Food safety and epidemiology: new database of functional food factors.

: Biofactors. 2004;22(1-4):213-9. Related Articles, Links
Watanabe S, Zhuo XG, Kimira M.

More than 600 functional non-nutrient food factors (FFFs) in vegetables and fruits are considered to be effective for health promotion and disease prevention. However, phytochemicals studied thus far have failed to yield predicted results in randomized intervention studies. To assess the health effects of phytochemicals, a breakthrough in epidemiological methods was necessary. We constructed a database of non-nutrient FFFs to estimate the chemical classes and total amount of FFF-intake in order to facilitate estimation and calculation for nutritional research. So far, flavonoids, terpenoids, carotenoids, and sulfur compounds are included in our FFF database. We calculated the intake of various phytochemicals per capita from 79 subjects’ dietary records by FFF database, and estimated that subjects ingested more than 10 micromole per day of phytochemicals such as catechin, isoflavones, isothiocyanate, ferulic acid, quercetin, cinnamic acid and chlorogenic acid. Chief component analysis yielded 12 factors (80%), of which only a few factors showed negative associations with serum cholesterol and LDL concentration. Many factors showed adverse relationships with liver function and serum triacylglycerol concentration. Weekly self-reported daily dietary records including name of dish, constituent foods and their amounts were separately collected for 6 months and analyzed in Kyoto women. Seasonal changes of phytochemical intake showed significant variation according to the seasonal consumption of fruits and vegetables. Lycopene increased in the summer due to watermelon and tomato intake in this season. Seasonal variation of FFF was quite large compared to the variation of macro- and micronutrients. Careful evaluation of the effects of FFF intake on health is necessary, especially when supplements are also consumed. The most effective combinations of FFF intake for human health could be elucidated by using our FFF-DB in conjunction with population-based cohort studies