The design and validation of novel exposure surrogates for past environmental exposures.

B47

The opportunity to measure relevant (i.e. past) exposures to environmental carcinogens has been lost to time, impeding establishing causal relationships, and underscoring the need to develop exposure surrogates. This study is based on the hypothesis that environmental exposures occur from polluted drinking water emanating from groundwater leaching off hazardous waste sites which occurred decades ago, consistent with the latency period of the disease under investigation. I constructed the surrogates using governmental records of hazardous waste sites. A hazard density score, HDS, (the number of hazardous waste sites per square mile), and the chemical contamination index (CCI), designed to reflect exposures to individual chemicals prevalent in waste sites, were constructed on a per water-district basis. The CCI was derived by comparing the total square footage of all waste sites in a water district with its size in square miles, representing the percentage of the water district's contaminated surface area. Water districts were chosen as the exposure boundary, as opposed to zip codes, as better reflecting relevant exposures. Nassau County, New York, was the test site area for validation. The county derives all of its drinking water from ground water, has an unusually high incidence of breast cancer and records of well measurements currently exist. The most frequently detected chemicals in Nassau County waste sites were: perchloroethylene, (PCE) trichloroethylene (TCE), dichloroethylene, (DCE) and trichloroethane (TCA). The same chemicals were found in drinking wells -- in identical order of detection, frequency and prevalence, indicating reliability of the HDS. The study also demonstrated that two pollutants, TCE and DCE, infiltrated wells at high levels, up to two orders of magnitude higher than TCA. An 86% correlation was found between the CCI for DCE (considered a high correlation by the scientific community) and heavily contaminated wells, while no correlation was found between heavily contaminated wells and TCA, indicating that TCA if found at all in drinking wells, did so at very low levels. The correlation between heavily contaminated wells and TCE/PCE was moderate. It appears that DCE in waste sits infiltrates wells at the highest levels of concentration; TCE and PCE, also found in wells at high levels were related to waste site content - but to a lesser degree, and TCA, while present in waste sites, did not infiltrate wells to any significant degree. The investigation of a dose response relationship with the prevalent exposurues, using contamination per square mile as the dose index, can now be undertaken. Further, more effective exposure surveillance through targeted drinking well testing and better groundwater detection can be implemented.