Atrial Fibrillation Prediction Model Following Aortic Valve Replacement Surgery.

Abstract

(1) Background: Postoperative atrial fibrillation (POAF) is the most common complication following cardiac surgery. It leads to increased perioperative morbidity and costs. Our study aimed to determine the incidence of new-onset POAF in patients undergoing isolated aortic valve replacement (AVR) and develop a multivariate model to identify its predictors. (2) Methods: We conducted a retrospective study including all consecutive patients who underwent isolated AVR at our institution between January 2010 and December 2022. Patients younger than 18, with a history of atrial fibrillation, previous cardiac surgery, or those who underwent concomitant procedures were excluded. Patients were dichotomized into POAF and No POAF groups. Multivariate logistic regression with backward elimination was utilized for predictive modeling. (3) Results: This study included 1108 patients, of which 297 (27%) developed POAF. The final multivariate model identified age, larger valve size, cardiopulmonary bypass time, delayed sternal closure, ventilation time, and intensive care unit stay as predictors of POAF. The model exhibited fair predictive ability (AUC = 0.678, p < 0.001), with the Hosmer-Lemeshow test confirming good model fit (p = 0.655). The overall correct classification percentage was 65.6%. (4) Conclusions: A POAF prediction model offers personalized risk estimates, allowing for tailored management strategies with the potential to enhance patient outcomes and optimize healthcare costs.