Development and validation of a discrimination model for benign prostatic hyperplasia: key influencing factors, interaction effects, and the protective role of g-tocopherol.

Background: Reports on the discrimination of benign prostatic hyperplasia (BPH) are rare. This study investigated the related factors and constructed a model to discriminate the risk of BPH, and also performed an in-depth analysis on key indicators. Methods: The data were collected from the National Health and Nutrition Examination Survey (NHANES) database. The logistic regression analysis was conducted on differential variables between the BPH and non-BPH groups to explore the independent factors of BPH. Then a nomogram discrimination model was constructed and validated, followed by discriminative performance assessment through receiver operating characteristic (ROC) curves, calibration curve, and decision curve analysis (DCA). Further, we deeply analyzed the action mode of core indicators by SHapley Additive exPlanations (SHAP), interaction effect, and subgroup analyses. Results: A total of 4,151 participants were included, with 820 having BPH. The logistic regression results showed that age, race, smoking, alcoholic drinks, waist circumference (WC), g-tocopherol, and free prostate-specific antigen (PSA) were independently related to BPH. A nomogram constructed by the seven factors had a great discriminative performance on BPH both in training sets and validation sets, and can obtain a favorable clinical net benefit. Within the discrimination model, age, PSA, and WC were the more core influencing factors, and they can jointly amplify the risk of BPH. The effect of g-tocopherol was regulated by these background factors. Conclusions: This study identified key factors and constructed a useful tool for discriminating BPH, and also revealed the action mode of core factors. It provides a scientific basis for risk discrimination and key mechanism analysis of BPH.