A meta-analysis of randomized controlled trials in critically ill patients to evaluate the dose-response effect of erythropoietin.

The use of erythropoietin in critically ill patients has been investigated in multiple randomized clinical trials and its role in decreasing the number of units of blood transfused has been demonstrated in some trials. A meta-analysis was conducted to determine the pooled estimate of the decrease in number of units of blood transfused with the use of erythropoietin and investigated its dose-response effect. A systematic search was performed of the MEDLINE, EMBASE, and the Current Controlled Trials Register to identify randomized clinical trials investigating the role of erythropoietin in critically ill patients. Of 664 studies identified in the search, 5 randomized clinical trials met the inclusion criteria. The pooled estimate of the decrease of number of units of blood transfused was -1.64 (95% CI -2.6 to -0.67). Sensitivity analysis to establish the influence of temporal bias, quality of the study and comorbidities such as age and Acute Physiology and Chronic Health Evaluation (APACHE) II score were undertaken and did not reveal a significant difference. The inclusion of studies with higher doses of erythropoietin revealed a greater decrease in the number of units of blood transfused (-2.15; 95% CI -3.06 to -1.24). Despite the limitations of a meta-analysis we believe that the use of erythropoietin significantly decreases the number of units of blood transfused per patient. Our study also reveals the possibility of a dose-response effect of erythropoietin in decreasing the number of units of blood transfused.