Allogeneic versus autologous blood transfusion and survival after radical prostatectomy.

ABSTRACT

BACKGROUND: Potential adverse effects of blood transfusion (BT) remain controversial, especially for clinical outcomes after curative cancer surgery. Some postulate that immune modulation after allogeneic BT predisposes to recurrence and death, but autologous superiority is not established. This study assessed whether BT is associated with long-term prostate cancer recurrence and survival with a large single-institutional radical prostatectomy (RP) database. STUDY DESIGN AND METHODS: Between 1994 and 2012, a total of 11,680 patients had RP with available outcome and transfusion data. A total of 7443 (64%) had complete covariate data. Clinical variables associated with biochemical recurrence-free survival (BRFS), cancer-specific survival (CSS), and overall survival (OS) were identified with Cox proportional hazards models for three groups no BT (reference, 27.7%, n = 2061), autologous BT only (68.8%, n = 5124), and any allogeneic BT (with or without autologous, 3.5%, n = 258). RESULTS: Median (range) follow-up was 6 (1-18) years. Kaplan-Meier analysis showed significantly decreased OS (but not BRFS or PCSS) in the allogeneic group versus autologous and no BT groups (p = 0.006). With univariate analysis, any allogeneic BT had a hazard ratio (HR) of 2.29 (range, 1.52-3.46; p < 0.0001) for OS, whereas autologous BT was not significant (HR, 1.04 [range, 0.82-1.32], p = 0.752). In multivariable models, neither autologous nor allogeneic BT was independently associated with BRFS, CSS, or OS, and a dose response was not observed for allogeneic units and BRFS. CONCLUSION: Although allogeneic but not autologous BT was associated with decreased long-term OS, after adjustment for confounding clinical variables, BT was not independently associated with OS, BRFS, or CSS regardless of transfusion type. Notably, no association was observed between allogeneic BT and cancer recurrence. Observed differences in OS may reflect confounding.