Purpose: This study examined the reliability and validity of the TriTrac-R3D triaxial accelerometer to estimate energy expenditure during various modes of exercise. Methods: Twenty subjects (age = 21.5 +/- 3.4 yr; body mass index = 23.3 +/- 3.6 kg.m(-2)) performed five exercises (treadmill walking, treadmill running, stepping, stationary cycling, and slideboard), with each lasting 20-30 min and workload increased at 10-min intervals. To test the inter-TriTrac reliability, two TriTrac-R3D accelerometers were worn during each exercise period, and to examine validity, a simultaneous measurement of energy expenditure was made using indirect calorimetry (SensorMedics 2900 Metabolic Cart). Results: Results showed a significant correlation between the two TriTrac-R3D accelerometers during all exercises. The difference in estimated energy expenditure between the two accelerometers during the walking, stepping, and slideboard exercises was less than 1 kcal.min(-1) but statistically significant (P < 0.05). There was also a significant correlation between energy expenditure estimated by each of the TriTrac-R3D accelerometers and indirect calorimetry during walking, running, stepping, and slideboard exercise (P < 0.05). The interaction of Method X Workload was significant (P < 0.05) for each exercise, indicating that the TriTrac-R3D underestimates energy expenditure and that the magnitude of this underestimation increases as workload increases. Conclusions: Therefore, energy expenditure estimated via triaxial accelerometry does not increase with increasing workloads. These results suggest that there are limitations to using triaxial accelerometry to quantify energy expenditure.
