Outcomes/Effectiveness Research

Risk models can help predict medical and surgical complications of carotid endarterectomy

Carotid endarterectomy (CEA), the surgical removal of plaque from the carotid artery so that it can better supply oxygenated blood to the brain, is performed to reduce the risk of stroke. Ethan A. Halm, M.D., M.P.H., of the Mount Sinai School of Medicine in New York, and colleagues compared the predictive ability of 6 risk indexes to predict complications among a group of 1,998 patients who underwent CEA at 6 hospitals. They evaluated the ability of four cardiac risk indexes—the Goldman, Detsky, and Revised Cardiac Risk (RCR) Indexes and the American Society of Anesthesiologists index, as well as two CEA-specific risk models (the Halm and Tu scores) to predict a broad range of complications within 30 days of surgery.

Overall, 3.2 percent of patients died or suffered a stroke, 4 percent developed cardiac complications, 3.2 percent developed noncardiac medical complications, 6.9 percent suffered from minor neurologic complications, and 6 percent developed wound complications. Patients with cardiac, noncardiac medical, minor neurologic, or wound complications had 3- to 16-fold greater likelihood of dying or suffering a stroke.

Overall, a CEA-specific risk model (Halm score) had the best overall performance. It was superior for predicting the most important complications of death and stroke. It also predicted all other medical, neurological and surgical complications. Among the generic cardiac risk assessment tools, all cardiac indices predicted cardiac outcomes equally, and the RCR index performed best in predicting the broad range of major and minor complications of CEA. The study was supported by the Agency for Healthcare Research and Quality (HS09754).

See "Predicting medical and surgical complications of carotid endarterectomy," by Matthew J. Press, M.D., Mark R. Chassin, M.D., M.P.H., Jason Wang, Ph.D., and others, in the April 24, 2006, Archives of Internal Medicine 166, pp. 914-920.

Return to Contents
Proceed to Next Article