Purpose: The portable metabolic analyzer COSMED K5 (Rome, Italy) allows for switching between breath-by-breath (BBB) and dynamic micro-mixing chamber (DMC) modes. This study aimed to evaluate the reliability and validity of the K5 in BBB and DMC at low, moderate, and high metabolic rates.
Methods: Two K5 simultaneously operated in BBB or DMC, whereas (i) a metabolic simulator (MS) produced four different metabolic rates (repeated eight times), and (ii) 12 endurance-trained participants performed bike exercise at 30%, 40%, 50%, and 85% of their individual power output at V˙O2max (repeated three times). K5 data were compared with predicted simulated values and consecutive Douglas bag measurements.
Results: Reliability did not differ significantly between BBB and DMC, whereas the typical error and intraclass correlation coefficients for oxygen uptake (V˙O2), carbon dioxide output (V˙CO2), and minute ventilation (V˙E) ranged from 0.27% to 6.18% and from 0.32 to 1.00 within four metabolic rates, respectively. Validity indicated by mean differences ranged between 0.61% and -2.05% for V˙O2, 2.99% to -11.04% for V˙CO2, and 0.93% to -6.76% for V˙E compared with MS and Douglas bag at low to moderate metabolic rates and was generally similar for MS and bike exercise. At high rates, mean differences for V˙O2 amounted to -4.63% to -7.27% in BBB and -0.38% to -3.81% in DMC, indicating a significantly larger difference of BBB at the highest metabolic rate.
Conclusion: The K5 demonstrated accurate to acceptable reliability in BBB and DMC at all metabolic rates. Validity was accurate at low and moderate metabolic rates. At high metabolic rates, BBB underestimated V˙O2, whereas DMC showed superior validity. To test endurance athletes at high workloads, the DMC mode is recommended.