CROI 2025: Metabolic and Other Complications of HIV Infection

Abstract

Aging-related comorbid conditions have major effects on health, quality of life, and survival in people with HIV (PWH). The 2025 Conference on Retroviruses and Opportunistic Infections (CROI) featured numerous studies about comorbid diseases in PWH. Cardiovascular diseases, including atherosclerosis and heart failure were important topics at the CROI, with ancillary analyses from REPRIEVE (the Randomized Trial to Prevent Vascular Events in HIV) and studies from lower- and middle-income countries. Numerous studies examined epigenetic markers of biologic aging in PWH and the effects of treatments, including glucagon-like peptide-1 receptor agonists. In a clinical trial, cytomegalovirus suppression was shown to decrease immune activation and systemic inflammation, as well as improve physical function. Large epidemiologic studies examining the effect of switching to integrase strand transfer inhibitors showed an increased risk of diabetes and hypertension, which was independent of weight gain. This review focuses on the abstracts presented at CROI 2025 in these areas, highlighting those with the most clinical impact.

Encouraging Findings Regarding Non-AIDS Events Among People With HIV in Spain

With effective antiretroviral therapy (ART), serious non-AIDS events (SNAEs) have eclipsed AIDS-defining events as the major source of morbidity and mortality in people with HIV (PWH). In many analyses, these events are higher in incidence than in similar populations without HIV. It is not clear how the incidence of these SNAEs have changed over time in PWH. With the aging of the population with HIV, one could hypothesize that the incidence of SNAEs would increase since the advent of effective ART. On the other hand, with the increased recognition of the importance of comorbidities in PWH among clinicians and the availability of effective treatments for various comorbidities, one could hypothesize that SNAEs would decrease over time.

Garcia and colleagues (Abstract 175) presented an analysis of the CoRIS (Cohort of the Spanish HIV/AIDS Research Network) cohort, in which the incidence of SNAEs (nonfatal myocardial infarction/nonfatal stroke/cardiovascular (CVD) death, non-AIDS malignancy, and non-AIDS, nonaccidental death) was assessed in three 6-year periods over 18 years, beginning in 2006 among 18,659 PWH. Over this period, the incidence of SNAEs decreased significantly. Compared with period 1 (2006 to 2011), SNAE incidence remained the same in period 2 (2012 to 2017; hazard ratio [HR], 0.98; 95% CI, 0.76-1.28). However, compared with period 1, incident SNAEs decreased by 32% in period 3(2018 to 2023; HR, 0.68; 95% CI, 0.50-0.92) after adjustment for age, sex, education, weight, HIV acquisition risk, geographic origin, CD4+ cell nadir, CD4+/CD8+ cell ratio, time from HIV diagnosis to ART initiation, and HIV RNA level at baseline. During this period, comorbid conditions were similar over time; however, polypharmacy increased. The underlying mechanisms for this decrease over time remain to be clarified but may suggest that the higher burden of medications to treat comorbidities (eg, statins) may be decreasing the risk of events. The contributions of temporal trends in health-related behaviors (eg, cigarette smoking) or the declines in the use of more toxic ART medications require further exploration. Also, the cohort was 80% male, so it is unclear if these findings can be generalized to women. Nevertheless, these findings suggest that over time, the healthspans of PWH are improving, along with survival.

Cardiovascular Diseases

Atherosclerosis and heart failure (HF) in PWH were major topics at CROI 2025. Although atherosclerosis in PWH in the US and Europe has been extensively studied, limited data exist in other regions of the world. Siedner and colleagues (Abstract 177) used a cross-sectional design to investigate factors related to subclinical cardiovascular disease in PWH and age- and sex-matched people without HIV living in Uganda. Coronary artery disease (CAD) was the primary outcome, and CAD included calcified and noncalcified plaque detected using coronary computed tomography angiography. The investigators enrolled 586 study participants, 49% of whom were PWH. Most of the PWH (95%) had viral suppression. The median 10-year atherosclerotic CVD (ASCVD) risk scores between the 2 groups were not significantly different. Similarly, the absolute difference in CAD prevalence was not significantly different between the 2 groups (absolute difference, 2.7%; 95% CI, -1.6 to 7.0%). Adjusting for CVD risk factors in multivariable regression models yielded a similar result, but the prevalence ratio for CAD between the groups was nonsignificant. Of note, CAD was detected in less than 10% of the study population. The authors concluded that CAD may not be a leading health problem among PWH in Uganda. A comparison can be made to studies in Europe and the US. The Swiss HIV Cohort Study (n = 704 participants)¹ demonstrated an overall prevalence of 55.4% for any coronary artery plaque. Similarly, in the Multicenter AIDS Cohort Study, the prevalence of any coronary plaque among men with and without HIV (n = 1001 participants) was 76.3%.² The study by Siedner and colleagues is an important contribution to our understanding of CAD in PWH in Uganda.

In the same theme of subclinical CAD, Lu and colleagues used data from a substudy of the REPRIEVE (Randomized Trial to Prevent Vascular Events in HIV) trial to investigate the associations of inflammatory markers and noncalcified plaque with major adverse cardiovascular events (MACE). Of the 804 participants in the substudy, 17 MACE occurred in the pitavastatin group and 21 MACE occurred in the placebo group. Noncalcified plaque was more prevalent in the participants with a MACE (62%) than in participants without a MACE (39%). Higher levels of baseline interleukin-6 (IL-6) (HR, 2.9; 95% CI, 1.2-6.9 for IL-6 levels ≥2.18 pg/mL) and high-sensitivity cardiac troponin T (hs-cTnT) (HR, 5.5; 95% CI, 2.0-15.4 for hs-cTnT levels ≥9.64 ng/L) were associated with MACE. The study authors noted that the effect of pitavastatin to reduce MACE was greater in those participants with elevated hs-cTnT levels more than 7.5 ng/L (than in those with lower hs-cTnT levels) and in those participants with noncalcified plaque (than in those without noncalcified plaque). The REPRIEVE trial was a landmark study for CVD risk reduction in PWH in its demonstration that pitavastatin reduces MACE in PWH with low to moderate risk of ASCVD. The study by Liu and colleagues increases the understanding of the mechanisms by which pitavastatin reduces ASCVD risk in PWH.

Heart Failure

With increasing evidence to support the use of sodium-glucose cotransporter-2 (SGLT-2) inhibitors and glucagon-like peptide-1 (GLP-1) receptor agonists in the treatment of HF in the general population, it is imperative to further elucidate the pathophysiology of HF in PWH, who have unique risk factors that predispose them to developing HF. Leveraging data from REPRIEVE, Watanabe and colleagues (Abstract 818) used the Predicting Risk of Cardiovascular Disease Events (PREVENT) equation to calculate 10-year risk of HF. The PREVENT equation calculates risk for CVD, including HF, in the general population and can also calculate 10-year HF risk alone. It can be used for individuals aged 30 to 79 years. Out of 7769 participants, the median 10-year PREVENT risk score for HF was 1.66%. Sixty-seven HF events were observed, which is similar to the expected number of 73 events. Factors associated with HF events included female sex, Black ethnicity in high-income countries, residency in sub-Saharan Africa, hypertension, and elevated body mass index (BMI). Of note, the PREVENT equation was developed using data from US adults; the REPRIEVE study was an international, multicenter study. Nonetheless, this study by Watanabe demonstrates that there may be utility in using the PREVENT equation, which was developed from data from the US general population without HIV, and may be potentially useful to predict HF risk in PWH.

Beydoun and colleagues (Abstract 814) characterized the spectrum of HF in PWH from 2 centers in the CNICS (Centers for AIDS Research [CFAR] Network of Integrated Clinical Systems) from 2010 to 2023. Incident HF was identified and classified by type and etiology. Of the 212 PWH with HF, 50% had HF with reduced ejection fraction and 50% had HF with preserved ejection fraction; 71% had nonischemic or mixed ischemic and nonischemic etiology of HF. Factors associated with HF included traditional cardiovascular risk factors, including older age, hypertension, diabetes, and renal insufficiency, and HIV-related risk factors, including elevated viral load and low CD4+ cell count. Understanding the types of HF and etiologies of HF that PWH are more likely to have may impact clinical care in the future, enabling clinicians to more readily recognize when PWH are at higher risk of developing HF.

Some data have demonstrated that PWH may not be receiving guideline-recommended HF treatment. The 2022 American Heart Association (AHA)/American College of Cardiology (ACC) guidelines³ outlined the use of medications from 4 different classes in the treatment of HF associated with reduced ejection fraction. Prior to the 2022 guidelines, the ACC had published guidelines in 2017⁴ for the treatment of HF. In 2023, the ACC published a consensus statement⁵ for treatment of HF with preserved ejection fraction. Cutshaw and colleagues (Abstract 812) further investigated the use of guideline-directed medical therapy (GDMT) in their study using data from the PATHWAYS (Pathways to Cardiovascular Disease Prevention and Impact of Specialty Referral in Under-Represented Racial/Ethnic Minorities With HIV) study from 2014 to 2020. They used a scoring system in which 100% of participants were defined as fully meeting GDMT. Participants who had HF at the time of or after diagnosis with HIV were included. Factors associated with a higher GDMT score included a history of diabetes, hypertension, elevated low-density lipoprotein cholesterol level, obesity, and history of myocardial infarction. However, the average GDMT score was low in the study population, at 30%. This study highlights the ongoing need to quantify the guideline-directed care that PWH receive for CVD and to understand the factors that prevent PWH from receiving optimal HF therapy.

HIV and Aging

Frailty as a Cardiovascular Disease Risk Factor

With effective ART, PWH are aging into their 60s, 70s, and beyond; however, although lifespan is lengthening, healthspan (ie, the period of comorbidity-free years) is still shorter in PWH than in individuals without HIV. In addition, evidence is mounting that aging-related phenotypes, including frailty, may be more common in PWH and may be a risk factor for negative health outcomes. In an analysis from REPRIEVE, Erlandson and colleagues (Abstract 179) examined whether frailty could predict MACE independent of age, sex, and ASCVD risk score. They examined 7769 individuals who participated in REPRIEVE and categorized them by their frailty status at baseline using a 32-item frailty index that predicted all-cause mortality. In this sample, 67% were nonfrail, 29% were prefrail, and 4% were frail.

Among these participants, 257 MACE endpoints were observed. Compared with individuals who were nonfrail at baseline, those who were prefrail had a 74% increased risk of MACE (HR, 1.76; 95% CI, 1.35-2.3), whereas the risk of MACE was more than double in those who were categorized as frail (HR, 2.14; 95% CI, 1.33-3.46). However, the effect of pitavastatin on MACE did not differ by baseline frailty status, suggesting that frailty status does not impact the benefits of statin therapy. The primary finding on the relationship between frailty and CVD outcomes suggests that the population with frailty and prefrailty should undergo more aggressive CVD risk-factor modification. Whether improvement in frailty can also improve CVD outcomes requires further examination.

Epigenetic Age

The biology underlying aging is complicated; one of the key hallmarks of aging is epigenetic changes, most commonly, DNA methylation changes. The methylation of numerous genes changes with increasing age and predicts negative health outcomes. From the DNA methylation patterns across the genome in various tissues, biologic age can be estimated and compared with chronologic age, which gives insights into factors that are associated with age acceleration. Numerous “epigenetic clocks” have been developed using these DNA methylation patterns to study aging in specific populations, factors related to accelerated aging, and interventions that can reverse this marker of accelerated aging. At CROI 2025, several studies explored epigenetic clocks in PWH.

Epigenetic Age is Accelerated in People With HIV and Predicts Aging-Related Outcomes

HIV and substance use have been associated with accelerated epigenetic age in previous studies. Sun and colleagues (Abstract 1169) examined 396 persons who inject drugs (including 127 PWH) participating in the ALIVE (AIDS Linked to the Intravenous Experience) study comprising a prospective cohort in Baltimore, Maryland. PWH exhibited more age acceleration than people without HIV, with the greatest age acceleration in those with a lower CD4+ cell count or uncontrolled HIV viremia (about 3-6 years). In addition, accelerated aging with these epigenetic clocks, particularly the PhenoAge clock, was associated with an increased risk of all-cause mortality.

Martinez-Martin and colleagues (Abstract 180) examined numerous epigenetic clocks to determine whether these clocks could predict clinical events (AIDS and non-AIDS) and aging-related diseases in a cohort of 120 PWH in Madrid (median age, 44 years; 77% male; median nadir CD4+ count, 182 cells/µL) observed over a 10-year period between 2014 and 2024. Six different epigenetic clocks were assessed: 3 first-generation clocks (Horvath, Hannu, SkinBlood) designed to predict chronologic age, and 3 second-generation clocks (PhenoAge, GrimAge, FitAge) designed to predict morbidity and mortality. Models were adjusted for sex, age, alcohol use, smoking, HIV transmission risk factors, previous AIDS diagnosis, CD4+ cell count nadir, time since diagnosis, and virologic failure. Over a 9.5-year period, 67 participants had 113 events (1 AIDS-related event, 38 SNAEs, 50 aging-related events [hypertension, osteoporosis, diabetes, cataracts, dementia], 21 with a SNAE and an aging-related event, and 16 deaths). None of the epigenetic clocks was associated with the composite endpoint. However, increased epigenetic age with the Horvath clock was associated with a 2-fold increased risk of an aging-related event (HR, 2.19; 95% CI, 1.18-4.05). When examining SNAEs, increased epigenetic age with the GrimAge and Hannum clocks showed statistically significant associations. For non-AIDS defining cancers, GrimAge, PhenoAge, and SkinBlood were statistically significantly associated, whereas for death, GrimAge showed the strongest associations (HR, 5.2; 95% CI, 1.1-24.4). Overall, it appeared that the GrimAge had the best predictive value for the most events. However, it should be noted that the cohort was quite small and the types of events were very heterogeneous. Nevertheless, despite its limitations, this study shows that these biomarkers may be useful prognosticators for negative health events in PWH.

Interventions Altering Epigenetic Age

If epigenetic age is older in PWH and predicts health outcomes, will interventions to treat various health conditions impact epigenetic age? Several studies at CROI examined this question. GLP-1 receptor agonists are in widespread use, and not only improve glucose control in persons with diabetes but lead to clinically significant decreases in weight and reductions in cardiovascular events. Two studies presented at CROI examined whether the GLP-1 receptor agonist semaglutide could alter epigenetic age in PWH.

The first study (Abstract 181) leveraged a completed randomized, placebo-controlled study of semaglutide in PWH with central lipohypertrophy⁶ and examined the effect of semaglutide on epigenetic age over a 32-week intervention period in 84 participants (45 receiving semaglutide, 39 receiving placebo). At baseline, participants showed an increase in epigenetic age of 3.74 years using first-generation clocks and 9 years using second-generation clocks; however, a third-generation clock that estimated the pace of aging was close to 1, indicating that the pace of aging was not accelerated. In the placebo group, these clocks increased, whereas reductions in epigenetic age were observed in the semaglutide group, with the difference reaching statistical significance for the Hannum clock (first generation), the PhenoAge and GrimAge clocks (second generation), and the DundinPACE clock (third generation). These results remained statistically significant after adjusting for sex, baseline BMI, and inflammatory markers (high-sensitivity C-reactive protein [hsCRP] and soluble CD163), yielding about a 3-year decrease in the second-generation epigenetic clock, and a 9% reduction in the pace of aging in the semaglutide arm compared with placebo.

In a separate analysis (Abstract 113), the effect of semaglutide on epigenetic age was examined in PWH participating in the SLIM LIVER (Open-Label Semaglutide Reduces Metabolic-Associated Steatotic Liver Disease in People With HIV) study. In this study, people with metabolic-associated liver disease (MASLD) received semaglutide for 24 weeks, which showed a 31% reduction in liver fat.⁷ In 41 participants, the FitAge epigenetic clock was assessed at baseline and 24 weeks. This clock examines aging-related DNA methylation changes and incorporates measures of physical fitness. At baseline, the FitAge clock estimated that age acceleration was 9.96 years. The analysis also examined the relationship between another epigenetic marker of handgrip and muscle function, called GripMax, to the changes in liver fat during the trial. Weaker GripMax scores at baseline were associated with greater reduction in liver fat with semaglutide. This suggests that this epigenetic marker may be useful in predicting clinical response to semaglutide in PWH with MASLD.

In another intervention study by Marcelo and colleagues (Abstract 876), PWH were randomly assigned to receive metformin 850 mg twice daily or placebo to determine whether metformin could alter epigenetic age using numerous epigenetic clocks. The metformin dose was 850 mg daily in persons receiving concomitant dolutegravir to compensate for the drug-drug interaction. Over the course of 96 weeks, changes in all epigenetic clocks were similar in the metformin group (n = 17) and placebo group (n = 18), suggesting that in this small population, metformin does not impact epigenetic aging; however, not all of the hallmarks of aging are linked to epigenetic aging.⁸

One important question with epigenetic clocks is whether these biomarkers of biologic aging also play a causal role in the aging process. It remains to be determined whether methylation of these genetic sites alters gene expression and cellular function, leading to aging-related changes at the cellular, tissue, or organism level. It is also not clear whether decreases in epigenetic age with an intervention like a GLP-1 receptor agonist actually reverse biologic aging. Nevertheless, these markers do offer some insight into biologic aging among PWH and interventions that may influence the aging process.

Cytomegalovirus Suppression and Immunologic and Functional Aging Outcomes

Interesting data were presented on letermovir, a cytomegalovirus (CMV)-directed antiviral drug, and its impact on immunologic measures and physical function (Abstract 182). One source of persistent systemic inflammation among PWH may be chronic concomitant infection with CMV. The ACTG (AIDS Clinical Trial Group) study A5383 randomly assigned PWH with undetectable HIV RNA level and CMV immunoglobulin G (IgG) antibody positivity to letermovir, a CMV terminase inhibitor, or placebo, and studied participants for 48 weeks, with the primary outcome of immune activation as estimated by the biomarker soluble tumor necrosis factor receptor 2 (sTNFR2). This study (presented at CROI 2024) was stopped after 40 of the projected 180 participants were enrolled, based on the results of a futility analysis showing an unexpected increase in sTNFR2 after 8 weeks. It was hypothesized that this effect of letermovir on sTNFR2 was due to the suppression of viral interleukin-10 (IL-10), a CMV immunomodulatory protein that decreases some pathways of inflammation, including the TNF-α pathway.

Overall, letermovir was well tolerated and reduced mucosal CMV viral shedding. After the initial increase, letermovir actually decreased sTNFR2 over 48 weeks of follow-up, with a similar pattern for other markers, such as IL-6, d-dimer, and CRP. In contrast, IL-1β showed early and sustained decreases in the letermovir arm compared with the placebo arm; the same pattern was observed with CMV-specific IgG titers. The CD4+/CD8+ cell ratio, a biomarker inversely related to immune activation, increased in the letermovir arm, with greater changes in women than in men (88% vs 15%, respectively). Along with these immunologic changes, chair rise time (ie, time to stand from a seated position 5 times) improved significantly in the letermovir arm compared with placebo and was correlated with the change in CD4 + /CD8+ cell ratio. These findings in a small group of PWH suggest that CMV treatment decreases markers of immune activation and inflammation and shows beneficial effects on a measure of physical function. Taken together, these results may indicate that chronic CMV infection plays an important role in aging-related changes in physical function among PWH. Whether letermovir can also improve cognitive function and other hallmarks of aging, including epigenetic age, requires further investigation.

Effect of Switch to Integrase Strand Transfer Inhibitors on Cardiometabolic Outcomes

In ART initiation studies and ART switch studies, integrase strand transfer inhibitor (InSTI) exposure has been associated with weight gain, although there is controversy regarding the direct impact of InSTIs relative to the effects of the comparison ART medications. There have been fewer studies that have examined the association between InSTI use and other cardiometabolic outcomes. In an analysis from the REPRIEVE study, Kileel and colleagues (Abstract 838) examined the effects of switching to an InSTI-based ART regimen on new-onset obesity, diabetes, metabolic syndrome, hypertension, and MACE. More than 5100 participants were included in at least 1 of the analyses, with 52% switching to an InSTI-containing ART regimen (82% dolutegravir). During the follow-up period (average, 3.4 years), 11% developed obesity, 5% diabetes, 9% hypertension, 19% metabolic syndrome, and 2% experienced a MACE. Individuals who switched to an InSTI had an increased hazard of developing obesity (HR, 1.34; 95% CI, 1.14-1.52), diabetes (HR, 1.41; 95% CI, 1.13-1.70), hypertension (HR, 1.44; 95% CI, 1.17-1.65), and metabolic syndrome (HR, 1.18; 95% CI, 1.04-1.33), but not MACE (HR, 1.06; 95% CI, 0.67-1.53). The authors reported that using models accounting for changes in BMI did not appreciably change the results.

A similar study from NA-ACCORD (the North American AIDS Cohort Collaboration on Research and Design) also examined the effect of switching to an InSTI-based regimen but focused on new-onset diabetes (Abstract 898). In this study, 13,071 participants receiving non-InSTI-containing regimens and without diabetes mellitus (DM) were included in the analysis and followed up for a median of 2.5 years. Switching to an InSTI was associated with a higher risk of incident DM (HR, 1.22; 95% CI, 1.03-1.45), than continuing a nonnucleoside analogue reverse transcriptase inhibitor (NNRTI)-based or protease inhibitor (PI)-based regimen. When weight gain in the first year after the switch was accounted for as a potential mediator, the effect was similar (HR, 1.18; 95% CI, 0.97-1.43). Taken together, these switch studies suggest that InSTI switching is associated with an increased risk of DM and other cardiometabolic problems, which do not appear to be driven by weight gain. These findings suggest that clinicians should have increased vigilance in monitoring for these conditions among PWH receiving InSTIs.

Incidence of Hepatic Steatosis in NA-ACCORD

MASLD is a CVD risk factor and is highly prevalent in PWH. As such, Price and colleagues (Abstract 761) investigated the incidence of hepatic steatosis within the NA-ACCORD study, which comprises 20 cohorts of PWH in North America. The study investigators also examined the relationships among hepatic steatosis, liver-related events (ie, ascites, encephalopathy, variceal bleeding, or hepatorenal syndrome) and mortality. The incident rate of hepatic steatosis increased from the periods 2000 to 2004 and 2005 to 2008, and then decreased slightly in the period 2009 to 2015. Of note, hepatic steatosis was not associated with liver-related events but was associated with mortality (adjusted HR [aHR], 1.02; 95% CI, 1.01-1.04), after adjusting for risk factors including age, dyslipidemia, and hypertension. The authors noted that future directions include determining the links among hepatic steatosis, future morbidity, and mortality.