Elderly patients with ischaemic heart disease are often treated more conservatively and for longer than younger patients, but this strategy may result in subsequent invasive intervention of more advanced and higher risk coronary disease. We performed a retrospective analysis of 109 patients aged > or = 70 years (mean age 74 years, 66% men), who presented with angina refractory to maximal medical treatment or unstable angina over a 2-year period (1988-1990), to compare the relative risks and benefits of myocardial revascularisation [coronary artery bypass grafting (CABG) and percutaneous transluminal coronary angioplasty (PTCA)] in this higher-risk age group. Sixty patients underwent CABG and 49 patients PTCA. There were eight periprocedural deaths in total (six in the CABG group, and two in the PTCA group, P = 0.29). Six patients in the CABG group suffered a cerebrovascular accident (two fatal). Acute Q-wave myocardial infarction occurred in one patient in the CABG group and in two patients in the PTCA group. The length of hospital stay was longer for the CABG group (CABG group 11.4 +/- 5.4 days, range 7-30 days, PTCA group 7.4 +/- 7.6 days, range 1-39 days, P = 0.01). Outcome was assessed using the major cardiac event rate (MACE; i.e. the rate of death, myocardial infarction, repeat CABG or PTCA). The cumulative event-free survival in the CABG group in 1, 2 and 3 years was 87, 85 and 85%, respectively. In contrast, in the PTCA group it was 55, 48 and 48% (P = 0.0001). Age, sex, number of diseased vessels, degree of revascularisation and left ventricular function were not predictive of the recurrence of angina in both groups. Actuarial survival (total mortality, including perioperative mortality) was lower at 1 year in the CABG group due to the higher perioperative mortality, but similar in both groups after the second year (P = 0.62). Elderly patients with refractory or unstable angina who are revascularised surgically have a better long-term outcome (less frequent event rate of the composite end-point--myocardial infarction, revascularisation procedures and death) compared with those who are revascularised with PTCA. This benefit is been realised after the second year. Total mortality is similar in both groups after the second year. Therefore elderly patients who are fit for surgery should not be denied the benefits of CABG. PTCA may be regarded as a complementary and satisfactory treatment, especially for those whose life expectancy is limited to less than 2 years. The use of stents may improve outcome in the PTCA group and this needs to be evaluated.