The research team analyzed 941 consecutive transplant recipients treated between 2015 and 2023, a cohort split almost evenly between adults (56.5 percent) and children (43.5 percent). Most patients (76.0 percent) had underlying malignant disease, with leukemia accounting for nearly half of all diagnoses, followed by lymphoma. Roughly three-quarters of the transplants were allogeneic, meaning donor-derived, and most used myeloablative conditioning regimens. This composition reflects the clinical reality of a major cancer center, where transplantation is often deployed with curative intent against hematologic malignancies and where the resulting immunosuppression creates fertile ground for viral reactivation.

Viral detection relied on quantitative real-time polymerase chain reaction (qPCR) of whole-blood samples, with positivity thresholds of at least 153 copies per milliliter for CMV and 433 copies per milliliter for EBV. CMV testing followed a standardized weekly schedule from engraftment until day 100 after transplantation, or longer when clinically indicated. EBV testing, by contrast, was performed only when a treating physician suspected reactivation, a protocol difference that would later prove consequential for interpreting the study’s findings. Pre-transplant IgG serostatus for either virus was not routinely assessed as part of post-transplant monitoring at the institution.

The divergence between the two viruses emerged immediately. Adults accounted for 89.1 percent of all CMV tests performed and 76.6 percent of the 137 CMV infections recorded, while pediatric patients made up 59.8 percent of the 127 EBV infections. CMV testing covered 89.1 percent of the cohort, whereas EBV testing reached only 65.5 percent of patients. These disparities in both testing intensity and infection distribution suggest that the risk landscape for each virus is shaped by fundamentally different host factors, and that surveillance strategies calibrated for one virus may be poorly suited to the other.

When the researchers examined CMV infection in isolation, the risk profile pointed squarely toward older patients and specific transplant types. Patients with CMV infection were on average seven years older than those without, with a median age of 30 years in the CMV-positive group versus 20 years in the CMV-negative group. Malignant disease was also strongly associated with CMV, comprising 84.7 percent of CMV-positive cases. Notably, the broad category of allogeneic versus autologous transplantation was not itself a significant risk factor for CMV; instead, the specific donor source mattered most, with haploidentical transplants—those from partially matched family donors—accounting for over half of all CMV infections. On multivariable logistic regression, adult age (adjusted odds ratio 3.60, 95 percent confidence interval 2.15 to 6.04, p<0.0001) and haploidentical donor type (adjusted odds ratio 3.20, 95 percent confidence interval 2.08 to 4.93, p<0.0001) remained independently associated with CMV infection.

EBV told a different story. Pediatric patients comprised 59.8 percent of EBV-positive cases, and the median age of EBV-infected patients was 15 years compared to 24 years among EBV-negative patients, a highly significant difference. Non-malignant underlying disease, which represented only a quarter of the overall cohort, accounted for 35.4 percent of EBV-positive cases. The most clinically significant finding was the association between EBV infection and subsequent acute graft-versus-host disease (GVHD), a condition in which donor immune cells attack the recipient’s tissues. Patients with EBV infection developed acute GVHD at a rate of 47.9 percent compared to 37.0 percent among those without EBV infection (p=0.027). On multivariable analysis, pediatric age (adjusted odds ratio 0.51 for adult versus pediatric, 95 percent confidence interval 0.33 to 0.81, p=0.0040), non-malignant disease (adjusted odds ratio 0.60 for malignant versus non-malignant, 95 percent confidence interval 0.38 to 0.94, p=0.0268), and acute GVHD (adjusted odds ratio 1.64, 95 percent confidence interval 1.08 to 2.49, p=0.0194) were independently associated with EBV infection.

Post-transplant lymphoproliferative disorder (PTLD), the most feared EBV-related complication, was extremely rare at KHCC and caused no EBV-related deaths in this cohort. The authors suggest that most EBV viremia cases represented reactivation of latent virus rather than persistence of a pre-existing active infection, consistent with the high baseline seroprevalence of EBV in the population served by the center. This observation aligns with prior reports that pediatric recipients face elevated PTLD risk due to more frequent EBV naivety and altered immune reconstitution patterns.

Simultaneous reactivation of both viruses—co-infection—occurred in 7.6 percent of patients (70 of 916 with available testing data). Co-infection clustered within a distinct high-risk phenotype: allogeneic transplantation (p=0.004), haploidentical donor type (p=0.010), reduced-intensity conditioning (p=0.017), and acute GVHD (p=0.028) were all significantly associated with co-infection on univariate analysis. Notably, underlying malignancy status showed no association (77.1 percent of co-infected patients had malignant disease versus 75.9 percent of non-co-infected patients, p=0.819), indicating that co-infection risk derives from the intensity of iatrogenic immunosuppression and alloimmune dysregulation rather than from the underlying diagnosis itself. On multivariable logistic regression, haploidentical donor type remained independently associated with co-infection (adjusted odds ratio 2.01, 95 percent confidence interval 1.18 to 3.41, p=0.0098), while acute GVHD showed a non-significant trend (adjusted odds ratio 1.66, 95 percent confidence interval 0.98 to 2.79, p=0.058).

Survival analysis revealed a nuanced picture. Pediatric patients demonstrated markedly superior three-year overall survival at 78.4 percent (95 percent confidence interval 73.7 to 82.8 percent) compared to 64.6 percent (95 percent confidence interval 59.8 to 69.3 percent) among adults, with a median follow-up of 41.8 months among surviving patients. However, the difference in three-year overall survival between co-infected patients (67.3 percent) and non-co-infected patients (70.9 percent) did not reach statistical significance (Kaplan-Meier log-rank p=0.4046). A crude mortality comparison similarly showed no significant difference (35.7 percent versus 29.6 percent, Pearson chi-square p=0.286). In multivariable Cox regression, chronic GVHD was the only independent predictor of overall survival (adjusted hazard ratio 0.016, 95 percent confidence interval 0.002 to 0.150, p=0.0003), while age group and malignancy status showed non-significant trends.

The authors are careful to caution against overinterpreting the absence of a survival signal. The observational design, the selection bias introduced by non-uniform EBV testing, the relatively short median follow-up, and the possibility of type II error all weigh against concluding that pre-emptive strategies have eliminated the survival impact of co-infection. Recent external data from a five-year single-center cohort study suggest that isolated, sustained CMV replication, rather than transient CMV/EBV co-reactivation, may be the dominant independent predictor of late mortality after allogeneic transplantation, raising the possibility that the clinical significance of co-reactivation is time-dependent and may only become apparent with longer observation. The study period also spans the pre- and post-letermovir era, a CMV prophylactic agent that became standard of care internationally after 2017-2018 but was not used at KHCC during this time, potentially limiting comparability with more recent published cohorts. Emerging evidence further suggests that letermovir-mediated CMV suppression may be associated with a compensatory increase in EBV reactivation, particularly after haploidentical transplantation, a dynamic that merits attention in future studies.

The practical implications point toward age- and risk-adapted viral monitoring. The findings support intensified CMV surveillance for older adults and recipients of haploidentical transplants, and proactive EBV surveillance—including routine weekly qPCR—for pediatric patients, those with non-malignant diseases, and patients who develop acute GVHD. The authors also emphasize that future prospective studies should implement universal, protocol-driven EBV testing to accurately define co-infection incidence, incorporate donor-recipient serostatus data, and extend follow-up to capture the potential late-phase survival impact of co-reactivation. As haploidentical transplantation continues to expand globally as a donor option, understanding how these two herpesviruses interact within the reconstituting immune system will remain central to improving outcomes for transplant recipients of all ages.

Subject of Research: CMV and EBV co-infection epidemiology and outcomes in adult and pediatric hematopoietic stem cell transplant recipients

Article Title: Divergent CMV and EBV co-infection impact in adult and pediatric hematopoietic stem cell transplantation recipients

Article References: Alsmadi, O., Hashem, H., Halahleh, K., Jarrah, J., Talab, Y., Sultan, I., & Al-Hussaini, M. (2026). Divergent CMV and EBV co-infection impact in adult and pediatric hematopoietic stem cell transplantation recipients. Frontiers in Cellular and Infection Microbiology, 16, Article 1793798. https://doi.org/10.3389/fcimb.2026.1793798

Image Credits: AI Generated

DOI: 10.3389/fcimb.2026.1793798

Keywords: CMV, EBV, hematopoietic stem cell transplantation, co-infection, graft-versus-host disease, haploidentical transplant, pediatric oncology, qPCR surveillance, viral reactivation, post-transplant lymphoproliferative disorder, letermovir, immunosuppression