Low Reproducibility Of Lactate Markers During Incremental Cycle Exercise

Abstract

PURPOSE: Previous reports on reproducibility of lactate markers during incremental exercise have usually considered only two trials, and these predominantly in running. This study was undertaken to determine the reproducibility of power output at four different markers in a group of subjects over at least six repeated incremental cycling trials held under tightly controlled conditions. METHODS: Ten subjects undertook six, and an eleventh undertook seven, repeated incremental exercises (50W start, +50W every 3-min to volitional exhaustion) on a cycle ergometer. At each trial blood lactate concentration was determined at rest and within the final 30s of each 3-min workload. The Rest+1 and 4.0 mMol.L-1 markers were determined by interpolation, and the blood lactate vs. power output relationship was modelled using a quadratic function, from which the nadir (lactate minimum) and D-max markers were determined. Analyses of variance were utilised to partition inter- and intra-subject variation in power output at these markers. Intraclass correlation coefficients were evaluated to assess reproducibility. Standard calculations were performed for assessment of statistical power. RESULTS: Power output at all markers differed significantly between subjects, but not between trials. Intraclass correlation coefficients were 0.799 for the Rest+1 point, 0.903 for D-max, 0.677 for the nadir, and 0.807 for the 4.0 mMol.L-1 point. Standard errors of measurement were 11.9, 2.5, 9.2 and 9.1 W respectively. Corresponding pooled within-subject standard deviation estimates were 26.1, 7.9, 16.2 and 20.8 W, yielding 95% confidence intervals of width 105.2, 31.8, 65.3 and 83.8 W respectively. At a significance level of 0.05 the calculated powers of detecting a 30 W increment in these markers are 0.31, 0.98, 0.58 and 0.42 respectively. CONCLUSIONS: These results suggest that during incremental cycling exercise the reproducibility of power output at lactate markers other than D-max are not as good as those previously reported for running. In addition, apart from D-max, their error margins are of such magnitude that in trained subjects only gross changes in training status can be identified with significant power.