VO2 max explained for runners, and what it cannot tell you

Your watch prints a number. It is not a measurement of anything, it is the output of a model that has never seen the inside of your lungs, and even a proper laboratory figure would not tell you the thing you actually want to know.

That is not a reason to ignore it. VO2 max explained honestly is one of the more useful ideas in endurance running, because the equation behind it says exactly which parts of your physiology are worth training and which are fixed. What it does not do is predict who wins, and the reason it does not is more interesting than the number itself.

The short version: maximal oxygen uptake sets the ceiling on how much aerobic work you can do. It says nothing about how much running you get for each litre of oxygen, and nothing about how long you can operate near the ceiling. Two runners with identical engines can be separated by minutes.

VO2 max explained: the equation it actually comes from

Maximal oxygen uptake is the largest rate at which your body can take oxygen from the air and consume it in working muscle. It is a rate, measured in millilitres per minute, and it comes out of a single identity.

Oxygen consumption equals cardiac output multiplied by the arteriovenous oxygen difference. Cardiac output is how much blood the heart moves per minute, which is heart rate multiplied by the volume ejected per beat. The arteriovenous difference is how much oxygen the tissues strip out of each litre of that blood. Delivery times extraction. Nothing else is in the equation.

Here is that arithmetic on invented round numbers, so the shape of it is visible.

• 190Maximum heart rate, beats per minute
• 160Stroke volume, millilitres per beat
• 30.4Cardiac output, litres per minute
• 69.5Relative VO2 max, millilitres per kilogram per minute

Every value here is invented and round, chosen so the arithmetic is legible. It describes nobody. Cardiac output is heart rate times stroke volume; oxygen uptake is cardiac output times the oxygen extracted per litre of blood; the relative figure divides by an invented 70kg body mass.

Now look at which terms in that equation can be changed.

Maximum heart rate is effectively fixed. It is not a fitness marker, it does not rise with training, and in trained athletes it tends to drift very slightly downward. It falls with age regardless of what anybody does. Any training programme that promises to raise your maximum heart rate is selling something.

Stroke volume is the term that moves. Endurance training increases the volume of blood the heart ejects per beat, through a larger ventricular chamber, more plasma and a heart that fills better during the brief diastole available at high rates. This is where most of the trainable improvement in VO2 max lives, and it is why the standard prescription is long, repeated work at high cardiac output rather than short maximal efforts.

The extraction term moves too, more slowly. More capillaries per muscle fibre, more mitochondria, more enzyme activity in the aerobic pathways: all of it increases how much oxygen the muscle can pull out of the blood delivered to it. This adaptation is largely local to the muscles you train, which is why the number is specific to the activity.

Where the ceiling sits, and it is not your lungs

Ask a room of runners what limits VO2 max and most will say breathing. In healthy people at sea level, it almost never is.

The lungs oxygenate blood so efficiently that arterial saturation stays high through maximal exercise for most people, and a system that is already near-saturated cannot be the bottleneck. There is an exception, and it is a revealing one: some very highly trained endurance athletes do desaturate at maximal effort, because their cardiac output has grown large enough that blood passes through the pulmonary capillaries too quickly for full equilibration. In those athletes the lung has become limiting precisely because everything downstream got better.

For everybody else the constraint is delivery. Manipulate the oxygen-carrying capacity of the blood and VO2 max moves with it, in both directions, which is the single most direct piece of evidence about where the limit sits. It is also, uncomfortably, the reason blood doping works and the reason the anti-doping system watches haematological markers over time rather than looking for a drug. The logic of monitoring an athlete's own blood profile for changes rather than testing for a substance follows directly from this physiology.

The mitochondria are generally not limiting either. Trained muscle has more oxidative capacity than the cardiovascular system can supply, which is why a small muscle group working alone can achieve oxygen consumption rates per kilogram that the whole body never approaches.

1. 1VentilationAir is moved into the lungs. Rarely limiting at sea level in healthy people; the respiratory system has spare capacity at maximal exercise, which is why breathing exercises do very little for this number.
2. 2Diffusion across the alveolar membraneOxygen crosses from the air spaces into the pulmonary capillaries. Fast and effective, except in some highly trained athletes whose cardiac output is large enough to shorten the time available for equilibration.
3. 3Binding to haemoglobinOxygen is loaded onto red cells. The total oxygen-carrying capacity of the blood is set here, which is why altering red cell mass moves the whole ceiling and why the anti-doping system watches this variable closely.
4. 4Cardiac outputHeart rate times stroke volume. In practice this is the limiting step, and stroke volume is the part of it that responds to training. Almost everything a distance programme does is aimed here.
5. 5Distribution to working muscleBlood is redirected away from tissues that are not working and towards the ones that are. Trainable, specific to the muscles used, and part of why a runner's figure on a bicycle is lower than on a treadmill.
6. 6Diffusion into the muscle fibreOxygen crosses from capillary to fibre. Capillary density per fibre rises with endurance training, shortening the diffusion distance and raising the extraction the muscle can achieve.
7. 7Mitochondrial consumptionOxygen is used at the end of the electron transport chain. Trained muscle generally has more capacity here than the delivery system can keep supplied, so this step is rarely the ceiling in a whole-body test.

Seven steps between the air and the mitochondrion. Only one of them is usually the limiting one in a healthy person at sea level, which is why training that targets the others produces so little.

Altitude is the cleanest proof that delivery is the limit

Take the same runner to a mountain and their maximal oxygen uptake falls. Nothing about their heart, their muscles or their training has changed in the hours it took to get there.

What has changed is the partial pressure of oxygen in the air, which lowers the amount carried per litre of blood. Delivery drops, and the ceiling drops with it, immediately and in proportion. If the limit sat in the mitochondria or in ventilation, a thinner atmosphere would not do that so cleanly.

The adaptations that follow are the same story running in reverse. Weeks at altitude increase red cell mass, restoring some of the oxygen content of each litre of blood, and a runner who returns to sea level carries that improved delivery with them for a period. That is the entire logic of altitude training, and it is why the protocol that survived scrutiny is to live high and train low: the adaptation is driven by the hours spent breathing thin air, while the quality of hard training is degraded by trying to do it up there.

It also explains why altitude helps some athletes and not others, and why the effect is measured in blood variables rather than in enthusiasm. If a camp does not move an athlete's red cell mass, it has not acted on the mechanism, and no amount of scenery compensates.

The denominator decides who looks good

VO2 max in litres per minute is a property of your body. VO2 max in millilitres per kilogram per minute is a property of your body divided by your mass, and that division does a lot of work.

Runners use the relative figure because they carry themselves. A cyclist on flat ground cares much more about the absolute one, because the bicycle carries the rider. Neither convention is wrong, and swapping between them changes who looks impressive.

Here is what the denominator does, using one fixed absolute value and nothing else changing.

Arithmetic on an invented constant: an absolute maximal oxygen uptake of 4.2 litres per minute, divided by body mass. Nothing physiological changes across this chart. Only the denominator does.

Show the numbers

The same absolute engine, divided by different body masses

Item	Value
50kg	84ml/kg/min
55kg	76.4ml/kg/min
60kg	70ml/kg/min
65kg	64.6ml/kg/min
70kg	60ml/kg/min
75kg	56ml/kg/min

Nobody in that chart got fitter. The figure moved by half again from one end to the other purely because of how much the athlete weighs.

There is a further complication that most articles skip. Dividing by mass raised to the power of one assumes oxygen uptake scales in direct proportion to body mass, and the physiological literature on allometric scaling argues it does not: the appropriate exponent is below one, with two thirds and three quarters both defended. If that is right, then simple ml/kg/min systematically flatters small athletes and penalises large ones, which matters in a sport where a marathon runner and a 400 metre runner can differ substantially in mass.

The practical version: comparing your relative figure with somebody of a different build tells you about the two builds as much as about the two engines.

What a maximal test actually involves, and why the number wobbles

A laboratory test is an incremental protocol, on a treadmill for runners, where speed or gradient rises in steps until the athlete cannot continue. Expired air is collected continuously and analysed for oxygen and carbon dioxide, and the highest sustained oxygen uptake is recorded.

The classical criterion for a true maximum is a plateau: oxygen uptake stops rising even though the workload does. That is what makes it a maximum rather than merely the highest value observed. The awkward fact is that plateaus frequently fail to appear, particularly in less-trained participants, and when no plateau appears the test has not demonstrated a ceiling at all.

Laboratories therefore fall back on secondary criteria, which are agreement markers rather than proof. A respiratory exchange ratio above a threshold value, a heart rate close to the age-predicted maximum, a blood lactate concentration above a set level, an exhaustion rating at the top of the scale. Meeting several of them makes it likely the athlete was at or near their maximum. It does not make it certain, which is why careful work reports VO2 peak rather than VO2 max when no plateau was seen.

The better modern practice is a verification bout: after a recovery period, the athlete performs a short, constant-intensity effort above the highest stage they reached. If oxygen uptake in the verification does not exceed the value from the ramp, the ramp value is confirmed.

Protocol changes the answer too. Ramp rate, stage length, treadmill gradient, whether the athlete holds the handrail, whether they are running or cycling: all of these shift the number. Two laboratories reporting different values for the same person may both be right about their own protocol. A figure without its protocol is a figure without units.

And the day matters. Sleep, heat, hydration, illness, time since the last hard session and simple motivation all move the result. Maximal oxygen uptake is the most reproducible of the endurance measures, and it is still not a constant.

What the watch is doing, and why it is not that

A wearable does not measure oxygen. It measures heart rate, or estimates it from a wrist optical sensor, and it measures pace, usually from GPS with an accelerometer filling the gaps.

From those two signals it builds the relationship between your heart rate and your speed during ordinary submaximal running, extrapolates that relationship out to an assumed maximum heart rate, and converts the resulting speed into an oxygen cost using an assumed running economy. Three assumptions, each of which can be wrong, stacked on two measurements, each of which can be noisy.

The failure modes follow directly. Heat pushes heart rate up at a given pace, so the estimate falls in summer and rises in autumn without any change in fitness. Dehydration does the same. Caffeine, poor sleep, illness and the first week after a hard block all move heart rate at a given pace. Hills and wind change the true oxygen cost while the watch still assumes flat, still air. GPS error inflates or deflates pace, particularly on twisting routes or under trees. And if the assumed maximum heart rate is wrong, every estimate is proportionally wrong in the same direction forever.

None of that makes it useless. A wrist estimate tracked over months, on similar routes in similar conditions, is a reasonable trend line for one person against their own history. What it cannot support is a comparison between two people, and it is precisely that comparison that people use it for.

VO2 max explained for running, not for cycling or rowing

Maximal oxygen uptake is mode-specific, and the reason is in the cascade above: distribution and extraction adapt in the muscles you actually train.

A trained runner tested on a bicycle will usually record a lower figure than on a treadmill, because their legs are not conditioned for that pattern of contraction and their body is not practised at directing blood to those muscles in that posture. A trained cyclist tested running will show the mirror image, with the added complication that running is mechanically harder on untrained legs.

This is why a runner's number and a cyclist's number are not comparable even before you get to the absolute-versus-relative problem. The cycling world has largely stopped arguing about maximal oxygen uptake for that reason and moved to sustainable power, which is measurable in the field every day rather than in a laboratory once a year. The reasoning behind why cyclists settled on a sustainable power figure instead of a maximal one applies almost unchanged to runners, and the sport has been slower to accept it only because a runner's power output is much harder to measure than a cyclist's.

The three-variable model, and why the other two do the work

Distance running performance is usually described by three physiological variables working together. Maximal oxygen uptake, which sets the ceiling. Fractional utilisation, which is the proportion of that ceiling you can hold for the duration of the event. And running economy, which is how much oxygen you use to travel a given distance.

The third one is the key to seeing why the first is oversold, because it converts oxygen into speed.

Running economy is normally expressed as the oxygen cost of covering a kilometre, per kilogram of body mass. Divide maximal oxygen uptake by that cost and you get the speed at which you would be consuming oxygen at your maximum: velocity at VO2 max, which is a far more useful number than either input on its own.

Now watch what happens with two constructed runners. The values are invented and round.

Runner A has a maximal oxygen uptake of 70 millilitres per kilogram per minute and an oxygen cost of 210 millilitres per kilogram per kilometre. Seventy divided by 210 is a third of a kilometre per minute, which is 20.0 kilometres per hour.

Runner B has a maximal uptake of 62 and an oxygen cost of 185. Sixty-two divided by 185 is 0.335 kilometres per minute, which is 20.1 kilometres per hour.

Runner B has a maximal oxygen uptake more than a tenth lower and reaches it at a marginally higher speed. The eight-point gap in the headline figure has vanished entirely.

Then add the third variable. Suppose Runner A can hold 80 per cent of that speed for a 10 kilometre race and Runner B can hold 85 per cent, which is a difference in threshold rather than in capacity.

• Runner A
• Runner B

Speed at maximal oxygen uptake20km/h20.1km/h

Speed at their own sustainable fraction16km/h17.1km/h

A constructed example using invented round values, not a description of any athlete. Runner A has a maximal uptake of 70 ml/kg/min and an oxygen cost of 210 ml/kg/km; Runner B has 62 and 185. The lower bars apply an invented sustainable fraction of 80 per cent for A and 85 per cent for B, treating oxygen cost as linear with speed, which is an approximation.

Show the numbers

Two constructed runners, the same arithmetic, opposite conclusions

Item	Runner A	Runner B
Speed at maximal oxygen uptake	20km/h	20.1km/h
Speed at their own sustainable fraction	16km/h	17.1km/h

Those race speeds are about 3 minutes 45 seconds and 3 minutes 31 seconds per kilometre. Over 10 kilometres that is the best part of two and a half minutes, in favour of the runner whose VO2 max is lower. Nothing in the arithmetic is exotic. The headline number is simply the least decisive of the three terms.

Why the number stops predicting anything at the sharp end

There is a statistical reason, quite separate from the physiology, why maximal oxygen uptake predicts performance well in the general population and badly among good runners.

A variable can only explain differences in outcomes to the extent that it varies. Take a large, mixed group containing sedentary people and club runners and internationals, and aerobic capacity will track race performance strongly, because it spans an enormous range across that group. Take the finalists of a national championship, where everybody has trained for a decade and the capacity values sit inside a narrow band, and the same variable explains very little. This is range restriction, and it is a property of the sample rather than a discovery about the body.

The consequence is one of the more useful things to understand in endurance sport. A high VO2 max is a qualifying condition. It gets you into the room. Once everybody in the room has one, the question of who wins is decided by the variables that still differ, which are economy, threshold and the ability to hold both together when tired.

This is also why the figure is a poor talent identification tool applied on its own and a reasonable one applied as a filter. It can tell you that a young athlete has the aerobic capacity to be a distance runner. It cannot tell you which of two athletes who both have it will be better.

Running economy is the variable with the most room in it

If maximal uptake is close to a ceiling in trained runners, economy is not. It varies widely between athletes with similar capacities, and it keeps improving for years after the ceiling has stopped moving.

What actually determines it is a long list, and that is part of why it responds to so many different interventions. Tendon stiffness and the elastic energy returned by the Achilles and the arch. The distribution of mass along the limb, which is why shoe weight at the foot costs more than the same weight at the waist. Body mass and anthropometry. Muscle fibre composition. Stride mechanics, ground contact time, vertical oscillation. Years of accumulated running, which improves economy through mechanisms nobody has fully separated.

The measurement has a subtlety that most popular accounts miss. Oxygen cost is only a valid proxy for energy cost if the respiratory exchange ratio is accounted for, because the energy released per litre of oxygen depends on whether the runner is burning fat or carbohydrate. Two runners with identical oxygen costs and different substrate mixes are not using the same energy. Careful work reports economy in kilojoules or kilocalories per kilogram per kilometre for exactly this reason, and comparisons between studies that use different conventions are shakier than they look.

Economy is also the variable most exposed to equipment, which is what made the last decade in road running so complicated. Shoe construction changes the oxygen cost of running at a fixed speed, and it does so without touching the athlete's physiology at all, which is why World Athletics ended up regulating sole thickness rather than trying to regulate the effect. A rule about millimetres of foam is really a rule about running economy.

Fractional utilisation, and where the threshold lives

The middle term is the one runners feel. It is the proportion of maximal oxygen uptake you can sustain, and it falls as the event gets longer: high for a 5 kilometre race, lower for a half marathon, lower again for a marathon.

What sets it is the point at which blood lactate begins to accumulate faster than it can be cleared, and the intensity a runner can hold indefinitely sits just below that. Raising it is what most serious endurance training is for, and the mechanism by which lactate is produced, cleared and eventually outpaced explains far more about a runner's race times than their maximal capacity does.

The relationship between the two is worth stating plainly. VO2 max is the size of the engine. Threshold is the fraction of it you can run at without the wheels coming off. A runner who improves their threshold from a lower to a higher percentage of an unchanged maximum has got faster without moving the headline number at all, and this is the single most common pattern in trained runners' physiology over a career.

It is also why the training that raises VO2 max and the training that raises threshold look different. The first involves repeated efforts near maximal oxygen uptake. The second involves longer, steadier work at or slightly below the threshold intensity, in volumes that only make sense if you understand what is being adapted.

The mechanism behind the first is worth stating precisely, because it is usually described as a session format when it is really a dose. What drives the adaptation is accumulated time spent at or close to maximal oxygen uptake, and the reason interval formats exist at all is that almost nobody can accumulate that time continuously. Oxygen uptake takes a minute or two to rise towards its ceiling once an effort begins, and it falls slowly during a short recovery, so a series of efforts with incomplete rest keeps a runner near the top of the range for far longer than one continuous effort at the same intensity could.

That is why arguments about whether four-minute repetitions beat thirty-second ones tend to go nowhere. Both formats work if they deliver the dose, and both fail if they do not. Repetitions that are too short with recoveries that are too long never let uptake climb. Repetitions that are too long force the pace down below the range that matters. The format is a delivery mechanism for time near the ceiling, and a coach who understands that can design a session in any shape they like.

The corollary is uncomfortable for anybody who enjoys hard sessions. Once a trained runner's ceiling has stopped moving, more of this work buys progressively less, while costing the same in recovery and in risk. That is the real reason elite endurance programmes look so unglamorous from the outside: they are spending their limited hard days on the variables that still respond.

Durability, the variable nobody was measuring

All three classic determinants are measured on a fresh athlete in a laboratory. Races are not run fresh.

Recent work has started asking what happens to those determinants after a runner has already been going for an hour or two, and the answer is that they change. Economy deteriorates. The heart rate and lactate response at a given speed shifts. The threshold intensity measured after prolonged running is not the threshold intensity measured at the start of it. A study examining well-trained marathon runners after 90 and 120 minutes of running found the determinants of endurance performance meaningfully altered by the prior work, which is a straightforward finding with awkward implications for every prediction model built on fresh values.

For marathon runners this is close to the whole problem. A runner whose fresh numbers are excellent and whose numbers at 30 kilometres are ordinary will run an ordinary marathon, and no fresh test would have predicted it. Whatever quality resists that deterioration has come to be called durability or physiological resilience, and it is trainable, and it is not what a treadmill ramp measures.

The practical version has been known to coaches for a long time without the vocabulary. It is why long runs finish with quality rather than starting with it, why marathon-pace work is placed late in a session rather than early, and why how effort is distributed across a long race matters more than any single physiological value the runner brings to the start line.

How much can you actually change it

The honest picture has two halves that get conflated.

An untrained person who begins endurance training will see a substantial rise in maximal oxygen uptake, mostly from stroke volume, over a period of months. This is real, reliable and one of the better-established effects in exercise physiology.

A trained runner will not. Once several years of consistent aerobic training are in the bank, the ceiling is close to wherever that individual's ceiling is, and further improvement is slow, small and inconsistent. Individual responsiveness varies a great deal, to the point where identical programmes produce large gains in some people and almost none in others, and nobody can currently predict in advance which you will be.

Age subtracts steadily. Part of that decline is falling maximum heart rate, which nothing prevents. Part is loss of muscle mass and changes in body composition, which training slows considerably. Part is simply doing less, which is the portion that is entirely within the athlete's control and the portion that most cross-sectional charts quietly attribute to ageing.

The reason this matters is that a trained runner spending their limited hard sessions chasing the ceiling is spending them on the variable with the least room. The same hours put into threshold work, volume and the technical and neuromuscular components of economy act on variables that are still moving.

Using the number honestly

There is a version of this figure that is genuinely useful, and it involves treating it as an input rather than a score.

Convert it into a speed. Velocity at maximal oxygen uptake is the number worth knowing, because it combines capacity and economy into something you can run at. Interval training prescribed as a percentage of that speed makes physiological sense in a way that intervals prescribed as a percentage of a heart rate do not.

Track the oxygen cost, not the capacity. For most runners the accessible proxy is heart rate at a fixed, repeatable submaximal pace on a known route in known conditions. If that drifts down over months, economy or capacity or both are improving. It is cruder than a laboratory and far more frequent, which is often the better trade.

Know your threshold pace and update it. It is the number that decides your racing over anything longer than about fifteen minutes, and it is the number most amateur runners have never actually established.

Test tired. How your pace at a given effort holds up in the last third of a long run tells you something no fresh measurement can, and it is the quality that decides marathons.

Ignore anybody else's figure. Different protocol, different body mass, different sport, different denominator, possibly a wrist estimate rather than a measurement.

The number was never the point. It describes the size of a room, and the events on a track or a road course are decided by how efficiently a runner moves within it and how long they can stay near the far wall. That is true of a 5000 metres, and it is true of an event where thirty-five obstacles make efficiency the whole discipline. The ceiling gets you in. Everything else decides what happens next.