Trichloroethylene and Ocular Malformations: Analysis of Extant Literature

Abstract

An evaluation of the scientific literature for trichloroethylene (TCE) identified two reports of ocular defects, specifically microphthalmia/anophthalmia, following prenatal TCE exposure in rats. Herein, these reports are analyzed in detail and interpreted within the context of other developmental TCE exposure studies. The ocular findings following prenatal TCE exposure are reported in studies that were not designed specifically for developmental safety assessment, and thus did not use standard experimental practices. Furthermore, these findings most commonly occurred at TCE doses associated with considerable maternal toxicity. Among the 18 published studies using developmental TCE exposures, only 3 used doses sufficiently high enough to result in maternal toxicity, and of these, only the 2 discussed in detail in this paper demonstrated ocular defects in the offspring. Furthermore, statistically significant effects were only observed at doses that were above the currently accepted guideline limit dose of 1000 mg/kg body weight. All other TCE developmental exposure studies failed to demonstrate ocular defects as a result of prenatal exposure. These results suggest that the micro-/anophthalmia findings were likely a consequence of delivery of an extremely high bolus TCE dose that is irrelevant to human environmental exposure scenarios.

Trichloroethylene (TCE), a halogenated hydrocarbon, is used predominantly as a metal degreasing agent, although it is also a chemical intermediate in the production of fluorochemicals and polyvinyl chloride. Over the past seven decades, TCE has had a wide variety of applications, including use as an anesthetic, an antiseptic, an extractant for spices, a solvent for decaffeinating coffee, and a dry-cleaning agent (Steinberg and DeSesso 1993). In humans, TCE undergoes oxidative biotransformation to chloral hydrate, a drug that is often prescribed to treat insomnia in elderly patients or to sedate young children undergoing ophthalmologic procedures or computer-assisted tomography (CAT) scans (Steinberg 1993). Complete oxidative biotransformation of TCE leads to formation of trichloroacetic acid (TCA) and dichloroacetic acid (DCA) (Goeptar et al. 1995). Due to its volatility, most TCE released into the environment evaporates. When TCE is found in groundwater, however, it has limited contact with air and can persist for years as a contaminant. It is TCE’s status as a widespread, persistent groundwater contaminant that has led to its evaluation as a potential toxicant, especially in the areas of carcinogenicity and developmental toxicity.

Many published studies have examined the potential association between TCE exposure in pregnant experimental animals and birth defects or other developmental toxicity in the offspring (Schwetz, Leong, and Gehring 1975; Dorfmueller et al. 1979; Hardin et al. 1981; Healy, Poole, and Hopper 1982; Manson et al. 1984; NTP 1985, 1986; Taylor et al. 1985; Noland-Gerbec et al. 1986; Isaacson and Taylor 1989; Dawson et al. 1990, 1993; Cosby and Dukelow 1992; Narotsky and Kavlock 1995; Narotsky et al. 1995; Carney et al. 2001, 2006; Fisher et al. 2001; Johnson et al. 2003; Warren et al. 2006). The vast majority of these papers indicate that maternal TCE exposure is not associated with developmental toxicity in mammals. Nevertheless, concerns have been raised regarding the possibility of TCE-associated cardiac teratogenicity. These concerns are based primarily on the research of one investigatory group suggesting possible congenital heart defects (Dawson et al. 1990, 1993; Johnson et al. 2003). More recent detailed analyses of both the epidemiologic and scientific literature (Hardin, Kelman, and Brent 2005; Watson et al. 2006), as well as a blinded study that was unable to replicate the results of this investigatory group (Fisher et al. 2001) have largely cast doubt on the validity of earlier reports and suggest that TCE exposure during pregnancy does not cause cardiac malformations in mammalian offspring.

More recently, the possibility that in utero TCE exposure may result in ocular defects has been raised. These concerns are based primarily on two published reports from a single laboratory (Narotsky and Kavlock 1995; Narotsky et al. 1995). In these studies, high-dose TCE exposures to pregnant animals were associated with microphthalmia (small eyes) and/or anophthalmia (no eyes) in the offspring. In humans, large-scale epidemiology studies on this topic have been negative (Bove, Shim, and Zeitz 2002; Shaw et al. 1992), although a retrospective observational study reported ocular-related defects in a small population in Woburn, MA, that had been exposed to a mixture of organic solvents, including TCE (Lagakos, Wessen, and Zelen 1986).

The purpose of the present report is to investigate whether TCE exposures during pregnancy may put human offspring at risk for developmental ocular defects. The two experimental animal studies (Narotsky and Kavlock 1995; Narotsky et al. 1995) reporting such findings are evaluated in detail and assessed in the context of other research involving in utero TCE exposures.

EVALUATION OF THE ANIMAL DATA

Two experimental animal studies have reported ocular defects in the offspring as a result of TCE exposure in pregnant animals (Narotsky and Kavlock 1995; Narotsky et al. 1995). In order to properly evaluate these studies, it is important to understand the context in which they were conducted. The safety assessment of new chemicals can be a long and laborious process, requiring testing in numerous toxicity assays. The above-mentioned two studies evaluated potential modifications to the standard toxicity screening assays that would allow for simultaneous assessment of developmental and adult endpoints in one assay (Narotsky and Kavlock 1995) and toxicity assessment of chemical mixtures consisting of three or more component chemicals (Narotsky et al. 1995). Therefore, although these studies evaluated developmental toxicity end points, they did not adhere to standard developmental testing procedures and were not meant to stand as definitive developmental toxicity assessments of TCE or any other chemical.

In order to determine how the data from these two studies may influence a human health risk assessment for TCE as it applies to prenatal development, each study is described in detail below. The intended purpose of each study is reported, the experimental methods described, and the overall results as they relate to in utero TCE exposures are provided.

Narotsky and Kavlock 1995

The purpose of the study by Narotsky and Kavlock (1995) was to evaluate whether modification of standard in-life animal screening assays could allow for the incorporation of developmental toxicity end points. Such a modification, if properly validated, could result in a simplified toxicity testing battery that would reduce the number of animals used in testing, diminish the required amount of technical labor, and reduce the time to market for new chemicals.

In keeping with the stated purpose of the study, the experiment did not follow standard developmental toxicity testing procedures. Rather, pregnant rats were exposed to test materials during the organogenesis stage of gestation, allowed to deliver, and pups were examined through postnatal day 6 (PND 6). A total of 10 chemicals was evaluated using the modified testing assay: carbaryl, triadimefon, chlordane, heptachlor, dichloromethane, carbon tetrachloride, TCE, tetrachloroethylene, di(2-ethylhexyl)phthalate, and phenol. The authors selected these compounds based on known systemic, neurological, and/or developmental effects as well as their prevalence in occupational settings or as environmental contaminants. The present analysis focuses solely on the results obtained following TCE administration.

Study Design

Mated female Fischer 344 rats were assigned to each dose group in such a way as to assure homogeneous distribution of body weights at the beginning of study. The day of mating was designated gestational day 0 (GD 0). Animals were treated by oral gavage on GD 6 to 15 with corn oil vehicle, 1125 mg TCE/kg/day (low dose), or 1500 mg TCE/kg/day (high dose) in a dose volume of 2 ml/kg administered once per day. Individual animal doses were based on GD 6 body weights and remained constant throughout treatment. Therefore, as animals gained weight during pregnancy, the mg/kg dose of TCE delivered per animal each day ordinarily would have gone down during gestation. However, because animals in the TCE treatment groups actually lost weight for much of gestation, this was not the case. The TCE high dose was selected based on the results of a 14-day repeat-dosing study conducted in nonpregnant female rats (Berman et al. 1995). It was anticipated that this dose would cause some overt maternal toxicity. The low dose was set at 75% of the high dose. In retrospect, the magnitude of the TCE doses and the spacing of doses do not comply with the U.S. Environmental Protection Agency’s health effects test guidelines for prenatal developmental toxicity studies. These guidelines recommend at least three doses of test agent, and two- to four-fold intervals for the spacing of such doses, which should not exceed 1000 mg/kg/day unless indicated based on human exposure data (US EPA 1998). Because this experiment was not meant to assess the developmental toxicity potential of TCE, however, the dosing methods used are deemed appropriate.

Maternal body weights were determined on GD 6, 8, 10, 13, 16, and 20, and dams were examined throughout study for clinical signs of toxicity. GD 22 was defined as postnatal day 1 (PND 1), regardless of whether animals delivered. Pups were examined and counted on PND 1, 3, and 6; pups in each litter were weighed collectively on PND 1 and 6. On PND 6, dams were sacrificed and uterine implantation sites were counted. Those females that did not deliver by GD 24 (PND 3) were sacrificed to assess pregnancy status. Staining with ammonium sulfide was used to detect potential cases of early full litter resorption.

Study Results

During the early treatment period (GDs 6 to 8), TCE caused marked maternal weight loss of approximately 8 and 14 g/animal in the low- and high-dose groups, respectively. These losses are in contrast to the approximate 6 g weight gain among control animals for the same time period. It is not known how animals fared at the other interim weighing times, as these data were not reported. However, the TCE-treated groups gained significantly less weight over GD 6 to 20 compared to controls: net maternal body weight gains (i.e., weights adjusted for live PND 1 litter weights) of 10 and 2 g in the low- and high-dose groups, respectively, versus a 17.5 g net body weight gain in the controls. Due to the decrease in body weights experienced by animals in both TCE treatment groups during the first days of treatment, these animals would have actually received greater than the target doses of 1125 and 1500 mg TCE/kg/day at the beginning of dosing, including during the major period of organogenesis for ocular structures in the rat, which occurs during GD 10 to 12.5 (DeSesso 2006).

Marked maternal weight loss was accompanied with dose-related transient ataxia and decreased motor activity. Rales and dyspnea were reported in high-dose animals and 2 of 17 females in this group died before the end of study. Reproductive data from this experiment are presented in Table 1. The high number of full litter resorptions in the TCE treatment groups (Table 1; 46.7% and 84.6% in the low- and high-dose groups, respectively) should be noted. Ammonium sulfide staining indicated that these resorptions occurred early in pregnancy and were likely associated with high maternal toxicity, which would not be expected to occur in humans as a result of typical environmental TCE exposures. The high dose caused greater than ten percent maternal mortality (2/17; 11.8%) and obvious maternal toxicity (11/13 full litter resorptions, rales, and dyspnea). The low dose caused 7/15 full litter resorptions. Guidelines state that the highest dose in a prenatal developmental toxicity test should not induce significant death or severe suffering; specifically, this dose should not cause greater than 10% maternal mortality (US EPA, 1998). Thus, the maternal mortality results, in combination with the clinical findings and reduced maternal weight gains, again suggest that the TCE doses selected for study were inappropriately high for developmental toxicity assessment.

In addition to effects on the dams, TCE treatment produced both a statistically significant, dose-related reduction in the number of live PND 1 pups (Table 1) and a dose-related decrease in surviving PND 1 pup weights (exact weights were not provided in the study report; rather, the data were depicted graphically). The authors also reported an unspecified incidence of micro-/anophthalmia; however, no numerical data were presented. From the methods description, it is not known how these malformations were confirmed. Because the size of the head is correlated with body weight and because pup weights were decreased by treatment, it is plausible that microphthalmia could have been misidentified in pups that were smaller than normal. Because individual pup weights were not recorded, this cannot be determined for the pups from this study; however, the total numbers of pups in the TCE treatment groups were very low (30 and 2 in the low- and high-dose groups, respectively, versus 146 in the control group; calculated from the mean number of pups per litter provided in Table 1), which makes direct comparison of the group means impractical.

Narotsky et al. 1995

This paper describes an experimental paradigm to test exposures to mixtures of chemicals. Guidelines are not available for the testing of chemical mixtures, although such exposures are typically experienced in the real world. While health effects associated with binary mixtures (mixtures of two chemicals) have been studied to some degree, research using tertiary or greater mixtures (mixtures of three or more chemicals) has been limited. Therefore, as a first step in designing possible future testing methods for the study of such mixtures, the investigators conducted a modified in-life developmental toxicity screen using TCE, di(2-ethylhexyl)phthalate (DEHP) and heptachlor, each administered at five different dose levels (5 × 5 × 5 design).

Study Design

Experimental methods were similar to those of the previously described paper (Narotsky and Kavlock 1995). Preliminary experiments were first conducted using each chemical separately to establish doses for the 5 × 5 × 5 mixtures study. Doses for the range-finding experiments were selected based on results of the previous study (Narotsky and Kavlock 1995). Mated female Fischer 344 rats were assigned to each of five different treatment groups (N = 7–12 animals per group). The day of mating was designated GD 0. Individual animal doses were based on GD 6 body weights and remained constant throughout treatment. In the preliminary study, rats were treated by oral gavage once per day on GD 6 to 15 with either corn oil vehicle or 475, 633, 844, or 1125 mg TCE/kg/day administered in a dose volume of 2 ml/kg. Based on results of the preliminary study, TCE doses of 0, 10.1, 32, 101, and 320 mg/kg/day were used in the definitive study (N = 9 to 15 animals per group).

Maternal body weights were determined on GD 6, 8, 10, 13, 16, and 20, and dams were examined throughout study for clinical signs of toxicity. GD 22 was defined as PND 1, regardless of when animals had delivered. Pups were examined, counted, sexed, and weighed on PND 1 and 6. These measurements were taken collectively in the preliminary study and individually in the definitive study. The incidence of micro-/anophthalmia was determined based on the presence of a reduced ocular bulge or the complete absence of such a bulge. After PND 6, pups were sacrificed and all eye defects were confirmed by dissection. Dams were killed and uterine implantation sites counted. Those females that did not deliver were assessed for pregnancy status via ammonium sulfide staining of uteri to detect cases of early resorption.

Study Results

TCE doses used in the definitive 5 × 5 × 5 mixtures study (0, 10.1, 32, 101, and 320 mg/kg/day) are all below the lowest TCE dose of 475 mg/kg/day used in the preliminary experiment; therefore, these results are presented first. The data from the definitive experiments were represented graphically rather than in a table, making it difficult to decipher; however, the results seem to indicate that, when TCE was administered alone (without heptachlor or DEHP) in the 5 × 5 × 5 study at doses up to 320 mg/kg/day, no maternal deaths or full litter resorptions were observed. Mean maternal weight gain on GD 6 to 8 was reduced in a dose-dependent manner, however, as were mean PND 1 pup weights. Micro-/anophthalmia incidence data in the definitive study were presented graphically, making it difficult to determine incidence rates in the TCE-only treatment groups. Based on data presented in Barton and Das (1996), however, non-significant increased total pup incidences of 4.41% and 3.66% were found in the 101 and 320 mg/kg/day treatment groups, respectively (note: these are total incidence rates and not per litter incidence rates).

Reproductive data from the preliminary experiment using TCE administered alone at doses of 475 to 1125 mg/kg/day were presented in the study report, as shown in Table 2. Transient ataxia associated with dosing was reported in animals receiving 633 mg TCE/kg/day or greater, but no other clinical signs of toxicity were reported. Two of 13 females (15.4%) in the 1125 mg TCE/kg/day treatment group were reported to have died on study, meaning that 11 dams survived; however, data from only 10 dams were reported. The status of the remaining dam or origin of the discrepancy in reported findings is not known. Five of the reported surviving dams administered 1125 mg TCE/kg/day experienced full litter resorptions. Full litter resorptions were also noted at 475 and 844 mg TCE/kg/day at rates equivalent to one full litter resorption per treatment group. Mean maternal weight gains for the GD 6 to 8 and 6 to 20 time periods were affected in a dose-dependent manner (Table 2). Because doses were based on GD 6 weights and did not change throughout the study, dams in the higher-dose treatment groups received greater than the intended target doses. This view, however, is based on an assumption that the authors expected maternal weight gain to be similar across groups, regardless of treatment.

Interestingly, maternal toxicity in the 1125 mg TCE/kg/day treatment group in this experiment (Narotsky et al. 1995) is considerably higher than that seen at the same TCE dose level in the previous study (Narotsky and Kavlock 1995). For example, maternal mortality was 15.4% at this dose in Narotsky et al. (1995), but no deaths were reported at the same dose in the previous study (Narotsky and Kavlock 1995). Additionally, weight loss over GDs 6 to 8 was greater (11.7 compared to 8 g) and the mean net body weight gain for GD 6 to 20 was smaller (1.4 versus approximately 10 g). The reason for the increased maternal toxicity at 1125 mg TCE/kg/day is not known. These results, however, in combination with the extremely high rate of full litter resorptions, again indicate that 1125 mg TCE/kg/day is an inappropriately high dose for assessing the risk for developmental toxicity.

Developmental data from the preliminary experiment using TCE administered alone were presented in the study report, as shown in Table 3. A statistically significant increase in postnatal pup loss was seen at 475 mg TCE/kg/day, but this effect was not observed at higher doses, and therefore may be a chance finding. Although not statistically different from control, mean PND 1 pup weights appear to decrease in a dose-dependent manner. Similarly, the mean percentage of pups per litter exhibiting eye defects increased with dose, and a statistically significant 30.0 % incidence of such defects was reported for pups in the high dose group of 1125 mg TCE/kg/day (Table 3). These incidence data were also presented as raw numbers in an analysis by Barton and Das (1996), as shown in Table 4.

With regard to the report of TCE-induced micro-/ anophthalmia, the methods by which these findings were diagnosed and confirmed are detailed in the methods section of the study report and appear to be adequate. Nevertheless, it is difficult to draw firm conclusions concerning these findings for several reasons. Microphthalmia and anophthalmia were grouped together for reporting purposes, so the separate incidences of each are not known. Nor is the defect incidence per litter reported; thus, the possibility of a litter effect cannot be assessed. However, the relatively high standard error relative to the mean value reported for the high dose group (30.0 ± 10.8) suggests that the incidence of eye defects may not be uniformly increased across litters, and may instead be clustered within a limited subset.

Historically, the incidence of micro-/anophthalmia has been low in control rats. For example, based on historical control data from WIL Research Laboratories for 1982 to 1997, the incidence of micro-/anophthalmia in Sprague Dawley rats has been 31/50,858 total fetuses (29/3926 total litters) (WIL Research Laboratories, personal communication). In this study (Narotsky et al. 1995), however, a case of micro-/anophthalmia was reported in the control group for an incidence rate of 1/197. This finding could be purely coincidental. Alternatively, the investigators may have had a low threshold for calling microphthalmia. Based on the data shown in Table 4, the incidences of micro-/anophthalmia are elevated at doses of ≥ 101 mg/kg/day and increase substantially at TCE doses of ≥475 mg/kg/day. Maternal toxicity is also noted at these higher doses (≥475 mg/kg/day), but not at the lower doses used in the definitive study. Furthermore, no statistically significant difference in incidence of eye defects is noted until a dose of 1125 mg/kg/day is administered. At this dose, however, the incidence of maternal toxicity is extremely high, which makes interpretation of the findings difficult.

Based on the foregoing analysis, the highest TCE doses selected for these studies were inappropriate for assessing the risk of developmental toxicity because they resulted in extreme maternal toxicity. As well, the total numbers of pups in the high dose groups were significantly lower than in controls, which might have contributed to artificially increased incidence percentages. It appears that, in these two studies, a statistically increased incidence of micro-/anophthalmia is only associated with the administration of high bolus doses of TCE that generally cause maternal toxicity. Such doses are apt to be irrelevant to understanding the possible outcomes associated with anticipated human TCE exposure scenarios. Furthermore, although increased incidences of micro-/anophthalmia were noted at lower doses (101 to 320 mg TCE/kg/day), these increases did not reach statistical significance.

EVALUATION IN CONTEXT OF OTHER TCE DEVELOPMENTAL TOXICITY STUDIES

In order to better understand the findings of the above described two studies (Narotsky and Kavlock 1995; Narotsky et al. 1995), it is useful to examine them in the context of other research investigating the developmental effects of in utero TCE treatment. Table 5 lists such studies to date (Schwetz, Leong, and Gehring 1975; Dorfmueller et al. 1979; Hardin et al. 1981; Healy, Poole, and Hopper 1982; Manson et al. 1984; NTP 1985, 1986; Taylor et al. 1985; Noland-Gerbec et al. 1986; Isaacson and Taylor 1989; Dawson et al. 1990, 1993; Cosby and Dukelow 1992; Narotsky and Kavlock 1995; Narotsky et al. 1995; Carney et al. 2001, 2006; Fisher et al. 2001; Johnson et al. 2003; Warren et al. 2006). Of the 18 studies listed, 6 were conducted via oral gavage, 5 administered TCE in the drinking water, 1 used intrauterine administration, 5 used inhalation, and 2 were dietary administration studies. Only the two studies described above reported micro-/anophthalmia in response to TCE treatment. It should be noted that the study by Warren et al. (2006) conducted specific, detailed examination of the eyes following prenatal TCE exposure and did not find evidence of ocular defects in the resulting offspring.

The drinking water, inhalation, and feed studies used TCE concentrations as high as 1250 mg TCE/L, 1800 ppm TCE, and 0.60% (6000 mg TCE/kg feed), respectively (Table 5). Because of factors affecting TCE absorption and distribution, however, none of these studies likely could achieve the elevated internal dose levels anticipated following high-dose oral gavage and intrauterine administration experiments. If it is assumed that the findings of micro-/anophthalmia are likely due to the administration of extremely high bolus doses of TCE, it is not surprising then that eye defects were not observed in the drinking water, inhalation, and feed studies listed in Table 5.

Focusing solely on the gavage and intrauterine administration studies, we find that only the above described two studies (Narotsky and Kavlock 1995; Narotsky et al. 1995) and another by Manson et al. (1984) demonstrated maternal toxicity in response to TCE treatment. In Manson et al. (1984), up to 1000 mg TCE/kg/day was administered by gavage to female Long-Evans hooded rats starting 2 weeks premating and continuing throughout pregnancy. Three maternal deaths were noted out of 23 animals treated at the high dose (one during the premating period and two during pregnancy). However, only 1/20 full litter resorptions were reported at the high dose in Manson et al. (1984), compared to rates of 11/13 and 5/10 at the highest doses in the other two studies (Narotsky and Kavlock 1995; Narotsky et al. 1995), respectively. Thus, the degree of maternal toxicity experienced in Manson et al. (1984) appears to be substantially less than that in the above-described studies (Narotsky and Kavlock 1995; Narotsky et al. 1995). The reason for the discrepancy among the three studies is not known; however, rat strain differences may play a role. Although the primary focus of Manson et al. (1984) was evaluation of TCE exposure on female reproductive function, the study authors specifically noted that no major malformations were observed upon gross external examination of the pups. The methods for this evaluation, however, are not provided.

The reason for the lack of maternal toxicity in the intrauterine study by Dawson et al. (1990) likely is not related to applied dose. Animals were administered 1500 ppm TCE in the high-dose group in this study; however, because the TCE was delivered directly into the uterine lumen (and not systemically), little of it was probably absorbed into the maternal blood stream. Therefore, maternal toxicity likely did not result.

In comparison to the studies discussed above, the gavage studies by Warren et al. (2006) and Cosby and Dukelow (1992) tested doses up to 500 and 240.2 mg TCE/kg/day, respectively, but did not demonstrate maternal toxicity (Table 5). The study by Warren et al. (2006) deserves special attention because it was specifically designed to assess the effects of gavage treatment during organogenesis (GD 6 to 15) on ocular development. Doses of 500 mg TCE/kg/day, 300 mg TCA/kg/day, 300 mg DCA/kg/day, and 15 mg retinoic acid (RA)/kg/day were administered and the effect on a number of ocular measurements, including lens area, globe area, medial canthus distance, and interocular distance, were measured. This study was conducted on a subset of fetuses examined for cardiac malformations in Fisher et al. (2001). Heads of GD 21 fetuses were fixed in Bouin’s solution and examined for gross external abnormalities. A dissection microscope and Leica Quantimet 570C Image Analysis System were used to measure the distance between the medial canthi of the eyes. Heads were then hand-sectioned using methods designed to capture sections through the center of each ocular globe. Sections were transferred to microscopic slides and digital images taken for further assessments. From the digital images, interocular distance, total area of the cut surface, areas of left and right lenses, and areas of left and right globes were measured.

No findings of micro-/anophthalmia or exencephaly were reported for the TCE- or TCA-treated groups. In the DCA-and RA-treated groups, 1% and 25.9% of fetuses, respectively, were determined to have micro-/anophthalmia. Also, 39.2% of RA-treated fetuses were exencephalic. A total of 1185 (71%) fetuses from all groups were selected for ocular examination, of which only 574 (48%) were deemed to have appropriate sections through the middle of the ocular globes. While reduced ocular measurements were found as a result of treatment with TCA (nonsignificant reductions in all four ocular measurements), DCA (statistically significant decreases in lens area, globe area, and interocular distance), and RA (statistically significant decreases in lens and globe areas), no significant ocular changes were observed in fetuses from dams treated with 500 mg TCE/kg/day.

Why the results of Warren et al. (2006) should differ significantly from the findings of micro-/anophthalmia noted in the previously discussed two studies (Narotsky and Kavlock 1995; Narotsky et al. 1995) is not known. Certainly, the latter study (Narotsky et al., 1995) finds a nonsignificant increased incidence of micro-/anophthalmia at TCE doses similar to that examined by Warren et al. (2006). As already suggested, it may be that the two different investigatory groups had different thresholds for making a call of micro-/anophthalmia. The degree of care taken by Warren et al. (2006), however, suggests that eye defects, if present, would not have been missed in this study. Alternatively, it may be that the Fischer 344 rats used in Narotsky and Kavlock (1995) and Narotsky et al. (1995) are substantially more sensitive to the development of eye defects upon insult compared to the Sprague-Dawley rats used in Warren et al. (2006). Of particular note, however, is the fact that maternal toxicity was not seen in the study by Warren et al. (2006). These results suggest that the internal TCE dose achieved in the Sprague Dawley rats, as compared to that of the Fischer 344 rats used in the other studies (Narotsky and Kavlock 1995; Narotsky et al. 1995), was not sufficient to cause micro-/anophthalmia in the resulting offspring.

Finally, although the results of Warren et al. (2006), as well as those of a study by Smith et al. (1992) suggest that prenatally administered DCA can alter eye development in pups, it is unlikely that the results of the studies by Narotsky and Kavlock (1995) and Narotsky et al. (1995) are due to DCA production because DCA is not a major TCE metabolite in rats or humans (Goeptar et al. 1995; Templin et al. 1995; Lash et al. 2000). On the other hand, the effects might be due to production of TCA, which has been shown to cause ocular defects (Warren et al. 2006; Smith et al. 1989). The question remains, however, whether the TCE doses causing micro-/anophthalmia in Narotksy and Kavlock (1995) and Narotsky et al. (1995) could result in high enough internal TCA concentrations to produce ocular defects. Based on a pharmacokinetic study of TCE exposure in the pregnant rat (Fisher et al. 1989), it can be estimated that 12% of a TCE dose will be converted to TCA in the body. Working backwards and assuming 100% bioavailability, it is estimated that a TCE dose of 1000 mg/kg/day is approximately equivalent to a TCA dose of 120 mg/kg/day, and the lowest TCE dose at which an increased incidence of micro-/anophthalmia was observed—101 mg/kg/day—is approximately equivalent to a TCA dose of 12 mg/kg/day. The doses of TCA at which ocular defects were observed in Warren et al. (2006) and Smith et al. (1989) were certainly higher at 300 and 800 mg/kg/day, respectively. Finally, the relevance of the findings of Narotsky and Kavlock (1995) and Narotsky et al. (1995) to human health risk assessment is questionable. The current maximum contaminant level for TCE in drinking water is set at 5 ppb. An average human of 70 kg drinking 2 L of water per day would consume 10 μg TCE per day or 0.14 μg/kg/day. This dose is 7.214 × 10⁵ fold lower than the lowest TCE dose that was reported to cause ocular defects in rats (101 mg/kg/day). This comparison indicates that ocular defects are not likely to occur in humans at environmentally relevant exposure levels.

CONCLUSIONS

Examination of the available developmental TCE research demonstrates that only two studies, which administered the highest doses of TCE, report possible ocular defects in resulting offspring (Narotsky and Kavlock 1995; Narotsky et al. 1995). Detailed analysis of these two studies establishes that they are non-traditional developmental toxicity screening studies for which there are few historical control data. Structural changes of the eyes in resulting offspring were consistently reported in conjunction with the administration of high-bolus TCE doses that also caused maternal toxicity. Additionally, dose-related reductions in fetal weights were observed in both of these studies—a finding that, if not taken into proper consideration, may have confounded the diagnosis of microphthalmia. Given the extreme bolus doses employed to achieve a statistically increased incidence of eye defects, the findings of micro-/anophthalmia reported in these two studies are of little to no relevance for assessing human risk at reasonably expected workplace and environmental exposure levels. In conclusion, based on a thorough analysis of the two above described studies and consideration of their findings in the context of other TCE developmental toxicity studies, ocular defects as a result of in utero TCE exposure in humans at environmentally relevant exposure levels are not anticipated.