Validation of combined tri-axial accelerometry and heart rate for predicting energy expenditure during walking in overweight and obese adults

Purpose: To validate previously established (Freedson and VM) and newly generated novel prediction equations in overweight and obese inidividuals during a 30 min treadmill walk and three km external self paced overground walk. Methods: Study 1: Twenty overweight and obese adults performed an incremental treadmill walk. EE was measured using breath-by-breath indirect calorimetry. Activity counts and HR were recorded throughout using an AT accelerometer. Study 2: Ten healthy adult males performed three discontinuous maximal tests comparing the laboratory based Oxycon Pro and K4b[sup]2. Study 3: Ten overweight and obese adults carried out a three km overground walk, whilst EE was neasured via the K4b[sup]2 and activity counts and HR were recorded via the AT. Results: The Freedson and VM prediction equations significant overestimated EE by ~40% and 31% respectively. There was good agreement between the Oxycon Pro and K4b[sup]2 when calibrated immediately before use with a mean difference of 38.4 ml.min[sup]-1 (VO[sub]2), 39.5 ml.min[sup]-1 (VCO[sub]2) and 0.76 L.min[sup]-1 (VE). Novel prediction equations combining VM, BM and HR provided the most accurate estimation of EE during indoor and outdoor walking with no difference between measured and estimated EE (P=0.114) with a SEE=0.82 kcal.min[-1] and showed a strong correlation with measured EE (r=0.891, p<0.001). Conclusion: Both the Freedson and VM equations are unsuitable to predict EE in overweight and obese inidividuals. During walking exercise, novel, multiple regression equations combining VM, BM and HR provide a more accurate estimate of EE in this population during treadmill and overground walking. Future directions should focus on the use of these prediction equations during weight loss intervention programmes.