Toxicological and epidemiological evidence on the association between perfluorooctanoic acid (PFOA) or perfluorooctane sulfonic acid (PFOS) and birth/fetal weight was assessed. An extensive search for toxicological information in rats and mice, and a systematic search for epidemiological evidence were conducted. The linear regression coefficient (LRC) of birth weight (BrthW) on PFOA/PFOS was considered, and separate random effects meta-analyses for untransformed (i.e. not mathematically transformed) and log-transformed values were performed. Toxicological evidence: PFOA: 12 studies (21 datasets) in mice showed statistically significant lower birth/fetal weights from 5 mg/kg body weight per day. PFOS: most of the 13 studies (19 datasets) showed lower birth/fetal weights following in utero exposure. Epidemiological evidence: Sixteen articles were considered. The pooled LRC for a 1 ng/mL increase in untransformed PFOA (12 studies) in maternal plasma/serum was -12.8 g (95% CI -23.2; 2.4), and -27.1 g (95% CI -50.6; -3.6) for an increase of 1 loge ng/mL PFOA (nine studies). The pooled LRC for untransformed PFOS (eight studies) was -0.92 g (95%CI -3.4; 1.6), and for an increase of 1 loge ng/mL was -46.1(95% CI -80.3; -11.9). No consistent pattern emerged for study location or timing of blood sampling.
Conclusions: Epidemiological and toxicological evidence suggests that PFOA and PFOS elicit a decrease in BrthW both in humans and rodents. However, the effective animal extrapolated serum concentrations are 102-103 times higher than those in humans. Thus, there is no quantitative toxicological evidence to support the epidemiological association, thus reducing the biological plausibility of a causal relationship.
Keywords: Birth weight; PFOA; PFOS; epidemiology; evidence; fetal growth; integration; meta-analysis; perfluoroalkyls; reproduction; systematic review; toxicology.