Comprehensive food and nutrient database of the DietSys system for diet surveys in Brazil, Argentina, and the United States

DEFAGO MD; ROSATTO S; FUNG T; PERDOMO M; WILLETT W; SAMPSON L

Buenos Aires

Congreso; IUNS 21st International Congress of Nutrition; 2017

Sociedad Argentina de Nutrición

databases. Background and objectives:  Theneed for information on food composition that is up to date, accurate andreliable is critical to establishing a food policy aimed at promoting health. Mostlocal food and nutrient databases lack food items and/or bromatologicalanalysis and are impractical for use in many countries owing to differences ofnutrient composition and eating habits. The DietSys diet-data processing systemallows use of customized databases to calculate food components. We aimed tomerge three major nutrient composition tables into a comprehensive food andnutrient database to be used in the DietSys system in Brazil, Argentina, andthe United States.Methods: Three nutrient composition tables were unified: the Brazilian foodcomposition table (TACO), the Argentinean food composition table (Argenfoods),and the U.S. Department of Agriculture, Agricultural Research Service (USDA),National nutrient database for standard reference, Release 27. We created 21food groups and used the NOVA system to create four food processing categories(FPC): unprocessed or minimally processed foods (UMP), processed culinaryingredients (PCI), processed foods (PF), and ultra-processed foods (UPF). Foreach FPC and food group, we calculated the number and proportion (%) of fooditems and the average density of energy (kcal/100 grams) and macronutrients(grams/100 grams). Results: The DietSys database comprised 9386 food items: 6.4% from TACO, xx%from Argenfoods, and 93.6% from the USDA. Of food items classified as UMP,49.6% were from TACO, while xx% were from Argenfoods and 22.6% from the USDA.The USDA contained 27% fewer UMP (N=1964 out of 8789), 22% more PF (N=2236),and 44% more UPF (N=4203) items than the TACO. Average density of energy(kcal/100 grams), proteins, lipids, and carbohydrates (grams/100 grams) of PCIwas 1%, 36%, 23% and 39% higher in TACO food items than the USDA. Some foodgroups included more than one FPC (e.g. in TACO 16.6% of meats and derivativeswere UMP, 35.7% were PF, and 3.7% were UPF; whereas in the USDA 40.4% of meatsand derivatives were UMP, 54.2% were PF, and 1.0% were UPF).   Conclusion: This food and nutrient database will allow the DietSys user tocalculate the intake of foods, food groups, and FPCs, and nutrient content forBrazilian, Argentinean, and North-American populations, and to prioritizeregional food items or to select alternative items. It is a dynamic andflexible tool that allows the addition of other foods and nutrients in thefuture.