Secondary Logo

Journal Logo

Knee Extensor Fatigue Threshold Is Related to Whole-Body V˙O2max


Medicine & Science in Sports & Exercise: July 2012 - Volume 44 - Issue 7 - p 1366–1374
doi: 10.1249/MSS.0b013e318249d701

Purpose Above a given exercise intensity, rapid muscle fatigue will occur. We explored the possibility of assessing torque threshold for peripheral fatigue during single-legged repetitive isometric knee extensor exercise. We hypothesized this fatigue threshold to be related to the general aerobic fitness level and the so-called “critical torque” (CT) established with a recently validated 5-min all-out test.

Methods Seventeen healthy men (V˙O2max = 44.7–69.6 mL·kg−1·min−1) performed six submaximal (20%–55% maximal voluntary contraction [MVC]) 5-min bouts of 60 repetitive contractions (3-s on, 2-s off). Torque was changed between bouts in steps of 5% MVC to estimate the highest intensity (fatigue threshold) at which average changes in rsEMG, EMG median power frequency, and tissue deoxygenation (near-infrared spectroscopy) of the three superficial knee extensor muscles were still <5%, signifying steady-state exercise with minimal peripheral fatigue. On another occasion, one bout was performed in an all-out manner with end-test torque representing CT.

Results Fatigue threshold (40.0% ± 8.1% MVC) was related (r2 = 0.57, P < 0.05) to CT (53.1% ± 10.0% MVC), but it was consistently lower (P < 0.05) and only fatigue threshold was significantly related to V˙O2max (r2 = 0.68), and the first (r2 = 0.45) and second (r2 = 0.63) ventilatory threshold obtained during cycle ergometry.

Conclusions Performing submaximal bouts of knee extensor contractions, while monitoring EMG and deoxygenation, seems a feasible manner to estimate an aerobic capacity–related exercise intensity of peripheral fatigue onset. This test may be used to evaluate changes in endurance capacity of single muscle groups, without the necessity for all-out testing, which could be problematic with frail subjects.

1Research Institute MOVE, VU University Amsterdam, THE NETHERLANDS; and 2Institute for Biomedical Research into Human Movement and Health, Manchester Metropolitan University, Manchester, UNITED KINGDOM

Address for correspondence: Cornelis J. de Ruiter, Ph.D., Research Institute MOVE, FBW, VU University Amsterdam, Van der Boechorststraat 9; 1081 BT Amsterdam, The Netherlands; E-mail:

Submitted for publication September 2011.

Accepted for publication January 2012.

During exercise involving submaximal repetitive muscle contractions, there is a given exercise intensity above which activated muscle fibers will soon start to fatigue, and torque, velocity, or power can only be maintained by recruiting additional motor units. This recruitment will follow the size principle (17); hence, the newly recruited motor units will, in general, be larger and contain fibers that have less resistance to fatigue than the already-activated fibers. The newly recruited fibers are therefore likely to be subject to fatigue themselves, necessitating recruitment of yet other (larger) motor units in an attempt to keep torque or power output at the same level. Consequently, whenever peripheral fatigue develops, a central neural drive on the motoneuron pool has to increase to maintain output and the exercise becomes increasingly harder for the subjects over time. Finally, subjects are unable to sufficiently activate the contractile machinery and output will decline. Concepts such as maximal lactate steady-state ventilatory threshold (3,25), critical power or velocity (20), or torque for isometric contractions (4) refer to the existence of a critical exercise intensity, above which rapid fatigue will occur and below which the exercise can be maintained for an extended period provided that other factors such as substrate depletion and hyperthermia do not occur (20).

Establishing these critical exercise intensities during whole-body activities such as running, swimming, and cycling can be used to design training programs and to evaluate their effects. Critical power, for example, has been related to V˙O2max and ventilatory threshold (29) and was found to increase after training (28). For research purposes and in rehabilitation, it would be useful to have a simple test that would reflect aerobic capacity of single muscle groups, which, in addition, could be related to (changes in) contractile parameters of that muscle group. It has been well established that short-term peripheral muscle fatigue occurs when the required energy flux surpasses a certain threshold, which, in general, occurs when muscle fibers operate at approximately 70% of their maximal aerobic capacity (7). Exercise above critical power was accompanied by fatigue-related non–steady-state metabolic responses such as increases in inorganic phosphate (Pi) and lactate and decreased phosphocreatine (PCr) and pH (21,27). In addition, during exercise above critical power, oxygen consumption steadily increases (13,27), which is accompanied by a small but consistent rise in muscle deoxygenation (2,15). This slow component of V˙O2 kinetics during “very heavy” exercise has been well documented, and progressive motor unit recruitment has been proposed as a potential underlying cause (e.g., Poole et al. [26]). Recent findings suggest that a reduction of mechanical efficiency of early recruited motor units may also contribute to the steady rise in V˙O2 (6). This is in line with data from isolated frog fibers showing reduced force production relative to oxygen consumption when slowly fatiguing fibers were fatigued (18). Clearly, this could also contribute to a rise in muscle deoxygenation above critical exercise intensity. In the present study, increases in deoxygenation, determined with near-infrared spectroscopy (NIRS), in addition to decreases in electrically induced twitch torques and changes in surface EMG parameters were used to determine the torque threshold for onset of peripheral fatigue.

We assessed this torque threshold during one-legged repetitive submaximal isometric knee extensor contractions, and we expected this threshold to be related to the aerobic capacity of the muscle. Obtaining the aerobic capacity of the knee extensors in isolation is not easy, but because knee extensors produce most of the work during cycling, we hypothesized that the torque threshold for fatigue during isolated knee extensor contractions would be positively related to V˙O2max and ventilatory thresholds assessed during cycle ergometry.

Recently, Burnley (4) established that, during 5 min of all-out repetitive maximal knee extensor contractions, end-test torque leveled off to values similar to the conventionally determined CT, which involves repetitive testing at different exercise intensities on different days to exhaustion. Clearly, the advantage of Burnley’s test is that it is short lasting, but 5 min of all-out maximal contractions is very hard exercise requiring very high motivation, and it may not be feasible with elderly or frail subjects. Because the subjects in the present study were healthy and fit, we could use Burnley’s recently validated all-out test for CT as a reference to which we could relate the torque at the threshold for peripheral fatigue established with our submaximal test. CT of an isolated muscle group such as the knee extensors has not yet been related to whole-body aerobic fitness, and although torque at fatigue threshold may not be exactly similar to CT, we expected both to be related and consequently hypothesized that similar to torque at peripheral fatigue threshold. CT would also have positive relations with parameters of aerobic fitness assessed during cycle ergometry.

Back to Top | Article Outline



Seventeen healthy male subjects (height = 183 ± 8 cm, body mass = 78.4 ± 9.2 kg, age = 27 ± 7 yr) with differing training backgrounds were recruited from the university to maximize the range in oxidative capacity within the group. Subjects participated in various sports (rowing, running, cycling, strength training, football, etc.) for 6.4 ± 4.1 h·wk−1 (range = 1–18 h·wk−1). We expected that a large variation in 2max among our subjects would help us to establish whether a relation between torque at fatigue threshold and V˙O2max would be present. The study was approved by the local ethics committee, and all participants signed informed consent forms. The participants visited our laboratory on three occasions with at least a 48-h break in between.

Back to Top | Article Outline

First Occasion

Familiarization trial on the dynamometer.

During this occasion, the subjects were familiarized with the equipment and the procedure during the actual experiments. First, maximal voluntary contraction (MVC) and voluntary activation of the dominant (kicking) leg were obtained in the dynamometer with knee and hip angle set at 90°. The participants were secured with tight straps around the shoulders and hips. The shin was protected by a shin guard and secured to a force transducer (KAP-E K.L.0.2; A.S.T. GmbH, Dresden, Germany). The measured forces were multiplied with the external moment arm (22–28 cm, adjusted for leg length) to calculate torques. Subsequently, participants were familiarized with (superimposed) electrical doublet stimulation. After shaving, roughening, and cleansing the skin with 70% ethanol, EMG surface electrodes (Blue Sensor, Ambu Ølstykke, Denmark) were placed on the vastus lateralis (VL), rectus femoris (RF), vastus medialis (VM), and biceps femoris (BF) muscle in a bipolar configuration, in line with the muscle fiber direction, with an interelectrode distance of 25 mm. Reference electrodes were placed on bone structures, on each patella, and on the lateral epicondyle of the femur. The location of each electrode was accurately marked with a waterproof felt tip for precise electrode reapplication in subsequent sessions.

For electrical stimulation, the anode (80 × 130 mm; Schwa-medico, Leusden, The Netherlands) was placed on the gluteal fold and the cathode (50 × 50 mm) was placed on the trigonum femorale. Electrical stimulation doublets (100 Hz) were applied using a computer-controlled stimulator (model DS7A; Digitimer Ltd., Welwyn Garden City, UK). The current was increased in steps of 50 mA until the torque leveled off and was then increased by an additional 50 mA.

After several submaximal voluntary contractions of increasing intensity, participants performed two MVCs (2-min rest) with superimposed doublet stimulation on the torque plateau while both were immediately followed by doublet stimulation on the relaxed muscle. During the 2-min rest, a flexion MVC was performed; this was used for normalization of flexion EMG during knee extensions, but the flexor EMG data did not provide any important additional information and will therefore not be presented.

The familiarization trial was also used to practice a bout of 5-min contractions similar to the bouts that would be used during the experimental trials on the second and third occasions. These were repetitive single-legged isometric contractions at 30% MVC, 3 s on 2-s rest in between, hence 60 contractions in total. On the last contraction, a superimposed doublet was applied, followed by a doublet on the relaxed muscle. After 2 min of rest, the subject was asked to perform another MVC. Subjects were not given an opportunity to practice the all-out test because, from pilot measurements, it seemed that subjects were negatively surprised by the exhaustive nature of the all-out test and we were afraid that they would adopt some pacing strategy if they had prior experience with the test. In retrospect, one may question this approach as we did not succeed in preventing pacing (see Results). Moreover, the subjects of Burnley (4) had similar performances in two all-out test occasions.

Back to Top | Article Outline

V˙O2max test.

After these familiarization measurements on the dynamometer, the participants performed a graded exercise test on a cycle ergometer (Lode, Groningen, The Netherlands) to obtain V˙O2max and to determine first and second ventilatory thresholds (VT1, VT2). Oxygen consumption was measured breath by breath, using open circuit spirometry (Oxycon alpha, Mijnhardt, The Netherlands) with a facemask covering the nose and mouth. Subjects were instructed to keep their pedaling rate at about 90 rpm. The subjects started with a warming up of 2-min unloaded cycling followed by 1 min at 25 W. After the warming up, the load increased every minute with steps of 20–35 W (depending on the weight, length, age, and fitness level) until exhaustion. The total duration of the test was 10–20 min. During the final stages of the test, the subjects were strongly encouraged to continue cycling for as long as possible.

Back to Top | Article Outline

Second and Third Occasions

On the second or third visits, the participants performed, in random order, either an all-out test or a series of six bouts of submaximal exercise (5 min each) consisting of repetitive isometric contractions in the dynamometer. On the second and third visits, in addition to EMG and electrical stimulation (see first occasion), NIRS was used.

Back to Top | Article Outline


The changes in muscle oxygenation of the VM, VL, and RF muscles during the exercise tests were determined using a continuous wave near-infrared spectrophotometer (Oxymon; Artinis Medical Systems, Zetten, The Netherlands), which generated light at 780 and 850 nm (31). Data were collected with a sample frequency of 10 Hz. The three optode sets were each fixed in a mold with an interoptode distance of 45 mm. The molds were secured to the thigh using elastic Velcro straps such that the optodes did not move during contraction. The optodes were positioned on the center of the muscle bellies in such a manner that the EMG electrodes on the knee extensor muscles were exactly in between both optodes of all three sets. Before the optodes were placed, the skin was shaved and cleaned with ethanol. The positions of the optodes were marked to guarantee that optode positions were similar during the second and third visits. NIRS signals were visible on line on a separate monitor.

With NIRS, the tissue oxygenation level can be measured noninvasively. The optical change in density of the tissue is measured and transformed into the change in concentrations of oxyhemoglobin ([O2Hb]) and oxymyoglobin ([O2Mb]) and deoxyhemoglobin and deoxymyoglobin ([HHb] and [HMb]) by modification of the Lambert–Beer law (24). Because of the overlap in the spectrum, hemoglobin and myoglobin cannot be measured separately; O2Hb and HHb will denote the oxygenated and deoxygenated forms, respectively, of both proteins in the present study and were expressed relative to maximal deoxygenation (9). Maximal deoxygenation is defined as the absolute difference in [O2Hb] (and [HHb]) when virtually all O2Hb is converted into HHb. In the present study, maximal deoxygenation was determined using 20-s knee extensor MVCs at the end of the second and third occasions. During long-lasting MVCs, internal muscles forces will totally occlude blood flow and oxygen will be consumed at a sufficiently high rate to reach maximal deoxygenation within 20 s (11).

Back to Top | Article Outline

All-out test.

Participants sat in the dynamometer, electrodes (EMG and stimulation) and NIRS optodes were placed, and supramaximal stimulation current was determined.

Although Burnley’s (4) original protocol did not include a warm-up, a warm-up trial of 2-min contractions (3-s on, 2-s off) at 30% of MVC was included, which was considered an appropriate intensity because, in pilot experiments, all subjects were able to complete 5 min at 30% MVC. After this, participants performed three maximal extension MVCs (3–5 s) with superimposed doublet stimulation at the plateau and doublet stimulation immediately after relaxation. There were 2-min rest in between these contractions, during which a flexion MVC was generated. Thereafter, after 5-min rests, the participants performed an all-out fatigue test (4): 60 maximal contractions (3-s duration) with 2-s rest in between, thus with a total duration of 5 min. Audio signals were presented throughout the test, and participants were instructed to start pushing when a 3-s beep was heard and to stop when the beep ended (this was practiced on the first occasion). Online visual feedback of the torque produced during each contraction was provided, whereas the torque produced during the previous contraction was also visible on the screen. The participants were verbally encouraged to reach (at least) the same torque as during the preceding contraction. During the test, electrical stimulation was applied on the second, sixth, and every after sixth contraction (every 30 s), both on the contracting muscle and as well as 1 s after the contraction ended on the relaxed muscle (4), to monitor voluntary activation and central and peripheral (muscle) fatigue. To discourage the use of any pacing strategies, participants were not informed about the elapsed time. After the test, the participants rested for 5 min, after which they produced one more MVC lasting 20 s to establish the maximal deoxygenation of the knee extensor muscles (see NIRS). During all measurements, the subjects were strongly encouraged to produce the highest torques possible.

Back to Top | Article Outline

Submaximal test.

Procedures were the same as described for the all-out test, but instead of the 5-min maximal test, participants now produced a series of six bouts of sixty repetitive submaximal contractions (3-s on, 2-s off, audio paced). There was a 1-min rest in between these 5-min bouts; again, subjects were not informed about the elapsed time during the bouts. On a monitor in front of the participants, the torque signal was shown together with a line denoting the target torque. The highest MVC produced on this experimental day was set as 100%. Doublet stimulation was applied on every last contraction of each 5-min bout followed by doublet stimulation immediately after muscle relaxation. If participants were not able to deliver the required torque for 5 min, they were instructed to produce maximal torque for three more contractions and doublet stimulation was applied on the final contraction and on the relaxed muscle immediately afterward. Subsequently, after the 1-min rest, the participant continued with the next bout. The first bout was always at 30% MVC, because based on pilot experiments, we anticipated our subjects to be able to complete a 30% bout with little or no peripheral fatigue (see Data Analysis). This 30% bout was followed by a second bout at 50% MVC. The third and fifth bouts were “recovery” bouts at 20% MVC, whereas the fourth and sixth bouts were at intensities that were chosen based on the signs of peripheral fatigue (see next paragraph) observed during the preceding bouts. Torque level between bouts was varied in steps of at least 5% MVC.

During post hoc data analysis, several objective criteria were used to determine whether muscle fatigue developed during each bout, but during the test, we could only use those indications of fatigue that were immediately available. We used a decrease in doublet torque by more than 10%; nonstabilizing (increasing) deoxygenation of any of the knee extensor muscles during the last 2 min of a bout; subjective indications provided by the participant who rated the intensities as easy, hard, or very hard exercise; and of course, failure to complete a bout at a certain intensity. After completion of the six 5-min bouts at various submaximal torque levels with 1-min rest in between, there was again a 5-min rest. Finally, and similar to the procedure after the all-out test, the participants produced one more MVC lasting 20 s to establish the maximal deoxygenation of the knee extensor muscles for normalization of deoxygenation data.

Back to Top | Article Outline

Data Analysis

Cycle ergometry.

During the V˙O2max test, data were sampled breath by breath. Afterward, they were averaged over 20-s intervals. V˙O2max (mL·min−1) was then calculated as the highest V˙O2 value over such an interval and corrected for body mass (mL·kg−1·min−1). Because V˙O2max may not be the best parameter to quantify aerobic fitness level (8), V˙O2 at VT1 and VT2 were also assessed (25). VT1 and VT2 have been documented as important determinants of endurance training status (25). At VT1, the ventilatory equivalent for oxygen (E/V˙O2) starts to increase while there is no change in the ventilatory equivalent for CO2 (E/V˙CO2). The end-tidal partial pressure of oxygen (PETO2) also starts to increase, whereas PETCO2 remains steady (25). VT1 was obtained by visual inspection of the V˙O2 versus V˙CO2 curve (V-slope method) (3). Two researchers individually examined the breakpoints in this curve. If there was no clear breakpoint, or if the researchers disagreed on this point by more than 100 mL·min−1, the E/V˙O2 and the E/V˙CO2 curves were inspected (ventilatory equivalent method), followed by the PETO2 and the PETCO2 curves.

VT2 was found at a higher exercise intensity and was determined by two researchers using the criteria of an increase in both E/V˙O2 and E/V˙CO2 and a decrease in PETCO2 (25).

Back to Top | Article Outline


All data were analyzed using custom-written MATLAB software (version R2008b). The EMG data (sampled at 1000 Hz) were rectified and filtered using a fourth-order band-pass filter of 10–400 Hz. Torque data (sampled at 1000 Hz) were low-pass-filtered at 50 Hz using a fourth-order Butterworth filter. All doublet torques were normalized to the mean torque of the three doublets applied after each MVC on each occasion. Rectified surface EMG amplitude (rsEMG) and median frequency as well as muscle deoxygenation (changes in [HHb] signal) during torque production were calculated for each of the muscles separately. For each muscle separately, deoxygenation was subsequently expressed as a percentage of maximal deoxygenation (maximal change in [HHb]) measured in each knee extensor muscle head during the 20-s MVC at the end of each experimental day. All torques and EMG signals were normalized to the values obtained during the highest MVC (1 s around peak torque) on the same experimental day (peak MVC). For every 3-s contraction, the average value over the middle 2 s was calculated for rsEMG, median frequency, deoxygenation, and torque. Torque time integrals of all contractions were also calculated, but because the outcome for torque time integral and mean torque were very comparable, the former are not presented. Finally, for rsEMG, EMG median frequency, and deoxygenation, the normalized values of VL, RF, and VM were averaged to obtain an indication of the respective average whole-knee extensor values, which are presented.

Voluntary activation (VA) was calculated as follows (4,12):

When doublet stimulation on a voluntary contraction does not cause any increase in torque output of the muscle, voluntary activation would be 100%. Note, however, that there are strong indications (see Results) that this method, in general, overestimates VA (23,30).

Back to Top | Article Outline

Determination of CT and fatigue threshold.

For the all-out test, CT was defined as the average torque produced over the contractions no. 54 to 59 (4). The last (60th) contraction was discarded for this calculation because of the superimposed stimulation that was delivered during this contraction.

The bouts of submaximal contractions were used to determine the torque at the threshold for peripheral fatigue onset (fatigue threshold). Fatigue threshold was expected to represent the highest torque level, which could be produced during the 5-min exercise without progressive fatigue. As explained in the introduction, fatigue was expected to be accompanied with non–steady-state metabolism and increased neural drive by the participant over time. Consequently, the fatigue threshold was established as the highest exercise intensity that could be maintained meeting all of the following three criteria. First, any increase in averaged (across three knee extensor muscle heads) rsEMG over the final 2 min of a 5-min exercise bout was <5%. Second, any increase in averaged deoxygenation over the final 2 min of a bout was <5%. Third, any decrease in averaged median frequency of the EMG over the final 2 min of a bout was <5%. This may seem to be a strict set of criteria, but in practice, because torque was varied with minimum steps of 5% MVC, when one of the parameters changed at a certain torque, this was usually by >5%. In addition, usually the change in at least one of the other two parameters was also >5%. The relative change in the above parameters during the last 2 min of a bout of 60 contractions was calculated as follows: [(average value of contractions 54 to 59) − (average value of contractions 30 to 35)] / (average value of contractions 30 to 35) × 100%.

Back to Top | Article Outline


Results are presented as mean ± SD. Paired t-tests or ANOVA was used (version 17.0; SPSS, Inc., Chicago, IL) when applicable. Pearson correlation coefficient was calculated to establish the significance of linear relations. Statistical significance was accepted at P < 0.05.

Back to Top | Article Outline


Maximal voluntary torque on the dynamometer (peak MVC) before the first test was 267 ± 86 N·m, which was similar (P = 0.27) to peak MVC before the second test (277 ± 78 N·m). V˙O2max established on the cycle ergometer was 54.4 ± 7.3 mL·kg−1·min−1 (range = 44.7–69.6 mL·kg−1·min−1), whereas VT1 and VT2 were 31.6 ± 6.5 and 44.0 ± 7.1 mL·kg−1·min−1, respectively. Together these numbers illustrate that the present subjects were fit individuals.

Back to Top | Article Outline

Submaximal test.

Figure 1 shows a typical example of a sequence of six 5-min bouts of submaximal contractions at different percentages of MVC. For this subject, signs of peripheral fatigue (increases in rsEMG and deoxygenation and decrease in median EMG frequency during a bout) were present during contractions at 50%, 40%, and 35% MVC. At 50% MVC, this subject even failed to complete the 5-min exercise bout. In this case, fatigue threshold was calculated to be 32.5% (average torque during the first bout). On average (n = 17), fatigue threshold was reached at 40.0% ± 8.1% of MVC.



In line with our hypothesis, torque at fatigue threshold was significantly (P < 0.05) related (n = 17) to V˙O2max (r2 = 0.46), VT1 (r2 = 0.36), and VT2 (r2 = 0.37). Moreover, the r2 values of these relations improved considerably, to 0.68, 0.45, and 0.63, respectively, when one outlier was not included (Fig. 2).



Back to Top | Article Outline

Five-minute all-out test.

The linear regression showed that the slope of torque during the last 60 s of the 5-min all-out test was 0.02% ± 0.09% MVC·s−1, which was not significantly different from zero (P = 0.40). During the final 30 s, torque had stabilized at 53.1% ± 10.0% MVC. This CT was significantly higher (P < 0.05) than the torque at fatigue threshold (40.0% ± 8.1% MVC) established with the submaximal test. Nevertheless, and as hypothesized, there was a significant relation (r2 = 0.57, P < 0.05) between both parameters (Fig. 3). However, and in contrast to torque at fatigue threshold, CT was not related (r2 < 0.14, P > 0.05) to V˙O2max, VT1, or VT2.



Torque in the first contraction of the all-out test was, on average, <100% peak MVC (82.3% ± 14.1%). Superimposed stimulation confirmed that VA was significantly <100% in the first contraction (95.2% ± 7.5%) of the test. VA decreased even further (to 87.7% ± 12.5%; paired t-test, P = 0.02) in the last contraction. Similarly, rsEMG, normalized to values obtained during peak MVC, was <100% during the first contraction (81.8% ± 14.3%), and it further decreased (to 72.2% ± 21.9%; paired t-test, P = 0.14) in the last contraction (Fig. 4). Neural activation during the voluntary all-out test could potentially affect CT. However, the less-than-maximal VA during the test did not seem to affect the value of CT in a systematic manner: there was no relation (r2 = 0.09, P > 0.05) between torque in the first contraction (%peak MVC) and CT. Moreover, there were relations between VA neither at the start nor at the end of the test (mean of last three measurements) and CT (r2 < 0.16, P > 0.05). Furthermore, rsEMG (mean of the last six contractions) was also not related to CT (r2 = 0.05, P > 0.05). In addition, and although percentage doublet torque at the end of the test (62.2% ± 23.3%) was not significantly different from CT, both parameters were not related (r2 = 0.20, P > 0.05). Finally, percentage deoxygenation reached during the all-out test was also not related to CT (r2 = 0.005, P > 0.05). Muscle deoxygenation increased over the first six contractions and subsequently leveled off at values near 70% of maximal deoxygenation (Fig. 4).



Back to Top | Article Outline


The main finding of the present study was that torque threshold for peripheral fatigue established with consecutive bouts of submaximal one-legged isometric knee extensor contractions had good relations with parameters of aerobic training status (V˙O2max, VT1, VT2) determined during a graded cycle ergometer test. The use of submaximal bouts of muscle contractions and signs of peripheral fatigue seems a promising method for assessing a fatigue threshold that reflects aerobic capacity of an isolated muscle group during standardized exercise. This may be useful for evaluating the effects of training and rehabilitation programs on aerobic capacity of separate muscle groups.

Back to Top | Article Outline

All-out test (CT).

In the current study, average torque at the start of the all-out test was 82.3% of MVC and only six subjects were able to start with contractions greater than 90% MVC. Clearly, this was related to submaximal muscle activation because rsEMG and VA at the start of the test were lower than during isolated MVCs before the test. Apparently, many of our subjects (subconsciously) adopted some form of pacing at the beginning of the all-out test. This is a well-known phenomenon related to the anticipation of fatigue and the knowledge that total exercise duration would be several minutes (in this, case 5 min) (22,32). The eight subjects in Burnleys’ study (4) did a familiarization all-out test, which may have contributed to their high torques (89%–104% MVC) at the start of the test. In the present study, pacing could have mitigated muscle fatigue and may have contributed to the relatively high values for CT (∼53%) compared with the Burnley study (∼30%) (4). However, the findings that torque (%MVC) at the start of the test was unrelated to CT and CT was related (r2 < 0.2) to VA neither at the start nor at the end of the test suggest that VA (pacing) did not affect CT in a systematic way. In addition, in the study that validated this all-out test for CT, VA and rsEMG also declined by approximately 25% during the test (4). The relatively high CT in the present compared to this recent study is in line with the relatively high values for electrically stimulated doublet torque at the end (% pretest) of the test in the present (∼63%) compared with the earlier study (∼46%) (4) and could be indicative for a better aerobic fitness of the present subjects. However, within the present study, CT was not related to the end-test doublet torque or the parameters of aerobic training status (V˙O2max, VT1, and VT2). Unfortunately, these relationships were not reported or investigated in the previous report (4).

We expected CT as well as torque at fatigue threshold to be related to aerobic fitness because aerobic capacity would determine the potential for metabolic recovery in between maximal muscle contractions of the all-out test. PCr recovery, for example, is dependent on oxygen availability (16). During maximal contractions, the high intramuscular pressure will completely occlude blood flow (11), but during the 2-s intervals between contractions, some reoxygenation and partial recovery of PCr can occur. Recently, during a similar all-out test, phosphocreatine (PCr) declined to about 20% and pH to about 6.64, whereas Pi and diprotonated phosphate concentrations increased substantially (5). Such metabolic changes may underlay peripheral fatigue (1). Therefore, the aerobic capacity–related potential for metabolic recovery in between contractions was expected to determine the extent of peripheral fatigue and CT. It cannot be excluded that the between-subject variation in contribution of central fatigue (VA) to end-test torque (CT) may have clouded any relationships between CT and the parameters of aerobic fitness.

Back to Top | Article Outline

Submaximal test (fatigue threshold).

Because of the nature of the test, subjects exercised slightly above fatigue threshold in some of the submaximal exercise bouts, but even then, most subjects were able to complete all 5-min bouts. This suggests that, in contrast to the all-out test, central fatigue did not affect the estimation of torque at fatigue threshold. However, to limit total test duration to 30 min, the first two bouts at 30% and 50% MVC, respectively, were used to obtain a first indication of the fatigue threshold. In retrospect, 50% MVC was considerably above fatigue threshold for several subjects, and it even led to bout termination sooner than 5 min in five subjects (e.g., Fig. 1). For those participants in particular, it may be worth considering whether 1 min of rest followed by a recovery bout at 20% MVC was enough to completely recover from the bout at 50% MVC. Indeed, peripheral fatigue seemed to have developed during the test because average doublet torque was about 20% lower after the last three compared with the first three bouts (data not presented). However, fibers that fatigued at bout intensities 10%–15% above fatigue threshold probably belonged to larger units that were less intensively recruited during subsequent bouts at intensities closer to fatigue threshold. Therefore, and although this is speculative, the effect of peripheral fatigue on determination of fatigue threshold may have been limited. Moreover, the good relations between torque at fatigue threshold and the parameters of aerobic fitness (Fig. 2) seem to indicate that torque at fatigue threshold is a meaningful parameter. Nevertheless, although the submaximal test seems promising, it remains to be seen if bout intensity order does affect torque at fatigue threshold. It may be feasible to start the first bout at 20% MVC and just increase intensity in steps of 5% MVC until fatigue threshold is reached. Clearly, further study is needed to determine whether the fatigue threshold can also be observed with such a staged incremental test.

By definition (see Methods), above fatigue threshold, rsEMG and muscle deoxygenation did not stabilize. Jones et al. (21) showed that exercising 10% above a critical exercise intensity resulted in non–steady-state metabolism (decline of pH and fall in [PCr] and increase in Pi), whereas a metabolic steady state was reached within 3 min during exercise 10% below critical exercise intensity. Poole et al. (27) showed that at an exercise intensity 5% below sustainable power output, blood lactate and V˙O2 reached constant values, whereas at a 5% higher exercise intensity, blood lactate and V˙O2 steadily increased. During exercise above sustainable power output, muscle deoxygenation has also been reported to increase (2,15). The latter is in accordance with the present finding that, during bouts above fatigue threshold, deoxygenation during the contractions gradually increased (Fig. 1). Although increases in deoxygenation may have several causes such as decreased muscle blood flow, the changes found in [HbO2] and [HHb] during steady force production suggest that oxygen consumption gradually increased during bouts above fatigue threshold. In general, changes in [HbO2] and [HHb] were mirror images signifying that total Hb concentration ([total Hb]) remained constant (e.g., bottom of Fig. 1, the last bout). Occasionally, [total Hb] concentration did change, for example, during the first part of the second bout shown in Fig. 1. However, note that, during the last contractions of the second bout, [total Hb] leveled off at the same time deoxygenation steeply increased. A gradual rise in oxygen consumption could be caused by increased rate coding and/or recruitment of additional motor units to compensate for the loss of force-producing capacity of fatiguing muscle fibers. Indeed, clear increases in deoxygenation were always accompanied by increases in rsEMG. Although great care should be taken to interpret changes in rsEMG in terms of motor unit recruitment and rate coding (14) when rsEMG does increase, it is usually paralleled by additional motor unit recruitment and/or increased rate coding (10,19). It is important to note that recent findings support the idea that reduced efficiency of early recruited motor units may also contribute to the rise in oxygen consumption during very heavy exercise (6,18). The concurrent decrease in mean power frequency during contractions above fatigue threshold may relate to decreases in muscle fiber action potential conduction velocity, which, in turn, have been linked to fatigue-related metabolic changes such as accumulation of potassium and decreased pH (33). It is noteworthy that the aforementioned signs of peripheral fatigue found above fatigue threshold in all subjects occurred slightly below the exercise intensity of CT (Fig. 3). This suggests that, despite the significant relation (r2 = 0.57, P < 0.05) between torque threshold for peripheral fatigue and CT, these parameters are partly affected by different physiological and/or neural (possibly voluntary activation) factors.

In conclusion, performing submaximal bouts of knee extensor contractions while monitoring EMG and deoxygenation seems a feasible manner to estimate an aerobic capacity–related exercise intensity of peripheral fatigue onset. This test may be used to evaluate changes in endurance capacity of single muscle groups. In frail subjects, in particular, this test has clear advantages because it does not involve all-out testing or repetitive testing to torque failure, which may be too strenuous.

The authors did not receive external funding for this study and have no professional relationships with companies or manufacturers.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Allen DG, Lamb GD, Westerblad H. Skeletal muscle fatigue: cellular mechanisms. Physiol Rev. 2008; 88 (1): 287–332.
2. Bailey SJ, Wilkerson DP, Dimenna FJ, Jones AM. Influence of repeated sprint training on pulmonary O2 uptake and muscle deoxygenation kinetics in humans. J Appl Physiol. 2009; 106 (6): 1875–87.
3. Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986; 60 (6): 2020–7.
4. Burnley M. Estimation of critical torque using intermittent isometric maximal voluntary contractions of the quadriceps in humans. J Appl Physiol. 2009; 106 (3): 975–83.
5. Burnley M, Vanhatalo A, Fulford J, Jones AM. Similar metabolic perturbations during all-out and constant force exhaustive exercise in humans: a 31P magnetic resonance spectroscopy study. Exp Physiol. 2010; 95 (7): 798–807.
6. Cannon DT, White AC, Andriano MF, Kolkhorst FW, Rossiter HB. Skeletal muscle fatigue precedes the slow component of oxygen uptake kinetics during exercise in humans. J Physiol. 2011; 589 (Pt 3): 727–39.
7. Conley KE, Kemper WF, Crowther GJ. Limits to sustainable muscle performance: interaction between glycolysis and oxidative phosphorylation. J Exp Biol. 2001; 204 (Pt 18): 3189–94.
8. Coyle EF. Integration of the physiological factors determining endurance performance ability. Exerc Sport Sci Rev. 1995; 23: 25–63.
9. de Ruiter CJ, de Boer MD, Spanjaard M, de Haan A. Knee angle–dependent oxygen consumption during isometric contractions of the knee extensors determined with near-infrared spectroscopy. J Appl Physiol. 2005; 99 (2): 579–86.
10. de Ruiter CJ, Elzinga MJH, Verdijk PWL, van Mechelen W, de Haan A. Voluntary drive-dependent changes in vastus lateralis motor unit firing rates during a sustained isometric contraction at 50% of maximum knee extension force. Pflügers Arch - Eur J Physiol. 2004; 447: 436–44.
11. de Ruiter CJ, Goudsmit JF, Van Tricht JA, de Haan A. The isometric torque at which knee-extensor muscle reoxygenation stops. Med Sci Sports Exerc. 2007; 39 (3): 443–53.
12. de Ruiter CJ, Vermeulen G, Toussaint HM, de Haan A. Isometric knee-extensor torque development and jump height in volleyball players. Med Sci Sports Exerc. 2007; 39 (8): 1336–46.
13. Endo MY, Kobayakawa M, Kinugasa R, et al.. Thigh muscle activation distribution and pulmonary V˙O2 kinetics during moderate, heavy, and very heavy intensity cycling exercise in humans. Am J Physiol Regul Integr Comp Physiol. 2007; 293 (2): R812–20.
14. Farina D, Merletti R, Enoka RM. The extraction of neural strategies from the surface EMG. J Appl Physiol. 2004; 96 (4): 1486–95.
15. Ferreira LF, Townsend DK, Lutjemeier BJ, Barstow TJ. Muscle capillary blood flow kinetics estimated from pulmonary O2 uptake and near-infrared spectroscopy. J Appl Physiol. 2005; 98 (5): 1820–8.
16. Haseler LJ, Hogan MC, Richardson RS. Skeletal muscle phosphocreatine recovery in exercise-trained humans is dependent on O2 availability. J Appl Physiol. 1999; 86 (6): 2013–8.
17. Henneman E, Somjen G, Carpenter DO. Functional significance of cell size in spinal motorneurons. J Neurophysiol. 1965; 28: 560–80.
18. Hepple RT, Howlett RA, Kindig CA, Stary CM, Hogan MC. The O2 cost of the tension–time integral in isolated single myocytes during fatigue. Am J Physiol Regul Integr Comp Physiol. 2010; 298 (4): R983–8.
19. Hunter SK, Duchateau J, Enoka RM. Muscle fatigue and the mechanisms of task failure. Exerc Sport Sci Rev. 2004; 32 (2): 44–9.
20. Jones AM, Vanhatalo A, Burnley M, Morton RH, Poole DC. Critical power: implications for determination of V˙O2max and exercise tolerance. Med Sci Sports Exerc. 2010; 42 (10): 1876–90.
21. Jones AM, Wilkerson DP, DiMenna F, Fulford J, Poole DC. Muscle metabolic responses to exercise above and below the “critical power” assessed using 31P-MRS. Am J Physiol Regul Integr Comp Physiol. 2008; 294 (2): R585–93.
22. Kooistra RD, Blaauboer ME, Born JR, de Ruiter CJ, de Haan A. Knee extensor muscle oxygen consumption in relation to muscle activation. Eur J Appl Physiol. 2006; 98 (6): 535–45.
23. Kooistra RD, de Ruiter CJ, de Haan A. Conventionally assessed voluntary activation does not represent relative voluntary torque production. Eur J Appl Physiol. 2007; 100 (3): 309–20.
24. Livera LN, Spencer SA, Thorniley MS, Wickramasinghe YA, Rolfe P. Effects of hypoxaemia and bradycardia on neonatal cerebral haemodynamics. Arch Dis Child. 1991; 66 (4 Spec No): 376–80.
25. Lucia A, Hoyos J, Perez M, Chicharro JL. Heart rate and performance parameters in elite cyclists: a longitudinal study. Med Sci Sports Exerc. 2000; 32 (10): 1777–82.
26. Poole DC, Barstow TJ, Gaesser GA, Willis WT, Whipp BJ. V˙O2 slow component: physiological and functional significance. Med Sci Sports Exerc. 1994; 26 (11): 1354–8.
27. Poole DC, Ward SA, Gardner GW, Whipp BJ. Metabolic and respiratory profile of the upper limit for prolonged exercise in man. Ergonomics. 1988; 31 (9): 1265–79.
28. 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 Occup Physiol. 1990; 59 (6): 421–9.
29. Smith JC, Dangelmaier BS, Hill DW. Critical power is related to cycling time trial performance. Int J Sports Med. 1999; 20 (6): 374–8.
30. Tillin NA, Pain MT, Folland JP. Short-term unilateral resistance training affects the agonist–antagonist but not the force–agonist activation relationship. Muscle Nerve. 2011; 43 (3): 375–84.
31. van der Sluijs MC, Colier WNJM, Houston RJF, Oeseburg B. A new and highly sensitive continuous wave near infrared spectrophotometer with multiple detectors. Proc SPIE. 1998; 3194: 63–72.
32. Wittekind AL, Micklewright D, Beneke R. Teleoanticipation in all-out short-duration cycling. Br J Sports Med. 2011; 45 (2): 114–9.
33. Zwarts MJ, Stegeman DF. Multichannel surface EMG: basic aspects and clinical utility. Muscle Nerve. 2003; 28 (1): 1–17.


©2012The American College of Sports Medicine