Abstract
INTRODUCTION: Atrial fibrillation is the most common atrial arrhythmia in the perioperative period and is associated with prolonged hospital stay, increased costs, and increased mortality. However, there are few data on the predictors and incidence of preoperative atrial fibrillation in hip fracture patients. Our aim was to identify predictors of preoperative atrial fibrillation and to propose a valid clinical prediction model. METHODS: Predictor variables included demographic and clinical variables. LASSO regression analyzes were performed to identify predictors of preoperative atrial fibrillation, and models were constructed and presented as nomograms. Area under the curve, calibration curve, and decision curve analysis (DCA) were used to examine the discriminative power, calibration, and clinical efficacy of the predictive models. Bootstrapping was used for validation. RESULTS: A total of 1415 elderly patients with hip fractures were analyzed. Overall, 7.1% of patients had preoperative atrial fibrillation, and they were at significant risk for thromboembolic events. Patients with preoperative AF had a significantly longer delay in surgery than those without preoperative atrial fibrillation (p < 0.05). Predictors for preoperative atrial fibrillation were hypertension (OR 1.784, 95% CI 1.136-2.802, p < 0.05), C-reactive protein at admission (OR 1.329, 95% CI 1.048-1.662, p < 0.05), systemic inflammatory response index at admission (OR 2.137, 95% CI, 1.678-2.721 p < 0.05), Age-Adjusted Charlson Comorbidity Index (OR 1.542, 95% CI 1.326-1.794, p < 0.05), low potassium(OR 2.538, 95% CI 1.623-3.968, p < 0.05), anemia(OR 1.542, 95% CI 1.326-1.794, p < 0.05). Good discrimination and calibration effect of the model was showed. Interval validation could still achieve the C-index value of 0.799. DCA demonstrated this nomogram has good clinical utility. CONCLUSION: This model has a good predictive effect on preoperative atrial fibrillation in elderly patients with hip fractures, which can help to better plan clinical evaluation.