Isoflavone and Soyfood Intake and Colorectal Cancer Risk: A Case-Control Study in Korea

We aimed to assess the relationship between dietary soyfood and isoflavone intake and colorectal cancer risk in a case-control study. A total of 901 colorectal cancer cases and 2669 controls were recruited at the National Cancer Center, Korea. A semi-quantitative food frequency questionnaire was used to assess the usual dietary habits, and the isoflavone intake level was estimated from five soyfood items. A high intake of total soy products, legumes, and sprouts was associated with a reduced risk for colorectal cancer in men and women, although the middle quartiles of intake of total soy products were associated with an elevated risk. In contrast, a high intake of fermented soy paste was associated with an elevated risk for colorectal cancer in men. The groups with the highest intake quartiles of isoflavones showed a decreased risk for colorectal cancer compared to their counterparts with the lowest intake quartiles in men (odds ratio (OR): 0.67, 95% confidence interval (CI): 0.51–0.89) and women (OR: 0.65, 95% CI: 0.43–0.99). The reduced risk for the highest intake groups persisted for distal colon cancer in men and rectal cancer in women. The association between soyfood intake and colorectal cancer risk was more prominent among post-menopausal women than pre-menopausal women. In conclusion, a high intake of total soy products or dietary isoflavones was associated with a reduced risk for overall colorectal cancer, and the association may be more relevant to distal colon or rectal cancers.


Introduction
Colorectal cancer is the third most common cancer in Korea, and the incidence rates have increased in recent decades. [1] Although the mortality rate of colorectal cancer in men has been stabile since 2002 and has decreased since 2004 in women, [2] colorectal cancer is the by Pearson's chi-square statistics. The residual method was used to adjust the total energy intake. [22] The intake levels of soyfoods and isoflavones were categorized into quartiles according to the distribution of the control groups, except for soy milk, which showed a low proportion of consumption. The association between dietary factors and colorectal cancer risk was assessed by analysis of logistic regression models with adjustment for potential confounding variables, and odds ratios (OR) and their 95% confidence intervals (CI) were calculated. For the subsite analysis, polytomous logistic regression models were used. All analyses were performed separately by gender, and SAS version 9.3 software (SAS Institute, Inc., Cary, NC, USA) was used.

Results
Although age was matched by a 10-year frequency, more male patients over 60 years of age were recruited. Male colorectal cancer patients were more likely to have less education and a lower household income. The patients were less like to participate in regular physical activity, and they had a high proportion of colorectal cancer in their family histories. Similar trends were observed for education, household income, and physical activity among the women. In addition, the female patients were more likely to have a body mass index of 25 kg/m 2 or more or to ever have smoked or consumed alcohol ( Table 1).
The most abundant dietary source of isoflavones for men and women was soybean curd (tofu), which amounted to 38~44% of the isoflavones consumed, followed by black soybeans, soybean paste, and soybean milk. Daidzein and genistein were the most abundant isoflavones in men and women (data not shown). For the total soyfood, the group with the second and the third quartiles of intake showed increased risk for colorectal cancer in men, whereas a reduced risk for colorectal cancer among the highest intake groups was observed for both men (OR: 0.67, 95% CI: 0.49-0.92) and women (OR: 0.62, 95% CI: 0.39-1.0) ( Table 2). Legumes and sprouts intake was inversely associated with colorectal cancer risk in both men and women, whereas fermented soy paste was associated with increased risk for colorectal cancer in men. Middle intake groups for tofu and soy milk showed increased risk for colorectal cancer in both men and women. Compared with the lowest intake quartiles group, the groups with the highest intake quartiles of total isoflavones showed a decreased risk for colorectal cancer in men (OR: 0.71, 95% CI: 0.52-0.97). A separate analysis of three isoflavones (daidzein, genistein, and glycitein) yielded results very similar to those for the total isoflavones.
In the subsite analysis, a high consumption of total soy products and legumes was associated with the risk for distal colon cancer in men (Table 3) and rectal cancer in women (Table 4). However, similar to the results for overall colorectal cancer, the second and the third quartiles of intake groups for total soy products or tofu showed an elevated risk for proximal colon cancer, and rectal cancer in men and distal colon cancer in women. Soy milk was associated with a reduced risk of cancer in all the subsites in men, whereas no apparent association was observed for each subsite in women. Sprouts were associated with a reduced risk for all the subsites in men and women. The reduced risk for the highest intake groups of total isoflavones, daidzein, genistein, and glycitein persisted for distal colon cancer in men (Table 3) and rectal cancer in women (Table 4).
In the analyses by the menopausal status of women, a decreased risk for colorectal cancer among the highest quartile intake groups for total soy products, legumes, sprouts, and glycitein was observed only among postmenopausal women (Table 5).

Discussion
Although the middle quartiles of intake was associated with an elevated risk, the highest intake quartile of total soy products or dietary isoflavones is associated with a reduced risk for overall  colorectal cancer, and the association might be more relevant to distal colon or rectal cancers. The reduced risks for colorectal cancer among high intake groups were most consistent for legumes and sprouts in both men and women and for all subsites.
A meta-analysis of epidemiological studies on soy food intake and colorectal cancer risk showed a 21% reduction in the colorectal cancer risk among the high intake groups in women (the combined risk estimate: 0.79, 95% CI: 0.65-0.97), whereas there was no apparent benefit from soy consumption on the colorectal cancer risk in men (the combined risk estimate: 1.10, 95% CI 0.90-1.33). [13] The subgroup analysis for the types of soy products or study population did not show significant heterogeneity in the pooled risk estimates. [13] Among soy foods,   tofu [23][24][25][26][27][28] and miso soup [14,25,[28][29][30] have been the most widely analyzed. We found a consistent association between a high intake of legumes and a reduced risk for overall colorectal cancer, and a high intake of fermented soy paste was associated with an increased risk for overall colorectal cancer as well as cancer of all the subsites in men. Legumes contain carbohydrates (6-62%), proteins (20-30%), and fiber (3-31%), depending on the type. [31] Legumes are important sources of minerals such as iron, zinc, calcium, and selenium as well as nonnutrient compounds such as lectins, agglutinins, protease inhibitors, bioactive peptides, phenolic compounds (tannins), and saponins. [31] Additionally, soybeans are an important source of linoleic and linolenic acids. [31] Fermented soy paste contains high salt, however, and there is some limited epidemiological evidence that a high salt intake has a role in colorectal carcinogenesis. [32] Genistein, the most predominant isoflavone, is known to inhibit prostate carcinogenesis in mouse models, to modulate genes that are involved in the control of the cell cycle and apoptosis, and to have antioxidant properties. [33] In addition, equol, a plant isoflavone metabolite produced by ruminal bacteria of mammalian, has a weak estrogenic effect, which has been linked to a protective effect against breast carcinogenesis. [34]Among nine epidemiological studies on isoflavone intake and the colorectal cancer risk, four case-control studies and one prospective cohort study reported a reduced risk for colorectal cancer among high intake groups, [15,16,18,35,36] whereas three cohort studies did not find any association. [14,17,37] A Japanese case-control study suggested a beneficial effect of isoflavones on the colorectal cancer risk only among postmenopausal women. [38] Two studies on gender-specific association, decreased risk of colorectal cancer was observed only among women for soy food. [17,39] In a case-control study conducted in Hong Kong, women who consumed soy products 4 times per week or more showed decreased risk for colorectal cancer (odds ratio (95% CI): 0.47 (0.28-0.81)), whereas statistical significance was not reached among men (odds ratio (95% CI): 0.66 (0.40-1.08). [39] In a Japanese cohort study, high intakes of total soy food (hazard ratio (95% CI): 1.24 (0.77-2.00) for men and 0.56 (0.34-   [17] Other studies, however, did not find a significant difference in risk between men and women. [14,27,37,38]A cohort study of Chinese women reported a high intake of soyfoods and a reduced risk of colorectal cancer only among postmenopausal women and in rectal cancer. [18] With the result from a Japanese case-control study, [38] a more prominent association between soy or isoflavone intake and colorectal cancer among postmenopausal women is consistent with the results of this study. The putative biological mechanism that connects postmenopausal colorectal cancer and soy foods is the estrogenic effect of soyfoods. [18] A diphenolic structure of isoflavones is similar to the structure of 17β-estradiol, and various isoflavone compounds bind to estrogen receptor α and ß. [40] An in vitro study showed that isoflavones act as estrogen antagonists with a premenopausal (high) dose of estradiol, whereas they act as estrogen agonist in a lowestrogen environment near the serum level of postmenopausal women. [40] Few studies have assessed the association between soyfoods or isoflavones and the colorectal cancer risk by subsites. [14,15] The Fukuoka Colorectal Cancer Study reported an inverse association of soyfoods and isoflavones with rectal cancer in men and overall colorectal cancer in postmenopausal women. [15] The subsites of colorectal cancer show different molecular profiles, such as CpG island methylator phenotypes and microsatellite instability. [41] In addition, the estrogen receptor ß expression in normal epithelium was higher in the ascending colon than in the descending colon, [42] and it might partly explain the right-side dominance in colon cancer in women and the difference in susceptibility to isoflavone intake. [1] There is substantial variation in the intake level of soyfood and isoflavones among different populations. [43] The median intake levels of isoflavones among the control groups were 12.47 mg/day for men and 13.46 mg/day for women. These intake levels are much higher than those observed in studies from Canada, [16] England, [37] and Italy [36] and comparable or lower than those in studies from Japan [14,15,17,38] and China. [18] The strengths of this study include detailed information on the main exposures such as different types of soyfoods and isoflavones as well as on the anatomical subsites of colorectal  cancer. The limitations include the use of only five soyfood items to calculate the dietary isoflavone intake. Limited numbers of premenopausal women included in our study might lead to low statistical power to detect meaningful association. The highest intake quartiles of soyfoods or isoflavones are associated with a reduced risk for overall colorectal cancer. In the subsite analysis, a reduced risk among the highest intake groups for soyfoods or isoflavones is more prominent for distal colon cancer in men and rectal cancer in women. Additionally, the protective effects of soyfoods are observed only in postmenopausal women. Although precaution is required for the probably elevated risk among middle intake quartiles, our results suggest a potential role of soyfoods in colorectal cancer prevention, and the effect might differ by colorectal cancer subsite, gender, and menopausal status of women.