Which risk score best predicts perioperative outcomes in nonvalvular atrial fibrillation patients undergoing noncardiac surgery?

Abstract

Patients with nonvalvular atrial fibrillation (NVAF) are at increased risk for adverse events after noncardiac surgery. The Revised Cardiac Index (RCI) is commonly used to predict perioperative events; however, the prognostic utility of NVAF risk scores (CHADS2, CHA2DS2-VASc, and R2CHADS2) has not been evaluated in patients undergoing noncardiac surgery. Using a population-based data set of NVAF patients (n = 32,160) who underwent major or minor noncardiac surgery between April 1, 1999, and November 30, 2009, in Alberta, Canada, we examined the incremental prognostic value of the CHADS2, CHA2DS2-VASc, and R2CHADS2 scores over the RCI using continuous net reclassification improvement (NRI). The primary composite outcome was 30-day mortality, stroke, transient ischemic attack, or systemic embolism. The median age was 73 years, 55.1% were male, 6.6% had a previous thromboembolism, 17% of patients underwent major surgery, and the median risk scores were as follows: RCI = 1, CHADS2 = 1, CHA2DS2-VASc = 3, and R2CHADS2 = 2. The incidence of our 30-day composite was 4.2% (mortality 3.3%; stroke, transient ischemic attack, or systemic embolism 1.2%); and c indices were 0.65 for the RCI, 0.67 for the CHADS2 (NRI 14.3%, P < .001), 0.67 for CHA2DS2-VASc (NRI 10.7%, P < .001), and 0.68 for R2CHADS2 (NRI 11.4%, P < .001). The CHADS2, CHA2DS2-VASc, and R2CHADS2 scores were also all significantly better than the RCI for mortality risk prediction (NRI 12.3%, 8.4%, and 13.3%, respectively; all Ps < .01). In NVAF patients undergoing noncardiac surgery, the CHADS2, CHA2DS2-VASc, and R2CHADS2 scores all improved the prediction of major perioperative events including mortality compared to the RCI.