P4574Predicting postoperative atrial fibrillation after cardiac surgery: validation and comparison of CHA2DS2-VASc, POAF and HATCH risk scoring systems

Abstract

Introduction: Postoperative atrial fibrillation (POAF) occurs in 20 to 40\% of patients following heart surgery. It leads to a longer hospital stay, and also entails a higher perioperative morbi-mortality. The POAF scores, CHA2DS2-VASc and HATCH were evaluated to predict AF in the postoperative period of cardiac surgery, and demonstrated a validated predictive value. There are no studies comparing and validating these scoring systems in Argentina. Purpose(s): The aim of the study was to validate and compare the performance of these scores in predicting the occurrence of AF in the postoperative period of cardiac surgery. Method(s): An analytical study of unicentric retrospective cohort was conducted, including consecutive postoperative cardiac surgery patients during the 2010- 2016 period. Patients with previous AF were excluded. The primary end point was POAF during in-hospital stay. The validity of the models was evaluated through its two components: discrimination through the area under the ROC curve (C statistic) and calibration using the Hosmer-Lemeshow (HL) test. We use the Youden index to establish the best cut-off point for each score. Each patient was assigned a CHA2DS2-VASc, POAF and HATCH score. CHA2DS2-VASc:Age {\textgreater}=75 2 points, 65-74 years 1 point; Woman 1 point, Heart failure 1 point, Hypertension, 1 point, Stroke/transient ischemic attack (TIA)/Thromboembolism History 2 points, dia-betes 1 point, Vascular Disease History 1 point. POAF: Age: 60-69: 1, 70-79: 2, {\textgreater}=80: 3, chronic obstructive pulmonary disease (COPD): 1; Glomerular Filtration Rate (GFR) {\textless}15 ml/min for 1.73 m3 or dialysis: 1; emergency operation: 1, preoperative intra-aortic balloon pump: 1, left ventricular ejection fraction {\textless}30\%: 1, valve surgery: 1 point. HATCH Hypertension 1 point, Age {\textgreater}=75 years: 1 point, Stroke/TIA: 2 points, COPD: 1 point, heart failure: 2 points. Result(s): During the study period 3113 patients underwent cardiac surgery: 45\% coronary artery bypass grafting (CABG), 24\% heart valve surgery, 15\% combined surgery (CABG and valve surgery) and 14\% other. New-onset POAF occurred in 654 patients (21\%). Median scores among patients with and without POAF were statistically higher in all scores (p{\textless}0,001). For predicting POAF, the CHAD2DS2-VASc score exhibited the highest discrimination as a continuous variable (c-statistic =0,77; 95\% confidence interval (CI): 0,75-0,79) while POAF and HATCH score had c-statistics of 0,71 95\% CI 0,69-0,73 and 0,70; 95\% CI: 0,67- 0,72, respectively. The HL test applied to all scores had a p{\textgreater}0.05, representing good calibration. Under univariate and multivariate analysis the three scores were significant predictors of POAF (P{\textless}0.05). Conclusion(s): These scoring systems present good discrimination and calibration to predict POAF. Amongst them, the CHA2DS2-Vasc score presents the highest discrimination of POAF risk, and has the advantage of being easily calculated, making it a very useful option.