Japanese Intake of Flavonoids and Isoflavonoids from Foods

The intake of flavonoids and isoflavonoids was estimated based upon a preliminary database of 40 food items, covering at least 80% of total food consumption. Fifty volunteer women in "ICity" recorded the weight of all dietary intake for 3 days in September 1996, and received a health check-up, as well as laboratory examination. The data was analyzed in relation to the various food factors. Average daily intake per capita of flavonoids was as follows: 4.9 mg kaempferol, 8.3 mg quercetin, 1.5 mg rutin, 0.6 mg myricetin, 0.3 mg luteolin, 0.01 mg myricitrin, 0.4 mg fisetin, and 0.3 mg eriodictyol. Total intake from vegetables and fruits was less than 10 mg. 16.2 mg (range: 3.18-35.61 mg) and 23.27 mg (4.62-52.12 mg) of isoflavones, such as daidzein and genistein, respectively, were taken per day, and total isoflavone intake was 39.46 mg (7.80-87.73 mg). Chief component analysis on ingested vitamins, flavonoids and isoflavonoids was carried out. Factor 1 was mainly composed of flavonoids and antioxidant vitamins. Factor 1 was positively associated with age and the level of HDL cholesterol and negatively related to the level of triglycerides. Factor 2, which was mainly composed of isoflavonoids, was positively associated with creatinine and uric acid levels. So far, these factors did not show a significant association with bone density and other health indices, such as BMI and blood pressure. J Epidemiol, 1998 ; 8 : 168-175.


Intake of Flavonoids and Isoflavonoids from Foods
Mitsuru Kimira 1, Yusuke Arai 1, Kayako Shimoi 2, and Shaw Watanabe 1 The intake of flavonoids and isoflavonoids was estimated based upon a preliminary database of 40 food items, covering at least 80% of total food consumption.Fifty volunteer women in "I-City" recorded the weight of all dietary intake for 3 days in September 1996, and received a health check-up, as well as laboratory examination.The data was analyzed in relation to the various food factors.
Chief component analysis on ingested vitamins, flavonoids and isoflavonoids was carried out.Factor 1 was mainly composed of flavonoids and antioxidant vitamins.Factor 1 was positively associated with age and the level of HDL cholesterol and negatively related to the level of triglycerides.Factor 2, which was mainly composed of isoflavonoids, was positively associated with creatinine and uric acid levels.So far, these factors did not show a significant association with bone density and other health indices, such as BMI and blood pressure.J Epidemiol, 1998 ; 8 : 168-175.
Non-nutrient food chemicals (factors) carry on various pharmacological activities in the human body.Antioxidant properties of flavonoids, such as quercetin and kaempferol, are considered to attribute to the low risk of coronary heart disease in western countries 1, 4 3).Japanese consume a lot of soy bean products, containing phytoestrogens, such as daidzein and genistein, which may lead to the low incidence of estrogenrelated cancers ', ).More than 600 food factors in vegetables are considered to be effective for cancer prevention through various influences on the metabolism inside the body.
Accurate estimation of these chemical intake becomes important, but an appropriate database for Japanese foods has not been available.A preliminary database was built in this study to estimate the amount of flavonoids and isoflavonoids ingested by Japanese.

MATERIALS AND METHODS
Intake of food factors (phytochemicals) was calculated from the field study in "I-City" of Iwate Prefecture.This field study has been continued since 1980 under collaboration with the local government.In early September, 1995, volunteer housewives in fifty families were asked to record the weight of all food intake and cooking methods for 3 consecutive days.Their age distribution was from 32 to 68 years of age.The dietary records were checked through interviews by trained dieticians on the fourth day when the physical examinations were scheduled.Samples of 10 ml of blood were also collected from each volunteer for blood chemistry analysis.
Laboratory of Nutrition and Epidemiology, Department of Applied Bioscience, Tokyo University of Agriculture.2Department of Food Science , Shizuoka Prefectural University.
Physical The intake of flavonoids and isoflavonoids was estimated by calculation from the database.Our own database of flavonoids was determined by Dr. Shimoi 0, and derived from flavonoid data in foods.Optimization of a quantitative HPLC determination was employed for measuring flavonoids *.Quercetin, kaempferol, myricetin, fisetin, eriodictyol and luteolin were measured by freeze-dried vegetables and fruits after acid hydrolysis of the parent glycosides.In some foods containing rutin and myricitrin, such as bean sprouts, grapefruits, moroheiya, onion, tomato and green tea, methanol extraction yielded more flavonoids; but we employed a value for acid hydrolyzates, because of the more plausible similarities in cooking and absorption into the body.Isoflavonoid value in  soybean paste (20.9%), natto (14.7%) and cooked soybean (14.8%).Foods containing daidzein were similar.Intake of flavonoids, isoflavonoids, vitamins and carotenoids are summarized in Table 6.
Intake of flavonoids showed significantly positive correlation with vitamin C intake, and carotene intake (Fig. 2ab).
Principle component analysis of these factors yielded several groups (Table 7).Factor 1 was mainly composed of retinol, caroteinoid, vitamin C, rutin, myricetin and kaempferol, suggesting a rich vegetable intake; factor 2 was composed of daidzein, genistein, myricitrin, eriodictyol, and tocopherol, suggesting a rich pulse intake; factor 3 showed positive direction for myricitrin and eriodictyol and negatively for daidzein and genistein.Association between these factors and food items is shown in Table 8.Factor 1 was significantly associated with vegetables, fish and potatoes; factor 2 with legumes and fat and oil; factor 3 showed positive correlation with noodles and milk, and negatively with legumes.Association between the above factors and physical and laboratory data of blood chemistry showed that factor 1 was positively related to age and HDL cholesterol, and negatively associated with triglycerides.Factor 2 was positively associated with creatinine and uric acid (Table 9).Factor 3 positively associated with body weight and negatively associated with GOT.
Intake of flavonoids and BMI did not show any significant association (Fig. 3a), but those who consumed more flavonoids belonged to the ideal range of BMI.Neither intake of isoflavonoids nor flavonoids was significantly associated with bone density (Fig. 3b).

DISCUSSION
The importance of non-nutrient phytochemicals for human health has been recognized).
In "I-City" of Iwate Prefecture, we have continued a nutritional survey for more than 15 years.The participants in the study were volunteers with high incentives for good nutrition, so the rates of nutrient intake satisfied the level of RDA for Japanese, except for calcium intake.SMR of cancer in "I-City" in 1990 was 0.93, but that of cerebrovascular diseases was 1.40.10).To understand the effects of phytochemical intake on human health, we tried to estimate the amount of intake of flavonoids in this area.This research also provides fundamental data for future study in different areas in Japan.Isoflavonoid intake was also estimated, because blood and urinary levels were very high compared to the Caucasians n) Data on the content of flavonoids and isoflavonoids in Japanese foods had been quite insufficient up to this point.Hertog et al 12) determined the content of quercetin, kaempferol, myricetin, apigenin, and luteolin among 28 vegetables and 9 fruits by Reversed phase-HPLC with UV detection for estimating flavonoids intake from foods.Quercetin levels in the edible parts of most vegetables were generally below 10 mg/kg except for onions (284-486 mg/kg), kale (110 mg/kg), broccoli (30 mg/kg), French beans (32-45 mg/kg), and slicing beans (28-30 mg/kg).Kaempferol could only be detected in kale (211 mg/kg), endive (15-91 mg/kg), leeks (11-56 mg/kg), and turnip tops (31-64 mg/kg).The content of myricetin, luteolin, and apigenin was below the limit of detec-tion (<lmg/kg) except for fresh broad beans (26 mg/kg myricetin) and red bell peppers (13-31 mg/kg luteolin).Seasonal variability was low for most vegetables except leafy vegetables which have the highest flavonoid levels in summer.
Our data base is still preliminary, but it covers 40 food items and would cover 77% of flavonoid and isoflavonoid intake from vegetables, 99.5% from legumes, 58.4% from potatoes, and 43.5% from fruits included in this study.In the Zutphen Elderly Study, mean baseline flavonoids intake was 25.9 mg per day).The major sources of intake were tea (61%), onions (13%), and apples (10%).Rimm et al 13) reported the average intake of flavonoids of Americans was 20 mg/day coming from black tea (25%), onions (25%), apples (10%) and broccoli (7%).
Japanese in our study consumed 16.22 mg of flavonoids, which seemed to be less than Caucasians.Although catechin was not included in the present study, average tea consumption was 300 ml, in which the amount of flavonoids would be less than 3 mg.On the contrary, isoflavonoid intake was much higher than flavonoids.The isoflavonoid level found in Caucasians was reported to be low, but the Japanese level was very high lu, although inter-individual differences were large.
The health effects of these chemicals are still under observation in Japan.These could be clarified by a nested-case control study in the large prospective study M).The average intake of flavonoids was inversely associated with mortality from coronary heart disease and provided an explanation for 25% of the variance in coronary heart disease rates in the 16 cohorts of the Seven Country Study 15).In multivariate analysis, intake of sat-urated fat (73%; p=0.001), flavonoid intake (8%, p=0.01), and percentage of smokers per cohort (9%; p=0.03) together explained 90% of the variance in coronary heart disease rates.These variables were independent of intake of alcohol and antioxidant vitamins.Average flavonoid intake may partly contribute to differences in coronary heart disease mortality across populations, but it does not seem to be an important determinant of cancer mortality.
The preventative properties of flavonoids are considered to be due to their strong antioxidant action 16).Hertog et al. 1) measured quercetin, kaempferol, myricetin, apigenin, and luteolin in the Zutphen Elderly Study.The relative risk of coronary heart disease mortality in the highest versus the lowest tertial of flavonoid intake was 0.42 (95% CI; 0.2-0.88).After adjustment for various factors, the risk was still significant (0.32 (0.15-0.71)).Intake of tea, onions and apples as major sources of flavonoids was also inversely related to coronary heart disease mortality, but this association was weaker.
In some reports, however, such benefit was not confirmed.Rimm et al 13) reported that the upper/lower flavonoid intake quintile RR of non-fatal cardiac infarct (496 cases) was 1.08 (95% CI; 0.81-1.43)among 34,789 males (40-75 y/o).Hertog et al 17) found in a Welsh population of men in the Caerphilly Study, that men with the highest consumption of tea (>1.2L, or >8 cups/d) had an RR of 2.4 (95% CI: 1.5-3.9) of dying by ischemic heart disease, during the follow-up period, compared with men consuming <300 ml/d (2 cups/d).They considered that adding milk to tea destroyed the plasma antioxidant-raising capacity of tea.Interaction of these phytochemicals during cooking and with other biomolecules in the body are considered important.
The mechanism of the biological effects of different metabolites also need to be clarified.Manach et al 18) studied the dietary quercetin recovery in rat plasma and found conjugated derivatives of isorhanmetin, 3'-O-methylated form of quercetin in the plasma from the rats fed 0.25% quercetin.After deconjugation, the concentration of aglycones in the plasma reached 120 ± 16 /1 mole/l, with an isorham-netin/quercetin ratio of about 5.In bile and in urine, where the 4-oxo-flavonoid concentration were 378 ± 42 and 128±19/1 mole/L, respectively.Watanabe et al. 19) reported the metabolism of isoflavonoids after intake of 60 g soy bean powder in men, and found that the blood level reached 4.0/1 mole/L which is much higher than that of the endogenous 17-beta estradiol level.This indicated complicated interaction of food-derived phytochemicals (factors) and body constituent biomolecules taking place inside the body.
Beta-carotene was once considered to be a representative antioxidant for cancer prevention, but interaction with other chemicals should occur.The distinction of various food factors should be determined as to whether they are real or surrogate markers for epidemiological study 9).

Table 2 .
Physical Characteristics of Volunteer Participants.N=49

Table S .
Consumption of Vegetables, Nuts.Seeds and Tubes.

Table 4 .
Intake of Foods in 13 Food Categories.

Table S .
Association between Factors of Principle Component Analysis and Food Categories.

Table 9 .
Association between Factors of Principle Component Analysis and Physical and Laboratory Data.