Predicting New-Onset Postoperative Atrial Fibrillation in Cardiac Surgery Patients

Abstract

Objective: To derive a simple clinical prediction rule identifying patients at high risk of developing new-onset postoperative atrial fibrillation (POAF) after cardiac surgery.

Design: Retrospective analysis on prospectively collected observational data.

Setting: A university-affiliated cardiac hospital.

Participants: Adult patients undergoing coronary artery bypass grafting and/or valve surgery.

Interventions: Observation for the occurrence of new-onset postoperative atrial fibrillation.

Measurements and main results: Details on 28 preoperative variables from 999 patients were collected and significant predictors (p<0.2) were inserted into multivariable logistic regression and reconfirmed with recursive partitioning. A total of 305 (30.5%) patients developed new-onset POAF. Eleven variables were associated significantly with atrial fibrillation. A multivariable logistic regression model included left atrial dilatation, mitral valve disease, and age. Coefficients from the model were converted into a simple 7-point predictive score. The risk of POAF per score is: 15.0%, if 0; 20%, if 1; 27%, if 2; 35%, if 3; 44%, if 4; 53%, if 5; 62%, if 6; and 70%, if 7. A score of 4 has a sensitivity of 44% and a specificity of 82% for POAF. A score of 6 has a sensitivity of 11% and a specificity of 97%. Bootstrapping with 5,000 samples confirmed the final model provided consistent predictions.

Conclusions: This study proposed a simple predictive score incorporating three risk variables to identify cardiac surgical patients at high risk of developing new-onset POAF. Preventive treatment should target patients ≥ 65 years with left atrial dilatation and mitral valve disease.

Keywords: atrial fibrillation; cardiac surgery; postoperative complication; prediction rule.