Five-Year Outcomes of the “Abatacept Combined with a Calcineurin Inhibitor and Methotrexate for Graft Versus Host Disease (GVHD) Prophylaxis: A Randomized Controlled Trial” (‘ABA2’)

Background: GVHD remains a major cause of non-relapse mortality (NRM) after hematopoietic cell transplant (HCT). Abatacept, a costimulatory blockade agent targeting the CD28:CD80/86 pathway, was FDA-approved in 2021 for GVHD prophylaxis in patients undergoing an 8/8 HLA-matched unrelated donor (MUD), or 7/8-HLA mismatched unrelated donor (MMUD) HCT for a hematologic malignancy, in part based on the ABA2 trial (NCT01743131). Here, we present 5-yr follow-up data for ABA2.

Methods: ABA2 patient eligibility criteria (described in detail in PMID: 33449816) included patients ≥ 6 yr with heme malignancies undergoing MUD or MMUD HCT after intensive conditioning. MUD patients were enrolled on the randomized, double-blind placebo-controlled arm, with patients randomized to either 4 doses of abatacept (10mg/kg/dose on Day -1, +5, +14 +28) or placebo. MMUD patients all received 4 doses of abatacept. 185 patients were enrolled: 69 MUD-placebo, 73 MUD-abatacept, 43 MMUD-abatacept. GVHD prophylaxis also included a calcineurin inhibitor (CNI, continued through Day +100 followed by a wean) and methotrexate (MTX, 15mg/m² on day +1 and 10mg/m² on Day +3, +6, +11). Both bone marrow and PBSC grafts were used.

All patient data were independently monitored through 5 yr by the NMDP CRO. Here we report cumulative incidence (CI) data at 2-, 3- and 5-yr for overall survival (OS), relapse, relapse-free survival (RFS), non-relapse mortality (NRM), systemic therapy (ST)-requiring chronic GVHD (cGVHD), and grade 3-4 acute GVHD (aGVHD)-free/ST-requiring cGVHD-free/relapse-free survival (GRFS).

Results: The median follow-up was 53.8 months, 56.1 months and 60 months for the MUD-placebo, MUD-abatacept, and MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr OS was 65.2%, 62.3% and 54.3% for MUD-placebo, 75.2%, 68.0% and 61.0% for MUD-abatacept, and 78.8%, 74.0% and 71.7% for MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr CI of relapse was 25.7%, 27.5% and 27.5% for MUD-placebo, 21.7%, 25.0% and 25.0% for MUD-abatacept, and 9.7%, 9.7% and 12.6% for MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr CI of NRM was 18.1%, 20.0%, 28.7% for MUD-placebo, 14.4%, 18.1% and 22.5% for MUD-abatacept, and 15.0%, 20.2% and 20.2% for MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr CI of RFS was 60.9%, 58.0% and 51.7% for MUD-placebo, 67.0%, 61.4% and 58.1% for MUD-abatacept, and 76.7%, 72.1% and 69.8% for MMUD-abatacept cohorts, respectively. The 1-yr CI of grade 3-4 acute GVHD was 14.8% for MUD-placebo, 8.4% for MUD-abatacept, and 2.3% for MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr CI of ST-requiring chronic GVHD was 49.6%, 52.0% and 52.0% for MUD-placebo, 54.1%, 58.5% and 58.5% for MUD-abatacept, and 72.5%, 75.2% and 78.0% for MMUD-abatacept cohorts, respectively. The 2-, 3- and 5-yr cumulative incidence of GRFS was 27.5%, 26.1%, and 26.1% for MUD-placebo, 27.4%, 21.9% and 21.9% for MUD-abatacept, and 23.3%, 20.9% and 16.3% for MMUD-abatacept cohorts, respectively.

Conclusions: Here we demonstrate that the central outcomes previously reported for ABA2 at 2 yr post-HCT (PMID: 33449816) were durable at 3- and 5-yr, with relatively stable OS, relapse, NRM, cGVHD and RFS. These data are important given new CIBMTR comparisons demonstrating that, for both MUD and MMUD recipients at 2 yr, abatacept/CNI/MTX is superior for OS and RFS compared to both CNI/MTX and CNI/MTX/ATG, and is similar to post-transplant cyclophosphamide (PT-Cy, Blood 2024, PMID:39028876). Moreover, the 3-yr outcome data for MMUD patients presented here for ABA2 also appear similar to the recently reported 3-yr outcomes for myeloablative conditioning (MAC) and PT-Cy for the MMUD-15 trial (PMID: 36584941, 5-yr outcomes not yet reported for MMUD-15). Most notable is the fact that despite increased cGVHD in ABA2 compared to MMUD-15, the GRFS is similar in both trials (20.9% GRFS at 3 yr for ABA2 versus 16.9% for PT-Cy after MAC HCT). These similar composite endpoints may be driven by lower apparent relapse rates in ABA2 (9.7% at 3-yr for ABA2 and 50.5% at 3-yr for MAC PT-Cy in MMUD-15). Together, these data lead to two key conclusions: (1) That the addition of abatacept continues to abrogate historic disparities in OS and RFS for HLA-mismatched vs HLA-matched HCT (extending our previous comparison (PMID:34753172) to 5-yrs post-HCT). (2) That a head-to-head randomized trial of abatacept- versus PT-Cy-based prophylaxis is warranted.