Validity Of The Actiheart Monitor For Predicting Physical Activity Energy Expenditure In Lower-limb Amputees.: 3549 Board #6 June 4, 9: 00 AM - 11: 00 AM.

Human physical activity energy expenditure (PAEE) is inherently difficult to measure in free-living conditions, particularly in populations where movements and movement patterns are atypical, such as prosthetic users following lower-limb loss. PURPOSE: To assess the validity of a multi-sensor device, which combines accelerometry and heart rate, to predict PAEE in functional amputees. METHODS: Ten unilateral (UNI) (33 ± 6 years), nine bilateral (BI) (29 ± 4 years) amputees, and eight non-injured healthy controls (CON) (31 ± 6 years) completed nine activities; resting in a prone position, standing, walking flat (1, 1.5, 2, 2.5, 3 mph) and on a gradient (3 and 5% gradient at 2mph) on a treadmill. An ActiheartTM device was worn on the chest. The relationships between predicted and criterion PAEE (measured using indirect calorimetry) were assessed. Bias and 95% limits of agreement (LoA) were calculated within each group. RESULTS: Predicted PAEE from the ActiheartTM device was significantly (p<0.001) associated with the criterion measure of PAEE in each group. Within the UNI and CON group there were negligible differences in the strength of relationships and associated error (r = .77, SEE = 1.32 kcal·min-1), with strongest relationship reported in the BI group (r = .84, SEE = 1.50 kcal·min-1). The BI group reported the widest LoA (-0.28 ± 3.98 kcal·min-1) followed by the UNI group (-0.85 ± 2.66 kcal·min-1) with the smallest LoA reported in the CON group (-0.19 ± 1.94 kcal·min-1). DISCUSSION: The smallest mean bias and LoA for the ActiheartTM, across a range of treadmill velocities, was in the CON group. This is perhaps unsurprising as the proprietary algorithms intrinsic to the device were developed on healthy non-injured humans. Wider LoA, demonstrating a greater magnitude of random error, is associated with increased functional impairment and injury severity. Therefore, it is likely that an increase in the number of amputated limbs is associated with a decrease in functional ability and inefficient gait mechanics. This was associated with a wide variability in the PAEE in the BI group when performing ambulatory tasks across a range of exercise intensities. It is possible that prediction of PAEE in these specific groups could be improved with population specific prediction algorithms or individual heart rate calibration.