Introduction: Severe aplastic anemia (SAA) is a life-threatening bone marrow failure disorder curable with allogeneic hematopoietic cell transplantation (HCT). Early efforts with conditioning using cyclophosphamide alone (Cy; 50 mg/kg/day x 4 days) demonstrated a rejection rate of approximately 36% among transfused patients who received unmodified blood products pre-HCT. The addition of horse anti-thymocyte globulin (ATG) then reduced rejection rates down to about 7%. Extensive preclinical studies from our group demonstrated that buffy coat (BC) depleted blood products—removal of the buffy coat layer via centrifugation—lowered the rejection risk in that model from 100% to 35%, and in vitro irradiation of unmodified products with 2,000 cGy reduced the risk from 100% to 10%. These interventions later became routine in blood banks. When combined, they theoretically reduce the rejection risk to approximately 3.5%—even without ATG. Given these advances, the continued need for ATG inclusion in SAA conditioning should be re-evaluated, especially since ATG is thought to not only reduce the risk of graft rejection but also the risk of graft-versus-host disease (GVHD). This analysis specifically addresses whether omitting ATG from conditioning regimens might increase the risk of GVHD.

Methods: We retrospectively identified 392 patients with SAA who received HLA-matched related bone marrow grafts at Fred Hutchinson Cancer Center between 1970 and 2024. GVHD severity was graded by two primary investigators using Seattle criteria for acute GVHD; NIH consensus criteria were used for chronic GVHD. We assessed the association between ATG use and GVHD outcomes using multivariable regression models (logistic for acute GVHD, Cox for chronic GVHD). Covariates in adjusted models included use of BC infusions in addition to marrow, age at transplant, GVHD prophylaxis strategy (monotherapy vs. combination cyclosporine [CSP] and methotrexate [MTX]), acute GVHD grader, and year of HCT. Additional analyses were performed in the subgroup receiving combination GVHD prophylaxis due to strong collinearity between ATG and prophylaxis regimen. Two patients were excluded from acute GVHD analyses due to missing acute GVHD data.

Results: Among 392 patients, 124 (32%) received ATG. The distribution of GVHD prophylaxis was as follows: 53% (208/392) received monotherapy without ATG, 31% (123/392) received ATG with dual-agent prophylaxis (CSP+MTX), 15% (60/392) received dual-agent prophylaxis without ATG, and one patient received ATG with monotherapy. After adjustment for the factors listed above, the odds of grades II-IV acute GVHD after ATG use was 1.48-times that among those who did not receive ATG (odds ratio (OR)=1.48 95% CI [0.50 -4.38]; p=0.483), and the adjusted OR for grades III-IV acute GVHD was 2.67 (95% CI [0.61-11.62]; p=0.190). The adjusted cause-specific hazard ratio for chronic GVHD was 0.89 (95% CI [0.41-1.94]; p=0.764). Among patients in the subgroup who received multi-agent acute GVHD prophylaxis, the above results were qualitatively similar. In particular, the adjusted OR for grades II-IV acute GVHD was 1.87 (95% CI [0.55-6.33]; p=0.312). Due to the limited number of cases of grades III-IV acute GVHD, the corresponding model was adjusted only for age and acute GVHD grader. The adjusted OR for grades III-IV acute GVHD was 2.83 (95% CI [0.64-12.49]; p=0.169). The adjusted cause-specific HR for chronic GVHD in the subgroup was 0.77 (95% CI [0.34-1.77]; p=0.54). The adjusted HR for mortality comparing ATG vs non-ATG regimens was 0.88 (95% CI [0.44-1.75]; p=0.712).

Conclusion: In this single-center cohort of HLA-matched related allo-HCT recipients with SAA, ATG use was associated with a numerically higher incidence of acute GVHD after adjusting for potential confounders. Although ATG was linked to a numerically lower incidence of chronic GVHD, the magnitude of this reduction was modest and may be attributable to chance. Given current practices of BC-poor and in vitro irradiated transfusion products, these findings suggest that the routine inclusion of ATG in the conditioning regimen may not be needed for preventing both graft rejection and GVHD, challenging long-held assumptions about its dual benefit in conditioning regimens. These data support reconsideration of conditioning regimens for HLA-matched related allogeneic marrow grafts in SAA that do not include horse ATG.

This content is only available as a PDF.
2025
Sign in via your Institution