Precision prediction of hyperhomocysteinemia development in perimenopausal women using LASSO regression

TL;DR

This study develops a predictive model using LASSO regression to identify risk factors for hyperhomocysteinemia in perimenopausal women, offering a potential tool for early detection and treatment.

Contribution

A novel LASSO-based predictive model for hyperhomocysteinemia in perimenopausal women, validated across multiple datasets.

Findings

Four predictors (egg consumption frequency, LDL, TP, and CysC) were identified for the nomogram model.

The model showed good predictive performance with AUCs of 0.765, 0.854, and 0.776 in training, internal, and external validation sets, respectively.

The nomogram demonstrated high clinical utility for HHcy risk prediction in perimenopausal women.

Abstract

Hyperhomocysteinemia (HHcy) is associated with an increased risk of cardiovascular diseases, particularly in perimenopausal women, who are more susceptible to metabolic disorders due to declining estrogen levels. This study aimed to identify risk factors and develop a predictive model for HHcy in this population.

A retrospective study included 687 perimenopausal women, divided into a training set (481) and an internal validation set (206). Demographic characteristics, pregnancy-related factors, lifestyles, and diet information were collected by questionnaire. 63 perimenopausal women hospitalized from March to June 2025 were selected as the external validation set. The least absolute shrinkage and selection operator (LASSO) regression was used to select variables. The logistic regression model was developed to predict HHcy risk, with results visualized using a nomogram. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

137 of 687 (19.94%) perimenopausal women had HHcy. Through Lasso regression and multifactor logistic regression, 4 predictors were identified, including egg consumption frequency, LDL, TP, and CysC for constructing the nomogram model. The AUC of the training set was 0.765 (95% CI = 0.708–0.822), for the internal validation set was 0.854 (95% CI = 0.781–0.928), and for the external validation set was 0.776 (95% CI = 0.603–0.949), indicating good predictive performance of the model.

The nomogram demonstrated high predictive accuracy and clinical utility, providing a potential tool for HHcy risk prediction and selection of treatment strategies in perimenopausal women.

Figures