Validity and reliability of physiological testing in cycling

This project evaluated physiological testing procedures in male road cyclists and focused on the following: (a) predictive validity of ventilatory (VT) and lactate thresholds (LT) for 40-km cycle time trial (40K) performance, (b) influence of testing protocol on VTs and the determination of their predictive validity for 40K mean power output (40Kavgwatts), and (c) the evaluation of validity and reliability of VT determination methods. (a) Fifteen male cyclists (VO2max 67.6 +/- 5.3 ml·kg -1·min-1) performed LT test, VT test, 5-km cycle time trial (5K), and 40K over an 8-day period. Key variable of interest were Watts at thresholds. For VT determination we used breakpoint of VE/VO2; breakpoint of VE/VCO2; V-slope; RER = 1; and RER = 0.95. For LT we used Stegmann's IAT; stage proceeding the second 0.5 mmol/L increase; 4 mmol/L; 1 mmol/L increase in 3 min; stage proceeding the first 1 mmol/L increase as criterion methods. Analyses also included peak-power during the incremental threshold tests and 5K performance. (b) Fifteen cyclists performed two incremental exercise tests (T50x3: 100 W + 50 W·3min-1, T25x1: 20 W + 25 W·min-1) and a 40K over an 8-day period. Key variable of interest were Watts at VT. VTs were assessed as described above. (c) VT test and 40K were performed on two occasions (February/September) by two subsets of cyclists (n = 15 each; VO2max 67.6 +/- 5.3/71.5 +/- 3.0 ml·kg-1·min-1 ). Key variables of interest were power output at VTs, MaxVT watts, 40Kavgwatts. VTs were assessed as described above. As a follow-up, test-retest reliability of VTs and at max. were examined in 20 cyclists (VO2max 63.4 +/- 8.0 ml·kg-1 ·min-1) on two occasions, separated by 48 h. Results for (a) showed that VT based variables were generally superior to LT variables relative to predicting 40Kavgwatts. On the basis of regression analyses and the absolute power-output, VE/VO2 appears to be the superior method. Findings for (b) showed that T25x1 based VT variables were superior to T50x3 variables regarding 40Kavgwatts prediction. Only VE/VO2 appears to be a protocol-independent 40Kavgwatts predictor. Results for (c) showed that VE/VO2 has a high test-retest reliability and the highest consistency to predict 40Kavgwatts over a competitive season. This project was the first to assess different aspects of reliability and validity of VT based performance predictors over a competitive season and one of the few to evaluate a variety of physiological testing procedures regarding their ability to predict endurance performance in terms of power output. The results showed that VE/VO2 is the superior method in cycling testing.