Open any wearable app or longevity lab portal and you get a wall of numbers. Biological age. Recovery score. Telomere percentile. Readiness. 🧬 A few of them carry real information about whether you reach 90 in good shape. Others are a horoscope with decimal points.
We went through the evidence for each, and the sorting is less mysterious than the industry wants it to be. The metrics that earn a spot share three traits: large human cohorts, hard outcomes like death or heart attack, and numbers you can move with behavior or treatment. The ones that fail tend to be noisy, expensive, or both. 🔬
One caveat applies to everything below. Nearly all of this evidence is observational. Fitter people live longer, but nobody has run a 30-year trial that randomly assigns VO2 max and counts the funerals. So treat these as strong signals, not proven levers. Which of these do you already track? Keep that in mind as you read.
Three numbers your body hands you for free
These cost little or nothing, and the evidence behind them is the largest we found. 💪
1. Cardiorespiratory fitness (VO2 max)
Higher fitness tracks with lower death risk in almost every cohort that has looked. VO2 max is the ceiling on how much oxygen your lungs, heart, blood vessels and muscle cells can deliver and use together, so a low number can flag a weak link anywhere in that chain. 🫀 In a 2026 observational analysis of the UK Biobank (71,782 participants), higher fitness was tied to a hazard ratio of 0.78 for all-cause mortality. An NHANES III cohort of 13,968 adults found that people with diabetes in the top fitness tertile had mortality rates as low or lower than less-fit people without diabetes.
Fitter people also differ in other ways, and no randomized trial has settled cause and effect. Still, this is the one number where I would pay for a proper test. Wrist estimates are probably fine for watching a trend line, and I would treat the absolute value with suspicion.
2. Grip strength
In the PURE study, 139,691 adults aged 35 to 70 across 17 countries squeezed a Jamar dynamometer. Each 5 kg drop in grip came with a 16% higher hazard of all-cause death. ⚡ Grip was also a stronger predictor of all-cause and cardiovascular mortality than systolic blood pressure, according to the investigators. Follow-up was only about four years.
The mechanism is probably simple. Grip is a fast readout of overall muscle mass and nerve-to-muscle function, and muscle is your reserve when illness or injury hits. Two flags: PURE was partly funded by unrestricted pharma grants (AstraZeneca and Novartis among them), and the authors say it is still unknown whether raising strength lowers mortality. Do you know your grip number? Most people don’t, and it takes ten seconds.
3. Daily steps
A dose-response meta-analysis of prospective studies with device-measured steps, summarized by Epocrates, found that 7,000 steps a day was linked to 47% lower all-cause mortality than 2,000 steps (HR 0.53, 95% CI 0.46 to 0.60). 🚶 The authors flag residual confounding, which is fair: people who can walk more are often healthier to begin with. Steps are a decent proxy for how much you move at all, and that is the thing worth tracking.
What to log:
VO2 max: get one proper lab or field test, retest about once a year, and use your watch for direction in between.
Grip strength: use the same dynamometer and posture each time, and log kilograms.
Steps: watch the weekly average, because one lazy Tuesday means nothing.
Two blood tests that change decisions
4. ApoB
Every atherogenic particle in your blood (LDL, VLDL, remnants) carries exactly one apoB molecule. ApoB counts particles. LDL-C measures the cholesterol cargo inside them. 🩸 A Mendelian randomization analysis across 63 studies and 654,783 people found that once apoB was accounted for, the links between LDL-C or triglycerides and coronary heart disease dropped to nothing, while apoB kept an odds ratio of 0.761.
There is a limit. Mendelian randomization captures lifelong genetic exposure, and extrapolating that to drug treatment is problematic, as one cardiology education group puts it. Even so, the 2026 ACC/AHA guideline now backs expanded use of apoB for risk stratification beyond LDL-C. It is one add-on to a blood draw you may already be getting. 💊
5. Lipoprotein(a)
Lp(a) is an LDL-like particle set almost entirely by your genes. Diet and exercise barely move it, so one test tells you your lifetime number. The 2026 ACC/AHA dyslipidemia guideline gives universal once-in-a-lifetime testing a Class I recommendation, the strongest tier. A level of 125 nmol/L (50 mg/dL) or higher carries roughly 1.4 times the risk of heart attack or stroke, and 250 nmol/L doubles it. About 1 in 5 people carry elevated levels, yet fewer than 2% of Americans have been tested. 🧬
Here is the honest part. Pelacarsen, an Lp(a)-lowering drug, reportedly missed its primary endpoint in a September 2026 trial announcement, per a testing company’s page. A high result today means tighter control of everything else, not a proven Lp(a) fix. Insurance coverage also lags the guideline.
What to ask for at your next draw:
An apoB, alongside the standard lipid panel.
An Lp(a), once. Write the number down and you are done with it.
Your LDL-C target for your risk level, since the 2026 guideline pushes for earlier, lower targets.
A note on whether your insurer covers the add-ons, so the bill does not surprise you.
Two daily signals worth a long view
6. Sleep regularity
In 60,977 UK Biobank participants who wore accelerometers for a week, sleep regularity predicted all-cause mortality. There were 1,859 deaths over about 6.3 years. In a conference abstract from the same group, one standard deviation of better regularity (10.7 points on the Sleep Regularity Index) came with a 20.4% lower hazard. 😴
The likely mechanism is circadian. When bed and wake times wander, your body clock and your behavior drift out of step, and hormone release, blood pressure and glucose handling get mixed signals. Probably, though that part is inference. The cohort was not diverse, and irregular sleep may be a marker of poor health rather than a cause of it. It is still the cheapest metric on this list to improve, and it costs nothing.
7. Resting heart rate (the 30-day trend)
Resting heart rate (RHR) reflects autonomic tone and how efficiently your heart moves blood, and it falls as fitness rises. A 2026 analysis of the KORA cohort in Scientific Reports tested RHR and many heart rate variability indices against cardiovascular and coronary mortality, and the two behave differently. The literature is not unanimous either: that paper’s own introduction notes one study that found no RHR link at all. 🫀
A May 2026 medRxiv preprint (not peer reviewed) found that in adults aged 20 to 49, a high RHR predicted death mainly through non-cardiovascular causes, and the signal was absent before age 35. Its authors suggest looking at distress, substance use, sleep disorders and inflammation. I read that as a prompt to look at your habits, not a diagnosis.
How to use them:
Fix your wake time first, seven days a week as closely as life allows. Bedtime follows.
Check bedtime variability in your sleep app, if it reports one.
Use one device for RHR and compare 30-day averages, never single mornings.
Treat a sustained climb as a reason to review sleep, alcohol, stress and illness.
Two tests that sell certainty they can’t deliver
Ignore 1: consumer biological age scores
Epigenetic clocks are real research tools, and at population level they predict mortality. The consumer product is a different animal. A Yale team reported in Nature Aging that the same blood sample could return ages nine years apart, 50 on one run and 59 on the next. 🔬 Newer versions cut that to under a year for most tests. But a July 2026 Aging Cell analysis found that biological reliability after meals, stress, pollution or altitude change was only moderate to good (ICC roughly 0.4 to 0.7), with no clock reaching “excellent.”
The tests are also not approved medical tests, and saliva versus blood can give different answers for the same person. If you’re following this area of longevity science professionally, LongevityHub Pro gives you the structured intelligence platform to go deeper than any single article can. For everyone else, my advice is to skip the $300 kit and track the seven numbers above.
Ignore 2: telomere length
Telomere shortening is a real aging mechanism, and very short telomeres cause rare diseases like pulmonary fibrosis. 💡 That is a clinical story about the extreme tail of the distribution. Mary Armanios, who runs the telomere center at Johns Hopkins, told Science News that the qPCR method behind most consumer kits has about 20% variability between tests, compared with about 5% for the clinical flow-FISH method. She also warns that the tests can suggest something is wrong when nothing is.
Before you buy any biological age test, ask:
How much does the number change if I test twice in one week?
Is the method the same one used in the studies that predicted outcomes?
What would I do differently if the result were good, and if it were bad?
Does the company report a percentile and a range, or only a single flattering age?
Three habits of measurement that mislead
Ignore 3: the 10,000-step finish line
That figure is popular with little evidence behind it, a point the authors of the Lancet Public Health meta-analysis of 15 cohorts make in their own abstract. Their data, funded by the US CDC, show mortality risk falling until about 6,000 to 8,000 steps in adults 60 and older, and until about 8,000 to 10,000 in younger adults. 🚶 Keep the steps. Drop the ritual of pacing your kitchen at 11 p.m. to hit a round number.
Ignore 4: daily HRV and recovery swings
A study of 557 college students with Fitbits (255,736 nights) found that going to bed 30 minutes later than usual was linked to a rise of 0.18 bpm in sleeping heart rate. That is tiny. ⚡ A small UNC undergraduate study of 48 people found no link between sleep regularity and HRV at all, though it was too small to settle anything. A red “low recovery” badge after a late night is mostly your tracker noticing you had a late night. Look at the monthly trend and leave the morning drama alone.
Ignore 5: hours slept as your only sleep score
Duration still matters, so this one is a demotion rather than a dismissal. But in the UK Biobank work, sleep regularity explained about 19% of the hazard reduction versus 12.4% for duration. A summary of the full paper also notes that irregular sleep predicted cancer mortality while short sleep did not. 😴 Eight perfect hours at wildly different times is a worse deal than the tracker’s score implies.
Here is a practical next step. At your next blood draw, ask for an apoB and a one-time Lp(a). This week, squeeze a dynamometer once, set a fixed wake time, and cancel any biological age kit sitting in your cart. Which of these twelve numbers will you stop checking first?