Breath-by-breath oxygen uptake during running: Effects of different calculation algorithms

What is the central question of this study? Breath-by-breath gas exchange analysis during treadmill exercise can be disturbed by different breathing patterns depending on cadence, and the flow sensor might be subjected to variable mechanical stress. It is still unclear whether the outcomes of the gas exchange algorithms can be affected by running at different speeds. What is the main finding and its importance? Practically, the three investigated breath-by-breath algorithms ('Wessel', 'expiration-only' and 'independent breath') provided similar average gas exchange values for steady-state conditions. The 'independent breath' algorithm showed the lowest breath-by-breath fluctuations in the gas exchange data compared with the other investigated algorithms, both at steady state and during incremental exercise.

Recently, a new breath-by-breath gas exchange calculation algorithm (called 'independent breath') was proposed. In the present work, we aimed to compare the breath-by-breath O₂ uptake ( $$ ) values assessed in healthy subjects undergoing a running protocol, as calculated applying the 'independent breath' algorithm or two other commonly used algorithms. The traces of respiratory flow, O₂ and CO₂ fractions, used by the calculation algorithms, were acquired at the mouth on 17 volunteers at rest, during running on a treadmill at 6.5 and 9.5 km h^-1 , and thereafter up to volitional fatigue. Within-subject averages and standard deviations of breath-by-breath $$ were calculated for steady-state conditions; the $$ data of the incremental phase were analysed by means of linear regression, and their root mean square was assumed to be an index of the breath-by-breath fluctuations. The average values obtained with the different algorithms were significantly different (P < 0.001); nevertheless, from a practical point of view the difference could be considered 'small' in all the investigated conditions (effect size <0.3). The standard deviations were significantly lower for the 'independent breath' algorithm (post hoc contrasts, P < 0.001), and the slopes of the relationships with the corresponding data yielded by the other algorithms were <0.70. The root mean squares of the linear regressions calculated for the incremental phase were also significantly lower for the 'independent breath' algorithm, and the slopes of the regression lines with the corresponding values obtained with the other algorithms were <0.84. In conclusion, the 'independent breath' algorithm yielded the least breath-by-breath O₂ uptake fluctuation, both during steady-state exercise and during incremental running.