Frequency, Predictors, and Neurologic Outcomes of Vaso-occlusive Strokes Associated With Cardiac Surgery in Children

OBJECTIVE: Our aim was to define the frequency, predictors, and outcomes of stroke associated with cardiac surgery in children with congenital heart disease. METHODS: We performed a case-control study of children (term birth to 18 years) with congenital heart disease who underwent cardiac surgery at the Hospital for Sick Children between January 1, 1992, and March 1, 2001. Case subjects experienced stroke within 72 hours after cardiac surgery, and control subjects (2 for each case subjects) had cardiac surgery and no stroke. The frequency of arterial ischemic stroke/cerebral sinovenous thrombosis was calculated among children who underwent cardiac surgery during the study period. Predictors for stroke, including age, gender, simple versus complex procedure, reoperation, bypass duration, circulatory arrest, postoperative hematocrit level, and intraoperative activated clotting time, were tested. The presence of clinical and radiologically defined stroke was the main outcome. Neurologic outcomes were assessed in case subjects with the Pediatric Stroke Outcome Measure. RESULTS: During the study period, 30 children with stroke (28 with arterial ischemic stroke and 2 with cerebral sinovenous thrombosis) were identified among 5526 children undergoing cardiac surgery. This yielded a risk for arterial ischemic stroke/cerebral sinovenous thrombosis of 5.4 strokes per 1000 children undergoing a cardiac operation. Univariate analysis revealed that older age at the time of the procedure, longer duration of cardiopulmonary bypass, number of days in the hospital postoperatively, and reoperation were associated with stroke. In multivariate analyses, only reoperation was associated with stroke. CONCLUSIONS: The frequency of vaso-occlusive stroke in children with congenital heart disease undergoing cardiac surgery was 5.4 cases per 1000 children. Age, duration of bypass, and reoperation may be associated with stroke risk.