Erythropoiesis-stimulating agents in myelodysplastic syndromes

Transfusion-dependent anemia is a very common feature of myelodysplastic syndromes (MDS), and often the main clinical problem in low-risk subtypes of the disease. Chronic transfusion need is associated with reduced quality of life, not only due to the low and undulating hemoglobin levels in these patients but also because of constrictions in lifestyle, frequent visits to the transfusion unit, and inability to travel freely. Moreover, chronic transfusion therapy will eventually lead to iron overload and a need for iron chelation treatment in a subset of patients [1].

None of the ‘erythropoietins’ is licensed for treatment of MDS-related anemia in any part of the world. The main reasons are that off-license use emerged early as a common therapeutic approach in countries where funding was possible, and that no study group or pharmaceutical company made the effort to perform a randomized controlled study with endpoint of survival or freedom from transfusions. The consequence is that a number of countries deny reimbursement of these drugs, in spite of the fact that both the European Medicines Agency (EMEA) and Food and Drug Administration (FDA) require the use of erythropoietins in the control arm of many clinical trials. Nonetheless, a large number of phase II trials have shown response rates to erythropoietin (EPO) alone of 20–40%, higher in more recent studies with patients treated closer to diagnosis [1]. Darbepoetin (DA) has shown at least equal and possibly better effects, with responses observed also in patients with no response to conventional EPO [2]. Furthermore, the effect of EPO can be enhanced by the addition of granulocyte-colony stimulating factor (G-CSF), which has a synergistic effect on response rate [2,3,4]. The addition of G-CSF is particularly effective in refractory anemia with ring sideroblasts (RARS, including refractory cytopenia with multilineage dysplasia and ringed sideroblasts (RCMD-RS)), and acts through the suppression of apoptosis of marrow erythroblasts [1,2].

The identification of subsets of patients with higher probabilities of responding to growth factor treatment is mandatory for clinical decision-making and a sound health economy. Serum EPO and the degree of red blood cell (RBC) transfusion need have in combination shown the strongest association with response to treatment in most studies. A predictive model for the response to recombinat human EPO (rHuEPO) + G-CSF in MDS was developed from two large phase II trials, and then validated in a prospective study [5]. This model identifies patients with serum EPO levels ≤500 U/L and a transfusion need of <2 units/month as having a high probability (74%) of response to treatment, while patients with EPO levels of >500 U/L and ≥2 units/month are likely to have a very poor response (7%). The identification of non-responders is important also because these patients show a poorer overall survival and a higher risk for progression to acute myeloid leukemia (AML), and may be subject to other treatment approaches, such as azacytidine or stem cell transplant [6].

While there is no doubt about the response rates obtained by EPO ± G-CSF, the effect of treatment on the long-term outcome of patients has been questioned. In particular, there has been a fear that growth factor treatment may enhance the risk for leukemic transformation. Recently, three major publications have, however, indicated a favorable effect on long-term outcome. The French Group compared a large cohort of EPO ± G-CSF treated patients with an International Prognostic Scoring System (IPSS) cohort and showed that responding patients survived significantly better than the untreated IPSS cohort [7]. The Nordic group compared 129 EPO–G-CSF treated patients with a similar but untreated cohort from the Pavia registry, and showed in multivariate analysis using treatment as a covariate that the Nordic cohort had a significantly better survival (p < 0.001) than the Italian cohort, irrespective of response to treatment [8]. The survival difference was only observed in patients with a moderate (<2 units/month) pretreatment transfusion need. Finally, a US randomized phase III study comparing EPO–G-CSF with supportive care confirmed a significantly better response rate in the treated arm and better survival in responding patients [9]. There was no overall effect on survival and transformation, but this part of the study was hampered by the cross-over design.

Taken together, these data suggest that treatment of anemia in MDS with EPO and G-CSF in selected patients is safe and has a positive impact on survival, although definite scientific proof would demand a prospective randomized study. Whether the effect on survival is mediated by limiting iron overload and other adverse effects of transfusions, or counteracting the negative effects of chronic severe anemia in itself, remains to be investigated. With this background, it is important to further explore the potentially positive effects of pro-erythroid growth factors in MDS, both on survival and on quality of life. Oliva et al. report in the current issue of Leukemia and Lymphoma on 41 patients receiving darbepoetin treatment for MDS-related anemia [10]. The response rate in both transfusion-independent patients and those with a moderate transfusion need was relatively high, >50%, and the study corroborates previous results that treatment is associated with decreased apoptosis of bone marrow progenitors. In addition, the study evaluated quality of life, and showed clear improvement over time in responding patients. Importantly, changes in transfusion requirements correlated with changes in functional well-being. The study adds to the current knowledge about the effects of erythropoietins in MDS and may further support the process to get them officially approved as treatment for these disorders.