The tolerable duration of high-intensity, constant-load exercise is a hyperbolic function of the power, with an asymptote termed the ‘critical power’ (CP) and a curvature constant (W Ì) equivalent to a constant amount of work. Exercise performed in the ‘very heavy’ domain, i.e. above CP, involves a progressive anaerobic contribution to metabolism, the gradual depletion of W Ì and inexorable fatigue. Fukuba & Whipp (1999) suggested that continued exercise after exhaustion (i.e. with W Ì depleted) would only be possible if the power output were reduced to a level below CP, where wholly aerobic energy transfer can occur, permitting W Ì repletion.
We tested this hypothesis on six healthy male volunteers (after informed consent had been provided, as approved by the Local Research Ethics Committee) using cycle ergometry and measuring ventilatory and pulmonary gas exchange variables breath-by-breath (with a turbine and mass spectrometer). Subjects initially performed an incremental work test for the determination of the peak ΩO2 and the estimation of lactate threshold (ã^L). The subjects also undertook three high-intensity square-wave tests to exhaustion to establish their CP and W Ì. Subsequently, the subjects cycled to the limit of tolerance (approximately 6 min) three times at a work rate in the ‘very heavy’ domain each being followed by a reduction of the work rate to: (a) 110 % CP – ‘very heavy’, (b) 90 % CP –‘heavy’ and (c) 80 % ã^L – ‘moderate’ intensity exercise for a 20 min target recovery period. The tests were performed in randomised sequence each on a different day.
When power output was reduced to 110 % CP (▓Dgr│Watts 34 ± 7, mean ± S.D.) exercise was sustained for only 30 s (± 12) beyond the point of initial fatigue, i.e. the point at which the cycling frequency could no longer be maintained above 60 r.p.m. This was equivalent to only 3.5 % (± 1.7) of WÌ. In contrast when the work rate was reduced to 80 % ã^L all subjects, as expected, completed the target 20 min recovery. Interestingly only two subjects completed 20 min at 90 % CP: the remaining four subjects fatigued before the target time (at 785 ± 400 s). Analysis of variance (ANOVA) and post-hoc (Neuman-Keuls) tests showed a significant difference between the recovery times (P < 0.01) at each of the three intensities.
Our results support the suggestions of Fukuba & Whipp (1999) in that, after a bout of fatiguing exercise, it is necessary to reduce the power output to a sub-CP level in order to sustain exercise longer than a ‘few’ seconds. Interestingly however, during recovery into the ‘heavy’ domain while exercise was sustained for over 10 min on average, fatigue occurred despite ΩO2 continuing to decrease, consistent with mechanisms such as CP being reduced consequent to the prior high-intensity bout and/or regional depletion of energy substrates, e.g. muscle glycogen (Bergstrom et al. 1967; Newsholme et al. 1992).