Florêncio Diniz-Sousa1, Lucas Veras2, José Carlos Ribeiro2, Giorjines Boppre2, Vítor Devezas3, Hugo Santos-Sousa3, John Preto3, Leandro Machado4, João Paulo Vilas-Boas4, José Oliveira2, Hélder Fonseca5. 1. Research Center in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sport, University of Porto, Porto, Portugal. Electronic address: joseflorenciosousa@gmail.com. 2. Research Center in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sport, University of Porto, Porto, Portugal. 3. General Surgery Department, São João Medical Center, Porto, Portugal. 4. Center of Research, Education, Innovation and Intervention in Sport (CIFI2D), Faculty of Sport, University of Porto, Porto, Portugal; Biomechanics Laboratory (LABIOMEP-UP), University of Porto, Porto, Portugal. 5. Research Center in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sport, University of Porto, Porto, Portugal; General Surgery Department, São João Medical Center, Porto, Portugal.
Abstract
BACKGROUND: Almost all accelerometer calibration studies were developed for non-obese people, which hampers an accurate prediction of energy expenditure (EE) and induces a misclassification of sedentary activity (SA) and physical activity intensities (PAI) in class II-III obese people. RESEARCH QUESTION: The purpose of this study was to develop regression equations to predict EE and cut-points to classify SA and PAI in severe obese people based on several metrics obtained from hip and back accelerometer placement data. METHODS: 43 class II-III obese participants performed a protocol that included sitting and standing positions and walking at several speeds. During the protocol participants wore an accelerometer at hip and back, and respiratory gas exchange was measured by indirect calorimetry. Accelerometer metrics analyzed were: activity counts, mean amplitude deviation and euclidean norm minus one. EE was predicted through linear mixed models while cut-points to classify SA and PAI were obtained applying receiver operating characteristic curves. Leave-one-out cross-validation data was used to calculate Bland-Altman plots, prediction accuracy, Kappa statistic and percent agreement. RESULTS: All prediction models presented a quadratic equation that had as predictors body mass and one of the accelerometer metrics. Predicted EE indicated a good agreement and a root mean square error below 1.02 kcal min-1. Global classification agreement from developed cut-points was categorized as almost perfect with a percent agreement above 84 %. Prediction accuracy and classification agreement were similar among accelerometer metrics in each position and between them in hip and back placement. SIGNIFICANCE: Hip and back accelerometer data collected in severe obese people allow to accurately estimate EE and to correctly classify SA and PAI. These results enable future studies to adopt appropriate regression equations and cut-points developed for class II-III obese people rather than those established for non-obese people.
BACKGROUND: Almost all accelerometer calibration studies were developed for non-obesepeople, which hampers an accurate prediction of energy expenditure (EE) and induces a misclassification of sedentary activity (SA) and physical activity intensities (PAI) in class II-III obesepeople. RESEARCH QUESTION: The purpose of this study was to develop regression equations to predict EE and cut-points to classify SA and PAI in severe obesepeople based on several metrics obtained from hip and back accelerometer placement data. METHODS: 43 class II-III obeseparticipants performed a protocol that included sitting and standing positions and walking at several speeds. During the protocol participants wore an accelerometer at hip and back, and respiratory gas exchange was measured by indirect calorimetry. Accelerometer metrics analyzed were: activity counts, mean amplitude deviation and euclidean norm minus one. EE was predicted through linear mixed models while cut-points to classify SA and PAI were obtained applying receiver operating characteristic curves. Leave-one-out cross-validation data was used to calculate Bland-Altman plots, prediction accuracy, Kappa statistic and percent agreement. RESULTS: All prediction models presented a quadratic equation that had as predictors body mass and one of the accelerometer metrics. Predicted EE indicated a good agreement and a root mean square error below 1.02 kcal min-1. Global classification agreement from developed cut-points was categorized as almost perfect with a percent agreement above 84 %. Prediction accuracy and classification agreement were similar among accelerometer metrics in each position and between them in hip and back placement. SIGNIFICANCE: Hip and back accelerometer data collected in severe obesepeople allow to accurately estimate EE and to correctly classify SA and PAI. These results enable future studies to adopt appropriate regression equations and cut-points developed for class II-III obesepeople rather than those established for non-obesepeople.
Authors: Ville Stenbäck; Juhani Leppäluoto; Rosanna Juustila; Laura Niiranen; Dominique Gagnon; Mikko Tulppo; Karl-Heinz Herzig Journal: J Clin Med Date: 2022-06-07 Impact factor: 4.964
Authors: L Veras; F Diniz-Sousa; G Boppre; V Devezas; H Santos-Sousa; J Preto; J P Vilas-Boas; L Machado; J Oliveira; H Fonseca Journal: Osteoporos Int Date: 2020-01-21 Impact factor: 4.507