VO2max: what it tells you (and what it doesn't)

VO2max is a real thing: the highest rate at which your body can take in and use oxygen. It travels well with endurance performance and with long-term health. It also travels less well from the lab to the finish line, and less well between two people, than most training articles suggest.

It is a predictor, not a prediction. Two people with the same VO2max can turn in very different race times — the missing variable is how much oxygen a given pace costs each of them.
The number in your watch is the least direct version. On average across fourteen validation studies the good algorithms sit close to the lab — but on you, personally, they can be off by nearly ten units either way.
It drops by about ten percent a decade in the average adult — and yes, that can be pushed back, though the field's training literature is thick on eight-week programmes for young men and thin on everything else.
Higher fitness travels with lower mortality — an association from thirty-three cohort studies, not a promise that raising your own number lowers your own risk.

Prefer to watch? The whole explainer in 9:50, narrated — it has its own page too, with chapters and a full transcript. Subscribe on YouTube if you want to hear about the next one.

Almost every article about training performance eventually points at this one number. It's shorthand for aerobic fitness, it comes with a satisfying decimal, and most modern watches will tell you yours. What it means, and how far its reach actually goes, is more interesting than the number itself.

The plan for this page: what VO2max is trying to measure, three different ways of arriving at it, what it says (and does not say) about your race time, whether you can raise it, how it drifts with age, and the long-term-health tail you may have read about.

1 What the number is trying to measure

VO2max is the highest rate at which your body can take in oxygen, deliver it to the muscles, and use it — the top of the aerobic engine.

It is written in millilitres of oxygen, per kilogram of body weight, per minute (mL/kg/min). The division by body weight is what makes the number comparable between a seventy-kilogram runner and a ninety-kilogram one: both bodies can absorb oxygen, but they also have different amounts of tissue to move. The clearest place this division earns its keep is on a climb — carrying an extra ten kilos up an alpine col costs about as much extra power as you'd expect, and a per-kilogram number keeps the accounting hones (figure 2)t. (If you want to see that trade-off in watts, our climb-time calculator is built out of exactly it.)

A diagram of the unit as a division. On the left, a pair of lungs with an arrow to a single muscle, labelled in through the lungs and used in the muscles, with the caption millilitres of oxygen per minute. A large division sign in the middle. On the right, two standing figures of the same height, one slim and one heavier, labelled 70 kg and 90 kg. — **Figure 2.** **The unit is a division, and both halves matter.** The top half is a rate: oxygen taken in through the lungs and used in the muscles, every minute. The bottom half is you. Dividing one by the other is what lets a 70 kg runner and a 90 kg one be compared at all, and it is why the same litres of oxygen produce a different number on two different bodies.

The number is a real physiological measurement, and it correlates with endurance performance and with long-term health. But it is an association with performance, not performance itself (figure 3) — a sentence worth repeating because most watch-generated charts nudge you the other way.

Diagram: two runners share the same VO2max of about 55 mL/kg/min, but on a 10K race day the more economical runner finishes in 38:20 and the less economical one in 42:05 — a difference of nearly four minutes. — **Figure 3.** **The same top of the engine, two different finish lines.** Aerobic capacity tells you how much oxygen your body can move. Running economy tells you how much oxygen a given pace costs you. The second number is doing much of the work the first one cannot see. Schematic; the principle, not measured data.

This is why a race-time predictor built from VO2max alone is a leaky machine. The pace you can actually hold depends on your ceiling, on how efficiently you turn oxygen into motion, on how long you can hold a given fraction of that ceiling, and — on the day — on the weather, your fuelling and your legs. The number in your watch answers one of those four questions.

2 How you get to the number

There are three routes to a VO2max value. They arrive at numbers with the same units and very different meanings, and knowing which one you have is half of reading the number well.

The reference version is a laboratory cardiopulmonary exercise test (CPX) — a treadmill or bike test to voluntary exhaustion while a mask analyses every exhalation. It has been the standard since the 1920s. The very first values were measured on an 84-metre grass track in England, into a Douglas bag — literally a canvas bag that collected the exhaled air for later analysis. Modern kit reads breath by breath instead of emptying a bag, but the question it answers is the same one that A.V. Hill and Hartley Lupton first asked in 1923 (figure 4).⁶

Two diagrams side by side. On the left, a 1923 arrangement: a person wearing a canvas bag on the back, breathing through a mouthpiece with two one-way valves, so that room air goes in and every exhaled breath goes into the sealed bag. On the right, the modern equivalent: a person in a mask connected by a hose to a trolley with a gas analyser and a screen, reading breath by breath. — **Figure 4.** **A century apart, the same question.** Both arrangements collect the air you breathe out and analyse what is in it. The bag was sealed and measured afterwards; modern kit reads each breath as it happens. This is the only route that measures oxygen rather than inferring it, which is why it is the one the other two are judged against.

The submaximal field test lets you skip the exhaustion. You do a shorter effort at a known power or pace, the watch or bike computer records heart rate, and the number is inferred using your age, weight and an estimated maximum heart rate. Sport physiology has used variants of this since the 1950s — the Åstrand-Ryhming step test and the cycle-ergometer variants of it — and cardiology uses much the same logic today (figure 5).

A cyclist wearing a heart-rate strap on the chest and riding a bike fitted with a power meter, next to a three-scale chart: a power scale on the left, a heart-rate scale on the right, and a slanted middle scale where the VO2max value is read off. A straight line drawn between a power value and a heart-rate value crosses the middle scale at the estimate. — **Figure 5.** **No gas is measured, so the number is worked out backwards.** You ride at a known power until your heart rate settles, and the two together are read off against a chart. It stops well short of exhaustion, which is the whole appeal. The original Åstrand-Ryhming nomogram is under copyright, so this is our own drawing of the principle rather than a reproduction of it.

The wearable estimate is what most people mean when they say "my VO2max". Newer watches run a version of the submaximal trick automatically: after a run where your heart rate goes above about seventy percent of its estimated maximum for at least ten minutes, the watch combines heart rate, pace and your entered profile (age, sex, weight) into a number. It does not measure oxygen. It fills in the gaps from who you told it you are (figure 6).

Figure 6. Three inputs, and none of them is your breath. Heart rate, pace and the profile you typed in are combined into a number after a run that was hard enough for long enough. Nothing in that chain measures the gas you breathe, which is what separates this rung from the first one.

Fourteen studies pooled by an international consortium have looked at how well those estimates land against a real lab test.³ On average the exercise-based algorithms barely miss (a bias of about −0.1 mL/kg/min); on average the resting-based ones sit about two units too high. The catch is not the average — it is the range around it (figure 7).

A comparison of two algorithm families for estimating VO2max from a wearable, showing the pooled bias and the 95% limits of agreement around it. Exercise-based algorithms sit almost on zero with a range from about minus ten to plus ten. Resting-based algorithms sit at plus two with a range from about minus thirteen to plus seventeen. — **Figure 7.** **A watch is roughly right about a crowd, roughly right or wrong about you.** Bias is the average error across the study; the band is the range that 95% of individual estimates fall in. Even the good algorithm is nearly ±10 mL/kg/min wide on one person. From Molina-Garcia 2022 (INTERLIVE meta, 14 studies).

One review's worth: the meta-analysis conclusion is that exercise-based wearables can support population-level assessment but not clinical individual assessment. That is the sentence to keep. For a study of a thousand people the errors average out. For your own reading, a band of nearly twenty units around the answer is exactly the number you cannot ignore.

3 VO2max is not your threshold

These two get confused constantly, and they measure different things.

VO2max is the ceiling. Your lactate threshold — the pace or power you can hold before fatigue builds sharply — is the fraction of that ceiling you can sustain for a while, and it is what actually decides your race pace. Two runners can share a VO2max and hold very different fractions of it (figure 8). That is a whole piece in itself; we have flagged it in the queue for a full explainer, called Thresholds, explained like a human being.

Two runners, each next to a tall outlined box. The box for runner A is filled almost to the top; the box for runner B is filled to about two thirds. Labels point out that the whole box is the VO2max ceiling and the shaded part is the lactate threshold, the fraction that can actually be held. — **Figure 8.** **The ceiling is one number; the part you can hold is another.** Both runners have the same roof. What differs is how much of it they can sit under before fatigue starts winning, and that is what sets race pace. The two fractions drawn here are illustrative rather than measured.

4 Can you raise it, and does it matter?

Short answers: yes, and it depends who you are and what you want.

A scoping review published in 2026 mapped six hundred and seventeen studies of aerobic training and VO2max.² That is the good news — a lot of people have looked. The less-good news is that the map is very lopsided (figure 9).

A summary of 617 aerobic-training studies: moderate continuous training (MICT) accounts for 363 of them (58.9%), high-intensity intervals for 102 (16.5%), sprint intervals for 70 (11.3%) and combined modalities for 82 (13.3%). Of the 586 that reported sex, 264 (45.1%) enrolled only men, 239 (40.7%) were mixed, and 83 (14.2%) were only women. Most interventions were between one and three months long. — **Figure 9.** **Six hundred training studies, most of them short and male.** The literature is thick on eight-week programmes for young, healthy men — and thin on almost everything else. From Wu et al. 2026, a scoping review of the aerobic-training literature.

Most studies use moderate continuous training; fewer use intervals or sprints. Two in every three that reported sex enrolled only men or a male-biased mix; only about one in seven was female-only. And most interventions ran three months or less — long-term studies (a year or more) exist only in the moderate-continuous group, and only nine of them in the entire sample. That does not mean training does not work. It means the picture of how much the average person can raise VO2max, and over how long, is drawn on a narrow slice of the possible sample.

The practical version: consistent aerobic training raises VO2max, and higher intensities raise it faster for a given time invested. How much you can raise it depends on how much headroom you started with — beginners see the biggest jumps, and closer to your ceiling the gains get harder to find and easier to lose. If you want the how-a-training-study-works reasoning behind that sentence, we wrote it up separately in how to read a training study.

5 Whose number is this for?

Both the elite runner and the recreational one — but the reasons are different, and it is worth being clear about which one applies to you.

For the sharp end of the sport, VO2max is a physiological ceiling that matters because everything else has to be paid from under it. Elite endurance is built on very high values, and small differences at that level track with real differences in race outcomes. That is what most textbook descriptions of "VO2max determines performance" are actually about.

For a recreational runner or cyclist, VO2max is a fitness indicator — where you sit against the population, whether the last training block moved you, and what the long-term trend looks like as you age. That is a legitimate use of the number, and a watch estimate is a reasonable tool for it (with the ten-unit health warning attached) (figure 10). What that same number is not built for is telling you your goal race pace. For that, threshold and running economy do most of the work.

Two panels side by side. The left, headed the sharp end, shows a lean competitive runner at racing speed with three lines: a physiological ceiling, everything is paid from under it, small differences decide races. The right, headed the rest of us, shows an ordinary runner jogging with three lines: a fitness indicator, where you sit and whether you moved, a watch estimate is fine for that. — **Figure 10.** **Same number, two different jobs.** At the sharp end the number is a ceiling that everything else has to be paid from under. For everyone else it is an indicator of where you sit and whether the last block moved you — a job a watch estimate can do. Neither reading gives you a goal race pace.

6 How the number changes with age

VO2max drops as you get older, in a pattern that is remarkably regular across large populations and much more variable in any one person.

The most-cited reference dataset in the world is FRIEND — the Fitness Registry and the Importance of Exercise: A National Database, built from 7,783 maximal treadmill tests across eight US labs and published in 2015.⁴ Its median values for men drop from 48.0 mL/kg/min in the twenties to 24.4 in the seventies; for women, from 37.6 to 18.3 across the same span. The rate of decline is about ten percent per decade in both sexes (figure 11).

Line chart of the 50th-percentile VO2max in the FRIEND reference dataset by ten-year age band, from the twenties to the seventies. Men drop from 48.0 mL/kg/min to 24.4, women from 37.6 to 18.3, in both sexes about ten percent per decade. — **Figure 11.** **What the middle of the population looks like at each age.** From Kaminsky, Arena and Myers (2015), FRIEND registry, table 3. Above the line is above-average for your age; below it, below.

The trajectory is flatter in people who keep training. A study published in 2026 looked at 340 multi-marathoners from 24 countries, with an average of 121 lifetime marathon completions and an average age of 52.¹ Their estimated VO2max was well above the FRIEND median in every age band, and the drop from decade to decade was more gradual (figure 12).

A comparison of two age curves. The solid line is the FRIEND reference median for men, dropping steadily from 48 to 24 across five decades. The dashed line sits higher and drops less steeply — a schematic of the flatter pattern reported in a survey of multi-marathoners. — **Figure 12.** **Both curves drop, one drops less steeply.** The reference line is lab-measured. The upper pattern is wearable-estimated in a very self-selected group, so read it as a shape rather than as tabled values. That paper is a nice example of the whole page, actually: the value is estimated, the sample is self-reported, and the health benefit is modelled — three of the ceilings this page discusses, in one place.

Two things follow. Consistent training does not stop the clock, but it appears to slow it down; and any single reading — yours or your neighbour's — sits somewhere on a wide distribution around the average, so a number below or above the median at your age is information, not a judgement.

7 Which method for which question?

Most people asking about their VO2max are asking one of three different questions, and the honest answer depends on which one (figure 13).

A guidance table comparing the three ways of arriving at a VO2max. A lab CPX earns five stars, suits an athlete targeting a big race, and costs a half-day and 150 to 300 euros. A submaximal field test earns three stars, suits a committed amateur tracking a season, and costs nothing beyond your own bike or track. A wearable estimate earns two stars, suits anyone with a running watch, and costs whatever the watch cost. — **Figure 13.** **The right method is the one that answers your question.** Almost nobody needs a lab test; almost everybody has a reason to look at the population-level number their watch already produces. The stars are our own reading of the accuracy each route can deliver for a typical amateur; the lab-test cost is the range that a European sports medicine clinic will typically quote.

The submaximal field test is under-used. It is free, it is repeatable, and if you do it on the same course under the same conditions you get a consistent-enough baseline to see whether your training is moving you — which is more actionable than the absolute number either way.

8 The mortality tail

You may have read that VO2max predicts how long you live. Here is what that actually says, and what it does not.

A large meta-analysis pooled 33 cohort studies and about 103,000 healthy adults, with 6,910 deaths recorded over follow-up.⁵ Each 1-MET higher aerobic capacity was associated with a relative risk of all-cause mortality of 0.87 — a 13 percent lower rate (figure 14). Because 1 MET equals 3.5 mL/kg/min, that works out at about 3.7 percent per single mL/kg/min, and it is the number the multi-marathoner paper leans on to model the mortality benefit of a higher VO2max.

A curve showing relative all-cause mortality risk against extra METs of aerobic capacity, starting at 1.0 at baseline and dropping to 0.87, 0.76, 0.66, 0.57 and 0.50 with each additional MET — a 13% lower risk per MET, or about 3.7% per mL/kg/min. — **Figure 14.** **Higher fitness travels with lower mortality.** From Kodama et al. (2009), pooling 33 cohort studies. This is an association from observing large groups over time — not a promise that raising your own VO2max lowers your own risk.

Two footnotes make that number honest. The first: these are observational cohorts, not trials. Fitter people are also, on average, thinner, less likely to smoke, and more likely to see a doctor early. Statistical methods can adjust for those things but not perfectly. The second: an association from a population does not tell you what happens when one person changes their number. Both directions are plausible — fitness may cause longer life, longer life may follow from whatever else lets you stay fit — and the truth is probably some of both.

None of which makes the association a mirage. Higher aerobic fitness is one of the best-corroborated markers of long-term health we have, and the number in your watch is a real (if noisy) proxy for it. It is worth knowing. It is not a medical destiny.

9 Seven things to remember

If you close this tab and see the number in your watch tomorrow morning, these are the things that were probably not on the same screen as it.

It is a real physiological measurement — of a ceiling, not of a race time. Aerobic capacity and pace at threshold are different quantities.
The number in your watch is inferred. The lab measures the gas you breathe; the watch fills in the gaps from your age, weight and heart rate.
On average, exercise-based watch estimates are close. On you, personally, they can be off by nearly ten units either way — a 95% range of about 20 units around the answer.
Two runners with the same VO2max can turn in very different race times. The missing variable is running economy — how much oxygen a given pace costs each of them.
Yes, you can raise it. Consistent aerobic training does the work; the literature is thick on eight-week programmes for young men and thin on almost everything else, which is worth knowing when you read what "the research says".
It drops about 10% per decade on average — and less steeply for people who keep training. Above or below the median at your age is information, not a judgement.
Higher VO2max travels with lower mortality — 13% per MET, or ~3.7% per single mL/kg/min, from a large observational meta-analysis. Association, not causation.

The number is a lens on your aerobic engine. It is not a photograph of your race day, and it is not a promise about your ninetieth birthday. Both of those depend on what you do with it.

Sources

This is an explainer rather than a review: the six sources below were picked to cover each part of the story — measurement, wearables, trainability, age norms, the mortality association, and the historical origin. All six were read in full at their own source.

T1 Lundy L, Reilly RB, Fleming N. VO2max ageing and all cause mortality in a global cohort of multi marathoners. Scientific Reports. 2026;16:21761. doi:10.1038/s41598-026-52475-x — read at source; the sample of 340 multi-marathoners across 24 countries with an average of 121 marathon finishes and mean age 52.2 comes directly from the abstract and Methods.
T1 Wu Z, Preobrazenski N, Renwick JRM, et al. Aerobic Exercise Training and VO2max: A Scoping Review of Study Populations and Protocols. Journal of Functional Morphology and Kinesiology. 2026;11(1):70. doi:10.3390/jfmk11010070 — read at source; all counts (617 studies; MICT 363, HIIT 102, SIT 70, multiple 82; male-only 264, female-only 83 of 586 with sex reported; interventions shorter than three months n=399) come from the Results.
T1 Molina-Garcia P, Notbohm HL, Schumann M, et al. Validity of Estimating VO2max by Consumer Wearables: A Systematic Review with Meta-analysis and Expert Statement of the INTERLIVE Network. Sports Medicine. 2022;52(7):1577–1597. doi:10.1007/s40279-021-01639-y — read at source; the pooled bias and 95% limits of agreement for exercise-based (-0.09; -9.92 to +9.74) and resting-based (2.17; -13.07 to +17.41) algorithms come from the meta-analytic Results.
T1 Kaminsky LA, Arena R, Myers J. Reference Standards for Cardiorespiratory Fitness Measured with Cardiopulmonary Exercise Testing Using Cycle Ergometry: Data From the FRIEND Registry. Mayo Clinic Proceedings. 2015;90(11):1515–1523. doi:10.1016/j.mayocp.2015.07.026 — read at source; the 50th-percentile values (men 48.0 → 24.4 mL/kg/min; women 37.6 → 18.3) and the ~10% per decade decline are from the abstract and Table 3.
T1 Kodama S, Saito K, Tanaka S, et al. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: a meta-analysis. JAMA. 2009;301(19):2024–2035. doi:10.1001/jama.2009.681. Read in full. The pooled effect we cite: each 1 MET of extra aerobic capacity came with a 13% lower rate of death from any cause, and the range the study is confident about runs from 16% to 10% lower. Across 33 cohort studies, 102,980 participants and 6,910 deaths.
T1 Robergs RA, O'Malley B, Torrens S, et al. Lessons from history for VO2max and the VO2 plateau, part 1, 1920–1961. Frontiers in Physiology. 2025;16:1688750. doi:10.3389/fphys.2025.1688750 — read at source; the Hill and Lupton 1923–1924 origin, the 84-metre grass-track measurements and the Douglas-bag method are from the historical review.

Found a mistake here? Tell us at info@enduranceproof.com — you do not need to be a scientist, and "this number looks off" is a perfectly good message. Anything we correct, and when, goes on our corrections page.

This is educational material about a physiological measurement, not training, medical or nutrition advice. Nothing here is a recommendation about what any individual should do.

VO2max: what it tells you, and what it doesn't