Biological Age Tests
Can You Reverse Your Biological Age?
A biological age score has dropped in several small human trials. No trial has shown a lower score means slower aging, less disease or a longer life.
A biological age test sells you a second age, one that isn’t printed on your driver’s license. Naturally, the next thing people type is “can you reverse your biological age?” Toast can’t be un-toasted. A DNA test result, apparently, is more negotiable, and researchers have run actual human trials to find out how much. Humanic Blue is looking at what changed in those trials—and what did not.
Six of those trials, published between 2019 and 2026, reported a lower score after an intervention: a drug cocktail, diet coaching, calorie restriction, a fasting-mimicking diet, omega-3 capsules and a yogurt-plus-exercise program. Where the change was counted in years, it ran from about three months to about three years. So yes, the number can go down.
What no trial has shown is that a lower score means slower aging, fewer diseases or a longer life. None followed people long enough to find out, and the scientists who built the clocks say so themselves. Here are the trials in order, with who ran each one and who paid.
Key Takeaways
- Clock scores fell in six small or secondary human trials, by amounts from a few months to about three years. Supported, as a change in the score.
- The randomized trials that ran two or three years found small effects: a 2–3% slower pace of aging on calorie restriction, and about three to four months over three years on omega-3. Promising, and small.
- No trial has tested whether a lower score means less disease or a longer life. Unknown.
- The same blood sample run twice can differ by years on standard clocks, which is bigger than most of the reported changes. Supported.
- Exercise, not smoking, blood pressure control and enough sleep have evidence on disease and death, not just on clocks. Established.
Which trials lowered a biological-age score?
An epigenetic clock is an algorithm that reads chemical tags (methylation) on your DNA and turns them into an age estimate. Pace of aging measures such as DunedinPACE give a speed instead, where 1.0 means one year of biological aging per calendar year. Neither is a diagnosis.
Each of the six trials below has a company, a product or a clock’s inventor attached to it. That doesn’t make a result wrong, but it’s worth reading alongside the numbers. See all six side by side
- TRIIM, 2019: Nine men completed a year of growth hormone, DHEA and metformin with no comparison group. Their average clock age fell, but the researchers said a follow-up large enough to test the result reliably was needed. The sponsor, patent and author ownership ties all ran through Intervene Immune. Preliminary.
- Fitzgerald, 2021: About 40 men were analyzed after eight weeks of lifestyle coaching, supplements and probiotics versus no intervention. The program group read about 3.2 years younger than controls on a saliva clock. The trial was funded by Metagenics, and two authors use or earn from the program. Preliminary.
- CALERIE, 2023: In adults assigned to calorie restriction or their usual diet for two years, DunedinPACE slowed 2–3%; PhenoAge and GrimAge did not. The effect was small, and senior author Daniel Belsky co-invented the commercially licensed clock. Promising for the biomarker; unknown for health or lifespan.
- Fasting-mimicking diet, 2024: An additional analysis of about 100 adults reported a nearly 2.5-year median drop on a blood-and-clinic score after three monthly diet cycles. About one-third of intervention participants scored older, and L-Nutra supplied the diet while holding licensed intellectual property with USC and founder Valter Longo. Preliminary.
- DO-HEALTH, 2025: Among about 780 older Swiss adults in a three-year randomized trial, omega-3 modestly slowed three clocks—about three to four months over three years. The clock analysis was added after the original trial design, and clock inventors and commercial donors were involved. Promising, and small.
- Yogurt trial, 2026: In about 50 older men with overweight, a 12-week package of exercise, diet advice and probiotic yogurt slowed DunedinPACE by an estimated 2.2%, but none of the age-clock results remained statistically significant after accounting for the many tests performed. All five authors worked for the company that funded and supplied the yogurt. Preliminary and industry-authored.
That is the whole arc: six scores moved, the clocks did not move consistently, and none of the trials followed disease or death. The designs, exact results and conflicts
What does a lower score not mean?
Steve Horvath developed the Horvath clock and, with Ake Lu, GrimAge. He’s a longtime UCLA professor and a former principal investigator at Altos Labs. On Eric Topol’s Ground Truths in 2024, he said: “An epigenetic age estimate is certainly not the same as a biologic age estimate.”
He also explained why individuals aren’t there yet: “at the moment, we don’t have validated interventions against accelerated epigenetic age.” His point concerns a validated intervention for an individual with an accelerated score. It does not contradict the small group-level score changes in the trials above.
Vadim Gladyshev’s lab at Harvard Medical School and Brigham and Women’s Hospital reported in Cell Metabolism that biological-age estimates can move quickly in either direction. In people or mice, clock ages rose during major surgery, pregnancy and severe COVID-19, then fell with recovery. A falling score can mean a stress ended, not that a body got younger.
A 2023 consensus paper in Cell from the Biomarkers of Aging Consortium, with Gladyshev as senior author, says the field still lacks agreed standards for deciding whether an aging biomarker is reliable enough for clinical validation.
Matt Kaeberlein, CEO of the longevity-medicine company Optispan and affiliate professor at the University of Washington, wrote in March 2026 that even accurate tests “would at best show whether your trajectory is improving or declining without providing any information about cause and effect.” A lower number doesn’t tell you what lowered it.
S. Jay Olshansky, professor of epidemiology and biostatistics at the University of Illinois Chicago, examined the prospects for large gains in human life expectancy. His 2024 Nature Aging analysis of the world’s longest-lived populations concluded that “unless the processes of biological aging can be markedly slowed, radical human life extension is implausible in this century.” That analysis concerns population survival, not clock scores. The three to four months reported in DO-HEALTH are a clock change, not months of life gained.
Eric Topol, a cardiologist and genomics professor at Scripps Research, told Scientific American: “There’s only one thing that’s been definitively shown to slow the epigenetic aging process, and that’s exercise.” He didn’t say that slowing the clock has been shown to help anyone live longer. The experts’ longer remarks
Which missing outcome would help distinguish a lower age score from reversal of aging?
Why can a score drop when nothing about you changed?
The measurement wobbles. A Yale team, with Albert Higgins-Chen as first author, reported in Nature Aging that the same blood, run twice, can land up to 9 years apart on six standard clocks. That’s larger than any change reported in years in the trials above. Their rebuilt versions agreed within 1.5 years for most repeat runs; metrics built with that method are licensed to Elysium Health through Yale.
Kaeberlein saw it firsthand: “the same sample sent to the same company produced biological age estimates nearly 15 years apart.” His comparison: a “biological age” number in a PDF is “no different than giving them a horoscope.”
High scores drift back. Adrian Barnett and colleagues, in the International Journal of Epidemiology, describe regression to the mean as a statistical effect that can make natural variation in repeated measurements look like real change. An unusually high first reading may partly reflect temporary variation. The next reading may be closer to the average even if an intervention did nothing. A trial with no comparison group, like TRIIM, or a six-person case series can’t rule it out.
The designs were loose. TRIIM had no comparison group, and Fitzgerald’s controls got nothing, so nobody was blinded. The fasting-diet and DO-HEALTH clock analyses were exploratory or added to trials originally designed for other questions. That makes their findings less conclusive than a trial designed specifically to test a clock result.
Someone had a stake. As the trial list shows, every one had a sponsor, a product or a clock inventor attached. The TRIIM, Fitzgerald and yogurt-trial authors each say larger studies are needed to confirm their results.
Idan Shalev, associate professor of biobehavioral health at Penn State, and Abner Apsley of the University of Illinois Urbana-Champaign wrote in The Conversation that the clocks are “highly effective research tools to study aging at the population level,” but “aren’t designed to make claims about the health of individuals.” Shalev was also a co-author of the CALERIE analysis. An average pace-of-aging change in a trial does not establish how precisely a test can track one person. Pace scores also use a different scale from age scores; the errors reported in years cannot simply be applied to a rate. The noise studies in detail
What comparison would separate an intervention effect from ordinary variation in a clock?
What has evidence for longer, healthier lives?
Kaeberlein points elsewhere: “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 of body composition and bone density; VO₂max is a treadmill or bike test of aerobic capacity.
Four habits have evidence on the things clocks are supposed to predict, which is living longer and staying well:
- Exercise. The CDC and NHS both advise 150 minutes of moderate or 75 minutes of vigorous activity a week, plus strength work on 2 days. The WHO says inactive people have a 20% to 30% higher risk of death.
- Not smoking. The CDC says quitting “can add as much as 10 years to life expectancy.”
- Blood pressure. In the US, the CDC, citing the 2025 AHA/ACC guideline, counts 130/80 or higher as high; the NHS uses 140/90 in a clinic or 135/85 at home.
- Sleep. The CDC recommends “7 or more hours” a night for adults aged 18 to 60, 7 to 9 hours at 61 to 64, and 7 to 8 hours from 65.
Notice the country split on blood pressure: the same reading can count as high in Chicago and not in Manchester. That’s a fact about guidelines, and a reason to ask which one your doctor uses. The guideline wording
What if you still want to take a test?
As of September 26, 2026, TruDiagnostic’s TruAge lists at $499 for a finger-prick blood kit, with results 3–4 weeks after the lab gets the sample. Elysium’s Index lists at $299 for saliva, with results in about six weeks.
Elysium also sells supplements, including Basis, Matter, Signal and a Senolytic Complex. The company that scores you sells things a lower score could be credited to.
Neither company says its test is FDA-reviewed. Elysium says Index “has not been reviewed or approved by the FDA,” and TruDiagnostic calls TruAge “not a diagnostic test.” The FDA’s 2024 attempt to regulate lab-developed tests as devices was vacated by a court, and in September 2025 the agency reverted its rule. Topol’s view of the market: “the lack of standards from one company to the next is disconcerting.”
Higgins-Chen’s replicate figure and Kaeberlein’s retest are the reasons to ask a lab for the named method’s published test-retest error in the units of its output—years for an age estimate or rate units for a pace estimate—before paying, and to use the same lab and sample type for any repeat. Prices and regulation in detail
What’s Interesting
DunedinPACE was built to track pace, and it changed in CALERIE, DO-HEALTH and the yogurt trial. Those results concern a rate score rather than years removed from a biological age. GrimAge, which its developers reported “strongly predicts lifespan,” didn’t change significantly in CALERIE and wasn’t among the clocks omega-3 slowed in DO-HEALTH.
The Gladyshev lab’s finding that scores rise under stress and fall on recovery reframes some “reversal” as recovery. Less cinematic, maybe more useful: it suggests the number is partly a snapshot of how your body is coping right now.
And every trial here reports averages across groups of volunteers. Your starting point, genes and health history may mean a change that moved a group’s number moves yours more, less or not at all.
What This Means in Practice
- If you track the research, watch for trials built from the start to test a clock and to follow people to disease and death. That’s what the CALERIE authors say a conclusive test needs.
- If your number came back higher than you expected, keep Higgins-Chen’s finding in mind: the same sample can land up to nine years apart on standard clocks.
- If you’re new to this, “pace of aging” is a research measure, not a verdict on your body.
- If you want to live longer and better, the four habits above have outcome evidence the clocks don’t.
- If you test, ask for the lab’s test-retest error, and keep the lab and sample type the same.
A primary care physician, or where available a geriatrician or preventive cardiologist, is the professional to ask. Kaeberlein’s list makes a good question: which of the tools already in the clinic should I be tracking, and what would a change in them mean?
The measurement question behind any before-and-after claim is examined in Can Biological Age Tests Track an Intervention?.
Bottom Line
A biological age score can go down. Six small or secondary human trials lowered one, and the two randomized trials that ran two or three years found small effects. Every trial had a sponsor, product or clock inventor attached, and standard clocks can wobble by years on the same sample. The open question is the one the CALERIE team set: whether a lower score means less disease or a longer life. Answering it takes a long trial, and none of these six was one.
A NOTE FROM LONGEVITY365
We read research, consult expert perspectives and work to make complicated health information easier to understand. What you’re reading here reflects how we’ve interpreted and organized that information. It is educational content—not personal medical advice, diagnosis or treatment guidance.
We’re not perfect, and neither is the available research. Evidence changes, new studies emerge and what’s appropriate for one person may not be appropriate for another.
Use this article as a starting point for your own research and conversations with qualified professionals. Always talk with your doctor or another trusted healthcare professional before making decisions that could affect your health, particularly if you take medications, are receiving treatment, have existing health conditions, are pregnant or have other circumstances that could affect what’s appropriate for you.
Reader Questions
A few questions to carry into the comments or your next conversation. Please leave out identifying or personal medical details.
- 01After reading this, what would you want a test result to actually help you decide?
- 02Which distinction changed how you read the evidence: "Which trials lowered a biological-age score" or "What if you still want to take a test"?
- 03What would you ask a clinician, lab, or qualified professional before letting a number change a health decision?
Extra Insights & Sources
For readers who want the details behind each claim above.
What We Know
- Several small or secondary human trials have lowered an epigenetic or blood-chemistry age score. Supported, as a change in the score.
- In CALERIE, two years of calorie restriction slowed DunedinPACE by 2–3%, with no significant change on PhenoAge or GrimAge. Promising.
- In DO-HEALTH, three years of omega-3 slowed three clocks by 2.9 to 3.8 months. Promising, and small.
- Technical noise on standard clocks can reach nine years between replicates of the same sample. Supported.
What We Don’t Know
- Whether any change in a clock score changes the risk of disease or death. No trial in this list followed people to those endpoints. Unknown.
- Whether the effects replicate without the sponsor or the clock’s inventor involved. Unknown.
- Whether a drop in one person’s score is signal or noise. Unknown.
The six trials in detail
| Year | Trial | n | Design | Duration | Measure | Reported change | Funding / COI |
|---|---|---|---|---|---|---|---|
| 2019 | TRIIM (Fahy) | 10 men enrolled, 9 completed, 51–65 | Open-label, no control, not preregistered | 12 months | Horvath, Hannum, PhenoAge, GrimAge | ~1.5 yr below baseline; −2.5 yr vs expected | Intervene Immune; authors hold shares or options; officers named on a patent application (Note 1) |
| 2021 | Fitzgerald | 43 men randomized, 38 analyzed, 50–72 | Randomized; controls got nothing | 8 weeks | Horvath 2013 clock, run on saliva | −3.23 yr vs controls (p = 0.018) | Metagenics grant; two authors use and earn from the program (Note 2) |
| 2023 | CALERIE (Waziry, Belsky) | 220 randomized, 197 with DNAm data, mean age 38 | Randomized controlled, post hoc analysis | 2 years | DunedinPACE, PhenoAge, GrimAge | Pace slowed 2–3%; PhenoAge and GrimAge not significantly changed | DunedinPACE inventors licensed it commercially (Note 3) |
| 2024 | FMD (Brandhorst, Levine, Longo) | 52 FMD vs 48 control, 18–70 | Randomized; secondary, exploratory analysis | 3 monthly cycles | Blood-chemistry age (not epigenetic) | Median −2.5 yr | L-Nutra supplied the diet; USC licensed IP to L-Nutra; Longo holds equity (Note 4) |
| 2025 | DO-HEALTH (Bischoff-Ferrari) | 777 of 2,157 adults, 70+ | Randomized, placebo-controlled, post hoc | 3 years | PhenoAge, GrimAge, GrimAge2, DunedinPACE | Omega-3: 2.9–3.8 months slower | EU grant plus Roche, NESTEC, Pfizer, others; Horvath’s clock foundation (Note 5) |
| 2026 | Nishimura | 48 overweight men randomized (24 + 24), 45 per-protocol, 50–74 | Exploratory randomized controlled | 12 weeks | DunedinPACE | 2.2% slower pace | All five authors employed by Morinaga Milk Industry, which funded it (Note 6) |
TRIIM (2019). Fahy GM and colleagues, Aging Cell. “Ten nominally healthy adult men from 51–65 years of age were recruited for the study.” One volunteer was removed after about a month (self-reported bradycardia and a belatedly disclosed family history of cancer). The trial was, “in accordance with FDA guidelines for pre‐Phase I exploratory studies… not preregistered on clinicaltrials.gov.” Headline result: “a mean epigenetic age approximately 1.5 years less than baseline after 1 year of treatment (−2.5‐year change compared to no treatment at the end of the study).” The four clocks were Horvath, Hannum, PhenoAge and GrimAge. Conflict statement: “GMF, RTB, JPW, and SH are shareholders in or have options to purchase shares in Intervene Immune, Inc., GMF and RTB are officers of Intervene Immune and are named in a related Intervene Immune patent application.” SH is Steve Horvath. (Note 1)
Fitzgerald (2021) and its 2024 correction. Fitzgerald KN and colleagues, Aging. From the abstract: “we report on a randomized controlled clinical trial conducted among 43 healthy adult males between the ages of 50-72. The 8-week treatment program included diet, sleep, exercise and relaxation guidance, and supplemental probiotics and phytonutrients. The control group received no intervention.” Result: “The diet and lifestyle treatment was associated with a 3.23 years decrease in DNAmAge compared with controls (p=0.018).” The analysis included 18 treated men and 20 controls. The original paper reported a within-group drop of 1.96 years (p = 0.066) in the treated men and a non-significant 1.27-year rise in controls (p = 0.153). A 2024 correction, made “to account for the participants’ chronological age increase (8+ weeks),” changed these to 2.04 years younger (p = 0.043) and 1.10 years older (p = 0.191); of the 18 treated men, “8 scored age reduction, 9 were unchanged, and 1 increased.” The between-group difference “is not changed by this correction.” A 2022 correction revised the data-sharing statement. Conflicts: “KF and RH declare that they use the intervention described here in clinical practice, are named in a related patent application, and receive earnings from educational products associated with its use.” The study was funded by “an unrestricted grant from Metagenics, Inc.” (Note 2) The 2023 follow-up is a case series of six women with no control group; it reported an average 4.60-year decrease on the Horvath clock in blood.
CALERIE (2023). Waziry R, Ryan CP (joint first authors), Belsky DW (senior author) and colleagues, Nature Aging: “n = 220 adults without obesity were randomized to 25% CR or ad libitum control diet for 2 yr.” The team analyzed 197 participants with DNA methylation data; mean age 38, 70% women. Achieved restriction was “mean = 11.9” percent. “CALERIE intervention slowed the pace of aging, as measured by the DunedinPACE DNAm algorithm, but did not lead to significant changes in biological age estimates measured by various DNAm clocks including PhenoAge and GrimAge.” The effects “correspond to a reduction in the pace of aging of 2–3%.” The authors also note that for some participants DunedinPACE rose. From the introduction: “Ultimately, a conclusive test of the geroscience hypothesis will require trials with long-term follow-up to establish effects of intervention on primary healthy-aging endpoints, including incidence of chronic disease and mortality.” Conflict statement: “D.L.C. and D.W.B. are listed as inventors on a Duke University and University of Otago invention, DunedinPACE, that was licensed to a commercial entity.” (Note 3)
Fasting-mimicking diet (2024). Brandhorst S and Levine ME (joint first authors), Longo VD (senior author) and colleagues, Nature Communications. The biological-age composite used “albumin, alkaline phosphatase, serum creatinine, C-reactive protein, Hba1c, systolic blood pressure, and total cholesterol.” “After 3 FMD cycles, median biological age in the 52 study participants decreased by nearly 2.5 years.” In the second trial, which compared the diet with a Mediterranean diet, 34 participants on the fasting-mimicking diet contributed data, with a median decrease of 2.7 years. The paper also reports 19 control participants whose biological age rose a non-significant 0.78 years. Sixteen of 52 (30.7%) in the first trial and 8 of 34 in the second showed increased biological age after the diet. The authors caution that “projections obtained by extrapolating the effects after 3–6 months of the FMD to a lifelong intervention should be cautiously interpreted.” Conflict statement: “The experimental FMD was provided by L-Nutra Inc. USC has licensed intellectual property to L-Nutra that is under study in this research. As part of this license agreement, the University has the potential to receive royalty payments from L-Nutra. V.D.L. and T.E.M., who have equity interest in L-Nutra, did not participate in the collection and analysis of the data. One-hundred percent of V.D.L.’s equity will be assigned to charitable organizations and companies including the Create Cures Foundation.” (Note 4)
DO-HEALTH (2025). Bischoff-Ferrari HA and colleagues (senior author Horvath S), Nature Aging. The trial tested vitamin D (2,000 IU per day), omega-3 (1 g per day) and a simple home exercise program (three times 30 minutes a week) in a 2 × 2 × 2 factorial design. Of 1,006 Swiss participants, 777 had consent and samples; the paper describes them as “a healthier and more active subgroup.” From the abstract: “Omega-3 alone slowed the DNAm clocks PhenoAge, GrimAge2 and DunedinPACE, and all three treatments had additive benefits on PhenoAge. Overall, from baseline to year 3, standardized effects ranged from 0.16 to 0.32 units (2.9–3.8 months).” The authors call this “a small protective effect of omega-3 treatment.” In the University of Zurich release, Bischoff-Ferrari said: “There is no generally accepted gold standard for measuring biological age,” and “However, we analyzed the best currently validated epigenetic clocks, which reflect the state of the art.” Conflict statement: “S.H. is a founder of the nonprofit Epigenetic Clock Development Foundation, which has licensed several patents for epigenetic clocks (including for GrimAge) from his former employer, UC Regents. Further, S.H. works for Altos Labs, Cambridge, UK. The other authors declare no competing interests.” Daniel Belsky is among the co-authors. (Note 5)
Two independent experts commented through the Science Media Centre. Partridge: “However, whether there is any direct link between the molecular events mediating the responses of the methylation clocks and the improvements in health awaits further study.” Kevin McConway, emeritus professor of applied statistics at the Open University: “Maybe one of the other clocks will eventually turn out to relate more closely to survival rates and future health. … We just can’t say yet.” McConway declared that he was previously a trustee of the SMC; Partridge declared no interests.
Nishimura (2026). Nishimura T and colleagues (senior author Odamaki T), Aging, titled as “an exploratory randomized controlled trial.” Forty-eight men were randomized (24 intervention, 24 control); per-protocol analyses covered 21 and 24. “The intervention group exhibited a significant deceleration in DunedinPACE, corresponding to an estimated 2.2% slower pace of aging, whereas no meaningful change was observed in the control group.” The full text adds a 2.3% slower pace relative to controls and that “no clock within the biological age remained significant after false discovery rate correction.” The authors write that the study “cannot attribute the observed epigenetic changes to any single component.” Conflict statement: “All authors are employed at Morinaga Milk Industry Co., Ltd., which funded this study and provided the test foods.” (Note 6) The full text was read from an archived copy of the journal page (July 28, 2026), because the journal site blocks automated access.
What the clock-makers and critics said
Horvath, Ground Truths (August 23, 2024). “An epigenetic age estimate is certainly not the same as a biologic age estimate.” On individual testing: “is it helpful for the individual should you get such a test? And I would say we are not quite there yet for a variety of reasons.” He continued: “The main reason being we don’t have good interventions against accelerated epigenetic age.” Horvath is a former principal investigator at Altos Labs (he left in February 2026) and a longtime UCLA professor. GrimAge was published in 2019 by Ake Lu, Horvath and colleagues in Aging under the title “DNA methylation GrimAge strongly predicts lifespan and healthspan.”
Poganik, Gladyshev and colleagues (2023), Cell Metabolism. The paper, titled “Biological age is increased by stress and restored upon recovery,” found transient changes “during major surgery, pregnancy, and severe COVID-19 in humans and/or mice.” The authors declare no competing interests.
Moqri and colleagues, Biomarkers of Aging Consortium (2023), Cell. Lead author Mahdi Moqri, senior author Vadim Gladyshev. Several authors, including Moqri and Gladyshev, have filed patents on measuring aging, according to the paper’s declaration.
Kaeberlein, lifespan.io (March 23, 2026). Full sentence: “Handing a patient a PDF with a ‘biological age’ number is no different than giving them a horoscope.” He continues: “Worse, it risks replacing evidence-based care with false reassurance or unnecessary fear.” His byline lists him as chief executive officer of Optispan, Inc. and affiliate professor at the University of Washington.
Topol, Scientific American (July 14, 2025). On consumer tests: “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.”
Shalev and Apsley, The Conversation (May 4, 2026). Full sentence: “But while epigenetic clocks are highly effective research tools to study aging at the population level, they aren’t designed to make claims about the health of individuals.” They add that “different clocks do not always agree with one another, even when used on the same person.”
Olshansky and colleagues (2024), Nature Aging. The analysis covered death rates in the eight countries with the longest-lived populations, plus Hong Kong and the United States, from 1990 to 2019. It estimated that survival to age 100 is unlikely to exceed 15% for women and 5% for men.
The noise studies in detail
Higgins-Chen and colleagues (2022), Nature Aging. First author Albert Higgins-Chen, Yale Department of Psychiatry; senior author Morgan Levine, Yale Department of Pathology. From the abstract: “technical noise produces deviations up to 9 years between replicates for six prominent epigenetic clocks.” Their principal-component versions “show agreement between most replicates within 1.5 years.” Every change reported in years in the trial table falls between about 0.24 years (2.9 months) and 3.23 years. Conflict statement: 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 provided paired blood and saliva replicate datasets reported in this study, but otherwise did not fund the study.” Levine “previously acted as a Scientific Advisor for, and received consulting fees from, Elysium Health, Inc.”; she was a paid Elysium adviser from July 2019 to October 2021.
Barnett and colleagues (2005), International Journal of Epidemiology. The authors define regression to the mean (RTM) as a statistical phenomenon that can make natural variation in repeated data look like real change: unusually large or small measurements tend to be followed by measurements closer to the average. They add that it should always be considered as a possible cause of an observed change.
The guidelines in detail
| Intervention | Guideline | The number | Outcome evidence |
|---|---|---|---|
| Exercise | CDC and NHS | 150 min moderate or 75 min vigorous per week, plus strength work on 2 days | WHO: “People who are insufficiently active have a 20% to 30% increased risk of death compared to people who are sufficiently active.” |
| Not smoking | CDC | Quitting “can add as much as 10 years to life expectancy” | Coronary heart disease risk “falling sharply 1-2 years after cessation” |
| Blood pressure | 2025 AHA/ACC via CDC; NHS | US: hypertension at 130/80 or higher; UK: 140/90 in clinic, 135/85 at home | Treatment trials measure strokes and heart attacks, not clock years |
| Sleep | CDC | Adults 18–60: “7 or more hours” a night; 61–64: 7–9 hours; 65 and older: 7–8 hours | CDC counts adults who report less than 7 hours as “getting insufficient sleep“; it says enough sleep can help lower the risk of type 2 diabetes, heart disease, high blood pressure and stroke |
The tests and the regulators
As of September 26, 2026: TruDiagnostic’s TruAge Complete Epigenetic Collection is $499, collected by finger prick, with results “3–4 weeks after our lab receives your sample,” reporting OMICmAge, SYMPHONYAge and DunedinPACE among other measures. Elysium’s Index is $299, collected from saliva: “Approximately six weeks after receiving your saliva sample.” Elysium’s page states: “Index has not been reviewed or approved by the FDA as it is not a medical device.” TruDiagnostic’s TruAge page says: “Because TruAge is not a diagnostic test, it is best positioned as an adjunct to clinical evaluation.” Elysium’s store lists Basis, Matter, Signal and Senolytic Complex supplements.
The FDA’s laboratory-developed tests page states that on May 6, 2024, the agency issued a final rule bringing laboratory-made tests under the device definition; “On March 31, 2025, a federal district court vacated that final rule. On September 19, 2025, the FDA issued a final rule reverting to the text of the regulation as it existed prior to the effective date of the May 2024 final rule.”
Sources
- Fahy GM, et al. Reversal of epigenetic aging and immunosenescent trends in humans (TRIIM), Aging Cell, 2019: PMC6826138 (PubMed 31496122)
- Fitzgerald KN, et al. Potential reversal of epigenetic age using a diet and lifestyle intervention, Aging, 2021: PMC8064200 (PubMed 33844651); corrections 2022 PMC9365547 and 2024 PMC10968716
- Fitzgerald KN, et al. Potential reversal of biological age in women (case series), Aging, 2023: DOI 10.18632/aging.204602 (PubMed 36947707)
- Waziry R, et al. Effect of long-term caloric restriction on DNA methylation measures of biological aging (CALERIE), Nature Aging, 2023: PMC10148951 (PubMed 37118425)
- Brandhorst S, Levine ME, et al. Fasting-mimicking diet causes hepatic and blood markers changes indicating reduced biological age and disease risk, Nature Communications, 2024: PMC10879164 (PubMed 38378685)
- Bischoff-Ferrari HA, et al. Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks (DO-HEALTH), Nature Aging, 2025: PMC11922767 (PubMed 39900648)
- Nishimura T, et al. Short-term responsiveness of DNA methylation-based aging biomarkers to a multimodal intervention (yogurt with B. longum BB536), Aging, 2026: DOI 10.18632/aging.206386 (PubMed 42377118); full text via archived copy
- Lu AT, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan, Aging, 2019: PubMed 30669119
- Higgins-Chen AT, et al. A computational solution for bolstering reliability of epigenetic clocks, Nature Aging, 2022: DOI 10.1038/s43587-022-00248-2 (PubMed 36277076)
- Poganik JR, et al. Biological age is increased by stress and restored upon recovery, Cell Metabolism, 2023: DOI 10.1016/j.cmet.2023.03.015 (PubMed 37086720)
- Moqri M, et al. Biomarkers of aging for the identification and evaluation of longevity interventions, Cell, 2023: DOI 10.1016/j.cell.2023.08.003 (PubMed 37657418)
- Olshansky SJ, et al. Implausibility of radical life extension in humans in the twenty-first century, Nature Aging, 2024: DOI 10.1038/s43587-024-00702-3 (PubMed 39375565)
- Barnett AG, et al. Regression to the mean: what it is and how to deal with it, International Journal of Epidemiology, 2005: DOI 10.1093/ije/dyh299 (PubMed 15333621)
- Eric Topol, Ground Truths: Steve Horvath interview (August 23, 2024)
- Matt Kaeberlein, lifespan.io: When Doctors Prescribe Horoscopes (March 23, 2026)
- Scientific American: Eric Topol on biological age clocks (July 14, 2025)
- Shalev I, Apsley A, The Conversation: Biological age tests (May 4, 2026)
- Science Media Centre: Expert reaction to DO-HEALTH omega-3 study (February 3, 2025)
- University of Zurich: Omega-3s slow the aging process (2025); University of Basel: Heike Bischoff-Ferrari profile
- NBC News: What is a biological age test? (April 5, 2026); Columbia Mailman School: Daniel Belsky profile
- L-Nutra: company site (founder)
- WHO: Physical activity fact sheet; CDC: Physical activity guidelines for adults; NHS: Physical activity guidelines for adults
- CDC: Benefits of quitting smoking; CDC: About high blood pressure (page dated January 28, 2026); NHS: High blood pressure
- CDC: About sleep; CDC: Sleep in adults, FastStats
- FDA: Laboratory Developed Tests
- TruDiagnostic: TruAge Complete and TruAge; Elysium Health: Index (prices as of September 26, 2026)
Notes
Note 1: The TRIIM trial was supported by Intervene Immune, Inc., with personal donations and a supplemental grant from the Life Extension Foundation, according to the paper.
Note 2: Funded by an unrestricted grant from Metagenics, Inc.
Note 3: Funded by grant R01AG061378 to Daniel Belsky, with additional grants to individual authors, including support from the American Brain Foundation and the Canadian Institutes of Health Research.
Note 4: Funded by the USC Edna Jones chair fund to Valter Longo and NIH/NIA grants and the Yale PEPPER Center to Morgan Levine and Eileen Crimmins; the paper says the funders had no involvement in the study. L-Nutra supplied the diet.
Note 5: DO-HEALTH was funded by the European Commission’s Seventh Framework Programme (grant 278588) and, within that framework, by the University of Zurich, DNP, Roche, NESTEC, Pfizer and Streuli; the Swiss National Science Foundation funded the clock assays. The paper says the funders had no role in the design, analysis or writing.
Note 6: Funded by Morinaga Milk Industry Co., Ltd., which also provided the test foods.
Photo credits
Images from Pexels.
- featured: Photo by Anastasia Shuraeva — original Pexels photo.
- inline-1: Photo by https://kaboompics.com/ — original Pexels photo.
- inline-2: Photo by Thirdman — original Pexels photo.
Leave a Reply