Field evaluation of energy expenditure in women using Tritrac accelerometers

To investigate the use of Tritrac accelerometers to measure energy expenditure (EE) of various activities for women in the field setting, as compared with portable indirect calorimetry. Twenty women (age 20-29) performed a choreographed routine of six activities (walking, jogging, stair climbing, walking on an incline, stationary cycling, and arm ergometry) while wearing a Tritrac-R3D accelerometer (Hemokinetics Inc., Madison WI) and the Cosmed K4b(2) portable metabolic cart (Cosmed, Rome, Italy). Comparing the mean error scores (K4b(2) - Tritrac), the Tritrac overestimated the EE (kcal x min(-1)) of walking (-1.45) and jogging (-1.75), whereas underestimating the EE of stair climbing (2.76), stationary cycling (2.75), and arm ergometry (1.20). Walking on an incline showed the lowest mean error score (-0.11). Intraclass correlations were moderate for walking (r = 0.568, < 0.05), jogging (r = 0.666, < 0.05), and stairs (r = 0.503, < 0.05) but for the other activities ranged from r = 0.290 ( > 0.05) to r = 0.480 ( < 0.05). The raw data from the Tritrac was applied to a previously developed nonlinear model to adjust the Tritrac scores to the standard of whole-room indirect calorimetry. This resulted in statistically significant improvements in the agreement between the adjusted Tritrac value and the K4b for walking, jogging, and walking on an incline ( < 0.05). When compared with portable indirect calorimetry, the Tritrac overestimates the EE of walking and jogging, whereas underestimating that of stair climbing, stationary cycling, and arm ergometry. This limits the use of such a technique to measure EE in the field. The main issues appear to be the type and intensity of the activity and the need for movement in order for the Tritrac to register EE. Activity specific linear regression equations are proposed as a tool to improve the measurement of EE using the Tritrac in the field.