What question did this study set out to answer?

To determine whether fat-free mass or body mass provides better predictions of maximum oxygen uptake.

March 14, 2026BMJ Open Sport & Exercise Medicine0 citationsOpen Access

Should we use fat-free mass or body mass and percentage body fat as separate predictors to predict maximum oxygen uptake?

ANALAN M. NEVILL MHMatthew P. Harber MWMatthew Wyon

Key Points

To determine whether fat-free mass or body mass provides better predictions of maximum oxygen uptake.
Utilized data from 3930 cardiopulmonary exercise tests from the FRIEND registry.
Employed allometric/power function models for prediction of VO2max using either fat-free mass or body mass.
Applied analysis of covariance to estimate parameters of the prediction equations.
Found that using body mass and percentage body fat separately increased explained variance to R2=0.733.
Confirmed AIC improved when incorporating separate body mass and percentage body fat terms.
Identified that fat-free mass alone fails to account for the negative effects of central adiposity.

Abstract

Background/objective To assess whether fat-free mass (FFM) or body mass (M) is the more appropriate body size variable to predict maximum oxygen uptake (VO 2max , L⋅min −1 ). Methods Data (3930 cardiopulmonary exercise tests) were provided from the FRIEND registry. Our prediction equations adopted the well-known allometric/power function model VO 2max (L⋅min −1 )= a · X b , using either FFM or M as the predictor variable ( X ). These models can be linearised with a log-transformation, and analysis of covariance (ANCOVA) can then be used to estimate the unknown parameters. Results Initially, when predicting Ln(VO 2max ) using only Ln(FFM) adjusted for age and sex, the explained variance was R 2 =0.718 (Akaike information criterion (AIC)=−1882.5), with the FFM exponent b=0.658. However, when predicting Ln(VO 2max ) using M AND bodyfat% separately, the explained variance increased to R 2 =0.733 (AIC=−2077.4), with the M exponent b=0.636. The difference in R 2 and AICs confirmed the benefit of predicting VO 2max using separate M and bodyfat% terms. The analysis identified an enlarged negative bodyfat% term that improved the prediction of VO 2max , explained latterly by central adiposity (waist circumference). These final, more inclusive M and FFM exponents were estimated to be b=0.67, suggesting that VO 2max should be normalised using VO 2max (mL·FFM −2/3 ·min −1 ) or preferably VO 2max (mL·M −2/3 ·min −1 ) rather than VO 2max (mL·FFM −1 ·min −1 ). We also found that linear prediction models systematically under-estimate the VO 2max of overweight and underweight individuals, but over-estimate the VO 2max of average-weight individuals. Conclusion Incorporating FFM into equations to predict VO 2max fails to explain the negative effect of central adiposity. However, by incorporating M and percentage body fat (BF%) separately into the allometric models, a greater/enlarged negative BF% term explains this apparent omission/absence.

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper