Meta-analysis of epoetin beta and darbepoetin alfa treatment for chemotherapy-induced anemia and mortality: Individual patient data from Japanese randomized, placebo-controlled trials

Abstract

Erythropoiesis-stimulating agents (ESA) reduce the need for transfusions and improve the quality of life in patients receiving chemotherapy, but several clinical trials have suggested that ESA might have a negative impact on survival. To evaluate the efficacy and safety of ESA, epoetin beta and darbepoetin alfa, including their impact on overall survival and thromboembolic events, we conducted an individual data-based meta-analysis of three randomized, placebo-controlled trials studying Japanese patients with chemotherapy-induced anemia. All trials were conducted in compliance with Good Clinical Practice. A total of 511 patients with solid tumor or lymphoma (epoetin beta or darbepoetin alfa, n = 273; placebo, n = 238) were included. The ESA significantly reduced the risk of transfusion (relative risk, 0.47; 95% confidence interval, 0.29-0.76). No significant effect of the ESA on overall survival was observed (unadjusted hazard ratio, 1.00; 95% confidence interval, 0.75-1.34). A prespecified subgroup analysis showed no strong interaction between the baseline hemoglobin concentration and the effect of ESA on overall survival. Among the ESA-treated patients, the highest hemoglobin achieved during the treatment period in each patient had no impact on mortality. No increase in thromboembolic events was observed in the ESA-treated patients (0.7% vs 1.7% placebo). The ESA reduced the risk of transfusion without a negative impact on the survival of patients with chemotherapy-induced anemia.

Figure 1

Overall survival. Kaplan–Meier curve for all patients treated with erythropoiesis‐stimulating agents (ESA) or placebo. The median follow‐up period for overall survival of all patients was 13.3 months (min.–max., 1.1–19.4 months). The unadjusted hazard ratio for overall survival of all patients was 1.00 (95% confidence interval [CI], 0.75–1.34) with a 1‐year survival rate of 64% for the ESA‐treated group and placebo‐treated group.

Figure 2

Mortality risk stratified by trials and patient characteristics. Forest plots of the hazard ratio (HR) for overall survival from the three trials for all patients and stratified by trial and baseline characteristics. The position of each diamond indicates the HR estimate. Horizontal lines indicate the 95% confidence interval (CI). None of the variables (including cancer type and hemoglobin concentration) influenced patient mortality. *Test for interaction. ECOG, Eastern Cooperative Oncology Group.