Abstract
OBJECTIVES: This study aims to research the clinical features of atrial thrombi in patients with nonvalvular atrial fibrillation (AF). METHODS: This study included 191 patients of AF who had atrial thrombi. One hundred and twenty-eight of them were assigned into nonventricular cardiomyopathy group (non-VCM), and the remaining 63 into ventricular cardiomyopathy group (VCM). After atrial thrombi diagnosed, all patients had taken oral anticoagulant therapy. The resolution rates of thrombi within 12 months were compared between the two groups, as well as the locations of thrombi. RESULTS: Of all 191 patients, 161 had thrombi only detected in left atrial appendage (LAA), 20 in both left atrium (LA) and LAA, six in LA only, and four in right atrium only. More patients had thrombi out of LAA in the VCM group than in the non-VCM group (30.2% vs 8.6%, P < .001). After propensity score matching, the atrial thrombi were resolved faster in the non-VCM group than in the VCM group (mean time length: 22 ± 2 weeks vs 30 ± 3 weeks, P = .038), and the resolution rate within 12 months was higher in the non-VCM group than in the VCM group (88.7% vs 61.4%, Log-rank, P = .038). In Cox proportional hazards model, absence of ventricular cardiomyopathy was an independent predictor for the resolution of atrial thrombus (hazard ratio: 1.76; P = .035). CONCLUSIONS: The patients of atrial fibrillation with ventricular cardiomyopathies have higher incidence of thrombosis in the body of left atrium or right atrium. And the resolution rate was lower in these patients.