Background: High-risk (HR) AML including secondary AML (sAML) or therapy-related (tAML) are associated with significantly lower complete remission (CR) rates and poor outcome, after upfront “3+7” and post-remission chemotherapy. No standard intensive treatment approach for relapsed/refractory AML is well established, even more so, in older patients, but “bridging” such patients in CR to allogeneic stem-cell transplantation (HSCT), whenever feasible, remain the only curative strategy, as in younger patients. Addition of cladribine, a purine analog, to cytarabine is an available option. Significant clinical activity of CLAG (-M) regimens is now well established, with a response rate (RR) of 60% in relapsed/refractory AML or AML patients who failed hypomethylating agents. A prospective frontline study of an intensified CLAG-M regimen has recently confirmed its association to a high response rate, that may particularly benefit HR-AML patients, as a bridge to HSCT. As other reports in the literature remain sparse, we report here our current experience with CLAG-M in relapsed/refractory and sAML patients. Methods: From January 2015 to July 2018, 20 consecutive patients with HR-AML, were treated in our center, with one course of CLAG-M (cladribine 5 mg/m2 /day (days 2-6), cytarabine 2g/m2/day (days 2-6) and reduced to 1g/m2/day for patients 65y+, filgrastim 300mcg/daily (days 1-6), and mitoxantrone 10mg/m2/day (days 2-4)). If eligible for HSCT, a second CLAG course was to be administered to patients in CR. Extended myeloid mutation analysis was performed, using a 37-gene NGS panel. Such patients were classified according to Lindsley et al.,Blood 2015. Results: Median age was 63.5 years (33-79) with a 3:1 M/F ratio. Four sAML patients and 4 tAML patients were treated upfront (5 CR, 1 early death (ED) and 2 treatment failures). ELN cytogenetics was adverse (n=5), intermediate (n=1) or failed (n=4). Twelve patients were all treated after frontline 7+3 and intensive consolidation courses (n=10) or azacytidine (n=2). Three patients were refractory to prior intensive chemotherapy or AZA and 9 were in first or subsequent relapse, at time of CLAG-M administration. Median time from first treatment for MDS/AML to CLAG-M onset was 17 mos (3-29). Initial/relapse ELN cytogenetics was adverse (n=5), intermediate (n=5) and failed (n=2). Of these 12 patients, 7 obtained a response (6 CR, one CRi), 3 failed to obtain a response and 2 early died from sepsis. Seventeen patients could be classified, according to Lindsley et al. The 3 patients with missing NGS data, all had adverse ELN cytogenetics (inv3q/MECOM1, MLLr by FISH analysis or monosomy 7, associated with an IDH2 mutation). After one course, 5/7 patients, classified as secondary AML, obtained a CR, including one CRi, 3/4 patients classified as pan AML obtained a CR, while only 2/6 patients with mutated TP53 alleles obtained a CR (3 failed to respond). Overall, 12 of the 20 patients obtained a complete response (11 CR and 1 CRi), despite adverse genetical characteristics and 12 of them being administered CLAG-M, during the late evolution of their disease. Three patients early died due to undocumented pneumonitis (n=2) or bacterial sepsis (n=1). Otherwise, observed treatment toxicities were mild, with no unusual infections seen after CLAG-M. Median duration of neutropenia (<0.5 G/l) and thrombocytopenia (<100 G/l) was 28 (12-47) and 30 (23-40) days, respectively. Seven patients in CR received a second CLAG course before HSCT, when 3 patients, aged more than 70 years, only less intensive courses. One CR patient did not received consolidation due to severe sepsis, as also did the only patient with CRi due to persisting cytopenias. Both patients underwent HSCT. Fifteen patients were deemed eligible at entry for HSCT, based on age and performance status. Of those, 8 achieved a CR/CRi and all of them proceeded to HSCT. Four such patients are currently alive (2 mo, 2.6 y, 3y, 3.4y), 2 patients died from early relapse (4 and 6 mos) and 2 patients from HSCT toxicity. Conclusions: In this real-life small study of HR-AML patients, predominantly older than 60 years of age, CLAG-M was also used at first or subsequent relapse post-intensive therapy. Despite very unfavorable characteristics, 60 % of them obtained a complete response and 50% of those initially eligible were effectively “bridged ”to HSCT, without unusually deleterious outcomes observed in this small cohort. DisclosuresPeffault De Latour:Amgen Inc.: Research Funding; Pfizer Inc.: Consultancy, Honoraria, Research Funding; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding. Braun:CELLIPSE: Research Funding.