Fludarabine finds its significant other?

Comment on de Lima et al, page 857, and de Lima et al, page 865

Fludarabine-based conditioning regimens for allogeneic stem cell transplantation offer distinct advantages for reducing treatment-related mortality; the problem is how to conserve low toxicity while maintaining antileukemic efficacy.

When bone marrow transplantation conditioning regimens were first used to treat leukemia, the transplants were intended primarily as a means to reconstitute marrow function after a massive leukemia-ablative treatment with “supralethal” doses of total body irradiation (TBI). The 1990s brought a realization that the donor graft could strongly attack leukemia through an alloimmune graft-versus-leukemia (GVL) effect. Transplantation conditioning regimens for malignant diseases are, therefore, now designed not only to cytoreduce leukemia but also to establish donor immunity to provide a GVL effect. This new thinking stimulated the M.D. Anderson and other marrow transplantation teams to explore reduced-intensity conditioning regimens, focusing more on the establishment of a GVL effect and less on the cytoreductive power of the regimen. Fludarabine has become a central component of these regimens because of its potent immunosuppressive but nonmyeloablative action. An immediate spin-off was the safer application of stem cell transplants to older and debilitated patients. However, these regimens often failed to control disease, and the search for safe antileukemic agents to combine with fludarabine has continued.

Two papers in this issue of Blood describe 3 fludarabine-based stem cell transplantation regimens studied at the M.D. Anderson Cancer Center. The first study from de Lima and colleagues compares outcomes of older patients receiving FAI (fludarabine, cytosine arabinoside, and idarubicin), a nonmyeloablative regimen, with FM (fludarabine and melphalan), a myeloablative regimen. Predictably, the FM patients had more treatment-related complications and higher treatment-related mortality (TRM), while the FAI patients relapsed about twice as frequently (despite having more favorable disease features). However, the reasons for these differences are not simply explained by differences in myelosuppressive intensity. Further inspection reveals that the FAI group may have had less opportunity to generate GVL; they mainly had bone marrow transplants from matched siblings, while the FM group mainly had peripheral blood transplants from unrelated donors, affecting the observed increased occurrence of graft-versus-host disease (GVHD) and consequent enhancement of GVL (reducing relapse but increasing transplant risk). Furthermore, FAI caused prolonged mixed chimeric states and some patients rejected the transplant, a provocative setting for disease relapse. Unfortunately, T-cell chimerism was not specifically measured, making it impossible to confirm the suspicion that a component of the failed leukemia control was incomplete eradication of recipient immunity, favoring tolerance and preventing GVH and GVL effects but improving survival. The message from this study is that it is not sufficient to label regimens according to their perceived ability to suppress marrow. Attention should also be paid to their capacity to immunoablate, since the rapid establishment of a donor immune system becomes the central means of controlling leukemia in such transplants. FAI appears to fail the test of being sufficiently immunoablative; combinations of fludarabine with cyclophosphamide are likely more immunosuppressive and possibly superior to FAI. On the other hand FM, while suppressing leukemia, was more likely to cause TRM. Again, we cannot simply blame increased regimen intensity per se; FM seems to have more power to establish the graft, thereby facilitating GVHD with its good (GVL) and bad (TRM) consequences.FIG1 

The second study by de Lima and colleagues describes a fludarabine-busulfan (FluBu) regimen, first developed by Russell and colleagues at the Tom Baker Cancer Institute in Canada. The schedule was thoughtfully constructed to follow each daily dose of Flu with intravenous Bu to exploit possible antileukemic synergy between these agents.¹  Fortuitously, the combination of Bu with Flu may also be less hepatotoxic than the widely used Bu-cyclophosphamide (BuCy) regimens. Thus the regimen potentially has the unusual desirable feature of being simultaneously more effective and less toxic. The patients studied had a similar spectrum of myeloid malignancies to the group in the first paper but were more than a decade younger (median age 45 years). Nevertheless, even accounting for the favorable impact of younger age, the regimen was exceptionally well tolerated; the TRM of 3% looks significantly lower than the 10% to 20% usually seen in similar groups of patients receiving standard TBI or BuCy regimens. FluBu may thus represent a real advance in preparative regimens, setting new standards for what is an acceptable TRM. FluBu merits further testing in patients of all age groups. However, as toxicity becomes less of an issue, we still have to confront the continuing problem of failure to prevent disease relapse in patients with high-risk leukemias. The relapse/disease progression rate of around 45% with FluBu is comparable to that seen with TBI regimens in such a patient population. Further intensification of the regimen is therefore not likely to provide much benefit and might risk increasing TRM. Safely improving disease control will require either targeted intensification—for example, with radiolabeled leukemia-specific antibodies or the enhancement of the GVL effect.