Reproducibility of 20 and 40km laboratory based simulated trials

Establishing the reproducibility of a measure is important to facilitate appropriate estimations of sample size for research and to make pragmatic assumptions concerning the usefulness of an intervention. The aim of this study was to quantify the reproducibility of laboratory based 20 and 40 km simulated time trials (TT), and the smallest worthwhile change in performance, in well-trained cyclists. With institutional ethics approval 11 welltrained male cyclists (age, 37.9¡6.2 years, mass, 78.8¡9.2 kg, VO2max, 4.78¡0.30 l min21) completed three consecutive TT at each distance, the order of which was randomised and counterbalanced, on an electromagnetically braked cycle ergometer (Velotron, RacerMate, Seattle). Participants were instructed to complete each TT as fast as possible and feedback was given on distance elapsed every 10% of the trial. Data were log transformed before analysis. Change in the mean was assessed for learning effects between TT 1–2 and 2–3. The mean typical error with 90% confidence intervals, were used to quantify reproducibility. Smallest worthwhile change in performance was quantified as 0.2 of the between-participant standard deviation. There were no trends for improved performance on repeat TTs (20 km 1–2=0.45%, 2– 3=0.21%. 40 km 1–2=23.2%, 2–3=2.2%). For 20 km, mean power was 269¡33 W, typical error was 2.9% (2.3 to 4.1%), and smallest worthwhile change was 2.4%. For 40 km, mean power was 235¡33 W, typical error was 3.7% (2.9 to 5.3%) and smallest worthwhile change was 2.7%. These data are comparable with previous reproducibility studies on 20 (Sporer and McKenzie, 2007) and 40 km TT’s. The results of the study can be used to determine sample sizes for future research, and for interpreting changes in performance in well-trained cyclists.