Association between volatile organic compounds exposure and infertility risk among American women aged 18-45 years from NHANES 2013-2020.

The risk of infertility is progressively escalating over the years, and it has been established that exposure to environmental pollutants is closely linked to infertility. As a prevalent environmental pollutant in daily life, there is still a lack of substantial evidence on the association between volatile organic compounds (VOCs) exposure and infertility risk. This study aimed to examine the association between VOCs exposure and the risk of female infertility in the United States. Participant data sets from three cycles (2013-2020) were collected and downloaded from the National Health and Nutrition Examination Survey (NHANES), including demographics, examination, laboratory and questionnaire data. The baseline characteristics of the included population were evaluated, and the weighted quartile logistic regression was used to analyze the association between the urinary metabolites of VOCs (mVOCs) levels and the risk of infertility. Further exploration of the relationship between mVOCs and infertility was conducted by using 35 and 25 as the cut-off points for age and BMI subgroup analyses, respectively. Restricted cubic spline (RCS) was employed to elucidate the nonlinear relationship between mVOCs and infertility risk. Additionally, the Bayesian kernel machine regression (BKMR) model with 20,000 iterations was applied to elucidate the link between mVOCs and the risk of infertility when exposed to mixed or individual mVOCs. A total of 1082 women aged 18 to 45 years were included in this study, with 133 in the infertility group and 949 in the control group. The analysis of baseline characteristics suggested that urinary 34MHA, AMCC and DHBMA levels were significantly higher in the infertility group compared to the control group (p < 0.05). Quartile logistic regression analysis indicated that AAMA (Q3), AMCC (Q4), CYMA (Q3) and HPMMA (Q3) were positively associated with infertility risk in all models (p < 0.05). Subgroup analysis revealed different risk factors for infertility among various subgroups, with CYMA consistently showing a positive correlation with infertility risk in two age subgroups (p < 0.05). Furthermore, the association between mVOCs and infertility was observed only in the subgroup with BMI ≥ 25 kg/m. RCS analysis indicated that 2MHA, ATCA, BMA, BPMA, CYMA, 2HPMA, 3HPMA and PGA exhibited linear dose-response relationships with infertility (p > 0.05), while the remaining variables showed nonlinear relationships (p < 0.05). The BKMR model demonstrated that the risk of female infertility exhibited an increasing trend with the accumulation of mVOCs co-exposure. A positive association between the exposure to mVOCs represented by 34MHA and AMCC and the risk of infertility was observed in this research. However, the inherent limitations associated with the cross-sectional study design necessitate the pursuit of additional prospective and experimental research to further elucidate and validate the relationships between various mVOCs exposure and female infertility.