Validity of Two Questionnaires and Two Prediction Equations for Estimating Total Energy Expenditure in Healthy Adults: A Doubly Labeled Water Study

Abstract

Abstract Objectives The doubly labeled water (DLW) method is the gold standard for assessing total energy expenditure (TEE), but is costly. Questionnaires and prediction equations for TEE are nearly cost-free but research on their validity is scarce. We evaluated the validity of TEE assessed by two questionnaires and two prediction equations in comparison with TEE assessed by DLW. Based on previous work, we hypothesized that the questionnaires would be valid at a group level, and that the prediction equations would be valid at an individual level. Methods Data from a 10-d observational study in 124 healthy, nonsmoking adults were used (63% F, aged 29.8 ± 12.2 y, BMI 24.5 ± 3.9 kg/m2 (Mean ± SD)). TEE was measured by DLW using a mixed oral 2H218O dose containing 0.15 g 2H218O and 0.07 g of 2H2O per kg body weight. Analysis of urine samples and calculations of TEE were carried out using standard methodology. TEE was estimated from the 7-day Physical Activity Recall (7dPAR; Sallis et al. 1985), the Block Work and Home Survey (BWHS; Block et al. 2009), Dietary Reference Intakes (DRI) equations for estimated energy requirements of adults with normal weight or overweight/obesity (IOM 2005), and the BOD POD air displacement plethysmograph. 7dPAR TEE was estimated by multiplying MET-min/day with resting metabolic rate estimated from DRI basal energy expenditure equations. The BOD POD measured body composition by densitometry and TEE was estimated from fat-free mass and fat mass (Nelson et al. 1992) and an activity factor (WHO 1985). Results TEE values were 2430 ± 535 (DLW), 2375 ± 445 (7dPAR), 2407 ± 750 (BWHS), 2335 ± 388 (DRI), 2134 ± 439 (BOD POD) kcal/d (Mean ± SD), with DRI and BOD POD significantly lower than DLW (P <0.01). Mean ± 2SD limits of agreement (kcal/d) between DLW and 7dPAR (−766, 877) and BWHS (−1420, 1468) were wider than those between DLW and DRI (−630, 822) and BOD POD (−463, 1057). The R2 and SEE of the method associations with DLW ranged from 0.17 to 0.54 and 264 to 688 kcal/d, respectively (all P = 0.000). Conclusions The 7dPAR and BWHS were valid for estimating TEE at a group level. While the DRI and BOD POD equations were more accurate at estimating TEE of individuals, none of the tools are recommended for individual assessment of TEE due to their low R2 and wide Bland-Altman limits of agreement with DLW. Funding Sources NIH R01 DK075862 and Purdue University.