The study was based on a cross-sectional study of people aged 40 - 69 years in Ohshima-gun in the Amami Islands, which are located in south-western Japan. Data and blood were collected from participants undergoing routine health checkups that were conducted by the local government or private companies, after receiving each person’s written informed consent. The survey comprised a questionnaire and blood collection was conducted in October 2005. In total there were 1131 participants (618 women and 513 men). Participants were excluded from the study if they had missing questionnaire data or biochemical parameters for blood. Finally, 1099 participants were included in the analysis.

This study was conducted with the approval of the Ethics Committee of Epidemiological Studies of Kagoshima University Graduate School of Medical and Dental Sciences (No. 76, approved on 7 September 2010) and in accordance with the principles contained within the Declaration of Helsinki.

We used the standardized questionnaire that was used in the J-MICC study [11]. The questionnaire included questions regarding smoking, alcohol, prescription medicine, and FFQ. The FFQ collected information on average food consumption of 43 food items and eight types of non-alcoholic drisnks (Table 1) [12]. To evaluate food consumption, the number of times each type of food was consumed was given a score: 1) none; 2) 1 - 3 times/month; 3) 1 - 2 times/week; 4)

Values ^†Serving frequency was expressed as below, 1) none; 2) 1 - 3 times/month; 3) 1 - 2 times/week; 4) 3 - 4 times/week; 5) 5 - 6 times/week: 6) 1 times/day; 7) 2 times/day; 8) >3 times/day. ^‡Drinking volumes was expressed as below, 1) none; 2) <2 cups/week; 3) 3 - 4 cups/week; 4) 5 - 6 cups/week; 5) 1 - 2 cups/day: 6) 3 - 4 cups/day; 7) >5 cups/day. One cup is 175 mL. *Significant difference between men and women, p < 0.05. ^‡Significant difference between non-smoking and smoking, p < 0.05. ^§Significant difference between premenopausal and postmenopausal, p < 0.05. Abbreviations: B-vegetables, Brightly colored vegetables; Fish egg, Processed foods from fish egg; V-juice, Vegetable juice; F-juice, Fruit juice; L-vegetables, Light-colored vegetables; S-fish, Steamed processed fish meet products; F-fish, Fried processed fish meet products..

3 - 4 times/week; 5) 5 - 6 times/week: 6) 1 time/day; 7) 2 times/day; 8) >3 times/day. For drinks, the number of times (measured in cups) each drink was consumed was given a score: 1) none; 2) <2 cups/week; 3) 3 - 4 cups/week; 4) 5 - 6 cups/week; 5) 1 - 2 cups/day: 6) 3 - 4 cups/day; 7) >5 cups/day. One cup was 175 mL. The accuracy of the FFQ was evaluated by weighed dietary records [13]. The methods used for the collection of physiological data and biochemical parameters of blood including Hcy have been described previously [13].

All experimental data were analyzed with JMP software (ver. 9; SAS Institute Japan Ltd., Tokyo, Japan). We first evaluated whether the Hcy concentration was effected by sex, smoking, menopause, or drug use. Sex, smoking in men, and menopause were associated with Hcy concentration but drug use was not. Therefore, we divided participants by these factors in the analysis. The unpaired Student t-test was used for comparisons between two groups. ANOVA was used for comparisons of physiological data among three Hcy concentrations. Stepwise regression was used for the analysis of the relationship between Hcy concentration and food consumption. Factors were chosen if they had an F-value > 2. Biological data such as HDL, FBS, BUN, and eGFR were added as factors in men. Age, BUN, eGFR, and lactate dehydrogenase (LDH) were added as factors in women. All data are expressed as the average ± SD. A p-value < 0.05 was considered statistically significant.

The Hcy concentration was significantly lower in women compared with men. This finding aligns with previous research [7]. Table 2 shows the analysis by sex. Age showed a weak positive correlation with Hcy concentration in women (r = 0.17, p < 0.0001) but not in men. Age has previously been noted to be associated with increased Hcy concentrations. All participants were more than 40 years old in the current study. It is possible that the absence of a strong correlation between Hcy and age in men resulted from the age of participants. The Hcy concentration may increase up until the age of ~40 years in men and then be maintained at the same level thereafter. Our previous unpublished results showed that the mean Hcy concentration was low in young people (10.8 ± 4.3 µM; 26.3 ± 6.1 years old; n = 9).

Values are means ± SD. ^†Significant difference between men and women, p < 0.05. ^‡Significant difference between non-smoking and smoking, p < 0.05. ^§Significant difference between premenopausal and postmenopausal, C0.05. ^¶mL/min/1.73 m².

The Hcy concentration was significantly high in men who smoked but not in women who smoked. As there were few women who smoked in the study population (n = 24), it is possible that this may have been the reason for no significant difference. Sobczak et al. reported that smoking is a strong risk factor for increased plasma Hcy concentrations [14].

It is reported that plasma Hcy concentrations increase in postmenopausal women [15]. In the current study, the Hcy concentration of premenopausal women was 1.6 μM higher than postmenopausal women. Several studies report that estrogen can reduce plasma Hcy concentrations, with low Hcy concentrations being observed in women [16]. It is, however, unclear how estrogen reduces the Hcy concentration.

Participants were divided into three groups by Hcy concentration: low, <15 μM; middle, ≤15 to <20 μM; and high, ≥20 μM [17]. We observed that BUN, Cr, and uric acid levels increased according to Hcy concentration in both women and men (Table 3). These indexes indicate kidney filtration function. Generally, eGFR calculated from plasma Cr, and adjusted by age and sex, is used to evaluate kidney filtration function. We found that increased Hcy concentrations were significantly associated with decreasing eGFR values. It is reported that homocysteinemia is associated with an accelerated decline in renal function [18] [19]. There is a tendency for increases in ALT but not AST to accompany increasing Hcy concentrations. ALT is a marker of liver inflammation. The liver, and particularly the kidney, are the major organs of Hcy metabolism [6]. Kidney function appears to be closely associated with homocysteinemia. In the current study, LDH was significantly associated with Hcy concentrations in women only. LDH is also one of the markers of liver injury.

We evaluated whether the Hcy concentration was affected by drug use and found that antihypertensive drugs, antidiabetic drugs, hypocholesterolemic drugs, painkillers, and sleeping pills did not affect the Hcy concentration in both women and men (data not shown).

We divided the population into non-smoking and smoking men and pre or postmenopausal women. Data of food consumption by group are shown in Table 1. Multivariate analysis was conducted using the stepwise method to determine the association between Hcy concentration and food consumption (Table 4). Several biomarkers associated with Hcy concentration (HDL, BUN, and eGFR in men, and age, BUN, eGFR, and LDH in women) were added as factors. We expected folate-rich food, such as bright or light-colored vegetables, to be associated with the Hcy concentration [8] [20] [21]. However, there was no common food associated with Hcy concentrations, except for in the non-smoking

Values are means ± SD. ^†mL/min/1.73 m². ^‡TBARS value was expressed as malondialdehyde equivalent. Effect of Hcy concentration in the linear regression model (Wald F test).

*p < 0.05; **p < 0.01; ***p < 0.001. Abbreviations: B-vegetables, Brightly colored vegetables; Fish egg, Processed foods from fish egg; V-juice, Vegetable juice; 4 F-juice, Fruit juice; L-vegetables, Light-colored vegetables. Data were adjusted by eGFR^†, or. eGFR and LDH^‡.

men group. In women, broccoli and vegetable juice showed a tendency to have a negative association between their consumption and the plasma Hcy concentration (p = 0.0548 and p = 0.0659, respectively). These results suggested that folate-rich foods were not strongly associated with a low Hcy concentration. Some results were unexpected. For example, consumption of cabbage in the smoking men group and in the postmenopausal women group was significantly associated with a high Hcy concentration, although cabbage is a folate-rich vegetable.

Consumption of eggs and processed foods from fish eggs in the smoking men group was significantly associated with a low Hcy concentration. Postmeno pausal women also showed the same tendency, but this was not significant (p = 0.0959). Generally, eggs are not folate-rich foods. Some reports have suggested that the bio-availability of folate is different among different food groups. Monoglutamyl folate is a major form of folate in eggs, whereas polyglutamyl folate is the major form in vegetables. To absorb folate, deconjugation of glutamate is needed [22]. Although there is no difference in bio-availability between polyglutamyl folate and monoglutamyl folate [23] , polyglutamyl folate is more easily broken down in the cooking process compared with monoglutamyl folate [24]. Therefore, the bio-availability of folate in eggs may be higher than that in vegetables. Eggs are also a rich source of choline in the form of a phosphatidylcholine. Choline is used in the methylation of Hcy to methionine [25].

In the current study, processed soybean foods, including Natto, Tofu, and Miso soup, were associated with low Hcy concentrations in women and men, but the types of processed soybean foods associated with Hcy concentrations were different among the groups. Soybeans fermented by Bacillus subtilis var. natto, called “Natto” in Japanese, is recognized as a significant factor associated with plasma Hcy concentrations in non-smoking men and premenopausal women. Tofu, which is coagulated soymilk using minerals, showed a slight association with low Hcy levels in the smoking men group (p = 0.0527). Miso, which is fermented by Aspergillus oryzae, is a Japanese seasoning and is used for Miso soup, and showed a significant association with low Hcy concentrations in postmenopausal women. Soybeans contain a small amount of folate compared with vegetables but also contain isoflavones. Isoflavones act in a similar way to estrogen. As mentioned, estrogen could result in a decrease in the Hcy concentration. We suspect that isoflavones in soybean products decrease the Hcy concentration. Song et al. conducted a meta-analysis on the relationship between soybean products and plasma Hcy concentrations. They concluded that soybean products did not have an association with the reduction in the Hcy concentration [26]. Their report did not contain any studies of Japanese populations. Further studies are needed to determine the relationship between soybean consumption and plasma Hcy concentrations.

Chinese tea showed a significant association with low Hcy concentrations in men but not in women. Oolong tea and jasmine tea were included in the Chinese tea category in the current study. There was no difference in consumption of Chinese tea between men (2.77 ± 1.74) and women (2.79 ± 2.04). Two studies have reported on the differences between men and women in terms of the health benefits of drinking Chinese tea, with the beneficial effects being more remarkable in men than in women [27] [28].

It is reported that catechins have an anti-folate activity because of their ability to inhibit dihydrofolate reductase [29]. Murakami et al. reported a relationship between a high intake of green or oolong tea and an increase in the plasma Hcy concentration [8]. No association was observed between green tea and Hcy concentration in the current study. Further studies are needed to determine the effects of catechins on plasma Hcy concentrations in elderly Japanese people.

Mushroom consumption showed a slight but not significant association with low Hcy concentrations in the smoking men group (p = 0.0795) and in the premenopausal women group (p = 0.0573). Yang et al. reported that Lentinus edodes (Shiitake mushroom), the most popular mushroom in Japan, was associated with a decrease in the Hcy concentration in a mouse model of homocysteinemia [30]. The mechanism behind the Hcy-reducing effect of Lentinus edodes has not been determined.

In a study examining the relationship between dietary habits and plasma Hcy concentrations in a United States population, Ganji and Kafai reported that seafood, eggs, rice, pasta, bread, starches, meat, and fat were not associated with Hcy concentrations [31]. In the current study, there was a positive correlation between plasma Hcy levels and butter, beef or pork, canned tuna, pumpkin, and potato. Soft drinks were also associated with a high plasma Hcy concentration. Ganji and Kafai suggested that there was little tendency for soft drinks to increase serum Hcy concentrations. They also reported that frequent consumption of milk and yoghurt were inversely association with the serum Hcy concentration. Our results showed that the consumption of yoghurt in postmenopausal women had a negative association with plasma Hcy concentrations. These differences may have resulted from the ages of the two studies, with their population being younger than the current study. Mean of age of participants was 39.9 years and the maximum mean for serum Hcy concentration was 10.6 µM in their analysis.

This study has several limitations that must be considered when interpreting the results. First, we did not convert food consumption data to nutrient content. Comparisons should be made between plasma Hcy concentration and intake of folate, vitamin B₁₂ and vitamin B₆, because they were associated with the plasma Hcy concentration. Second, dietary habits in the Amami Islands may be different from other regions of Japan. The Amami Islands are known to be a region of high longevity. While the variety of foods available in the Amami Islands could be different to the mainland, transportation means that almost all food types can be found in grocery stores in Japan. Third, the number of participants in each study group was not large, as the total population was divided into groups by sex, smoking, and menopausal state.