Effects of Sodium Bicarbonate on VO2 Kinetics During Heavy Exercise

Fred W. Kolkhorst; Robert S. Rezende; Susan S. Levy; Michael J. Buono


Med Sci Sports Exerc. 2004;36(11) 

In This Article


Ten active individuals (nine males, one female; 28 ± 9 yr; 82.4 ± 11.2 kg) volunteered to serve as subjects. All subjects provided informed consent to participate in the study, which was approved by the Institutional Review Board of San Diego State University.

As is typically done in O2 kinetics studies on heavy exercise, subjects exercised at a relative percentage of power between that at VT and O2max. Because several subjects were over 40 yr of age, and because of the institution's IRB restrictions on maximal testing of older subjects without a physician present, maximal testing was not performed. Rather, subjects exercised at an absolute power (25 W) above their VT, which, from a review of the literature on O2 kinetics, was a typical power that was 50% between that at VT and O2max. VT was determined at the initial visit using a continuous, incremental cycling protocol. Exercise began at 50 W (Lode Excaliber, Groningen, The Netherlands). Work rate was increased 25 W·min-1 until at least one workload above the VT, which was determined by visually inspecting for the rapid increase in the ventilatory equivalent of oxygen ( E/ O2) graphed against the ventilatory equivalent of carbon dioxide ( E/ CO2).[23]

On two separate occasions, each subject completed one 6-min cycling bout at a power output predicted to correspond to 25 W greater than power output at VT. Upon reporting to the laboratory, the subject was given either 0.3 g·kg-1 body weight of sodium bicarbonate dissolved in 1 L of water or water only. An hour after consuming the solution, a hand was immersed in hot water for 2-3 min to obtain arterialized blood from a finger stick. Blood was analyzed for pH and bicarbonate concentration (i-STAT, East Windsor, NJ). Afterward, subjects completed the cycling bout. A 5-min warm-up of unloaded cycling preceded each bout, and the increase in workload occurred without warning to the subject so as to avoid an anticipatory ventilatory response. O2 was measured using a TrueMax 2400 system (Parvo Medics, Sandy, UT) in the breath-by-breath mode.

Prior investigations of O2 kinetics average data from multiple trials for each subject in order to reduce non-Gaussian noise (for instance,[4,13]) However, because of concern for subjects from repeated discomfort with multiple trials, subjects were asked to perform only one trial for each treatment. Extraneous O2 values, defined as being outside three SD from the average of the preceding and subsequent three data points, were considered outliers and eliminated, and the remaining data were interpolated to 1-s intervals. The data were time aligned to the beginning of exercise, and, because subjects performed only one bout for each trial, the data were averaged to 5-s intervals to further reduce noise.

O2 data from the cycling trials were fitted to a three-component, nonlinear least regression model (equation 1) (see Fig. 1).


The model included a baseline term ( O2base), three asymptotic amplitude terms (A1, A2, and A3), three time constants (τ1, τ2, and τ3), and three time delays (TD1, TD2, and TD3). O2(t) is the time-dependent variation of O2. Fitting for the second component was not begun until the first component was complete; likewise, the third component was not begun until the second component was finished.

A three-component exponential model for O2 kinetics applied in response to application of square-wave work above the ventilatory threshold. O2base is the steady-state O2 at the onset of exercise; A'1, A'2, and A'3 are amplitudes; TD2 and TD3 are time delays; and τ1, τ2, and τ3 are time constants (i.e., the time required to achieve 63% of the amplitude) for the three phases. Subscripts indicate the cardiodynamic (1), rapid (2), and slow (3) components of O2 kinetics.

The curve fittings were iterated using a nonlinear regression program (NLREG v5.2, Phillip H. Sherrod) until achieving a best fit of the residual sum of squares between the predicted and actual O2. Convergence was achieved for each model fit. To ensure that the predicted parameters remained within a physiological range, parameters were constrained within a wide range during iterations. A typical O2 response and curve fitting are shown in Figure 2.

O2 responses of a representative subject to square-wave exercise at 25 W above the ventilatory threshold for the control trial. The heavy line indicates O2 predicted from a three-component, nonlinear regression model.

The difference in O2 from the end of the cardiodynamic phase (i.e., at TD2) and O2base was reported as A'1; likewise, the difference in O2 from the end of the rapid phase (i.e., at TD3) and O2base was reported as A'2. The change in O2 from the end of the rapid component and the end of exercise was reported as A'3. An additional measure of the slow component was computed as the difference in O2 between minutes 6 and 3 (Δ O2 6–3). This was used as a comparison to the slow component amplitude as determined from the nonlinear regression model, and it was calculated as the difference in O2 from the means of the last 30 s of exercise and a 15-s equidistant bin around minute 3. Also, an overall mean response time (MRT) was determined by fitting the data to a monoexponential model that began at the onset of exercise and utilized a single amplitude and time constant.

To examine differences in O2 kinetic parameters, the data were analyzed by repeated measures ANOVA with alpha set at 0.05 for all tests of significance. SPSS v11.5 (SPSS, Inc., Chicago, IL) was used for all the data analyses. Data are reported as means ± SE.