Königstein K, Klenk C, Rossmeissl A, Baumann S, Infanger D, Hafner
B, Hinrichs T, Hanssen H, Schmidt-Trucksäss A

Obesity (Silver Spring). 2018 Feb;26(2):291-298. doi: 10.1002/oby.22078. Epub
2017 Dec 12.

OBJECTIVE: Cardiopulmonary exercise testing is clinically used to estimate
cardiorespiratory fitness (CRF). The relation to total body mass (TBM) leads to
an underestimation of CRF in people with obesity and to inappropriate prognostic
and therapeutic decisions. This study aimed to determine body composition-derived
bias in the estimation of CRF in people with obesity.
METHODS: Two hundred eleven participants (58.8% women; mean BMI 35.7 kg/m2
[± 6.94; 20.7-58.6]) were clinically examined, and body composition (InBody720;
InBody Co., Ltd., Seoul, South Korea) and spiroergometrical peak oxygen
consumption (VO2 peak) were assessed. The impacts of TBM, lean body mass (LBM),
and skeletal muscle mass (SMM) on CRF estimates were analyzed by the application
of respective weight models. Linear regression and plotting of residuals against
BMI were performed on the whole study population and two subgroups (BMI < 30
kg/m2 and BMI ≥ 30 kg/m2 ).
RESULTS: For every weight model, Δmean VO2 peak (expected - measured) was
positive. LBM and SMM had a considerable impact on VO2 peak demand (P = 0.001;
ΔR2  = 2.3%; adjusted R2  = 56% and P = 0.001; ΔR2 = 2.7%; adjusted R2  = 56%),
whereas TBM did not. Confounding of body composition on VO2 peak did not differ
in LBM and SMM.
CONCLUSIONS: TBM-adjusted overestimation of relative VO2 demand is much higher in
people with obesity than in those without. LBM or SMM adjustment may be superior
alternatives, although small residual body composition-derived bias remains.