Estimation of nutrient intake by the new version of Japanese food composition tables in comparison with that by the previous version

This study was initiated to examine if the revision of the food composition tables from the 4th version to the 5th version has induced any substantial changes in nutrient intake estimation, and if so, in which nutrient item and to what extent. For this purpose, 24-hour duplicates of food samples were collected in 1996 to 1998 from 71 girl university students, together with food menus of the day. The nutrients in the duplicates were estimated by use of the 4th and the 5th version food composition tables in parallel, with supplements of other databases. The results of the estimation with the two databases were compared by Student's paired t-test. Compared with the results with the 4th version, the estimation utilizing the 5th version gave 3 to 8% increases in intakes of major nutrients including energy (except for protein), a 20% increase in vitamin A, and decreases in iron (-13%) and NaCl (-3%), whereas there were only small or essentially no changes in other minerals and vitamins. The largest increment in energy through the recalculation came from the cereals (especially rice), and it was meats for the increment in lipid intake. The increment in vitamin A with the 5th version is primarily due to the re-evaluation of fruits and vegetables. The present observation suggests that care should be practiced in examining secular trends in nutrition when the basis of estimation is shifted from one database of the 4th version of food composition tables to another of the 5th version. Further study of a larger scale is apparently warranted to confirm the conclusions.