Low-Dose Azacitidine for Relapse of MDS/AML after Unrelated Donor Peripheral Blood Stem Cell Transplantation.

Background: Post-transplant relapse in patients with myelodysplastic syndrome (MDS) and acute myelogenous leukemia (AML) is difficult to manage. Cytogenetic relapse and decreasing donor chimerism often precedes morphological relapse. Weaning of immunosuppressive agents and donor lymphocyte infusion (DLI) may induce graft versus tumor effect (GVT), but is of limited value for patients with existing GVHD. In addition, the degree of GVT in high-risk myeloid disease is suboptimal. We present our experience using low-dose azacitidine (AZA) for cytogenetic relapse post-transplantation.

Methods: Six patients with high-risk myeloid malignancy with cytogenetic relapse after matched unrelated donor peripheral blood stem cell transplantation were treated with low-dose AZA at 25 mg/m² SC or IV for 5 days. An average of 2 cycles of AZA were given (range = 1 to 3). There were 3 females and 3 males with a mean age of 48 years (range = 31 to 59 years). Four had high-grade MDS (including 2 with treatment related disease) and 2 had high-risk AML. Conditioning consisted of fludarabine and ablative doses of busulfan in all patients. Cytogenetic relapse was seen by FISH testing within 187 days post-transplant (range = 30 to 730 days). AZA was given after an initial attempt to wean immunosuppression, which was not possible in 3 patients due to existing GVHD. All patients tolerated AZA well, without major toxicity. DLI was possible in 3 patients following AZA. A reduction in cytogenetic abnormalities (by FISH) and increase in donor chimerism (by FISH or STR) was observed in 5 out of 6 patients (83%) within 21 days post-AZA (range = 7 to 71 days). Three of the 5 responders demonstrated improvement after 1 cycle. The other 2 responders improved after 2 cycles given 28 and 60 days apart, respectively. One of these patients responded to a second AZA cycle after failing DLI. AZA did not appear to induce or worsen GVHD in any patient. One patient remains in CR 4 months after 1 cycle of AZA. Another patient demonstrated ongoing improvement in chimerism until her death from previously existing GVHD 20 days after a third cycle of AZA. The remaining 3 responders relapsed within 30 days from time of first response (range = 17 to 43 days).

Conclusions: Low-dose AZA appears to have activity in post-transplant relapse of MDS and AML. This low-dose regimen appears to be well tolerated, however, response to AZA is short-lived in the majority of patients. Further investigation is planned to improve the durability of response by giving AZA at regular intervals from the time of early relapse. The utility of AZA as a preparatory regimen pre-DLI should also be explored.