Sirolimus-Containing Graft-Versus-Host Disease Prophylaxis and High- Resolution HLA Typing Improves the Outcome of Mismatched Unrelated Donor Allogeneic Hematopoietic Stem Cell Transplantation.

Allogeneic hematopoietic stem cell transplantation (HSCT) from unrelated donors can cure a variety of hematologic diseases; however, decreased survival and graft versus host disease (GVHD) are associated with mismatches at HLA class I and class II loci. Beginning in 2004, we have employed a modified three-drug sirolimus-based GVHD prophylactic regimen (

Median age (45 vs. 51), disease status, donor gender, and type of transplant conditioning (reduced-intensity, 9 vs. 8) were similar between the two groups. Seven of 14 patients in the sirolimus (−) group received bone marrow grafts whereas all patients in the sirolimus + group received peripheral blood stem cells. Nine patients in the sirolimus – group had antigen-level typing at HLA A, B, C and DQ loci and allele level typing at HLA DRB1 and 5 patients had allele-level typing at all loci. Patients in the sirolimus+ group had allele-level typing at all loci. One subject was mismatched at HLA DRB1 and DQ at the allele level; the remaining subjects had a single allele or antigen mismatch. GVHD prophylaxis for the sirolimus(−) group included tacrolimus and mycophenolate (n=11) or tacrolimus and low-dose methotrexate x 3 doses (n=3). GVHD prophylaxis for the sirolimus (+) group included sirolimus, tacrolimus and post- transplant methotrexate for 2 or 3 doses, totaling 15mg/m2.

Median follow-up for surviving patients was 1921 days (1781–2295) in the sirolimus-group and 447 days (53–1122) in the sirolimus(+) group. The day +100 cumulative incidence of grades II-IV and III-IV acute GVHD were, respectively, 71% and 43% in the sirolimus(−) group, and 14% and 5% in the sirolimus(+) groups (p=0.002; p=0.02). Kaplan-Meier estimates of the probability of chronic GVHD or late onset acute GVHD at two years were 51% in the sirolimus(–) group and 70% in the sirolimus(+) group.

Graft failure occurred in 2 patients each in both groups and was successfully treated in two patients by retransplantation or pentostatin plus donor lymphocyte infusion. Adverse effects of sirolimus included 8 patients with renal dysfunction (36%), 4 with thrombotic microangiopathy (TMA) (18%), and 3 with dyslipidemia requiring medical therapy (14%). Deaths in the sirolimus(−) group were due to disease recurrence (n=3), graft failure (n=1), and GVHD (n=7). Deaths in the sirolimus(+) group were due to disease recurrence (n=2), TMA (n=1) and GVHD (n=4). The Kaplan-Meier estimates of disease free (DFS) and overall survival (OS) at 3 years were 21% and 21% for the sirolimus(−) group and 53% and 56% for the sirolimus(+) group. The Kaplan-Meier estimates of non-relapse mortality (NRM) at day +100 and day +365 were 22% and 50% for the sirolimus (−) group and 5% and 13% for the sirolimus(+) group.

Summary: High-resolution HLA typing and GVHD prophylaxis with sirolimus, tacrolimus, and low-dose methotrexate in patients undergoing MMUD allogeneic HSCT is associated with low rates of severe acute GVHD and non-relapse mortality. Renal dysfunction and TMA appear more common with sirolimus, and, as in prior studies with this regimen, a lower rate of acute GVHD did not result in less chronic GVHD. Nonetheless, DFS and OS appear to compare favorably to recently reported registry data for MMUD allograft recipients (Lee SJ et al., Blood;110:4576), and a prospective trial has been initiated to further study this regimen in MMUD recipients.