Transfusion reactions: a comparative observational study of blood components produced before and after implementation of semiautomated production from whole blood

BACKGROUND: A semiautomated method of component production from whole blood was implemented at Canadian Blood Services. To assess safety of the new components, the frequency of adverse transfusion events (ATEs) to platelet components (PCs) and red blood cell (RBCs) produced before and after implementation of the new method was surveyed and compared. STUDY DESIGN AND METHODS: This retrospective, observational, noninferiority study was conducted in 12 sentinel hospitals across Canada. The control group received RBCs in additive solution-3 (AS-3) and platelet-rich plasma (PRP)-produced platelets (PLTs) for 3 to 11 months before implementation of semiautomated production, and the study group received RBCs in saline-adenine-glucose-mannitol (SAGM) and buffy coat (BC)-produced PLTs for 3 to 11 months after implementation. ATE definitions at each hospital and standard practice for reporting did not change between control and study periods. Data for analysis were obtained from databases and original report forms. RESULTS: The pooled risk ratio of a reaction to SAGM versus AS-3 RBCs was 0.77 (95% confidence interval [CI], 0.66-0.90), suggesting that SAGM products had significantly lower reaction rates than AS-3 products (p < 0.01). Reported allergic reactions to RBCs decreased from 0.07% (AS-3) to 0.04% (SAGM). For PLTs, the difference in reaction rates between BC and PRP was not significant (p = 0.37), and the pooled risk ratio of BC versus PRP was 1.14 (95% CI, 0.86-1.50). CONCLUSION: The change in manufacturing method was associated with lower reaction rates to SAGM RBCs than to AS-3 RBCs. Pooled BC PLTs were noninferior to random-donor PRP PLTs with respect to ATEs.