Eur J Appl Physiol (1999) 79: 290 293

Springer-Verlag 1999

SHORT COMMUNICATION

D.W. Hill C.S. Ferguson

A physiological description of critical velocity

Accepted: 24 July 1998

Abstract Although critical velocity (CV) provides a valid index of aerobic function, the physiological signicance of CV is not known. Twelve individuals performed exhaustive runs at 95% to 110% of the velocity at which VO2max was attained in an incremental test. 2max was elicited in each run. Using the time to exVO haustion at each velocity, CV was calculated for each participant. Using the time to achieve VO2max at each velocity, which was shorter at higher velocities, a parameter we have designated as CV was calculated for each participant. During exercise at or below CV, VO2max cannot be elicited. CV (238 24 m minA1) and CV (239 25 m minA1) were equal (t 0.60, p 0.56) and correlated (r 0.97, p < 0.01). These results demonstrate that CV is the threshold intensity above which exercise of sucient duration will lead to attainment of VO2max. Key words Critical power Slow component O2max

Introduction
In cycle ergometer exercise, the relationship between power and time to exhaustion (TTE) is described by a hyperbolic model (Moritani et al 1981). This is the basis of the critical power concept (Hill 1993). For running, the relationship between intensity (i.e., velocity) and TTE is also described by a hyperbolic model (Hughson et al. 1984, Housh et al. 1992, Pepper et al. 1992). TTE is a function of anaerobic work capacity (AWC) and critical velocity (CV) in the form TTE AWC velocity CV1 1X

CV (or CP) is related to indices of aerobic tness, such as the anaerobic threshold or VO2max (Moritani et al. 1981, Housh et al. 1991). However, the physiological basis for CV or CP is not known. Only Poole and colleagues (1988, 1990) have attempted to elucidate a basis, and their hypothesis has been presented in a recent review (Gaesser & Poole 1996). At the onset of exercise below the lactate threshold, VO2 increases rapidly and then levels o (Whipp & Wasserman 1972). Above the lactate threshold, the rapid initial increase in VO2 is followed by a second ``slow'' component which results in a delayed steady state (Henson et al. 1989). Gaesser and Poole (1996) proposed that, during exercise above CP, the slow component of the VO2 response will drive VO2, not to a delayed steady state but, to VO2max. Clearly, VO2max can be elicited during constant power exercise at a range of intensities (Gaesser & Poole 1996, Sloniger et al. 1996). Evidence that VO2max will be attained only during exercise above CP was provided by Poole et al. (1988, 1990) who showed that, before and after training, during exercise at CP, VO2 reached a steady state at $75% of VO2max. But, during exhaustive exercise at power outputs that were only 8 to 11% (16 to 23 W) above CP ($75% of the work rate associated with attainment of VO2max in an incremental test), VO2 reached VO2max. We hypothesized that, just as the relationship between intensity and TTE can be t to the hyperbolic model, so can the relationship between intensity and time to achieve VO2max (TTmax). In this case, the relationship is TTmax AWCH velocity CVH
1

2X

D.W. Hill (8) C.S. Ferguson Department of Kinesiology, University of North Texas PO Box 311337, Denton, TX 76203-1337 USA Tel: +940-565-2252 Fax: +940-565-4904 e-mail: dhill@coefs.coe.unt.edu

CV is the highest submaximal exercise intensity that can be sustained without eliciting VO2max and AWC is the amount of work performed anaerobically before VO2max is attained. The rst purpose of this study was to test the hy pothesis that VO2max would be achieved faster at faster running velocities and that the relationship between

291

intensity and TTmax could be described by a hyperbolic model. The second purpose was to compare CV and CV.

and distance CV TTE AWC 4X Parameters were derived using iterative nonlinear regression or linear regression on SPSS (Chicago, IL, USA). For each participant, regressions were performed using all four velocities and all possible combinations of three velocities. The data set for which there was the least variability among the CV estimates generated by the three models was selected (see Hill, 1993). The estimate generated using the nonlinear velocity-time model with this data set was recorded as the criterion measure of CV. Velocity and TTmax data were also t to three models. There was a hyperbolic model (equation 2), and two linear models velocity AWCH TTmax1 CVH and distance CVH TTmax AWCH 6X Estimates of AWC and CV were derived as described for AWC and CV. Statistical analyses Responses to exercise at 95%, 100%, 105%, and 110% of Vmax were compared using a repeated measures analysis of variance (ANOVA) with signicance accepted at the 0.05 level. The significance of the linear eect of velocity was tested at the 0.05 level. Post hoc comparisons of least square adjusted means (``contrasts'') were performed with signicance accepted at the 0.01 level. For VO2max, a one-way ANOVA was also carried out over values from ve tests, the four constant velocity tests plus the incremental test. To evaluate the physiological signicance of CV, CV was compared with CV using a paired-means t-test and a Pearson product moment correlation. 5X

Methods
Participants Twelve individuals volunteered to participate in the study which had been approved by the Institutional Review Board at the University of North Texas. There were 5 men, of mean ( SD) age, 33 7 year, height, 179 5 cm, and weight 69.7 7.8 kg, and there were 7 women, of mean age, 24 3, height, 166 4 cm, and weight 60.5 5.0 kg. All were involved in personal tness programs and were running 40 km wkA1, and all consented to adjust their training in order to be ``well rested'' for the laboratory testing. Overview Participants performed six exhaustive treadmill tests in a climatecontrolled laboratory (20 to 22C). The purpose of the rst test session was to determine VO2max and the velocity at which VO2max was elicited (Vmax). At the next ve sessions, constant velocity tests were performed at dierent percentages of Vmax. During each test, expired gases were analyzed using a MedGraphics (St. Paul, Minnesota, USA) CPX metabolic cart. The cart was calibrated before each test using gases of known O2 and CO2 concentration and a syringe of known volume. Calibration was veried after the tests. Sessions were separated by at least 24 h, and were scheduled at the same time of day. Using TTE and TTmax at each velocity, values for CV and CV were derived. Determination of VO2max and Vmax Velocities of the rst two stages of the incremental test diered by 40 m minA1 (1.5 km hrA1) and were individually selected to be below each participant's reported training paces for distance runs. After two 5-min stages, each followed by a 5-min recovery, there were 2-min stages with increments of 21 m minA1 (0.8 km hrA1) and no rest periods. VO2max was the highest 30-s average determined from rolling averages of 15-s means. Vmax was the velocity at which VO2max was rst attained. Constant velocity tests Participants performed exhaustive constant velocity tests at 100% of Vmax (practice trial, results not used in analyses), and 110%, 105%, 95%, and 100% of Vmax. Tests were preceded by 6-min warmups at velocities below the participants' individually-determined ventilatory anaerobic threshold, and a 5-min recovery. Par ticipants received verbal encouragement during the tests. VO2max was determined as above. TTmax was determined to the nearest second as the time from the onset of exercise until the middle of the rst 15-s period when VO2 reached or exceeded the 30-s value for VO2max achieved in that test. Participants were not given access to any information about their elapsed time or any physiological responses until all data collection had been completed. Derivation of parameter estimates Estimates of AWC and CV were generated using the three dierent forms of the velocity-time relationship. The three models are mathematically equivalent forms of the hyperbolic function (equation 1), which may be rewritten as two linear relationships velocity AWC TTE1 CV 3

Results
The mean ( SD) value for VO2max determined in the incremental tests was 3.39 0.93 l minA1 A1 A1 (52.7 ml kg min ). Vmax was 270 23 m minA1 (16.2 km hrA1). Responses in the constant velocity tests are in Table 1. Results of the ANOVA and contrasts revealed that VO2max in these tests were not dierent from the value in the incremental tests. Results of the ANOVA and test for linear eects revealed that TTE, TTmax, and the time that VO2max was sustained (TAmax) were longer in the lower velocity tests. The mean CV was 238 24 m minA1 A1 (14.3 km hr ), which was 88% of Vmax. The mean CV was 239 25 m minA1 (14.5 km hrA1), or 89%
Table 1 Responses in the constant velocity tests Velocity expressed as a percentage of Vmax 95% TTE (s) TTmax TAmax VO2max 468a 241 367a 200 301a 114 3.41a 0.95 100% 290b 247b 44ab 3.49a 105% 143 189bc 113 154bc 45 35ab 0.92 3.43a 110% 58 129c 37 66 110c 29 27 19b 20 0.90 3.32a 0.89

Note: for each variable, means with the same superscript were not dierent

292

of Vmax. CV and CV were not dierent (t 0.60, p 0.56) and they were highly correlated (r 0.97, p < 0.01). Values for AWC (mean, 124 43 m) tended to be higher (p 0.07) than values for AWC (mean, 110 38 m), and the two were correlated (r 0.71, p < 0.01).

Discussion
The rst purpose of this study was to determine if the relationship between running velocity and TTmax could be described by a hyperbolic function. We found that (a) VO2max was elicited during exercise at 95% to 110% of Vmax, (b) TTE, TTmax, and TAmax were shorter at higher velocities, and (c) the relationship between velocity and TTmax could, indeed, be described by a hyperbolic function. Previously, much of the study about the eect of exercise intensity on VO2 kinetics has focussed on differences during exercise below or above the lactate threshold, not on the eects of exercise intensity per se. We found an eect of velocity on the rate of response of aerobic energy production over a narrow range of supra lactate threshold intensities. VO2max was always achieved, but faster at higher velocities. As in other studies that have involved running exercise (Hughson et al. 1984, Housh et al. 1992, Pepper et al. 1992), the relationship between velocity and TTE was well described by equations 1, 3, and 4. For each model, R2 averaged 0.965 and SEE averaged 5 m minA1 (2% of the mean). Mean values obtained using each model were 238 m minA1, 239 m minA1, and 240 m minA1. The SEE and R2 were similar to previously reported values (Poole et al. 1988, Housh et al. 1989, 1991, 1992, Jenkins & Quigley 1990). We also found a relationship between velocity and TTmax. Data were t using equations 2, 5, and 6. For each model, the R2 averaged 0.955 and the SEE averaged 5 m minA1 (2% of the mean). Mean values obtained using the three models were 239 m minA1, 240 m minA1, and 242 m minA1. Based on the low SEE and the high R2 for each model, and on the agreement between estimates generated by the three models, it was concluded that the relationship between velocity and time to achieve VO2max could be described well by a hyperbolic model (and the two linear forms of the relationship). The second purpose of this study was to investigate the physiological importance of CV. To date, CV has been determined, dened, and described based purely on statistical evaluation of the velocity-time relationship, and only one group of investigators has attempted to provide a physiological basis for CV. Poole et al. (1988, 1990) hypothesized that CV (CP in their study) repre sented the threshold intensity above which VO2 would gradually increase to VO2max. We calculated the value of CV and also the value of CV, which theoretically represents the velocity at or

below which VO2max cannot be elicited. CV was equal to CV. Thus, the results of the present study suggest that, during exercise at an intensity above CV, VO2 can gradually increase to VO2max. This supports the hypothesis of Poole et al. (1988, 1990) that CV is the highest velocity at which VO2, and presumably other metabolic indices, such as blood lactate concentration, lactate-to-pyruvate ratio, bicarbonate concentration, and blood pH, can achieve a steady-state. As such, CV represents an upper limit for sustainable power during running. Gaesser and Poole (1996) have dened the ``severe'' exercise domain as intensities associated with attainment of VO2max. Thus, they proposed that CP (or CV) demarcates the lower bound of the severe intensity exercise domain. However, they did not provide an upper bound. An upper bound must occur at the intersection of the velocity-TTE and velocity-TTmax curves. Further research is needed to accurately describe responses to exercise within the severe intensity domain and to identify the bounds to the domain. In addition, study is needed to determine the signicance of AWC, to determine if CP and CV really are interchangeable parameters as we have assumed, and to evaluate responses at CP (or CV) to directly ascertain if, as suggested by the results of this study and studies by Poole et al. (1988, 1990), CP (or CV) is, in fact, the threshold intensity above which VO2max can be elicited. In summary, we have shown that VO2max can be achieved over a range of running velocities, it is achieved faster at higher velocities, and the relationship between velocity and TTmax is described by a hyperbolic function. Second, we have investigated the CV parameter, a proxy for CP when the exercise mode is running. CV was equal to CV. Thus, CV represents the threshold velocity above which VO2max can be elicited during exercise of sucient duration.

References
Gaesser GA, Poole DC (1996) The slow component of oxygen uptake kinetics in humans. In: Holloszy JO (Ed), Exercise and sport sciences reviews. Williams & Wilkins, Baltimore, MD, pp 3570 Henson LC, Poole DC, Whipp BJ (1989) Fitness as a determinant of oxygen uptake response to constant-load exercise. Eur J Appl Physiol 59:2128 Hill DW (1993) The critical power concept: A review. Sports Med 16:237254 Housh DJ, Housh TJ, Bauge SM (1989) The accuracy of the critical power test for predicting time to exhaustion during cycle ergometry. Ergonomics 32:9971004 Housh TJ, deVries HA, Housh DJ, Tichy MW, Smyth KD, Tichy AM (1991) The relationship between critical power and the onset of blood lactate accumulation. J Sports Med Phys Fit 31:3136 Housh TJ, Johnson GO, McDowell SL, Housh DJ, Pepper ML (1992) The relationship between anaerobic running capacity and peak plasma lactate. J Sports Med Phys Fit 32:117122 Hughson RL, Orok CJ, Staudt LE (1984) A high velocity treadmill running test to assess endurance running potential. Int J Sports Med 5:2324

293 Jenkins DG, Quigley BM (1990) Blood lactate in trained cyclists during cycle ergometry at critical power. Eur J Appl Physiol 61:278283 Moritani T, Nagata A, deVries HA, Muro M (1981) Critical power as a measure of physical work capacity and anaerobic threshold. Ergonomics 24:339350 Pepper ML, Housh TJ, Johnson GO (1992) The accuracy of the critical velocity test for predicting time to exhaustion during treadmill running. Int J Sports Med 13:121124 Poole DC, Ward SA, Gardner G, Whipp BJ (1988) A metabolic and respiratory prole for prolonged exercise in man. Ergonomics 31:12651279 Poole DC, Ward SA, Whipp BJ (1990) The eects of training on the metabolic and respiratory prole of high-intensity cycle ergometer exercise. Eur J Appl Physiol 59:421429 Sloniger MA, Cureton KJ, Carrasco DI, Prior BM, Rowe DA, Thompson RW (1996) Eects of the slow-component rise in oxygen uptake on VO2max. Med Sci Sports Exerc 28:7278 Whipp BJ, Wasserman K (1972) Oxygen uptake kinetics for various intensities of constant-load work. J Appl Physiol 33:351 356