51 studies found that aging clocks respond differently to drugs and lifestyle changes
51 studies found that aging clocks respond differently to drugs and lifestyle changes
On August 21, Nature Medicine published a pooled analysis of 51 longitudinal studies involving 3 128 blood samples. The authors applied the same methods to calculate 16 epigenetic clocks across these datasets. These clocks use chemical marks on DNA to estimate measures associated with age, mortality risk, or the pace of aging.
Clinical outcomes from anti-aging interventions can take years to emerge because disease and mortality develop slowly. Researchers therefore collect blood before and after an intervention and use the clocks as faster measures of change. A useful clock should detect changes and produce consistent results when measurements are repeated.
In an earlier analysis of repeat blood draws, the same group showed that repeated analysis of a single sample can be precise, while a new blood draw from the same person can produce a different result. The new study addresses the next question: which reliable measure can detect a change after a specific intervention?
The authors assembled paired samples collected before and after interventions in TranslAGE, a database of such studies. They standardized information about participants and follow-up periods, recalculated the same panel of clocks for every sample, adjusted the measures for chronological age, and compared each participant’s paired results. This approach allowed them to assess which measures changed across different interventions and participant groups.
Measures developed to estimate the pace of aging or mortality risk more often produced consistent results. DunedinPACE, a measure of the pace of aging, decreased statistically after 16 interventions and increased after one. Across the complete set of studies, PCGrimAge, PCPhenoAge, GrimAgeV2, and SystemsAge also showed statistically significant average decreases.
The type of intervention affected which measure performed best. Following lifestyle changes, DunedinPACE decreased statistically in eight of 15 studies. Following drug interventions, GrimAgeV2 decreased statistically in eight of 14. Among participants with diseases, PCPhenoAge, SystemsAge, and PCGrimAge showed larger changes, while DunedinPACE changed similarly in healthy and clinical groups.
A new trial should select several clocks in advance based on the intervention type and the participants’ health status. If several measures show the same change and an independent study of the same intervention reproduces it, that result provides stronger evidence than a single value in the final table.