Should Time Trial Performance Be Predicted From Three Serial Time-to-Exhaustion Tests?

Dear Editor-in-Chief: Hinckson and Hopkins (1) attempt to convince the reader that time to exhaustion (TTE) is a reliable measure of endurance performance on the basis of predictions of time trial times from three serial runs to exhaustion by means of mathematical modeling. We maintain that the conclusions have been applied beyond the scope of the data that have been collected. Hinckson and Hopkins (1) correctly comment that TTE has traditionally been used as an indicator of endurance performance, but that more recently there has been a shift toward the use of time trial protocols. This change was probably triggered by a paper, published in Medicine & Science in Sport & Exercise® in 1996 (2). In this paper, it was clearly demonstrated that time trials (TT) have less day-to-day variation than TTE protocols, although TTE may be more appropriate in some conditions depending on the research question. In their introduction the authors (1) focus on endurance exercise and they specifically quote the work of Maughan and colleagues, who generally employed TTE tests lasting 40–180 min. However, the study performed by Hinckson and Hopkins (1) does not focus on this duration of exercise, since they purposely restricted TTE to less than 10 min. The study conclusions become even more questionable when one scrutinizes the data more closely. There is an increase in the coefficient of variation for time to exhaustion as the test becomes longer (9.2% for 1.8 min and 16% for 7.8 min). It would be interesting to know what the variation is for TTE tests, which are in the typical “endurance exercise” range. Therefore, irrespective of the outcome of their study, it is problematic to extrapolate their results to exercise of longer duration. We agree with Hinckson and Hopkins (1) that a reliable test (i.e., time trial) does not necessarily mean that differences due to some intervention can be detetcted with this test. On the other hand a test with a large day-to-day variation does not exclude that differences can be detected as a result of an intervention. However, this issue of sensitivity or signal-to-noise ratio was not directly addressed in their study. In order to study the sensitivity of a test the intraindividual variation (noise) as well as the performance effect size (signal) should have been measured for both a TTE test and a TT. In many cases, time trials will be the preferred choice of test because they have greater logical validity. Pacing strategy is an inherent component of real performance rather than something that should be omitted from a performance test. Perhaps most importantly, it seems impractical and illogical to perform two or three separate tests to predict performance with mathematical modeling if it can also be measured directly. The key to a good performance test is the signal-to-noise ratio. If the signal does not exceed the noise it is impossible to pick up differences as a result of an intervention. It is this sensitivity issue that should receive more attention in future studies. Asker E. Jeukendrup Kevin Currell School of Sport and Exercise Sciences University of Birmingham Birmingham, United Kingdom