Predictive performance of multiple anthropometric obesity indicators for female infertility risk: a cross-sectional...

Various obesity metrics are used to assess female infertility risk, but their relative prognostic value remains unclear. This cross-sectional study systematically evaluated 16 anthropometric indices to identify the most effective predictors.

Using National Health and Nutrition Examination Survey (NHANES) 2017-2020 data (n = 1154 reproductive-aged women), we applied a three-phase framework: (1) weighted restricted cubic spline (RCS) models were applied to examine potential non-linear associations between 16 anthropometric obesity indicators and female infertility risk, with univariate analyses used to assess their statistical associations, (2) two model states ('real-world settings' and 'ideal-world settings') were constructed based on the union and intersection of indicator datasets, and weighted logistic regression models were developed for each state and (3) predictive performance and clinical utility were assessed using receiver operating characteristic (ROC) curves, calibration curves and decision curve analysis (DCA), with key diagnostic metrics compared to identify optimal predictors.

Thirteen obesity indicators showed predictive value for infertility; Anthropometric Body Shape Index (ABSI) exhibited a U-shaped association, whereas the others demonstrated positive linear relationships. Body mass index (BMI), Weight-Adjusted Waist Index (WWI) and Conicity Index (CI) demonstrated relatively better predictive performance (AUC > 0.65), with WWI showing slightly stronger association with infertility than the other indicators in both model states (adjusted OR = 1.40, 95% CI: 1.10-1.77; AUC = 0.656; results were consistent across the two model states).

Thirteen obesity indicators predicted infertility; ABSI was U-shaped, others were linear. WWI was slightly superior, though differences were modest.