Resting Energy Expenditure of Black Overweight Women in South Africa Is Lower than of White Women

Abstract

Background: Overweight affects 65% of black South African women. Effective weight management requires accurate measurement or estimation of energy expenditure. Aims: The study aimed to determine, among overweight women, whether measured resting energy expenditure (REE) differs between black and white participants, and the performance of REE estimation equations. Methods: The REE of 44 black (age 39.6 ± 9.7 years, body mass index (BMI) 35.1 ± 6.2 kg/m²) and 41 white (age 38.0 ± 11.6 years, BMI 33.9 ± 7.6 kg/m²) women was measured with indirect calorimetry and estimated with equations. Body composition was assessed with multi-frequency bioelectrical impedance analysis. Differences in REE were determined with t tests (Welch), and included adjustment for fat free mass (FFM) and BMI, and for FFM index (FFMI). Results: Measured REE was 585 kJ/day (95% CI 264-905; p = 0.0005) and 861 kJ/day (95% CI 499-1,221; p < 0.0001) lower in black than in white women when adjusted for FFM and BMI, and FFMI, respectively. Out of 14 equations, 13 underestimated REE (error range 2,261 ± 727 kJ/day (Bernstein equation, white women) to 8 ± 782 kJ/day (BMI equation, black women)). Conclusions: Black overweight women have significantly lower REE than their white counterparts. No tested estimation equation provided satisfactory results across race/ethnicity. REE measurements or development of overweight- or race/ethnicity-specific estimation equations are recommended.

Resting Energy Expenditure and Related Factors in 6- to 9-Year-Old Southern African Children of Diverse Population Groups

“…Regardless of factors contributing to REE, acknowledging differences in REE between population groups is essential for acceptable estimation of REE in the prevention and management of childhood obesity. Previous studies at our institution [16,17] identified a lower REE in the black than white population group, specifically for women, which may contribute to the high prevalence of overweight or obesity in black South African women [4]. Our study results suggest that sex-associated population differences in REE may only occur after middle childhood but prior to adulthood and constitutes a future research question.…”

“…Com-parison of population groups with respect to REE, overall and within sex categories, was considered for sample size calculations. To detect a clinically relevant difference in REE of 75 kcal/day between groups and assuming a common SD of 84 kcal/day [16,17], a sample of 28 participants per group was calculated to have 90% power when using a two-sided two-group Student's paired t-test at the 0.05 level of significance. More accurate calculation of sample size before the onset of the study, and the estimation of oversampling requirements to overcome sampling challenges and non-compliance, were not possible since REE research in South.…”

“…Research to investigate the REE in Sub-Saharan population groups is limited [14]. In South Africa, two recent studies found lower REE in black adults compared to their white counterparts [16,17]. Studies to determine REE differences between children of various population groups are also limited to the African American and white populations and have indicated varied results [8,9,11,12,[18][19][20].…”

Resting energy expenditure of a diverse group of South African men and women

“…Weight and height were measured with a digital scale (Sensa 804; Seca) and a stadiometer (Seca), respectively. The standardised protocol followed for measurement of REE including pretesting of the metabolic cart, achieving and identifying steady state, body composition and anthropometry has been described previously 10 . All measurements were taken by trained dietitians.…”

“…For REE outcome variables, a quantile regression model with bootstrap estimates race/ethnicity by sex was employed with covariates (five models for REE adjusted for age plus (a) FFM, (b) FFMI, (c) FFM plus FM, (d) FFMI plus FMI and (e) height, respectively). The combinations for these models were guided by factors influencing REE of race/ethnic groups and the findings of our previous research 10 . Reliability/agreement of REE estimations was conceptualised in accordance with ‘GRRAS’ guidelines 12 .…”