Predicting Activity Energy Expenditure by accelerometry in children and adolescents

Abstract

Purpose of this review was to describe the validity en reproducibility of accelerometry based prediction models to estimate activity energy expenditure (AEE) in children and adolescents. Databases CINAHL, EMBASE, PsycINFO and PubMed/MEDLINE were searched. References of included articles were screened for additional relevant studies. Inclusion criteria were set for aim of the study: development or validation of a accelerometer based prediction model for the estimation of AEE in healthy children or adolescents (6-18 years), criterion measure: indirect calorimetry, or doubly labelled water, and language: Dutch, English or German. Studies concerning pedometers were excluded. Methodological issues were evaluated by a modified checklist. Data were extracted in a table with items considered relevant for this topic by known authors (Puyau and Trost). Nine studies were included. The median methodological quality was 6.0 ± 1.6 (out of 10 points). Prediction models combining heart rate and counts explained 86-91% of the variance in measured AEE. A prediction model based on a triaxial accelerometer explained 90% of the variance in measured AEE for indoor activities. Models derived during free-living explained up to 45%. Accelerometry based prediction models may provide an accurate estimate of AEE in children on a group level. The estimation of AEE is more accurate when the model is derived (and used) in a laboratory setting. Best results are retrieved when the model combines accelerometer counts with heart rate or a triaxial accelerometer is used. Future development of equations applicable to free-living scenarios is needed.

Purpose of this study was to test the validity of activity energy expenditure (AEE) estimated from existing prediction equations solely based on accelerometry (Actical & Actiheart), solely based on heart rate (Actiheart) and a prediction equation combining accelerometry and heart rate (Actiheart) in children with Spina Bifida (SB). Twenty-seven children aged 6-16 year with SB and at least community ambulatory according the adapted Hoffer scale participated in this study. The test protocol included: 10 minutes rest, 6 minutes walking, 15 minutes recovery period and an incremental treadmill test. During the protocol AEE was measured by indirect calorimetry and compared with predicted AEE by Actical and Actiheart. Correlation between counts and measured AEE varied (0.162 to 0.737). All prediction equations overestimated AEE. Actiheart equations showed substantial agreement for the 6 minutes walk test (0.646 and 0.666) or the treadmill test overall (0.617). AEE predicted by the Actical showed overall low agreement with measured AEE (0.114 to 0.393). Actiheart Combined, a prediction equation combining accelerometer counts and heart rate provided the most accurate prediction of AEE (0.666) with a substantial degree of error (coefficient of variance 21.2%). Tested prediction equations provided no valid estimation of AEE in children with SB. Individual calibrated prediction equations combining accelerometry and heart rate might be the most appropriate for children with SB. The Actiheart seems the most promising accelerometer for this group.