EBV-related post-transplant lymphoproliferative disease (EBV-PTLD) is a serious complication of allo-HCT, for which antithymocyte globulin (ATG) administration has been identified as the most important predisposing factor. Although the development of EBV-PTLD used to be detrimental to patients, use of rituximab has largely changed both the incidence and the outcome of this complication, either administered as preemptive treatment based on peripheral blood (PB) EBV titles, or for the management of the disease itself.

We determined the incidence of EBV-PTLD through chart review of 797 consecutive allo-HCT recipients transplanted in our center (7/1990-7/2018) and evaluated factors potentially influencing EBV-PTLD occurrence and outcome.

Among 797 allo-HCT recipients (n=465 sibling, n=277 MUD, n=47 haploidentical, n=2 twin and n=6 cord blood), 14 (1.7%) patients developed EBV-PTLD. The diagnosis was confirmed by biopsy in 9/12 cases (7 monomorphic, 1 polymorphic and 1 HL-like PTLD), 1 case developed sole EBV encephalitis, 1 case developed CNS lymphoma, and the remaining 3 cases were clinically diagnosed on the basis of generalized lymphadenopathy coupled with high PB EBV title. The patients suffered from ALL (n=6), AML (n=2), CML (n=2), severe aplastic anemia (n=2), MDS (n=1) and plasmacytic leukemia (n=1). Notably, 6/14 patients had high tumor burden at the time of transplantation [refractory AML (n=1), relapsed refractory ALL (n=2), CML blastic crisis (n=2), refractory MDS RAEB II (n=1)]. EBV-PTLD incidence was significantly higher in MUD versus sibling allo-HCT (3.4% versus 0.6%, p=0.006), in haploidentical versus sibling allo-HCT (6.4% versus 0.6%, p<0.001), and in haploidentical versus conventional allo-HCT (6.4% versus 1.5%, p=0.01). Among the haploidentical allo-HCTs complicated with EBV-PTLD, 2 were T-cell depleted with add-back infusions of transduced donor lymphocytes, and 1 was T-cell replete. These results possibly reflect the standard use of ATG as acute GVHD prophylaxis in MUD and haploidentical allo-HCT. Indeed, 12/14 (85.7%) patients who developed EBV-PTLD had received ATG as part of their conditioning regimen, at doses ranging from 5-10mg/kg (median 5mg/kg), and 1 additional patient had received ATG 20mg/kg as 2nd-line treatment for hyperacute GVHD. The conditioning intensity was not significantly associated with EBV-PTLD development [classic myeloablative (n=8), reduced-toxicity (n=3), reduced-intensity (n=3)]. EBV-PTLD occurred early in the post-transplant period (median: 73, range 41-603 days). In 7/14 cases, it was preceded by the onset of aGVHD (median interval from aGVHD diagnosis: 58 days), in one case it coincided with aGVHD diagnosis +45 days from allo-HCT, and in one case it developed soon (+55 days) after induction of GVHD through DLIs. At EBV-PTLD diagnosis, the median EBV title in PB was 78,400 copies/ml (range 84-2,860,000). Of note, one patient developed sole CNS EBV-PTLD with extremely low PB EBV viral load. This patient had received preemptive rituximab 500mg 7 months prior to CNS EBV-PTLD development, on the basis of elevated PB EBV title (70,900 copies/ml). Since then, PB EBV title was always tested below <500 copies/ml by conventional PCR, and was practically negative at the time of EBV-PTLD diagnosis (84 copies/ml). Ten patients were treated with intravenous rituximab: in six patients the disease resolved, but 3/4 cases with CNS involvement succumbed. The fourth patient with CNS involvement is currently under treatment with high-dose MTX and intravenous rituximab and has achieved partial remission. The remaining four patients were unsuccessfully treated with combinations of bleomycin/vindesine/IFNa/immunoglobulins/DLIs.

In conclusion, EBV-PTLD is an early complication of alloHCT, associated mainly with defects in T cell immunity. Selective ATG administration to MUD and haploidentical allo-HCT may explain the higher incidence of EBV-PTLD as compared to sibling allo-HCT, yet we identified higher incidence among haploidentical alloHCTs that cannot be attributed to a particular lymphodepletion scheme. Prompt administration of intravenous rituximab is very effective, except for CNS involvement. Importantly, PB EBV monitoring may be misleading in CNS EBV-PTLD, particularly following preemptive intravenous rituximab administration. Taking into account the dismal outcome, intrathecal rituximab should be considered.

Disclosures

Vardi:Gilead: Research Funding; Janssen: Honoraria. Gavriilaki:European Hematology Association: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution