Lower reference limit for maximal oxygen uptake in men and women.

Abstract

Maximal oxygen uptake (VO2max) is an important measure of exercise tolerance and low values may have clinical significance. Our purpose was to develop the necessary statistics--prediction equations and standard errors of estimate (SEE)--so that the lower reference limit for VO2max can be predicted for men and women. The subjects were healthy, non-smoking, sedentary men (n = 115) and women (n = 115) aged 20-70 years who performed 15 W min-1 cycle ergometer exercise tests. Three equations were developed for each gender using multiple linear regression with the non-exercise predictor variables of age and height, age and mass, and age and fat-free mass (FFM). The assumptions of regression analysis were examined and the predicted residual sum of squares (PRESS) method was used to cross-validate each equation. Healthy and diseased individual subject data from the literature were used to externally validate our lower reference limit statistics. The equations developed meet the assumptions of regression analysis and have an accuracy similar to the non-exercise prediction equations in the literature with R2 values of approximately 0.581. The PRESS method revealed that the equations are generalizable, i.e. may be used in future studies without a significant loss of accuracy. The lower reference limit predictions for the healthy and diseased individual subject data from the literature produced few miscategorizations unless the subjects were obese and mass was used as a predictor variable. In conclusion, the equations and SEE generated in this study can be used to predict an accurate and valid VO2max lower reference limit for a given subject.