Guest Editorial: Biomarkers of Perfluorinated Chemicals and Birth Weight

appearing in this issue of Environmental Health Perspectives evaluate the relationship of perfluo-rooctanoate (PFOA) and perfluororooctane sulfonate (PFOS) with birth weight; they constitute the first rigorous epidemiologic studies to address this topic. The ubiquity of these chemicals in the environment and in humans and the growing concerns about potential health effects from such toxicants make these studies a valuable addition to the health effects literature. The ultimate implications of these studies concerning etiologic relationships with human health remain to be elucidated, but they raise two important issues: the interpretation of small shifts in birth weight, and the potential for shared determinants of exposure biomarkers and biological indicators of health outcome. Birth weight offers many advantages as an end point for studies of environmental agents. It is measured easily and accurately, shows substantial variability, and has been proven to be sensitive to environmental insults, most notably tobacco smoke (Wilcox and Russell 1983; Windham et al. 2000). Because birth weight is measured on a continuous scale, even moderate-size studies often have excellent statistical power to detect small decrements. On the other hand, such shifts in birth weight, even if ultimately proven to be causal, reflect variation within the normal range of the distribution, in which there appears to be few or no direct consequences for infant mortality or morbidity (Wilcox 2001). Even when there are relatively large shifts in birth weight, as associated with altitude or cigarette smoking, there do not appear to be adverse effects mediated by the shift in mean birth weight per se. Although shifts in the entire distribution would be expected to increase the number of births at the extreme low end, in fact it seems that influences on the dominant part of the birth weight distribution operate separately from determinants of the extreme residual end of the scale (Wilcox 2001; Wilcox and Russell 1983). When studies measure changes in mean birth weight, the results reflect only the dominant distribution—the normal ranges. The same is true for gestational age: Changes in mean gestational age at birth tell us little about changes in the risk of preterm birth, and instead reflect changes in the normal 38-to 40-week range where most births occur. The number of births falling into the residual tail of the distribution determines the clinical and public health outcomes of importance. In contrast, blood pressure, body mass index (BMI), and blood glucose levels appear to affect risk …