Background
There is interest in metformin as a drug to delay aging in the general population. This is the first trial to evaluate the impact of metformin on biological aging measured by epigenetic clocks with prognostic value in PWH.
Methods
Phase II randomized, double-blind, pilot clinical trial comparing metfomin 850 mg BID vs placebo. Main inclusion criteria included age ≥ 50, HIV viral load < 50 for ≥ 1 year, CD4 > 500, HOMA-IR ≤ 2.6, BMI <30, and normal B12 vitamin levels. Key exclusion criteria included diabetes, comorbidities (stroke, heart failure, dementia, cancer or renal impairment) and pregnancy. Objectives were to assess metformin safety, changes in Epigenetic Age Acceleration (EAA) measured by Horvath, Hannum, PhenoAge, GrimAge, and DNAmfitAge epigenetic clocks in whole blood, and changes in blood telomere length measured by DNA methylation (DNAmTL) after 96 weeks of follow-up. EAA was calculated as the residuals from the regression of epigenetic age on chronological age. We compared changes between treatment groups using both linear mixed models with random intercepts and slopes and Wilcoxon signed-rank test.
Results
We included 40 participants, of whom 35 (17 metfomin and 18 placebo) maintained their randomized treatment until week 96. Discontinuations were not attributed to adverse events related to the study medication. At baseline, groups were balanced, with a mean chronological age of 57 years, 30% female, BMI of 24.1, HOMA IR of 1.39, CD4 counts (current/nadir) of 855/291 cells/µl, CD4/CD8 of 1.3, and with no differences in EAA measures and DNAmTL. Participants receiving metformin demonstrated a significant increase in serum GDF15, confirming metformin adherence. At week 96, there were no significant differences between groups in CD4 counts, and renal function, and all participants remained virologically suppressed. Median decreases in EAA were numerically larger in metformin for 3 of the 5 clocks, without statistically significant differences. Linear mixed models showed a similar evolution between groups in EAA. A trend towards a potential effect on DNAmTL was observed with Wilcoxon (p = 0.077), although this was not confirmed by linear mixed model.
Conclusions
In this pilot trial, metformin treatment for 96 weeks was safe and well-tolerated in PWH without diabetes. While no trend favoring metformin was observed across the epigenetic clocks analyzed, the trend in DNAmTL suggests a potential positive biological effect, warranting further investigation in larger studies.