Activity and Tolerability of Azacitidine in Patients Who Relapse after Allogeneic Stem Cell Transplantation (allo-SCT) for Acute Myeloid Leukemia (AML) and Myelodysplasia (MDS): a Survey from the European Society for Blood and Marrow Transplantation (EBMT

Disease relapse is the most common cause of treatment failure after allo-SCT for high risk AML and MDS. Treatment options for patients with recurrent disease are extremely limited and re-induction chemotherapy, when administered, is often either poorly tolerated or ineffective. Azacitidine (AZA) is a DNMT inhibitor which is well tolerated and clinically active in high risk AML/MDS. Of interest AZA also up-regulates the expression of tumor antigens and plausibly augments a graft-versus-leukemia effect. A number of small studies have suggested clinical activity of AZA in patients who relapse after an allograft for AML/MDS but both overall response rate and predictors of response remain unknown. We report the first systematic study of the activity and tolerability of AZA in patients who relapsed after allo-SCT for AML/MDS. 204 patients who relapsed at a median of 6.5 months (range, 1-49) after an allograft for AML (n=130) or MDS (n=74) were studied. The median age was 58 years (range 22-76). 89 patients were transplanted using a matched sibling donor and 115 from an adult unrelated donor. 47 patients received a myeloablative and 157 a reduced intensity conditioning regimen. AZA was administered for 5-7 consecutive days every month. The median duration of AZA treatment was 68 days (inter-quartile range 24-154 days). 66 patients received additional donor lymphocyte infusions (DLI) at a median of 43 days after commencement of AZA. AZA was well tolerated in the majority of patients. 57 patients developed Grade 3-4 non-hematological toxicities 47 of which were infectious complications and likely also attributable to relapsed disease. 4 patients developed Grade 3-4 acute GVHD after AZA treatment. 45 (22%) patients achieved a complete remission (CR) or partial remission after AZA administration at a median of 114 days after commencement of treatment. 31 (15%) patients achieved a CR. The median number of courses of AZA to achieve a clinical response was three. In multivariable analysis the only significant factor determining improved response to AZA was relapse occurring more than 12 months post-transplant. The median overall survival (OS) for all patients was 6 months after the commencement of AZA therapy. In patients who achieved a CR the 2 year OS after commencement of AZA was 38.5% versus 11% for the whole population (p= 0.001). In multivariable analysis OS was determined by the occurrence of disease relapse more than 6 months post-transplant and achievement of a CR after AZA therapy. Of note, there was no significant difference in response rates to AZA between patients with relapsed AML or MDS. Concurrent administration of DLI did not improve either response or survival rates. In conclusion, these data confirm the ability of AZA to salvage a proportion of patients with AML or MDS who relapse after an allogeneic SCT and identify prognostic factors of response. The response and survival rates achieved with salvage AZA are comparable to those previously reported with either intensive chemotherapy or DLI. We conclude AZA represents an important and relatively well-tolerated new treatment option in the management of selected patients with AML and MDS who relapse after allo-SCT.