Development and validation of a predictive model for carotid atherosclerosis in postmenopausal women

With the global aging of the population, menopausal women face higher cardiovascular disease (CVD) risks, with carotid atherosclerosis as the primary pathological basis. However, no effective tools exist for assessing carotid atherosclerosis risk, and this study fills the gap in predictive tools in this field. Using data from 4,446 menopausal women in Shenzhen, we developed and validated a Nomogram model for carotid atherosclerosis risk. The sample was divided into training (2,178), internal validation (934), and external validation (1,334) sets. Variables were selected using logistic regression and LASSO, including age, systolic blood pressure (SBP), lipoprotein a (LPa), non-HDL cholesterol (Non-HDL-C), TC/HDL-C ratio, glycosylated hemoglobin (HbA1c), and blood glucose (GLU). Random Forest validation confirmed the model's robustness. The Nomogram's C-index was 0.706 (training), 0.664 (internal), and 0.668 (external), with Random Forest results of 0.721, 0.662, and 0.661, respectively. Calibration and decision curve analyses demonstrated the model's accuracy and clinical utility. Additionally, a slight negative correlation between age and GLU (OR = 0.689, P = 0.068) suggested reduced glycemic risk with age. This model provides a scientific basis for early risk assessment and personalized interventions for menopausal women, guiding future research on related biological mechanisms.