Critical Power Closely Approximates the Power Output at the Estimated Maximal Metabolic Steady State in Trained and Untrained Participants.

Purpose: This study compared estimations of critical power (CP) to maximal metabolic steady state (MMSSest) to see if the differences in the predictions were affected by training status.

Methods: Twelve trained (6 females) and 12 untrained and not experienced with maximal testing (5 females) participants underwent: i) a Step-Ramp-Step (SRS) test to task failure to determine maximal oxygen consumption (V̇O2max) and peak power output (POpeak); ii) 4-5 time to task failure (TTF) trials at average power outputs (PO) ranging from 70 to 90% of POpeak for CP estimations; iii) 2-3 30-min constant PO rides to establish MMSSest as the highest PO at which oxygen consumption (V̇O2) and blood lactate concentrations ([La-]b) are stable.

Results: The PO associated with CP was significantly greater than that associated with MMSSest in both untrained (155 ± 39 W vs. 147 ± 34 W, respectively) and trained (233 ± 37 W vs. 225 ± 39 W, respectively) individuals (p < 0.001). Both the untrained and trained groups displayed a similar and significant bias for MMSSest compared to CP (i.e., 7.5 W; p < 0.05), with 95% limits of agreement from -13 to 28 W, and -11 to 26 W for untrained and trained, respectively.

Conclusions: These findings indicate that, despite a significant (albeit small) difference between CP and MMSSest, the CP model provided a close approximation of the PO associated with MMSSest in both untrained and trained participants, as the difference in PO was within the expected measurement error. Therefore, our results showed that, despite some small discrepancies between groups, the CP model fitting was not affected by training status and that previous testing experience with highly demanding exercise is not a key component of the quality of the prediction model.