Comparison of Research- and Consumer-grade Energy Expenditure Estimation Methods during 10 Days of Military Training

Abstract

PURPOSE: Wearable physical activity monitoring devices have improved the ability to estimate free-living total energy expenditure (TEE) but their application during arduous military training alongside more well-established techniques has not been widely documented. This study aimed to assess the validity of two wrist-worn activity monitors to estimate TEE by evaluating performance against doubly labelled water (DLW) during British Army Officer Cadet (OC) training. METHODS: Twenty (10 male and 10 female) OCs (mean ± SD: age 22 ± 1 years, height 1.73 ± 0.08 m, body mass 77.0 ± 9.3 kg) wore one research-grade accelerometer (GENEActiv, Cambridge, UK; “RES”) on the dominant wrist and one commercially available (FITBIT SURGE, USA; “COM”) monitor on the non-dominant wrist for 10 days of training. Immediately prior to this 10-day period, participants consumed a bolus of DLW and provided daily urine samples, which were analysed by mass spectrometry to determine TEE. Bivariate correlations and limits of agreement were calculated to compare the 10-day mean TEE from DLW with both activity monitors to evaluate device performance. RESULTS: TEE (mean ± SD) from DLW, RES and COM were 17.2 ± 2.7 MJ.day-1 (4112 ± 652 kcal.day-1), 17.3 ± 2.8 MJ.day-1 (4129 ± 677 kcal.day-1) and 15.1 ± 3.7 MJ.day-1 (3607 ± 888 kcal.day-1), respectively. TEE from DLW was linearly correlated with both RES (r=0.786, p<0.001) and COM (r=0.888, p<0.001). Despite a stronger association with DLW however, COM tended to underestimate TEE (mean bias [95% CI]) by -2.1 [-5.6-1.4] MJ.day-1 (-505 [-1348-339] kcal.day-1; p<0.05). In contrast, mean TEE from RES was similar to DLW (-0.05 [-3.6-3.5] MJ.day-1; -11 [-867-845] kcal.day-1; p>0.05). CONCLUSION: Wearable physical activity monitors provides a cheaper and more practical method for estimating free-living TEE than DLW, and could be useful for military populations. However, this study suggests a consumer monitor may underperform, by underestimating TEE, during physically demanding training in comparison to a research- grade device.