Development of a prediction equation for estimating resting metabolic rate in adult African-American females

Resting Metabolic Rate (RMR) accounts for 60--75% of total energy expenditure in humans and is used by clinicians to estimate daily energy requirements and by health organizations to define population energy requirements. Direct measurement of RMR is not practical in these cases and thus, prediction equations have been developed to estimate RMR. Harris and Benedict developed the most widely used equations in 1919 and these equations are still widely used today. Questionable data collection techniques sparked several researchers to develop additional equations referred to in this investigation as the FAO equations. Both sets of equations have been reported to be poor predictors of RMR in various ethnic populations. It is well documented that the prevalence of obesity is greater in African-Americans (AA) and that RMR is lower in AA compared to Caucasians. The purpose of this investigation was to develop two equations for predicting RMR in AA females and to compare the results of these equations and the results of Harris-Benedict and FAO equations to RMR as estimated using indirect calorimetry.; One hundred AA females between the ages of 18 and 40 with various Body Mass Indexes (BMI) participated in this study. Fifty were randomly selected from the study group to develop the equation and the remaining fifty were used to test the equations. Participants' height (ht), weight (wt), and age were obtained and fat-mass (fm) and fat-free mass (ffm) were measured by Dual-Energy X-ray Absorptiometry (DXA) and Bioelectrical Impedance (BIA). RMR was measured by indirect calorimetry using a canopy system.; A forward stepwise multiple regression analysis was performed using SPSS software with measured RMR as the dependent variable and age, ht, and wt serving as independent variables for the Wells-Valliant equation; and age, ht, wt, FM and FFM serving as independent variables for the Wells-ValliantBIA equation. Variables were selected for inclusion in the equations if the variables significantly (p<0.05) correlated with RMR. The mean RMR as estimated by indirect calorimetry of the 50 participants in the test group was compared to the mean RMRs predicted by each equation using ANOVA. When significant F values were obtained, Tukey's post hoc test was used to identify significant differences between the means. Bland-Altman analysis was used to compare measured and predicted RMR.; The Wells-ValliantBIA equation was the only equation considered in this investigation that did not differ significantly from mRMR. In addition, the Wells-ValliantBIA equation had the largest R2 (.59) and smallest SEM (26.64) and therefore is recommended for use when predicting RMR in adult AA females.