Classification of cardiorespiratory fitness without exercise testing

Purpose: We examined the ability of a nonexercise based (V) over dot O-2max , prediction model to classify cardiorespiratory fitness (CRF) in a populat ion of men and women aged 19-79 yr of age (N = 799). Methods: A (V) over do t O-2max (mL.kg(-1).min(-1)) prediction model was developed in the study gr oup using multiple linear regression from the independent variables age, ag e(2), gender, physical activity status, height, and body mass. The classifi cation accuracy of this model was examined by cross-tabulating age and gend er specific quintiles of measured and predicted CRF. Results: Overall class ification accuracy of the model was modest (36%); however, 83% of all subje cts were either classified correctly or within one quintile of measured CRF . Extreme misclassification (e.g., misclassifying a low fit individual as h igh fit) was only rarely observed (0.13%). Conclusions: The present results support the concept that CRF prediction models can be used to reasonably c haracterize the fitness level of a cohort using data that can be obtained f rom a questionnaire. Accordingly, predicted CRF values may be useful as an exposure variable in large epidemiologic studies in which exercise testing is not feasible.