Epstein-Barr Virus Seropositivity, Immune Dysregulation, and Mortality in Pediatric Sepsis

Abstract

Importance: Epstein-Barr virus (EBV) seropositivity is associated with chronic immune dysregulation conditions, including multiple sclerosis, systemic lupus erythematosus, post-COVID-19 condition, and multiple cancers. Sepsis is an acute immune dysregulation condition attributed to 1 of 5 global deaths.

Objective: To assess causal associations among EBV seropositivity, immune dysregulation, and mortality in children with sepsis.

Design, setting, and participants: This cohort study analyzed 320 children with sepsis in the 9-center Eunice Kennedy Shriver National Institutes of Child Health and Development Collaborative Pediatric Critical Care Research Network Phenotyping Pediatric Sepsis-Induced Multiple Organ Failure (PHENOMS) study who had not previously received intravenous immune globulin. Blood samples and clinical data were collected from January 1, 2015, to December 31, 2018, and assayed from January 1, 2019, to December 31, 2022. Causal algorithms were modeled in directed acyclic graphs and subsequent sensitivity and mediation analyses applied with further confirmation by structural equation modeling. Data analysis was performed from May 2022 to January 2025.

Intervention: Blood sample collected at 24 to 48 hours of sepsis.

Main outcomes and measures: Circulating biomarkers of inflammation (C-reactive protein, ferritin, and 32 cytokines), immune depression (ex vivo tumor necrosis factor response to endotoxin < 200 pg/mL), thrombotic microangiopathy (ADAMTS13 activity <57%), and EBV seropositivity (viral capsid IgG) were measured. Causal inference analysis identified causal associations between EBV seropositivity, immune dysregulation biomarkers, macrophage activation syndrome, and death.

Results: Of the 320 children (median [IQR] age, 6 [1-12] years; 172 [53.8%] male), 150 (46.9%) were previously healthy, and 72 (22.5%) had immunocompromise at admission. A total of 172 (53.8%) had causal associations with death directly and through the mediators hyperferritinemia and macrophage activation syndrome (MAS) and also had direct causal associations with increased C-reactive protein, ferritin, and interleukin 18 binding protein, which in turn had direct causal associations with decreased ADAMTS 13 activity and decreased whole blood ex vivo tumor necrosis factor response to endotoxin. Mediation analysis found that EBV seropositivity was associated with mortality (estimate [SE], 1.86 [0.55]; P < .001). With both EBV seropositivity and ferritin included in the model, the effect of EBV seropositivity on death remained (estimate [SE], 1.52 [0.57]; P = .007), as did the ferritin effect (estimate [SE], 0.50 [0.15]; P = .001). EBV seropositivity remained significantly associated with death even after adjustment for MAS (estimate [SE], 1.78 [0.56]; P = .001).

Conclusions and relevance: In this cohort study of pediatric sepsis, EBV seropositivity was associated with immune dysregulation and mortality. Further study is warranted to address the possibility that latent EBV infection immune reprogramming poses an important public health problem that contributes to not only chronic disorders of immune dysregulation but also acute disorders of immune dysregulation, such as sepsis.

Figure 1.. Cytokine Heatmap of Groups With Epstein-Barr Virus Positive and Negative Serologic Test Results

The heatmaps in both panels show the log ratio of the median biomarker values for various markers of the host response and their hierarchical cluster associations. Red represents a greater median biomarker value for that group compared with the median for the entire study cohort, whereas blue represents a lower median biomarker value compared with the median for the entire study cohort. BP indicates binding protein; CRP, C-reactive protein; IFN, interferon; IL, interleukin; IP-10, interferon gamma-induced protein 10; MCP, monocyte chemoattractant protein; M-CSF, monocyte colony-stimulating factor; MIG, monokine induced by γ-interferon; MIP, macrophage inflammatory protein; sCD163, soluble CD163; SCF, stem cell factor; sFASLg, soluble FAS ligand; TNF, tumor necrosis factor.

Figure 2.. Abridged Causal Association Network for Epstein-Barr Virus (EBV) Seropositivity Phenotype in All 320 Patients

Orange arrows represent positive effect sizes, and blue arrows represent negative effect sizes. See eFigure 3 in Supplement 1 for the full network. BP indicates binding protein; CRP, C-reactive protein; IL, interleukin; MAS, macrophage activation syndrome; M-CSF, monocyte colony-stimulating factor; sCD163, soluble CD163; SCF, stem cell factor; sFASLg, soluble FAS ligand; TNF, tumor necrosis factor.

Figure 3.. Abridged Causal Association Network for Epstein-Barr Virus (EBV) Seropositivity Phenotype in the 218 Patients Without Prior Transfusion

Orange arrows represent positive effect sizes, and blue arrows represent negative effect sizes. See eFigure 4 in Supplement 1 for full network. BP indicates binding protein; CRP, C-reactive protein; IL, interleukin; MAS, macrophage activation syndrome; sCD163, soluble CD163; sFASLg, soluble FAS ligand; TNF, tumor necrosis factor.

Figure 4.. Outcome Curves for 28 Days in Groups With Positive and Negative Serologic Test Results

The mean numbers of organ failures and 95% CIs (shaded areas) were calculated each day by nonnested observation; the Organ Failure Index is not carried forward at the time the patient leaves the pediatric intensive care unit alive or dead.