ALBERT

Description: Purpose: The FB2A2 (fludarabine, intermediate doses of busulfan and ATG) reduced-intensity conditioning (RIC) regimen is considered as a standard RIC regimen in many centers worldwide. Recently, we have reported the prospective good results of a clofarabine-busulfan containing RIC regimen (CloB2A2) in adults with high-risk acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) in complete remission (CR) at time of transplant (Chevallier et al, Haematologica, 2014). Thus, this regimen may prove to be superior to the FB2A2 regimen in patients with AML/MDS. Patients and Methods: The aim of this study was to compare outcomes between adult AML/MDS patients who have received, between 2009 and 2015, in 26 French centers, either a FB2A2 RIC regimen (n=170, male 61%, median age: 58 years, AML 86%, CR1 79%) or the CloB2A2 RIC regimen (n=39, including the 16 cases treated within the prospective trial mentioned above, male 62%, median age: 61 years, AML 62%, CR1 64%). The FB2A2 and CloB2A2 regimens consisted of either 30 mg/m²/day Fludarabine for 5 days or 30mg/m²/day Clofarabine for 4 or 5 days, each combined with 3.2 mg/kg/day Busulfan for 2 days and 2.5 mg/kg/day Anti-thymocyte globulin (ATG, Thymoglobuline) for 2 days. As GVHD prophylaxis, cyclosporine (CsA) alone was used in case of related donor in both groups, and for the 16 CloB2A2 patients treated within the prospective trial, while CsA+ MMF were used in case of unrelated donors. The two groups were not statistically different in term of gender, median age and performans status at transplant, median white blood count at diagnosis, median time between diagnosis and transplant, type of donors or cytogenetics for AML patients. Conversely, there were more AML patients (86% vs 62%, p=0.0004) and more patients in CR1 (79% vs 64%, p=0.04) in the FB2A2 group. Also, CloB2A2 patients were transplanted more recently (median year of transplant: 2014 vs 2011, pmedian) (HR: 1.76; 95%CI: 1.10-2.82, p=0.01) were also significantly associated with lower LFS. However, when considering AML and MDS patients separately, benefit of CLOB2A2 RIC regimen appears to be restricted to AML patients (2-year OS FB2A2: 58.1% vs CloB2A2: 80.2%; HR: 2.45; 95%CI: 1.08-5.55, p=0.03; and 2-year LFS FB2A2: 53.6%, vs CloB2A2: 76.9%; HR: 2.32; 95%CI: 1.12-4.79, p=0.02). Conclusion: Thisretrospective comparison suggests thattheCloB2A2 RIC regimen can likely provide a higher survival compared to the use of a FB2A2 RIC regimen for AML patients. A prospective phase 3 randomized study is warranted. Disclosures Deconinck: JANSSEN: Other: Travel for international congress; NOVARTIS: Other: Travel for international congress; ALEXION: Other: Travel for international congress; ROCHE: Research Funding; PFIZER: Research Funding; CHUGAI: Other: Travel for international congress; LFB loboratory: Consultancy. Mohty:Janssen: Honoraria; Celgene: Honoraria.

Description: Background Efficiency measurement is an important indicator to improve quality of PBSC harvests. We have recently changed our apheresis device for PBSC collections: from the Cobe Spectra (Terumo BCT) to the Optia Spectra (Terumo BCT). We have further to this change decided to follow prospectively the harvests to estimate the efficiency of the new apheresis device. Material and Method Between November 2013 and April 2014, 21 consecutive harvests were conducted on Optia Spectra for ten consecutive patients/healthy donors. Six patients (3 multiple myeloma and 3 NHL) gave autologous PBSC. They were all mobilized with filgrastim (Neupogen, Amgen) which was associated with plerixafor (Mozobil, Genzyme) for two patients and pegfilgrastim (Neulasta, Amgen) for one patient. Four healthy donors gave PBSC for allogeneic HSCT; three were mobilized with filgrastim (Neupogen, Amgen) and one was mobilized with lenograstim (Granocyte, Chugai). Results The data of our 21 consecutive harvests showed a median PB CD34+ count of 20/mL [range, 7-132] and a median collection of 1.94 CD34 +/kg [range, 0.48-8.33]. The correlation between the PB CD34+ count and the number of CD34+ collected was strong (R2=0.83) (fig.1). A median of 3 and 2 collections were necessary in order to obtain a suitable graft for autologous and allogeneic HSCT respectively. In this study, the efficiency of leukapheresis was only calculated for harvests with a PB CD34+ count 〉 or = 20/µL (n=11), yielding a median efficiency of 30% (range, 13-52). These results were compared with those of our historic cohort of harvests achieved with the Cobe Spectra apheresis machine which yielded a median efficiency of 50% [range, 20-70]. In May 2014, in view of our lower efficiency results with the new device after 6 months of prospective study, we asked the Terumo BCT Company for an analysis of the harvests' data recorded in the Optia Spectra apheresis machine. This analysis revealed a systematic accumulation of the buffy coat layer during the procedures, which led to the low efficiency of our harvests. Awareness of the fast accumulation of the buffy coat layer in the collection chamber, especially when the donor's white blood cell count is 〉 30 G/L, is of utmost importance. The correction has to be made quickly by reducing the depth of collection from 60 to 40 (or 40 to 30) and by decreasing the speed of the pump to 3mL/min if the initial speed is 〉 3 mL/min. Theoretically, these two operations should allow the elimination of the buffy coat layer and improve the efficiency of the harvests. The outcomes of our patients are described in Table 1. Conclusion After 6 months of prospective study, the median efficiency of our PBSC harvests with the new device was 30%, that is below our expected value. Repeated procedures were necessary to collect enough CD34+ cells for grafting. For this reason, we requested that Terumo BCT make a retrospective analysis of the collection data recorded in the Optia Spectra system. Their audit showed a systematic accumulation of the buffy coat layer in the collection chamber. Now corrective measures have been implemented and we are going to continue this study over the next 6 months (June 2014 - November 2014) to estimate our new PBSC collections' efficiency. Figure 1 Correlation between CD34+ cells collected and CD34+ peripheral blood stem cell (PB) Count (R2= 0.83) Figure 1. Correlation between CD34+ cells collected and CD34+ peripheral blood stem cell (PB). / Count (R2= 0.83) Abstract 5834 Table 1 Patients outcome analysis n Nber of collec- tion(s) CD34+ collected Disease Auto/alloHSCT Aplastic outgoing delayed (neutro or plat) Disease Status at+3Months Disease Status at +6Months alive at Last follow-up Cause of death 1 4 3.46 MM auto No PR PR Yes Disclosures No relevant conflicts of interest to declare.

Description: Background Invasive aspergillosis (IA) during induction chemotherapy of acute leukemia or allogeneic HSCT has a negative impact on the outcome of patients and is a major concern in haematology department. The impact of air treatment on invasive aspergillosis since the introduction of posaconazole prophylaxis is not very well studied at this time. In our center, the haematological intensive care unit contain ten single rooms each, of which 4 rooms are equipped with laminar airflow named “Controlled Air Zone” (CAZ) and 6 rooms have no specific air treatment (NSAT). Objective Our objective in this study was to compare the development of IA during hospitalization in the two areas of the unit and to analyze the impact of posaconazole prophylaxis on the short- and long-term survival. We compared the outcome of 56 consecutive patients hospitalized for induction treatment of ALL, AML or underwent allogeneic HSCT for AML/ALL in CR1 or had been hospitalized for GvHD treatment between the years 2009 and 2013. Posaconazole prophylaxis was introduced in 2013. Results In this study, there were 29 (52%) male and 27 (48%) female with a median age of 53 years (range, 20-64), diagnosis were AML for 35 (62.5%) patients, ALL for 11 (19.65%) patients, allogeneic HSCT for 4 (7.14%) patients (AML in CR1 for 3 and ALL in CR1 for 1), episode of aGvHD for 4 (7.14) patients and biphenotypic AL for 2 (3.57%) patients. All patients with GvHD had a grade III-IV of acute GvHD with digestive involvement. Fourty (71%) patients were hospitalized in the CAZ and 16 (29%) patients in the NSAT zone. Eleven (19%) patients received posaconazole and 45 (81%) patients received fluconazole (p=0.007). IA was observed in 19 (34%) patients: 12 (21%) patients in the CAZ and 7 (12%) patients in the NSAT zone with a median time of 2.7 months (range, 0.3-20.50) since disease diagnosis. Twenty-five (44.6%) patients died in this study: 21(37.5%) from relapse and 4 (7%) from aspergillosis. After a median follow-up of 12.23 months (range, 0.4-54), the probability of survival at 6 months and 12 months was for patients in CAZ: 72.5% and 64.44% respectively and for patients not in CAZ: 67.71% and 54.17% respectively. The probability of survival at 6 months and 12 months for posaconazole patients was 90% for both, and for fluconazole patients: 66.67% and 54.17% respectively. The survival according to treatment and zone is presented in Figure 1. The cumulative incidence of aspergillosis at 6, 12 and 24 months for fluconazole patients: 28.89%, 28.89% and 35.88% respectively and for patients with posaconazole prophylaxis: 9.09%, 9.09% and 50%. All patients who had AI in the posaconazole group had aGvHD with digestive involvement and were hospitalized during this event in the NSAT zone. The incidence of aspergillosis according to zone was 25% and 46.88% at 6 and 12 months respectively for patients not in CAZ and for patients in CAZ: 25%, 25% and 34% at 6, 12 and 24 months respectively. For posaconazole patients in CAZ, there were no event of AI. The incidence of aspergillosis is presented according zone and treatment in Figure 2. In univariate analysis, we found an impact of gender (p=0.045), ALL (p

Description: Introduction Clofarabine (Clo) is a purine analogue which was developed to overcome limitations and to incorporate the favorable pharmakokinetic properties of fludarabine and cladribine. This agent has also a significant antileukemic activity. Thus, can one exploit both the antileukemic and immunosuppressive effects of Clo for further improving outcome after RIC allo-SCT for patients with high-risk myelodysplastic syndrome (MDS) or acute leukemia. Here we report the results of a prospective multicentre trial testing the use of Clo in replacement of fludarabine in combination with i.v. Busulfan (Bu) and ATG in 30 patients with high-risk MDS/acute leukemia (clinicaltrials no. NCT00863148). Patients and Methods Thirty patients (male n=18, female n=12) from 6 centres were included in this study between October 2009 and August 2012. Sixteen patients were diagnosed with high-risk MDS (n=5) or acute myeloid leukemia (AML, n=11), while 13 patients had high-risk acute lymphoblastic leukemia (ALL, Ph+ n= 2, Ph- n=11) and 1 patient a biphenotypic leukemia. All patients were in first (AML/MDS, n= 10; ALL/biphenotypic n=10) or second (AML n= 3; ALL n= 4) complete remission, or in stable disease (MDS n=3) at time of transplant. Median age at transplant was 58.8 years (range: 20.5-64.5). The median interval between diagnosis and transplant was 6 months (range: 3.8-124). Karnofsky’s performans status at transplant was: 100% (n=19); 90% (n=6); and 80% (n=4).The RIC regimen consisted of: i.v. Clo 30 mg/m²/day for 4 days (day-8 to day-5), i.v. Bu 3.2 mg/Kg/day for 2 days (day-4 and day-3) and ATG (Thymoglobuline) 2.5 mg/kg/day for 2 days (day -2 and day-1). All patients received G-CSF-mobilized PBSCs and cyclosporine alone for GVHD prophylaxis, irrespective of the type of donor (sibling donors n=14; 10/10 MUD, n=16). For the purpose of this study, the single case of binephotypic leukemia was considered as ALL for comparison between AML/MDS and ALL patients. The primary endpoint of the trial was the assessment of leukemia-free survival (LFS) at one year after allo-SCT. Results Engraftment was observed in all patients (100%). Median time for neutrophils (〉500 /µL) and platelets (〉50.000/µL) recovery was 18 (range: 14-26) and 12 (range: 0-23) days, respectively. With a median follow-up of 23 months (range: 12-37), the 1-year and 2-year overall survival (OS), leukemia-free survivals (LFS), relapse incidence (RI) and non relapse mortality (NRM) rates were 63±9% and 58+-10%, 57±9% and 53+-9%, 40±9% and 44+-9%, and 3.3±3% and 3.3±3%, respectively. Thirteen patients relapsed (43%) at a median time of 3.5 months (range: 2.3-13.1) after allo-SCT. Overall, 13 patients died, with the cause of death being relapse in 11 and GVHD in 2 (including one after donor lymphocyte injection). 1-year and 2-year OS were significantly higher for AML/MDS patients compared to ALL patients (75±10% vs 50±13%, and 75±10% vs 37±14%, p=0.04). There were trends for higher 1-year and 2-year LFS (69±12% vs 43±13%, and 69±12% vs 34±13%, p=0.08) and lower RI (57±14% vs 25±11%, and 66±14% vs 25±11%, p=0.05) for AML/MDS patients compared to ALL patients. Finally, 1-year and 2-year NRM were similar between both groups (AML/MDS: 6±6% vs 0%, p=0.36). Conclusion This phase 2 prospective multicentre trial shows that a Clo-i.v. Bu-ATG RIC regimen prior to allo-SCT in high-risk MDS/leukemia is feasible allowing for full engraftment and very low toxicity. Disease control appears to be satisfactorily, especially in AML/MDS, warranting a prospective comparison with other widely used fludarabine-based RIC regimens (e.g. Fludarabine, i.v. Bu, ATG versus Clo, i.v. Bu, ATG). Genzyme/Sanofi provided clofarabine and financial support for the study. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 3087 Introduction Allogeneic hematopoietic stem cell transplantation (allo-HSCT) represents the only potential to cure wide types of hematological diseases. A patient has 30% of chance to find a HLA-identical sibling donor while the rest of patients should find an alternative unrelated donor. The use of 10/10 HLA matched unrelated transplants has been used as a main alternative and with its unavailability, when available, a 9/10 HLA mismatched unrelated transplant has been used. The outcome of this last mismatched transplant is not very clear and its use according to patient and disease conditions has not been well defined yet. Aims To evaluate the outcome of allo-HSCT from 9/10 HLA mismatched unrelated donors compared to those from 10/10 HLA identical unrelated donors and siblings; and to define which category of patients can benefit the more in each alternative. Material and methods We have retrospectively studied the outcome of 213 patients who received allo-HSCT for different hematological malignancies, 121 (57%) from HLA identical siblings, 63 (29%) from 10/10 HLA identical unrelated donors and 29 (14%) from 9/10 HLA mismatched unrelated donors treated during the same period of time between 2006 and 2011 at our institution. In the mismatched group, 12 patients had the mismatch at HLA-A locus, 7 at the HLA-B, 7 at the HLA-C and 3 at the HLA-DQ. Characteristics between the 3 groups were comparable except for: disease type between the 2 unrelated groups, sex-matching, CMV-matching and ABO-matching. The different characteristics are detailed in Table 1. Results After HSCT, engraftment was significantly lower in the 9/10 HLA group (90%) than in the 10/10 HLA group (95%) than in the sibling group (99%), (p=0.03); the cumulative incidence of acute GVHD ≥2 at 3 months was 32% (23–41), 20% (15–26) and 27% (23–32) respectively; the cumulative incidence of extensive chronic GVHD at one year was 21% (13–30), 9% (5–13) and 17% (14–21) for the 3 groups respectively. After a median follow-up of 8 months (0–54) in the 9/10 HLA group, 10 months (0–60) in the 10/10 HLA group and 18 months in the siblings group, the median overall survival (OS) was 10 months (5–21), 18 months (11-NR) and 60 months (31-NR) respectively with a 2-years probability of 19% (8–44), 43% (31–59) and 63% (54–74) respectively. There was a higher but not significant relapse incidence at one year in the 9/10 HLA group compared to other groups while the transplant related mortality was significantly higher with a cumulative incidence at 1 year of 45% (35–55), (p

Description: Abstract 1957 Introduction: The FLAMSA sequential chemotherapy regimen followed by RIC for allo-HSCT and prophylactic donor lymphocyte transfusion (pDLT) was introduced a few years ago as a salvage therapy for patients with refractory or high risk AML/MDS. This retrospective analysis aimed to assess the outcome of 40 patients with refractory or high risk AML/MDS who received a salvage FLAMSA sequential chemotherapy (including a subgroup of patients who received a modified FLAMSA regimen incorporating Busulfan instead of total-body irradiation (TBI)). Patients and methods: This series included 30 males and 10 females with a median age of 52 years (range, 32–66) treated in 2 transplant centres in France. Diseases characteristics were as follow: progressive or refractory disease after rescue treatment for first relapse (n=21), early relapse without any further salvage therapy (n=4), and primary induction failure (PIF; n=4). The series also included 7 patients with high risk MDS (IPSS ≥ 1.5), and another 4 patients in first CR but having a very poor prognosis based on cytogenetics features (CR group hereinafter). According to cytogenetics and molecular markers, 12 patients were classified in the “standard risk” cytogenetics risk group, while 21 patients had an unfavourable/poor profile. Data were not available for 7 patients. The FLAMSA regimen included Fludarabine (30 mg/m2/d), cytarabine (2 g/m2/d) and amsacrine (100mg/m2/d) from day −12 to day −9. In addition, priming with G-CSF once daily was done in 12 patients. After 3 days of rest, a RIC regimen was administered. In 28 patients, the RIC regimen included 4 Gy. TBI, ATG 5 mg/kg total dose, and cyclophosphamide (40 mg/kg in case of matched related donors, and 60 mg/kg for unrelated or mismatched donors). In the remaining 12 patients, TBI was replaced by I.V. Busulfan 3.2 mg/kg/d for 4 days. Per protocol, pDLT was planned to be given from day +120 in patients who were not receiving immunosuppression and were free of GvHD. Eighteen patients were transplanted using an HLA identical sibling donor, and 22 received transplant from an unrelated donor (14 matched 10/10 and 8 mismatched 9/10). Results: After allo-HSCT, 39 patients (97.5%) engrafted. In the CR group (n=4), after a median follow-up of 5 months (range, 3–31) all patients were still alive in CR at last follow-up. In the remaining 36 patients, 9 patients developed acute GVHD ≥2 with a cumulative incidence at 3 months of 18% (95%CI, 10–26). At day 90 post HSCT, 23 (64%) patients could achieve hematological CR, and 14 of the 23 remained in CR at last follow-up. After a median follow-up of 6 months (range, 1–60), the 2-years probability of OS was 30% (95%CI, 17–52), and the 2-years probability of PFS was 29% (95%CI, 17–50). The cumulative incidence of disease progression at 1 year was 25% (95%CI, 18–33). Interestingly, none of the patients who received Busulfan instead of TBI as part of the RIC, relapsed. The cumulative incidence of TRM at 3 months and 1 year were 14% (95%CI, 8–20) and 22% (95%CI, 15–29), respectively. In the multivariate analysis (taking into account all relevant parameters such as priming with G-CSF, conditioning with or without TBI, related or unrelated donors, HLA matching and number of previous treatment lines), there was a significantly worsened PFS in patients who received transplant from a mismatched donor (HR=3.6; [95%CI, 1.3–10] p=0.01; below Figure). Also, when considering disease relapse, there was a highly significant impact of the type of RIC regimen (in favour of a FLAMSA regimen without TBI (p

Description: The two first authors contributed equally. Introduction There is currently little data on the outcome of AML in adolescents and young adults (AyAs) after allogeneic HSCT. In this retrospective study from the SFGM-TC registry, we analyzed the outcome of AML patients classified in 3 groups according to the age at transplantation: children (

Description: Abstract 4540 Introduction The impact of HLA DRB3 and DRB4 allele mismatch after allogeneic HSCT using unrelated donors is unclear. We therefore examined retrospectively the outcome of 35 patients who received HLA-10/10 unrelated hematopoietic stem cell transplantation with a DRB3 or DRB4 mismatch between 2005 and 2011. This cohort of 35 patients was a part of a cohort of 132 consecutive patients who underwent allogeneic HSCT between 2005–2011 with a 10/10-HLA matched donor. There were 18 males (51.4%) and 17 females (48.6%) with a median age of 48 years (range, 6–64), there were 13 (37%) AML, 9 (26%) ALL, 4 (11.5%) MDS, 3 (8.5%) multiple myeloma and 6 (5.7%) other (CML, CLL, NHL). Twenty patients (57%) received a myeloablative conditioning (MAC) and 15 (43%) received a reduced intensity conditioning (RIC). At transplantation, 21 patients (60%) were in complete remission (CR), 4 patients (11.5%) in partial remission (PR) and 10 (28.5%) in relapse; 13 (37%) patients received peripheral blood stem cell (PBSC) and 22 (63%) received bone marrow (BM). Twelve (34%) patients had a mismatched DRB4 donor and 23 (66%) patients had a mismatched DRB3 donor. In the remains of 97 patients, there were 55 male (57%) and 42 female (43%), 28 (29%) patients received a MAC and 69 (71%) a RIC as regimen before allogeneic HSCT. The stem cell source was BM for 32 (34%) patients and PBSC for 65 (66%). At transplantation, 34 (35%) patients are in CR and 63 (65%) were in PR. The distribution of diagnosis was acute leukaemia and MDS for 44 (45%), CLL for 2 (2.5%) and other diagnosis (aplastic anemia, NHL, CML, MPS) for 51 patients (52.5%). Results After HSCT, 124 (94%) patients engrafted. After a median follow-up of 11.5 months (range, 0–76), the cumulative incidence of acute GvHD≥2 at 3 months was 20% (95%CI,16.5–24) and the cumulative incidence of chronic GvHD at one year was 19 % (95%CI, 15–22). In univariate analysis, the mismatch DRB3 or DRB4 had no effect on engraftment and no effect on acute GvHD (p=0.08) or chronic GvHD (p=0.63). There was no impact of DRB3 or DRB4 mismatch on relapse (p=0.33 and p=0.53, respectively) and on PFS (p=0.63 and p=0.07, respectively). We found an impact of the DRB4 mismatching (p=0.016) on overall survival. The median survival for patient without DRB3 or DRB4 mismatch was 23 months (14-NR), for patients with DRB3 mismatch 32 months (12-NR), and for DRB4 mismatched patients 5 months (3-NR). The probability of survival at 24 months, for patients without mismatch DRB3 or DRB4 is 47% (36–61), for patients with DRB3 mismatch 51% (32–82) and for DRB4 mismatched patients 19% (6–66%). (figure1). The multivariate analysis that studied age, type of disease, DRB3 or DRB4 mismatch, sexmatching, TBI, ATG, disease status at transplantation and type of conditioning and stem cell source showed a significant impact of mismatch DRB4 on survival (HR= 2.5 [95%CI, 1.2–5.5] p=0.019); there was no impact for DRB3 mismatch (HR= 1.3 (95%CI,0.5–3.9 p=0.58). We found also an impact of the DRB4 mismatch on TRM (HR= 3.5; [95%CI, 1.6 –8] p= 0.026). The incidence of TRM at 24 months for patients without DRB3 or DRB4 mismatch is 29% (24–34), for patients with DRB3 mismatch 17% (9–26%) and for DRB4 mismatched patients 50% (34–66%). (figure 2). Conclusion The HLA DRB4 matching donor is relevant for survival of patients who undergo allo-HSCT from unrelated donor in the HLA-10/10 matching settings. In view of the important impact of these loci mismatches on clinical outcome, it seems to be important to consider this matching loci in the unrelated donor selection. Disclosures: No relevant conflicts of interest to declare.

Description: Iron overload (IO), primarily related to multiple red blood cell transfusions, is a relatively common complication in allogeneic hematopoietic stem cell transplant (allo-HSCT) recipients. Elevated pre-transplant ferritin level, a surrogate marker of iron overload, was demonstrated to be an important cause of mortality and morbidity in patients who have undergone allo-HSCT. Excessive iron accumulation results in tissue damage and organ failure, mainly as a result of the generation of free radicals that cause oxidative damage and organ dysfunction. Iron chelators have been widely used leading to normalisation for ferritine level and lower IO-related complications. As iron has a fundamental role in cell survival affecting pathways involved in DNA synthesis, cell differentiation, and apoptosis, some studies evaluated the anti-proliferative activity of iron chelators in cancer and leukemia patients on disease recurrence. The objective of this study was to determine at a first time the impact of serum ferritin level measured at time of allogeneic HSCT in adult patients with hematological disorders on the different outcomes and to investigate at a second time the role of iron chelation on relapse incidence. We included 158 patients, 100 males and 58 females with a median age of 45 years (18-67) who underwent allo-HSCT between 2002 and 2010. There were 83 acute myeloid leukemias, 10 chronic myeloid leukemias, 11 myelodysplastic syndromes, 7 myeloproliferative disorders, 19 myelomas, 9 non-Hodgkin lymphomas, 6 Hodgkin diseases, 5 aplastic anemias and 3 hemoglobinopathies. Sixty-seven (42%) patients were sex mismatched (F→M:37; M→F:30); for ABO compatibility, 61% were compatible, 18% had minor incompatibility and 21% had major incompatibility. Concerning the HSCT procedures, 60 patients (38%) received peripheral blood stem cell and 98 (62%) received bone marrow from 97 (61%) HLA related donors [matched, n=76; mismatched, n=21], and 61 (39%) HLA unrelated donors [matched, n=36; mismatched, n=25] after myeloablative [n=64, (41%)] or reduced intensity conditioning [n=94, (59%)]. At transplantation, 91 (58%) were in complete remission (CR) or chronic phase [CR1: n=61 (67%); ≥CR2: n=30 (33%)]. The median serum ferritin level at HSCT was 1327 microg./l (26-14136); 31(20%) patients had a level 26-500, 33 (21%) had a level 500-2500, and 94 (59%) 〉2500. There was no significant correlation between the different ferritin levels, disease kind and status at HSCT. After transplantation, 23 patients received iron chelating agents after a serum ferritin level of 1000 microg/l and stopped when the level decreased below 1000. The cumulative incidence of acute GVHD ≥ II at 3 months was 14% (11-16.5) with 10.5% (8-13) for grade III and 7% (5-9) for grade IV; the 1 year cumulative incidence of limited and extensive chronic GVHD were 4% (2-6) and 12.4% (9-16) respectively. After a median follow-up of 18 months (1-106), the 5 years OS probability was 65% for patients with ferritin level below 500 microg./l, 39% for level between 500 and 2500 microg./l and 28% for level 〉 2500 micog./l, [Hazard ratio= 3.5 (1.5-8.1), p=0.002]; this was explained by a significant higher TRM in patients with level 〉2500 [Hazard ratio= 4.3 (1.02-18), p=0.04]. Interestingly, we found in multivariate analysis that patients receiving iron chelators had significantly better OS [5 years OS= 59% vs. 34% for non-chelated patients, Hazard ratio= 0.34 (0.15-0.76), p=0.008], (Figure 1a), and experienced less disease relapse [5 years relapse incidence= 18% vs. 41% for non-chelated patients, Hazard ratio= 0.22 (0.07-0.73), p=0.012], (Figure 1b). In conclusion, we confirmed the negative impact of iron overload on the outcomes allo-HSCT recipients. More importantly, we demonstrated that iron chelators have a positive impact in reducing disease relapse by the possible mechanism of iron deprivation in leukemic cells. This clinical observation needs to be confirmed by prospective randomized trials.Figure 1a: Overall survival probability and b: relapse incidence in patients with or without iron chelationFigure 1. a: Overall survival probability and b: relapse incidence in patients with or without iron chelation Disclosures: Michallet: Novartis: Honoraria, Research Funding. Nicolini:Novartis: Consultancy, Honoraria, Research Funding.

Description: Abstract 5070 In a recent phase III trial, azacitidine was demonstrated to significantly prolong OS compared with conventional care regimens in patients classified in intermediate-2 and high-risk myelodysplastic syndromes (MDS) according to the International Prognostic Scoring System (IPSS) (Fenaux et al. 2009). This study used the French- American-British (FAB) classification for MDS and included approximately one third of patients with refractory anemia with excess blasts in transformation (RAEB-t; 20% to 30% bone marrow blasts). WHO criteria now define AML as ≥20% BM blasts. Using those criteria, RAEB-t is now considered as AML. We conducted a retrospective analysis on patients who received azacitidine between August 2005 and November 2011 at our institution for MDS or AML. Patients were identified through the hospital database and individual charts were reviewed. The primary objective was to investigate the outcome of patients receiving azacitidine in a daily clinical practice in high risk MDS and AML patients and to evaluate its impact on overall survival (OS). Secondary objectives were hematological response rate and transfusion spare. Patients were included if they received at least one cycle of azacitidine. Disease status was defined by both the French American British (FAB) and the World Health Organization (WHO) classification systems, and risk was scored by the International Prognostic Scoring System (IPSS). All analyses were conducted using R statistical software. Descriptive statistics were used for baseline characteristics. Kaplan-Meier estimates were used to calculate overall survival (OS). There were 79 patients, 51 (65%) males and 28 (35%) females with a median age of 70 years (32–85). The indication of azacitidine was the first line treatment use for MDS, mainly refractory anemia with excess blasts, in 40 (51%) patients (group1) and treatment for patients who had AML and transformed, in 39 (49%) patients (group2). (post chemotherapy: n=16, first line: n=23). Patient characteristics, prognostic factors according to FAB classification, ISPP risk and cytogenetics for both groups are shown in table1. The median number of azacitidine cycles in groups 1 and 2 was 8 (1–30) and 3 (1–29) respectively. Evaluation after 6 cycles showed 55% of responders in group 1 and 31% in group 2; the rest of patients have progressed. The median OS for the group 1 was 24. 5 months (17. 8-NR) while in group2; it was 15. 5 months (11. 2-NR) for patients who received AZA in first line and 6 months (3. 9-NR) for patients with previous chemotherapy. In terms of transfusions number, we did not find any significant spare in terms of both RBC and platelets transfusion in group1 while there was a significant spare of 33% of red blood cells transfusions (p=0. 05) and 42% of platelets transfusions only in group 2 (p=0. 04). The multivariate analysis studying the impact of different variables on OS showed: a worse OS in AML patients with previous chemotherapy (HR= 9. 84 [ 3. 56 – 27. 19 ], p〈 0. 001), a worse OS in patients with unfavorable caryotype (HR= 7. 30 [ 2. 13 – 24. 98 ], p〈 0. 001), and a better OS in female patients (HR= 0. 31 [ 0. 14 – 0. 68 ], p= 0. 003). Our study confirmed results from previous prospective study in MDS patients while AML patients not receiving azacitidine in first line do not seem to benefit from this treatment. Cytogenetics remain a major factor impacting OS with no significant impact of IPSS. Disclosures: Nicolini: Novartis, Bristol Myers-Squibb, Pfizer, ARIAD, and Teva: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding.