A Comparison Of Female Only Predictive Equations Versus Measured Resting Metabolic Rate

Abstract

Established prediction equations, such as Dore, Garrow and Weber (GW), and Cunningham (CH) were developed to predict resting metabolic rate (RMR) in females. However, these female prediction equations should be reexamined to determine if an update is required. PURPOSE: To examine several well-known female only RMR prediction equations (RMRpredict) against measured (RMRmeas) values in healthy females. METHODS: Fourteen female participants (36.5 ± 16.2 yrs, 166.8±3.6 cm, and 62.7±33.2kg) participated in this study. Each participant followed standard pre-test guidelines and underwent a single RMR assessment using a metabolic cart calibrated before each test. Each participant laid motionless in a supine position under a ventilated canopy for 25-30 minutes. The most stable 5 minutes of resting data within ≤5% of coefficient of variation was defined as the measured resting metabolic rate. A one-way repeated measures ANOVA was used to compare the RMRmeas against RMRpredict (Dore, GW, and CH). Additionally, Bland Altman Limits of Agreement (LA) were reported as frequency of subjects outside of agreement compared to RMRmeas. For significant differences, Cohen’s d effects sizes were reported. All results are expressed as M ± SD with significance set at p ≤ 0.05. RESULTS: The GW (1408.3 ± 149.3 kcal/day) and CH (1497 ± 135.8 kcal/day) prediction equations did not differ from the RMRmeas (1485.4 ± 247.2 kcal/day) (p > 0.05). However, significant differences were observed for Dore (1104.5 ± 89.3 kcal/day; d = 2.71) prediction equation when compared to RMRmeas. Dore under-predicted 100% of cases under the LA and the Garrow had the highest variability over-predicting and under-predicting 50% and 21.4%, respectively. The Dore equation significantly underestimated RMR when compared to RMRmeas by 23.8 ± 9.3 %. CONCLUSION: Our preliminary data shows that the use of the Dore prediction equation underestimated RMR by 380.9± 157.9 kcal/day. Therefore, using this equation to estimate calories in females should be interpreted with caution. Although not statistically different from the RMRmeas, the GW equation had the largest variability of RMR estimates. Our data supports the continued use of the CH prediction equation, as no significant differences were observed compared to RMRmeas in a healthy population of females.