Between 1970 and 1985, 1801 patients underwent coronary-artery surgery without associated valvular surgery. Eighty-four per cent of patients were male and the mean (+/- SD) age was 55.7 +/- 8.3 years. Of the patients, 18.7% were from the Coronary Care Unit and 6.5% had diabetes. The hospital mortality rate for the whole group was 3.5%. Patients from the Coronary Care Unit had the highest (8.9%) hospital mortality rate compared with those patients who were not from the Coronary Care Unit (2.1%; P less than 0.001). Other factors which increased the hospital mortality rate significantly were the number of diseased vessels (P less than 0.01), the degree of left main coronary-artery stenosis (P less than 0.001), an earlier year of surgery (P less than 0.01) and female sex (P less than 0.01). After these were taken into account, no other factors (for example, age, preoperative infarction, presence of left-ventricular aneurysm, left-ventricular end-diastolic pressure, diabetes, use of mammary-arterial grafts or the need for endarterectomy) affected the mortality rate. Patients were followed-up for a mean (+/- SD) of 4.4 +/- 2.8 years. The five-year survival rate for all patients was 88% and the 10-year survival rate was 65%. Cox regression analysis showed that the significant indicators of decreased long-term survival were undergoing operation directly from the Coronary Care Unit (P less than 0.001), left main coronary-artery stenosis (P less than 0.01), the number of grafted vessels (P less than 0.01), concomitant surgery for aneurysm (P less than 0.001), year of surgery (P less than 0.01). Seventy-nine per cent of patients were free of angina pectoris at five years after operation. The year of surgery (P less than 0.001) and preoperative myocardial infarction (P less than 0.05) were the best predictors of recurrent angina. In the long term, recurrent angina remains a problem, although this may change with the increased use of mammary-arterial grafts.