Why Your Estimated VO2 Max Might Drop Without Your Fitness Actually Changing

Covers wrist-based wearable VO2max estimation accuracy and reliability relative to laboratory gas-analysis testing. Does not cover diagnosed cardiopulmonary conditions or clinical cardiopulmonary exercise testing.

A VO2 max estimate that drops on a watch or app does not necessarily mean cardiorespiratory fitness actually declined. Wearable VO2max is a formula built from heart rate and pace data, not a direct physiological measurement, and its own validation research describes a real, documented reliability problem.

The Watch's Own Validation Study Called Its Reliability Poor

A validation study compared the Apple Watch Series 7's estimated VO2max against a metabolic gas analyzer, the laboratory gold standard, in 19 participants completing a graded cycle ergometer test to exhaustion. The watch's estimates carried a mean absolute percentage error of 15.79 percent against the lab measurement.

The same study calculated test-retest reliability using an intraclass correlation coefficient, a standard statistic for how consistently a measurement reproduces itself. But the result, 0.47, fell in the range the researchers themselves described as poor reliability. The lab-measured average VO2max was 45.88 mL/kg/min; the watch's estimate averaged 41.37, an underestimate on the group level.

A separate validation study using a different reference method, indirect calorimetry with 30 participants on a treadmill protocol, found the same device underestimated VO2max again, by a mean of 6.07 mL/kg/min. Two different labs. Two different protocols. Both point the same direction. How wearables actually build a VO2max estimate from heart rate and pace gets into why that consistency exists.

What Poor Reliability Actually Means for a Week-to-Week Change

An ICC of 0.47 does not describe a device that is simply a little bit off in a fixed way. It describes a measurement that does not reproduce itself consistently even when the underlying fitness has not changed, which is a different and larger problem than a stable bias would be.

A single point drop, or even several points, in an estimate with documented poor test-retest reliability is weak evidence of an actual fitness change on its own. Real cardiorespiratory fitness does change, through training, illness, detraining, or age, but it does not typically move in the kind of week-to-week swings that a formula this noisy can produce without any real change underneath it.

When a Fitness Change Is Worth Bringing to a Doctor

A wearable VO2max number that drops without any change in training, weight, or health fits the pattern this evidence describes: a noisy estimate, not a diagnosis. Other patterns don't. New shortness of breath during activities that used to feel easy, chest pain or tightness with exertion, or a real, measured decline in exercise capacity confirmed outside the watch are factual reasons to raise it with a doctor.

A formula built from heart rate and pace cannot diagnose a cardiac or respiratory change, it was never validated to. It can only estimate a number, with a documented margin around it, nothing more.

Common questions