Purpose: This study examined the partial reconstitution of the work capacity above critical power (W') following successive bouts of maximal exercise using a new repeated ramp test, against which the fit of an existing W' balance ( ) prediction model was tested.
Methods: Twenty active adults, consisting of trained cyclists (n = 9; age 43 [15] y, 61.9 [8.5] mL·kg-1·min-1) and untrained cyclists (n = 11; age 36 [15] y, 52.4 [5.8] mL·kg-1·min-1) performed 2 tests 2 to 4 d apart, consisting of 3 incremental ramps (20 W·min-1) to exhaustion interspersed with 2-min recoveries.
Results: Intratrial differences between recoveries demonstrated significant reductions in the amount of W' reconstituted for the group and both subsets (P < .05). The observed minimal detectable changes of 475 J (first recovery) and 368 J (second recovery) can be used to monitor changes in the rate of W' reconstitution in individual trained cyclists. Intertrial relative reliability of W' reconstitution was evaluated by intraclass correlation coefficients for the group (≥.859) and the trained (≥.940) and untrained (≥.768) subsets. Absolute reliability was evaluated with typical error (TE) and coefficient of variation (CV) for the group (TE ≤ 559 J, CV ≤ 9.2%), trained (TE ≤ 301 J, CV ≤ 4.7%), and untrained (TE ≤ 720 J, CV ≤ 12.4%).
Conclusions: The reconstitution of W' is subject to a fatiguing effect hitherto unaccounted for in prediction models. Furthermore, the model did not provide a good fit for the repeated ramp test, which itself proved to be a reliable test protocol.
Keywords: critical power; fatigue; modeling; testing.