Introduction: Despite therapeutic advances in recent years, disease relapse is still observed in up to 50-60% of the patients (pts) receiving allogeneic transplantation for high risk AML and MDS. In an initial Phase II trial of 30 pts treated with prophylactic low-dose AZA and escalated doses of DLI to prevent relapse, we reported that such prophylactic treatment can be safely administered and compared favorably to patients receiving no post-transplant maintenance. We here update 16 pts from our center, included in that protocol, as well as a larger cohort of additional patients treated with combined epigenetic and cellular therapy and the resultant effect on their event-free survival and their overall survival. Patients and methods: Fifty-nine pts (median age 59 yrs (range, 37-70); 30 M/29 F), comprising 40 pts with AML and 19 pts with MDS, were treated. Among the AML pts (ELN classification FAV n=2, INT n=17, ADV n=21), high risk criteria were as follows: 15 had complex karyotypes, 7 Flt3-ITD, 1 MECOM (EVI1) rearr., 3 monosomies, 2 t(9;11), and 1 t(11;19)); 9 were in CR≥2, 4 were refractory, and 13 had 2ary AML to MDS/chemotherapy/or MPS. Among the MDS pts, high risk criteria included 6 pts with complex caryotypes, 2 monosomies, 1 MECOM (EVI1) rearr., and 4 had 2ary MDS. IPSS-R median was 5.5 (3 - 7.5). MDS pre-transplant status included 7 CR1, 1 CR2, 5 PR, 1 relapse, 4 upfront, and 1 refractory. The therapeutic protocol administered to the initial 16 pts consisted of AZA, starting between d56 and d100 post-transplant, at a dose of 32 mg/m²/d SC, for 5 consecutive days, every 28 days, for up to a total of 12 cycles followed by DLI commencing after 3 cycles of AZA and 4 weeks following discontinuation of immunosuppressive prophylaxis. Two additional DLI were scheduled every 8 weeks following the 1st DLI. The doses of DLI 1, 2 and 3 were, respectively, 5x106, 1x107, and 5x107 CD3+ cells/kg for matched related donor (MRD), and 1x106, 5x106, and 1x107 CD3+ cells/kg for unrelated donor (UD). In the extended patient cohort, presented here, we slightly modified the schedule: DLI could be given earlier (after one or two cycles of AZA) and haploidentical pts could receive AZA/DLI as well, with DLI at the escalated doses of 1x105, 5x105, and 1x106 CD3+ cells/kg. Immunosuppression included cyclosporine A (CsA) in case of MRD, CsA and mycophenolate mofetil (MMF) in case of UD or haplo. MMF was progressively reduced during the first 2 months. CsA was tapered starting at day 60-100 post-transplant. Patients could start AZA while they were still receiving CsA. The conditioning regimens consisted of MAC for 5 pts, RIC for 46 pts and sequential for 8 pts. Donors were MRD, MUD, UD 9/10, and haplo for 20, 30, 1 and 8 pts respectively. Results: The median number of cycles AZA was 7 (range 1-12), 17 pts received 12 cycles of AZA, while 42 pts (71%) received at least one DLI. The median number of DLI was 1 (range 0-4), 15 pts received ≥ 3 DLI. The median time for the first post-transplant AZA injection was 83 days (range 56-145 d) and the median time post-transplant for those pts who received a first DLI was 148 days (range 78-245 d). The median pt-follow-up was 17 months (range 2-83). Eight pts (13%) relapsed (7 AML and 1 MDS). The cumulative incidence of relapse and NRM at 1 year was 12% and 11.5% respectively. Fifteen pts (25%) died. Causes of death included relapse in 6 pts, infection in 7 pts, myocardial infarction in 1 pt, and GvHD in 1 pt. Overall survival and event free survival for the entire group at 2 years was 71% (95% CI 54-82%) and 67% (95% CI 50-80%) respectively. The cumulative incidence of acute GvHD grade 1-4 and chronic GvHD were 32% (95% CI 16-49%) and 39% (95% CI 20-58%). Among the 8 pts who developed grade 3 acute GvHD, 4 occurred in pts who had received DLI. One pt died from grade 4 digestive GvHD post-AZA alone. Conclusion: We conclude that, despite their very high risk disease, prophylactic/preemptive low-dose AZA and DLI can be readily and safely administered with an acceptable incidence of subsequent GvHD to the AML and MDS pts described here. Significantly, in the poor-prognosis pts under study, the incidence of relapse is lower than expected. Furthermore, considering the immune escape mechanisms prophylactic or pre-emptive either AZA or DLI might be associated to additive non-cross reactive interventions, such as checkpoint inhibitors, in order to further decrease disease relapse and increase overall survival. Disclosures Peterlin: Jazz Pharma: Consultancy; Astellas: Consultancy; Daiichi-Sankyo: Consultancy; AbbVie Inc: Consultancy. Orvain:Novartis: Honoraria; Incyte: Honoraria. Chevallier:Jazz Pharmaceuticals: Honoraria; Daiichi Sankyo: Honoraria; Incyte: Consultancy, Honoraria.