Potential Bias of Doubly Labeled Water for Measuring Energy Expenditure Differences Between Diets Varying in Carbohydrate

Abstract

BackgroundVery low-carbohydrate diets have been reported to substantially increase human energy expenditure as measured by doubly labeled water (DLW) but not by respiratory chambers. Do the DLW data reflect true physiological differences that are undetected by respiratory chambers? Alternatively, are the apparent DLW energy expenditure a consequence of failure to fully account for respiratory quotient (RQ) differences between diets?\n\nObjectiveTo examine energy expenditure differences between diets varying drastically in carbohydrate and to quantitatively compare DLW data with respiratory chamber and body composition measurements within an energy balance framework.\n\nDesignDLW measurements were obtained during the final two weeks of month-long baseline (BD; 50% carbohydrate, 35% fat, 15% protein) and isocaloric ketogenic diets (KD; 5% carbohydrate, 80% fat, 15% protein) in 17 men with BMI 25-35 kg/m2. Subjects resided 2d/week in respiratory chambers to measure energy expenditure (EEchamber). DLW expenditure was calculated using chamber-determined respiratory quotients (RQ) either unadjusted (EEDLW) or adjusted (EEDLW{Delta}RQ) for net energy imbalance using diet-specific coefficients. Accelerometers measured physical activity. Body composition changes were measured by dual-energy X-ray absorptiometry which were combined with energy intake measurements to calculate energy expenditure by balance (EEbal).\n\nResultsAfter transitioning from BD to KD, neither EEchamber nor EEbal were significantly changed ({triangleup}EEchamber=24{+/-}30 kcal/d; p=0.43 and {triangleup}EEbal=-141{+/-}118 kcal/d; p=0.25). Similarly, physical activity (-5.1{+/-}4.8%; p=0.3) and exercise efficiency (-1.6{+/-}2.4%; p=0.52) were not significantly changed. However, EEDLW was 209{+/-}83 kcal/d higher during the KD (p=0.023) but was not significantly increased when adjusted for energy balance (EEDLW{Delta}RQ =139{+/-}89 kcal/d; p=0.14). After removing 2 outliers whose EEDLW were incompatible with other data, EEDLW and EEDLW{triangleup}RQ were marginally increased during the KD by 126{+/-}62 kcal/d (p=0.063) and 46{+/-}65 kcal/d (p=0.49), respectively.\n\nConclusionsDLW calculations failing to account for diet-specific energy imbalance effects on RQ erroneously suggest that very low carbohydrate diets substantially increase energy expenditure.