If ABO antigens/antibodies play any role in the pathogenesis of acute graft-versus-host disease (GVHD), one would expect the highest incidence of GVHD in recipients of minor ABO-mismatched grafts, followed by ABO-matched grafts, and the lowest incidence in major ABO-mismatched transplants. To test this hypothesis 174 patients receiving an HLA-identical allogeneic bone marrow transplant (BMT) for aplastic anemia (n = 32) or leukemia (n = 142) were analyzed for factors associated with acute GVHD. Variables analyzed included diagnosis, sex, age, blood group of donor and recipient, ABO compatibility, Rhesus compatibility, sex compatibility, number of bone marrow cells given at BMT, year of transplant, day of engraftment, and GVHD prophylaxis. We first carried out an exploratory contingency table analysis: minor ABO incompatibility was associated with a significantly higher risk of severe acute GVHD when compared with ABO-matched and major-ABO mismatched pairs (P = 0.003): 14/9, 57/67, and 5/22 patients developed, respectively, 0-I/II-IV acute GVHD in ABO major-mismatched, matched, and minor-mismatched pairs. Donors of group 0, (P = 0.06), older recipient's age (P = 0.08), fast engraftment (P = 0.03), and older donor's age (0.08) were also associated with a higher risk of GVHD. Recipient's ABO group, diagnosis, year of transplant, Rhesus group of donor or recipient, Rhesus compatibility, sex of donor or recipient, sex compatibility, and type of GVHD prophylaxis were not predictive of GVHD. A Cox multifactorial proportional hazards analysis confirmed that ABO matching was the single most significant factor associated with GVHD (P = 0.006). The cumulative incidence of GVHD grade II+ was 39%, 54%, and 82% for ABO major-mismatched, matched, and minor-mismatched pairs (P = 0.01). This study suggests that ABO antigens may play a role in the development of acute GVHD.