This chapter comments on the discussion of neurotoxicity in the U.S. Environmental Protection Agency (EPA 2001b) draft risk assessment of trichloroethylene and reviews information on the effects of trichloroethylene on the nervous system generated since that document was released. Other recent reviews are considered, including those of the Agency for Toxic Substances and Disease Registry (ATSDR 1997a) and the New York State Department of Health (NYDOH 2005b). The chapter also addresses (1) information about the effects on complex cognitive functions, (2) sensitive populations, (3) known interactions of trichloroethylene with other exposures that may affect the risk for neurotoxicity, (4) the role of trichloroethylene concentrations in the brain, (5) the potential role of trichloroethylene in the development of neurodegenerative diseases, (6) potential mechanisms of effect and their implications for complex behavioral function, and (7) research needs.
In the past, trichloroethylene was widely used as an anesthetic at concentrations of approximately 2,000 parts per million (ppm). That use was generally restricted around 1977 because of adverse effects associated with such treatments (ATSDR 1997a). Given trichloroethylene’s anesthetic uses and its widespread use in occupational settings, significant information is available on the acute toxicity of trichloroethylene and its metabolites. Surprisingly, little information exists on the effects of more protracted exposures on the central nervous system, either in humans or in experimental
models, particularly at lower concentrations of exposure. Where studies are available, information from human populations often relies on estimated rather than actual concentrations of exposure, making it difficult to evaluate risks to health. In addition, much of the literature related to trichloroethylene exposure in humans includes exposures to mixtures of solvents, so that it is difficult to evaluate the specific contribution of trichloroethylene to health outcomes.
Experimental studies of acute exposures in rats have shown behavioral alterations across several functional domains at a range of concentrations that overlap with those associated with effects in humans. Most of these studies involved inhalation exposures. At higher concentrations of exposure (e.g., 1,000-4,000 ppm), reported effects include hearing loss, impaired oculomotor control, seizures, decreased wakefulness, and anesthetic effects such as lethargy and ataxia.
Auditory deficits have been observed in several studies at comparable exposure concentrations and in different strains of rats, attesting to the generality of the effects. These studies show auditory effects to occur primarily for the midfrequency tone range (Mattsson et al. 1993; Crofton and Zhao 1993; Jaspers et al. 1993; Rebert et al. 1993; Crofton et al. 1994). Studies have indicated the persistence of some adverse auditory effects, as evidenced 14 weeks postexpsoure to trichloroethylene at 4,000 ppm for 6 hr/day for 5 days (Crofton and Zhao 1993). Apparently, this outcome has not been studied after acute exposures of humans to high concentrations of trichloroethylene.
In rat models, high doses of trichloroethylene administered orally (2,500 mg/kg per day, 5 days per week for 10 weeks) result in morphologic changes in nerves, including alterations in myelination characteristics of the trigeminal nerve (Barret et al. 1991, 1992). These findings are consistent with reports of cranial nerve damage in humans (e.g., Cavanagh and Buxton 1989). A role was noted for the trichloroethylene degradation byproduct dichloroacetylene in eliciting these effects.
Effects on behavior at lower trichloroethylene concentrations in experimental studies have included impaired effortful motor response in rodents (measured by swimming performance) and decreased response of rats to avoid electric shock after a 4-hour exposure to trichloroethylene at 250 ppm (Kishi et al. 1993). The concentrations at which Kishi et al. (1993) observed effects are similar to those noted by Stewart et al. (1970) in humans reporting headaches, fatigue, and drowsiness after exposure to trichloroethylene
for 7 hr/day for 5 days. ATSDR (1997a) used these studies to formulate a minimum risk level for acute duration inhalation in humans.
A newer study by Ohta et al. (2001), not available at the time of the EPA or the ATSDR review, examined the effects of trichloroethylene on long-term potentiation (an enduring increase in the efficacy of specific brain pathways), one of the hypothesized neurophysiologic mechanisms for learning. They evaluated measurements of long-term potentiation in hippocampal slices in mice 24 hours after single exposures to trichloroethylene. They observed dose-related decreases in potentiation of the action potentials of a population of neurons (population spikes) after tetanus treatment, with reductions of 15% at 300 mg/kg and of 26% at 1,000 mg/kg. The size of the area responsive to potentiation was also reduced by trichloroethylene exposure. The animals did not appear to be anesthetized by this dose. One difficulty in comparing exposures for the effects observed by Ohta et al. (2001) with those from other studies of acute exposures is the difference in route of exposure. Ohta et al. (2001) used intraperitoneal injections and did not provide any information about peak brain concentrations of trichloroethylene produced by this exposure. However, efforts should be made to estimate from physiologically based pharmacokinetic models what the peak brain concentrations would be in this study and how they might compare with other routes for potential utility in evaluating acute-exposure risk assessment, given the nature and magnitude of the reported effect and its observation 24-hours postexposure.
Significantly more information is available with regard to inhalational exposures in rats. Many studies have focused on sensory-based alterations in response to trichloroethylene. Reported effects include changes in the amplitude of flash-evoked potentials (visual function) (Blain et al. 1992; Mattsson et al. 1993), reduced acoustic startle response, and auditory-evoked potentials (auditory function) (Rebert et al. 1991; Jaspers et al. 1993), consistent with the auditory effects described above. As with higher concentrations, lower concentrations alter the shock-avoidance response (Goldberg et al. 1964).
Among studies using intermediate subchronic exposures, the lowest concentration of trichloroethylene associated with effects on the nervous system comes from a report by Arito et al. (1994). Rats exposed for 8 hours a day, 5 days a week for 6 weeks showed decreased wakefulness and increased slow-wave sleep during the period of exposure. When measured 22 hours after exposure, the rats showed decreased heart rates during sleep. The effects on wakefulness and sleep were observed at exposures of 50 ppm, as well as at 100 and 300 ppm, and were not dose related. Moreover, they
persisted over the 6 weeks of the study (no adaptation was observed). These effects might relate to the fatigue and lethargy associated with exposure to trichloroethylene in human studies. ATSDR (1997a) used data from the 50-ppm exposure to trichloroethylene by Arito et al. (1994) to determine an intermediate-duration inhalation minimal risk level of 0.1 ppm. EPA (2001b) used a lowest-observed-adverse-effect level (LOAEL) of 50 ppm from this study to derive a pharmacokinetic-adjusted human equivalent concentration of 9 ppm and a benchmark dose associated with a 10% response (BMD10) of 5 ppm. These levels of effects are directly comparable to LOAEL values determined from human studies based on chronic exposures (described later in this chapter).
Isaacson and Taylor (1989) studied the effects of trichloroethylene on rats exposed during development (gestation and lactation) at concentrations of 312, 625, and 1,250 mg/L in drinking water. These exposures were reported to increase exploratory behavior in 60- and 90-day-old offspring, with the highest exposure concentration increasing locomotor activity at 60 days of age. These exposures likewise resulted in a 40% reduction in the number of myelinated fibers in the hippocampus, a region critical to complex cognitive function (both 312 and 625 mg/L or, equivalently, 4.0 and 8.1 mg of trichloroethylene per day, respectively). It is not known, however, whether these concentrations were associated with effects on maternal weight gain during pregnancy or on litter size and pup brain and body weights. Also, doses to the pups are not known, making extrapolation difficult. Nevertheless, these effects speak to the potential for permanent damage resulting from trichloroethylene exposure during development.
Studies in gerbils reported changes in the expression of protein concentrations in the brain at lower concentrations of trichloroethylene. This includes inhalation exposures to trichloroethylene at 60 or 320 ppm for 3 months, followed by a 4-month postrecovery period, after which increases in proteins appeared in multiple brain regions, even in response to the lower concentration. DNA was elevated in two regions at 320 ppm, with a LOAEL of 60 ppm (Haglid et al. 1981). A decrease in S100 protein concentrations, thought to be a marker of brain damage, was observed after exposure to trichloroethylene at 170 ppm or after intermittent exposure at 500 ppm for 5 months, with no postexposure recovery period (Kyrklund et al. 1984). These seemingly opposite effects of exposure to trichloroethylene could reflect the dynamics of the protein response over the period of exposure and recovery rather than discrepancies in outcome. Although these findings are potentially interesting, it is difficult to extrapolate from gerbils to humans, because kinetic characteristics of trichloroethylene in gerbils are unknown relative to rats and mice and there have been no follow-up reports in rats or mice.
Studies of subchronic exposure to trichloroethylene in rats reported after the EPA (2001b) draft risk assessment include that of Poon et al. (2002) and Oshiro et al. (2004). Poon et al. (2002) examined the effects of oral exposures to trichloroethylene at 0, 0.2, 2, 20, and 200 ppm in male and female Sprague-Dawley rats for 13 weeks. Of relevance to neurotoxicity were measures of histologic changes in the myelin sheath of the optic nerves and concentrations of biogenic amines, determined in several different brain regions in a subset of males. In the absence of reductions in food or water intake or in body weight gain, the authors reported a mild vacuolation of the myelin sheath at the highest concentration (200 ppm) in 30%-70% of the animals examined. However, there was no associated axonal degeneration or lymphoid infiltration, making interpretation of these findings difficult. It was not clear how many animals were examined for this effect, how the 30%-70% range was determined, or whether these tissues were examined in a blinded fashion. Because measurement was taken at a single time point during the exposure, it is difficult to determine the degree of damage it signifies and whether such effects were progressive with time or represent early exposure effects. No changes in biogenic amines were reported by Poon et al. (2002), as examined in frontal cortex, caudate nucleus, nucleus accumbens, hippocampus, and substantia nigra. However, the sample sizes (n = 5) used for this component of the study compromise the ability to detect such changes, which normally would require sample sizes up to double those used here. This problem is clear from the measures of variability presented for these data, with standard deviations as high as 50% for the control group in some cases. Thus, the general absence of effects in this study may reflect experimental parameters rather than an insensitivity of the nervous system to these concentrations of trichloroethylene. Given these limitations, the utility of these data for risk assessment is questionable despite the focus on low doses of trichloroethylene.
It appears that repeated exposures to trichloroethylene can impair sustained attention but do not seem to produce a residual impairment in this behavioral process. Bushnell and Oshiro (2000) reported that exposures to trichloroethylene at 2,000 or 2,400 ppm via inhalation for 9 days disrupted performance on a sustained attention task, decreasing the probability of a hit (correct response to a signal) and increasing response time as well as the number of response failures. Tolerance to these impairments developed over the course of exposure, however, and additional work is required to determine whether this reflected metabolic or behavioral tolerance.
In a follow-up study, Oshiro et al. (2004) examined the residual neurological effects of exposure to trichloroethylene at 0, 1,600, or 2,400 ppm for 6 hours a day for 20 days in adult male rats, with evaluation of learning a sustained attention task beginning 3 weeks postexposure. No exposure-
related effects were found. Both ethanol and d-amphetamine impaired performance on the task, with a more pronounced reduction of the probability of a correct response to a signal at the highest dose of amphetamine in the group exposed to trichloroethylene at 2,400 ppm. The authors suggested that the discrepancy between the findings of this study and those of human studies that found residual deficits in cognitive function might be due to differences in the duration and number of exposures to trichloroethylene. Human studies that found residual effects involved exposure durations ranging from a mean of 3 years to 24.5 years, whereas the exposures in the Oshiro et al. study were estimated to be equivalent to 2.6 and 3.8 years of exposure for humans. Furthermore, many previous reports of attention effects reflect exposures to mixtures of solvents rather than trichloroethylene alone, raising questions about the specific components of the exposures that would have contributed to the effects.
Although these findings suggest no residual learning deficits after trichloroethylene exposure, the differential effects of amphetamine in control versus trichloroethylene-treated animals could indicate residual effects on brain dopamine neurotransmitter systems after trichloroethylene exposure. Dopamine pathways of the central nervous system are critical for cognitive and executive functions. Moreover, they further support the potential for trichloroethylene to have protracted effects even after exposure ceases.
Waseem et al. (2001) examined neurobehavioral effects of trichloroethylene in rats after oral administration for 90 days of 350, 700, or 1,400 ppm or after inhalation exposure of 376 ppm for 4 hours per day, 5 days per week, for a total of 180 days. Locomotor activity was measured in addition to “cognition” which was evaluated using acquisition of a conditioned shock-avoidance response. Neither oral nor inhalation exposure resulted in differential effects on acquisition of the response as measured for 7 days immediately after 90 days of oral exposure or 180 days of inhalation exposure. One problem with interpretation of these studies was the experimental design in which acquisition was studied after trichloroethylene exposure. The acquisition of shock avoidance is critically dependent on the intensity of the shock stimulus. It is possible that exposure to trichloroethylene altered shock sensitivity per se. If shock sensitivity were actually reduced, one might expect a lower rate of acquisition. Thus, rates of acquisition would have to be “normalized” to shock sensitivity. Differences in sensitivity to shock per se were never compared between the two groups. In addition, rats exposed to trichloroethylene had higher levels of locomotor activity. The increased levels of motor activity also could have contributed to levels of shock avoidance, causing higher levels of movement between the chambers of the shuttle box used to measure avoidance. Although these increases were stated to be significant at days 30 and 90 of exposure and not statisti-
cally significant at day 180, the trends were still evident at 180 days, and the small number of animals used in the experiments (six per group) would likely have precluded the ability to statistically confirm such differences, particularly as locomotor activity varies substantially among animals. For these reasons, it is not possible to determine whether there were differences in learning between trichloroethylene-exposed and control animals in these experiments. Even if there were differences, the values here were not below current LOAEL values.
Few experimental studies examining the effects of chronic exposure to trichloroethylene (>365 days) have been reported. In one study (NTP 1988), rats were administered trichloroethylene at 500 or 1,000 mg/kg per day for 103 weeks via gavage. The report described (but did not quantify) transient postdosing effects in rats that are consistent with previous reports in human and experimental exposures (lethargy, ataxia, and convulsions). One notable observation from that study suggesting a sensitization effect was that some convulsions occurred before dosing, during the weighing period. A study of mice exposed for 54 weeks to 2,400 mg/kg per day (males) or to 1,800 mg/kg per day (females) reported nonquantified observations of excitation immediately after dosing followed by anesthetic-type effects (Henschler et al. 1984). These reports indicate a consistency of trichloroethylene’s effects across species but do not provide information of particular use to the risk assessment process. Moreover, these studies examined relatively high concentrations of trichloroethylene as they were carried out in the context of carcinogenicity evaluations.
Not surprisingly, given its anesthetic properties, chronic exposures to trichloroethylene have been shown to alter neurotransmitter functions. In a study of gerbils exposed to trichloroethylene for 12 months via inhalation at 50 and 150 ppm, dose-dependent increases (52% and 97% for glutamate; 69% and 74% for γ-aminobutyric acid [GABA], respectively) were observed in the uptake of glutamate and GABA in the posterior cerebellar vermis but not in the hippocampus; perchloroethylene did not produce corresponding changes, suggesting some specificity of the effect (Briving et al. 1986). These effects were seen in the absence of any changes in body or whole brain weights and so do not appear to reflect systemic toxicity. The LOAEL for this study was 50 ppm. Difficulties in using these data include extrapolation to human exposure, given differences between the gerbil and more standard rat and mouse models for which toxicokinetic parameters are well described. Nevertheless, they appear to support the EPA-derived LOAEL of 50 ppm from Arito et al. (1994).
Effects from acute (<14 days) exposure to trichloroethylene are widely reported in humans. At lower exposures (50-300 ppm), headache, fatigue, drowsiness, and inability to concentrate are reported. As the trichloroethylene concentrations increase, dizziness, loss of facial sensation and unconsciousness can occur. With acute exposures to high concentrations (albeit highly unspecified [e.g., 1,000 ppm and above; anesthetic use was approximately 2,000 ppm]), trichloroethylene has been associated with dizziness, headache, euphoria, sleepiness, nausea, confusion, and visual and motor disturbances. Acute exposures to high concentrations, most often due to accidental occupational exposures at unspecified concentrations, have been associated with nerve damage (typically cranial nerves) and residual neurological deficits, including memory loss when measured as long as 12-18 years later (e.g., Feldman et al. 1985). A remaining uncertainty is whether the reported nerve damage results from trichloroethylene or from a metabolite.
In many reports, actual concentrations of trichloroethylene are unspecified. In a controlled exposure experiment using trichloroethylene at 100 ppm for 6 hours a day for 5 consecutive days, Triebig et al. (1977) reported no statistically significant differences between exposed and control subjects in standardized achievement tests and self-reporting scales. Stewart et al. (1970) used human volunteers exposed to trichloroethylene at defined concentrations for specified durations to evaluate changes in motor function. Outcomes in these tests were normal in response to 200 ppm, but subjects complained of fatigue and drowsiness as well as a need to exert greater mental effort on the tests, an effect that may reflect the symptoms described. In a review of the literature, ATSDR (1997a) used the study by Stewart et al. (1970) to derive an acute-duration inhalation minimum risk level. This was later adjusted to determine an intermediate-duration inhalation minimal risk level of 0.1 ppm.
Intermediate Chronic Duration Exposure
The assessment of intermediate subchronic exposures (15-364 days) reviewed by both ATSDR (1997a) and EPA (2001b) focused on studies in rats because most human studies involved chronic exposures. In general, effects from intermediate subchronic exposures are similar to those reported at higher concentrations but occur at lower trichloroethylene concentrations with more protracted duration exposures.
In a review of the published literature, ATSDR (1997a) did not cite any long-term exposure studies in humans, because exposures to trichloroethylene in these studies are unspecified (estimated rather than empirically measured). EPA (2001b) adopted a different approach, validly recognizing the adverse effects reported for humans experiencing chronic exposure. The approach reflects a weight of evidence of effects from low-dose exposures that included adverse outcomes on the central nervous system, as well as other target organs, and involved development of reference doses and reference concentrations (RfCs) and pharmacokinetic modeling. Thus, in its evaluation, EPA was more inclusive in its use of information related to longer-term exposure.
Studies of central nervous system toxicity identified by EPA include effects that are also reported in response to shorter durations and higher concentrations of trichloroethylene and, thus, represent a continuum of these effects along the dose and exposure duration-response curve, with trichloroethylene effects appearing at lower concentrations when exposure durations are longer. In addition, the LOAELs from four different animal studies examining nervous system effects show a high degree of correspondence, ranging from 20 to 50 ppm, with corresponding human equivalent concentrations of 7-16 ppm.
Among the human studies, the report by Ruijten et al. (1991) demonstrates changes in trigeminal nerve function measured using the massiter reflex latency in workers whose exposure was below the threshold limit value (35 ppm) and whose exposure duration averaged 16 years. These findings correspond to reports from studies cited above of trigeminal and cranial nerve damage at shorter-duration exposure to high concentrations of trichloroethylene. In addition, this study noted slight reductions in the sensory nerve conduction velocity and the sensory refractory period of the sural nerve, consistent with preclinical peripheral nervous system impairments and with corresponding reports from experimental studies. Because all workers had been employed as printers in the same workplace and factory, the exposures were likely highly homogenous within this group. A human equivalent concentration LOAEL of 16 ppm was determined by EPA from this report using RfC methodology for a category 3 gas, extrathoracic effects (EPA 1994b).
Rasmussen et al. (1993) examined changes in cranial nerve function, motor coordination, and vibration sensitivity in metal degreasers whose primary exposure was to trichloroethylene (determined from biomonitoring data from the Danish Labour Inspection Service). Exposure durations were shorter than in the Ruijten et al. (1991) study. Highly significant dose-related increases were seen in motor dyscoordination. These findings are impressive because they were measured by clinical neurological examination, a far less
sensitive approach than is available with more sophisticated technologies (e.g., Weiss and Cory-Slechta 2001), although it is not clear whether the examiner was blind to the exposure categorization of each worker. Abnormal olfactory (cranial nerve) function was also dose related, with similar, but not significant, trends for trigeminal nerve sensory function and facial nerve function measured via taste. The human equivalent LOAEL determined by EPA from these studies was 7 ppm using RfC methodology for a category 3 gas, extrathoracic effects (EPA 1994b).
These reports are further supported by early studies reporting symptoms of drowsiness, fatigue, headaches, and nausea in response to occupational inhalation exposures to trichloroethylene over a mean of 7-8 years, with human pharmacokinetic adjusted LOAELs using RfC methodology for a category 3 gas, extrathoracic effects (EPA 1994b), of 7-11 ppm (Okawa and Bodner 1973; Vandervort and Polakoff 1973). Moreover, all four human studies demonstrated effects in the same exposure range as the report of decreased wakefulness by Arito et al. (1994) in rats after subchronic inhalation exposures (LOAEL of 50 ppm with human pharmacokinetic adjusted value of 9 ppm and a human pharmacokinetic adjusted BMD10 value of 5 ppm).
Three new human chronic exposure studies have appeared since the reviews by ATSDR and EPA. Two of them examine the impact of environmental trichloroethylene exposures on neurobehavioral function (Kilburn 2002a,b). Exposures were estimated based on groundwater plumes measured during a 3-month period. Concentrations of trichloroethylene measured in well water ranged from 0.2 to 10,000 parts per billion (ppb). Exposed subjects (n = 236) lived near two electronic manufacturing plants and were involved in litigation related to these exposures; referents matched on a number of other factors (n = 161) lived in a town without contaminated water located 88 km upwind from the exposed subjects. Additional reference subjects (n = 67) were from the same geographic area as the subjects but had never lived in the exposure zone. In the first of these studies, exposed subjects were reported to have delayed simple and choice reaction times; impaired balance; delayed blink reflex latency; abnormal color discrimination; and impaired cognitive function, attention, recall, and perceptual speed (Kilburn 2002a). This study has many limitations. Exposures are estimated, not directly measured, and involve mixed solvent exposures (although the author states that the primary toxicant was trichloroethylene), examiners did not appear to have been fully blinded to treatment conditions, and the period when trichloroethylene was measured was brief. Further, subjects were involved in a lawsuit related to this exposure, introducing the potential for bias. In addition, all relevant comparisons were not made (e.g., the two reference groups were never shown to be comparable), and the pattern
of differences between referents and subjects was not the same in the two groups (e.g., subjects versus local referent outcomes was not the same as subjects versus referents living 88 km distant). Thus, the reliability and utility of these findings is questionable and their relationship to exposure levels is unknown.
A second study published by the same author attempted to address the issue of the potential bias introduced by the subjects being involved in ongoing litigation (Kilburn 2002b). In this case, the 236 subjects were compared with 58 nonclaimants within the three residential areas in the exposure zone. In addition, subjects were divided into two groups based on duration of exposure (and, presumably, years of litigation as well). In addition to having the same study limitations noted above, other inconsistencies were noted. For example, subjects with shorter exposure durations to trichloroethylene had significantly abnormal sway (balance) relative to subjects with longer exposures, despite the fact that they were also 10 years younger. To examine the impact of litigation, subjects and referents were divided into three groups, each related to the area where they lived. By adopting this approach, the sample size, and thus the power to detect effects, was considerably diminished and therefore does not represent a true assessment of the impact of litigation. For example, comparisons in zone A involved 9 nonclients versus 100 clients; corresponding figures for zone B were 18 versus 16 and for zone C were 15 versus 11. The study reports mean values but not information on variability around the mean; in all other presentations, the standard deviations were shown. Thus, these two studies do not seem adequately suited to calculate trichloroethylene risk arising from chronic exposures.
A cross-sectional study of human environmental exposure was reported by Reif et al. (2003) based on residence in a community where the drinking water had been contaminated with trichloroethylene and related chemicals between 1981 and 1986 (Rocky Mountain Arsenal Superfund site). Tests of behavior, visual contrast sensitivity, and mood were carried out for estimated exposures of ≤5, >5-10, >10-15, and >15 ppb, with 5 ppb representing the maximum contaminant level for drinking water as defined by the EPA Office of Drinking Water. Testing occurred 6 years after peak concentrations of trichloroethylene. Subjects in the study (mean ages 48.6 to 55.8 years) resided in this area for a minimum of 2 years. In this analysis, trichloroethylene at >15 ppb affected visual function (contrast sensitivity) and increased scores for confusion, depression, and tension. For behavioral function, measured using the Neurobehavioral Core Test Battery, poorer performance on the digit symbol substitution test was reported. All these effects, however, were of marginal statistical significance. Further, these studies did not directly measure exposures but were based on estimates
from geographic information systems. It is also not clear to what extent the participants were aware of exposures; cleanup began in 1986. In addition, there is no information on out-migration of the population (e.g., affected individuals who may have moved from the area). The role of duration of exposure was not evaluated. Moreover, the wells were contaminated with other organic solvents, although trichloroethylene was stated to be the primary contaminant and was present at high concentrations. Collectively or individually, these limitations could increase or decrease the sensitivity of this study to detect effects. For these reasons, including these data in the trichloroethylene risk assessment should be considered cautiously.
One interesting aspect of this study, however, if reliability of the trichloroethylene exposure assessments is assumed, is the strong interactions that emerged between trichloroethylene exposure and alcohol consumption. In the group exposed to trichloroethylene at >15 ppb, the impairments in the digit span test (considered a measure of memory) were highly significant among individuals reporting alcohol use of at least one drink per month compared with no alcohol use. In addition, these individuals showed longer simple reaction time values. Deficits in memory and response time are also characteristic of higher concentration, shorter-duration exposures. Thus, alcohol appeared to exaggerate some of the behavioral effects of trichloroethylene. These findings of alcohol-trichloroethylene interactions (see also Chapter 10), while intriguing, nevertheless require caution in interpretation based on the study limitations noted above.
MODE OF ACTION
As the current literature indicates, trichloroethylene has a wide array of effects on the nervous system that may involve different mechanisms of action. For example, changes in learning and memory might be related to the impact of trichloroethylene on long-term potentiation, considered to be a neurophysiological basis of learning. Current evidence also shows that trichloroethylene affects various neurotransmitter systems. Alterations in susceptibility to chemoconvulsants after trichloroethylene exposure implicates the involvement of GABA(A) receptors (Shih et al. 2001). Studies also show effects of trichloroethylene on serotonin neurotransmitter systems (Gerlach et al. 1998; Lopreato et al. 2003). Dopaminergic consequences (e.g., Oshiro et al. 2004) likely contribute to motor deficits associated with trichloroethylene. Mechanisms of noncarcinogenic action for other target organs may likewise be operative in the brain, including oxidative stress. As a compound that can act on peroxisome proliferator-actived receptor α (PPARα), it is important to note the existence of such receptors in brain along with their functional roles as currently established (see Appendix E for some background information on PPARα agonism).
Trichloroethylene and Cognitive Function
The relationship between trichloroethylene exposure parameters and impairments of complex cognitive function remain unclear. White et al. (1997) described evidence in support of impaired cognitive function. In that study, neuropsychological testing was carried out in groups of exposed individuals from three different locations (Woburn, Massachusetts; Alpha, Ohio; and St. Paul-Minneapolis, Minnesota) and percentages of each group affected on different domains (relative to normative scores) were reported. Consistently affected across all three groups were attention and executive function and memory. Many details of this study were not reported (e.g., how subjects were recruited, awareness of exposure, litigation issues), nor were individual exposure measurements available for all subjects. Although White et al. (1997) stated that cognitive deficits were more pronounced the earlier in life the exposure occurred (developmental exposures are associated with more pronounced effects), they presented no data to support that assertion. The authors also noted that effects of these environmental (oral) exposures occurred at lower concentrations than anticipated from their experience with occupational cohorts; again, comparative data were not provided. Thus, while intriguing, it is difficult to determine the significance of the findings at the current time.
Studies reported in addition to those cited above (Oshiro et al. 2001, 2004) do little to clarify the question of the parameters of trichloroethylene exposure and impaired cognition. Isaacson et al. (1990) cited improvements rather than impairments in learning, here measured using a spatial learning paradigm in young male rats exposed to trichloroethylene in drinking water. Most improved were rats that had been exposed from day 21 to day 48 and again from day 63 to day 78 of age. Estimated intake of trichloroethylene in these studies averaged 5.5 mg/day for 28 days followed by 8.5 mg/day during the second exposure interval. Similarly, an unpublished study cited by Isaacson et al. (1990) apparently observed facilitation of learning in rats exposed to trichloroethylene during development.
Many noncognitive behavioral functions can indirectly influence measurement of learning (Cory-Slechta 1989). Alterations in motor function can alter the topography or effortfulness of responding. Sensory alterations can change the discriminability of environmental signals. Alterations in motivational state can influence either the salience or the potency of a reward. All these factors must also be controlled or explored in evaluating the outcome of learning paradigms. It is not possible to determine from the experimental description whether the facilitation noted by Isaacson et al. (1990) represents a true improvement in learning, or whether it is an indirect
consequence of changes in other behavioral domains (e.g., faster swim time and thus shorter delay to reward).
An example described above comes from a report by Waseem et al. (2001) that examined neurobehavioral effects of trichloroethylene in rats after oral administration for 90 days of trichloroethylene at 350, 700, or 1,400 ppm or inhalation exposure at 376 ppm for 4 hours per day, 5 days per week for a total of 180 days. The report cited a lack of effect of exposure on acquisition of a conditioned shock-avoidance response. However, this study did not control for the potential of trichloroethylene to alter levels of shock sensitivity per se, which would thereby influence the rate of acquisition of this response.
In summary, it is not yet possible to ascertain the extent of trichloroethylene-induced impairment of complex functions such as learning, memory and attention, preferential vulnerability to trichloroethylene across these domains, the exposure parameters that might be associated with any adverse effects, the extent of their reversibility, and the impact of developmental period of exposure on such effects.
Evidence to determine the extent to which trichloroethylene exposures during development or advanced age could enhance its adverse effects on the nervous system is limited. As noted above, White et al. (1997) reported more pronounced effects of environmental trichloroethylene exposures in younger humans, but they provided no data to support these statements. Experimental studies in which the effects of developmental exposures and adult trichloroethylene exposures are directly compared have not been reported. A study by Moser et al. (1999) of oral exposure to dichloroacetic acid, a metabolite that can be formed via mixed function oxidase metabolism of trichloroethylene, does include some comparisons of weanling versus adult rats. The results presented in the paper suggest comparable effects in the end points shown, although the authors noted that neuromuscular toxicity effects appeared to be somewhat greater in rats exposed as weanlings than in those exposed as adults. Generally speaking, there is insufficient evidence to ascertain whether there are developmental differences in sensitivity to trichloroethylene-induced neurotoxicity.
Aging does appear to enhance sensitivity to the adverse effects of trichloroethylene on the nervous system. A study by Arito et al. (1994) compared the responses of 2-, 13-, 20-, and 26-month-old rats to trichloroethylene at 300 ppm for 8 hours, followed by 1,000 ppm for 8 hours, after an intervening period of clean air for 7 days. In this study, the number of incidents of spontaneous bradyarrhythmia episodes during the 28-hour
period after cessation of exposure to trichloroethylene at 300 or 1,000 ppm compared with those occurring during the corresponding period of exposure to clean air was significantly greater in 20- and 26-month-old rats than in the 2- or 13-month-old rats. Measurements of trichloroethylene in brain and blood also revealed a prolonged half-life and delayed clearance with advancing age, leading the authors to posit that pharmacokinetic differences during aging may contribute to this enhanced sensitivity. Certainly, aging needs to be considered in the uncertainties associated with the risk assessment calculations.
The extent to which trichloroethylene neurotoxicity may be altered by coexposures with other environmental or dietary constitutents is not fully elaborated. One risk modifier for neurotoxicity described by Reif et al. (2003) is alcohol. As noted above, impairments in the digit span test in response to trichloroethylene were highly significant among individuals reporting alcohol consumption of at least one drink per month, whereas no effects were observed in individuals reporting no alcohol consumption. In addition, individuals who consumed alcohol also showed longer simple reaction time values in response to trichloroethylene exposure. Thus, alcohol was noted to potentiate the effects of trichloroethylene on a measure of attention and memory. As noted previously, however, limitations of this study necessitate caution in interpreting the validity of these findings.
Significance of Brain Concentrations of Trichloroethylene
Boyes et al. (2000, 2003) explored the relationship between exposure concentration and duration under conditions of acute exposure in predicting risk of trichloroethylene neurotoxicity. The first of these studies (Boyes et al. 2000), examining hearing loss, signal detection behavior, and visual function, demonstrated that Haber’s law would overestimate when extrapolating from shorter to longer trichloroethylene exposure durations and would underestimate when extrapolating from longer to shorter exposures. This study also showed that, instead, the estimated peak blood concentration of trichloroethylene at the time of testing accurately predicted the magnitude of effect on visual function and on signal detection (these neurotoxic effects reflect momentary tissue concentrations of trichloroethylene). A second study (Boyes et al. 2003) used visual evoked potentials as an outcome measure and demonstrated that the exposure metric of area under the curve was not an accurate predictor of effect; instead, brain concentration of trichloroethylene at the time of visual evoked potential testing predicted the effects
of trichloroethylene across exposure concentration-duration parameters. Still questionable, however, is the extent to which this relationship would generalize to longer-term exposures.
Some reports have suggested a link between trichloroethylene exposure and Parkinson’s disease. Among these are two case reports. Guehl et al. (1999) described a case of Parkinson’s disease in a 47-year-old woman who had 7 years of exposure to trichloroethylene. This case is notable given the young age relative to typical onset and the fact that the subject was a woman, because time to onset is longer and Parkinson’s disease incidence is lower in women than in men. Unknown, however, is the specificity of the exposure and the genetic background of the subjects. In another report, Kochen et al. (2003) cited the onset of Parkinson’s disease in three individuals chronically exposed to trichloroethylene during the postexposure period. Following the case report observation, Guehl et al. (1999) also described the loss of dopamine neurons in substantia nigra pars compacta (the hallmark of Parkinson’s disease) after intraperitoneal injections of trichloroethylene at 400 mg/kg per day, 5 days a week for 4 weeks, to mice. No follow-up studies to this report have been described.
Although it is not clear whether assessment of the incidence of Parkinson’s disease has been examined in trichloroethylene-exposed populations, a biological basis for its potential contribution to this disease has been suggested by Riederer et al. (2002) to be based on the formation of TaClo (1-trichloromethyl-1,2,3,4-tetrahydro-β-carboline), a potent dopaminergic neurotoxin that can be formed endogenously after exposure to the sedative chloral hydrate or after exposure to trichloroethylene. As the authors note, trichloroethylene has an estimated half-life in humans in venous blood of 21.7 hours, sufficient for the appropriate in vivo condensation reactions that would be involved in TaClo formation. Indeed, significant amounts of TaClo (approximately 200 ng per 10 mL samples) were detected in both serum and clot in Parkinson’s disease patients that had been treated for several days with 500 mg of chloral hydrate (a metabolite of trichloroethylene), with blood sampled on the final day of treatment (Bringmann et al. 1999).
Additionally, there are significant structural similarities between TaClo and MPP+ (1-methyl-4-phenylpyridinium ion), the most widely used experimental model for Parkinson’s disease. As noted in the review by Riederer et al. (2002), like MPP+, TaClo specifically inhibits the electron transfer from complex I to ubiquinone in the mitochondrial respiratory chain in both rat brain homogenate and rat liver submitochondrial particle preparations. TaClo was shown to reduce dopamine uptake and the number and size of tyrosine-hydroxylase-positive cells in C56/BL6 mouse primary cell cultures.
Injection of TaClo directly into the rat substantia nigra pars compacta decreases both neuronal density and number of neurons in this region. This treatment also resulted in a progressive decline in concentrations of the dopamine metabolite 3,4-dihydroxyphenylacetic acid over 6 weeks after a single injection, consistent with some other models of the Parkinson’s disease phenotype. While an intriguing series of studies, it is unknown whether TaClo can be formed following trichloroethylene exposure per se.
FINDINGS AND RECOMMENDATIONS
With respect to the EPA (2001b) draft risk assessment, several neurotoxicity studies contributed to the derivation of an inhalation RfC. In general, these studies report effects in humans and in experimental models (rat) at very similar concentrations. In addition, common effects are seen across these studies, and the nature of the effects described are comparable to or consistent with those reported in response to acute exposures to higher concentrations. Those studies utilized to derive the RfC include reports in humans of changes in trigeminal nerve function (measured using the massiter reflex latency) and motor incoordination at human equivalent LOAEL concentrations of 7-16 ppm (Ruijten et al. 1991; Rasmussen et al. 1993) and symptoms including nausea, drowsiness, and fatigue (Okawa and Bodner 1973; Vandervort and Polakoff 1973). Studies in rats showed changes in heart rate and wakefulness at a human pharmacokinetic adjusted LOAEL of 9 ppm (Arito et al. 1994). This appears to be a valid and standard approach taken to evaluate risk.
Furthermore, as is clear from the discussions above, new information on trichloroethylene published since the EPA (2001b) review is limited and thus may offer little in the way of amendment to the current RfC:
The effects Ohta et al. (2001) described in mice after single intraperitoneal injections on long-term potentiation in hippocampal slices may be significant to any derivation of acute-exposure risk assessment, particularly considering the critical nature of the effect and its implications for complex cognitive function. However, the intraperitoneal route of administration makes extrapolation of these findings to humans difficult.
Although the exposure concentrations (0.2-200 ppm) used in the subchronic study of rats by Poon et al. (2002) are low and thus would be of interest, small sample sizes used in components of this study may have precluded the ability to detect effects; histologic changes are hard to interpret given measurement at a single time point. Thus, the absence of effects reported in this study may reflect experimental inadequacies rather than representing actual no-observed-adverse-effect levels.
The two studies by Kilburn (2002a,b) of chronic environmental
exposures in humans have major limitations, including potential for bias, inconsistency of effects, absence of appropriate comparisons, and others as noted above. For these reasons, it is not clear that the studies can be used in the trichloroethylene risk assessment.
A study by Reif et al. (2003) reporting that low (and high) concentrations of trichloroethylene affect memory, as well as a potentiation of such effects by alcohol (an interaction supported by the experimental literature) has limitations that include potential misclassification of exposures and bias, and thus must be interpreted with caution.
Certainly of potential relevance to risk assessment are studies that suggest protracted effects of trichloroethylene after cessation of exposure, such as described by Oshiro et al. (2004) for dopaminergic systems and by Haglid et al. (1981) for alterations in protein levels in multiple brain regions.
One thing made clear by any assessment of the trichloroethylene literature as it relates to the nervous system is the paucity of data available to define the extent of its neurotoxicity and the parameters and conditions of exposure under which it occurs. For example, studies of chronic exposure are limited. The realities of evaluating the impact of human environmental exposures generally mean that measures of trichloroethylene exposures in such studies are estimated and not empirically determined, leaving open the possibility of misclassification, a problem that can increase or decrease the probability of detecting effects. In some other human studies, the effects are confounded by involvement of the subjects in litigation. When experimental studies were carried out, they were generally not lifetime studies and the extent to which they evaluated behavioral and neurological function was limited, because most of the emphasis was on carcinogenicity.
Long-term studies, human and experimental, are critically needed to evaluate the effect of trichloroethylene on the central nervous system. For human studies, measurement or better estimates of exposure are necessary.
Another gap in the literature is the extent to which development represents a period of enhanced susceptibility to the neurotoxic effects of trichloroethylene. The only comparative data available at the current time appears to be from the study of Moser et al. (1999) examining one metabolite of trichloroethylene, dichloroacetic acid, where the evidence in support of enhanced susceptibility of younger rats is limited. Statements of greater sensitivity of children than of adults exposed to trichloroethylene
are also described by White et al. (1997), but again the supporting data are not presented.
The one study published to date clearly demonstrates an enhanced sensitivity of aging rats relative to young rats to the effects of trichloroethylene, measured in that study as changes in heart rate (Arito et al. 1994). These changes were shown to be due to changes in pharmacokinetics with age, with older rats exhibiting higher brain concentrations of trichloroethylene as well as longer exposure (delayed clearance) to those doses. One would predict that this enhanced toxicity should generalize to other behavioral and neurological consequences of trichloroethylene as well, because functionally it represents a higher dose to the brain, particularly if peak blood trichloroethylene concentrations are critical to adverse effects (Boyes et al. 2003). This also means that aging and exposure duration may be related in chronic exposure scenarios in humans.
More research is needed to assess different life stages at which humans might be more susceptible to the neurotoxic effects of trichloroethylene.
Another area of interest is the possibility of permanency versus reversibility of effects and the conditions under which this could occur. To date, the evidence is conflicting and undoubtedly would reflect the parameters of exposure, but some studies document protracted effects of trichloroethylene on the nervous system (e.g., Haglid et al. 1981; Oshiro et al. 2004; Crofton and Zhao 1993). It is clear from studies reported many years ago that acute exposures to high concentrations of trichloroethylene, in occupational or experimental contexts, can produce permanent changes in the nervous system. What is not yet known is the exposure conditions, particularly repeated exposures, under which effects would no longer be reversible. A related issue is whether effects of exposure can be progressive even after that exposure has terminated.
Additional research examining the extent to which observed effects are permanent versus reversible would be of relevance to risk evaluation.
Despite the associations of occupational exposures with memory loss and other cognitive deficits, the nature of such effects and the exposure conditions with which they can be associated have not been elaborated. A report by White et al. (1997), despite its deficiencies, clearly shows common effects on complex cognitive functions across three different populations
exposed to trichloroethylene environmentally. Experimental studies have been less clear about such effects, but the extent to which this has been addressed is limited and, in some published reports, is not interpretable with respect to outcome.
A related function that clearly seems to be affected by trichloroethylene is motor function, as has been demonstrated in experimental studies as well as in occupational cohorts. As with other behavioral functions, the trichloroethylene exposure conditions under which such effects occur are not yet known. It may be important to define such conditions, particularly if, as suggested by other reports, trichloroethylene might contribute to neurodegenerative disorders such as Parkinson’s disease. The earliest signs of motor dysfunction could serve as biomarkers of such a contribution.
Many neurological and behavioral disorders represent complex multifactorial etiologies. Given the broad spectrum of its effects across behavioral domains as well as neurotransmitter systems (and other as yet unknown mechanisms), it is possible that trichloroethylene may contribute as a risk factor to other neurodegenerative and behavioral diseases or dysfunctions, acting in conjunction with other risk modifiers that may include genetic background (P-450 polymorphisms) and lifestyle factors (e.g., alcohol consumption), aging, and other factors that are undetermined.
Studies of additional functional end points, including cognitive deficits and motor and sensory function in response to chronic exposures to trichloroethylene would be of value to risk assessment.
Current evidence suggests the possibility of multiple mechanisms by which trichloroethylene may act, with the recognition that these mechanisms may also depend on the parameters of exposure. Extant literature already documents changes in long-term potentiation as well as alterations in functions of several neurotransmitter systems, a basis from which complex cognitive functions as well as other behavioral domains could be impaired.
Additional research is required to elucidate the underlying modes of action of trichloroethylene-induced neurotoxicity.