out of the
297 births were identified as exhibiting congenital malformations or abnormalities (confirmed by
a pediatrician) with five major malformations noted. Two of the five major malformations in the
study population were determined by the authors to likely be due to hereditary factors.
The observed rate of congenital malformations in the study population (5/297) was
compared to the expected rate of 0.8/297 births based on annual statistics from the North Rhine-
Westphalia region of Germany for 1974-1978. Congenital malformations in the reference
population were thought to be underreported because reporting of birth defects is not required on
birth certificates in that area of Germany. The study authors noted that other investigations
reported an incidence of 2-3% of birth defects among live births, a value that is consistent with
1.7% incidence of birth defects in the study population (5/297 for major malformations) and
3.7% (11/297) for all malformations. The study authors concluded that a causal relationship
between thallium exposure and congenital malformations in this population was unlikely.
However, study deficiencies, including lack of information on exposure to thallium at the time of
pregnancy, limit the strength of this study.
Navas-Acien et al. (2005) examined the association between urinary levels of various
metals, including thallium, with peripheral arterial disease (PAD) in a cross-sectional analysis of
790 participants in NHANES 1999-2000. Thallium was not associated with PAD in this sample
of the U.S. population.
4.1.3. Occupational Exposure
Schaller et al. (1980) examined 128 men (ages 16-62 years) who were exposed to
thallium for 1-42 years in three cement manufacturing plants in the Franconia region of
Germany. Health effects were determined through medical histories and a physical examination
for symptoms. Information on the scope of the physical examinations was not provided.
Analyses of roasted pyrites and electro-filter dust confirmed the presence of thallium in various
production areas in the plants. The median concentration of thallium in the urine in exposed
workers was 0.8 ug/g Tl creatinine with a range of <0.3-6.3 ug/g Tl creatinine. The range in
20 individuals without known occupational exposure was <0.3-1.1 ug/g Tl creatinine (median
concentration not reported). Medical histories and physical examinations did not indicate
thallium poisoning. The health status of exposed workers, however, was not compared with an
unexposed reference population, and a single measurement of urinary thallium did not provide a
measure of past exposures.
Thirty-six cement plant workers (presumably in Germany) were examined for clinical
and electrophysiological parameters (Ludolph et al., 1986). Thallium levels were found to be
17
-------
elevated in the blood of 16 workers, urine of 5 workers, and hair of 5 workers. It was not noted
if these were all separate cases or if elevations in all three parameters occurred in the same
individuals. The study determined that 28-39% of the individuals had some form of peripheral
and central motor and sensory impairment. The neurological impairments could not
conclusively be attributed to thallium exposure because half the patients suffered from
concurrent diseases (including peptic ulcer, diabetes, disorders of joints and connective tissue,
and hypertensive vascular disease), which could possibly cause neuromuscular impairment. No
controls were employed, and no correlations were made with the levels of thallium in individuals
and their disease states.
In another occupational study, Marcus (1985) examined medical records for 86 workers
(sex not reported) occupationally exposed to thallium at a magnesium seawater battery factory.
Exposure was determined by measuring thallium in urine samples. Marcus also examined the
records of 79 unexposed workers matched for age, length of employment, shift pattern, and type
of work. Exposed workers did not have an increase in incidence of benign neoplasms or any
other clinical diagnoses when compared with unexposed workers. This study is limited by lack
of exposure quantitation, the size of the cohort, and unknown length of follow-up.
Although there are many case reports of thallium poisoning in the literature, the doses
were largely unknown because ingestion was accidental or occurred through criminal poisoning.
Given the severity of reported symptoms, most of the exposures were likely to have been
relatively large. The few epidemiology studies that looked at populations surrounding a cement
factory that released thallium only attempted to compare thallium exposure with congenital
malformations or surveyed symptoms. None of the studies specifically studied cancer as an
endpoint. Overall, the available epidemiology literature is considered limited and inconclusive.
4.2. LESS-THAN-LIFETIME AND CHRONIC STUDIES AND CANCER BIOASSAYS
IN ANIMALS—ORAL AND INHALATION
4.2.1. Oral Exposure
4.2.1.1. Acute and Subchronic Studies
4.2.1.1.1. Rats. In a study performed by Midwest Research Institute (MRI) (1988) for EPA's
Office of Solid Waste, male and female Sprague-Dawley rats (45 days old, 20/sex/group) were
administered 0 (untreated and vehicle controls), 0.01, 0.05, or 0.25 mg/kg-day of an aqueous
solution of thallium (I) sulfate (approximately 0, 0.008, 0.04, or 0.20 mg/kg-day Tl) by gavage
for 90 days. The study was conducted in compliance with EPA good laboratory practice (GLP)
mandates. The MRI (1988) study is an unpublished study; accordingly, an external peer review
was initiated by EPA in November 2006. Body weight, food consumption, hematologic and
clinical chemistry parameters, ophthalmologic examinations, gross pathological observations,
and organ weights (liver, kidneys, brain, gonads, spleen, heart, and adrenals) were recorded for
all animals. Neurotoxicological examinations (three times/week) were performed on six
18
-------
rats/sex/group; these examinations were apparently observational (further details were not
provided in the study report). Tissues from three rats/sex/group were prepared for
neuropathologic examination. Complete histopathologic examinations (including
neuropathologic examinations) were conducted for the vehicle control and 0.2 mg/kg-day Tl
groups only; for the other three groups, only the livers, lungs, kidneys and gross lesions were
examined histopathologically. Neuropathologic examinations included the following: dorsal and
ventral root fibers of the spinal nerves, dorsal root ganglia, spinal cord at C3-C6 and L1-L4, and
six sections of the brain.
There were no statistically significant differences in body weight, food consumption, or
absolute and relative organ weights among control groups and groups receiving thallium (I)
sulfate. The study authors concluded that the histopathologic examination did not reveal any
treatment-related effects.
Lacrimation (secretion of tears), exophthalmos (abnormal protrusion of the eyeball), and
miosis (contraction of the pupil) were observed at higher incidences in the treated male and
female rats compared with both untreated and vehicle controls (Table 4-2). Ophthalmologic
examination and gross and histopathologic examination of the eyes, however, revealed no
treatment-related abnormalities. The incidence of clinical observations related to the coats
(including rough coat, piloerection, shedding, and alopecia) and behavior (including aggression,
tension/agitation, hyperactivity, vocalization, and self-mutilation) were also elevated in male and
female rats at the higher doses (Table 4-2).
19
-------
Table 4-2. Selected clinical observations in Sprague-Dawley rats treated
with thallium sulfate for 90 days
Observation11
Untreated
control
Vehicle
control
0.008 mg/kg-day
0.04 mg/kg-day
0.2 mg/kg-day
Male
Coat/skin
Rough coat
Piloerection
Shedding
Alopecia
1/20
0/20
0/20
2/20
3/20
0/20
0/20
1/20
11/20
1/20
4/20
4/20
16/20
4/20
10/20
9/20
19/20
13/20
8/20
4/20
Eyes
Lacrimation
Exophthalmos
Miosis
Behaviorb
1/20
1/20
0/20
3/20
6/20
5/20
1/20
0/20
19/20
12/20
5/20
7/20
20/20
20/20
7/20
6/20
20/20
20/20
15/20
7/20
Female
Coat/skin
Rough coat
Piloerection
Shedding
Alopecia
1/20
0/20
0/20
4/20
0/20
0/20
0/20
1/20
1/20
0/20
2/20
4/20
5/20
3/20
3/20
9/20
11/20
8/20
13/20
12/20
Eyes
Lacrimation
Exophthalmos
Miosis
Behaviorb
7/20
5/20
2/20
2/20
6/20
6/20
3/20
2/20
20/20
19/20
1/20
0/20
20/20
20/20
11/20
1/20
20/20
20/20
8/20
7/20
aListed as number of animals with the sign observed at least once during the 90-day study.
bAnimals exhibiting one or more behavioral observations at least once during the 90-day study, including the
following: aggression, tension/agitation, hyperactivity, vocalization, serf-mutilation.
Source: MRI(1988).
As noted above, the incidence of alopecia was increased, particularly in female rats (see
Table 4-3). Examination of individual animal clinical observation data for female rats from the
MRI (1988) study showed that alopecia was first observed in control and treated groups
anywhere from study day 44 to 60. Based on a statistical analysis performed by the U.S. EPA1,
the incidence of alopecia (total number of cases in each dose group) was statistically
1 A statistical analysis of the incidence of alopecia (based on the total number of cases of alopecia in each dose
group) was performed by EPA by using Fisher's exact test. Incidence in the treated groups was compared with
incidence in the untreated control, vehicle control, and pooled control.
20
-------
significantly elevated relative to controls in mid-dose males and mid- and high-dose females.
Most instances of alopecia in females were attributed to barbering behavior (where fur was
present but cropped short). Of the 12 high-dose females with alopecia, 5 instances were not
totally attributed to barbering behavior. Histopathologic examination revealed atrophy of the
hair follicles in two high-dose female rats. Because the skin was examined for histopathologic
changes only in the vehicle control and high-dose groups, no information on dermal
histopathology was available for the low- and mid-dose groups. The two high-dose females with
atrophy of the hair follicles also had alopecia; whether the hair follicle atrophy and alopecia
occurred at the same location on the rats could not be determined from the study report. The
study authors concluded that the alopecia was attributable to the cyclic pattern of hair growth in
rodents. Consequently, the authors did not consider these findings to be lexicologically
significant.
Table 4-3. Incidence of alopecia in rats
Dose
(mg/kg-day Tl)
0 (untreated control)
0 (vehicle control)
0.008
0.04
0.2
Males
Alopecia3' b
2/20
1/20
4/20
9/20e
4/20
Hair follicle atrophy0
d
0/20
d
d
0/20
Females
Alopecia3' b
4/20
1/20
4/20
9/20f
12/20e
Hair follicle atrophy0
d
0/20
d
d
2/20
aNumber of animals with alopecia at least once during the 90-day study based on clinical observations.
bOf the animals with alopecia, the following are the numbers of cases in each dose group that the study authors
stated were "not totally attributed to barbering behavior":
Males: untreated control, 1; vehicle control, 0; 0.008 mg/kg-day, 2; 0.04 mg/kg-day, 4; 0.2 mg/kg-day, 1.
Females: untreated control, 0; vehicle control, 0; 0.008 mg/kg-day, 1; 0.04 mg/kg-day, 3; 0.2 mg/kg-day, 5.
°Based on histopathologic observation.
dSkin was not examined for histopathologic lesions.
Incidence of alopecia (total number of cases) was statistically significantly elevated (p < 0.05) relative to incidence
in vehicle control, incidence in untreated control, and pooled incidence of vehicle and untreated control, based on
Fisher's exact test performed by EPA.
Incidence of alopecia (total number of cases) was statistically significantly elevated (p < 0.05) relative to incidence
in vehicle control and pooled incidence of vehicle and untreated control, based on Fisher's exact test performed by
EPA.
Source: MRI(1988).
Subtle but statistically significant changes were observed in several blood chemistry
parameters that the investigators considered probably treatment related. Specifically, dose-
related increases in AST, lactate dehydrogenase (LDH), and sodium levels and decreases in
blood sugar levels were detected in male and female rats after 30 and 90 days of exposure.
Reported values for the selected blood chemistry parameters are summarized in Table 4-4. Other
21
-------
changes in blood chemistry parameters were less consistent across species, dose groups, and
exposure durations.
At 90 days, the differences in AST, LDH, sodium, and blood sugar levels in dosed male
and female rats were no greater than +31, +38, +4, and -21%, respectively, of the vehicle control
group values. The investigators observed that the increases in AST and LDH levels could
indicate a possible effect of treatment on cardiac function, that increases in LDH coupled with
subtle changes in electrolytes could indicate an effect on renal function, and that, in rare
instances, a decrease in blood sugar coupled with an increase in sodium occurs as a defense
mechanism for maintaining cellular integrity. The investigators concluded that none of the
changes observed in the blood chemistries of male or female rats during the study were of
sufficient magnitude to significantly affect the health status of the animals. Further,
histopathologic evaluation did not confirm any cellular damage suggested by the clinical
chemistry findings.
Table 4-4. Selected blood chemistry values
Endpoint
Study
day
Untreated
control
Vehicle control
0.008 mg/kg-day
0.04 mg/kg-day
0.2 mg/kg-day
Males'1
AST
(I.U.)
LDH
(I.U.)
Na
(meq/L)
Blood sugar
(mg/100 mL)
30
90
30
90
30
90
30
90
91 ±26.5
77 ± 19.7
795 ± 322
587 ± 305
148 ±1.3
144 ±1.6
100 ±22.1
158 ±15.6
108 ± 18.6
87 ± 17.8
1206 ± 424b
856 ± 385
149 ±2.4
147 ± 2.0b
97 ±18.1
138 ± 16.8b
128 ± 24.5b
99 ±20.4
1333±340b
1003 ± 363b
152 ± 4.0b
147±1.9b
93 ± 10.0
131±17.6b
134 ± 29.0b'c
113±27.0b'c
1396 ± 407b
1071 ± 507b
154 ± 2.5b'c
149 ± 2.0b'c
90 ±18.3
121±15.7b
152±20.1b'c
114±31.1b'c
1802 ± 341b'c
1119±477b
153±2.1V
151±2.2b'c
62 ± 14.8b'c
113±22.4b'c
Females"
AST
(I.U.)
LDH
(I.U.)
Na
(meq/L)
Blood sugar
(mg/100 mL)
30
90
30
90
30
90
30
90
95 ±22.8
77 ±19.2
1047 ±335
745 ± 320
148 ±1.7
146 ±1.8
103 ±23. 9
110 ±28.7
115 ±30.3
90 ±19.1
1277 ± 495
881 ±273
150 ± 1.9
146 ± 1.0
80±13.3b
89 ±15. 9
127 ± 27.8b
93 ±33.1
1402 ±501
823 ± 354
153±4.1b'c
148±1.8b'c
80 ± 9.0b
103 ± 19.9
149 ± 26.8b'c
lll±30.7b
1763 ± 370b'c
1044 ± 436
154 ± 2.8b'c
150 ± 2.0b'c
67 ± 20.0b
88 ± 20.4
154 ± 18.2b'c
112±31.0b
1764±361V
1219±338b
155±2.5b'c
152±1.0b'c
50±11.8b'c
70 ± 18.0b
aMean ± standard deviation of 7-10 rats.
bSignificantly different (p < 0.05) from the untreated control group.
Significantly different (p < 0.05) from the vehicle control group.
Source: MRI(1988).
The authors concluded that the minor dose-related changes in this study did not affect the
health status of the treated animals and therefore were not lexicologically significant and
22
-------
identified the highest dose, 0.25 mg/kg-day thallium (I) sulfate (0.20 mg/kg-day Tl), as a no-
observed-effect level (NOEL). However, upon further analysis by EPA of the MRI (1988)
findings as part of this health assessment, a different determination was reached regarding the
no-observed-adverse-effect level (NOAEL) and lowest-observed-adverse-effect level (LOAEL)
(see the discussion in Section 5.1.1).
Manzo et al. (1983) administered drinking water containing thallium (I) sulfate at a
concentration of 10 mg/L Tl (approximately equivalent to a dose of 1.4 mg/kg-day Tl based on
reported thallium intakes and an assumption that the rats weighed 200 g) to 80 female Sprague-
Dawley rats for 36 weeks. Mortality was 15 and 21% after 40 and 240 days of treatment,
respectively. After 4+ weeks (32 days) of treatment, hair loss appeared and involved about 20%
of the animals thereafter. Functional and histopathologic changes were observed in the
peripheral nerves, including changes in motor and sensory action potentials and histopathologic
changes in the sciatic myelin sheath and axonal destruction characterized by Wallerian
degeneration (degeneration of the axon and its myelin sheath distal to a site of injury),
mitochondrial degeneration, neurofilamentous clustering, and elevated lysosomal activity.
Ten adult male albino rats were administered 0.8 mg/kg (l/20th of the LD50) of thallium
(I) sulfate orally (presumably via gavage) for 3 months (El-Garawany et al., 1990). Blood
samples were obtained initially and at monthly intervals. At all three monthly intervals, the
treated group had statistically significantly (p < 0.001) increased levels of blood urea, serum
creatinine, serum bilirubin, and serum ALT. The largest increase (<90%) in these parameters
occurred in the first month, with smaller increases (>15%) occurring for each additional month.
Mourelle et al. (1988) examined the effects of silymarin, an antioxidant, on various
biochemical indicators of liver damage in male Wistar rats (200-250 g) induced by oral
administration (gavage) of thallium (I) sulfate (10 mg/kg) dissolved in water. The controls were
given vehicle only. Ten rats per group were sacrificed at 0, 24, 48, 72, and 96 hours and 5, 10,
and 20 days after treatment. Without silymarin administration, thallium administration produced
a statistically significant (p < 0.05) decrease in the content of glycogen and reduced glutathione
(GSH) and caused a statistically significant (p < 0.05) increase in malondialdehyde (MDA)
production and triglycerides in the liver 48 hours after treatment. (MDA production and GSH
content in the liver served as indicators of lipid peroxidation.) Levels of serum alkaline
phosphatase were increased and liver cell membrane alkaline phosphatase activity was decreased
after 24 hours and remained unchanged for 5 days. Furthermore, Na+/K+-ATPase activity in the
liver cell membranes was rapidly reduced within 24 hours of thallium treatment; the decrease
persisted through day 5 and began to rebound by day 10, with values similar to the control by
day 20. Serum and liver cell membrane gamma-glutamyl transpeptidase and serum ALT were
significantly (p < 0.001) elevated by 24 hours and remained elevated through day 5.
Administration of silymarin (100 mg/kg i.p.) completely prevented these biochemical changes.
The authors suggested that silymarin acted by stabilizing membranes via some antioxidant
23
-------
property. During the 20 days, none of the rats treated with thallium alone died, but the rats
exhibited signs of toxicity that included hypomotility and piloerection.
Downs et al. (1960) fed groups of Wistar-derived albino rats (5/sex/dose) diets containing
nominal concentrations of 0, 5, 15, or 50 mg thallium (I) acetate/kg (or ppm) in the diet
(corresponding to approximately 0, 0.4, 1.2, or 3.9 mg/kg-day Tl for 100 g rats, assuming food
consumption of 10 g/day). Animals were allowed ad libitum access to these diets for 15 weeks.
At the 50 ppm dose level, mortality was 100% by week 5 in males and by week 13 in females.
By week 15, 4/10 control animals died (2/sex), making interpretation of survival in the remaining
dose groups difficult (15 ppm, 3/5 males and 1/5 females died; 5 ppm, 2/6 males and 0/4 females
died). An additional treatment group (30 ppm) and control group (corresponding to 0 and
2.4 mg/kg-day Tl) were added 6 weeks after the study had been initiated and were maintained on
the diet for 9 weeks. At the end of the 9 weeks, 2/5 male and 1/5 female controls were dead and
4/5 males and 3/5 females at 30 ppm were dead. At termination, the only gross finding was
alopecia in the 15 and 30 ppm groups. The alopecia was noted beginning 2 weeks after
commencement of the diet, with the rats nearly free of hair at termination. The authors reported
a slight increase in kidney weight (doses not specified; data not shown). The authors also
reported that histopathologic evaluations did not indicate treatment-related pathology, but they
did not prepare skin sections. The study findings for alopecia suggest a NOAEL and LOAEL of
0.4 mg/kg-day Tl and 1.2 mg/kg-day Tl, respectively, for this endpoint. Because mortality
occurred in rats in both the control and treated groups, it is not possible to determine whether the
deaths in low-dose (5 ppm) male rats were related to thallium exposure. Therefore, a NOAEL
and LOAEL cannot be reliably established for this study.
Downs et al. (1960) also examined the effects of thallium (III) oxide on weanling Wistar-
derived albino rats (five rats/sex/treatment). Rats received 0, 20, 35, 50, 100, or 500 mg/kg (or
ppm) thallium (III) oxide in the diet for 15 weeks. This was equivalent to doses of 0, 1.8, 3.1,
4.5, 9.0, or 44.8 mg/kg-day Tl, respectively. All rats (males and females) treated with 50 ppm
and greater in the diet died before 8 weeks. The mortality rates in the remaining groups at
15 weeks were as follows: 1/5 control males, 0/5 males treated with 20 ppm, and 4/5 males
treated with 35 ppm; 0/5 control females, 2/5 females treated with 20 ppm, and 2/5 females
treated with 35 ppm. Thallium (III) oxide caused a dose-related decrease in body weight at
15 weeks. Body weight reductions relative to the control were 50 and 180 g in males treated
with 20 and 35 ppm dietary doses, respectively, and 50 g in females treated with 35 ppm in the
diet. Males treated with either 20 or 35 ppm in the diet had marked hair loss beginning around
4 weeks, with near complete hair loss after 6 weeks; females were less affected.
There was a statistically significant (p < 0.05) increase in absolute kidney weights in
males and females treated with 20 ppm and females treated with 35 ppm and a dose-response
trend in kidney to body weight ratio. Histopathologic examination did not reveal any alterations
in the kidney related to thallium treatment. Histopathologic evaluation of the skin revealed a
24
-------
decrease in the number of hair follicles and hair shafts, atrophy of the remaining follicles,
decrease in the size of the sebaceous glands, and hyperkeratinized epidermis. However, the
incidence by dose was not presented. The lowest level tested, 1.8 mg/kg-day Tl (20 ppm
thallium (III) oxide in the diet), is considered to be a LOAEL based on findings of alopecia and
significant elevations in kidney weights for male and female rats. A NOAEL was not identified
for this study.
Leloux et al. (1987) investigated the acute toxicity of oral exposure to thallium (I) nitrate
in the adult Wistar rat. In the first experiment, a single dose of 20 mg/kg thallium (I) nitrate was
administered via gavage to male and female rats (three per sex); all males and females were
found dead within 40 and 54 hours post-dosing, respectively. Increases in absolute kidney (36%,
females; 61%, males) and adrenal (47%, females; 100%, males) weights were observed
following the single exposure. The second experiment involved administering four daily gavage
doses of 1 mg/kg-day thallium (I) nitrate to 20 animals of each sex. Male rats treated with four
doses began to lose their hair 96 hours after the first exposure. All treated animals had diarrhea.
After the fourth gavage dose, 2/20 males and 2/20 females died. Two more females died within
126 hours, and 11 females and 15 males died within 168 hours. Three rats of each sex were
sacrificed at 126 hours post-dosing for gross pathological examination and organ weight
changes. The remaining two females were sacrificed at 192 hours post-dosing. Treated animals
weighed less than the untreated controls. The tissues did not demonstrate any macroscopic
degenerative changes, but there was an increase in the absolute weights of the kidneys (33%,
females; 48%, males) and eyes (54%, females; 34%, males). Histopathology was not performed.
4.2.1.1.2. Dogs. Reports of thallium toxicity in dogs are limited to a few cases in the literature
of accidental exposure. A 9-month-old Doberman pinscher accidentally consumed mole bait
containing 1% thallium (Waters et al., 1992). Two days later the dog was lethargic, vomited
blood, and had bloody feces. The dog had moderate hypoproteinemia and a slight prolonged
activated clotting time. The dog's condition was improved by the third day following supportive
care, including treatment with activated charcoal.
Thomas and McKeever (1993) reported a case of a 1-year-old neutered male miniature
schnauzer that had ingested an unknown amount of bread soaked in thallium (the level in one
piece of bread was 1.6 ppm). Beginning symptoms were lethargy, followed 2 weeks later by
severe, rapidly progressing alopecia. No abnormalities were found in a complete blood count
test, serum chemistry profile, urinalysis, or abdominal radiographs. Diphenylthiocarbazone
treatments (40 mg/kg three times daily) were started upon establishing thallium toxicity. On the
second day of veterinary treatment, the dog showed signs of respiratory distress and was
euthanized due to its poor condition. An autopsy revealed severely congested and edematous
lungs, congestion of the liver and kidneys, and areas of congestion and hemorrhage in the
25
-------
pancreas. Histologic evaluations demonstrated abnormalities in the lungs, kidneys, liver, and
pancreas. Thallium was detected in the liver (11 ppm), kidneys (12 ppm), and spleen (7 ppm).
Histopathology of the skin from 13 cases of thallium poisoning in dogs revealed
dyskeratotic and necrolytic changes in the skin and hair follicles (Schwartzman and Kirschbaum,
1961). The most prominent features were massive parakeratosis, spongiform abscess formation,
and induction of telogen follicles.
4.2.1.2. Chronic Studies and Cancer Bioassays
There are no chronic animal studies or cancer bioassays for thallium reported in the
literature.
4.2.2. Inhalation Exposure
No studies were identified that examined the effects of inhaled thallium in animal
models.
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES
4.3.1. Reproductive Toxicity
Effects of thallium on male reproduction have been investigated in rats (Formigli et al.,
1986; Gregotti et al., 1985; Zasukhina et al., 1983) and mice (Wei, 1987). These studies suggest
that thallium exposure can produce effects on the testes and sperm. None of the available
reproductive toxicity studies, however, used standard protocols for evaluating reproductive
endpoints. No studies of the potential reproductive toxicity of thallium in female experimental
animals were identified.
Male Wistar rats (10/group) were administered drinking water containing 10 ppm
thallium (I) sulfate (approximately 0.7 mg/kg-day Tl based on reported daily thallium
consumption [270 ug Tl/rat] and initial body weights [350-380 g]) (Formigli et al., 1986). The
compound was administered for 30 and 60 days. Although the study authors stated that the
controls were pair-fed, they also stated that food was available ad libitum and that thallium did
not affect food consumption. No abnormalities were observed after 30 days of treatment.
However, after 60 days of treatment, the following testicular effects were observed:
disarrangement of the tubular epithelium, cytoplasmic vacuolation and distention of smooth
endoplasmic reticulum of the Sertoli cells, reduced testicular p-glucuronidase activities (an
enzyme primarily located in the Sertoli cell and spermatogonia), high concentrations of thallium
in the testes, and reduced sperm motility. Plasma testosterone levels were within normal limits.
p-Glucuronidase activity was also affected after 60 days of treatment and ultrastructural changes
were observed in the Sertoli cells. From these results, a LOAEL of 0.7 mg/kg-day Tl (10 mg/L)
was identified.
26
-------
Gregotti et al. (1985) also reported p-glucuronidase activity and ultrastructural changes in
the Sertoli cells after 60 days of treatment. Gregotti et al. (1992) further examined this effect
in vitro and demonstrated that thallium (even at the lowest concentration) causes a dose- and
time-dependent detachment of germ cells from Sertoli cells when testicular cells were treated
with thallium concentrations corresponding to 1.4, 7, and 35 ug/g testis, estimated from protein
content of cultures.
Zasukhina et al. (1983) performed a dominant lethal test with male rats that were given
daily oral doses of thallium (I) carbonate (0.005, 0.05, and 0.5 ug/kg-day) for 8 months and
subsequently mated with untreated females. The authors concluded that thallium carbonate
produced a treatment-related enhancement of embryonic mortality. The authors' conclusion was
not, however, supported by the data presented in the paper. The number of implantation sites
(10.11 ± 0.48 to 11.05 ± 0.49), number of "yellow bodies" in pregnant rats (11.11 ± 0.35 to
11.75 ± 0.33), and number of live embryos (9.77 ± 0.48 to 10.87 ± 0.37) were similar across the
control and treated groups. ("Yellow bodies" were not further defined in the paper but may
mean corpora lutea.) Because the paper lacked statistical analysis, it was unclear whether the
differences were statistically significant. Only the number of resorptions in the control group
(0.87 ±0.13) was appreciably higher than that in the treated groups (0.22 ± 0.12 to 0.33 ± 0.13).
In addition, the authors' calculation of total embryonic deaths could not be reproduced from the
data provided. Finally, overall confidence in the reported findings was low because of
inadequate reporting (e.g., the number of male rats exposed and the rat strain were not reported),
the use of nonstandard terminology, the relatively small number of pregnant females (16-18 per
group), and lack of statistical analysis.
Wei (1987) administered 0, 0.001, 0.01, 0.1, 1.0, or 10.0 mg/L thallium (I) carbonate in
drinking water for 6 months to groups of male Kunming mice (20/group) weighing 15-20 g at
study initiation. (Assuming an average body weight of 20 g over the exposure period and
drinking water ingestion rate of 6 mL/day [Derelanko and Hollinger, 1995], these exposure
levels are approximately equivalent to doses of 0.0003, 0.003, 0.03, 0.3, and 3 mg/kg-day Tl.
These doses are approximate only because the age of animals at study initiation, terminal body
weights, and water consumption were not reported.) At the end of the exposure period, half of
the male mice (10/group) were sacrificed for epididymal sperm examination. The remaining
males (10/group) were housed with untreated females (1:2 ratio) for 1 week to evaluate male
reproductive function. At the end of the 1-week mating period, these male mice were also
sacrificed for epididymal sperm examination. Untreated female mice were sacrificed on
gestation day 20 and evaluated for the following measures of reproductive function: number of
pregnant female mice, number of live and dead fetuses, number of implantations, and number of
early resorptions. Water intake, body weights, behavior, and animal health were reportedly
assessed; however, this information was not provided in the study report. The author reported
that sperm motility (rapid speed, sperm immobility) was affected at the lowest drinking water
27
-------
concentration (0.001 mg/L) tested. Effects were shown to increase with increasing
concentration, thus indicating a dose-response relationship. At 0.01 mg/L and higher, the
number of dead sperm was statistically significantly increased. Sperm count was statistically
significantly reduced and the percent of deformed sperm was increased at concentrations of
0.1 mg/L and higher. The author observed that there was an adverse effect on sperm quality
(motility) at low doses, and, as the dose increased, there was an accompanying decrease in sperm
count in addition to the motility change. Examination of reproductive function in the group of
mice housed for 1 week with untreated females revealed that the reproductive index (number of
pregnant female mice/number of mated female mice) and the number of implantations were not
statistically different between treated and control animals. The mean number of live fetuses was
statistically significantly increased at concentrations of 0.01 mg/L and above. The percent of
dead fetuses was significantly lower than in the control group at concentrations of 0.001, 0.01,
and 0.1 mg/L but was increased at the two highest concentrations (1 and 10 mg/L). Review of
the reported results reveals that a number of male mice were not accounted for at study
termination. Of the initial 20 male mice/group, sperm results were provided for only 12 mice in
the 0.1 mg/L group, 10 mice in the 10 mg/L group, and 16 mice in the remaining groups. No
explanation is provided for the loss of animals over the 6-month study. The author concluded
that the lowest dose tested (0.001 mg/L thallium (I) carbonate) was a LOAEL, causing
reproductive effects in male mice.
Table 4-5 summarizes thallium toxicity in animals following oral exposure.
28
-------
Table 4-5. Thallium toxicity in animals following oral exposure
Reference
Species
Age
Sex
Route
Dose and duration
NOAEL
LOAEL
Effect
Acute studies
Leloux et al.
(1987)
Leloux et al.
(1987)
Mourelle et al.
(1988)
Rat
3/sex
Rat
10/sex/
group
Rat
10/group
Adult
Adult
NSb
Both
Both
Male
Oral
(gavage)
Oral
(gavage)
Oral
(gavage)
20 mg/kg thallium
(I) nitrate; single
dose
0, 1 mg/kg thallium
(I) nitrate; once daily
for 4 days
0, 10 mg/kg thallium
(I) surf ate; single
dose. Sacrificed at
24 hours to 2 days
after dosing
NIa
NI
NI
15 mg/kg Tl
0.77 mg/kg Tl
8.1 mg/kg Tl
Difficulty breathing; rough coat; increased
absolute kidney, adrenal weights; death
Alopecia; diarrhea; increased absolute
kidney, eye weights; death
Liver changes: increased triglycerides and
lipid peroxidation; decreased glutathione and
glycogen; increased alkaline phosphatase in
serum and liver cell membranes
Subchronic studies
Downs et al.
(1960)
Downs et al.
(1960)
El-Garawany
et al. (1990)
Manzo et al.
(1983)
Rat/
5/sex/
group
Rat
5/sex/
group
Rat
n=10
Rat
n=80
NS
Weanling
NS
NS
Both
Both
Male
Female
Oral
(feed)
Oral
(feed)
Oralc
Oral
(DWd)
0, 5, 15, or 50 ppm
thallium (I) acetate
(corresponding to 0,
0.4, 1.2, or
3.9mg/kg-dayTl);
15 weeks
0 or 30 ppm
(corresponding to 0
or 2.4 mg/kg-day
Tl); 9 weeks
0, 20, 35, 50, 100,
and 500 ppm
thallium (III) oxide
(corresponding to 0,
1.8, 3.1,4.5, 9.0, and
44.8 mg/kg-day Tl);
15 weeks
0.8 mg/kg thallium
(I) sulfate; 90 days
10 mg/L Tl as
thallium (I) sulfate;
36 weeks
0.4 mg/kg-
day Tl*
NI
NI
NI
1.2 mg/kg-
day Tl*
1.8 mg/kg-
day Tl (20
ppm)
0.65 mg/kg-
day Tl
1.4 mg/kg-
day Tl
Alopecia; increased kidney weight; mortality
in treated and control groups.
*The NOAEL and LOAEL are for alopecia.
Because of reported mortality in the control
and treated groups, a study NOAEL and
LOAEL cannot be reliably determined.
Reduced body weight; alopecia; increased
mortality; increased absolute and relative
kidney weights
Increased blood urea; serum creatinine;
serum bilirubin; serum ALT
Nerve histopathology; alopecia; mortality
29
-------
Table 4-5. Thallium toxicity in animals following oral exposure
Reference
MRI (1988)
Species
Rat
20/sex/
group
Age
45 days
Sex
Both
Route
Oral
(gavage)
Dose and duration
0,0.01,0.05, or 0.25
mg thallium (I)
sulfate/kg
(corresponding to 0,
0.008, 0.04, or
0.20 mg/kg-day Tl);
90 days
NOAEL
NI
LOAEL
0.008 mg/kg-
day Tle
Effect
Increased incidence of alopecia and other
observations related to coat (rough coat,
piloerection, shedding); lacrimation,
exophthalmos, and miosis; and various
behavioral observations; statistically
significant increases in AST, LDH, and
sodium levels; decreased blood sugar levels.
The study authors identified 0.2 mg/kg-day
Tl as the NOAEL.
Reproductive and developmental toxicity
Formigli et al.
(1986);
Gregotti et al.
(1985)
Wei (1987)
Rossi et al.
(1988)
Rat
10/group
Mouse
Rat
Adult
NS
Perinatal
Male
Male
Both
Oral
(DW)
Oral
(DW)
Mother's,
then
pup's DW
(DW)
0, 10 ppm thallium
(I) sulfate; 30 or 60
days
0,0.001,0.01,0.1,
1.0, andlOmg/L
thallium (I)
carbonate
(corresponding to 0,
0.0003, 0.003, 0.03,
0.3, and 3 mg/kg-
day Tl); 6 months
0, 1 mg/dL of
thallium (I) sulfate
Day 1 of gestation to
weaning then thru 60
days
NI
NI
NI
0.7 mg/kg-
day Tl
0.0003 mg/kg
-day Tl
NI
Testicular effects: tubular epithelium
disarrangement; cytoplasmic vacuolation;
reduced sperm motility; distention of smooth
endoplasmic reticulum of Sertoli cells;
reduced p-glucuronidase activity
Decreased sperm motility and counts;
increase in deformed sperm; decrease in live
fetuses.
Dose estimated from an assumed average
body weight of 20 g and drinking water
ingestion rate of 6 mL/day.
Prenatal and postnatal exposure caused a
delay in the development of the pilus
apparatus by 50 days; reduction of the a- and
(3-adrenergic and muscarinic vasomotor
reactivity noted.
aNI = not identified.
bNS = not specified.
Presumably via gavage.
dDW = drinking water.
eSee discussion of the NOAEL and LOAEL determination in Section 5.1.1.
30
-------
4.3.2. Developmental Toxicity
Developmental toxicity studies in the rat (Barroso-Moguel et al., 1992; Rossi et al., 1988;
Gibson and Becker, 1970) and chicken embryo (Hall, 1985, 1972; Karnofsky et al., 1950)
provide evidence that thallium exposure during development can produce abnormalities
(including effects on the developing vascular autonomic nervous system and bones) and reduced
fetal body weight. Of the studies in rats, only one involved oral drinking water exposure to
thallium (Rossi et al., 1988); in other developmental rat studies, dams were exposed by i.p.
injection.
A group of NOS albino male and female rats was administered 1 mg/dL of thallium (I)
sulfate from day 1 of gestation to weaning (22 days after birth) via the dams' drinking water then
through their own drinking water until 60 days of age (Rossi et al., 1988). These rats were
considered prenatally exposed. Another group of NOS albino male and female rats was exposed
to 1 mg/dL of thallium (I) sulfate via the dams' drinking water from birth until weaning (22 days
after birth) then through their own drinking water until 60 days of age. These rats were
considered postnatally exposed. Both situations (pre- and postnatal exposure) caused a delay in
the development of the pilus apparatus by 50 days. A reduction of the a- and p-adrenergic and
muscarinic vasomotor reactivity also was noted. Authors noted that this reduction may be due to
one of the following mechanisms: probable reduction in the number and/or sensitivity of both a-
and P-adrenergic and muscarinic receptors or a change of cell membrane in relation to a possible
modification of potassium cell concentration.
Gibson and Becker (1970) administered thallium (I) sulfate (i.p.) to pregnant Simonsen
Sprague-Dawley rats during early (2.5 mg/kg on days 8, 9, and 10) or late (2.5 or 10 mg/kg on
days 12, 13, and 14) gestation. Fetuses were examined for abnormalities. All three thallium
treatments caused a statistically significant (p < 0.05) reduction in fetal body weight. Thallium
treatment (2.5 mg/kg) during early gestation caused a slight (not statistically significant) increase
in the incidence of hydronephrosis (29% in treated versus 16% in control) and missing or non-
ossified vertebral bodies (36% in treated versus 17% in controls). The 2.5 mg/kg thallium (I)
sulfate treatment administered during late gestation caused a statistically significant (p < 0.05)
increase in the incidence of hydronephrosis (47% in treated versus 16% in control) and missing
or non-ossified vertebral bodies (60% in treated versus 17% in controls). Increasing the dose to
10 mg/kg thallium (I) sulfate during late gestation did not increase the incidence of
developmental abnormalities. In fact, 10 mg/kg thallium (I) sulfate administered during late
gestation had no effect on the incidence of hydronephrosis and was comparable to the 2.5 mg/kg
dose administered during late gestation in the induction of missing or non-ossified vertebral
bodies (i.e., 60% in treated versus 17% in controls). Maternal toxicity (diarrhea, lethargy,
irritability, poor hair luster, and hair loss) was noted.
Barroso-Moguel et al. (1992) administered a single i.p. injection of 32 mg/kg aqueous
thallium (I) acetate to 20 newborn (24-hour-old) Wistar rats. Results were compared with those
31
-------
of five vehicle controls. Rats (four treated and one control per time point) were sacrificed at 24,
48, and 72 hours and at 7 and 50 days. Cartilaginous and osseous tissue alterations were noted.
Diarrhea was observed through 72 hours postinjection. Two rats surviving to 50 days
postinjection had persisting alopecia (one irreversible and one with discrete recovery). Although
skeletal images of 72-hour-old animals did not show any differences when compared with the
control, microscopic images of the distal third of the tibia showed disorganization and edema of
the fibroblasts of the fibrous layer. By day 7, delays in ossification in the right forelimb were
noted. Microscopic examination demonstrated a majority of pyknotic chondrocytes and the lack
of bone trabeculae calcification. Profound skeletal alterations were noticeable 50 days after
injection. Many of the cartilaginous cells were altered or dead, leading to a decrease of the
growth cartilage and scanty bone trabeculae with few osteoblasts. The bone marrow also had
few myeloblasts and megakaryocytes.
Hall (1972) incubated chick embryos in a forced-draft Humidaire incubator and injected
0.6 mg thallium (I) sulfate/0.5 mL saline into each embryo via the chorioallantoic membrane at
7 days of incubation. This dose caused a minimal lethal effect with survival varying from 94 to
100%. Treated embryos were smaller than controls from 10 days of incubation onward and by
18 days were 26% smaller than controls. Treated embryos failed to commence ossification or
had not progressed to similar developmental stages observed in the control. The long bones of
treated embryos were smaller (the tibia to a greater degree than the femur), contained less
organic material, and contained more water (as a percent of dry weight) than the untreated
control embryos. An abnormal distribution of the acid mucopolysaccharides and necrotic areas
in maturing hypertrophic chondrocytes were detected histologically. In addition, biochemical
assays verified the reduced acid mucopolysaccharide activity. Hall (1985) further demonstrated
that the critical period of susceptibility ended at 8 2/3 days of incubation. Tibial growth was
inhibited by thallium sulfate in 8-day-old embryos but not in 9-day-old embryos.
Achondroplasia (a birth defect characterized by imperfect bone formation) was also
induced in embryonic chicks via in vitro cultures with injection into the chorioallantoic
membrane (Hall, 1985, 1972) or injection into the yolk sac (Karnofsky et al., 1950). Karnofsky
et al. (1950) determined that thallium (I) sulfate was lethal to 2-day embryos at a dose lower than
would be necessary to induce achondroplasia. Treatment of 4-day-old chick embryos with 0.2,
0.5, 1.0, or 2.0 mg/egg induced 0, 45, 92, and 100% incidence of achondroplasia, respectively.
Although the data were not presented, the study authors reported that thallium (I) nitrate
produced achondroplasia at similar doses.
4.4. OTHER ENDPOINT-SPECIFIC STUDIES
A number of investigators have specifically examined the effect of thallium compounds
administered to experimental animals by injection (s.c., i.p., or i.v.) and reported effects on the
liver, kidneys, heart, and nervous system.
32
-------
4.4.1. Liver and Kidney Toxicity
Liver and kidney were among the organs affected when male and female Sprague-
Dawley rats were given s.c. injection of thallium (I) acetate as acute (single dose of 20-
50 mg/kg), subacute (2-3 weekly injections of 10-15 mg/kg), or chronic (10-20 mg/kg, followed
by weekly injections of 5 mg/kg or occasionally 2.5 mg/kg for up to 26 weeks) exposures
(Herman and Bensch, 1967). Toxicity was observed in all treatment groups. Animals dosed
acutely displayed the symptoms earlier than those on subacute or chronic dosing schedules.
Animals were sacrificed when signs of toxicity became apparent.
Acutely exposed animals had the following changes observed via light microscopy:
eosinophilic granular casts in 50-75% of the renal proximal and distal tubules, mild to moderate
enteritis, moderate to severe colitis, and dense infiltration of polymorphonuclear leukocytes and
lymphocytes that extended through all layers of the wall of the large intestine. Electron
microscopy revealed severe degenerative changes in the mitochondria of the renal tubular cells
and hepatocytes. In rats that received subacute injections of thallium acetate, eosinophilic
granular casts occurred in one-third of the proximal and distal tubules. Electron microscopy
revealed moderately prominent mitochondrial granules in the kidney, dense bodies in the
cytoplasm of the cells in the loops of Henle, and distal convoluted tubules. Mitochondrial
granules of hepatocytes were slightly enlarged and lacked electron-lucent cores.
In the chronically exposed rats, electron microscopy revealed increased size of
mitochondrial granules in the proximal convoluted tubules and an increase in cup-shaped
mitochondria in the distal convoluted tubules. Hepatocytes had increased numbers of large
complex residual bodies and lipid droplets, and the mitochondria were swollen with enlargement
of mitochondrial granules.
Yoshida et al. (1997) administered a single i.p. injection of thallium (I) sulfate
(25 mg/kg) to ICR mice (30-35 g). As was observed in Mourelle et al. (1988) after a single oral
dose of thallium (I) sulfate (10 mg/kg) to Wistar rats, liver Na+/K+-ATPase was statistically
significantly decreased by 12 hours. However, the rebound occurred by 24 hours instead of the
5 days observed in the Mourelle et al. (1988) study. While the Na+/K+-ATPase activity was
decreased, the ATP levels were increased and returned to control values by 12 hours post-dosing.
The effects were slightly different in the kidney. At 6 hours Na+/K+-ATPase activity was
statistically significantly decreased but had rebounded by 12 hours. The ATP levels were
significantly (p < 0.01) increased through 12 hours then decreased to levels significantly
(p < 0.01) lower than controls by 24 hours; ATP levels did not rebound until 240 hours after
thallium administration.
Male albino rats (210-260 g) receiving a single i.p. injection of 30, 60, or 120 mg/kg
thallium (I) sulfate had statistically significant (p < 0.05) increased levels of AST and ALT
above controls (0.5 mL of 0.9% saline) 16 hours after treatment, regardless of dose (Leung and
Ooi, 2000). There was a dose-dependent increase in ALT and AST between 30 and 60 mg/kg,
33
-------
but the levels did not increase from 60 to 120 mg/kg. Some 30 mg/kg animals exhibited
weakness, sluggishness, loss of hair, ptosis of the eyelids, diarrhea, and respiratory difficulty;
they were sacrificed 4 days post-dosing. The ALT and AST levels in these animals were still
elevated over the controls by a factor of 2-2.5. Serum creatinine levels also were elevated by
approximately 2.5 times over the control (0.5 mg/dL control versus 1.33 mg/dL treated).
Histologic evaluation confirmed damage to both the kidney and the liver. The damage in the
liver consisted of necrosis and swollen and vacuolated cells, which appeared to reduce the
sinusoidal space. Kidney tubules were atrophied and vacuolated with cell outlines less distinct
and cells containing many pyknotic nuclei. Amorphous material was apparent in the lumen of
the proximal tubules, and the brush borders were disorganized.
Appenroth et al. (1995) examined the effects of a single i.p. dose of thallium (I) sulfate
(5, 10, 15, or 20 mg/kg) on renal function and morphology in adult female Wistar rats. Low
doses (i.e., 5 and 10 mg/kg) of thallium (I) sulfate increased the volume of urine (measured on
day 2) but did not affect the protein level. Higher doses (i.e., 15 and 20 mg/kg) caused a
reduction in urinary volume along with an increase in the urinary protein concentration. The
glomerular filtration rate was statistically significantly (p < 0.05) reduced at 2 days after
treatment with 20 mg/kg but had returned to control levels by day 10. Blood urea nitrogen
(BUN) levels were significantly (p < 0.05) increased 2 days after treatment but had returned to
normal values by day 10. Histopathology showed a thickening of ascending limb of the loop of
Henle notable on day 2 after treatment and resolved by day 10. Na+/K+-ATPase activity was
significantly (p < 0.05) increased in the medulla on day 2 but was significantly (p < 0.05)
reduced by day 5. No changes in Na+/K+-ATPase activity were noted in the cortex.
In follow-up studies, Appenroth et al. (1996) and Fleck and Appenroth (1996) examined
the age-related nephrotoxicity of thallium (I) sulfate. Both studies demonstrated that
nephrotoxicity was more severe in the adult rat (Wistar) than in young rats (10 or 20 days old)
after a single i.p. injection of 20 mg/kg thallium (I) sulfate. Appenroth et al. (1996) determined
that there were several biochemical differences in kidney function between 10- and 20-day-old
rats (Wistar) administered 20 mg/kg thallium (I) sulfate, but there were no structural changes
indicating kidney damage in either group of young rats. In comparison, the thick ascending limb
of loop of Henle in 55-day-old rats showed damage. Thallium did not affect the glomerular
filtration rate in either 10- or 20-day-old rats but caused a significant reduction in the rate in
55-day-old rats (Appenroth et al., 1996; Fleck and Appenroth, 1996). Thallium caused a
statistically significant (p < 0.05) increase in the fractional excretion of a few amino acids (i.e.,
B-alanine, taurine, and 1-methylhistidine), a statistically significant (p < 0.05) decrease in the
fractional excretion of glycine in 10-day-old rats, and a statistically significant (p < 0.05)
increase in the fractional excretion of 13 amino acids in 55-day-old rats. Fleck and Appenroth
(1996) determined that thallium affects renal tubular amino acid resorption and causes kidney
damage only when mature kidney function is present.
34
-------
Woods and Fowler (1986) examined the effects of a single i.p. dose of thallium (III)
chloride (TlCb) at doses of 50, 100, or 200 mg/kg on liver structure and function in male
Sprague-Dawley rats (CD strain; 150-200 g) 16 hours after treatment. They determined a dose-
related effect on the volume density of mitochondria (increased), rough endoplasmic reticulum
(increased), lysosomes (increased), and cytoplasm (decreased). The surface densities of the
inner cristae of mitochondria and the rough endoplasmic reticulum also were determined to
increase in a dose-dependent manner. Statistically significant (p < 0.05) increases occurred in
monoamine oxidase (MAO) (100 and 200 mg/kg) and ferrochelatase (50, 100, and 200 mg/kg).
Statistically significant (p < 0.05) decreases occurred in aminolevulinic acid (ALA) synthetase
(50, 100, and 200 mg/kg), aminopyrine demethylase (200 mg/kg), aniline hydroxylase (50, 100,
and 200 mg/kg), and NADPH cytochrome c (P450) reductase (50, 100, and 200 mg/kg). In
addition, in vitro studies using 50, 100, or 200 ug/mL thallium (III) chloride demonstrated a
significant (p < 0.05) reduction in ALA synthetase, ferrochelatase, aniline hydroxylase, ALA
dehydratase, and NADPH cytochrome c (P450) reductase for all dose concentrations. A dose-
related loss of ribosomes from the smooth endoplasmic reticulum and proliferation of the rough
endoplasmic reticulum were observed through ultrastructural examination. Also observed were
generalized mitochondrial swelling and increased numbers of electron-dense autophagic
lysosomes.
Table 4-6 summarizes toxicity data from animal studies involving i.p. or i.v. injection.
35
-------
Table 4-6. Thallium toxicity in animals via injection
Reference
Species
Age or
weight
Sex
Route/exposure
period
Doses
NOAEL/LOAEL
Study type/effect
Acute studies (single dose)
All etal. (1990)
Appenroth et al.
(1995)
Barroso-Moguel
etal. (1990)
Barroso-Moguel
et al. (1992)
Barroso-Moguel
et al. (1996)
Kuperberg et al.
(1998)
Lameijer and
van Zwieten
(1976)
Rat
Rat
Rat
Rat
Rat
Rat
Rat
Adult
Adult
Newborn
Newborn
Newborn
250-
300 g
Young
adult
Male
Female
Both
Both
Both
Male
Male
i.p./single injection
i.p./single injection
i.p./single injection
i.p./single injection
i.p./single injection
i.p./single injection
i.v./single injection
20 mg/kg Tl as
thallium (I) acetate
5, 10, 15, and
20 mg/kg thallium
(I) sulfate
32 mg/kg thallium
(I) acetate
32 mg/kg thallium
(I) acetate
16 mg/kg thallium
(I) acetate
25 mg/kg thallium
(I) acetate
3-100 mg/kg
thallium (I) sulfate
LOAEL = 20 mg/kg
Tl
LOAEL = 12 mg/kg
Tl
LOAEL = 25 mg/kg
Tl
LOAEL = 25 mg/kg
Tl
LOAEL = 12 mg/kg
Tl
LOAEL = 19 mg/kg
Tl
LOAEL = 24 mg/kg
Tl (30 mg/kg thallium
(I) sulfate)
Neurological toxicity:
neurochemical changes in the
brain that were resolved within 24
hrs
Kidney toxicity: decreased urine
volume; increased urine protein;
thickened ascending limb of the
loop of Henle; changes in brain
Na+/K+-ATPase
Neurological toxicity: neuronal
and vascular damage in the brain
Developmental toxicity: diarrhea;
alopecia; cartilaginous and osseous
tissue alterations; profound
skeletal alterations
Developmental toxicity: diarrhea;
muscle atrophy; small size;
alopecia; death; interstitial edema
between myelin sheaths; thinner
muscle fibers; nerve fiber damage
Bladder and neurological toxicity:
distended bladder; reduced acetyl
cholinesterase activity in the brain
and bladder; increased choline
acetyltransferase activity in the
brain and bladder
Cardiotoxicity: hypertension
36
-------
Table 4-6. Thallium toxicity in animals via injection
Reference
Leung and Ooi
(2000)
Osorio-Rico et
al. (1995)
Woods and
Fowler (1986)
Species
Rat
Rat
Rat
Age or
weight
210-
260 g
200-
250 g
Young
adult
Sex
Male
Male
Male
Route/exposure
period
i.p./single injection
i.p./single injection
i.p./single injection
Doses
30, 60, 120 mg/kg
thallium (I) sulfate
30 or 50 mg/kg
thallium (I) acetate
50, 100, and 200
mg/kg thallium
(III) chloride
NOAEL/LOAEL
LOAEL = 24 mg/kg
Tl
LOAEL = 23 mg/kg
Tl
LOAEL = 42 mg/kg
Tl
Study type/effect
General toxicity: increased AST
and ALT; weakness; sluggishness;
alopecia; ptosis of the eyelids;
diarrhea; respiratory difficulty;
liver necrosis; kidney damage
Neurological toxicity: increase in
MAO and 5-hydroxytryptamine in
the brain
Liver toxicity: increased volume
density of mitochondria,
lysosomes, and rough endoplasmic
reticulum of the liver; decreased
cytoplasm in the liver; increased
MAO and ferrochelatase; changes
in several liver enzymes
Acute studies (3-10 doses)
Brown et al.
(1985)
Gibson and
Becker (1970)
Gibson and
Becker (1970)
Rat
Rat
Rat
250 g
Fetus; 8,
9, and 10
or 12, 13,
and 14
days of
gestation
Pregnant
Male
Both
Female
i.p./6 days
Transplacental via
i.p. injection to
dam/3 days
i.p./3 days
4 or 8 mg/kg-day
thallium (I) acetate
2.5 or 10 mg/kg-
day thallium (I)
sulfate
2.5 or 10 mg/kg-
day thallium (I)
sulfate
LOAEL = 3.1 mg/kg-
day Tl
LOAEL = 1.9 mg/kg-
day Tl
LOAEL = 1.9 mg/kg-
day Tl
Neurological toxicity: increased
lipid peroxidation in the brain;
increased p-galactosidase activity
in the brain; behavioral changes
Developmental toxicity: reduced
fetal body weight; increase in
hydronephrosis; increase in
missing or non-ossified vertebral
bodies
General toxicity: diarrhea;
lethargy; irritability; poor hair
luster; alopecia
37
-------
Table 4-6. Thallium toxicity in animals via injection
Reference
Hasan et al.
(1977)
Hasan et al.
(1978)
Hasan and All
(1981)
Hasan and
Haider (1989)
Kuperberg et al.
(1998)
Species
Rat
Rat
Rat
Rat
Rat
Age or
weight
-150 g
-150 g
-150 g
-150 g
250-
300 g
Sex
Male
Male
Male
Male
Male
Route/exposure
period
i.p./7 days
i.p./7 days
i.p./7 days
i.p./6 days
i.p./5 days
Doses
5 mg/kg thallium
(I) acetate
5 mg/kg thallium
(I) acetate
5 mg/kg thallium
(I) acetate
5 mg/kg thallium
(I) acetate
0,0.1, 1.0, or
5.0 mg/kg thallium
(I) acetate
NOAEL/LOAEL
LOAEL = 3.9mg/kg-
dayTl
LOAEL = 3.9mg/kg-
dayTl
LOAEL = 3.9mg/kg-
dayTl
LOAEL = 3.9mg/kg-
dayTl
LOAEL =
0.08 mg/kg-day Tl
Study type/effect
General toxicity: anorexia; poor
hair luster; diarrhea; difficulty
walking; abnormal head rotation;
lethargy; death; changes in Golgi
complexes and smooth cisternae/
vesicles of hypothalamic neurons;
decreased succinic dehydrogenase
and guanine deaminase activities
in the brain
Neurological toxicity: decreased
dopamine, norepinephrine, and
5-hydroxytryptamine in the brain
General toxicity: anorexia; poor
hair luster; diarrhea; difficulty
walking; abnormal head rotation;
lethargy; increased lipid
peroxidation; aggregation of
lipofuscin granules in the
perikarya of cerebellar neurons
Neurological toxicity: GSH
Neurological toxicity: difficulty
walking and maintaining pressure
on the hind paws; loss of
coordination in motor activity;
lethargy; reduced food
consumption; distended bladder;
decreased acetyl cholinesterase
activity in the bladder
Subchronic studies
Galvan-Arzate
et al. (2000)
Rat
200-
250 g
Male
i.p./30 days
0.8orl.6mg/kg-
day thallium (I)
acetate
LOAEL = 0.6 mg/kg-
day Tl
Neurological toxicity: increased
lipid peroxidation in the brain
38
-------
Table 4-6. Thallium toxicity in animals via injection
Reference
Kennedy and
Cavanagh
(1977)
Species
Cat
Age or
weight
Not
specified
Sex
Female
Route/exposure
period
s.c.74-26 weeks
Doses
2.3-4.5 mg/kg per
week thallous
acetate
(unclear if dose
was reported as
thallous acetate or
as thallium)
NOAEL/LOAEL
0.32 mg/kg-day
(unclear if dose was
reported as thallous
acetate or as thallium)
Study type/effect
Dying back of central and
peripheral sensory neurons; no
motor nerve fiber degeneration
39
-------
4.4.2. Cardiotoxicity
Male Wistar rats (180 to 220 g) injected i.v. with 3 to 100 mg/kg thallium (I) sulfate,
while under pentobarbital anesthesia, rapidly developed hypotension with the lowest blood
pressures reached within 3 to 5 minutes (Lameijer and van Zwieten, 1976). Blood pressures
dropped in a dose-dependent manner with doses of 30-100 mg/kg causing a drop of 20-40%
from initial values and a maximum effect achieved in the 50 to 100 mg/kg range. Thallium had a
greater effect on the diastolic pressure. After 10 minutes, animals injected with doses ranging
from 3-40 mg/kg had blood pressures resembling those prior to thallium injection. The higher
doses had a more permanent effect on blood pressure. In addition to lower blood pressure, the
rats had a dose-dependent decrease in heart rate with no maximum achieved. Rats treated with
100 mg/kg had heart rates that were one-third their pre-injection rates. The same effects were
observed in anesthetized cats injected i.v. with monovalent thallium but not when the thallium
was infused into the left vertebral artery (Lameijer and van Zwieten, 1976).
4.4.3. Neurotoxicity
No studies of thallium neurotoxicity following exposures by the inhalation or dermal
routes of exposure were identified. Manzo et al. (1983) demonstrated functional and
histopathologic changes in peripheral nerves in rats that received thallium sulfate in drinking
water at a dose equivalent to approximately 1.4 mg/kg-day Tl (see Section 4.2.1.1). Study
findings reported below involved administration of thallium compounds by injection (i.p., s.c.,
i.v.).
A single i.p. injection of 20 mg/kg thallium (I) acetate in 8-12 (exact number per group
not specified) adult male Sprague-Dawley rats (-300 g) resulted in a statistically significant
(p < 0.02) decrease in aspartate and taurine in the hippocampus 6 hours after treatment that was
resolved by 24 hours (Ali et al., 1990). Glutamine and taurine levels were statistically
significantly increased (p < 0.05) in the frontal cortex at 6 hours. While the glutamine returned
to control levels by 24 hours, the taurine was still significantly elevated. A dose-dependent
decrease in dopamine and muscarinic cholinergic receptor binding in caudate nucleus was not
reported in these treated rats but was observed in caudate nucleus incubated in vitro with
thallium. The study report indicated that the effect was not observed 24 hours after the last
subacute dose (5 mg/kg daily for 10 days, i.p.) of thallium (I) acetate in a separate study, but the
results were not presented. The subacute study demonstrated a statistically significant (p < 0.05)
increase in dopamine, 3,4-dihydroxyphenylacetic acid, and 5-hydroxytryptamine (5-HT
[serotonin]) levels in the amygdala nucleus, as well as an increase in 5-HT in the hypothalamus.
Dopamine, 3,4-dihydroxyphenylacetic acid, and 5-HT remained similar to control concentrations
in the caudate nucleus, frontal cortex, and hippocampus.
40
-------
Kennedy and Cavanagh (1977) examined the neurotoxic properties of thallous acetate in
16 female cats. Cats were injected subcutaneously with thallous acetate in distilled water once
per week (or in some cases every other week or every few weeks) at mean weekly doses of 2.3-
4.5 mg/kg per week for 4-26 weeks. (It was unclear if doses were reported as thallous acetate or
thallium). Six cats died during the study without opportunity for neurological assessment.
Remaining animals were observed daily for clinical signs. At intervals from 4-12 weeks after
exposure was initiated, nine animals were sacrificed and nervous system tissues were subject to
perfusion fixation; one cat was exposed weekly for 26 weeks. Ataxia and hypotonia were
observed in nearly all the cats. Histopathologic examination revealed a distal degeneration
confined to the central and peripheral axons of primary sensory neurons. No motor nerve fiber
degeneration was found. The investigators proposed that muscle tissue, which has the largest
reservoir of K+ ions in the body, may serve as a sink for thallium ions such that thallium ions
never reach a concentration in the extracellular space sufficient to damage nerve fibers.
Experimental evidence for this hypothesis, however, was not available.
A set of companion studies in newborn Wistar rats (Barroso-Moguel et al., 1996, 1990)
demonstrated severe and progressive lesions in nerve fibers following i.p. administration of
thallium (I) acetate. In the first study (Barroso-Moguel et al., 1990), 32 mg/kg thallium (I)
acetate was given to 15 newborn Wistar rats. Equal numbers were sacrificed at 24, 48, and
72 hours and on days 7 and 51 (three rats/time point). Alterations in the capillary vessel walls of
the brain, observed within the first hours after thallium injection, progressed to irregular
thickened walls and fibrotic sclerosis, which obstructed the lumen by 51 days postexposure.
Other changes in the brain began in a diffuse manner with all sections affected. Cortical neurons
developed the first and most intense lesions; by 51 days, the cortical neurons had mostly
disappeared and lesions were apparent in the central grey nuclei.
In a follow-up study (Barroso-Moguel et al., 1996), 16 mg/kg thallium (I) acetate was
administered via i.p. injection to 20 newborn Wistar rats (10 each sacrificed at 8 and 50 days of
age). General toxicity was manifested through diarrhea, progressive muscular atrophy, small
body size, and persistent alopecia. Eight of the 20 thallium-treated rats died during the study.
By 8 days of age, thallium-treated rats had interstitial edema between the myelin sheaths
(causing separation of the nerve fibers), edema around some axons and within the myelin, and, in
some cases, initial damage and degeneration of the myelin sheaths. In addition, muscle fibers
were thinner and showed signs of beginning progressive muscular atrophy. Hemorrhage,
necrosis, and destruction of the striation were present in some areas of the muscle. By 50 days of
age, the rats developed nerve damage with progressive disappearance of nerve fibers and
granular, filiform, and amorphous inclusions in abnormal axons and collapsing myelin sheaths.
At this time, the muscle fibers lost their transverse striation. Other muscle fibers were observed
41
-------
to be atrophic, fragmented, and exhibiting hyaline degeneration and initial fibroblast reaction;
further, some were infiltrated with phagocytic macrophages.
Osorio-Rico et al. (1995) measured MAO activity and 5-HT turnover rates in different
regions of the brain in 127 male Wistar rats (200-250 g) 24 hours after an i.p. administration of
30 or 50 mg/kg aqueous thallium (I) acetate. Results demonstrated MAO was significantly
increased (p < 0.05) at 30 mg/kg in the midbrain (27.7% over controls) and pons (37% over
controls) sections. MAO increases also were observed at 50 mg/kg in the midbrain (48% over
controls) and pons (47%). 5-HT turnover was significantly increased in the pons (172% over
controls;/* < 0.001) after 30 mg/kg treatment and in the pons (166.7% over controls;/? < 0.001)
and midbrain (56% over controls;/* < 0.01) after 50 mg/kg treatment. No significant changes
were observed in the dopamine turnover rate.
Subcutaneous injection of thallium (I) acetate as acute (single dose of 20-50 mg/kg),
subacute (2-3 weekly injections of 10-15 mg/kg), or chronic doses (10-20 mg/kg, followed by
weekly injections of 5 mg/kg or occasionally 2.5 mg/kg for up to 26 weeks) to male and female
Sprague-Dawley rats (250-500 g) caused toxicity in all treatment groups (Herman and Bensch,
1967). Symptoms reported in all groups included diarrhea, marked weight loss, anorexia, and
lethargy. Animals dosed acutely displayed the symptoms earlier than those on subacute or
chronic dosing schedules. Chronically exposed animals also had hair loss (maximal at 2-
4 weeks after initial injection), irritability, and dragging of the hind limbs. Animals were
sacrificed when signs of toxicity became apparent.
In acutely exposed animals, the mitochondria of the brain were frequently filled with an
overabundance of stacked mitochondrial cristae. Three of four animals administered thallium (I)
acetate subacutely had changes in the brain, including occasional foci of perivascular cuffing
with lymphocytes and hemosiderin-filled macrophage, acute necrosis, and swollen histiocyte-
like cells. Peripheral nerves had occasional dense bodies in an unmyelinated nerve plexus.
Brain neurons had numerous lipofuscin bodies as did the neurons of the chronically exposed
animals.
Twenty-four hours after a single dose of 25 mg/kg thallium (I) acetate, acetyl
cholinesterase (AchE) activity was reduced in the hypothalamus and nucleus accumbens (NA)
regions of the brain and the activity of choline acetyltransferase (ChAT) activity was
significantly (p < 0.05) increased (Kuperberg et al., 1998). At 48 hours, the AchE in the
hypothalamus was still significantly (p < 0.05) reduced, and the AchE in the NA region of the
brain was back to control levels. AchE activity also was reduced in the duodenum and the
sphincter-trigon region of the bladder following this single high dose, while ChAT activity was
significantly (p < 0.05) increased in the ileum, duodenum, and both regions of the bladder. After
48 hours, the AchE levels in the duodenum and the sphincter-trigon and detrusor regions of the
bladder were still significantly (p < 0.05) reduced.
42
-------
Adult male Sprague-Dawley rats (250-300 g) administered 0.1, 1.0, or 5.0 mg/kg
thallium (I) acetate (i.p.) daily for 5 days exhibited difficulty walking and maintaining pressure
on their hind paws, loss of coordination in motor activity, lethargy, and reduced food
consumption (Kuperberg et al., 1998). Most of the rats (6/8) treated at the high dose died by
48 hours posttreatment. The only significant changes observed 24 hours after the last dose were
a reduction of AchE levels in the NA region of the brain with the 1.0 mg/kg dose and an increase
in the AchE levels in the striatum and midbrain of rats treated with 5 mg/kg. There was a
decrease in AchE activity in the sphincter-trigon region of the bladder at 24 hours in all of the
repeat-dose groups (i.e., 0.1, 1.0, and 5.0), which returned to control values by 48 hours.
Decreased bladder AchE levels were also observed in the detrusor region of the bladder 24 hours
after 1.0 or 5.0 mg/kg doses and 48 hours after the 0.1 mg/kg dose. The bladders of these rats
were distended and contained twice the amount of urine seen in the controls.
Brown et al. (1985) examined lipid peroxidation in the brain after thallium exposure. In
this study, groups of male Sprague-Dawley rats (250 g) were administered daily i.p. injections of
4 or 8 mg/kg thallium (I) acetate in water for 6 days. Controls received saline. Twenty-four
hours after the final injection behavioral analysis was performed and the following morning, rats
were sacrificed. A dose-dependent increase in lipid peroxidation was observed in the
cerebellum, brain stem, and striatum but not in the midbrain and hippocampus. The 8 mg/kg
dose also caused a statistically significant (p < 0.05) increase in the lipid peroxidation of the
cortex. Beta-galactosidase activity followed a dose-dependent increase in the cerebellum, brain
stem, and cortex. The 8 mg/kg dose also caused a statistically significant (p < 0.05) increase in
beta-galactosidase activity in the midbrain and hippocampus. The beta-galactosidase activity in
the striatum was not statistically significantly changed by either dose. In general, thallium
decreased the frequency of grooming behavior while the frequency of exploratory and attention
behaviors was increased. However, the changes were not dose dependent.
Hasan et al. (1977) administered 5 mg/kg thallium (I) acetate i.p. for 7 days to albino
male rats (weighing approximately 150 g). Controls received sodium acetate solution in equal
volumes with the same molar concentration. Clinical symptoms included anorexia, failure to
gain weight, irritability, tenderness during handling, poor hair luster, lethargy, diarrhea, dragging
hind limbs, and fits of abnormal rotation of head and neck, curving the body. Eight of 55 treated
rats died by day 7. Electron microscopy showed an increased incidence of well-developed Golgi
complexes and curved conformation of smooth cisternae and vesicles of the neurons in the
hypothalamus. More significantly, there was a peculiar isolation of axonal endings of the
anterior hypothalamus by membranous circumferential lamellae, which appeared to have arisen
from the neighboring astrocytic processes. In addition, the succinic dehydrogenase and guanine
deaminase activities in the cerebrum were significantly decreased in thallium-treated rats.
Protein levels, MAO, adenosine triphosphate, and protease levels in the cerebrum were
43
-------
unaffected. Mitochondrial succinic dehydrogenase also was decreased in the cerebrum of
thallium-treated rats.
In a study by the same group of investigators (Hasan and Ali, 1981), male Charles Foster
rats (approximately 150 g) injected with 5 mg/kg thallium (I) acetate daily for 7 days showed
clinical symptoms similar to those reported by Hasan et al. (1977). All rats were lethargic after
4-5 days of treatment. Rats were sacrificed the day after the final dose (day 8), and their brains
were removed and separated into sections. A statistically significant (p < 0.001) increase in lipid
peroxidation was reported in the cerebral hemisphere, cerebellum, and brain stem by 49, 142,
and 116%, respectively. Electron microscopy demonstrated prominent aggregation of lipofuscin
granules in the perikarya cerebellar neurons (cell body of neurons in the brain) in thallium-
treated rats that were hardly discernible in control rats. Comparisons made with nickel- and
cobalt-treated rats demonstrated differences in the areas of increased lipid peroxidation. Nickel
and cobalt both had the greatest impact on the brain stem, whereas thallium had a greater effect
on the cerebellum. Although Hasan and Ali (1981) could not relate this to the differences in
behavioral observations, they did note that nickel- and cobalt-treated rats were irritable and
restless; these symptoms were not observed in the thallium-treated rats.
Hasan and colleagues also examined the effects of thallium (I) acetate on
neurotransmitter levels and sulfhydryl groups in the brain. Charles Foster rats (approximately
150 g) were administered 5 mg/kg thallium (I) acetate i.p. for 6 days (Hasan and Haider, 1989)
or 7 days (Hasan et al., 1978). Dopamine, norepinephrine, and 5-HT were reduced in the four
sections of the brain examined (hypothalamus and limbic area, corpus striatum, cerebellum, and
brain stem), but not all reductions were statistically significant. Dopamine was significantly
reduced in the hypothalamus and limbic area (50%; p < 0.05) and corpus striatum (64%;
p < 0.01). Norepinephrine was reduced by 9-33%, depending on the brain region, but none of
these reductions were statistically significant. 5-HT was significantly reduced in the corpus
striatum (53%; p < 0.001), cerebellum (36%; p < 0.05), and brain stem (66%; p< 0.001) (Hasan
et al., 1978). Glutathione was significantly (p < 0.001) reduced in the cerebrum (56%),
cerebellum (62%), and brain stem (74%), and sulfhydryl radicals were significantly (p < 0.05)
reduced in the cerebellum (25%) and brain stem (32%) (Hasan and Haider, 1989).
Galvan-Arzate et al. (2000) administered 0.8 mg/kg (considered 1/40 of the median lethal
dose [LDso]) or 1.6 mg/kg (considered 1/20 of the LDso) thallium (I) acetate in deionized water
via i.p. injection for 30 days to male Wistar rats (200-250 g). Three days after treatments ended,
rats were sacrificed and their brains were dissected into five different regions (hypothalamus,
cerebellum, frontal cortex, hippocampus, and corpus striatum). In each region, with the
exception of the cerebellum, significant (p < 0.01) increases in thallium content were observed
after administration of 1.6 mg/kg compared to administration of 0.8 mg/kg. There were no
statistically significant differences in the deposition of thallium within each region of the brain
44
-------
for each dose. The rate of lipid peroxidation, a marker of oxidative stress, was increased
significantly (p < 0.01) in the corpus striatum (182%) and cerebellum (130%) after treatment
with 0.8 mg/kg. At 1.6 mg/kg, all five regions exhibited statistically significant increases in lipid
peroxidation over controls (corpus striatum, 161% increase,/* < 0.05; hippocampus, 114%
increase,/? < 0.01; hypothalamus, 100% increase,p < 0.01; cerebellum, 81% increase,/? < 0.01;
and frontal cortex, 80% increase,/? < 0.05). The two regions affected at 0.8 mg/kg (i.e., corpus
striatum and cerebellum) were not affected to a greater extent at the higher dose. Lipid
peroxidation was measured to determine if oxidative stress plays a role in thallium's toxicity.
The study authors concluded that additional studies need to be performed to establish the precise
mechanism of neurotoxicity.
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
ACTION
Several, possibly related, mechanisms have been postulated for the toxic action of
thallium; however, the exact mechanism or mechanisms of toxicity are unknown.
4.5.1. Interference with Potassium Transport
Monovalent thallium is similar to potassium (K+) in ionic radius and electrical charge,
which may contribute to its toxic properties.
Monovalent thallium has been demonstrated to have a 10-fold higher affinity for rabbit
kidney Na+/K+-ATPase than does potassium (Britten and Blank, 1968). Barrera and Gomez-
Puyou (1975) reported that monovalent thallium also inhibits the influx and efflux of potassium
in rat liver mitochondria at concentrations (10-15 nmol bound monovalent thallium per mg of
mitochondrial protein) that do not affect oxidative phosphorylation. This inhibitory effect of
thallium seemed to be specific for potassium since it did not affect the movement of sodium.
Monovalent thallium was completely equilibrated by human red blood cells in 30 minutes
in a high (140.5 mM)-sodium (Na+) medium (Cavieres and Ellory, 1974) but was equilibrated
even faster when the medium contained a low concentration (5 mM Na+). In the high-Na+
medium containing 1 mM external potassium, monovalent thallium caused a dose-dependent
decrease in the ouabain-sensitive potassium influx. Monovalent thallium had a different effect in
a medium containing 0.17 mM potassium; low (0.2 mM or less) monovalent thallium ion
concentrations stimulated the ouabain-sensitive potassium influx but inhibited it at higher
concentrations. Monovalent thallium also had an inhibitory effect on the ouabain-sensitive
sodium efflux. It was suggested that the effects on ouabain-sensitive sodium efflux and
potassium influx are related to thallium's high-affinity substitution of potassium at the external
potassium sites of the sodium pump, which is actively transporting monovalent thallium ions in
while pumping sodium ions out.
45
-------
Tao et al. (2008) reported that thallium (I) could functionally replace potassium ion when
applied in an in vitro system and that glutamate transporters could interact with thallium (I).
4.5.2. Disturbance of Mitochondrial Function and Energy Generation
Thallium may exert toxicity by disturbing mitochondrial function. Thallium (I) acetate
caused an uncoupling of oxidative phosphorylation and swelling of isolated mitochondria and
induced an increase in oxygen consumption and lactic acid production in ascites tumor cells in
vitro (TPCS, 1996).
There is evidence that thallium reduces the available energy within peripheral nerves
(resulting in a "dying back" type of neuronal degeneration) and other metabolically highly
demanding cells (Kazantzis, 2007; Cavanagh, 1991). Cavanagh (1991) proposed that thallium in
tissues produces a tissue deficiency of available riboflavin, leading to disturbances of metabolic
reactions dependent on flavoproteins. Among these reactions would be steps in the passage of a
number of important intermediate metabolites in the electron transport chain. This metabolic
disruption could lead to substantial impairment of energy production in the cells. Cavanagh
(1991) noted that support for this hypothesis includes the following: (1) riboflavin deficiency in
the rat will lead to peripheral neuropathy and in the primate to loss of hair with circumoral skin
lesions, and thallium toxicity in the primate produces a similar clinical picture to riboflavin
deficiency; (2) the testis, which is peculiarly dependent on glucose as a substrate for energy
metabolism, is sensitive to thallium toxicity; and (3) heart muscle, also known for high energy
utilization, is also damaged as a result of chronic thiamine deficiency.
4.5.3. Induction of Oxidative Stress
Other research suggests that thallium may trigger toxicity through induction of oxidative
stress. Hanzel et al. (2005) investigated effects of thallium (III) hydroxide on metabolism of
GSH, which plays a key role in the regulation of cell redox state, in an in vitro system using rat
brain cytosolic fractions. Thallium hydroxide decreased the content of GSH and inhibited
glutathione peroxidase and glutathione reductase activity, suggesting that thallium impairs the
glutathione-dependent antioxidant defense system. Using rat pheochromocytoma (P12) cells
in vitro incubated with both thallium (I) nitrate and thallium (III) nitrate, Hanzel and Verstraeten
(2006) found a concentration- and time-dependent decrease in cell viability, decreased
mitochondrial membrane potential, increased steady-state levels of mitochondrial hydrogen
peroxide (H2O2, a product of partial reduction of molecular oxygen whose generation is
enhanced when electron transport is impaired), and GSH content. These investigators postulated
that both ionic species of thallium enhance reactive oxygen species production in the cell,
decreasing mitochondrial functionality and cell viability. In a follow-up study with P12 cells in
vitro, Hanzel and Verstraeten (2009) observed that thallium (I) and (III) do not cause cell
46
-------
necrosis, but significantly increased the number of cells with apoptotic features. The oxidatation
state (i.e., I vs III) appeared to influence the apoptotic pathways are involved. Thallium (I)-
mediated cell apoptosis was mainly associated with mitochondrial damage, whereas thallium
(III) showed a mixed effect triggering both the intrinsic and extrinsic pathways of apoptosis.
Galvan-Arzate et al. (2005) investigated the effects of a single dose (8 or 16 mg/kg, i.p.)
of thallium acetate on lipid peroxidation in different brain regions of Wistar rats (as an indicator
of oxidative damage) and alterations in endogenous antioxidant systems. Lipid peroxidation was
increased in three of five brain regions from animals that received 16 mg/kg thallium acetate at
day 7 postexposure (but not at days 1 and 3); antioxidants GSH and superoxide dismutase
showed only a modest depletion in only one or two brain regions.
4.5.4. Reaction with Thiol Groups
The capacity of thallium to react with thiol groups, thereby interfering with a variety of
processes, is postulated as another mechanism of toxicity, although interference with the
metabolism of sulfur-containing amino acids does not seem to be directly involved in toxicity
(IPCS, 1996). Thallium (I) chloride formed complexes with a number of sulfur-containing
amino acids (i.e., L-cysteine, DL-penicillamine, N-acetyl-L-cysteine, and N-acetyl-DL-
penicillamine) in aqueous solution (Bugarin et al., 1989). Because the thallium (I) complexes
formed were weaker than those formed using dimethylthallium (III), the study authors concluded
that this was unlikely to be the major mechanism of toxicity.
4.5.5. Other Endpoint-specific Mechanistic Data
4.5.5.1. Cardiotoxicity
Monovalent thallium ions caused a dose-dependent decrease in the heart rate and
contractile force of spontaneously beating atria of guinea pigs (Lameijer and van Zwieten, 1976).
At a concentration of 10~3 M, monovalent thallium reduced the heart rate by approximately 60%.
When the isolated atria were electrically driven, monovalent thallium ions at doses up to 10~3 M
were not able to significantly decrease the amplitude of contraction. Neither potassium (ranging
from 2.4 to 9.4 mM) nor cocaine (10"6 or 10~5 M) were able to influence the reduction in heart
rate or contractile force of spontaneously beating guinea pig atria caused by monovalent thallium
ions (0.5 mM) (Lameijer and van Zwieten, 1976).
Isolated rat heart (from albino rats of both sexes) perfused with a nitrate-Krebs solution
containing monovalent thallium ions (as thallium (I) nitrate) in place of potassium had a rapid
decrease in beat frequency (Hughes et al., 1978). The heart stopped completely in an average of
7 minutes. Placing the hearts in normal or nitrate-Krebs saline allowed for some recovery in
approximately 10 minutes. The monovalent thallium ions were still present in the heart tissue
(9 mmol Tl+/kg wet tissue) after 30 minutes in thallium-free solution; this residual thallium
47
-------
caused the heartbeat frequency and amplitude to remain low. When only a portion of the
potassium was replaced with monovalent thallium ions, there was a concentration ratio and time
dependency on the reduction of the heart rate. In a separate experiment, it was noted that
injecting thallium (I) nitrate into the perfusion stream close to the heart caused an initial
acceleration of the heartbeat followed by a reduction in amplitude until at certain concentrations
the heart stopped. The heartbeat could recover somewhat following nitrate-Krebs perfusion, but
the heartbeat frequency and amplitude did not return to initial values and were instead
comparable to those values noted prior to the heart stopping. There was a dose-dependent effect
on the length of the cardiac paralysis (5-25 umol monovalent thallium ions injected). The use of
potassium nitrate instead of thallium (I) nitrate also caused a dose-dependent increase in the
length of cardiac paralysis, but the length of paralysis was shorter than that with thallium (I)
nitrate. Furthermore, potassium nitrate-treated hearts recovered completely when washed with
nitrate-Krebs solution (Hughes et al., 1978).
4.5.5.2. Neurotoxicity
Wiegand et al. (1984) recorded the frequencies (reflects presynaptic processes) and
amplitudes (reflects postsynaptic processes) of miniature endplate potentials (MEPPs) from
neuromuscular junctions of rat (strain not specified) phrenic nerve (of the diaphragm)
preparations. Investigators reported a gradual increase in the frequency of MEPPs by a factor of
10 within 30 and 180 minutes at concentrations of 1 x 10~3 and 5 x 10~4 M thallium (I) acetate,
respectively, which was reversible. The amplitude was unchanged. Therefore, it was concluded
that thallium interfered presynaptically with spontaneous transmitter release. A follow-up
experiment using triangularis sterni muscles of adult mice (strain not specified) demonstrated
that, although thallium disturbed the presynaptic transmission in a manner similar to divalent
metal cations, thallium (monovalent; compound used not specified) acted via a different
mechanism than either divalent cobalt or cadmium (Wiegand et al., 1990). This study also
demonstrated that thallium did not influence the presynaptic potassium or calcium channels.
Windeback (1986) used an in vitro model system to compare the inhibitory effects of four
metals on neurite outgrowth. Metals were added to cultures of El 5 rat embryo dorsal root
ganglion neurons; neurite outgrowth was measured during the linear phase of growth after
24 and 40-80 hours. Compared to mercury and arsenic, which inhibited neurite outgrowth at
very low concentrations (50% inhibition at 3.9 and 9.6 x 10"6 M, respectively), neurons were
relatively resistant to thallium (50% inhibition at 1.3 x 10^ M), suggesting that populations of
neurons have different susceptibilities to various metals. The basis for these different
sensitivities was not explored in this study.
Hippocampal slices from adult guinea pigs or rats (strain not specified) were used to
examine the effects of thallium on central neuronal activity (Lohmann et al., 1989). The study
48
-------
authors did not note any differences between the guinea pig and rat results and appear to have
combined the results. Light microscopy did not show any morphologic changes in thallium-
treated hippocampal slices even with a high concentration (1-1.2 mM) for 6 hours. Thallium
was determined to reversibly reduce the amplitudes of the compound action potential of CA1
pyramidal cells in a dose- and time-related manner. Thallium did not alter intracellular response
parameters, indicating that membrane potential and input resistance were not affected, but
postsynaptic potentials were inhibited. The study authors concluded that in the hippocampal
slices thallium reacts mainly with postsynaptic target sites and exerts an unknown influence on
intracellular metabolism of CA1 pyramidal cells (Lohmann and Wiegand, 1996; Lohmann et al.,
1989).
Diaphragms from male and female albino rats perfused with a nitrate-Krebs solution
containing monovalent thallium (as thallium (I) nitrate) in place of potassium (K+) had an initial
(1-2 minute) increase in the contraction amplitude, followed by a steady decline in response
(Hughes et al., 1978). Four experiments were performed with the sequence taking 20-
40 minutes to block indirect (nerve) stimulation and 70-100 minutes to block direct (muscle)
stimulation. After returning the diaphragms to normal nitrate-saline, the block in response was
reversed, but contraction amplitudes were 30% of the original response for both nerve and
muscle after a 75-minute perfusion. The presence of monovalent thallium in the solution,
whether as a replacement or an addition to potassium, caused a dose dependency to the response.
Diaz and Monreal (1994) examined the effects of thallium compounds (thallium (III)
chloride, thallium (III) nitrate, and thallium (I) acetate) on proton and chloride permeabilities
through myelin lipid bilayers using an in vitro system of liposomes prepared with lipids from
brain myelin. Trivalent thallium, but not monovalent thallium, mediated a rapid
chloride/hydroxyl ion exchange through the lipid bilayers. Trivalent thallium in the presence of
reducing agents did not have the same reaction. The ion exchange was faster with trivalent
thallium than with mercury (Hg+2). In addition, the reaction occurred with a 10-fold lower
concentration of trivalent thallium than of mercury.
4.5.5.3. Dermal Toxicity
Arbiser et al. (1997) examined the effects of thallium acetate on three types of skin cells,
human keratinocytes, primary endothelial cells, and melanoma cells, to determine whether
thallium affected cell growth and differentiation in vitro. Inhibition of proliferation of all three
cell types was observed. In melanoma cells, thallium caused dose-dependent decreases in cell
dendricity and shape but not cellular motility. In normal human keratinocytes, thallium appeared
to interfere with the normal program of cutaneous keratinization. In an in vivo study by the same
investigators using piebald LPJ mice with both melanin-rich and -poor areas in the same animal,
one week administration of thallium acetate (5 mg/kg daily) by i.p. injection produced evidence
49
-------
of lipid peroxidation in skin in a perifollicular distribution (Arbiser et al., 1997). (The
investigators noted that lipid peroxidation in vivo results in oxidation of lipid membranes,
resulting in increased concentrations of aldehydes, which can react with the Schiff reagent,
thereby producing a colored product.) It was suggested that lipid peroxidation may result in cell
death due to membrane damage and may partly account for thallium-induced alopecia.
4.5.6. Genotoxicity
Positive results were obtained for thallium (I) nitrate (1 mM) in the recombination-repair
(Rec) assay using Bacillus subtilis strains H17 and M45; whether or not hepatic homogenates
were used was not specified (Kada et al, 1980; Kanematsu et al., 1980). These positive results
were obtained using "cold incubation," which increases the sensitivity of the assay by 20-
50 times for many drugs. In this test, plates containing the bacteria with a 10 mm filter paper
disk containing the metal solution (0.05 mL) were incubated at 4°C for 24 hours prior to being
incubated at 37°C overnight. The differences in the inhibition of growth between the Rec+ strain
and the Rec" strain were measured. Thallium (I) nitrate was not mutagenic in reverse mutation
assays using Salmonella typhimurium strains TA98, TA100, TA1535, TA1537, and TA1538
(histidine reversions) and Escherichia coli strains B/r WP2 tr~and WP2 her" tr~; whether or not
hepatic homogenates were used was not specified (Kanematsu et al., 1980). Negative results
were obtained in a screening assay for the induction of mitogenic gene conversion and reverse
mutation in the yeast Saccharomyces cerevisiae at a 0.1 M concentration of thallium (I) nitrate
(Singh, 1983).
Thallium (I) nitrate did not affect cell division in S. cerevisiae (isolated from baker's
yeast) and E. coli (strain B) but proved toxic to the aerobic growth processes of S. cerevisiae
(Loveless et al., 1954). A dose of 250 ug/mL thallium (I) nitrate caused a 50% reduction in
aerobic growth processes. The report did not identify specific doses of thallium (I) nitrate tested.
The organisms were treated under conditions of logarithmic phase growth. S. cerevisiae was
incubated for 4 hours with thallium (I) nitrate, and E. coli was incubated for 1.5 hours. The
study authors noted that several of the other compounds tested (e.g., iodoacetamide) that were
specific inhibitors of sulfhydryl groups also reduced the growth processes with no effect on
cellular division.
A concentration of 1,000 ug/mL thallium (I) acetate reduced viability of Chinese hamster
ovary (CHO) cells in culture to 20% with a concomitant decrease in DNA synthesis (i.e., 1% of
control values) (Garrett and Lewtas, 1983). The ECso (effective concentration necessary to
produce a 50% response) values were 307 ug/mL for viability and 18 ug/mL for DNA synthesis.
Thallium acetate also depressed ATP and protein synthesis in culture.
Single-strand DNA breaks occurred in cell cultures of C57BL/6 mouse and rat embryo
fibroblasts exposed to thallium (I) carbonate at both concentrations tested in mouse fibroblasts
50
-------
(i.e., 10 5 and 10 4M) and all three concentrations tested in rat fibroblasts (i.e., 10"6, 10 5, and
10~4M) (Zasukhina et al., 1983). However, thallium (I) carbonate did not induce single-strand
DNA breaks in CBA mouse fibroblasts after treatment with 10~4-10~6M concentrations.
Zasukhina et al. (1983) performed a dominant lethal test on male white rats that received
daily oral doses of thallium (I) carbonate (0.005-0.5 ug/kg-day) for 8 months and were
subsequently mated with untreated females. Female rats were sacrificed on day 20, and
mutagenic potential was evaluated based on evidence of embryotoxicity. The investigators
reported an increase in embryonic death, suggestive of a dominant lethal effect. As reported
previously (Section 4.3), however, confidence in this study is low. The number of resorptions
was highest in the control group. In addition, methods were not adequately reported and results
were not analyzed statistically.
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS
4.6.1. Oral
Thallium is readily absorbed through the GI tract and distributed throughout the organs
and tissues of the body. Although thallium is not metabolized, it occurs in two valence states. If
or how the body modifies the valence state of thallium is unknown, but orally administered
monovalent thallium and trivalent thallium appear to be distributed in a similar manner
throughout the body (Sabbioni et al., 1980a, b). Once thallium is distributed, elimination occurs
mainly in the urine and feces with the amounts in each varying by species.
Most of the available human case reports are the results of poisonings, suicide attempts,
or accidental ingestion of rodenticides. The lowest known dose to cause symptoms is a single
dose of 0.31 g; the patient recovered after treatment (Cavanagh et al., 1974). The only studies of
repeated oral exposure to thallium were two surveys of populations exposed to thallium through
contaminated homegrown foods (Dolgner et al., 1983; Brockhaus et al., 1981). Limitations in
these epidemiology studies included the lack of objective tests for toxicity, reliance on the
incidence of symptoms obtained from questionnaires, and characterization of chronic thallium
exposure by measuring the levels in urine and hair at a single point in time. Three studies of
occupationally exposed populations (Ludolph et al., 1986; Marcus, 1985; Schaller et al., 1980)
did not conclusively establish an association between thallium exposure and impaired health
status; however, all three studies were limited in terms of size of the study population and study
design.
Symptoms of thallium toxicity are diverse in both humans and animals. The triad of
gastroenteritis, polyneuropathy, and alopecia has been regarded as the classic syndrome of
thallium poisoning (IPCS, 1996), although not all three of these effects are observed in all
poisoning cases, and other symptoms develop in varying sequence depending on the magnitude
and duration of thallium exposure.
51
-------
The nervous system as a target organ of thallium is supported by observations from
human case reports and animal studies. Relatively high doses of thallium cause neurological
symptoms in humans (e.g., paresthesia of the hands and feet, weakness, tremors, coma, and
convulsions) (see Table 4-1). Some of these neurological symptoms (e.g., paresthesia and
weakness) were reversible, although recovery was slow. Other effects, including mental and/or
psychological problems, were more persistent. Neurological symptoms have also been
associated with chronic exposure to thallium in humans. These symptoms include sleep
disorders, tiredness, weakness, nervousness, headache, psychological alterations, and
neurological and muscular problems.
Although paresthesia of the hands and feet are trademark symptoms of thallium toxicity,
it is generally alopecia that leads to a diagnosis of thallium poisoning in humans. Alopecia
occurs about 2 weeks after exposure and is reversible after exposure to thallium is discontinued.
Alopecia has also been repeatedly observed in experimental animals exposed to thallium
compounds.
Thallium exposure in humans has been associated with respiratory effects and GI effects,
including diarrhea and vomiting. Other toxic effects associated with oral thallium exposure in
humans and animals are changes in blood pressure (high, low, and fluctuating values have all
been noted) and liver and kidney damage (kidney damage is age dependent and occurs only in
mature kidneys), all of which appear to be reversible with the removal of thallium exposure.
Doses that do not affect survival have been shown to affect clinical chemistry parameters such as
ALT, AST, BUN, blood glucose, and blood sodium levels, indicating liver and kidney damage
with subchronic exposures (Leung and Ooi, 2000; Appenroth et al., 1996, 1995; Fleck and
Appenroth, 1996; El-Garawany et al., 1990; Mourelle et al., 1988).
In experimental animal studies, thallium exposure has been associated with biochemical
changes, lipid peroxidation, and histopathologic changes in the brain and functional and
histopathologic changes in peripheral nerves (see Section 4.4.3). The areas affected in the brain
differ with the age of the treated animal; nevertheless, all measured endpoints (symptoms,
biochemical measurements, and histopathology) indicate that high doses (close to lethal doses)
of thallium induce significant degradation of the nervous system. Results from in vitro studies
further confirm these observations.
Despite the fact that the nervous system is a known target of thallium toxicity, studies
using standard measures of neurobehavioral toxicity have not been performed. In a 90-day oral
gavage study, MRI (1988) observed rats daily for clinical signs. Rats exposed to thallium sulfate
showed consistently increased incidences of clinical observations related to the coat (rough coat,
piloerection, shedding, and alopecia), eyes (lacrimation, exophthalmos, and miosis), and
behavioral signs compared with untreated or vehicle controls. The underlying mode of action for
these clinical observations is not known. Collectively, however, these observations suggest a
52
-------
treatment-related effect of thallium on the rat and possibly an indirect measure of stress or other
effects on the nervous system. For example, it has been suggested that barbering (or
overgrooming) in rodents may represent a stress-evoked behavioral response or other nervous
system dysfunction (Welch et al., 2007; Kalueff et al., 2006; Kalueff and Tuohimaa, 2005; Greer
and Capecchi, 2002).
Thallium salts have been shown to affect reproductive function. A dose as low as
0.7 mg/kg-day Tl (10 ppm of thallium (I) sulfate) resulted in testicular damage and reduced
sperm motility in male Wistar rats within 60 days (Formigli et al., 1986). Wei (1987) reported
that doses as low as 0.001 mg/L in the drinking water for 6 months in Kunming mice reduced
sperm motility (rapid speed only), and 0.01 mg/L reduced overall sperm motility and sperm
counts and caused a reduction of live offspring. However, confidence in this study is low due to
the non-reporting of several observations (water intake, body weights, behavior, and animal
health). In addition, the numbers of animals examined for sperm and reproductive endpoints
were ambiguously reported. Half of the animals were sacrificed at the end of the exposure
period, while the other half were allowed to mate for 1 week and then sacrificed. The study
results for male mice used for analysis of sperm and reproductive endpoints were reported
together. That is, results for sperm count, sperm mobility, and all other epididymal sperm
analysis for males sacrificed at the end of the exposure period were combined with the results for
males that were allowed to mate and thus permitted a one-week recovery period prior to
sacrifice. In addition, at the initiation of the study each group consisted of 20 animals. In the
reporting of the results, however, sperm data were missing for 4 mice from each of the control,
0.001, 0.01, and 1 mg/L groups, for 8 mice from the 0.1 mg/L group, and for 10 mice from the
10 mg/L group. No explanation for the loss of male mice (ranging from 20 to 50%) is provided.
Limited data in humans and experimental animals suggest that thallium may produce
developmental toxicity. A review of case studies of women exposed orally to high levels
(approximately 120-1,100 mg) of thallium during pregnancy suggested a trend toward premature
and low-birth-weight infants, especially if exposure took place in the last trimester; no other
developmental abnormalities were identified (Hoffman, 2000). Dolgner et al. (1983) examined
birth defects in a German population living near a cement plant emitting thallium dusts during
the mid-1970s and found a higher incidence of congenital malformations than the incidence
documented in the government birth records from the area. The association between the number
of birth defects and thallium exposure was weak, however, because two of the malformations
were considered hereditary and the incidence of birth defects, although greater than that
determined from civil records, was consistent with that reported in the literature. Confidence in
this study was limited by lack of exposure data during pregnancy and possible underreporting in
controls. In vivo data in rats support an association between i.p. thallium exposure and low birth
53
-------
weight, although such an association has not been reported with orally administered thallium.
In vitro data demonstrated an increase in bone malformations in both rat and chick embryos.
4.6.2. Inhalation
There are currently no studies that examine the effects of inhaled thallium. A few case
reports (Hirata et al., 1998; Ludolph et al., 1986) suggest an association between occupational
exposure and toxicity (including alopecia, GI symptoms, and neuropathy), but the route or routes
of exposure in these workplace setting could not be established. A study of a population living
near a cement factory emitting thallium (Dolgner et al., 1983) determined that thallium exposure
occurred via consumption of plants grown in thallium-contaminated soil to a greater extent than
via inhalation.
4.6.3. Mode-of-Action Information
The underlying mode of action for clinically observed effects of thallium has been
studied. Verstraeten (2006) determined that T11+ can oxidize membrane fatty acids, causing an
increased fluidity in the membrane and an increased concentration of cellular oxidants associated
with thallium toxicity. Ensuing axonal degeneration characterized by a decrease in density of
large myelinated fibers and loss of epidermal nerves, indicating small sensory nerve involvement
in thallium toxicity, has been reported (Kuo et al., 2005). The skin experiences parakeratosis and
vacuolar degeneration of the basal layer, resulting in a loss of epidermal nerves and persistent
damage to sensory nerve endings (Lu et al., 2007). Collectively, these observations suggest a
treatment-related effect of thallium on the rats and direct effects on the nervous system.
Both potassium and thallium are monovalent cations with similar atomic radii (Tl+:
1.50 A; K+: 1.38 A) (Ibrahim et al., 2006). Thallium has been shown to replace potassium in the
reaction of Na+/K+-ATPase (Barrera and Gomez-Puyou, 1975; Britten and Blank, 1968) and to
mimic the biological actions of potassium. Monovalent thallium has been shown to have a
10-fold higher affinity than potassium for Na+/K+-ATPase and thus replaces potassium as a
substrate for this enzyme (Barrera and Gomez-Puyou, 1975; Britten and Blank, 1968). In other
studies it caused a decrease in Na+/K+-ATPase activity in the liver and kidney (Yoshida et al.,
1997; Mourelle et al., 1988). After a single oral or i.p. dose of thallium (10 mg/kg orally or
25 mg/kg i.p.), the disruption of Na+/K+-ATPase activity was found to be reversible.
In addition to effects related to interference with potassium transport and function, there
is evidence that thallium uncouples oxidative phosphorylation, adversely affects protein
synthesis, reduces the available energy within peripheral nerve axons, and inhibits a number of
enzymes, including alkaline phosphatase and succinic dehydrogenase (Kazantzis, 2007).
Thallium's activity as a Lewis acid with an affinity for organosulfur compounds (Lewis
bases) may account for its adverse effect on hair production. Keratin is the primary protein
54
-------
found in hair. It is rich in the amino acid cysteine and its low solubility is, in large part, the
product of the formation of inter-polypeptide cysteine-cysteine cross-links during
posttranslational modification of the nascent polypeptides. Thallium prevents keratinization of
hair proteins by binding with cysteine and preventing the formation of the cross-linking bonds
(Mulkey and Oehme, 1993), a property that may be related to the alopecia observed in humans
and animals following thallium exposure. Binding to cysteine may also account for inhibition of
enzymes with active site cysteine residues and increases oxidative stress as a result of GSH
modification (Mulkey and Oehme, 1993).
4.7. EVALUATION OF CARCINOGENICITY
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), there is
"inadequate information to assess the carcinogenic potential" for thallium and thallium
compounds. There are presently no studies that evaluate the carcinogenic potential of thallium in
animals and no adequate studies of humans chronically exposed to thallium.
Two studies of chronic health effects in workers exposed to thallium are available
(Marcus, 1985; Schaller et al., 1980), but these studies are inadequate for the assessment of
carcinogenicity. The study by Marcus (1985) is limited by the examination of medical records
only, lack of exposure quantitation, small cohort size, and the unknown length of observation.
Schaller et al. (1980) identified health effects in a worker population at a single time point
through medical histories and physical examinations for unspecified symptoms. Worker
exposures to thallium were limited to a single measure of urinary thallium, which would not
provide an adequate measure of past exposure. This health evaluation was not adequate to detect
any carcinogenic response.
Relatively few studies have examined the genotoxicity of thallium compounds; these
studies provide inconsistent evidence for genotoxicity. Positive results were obtained at 0.001M
for thallium (I) nitrate in the Rec assay using B. subtilis strains H17 and M45 (Kanematsu et al.,
1980). However, negative results were obtained in reverse mutation assays using several
S. typhimurium and E. coli strains and mitogenic gene conversion and reverse mutation tests in
yeast. Cytotoxic levels (1,000 ug/mL) of thallium (I) acetate caused depressed DNA synthesis in
CHO cells (Garrett and Lewtas, 1983). Single-strand DNA breaks occurred in C57B1/6 mouse
and rat embryo fibroblasts exposed to thallium carbonate but not in similarly exposed CBA
mouse fibroblasts (Zasukhina et al., 1983). A dose of 200 mg thallium sulfate caused a slight
increase in SCEs in peripheral blood lymphocytes taken from a 48-year-old man on day 1 and
day 15 postexposure and caused a 3.5-fold increase in binucleated cells with micronuclei
(Hantson et al., 1997).
55
-------
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
There are little or no data available to establish a particular subpopulation as being
particularly susceptible to the toxic effects of thallium compounds.
4.8.1. Possible Childhood Susceptibility
Exposures in human case reports generally are poorly characterized. Therefore, no
comparison can be made between children and adults with regard to susceptibility. In rats, doses
of thallium that caused maternal toxicity have been demonstrated to affect the developing fetus
(Gibson and Becker, 1970). The only studies that examined the toxic effects of thallium at
different ages were those of Appenroth et al. (1996) and Fleck and Appenroth (1996), which
focused on the age-related effects on nephrotoxicity. In these studies, mature rats were
determined to be more susceptible to kidney damage than young (10- or 20-day-old) rats,
because mature kidney function appeared necessary for thallium to adversely affect the kidney.
4.8.2. Possible Gender Differences
Leloux et al. (1987), MRI (1988), and Downs et al. (1960) are the only available toxicity
studies of thallium compounds that used both male and female rats. Leloux et al. (1987)
administered only a single lethal dose of thallium (I) nitrate to rats and thus did not provide
findings useful for discerning possible gender differences. Downs et al. (1960) reported slight
differences in thallium (III) oxide toxicity between the sexes, with males dying earlier, exhibiting
greater and more severe alopecia, and having more profound decreases in body weight than
females. No sex-related differences in response were noted in rats treated with thallium (I)
acetate (Downs et al., 1960). No marked differences in response were noted in male and female
rats after exposure to thallium (I) sulfate (MRI, 1988); however, high-dose female rats exhibited
a higher incidence of alopecia than males, and hair follicle atrophy was observed in females
only. Overall, the limited data do not identify any consistent pattern of gender-related
differences in response to thallium exposure.
56
-------
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
As discussed in Section 4.6.1., most information on thallium toxicity in humans comes
from poisonings, suicide attempts, or accidental exposures. Epidemiologic studies of either the
general population or occupationally exposed groups are limited by inadequate study design
and/or insufficient exposure characterization. Thus, available human studies do not provide data
useful for dose-response analysis.
There are only four repeat-dose oral toxicity studies of thallium that used more than one
dose level: MRI (1988), Wei (1987), Zasukhina et al. (1983), and Downs et al. (1960). Of these
studies, the subchronic (90-day) toxicity study of thallium (I) sulfate in Sprague-Dawley rats by
MRI (1988) is the most comprehensive study of thallium toxicity. In the Downs et al. (1960)
study of thallium (I) acetate, mortality in two control groups was 30-40%, complicating
interpretation of the findings in treated rats (mortality and alopecia). A study of thallium (III)
oxide by the same investigators (Downs et al., 1960) failed to identify a NOAEL. Further, death
in 2/5 female rats at the lowest dose tested in this study was observed and may have been
treatment related. The repeat-dose dominant lethal study of Zasukhina et al. (1983) suffered
from critical reporting deficiencies and internal inconsistencies (see Section 4.3). Wei (1987)
reported effects on sperm count, motility, and viability in Kunming mice exposed to drinking
water concentrations of thallium (I) carbonate as low as 0.001 mg/L; however, when the treated
males were mated with untreated females, the reproductive index and number of implantations
were not affected at drinking water concentrations up to 10 mg/L (10,000-fold higher than the
concentration identified by the study investigators as the LOAEL for sperm effects). The percent
of dead fetuses was significantly lower than in the control group at concentrations of 0.001, 0.01,
and 0.1 mg/L but was increased at the two highest concentrations (1 and 10 mg/L). Although the
effects on sperm appeared to be dose related, some uncertainty is associated with the reporting of
study findings (see Sections 4.3 and 4.6.1). Further, the unexplained absence of reported sperm
results for 20-50% of the male mice in each of the treatment groups lowers confidence in these
findings. Supporting literature on the reproductive toxicity of thallium is limited. Formigli et al.
(1986) reported adverse effects on the testes and sperm of rats exposed to thallium sulfate in
drinking water but at a drinking water concentration 10,000-fold higher than the LOAEL
reported by Wei (1987). In the absence of confirmatory findings of sperm effects at the low
drinking water concentrations used in the Wei (1987) study, the study findings were not
considered sufficiently reliable to serve as the basis for RfD derivation.
57
-------
Other repeat-dose studies of thallium oral toxicity used study designs that included only a
single-dose group and thus did not provide data useful for dose-response analysis (El-Garawany
et al., 1990; Rossi et al., 1988; Formigli et al., 1986; Gregotti et al., 1985; Manzo et al., 1983).
Manzo et al. (1983) reported mortality at the only dose tested, making this study unsuitable for
RfD derivation. Finally, doses in mg/kg-day could not be estimated from the study information
provided in Rossi et al. (1988).
Based on the above considerations, the 90-day MRI (1988) study was selected as the
principal study for derivation of the RfD. In the MRI (1988) study, rats (20/sex/group) were
treated by gavage daily for 90 consecutive days with 0, 0.01, 0.05, or 0.25 mg/kg-day of an
aqueous solution of thallium (I) sulfate (approximately 0, 0.008, 0.04, or 0.20 mg/kg-day Tl).
No differences in body weight, body weight gains, food consumption, or absolute and relative
organ weights were observed among control groups and groups receiving thallium sulfate.
Lacrimation, exophthalmos, and miosis were observed at higher incidences in the treated male
and female rats at all doses compared with both untreated and vehicle controls (see Table 4-2);
however, ophthalmologic examination and gross and histopathologic examination of the eyes
revealed no treatment-related abnormalities. The incidence of clinical observations related to the
coat and skin (including rough coat, piloerection, shedding, and alopecia) and behavioral
changes (including combined incidences of aggression, tense/agitated behaviors, hyperactivity,
vocalization, and self-mutilation) were also elevated in treated male and female rats at all doses
(see Table 4-2). The investigators attributed most, but not all, instances of alopecia to barbering
behavior. For example, of the 12 high-dose females with alopecia, 5 instances were not totally
attributed to barbering. Review of individual animal data revealed no discernable differences in
either the severity or distribution pattern of alopecia across control and treated groups.
Histopathologic examination did not reveal any statistically significant treatment-related effects,
although it was noted that atrophy of the hair follicles occurred in two high-dose female rats.
Tissue samples of the skin from low- and mid-dose groups, however, were not examined for
histopathologic changes. Subtle, but statistically significant, changes were observed in several
blood chemistry parameters (AST, LDH, sodium and blood sugar levels) (see Table 4-4). These
changes were not confirmed by any histopathologic findings and were not considered by the
investigators to be lexicologically significant.
Although the study authors identified the highest dose (0.20 mg/kg-day Tl) as a NOAEL
based on lack of biological significance of the observed effects, EPA reached the determination
that histopathologic findings in the skin of female rats and dose-related clinical findings, when
considered collectively, may reflect an adverse effect on the health of the rats.
As noted above, histologic examination of skin samples from two high-dose females
showed atrophy of hair follicles. These two animals also exhibited alopecia. Hair loss (alopecia)
is characteristic of thallium poisoning in humans and experimental animals (Ibrahim et al., 2006;
58
-------
Galvan-Arzate and Santamaria, 1998), and typically occurs in humans within two weeks of
exposure. It is hypothesized that thallium's affinity for sulfhydryl groups may be responsible for
alopecia; thallium prevents keratinization of hair proteins by binding with cysteine. Skin
biopsies have been taken from a limited number of patients with alopecia and other symptoms of
thallium poisoning; these biopsies have revealed atrophic and necrotic changes of the skin (Lu et
al., 2007; Heyl and Barlow, 1989; Saddique and Peterson, 1983). For example, skin biopsy
findings from two patients who ingested water that contained thallium included parakeratosis,
dilated hair follicles filled with keratin and necrotic sebaceous materials, mild epidermal atrophy,
and vacuolar degeneration of the basal layer (Lu et al., 2007). Thus, the finding of two cases of
atrophy of the hair follicles in high-dose female rats with alopecia is consistent with the atrophic
changes observed in cases of human thallium poisoning, and suggests that alopecia at the high-
dose (0.2 mg/kg-day Tl) may be related to thallium exposure. This endpoint was considered a
candidate critical effect for derivation of the RfD.
The clinical observations from the MRI (1988) study were also considered as candidate
critical effects for derivation of the RfD. As discussed in Section 4.6.1, dose-related increases in
the incidence of lacrimation, exophthalomos, and miosis, coat-related findings, and behavioral
observations suggest a possible effect on the health of the exposed rats from exposure to
thallium, although the underlying basis for these observations is unknown. Increased incidences
of barbering and behavioral changes are consistent with possible effects on the nervous system, a
known target of thallium.
Clinical chemistry changes were not considered as the basis for determining the point of
departure (POD) for the RfD. At 90 days, AST, LDH, sodium, and sugar levels in the blood of
high-dose male and female rats differed by +31, +38, +4, and -21%, respectively, from vehicle
control group values. Even with these changes, these clinical chemistry parameters are all well
within two standard deviations of the mean for control Sprague-Dawley rats as reported by
Petterino and Argentino-Storino (2006) and Matsuzawa et al. (1993). In the low-dose group,
none of the blood chemistry parameters was statistically significantly different from the vehicle
control group with the exception of the sodium level in female rats, which was increased over
control values by only 1.4%. In light of these modest changes in blood chemistry parameters and
the lack of other findings in exposed animals to confirm the biological significance of these
changes, clinical chemistry parameter data were not selected for dose-response modeling.
In summary, two endpoints were considered as potential critical effects for determination
of the POD for the RfD: (1) hair follicle atrophy in female rats that also had alopecia, and
(2) clinical observations, including those related to animal coat (rough coat, piloerection,
shedding, and alopecia), eyes (including lacrimation, exophthalmos, and miosis), and behavior.
59
-------
5.1.2. Methods of Analysis
Hair Follicle Atrophy
The NOAEL-LOAEL approach was used for dose-response analysis of hair follicle
atrophy in rats with alopecia. A benchmark dose (BMD) analysis was not conducted because the
incidence of histopathologically-determined hair follicle atrophy was not considered amenable to
BMD methods. Histopathological examination of the skin was performed for two groups only—
the high-dose group and vehicle control. Two of 20 female rats in the high-dose group (10%)
had hair follicle atrophy and alopecia that is consistent with thallium toxicity in both animals and
humans and was thus characterized by EPA as possibly treatment-related. The high dose
(0.2 mg/kg-day Tl) was therefore characterized as a LOAEL. Because skin tissue from rats in
the low- and mid-dose groups was not examined for histopathologic changes, the NOAEL for
this endpoint cannot be determined with certainty. Given the low incidence of hair follicle
atrophy in females in the high-dose group and absence of cases of hair follicle atrophy in male
rats, the mid-dose can reasonably be assumed to approximate a NOAEL for skin histopathology.
Thus, an estimated NOAEL of 0.04 mg/kg-day Tl was used as the POD for hair follicle atrophy
from the MRI (1988) study.
Clinical Observations
BMD modeling (U.S. EPA, 2000b) was used to analyze the clinical observation data
from MRI (1988). Incidence data for the selected clinical observations in male and female rats
are summarized in Table 5-1. All of the available dichotomous models in U.S. EPA's BMDS
(version 1.4.1) (U.S. EPA, 2007) were fit to these incidence data.
60
-------
Table 5-1. Incidence data and BMD modeling results for selected clinical observations in
Sprague-Dawley rats treated with thallium sulfate for 90 daysa
Observation
Dose (mg/kg-day)
Untreated
control
Vehicle
control
0.008
0.04
0.2
BMD10b
(mg/kg-day)
Male
BMDL10b
(mg/kg-day)
Coat/skin
Rough coat
Piloerection
Shedding
Alopecia
1/20
0/20
0/20
2/20
3/20
0/20
0/20
1/20
11/20
1/20
4/20
4/20
16/20
4/20
10/20
9/20
19/20
13/20
8/20
4/20
BMD and BMDL well below experimental
range.d
0.020
0.013
All dichotomous models in BMD S exhibited
statistically significant lack of fit.e
All dichotomous models in BMD S exhibited
statistically significant lack of fit.e
Eyes
Lacrimation
Exophthalmos
Miosis
Behavior0
1/20
1/20
0/20
3/20
6/20
5/20
1/20
0/20
19/20
12/20
5/20
7/20
20/20
20/20
7/20
6/20
20/20
20/20
15/20
7/20
BMD and BMDL well below experimental
range.d
BMD and BMDL well below experimental
range.d
0.0069
0.0039
All dichotomous models in BMD S exhibited
statistically significant lack of fit.e
Female
Coat/skin
Rough coat
Piloerection
Shedding
Alopecia
1/20
0/20
0/20
4/20
0/20
0/20
0/20
1/20
1/20
0/20
2/20
4/20
5/20
3/20
3/20
9/20
11/20
8/20
13/20
12/20
0.021f
0.043f
0.020
0.018f
Eyes
Lacrimation
Exophthalmos
Miosis
Behavior0
7/20
5/20
2/20
2/20
6/20
6/20
3/20
2/20
20/20
19/20
1/20
0/20
20/20
20/20
11/20
1/20
20/20
20/20
8/20
7/20
0.013f
0.026f
0.013
0.010f
All dichotomous models in BMD S exhibited
statistically significant lack of fit.e
BMD and BMDL well below experimental
range.d
All dichotomous models in BMD S exhibited
statistically significant lack of fit.e
0.14f
0.054f
61
-------
Table 5-1. Incidence data and BMD modeling results for selected clinical observations in
Sprague-Dawley rats treated with thallium sulfate for 90 daysa
Observation
Dose (mg/kg-day)
Untreated
control
Vehicle
control
0.008
0.04
0.2
BMD10b
(mg/kg-day)
BMDL10b
(mg/kg-day)
a Listed as number of animals with the sign observed at least once during the 90-day study.
b BMD10 = Benchmark dose corresponding to a 10% extra risk; BMDL10 = 95% lower bound on the dose corresponding to
a 10% extra risk.
0 Animals exhibiting one or more behavioral observations at least once during the 90-day study, including the following:
aggressive, tense/agitated, hyperactive, vocalization, self-mutilation.
d Data set not amenable to modeling because of steep slope of the dose-response curve in low-dose region; BMD and
BMDL well below experimental range.
e Some data sets shown in this table exhibited a statistically significant lack of fit with the models available in BMDS.
Where lack of fit is due to characteristics of the dose-response data for high doses, one option sometime used is to adjust
the data set by eliminating the high-dose group(s) (U.S. EPA, 2000b). In the absence of a mechanistic understanding of
the biological response or other observed toxicity in the thallium-exposed animals that could explain the dose response at
the higher doses, dose-response assessment with adjusted data sets (i.e., removing the high-dose group(s)) was not
performed.
f Reported BMD10 and BMDL10 represent averages of results from several similar model fits. See Appendix B for more
details.
Source: MRI (1988).
Consistent with EPA (2000b) BMD technical guidance, consideration was given to
identifying biologically relevant response levels for developing the RfD. A BMR of 10% is
generally used to facilitate a consistent basis of comparison across assessments and was used for
these endpoints in the absence of information regarding the level of change considered to be
biologically significant. Doses (i.e., BMDio [BMD corresponding to a 10% extra risk] and
BMDLio [95% lower bound on the BMD corresponding to a 10% extra risk]) associated with a
BMR of 10% extra risk are presented in Table 5-1.
Details of the BMD modeling conducted for each endpoint presented in Table 5-1 are
provided in Appendix B. In general, model fit was assessed by a chi-square goodness-of-fit test
(i.e., models with/? < 0.1 failed to meet goodness-of-fit criterion) and visual inspection of the
respective plots of observed versus predicted values from the various models. BMDLio
estimates from these models that were within a factor of three of each other suggested no
appreciable model dependence. Fitted models exhibiting adequate fit (i.e., p > 0.1) with Akaike
Information Criterion (AIC) values within two units of the lowest AIC were considered
indistinguishable from one another (Burnham and Anderson, 2002), and thus BMDio and
BMDLio values from these models were averaged. Model fits that yielded the same
mathematical model were counted as a single model for averaging purposes (see Appendix B for
details).
As Table 5-1 shows, clinical observation data sets for female rats were generally more
amenable to BMD modeling than the male rat data sets. The BMDLio values for endpoints
62
-------
related to the coat/skin in female rats ranged from 0.010 to 0.026 mg/kg-day, a range of only
2.6-fold. The lowest BMDLio value for female rats (0.010 mg/kg-day) comes from averaging
the BMDLio values from several similar model fits of incidence data for alopecia. As discussed
in Section 5.1.1, alopecia is a hallmark of thallium poisoning. To the extent that instances of
alopecia resulted from barbering, the occurrence of alopecia could also be consistent with
possible effects of thallium on the nervous system, although no direct evidence for neurotoxicity
was provided.
In male rats, the smallest BMDLio (0.0039 mg/kg-day) is based on data for miosis.
While elevated in thallium-treated animals, the biological significance of miosis is less clear than
the collection of effects related to the rats coat and skin. Review of individual animal data from
the MRI (1988) study revealed that at the high dose, miosis was observed between days 6 to 10
only, and in the majority of animals (approximately 70%) miosis was observed on only one or
two days.
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs)
5.1.3.1. Soluble Thallium Salts: Acetate, Carbonate, Chloride, Nitrate, andSulfate
The principal study considered for RfD derivation - MRI (1988) - involved
administration of thallium (I) sulfate. There are no studies of thallium (I) acetate, thallium (I)
carbonate, thallium (I) chloride, or thallium (I) nitrate that are appropriate to use as the basis for
an RfD. For the following reasons, it was considered appropriate to treat these monovalent
thallium salts as lexicologically equivalent to thallium (I) sulfate when expressed in terms of
thallium. It is likely that the mechanism of toxicity is the same for these salts due to the fact that
they all contain monovalent thallium ions and are water soluble. Only small differences in the
toxicity of various water-soluble thallium (I) salts exist in mice, rats, rabbits, and dogs. In
general, for most laboratory species at an observation period of approximately 1-2 weeks, the
LDso or minimum effective dose values range between 10 and 30 mg/kg body weight for
thallium (I) salts, independent of the exposure route (IPCS, 1996). Therefore, the use of thallium
(I) sulfate as a surrogate for the other thallium salts is considered appropriate.
Candidate RfD values for thallium were derived using PODs for two endpoints from the
MRI (1988) study - a NOAEL of 0.04 mg/kg-day Tl for hair follicle atrophy and an average
BMDLio of 0.01 mg/kg-day Tl for alopecia (as representative of clinical observations more
generally).
Some degree of uncertainty is associated with both PODs. Hair follicle atrophy was
reported in the MRI (1988) study only in high-dose female rats. Because skin tissue from the
low- and mid-dose groups was not examined histopathologically, the NOAEL for this endpoint
could only be estimated. The biological significance of the dose-related clinical observations in
rats in the MRI (1988) study is unclear. In general, the collection of clinical observations data is
63
-------
relatively subjective and less rigorously measured than endpoints such as clinical chemistry or
histopathology. While clinical observations reported in the MRI (1988) study could represent an
indirect measure of stress or possible effects on the nervous system, no direct support for these
etiologies is available. The increased incidence of alopecia in female rats was dose-related (less
clearly so in male rats); however, the background incidence of alopecia in control animals and
attribution of some cases of alopecia by the study authors to barbering behavior introduce
uncertainty in dose-response analysis of this endpoint. Furthermore, examination of individual
animal data from the MRI (1988) study revealed no discernable difference in the severity or
distribution of alopecia across control and treated groups. Effects on the eyes based on clinical
observations (in particular lacrimation and exophthalmos) were not supported by ophthalmic
examination or by gross or histopathological examination of the eyes.
A total uncertainty factor (UF) of 3,000 (10 for interspecies extrapolation, 10 for
intraspecies extrapolation, 3 for extrapolation from a subchronic to a chronic study, and 10 for
database deficiencies) was applied to both PODs to estimate candidate oral RfD values for
soluble thallium salts.
• A default interspecies UF of 10 was applied for extrapolation from laboratory animals
to humans. No information was available to characterize the toxicokinetic or
toxicodynamic differences between experimental animals and humans.
• A default intraspecies UF of 10 was applied to account for variation in human
susceptibility in the absence of information on the variability of response to thallium
in the human population.
Because no chronic toxicity studies for thallium are available, an UF of 3 was applied
to account for extrapolation from subchronic to chronic exposure duration. Oral
toxicity data for thallium suggest that an UF of 10 would overestimate the difference
in response following subchronic and chronic oral exposures. Effects on the coat/skin
as well as other clinical observations occur within weeks of exposure to thallium (i.e.,
these sensitive effects do not require chronic exposure in order to manifest).
• An UF for LOAEL to NOAEL extrapolation was not needed for hair follicle atrophy
in rats with alopecia because a NOAEL was estimated from data provided in the
principal study.
Similarly, a UF to account for extrapolation from a LOAEL to NOAEL was not used
for alopecia (as representative of clinical observations) because BMD methods were
64
-------
used for dose-response analysis of these endpoints. The current approach is to
address this extrapolation as one of the considerations in selecting a BMR for BMD
modeling. In this case, a BMR of 10% increase in the incidence of alopecia was
selected under the assumption that it represents a minimally biologically significant
change.
• The thallium database includes several subchronic oral toxicity studies in rats.
Studies of reproductive and developmental toxicity of thallium compounds in rats and
mice are available; however, these studies used nontraditional study designs that did
not provide adequate testing of reproductive or developmental endpoints. In
reproductive toxicity studies by Gregotti et al. (1985) and Formigli et al. (1986), male
rats only were exposed for periods up to 60 days and evaluation of reproductive
toxicity was limited to examination of male reproductive organs. In two other
reproductive toxicity studies, male rats (Zasukhina et al., 1983) or mice (Wei, 1987)
only were exposed to thallium compounds for 6 to 8 months and mated with untreated
females. Confidence in these latter two studies was low (see Sections 4.3 and 4.6.1).
No studies of reproductive toxicity of thallium in females or multigeneration
reproductive toxicity studies were identified. Developmental toxicity by the oral
route was limited to Rossi et al. (1988) in which rats were exposed to thallium sulfate
from gestation day 1 through 60 days of age or from birth through 60 days of age. No
investigation of developmental endpoints at the end of the gestation period was
performed. Despite the limitations in the available reproductive and developmental
toxicity studies, these studies provide suggestive evidence that thallium compounds
can adversely affect male reproductive organs and the developing fetus and highlight
the deficiencies in the current thallium database. Limited neuropathologic
examinations were included in the subchronic toxicity studies by MRI (1988) and
Manzo et al. (1983), but no standard neurobehavioral studies were identified.
Because the nervous system is a sensitive target of thallium toxicity, the limited
investigation of thallium neurotoxicity represents a data deficiency. Thus, a database
UF of 10 was applied to account for a lack of adequate developmental toxicity studies
and a two-generation reproductive toxicity study, and additional uncertainty
associated with the limited data available on neurotoxicity in light of the potential for
neurotoxicity to represent a sensitive endpoint for thallium exposure.
Thus, the candidate RfDs for thallium (I) are calculated as the POD (NOAEL or BMDL)
divided by the composite UF, as follows:
65
-------
Derivation of candidate RfD based on hair follicle atrophy in animals with alopecia:
0.04 mg/kg-day Tl H- 3000 = 1 x 10"5 mg/kg-day Tl
Derivation of candidate RfD based on clinical observations/alopecia:
0.01 mg/kg-day Tl - 3000 = 3 x 10"6 mg/kg-day Tl
The available toxicity database for thallium contains studies that are generally of poor
quality. The MRI (1988) study that was selected as a candidate principal study suffers from
certain critical limitations (e.g., high background incidence of alopecia, lack of histopathological
examination of skin tissue in low- and mid-dose groups, and inadequate examination of objective
measures of neurotoxicity), and there are particular difficulties in the selection of appropriate
endpoints. Therefore, even though an RfD would generally be derived with a combined
uncertainty factor of 3000, an RfD for soluble thallium salts is not derived in this specific case.
5.1.3.2. Insoluble Thallium Salts: Thallium (III) Oxide
No oral studies of thallium (III) oxide were found that are adequate to support derivation
of an RfD. Downs et al. (1960) administered thallium (III) oxide to Wistar-derived albino rats
via the diet for 15 weeks at concentrations of 0, 20, 35, 50, 100, and 500 ppm. The study
reported only body weights, mortality, kidney weights, and limited histopathology. Because all
relevant endpoints were not evaluated, treatment group sizes were small (five per group per sex),
and few animals survived to termination, this study could not be used to derive an RfD.
Because of differences in the physical-chemical properties of thallium (I) sulfate and
thallium (III) oxide, toxicity information for thallium (I) sulfate cannot be used to inform the
toxicity of thallium (III) oxide. Unlike thallium (I) sulfate, thallium (III) oxide is insoluble in
water. Thallium (I) sulfate contains monovalent thallium (Tl+1), while thallium (III) oxide
contains trivalent thallium (Tl+3). Although the gastric environment may increase the solubility
of thallium (III) oxide, no studies are available that examine the effects of the gastric pH and
gastric environment on the valence state of thallium (III) oxide.
Limited evidence from the toxicology literature suggests that the distribution of
thallium (I) and thallium (III) compounds and the lethality of these two compounds may be
comparable. Sabbioni et al. (1980a) determined that, following oral administration of either
inorganic monovalent thallium (Tl+1) as thallium (I) sulfate or trivalent thallium (Tl+3) as
thallium (III) chloride, a similar distribution of thallium in the tissues was seen (valence state
could not be determined) at 16 hours and 8 days after administration, indicating that the valence
state of thallium did not affect tissue distribution (and presumably uptake). Downs et al. (1960)
demonstrated similar oral 7-day LDso values for thallium (I) acetate (32 mg/kg Tl) and thallium
(III) oxide (39 mg/kg Tl) in female rats, indicating that lethality may be independent of valence
66
-------
state. Whether other endpoints would respond similarly to different valence states of thallium is
unknown. Monovalent thallium compounds have been determined to behave similarly to
potassium (K+), thus disrupting Na+/K+-ATPase and the systems dependent on this transporter
(e.g., liver and kidney). Trivalent thallium has not been demonstrated to behave in the same
manner. A single in vitro study of mono- and trivalent thallium compounds (Diaz and Monreal,
1994) suggested that biological responses to thallium in the (I) and (III) valence states may
differ. Using an in vitro system with liposomes prepared with lipid from brain myelin, these
investigators reported that trivalent thallium, but not monovalent thallium, mediated a rapid
chloride/hydroxyl ion exchange through the lipid bilayers.
Overall, the available toxicity information for thallium (III) oxide specifically and
thallium compounds more generally is insufficient to support derivation of an RfD for thallium
(III) oxide.
5.1.3.3. Thallium (I) Selenite
No toxicity studies of thallium (I) selenite are available. Thallium (I) selenite contains
monovalent thallium as does thallium (I) sulfate. No information could be found in the
literature, however, on the water solubility of thallium (I) selenite. In the absence of solubility
information, toxicity information for thallium (I) sulfate cannot be used to inform the toxicity of
thallium (I) selenite. Accordingly, the available data do not support derivation of an RfD for
thallium (I) selenite.
5.1.4. Candidate RfD Comparison Information
PODs and candidate RfD values based on selected studies included in Table 4-5 are
arrayed in Figures 5-1 to 5-5 and provide perspective on the magnitude of PODs and candidate
RfD values associated with different toxicity endpoints. These figures should be interpreted with
caution because the PODs across studies are not necessarily comparable, nor is the confidence in
the data sets from which the PODs were derived the same. PODs in these figures may be based
on a LOAEL or a BMDL, and the nature, severity, and incidence of effects occurring at a
LOAEL are likely to vary. To some extent, the confidence associated with the data sets is
reflected in the magnitude of the total UF applied to the POD (i.e., the size of the bar); however,
the text of Sections 5.1.1 and 5.1.2 should be consulted for a more complete understanding of the
issues associated with each data set.
Thallium causes toxicity in a wide range of target organs, including the nervous system,
kidney, cardiovascular system, liver, skin, reproductive system, and possibly the developing
fetus. Figure 5-1 provides a graphical display of dose-response information for skin
histopathology (hair follicle atrophy in female rats with alopecia) based on the 90-day MRI
(1988) study. Uncertainties in this data set are discussed in Section 5.1.3.1.
67
-------
Figure 5-2 provides a graphical display of dose-response information for clinical
observations from the 90-day study conducted by MRI (1988). Collectively, the dose-related
increased incidences of effects on the coat, eyes, and behavior suggest a potential treatment-
related effect of thallium, possibly including some measure of stress or other effect on the
nervous system; however, the underlying basis for these clinical observations is unknown. It
should be noted that Figure 5-2 includes PODs derived using BMDLio values as the POD for
those data sets for which a model fit could be obtained from the models in BMDS, as well as a
reference value derived using the LOAEL as the POD (since some data sets were not fit by the
models in BMDS). Uncertainties in these data sets are discussed in Section 5.1.3.1.
In some studies, doses of thallium that do not affect survival have been shown to affect
clinical chemistry parameters (e.g., BUN, serum creatinine, bilirubin, and ALT) that may be
indicative of effects on the kidney and liver. Changes in clinical chemistry parameters observed
in El-Garawany et al. (1990) are plotted in Figure 5-3. It should be noted that application of a
composite UF of 10,000, reflecting five areas of uncertainly, indicates that derivation of a
reference value using such a data set is highly uncertain. In such cases, U.S. EPA (2002)
recommends that an RfD not be derived.
Evidence from experimental animal studies suggests that thallium may produce effects on
reproductive function. Figure 5-4 includes plots of LOAELs and associated UFs for data sets
from Formigli et al. (1986) and Wei (1987). Formigli et al. (1986) reported testicular effects in
rats exposed to thallium sulfate for 60 days but used only a single dose level of thallium.
Therefore, this study has limited utility for dose-response analysis. Wei (1987) found effects on
sperm in mice exposed to thallium carbonate at drinking water concentrations as low as
0.001 mg/L. As discussed in Sections 4.6.1 and 5.1.1, confidence in this study is low because of
uncertainties associated with estimates of exposure, reporting inconsistencies, and unexplained
absence of sperm data for male mice in all dose groups (ranging from 20 to 50%). Composite
UFs of 100,000 and 10,000 were applied to the LOAELs from the Formigli et al. (1986) and Wei
(1987) data sets. As with the data from El-Garawany et al. (1990), the magnitude of the
uncertainty associated with these data sets indicates that they are insufficient to support
derivation of a reference value.
In humans, numerous case reports of neuropathy following thallium poisoning have been
reported. In experimental animals, effects on the nervous system have been most clearly
observed following oral exposure at levels also associated with increased mortality (e.g., Manzo
et al. [1983]) or following injection exposure (see Section 4.4.3). Therefore, although the
nervous system is known to be a target of thallium toxicity, the available studies do not provide
adequate dose-response information on standard measures of nervous system toxicity at
nonlethal doses that are appropriate for defining a POD.
Figure 5-5 displays PODs for the major targets of toxicity (from Figures 5-1 to 5-4)
associated with oral exposure to thallium, including skin histopathology, alopecia and other
68
-------
general clinical observations, changes in clinical chemistry possibly indicative of the kidney and
liver as targets of thallium toxicity, and the reproductive system, along with the UFs that would
be applied for candidate RfD derivation.
69
-------
0.1 -!
0.01
=• 0.001
I-
>.
t
O)
0.00001
0.000001
MRI (1988); 90-d rat study (gavage); hair follicle atrophy; NOAEL
O
POD
UF-animal
to human
UF—human
variation
UF-LOAELto
NOAEL
UF-subchronic
to chronic
UF-database
deficiencies
Derived
candidate RfD
Figure 5-1. POD (mg/kg-day) with corresponding derived candidate reference value that would result if histopathologic
changes of the skin (hair follicle atrophy) were used as the critical effect.
70
-------
1.0E-01
1.0E-02
1.0E-03
-------
1 .OE+00
1.0E-01
p 1.0E-02
ro
O)
1 .OE-03
1 .OE-04
1 .OE-05
EI-Garawanyetal. (1990); 90-d rat study (gavage); clinical chem; LOAEL
O
POD
UF-animal
to human
UF-human
variation
UF-LOAEL to
NOAEL
UF—subchronic
to chronic
UF-database
deficiencies
Derived
candidate RfD
Figure 5-3. POD (mg/kg-day) with corresponding derived candidate reference value that would result if clinical chemistry
changes (suggesting the liver or kidney as a target) were used as the critical effect.
72
-------
1 .OE+00
1 .OE-01
1 .OE-02
1 .OE-03
ro
01
Q
1.0E-04
1 .OE-05
1 .OE-06
1 .OE-07
1 .OE-08
o
POD
UF—animal
to human
UF-human
variation
UF-LOAEL to
NOAEL
UF—subchronic
to chronic
UF—database
deficiencies
Derived
candidate RfD
Formigli etal. (1986); 60-d rat study (DW);
testicular effects; LOAEL
Wei (1987); 6-month mouse study (DW);
sperm effects; LOAEL
Figure 5-4. PODs (mg/kg-day) with corresponding derived candidate reference values that would result if reproductive
toxicity endpoints were used as the critical effect.
73
-------
1.0E+00
1.0E-01
1.0E-02
~ 1.0E-03
t
"3>
O
o
1.0E-04
1.0E-05
1.0E-06
1.0E-07
1.0E-08
Skin histopathology
Ben avi or/stress/gen eral
health
Reproduction
Clinical chemistry (liver,
kidney)
• POD
njT| UF-animal
LLLU to human
II UF-human
I—I variation
tffl UF-LOAELto
^ NOAEL
I I UF-subchronic
'—' to chronic
UF-database
deficiencies
Derived
candidate RfD
Figure 5-5. PODs (mg/kg-day) with corresponding derived candidate reference values that would result if alternative
endpoints were used as the critical effect.
74
-------
5.1.5. Previous RfD Assessment
The previous IRIS RfD values for thallium (I) acetate, thallium (I) carbonate, thallium (I)
chloride, thallium (I) nitrate, and thallium (I) sulfate were posted to the IRIS database in
September 1988. These assessments were based on the same principal study considered for the
principal study in the current assessment (MRI, 1988). (The principal study was previously cited
as U.S. EPA [1986c]. MRI subsequently issued a revised final report in 1988, which is the basis
for the candidate RfD values developed in the current assessment. There are no substantive
differences in the findings and conclusions between the 1986 and 1988 versions of the MRI
report.) Previous RfD values (adjusted for differences in molecular weight) ranged from 8 x 1CT5
to 9 x 1CT5 mg/kg-day. These RfD values were based on a NOAEL of 0.25 mg/kg-day thallium
sulfate, the highest dose tested by MRI (1988), and application of a composite UF of 3,000 (10 to
extrapolate from subchronic to chronic data, 10 for intraspecies extrapolation, 10 to account for
interspecies variability, and 3 to account for lack of reproductive and chronic toxicity data).
Although both the previous and current assessments for soluble thallium salts (acetate, carbonate,
chloride, nitrate, and sulfide) identified the same study as principal, the assessments differed in
terms of the interpretation of the MRI (1988) findings and method for dose-response analysis.
Most significantly, the current assessment does not recommend a value for the RfD.
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
Information on the inhalation toxicity of thallium is insufficient to derive an inhalation
RfC. No studies of inhaled thallium in experimental animals were identified, and occupational
epidemiology studies involving possible inhalation exposures to thallium were limited and
inconclusive.
5.3. CANCER ASSESSMENT
There are no human or animal studies available that are adequate to assess the
carcinogenic potential of thallium salts.
75
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
Thallium is well absorbed from the GI tract, skin, and respiratory tract and is distributed
throughout the body. Thallium, as an element, is not metabolized, but the extent to which it may
be converted from one valence state to another (i.e., Tl+1 and Tl+3) in the body is not known.
Excretion of thallium occurs mainly through the urine and feces, but the amounts are species
dependent. Humans mainly excrete thallium through the urine, but it also has been detected in
the hair, sweat, tears, and breast milk.
Thallium salts cause a wide spectrum of adverse effects in humans and animals.
Alopecia is an effect that is characteristic of thallium exposure. Alopecia generally occurs
within 2 weeks of exposure and is reversible when thallium exposure is removed. Only one
epidemiologic study reported a negative correlation between thallium exposure and hair loss
(Brockhaus et al., 1981); however, this study lacked measures of chronic exposure to thallium
and was limited by reliance on questionnaires to determine symptomology in thallium-exposed
individuals.
Some study observations suggest that the nervous system may be the primary target
organ for thallium salt toxicity. Neurotoxicity has been observed after a single dose of 0.31 g
thallium acetate in an adult male or 1 mg/kg thallium (I) nitrate for 4 days in adult rats. A
variety of neurological effects have been reported in humans, including lethargy, back pain,
paresthesia of the hands and feet, weakness (including facial weakness), inability to walk, and
prolonged mental defects. Some of the effects are reversible depending on the severity, while
others are irreversible and may require long-term care. Animal studies support these findings in
humans.
Kidney damage in humans has been noted by increases in BUN levels and serum
creatinine and is reversible with treatment and/or discontinued exposure to thallium. Data from
animal studies support those in humans with regard to kidney damage; thallium-related effects
include increased or decreased urine output (depending on dose), protein in the urine, and
changes in electrolyte balance. Histopathologic examination of thallium-exposed animals
revealed changes in kidney morphology, including atrophied and vacuolated kidney tubules,
amorphous material in the lumen of the proximal tubules, disorganized brush borders, and
thickening ascending limb of the loop of Henle. Na+/K+-ATPase activity in the medulla also was
significantly reduced. None of these changes were dose related and were generally seen with
large doses of thallium. Animal studies also suggest that kidney toxicity requires mature kidney
function. The subchronic (90-day) toxicity study in rats used to derive the RfD showed moderate
increases in AST and LDH, which are general indicators of tissue damage (MRI, 1988).
76
-------
However, a specific relation to possible kidney damage was not indicated due to the lack of
changes in other clinical chemistry parameters (i.e., BUN and creatinine) and histopathologic
changes in the kidney.
Cardiotoxicity findings in thallium-exposed humans are variable. Many case reports
indicate hypertension, while a few reported hypotension. Animal studies indicate that thallium
affects heart rate and causes a decrease in blood pressure, which also can be related to kidney
effects and is supported by in vitro results.
Many of the case reports in humans reported increased ALT and/or AST levels,
indicating liver toxicity. These effects returned to normal after patients received medical care for
thallium exposure. Various indicators of liver damage, including increases in ALT, AST, serum
bilirubin, lipid peroxidation, triglycerides, and serum alkaline phosphatase and decreases in
glycogen, glutathione, and liver Na+/K+-ATPase, have been reported in animal studies. In
addition, histopathology revealed swollen and vacuolated cells and swollen mitochondria. Many
of these effects were reversible in animals after a single dose of thallium. Statistically significant
increases in AST and LDH were observed in the 90-day MRI (1988) toxicity study in rats;
however, the increases remained within the range reported in control rats and were not
accompanied by histopathologic changes of the liver or changes in related clinical chemistry
parameters (i.e., ALT).
Low birth weight is a likely adverse effect of thallium exposure in females (humans and
animals) exposed during pregnancy. Male mice exposed to thallium were reported to have low
sperm counts, low sperm motility, and increases in the number of deformed sperm. Testicular
effects observed in animals included disarrangement of the tubular epithelium, cytoplasmic
vacuolation and distention of smooth endoplasmic reticulum of the Sertoli cells, and reduced
beta-glucuronidase activities. Mating exposed male mice to unexposed female mice appeared to
cause a decrease in the number of live fetuses and an increase in the overall rate of dead fetuses;
however, critical information concerning the doses of thallium administered makes it difficult to
assess these results.
There are no human studies relating thallium exposure to developmental toxicity. A
literature review of pregnant women exposed to thallium during various stages of pregnancy
reported an association with low birth weight. A survey of children born near a cement plant
emitting thallium found an increase in congenital malformations over those reported to the
government; however, the study authors did not consider these malformations to be related to
thallium exposure because two of the cases were considered attributable to hereditary factors and
the incidence was similar to that reported in the literature. Because of various study limitations,
the findings are considered inconclusive. In rats exposed transplacentally and/or via mother's
milk, then via the drinking water until 60 days of age, thallium (1 mg/dL in the drinking water of
dams' then offspring) affected bone development and vasomotor reactivity. Chick embryos
77
-------
exposed to thallium developed achondroplasia; these data further support the potential role of
thallium in the disruption of bone development.
There are no studies available to determine the carcinogenic potential of thallium in
animals and no adequately conducted studies in humans. The limited number of studies of the
genotoxicity of thallium compounds provides inconsistent evidence of genotoxic potential.
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the Agency
concluded that there is "inadequate information to assess the carcinogenic potential."
6.2. DOSE RESPONSE
The 90-day oral toxicity study of thallium (I) sulfate in Sprague-Dawley rats (MRI, 1988)
was selected as a potential principal study. PODs for two alternative endpoints from the MRI
(1988) study were selected for dose-response analysis - (1) hair follicle atrophy in female rats
with alopecia, and (2) clinical observations, including those related to animal coat (rough coat,
piloerection, shedding, and alopecia), eyes (including lacrimation, exophthalmos, and miosis),
and behavior. Using hair follicle atrophy, the high dose (0.2 mg/kg-day Tl) was identified as a
LOAEL; the mid dose (0.04 mg/kg-day Tl) was considered a NOAEL and was used as a
candidate POD. For clinical observations, an average BMDLio of 0.01 mg/kg-day Tl was
derived using BMD modeling methods.
A total uncertainty factor of 3,000 (10 for interspecies extrapolation, 10 for intraspecies
extrapolation, 3 for extrapolation from a subchronic to chronic study, and 10 for database
deficiencies) was applied to the PODs to yield candidate RfD values for thallium in the form of
soluble thallium salts of 1 x 10"5 mg/kg-day Tl (for hair follicle atrophy) or 3 x 10"6 mg/kg-day
Tl (for clinical observations). The available toxicity database for thallium contains studies that
are generally of poor quality. The MRI (1988) study that was selected as a candidate principal
study suffers from certain critical limitations (e.g., high background incidence of alopecia, lack
of histopathological examination of skin tissue in low- and mid-dose groups, and inadequate
examination of objective measures of neurotoxicity), and there are particular difficulties in the
selection of appropriate endpoints. Therefore, even though an RfD would generally be derived
with a combined uncertainty factor of 3000, an RfD for soluble thallium salts is not derived in
this specific case.
The available data are not adequate to derive an RfD for thallium (III) oxide or thallium
(I) selenite. Inhalation toxicity studies are not available to support derivation of RfCs for any
thallium compounds.
78
-------
7. REFERENCES
All, SF; Jairaj, K; Newport, GD; et al. (1990) Thallium intoxication produces neurochemical alterations in rat brain.
Neurotoxicology 11:381-390.
Ammendola, A; Ammendola, E; Argenzio, F; et al. (2007) Clinical and electrodiagnostic follow-up of an adolescent
poisoned with thallium. Neurol Sci 28:205-208.
Andre, T;Ullberg, S; Winqvist, G. (1960) The accumulation and retention of thallium in tissues of the mouse. Acta
Pharmacol Toxicol 16:229-234. (As cited in CalEPA, 1999).
Appenroth, D; Gambaryan, S; Winnefeld, K; et al. (1995) Functional and morphological aspects of thallium-induced
nephrotoxicity in rats. Toxicology 96(3):203-215.
Appenroth, D; Tiller, S; Gambaryan, S; et al. (1996) Ontogenetic aspects of thallium-induced nephrotoxicity in rats.
J Appl Toxicol 16(3):235-243.
Aoyama, H. (1989) Distribution and excretion of thallium after oral and intraperitoneal administration of thallous
malonate and thallous sulfate in hamsters. Bull Environ Contam Toxicol 42:456-463.
Arbiser, JL; Alani, R; Flynn, E; et al. (1997) Effects of thallium ion on cellular components of the skin. J Dermatol
24:147-155.
ATSDR (Agency for Toxic Substance and Disease Registry). (1992) Toxicological profile for thallium. Public
Health Service, U.S. Department of Health and Human Services, Atlanta, GA. Available online at
http://www.atsdr.cdc.gov/toxpro2.html.
Atsmon, J; Taliansky, E; Landau, M; et al. (2000) Thallium poisoning in Israel. Am J Med Sci 320(5):327-330.
Barclay, RK; Peacock, WC; Karnofsky, DA. (1953) Distribution and excretion of radioactive thallium in the chick
embryo, rat, and man. J Pharmacol Exp Ther 107:178-187. (As cited in CalEPA, 1999).
Barrera, H; Gomez-Puyou, A. (1975) Characteristics of the movement of K+ across the mitochondrial membrane and
the inhibitory action of Tl+. J Biol Chem 250(14):5370-5374.
Barroso-Moguel, R; Rios-Castaneda, C; Villeda-Hernandez, J; et al. (1990) Neurotoxicity of thallium biochemical
and morphological study of organic lesions. Arch Invest Med 21: 115-122.
Barroso-Moguel, R; Villeda-Hernandez, J; Mendez- Armenia, M; et al. (1992) Osteochrondric lesions in developing
rats intoxicated with thallium twenty -four hours afterbirth. ArchMedRes 23(3):129-133.
Barroso-Moguel, R; Mendez-Armenta, M; Villeda-Hernandez, J; et al. (1996) Experimental neuromyopathy induced
by thallium in rats. J Appl Toxicol 16(5):385-389.
Britten, JS; Blank, M. (1968) Thallium activation of (Na+-K+)-activated ATPase of rabbit kidney. Biochim Biophys
Acta 159: 160-166.
Brockhaus, A; Dolgner, R; Ewers, U; et al. (1981) Intake and health effects of thallium among a population living in
the vicinity of a cement plant emitting thallium containing dust. Int Arch Occup Environ Health 48:375-389.
Brown, DR; Callahan, BG; Cleaves, MA; et al. (1985) Thallium-induced changes in behavioral patterns: correlation
with altered lipid peroxidation and lysosomal enzyme activity in brain regions of male rats. Toxicol Ind Health
Bugarin, MG; Casa, JS; Sordo, J; et al. (1989) Thallium (I) interactions in biological fluids: a potentiometric
investigation of thallium (I) complex equilibria with some sulphur-containing amino acids. J Inorg Biochem 35:95-
105.
79
-------
Burnham, KP; Anderson, DR. (2002) Model selection and multimodel interference: a practical information-theoretic
approach. 2nd edition. New York, NY: Springer-Verlag.
Careaga-Olivares, J; Gonzalez-Ramirez, D. (1995) Penicillamine produces changes in the acute blood elimination
and tissue accumulation of thallium. Arch Med Res 26(4):427-430.
Cavanagh, JB. (1991) What have we learnt from Graham Frederick Young? Reflections on the mechanism of
thallium neurotoxicity. Neuropathol Appl Neurobiol 17:3-9.
Cavanagh, JB; Fuller, NH; Johnson, HRM; et al. (1974) The effects of thallium salts, with particular reference to the
nervous system changes. Q J Med 43:293-319.
Cavieres, JD; Ellory, JC. (1974) Thallium and the sodium pump in human red cells. J Physiol 243:243-266.
CalEPA (California Environmental Protection Agency) (1999) Public health goal for thallium in drinking water.
Office of Environmental Health Hazard Assessment, California Environmental Protection Agency, Sacramento, CA.
Available online at http://oehha.ca.gov/water/phg/pdf/thal_f.pdf.
CDC (Centers for Disease Control and Prevention). (2005) Third national report on human exposure to
environmental chemicals. National Center for Environmental Health, Public Health Service, Department of Health
and Human Services, Atlanta, GA; NCEH Pub. No. 05-0570. Available online at
http://www.cdc.gov/ExposureReport/pdf/thirdreport.pdf.
Davis, LE; Standefer, JC; Kornfeld, M; et al. (1981) Acute thallium poisoning: lexicological and morphological
studies of the nervous system. AnnNeurol 10:38-44.
Derelanko, MJ; Hollinger, MA; eds. (1995) CRC handbook of toxicology. Boca Raton, FL: CRC Press.
Diaz, RS; Monreal, J. (1994) Thallium mediates a rapid chloride/hydroxyl ion exchange through myelin lipid
bilayers. Mol Pharmacol 46:1210-1216.
Dolgner, R; Brockhaus, A; Ewers, U; et al. (1983) Repeated surveillance of exposure to thallium in a population
living in the vicinity of a cement plant emitting dust containing thallium. Int Arch Occup Environ Health 52:79-94.
Downs, AJ, ed. (1993) Chemistry of aluminium, gallium, indium, and thallium. London, UK: Blackie Academic &
Professional.
Downs, WL; Scott, JK; Steadman, LT; et al. (1960) Acute and sub-acute toxicity studies of thallium compounds.
Industr Hygiene J 21:399-406.
Elsenhans, B; Schumann, K; Forth, W. (1991) Toxic metals: interactions with essential metals. In: Rowland, IR; ed.
Nutrition, toxicity, and cancer. Boca Raton: CRC Press; pp. 223-258.
El-Garawany, AA; Samaan, HA; Sadek, M. (1990) Comparative hepatorenal toxicity of some commonly used
chemical environmental pollutants. Egypt J Pharm Sci 31(1^):331-336.
Feldman, J; Levisohn, DR. (1993) Acute alopecia: clue to thallium toxicity. Pediatr Dermatol 10(1):29-31.
Fleck, C; Appenroth, D. (1996) Renal amino acid transport in immature and adult rats during thallium-induced
nephrotoxicity. Toxicology 106:229-236.
Formigli, L; Scelsi, R; Poggi, P; et al. (1986) Thallium-induced testicular toxicity in the rat. Environ Res 40:531-
539.
Forth, W; Rummel, W. (1975) Gastrointestinal absorption of heavy metals. In: Forth, W; Rummel, W; Aguiar, AJ;
eds. International encyclopedia of pharmacology and therapeutics: Section 39B. Pharmacology of intestinal
absorption: gastrointestinal absorption of drugs. Vol. II. Oxford, New York, NY: Pergamon Press; p. 647.
80
-------
Galvan-Arzate, S; Rios, C. (1994) Thallium distribution in organs and brain regions of developing rats. Toxicology
90:63-69.
Galvan-Arzate, S; Santamaria, A. (1998) Thallium toxicity. Toxicol Lett 99:1-13.
Galvan-Arzate, S; Martinez, A; Medina, E; et al. (2000) Subchronic administration of sublethal doses of thallium to
rats: effects on distribution and lipid peroxidation in brain regions. Toxicol Lett 116:37-43.
Galvan-Arzate, S; Pedraza-Chaverri, J; Medina-Campos, ON; et al. (2005) Delayed effects of thallium in the rat
brain: regional changes in lipid peroxidation and behavioral markers, but moderate alterations in antioxidants, after a
single administration. Food Chem Toxicol 43:1037-1045.
Garrett, NE; Lewtas, J. (1983) Cellular toxicity in Chinese hamster ovary cell cultures. I. Analysis of cytotoxicity
endpoints for twenty-nine priority pollutants. Environ Res 32:455-465.
Gefel, A; Liron, M; Hirsch, W. (1970) Chronic thallium poisoning. Israel J Med Sci 6:380-382.
Gibson, JE; Becker, BA. (1970) Placenta! transfer, embryotoxicity, and teratogenicity of thallium sulfate in normal
and potassium-deficient rats. Toxicol Appl Pharmacol 16:120-132.
Gosselin, RE; Smith, RP; Hodge, HC. (1984) Clinical toxicology of commercial products. 5th edition. Baltimore,
MD: Williams and Wilkins; pp. III-379-III-383. (As cited in CalEPA, 1999).
Greer, JM; Capecchi, MR. (2002) HoxbS is required for normal grooming behavior in mice. Neuron 33:23-34.
Gregotti, C; Di Nucci, A; Formigli, L; et al. (1985) Altered testicular enzyme patterns in rats after long-term
exposure to thallium sulphate. J Toxicol Clin Exp 5(4):265-271.
Gregotti, C; Di Nucci, A; Costa, LG; et al. (1992) Effects of thallium on primary cultures of testicular cells. J
Toxicol Environ Health 36:59-69.
Hall, BK. (1972) Thallium-induced achondroplasia in the embryonic chick. Dev Biol 28:47-60.
Hall, BK. (1985) Critical periods during development as assessed by thallium-induced inhibition of growth of
embryonic chick tibiae in vitro. Teratology 31:353-361.
Hantson, P; Desoir, R; Leonard, ED; et al. (1997) Cytogenetic observations following thallium poisoning. J Toxicol
Environ Health 50:97-100.
Hanzel, CE; Verstraeten, S V (2006) Thallium induces hydrogen peroxide generation by impairing mitochondrial
function. Toxicol Appl Pharmacol 216:485-492.
Hanzel, CE; Verstraeten, SV. (2009) T1(I) and Tl(III) activate both mitochondrial and extrinsic pathways of
apoptosis in rat pheochromocytoma (PC12) cells. Toxicol Appl Pharmacol 236:59-70.
Hanzel, CE; Villarverde, MS; Verstraeten, SV. (2005) Glutathione metabolism is impaired in vitro by thallium(III)
hydroxide. Toxicol 207:501-510.
Hasan, M; Ali, SF. (1981) Effects of thallium, nickel and cobalt administration on the lipid peroxidation in different
regions of the rat brain. Toxicol Appl Pharmacol 57:8-13.
Hasan, M; Haider, SS. (1989) Acetyl-homocysteine thiolactone protects against some neurotoxic effects of thallium.
Neurotoxicology 10:257-262.
Hasan, M; Chandra, SV; Bajpai, VK; et al. (1977) Electron microscopic effects of thallium poisoning on the rat
hypothalamus and hippocampus: biochemical changes in the cerebrum. Brain Res Bull 2:255-261.
81
-------
Hasan, M; All, SF; Tariq, M. (1978) Levels of dopamine, norepinephrine and 5-hydroxytryptamine in different
regions of the rat brain in thallium toxicosis. Acta Pharmacol Toxicol 43:169-173.
Herman, MM; Bensch, KG. (1967) Light and electron microscopic studies of acute and chronic thallium
intoxication in rats. Toxicol Appl Pharmacol 10:199-222.
Heyl, T; Barlow, RJ. (1989) Thallium poisoning: a dermatological perspective. Br JDermatol 121:787-791.
Hirata, M; Taoda, M; Ono-Ogasawara, M; et al. (1998) A probable case of chronic occupational thallium poisoning
in a glass factory. Ind Health 36:300-303.
Hoffman, RS. (2000) Thallium poisoning during pregnancy: a case report and comprehensive literature review. Clin
Toxicol 38(7):767-775.
Hughes, MN; Man, WK; Whaler, BC. (1978) The toxicity of thallium (I) to cardiac and skeletal muscle. Chem Biol
Interact 23:85-97.
Ibrahim, D; Frobert, B; Wolf, A; et al. (2006) Heavy metal poisoning: clinical presentations and pathophysiology.
Clin Lab Med 26:67-97.
IPCS (International Programme on Chemical Safety). (1996) Thallium. Environmental health criteria. Vol. 182.
World Health Organization, Geneva, Switzerland. Available online at http://www.inchem.org/
documents/ehc/ehc/ehc 182 .htm.
Kada, T; Hirano, K; Shirasu, Y. (1980) Screening of environmental chemical mutagens by the Rec-assay system
with Bacillus subtilis. In: de Serres, FJ; Hollaender, A; eds. Chemical mutagens: principles and methods for their
detection. New York, NY: Plenum Press; pp. 149-173.
Kalueff, AV; Tuohimaa, P. (2005) Contrasting grooming phenotypes in three mouse strains markedly different in
anxiety and activity (129S1, BALB/c and NMRI). Behav Brain Res 160:1-10.
Kalueff, AV; Minasyan, A; Keisala, T; et al. (2006) Hair barbering in mice: implications for neurobehavioral
research. Behav Proc 71:8-15.
Kanematsu, N; Kara, M; Kada, T. (1980) REC assay and mutagenicity studies on metal compounds. Mutat Res
77:109-116.
Karnofsky, DA; Ridgway, LP; Patterson, PA. (1950) Production of achondroplasia in the chick embryo with
thallium. Proc Soc Exp Biol Med 73:255-259.
Kazantzis, G. (2007) Thallium. In: Nordberg, GF; Fowler, BA; Nordberg, M; et al.; eds. Handbook on the
toxicology of metals. New York, NY: Elsevier; pp. 827-837.
Kennedy, P, Cavanagh, JB. (1977) Sensory neuropathy produced in the cat with thallous acetate. Acta Neuropathol
(Berl) 39:81-88.
Kuo, HC; Huang, CC; Tsai, YT; et al. (2005) Acute painful neuropathy in thallium poisoning. Neurology 65:302-
304.
Kuperberg, JM; Ngong, JM; Rutledge, LP; et al. (1998) Central and peripheral alteration of the cholinergic enzyme
activities of the rat in response to repeated and acute thallium exposure. Toxic Subst Mech 17:285-298.
Lameijer, W; vanZwieten, PA. (1976) Acute cardiovascular toxicity of thallium (I) ions. Arch Toxicol 35:49-61.
Lameijer, W; van Zwieten, PA. (1977) Kinetic behavior of thallium in the rat: accelerated elimination of thallium
owing to treatment with potent diuretic agents. Arch Toxicol 37:265-273.
82
-------
Lehmann, FPA; Favari, L. (1985) Acute thallium intoxication: kinetic study of the relative efficacy of several
antidotal treatments in rats. Arch Toxicol 57:56-60. (As cited in CalEPA, 1999).
Leloux, M-S; Lich, NP; Claude, J-R. (1987) Experimental studies on thallium toxicity in rats. I. Localization and
elimination of thallium after oral acute and sub-acute intoxication. J Toxicol Clin Exp 7(4):247-257.
Leonard, A; Gerber, GB. (1997) Mutagenicity, carcinogenicity and teratogenicity of thallium compounds. Mutation
Research 387:47-53.
Leopold, G; Furukawa, E; Forth, W; et al. (1968) [Comparative studies of the absorption of heavy metals in vivo
and in vitro]. Naunyn Schmiedebergs ArchPharmacol263(l):275-276. (As cited in Elsenhans etal., 1991).
Leung, KM; Ooi, VEC. (2000) Studies on thallium toxicity, its tissue distribution and histopathological effects in
rats. Chemosphere 41:155-159.
Lie, R; Thomas, RG; Scott, JK. (1960) The distribution and excretion of thallium-204 in the rat, with suggested
MFCs and abio-assay procedure. Health Phys 2:334-340. (As cited in U.S. EPA, 1991b).
Limos, CL; Ohnishi, A; Suzuki, N; et al. (1982) Axonal degeneration and focal muscle fiber necrosis in human
thallotoxicosis: histopathological studies of nerve and muscle. Muscle Nerve 5:598-706.
Lohmann, H; Wiegand, H. (1996) Reduced probability of orthodromically evoked action potential firing in CA1
pyramidal cells of guinea pig hippocampal slices after acute thallium exposure. Arch Toxicol 70:430-439.
Lohmann, H; Csicsaky, M; Wiegand, H. (1989) The action of thallium on the excitability of CA1 pyramidal cells in
hippocampal slices. Neurotoxicol Teratol 11:545-549.
Loveless, LE; Spoerl, E; Weisman, TH. (1954) A survey of effects of chemicals on division and growth of yeast and
Escherichia coli. J Bacteriol 68:637-644.
Lu, CI; Huang, CC; Chang, YC; et al. (2007) Short-term thallium intoxication. Arch Dermatol 143:93-98.
Ludolph, A; Elger, CE; Sennhenn, R; et al. (1986) Chronic thallium exposure in cement plant workers: clinical and
electrophysiologicaldata. TraceElemMed3:121-125.
Lund, A. (1956) Distribution of thallium in the organism and its elimination. Acta Pharmacol Et Toxicol 12:251-
259.
Manzo, L; Scelsi, R; Moglia, A; et al. (1983) Long-term toxicity of thallium in the rat. In: Brown, SS; Savory, J;
eds. Chemical toxicology and clinical chemistry of metals. London: London Academy Press; pp. 401-405.
Marcus, RL. (1985) Investigation of a working population exposed to thallium. J Soc Occup Med 35:4-9.
Matsuzawa, T; Nomura, M; Unno, T. (1993) Clinical pathology reference ranges of laboratory animals. J Vet Med
Sci 55:351-362.
Mourelle, M; Favari, L; Amezcua, JL. (1988) Protection against thallium hepatotoxicity by silymarin. J Appl
Toxicol 8(5):351-354.
MRI (Midwest Research Institute). (1988) Toxicity of thallium (I) sulfate (CAS No. 7446-18-6) in Sprague-Dawley
rats. Vol. 2. Subchronic (90-day) study [revised final report]. Prepared by Dynamac Corporation, Rockville, MD,
for the Office of Solid Waste, Washington, DC; Project No. 8702-L(18); Work Assignment No. 111148-008. [An
external peer review was conducted by EPA in November 2006 to evaluate the accuracy of experimental procedures,
results, and interpretation and discussion of the findings presented. A report of this peer review is available through
the EPA's IRIS Hotline, at (202) 566-1676 (phone), (202) 566-1749 (fax), or hotline.iris@epa.gov (e-mail address)
and at www.epa.gov/iris.]
Mulkey, JP; Oehme, FW. (1993) A review of thallium toxicity. Vet Human Toxicol 35(5):445-454.
83
-------
Munch, JC. (1934) Human thallotoxicosis. J Am Med Assoc 102:1929-1934.
Navas-Acien, A; Silbergeld, EK; Sharrett, AR; et al. (2005) Metals in urine and peripheral arterial disease. Environ
Health Perspect 113:164-169.
NLM (National Library of Medicine). (1998) Thallium. HSDB (Hazardous Substances Data Bank). National
Institutes of Health, U.S. Department of Health and Human Services, Bethesda, MD. Available online at
http://toxnet.nlm.nih.gov.
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press.
Oh, HS; Smart, RC. (1996) An estrogen receptor pathway regulates the telogen-anagen hair follicle transition and
influences epidermal cell proliferation. Proc Natl Acad Sci USA 93(22): 12525-12530.
Osorio-Rico, L; Galvan-Arzate, S; Rios, C. (1995) Thallium increases monoamine oxidase activity and serotonin
turnover rate in rat brain regions. Neurotoxicol Teratol 17(1): 1-5.
Pearson, RG. (1963) Hard and soft acids and bases. J Amer Chem Soc 85:3533-3539.
Petterino, C; Argentino-Storino, A. (2006) Clinical chemistry and haematology historical data in control Sprague-
Dawley rats from pre-clinical toxicity studies. Exp Toxicol Pathol 57:213-219.
Reed, D; Crawley, J; Faro, SN; et al. (1963) Thallotoxicosis. JAMA 183(7):516-522.
Rios, C; Galvan-Arzate, S; Tapia, R. (1989) Brain regional thallium distribution in rats acutely intoxicated with
T12SO4. Arch Toxicol 63:34-37.
Roby, DS; Fein, AM; Bennet, RH; et al. (1984) Cardiopulmonary effects of acute thallium poisoning. Chest
85:236-240.
Rossi, F; Marrazzo, R; Berrino, L; et al. (1988) Prenatal and postnatal thallium exposure in rats: effect on
development of vasomotor reactivity in pups. Teratog Carcinog Mutagen 8:13-23.
Rusyniak, DE; Furbee, RB; Kirk, MA. (2002) Thallium and arsenic poisoning in a small Midwestern town. Ann
EmergMed39(3):307-311.
Sabbioni, E; Goetz, L; Marafante, E. (1980a) Metabolic fate of different inorganic and organic species of thallium in
the rat. Sci Total Environ 15:123-135.
Sabbioni, E; Marafante, E; Rade, J; et al. (1980b) Metabolic patterns of low and toxic doses of thallium in the rat.
Dev Toxicol Environ Sci 8:559-564.
Saddique, A; Peterson, CD. (1983) Thallium poisoning: a review. Vet Hum Toxicol 25:16-22.
Saha, A; Sadhu, HG; Karnik, AB; et al. (2004) Erosion of nails following thallium poisoning: a case report. Occup
Environ Med 61:640-642.
Schaller, KH; Manke, G; Raithel, HJ; et al. (1980) Investigation of thallium-exposed workers in cement factories.
Int Arch Occup Environ Health 47:223-231.
Schoer, J. (1984) Thallium. In: Hutzinger, O; ed. The handbook of environmental chemistry. Vol. 3. Anthropogenic
compounds. Part C. Berlin: Springer-Verlag; pp. 143-214. (As cited in CalEPA, 1999).
Schwartzman, RM; Kirschbaum, JO. (1961) The cutaneous histopathology of thallium poisoning. J Invest Dermatol
39:169-173.
84
-------
Sharma, AN; Nelson, LS; Hoffman, RS. (2004) Cerebrospinal fluid analysis in fatal thallium poisoning: evidence
for delayed distribution into the central nervous system. Am J Forensic Med Pathol 25:156-158.
Shaw, PA. (1933) Toxicity and deposition of thallium in certain game birds. J Pharmacol Exp Ther 48:478-487.
(As cited in U.S. EPA, 199Ib)
Singh, I. (1983) Induction of reverse mutation and mitotic gene conversion by some metal compounds in
Saccharomyces cerevisiae. MutatRes 117:149-152.
Talas, A; Wellhoner, HH. (1983) Dose-dependency of Tl kinetics as studied in rabbits. Arch Toxicol 53:9-16.
Talas, A; Pretschner, DP; Wellhoner, HH. (1983) Pharmacokinetic parameters for thallium (I) ions in man. Arch
Toxicol 53:1-7.
Tao, Z; Gameiro, A; Grewer, C. (2008) Thallium ions can replace both sodium and potassium ions in the glutamate
transporter excitatory amino acid carrier I. Biochemistry 47:12923-30.
Thomas, ML; McKeever, PJ. (1993) Chronic thallium toxicosis in a dog. J Am Anim Hosp Assoc 29:211-215.
Tsai, Y-T; Huang, C-C; Kuo, H-C; et al. (2006) Central nervous system effects in acute thallium poisoning.
Neurotoxicology 27:291-295.
U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
mixtures. Federal Register 51(185):34014-34025. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (1986b) Guidelines for mutagenicity risk assessment. Federal Register 51(185):34006-34012. Available
online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (1986c) Subchronic (90-day) toxicity of thallium (I) sulfate in Sprague-Dawley rats. Prepared by the
Midwest Research Institute, Kansas City, MO for the Office of Solid Waste, Washington, DC.
U.S. EPA (1988) Recommendations for and documentation of biological values for use in risk assessment.
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment, Cincinnati, OH;
EPA/600/6-87/008. Available from the National Technical Information Service, Springfield, VA, PB88-179874/AS,
and online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
U.S. EPA (1991a) Guidelines for developmental toxicity risk assessment. Federal Register 56(234):63798—63826.
Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (1991b) Drinking water health advisory for thallium. Office of Water, Washington, DC. Available from
the National Technical Information Service, Springfield, VA; PB92-135524.
U.S. EPA (1994a) Interim policy for particle size and limit concentration issues in inhalation toxicity: notice of
availability. Federal Register 59(206):53799. Available online at http://www.epa.gov/EPA-
PEST/1994/October/Day-26/pr-ll.html.
U.S. EPA (1994b) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry. Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH; EPA/600/8-90/066F. Available from the National Technical Information Service, Springfield, VA,
PB2000-500023, and online at http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=71993.
U.S. EPA (1995) Use of the benchmark dose approach in health risk assessment. Risk Assessment Forum,
Washington, DC; EPA/630/R-94/007. Available from the National Technical Information Service, Springfield, VA,
PB95-213765, and online athttp://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (1996) Guidelines for reproductive toxicity risk assessment. Federal Register 61(212):56274-56322.
Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
85
-------
U.S. EPA (1997) Exposure factors handbook. National Center for Environmental Assessment, Office of Research
and Development, Washington, DC; EPA/600-P-95-002Fa,b,c. Available online at
http://www.epa.gov/ncea/pdfs/efh/front.pdf.
U.S. EPA (1998) Guidelines for neurotoxicity risk assessment. Federal Register 63(93):26926-26954. Available
online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (2000a) Science policy council handbook: risk characterization. Office of Science Policy, Office of
Research and Development, Washington, DC; EPA/100-B-00-002. Available online at
http://www.epa.gov/OSA/spc/pdfs/prhandbk.pdf.
U.S. EPA (2000b) Benchmark dose technical guidance document [external review draft]. Risk Assessment Forum,
Washington, DC; EPA/630/R-00/001. Available online at http://cfpub.epa.gov/ncea/cfm/
nceapublication.cfm?ActType=PublicationTopics&detype=DOCUMENT&subject=BENCHMARK+DOSE&subjty
pe=TITLE&excCol=Archive.
U.S. EPA (2000c) Supplementary guidance for conducting health risk assessment of chemical mixtures. Risk
Assessment Forum, Washington, DC; EPA/630/R-00/002. Available online at
http://cfpub.epa. gov/ncea/raf/chem_mix.cfm.
U.S. EPA (2002) A review of the reference dose concentration and reference concentration processess. Risk
Assessment Forum, Washington, DC; EPA/630/P-02/002F. Available online at
http://cfpub.epa. gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (2005a) Guidelines for carcinogen risk assessment. Federal Register 70(66): 17765—18717. Available
online at http://www.epa.gov/cancerguidelines.
U.S. EPA (2005b) Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens. Risk
Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available online at
http://www.epa.gov/cancerguidelines.
U.S. EPA (2006a) Science policy council handbook: peer review. 3rd edition. Office of Science Policy, Office of
Research and Development, Washington, DC; EPA/100/B-06/002. Available online at
http://www.epa.gov/OSA/spc/2peerrev.htm.
U.S. EPA (2006b) A framework for assessing health risk of environmental exposures to children. National Center
for Environmental Assessment, Washington, DC; EPA/600/R-05/093F. Available online at
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=158363.
U.S. EPA (2007) Benchmark dose software (BMDS) version 1.4.1. (last modified February 2007). Available online
at http://www.epa.gov/ncea/bmds/index.html.
Valentin, J; ed. (2003) Basic anatomical and physiological data for use in radiological protection: reference values.
ICRP publication 89. Oxford; New York, NY: Pergamon Press.
Verstraeten, SV. (2006) Relationship between thallium(I)-mediated plasma membrane fluidification and cell
oxidants production in Jurkat T cells. Toxicology 222(l-2):95-102.
Waters, CB; Hawkins, EC; Knapp, DW. (1992) Acute thallium toxicosis in a dog. JAMA 201(6):883-885.
Wei, Q. (1987) Studies on spermotoxicity of thallium carbonate in drinking water and its effect on reproductive
function of mice. Zhonghua Yu Fang Yi Xue ZaZhi 21(3): 141-143.
Welch, JM; Lu, J; Rodriguiz, RM; et al. (2007) Cortico-striatal synaptic defects and OCD-like behaviours in
SapapS-mutant mice. Nature 448:894-900.
Wiegand, H; Papadopoulos, R; Csicsaky, M; et al. (1984) The action of thallium acetate on spontaneous transmitter
release in the rat neuromuscular junction. Arch Toxicol 55:253-257.
86
-------
Wiegand, H; Uhlig, S; Gotzsch, U; et al. (1990) The action of cobalt, cadmium and thallium on presynaptic currents
in mouse motor nerve endings. Neurotoxicol Teratol 12:313-318.
Windebank, AJ. (1986) Specific inhibition of myelination of lead in vitro; comparison with arsenic, thallium and
mercury. ExperNeurol 94:203-212.
Woods, JS; Fowler, BA. (1986) Alteration of hepatocellular structure and function by thallium chloride:
Ultrastructural, morphometric, and biochemical studies. Toxicol Appl Pharmacol 83:218-229.
Yokoyama, K; Araki, S; Abe, H. (1990) Distribution of nerve conduction velocities in acute thallium poisoning.
Muscle Nerve 13:117-120.
Yoshida, M; Igeta, S; Kawashima, R; et al. (1997) Changes in adenosine triphosphate (ATP) concentration and its
activity in murine tissues after thallium administration. Bull Environ Contam Toxicol 59(2):268-273.
Zasukhina, GD; Vasilyeva, IM; Sdirkova, NI; et al. (1983) Mutagenic effect of thallium and mercury salts on rodent
cells with different repair activities. MutatRes 124(2): 163-173.
87
-------
APPENDIX A
Summary of External Peer Review and Public Comments and Disposition
The Toxicological Review of Thallium and Compounds (U.S. EPA, 2009) has undergone
a formal external peer review performed by scientists in accordance with EPA guidance on peer
review (U.S. EPA, 2006a, 2000a). The external peer reviewers were tasked with providing
written answers to general questions on the overall assessment and on chemical-specific
questions in areas of scientific controversy or uncertainty. A summary of significant comments
made by the external reviewers and EPA's responses to these comments arranged by charge
question follow. In many cases the comments of the individual reviewers have been synthesized
and paraphrased in development of Appendix A. EPA also received scientific comments from
the public. These comments and EPA's responses are included in a separate section of this
appendix.
EXTERNAL PEER REVIEW PANEL COMMENTS
EPA revised the Toxicological Review to present dose-response analyses for two data sets
from MRI (1988), hair follicle atrophy and clinical observations, and derived candidate RfDs
taking the peer reviewers recommendations into account. However, as pointed out by the peer
reviewers, the MRI (1988) study that was selected as a principal study suffers from certain
critical limitations (e.g., high background incidence of alopecia, lack of histopathological
examination of skin tissue in low- and mid-dose groups, and inadequate examination of objective
measures of neurotoxicity), and there are particular difficulties in the selection of appropriate
endpoints. For these reasons, an RfD for soluble thallium salts was not derived in this specific
case. Text explaining the basis for EPA's decision was added to Section 5.1.3.1. This decision
is considered consistent with the peer reviewers observations that the toxicity database for
thallium is weak. EPA has provided responses to the reviewers' comments in detail below,
because the derivations of the candidate RfDs were retained in the assessment.
The reviewers made several editorial suggestions to clarify specific portions of the text.
These changes were incorporated in the document as appropriate and are not discussed further.
I. General Comments
1. Is the Toxicological Review logical, clear, and concise? Has EPA accurately, clearly, and
objectively represented and synthesized the scientific evidence for noncancer and cancer
hazard?
Comments: Four reviewers considered the Toxicological Review to be clearly written and/or to
present a critical evaluation of the published literature. One of these reviewers stated that despite
A-l
-------
considerable shortcomings in the overall database for thallium (which are acknowledged in the
report), the authors of the review did a good job at synthesizing the available information and
presenting it in a logical fashion. One reviewer observed that the synthesis was somewhat less
than ideal in that it was interjected with editorial comments (some from original authors) and that
it might have been more desirable to present the facts and leave the interpretation to the reader.
This reviewer also believed that the ToxicologicalReview did not provide a sufficient review of
the foreign language material or of environmental surveys. One reviewer stated that the reasons
for EPA's reevaluation of the data used as the basis for the RfD currently on IRIS and the
different conclusion regarding the NOAEL and LOAEL from the MRI (1988) study were not
fully explained. This reviewer did not consider the comparison of the LOAEL in the MRI study
with LOAELs from other studies that were three- to ninefold higher to be a "logical argument."
One reviewer noted that EPA's case for thallium-induced alopecia could have been
strengthened if the Toxicological Review had included a discussion of the hair growth cycle in
rats and how a clinical condition such as alopecia can be distinguishable from hair loss
associated with normal growth cycle in rats. One reviewer stated that the MRI (1988) study
could have been better described and evaluated. Two reviewers criticized the Toxicological
Review for rejection of studies (including Zasukhina et al. [1983] and Wei [1987]) that may have
provided useful information for reasons that were inadequate. One reviewer noted that the title
Toxicological Review of Thallium and Compounds was a bit unclear and confusing and
suggested as potential alternatives: Toxicological Review of Thallium and Its Salts or
Toxicological Review of Thallium and Thallium Compounds.
Other comments were provided in response to this general charge question that were
repeated in response to other more specific charge questions. In these cases, the peer reviewer
comments (and EPA's responses) were summarized under the more specific charge question.
Response: The general structure of a toxicological review is to present a factual summary of
toxicity studies in Section 4 and critical interpretation/synthesis in Section 5. The study authors'
conclusions were presented in Sections 4 and 5 only where necessary (e.g., where it was
necessary to document that EPA's interpretation of the study findings differed from the
authors'.) The comment regarding insufficient consideration of the foreign language literature is
addressed under general charge question 2.
The health effects of thallium exposure were reevaluated in light of the availability of
new data and the need for an up-to-date assessment for regulatory purposes. The comment
related to determination of the NOAEL and LOAEL from MRI (1988) is considered further
under charge question A3. Comparison to LOAELs from other studies was removed.
Alopecia was one of a number of potential critical effects considered for dose-response
analysis. The POD derived from alopecia using BMD modeling was consistent with the PODs
derived from other related clinical effects. EPA did not consider a discussion of the hair growth
A-2
-------
cycle in rats and how alopecia can be distinguished from normal cyclical hair loss to be
necessary.
The Toxicological Review was revised to include an expanded discussion of the findings
from MRI (1988) (see Section 4.2.1.1).
The studies by Zasukhina et al. (1983) and Wei (1987) were critically reevaluated. EPA
again reached the conclusion that the Zasukhina et al. (1983) study suffered from critical
deficiencies and did not provide a scientifically defensible basis for the oral RfD. More thorough
justification for this determination was added to Section 4.3. The discussion of the Wei (1987)
study was revised (Section 4.3 and entry in Table 4-5). In light of the uncertainties associated
with incomplete reporting and reporting inconsistencies and lack of confirmation of these low-
exposure findings on sperm, the Wei (1987) study was not used as the principal study (see
discussions in Sections 4.6.1 and 5.1.1). Nevertheless, the study was included in Section 5.1.5,
Candidate RfD Comparison Information. See also the response to charge question A.I for
additional response to this comment related to consideration of other studies.
The title, Toxicological Review of Thallium and Compounds, was retained for
consistency with other similar assessments on the IRIS database.
2. Please identify any additional studies that should be considered in the assessment of the
noncancer and cancer health effects of thallium and thallium compounds.
Comments: Three reviewers did not identify any additional studies that should be included in
the assessment. One of these reviewers noted that a rapid Medline search identified several
articles dealing with thallium toxicity that were not cited in the Toxicological Review; however,
this reviewer did not review the studies to determine whether they provided important new
additional information. Other reviewers identified the following thallium studies that were not
cited in the Toxicological Review:
• Cavanagh, JB. (1988) Book chapter in Recent Advances in Nervous System Toxicology, Vol. 100, pp.
177-202 Galli, GL; Manzo, L; Spencer, PS, eds. New York: Plenum Press.
• Kazantzis, G. (2007) Thallium. In: Nordberg, GF; Fowler, BA; Nordberg, M; et al.; eds. Handbook on the
toxicology of metals. New York, NY: Elsevier; pp. 827-837.
• Granero, S; Domingo, JL. (2002) Levels of metals in soils of Alcala de Henares, Spain: human health
risks. Environ Int 28(3): 159-164.
• Kennedy, P; Cavanagh, JB. (1977) Sensory neuropathy produced in the cat with thallous acetate. Acta
Neuropathol 39:81-88.
• Windebank, AJ. (1986) Specific inhibition of myelination of lead in vitro; comparison with arsenic,
thallium and mercury. Exp Neurol 94:203-212.
One of the reviewers noted that the foreign literature may be very significant. This
reviewer identified the following papers and suggested these and other foreign language
literature be translated and included in the review:
• Kamil'dzhanov, AKh. (1993) [Experimental substantiation of maximum permissible concentration of
thallium carbonate in environmental air]. Gig Sanit 5:8-10.
A-3
-------
• Gerasimova, IL. (1991) [Establishment of MPEL for thallium iodide activated cesium iodide in the
working zone air]. Gig Tr Prof Zabol 1:31.
• Viereck, L; Kramer, M; Eikmann, T; et al. (1990) [Determining guidelines for metals in children's
playgrounds in North Rhine-Westphalia]. Offentl Gesundheitswes 53(1):7-15.
• Krasovskii, GN; Kenesariev, UI. (1984) [Methodological outline for the experimental substantiation of a
system of indices of the adverse effect of metals on the health status of the population (the example of
thallium)]. Gig Sanit 2:22-25.
• Zasukhina, GD; Vasil'eva, IM; Sdirkova, NI.(1980) [Approach to the determination of the mutagenic
potential of environmental pollutants with the example of detecting the mutagenic action of thallium
carbonate]. Dokl Akad Nauk SSSR 250(3):766-768.
One reviewer noted that "over-grooming" in any species may be a sign of stress, pain, or
other changes in the central or peripheral nervous systems and provided the following references
that address this issue:
• Greer, JM; Capecchi, MR. (2002) HoxbS is required for normal grooming behavior in mice. Neuron
33:23-34.
• Kalueff, AV; Tuohimaa, P. (2005) Contrasting grooming phenotypes in three mouse strains markedly
different in anxiety and activity (129S1, BALB/c and NMRI). Behav Brain Res 160:1-10.
• Kalueff, AV; Minasyan, A; Keisala, T; et al. (2006) Hair barbering in mice: implications for
neurobehavioral research. Behav Proc 71:8-15.
• Welch, JM; Lu, J; Rodriguiz, RM; et al. (2007) Cortico-striatal synaptic defects and OCD-like behaviours
in SapapS-mutant mice. Nature 448:894-900.
In addition, one reviewer noted that some of the cited references that include
experimental studies using more than the LD50 values have little significance to the toxicity of
thallium because of the high dose and suggested deleting most of them from the Toxicological
Review.
Response: Information from the papers on thallium toxicity by Cavanagh and colleagues,
Windebank, and Kazantzis was added to relevant sections of the Toxicological Review. A
discussion of the evidence that over-grooming may be indicative of effects on the nervous
system was also added, including citations to the papers identified above.
The paper by Granero and Domingo (2002) presented a risk assessment for 12 metals in
soil in a region of Spain. For thallium, the EPA RfD previously on IRIS was used in the risk
assessment. No further discussion of health effects information for thallium was presented.
Three of the studies from the foreign literature identified by one reviewer appear to
address the derivation of regulatory limits (Kamil'dzhanov, 1993; Gerasimova, 1991; Viereck et
al., 1990); limits developed by other regulatory agencies are generally not included in IRIS
toxicological reviews. Therefore, these papers from the foreign language literature were not
considered useful additions to the thallium toxicological review. The paper (in Russian) by
Zasukhina et al. (1980) appears to be the same as the Zasukhina et al. (1983) paper published in
English.
In general, a toxicological review is not intended to be a comprehensive treatise (see
Foreword to the thallium Toxicological Review}. EPA acknowledges that additional literature on
A-4
-------
thallium may be available; however, EPA considers all literature that has a bearing on the
derivation of toxicity reference values to have been included in the Toxicological Review.
3. Please discuss research that you think would be likely to increase confidence in the database
for future assessments of thallium and compounds.
Comments: Several reviewers noted that the toxicity database for thallium is very limited.
Research suggested by the reviewers to increase the confidence in the thallium database included
the following study types and suggestions for investigation of specific endpoints:
• Acute dose-range finding studies and, if needed, a chronic inhalation study to identify the
hazard of inhaled thallium.
• A chronic bioassay that would provide information on chronic noncancer effects and
cancer endpoints. (One reviewer noted that such studies are expensive and time-
consuming, and sources of funding may be difficult to find.)
• Studies of general reproductive toxicity (in particular to improve the ability to interpret
the results in Kunming mice observed by Wei).
• Studies of developmental toxicity and developmental neurotoxicity.
• A modern in vivo genotoxicity evaluation. Depending on the outcome of a genotoxicity
assessment, the value of conducting a lifetime chronic bioassay could be weighed.
• A much larger human investigation of an exposed population. (The reviewer noted that a
human investigation would be very costly and that an animal model might be more
practical.)
• Endpoints that the various reviewers suggested would be better characterized in studies of
subchronic or chronic duration included alopecia and its sequelae and neurological,
reproductive, endocrine, and cardiovascular endpoints.
• Additional clinical chemistry, functional, and histopathologic assessments to determine
the source and intensity of the clinical chemistry changes observed in the MRI (1988)
subchronic study.
• Studies that more accurately define dose-response relationship for males and females
with appropriate dose ranges that capture the alopecia endpoint, including supportive
histopathologic examination of alopecia skin and apparently unaffected skin areas of all
treated and untreated animals and supportive in vivo and in vitro experiments
demonstrating thallium interaction with hair follicles at various stages of the hair cycle in
both males and females.
• More information on the interaction of thallium with trace elements such as selenium and
potassium.
A-5
-------
• More studies on the potential mechanisms of action of thallium, including alopecia and
neurotoxicity (as perhaps the most economical way of addressing uncertainties about the
target organs for thallium and the plausibility of some of the responses that have been
observed).
• Studies to establish whether absorption, distribution, and elimination of thallium are
linear following oral and dermal absorption of thallium salts.
• Studies on metabolism of thallium that address the question as to whether the body has
the capability to convert thallium from one valence state to the other to determine the
extent to which studies on monovalent thallium are relevant to the toxicity of trivalent
thallium. (The reviewer noted that if there is little or no conversion, then it will be
important to characterize the toxicity of trivalent thallium separately.)
• Pharmacokinetic studies that would support the development of a physiologically based
pharmacokinetic model for thallium (to better understand thallium dosimetry) as well as
relative toxicity and pharmacokinetics of the various thallium salts.
Response: No response required.
4. Please comment on the identification and characterization of sources of uncertainty in
Sections 5 and 6 of the assessment document. Please comment on whether the key sources
of uncertainty have been adequately discussed. Have the choices and assumptions made in
the discussion of uncertainty been transparently and objectively described? Has the impact
of the uncertainty on the assessment been transparently and objectively described?
Comments: Three reviewers observed that the sources of uncertainty were adequately addressed
and that the discussion of uncertainty was comprehensive. One reviewer considered alopecia to
be an acceptable marker of thallium toxicity until better markers become available, at least for
female rats, but further noted that it may be difficult to use alopecia as a reliable marker in male
rats in the absence of a statistically significant dose response effect for alopecia in males.
One reviewer observed that the very limited data set for thallium can be further
appreciated by comparing the proposed RfD of 1 x 1CT5 mg/kg-day (0.01 |ig/kg-day) to
background exposure to environmental levels of thallium. This RfD suggests that a 70 kg adult
can consume 0.7 jig/day. The 90th percentile adult urinary thallium elimination from NHANES
is 0.380-0.390 |ig/L, which is about 0.760-0.780 |ig/day in the urine alone. Because thallium
has substantial fecal elimination, this suggests that, if the RfD is adopted, greater than 10% of
Americans, and in reality probably close to 50% of Americans, would ingest more than the RfD.
Because there is no evidence that thallium in the current U.S. diet poses any threat and because
there is no possible remediation, this reviewer noted that adoption of this RfD would produce
unnecessary concern if the above calculation were correct. This reviewer further noted that a
A-6
-------
more thorough use of the MRI (1988) study would probably further lower the RfD, suggesting
possibly that the entire U.S. population is exposed above threshold. This reviewer considered
such an analysis to be an example of how poor the existing data are and further questioned the
validity of the analysis.
Five reviewers agreed that the Toxicological Review correctly characterized confidence in
the RfD for thallium as low and/or that the thallium database is weak. In light of this low
confidence, one of these reviews offered the opinion that it may be better to suggest a range of
RfD values for thallium by using various UFs because of uncertain endpoint, inadequate
scientific data, and high UFs. A second reviewer stated that, in light of the weakness of the data
set, it is important to openly discuss in the document the pros and cons of calculating an IRIS
RfD for thallium. A third reviewer noted that the available data are so poor that it was
questionable whether such an analysis is even valid. A fourth reviewer considered the low
confidence to be a reflection of the limited database, including the lack of studies addressing the
known toxic effects of thallium, particularly neurotoxicity, developmental toxicity, and
endocrine effects, failure of the MRI (1988) study to identify aNOAEL if all relevant endpoints
are considered, and consideration of other studies rejected by EPA that suggest that effects may
occur at lower doses than in the MRI study.
Other comments offered on specific UFs are summarized under charge question A5 along
with EPA's response.
Response: In response to the reviewer who suggested that different UFs be used for males and
females, EPA observes that the overall thallium database does not suggest appreciable gender
differences in response to thallium (see Section 4.8.2) and that the development of gender-
specific reference values is not supported.
EPA disagrees with the suggestion that a range of RfD values be developed by using
various UFs because of uncertain endpoint, inadequate scientific data, and high UFs. Comments
from the peer reviewers in response to charge questions A5 and A6 provide little agreement on
alternative UFs. EPA considers the limitations in the database for thallium not to support an UF
smaller than 3,000, and therefore a range of RfD values based on different composite UFs is not
scientifically justified.
A summary of NHANES biomonitoring data was added to Section 3.4. Because an
assessment of potential exposure is generally outside the scope of an IRIS assessment, a
discussion of estimated intakes from background sources of thallium was not included in the
Toxicological Review.
A-7
-------
II. Chemical-Specific Charge Questions
A. Oral Reference Dose (RfD) for Thallium
Al. The 90-day oral gavage study by MRI (1988) was selected as the basis for the RfD. Please
comment on whether the selection of this study as the principal study has been
scientifically justified. Has this study been transparently and objectively described in the
document? Please identify and provide the rationale for any other studies that should be
selected as the principal study.
Comments: Four reviewers generally agreed with the selection of the MRI (1988) study as the
most appropriate basis for the RfD. One reviewer noted that this study may be an acceptable
basis for the RfD, although this was not clear without further analysis of other studies. One other
reviewer raised significant concerns about the study, including the fact that the choice of animal
model, the doses given, and the duration of investigation were not well substantiated; sample
size, statistical analysis, and calculations were not presented; and there was lack of publication in
a peer-reviewed journal. This reviewer did not identify an alternative study to use as the basis
for the RfD.
Several reviewers recommended that the summary of the MRI (1988) study be expanded
to include additional study details. One reviewer suggested that the ToxicologicalReview
provide a better time line for the appearance of alopecia as well as hair loss associated with the
normal hair cycle in the rat and a better description of the distribution pattern of alopecia. Four
reviewers indicated that other toxicological endpoints (including LDH activity, exophthalmos,
lacrimation, and miosis) needed to be better presented/explained.
One reviewer noted that the Toxicological Review considered one or two other studies for
use as the principal study and that a good rationale was provided as to why they were not as good
a choice as the MRI study. A second reviewer stated that he was not aware of any other study
that could have been used as a primary study to derive the POD and the RfD but, given the
paucity of data available, suggested that other studies could be analyzed more in depth and
information compared to that obtained from the MRI study. Three reviewers indicated that
EPA's justification for excluding other in vivo studies as the principal studies were not
sufficient. One of these reviewers recommended that RfDs be derived from the other endpoints
identified in the MRI study and from the Wei (1987) and Zasukhina et al. (1983) studies and the
results compared to determine where the MRI study lies with respect to sensitivity. One
reviewer observed that epidemiologic studies would be ideal, but all available epidemiologic
studies provided limited evidence of cause and effect associated with exposure in food or
inhalation.
A-8
-------
Response: Consistent with the feedback from the majority of reviewers, the MRI (1988) study
was retained as a potential principal study. This study was subject to an independent peer review
(as noted in Section 4.2.1.1). None of these independent reviewers considered the study to be
inadequate as the basis for the RfD.
The summary of the MRI (1988) findings, including LDH activity, exophthalmos,
lacrimation, and miosis, was expanded in the ToxicologicalReview, Section 4.2.1.1. Information
on the time that alopecia was first observed was also added to this section based on examination
of individual animal observational findings in the MRI (1988) study. Across control and treated
groups, alopecia was first observed somewhere between study days 44 and 60. Severity scores
across control and treated groups ranged from 1 to 5. There were no discernable differences in
either the severity or distribution pattern of alopecia across control and treated groups.
Further consideration was given to the use of other in vivo studies as the basis for the
RfD (see Sections 5.1.1 and 5.1.5). Where certain in vivo studies were not considered by EPA to
be sufficiently reliable as the basis for the POD, more thorough justification was provided (see
Section 5.1.1). Expanded discussion of the limitations of the Zasukhina et al. (1983) study was
added to Section 4.3. As noted in response to general charge question 1, the discussion of the
Wei (1987) study was revised (Section 4.3), and the study findings were further considered as
the potential basis for the POD in Sections 5.1.1 and 5.1.5. For the reasons discussed in Sections
4.6.1 and 5.1.1, the Wei (1987) study was not considered sufficiently reliable as the basis for the
RfD.
A2. Alopecia (hair loss) was selected as the most appropriate critical effect for the RfD. EPA
characterizes alopecia as being an adverse effect. Please comment on whether the science
and mode-of-action information supports alopecia as an adverse effect. EPA has stated:
"Whether alopecia is itself an adverse effect merits consideration. In humans, alopecia is
generally reversible upon cessation of thallium exposure. Alopecia, however, appears to be
a part of a continuum of dermal changes observed following thallium exposure, as well as
one of a spectrum of effects on target organs that include the nervous and gastrointestinal
systems. For these reasons, alopecia supported by two cases of hair follicle atrophy is
considered an adverse effect." Please comment on whether the selection of this critical
effect has been scientifically justified. Is EPA's choice transparently and objectively
described in the document? Please provide a detailed explanation. Please identify and
provide the rationale for any other endpoints that should be used instead of alopecia to
develop the RfD.
Comments: Three reviewers agreed that alopecia should be considered an adverse effect in and
of itself. One reviewer considered alopecia to be undesirable. Another reviewer considered
alopecia with histopathology of the hair follicle to be adverse. A final reviewer stated that use of
A-9
-------
alopecia "as a biomarker" is justified until a better marker for adverse effect such as neurotoxic
effects is identified. Several reviewers observed that the fact that alopecia may be reversible
does not influence its adversity, whereas one review stated that a drawback to the use of alopecia
as an endpoint for RfD is that it is a reversible effect. Reviewers pointed to other limitations in
the use of alopecia as the critical effect. One reviewer considered limitations to the alopecia data
set to be the facts that histopathology was only noted in the assigned LOAEL dose group and
was not examined in the assigned NOAEL and lower dose groups and that there is no evidence
that the skin sections examined were from the alopecia areas of the skin. A second reviewer
considered a drawback to be that the MRI (1988) study did not show a statistically significant
dose-response effect for alopecia in male rats. One reviewer suggested that the Toxicological
Review could provide a more detailed discussion of mechanisms of alopecia, particularly with
regard to thallium.
Two reviewers disagreed with the use of alopecia as the critical effect for RfD
determination for thallium. One reviewer based this opinion on the fact that it was not clear to
this reviewer from either the description of the study or the peer review of the study as to
whether this effect was really treatment related. This reviewer noted that there was a high
background incidence of alopecia in the study population and that the observation in the study of
"barbering behavior" in the animals may have accounted for much of the alopecia. It was this
reviewer's experience with lab animals that there can be any number of reasons for hair loss,
including changes in caging or husbandry, and the fact that there was such a high background
incidence of alopecia suggested that the effect could have been unrelated to treatment. A second
reviewer disagreed with the choice of alopecia in the MRI (1988) study because several findings
(changes in biochemical parameters, exophthalmos, miosis, and changes in coat) appear to show
better dose-response effects than alopecia.
In response to this and charge question A3, four reviewers suggested that other effects be
considered as the basis for the RfD, including clinical chemistry, exophthalmos, miosis, changes
in coat, lacrimation, and behavioral effects that appear to show dose-response effects and could
be more defensible than alopecia as the critical effect.
One reviewer stated that data from several limited subchronic studies (Wei, 1987;
Formigli et al., 1986) and in vitro studies (Gregotti et al., 1992) suggested that reproductive
(especially male) endpoints are worthy of further investigation. This reviewer noted that these
studies were poorly reported by the authors, and a dose-response relationship was difficult to
ascertain, but that this was "however a poor excuse for not attempting to determine a dose-
response relationship and providing NOAELs and BMDLs." This reviewer pointed to the study
by Galvan-Arzate et al. (2005) that describes a dose-related increase in thallium deposition in the
brain and lipid peroxidation in male Wistar rats after 30 day exposure to 0.8 mg/kg or 1.6 mg/kg
doses. This reviewer also suggested that EPA provide a more critical review of the 36-week oral
A-10
-------
study by Manzo et al. (1983) that described in some detail functional and histopathologic
changes in the peripheral nervous system.
Response: Consistent with the position of three reviewers who agreed that alopecia is an adverse
effect, alopecia was considered along with other clinical observations that showed an increased
incidence with dose (see Table 5-1) as a candidate critical effect for establishing the POD for the
RfD and as part of a spectrum of clinical observations in thallium-exposed rats. The limitations
in the skin histopathology data were recognized (see Sections 4.2.1.1 and 5.1.1). EPA
acknowledges that the incidence of alopecia in male rats does not increase monotonically with
dose, although the incidence is higher in all groups of treated male rats compared with untreated
and vehicle controls. Further, the incidences of other clinical observations were increased in
treated male rats over the control groups, consistent with a treatment-related effect in male rats.
Available information on the mechanism by which thallium induces alopecia is presented in
Sections 4.5.5 and 4.6.3.
Consistent with the input from the reviewers, EPA expanded the analysis of the critical
effect to include other clinical observations that showed a dose-response effect, including
observations related to coat, lacrimation, exophthalmos, miosis, and behavior. For the reasons
presented in Section 5.1.1, changes in clinical chemistry parameters were not considered as the
basis for the POD. EPA acknowledges that other factors such as caging and husbandry can
cause alopecia in laboratory rodents; however, the incidence was clearly elevated in both male
and female rats over controls. Further, to the extent that alopecia was due to barbering, research
has shown that barbering in rodents can reflect a stress-evoked behavioral response (see Section
4.6.1). Accordingly, EPA retained alopecia as one of a set of clinical observations that were
indicative of a potential treatment-related effect in rats. Limitations in clinical observation data
were further discussed in Section 5.1.3.1.
EPA reconsidered the findings from other studies, including Galvan-Arzate et al. (2000),
Gregotti et al. (1992, 1985), Wei (1987), Formigli et al. (1986), and Manzo et al. (1983), as the
basis for the critical effect. As discussed in Section 5.1.1, Formigli et al. (1986), Gregotti et al.
(1985), and Manzo et al. (1983) were all single-dose studies and thus did not provide data useful
for dose-response analysis. In addition, these studies administered doses higher than those used
in the MRI (1988) study and therefore provided less information on low-dose toxicity. In vitro
studies such as Gregotti et al. (1992) and single administration studies such as Galvan-Arzate et
al. (2005) do not provide information appropriate for characterization of long-term oral exposure
to thallium. The Wei (1987) study was considered further as the basis for the POD for the oral
RfD (see Section 5.1.5). In response to one reviewer comment, EPA reexamined the discussion
of the Manzo et al. (1983) study in Section 4.2.1.1 and determined it to be adequate. Given the
increased mortality in thallium-exposed rats at 40 and 240 days after treatment and the fact that
A-ll
-------
only a single dose level was used in this study, Manzo et al. (1983) was not considered
appropriate as a principal study for RfD derivation.
A3. At the high dose in the MRI (1988) study, two female rats exhibited moderate to severe
alopecia that could not be attributed to self-barbering or normal cyclic hair growth.
Histologic examination of skin samples from these high-dose females showed atrophy of
hair follicles. EPA considered these findings to be adverse and thus the high dose in this
study (0.25 mg/kg-day thallium sulfate) to be the lowest-observed-adverse-effect level
(LOAEL). The mid-dose group (0.05 mg/kg-day thallium sulfate) was identified as the no-
observed-adverse-effect level (NOAEL). Is EPA's interpretation of the study findings
scientifically justified? Has this interpretation of the findings been transparently and
objectively described in the document?
As part of the evaluation of alopecia as a critical effect for the RfD, EPA performed a
series of Fisher's exact tests to determine if the incidence of alopecia in any of the three
dose groups was statistically significantly elevated above controls by using all cases of
alopecia reported by MRI (1988). Please comment on whether EPA chose the appropriate
data set and the appropriate statistical test for this analysis.
The study investigators reached a different interpretation of the study findings than did
EPA. The investigators considered alopecia to be attributable to the cyclic pattern of hair
growth in rodents and, consequently, did not consider these findings to be biologically
significant. The high dose (0.25 mg/kg-day thallium sulfate) was identified in the MRI
(1988) study as the NOAEL. Is the study authors' conclusion that the high dose (0.25
mg/kg-day thallium sulfate) represents a NOAEL justified and supported by the study data?
Comments: One reviewer offered the opinion that EPA correctly identified an adverse effect that
cannot be "attributable to the cyclic pattern of hair growth in rodents." This reviewer stated that
EPA needs to clearly define what a NOAEL is versus a LOAEL as it pertains to alopecia; the
Toxicological Review needs to make clear whether a dose of 0.04 mg/kg that caused significant
incidence of alopecia with no evidence of histopathologic change in skin/hair follicle is a
NOAEL. A second reviewer could not fully agree with lowering the NOAEL value to
0.05 mg/kg because it was based on a change in interpretation of the data rather than any new
scientific evidence and because male rats did not show a statistically significant dose response
for alopecia.
One reviewer considered the high dose of thallium (0.2 mg/kg-day), and possibly the
middle dose (0.04 mg/kg), to be a LOAEL rather than a NOAEL. This reviewer also considered
data on the incidence of alopecia not attributable to barbering to be amenable to BMD modeling.
A-12
-------
One reviewer found it difficult to conclude definitively that the alopecia in two animals
with hair follicle atrophy was in fact treatment related. This reviewer noted that hair follicles go
through a natural cycle of activity and inactivity; it was not clear from the descriptions of atrophy
that what was being observed was anything more than an observation of a normal condition in
these animals; there was no systematic evaluation of dermal tissue from other areas in the same
animals; and there was no evaluation of hair follicle status in all dose groups. Because this is not
a standard assessment in subchronic studies, it was not possible to know whether what was
observed is within the range of normal.
One reviewer found the use of alopecia findings to be questionable, noting that the mid-
dose group also had alopecia and, although it may have been a result of barbering, barbering is
not a normal behavior and suggests that the animals were under stress, possibly in pain; thus, this
dose level could not be a NOAEL. This reviewer also stated that it was not valid to discount a
sex difference because sex may affect sensitivity to a given toxin. Finally, this reviewer noted
that other parameters that show a dose-response relationship might be amenable to BMD
analysis.
One reviewer observed that the incidences of alopecia in middle- and high-dose groups
were both significantly different from controls, whereas EPA considered the middle dose to
represent a NOAEL with no explanation. It appeared to this reviewer that EPA considers hair
follicle atrophy to be the adverse effect. Because histopathologically observed atrophy was
examined only at the high dose, this reviewer believed that the high dose should be considered a
LOAEL for hair follicle atrophy and a UF of 10 applied. More importantly, this reviewer
believed that barbering represents a change in behavior (e.g., caused by pain, an increased
arousal level, or stress) and should be considered an adverse effect. This reviewer noted that
barbering increased in a dose-dependent manner. This reviewer considered many of the
endpoints in the rat study to be consistent with neurological effects in humans and that they
should not have been dismissed. This reviewer considered the lowest dose to be a LOAEL.
One reviewer suggested that EPA recognize in their review that there are possible
estrogen receptor pathways within the dermal papilla that regulate the telogen-anagen follicle
transition and that diffusible factors associated with the anagen follicle influence cell
proliferation in the epidermis (Oh and Smart, 1996). This may explain male versus female
differences observed in the MRI study.
Response: In response to peer reviewer comments, EPA broadened the evaluation of the MRI
(1988) study findings beyond alopecia in the presence of hair follicle atrophy to include all
clinical observations in male and female rats and presented the possibility that the dose-related
increase in the incidence of clinical effects indicated a possible treatment-related effect. EPA
identified two possible critical effects from the MRI (1988) study - hair follicle atrophy and
clinical observations - reflecting different interpretations of MRI study findings. With respect to
A-13
-------
hair follicle atrophy, EPA in Section 5.1.2 clarified that skin tissue was not examined for
histopathological changes in the low- and mid-dose groups. The mid-dose group was assumed to
approximate a NOAEL for this endpoint in light of the low incidence of hair follicle atrophy in
high-dose female rats and absence in male rats at all dose levels. EPA acknowledges this to be a
change in the interpretation of the data that was made in light of all relevant literature on
thallium toxicity and considered this to be an appropriate basis for reaching a different
determination of the NOAEL and LOAEL from the authors of the MRI (1988) study.
In the revised discussion of alopecia and other clinical observations, EPA considered the
possibility that barbering may indicate that animals were under stress or that barbering (and other
clinical observations) may reflect some other change in behavior (e.g., see Section 4.6.1).
As suggested by several reviewers, BMD methodology was applied to the various clinical
observation data sets and was used to derive the potential PODs, including the POD for alopecia.
EPA appreciated one reviewer's suggestion of a possible estrogen receptor pathway as an
underlying basis for alopecia; however, there is no evidence in the thallium toxicity literature for
such a mode of action. Because of the lack of data to support this hypothesis, it was not added to
the Toxicological Review.
In Section 4.8.1, EPA concluded that the available thallium toxicity literature as a whole
showed no consistent pattern of gender-related difference in toxicity.
Comments: Two reviewers considered the nonparametric statistical analysis using Fisher's exact
test to be appropriate. One of the reviewers, however, noted that the choice to include all
animals that had alopecia is problematic, especially given that the study investigators did not
attribute all instances of alopecia to thallium treatment. A third reviewer stated that a Fisher's
exact test is a standard procedure for pair-wise comparisons but fails to take advantage of the fact
that there are three dose groups. This reviewer suggested that a more appropriate analysis would
be a trend analysis. A fourth reviewer stated that, whereas a Fisher's exact test is specifically
designed for a simple comparison, a more complex analysis is generally performed in a dose-
response relationship and it is quite likely that there would be no significance had that approach
been taken.
Response: As noted previously, EPA revised the dose-response analysis to include consideration
of a range of clinical observations and used a BMD approach in place of a NOAEL/LOAEL
approach for identifying a POD. Therefore, less importance was placed on pair-wise comparison
as the basis for identifying a NOAEL and LOAEL from the MRI (1988) study.
Comments: Two reviewers specifically commented on the study authors' identification of the
NOAEL. One reviewer stated that the conclusion as to whether the observation of alopecia is an
adverse one is a matter of judgment and that the investigators who wrote the original study report
A-14
-------
had a different interpretation of the alopecia; however, it was less clear to this reviewer why they
concluded that the clinical chemistry observations or exophthalmos were not adverse. A second
reviewer noted that the presence of follicle changes suggests that this was truly a toxic effect and
that barbering represented stress and possibly pain. This reviewer further noted that the study
investigators' interpretation of these data added considerable doubt on their ability to interpret
the other findings in their study and further calls into question the choice of this study as the
critical data for determining the RfD.
Response: EPA disagrees with the reviewer who called into question the choice of the MRI
(1988) study as a potential principal study because of doubt about the ability of the study authors
to interpret study findings. EPA notes that this study was subject to an independent peer review.
The peer reviewers of the MRI (1988) study generally found the study to have been conducted
according to guideline protocols in place at the time. Further, EPA reached its own conclusions
based on an independent review of study findings and did not rely only on the authors'
interpretation of the data. Accordingly, EPA considers the selection of the MRI (1988) study as
a principal study to be supported, although it suffers from some limitations. Limitations in study
design and uncertainties in data sets from MRI (1988) were addressed in Section 5.1.3.1.
Additional comments in response to this charge question, related to the consideration of
other endpoints as critical effects, are summarized under charge question A2.
A4. The traditional NOAEL-LOAEL approach was used to define the point of departure (POD)
for the RfD. A benchmark dose (BMD) analysis was considered but was not conducted
because of the nature of the data set for alopecia. Please provide comments with regards to
whether a NOAEL-LOAEL approach is the best approach for determining the POD. Has
the approach been scientifically justified? Is it transparently and objectively described?
Please identify and provide a rationale for any alternative approaches for the determination
of the POD and if such approaches are preferred to EPA's approach.
Comments: Five reviewers agreed that the NOAEL-LOAEL approach is the most appropriate
method to define the POD if it is derived using data for alopecia. One reviewer stated that, if the
LOAEL-NOAEL approach from the MRI (1988) study is used, the lowest dose should be the
POD as a LOAEL.
Five reviewers noted that there are data sets for other endpoints from the MRI (1988)
study (e.g., exophthalmos, lacrimation, blood chemistry) for which BMD modeling should be
considered. One of these reviewers also noted that the BMD approach should be tried for each
of the endpoints that exhibits a dose-response relationship in the Wei (1987) and Zasukhina et al.
(1983) studies. In this case, the POD used to derive the RfD may be the lowest BMDL, the one
that EPA thinks is the most reliable or most orderly under analysis, or some kind of average.
A-15
-------
One reviewer suggested deriving a range of RfD values rather than a single RfD value.
Response: BMD modeling of all appropriate data sets from the MRI (1988) study was
conducted and presented in Appendix B of the lexicological Review. BMD modeling was not
performed for the Zasukhina et al. (1983) study because, as discussed in the response to charge
question Al comments, the study results were not considered scientifically reliable. The Wei
(1987) study was included in Section 5.1.5, Candidate RfD Comparison Information, and the
LOAEL was considered as a potential POD. Given the uncertainties associated with incomplete
reporting and the lack of confirmatory findings at dose levels close to those used in the Wei
(1987) study, data sets from this study were not analyzed using BMD methods.
As noted in response to general charge question 4 comments, it is EPA's general practice
in developing IRIS reference values to derive a single value and not a range.
A5. Please comment on the selection of the uncertainty factors applied to the POD for the
derivation of the RfD. For instance, are they scientifically justified and transparently and
objectively described in the document? If changes to the selected uncertainty factors are
proposed, please identify and provide a rationale(s).
Comments: None of the reviewers disagreed with the UF applied for interspecies or LOAEL to
NOAEL extrapolation. Two reviewers questioned the need for application of an intraspecies UF.
(One of these comments was offered in response to general charge question 4 but is summarized
here.) In one case, this opinion was based on the fact that metals such as thallium are not
metabolized and that differences in the sensitivity to thallium in humans will therefore be
insignificant. In the second case, the reviewer identified limited dose-response data for human
dermal exposure (resulting from the routine practice of dermal administration of thallium salts to
children with ringworm) and, in particular, 1930s papers from Munch (1934) that demonstrated
the distinction between epilation and toxicity and suggested a possible range for which animal
effects can be correlated with human effects. This reviewer noted that it was unclear whether the
addition of these data would improve uncertainty.
Only one reviewer questioned the subchronic to chronic UF of 3. It was this reviewer's
opinion that there is little or no uncertainty when alopecia occurs following oral exposure to
thallium. The effect occurs in less than a subchronic time frame, and therefore the UF for
subchronic to chronic extrapolation may not be necessary. This reviewer also stated that EPA
needs to counter this argument if they are concerned about toxic effects other than alopecia
following chronic exposure.
Input on the database UF is discussed under charge question A6 below.
A-16
-------
Response: Consistent with the input from the peer reviewers, the UFs for interspecies and
LOAEL to NOAEL extrapolation were retained.
The paper by Munch (1934) was reviewed for information that would support reducing
the uncertainty associated with intraspecies variability. This paper, which generally reported
symptoms of overt toxicity, including death, was determined not to be useful for characterizing
variation in intraspecies sensitivity to environmental exposures to thallium. Accordingly, the
intraspecies UF of 10 was retained.
Consistent with the majority of reviewers, a subchronic to chronic UF of 3 was retained.
One reviewer suggested that a subchronic to chronic UF might not be necessary; however, the
UF of 3 was retained because, in the absence of any chronic toxicity information, it is unknown
whether effects other than alopecia and other clinical observations could manifest at low doses.
A6. Please comment specifically on the database uncertainty factor of 10 applied in the RfD
derivation. Please comment on the use of the database uncertainty factor specifically for
the lack of adequate developmental toxicity studies and a two-generation reproductive
toxicity study and additional uncertainty associated with the limited data available on
neurotoxicity in light of the potential for neurotoxicity to represent a sensitive endpoint for
thallium exposure. Please comment on whether the selection of the database uncertainty
factor for the RfD has been scientifically justified. Has this selection been transparently
and objectively described in the document?
Comments: Three reviewers considered EPA's database UF to be appropriate. One of these
reviewers stated that the absence of specific dose-response studies addressing the issues of
neurotoxicity, reproductive toxicity, and developmental toxicity of thallium suggests that a UF to
account for an incomplete database is appropriate. Use of a database UF was supported by
evidence that neurotoxicity is seen in humans upon (high) exposure to thallium and in animal
studies (e.g., Manzo et al., 1983) and that reproductive toxicity is also seen in animal studies
(Wei, 1987) at doses below those used in the MRI (1988) study. A second reviewer similarly
noted that neurotoxic effects and reproductive effects were not adequately evaluated in the
available animal studies and that available developmental toxicity studies, although not state-of-
the-art, appear adequate to conclude that developmental toxicity would not drive the risk
assessment for thallium. A third reviewer considered the database to be so weak that a UF of 10
must be utilized and that this choice was clearly described and appropriate.
One reviewer stated that a UF of 10 for deficiencies in the thallium toxicity database may
be deemed to be too conservative. Another reviewer noted that a database UF of 10 may not be
necessary if all the endpoints in the MRI (1988) study are considered, some of which may be
indicative of neurotoxicity. This reviewer noted that the lack of a robust database on
developmental toxicity is problematic, that developmental neurotoxicity and endocrine studies
A-17
-------
have apparently not been performed, and that there are some data providing evidence for
reproductive toxicity, but only for one sex. This reviewer suggested that the three studies should
be used to generate sample RfDs from the endpoints that exhibit a dose-effect function; in this
case, a database UF of 3 may be more appropriate.
One reviewer did not take issue with the database UF but proposed changing the single
RfD to a range of RfDs.
Response: Consistent with the opinion of three of the external peer reviewers that the database
UF of 10 is appropriate, the opinion of essentially all of the reviewers that the limited database
for thallium toxicity results in an RfD of low confidence, the lack of adequate developmental
toxicity studies and a two-generation reproductive toxicity study, and the uncertainty associated
with the limited neurotoxicity data in light of the potential for neurotoxicity to represent a
sensitive endpoint for thallium exposure, the full default database UF of 10 was retained. One
reviewer observed that, based on the available studies, developmental toxicity would not likely
drive the risk assessment for thallium. EPA does not believe that this observation should change
the size of the database UF since the database deficiencies associated with reproductive toxicity
and neurotoxicity also support a database UF of 10.
B. Inhalation Reference Concentration (RfC) for Thallium
B1. Has the rationale and justification for not deriving an RfC for thallium been transparently
described in the document?
Comments: Three reviewers agreed that the available data on inhalation exposure to thallium
were insufficient to support derivation of an RfC. One reviewer noted that an Occupational
Safety and Health Administration (OSHA) threshold limit value of 0.10 mg/m3 Tl is available
and should be reported in the Toxicological Review.
Response: IRIS toxicological reviews generally do not include a summary of other Agency
regulatory guidelines and standards. Regulatory values included in an IRIS assessment could
become irrelevant with time as they are updated. Further, OSHA occupational limits take into
consideration technological achievability and economic costs in addition to toxicity and therefore
are not comparable to IRIS reference values. Therefore, the OSHA occupational limit was not
included in the thallium Toxicological Review.
A-18
-------
C. Carcinogenicity of Thallium and Compounds
C1. Under the EPA's (2005a) Guidelines for Carcinogen Risk Assessment
(www.epa.gov/iris/backgr-d.htm), the Agency concluded that there is "inadequate
information to assess the carcinogenic potential" of thallium and compounds. Please
comment on the scientific justification for the cancer weight-of-evidence characterization.
A quantitative cancer assessment was not derived for thallium. Has the scientific
justification for not deriving a quantitative cancer assessment been transparently and
objectively described?
Comments: All six reviewers agreed that the available literature does not support a
characterization of the carcinogen!city of thallium and that the rationale for not performing a
quantitative risk assessment for cancer was appropriate.
Response: No response needed.
PUBLIC COMMENTS
Two submissions from the public were received. The submissions received during the
official public comment period were distributed to the external peer review panel prior to the
public meeting and discussion of the assessment. Submissions offered some editorial comments
and suggestions for clarification of specific portions of the text. Changes were incorporated in
the ToxicologicalReview as appropriate and are not discussed further in this appendix.
A. General Comments
Comment: One public commenter considered thallium to be a very dangerous substance and
urged EPA to retain the IRIS assessment for thallium on the IRIS database.
Response: EPA intends to retain an assessment of the health effects of thallium and compounds
on the IRIS database. The draft Toxicological Review provided for public comment and external
peer review represented an effort to update the health assessment for thallium and thallium
compounds to reflect the current available science on this compound. However, EPA has not
recommended an RfD for thallium compounds in this specific case. Justification for this
decision is provided in Section 5.1.3.1.
A-19
-------
B. Oral Reference Dose (RfD) for Thallium
Comment: One public commenter did not consider the 0.25 mg/kg-day dose to be a biologically
significant LOAEL because there is inadequate evidence to link the high dose alopecia with
biologically significant effects. Specific reasons for failing to establish this link included the
following: (1) complete histopathologic examinations were conducted only in vehicle control
and high-dose groups; therefore, the dose-response relationship of this effect was not established;
(2) background occurrence of alopecia in study animals and the potential for misclassification
adds uncertainty regarding the incidence of treatment-related alopecia in study animals;
(3) patterns of alopecia and hair follicle atrophy were inconsistent between male and female rats;
increases in the incidence of alopecia in male rats were not dose related; and atrophy of the hair
follicles in male rats was absent; and (4) there is incomplete information on, and inconsistencies
in, anatomical patterns of effects.
Response: As discussed in response to peer reviewer comments on charge question A2, EPA
expanded the analysis of potential critical effects to include other clinical observations that
showed a dose-response effect, including observations related to coat, lacrimation,
exophthalmos, miosis, and behavior. Collectively, these effects were considered evidence of a
possible treatment-related effect and to be biologically significant. Uncertainties associated with
these endpoints were also discussed (Section 5.1.3.1).
Comment: One public commenter stated that EPA failed to adequately support its determination
that the study dose selected as the POD constitutes a biologically significant endpoint. This
commenter observed that alopecia should not be considered adverse because it is reversible after
exposure ceases and noted that one reviewer of the MRI (1988) study described alopecia as an
"early (low grade) lesion" and added that "it is not a lesion that could be considered as an
adverse health effect at that stage, and it certainly is not life threatening."
Response: External peer reviewers generally considered alopecia to be an adverse effect. Several
reviewers strongly recommended, however, that EPA consider other effects from the MRI (1988)
study in addition to (or in place of) alopecia because these other effects were either more
sensitive than alopecia, displayed a better dose response, and/or were more clearly treatment
related. A more rigorous dose-response analysis of data sets for these other effects was added to
the Toxicological Review.
Comment: One public commenter observed that available human data do not provide support for
the choice of the LOAEL or NOAEL established by EPA, noting that, although limited in
number and methodology, population-based surveys do not provide evidence for alopecia in
A-20
-------
thallium-exposed populations. This commenter also noted that the draft RfD is inconsistent with
the World Health Organization evaluation of thallium toxicity (presented in IPCS [1996]) that
considers urinary concentrations <5 ug/L (assumed to be associated with a daily intake of 10 ug
thallium) unlikely to cause adverse health effects based on the Brockhaus et al. (1981) study of
residents living near a thallium-emitting cement plant in Germany. The commenter further
observed that thallium exposures up to 100 times greater than the draft RfD did not include
effects consistent with the draft RfD.
Response: EPA considered the use of the Brockhaus et al. (1981) and other population-based
surveys as a basis for the RfD; however, the available studies were not considered sufficiently
rigorous for reference value derivation.
Comment: One commenter considered the draft RfD (0.7 ug/day Tl for a 70 kg adult) to be of
questionable relevance for protection of public health because intake of thallium from background
sources (according to ATSDR [1992] as high as 5 ug/day from food and 2 ug/day from drinking
water) exceeds levels of thallium oral exposure corresponding with the draft RfD. The commenter
noted that, although some of this intake may be in less soluble and less well absorbed forms of
thallium than the soluble salts upon which the draft RfD is based, these forms may better represent
the forms of thallium present in soil and produce at many contaminated sites.
Response: Because the scope of an IRIS assessment does not generally include an analysis of
potential exposure, a discussion of intake of thallium compounds from background sources was
not included in the Toxicological Review. As noted in response to general charge question 4,
EPA is not recommending an RfD value for thallium compounds.
Comment: One public commenter observed that the draft RfD applies only to soluble forms of
thallium and strongly urged EPA to restrict application of the draft RfD for soluble salts to only
those cases where these forms of thallium are known to be present.
Response: EPA is not recommending an RfD for soluble thallium salts. Therefore, restrictions
on the application of an RfD is not an issue at this time.
Comment: One public commenter observed that, although human biomonitoring studies provide
direct evidence of widespread, low-level thallium exposures in the general population,
biomonitoring estimates from these studies describe uptake rather than intake, do not account for
the unabsorbed fraction of the ingested dose, and need to be coupled with a better understanding
of the absorption fraction for dietary thallium and fraction excreted in urine in order to estimate
daily intakes more reliably.
A-21
-------
Response: A summary of NHANES biomonitoring data was added to Section 3.4. Because an
assessment of potential exposure is generally outside the scope of an IRIS assessment, an
analysis of thallium intake based on available biomonitoring data was not included in the
Toxicological Review.
A-22
-------
APPENDIX B
Documentation of Benchmark Dose Modeling
B.I. Summary of BMDS Modeling Results for Clinical Observations in the Male Rat
(MRI, 1988)
Table B-l. A summary of BMDS (version 1.4.1) modeling results based on
incidence of rough coat in male Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0000
0.0000
0.8762
0.0000
0.0000
0.0008
0.0000
AIC
95.1404
104.547
85.7472
95.1404
106.954
92.4736
95.1404
BMD10
(mg/kg-day)
0.003737
0.009731
0.001063
0.003737
0.01361
0.004137
0.003737
BMDL10
(mg/kg-day)
0.002460
0.006365
0.0005675
0.002460
0.009597
0.002679
0.002460
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI (1988).
Only the log-logistic model had ap > 0.1; however, the data set was not considered
amenable to modeling because of the steep slope of the dose-response curve in the low-dose
region and because the BMDio and BMDLio were well outside experimental range of the data.
BMDS output from the log-logistic model is provided below.
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\ROUGH_COAT-MALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\ROUGH_COAT-MALE.pit
Fri Jun 13 08:32:57 2008
The form of the probability function is:
P[response] = background+(1-background)/[l+EXP(-intercept-slope^Log(dose))
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
B-l
-------
Default Initial Parameter Values
background = 0.1
intercept = 4.47108
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
background
intercept
intercept
-0.26
1
Parameter Estimates
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
85.7472
Goodness of Fit
Est._Prob. Expected Observed Size
4
11
16
19
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.00106311
BMDL = 0.000567469
B-2
-------
Log-Logistic Model with 0.95 Confidence Level
0.8
0.2
Log-Logistic
'felj/IDL.BMD
0.05
0.1
dose
0.15
0.2
08:3206/132008
Table B-2. A summary of BMDS (version 1.4.1) modeling results based on
incidence of piloerection in male Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
/7-valuea
0.9976
0.0868
0.9609
0.9976
0.1124
0.3004
0.9976
AIC
55.896
63.7889
57.9327
55.896
63.1776
60.4202
55.896
BMD10
(mg/kg-day)
0.01958
0.06642
0.01813
0.01958
0.0596
0.03242
0.01958
BMDL10
(mg/kg-day)
0.01339
0.04852
0.008735
0.01339
0.04436
0.02167
0.01339
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
All models with the exception of the logistic model provided adequate fits of the data as
assessed by a chi-square goodness-of-fit test (p > 0.1) and visual inspection of the respective
plots of observed versus predicted values from the various models. The gamma, multistage, and
Weibull models provided the same fit of the data and were judged to provide the best model fit
based on the lowest AIC value. The BMDS output for the gamma model is provided below.
B-3
-------
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\PILOERECTION-MALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\PILOERECTION-MALE.pit
Fri Jun 13 09:33:49 2008
BMDS MODEL RUN
The form of the probability function is:
Dependent variable = #Piloerection
Independent variable = Dose(mg/kg-d)
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0121951
Slope = 6.76558
Power = 1.26625
Asymptotic Correlation Matrix of Parameter Estimates
Variable
Background
Slope
Power
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-26.9273 4
-26.948 1 0.0414636 3
-47.1393 1 40.4241 3
P-value
Est. Prob.
0
1
4
13
B-4
-------
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0195751
BMDL = 0.0133909
Gamma Multi-Hit Model with 0.95 Confidence Level
0.8
-a 0.6
0)
t3
I
c 0.4
o
'
ro
0.2
Gamma Multi-Hit
0.05
0.1
dose
0.15
0.2
09:3306/132008
Table B-3. A summary of BMDS (version 1.4.1) modeling results based on
incidence of shedding in male Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0001
0.0000
0.0018
0.0001
0.0001
0.0000
0.0001
AIC
97.3226
101.522
89.706
97.3226
101.233
103.483
97.3226
BMD10
(mg/kg-day)
0.02650
0.06984
0.009591
0.02650
0.06524
0.06468
0.02650
BMDL10
(mg/kg-day)
0.01557
0.04768
0.005740
0.01557
0.04472
0.02975
0.01557
aChi-square />-value = p value from the Chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
This data set was not fit by any of the models in BMDS.
B-5
-------
Table B-4. A summary of BMDS (version 1.4.1) modeling results based on
incidence of alopecia in male Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0006
0.0029
0.0032
0.0031
0.0030
0.0006
0.0031
AIC
106.08
103.671
103.436
103.514
103.655
106.08
103.514
BMD10
(mg/kg-day)
201851
0.2356
0.1419
0.1701
0.2283
0.8226
0.1701
BMDL10
(mg/kg-day)
0.04072
0.07835
0.02992
0.04426
0.07445
0.1092
0.04426
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
This data set was not fit by any of the models in BMDS.
Table B-5. A summary of BMDS (version 1.4.1) modeling results based on
incidence of lacrimation in male Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 day
Model
Gamma
Logistic
Log-logistic
Multistage (2°)
Probit
Log-probit
Weibull
Chi-square
p-valuea
0.9986
1.0000
0.9996
1.0000
1.0000
0.9990
1.0000
AIC
51.0387
49.0387
51.0387
51.0387
49.0387
51.0387
49.0388
BMD10
(mg/kg-day)
0.0004624
0.0008749
0.004725
0.000374
0.000879
0.002302
0.0003007
BMDL10
(mg/kg-day)
0.0001676
0.0005251
6.03 x KT6
0.0001676
0.0005970
0.0001857
0.0001676
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
This data set was not considered amenable to modeling because of the steep slope of the
dose-response curve in the low-dose region and because the BMDio and BMDLio values were
well outside the experimental range of the data. The BMDS output for the logistic model is
presented below.
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\LACRIMATION-MALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\LACRIMATION-MALE.pit
Fri Jun 13 09:43:31 2008
B-6
-------
The form of the probability function is:
Dependent variable = #Lacrimation
Independent variable = Dose(mg/kg-d)
Slope parameter is not restricted
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0 Specified
intercept = 1.2157
slope = 14.6477
Asymptotic Correlation Matrix of Parameter Estimates
Variable
intercept
slope
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-2.36618 -0.735007
290.632 833.127
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-22.5194 4
-22.5194 2 3.27387e-008 2
-64.1035 1 83.1684 3
Est.Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.000874909
BMDL = 0.000525133
B-7
-------
Logistic Model with 0.95 Confidence Level
0.8
£ 0.6
-------
The form of the probability function is:
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Default Initial Parameter Values
Background = 1
Beta(l) = 4.41266e+020
Beta(2) = 0
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Beta(l)
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
Beta(2)
-0.44
1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.15 * * *
0 * * *
11777.7 * * *
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -30.3686 4
Fitted model -30.3686 2 2.2259e-007 2
Reduced model -68.0292 1 75.3212
64.7372
Goodness of Fit
Est._Prob. Expected Observed Size
3
Chi^2 = 0.00 d.f. = 2
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
B-9
-------
Multistage Model with 0.95 Confidence Level
Multistage
0.8
T3
3
I 0.6
o
ro 0.4
LL
0.2
OBlt/IDL, BMP
0
09:5606/132008
0.05
0.1
dose
0.15
0.2
Table B-7. A summary of BMDS (version 1.4.1) modeling results based on
incidence of miosis in male Sprague-Dawley rats exposed to thallium sulfate
via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0924
0.0201
0.3209
0.0924
0.0218
0.0214
0.0924
AIC
88.4982
93.2144
86.2645
88.4982
92.9002
91.7394
88.4982
BMD10
(mg/kg-day)
0.01341
0.03912
0.006854
0.01341
0.03674
0.02461
0.01341
BMDL10
(mg/kg-day)
0.009005
0.02853
0.003910
0.009005
0.02777
0.01547
0.009005
"Chi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
The log-logistic model was the only model that did not exhibit statistically significant
lack of fit. The BMDS output from this model is provided below.
B-10
-------
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\MIOSIS-MALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\MALE\MIOSIS-MALE.pit
Fri Jun 13 10:00:04 2008
BMDS MODEL RUN
Dependent variable = #Miosis
Independent variable = Dose(mg/kg-d)
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0.025
intercept = 2.67092
slope = 1
background intercept
background 1 -0.38
intercept -0.38 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
background
intercept
slope
Param's Deviance Test d.f. P-value
4
2 2.02732 2
1 38.3534 3
B-ll
-------
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
Log-Logistic Model with 0.95 Confidence Level
I
0.6
0.4
0.2
Log-Logistic
BMQL BMP,
o
10:0006/132008
0.05
0.1
dose
0.15
0.2
Table B-8. A summary of BMDS (version 1.4.1) modeling results based on
incidence of behavioral findings in male Sprague-Dawley rats exposed to
thallium sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0271
0.0251
0.0283
0.0271
0.0253
0.0191
0.0271
AIC
108.74
109.105
108.549
108.74
109.067
109.827
108.74
BMD10
(mg/kg-day)
0.06920
0.09757
0.05662
0.06920
0.09430
0.1271
0.06920
BMDL10
(mg/kg-day)
0.03038
0.05538
0.02007
0.03038
0.05269
0.06223
0.03038
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
This data set was not fit by any of the models in BMDS.
B-12
-------
B.2. Summary of BMDS Modeling Results for Clinical Observations in the Female Rat
(MRI, 1988)
Table B-9. A summary of BMDS (version 1.4.1) modeling results based on
incidence of rough coat in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.6681
0.0831
0.8772
0.6681
0.1033
0.2042
0.6681
AIC
72.0684
75.8559
71.5823
72.0684
75.4298
74.1826
72.0684
BMD10
(mg/kg-day)
0.02434
0.06646
0.01778
0.02434
0.06033
0.03935
0.02434
BMDL10
(mg/kg-day)
0.01597
0.04900
0.01008
0.01597
0.04503
0.02596
0.01597
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI (1988).
All models with the exception of the logistic model provided adequate fits to the data as
assessed by a chi-square goodness-of-fit test (p > 0.1) and visual inspection of the respective
plots of observed versus predicted values from the various models. BMDLio estimates from
these models were within a factor of three of each other, suggesting no appreciable model
dependence. Fitted models exhibiting adequate fit with AIC values within two units of the
lowest AIC were considered indistinguishable from one another; thus, BMDio and BMDLio
values from these models were averaged to derive the POD. Model fits that yielded the same
mathematical model were counted as a single model for averaging purposes (these models
included gamma, multistage, and Weibull). Therefore, the BMDio and BMDLio values for these
models were averaged as follows:
Average BMDio = (0.02434 + 0.01778) •*• 2 = 0.02106 mg/kg-day
Average BMDLio = (0.01597 + 0.01008) •*• 2 = 0.01303 mg/kg-day
BMDS outputs from the gamma and log-logistic models are presented below.
B-13
-------
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ROUGH_COAT-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ROUGH_COAT-
FEMALE.pit
Fri Jun 13 08:50:07 2008
The form of the probability function is:
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0365854
Slope = 8.19903
Power = 1.3
Background
Background 1
Slope -0.27
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.0268981 0.0242052 -0.0205432 0.0743394
4.32831 1.21395 1.94901 6.70761
1 NA
NA - Indicates that this parameter has hit a bound implied by some ineguality constraint and thus
has no standard error.
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -33.6561 4
Fitted model -34.0342 2 0.756278 2
Reduced model -47.1393 1 26.9666 3
Est. Prob.
11
B-14
-------
Chi"2 = 0.81 d.f. = 2
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0243422
BMDL = 0.0159664
Gamma Multi-Hit Model with 0.95 Confidence Level
Affected
.2
t3
ro
u_
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
f^n-in-i-^ Miilti Wit :
oamma iviuiii-riii
^^ I
^^-^^^ \
^^^ \
,~^ -
^ \
_ •"" j
,BMDL BMD , , , ,
0 0.05 0.1 0.15 0.2
dose
08:5006/132008
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ROUGH_COAT-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ROUGH_COAT-
FEMALE.pit
Fri Jun 13 08:52:08 2008
The form of the probability function is:
P[response] = background+(1-background)/[l+EXP(-intercept-slope^Log(dose))
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log trans formed model
Default Initial Parameter Values
background = 0.025
intercept = 2.22927
slope = 1.17797
Asymptotic Correlation Matrix of Parameter Estimates
B-15
-------
background
intercept
Variable
background
intercept
slope
intercept
-0.26
1
Parameter Estimates
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -33.6561 4
Fitted model -33.7911 2 0.270147 2 0.8737
-47.1393 1 26.9666 3 <.0001
71.5823
Goodness of Fit
Dose Est. Prob. Expected Observed
1
1
0.
11
Chi'
d.f.
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.1
Extra risk
0. 95
0.0177817
0.0100819
Log-Logistic Model with 0.95 Confidence Level
1
li
<
.1
•5
ro
""
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
.....' ' '
Log-Logistic
\
^
^^~-^~~^^ :
^^-^^^ :
/^ j
x^ :
J
BMDL BMD :
0 0.05 0.1 0.15 0.2
dose
08:5206/132008
B-16
-------
Table B-10. A summary of BMDS (version 1.4.1) modeling results based on
incidence of piloerection in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.8149
0.0484
0.6738
0.8149
0.0600
0.2492
0.8149
AIC
47.1544
54.0549
48.9941
47.1544
53.5635
49.1272
47.1544
BMD10
(mg/kg-day)
0.03857
0.1008
0.03656
0.03857
0.09134
0.05296
0.03857
BMDL10
(mg/kg-day)
0.02426
0.07462
0.01845
0.02426
0.06737
0.03658
0.02426
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
All models with the exception of the logistic and probit models provided adequate fits to
the data as assessed by a chi-square goodness-of-fit test (p > 0.1) and visual inspection of the
respective plots of observed versus predicted values from the various models. BMDLio
estimates from these models were within a factor of three of each other, suggesting no
appreciable model dependence. Fitted models exhibiting adequate fit with AIC values within
two units of the lowest AIC were considered indistinguishable from one another; thus BMDio
and BMDLio values from these models (gamma, log-logistic, multistage, log-probit, and
Weibull) were averaged to derive the POD. Model fits that yielded the same mathematical
model were counted as a single model for averaging purposes (these models included gamma,
multistage, and Weibull). Therefore, the BMDio and BMDLio values for these models were
averaged as follows:
Average BMDio = (0.03857 + 0.03656 + 0.05296) •*• 3 = 0.04270 mg/kg-day
Average BMDLio = (0.02426 + 0.0184 + 0.03658) •*• 3 = 0.02641 mg/kg-day
BMDS outputs from the gamma, log-logistic, and log-probit models are presented below.
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\PILOERECTION-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\PILOERECTION-
FEMALE. pit
Fri Jun 13 08:56:08 2008
BMDS MODEL RUN
The form of the probability function is:
P[response]= background+(1-background)^CumGamma[slope^dose,power]
B-17
-------
where CumGamma(.) is the cummulative Gamma distribution function
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0121951
Slope = 5.0954
Power = 1.3
The model parameter(s) -Background -Power
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
NA - Indicates that this parameter has hit a bound implied by some ineguality constraint and thus
has no standard error.
Model
Full model
Fitted model
Reduced model -34.6515 1
47.1544
Goodness of Fit
Dose Est._Prob. Expected Observed Size
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0385659
BMDL = 0.024262
B-18
-------
Gamma Multi-Hit Model with 0.95 Confidence Level
0.7
0.6
0.5
T3
-------
Variable
background
intercept
slope
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-21.9144 4
-22.497 2 1.16526 2
-34.6515 1 25.4742 3
P-value
d.f.
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.7
0.6
0.5
T3
-------
Probit Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\PILOERECTION-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\PILOERECTION-
FEMALE. pit
Fri Jun 13 08:59:17 2008
The form of the probability function is:
where CumNorm(.) is the cumulative normal distribution function
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0
intercept = 1.70413
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
Variable
background
intercept
slope
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-21.9144 4
-23.5636 1 3.29839 3
-34.6515 1 25.4742 3
P-value
Est. Prob.
B-21
-------
d. f.
P-value =
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.7
0.6
0.5
o.i
Extra risk
0. 95
0.0529554
0.0365797
Probit Model with 0.95 Confidence Level
T3
CD
•Q 0.4
f
g 0.3
=8
ro
it 0.2
0.1
0
Probit
BMDL
BMD
0.05
0.1
dose
0.15
0.2
08:5906/132008
Table B-ll. A summary of BMDS (version 1.4.1) modeling results based on
incidence of shedding in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.5894
0.1620
0.4442
0.5894
0.1783
0.1332
0.5894
AIC
59.2904
64.9908
61.4053
59.2904
64.6269
64.1872
59.2904
BMD10
(mg/kg-day)
0.01968
0.06624
0.01453
0.01968
0.05923
0.03710
0.01968
BMDL10
(mg/kg-day)
0.01347
0.04847
0.008519
0.01347
0.04416
0.02449
0.01347
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
B-22
-------
All models provided adequate fits of the data as assessed by a chi-square goodness-of-fit
test (p > 0.1) and visual inspection of the respective plots of observed versus predicted values
from the various models. The gamma, multistage, and Weibull models provided the same fit of
the data and were judged to provide the best model fit based on the lowest AIC value. The
BMDS output for the gamma model is provided below.
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\SHEDDING-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\SHEDDING-FEMALE.pit
Fri Jun 13 09:02:02 2008
BMDS MODEL RUN
The form of the probability function is:
Dependent variable = #Shedding
Independent variable = Dose(mg/kg-d)
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0121951
Slope = 8.9948
Power = 1.64923
The model parameter(s) -Background -Power
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
Slope
1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Slope £
Power
Model
Full model
Fitted model -28.6452 1 1.48083 3
Reduced model -47.1393 1 38.4691 3
59.2904
Goodness of Fit
Est._Prob. Expected Observed Size
B-23
-------
d. f.
P-value = 0.5894
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0196797
BMDL = 0.0134653
Gamma Multi-Hit Model with 0.95 Confidence Level
0.8
0.6
§ °-4
0.2
Gamma Multi-Hit
0.05
0.1
dose
0.15
0.2
09:0206/132008
Table B-12. A summary of BMDS (version 1.4.1) modeling results based on
incidence of alopecia in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.2392
0.0863
0.4922
0.2392
0.0924
0.0695
0.2392
AIC
111.335
113.282
109.978
111.335
113.141
113.624
111.335
BMD10
(mg/kg-day)
0.02280
0.04556
0.01338
0.02280
0.04351
0.04057
0.02280
BMDL10
(mg/kg-day)
0.01389
0.03215
0.00665
0.01389
0.03125
0.02328
0.01389
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
B-24
-------
The gamma, log-logistic, multistage, and Weibull models provided adequate fits to the
data as assessed by a chi-square goodness-of-fit test (p > 0.1) and visual inspection of the
respective plots of observed versus predicted values from the various models. BMDLio
estimates from these models were within a factor of three of each other, suggesting no
appreciable model dependence. Fitted models exhibiting adequate fit with AIC values within
two units of the lowest AIC were considered indistinguishable from one another; thus, BMDio
and BMDLio values from these models (gamma, log-logistic, multistage, and Weibull) were
averaged to derive the POD. Model fits that yielded the same mathematical model were counted
as a single model for averaging purposes (these models included gamma, multistage, and
Weibull). Therefore, the BMDio and BMDLio values for these models were averaged as follows:
Average BMDio = (0.02280 + 0.01338) •*• 2 = 0.01809 mg/kg-day
Average BMDLio = (0.01389 + 0.00665) •*• 2 = 0.01027 mg/kg-day
BMDS outputs from the gamma and log-logistic models are provided below.
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ALOPECIA-
FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ALOPECI/
FEMALE.pit
Fri Jun 13 09:06:01 2008
The form of the probability function is:
P[response]= background+(1-background)*CumGamma[slope*dose,power]
where CumGamma(.) is the cummulative Gamma distribution function
Dependent variable = #Alopecia
Independent variable = Dose(mg/kg-d)
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
B-25
-------
Variable
Background
Slope
Power
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.0604373 0.259946
1.49 7.75156
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-52.3019 4
-53.6677 2 2.73165 2
-61.0864 1 17.5691 3
P-value
0.
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0228015
BMDL = 0.0138895
I
c
o
't3
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Gamma Multi-Hit Model with 0.95 Confidence Level
Gamma Multi-Hit
BMDL , BMD
0
0.05
0.1
dose
0.15
0.2
09:0606/132008
B-26
-------
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ALOPECIA-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\ALOPECIA-FEMALE.pit
Fri Jun 13 09:07:14 2008
BMDS MODEL RUN
Dependent variable = ^Alopecia
Independent variable = Dose(mg/kg-d)
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.125
intercept = 2.11821
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
background intercept
background 1 -0.48
intercept -0.48 1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
background
intercept
slope
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -52.3019 4
-52.9891 2 1.3744
-61.0864 1 17.5691
AIC: 109.978
Goodness of Fit
Est._Prob. Expected Observed
4
9
12
B-27
-------
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
Log-Logistic Model with 0.95 Confidence Level
0.8
0.7
15 °'6
| 0.5
<
| °-4
't3
2 0.3
LJ-
0.2
0.1
0
Log-Logistic
BMDL BMD
0
0.05
0.1
dose
0.15
0.2
09:0706/132008
Table B-13. A summary of BMDS (version 1.4.1) modeling results based on
incidence of lacrimation in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (2°)
Probit
Log-probit
Weibull
Chi-square
p value"
.0000
.0000
.0000
0.9996
.0000
.0000
.0000
AIC
54.4465
54.4465
54.4465
56.4465
54.4465
54.4465
54.4465
BMD10
(mg/kg-day)
4.808 x 10~5
0.0001258
1.517 x 10~n
4.882 x 10~5
0.00024051
9.098 x KT6
4.72 x 10~5
BMDL10
(mg/kg-day)
BMD computation failed. Lower
limit includes zero.
8.1310 x 10~9
BMD computation failed. Lower
limit includes zero.
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
All models in BMDS failed to estimate useful BMD and BMDL values without excessive
extrapolation.
B-28
-------
Table B-14. A summary of BMDS (version 1.4.1) modeling results based on
incidence of exophthalmos in female Sprague-Dawley rats exposed to
thallium sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (2°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.9983
1.0000
0.9996
1.0000
1.0000
0.9969
0.9962
AIC
60.9941
58.9941
60.9941
60.9941
58.9941
60.9941
60.9942
BMD10
(mg/kg-day)
0.0005049
0.0006936
0.004856
0.001588
0.0007337
0.002044
0.0003155
BMDL10
(mg/kg-day)
0.0001719
0.0004144
6.86 x IQ-6
0.0001719
0.0004976
0.0001944
0.0001719
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI(1988).
The data set was not considered amenable to modeling because of the steep slope of the
dose-response curve in the low-dose region and because the BMDio and BMDLio values were
well outside the experimental range of the data. Of the models with the lowest AIC values, the
one with the lowest BMDLio (i.e., the logistic model) is presented below.
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\EXOPHTHALMOS-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\EXOPHTHALMOS-
FEMALE.plt
Fri Jun 13 09:15:35 2008
The form of the probability function is:
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0 Specified
intercept = 1.43545
slope = 13.3397
B-29
-------
Variable
intercept
slope
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-1.66345 -0.275389
223.326 755.172
Model
Full model
Fitted model
Reduced model
# Param's Deviance Test d.f. P-value
4
2 3.53613e-007 2 1
1 67.1787 3 <.0001
Prob.
Specified effect =
Risk Type =
Confidence level =
BMD =
BMDL =
.Q
"5
0.8
< 0.6
0.4
0.2
BiyiDLBMD
0.1
Extra risk
0.95
0. 000693631
0. 000414359
Logistic Model with 0.95 Confidence Level
0.05
0.1
dose
0.15
0.2
09:1506/132008
B-30
-------
Table B-15. A summary of BMDS (version 1.4.1) modeling results based on
incidence of miosis in female Sprague-Dawley rats exposed to thallium
sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (1°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.0010
0.0004
0.0028
0.0010
0.0004
0.0001
0.0010
AIC
109.237
111.015
107.854
109.237
110.867
112.733
109.237
BMD10
(mg/kg-day)
0.03729
0.07143
0.02252
0.03729
0.06780
0.08222
0.03729
BMDL10
(mg/kg-day)
0.02024
0.04590
0.01048
0.02024
0.04360
0.03640
0.02024
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI (1988).
This data set was not fit by any of the models in BMDS.
Table B-16. A summary of BMDS (version 1.4.1) modeling results based on
incidence of behavioral findings in female Sprague-Dawley rats exposed to
thallium sulfate via gavage for 90 days
Model
Gamma
Logistic
Log-logistic
Multistage (2°)
Probit
Log-probit
Weibull
Chi-square
p value"
0.1256
0.2197
0.1256
0.2766
0.2100
0.1256
0.1256
AIC
69.3045
67.905
69.3045
67.482
68.0129
69.3045
69.3045
BMD10
(mg/kg-day)
0.1594
0.1009
0.1736
0.1081
0.09473
0.1541
0.1760
BMDL10
(mg/kg-day)
0.04585
0.07207
0.04377
0.04445
0.06621
0.05892
0.04585
aChi-square p value = p value from the chi-square test for lack of fit. Values <0.1 fail to meet conventional
goodness-of-fit criteria.
Data source: MRI (1988).
All of the models provided adequate fits of the data as assessed by a chi-square goodness-
of-fit test (p > 0.1) and visual inspection of the respective plots of observed versus predicted
values from the various models. BMDLio estimates from these models were within a factor of
three of each other, suggesting no appreciable model dependence. Fitted models exhibiting
adequate fit with AIC values within two units of the lowest AIC were considered
indistinguishable from one another; thus BMDio and BMDLio values from all seven models were
averaged to derive the POD. Therefore, the BMDio and BMDLio values for these models were
averaged as follows:
B-31
-------
Average BMDio = (0.1594 + 0.1009 + 0.1736 + 0.1081 + 0.09473 + 0.1541 + 0.1760) •*• 7
= 0.1381 mg/kg-day
Average BMDLio = (0.04585 + 0.07207 + 0.04377 + 0.04445 + 0.06621 + 0.05892 +
0.04585) -T- 7 = 0.05387 mg/kg-day
BMDS outputs from each model are provided below.
Gamma Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:23:45 2008
BMDS MODEL RUN
The form of the probability function is:
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
Parameter Convergence has been set to: le-008
Asymptotic Correlation Matrix of Parameter Estimates
Background Slope Power
Background 1 -0.0024 -0.0022
Slope -0.0024 1 1
Power -0.0022 1 1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0.0625 0.0270635 0.00945649 0.115543
Slope 59.7049 4193.5 -8159.41 8278.82
Power 14.0574 900.278 -1750.45 1778.57
# Param's Deviance Test d.f. P-value
4
Fitted model -31.6523 3 3.45942 1
Reduced model -36.6925 1 13.5399 3
B-32
-------
Est. Prob.
d.f. =1
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.1
Extra risk
0. 95
0.159402
0.0458492
Gamma Multi-Hit Model with 0.95 Confidence Level
1
1
0
't3
LL
0.6
0.5
0.4
0.3
0.2
0.1
0
' . . . . ... ' ' ' -
oamma Multi-nit :
\
\
/ \
/
/
/ :
/ -
./ '-_
^^^^^^^^ ~~-
:
BMDL BMD :
0 0.05 0.1 0.15 0.2
dose
09:2306/132008
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:24:08 2008
The form of the probability function is:
Dependent variable = ^Behavior
Independent variable = Dose(mg/kg-d)
Slope parameter is not restricted
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
B-33
-------
Default Initial Parameter Values
background = 0 Specified
intercept = -2.96071
slope = 11.6261
intercept
slope
intercept
1
-0.74
slope
-0.74
1
Variable
intercept
slope
Parameter Estimates
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-3.77147 -1.85275
3.89255 17.6234
Param's Deviance Test d.f.
4
2 4.05988 2
1 13.5399 3
P-value
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.10089
BMDL = 0.0720676
B-34
-------
Logistic Model with 0.95 Confidence Level
T3
Fraction Affe
0.6
0.5
0.4
0.3
0.2
0.1
0
1 - . . . ' ' ' ' :
LOCJISTIC :
1
1
^^^^^ I
^^ — — — ~~ ~.
;
BMDL BMD :
0 0.05 0.1 0.15 0.2
dose
09:2406/132008
Logistic Model. (Version: 2.9; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:24:29 2008
The form of the probability function is:
P[response] = background+(1-background)/[l+EXP(-intercept-slope^Log(dose))
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.1
intercept = 0.293481
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 0.0045 0.0046
intercept 0.0045 1 1
slope 0.0046 1 1
B-35
-------
Std. Err.
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-29.9226 4
-31.6523 3 3.45942 1
-36.6925 1 13.5399 3
= 1
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.6
0.5
"S 0.4
"5
I
< 0.3
.o
2 0.2
LJ-
0.1
0
0.1
Extra risk
0. 95
0.173579
0.0437732
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
BMDL
0.05
0.1
dose
0.15
BMD
0.2
09:2406/132008
B-36
-------
Multistage Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:25:07 2008
BMDS MODEL RUN
The form of the probability function is:
The parameter betas are restricted to be positive
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Default Initial Parameter Values
Background = 0.0459871
Beta(l) = 0
Beta(2) = 9.5796
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Beta(l)
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
Background
Background 1
Beta(2) -0.4
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(l)
Beta(2)
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -29.9226 4
Fitted model -31.741 2 3.63685 2
Reduced model -36.6925 1 13.5399 3
B-37
-------
d. f.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
BMDL =
BMDU =
Taken together, (0.0444457, 0.173737) is a 90
interval for the BMD
Multistage Model with 0.95 Confidence Level
0.6
0.5
"S 0.4
"5
I
< 0.3
.o
2 0.2
LJ-
0.1
0
Multistage
BMDL
BMD
0.05
0.1
dose
0.15
0.2
09:2506/132008
Probit Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:25:44 2008
The form of the probability function is:
P[response] = CumNorm(Intercept + Slope*Dose) ,
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = ^Behavior
Independent variable = Dose(mg/kg-d)
Slope parameter is not restricted
B-38
-------
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
background = 0 Specified
intercept = -1.68929
slope = 6.25307
The model parameter(s) -background
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
intercept
intercept 1
slope
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-2.04084 -1.12985
2.08512 9.57595
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-29.9226 4
-32.0065 2 4.16782 2
-36.6925 1 13.5399 3
Dose
Scaled
Residual
Chi'
d.f.
P-value =
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0947316
BMDL = 0.0662085
B-39
-------
Probit Model with 0.95 Confidence Level
0.6
0.5
< 0.3
o
2 0.2
LJ-
0.1
0
Probit
,BMDL BMP
0.05
0.1
dose
0.15
0.2
09:2506/132008
Probit Model. (Version: 2.8; Date: 02/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:26:03 2008
The form of the probability function is:
where CumNorm(.) is the cumulative normal distribution function
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
background = 0.1
intercept = 1.18165
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 -0.001 -0.00087
intercept -0.001 1 1
slope -0.00087 1 1
B-40
-------
Variable
background
intercept
slope
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.00945646 0.115544
-703.162 711.741
-436.585 442.543
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-29.9226 4
-31.6523 3 3.45942 1
-36.6925 1 13.5399 3
Goodness of Fit
Dose Est._Prob. Expected Observed Size
d.f. = 1
P-value = 0.1256
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.154125
BMDL = 0.0589218
Probit Model with 0.95 Confidence Level
T,
!
t
<
c
0
'o
LJ-
0.6
0.5
0.4
0.3
0.2
0.1
0
F
Vi-ihit
[Quit :
^
/ '.
/ '-
/ '.
y/ -
/^ :
^^^^ :
BMDL BMD :
0 0.05 0.1 0.15 0.2
dose
09:2606/132008
B-41
-------
Weibull Model using Weibull Model (Version: 2.7; Date: 2/20/2007)
Input Data File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.(d)
Gnuplot Plotting File: M:\IRIS CHEMICALS\THALLIUM\BMD\POST EPR 6-2008\FEMALE\BEHAVIOR-FEMALE.pit
Fri Jun 13 09:26:52 2008
BMDS MODEL RUN
Dependent variable = ^Behavior
Independent variable = Dose(mg/kg-d)
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.109756
Slope = 9.64766
Power = 2.10563
Asymptotic Correlation Matrix of Parameter Estimates
Parameter Estimates
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.00944955 0.11555
-2.05203e+011 2.05207e+011
-55323 55342.4
Model
Full model
Fitted model
Reduced model
1
69.3045
Goodness of Fit
Est._Prob. Expected Observed Size
d.f. =1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
B-42
-------
Confidence level =
BMD =
BMDL =
0.6
0.5
"8 0.4
< 0.3
.o
2 0.2
LL
0.1
0
Weibull
Weibull Model with 0.95 Confidence Level
BMDL
BMD
0.05
0.1
dose
0.15
0.2
09:2606/132008
B-43
------- |