The reproducibility of estimates of critical power and anaerobic work capacity in upper-body exercise.

The purposes of this study were to apply the linear power versus inverse of time relationship to high-intensity upper-body exercise and to assess the repeatability of the parameters critical power (CP) and anaerobic working capacity (AWC), using limits of agreement (Bland and Altman 1986). Sixteen active male subjects (aged 20-34 years), performed two sets of five constant-power exercises on an adapted cycle ergometer. There were no significant differences between mean estimates of CP [96 (16) W and 95 (17) W] and AWC [7457 (2011) J and 7608 (1684) J] from the first and second sets of bouts. Despite the lack of systematic bias, there was evidence of large random error. Ratio limits of agreement for time to exhaustion during constant-power exercises suggested that a repeat measurement might be expected in 95% of cases to be between 0.64 and 1.59 times the original measurement. The 95% limits of agreement for CP were -15 W to +17 W. The ratio limits of agreement for AWC suggest that in 95% of cases a repeat measurement might be between 0.57 and 1.67 times the original estimate. The results of this study suggest a poor repeatability of constant-power upper-body exercises to exhaustion, which may contribute to a poor repeatability of CP and AWC determined from the linear power versus inverse of time model.