OBJECTIVETo investigate the role of left atrial appendage (LAA) closure for cerebral ischemic stroke prevention following mitral valve replacement.

METHODSRetrospective data on 860 consecutive adult patients undergoing mitral valve replacement between January 2008 and January 2013 were analyzed. There were 414 male and 446 female patients, with a mean age of (53 ± 12) years. The patients were divided into two groups according to whether the left atrial appendage was closed during operation: LAA closure group (n = 521) and non-LAA closure group (n = 339).Early mortality, postoperative cerebral ischemic stroke and the risk factors for cerebral ischemic stroke were assessed. Multivariate analysis was performed using logistic regression analysis.

RESULTSCompared with non-LAA closure group, LAA closure group had higher proportion of female gender, higher percentage of patients with cardiac insufficiency, pulmonary hypertension and left atrial thrombus, higher incidence of mechanical valve implantation and concurrent tricuspid surgery, and larger preoperative diameter of left atrium, but lower proportion of hypertension and patients undergoing coronary artery bypass surgery, and shorter aorta cross clamping time (χ² = 6.807 to 122.576, t = -2.818 and 3.756, all P < 0.05). There were no differences in exploratory thoracotomy for bleeding and in-hospital mortality between the two groups. Postoperative cerebral ischemic stroke occurred in 12 patients (1.4%). The incidence of cerebral ischemic stroke in LAA closure group was significantly lower than in non-LAA closure group (0.6% vs.2.7%, χ² = 6.452, P = 0.011).Logistic regression analysis showed that LAA closure was a significant protective factor for postoperative cerebral ischemic stroke (OR = 0.189, 95% CI: 0.039 to 0.902, P = 0.037) while history of cerebrovascular disease (OR = 4.326, 95% CI:1.074 to 17.418, P = 0.039) and preoperative diameter of left atrium (OR = 1.509, 95% CI: 1.022 to 1.098, P = 0.002) being the independent risk factors for postoperative cerebral ischemic stroke. The subgroup analysis showed that, for atrial fibrillation patients, LAA closure was a strong protective factor (OR = 0.064, 95% CI: 0.006 to 0.705, P = 0.025), but LAA closure was not a significant predictive factor (OR = 1.902, 95% CI: 0.171 to 21.191, P = 0.601) in non-atrial fibrillation patients.

CONCLUSIONConcurrent LAA closure during mitral valve replacement is safe and effective to reduce the early postoperative risk of cerebral ischemic stroke in atrial fibrillation patients.