Abstract
Background
People with HIV (PWH) are at increased risk for non-AIDS comorbidities despite suppressive ART. Chronic co-infections with CMV and EBV may contribute to systemic immune activation and comorbidities, but the relative contribution of viral activity vs host immune response is unclear.
Methods
We evaluated associations between viral measures and the Veterans Aging Cohort Study (VACS) Index 2.0, a validated 5-year mortality risk score for PWH. Plasma CMV and EBV IgG were quantified by ELISA, and CMV, EBV, and HIV cell-associated DNA (CA-DNA) were measured in peripheral blood mononuclear cells by droplet digital PCR. Total globulin levels were abstracted from clinical panels as a marker of generalized immune activation. Regression models were adjusted for sex, race, ethnicity, HIV duration, and history of AIDS.
Results
Among 485 ART-suppressed PWH (mean age 54 years; 17% women; 59% White), 96.5% were CMV seropositive, 100% EBV seropositive, CMV CA-DNA was detected in 45.9%, EBV in 95.4%, and HIV in 99.0%. Higher VACS scores were associated with higher CMV IgG, EBV IgG, EBV CA-DNA, HIV CA-DNA, and total globulin in adjusted models. In multivariable models, CMV IgG, EBV CA-DNA, and total globulins remained independently associated with VACS Index 2.0. HIV CA-DNA and sex were retained in the best-fit model with trend-level significance.
Conclusions
These findings suggest distinct mechanisms by which CMV, EBV, and HIV may contribute to morbidity in PWH on ART: CMV through immune activation, and EBV and HIV through viral persistence. Longitudinal studies are needed to clarify causal pathways and guide interventions.
Keywords: HIV, Cytomegalovirus (CMV), Epstein-Barr virus (EBV), latent herpesviruses, Veterans Aging Cohort Study (VACS) Mortality Risk Index
In people with HIV on virally suppressive ART, higher CMV IgG and EBV cell-associated DNA levels were associated with increased VACS Index 2.0 mortality risk scores, suggesting distinct mechanisms by which chronic herpesvirus co-infections may contribute to morbidity and mortality risk.
(See the Editorial Commentary by Lichtner and Vita on pages 323–5.)
Despite effective ART, people with HIV (PWH) remain at increased risk of non-AIDS morbidity and mortality compared to people without HIV [1–3]. Persistent immune activation and systemic inflammation are key contributors to this excess risk [3]. Chronic co-infections with herpesviruses, particularly CMV and EBV, are nearly universal in PWH and represent potentially modifiable drivers of immune dysregulation [4, 5].
CMV establishes lifelong infection and undergoes frequent subclinical reactivations, resulting in persistent antigenic stimulation and broad immune system engagement [6]. CMV encodes viral cytokines and cytokine receptor homologs that modulate host signaling pathways, producing mixed pro- and anti-inflammatory effects [7–15]. CMV also promotes expansion of highly differentiated, senescent CD8+ T cells (“memory inflation”), reducing immunologic diversity and contributing to immune aging [16]. Additionally, CMV has been linked to endothelial dysfunction and increased cardiovascular disease risk, even in immunocompetent individuals [17–19]. In PWH on suppressive ART, CMV-specific immune responses and subclinical viral activity, independent of overt viremia or end-organ disease, have been associated with inflammation-related comorbidities and non-AIDS mortality [20–24].
EBV, another ubiquitous herpesvirus, establishes latency in memory B cells and persists through intermittent reactivation, B-cell clonal expansion, and immortalization [25–27]. EBV-driven B-cell activation and dysregulation can sustain systemic immune perturbation. In PWH, EBV co-infection has been associated with increased HIV reservoir size and diversity, likely via chronic immune activation and impaired immune surveillance [13, 24]. Compared with CMV, fewer studies have evaluated EBV's contribution to non-AIDS comorbidities.
A critical challenge is disentangling whether adverse health outcomes are driven by direct viral activity, host immune responses (general or virus specific), or their interactions. While higher IgG levels are often assumed to reflect more frequent reactivation, several studies show no associations between CMV IgG and DNA levels or reactivation events [4, 28–31]. A prior analysis of men with HIV even found that higher CMV IgG levels were associated with less frequent detection of CMV in seminal fluid and blood [13]. Other studies have found stronger associations between CMV-specific immune responses (CD8+ T cells and total IgG) and adverse outcomes than with measures of viral replication, suggesting that immune activation may be the key mediator [30, 32, 33]. This relationship is even less clear for EBV. Clarifying whether viral burden or host responses are more clinically relevant is critical for informing clinical trial design and therapeutic strategies.
Most prior studies have focused on a single virus or immunologic parameter, limiting insights into the combined and virus-specific effects of chronic co-infections. Few studies have incorporated general markers of immune activation to contextualize virus-specific findings. Total globulin levels (or “gamma gap”) represent a nonspecific but readily available marker of polyclonal immune activation and have been associated with increased morbidity and mortality risk across multiple chronic disease settings [34–38]. We included total globulin levels to help distinguish general immune activation from virus-specific effects.
To assess the clinical significance of chronic viral burden and immunologic markers, we leveraged the Veterans Aging Cohort Study (VACS) Mortality Index 2.0, a validated composite risk score incorporating HIV-specific parameters alongside general indicators of physiologic dysfunction [39]. The VACS Index 2.0 provides a robust surrogate for global morbidity and mortality risk in PWH, making it a valuable tool for evaluating links between chronic viral burden, immune dysregulation, and long-term health outcomes [40–42].
In this study, we examined the relationships between VACS Index 2.0 scores and quantitative measures of CMV, EBV, and HIV-1, including CA-DNA and IgG titers, in a large cohort of ART-suppressed PWH. We hypothesized that stronger immune responses (IgG) and higher levels of viral DNA would be independently associated with increased VACS scores, reflecting greater risk of non-AIDS comorbidity and mortality.
METHODS
Study Design and Participants
This was a retrospective cross-sectional analysis incorporating additional viral and serological testing on stored biospecimens from 2 large observational cohort studies of PWH in the United States: CHARTER (CNS HIV ART Effects Research) and HNRP (HIV Neurobehavioral Research Program) [43, 44]. A total of 485 participants were included based on the following criteria: (1) availability of stored plasma specimens collected during a comprehensive clinical research assessment between 2003 and 2019; (2) documented use of ART at the time of specimen collection, and (3) plasma HIV-1 RNA <200 copies/mL.
HNRP assessments [2] were completed at the University of California San Diego, and CHARTER visits were completed at 1 of 6 US sites (Baltimore, Galveston, New York City, San Diego, Seattle, and St Louis) [45]. Original cohort enrollment excluded participants with active opportunistic infections, severe psychiatric disorders (eg, untreated schizophrenia), substance use that would interfere with participation, or inability to complete assessments [43]. All projects were approved by the Human Research Protections Program at the University of California San Diego, and all participants provided written informed consent for the study procedures, including future use of their data and stored biospecimens for retrospective analyses.
Laboratory Assessments
All participants had a confirmed HIV-1 diagnosis by ELISA. HIV-1 RNA levels in plasma were quantified by a commercial PCR assay; clinical viral suppression was defined as <200 copies/mL as per United States Department of Health and Human Services guidelines. Peripheral blood T-cell subsets were measured by flow cytometry. Routine clinical laboratory tests (eg, metabolic panels and T-cell subsets) were performed by a Clinical Laboratory Improvement Amendments–certified laboratory. Total globulin levels are calculated from routine metabolic panels as total protein − albumin (reported in g/dL).
Viral Cell–Associated DNA and IgG Antibody Measurements
Cell-associated (ca-) DNA for CMV, EBV, and HIV-1 was extracted from 5 million peripheral blood mononuclear cells (PBMC)s using the AllPrep DNA/RNA minikit (Qiagen, Carlsbad, CA). Viral DNA was quantified by droplet digital PCR (ddPCR) as previously described [13, 29, 46]. Copy numbers were normalized to cell input using the RPP30 reference gene.
Quantitative CMV and EBV IgG levels were measured in plasma by commercially available enzyme immunoassay kits: Epstein–Barr Virus VCA IgG ELISA Kit # KA1444 (Abnova, Taoyuan City, Taiwan) and CMV IgG ELISA Kit #GWB-892399 (Genway Biotech, San Diego, CA). Antibody concentrations were calculated by interpolation from a standard curve, as per manufacturer protocol.
The VACS Index 2.0
The VACS Index, originally developed to predict all-cause mortality in PWH, has since been updated (VACS Index 2.0) to improve prognostic accuracy, particularly in older individuals with suppressed HIV viral loads [39, 42, 47, 48]. VACS Index 2.0 incorporates both HIV-specific variables and markers of systemic health, including age, BMI, CD4+ T-cell counts, HIV RNA viral load, hemoglobin, platelets, white blood cell count, albumin, aspartate and alanine transaminases (AST, ALT), creatinine, and hepatitis C virus serostatus. Composite indices within VACS Index 2.0 include the FIB-4 index calculated as [FIB 4: (age AST)/(platelet count ALT1/2)] to estimate liver fibrosis risk, and eGFR, calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI 2009) equation.
Statistical Analyses
Continuous variables were summarized using medians and interquartile ranges, and categorical variables as counts and percentages. CA-DNA and EBV IgG values were log10 transformed to reduce skewness and improve model assumptions. Relationships between pairs of continuous viral, immunologic and clinical variables were assessed using Spearman correlation matrices.
Associations between individual viral or immunologic markers and VACS Index 2.0 scores were first evaluated separately using unadjusted and adjusted linear regression models; adjusted models forced inclusion of prespecified covariates (sex assigned at birth, race, ethnicity, estimated duration of HIV infection [EDI], and prior AIDS diagnosis [49]). Then, to assess combined and relative contributions, additional multivariable linear regression was conducted in 2 stages: Model 1: the 6 virologic and serologic predictors of interest simultaneously and Model 2: Backward selection based on Akaike Information Criterion (AIC) was used to derive a parsimonious best-fit model with the 6 predictors of interest and the 5 prespecified covariates (sex, race, ethnicity, EDI, and prior AIDS diagnosis) as candidates. Variance inflation factors (VIF) were examined to assess multicollinearity. To evaluate potential confounding of the CMV IgG association by age (excluded from primary models as age is a component of the VACS Index 2.0) sensitivity analysis of the VACS Index 2.0 association with CMV IgG was performed by adding age alone and age plus total globulins. False discovery rate correction was applied to adjust for multiple comparisons. A 2-sided significance level of α = .05 was used. Post hoc power calculations based on the final sample size (n = 485) confirmed that the study was well powered (≥87%) to detect small-to-moderate effect sizes in the primary univariable regression analyses. Analyses were conducted using R software (version 4.5.1) and JMP Pro (version 17.2.0).
RESULTS
Participant Characteristics
This cross-sectional analysis included 485 virally suppressed individuals with HIV-1. Participants were predominantly White (59%) men (83%) with a median age of 54 years and had HIV for a median of 20.3 years (Table 1). The median nadir CD4+ T-cell count was 159 cells/µL, and the median CD4+ T-cell count at the time of the assessment was 616 cells/µL. Thirty percent were HCV seropositive. All participants were on virally suppressive ART with HIV-1 RNA <200 copies/mL.
Table 1.
Demographics, HIV Disease Characteristics and CMV, EBV, HIV Measures.
| N (%) or Median [IQR] | % Missing | |
|---|---|---|
| Demographic characteristics | ||
| Age (y) | 54 [47, 60] | 0.0 |
| Sex (female) | 84 (17.3) | 0.0 |
| Gender (woman) | 87 (17.9) | 0.0 |
| Race | … | 1.2 |
| White | 283 (59.1) | … |
| Black | 144 (30.1) | … |
| Multiracial | 24 (5.0) | … |
| American Indian | 11 (2.3) | … |
| Does not identify | 11 (2.3) | … |
| Asian/PI | 6 (1.3) | … |
| Ethnicity (Hispanic) | 89 (18.4) | 0.4 |
| HIV and clinical characteristics | ||
| CD4+ T-cell count (cells/μL) | 616 [431, 825] | 3.5 |
| CD4+ T-cell Nadir (cells/μL) | 159 [36, 300] | 2.3 |
| CD4+:CD8+ ratio | 0.77 [0.5, 1.2] | 12.2 |
| AIDS diagnosis (ever) | 320 (66.1) | 0.2 |
| Estimated duration of HIV (y) | 20.3 [13.3, 25.8] | 1.0 |
| Duration of all ART (m) | 165.6 [93.6, 231.1] | 0.0 |
| HCV seropositive | 144 (29.7) | 0.0 |
| Total globulins (globulin gap) | 2.8 [2.5, 3.3] | 0.1 |
| VACS Index 2.0 | 37 [29, 47] | 6.5 |
| Herpesvirus IgG and cell-associated viral DNA measures | ||
| CMV seropositive | 445 (97) | 4.9 |
| CMV IgG (U/mL) | 38 [29, 44] | 4.9 |
| CMV DNA detectable | 230 (47.8) | 0.8 |
| CMV DNA copies/106 cells | 0 [0.0, 5.3] | 0.8 |
| CMV DNA/106 cells (among detectable only) | 5.9 [2.8, 12.3] | N/A |
| EBV seropositive | 461 (100) | 4.9 |
| EBV IgG (U/mL) | 835 [387, 1710] | 4.9 |
| EBV DNA detectable | 460 (95.6) | 0.8 |
| EBV DNA copies/106 cells | 102 [27, 605] | 0.8 |
| HIV DNA detectable | 481 (99.2) | 0.0 |
| HIV DNA copies/106 cells | 186.3 [82.2, 342.8] | 0.0 |
All participants reported ART use and had plasma HIV RNA ≤200 copies/mL. Values reported are either number (%) or median [IQR]. Total N = 485.
The median total globulin level was 2.8 g/dL. CMV seropositivity was 96.5% with a median CMV IgG level of 38 U/mL, and EBV seropositivity was 100% with a median IgG level of 835 U/mL. For CA-DNA obtained by ddPCR in PBMCs, CMV CA-DNA was detected in 45.9%, EBV CA-DNA in 95.4%, and HIV CA-DNA in 99.0% of participants. The median VACS Index 2.0 score was 37, corresponding to an estimated 5-year mortality risk of 5.3%.
Relationships Among Viral Measures, Immune Markers, and Clinical Characteristics
To characterize relationships among continuous viral, immunologic, and clinical variables, we generated a Spearman correlation matrix (Figure 1). Most correlations were statistically significant but modest (ρ < 0.3). EBV IgG showed the strongest correlations, including positive correlations with EBV CA-DNA (ρ = 0.6, P < .0001), total globulins (ρ = 0.3, P < .001) and CMV IgG (ρ = 0.24, P < .001), and an inverse correlation with the CD4:CD8 ratio (ρ = −0.3, P < .001).
Figure 1.
Spearman correlation matrix between total globulins, IgG (EBV and CMV) titers, cell-associated viral DNA (EBV, CMV, and HIV) levels, and clinical characteristics. P value: *<.05; **<.01; ***<.001.
CMV IgG was also inversely correlated with the CD4:CD8 ratio (ρ = −0.3, P < .001); however, in contrast to EBV, CMV IgG was not correlated with CMV CA-DNA in the full cohort (P = .90) or when restricting to participants with detectable CMV CA-DNA (P = .60). HIV CA-DNA was inversely correlated with nadir CD4+ T-cell count (ρ = −0.2, P < .0001). CMV CA DNA showed weak inverse correlations with EDI (ρ = −0.2, P < .0001).
We also evaluated whether viral and immunologic measures differed by sex, race, or ethnicity using ANOVA with Tukey post hoc testing. Total globulin levels were higher in women than men (P = .036). EBV IgG levels differed by race (overall P = .01), with the highest levels among participants identifying as Black and the lowest among those identifying as Asian/Pacific Islander. Total globulins also varied by race (overall P < .0001), driven by higher levels in Black compared with White participants. No significant differences by ethnicity were observed for CMV IgG, EBV IgG, CMV, EBV, or HIV CA-DNA, or for total globulins.
Associations Between VACS Index 2.0 and Viral and Immunologic Measures
To test the hypothesis that chronic viral burden and immune activation contribute to increased mortality and morbidity risk in PWH, we performed unadjusted and adjusted linear regression analyses using the VACS Index 2.0 score as the outcome and each viral or immunologic marker as a predictor (Figure 2). All adjusted models included prespecified covariates: sex assigned at birth, race, ethnicity, EDI, and prior AIDS diagnosis.
Figure 2.
Associations between VACS Index 2.0 and viral and immunologic measures—unadjusted and adjusted models. Forest plot of standardized beta coefficients and P values. Horizontal bars indicate 95% CIs. Each model contains a single viral or immunologic measure as predictor and VACS Index 2.0 as outcome. Adjusted models also include sex, race, ethnicity, estimated duration of infection (EDI), and prior AIDS diagnosis.
In both unadjusted and adjusted models, higher VACS Index 2.0 scores were associated with higher CMV IgG (standardized β = .18, P < .001), EBV IgG (std β = .15, P < .001), EBV CA-DNA (std β = .18, P = .001), HIV CA-DNA (std β = .11, P = .025), and total globulins (std β = .16, P < .001). CMV CA-DNA was inversely associated with VACS Index 2.0 in unadjusted analyses (std β = −.11, P = .023) but not after adjustment (P = .76), suggesting confounding by demographic or HIV-related factors.
CMV IgG showed a weak but statistically significant correlation with age (ρ = 0.16, P = .001); because age is a component of the VACS Index 2.0 and was therefore excluded from primary adjusted models, we performed sensitivity analyses to confirm the CMV IgG association was not confounded by age. CMV IgG remained independently associated with VACS Index 2.0 when age was added alone (std β = .18, P < .001; R2 = 0.257) and when age and total globulins were added simultaneously (std β = .13, P = .002; R2 = 0.339), with no meaningful multicollinearity (VIF < 1.06 for all variables).
Relative and Combined Contributions of CMV, EBV, and HIV Measures in Multivariable Models
To assess the combined and independent contributions of viral and serologic markers, we constructed multivariable linear regression models including the 6 viral and serologic predictors of interest (CMV IgG, EBV IgG, total globulins, CMV, EBV, and HIV CA-DNA; Table 2). These analyses were conducted in 2 stages. Model 1 evaluated the 6 predictors simultaneously, and Model 2 incorporated these predictors together with prespecified covariates (sex assigned at birth, race, ethnicity, EDI, and prior AIDS diagnosis) as candidate variables for backward selection to identify the best-fitting model.
Table 2.
Two-Stage Multivariable Regression Analyses.
| Model 1 R2 = 0.179, P = 1.3e−15 |
Model 2: Best-Fit Model R2 = 0.271, P = 7.4e−24 |
||||||
|---|---|---|---|---|---|---|---|
| Std β | P Value | FDR P Value |
Std β | P Value | FDR P Value |
Risk Direction |
|
| CMV IgG U/mL | .1828 | <.001 | <.001 | .1407 | .002 | .003 | Higher |
| EBV CA-DNAa | .1761 | .001 | .002 | .1321 | .003 | .005 | Higher |
| Total globulins | .1635 | <.001 | .002 | .1884 | <.001 | <.001 | Higher |
| HIV CA-DNAa | .1146 | .012 | .017 | .0648 | .142 | .142 | Higher |
| CMV CA-DNAa | −.087 | .054 | .065 | … | … | … | Lower |
| EBV IgG U/mLb | .0379 | .503 | .503 | … | … | … | ns |
| EDI (y) | N/A | … | … | .2172 | <.001 | <.001 | Longer |
| AIDS status | N/A | … | … | .3743 | <.001 | <.001 | AIDS |
| Sex | N/A | … | … | .2022 | .070 | .082 | Female |
Significant p-values (P < .05) shown in bold text.
Model 1: Includes the 6 viral and serologic predictors of interest (CMV IgG, EBV IgG, total globulins, CMV CA-DNA, EBV CA-DNA, and HIV CA-DNA) entered simultaneously into the regression model.
Model 2: Best-fit model using backward selection based on Akaike Information Criterion (AIC). Candidate covariates included the 6 viral and serologic predictors from Model 1 (CMV IgG, EBV IgG, total globulins, CMV CA-DNA, EBV CA-DNA, and HIV CA-DNA) and prespecified clinical and demographic covariates (estimated duration of infection [EDI], AIDS status, sex, race, and ethnicity).
aCA-DNA values are all log10 values of DNA copies per million cells.
bEBV IgG is log10 transformed to reduce skewness. Race and ethnicity were not significantly associated with VACS Index 2.0 and not shown.
Model 1: Viral and Serologic Predictors Only
Model 1 included the 6 primary predictors of interest (CMV IgG, EBV IgG, total globulins, CMV DNA, EBV DNA, and HIV DNA) and explained 17.9% of the variance in VACS Index 2.0 scores (R2 = 0.179, P = 1.3 × 10−15). Higher VACS Index 2.0 scores were independently associated with CMV IgG (β = .183, P < .001), EBV CA-DNA (β = .176, P = .001), total globulins (β = .164, P < .001), and HIV CA-DNA (β = .115, P = .012). CMV CA-DNA showed a nonsignificant trend toward a negative association (β = −.0870, P = .054). EBV IgG was no longer independently associated (P = .50) with VACS Index 2.0 in multivariable regression. VIF were <2 for all variables, indicating minimal multicollinearity.
Model 2: Best-Fit Model
Using backward selection based on AIC, a model including the 6 viral/serologic predictors and 5 prespecified covariates (sex, race, ethnicity, EDI, and prior AIDS diagnosis) as candidate covariates was evaluated. The best-fit model retained CMV IgG, EBV CA-DNA, total globulins, HIV CA-DNA, EDI, prior AIDS diagnosis, and sex. Race and ethnicity were not retained. This model explained 27.1% of the variance in VACS Index 2.0 scores (R2 = 0.271, P = 7.4 × 10−24). CMV IgG (β = .1407, P = .002), EBV CA-DNA (β = .1321, P = .003), and total globulins (β = .1884, P < .001) remained significantly associated with higher VACS Index 2.0 scores. HIV CA-DNA (β = .0648, P = .142) and sex (β = .2022, P = .070) did not reach statistical significance but improved overall model fit and were retained. Both EDI (β = .2172, P < .001) and prior AIDS diagnosis (β = .3743, P < .001) were strongly associated with higher VACS Index 2.0 scores.
Integrated Summary of Multivariable Models
Across all models, CMV IgG, EBV CA-DNA, and total globulin levels emerged as the most consistent and robust predictors of higher VACS Index 2.0 scores, supporting their independent contributions to morbidity risk in ART-suppressed PWH. Although HIV CA-DNA was associated with VACS scores in unadjusted analyses, this association was attenuated after adjustment. CMV CA-DNA and EBV IgG were not independently associated with VACS scores in any model. The best-fit model highlights the additive influence of virus-specific immune responses, viral persistence, and generalized immune activation on morbidity risk.
DISCUSSION
The mechanisms and relative contributions of CMV and EBV co-infections to excess morbidity and mortality risk in PWH on suppressive ART remain incompletely understood. In this study, we examined associations between the VACS Index 2.0—a validated composite mortality risk score—and quantitative CMV, EBV, and HIV markers (CA-DNA and IgG) in a well-characterized cohort of 485 ART-suppressed PWH. By adjusting multivariable models for total globulin levels, we aimed to distinguish virus-specific effects from generalized immune activation. We found that higher CMV IgG and EBV CA-DNA were independently associated with higher VACS Index 2.0 scores, whereas CMV CA-DNA and EBV IgG were not. These findings suggest that, in people with well controlled HIV on ART, CMV-associated risk is primarily associated with host immune responses, whereas EBV-associated risk reflects more direct effects of cellular viral persistence.
CMV and EBV: Distinct Mechanisms of Risk
The association between CMV IgG—but not CMV CA-DNA in PBMCs—and mortality risk suggests that host humoral responses may better capture clinical vulnerability than circulating PBMC CA-DNA, at least in PWH on stable ART. This association remained robust across adjusted models and sensitivity analyses accounting for age and total globulins, supporting a virus-related contribution beyond generalized immune activation. The low CMV CA-DNA copy numbers and the lack of association with either VACS scores or CMV IgG are consistent with PBMC-associated CMV DNA reflecting latent genomes rather than active replication in this population. This aligns with prior evidence that PBMC-based CMV DNA associations with inflammation are lost after ART suppression [31]. This is consistent with CMV reactivation occurring predominantly in tissues such as salivary glands, the genital tract, and vascular endothelium, rather than in circulating leukocytes. CMV IgG likely reflects both CMV-specific humoral immunity and some nonspecific immune dysregulation warranting cautious interpretation. As a stable, blood-based marker integrating cumulative host–virus immune engagement, it offers practical advantages over more technically demanding, invasive, or transient measures such as CMV-specific T- and NK-cell responses or mucosal viral shedding [50]. As with all cross-sectional findings, directionality cannot be established—greater comorbidity burden could drive higher CMV IgG rather than the reverse—and longitudinal studies are needed to clarify causal pathways and identify which CMV-specific immune correlates best predict clinical risk in PWH on ART.
In contrast, EBV CA-DNA in PBMCs remained significantly associated with VACS Index 2.0 scores across all models, while EBV IgG did not, suggesting that EBV-associated risk is more closely linked to cellular viral persistence than to humoral immune responses. That EBV CA-DNA was the more specific predictor—despite its strong correlation with EBV IgG—likely reflects direct effects of EBV persistence and activity in circulating B cells on immune activation and increased HIV transcription, reservoir size, and diversity [24, 29]. EBV remains less studied than CMV, and our findings support further investigation.
HIV CA-DNA
Although HIV CA-DNA was detectable in PBMCs in nearly all participants, its association with VACS Index 2.0 was modest and attenuated after adjustment. HIV CA-DNA was retained in the best-fit model with trend-level significance, likely reflecting limitations of total HIV CA-DNA as a surrogate for replication-competent reservoir [24]. Nevertheless, residual HIV DNA may still mark prior immune injury or ongoing low-level inflammation, partially explaining the weaker associations observed.
Total Globulins and Nonspecific Immune Activation
Total globulin levels were significantly associated with VACS Index 2.0 scores across all models, reinforcing their value as a marker of generalized immune activation and polyclonal B-cell stimulation [35, 38]. Elevated globulin levels have been linked to increased morbidity and mortality risk in multiple chronic disease contexts [34–38]. The retention of CMV IgG, EBV CA-DNA, and total globulins as independent predictors in multivariable models suggests that these measures capture overlapping yet distinct aspects of immune dysregulation, likely reflecting bidirectional relationships between chronic viral infections and host immune activation.
Notably, total globulin levels were higher in female and Black participants in this cohort; whether this reflects differences in immune activation, higher sex hormone binding globulin, or other factors warrants further investigation in cohorts with more balanced demographic representation.
Limitations and Future Directions
This study has several limitations. Its cross-sectional design precludes causal inference. The use of IgG titers targeting immunodominant antigens as markers of antiviral immunity underrepresents the complexity of host–virus interactions, as these measures do not capture cellular immunity, broader antibody repertoires, or functional T-, B-, or NK-cell responses. Viral quantification by ddPCR does not distinguish latent genomes from active replication for CMV and EBV, nor replication competent from defective HIV proviruses. While the VACS Index 2.0 is a validated surrogate for morbidity and mortality risk, it remains an indirect, composite measure and may not fully reflect organ-specific disease processes. Generalizability is limited, as participants were predominantly White men with long-standing HIV infection and substantial historical immune compromise; whether similar associations exist among PWH who initiated ART during early infection or groups underrepresented in our sample, including women and some racial/ethnic minorities, remains uncertain. Additionally, future studies should incorporate other herpesviruses, particularly given emerging evidence linking VZV and HSV to aging and dementia risk.
Future studies should prioritize longitudinal designs linking serial viral and immunologic measures to incident non-AIDS outcomes to clarify causal pathways. More detailed immune profiling will be necessary to distinguish host–viral interactions and to identify determinants of interindividual variability in risk. Mucosal or tissue-based assessments of CMV and EBV activity may better capture biologically relevant viral effects than PBMC-based measures alone. Ultimately, interventional studies will be needed to determine whether targeting herpesvirus-associated immune activation—through antiviral, immunomodulatory, or vaccine-based strategies—can meaningfully reduce comorbidity and mortality risk, and to identify which patients are most likely to benefit from personalized approaches.
CONCLUSIONS
CMV IgG, EBV CA-DNA, and total globulins are independently associated with mortality risk in ART-suppressed PWH, suggesting CMV and EBV contribute through distinct mechanisms: CMV primarily through host immune responses, EBV through direct viral effects. These distinct host- and virus-specific patterns underscore the complexity of herpesvirus-associated morbidity in PWH and highlight the need for longitudinal and mechanistic studies to determine whether targeting herpesvirus-driven immune dysregulation can meaningfully reduce comorbidity risk.
Notes
Acknowledgments. The authors gratefully acknowledge support from the National Institutes of Health funded CNS HIV Anti-Retroviral Therapy Effects Research (CHARTER) and the HIV Neurobehavioral Research Center (HNRC) research cohorts and biorepositories.
Data availability. The data underlying this article will be shared on reasonable request with the corresponding author.
Financial support. This study was made possible by grants from the National Institutes of Health, including R01 MH107345, R01 MH125720, P30 MH062512, and P30 AI036214. P. K. R. also received support from K12 TR001444, T32 AI007384, and CFAR P30 AI03621. S. L. L. received funding from NIH P30MH062512, R01MH125720, and R24MH129166. R. J. E. received funding from P30MH062512 and R01MH125720. The CNS HIV Anti-Retroviral Therapy Effects Research (CHARTER) is funded by the National Institutes of Mental Health (R01 MH107345); is affiliated with Johns Hopkins University; the Icahn School of Medicine at Mount Sinai; University of California, San Diego; University of Texas, Galveston; University of Washington, Seattle; Washington University, St Louis; and is headquartered at the University of California, San Diego and includes Robert K. Heaton, PhD, S. L. L., MD, R. J. E., MD, PhD, Christine Fennema-Notestine, PhD, Rebecca Theilmann, PhD, Amanda Bischoff-Grethe, Bin Tang, Clint Cushman, Debra Cookson, Matthew Dawson, and D. F., Jr; Johns Hopkins University site: Leah Rubin, PhD (P.I.); Icahn School of Medicine at Mount Sinai site: Susan Morgello, MD (P.I.), Letty Mintz, NP; University of California, University of Washington, Seattle site: Christina Marra, MD (P.I.), Payal Patel, MD; University of Texas, Galveston site: Benjamin Gelman, MD, PhD (P.I.), Eleanor Head, RN, BSN; and Washington University, St Louis site: David B. Clifford, MD (P.I.), Beau Ances, MD, PhD, and Mengesha Teshome, MD.The HIV Neurobehavioral Research Center (HNRC) is supported by center award P30 MH062512 from the National Institutes of Mental Health. The San Diego HIV Neurobehavioral Research Center (HNRC) group is affiliated with the University of California, San Diego, the Naval Hospital, San Diego, and the Veterans Affairs San Diego Healthcare System, and includes Director: David J. Moore, PhD, Co-Director: Igor Grant, MD; Associate Directors: R. J. E., MD, PhD, and S. L., MD; Center Manager: Jennifer Iudicello, PhD; D. F., Jr; Leyhda Acol; NeuroMedical Core: R. J. E., MD, PhD (Director/NeuroMedical Unit Head), S. L. L., MD (Co-I/Laboratory Unit Head), Christine Fennema-Notestine, PhD (Co-I/Neuroimaging Unit Head); Debra Rosario, MPH, Neurobehavioral & Psychiatry Core: Jennifer Iudicello, PhD (Co-Director/Neurobehavioral Unit Head), Murray B. Stein, MD (Co-Director/Psychiatry Unit Head), Andrew H. Miller, Matthew Dawson, NeuroVirology & Biology Core: Sara Gianella Weibel, MD (Co-Director/NeuroVirology Unit Head), Sarah A. LaMere, DVM, PhD (Associate Unit Head, NeuroVirology Unit), Adam Fields, PhD (Co-Director/Neurobiology Unit Head); Microbiome Core: Rob Knight, PhD (Co-Director), Pieter Dorrestein, PhD (Co-Director); Developmental Core: S. L. L, MD (Director), Ajay Bharti, MD (Co-I), J. Allen McCutchan, MD, Christine Fennema-Notestine, PhD; Administrative Core: Robert K. Heaton, PhD (Director/Coordinating Unit Head), Participant Accrual and Retention Unit: J. Hampton Atkinson, MD (Unit Head), Jennifer Marquie-Beck, MPH; Data Management and Information Systems Unit: Ian Abramson, PhD (Unit Head), Clint Cushman; Statistics Unit: Florin Vaida, PhD (Unit Head), Anya Umlauf, MS, and Bin Tang, PhD.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Contributor Information
Patricia K Riggs, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
Gordon Honerkamp-Smith, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
Milenka Meneses, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
Gemma Caballero, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
Antoine Chaillon, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
Donald Franklin, Department of Psychiatry, University of California San Diego, San Diego, California, USA.
Ronald J Ellis, Department of Neurosciences, University of California San Diego, San Diego, California, USA.
Scott L Letendre, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA; Department of Psychiatry, University of California San Diego, San Diego, California, USA.
Sara Gianella, Division of Infectious Diseases and Global Public Health, Department of Medicine, University of California San Diego, San Diego, California, USA.
References
- 1. Marcus JL, Leyden WA, Alexeeff SE, et al. Comparison of overall and comorbidity-free life expectancy between insured adults with and without HIV infection, 2000-2016. JAMA Netw Open 2020; 3:e207954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Rodriguez-Penney AT, Iudicello JE, Riggs PK, et al. Co-morbidities in persons infected with HIV: increased burden with older age and negative effects on health-related quality of life. AIDS Patient Care STDS 2013; 27:5–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gabuzda D, Jamieson BD, Collman RG, et al. Pathogenesis of aging and age-related comorbidities in people with HIV: highlights from the HIV ACTION workshop. Pathog Immun 2020; 5:143–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Hunt PW, Martin JN, Sinclair E, et al. Valganciclovir reduces T cell activation in HIV-infected individuals with incomplete CD4+ T cell recovery on antiretroviral therapy. J Infect Dis 2011; 203:1474–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Vasilieva E, Gianella S, Freeman ML. Novel strategies to combat CMV-related cardiovascular disease. Pathog Immun 2020; 5:240–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Forte E, Zhang Z, Thorp EB, Hummel M. Cytomegalovirus latency and reactivation: an intricate interplay with the host immune response. Front Cell Infect Microbiol 2020; 10:130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Smith NA, Chan GC, O'Connor CM. Modulation of host cell signaling during cytomegalovirus latency and reactivation. Virol J 2021; 18:207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Lisco A, Vanpouille C, Margolis L. War and peace between microbes: HIV-1 interactions with coinfecting viruses. Cell Host Microbe 2009; 6:403–8. [DOI] [PubMed] [Google Scholar]
- 9. Chang WLW, Baumgarth N, Yu D, Barry PA. Human cytomegalovirus-encoded interleukin-10 homolog inhibits maturation of dendritic cells and alters their functionality. J Virol 2004; 78:8720–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Freeman ML, Lederman MM, Gianella S. Partners in crime: the role of CMV in immune dysregulation and clinical outcome during HIV infection. Curr HIV/AIDS Rep 2016; 13:10–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Gianella S, Letendre S. Cytomegalovirus and HIV: a dangerous pas de deux. J Infect Dis 2016; 214 Suppl 2:S67–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Gianella S, Massanella M, Wertheim JO, Smith DM. The sordid affair between human herpesvirus and human immunodeficiency virus. J Infect Dis 2015; 212:845–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Gianella S, Morris SR, Tatro E, et al. Virologic correlates of anti-CMV IgG levels in HIV-1-infected men. J Infect Dis 2014; 209:452–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Berry R, Watson GM, Jonjic S, Degli-Esposti MA, Rossjohn J. Modulation of innate and adaptive immunity by cytomegaloviruses. Nat Rev Immunol 2020; 20:113–27. [DOI] [PubMed] [Google Scholar]
- 15. Poole E, Neves TC, Oliveira MT, Sinclair J, da Silva MCC. Human cytomegalovirus interleukin 10 homologs: facing the immune system. Front Cell Infect Microbiol 2020; 10:245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Suni MA, Ghanekar SA, Houck DW, et al. CD4(+)CD8(dim) T lymphocytes exhibit enhanced cytokine expression, proliferation and cytotoxic activity in response to HCMV and HIV-1 antigens. Eur J Immunol 2001; 31:2512–20. [DOI] [PubMed] [Google Scholar]
- 17. Hsue PY, Hunt PW, Sinclair E, et al. Increased carotid intima-media thickness in HIV patients is associated with increased cytomegalovirus-specific T-cell responses. AIDS 2006; 20:2275–83. [DOI] [PubMed] [Google Scholar]
- 18. Wang H, Peng G, Bai J, et al. Cytomegalovirus infection and relative risk of cardiovascular disease (ischemic heart disease, stroke, and cardiovascular death): a meta-analysis of prospective studies up to 2016. J Am Heart Assoc 2017; 6:e005025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Garg A, Gianella S, Nakazawa M, Trout R, Spector SA. Association of cytomegalovirus DNA and immunologic markers of cardiovascular disease. Open Forum Infect Dis 2019; 6:ofz113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Hodowanec AC, Lurain NS, Krishnan S, Bosch RJ, Landay AL. Increased CMV IgG antibody titer is associated with non-AIDS events among virologically suppressed HIV-positive persons. Pathog Immun 2019; 4:66–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Patel EU, Gianella S, Newell K, et al. Elevated cytomegalovirus IgG antibody levels are associated with HIV-1 disease progression and immune activation. AIDS 2017; 31:807–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Lichtner M, Cicconi P, Vita S, et al. Cytomegalovirus coinfection is associated with an increased risk of severe non-AIDS-defining events in a large cohort of HIV-infected patients. J Infect Dis 2015; 211:178–86. [DOI] [PubMed] [Google Scholar]
- 23. Tavenier J, Margolick JB, Leng SX. T-cell immunity against cytomegalovirus in HIV infection and aging: relationships with inflammation, immune activation, and frailty. Med Microbiol Immunol 2019; 208:289–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Chaillon A, Nakazawa M, Rawlings SA, et al. Subclinical cytomegalovirus and Epstein-Barr virus shedding is associated with increasing HIV DNA molecular diversity in peripheral blood during suppressive antiretroviral therapy. J Virol 2020; 94:. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Cohen JI. Epstein-Barr virus infection. N Engl J Med 2000; 343:481–92. [DOI] [PubMed] [Google Scholar]
- 26. Babcock GJ, Decker LL, Volk M, Thorley-Lawson DA. EBV persistence in memory B cells in vivo. Immunity 1998; 9:395–404. [DOI] [PubMed] [Google Scholar]
- 27. Young LS, Rickinson AB. Epstein-Barr virus: 40 years on. Nat Rev Cancer 2004; 4:757–68. [DOI] [PubMed] [Google Scholar]
- 28. Mansfield SA, Dwivedi V, Elgharably H, et al. Cytomegalovirus immunoglobulin G titers do not predict reactivation risk in immunocompetent hosts. J Med Virol 2019; 91:836–44. [DOI] [PubMed] [Google Scholar]
- 29. Gianella S, Anderson CM, Var SR, et al. Replication of human herpesviruses is associated with higher HIV DNA levels during antiretroviral therapy started at early phases of HIV infection. J Virol 2016; 90:3944–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Lurain NS, Hanson BA, Hotton AL, Weber KM, Cohen MH, Landay AL. The association of human cytomegalovirus with biomarkers of inflammation and immune activation in HIV-1-infected women. AIDS Res Hum Retroviruses 2016; 32:134–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Gianella S, Moser C, Vitomirov A, et al. Presence of asymptomatic cytomegalovirus and Epstein–Barr virus DNA in blood of persons with HIV starting antiretroviral therapy is associated with non-AIDS clinical events. AIDS 2020; 34:849–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Margolick JB, Bream JH, Nilles TL, et al. Relationship between T-cell responses to CMV, markers of inflammation, and frailty in HIV-uninfected and HIV-infected men in the multicenter AIDS cohort study. J Infect Dis 2018; 218:249–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Gomez-Mora E, Massanella M, García E, et al. Elevated humoral response to cytomegalovirus in HIV-infected individuals with poor CD4+ T-cell immune recovery. PLoS One 2017; 12:e0184433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Avramovski PJ, Petlichkovski A, Avramovska M, et al. The gamma gap predicts all-cause mortality in chronic dialysis patients. Indian J Nephrol 2021; 31:212–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Juraschek SP, Moliterno AR, Checkley W, Miller ER, 3rd. The gamma gap and all-cause mortality. PLoS One 2015; 10:e0143494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Liu GY, Tang O, Brotman DJ, Miller ER, 3rd, Moliterno AR, Juraschek SP. Gamma gap thresholds and HIV, hepatitis C, and monoclonal gammopathy. PLoS One 2020; 15:e0224977. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Suleman A, Cameron DW, Corrales-Medina V, McCudden C, Cowan J. Evaluation of the protein gap for detection of abnormal serum gammaglobulin level: an imperfect predictor. Clin Chem Lab Med 2021; 59:869–74. [DOI] [PubMed] [Google Scholar]
- 38. Yang M, Xie L, Liu X, Hao Q, Jiang J, Dong B. The gamma gap predicts 4-year all-cause mortality among nonagenarians and centenarians. Sci Rep 2018; 8:1046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. McGinnis KA, Justice AC, Moore RD, et al. Discrimination and calibration of the veterans aging cohort study Index 2.0 for predicting mortality among people with human immunodeficiency virus in North America. Clin Infect Dis 2022; 75:297–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Marquine MJ, Montoya JL, Umlauf A, et al. The Veterans Aging Cohort Study (VACS) Index and neurocognitive change: a longitudinal study. Clin Infect Dis 2016; 63:694–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Shahani L, Breaux K, Lin M, Marcelli M, Rodriguez-Barradas MC. Veterans aging cohort study Index as a marker of bone disease in HIV-infected patients. AIDS Res Hum Retroviruses 2019; 35:1143–7. [DOI] [PubMed] [Google Scholar]
- 42. Hernández-Favela CG, Hernández-Ruiz VA, Bello-Chavolla OY, et al. Higher veterans aging cohort study 2.0 Index score predicts functional decline among older adults living with HIV. AIDS Res Hum Retroviruses 2021; 37:878–83. [DOI] [PubMed] [Google Scholar]
- 43. Heaton RK, Clifford DB, Franklin DR, Jr, et al. HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER study. Neurology 2010; 75:2087–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Morgan EE, Iudicello JE, Weber E, et al. Synergistic effects of HIV infection and older age on daily functioning. J Acquir Immune Defic Syndr 2012; 61:341–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Heaton RK, Franklin DR, Jr, Deutsch R, et al. Neurocognitive change in the era of HIV combination antiretroviral therapy: the longitudinal CHARTER study. Clin Infect Dis 2015; 60:473–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Massanella M, Gianella S, Lada SM, Richman DD, Strain MC. Quantification of total and 2-LTR (long terminal repeat) HIV DNA, HIV RNA and herpesvirus DNA in PBMCs. Bio Protoc 2015; 5:e1492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Justice AC, Modur SP, Tate JP, et al. Predictive accuracy of the veterans aging cohort study index for mortality with HIV infection: a North American cross cohort analysis. J Acquir Immune Defic Syndr 2013; 62:149–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Tate JP, Sterne JAC, Justice AC; Veterans Aging Cohort Study (VACS) and the Antiretroviral Therapy Cohort Collaboration (ART-CC) . Albumin, white blood cell count, and body mass index improve discrimination of mortality in HIV-positive individuals. AIDS 2019; 33:903–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. From the Centers for Disease Control and Prevention . 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. JAMA 1993; 269:729–30. [PubMed] [Google Scholar]
- 50. Schnittman SR, Hunt PW. Clinical consequences of asymptomatic cytomegalovirus in treated human immunodeficency virus infection. Curr Opin HIV AIDS 2021; 16:168–76. [DOI] [PMC free article] [PubMed] [Google Scholar]