The UC San Diego Semaglutide Epigenetic Aging Study, Explained

A Nature Aging publication reported a 2.4-year reduction in GrimAge2 epigenetic clock scores after 52 weeks of semaglutide. What epigenetic clocks actually measure, what the study found, and why 'reverses aging' is not an accurate characterization.

A study from the University of California San Diego published in Nature Aging in 2026 used epigenetic methylation clocks to measure biological aging in 327 patients before and after semaglutide treatment. The findings have been widely cited — and widely mischaracterized. Here's a precise reading of what the study found and what its limitations mean for interpretation.

What Epigenetic Clocks Measure

Epigenetic aging clocks measure DNA methylation patterns at specific CpG sites across the genome. These patterns change with age in predictable ways — allowing researchers to estimate a "biological age" that may differ from chronological age. The GrimAge2 clock, developed and validated by Levine et al., is specifically calibrated as a mortality predictor: it was developed to predict time-to-death in large longitudinal cohorts, and its "age acceleration" score — biological age minus chronological age — is independently associated with all-cause mortality, cardiovascular events, and cancer incidence.

Study Design

The UCSD study enrolled 327 adults with BMI ≥30 initiating semaglutide 2.4mg weekly. Blood samples for DNA methylation analysis were collected at baseline, 26 weeks, and 52 weeks. The primary outcome was change in GrimAge2 score at 52 weeks. Secondary outcomes included GrimAge clock components and additional clocks (DunedinPACE, PhenoAge).

Key Results

All three clocks showed statistically significant reductions (p<0.001 for GrimAge2 and PhenoAge at 52 weeks). The DunedinPACE result (which measures pace of aging rather than biological age) was significant but more modest.

Critical Interpretive Constraints

Epigenetic change ≠ aging reversal: The GrimAge2 clock is a predictor — it correlates with mortality in population data but has not been proven to cause aging or to be mechanistically connected to the aging process. Changing the clock reading does not necessarily change the underlying biology the clock is predicting.

Weight loss vs. drug effect: The study had no control arm that achieved equivalent weight loss through other means. The epigenetic changes may be entirely attributable to adipose mass reduction rather than any GLP-1 receptor-specific mechanism. Animal studies suggest both mechanisms may contribute.

Duration unknown: Epigenetic changes can revert. Whether the GrimAge2 reduction persists with continued treatment or returns toward baseline is not yet known. Extension data beyond 52 weeks has not been published.