Introduction
EBV reactivation post allogeneic stem cell transplantation (alloSCT) occurs due to profound suppression of T-cell number or function and subsequent reactivation of latent host or recipient EBV infection through loss of immune-surveillance. The relationship of EBV reactivation to patient outcomes has not been reported in patients who have received T-cell depletion (TCD) with Thymoglobulin. This analysis assessed the outcomes in patients who had previously received an alloSCT with TCD using Thymoglobulin and who developed EBV reactivation that either required treatment (EBV Treatment) vs those who did not require treatment (EBV No Treatment) vs those who had no evidence of EBV reactivation (No EBV).
Methods
This was a retrospective review of all patients who received TCD with Thymoglobulin prior to alloSCT at the Royal Melbourne Hospital between 2006-2017. Eligible patients received reduced intensity conditioning (RIC) or myeloablative conditioning (MAC) for acute leukemia (AL) or myelodysplastic syndrome (MDS). EBV was monitored weekly via semi-quantitative PCR . Preemptive rituximab was given based on peak viral loads, rate of rise in viral loads or biopsy proven post-transplant lymphoproliferative disorder (PTLD). Categorical and continuous baseline characteristics of the cohorts were compared using chi-squared and Kruskal-Wallis tests respectively. EBV Treatment was treated as a time-dependent variable and evaluated for its relationship to overall survival (OS) using Cox proportional modelling, adjusted for age and high or very high Disease Risk Index (DRI) and grade III-IV acute GVHD. Survival estimates were performed by Kaplan-Meier product-limit method. Cumulative incidence (CI) of relapse and non-relapse mortality (NRM) were calculated by the Fine-Gray method.
Results
Of 165 eligible patients, 101 (61%) had an episode of EBV reactivation at a median of 46 days post-transplant (range 10-221 days). 24 received therapy for EBV. Of these, 4 had biopsy-proven PTLD, and 23 received rituximab alone with 1 receiving R-CHOP for PTLD. An undetectable EBV viral load was achieved in 20 out of 24. The baseline characteristics of the subgroups are presented in Table 1.
The median survival of the whole cohort was 79 months with an estimated 2-year OS of 60% (95%CI 56-68) and a median follow up of time of 81 months in survivors. The median survival of the EBV Treatment group was 14 months, compared with 121 months and 81.1 months in the EBV No Treatment and No EBV cohorts respectively (Figure 1). Table 2 lists causes of death per group with relapse and complications of poor graft function (PGF) being the leading causes of death in the EBV Treatment group. The incidence of grade III-IV acute GVHD at 100 days was similar between groups as was the incidence of any chronic GVHD at 6 and 12 months. EBV Treatment was significantly associated with poorer OS even when adjusting for age, high/very high DRI and presence of acute GVHD (HR 2.47 CI 1.46-4.17; p=0.0007). The EBV Treatment group had a significantly higher 2 year CI of NRM compared to the EBV Treatment and No EBV groups (38% vs 11% vs 18% p=0.006) but similar incidence of relapse (Figure 2) .
Conclusion
In this cohort of AL/MDS patients undergoing TCD alloSCT, those who required treatment for EBV reactivation experienced poorer survival due to increased NRM. This increase in NRM may reflect impaired immune reconstitution of which significant EBV reactivation is a manifestation. Further studies to understand the underlying immune defects post alloSCT that predispose patients to significant EBV reactivation and associated complications is warranted to identify immune recovery strategies that may be employed in this poor prognosis group.
No relevant conflicts of interest to declare.
