Technological and Biological Reliability, and Validity of Five Different CPET Systems During Simulated and Human Exercise

ABSTRACT The validity and between‐day reliability of cardiopulmonary exercise testing (CPET) systems remain largely unexplored. We therefore evaluate the validity and between‐day technological and biological reliability of five popular CPET systems for assessing respiratory variables, substrate use, and energy expenditure during simulated and real human exercise. The following systems were assessed: Vyntus CPX, Oxycon Pro, VO2 Master, KORR, and Calibre. A metabolic simulator was used to simulate breath‐by‐breath gas exchange. The values measured by each system (minute ventilation (V̇E), breathing frequency (BF), oxygen uptake (V̇O 2 ), carbon dioxide production (V̇CO 2 ), respiratory exchange ratio (RER), energy from carbohydrates and fats, and total energy expenditure) were compared to the simulated values to assess the validity. Six well‐trained participants cycled 5% below their first ventilatory threshold on 2 days to verify the validity in human exercise. Between‐session reliability was assessed in both the simulation and human experiments to determine technological and biological variability. Absolute percentage errors during the simulations ranged from 0.69% to 5.56% for V̇E, 0.92% to 1.44% for BF, 3.12% to 7.86% for V̇O 2 , 4.07% to 12.1% for V̇CO 2 , 1.21% to 6.94% for RER, 2.83% to 48.8% for Kcal from carbohydrates, 14.1% to 50.3% for Kcal from fats, and 4.21% to 6.98% for total energy expenditure. Between‐session variability during simulation (i.e., technological variability) ranged from 0.46% to 3.15% for V̇O 2 and 0.71% to 4.99% for V̇CO 2 . The error and between‐day variability of the error for respiratory gas variables, substrate, and energy use differed substantially between systems. Biological and technological V̇O 2 and V̇CO 2 variability, respectively, accounted for ~60%–70% and 40%–30% of the variability in repeated human testing.