Abstract
The reproducibility, or reliability, of exercise tests is essential in detecting changes in performance. The reliability of cycling performance is difficult to measure in the field for several reasons, especially wind. Therefore, performance measures of cycling are often conducted in a laboratory setting. However, standard cycle ergometers do not allow cyclists and triathletes to adjust their riding position to the one normally used on their own bicycle, which may compromise performance and reliability. In addition, ergometers that do allow for riding position adjustment are very expensive. PURPOSE: To evaluate the reproducibility of a simulated 20-km time trial (TT) in competitive cyclists and triathletes using the athlete's own bicycle mounted to an electronic load generator (ELG). METHODS: 1 1 competitive road cyclists and triathletes (8 men, 3 women) who were currently in their off-season, but regularly cycle training, were recruited as subjects. Each subject completed three identical TT using their own bicycle, each separated by 2 or 3 days. Subjects were asked to refrain from alcohol consumption and vigorous exercise for 48 h prior to testing, and report in a fasted state (about 4 hours after the last meal and caffeine ingestion). After a standard warm-up, the subject's rear bicycle tire was inflated to a standard pressure (7.58 BAR) and the ELG was calibrated per manufacturer's instructions. Subjects then cycled 20-km as fast as possible on their own bicycle, mounted on the ELG. The ELG was connected to a computer that displayed a simulated outdoor course on the monitor, and recorded power output (PO), time, and distance cycled. During the TT, subjects were allowed to drink water ad libitum, and knew how far they had cycled. Heart rate (HR) was measured with a HR monitor, however, subjects were not allowed to know their HR, elapsed time, or PO. Plasma lactate (PL) concentrations were measured via finger stick at rest, 1 0-km and 20-km during the TT, and 3 min post-TT. Velocity (V) was calculated from the PO and time to complete the TT measured by the ELG. Differences between trials were analyzed using a repeated measures ANOVA (p<0.05) and intraclass correlations (ICC) were calculated to analyze the agreement between repeated trials. RESULTS: Of the 1 1 subjects, only 8 (5 men, 3 women, mean ± SD, age = 34 ± 9 yr, body weight = 68.2 ± 8.3 kg, body fat = 1 5.3 ± 9.4 %, PO = 1 99 ± 69.4 watts) had performance data from all 3 trials. For these 8 subjects, there were no significant differences (p>0.05) for PO, time to complete TT, or V across the trials, with ICC values of R = 0.99 for each variable. HR responses were not different across the trials at 10-km or 20-km (p>0.05) with ICC values of R = 0.85 and 0.94, respectively. Of these 8 subjects, only 6 had PL data from all 3 trials. For these 6 subjects, there was no significant difference (p>0.05) for PL at 1 0-km, 20-km, or post-TT, with ICC values of R = 0.89, 0.97, and 0.96, respectively. CONCLUSION: If strict calibration methods are followed, both performance and physiologic data are highly reproducible when testing competitive cyclists using their own bicycle and an ELG controlled by a computer. PRACTICAL APPLICATIONS: Using a reliable ELG that allows cyclists to use their own bicycle provides researchers, coaches, and athletes with a less expensive option for performance testing of athletes that compete on a variety of bicycles with varying riding positions. [PUBLICATION ABSTRACT]