Dietary Fibre and the Risk of Colorectal Cancer : a Case-Control Study

Colorectal cancer is one of the most commonly occurring cancers in the world. The incidence rate has been on the rise; from 2003 to 2007, it was 28.08 per 100,000 males and ranked third among all cancer cases in China (Chen et al., 2012). Etiological studies have shown that diet is a very important factor in colorectal carcinogenesis (Armstrong et al., 1975). Although some studies have been conducted on dietary fibre and the risk of colorectal cancer in western countries, the evidence has been inconsistent (Burkitt, 1971; Negri et al., 1998; Jansen et al., 1999; Slattery et al., 2004; Lin et al., 2005; Michels et al., 2005). Burkitt hypothesised that dietary fibre might be a protective factor for colorectal cancer (Burkitt, 1971). Some studies also reported an inverse association about fibre and colorectal cancer (Negri et al., 1998; Slattery et al., 2004). However, some recent studies have not shown a protective effect of dietary fibre on colorectal cancer (Lin et al., 2005; Michels et al., 2005). Fibre can be classified into soluble fibre and insoluble


Introduction
Colorectal cancer is one of the most commonly occurring cancers in the world. The incidence rate has been on the rise; from 2003 to 2007, it was 28.08 per 100,000 males and ranked third among all cancer cases in China (Chen et al., 2012). Etiological studies have shown that diet is a very important factor in colorectal carcinogenesis (Armstrong et al., 1975). Although some studies have been conducted on dietary fibre and the risk of colorectal cancer in western countries, the evidence has been inconsistent (Burkitt, 1971;Negri et al., 1998;Jansen et al., 1999;Slattery et al., 2004;Lin et al., 2005;Michels et al., 2005). Burkitt hypothesised that dietary fibre might be a protective factor for colorectal cancer (Burkitt, 1971). Some studies also reported an inverse association about fibre and colorectal cancer (Negri et al., 1998;Slattery et al., 2004). However, some recent studies have not shown a protective effect of dietary fibre on colorectal cancer (Lin et al., 2005;Michels et al., 2005).
Fibre can be classified into soluble fibre and insoluble

Dietary Fibre and the Risk of Colorectal Cancer: a Case-Control Study
Y Song 1& *, M Liu 1& , FG Yang 2 , LH Cui 1 , XY Lu 3 , C Chen 1 fibre, according to differences in water solubility. According to food sources, fibre can be further classified into grain, soy, vegetable, and fruit fibre and so on. One study showed that vegetable fibre can reduce the risk of colorectal cancer (Nomura et al., 2007), whereas other studies reported that vegetable fibre was not associated with a decreased risk of colorectal cancer (Terry et al., 2001;Lin et al., 2005). Some animal studies also have shown that various types of fibre have different effects on colon cancer (Bingham, 1990;Slavin et al., 1990;Schatzkin, 2000). Therefore, the relationship between dietary fibre and colorectal cancer is a conflicting subject. Moreover, few studies have been conducted on the associations between soluble and insoluble fibre and To further understand the association among food groups, fibre and colorectal cancer, we conducted a case-control study in Qingdao. We also analysed the relationships between fibre from different foods and colorectal cancer.

Study population
Cases were admitted to Affiliated Hospital of Qingdao University Medical College, Qingdao Municipal Hospital, Qingdao Central Hospital from October 2013 to June 2014. Controls were individuals who had medical examinations in Qingdao Huici Heath Examination Centre within the same time frame. Ethical approval was obtained from Registered Ethic Committee of the Affiliated Hospital of Qingdao University Medical College.
The inclusion criteria for selecting cases were as follows: newly diagnosed with electronic colonoscopy and pathology report, aged 30-70 years, living in Qingdao, and able to complete the interview independently. Ineligible cases were excluded for the following reasons: (i) suffered from cancer previously (n=12); (ii) in-complete or inconsistent dietary data (n=20); or (iii) familial adenomatous or hereditary non-polyposis colon cancer (n=15). Overall, 265 (colon cancer: 105; rectal cancer: 144; colon and rectal cancer: 16) eligible cases were included in the study.
The inclusion criteria for selecting controls were as follows: (i) matched for age (≤3 years) and sex with cases; (ii) no disease of intestine after medical check-up and no history of intestinal disease; and (iii) no family history of colorectal cancer in a first-degree relative. We selected 271 controls; of these, 19 were ineligible because of incorrect information. In total, 252 controls were included for the analysis.

Data collection
Data were collected through face-to-face interviews conducted by trained investigators. All participants completed a questionnaire that included two sections: a health-related factors questionnaire and a semi-quantitative food frequency questionnaire. Cases were asked to recall their consumption frequency and the typical amount consumed of various foods one year before diagnosis, and controls were asked to report the same information about the one year before the interview date. The healthrelated factors questionnaire included sociodemographic characteristics and lifestyle habits. The semi-quantitative food frequency questionnaire included 121 food items, which were categorised into 9 groups: cereal and cereal products; soy and soy products; vegetables; fruit; nut; fish; meat (pork, beef, lamb, poultry); eggs; and milk and dairy products. We provided colour photos of food to aid recall. The interview took approximately half an hour. The energy and fibre intakes were estimated using the 2004 China Food Composition Table (Yang et al., 2004). Soluble and insoluble fibre intakes were calculated according to a Chinese study (Yin et al., 2004).

Statistical analysis
Data were analysed with SPSS version 17.0. Differences between continuous variables were examined using t-tests. Chi-square tests were used to examine differences among categorical variables. T-tests were also used to assess the differences in mean intakes of some food groups between cases and controls. To estimate the risk of colorectal cancer from dietary fibre, we used unconditional logistic regression models. Fibre intake was grouped into four categories using quartiles as cutoff points based on the distribution of controls. The odds ratios (ORs) and 95% confidence intervals (CI) for each category used the lowest categories as the reference. Age, sex, smoking habits, drinking habits, physical activity, body mass index (BMI), and total energy were selected as potential confounding factors. Tests for trends were conducted using chi-square tests. p<0.05 was considered statistically significant.

Results
The characteristics of the study population are described in Table 1. There were no meaningful differences in age and gender because controls were matched with cases. No significant differences were observed in BMI, physical activity, total energy, smoking habits or drinking habits between controls and cases with colorectal cancer or colon cancer. A significant difference was observed in drinking habits between controls and cases with rectal cancer (p<0.05). Table 2 shows the mean intake of various food groups in cases and controls. For grains, fruit, meat, and sea-food, we did not find significant differences among colorectal, colon, rectal cancer and controls (p>0.05). For vegetables and total fibre, controls reported higher intakes than did cases with colorectal, colon and rectal cancer (p<0.05). For soy food intake, significant differences were observed between controls and cases with colorectal and rectal cancer, but not colon cancer.
Associations between various types of fibre and colorectal cancer are shown in Table 3. Similar inverse associations were also found between the highest intakes of vegetable fibre, soluble dietary fibre, insoluble dietary fibre, total fibre and the risk of colorectal cancer (Q4   Associations between various types of fibre and rectal cancer are shown in Table 5. For rectal cancer, the risk decreased with increasing intakes of vegetable fibre and total fibre (Q4 vs Q1: OR=0.53, 95%CI: 0.29-0.97; OR=0.52, 95%CI: 0.29-0.91). Rectal cancer was not associated with grain fibre, soy fibre, fruit fibre, soluble fibre, fibre, soluble fibre or insoluble fibre. Though no significant relationship was observed between soluble fibre and rectal cancer, the risk of rectal cancer displayed a decreasing trend with increased soluble fibre intake (p<0.05).

Discussion
Previous studies on intake of food groups and the risk of colorectal cancer have had inconsistent findings (Pietinen et al., 1999;Michels et al., 2000;Sandhu et al., 2001;Tabatabaei et al., 2011). In our study, the intakes of grain products, meat, sea-food and fruit did not differ   significantly between cases and controls. We also did not find differences when the results were stratified by tumour site. The findings of our study suggest that intake of vegetables reduced the risk of colorectal cancer. Similar results were found in colon and rectal cancer. We observed that soy food could decrease the risk of colorectal cancer and rectal cancer.
The protective effect of vegetables observed in our study was consistent with some previous findings. A study conducted by Marta Banque et al. (2012) showed an inverse association between intake of vegetables and colorectal cancer. However, intake of fruit was not related to the risk of colorectal, colon and rectal cancer. A study in Japan also did not support the inverse association between fruit and colorectal cancer (Sato et al., 2005). Fruit may lack some beneficial factors, such as low energy and low sugar content (Franceschi et al., 1997). A study carried out in Shanghai showed that soy food reduced the risk of colorectal cancer in women (Yang et al., 2009), similar to our findings. It may be that soy foods are rich in folic acid, calcium, isoflavones and so on. These nutrients have protective effects against colorectal cancer. Previous studies have shown that these nutrients can maintain DNA synthesis and methylation (Mason et al., 1996), modulate signal pathways in malignant transformation (Li et al., 2005), inhibit cell growth and induce apoptosis (Bennink, 2001). Foods that lack these nutrients are related to an increased risk of colorectal cancer (Potter, 1999). Although some studies have shown that sea-food intake may protect against colon cancer (Franceschi et al., 1997;Kato et al., 1997), our study did not find an effect on colon or rectal cancer. However, we cannot rule out all confounding factors, including recall bias. In western countries, some studies have shown that high intake of meat may increase the risk of colorectal cancer (Larsson et al., 2005;Norat et al., 2005). However, some studies in Japan did not show clear associations between colorectal cancer, colon cancer and rectal cancer (Kojima et al., 2004;Sato et al., 2006). Similar to studies in Japan, the mean intake of meat did not differ between cases and controls.
In our study, we observed inverse associations between total fibre and vegetable fibre intake and the risk of colorectal cancer. The protective effect of fibre could be due to various mechanisms. (i) In the large bowel, fibre can increase stool bulk, dilute faecal carcinogens, shorten faecal transit time and reduce the contact of carcinogens with the colon epithelium (Bingham, 1990;Schatzkin, 2000). (ii) Fibre can bind to the bile acid, which can produce carcinogens (Wakai et al., 2007). (iii) Fibre can be fermented by gut flora to short-chain fatty acids. Butyrate is a one of the main short-chain fatty acids in the human colon, which has the ability to inhibit carcinogenesis (Goncalves et al., 2013). (iv) Fibre can reduce hyperinsulinemia by delaying the starch absorption (Hawk et al., 2002). Moreover, one study has shown that hyperinsulinemia is related to colorectal cancer (Ma et al., 2004). Inverse associations were observed between soluble fibre and insoluble fibre and colon cancer but not rectal cancer. The results of a Japanese collaborative study are consistent with our findings (Wakai et al., 2007). The reasons may be that rectum is empty in most of the time, thus reducing the effect of fibre (McNeil et al., 1981) or because the fermentable rates of fibre and bacterial growth are highest in the colon (Mai et al., 2003). However, one study showed that insoluble fibre had a strong inverse relationship with rectal cancer (Slattery et al., 2004). The reason may be that it is not possible to absolutely separate the effect of the two fibres in an observational study.
In terms of fibre from different food sources, we find an inverse association between vegetable fibre and the risk of colon or rectal cancer, similar to the result of a study conducted in Hawaii and Los Angeles (Nomura et al., 2007). However, this study (Nomura et al., 2007) reported an inverse association between fruit fibre and colorectal cancer risk (Q5VSQ1: OR=0.59, 95%CI: 0.40-0.71). This discrepancy in findings may be due to the different types of fruits people eat in Qingdao, as different fruit fibres may have different effects on colorectal cancer. Therefore, the different proportion of fruit fibre may lead to these observed differences. Soy food intake was inversely related to colorectal cancer and rectal cancer, but we did not find inverse associations between soy fibre intake and the risk of colorectal cancer. Another Chinese study (Zhong et al., 2014) reported that soy fibre did not reduce the risk of colorectal cancer and suggested that some nutrients contained in soy food have protective effects against colorectal cancer. However, another study found that blue lupin kernel fibre could improve colonic function and had beneficial effects on decreasing the risk of colorectal cancer (Fechner et al., 2013). Grain fibre did not show an inverse association with colorectal cancer in our study, and a prospective study of 88757 women also did not report a significant relationship between grain fibre and the risk of colorectal cancer (Fuchs et al., 1999). However, another cohort study reported a decrease in the risk of colorectal cancer associated with the intake of grain fibre (Schatzkin et al., 2007).
In conclusion, the present case-control study suggests that vegetables and fibre (including vegetable fibre, soluble fibre, and insoluble fibre) are associated with a decreased risk of colorectal cancer and colon cancer. Intake of vegetables and soy foods was inversely related to rectal cancer risk. More in-depth studies are needed to confirm the relationship between fruit and colorectal cancer stratified by gender.