Dietary energy density determined by eight calculation methods in a nationally representative United States population.

Dietary energy density [kcal/g (kJ/g)] influences energy intake under controlled laboratory conditions. Little is known about the energy density of the diets of free-living persons. Because energy density investigations are a relatively new endeavor, there are neither standard calculation methods nor published nationally representative values. This paper examines the calculation of energy density based on systematic exclusion of beverage categories, presents data on variability, and compares values by sex, age, and race/ethnicity in a representative sample of U.S. adults. Mean daily dietary energy density values for adults (aged >19 y) were calculated using two 24-h recalls from the Continuing Survey of Food Intakes by Individuals 1994-1996 based on food, food and liquid meal replacements, food and alcohol, food and juice, food and milk, food and juice and milk, food and energy-containing beverages, and food and all beverages. Energy density varied by calculation method, ranging from 0.94 to 1.85 kcal/g (3.93-7.74 kJ/g). Intraindividual-to-interindividual CV ratios were highest for the food and energy-containing beverages calculation. Men reported diets with a higher energy density than women for all calculation methods (P < 0.0001). There were differences by race/ethnicity and an inverse linear trend for age. These data indicate that beverage inclusion schemes should be clearly defined when reporting energy density values. In epidemiologic studies, calculations based on food and all beverages and food and energy-containing beverages may diminish associations with outcome variables. These nationally representative data, which provide an important frame of reference for other studies, indicate that dietary energy density differs by sex, age, and race/ethnicity.