AbstractAbstract 2314Allogeneic stem cell transplantation (SCT) from matched unrelated donors (MUD) is a potentially curative approach for patients (pts) with hematologic malignancies and no sibling donors. Large progress has been made in MUD SCT, in recent years with better tissue typing and supportive care, however GVHD continues to be a major obstacle to cure. Several GVHD prevention regimens have been used, but there is currently no data to support one over the others. Over the last 10 years, 279 pts were given MUD SCT in a single institution. We historically used cyclosporine A and short-term methotrexate (15 mg/m2 on day1 and 10 mg/m2 on days 3 and 6, CSA/MTX, n=187). Over the last 3 years we used CSA with oral mycophenolate mofetil (MMF, 1 gr q12h, from day 0, n=92). We analyzed outcomes retrospectively. Acute and chronic GVHD were graded by the NIH consensus criteria. The median age was 52 years (range, 18–73). Diagnoses included AML/MDS (n=152), ALL (n=30), CML (n=12), myeloma (n=24), lymphoma (n=40) and others (n=21). Disease status at SCT was early (n=79), intermediate (n=81) or advanced stage (n=119). The donor was 10/10 (n=191), 9/10 (n=54) or ≤8/10 match (n=34). Conditioning was myeloablative (n=58), reduced intensity (RIC, n=145) or reduced toxicity myeloablative (RTC, n=76). All pts were given ATG during conditioning (15 mg/kg). The CSA/MMF regimen was less toxic than the CSA/MTX regimen. It allowed faster engraftment, day +11 and +14, respectively (p<0.001). Only 8% Vs 24% had ≥grade III mucositis (p=0.001) and 28% Vs 41% ≥grade III liver toxicity (p=0.04), respectively. Day 30 mortality, that is mostly attributed to regimen toxicity was 2.2% and 10.7%, respectively (p=0.04). Multivariate analysis (MVA) identified high comorbidity index (HR 2.5, p=0.05), advanced disease (HR 6.2, p=0.005), HLA mismatched donor (HR 2.4, p=0.03) and lymphoid malignancies (HR 2.8, p=0.03) as adverse prognostic factors for regimen-related mortality. GVHD prevention with CSA/MMF was protective (HR 0.4, p=0.02). However, CSA/MMF was less effective in preventing acute GVHD. The cumulative incidence of grade III-IV acute GVHD was 29% and 18%, respectively (p=0.005), with related deaths occurring in 22% and 9% of all pts, respectively (p=0.005). MVA identified HLA mismatched donor (HR 3.4, p=0.001), CSA/MMF (HR 2.4, p=0.004), and male gender (HR 2.5, p=0.05) as adverse factors, while RIC/ RTC was protective from severe acute GVHD and GVHD-related deaths (HR 0.4, p=0.01). The net effect was that the cumulative incidence of non-relapse mortality (NRM) with both regimens was equivalent. The most important factor predicting for overall survival (OS) was disease status at SCT with estimated 5-year OS rates of 52%, 40% and 14%, for pts with early, intermediate and advanced disease status, respectively (p<0.001). MVA identified advanced disease (HR 2.9, p<0.001), HLA mismatched donor (HR 1.4, p=0.05) and age>55 (HR 1.4, p=0.08) as adverse factors while RIC/RTC was favorable in this dataset (HR 0.7, p=0.05). The GVHD prevention regimen had no impact on OS. However, when pts were stratified based on disease status, the GVHD prevention regimen had a major impact. In early stage disease, OS was 65% and 42%, after CSA/MTX and CSA/MMF, respectively (p=0.05), with CSA/MMF identified as a poor prognostic factor in MVA; HR 1.9 (p=0.06). This was predominantly due to excess deaths of GVHD in this group. On the other hand, in pts in advanced disease, OS was better with CSA/MMF; 20% and 12%, respectively (p=0.04). MVA identified age>55 (HR 1.6, p=0.04), mismatched donor (HR 2.2, P=0.007) and lymphoid malignancies (HR 1.9, p=0.03) as poor prognostic factors, while CSA/MMF was favorable (HR 0.5, P=0.005). This was predominantly due to less deaths due to relapse in this subgroup. In conclusion, the GVHD prevention regimen has a major impact on outcome after MUD SCT. CSA/MMF is a less toxic regimen and allows prompt engraftment, but is less effective in preventing GVHD. In early stage disease, where outcome is dominated by NRM, a more effective GVHD prevention regimen such as CSA/MTX is needed. However in pts with advanced disease, who are also more susceptible to regimen toxicity, a less intensive regimen, that also better preserves GVL, such as CSA/MMF, may be associated with a better outcome. The GVHD prevention regimen selected may need to be tailored to pt and disease characteristics. These observations merit further investigation in randomized prospective studies. Disclosures:No relevant conflicts of interest to declare.