Validity of SenseWear® Armband v5.2 and v2.2 for estimating energy expenditure

We compared SenseWear Armband versions (v) 2.2 and 5.2 for estimating energy expenditure in healthy adults. Thirty-four adults (26 women), 30.1 +/- 8.7years old, performed two trials that included light-, moderate- and vigorous-intensity activities: (1) structured routine: seven activities performed for 8-min each, with 4-min of rest between activities; (2) semi-structured routine: 12 activities performed for 5-min each, with no rest between activities. Energy expenditure was measured by indirect calorimetry and predicted using SenseWear v2.2 and v5.2. Compared to indirect calorimetry (297.8 +/- 54.2kcal), the total energy expenditure was overestimated (P<0.05) by both SenseWear v2.2 (355.6 +/- 64.3kcal) and v5.2 (342.6 +/- 63.8kcal) during the structured routine. During the semi-structured routine, the total energy expenditure for SenseWear v5.2 (275.2 +/- 63.0kcal) was not different than indirect calorimetry (262.8 +/- 52.9kcal), and both were lower (P<0.05) than v2.2 (312.2 +/- 74.5kcal). The average mean absolute per cent error was lower for the SenseWear v5.2 than for v2.2 (P<0.001). SenseWear v5.2 improved energy expenditure estimation for some activities (sweeping, loading/unloading boxes, walking), but produced larger errors for others (cycling, rowing). Although both algorithms overestimated energy expenditure as well as time spent in moderate-intensity physical activity (P<0.05), v5.2 offered better estimates than v2.2.