Objective: Assessing energy expenditure (EE) is important for the control of obesity. Daily step counts have become popular and constitute one practical technique for evaluating the physical activity (PA) in large population studies. However, information on the capacity of pedometers to track EE in free-living conditions remains scanty.
Subjects and methods: The 24-h EE of 71 healthy adults was measured by indirect calorimetry in a large respiratory chamber. Two accelerometers were attached to the waist, one for counting the total daily steps (ACC(STEP)) and another for measuring the anteroposterior whole body acceleration calculated as the root mean square of the acceleration signal at every second (ACC(RMS)).
Results: The ACC(STEP) was not associated with PA-related EE (PAEE) or 24-h EE. Body weight (BW) was the main determinant of both the values (explaining 30 and 75% of the variance, respectively). Approximately 8% (P < 0.001) of the variance in PAEE was attributed to the ACC(RMS) after BW was accounted for, whereas the ACC(STEP) did not explain any additional variance. A multiple stepwise regression analysis revealed that BW, height and ACC(RMS) were highly significant determinants of 24-h EE and accounted for as much as 83% of the total variance.
Conclusions: Recording the number of steps per day does not provide accurate information on EE, and at best is only a crude predictor of the general PA in terms of displacement. In contrast, accelerometry signals are considered to be a more meaningful factor in the assessment of EE rather than step counts under sedentary conditions. European Journal of Clinical Nutrition (2009) 63, 1425-1432; doi:10.1038/ejcn.2009.108; published online 26 August 2009
