VO2 Max Explained: What It Measures and How Much of It You Can Change

VO2 max is the largest volume of oxygen your body can take in, transport and use per minute during progressively harder exercise. It is expressed in millilitres of oxygen per kilogram of body weight per minute, and it is the closest thing physiology has to a single number for aerobic capacity.

Two things about it are true at once, and holding both is the whole trick. It is among the strongest predictors of all-cause mortality ever measured in large cohorts, stronger in some analyses than smoking or diabetes. And it is substantially genetic, to the point where two people completing the identical training programme can post improvements that differ by an order of magnitude. What follows is what the number actually describes, what the population data look like, how much of it responds to work, and what to do about the part that does not.

What the measurement describes

The concept dates to Hill and Lupton, writing in the Quarterly Journal of Medicine (16:135-171, 1923), who observed that as running speed increased, oxygen uptake rose and then flattened while the runner kept accelerating. That plateau is the definitional feature. Above it, additional work is paid for anaerobically, which is why it cannot last.

The physiological ceiling is described by the Fick principle: oxygen consumption equals cardiac output multiplied by the arteriovenous oxygen difference. Cardiac output is heart rate times stroke volume. The arteriovenous difference is how much oxygen the working muscle strips out of the blood passing through it.

That equation is useful because it shows you which terms training can move. Maximum heart rate is essentially fixed and drifts down with age, as the trouble with age-predicted maximum heart rate sets out. Stroke volume, by contrast, is highly trainable: endurance training expands plasma volume, increases left ventricular chamber size and improves ventricular filling, so each beat ejects more blood. On the peripheral side, capillary density around each muscle fibre increases, mitochondrial content rises, and the muscle becomes better at extracting and using what arrives.

Which of those two ends actually limits the number has been argued for decades. Levine's review in the Journal of Physiology (586:25-34, 2008) makes the case that in healthy people at sea level, oxygen delivery (cardiac output) is the binding constraint rather than the muscle's capacity to consume. This matters practically: it is why VO2 max responds to training that stresses the heart at high stroke volumes, and why it improves less from work that fatigues muscle without loading the central circulation.

The Dallas Bed Rest and Training Study gives the clearest illustration of how plastic the system is. Five young men lost around a quarter of their VO2 max after three weeks of bed rest in 1966. When McGuire and colleagues followed them up in Circulation (104:1350-1357, 2001), thirty years of ageing had cost them roughly what three weeks of lying still once had. Inactivity is not a slower version of ageing, but on this particular measure it is a remarkably efficient imitation of it.

What the numbers look like across a population

Absolute values depend heavily on age, sex and training history. Broad reference ranges give useful orientation without pretending to precision.

Two structural features of that table are worth naming. The sex difference of roughly ten percent persists after adjusting for body mass, driven mainly by differences in haemoglobin concentration and heart size, and it is a population characteristic rather than a statement about any individual. And the age decline runs at roughly eight to ten percent per decade from the late twenties in sedentary people, but a good deal slower in people who keep training. Much of what is usually filed under ageing is deconditioning that happened to occur while someone got older.

Normalising by body weight also has a distorting effect. Dividing by kilograms is right for running, where you carry your mass, and wrong for rowing or cycling, where you largely do not. It also penalises muscular bodies: adding lean mass can lower your VO2 max in mL/kg/min while raising your absolute oxygen uptake in litres per minute. If you lift as well as run, expect the ratio to move against you for reasons that have nothing to do with your heart, a tension explored in what cardio and strength each do.

The mortality evidence is unusually strong

The reason clinicians pay attention to this number is not athletic performance. It is that cardiorespiratory fitness tracks survival better than most things routinely measured in a clinic.

Kodama and colleagues pooled 33 cohort studies in JAMA (301:2024-2035, 2009) and found each one-MET increment of fitness (about 3.5 mL/kg/min) was associated with roughly 13 percent lower all-cause mortality and 15 percent lower risk of cardiovascular events. Blair and colleagues had established the shape of the relationship two decades earlier in JAMA (262:2395-2401, 1989), where the steepest reduction in risk occurred between the least fit fifth of the population and the next fifth up.

Mandsager and colleagues extended it in JAMA Network Open (1:e183605, 2018) using treadmill test results from 122,007 patients. They found no upper threshold beyond which more fitness stopped helping, and the adjusted mortality difference between the lowest and highest fitness categories was larger than the differences associated with smoking, diabetes or coronary artery disease.

Three cautions belong with those figures. They are observational, so reverse causation is a live concern. Undiagnosed illness lowers fitness as well as raising mortality, though the better studies exclude early deaths to address this. The largest absolute gains sit at the bottom of the distribution, which is where most people are and where modest training does the most good. And these are population associations, not personal forecasts.

How much of it you can actually change

This is where the honest answer becomes uncomfortable.

The HERITAGE Family Study put 481 sedentary adults from 98 families through an identical, fully supervised 20-week cycling programme. Bouchard and colleagues reported the results in the Journal of Applied Physiology (87:1003-1008, 1999). The average improvement in maximal oxygen uptake was around 400 mL/min, roughly 17 percent. The range ran from essentially no change at all to more than a litre per minute. That is a threefold spread in a programme where every participant did the same sessions under supervision.

The critical finding was that the response clustered by family. Trainability aggregated within families to a degree implying heritability of roughly 47 percent for the size of the training response, quite separately from the heritability of baseline fitness. Bouchard and colleagues later identified panels of genetic variants accounting for a substantial share of that variance in the same journal (110:1160-1170, 2011).

So there are genuine high responders and genuine low responders, and which you are was largely decided before you started. Three things follow from that, and none of them is discouraging when read properly.

The first is that low responders on VO2 max are not low responders on everything. People whose maximal oxygen uptake barely budges still improve submaximal endurance, threshold pace, blood pressure, insulin sensitivity and body composition. VO2 max is one adaptation among many, and it is the one with the least favourable trainability profile.

The second is that starting position dominates. Someone beginning at 30 mL/kg/min has far more available headroom than someone at 55, and the health-relevant part of the curve is exactly where untrained people start.

The third is that the comparison worth making is against your own trajectory, not against a friend's. A programme that produced a 5 percent gain in you and 20 percent in someone else was not a failure; it was the same programme meeting different genetics.

What raises it fastest

Intensity does most of the work. Milanović, Sporiš and Weston's meta-analysis in Sports Medicine (45:1469-1481, 2015) compared high-intensity interval training against moderate continuous training and found intervals produced larger VO2 max improvements, with the effect holding across the trials they pooled. Bacon and colleagues reached a compatible conclusion in PLoS One (8:e73182, 2013), noting that the protocols producing the largest gains involved intervals of several minutes at close to maximal aerobic intensity, and that improvements of around half a litre per minute were achievable in previously untrained people.

The mechanism fits Levine's delivery-limited account. Intervals lasting three to five minutes at an intensity that drives you near maximal oxygen uptake spend more total time with the heart working at high stroke volume than an easy hour does, and that stress is what remodels the ventricle.

That is not an argument for training hard all the time. Sessions at that intensity are expensive, tolerated only a couple of times a week, and they sit on top of an aerobic base built at lower intensities. The adaptations that base produces, and the way it gets oversold, are both examined in the case for and against Zone 2. The distribution most trained endurance athletes converge on is a large majority of easy volume with a small, protected fraction of genuinely hard work.

For structuring that, the heart rate zone calculator gives serviceable intensity bands, with the caveat that they inherit whatever error sits in your estimated maximum. For a sense of the energy cost of the work you are doing, the calories burned calculator uses MET values from the Compendium of Physical Activities, which are population averages that take no account of your efficiency or your terrain.

Measuring it outside a laboratory

A true VO2 max test requires a metabolic cart, a mask, and a graded protocol taken to volitional exhaustion. It is a clinical or laboratory procedure, and in anyone with known cardiovascular, metabolic or renal disease, or symptoms suggestive of them, the decision to perform a maximal test belongs to a clinician rather than to a calculator or a coach.

Field estimates exist and are reasonable for tracking. Timed distance tests predict VO2 max from performance: a 12-minute run, a 1.5-mile run, a 20-metre shuttle test. Their error against laboratory values is typically in the range of 10 to 15 percent. Submaximal protocols like the Rockport walk test avoid maximal effort at the cost of more error. Wearable estimates work by modelling the relationship between your pace and your heart rate, which means they inherit every problem in your estimated maximum heart rate and tend to compress the extremes, reading the unfit too high and the very fit too low.

For practical purposes, the trend matters and the absolute value does not. A watch estimate that climbs three points across a training block is telling you something real even if the underlying figure is off by five. Compare it against itself, measured the same way, and treat any single reading as an approximation.

What the number will not tell you

Among trained endurance athletes, VO2 max is a poor predictor of who wins. Once a group of runners all sit in the same broad range, performance is determined by the fraction of that maximum they can hold for the duration of the event, by movement economy, and by pacing. Threshold pace and running economy separate athletes whose maximal oxygen uptake is effectively identical, and no amount of raising the ceiling helps if you cannot spend time near it.

It also says nothing about strength, mobility, bone density or metabolic health beyond what it correlates with, and nothing at all about body composition. It is one instrument reading one property of one system. On the question of long-term health, that property happens to be an unusually informative one, which is reason enough to know roughly where you stand and to spend some of your training on moving it.