VO(2) slow component: to model or not to model?

PURPOSE: The purpose of this study was to compare several techniques often used in the literature for measuring the amplitude of the slow component of oxygen uptake kinetics. METHODS: Eight healthy male volunteer cyclists performed two identical bouts of square wave cycle ergometry, from a VO(2) of 60% of the lactic acid threshold (LAT) to 30% of the difference between LAT and VO(2) peak. Predetermined intervals (3--6 and 3--10 min) were chosen to reflect those often used in the literature, namely 3-6 min and 3 min to the end of exercise. Several procedures were used to estimate the 3, 6, and 10-min VO(2) values (20-s averaging, 60-s averaging, and mono-exponential modeling). These were compared with the modeled slow component amplitude using a two-phase model with independent time delays: VO(2)(t) = B VO(2) + A(1)(1 -- e(-(t-TD1)/tau(1)) + A(2)(1 -- e(-(t-TD2)/tau(2)). CONCLUSIONS: The results showed a significant underestimation for all methods of slow component amplitude estimation (P < 0.05) when compared with the actual (modeled) amplitude. In so far as research on oxygen uptake kinetics is used to understand the underlying physiology, it is imperative that the components of the kinetics be determined accurately. The use of a predetermined time frame for estimation of the amplitude of the slow component is not supported by this study. Future investigations should consider these results and make every effort to model the underlying response.