Evidence from a nonrandomized study that coronary surgery prolongs survival in patients with two-vessel coronary disease.

Within the larger Seattle Heart Watch arteriography registry, surgically treated patients non randomly selected for direct myocardial revascularization were matched to medically treated patients such that each of the 287 pairs was identical in seven variables (ejection fraction, ventricular arrhythmia on resting electrocardiogram, age, heart murmur, stenosis of left main coronary artery greater than or equal to 50%, number of vessels with stenosis greater than or equal to 70%, and use of diuretics) previously demonstrated to be independently predictive of survival. Actuarial survival analyses based on cardiovascular deaths (average follow-up 3.5 years) indicate improved survival for the entire surgical matched pair cohort (p = 0.008) and for the surgically treated subgroup with two-vessel disease (p = 0.0002) when compared to the medical cohort. These results were confirmed by examination of the entire arteriography registry (n = 1524) in which these seven variables were known, using Cox's model for survival analysis. This multivariate, statistical technique indicated that the surgical mode of therapy was significantly predictive of improved survival in surgically treated patients for the entire registry (p = 0.008) and for the subgroup with two-vessel disease (p = 0.0005).