REF-169 · Van Hooren et al., 2024 · Scandinavian Journal of Medicine & Science in Sports, · Paper

Accuracy of Respiratory Gas Variables, Substrate, and Energy Use from 15 CPET Systems During Simulated and Human Exercise

Blimited evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

Van Hooren et al. tested the accuracy of 15 widely used CPET systems by connecting each to a metabolic simulator that mimics breath-by-breath gas exchange; they then checked the simulation results with people cycling and, for some of the systems, measured reliability between sessions.

Systems (manufacturer in brackets): Quark CPET and K5 (COSMED); MetaLyzer 3B and MetaMax 3B (Cortex); Vyntus CPX and Oxycon Pro (Vyaire); Omnical (Maastricht Instruments); Ergocard Clinical, Ergocard Pro and Ultima (MGC Diagnostics); PowerCube Ergo (Ganshorn/Schiller); Ergostik (Geratherm); VO2masterPro (VO2master); PNOĒ; Calibre (Calibre Biometrics).

Protocol

Simulated breath-by-breath gas exchange; each system's minute ventilation, breathing frequency, VO2, VCO2, RER, energy from carbohydrate and fat, and total energy expenditure compared to the simulated values.

Key Results — absolute percentage errors during simulation (range across systems)

VE: 1.15–44.3%
Breathing frequency: 1.05–3.79%
VO2: 1.10–13.3%
VCO2: 1.07–18.3%
RER: 0.62–14.8%
Kcal from carbohydrate: 5.52–99.0%
Kcal from fat: 5.13–133%
Total energy expenditure: 0.59–12.1%
Between-session variation: 0.86–21.0% for VO2, 1.14–20.2% for VCO2

Conclusion (abstract): error differed substantially between systems, with only a few systems showing a consistently acceptable error. Which systems those were is not named in the abstract.

Practical Application

Fifteen CPET systems tested against a metabolic simulator: VO2 error ranged 1.1-13.3% across systems, VCO2 1.1-18.3%, RER 0.6-14.8%, and the substrate estimates were far worse — 5.5-99.0% for carbohydrate kcal and 5.1-133% for fat kcal; between-session variation reached 21.0% for VO2 on some systems. Only a few systems were consistently acceptable. A VO2max or fat-oxidation number is only as good as the analyser it came from.
Substrate oxidation from gas analysis: treat absolute values with suspicion — on some systems the fat/carbohydrate kcal error exceeds the value itself (up to 133%). Not from this paper; the base's own practice: set fat-oxidation training zones from lactate and RPE, use CPET substrate data qualitatively.
Not from this paper; the base's own practice: retest on the same system with the same calibration routine; read relative change, not absolute values, and do not compare VO2max across laboratories or devices as if they were interchangeable — the between-system spread here can exceed a season's real improvement.
The abstract does not rank or name the acceptable systems; which device is "validated" must be checked in the full paper, not assumed from brand.

Limitations

Simulator errors were verified in human cycling, but the abstract gives the numbers for the simulation only.
Between-session reliability assessed for a subset of systems.
Firmware/calibration versions and later product updates can change device accuracy.
How this entry was checked and changed
Why this level

regraded 2026-09-27 from A: test/equipment validation study, no performance outcome

Content correction 2026-09-27
"Top systems: COSMED Quark CPET, Cortex MetaLyzer, Vyaire Vyntus CPX" and "use ONLY validated systems (COSMED, Cortex)" -> the abstract names no acceptable systems; deleted.
"Difference between systems can be 5-10%" -> abstract range for VO2 is 1.1-13.3%; house figure removed. "Week 4 vs Week 10" -> relabelled base practice without weeks.
VE, breathing-frequency, total-energy and between-session ranges added from the abstract.
Level A -> B (equipment validation study, per rubric).