Methods

Subjects

Eight elite cyclists (age 26.1 years, SD 5.3; height 181.0 cm, SD 7.1; body mass 71.6 kg, SD 8.1; peak power output 437 W, SD 38; peak oxygen consumption (Vo_2peak) 70.6 ml/kg per minute, SD 6.8) were recruited for this study. Before participation in the study, the subjects were informed of the risks associated with the study. Informed consent was obtained in writing before the initiation of the study. All procedures conformed with the declaration of Helsinki.21 The Research and Ethics Committee of the Faculty of Health Sciences of the University of Cape Town Medical School approved the study.

Testing procedure

Subjects reported to the laboratory on four occasions. During the first two visits, subjects underwent anthropometric measurements, familiarisation trials and preliminary testing for measurement of Vo_2peak. The subject’s height (cm) and body mass (kg) were measured using a precision stadiometer and balance (model 770; Seca, Bonn, Germany; accurate to 10 g). Percentage body fat was calculated according to the equation of Durnin and Wormersley from skinfold measurements taken at seven sites.22 During the third and fourth laboratory visits (experimental trials), subjects performed two cycling trials in random order, after the ingestion of either 10 mg methylphenidate or a placebo. In addition to power output, heart rate and electromyographic measurements, arterial blood was drawn from the radial artery (blood trials; n  =  4) or respiratory gasses were collected (respiratory trials; n  =  4) for measurement at 10-minute intervals. Subjects were asked to refrain from eating or drinking for at least 6 h before each of the performance tests. One and a half hours before each test, subjects were asked randomly to ingest either 10 mg methylphenidate or placebo before consuming a standardised meal comprising a 150 g bran muffin and 250 ml orange juice. Each subject was asked to refrain from training for 24 h before the Vo_2peak test and to perform a 90-minute low-intensity “recovery ride” 24 h before the two experimental trials. They were also asked to refrain from consuming any caffeine or other stimulants on the day of each performance test. Before each testing session subjects were questioned to confirm that they had adhered to these instructions.

Preliminary testing

Tests were performed using either a Kingcycle or Computrainer Ergometer. Four subjects were tested using the Kingcycle ergometer. Due to the availability of the Computrainer ergometer, the remaining four subjects were tested using this device due to its greater capture rate, accuracy and reliability.23 Each device was calibrated according to the manufacturer’s guidelines before each testing session. Subjects completed a self-paced warm-up for 15 minutes before each testing session. During both the progressive exercise test and the familiarisation trials, the subjects were familiarised with the Borg RPE scale, and a standard set of instructions was given for the subsequent trials.

Progressive exercise test

Testing for Vo_2peak was performed at a starting work rate of 2.50 W/kg body mass. The load was increased incrementally at a rate of 20 W every 60 s until the subject could not sustain a cadence greater than 70. During the progressive exercise test, ventilation volume (V_E), oxygen consumption (Vo₂) and carbon dioxide production (Vco₂) were averaged over 15-s intervals using an on-line breath-by-breath gas analyser and pneumotach (Oxycon; Viasis, Hoechberg, Germany). Calibration of this device was performed before testing according to the manufacturer’s instructions. Vo_2peak was recorded as the highest Vo₂ reading recorded for 30 s during the test. Peak power output (PPO) was calculated by averaging the power output for the final minute of the Vo_2peak test. Subjects were requested to refrain from standing on the pedals throughout the test.

Familiarisation trial

Within a week of performing the Vo_2peak test, subjects reported to the laboratory for a familiarisation session, during which they underwent procedures identical to the experimental trials. Subjects completed a cycling trial at a fixed RPE under the same conditions and procedures as the experimental trials (described below).

Cycle trials at a fixed RPE

During experimental trials, subjects cycled at a fixed RPE. Subjects were instructed to cycle from the outset at a power output that they perceived represented an RPE of 16 on the Borg 6 to 20 RPE scale.24 This rating corresponded to the verbal cue of between “hard” and “very hard” on the Borg scale. The power output measured during the first 3 minutes of the trial was averaged to calculate an initial value. The trial was terminated when the power output fell below 70% of the initial value for two consecutive minutes. No feedback in terms of distance covered, time elapsed, power output or heart rate was provided to the subject at any time during any trial. Subjects were allowed to ingest water freely during the trials.

Electromyographic testing

Before each experimental session, subjects performed three maximum voluntary contraction (MVC) for normalisation of the electromyographic signal that was obtained from the vastus lateralis muscle during the subsequent cycling bouts. We have previously used this method for normalisation of the electromyographic signal obtained during submaximal6 18 cycling exercise. Electromyographic activity was recorded from the right vastus lateralis muscle after attachment of a triode electrode (Thought Technology, West Chazy, New York, USA) was placed over the muscle belly and connected to a pre-amplifier. Outputs from the pre-amplifier were relayed to a Flexcomp/DSP electromyography apparatus (Thought Technology) via a fibreoptic cable and stored by an on-line computer. Electromyographic signals were captured at 1984 Hz and analysed for 5-s periods during the MVC and for 5-s periods at each measurement period during the trials. Electromyographic data were analysed according to previously described methods.18 25 All processed integrated electromyographic data were normalised by dividing the integrated electromyography obtained at each workload during the trials by the integrated electromyography obtained during the MVC performed before the start of the trial. Electromyographic data are therefore expressed as a percentage of the integrated electromyography measured during the MVC. We have previously shown that this method of electromyographic normalisation is reliable and valid for use in cycling trials.18 25

Arterial blood samples

The subject’s left or right radial artery was cannulated and connected to a three-way tap and thin extension tubing containing diluted heparin (10 iu/ml).

Twenty millilitres of arterial blood was drawn before the start of trials and at 10-minute intervals during each trial for measurement of blood glucose, lactate, partial pressure oxygen, partial pressure carbon dioxide and pH.

Respiratory gas samples

Before the start and at 10-minute intervals during each trial, V_E, Vo₂ and Vco₂ were measured over 45-s intervals using an on-line breath-by-breath gas analyser and pneumotach (Oxycon; Viasys, Hoechberg, Germany) as described for the progressive exercise test.

Additional measurements

Heart rates were measured at the start of the trial and at 15-s intervals throughout the duration of the trial, using a Polar S720i heart rate monitor (Polar Electro OY, Kempele, Finland).

Power output readings were recorded at 30-s intervals by the Kingcycle ergometer and at 1-s intervals by the Computrainer ergometer. Power outputs were averaged for each 1 minute epoch before analysis.

RPE was recorded every 2 minutes. As required by the RPE clamp, subjects were required to ride constantly at an RPE of 16 using the Borg 6 to 20 RPE scale.24

Statistical analysis

All statistical analyses were performed using STATISTICA version 7.0. Power output, physiological variables and integrated electromyography were analysed using repeated measures analysis of variance. When a significant difference was found for either main effect (trial or time), a Tukey post hoc analysis was performed. Physiological variables were plotted against the percentage of peak power values for each subject using GRAPHPAD PRISM version 3.0. The slope and intercept values of the regression lines for each physiological variable were compared using a repeated-measures analysis of variance. Statistical significance was accepted when p<0.05. A significant interaction was interpreted as meaning that the trials responded differently over time for that variable. All data are expressed as means (SD).

Results

Exercise performance

Initial power outputs (305 W (SD 33) vs 303 W (SD 46); NS; methylphenidate vs placebo) were the same in both trials. Power output fell significantly more slowly in the methylphenidate compared with the placebo trial (1.02 W/minute (SD 0.28) vs 1.31 W/min (SD 0.42); p<0.05; fig 1). As a result, total trial duration (terminated when power output fell below 70% of the starting value for more than 2 minutes) was significantly longer in the methylphenidate compared with the placebo trial (88.0 minutes (SD 23.1) vs 68.3 minutes (SD 16.9); p<0.004).

Furthermore, the power output was significantly higher in the methylphenidate trial at the equivalent time (68.3 minutes, SD 16.9) at which the subject terminated in the placebo trial (251 W (SD 33) vs 211 W (SD 31); p<0.001; table 1).

Physiological variables

Heart rate

Mean heart rates were significantly higher in the methylphenidate than in the placebo trial (167 beats/minute (SD 7) vs 158 beats/minute (SD 10); p<0.01; fig 2A). Expressed relative to trial duration, heart rate decreased significantly in both trials (p<0.01; fig 2A).

Heart rates were significantly higher in the methylphenidate trial at the equivalent time at which subjects terminated the placebo trial (165 bpm (SD 11) vs 148 bpm (SD 10); p<0.005; table 1).

Vo₂, ventilatory volume and cycling economy

Vo₂ was significantly higher in the methylphenidate trial at the equivalent time at which subjects terminated the placebo trial (3865 ml/minute (SD 610) vs 3428 ml/minute (SD 500); p<0.05; table 1).

Ventilatory volumes (V′E) were significantly higher in the methylphenidate trial at the equivalent time at which subjects terminated the placebo trial (93 l/minute (SD 23) vs 83 l/minute (SD 22); p<0.05; table 1).

Mean values for cycling economy did not differ between the trials (14.4 ml/W (SD 0.9) vs 14.1 ml/W (SD 0.6); methylphenidate vs placebo). The relationship between Vo₂ and power output was not significantly different between groups (fig 3A), indicating that cycling economy was not altered by methylphenidate.

Arterial blood lactate concentrations

Arterial blood lactate concentrations did not differ significantly between the trials. When expressed as a percentage of completed distance, lactate concentrations fell significantly over time in both trials (p<0.001; fig 2B), reaching the lowest levels at exercise termination. Methylphenidate did not alter the relationship between arterial lactate concentration and percentage PPO (fig 3B).

Arterial lactate concentrations were significantly higher in the methylphenidate trial at the equivalent time at which subjects terminated the placebo trial (1.75 mmol/l (SD 1.06) vs 1.23 mmol/l (SD 0.74); p<0.05; table 1).

Arterial blood glucose concentrations

Arterial blood glucose concentrations did not differ significantly between trials. A significant time effect was observed (p<0.01). Arterial glucose values initially increased until approximately 40% of trial duration and subsequently decreased (fig 2C).

Substrate utilisation

Starting values for respiratory exchange ratios (RER) were the same but fell significantly (p<0.001) with time in both trials (fig 2D). RER values were significantly lower for the final 20% of the methylphenidate than for the placebo trial (p<0.05; fig 2D).

Arterial Po₂, Pco₂ and pH

The plots of arterial pH (fig 3C), Po₂ (fig 3D) and Pco₂ (fig 3E) against percentage PPO were not significantly different between trials.

Integrated electromyographic activity

Integrated electromyographic amplitude, expressed as a percentage of MVC, was not different at the start of the trials and fell significantly (p<0.001) with exercise duration (fig 2E). Although there were no differences in integrated electromyographic values between trials at the time at which subjects terminated the placebo trial, at 100% of trial duration integrated electromyographic amplitude was lower in the methylphenidate than in the placebo trial (p<0.01; fig 2E).

Discussion

Our first important finding was that methylphenidate slowed the rate at which power output fell when exercising at a fixed RPE (fig 1). As a result, power output was approximately 19% higher and exercise duration 32% longer in the methylphenidate than in the placebo trial. This occurred even though, by design, subjects perceived both exercise bouts to be equally demanding (table 1).

Second, this ergogenic effect of methylphenidate was not due to any reduction in metabolic or cardiorespiratory stress during exercise. Instead Vo₂ (13%), heart rates (12%), ventilatory rates (11%), and arterial lactate concentrations (43%) were all significantly higher during the methylphenidate trial although arterial Po₂, Pco₂ and pH were not altered (table 1).

Methylphenidate thus allowed subjects to sustain higher work rates and greater levels of metabolic and cardiorespiratory stress for longer, while perceiving the exercise stress to be identical. This indicates that subjects terminated the placebo trial with a metabolic and cardiorespiratory reserve that was not accessible without the use of methylphenidate. Methylphenidate allowed subjects to utilise this reserve presumably by removing the restraining effects of a CNS regulator1 as first proposed by Lehman et al26 70 years ago. This finding proves that amphetamines do not enhance performance simply by preventing “peripheral fatigue” but is explained by a model of exercise1 4 in which there is metabolic and cardiorespiratory reserve to ensure that whole-body homeostasis is always preserved.27

The more popular theory that peripheral fatigue causes a progressive mechanical failure28 29 30 predicts that methylphenidate could improve exercise performance by increasing the heart rate14 31 32 and thus the cardiac output. By increasing oxygen delivery to the active muscles, methylphenidate might reverse “anaerobic” conditions in the muscles, thereby allowing exercise to continue at a higher Vo₂. Four important findings disprove this interpretation.

First, arterial lactate concentrations were increased, not decreased, in the methylphenidate trials (table 1) suggesting that if lactate production was caused by anaerobiosis (a theory that has been disproved over time),33 then methylphenidate increased, rather than reduced such anaerobiosis. Furthermore, arterial lactate concentrations fell exponentially with exercise duration (fig 2B), dissociating (falling) arterial lactate concentrations as a cause of the linear reduction in power output as exercise progressed.

Second, integrated electromyographic activity fell during both trials (fig 2E) in keeping with our model, which predicts that changes in power output during prolonged exercise are due principally to altered central motor output to the exercising limb. This linear fall in power output was thus due to a progressive reduction in the number of motor units recruited in the exercising limbs, as predicted.1 4 In contrast, if the linear fall in power output had been due solely to “peripheral fatigue”, then integrated electromyographic activity would have either increased or stayed the same. Accordingly, this study proves that the reduction in power output during exercise was associated with reduced central motor command to the exercising limb to which any peripheral fatigue, secondary to noxious biochemical changes,30 may or may not have contributed.

Furthermore, the finding that the electromyographic activity was never greater than 45% of that achieved during a MVC indicates the presence of substantial motor recruitment reserve in the exercising limbs. Indeed, if peripheral fatigue was the sole cause of the impaired performance during prolonged exercise, the CNS should increase the number of motor units active in the exercising limb, thereby delaying the onset of exhaustion until 100% of available motor units were recruited. The finding that the CNS does not adopt this policy indicates that the maximisation of human exercise performance, regardless of cost, is not its principal priority. As the CNS chooses to reduce the number of motor units recruited at the same RPE during prolonged exercise, its principal priority must rather be the protection of skeletal muscle integrity and the prevention of a catastrophic mechanical failure.

Although integrated electromyographic activity was not significantly different at the time at which subjects terminated the placebo trial, integrated electromyography was not increased in the methylphenidate trial even though the power output was increased. Similarly, integrated electromyography and RER values recorded during the final 10% of the methylphenidate trials were significantly lower than those recorded at similar power outputs during the placebo trials. These values occurred at time points that exceeded the duration of the placebo trials. There are a number of possible explanations for this unexpected finding.

Either muscle contractility is increased by methylphenidate or these differences may be related to altered muscle recruitment patterns, for example a change from predominantly type II to type I muscle fibre recruitment secondary to glycogen depletion.34 35 36 37 38 Or as a number of the methylphenidate trials exceeded 90 minutes in duration, there may have been an alteration in whole muscle recruitment patterns in these very prolonged trials.39 40 Although we did not measure this, increases in muscle temperature induced by altered thermoregulatory control in the methylphenidate trials may also have reduced integrated electromyographic amplitude during the later stages of these trials.41 42 Furthermore, the greater duration of the methylphenidate trials may have affected integrated electromyographic amplitude due to increased sweat accumulation, electrode detachment or other factors affecting electromyographic signal strength.43 Finally, the absolute difference between the control group and the methylphenidate group of approximately 40 W may have been insufficient to measure differences in muscle recruitment between groups. In previous studies a difference of less than 50 W has not been associated with statistically significant differences in integrated electromyographic amplitude.18 44

Our third relevant finding was that methylphenidate did not improve exercise performance either by altering cycling economy (Vo₂/power relationship (fig 3A) was not altered), or by improving respiratory efficiency or gaseous exchange (blood gas measurements and the rate of ventilation were unchanged (fig 3C, D, E), or by altering metabolic substrate utilisation as the RER was the same in both trials (fig 2D).

Fourth, we found that the initial power outputs were not different between trials (table 1, fig 1). As the power output was selected on the basis of each subject’s perception of effort, this finding indicates that amphetamines do not alter those feed-forward controls that determine the choice of power output at the onset of exercise. However, methylphenidate caused the power output to fall less rapidly, presumably because it altered the manner in which the central regulator interpreted the continual afferent feedback from the numerous homeostatic sensors in the body. This allowed a higher power output and greater metabolic and cardiorespiratory stress to be tolerated at the same RPE.

Interestingly, amphetamines appear to have a greater ergogenic effect during prolonged submaximal exercise13 14 than during either maximal or short duration exercise.15 31 This suggests that amphetamines may selectively influence the central interpretation of afferent sensory feedback from some, but not all, homeostatic regulators. Afferent sensory feedback related to thermoregulation appears particularly sensitive to this effect. Amphetamines improved performance during maximal 20 km cycling time trials in hot environmental conditions,31 and a number of case reports describe the development of exertional heat stroke in athletes ingesting amphetamines.45

It might be argued that individuals do not normally exercise at a fixed RPE; rather the RPE typically rises as a function of the exercise duration that remains,5 whereas the work rate tends to show a biphasic pattern in which, following a fall, there is an increase near the end of exercise—the “end spurt”. However, we controlled RPE as we have previously shown that the RPE is directly involved in the regulation of exercise performance.6 18 We thus chose to investigate the effects of methylphenidate on power output independent of any effect on RPE.

Finally, we acknowledge that in high-intensity exercise of short duration (<2 minutes) there may be a significant contribution of peripheral regulation to the impairment of exercise performance.46

Conclusions

In summary, we confirm previous studies12 13 14 15 16 showing that human exercise performance is substantially prolonged by a stimulant acting principally on the CNS. This proves that performance in the natural state can be substantially improved by chemical modulation of those unknown processes in the CNS that regulate exercise performance. The finding that humans exercise with a metabolic and cardiorespiratory reserve in the natural state indicates that the protection of homeostasis and the avoidance of mechanical failure in the exercising limbs is the more important consideration during exercise, as it is at rest.1 27 Therefore, human exercise performance is not normally regulated with the aim of achieving an ultimate athletic performance without any concern for the biological consequences.