Associations between per- and polyfluoroalkyl substances and menstrual cycle regularity in reproductive-aged female: A cross-sectional study

Abstract

Menstrual cycle characteristics constitute one of the significant female fertility indicators. Previous epidemiological studies have shown that exposure to environmental chemicals could affect menstrual cycle characteristics, but the knowledge remains limited overall. Per- and polyfluoroalkyl substances (PFAS) have been identified as potential reproductive toxicants, while previous studies mainly focused on several legacy PFAS chemicals but generally failed to explore the outcomes from exposure to a complex mixture of both legacy and emerging PFAS. Besides, the modification effect of physical activity is rarely considered. In the present study, we explored the associations of exposure to a suite of legacy and emerging PFAS and menstrual cycle regularity as well as the potential modification by physical activity based on a pre-conception cohort in Shanghai (China) with the participation of 1001 reproductive-aged women. A total of 20 PFAS chemicals with detection frequency >80 %, which were derived from the PFAS exposure profile of the same population in our previous study, were included in the confounder-adjusted logistic regression and Bayesian kernel machine regression (BKMR) analysis. In individual PFAS analysis, after adjustment of the covariates, ∑2m-PFOS (the sum of all perfluoro-dimethylhexane sulfonates) was significantly associated with menstrual cycle irregularity with an odds ratio (OR) of 1.35 (95 % confidence interval, CI: 1.09, 1.67) as well as long cycles (OR = 1.37; 95 % CI: 1.08, 1.70). In addition, a significant positive association was also found between perfluoro-n-nonanoic acid (PFNA) and long cycles (OR = 1.40; 95 % CI: 1.06, 1.86). No significant associations were found between the PFAS mixture and the menstrual cycle characteristics as revealed by BKMR analysis, while the significant association between ∑2m-PFOS and menstrual cycle irregularity was also observed in the mixture exposure model. Subgroup analysis stratified by physical activity level showed that the associations between ∑2m-PFOS and menstrual cycle irregularity as well as long cycles were more pronounced in the inactive physical activity subgroup. This study suggested that branched PFOS (i.e., ∑2m-PFOS) might act as the predominant risk factor for menstrual cycle irregularity, and physical activity could influence the risks.