Key points are not available for this paper at this time.
Twenty-four coded graph sets of gas exchange variables and blood lactate concentration (LA) plotted against time at 15-s intervals were analyzed by nine evaluators who determined the gas exchange (ATGE) and LA (ATLA) anaerobic thresholds. In addition, ATGE and ATLA were determined by a linear regression computer program. Agreement between ATGE and ATLA was poor; the median intraclass correlation coefficient (ri) was 0.53. Among evaluators, ATLA agreement (median ri = 0.81) was better than ATGE agreement (median ri = 0.70). In general, the ability of any evaluator to choose similar values from duplicate plots for either ATGE (median ri = 0.97) or ATLA (median ri = 0.995) was good. There was better agreement between the mean ATLA of the evaluators and the computer ATLA (ComLA) (ri = 0.88) than between the mean ATGE of the evaluators and the computer ATGE (ComVE), (ri = 0.58). Agreement between ComVE and ComLA was poor (ri = 0.29). These results suggest that ATGE does not accurately predict ATLA and that different evaluators choose different thresholds from the same data. Further assessment of the validity and precision of ATGE based on breath-by-breath and minute-by-minute data is needed.
Gladden et al. (Sat,) studied this question.