Serum Perfluorooctanoic Acid and Birthweight: An Updated Meta-analysis With Bias Analysis

Background: A recent meta-analysis of 15 studies found a change in birthweight of -12.8g (95% CI = -23.1, -2.38) per ng/ml PFOA in maternal or cord blood and -27.1g (-50.6, -3.6) per log ng/ml PFOA. Almost all studies were done in low-exposed populations. There are nine new studies, adding 6,019 births to the previous 6,937 births. Methods: We conducted a meta-analysis of 24 studies. To combine all results, we approximated results for untransformed PFOA from nine studies using log-transformed PFOA. We also included another large study, excluded from previous analyses, in a sensitivity analysis. Results: We found a change of birthweight of -10.5g (-16.7, -4.4) for every ng/ml PFOA in maternal or cord blood. After adding one previously excluded large study, we found little evidence of an association (-1.0g; 95% CI = -2.4, 0.4). Restricting to studies where blood was sampled from mothers early in the pregnancy or shortly before conception (5,393 births), we found little association of PFOA with birthweight (-3.3g [-9.6, 3.0]). In studies where blood was sampled late in the pregnancy (7563 pregnancies), lower birthweight was associated with higher PFOA (-17.8 [-25.0, -10.6]). Conclusion: Present human evidence provides only modest support for decreased birthweight with increasing PFOA. Studies with a wide range of exposure, and studies with blood sampled early in pregnancy, showed little or no association of PFOA with birthweight. These are studies in which confounding and reverse causality would be of less concern.