ALBERT

Description: Abstract 5178 Background: Natural Killer cells (NK cells) are part of the innate immune system. These cells have the ability to recognise and kill malignant cells, like myeloma cells. NK cell activation is tightly regulated by different activating or inhibiting receptors. Killer immunoglobulin like receptors (KIR, CD158) are a family of receptors which have activating as well as inhibitory function. KIR molecules are thought to recognize the HLA-C molecules, which then lead to NK cell signal transduction. Our knowledge about KIR expression and impact on tumor cell control has developed over the last years, but still our understanding of how the receptors are activated in multiple myeloma is limited. We therefore, investigated three different model systems for NK cell alloreactivity: 1) HLA-C/ HLA-C interaction model (KIR-Ligand model), 2) HLA-C/ KIR receptor interaction model, and 3) impact of donor KIR haplotype. Material and Methods: Three different myeloma cell lines (KMS12BM [C1/C1], MOLP8 [C1/C2] and RPMI8266 [C2/C1]) and a NK cell line (NKL) were cultured under standard conditions. NKL cells were transfected with human KIR2DL1 and KIR2DL3 alleles, respectively. For RNAi experiments two siRNA and two control siRNA were used. NK cells from healthy donors were isolated by magnetic end labeling. Enriched NK cells were HLA-typed for expression of HLA-C molecules, KIR receptor expression or KIR haplotype. Thereafter, NK cells were transiently transfected with the siRNA or control siRNA against the KIR2DL1 or KIR2DL3 receptors for up to 48 h. Functional analysis of the NK cell cytotoxicity was measured using a LDH release assay, based on their killing ability against the three fore mentioned myeloma cell lines. Results: Using NK cells that have been HLA-C genotyped as HLA-C1/C1, we observed a rescue of C1/C1 positive meyeloma cell lines in contrast to the C2/C1 myeloma cell line (12% cytotoxicity vs. 38% and 45% cytotoxicity, respectively, p

Description: Abstract 1201 Poster Board I-223 Introduction: Autologous stem cell transplantation followed by a dose-reduced conditioning and allogeneic stem cell transplantation from HLA-identical siblings has become a treatment option for patients with multiple myeloma. However, only a minority of the patients with multiple myeloma has an HLA-identical sibling and the experience using unrelated donor in this setting is limited. Patients and Methods: From 1997 to 2007, 73 patients (male:45; female:28) with multiple myeloma stage II/III and a median age of 49 years (r, 29-64) were included in a prospective trial to determine the efficacy of a tandem auto-allogeneic stem cell transplantation SCT) from HLA-identical sibling (n=24) or unrelated donors (n=45). Unrelated donor were either fully HLA matched (n=29) or had one mismatch (n=16).Deletion 13q14 could be analyses in 64 pts was found to be positive in 66% of the pts. Del13q14 was more present in patient with unrelated (n=42) than with related (n=22) donors. Stem cell source was PBSC (n=69) or bone marrow (n=4). Induction-chemotherapy consisted of a median of 4 cycles anthracycline-based therapy in 60 pts, or of thalidomide- (n=3) or bortezomib- (n=8) based regimen. 6 pts did not respond to induction therapy and received salvage chemotherapy before autologous SCT. Conditioning prior auto SCT consisted of melphalan 200mg/m2. After a median of 110 days (range 39-228) patients received a reduced intensity regimen with melphalan (140 mg/m2)/fludarabine regimen followed by allogeneic SCT from related (n=24) or unrelated (n=45) donors. GvHD prophylaxis consisted of anti-lymphocyte globulin (ATG-Fresenius®), cyclopsorin A and short course of MTX. Results: No primary or secondary graft failure was observed and leukocyte engraftment was achieved after a median of 15 days (range, 9-27), respectively. Acute graft-versus-host disease (GvHD) grade II to IV occurred in 38% and chronic GvHD in 22% of the patients. Limited GvHD was seen in 16 % and extensive GvHD was seen in 6 % of the patients. There was no difference regarding incidence of GvHD between HLA-identical sibling and unrelated donors. Overall response rate at day 100 was 94% including 55% complete remission (CR) and did not differ between related and unrelated SCT. Cumulative incidence (CI) of non-relapse mortality at one year was 20% (95% CI:11-29%) and did not differ between MUD and MRD (21 vs 17%, p 0.35). The cumulative incidence of relapse at 3 and 5 years was 30% (95% CI:19-41%) and 42% (95% CI: 29-55%), respectively with no difference between related and unrelated SCT at 5 years: 36 vs 44%(p= 0.6). The only significant factor for higher relapse incidence at 5 years was the presence of del13q14 (60 vs 20%, p= 0.007). After a median follow up of 40 months (r., 26-100), the estimated 5-year progression-free (PFS) and overall survival (OS) rates were 31% (95%CI: 19-43%) and 54% (95% CI: 42-64%), respectively, with no difference between related and unrelated SCT. Due to the higher relapse incidence only presence of del13q resulted in a significant worse 5- year OS and DFS (45 vs 77%, p=0.02 and 18 vs 57%, p=0.04). Conclusions: Unrelated donors as stem cell source for auto-allogeneic tandem stem cell transplantation for newly diagnosed myeloma patients resulted in similar NRM, relapse-incidence, DFS and OS than HLA-identical sibling transplantation and can therefore be used as alternative stem cell source. Outcome after transplantation is better for patients lacking del 13q14. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 3055 Acute and chronic graft versus host disease (GvHD) are severe complications after allogeneic stem cell transplantation but occurrence of GvHD is associated with reduced risk of relapse. Therefore, strategies to prevent GvHD without increasing the risk of relapse are urgently needed. We investigate retrospectively whether anti-lymphocyte globulin (ATG Fresenius®, Fresenius Biotech, Graefelfing, Germany) as part of the conditioning regimen may reduce the risk of GvHD without increasing the risk of relapse after peripheral blood stem cell transplantation from HLA-identical siblings or HLA-compatible relatives. Out of 462 patients who received HLA-identical stem cell transplantation between 1990 and 2011 in our institution, we selected 238 consecutive patients who received allogeneic peripheral blood stem cell grafts after the year 2000. The median age of the patients was 48 years (r.,18–73y) and diagnosis were: AML (n=93), ALL (n=24), CML (n=25), MDS (n=23) or lympho-proliferative disorders (n=73). Patients were classified as good risk (n=95) or bad risk (n=143). 79 patients did receive ATG within the conditioning regimen with a median dose of 30mg/kg (r., 20–90mg/kg) and 159 patients did not receive ATG. In the ATG group there were more HLA mismatch donors (6% vs. 1%, p=0.02), more bad risk patients (70% vs. 50%, p=0.04), more reduced intensity conditioning regimens (65% vs. 34%, p〈 0.001) and older patients (median age 50 vs. 46 years, p=0.03). Acute GvHD grade I to IV was less observed in the ATG group (27% vs. 40%, p= 0.04), but the difference in severe GvHD grade III and IV did not show statistical significance (10% vs. 18%, p=0.1). Chronic GvHD was less observed in the ATG group (30% vs. 52%, p=0.002), which was most obvious for extensive chronic GvHD (14% vs. 40%, p

Description: Abstract 3024FN2 Reduced-intensity conditioning (RIC) regimen followed by allogeneic stem cell transplantation (SCT) has become a reasonable treatment option for patients with multiple myeloma who are refractory or have relapse to an autograft. However, in comparison to standard myeloablative conditioning RIC resulted in higher risk of relapse. Maintenance therapy after autologous transplantation has shown to improve survival while after allogeneic SCT data are lacking so far. We here report the results of myeloablative toxicity-reduced allograft consisting of intravenous busulfan (12.2 mg/kg) and cyclophosphamide (120 mg/kg) followed by lenalidomide maintenance in 33 patients with multiple myeloma who relapsed to an autograft. A total of 32 patients had received one (n=16), two (n=15), or even three (n=1) autografts, and 1 patient was refractory to 2 induction therapies and failed to collect autologous stem cells. The median duration of remission after the autograft was 12 months. The primary endpoint of the study was non-relapse mortality at 1 year and secondary objectives were evaluation of response, incidence of acute and chronic graft-versus-host disease as well as progression and overall survival. To prevent graft-versus-host disease antithymocyte globulin (ATG Fresenius®) was given at a median dose of 20 mg/kg on day -3, -2, and -1. Lenalidomide was started earliest at day 120 after transplantation if there were no signs of infection or graft-versus-host disease. The median time between last autograft and allogeneic transplantation was 20 months. 19 patients were treated with fully HLA-matched unrelated donor, 8 patients had a mismatch donor, and 6 patients were transplanted from an HLA-identical sibling. 2 patients died of treatment-related complications resulting in a cumulative incidence of non-relapse mortality at 1 year of 6% (95% CI: 0–14%). After transplantation 27% developed grade II graft-versus-host disease, and severe grade III graft-versus-host disease was seen in 6% of the patients. Complete remission was noted in 46% of the patients, partial remission was seen in 48% and stable disease in 3%. The median interval between allogeneic transplant and start of lenalidomide was 168 days. The median starting dose was 5 mg (range 5–15 mg) without dexamethasone for 21 day followed by 1 week rest. 9 patients did not receive lenalidomide maintenance due to ongoing graft-versus-host disease, cytopenia or patient's wish. The median number of lenalidomide cycles was 6 (range 1–30). During follow-up 13 patients discontinued lenalidomide treatment due to progressive disease (n=6), GvHD (n=3), thrombocytopenia (n=2), or fatigue (n=2). In 10 patients lenalidomide dose could be increased to 10 or 15 mg, respectively. The major toxicities of lenalidomide were acute graft-versus-host disease grade I – III (21%), viral reactivation (12%), thrombocytopenia grade III-IV (12%), neutropenia grade III-IV (6%), peripheral neuropathy grade I-II (12%), or other infectious complications (6%). During follow-up 9 patients experienced relapse resulting in a cumulative incidence of relapse at 3 years of 42% (95% CI: 18–66%). The 3 year estimated probability of progression-free and overall survival was 52 % (95% CI: 28–76%) and 79 % (95% CI: 63–95%), respectively. In the current trial neither the deletion 13q14 nor the use of mismatch donor nor the chemosensitivity prior allogeneic SCT could be identified as risk factor for survival. This study showed that toxicity-reduced myeloablative conditioning regimen is feasible and highly effective in relapsed patients with multiple myeloma resulting in an acceptable treatment-related mortality. Lenalidomide as maintenance therapy is feasible early after transplantation but toxicity especially the induction of graft-versus-host disease should be considered. Disclosures: Kröger: Celgene: Research Funding. Kropff:Janssen-Cilag: Consultancy, Honoraria; Celgene: Consultancy, Honoraria.

Description: Abstract 3387 Poster Board III-275 Introduction: Allogeneic stem cell transplantation after a dose-reduced conditioning has become a reasonable treatment option for elderly patients with MDS/sAML. For patients with high number of blasts prior transplantation, the risk of relapse is considerably. To reduce the risk of relapse after dose-reduced allograft we performed a study using an anthracycline based induction chemotherapy (amsacrine, cytosine-arabinoside, fludarabine) followed immediately by a reduced intensity conditioning therapy consisting of busulfan (8mg/kg). Patients and Methods: Between November 2005 and November 2008, 49 patients with MDS (n = 24), CMML (n = 8) and sAML (n = 17) and a median age of 61 years (r: 26 – 73) and a median number of 13% blasts were included. Stem cell source were unrelated (n = 43) or related donor (n = 6). Results: No graft failure was observed and the median time to leukocyte engraftment (〉 1.0 × 109 /l) was only 10 days (r: 7 – 32). The incidence of acute graft-versus-host disease grade II to IV was 39 % and of grade III / IV was 14 %. Chronic GvHD was noted in 57 % of the patients, which was limited in 35 % and extensive in 32 % of the patients. After a median follow-up of 15 months (r: 3 – 35) the two-years estimated disease-free and overall survival was 49 % (95 % CI 33 – 65 %) and 54 % (95 % CI 39 – 69 %), respectively. The 1 year cumulative incidence of treatment-related mortality was 29 % (95 % CI 15 – 43 %). The 2 year cumulative incidence of relapse was 18 % (95 % CI 6 – 30 %). Patients with fully matched related or unrelated donor had a better survival than patients transplanted from mismatched donor (69 % vs. 37 %; p=0.06). Conclusions: A sequential approach using anthracycline based induction chemotherapy followed immediately by a busulfan based reduced conditioning regimen and allogeneic stem cell transplantation from related and unrelated donors resulted in a fast engraftment and a relative low risk of relapse in elderly patients with advanced MDS or sAML. To lower the therapy related mortality a careful donor selection is mandatory. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 2373 NK-cell alloreactivity determined by donor killer-cell immunoglobulin-like receptors (KIRs) has been shown to be predictive for relapse-free survival in myeloid leukemia after allogeneic stem cell transplantation. The role role of NK-cell allo reactivity in patients with multiple myeloma is unclear. Although the ligands for activating KIRs remain elusive, inhibitory KIRs on donor derived NK-cells and HLA class I molecules on the recipient cell determine NK cell alloreactivity. KIR genes on chromosome 19 segregate independently from HLA genes KIR haplotypes can be distinguished in two groups: group A has a fixed number of inhibitory receptors and only one activating receptor (KIR2DS4), while group B haplotype includes more activating receptor genes. Therefore individuals can be categorised as having A-KIR haplotype (A/A), or B-KIR haplotype which contains either A/B heterocygotes or B/B homocygotes KIRs. To analyse the effect of donor KIR-haplotype on outcome in multiple myeloma, we analysed 118 patients with a median age of 51 years (r: 29 – 68) who underwent allogeneic stem cell transplantation. Gender of the study population was predominantly male (n = 75, female: n = 33). Stem cell source was mainly peripheral blood stem cells (n = 105) and derived from unrelated donors (n = 81) or HLA-identical sibling (n = 37). 46 patients received graft from HLA-mismatched donor. The majority of the patients received a melphalan (100 – 140 mg/m2)/fludarabine-based reduced conditioning regimen (N = 106), either as part of an auto-allo-tandem protocol or as an allogeneic salvage protocol after relapse to an autograft. 46 patients had experienced relapse to an autograft prior to allogeneic stem cell transplantation. 79 patients were transplanted from KIR haplotype Bx and 39 from KIR haplotype A. Both groups were well balanced regarding age, HLA match, stem cell source, CMV seropositivity, remission status prior to transplantation, del 13q14, and relapse to prior autograft. After a median follow-up of 5 years the 5-year estimated disease-free and overall survival was better for KIR haplotype B donors in comparison to KIR haplotype A donors (30 % vs. 14 %, p = 0.008, and 49 % vs. 36 %, p = 0.07). The disease-free survival benefit for KIR haplotype B was mainly seen in HLA-matched patients (4y DFS: 39 % vs. 18 %, p = 0.005), while in HLA mismatch transplantation no difference according to KIR haplotype could be detected (26 % vs. 18 %, p = 0.5).The difference in survival in the HLA-matched group was due to higher risk of relapse for KIR haplotype A in the HLA-matched group (cumulative incidence at 1 year 46 % vs. 17 %, p = 0.005). In a multivariate Cox model KIR haplotype A remained a significant factor for worse disease-free (HR: 1.82, p = 0.008) and overall survival (HR: 1.69, p = 0.042). Other significant factors in the multivariate analysis were female donor sex (HR: 1.67, p = 0.04) and chemo-refractory disease at transplantation (HR: 1.76, p = 0.03) for OS and factors for DFS were unrelated donor (HR: 1.83, p = 0.02) and patient's age as continous variable (HR: 1.03, p =0.008) and donor female sex (HR: 1.62, p = 0.03). We conclude that female donor sex and KIR haplotype A are important risk factors for outcome of allogeneic SCT in multiple myeloma and should be considered for donor selection. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 2376 Allogeneic stem cell transplantation (HSCT) for multiple myeloma is a potential curative treatment approach. A high number of relapses after allogeneic stem cell transplantation after reduced intensity conditionings underlying the need of post transplant strategies to improve remission rates and disease free survival. Lenalidomide is an effective drug in treatment of multiple myeloma patients. The efficacy and the immunmodulatory properties on T- and NK- cells may augment the Graft versus Myeloma effect after HSCT, but myelosuppression as well as induction of GvHD is a concern of using the drug early after allogeneic stem cell transplantation. Study objective was to determine the maximal tolerable dose (evaluating three dose levels) of lenalidomid after HSCT in patients with multiple myeloma. A total of 18 patients with multiple myeloma were enrolled so far. Lenalidomide as single agent maintenance treatment was started between 100 and 180 days after HSCT. In the first subgroup three patients started with dose of 5 mg daily from day 1 till 21 for duration of four cycles. In this group dose limited toxicities not appear. The next higher dose level was 10 mg/d lenalidomide for a total of 6 patients. In this cohort of patients, 3 patients showed dose limiting toxicities, which were caused by an acute pancreatitis in one case, elevated liver enzymes (CTC grade III), attribute to liver GvHD in one case and renal insufficiency in one patient. Because 3 patients in 10 mg cohort developed DLT, the maximum tolerate dose has been declared as 5 mg. So far 6 additional patients were treated with this dose level (5 mg/d, day 1–21). In these cohort four patients experienced CTC grad III toxicity was observed: two cases with acute GvHD, one case with elevated liver enzymes and one case with intolerance and dizziness. Beside lenalidomide’s effect on myeloma cells, the drug is a known immune modulator. We therefore, analysed T- and NK cell subsets of patient peripheral blood by flow cytometry after lenalidomide treatment. An increase of 6% in activated CD3 cells was observed, as indicated by expression of HLA-DR molecules. Furthermore, except for the increase of whole CD8 cell number, we also observed increased proinflammatory CD8/INFg+ T cell numbers (1,5% to 4,5% p=ns). Along with the T cell date, NK cells expressed more activating receptors, like NKp44 and less inhibitory receptors, like NKG2A, support the immunmodulatory efficacy of lenalidomide. We concluded 5 mg lenalidomide as a single agent is the maximum tolerate dose if used early after allogeneic stem cell transplantation (day 100 – 180). Lenalidomide has high immune modulatory properties that might increase the risk of GvHD by activating CD8 cell. Combining lenalidomide with immunosuppressive drugs such as dexamethasone or bortezomib may reduce the risk of GvHD. Disclosures: Schonland: Celgene: Research Funding. Kröger: celgene: Research Funding.

Description: Abstract 148 Multiple Myeloma (MM) is considered to be an incurable disease, but in some patient long-term survival can be achieved after high dose chemotherapy followed by autologous or allogeneic stem cell transplantation. Those patients, who achieved molecular remission, have a high probability of long-term disease-freedom and of cure. More recently the introduction of dose-reduced conditioning and the use of auto-allo tandem approach has broaden the use of allogeneic stem cell transplantation in MM, but the results of prospective trials are controversial. We here investigated within a prospective protocol the incidence and impact of molecular remission on long-term outcome obtained after auto-allo tandem approach in MM. (study registration NCT 00781170). From 4/2000 to 10/2008, 73 patients with a median age of 50 years (r 29–64) and advanced stage II/III MM were included. The conditioning regimen for autologous transplantation consisted of melphalan 200 mg/m2 given divided over 2 days. After an interval of 2–3 months, the patients received a dose-reduced conditioning consisting of melphalan 140 mg/m2, fludarabine 180 mg/m2 and ATG Fresenius® (Fresenius, Bad Homburg, Germany) at a dose of 10 mg/kg for related and 20 mg/kg for unrelated donors on day -3, -2, and -1 followed by allogeneic stem cell transplantation. Remission was defined according to modified EBMT criteria. In 46 patients with CR or nCR minimal residual disease could be monitored by myeloma-specific primers (allele specific oligonucleotides, ASO) targeting IgH gene rearrangements (n=20) or by highly sensitive plasma cell chimerism using real-time PCR after magnet-activated cell sorting of CD138+ cells (n=26). The sensitivity of plasma cell chimerism was 10−4 using real-time PCR and donor/patient specific polymorphisms and 10−5 in case of Y-PCR. Sensitivity for nested PCR with myeloma-specific primers was 10−5. In 5 patients only 1 MRD measurement was available and those patients were excluded from MRD sub-analysis. Molecular complete remission was defined as by least two negative PCR results with ASO primers or achievement of at least two times ≥99.9 % donor chimerism of plasma cells by real-time PCR from bone marrow samples. Sustained MRD negativity was defined by at least 4 negative molecular results either by ASO primer or plasma-cell chimerism. Mixed MRD was defined by negative molecular markers for at least two times, but intermittent positivity was defined by myeloma-specific primers or plasma cell chimerism. Overall 44 patients (60 %) achieved a CR according to EBMT criteria with negative immunofixation. 8 % achieved a VGPR and 18 % a partial remission. 3 % had progressive disease and 11 % were not evaluable for determination of remission. Molecular remission was observed in 30 patients, in 15 patients the molecular markers were sustained negative while in 15 patients molecular markers were only intermittent negative, resulting in an overall complete molecular remission rate of 46 % and a stable sustained complete molecular remission rate of 23 %. After a median follow-up at 7 years the 5 year progressive-free survival was 29 % (95% CI: 17–41%). Patients without del(13q14) had a significantly better PFS than those with del(13q14) (5 y: 56% vs. 17%, p=0.02). Patients who achieved CR after transplantation had improved PFS in comparison to non-CR patients (5 y: 41% vs. 28%, p=0.008). Patients with sustained negative molecular remission had a 5 year PFS of 85 % vs. 31 % for mixed molecular remission (p= 0.003). The 5 year overall survival was 52 % (95% CI: 40–64%). Patients without del(13q14) had significantly improved 5 year survival (75% vs. 40%. p=0.02). Patients who achieved a sustained molecular remission had a 5 year OS of 91 % while those with mixed molecular remission resulted in a 5 year OS of 87 % (p=0.06). The study underlines the importance of the depth of remission and shows that achieving molecular remission as determined by myeloma-specific IgH gene rearrangements and plasma cell chimerism is associated with long-term freedom from disease and potential cure of MM in an auto-/allo SCT approach. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 4147 Introduction: AML is one of the most common indications for allogeneic hematopoietic stem cell transplantation (allo-HSCT). A matched sibling donor (matched related donor, MRD) is generally considered preferable to an unrelated donor (UD). It has been reported before that donor age may influence outcome of allo-HSCT. Sibling donors of older patients will generally be old as well. How a young unrelated donor compares with an old sibling donor is unclear. In this retrospective study we evaluate the influence of donor age on outcome of MRD and UD allo-HSCT and compare the impact of young unrelated donors vs. old related donors on overall survival (OS). Patients and Methods: We performed a retrospective analysis of 298 consecutive patients transplanted for AML at our center between January 2000 and December 2009. Median follow-up was 56 months. The median age of donors and patients were 39 and 50 yrs. respectively. These two ages were used as cut-off points for young and old donors and patients respectively. Accordingly, four donor aged groups were assigned: UD ' 39 years, UD 〉 39 yrs, MRD ≤ 39 yrs and MRD 〉 39 yrs. Analyses were performed in the more homogeneous population of patients in CR at time of transplant (n = 170) as well for patients older than 50 years and in CR at time of transplant (n = 77). Patient characteristics for the 170 Patients in CR are summarized in table 1. Results: In the cox-regression multivariate analysis of all 298 patients including factors significant in the univariate analysis (conditioning intensity, cytogenetic risk group, remission status at time of transplant and donor age) remission status (RR: 1.88, p 〈 0.001) and cytogenetic risk (RR: 5.50, p = 0.004) significantly impacted overall survival (OS). Donor age (p = 0.08) and patient age (p = 0.063) tended to influence OS in this group. In the more homogenous group of 170 patients who were in CR at time of transplantation the Kaplan Meier estimated 5 yr OS was 66% (95% CI: 54%-78%) for patients transplanted from UD ≤ 39 yrs, 41% (95% CI: 25–57%) for UD 〉39 yrs, 61% (41–81%) for MRD ≤ 39 yrs and 33% (19–47%) for MRD 〉 39 yrs (p =0.002 comparing all 4 groups), fig. 1. OS of patients with UD ≤ 39yrs was significantly better than for those with MRD 〉 39 yrs (66% vs. 33%, p = 0.001). In the multivariate cox-regression analysis (including donor age, patient age and cytogenetic risk) only donor age (p = 0.001) and cytogenetic risk (p = 0.011) significantly impacted OS whereby MRD 〉 39 yrs. was associated with poorer OS compared to UD ≤ 39 yrs (RR: 3.07, p〈 0.001). Further subgroup analyses of patients 〉 50 years old and in CR at time of transplant (n = 77) revealed similar findings with 5 yr OS of 62% for UD ≤ 39 yrs and 25% for MRD 〉39 yrs (p = 0.016). Conclusions: In patients undergoing allo-HSCT for AML our findings suggest that young unrelated donors may improve survival outcome as compared to old related donors. Further studies are necessary to confirm these findings and better define possible age limits for choosing young unrelated donors vs. older sibling donors. Disclosures: No relevant conflicts of interest to declare.