Accuracy of respiratory gas variables, substrate, and energy use from 15 CPET systems during simulated and human exercise

Purpose: Various systems are available for cardiopulmonary exercise testing (CPET), but their accuracy remains largely unexplored. We evaluate the accuracy of 15 popular CPET systems to assess respiratory variables, substrate use, and energy expenditure during simulated exercise. Cross-comparisons were also performed during human cycling experiments (i.e., verification of simulation findings), and between-session reliability was assessed for a subset of systems.Methods: A metabolic simulator was used to simulate breath-by-breath gas exchange, and the values measured by each system (minute ventilation [V?E], breathing frequency [BF], oxygen uptake [V?O-2], carbon dioxide production [V?CO2], respiratory exchange ratio [RER], energy from carbs and fats, and total energy expenditure) were compared to the simulated values to assess the accuracy. The following manufacturers (system) were assessed: COSMED (Quark CPET, K5), Cortex (MetaLyzer 3B, MetaMax 3B), Vyaire (Vyntus CPX, Oxycon Pro), Maastricht Instruments (Omnical), MGC Diagnostics (Ergocard Clinical, Ergocard Pro, Ultima), Ganshorn/Schiller (PowerCube Ergo), Geratherm (Ergostik), VO2master (VO2masterPro), PNOE (PNOE), and Calibre Biometrics (Calibre).Results: Absolute percentage errors during the simulations ranged from 1.15%-50.3% for V?E, 1.05-3.79% for BF, 1.10%-17.5% for V?O-2, 1.07%-18.3% for V?CO2, 0.62%-14.8% for RER, 5.52%-99.0% for Kcal from carbs, 5.13%-133% for Kcal from fats, and 0.59%-12.1% for total energy expenditure. Between-session variation ranged from 0.86%-22.4% for V?O-2 and 1.14%-20.2% for V?CO2, respectively.Conclusion: The error of respiratory gas variables, substrate, and energy use differed substantially between systems, with only a few systems demonstrating a consistent acceptable error. We extensively discuss the implications of our findings for clinicians, researchers and other CPET users.