Background: Abdominal aortic aneurysm surgery is a major vascular procedure with a considerable risk of (mainly cardiac) mortality.
Objective: To estimate elective perioperative mortality, we developed a clinical prediction rule based on several well-established risk factors: age, gender, a history of myocardial infarction, congestive heart failure, ischemia on the electrocardiogram, pulmonary impairment, and renal impairment.
Methods: Two sources of data were used: (1) individual patient data from 246 patients operated on at the University Hospital Leiden (the Netherlands) and (2) studies published in the literature between 1980 and 1994. The Leiden data were analyzed with univariate and multivariate logistic regression. Literature data were pooled with meta-analysis techniques. The clinical prediction rule was based on the pooled odds ratios from the literature, which were adapted by the regression results of the Leiden data.
Results: The strongest adverse risk factors in the literature were congestive heart failure and cardiac ischemia on the electrocardiogram, followed by renal impairment, history of myocardial infarction, pulmonary impairment, and female gender. The literature data further showed that a 10-year increase in age more than doubled surgical risk. In the Leiden data, most multivariate effects were smaller than the univariate effects, which is explained by the positive correlation between the risk factors. In the clinical prediction rule, cardiac, renal, and pulmonary comorbidity are the most important risk factors, while age per se has a moderate effect on mortality.
Conclusions: A readily applicable clinical prediction rule can be based on the combination of literature data and individual patient data. The risk estimates may be useful for clinical decision making in individual patients.