Glycated hemoglobin and risk of first episode stroke in diabetic patients with atrial fibrillation: A cohort study.

Abstract

Diabetes is associated with an increased risk of stroke in patients with atrial fibrillation. Whether glycemic control, evaluated by glycated hemoglobin, affects the risk of stroke in diabetic patients with atrial fibrillation remains unknown. To examine the association between glycated hemoglobin and risk of first episode stroke in diabetic patients with atrial fibrillation. By using the computerized database of the largest health maintenance organization in Israel, we identified a cohort of adults (age ≥20 years) in whom atrial fibrillation was diagnosed before January 1, 2012. Eligible subjects had no previous stroke or transient ischemic attack and were not on anticoagulants at baseline. The cohort (37,358 subjects) was followed for the first episode stroke or transient ischemic attack up to December 31, 2012. A total of 1052 subjects had stroke during 35,278 person-years of follow-up (stroke rate 2.98 per 100 person-years). Cox proportional hazards regression analysis adjusting for CHA2DS2-VASc score risk factors revealed that compared with subjects without diabetes, the hazard ratio for stroke was 1.04 (95% confidence interval [CI] 0.83-1.30) for the lowest glycated hemoglobin quartile (<6.35%), 1.14 (95% CI 0.92-1.42) for the second quartile (6.35%-6.90%), 1.46 (95% CI 1.19-1.79) for the third quartile (>6.90%-7.70%), and 1.63 (95% CI 1.33-2.00) for the highest quartile(>7.70%) (for trend, P < .001). In diabetic patients (n = 11,176), the hazard ratio for stroke was 1.17 (95% CI 1.09-1.26) for every 1% increment in glycated hemoglobin level. The area under the receiver operating characteristic curve was 0.585 for the CHA2DS2-VASc score, which increased to 0.604 when glycated hemoglobin was included in the model (P = .038). Glycated hemoglobin is directly associated with stroke risk, and it improves the predictive accuracy for stroke in diabetic patients with atrial fibrillation.