Background
Epigenetic age acceleration has prognostic value in the general population. However, its role in predicting clinical outcomes in people with HIV (PWH) with long-term virological suppression remains unclear
Methods
Retrospective cohort of PWH with long-term virological suppression, followed at a single HIV unit. Blood DNA methylation was assessed using the Illumina Infinium MethylationEPIC BeadChip, and six epigenetic biomarkers of aging were evaluated (Horvath´s and Hannum´s clocks, PhenoAge, GrimAge, DNAmFitAge, and DNAmTL). EAA was calculated as the residuals resulting from the regression of epigenetic age on chronological age. After a 10-year follow-up period, serious non-AIDS events (SNAEs), and aging-related events (diabetes, hypertension, osteoporosis, cataracts, dementia) were recorded. Multivariant adjusted Cox proportional hazards regression models were employed to assess the association between EAA and clinical events
Results
120 participants, mostly male (76.6%). At the moment of blood sample collection median age 44.6 years (IQR: 45-55), median time since HIV diagnosis 16.7 years (IQR: 11.6-20.7), median duration of virological suppression 7.5 years (IQR: 7,9-6,3), CD4 count 740 cells/mm3 (IQR: 496-1000) and 77% were current or former smokers. During follow-up, virological failure occurred in 5 patients and 70 suffered an event: 23 age-related events, 20 SNAEs and 27 had both types of events. In adjusted Cox regression models, a one-year increase in GrimAge acceleration (GrimAA) and PhenoAge acceleration (PhenoAA) was associated with a 9% (HR: 1.09; 95% CI: 1.018-1.17; p=0.014) and a 6.7% increased hazard (HR: 1.07; 95% CI: 1.006-1.13; p=0.04) of SNAEs respectively. Specifically for non-AIDS-defining cancers (NADCs), having a positive PhenoAA or GrimAA was associated with an almost four-fold (HR: 3.73; 95% CI: 2.41-5.04; p=0.014) and more than six-fold increased risk (HR: 6.5; 95% CI: 4.63-8.37; p=0.003) respectively, whereas a positive GrimAA was linked to a five-fold increased risk of death (HR: 5.32; 95% CI: 3.75-6.88; p= 0.032) (Fig. 1). No significant associations were observed for any biomarker with all events combined or with age-related events alone. Horvath’s and Hannum’s clocks, DNAmFitAge and DNAmTL were not associated with the occurrence of any event
Conclusions
GrimAA has significant prognostic value for SNAEs, NADCs and mortality in people with well-controlled HIV and high CD4 cell counts, highlighting its potential as a biomarker for long-term health risks
Table or Figure
