Fitness

VO2 max and longevity: what the mortality research shows

Cardiorespiratory fitness is one of the most consistent predictors of mortality in observational research. How large the association is, what it is compared with, and why it is not a guaranteed number of extra years.

By The SnapLongevity Team10 October 20266 min read

Fitness studies measure what the body can do at maximum effort. What that means for lifespan depends on how the evidence is read.

VO2 max is the highest rate at which the body can take in and use oxygen during maximal exercise. It is the standard laboratory measure of cardiorespiratory fitness, and studies of fitness and mortality have produced some of the steepest gradients in the epidemiology of health behaviour. The finding is robust. What it means for any one person is narrower than the headlines suggest.

0.87
Relative risk of death per 1-MET higher fitness: 1.70
Low vs high fitness, relative risk of death: 122,007
Patients in the treadmill cohort

Figures as reported in each study's abstract. Relative risks compare groups within the studied populations; they are not absolute risks or individual predictions.

In plain English

VO2 max: Maximal oxygen uptake, in millilitres of oxygen per kilogram of body weight per minute, measured with gas analysis during a graded exercise test.
MET: Metabolic equivalent: one MET is roughly the oxygen used at rest. Many studies report fitness as the peak METs reached on a treadmill rather than measured VO2 max.Why it matters: Most large mortality studies estimated fitness from treadmill performance, not from direct gas analysis.

What the major studies found

A 2009 meta-analysis in JAMA pooled 33 observational cohorts of healthy adults, covering 102,980 participants and 6,910 deaths. Each 1-MET higher level of maximal aerobic capacity was associated with a 13% lower risk of death from any cause (relative risk 0.87). People in the lowest fitness category had a 70% higher risk than those in the highest.

A 2018 study in JAMA Network Open followed 122,007 patients referred for treadmill testing at a US clinic, with 13,637 deaths. In adjusted models, below-average fitness compared with above-average carried a hazard ratio of 1.41 — about the same as smoking (1.41) or diabetes (1.40) in the same study. These are observational associations, not proof that changing fitness alone would change risk by that amount. Mortality was lowest in the highest-fitness ('elite') group, and its advantage over the next group was also seen in patients aged 70 or older.

How large is the association, and compared with what?

Adjusted hazard ratios for all-cause death in one cohort: lower fitness and three clinical risk factors

Filled square: fitness comparison. Open squares: clinical risk factors reported in the same study, for scale. Each row compares a group with its own reference group, shown in the label.

Source: Mandsager et al., JAMA Network Open 2018 (abstract, Results). 122,007 patients referred for treadmill testing at one US clinic (mean age 53), 13,637 deaths; risk-adjusted hazard ratios with 95% confidence intervals, all from the same study. Observational associations: they compare groups, are not absolute risks or individual predictions, and do not prove that changing fitness would change risk by this amount. The patients were referred for testing, so they are not a general-population sample. Log scale.

Association is not cause

Every figure above is observational. Fitness partly reflects training and partly reflects health that is not otherwise measured: people with undiagnosed heart, lung or metabolic disease reach lower treadmill levels and also die sooner. Fitness is also partly inherited. Studies adjust for known risk factors, but they cannot remove confounding they did not measure, and no trial has randomised people to lifelong fitness levels and counted deaths.

Relative risk

How much higher the rate of death is in one group than another. A relative risk of 1.70 means 70% higher, whatever the starting level.

Absolute risk

The actual chance of dying over a period. If a group's baseline risk is low, even a large relative difference can be a small absolute one.

The cited abstracts report relative measures. Translating them into years of life for an individual would require assumptions these studies do not test.

Laboratory tests and wearable estimates

Many watches now display an estimated VO2 max. A 2022 meta-analysis of 14 validation studies found that devices using resting data overestimated VO2 max on average by about 2.2 ml/kg/min, with wide individual error; devices using exercise data were closer on average but still had large individual error. The authors judged them reasonable for population-level estimates but not yet precise enough for clinical use.

Laboratory VO2 max uses gas analysis at maximal effort and is the reference standard.
Treadmill METs, used in most mortality studies, estimate fitness from speed and incline reached.
Wearable estimates are algorithms, and their error for one person can be large.
Associations reported for measured fitness should not be assumed to apply with the same strength to a device estimate.

Is VO2 max the best predictor of longevity?

Some coverage calls VO2 max the single best predictor of how long you will live. The cited studies do not test that precise claim. They show that measured fitness is strongly associated with mortality, and in one clinical cohort its association was comparable to or greater than several traditional risk factors. That is a strong finding, but it is not a ranking of every possible predictor.

For a shorter summary alongside the other evidence factors, see cardiorespiratory fitness.

Mortality associations do not measure years lived in full health. WHO's population-level HALE is a different measure, compared with total life expectancy at birth and at 60. Compare life expectancy and healthy years.