Journal Logo


Prior Upper Body Exercise Reduces Cycling Work Capacity but Not Critical Power

Johnson, Michael A.1; Mills, Dean E.1; Brown, Peter I.2; Sharpe, Graham R.1

Author Information
Medicine & Science in Sports & Exercise: April 2014 - Volume 46 - Issue 4 - p 802-808
doi: 10.1249/MSS.0000000000000159
  • Free


The duration for which severe-intensity constant power exercise can be tolerated increases as a hyperbolic function of decreasing power (19,22). This power–duration relationship is characterized by two parameters: a power asymptote termed critical power (CP) and a curvature constant termed W′. The CP represents the lower boundary of the severe-intensity exercise domain (23,28) and, thus, the power that evokes the highest sustainable rate of oxidative metabolism. Exercise above CP is thus characterized by an inexorable accumulation of fatigue-related metabolites (e.g., La, H+, and inorganic phosphate), a continual decline in intramuscular phosphocreatine concentration, and an increasing pulmonary oxygen uptake (V˙O2) toward V˙O2max (22,23,28,33).

The W′ reflects the maximum amount of work that can be performed above CP irrespective of the rate of W′ utilization (17). Once W′ is expended, exhaustion will ensue unless exercise intensity is reduced below CP to allow restoration of W′ (10,11). However, compared with CP, the mechanistic bases of W′ are less well defined. The W′ is commonly described as a finite energy store determined by oxygen bound to myoglobin, intramuscular phosphocreatine, and glycogen (22,25,31). In support, and in the absence of any change in CP, oral creatine supplementation increases W′ (31), whereas glycogen depletion decreases W′ (25). Partial depletion of intramuscular phosphocreatine may also explain, in part, why prior exercise at powers above CP reduces W′ (13,14,27,34). However, additional mechanisms are likely to exist because the recovery kinetics of V˙O2 (a proxy for intramuscular phosphocreatine recovery) are faster than the recovery kinetics of W′ (13). There is growing support for the notion that W′ may thus also depend on the accumulation of fatigue-related metabolites to a critical tolerable limit, which occurs in proportion to the rate of W′ utilization (13,14,21,22,33). The reduction in W′ due to prior exercise is therefore difficult to interpret because all exercise was performed using the same muscle groups and thus energy store depletion presumably coincided with metabolite accumulation.

The influence of metabolite accumulation on exercise tolerance and W′ may be examined more discretely by performing upper body exercise before the criterion bout of leg cycle exercise (3,24,26). Severe-intensity upper body exercise elevates blood and muscle [La] and [H+] without affecting leg muscle concentrations of ATP, phosphocreatine, and glycogen (2,3). Furthermore, during subsequent leg exercise, K+ efflux from the active leg muscle, and increases in interstitial [K+] are accelerated and exercise tolerance is reduced (3,26). Prior upper body exercise, thus, allows the effects of metabolite accumulation on W′ to be examined without the confounding, concomitant influence of intramuscular energy store depletion.

Therefore, the aim of this study was to investigate the effects of metabolite accumulation, induced by prior severe-intensity upper-body exercise on parameters of the power–duration relationship for leg cycle ergometry. We hypothesized that prior upper body exercise would reduce W′ without affecting CP.



Seven healthy, nonsmoking, moderately trained males (age = 26 ± 4 yr, height = 182 ± 4 cm, body mass = 83 ± 4 kg) provided written informed consent to participate in the study. Participants refrained from caffeine on test days and alcohol and strenuous exercise the day preceding and day of a test. Participants reported to the laboratory at least 2 h postprandial. The study was approved by the Nottingham Trent University Human Ethics Committee, and all procedures were conducted in accordance with the Declaration of Helsinki.

Experimental design

Participants attended the laboratory on ten separate occasions, at a similar time of day, separated by at least 48 h. The initial five visits composed a maximal incremental cycling test and four constant power cycling tests for determination of the power–duration relationship. All cycling tests were performed to the limit of tolerance. The cycling tests were then repeated, in randomized order, during the subsequent five laboratory visits, with each test preceded by severe-intensity intermittent arm-cranking exercise. Hereafter, incremental and constant power cycling tests performed without and with prior arm-cranking exercise are called LINC and ALINC, and LCONST and ALCONST, respectively. Cycling tests during LINC and LCONST trials were preceded by a 20.5-min rest period, which matched the experimental protocol duration preceding the onset of cycling exercise in ALINC and ALCONST.

Equipment and measurements

Measurements were taken using equipment and techniques described previously (7,20). Exercise was performed using electromagnetically braked cycle (Excalibur Sport; Lode, Groningen, the Netherlands) and arm-cranking (Angio; Lode) ergometers. During all tests, participants wore a facemask (model 7940; Hans Rudolph, Kansas City, MO) connected to a flow sensor (ZAN variable orifice pneumotach; Nspire Health, Oberthulba, Germany) that was calibrated using a 3-L syringe. Gas concentrations were measured using fast responding laser diode absorption spectroscopy sensors, which were calibrated using gases of known (5% CO2, 15% O2, balance N2) concentration (BOC, Guilford, UK), and ventilatory and pulmonary gas exchange variables were determined breath by breath (ZAN 600USB; Nspire Health). During all tests, V˙O2peak was defined as the highest recorded value for any 30-s period. Heart rate was measured using short-range telemetry (Polar S610; Polar, Kempele, Finland) and arterial oxygen saturation was estimated (SpO2) using a finger pulse oximeter (Model 8500; Nonin Medical, Plymouth, MN). Arterialized venous blood (6 mL) was drawn from a heated dorsal hand vein via an indwelling 21-G cannula. Blood was analyzed immediately for PCO2 and pH (ABL520; Radiometer, Copenhagen, Denmark), and values were corrected for changes in rectal temperature (1000 Series Squirrel; Grant Instruments, Cambridge, UK). PCO2 and pH were used to calculate plasma bicarbonate concentration (

) using the Henderson–Hasselbalch equation:

Plasma acid–base balance was examined using the physicochemical approach (20,32), which describes the dependency of [H+] and

on the three independent physicochemical variables: strong ion difference ([SID]), PCO2, and the total concentration of weak acids ([Atot]). Thus, a portion (5 mL) of each blood sample was immediately centrifuged for 10 min at 3000g, and the plasma supernatant was removed. Plasma [La] was subsequently determined using an automated analyzer (Biosen C_line Sport; EKF Diagnostics, Barleben, Germany). Plasma [Na+], [K+], and [Cl] were determined using ion selective electrodes, and total protein concentration ([PPr]) was assayed by immunoturbidimetry (ABX Pentra 400; Horiba, Northampton, UK). [Atot] was calculated as 2.45[PPr] (30). Plasma strong ion difference ([SID]) was calculated as the sum of the strong cations minus the sum of the strong anions (32):

During all trials, blood samples were taken, and heart rate and SpO2 were recorded, at rest, immediately before the prescribed cycling test (CYCONSET), and at the limit of cycling exercise tolerance (CYCEND).

Maximal incremental cycling test

Participants performed an incremental cycling test to the limit of tolerance, which was defined as the point at which cycling cadence fell lower than 60 rpm. Tests began at 0 W, and power was increased by discrete 20 W increments every 60 s. Cycling cadence was self-selected and matched during LINC and ALINC. Ventilatory and pulmonary gas exchange variables were averaged for 10-s periods, and the functional gain (i.e., slope of ΔV˙O2W) was determined, using linear regression, from 1 min into the incremental test up to either V˙O2 peak or where V˙O2 began to plateau (4). Maximum power output (W˙max) was calculated as the sum of the power output in the last completed stage plus the product of ramp increment (20 W) and the fraction of the final stage actually completed.

Power–duration relationship

The power–duration relationship was determined from four constant-power cycling tests performed to the limit of tolerance. Each participant adopted the same self-selected cycling cadence for all tests, which were terminated when cadence fell lower than 60 rpm. The initial LCONST test was performed at 85% of the W˙max achieved during the preliminary LINC test, and subsequent tests were performed at powers prescribed to elicit exercise intolerance for a range of times between approximately 3 and 15 min (19). Identical cycling powers were used during LCONST and ALCONST trials. CP and W′ were estimated using the nonlinear power–time model, and the linear work–time and power–(1/time) models. The power–(1/time) model was associated with the lowest SEE for the parameter estimates and was therefore chosen for further analysis (18).

Arm-cranking protocol

The arm-cranking protocol was adapted from that described previously (3,26). After a 5-min rest period, participants performed eight 1-min arm-cranking exercise bouts, interspersed with 30-s rest, at a work rate of 1.5–2.0 W·kg−1 body mass. The center of the arm-crank shaft was aligned to shoulder level, and subjects were seated in an upright position so that the elbow was slightly flexed when the hand was most distal. Cadence was maintained between 90 and 100 rpm. Consistent with the procedures of Nordsborg et al. (26), the final arm-cranking exercise bout was followed by a 4-min rest period, during which participants immediately transferred to the cycle ergometer in preparation for the prescribed cycling test. Ventilatory and pulmonary gas exchange variables were averaged for the final 30 s of each arm-cranking exercise bout and for the final 30 s of each minute during the 4-min rest period prior to the prescribed cycling test.

Statistical analyses

Data were analyzed using a two-way (trial × time) repeated-measures ANOVA and Student’s paired t-tests, as appropriate. Relationships between variables were examined using Pearson’s product–moment correlation coefficient (r). Statistical significance was set at P < 0.05. Results are presented as mean ± SD unless otherwise stated.


Physiological effects of arm-cranking exercise

All participants successfully completed the arm-cranking protocol. Physiological data at rest were pooled from all trials. Repeated-measures ANOVA revealed no between-test differences in the ventilatory and pulmonary gas exchange responses to arm-cranking exercise (P > 0.05), and therefore, these data were pooled. Furthermore, repeated-measures ANOVA revealed no between-test differences in physiological responses at CYCONSET during L and AL trials (P > 0.05), and therefore, data from L and AL trials were pooled separately. Ventilatory and pulmonary gas exchange responses during intermittent arm-cranking exercise and during the 4-min rest period preceding the subsequent cycling test are shown in Figure 1. During AL trials, V˙E, V˙O2, and V˙CO2 were still elevated above rest at CYCONSET (P < 0.01). Heart rate, SpO2, and plasma acid–base balance responses at CYCONSET are shown in Table 1. Heart rate was higher at CYCONSET during AL compared with L (P < 0.01), whereas SpO2 was not different between trials. Arm-cranking resulted in different plasma acid–base balance responses between L and AL trials at CYCONSET. Specifically, at CYCONSET, [Na+] and [La] were 3 and 10.4 mEq·L−1 higher (P < 0.05 and 0.01, respectively), [Cl] was 2 mEq·L−1 lower (P < 0.05), and [PPr] was 0.7 g·dL−1 higher (P < 0.01) during AL compared with L. These differences in plasma ions and [PPr] affected the independent acid–base variables: [SID] was 5.4 mEq·L−1 lower and [Atot] was 2.0 mEq·L−1 higher during AL compared with L (P < 0.01). These differences in the independent acid–base variables also affected the dependent acid–base variables: [H+] was 12.7 nEq·L−1 higher and

was 7.4 mEq·L−1 lower during AL compared with L (P < 0.01).

Physiological responses at rest, immediately before cycling exercise (CYCONSET), and at the limit of cycling exercise tolerance (CYCEND) during incremental (INC) and constant power (CONST) exercise.
Ventilatory and pulmonary gas exchange responses to intermittent arm-cranking exercise. Dashed vertical lines represent the start and end of the arm-cranking protocol. Data points are mean ± SD and reflect the mean responses for the final 30 s of each arm-cranking exercise bout and for the final 30 s of each minute during the 4-min rest period before the prescribed cycling test.

There was a tendency for the ΔV˙O2W slope to be lower during ALINC (9.3 ± 0.6 mL·min−1·W−1) compared with LINC (10.5 ± 1.3 mL·min−1·W−1) (P = 0.06). Exercise duration (17.9 ± 0.8 vs 16.6 ± 1.0 min), W˙ max (358 ± 15 vs 332 ± 21 W), and V˙O2 peak (4.31 ± 0.36 vs 3.71 ± 0.44 L·min−1) were lower during ALINC compared with LINC (P < 0.05). That a maximal effort was exerted during ALINC is evidenced by all participants demonstrating a plateau in V˙O2, defined as an increase in V˙O2 <50% of the expected increase for a 20-W increment as determined from each participant’s ΔV˙O2W slope (29). The reduction in V˙O2 peak during ALINC was not correlated with the reduced exercise duration (r = 0.52, P = 0.23) or W˙max (r = 0.54, P = 0.22) but was correlated with the reduced ΔV˙O2W slope (r = 0.75, P < 0.05). A representative example of the V˙O2 response to incremental exercise is shown in Figure 2.

V˙O2 responses from a representative participant during LINC (▪) and ALINC (□). Note the lower V˙O2 slope and V˙O2peak during ALINC compared with LINC.

At CYCEND, heart rate was higher during LINC compared with ALINC (P < 0.01), whereas SpO2 was not different between trials (Table 1). [La] and [K+] were 2.4 and 0.48 mEq·L−1 higher during LINC compared with ALINC (P < 0.05), whereas there were no differences between trials for the independent acid–base variables [SID], [Atot], and PCO2. The dependent acid–base variable [H+] was 9.9 nEq·L−1 higher (P < 0.01), whereas

tended to be lower (P = 0.08), during LINC compared with ALINC.

Power–duration relationship and physiological responses at CYCEND during constant power exercise

Constant power exercise duration was 35 ± 15% shorter during ALCONST compared with LCONST trials (P < 0.01). The power–duration relationship was well described by the power–(1/time) model after both LCONST (r2 = 0.996 ± 0.003) and ALCONST (r2 = 0.993 ± 0.002) trials. CP was not different after LCONST (267 ± 19 W, 95% confidence interval = −8 to 8 W) and ALCONST (264 ± 20 W, 95% confidence interval = −10 to 11 W) trials. Conversely, W′ was 32 ± 6% lower after ALCONST (11.8 ± 4.2 kJ, 95% confidence interval = −2.7 to 2.6 kJ) compared with LCONST (17.3 ± 5.7 kJ, 95% confidence interval = −3 to 3 kJ) trials (P < 0.01) (Fig. 3). The SEE was low for both CP (2 ± 2 and 3 ± 1 W, representing 0.9 ± 0.7 and 1.1 ± 0.5% of the mean CP after LCONST and ALCONST trials, respectively) and W′ (0.93 ± 0.69 and 0.77 ± 0.42 kJ, representing 4.9 ± 2.1 and 6.3 ± 1.3% of the mean W′ after LCONST and ALCONST trials, respectively). Furthermore, estimates of CP and W′ from the power–(1/time) model were not different from those determined from the nonlinear power–time model (LCONST = 268 ± 21 W and 16.9 ± 6.4 kJ, ALCONST = 262 ± 22 W and 12.7 ± 4.7 kJ) and linear work–time model (LCONST = 267 ± 20 W and 17.0 ± 5.9 kJ, ALCONST = 263 ± 21 W and 12.1 ± 4.5 kJ), and each pair of values was highly correlated after LCONST (CP = r = 1.00, W′ = r ≥ 0.97, P < 0.01) and ALCONST (CP = r = 1.00, W′ = r ≥ 0.99, P < 0.01) trials. The parameter estimates were therefore associated with low levels of uncertainty (19,20).

The power–duration relationship in a representative participant after LCONST (•) and ALCONST (○) trials. CP and W′ are denoted by the y-intercept and slope, respectively, of the linear regression.

The mean V˙O2 peak was not different between LCONST (4.11 ± 0.19 L·min−1) and ALCONST (3.95 ± 0.35 L·min−1). Heart rate and SpO2 at CYCEND were not different between LCONST and ALCONST (Table 1). Conversely, [K+] was 0.14 mEq·L−1 lower, and [La] was 0.9 mEq·L−1 higher during ALCONST compared with LCONST (P < 0.05). The absolute increase in [K+] from CYCONSET to CYCEND was similar between ALCONST (0.61 ± 42 mEq·L−1) and LCONST (0.66 ± 41 mEq·L−1), although the shorter exercise duration in ALCONST meant that the rate of increase in [K+] was greater in ALCONST (0.14 ± 0.13 mEq·L−1·min−1) compared with LCONST (0.09 ± 0.08 mEq·L−1·min−1) (P < 0.05). At CYCEND [PPr] was 0.3 g·dL−1 higher during ALCONST compared with LCONST, which resulted in a 0.7 mEq·L−1 higher [Atot] (P < 0.05). There were no differences between LCONST and ALCONST for the independent acid–base variables [SID] and PCO2, or the dependent acid–base variables [H+] and



Consistent with our hypothesis, the major finding of the present study was that prior severe-intensity upper body exercise reduced leg cycling W′ without affecting CP. A novel aspect of the present study was that our experimental model allowed us to manipulate plasma, and presumably leg muscle, metabolite accumulation by performing prior upper body exercise. Although not measured in the present study, previous studies have reported constancy in leg intramuscular energy stores (ATP, phosphocreatine, and glycogen) after severe-intensity upper body exercise (2,3). Therefore, the reduction in W′ due to prior upper body exercise provides novel empirical support for the notion that the magnitude of W′ is partly dependent on metabolite accumulation. Furthermore, the constancy of CP means that the reduced exercise tolerance during ALCONST was exclusively dependent on the reduction in W′ and consistent with previous studies (13,14,27,34) that the physiological bases of CP are insensitive to metabolite accumulation.

Existing empirical support for the notion that W′ may depend on metabolite accumulation rather than intramuscular energy stores per se resides in a limited number of indirect observations. First, during severe-intensity exercise intramuscular phosphocreatine concentration may decline to a minimum well before exercise intolerance ensues (33), and at the limit of severe-intensity exercise tolerance considerable reserve exists in intramuscular phosphocreatine (∼10–40% of baseline) and ATP (∼83% of baseline) concentrations (10,13,23,33) (although depletion of individual muscle fibers is possible). Furthermore, continuation of exercise (via restoration of W′) after the limit of severe-intensity exercise tolerance has been reached is only possible if work rate is reduced below CP (10,11). Presumably this is because net clearance of fatigue-inducing metabolites can only occur at work rates below CP (10,11), although restoration of intramuscular phosphocreatine may also play a role. Second, irrespective of work rate, the limit of tolerance during severe-intensity exercise is associated with a consistent and, thus, potentially “critical” intramuscular pH and concentrations of inorganic phosphate and ADP (33). Third, the recovery kinetics of V˙O2 (a proxy for intramuscular phosphocreatine recovery) after severe-intensity exercise are slower than the recovery kinetics of W′ (13). Lastly, although leg intramuscular energy stores are unaffected by inspiratory muscle training, blood [La] and [H+] are attenuated (7), and W′ is increased in the absence of a change in CP (21).

Although these observations collectively suggest that W′ may depend on metabolite accumulation, to our knowledge, no previous study has characterized the power–duration relationship after the discrete manipulation of fatigue-inducing metabolites. Interestingly, when prior severe-intensity cycling exercise was performed before the criterion cycling exercise (i.e., the same muscle groups were used for both prior and criterion exercise), W′ was reduced by broadly the same extent (−34%) as the current findings (−32%), and CP was also unchanged (14). Despite dissimilar prior exercise protocols, the reductions in W′ followed broadly similar changes in the metabolic milieu: immediately prior to the constant power cycling tests used to determine the power–duration relationship plasma [La] was 11.6 mEq·L−1 in the current study, whereas whole blood [La] was 8.6 mEq.L−1 in the study of Ferguson et al. (14). It may seem surprising, therefore, that W′ was not reduced to a greater extent after prior exercise using the same muscle groups because in addition to metabolite accumulation, partial depletion of leg intramuscular energy stores must have also occurred. Consequently, resolving the relative impact of these two factors on reducing W′ is not possible and represents a limitation of the work of Ferguson et al. (14). Comparison of these studies is further complicated because “priming” effects resultant from prior exercise differ depending on whether the same (large influence) or different (negligible influence) muscle groups are used in the priming and criterion exercise bouts (16). Our experimental model allowed us to avoid the priming effect associated with prior exercise using the same muscles and presumably retain the leg intramuscular energy stores at CYCONSET (2,3). Therefore, by discretely manipulating the temporal profile of plasma and, presumably, leg muscle metabolite accumulation during subsequent cycling exercise, our findings provide novel empirical support for the notion that W′ at least partially depends on the accumulation of fatigue-inducing metabolites.

The mechanism(s) by which prior upper body exercise affects leg cycling exercise tolerance and hence W′ may partly reside in the effect of elevated plasma metabolites on previously resting leg muscle function (8). Although intracellular acidosis has long been considered a key mediator of muscle fatigue during severe-intensity exercise (15), this view has been challenged (3,8,26). Conversely, muscle fatigue during severe-intensity exercise has been causatively linked with an increased interstitial [K+], which induces a loss of excitability and contractility (8). Using the microdialysis technique, Nordsborg et al. (26) demonstrated a similar interstitial [K+] at the onset of single leg knee extensor exercise during L and AL. However, during leg exercise, K+ efflux from the active muscle and increases in interstitial [K+] were accelerated during AL compared with L and exercise tolerance was reduced. Consistent with these observations, we observed an accelerated increase in plasma [K+] during ALCONST compared with LCONST. However, muscle fatigue is a multifaceted process (1,8,15) that is difficult to resolve based on humoral measures per se, and greater insight into the mechanism(s) by which W′ is reduced during AL would come from studies using interstitial measurements, muscle biopsies, or 31P magnetic resonance spectroscopy. Using the latter technique, Vanhatalo et al. (33) have shown that the limit of tolerance during severe-intensity knee extensor exercise coincides, irrespective of the work rate, with the attainment of consistently low values of intramuscular phosphocreatine concentration and pH. Whether the limit of severe-intensity cycling exercise tolerance, after prior upper body exercise, is also associated with a consistent “critical” intramuscular milieu thus provides an interesting avenue for future investigation.

Exercise duration was 7% shorter during ALINC compared with LINC, which is considerably less than the 35% shorter exercise duration observed during ALCONST compared with LCONST. This difference may be attributed to the duration spent at sub-CP exercise intensities during the incremental exercise test, which would have prolonged the recovery period and thus increased restoration of W′ (13). Nevertheless, exercise duration/W˙max, V˙O2peak, and the ΔV˙O2W slope were still lower during ALINC compared with LINC. These findings contrast those of Boone et al. (5), who reported no change in these parameters during incremental cycling exercise preceded by maximal incremental arm-cranking exercise. However, compared with the present study, Boone et al. (5) used a longer intervening recovery period (6 min rest followed by 3 min of cycling at 50 W), and blood [La] at CYCONSET (8.4 mEq·L−1) was lower, which may explain these differences.

Elucidating the physiological mechanisms responsible for the lower ΔV˙O2W slope and V˙O2peak during ALINC compared with LINC was beyond the scope of the present study, and therefore the reasons for these observations remain unclear. The V˙O2 response to incremental cycling exercise is known to depend on changes in muscle blood flow (i.e., oxygen transport) and muscle fiber recruitment (i.e., oxygen utilization) (6). Indeed, during incremental exercise, ΔV˙O2W and V˙O2peak are reduced when oxygen transport is limited by breathing hypoxic air (35), whereas V˙O2 during constant power exercise is reduced by prior preferential fatigue or glycogen depletion of type II muscle fibers (9,12). These observations indicate that the lower V˙O2 response during ALINC compared with LINC may be explained by a limitation in oxygen transport and/or utilization, although further research is necessary to elucidate their relative contributions and the mechanism(s) by which they are influenced by prior upper body exercise.

In conclusion, prior severe-intensity upper body exercise reduced leg cycling W′ without affecting CP. This finding therefore provides novel empirical support for the notion that the magnitude of W′ is partly dependent on metabolite accumulation rather than a finite energy store per se.

The authors declare no conflict of interest.

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


1. Amann M. Central and peripheral fatigue: interaction during cycling exercise in humans. Med Sci Sports Exerc. 2011; 43 (11): 2039–45.
2. Bangsbo J, Aagaard T, Olsen M, Kiens B, Turcotte LP, Richter EA. Lactate and H+ uptake in inactive muscles during intense exercise in man. J Physiol. 1995; 488 (Pt 1): 219–29.
3. Bangsbo J, Madsen K, Kiens B, Richter EA. Effect of muscle acidity on muscle metabolism and fatigue during intense exercise in man. J Physiol. 1996; 495 (Pt 2): 587–96.
4. Barstow TJ, Jones AM, Nguyen PH, Casaburi R. Influence of muscle fibre type and fitness on the oxygen uptake/power output slope during incremental exercise in humans. Exp Physiol. 2000; 85 (1): 109–16.
5. Boone J, Bouckaert J, Barstow TJ, Bourgois J. Influence of priming exercise on muscle deoxy[Hb + Mb] during ramp cycle exercise. Eur J Appl Physiol. 2012; 112 (3): 1143–52.
6. Boone J, Koppo K, Barstow TJ, Bouckaert J. Effect of exercise protocol on deoxy[Hb + Mb]: incremental step versus ramp exercise. Med Sci Sports Exerc. 2010; 42 (5): 935–42.
7. Brown PI, Sharpe GR, Johnson MA. Loading of trained inspiratory muscles speeds lactate recovery kinetics. Med Sci Sports Exerc. 2010; 42 (6): 1103–12.
8. Cairns SP, Lindinger MI. Do multiple ionic interactions contribute to skeletal muscle fatigue? J Physiol. 2008; 586 (Pt 17): 4039–54.
9. Carter H, Pringle JS, Boobis L, Jones AM, Doust JH. Muscle glycogen depletion alters oxygen uptake kinetics during heavy exercise. Med Sci Sports Exerc. 2004; 36 (6): 965–72.
10. Chidnok W, Fulford J, Bailey SJ, et al. Muscle metabolic determinants of exercise tolerance following exhaustion: relationship to the ‘critical power.’ J Appl Physiol. 2013; 115 (2): 243–50.
11. Coats EM, Rossiter HB, Day JR, Miura A, Fukuba Y, Whipp BJ. Intensity-dependent tolerance to exercise after attaining V˙(O2) max in humans. J Appl Physiol. 2003; 95 (2): 483–90.
12. Deley G, Millet GY, Borrani F, Lattier G, Brondel L. Effects of two types of fatigue on the V˙O(2) slow component. Int J Sports Med. 2006; 27 (6): 475–82.
13. Ferguson C, Rossiter HB, Whipp BJ, Cathcart AJ, Murgatroyd SR, Ward SA. Effect of recovery duration from prior exhaustive exercise on the parameters of the power–duration relationship. J Appl Physiol. 2010; 108 (4): 866–74.
14. Ferguson C, Whipp BJ, Cathcart AJ, Rossiter HB, Turner AP, Ward SA. Effects of prior very-heavy intensity exercise on indices of aerobic function and high-intensity exercise tolerance. J Appl Physiol. 2007; 103 (3): 812–22.
15. Fitts RH. The cross-bridge cycle and skeletal muscle fatigue. J Appl Physiol. 2008; 104 (2): 551–8.
16. Fukuba Y, Hayashi N, Koga S, Yoshida T. V˙O(2) kinetics in heavy exercise is not altered by prior exercise with a different muscle group. J Appl Physiol. 2002; 92 (6): 2467–74.
17. Fukuba Y, Miura A, Endo M, Kan A, Yanagawa K, Whipp BJ. The curvature constant parameter of the power–duration curve for varied-power exercise. Med Sci Sports Exerc. 2003; 35 (8): 1413–8.
18. Hill DW, Smith JC. A method to ensure the accuracy of estimates of anaerobic capacity derived using the critical power concept. J Sports Med Phys Fitness. 1994; 34 (1): 23–37.
19. Hill DW. The critical power concept. A review. Sports Med. 1993; 16 (4): 237–54.
20. Johnson MA, Mills DE, Brown DM, Bayfield KJ, Gonzalez JT, Sharpe GR. Inspiratory loading intensity does not influence lactate clearance during recovery. Med Sci Sports Exerc. 2012; 44 (5): 863–71.
21. Johnson MA, Sharpe GR, Brown PI. Inspiratory muscle training improves cycling time-trial performance and anaerobic work capacity but not critical power. Eur J Appl Physiol. 2007; 101 (6): 761–70.
22. 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.
23. 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.
24. Karlsson J, Bonde-Petersen F, Henriksson J, Knuttgen HG. Effects of previous exercise with arms or legs on metabolism and performance in exhaustive exercise. J Appl Physiol. 1975; 38 (5): 763–7.
25. Miura A, Sato H, Sato H, Whipp BJ, Fukuba Y. The effect of glycogen depletion on the curvature constant parameter of the power–duration curve for cycle ergometry. Ergonomics. 2000; 43 (1): 133–41.
26. Nordsborg N, Mohr M, Pedersen LD, Nielsen JJ, Langberg H, Bangsbo J. Muscle interstitial potassium kinetics during intense exhaustive exercise: effect of previous arm exercise. Am J Physiol Regul Integr Comp Physiol. 2003; 285 (1): R143–8.
27. Parker Simpson L, Jones AM, Vanhatalo A, Wilkerson DP. Influence of initial metabolic rate on the power–duration relationship for all-out exercise. Eur J Appl Physiol. 2012; 112 (7): 2467–73.
28. 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.
29. Poole DC, Wilkerson DP, Jones AM. Validity of criteria for establishing maximal O2 uptake during ramp exercise tests. Eur J Appl Physiol. 2008; 102 (4): 403–10.
30. Rossing TH, Maffeo N, Fencl V. Acid–base effects of altering plasma protein concentration in human blood in vitro. J Appl Physiol. 1986; 61 (6): 2260–5.
31. Smith JC, Stephens DP, Hall EL, Jackson AW, Earnest CP. Effect of oral creatine ingestion on parameters of the work rate-time relationship and time to exhaustion in high-intensity cycling. Eur J Appl Physiol Occup Physiol. 1998; 77 (4): 360–5.
32. Stewart PA. Modern quantitative acid–base chemistry. Can J Physiol Pharmacol. 1983; 61 (12): 1444–61.
33. Vanhatalo A, Fulford J, DiMenna FJ, Jones AM. Influence of hyperoxia on muscle metabolic responses and the power–duration relationship during severe-intensity exercise in humans: a 31P magnetic resonance spectroscopy study. Exp Physiol. 2010; 95 (4): 528–40.
34. Vanhatalo A, Jones AM. Influence of prior sprint exercise on the parameters of the ‘all-out critical power test’ in men. Exp Physiol. 2009; 94 (2): 255–63.
35. Walsh ML, Banister EW. The influence of inspired oxygen on the oxygen uptake response to ramp exercise. Eur J Appl Physiol Occup Physiol. 1995; 72 (1–2): 71–5.


© 2014 American College of Sports Medicine