Validation of the Italian food composition database of the European institute of oncology.

ABSTRACT

OBJECTIVE: To compare nutrient intakes obtained by chemical analysis of food composite or duplicate portion of diets with those obtained by weighed record method using the database of the European Institute of Oncology (EIO). SETTING: Nutrition Section, Department of Internal Medicine, University of Perugia, Italy. SUBJECTS: Fifteen subjects aged 40-59 y in 1960 (41 observations in three seasons), twenty-six subjects in 1965, and only nine remaining subjects in 1970 and 1991 were examined in Crevalcore. In Montegiorgio sixteen subjects aged 40-59 y in 1960 (39 observations in three seasons), thirty-two in 1965, twenty in 1970 and nine in 1991 were assessed. Forty-four subjects in Gubbio area (Biscina, Belvedere and Scritto; 21 males, 23 females; age 56.2+/-14.4 y) were evaluated in 1993 and 1994. METHODS: For dietary appraisal the individual weighed record method was used for 7, 3 or 2 days. Equivalent food composites were made up from local foodstuffs and the duplicate portions were chemically analysed for total nitrogen, fat, saturated and polyunsaturated fatty acids, carbohydrates, retinol, beta-carotene, thiamin and riboflavin. RESULTS: In Crevalcore, a significant difference for protein intake was found between analysis and calculation with EIO database in 1965 and 1991 (P<0.05). Fat intake was significant different for EIO database compared to analysis in 1965 survey (P<0.05), but not for other years. In Montegiorgio, there was a significant difference for protein intake between analysis and calculation with EIO database in 1970 and 1991 (both P<0.001). EIO database showed a significant difference in regard to analysis for fat intake in 1960 IV, 1965, 1970 and 1991 (P<0.05). In both areas there was a significant difference between analysis and EIO database for starch and fibre, but not for polyunsaturated fatty acids and soluble carbohydrates (all P<0.05). In Gubbio area, a significant difference was found between analysis and calculation with EIO database for fat, retinol, beta-carotene and riboflavin intakes (all P<0.05). CONCLUSIONS: According to previous and present studies food composition tables and databases, such as the EIO database, cannot be considered a reliable method to determine nutrient intakes, particularly for some vitamins.