Comparison of computerized methods for detecting the ventilatory thresholds

The aim of this study was to compare computerized automatic methods to detect the ventilatory threshold (VT). Thirty apparently healthy and physically active volunteers [22.5 (6.5) years; 1.72 (0.08) m; 71.9 (8.5) kg] were submitted to a progressive and maximal cycle exercise. The gas exchange was monitored breath-by-breath with a fast gas analyser. The VT and respiratory compensation (RC) were automatically detected based on the respiratory exchange ratio, the ventilatory equivalent for O2 and the ventilatory equivalent for CO2, pulmonary ventilation, end-tidal PO2 and PCO2, and v-slope. In addition, VT and RC were also determined independently by visual inspection by two experienced investigators, and the results were compared with those of the automatic procedures. The automatic VT averaged 77% of the maximal V̇O2 and the RC 88%. The agreement between the experienced observers was very close [mean difference: 44.4 (16.1) ml, r=0.94, not significant]. Data were expressed as the mean value together with the standard deviation in each case. The automatic and visual inspection procedures did not present significant differences, resulting in 29.6 (29.6) ml with a reliability of r=0.86. All methods were significantly correlated for VT and RC (r=0.93 on average, P<0.01). ANOVA did not show differences between either the VT methods (P=0.131) or the RC methods (P=0.41). In conclusion, the present study has compared several simultaneous breath-by-breath ergospirometric methods that are used to describe the anaerobic threshold, showing high confidence when compared to visual inspection. No statistical differences were found between the VT and RC techniques for physically active subjects indicating that these methods may be equally effectively employed.