United States
Environmental Protection
1=1 m m Agency
EPA/690/R-08/008F
Final
8-25-2008
Provisional Peer Reviewed Toxicity Values for
Cobalt
(CASRN 7440-48-4)
Superfund Health Risk Technical Support Center
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268

-------
Acronyms and Abbreviations
bw
body weight
cc
cubic centimeters
CD
Caesarean Delivered
CERCLA
Comprehensive Environmental Response, Compensation and Liability Act

of 1980
CNS
central nervous system
cu.m
cubic meter
DWEL
Drinking Water Equivalent Level
FEL
frank-effect level
FIFRA
Federal Insecticide, Fungicide, and Rodenticide Act
g
grams
GI
gastrointestinal
HEC
human equivalent concentration
Hgb
hemoglobin
i.m.
intramuscular
i.p.
intraperitoneal
IRIS
Integrated Risk Information System
IUR
inhalation unit risk
i.v.
intravenous
kg
kilogram
L
liter
LEL
lowest-effect level
LOAEL
lowest-observed-adverse-effect level
LOAEL(ADJ)
LOAEL adjusted to continuous exposure duration
LOAEL(HEC)
LOAEL adjusted for dosimetric differences across species to a human
m
meter
MCL
maximum contaminant level
MCLG
maximum contaminant level goal
MF
modifying factor
mg
milligram
mg/kg
milligrams per kilogram
mg/L
milligrams per liter
MRL
minimal risk level
MTD
maximum tolerated dose
MTL
median threshold limit
NAAQS
National Ambient Air Quality Standards
NOAEL
no-ob served-adverse-effect level
NOAEL(ADJ)
NOAEL adjusted to continuous exposure duration
NOAEL(HEC)
NOAEL adjusted for dosimetric differences across species to a human
NOEL
no-ob served-effect level
OSF
oral slope factor
p-IUR
provisional inhalation unit risk
p-OSF
provisional oral slope factor
p-RfC
provisional inhalation reference concentration
1

-------
p-RfD
provisional oral reference dose
PBPK
physiologically based pharmacokinetic
ppb
parts per billion
ppm
parts per million
PPRTV
Provisional Peer Reviewed Toxicity Value
RBC
red blood cell(s)
RCRA
Resource Conservation and Recovery Act
RDDR
Regional deposited dose ratio (for the indicated lung region)
REL
relative exposure level
RfC
inhalation reference concentration
RfD
oral reference dose
RGDR
Regional gas dose ratio (for the indicated lung region)
s.c.
subcutaneous
SCE
sister chromatid exchange
SDWA
Safe Drinking Water Act
sq.cm.
square centimeters
TSCA
Toxic Substances Control Act
UF
uncertainty factor
l^g
microgram
[j,mol
micromoles
voc
volatile organic compound
11

-------
8-25-2008
PROVISIONAL PEER REVIEWED TOXICITY VALUES FOR
COBALT (CASRN 7440-48-4)
Background
On December 5, 2003, the U.S. Environmental Protection Agency's (EPA's) Office of
Superfund Remediation and Technology Innovation (OSRTI) revised its hierarchy of human
health toxicity values for Superfund risk assessments, establishing the following three tiers as the
new hierarchy:
1.	EPA's Integrated Risk Information System (IRIS).
2.	Provisional Peer-Reviewed Toxicity Values (PPRTV) used in EPA's Superfund
Program.
3.	Other (peer-reviewed) toxicity values, including:
~	Minimal Risk Levels produced by the Agency for Toxic Substances and Disease
Registry (ATSDR),
~	California Environmental Protection Agency (CalEPA) values, and
~	EPA Health Effects Assessment Summary Table (HEAST) values.
A PPRTV is defined as a toxicity value derived for use in the Superfund Program when
such a value is not available in EPA's Integrated Risk Information System (IRIS). PPRTVs are
developed according to a Standard Operating Procedure (SOP) and are derived after a review of
the relevant scientific literature using the same methods, sources of data, and Agency guidance
for value derivation generally used by the EPA IRIS Program. All provisional toxicity values
receive internal review by two EPA scientists and external peer review by three independently
selected scientific experts. PPRTVs differ from IRIS values in that PPRTVs do not receive the
multi-program consensus review provided for IRIS values. This is because IRIS values are
generally intended to be used in all EPA programs, while PPRTVs are developed specifically for
the Superfund Program.
Because new information becomes available and scientific methods improve over time,
PPRTVs are reviewed on a five-year basis and updated into the active database. Once an IRIS
value for a specific chemical becomes available for Agency review, the analogous PPRTV for
that same chemical is retired. It should also be noted that some PPRTV manuscripts conclude
that a PPRTV cannot be derived based on inadequate data.
Disclaimers
Users of this document should first check to see if any IRIS values exist for the chemical
of concern before proceeding to use a PPRTV. If no IRIS value is available, staff in the regional
Superfund and RCRA program offices are advised to carefully review the information provided
in this document to ensure that the PPRTVs used are appropriate for the types of exposures and
circumstances at the Superfund site or RCRA facility in question. PPRTVs are periodically
1

-------
8-25-2008
updated; therefore, users should ensure that the values contained in the PPRTV are current at the
time of use.
It is important to remember that a provisional value alone tells very little about the
adverse effects of a chemical or the quality of evidence on which the value is based. Therefore,
users are strongly encouraged to read the entire PPRTV manuscript and understand the strengths
and limitations of the derived provisional values. PPRTVs are developed by the EPA Office of
Research and Development's National Center for Environmental Assessment, Superfund Health
Risk Technical Support Center for OSRTI. Other EPA programs or external parties who may
choose of their own initiative to use these PPRTVs are advised that Superfund resources will not
generally be used to respond to challenges of PPRTVs used in a context outside of the Superfund
Program.
Questions Regarding PPRTVs
Questions regarding the contents of the PPRTVs and their appropriate use (e.g., on
chemicals not covered, or whether chemicals have pending IRIS toxicity values) may be directed
to the EPA Office of Research and Development's National Center for Environmental
Assessment, Superfund Health Risk Technical Support Center (513-569-7300), or OSRTI.
INTRODUCTION
The Integrated Risk Information System (IRIS) does not report a Reference Dose (RfD)
for cobalt (U.S. EPA, 2007). The Health Effects Assessment Summary Tables (HEAST)
(U.S. EPA, 1997a) and Drinking Water Standards and Health Advisories list (U.S. EPA, 2004)
likewise do not contain an RfD for cobalt. The Chemical Assessments and Related Activities
(CARA) lists (U.S. EPA, 1991, 1994a) report a Health Effect Assessment (HEA) for cobalt
(U.S. EPA, 1987). The 1987 HEA derived a chronic RfD of 0.005 mg cobalt/kg-day based on a
no-observed-adverse-effect level (NOAEL) of 5 mg cobalt/kg-day for testicular effects in a
subchronic rat study (Nation et al., 1983). The Agency for Toxic Substances and Disease
Registry (ATSDR) Toxicological Profile for cobalt and its compounds reports an oral Minimal
Risk Level (MRL) for intermediate exposure of lxlO"2 mg/kg-day (ATSDR, 2004), based on a
lowest-observed-adverse-effect level (LOAEL) of approximately 1 mg cobalt/kg-day for
polycythemia in humans (Davis and Fields, 1958). ATSDR (2004) did not derive an oral MRL
for chronic exposure. This MRL for intermediate exposure was based on the polycythemic
effect of cobalt exposure (1 mg cobalt/kg-day, Davis and Fields, 1958) by application of an UF
of 10 for a LOAEL and an UF of 10 for human variability. The World Health Organization
(WHO, 2005) has not published an Environmental Health Criteria (EHC) document on cobalt.
An International Agency for Research on Cancer (IARC) Monograph on cobalt and its
compounds (IARC, 2006) and the National Toxicology Program (NTP) Status Reports (NTP,
2005) were searched for relevant information.
IRIS (U.S. EPA, 2007) does not report a Reference Concentration (RfC) for cobalt. The
HEAST (U.S. EPA, 1997a) likewise does not list an RfC for cobalt. The cobalt HEA (U.S. EPA,
1987) derived a subchronic inhalation RfC of 9xl0"5 mg/m3 based on a LOAEL of 0.1 mg/m3 for
2

-------
8-25-2008
respiratory effects in a 3-month study in swine (Kerfoot et al., 1975). A chronic inhalation RfC
of 9x10~6 mg/m3 was derived from the same study. The ATSDR Toxicological Profile for cobalt
and its compounds reports an inhalation MRL for chronic exposure of lxlO"4 mg/m3 (ATSDR,
"3
2004),	based on a NOAEL of 0.0053 mg cobalt/m for decreased pulmonary function in humans
(Nemery et al., 1992). The American Conference of Governmental Industrial Hygienists
(ACGIH, 2004) has set a Threshold Limit Value-Time-Weighted Average (TLV-TWA) of
0.02 mg/m3 for cobalt and inorganic cobalt compounds, expressed as cobalt, based on respiratory
and cardiovascular effects. The National Institute for Occupational Safety and Health (NIOSH,
2005)	Recommended Exposure Limit (REL) TWA for cobalt is 0.05 mg/m3, based on effects in
the respiratory system. The Occupational Safety and Health Administration (OSHA, 2005)
Permissible Exposure Limit (PEL) is 0.1 mg/m3.
IRIS (U.S. EPA, 2007) does not report a cancer classification, slope factor or unit risk for
cobalt. The HEAST (U.S. EPA, 1997a) and Drinking Water Standards and Health Advisories
list (U.S. EPA, 2004) likewise do not report carcinogenicity assessments for cobalt. The CARA
lists (U.S. EPA, 1991, 1994a) do not report a cancer classification or an estimate of the
carcinogenic potency of stable cobalt compounds due to a lack of pertinent data. An IARC
Monograph on cobalt and its compounds (IARC, 2006) classified cobalt sulfate and other soluble
cobalt (II) salts as "possibly carcinogenic to humans." ACGIH (2004) has classified cobalt in
category A3 - confirmed animal carcinogen with unknown relevance to humans.
Literature searches for studies relevant to the derivation of provisional toxicity values for
cobalt were conducted initially through 2000 in TOXLINE (supplemented with BIOSIS and
NTIS updates), MEDLINE, TSCATS, RTECS, CCRIS, DART, EMIC/EMICBACK, HSDB,
GENETOX and CANCERLIT and subsequently from 2000 to August 2005 in MEDLINE,
TOXLINE (NTIS subfile), TOXCENTER, TSCATS, CCRIS, DART/ETIC, GENETOX, HSDB,
RTECS and Current Contents. An updated literature search was performed in MEDLINE from
2005 to June 2008.
REVIEW OF PERTINENT DATA
Human Studies
Overview
Indicators of adverse health effects in humans following oral exposure to cobalt include
increased erythrocyte number and hemogloblin (Taylor et al., 1977; Duckham and Lee, 1976;
Davis and Fields, 1958), cardiomyopathy (Morin et al., 1971; Alexander, 1969, 1972) and
decreased iodine uptake by the thyroid (Roche and Layrisse, 1956). Cardiomyopathy is an
endpoint of concern for cobalt in humans; however, it is highly likely that alcohol consumed in
"beer-cobalt cardiomyopathy," as well as other factors, such as smoking, played a role in the
effects that were observed. Cobalt is a sensitizer in humans by any route of exposure. Sensitized
individuals may react to inhalation of cobalt by developing asthma; ingestion or dermal contact
with cobalt may result in development of dermatitis. Several studies have suggested that
3

-------
8-25-2008
cross-sensitization may occur between cobalt and nickel (Shirakawa et al., 1990; Lammintausta
et al., 1985; Bencko et al., 1983; Rystedt and Fisher, 1983).
Respiratory effects, including respiratory irritation, wheezing, asthma, pneumonia and
fibrosis, have been widely reported in humans exposed to cobalt by inhalation (for review, see
Barceloux, 1999; Lison, 1996). Epidemiology studies show decreased pulmonary function in
workers exposed to inhaled cobalt (Nemery et al., 1992; Gennart and Lauwerys, 1990). Results
of studies investigating cancer incidence in workers exposed to inhaled cobalt are suggestive of a
possible association between exposure to cobalt and respiratory tumors (Tuchsen et al., 1996;
Mur et al., 1987; Morgan, 1983).
Oral Exposure
In humans, cobalt stimulates production of red blood cells through increased production
of the hormone erythropoietin and has been explored for use in the treatment of anemia (Smith
and Fisher, 1973; Duckham and Lee, 1976). Increases in red blood cell counts and blood
hemoglobin have been reported in non-anemic volunteers (Davis and Fields, 1958) and in
anephric anemic patients (Taylor et al., 1977; Duckham and Lee, 1976).
Reversible polycythemia (increase in blood cell number) was reported (see Table 1) in
six healthy adult males following treatment with 150 mg cobalt chloride per day for 22 days
(Davis and Fields, 1958). Five subjects received 150 mg cobalt chloride/day for the entire
exposure period and a sixth subject initially received 120 mg cobalt chloride/day, which was
later increased (time not specified) to 150 mg/day. Cobalt chloride was administered as a 2%
solution diluted in either water or milk. Assuming an average body weight of 70 kg, 150 mg
cobalt chloride/day corresponds to approximately 1 mg cobalt/kg-day. Outcomes assessed in
this study were red blood cell count, hemoglobin percentage, leukocyte count, reticulocyte
percentage and thrombocyte count. Polycythemia was observed in all six patients within 7 to
22 days of treatment as demonstrated by increases in red blood cell counts ranging from 0.5 to
1.19 million (approximately 16-20% increase above pre-treatment levels) and increases in
hemoglobin levels ranging from 6 to 11% above pretreatment values. In five of the six subjects,
reticulocyte levels were elevated, reaching at least twice the pre-experiment values.
Thrombocyte and total leukocyte counts were not significantly different from pretreatment
values. Erythrocyte counts returned to pre-treatment levels within 9 to 15 days after cobalt
administration was discontinued. The fact that leucocyte counts remained relatively constant
throughout the experiment supports the concept that this is a true polycythemia. As such, based
on the results of this study, 1 mg cobalt/kg-day was identified as a LOAEL for cobalt-induced
polycythemia in humans.
4

-------
8-25-2008
Table 1. Hematopoietic, Thyroid and Developmental Effects of Cobalt (Co) via Oral
Route
Target Organ
Species
Effect
Dosage
(mg Co/kg-day)
Hematopoietic Effects
Human
Human
Rat
Reversible Effect (Polycythemia)
t Hemoglobin and | RBC
Hematopoietic effect
1.0
0.16-0.32*
0.5-32.0
Thyroid
Human
Mice
i Iodine uptake
Histopathological changes in thyroid
1.0
48.0
Fetus
Rat
Developmental toxicity
5.2-21.0
Heart
Rat
I Myocardial function
8.0
*Therapeutic doses for anemic patients
Duckham and Lee (1976) treated 12 anephric patients on dialysis with 25 to 50 mg cobalt
chloride daily for approximately 12 weeks. Assuming an average body weight of 70 kg, doses of
25 and 50 mg cobalt chloride/day are equivalent to 0.16 and 0.32 mg cobalt/kg-day, respectively.
During the exposure period, patients also received daily treatment with 100 mg ferrous sulfate
and 50 mg ascorbic acid. Within approximately 2 months of initiation of treatment with cobalt,
an increase in hemoglobin of 26-70% was observed in patients treated with 0.32 mg
cobalt/kg-day. Serum cobalt levels appeared to reach steady state within 2 months of exposure
(approximately 40-100 |ig cobalt/100 mL). In a subgroup of three patients, continuation of
treatment with 0.16 mg cobalt/kg-day for approximately 3 months maintained elevated
hemoglobin levels. Hemoglobin levels decreased rapidly when cobalt therapy was discontinued.
The authors did not report whether therapy with ferrous sulfate and ascorbic acid was
discontinued at the same time. Results of this study are difficult to interpret because patients
were anephric and on dialysis, which may have altered cobalt pharmacokinetics and dose-effect
relationships. Furthermore, since it is well established that treatment with ferrous sulfate alone
increases hemoglobin concentration (Hillman, 2001), concomitant therapy with iron is a
confounding factor. Since this study did not evaluate the response of patients treated with
ferrous sulfate alone, it is not possible to determine the relative contributions of iron and cobalt
to the observed increases in hemoglobin. Thus, adverse effect levels cannot be confidently
determined for cobalt. In a separate study, a group of eight anephric patients with refractory
anemia were treated with 25 to 50 mg cobalt chloride daily for 12 to 36 weeks (Taylor et al.,
1977). Increased hemoglobin concentration and decreased requirement for blood transfusions
were observed (Taylor et al., 1977). Data on hemoglobin concentrations (or other indicators of
polycythemia) were not reported.
Pregnant women given 75 to 100 mg cobalt chloride/day with no other treatment for
90 days to 6 months did not experience pregnancy-induced reductions in hematocrit and
hemoglobin levels, compared to untreated controls (Holly, 1955). However, daily treatment with
1 g ferrous sulfate alone or combined daily treatment with 60 to 90 mg cobalt chloride and 0.8 to
1.2 g ferrous sulfate prevented pregnancy-related decreases in hematocrit and hemoglobin levels.
The response to combined cobalt chloride and iron therapy was more pronounced than the
5

-------
8-25-2008
response to iron therapy alone. In patients treated with iron only, decreases in hemoglobin and
hematocrit were prevented in approximately 80% of patients, compared to 100% of patients
treated with combined cobalt chloride and iron.
Cardiomyopathy has been observed in association with consumption of large quantities
of beer containing cobalt chloride (introduced into the beer to stabilize the foam) (Alexander,
1969, 1972; Morin et al., 1971). Exposure estimates in reported cases range from 0.04 to
0.14 mg cobalt/kg-day (corresponding to approximately 8-30 pints of beer daily) over a period of
years (Alexander, 1969, 1972; Morin et al., 1971). The cardiomyopathy in the beer drinkers,
referred to in the literature as "beer-cobalt cardiomyopathy," was fatal to 43% of the subjects
within several years, with approximately 18% of these deaths occurring within the first several
days following diagnosis. Beer-cobalt cardiomyopathy appeared to be similar to alcoholic
cardiomyopathy and beriberi; however, the onset of the beer-cobalt cardiomyopathy was much
more abrupt. The practice of adding cobalt to beer to stabilize the foam has been discontinued.
It should be noted, however, that the cardiomyopathy may also have been due to the fact that the
beer drinkers had protein-poor diets and may have had prior or concurrent cardiac and hepatic
damage from alcohol abuse. Due to the potential adverse effects of poor nutrition and/or chronic
ethanol exposure on cardiovascular health, it is difficult to delineate the contribution of oral
cobalt exposure to the observed cardiomyopathy. As such, no adverse effects levels can be
determined for cobalt-induced cardiotoxicity.
The thyroid also appears to be a target organ for cobalt (see Table 1). Treatment of
12 euthyroid (normal thyroid) patients with 150 mg cobalt chloride/day (equivalent to 1 mg
cobalt/kg-day, assuming a body weight of 70 kg) for 2 weeks resulted in a greatly reduced
uptake of 48-hour radioactive iodine by the thyroid when measured after 1 week of exposure to
cobalt, with uptake nearly abolished completely by the second week of exposure to cobalt
(Roche and Layrisse, 1956). It should be noted that when cobalt treatment was discontinued,
iodine uptake returned to pre-treatment reported values. No other clinical details were provided
for the human subjects. Therefore, based on the results of this study, a LOAEL of 1 mg
cobalt/kg-day was identified for decreased radioactive iodine uptake in human thyroid following
oral cobalt exposure. In another small clinical study (Paley et al., 1958), decreased radioactive
iodine uptake was reported in two of four (3 males, 1 female) euthyroid patients orally
administered 37.5 mg cobalt/day as cobalt chloride (equivalent to 0.54 mg cobalt/kg-day,
assuming a body weight of 70 kg) for 10 to 14 days. One of the two subjects with reported
decreased iodine uptake had received i.v. cobalt in addition to oral cobalt intake, and had been
previously diagnosed with hyperthyroidism (although was clinically euthyroid at the time of
study). The i.v. dosing may have raised the internal cobalt concentration to a level greater than
the reported 0.54 mg dosage based upon oral dosing of 37.5 mg/day in other subjects that did not
receive i.v. cobalt. Of the remaining three subjects, 24-hour iodine uptake was not significantly
decreased following oral cobalt exposure compared to corresponding pre-treatment values (based
on pairwise t-test). The oral cobalt dose of 0.54 mg cobalt/kg-day represents a NOAEL for
thyroid effects in humans. It should be noted that the Roche and Layrisse (1956) and Paley et al.
(1958) studies lack details pertinent to other clinical conditions (e.g. including effects on thyroid
stimulating hormone [TSH]) of these patients; thus the mechanism for the effect of cobalt on
thyroidal iodine uptake cannot be ascertained. However, cobalt appears to increase thiocyanate-
6

-------
8-25-2008
induced release of radioiodine from the thyroid, suggesting a possible effect on binding of iodine
(e.g., iodination of thyroglobulin) in the thyroid gland.
Cobalt has been found to be a sensitizer in humans. Individuals are sensitized following
dermal or inhalation exposure, but flares of dermatitis may be triggered following cobalt
ingestion. In a small clinical study, several patients with eczema of the hands were challenged
orally with 1 mg cobalt sulfate (0.005 mg cobalt/kg-day, assuming a body weight of 70 kg) in
tablet form once per week for 3 weeks; this translates to an estimated average daily dose of
0.0007 mg cobalt/kg-day (1 day a week/7 days a week x 0.005 mg cobalt/kg-day).
28/47 patients had a flare of dermatitis following the oral challenge (Veien et al., 1987). All
47 patients had positive dermal patch tests to cobalt (13 to cobalt alone and 34 to nickel and
cobalt) and 7 of the 13 patients who had patch-tested positive to cobalt alone reacted to the oral
challenge. These results suggest that cobalt allergy can be induced from oral ingestion exposures
to cobalt. Although the exposure levels associated with sensitization to cobalt following
inhalation or dermal exposure have not been established, interrelationships have been found to
exist between cobalt and nickel sensitization (Bencko et al., 1983; Rystedt and Fisher, 1983;
Veien et al., 1987). In guinea pigs, nickel and cobalt sensitization appear to be interrelated and
mutually enhancing (Lammintausta et al., 1985). Therefore, it is possible that in people
sensitized by nickel, exposure to cobalt may result in an allergic reaction.
Inhalation Exposure
Numerous studies have investigated health effects in workers occupationally exposed to
cobalt-bearing dust (Linna et al., 2003; Swennen et al., 1993; Auchincloss et al., 1992; Cugell,
1992; Nemery et al., 1992; Prescott et al., 1992; Gennart and Lauwerys, 1990; Meyer-Bisch et
al., 1989; Raffn et al., 1988; Shirakawa et al., 1988, 1989; Sprince et al., 1988; Kusaka et al.,
1986a,b; Demedts et al., 1984; Davison et al., 1983). However, many of these studies are of
limited utility for risk assessment due to inadequate characterization of exposure and/or effects.
Four studies were considered to be potentially suitable for RfC derivation. Two of these focused
exclusively on respiratory effects (Nemery et al., 1992; Gennart and Lauwerys, 1990); one
studied only thyroid effects (Prescott et al., 1992) and one considered multiple endpoints
(Swennen et al., 1993). The populations studied included diamond-cobalt saw manufacturers,
diamond polishers, plate painters and cobalt production workers. All four studies were cross-
sectional design.
Several studies have examined the effects of hard metal, a mixture containing
approximately 20% cobalt with the remainder being primarily tungsten carbide. Exposure of
humans to hard metal has been shown to result in an increase in cancer mortality (Moulin et al.,
1998; Lasfargues et al., 1994) as well as a number of other diseases, including asthma and
pulmonary fibrosis (for reviews, see Barceloux, 1999; Lison, 1996). There is substantial
evidence from animal studies that tungsten, although it acts as an inert dust by itself, can
potentiate the effects of cobalt on the respiratory tract (Lasfargues et al., 1995; Lison et al., 1995,
1996; Swennen et al., 1993). For this reason, studies of hard metal were not given further
consideration.
7

-------
8-25-2008
Gennart and Lauwerys (1990) studied ventilatory function in workers at a plant
producing diamond-cobalt circular saws. The form of cobalt used in diamond polishing is
primarily metallic cobalt powder; specific cobalt species contained in this powder were not
identified. The exposed population consisted of 48 workers (34 males and 14 females) who
agreed to participate in the study (an additional 27 workers declined). Exposure duration for
these workers ranged from 0.1 to 32 years, with an average of approximately 6 years. The work
involved weighing and mixing cobalt powder and microdiamond particles (and possibly small
amounts of other undisclosed substances), cold pressing, heating and hot pressing. After
sintering, the pieces were welded onto steel disks. These operations were performed in two
rooms called the mixing room and the oven room, where all the examined workers spent most of
their time. Controls consisted of 23 workers (11 males and 12 females) from other factories in
the same area who were not exposed to known pneumotoxic chemicals. Personal air samples
were collected at different workplaces during half a workshift. Subjects filled out a
questionnaire regarding occupational and medical histories, smoking habits and pulmonary
symptoms; gave a urine sample for cobalt determination; and participated in lung function tests.
"3
Cobalt concentrations varied from 9.4 to 2875 jig/m in the mixing room (geometric
mean=135.5 (J,g/m3) and from 6.2 to 51.2 (J,g/m in the oven room (geometric mean=15.2 (j,g/m3).
The prevalence of respiratory symptoms, such as cough, sputum and dyspnea, were significantly
increased in the exposed workers compared to the control group (numeric data not reported).
Mean predicted values of FEVi (forced expiratory volume in 1 second adjusted for body size)
and FVC (forced vital capacity) were significantly lower, and the prevalence of abnormal values
was higher in the cobalt exposed workers (both smokers and non-smokers) compared to the
control group. In controls, FEVi and FVC were 95.4 and 101.6 percent of predicted values,
respectively. Mean percent predicted FEVi and FVC in exposed non-smokers were 87.1 and
92.3, respectively, and in exposed smokers were 83.9 and 93.4, respectively. Among
non-smokers, all measures of pulmonary function were lower in workers exposed for 5 years or
more than in those exposed to cobalt for a shorter period of time.
Nemery et al. (1992) conducted a cross-sectional study of cobalt exposure and respiratory
effects in diamond polishers who were primarily exposed to metallic cobalt-containing dust;
species of cobalt in the dust samples were not identified. The study group was composed of
194 polishers working in 10 different workshops. In two of these workshops (#1, 2), the workers
used cast iron polishing disks almost exclusively, and in the others, they primarily used
cobalt-containing disks. The number of subjects from each workshop varied from 6 to 28 and
the participation rate varied from 56 to 100%. The low participation in some workshops reflects
the fact that only workers who used cobalt disks were initially asked to be in the study; low
participation is not due to a high refusal rate (only eight refusals were documented). More than a
year after the polishing workshops were studied, an additional three workshops with workers
engaged in sawing diamonds, cleaving diamonds or drawing jewelry were studied as an
unexposed control group (n=59 workers). Subjects were asked to fill out a questionnaire
regarding employment history, working conditions, medical history, respiratory symptoms and
smoking habits; to give a urine sample for cobalt determination; and to undergo a clinical
examination and lung function tests. Both area air samples and personal air samples were
collected (always on a Thursday). Sampling for area air determinations started 2 hours after
work began and continued until 1 hour before the end of the work day. Personal air samples
were collected from the breathing zone of a few workers per workshop for four successive
8

-------
8-25-2008
1-hour periods. Air samples were analyzed for cobalt and iron. In addition, personal air
samplers were used to sample the air 1 cm above the polishing disks. These samples were
analyzed for the entire spectrum of mineral and metallic compounds. Air samples were not
obtained at one of the polishing workshops (#4); however, this workshop was reported to be
almost identical to an adjoining workshop (#3) for which samples were obtained. Urinary cobalt
levels were similar between workers in these two workshops, so exposure was considered to be
similar as well.
Results of area and personal air sampling were strongly correlated (R=0.92), with area air
sampling reporting lower concentrations than personal air samples in all workshops except one
(#9) (Nemery et al., 1992). In this workshop, personal air samples appeared to be artificially low
in comparison to area air samples and urinary cobalt levels of the workers. When this workshop
was excluded, a strong correlation (R=0.85-0.88) between urinary cobalt and cobalt in the air
was observed. Based on urinary cobalt levels, the predicted concentration of cobalt expected in
"3
personal air samples from workshop #9 was approximately 45 (J,g/m (the mean value actually
reported was 6 (J,g/m3). The polishing workshops were divided into two groups: those with low
exposure to cobalt (#1-5, n=102) and those with high exposure to cobalt (#6-10, n=91). Mean
cobalt exposure concentrations were 0.4, 1.6 and 10.2 (J,g/m3 by area air sampling and 0.4, 5.3
"3
and 15.1 (J,g/m by personal air sampling in the control, low-exposure and high-exposure groups,
respectively. The inclusion of the apparently biased personal air samples from workshop #9
means that the reported mean cobalt exposure in the high-exposure group obtained by personal
air sampling (15.1 (J,g/m3) may be lower than the true value. Air concentrations of iron were
highest in the two polishing workshops that used iron disks and the sawing workshop (highest
value=62 (j,g/m3), and were not correlated with cobalt levels. Analysis of samples taken near the
disks showed the presence of cobalt, with occasional traces of copper, zinc, titanium, manganese,
chromium, silicates and silicon dioxide. No tungsten was detected. Some workers may have
previously been exposed to asbestos since pastes containing asbestos had been used in the past to
glue the diamonds onto holders. However, since the asbestos was in its non-friable form,
exposure was insufficient to produce functional impairment. Smoking habits were similar in
workers from the high-exposure, low-exposure and control groups. Duration of exposure was
not discussed.
Workers in the high-exposure group were more likely than those in the other groups to
complain about respiratory symptoms; the prevalences of eye, nose and throat irritation and
cough, and the fraction of these symptoms related to work, were significantly increased in the
high-exposure group (Nemery et al., 1992). Workers in the high-exposure group also had
significantly lower lung function compared to controls and low-exposure group workers, as
assessed by FVC, FEVi, MMEF (forced expiratory flow between 25 and 75% of the FVC) and
mean PEF (peak expiratory flow rate), although the prevalence of abnormal values did not differ
significantly between exposure categories. In controls, FVC, FEVi and MMEF were
approximately 110, 107 and 94 percent of predicted values, respectively, compared to
approximately 105, 104 and 87 percent of predicted values, respectively, in the high-exposure
group workers. Results in the low-exposure group did not differ from controls. The effect on
spirometric parameters in the high exposure group was present in both men and women. Women
seemed to be affected more than men; however, the interaction between exposure and sex was
not significant (two-way analysis of variance). Smoking was found to exert a strong effect on
9

-------
8-25-2008
lung function; however, lung function level remained negatively correlated with exposure to
cobalt, independent of smoking.
A cobalt dose-effect relationship is evident from the Nemery et al. (1992) study, based on
a multivariate regression analysis of urinary cobalt and lung function measurements. Increasing
urinary cobalt concentration (approximate range <1-70 |ig cobalt/g creatinine) was significantly
(/;<0,05) associated with co-variate-adjusted decreasing forced expiratory volume (FEV1%) and
forced vital capacity (FVC%). Significant co-variates retained in the regression analysis
included gender and smoking. The model predicted 3% and 4% decreases in FEV1% and
FVC%, respectively, in association with a 10-fold increase in urinary cobalt concentration. The
approximate mean urinary cobalt levels of the control and high exposure groups were 2 and
20 |j,g cobalt/g creatinine, respectively. The magnitude of the cobalt effect was similar to the
predicted effect of smoking, approximately 3-4% decrease in FEV1% and FVC%. Cobalt
concentration determined from personal air sampling may be more representative of airborne
cobalt exposure than area sampling. As such, 5.3 (j,g/m3 and 15.1 (j,g/m3 represent a NOAEL and
LOAEL, respectively, for decreased pulmonary function and increased symptoms of airway
irritation.
Swennen et al. (1993) conducted a cross-sectional study of workers exposed to metallic
cobalt and various inorganic cobalt salts and oxides (specific species not identified) at a cobalt
plant producing these materials from cobalt metal cathodes and scrap metal. The study group
included 82 male workers from the cobalt plant who had no history of lung disease prior to
employment and who had never been exposed to other pneumotoxic chemicals. Methods for
selection or exclusion of subjects in constructing the cohort and participation were not reported.
The control group comprised 82 age-matched workers from the mechanical workshop of a
nearby plant owned by the same company. Workers filled out a questionnaire regarding
occupational history, respiratory complaints and smoking habits; received a routine clinical
examination; participated in lung function tests; had a chest radiograph taken; and gave blood
and urine samples (before and after working on Monday and Friday of one week) for
determination of cobalt content as well as hematological and serum chemistry analyses.
Exposure was monitored by personal air samplers worn by each cobalt worker for 6 hours on
both Monday and Friday.
Workers in the cobalt plant were exposed to cobalt concentrations ranging from 1 to
7772 (J,g/m3 (Swennen et al., 1993). The geometric mean exposure concentration was 125 (J,g/m3.
Exposure duration ranged from 0.3 to 39.4 years, with an average exposure of 8.0 years. A
significantly higher number of exposed workers reported dyspnea than did controls. The
increase occurred primarily among smokers although no significant interaction was found
between smoking and exposure to cobalt. Based on a logistic regression model, the probability
of dyspnea during exercise was significantly associated with increasing cobalt concentration in
the air or urine. The parameters of the model were not reported. The clinical examinations
detected significantly increased prevalence of skin disorders (eczema, erythema) (51 vs. 25%)
and wheezing (16 vs. 6%) in the exposed group compared to controls. Lung function tests did
not differ between the two groups; however, a few significant trends were noted: the FEVi/VC
(forced expiratory volume in one second/vital capacity) ratio decreased with increasing
concentration of cobalt in the air and urine, and the RV (residual volume) and TLC (total lung
10

-------
8-25-2008
capacity) increased with increasing duration of exposure. No lung abnormalities were found by
chest radiographs in either group. Blood analyses did not show polycythemia, and in fact, there
were slight, but significant, decreases in red blood cell count, hemoglobin and hematocrit in the
exposed workers. White blood cell counts were significantly increased. Serum levels of the
thyroid hormone T3 (triiodothyronine) were slightly (7%), but significantly, decreased in the
exposed group, while T4 (thyroxine) and TSH (thyrotropin) were not affected. Serum markers
for cardiomyopathy (i.e., myocardial creatine kinase) were unchanged.
Prescott et al. (1992) conducted a cross-sectional study to investigate the effects of cobalt
exposure on thyroid volume in female plate painters. The test group included 61 female plate
painters exposed to cobalt blue dyes in two porcelain factories. The control group consisted of
48 unexposed women working at the same factories. The dyes used in the two factories differed;
factory I (36 workers) used cobalt aluminate, which is insoluble, and factory II (25 workers) used
cobalt-zinc silicate, which was reported to be "semi-soluble." Workers were exposed to cobalt
during the painting procedure when the plates were spray-painted (under a fume hood) two or
three times with the water-based cobalt blue underglaze and when the excess color was removed
"3
with a brush after drying. Cobalt concentrations were reported to be approximately 0.05 mg/m
in the workplaces (no further details on air levels were reported). The average duration of
exposure was 14.6 years in group I workers and 16.2 years in group II workers. Subjects filled
out a questionnaire regarding health, use of medicines, day of menstrual cycle, employment
information and smoking habits and agreed to give blood and urine samples for determination of
thyroid hormone levels (e.g. thyroxine (T4), triiodothyronine (T3), and thyroid stimulating
hormone) and cobalt concentration, respectively, and to undergo ultrasonography to determine
volume of the thyroid gland.
Urinary cobalt levels were similar in group I exposed workers and controls (Prescott et
al., 1992). Group II workers exposed to semi-soluble cobalt-zinc silicate had urinary cobalt
levels that were approximately 10-fold higher than controls. Group I workers did not differ from
controls for any of the thyroid parameters measured; however, Group II workers had a
significant 22% increase in serum T4 (thyroxine) levels. Mean thyroid volume was lower in this
group as well, although the difference from controls (16.1 mL in group II vs. 19.2 mL in controls
and 18.7 mL in group I) was not statistically significant. The occurrence of respiratory effects in
these workers was not reported.
Results of three studies investigating cancer incidence in workers exposed to cobalt by
the inhalation route (Tuchsen et al., 1996; Mur et al., 1987; Morgan, 1983) are suggestive of a
possible association between exposure to cobalt and respiratory tumors. Morgan (1983)
investigated the health and causes of death of 49 men occupationally exposed to cobalt salts and
oxides (specific species not identified) in a manufacturing plant in South Wales. During the
study period, 33 men died (five with lung cancer and three with cancer at other sites). The
expected number of deaths was 3.0 for lung cancer and 4.1 for cancers at other sites, based on
national statistics, resulting in mortality ratios of 1.7 and 0.73, respectively (statistical analysis of
data not reported).
Mur et al. (1987) analyzed the mortality of a cohort of 1143 workers in a plant that
refined and processed cobalt and sodium. The plant workers may have been involved in multiple
processing applications utilizing different forms of cobalt including cobalt chloride, oxides and
11

-------
8-25-2008
other salts (specific species not identified). An increase in deaths [Standard Mortality Ratio
(SMR) = 4.66; 95% confidence interval (CI) = 1.46-10.64] resulting from lung cancer was
observed in workers based on four cases observed in the exposed group and one case expected
based on French national statistics. In a study within the cohort that controlled for age and
smoking habits, 44% (four workers) in the exposed group and 17% (three workers) in the control
group died of lung cancer. The authors indicated that the differences were not statistically
significant and that the workers were exposed to arsenic and nickel in addition to cobalt. The
exposure levels of cobalt were not reported.
Tuchsen et al. (1996) analyzed the cancer incidence of a cohort of 874 women who
worked in one of two factories (382 from one factory, 492 from a second factory) applying a
cobalt-based (cobalt-aluminate spinel) plate underglaze. From unexposed areas of factory I,
520 referents were selected. Both groups were compared to statistics for all Danish women in
the same calendar year. During the 5-year follow-up period, the overall cancer incidence was
only slightly elevated in exposed workers, while the incidence of lung cancers was significantly
increased [Standard Incidence Ratio (SIR) = 2.35; 95% CI = 1.01-4.6], The incidence of lung
cancers in the referents (not exposed to cobalt) was greater than that of all Danish women, but
the difference was not statistically significant. Exposure characterization prior to 1980 was not
described, while exposures after 1980 were variable and reported as a mean concentration for a
given year. Exposures were generally in the range of 0-1 mg cobalt/m3 except for 2 years,
during which they were greater.
Animal Studies
Overview
Studies in animals show that oral exposure to cobalt produces effects similar to those
observed in humans, including increases in red blood cells and hemoglobin (Domingo et al.,
1984; Krasovskii and Fridlyand, 1971; Murdock, 1959; Holly, 1955; Stanley etal., 1947),
thyroid effects (Shrivastava et al., 1996) and cardiac effects (Haga et al., 1996; Pehrsson et al.,
1991; Mohiuddin et al., 1970). Other findings in animals not reported in humans include
neurobehavioral changes (Singh and Junnarkar, 1991; Bourg et al., 1985; Krasovskii and
Fridlyand, 1971) and testicular toxicity (Anderson et al., 1992, 1993; Pedigo et al., 1988; Corrier
et al., 1985; Mollenhauer et al., 1985; Domingo et al., 1984; Nation et al., 1983). Developmental
toxicity studies in rats and mice provide evidence that high oral doses of cobalt may produce
developmental effects in animals, in some cases in the absence of overt maternal toxicity
(Szakmary et al., 2001; Paternain et al., 1988; Domingo et al., 1985).
Animal data support the conclusion that the respiratory tract is the critical target for
inhaled cobalt (NTP, 1991; Bucher et al., 1990; Wehner et al., 1977). Subchronic inhalation
exposure to cobalt resulted in cytotoxicity and reparative proliferation in all regions of the
respiratory tract in rats and mice (NTP, 1991; Bucher et al., 1990). Available chronic animal
studies have demonstrated the carcinogenic potential of inhaled cobalt in male and female rats
and mice, with alveolar and bronchiolar tumors being the most prevalent (Bucher et al., 1999;
NTP, 1998).
12

-------
8-25-2008
Oral Exposure
Studies in rats show that subchronic oral exposure to cobalt chloride increases red blood
cell counts and hemoglobin levels with NOAELs ranging from 0.05 to 0.62 mg cobalt/kg-day
(Krasovskii and Fridlyand, 1971; Stanley et al., 1947) and LOAELs ranging from 0.5 to 32 mg
cobalt/kg-day (Domingo et al., 1984; Krasovskii and Fridlyand, 1971; Murdock, 1959; Holly,
1955; Stanley et al., 1947). In general, effects in animal studies were observed at higher
exposure levels than those reported in humans.
Effects of cobalt on red blood cells and hemoglobin were investigated in Sprague-Dawley
rats treated with 2.5, 10, and 40 mg cobalt chloride hexahydrate/kg-day (equivalent to 0.62, 2.5,
and 9.9 mg cobalt/kg-day, respectively) for 8 weeks (Stanley et al., 1947). After 8 weeks of
exposure, increases in hemoglobin and red blood cell number were observed in the 2.5 and
9.9 mg cobalt/kg-day treatment groups. Statistical significance was not reported.
Hemoglobin and hematocrit were significantly increased in male Sprague-Dawley rats
exposed to 500 ppm cobalt chloride in drinking water, equivalent to approximately 32 mg
cobalt/kg-day (assuming a water intake of 0.139 L/kg-day for male Sprague-Dawley rats;
U.S. EPA, 1988), for 3 months (Domingo et al., 1984). Compared to controls, hematocrit and
hemoglobin were both increased by approximately 30% at the end of the 3-month exposure
period, with increases observed within the first 2 weeks of exposure (numeric data not
presented). Following the 3-month exposure period, histopathological examination showed no
treatment-related morphological or ultrastructural changes to any organ. Increased tissue
weights were observed for spleen, heart and lungs, and testicular weight was decreased
compared to controls. Based on the results of this study, 32 mg cobalt/kg-day was identified as a
subchronic LOAEL for increased hematocrit and hemoglobin and decreased testicular weight in
rats.
In rats exposed to 40 mg cobalt chloride/kg-day (equivalent to 18 mg cobalt/kg-day) for
4 months, hemoglobin and red blood cell count were increased by 37 and 21%, respectively,
compared to controls (Holly, 1955). Similar effects were observed following concomitant
administration of 40 mg cobalt chloride/kg-day and 200 mg ferrous sulfate, with increases of
30%) for hemoglobin and 32%> for red blood cell count, compared to controls. Statistical
significance was not reported.
Oral exposure of rats to 10 mg cobalt/kg-day (as cobalt chloride) for 5 months resulted in
increases in hemoglobin, hematocrit and red blood cell count compared to untreated controls,
with effects reaching a plateau after approximately 60 days of exposure (Murdock, 1959).
Statistical significance was not reported. No changes were observed for mean corpuscular
hemoglobin concentration and mean cell volume compared to untreated controls, indicating that
stimulation of erythropoiesis by cobalt did not result in the production of abnormal red blood
cells.
The effects of exposure to 0.05, 0.5, and 2.5 mg cobalt/kg-day (as cobalt chloride) for
7 months were examined in rats (Krasovskii and Fridlyand, 1971). Treatment with 0.5 and
2.5 mg cobalt/kg-day, but not 0.05 mg cobalt/kg-day, for 7 months increased red blood cells and
hemoglobin. Stimulation of hematopoiesis was more pronounced in the 2.5 mg cobalt/kg-day
13

-------
8-25-2008
group than in the 0.5 mg cobalt/kg-day group, with polycythemia in the 0.5 mg cobalt/kg-day
group described as mild and transient. Results of this study are difficult to evaluate since
numeric data and statistical analyses were not reported.
Studies in animals have noted cardiac effects following cobalt (cobalt sulfate) exposure
(Haga et al., 1996; Pehrsson et al., 1991; Mohiuddin et al., 1970), although at higher exposure
levels than observed in human studies. The effect of cobalt on myocardial function was
examined in rats exposed to 8.4 mg cobalt/kg-day for 16 or 24 weeks (Haga et al. 1996). After
24 weeks of exposure, decreased left ventricular systolic and diastolic function was observed.
An increase in the ventricular weight to body weight ratio indicates that left ventricular
hypertrophy is a contributory factor in cobalt-induced myocardial dysfunction although a
mechanism was not identified. Significant effects on cardiac function were not observed
following 16 weeks of exposure. In guinea pigs, exposure to 20 mg cobalt/kg-day as cobalt
sulfate in the diet for 5 weeks resulted in decreased absolute and relative heart weights and a
greater incidence of abnormal electrocardiograms compared to animals fed on diets not
supplemented with cobalt (Mohiuddin et al., 1970). Cardiac arrhythmias, including bradycardia,
and repolarization abnormalities, were observed in 65% of cobalt-treated animals compared to
5% of control animals. Cellular alterations, observed at the light and electron microscopic
levels, in cardiac tissues included pericardial thickening and inflammation, myocardial
degeneration and vacuolization, endocardial thickening and myofibrillar damage. In contrast, no
effects on cardiac function were observed in male rats (12/group) exposed to protein-restricted
diets containing 8.4 mg cobalt/kg-day for 8 weeks (Pehrsson et al., 1991). Treated rats showed a
significant decrease in body weight but no differences in left ventricular function relative to
animals treated with protein-restricted diets without added cobalt. Although the results from the
Pehrsson et al. (1991) and Haga et al. (1996) rat studies conflict, it appears that oral
cobalt-induced myocardial injury/dysfunction may have a significant time-dependence. Oral
cobalt (as cobalt sulfate) at the same dose level (8.4 mg cobalt/kg-day) did not appear to alter
cardiac structure or function following exposure for up to 16 weeks (Pehrsson et al., 1991; Haga
et al., 1996). However, ventricular hypertrophy with a concomitant decrease in left ventricular
systolic and diastolic function was observed in rats after 24 weeks of oral cobalt (Haga et al.,
1996). Thus, based on the results of this study, 8.4 mg cobalt/kg-day represents a subchronic
LOAEL for myocardial toxicity in rats; Based on the results of the Mohiuddin et al. (1970)
study, a LOAEL of 20 mg cobalt/kg-day was identified for myocardial toxicity in guinea pigs.
Histopathological changes in the thyroid gland have been observed following exposure of
female mice to 400 ppm cobalt chloride (-48 mg cobalt/kg-day, assuming an average water
intake of 0.265 L/kg-day for female mice; U.S. EPA, 1988) in drinking water for 15 to 45 days
(Shrivastava et al., 1996). The severity of effect increased with exposure duration. After
15 days of exposure, a reduction in thyroid epithelial cell height with degenerated nuclei and
reduced amount of colloid with peripheral resorption vacuoles was observed, with more
pronounced effects after 30 days of exposure. More significant degenerative changes were
observed after 45 days of exposure, including necrotic epithelial cells, reduced connective tissue
between follicles, lymphocytic infiltrate and larger amounts of colloid within the lumen. Based
upon significant thyroid toxicity observed in this study, a LOAEL of 48 mg cobalt/kg-day was
identified in mice.
14

-------
8-25-2008
Developmental effects of orally administered cobalt have been studied in rats, rabbits and
mice (Szakmary et al., 2001; Pedigo and Vernon, 1993; Paternain et al., 1988; Seidenberg et al.,
1986; Domingo et al., 1985; Elbetieha et al., 2008). Szakmary et al. (2001) evaluated the
developmental effects of oral cobalt sulfate exposure in rats, mice and rabbits. Exposure of
pregnant rats to 5.2-21.0 mg cobalt/kg-day (oral gavage) decreased perinatal growth and
survival, retarded skeletal development and produced skeletal and urogenital malformations,
with a LOAEL of 5.2 mg cobalt/kg-day. Maternal toxicity (increased relative liver, adrenal,
spleen weights; increased BUN, serum creatinine) was only observed at the highest dose
(21.0 mg cobalt/kg-day). Thus, embryotoxicity in rats was observed at exposure levels below
the LOAEL for maternal toxicity. In pregnant mice exposed to 10.5 mg cobalt/kg-day, retarded
skeletal development and malformations of the eye, kidney and skeleton were observed in the
absence of maternal toxicity. In pregnant rabbits exposed to 4.2 mg cobalt/kg-day, 20%
mortality was observed in dams. Fetal resorptions were observed in 30% of surviving dams.
Results of the studies in rats and mice provide evidence that adverse developmental effects can
occur in the absence of maternal toxicity, and that rabbits are more sensitive to oral cobalt.
Domingo et al. (1985) treated pregnant female rats (15 animals/group) with 5.4 to
21.8 mg cobalt/kg-day as cobalt chloride from gestation day 14 through lactation day 21.
Offspring were examined for mortality, body weight, body and tail length and general signs of
toxicity after 1, 4 and 21 days of nursing. In contrast to the study by Szakmary et al. (2001),
results of the Domingo et al. (1985) study reported maternal toxicity at all doses that produced
adverse developmental effects (specific maternal effects observed were not reported). Fetal
effects at 5.4 mg cobalt/kg-day included stunted growth of the pups of both sexes, decreased
body length and tail length in male offspring and decreased spleen and liver weight in female
offspring. Effects at the 10.9 mg cobalt/kg-day dose included decreased body weight in female
pups, while at 21.8 mg cobalt/kg-day, decreased number of living young and decreased survival
were seen. Blood parameters (liver enzymes, bilirubin, total protein, uric acid, urea, creatinine,
hemoglobin and hematocrit) in pups did not show any treatment-related changes. No signs of
toxicity were observed in surviving pups in any of the cobalt exposure groups.
No significant effects on fetal growth or survival were found in rats exposed to 6.2 to
24.8 mg cobalt/kg-day as cobalt chloride (oral gavage) during gestation days 6-15 (Paternain et
al., 1988). The incidence of stunted fetuses was higher in the animals treated with 12.4 or 24.8
mg cobalt/kg-day (0.3 stunted fetuses per litter in the 12.4 mg cobalt/kg-day group; 1.0 stunted
fetuses per litter in the 24.8 mg cobalt/kg-day group) compared to the control group (0 stunted
fetuses per litter); however, the differences were not statistically significant. No treatment-
related effects were observed for the number of corpora lutea, total implants, resorptions, the
number of dead and live fetuses or fetal size parameters. No gross external abnormalities,
skeletal malformations or other signs of fetal toxicity were observed. Maternal effects, including
reduced body weight gain and food consumption and altered hematological parameters
(increased hematocrit, hemoglobin and reticulocytes), were reported at all exposure levels. No
fetal effects were reported in mice exposed to 81.7 mg cobalt/kg-day (oral gavage) during
gestation days 8-12 (Seidenberg et al., 1986), but a significant (p<0.05) decrease in maternal
weight was found. Additional details were not reported.
15

-------
8-25-2008
Pedigo and Vernon (1993) exposed male B6C3F1 mice to 400 ppm cobalt chloride
(-45 mg cobalt/kg-day, assuming a water intake of 0.247 L/kg-day for male B6C3F1 mice;
U.S. EPA, 1988) in the drinking water for 10 weeks, after which the males were mated with
control females to examine for dominant lethal effects. Relative to the control group, the cobalt
treatment group had a lower percentage of pregnant females (control, 29/32; cobalt, 18/31),
lower number of implantations per female (control, 8.3; cobalt, 6.5) and higher preimplantation
losses (control, 0.43; cobalt, 2.4). At the end of the 10-week treatment period, sperm
concentration was decreased to 15.3% and motility decreased to 18.3% of controls. Several
measures of sperm velocity were also depressed relative to controls. All sperm parameters,
except sperm concentration, returned to control levels 8 weeks after the cobalt exposure was
terminated. The increase in preimplantation losses in the dominant lethal assay appears related
to adverse effects on spermatogenesis rather than to effects on preimplantation development of
embryos.
Several studies reported testicular degeneration and atrophy in rats exposed to 11.7 to
46.9 mg cobalt/kg-day as cobalt chloride for 2-3 months in the diet or in the drinking water
(Anderson et al., 1992, 1993; Pedigo et al., 1988; Corner et al., 1985; Mollenhauer et al., 1985;
Domingo et al., 1984; Nation et al., 1983). Pedigo et al. (1988) exposed male CD-I mice to 100,
200 or 400 ppm of cobalt chloride (~ 11.7, 23.4 or 46.9 mg cobalt/kg-day, respectively, assuming
an average water intake of 0.258 L/kg-day for male mice; U.S. EPA, 1988) in the drinking water
for 13 weeks. High-dose animals showed a significantly decreased testicular weight beginning at
week 9 of treatment and a decreased epididymal sperm concentration by week 11 of treatment.
All dose groups showed significantly decreased testicular weight and epididymal sperm
concentration and increased serum testosterone levels by week 12 of exposure, with the
magnitude increasing with dose. Effects on serum testosterone levels may be secondary to
effects on spermatogenesis and related to inhibition of local inhibitory feed-back mechanisms.
Based on the results of this study, 11.7 mg cobalt/kg-day was identified as a subchronic LOAEL
for decreased testicular weight and epididymal sperm concentration in male rats.
Anderson et al. (1992, 1993) exposed groups of male CD-I mice to 400 ppm of cobalt
chloride (-46.9 mg cobalt/kg-day, assuming an average water intake of 0.258 L/kg-day for male
mice; U.S. EPA, 1988) in the drinking water for up to 13 weeks. A decrease in testicular weight
and a progressive degeneration of the seminiferous tubules were seen beginning at 9 weeks of
exposure. Initial changes were vacuolization of Sertoli cells and abnormal spermatid nuclei,
followed by sloughing of cells, shrinkage of tubules and thickened endothelium. No recovery
was reported after a 20-week non-exposure recovery period. Co-administration of 800 ppm of
zinc chloride provided a partial protection against the effects of cobalt. Based on the results of
this study, 46.9 mg cobalt/kg-day was identified as a subchronic LOAEL for decreased testicular
weight and degeneration of seminiferous tubules in male mice. Similar histology (degeneration
of the testes, particularly the seminiferous tubules) was noted in Sprague-Dawley rats exposed to
20 mg cobalt/kg-day in the diet for up to 98 days (Corner et al., 1985; Mollenhauer et al., 1985).
Decreased testicular weight was seen in Sprague-Dawley rats exposed to 500 ppm cobalt
chloride (-32 mg cobalt/kg-day, assuming a water intake of 0.139 L/kg-day for male Sprague-
Dawley rats; U.S. EPA, 1988) for 3 months (Domingo et al., 1984).
16

-------
8-25-2008
Elbetieha et al. (2008) examined the potential effects of cobalt on male fertility in forty
adult (60 day-old) male Swiss mice exposed to cobalt chloride hexahydrate via drinking water at
concentrations of 200, 400, or 800 ppm for 12 weeks. Based on daily water intake reported in
the study, daily average doses of cobalt chloride were estimated at 26, 47, or 93 mg/kg-day
(equivalent to 6.5, 11.7, or 23 mg cobalt/kg-day); control animals received untreated tap water.
Mice were observed daily for signs of clinical toxicity during the exposure period. At the end of
the 12-week cobalt exposure period, male mice were separated into individual cages containing
two virgin Swiss female mice and given ad libitum access to food and untreated tap water. Mice
were cohabitated for 10 days during which it was estimated that the females completed two
estrus cycles. Male control and cobalt-treated mice were necropsied after day 10 of cohabitation
and testes, seminal vesicles, epididymides and preputial glands were harvested, weighed, and
prepared for analysis. The left testis and epididymis from each male mouse was processed for
determination of sperm count, while the right testis was processed for histopathology. Ten days
later, female mice were necropsied and examined for number of pregnancies, number of
implantation sites, number of viable fetuses, total number of resorptions, and incidence rate of
resorptions.
Ingestion of cobalt chloride was associated with 1/10 and 2/10 deaths in the mid- and
high-dose treatment groups, respectively, during week 10 of exposure. Average body weight
gain was significantly reduced in all cobalt treatment groups (p < 0.01). No other signs of
clinical toxicity were observed in surviving male mice. Relative to the control group, the
number of pregnant females mated with male mice from the mid- and high-dose groups was
significantly (p < 0.05) reduced (control, 19/20; mid-dose, 12/18; high-dose, 7/16). The number
of implantation sites was significantly (p < 0.01) reduced in females mated with low- and mid-
dose males (control, 7.89; low-dose, 5.67; mid-dose, 5.42), and the number of viable fetuses was
significantly (p < 0.05) reduced in females mated with males from all cobalt treatment groups
(control, 7.74; low-dose, 5.0; mid-dose, 4.67; high-dose, 5.83). In addition, the total number of
resorptions (control, 3/150; low-dose, 9/81; mid-dose, 9/65; high-dose, 10/45) and the number of
animals with resorptions (control, 3/19; low-dose, 10/15; mid-dose, 10/16; high-dose, 5/7) were
significantly (p < 0.05) increased in females mated with males from all three cobalt-treatment
groups. Analysis of male reproductive organs revealed a significant (p < 0.005) decrease in
absolute epididymal weight in mice of the high-dose treatment group. Testes weights were
significantly (p < 0.01) reduced in males at all doses of cobalt, and a significant (p < 0.005)
increase in the absolute weight of seminal vesicles of the mid- and high-dose males only.
Compared to controls, testicular sperm counts and daily sperm production were decreased in the
mid- and high-dose males, but not in the low-dose animals. Epididymal sperm counts were
decreased in male mice from all three cobalt treatment groups. Histopathological examination of
testis tissue from males of the mid- and high-dose revealed a number of abnormalities including
necrosis of the seminiferous tubules and interstitium, congested blood vessels, hypertrophy of the
interstitial Leydig cells, and degeneration of the spermatogonial cells; incidence rate of these
observations was not reported. These testicular histopathologies were not observed in the testes
of control and low-dose treated males. Based on the results of this study, a LOAEL of
6.5 mg/kg-day was identified for decreased testicular weight, epididymal sperm counts, and
associated reproductive abnormalities in pregnant females. A NOAEL was not identified.
17

-------
8-25-2008
Nation et al. (1983) exposed groups (n=6) of male Sprague-Dawley rats (weighing
200-210 g) to diets containing 0, 5 or 20 mg cobalt/kg-day as cobalt chloride for a total of
69 days. Following 14 days of exposure, animals were trained for scheduled (operant) or
conditioned suppression neurobehavioral tests. Other than two seizures in the same high-dose
animal, no overt signs of neurotoxicity were reported at any exposure level. A trend toward a
decreased response rate in the schedule training behavior was observed in both the exposed
groups but only attained statistical significance in the high-dose animals near the end of the
operant testing period (sessions 28-35, on exposure days 44-51). A trend toward decreased
conditioned suppression behavior did not attain statistical significance in either group. Animals
exposed to 20 mg cobalt/kg-day, but not 5 mg cobalt/kg-day, showed a significantly decreased
weight of the testes following 69 days of exposure. Based on the results of this study, a NOAEL
of 5 mg cobalt/kg-day and a LOAEL of 20 mg cobalt/kg-day was identified for decreased
testicular weight and changes in operant behavior in male Sprague-Dawley rats.
Several other studies have examined the effects of cobalt on neurobehavioral parameters
(Singh and Junnarkar, 1991; Krasovskii and Fridlyand, 1971; Bourg et al., 1985). In groups of
male Sprague-Dawley rats (n=8) exposed to 20 mg cobalt/kg-day as cobalt chloride for 57 days
in the drinking water, cobalt enhanced behavioral reactivity to stress (the animals were less likely
to descend from a safe platform to an electrified grid) (Bourg et al., 1985). Singh and Junnarkar
(1991) reported a moderate reduction in spontaneous activity and mild hypothermia in rats
exposed orally to cobalt chloride (approximately 8 mg cobalt/kg-day) or cobalt sulfate
(approximately 35 mg cobalt/kg-day). Krasovskii and Fridlyand (1971) exposed groups of rats
(number and sex not specified) to 0.05, 0.5 or 2.5 mg cobalt/kg-day as cobalt chloride for up to
7 months. Neurobehavioral tests showed that treatment with cobalt resulted in a significant
(p<0.05) increase in the latent reflex period at 0.5 mg cobalt/kg and above, and a pronounced
neurotropic effect (disturbed conditioned reflexes) at 2.5 mg cobalt/kg.
Inhalation Exposure
In a subchronic inhalation study, groups of 10 F344/N rats and 10 B6C3F1 mice of each
sex were exposed to cobalt sulfate hexahydrate aerosol (MMAD=0.83-1.10 [j,m; og not reported)
at concentrations of 0, 0.3, 1, 3, 10 or 30 mg/m3 (equivalent to 0, 0.067, 0.22, 0.67, 2.2 or 6.7 mg
cobalt/m3) 6 hours/day, 5 days/week for 13 weeks (Bucher et al., 1990; NTP, 1991). Although
this report indicates that exposure was to cobalt sulfate heptahydrate aerosol, detailed analysis of
the cobalt aerosol in the 2-year continuation study (Bucher et al., 1999; NTP, 1998) reports that
the aerosol was actually composed of cobalt sulfate hexahydrate; thus, exposure to the
hexahydrate form is assumed for the 13-week study. Animals were monitored for body weight
and observed for clinical signs during the exposure period. Urine samples for urinalysis and
cobalt determination were collected from rats prior to sacrifice. Following termination of
exposure, all animals were sacrificed and necropsied. Blood samples were collected and
analyzed for hematological parameters (rats and mice) and serum chemistry and thyroid function
parameters (rats only). The major organs were weighed. Animals from the control and
high-dose groups received comprehensive histopathological examinations, while those from the
lower dose groups received more limited examinations focused on the respiratory tissues.
18

-------
8-25-2008
All rats survived until scheduled necropsy (NTP, 1991; Bucher et al., 1990). Gross
"3
evidence of toxicity was noted only in rats exposed to 6.7 mg cobalt/m , and they displayed
clinical signs of toxicity (ruffled fur, hunched posture) and reduced body weights. Polycythemia,
indicated by significant increases in red blood cell count, hemoglobin and hematocrit, was noted
in males exposed to >0.67 mg cobalt/m3 and females exposed to >2.2 mg cobalt/m3. In addition,
"3
platelets were significantly reduced in rats of both sexes at >2.2 mg cobalt/m and reticulocytes
were increased in females at 6.7 mg cobalt/m3. Leukocyte counts and differentials were
"3
unaffected. Serum cholesterol was significantly reduced in males at >2.2 mg cobalt/m and
females at 6.7 mg cobalt/m3. No other serum chemistry parameters were affected, including
creatine kinase isozymes indicative of damage to cardiac muscle cells. Among the thyroid
hormones, T3 (triiodothyronine) was significantly reduced in females at 2.2 mg cobalt/m3 (83%
"3
of control) and males at 6.7 mg cobalt/m (62% of control) and TSH (thyrotropin) was
significantly reduced in males at 6.7 mg cobalt/m3 (30% of control), but T4 (thyroxine) was not
affected in either sex at any dose and the researchers concluded that thyroid function was not
consistently affected in this study. Urinalysis revealed a dose-related increase in the number of
epithelial cells and granular casts in the urine of many exposed male rats (3-7 per group exposed
to >0.67 mg cobalt/m3) but not in the urine of control male rats. The researchers interpreted this
finding as indicating minimal nephropathy in exposed male rats although histopathological
lesions were not detected in the kidney. No effects on sperm counts, sperm motility or the
incidence of abnormal sperm were noted. Average estrus cycle of females exposed to 6.7 mg
cobalt/m3 was slightly longer than controls, but the difference was not significant. Absolute and
relative lung weights were significantly increased in both male and female rats at >0.22 mg
cobalt/m3. Other organ weights were not affected by treatment. Compound-related lesions were
found only in the respiratory tissues of exposed rats. Degenerative, inflammatory and
regenerative lesions were found throughout the respiratory tract (see Table 2). Incidence and
severity of lesions were similar in males and females. The most sensitive tissue was the larynx,
with squamous metaplasia present at all exposure levels.
"3
Among mice, 2/10 males exposed to 6.7 mg cobalt/m died during the study (NTP, 1991;
Bucher et al., 1990). The only clinical signs of toxicity observed were rapid breathing and skin
discoloration in one of the mice that died. Body weights were reduced throughout the study in
both males and females exposed to 6.7 mg cobalt/m3. No dose-related hematological effects
were found. Absolute and relative lung weights were significantly increased in male and female
mice exposed to >2.2 mg cobalt/m3. Respiratory lesions were similar to those observed in rats.
As with rats, the most sensitive tissue was the larynx, with squamous metaplasia present at all
exposure levels. Reproductive system effects were more prominent in mice than rats. Males had
significantly decreased testicular weight (48% compared to control), decreased epididymal
weight (81%) compared to control), testicular atrophy consisting of loss of germinal epithelium in
the seminiferous tubules and foci of mineralization and an increased percentage of abnormal
sperm at 6.7 mg cobalt/m3 (295% compared to control). Significant reductions in sperm motility
"3
of 90, 87 and 54% were observed in the 0.67, 2.2 and 6.7 mg cobalt/m exposure groups,
respectively (lower doses were not tested). Females had a significantly increased length of the
"3
estrus cycle at 6.7 mg cobalt/m (119%> longer compared to control).
19

-------
8-25-2008
Table 2. Rats with Selected Lesions in the 13-Week Cobalt Sulfate Inhalation Studya


Exposure Group (mg Cobalt (Co) per m3)
Site
Lesion
Control
0.067
mg Co/m3
0.22
mg Co/m3
0.67
mg Co/m3
2.2
mg Co/m3
6.7
mg Co/m3
Larynx
Inflammation
M: 0
F: 1
M: 2
F: 2
M: 8C
F: 7 0
M: 9C
F: 10 c
M: 9C
F: 10 c
M: 9C
F: 10c

Squamous metaplasia
M: 0
F: 1
M: 9C
F: 7 0
M: 10 c
F: 10 c
M: 10 c
F: 10 c
M: 10 c
F: 10 c
M: 10 c
F: 10 c
Lung
Inflammation
M: 0
F: 0
M: 0
F: 0
M: 6b
F: 2
M: 10 c
F: 9C
M: 10 c
F: 10 c
M: 10 c
F: 10 c

Fibrosis
M: 0
F: 0
M: 0
F: 0
M: 0
F: 0
M: 0
F: 1
M: 1
F: 4b
M: 10c
F: 5 b

Bronchiolar epithelium
regeneration
M: 0
F: 0
M: 0
F: 0
M: 0
F: 0
M: 0
F: 0
M: 0
F: 0
M: 7 0
F: 5b
M: number of males with lesions out of 50 animals.
F: number of females with lesions out of 50 animals.
aNTP, 1991; Bucher etal., 1990
b/?<0.05 vs controls by Fisher exact test
c/?<0.01 vs controls by fisher exact test
Other studies in animals have also reported respiratory lesions and altered respiratory
function following inhalation exposure to cobalt. Kyono et al. (1992) observed mild pulmonary
"3
lesions in rats exposed to 2.12 mg/m of cobalt aerosols (generated from an aqueous suspension
of ultrafine metallic cobalt particles) 5 hours/day for 4 days. Lesions were characterized by focal
hypertrophy of the epithelium, abnormal macrophages, vacuolization of type I epithelial cells
and proliferation of type II epithelial cells, which are indicative of an initial inflammatory
"3
response. Kerfoot et al. (1975) exposed groups of five miniature swine to 0, 0.1 or 1.0 mg/m of
pure cobalt metal powder for 6 hours/day, 5 days/week for 3 months. Wheezing was observed in
animals from both cobalt groups after 4 weeks of exposure (numeric data not reported). Tidal
volume was decreased to 73% and 64% of controls in the low and high dose groups,
respectively, and total respiratory compliance was decreased relative to controls (low dose, 66%
of control; high dose, 56% of control). Statistical significance was not reported. Examination of
lung tissue by electron microscopy revealed septa thickened by collagen, elastic tissue and
fibroblasts in both exposure groups, with more pronounced effects in the high dose group.
"3
Johansson et al. (1987) exposed rabbits (8/group) to 0.4 or 2 mg cobalt/m as cobalt chloride,
6 hours/day, 5 days/week for 14-16 weeks. Nodular accumulation of alveolar type II cells
(8/8 rabbits in both cobalt groups), abnormal accumulation of enlarged, vaculolated alveolar
macrophages (5/8 in the low dose group and 8/8 in the high dose group) and interstitial
inflammation (4/8 rabbits in the low dose group and 8/8 rabbits in the high dose group) were
observed, with more pronounced effects in the high dose group.
The carcinogenicity of inhaled cobalt was investigated in groups of 50 F344/N rats and
50 B6C3F1 mice of each sex exposed to cobalt sulfate hexahydrate aerosol
(MMAD=1.4-1.6 [j,m; og=2.1-2.2) at concentrations of 0, 0.3, 1 or 3 mg/m3 (equivalent to 0,
20

-------
8-25-2008
0.067, 0.22 or 0.67 mg cobalt/m3) 6 hours/day, 5 days/week for 105 weeks (Bucher et al., 1999;
NTP, 1998). Animals were monitored for body weight and observed for clinical signs during the
exposure period. Following termination of exposure, all animals were sacrificed and necropsied.
At necropsy, all organs and tissues were examined for gross lesions, trimmed and examined
histologically.
In F344 rats, there were no changes in survival or mean body weights in males or females
of any exposure group (Bucher et al., 1999; NTP, 1998). Irregular breathing was noticed more
"3
frequently in female rats exposed to 0.67 mg cobalt/m than in controls or other treatment
groups; no changes in clinical signs were noted in any of the treated male rats. Incidence of
selected neoplasms and nonneoplastic lesions of the lung in rats is summarized in Table 3. Both
male and female rats in all exposure groups showed a high incidence (94% or greater) of
squamous metaplasia of the alveolar epithelium, fibrosis of the pulmonary interstitium and
granulomatous inflammation, with all lesions increasing in severity with increasing exposure
level. Significant increases in alveolar/bronchiolar adenomas or carcinomas were seen in
high-dose male rats, while significant increases in alveolar/bronchiolar adenomas or carcinomas
were seen in the mid- and high-dose female rats. The combined incidence of
alveolar/bronchiolar neoplasms (adenoma and carcinoma) in male rats and female rats was
significantly greater than that in control animals, and a significant linear trend occurred in both
sexes. Rats of both sexes showed treatment-related increases in hyperplasia of the lateral nasal
wall, atrophy of the olfactory epithelium and squamous metaplasia of the larynx. A significant
increase in the incidence of pheochromocytoma in 0.67 mg cobalt/m3 dosed females was also
noted (2/48, 1/49, 4/50 and 10/50 in control, 0.067, 0.22 and 0.67 mg cobalt/m3 groups,
respectively). A marginally increased incidence of pheochromocytoma in males exposed to
3	3
0.22 mg cobalt/m , but not in those exposed to 0.67 mg cobalt/m , was considered by the study
authors not to be related to treatment.
In B6C3F1 mice, no changes in survival were observed in any exposure group (Bucher et
al., 1999; NTP, 1998). Male mice exposed to 0.67 mg cobalt/m3 showed a decreased mean body
weight relative to controls from week 96 through the end of the study (105 weeks). Mean body
weights of exposed female mice were generally greater than those of controls throughout the
study. Irregular breathing was noted slightly more frequently in female mice exposed to 0.22 mg
cobalt/m3 than in controls or other exposed groups. Incidence of selected neoplasms and
nonneoplastic lesions of the lung in mice is summarized in Table 4. A dose-related increase in
the occurrence of cytoplasmic vacuolization of the bronchus was seen in both sexes of mice, with
incidences at all exposure levels being significantly different from controls. As in rats, both
sexes of mice showed a significant linear trend toward increased alveolar/bronchiolar tumors,
3	3
with the 0.67 mg cobalt/m male and the 0.22- and 0.67 mg cobalt/m female groups attaining
statistical significance. Mice of both sexes showed significantly increased incidences of
squamous metaplasia of the larynx (/;<0,05) at all exposure levels examined. In male mice, but
not in females, the incidence of hemangiosarcoma was significantly elevated in animals exposed
"3
to 0.22 mg cobalt/m , but not in other exposure groups (2/50, 4/50, 8/50 and 7/50 in the control,
0.067, 0.22 and 0.67 mg cobalt/m3 groups, respectively).
21

-------
8-25-2008
Table 3. Incidence of Selected Neoplasms and Nonneoplastic Lesions in the Respiratory
Tract of Rats in the 2-Year Inhalation Study of Cobalt Sulfatea



Exposure Group (mg Cobalt (Co) per m3)



0.067
0.22
0.67
Site
Lesion Type
Control
mg Co/m3
mg Co/m3
mg Co/m3
Lung
Alveolar epithelium hyperplasia
M: 9
M: 20 b
M: 20 b
M: 23c


F: 15
F: 7
F: 20
F: 33 c

Alveolar epithelium metaplasia
M: 0
M: 50 c
M: 48 c
M: 49 c


F: 2
F: 47 c
F: 50 c
F: 49 c

Inflammation granulomatous
M: 2
M: 50 c
M: 48 c
M: 50 c


F: 9
F: 47 c
F: 50 c
F: 49 c

Alveolar/bronchiolar adenoma
M: 1
M: 4
M: 1
M: 6


F: 0
F: 1
F: 10 c
F: 9C

Alveolar/bronchiolar carcinoma
M: 0
M: 0
M: 3
M: 1


F: 0
F: 2
F: 6b
F: 6b

A/B adenoma or carcinoma
M: 1
M: 4
M: 4
M: 7b


F: 0
F: 3
F: 15 c
F: 15 c

Squamous cell carcinoma
M: 0
M: 0
M: 0
M: 0


F: 0
F: 0
F: 1
F: 1
Nose
Lateral wall hyperplasia
M: 2
M: 14 c
M: 21c
M: 21c


F: 1
F: 8b
F: 26 c
F: 38 c

Olfactory epithelium atrophy
M: 8
M: 24 c
M: 42 c
M: 48 c


F: 5
F: 29 c
F: 46 c
F: 47 c
Larynx
Squamous metaplasia
M: 0
M: 10 c
M: 37 c
M: 50 c


F: 1
F: 22 c
F: 39 c
F: 48 c
M: Incidence of lesions in male rats out of 50 animals.
F: Incidence of lesions in female rats out of 50 animals.
a Bucher et al., 1999; NTP, 1998
b/?<0.05 compared to control by logistic regression test
0 /?<(),()1 compared to control by logistic regression test
22

-------
8-25-2008
Table 4. Incidence of Selected Neoplasms and Nonneoplastic Lesions in the Respiratory
Tract of Mice in the 2-Year Inhalation Study of Cobalt Sulfatea


Exposure Group (mg cobalt (Co) per cubic meter)
Site
Lesion Type
Control
0.067
mg Co/m3
0.22
mg Co/m3
0.67
mg Co/m3
Lung
Bronchus cytoplasmic vacuolization
M: 0
F: 0
M: 18 c
F: 6b
M: 34 c
F: 31c
M: 38 c
F: 43 c

Alveolar/bronchiolar adenoma
M: 9
F: 3
M: 12
F: 6
M: 13
F: 9
M: 18 b
F: 10 b

Alveolar/bronchiolar carcinoma
M: 4
F: 1
M: 12
F: 1
M: 13
F: 4
M: 18 b
F: 9C

A/B adenoma or carcinoma
M: 11
F: 4
M: 14
F: 7
M: 19
F: 13 c
M: 28 c
F: 18 c
Nose
Olfactory epithelium atrophy
M: 0
F: 0
M: 0
F: 2
M: 28 c
F: 12 c
M: 48 c
F: 46 c

Hyperplasia
M: 0
F: 0
M: 0
F: 0
M: 0
F: 0
M: 10 c
F: 30 c
Larynx
Squamous metaplasia
M: 0
F: 0
M: 37 c
F: 45 c
M: 48 c
F: 40 c
M: 44 c
F: 50 c
M: Incidence of lesions in male mice out of 50 animals.
F: Incidence of lesions in female mice out of 50 animals.
a Bucher et al., 1999; NTP, 1998
b/?<0.05 compared to control by logistic regression test
0 /?<(),()1 compared to control by logistic regression test
Wehner et al. (1977, 1979) exposed 2-month-old male Syrian golden hamsters to inhaled
cobalt oxide at 0 or 10 mg/m3 (51 animals/group), 7 hours/day, 5 days/week for approximately
15 months. The incidence of tumors in treated hamsters was not statistically different from
controls. There was "limited" histopathologic and ultrastructural examination in the study. No
developmental toxicity studies were located following inhalation exposure to cobalt.
Other Studies
Parenteral Administration
Heath (1956) injected groups of 10 male and 20 female rats with a single intramuscular
28 mg dose of powdered cobalt in the thigh. Injection-site sarcomas appeared in 18 (60%) of the
treated rats within 5-12 months. Similar results were observed in Wistar rats by Gilman (1962)
and Gilman and Ruckerbauer (1962), with single intramuscular doses of 20 mg of cobalt oxide
and cobalt sulfide. Cobalt oxide and cobalt sulfide given intramuscularly at doses twice those
used in rats did not induce sarcomas in mice (Gilman and Ruckerbauer, 1962). Shabaan et al.
(1977) observed a high incidence of fibrosarcomas in rats given subcutaneous injections of
cobalt chloride at 40 mg/kg-day for 10 days. Tumors developed in 8-12 months. Stoner et al.
(1976) tested cobalt acetate in the strain A mouse pulmonary tumor test. Groups of 20 mice/sex
23

-------
8-25-2008
received three times per week intraperitoneal injections for a total of 19 cumulative doses of 0,
95, 237 or 475 mg/kg. Survival was high over the 30-week observation period, and the
incidence of lung tumors in treated mice was not statistically different from controls.
Genotoxicity Studies
The genetic toxicity of cobalt was reviewed by Beyersman and Hartwig (1992) and more
recently by De Boeck et al. (2003b), Hartwig and Schwerdtle (2002) and Lison et al. (2001).
Cobalt compounds have generally tested negative in bacterial mutagenicity assays, with
occasional positive results occurring with the addition of an exogenous metabolic system. In
contrast, cobalt compounds have generally tested positive in yeast and plant cells. In mammalian
cell systems, cobalt has been shown to induce DNA strand breaks, sister-chromatid exchanges
and morphological cell transformation.
Results of in vitro studies using human peripheral blood mononucleated cells show that
cobalt metal and cobalt chloride induced DNA strand breaks at non-cytotoxic concentrations (De
Boeck et al., 1998, 2003a). Evidence demonstrating mutagenic activity of cobalt in vivo in
humans is lacking. No significant change in DNA strand breaks were observed in lymphocytes
from nonsmoking workers who had been occupationally exposed to cobalt or hard metal dust
although a positive association was observed between DNA strand breaks and smoking (De
Boeck et al., 2000).
Experimental data in animals provide evidence of genotoxicity following in vivo
exposure to cobalt. Single oral exposure of male Swiss mice to 0, 4.96, 9.92 or 19.8 mg
cobalt/kg-day, as cobalt chloride, resulted in significantly increased percentages of both
chromosomal breaks and chromosomal aberrations in bone marrow cells, with significant linear
trends toward increasing aberrations with increased exposure (Palit et al., 1991a,b,c,d). Thirty
hours following single intraperitoneal injection of cobalt chloride at doses of 6.19, 12.4, or
22.3 mg cobalt/kg in BALB/c mice, an increase in micronucleus formation was seen in the mid-
and high-dose mice but not in low-dose mice (Suzuki et al., 1993). Single injection of 12.4 mg
cobalt/kg resulted in significantly increased micronucleus formation at 24 hours post-injection
but not at 12, 48, 72 or 96 hours. Pedigo and Vernon (1993) reported that treatment with
400 ppm cobalt chloride (-45 mg cobalt/kg-day, assuming a water intake of 0.247 L/kg-day for
male B6C3F1 mice; U.S. EPA, 1988) in the drinking water of male B6C3F1 mice for 10 weeks
resulted in an increase in dominant lethal effects as indicated by changes in the number of
pregnant females, percentage of live embryos and number of pre-implantation losses per female.
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC
ORAL RfD VALUES FOR COBALT
Indicators of human health effects following oral exposure to cobalt (Co) include
increased erythrocyte production and hemogloblin levels, decreased iodine uptake by the thyroid
gland, elicitation of dermatitis in sensitized individuals and cardiomyopathy. Observations in
humans for effects on the heart, blood and the thyroid gland are supported by results of studies in
animals. Other effects, including neurobehavioral, developmental and testicular toxicity were
24

-------
8-25-2008
observed in animals and at relatively high doses; these endpoints were not considered further for
the development of the subchronic or chronic provional RfD (p-RfD).
Cardiomyopathy was considered as an endpoint of concern for cobalt exposure in
humans; however, it is probable that alcohol consumed in "beer-cobalt cardiomyopathy," as well
as other associated factors such as nutritional deficiency, played a role in the cardiotoxic effects
observed. Therefore, a dose-response relationship could not be determined for cobalt exposure
from these studies. Studies in animals have noted cardiac effects following cobalt exposure at
higher exposure levels than observed in human studies of "beer-cobalt cardiomyopathy." On this
basis, cardiomyopathy was not selected as the critical endpoint for p-RfD derivation.
Allergic response in cobalt-sensitized workers was considered as a potential critical
endpoint for the derivation of an oral p-RfD. However, the available data provide no
information on the dose-response relationship of cobalt sensitization, nor is a no observable
adverse effect level (NOAEL) for the elicitation of an allergic response in humans defined.
Interrelationships also exist between cobalt and nickel (Ni) sensitization so that people sensitized
by (Ni) may have an allergic reaction following cobalt exposure. Allergic response was,
therefore, not chosen as the critical effect for p-RfD derivation.
Cobalt has been shown to induce polycythemia which is characterized by an increase in
erythrocyte number and hemoglobin levels through stimulation of erythropoietin, a hormone
produced primarily in the kidney. The hematological effects of cobalt treatment have been
reported in healthy, non-anemic adults (Davis and Fields, 1958) and in anephric anemic dialysis
patients (Taylor et al., 1977; Duckham and Lee, 1976). However, the effects observed in healthy
adults were reversible and erythrocyte counts returned to pre-treatment levels within 9 to 15 days
after cobalt administration was discontinued. In anephric dialysis patients, treatment with cobalt
resulted in an increase in hemoglobin from levels clinically described as "anemic" to levels at or
near "normal." Thus, the effect of cobalt administration in these patients was clinically
beneficial. Furthermore, the results of this study are difficult to interpret due to confounding
factors, including the anephric status of patients and the concomitant administration of iron.
Hematologic effects of cobalt were also found in several studies in rats (Domingo et al., 1984;
Krasovskii and Fridlyand, 1971; Murdock, 1959; Holly, 1955, Stanley et al., 1947), supporting
the plausibility for the effects observed in humans. However, the effects in animals were
generally observed at higher doses than that used in the Davis and Fields (1958) human study. It
is not known whether cobalt exposure in humans at higher dose levels would increase
erythrocytes sufficiently above normal physiological levels to significantly increase the risk of
cardiovascular effects. Therefore, polycythemia was not chosen as the critical effect for p-RfD
derivation.
Effects of cobalt on thyroidal iodine uptake were identified as an endpoint of concern in
humans, based on a preliminary report by Roche and Layrisse (1956). This report showed that
oral exposure to cobalt (1 mg cobalt/kg-day) for 2 weeks markedly inhibited radioactive iodine
uptake in the human thyroid. In a smaller human clinical study, reduced iodine uptake was
reported in 2 of 4 euthyroid patients exposed to 0.54 mg cobalt/kg-day by the oral route for up to
14 days (Paley et al., 1958). A confounding factor in this study is that one of the two subjects
reported to have reduced iodine uptake had received intravenous (i.v.) cobalt in addition to oral
cobalt intake. The i.v. loading dose regimen may have raised the internal concentration of cobalt
25

-------
8-25-2008
to a level greater than the estimated 0.54 mg cobalt/kg-day based on oral intake alone, rendering
the Paley et al. (1958) study inappropriate for consideration. Importantly, long-term cobalt
exposure (up to 7 months) at 2-4 mg/kg-day in anemic children has been reported to cause goiter
(Gross et al., 1954; Kriss et al., 1955; Little and Sunico, 1958). Therefore, while reduced iodine
uptake is reported in humans following short-term exposures at low doses (Roche and Layrisse,
1956; Paley et al., 1958), potentially more severe thyroid lesions may occur as a function of
increased duration or dose. Based on observations from rodent models of cobalt exposure, the
severity of thyroid toxicity appears to be related to duration of exposure. Indeed, necrosis and
inflammation of the thyroid has been reported in mice exposed to approximately 48 mg
cobalt/kg-day with an increase in severity over a period of 15-45 days (Shrivastava et al., 1996).
Subchronic provisional RfD
Although cobalt exposure induces decreased radioactive iodine uptake in the thyroid
(Roche and Layrisse, 1956), and polycythemia (Davis and Fields, 1958) in humans at similar
daily exposure levels (1 mg/kg-day and 0.97 mg/kg-day, respectively), thyroid toxicity is chosen
as the critical effect for derivation of provisional oral reference values. Cobalt-induced
polycythemia and decreased iodine uptake by the thyroid were reversible following relatively
short-term exposure in humans, however supporting studies indicate the potential for more
severe thyroid effects (e.g., Kriss et al., 1955). The point of departure (POD) of 1 mg
cobalt/kg-day for decreased iodine uptake in human thyroid is the LOAEL; dividing this POD by
a composite uncertainty (UF) of 300 yields a subchronic p-RfD of 3E-3 mg/kg-day as follows:
Subchronic p-RfD = LOAEL UF
= 1 mg/kg-d -^300
= 0.003 or 3E-3 mg/kg-day
The composite UF of 300 is composed of three uncertainty factors: An UF of 10 for LOAEL to
NOAEL extrapolation was applied because the POD is based on a LOAEL. An UF of 10 was
applied due to the lack of data regarding inter-individual human variability or information on
sensitive subpopulations. Specifically, because the critical study (Roche and Layrisse, 1956) for
oral cobalt was based on healthy (euthyroid) adults, an UF of 10 was applied to protect sensitive
human populations. The available database includes several short-term human studies, multiple
developmental studies in animals and animal studies investigating hematological, cardiac,
neurological, neurobehavioral, and thyroid endpoints. The lack of a multi-generation
reproductive toxicity study is of particular concern because the database includes several animal
studies indicating effects on sperm function and testicular degeneration which raises concerns
that cobalt exposure may affect reproductive capability. Therefore, an UF of 3 was applied to
account for lack of a multi-generation toxicity study.
Chronic provisional RfD
Using the same LOAEL of 1 mg/kg-day for decreased iodine uptake in humans, and an
additional UF of 10 for extrapolating from subchronic to chronic duration (composite UF of
3000), a chronic p-RfD of 3E-4 mg/kg-day is derived as follows:
26

-------
8-25-2008
p-RfD = LOAEL-UF
= 1 mg/kg-d ^ 3000
= 0.0003 or 3E-4 mg/kg-day
An UF of 10 for extrapolation from subchronic to chronic duration was applied because the
critical effect was chosen from a principal study of a relatively short duration (2 weeks) of oral
exposure in humans. The temporal relationship between cobalt-induced decreased radioactive
iodine uptake and more severe thyroid toxicity should be considered carefully. One postulated
temporal relationship is that chronic exposure may have no greater effect than that resulting from
short-term exposure, because if the precursor event of inhibition of iodine uptake does not occur,
then there may be no change in thyroid function in the short- or long-term. Prolonged cobalt
exposure could have less of an effect because of the compensatory response of the pituitary-
thyroid axis to iodine deficiency, via increasing iodine uptake. However, although plausible,
there are no data to suggest that this postulated temporal relationship exist for cobalt-induced
thyroid toxicity. Indeed, a limited number of clinical observations primarily in children exposed
to oral cobalt at doses of 2-4 mg/kg-day for up to 7 months suggest the potential for more severe
thyroid toxicity (e.g., Kriss et al., 1955). In addition, cobalt may not be readily eliminated from
the body; for example, the biological half-life of cobalt chloride in rats is 25 hours (Rosenberg,
1993). Therefore, an UF of 10 for extrapolation from subchronic to chronic duration was
applied.
Confidence in the principal study is low-to-medium. Roche and Layrisse (1956)
examined twelve subjects over a two-week exposure period. Since only a single dose level was
evaluated, a NOAEL for decreased iodine uptake was not identified. Other human and animal
studies support the plausibility of cobalt producing thyroid toxicity (Paley et al., 1958; Prescott et
al., 1992; Shirivistava et al., 1996). Confidence in the database is low-to-medium. Although
some studies (Gross et al., 1954; Kriss et al., 1955; Little and Sunico, 1957) of longer duration
reported increased severity of thyroid effects (e.g., goiter) in children exposed to cobalt at higher
doses (2-4 mg cobalt/kg-day), critical details of these studies are unavailable for assessment.
Therefore, a temporal relationship between prolonged oral cobalt exposure and increased
severity of thyroid effects in humans (or experimental animals) is not clear, based upon available
data. As such, a low confidence in the provisional subchronic and chronic RfDs results.
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC
INHALATION RfC VALUES FOR COBALT
The human and animal database indicates that respiratory effects are sensitive endpoints
of inhaled cobalt. Symptoms of respiratory tract irritation and altered pulmonary function have
been widely reported in workers exposed to cobalt-containing airborne media. Of the four
human epidemiology studies discussed above, the study by Nemery et al. (1992) provides the
strongest basis for derivation of a provisional RfC (p-RfC). Workers in this study were exposed
to lower air concentrations of metallic cobalt dust than in the studies by Gennart and Lauwerys
(1990), Prescott et al. (1992) and Swennen et al. (1993). The values obtained from personal air
"3
samples from the Nemery et al. (1992) study, indicate a NOAEL of 5.3 (J,g/m and a LOAEL of
15.1 (J,g/m3. Furthermore, the Nemery et al. (1992) study demonstrated a dose-effect relationship
27

-------
8-25-2008
on lung function which correlated with urinary cobalt-levels, after adjusting for effects of
smoking and gender.
Animal data support the conclusion that the respiratory tract is the critical target for
inhaled cobalt (NTP, 1991; Bucher et al., 1990; Wehner et al., 1977). Subchronic and chronic
inhalation exposure to cobalt resulted in inflammation, fibrosis, and bronchiolar regeneration in
all regions of the respiratory tract in both rats and mice (NTP, 1991, 1998; Bucher et al., 1990,
1999) at doses higher than those identified in the Nemery et al. (1992) study. The NTP (1991)
study further demonstrated that cobalt can produce testicular effects in male mice following
inhalation exposure, but the effects were produced only at high dose levels. Oral studies have
also identified the testes as a target for cobalt toxicity. Multi-generation reproduction studies
following inhalation or oral exposure to cobalt are not available. Although developmental
toxicity studies following inhalation exposure to cobalt are not available, oral studies provide
evidence that high oral doses of cobalt may produce developmental effects in animals (Szakmary
et al., 2001; Paternain et al., 1988; Domingo et al., 1985).
Decreased pulmonary function and respiratory tract irritation were identified as the
co-critical effects for derivation of the subchronic and chronic p-RfCs. Assuming the personal
air samples to be more representative of worker exposure than the area air samples, the study by
Nemery et al. (1992) identified aNOAEL of 5.3 (J,g/m3 and a LOAEL of 15.1 (J,g/m3 for metallic
cobalt for effects on pulmonary function (e.g. forced expiratory volume (FEV), forced vital
capacity (FVC) and forced expiratory flow [referred to as MMEF]) and an increased prevalence
of symptoms of respiratory tract irritation (e.g. nose/throat irritation, cough, phlegm, dyspnea).
Although the LOAEL may be biased low due to inclusion of data from workshop #9, this does
not affect the p-RfC derivation. A NOAEL/LOAEL approach is taken for the derivation of
inhalation RfC values because the critical effect data are not amenable to benchmark dose
modeling. For example, workers in the low cobalt exposure group experienced a slight but
non-statistically significant increase in ventilatory function compared to controls, whereas a
significant decrease in ventilatory function was observed in the high cobalt exposure group
compared to both the control and the low cobalt exposure groups. The NOAEL for occupational
exposure was adjusted to continuous exposure as follows:
5.3 (J,g/m3 (10 m3/day / 20 m3/day) (5 days / 7 days) =1.9 (J,g/m3
"3
Using the NOAELadj of 1.9 (J,g/m as the POD, the subchronic p-RfC and chronic p-RfC for
cobalt was derived as shown below.
Subchronic p-RfC
"3
Dividing the NOAELadj of 1.9 (J,g/m by a composite UF of 100 yields a subchronic
p-RfC of 2E-5 m«/m3 for metallic cobalt as follows:
Subchronic p-RfC = NOAELadj ^ UF
= 1.9 (J,g/m3 100
= 0.00002 or 2E-5 mg/m3
28

-------
8-25-2008
The composite UF of 100 is composed of two uncertainty factors: 10 for database insufficiencies
and 10 for inter-individual variability. Nemery et al. (1992) did not report exposure duration for
any worker in this study; an assumption is made that worker exposure was at least of subchronic
duration. A factor of 10 was applied to account for database insufficiencies due to the lack of
inhalation developmental toxicity studies and a multi-generation reproduction study. A factor of
10 was applied to account for human variability, including sensitive subgroups. Individuals with
underlying respiratory diseases (asthma, chronic obstructive pulmonary disease) may be more
sensitive to the respiratory effects of inhaled cobalt. This subchronic p-RfC may not be
protective for people with hypersensitivity to cobalt.
Chronic p-RfC
"3
Dividing the NOAELadj of 1.9 (J,g/m by a composite UF of 300 yields a chronic p-RfC
-2
of 6E-6 mg/m for metallic cobalt as follows:
Chronic p-RfC = NOAELADj - UF
= 1.9 (J,g/m3 300
= 0.000006 or 6E-6 mg/m3
The composite UF of 300 is composed of three uncertainty factors: 3 to account for extrapolating
from an assumed subchronic exposure duration to a chronic exposure duration, 10 for database
insufficiencies and 10 for human inter-individual variability. A factor of 3 is applied to account
for extrapolating from an assumed subchronic to chronic exposure duration. Since Nemery et al.
(1992) did not report duration for any worker in this study, it is possible that exposure duration
may have been subchronic or longer for some workers. A factor of 10 is applied to account for
database insufficiencies due to the lack of inhalation developmental toxicity studies and a
multi-generation reproduction study. A factor of 10 is applied to account for human variability,
including sensitive subgroups. Individuals with underlying respiratory diseases (asthma, chronic
obstructive pulmonary disease) may be more sensitive to the respiratory effects of inhaled cobalt.
This chronic p-RfC may not be protective for people with hypersensitivity to cobalt.
Confidence in the key study (Nemery et al., 1992) is low because this cross-sectional study:
•	looked at only respiratory endpoints;
•	included a control group that was studied more than 1 year after the exposed
population;
•	included a study group exposed to iron and diamond dust in addition to cobalt
(and possibly to asbestos in the past);
•	did not report duration of exposure; and
•	encountered a number of procedural difficulties during its course (e.g.,
construction of control group).
Confidence in the database is medium. The choice of the critical endpoint is well supported by
other studies in humans and animals. Subchronic exposure studies in rats and mice (NTP, 1991)
found histopathological changes in the upper respiratory tract. Other studies in animals support
these findings. Reproductive and developmental effects have not been adequately studied.
Furthermore, oral studies reported large doses were required to produce reproductive or
29

-------
8-25-2008
developmental effects. It would be difficult to get a large enough internal dose via inhalation to
produce these effects. For these reasons, there is medium-to-low confidence in the subchronic
and chronic p-RfCs.
PROVISIONAL CARCINOGENICITY ASSESSMENT FOR COBALT
Weight-of-Evidence Descriptor
Under the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), cobalt
sulfate (soluble) is described as "likely to be carcinogenic to humans by the inhalation route,"
based on both the limited evidence of carcinogenicity in humans and sufficient evidence of
carcinogenicity in animals as shown by a statistically significant increased incidence of
alveolar/bronchiolar tumors in both sexes of rats and mice, pheochromocytomas in female rats,
and hemangiosarcomas in male mice (Bucher et al., 1999). While available studies in humans
have suggested a possible association between exposure to cobalt and respiratory tumors in
cobalt workers (Tuchsen et al., 1996; Mur et al., 1987; Morgan et al., 1983), limitations within
these studies, including small numbers of subjects, inadequate exposure assessment and potential
exposure to other chemicals make them inadequate for assessing the carcinogenic potential of
cobalt. Studies for evaluation of the oral carcinogenic potential for cobalt were not located.
Mode-of-Action Discussion
The U.S. EPA (2005a) Guidelines for Carcinogen Risk Assessment defines mode of
action as "a sequence of key events and processes, starting with the interaction of an agent with a
cell, proceeding through operational and anatomical changes, and resulting in cancer formation."
Examples of possible modes of carcinogenic action, in general, include mutagenic, mitogenic,
anti-apoptotic (inhibition of programmed cell death), cytotoxic with reparative cell proliferation,
and immunologic suppression.
While the mode of action of cobalt-induced carcinogenicity has not been determined,
data suggests a number of potential biological events that might be involved including
non-mutagenic genotoxicity (e.g. clastogenicity). A recent review by Lison et al. (2001) of in
vitro and in vivo experiments in animal models indicates that two different mechanisms of
genotoxicity may contribute to the carcinogenic potential of cobalt compounds: DNA strand
breakage and inhibition of DNA repair. DNA strand breaks have been reported at non-cytotoxic
concentrations in human peripheral blood monocytes (De Boeck et al., 1998, 2003a).
Furthermore, oral exposure of mice to cobalt chloride resulted in significantly increased
percentages of both chromosomal breaks and chromosomal aberrations in bone marrow cells
(Palit et al., 1991a,b,c,d). Mechanistic studies suggest cobalt-induced oxidative stress may be
involved. Exposure to cobalt compounds increases indices of oxidative stress, including
diminished levels of reduced glutathione, increased levels of oxidized glutathione, increased
levels of oxygen radicals and increased free-radical-induced DNA damage (Kawanishi et al.,
1994; Lewis et al., 1991; Kadiiska et al., 1989; Zhang et al., 1998; Moorehouse et al., 1985). To
compound the potential DNA strand breaking effects of cobalt, it appears that cobalt may also
inhibit the repair of such genetic damage. A review by Hartwig and Schwerdtle (2002)
30

-------
8-25-2008
concluded that cobalt may specifically target zinc finger structures in DNA repair proteins,
interfering with base and nucleotide excision repair. Collectively, while data indicate that cobalt
induces DNA damage and repair inhibition, there is weak evidence to suggest direct or indirect
mutagenicity in bacterial or mammalian systems.
Potential for a Mutagenic Mode of Action
Key events
The precise mechanism of cobalt-induced carcinogenicity has not been fully determined.
There is evidence that cobalt is capable of eliciting genotoxic effects. While evaluations for
mutagenic effects in bacteria have generally yielded negative results, results in several
mammalian cell systems have suggested that cobalt is genotoxic in mammalian cells. Limited
data from in vivo animal studies show that cobalt induces genotoxic effects, including
chromosomal breaks, chromosomal aberrations and micronucleus formation. The most likely
mechanisms for the genotoxic effects of cobalt are DNA strand breakage and the inhibition of
DNA repair.
Strength, consistency, specificity of association
Although the carcinogenic potential of inhaled cobalt has been demonstrated in rats and
mice by increased incidence of alveolar/bronchiolar tumors (Bucher et al., 1999; NTP, 1998),
direct evidence demonstrating that cobalt can induce mutagenic changes in cells of the
respiratory tract is lacking. In vivo exposure to hard metal dust containing 6.3% cobalt, 84%
tungsten and 5.4% carbon induced DNA strand breaks in rat type II epithelial lung cells (De
Boeck et al., 2003c). Chromosome/genome mutations were observed within 12 hours of
exposure to a single intratracheal instillation of 16.6 mg hard metal dust/kg body weight. Since
the mutagenic potential of cobalt alone was not evaluated in this study, a causal relationship
between type II epithelial cell mutations and cobalt exposure could not be established. Potential
mutagenic changes in respiratory tract cells could also be mediated through activated oxygen
species released by inflammatory cells (e.g., macrophages, polymorpohnuclear neutrophils),
rather than directly by cobalt (Lison et al., 2001).
Dose-response concordance
A dose-response concordance has not been established between the development of
bronchoalveolar tumors and mutagenesis following inhalation exposure to cobalt.
Dose-response information on mutagenicity is available for acute oral and parenteral exposure to
cobalt in mice (Suzuki et al., 1993; Palit et al., 1991a,b,c,d). No carcinogenicity data are
available for the oral or parenteral routes upon which to base a dose-response concordance.
Furthermore, no data are available on the mutagenic potential of cobalt in respiratory tract cells
following in vitro or in vivo exposure.
31

-------
8-25-2008
Temporal relationships
In vivo studies in animals show that acute oral and parenteral exposure to cobalt produces
genotoxicity to bone marrow cells (Suzuki et al., 1993; Palit et al., 1991a,b,c,d). Due to the lack
of data on the mutagenic potential of cobalt in respiratory tract cells, the temporal relationship
between potential mutagenic mechanisms and the development of bronchoalveolar tumors
cannot be assessed. Development of lung tumors in animals exposed to cobalt occurred
following chronic exposure (NTP, 1998).
Biological plausibility and coherence
In vivo mutagenicity studies in mice show that oral and intraperitoneal exposure to single
doses of cobalt chloride induced mutagenic changes in bone marrow cells (Suzuki et al., 1993;
Palit et al., 1991a,b,c,d). Although it has been hypothesized that the bronchoalveolar tumors are
the result of genotoxicity (De Boeck et al., 2003b; Hartwig and Schwerdtle, 2002; Lison et al.,
2001), no direct evidence is available linking cobalt-induced mutagenesis to the development of
cancer. Carcinogenicity through an indirect mutagenic mode of action may be mediated by
activated inflammatory cells (macrophages, polymorpohnuclear neutrophils) (Lison et al., 2001).
Other Potential Mode(s) of Action: Cytotoxicity and Cellular Regeneration
Subchronic and chronic inhalation studies (Bucher et al., 1990, 1999; NTP, 1991, 1998)
in rodents provide some evidence that cobalt causes cell injury with subsequent reparative cell
proliferation, which may be involved in the development of bronchoalveolar tumors. Following
"3
inhalation exposure to cobalt sulfate hexahydrate aerosol at concentrations of 0.3 to 30 mg/m
(equivalent to 0.067 to 6.7 mg cobalt/m3) for 3 months, rats and mice developed several lesions
indicative of cell damage and proliferation throughout the entire respiratory tract, including nasal
epithelial degeneration and metaplasia, laryngeal inflammation and metaplasia, bronchiolar
epithelial regeneration and ectasia, alveolar hyperplasia and lung fibrosis (NTP, 1991; Bucher et
al., 1990). Squamous hyperplasia of the larynx was the most sensitive effect (LOAEL=0.067 mg
"3
cobalt/m ). The results of the 2-year carcinogenesis study (Bucher et al., 1999; NTP, 1998) in
rats and mice revealed a statistically significant increase in combined alveolar/bronchiolar
"3
adenomas and carcinomas in the 0.67 mg cobalt/m group, but not in the 0.067 and 0.22 mg
cobalt/m3 groups for male rats and mice. In female rats and mice, a statistically significant
increase in combined alveolar/bronchiolar adenomas and carcinomas was observed in the
3	3
0.22 and 0.67 mg cobalt/m groups, but not in the 0.067 mg cobalt/m group. In this same study,
granulomatous inflammation of the lung was observed at all exposure levels (0.067, 0.22 and
0.67 mg cobalt/m3) in rats. Other markers of cell damage and proliferation, including
"3
hyperplasia, metaplasia and fibrosis, were observed in the 0.22 and 0.67 mg cobalt/m exposure
groups. Compared to rats, mice appeared to be less sensitive to cobalt-induced cytotoxic
changes. Results of this study show that bronchoalveolar tumors develop at exposure levels that
also produce cell damage and reparative proliferation, although cell damage and repair are also
observed at lower exposure levels than tumorigenesis. These observations suggest the possibility
that cell injury in the respiratory tract may have preceded the development of cancers although
direct evidence for this assertion is lacking.
32

-------
8-25-2008
Although limited evidence of carcinogenicity in humans is available, results of several
epidemioloic studies suggest a possible association between exposure to cobalt and respiratory
tumors (Tuchsen et al., 1996; Mur et al., 1987; Morgan, 1983). Subchronic exposure studies in
cobalt workers show an association between cobalt exposure and diminished pulmonary function
(Nemery et al., 1992; Gennart and Lauwerys, 1990). Taken together, results of studies in rodents
and humans suggest that inhaled cobalt may produce a cytotoxic response in the respiratory tract
that may contribute to decreases in pulmonary function and the development of bronchoalveolar
tumors.
Sustained cell proliferation, in response to cytotoxicity, can be a significant risk factor for
cancer (Correa, 1996). Sustained cytotoxicity and regenerative cell proliferation may result in
the perpetuation of mutations (spontaneous or directly or indirectly induced by the chemical),
resulting in uncontrolled growth. It is also possible that continuous proliferation may increase
the probability that damaged DNA will not be repaired. No data on cobalt are available to
directly evaluate the relationship between cell damage and reparative proliferation and the
development of bronchoalveolar tumors.
Conclusions Regarding Cancer Mode of Action
Limited evidence supports genotoxicity and cytotoxicity followed by cellular
regeneration as potential modes of action for cobalt tumorigenicity. In vitro and in vivo studies
provide evidence that cobalt is capable of eliciting genotoxic effects in mammalian cells;
however, two key uncertainties remain:
(1) No direct evidence linking cobalt-induced mutagenesis to the development of cancer is
available and (2) the mutagenic potential of cobalt in respiratory cells has not been evaluated.
Results of the 3-month and 2-year inhalation studies in rats and mice (Bucher et al., 1990,
1999; NTP, 1991, 1998) are also consistent with the hypothesis that cobalt acts through a mode
of action involving cytotoxicity and cellular regeneration, based on the observations that these
effects occur following subchronic exposure and bronchoalveolar tumors develop at exposure
levels that produce cytotoxicity and reparative proliferation. These observations suggest the
possibility that cell injury in the respiratory tract may have preceded the development of cancers
although direct evidence for this assertion is lacking. No mode of action data are available to
explain the statistically significant increases in the incidences of pheochromocytomas and
hemangiosarcomas that were observed in female rats and male mice, respectively.
Because a mutagenic mode of action is plausible, but cannot be clearly established for
carcinogenicity of inhaled cobalt, it is recommended that an age-dependent adjustment factor not
be applied to the unit risk to account for possible age-dependence of carcinogenic potency as
described in U.S. EPA (2005b).
33

-------
8-25-2008
Quantitative Estimates of Carcinogenic Risk
Oral Exposure
Human or animal studies examining the carcinogenicity of cobalt following oral exposure
were not located. Therefore, derivation of an oral slope factor is precluded.
Inhalation Exposure
As available human inhalation studies were not sufficiently detailed, particularly with
regards to analysis of exposure, the NTP (1998; Bucher et al., 1999) 2-year carcinogenicity study
in rats and mice was chosen as the principal study for the derivation of an inhalation unit risk,
based on the dose-response relationship for statistically significant increased incidences of
alveolar/bronchiolar (A/B) neoplasms (adenoma and carcinoma). Although statistically
significant increases in the incidences of pheochromocytomas and hemangiosarcomas were
observed in female rats and male mice, respectively, these tumors were not considered for the
derivation of the inhalation unit risk because a higher and more consistent response across
species was observed for alveolar/bronchiolar tumors. The exposure concentrations in this study
were adjusted to continuous exposure as follows:
5 days I week 6 hours I day
Cone n = Concx —		-		 x			——
L J	1 days/ week 24 hours/ day
This adjustment resulted in duration-adjusted concentrations of 0, 0.012, 0.040 and 0.120 mg
3	3
cobalt/m , respectively, for exposure to cobalt sulfate hexahydrate at 0.0, 0.3, 1.0 and 3.0 mg/m
exposure levels. Using the RDDR computer program, as specified in the RfC guidelines
"3
(U.S. EPA, 1994b), human equivalent concentrations (HECs, in mg cobalt/m ) were calculated at
each exposure level for each species and sex using body weight default values (U.S. EPA,
1994b), assuming exposure to particulates (MMAD=1.5 [j,m, og=2.2) with effects occurring in
the thoracic region of the respiratory tract. Table 5 shows the resulting HECs.
Table 5. Human Equivalent Concentrations (mg Cobalt/m3) Corresponding to Exposure
Concentrations in the NTP (1998; Bucher et al., 1999) Chronic Cancer Bioassay
Study
Male Rat
Female Rat
Male Mouse
Female Mouse
RDDR
Multiplier
0.83
0.79
1.48
1.44
Control
0
0
0
0
Low
0.010
0.0095
0.018
0.017
Medium
0.033
0.032
0.059
0.058
High
0.10
0.095
0.18
0.17
34

-------
8-25-2008
All models for quantal data in the U.S. EPA Benchmark Dose (BMD) software
(version 1.3.2) were fit to incidence for tumors (combined A/B adenomas and carcinomas), in
rats and mice; males and females were modeled separately. All data sets modeled showed a
statistical trend for increased tumor incidence with increasing exposure concentration. In
accordance with the U.S. EPA (2000) BMD methodology, the default benchmark response
(BMR) of 10% increase in extra risk was used as the basis for the BMD, with the BMDL
represented by the 95% lower confidence limit on the BMD. Models were run using the default
restrictions on parameters built into the BMD software. Table 6 shows the exposure
concentration and incidence data that were modeled.
Table 6. Neoplasm Incidence Observed in the NTP (1998; Bucher et al., 1999) Chronic
Cancer Bioassay
Animal/Strain/Site
Incidence of Neoplasms
F-344 Rats (male)
Human Equivalent Concentration of Cobalt (mg/m3)
0
0.010
0.033
0.10
Lung: A/B adenoma or carcinoma
1/50
4/50
4/48
7/50
F-344 Rats (female)
Human Equivalent Concentration of Cobalt (mg/m3)
0
0.0095
0.032
0.095
Lung: A/B adenoma or carcinoma
0/50
3/49
15/50
15/50
B6C3F1 Mice (male)
Human Equivalent Concentration of Cobalt (mg/m3)
0
0.018
0.059
0.18
Lung: A/B adenoma or carcinoma
11/50
14/50
19/50
28/50
B6C3F1 Mice (female)
Human Equivalent Concentration of Cobalt (mg/m3)
0
0.017
0.058
0.17
Lung: A/B adenoma or carcinoma
4/50
7/50
13/50
18/50
Table 7 summarizes the BMD modeling results. BMDLs shown were derived from
acceptable model fits (p>0.5). As is shown in Table 7, BMDLs were similar across study groups
"3
(range: 0.011-0.035 mg/m ). Lung tumors in female rats were chosen as the endpoint for use as
a point of departure for derivation of the inhalation unit risk. The BMDL for this endpoint was
the lowest for all study groups (i.e., male and female rats and mice) and was based on a model
that showed a good fit to the data (p=0.84), as reflected in the proximity of the BMDL to the
BMD, after dropping the high exposure group. Dropping the high exposure group is
recommended according to U.S. EPA (2000) procedure when no models achieve adequate fit
using all exposure levels. Although this left only two exposure levels (in addition to the control),
these exposure levels are in the low-dose portion of the curve within the region of the
dose-response relationship in which response is increasing with exposure level (i.e., the region of
interest for deriving the point of departure) and bracket the derived BMD. Appendix A presents
the results from all model runs used to support this toxicity assessment.
35

-------
8-25-2008
Table 7. Summary of BMD Modeling Results for Cobalt Cancer Data
Tumor
Species
Sex
BMD
(mg/m3)
BMDL
(mg/m3)
Lung: A/B adenoma or carcinoma
rat
male
0.085
0.035
Lung: A/B adenoma or carcinoma
rat
female
0.014a
o.oir
Lung: A/B adenoma or carcinoma
mouse
male
0.026
0.015
Lung: A/B adenoma or carcinoma
mouse
female
0.038
0.023
a Based on control, low and middle exposure levels; high exposure level was dropped due to failure of models to
achieve adequate fit using all exposure levels.
In the absence of mode of action data to inform the low dose extrapolation for cobalt, an
inhalation cancer unit risk was calculated by linear extrapolation of the BMDL to zero exposure
3 1
level (U.S. EPA, 2005a). The provisional inhalation unit risk of 9 (mg/m )" for cobalt sulfate
(soluble) was calculated as follows:
Provisional Unit Risk = BMR / BMDL
0.1/0.011
= 9 (mg/m3)"1
Table 8 shows continuous life-time exposure concentrations that correspond with specified risk
levels (i.e., lxlO"4, lxlO"5, lxlO"6).
Table 8. Continuous Life-time Exposure Concentrations Corresponding to
Specified Cancer Risk
Exposure Concentration at lxlO"4 Risk
l.lxlO5 mg/m3
Exposure Concentration at lxlO"5 risk
l.lxlO6 mg/m3
Exposure Concentration at lxlO"6 Risk
l.lxlO7 mg/m3
REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). 2004. TLVs® and
BEIs®: Threshold Limit Values for Chemical Substances and Physical Agents, Biological
Exposure Indices. Cincinnati, OH.
Alexander, C.S. 1969. Cobalt and the heart. Ann. Int. Med. 70:411-413.
Alexander, C.S. 1972. Cobalt-beer cardiomyopathy: A clinical and pathologic study of twenty-
eight cases. Am. J. Med. 53:395-417.
36

-------
8-25-2008
Anderson, M.B., N.G. Pedigo, R.P. Katz and W.J. George. 1992. Histopathology of testes from
mice chronically treated with cobalt. Reprod. Toxicol. 6:41-50.
Anderson, M.B., K. Lepak, V. Farinas and W.J. George. 1993. Protective action of zinc against
cobalt-induced testicular damage in the mouse. Reprod. Toxicol. 7:49-54.
ATSDR (Agency for Toxic Substances and Disease Registry). 2004. Toxicological Profile for
Cobalt. U.S. Public Health Service. Atlanta, GA.
Auchincloss, J.H., J.L. Abraham, R. Gilbert et al. 1992. Health hazard of poorly regulated
exposure during manufacture of cemented tungsten carbides and cobalt. Br. J. Ind. Med.
49:832-836.
Barceloux, D.G. 1999. Cobalt. Clin. Toxicol. 37(2):201-216.
Bencko, V., V. Wagner, M. Wagnerova et al. 1983. Immuno-biochemical findings in groups of
individuals occupationally and non-occupationally exposed to emissions [sic] containing nickel
and cobalt. J. Hyg. Epidemiol. Microbiol. Immunol. 27:387-394.
Beyersmann, D. and A. Hartwig. 1992. The genetic toxicology of cobalt. Toxicol. Appl.
Pharmacol. 115:137-145.
Bourg, W.J., J.R. Nation and D.E. Clark. 1985. The effects of chronic cobalt exposure on
passive-avoidance performance in the adult rat. Bull. Psychonom. Soc. 23(6):527-530.
Bucher, J.R., M.R. Elwell, M.B. Thompson et al. 1990. Inhalation toxicity studies of cobalt
sulfate in F344/N rats and B6C3F1 mice. Fundam. Appl. Toxicol. 15:357-372.
Bucher, J.R., J.R. Hailey, J.R. Roycroft et al. 1999. Inhalation toxicity and carcinogenicity
studies of cobalt sulfate. Toxicol. Sci. 49:56-67.
Correa, P. 1996. Morphology and natural history of cancer precursors. In: Cancer
Epidemiology and Prevention. D. Schottenfield and J.F. Fraumeni, Eds. New York: Oxford
University Press.
Corrier, D.E., H.H. Mollenhauer, D.E. Clark et al. 1985. Testicular degeneration and necrosis
induced by dietary cobalt. Vet. Pathol. 22:610-616.
Cugell, D.W. 1992. The hard metal diseases. Clinics in Chest Medicine 13:269-279.
Dabeka, R.W. and A.D. McKenzie. 1995. Survey of lead, cadmium, fluoride, nickel, and cobalt
in food composites and estimation of dietary intakes of these elements by Canadians in 1986-
1988. J. AO AC Int. 78(4):897-909.
Davis, J.E. and J.P. Fields. 1958. Experimental production of polycythemia in humans by
administration of cobalt chloride. Proc. Soc. Exp. Biol. Med. 37:96-99.
37

-------
8-25-2008
Davison, A.G, P.L. Haslam, B. Corrin et al. 1983. Interstitial lung disease and asthma in hard-
metal workers: Bronchoalveolar lavage, ultrastructural and analytical findings and results of
bronchial provocation tests. Thorax. 38:119-128.
De Boeck, M., D. Lison and M. Kirsh Volders. 1998. Evaluation of the in vitro direct and
indirect genotoxic effects of cobalt compounds using the alkaline comet assay. Influence of
interdonor and interexperimental variability. Carcinogenesis. 19:2021-2129.
De Boeck, M., S. Lardao, J. Buchet et al. 2000. Absence of significant genotoxicity in
lymphocytes and urine from workers exposed to moderate levels of cobalt-containing dust: a
cross-sectional study. Environ. Mol. Mutagen. 36(2):151-60.
De Boeck, M., N. Lombaert, S. De Backer, R. Finsey, D. Lison and M. Kirsch-Volders. 2003a.
In vitro genetoxic effects of difference combinations of cobalt and metallic carbide particles.
Mutagenesis. 18(2): 177-186.
De Boeck, M., M. Kirsch-Volders and D. Lison. 2003b. Cobalt and antimony: genotoxicity and
carcinogenicity. Mutat. Res. 533:135-153.
De Boeck, M., P. Hoet, N. Lombaert, B. Nemery, M. Kirsch-Volders and D. Lison. 2003c. In
vivo genotoxicity of hard metal dust: induction of micronuclei in rat type II epithelial lung cells.
Carcinogenesis. 24(11): 1793-1800.
Demedts, M., B. Gheysens, J. Nagels et al. 1984. Cobalt lung in diamond polishers. Am. Rev.
Respir. Dis. 130: 30-135.
Domingo, J.L., J.M. Llobet and R. Bernat. 1984. A study of the effects of cobalt administered
orally to rats. Arch. Farmacol. Toxicol. 10:13-20.
Domingo, J.L., J.L. Paternain, J.M. Llobet et al. 1985. Effects of cobalt on postnatal
development and late gestation in rats upon oral administration. Rev. Esp. Fisiol. 41:293-298.
Duckham, J.M. and H.A. Lee. 1976. The treatment of refractory anemia of chronic renal failure
with cobalt chloride. Q. J. Med. 178:277-294.
Elbetieha, A., A.S. Al-Thani, R.K. Al-Thani, H. Darmani and W. Owais. 2008. Effects of
chronic exposure to cobalt chloride on the fertility and testes of mice. J. Appl. Biol. Sci. 2:1-6.
Gennart, J.P. and R. Lauwerys. 1990. Ventilatory function of workers exposed to cobalt and
diamond containing dust. Int. Arch. Occup. Environ. Health. 62:333-336.
Gilman, J.P.W. 1962. Metal carcinogenesis. II. A study on the carcinogenic activity of cobalt,
copper, iron, and nickel compounds. Cancer Res. 22:158-162.
Gilman, J.P.W. and G.M. Ruckerbauer. 1962. Metal carcinogenesis. I. Observations on the
carcinogenicity of a refinery dust, cobalt oxide, and colloidal thorium dioxide. Cancer Res.
22:152-157.
38

-------
8-25-2008
Gross, R.T., J.P. Kriss and T.H. Spaet. 1954. The haematopoietic and goitrogenic effects of
cobaltous chloride in patients with sickle cell anemia. Am. Med. Assoc. J. Dis. Child. 88:284-
290.
Haga, Y., N. Cline, N. Hatori et al. 1996. Impaired myocardial function following chronic
cobalt exposure in an isolated rat heart model. Trace Elem. Elect. 13:69-74.
Hartwig, A. and T. Schwerdtle. 2002. Interactions by carcinogenic metal compounds with DNA
repair processes: toxicological implications. Toxicol Lett. 127(l-3):47-54.
Heath, J.C. 1956. The production of malignant tumors by cobalt in the rat. Brit. J. Cancer.
10:668-673.
Hillman, R.D. 2001. Hematopoietic agents: growth factors, minerals, and vitamins. In:
Goodman and Gilman's The Pharmological Basis of Therapeutics, Tenth Edition. J.G.
Hardman, L.E. Limbird, and A.G. Gilman, Eds. McGraw-Hill Medical Publishing Division.
Holly, R.G. 1955. Studies on iron and cobalt metabolism. J. Am. Med. Assoc. 158:1349-1352.
IARC (International Agency for Research on Cancer). 1991. IARC Monographs on the
Evaluation of Carcinogenic Risks to Humans. Volume 52: Chlorinated Drinking-Water;
Chlorination By-Products; Some Other Halogenated Compounds; Cobalt and Cobalt
Compounds, p. 363-472.
ICRP (The International Commission on Radiological Protection). 1993. Age-dependent doses
to members of the public from intake of radionuclides: Part 2 ingestion dose coefficients. The
International Commission on Radiological Protection. ICRP publication 67. New York, NY:
Pergamon Press.
Johansson, A., B. Robertson and P. Camner. 1987. Nodular accumulation of type II cells and
inflammatory lesions caused by inhalation of low cobalt concentrations. Environ. Res. 43:227-
243.
Kadiiska, M.B., K.R. Maples and R.P. Mason. 1989. A comparison of cobalt(II) and iron(II)
hydroxyl and superoxide free radical formation. Arch. Biochem. Biophys. 275(1):98-111.
Kawanishi, S., S. Inoue and K. Yamamoto. 1994. Active oxygen species in DNA damage
induced by carcinogenic metal compounds. Environ. Health. Perspect. 102(Suppl. 3): 17-20.
Kerfoot, E.J., W.G. Fredrick and E. Domeier. 1975. Cobalt metal inhalation studies on
miniature swine. Am. Ind. Hyg. Assoc. J. 36:17-25.
Krasovskii, G.N. and S.A. Fridlyand. 1971. Experimental data for the validation of the
maximum permissible concentration of cobalt in water bodies. Hyg. Sanit. 36:277-279.
Kriss, J.P., W.H. Carnes and R.T. Gross. 1955. Hypothyroidism and thyroid hyperplasia in
patients treated with cobalt. JAMA 157:117-121.
39

-------
8-25-2008
Kusaka, Y., K. Yokoyama, Y. Sera et al. 1986a. Respiratory disease in hard metal workers: An
occupational hygiene study in a factory. Br. J. Ind. Med. 43:474-485.
Kusaka, Y., Y. Ishikawa, T. Shirakawa et al. 1986b. Effect of hard metal dust on ventilatory
function. Br. J. Ind. Med. 43:486-489.
Kyono, H., Y. Kusaka, K. Homma et al. 1992. Reversible lung lesions in rats due to short-term
exposure to ultrafine cobalt particles. Ind. Health. 30:103-118.
Lammintausta, K., O.P. Pitkanen, K. Kalimo et al. 1985. Interrelationship of nickel and cobalt
contact sensitization. ContactDerm. 13:148-152.
Lasfargues, G., D. Lison, P. Maldague and R. Lauwerys. 1992. Comparative study of the acute
lung toxicity of pure cobalt powder and cobalt-tungsten carbide mixture in rat. Toxicol. Appl.
Pharmacol. 112:41-50.
Lasfargues, G., P. Wild, J.J. Moulin et al. 1994. Lung cancer mortality in a French cohort of
hard-metal workers. Am. J. Ind. Med. 26:585-595.
Lasfargues, G., C. Lardot, M. Delos et al. 1995. The delayed lung responses to single and
repeated intratracheal administration of pure cobalt and hard metal powder in the rat. Environ.
Res. 69:108-121.
Lewis, C.P.L., M. Demedts and B. Nemery. 1991. Indices of oxidative stress in hamster lung
following exposure to cobalt(II) ions: In vivo and in vitro studies. Am. J. Resp. Cell Mol. Biol.
5:163-169.
Linna, A., P. Oksa, P. Palmroos, P. Roto, P. Laippala and J. Uitti. 2003. Respiratory health of
cobalt production workers. Am. J. Ind. Med. 44(2): 124-132.
Lison, D. 1996. Human toxicity of cobalt-containing dust and experimental studies of the
mechanism of interstitial lung disease (hard metal disease). Crit. Rev. Toxicol. 26(6):585-616.
Lison, D., P. Carbonnelle, L. Mollo et al. 1995. Physicochemical mechanism of the interaction
between cobalt metal and carbide particles to generate toxic activated oxygen species. Chem.
Res. Toxicol. 8:600-606.
Lison, D., R. Lauwerys, M. Demedts et al. 1996. Experimental research into the pathogenesis of
cobalt/hard metal lung disease. Eur. Respir. J. 9:1024-1028.
Lison, D., M. De Boeck, V. Verougstraete and M. Kirsch-Volders. 2001. Update on the
genotoxicity and carcinogenicity of cobalt compounds. Occup. Environ. Med. 10:619-625.
Little, J. A. and R. Sunico. 1958. Cobalt-induced goiter with cardiomegaly and congestive
failure. J. Pediatr. 52:284-288.
Meyer-Bisch, C., Q.T. Pham, J-M. Mur et al. 1989. Respiratory hazards in hard metal workers:
a cross sectional study. Br. J. Ind. Med. 46:302-309.
40

-------
8-25-2008
Mohiuddin, S.M., P.K. Taskar, M. Rheault et al. 1970. Experimental cobalt cardiomyopathy.
Am. Heart. J. 80:532-543.
Mollenhauer, H.H., D.E. Corrier, D.E. Clark et al. 1985. Effects of dietary cobalt on testicular
structure. Virchows Arch. B. Cell Pathol. Incl. Mol. Pathol. 49:241-248.
Moorehouse, C.P., B. Halliwell, M. Grootveld et al. 1985. Cobalt(II) ion as a promoter of
hydroxyl radical and possible 'crypto-hydroxyl' radical formation under physiological
conditions. Differential effects of hydroxyl radical scavengers. Biochim. Biophys. Acta.
843:261-268.
Morgan, L.G. 1983. A study into the health and mortality of men exposed to cobalt and oxides.
J. Soc. Occup. Med. 33:181-186.
Morin, Y., A. Tetu and G. Mercier. 1971. Cobalt cardiomyopathy: Clinical aspects. Br. Heart.
J. 33:175-178.
Moulin, J.J., P. Wild, S. Romazini et al. 1998. Lung cancer risk in hard-metal workers. Am. J.
Epidemiol. 148:241-248.
Mur, J.M., J.J. Moulin, M.P. Charruyer-Steinerra et al. 1987. A cohort mortality study among
cobalt and sodium workers in an electrochemical plant. Am. J. Ind. Med. 11:75-82.
Murdock, H.R. 1959. Studies on the pharmacology of cobalt chloride. J. Am. Pharm. Assoc.
Sci. Ed. 48:140-142.
Murthy, G.K., U. Rhea and J.T. Peeler. 1971. Levels of antimony, cadmium, chromium, cobalt,
manganese, and zinc in institutional total diets. Environ. Sci. Technol. 5:436-442.
Nation, J.R., A.E. Bourgeois, D.E. Clark et al. 1983. The effects of chronic cobalt exposure on
behavior and metallothionein levels in the adult rat. Neurobehav. Toxicol. Teratol. 5:9-15.
Nemery, B., P. Casier, D. Roosels et al. 1992. Survey of cobalt exposure and respiratory health
in diamond polishers. Am. Rev. Resp. Disease 145:610-616.
NIOSH (National Institute for Occupational Safety and Health). 2005. NIOSH Pocket Guide to
Chemical Hazards. Index by CASRN. Online, http://www.cdc.gov/niosh/npe/npedcas.html.
NRC (National Research Council). 1989. Recommended Daily Allowances. Tenth Edition.
Subcommittee on the Tenth Edition of the RDAs. Food and Nutrition Board.
NTP (National Toxicology Program). 1991. Toxicity Studies of Cobalt Sulfate Heptahydrate in
F344/N Rats and B6C3F1 Mice (Inhalation Studies). U.S. Dept. of Health and Human Services,
Public Health Service, National Institutes of Health. NTP TOX 5.
41

-------
8-25-2008
NTP (National Toxicology Program). 1998. Toxicology and Carcinogenicity Studies of Cobalt
Sulfate Heptahydrate (CAS No. 10026-24-1) in F344/N Rats and B6C3F1 Mice (Inhalation
studies). U.S. Dept. of Health and Human Services, Public Health Service, National Institutes of
Health. NTP Technical Report Series, No. 471.
NTP (National Toxicology Program). 2005. Management Status Report. Online. http://ri.tp-
server.niehs.nih.eov/cgi/iH Indexes/Res Stat/iH Res Stat Frames.html
OSHA (Occupational Safety and Health Administration). 2005. Table Z-l limits for air
contaminants. Occupational Safety and Health Standards. Washington, DC. Online.
http://www.osha.gov/pls/oshaweb/owadisp.show document?p table=STANDARDS&p id=999
2.
Paley, K.R., E.S. Sobel and R.S. Yallow. 1958. Effect of oral and intravenous cobaltous
chloride on thyroid function. J. Clin. Endocrinol. Metab. 18:850-859.
Palit, S., A.K. Ghosh, A. Sharma et al. 1991a. Modification of the clastogenic effects of cobalt
by calcium in bone marrow cells of mice in vivo. Cytologia. 56:373-377.
Palit, S., A. Sharma and G. Talukder. 1991b. Chromosomal aberrations induced by cobaltous
chloride in mice in vivo. Biol. Trace Elem. Res. 29:139-145.
Palit, S., A. Sharma and G. Talukder. 1991c. Cytotoxic effects of cobalt chloride on mouse
bone marrow cells in vivo. Cytobios 65:85-89.
Palit, S., A. Sharma and G. Talukder. 199Id. Protection by chlorophyllin against induction of
chromosomal aberrations by cobalt in bone marrow cells of mice in vivo. Fitoterapia 62(5):425-
428.
Paternain, J.L., J.L. Domingo and J. Corbella. 1988. Developmental toxicity of cobalt in the rat.
J. Toxicol. Environ. Health. 24:193-200.
Pedigo, N.G. and M.W. Vernon. 1993. Embryonic losses after 10-week administration of cobalt
to male mice. Reprod. Toxicol. 7:111-116.
Pedigo, N.G., W.J. George and M.B. Anderson. 1988. Effects of acute and chronic exposure to
cobalt on male reproduction in mice. Reprod. Toxicol. 2:45-53.
Pehrsson, S.K., N. Hatori, N. Clyne et al. 1991. The effect of chronic cobalt exposure on
cardiac function in rats. Trace Elem. Med. 8:195-198.
Pennington, J.A. and J.W. Jones. 1987. Molybdenum, nickel, cobalt, vanadium, and strontium
in diets. J. Am. Diet. Assoc. 87(12): 1644-1650.
Prescott, E., B. Netterstrom, J. Faber et al. 1992. Effect of occupational exposure to cobalt blue
dyes on the thyroid volume and function of female plate painters. Scand. J. Work Environ.
Health. 18:101-104.
42

-------
8-25-2008
Raffn, E., S. Mikkelsen, D.G. Altman et al. 1988. Health effects due to occupational exposure
to cobalt blue dye among plate painters in a porcelain factory in Denmark. Scand. J. Work
Environ. Health. 14:378-384.
131
Roche, M. and M. Layrisse. 1956. Effect of cobalt on thyroidal uptake of I .J. Clin.
Endocrinol. Metab. 16:831-833.
Rystedt, I. and T. Fisher. 1983. Relationship between nickel and cobalt sensitization in hard
metalworkers. Contact Derm. 9:195-200.
Seidenberg, J.M., D.G. Anderson and R.A. Becker. 1986. Validation of an in vivo
developmental toxicity screen in the mouse. Teratog. Carcinog. Mutagen. 6:361-374.
Shabaan, A. A., V. Marks, M.C. Lancaster and F.N. Dufeu. 1977. Fibrosarcomas induced by
cobalt chloride (C0CI2) in rats. Lab. Anim. 11:43-46.
Shirakawa, T., Y. Kusaka, N. Fujimura et al. 1988. The existence of specific antibodies to
cobalt in hard metal asthma. Clin. Allergy. 18:451-460.
Shirakawa, T., Y. Kusaka, N. Fujimura et al. 1989. Occupational asthma from cobalt sensitivity
in workers exposed to hard metal dust. Chest 95:29-37.
Shirakawa, T., Y. Kusaka, N. Fujimura et al. 1990. Hard metal asthma: cross immunological
and respiratory reactivity between cobalt and nickel? Thorax. 45 267-271.
Shrivastava, V.K., C.V. David, N. Khare et al. 1996. Cobalt chloride induced histopathological
changes in thyroid gland of female mice, Mus musculus (P.). Pollut. Res. 15(3):307-309.
Singh, P.P. and A.Y. Junnarkar. 1991. Behavioral and toxic profile of some essential trace
metal salts in mice and rats. Indian J. Pharmacol. 23(3): 153-159.
Smith, R.J. and Fisher, J.W. 1973. Effects of cobalt on the renal erythropoietic factor kidney
hydrolase activity in the rat. Blood 42(2):893-905.
Sprince, N.L., L.C. Oliver, E.A. Eisen et al. 1988. Cobalt exposure and lung disease in tungsten
carbide production: A cross-sectional study of current workers. Am. Rev. Respir. Dis.
138:1220-1226.
Stanley, A.J., H.C. Hopps and A.M. Shideler. 1947. Cobalt polycythemia. II. Relative effects
of oral and subcutaneous administration of cobaltous chloride. Proc. Soc. Exp. Biol. Med.
66:19-20.
Stoner, G.D., M.B. Shimkin, M.C. Troxell et al. 1976. Test for carcinogenicity of metallic
compounds by the pulmonary tumor response in strain A mice. Cancer. Res. 36:1744-1747.
Suzuki, Y., H. Shimizu, Y. Nagae et al. 1993. Micronucleus test and erythropoiesis: Effect of
cobalt on the induction of micronuclei by mutagens. Environ. Mol. Mutagen. 22:101-106.
43

-------
8-25-2008
Swennen, B., J-P. Buchet, D. Stanescu et al. 1993. Epidemiological survey of workers exposed
to cobalt oxides, cobalt salts, and cobalt metal. Br. J. Ind. Med. 50 835-842.
Szakmary E., G. Ungvary, A. Hudak, E. Tatrai, M. Naray and V. Morvai. 2001. Effects of
cobalt sulfate on prenatal development of mice, rats, and rabbits, and on early postnatal
development of rats. J. Toxicol. Environ. Health A. 9;62(5):367-86.
Taylor, A., V. Marks, A.A. Shabaan et al. 1977. Cobalt-induced lipaemia and erythropoiesis.
Dev. Toxicol. Environ. 1:105-108.
Tuchsen, F., M.V. Jensen, E. Villadsen and E. Lynge. 1996. Incidence of lung cancer among
cobalt-exposed women. Scand. J. Work. Environ. Health. 22(6):444-450.
U.S. EPA. 1986. Guidelines for Carcinogen Risk Assessment. Federal Register.
51(185):33992-34003.
U.S. EPA. 1987. Health Effects Assessment for Cobalt. Prepared by the Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH for
the Office of Solid Waste and Emergency Response, Washington, DC. ECAO-CIN-H120.
U.S. EPA. 1988. Recommendations for and documentation of biological values for use in risk
assessment. EPA/600/6-87/008. NTIS PB88-179874/AS.
U.S. EPA. 1991. Chemical Assessments and Related Activities. Office of Health and
Environmental Assessment, Washington, DC. April.
U.S. EPA. 1994a. Chemical Assessments and Related Activities. Office of Health and
Environmental Assessment, Washington, DC. December.
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, Washington, DC. October, 1994. EPA/600/8-90/066F.
U.S. EPA. 1997a. Health Effects Assessment Summary Tables. FY-1997 Update. Office of
Research and Development, Office of Emergency and Remedial Response, Washington, DC.
July, 1997. EPA/540/R-97/036. NTIS PB 97-921199.
U.S. EPA. 1997b. Exposure Factors Handbook. U.S. Environmental Protection Agency, Office
of Research and Development, National Center for Environmental Assessment, Washington, DC.
Online, http://www.epa.gov/ncea/pdfs/efh/front.pdf.
U.S. EPA. 2000. Benchmark Dose Technical Guidance Document [external review draft],
EPA/630/R-00/001. Online, http://www.epa.gov/iris/backgr-d.htm.
U.S. EPA. 2004. 2004 Edition of the Drinking Water Standards and Health Advisories. Office
of Water, Washington, DC. EPA/822/R-02/038. Online.
http://www.epa.gov/waterscience/drinking/standards/dwstandards.pdf.
44

-------
8-25-2008
U.S. EPA. 2005a. Guidelines for carcinogen risk assessment. Risk Assessment Forum,
Washington, DC; EPA/630/P-03/001F. Federal Register 70(66): 17765-17817. Online.
http://www.epa.gov/raf.
U.S. EPA. 2005b. Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens. EPA/630/P-03/003F. Online.
http: //cfpub. epa. gov/ncea/ cfm/recordi spl ay. cfm ? dei d= 116283.
U.S. EPA. 2007. Integrated Risk Information System (IRIS). Online. Office of Research and
Development. National Center for Environmental Assessment, Washington, DC.
www.epa.gov/iris.
Veien, N.K., T. Hattel, O. Justesen et al. 1987. Oral challenge with nickel and cobalt in patients
with positive patch tests to nickel and/or cobalt. Acta. Derm. Venereol. 67:321-325.
Wehner, A.P., R.H. Busch, R.J. Olson and D.K. Craig. 1977. Chronic inhalation of cobalt oxide
and cigarette smoke by hamsters. Am. Ind. Hyg. Assoc. J. 38:338-346.
Wehner, A.P., B.O. Stuart and C.L. Sanders. 1979. Inhalation studies with Syrian golden
hamsters. Prog. Exp. Tumor Res. 24:177-198.
WHO (World Health Organization). 2005. Online catalogs for the Environmental Health
Criteria series. Online, http://www.who.int/dsa/cat97/zehc.htm and
http://www.who.int/dsa/iustpub/add.htm.
Zhang, Q., Y. Kusaka, K. Sato et al. 1998. Differences in the extent of inflammation caused by
intratracheal exposure to three ultrafine metals: Role of free radicals. J. Toxicol. Environ.
Health, Part A. 53:423-438.
45

-------
8-25-2008
APPENDIX A: SUMMARY OF BMD MODELING OF TUMOR INCIDENCE DATA IN
MALE AND FEMALE RATS AND MICE (NTP, 1998; BUCHER ET AL., 1999)
Male rat - A/B adenoma or carcinoma:
All models show acceptable fit (p > 0.1)
Log-logistic model yielded best fit (highest p-value and lowest AIC)
Best estimate of BMDL = 0.035 mg/m3
Model
P
AIC
BMD
BMDL



mg/m3
mg/m3
gamma (power >1)
0.502
111.12
0.087
0.043
logistic
0.446
111.52
0.099
0.066
log logistic (slope >1)
0.510
111.07
0.085
0.035
2 degree polynomial (pos betas)
0.502
111.12
0.087
0.043
1 degree polynomial (pos betas)
0.502
111.12
0.087
0.043
probit
0.453
111.47
0.098
0.063
log probit (slope >1)
0.357
112.11
0.104
0.064
quantal linear
0.502
111.12
0.087
0.043
quantal quadratic
0.373
111.99
0.010
0.069
weibull (power >1)
0.502
111.12
0.087
1.043
Output from BMD vl.3.2 is shown below:
Logistic Model $Revision: 2.1 $ $Date: 2000/02/26 03:38:20 $
Input Data File: C:\PROJECTS\COBALT\BMDS\RAMALULOG.(D)
Gnuplot Plotting File: C:\PROJECTS\COBALT\BMDS\RAMALULOG.plt
Fri Sep 09 11:46:38 2005
BMDS MODEL RUN
The form of the probability function is
P[response] = background+(l-background)/[l+EXP(-intercept-slope*Log(dose))]
Dependent variable = INRM
Independent variable = ECRM
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
46

-------
8-25-2008
Parameter Convergence has been set to le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.02
intercept = 0.683504
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope
have been estimated at a boundary point or have been specified by the user
and do not appear in the correlation matrix )
background intercept
background	1 -0.63
intercept -0.63	1
Parameter Estimates
Variable	Estimate	Std. Err.
background	0.0398603	0.0231667
intercept	0.272287	0.592931
slope	1	NA
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
Model Log(likelihood)
Full model -52.8567
Fitted model -53.5353
Reduced model -55.5862
Deviance TestDF p-value
1.35715 2 0.5073
5.45902 3 0.1411
47

-------
8-25-2008
AIC:
111.071

Goodness
of Fit







Scaled

Dose
Est.Prob.
Expected
Observed
Size Residual
0.0000
0.0399
1.993
1
50
-0.7178
0.0100
0.0523
2.615
4
50
0.8797
0.0330
0.0797
3.827
4
48
0.09207
0.1000
0.1513
7.565
7
50
-0.2228
Chi-square =1.35 DF = 2 p-value = 0.5099
Benchmark Dose Computation
Specified effect
Risk Type
Confidence level
BMD
BMDL
= 0.1
= Extra risk
= 0.95
= 0.0846262
= 0.0394914
48

-------
8-25-2008
Log-Logistic Model with 0.95 Confidence Level
dose
09:11 09/102005
49

-------
8-25-2008
Female rat - A/B adenoma or carcinoma:
Most models showed poor fit (p < 0.05) with highest exposure level included (no increase in
incidence at the highest exposure level.
The log-logistic model showed the best fit (p=0.11, lowest AIC)
Model
P
AIC
BMD
BMDL



mg/m3
mg/m3
gamma (power >1)
0.025
155.21
0.018
0.043
logistic
0.000
167.02
0.045
0.036
log logistic (slope >1)
0.090
152.86
0.015
0.011
2 degree polynomial (pos




betas)
0.025
155.21
0.018
0.014
1 degree polynomial (pos




betas)
0.025
155.21
0.018
0.014
probit
0.000
166.25
0.042
0.033
log probit (slope >1)
0.000
166.51
0.032
0.023
quantal linear
0.025
155.21
0.018
0.014
quantal quadratic
0.000
170.16
0.052
0.040
weibull (power >1)
0.025
155.21
0.018
0.014
Omitting the data from the highest exposure level improved fit of all models (p > 0.1)
Log-probit model yielded best fit (highest p-value and lowest AIC)
Best estimate of BMDL=0.011 mg/m3
Model
P
AIC
BMD
BMDL



mg/m3
mg/m3
gamma (power >1)
1.000
87.66
0.013
0.0077
logistic
0.242
89.69
0.020
0.0164
log logistic (slope >1)
1.000
87.66
0.013
0.0071
2 degree polynomial (pos betas)
0.710
87.66
0.014
0.0077
1 degree polynomial (pos betas)
1.000
86.40
0.011
0.0073
probit
0.289
89.31
0.019
0.0152
log probit (slope >1)
0.843
85.98
0.014
0.0110
quantal linear
0.710
86.40
0.011
0.0073
quantal quadratic
0.535
86.70
0.017
0.0139
weibull (power >1)
1.000
87.66
0.014
0.0077
Output from BMD vl.3.2 (all data included) is shown below:
Logistic Model $Revision: 2.1 $ $Date: 2000/02/26 03:38:20 $
Input Data File: C:\PROJECTS\COBALT\BMDS\RAFELU\RAFELULOGLOG.(D)
Gnuplot Plotting File:
C:\\PROJECTS\COBALT\BMDS\RAFELU\RAFELULOGLOG.plt
50

-------
8-25-2008
Fri Sep 09 16:34:38 2005
BMDS MODEL RUN
The form of the probability function is
P[response] = background + (l-background)/[l+EXP(-intercept-slope*Log(dose))]
Dependent variable = INRF
Independent variable = ECRF
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
intercept = 1.93572
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background -slope
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
Parameter Estimates
51

-------
8-25-2008
Variable
background
intercept
slope
Estimate
0
1.98253
1
Std. Err.
NA
0.20995
NA
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
Model Log(likelihood) Deviance TestDF P-value
Full model -72.3723
6.1152 3 0.1061
34.0413 3 <0001
Fitted model -75.4299
Reduced model -89.3929
AIC:
152.86

Goodness
of Fit






Scaled


Dose
Est.Prob.
Expected
Observed
Size
Residual
0.0000
0.0000
0.000
0
50
0
0.0095
0.0645
3.162
3
49
-0.09415
0.0320
0.1885
9.427
15
50
2.015
0.0950
0.4082
20.411
15
50
-1.557
Chi-square = 6.49 DF = 3 p-value = 0.0900
Benchmark Dose Computation
Specified effect =0.1
Risk Type	= Extra risk
Confidence level =0.95
BMD
BMDL
= 0.0153022
= 0.0109172
52

-------
8-25-2008
Log-Logistic Model with 0.95 Confidence Level
dose
09:21 09/102005
Output from BMD vl.3.2 (highest exposure level excluded) is shown below:
Probit Model $Revision: 2.1 $ $Date: 2000/02/26 03:38:53 $
Input Data File:
C:\PROJECTS\COB ALT\BMDS\RAFELUSE\RAFELUSEPROLOG.(D)
Gnuplot Plotting File:
C:\PROJECTS\COBALT\BMDS\RAFELUSE\RAFELUSEPROLOG.plt
Fri Sep 09 16:41:28 2005
BMDS MODEL RUN
The form of the probability function is:
P[response] = Background
+ (1-Background) * CumNorm(Intercept+Slope*Log(Dose)),
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = INRF
53

-------
8-25-2008
Independent variable = ECRF
Slope parameter is restricted as slope > 1
Total number of observations = 4
Total number of records with missing values = 1
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 = 3.0285
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background -slope
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
Parameter Estimates
Variable Estimate
background 0
intercept	2.97347
Std. Err.
NA
0.157916
NA
slope	1
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
54

-------
8-25-2008
Model Log(likelihood)
Full model -41.8291
Fitted model -41.9887
Reduced model -54.9105
AIC: 85.9774
Deviance TestDF P-value
0.319256 2 0.8525
26.1628 2 <0001
Goodness of Fit
Scaled
Dose Est.Prob. Expected Observed Size Residual
0.0000
0.0000
0.000
0
50
0
0.0095
0.0462
2.263
3
49
0.5015
0.0320
0.3197
15.985
15
50
-0.2987
Chi-square= 0.34 DF = 2 p-value = 0.8434
Benchmark Dose Computation
Specified effect =0.1
Risk Type	= Extra risk
Confidence level =0.95
BMD	=0.0141927
BMDL	= 0.0109984
55

-------
8-25-2008
Probit Model with 0.95 Confidence Level
dose
09:26 09/102005
56

-------
8-25-2008
Male mouse - A/B adenoma or carcinoma:
All models show acceptable fit (p > 0.1)
Log-logistic model yielded best fit (highest p-value and lowest AIC)
Best estimate of BMDL = 0.015 mg/m3
Model
P
AIC
BMD
BMDL



mg/m3
mg/m3
gamma (power >1)
0.944
251.10
0.033
0.0215
logistic
0.759
251.54
0.048
0.0359
log logistic (slope >1)
0.999
250.99
0.026
0.0150
2 degree polynomial (pos betas)
0.944
251.10
0.033
0.0215
1 degree polynomial (pos betas)
0.944
251.10
0.033
0.0215
probit
0.775
251.50
0.046
0.0349
log probit (slope >1)
0.594
252.03
0.059
0.0397
quantal linear
0.944
251.10
0.033
0.0215
quantal quadratic
0.412
252.76
0.080
0.0633
weibull (power >1)
0.944
251.10
0.033
0.0215
Output from BMD vl.3.2 is shown below:
Logistic Model $Revision: 2.1 $ $Date: 2000/02/26 03:38:20 $
Input Data File: C:\ROJECTS\COBALT\BMDS\MOMALU\MOMALULOGLOG.(D)
Gnuplot Plotting File:
C:\PROJECTS\COBALT\BMDS\MOMALU\MOMALULOGLOG.plt
Fri Sep 09 16:57:38 2005
BMDS MODEL RUN
The form of the probability function is
P[response] = background+(l-background)/[l+EXP(-intercept-slope*Log(dose))]
Dependent variable = INMM
Independent variable = ECMM
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
57

-------
8-25-2008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.22
intercept = 1.47367
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
background intercept
background	1 -0.62
intercept -0.62	1
Parameter Estimates
Variable	Estimate	Std. Err.
background	0.22179	0.0478621
intercept	1.45848	0.375385
slope	1	NA
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
Model Log(likelihood) Deviance TestDF	p-value
Full model -123.493
Fitted model -123.494 0.00271986 2	0.9986
Reduced model -130.684 14.3818 3	0.002429
58

-------
8-25-2008
AIC: 250.988

Goodness
of Fit





Scaled

Dose
Est.Prob.
Expected
Observed
Size Residual
0.0000
0.2218
11.090
11
50 -0.03047
0.0180
0.2777
13.884
14
50 0.03649
0.0590
0.3793
18.963
19
50 0.0109
0.1800
0.5613
28.065
28
50 -0.0185
Chi-square = 0.00 DF = 2 p-value = 0.9986
Benchmark Dose Computation
Specified effect =0.1
Risk Type	= Extra risk
Confidence level =0.95
BMD	= 0.0258434
BMDL	= 0.0149697
59

-------
8-25-2008
Log-Logistic Model with 0.95 Confidence Level
dose
09:32 09/10 2005
60

-------
8-25-2008
Male mouse - A/B adenoma or carcinoma:
All models show acceptable fit (p > 0.1)
Log-logistic model yielded best fit (highest p-value and lowest AIC)
Best estimate of BMDL = 0.023 mg/m3
Model
P
AIC
BMD
BMDL



mg/m3
mg/m3
gamma (power >1)
0.571
196.12
0.0455
0.0296
logistic
0.273
197.61
0.0735
0.0562
log logistic (slope >1)
0.700
195.72
0.0384
0.0231
2 degree polynomial (pos betas)
0.571
196.12
0.0455
0.0296
1 degree polynomial (pos betas)
0.571
196.12
0.0455
0.0296
probit
0.300
197.42
0.0697
0.0528
log probit (slope >1)
0.167
198.57
0.0768
0.0524
quantal linear
0.571
196.12
0.0455
0.0296
quantal quadratic
0.117
199.26
0.0959
0.0739
weibull (power >1)
0.571
196.12
0.0455
0.0296
Output from BMD vl.3.2 is shown below:
Logistic Model $Revision: 2.1 $ $Date: 2000/02/26 03:38:20 $
Input Data File: C:\PROJECTS\COBALT\BMDS\MOFELU\MOFELULOGLOG.(D)
Gnuplot Plotting File:
C:\PROJECTS\COBALT\BMDS\MOFELU\MOFELULOGLOG.plt
Fri Sep 09 17:05:08 2005
BMDS MODEL RUN
The form of the probability function is
P[response] = background+(l-background)/[l+EXP(-intercept-slope*Log(dose))]
Dependent variable = INMF
Independent variable = ECMF
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
61

-------
8-25-2008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.08
intercept = 1.17812
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
background intercept
background	1 -0.6
intercept -0.6	1
Parameter Estimates
Variable	Estimate	Std. Err.
background	0.0920048	0.035283
intercept	1.06119	0.354864
slope	1	NA
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
Model Log(likelihood) Deviance TestDF	p-value
Full model -95.5104
Fitted model -95.8619 0.702985 2	0.7036
Reduced model -102.791 14.5619 3	0.002232
62

-------
8-25-2008
AIC:
195.724
Goodness of Fit
Scaled
Dose Est.Prob. Expected Observed Size Residual
0.0000
0.0920
4.600
4
50
-0.2937
0.0170
0.1345
6.726
7
50
0.1135
0.0580
0.2223
11.117
13
50
0.6403
0.1700
0.3911
19.556
18
50
-0.451
Chi-square= 0.71 DF = 2 p-value = 0.7003
Benchmark Dose Computation
Specified effect =0.1
Risk Type	= Extra risk
Confidence level =0.95
BMD
BMDL
= 0.0384492
= 0.0231
63

-------
8-25-2008
Log-Logistic Model with 0.95 Confidence Level
dose
09:36 09/102005
64

-------