VO2 Max Lab Test vs. Wearable Estimate: What to Trust

What a lab VO2max test actually measures

A laboratory VO2max test is a form of cardiopulmonary exercise testing (CPET). You run on a treadmill or pedal a stationary bike while breathing through a mask or mouthpiece connected to a metabolic cart, a machine that performs indirect calorimetry: it measures the volume and gas concentration of every breath in and out, then calculates exactly how much oxygen your body is using per minute. The 2003 ATS/ACCP Statement on Cardiopulmonary Exercise Testing, a joint clinical statement from the American Thoracic Society and the American College of Chest Physicians, describes this as the reference method for measuring aerobic capacity because it captures the real physiological ceiling rather than inferring it.

The test itself follows a graded protocol: workload increases every one to three minutes until you can no longer continue. A classic criterion for confirming you actually hit a true VO2max, rather than just stopping from fatigue or motivation, comes from Henry Taylor, Elsworth Buskirk, and Ancel Henschel's 1955 study in the Journal of Applied Physiology. They defined a genuine plateau as an increase in oxygen uptake of less than about 150 mL per minute (roughly 2.1 mL/kg/min) despite a further rise in workload, and found that about three out of four subjects met that strict criterion. When someone stops before a plateau shows up, exercise physiologists call the result VO2peak rather than VO2max, since it is possible the true ceiling was never reached.

How a wearable estimates VO2max instead

A consumer wearable has no way to measure the gas you breathe, so it takes an entirely different approach: it estimates VO2max from the relationship between your heart rate and your pace or power output. The core idea, used in some form by Garmin, Apple, and most other fitness watches, is that a fitter cardiovascular system moves you at a given pace with a lower heart rate. Garmin's implementation runs on technology from Firstbeat Analytics, a Finnish physiology company Garmin acquired in 2020, whose published methodology builds a submaximal prediction model from heart rate, pace, and pace variability during a qualifying outdoor run of roughly ten minutes or more at an elevated heart rate. Apple Watch uses a comparable submaximal approach: it needs GPS and heart rate data from an outdoor walk, run, or hike, with your heart rate elevated to roughly 30 percent above resting, and combines that with your resting heart rate to fit a model rather than measuring gas exchange directly.

Neither approach asks you to go anywhere near exhaustion. That is the entire point of a submaximal estimate: it trades the effort, cost, and equipment of a true VO2max test for a number you can get from an ordinary training run, at the cost of relying on a population-level statistical model instead of a direct measurement of your own physiology.

How accurate wearable VO2max estimates actually are

The most complete look at this question is a 2022 systematic review and meta-analysis in Sports Medicine by Pablo Molina-Garcia and colleagues, done for the INTERLIVE network, which pooled 14 validation studies. It found that wearables using resting-condition data in their algorithm significantly overestimated lab-measured VO2max, with a pooled bias of 2.17 mL/kg/min. Devices that instead used exercise-based data, the Garmin and Apple approach described above, performed better on average, but the review's authors were explicit that exercise-based estimation looked reasonably accurate at the population level while the error for any single individual remained large, and concluded these methods still need improvement for sport or clinical use.

Two independent, peer-reviewed studies on the Apple Watch specifically illustrate what that individual-level error looks like. Sinead Lambe and colleagues published a 2025 validation study in PLOS ONE comparing Apple Watch VO2max estimates against a supervised maximal treadmill test in 30 participants. The watch underestimated VO2max by a mean of 6.07 mL/kg/min, with a mean absolute percentage error of 13.31 percent (95% CI 10.01 to 16.61). A separate 2024 validation study in JMIR Biomedical Engineering by Polona Caserman and colleagues, using a metabolic gas analyzer as the lab reference in 19 participants, found the pattern ran in both directions: lab-measured VO2max averaged 45.88 mL/kg/min against a watch-predicted average of 41.37 mL/kg/min, and the watch tended to overestimate people with lower fitness and underestimate people with higher fitness, a regression-to-the-mean pattern common to submaximal prediction models.

Garmin's own technical documentation for the Firstbeat-derived method reports considerably tighter agreement under the conditions it was built and tested in: a company white paper describes about 5 percent mean absolute percentage error for running, based on roughly 2,690 runs from 79 marathon-training runners whose lab VO2max was retested four times over six to nine months. That figure comes from the manufacturer's own validation rather than an independent peer-reviewed trial, and it reflects trained runners using a controlled protocol, so it should be read alongside the independent studies above rather than in place of them. The honest summary across all of it: a wearable estimate correlates well with lab VO2max as a fitness trend across large groups of people, but for any one person on any one day, the number can be off by a meaningful margin, and that margin tends to grow at the low and high ends of the fitness range.

The common misconception

The mistake runs in both directions. Treating the wearable number as clinically precise, or comparing your watch's figure against a friend's different brand, ignores how these algorithms are built and validated. But dismissing the number as useless goes too far the other way. Studies like Molina-Garcia and colleagues' 2022 meta-analysis found that exercise-based wearable estimates track reasonably well with lab VO2max at the population and trend level, which is exactly the use case most people actually need: is your aerobic fitness improving over months of training, not what your exact number is compared to an elite athlete's.

What throws off a wearable's estimate

Because the algorithm is built on population averages, anything that makes your physiology diverge from the average, or anything that feeds it noisy input data, widens the gap between the estimate and your true VO2max.

What to actually do with each number

Frequently asked questions