“Fairest of them all” for myelodysplastic syndromes

Raza and colleagues report that lenalidomide reduced red cell (RBC) transfusion requirements ≥ 50% in 43% of 214 patients with non–deletion 5q, primarily low-risk, myelodysplastic syndrome (MDS); 26% (95% CI, 20%-32%) became transfusion independent for a median of 41 weeks. The results, although not as dramatic as those noted with lenalidomide in similar patients with deletion 5q, are perhaps more significant, since most MDS patients do not have deletion 5q.

Medicine is fundamentally concerned with choosing between different therapeutic options. Several exist for patients like those treated by Raza et al. While epoetin or darbepoetin, possibly with granulocyte-colony stimulating factor, may produce prolonged transfusion independence (TI) in two-thirds of patients who require fewer than 2 units of red blood cells (RBCs) monthly and have serum erythropoietin (EPO) levels below 500,¹  only 35% of Raza et al's patients had such low transfusion requirements, and probably even fewer had such low EPO levels. A similarly small proportion is likely to respond to immunosuppressive therapy.²  A third option is a clinical trial. However, within the past 3 years, the Food and Drug Administration has approved lenalidomide, azacitidine, and decitabine for use in MDS, and many patients might prefer these options. Thus, practically speaking, physicians seeing patients with RBC transfusion–dependent low-risk non–deletion 5q MDS patients must decide which of these 3 drugs most effectively reduces the need for transfusions, which are typically these patients' principal clinical problem. However, attempts to assess which drug is the “fairest of them all” are complicated by the variability in the design, as well as the reports, of the relevant trials. For example, the literature suggests that azacitidine produced TI more frequently (45%; 95% CI, 32%-57%) and for as long as lenalidomide.³  But while a need for more than 4 units of red cells in the 8 weeks preceding lenalidomide (the median requirement) predicted a lower subsequent rate of TI, the azacitidine literature does not note pretransfusion requirements, so the nominally higher rate of TI might merely reflect the frequent inclusion of patients with low transfusion requirements. The effects of the 3 drugs on RBC transfusion needs may be similarly confounded by the much higher proportion of high-risk (International Prognostic Scoring System [IPSS] int-2 or high) patients in the azacitidine³  and decitabine⁴  studies, reports of which do not analyze rates of TI by IPSS score. On the other hand, the latter 2 trials included a randomized, supportive-care-only group, thus allowing the occasional “spontaneous” response and “toxicity” due to disease rather than treatment to be taken into account.

My first choice would be lenalidomide, simply because median time to TI was 5 weeks, compared with 3 to 6 months for azacitidine or decitabine. But the practicing physician can be forgiven for wishing for a trial randomizing among lenalidomide, azacitidine, and decitabine to address the issues noted above. Such a trial is unlikely to be conducted, and perhaps there are greater priorities. Nonetheless, much as the queen in “Snow White” needed a mirror to determine the “fairest of them all,” perhaps we need a mirror in the form of standardized reporting requirements from single-arm clinical trials in this population.