Surgical survival in the Coronary Artery Surgery Study (CASS) registry

Abstract

The overall surgical survival data in the Coronary Artery Surgery Study (CASS) registry have not been published to date, pending the report of the randomized medical-surgical comparison (CASS randomized trial). Non-randomized surgical survival data from the CASS registry are given in this article. The overall medical survival data from the registry were reported previously as a natural history study. There were 8,991 patients in the registry portion of CASS who had primary isolated coronary artery bypass grafting and 8,971 with follow-up of more than 30 days. The 5-year survival for all 8,971 patients was 90%, and the operative mortality was 2.37%. Patients with left main coronary artery disease had an operative mortality of 3.84% and a 5-year survival of 85%, while patients with lesions in other vessels had an operative mortality of 2.12% and a 5-year survival of 91%. Among patients without left main coronary disease, the 5-year survival was 93% in those with single-vessel and 92% in those with double-vessel disease (operative mortality was 1.50% and 1.92%, respectively) and 88% in patients with triple-vessel disease (operative mortality was 2.62%; p = 0.009). When results for patients with left main coronary artery obstruction were compared with those for triple-vessel disease, the 5-year survival figures were 85% and 88%, respectively (p = 0.02) and the operative mortality, 3.84% and 2.62%, respectively (p = 0.03). Patients with normal or nearly normal left ventricular (LV) function (i.e., LV segmental wall motion scores ranging from 5 through 11) had a 5-year survival of 92% and an operative mortality of 1.97%. Patients with moderate impairment (LV score range, 12 through 16) had a 5-year survival of 80% and an operative mortality of 4.21%. In those with poor ventricular function (LV score of 17 or greater), the 5-year survival was 65% and the operative mortality was 6.21%. The difference in survival among the three groups was significant (p less than 0.0001). Of 29 variables used in a stepwise Cox regression analysis, LV wall motion score, congestive heart failure score, age, number of operable vessels, smoking history, LV end-diastolic pressure, and percent of left main coronary artery stenosis were found to have a significant effect on long-term survival (excluding 30-day mortality), and these variables plus surgical priority and height influenced surgical mortality. When height was used in the Cox proportional hazards model, female sex was no longer a significant variable.