Cardiac surgery in patients on dialysis:: Decreased 30-day mortality, unchanged overall survival

Background. The risk of cardiac surgery in dialysis-dependent patients is high, but little is known about the determinants of survival. We initiated a retrospective multicenter study to overcome this limitation. Methods. Nine centers provided data on 522 patients (70% male, aged 61 +/- 11 years) who had chronic dialysis-dependent renal failure. A 14-year period was covered. Most patients had coronary artery bypass grafting, either with (n = 103) or without (n = 326) valve surgery. Multivariable analysis of survival was explored using Cox models. Results. The proportion of patients with diabetes mellitus increased significantly (from 17%, 1989 to 1993, to 32%, 2000 to 2003; p = 0.021) and was independently associated with 30-day mortality (odds ratio = 3.30, p = 0.001) The mean 30-day mortality was 12% (n = 60), but declined significantly during the study period (from 28%, 1989 to 1993, to 7%, 2000 to 2003; p = 0.003). The 5-year survival probability was 42% (95% confidence interval: 36% to 47%). Patients who had renal transplantation during follow-up (n = 17) had the best survival probability (hazard ratio [HR] = 0.14, p = 0.007). Sinus rhythm (HR = 0.48, p < 0.001) and use of internal thoracic artery grafts (HR = 0.67, p = 0.006) proved beneficial for long-term survival. Predictors of death during long-term follow-up were emergency surgery (HR = 2.25, p = 0.001), diabetes mellitus (HR = 1.46, p = 0.020), number of allogenic transfusions (HR = 1.03/unit, p = 0.015), and age (HR = 1.04/year, p < 0.001). Conclusions. In dialysis-dependent patients, cardiac surgery has become significantly safer in recent years, but the overall prognosis of the patients remains poor. The observed improvements in the perioperative survival do not necessarily translate into an improved longterm prognosis. Diabetes mellitus is an important independent risk factor for perioperative mortality and death during follow-up.