Identification of Important Risk Factors for Hypoproteinemia in Polycystic Ovaries Patients

To identify the significant influencing factors of hypoproteinemia in patients with polycystic ovary syndrome.

This cross-sectional investigation utilized data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Multifactorial logistic regression analysis was employed to explore factors influencing hypoproteinemia. Receiver operating characteristic (ROC) analysis and decision curve analysis (DCA) were conducted to compare the discrimination performance and net clinical benefit of influencing factors for hypoproteinemia. The XGBoost algorithm was used to identify crucial variables. The relationships between key factors and hypoproteinemia were explored using restricted cubic spline (RCS) analysis. The stability of this relationship was verified through sensitivity analyses, including trend regression analysis and interaction tests.

115 patients with polycystic ovary syndrome or polycystic ovaries were included in this study, of whom 24.348% were hypoproteinemia. Calcium, prothrombin time, and hyperlipidemia were the most stable independent associations with hypoproteinemia. Among the three, calcium had the highest AUC (0.737) and sensitivity (0.839), while hyperlipidemia showed the highest specificity (0.954). The combination of hyperlipidemia and calcium significantly enhanced the net clinical benefit compared to single hyperlipidemia or calcium. Calcium may be the more critical factor with a stable influence on hypoproteinemia than hyperlipidemia.

The combination of calcium and hyperlipidemia may serve as a superior discriminative factor of hypoproteinemia, with calcium potentially being the more crucial factor. A value below the normal range threshold for calcium levels is a good early warning reference for hypoproteinemia.