A logistic regression model predicting postoperative atrial fibrillation after coronary artery bypass surgery yielded an AUC of 0.682 (95% CI 0.663-0.701; p<0.001), lacking sufficient clinical accuracy.
Cohort (n=4,783)
Can a logistic regression model accurately predict the risk of postoperative atrial fibrillation in patients undergoing coronary artery bypass grafting?
A logistic regression model based on standard clinical parameters is not sufficiently accurate to predict postoperative atrial fibrillation after CABG, though the occurrence of AF is significantly associated with worse clinical outcomes.
Effect estimate: AUC 0.682 (95% CI 0.663-0.701)
p-value: p=<0.001
BACKGROUND AND AIMS: The aim of this study was to determine the risk factors of postoperative atrial fibrillation (AF) after coronary artery bypass grafting and to create predictive model and to evaluate the effects of AF on patients outcome. MATERIAL AND METHODS: Data of 3,676 consecutive patients were analysed to identify the predictors of AF. Multivariate logistic regression model was validated prospectively in 1,107 patients. RESULTS: Increasing age (p < 0.001), preoperative use of digoxin (p = 003), need of intra-aortic balloon pump or inotropic medication in the weaning off cardiopulmonary by pass or during the first 24 hours postoperatively (p = 0.013), increasing body surface area (p = 0.006) and lower ejection fraction (p = 0.048) were independent risk factors for postoperative AF. The predictive model gave area under the receiver-operating characteristic (ROC) curve 0.682, 95% confidence interval 0.663-0.701, and p < 0.001. The patients with AF incidence had more postoperative stroke (p = 0.008), confusion (p < 0.001) severe gastrointestinal complications (p = 0.005), readmission to ICU (p < 0.001), longer ICU (p < 0.001) and hospital stay (p < 0.001) when compared with the patients who remained in sinus rhythm. CONCLUSION: Logistic regression model with the parameters used was not accurate enough for clinical purposes. Postoperative AF is associated with postoperative stroke, severe gastrointestinal complications, readmission to ICU, and longer ICU and hospital stay.
Hakala et al. (2002) conducted a cohort in Postoperative atrial fibrillation after coronary artery bypass grafting (n=4,783). Logistic regression predictive model was evaluated on Predictive accuracy of the logistic regression model for postoperative AF (Area under the ROC curve) (AUC 0.682, 95% CI 0.663-0.701, p=<0.001). A logistic regression model predicting postoperative atrial fibrillation after coronary artery bypass surgery yielded an AUC of 0.682 (95% CI 0.663-0.701; p<0.001), lacking sufficient clinical accuracy.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: