Is electrical cardioversion independently associated with infarcts on brain magnetic resonance imaging or clinical outcomes in patients with atrial fibrillation?

Background Electrical cardioversion (ECV) is frequently performed in symptomatic atrial fibrillation. Objective This study aimed to assess the association of ECV with infarcts on brain magnetic resonance imaging (bMRI) and clinical outcomes. Methods The Swiss Atrial Fibrillation Cohort Study included 2386 patients; 1731 patients were evaluated by bMRI. ECVs were recorded by questionnaire. Patients were assigned to categories by number of ECVs performed before enrollment (0, 1, >= 2). A bMRI study was conducted at baseline and after 2 years (n = 1227) and analyzed for large noncortical or cortical infarcts and small noncortical infarcts. Clinical outcomes were recorded during follow-up. Associations of ECV and outcome measures were assessed by multivariate analyses. Results There was no independent association between the number of ECVs and infarct prevalence (large noncortical or cortical infarcts and small noncortical infarcts) on baseline bMRI (ECV 1 vs 0: odds ratio [OR], 0.95 [95% CI, 0.68-1.24]; ECV >= 2 vs 0: OR, 1.04 [0.72-1.44]) or between ECVs performed during follow-up and new infarcts on bMRI at 2 years (OR, 1.46 [0.54-3.31]). ECVs were not associated with overt stroke or transient ischemic attack (ECV 1 vs 0: hazard ratio [HR], 1.36 [0.88-2.10]; ECV >= 2 vs 0: HR, 1.53 [0.94-2.48]), hospitalization for heart failure (ECV 1 vs 0: HR, 1.06 [0.82-1.37]; ECV >= 2 vs 0: HR, 1.03 [0.77-1.38]), or death (ECV 1 vs 0: HR, 0.90 [0.70-1.15]; ECV >= 2 vs 0: HR, 0.91 [0.69-1.20]). Conclusion There was no association between ECV performed before enrollment and cerebral infarcts on baseline bMRI or between ECV performed during follow-up and new infarcts at 2 years. Moreover, ECV was not associated with clinical events.