We analyzed dietary data for the general population to examine differences in nutrient intake values calculated using the 2015 (seventh) and 2020 (eighth) editions of the Standard Tables of Food Composition in Japan. The subjects were 7,596 men and 7,566 women who had completed the 3-day dietary record using the weighing method from the first to the seventh wave of the Longitudinal Epidemiological Study of Aging. There was a significant correlation between nutrient intakes estimated using the seventh and eighth editions, but there were differences in the mean percentages of energy (5.1％), carbohydrate (5.8％), and protein (6.0％) calculated as the sum of amino acids, for which the measurement method had been changed; there were also differences in nutrient intake values for which the numbers of components, such as organic acids, had been changed. These results indicate the possibility of systematic errors due to database revision. In addition, our results showed sex-related differences. The degree of influence of the two editions may differ depending on the diet quality of individual subjects. These findings suggest that it is important to revise the nutrition assessment databases carefully. However, in epidemiological studies, such as those in which subjects are ranked by nutrient intake, the impact of database revision might be relatively small.
