Disinfection by-products and prenatal development

Abstract

While toxicologic studies have suggested modes of action for disinfection by-product (DBP) toxicity or teratogenicity, there is little evidence for a causal relationship between DBPs and induction of adverse outcomes in humans. This may be due, in part, to biases inherent to exposure assessment or analytical methodologies used in past studies. This dissertation presents three approaches designed to develop and evaluate new methods of minimizing exposure misclassification and improving risk estimation for studies of DBPs and prenatal development. These approaches are described below. Past epidemiological studies of DBPs and birth outcomes have used exposure measurements for study participants that are susceptible to misclassification due in part to intra-system spatial variation. Spatial variability occurs when DBP contaminants in treated drinking water do not occur uniformly across individual public water distribution systems. We conducted a feasibility study to identify and test methods that can be used to select communities that are served by water distribution systems with low spatial variability for epidemiologic investigations. We also aimed to identify guidelines for characterization of brominated compounds. These methods were tested on quarterly DBP concentrations from the distribution system of 198 facilities, listed in USEPA's Information Collection Rule. We identified sites with low spatial variability that had high overall levels of brominated trihalomethanes (THMs), high overall levels of chlorinated THMs and low levels of THM species. We classified spatial variation of THMs within these distribution systems using methods based upon: overall levels of chlorinated THMs and low levels of total THM species. We classified spatial variation of THMs within these distribution systems using methods based upon: (1) two-way analysis of variance, and (2) outpoints deemed biologically important in epidemiologic literature. The method based upon epidemiologic literature was determined to be superior, as the two-way ANOVA method utilized arbitrary significance levels to ascertain low spatial variability. Using the epidemiologic literature method, we identified twenty sites with low spatial variability, one of which had consistently high THM levels that were primarily brominated. This study presents a simple method for a priori selection of sites with low spatial variability as a means to reduce misclassification in exposure assessment for epidemiologic studies of DBPs. Previous epidemiologic studies of DBPs and growth-related birth outcomes such as preterm birth or low birth weight have not attempted to estimate exposure over the third trimester, the time period of biological importance to induction of these outcomes. We conducted a study to identify the existence and magnitude of biases that arise from use of variable third trimester lengths and regression techniques (Dodds et al., 1999; King et al., 2000; Dodds and King, 2001), when temporally variable environmental exposures, such as DBPs, are evaluated against time-dependent outcomes. Using examples from a simulated population modeled after the U.S. distribution of normal and preterm births, we evaluated how selection of participants and cutpoint use affected risk estimation. We found that failure to adjust for variable length of third trimesters and use of cutpoints with regression can potentially lead to significant bias away from or towards the null, depending on peak or trough exposure and placement of outpoint, and further complicate estimation of true risk in epidemiologic studies that likely contain biases from other sources such as inadequate or improper methods to assess exposure. Toxicological studies have demonstrated associations between exposure to high concentrations of DBPs and several teratogenic outcomes. Exposure to high doses of haloacetic acids (HAAs), the second most commonly occurring group of DBPs, causes cardiovascular and orofacial malformations, such as cleft palate, in laboratory studies. Human studies have focused primarily on low-dose exposures to THMs. Most epidemiological studies have not evaluated exposure over time periods of pregnancy that are specific to induction of the specific outcome. We conducted a retrospective cohort study of cleft palate to determine whether a relationship exists with exposure to THMs or the five most prevalent individual haloacetic acids (HAA5), which include trichloroacetic acid (TCAA). The study population consisted of all singleton births in a large community served by three water treatment plants. This community was chosen since THMs and HAAs varied by season with levels ranging from low to moderate, but exhibited little spatial variability within the distribution system. Birth certificate data (n=51,717) were obtained and logistic regression techniques were used to estimate the relationship between mean exposures to DBPs over the first trimester, second month, and 6th, 7th and 8th weeks of pregnancy and cleft palate. We found no increased risk with categorical levels of exposure to THMs or HAAs during critical time windows. In addition, no associations were observed for any DBP evaluated as a continuous variable. However, the power of this study was low due to a small number of cleft palate cases.