Early Development of Immunity to CMV Following Hematopoietic-Stem Cell Transplantation Is Associated with Graft-Versus Leukaemia Effect

Abstract 4214

Relapse of disease remains the most common cause of failure in children undergoing hematopoietic-stem cell transplantation (HSCT) for acute leukaemia. A favourable impact, on graft-versus leukaemia effect (GVL), from recipient CMV-seropositivity before HSCT and from CMV-reactivations in the recipient after HSCT has been suggested. The potential role, in this process, for the immune response triggered by CMV has not been directly addressed so far.

108 children (median age 8 years) were included at HSCT following myeloablative conditioning for primary acute leukaemia (lymphoblastic 54%, myeloblastic 42%, biphenotypic 4%). HSCT were from HLA-matched related (41%) or unrelated (44%) donors. 15% patients received cord blood units. CMV- DNAemia was programmed weekly for at least 3 months post-HSCT. When PCR showed ≥1000 copies/ml, patients received ganciclovir as pre-emptive therapy. Immunity to CMV was evaluated sequentially since the first month post-HSCT using 3H-Thymidine incorporation assay (T-cell proliferation) and intracytoplasmic cytokine accumulation assay (IFNg secretion).

Median follow-up from transplant was 40 months. 57% of patients were CMV-seropositive before HSCT. Cumulative incidence of recipients with CMV-DNAemia at day 120 was 31% (median time to onset: 26 days). Cumulative incidence of recipients with immunity to CMV (T-cell proliferation and/or IFNγ secretion assays) at 1 year was 38% (median time to onset: 2 months). As expected, recipient CMV seropositivity represented a major factor contributing to DNAemia (p<0.0001) and to immunity (p<0.001) occurrence as well. The 2 years cumulative incidence of relapse was 26% Among the 89 recipients free of relapse at day +120 and evaluated for immunity before that time, the 2 years cumulative incidence of relapse was 15%. Multivariable analysis revealed a different risk of leukemic relapse according to immunity to CMV and CMV-reactivation (p=0.039), with the lowest risk in recipients with early (before day 120) immunity to CMV but no CMV-reactivation and the highest risk in recipients with early (before day 120) CMV-reactivation but no immunity to CMV (HR: 0.06, 95%CI 0.01–0.69, p=0.024).

In conclusion, early development of immunity to CMV rather than early viremia had a favourable impact on GVL in this pediatric series.