Evaluation Of Energy Intake At A Range Of Energy Flux Derived From Self-report, Doubly Labeled Water, And Activity Monitors

Abstract

PURPOSE: Self-report estimates of energy intake (EI) are adversely affected by measurement error. Recently, a simple equation has been developed and validated to estimate EI based on the energy balance equation (EI= energy storage + energy expenditure). Our goal was to compare self-report estimates of EI against two alternative values calculated from the energy balance equation. METHODS: Body composition of participants (N=195, mean age= 27.9 years, 46% female) was measured at the beginning and end of a two-week assessment period using dual-energy X-ray absorptiometry. Resting metabolic rate (RMR) was measured by indirect calorimetry. Energy expenditure (EE) was assessed using the doubly-labeled water technique and an arm-based activity monitor (Sensewear® Mini, SWA). Self-reported EI was calculated using dietician-administered 24-hour dietary recalls (three days, including at least one weekend day). Two estimates of EI were calculated using a validated equation: energy equivalent of changes in fat mass and fat-free mass occurring over the assessment period, plus EE from either DLW or SWA. To evaluate estimates of EI, reporting bias ((estimated EI - EE from DLW)/EE from DLW x 100) and Goldberg Cut-offs (estimated EI/RMR) were calculated. RESULTS: Mean±SD EE values from DLW and SWA were 2731±494 and 2729±559 kcals/day, respectively. Self-reported EI was 2113±638 kcals/day, EI derived from DLW was 2723±469, and EI derived from SWA was 2720±730. Reporting bias for self-reported EI, DLW EI, and SWA EI are as follows: -21.5±22.2%, -0.7±18.5%, and 0.2±20.8%, respectively. Goldberg Cut-offs for self-reported EI, DLW EI, and SWA EI are as follows: 1.39±0.39, 1.77±0.38, 1.77±0.38, respectively. When examined by tertiles of EE, bias of EI derived from self-report, DLW, and SWA varied between low (-17.4±27.5%, -2.9±18.1%, and 4.5±22.3%, respectively) and high groups (-25.6±19.5%, 0.4±17.5%, -4.4±17.6%, respectively). CONCLUSION: These results suggest that estimates of EI based on the energy balance equation can provide reasonable estimates of group mean EI. When compared by different levels of energy flux, estimates from all methods varied by EE level, indicating a potential source of bias that requires further investigation.