Leukoaraiosis and stroke recurrence risk in patients with and without atrial fibrillation

Objective: We aimed to investigate the association between leukoaraiosis and long-term risk of stroke recurrence adjusting for clinical scores developed and validated for the prediction of stroke risk, such as CHADS(2) (congestive heart failure, hypertension, age >= 75 years, diabetes mellitus, and stroke or TIA) and CHA(2)DS(2)-VASc (congestive heart failure, hypertension, age >= 75 years, diabetes mellitus, stroke or TIA, vascular disease, age 65-74 years, sex category). Methods: Study population was derived from the Athens Stroke Registry and was categorized in 2 subgroups according to the presence of atrial fibrillation (AF). Cox proportional hazards analysis was performed to assess the independent predictors of stroke recurrence. To investigate whether leukoaraiosis adds to the prognostic accuracy of CHADS(2) and CHA(2)DS(2)-VASc scores, we used the likelihood ratio test. Overall model assessment was performed with Nagelkerke R-2 and Harrell C statistic. Kaplan-Meier analyses were also performed. Results: Among 1,892 patients, there were 320 (16.9%) with leukoaraiosis and 670 (35.4%) with AF. In the Kaplan-Meier analysis, there was significant difference in cumulative probability of stroke recurrence between patients with and without leukoaraiosis in the non-AF group (p < 0.01), but not in the AF group (p = 0.46). On Cox multivariate analysis, leukoaraiosis was found to be a significant independent predictor of stroke recurrence only in the non-AF group, in the models adjusting for CHADS(2) (hazard ratio: 1.86, 95% confidence interval: 1.35-2.56) and CHA(2)DS(2)-VASc (hazard ratio: 1.82, 95% confidence interval: 1.32-2.51) scores. Leukoaraiosis was not a predictor of stroke recurrence in the AF group. Leukoaraiosis did not improve the predictive accuracy of the 2 scores, whether in the non-AF group (Harrell C statistic: 0.56 vs 0.59 [p = 0.31] for the model including CHADS(2); 0.56 vs 0.59 [p = 0.44] for the model including CHA(2)DS(2)-VASc) or the AF group (Harrell C statistic: 0.63 vs 0.62 for the model including CHADS(2); 0.64 vs 0.64 for the model including CHA(2)DS(2)-VASc). Conclusions: Leukoaraiosis is an independent predictor of stroke recurrence in non-AF stroke patients. However, leukoaraiosis did not increase the accuracy of the CHADS(2) and CHA(2)DS(2)-VASc scores to predict stroke recurrence in AF or non-AF stroke patients.