Treating patients of myelodysplastic syndrome with antithymocytic globulin—should we be more selective?

Madkaikar and Ghosh raise important points about the cost and toxicity of antithymocyte globulin (ATG) and thereby underscore one of the conclusions of our recently terminated clinical trial.¹  However, we feel that the use of ATG and similar therapies for some patients with myelodysplastic syndrome (MDS) should not be dismissed prematurely. While the series of ATG-treated MDS patients recently reported by Molldrem et al (the largest to date) may include individuals atypical of those commonly encountered in community clinical practice (eg, selected by their motivation to participate in a clinical trial and ability to travel to a referral center for therapy), the response rate observed (34% red cell transfusion independence, with improvement in approximately half of patients with severe thrombocytopenia and neutropenia) suggests that there is indeed a subset of MDS patients whose ineffective hematopoiesis is at least partially immune mediated.²  Such patients may attain palliative benefit from ATG or other forms of immunomodulatory therapy. The challenge to clinicians is identifying these individuals prospectively, so that the expense and toxicity of the drug can be avoided in patients who cannot be expected to respond—especially in settings where health care resources are limited.

MDS is an extraordinarily heterogeneous disorder, and patients seen in the clinic with this catch-all diagnosis likely suffer from a very diverse set of maladies.³  Recently, Saunthararajah et al devised a method based on a logistical regression model to predict which patients classified as having MDS are most likely to respond to ATG.⁴  Important positive predictive factors in this analysis are younger age, shorter duration of red cell transfusion dependence, and HLA-DR15 (DR2) positivity. These surrogate markers for immune-mediated marrow failure will need to be validated in a prospectively treated cohort of MDS patients, and we can expect refinement of the model as experience with the drug accumulates.

Madkaikar and Ghosh also raise the issue that 6 of the 8 nonresponding patients in our series had refractory anemia with excess blasts (RAEB). Indeed, Molldrem et al demonstrated that patients with refractory anemia (RA) are much more likely to respond to ATG than those with RAEB.²  This is not surprising; in the chronic myeloid disorders, it is suspected that by the time marrow myeloblasts begin to accumulate, multiple genetic injuries have occurred at the level of progenitor cells and the immune system is no longer likely to be a major player in sustaining ineffective hematopoiesis. Notably, however, all of the patients in our series had 10% blasts or less, and the International Workshop on Prognostic Factors in MDS identified 10% blasts as an important predictor of survival and rate of leukemic transformation.⁵  The “10% cutoff” separating low-risk from high-risk disease has been confirmed by others and is reflected in the new World Health Organization (WHO) classification of myeloid neoplasms that subdivides RAEB into 2 subtypes, RAEB-1 and RAEB-2, with 10% blasts as the distinction between them.⁶  The ATG trial of Molldrem et al preceded the WHO classification and the recently reported British antilymphocyte globulin (ALG) trial⁷  also used the older French-American-British MDS classification, so it is not clear how refractory cytopenia with multilineage dysplasia patients (RCMD—another new WHO category⁶ ) will respond to ATG/ALG compared with RA patients, nor is it known whether RAEB-1 patients might fare better than RAEB-2.

Finally, we note that several different types of ATG and antilymphocyte globulin (including those derived from horses, rabbits, and goats) are currently available for use around the world and that the response patterns to one preparation may differ from those observed with another.