Background: Stroke is a major complication of coronary operation, with reported rates of postoperative cerebral dysfunction ranging from 0.4% to 13.8%. In this report, the incidence, correlates, outcomes, and costs of stroke in coronary operation were evaluated at Emory University between 1988 and 1996.
Methods: Data were entered prospectively into a dedicated computerized database at Emory University and analyzed retrospectively. Univariate and multivariate analyses were utilized where appropriate.
Results: Data from 10,860 patients undergoing primary coronary operation between 1988 and 1996 were analyzed. There were 250 patients not entered into the database. Stroke occurred in 244 (2.2%). Univariate predictors of stroke (p<0.05) included age, female gender, hypertension, diabetes, prior stroke, prior transient ischemic attack (TIA), and carotid bruits. Multivariate correlates included age (odds ratio 1.07) previous TIA (odds ratio 2.2), and carotid bruits (odds ratio 1.9), although the area under the Receiver Operating Characteristics (ROC) curve was only 0.69, suggesting limited ability to predict stroke. One and 5 year survival rates were 64% and 44% with stroke, and 94% and 81% without stroke, respectively. Among the stroke group, 23% of the patient population died before hospital discharge. The stroke group had a significantly longer length of hospital stay, as well as higher costs.
Conclusions: Stroke is a devastating complication of coronary operation, significantly increasing morbidity, mortality, and cost. Three independent variables were identified for predicting stroke, including age, previous TIA, and carotid bruits. Patients should be carefully screened for cerebrovascular disease to help prevent stroke and its associated morbidity.