Estimation of Energy Expenditure Using Accelerometry in Individuals with Locomotor Dysfunction

Abstract

Accelerometry is widely used to determine energy expenditure (EE) and daily physical activity patterns in a variety of able-bodied populations. The reliability and validity of these devices are well documented in the able-bodied population; however the assumption is that the side of the body and location to which the accelerometer is placed is not critical due to the symmetry of normal gait. PURPOSE: The purpose of this study was twofold: 1) to analyze accelerometer activity counts (AC) from a 6 minute walk (6MW) test in individuals with locomotor dysfunction (LD) and determine if differences exist between sides or locations of accelerometer placement and 2) to determine if current EE equations are appropriate for individuals with LD. METHODS: Prior to the 6MW, Actical accelerometers (Mini Mitter Co, Bend, OR) were place on bilateral ankles, wrists, and hips of 6 women and 10 men (Mean±SD: 61±16 yrs, 92.6±19.2 kg) with LD. Mean activity energy expenditure (AEE, kcals/ kg/min) was determined from the last two mins of the 6MW using continuous VO2 measurements from an ambulatory metabolic analyzer. Mean predicted AEE (PEE), as well as mean AC, were derived from the last minute of the 6RM for each of the 6 Actical wearing locations. Mean AEE and PEE at each site were compared using a 1-factor RMANOVA. Correlations of mean AEE with PEE and AC at each site for all wearing locations were computed. All analyses performed at 0.05 alpha level. RESULTS: Mean AEE (Mean±SE: 0.058±0.005 kcals/kg/min) was significantly lower than PEE at either left or right wrist (0.021±0.003 and 0.029±0.004; P<0.05) or hip (0.035±0.003 and 0.032±0.003; P<0.5), but statistically similar to PEE for both left (0.064±0.004) and right ankles (0.063±0.003). Correlations of AEE with PEE (r = −0.10 to 0.27) and AC (r = −0.13 to 0.35) across all 6 locations were nonsignificant (p<0.05). AC at the bilateral wrist and ankle sites were poorly correlated. CONCLUSIONS: In contrast to what is reported for able-bodied adults during locomotion, we found that AEE was consistently under predicted and correlated poorly with AC regardless of measurement location. The data from this study confirm the clinical observation that left and right side of the body are moving differently in people with LD. Subject-specific algorithms for converting AC to EE may be required for this population.