Studies that Use Epigenetic Clocks Must Obtain Other Health Data as Well
Epigenetic clocks produce a value that correlates well with chronological age, and is thought to reflect biological age, in that people with higher epigenetic ages appear to have a worse risk of age-related disease. What underlying processes produce the characteristic epigenetic changes measured by the clocks, however? Without knowing this, it is hard to take clock data seriously as an assessment of the potential for any given novel intervention to slow or reverse aging. Perhaps the clock places too much weight on one specific mechanism of aging, or is insensitive to another, which would distort the outcome for potential therapies that targeted those mechanisms. Thus as researchers add clock data to their studies, it remains important to also collect other measures of health to corroborate or dispute the observed changes in epigenetic age.
In this work, we refer to the measurement made by an aging clock as biological age (BA) given that the disparity between BA and chronological age (CA) significantly correlates with age-related health outcomes such as mortality and disease burden. Whether or not the metric provided by an aging clock truly represents BA is, however, debatable. Ultimately, these clocks make a calculation based on a set of inputs, which are typically molecular in nature and predictably vary with age in a population. In the case of epigenetic models, the methylation status (i.e., methylated or demethylated) of CpGs is utilized. If an intervention decreases the number outputted by an epigenetic clock, this means that the status of specific DNA methylation sites resembles that of a younger individual. While such a change may indicate that an individual has become biologically younger, it is feasible that a more youthful epigenetic signature can be induced irrespective of BA. One way to explore these two possibilities would be to determine if inputs used by aging clocks represent downstream biomarkers or instead causally contribute to age-related dysfunction.
Future trials using aging clocks should also take care to make traditional clinical measurements. Tests that assess functional performance in older adults - such as grip strength, gait speed, the 6-min walk test, and the timed up-and-go test - are especially valuable. In addition to estimating BA, it would be helpful to measure classical clinical biomarkers that are known to associate with lifespan and healthspan. Ultimately, the utility of BA being reduced without a concomitant functional improvement and/or a decreased risk of mortality is questionable. Conversely, a reduction in BA that is tethered to a clear enhancement in health and/or longer life is of interest. Long-term, longitudinal trials in older populations would be exceptionally valuable and offer insight into how a change in BA alters mortality-risk on an individual level. As more trials are published, we will gain a more thorough understanding of how clinically significant altering an aging clock is.