Associations between internal concentrations of fluorinated and organochlorinated chemicals in women and in vitro fertilization outcomes: A multi-pollutant study

Abstract

The impact of persistent organic pollutants (POPs) on reproductive health is still poorly understood, even though infertility management has high associated societal and economical costs. The aims of this study were to characterize the internal levels of polychlorinated biphenyls (PCBs), organochlorine pesticides (OCPs) and perfluoroalkylated substances (PFAS) in women undergoing in vitro fertilization (IVF); and evaluate their association with IVF outcomes, individually and as mixtures in a combined multipollutant approach. Thus, 136 women undergoing IVF treatment at Nantes University Hospital (France) were prospectively recruited between 2019 and 2020. Serum samples were analyzed using liquid chromatography with tandem-mass spectrometry for 14 PFAS. Follicular fluid was analyzed with gas chromatography coupled to high resolution mass spectrometry for 14 PCBs and 25 OCPs. Intermediate and clinical IVF outcomes were ascertained by embryologists and clinicians using standardized protocols. Multivariate Poisson regression models and Bayesian Kernel Machine Regressions (BKMR) were used to identify individual and joint associations between POPs and IVF outcomes adjusting for age, body mass index (BMI) and anti-Müllerian hormone. The results showed that most POPs were widely present in women, and globally not associated with clinically relevant IVF outcomes, like live birth rates. Nonetheless, negative associations between PCB138 and trans-nonachlor with useable blastocysts were identified, β −0.28 (95%CI [-0.52; −0.04] p = 0.02) and β −0.22 (95%CI [-0.40; −0.03] p = 0.02). Conversely, PCB28 showed positive associations with the number of useable blastocysts, pregnancy rate and live birth rate. The BKMR analysis suggested the lack of association of the mixture with intermediate and clinical outcomes. The study supports the need of conducting further studies in a larger population sample in order to ensure sufficient statistical power to identify modest effects and a robust stratification analysis to account for the large underlying disease heterogeneity.