A nomogram prediction model for perioperative heart failure in elderly patients with intertrochanteric femoral fracture based on lasso-logistic regression.

OBJECTIVE: To explore a nomogram prediction model for perioperative heart failure (HF) in elderly patients with intertrochanteric femoral fracture (ITF) based on Lasso-Logistic regression. METHODS: The clinical data of 345 elderly patients with ITF admitted to our hospital were retrospectively collected (as the training set). In addition, the clinical data of 251 elderly patients with ITF admitted to Hunan University of Medicine General Hospital were collected (as the validation set). Patients were separated into the HF group and the non-HF group according to whether HF occurred during the perioperative period. The clinical data of patients were collected. RESULTS: According to Lasso-Logistic analysis, age, history of heart disease, Garden typing, preoperative electrolyte imbalance, 24 h positive balance of intake and output, and Hb were the influencing factors of HF occurrence during the perioperative period in elderly ITF patients (P < 0.05). The ROC curve showed that the AUC of this model was 0.886, the C-index was 0.857, the H-L test was χ 2 = 7.643, P = 0.711, and its predicted risk of HF occurrence was highly consistent with the actual result. The DCA curve showed a relatively good positive net benefit within the probability range of 0.15∼0.86. The external validation ROC curve showed that the AUC of this model was 0.911, the C-index was 0.834, the H-L test was χ 2 = 7.564, P = 0.705, and the predicted risk of HF was highly consistent with the actual result. The DCA curve showed a good positive net benefit within the probability range of 0.15∼0.73. CONCLUSION: The nomogram model constructed based on Lasso-Logistic regression in this study can assist clinicians in assessing the risk of HF in elderly ITF patients during the perioperative period, thereby implementing effective preventive measures.