Treatment of myelodysplastic syndrome patients with erythropoietin with or without granulocyte colony-stimulating factor: results of a prospective randomized phase 3 trial by the Eastern Cooperative Oncology Group (E1996)

Abstract

This phase 3 prospective randomized trial evaluated the efficacy and long-term safety of erythropoietin (EPO) with or without granulocyte colony-stimulating factor plus supportive care (SC; n = 53) versus SC alone (n = 57) for the treatment of anemic patients with lower-risk myelodysplastic syndromes. The response rates in the EPO versus SC alone arms were 36% versus 9.6%, respectively, at the initial treatment step, 47% in the EPO arm, including subsequent steps. Responding patients had significantly lower serum EPO levels (45% vs 5% responses for levels < 200 mU/mL vs > or = 200 mU/mL) and improvement in multiple quality-of-life domains. With prolonged follow-up (median, 5.8 years), no differences were found in overall survival of patients in the EPO versus SC arms (median, 3.1 vs 2.6 years) or in the incidence of transformation to acute myeloid leukemia (7.5% and 10.5% patients, respectively). Increased survival was demonstrated for erythroid responders versus nonresponders (median, 5.5 vs 2.3 years). Flow cytometric analysis showed that the percentage of P-glycoprotein(+) CD34(+) marrow blasts was positively correlated with longer overall survival. In comparison with SC alone, patients receiving EPO with or without granulocyte colony-stimulating factor plus SC had improved erythroid responses, similar survival, and incidence of acute myeloid leukemia transformation.

Figures

Figure 1

Overall survival of patients randomized to the supportive care alone (arm A, n = 57) and EPO/G-CSF (arm B, n = 53) arms. No statistically significant difference in overall survival between these 2 groups of patients was demonstrated (P = .28).

Figure 2

Forest plot of hazard ratios (EPO/SC) for overall survival related to treatment for clinical subgroups. A univariate Cox proportional hazard model, stratified on randomization features, estimated hazard ratios, and significance for overall survival. The horizontal lines provide the 95% confidence interval for the ratios. The dotted vertical line represents the overall hazard ratio (0.77).

Figure 3

Time to leukemic transformation of patients randomized to the SC alone (arm A, n = 57) and EPO/G-CSF (arm B, n = 53) arms. No statistically significant difference in time to transformation between these 2 groups of patients was demonstrated (P = .83).

Figure 4

Cumulative incidence of leukemic transformation, with death as a competing risk event. No statistically significant difference in time to transformation between these 2 groups of patients was demonstrated when the competing risk of death was included (P = .34, cumulative incidence analysis).