Reliability of assessments of ventilatory thresholds

The reproducibility of the assessment of ventilatory thresholds was investigated in two test‐retest experiments, one performed on a cycle ergometer with 21 moderately active male subjects and the other on a treadmill with 20 well‐trained male subjects. The first (VT‐1) and the second (VT‐2) nonlinear increases in ventilation (V E) relative to O2 consumption were determined (a) by three independent evaluators coding separately (OIE), (b) as the mean of three independent evaluators (TIE) and (c) by two dependent evaluators (TDE). One of the evaluators repeated the assessment four months later (SELF). The VT‐1 and VT‐2 were also assessed from the graph of V E /VO2 relative to VO2 . Under the SELF condition VT‐1 and VT‐2 in ml O2 per kg min‐1 proved to be reliable measurements with intraclass correlations of 0.84 and 0.91 respectively. Independent evaluators were individually reliable assessors of VT‐1 and VT‐2, with the exception of VT‐1 in the treadmill group in terms of ml O2 per kg min‐1 with coefficient ranging from 0.71 to 0.94. Similar results were obtained under the TDE condition. The VT‐1 assessments in the treadmill group were generally less reproducible than VT‐2 and less reproducible than either VT‐1 or VT‐2 in the bicycle group. The VT‐1 and VT‐2 expressed as percentages of VO2 max were not reproducible measurements under the conditions of this study. It is concluded that the condition described as TDE offers the most advantageous procedure for the determination of VT‐1 and VT‐2.