PCBs, p,p’-DDE and birth outcomes in 11 European birth cohorts: dose-response relationship and effect modifiers

Abstract

Background: Low-level exposure to polychlorinated biphenyls (PCB) is inversely associated with fetal growth, but the dose-response relationship and effect modifiers of such association are not well established. This study is an extension of an earlier European meta-analysis. Aims: To explore the dose-response relationship between PCB-153 and p-p’-DDE and selected pregnancy outcomes, and to evaluate if no-effect threshold and susceptible subgroups exist. Methods: We used a pooled dataset of 8702 mother-child pairs enrolled in 14 study populations from 11 European birth cohorts. General additive models were used to evaluate the shape of the relationship between organochlorine compounds and pregnancy outcomes and analysis of interactions were performed. Results: We observed an inverse linear dose-response relationship between prenatal exposure to PCB-153 at levels below 1200 ng/L and birth weight (?=-200 g/µg PCB-153, 95%CI: -338, -62). No threshold or no-effect levels at very low PCB-153 concentrations were identified. The greatest reductions on birth weight were found in girls and in children born from smoking and non-Caucasian mothers. We also found a small but statistically significant alteration in gender ratio for PCB-153 (OR: 1.019; 95%CI: 1.001, 1.038) and p,p’-DDE exposure (OR: 0.992; 95%CI: 0.999, 0.996), corresponding to an increase and a decrease in the proportion of girls, respectively. Conclusions: This study suggests that the association between low-level exposure to PCB-153 and birth weight follows an inverse linear relationship. Our study also suggests that gender, maternal smoking, and ethnicity are effect modifiers of such association. Whether weak associations between PCB-153/p,p’-DDE and gender ratio reflects biological mechanisms remains to be established.