Accuracy of a Clinical Applicable Method for Prediction of VO2max Using Seismocardiography

Abstract

Cardiorespiratory fitness measured as V̇O2max is considered an important variable in the risk prediction of cardiovascular disease and all-cause mortality. Non-exercise V̇O2max prediction models are applicable, but lack accuracy. Here a model for the prediction of V̇O2max using seismocardiography (SCG) was investigated. 97 healthy participants (18-65 yrs., 51 females) underwent measurement of SCG at rest in the supine position combined with demographic data to predict V̇O2max before performing a graded exercise test (GET) on a cycle ergometer for determination of V̇O2max using pulmonary gas exchange measurements for comparison. Accuracy assessment revealed no significant difference between SCG and GET V̇O2max (mean±95% CI; 38.3±1.6 and 39.3±1.6 ml·min-1·kg-1, respectively. P=0.075). Further, a Pearson correlation of r=0.73, a standard error of estimate (SEE) of 5.9 ml·min-1·kg-1, and a coefficient of variation (CV) of 8±1% were found. The SCG V̇O2max showed higher accuracy than the non-exercise model based on the FRIENDS study when this was applied to the present population (bias=-3.7±1.3 ml·min-1·kg-1, p<0.0001. r=0.70. SEE=7.4 ml·min^-1·kg^-1, and CV=12±2%). The SCG V̇O₂max prediction model is an accurate method for the determination of V̇O₂max in a healthy adult population. However, further investigation on the validity and reliability of the SCG V̇O₂max prediction model in different populations is needed for consideration of clinical applicability.