Stroke after coronary artery bypass grafting: preoperative predictive accuracies of CHADS2 and CHA2DS2VASc stroke risk stratification schemes.

OBJECTIVE

Neurologic events after coronary artery bypass grafting are an infrequent but devastating complication. This study analyzed the preoperative predictive abilities of the CHADS(2) and CHA(2)DS(2)VASc stroke scores in patients undergoing isolated coronary artery bypass grafting.

METHODS

Included in the study were 2910 patients who underwent isolated coronary artery bypass grafting during a 19-year period. CHADS(2) and CHA(2)DS(2)VASc scores were computed for all patients, and outcomes were evaluated in terms of perioperative stroke and compared with 2 specific models for predicting surgical coronary artery bypass grafting stroke (Northern New England Cardiovascular Disease Study Group and Multicenter Study of Perioperative Ischemia Research Group). Perioperative stroke discrimination was quantified by computing the area under the receiver operating characteristic curve.

RESULTS

Overall, 62 (2.1%) had perioperative strokes. Areas under the curve were 0.71 (95% confidence interval, 0.64-0.78) for CHADS(2), 0.72 (95% confidence interval, 0.65-0.79) for CHA(2)DS(2)VASc, 0.69 (95% confidence interval, 0.61-0.76) for Northern New England Cardiovascular Disease Study Group, and 0.73 (95% confidence interval, 0.67-0.80) for Multicenter Study of Perioperative Ischemia Research Group scores. Northern New England Cardiovascular Disease Study Group and CHA(2)DS(2)VASc scores were better at discriminating patients with particularly low or high risk of stroke.

CONCLUSIONS

CHADS(2) and CHA(2)DS(2)VASc scores predicted perioperative coronary artery bypass grafting strokes with discriminatory abilities similar to those of specific predictive surgical coronary artery bypass grafting stroke models. All schemes tested showed similar limitations in discriminating patients with high postoperative stroke risk, with a high proportion being classified as having intermediate stroke risk.