Prophylactic LTV was implemented at Nanfang Hospital, Southern Medical University from March 2022 for CMV-seropositive allo-HSCT recipients. LTV prophylaxis was started on the second day of neutrophil engraftment and was continued until day 100 with a dosage of 480 mg daily or 240 mg if concurrent cyclosporin A was used. Consecutive CMV-seropositive haplo-HSCT recipients who received LTV primary prophylaxis for CMV prophylaxis from March 2022 to December 2023 were analyzed in this retrospective study as the LTV cohort. For the non-LTV cohort, propensity score matching (PSM) for baseline variables was used and CMV-seropositive haplo-HSCT recipients without LTV between January 2020 and November 2023 were included. This study was approved by the Medical Ethics Committee of Nanfang Hospital.

For all the recipients, the CMV and EBV-DNA loads in the blood were measured regularly by real-time quantitative polymerase chain reaction weekly for the first 3 months after transplantation, once every 2 weeks from the 4^th to the 9^th month post-transplantation and then once per month from the 10^th to the 12^th month. Once CMV or EBV-DNA in the blood was positive, the viral loads were detected once again the next day. If positive, viral loads were monitored twice a week. The preemptive treatment threshold is defined as either two consecutive positive CMV PCR results within one week or a single result >500 copies/mL. First-line therapies include ganciclovir and valganciclovir, while second-line options consist of foscarnet, cidofovir, immunoglobulin and cytotoxic T lymphocytes (CTLs) (2022;Xuan et al., 2012). Upon initial detection of positive EBV-DNA in blood (>500 copies/mL), repeat viral load testing was performed the following day. For patients with two consecutive positive EBV-DNA results, the EBV preemptive strategy was initiated, including antiviral therapy (ganciclovir, acyclovir, or foscarnet), intravenous immunoglobulin (0.4 g/kg/day × 3 days), or immunosuppression reduction. In cases of persistently positive EBV-DNA across four consecutive tests with an upward trend, weekly rituximab (375 mg/m²) was administered until viral clearance or for a maximum of 4 weeks (Xuan et al., 2012). ATG (ImtixSangstat, Lyon, France) was administered at 2.5 mg/kg/day from days -3 to -1.

T cell and NK cell subset reconstitution was analyzed using flow cytometry at +1, +2, and +3 months after allo-HSCT in the PSM-matched population to evaluate the overall immune reconstitution process. Besides, functional analysis of cellular immunity was performed to assess the potential impact of LTV prophylaxis on immune function.

Peripheral blood mononuclear cells were isolated from fresh anticoagulated blood via Ficoll density gradient centrifugation and assayed via a FACS CANTO II flow cytometer (BD Biosciences). T cell phenotyping was performed using directly conjugated monoclonal antibodies CD3 (PerCP), CD8 (FITC-A), and CD4 (FITC-A). For NK cell surface staining to identify cell subsets, CD3 (PerCP), CD56 (PE-A), and CD16 (FITC) were used. Concurrently with T cell and NK cell phenotyping, monoclonal antibodies PD-1 (BV421), TIM-3 (PE-Cy-7-A), and CTLA-4 (APC) were used to detect T cell and NK cell exhaustion markers.

T cell and NK cell functional assays involved non-specific stimulation with phorbol myristate acetate (PMA, 50 ng/mL) and ionomycin (Ino, 1 μg/mL) to induce the secretion of intracellular cytokines. After permeabilization, intracellular cytokines interferon-γ (IFN-γ, PE-Cy-7-A), tumor necrosis factor-α (TNF-α, APC), granzyme B (APC), and perforin (PE-Cy-7-A) were detected. The acquired data were further analyzed using BD-FACSDiva™ software. The flow cytometric results are presented as the percentage of positive cells ( Additional File: Figures S1, S2 ).

CMV and EBV viremia were both defined as the presence of more than 500 copies/mL in the blood twice consecutively (Lin et al., 2019). The diagnosis of EBV- and CMV- associated diseases were according to the guidelines and our previous description (Lin et al., 2019; Ljungman et al., 2002). Late-onset CMV reactivation was defined as CMV reactivation occurred after 100 days after transplantation (Rowe et al., 2013; Liu et al., 2022). Breakthrough CMV infection was defined as an infection occurring during LTV administration (Perchetti et al., 2023). Patients’ COVID-19 status before HSCT is determined by pre-transplant serological antibody testing and nucleic acid testing (Nuccetelli et al., 2020). Engraftment, relapse, non-relapse mortality (NRM), overall survival (OS), and disease-free survival (DFS) were assessed as previously described (Zeng et al., 2021; Hu et al., 2024). GVHD-free and relapse-free survival (GRFS) was defined as survival without the following events: grade III-IV acute GVHD (aGVHD), severe chronic GVHD (cGVHD), disease relapse, or death from any cause after haplo-HSCT (Li et al., 2023). Organ scoring and global assessment of cGVHD were performed according to the 2014 National Institutes of Health consensus criteria (Filipovich et al., 2005), while aGVHD was evaluated based on the criteria established by the Mount Sinai Acute GVHD International Consortium (Schoemans et al., 2018).

PSM for baseline variables (recipient age and sex, disease, graft type, conditioning regimen, human leukocyte antigen, pretransplant remission status, donor age and sex, CMV/EBV serostatus, mononuclear cells graft) used Logistic regression with the nearest-neighbor method. Caliper settings were set to less than 0.2 to restrict the distance between matched units. Matching ratio was 1:2 for each LTV-non-LTV cohort. Postmatching balance was evaluated with the standardized mean difference, and the optimal balance was considered as <0.2. The chi-square test and Fisher’s exact test were used for categorical or hierarchical features, and Wilcoxon rank-sum test was used to compare continuous variables. Relapse, non-relapse mortality and viral infections were evaluated using the Fine-Gray method (Austin and Fine, 2017) (package cmprsk of R), taking into account cumulative incidence. Correlated risks were estimated with Competing risk model (package FGR of R). Competing events were defined as follows: for relapse, death without relapse; for non-relapse mortality, relapse/progression; for viral infection, death without viral infection. The Kaplan-Meier survival curve and Log-rank test are used for OS, DFS and GRFS. A Cox proportional hazards model was used to evaluate the associations of patient and transplant characteristics with outcomes in a multivariate analysis. The level of statistical significance was set at P < 0.05. Statistical analyses were performed using R version 4.3.3 and SPSS 26.0.

A total of 462 patients were screened in this study including 106 in the LTV group and 356 in the non-LTV group. The clinical characteristics are shown in Table 1 . Among the overall cohort, the patient and transplant characteristics were comparable between the LTV group and the non-LTV group, except for differences in graft type (P = 0.016). After PSM, 212 of the 356 patients in the non-LTV group were randomized. The patient characteristics were well-balanced with PSM, and there were no significant differences between the groups ( Table 1 ; all P > 0.05, SMD < 0.2). In the LTV group, LTV was started at a median of 13 days (range, 9-24) after haplo-HSCT and was administered for a median duration of 84 days (range, 58-103) post-HSCT.

The median follow-up was 761.5 days (range, 1–1807) after transplantation. A total of 133 recipients (41.8%) experienced CMV viremia at a median of 40 days (range, 6–366) following haplo-HSCT. The 100-day cumulative incidence of CMV viremia was 24.5% (95% confidence interval [CI], 16.8–33.1%) in the LTV group and 45.8% (95% CI, 38.9–52.3%) in the non-LTV group (P < 0.001). The 1-year cumulative incidence of CMV viremia was significantly lower in the LTV group than in the non-LTV group [32.1% (95% CI, 23.4–41.1%) vs 46.2% (95% CI, 39.4–52.8%); P= 0.009] ( Figure 1A ). Sixteen patients experienced breakthrough CMV viremia at a median of 29.5 days (range, 20-68) after the initiation of LTV prophylaxis. Following first-line treatment, 12 patients (75.0%) achieved CMV-DNA negativity. The remaining 4 patients (25.0%) had persistent CMV infection but successfully cleared the virus after second-line therapy. No patients developed CMV disease or died with sustained CMV-DNA positivity. The 1-year cumulative incidences of late-onset CMV viremia after HSCT in the LTV and non-LTV groups were 8.5% (95% CI, 4.2-14.9%) and 0.5% (95% CI, 0-2.4%), respectively (P <0.001).

During the follow-up period, one patient in the LTV group developed CMV pneumonia at 115 days after haplo-HSCT. In the non-LTV group, eight patients developed CMV-associated diseases including 4 with enteritis, 3 with pneumonia, and 1 with encephalitis ( Figure 1B ). The median time to onset of CMV-associated diseases in the non-LTV group was 52 days (range, 41–268) after haplo-HSCT. The 1-year cumulative incidence of CMV-associated diseases was 0.9% (95% CI: 0.1–4.7%) and 3.8% (95% CI: 1.8–7.0%) in the LTV and non-LTV groups, respectively (P = 0.150). The detailed characteristics of CMV infection are summarized in Table 2 .

A total of 70 recipients (22.0%) experienced EBV viremia at a median of 51.5 days (range, 18–275) following haplo-HSCT. The 100-day cumulative incidence of EBV viremia was 34.0% (95% CI, 25.1–43.0%) in the LTV group and 11.3% (95% CI, 7.5–16.0%) in the non-LTV group (P < 0.001). The 1-year cumulative incidence of EBV viremia was significantly higher in the LTV group than in the non-LTV group [38.7% (95%CI, 29.4-47.9%) vs. 13.7% (95%CI, 9.5-18.7%), P < 0.001] ( Figure 1C ). In the LTV group, 16 patients with EBV viremia received rituximab preemptive therapy and 1 received EBV-CTL therapy. In the non-LTV group, 15 patients with EBV viremia received rituximab pre-emptive therapy and 6 received EBV-CTL therapy. The overall response rates to therapy were 86.7% and 72.2% for patients with EBV viremia in the LTV and non-LTV groups, respectively (P = 0.413).

During the follow-up period, two patients in the LTV group developed EBV-associated diseases including 1 hemophagocytic lymphohistiocytosis (HLH) and 1 with encephalitis while 8 in the non-LTV group developed EBV-associated diseases including 3 posttransplant lymphoproliferative disorder (PTLD), 3 with enteritis, 1 with encephalitis and 1 with pneumonia. The median time to onset of EBV-associated diseases was 66 days (range, 65-67) in the LTV group and 91 days (range, 33–357) in the non-LTV group. The 1-year cumulative incidence of EBV-associated diseases was 1.9% (95% CI: 0.4–6.1%) and 3.3% (95% CI: 1.5–6.4%) in the LTV and non-LTV groups, respectively (P = 0.370) ( Figure 1D ). Among the 2 patients who developed EBV- associated disease in LTV group, both received rituximab and EBV-CTL therapy but ultimately died due to disease severity. In contrast, although 8 patients experienced EBV- associated disease in non-LTV group, outcomes were comparatively better—only 1 patient with EBV pneumonia died from sepsis, while the remaining cases were controlled or cured. The detailed characteristics of the EBV infection are summarized in Table 2 .

Univariate and multivariate analyses of the risk factors for EBV and CMV infections post-transplantation are shown in Table 3 . On multivariate analysis, LTV prophylaxis was a protective factor for CMV viremia (HR = 0.54, 95%CI, 0.33-0.88, P = 0.014) but a risk factor for EBV viremia (HR = 2.69, 95%CI, 1.56–4.64, P<0.001). The female donor served as a protective factor for EBV viremia. Grade III-IV aGVHD was identified as a risk factor for CMV viremia, CMV-associated diseases and EBV-associated diseases. Patient sex and age, donor age, conditioning regimen, graft type, HLA typing, patients’ COVID-19 status pre-HSCT and EBV/CMV serostatus did not show any significant influence on the risk of EBV and CMV infections.

To further examine the influence of LTV prophylaxis on EBV viremia post-transplantation, we performed subgroup analyses stratified by patient sex, patient age, donor sex, donor age, condition regimen, graft type, HLA typing, aGVHD before EBV viremia, patients COVID-19 status pre-HSCT and EBV serostatus. The 1-year incidence of EBV viremia was significantly higher in the LTV cohort than in the non-LTV cohort across all subgroups except for patients with Grade III-IV aGVHD before EBV viremia ( Additional File: Figure S3 ).

There were no significant differences in the cumulative incidence of grades I to IV aGVHD between the LTV group and the non-LTV group within 100 days [42.5% (95%CI, 32.9-51.7%) vs. 41.5% (95%CI, 34.8-48.1%), P = 0.780]. The cumulative incidence of grades II to IV acute GVHD for patients with LTV and patients without LTV were 22.8% (95%CI, 15.3-31.2%) and 19.4% (95%CI, 14.4-25.0%), respectively (P = 0.450). No significant differences in grades III and IV acute GVHD were detected between LTV group and non-LTV group [5.7% (95%CI, 2.3-11.2%) vs. 6.6% (95%CI, 4.0-10.5%), P = 0.738]. The cumulative incidence of moderate to severe cGVHD by 1-year post-HSCT was similar between patients with and without LTV, with rates of 17.9% (95%CI, 11.3-25.8%) and 16.5% (95%CI, 11.9-21.8%), respectively (P = 0.730).

During the follow-up period, 272 patients survived and 49 died, of whom 15 were in the LTV group and 34 were in the non-LTV group. The causes of death are presented in Additional File: Figure S4 . The 1-year incidence of NRM was 9.4% (95% CI, 4.8–15.9%) in the LTV group and 12.3% (95% CI, 8.3–17.2%) in the non-LTV group (P = 0.410, Figure 2A ). The 1-year incidence of relapse was 13.2% (95% CI, 7.6–20.4%) in the LTV group and 6.1% (95% CI, 3.4–9.9%) in the non-LTV group (P = 0.032, Figure 2B ). The 1-year incidence of OS was 85.9% (95% CI, 78.0–91.2%) and 85.4% (95% CI, 80.0–89.5%), DFS was 77.4% (95% CI, 68.5–84.3%) and 82.1% (95% CI, 76.4–86.7%), and GRFS was 74.5% (95% CI, 65.5–81.9%) and 76.9% (95% CI, 70.8–82.1%), respectively, in the LTV and non-LTV groups (OS: P = 0.826, Figure 2C ; DFS: P = 0.439, Figure 2D ; GRFS: P = 0.759, Figure 2E ). In multivariate analysis, LTV prophylaxis was the independent risk factor for relapse (HR, 2.56; 95% CI, 1.13-5.80 P = 0.024). EBV viremia was an independent risk factor for NRM, while Grade III–IV aGVHD was an independent risk factor for NRM, OS, and DFS ( Table 4 ; Supplementary Table S1 ).

Immune reconstitution was analyzed in 272 patients who had continuous and complete immune reconstitution data including 95 in the LTV group and 177 in the non-LTV group ( Figure 3 and Additional File: Table S2 ). At 1 and 2 months post-transplant, the median percentage of CD8 T cells was significantly lower in the LTV group compared with the non-LTV group (both P < 0.001, Figure 3A ). No significant difference was observed in the median percentage of CD4 T cells and NK cells between the two groups at 1, 2, 3 months post-transplant. ( Figure 3A ; C). At 2 month post-transplant, the median counts of lymphocytes were significantly lower in the LTV group compared to the non-LTV group (P < 0.001, Figure 3B ). At 1 and 2 months post-transplant, the median counts of CD8 T cells were significantly lower in the LTV group compared to the non-LTV group (both P < 0.001, Figure 3B ). At 2 month post-transplant, the median counts of CD4 T cells were significantly lower in the LTV group compared to the non-LTV group (P < 0.001, Figure 3B ).

Further analysis of T cell function was performed in 29 patients including 17 LTV group and 12 in non-LTV group. The results revealed that, at 2 months post-transplant, the Granzyme B (P = 0.014; Figure 3E ) and perforin (P = 0.040; Figure 3E ] expression in CD8 T cells in the LTV group was significantly lower than in the non-LTV group. The expression of PD-1, TIM-3, and CTLA-4 on CD8 and CD4 T cell subsets were similar between the two groups at 1, 2, 3 months post-transplant ( Figure 3D ).

Among patients with immune reconstitution cell count analysis, 51 (18.75%) developed EBV reactivation (EBV+ group); while among patients with immune function assessment, 16 (55.17%) experienced reactivation. Comparative analysis revealed that 1 month post-transplant, the EBV+ group exhibited significantly lower absolute counts of lymphocytes (P <0.001)), CD8 T cells (P <0.001), and CD4 T cells (P = 0.037) compared to the EBV- group, with CD8 T cell counts remaining significantly lower at 2 month post-transplant (P <0.001). Regarding immune function, the EBV+ group showed markedly reduced expression of granzyme B (P = 0.013) and perforin (P = 0.028) by CD8⁺ T cells at month 1 post-transplant relative to the EBV- group ( Supplementary Figure S5 ).