In endurance sport, exercise intensity has traditionally been prescribed using intensity zones on the basis of discrete heart rate and lactate concentrations associated with variations in work rate (power). Recently, bicycle-mounted power meters capable of measuring power output in the field have become readily available and increasingly economical. Training with a power meter can be performed using power-based intensity zones targeting specific energy systems in lieu of heart rate or lactate concentration (1). Although most coaches create five or more exercise intensity zones, the fundamental basis of these zones can be traced back to the three exercise intensity domains of moderate, heavy, and severe, identified using measures of oxygen consumption (V˙O2) and blood lactate concentration (10,16,17). Work rates in the moderate domain are defined as those at which V˙O2 reaches a steady state within approximately 3 min after the initiation of exercise without a sustained rise in blood lactate concentration (10). The heavy domain begins with work rates at which blood lactate production begins to exceed its rate of removal (lactate threshold) and ends at the highest work rate where blood lactate concentration can be stabilized (maximum lactate steady state (MLSS)) (3,10). Finally, work rates performed in the severe domain will induce a continuous rise in both V˙O2 and blood lactate concentration until V˙O2 reaches V˙O2max and eventual fatigue follows (10).
A 3-min all-out cycling test (3MT) (7,18) has been proposed to identify the transition between the heavy and the severe exercise domains (10,16,17) by solving for the constants of the power-time relationship shown in equation 1 (7,12,16,18).
Power limit (Plim, in watts) is the maximum power output sustainable over a given duration (tlim, in seconds). Critical power (CP, in watts) represents the asymptote of the power-duration relationship (12,14). W′ (in joules) is a measure of energy production primarily related to the immediate energy stores available to the working muscles (12,14). Plim approaches CP as tlim approaches infinity, which is the argument for CP being a sustainable power output. Alternatively, if W′ is reduced to zero, the maximum power output for any tlim would be CP. Hypothetically, if there was a method to expend all immediate energy stores, W′ would reduce to zero, and any sustained power output would be equivalent to CP. In the 3MT, subjects produce an all-out effort for the entire 3 min to expend the immediate energy stores, minimizing the work contribution related to W′. In theory, if this effort reduces W′ to zero, an aerobically sustainable power output should be observable that is equivalent to CP. The mean power output over the last 30 s of the 3MT (end-test power (EP)) has been shown to approximate power output at MLSS and CP (7,18). Typical tests for both MLSS power output and CP require a series of efforts performed over multiple days (3,12), whereas the EP of the 3MT has been shown to provide a convenient single session estimate of these measures (7,18).
One purpose of this study was to develop a modified version of the 3MT performed on the subjects' own bicycles equipped with a power meter and mounted to a progressive resistance trainer. In this configuration, the workload is based on the speed of the bicycle's rear wheel, allowing subjects to vary workload primarily by gear selection and secondarily by slight variations in pedal cadence. In comparison with the established test protocol (7), our modified 3MT provides a more natural test for experienced cyclists, minimizes the relationship between pedal cadence and ergometer load, which has been shown to influence the EP (19), and uses equipment readily available to the cycling community.
A second purpose of this study was to identify the established exercise intensity domains using the EP of the 3MT. Power outputs at common markers of exercise intensity (lactate threshold, ventilatory threshold, onset of blood lactate accumulation, and V˙O2peak) were assessed to help define where EP lies in the exercise intensity spectrum. The demarcation between moderate and heavy exercise intensity domains has been typically established using lactate or ventilatory thresholds (16), whereas the 3MT has been shown to provide an estimate of power at the heavy to severe transition (7,18). On the basis of previous findings, we hypothesized that the proposed 3MT could be used as a single session test to establish the transitions between both the moderate to heavy and the heavy to severe exercise intensity domains.
Fourteen male and two female licensed competitive road cyclists (mean ± SD: age = 32.4 ± 8.7 yr, height = 176.4 ± 9.8 cm, mass = 73.6 ± 11.3 kg, racing experience = 8.3 ± 6.5 yr, V˙O2peak = 60.3 ± 8.3 mL·kg−1·min−1) participated in this study. The University of New Hampshire's institutional review board approved the use of human subjects, and written informed consent of each subject was obtained before their participation.
Subjects visited the laboratory three times, with a minimum of 48 h of recovery between each session. All three visits occurred within a 14-d period. The day before each test, subjects were asked to refrain from heavy training. On test days, before each test, subjects were asked to refrain from training, be adequately hydrated, and not have consumed food or caffeine for 3 h. Between test days, subjects were asked to maintain their existing training and nutrition schedule. Visit 1 consisted of an incremental load test for the measurement of V˙O2peak and lactate threshold (see next sections). A familiarization trial for the 3MT was performed during visit 2. Although all subjects were experienced cyclists, a continuous all-out effort for 3 min is an exercise modality many cyclists are not familiar with. In performing the 3MT, the maximal effort must continue throughout the test, even as the subjects fatigue and work rates fall. Because of the high level of motivation required to properly perform the 3MT, subjects were not informed that the first trial was for familiarization purposes only. A final 3MT was performed during visit 3 (see next sections).
Subjects used their own bicycles during the tests. Bicycles were fitted with a laboratory provided rear wheel containing a power-measuring hub and compatible bicycle computer (PowerTap SL 2.4 Wireless 7205S; Saris Cycling Group, Inc., Madison, WI). To minimize problems with drive train compatibility, subjects' own rear gears were removed from their rear wheels and used on the power-measuring wheel. The bicycle computer recorded power output and cadence at a 1-Hz sampling rate. Power output and cadence data were downloaded from the bicycle computer and saved on a personal computer using power meter analysis software (WKO+; Peaksware, LLC, Lafayette, CO). During the incremental load test, subjects' bicycles were mounted to an electronically controlled bicycle ergometer (CompuTrainer Pro Model 8001; RacerMate Inc., Seattle, WA). The 3MT was performed with subjects' bicycles mounted to a progressive resistance wind trainer (Wind 9202, Saris Cycling Group, Inc.).
An L-lactate analyzer (YSI 1500 Sport; YSI Inc., Yellow Springs, OH) was used to immediately analyze capillary blood samples (duplicates) taken from the subjects' fingertips. During all tests, subjects wore a mouth breathing face mask and head cap (7930 series; Hans Rudolph Inc., Kansas City, MO). A flow sensor (Vmax Sensor 770279; SensorMedics Corp., Yorba Linda, CA) inserted into the face mask was connected to an indirect calorimeter (Vmax 229LV Lite, SensorMedics Corp.) for breath-by-breath analysis of gas exchange and flow, which were subsequently averaged over each 20-s period.
Incremental load test.
Subjects began a 15-min warm-up at 100 W. Gear ratios were selected such that at the subjects' preferred pedal cadences, the speeds of the rear wheels were greater than 32 kph. These gear ratios, once set during warm-up, were not adjusted further. Immediately after the warm-up and every 4 min thereafter, the ergometer load was increased by 25 W. Subjects received verbal encouragement and continued until volitional termination, at which the ergometer load was reduced to 50 W, and subjects continued pedaling for active recovery until blood lactate concentration began to recover (see below).
Blood samples were taken at rest, during the last 2 min of warm-up, and during the last minute of each 4-min work interval. To identify peak lactate concentration, blood samples were taken upon volitional termination and every 2 min during the active recovery. Active recovery continued until blood lactate concentration decreased for two consecutive 2-min intervals. V˙O2peak was recorded as the maximum average V˙O2 measured during any 20-s period. Power output at V˙O2peak was established as the mean power observed during the stage at which V˙O2peak was recorded. [BLa]peak was the maximum blood lactate concentration.
Determination of threshold power outputs.
A third-order polynomial was fit to the individual's blood lactate concentration versus stage power data using the method of least squares. Gas exchange data were averaged across the last minute of each 4-min stage. Mathematical analysis software (Matlab; The MathWorks, Inc., Natick, MA) was used to determine the power outputs at lactate and ventilatory thresholds according to the following parameters:
1. Lactate threshold: The point at which there was a 1-mmol·L−1 increase in blood lactate concentration above exercise baseline, where exercise baseline was defined as a horizontal line with a Y-intercept calculated as the average blood lactate concentration of the first three stages (15).
2. Onset of blood lactate accumulation (OBLA): The point at which blood lactate concentration was 4 mmol·L−1 (11).
3. Ventilatory threshold: determined using the V-slope method (21).
3MT and determination of EP.
The initiation of the 3MT begins with an all-out sprint-like effort. Before the 3MT, subjects identified gear ratios appropriate for the starting effort. During the 3MT, subjects were then able to self-select the work rate by gear selection and variations in pedal cadence. Similar to the methods of Burnley et al. (7), the subjects performed a 5-min warm-up at 100 W followed by 5 min of rest. The original and modified 3MT provide minimal warm-up to minimize the impact on W′ (equation 1). During rest, the subjects switched to the gear ratios preselected for the start of the all-out effort, and calibration of the power-measuring wheel was verified.
Immediately after the resting phase, the 3MT began with subjects pedaling to accelerate as fast as possible, shifting gears or pedaling cadence as necessary to obtain the highest possible work rate throughout the test. The subjects received consistent verbal encouragement but were not informed of the elapsed time or power output to minimize pacing of effort. At the end of 3 min, subjects performed an active recovery at 50 W. EP was determined by averaging power output over the last 30 s of the 3MT (Fig. 1A).
Statistical analyses were performed using SPSS Version 17.0 (SPSS Inc., Chicago, IL). One-way repeated-measures ANOVA was used to determine significant differences between EP, power outputs at lactate threshold, OBLA, ventilatory threshold, and V˙O2peak as well as differences between measures of V˙O2peak and [BLa]peak between the incremental load test and the 3MT. Main effects were compared using the least significant difference test. Independent sample t-tests were performed to determine significant differences between the mean power outputs over 30-s intervals during 3MT, the V˙O2 at peak and the end of the 3MT and to compare mean pedal cadences from the incremental load test to those observed during the 3MT. Bivariate correlation analysis was performed between EP and power outputs at lactate threshold, OBLA, ventilatory threshold, and V˙O2peak. Linear regression through the origin was used to provide estimates for power outputs at lactate threshold, OBLA, ventilatory threshold, and V˙O2peak using EP. For all tests, statistical significance was accepted at the P < 0.05 level.
V˙O2peak was similar between the incremental load test and the 3MT (60.3 ± 8.3 vs 59.6 ± 8.0 mL·kg−1·min−1, respectively, P = 0.56). Absolute V˙O2peak measures during the incremental load test and the 3MT were 4.43 ± 0.88 and 4.36 ± 0.76 L·min−1, respectively. [BLa]peak during the incremental load test was significantly lower than the [BLa]peak measured during the 3MT (10.1 ± 2.3 vs 15.6 ± 1.8 mmol·L−1, respectively, P < 0.001).
The group mean power output profile across the 3MT is shown in Figure 1A. To characterize the work rate trend, mean powers over each 30-s interval of the 3MT were compared (Fig. 1B). Although the mean power output significantly declined over the first 90 s of the 3MT, power output leveled off over the last 90 s. V˙O2peak during the 3MT was attained at a test time of 128 ± 41 s and was not significantly different than V˙O2 over the last 20 s of the 3MT (57.0 ± 7.2 mL·kg−1·min−1).
Pedal cadence comparisons.
Mean pedal cadence during the incremental load test was similar to the mean pedal cadence during the last 30 s of the 3MT (92 ± 5 vs 94 ± 8 rpm, respectively, P = 0.360). However, mean pedal cadence throughout the 3MT (98 ± 12 rpm) was significantly higher than both the cadences measured during the incremental load test (P < 0.01) and the last 30 s of the 3MT (P < 0.001). Peak cadence during the 3MT was 128 ± 10 rpm.
A one-way repeated-measures ANOVA indicated a significant effect between power outputs at lactate threshold, OBLA, ventilatory threshold, V˙O2peak, and EP (P < 0.001). Mean power output data and pairwise comparisons for significant effects between the five power output measures are shown in Table 1. Ventilatory threshold and OBLA power outputs were significantly greater than lactate threshold power output. EP was significantly greater than the three threshold power measures but was significantly less than power output at V˙O2peak. EP was significantly correlated with all power output measures (Table 1). EP was found to be 95% ± 5% of power output at V˙O2peak and 121% ± 18% of power output at ventilatory threshold. Using linear regression through the origin, percentages of EP were determined that could be used to estimate other power outputs and are presented along with standard errors of the estimate and actual power outputs in Table 1. The actual versus predicted power outputs are illustrated in Figures 2A-D.
The study revealed two key findings. First, it introduced a modified version of the 3MT (7,18) that used equipment readily available to cyclists and one that allowed participants to select their own pedal cadences throughout the duration of the test. Second, it showed that the mean power over the last 30 s of the 3MT (EP) could be used to predict power output at lactate threshold, the demarcation between moderate and heavy exercise intensity domains. Because the prior research had shown EP could be used to establish the transition from heavy to severe exercise intensity domains (7,18), these results extend the utility of the 3MT, such that the transitions between the moderate to heavy and heavy to severe domains could be established with a single session test.
The original 3MT is performed on relatively expensive laboratory grade equipment (7), where the modified 3MT can be performed using any device capable of measuring variations in power output while cycling. To assign the load to the rear wheel, a progressive resistance wind trainer was chosen primarily on the basis of mechanical simplicity and secondarily on cost. The trainer workload was established by the speed of the bicycle's rear wheel, which allowed subjects to vary workload naturally by gear selection and slight variations in pedal cadence. Subjects were able to self-select pedal cadences throughout the 3-min duration, whereas in the original 3MT, workload was mathematically related to pedal cadence (7).
Cyclists in training and competition normally select pedal cadences near 90 rpm (13). Measures of W′ and EP have been shown to be lower when subjects performed the original 3MT using a higher than preferred cadence (19). Peak cadence measured during the original 3MT was 150 ± 14 rpm (7), approximately 17% higher than our observed peak cadence of 128 ± 10 rpm. In addition, mean cadence over the end of the 3MT in the original work was 88 ± 5 rpm (7), in contrast to our observed mean of 94 ± 8 rpm. Our study observed lower peak yet higher end cadence, limiting the extremes of cadence to what might more normally be observed in training and competition.
Although it should be pointed out that this study did not specifically validate the modified 3MT methodologies against those of the prior research (7,18), the general performance parameters between the original and the modified 3MT can be compared to establish some level of confidence. The mean power profile of the modified 3MT (Fig. 1A) shows the initial high work rate depleting W′, eventually leading to a work rate leveling identified as EP, and is similar to that reported in the prior research (7). The EP of the current study (273 ± 52 W) was similar to that of the original work (257 ± 49 W) while using comparatively similar subjects (7).
Interesting differences are noted between studies when EP is reported as a percentage of either the power output at V˙O2peak or the power output at ventilatory threshold. In the current study, EP was 95% ± 5% of the power output at V˙O2peak, whereas in the original work, it was 70% ± 4% of the power output at V˙O2peak. Also different, EP was 121% ± 18% of the power output at ventilatory threshold in this study and was 154% ± 25% of ventilatory threshold power in the original work. These differences can simply be explained by a lower power output at V˙O2peak (288 ± 56 vs 368 ± 73 W) and a higher power output at the ventilatory threshold (232 ± 64 vs 169 ± 55 W) in this research, likely because of methodological differences between studies. For example, the original work established these measures using a ramp protocol with an increase in work rate of 30 W·min−1, which resulted in an estimated time of 12 min to volitional termination (7,18). In contrast, the protocol used in this study started at 100 W and increased the work rate by 25 W every 4 min with an estimated time to volitional termination of 30 min. The longer duration of each stage in our protocol allowed the blood lactate concentration to stabilize but resulted in a greater influence of the slow component of V˙O2 (10) such that V˙O2peak was achieved at a lower work rate (4).
The lower ratio of EP to power output at ventilatory threshold observed in this study may be due to differences in either protocol or subject conditioning. Both studies used the V-slope method to find power output at ventilatory threshold, but because of the incremental step nature of our protocol, there would be some loss of resolution compared with a constant ramp test. The subjects in this study exhibited a V˙O2peak approximately 15% higher than the original study's subjects (4.43 ± 0.88 vs 3.84 ± 0.79 L·min−1), also indicating training state as a contributing factor for the difference in the ratio of EP to power output at ventilatory threshold. The power output at ventilatory threshold in this study was 80% of the power at V˙O2peak, which is typical for competitive cyclists (8), whereas the power output at ventilatory threshold in the original work was lower than expected (46% of power output at V˙O2peak).
Defining intensity domains.
The heavy to severe demarcation has already been established as EP itself (7,18). The demarcation between moderate to heavy exercise intensity domains has been identified as the "lowest work rate at which blood lactate appearance exceeds its rate of removal" (10), typically established using lactate or ventilatory thresholds (10,16). Across subjects, lactate threshold power output was consistently a percentage of EP, making these two variables highly correlated (Table 1). EP is sustainable for nearly 30 min (7), whereas intensities near lactate and ventilatory thresholds can be maintained from 60 to 180 min (2,5,9) such that lactate and ventilatory power outputs should be less than EP. As a result, a percentage of EP could be used to approximate the demarcation between moderate and heavy exercise intensity domains.
After performing the 3MT, estimates to define the three exercise intensity domains may be calculated from EP. To simplify estimate calculations, the linear regression between these variables was constrained to pass through the origin. The heavy to severe demarcation is established at 100% of EP, whereas the moderate to heavy demarcation can be approximated at 76% of EP (Table 1). The estimate for this transition (power output at lactate threshold) occurred at 72% of the power output at V˙O2peak, similar to what is typically observed in trained cyclists (8). A prediction error of ±28 W (approximately ±15%) in determining the power output at the moderate to heavy demarcation suggests this technique is adequate for applications that do not require exact determination of the transition. When greater levels of precision are required, the use of lactate or ventilatory threshold may be more appropriate.
One application of the 3MT is the use as a benchmark test to monitor adaptations to training in a controlled environment. For example, tracking changes in EP over time has been shown to reflect training induced changes in critical power (20). Another application is to create power-duration-based training intensity zones using EP. Measures examined in this study have been previously associated with durations of sustainable power output with the power output at lactate threshold being sustainable for 3 h or more (9); power outputs at ventilatory threshold and OBLA, between 60 and 90 min (2,5); EP, 30 min (7); and power output at V˙O2peak, 4 min (6). Using the estimates between EP and these measures (Table 1), the 3MT could be used to create training intensity zones as shown in Figure 3. Because constant monitoring of actual performance is one of the benefits of on-bike power output measurement, it should be noted that estimates from any test used to predict power-duration relationships may become less relevant once actual performance data have been obtained (e.g., through races, time trials, etc.).
The most prominent limitation of the current study was that the EP of the modified 3MT was not validated against the critical power, the power at MLSS, or the original 3MT. Subjects were deliberately chosen from a relatively homogeneous population of experienced cyclists; findings may have limited applications outside of this group and need to be validated with a larger, more heterogeneous sample. Guidelines were provided but subjects' activities, diet, and hydration status were not strictly controlled. The incremental load test was rather long for highly trained subjects, up to 62 min, resulting in lower than expected V˙O2peak measures for those individuals. It is not clear how a minimal warm-up affected the EP. Future work could examine the influence of warm-up protocol on the performance of the 3MT, especially if the primary goal is determination of EP with less emphasis on W′.
The 3MT can be performed using equipment readily available to cyclists without the need for indirect calorimetry or blood sampling. The average power output over the last 30 s of the 3-min test is correlated with power outputs at lactate and ventilatory threshold as well as power output at V˙O2peak. The EP from a single session 3MT can be used to estimate the demarcations between both moderate to heavy and heavy to severe exercise intensity domains.
This study was conducted with no external funding.
The authors thank the subjects who participated in this study; Dr. Heather Barber, Kelly Cote, Shauna Dempsey, and Kerry Litka for their assistance and guidance; and Dr. Mark Burnley for correspondence regarding the original work on the 3MT.
The results of this study do not constitute endorsement by the American College of Sports Medicine.
1. Allen H, Coggan A. Training and Racing with a Power Meter
. Boulder (CO): VeloPress; 2006. p. 5-20, 41-65.
2. Amann M, Subudhi AW, Foster C. Predictive validity of ventilatory and lactate thresholds for cycling time trial performance. Scand J Med Sci Sports
3. Beneke R, von Duvillard SP. Determination of maximal lactate steady state response in selected sports events. Med Sci Sports Exerc
4. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis: implications for performance diagnostics in endurance athletes. Sports Med
5. Bentley DJ, McNaughton LR, Thompson D, Vleck VE, Batterham AM. Peak power output, the lactate threshold, and time trial performance in cyclists. Med Sci Sports Exerc
6. Billat V, Faina M, Sardella F, et al. A comparison of time to exhaustion at VO2
max in élite cyclists, kayak paddlers, swimmers and runners. Ergonomics
7. Burnley M, Doust JH, Vanhatalo A. A 3-min all-out test to determine peak oxygen uptake and the maximal steady state. Med Sci Sports Exerc
8. Coyle EF, Feltner ME, Kautz SA, et al. Physiological and biomechanical factors associated with elite endurance cycling performance. Med Sci Sports Exerc
9. Coyle EF. Integration of the physiological factors determining endurance performance ability. Exerc Sport Sci Rev
10. Gaesser GA, Poole DC. The slow component of oxygen uptake kinetics in humans. Exerc Sport Sci Rev
11. Heck H, Mader A, Hess G, Mücke S, Müller R, Hollmann W. Justification of the 4-mmol/L lactate threshold. Int J Sports Med
12. Hill DW. The critical power concept. A review. Sports Med
13. Lucía A, Hoyos J, Chicharro JL. Preferred pedalling cadence in professional cycling. Med Sci Sports Exerc
14. Monod H, Scherrer J. The work capacity of a synergic muscular group. Ergonomics
15. Pfitzinger P, Freedson PS. The reliability of lactate measurements during exercise. Int J Sports Med
16. Poole DC, Ward SA, Gardner GW, Whipp BJ. Metabolic and respiratory profile of the upper limit for prolonged exercise in man. Ergonomics
17. Poole DC, Ward SA, Whipp BJ. The effects of training on the metabolic and respiratory profile of high-intensity cycle ergometer exercise. Eur J Appl Physiol
18. Vanhatalo A, Doust JH, Burnley M. Determination of critical power using a 3-min all-out cycling test. Med Sci Sports Exerc
19. Vanhatalo A, Doust JH, Burnley M. Robustness of a 3 min all-out cycling test to manipulations of power profile and cadence in humans. Exp Physiol
20. Vanhatalo A, Doust JH, Burnley M. A 3-min all-out cycling test is sensitive to a change in critical power. Med Sci Sports Exerc
21. Wasserman K, Hansen J, Sue D, Stringer W, Whipp B. Principles of Exercise Testing and Interpretation
. Philadelphia (PA): Lippincott Williams & Wilkins; 2005, p. 88-9.
Keywords:©2010The American College of Sports Medicine
BICYCLE; FIELD TEST; POWER METER; LACTATE THRESHOLD; TRAINING ZONES