According to the “catastrophic” model of fatigue,1 fatigue occurs only after bodily system failure has occurred, leading rapidly to exhaustion. Hence, accumulation of metabolites such as lactic acid, or substrate depletion, induces catastrophic failure of homeostasis in the exercising muscles, leading directly to exercise termination.

Recently, Noakes and colleagues proposed a model of integrative central neural regulation of effort.2^–7 They claimed that there is no convincing evidence for failure of any of the major organ systems at exhaustion in healthy athletes. However, this model is not a simple reworking of the old central versus peripheral models. Rather, the authors explain that all physiological systems must be homoeostatically regulated specifically to prevent catastrophic failure, including irreversible widespread cell damage or even death. This model proposes that physical activity is controlled by a central governor in the brain so that the human body functions as a complex system during exercise. They propose that, during self-paced exercise, the central nervous system (CNS) continuously modifies the pace as part of a complex, non-linear dynamic system.8 9 Metabolic rate, the size of the fuel reserves and the rates of heat accumulation, amongst many other physiological variables, would be monitored in order to generate subconscious calculations that create a continuously adjusting power output during exercise.9 Indeed, according to this new model, biological changes in peripheral systems act as afferent signals to modulate control processes in the brain in a dynamic, non-linear, integrative manner. The brain achieves this control by continuously varying the work rate and metabolic demand.3

Thus, on the basis of feedforward control modified by afferent feedback from many physiological systems, the brain continuously alters pacing strategy, with the sensation of fatigue being the conscious interpretation of these homoeostatic, central governor control mechanisms.7 According to this novel explanation, the rising perception of discomfort progressively reduces the conscious desire to override this control mechanism, leading either to reduced exercise intensity or to the termination of exercise if the intensity cannot be reduced. Hence, fatigue is the conscious manifestation of the subconscious CNS processes.10 11

Recently Noakes12 has shown that marathoners finished races without evidence for a catastrophic failure of homoeostasis, confirming the anticipatory mechanism of central neural control in accordance with the central governor model. Nevertheless, to the best of our knowledge, no study has verified whether this model could be applied to exercises performed at higher intensities.

Hence, rather than trying to provide evidence for a “grand unified theory of fatigue”, the purpose of our present study was to observe physiological responses and the rating of perceived exertion (RPE)13 during endurance exercise performed until exhaustion to verify whether the central governor model could predict how fatigue develops at the maximal lactate steady state (MLSS) intensity that corresponds to the transition between aerobic and “anaerobic metabolism”.14

METHODS

Subjects and design

Eleven well-trained male subjects aged 20 (SD 2) years volunteered for this study. Mean height and body mass values were 1.80 (0.03) m and 73.5 (5.2) kg, respectively. All the subjects performed six tests on a cycle ergometer (Ergo-metrics 800, Ergoline, Germany) in an air-conditioned room (20°C) within a 15 day period as follows.

Methodology

Test 1

An incremental continuous cycling test was performed for the measurement of maximal aerobic power (MAP) and maximal oxygen uptake (VO_2max). This maximal incremental test was preceded by a 3 min warm-up period at 75 W. This initial workload was then increased by 25 W/min. To ensure that VO_2max was reached during the test, subjects were verbally encouraged to continue as long as possible. The test ended at the point of voluntary exhaustion. During the test, the subjects breathed through a face mask. Oxygen uptake (VO₂), carbon dioxide output (VCO₂), minute ventilation (VE) and respiratory rate (RR) were measured breath by breath using an open circuit system (CPX Medical Graphics, St Paul, Minnesota). The values were averaged for a 15 s period. The mean respiratory exchange ratio (RER) values were calculated from the recorded measurements. VO_2max and MAP were defined as the highest VO₂ and the highest power output attained at exhaustion. A 10-lead ECG (Quinton Q 810, Seattle, Washington) was continuously recorded during tests to determine heart rate (HR).

Tests 2, 3, 4, 5

Four 30 min tests were randomly performed at power output corresponding to 70, 75, 80 and 85% of the VO_2max measured in test 1. Tests were separated by 1 day of total rest and were randomised. These intensities were chosen in order to determine the single intensity corresponding to the MLSS for each subject. Finger-tip whole blood samples were taken every 5 min during each test and were analysed using the photometric method (Dr Lange miniphotometer plus LP20, GmbH & Co. KG, Germany) in order to determine the capillary blood lactate concentration ([La]_cap). MLSS was determined for each subject as the highest workload that could be maintained with a [La] increase lower than 1.0 mmol/l during the final 20 min of the appropriate 30 min tests.15 The [La] value at MLSS was calculated as the average [La] measured at the 10th, 15th, 20th, 25th and 30th min of the test.

Test 6

All subjects were asked to perform a test until exhaustion at the intensity corresponding to MLSS previously determined. VO₂, VCO₂, RER, VE, RR, and HR were continuously measured according to the procedures used in test 1. Nevertheless, because of the different times to exhaustion of the 11 subjects, these respiratory parameters and HR were expressed and analysed as percentage of time to exhaustion between 10% and 100% of time to exhaustion (t_10%, t_20%, t_30%, t_40%, t_50%, t_60%, t_70%, t_80%, t_90% and t_100%). During these tests, arterial blood was sampled from the radial artery through a catheter in order to determine arterial plasma lactate ([La]) and pyruvate ([Pyr]) concentrations, redox state ([La]/[Pyr]) (KONE LAB 30, Eragny, France), and ammonia concentrations ([NH₄]) (Synchron Beckman Coulter LX 20, Paris, France). In the same way, pH, arterial oxygen pressure (pO₂), arterial carbon dioxide pressure (pCO₂), arterial oxygen saturation (SO₂), haematocrit (Ht), and haemoglobin concentration ([Hb]) were measured, using electrochemical methodology (ABL 520, Radiometer Medical A/S, Bronshoj, Danemark) whereas bicarbonate concentration ([HCO₃⁻]) and base excess (BE) were calculated.

All these blood variables were measured at rest (t₀), and at the 10th, 20th, 30th min and at exercise termination (t₁₀, t₂₀, t₃₀, and t_end). The RPE was measured using Borg’s category ratio scale,13 which consists of 12 statements ranging from 0 to 10 (from “nothing” to “maximal”). Similarly general, muscular and ventilatory perceptions of exertion (genRPE, muscRPE, and ventilRPE, respectively) were measured on the basis of their responses using a modified Borg’s category ratio scale13 presented to the subjects at t₁₀, t₂₀, t₃₀, and t_end. Tympanic temperature was measured using an eardrum thermometer (Braun Thermo Scan ear thermometer, Gmbh, Kronberg/TS, Germany) at t₀ and t_end.

Statistical analysis

All calculations were performed using STATISTICA software (Statsoft 2000). Standard statistical methods were used for the calculation of means and standard deviations. Normal Gaussian distribution of the data was verified by the Shapiro–Wilks’ test. ANOVA with repeated measurements was used to compare the changes in physiological parameters during test 6. All data were compared at t₁₀, t_20, t₃₀, and t_end, except for heart rate and respiratory parameters, which were analysed minute-by-minute and expressed relative to the percentage of time to exhaustion. Compound symmetry, or sphericity, was verified by the Mauchley test. When the assumption of sphericity was not met, the significance of F ratios was adjusted according to the Greenhouse−Geisser procedure. Multiple comparisons were made with the Tukey HSD post hoc test when the Greenhouse–Geisser epsilon correction factor was >0.50, or with the Bonferroni post hoc test when the epsilon was <0.50. Statistical significance was set at p = 0.05 level for all analyses.

RESULTS

Incremental continuous test (Test 1)

Mean maximal aerobic power value reached by the subjects was 323.2 (30.2) W and corresponded to a VO_2max of 54.8 8.2 ml/min^/kg (table 1). HR_max was 203 (SD 18) beats/min. The maximal values of HR, VO₂, VCO₂, RER and VE attained during the test (HR_max, VO_2max, VCO_2max, RER_max, VE_max) are reported in table 1.

Submaximal tests (Tests 2, 3, 4, 5)

In each subject, [La]_cap increased more than 1 mmol/l between t₁₀ and t₃₀ in at least one of the tests. Mean MLSS intensity was determined using the different submaximal tests corresponding to 71.3% (5.2%) of VO_2max. Mean capillary blood lactate concentration value corresponding to MLSS was 5.5 (1.5) mmol/l and ranged from 3.6 to 7.9 mmol/l.

Test at MLSS (Test 6)

Time to exhaustion

Subjects were able to maintain the exercise intensity corresponding to MLSS for 55.0 (SD 8.5) min.

Cardiorespiratory and thermoregulatory parameters

All the mean values reported during the test were significantly higher than those measured at rest (p<0.05) (table 2). Mean values of VO₂, VCO₂ and RER did not vary between t_10% and t_100% (2.7389 (SD 0.5594) vs 2.8913 (SD 0.3438) l/min, 2.6697 (SD 0.5893) vs 2.7586 (SD 0.3678) l/min and 0.97 (SD 0.04) vs 0.95 (SD 0.05), respectively, p>0.05; table 2) whereas RR and VE increased significantly (28.4 (SD 4.8) vs 47.4 (SD 7.6) cycles/min and 64.3 (SD 17.2) vs 85.4 (SD 14.8) l/min respectively, p <0.05). The Bonferroni post hoc test showed that RR increased after t_40%, whereas VE increased between t_10% and t_60%, t_10% and t_80%, and t_10% and t_100%, but did not evolve after t_20%. Some of these findings are presented graphically in fig 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Respiratory parameters (oxygen uptake (VO₂), carbon dioxide output (VCO₂), minute ventilation (VE), and respiratory rate (RR)) during the test performed to exhaustion at MLSS (Test 6). *significant variation (p<0.05) between t_10% and t_100% of time to exhaustion.

HR increased between t_10% and t_100% (151 (SD 19) vs 175 (SD 14) beats/min, p<0.05; fig 2). The Bonferroni post hoc test showed that HR values increased significantly between t_10% and t_40%, t_10% and t_60%, t_10% and t_80%, t_10% and t_90%, and t_10% and t_100%. Even though exercise was performed in an air-conditioned environment (20°C), body temperature significantly increased between t₀ and t_end (37.3 (SD 0.5) vs 38.4 (SD 0.7)°C, p<0.05).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Heart rate (HR) during the test performed to exhaustion at MLSS (test 6). *significant variation (p<0.05) between t_10% and t_100% of time to exhaustion.

Blood parameters and RPE

Mean values and standard deviations of blood parameters at t₀, t₁₀, t₂₀, t₃₀, and t_end are presented in table 2. With the exception of pCO₂ and pH values, which were significantly lower (p <0.05), all the mean values reported at t₁₀, t₂₀, t₃₀ and t_end were significantly higher during exercise than were values measured at rest. No blood parameters varied significantly between t₁₀ and t_end except [NH₄], BE and pH values, which increased significantly (p <0.05), and [La] and PCO₂ values, which decreased significantly (p <0.05). Mean [Pyr] and [La]/[Pyr] values at t₁₀, t₂₀, t₃₀, and t_end were 18.56 (SD 3.76) mmol/l and 28.1 (SD 3.7) mmol/l, respectively. Mean pO₂, SO₂ and [HCO₃⁻] values at t₁₀, t₂₀, t₃₀, and t_end were 91.86 (SD 6.26) mm Hg, 95.45% (SD 0.97%) and 20.88 (SD 2.39) mmol/l, respectively. General RPE (genRPE), muscular RPE (muscRPE) and ventilatory RPE (ventilRPE) values at t₁₀, t₂₀, t₃₀, and t_end are also presented in table 2. Whereas genRPE and muscRPE significantly increased between t₁₀ and t_end (3 (SD 1) vs 6 (SD 2) and 3 (SD 2) vs 6 (SD 2), respectively, p<0.05), ventilRPE remained stable (2 (SD 1), p>0.05).

DISCUSSION

Similar to the data reported in the literature, the MLSS determined from tests 2, 3, 4, and 5 occurred at 71.3% (SD 5.2%) of VO2_max14 15 and time to exhaustion during the test performed at MLSS was 55.0 (SD 8.5) min.16

Accordingly, our first important finding was that termination of exercise at the intensity corresponding to the MLSS study occurred without any clear evidence for a failure of homoeostasis in either the cardiorespiratory, metabolic or acid–base systems:

Cardiovascular, respiratory and thermoregulatory parameters

VO₂, VCO₂, RER, RR, and VE values obtained at the end of the MLSS test remained lower than maximal values reached during incremental test (p >0.05) and respiratory parameters could not be considered as factors causing the termination of exercise during the MLSS test.

The increase in HR values between t_10% and t_100% (fig 2, p<0.05) could be explained by an increase in sympathetic nervous system activity and an increase in circulating norepinephine concentrations [NE],15 but also by hyperthermia and associated mechanisms to maintain cardiac output17 18 and to dissipate the heat.19 20

The MLSS test was performed in an air-conditioned environment (20°C) and the body temperature at the end of exercise was only 38.4 (SD 0.7)°C, well below values of >40°C known to cause the termination of exercise in the heat for well trained subjects.21^–23 Thus the achievement of a critical body temperature was not likely the cause of exercise termination.

In addition, whereas [Hb], [Ht] and osmolarity increased between t₀ and t₁₀, probably as a result of fluid shifts during the early part of exercise,24 25 there was no further change in any of these variables between t₁₀ and t_end (table 2). Hence, a progressive loss of circulating blood volume did not occur and could therefore also not have explained either the progressive rise in HR or the termination of exercise.

In addition, the final HR at the end of the MLSS was significantly less than the HR_max during the maximum exercise test (175 (SD 14) beats/min vs 203 (SD 18) beats/min, p<0.05) and can also not be considered as a factor causing the termination of exercise during the MLSS test. Likewise, SO₂ and PO₂ did not decrease significantly during the MLSS test and therefore also cannot be considered as factors causing the termination of exercise.

Metabolic parameters

The stability of arterial [La] between t₁₀ and t₃₀ (p>0.05) of the MLSS test (table 2) confirmed that the intensity imposed during this test produced a stable arterial lactate concentration. [Pyr] and the [La]/[Pyr] ratio also produced a steady state in the redox state.15 26

In fact, arterial [La] decreased slightly between t₂₀ and t_end during the MLSS, indicating either reduced lactate production or increased peripheral utilisation.27 More to the point, however, the absence of a progressive increase in the arterial [La] rules out the arterial lactate as a cause of exercise termination in this study, in line with the more modern interpretation that lactate is not the cause of fatigue during this form of exercise.28 29

[NH₄] responses are different from [La] responses during the MLSS test, in accordance with the study of Ogino et al.30 The progressive increase in arterial [NH₄] during the MLSS test might have been due to deamination of AMP to IMP and ammonia in the purine nucleotide cycle,31 and by the catabolism of amino acids by active skeletal muscle. Alternatively, others have suggested that the increase in blood [NH₄] during prolonged exercise may be caused by hyperthermia32 and falling muscle glycogen concentrations.33 Nevertheless, if partial glycogen depletion is observed during prolonged exercise,34 this mechanism cannot be held as directly responsible for exhaustion.2

Although an increase in blood [NH₄] concentrations has been associated with muscle fatigue and exhaustion, including motor dysfunction, perhaps as a result of so-called central fatigue,35 36 as yet no direct causal relationship between this blood variable and fatigue has been proven. In contrast, Baron et al37 have already shown that exercise performed at a higher exercise intensity (Critical Power: 85.4% (SD 4.8%) of VO_2max) could be maintained even though arterial [NH₄] was higher than that observed at MLSS.

Hence, there is presently no evidence to conclude that an increase in blood [NH₄] is the sole and exclusive cause of exercise termination. Furthermore, if [NH₄] does indeed play a role in the development of fatigue during exercise, its effect would likely be as a result of a central (brain) effect rather than one in the peripheral tissues.35

Acid–base status

Even though arterial pH decreased between t₀ and t₁₀ (p <0.05), it increased slightly after t₁₀, whereas paCO₂ decreased and [HCO₃⁻] did not change (table 2), suggesting a ventilatory compensation for a mild exercise-induced metabolic acidosis15 or hyperthermically induced hyperventilation.38 Furthermore, arterial pH at the termination of exercise (∼7.37 units) was much above values considered to cause fatigue by a direct effect in the exercising muscles.39 40

In summary, exhaustion at MLSS occurred whilst homoeostasis was still present in all the major bodily systems that we studied, according to the central governor model,2^–7 whereas the perceptions of discomfort increased progressively until exhaustion, similar to the findings of Baron et al16 during a 2 h swimming test. Hence, genRPE and muscRPE increased whereas ventilRPE remained constant.

CONCLUSION

The results of this study show that, during exercise performed at MLSS, exhaustion occurred while physiological reserve capacity still existed, but in association with an increase in the ratings of perceived exertion, as predicted by the central governor model. Exercise termination may be induced by an integrative homoeostatic control of the peripheral physiological system specifically to ensure the maintenance of homeostasis.