HIV, Epigenetic Aging, and Why ART Leaves a Gut Gap

HIV appears to accelerate DNA-methylation measures of biological aging in most studies, and suppressive antiretroviral therapy (ART) slows that trajectory. It does not reliably erase it. A 2025 systematic review found accelerated epigenetic aging in 22 of 25 studies comparing people living with HIV with HIV-negative controls. The remaining gap may involve persistent intestinal barrier dysfunction and microbial translocation even after blood viral suppression.

That conclusion needs boundaries. Epigenetic clocks are biomarkers calculated from DNA methylation patterns; they are not a direct reading of a person's remaining lifespan, nor proof that changing a clock will prevent heart disease, frailty, or death. The HIV literature is also heterogeneous: it uses different clocks, tissues, populations, and methods for handling smoking, body weight, diabetes, hepatitis co-infection, and substance use. Still, the consistency of the signal makes HIV a useful human model for studying how chronic inflammation and gut-derived immune activation may affect aging biology.

Untreated HIV tracks with faster epigenetic aging

The strongest longitudinal estimate comes from Schoepf et al.'s 17-year Swiss cohort study. Researchers followed 80 people living with HIV at four time points and estimated within-person changes before and after viral suppression. During untreated infection, excess epigenetic age accumulated by 0.47 years per calendar year with the Horvath clock (95% CI, 0.37 to 0.57), 0.43 with Hannum, 0.36 with SkinBlood, and 0.69 with PhenoAge.

Adding the one calendar year everyone accumulates produces estimated biological-aging rates of about 1.47 years per year for Horvath and 1.69 for PhenoAge. Those figures describe clock trajectories in this cohort, not a literal claim that every organ aged at the same speed. The cohort was 95% white and 65% male, which limits how confidently its exact rates apply to the global HIV population.

A separate NEAT001/ANRS143 randomized-trial substudy supplies a complementary snapshot. At ART initiation, 168 ART-naive participants with HIV had higher epigenetic age acceleration than 44 HIV-negative controls: +2.5 years for Horvath (95% CI, 1.89 to 3.22; p=0.0008), +2.8 for GrimAge (95% CI, 1.97 to 3.68; p=0.0021), and +7.3 for PhenoAge (95% CI, 6.40 to 8.13; p<0.0001). Hannum acceleration was +1.4 years but did not reach conventional statistical significance (95% CI, 0.74 to 1.99; p=0.059).

The pattern also tracks immune damage. In the 2025 systematic review by Bożejko et al., studies that reported immune stratification found more acceleration with CD4 counts below 200 cells/µL and, for PhenoAge, viral loads above 100,000 copies/mL. This supports a relationship between uncontrolled infection, immune injury, and methylation-based aging. It does not isolate HIV from every correlated exposure.

ART slows the clocks, but residual differences remain

In the Swiss cohort, suppressive ART coincided with annual reductions in excess epigenetic age: −0.35 years per year for Horvath and −0.49 for PhenoAge. Expressed using the study's modeled rates, that translates to roughly 0.65 and 0.51 biological years per calendar year, respectively. These sub-chronological estimates are intriguing, but they are not proof of wholesale age reversal. The authors described the separate contributions of HIV status, ART duration, and immune variables as small in absolute magnitude. Regression toward the mean, selection of people who achieved durable suppression, and CD4 recovery could contribute.

After two years of ART in NEAT001/ANRS143, Horvath and Hannum differences versus HIV-negative controls had narrowed to non-significance. PhenoAge acceleration remained +3.69 years (95% CI, 1.77 to 5.61; p=0.0002), while GrimAge remained +2.20 years (95% CI, 0.47 to 3.99; p=0.013). Long-term follow-up in aviremic people with HIV likewise found that epigenetic-age trajectories continued to change during prolonged ART and that residual gaps persisted.

The practical inference is modest but important: early diagnosis and durable ART adherence likely limit the period of uncontrolled viral replication associated with the largest clock acceleration. ART remains the intervention with the clearest clinical benefit here because it suppresses HIV and prevents HIV-related illness; its epigenetic-clock effects are secondary biomarker findings.

The gut may help explain what suppression does not fix

Blood viral suppression does not necessarily restore the intestinal barrier. In a 2024 cross-sectional study by Singh et al., researchers studied 25 ART-suppressed people living with HIV and 23 HIV-negative controls using stool, blood, and ileum and colon biopsies. Participants with HIV showed persistent barrier dysfunction, microbial translocation, and gut microbial differences that included fewer short-chain-fatty-acid-producing bacteria and more pro-inflammatory taxa. These features correlated with accelerated biological aging in intestinal tissue and blood.

This is unusually valuable because it measured biological aging in gut biopsies rather than blood alone. It also cannot establish that dysbiosis caused the aging signal. The groups differed sharply in potentially relevant exposures: 72% of participants with HIV were men who have sex with men, versus none of the controls, and hepatitis co-infection occurred in 32% versus 4%. Sexual behavior and hepatitis co-infection may affect the intestinal microbiome independently of HIV. The study identifies a plausible pathway, not a settled causal chain.

The gut hypothesis still has practical relevance. Microbial translocation can sustain immune activation after ART has suppressed circulating virus, which offers a biologically coherent explanation for persistent methylation differences. A 2026 randomized exercise trial in 65 people with HIV found that a six-month cardiorespiratory and resistance program altered gut microbiome structure compared with a stretching control, although it did not test epigenetic aging as its endpoint. Exercise is therefore a low-risk health intervention with an adjacent microbiome signal, not a demonstrated way to reverse HIV-associated epigenetic aging.

Metformin changed monocyte clocks in a very small trial

Corley et al. tested metformin in euglycemic, virally suppressed older people living with HIV. In the 24-week randomized study, six participants received extended-release metformin, escalated from 500 mg to 1,000 mg, and five were observed without treatment. In monocytes, PCGrimAge decreased by 1.84 years (95% CI, −3.03 to −0.64; p=0.01) and PCPhenoAge by 3.53 years (95% CI, −6.06 to −0.99; p=0.01). Neither clock changed significantly in the observation group.

Those results are proof of concept, not a prescription. The effect estimates come from 11 evaluable participants, and the study did not establish whether the changes lasted, improved immune function, lowered inflammation, or changed clinical outcomes. The same paper found no significant clock changes in CD8+ T cells in an eight-week, single-arm trial of eight participants. It also found significant increases in the monocyte mitotic clocks EpiTOC (p=0.003) and EpiTOC2 (p=0.01), a finding whose biological meaning remains unclear.

Cell type matters. Monocytes turn over much faster than long-lived memory CD8+ T cells, so a 24-week methylation shift in monocytes may be easier to detect than an eight-week shift in T cells. It does not show a body-wide reversal of biological aging. Metformin has established medical uses, risks, and drug-interaction considerations; people without a cardiometabolic indication should not start it solely to improve a methylation-clock result.

Clock choice changes the story

Different clocks answer different questions. First-generation clocks such as Horvath and Hannum were trained largely to predict chronological age. PhenoAge and GrimAge incorporate methylation signatures linked more closely to mortality risk and inflammation-related biology. In the metformin study, PC-corrected versions of GrimAge and PhenoAge detected monocyte changes while first-generation clocks and DunedinPACE did not.

GrimAge also illustrates why results should not be averaged into a single number. It showed +2.8 years of acceleration at baseline in NEAT001/ANRS143, but in the Swiss longitudinal study it accumulated only 0.10 excess years per untreated year (95% CI, 0.02 to 0.19) and showed no statistically significant change during ART. Acute or chronic inflammatory signals embedded in the clock may create a large cross-sectional contrast without matching the rate seen within individuals over time. That is an unresolved measurement issue, not a reason to ignore either result.

What this means now

For people living with HIV, prompt ART initiation and sustained viral suppression remain the evidence-based way to reduce HIV-related health risk and likely reduce the period associated with the fastest epigenetic-age acceleration. A residual biomarker gap after suppression is real in several studies, especially with PhenoAge and GrimAge, but its clinical meaning remains uncertain.

The gut findings justify better research on diet, exercise, microbial metabolites, and barrier-targeted therapies, with tissue-specific and clinical endpoints. They do not justify treating probiotics, resistant starch, or microbiome tests as proven anti-aging therapies for HIV. Metformin's monocyte results deserve replication in a much larger trial with functional and clinical outcomes. Until then, the most defensible lesson is almost boring: suppress the virus early, manage cardiometabolic risk well, and treat methylation clocks as useful research instruments rather than a report card on your future.