Biological Age Testing: What Epigenetic Clocks Measure, What They Predict, and What They Cannot
Medically reviewed by Dr. Mia Chorney, DNP, FNP-BC, MSCP — Cofounder & Chief Medical Officer, Amsara Health
Your driver’s license says 52. A blood test says 47, or 58. Which one is right? Both, in a sense, and the gap between them is the most useful number in longevity medicine right now, provided you understand what it measures and how much noise it carries. Biological age testing has moved from research curiosity to something you can order online for a few hundred dollars, and the marketing has outrun the science. Here is the science.
What is an epigenetic clock?
Your DNA sequence does not change with age, but the chemical marks on top of it do. Methyl groups attach to specific sites on the genome and switch genes up or down; the pattern shifts in predictable ways as tissues age. In 2013, Steve Horvath at UCLA showed that reading 353 of those sites could predict a person’s calendar age to within about three and a half years across almost every tissue in the body. That was the first epigenetic clock.
The clocks that matter clinically came next. PhenoAge (2018) was trained not on birthdays but on a composite of clinical biomarkers, so it predicts health rather than time. GrimAge (2019) reads methylation signatures of plasma proteins and smoking exposure, and is the strongest predictor of mortality yet built. DunedinPACE (2022) measures something different again: not how old you are, but how fast you are aging, expressed as biological years per calendar year, where 1.0 is average and 0.9 means you are aging about 10 percent slower than your peers. Second-generation and pace clocks are what a serious practice uses. First-generation clocks, including many of the cheaper consumer tests, predict birthdays well and outcomes poorly.
Does biological age predict anything real?
Yes, and the effect sizes are not small. In a seventeen-year follow-up of 2,105 US adults, every five years of GrimAge acceleration carried a 51 percent higher risk of death from any cause and a 61 percent higher risk of cardiovascular death. In the Generation Scotland cohort of 18,859 people, GrimAge predicted mortality and DunedinPACE predicted new diabetes, with first-generation clocks roughly half as strong. In the Framingham Offspring study, each standard deviation of faster DunedinPACE meant a 39 percent higher risk of cardiovascular disease and a 65 percent higher risk of death. Among 6,069 postmenopausal women followed for nine years, faster GrimAge was associated with incident mild cognitive impairment and dementia.
The clocks are also, notably, sensitive to menopause. Morgan Levine and colleagues showed in 2016 that earlier menopause and surgical removal of the ovaries were associated with accelerated epigenetic aging in blood, with genetic evidence pointing toward causation rather than coincidence, and that women using hormone therapy had younger epigenetic age in cheek tissue. This is one reason we consider biological age a women’s health measurement, not only a longevity one.
Can you change your biological age?
In the direction you would hope, yes, though the honest word is “modestly” and the honest timeframe is “years, not weeks.” The best evidence is the CALERIE trial, a two-year randomized study in which a roughly 12 percent calorie reduction slowed DunedinPACE by 2 to 3 percent, while PhenoAge and GrimAge did not move. A six-month cycling program that raised VO2max 20 percent lowered GrimAge by about seven months relative to expectation, and the change tracked fitness gains. Quitting smoking reverses GrimAge acceleration steadily over the first decade. A 2026 meta-analysis of 51 intervention studies found DunedinPACE and the newer principal-component version of GrimAge the most responsive clocks, with first-generation clocks responding sporadically.
You will also see a widely shared 2021 pilot claiming a three-year reversal in eight weeks from diet and lifestyle. It enrolled 43 men, used a first-generation clock, and has since been corrected twice. Treat it as a hypothesis, not a promise.
How accurate is a single biological age result?
Less than the number on the report implies, and this is the fact most consumer tests bury. When the same blood sample is run twice, first-generation clocks can differ by as much as nine years; newer principal-component clocks reduce that to roughly a year and a half. DunedinPACE has excellent test-retest reliability. No clock is standardized across laboratories, none is an FDA-cleared diagnostic, and none is yet an accepted clinical endpoint. Horvath himself has said the results should be read only alongside blood pressure, glucose, and lipids. That is not a reason to skip the test. It is a reason to buy the right one and read it correctly.
What about telomere testing?
Telomeres, the protective caps on chromosomes, shorten with age, and telomere length was the original biological age biomarker. It remains a legitimate measure of cellular aging, and we offer it. It is also the noisier one: telomere length is a weaker and less consistent predictor of mortality than GrimAge or DunedinPACE, and results shift measurably with laboratory method. We think of it as a supporting dimension rather than the headline number.
How does Amsara use biological age testing?
As a benchmark and a scoreboard, in that order. Your first result, drawn with your comprehensive labs, establishes where your biology stands relative to your calendar. We read it next to the markers that actually drive the number: ApoB and lipoprotein(a), HbA1c and fasting insulin, hs-CRP, blood pressure, body composition, and, when you are ready, VO2max. Then we build the plan, which almost always centers on the unglamorous levers the trials keep confirming: aerobic fitness, muscle, protein, sleep, alcohol, and, for many women, hormone therapy that is right for them. We repeat the test, from the same laboratory using the same sample type, after enough time has passed for real change to exceed the measurement noise, typically a year. A one-year move of a few tenths on DunedinPACE is meaningful. A three-month move is mostly weather.
The test is on our Tests & Services menu, ordered by your provider and reviewed with you. If longevity is your goal, it belongs in the same conversation as multi-cancer screening and functional genomics: three different windows into the same body, none of which should be read alone.
Frequently asked questions
Is my biological age result accurate to the year?
No. Depending on the clock, technical noise ranges from roughly a year and a half to as much as nine years on the same sample. Second-generation and pace-of-aging clocks such as PC-GrimAge and DunedinPACE are the most reliable; first-generation consumer clocks are the least.
If my biological age is higher than my calendar age, does that mean I will die sooner?
It means your risk is higher than average, not that your outcome is fixed. Every five years of GrimAge acceleration is associated with about a 50 percent higher risk of death in long-term studies, and the same studies show the number responds to fitness, smoking cessation, and diet.
Can I lower my biological age, and how fast?
Yes, modestly. Calorie restriction slowed pace of aging by 2 to 3 percent over two years in a randomized trial; six months of aerobic training lowered GrimAge by about seven months. Expect meaningful change over a year, not eight weeks.
Did menopause make me age faster, and does hormone therapy help?
Earlier menopause and surgical menopause are associated with accelerated epigenetic aging in blood. Hormone therapy users showed younger epigenetic age in some tissues in the 2016 study; the evidence is suggestive rather than settled, and it is one factor your provider weighs when discussing hormone therapy.
Does it matter whether the test uses blood or saliva?
Yes. Results differ by tissue, so baseline and repeat tests must use the same sample type from the same laboratory. Amsara draws biological age testing with your comprehensive blood work.
Is biological age testing FDA approved or covered by insurance?
No. Epigenetic clocks are not FDA-cleared diagnostics and are not covered by insurance. Amsara members receive 20 percent off non-insurable testing; current pricing is on the Tests & Services page.
Should I get telomere testing instead?
Methylation clocks are stronger predictors of health outcomes than telomere length, which varies more with laboratory method. Telomere testing is a reasonable supporting measure, not a substitute.
How often should I retest, and what counts as real change?
Roughly once a year, from the same lab and sample type. A change larger than the clock’s known technical noise, about a year and a half for principal-component clocks, is meaningful; smaller moves over shorter periods usually are not.
This article is for education only and is not medical advice. Talk with a qualified clinician about your individual health before starting or stopping any therapy.
Curious what your biology says about your age?
Unhurried appointments, advanced diagnostics, and a plan built for you.
Already a patient? Visit your Amsara Patient Portal →
