Resumen de Predicting VO2max From Treadmill Performance in American-Style Football Athletes

Stephen F. Crouse, Homer Tolson, Jason Lytle, Kalen A. Johnson, Steven E. Martin, John S. Green, Jonathan M. Oliver, Aaron F. Carbuhn, Bradley Lambert, Joe P. Bramhall

• Prediction equations are often used to estimate V[Combining Dot Above]O2max in the general population but are lacking for American-style football (ASF) athletes. We sought to develop a regression model to estimate V[Combining Dot Above]O2max from treadmill exercise time in ASF athletes and compare our football V[Combining Dot Above]O2max model with 2 published prediction equations (Foster et al., 1984, and Bruce, 1973). American-style football athletes (N = 472, age = 18 +/- 1 year, height = 186.1 +/- 8.2 cm, and body mass = 101.8 +/- 20.4 kg) underwent treadmill exercise to voluntary exhaustion (Bruce protocol). Maximal exercise time was recorded in minutes (Tmin), and V[Combining Dot Above]O2max was simultaneously measured (M-V[Combining Dot Above]O2max, mlO2[middle dot]kg-1[middle dot]min-1) by an automated gas-analysis system. Athletes were then randomly divided into validation and cross-validation groups (n = 236). Linear regression yielded estimates of V[Combining Dot Above]O2max from Tmin as follows: validation V[Combining Dot Above]O2max = 4.012 x Tmin - 4.628 (r2 = 0.678, p < 0.001, and SEE = 4.07); cross-validation V[Combining Dot Above]O2max = 4.025 x Tmin - 4.693 (r2 = 0.661, p < 0.001, and SEE = -4.16). These equations had a cross-validation coefficient of 0.813 and a double cross-validation coefficient of 0.823. Differences between the slopes of the 2 equations were not significant (t-test, p = 0.9603). Because validation and cross-validation groups were not statistically different on any variables measured (multivariate analysis of variance, p > 0.05), all athletes were combined to yield our final prediction equation: football V[Combining Dot Above]O2max = 4.017 x Tmin - 4.644 (r2 = 0.670, p < 0.001, and SEE = 4.11). Repeated-measures analysis of variance demonstrated significant differences (p < 0.001) in estimates of V[Combining Dot Above]O2max among Foster (44.1 +/- 6.1), Bruce (47.1 +/- 5.5), and our football (45.1 +/- 5.8) equations. Foster and Bruce V[Combining Dot Above]O2max estimates were also significantly different from M-V[Combining Dot Above]O2max (diff = -0.975 and 1.995, respectively, p < 0.001). V[Combining Dot Above]O2max of ASF athletes can be reasonably estimated by our football prediction equation using maximal treadmill time as the predictor.