Longevity biomarkers, ranked by how well they actually predict death

If you could run only one test to estimate how long someone will live, most people would reach for a blood panel — cholesterol, glucose, maybe blood pressure. It’s the intuitive move: bad blood, bad outcome. Turns out that’s not where the strongest signal lives. The numbers that best separate the people who die from the people who don’t over the next decade aren’t in a vial. They’re in how fast you walk, how hard you can squeeze, and how much air your body can move.

Here’s the thing the longevity-content firehose rarely says out loud: these biomarkers are not equal. Some of them carry enormous predictive weight, and some of the ones people obsess over on their wearables are, frankly, noise dressed up as insight. So I did the boring, nerdy thing and lined them up by the one axis that matters — how well each actually predicts all-cause mortality in the published cohorts — and the ranking is genuinely not what most people expect.

The thesis, up front: functional fitness measures — VO2 max, grip strength, walking pace — predict who dies far more powerfully than the blood markers and gadget metrics that get all the attention. The best predictor is also the hardest to measure, the cheapest test outperforms an expensive one, and the muscle-mass number everyone chases matters less than whether that muscle can actually do anything.

How I ranked them

One criterion: discriminatory power for all-cause mortality — how much the risk of dying changes as the metric moves, in large prospective cohorts, adjusted for the obvious confounders. Not “is it a real risk factor” (plenty are), and not “can you treat it” (a separate, important question I’ll come back to). Purely: if I only knew this one number about you, how well could I guess your odds?

A quick caution before the list, because it’s the honest part. High predictive power is not the same as a causal lever you can yank. Some of these numbers are low because a person is already ill — reverse causation — so they’re partly a symptom, not just a cause. I’ll flag where that bites. And “predicts mortality” in a population is not a personal prophecy; it’s a probability across thousands of people, not a countdown clock for you.

Right, the ranking.

1. VO2 max (cardiorespiratory fitness) — the strongest, and it’s not close

The 2018 JAMA Network Open study by Mandsager and colleagues put 122,007 patients on a treadmill and followed them. The least-fit group had roughly five times the mortality risk of the elite-fit group — and, unusually, there was no upper limit of benefit. Fitter was better all the way up. That spread is larger than the mortality gap from smoking, diabetes or hypertension. No other single metric on this list separates the living from the dying as cleanly.

The catch is measurement. A true VO2 max needs a mask and a graded test to exhaustion. Your watch’s estimate is a rough proxy and a submaximal test gets you a usable trend, but the best predictor is annoyingly the least convenient to pin down. Worth it anyway — this is why VO2 max is the strongest predictor of longevity, and it’s the one number I’d fight to keep if I could only track one.

2. Grip strength — the cheap test that beats blood pressure

The PURE study (Leong et al., Lancet 2015) measured grip in ~139,000 people across 17 countries. Each 5 kg drop in grip strength was associated with a 16% higher risk of all-cause death — and here’s the kicker, grip predicted cardiovascular death better than systolic blood pressure did. A €20 hand dynamometer out-forecasting the cuff at your GP’s office is exactly the kind of thing that makes wellness culture sheepish.

Grip isn’t magic; it’s a proxy. It stands in for whole-body strength, neuromuscular integrity and how well you’ve held onto muscle as you’ve aged. That’s why grip strength says so much about how you’re aging despite being, on paper, a slightly absurd thing to measure.

3. Walking pace — the one you can measure without noticing

Ask people to report their usual walking pace — slow, steady, brisk — and that self-reported answer turns out to be one of the strongest predictors of five-year mortality in the UK Biobank (Ganna & Ingelsson, Lancet 2015). Measured gait speed does the same job: a pooled analysis of 34,485 older adults (Studenski et al., JAMA 2011) found survival rose steadily with walking speed across every age band. There’s even a study cheekily titled around whether you walk faster than the Grim Reaper (Stanaway et al., BMJ 2011) — the fast walkers, reassuringly, do.

Gait speed is an integrated readout: it needs a functioning heart, lungs, legs, nerves and brain all at once, so when it drops, something upstream is usually giving way. And you can measure it for free. Full detail on how fast you walk and how long you’ll live.

4. Heart-rate recovery — how fast you come back down

After you stop exercising, how many beats your heart rate falls in the first minute is a window onto your autonomic nervous system. In the classic Cole et al. study (NEJM 1999), an abnormal recovery — a drop of 12 beats or fewer at one minute — roughly doubled the risk of death over six years after adjusting for age, medications and standard risk factors (the raw, unadjusted gap was closer to fourfold). It’s measurable with any heart-rate strap and a stopwatch, which puts it firmly in the “no excuses” tier for accessibility (Tessa’s note: I hate that phrase, but the test genuinely is that easy). More on heart-rate recovery, the post-workout number that predicts survival.

5. The one-leg stand — a free test that embarrasses expensive ones

Can you balance on one leg for 10 seconds, phone down, no wall? In Araújo et al. (Br J Sports Med 2022), middle-aged and older adults who couldn’t had an 84% higher risk of all-cause death over the following ~7 years (hazard ratio 1.84). A 10-second test, no equipment, out-predicting a fair few things you’d pay for. Balance is another integrated measure — it folds in strength, proprioception, vestibular function and neural speed. Here’s what standing on one leg says about how you age.

6. Resting heart rate — useful, but softer

A high resting heart rate travels with higher mortality — across cohorts, the extremes (say, above 90 vs below 50 bpm) differ by roughly two- to threefold in risk. It’s continuous, trivially measurable on any wearable, and moves in the right direction with fitness. But the effect size per beat is modest and it’s easily perturbed by caffeine, stress, sleep and a bad night’s data, so as a standalone predictor it sits mid-table. Good for trends, weak for one-shot fortune-telling — what your resting heart rate actually means.

7. Muscle strength over muscle mass — and this is the surprising one

Everyone chasing a DEXA muscle-mass number, look away now. In the Health ABC study, muscle strength predicted mortality in older adults, but muscle mass and cross-sectional area largely did not once strength was accounted for. A large muscle you can’t generate force with doesn’t buy you much. This is exactly why grip (a strength measure) sits at #2 while raw mass lands down here. Mass still matters — it’s metabolic reserve and a buffer against frailty — but as a predictor, function wins. The nuance is worth the read: muscle mass, the organ you’re not tracking, with the strength caveat attached.

8. Heart-rate variability — the trendiest, and the noisiest

HRV has the biggest gap between hype and predictive weight. Lower HRV does track with higher mortality (Tsuji et al., Circulation 1994, from Framingham), on the order of ~1.5x per standard deviation. But it’s enormously sensitive to measurement conditions, position, breathing, alcohol, illness and which device you used — the day-to-day noise often swamps the between-person signal. Brilliant for tracking your own recovery trend, genuinely weak for ranking one person against another. It’s on your dashboard because it’s easy to capture, not because it’s the sharpest predictor.

The blood panel, and why it’s not on the podium

Notice what’s missing from the top: cholesterol, blood pressure, glucose. Not because they don’t matter — they’re causal, modifiable risk factors and you should absolutely manage them. But as standalone discriminators of who dies across a whole population, each one individually is weaker than whole-body fitness. Blood pressure is a single organ system’s pressure gauge; VO2 max and gait speed are the entire machine reporting in at once. Treat the blood markers; just don’t mistake a normal panel for proof you’re aging well. (For the bigger picture of what “aging well” even means as a number, see what biological age actually is.)

The catch, stated plainly

Here’s where I stop selling. Every ranking above is about prediction, and prediction hides two traps.

Reverse causation. The sickest people have low VO2 max, weak grip and slow gait partly because they’re ill. Some of the predictive power is the metric acting as a symptom, not a cause. That doesn’t make the numbers useless — a symptom that shows up years early is exactly what you want — but it means “raise the number” isn’t automatically “add years.”

Gaming the test buys nothing. Predictive power lives in what the number represents — integrated organ reserve. Practise the one-leg stand until you pass and you’ve improved your balance-test score, not necessarily your survival. But — and this is the part that redeems the whole exercise — the training that genuinely moves these markers is the same training that lowers real mortality risk. Build the aerobic base and the intervals that raise VO2 max, the strength work that raises grip, the fitness that speeds your gait, and you’re not gaming anything. You’re doing the actual thing. Which is why the four pillars of longevity training point at exactly the levers this ranking rewards.

How to actually use this

Don’t track all eight. Pick a small basket weighted toward the top of the list and toward what you’ll realistically measure:

• Free, no kit, do it this week: usual walking pace (honest self-assessment), the 10-second one-leg stand, and — if you’ve a heart-rate strap — one-minute heart-rate recovery after a hard effort.
• Cheap, worth it: a ~€20 hand dynamometer for grip. Best predictive bang for the fewest euros on this list.
• The big one, worth the hassle: a VO2 max estimate. Watch figure for the trend, a proper submaximal or lab test once a year if you can.
• Track for your own trend, not for comparison: resting heart rate and HRV. Useful as a personal baseline, weak as a verdict.

The uncomfortable, freeing takeaway is that the metrics with the most predictive weight are mostly the cheap, physical, slightly undignified ones — squeeze hard, walk fast, stand on one leg, get your breath back quickly. Genuinely more informative than most of what a wearable will upsell you. Right, so: fewer dashboards, more of the four or five numbers that actually move the needle.