Caffeine lowers perceptual response and increases power output during high-intensity cycling

The aim of this study was to determine the effects of caffeine ingestion on a ‘preloaded’ protocol that involved cycling for 2?min at a constant rate of 100% maximal power output immediately followed by a 1-min ‘all-out’ effort. Eleven male cyclists completed a ramp test to measure maximal power output. On two other occasions, the participants ingested caffeine (5?mg?·?kg^?1) or placebo in a randomized, double-blind procedure. All tests were conducted on the participants' own bicycles using a Kingcycle? test rig. Ratings of perceived exertion (RPE; 6–20 Borg scale) were lower in the caffeine trial by approximately 1 RPE point at 30, 60 and 120?s during the constant rate phase of the preloaded test (P?<0.05). The mean power output during the all-out effort was increased following caffeine ingestion compared with placebo (794±164 vs 750±163?W; P?=?0.05). Blood lactate concentration 4, 5 and 6?min after exercise was also significantly higher by approximately 1?mmol?·?l^?1 in the caffeine trial (P?<0.05). These results suggest that high-intensity cycling performance can be increased following moderate caffeine ingestion and that this improvement may be related to a reduction in RPE and an elevation in blood lactate concentration.     

Caffeine withdrawal and high-intensity endurance cycling performance

In this study, we investigated the impact of a controlled 4-day caffeine withdrawal period on the effect of an acute caffeine dose on endurance exercise performance. Twelve well-trained and familiarized male cyclists, who were caffeine consumers (from coffee and a range of other sources), were recruited for the study. A double-blind placebo-controlled cross-over design was employed, involving four experimental trials. Participants abstained from dietary caffeine sources for 4 days before the trials and ingested capsules (one in the morning and one in the afternoon) containing either placebo or caffeine (1.5 mg · kg(-1) body weight · day(-1)). On day 5, capsules containing placebo or caffeine (3 mg · kg(-1) body weight) were ingested 90 min before completing a time trial, equivalent to one hour of cycling at 75% peak sustainable power output. Hence the study was designed to incorporate placebo-placebo, placebo-caffeine, caffeine-placebo, and caffeine-caffeine conditions. Performance time was significantly improved after acute caffeine ingestion by 1:49 ± 1:41 min (3.0%, P = 0.021) following a withdrawal period (placebo-placebo vs. placebo-caffeine), and by 2:07 ± 1:28 min (3.6%, P = 0.002) following the non-withdrawal period (caffeine-placebo vs. caffeine-caffeine). No significant difference was detected between the two acute caffeine trials (placebo-caffeine vs. caffeine-caffeine). Average heart rate throughout exercise was significantly higher following acute caffeine administration compared with placebo. No differences were observed in ratings of perceived exertion between trials. A 3 mg · kg(-1) dose of caffeine significantly improves exercise performance irrespective of whether a 4-day withdrawal period is imposed on habitual caffeine users.     

Caffeine supplementation and peak anaerobic power output

Abstract

The aim of this study was to investigate the effects of caffeine supplementation on peak anaerobic power output (W_max). Using a counterbalanced, randomised, double-blind, placebo-controlled design, 14 well-trained men completed three trials of a protocol consisting of a series of 6-s cycle ergometer sprints, separated by 5-min passive recovery periods. Sprints were performed at progressively increasing torque factors to determine the peak power/torque relationship and W_max. Apart from Trial 1 (familiarisation), participants ingested a capsule containing 5 mg·kg^?1 of caffeine or placebo, one hour before each trial. The effects of caffeine on blood lactate were investigated using capillary samples taken after each sprint. The torque factor which produced W_max was not significantly different (p ≥ 0.05) between the caffeine (1.15 ± 0.08 N·m·kg^?1) and placebo (1.13 ± 0.10 N·m·kg^?1) trials. There was, however, a significant effect (p < 0.05) of supplementation on W_max, with caffeine producing a higher value (1885 ± 303 W) than placebo (1835 ± 290 W). Analysis of the blood lactate data revealed a significant (p < 0.05) torque factor × supplement interaction with values being significantly higher from the sixth sprint (torque factor 1.0 N·m·kg^?1) onwards following caffeine supplementation. The results of this study confirm previous reports that caffeine supplementation significantly increases blood lactate and W_max. These findings may explain why the majority of previous studies, which have used fixed-torque factors of around 0.75 N·m·kg^?1 and thereby failing to elicit W_max, have failed to find an effect of caffeine on sprinting performance.     

The influence of cadence and power output on force application and in-shoe pressure distribution during cycling by competitive and recreational cyclists

The aim of this study was to determine the response of cyclists to manipulations of cadence and power output in terms of force application and plantar pressure distribution. Two groups of cyclists, 17 recreational and 12 competitive, rode at three nominal cadences (60, 80, 100 rev min -1 ) and four power outputs (100, 200, 300, 400 W) while simultaneous force and in-shoe pressure data were collected. Two piezoelectric triaxial force transducers mounted in the right pedal measured components of the pedal force and orientation, and a discrete transducer system with 12 transducers recorded the in-shoe pressures. Force application was characterized by calculating peak resultant and peak effective pedal forces and positive and negative impulses. In-shoe pressures were analysed as peak pressures and as the percent relative load. The force data showed no significant group effect but there was a cadence and power main effect. The impulse data showed a significant three-way interaction. Increased cadence resulted in a decreased positive impulse, while increased power output resulted in an increased impulse. The competitive group produced less positive impulse but the difference became less at higher cadences. Few between-group differences were found in pressure, notable only in the pressure under the first metatarsal region. This showed a consistent pattern of in-shoe pressure distribution, where the primary loading structures were the first metatarsal and hallux. There was no indication that pressure at specific sites influenced the pedal force application. The absence of group differences indicated that pressure distribution was not the result of training, but reflected the intrinsic relationship between the foot, the shoe and the pedal.     

The influence of cadence and power output on force application and in-shoe pressure distribution during cycling by competitive and recreational cyclists

The aim of this study was to determine the response of cyclists to manipulations of cadence and power output in terms of force application and plantar pressure distribution. Two groups of cyclists, 17 recreational and 12 competitive, rode at three nominal cadences (60, 80, 100 rev x min(-1)) and four power outputs (100, 200, 300, 400 W) while simultaneous force and in-shoe pressure data were collected. Two piezoelectric triaxial force transducers mounted in the right pedal measured components of the pedal force and orientation, and a discrete transducer system with 12 transducers recorded the in-shoe pressures. Force application was characterized by calculating peak resultant and peak effective pedal forces and positive and negative impulses. In-shoe pressures were analysed as peak pressures and as the percent relative load. The force data showed no significant group effect but there was a cadence and power main effect. The impulse data showed a significant three-way interaction. Increased cadence resulted in a decreased positive impulse, while increased power output resulted in an increased impulse. The competitive group produced less positive impulse but the difference became less at higher cadences. Few between-group differences were found in pressure, notable only in the pressure under the first metatarsal region. This showed a consistent pattern of in-shoe pressure distribution, where the primary loading structures were the first metatarsal and hallux. There was no indication that pressure at specific sites influenced the pedal force application. The absence of group differences indicated that pressure distribution was not the result of training, but reflected the intrinsic relationship between the foot, the shoe and the pedal.     

The influence of cadence and power output on the biomechanics of force application during steady-rate cycling in competitive and recreational cyclists

The intent of this study was two-fold. The first aim was to investigate how cyclists orient forces applied by the feet to the pedals in response to varying power output and cadence demands, and the second was to assess whether competitive riders responded differently from recreational riders to such variations. One group consisted of US Cycling Federation category II licensed competitive cyclists (n = 7) and the second group consisted of recreational cyclists with no competitive experience (n = 38). The subjects rode an instrumented stationary 10-speed geared bicycle mounted on a platform designed to provide rolling and inertial resistance for six pedal rate/power output conditions for a minimum of 2 min for each ride. The pedalling rates were 60, 80 and 100 rev min-1 and the power outputs 100 and 235 W. All rides were presented in random order. The forces applied to the pedals, the pedal angle with respect to the crank and the crank angle were recorded for the final 30 s of each ride. From these data, a number of variables were computed including peak normal and tangential forces, crank torque, angular impulse, proportion of resultant force perpendicular to the crank, and pedal angle. Both the competitive and recreational groups responded similarly to increases in cadence and power output. There was a decrease in the peak normal forces, whereas the tangential component remained almost constant as cadence was increased. Regardless of cadence, the riders responded to increased power output demands by increasing the amount of positive angular impulse. All the riders had a reduced index of effectiveness as cadence increased. This was found to be the result of the large effect of the forces during recovery on this calculation. There were no significant differences between the two groups in each of these variables over all conditions. It was concluded that the lack of difference between the groups was a combined consequence of the limited degrees of freedom associated with the bicycle and that the relatively low power output for the competitive riders was insufficient to discriminate or highlight superior riding technique.     

Optimising distribution of power during a cycling time trial

A simple mathematical model is used to find the optimal distribution of a cyclist’s effort during a time trial. It is shown that maintaining a constant velocity is optimal if the goal is to minimise the time taken to complete the course while fixing amount of work done. However, this is usually impractical on a non-flat course because the cyclist would be unable to maintain the power output required on the climbs. A model for exertion is introduced and used to identify the distribution of power that minimises time while restricting the cyclist’s exertion. It is shown that, for a course with a climb followed by a descent, limits on exertion prevent the cyclist from improving performance by shifting effort towards the climb and away from the descent. It is also shown, however, that significant improvement is possible on a course with several climbs and descents. An analogous problem with climbs and descents replaced by headwinds and tailwinds is considered and it is shown that there is no significant advantage to be gained by varying power output. Lagrange multipliers are used solve the minimisation problems.     

Knee extensor fatigue developed during high-intensity exercise limits lower-limb power production

We investigated the association between changes in vastii electromyography (EMG) and knee extensor fatigue during high-intensity cycling, and the subsequent effect on lower-limb power and intermuscular coordination during all-out cycling. On two separate days, participants completed 30-s all-out cycling or 10-min of high-intensity cycling followed by 30-s all-out cycling. EMG for gluteus maximus (GMAX), rectus femoris (RF), vastii (VAS), hamstrings (HAM) and gastrocnemius (GAS); co-activation for GMAX/RF, VAS/HAM and VAS/GAS; isometric maximal voluntary force (IMVF) and resting twitch (RT) of the knee extensors were measured. VAS EMG increases during high-intensity cycling (6% to 14%, P < 0.05) were negatively correlated (r = ?0.791, P < 0.05) with knee extensor IMVF decreases (?2% to?36%, P < 0.05) following the exercise. Knee extensor IMVF decreases were positively correlated (r = 0.757, P < 0.05) with all-out cycling power reductions (0% to ?27%, P < 0.05). VAS/GAS co-activation did not change (P > 0.05) during all-out cycling while VAS and GAS EMG decreased. Larger increase in VAS EMG during high-intensity cycling was associated with greater knee extensor fatigue and larger power reduction during all-out cycling. High VAS/GAS co-activation potentially limited power reduction induced by knee extensor fatigue during all-out cycling.     

Effects of high-intensity intermittent priming on physiology and cycling performance

The pre-event warm-up or “priming” routine for optimising cycling performance is not well-defined or uniform to a specific event. We aimed to determine the effects of varying the intensity of priming on 3 km cycling performance. Ten endurance-trained male cyclists completed four 3 km time-trials (TT) on four separate occasions, each preceded by a different priming strategy including “self-selected” priming and three intermittent priming strategies incorporating 10 min of constant-load cycling followed by 5 × 10 s bouts of varying relative intensity (100% and 150% of peak aerobic power, W_peak, and all-out priming). The self-selected priming trial (379 ± 44 W) resulted in similar mean power during the 3 km TT to intermittent priming at 100% (376 ± 45 W; ?0.7%; unclear) and 150% (374 ± 48 W; ?1.5%, unclear) of W_peak, but significantly greater than all-out priming (357 ± 45 W; ?5.8%, almost certainly harmful). Differences between intermittent and self-selected priming existed with regards to heart rate (6.2% to 11.5%), blood lactate (?22.9% to 125%) and VO₂ kinetics (?22.9% to 8.2%), but these were not related to performance outcomes. In conclusion, prescribed intermittent priming strategies varying in intensity did not substantially improve 3 km TT performance compared to self-selected priming.     

Effects of magnitude and frequency of variations in external power output on simulated cycling time-trial performance

Abstract

Mechanical models of cycling time-trial performance have indicated adverse effects of variations in external power output on overall performance times. Nevertheless, the precise influences of the magnitude and number of these variations over different distances of time trial are unclear. A hypothetical cyclist (body mass 70 kg, bicycle mass 10 kg) was studied using a mathematical model of cycling, which included the effects of acceleration. Performance times were modelled over distances of 4–40 km, mean power outputs of 200–600 W, power variation amplitudes of 5–15% and variation frequencies of 2–32 per time-trial. Effects of a “fast-start” strategy were compared with those of a constant-power strategy. Varying power improved 4-km performance at all power outputs, with the greatest improvement being 0.90 s for ± 15% power variation. For distances of 16.1, 20 and 40 km, varying power by ± 15% increased times by 3.29, 4.46 and 10.43 s respectively, suggesting that in long-duration cycling in constant environmental conditions, cyclists should strive to reduce power variation to maximise performance. The novel finding of the present study is that these effects are augmented with increasing event distance, amplitude and period of variation. These two latter factors reflect a poor adherence to a constant speed.     

Ischemic preconditioning enhances critical power during a 3 minute all-out cycling test

This study tested the hypothesis that ischemic preconditioning (IPC) would increase critical power (CP) during a 3 minute all-out cycling test. Twelve males completed two 3 minute all-out cycling tests, in a crossover design, separated by 7 days. These tests were preceded by IPC (4 x 5 minute intervals at 220 mmHg bilateral leg occlusion) or SHAM treatment (4 x 5 minute intervals at 20 mmHg bilateral leg occlusion). CP was calculated as the mean power output during the final 30 s of the 3 minute test with W′ taken as the total work done above CP. Muscle oxygenation was measured throughout the exercise period. There was a 15.3 ± 0.3% decrease in muscle oxygenation (TSI; [Tissue saturation index]) during the IPC stimulus, relative to SHAM. CP was significantly increased (241 ± 65 W vs. 234 ± 67 W), whereas W′ (18.4 ± 3.8 vs 17.9 ± 3.7 kJ) and total work done (TWD) were not different (61.1 ± 12.7 vs 60.8 ± 12.7 kJ), between the IPC and SHAM trials. IPC enhanced CP during a 3 minute all-out cycling test without impacting W′ or TWD. The improved CP after IPC might contribute towards the effect of IPC on endurance performance.     

Determination of the peak power output during maximal brief pedalling bouts

The purpose of this study was to propose an optimization procedure for determining power output during very brief maximal pedalling exercise. Twenty-six healthy male students (21-28 years) performed anaerobic tests on a Monark bicycle ergometer with maximal effort for less than 10 s at eight different loads ranging from 28.1 to 84.2 Nm in pedalling moment. The maximal pedalling rate was determined from the minimal time required for one rotation of the cycle wheel. Pedalling rate decreased linearly with the load. The relationship between load and pedalling rate was represented by two linear regression equations for each subject; one regression equation was determined from eight pairs of pedalling rates and loads (r less than -0.976) and the other from three pairs (at 28.1, 46.8, 65.5 Nm; r less than -0.969). The two regression coefficients of the respective regression equations were almost identical. Mean +/- S.D. of maximal power output (Pmax) which was determined for each subject based on the two linear regression equations for eight pairs and three pairs of pedalling rates and loads was 930 +/- 187 W (13.4 +/- 1.6 W kgBW-1) and 927 +/- 187 W (13.4 +/- 1.6 W kgBW-1), respectively. There was no statistically significant difference between the values of Pmax which were obtained from each equation. It was concluded that maximal anaerobic power could be simply determined by performing maximal cycling exercise at three different loads.     

Reliability of power output during rowing changes with ergometer type and race distance

Coaches, sport scientists and researchers assess rowing performance on-water and on a variety of ergometers. Ergometers are frequently used because of the easier assessment environment. However, there is limited information on the ability of rowers to reproduce mean power or time-trial time when using different rowing ergometers (Concept II and RowPerfect) or completing tests over different distances (500 m versus 2000 m races). To test the efficacy of an intervention on a rower's ability to produce power, or to monitor that ability, it is essential to determine a reliable rowing performance test. The per cent standard error of measurement in performance (assessed by mean power and time-trial time) of fifteen national standard rowers was determined for five repeated 500 m and two repeated 2000 m races on a Concept II and RowPerfect ergometer. The per cent standard error of measurement (% SEM) in mean power between 5x500m races, regardless of gender, was 2.8% (95% confidence limits (CL)=2.3 to 3.4%) for the Concept II ergometer and 3.3% (95% CL=2.5 to 3.9%) for the RowPerfect ergometer (n = 15). Over 2000 m the per cent standard error of measurement in mean power was 1.3% (95% CL 0.9 to 2.9%) for the Concept II ergometer and 3.3% (95% CL 2.2 to 7.0%) for the RowPerfect ergometer The results highlight an increase in per cent standard error of the mean during performance races of less than 2000m on the Concept II ergometer, and performance races on the RowPerfect ergometer compared with the Concept II ergometer over 500 m and 2000 m. The most appropriate protocol for testing the influence of an intervention on the ability of a rower to produce power would be 2000 m races on a Concept II ergometer.     

Reliability of power output measurements during repeated treadmill sprinting in rugby players

The non-motorized treadmill system initially reported by Lakomy in 1984 has been used extensively to assess sprinting performance. However, there has been limited research into the reliability of power output measurement using such systems. The aim of this study was to design a system and protocol capable of measuring treadmill sprinting performance in rugby players and to assess the reliability of this system for measuring power output. Twenty-seven rugby players, all of whom were familiar with treadmill sprinting, performed three maximal 6 s sprints with 2 min recovery between sprints, on two occasions 1 week apart. Both tests were performed on a non-motorized Woodway tramp treadmill, interfaced to a data acquisition system. There were no significant differences (P > 0.05) between power output for repeated trials on the same day (between trials) or for repeated trials on different days (between days). Limits of agreement for maximum average power (the average of 100 readings per second) were 4+/-98 and 30+/-157 W for between trials and between days, respectively. When reported as ratio limits of agreement, these were 1.07 (*/divided 1.12) and 1.03 (*/divided 1.16), respectively. The limits of agreement for maximum instantaneous power (the highest of 100 readings per second) were 51+/-464 and 105+/-588 W for between trials and between days, respectively. When reported as ratio limits of agreement, these were 1.02 (*/divided 1.20) and 1.04 (*/divided 1.21) for between trials and between days, respectively. The coefficients of variation for all measures of power output were less than 9.3%. Hence, the treadmill system and protocol developed in this study provide a reliable measure of power output for rugby players.     

Reliability of power output measurements during repeated treadmill sprinting in rugby players

The non-motorized treadmill system initially reported by Lakomy in 1984 has been used extensively to assess sprinting performance. However, there has been limited research into the reliability of power output measurement using such systems. The aim of this study was to design a system and protocol capable of measuring treadmill sprinting performance in rugby players and to assess the reliability of this system for measuring power output. Twenty-seven rugby players, all of whom were familiar with treadmill sprinting, performed three maximal 6 s sprints with 2 min recovery between sprints, on two occasions 1 week apart. Both tests were performed on a non-motorized Woodway tramp treadmill, interfaced to a data acquisition system. There were no significant differences ( P > 0.05) between power output for repeated trials on the same day (between trials) or for repeated trials on different days (between days). Limits of agreement for maximum average power (the average of 100 readings per second) were 4 - 98 and 30 - 157 W for between trials and between days, respectively. When reported as ratio limits of agreement, these were 1.07 (*/ 1 1.12) and 1.03 (*/ 1 1.16), respectively. The limits of agreement for maximum instantaneous power (the highest of 100 readings per second) were 51 - 464 and 105 - 588 W for between trials and between days, respectively. When reported as ratio limits of agreement, these were 1.02 (*/ 1 1.20) and 1.04 (*/ 1 1.21) for between trials and between days, respectively. The coefficients of variation for all measures of power output were less than 9.3%. Hence, the treadmill system and protocol developed in this study provide a reliable measure of power output for rugby players.     

Anaerobic work and power output during cycle ergometer exercise: effects of bicarbonate loading

Eight trained male cyclists who competed regularly in track races, were studied under control, alkalotic (NaHCO3) and placebo (CaCO3) conditions in a laboratory setting to study the effect of orally induced metabolic alkalosis on 60 s anaerobic work and power output on a bicycle ergometer. Basal, pre- and post-exercise blood samples in the three conditions were analysed for pH, pCO2, pO2, bicarbonate, base excess and lactate. All blood gas measurements were within normal limits at basal levels. There were significant differences in the amount of work produced, and in the maximal power output produced by the cyclists in the experimental condition when compared to the control and placebo conditions (P less than 0.01). The post-exercise pH decreased in all three conditions (P less than 0.05) and post-exercise pCO2 increased significantly in the alkalosis trial (P less than 0.01). In the alkalotic condition, the pre-exercise base excess and HCO3- levels were both higher (P less than 0.05) than the basal levels, suggesting that the bicarbonate ingestion had a significant increase in the buffering ability of the blood. Post-exercise lactate levels were significantly higher (P less than 0.05) after the alkalotic trial when compared to the other two conditions, immediately post-exercise and for the next 3 min. Post-exercise lactate levels were higher than basal or pre-exercise levels (P less than 0.001). This was true immediately post-exercise and for the next 5 min. The results of this study suggest that NaHCO3 is an effective ergogenic aid when used for typically anaerobic exercise as used in this experiment. We feel that this ergogenic property is probably due to the accelerated efflux of H+ ions from the muscle tissue due to increased extracellular bicarbonate buffering.     

Sweat lactate response during cycling at 30 degrees C and 18 degrees C WBGT

Sweat lactate reflects eccrine gland metabolism. However, the metabolic tendencies of eccrine glands in a hot versus thermoneutral environment are not well understood. Sixteen male volunteers completed a maximal cycling trial and two 60-min cycling trials [30 degrees C = 30 +/- 1 degrees C and 18 degrees C = 18 +/- 1 degrees C wet bulb globe temperature (WBGT)]. The participants were requested to maintain a cadence of 60 rev min(-1) with the intensity individualized at approximately 90% of the ventilatory threshold. Sweat samples at 10, 20, 30, 40, 50 and 60 min were analysed for lactate concentration. Sweat rate at 30 degrees C (1380 +/- 325 ml x h(-1)) was significantly greater (P < 0.05) than at 18 degrees C (632 +/- 311 ml x h(-1)). Sweat lactate concentration was significantly greater (P < 0.05) at each time point during the 18 degrees C trial, with values between trials tending to converge across time. During the 30 degrees C trial, both heart rate (20, 30, 40, 50 and 60 min) and rectal temperature (30, 40, 50 and 60 min) were significantly higher than in the 18 degrees C trial. Higher sweat lactate concentrations coupled with lower sweat rates may indicate a greater relative contribution of oxygen-independent metabolism within eccrine glands during exercise at 18 degrees C. Decreases in sweat lactate concentration across time suggest either greater dilution due to greater sweat volume or increased reliance on aerobic metabolism within eccrine glands. The convergence of lactate concentrations between trials may indicate that time-dependent modifications in sweat gland metabolism occur at different rates contingent partially on environmental conditions.     

Neuromuscular control and coordination during cycling

The neuromuscular control aspect of cycling has been investigated through the effects of modifying posture and cadence. These studies show that changing posture has a more profound influence on neuromuscular coordination than does changing slope. Most of the changes with standing posture occur late in the downstroke: increased ankle and knee joint moment, reduced hip joint moment and greater activity in specific muscles. Due to the influence of lower extremity inertial properties, higher pedaling frequency induces more neuromuscular changes at the hip than at the knee or ankle joints. These neuromuscular adaptations to environmental and task constraints are discussed with regard to the contributions of the central nervous system and the solution provided by peripheral anatomical structure--mono- and biarticular muscles. The results indicate that training and related movement analysis should be specific to the motion, supporting the notion of task-specific training.     

Cycling power output produced during flat and mountain stages in the Giro d'Italia: a case study

Until recently, the physiological demands of cycling competitions were mostly reflected by the measurement of heart rate and the indirect estimation of exercise intensity. The purpose of this case study was to illustrate the varying power output of a professional cyclist during flat and mountain stages of a Grand Tour (Giro d'Italia). Nine stage recordings of a cyclist of the 2005 Giro d'Italia were monitored using a mobile power measurement device (SRM Trainingssystem, Julich, Germany), which recorded direct power output and heart rate. Stages were categorized into flat (n = 5) and mountain stages (n = 4). Data were processed electronically, and the overall mean power in flat and mountain stages and maximal mean power for various durations were calculated. Mean power output was 132 W +/- 26 (2.0 W x kg(-1) +/- 0.4) for the flat and 235 W +/- 10 (3.5 W x kg(-1) +/- 0.1) for the mountain stages. Mountain stages showed higher maximal mean power (367 W) for longer durations (1800 s) than flat stages (239 W). Flat stages are characterized by a large variability of power output with short bursts of high power and long periods with reduced intensity of exercise, whereas mountain stages mostly require submaximal, constant power output over longer periods.