We carried out a systematic overview using individual patient data from the seven randomised trials that have compared a strategy of initial coronary artery bypass graft (CABG) surgery with one of initial medical therapy to assess the effects on mortality in patients with stable coronary heart disease (stable angina not severe enough to necessitate surgery on grounds of symptoms alone, or myocardial infarction). 1324 patients were assigned CABG surgery and 1325 medical management between 1972 and 1984. The proportion of patients in the medical treatment group who had undergone CABG surgery was 25% at 5 years, 33% at 7 years, and 41% at 10 years: 93.7% of patients assigned to the surgery group underwent CABG surgery. The CABG group had significantly lower mortality than the medical treatment group at 5 years (10.2 vs 15.8%; odds ratio 0.61 [95% CI 0.48-0.77], p = 0.0001), 7 years (15.8 vs 21.7%; 0.68 [0.56-0.83], p < 0.001), and 10 years (26.4 vs 30.5%; 0.83 [0.70-0.98]; p = 0.03). The risk reduction was greater in patients with left main artery disease than in those with disease in three vessels or one or two vessels (odds ratios at 5 years 0.32, 0.58, and 0.77, respectively). Although relative risk reductions in subgroups defined by other baseline characteristics were similar, the absolute benefits of CABG surgery were most pronounced in patients in the highest risk categories. This effect was most evident when several prognostically important clinical and angiographic risk factors were integrated to stratify patients by risk levels and the extension of survival at 10 years was examined (change in survival -1.1 [SE 3.1] months in low-risk group, 5.0 [4.2] months in moderate-risk group, and 8.8 [5.4] months in high-risk group; p for trend < 0.003). A strategy of initial CABG surgery is associated with lower mortality than one of medical management with delayed surgery if necessary, especially in high-risk and medium-risk patients with stable coronary heart disease. In low-risk patients, the limited data show a non-significant trend towards greater mortality with CABG.