Energy cost of walking in older adults: accuracy of the ActiGraph accelerometer predictive equations

BACKGROUND/OBJECTIVES: Various accelerometer equations are used to predict energy expenditure (EE). On the other hand, the development of these equations and their validation studies have been conducted primarily without including older adults. This study assessed the accuracy of 8 ActiGraph accelerometer equations to predict the energy cost of walking in older adults. SUBJECTS/METHODS: Thirty-one participants with a mean age of 74.3 ± 3.3 yrs were enrolled in this study (20 men and 11 women). The participants completed 8 walking activities, including 5 treadmill and 3 self-paced walking activities. The EE was measured using a portable indirect calorimeter, with each participant simultaneously wearing the ActiGraph accelerometer. Eight ActiGraph equations were assessed for accuracy by comparing the predicted EE with indirect calorimetry results. RESULTS: All equations resulted in an overall underestimation of the EE across the activities (bias -1 to -1.8 kcal·min^-1 and -0.7 to -1.8 metabolic equivalents [METs]), as well as during treadmill-based (bias -1.5 to -2.9 kcal·min^-1 and -0.9 to -2.1 METs) and self-paced (bias -1.2 to -1.7 kcal·min-1 and -0.2 to -1.3 METs) walking. In addition, there were higher rates of activity intensity misclassifications, particularly among vigorous physical activities. CONCLUSIONS: The ActiGraph equations underestimated the EE for walking activities in older adults. In addition, these equations inaccurately classified the activities based on their intensities. The present study suggests a need to develop ActiGraph equations specific to older adults.