The prevalence of obesity has increased in virtually every sub-population in the United States over the past few decades. Although obesity and its implications for health are well described, the associated epidemiologic determinants have not been defined for either individuals or populations. Obesity is a complex disorder with multiple etiologies; nonetheless, the underlying process must be a result of excessive energy intake relative to energy expenditure. The importance of the components of the energy budget, i.e., total daily energy expenditure (TDEE), resting metabolic rate (RMR), physical activity (i.e., non-resting EE), and energy intake (EI), is not well understood, however. To date, methodological constraints have made it difficult to unravel the epidemiologic determinants of obesity. Simply put, there are no sensitive measures of TDEE, physical activity or EI presently being used in epidemiologic studies. Doubly labeled water (DLW) does provide a valid and reliable method of measuring TDEE in free-living individuals. Obesity and its sequelae are very prevalent among African-American women, but infrequent among the genetically related population in Nigeria. This natural contrast provides a unique opportunity to study the evolution of obesity risk. We will examine the relationships between the energy budget and weight gain at the individual level in each of these populations, and then compare the patterns across groups. We will also determine the relationship between the various components of the energy budget and associated secondary risk factors for cardiovascular disease. This study will test the following primary hypothesis: low physical activity levels, (i.e., non-resting EE) place women at risk for weight gain and increase in body fat. To test this hypothesis, we will use DLW and RMR measurements to determine the energy budget in population-based samples of 200 African-American and 200 Nigerian women. The relationship between non-resting EE and RMR and weight gain will then be examined in a 3-year prospective study. We will also obtain complementary anthropometrics and physiologic data using routine methods such as questionnaires and blood tests. The proposed study will not use an ecologic design, since conclusions will be based relationships defined among individuals, not the comparison of group means.
