Chronic Graft-Versus-Host Disease and Its Association with Treatment-Related Mortality, Relapse, Leukemia-Free and Overall Survival After Umbilical Cord Blood Transplantation (UCBT) In Children and Adolescents with Acute Leukemia

Abstract 213

Chronic graft-versus-host disease (cGVHD) is the leading cause of late transplant-related mortality (TRM) after unrelated adult donor transplantation. However, limited, mild or moderate cGVHD may be associated with higher survival in part due to an association with lower relapse risk in adults with acute and chronic leukemia. While cGVHD rates are generally lower in children, we sought to determine the impact of cGVHD after UCBT on risk of relapse, TRM, leukemia-free survival (LFS) and overall survival (OS) as well as to identify risk factors for cGVHD after UCBT. Patients <18 years with acute lymphoblastic or acute myeloblastic leukemia transplanted with a single UCB unit after myeloablative conditioning between 1995–2007 who survived without recurrent leukemia for more than 3 months were eligible for analysis. Of the 657 patients, 59% were male, 26% were >10 years of age, 24% were transplanted in third remission or beyond, 74% were conditioned with TBI, 83% had pre-transplant anti-thymocyte globulin and 11% received tacrolimus, rather than cyclosporine, for acute GVHD prophylaxis. The cumulative incidence of cGVHD was 25% at 5 years with mild, moderate and severe disease in 91, 38 and 19 patients, respectively. cGVHD rates did not vary with donor-recipient HLA disparity. We constructed Cox proportional hazard-regression models to examine the effects of cGVHD on TRM, relapse, treatment failure (relapse or death; inverse of LFS) and overall mortality adjusting for age, disease, disease status, transplantation period, GVHD prophylaxis and donor-recipient HLA match. The risks of TRM (hazard ratio [HR] 9.33, p<0.0001), treatment failure (HR 2.96, p<0.0001) and overall mortality (HR 4.79, p<0.0001) were significantly higher in patients who developed cGVHD compared to those who did not develop cGVHD. In contrast to reports in adults transplanted with bone marrow or peripheral blood, mortality risk did not vary in those with mild, moderate or severe cGVHD after UCBT. Notably, relapse risk was not altered by the development of cGVHD (HR 0.73, p=0.38). In multivariate analysis, after adjusting for age and transplantation period, cGVHD risks were higher in patients who received tacrolimus (HR 3.12, p<0.0001) and in those with a prior history of grade 2 acute GVHD (HR 2.04, p=0.0001) or grade 3–4 acute GVHD (HR 5.51, p<0.0001) with a higher risk of cGVHD in patients with grade 3–4 versus grade 2 acute GVHD (HR 2.70, p<0.0001). These data suggest cGVHD of any severity has a markedly negative impact on survival after UCBT in children with acute leukemia without the associated graft-versus-leukemia effect previously reported for other stem cell sources. This analysis compels us to reevaluate our current acute GVHD prophylaxis strategies, including how tacrolimus is typically used, in an attempt to reduce the risk of cGVHD and its deleterious effect on LFS and OS after UCBT.