Key points
Aging biomarkers can capture different aspects of aging: biological age measures attempt to quantify the cumulative effects of aging, whereas aging rate indicators (ARIs) aim to measure how quickly aging is occurring at a specific time.
Because biological age may take a substantial portion of the lifespan to diverge from the expected (chronological) age, validated ARIs could enable more rapid assessment of candidate aging interventions.
Establishing clinically useful ARIs will be challenging, and premature adoption carries risks, including the possibility of narrowing the field toward interventions that affect only the pathways captured by existing ARIs.
Introduction: age versus aging rate
Can the right biomarkers provide a snapshot of how fast someone is aging at a given moment? In their recent Frontiers in Science article, Austad, Kaeberlein, and Miller argue that such biomarkers, termed aging rate indicators (ARIs), probably exist, and that identifying and testing them is a critical priority for aging research (). ARIs differ from traditional biomarkers of biological age, which attempt to measure the cumulative effects of aging on the body and compare them with expected values for a given chronological age. To use an analogy, when purchasing a used car, one can check the year (chronological age) and odometer (biological age), but to know how much wear is currently being put on the car, it is more useful to check the speedometer (ARIs).
The advantages of ARIs become especially clear when considering candidate longevity interventions. Testing whether an intervention affects lifespan requires waiting a lifespan. Biological age measures improve the situation by allowing proxy measurements in living individuals, but they still require enough time for a detectable difference in biological age to emerge. In contrast, ARIs could quickly reveal whether an intervention has changed the current pace of aging—similar to how pressing the brake or accelerator rapidly changes the dial on a car’s speedometer.
Finding an ARI
While the idea of a “pure” ARI is attractive on paper, real data are messy, and candidate ARIs are likely to remain controversial for some time. Austad and colleagues propose 12 changes in tissue protein levels as candidate ARIs based on outcomes from the Interventions Testing Program and studies of genetically modified mice (). This list begins to highlight some of the ambiguities and questions that will need to be addressed.
Many candidate ARIs, while responding to interventions that promote longevity, will also be affected by factors unrelated to longevity. For example, one candidate, uncoupling protein 1 (UCP1) in adipose tissue, responds strongly to cold (), but cold exposure does not positively affect longevity in rodents (). Candidate ARIs are expected to respond to multiple, but not all, longevity interventions, raising the question of how many ARIs would need to respond for a result to be considered positive. It is also not clear whether the proposed ARIs can distinguish between male and female responses to drugs that have a sex-specific effect on longevity, such as acarbose, canagliflozin, and 17α-estradiol (). Each potential explanation for this highlights another consideration: correct predictions from ARIs could be masked by a separate, harmful effect of the same drug, or ARIs may be specific to a sex or genetic background.
Finally, there is no guarantee that a given biomarker will be purely an ARI or purely an indicator of biological age. Parameters such as blood pressure and stress hormones worsen with age on average, serving as conventional biomarkers (albeit poor ones), but may simultaneously act as ARIs. Although these examples are based on physiology, an ARI need not change with age or be directly connected to pathology; it is a statistical construct.
As a case in point, Belsky and colleagues established that an ARI-like measure, termed DunedinPACE, can be extracted from DNA methylation data obtained from a blood draw, using data collected over two decades from 1,037 participants born in 1972–1973 in Dunedin, New Zealand (). Importantly, the authors demonstrated associations between DunedinPACE values and morbidity, disability, and mortality in both the Normative Aging Study and Framingham Heart Study cohorts, whose measurements were obtained far from the location of the training data. Whether this measure will successfully predict the results of acute interventions, and whether the strategy can be applied to other types of “clock” data (i.e., mRNA, protein, and metabolite levels) remain to be seen.
Recent evidence suggests an additional caveat: all organs may not “age” at the same rate within an individual, and the most rapidly “aging” organ or organs may vary from person to person (). This raises questions not only about the sensitivity of ARIs to changes in aging rates, but also about their specificity: which organs’ aging rates are represented by a given ARI? The best individual ARIs will predict morbidity, mortality, and disability at the scale of the whole organism, but an aggregate of tissue-specific ARIs may ultimately have greater predictive value and clinical utility than a single whole-organism marker. With a bewildering and growing array of potential ARIs available, prioritization and validation will be significant challenges.
Validation of ARIs
Ultimately, an ARI would be validated by its successful use to identify a new longevity intervention, but gaining the confidence to move forward with such an experiment is a difficult challenge. Austad, Kaeberlein, and Miller suggest a subset of molecular markers based on their consistency across several models of slow-aging mice (). This approach could be extended to find markers that correlate with longevity across an increasing range of conditions. The current most promising set, or a more refined future set, could then be used to screen a variety of treatments in mice and prioritize those with strong ARI responses for testing in conventional lifespan and healthspan assays. However, whether the ARIs themselves—or the interventions they identify—will be relevant to humans will need to be tested separately.
Alternatively, starting with human data sacrifices the ability to use data from validated longevity interventions, but allows candidate ARIs to be tested directly in human datasets. DunedinPACE provides a good example of extracting a potential ARI from natural variation rather than intervention. Although it has already been replicated in independent cohorts, this measure, like any candidate ARI, should be validated with as many genetic ancestries, environmental exposures, and lifestyles represented as possible. Additionally, an ideal ARI would indicate that the “pace of aging” is much faster in patients with genetic causes of rapid aging, such as progeria, whose rates of functional decline, morbidity, and mortality are all much greater at younger ages. Although genetic diseases of rapid aging may not affect all processes involved in typical aging, they may serve as a natural experiment to help validate whether putative ARI markers do indeed predict lifespan, disease risk, and functional decline.
Finally, to be useful in the clinic, any ARI should ideally be detectable in an accessible tissue or biofluid, such as skin, saliva, or blood. The bar is high for developing a clinically useful ARI, but any candidate marker that meets these criteria could greatly benefit patients, scientists, and therapeutic development both in and beyond aging and translational geroscience.
Risks of ARIs
While the development of ARIs is an exciting prospect, their risks should not be overlooked. Using ARIs that are not sufficiently validated could waste more time than they were intended to save. In addition, the short-term nature of ARI-based experiments runs the risk of failing to identify long-term harms. For example, drugs that mildly accelerate cancer or neurodegeneration would be unlikely to raise red flags in an ARI-based screen.
Perhaps the most insidious risk, however, is siloing. If we develop ARIs based only on the pathways that we already know influence aging, we may blind ourselves to interventions that work through other mechanisms. We do not know what we do not know, and limiting biomarkers at this stage could stifle further discoveries if we assume that a lack of change in one ARI, or even a panel of ARIs, indicates no change in the rate of aging.
Therefore, we would advocate for ARIs as an important complement to, but not a replacement for, conventional studies of healthspan and longevity; any ARI-based signal should ultimately be corroborated by direct measurements of clinical benefit in populations of interest.
Conclusion
Austad, Kaeberlein, and Miller () make a compelling argument for the concept and utility of ARIs, but much work remains to be done before such markers can be used successfully to validate gerotherapeutics in the clinic. We applaud them for opening this important conversation and look forward to the essential work and discussion ahead, as the field endeavors to identify and test putative ARIs in the pursuit of our shared goal of extending healthy lifespan.
Statements
Author contributions
JPM: Writing – original draft, Conceptualization, Writing – review & editing.
JAB: Writing – original draft, Conceptualization, Supervision, Writing – review & editing.
Funding
The authors declared that financial support was received for this work and/or its publication. The authors are supported by the National Institutes of Health (NIH) grant no. R01 DK098656. The funder was not involved in the interpretation of data, the writing of this article, or the decision to submit it for publication.
Conflict of interest
The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
At the time of review, the handling editor FS declared they belong to two advisory boards, National University of Singapore; Infinita, to which the Lead Article author MK also belongs.
The author JAB declared that they were an editorial board member of Frontiers at the time of submission. This had no impact on the peer review process and the final decision.
Generative AI statement
The authors declared that no generative AI was used in the creation of this manuscript.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
AustadSNKaeberleinMMillerRA. Aging rate indicators and the search for anti-aging drugs. Front Sci (2026) 4:1821393. doi: 10.3389/fsci.2026.1821393
2
CutlerHBJall-RoggSThillainadesanSCookeKCMassonSWCSligarJMet al. Cold exposure stimulates cross-tissue metabolic rewiring to fuel glucose-dependent thermogenesis in brown adipose tissue. Sci Adv (2025) 11(24):eadt7369. doi: 10.1126/sciadv.adt7369
3
VaanholtLMDaanSSchubertKAVisserGH. Metabolism and aging: effects of cold exposure on metabolic rate, body composition, and longevity in mice. Physiol Biochem Zool (2009) 82(4):314–24. doi: 10.1086/589727
4
BelskyDWCaspiACorcoranDLSugdenKPoultonRArseneaultLet al. DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife (2022) 11:e73420. doi: 10.7554/eLife.73420
5
OhHS-HRutledgeJNachunDPálovicsRAbioseOMoran-LosadaPet al. Organ aging signatures in the plasma proteome track health and disease. Nature (2023) 624:164–72. doi: 10.1038/s41586-023-06802-1
Summary
Keywords
aging, aging rate indicator, biomarker, biomarkers of aging, geroscience, gerotherapeutics
Citation
McGaunn JP and Baur JA (2026) Catching aging in motion. Front Sci 4:1939931. doi: 10.3389/fsci.2026.1939931
Received
16 July 2026
Accepted
10 August 2026
Published
17 September 2026
Volume
4 - 2026
Edited and reviewed by
Felipe Sierra, Geroscience Consulting LLC, United States
Updates
Copyright
© 2026 McGaunn and Baur.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Joseph A. Baur, baur@pennmedicine.upenn.edu
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.