Submitted to the 2008 Annual Meeting of the American College of Sports Medicine for presentation

　

PREDICTING ACTIVITY ENERGY EXPENDITURE WITH ACTIVITY MONITORS IN ADULTS WITH PERIPHERAL NEUROPATHY

Daniel P. Heil, FACSM, Corey Black, Megan Duet, Li Li, FACSM

Montana State University, Bozeman, MT

Louisiana State University, Baton Rouge, LA

Over 20 million U.S. adults suffer from peripheral neuropathy which causes locomotion and balance problems due to pain and lack of peripheral sensations. These impairments led to extremely low physical activity (PA) and challenge the existing PA measurement methods. Purpose: This study compared measured energy expenditure (EE) with EE predicted using previously published prediction algorithms from accelerometry-based activity monitors at different body locations to determine accuracy of the predictio for this population. Methods: Five men (Mean±SD: 71±17 yrs age, 85.6±10.3 kg body mass, 28.6±3.2 kg/m² BMI) and 12 women (72±6 yrs, 71.1±12.0 kg, 27.5±4.3 kg/m² BMI), all diagnosed with peripheral neuropathy by a physician, volunteered to perform a series of structured activities: Sitting quietly (to estimate resting metabolic rate, RMR), newspaper reading while sitting, floor sweeping, table dusting, laundry folding, bagging groceries, as well as self-selected “slow” (52.6±15.1 m/min) and “brisk” (68.6±14.8 m/min) overground walking. EE was measured using a portable metabolic system worn using a modified backpack. Activity monitors were worn on the front beltline, the non-dominant wrist and corresponding ankle. Measured steady-state EE for each activity was transformed into activity EE (AEE, kcals/kg/min) by subtracting subjects’ estimated RMR. Predicted AEE based on the hip (AEE_H), wrist (AEE_W), and ankle (AEE_A) activity monitors were computed using previously published “1R” and “2R” algorithms (Heil RQES 77:64-80, 2006). Measured AEE was compared to 6 estimated AEE (3 monitor locations x 2 algorithms) using repeated measures ANOVA. Results: AEE_H for both 1R and 2R algorithms were significantly lower (P<0.05) than AEE for all activities except walking. 1R and 2R algorithms over predicted (P<0.05) AEE_W for only 3 activities (dusting, grocery bagging, while AEE_A was statistically similar to AEE for floor sweeping (1R and 2R) and walking (2R only). Conclusions: The hip and ankle monitor locations demonstrated the least promise for assessing AEE in this population, while the wrist monitor accurately predicted AEE for 4 of the 7 activities. Future studies should focus on revising the AEE prediction algorithms for this population with an emphasis on the wrist monitor location.