This chapter reviews information presented in the Environmental Protection Agency (EPA) draft Integrated Risk Information System (IRIS) assessment of the effects of tetrachloroethylene on the nervous system. It considers first the human evidence, including an evaluation of EPA’s selection of the most critical study on which to base its reference values, and then the evidence from experimental animal studies. The implications of the committee’s evaluation on the derivation of EPA’s reference values for tetrachloroethylene are discussed in Chapter 10.
The epidemiologic studies available for evaluating the neurotoxic effects of tetrachloroethylene were generally cross-sectional. Only one study (Gobba et al. 1998) had outcome measures at two times. Although the cross-sectional study design is limited in establishing temporality in a causal association, the combination of the results of such studies with other information can help to establish an exposure-effect relationship.
In evaluating the human evidence, the committee applied several criteria for determining which studies were the most useful in establishing a reference concentration (RfC) for tetrachloroethylene. The criteria included three general characteristics: the validity of individual studies, the internal consistency of results (for example, Is there an association in the low-exposure group but not in the high-exposure group?), and the consistency of the findings with what is known from other sources (how the study fits into the overall picture of what is known). In selecting studies, the committee considered the target population, the study population, potential confounders, and possible selection or information biases. Statistical issues were also considered. Each study was looked at in the light of those factors, and studies were neither chosen nor rejected on the basis of their results. The selection criteria included consideration of the following factors and questions:
Populations: Are the target and study populations well defined and described? Is the referent group representative of either the unexposed population (in a cross-sectional or cohort design) or of the source population (in a case-control design)? Studies with an inappropriate referent population were given less weight.
Selection of participants: Are the methods for recruiting and enrolling study participants well described? Is there evidence of selection bias? If so, have the authors provided information on the magnitude of the bias? Whether an “effect” is observed in the exposed group is strongly influenced by the choice of the comparison or control group. Thus, the selection and composition of the comparison group is extremely important and in part determines the internal validity of the study. In some cases, there were clear selection biases (for example, selecting comparison groups for the exposed group that did not represent the counterfactual example). That introduces the possibility of selection biases that could easily create the appearance of differences, especially subtle ones, when differences do not exist.
Exposure assessment: How well do the measurements used characterize tetrachloroethylene exposure? How are exposure groups defined? If individual exposure data were available, were they used, or was assignment to exposure groups based on ecologic criteria? In most cases, exposure was estimated at the time of a study. If it is assumed that exposure has only acute, reversible effects, cross-sectional studies are more appropriate. However, if occurrence of an effect when exposure concentrations are low requires long-term exposure, it is important to consider past exposure as well. Exposure assessment ranged from biologic measurements of tetrachloroethylene exposure to environmental exposure assessments. Studies that included measurements and analyses of exposure at the individual level were given greater weight.
Assessment of neurologic outcomes: The end points that were measured in terms of relevance to the visual system and the degree to which the measures are influenced by cognitive function were considered. Studies that used less sensitive measures were given less weight, as were studies that used outcome measures that were more susceptible to observer bias or potential individual confounders (such as ability to follow instructions).
Confounding: Observational studies are always subject to confounding when the exposed and referent groups are imbalanced with respect to factors that are not a result of the exposure but that are also related to the outcome. The committee considered the potential for differences in age, education, learning disabilities, and other variables to confound associations. If the potential for confounding was present and the effects of the confounding were not addressed by the study design or analytic methods, the results of the study were considered to be less credible.
Statistical analysis: Statistical issues were considered, particularly whether the sample size was adequate and whether the approach to analysis was appropriate. Did the studies provide adequate information about the distribution
of exposure levels or results of outcome testing? Were the results influenced by only a few extreme values? If so, was that considered? If continuous data were available, were they used or collapsed as a binary variable, making dose-response analysis or assessment of thresholds impossible? Were tests for interaction of tetrachloroethylene exposure with other variables done? If so, were they properly interpreted?
Having applied those criteria, the committee disagreed with EPA’s selection of the study by Altmann et al. (1995) as the critical study on the basis of which exposure limits should be estimated. EPA selected Altmann et al. (1995) because the data in it represent an environmental rather than an occupational exposure and because a standardized computer-assisted testing battery was used. Although those are reasonable considerations, they are not the most relevant for selecting a critical study. The committee concluded that the validity of the results of Altmann et al. (1995) was seriously compromised by the following methodologic deficiencies.
The reference group was inappropriate, because it did not represent the counterfactual example. The reference group included employees of the Public Health Office or the Medical Institution of Environmental Hygiene, none of whom resided at their place of employment and who may have lived outside the commercial city center. Personal characteristics as well as differences in exposures in the ambient environment may have confounded the analyses of exposure and neurobehavioral outcomes. Evidence of this selection bias is that although matched by age and sex, the referent group was clearly more educated than the exposed group. The distribution of the 14 exposed participants in the low, medium, and high education categories was four, eight, and two, respectively, and that of the 23 controls, one, 12, and 10. The effect of these differences on the study results could not be evaluated, however, because the numbers of years of education represented in the categories were not provided. Adjusting for education with broad categories rather than years of education is not adequate and can easily result in residual confounding by education. Evidence for residual confounding by education can be seen in the variability of results reported by Altmann et al. (1995) depending on the outcome measure. For example, no association between tetrachloroethylene and visual evoked potentials (VEPs) was found. That is important because changes in the visual system and abnormalities in VEPs have been associated with exposure to tetrachloroethylene and chemically related solvents (Bushnell and Crofton 1999; Gobba 2003; Bushnell et al. 2007; Benignus et al. 2009) and selected organic solvents (Benignus et al. 2009) and are unrelated to education. Measures of vigilance, attention, and visual memory are strongly associated with education and premorbid intelligence (Lezak et al. 2004). Those measures showed poorer performance in the exposed group, whereas measures of eye-hand coordination and finger tapping, which are weakly related to education and premorbid intelligence, were similar in the two groups.
The Neurobehavioral Evaluation System (NES) battery used to assess brain dysfunction related to exposure appropriately included four subtests that have been shown in other research to be associated with solvent exposure. However, the battery has no norms for this population, and some of the tests have not been well validated with regard to what they reveal about brain damage from any cause. The absence of norms makes it especially important to have standardized measures of intellectual function that can be used to characterize the native intellectual capacity of the two groups. Examples of such tests are the NES Vocabulary subtest, the Wide Range Achievement Test Reading subtest, and the Wechsler Adult Intelligence Scale Information subtest. Tests of native intellectual function like those are important to include in a battery used to assess neurocognitive outcomes because they are resistant to the effects of central nervous system insults from neurotoxic exposure. They can be used to control statistically for differences in premorbid function between exposed and control groups. Failure to use such measures can cause investigators to conclude that measured group differences in cognitive function are due to exposure when in reality they might exist without any exposure.
The authors indicated that there were 92 potentially eligible subjects, of whom 19 were selected as participants. It was unclear whether the 19 were selected because they were the only ones who had blood tetrachloroethylene over 2 μg/L, lived next to a dry-cleaning facility for at least 1 year, and had no occupational exposure to organic solvents. Even though a blood tetrachloroethylene concentration of over 2 µg/L was required for entry into the study, no concentrations were reported for five subjects (subjects 10-14) taken in their apartments (Figure 1A of Altmann et al. ). Without those specifications, it is impossible to determine whether the sample was biased (that is, whether others were excluded for reasons other than study design).
Tetrachloroethylene was measured in air samples from homes for 7 days. Figure 1B of the paper purports to show indoor air concentrations for exposed participants and controls, but no concentrations are shown for the referent group. For subject 13 of the exposed group, there was no indoor air measurement, there was no tetrachloroethylene concentration in blood drawn in the apartment, and the blood concentration obtained at the time of testing was at the limit of detection (0.5 µg/L). Duration of residence of the 14 exposed ranged from 1 to 30 years; only mean duration was reported, not median. Given only a mean value, there is no way to know whether most of the exposed subjects had relatively short exposures and just a few had long exposures. The amount of time that residents spent in their apartments is unknown. Time out of the apartments before neurobehavioral testing was unknown but was believed to account for the lower blood tetrachloroethylene concentrations before testing. Two exposed subjects had blood tetrachloroethylene concentrations at the limit of detection when tested, whereas the blood concentrations of subject 4 were 30 µg/L in the apartment and 200 µg/L at the time of testing.
In the analyses, exposure is defined by group membership (yes or no) rather than by individual markers of exposure, so a dose-effect relationship could not be assessed. As stated above, group differences in neurobehavioral performance were more likely to be related to residual confounding by education or pre-exposure intellectual capacity than to exposure.
Another paper cited in the draft IRIS assessment that associated environmental tetrachloroethylene exposure with visual-contrast sensitivity (VCS) dysfunction reported on a pilot study by Schreiber et al. (2002). The study also suffered from important methodologic problems that limit its usefulness, including the criteria used to select the exposed group, selection of a noncomparable referent group, and errors in analysis and interpretation. It has been suggested that the significant results reported by Schreiber et al. were influenced largely by two exposed children who had diagnoses of developmental disorders (Storm and Mazor 2004). The total sample in the study was 17, of whom four were children; when the children were excluded from analyses, no significant associations were observed. Given the cross-sectional design of the Schreiber et al. study, it cannot be determined whether exposure preceded the developmental disorders. The small sample makes results highly sensitive to a few observations.
The published papers that the committee judged to be more appropriate to use as a point of departure for derivation of the RfC and reference dose (RfD) were Echeverria et al. (1995), Cavalleri et al. (1994) in combination with Gobba et al. (1998) and Altmann et al. (1990). The reasons for the selections are given below.
Echeverria et al. (1995) conducted a well-designed study of the relationship between acute and cumulative tetrachloroethylene exposure in dry-cleaning shops in Detroit, Michigan, and performance on a neuropsychologic battery. There was no “unexposed” group, but the referent group (lowest exposed; mean air tetrachloroethylene concentrations, not greater than 11.4 ppm) was in the same cohort of dry-cleaning shops as the “exposed” group (mean air tetrachloroethylene concentrations, not greater than 40.8 ppm). Using an internal referent group reduced the potential for the types of selection bias present in many other studies. In the analyses, several potential confounders were considered, including, age, education, verbal skill, alcohol consumption, and prior intoxicant exposure. The authors used a stepwise selection procedure for adjustment, but it is not clear which variables were ultimately used. After adjustment for the covariates, performance on tests for Wechsler Memory Scale Visual Reproduction, NES Pattern Memory, and NES Pattern Recognition was significantly poorer in workers who had a high index of lifetime tetrachloroethylene exposure than in workers who had a low index of lifetime tetrachloroethylene exposure (Table 3-1). Estimated lifetime tetrachloroethylene exposure was positively associated with self-reported “tension” (on the Profile of Mood States) and inversely associated with NES Pattern Recognition scores. Subanalysis
demonstrated some similarity in the test results affected by tetrachloroethylene and alcohol consumption: Visual Reproduction, Pattern Memory, and Pattern Recognition. This similarity underscores the importance of adjusting for alcohol use in analyses of effects of tetrachloroethylene. The study is not without limitations in that recruitment was influenced by the lowering of the permissible exposure limit from 50 ppm to 25 ppm and by owners’ emphasizing the cost of such a change for relatively little effect on health status; therefore, only 23 of a potentially eligible 125 shops participated, for a total of 65 exposed workers.
Cavalleri et al. (1994) examined color-vision loss in 35 dry-cleaning workers in 12 small dry-cleaning shops in Modena, Italy, and in controls who had no solvent exposure and were matched by age, sex, alcohol use, and cigarette-smoking. Inclusion criteria were “apparently healthy,” average daily alcohol intake under 50 g/day, smoking fewer than 30 cigarettes/day, and corrected visual acuity of at least 6/10. Color vision was evaluated with the Lanthony 15 Hue desaturated panel, which was repeated 10 times. Few exposed or control workers were able to perform the test without error. Results wereexpressed as a color-confusion index (CCI) with errors in blue-yellow color vision. Tests were performed monocularly, and the mean CCI for both eyes was used in the analyses, although CCI may be affected in only one eye after tetrachloroethylene exposure. Air tetrachloroethylene concentrations obtained with personal passive sampling for 1 day produced a mean time-weighted average (TWA) for drycleaners of 7.27 ± 8.19 ppm (range, 0.38-31.19 ppm). The mean CCI for the drycleaners was significantly higher (1.192 ± 0.133) than that of controls (1.089 ± 0.117). The statistically significant relationship between TWA of tetrachloroethylene exposure and CCI depended on two extreme values. CCI was not related to duration of exposure or to an integrated index of exposure; only current exposure was known, and there were no data on tetrachloroethylene concentrations in previous years. The study established the protocol and baseline for the Gobba et al. (1998) study 2 years later, which was of greater interest to the committee.
TABLE 3-1 Estimated Meana Neuropsychologic Test Results by Lifetime Exposure to Tetrachloroethylene in Study by Echeverria et al. (1995)
Gobba et al. (1998) re-examined 33 of the workers from the Cavalleri et al. study for color-vision loss after an interval of 2 years. This study was unique in that it examined the same workers at two times. Overall, tetrachloroethylene concentrations remained unchanged for the whole group, but 19 workers (group A) had exposure to significantly increased tetrachloroethylene concentrations at the time of the second assessment, and the remainder (group B) had exposure to significantly lower concentrations because of changes in the processes used in their dry-cleaning shops. Demographic information was provided on the group as a whole but not the two subgroups. The mean CCI increased significantly over the 2 years in group A (from 1.16 ± 0.15 to 1.26 ± 0.18) but remained unchanged in group B (1.15 ± 0.14 and 1.15 ± 0.13). In comparison, the control group from the Cavalleri et al. study, which was not re-examined in the Gobba et al. study, had a mean CCI of 1.08 ± 0.10. The clinical significance of these CCI changes is uncertain. The participants in the Gobba et al. study had exposure concentrations closer to those reported in environmental studies. That the CCI did not improve in the group with lower tetrachloroethylene exposure might be because improvement in workplace conditions had been in place for only a short time or because the visual changes are not reversible.
Altmann et al. (1990) randomly allocated 22 healthy young male subjects to exposure to tetrachloroethylene at 10 ppm or 50 ppm in a chamber for 4 hours on 4 consecutive days, and blood samples were taken for tetrachloroethylene testing and visual and neurophysiologic tests were performed. All subjects had normal visual acuity and no previous solvent exposure. Increased latency in VEPs was observed in subjects exposed to tetrachloroethylene at 50 ppm, and decreased latency at 10 ppm; the greatest effect was observed on the last day of exposure. VEPs with the smallest visual angle and on the last day of exposure provided the greatest intergroup differences. VCS tests on five subjects (two at 50 ppm and three at 10 ppm) showed improvement at the low and intermediate spatial frequencies in the 10-ppm group but loss in the 50-ppm group. Brainstem auditory evoked potentials were not associated with tetrachloroethylene exposure. The lowest observed-adverse-effect level (LOAEL) appeared to be 10 ppm for VEP outcomes.
A second paper (Altmann et al. 1992) published on the above study summarized data on neurobehavioral outcomes but is not recommended for use in determining reference values. Performance during 4 days of exposure was compared with performance obtained on day 1 in the chamber, when there was no exposure. The NES subtests measuring mood and “cognitive function” showed no decrement in performance with days of exposure, but the continuous performance test, tracking task (hand-eye coordination subtest), and simple reaction time task showed improvement over time that was more pronounced in the 10-ppm control group than in the 50- ppm exposure group. However, the measure of premorbid function used in the study (a vocabulary test) was not included as a control measure in the data analyses; it might have affected the outcomes on all NES subtests, especially those of learning and memory. Some NES subtests were given only twice and some at every session; it is not clear which were
given when, but it might have influenced which test outcomes had significant results because of differences in practice effects.
This section describes controlled-exposure studies of experimental animals. As noted in the draft IRIS assessment, most animal studies have involved inhalation exposures to tetrachloroethylene at concentrations of about 30 ppm to over 1,000 ppm or administration by noninhalation routes of tetrachloroethylene at 100-to 4,000 mg/kg. Because of the relevance of the exposure regimen, the inhalation studies are emphasized here. However, it should be noted that studies like that of Warren et al. (1996) and Moser et al. (1995) deliver a known amount of tetracholorethylene by other routes (for example, by gavage) and also support tetrachloroethylene’s neurotoxicity. Warren et al. reported effects on a refined end point, schedule-controlled behavior, and linked behavioral deficits to blood and brain concentrations. Moser et al. (1995) used a broad range of doses administered acutely or “sub-acutely” (14 days) and reported LOAELs and noobserved-adverse-effect levels (NOAELs) on a well-characterized Functional Observational Battery.
Incorporating the animal literature into an assessment of tetrachloroethylene’s neurotoxicity has several advantages. The animal literature can demonstrate the plausibility of claims that neurotoxicity occurs, identify the role of dose and duration of exposure in neurotoxicity, discover neurotoxic effects for further study in humans, confirm with controlled exposures that neurotoxicity occurs in a specific domain, link effects to tissue concentrations, and determine mechanisms of action and similarities and differences between other compounds in the same class. The animal studies entail known histories and living conditions and controlled exposure conditions, usually over a range of doses or concentrations; this allows assessment of dose-effect relationships under conditions that are less influenced by the covariates and biases that hamper the interpretation of human exposures.
The literature describing controlled acute and subchronic inhalation exposures of laboratory animals is summarized in the EPA document. The end points affected include neurotransmitter or neurochemical concentrations (Honma et al. 1980; Nelson et al. 1979; Briving et al. 1986; Karlsson et al. 1987), long-chain fatty acid concentrations (Kyrklund et al. 1984, 1987), RNA expression (Savolainen et al. 1977), DNA expression and brain weight (Rosengren et al. 1986; Wang et al. 1993), electrophysiologic measures and evoked potentials (Mattsson et al. 1998), and locomotor activity (Savolainen et al. 1977; Kjellstrand et al. 1985; Szakmary et al. 1997), all of which indicate tetrachloroethylene’s neurotoxcity. Some studies published after the draft IRIS assessment was written have applied physiologically based pharmacokinetic (PBPK) modeling to characterize not only the dose to which an animal is exposed but the concentration at the target tissue for neurotoxicity, the brain (e.g,, Boyes et al. 2009).
The incorporation of PBPK modeling will facilitate generalization among species and among routes of exposure. The process can contribute to the identification of mechanisms and modes of action and can enhance understanding of the comparative toxicity of different solvents.
The animal studies have limitations. Most notably, as in the studies of controlled human exposure, they use concentrations that are much higher and durations that are much shorter than those experienced environmentally or occupationally. Incorporating their results into a risk assessment must entail the application of uncertainty factors to identify hazard at environmentally, or even occupationally, relevant concentrations. In addition, the dependent measures in most studies differed from those identified in the human literature as particularly sensitive to tetrachloroethylene exposure. In contrast, recently published papers, such as those by Oshiro et al. (2008) and Boyes et al. (2009), use end points that are directly relevant to humans.
The draft IRIS assessment reviews two papers by Kjellstrand et al. (1984, 1985 [see Table 4-6, page 4-409 of EPA 2008]) for neurotoxicity. However, the 1984 study is not appropriate for assessing neurotoxicity; its strengths are that it involved doses that ranged from 9 to 3,600 ppm and durations that ranged from 1 to 120 days and continuous exposure or exposure for a different number of hours per day, but no central nervous system end points were examined. EPA reports that brain butyrylcholinesterase activity was affected, but plasma was analyzed, so the relevance to neurotoxicity is unclear. Some mice were examined for locomotor activity, but exposure and effects are poorly described and unusable. Although the exposure was acute, the relationship between locomotor activity and exposure is described better in the 1985 paper.
Overall, the animal studies support the conclusion that tetrachloroethylene is neurotoxic, but, except for the study by Mattsson et al. (1998), the end points used in the animal studies that were reviewed by EPA were nonspecific and not directly related to the visual or cognitive effects reported in the human literature. The studies therefore provide only indirect support for EPA’s conclusions. The studies by Mattsson et al. entailed exposure 6 hours/day 5 days/week for 13 weeks and examined VEP and other functional effects, so their results are directly pertinent to human exposures. A NOAEL and a LOAEL were identified. Several related reports have been published since the draft IRIS assessment was written (for example, Boyes et al. 2009; Oshiro et al. 2008); they describe dose-effect relationships, spanning a broad range of doses, between acute exposure and visual and signal-detection end points.
In the Boyes et al. (2009) study, rats were exposed head-only to tetrachloroethylene while VEPs were recorded. Exposures were to concentrations of tetrachloroethylene ranging from 1,000-4,000 ppm for 1-2 hours, using concentration and time combinations derived from kinetic analyses. The most sensitive end point was the F2 (frequency-doubling) component of the evoked potential spectrum, a measure thought to reflect the activity of cortical neurons that respond to both stimulus offset and onset. Boyes et al. also conducted a toxicokinetic analysis relating exposure concentration (250-4,000 ppm) and duration (1
hour followed by a 6-hour washout period) to brain concentration. From this analysis, the investigators were able to link brain concentrations of tetrachloroethylene to visual function and to estimate an ED10 of 0.68 mg/L and ED50 of 47 mg/L.
In the study by Oshiro et al. (2008), rats were exposed by inhalation to tetrachloroethylene at 500, 1,000, and 1,500 ppm for 1 hour, during which a visual signal detection task was performed. Rats were trained to indicate the occurrence or nonoccurrence of a light flash during a trial period that lasted from 0.3 to 24.39 seconds, and individual trial durations were random. Exposure to tetrachloroethylene did not change the number of “correct” detections, but significantly increased the number of times that the rats incorrectly indicated a signal (false alarm), increased response time, and decreased the number of trials completed. The false-alarm rate was affected at the lowest concentration (500 ppm) and a NOAEL was not identified. The authors concluded that the results suggest attention deficits.
EPA also reviewed animal studies conducted with intraperitoneal or oral exposure. The studies of exposure of adults included functional observational batteries (Moser et al. 1995), locomotor activity (Fredriksson et al. 1993; Motohashi et al. 1993), and schedule-controlled operant behavior (Warren et al. 1996). EPA did not use the studies in establishing an oral RfD for chronic adult exposures, because effects occurred at high doses (150 mg/kg per day or higher) in the well-controlled studies.
The mode of action for tetrachloroethylene’s neurotoxicity is discussed in a separate section of the draft IRIS assessment (Section 4.6.4). The assessment notes that while the mechanism by which tetrachlorethylene acts is unknown, the evidence is good that it acts on ligand-gated ion channels like other organic solvents. EPA correctly notes that solvents act similarly to ethanol on GABAA receptors and that there are orderly structure-activity relationships, but the citation in support of this observation (Mihic 1999) reviews ethanol and not other solvents. As implied in the IRIS assessment, tetrachloroethylene’s effects on brain fatty acids are interesting but its functional significance is not clear. A weakness of the IRIS assessment’s treatment of the evidence on tetrachloroethylene’s mechanism of neuorotoxic action is that it is entirely descriptive and isolated from the rest of the document. Specifically, the implication that it resembles other volatile organic solvents is not used elsewhere in the document in support of tetrachloroethylene’s toxicity to the adult or the developing nervous system. In light of the importance of neurotoxicity to the development of the RfC, this is surprising.
The literature on developmental neurotoxicity is limited. EPA’s discussion of this important issue is distributed between the sections on neurotoxicity and reproductive toxicity. In light of the sensitivity of the developing nervous system
to neurotoxicants, including solvents (Costa et al. 2004; Grandjean and Landrigan 2006; Slikker 1994), the topic should have been given separate treatment. The EPA document appropriately raises concerns that the studies of tetrachloroethylene-exposed children are small or sufficiently problematic that firm conclusions cannot be drawn from them. Several effects have been reported, including alterations in sensorimotor function (Nelson et al. 1979; Umezu et al. 1997), brain neurochemistry (Nelson et al. 1979), and locomotor activity (Fredriksson et al. 1993; Motohashi et al. 1993; Nelson et al. 1979; Szakmary et al. 1997). Some of these studies used very high concentrations, but others involved concentrations relevant to potential human exposures.
Nelson et al. (1979) exposed pregnant rats to tetrachloroethylene at 900 ppm on gestational days 7-13 or 14-20 or at 100 ppm on days 14-20. No significant tetrachloroethylene-related effects were reported in the animals exposed at 100 ppm, but effects were noted in those exposed at 900 ppm. The tetrachloroethylene-exposed dams consumed less feed and gained less weight than air-exposed controls. No significant differences in growth were noted in offspring, but the draft IRIS assessment incorrectly states that diminished weight gain in offspring was reported. Offspring showed deficits in neuromuscular and sensorimotor functions and increases in locomotor activity.
Fredriksson et al. (1993) also reported changes in locomotor activity in 60-day-old rats after oral exposure to tetrachloroethylene administered (at 5 and 320 mg/kg) on postnatal days 16-20; the effects were not dose-related. The draft IRIS assessment appropriately raised a concern about adequate control for litter effects in the study. It is widely accepted that litter effects must be controlled for in analyses of developmental exposure. Usually litter effects are handled by including only one pup, or one pup per sex, from each litter in studies of prenatal or perinatal exposures. That is, to avoid “litter effects,” the litter should be the statistical unit. A failure to follow that convention inflates the type I error rate. Fredriksson et al. (1993) did not follow it but instead assigned pups to treatment groups randomly, so some treatment groups contained siblings. Some of the authors of the paper have argued that their approach is appropriate and does not inflate the type I error rate (Ericksson et al. 2005); their discussion is also cited in the draft IRIS assessment. Because exposures took place on postnatal days 16-20, the extent to which litter effects confounded the results in the 1993 Fredriksson et al. study is unclear. Nonetheless, the absence of a dose-effect relationship is of concern.
In a short communication, Kyrklund and Hagid (1991) described changes in brain fatty acids of neonatal guinea pigs exposed to tetrachloroethylene at 160 ppm during gestation, but the samples were very small, and many important details were lacking. As noted in the draft IRIS assessment, there was evidence of litter effects in this study, and EPA correctly notes that there are concerns about the absence of a dose-effect relationship and of important methodologic considerations, such as use of non-blinded observers on end points that involved subjective observations and difficulty in relating intraperitoneal routes of administration to oral or inhalation routes.
As noted in the draft IRIS assessment (section on “Mode of Action for Neurotoxic Effects” [4.6.4]), tetrachloroethylene has much in common with other volatile organic solvents, anesthetics, and alcohols. These shared mechanisms, coupled with similarities in the kinetics of these compounds and the high vulnerability of the developing brain to organic solvents and alcohols, raise concerns about the vulnerability of the developing organisms to tetrachloroethylene. The material on developmental neurotoxicity, while identifying the studies directly pertinent to tetrachloroethylene, omits mention of evidence that might be derived from similarly acting compounds. A separate section might have addressed these issues more thoroughly.
FINDINGS AND RECOMMENDATIONS
EPA’s selection of neurotoxicity with emphasis on the outcomes of cognitive and visual dysfunction in adults is appropriate as an end point for deriving a point of departure for development of its reference values. However, the committee disagrees with EPA that the study by Altmann et al. (1995) should be the basis for the noncancer risk values. The committee recommends the use of studies by Altmann et al. (1990), Cavalleri et al. (1994) as a baseline for Gobba et al. (1998), and Echeverria et al. (1995). A new animal study by Boyes et al. (2009) also provides a strong basis for a point of departure. Those five studies provide a stronger scientific basis for deriving the RfC and RfD. Despite the importance of the developing nervous system, the literature on potential neurodevelopmental effects is not sufficient to support the derivation of an RfC. This does not mean that developmental neurotoxicity is unlikely. The broader solvent literature raises significant concern about potential developmental neurotoxicity. While the draft IRIS assessment notes that tetrachloroethylene enters the developing brain, it appears to dismiss the potential for developmental neurotoxicity independent of reproductive or maternal toxicity.
Additional research may help to fill gaps in the evidence. For example, studies of developmental neurotoxicity are needed to fill an important gap in the database on tetrachloroethylene. Well-designed epidemiology studies of tetrachloroethylene and neurological end points that characterize both past and current exposure would be helpful. These studies should be done in populations with a range of exposures (such as occupational studies with a wide distribution of exposure and environmental exposures via both air and water).