Predicting death from ruptured abdominal aortic aneurysms

BACKGROUND: We have previously reported preoperative and immediate postoperative formulae to estimate mortality in patients with ruptured abdominal aortic aneurysms (rAAA). In this study, we prospectively compared these formulae in patients with rAAA with their actual outcomes. METHODS: Information was collected on 134 patients from two centers over a 3-year period. Preoperative mortality risk was estimated using coefficients for age, level of consciousness, and cardiac arrest. Mortality risk in the immediate postoperative state was based on the presence of coagulopathy, ischemic colitis, prolonged requirement for inotropes, time from arrival at hospital to surgery, patient age, perioperative myocardial infarction, renal failure, and pre-operative hemoglobin level. RESULTS: The average age was 73 years (range 30 to 92 y) and 20 of 134 (15%) patients were women. Sixty-three patients (47%) survived. For patients with a calculated preoperative mortality risk of >90%, the sensitivity, specificity, and positive and negative predictive values were 25%, 98%, 95%, and 54%, respectively. For a mortality risk >80%, these values were 37%, 94%, 87%, and 57%, respectively. For patients with an estimated immediate postoperative mortality risk > or = 90%, the sensitivity, specificity, and positive and negative predictive values were 17%, 87%, 60%, and 49%, respectively. For a predicted mortality > or = 80%, these values were 22%, 84%, 60%, and 50%, respectively. CONCLUSIONS: Our formula for predicting mortality for preoperative rAAA patients may be useful for patients with an estimated mortality risk >/=90%, based on the high positive predictive value. The formula for immediate postoperative rAAA patients was not useful in predicting death.