Health Risk
Assessment/Characterization
i
of the
Drinking Water Disinfection
Byproducts Chlorine Dioxide
& Chlorite
This Document was Prepared by:
Toxicology Excellence for Risk Assessment
4303 Hamilton Avenue.
Cincinnati, OH 45223
Under the Direction of:
Health and Ecological Criteria Division
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency
Washington, DC 20460
Under Purchase Order No.
8W-0766-NTLX
October 15,1998
-------
FOREWORD
The purpose of this document is to provide scientific support and rationale for the hazard
identification and dose-response information pertaining to chronic oral exposure to chlorine dioxide
and chlorite. It is not intended to be a comprehensive treatise on the chemical or toxicology of
chlorine dioxide or chlorite. Matters considered in this risk characterization include knowledge gaps,
uncertainties, quality of data and scientific controversies. This characterization is presented in an
effort to make apparent the limitations of the assessment and to aid and guide the risk assessor in the
ensuing steps of the risk assessment process.
An earlier draft of this document underwent external peer review by three independent
experts and experts within EPA. The charge to external peer reviewers and their comments are
presented in the Appendix. Reviewers' comments were considered in preparing the final version of
this document.
Chlorine Dioxide and Chlorite 2 3/13/98
Risk Chracterization
TERA
-------
FOREWORD 2
OUTLINE 3
LIST OF TABLES 4
LIST OF FIGURES 4
1.0 Introduction 5
2.0 Hazard Assessment/ Characterization 6
2.1 Studies in Human 6
2.2 Studies in Animals 8
2,2.1Toxicokinetics 8
2.2.2Noncancer Effects 9
2.2.2.1 Systemic Toxicity 9
2.2.2.2Reproductive and Developmental Studies 14
2.2.3 Carcinogenicity 21
2.3 Other Studies 22
2.4 Synthesis and Evaluation of Major Noncancer Effects and
Mode of Action 22
2.4.1 Major Noncancer Effects 22
2.4.2Mode of Action 26
2.5 Hazard Assessment Issues 26
2.5.1 Strengths and Weaknesses of Evidence 26
2.5.2 Susceptible Populations 27
3.0 Dose-Response Assessment
(RID DerivationyCharacterization 28
3.1 Choice of Pricipal Study and Critical Effect 28
3.2 Oral Reference Dose Derivation 29
4.0 Risk Characterization Summary 31
5.0 References 32
Chlorine Dioxide and Chlorite 3 3/13/98
Risk Characterization
TERA
-------
LIST OF TABLES
Table 1. Summary of Subchronic and Chronic Toxicity Studies for Chlorine Dioxide 10
Table 2. Summary of Subchronic and Chronic Toxicity Studies for Chlorite 11
Table 3. Summary of Reproductive and Developmental Studies for Chlorine Dioxide 16
Table 4. Summary of Reproductive and Developmental Studies for Chlorite 17
LIST OF FIGURES
Figure 1. Chlorine Dioxide: Developmental Studies 23
Figure 2. Chlorite: Developmental Studies 24
Chlorine Dioxide and Chlorite 4 3/13/98
Risk Chracterization
TERA
-------
1.0 Introduction
Chlorine dioxide (C1O2 ; CASRN 10049-04-4) is used as a drinking water
disinfectant. Chlorine dioxide is a strong oxidizing agent, that under oxidant demand
conditions, is readily reduced to chlorite (C1C-2-; CASRN 7758-19-2), another strong
oxidizing agent. The Drinking Water Criteria Document on Chlorine Dioxide, Chlorite
and Chlorate (U.S. EPA, 1994) provided the health effects data that support establishing
derivation of a MCLG under Section 1412 (b)(3)(A) of the Safe Drinking Water Act, as
amended in 1986. This document updates and evaluates information from 1994 to 1997
relevant to the health risk assessment and characterization of the drinking water
disinfectant, chlorine dioxide, and the degradation byproduct, chlorite. In particular, a
two-generation reproduction study in rats conducted by the Chemical Manufacturers
Association (CMA, 1996) is considered. The primary emphasis of this document is to
integrate the noncancer oral toxicity data for key endpoints relevant to human health and
to provide a hazard assessment/characterization, dose-response
assessment/characterization, and risk characterization for chlorine dioxide and chlorite to
support the Stage 1 Disinfectant/Disinfection By-Products (DBF) Rule.
Chlorine dioxide and chlorite are characterized together in this report since studies
conducted with chlorite, the predominant degradation product of chlorine dioxide, are
likely relevant to characterizing the toxicity of chlorine dioxide. In addition, studies
conducted with chlorine dioxide may be relevant to characterizing the toxicity of chlorite.
Chlorine dioxide is fairly unstable and rapidly dissociates predominantly into chlorite
and chloride, and to a lesser extent, chlorate. There is a ready interconversion among
these species in water (before administration to animals) and in the gut (after ingestion)
(U.S. EPA, 1994). Therefore, what exists in water or the stomach is a mixture of these
chemical species (i.e., chlorine-dioxide, chlorite, chlorate) and possibly their reaction
products with the gastrointestinal contents. As a result, the toxicity data for one
compound are considered applicable for addressing toxicity data gaps for the other.
The key endpoints associated with chlorine dioxide and chlorite most relevant to
human health effects are reproductive and neurodevelopmental toxicity (U.S. EPA,
1994). New data for these endpoints are reviewed in light of the new U.S. EPA
guidelines for reproductive and developmental toxicity risk assessment (U.S. EPA, 1991,
1996a), and for neurotoxicity risk assessment (U.S. EPA, 1997a). An earlier draft of this
document underwent peer review by three independent external experts and experts
within EPA. Reviewers' comments (presented in the Appendix) were taken into
consideration during preparation of this document
Chlorine Dioxide and Chlorite 5 • 3/13/98
Risk Chracterizatron
TERA
-------
2.0 Hazard Assessment/Characterization
In general, short-term exposure to chlorine dioxide and chlorite induces adverse
hematological effects, including methemoglobinemia. These effects are described in
detail in U.S. EPA (1994). Available studies in humans do not provide much insight into
the health effects of long-term exposure to chlorine dioxide and chlorite. Ecological
epidemiology studies are fraught with methodological problems that limit interpretation
of the results. Animal studies consistently identify adverse effects on developmental and
neurodevelopmental outcome. Carcinogenicity data from animal studies are largely
inadequate to discern a carcinogenic hazard to humans.
2.1 Studies in Humans
Clinical reports of poisonings are not available for chlorine dioxide or chlorite
(U.S. EPA, 1994). There are also no adequate long-term epidemiological studies to
assess carcinogenicity of these chemicals (U.S. EPA, 1994; I997c).
Short-term studies in humans reviewed in U.S. EPA (1994) include those by
Lubbers et al. (1981,1982,1983) and Bianchine et al. (1981) wherein no alterations in
hematological or urine chemistry or in physical symptoms were found in human
volunteers administered up to 0.34 mg/kg chlorine dioxide in drinking water for one day
or administered 0.036 mg/kg-day of chlorine dioxide or chlorite in drinking water for 84
days. In an epidemiological study of a community using chlorite as a drinking water
disinfectant, adult exposures ranged 0.007-0.03 mg/kg-day for chlorine dioxide and 0.09-
0.020 for chlorite for 12 weeks, and no consistent alterations in hematological parameters
were reported (Michael et al., 1981). Tuthill et al. (1982) retrospectively compared
morbidity and mortality data for a community that had utilized chlorite as a disinfectant
with a neighboring community and found a greater postnatal weight loss in infants from
the exposed community and no increase in the proportion of premature births when the
age of the mother was controlled.
The study by Kanitz et al. (1996) is the only new epidemiologic study on chlorine
dioxide or chlorite since the drinking water criteria document (U.S. EPA, 1994).
Prompted by suggestions of disinfectant byproduct-induced adverse
reproductive/developmental outcome from laboratory research and the results of Tuthill et
al. (1982), Kanitz et al. (1996) conducted a cross-sectional study in Italy on the
association between somatic parameters in infants at birth and drinking water disinfection
with chlorine dioxide and/or sodium hypochlorite. This study, however, because of
methodological problems, does not contribute new information as to whether there is an
Chlorine Dioxide and Chlorite 6 3/13/98
Risk Chracterization
TERA
-------
association between adverse reproductive/developmental outcome in human populations
and exposure to chlorine dioxide and/or chlorite in the drinking water.
During 1988-1989, Kanitz et al. (1996) followed 548 births at Galliera Hospital,
Genoa, and 128 births at Chiavari Hospital, Chiavari, Italy. Data on infant birth weight,
body length, cranial circumference and neonatal jaundice, and maternal age, smoking,
alcohol consumption, education and pre-term delivery, were collected from hospital
records. Women in Genoa were exposed to filtered water disinfected with chlorine
dioxide, sodium hypochlorite or both; trihalomethane levels varied from 8 to 16 ppb in
chlorite-treated water and 1 to 3 ppb in chlorine dioxide disinfected water. Women
residing in Chiavari used water pumped from wells, without any disinfection treatment,
and served as the comparison group (controls). Odds ratios were determined for the
somatic parameters by comparison of groups exposed to chlorine dioxide, sodium
hypochlorite, or both, with controls, and adjusted for maternal education level, income,
age, smoking and sex of the child. Neonatal jaundice occurred more frequently (odds
ratio, OR = 1.7; 95% confidence interval, CI = 1.1-3.1) in infants whose mothers resided
in the area where surface water was disinfected with chlorine dioxide, when compared to
infants with mothers using nondisinfected well water. Infants born to mothers residing in
areas where surface water was disinfected had smaller cranial circumference (< 35 cm;
OR= 22; 95% CI= 1.4-3.9 for chlorine dioxide; OR= 3.5; 95% CI = 2.1-8.5 for sodium
hypochlorite vs. untreated well water;'OR = 2.4; 95% CI = 1.6-5.3 for both vs. untreated
well water). In addition, these infants had a smaller body length (< 49.5 cm; OR = 2.0;
95% CI = 1.2-3.3 for chlorine dioxide vs. untreated well water, OR = 2.3; 95% CI = 1.3-
4.2 for chlorite vs. treated'well water). Risks for low-birth-weight infants (< 2500 g)
were reported to be increased in mothers residing in areas using water disinfected with
chlorite and chlorine dioxide, but these associations were not statistically significant. For
pre-term delivery (< 37 weeks), small but not statistically significant increased risks were
found among mothers residing in the area using chlorine dioxide. The authors concluded
that infants of women who consumed drinking water treated with chlorine compounds
during pregnancy were at higher risk for neonatal jaundice, cranial circumference £.35
cm and body length < 49.5 cm.
The interpretability of the results of Kanitz et al. (1996) is limited by lack of •
consideration of exposure and potential confounding variables, such as the quantity of
water consumed during pregnancy, lack of quantitative exposure information, exposure
to other chemicals in the water, and nutritional habits, amount of smoking and age
distribution of the women. In addition, baseline values for the infant sex ratio and
percentage of low-weight births for the comparison group deviate from values presented
by the World Health Organization for Italy. For example, the sex ratio used in the study
for the comparison group was 86, but most recent data (1991) for Italy indicate a sex ratio
Chlorine Dioxide and Chlorite 7 3/13/98
Risk Chracterization
TERA
-------
value of 112. While the percentage of low-weight births in the control group for the
Kanitz et al. (1996) study was 0.8%, in Italy for 1992. the percentage of low-weight births
(< 2,500 grams) is 6.7%. The quality of the untreated well water is not known, i.e.,
whether it contained any chemical or biological contaminants. The atypical baseline data
raise concerns about the control population selected for this study and render any
comparison to them by the exposed group difficult to interpret, thereby precluding the
ability to draw conclusions (Selevan, 1997).
Short-term drinking water studies with chlorine dioxide or chlorite in human
volunteers and ecological studies in communities using these chemicals for disinfection
of water supplies suggest that additional research into potential adverse effects, including
hematologic, reproductive and developmental outcome in human populations is needeoX
The studies thus far do not form an adequate basis to conclude that exposure to chlorine
dioxide and/or chlorite in drinking water results in adverse reproductive/developmental
health effects in humans (U.S. EPA, 19971)). Methodological problems, such as control
for confounding variables, identification of appropriate control or comparison groups,
quantification of exposure, and simultaneous exposure to multiple chemicals, are among
the factors that limit the interpretation of the results.
2.2 Studies in Animals
2.2.1 Toxicokinerics
It is difficult to obtain data on the toxicokinetics of chlorine dioxide and chlorite
because of their chemical reactivity. However, a series of experiments in rats by Abdel-
Rahman and coworkers (Abdel-Rahman, 198S; Abdel-Rahman et al., 1980a,b; 1982;
1984a; Sun and Abdel-Rahman, 1983; reviewed in U.S. EPA, 1994) using radioactive
chlorine as a tracer suggest that following oral exposure, chlorine dioxide and'chlorite are
rapidly absorbed (approximately 35-40% absorbed), widely distributed in the body, and
are rapidly and predominantly .excreted in the urine in the form of chloride (C1-).
Chlorine dioxide appears to be altered after coming into contact with organic material in
the gastrointestinal tract and may also be reduced by saliva. It is not clear whether it is
the parent compounds-or degradation products that are actually absorbed. There is no
evidence that these chemicals bioaccumulate.
Chlorine Dioxide and Chlorite 8 3/13/98
Risk Chracterization
TERA
-------
2.2.2 Noncancer Effects
2.2.2.1 Systemic Toxicity
U.S. EPA (1994) discusses at length the effects observed in chlorine dioxide and
chlorite studies utilizing various exposure durations. Tables 1 and 2 summarize the
subchronic and chronic toxicity information presented in U.S. EPA (1994); the reader is
directed to the criteria document for detailed information about these studies. In general,
the systemic toxicity studies presented in the criteria document are of limited quality for
risk assessment.
Groups of Sprague-Dawley rats (10/sex) were administered chlorine dioxide in
drinking water for 90 days at concentrations of 0, 25,50,100 or 200 mg/L (Daniel et al.,
1990). These concentrations correspond to doses of 0,2,4,6 or 12 and 0,2, 5, 8 and 15
mg/kg-day C1O2 for males and females, respectively (conversion provided by authors).
A significant increase in the incidence of nasal lesions (goblet cell hyperplasia and
inflammation of nasal turbinates) was found in the exposed animals. However, the
authors postulated that these lesions were most likely due to inhalation of C1O2 vapors at
the drinking water sipper tube or from off-gassing of the vapors after drinking rather than
ingestion of the drinking water. Thus, the 2 mg/kg-day dose group could be described as
a LOAEL, but the toxicological significance of the findings was inconclusive since the
lesions were not reported in any other studies and may possibly be an artifact of
treatment.
Revis et al. (1986) evaluated the effects of chlorine dioxide and chlorite on male
pigeons. Pigeons were exposed to water concentrations of either 0,0.07, or 0.5
mg/kg-day chlorine dioxide or chlorite for three months. The birds were maintained on
either a low calcium or low calcium/high Itpid diet; no exposed animals were maintained
on a normal diet. Increased serum cholesterol and decreased thyroid hormone levels were
observed in some groups, but there was no dose-response effect. Statistically significant
increases in aortic plaque size were observed in the 0.07 mg/kg-day group fed the
low-calcium diet. Therefore, the 0.07 mg/kg-day dose could be described as a LOAEL;
however, the relevance of the findings to animals on a normal diet is unclear. In order to
answer this question of relevance, Penn et al. (1990) repeated the study using chlorite
doses of 0,0.26, and 2 mg CIO2 mg/kg-day, maintaining the pigeons on a low calcium
diet. There was no significant difference in any of the parameters measured that could be
attributed to the ingestion of chlorite. Therefore, the 2 mg/kg-day dose is considered a
NOAEL.
Chlorine Dioxide and Chlorite 9 3/13/98
Risk Chracterization
TERA
-------
Table 1. Summary of Subchronic and Chronic Toxicity Studies for Chlorine Dioxide
Reference
Daniel et al.
(1990)
Revis et al.
(1986)
Abdel-Rahma
netal.
(1980a); Court
and
Abdel-Rahma
n (1980);
Abdel-Rahma
netal.
(1984b)
Haag(1949)
Species,
Sex
Rat, M
F
Pigeon, M
Rat,M
Mice,M
Ral.M,F
Route,
Doses
(mg/kg-d)
water,
0,2,4,6,
12
0,2,5.8,
13
water,
0,0.07,
0.5
water,
0,0.1, 1,
10
water.
0,0.18,
1.8,19
water.
0,0.07,
0.13,0.7,
U, 13
Duration
90 days
3 months
up to 1
year
Zyetrs
NOAEL
(mg/kg-d)
—
^•w
—
MD
1J
LOAEL
(mg/kg-d)
2
2
0.07
ND
13 (PEL)
Endpoint
nasal goblet cell
hyperplasia.
inflammation of nasal
turbinates; concern that
results not related to
ingestion of chlorine
dioxide
same as males antk
reduced spleen weight
elevated cholesterol,
decreased thyroid
hormones; not
dose-dependent
increased aortic plaque
size
all treated groups were
fed either low calcium
or low calcium-high
lipid diet; relevance of
findings to animals with
normal diet not clear
hematological effects
• (including increased
catalas* activity,
increased resistance of
RBCtohemoIysis);
effects not dose-related;
unclear if effects are
either statistically or
biologically significant
decreased survival;
study limited by
insufficient numbers of
animals and inadequate
evaluation of sensitive
parameters
PEL- frank .effect level
ND= not determined
Chlorine Dioxide and Chlorite
Risk Chracterization
TERA
10
3/13/98
-------
Table 2. Summary ofSubchronic and Chronic Toxicity Studies for Chlorite
Reference
Revis et al.
(1986)
Pennetal.
(1990)
Couriand
Abdel-
Rahman
(1980);
Abdel-
Rahmanet
al. (1984b)
Haag (1949)
RJdgway
(1992);
Harrington
etal.
(1995a)
Species,
Sex
Pigeon,
M
Pigeon,
M
Rat,M
Rat,M,
F
Rat,M,
F
Route, Doses
-------
A series of studies was conducted to examine the effects of drinking water
exposure to chlorine dioxide and chlorite on hematological parameters (Abdel-Rahman
et al., 1980a: Couri and Abdel-Rahman (1980), Abdel-Rahman et al., 1984b). Male rats
were exposed for up to one year to chlorine dioxide doses of 0,0.1, 1, or 10 mg/kg-day,
or to chlorite doses of 0, 1 or 10 mg/kg-day. A number of hematological parameters were
measured including red blood cell counts, osmotic fragility, hematocrit, and levels of the
following: hemoglobin, glutathione reduetase, glutathione peroxidase, catalase,
glutathione and methemoglobtn. Some alterations in most parameters were observed.
For chlorine dioxide, the effects included increased catalase activity and increased
resistance of red blood cell (RBC) to hemolysis. For chlorite, the observed effects
include decreased glutathione (GSH) and decreased osmotic fragility. However, there
was no consistent relationship with dose and effects were not observed consistently
throughout the period of exposure. Also, it was unclear if any of the effects were
statistically or biologically significant. Therefore, this series of studies was insufficient
for determining a NOAEL or LOAEL.
Haag (1949) conducted the only chronic study that evaluated the effects of
chlorine dioxide or chlorite in drinking water. Chlorine dioxide at concentrations of 0,
0.5,1, 5,10 or 100 mg/L (0,0.07,0.13,0.7,1.3 or 13 mg/kg-day) or chlorite at
concentrations of 0,1,2,4,8,100 or 1000 mg/L (0,0.09,0.18,0.35,0.7,9.3 or 81
mg/kg-day) was administered to rats (7/sex/ group) for two years. For chlorine dioxide,
the survival of animals consuming.water containing 10 mg C1O2/L (1.3 mg/kg-day) or
below was not significantly affected; survival in the 100 mg/L group was significantly
decreased. Histopathologic studies were also performed on representative animals from
each dose group, but no correlation was observed between treatment and any pathological
finding. This study identified a NOAEL of 1.3 mg ClO2/kg-day and a FEL (based on
decreased survival) of 13 mg C102/kg-day. Animals exposed to chlorite concentrations of
100 or 1,000 mg/L (9.3 or 81 mg ClO2-/kg-day) exhibited treatment-related renal
pathology, characterized by distention of the glomerular capsule and appearance of a pale
pinkish staining material in the renal tubules. These effects were also observed in a group
of animals that had been administered sodium chloride at a concentration equimolar to
1,000 mg NaClO2/L. The author concluded that renal pathology was a nonspecific salt
effect, but this observation does not alter the observation that concentrations of 100 mg/L
or higher lead to adverse effects. Based on renal effects, this study identified a NOAEL
of 0.7 mg ClO2-/kg-day and a LOAEL of 9.3 mg ClO2-/kg-day. The study was limited
because there was an insufficient number of animals tested per group, pathology was
conducted on a small number of animals, and it did not provide adequate evaluations of
more sensitive parameters, which would have been more useful in the overall assessment
of chronic toxicity.
Chlorine Dioxide and Chlorite 12 3/13/98
Risk'Chracterization
TERA
-------
One subchronic toxicity study of sodium chlorite in rats, a 13-week gavage study
(Harrington et al.. 1995a), has been published since the drinking water criteria document
(U.S. EPA, 1994). This study was previously described in U.S. EPA (1994) as Ridgway
(1992), based on laboratory report of the study.
Harrington et al. (1995a), treated Crl:CD (SD) BR rats (15/sex/group) with
sodium chlorite via daily gavage for 13 weeks at dose levels of 0,10,25, or 80 mg/kg-day
(equivalent to 0, 7.4, 18.6, or 59.7 mg ClO2-/kg-day, respectively), followed by terminal
sacrifice. In the 59.7 mg/kg-day group, four animals died during treatment and both
sexes exhibited salivation, significantly decreased RBC counts, and decreased total serum
protein levels. The males receiving 59.7 mg/kg-day exhibited significantly decreased
hematocrit and hemoglobin levels and increased methemoglobin and neutrophil levels,
while in the females, methemoglobin levels were significantly decreased. Possible
reasons for the decrease in methemoglobin in females, which is unexpected considering
the known oxidative effects of sodium chlorite, was not discussed by the authors.
Morphological changes in erythrocytes in some animals of both sexes and significant
increases in relative adrenal and spleen weights in the males and absolute and relative
spleen and adrenal weight, and relative liver and kidney weights in the females were also
noted at 59.7 mg/kg-day. Body weight and food consumption were not affected by
treatment. Histopathologic alterations in the 59.7 mg/kg-day group included squamous
epithelial hyperplasia, hyperkeratosis, ulceration, chronic inflammation, and edema in the
stomach of seven males and eight females. At 18.6 mg/kg-day, the following alterations
were reported: occasional salivation in two males, hematological alterations (increased
methemoglobin levels and neutrophil count, decreased lymphocyte count) in males,
increases in absolute and relative spleen' and adrenal weights in the females, and
histologic alterations in the stomach of two males, similar to that seen in the high-dose
group. The increase in absolute splenic weight was attributed to morphological
alterations in erythrocytes but no explanation was provided for alterations in absolute
adrenal weight. The NOAEL in this study was determined to be 7.4 mg/kg-day, and the
LOAEL is 18.6 mg/kg-day for rats subchronicaUy treated with sodium chlorite.
In conclusion, no subchronic or chronic studies adequately assess die systemic toxicity of
chlorine dioxide. The study by Haag (1949) identifies a NOAEL of 1.3 and a LOAEL of 13 mg
CKVkg-day. However, die adequacy of this study is questionable. For chlorite, one adequate
subchronic study (Harrington cl al., 1995a) identified a NOAEL of 7.4 and a LOAEL ol 18.6 mg
ClOr/kg-day, respectively, based on increases in absolute and relative spleen and adrenal weights
and histopathologic changes in die gastric mucosa.
Chlorine Dioxide and Chlorite 13 3/13/98
Risk Chracterization
TERA
-------
2.2.2.2 Reproductive and Developmental Studies
Discussed at length in the criteria document (U.S. EPA, 1994), are numerous
reproductive and developmental studies for chlorine dioxide and chlorite. Rats exposed
to chlorine dioxide prior to mating and throughout gestation (Carlton et al.T 1991; Mobley
et al., 1990; Orme et al., 1985; Suh et al.f 1983; Taylor and Pfohl, 1985) produced pups
with alterations in brain development (decreased brain weight, decreased brain cell
number), thyroxine levels, and behavior (decreased exploratory behavior, locomotor
activity). Similar studies with chlorite found either no significant treatment-related
effects on rat development (Suh et al., 1983), or decreased thyroxine levels (Carlton and
Smith, 1985), decreased pup weight and growth (Moore and Calabrese, 1982), or
decreased exploratory behavior in pups (Mobley et al., 1990).
Studies looking at reproductive endpoints in males showed decreased testicular
DNA synthesis (Suh et al., 1984; Abdel-Rahman et al., 1984b) and abnormal sperm
morphology (Meier et al., 1985) for both chemicals, and decreased sperm motility with
chlorite (Carlton and Smith, 1985; Carlton et al., 1987).
Subsequent to the 1994 criteria document, the U.S. EPA re-evaluated the Mobley
et al. (1990) study in light of the weight of evidence of the newer studies. Two
reproductive/ developmental studies with sodium chlorite in drinking water have recently
been published. Harrington et al. (1995b) conducted a teratology study in female rabbits.
This study was previously described in U.S. EPA (1994) as Irvine (1990), based on the
laboratory report of the study. Chemical Manufacturers Association conducted a 2-
generation study in rats that also included a neurotoxicity assessment (CMA, 1996).
These reproductive and developmental studies for chlorine dioxide and chlorite are
summarized in Tables 3 and 4, respectively.
Mobley et al. (1990) exposed female rats for 9 weeks to drinking water
containing.0,20 or 40 ppm sodium chlorite (0,3 or 6 mg CK>2-/kg-day) beginning 10
days prior to breeding with untreated males and until the pups were sacrificed at 35-42
days postconception (PCD). Slight delays in exploratory behavior on PCD 36-37 were
observed in rat pups exposed to 3 mg/kg-day. Animals exposed to 6 mg/kg-day exhibited
a more consistent and significant depression in exploratory behavior occurring on PCD
36-39. Exploratory activity was comparable between the treated and control groups after
PCD 39. Reviewing the results of this study relative to the findings of the newer
developmental studies in the database, suggests the NOAEL for neurodevelopmental
behavior effects in rats for this study is approximately 3 mg/kg-day and the LOAEL is 6
mg/kg-day.
Chlorine Dioxide and Chlorite 14 3/13/98
Risk Chracterization
TERA
-------
Harrington et al. (I995b) treated groups of 16 female New Zealand white rabbles
with technical grade sodium chlorite (80.6% purity) via their drinking water at levels of 0.
200, 600 or 1200 ppm from days 7 to 20 of pregnancy (gestational days, GD 7-20),
followed by terminal sacrifice at day 28. Water concentrations were maintained at the
same levels throughout pregnancy and were not adjusted for changes in volume of water
consumed. Based on measured water consumption, the authors calculated a mean daily
intake of approximately 0,10,26 or 40 mg ClO2-/kg-day (corrected for purity and
adjusted by the weight of the salt). Clinical condition, maternal body weight, and food
and water consumption were measured daily. At necropsy, gravid uterine weights,
number of corpora lutea, number of implantation sites, and live and dead fetuses were
recorded. Live fetuses were weighed, examined for external abnormalities, sexed,
dissected and a gross visceral examination performed. Skeletal examinations were alsov\
performed. Abnormalities were categorized as minor or major, the latter were thought to
impair survival or fitness; commonly observed variations were also" recorded. The
authors did not state which malformations fell into each of these categorizes. There was
no mortality, although two rabbits (one from each the control and 26 mg/kg-day groups)
were sacrificed in extremis due to a clinical condition unrelated to treatment. A
significant decrease in water consumption during the treatment period was observed in
the 26 and 40 mg/kg-day groups, and a decrease during treatment days 16-20 of
pregnancy was observed in the 10 mg/kg-day group. The authors attributed the decreases
in consumption to lack of palatability of the drinking water solution, although no
supporting data were presented. Food consumption was decreased in the 26 and 40
mgyTcg-day groups during days 7-11 of pregnancy. Body weight gain of treated animals
was decreased on days 7-19, although by day 26 these groups showed no differences from
controls in body weight gain. The authors concluded there were no treatment-related
effects on pregnancy incidence, number of implantations, number of preimplantation
losses, fetal sex ratio, number of live fetuses or fetal visceral or structural abnormalities.
Data for specific malformations and variations were not shown, instead data were
presented as the number or mean percentage of fetuses with major or minor external and
visceral or skeletal abnormalities. The number and mean percentage of major external
and visceral, and skeletal abnormalities were increased in the 26 and 40 mg/kg-day
groups (external/visceral: 6.6 and 2.9%, respectively, vs. 1.5% in controls; skeletal: 5.4
and 0%, respectively, vs. 0% in controls). Mean fetal weights in the 26 and 40 mg/kg-day
groups were slightly decreased (<9% relative to controls). In the 26 and 40 mg/kg-day
groups, the incidence of minor skeletal abnormalities (13.9 and 14.2 for the 26 and 40
mg/kg-day groups, respectively, vs. 7.7 % in controls) and skeletal variants Delated to
incomplete fetal bone ossification (such as of the pubis and stemebrae) is higher than
controls. The authors state in their discussion that these alterations in fetal body weight.
Chlorine Dioxide and Chlorite 15 3/13/98
Risk Chracterization
TERA
-------
Tabte 3. Summary of Reproductive and Developmental Studies Tor Chlorine Dioxide
Reference Species, Route,
Sex Doses
(mg/kg-d)
Suhetal. Rat, F water,
(1983) 0,0.1,1,10
Orme et al. Rat, F water,
(1985) 0,1,3,14
Rat, M gavage,
0,14
Taylor & Rat, F water,
Pfohl 0,14
(1985)
Rat, M gavage,
0,14
Toth et aL Rat, gavage,
(1990) pups 0,14
Mobley et Rat, F water.
al. (1990) 0,14
Carlton et Rat, M/F gavage,
al.(1991) 0.2.5,5,10
Suh et al. Rat, M water,
(1984) 0,1,10
Abdel- Rat, M water,
Rahman et 0,1,10
al. (1984b)
Duration
NOAEL LOAEL Endpoint
(mg/kg-d) (mg/kg-d)
Meier et Mice,M
aL(1985) 0,3.2,8,16
PCD= post conception day
PND= post natal day
2.S mos, then 1
gestation
2 wks prior mating 3
through lactation
PND5-20
2 wks prior mating —
through lactation
FND5-20
PND 1-20
10 d prior mating, —
gestation, lactation to
PCD 35-42
M: 56 d prior mating 10
F: 2 wk prior mating
through lactation (73
d total)
3 wks 1
3 months
Sd 16
10
14
14
14
14
14
14
10
decreased numbers of
implants & live births
decreased exploratory &
locomotor activity,
decreased T3&T4 in
pups
delayed development,
decreased T3 & T4,
decreased exploratory &
locomotor activity
decreased brain weight &
cell number, decreased
exploratory behavior in
pups
decreased brain cell
number & running wheel
activity
decreased forebrain
weight & protein content
decreased litter weight &
exploratory activity
(PCD 36-39)
no effect on reproductive
parameters or thyroid
hormone levels in pups or
adults
decreased testicular
DNA synthesis; biological
•significance unclear
decreased testicular DNA
thymidine incorporation;
biological significance
unclear
no sperm abnormalities 1,
3,5 wks after dosing
-------
Table 4. Summary of Reproductive and Developmental Studies for Chlorite
Reference
Moore &
Calabtese
(1982)
Suh et aJ.
(1983)
Moblcyctai.
(1990)
Irvine (1990);
Harrington et
al.(l995b) '
CMA (1996)
Cariton &
Smith (1985);
Cariton et al.
(1987)
Suh et al.
(1984)
Abdel-
Rahman et al.
(1984b)
Meier etal.
(1985)
Species,
Sex
Mice,F
Rat,F
Rat,F
Rabbit, F
Rat,M/F
Rat,F
M
Rat,M
Rat.M
Mice.M
Route,
Doses
(CIO,-)
(mg/kg-d)
water.
0,22
water.
0.0.1,1
water,
0,3.6
water,
0,10,26.40
water,
0, 2.9-3.8,
5.6-7.9,20-
28.6
water,
0,0.075.
0.75, 7.5
water,
0,0.075,
0.75.7.5.27
water,
0,0.1,1 .
water,
0. 1, 10
Savage,
0.8.20,4a
Duration
mating through
weaning
2.5 mos prior to
breeding then
— . — ? —
gestation
10 d prior
mating,
gestation,
lactation to
PCD 35-42
gestation days
7-20
2* generation
study- .
prorating,
gestation,
lactation
14 d prior
mating through
lactation
72-76 d
3wks
3 months
5d
NOAEL
(mg/kg-d)
—
1
3
10
2.9
0.75
0.75
0.1
—
40
LOAEL
(mg/kg-d)
22
—
6
26
5.9
7.5
7.5
1
1
—
Endpoint
decreased pup weight & growth
rate
no skeletal & soft tissue fetal
anomalies
decreased exploratory activity
(PCD 36-39); increase free T4 in
pups
decreased fetal weight & delayed
ossification;
decreased body weight gain &
food & water consumption in
dams
decreased absolute brain weight '
Fl & F2; lowered auditory start
in Fl & F2; altered liver weights
inFO&FI
decreased T3 & T4 in pups
decreased sperm modify &
increased abnormal morphology
alterations in testicular ft liver
DNA synthesis; biological
significance unclear
decreased testicular DNA
thymidine incorporation ;
biological significance unclear
no sperm abnormalities 1, 3, 5
wks after dosing
ND= not determined; PCD- post conception day; PND- post natal day
Chlorine Dioxide and Chlorite
Risk Characterization
TERA
17
3/13/98
-------
and delayed ossification are an indication of embryonic growth retardation. The NOAEL
for this study is 10 mg ClO2-/kg-day and the LOAEL is 26 mg/kg-day based on
decreased fetal weight and delayed skeletal ossification, and decreases in food and water
consumption, and decreased body weight gain in the dams.
While this study employed sufficient numbers of animals and administered
chlorite by a route relevant to human exposure, uncertainties exist in the interpretation of
the results due to inadequate reporting of the number and types of specific abnormalities
and variations. There is additional uncertainty as to whether the decreases in food and
water consumption and body weight gain in the dams are due to unpalatability or a direct
toxic effect of the chlorite.
CMA (1996) conducted a two-generation study to better define the NOAEL and
LOAEL for reproductive, developmental neurotoxicity and hematological endpoints.
Thirty males and 30 female Sprague-Dawtey rats of the OFA(SD)IOPS-Caw strain (FO)
generation received drinking water containing 35, 70 or 300 ppm sodium chlorite
(concentrations of sodium chlorite in the drinking water were apparently adjusted to
compensate for the 81.4% purity of the test material) for 10 weeks and were then paired
(1M: IF) for mating. A similar group received purified water and served as controls.
Males were exposed throughout mating and then were sacrificed. Exposure for the
females continued through mating, pregnancy, lactation and until necropsy following
weaning of their litters. Sodium chlorite concentrations were adjusted downward during
lactation to offset increases in the volume of water consumed so that a constant intake
(mg/kg-day) could be maintained. Twenty-five males and females from each of the first
25 litters to be weaned in a treatment group were chosen to produce the Fl generation.
The Fl pups were continued on the same treatment regimen as their parents. At
approximately 14 weeks of age, they were mated to produce the F2a generation. Due to a
reduced number of litters in the 70 ppm Fl-F2a generation, the Fl animals were remated
following weaning of the F2a to produce the F2b generation. Pregnant Fl females were
allowed to litter and rear the F2a and F2b generations until weaning at postnatal day
(PND) 21. Based on water consumption measured by the authors, and adjusting for the
molecular weight of sodium in sodium chlorite, EPA-calculated doses for the FO animals
were 0, 3.0,5.6,20.0 and 0,3.8,7.5, and 28.6 mg ClO2-4cg-day for males and females,
respectively. For the Fl animals, doses were 0,2.9,5.9, and 22.7 mg ClO2-/kg-day for
the males and 0,3.8,7.9, and 28.6 mgClO2-/kg-day for the females. Numerous
parameters were measured or calculated, including the following: body weight, food
consumption, water consumption, estrus cycle in the FO and Fl, hematology and
triidothyroxine and thyroxine levels in the Fl (blood samples collected from one male
and one female from the first 20 Fl litters at age PND 25 and another group at 13 weeks),
Chlorine Dioxide and Chlorite 18 3/13/98
Risk Chracterization
TERA
-------
gestation duration, liner size, pup sex, pup body weight, pup developmental landmarks.
number alive/dead pups in the Fl and F2 generation, total caudal sperm number and
percent motile, and morphology by computer assisted sperm motility analysis in the FO
and Fl, and organ weight and histopathological examination of the brain, pituitary gland.
liver, adrenal, spleen, thymus, kidneys, and reproductive organs of all FO and Fl controls
and high-dose animals. An additional group of Fl pups was chosen for
neurohistopathology on PND 11 (examination of the brain and spinal cord) or PND 60
(sensory ganglia, dorsal and ventral nerve roots, and several peripheral nerves and
muscles). Another group of Fl rats were examined for neurotoxicological endpoints
(motor activity in a Figure 8 Activity System and neuropathology on PND 60, auditory
startle in the' SR-Screening System, learning and memory retention in a water E-maze).
A functional observational battery (FOB) was also conducted on the pups undergoing trie
auditory and learning assessments. This group was composed of 2 males and 2 females
from twenty litters and exposure was discontinued after weaning. A re-evaluation of the
auditory startle response was conducted in 20 males and 20 females in the F2a and F2b
generations.
There were reductions in water consumption, food consumption and in body
weight gain in both sexes in all generations at various times throughout the experiment
(e.g., during premating, pregnancy, gestation, post-weaning), primarily in the 70 and 300
ppm groups. The authors attributed these reductions to a lack of palatability of the
drinking water solution, but did not show data to support this contention. Significant
alterations related to treatment at 300 ppm include: reductions in absolute and relative
liver weight in FO females and Fl males and females, reduced pup survival (increase in
number of pups found dead and/or killed prematurely during lactation) and reduced body
weight at birth and throughout lactation in Fl and F2, lower thymus and spleen weight in
both generations, lowered incidence of pups exhibiting a normal righting reflex and with
eyes open on PND IS, alteration in clinical condition in F2 animals chosen for
neurotoxicity, decreases in absolute brain weight for Fl males and F2 females, delays in
sexual development in males (preputial separation) and females (vaginal opening) in Ft
and F2, and lower red blood cell parameters in Fl. It is possible that the reported
alterations in pup sexual maturation measures may be due to reduced pup body weight,
but a definitive conclusion cannot be drawn. In the 70 ppm groups, reduced absolute and
relative liver weight in FO females and Fl males and females, reduced absolute brain
weight in Fl and F2,- and minor changes in red blood cell parameters in F1. In addition,
a significant decrease in maximum response to an auditory startle stimulus, was noted in
the 70 and 300 ppm groups on PND 24, but not on PND 60. Analysis of the E-maze data
by EPA personnel indicate possible alterations in learning behavior in the 70 ppm group,
but the differences from the conclusions of the report could not be resolved. Minor
changes in red blood cell parameters were apparent in the Fl at 35 ppm.
Chlorine Dioxide and Chlonte 19 3/13/98
Risk Chracterization
TERA
-------
This is an adequate study since it was conducted with sufficient numbers of
animals of both sexes and examined numerous endpoints. The study is acceptable and
consistent with the U.S. EPA testing guidelines that were in effect at the time of the study
(U.S. EPA, 1992). However, there are several limitations with the study. Lack of pair-
watered and fed control animals confounds the results and precludes making definitive
conclusions as to whether the alterations in food and water consumption and body weight
are related to water payability or a direct toxic effect of the agent Developmental
landmarks (e.g., vaginal opening in F2a group) were not reported for all groups. Grip
strength and landing foot splay were not included in the functional observational battery
(FOB). Discontinuation of exposure for the animals undergoing neurotoxicity testing
minimizes the likelihood of finding a positive effect and precludes comparison of the data
with that of other rats with continued exposure. While the study employed an exposure
regimen consistent with testing guidelines and should potentially detect adverse effects on
the developing nervous system, discontinuation of exposure after weaning reduces the
opportunity to detect neurological effects from continuous or lifetime exposures similar to
those expected from lifetime drinking water exposure in humans.
Interpretation of the neurobehavioral tests are limited. The report lacks detailed
descriptions of experimental methods (e.g., size of the arena, length of observations) and
positive control data (including estimates of variability) for the FOB. Positive control
studies for the motor activity and E-maze studies used high doses of the validation
chemicals and were not adequate to show the sensitivity of the methods, and showed only
that effects of the chemicals af maximally toxic doses could be recognized. Variability hi
the startle response data was high. The high variability and problems in calibrating and
operating the automated startle apparatus (as presented in the report) would tend to
decrease the sensitivity of the test to detect a difference between control and treated
groups, since differences in startle amplitude would have to be larger to attain statistical
significance. In some cases, inappropriate statistical analysis were applied. For example,
repeated-measures techniques were apparently not used to account for the fact that the
rats were tested repeatedly, and it is not clear how non-parametric rank data were
analyzed or why a log transformation was applied to the learning data.
Minor, statistically significant changes in hematological data at the 3 5 and 70
ppm concentrations (generally 1-7%) appear to be within normal ranges based on
historical data and are, therefore, not considered clinically or biologically significant or
adverse. The NOAEL for these hematological effects is considered'to be 7Q ppm.
The NOAEL for this study is 35 ppm (2.9 mg ClO2-/kg-day) and the LOAEL is
70 ppm (5.9 mg ClO2-/kg-day) based on lowered auditory startle amplitude, decreased
absolute brain weight, and altered liver weights in two generations.
Chlorine Dioxide and Chlorite 20 3/13/98
Risk Chracterization
TERA
-------
2.2.3. Carcinogenicity
There have been no Jong term oral bioassays for the carcinogenicity of chlorine
dioxide. As reviewed by U.S. EPA (1994), short-term assays with oral administration of
concentrates from drinking water treated with chlorine dioxide did not induce lung
adenomas in Strain A mice, did not increase skin tumor frequency in an initiation-
promotion assay, nor increase gamma glutamyl transpeptidase-positive foci in rat liver
slices (Miller et al., 1986). Robinson et al. (1986) demonstrated induction of hyperplastic
responses in the epidermis of SENCAR mice following 10-minute immersion in >300 mg
C1O2/L solutions.
Long-term studies in rats and mice do not provide sufficient evidence to draw
conclusions as to the carcinogenic potential of chlorite in humans. No increase in tumor
incidence relative to concurrent controls was found in male or female F344 rats drinking
0,300 or 600 ppm (0,28 or 32 and 0,28 or 41 mg Cl(>2-/kg-day in males and females,
respectively) sodium chlorite for 85 weeks (Kurokawa et al., 1986). In another study,
when Kurokawa et al. (1986) administered 0,250 or 500 ppm sodium chlorite (0, 36 or
71 mg ClO2-/kg-day) in drinking water to B6C3F1 mice of both sexes for 85 weeks,
increased liver tumors (hyperplastic nodules or hepatocellular carcinomas) and lung
adenomas in males were found relative to concurrent controls, but not-historical controls.
Similarly, Yokose et al. (1987) did not show clear evidence of carcinogenicity in B6C3F1
mice administered 0,250 or 500 ppm of chlorite for 80 weeks. Application of 100 mg
ClO2-/kg twice weekly for 51 weeks to the skin of female SENCAR mice did not induce
skin tumors. However, application of chlorite to skin following a single dermal
application of DMBA induced skin tumors, primarily squamous cell carcinomas, in 6/20
treated vs. 0/20 controls (Kurokawa et al., 1984).
As reviewed in U.S. EPA (1994) concentrates of chlorine dioxide-treated
drinking water generally did not induce a positive mutagenic responses in bacterial
systems nor did chlorine dioxide or chlorite induce clastogenic activity in mammalian
assays (micronucleus test, sperm head morphology or bone marrow chromosomal
aberrations).
In accordance with the 1986 cancer guidelines, chlorine dioxide and chlorite were
categorized in Group D, Not classifiable as to human carcinogenicity (U.S. EPA, 1986).
In accordance whh the 1996 proposed cancer guidelines, the carcinogenicity of these
chemicals are considered Cannot be determined (U.S. EPA, 1996b).
Chlorine Dioxide and Chlorite 21 3/13/98
Risk Chracterization
TERA
-------
2.3 Other Studies
The criteria document presented numerous studies examining the role of chlorite
in producing oxidative damage to the erythrocyte and methemoglobin formation. The
study by French et al. (1995) is a comparative study among several methemoglobin-
forming compounds, but does not shed any new light on mechanistic information.
French et al. (1995) examined the in vitro methemoglobin-forming potentials of
chlorite in Dorset sheep erythrocytes, and compared it with four other direct-acting
methemoglobin-forming agents in the same system. Samples were incubated for 1 hour
with doses of 0.0005,0.005,0.01 or 0.02 mM chlorite. Chlorite was ranked as among the
least potent agents compared top-dinitrobenzene, o-dinitrobenzene, copper, and nitrate*:
2.4 Synthesis and Evaluation of Major Noncancer Effects and Mode of Action
No subchronic or chronic studies adequately assess the systemic toxicity of
chlorine dioxide. The study by Haag (1949) identifies a NOAEL of 1.3 and a LOAEL of
13 mg/kg-day for chlorine dioxide, but is not considered of adequate design. For chlorite,
one adequate subchronic study (Harrington et al., 1995a) identified a NOAEL of 7.4
mg/kg-day and a LOAEL of 18.6 mg/kg-day based on effects on spleen and adrenal
weights and histopathologic changes in the gastric mucosa. However, the NOAEL
identified by Harrington et al. (1995a) is higher than the LOAEL for
developmental/neurodevelopmental effects (6 mg ClO2-/kg-day) identified by CMA
(1996). Mechanisms or mode(s) of action by which chlorine dioxide and chlorite exert
toxic effects are largely not understood.
2.4.1 Major Noncancer Effects
Developmental and neurodevelopmental effects are the critical effects) for both
chlorine dioxide and chlorite. A consistent finding is alteration in pup brain weight and
neurobehavioral endpoints. These effects tend to occur at dosages greater than 10 mg/kg-
day for chlorine dioxide and 6 mg/kg-day for chlorite (See Figures 1 and 2).
NOAEL of 3 mg/kg-day (20 ppm in drinking water) and a LOAEL of 14 mg/kg-
day (100 ppm) for chlorine dioxide were identified in rats based on decreased thyroid
hormone levels, and decreased exploratory, locomotor and running wheel behavior in
pups after treatment of dams from 2 weeks prior to breeding* through gestation and
lactation (Orme et al., 1995). A LOAEL of 14 mg/kg-day (100 ppm in drinking water),
based on decreased brain weight, brain cell number and protein content, and decreased
exploratory behavior in pups, was identified in rats in several studies of chlorine dioxide
Chlorine Dioxide and Chlorite 22 3/13/98
Risk Chracterization
TERA
-------
Figure 1. Chlorine Dioxide: Developmental Studies
jf> J&* _jj?> «*?• jJ1
& ^r ^ ^
*-* ^ >*" >* >
-------
Figure 2. Chlorite: Developmental Studies
30 T
29
28
-------
administered on similar breeding, gestation and/or lactational regimens (Taylor and
Pfohl, 1985: Toth et al., 1990. Mobley et al., 1990). Collectively these studies formed a
weight of evidence basis for the previous derivation of the RfD (U.S. EPA, 1994).
Carlton et al. (1991) reported 10 mg/kg-day as a NOAEL for reproductive parameters and
thyroid hormone levels in pups in a study employing gavage treatment of rats of both
sexes prior to mating, and females throughout gestation and lactation. Suh et al. (1983),
however, reported a LOAEL at 10 mg/kg-^ay for decreases in live births and number of
implants in rats drinking water for approximately 75 days and then during gestation.
For chlorite, a LOAEL of 3 mg/kg-day for decreased exploratory activity from
Mobley et al. (1990), supported by similar behavioral and neurohistopathology changes
from the chlorine dioxide studies (Oime et at., 1985; Taylor and Pfohl, 1985; Toth et afcx
1990), was used as the basis of the previously-derived RfD (U.S. EPA, 1994). The 3
mg/kg-day dose in Mobley et al. (1990) showed small changes, whereas changes
observed at 6 mg/kg-day were more consistent with findings from several other studies.
Similarly, lowered auditory startle response, reduced absolute brain weight, and reduced
liver weight were observed at 5.9 mg ClO2-/kg-day (LOAEL), but not at 2.9 mg C1O2-
/kg-day (NOAEL), in rats in a 2-generation study (CMA, 1996). Adverse effects in
developmental studies in other species occur at higher doses. A LOAEL of 22 mg/kg-day
(100 ppm chlorite in drinking water) was identified in mice based on a 14% decrease in
pup weight after treatment of dams through gestation and lactation; a NOAEL was not
identified (Moore and Calabrese, 1982). Irvine (1990); Harrington et al., (1995b) showed
decreased fetal weight and delayed ossification in rabbit pups exposed to 26 mg/kg-day
(LOAEL) but not to 10 mg/kg-day (NOAEL).
Data for reproductive endpoints for both chemicals are not as consistent, but
suggest there may be alterations in male reproductive toxicity. These data, however, do
not clearly define a threshold or NOAEL/LOAEL boundary. At 1 .0 mg/kg-day ( 10 ppm
in drinking water), for both chlorine dioxide and chlorite, reduced incorporation of
radiolabeled thymidine into testicular DNA after a 3- month treatment was identified in
male rats (Abdel-Rahman et al., 1984b). Suh et al. (1984) showed decreased testicular
DNA synthesis in rats administered 1 mg/kg-day chlorine dioxide. These studies are not
particularly useful for risk assessment without concomitant functional endpoints of male
reproductive toxicity. While Carlton and Smith (1985) and Carlton et al. (1987)
identified decreased sperm motility and increased abnormal sperm morphology in male
rats exposed for 72-76 days to 7.5 mg ClO2-/kg-day (LOAEL), but not 0.75 mg/kg-day
(NOAEL), CMA (1996) showed no alterations in sperm indices (motility, concentration,
morphology) up to 22 mg/kg-day. Meier et al. (1985) showed no alterations in sperm
morphology in mice for either chemical up to 16 mg/kg-day by gavage for 5 days.
Chlorine Dioxide and Chlorite 25 3/13/98
Risk Chracterization
TERA
-------
2.4.2. Mode of Action
The drinking water criteria document (U.S. EPA, 1994) discussed possible modes
of action whereby C102 and C1O2- produce hematological and systemic effects. The
mechanism(s) are still incompletely understood.
Oxidative damage to the erythrocyte and production of methemoglobin are most
likely related to their properties as oxidants (U.S. EPA, 1994). Chlorite is thought to be
the intermediate species responsible in many of the hematological effects of chlorine
dioxide because of its more efficient production of methemoglobin, depletion of RBC
glutathione and alteration of erythrocyte fragility.
In a series of experiments, Bercz and coworkers (Bercz et al.. 1982, 1986;
Harrington et al., 1985) have suggested that chlorine dioxide increases binding of dietary
iodide to gastrointestinal tissue and contents, producing a functional iodide deficiency.
Bercz et al. (1982) found decreased levels of circulating thyroxine in monkeys drinking
water .containing > 9.5 C1C>2 mg/kg-day, but not 44 mg ClC>2-/kg-day, for 4-6 weeks. In
a follow-up study, Harrington et al. (1986) demonstrated increases in thyroid iodide
uptake and a rebound in thyroxine levels in monkeys 1 year after an 8 weeks exposure to
approximately 5 mg ClO2/kg-day in drinking water. Unlike monkeys, rats showed dose-
related declines in thyroxine levels and no alteration in thyroid iodide uptake following an
8-week exposure to 10 mg ClC>2/kg-day in drinking water.
Whether either, or both, of these mechanisms are operable in inducing
reproductive, developmental and neurodevelopmental effects is not known. One could
also speculate that hypothyroidism, induced by chlorine dioxide-alteration of iodide
uptake in the gastrointestinal tract, might contribute to alterations in maternal or neonatal
behavior. Alternate, as yet unknown, mechanisms are also plausible as little definitive
mechanistic data are available. Additional research is needed to understand how chlorine
dioxide and chlorite induce alterations in fetal/neonatal neurodevelopment and behavior.
2.5 Hazard Assessment Issues
2.5.1 Strengths and Weaknesses of Evidence
Epidemiologic information comes from community ecological studies and has
been fraught with methodological problems, such as the lack of characterization of
exposure to other agents in the drinking water and controlling for confounding factors.
These studies do little to confirm a possible association between exposure to chlorine
dioxide and chlorite and adverse reproductive or developmental outcome in humans.
Chlorine Dioxide and Chlorite 26 3/13/98
Risk Chracterizaticn
TERA
-------
The animal toxicity data base for chlorine dioxide and chlorite are fairly
comprehensive, comprised of subchronic and chronic studies, reproductive and
developmental studies, and toxicokinetic and mechanistic information.
Multiple animal studies have shown similar alterations in neurodevelopmental
endpoints, such as brain weight and behavioral measures. The majority of these studies
have used sufficient numbers of animals and employed routes of exposure (gavage and
drinking water) relevant to human exposure. The majority of the developmental studies
have utilized rats and have shown a fairly consistent definition of the NOAEL/LOAEL.
For chlorine dioxide, all the developmental studies were conducted in rats. For chlorite,
there is one developmental study in mice (Moore and Calabrese, 1982), one
developmental study in rabbits (Irvine, 1990; Harrington et al., 1995b), and numerous
developmental studies in rats.
Reproductive studies in male animals are not consistent in demonstrating
alterations in spermatogenic indices, i.e., abnormal morphology or motility; however,
reported effects seem to appear at doses higher than the adverse developmental effects.
Similarly, clinically or toxicologically significant alterations in hematological parameters
occur at higher doses.
The mode of action for induction of adverse neurodevelopmental effects is not
known. It is also not known whether the rat is an adequate model for toxicity of chlorine
dioxide and chlorite in humans. However, this species is widely used to characterize
reproductive and developmental effects in humans.
2.5.2 Susceptible Populations
No subpopulation is recognized as uniquely susceptible to chlorine dioxide and
chlorite exposure. Developmental effects were noted in animal studies following both in
utero and postnatal exposure, suggesting infants and children may be more likely than
adults to experience adverse effects following exposure to chlorine dioxide and chlorite,
although the reasons for this increased sensitivity are not fully understood. It is well-
recognized that neurological development continues after birth and that gastrointestinal
uptake of many nutrients and chemicals is greater in the neonate than the adult.
Chlorine Dioxide and Chlorite 27 3/13/98
Risk Chracterization
TERA
-------
3.0 Dose-Response Assessment (RfD Derivation)/Characterization
3.1. Choice of Principal Study and Critical Effect
The CMA (1996) two-generation reproduction study was selected as the critical
study for the development of an RfD for both chlorine dioxide and chlorite. Since
chlorine dioxide readily degrades to chlorite, the critical effect(s) and dose at which those
effects are seen are similar. Therefore, the toxicity information for chlorite is relevant to
deriving the RfD for chlorine dioxide. The CMA (1996) rat study was designed to
evaluate the effects of chlorite (sodium salt) on reproduction and pre- and post-natal
development when administered orally via drinking water for two successive generations.
Developmental neurotoxicity and hematological toxicity were also evaluated in this
study.
Sodium chlorite was administered at 0, 35, 70, and 300 ppm via drinking water to
male and female Sprague Dawley rats (FO generation) for ten weeks prior to mating.
Dosing continued during the mating period, pregnancy and lactation. Reproduction,
fertility, clinical signs, and histopathology was conducted on FO and Fl males and
females. Fl and F2 pups were evaluated for growth and development, clinical signs, and
histopathology. In addition, animals from each dose group in the Fl were allocated for
neurotoxicity assessment (e.g., neurohistopathology, motor activity, learning ability and
memory retention, functional observations, auditory startle response). Limited
neurotoxicological evaluations were conducted on F2a and F2b pups.
Alterations in multiple endpoints define the LOAEL-NOAEL boundary in the
CMA (1996) study. The effects observed included statistically significant decreases in
pup body weight, absolute brain weight, pup survival, liver weight, and lowered startle
amplitude at the 300 ppm dose. Statistically significant decreases in auditory startle
amplitude, liver weight and absolute brain weight occurred at 70 ppm. Although different
responses were found for auditory startle (as indicated by measures of amplitude, latency
and habitation), mis is not unexpected given that these measures examine difference
aspects of nervous system function, and thus can be differently affected. Transient
alterations in neurofunctional (or neurochemical) measures, such as in the auditory startle
response, can occuc without neuropathological changes, and are considered of neurotoxic
concern (U.S. EPA, 1997a). Some of effects observed at 70 and 300 ppm occurred in
both sexes and in more than one generation, and are considered toxicologically
significant, which is consistent with U.S. EPA guidelines for reproductive,
developmental, and neurotoxicity risk assessment (U.S. EPA, 1991,1996a, 1997a). The
NOAEL for this study is 35 ppm (2.9 mg ClO2-/kg-day) and the LOAEL is 70 ppm (5.9
Chlorine Dioxide and Chlorite 28 3113/98
Risk Chracterization
TERA
-------
mg ClCb'/kg-day) based on lowered auditory startle amplitude, decreased absolute brain
weight and decreased liver weight.
While the CMA (1996) study is adequate and conducted with sufficient numbers
of animals of both sexes at multiple dose levels showing a range of effects, and examined
numerous endpoints, there are several limitations. Lack of pair-watered and pair-fed
control animals confounds the results and precludes making definitive conclusions as to
whether the alterations in food and water consumption and body weight are related to
water palatability or a direct toxic effect of agent. Discontinuation of exposure for the
animals undergoing neurotoxicity testing limits the likelihood of finding a positive effect,
precludes comparison of the data with that of other rats with continued exposure, and
does not reflect the expected lifetime exposure by humans to these chemicals in drinking
water. In addition, a lack of detailed description of experimental methods and positive
control data (including estimates of variability), and in some cases, inappropriate
statistical analysis, limit interpretation of the neurobehavioral tests.
The principal study is supported by the developmental studies by Orme et al.
(1985), Taylor and Pfohl (1985), Tom et al. (1990) and Mobley et al. (1990) wherein rats
were administered chlorite or chlorine dioxide at similar dosages in drinking water also
showed alterations in exploratory and locomotor behavior and reduced brain weights
(NOAELs of 3 mg/kg-day; LOAELs of 6-14 mg/kg-day).
The BID is derived for chlorine dioxide and chlorite from the NOAEL from the
CMA study (1996) of 2.9 mg ClC>2-/kg-day. This dose was determined from the nominal
water concentration based on measured water consumption and adjusted for the molecular
weight of the salt, so that doses are expressed as the chlorite ion. (For example, males
administered 35 ppm, had intakes of sodium chlorite equivalent to 3.9 mg/kg-day.
Adjusting for the molecular weight of sodium chlorite [MW= 90.5] relative to the chlorite
ion [MW= 67.5 J, gives the NOAEL dose of 2.9 mg C1C>2- /kg-day).
3.2 Oral Reference Dose Derivation
An uncertainty factor of 100 is appropriate for developing the RfD for chlorite and
chlorine dioxide. This composite factor includes a factor of 10 to account for
uncertainties associated with inter-species (extrapolating from experimental animals to
humans) and a factor of 10 for differences in response to toxicity within the human
population (to protect sensitive subpopulations). Because the critical effect is a
developmental effect in a database that includes chronic studies, it is not necessary to use
an uncertainty factor to account for use of a less-than-lifetime study. A default modifying
factor of I is applied.
Chlorine Dioxide and Chlorite 29 3/13/98
Risk Chracterization
TERA
-------
In derivation of the RfD in U.S. EPA (1994), an uncertainty factor of 3, in addition
to the composite 100 for inter- and intra-species variability, was applied for chlorine
dioxide to account for a database deficiency due to the lack of a multigenerational
reproductive study. For the derivation of the RfD for chlorite an uncertainty factor of 10,
to account both for the same data base gap (lack of a multigeneration reproductive study)
and extrapolation from a LOAEL to a NOAEL, was applied in addition to the composite
100 for inter- and intra-species variability. These uncertainty factors are no longer
needed since the CMA study (1996) fills the database gap for a multigenerational study
and the chlorite RfD is now based on the NOAEL from CMA (1996).
The RfD is obtained by rounding the NOAEL of 2.9 mg/kg-day to 3 mg/kg-day,
then dividing by the product of the uncertainty factors and the modifying factor (100).
The resulting RfD is 0.03 mg/kg-day for chlorine dioxide as well as for chlorite.
The RfD of 0.03 mg/kg-day is considered to be protective of susceptible groups,
including children, given that the RfD is based on a NOAEL from a two-generation
reproductive rat study. A two-generation reproductive study evaluates the effects of
chemicals on the entire developmental and reproductive life of the organism.
Additionally, current methods for developing RfDs are designed to be protective for
sensitive populations. In the case of chlorine dioxide and chlorite, an additional factor of
10 was used to account for variability between the average human response and the
response of more sensitive individuals.
Chlorine Dioxide and Chlorite 30 3/13/98
Risk Chracterization
TERA
-------
4.0 Risk Characterization Summary
Confidence in the critical study, CMA (1996) is medium. Although the study
design and analytical approaches are consistent with U.S. EPA testing guidelines, some
limitations in the design and conduct of the study exist. Confidence in the database is
high as there are studies in multiple species, chronic duration studies in males and
females, reproductive/developmental toxicity studies and a multigenerational study. The
threshold for adverse effects is consistently defined amongst the animal studies.
Resulting confidence in the RfD is medium-to-high.
Remaining areas of scientific uncertainty include the mode of action of chlorine
dioxide and chlorite in producing adverse effects on multiple organ systems, including
reproductive, developmental, and hematological effects. Inherent in the uncertainty with
the mode of action is identification of the susceptible populations or subgroups, since
additional research in this area would aid in better quantifying the additional risk to these
groups. Well-designed and conducted epidemiological studies in communities with
drinking water disinfected with these chemicals would decrease uncertainty in the
utilization of animal models for determination of human health effects.
Chlorine Dioxide and Chlorite 31 3/13/98
Risk Chracterization
TERA
-------
5.0 References
Abdel-Rahman, M.S. 1985. Pharmacokinetics of chlorine obtained from chlorine
dioxide, chlorine, chloramine and chloride. In: Jolley, R.L., et al., eds. Water
chlorination: Environmental impact and health effects. Vol. 5. Chelsea, MI: Lewis
Publ. Inc., pp. 281-293.
Abdel-Rahman, M.S., D. Couri and RJ. Bull. 1980a. Kinetics of C102 and effects of
C102, C102- and C103 in drinking water on blood glutathione and hemolysis in rat and
chicken. J. Environ. Pathol. Toxicol. 3:431-449.
Abdel-Rahman, M.S., D. Couri and J.D. Jones. 1980b. Chlorine dioxide metabolism in
rat. J. Environ. Pathol. Toxicol. 3:421-430.
Abdel-Rahman, M.S., D. Couri and RJ. Bull. 1982. Metabolism and phannacokinetics
of alternate drinking water disinfectants. Environ. Health Persp. 46:19-23.
Abdel-Rahman, M.S., Couri, D. and R.J. Bull. 1984. Toxicity of chlorine dioxide in
drinking water. J. Am. Coll. Toxicol. 3:277-284.
Abdel-Rahman, M.S., D. Couri and RJ. Bull. 1984a. The kinetics of chlorite and
chlorate in the rat. J. Am. Coll. Toxicol. 3:261-267.
Abdel-Rahman, M.S., D. Couri and RJ. Bull. 1984b. Toxicity of chlorine dioxide in
drinking water. J. Am. Coll. Toxicol. 3:277-284.
Bercz, J.P., L. Jones, L. Garner, D. Murray, D.A. Ludwig and J. Boston. 1982.
Subchronic toxicity of chlorine dioxide and related compounds in drinking water in the
nonhuman primate. Environ. Health Persp. 46:47-55.
Bercz, J.P., L.L. Jones, R_M. Harrington, R. Bawa and L. Condie. 1986. Mechanistic
aspects of ingested chlorine dioxide on thyroid function: Impact of oxidants on iodide
metabolism. Environ. Health Persp. 69:249-255.
Bianchine, J.R., J.R. Lubbers, S. Chauhan and J. Miller: 1981. Study of chlorine dioxide
and its metabolites in man. Final report on EPA grant No. 805643. EPA-600/1-81-068.
NTIS PB82-109356.
Chlorine Dioxide and Chlorite 32 3/13/98
Risk Chracterization
TERA
-------
Carlton, B.D. and M.K. Smith. 1985. Reproductive effects of alternate disinfectants and
their by-products. In: Jolley, R.L., et al., Eds. Water Chlorination: Environmental Impact
and Health Effects, Vol. 5. Chelsea, MI: Lewis Pubi. Inc. pp. 295-305.
Carlton, B.D.. D.L. Habash, A.H. Basaran, et al. 1987. Sodium chlorite administration in
Long-Evans rats: Reproductive and endocrine effects. Environ. Res. 42:238-245.
Carlton, B.D., A.H. Basaran, L.E. Mezza, E.L. George and M.K. Smith. 1991.
Reproductive effects in Long-Evans rats exposed to chlorine dioxide. Environ. Res.
56(2): 170-177.
Couri, D. and M.S. Abdel-Rahman. 1980. Effect of chlorine dioxide and metabolites on
glutathione dependent system in rat, mouse and chicken blood. J. Environ. Pathol.
Toxicol. 3:451-460.
CMA. 1996. Chemical Manufacturers Association. Sodium Chlorite: Drinking Water
Rat Two-Generation Reproductive Toxicity Study. Quintiles Report Ref. CMA/17/96.
Daniel, F.B., L.W. Condie, M. Robinson, J.A. Stober, R.G. York, G.R. Olseh and S.R.
Wang. 1990. Comparative subchronk toxicity studies of three disinfectants. Journal
American Water Works Association. 82:61-69.
French, C.L., Yaun, S.-S., Baldwin, L.A., Leonard, D.A., Zhao, X.Q. and EJ. Calabrese.
1995. Potency ranking of methemoglobin-forming agents. J.Appl. Toxicol. 15:167-174.
Haag, H.B. 1949. The effect on rats of chronic administration of sodium chlorite and
chlorine dioxide in the drinking water. Report to the Mathieson Alkali Works from H.B.
Haag of the Medical College of Virginia. Feb. 7,1949.
Harrington, R.M., H.G. Shertzer and J.P. Bercz. 1985. Effects of CK>2 on the absorption
and distribution of dietary iodide in the rat Fund. Appl. Toxicol. 5:672-678.
Harrington, R.M., H.G. Shertzer and J.P. Bercz. 1986. Effects of chlorine dioxide on
thyroid function in the African green monkey and the rat J. Toxicol. Environ. Health.
19:235-242.
Harrington, R.M., R.R. Romano, D. Gates and P. Ridgway. 1995a. Subchronic Toxicity
of Sodium Chlorite in the Rat Journal of the American College of Toxicology. 14(1):
21-33.
Chlorine Dioxide arid Chlorite 33 3/13/98
Risk Chracterization
TERA
-------
Harrington, R.M.. R.R. Romano and L. Irvine. 1995b. Developmental Toxicity of
Sodium Chlorite in the Rabbit. Journal of the American College of Toxicology. 14(2):
109-118.
Irvine, L.F.H. 1990. Sodium chlorite: Rabbit teratology study (drinking water
administration). EPA MRID #41715701. CMD-/3/R. Sept. 21. An unpublished study
submitted to EPA.
Kanitz, S., Y. Franco, V. Patrone, M Caltabellotta, et al. 1996. Association Between
Drinking Water Disinfection and Somatic Parameters at Birth. Environ. Health
Perspectives, 104(5), 516-520.
Kurokawa, Y., N. Takamura, Y. Matsushima, T. Imazawa and Y. Hayashi. 1984.
Studies on the promoting and complete carcinogenic activities of some oxidizing
chemicals in skin carcinogenesis. Cancer Letters. 24:299-304.
Kurokawa Y, S. Takayama, Y. Konishi, Y. Hiasa, S. Asahina, M. Takahashi, A.
Maekawa and Y. Hayashi. 1986. Long-term in vivo carcinogenicity tests of potassium
bromate, sodium hypochlorite and sodium chlorite conducted in Japan. Environ. Health
Persp. 69:221-235.
Lubbers, J.R., S. Chauhan and J.R. Bianchine. 1981. Controlled clinical evaluations of
chlorine dioxide, chlorite and chlorate in man. Fundam. AppL. Toxicol. 1:334-338.
Lubbers, J.R., S. Chauhan and J.R. Bianchine. 1982. Controlled clinical evaluations of
chlorine dioxide, chlorite and chlorate in man. Environ. Health Persp. 46:57-62.
Lubbers, J.R., J.R. Bianchine and R.J. Bull. 1983. Safety of oral chlorine dioxide,
chlorite, and chlorate ingestion in man. In: Jolley, R.L., et al., eds. Water chlorination:
Environmental impact and health effects, vol. 4(2). Ann Arbor, MI: Ann Arbor Science
Publishers, pp. 1335-1341.
Meier, J.R., R.J. Bull, J.A. Stober and M.C. Cimino. 1985. Evaluation of chemicals used
for drinking water disinfection for production of chromosomal damage and sperm-head
abnormalities in mice. Environ. Mutagen. 7:201-211.
Michael, G.E., R.K. Miday, J.P. Bercz, R.G. Miller, D.G. Greathouse, D.F. Kraemer and
J.B.Lucas. 1981. Chlorine dioxide water disinfection: A prospective epidemiology
study. Arch. Environ. Health. 36:20-27.
Chlorine Dioxide and Chlorite 34 3/13/98
Risk Chracterization
TERA
-------
Miller. R.G., F.C. Kopfler. L.W. Condie, M.A. Pereira. J.R. Meier, H.P. Ringhand. M.
Robinson and B.C. Casto. 1986. Results of toxicological testing of Jefferson Parish pilot
plant samples. Environ. Health Persp. 69:129-139.
Mobley, S.A., Taylor, D.H.. Laurie, R.D. and RJ. Pfohl. 1990. Chlorine dioxide
depresses T3 uptake and delays development of locomotor activity in young rats. In:
Jolley, R.L., et al., Eds. Water Chlorinatipn: Chemistry, Environmental Impact and
Health Effects, Vol. 6. Chelsea, MI: Lewis Publ., Inc. pp. 347-358.
Moore, G.S. and E.J. Calabrese. 1982. Toxicological effects of chlorite in the mouse.
Environ. Health Perspect. 46:31-37.
Orme, J., Taylor, D.H., Laurie, R.D. and R.J. Bull. 1985. Effects of chlorine dioxide on
thyroid function in neonatal rats. J. Toxicol. Environ. Health 15:315-322.
Penn, A., M-Xu Lu and J.L. Parker. 1990. Ingestion of chlorinated water has no effect
upon indicators of cardiovascular disease in pigeons. Toxicol. 63:301-313.
Revis, N.W., P. McCauley, R. Bull and G. Holdsworth. 1986. Relationship of drinking
water disinfectant to plasma cholesterol and thyroid hormone levels in experimental
studies. Proc.Natl.Acad.Sci. 83:1485-1489.
Ridgway P. 1992. Sodium chlorite: 13 Week oral (gavage) toxicity study in the rat.
Report to the Chlorine Dioxide Panel of the Chemical Manufacturers Association,
Washington D.C. by Toxicol Laboratories, Ltd., Ledbury, England (Laboratory Project
I.D. CMA/13/R).
Robinson, M., RJ. Bull, M. Schamer and R.F. Long. 1986. Epidermal hyperplasia in the
mouse skin following treatment with alternate drinking water disinfectants. Environ.
Health Persp. 69:293-300.
Selevan, S. 1997. Comments on Italian study: "Association between drinking water
disinfection and somatic parameters" by Kanitz et al., EHP 104(5):516-520,1996.
Memorandum to JF. Wiltse. Washington, DC: U.S. Environmental Protection Agency.
May 7.
Suh, D.H. and M.S. Abdel-Rahman. 1983. Kinetics study of chloride in the rat J.
Toxicol. Environ. Health. 12:467-473.
Chlorine Dioxide and Chlorite 35 3/13/98
Risk Chracterization
TERA
-------
Suh, D.H., Abdel-Rahman, M.S. and R.J. Bull. 1983. Effect of chlorine dioxide and its
metabolites in drinking water on fetal development in rats. J. Appl. Toxicol. 3:75-79.
Suh, D.H., M.S. Abdel-Rahman and R.J. Bull. 1984. Biochemical interactions of
chlorine dioxide and its metabolites in rats. Arch. Environ. Contain. Toxicol. 13:163-
169.
Taylor, D.H. and R.J. Pfohl. 1985. Effects of chlorine dioxide on the neurobehavioral
development of rats. In: Jolley, R.L., et al., Eds. Water Chlorination: Environmental
Impact and Health Effects, Vol. 5. Chelsea, MI: Lewis Publ., Inc. pp. 355-364.
Toth, G.P., R.E. Long, T.S. Mills and M.K. Smith. 1990. Effects of chlorine dioxide of)
the developing rat brain. J. Toxicol. Environ. Health. 31:29-44.
Tuthill, R.W., R.A. Giusti, G.S. Moore and E.J. Calabrese. 1982. Health effects among
newborns after prenatal exposure to C102-disinfected drinking water. Environ. Health
Persp. 46:39-45.
U.S. EPA. 1986. Guidelines for Carcinogen Risk Assessment. Federal Register. 51
(185): 33992-34003.
U.S. EPA. 1991. U.S. Environmental Protection Agency. Guidelines for Developmental
Toxicity Risk Assessment. December 5,1991. Federal Register 56: 63798-63826.
U.S.EPA. 1992. U.S. Environmental Protection Agency. Guidelines for reproductive
testing. CFR 798.4700. July 1, 1992.
U.S. EPA. 1994. U.S. Environmental Protection Agency. Drinking Water Criteria
Document on Chlorine Dioxide, Chlorite and Chlorate. Office of Water. Washington.
DC.
U.S. EPA. 1996a. U.S. Environmental Protection'Agency. Guidelines for Reproductive
Toxicity Risk Assessment. October 31,1996. Federal Register 61 (212): 56274-56322.
U.S.EPA. 1996b. Proposes Guidelines for Carcinogen Risk Assessment. Federal
Register. 61(79): 17959-18011.
U.S.EPA. 1997a. U.S. Environmental Protection Agency. Draft Final Guidelines for
Neurotoxicity Risk Assessment External Review Draft July 18,1997. Risk
Assessment Forum. Washington, D.C.
Chlorine Dioxide and Chlorite 36 313/98
Risk Chracterization
TERA
-------
U.S. EPA. 1997b. U.S. Environmental Protection Agency. Panel report and
recommendations for conducting epidemiological research on possible reproductive and
developmental effects of exposure to disinfected drinking water. Research Triangle Park,
NC: U.S. EPA National Health and Environmental Effects Research.
U.S. EPA. 1997c. National Primary Drinking Water Regulations: Disinfectants and
Disinfection Byproducts: Notice of Data Availability; Proposed Rule. Federal Register.
62 (212): 59387-59484. November 3, 1997.
Yokose, Y., K. Uchida, D. Nakae, et al. 1987. Studies of carcinogenicity of sodium
chlorite in B6C3F1 mice. Environ. Health Persp. 76:205-10.
Chlorine Dioxide and Chlorite 37 3/13/98
Risk Chracterization
TERA
-------
APPENDIX: SUMMARY OF REVIEWERS' COMMENTS
Statement of Work
TITLE: Peer review of Health Risk Characterization Report on Chlorite
BACKGROUND;
The mission of the United States Environmental Protection Agency's (EPA) Office of
Water (OW) is to protect public health and the environment from adverse effects of contaminants
in media such as ambient water, drinking water, waste water, sewage sludge and sediments. This
procurement relates to the peer review of a health risk assessment on the disinfection by product,
chlorite. This risk assessment will be used in support of EPA's stage 1 disinfection by product
rule which is scheduled to be final in November 1998. The Safe Drinking Water Act
Amendments of 1996 emphasize that "the best peer review science" be used in carrying out
SDWA regulations.
PURPOSE!
A cancer risk assessment/characterization document has been recently prepared that cites
and updates EPA's 1994 assessment on chlorite. This 1998 document considers a new study by
the Chemical Manufacturers Association on the reproductive and neurodevelopmental effects of
sodium chlorite.
TASK DESCRIPTION:
This purchase will procure a peer review on the 1998 EPA chlorite risk assessment
report. EPA has attached the 1998 chlorite risk assessment document, consisting of
approximately 20-25 pages, to be reviewed (Attachment 1), as well as supporting materials;
EPA's 1994 Criteria* Document on chlorite (Attachment 2), and a summary of the CMA study
with the accompanying peer review comments (Attachment 3). The peer reviewer shall submit
written comments that are clear/transparent, and constructive. They shall comment on whether
the document clearly and adequately discusses:
- the weight of the evidence
- the critical effect
- the key lines of evidence
- uncertainties in the risk assessment
- scientific basis for the derivation of the reference dose
The peer reviewers shall indicate where they are in agreement with the report and where they
disagree. If they disagree with any part of the report or find a weakness in the report, they shall
recommend explicit guidance on revising the report They shall provide comments mat include
an overall general summary on the acceptability and adequacy of the risk assessment, and
specific comments as needed (comment one on page X, paragraph X, line X).
-------
PEER REVIEWERS:
Dr. Jay Gandy
Center for Toxicology and Environmental Health
4301 W. Markham St.
Mail Slot 767
Little Rock, AR 72205
Tel No: 501/686-5239
FAX No: 501/686-8970
Dr. Rochelle Tyl
Research Triangle Institute
Center for Life Sciences and Toxicology
3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
Tel. No: 919/541-5972
FAX No: 919/541-6499
Dr. Paul Foster
cnr
6 Davis Drive
Research Triangle Park, NC 27709
Tel 919/558-1274
Fax 919/558-1300
-------
Summary of External Reviewer Comments
February 1998 Draft Chlorine Dioxide/Chlorite Risk Characterization
September 1998
Patricia McGinnis
Reviewers: 1= Foster
2=Gandy
3=Tyl
Comments below are paraphrased from the reviewers' comments.
Editorial revisions were made in response to comments by the reviewers.
Introduction
1. Comment (1): A reference is needed for the statement that chlorine dioxide dissociates under
aqueous conditions and in the GI tract of animals; this is a key assumption.
Response: Reference (DWCD, 1994) added.
Assessment
2. Comment (1): It is well understood that these agents can cause methemoglobin, etc.
(hematological effects). The available human data on hematological parameters should be
included.
Response: Section 2.0 revised to clarify hematological effects data for short-term exposure.
Reference to the DWCD (U.S. EPA, 1994) was added.
3. Comment (3): The status of the referent "neighboring" community as not exposed to chlorite,
or any drinking water disinfectant, should be made explicit.
Response: comment refers to Kanttz et al. (1996) —In section 2. 1, clarified that the community
used water pumped from wells without any water disinfection treatment.
4. Comment (1, 3): The section (Section 2.2.2. 1) describing the nasal lesions and potential
confounding with direct exposure to drinking water could be expanded to explain why this
potential artifact of treatment was noted. A fuller description of the nasal lesion could convince
the reader whether inadvertent treatment to the nasal passages was likely to happen; this is an .
important point to place these lesions in perspective for risk assessment
Response: Comment refers to Daniel etal. (1990). Authors do not provide further detail of the
nasal lesions. The sentence was clarified to the extent possible based on the authors' discussion.
-------
5 Comment (1) In Table 1, additional comments should be made to the endpoints column to
clarify why the observations are considered adverse
Response: Table 1 summarizes systemic toxicity data and lists the NOAEL and LOAEL.
Endpoints listed are for the LOAEL (therefore, considered adverse), unless no LOAEL is stated.
In the case where there is no LOAEL, the endpoint examined is the NOAEL and the column
states "not dose-dependent" or "no effects on ..." to clarify the non-adversity of the endpoint.
6. Comment (1): Was water consumption presented in the Harrington study? A presentation of
water consumption data or a summary statement is important because there is a tremendous
physiological demand for fluids during pregnancy.
Response: This comment refers to the Harrington et al. (199Sb) rabbit teratology study. The
study description includes a summary statement about alterations in water consumption; data
were not shown by the authors and additional details cannot be provided.
7. Comment (1): Why is significant growth retardation in the 26 and 40 mg/kg/d groups (in
Harrington et al., 1995b) not considered adverse? An alteration of growth, even in the presence
of maternal toxicity, should be considered adverse according to the guidelines.
Response: Results are considered adverse. Study text was revised to better explain the results for
the fetuses and to show the LOAEL at 26 mg/kg/d for maternal and fetal endpoints.
8. Comment (3): It should be clarified as to which sexes and doses as well as when reductions in
water consumption, food consumption and body weight gain were observed On the CMA, 1997,
study) so that palatability and toxicity may be discerned.
Response: Clarification added in se- -nd CMA paragraph, section 2.2.2.2- various times in both
sexes in all generations. Too cumK: iome to add all that data and its not the point of risk
characterization document, particularly when these endpoints are not the critical effect.
9. Comment (3): It should be confirmed that the NOAEL is 35 ppm for sodium chlorite, since
the purity of the test material was presented as 81.4%. Did the study authors make an adjustment
to the concentration of the test material in the drinking water to compensate for the purity or was
the calculated intake (in mg/kg/day) corrected for purity? The former may have contributed to
the observed "palatability problems7* (especially the sodium component).
Response: Apparently the study authors adjusted the concentration of sodium chlorite to
compensate for the purity of the test material- clarification added to first CMA paragraph,
section 2.2.2.2.
10. Comment (1): The statement regarding the CMA (1997) study that discontinuation of
exposure for animals undergoing neurotoxicity testing minimizes the likelihood of finding a
positive effect and precludes comparison is perplexing: The guidelines indicate that exposure
should be from gestation day 6 until postnatal day (PND) 10 with examination postexposure to
PND 60, and that offspring from a multigeneration study is acceptable. The study meets the
-------
requirements and can be appropriately compared with others; positive control data were
provided
Response. Additional text added to discussion of limitations of the CMA study to clarify this
statement (third paragraph on CMA study, section 2.2.2.2). While positive control data provided,
still some limitations (see next paragraph).
11. Comment (1): The statement (Section 2.2.3) that carcinogenicity data from animal studies
are largely inadequate should be qualified, as the mouse study in the second paragraph does not
seem to have been taken into consideration.
Response: The statement in the first paragraph referred to chlorine dioxide; there are no long
term bioassays for chlorine dioxide. The mice in the 85-week study, described in the second
paragraph, were administered sodium chlorite.. Introductory sentence added at beginning of
second paragraph referring to chlorite to clarify for reader.
12. Comment (3): The difference in sperm morphology outcomes between the two studies may
be due to differences in study duration as well as species.
Response: refers to last paragraph, section 2.4.1, Carlton and Smith (1985), Carlton et al. (1987)
and Meier et al. (1985). Sentences reworked- CMA male rat results added and compared with
Carlton studies. Meier mouse study sentence revised.
13. Comment (1): The statement about the same route and dosage for the rat and rabbit studies
during the same period of organogenesis is important since often the exposure period is of equal
importance to dose level in developmental studies.
Response: refers to sectioa 2.5.1. Section revised because Irvine (1990) and Harrington et al.
(1995b) found to be same study and NOAEL/LOAEL for Harrington 1995b clarified (see
comment re: Harrington study above).
14. Comment (2): The identification of the last two potentially susceptible populations seems
speculative because it is based on hypothetical mechanisms. The level of confidence for
identifying these subpopulations as uniquely susceptible is very low.
Response: Refers to Section 2.5.2. The point is well-taken. Section was revised.
-------
18 Comment (1). Was the whole reproductive and developmental data considered adequate for
risk assessment? Should additional (uncertainty) factors (UF) be invoked for "child health
issues"? This reviewer believes these are adequate and thr-- a further UF is not required.
Response: Yes- whole database considered and with new 2-generation study (CMA, 1997), there
are no database gaps. No further UF needed. Additional explanation was added 4th paragraph,,
section 3.2.
19. Comment (2,3): The approach for RfD calculation is valid and appropriate. The UF of 100
is appropriate (the UF of 3 is no longer needed). (However, one of the reviewers disagrees with
NOAEL -see comment above).
Response: agree
Risk Characterization Summary
20. Comment (1): Can confidence in the CMA study be increased upwards from medium (see
comment above on neurotoxicity)?
Response: No, there were limitations in study design and conduct that preclude upgrading
confidence.
21. Comment (3): Reviewer concurs with "medium" confidence in the critical study (CMA,
1997) in recognition of the limitations to study design, performance and final report presentation.
Reviewer agrees with "medium- to-high" confidence in the RfD based on the database.
Response: agree
-------
Dose-Response fRfD DerivationVCharacterization
15 Comment (1,3): Report clearly discusses the strengths and weaknesses of the weight of
evidence; reviewer concurs with critical study and the NOAEL of 35 ppm.
Response: agree
16. Comment (3): The critical study (CMA, 1997) should be considered only "adequate" as
there were problems with the study design and conduct The reviewer considers the following as
some of the problems: purity (low) of the test material; (presumed) method of correcting for test
material purity; discontinuation of exposure after weaning and testing of pups on PND 60; lack
of evaluations of F2a males and females for preputial separation; lack of collection of
coagulating gland weight; and in assignment of 300 ppm as reproductive and developmental
NOEL and 70 ppm as hematological NOAEL, the laboratory has discounted their data without
explanation.
Response: section 3.1, 4th paragraph. Revised to state study adequate; agree. NOAEL and agree
with reviewer that there are limitations with the study (some so noted above by reviewer and
others also noted in the text of section 3.1).
17. Comment (2): Choice of the reproductive/developmental endpoint and critical study is
appropriate but NOAEL and LOAEL should be 70 and 300 ppm, respectively, not 35 ppm and
70 ppm. A treatment-related decrease in offspring viability is not supported; variations were
small. The NOAEL should be 300 (or possibly 70 ppm), not 35 ppm, for offspring viability.
Significant differences in offspring body weight were detected only in 300 ppm of either Fl or
F2 generations; LOAEL should be 300 ppm and NOAEL 70 ppm for offspring body weight.
There is no convincing evidence that a treatment-related, biplogically relevant, adverse effect
was measured in auditory startle amplitude. (The reviewer constructed tables for males and
females for PND 24). The results for maximum amplitude of response are inconsistent, the
effect of maximum startle amplitude was transitory (there was no effect on PND 60), and there
was no supporting histopathotogical evidence. The-interpretation of the auditory startle-
habituation test results seems equivocal, at best There is no evidence for a NOAEL of 35 ppm
based on "possible" alteration in learning behavior. Likewise, any small decrease in absolute
brain wet weights are more likely a function of smaller body weight rather than a treatment-
related effect to the CNS. An argument can be made that the effects in treated rats are most
likely due to decreased food and -water intake due to altered palatability. A NOAEL of 70 ppm is
prudent and protective, since the true value is likely higher. More description and justification is
needed for the selection of the NOAEL and LOAEL.
Response: Agree with lack of treatment-related effect in offspring viability and body weight-
sections 2.2.2.2 CMA study and 3.1 revised Appreciate reviewer's analyses of data but disagree
with reviewer- there is an adverse effect on auditory startle, possibly learning and absolute brain
weight Section 3.1 expanded to explain transient alterations and different responses for
auditory startle are still considered indicative of taricity and adverse effects.
-------
Peer Review of the Draft Document Entitled
"Health Risk Assessment/Characterization of the
Drinking Water Disinfectant Chlorine
Dioxide and the Degradation By-product Chlorite"
Dated February 6,1998
Prepared for
Health and Ecological Criteria Division
Office of Science and Technology
Office of Water
U. S. Environmental Protection Agency
Washington, DC 20460
By Toxicology Excellence for Risk Assessment (TERA)
Cincinnati, Ohio 45223
I have received and examined the following documents:
1. The draft document specified above under review
2. External peer review of CMA Study 2-Generation (October 9,1997)
3. Response to review comments by CMA Chlorine Dioxide Panel (January 26,1998)
4. Volume 1 (of 6) of the study entitled 'Sodium Chlorite: Drinking Water Rat Two-Generation
Reproductive Toxicity Study" (March, 1996) performed by Quintites England, Ltd., for the
Chloride Dioxide Panel of CMA, Qulntites Report Reference No. "CMA/17/96" (and a
separate copy of title page [p1J and summary pp. 25-27])
5. The charge to the Peer Review Panel by Dr. Vlctd Deltarco, U.S. EPA (dated February 4.
1998)
6. Final Draft for the Drinking Water Criteria Document on Chlorine Dioxide, Chlorite and
Chlorate (dated March 31,1994), prepared for Health and Ecological Criteria Division, Office
of Science and Technology. Office of Water, U.S. EPA, Washington, DC, by Clement
International Corporation.
The peer reviewers have been asked to critique the draft risk assessment document In
terms of (1) the weight of the evidence, (2) the critical effect. (3) the key lines of evidence,
(4) uncertainties in the risk assessment. (5) the scientific basis for the derivation of the reference
dose, as presented in the document, and (6) the quality of the risk assessment document, and
(7) the quality of the critical study (CMA two-generation).
The comments of this reviewer, on whether the document dearly and adequately
discusses the seven aspects specified above, and whether she agrees or disagrees with the
report, are as follows.
1. The Weight of the Evidence
The report clearly discusses the strengths and weaknesses of the weight of the evidence. I
concur with the assessment as presented throughout the report, and especially on p. 24 and
following.
-------
EPA Review
Page 2
2. The Critical Effect
I agree with the report (Section 3.0, pages 25 and following) that the critical study is the
CMA (1996) two-generation study of sodium chlorite in the drinking water in rats. I concur
with the discussion in the report on its strengths and weaknesses (report p. 26, last
paragraph). I would request confirmation from the authors of the study that the NOAEL of
35 ppm is in feel sodium chlorite, since the purity of the test material was presented as
81.4% (Study Report Volume 1, p. 29). The impurities are presented in Volume 1 of the
Study Report, on page 118, and the text of the study report (p. 31) Indicates 'An adjustment
for purity was made using a conversion factor of 1.23.' I interpret this to mean the
concentration of the test material in the drinking was increased to compensate for the purity.
rather than that the calculated intake in mg/kg/day was corrected for purity. The increased
amount of test material (especially the sodium component) may nave contributed to the
observed "palatability problems.* The report assesses the critical study as "wen-designe^..
and acceptable" (p. 17). This reviewer would assess the study only as "adequate."
I concur with the report that the NOAEL for developmental effects is 35 ppm, equivalent to
2.9 mg C1p2"/kg body weight/day.
3. The Kev Lines of Evidence
I concur with the report discussion on the key lines of evidence, building on the Final 1994
Draft of the Drinking Water Criteria Document on Chlorine Dioxide. Chlorite and Chlorate,
and using the CMA study as the critical study.
4. Uncertainties in the Risk Assessment
The report indicates that the confidence In the critical study is 'medium' (report, p. 78), and
Indicates limitations in the design and conduct of the CMA two-generation study. I concur
with the confidence level and wtth the recognition of limitations to study design, performance
and (in my opinion) final report presentation (see this reviewer's Hem no. 7).
The report also indicates that the confidence in the RfD is "medlum-to-high." Based on the
database, I concur.
5. Scientific Basis for the Derivation of the Reference. Dose
The report uses the 2.9 mg/kg/day chlorite, the NOAEL from the two-generation study, and
applies the standard default uncertainty factors of 10 (intfifspedes extrapolation from rats to
humans) x 10 (intr§species extrapolation to protect sensitive subpopulations) x 1 (modifying
factor, as it was deemed unnecessary to account for use of a fess-than-lifetime exposure
since the critical effect was a developmental effort). Therefore, the report rounds the 2.9
mg/kg/day to 3 mg/kg/day. divides by 100 (10 x 10 x 1) to obtain an RfD (reference dose) of
0.03 mg/kg/day "for chlorine dioxide as well as for chlorite.* (p. 27).
This derivation is standard, relatively conservative and uses accepted default values for
uncertainty factors. I concur with the basis, approach and conclusion.
-------
EPA Review
Page 3
6. Quality of the Risk Assessment Document
The document is adequate. There are many typographical errors, some trivial, some
substantive, and some areas when the logic trail/discussion should be explicit, not implicit
Examples follow:
A. Typographical errors
i. Table 1, page 7, Revis et a/. (1986) doses were 0,O.fiZ and 0.5 (not 0. QjT and 0.5)
Rx endpoints to fine up with sexes in Daniel et a/. (1990) reference.
ii. Table 2, page 8, Revis et al. (1986) doses were 0,0.QZ and 0.5 (not 0, fi.1 and 0.5)
iii. Page 5, first full paragraph, sex ratios are usuaRy expressed as > or < 1.0, Le., 0.86
versus 1.12 (not 86 versus 112)
iv. Page 5. line 11, "...contained any.chemical..." (DfiJ "contained and chemical...-)
B. Logic Trail/Discussion not explicit
i. Page 4, Section 2.1, line 13. The status of the referent "neighboring community" as
not exposed to chlorite should be made explicit Were they exposed to any drinking
water disinfectant?
il. Page 4, Section 2.2.2.1, fines 10-14. Statement that the nasal lesions were "most
likely the result of direct contact with the chlorine dioxide solution rather than
Ingestion of the drinking water* and may be an "artifact of treatment" needs some
additional explanatory text
Hi. Section 2.2.2.2, page 16, last paragraph, lines 1-3. The 'reductions in water
consumption, food consumption and body weight gain in all generations' should be
Identified as to which sexes, which doses, and when the effects were observed. This
reviewer's concern is that the typical pattern of palatabittty problems with rats (from
dosed feed or dosed water) is usually initially low intake values which rise over time
as the animals adjust (rats are awesomely practical animals), versus the pattern of
toxitity which is usually initially normal intake values which drop over time as, e.g.,
body weights/weight gains drop, and/or other symptomology appears which is
characteristic of the toxidty. (Volume 1 of the study report does not include this
information.)
iv. Figure 2. Suggest cluster the bar graphs by spedes, i.e., rat, mouse, rabbit, and in
accompanying text on p. 22 also suggest group by species.
v. Section 2.4.1 .top page 23. The difference In sperm morphology outcomes between
the two studies may be due to differences in study duration (72-76 days with effects
versus 5 days with no effects) as weH as to differences in species (rat versus
mouse).
vi. The CMA study is routinely indicated as the CMA (1996)" (e.g.. p. 26 and 27); it
should be "the CMA Study (1996)."
vii. Check following references:
a. Abdel-Rahman et al.".. .C102. C102 and 0103' (p. 29)
b. Add space between the first two Suh et a/, references (p. 32)
c. I could not find "US EPA Guidelines for Developmental Toxkaty Risk
Assessment" (1996). I searched EPA's web site and used our professional
-------
EPA Review
Page 4
''search group. I did find the Reproductive Toxicity Risk Assessment Guidelines
(1996) and the U.S. EPA Guidelines for Developmental Toxicity Risk
Assessment, Federal Register 56(234), 63798-63826 (December 5.1991) (p. 33
of Report).
d. Please complete reference. U.S. EPA National Primary Drinking Water
Regulations (p. 34).
7. Comments on the Critical Studv: CMA M9fl61 Two-Generation Study of Sodium Chlorite
Administered in the Drinking Water to Rats
This reviewer considers the study "adequate"; she has major problems with the design and
conduct of this study, as follows:
1. The test material was 81.4% pure. Analysis including identity of the impurities is
provided (p. 118), jjuj there is no explicit rationale for the use of a test material with this
purity. For example, is this the 'typical' profile of water disinfected with sodium chlorite?
2. The level of purity was corrected for by adjusting the concentration of test material in the
drinking water by a factor of 1.23 (p. 31). It Is hoped that the actual concentrations of
test material are listed somewhere in the report (I only have Vol. 1 of 6).
3. Pups selected for neurobehavioral analyses were noj exposed to the dosed drinking
water after weaning (postnatal day 21) until all the tests were performed, on or about
postnatal day 60 (noted in the risk assessment document and "explained' by the CMA
response) thereby reducing the opportunity to produce and detect effects.
4. The F2a females were not evaluated for prepuHal separation. The text of the report
indicates that "with the exception of 20 males and 20 female pups retained for evaluation
of auditory startle response, the F2a pups were kilted on post natal day (PND) 24 and
subjected to necropsy." (p. 25). BUT Section 16.0. Deviations from the Protocol, item
no. 24 indicates *F2a generation vaginal preforation was not recorded until the majority
of the females had already achieved the criteria in error." (p. 183). The F2a females
must have been older than pnd 24 for the "majority of females" to achieve vaginal
patency since the mean age at acquisition is 33-34 days (CMA response).
5. According to Section 16.0. Deviations from the Protocol, no. 24 (page 25): "For the
[F2a] males there was a delay of four days prior to the commencement of observations
for preputial separation. However, this was considered to have a minimal effect and
hence the data was reported." The report does not specify at what age the F2a male
rats began to be evaluated, but this error may have been responsible, at least in part, for
the apparent lack of an effect on this parameter in F2a males, i.e., the acquisition in
controls may have been earlier than indicated.
6. Section 16.0. Deviations from the Protocol, item no. 17, indicates that although the
protocol specifies collection of weight of the coagulating gland, it was not done. The
reason as indicated was: "However, as this is part of the seminal vesicles and as such
cannot be weighed separately, the weight of the seminal vesicles Included that of the
coagulating gland." This reason Is not acceptable. My laboratory (and many others)
separates the coagulating glands from the seminal vesldes routinely for weighing and
histopathology. This apparent ignorance of relatively standard procedures in
-------
EPA Review
PageS
reproductive toxicity studies causes concern for performance of other aspects of the
study.
7. The performing laboratory considered 300 ppm as the NOEL (no observable effect level)
for reproductive toxicity of the parents and offspring and for developmental neurotoxlc
effects and 70 ppm the NOAEL (no observable adverse effect level) for hematologic
effects (with the NOEL as at "or closely approximate to' 35 ppm) (page 27,
Conclusions). They appear to be discounting their own data without suitable explanation
or rationale.
-------
Peer Review of the Draft Health Risk Assessment/Characterization of the Drinking
Water Disinfectant Chlorine Dioxide and the Degradation Byproduct Chlorite.
In general, the Draft report is well written and provides a good overview of the chlorine
dioxide and chlorite toxiclty studies. The key endpoints most relevant to human heath
are properly identified as potential reproductive and neurodevetopmental toxicriy, based
on several lines of evidence. Previous studies are of limited quality for use in risk
assessment
Identification of Critical Effactfs) and Setting a NOAEL
A recent CMA-sponsored 2-generation reproductive and developmental tadcfty study is
appropriately Identified as the critical study for development of a RfD for both chlorine
dioxide and chlorite. However. I disagree with the NOAEL of 35 ppm (3 ClOr mg/kg-
day) used in this Draft Health Risk Assessment/Characterization document
Considerably more description is needed in the Draft Health Risk Assessment/
Characterization document to justify the selection of the NOAEL and LOAEL used.
The document lists the following significant alterations related to treatment, (listed on page
18):
300 ppm
• reduced Hver weight in F0$, and F1tf&$
reduced pup survival (Increased number of pups found dead and/tor kOfed during lactation)
reduced pup body weight at birth and throughout lactation in F1 and F2
lower tnymus weight In F1 and F2
lower spleen weight in F1 and F2
lower Incidence of pups exhibiting a normal righting
lower incidence of pups with eyes open on PND15
altered clinical condition in F2 animals chosen for neurotoxicity
decreased pup brain weight
delays in sexual development in <5 (preputial separation) and $ (vaginal opening)
in F1 and F2
• lower RBC parameters.
70 ppm
reduced liver weight in FO 9. and F1
-------
Draft Chlorite RfD Peer Review
The Draft Health Risk Assessment/Characterization document (page 17) then identifies a
Developmental NOAEL of 35 and a developmental LOAEL of 70 ppm based on:
1. decreased offspring viability
2. decreased offspring body weight
3. lowered auditory startle amplitude
4. possible alterations in learning behavior
The Draft Health Risk Assessment Characterization document identifies a Systemic
NOAEL of 35 ppm and systemic LOAEL off 70 ppm based on:
1. altered organ weights in the dams
However, when one considers each of the above parameters used as a basis for the
NOAEL and LOAEL, considerable questions are raised. Each of these parameters is
considered below:
Decreased offspring viability:
Results from the CMA-sponsored 2-generation reproductive and developmental toxicity
study of 1996 give little Indication of dose-dependent decreased offspring viability in the
F1 or F2 generations. After reviewing Tables 17,51, and 60 of the CMA-study report I
fail to understand how a NOAEL of 35 ppm and a LOAEL of 70 ppm were derived based
on decreased offspring viability. Comparison of the Gestation Indexes for F1, F2a and
F2b groups does not support a treatment related effect at 70 ppm (93.1,100,100
Gestation Index, for the PI. F2a, and F2b offspring, respectively, that were exposed in
ute/Dto70ppm).
No statistically significant between-group differences were measured In Mean Viability
Indexes, Mean Number of Pups Bom, or Mean Cumulative Survival Indexes by Analysis
of Variance/Kruskal-Wallis analysis. If further statistical analyses were conducted on
these data, they were not made available to this reviewer.
The variations from control groups were small. It is not apparent that any small
decreases from control were treatment related. Therefore, based on these results, the
NOAEL for offspring viability should be 300 (or possibly 70 ppm) rather than 35 ppm.
Decreased offspring body weight:
Review of Tables 19,53 and 62 of the CMA-study report indicates that statistically
significant differences in pup bodyweight at parturition or at various days post-partum
only occurred in the highest treatment group (300 ppm). In the F1 generation, a small
significantly lower mean bfrthweight was reported in both male and female pups, and
this lower weight persisted throughout 24 days post-partum. No statistically significant
differences from control values were reported for the 35 ppm or 70 ppm groups.
In the F2a and F2b generations, only F2a males in the 300 ppm treatment group had
significantly lower birthweight than the corresponding control group. No statistically
significant differences from control values were reported for the 35 ppm. 70 ppm, or 300
ppm F1 a female, F2b male or F2b female groups.
-------
Draft Chlorite RID Pe«r Review
Therefore, significant differences in offspring bodyweight were detected only in the
highest treatment group of 300 ppm of either F1 or F2 generations. Subsequently,
based on this parameter, the LOAEL should be 300 ppm and the NOAEL should be 70
ppm. It is not clear how the authors selected a NOAEL of 35 ppm using offspring
bodyweight as a parameter of adverse effects.
Lowered auditory startle amplitude:
The Draft Health Risk Assessment/Characterization Report states that lowered auditory
startle amplitude was measured in chlortte-treated groups. Indeed, the CMA-study
report indicates,
"On day 24 post-partum the maximum startle amplitude tended to be decreased compared
to control in most trial blocks for groups 3 and 4. These differences achieved statistical
significance when compared with the controls. There were no statistically significant
changes observed in the Day 60 post-partum assessments."
However, from the results resented in the CMA-study report there is no convincing
evidence that a treatment related, biologically relevant, adverse effect was measured.
There are several reasons to reject the acceptance of 35 ppm as the NOAEL based on
the auditory startte-habltuattan response test results:
• Three parameters were measured in the auditory startJe-habHuation response, time
to maximum amplitude of response, maximum amplitude of response, and
habituation to response. No differences were measured between control and treated
groups for measurements of maximum amplitude of response or habituation of the
startle response. Only the maximum startle amplitude was described as statistically
significant in the two highest dose groups. The mean data for the auditory startle
amplitude response are presented In Tabte 43. page 1194. of the CMA-study report
Review of these results raises several questions.
1. It is not clear how the authors of the CMA-study report determined that the
differences In the two highest doses achieved statistical significance. None
of the maximum amplitude values in Table 43 are designated with an
asterisk; hence no indications of statistical significance are given in the table.
2. No consistent dose-dependent decreases were seen between the different
trial blocks in either the mate or females. For example, when the data are
expressed as percent of the control group within each trial block:
Maximum Amplitude of Response to Stimulus: Males, day 24 post-partum
Trials
2-10
11-20
21-30
31-40
41-50
35 ppm
0.95
0.84
0.85
0.83
0.89
70 ppm
0.85
0.84
0.88
0.90
0.89
300 ppm
0.85
0.82
0.87
0.86
0.91
Comment
Any real significant deferences
betwemdoses?
Ho reaTignfccantdOTwsncee between
greatest decrease In madman
anvftMto.
lowest exposure group has greatest
decrease m maximum amplitude.
Mo dose dependent effect Any real .
stanHcant deferences between doses?
-------
Draft Chlorite RID Peer Re\ierr
Maximum Amplitude of Response to Stimulus: Females, day 24 post-partum
Trials
2-10
11-20
21-30
31-40
41-50
35 ppm
1.14
1.19
0.98
0.91
0.94
70 ppm
0.92
0.77
0.74
0.72
0.68
300 pern
0.9b
0.95
0.80
0.72
0.70
Comment
No consistent duwtopondont
decrease. Mid-exposure group haa
Qraatnt decrease In rraxbnun
anwMude.
Any real slgnlflcant differences
Control and high dose? MM-
exposure Qroup tt88 cjreatest
decnaae In maximum amplitude.
No consistent dosfrdependent
decrease. Md-expcsure group has
QniBlMtdacreasa In nwxbreim
A^^^^feu^K
•HJiHIHe.
decrease. .
No consMent dose-dependent
montert decrease in maximum
jumHudB.
These results are very inconsistent In none of the trial blocks does
maximum amplitude decrease with increasing dose. Within three trial blocks
a decreased maximum amplitude in the two highest treatment groups
suggest a possible threshold effect between 35 ppm and 70 ppm (Trials 2-10
in males and trials 31-40 and 41-50 in ferrates), but this is so inconsistent
that no conctuskms can be made. Therefore, these results are too
inconsistent and the interpretation too equivocal to base any NOAEL or RfD
value.
• The effect on maximum startle amplitude was transitory (if there really was an effect).
Maximum startle amplitude responses at 60 days post-partum were not significantly
different.
• No histopathological evidence was detected to support a conclusion of
developmental neurotoxicity.
Therefore, interpretation of the auditory startle-habituation response test results as a
measure of developmental neurotoxicity seems equivocal, at best Unless further
analysis and discussion are provided, I do not accept the auditory startle amplitude data
as a basis for setting a NOAEL at 35 ppm.
Possible alterations In learning behavior:
After reviewing Tables 24-31 of the CMA-study report, I fail to find evidence for a NOAEL
of 35 ppm based on "possible" alteration in learning behavior.
Likewise, any small decrease in absolute brain wet weights are more likely a function of
smaller body weights rather than a direct treatment related effect to the CNS. No
significant differences were detected in brain weights when the data are expressed as a
ratio of organ weight to body weight
Therefore, I reject the NOAEL of 35 ppm tor the chlorite 2-generatton
reproductive/developmental toxicity stVdy of 1996. The LOAEL for this study should be
-------
Draft Chlorite RID Peer Review
300ppm and the NOAEL for this study should be 70ppm (5.9 CIOz- mg/kg/day. rounded
to 6 ClOr mg/kg/day).
An argument can still be made that the effects measured in treated rats of the CMA-
study were most likely due to decreased food and water intake caused by altered
palatability of the chtorite-containing water supply. As pointed out by one of the original
peer reviewers of the CMA-study (and tt was not me), it is quite possible that the true
NOAELs from direct chemical exposure, other than palatability, are higher than indicated
by the data. While this possibility may add some degree of uncertainty to the risk
assessment of chortle, tt errors on the side of caution. Therefore, a NOAEL of 70 ppm is
a prudent, protective value, since the true value likely is higher.
Calculation of Oral RID
The approach for calculating the oral RfD for chlorite is valid and appropriate. An
uncertainty factor of 100 is appropriate for calculating the RfD from the animal study.
The additional factor of 3 is no longer needed to account for the lack of a
multigeneratlonai reproductive study. While the approach is appropriate, the number for
the NOAEL needs to be revised to 6 CKfe- mg/kg-day, as recommended above,
resulting in an oral RfD of 0.06 mg/kg-day.
Othar comments:
The identification of the last two potentially susceptible populations on page 25 seems
highly speculative. While it is true that G-6-PDH-defident individuals and NADH-
dependent methemoglobin reductase-defteient individuals may be more susceptible to
oxldative agents, it Is not at all dear that this Is the mechanism by which chlorite exerts
any reproductive or neurodevelopmental taddty. The identification of these sensitive
populations is based on supposed mechanistic insights. The potential mechanisms
discussed on page 23 are hypotheses. Subsequently, the Identification of these
populations as potentially more susceptible to reproductive or developmental effects of
chlorite seems to be speculation based on speculation. The level of confidence for
identifying these subpopulatlons as uniquely susceptible is very low.
Page 5. first full paragraph, next to last sentence: The quality of the untreated well . ..it
contained and chemical../ Change 'and' to '
Page 5 first full paragraph, last sentence: The atypical baseline data raise concerns
about control.. ." Insert "the" after "about".
Table 2. page 8. last column for the Haag (1949) reference. I suggest you indicate that
the authors attributed the renal effects to a nonspecific salt effect
Page 1 5. line 2. "What malformations raft. . .' Change "What: to Which'. Better yet,
revise the sentence to read, The authors did not state which malformations fell into
each of these categories."
-------
Draft Chlorite RfD Peer
Page 15, last paragraph: "Males were exposed throughout mating then..." Insert "and"
after "mating".
Page 15, last full sentence: "Water concentrations of sodium chlorite..." Revise to read,
"Sodium chlorite concentrations were adjusted downward. .."
Page 23, third full paragraph, first line: Correct "Harringon" to'HarringJon".
Qvarall Summary:
The overall report is acceptable and adequate for risk assessment for chlorine dioxide
and chlorite. My only major disagreement is with the choice of the NOAEL used to
calculate the RID, as described above.
-------
Comments on
Draft Health Risk Assessment/ Characterization Of The
Drinking Water Disinfectant Chlorine Dioxide And The
Degradation Byproduct Chlorite
General Comments
Overalf I found this a reasonably balanced report based on the supporting materials
provided. I believe that the study chosen as being critical (the CMA 2-generation
reproduction study) was the correct study to select for risk assessment purposes,
based on the changes noted and the overall quality and age of the study.
It is also this reviewer's understanding that a complete review of the critical study is
NOT requested, only a review of the risk assessment document.. Therefore, I have
only used the report to clarify certain statements and have not examined the reviews
for .the scientific critique of the report itself.
The comments that I have do not change the approach; study selection or overall RfD
produced. They are more matters of clarification and where more detail could be
provided to sustain and improve the case being proposed.
Specific comments
1. Introduction..
The statement-is made (p3, para 2) regarding the dissociation of chlorine dioxide
under aqueous conditions and in the Gl tract of animals. Can the reviewers provide
-------
references to sustain this view, tt is important since in the last sentence of the
paragraph the key point is made that the toxicity data for one compound are
considered applicable for addressing toxicity data gaps for others. This is a major
assumption and should be qualified to the maximum extent possible.
2. Hazard assessment
The statement (p4, 2nd line) that carcinogenicity data from animal studies are largely
inadequate could be qualified. (See later comments).
3. Studies in humans
I would like to see which human data are available on hematological parameters.
That these agents can cause methemoglobin etc. seems to be well understood but
does not feature in this section - it should.
4. Studies in animals - systemic toxicity.
Nasat lesions and potential confounding with direct exposure to drinking water (p6,4th
para). This section could be expanded a little to explain why this potential artifact of
treatment was noted. There are examples of chemicals that produce nasal lesion via
true oral exposure. I would imagine that a fuller description of the nasal lesion could
convince the reader whether inadvertent treatment to the nasal passages was likely to
happen (e.g. how deep into the turbfnates was the lesion detected etc.) This is an
important point if these nasal observations are to be placed in the appropriate
perspective, for potential risk assessment.
Table 1. (p7)
-------
Could some addition be made to the end points section as to whether the
observations are considered to be adverse effects (e.g. hematological changes etc.).
A laundry list is not helpful and some simple clarification would help.
5. Reproductive and developmental studies
Harrington study. (P11,5th para)
Was water consumption presented? There is a tremendous physiological demand .on
the pregnant dam to drink fluids, whether the water is palatable or not. A presentation
of water consumption data, or summary statement would help.
Why in the 26 and 40 mg/kg/d groups does the finding of significant growth retardation
(-. decreased fetal weights, incomplete fetal ossification), not considered an adverse
event ? The developmental risk assessment guidelines clearly state that an alteration
of growth, even in the presence of maternal toxicrty, should be considered adverse.
unless evidence is presented to satisfy that the events were due to a secondary
(maternal) event. This reviewer would argue that the growth retardation should not be
excluded, particularly if it was consistent and dose -related and would therefore affect
the assignment of a developmental NOAEL for the study.
CMA study (p15,3rd para)
I am perplexed by the statement (p17 2nd para and p26 3rd para) - "Discontinuation of
exposure for the animals undergoing neurotoxicity testing minimizes the likelihood of
finding a positive effect." And precludes comparison etc. The developmental
neurotoxicity testing guidelines indicate exposure should be from gestation day 6 until
postnatal day (pnd) 10 with examination of offspring at a number of times post
exposure up to pnd 60. Moreover, the guidelines state that that using offspring from a
multigeneration study is acceptable. Therefore, the study meets all of the
requirements for a developmental neurotoxicity study and should be -used
-------
appropriately and can be compared to other tests using exactly the same dosing
regime, including stand alone studies. I understand from the accompanying material
that positive control data were provided to enable such a comparison to be made.
6. Carcinogenicity
The initial statement indicates that there are no long term oral bioassays (p18, 1st
para), but in the next paragraph indicate a mouse study over 85 weeks. Why is this not
considered adequate ? 18 month studies in mice are normally considered to be of
utility in the assessment of carcinogenic potential in this species. These data do not
appear to be have taken into consideration (and there may be flaws with this study that
this reviewer cannot ascertain) but to make such bold statements appears
inappropriate, else more qualification is required of the studies presented.
7. Hazard assessment issues
p24, 3rd para animal studies.. A comment is made "While exposure po rats and
rabbits] was by the same route to the same dosage range for essentially the same
days of gestation (my emphasis). It is essential that the reviewers understand
that the period of exposure is the same developmental period (i.e. organogenesis). In
these types of studies, it is often the case that the exposure period (during critical
windows of development) is of equal importance to the administered dose level.
Based on the comments above oh the neurotoxicity evaluation does this increase the
reviewers confidence in the CMA study from medium upwards ?
8. Oral reference dose (p27, section 3.2)
Did the reviewers consider the whole reproductive and developmental database to be
adequate for Risk Assessment purposes ? Should additional factors be invoked for
"child health issues" ? This reviewer believes that these are adequate and that a
-------
further uncertainty factor is not required. However this should be explicitly stated in the
report to avoid ^any confusion.
------- |