Meta-analysis of clinical studies of the efficacy of granulocyte transfusions in the treatment of bacterial sepsis.

Abstract

Meta-analysis was used to explain disagreements across controlled clinical studies of the efficacy of granulocyte transfusions (GTX) in the treatment of bacterial sepsis. Studies published in English in 1970-1994 were retrieved. Seven studies of adults and five of neonates were eligible for analysis. Summary relative odds (RR) of survival in treated patients versus controls were computed for patient subsets defined on the basis of microbiologic proof of infection, recovery of bone marrow function, method of procurement of granulocytes for transfusion, dose of granulocytes transfused, assessment of leukocyte compatibility, and survival rate of controls. The random-effects method was used for all analyses. Differences between the reviewed studies in the dose of granulocytes transfused and the survival rate of controls were primarily responsible for the disagreements across the reports. Adults (RR = 4.2) and neonates (RR = 18.0) receiving adequate doses of granulocytes and adults transfused in centers with a low survival rate of controls (RR = 8.9) experienced a significant (P < .05) benefit from GTX. GTX of adequate dose may be indicated in the 1990s for the treatment of sepsis in neonates, and perhaps also adults admitted to centers with an unusually high mortality rate of untransfused controls. More research is needed to reassess the proper role of GTX, in the light of modern transfusion medicine technology and the presently available options for the treatment of sepsis.