Stress Hyperglycemia Ratio and Preoperative Acute Heart Failure in Elderly Hip Fracture Patients with Chronic Cerebrovascular Diseases

Background: The prognostic significance of the stress hyperglycemia ratio (SHR), calculated using admission blood glucose (ABG) and glycated hemoglobin (HbA1c), remains unclear in the context of perioperative complications associated with hip fractures. This study evaluates the association and predictive capability of the SHR for preoperative acute heart failure in hip fracture patients with chronic cerebrovascular disease. Methods and Results: A total of 515 patients were analyzed, revealing a 16.8% incidence of preoperative HF. Significant intergroup differences in age (P < 0.001), ASA classification (P < 0.001), coronary heart disease (P=0.025), serum albumin level (P < 0.001), and admission SHR (P=0.033) were identified between patients with and without preoperative heart failure. Logistic regression identified SHR, CHD, ASA, neutrophils, and hypoproteinemia as independent predictors of preoperative HF, leading to the development of a nomogram that demonstrated good predictive performance (AUC 0.72). Calibration curves and decision curve analysis indicated the model’s clinical utility. Machine learning algorithms, especially the XGBoost model, surpassed traditional models by achieving an AUC of 0.787, demonstrating excellent calibration and clinical applicability. SHAP analysis validated the stable contribution of SHR in risk prediction. Conclusions: This study highlights the SHR’s strong predictive ability for preoperative HF in hip fracture patients with chronic cerebrovascular disease. The XGBoost machine learning model could improve risk stratification and clinical decision-making for this at-risk group.