Treatment of anemia in myelodysplastic syndromes with granulocyte colony-stimulating factor plus erythropoietin: results from a randomized phase II study and long-term follow-up of 71 patients.

Treatment with erythropoietin (epo) may improve the anemia of myelodysplastic syndromes (MDS) in approximately 20% of patients. Previous studies have suggested that treatment with the combination of granulocyte colony-stimulating factor (G-CSF) and epo may increase this response rate. In the present phase II study, patients with MDS and anemia were randomized to treatment with G-CSF + epo according to one of two alternatives; arm A starting with G-CSF for 4 weeks followed by the combination for 12 weeks, and arm B starting with epo for 8 weeks followed by the combination for 10 weeks. Fifty evaluable patients (10 refractory anemia [RA], 13 refractory anemia with ring sideroblasts [RARS], and 27 refractory anemia with excess blasts [RAEB]) were included in the study, three were evaluable only for epo as monotherapy and 47 for the combined treatment. The overall response rate to G-CSF + epo was 38%, which is identical to that in our previous study. The response rates for patients with RA, RARS, and RAEB were 20%, 46%, and 37%, respectively. Response rates were identical in the two treatment groups indicating that an initial treatment with G-CSF was not neccessary for a response to the combination. Nine patients in arm B showed a response to the combined treatment, but only three of these responded to epo alone. This suggests a synergistic effect in vivo by G-CSF + epo. A long-term follow-up was made on 71 evaluable patients from both the present and the preceding Scandinavian study on G-CSF + epo. Median survival was 26 months, and the overall risk of leukemic transformation during a median follow-up of 43 months was 28%. Twenty patients entered long-term maintenance treatment and showed a median duration of response of 24 months. The international prognostic scoring system (IPSS) was effective to predict survival, leukemic transformation, and to a lesser extent, duration of response, but had no impact on primary response rates.