A New Generalized Cycle Ergometry Equation for Predicting Maximal Oxygen Uptake: The Fitness Registry and the Importance of Exercise National Database (FRIEND)
journal contributionposted on 18.10.2018, 00:00 authored by Peter Kokkinos, Leonard A. Kaminsky, Ross Arena, Jiajia Zhang, Jonathan Myers
Background: To develop a clinically applicable equation derived from direct assessment of maximal oxygen uptake (VO2 max) to predict VO2 max assessed indirectly during cycle ergometry. Design: VO2 max was assessed by open-circuit spirometry during a graded maximal exercise test using cycle ergometry. Multivariable linear regression analysis was applied to identify the most relevant variables and construct the best prediction model for VO2 max using a random sample of 70% from each of the following age categories: <40, 40 to 50, 50 to 70 and >70 years; the remaining 30% was used for validation. Work rate (Watts*6.12/kg of body weight) and gender were considered in the final regression model as predictors of measured VO2 max and the resulting equation was compared to the traditional American College of Sports Medicine (ACSM) equation. Methods: Participants were part of the Fitness Registry and the Importance of Exercise National Database (FRIEND), a multi-institutional initiative with the primary objective of establishing normative VO2 max values across the adult lifespan. The final cohort consisted of 5,100 (3.378 males; mean age 35.9±12.1) and 1722 females; mean age=47.5±14.0 years). Results: The following equation was generated: VO2 max in ml O2●kg-1●min-1= 1.74* (Watts*6.12/kg of body weight) + 3.5. The derived FRIEND-ergometry equation predicted VO2 max with an overall relative bias of 0.51% ± 0.11) compared to a 15.46% ± 0.13 associated with the traditional ACSM equations (p<0.001). This predictive value was independent of gender, race, cardiac risk factors and cardiac; antihypertensive; metabolic; and/or lipid-lowering medication. Conclusion: The FRIEND-ergometry equation is considerably more precise than the traditional ACSM equation with an overall error >30 times lower than that associated with the ACSM equation.