Comparisons of the effects on survival after coronary artery surgery in subgroups of patients from the Seattle Heart Watch.

We compared the survival of medically and surgically treated coronary artery disease patients in subgroups of patients to determine those most or least likely to benefit from surgery after an average of 5.5 years of follow-up. Cox's regression model for survival analysis was used in conjunction with data from all patients to estimate and test for the significance of the effects of surgery on survival in subgroups of patients, defined by one or more of the following variables: number of stenotic vessels (greater than or equal to 70%), ejection fraction, age, heart murmur, diuretic therapy, ventricular arrhythmia on resting ECG, left main coronary artery stenosis greater than or equal to 50%, previous myocardial infarction, cardiomegaly, congestive heart failure, unstable angina, and functional class. The Cox model adjusts for differences between medical and surgical patients in variables shown to be predictive of survival. A statistically beneficial effect of surgery on survival was seen in patients with two- or three-vessel disease, ejection fraction greater than or equal to 30%, age greater than or equal to 48 years, no heart murmur, no diuretic therapy, no ventricular arrhythmia on resting ECG, left main coronary artery stenosis less than 50%, no cardiomegaly, and no congestive heart failure. The converse subgroups defined by these variables did not show a significant beneficial effect from surgery. However, patient subgroups defined by presence or absence of prior myocardial infarction or unstable angina and New York Heart Association functional class I-II vs III-IV all showed beneficial effects from surgery.