The results of this study suggest that the risk of certain mental illnesses is increased among individuals who were exposed to PCE-contaminated drinking water in early life. In particular, subjects with any exposure during gestation and early childhood had elevations in the risk of bipolar disorder and post-traumatic stress disorder that were further increased among subjects with the highest exposures. While the risk of schizophrenia was also elevated among exposed subjects, the number of cases was too small to draw reliable conclusions. In contrast, the risk of depression was not associated with prenatal and childhood PCE exposure.
The observed associations should be judged in light of the study limitations. First, the results are likely affected by exposure misclassification. Because historical exposure measurements were unavailable, we estimated the mass of PCE delivered to each subject's residence using EPANET water distribution modeling software that incorporated a leaching and transport model [13, 15]. The model did not incorporate information on water consumption and bathing habits due to poor maternal recall of these behaviors. While the model also necessitated many assumptions about the water distribution system (for example, that all users drew the same amount of water), results from validation studies show good correlation between our exposure estimates and PCE concentrations in historical water samples (Spearman correlation coefficient (ρ) = 0.65, P < 0.00010) .
Because exposure misclassification was likely to be present for all subjects, regardless of their mental health status, the limitations in the exposure assessment likely attenuated the findings from dichotomous comparisons . The direction of bias for comparisons involving the exposure tertiles is harder to predict, but risk ratios among subjects in the highest tertile are likely biased towards the null while those in the middle category may be biased either towards or away from the null.
Still another limitation arises the use of self-reports as the source of information on the mental illnesses. While subjects were asked if a health care provider ever said that they had each mental illness, it is unknown if the provider made the diagnosis based on criteria from the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders . Furthermore, the prevalence of the illnesses among study subjects was slightly lower than that observed in general population surveys conducted in Massachusetts  and the United States [25, 26]. For example, in 2001-2003 the lifetime prevalence of depression, bipolar disorder, and PTSD were 22.7%, 4.5% and 8.2%, respectively, in a nationally representative sample of adults aged 30-44 years .
Nonetheless, it is unlikely that inaccuracies in the diagnosis and reporting of the mental illnesses were related a subject's PCE exposure history because most subjects (and, by extension, their physicians) did not know this information. In particular, only 7% of subjects considered exposed by the modeled assessment believed that their drinking water was contaminated, whereas 29% believed that their water was not contaminated and 64% were unsure. Similarly, 31% of subjects considered unexposed by the modeled assessment believed that their drinking water was not contaminated while 5% believed that their drinking water was contaminated and 63% were unsure. The subjects' lack of knowledge about their exposure likely stems from their young age when the PCE contamination was publicized in the early 1980s.
A further limitation stems from possible residual confounding because of missing data on several risk factors for these mental illnesses, such as social supports and stressors . However, in order to confound the associations observed in this study, these factors would need to be tightly correlated with PCE exposure, an unlikely circumstance given the uneven distribution of PCE contamination across Cape Cod. In fact, our current and prior studies of this cohort have found little or no confounding for the associations under investigation by variables that were collected [12, 28].
Another limitation stems from the study's low response rate which reduced the statistical power of the study. It was also quite possible that individuals with mental illnesses, particularly severe mental illnesses, preferentially ignored our requests for participation and so reduced the number of cases in the final analytic population. However, available evidence suggests that this did not produce selection bias. First, similar proportions of participants and non-participants were exposed to PCE (36.7% of participants vs. 38.8% of non-participants). Second many characteristics of participants and non-participants were alike, including age, race, duration of gestation, birth weight, and birth order. Third, while participants were more likely to be female (60.7% of participants vs. 43.5% of non-participants) and have mothers with some college education (61.0% for participants vs. 49.3% for non-participants), these differences were present for exposed and unexposed non-participants. Lastly, losses stemming from the death of potential participants were small and unrelated to initial PCE exposure status (n = 111, Table 1). Also, our review of death records from the Massachusetts Registry of Vital Records and Statistics and the National Death Index found only one notation of a mental illness (acute psychosis induced by LSD) as a contributing cause of death.
PCE's potential to cause neurotoxic effects has been established through numerous animal and human studies . Because PCE is a relatively small fat-soluble molecule, it easily crosses the blood brain barrier and has a high affinity for the lipophilic tissues of the central nervous system. While the mechanism by which PCE might cause neurotoxic effects is currently unknown , there is evidence to support mechanisms involving the peroxidation of cell membrane lipids , alterations in the fatty acid profile of the brain , and loss of myelin , and interactions with neuronal receptors .
High levels of prolonged occupational exposure to PCE and other solvents have been associated with chronic solvent induced encephalopathy  and, in some cases, schizophrenia among adults [34–36]. Solvent levels above the accepted threshold have also been associated with increases in anger and confusion among child workers aged 10-17 years . Lower levels of occupational exposure have been also associated with increases in anxiety and depression among adults [4–8]. Because most of these occupational studies used employment information (such as job title and work practices) rather than industrial hygiene measurements to characterize individual exposure levels, it is difficult to compare directly the exposure levels in our population with those experienced by these worker groups. However, workplace settings typically result in higher exposure levels than environmental ones.
One community-based study of trichloroethylene (TCE)-contaminated drinking water also found higher levels of depression/dejection, confusion/bewilderment, and tension/anxiety among highly exposed adults; however, the effect was mainly attributable to exposed subjects who also consumed alcoholic beverages . Highly exposed subjects in this study received drinking water with estimated TCE levels that were greater than 15 ppb (the uppermost value was not stated). In comparison, PCE levels in the drinking water supplies of our study population ranged from 1.5 to 7,750 ppb.
To the best of our knowledge, only one prior study has examined the impact of early life exposure to solvents on subsequent mental health. This study assessed the risk of schizophrenia among Israeli children whose parents worked in dry cleaning occupations by linking information on parental occupations from birth records with information on mental illness from a national psychiatric registry . Four cases of schizophrenia were observed among 144 offspring with at least one parent working in dry cleaning (prevalence = 2.8%) for a crude relative risk of 3.4 (95% CI: 1.3-9.2). Adjustment for confounding factors did not appreciably alter these results.
In summary, the results of this study suggest that the risks of certain mental illnesses, particularly bipolar disorder, post-traumatic stress disorder, and schizophrenia are increased among adults exposed to PCE during gestation and early childhood. Independent investigations of similarly exposed populations are needed to corroborate these findings. Because PCE remains a commercially ubiquitous solvent and common contaminant of drinking water supplies [38, 39], it is important to determine its impact on the health of vulnerable populations.