Reliability of performance and associated physiological responses during simulated sprint-distance triathlon

Abstract

Many different methods of simulating triathlon performance in controlled conditions have been developed without establishing the reliability of these assessments. The aim of this study was to determine the reliability of performance and physiological measures during simulated triathlon. Seven trained male triathletes completed initial familiarization, followed by three separate simulated sprint-distance triathlon trials (750 m swim, 500 kJ bike, 5 km run), using a 25 m pool, an electromagnetically braked cycle ergometer and motorized treadmill. Performance (time and mean cycling power) and physiological variables (oxygen uptake, ventilation, heart rate and blood lactate concentration) were measured throughout. Reliability between trials was assessed using one-way analysis of variance (ANOVA), coefficient of variation (CV), intraclass correlation coefficient (ICC) and ratio limits of agreement (LoA). No significant differences were found in performance or physiological variables measured across simulated triathlon trials. High levels of reliability (CV <10% and ICC >0.8) were observed for all performance measures (except transitions) and a majority of physiological variables. Measurement of blood lactate concentration displayed the poorest reliability throughout, with CV’s up to 17.3% and ICC’s as low as 0.4. Ratio LoA for total performance time were similar between trials 1-2 (1.008 */÷ 1.077) and trials 2-3 (1.004 */÷ 1.064). Based on these results simulated sprint-distance triathlon allows for reliable measurement of performance parameters and associated physiological responses in a controlled environment. This reliability data should be considered by simulated triathlon studies when determining statistical power and sample sizes, to allow for more rigorous detection of genuine changes between trials.