Journal Logo

Special Communications: Methods

A new method for the calculation of constant supra-˙VO2peak power outputs


Author Information
Medicine & Science in Sports & Exercise: December 1996 - Volume 28 - Issue 12 - p 1505-1509
  • Free


To examine mechanisms contributing to fatigue during high intensity exercise or to evaluate the influence of a particular intervention on performance or physiological variables, researchers must first select the appropriate physical dimensions of exercise (i.e., intensity, duration, and/or work pattern). This power output will typically be either constant power and time, variable power and time, or a combination of these variables. The choice of intensity is often a key consideration and might be maximal, requiring an all-out effort, or less than maximal but requiring a constant power output(i.e., constant-load). This latter method is often used in research examining the effect of some form of intervention (e.g., training, supplementation) on physiological variables (1,3,6) as the same amount of work is performed on all test occasions (assuming a fixed-time period). Constant-load supra-˙VO2peak tests are also commonly used for estimating the maximal accumulated oxygen deficit or some similar measure of anaerobic capacity either as a single measurement or for comparisons before and after an intervention is applied(9,13,14).

At present there is no method of calculating a constant supra-˙VO2peak intensity that accounts for different aerobic and anaerobic performance abilities. The most common means of locating a supra-˙VO2peak power output is to express the intensity relative to˙VO2peak (e.g.,120% ˙VO2peak). Yet the selection of supra-˙VO2peak power outputs based on the extrapolation of the submaximal relationship between power output and oxygen consumption fails to account for individual differences in anaerobic work capacity. Research has shown that exercise lasting 2 to 3 min may involve a 30% contribution of anaerobic metabolism to the total energy requirements(8,14,17). Moreover, aerobic capacity appears to be negatively correlated with supramaximal performance(4,5,7). Thus, given that the physiological demands and relative effort at, for example, 120% ˙VO2peak, could vary considerably between individuals, one may question whether supra-˙VO2peak power output calculation is best based solely on an aerobically-related variable such as ˙VO2peak.

A standard means of calculating constant supramaximal exercise intensities using some measure of anaerobic work capacity may provide parity in the relative physiological demands (both aerobic and anaerobic) imposed on subjects during high intensity exercise. Reduced between-subject variance in times to exhaustion would reflect a more equitable physiological cost between individuals and allow the influence of an independent variable on an athlete's response to exercise to be more sensitively compared.

The present investigation examined the validity of calculating supramaximal power outputs using submaximal and maximal data in combination with data derived from a 30-s all-out sprint. The purpose of this investigation was to compare the between-subject variance in the times to exhaustion on a cycle ergometer using four different methods of supra-˙VO2peak power output prediction.


Subjects. Ten male students of mixed athletic ability were recruited for this study. Subject characteristics are presented inTable 1. Participation was voluntary and without remuneration. Experimental procedures were approved by The University of Queensland Medical Research Ethics Committee prior to beginning the study. Before giving written consent, all subjects were fully informed of the study's purpose, possible risks, and the benefits of their participation. The study was conducted in accordance with the policy statements of the American College of Sports Medicine.

Pre-experimental testing. In the week prior to completing four rides to exhaustion, subjects undertook two 30-s all-out cycle sprints(separated by a recovery period of 1 h), one incremental submaximal ride, and one incremental ride to volitional fatigue. Subjects completed these tests in the same order and exercise was separated by at least 24 h. The 30-s sprints were performed on a calibrated, multigeared air braked cycle ergometer (South Australian Sports Institute - SASI, Adelaide, Australia) fitted with dropped handlebars, toe-clips/clipless pedals, adjustable seat pillar and head stem, and a racing saddle. The sprints were performed in a gear ratio which elicited 8.87 flywheel revolutions per pedal crank revolution; previous work had shown this gear ratio to optimize both peak power output (PPO) and mean power output(MPO) during a 10-s all-out test (2). Subjects warmed up for 5 min at approximately 150 W during which three short (2-3 s) all-out sprints were performed. Approximately 1 min after the warm-up, the subject positioned the crank in the starting position (45° above the horizontal) and then on the appropriate command (“GO”), pedaled maximally(all-out) for 32 s. The test was extended to 32 s to minimize any pacing in the latter stages which would affect MPO; however, data were collected for the first 30 s only. Software developed by SASI (Cycletest version 3.1b) was used to calculate and record PPO and MPO for each 30-s sprint; sampling occurred at 10 Hz. To ensure that the subject remained seated throughout the sprint, a harness was fitted to his waist and secured to the ground.

The submaximal and maximal rides were performed on an electrically braked cycle ergometer (Excalibur Sports, Lode, Groningen, The Netherlands). This ergometer was also fitted with dropped handlebars, toe-clips/clipless pedals, adjustable seat pillar and head stem, and a racing saddle. The submaximal ride consisted of six stages of 5-min duration; the initial power output was 115 W, and 30 W increments were imposed every 5 min. Cadence was self selected during the submaximal ride with the chosen cadence being repeated during the maximal ride.

The maximal ride consisted of six stages of 5-min duration following which the power output was increased each minute until volitional fatigue. The initial power output in this instance was 100 W and increments were 30 W. During the submaximal and maximal rides, expired air was collected in Douglas bags during the final 2 min of each 5-min submaximal work stage and for at least 2 min prior to volitional fatigue. Inspired volume was calculated by a turbine ventilometer (Morgan Mark 2, Chatham, Kent, England), calibrated prior to each test, and attached to a two-way breathing valve (Hans Rudolph, Kansas City, KA). Expired O2 and CO2 concentrations were measured using Ametek electronic gas analyzers (S-3A/1 and CD3A, Pittsburgh, PA) which were calibrated prior to each test with known gas concentrations (Commonwealth Gases, Brisbane, Australia). Submaximal ˙VO2 values were calculated as the average of the two 1-min submaximal values collected at each 5-min stage, while ˙VO2peak was recorded as the highest˙VO2.

Calculation of power outputs. The ˙VO2 responses to the twelve 5-min submaximal power outputs (obtained from the submaximal and maximal rides) were used to generate the ˙VO2-power regression for each individual. This relationship was then used to determine the power output corresponding to both ˙VO2peak and 120% ˙VO2peak. It is important to note that this method for determining the power output at 100%˙VO2peak differs from that which may define such a power output as the lowest work rate that will elicit ˙VO2peak on a constant-load test (12). The PAS was calculated by subtracting the power output at 100% ˙VO2peak from the higher of the two measures of PPO obtained during the 30-s sprints (Fig. 1). Similarly, the MAS was calculated by subtracting the power output at 100%˙VO2peak from the higher of the two MPO values recorded during the 30-s sprints (Fig. 1). Ten and 20% of the PAS and MAS, respectively, were then added to the power output corresponding to 100%˙VO2peak. We chose 10 and 20% of PAS and MAS, respectively, as pilot work had shown these, when added to 100% ˙VO2peak, would induce exhaustion within 2-3 min of exercise. The mass of each subject wearing shorts and socks was determined and used to calculate the power output equivalent to 6 W·kg-1. The four supra-˙VO2peak power outputs examined were 120% ˙VO2peak, 6 W·kg-1, the power output at ˙VO2peak combined with 10% peak anaerobic scope, and the power output at ˙VO2peak combined with 20% mean anaerobic scope.

Experimental protocol. Each subject performed the four power outputs to exhaustion over a 1-wk period. The order of tests was randomized and each was separated by 24 h of rest. On the days of testing, subjects arrived at the laboratory hydrated and in a 3-h post-absorptive state. They had refrained from vigorous exercise 24 h prior to testing and had replicated their dietary intake in each 24 h pre-exercise period. In addition, each subject's diet was supplemented with a 120-g carbohydrate solution the night prior to each trial and immediately after exercise in an effort to maintain adequate muscle glycogen stores.

All trials were performed on an electrically braked ergometer with the load being applied within 3-5 s. The supra-˙VO2peak rides to exhaustion required the subjects to maintain 90 rpm for as long as possible with exhaustion being marked by a drop in cadence below 50 rpm. They were verbally encouraged to maintain 90 rpm throughout the rides. Subjects remained blind to the power outputs and times to exhaustion throughout the experimental period.

Blood samples (4 ml) were drawn into a heparinized syringe from a 20-gauge catheter (Insyte, Becton Dickinson, Rutherford, NJ) inserted at rest into an antecubital vein. The catheter was kept patent by the infusion of saline (0.9% sodium chloride), while flow was regulated by a three-way stopcock. Samples were obtained at rest, and at 0, 5, and 8 min post-exercise. Of the 4 ml collected, 3.5 ml was immediately expelled into a tube that was centrifuged at 3,000 rpm for 10 min. The supernatant was then drawn and frozen at -15°C for later biochemical analysis of lactate (10) using a Cobas® Mira autoanalyzer (Roche, Basle, Switzerland). Of the 0.5 ml remaining in the syringe, 85 μl was immediately injected into a blood gas analyzer (ABL300, Radiometer, Copenhagen, Denmark) for determination of blood hydrogen ion concentration ([H+]) (15).

Statistical analysis. Bartlett's Test for Homogeneity of Variances was used to determine if there were any significant differences in the variances of the times to exhaustion for the four methods of supramaximal load prediction. The Games and Howell method was then used to test equality of the means. Repeated measures analysis of variance compared the blood data, while the Tukey post-hoc test was used to locate differences. The level of significance was set at P < 0.05.


Performance variables. The highest PPO and MPO for all subjects are presented in Table 2. The mean for the˙VO2-power regressions was represented by a slope of 10.0 ml·min-1·W-1 (± 0.26), a y-intercept 6.79 ml·kg-1·min-1 (± 0.33) and Pearson-product moment correlation coefficient of 0.990.

Times to exhaustion. The mean times to exhaustion were not significantly different between the four methods of power output calculation. The variance in time to exhaustion when the power output was calculated using body weight (i.e., 6 W·kg-1) was greater (P < 0.05) than the remaining three methods of power output calculation(Table 3). Variance in time to exhaustion at the power output equivalent to 120% ˙VO2peak was significantly higher(P = 0.0173) than the variance at 100% ˙VO2peak + 20% MAS(Table 3).

Blood values. Plasma lactate concentrations did not significantly differ between the four methods of power output calculation. However, lactate concentrations at the three post-exercise time points were significantly higher than at rest (P < 0.05), and the concentrations at 5 and 8 min post-exercise were significantly higher than zero min post-exercise(Fig. 2). Values at 5 and 8 min post-exercise were not significantly different from each other.

Blood [H+] did not significantly differ between the four methods of power output calculation. Compared with resting levels, the blood [H+] values were significantly higher (P < 0.05) at all time points post-exercise (Fig. 3). Measures of post-exercise blood[H+] were not significantly different at 0, 5, and 8 min post-exercise.


The major finding of this investigation was that the variance in time to exhaustion at a constant supramaximal power output calculated by adding 20% of MAS to 100% ˙VO2peak was significantly less (P = 0.0173) than the variance in the time to exhaustion when the power output was calculated by extrapolating the submaximal ˙VO2-power relationship to an intensity corresponding to 120% ˙VO2peak. There were no significant differences between the two methods in mean time to exhaustion, plasma lactate concentration, and blood [H+]. No significant difference in the variance in times to exhaustion between 120% ˙VO2peak and 10% PAS + 100% ˙VO2peak was found.

Although calculation of supra-˙VO2peak exercise intensities using measures of submaximal oxygen consumption is common, it fails to account for an individual's anaerobic work capacity. As a result the relative physiological demands between sprint- and endurance-trained subjects at a power output calculated using the ˙VO2-power regression alone could differ, leading to a wide range in times to exhaustion. Antithetically, a power output equivalent to 100% ˙VO2peak in addition to 20% MAS is based not only on the aerobic capacity of the individual but also considers anaerobic work capacity. Hence, the relative physiological stress should be more equitable for all individuals and thus reduce the between-subject variance in times to exhaustion as was evident in the present findings.

A method that relied on a percentage of PAS was examined because PPO can be measured by all-out tests lasting only a few seconds. If both the PAS and MAS methods had proved equally superior to the 120% ˙VO2peak method, then pre-experimental testing need only have included short all-out efforts(e.g., 5-6 s) rather than the longer and more stressful 30-s efforts. This, however, was not the case. The variance in time to exhaustion of the method calculated by the addition of 10% PAS to 100% ˙VO2peak was not significantly different from the variance in time to exhaustion at 120%˙VO2peak. The advantage of the MAS method may be that it is partially based on a measure of anaerobic work capacity, which is important in predicting the ability of an individual to endure an exercise task lasting 2-3 min (8,9).

The method of supramaximal load prediction which relied on body weight(i.e., 6 W·kg-1) yielded the greatest variance in times to exhaustion. This is not surprising as calculating a power output by body weight alone does not account for inter-subject variation in either aerobic or anaerobic work capacities.

The calculation of 100% and 120% ˙VO2peak in the present investigation relied solely on extrapolation of the submaximal˙VO2-power regression. Although this is the conventional method of power output prediction, one could argue that it does not truly represent the respective maximal and supramaximal power outputs during constant-load exercise. Research has shown that during constant-load exercise, power outputs well below 100% ˙VO2peak will result in ˙VO2peak being reached (11,12). This is a result of the slow component of the O2 uptake (16). Thus, during constant-load testing, it may be more correct to refer to the power output of 100% ˙VO2peak as the lowest power output which elicits˙VO2peak.

In conclusion, a method of constant load prediction that incorporates measures of both anaerobic work capacity and aerobic ability (20% MAS + 100%˙VO2peak) reduces the variance in times to exhaustion during a supramaximal exercise task lasting between 2 and 3 min. Methods of supramaximal load prediction that rely solely on submaximal ˙VO2 data may indeed be adequate if the sample population is of similar athletic ability. If, however, the group of subjects is heterogenous, then the use of a power output such as 20% MAS + 100% ˙VO2peak, which relies on the contributions of both the aerobic and anaerobic energy systems, will impose more equitable physiological demands on a group of subjects. It is important to note that the appropriate percentage of MAS will depend on the cycle ergometer and method used to measure and calculate MPO.

Figure 1-Graphic representation of the calculation of peak anaerobic scope (PAS) and mean anaerobic scope (MAS). Mean power output is obtained during a 30-s cycle sprint.
Figure 2-Plasma lactate concentrations [PLa] for the 10 subjects at rest and at 0, 5, and 8 min post-exercise. Values represent averages from the four power outputs; measures were no different across the four protocols.* = significantly less than all others (:
P < 0.05). † = significantly less than 5′ and 8′ post-exercise( P < 0.05).
Figure 3-Measures of blood hydrogen ion concentration [H+] for the ten subjects at rest and at 0, 5, and 8 min post-exercise. Values represent averages from the four power outputs; measures were no different across the four protocols. * = significantly less than all others(:
P < 0.05).


1. Barnett, C., D. L. Costill, M. D. Vukovich, et al. Effect of L-carnitine supplementation on muscle and blood carnitine content and lactate accumulation during high- intensity sprint cycling exercise.Int. J. Sport Nutr. 4:280-288, 1994.
2. Barnett, C., D. G. Jenkins, and L. T. Mackinnon. Relationship between gear ratio and 10 s cycling on an air-braked ergometer.Eur. J. Appl. Physiol. 72:509-514, 1996.
3. Bond, V., R. G. Adams, R. J. Tearney, K. Gresham, and W. Ruff. Effects of active and passive recovery on lactate removal and subsequent isokinetic muscle function. J. Sports Med. Phys. Fitness 31:357-361, 1991.
4. Craig, N. P., K. I. Norton, P. C. Bourdon, et al. Aerobic and anaerobic indices contributing to track endurance cycling performance.Eur. J. Appl. Physiol. 67:150-158, 1993.
5. Crielaard, J. M. and F. Pirnay. Anaerobic and aerobic power of top athletes. Eur. J. Appl. Physiol. 47:295-300, 1981.
6. Daihyuk, C., K. J. Cole, B. H. Goodpaster, W. J. Fink, and D. L. Costill. Effect of passive and active recovery on the resynthesis of muscle glycogen. Med. Sci. Sports Exerc. 26:992-996, 1994.
7. Katch V. L. and A. Weltman. Interrelation between anaerobic power output, anaerobic capacity and aerobic power.Ergonomics 22:325-332, 1979.
8. Medbo, J. I. and I. Tabata. Relative importance of aerobic and anaerobic energy release during short-lasting exhausting bicycle exercise. J. Appl. Physiol. 67:1881-1886, 1989.
9. Medbo, J. I. and I. Tabata. Anaerobic energy release in working muscle during 30 s to 3 min of exhausting cycling. J. Appl. Physiol. 75:1654-1660, 1993.
10. Noll, F. L-(+)-Lactate: determination with LDH, GPT and NAD. In: Methods of Enzymatic Analysis, (2nd Ed.). New York and London: Verlag Chemie, Weinham and Academic Press, Inc., 1976, p. 1475.
11. Poole, D. C., W. Schaffartzik, D. R. Knight, et al. Contribution of exercising legs to the slow component of oxygen uptake kinetics in humans. J. Appl. Physiol. 71:1245-1253, 1991.
12. Poole, D. C., S. A. Ward, G. W. Gardner, and B. J. Whipp. Metabolic and respiratory profile of the upper limit for prolonged exercise in man. Ergonomics 31:1265-1279, 1988.
13. Ramsbottom, R., A. M. Nevill, M. E. Nevill, S. Newport, and C. Williams. Accumulated oxygen deficit and short-distance running performance. J. Sports Sci. 12:447-453, 1994.
14. Scott, C. B., F. B. Roby, T. G. Lohman, and J. C. Bunt. The maximally accumulated oxygen deficit as an indicator of anaerobic capacity. Med. Sci. Sports Exerc. 23:618-624, 1991.
15. Siggaard-Anderson, O. The Acid-Base Status of Blood, 4th Ed. Copenhagen: Munksgaard, 1974, pp. 154-166.
16. Whipp, B. J. The slow component of O2 uptake kinetics during heavy exercise. Med. Sci. Sports Exerc. 26:1319-1326, 1994.
17. Withers, R. T., W. M. Sherman, D. G. Clark, et al. Muscle metabolism during 30, 60 and 90 s of maximal cycling on an air-braked ergometer. Eur. J. Appl. Physiol. 63:354-362, 1991.


©1996The American College of Sports Medicine