The Social Epidemiology of Bisphenol A Exposure

Abstract

ISEE-0385 Background and Objective: Socioeconomic disparities in exposure to endocrine-disrupting chemicals such as bisphenol A (BPA) have not been well characterized, nor has the complex relationship between socioeconomic position (SEP) and race/ethnicity. Studying these differences may reveal that certain groups bear a disproportionate burden of exposure and may generate useful hypotheses about sources of exposure. Methods: We conducted a cross-sectional study of 2,517 participants in the 2003-2004 National Health and Nutrition Examination Survey with urinary BPA measurements. Using multiple regression models and controlling for age, gender, and creatinine, we investigated the association between log-transformed BPA levels and numerous self-reported socioeconomic indicators, which capture slightly different aspects of social stratification. We attempted to disentangle effects of race/ethnicity and SEP. Results: Measuring SEP by income, education, and occupation revealed somewhat similar inverse relationships with BPA levels. Income showed the strongest association; people with poverty income ratios (PIR) less than one had BPA levels 23% (95% CI 11-37%) higher than those with PIRs greater than three. Income measures incorporating household size were better predictors, as was education compared to occupation. Education effects remained after controlling for income. Other indicators such as food security also had residual effects. In models not adjusted for SEP, Non-Hispanic Blacks had BPA levels that were 10% higher than Non-Hispanic Whites. With SEP adjustment, this difference dropped to 5%. Mexican Americans, on the other hand, had lower BPA levels than Non-Hispanic Whites, a difference that increased from 21 to 27% with control for SEP. Conclusion: Lower SEP, measured in different ways, was associated with higher urinary BPA levels. Dietary patterns, known to differ by SEP, may contribute to this disparity. For Non-Hispanic Blacks, SEP appeared to mediate some of the observed difference by race/ethnicity, but not for Mexican Americans, indicating possibly different sources of exposure in these two groups.