Absence of posttransplantation lymphoproliferative disorder after allogeneic blood or marrow transplantation using posttransplantation cyclophosphamide as graft-versus-host disease prophylaxis

Abstract

Immunosuppressive regimens that effectively prevent graft-versus-host disease (GVHD) after allogeneic blood or marrow transplantation (alloBMT) have been associated with an increased incidence of posttransplantation lymphoproliferative disorder (PTLD) in the first year after transplantation. We evaluated the incidence of PTLD associated with the use of high-dose, posttransplantation cyclophosphamide (PTCy) as GVHD prophylaxis. From 2000-2011, 785 adult alloBMT patients received PTCy as GVHD prophylaxis at the Johns Hopkins Hospital, including 313 who received PTCy as sole GVHD prophylaxis. HLA-haploidentical or unrelated donor grafts were used for 526 (67%) patients. There were no cases of PTLD in the first year after alloBMT. PTLD is rare after alloBMT using PTCy, even in high-risk alternative donor transplants.

Introduction

The approach to graft-versus-host disease (GVHD) prophylaxis, including methods of graft manipulation and pharmacologic immunosuppression, impacts immune reconstitution after allogeneic blood or marrow transplantation (alloBMT) and modulates the risk for posttransplantation lymphoproliferative disorder (PTLD). PTLD risk peaks in the early months after alloBMT and the vast majority of cases occur within the first year after transplantation.[1] PTLD lesions that arise after alloBMT typically harbor Epstein-Barr virus (EBV).

In a multi-institutional analysis of PTLD incidence among 26,901 alloBMTs performed at 271 institutions worldwide, selective T-cell depletion of grafts and to a lesser extent other approaches to lymphocyte depletion were associated with increased risk for PTLD.[1] Patients who receive anti-thymocyte globulin (ATG) as GVHD prophylaxis are also at higher risk for PTLD.[1, 2] The use of alternative donors, including unrelated (URD), HLA-haploidentical (haplo), and umbilical cord blood (UCB), has been associated with higher rates of PTLD. This heightened risk for PTLD seems mostly related to the intensified GVHD prophylaxis strategies often employed with alternative donor grafts. Patients receiving URD or haplo grafts accompanied by ATG and/or selective T-cell depletion have an incidence of PTLD of 4-8%.[1] The incidence of PTLD after UCB alloBMT is 2-7%, but rates upwards of 17% have been reported with UCB alloBMTs using ATG.[2-5]

The use of posttransplantation cyclophosphamide (PTCy) as GVHD prophylaxis after alloBMT is associated with comparatively low rates of severe acute and chronic GVHD, even with URD or haplo grafts.[6] However, given that approaches that are highly effective in preventing GVHD can be associated with higher rates of PTLD, we have evaluated the incidence of PTLD associated with PTCy at the Johns Hopkins Hospital (JHH).

Materials and Methods

After JHH Institutional Review Board approval, the JHH Transplant Research databases were queried for patients age 18 or older who received PTCy as GVHD prophylaxis at JHH between January 1, 2000 to December 31, 2011. Clinical notes, pathology reports, and radiology reports up through the first year after alloBMT were reviewed.

For myeloablative conditioning, patients received a regimen of busulfan (Bu) and cyclophosphamide (Cy) as previously described.[7] Reduced-intensity conditioning (RIC) regimens were fludarabine (Flu)-based, with the majority (n=354) consisting of Flu/Cy/total body irradiation (TBI).[8] Other RIC regimens were Flu/TBI (n=72), Bu/Flu (n=14), ATG/Flu/Cy/TBI (n=14),[9] or alemtuzumab/Flu/TBI (n=2). All grafts were T-cell replete.

When used as sole GVHD prophylaxis, PTCy 50 mg/kg/day was given intravenously on days 3 and 4 after alloBMT,[7] with the exception of 11 patients who received PTCy on day 3 only. Most recipients of haplo grafts and/or RIC received additional immunosuppression following PTCy on days 3 and 4 after alloBMT, with the majority receiving mycophenolate mofetil and tacrolimus. GVHD prophylaxis regimens are shown in Table 1.

Table 1.

Patient and transplantation characteristics.

	AlloBMTs n=785
Age, median (range)	52 (18-75)
Male sex, n (%)	448 (57%)
Diagnosis, n (%)
Acute myeloid leukemia or myelodysplastic syndrome	308 (39%)
Mature B-cell neoplasm	236 (30%)
Acute lymphoblastic leukemia	69 (9%)
Myeloproliferative neoplasm, including CML	52 (7%)
Hodgkin lymphoma	48 (6%)
Mature T-cell or NK-cell neoplasm	36 (5%)
Hemoglobinopathy or thalassemia	17 (2%)
Aplastic anemia or paroxysmal nocturnal hemoglobinuria	9 (1%)
Bilineage leukemia	7 (1%)
Histiocytic or dendritic cell neoplasm	2
Lymphomatoid granulomatosis	1
Conditioning regimen, n (%)
Reduced-intensity	456 (58%)
Myeloablative	329 (42%)
ATG-containing conditioning regimen, n (%)	14 (2%)
Donor, n (%)
HLA-haploidenticala	396 (50%)
HLA-matched, related	259 (33%)
HLA-matched or 1 antigen mismatched, unrelated	130 (17%)
Allograft source, n (%)
Marrow, T-cell replete	748 (95%)
Peripheral blood, T-cell replete	37 (5%)
GVHD prophylaxis, n (%)
PTCy alone	313 (40%)
PTCy on days +3 and +4	302
PTCy on day +3	11
PTCy on days +3 and +4/MMF/tacrolimusb	382 (49%)
PTCy on days +3 and +4/MMFc	55 (7%)
PTCy on day +3/MMF/tacrolimusd	20 (2%)
PTCy on days +3 and +4/MMF/sirolimuse	15 (2%)
Rituximab ≤ 12 months after alloBMT, n (%)	63 (8%)
HLA-haploidentical	55/396 (14%)
HLA-matched, related	8/259 (3%)
HLA-matched or 1 antigen mismatched, unrelated	0/130
DLI ≤ 12 months after alloBMT, n (%)	57 (7%)
HLA-haploidentical	13/396 (3%)
HLA-matched, related	41/259 (16%)
HLA-matched or 1 antigen mismatched, unrelated	3/130 (2%)

Abbreviations: alloBMT, allogeneic blood or marrow transplantation; CLL, chronic lymphocytic leukemia; CML, chronic myelogenous leukemia; NK, natural killer; ATG, anti-thymocyte globulin; HLA, human leukocyte antigen; GVHD, graft-versus-host disease; PTCy, posttransplantation cyclophosphamide; MMF, mycophenolate mofetil; DLI, donor lymphocyte infusion

^a395 related donors, 1 unrelated donor

^bMMF 15 mg/kg, 3 times daily up to 1000 mg/dose on days 5-35 and tacrolimus (goal trough 5-15 ng/mL) on days 5-180. Twenty patients received MMF 15 mg/kg twice daily on days 4-33.

^cMMF 15 mg/kg, twice daily on days 4-33

^dMMF 15 mg/kg twice daily on days 4-35, tacrolimus (goal trough 5-15 ng/mL) on days 4-50 (n=10) or days 4-180 (n=10)

^eMMF 15 mg/kg, 3 times daily up to 1000 mg/dose on days 5-35 and sirolimus daily (goal trough 5-15 ng/mL) on days 5-365

PTLD was defined using the 2008 World Health Organization classification.[10] Data on the use of post-alloBMT therapies that could modulate PTLD risk, including donor lymphocyte infusion (DLI) and rituximab therapy, were compiled. Death was considered a competing risk for PTLD. The R program, version 2.15.2 (R Core Development Team, Vienna, Austria) was used to estimate the cumulative incidence of death.

Results

From 2000-2011, 785 alloBMTs were performed using PTCy as GVHD prophylaxis (Table 1). PTCy was used as sole GVHD prophylaxis in 313 (40%) alloBMTs, 295 of which utilized myeloablative conditioning and HLA-matched grafts. Haplo or URD grafts were used in 67% (n=526) of the alloBMTs.

Almost all patients (n=762, 97%) had sufficient follow-up to assess for PTLD in the year following alloBMT, with 505 followed for at least one year by JHH physicians, 203 surviving less than one year but followed until death or hospice enrollment, and 54 resuming care with an outside oncologist before one year post-alloBMT, but corresponding with JHH when clinical issues arose. The cumulative incidence of death at one year was 34% (95% confidence interval (CI) 29-38%) for haplo alloBMTs, 30% (95% CI 22-38%) for URD alloBMTs, and 29% (95% CI 23-34%) for matched, related donor alloBMTs (Figure 1). There were no cases of PTLD identified among the 762 patients with adequate follow-up. Twenty-three patients had insufficient follow-up to assess for PTLD through one year post-alloBMT, but no cases of PTLD were identified during the available follow-up (median 3.7 months, range 1.6-11.9 months).

Figure 1. Cumulative incidence of death by allograft type.

At one year, the cumulative incidence of death was similar for recipients of HLA-haploidentical, unrelated, and HLA-matched related donor allografts at 34% (95% CI 29-38%), 30% (22-38%), and 29% (23-34%), respectively. The cumulative incidence of PTLD is not shown given the absence of events.

Therapies such as DLI, adoptive EBV-specific cellular immunotherapy, and rituximab are effective in treating PTLD. In our cohort, DLI was used in 57 patients within the year following alloBMT, but in no case was DLI given for suspected or confirmed PTLD. Adoptive immunotherapy with EBV-specific cytotoxic T-lymphocytes was not used in this cohort. Sixty-three patients received rituximab during the year following alloBMT. These included 43 patients with CD20(+) lymphomas who received 8 weekly doses of rituximab upon engraftment after RIC alloBMT (41 haplo and 2 HLA-matched related donor grafts) as part of a clinical protocol, 19 patients who received rituximab as part of treatment for relapsed B-lineage malignancy, and one patient who received one dose of pre-emptive rituximab for an asymptomatic, modest elevation of EBV-DNA copy number in the blood 4 months after alloBMT at an outside institution in keeping with that institution's protocol for post-alloBMT EBV monitoring. The patient's elevation in EBV-DNA copy number was not associated with fever, lymphadenopathy, organ dysfunction or other clinical findings concerning for PTLD. EBV DNA monitoring has continued and viral DNA is intermittently detected but the patient remains without signs or symptoms of EBV-associated disease. The patient remains healthy nearly 3 years after alloBMT.

Discussion

The absence of PTLD after alloBMT using PTCy is noteworthy, particularly given that no routine EBV monitoring was performed. Many transplant centers employ EBV monitoring and administer rituximab preemptively in attempts to decrease the risk for PTLD after alloBMT, particularly in high-risk patients.[3, 5, 11-13] Because graft elutriation was used at JHH prior to the implementation of PTCy-based approaches and the associated rate of PTLD was very low,[14, 15] the monitoring of EBV DNA in the blood after alloBMT has not been an institutional standard.[16] As EBV DNA is commonly detected in the blood after alloBMT with wide variation in copy number among patients and in individual patients over time, elevations of EBV DNA in the blood after low-risk alloBMTs should be interpreted with caution.[17, 18] In asymptomatic patients, even sustained elevations of EBV DNA in the blood after alloBMT lack specificity for predicting the development of PTLD.[18]

In reports from other centers that use PTCy, collectively comprising 162 alloBMTs, no cases of PTLD have been reported.[19-24] Of these, 63 alloBMTs were performed at centers that monitor EBV.[20, 24] One center reported that 1 of 13 patients transplanted with PTCy developed elevated EBV-DNA in the blood and no patient developed PTLD, contrasting this finding with a 12% incidence of PTLD at their center with other approaches.[24] The second center reported that 4 of 50 patients developed elevated EBV-DNA in the blood, one of whom was found to have EBV gastritis.[20, 25] While the data are limited, interpretation of EBV-DNA copy number in the blood after alloBMT with PTCy remains uncharted territory.

The absence of PTLD in the year following alloBMT in this series suggests that effective GVHD prophylaxis need not be associated with an increase in PTLD. Our institution and others have reported the effectiveness of PTCy as GVHD prophylaxis in patients receiving HLA-matched related, HLA-matched URD, or haplo grafts, with the cumulative incidence of grade III-IV acute GVHD ranging from 6-12%.[6-8, 20, 21, 23] Thus, it appears that the prevention of PTLD with PTCy does not come at the cost of higher GVHD rates.

In considering death as a competing risk for PTLD, a large majority of patients in this series survived at least one year post-alloBMT, including patients at higher risk for PTLD (recipients of haplo or URD grafts). Furthermore, the vast majority was alive in the first 6 months after alloBMT, when the risk for PTLD is highest. Given that this series includes hundreds of patients who survived more than one year and thus remained at risk for PTLD, it seems unlikely that early deaths would completely account for the absence of PTLD in this series.

Several hypotheses might be considered to explain the lack of PTLD with PTCy, including destruction of donor and host EBV-infected B-cells, relative sparing of EBV-specific memory T-cells, or rapid immune reconstitution, and investigations into these mechanisms including EBV-specific T-cell recovery after alloBMT using PTCy are ongoing. Whatever the explanation(s), PTCy results in very low to absent rates of PTLD, even with unrelated or HLA-mismatched donors.