AUTHOR=Gao Feng , Chan-Yu Yujing , Li Zhi , Liu Yongfu , Liu Fei , Liu Di , Yu Wenjun , Chen Weiwei , Wang Junhua , Le Shenglong 

TITLE=Comparison of two portable metabolic systems for measuring energy expenditure at rest and during exercise in untrained women

JOURNAL=Frontiers in Physiology

VOLUME=Volume 16 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2025.1583703

DOI=10.3389/fphys.2025.1583703

ISSN=1664-042X

ABSTRACT=PurposePortable metabolic systems are used as the “gold standard” for measuring energy expenditure (EE) in the development and validation of wearable devices. This study aimed to compare EE measurements obtained using the COSMED K5 (K5) and CORTEX METAMAX 3B (M3B) during the resting state and submaximal-intensity exercise in women without self-reported regular exercise training.MethodsTwenty women aged 21.4 ± 1.5 years completed two measurements, including resting in a seated position and cycling on a simple upright ergometer at 30 W, 40 W, 50 W, and 60 W. Average EE and other metabolic parameters were compared between K5 and M3B. Differences between K5 and M3B were assessed using the paired-samples t-test, and the effect size was calculated as Cohen’s d. Agreement between the two systems was evaluated by calculating Pearson correlation coefficients and visually examining Bland–Altman plots.ResultsThe number of participants who completed resting and exercise measurements was 18 and 19, respectively. For resting EE, the mean values measured using K5 were 33.4% higher than those measured using M3B (p < 0.001, Cohen’s d = 1.47). Similar differences were observed for cycling at 30 W (15.8%, p < 0.001, Cohen’s d = 1.50), 40 W (16.1%, p < 0.001, Cohen’s d = 1.68), 50 W (14.8%, p < 0.001, Cohen’s d = 1.28), and 60 W (14.6%, p < 0.001, Cohen’s d = 1.29). Pearson correlation coefficients between EE measured using K5 and M3B was 0.66 for 30 W cycling (p = 0.002) and 0.62 for 40 W cycling (p = 0.005).ConclusionK5 and M3B show significant differences in EE measurements during rest and exercise among untrained female individuals, indicating systematic bias in EE measurement between the two systems. Thus, careful consideration is essential when interpreting the results of wearable device studies that use different automated metabolic systems.