A Correction Factor Improves Met Estimates Of Physical Activities

Abstract

The standard 1 MET = 3.5 ml·kg-1·min-1 generally overestimates resting metabolic rate (RMR) for low-fit, overweight, older and female individuals. It is unknown what effect this systematic error has on MET estimates of physical activities in self-report instruments (SRI). PURPOSE: To examine differences in activity METs estimated using measured RMR (measMET) compared to 3.5 ml·kg-1·min-1 (calcMET) and corrected METs (corrMET= mean calcMET *(3.5/Harris-Benedict RMR)) in sub-groups by age group, BMI category, gender and fitness. METHODS: 252 subjects were classified as: normal weight (NW = BMI < 25 kg·m-2 n=159) or overweight (OW= BMI ≥25 kg·m -2 n=93), male (n=118) or female (n=134), age group (by decade 20,30,40, or 50 y n=70, 69, 53 and 59, respectively) and fitness quintiles (low to high fit; n= 30, 31, 73, 88 and 39, respectively). RMR was measured after a 15 minute rest following a 4-hr restriction of food, caffeine and exercise. Subjects completed 6 treadmill activities at 3 speeds (1.34, 1.56, 2.23 m·sec-1) at 0% and 3% grade and 5 self-paced activities of daily living (ADL). Each activity was performed for 7 minutes while wearing a portable metabolic unit to measure VO2. Average MET differences (calcMET-measMET and corrMET-measMET) were computed by sub-group. RESULTS: The (mean ±SD) measMET was 6.1 ±2.64 (n=2555). CalcMET (Mean (95% CI)) was (0.51(0.59,0.42) METs) less than measMET. CorrMET was not statistically different from measMET (0.02 (-0.11,0.06) METs). calcMET underestimated measMET for most individuals who were low fit (-1.04(-1.38,-0.71) METs), OW (-0.65(-0.81,-0.49) METs), older (50-60yrs,-0.67(-0.87,0.47)METs) and female (-0.55(-0.68,-0.41) METs). Differences between corrMET and measMET were less than calcMET and measMET for all subgroups. CONCLUSION: SRI's that transform activity into MET metrics (e.g. MET hrs·day-1) using the Compendium of Physical Activities which assumes the 3.5 ml·kg-1·min-1 baseline, will lead to less precise estimates of self-reported physical activity particularly in individuals who are overweight, older, low fit and women. The Harris-Benedict based correction factor improves MET estimates and reduces the systematic error observed within sub-groups. Supported by NIH RO1 CA121005