The results of this study suggest that prenatal exposure to PCE increases the risk of certain kinds of congenital anomalies. Prenatal exposure was associated with large increases in the risk of gastrointestinal defects (particularly oral clefts), neural tube defects (particularly anencephaly) and, modest increases in the risk of genitourinary defects (particularly hypospadias). No meaningful increases in risk were seen for cardiac, musculoskeletal and chromosomal anomalies, and there were too few exposed cases to estimate odds ratios for eye; ear, face, and neck; respiratory; and other anomalies. An exposure-response relationship was observed for oral clefts but not for neural tube or genitourinary defects.
Several limitations of this study influence the interpretation of these results. First, these findings results are likely affected by exposure misclassification. Because individual level exposure measurements were not available for the study period, we estimated historical PCE exposures using a leaching and transport model developed by Webler and Brown  that predicted the annual mass of PCE delivered to each residence. The model was applied to water distribution system conditions in 1980 and was assumed to be representative of the entire study period. Furthermore, information on water consumption and bathing habits could not be recalled by a sizable number of mothers and so we were unable to incorporate these behaviors into our exposure assessment and data analysis.
Thus, while results from two validation studies indicate good correlation between PCE concentrations in historical water samples and exposure estimates based on the original Webler-Brown algorithm (Spearman correlation coefficient = 0.48, p < 0.0001) , and exposure estimates based on the EPANET water distribution system modeling software (Spearman correlation coefficient = 0.65, P < 0.001) , non-differential misclassification of the exposure remains likely.
The study was also limited by low statistical power stemming from the small number of congenital anomalies reported by the mothers. While the small number also made it difficult to control for confounders, the irregular geographic pattern of PCE exposure made it unlikely that exposed and unexposed mothers differed on both known and unknown risk factors for congenital anomalies. In fact, comparison of available crude and adjusted analyses in the present analysis and prior analyses of the study population [11, 12] provide evidence of limited or no confounding.
Still another limitation stems from the use of maternal reports as the source of information on the congenital anomalies. While comparison between the questionnaires and birth records suggest that study mothers were good reporters of pregnancy-related information, the births occurred many years before the study's data collection and the mothers had no training in teratology, and so it is likely that some anomalies, particularly minor ones, were not reported. Because pediatric records were unavailable, the presence of under-reporting could not be directly verified; however, its likelihood is supported by comparing the prevalence of anomalies in our study population with those observed in the 1960s Collaborative Perinatal Project (CPP), a prospective study of 50,000 women and their pregnancies. Using a comprehensive medical surveillance system, the CPP observed a prevalence of 45.3 malformations per 1,000 live- and still births  as compared to a prevalence of 33.5 per 1,000 in our study population. Surveillance data from the Metropolitan Atlanta Congenital Defects Program, and the CDC Birth Defects Monitoring Program during the 1970s and 1980s also support this notion [e.g., [29–32]].
While under-reporting likely affected the statistical stability of our findings, there is no evidence that it was more or less likely to occur among exposed mothers, because, as described in the results section, most mothers did not know their "true" exposure status. Thus, because the risk of congenital anomalies was small (<10%), the likely impact was a small bias towards the null .
Lastly, while non-participating mothers were younger and less educated than participating mothers, these differences were present for both exposed and unexposed non-participants, and so it is unlikely that selection bias influenced the current results. However, because birth certificates were used to identify study mothers, our results may not be generalizable to women who have never achieved a live birth.
Numerous animal experiments have observed teratogenic effects of prenatal exposure to PCE, TCE, and their metabolite trichloroacetic acid (TCA) that appear to be species- and dose-dependent. Increased rates of several types of malformations, including cardiac, ocular, and skeletal defects, have been seen in chick embryo cultures over a wide range of exposure levels, including doses as low as 1 umol [7, 34, 35]. Increased rates of malformations in the central nervous, musculoskeletal, and cardiovascular systems have also been observed among rats exposed to a wide range of TCE, PCE, and TCA exposures [8, 36–39]. In fact, cardiac malformations were seen among offspring of pregnant rats who ingested drinking water with TCE levels as low as 250 ppb . In contrast, no evidence of teratogenicity has been observed in experimental studies of exposed mice and rabbits [40, 41].
Several epidemiological studies have also found associations between maternal occupational exposure to organic solvents and congenital anomalies . A recent meta-analysis found a statistically significant association with major malformations (summary OR, 1.6; 95% CI, 1.2-2.3) . However, these results are difficult to interpret regarding the risk associated with PCE exposure because a combination of many solvents were examined and because organ system-specific analyses were not conducted. Studies among dry cleaners exposed to PCE have found no increases in the risk of congenital anomalies, but small sample sizes limited their statistical power to examine rare outcomes like anomalies. The total number of cases in these studies ranged from three to thirty-eight [44–47].
In contrast, three drinking water studies have found positive associations for congenital anomalies. Goldberg et al.  found that the prevalence of cardiac anomalies in Tuscon, Arizona was three times higher among children of parents who had contact with TCE contaminated drinking water as compared to unexposed parents (p < 0.005). The contaminants present in the Tuscon Valley water supply included TCE (6 to 239 ug/dL), chromium (below the action level of 0.1 mg/L, and dichloroethylene (concentrations typically between 5-10% of trichloroethylene levels). Lagakos et al.  found that the offspring of women exposed to well water contaminated with TCE (267 ppb), PCE (21 ppb), trichlorotrifluoroethane (23 ppb) and dichloroethylene (28 ppb) in Woburn, Masssachusetts had an increased prevalence of eye and ear anomalies combined, and central nervous system, chromosomal, and oral cleft malformations combined. While many questions have been raised about this study, including the unusual malformation groupings [e.g., [49, 50]], the results are consistent with the animal studies described above and an epidemiological study by Bove et al . The latter New Jersey study found that PCE drinking water levels > 10 ppb were associated with a 3.5-fold increased risk of oral clefts (90% CI: 1.3 - 8.8), that TCE drinking water levels > 5 ppb were associated with a 2.2 fold increased risk of oral clefts (90% CI: 1.2 - 4.2), and that TCE levels > 10 ppb were associated with a 2.5-fold increased risk of neural tube defects (90% CI: 0.9 - 6.4). In contrast, a follow-up study in Woburn found no increases in the risk of cardiac, genital or musculoskeletal defects among children with prenatal exposure to contaminated well water . While the authors concluded that the number of children with other defects was too small for meaningful analysis, they did observe a statistically unstable three to four-fold increased prevalence of eye anomalies based on four exposed cases. Reports of congenital anomalies were obtained from vital records and a more sensitive model of the water distribution system estimated prenatal exposures in the follow-up study.
Taken together, the results of the present and prior studies provide mounting evidence of an increased risk of oral clefts, and perhaps neural tube defects and other anomalies in relation to prenatal PCE exposure from drinking water. However, weaknesses in these studies, including the present one, make it important to confirm these findings in a follow-up investigation with a large number of affected children who are identified using medical or vital records. Because PCE remains a commonly used solvent and frequent contaminant of ground and drinking water supplies [6, 52], it is important to understand its impact on the occurrence of congenital anomalies.