What Is My Biological Age?

Extra Insights & Sources

For readers who want the details behind each claim above.

What We Know

• Epigenetic clocks predict mortality and disease at the population level. Levine’s PhenoAge showed a 4.5% rise in all-cause mortality per year of methylation age, pooled across several study groups; GrimAge “stands out” for time-to-death across 6,935 people. Supported.
• DunedinPACE repeats almost exactly in the lab (ICC 0.96) and moved by 2–3% in one two-year randomized trial. Supported for technical reliability; Preliminary as an intervention endpoint.
• Grip strength, gait speed and treadmill fitness predict survival in cohorts of 34,000 to 139,000 people, followed for a median of 4 years (grip) up to as long as 21 years (walking speed). Established.
• Technical noise on uncorrected clocks produced deviations of up to 9 years between replicates; principal-component versions agree within about 1.5 years. Supported.

What We Don’t Know

• What a single person’s clock reading means for that person. Herzog: “We don’t know what [a clock result] means for a person.” Unknown.
• Whether lowering an epigenetic age number changes when you die or how long you stay healthy. No trial has tested it. Unknown.
• Which clock is “right.” Horvath says the field lacks “reference standards”; MIT Technology Review reports that one sample through five clocks gives “five wildly different results.” Contested.
• How much day-to-day biology, rather than the lab, moves your number. Sehgal’s group reports that clocks can be precise in the lab yet unstable in the body. Preliminary.

The five kinds of test, side by side

How each clock was built

Horvath’s clocks. NBC reports that Horvath helped develop the first epigenetic clock in 2011 using saliva, and that it holds a US patent. His full comment: “nobody uses this clock — nobody, including me,” Horvath said. “It just wasn’t good enough.” His 2013 Genome Biology paper built a multi-tissue age predictor from 8,000 samples in 82 datasets, covering 51 healthy tissues and cell types; the clock uses 353 CpG sites, per the paper’s abstract. NBC describes Horvath as a professor of human genetics and biostatistics at UCLA’s David Geffen School of Medicine; The Scientist described him as a principal investigator at Altos Labs when he spoke at the Precision Medicine World Conference 2026.

PhenoAge (Levine et al., Aging, 2018). Step one used NHANES III, a nationally representative US survey with more than 23 years of mortality follow-up: 9,926 adults with complete biomarker data. Researchers used a statistical model to examine how 42 clinical markers and chronological age related to deaths during follow-up. They selected nine markers for the phenotypic-age calculation. That blood-based phenotypic age was validated in 6,209 NHANES IV adults with up to 12 years of mortality follow-up, where it correlated with chronological age at r = 0.94. The methylation version, DNAm PhenoAge, was then tested in other cohorts: “A one-year increase in DNAm PhenoAge is associated with a 4.5% increase in the risk of all-cause mortality” (fixed-effect meta-analysis 1.045, p = 7.9E-47, adjusted for chronological age). Conflicts of interest, per the paper: “The Regents of the University of California is the sole owner of a provisional patent application directed at this invention for which MEL, SH are named inventors.”

GrimAge (Lu et al., Aging, 2019). GrimAge is “a composite biomarker based on the seven DNAm surrogates and a DNAm-based estimator of smoking pack-years.” The surrogates are methylation patterns that stand in for seven plasma proteins, including PAI-1 and GDF-15. Validation used 7,375 blood methylation arrays from 6,935 people in five cohorts (about 50% European ancestry, 40% African American and 10% Hispanic). The authors write that it “stands out among existing epigenetic clocks in terms of its predictive ability for time-to-death” (Cox regression p = 2.0E-75). The paper states the University of California owns a provisional patent application on it, with Lu and Horvath named as inventors.

DunedinPACE (Belsky et al., eLife, 2022). The Dunedin Study follows all 1,037 babies born in Dunedin, New Zealand, between April 1972 and March 1973 (91% of eligible births; 52% male). The team modeled the decline in 19 indicators of organ-system integrity across four time points spanning two decades, then distilled it into a single-time-point blood test and evaluated it in five more datasets. On technical replicates, “reliability was excellent (ICC = 0.96 [0.93–0.98]),” compared with 0.89 for the earlier DunedinPoAm. ICC, the intraclass correlation coefficient, is a score from zero to one for how closely repeat measurements agree. Funding came from the US National Institute on Aging and the UK Medical Research Council. Competing interests, per the paper: Belsky, Moffitt, Caspi, Sugden and co-authors are “listed as inventors on a Duke University and University of Otago invention that was licensed to a commercial entity.” NBC reported in April 2026 that DunedinPACE “is now exclusively licensed to TruDiagnostic, a company that sells biological age tests and for which Belsky is an occasional adviser.”

The physical tests in detail

Walking speed (Studenski et al., JAMA, 2011). A pooled analysis of nine cohort studies collected between 1986 and 2000: 34,485 community-dwelling adults aged 65 or older, mean age 73.5, 59.6% women and 79.8% white, followed for 6 to 21 years; 17,528 died. Gait speed was associated with survival in every study: pooled hazard ratio 0.88 per 0.1 m/s (95% CI, 0.87–0.90; P < .001). A hazard ratio compares one group’s risk with another’s over the study; below one means lower risk. At age 75, predicted 10-year survival across the range of gait speeds ran from 19% to 87% in men and 35% to 91% in women. Several authors reported grants or consulting fees from drug companies, including Merck.

Grip strength (Leong et al., PURE, The Lancet, 2015). Of 142,861 people enrolled in 17 countries between 2003 and 2009, 139,691 aged 35–70 with known vital status were analyzed; over a median of 4.0 years, 3,379 (2%) died. Each 5 kg lower grip strength carried a hazard ratio of 1.16 for all-cause death (95% CI 1.13–1.20) and 1.17 for cardiovascular death. The authors reported that grip strength was a stronger predictor of all-cause and cardiovascular mortality than systolic blood pressure, and called for research testing whether improving strength reduces mortality and cardiovascular disease. (Note 1)

Fitness (Mandsager et al., JAMA Network Open, 2018). A retrospective Cleveland Clinic cohort of 122,007 consecutive adults referred for treadmill testing between 1991 and 2014 (mean age 53.4; 59.2% male), followed for a median of 8.4 years; 13,637 died. Fitness was measured as peak estimated metabolic equivalents on the treadmill, not directly measured VO2max. Elite versus low performers had an adjusted hazard ratio of 0.20 (95% CI 0.16–0.24). In the paper’s key points, “cardiorespiratory fitness was inversely associated with all-cause mortality without an observed upper limit of benefit.” The authors reported no conflicts of interest.

What each wearable uses

Oura says Cardiovascular Age is “an estimate of the health of your cardiovascular system in relation to your actual age,” calculated from an estimated pulse wave velocity and age-related changes in the shape of the heartbeat signal; the ring does not measure pulse wave velocity directly. The app shows a chart against your actual age and three levels: Below (6 or more years below your actual age), Aligned (up to 5 years above or below) and Above (6 or more years above). The page adds that “Oura Ring is not a medical device and is not intended to diagnose, treat, cure, monitor, or prevent medical conditions or illnesses.”

WHOOP’s 2025 Healthspan white paper says it “was made possible through the collaboration and support of the Buck Institute for Research on Aging,” with a foreword by the institute’s CEO, Eric Verdin, MD. It states: “WHOOP uses 9 key metrics spanning 3 domains to calculate your WHOOP Age,” namely sleep duration and consistency; steps, heart-rate-zone time and strength activity; and VO2 Max, resting heart rate and lean body mass percentage, based on the last six months of data. The paper says WHOOP Age is determined by estimating the all-cause mortality risk associated with a member’s data, and it describes the output as a single “‘physiological’ or ‘functional’ age.”

The free route, the paid kits and the regulators

The nine PhenoAge markers (Table 1 of Levine 2018): albumin, creatinine, serum glucose, C-reactive protein, lymphocyte percentage, mean cell volume, red cell distribution width, alkaline phosphatase and white blood cell count, plus chronological age. Matching them to a comprehensive metabolic panel, a CBC with differential and a CRP is our mapping, not a quoted source.

BioAge (Kwon and Belsky, GeroScience, 2021). The R package, at github.com/dayoonkwon/BioAge, implements Klemera-Doubal biological age, PhenoAge and homeostatic dysregulation, and is licensed under the GNU General Public License v3.0. The authors declared no competing interests.

InsideTracker InnerAge. The store page lists a $149-a-year membership and says “InnerAge analysis is $99 per calculation.” At minimum it needs “Albumin (Serum protein), Glucose (Blood sugar), LDL cholesterol (Cholesterol transporter), Triglycerides (Fats for energy storage).”

TruDiagnostic TruAge ($499). The product page describes “a quick, painless finger-prick” and results “within 3–4 weeks after our lab receives your sample.” Reports include OMICmAge (biological age), DunedinPACE (pace) and SYMPHONYAge (11 organ-system ages). A TruAge plus TruHealth combination costs $849.

Elysium Index ($299). Described as “a saliva-based DNA test” reporting a biological age, a cumulative rate of aging and nine system age scores. Elysium also sells supplements, including Basis.

UK prices (Lola Health, rechecked September 25, 2026; the page is undated). GlycanAge about £285 (an IgG-glycan test); Paar London £690 excluding VAT, including a consultation; epiAge about £170 (EUR 199) from saliva; StrideBioage about £699 a year, membership only.

Horvath on saliva and standards (The Scientist, 2026). “Saliva, by contrast, is a mixture of approximately 65 percent immune cells and 35 percent buccal epithelial cells,” he said in a follow-up interview. At the conference panel: “There’s no question that epigenetic clocks perform very well in epidemiological studies,” and “But what the field has not yet achieved is really what one would call reference standards. It’s not yet possible that you can go to different validated labs and get the same readout.” Herzog told MIT Technology Review that there are hundreds or even thousands of clocks out there.

The FDA’s wording (page content current as of December 20, 2019): “In general, direct-to-consumer tests for non-medical, general wellness, or low risk medical purposes are not reviewed by the FDA before they are offered.”

Topol on price (Scientific American, 2025): “You can run an epigenetic test in a very simple way for $10 or $20, but some of these companies are charging $200.”

The experts in their own words

Belsky (NBC News, April 5, 2026): “Most people’s values fall between 0.5 and 1.5. Even a 10% difference in the face of aging is a big deal. If you only age 11 months for every 12 that you live, that’s not bad.” On DunedinPACE: “[DunedinPACE] is your speedometer, tells you how fast you’re going.”

Vaughan (NBC News): “If you do a test and you find out that your biological age is 10 years younger than your chronological age, well, that makes for great cocktail party conversation. But what’s your cholesterol? What’s your blood pressure? If those kinds of things aren’t addressed, you’re going to have a problem.” And: “The data and the value of those kinds of tests is quite uncertain.”

Topol (Scientific American, July 14, 2025, interview by Lauren J. Young): “I haven’t seen their publications to be able to say with confidence that they are doing things right, and the lack of standards from one company to the next is disconcerting. They don’t want to shock customers by telling them that they’re 10 years older than their chronological age.”

Herzog (MIT Technology Review, October 14, 2025): “They’re not yet individually predictive. We don’t know what [a clock result] means for a person, [or if] they’re more or less likely to develop disease.”

Kaeberlein (Lifespan.io, March 23, 2026): epigenetic tests “don’t measure biological age. They measure DNA methylation status, which is correlated with chronological age (or disease risk or mortality), each of which is, in turn, correlated with biological age. Biological age tests are, at best, correlations to correlations. More importantly, outside of controlled research, their accuracy and precision are unknown.” His alternative: “We already have outstanding tools: glucose, insulin, lipids, DEXA, cardiac scans, VO₂max, strength and agility tests, cognitive assessments… the list goes on and on.” DEXA is a scan that measures bone density and body composition. His byline lists him as CEO of Optispan, Inc., and affiliate professor at the University of Washington.

Olshansky (Nature Aging, 2024, via AFAR). The analysis used national vital statistics from 1990 to 2019 for the eight countries with the longest-lived populations, plus Hong Kong and the United States. It found that survival to age 100 is unlikely to exceed 15% for females and 5% for males. The paper declares no competing interests; AFAR lists Olshansky as co-founder and chief scientist of Lapetus Solutions, Inc. The Medscape remark (June 20, 2025) is reported as a paraphrase; Medscape requires a login.

How the trials worked

CALERIE (Waziry et al., Nature Aging, 2023). CALERIE Phase 2 ran at three US centers. Of 238 eligible adults, 220 were randomized 2:1 to a calorie-restriction program (145) or their usual diet (75); men were 21–50 and premenopausal women 21–47, with body mass index 22.0–27.9. The prescribed target was 25% below each person’s baseline intake. DunedinPACE slowed by 2–3% at 12 and 24 months (standardized effects of about −0.29 and −0.25); PhenoAge and GrimAge did not change significantly. Columbia’s release adds that, in other studies, a change of that size “translates to a 10-15 percent reduction in mortality risk”; CALERIE itself did not measure deaths. Competing interests, per the paper: co-authors D.L. Corcoran and D.W. Belsky “are listed as inventors on a Duke University and University of Otago invention, DunedinPACE, that was licensed to a commercial entity.”

COSMOS (Li et al., Nature Medicine, 2026). A prespecified ancillary study of 958 participants (482 women, 476 men) tested a daily multivitamin-multimineral (Centrum Silver) and cocoa extract (500 mg cocoa flavanols a day) against placebo for two years, on five clocks: PCHannum, PCHorvath, PCPhenoAge, PCGrimAge and DunedinPACE. The multivitamin slowed yearly change by 0.113 years on PCGrimAge (95% CI −0.205 to −0.020; P = 0.017) and 0.214 years on PCPhenoAge (−0.410 to −0.019; P = 0.032), with a larger effect in people whose biological aging was already accelerated. Cocoa extract had no effect on any of the five. The authors described the effects as “statistically significant but small” and said further studies are needed to determine their clinical relevance. Per the paper, Pfizer Consumer Healthcare (now Haleon) and Mars Edge donated study pills and packaging, and FOXO Technologies provided in-kind donations to generate the DNA methylation data.

Retest numbers by clock

Higgins-Chen et al., Nature Aging, 2022 (senior author Morgan Levine): “technical noise produces deviations up to 9 years between replicates for six prominent epigenetic clocks.” Principal-component clocks, which use a statistical method to separate shared signal from noise, showed “agreement between most replicates within 1.5 years.” An earlier 2021 conference abstract from the same group put the spread at 3 to 9 years; one of its co-authors was at Elysium Health. Competing interests, per the 2022 paper: Levine and Higgins-Chen “have built epigenetic aging metrics involving the technology described in the present manuscript, and these metrics are licensed by Elysium Health through Yale University”; Elysium supplied paired blood and saliva replicate datasets but did not fund the study; Levine previously acted as a scientific adviser to Elysium and received consulting fees; Higgins-Chen received consulting fees from FOXO Technologies for unrelated work.

Sehgal et al., Aging Cell, 2026 (18 clocks). On technical replicates, the authors reported that most clocks reached excellent technical reliability (ICC above 0.9), with SystemsAge, GrimAge and the PC-based versions among the most stable, while earlier-generation clocks such as Hannum, PhenoAge and PCPhenoAge sat in the good range (ICC about 0.7–0.8). On biological repeats taken around meals, stress, pollution and altitude change, most clocks reached only moderate to good reliability (ICC about 0.4–0.7), no clock reached excellent, and GrimAge and DunedinPACE clustered near the lower end. The authors concluded that a clock can be precise in the lab yet biologically unstable, with values that swing widely after short-lived, reversible and minor exposures, such as one stressful event or simply a different time of day. Conflicts, per the paper: Sehgal and Higgins-Chen are co-inventors of the SystemsAge framework, which is the subject of a patent application; Higgins-Chen has received consulting fees from TruDiagnostic and FOXO Biosciences, and Sehgal from TruDiagnostic, LongevityTech.fund and Cambrian BioPharma.

Kaeberlein (Lifespan.io): “In my own experiment, the same sample sent to the same company produced biological age estimates nearly 15 years apart. Across multiple companies, the discrepancies were even larger.”

Shalev and Apsley (The Conversation, May 4, 2026; Shalev is an associate professor of biobehavioral health at Penn State, Apsley a postdoctoral researcher in molecular biology at Illinois): “Testing epigenetic age in saliva versus blood samples can yield substantially different results for the same person.”

Hine (NBC News): “Only doing it once isn’t so telling, because you don’t know if you’re [on] an upward trajectory or downward trajectory in terms of the rate of your aging.” NBC also reports his point that something as simple as a cold might influence results.

Sources

• Horvath S. DNA methylation age of human tissues and cell types, Genome Biology, 2013: doi 10.1186/gb-2013-14-10-r115
• Levine ME, et al. PhenoAge, Aging, 2018: PMC5940111
• Lu AT, et al. GrimAge, Aging, 2019: PMC6366976
• Belsky DW, et al. DunedinPACE, eLife, 2022: elifesciences.org/articles/73420
• Studenski S, et al. Gait speed and survival, JAMA, 2011: jamanetwork.com
• Leong DP, et al. PURE grip strength, The Lancet, 2015: thelancet.com
• Mandsager K, et al. Fitness and mortality, JAMA Network Open, 2018: PMC6324439
• Kwon D, Belsky DW. BioAge toolkit, GeroScience, 2021: PMC8602613
• Higgins-Chen AT, et al. Principal-component clocks, Nature Aging, 2022: nature.com; conference abstract, Innovation in Aging, 2021: PMC8679190
• Sehgal R, et al. Biological versus technical reliability of epigenetic clocks, Aging Cell, 2026: PMC13418614
• Waziry R, et al. CALERIE and DNA methylation aging, Nature Aging, 2023, via Columbia Mailman School release: ScienceDaily
• Li S, et al. COSMOS epigenetic clocks, Nature Medicine, 2026: PubMed 41803341
• Olshansky SJ, et al. Implausibility of radical life extension, Nature Aging, 2024, via AFAR: afar.org
• NBC News (Lindsey Leake), April 5, 2026: nbcnews.com
• Scientific American (Lauren J. Young), July 14, 2025: scientificamerican.com
• Kaeberlein M. Lifespan.io op-ed, March 23, 2026: lifespan.io
• The Scientist, 2026: the-scientist.com
• MIT Technology Review (Jessica Hamzelou), October 14, 2025: technologyreview.com
• Medscape, June 20, 2025: medscape.com
• Shalev I, Apsley A. The Conversation, May 4, 2026: theconversation.com
• US Food and Drug Administration: Direct-to-Consumer Tests
• NHS: NHS Health Check
• The Online GP: Epigenetic testing UK guide
• Lola Health: Best biological age tests UK
• TruDiagnostic: TruAge Test
• Elysium Health: Index
• InsideTracker: Membership
• Oura: Cardiovascular Age
• WHOOP: 2025 Healthspan white paper