Is Transfusion in Coronary Artery Surgery a Predictor or a Cause of Reduced Long-Term Survival?

Background: Transfusion is common after coronary bypass surgery. Transfused patients present with higher operative risk and increased hazard ratio for curtailed long-term survival. There is debate as to whether transfusion itself may further exacerbate late mortality. Methods: Long-term survival was studied in 2550 survivors following coronary revascularization in this retrospective, observational study. Kaplan-Meier survival curves were constructed to compare all transfused and non-transfused patients, as well as survival in propensity-matched transfused and non-transfused patients. Results: Operative mortality was 1.05% (original cohort 2577). Maximum follow-up was 23 years (mean 11.8, median 12.4 years). 34.7% of patients received a transfusion (mean 2 units pack red blood cells). Baseline risk characteristics (age, female gender, small body habitus, risk stratification scoring, diabetes, hypertension and reduced stroke volume) operative parameters (urgency and no internal thoracic graft) as well as post-operative parameters (intensive care, hospital stay and ventilation time) and complications (haemorrhage, intra-aortic balloon, ventricular arrhythmias, prolonged inotropic support, atrial fibrillation, dialysis, doubling of creatinine and resternotomy) were higher in the transfused patients. The long-term survival of these patients was significantly reduced when compared with that of non-transfused patients (log rank test p<0.001). When analyzed as a sole risk factor, transfusion was associated with reduced long-term survival (log rank test p<0.001) but when analyzed collectively with other risk factors, transfusion failed to demonstrate a causative effect (p=0.953). When propensity matched groups were compared (612 transfused versus 1222 non-transfused patients) long-term survival was similar (log rank test p=0.554). Conclusions: Transfusion was required in higher risk patients undergoing coronary revascularization. Long-term survival was curtailed in this group but this was due to preoperative risk and not directly to transfusion. Transfusion was a predictor but not a cause of reduced long-term survival.