REF-1284 · Kaminsky et al., 2022 · Mayo Clinic Proceedings · Paper · PubMed

Is my VO2max good for my age?

FRIEND ranks a lab VO2max among 22,379 healthy US adults' tests by decade, sex and mode (cycle runs lower than treadmill); each older decade's median is ~4 mL/kg/min lower. Not athletes; tables not read.

Main caveat: Read from the abstract only; the percentile tables, the laboratories' protocols and the reasons people were tested could not be checked.
moderate evidenceMeasured on non-athletes and recreational people

Kaminsky and colleagues updated the FRIEND reference standards for cardiorespiratory fitness in the United States. Thirty-four laboratories contributed 22,379 cardiopulmonary exercise tests (16,278 on a treadmill and 6,101 on a cycle ergometer) carried out between 1 January 1968 and 31 March 2021 in apparently healthy adults aged 20 to 89.

Percentiles of peak oxygen uptake were calculated for men and women for each decade of age and each mode, and again with maximal effort defined as a respiratory exchange ratio of at least 1.0 or at least 1.1.

The 50th percentile fell with age in both sexes and both modes, was higher in men at every age and higher on the treadmill than on the cycle; over six decades it fell on average 13.5% (4.0 mL/kg/min) per decade on the treadmill and 16.4% (4.3) per decade on the cycle. Mean peak VO2 was similar under either RER criterion across modes, ages and sexes.

The new treadmill standards are 1.5 to 4.6 mL/kg/min lower than the 2015 FRIEND standards, while the cycle standards are generally comparable to those of 2017. The authors conclude the update represents the US population better.

Our reading: the base's only VO2max norm, useful for "where does my lab VO2max sit for my age and sex", with three caveats an assistant must carry: it ranks people who were tested in US laboratories, not athletes; the decline per decade is cross-sectional, not anyone's rate of loss; and the actual percentile values were not read for this entry.

Protocol

Registry: FRIEND; 34 US laboratories contributed tests.
Tests: 22,379 cardiopulmonary exercise tests with measured gas exchange, 16,278 treadmill and 6,101 cycle ergometer, from 1 January 1968 to 31 March 2021.
Who: apparently healthy adults aged 20-89; how they came to be tested (research volunteers, clinical or fitness referrals), their training status, ethnicity and the protocols used are not in the abstract.
Analysis: percentiles of peak VO2 for men and women for each decade from 20 to 89, separately for treadmill and cycle; repeated with maximal effort defined as RER ≥ 1.0 and RER ≥ 1.1; comparison with the 2015 (treadmill) and 2017 (cycle) FRIEND standards.
Funding and interests: not in the abstract; Crossref names no funder, and PubMed names none but tags the paper "Research Support, Non-U.S. Gov't". Not checked in the full text.

Key Results

Item (abstract)TreadmillCycle ergometer
Tests16,2786,101
Average fall in the 50th percentile per decade, over six decades13.5% (4.0 mL/kg/min)16.4% (4.3 mL/kg/min)
Against the earlier FRIEND standards1.5-4.6 mL/kg/min lower than 2015generally comparable to 2017
Sex and mode: the 50th percentile was higher in men in every age group and higher for treadmill than for cycle tests.
Effort criterion: mean peak VO2 was observationally similar with RER ≥ 1.0 or ≥ 1.1 across modes, ages and sexes.
Percentile values by decade, sex and mode: in the full-text tables, not read for this entry.

Mechanisms

Measured: lower peak VO2 in older decades, in women and on the cycle ergometer.
Authors' interpretation: the updated standards improve the representativeness of the US population compared with the original standards; why the treadmill values came out lower is not explained in the abstract.
Inference: cycle tests give lower peak VO2 than treadmill tests in the same kind of people, so a cyclist's lab value belongs in the cycle table, and comparing a cycle value with treadmill norms understates the rank.
Inference: the fall per decade compares different people born decades apart and tested between 1968 and 2021; it is not an individual's rate of decline, which depends on what that person keeps doing (REF-595, REF-1138).

Practical Application

Is my VO2max good for my age? FRIEND ranks a lab VO2max among 22,379 healthy US adults' tests by decade, sex and mode (cycle runs lower than treadmill); each older decade's median is ~4 mL/kg/min lower. Not athletes; tables not read.
Give the rank with what it is: a percentile among apparently healthy adults tested in US laboratories, in the matching mode; athletes sit high in it by construction. It is a description, not a forecast or a training target (B+, from the abstract).
An estimated VO2max needs its error first: from power, a ramp test or a watch, quote SYNTH-78's range (at least ±10 mL/kg/min) before any percentile; a percentile from an estimate can move by several bands within that range (inference).
Do not turn the decade decline into a personal forecast: 4.0-4.3 mL/kg/min a decade is a cross-sectional median (SYNTH-35: do not quote a fixed loss per decade at an individual). Not from this paper; the base's own practice: an athlete's own repeated lab test is the comparison that matters.

Limitations

Read from the abstract only; the percentile tables, the laboratories' protocols and the reasons people were tested could not be checked.
A registry of laboratory tests in people who came to be tested (volunteers or referrals; the abstract does not say), not a random population sample; "apparently healthy" is not "trained".
Tests span 1968-2021; secular changes in fitness and in equipment are mixed into the percentiles.
Peak, not necessarily maximal, VO2: RER ≥ 1.0 is a lenient criterion, although the authors report similar means with ≥ 1.1.
US adults only; ethnicity and body size distributions are not in the abstract.
Funding and conflicts of interest not in the abstract or the Crossref and PubMed records.
How this entry was checked and changed
Why this level

large cross-sectional measurement registry (the rubric's B+ for 100 or more), from the abstract (the article is open access on the publisher's site under CC BY-NC-ND, but that site refused a plain download, and no Europe PMC or repository copy exists). Not higher: a convenience registry of tests done in 34 laboratories over five decades, whose reasons for testing and protocols are not in the abstract; descriptive norms, not a test of anything