Clinical characteristics and outcomes of Epstein-Barr virus viral load after allogeneic hematopoietic stem cell transplantation

This was a single-center, retrospective, observational study. A total of 121 consecutive patients who underwent allo-HSCT between December 2011 and February 2022 and EBV DNA quantification post-transplantation were included. Data on patients’ clinical findings were obtained from electronic medical records. The dataset was locked on February 28, 2022. Only cases in which the EBV viral load was measured over time after transplantation were included. This study was conducted in accordance with the Declaration of Helsinki 1964 and its later amendments or comparable ethical standards. This was approved by the Ethical Review Committee of the Japanese Red Cross Narita Hospital. This study was a retrospective observational study and obtaining written consent was not mandatory. All participants did not express any refusal to the study documents published on the Web using the opt-out method. The opt-out method was approved by our institutional review board.

Antibody titers of EBV prior to transplantation were checked to confirm previous infections in all patients. EBV DNA was quantified in whole blood. The basic measurement frequency is once every 2 weeks or once a week until 100 days post-transplant, and once a month after 100 days. On the other hand, in cases with frequent visits or those receiving large amounts of immunosuppressants continuously due to GVHD, EBV was sometimes measured weekly, even after 100 days post-transplantation.

In some cases, EBV-infected cell identification analysis of peripheral blood was performed using multicolor flow cytometry (FCM) to test whether the infected cells were of B or T/NK cell origin. Identification of infected cells by FCM was performed at the discretion of the attending physician. Specifically, FCM is used to test for the loss of CD20 expression for the identification of EBV-infected cells. This was performed in some cases that did not respond well to rituximab.

Diseases that required and actually performed transplantation were included in the study. Malignant diseases included leukemia, myelodysplastic syndrome, myeloproliferative neoplasms, and malignant lymphoma, while benign diseases included aplastic anemia.

Patients who underwent allo-HSCT for chronic active EBV infection and EBV-HLH, wherein EBV initially infected T/NK cells, were excluded. Donor sources included bone marrow, peripheral blood stem cells, and cord blood (CB). There were no 8/8 HLA match in CB and all CB were counted as mismatched unrelated donor (MMUD) for donor type.

We grouped patients on whether EBV DNA levels reached > 1000 copies/mL in whole blood during follow-up (N = 50) or not (N = 71). Previous reports have used a cutoff of 1000 to over 40,000 copies/mL, which is an extensive range, and we used 1000 copies/mL as the cutoff to examine trends in a larger number of patients. We analyzed the relationship between the rate of ISS reduction and prognosis for the group that had > 1000 copies/mL. We also examined the ISS dose that could be reduced after the EBV viral load reached its maximum level. In particular, we examined whether existing ISS could be reduced to 50% of the dose within 3 months, considering the period when post-transplant EBV-LPD (PT-EBV-LPD) is most likely to occur late in the course of post-allo-HSCT. Even if the ISS dose could be reduced by 50% within 3 months, if the dose was subsequently increased after dose reduction, it was considered a case of immunosuppressive dose reduction failure. Although calcineurin inhibitors were the main immunosuppressive drugs, we also assessed whether the dosage of steroids and MMF could be reduced by 50% in 3 months. If the dosage was not reduced by 50%, it was considered as an ISS dose reduction failure. The decision to reduce the ISS dose was at the physician’s discretion and undertaken considering the EBV-viral load status, GVHD status, infection, and risk of disease recurrence. As a rule, ISS dose reductions were actively implemented when GVHD findings improved or disappeared, when severe viral infections occurred, or when signs of disease recurrence appeared. The use of rituximab was considered when the EBV viral load was particularly high, such as > 100,000 copies/mL, but it has also been used in cases of aggressive LPD like malignant lymphoma.

Disease-free survival (DFS) was calculated as the time interval from the date of transplantation to the date of first progression, recurrence, or death without progression. Overall survival (OS) was calculated as the time interval from the date of transplantation to the date of death from any cause.

Baseline clinical characteristics were compared between the groups using the Mann–Whitney U test for continuous variables and Fisher’s exact test for categorical variables. Survival was analyzed using the Kaplan–Meier method and compared using the log-rank test to evaluate OS and DFS. Statistical significance was set at p < 0.05. Aplastic anemia was excluded from the prognosis stratified by complete remission (CR)/not CR. Since there are many cases of early death after transplantation, OS/DFS analysis by the landmark method at 3 months/6 months was also performed in some analysis. The prognostic impact was evaluated using univariate Cox proportional hazard analyses. The Scatter Plot was used for assessing the relationship between the two continuous variables. All statistical analyses were conducted using the EZR software (Saitama Medical Center, Jichi Medical University) [14].

The clinical characteristics of patients based on the presence of a high EBV viral load (> 1000 EBV copies/mL) are summarized in Table 1. The median age of all patients was 53.0 years, and 64.5% were male. EBV-DNA in 50 patients (41.3%) was > 1000 EBV copies/mL, while 71 (58.6%) did not. The underlying diseases were acute myeloid leukemia/myelodysplastic syndrome (68.6%), chronic myeloid leukemia (2.5%), myeloproliferative neoplasms (3.3%), acute lymphoblastic leukemia (16.5%), non-Hodgkin lymphoma/Hodgkin lymphoma (6.6%), and aplastic anemia (2.5%). Complete remission (CR) was achieved in 51.2% of the patients at the time of transplantation. Among patients with > 1000 EBV copies/mL, 48% achieved CR at the time of transplantation, compared to 53.5% in patients with < 1000 EBV copies/mL. There were no differences in conditioning, Eastern Cooperative Oncology Group (ECOG)- performance status (PS), chronic GVHD, or cytomegalovirus activation after transplantation between the patients with > 1000 EBV copies/mL and patients with < 1000 EBV copies/mL. In contrast, patients with > 1000 EBV copies/mL had a higher frequency of anti-thymocyte globulin (ATG) haplo-transplantation, acute GVHD (grade 2 or higher), and peripheral blood stem cells as the graft source.

We then performed an additional analysis of the 50 patients with a high EBV viral load (> 1000 copies/mL) after transplantation (Table 2). Sixty percent of the patients reached the maximum EBV DNA level before post-transplant day 100, while 38% reached the maximum level between days 100 and 1000, and 2% reached maximum level after day 1000. The viral load was 3log (> 1000 copies/mL) in 52%, 4log (> 10,000 copies/mL) in 32%, 5log (> 100,000 copies/mL) in 8%, and 6log (> 1,000,000 copies/mL) in 8% of the patients. PT-EBV-LPD was diagnosed using lymph node biopsy or based on new multiple lymphadenopathies on CT in 10% of cases. The median viral load in patients with LPD was 1,500,000 EBV copies/mL (range 76,000–4,900,000 copies/mL). Tacrolimus and cyclosporine were used in 28% and 56% of the cases, respectively, when EBV reached the maximum viral load. Following maximum EBV increase, rituximab was used in six cases (12%), and 30 cases (60%) showed ISS dose reduction to 50% within 3 months. The percentage of patients with lymphocytes at least 1000/µL at maximum detection was 22/50 (44.0%). No correlation was observed between lymphocyte counts and maximum EBV-DNA levels, with a Spearman’s rank correlation coefficient − 0.066 (Online Resource.1).

The median follow-up period was 16.2 (1.17–132.17/SD 34.6) months. EBV-PTLD began to appear as early as 3–6 months after transplantation in adults [15], and the previous report showed 3-year OS 47.3% for EBV-LPD after transplantation [16]; thus, we set the OS time point at approximately 20 months, which is close to the median observation period.

The median and 20-month OS rate of all patients were “not reached” and 58.7% (95% confidence interval [CI], 49.3–67.0%), respectively (Fig. 1A). The median and 20-month DFS rate of all patients were 74.8 months and 54.6% (95% CI, 45.1–63.0%), respectively (Fig. 1B). Patients who achieved CR before allo-HSCT showed significantly longer OS and DFS than those in patients who did not (20-month OS, 73.5% vs. 42.0% [95% CI, 60.4–82.9 vs. 28.9–54.6]; p = 0.0003; Fig. 2A; 20-month DFS, 68.3% vs. 38.5% [95% CI, 54.8–78.5 vs. 25.8–51.1]; p = 0.0002; Fig. 2B).

Patients with good PS (PS < 1) before allo-HSCT showed significantly longer OS and DFS than patients with poor PS (PS ≧ 1) before allo-HSCT (20-month OS, 69.0% vs. 45.1% [95% CI, 56.5–78.6 vs. 31.2–58.0]; p = 0.0189; Fig. 2C; and 20-month DFS, 65.8% vs. 39.2% [95% CI, 53.1–75.9 vs. 26.0–52.2]; p = 0.0021; Fig. 2D).

The 20-month OS (Fig. 3A) and DFS (Fig. 3B) were not significantly different between patients with < 1000 EBV copies/mL and patients with > 1000 EBV copies/mL (20-month OS, 56.0% vs. 60.6% [95% CI, 41.2–68.4% vs. 48.0–71.1%]; p = 0.503; 20-month DFS, 50.0% vs. 57.7% [95% CI, 35.6–62.8% vs. 45.0–68.4%]; p = 0.179).

Because patients could have died or showed progression/relapse before reaching their highest EBV viral load, a landmark OS analysis was performed at 3/6 months in the EBV viral load analysis. Patients with > 1000 EBV copies/mL were not significantly different in OS or DFS compared to patients with < 1000 EBV copies/mL (3-month landmark/20-month OS, 63.6% vs. 70.4%, p = 0.291, 3-month landmark/20-month DFS, 61.0% vs. 71.8%, p = 0.051, and 6-month landmark/20-month OS, 80.0% vs. 75.5%, p = 0.849, 6-month landmark/20-month DFS, 78.1% vs. 78.9%, p = 0.273) (Online Resource.2).

Further focusing on patients with > 1000 EBV copies/mL, those who achieved 50% ISS dose reduction, including not restarting ISS, within 3 months after HSCT did not have a significant difference in the frequency of chronic GVHD (10.0% vs. 15.0%; p = 0.672) or rituximab use (56.7% vs. 45.0%; p = 0.565) than those who did not achieve 50% dose reduction. However, these patients showed significantly longer OS and DFS than those who did not achieve 50% dose reduction (20-month OS, 80.0% vs. 20.0% [95% CI, 60.8–90.5 vs. 6.2–39.3]; p < 0.0001; Fig. 4A; and 20-month DFS, 70.0% vs. 20.0% [95% CI, 50.3–83.1 vs. 6.2–39.3]; p < 0.0001; Fig. 4B).

In considering dose reduction of ISS, an additional landmark analysis was performed because there were cases of death/progression before ISS dose reduction. Additional analysis focused only on patients who developed EBV reactivation, particularly within 12 months of transplantation. Those who achieved 50% ISS dose reduction showed significantly longer OS and DFS than those who did not achieve 50% dose reduction (3-month landmark/median OS, NR vs. 6.9 months, p = 0.001, 3-month landmark/median DFS, 88.4 months vs. 7.6 months, p = 0.007, and 6-month landmark/median OS, NR vs. 12.6 months, p = 0.025, 6-month landmark median DFS, 88.4 months vs. 12.1 months, p = 0.018) (Online Resource. 3).

Patients who reached the maximum EBV-DNA level over 100 days post-transplantation had better OS and DFS than those who reached the maximum EBV DNA level within 100 days post-transplantation (20-month OS, 75.0% vs. 43.3% [95% CI, 50.0–88.7 vs. 25.6–59.9]; p = 0.015; Fig. 4C; and 20-month DFS, 73.5% vs. 43.3% [95% CI, 47.5–88.1 vs. 25.6–59.9]; p = 0.013; Fig. 4D).

Univariate analysis was performed to identify factors independently associated with survival in patients with > 1000 EBV copies/mL after transplantation (Table 3). In the univariate analysis, achieving 50% dose reduction or not restarting ISS within 3 months after the maximum EBV increase, PS, and disease status were significantly associated with OS. ATG use, achieving 50% dose reduction or not restarting ISS within 3 months after the maximum EBV increase, and PS and disease status were significantly associated with DFS.

In five cases, the amount of EBV DNA in the peripheral blood continued to increase even after ISS ended (Table 4). The age at transplantation ranged from 47 to 66 years, and the background diseases included acute lymphoblastic leukemia (one case), myeloproliferative neoplasms (one case), non-Hodgkin lymphoma (one case), and myelodysplastic syndrome (two cases). Conditioning included myeloablative conditioning in two cases and reduced intensity conditioning in three cases. GVHD prophylaxis included ATG in only one case, and no post-transplant cyclophosphamide haplo-transplantation was done. None of the patients had post-transplant relapse or progression to PT-EBV-LPD. All patients were followed up without rituximab. In all cases, EBV did not disappear, and a high viral load (> 10,000 copies/mL) was maintained.