Risk Factors for Graft Failure and Mortality after HLA-Matched Sibling Donor Transplant for Severe Aplastic Anemia in Brazil.

Abstract

Listen

Translate article

Higher rates of graft failure observed after conventional cyclophosphamide (Cy) at 200mg/kg and HLA-matched sibling bone marrow transplant (BMT) for severe aplastic anemia (SAA) in South America prompted the use of a different conditioning regimen: busulfan (Bu) at 12 mg/kg plus Cy at 120 mg/kg. This change was instituted to provide greater immunosuppression for heavily transfused patients (>15 blood transfusion units pre-BMT). We report transplant outcomes in 269 SAA patients who received their HLA-matched BMT at a single center in Brazil between 1990 and 2003. Median age at transplant was 19 years (range 2 – 45). Median time from diagnosis to transplant was 3 months; 78% were transplanted ≤ 6 months from diagnosis, 11%, 7–24 months and 11% > 24 months. 128 patients received Cy alone, 2, Cy plus anti-thymocyte globulin and 139, Bu plus Cy. Eighty-one patients (62%) who received Cy conditioning also received ≤15 blood transfusion units and 49 (38%) received >15 blood transfusion units pre- BMT; all patients who received Bu plus Cy conditioning received >15 blood transfusion units. All patients received T-cell replete bone marrow grafts and almost all patients received cyclosporine and short course methotrexate for graft-versus-host disease prophylaxis. Median follow-up of surviving patients after Cy is 9 years and after Bu plus Cy, 6 years. This reflects the change in practice after 1994, when Bu plus Cy was favored for patients who received >15 blood transfusion units pre-BMT. Most patients achieved neutrophil recovery; the day-60 probabilities of recovery were 95% and 86% after Cy conditioning (≤15 and >15 blood transfusion units, respectively) and 92% after Bu plus Cy. Corresponding day-100 probabilities of platelet recovery were 95%, 86% and 87%. Thirty-nine patients subsequently lost their graft; failure rates were similar after Bu plus Cy and Cy conditioning in patients who received ≤15 blood transfusion units (13% vs. 17%; RR 0.97, p=0.924). Though failure rates were higher (10 of 42, 24%) after Cy conditioning in patients who received >15 blood transfusion units this did not attain statistical significance when compared to those who received Bu plus Cy (16 of 128, 13%, p=0.077). This may be explained by limited numbers of patients in the heavily transfused Cy group. Younger age (≤10 years) was the only factor that was associated with higher rates of secondary graft failure (RR 2.91, p=0.001). As expected, the 8-year probability of overall survival was highest (87%) after Cy conditioning in patients who received ≤15 blood transfusion units. Mortality rates were higher after Bu plus Cy conditioning (RR 3.18, p<0.001) and Cy conditioning in patients who received > 15 blood transfusion units (RR 4.66, p<0.001). The corresponding 8-year probabilities of overall survival were 67% and 51% (p=0.059). The data suggest secondary graft failure rates are similar after Bu plus Cy conditioning in patients who received >15 blood transfusion units compared to those who received Cy conditioning and fewer transfusions. Though survival rates are generally lower in patients who receive >15 blood transfusion units, the use of Bu plus Cy appears to confer a survival advantage. Patients with SAA should be referred for transplantation as soon as the diagnosis is confirmed to avoid excess blood transfusion pre-BMT.