A large epidemiological data set was used to identify 115 patients who died from a cardiovascular cause within 30 days of elective surgery under general anaesthesia. For each patient, a control was identified, matched for age (within 5 yr of the patient), sex, operation and consultant. Patients and controls were compared for cardiovascular risk factors in a matched analysis using conditional logistic regression, and a prognostic model was generated. Three risk factors were included in the final model: previous myocardial infarction (odds ratio 3.18 (95% confidence intervals (CI) 1.22-8.28), P = 0.018), history of hypertension (odds ratio 1.90 (0.99-3.62), P = 0.047) and renal failure (odds ratio 3.56 (1.04-12.10), P = 0.043).