Urinary concentrations of benzophenone-3 and reproductive outcomes among women undergoing infertility treatment with assisted reproductive technologies

Abstract

Benzophenone-3 is used in a variety of cosmetic products as a sunscreen, and has shown weak estrogenic and antiandrogenic activity in animal and in vitro studies. Few studies have evaluated whether benzophenone-3 is associated with reproductive outcomes among women. We studied 304 women undergoing infertility treatment (2007–2017) in the prospective Environment and Reproductive Health cohort study and who underwent 449 treatment cycles (n = 788 urines). Generalized linear mixed models were used with random intercepts to account for multiple cycles, and adjusting for confounders including physical activity. Analyses were also stratified by self-reported moderate/heavy outdoor work. The cycle-specific median (IQR) urinary benzophenone-3 concentration was 147 (58, 462) μg/L, and 98% samples had detectable concentrations. Self-reported sunscreen use, physical activity, and time spent on moderate/heavy outdoor work were positively associated with urinary benzophenone-3. Adjusted probabilities of implantation, clinical pregnancy and live birth were higher in increasing quartiles of benzophenone-3, but these associations were restricted to women who reported spending time outdoors performing moderate/heavy work. Specifically, among these women, those in the highest quartile of benzophenone-3 concentrations had 51% higher implantation (p,trend = 0.02), 68% higher clinical pregnancy (p,trend = 0.01) and 75% higher live birth (p,trend = 0.02) adjusted probabilities than women in the lowest quartile. Benzophenone-3 was unrelated to these outcomes among women who did not report doing moderate/heavy work outdoors. These results confirm that sunscreen use is a source of benzophenone-3 exposure, and show positive associations between benzophenone-3 and pregnancy outcomes, especially among women who reported engaging in outdoor work. Since these associations may be subject to important residual confounding by lifestyle factors, further research is needed to confirm these novel results in other populations, and to investigate whether other factors may be affecting the relation of benzophenone-3 with fertility and other health outcomes.