Fiber
We usually get our daily fiber from fruit, vegetable and grain cereals. Dating back to 1970s, Burkitt who worked in Africa in 1950s suggested that high fiber diet protected from CRC based on his note of the low incidence of CRC among African people who consumed large amount of high fiber food (Burkitt, 1971). This was confirmed by other studies (Nomura et al., 2007; McCullough et al., 2003a). In 1989, Heilbrun et al. (1989) found a significant negative association of dietary fiber and colon cancer risk among low fat intake men. In this subgroup, the men consuming <7.5 g/d of dietary fiber had an adjusted relative risk for colon cancer of 2.28 (95% CI 0.93–5.60), compared to those consuming ≥14.8 g/d of dietary fiber. Steinmetz et al. (1994) also observed the inverse associations of intakes of all vegetables and dietary fiber, and colon cancer in a prospective cohort study of 41,837 women aged 55–69. In 2003, another observational EPIC study (Bingham et al., 2003) discovered that dietary fiber in foods was inversely related to incidence of large bowel cancer (adjusted relative risk 0.75 [95% CI 0.59–0.95] for the highest versus lowest quintile of intake), the protective effect being greatest for the left colon, and least for the rectum. However, no significant difference among various fiber sources on protection against CRC was observed. In 2011, with longer follow-up of EPIC study, the protective effect of dietary fiber was strengthened, and indicated that total dietary fiber was still inversely associated with colorectal cancer after a mean follow-up of 11 years (Dahm, 2011). In another large prospective cohort study that consisted of 291,988 men and 197,623 women (Schatzkin et al., 2007), although the protective effect of total dietary fiber intake was not found, whole-grain consumption was associated with a modest reduced risk of developing CRC. In a nested case-control study using food diaries, Dahm et al. (2010) discovered that intakes of absolute fiber and fiber intake density were significantly inversely associated with the risk of colorectal cancer in both age-adjusted models and multivariable models that adjusted for age, anthropomorphic and socioeconomic factors, and dietary intakes of folate, alcohol, and energy. In 2012, similar conclusion was made by a prospective Scandinavian HELGA cohort study indicating that intake of dietary fiber, especially from cereal foods, was associated with lower incidence of colon cancer (Hansen et al., 2012). The mechanisms behind these findings may include the followings. Firstly, fiber reduces concentrations of intestinal carcinogens due to the increased stool mass, increased bowel motility, and increased bacterial fermentation of resistant starch to SCFAs especially the butyrate, which keeps the health of colonocytes and enhances apoptosis and inhibits cancer cell proliferation (Burkitt, 1971; Bergman, 1990; Hamer et al., 2008). Secondly, long term fiber dominant diet may shift the gut microbiota with increased density of Firmicutes, which may have impact on host immune response and function as immune modulatory and anti-inflammatory effects (Fung et al., 2012; Neish, 2009).
Although there are abundant evidence supporting that fiber is beneficial to colon cells and could reduce colorectal cancer risk, there are other voices against it. Several cohort studies did not reveal the benefit of fiber in reducing CRC risk (Butler et al., 2008; Mai et al., 2003; Sanjoaquin et al., 2004; Michels et al., 2005; Otani et al., 2006; Kabat et al., 2008). In a population-based prospective cohort study conducted among Chinese women in urban Shanghai, no obvious association between fiber intake and CRC risk was observed (Shin et al., 2006). Similarly, two cohort studies conducted in Northern European population concluded that intake of dietary fiber was not associated with decreased risk of CRC (Pietinen et al., 1999; Gaard et al., 1996). Neither did the New York university women’s health study (Kato et al., 1997). The main reasons behind the discrepancy on this issue may include the followings. Firstly, there may be too many confounding factors like lifestyle, physical activity, intake of Vitamin D and so on involved in these cohort studies. These are the factors that may substantially influence the inverse association between fiber and CRC. Secondly, the intake of fiber might not reach the threshold level to produce high level of butyrate to protect against the inflammation, neoplasm and environmental carcinogens. Hence, more cohort studies with high level of fiber intake intervention are warranted in the near future.
Fat
In 1973, Drasar and Irving found that the cancer of colon was highly correlated with fat (Drasar and Irving, 1973). This was the first time that the association between fat intake and CRC was established. Since then, the hypothesis was tested by many investigators via epidemiological studies and animal experiments. Several animal studies indicated that high fat diet could stimulate bile acids secretion and cause the regression of epithelium, damage colonic mucosa, and increase the risk of developing CRC (Chomchai et al., 1974; Narisawa et al., 1974; Rijnkels et al., 1997). However, human prospective cohort studies (Pietinen et al., 1999; Gaard et al., 1996; Kato et al., 1997; Stemmermann et al., 1984; Giovannucci et al., 1994; Bostick et al., 1994; Goldbohm et al., 1994; Chyou et al., 1996; Terry et al., 2001; Jarvinen et al., 2001b; Flood et al., 2003; Brink et al., 2004; Lin et al., 2004; Oba et al., 2006; Weijenberg et al., 2007; Butler et al., 2009) did not reveal the equivalent findings except for one study (Willett et al., 1990). In 1990, a conclusion was obtained from the nurses’ health study that a high intake of fat especially animal saturated and monounsaturated fat rather than vegetable fat, linoleic acid or cholesterol, increased the risk of developing colon cancer (Willett et al., 1990). In contrast, a significantly negative association between colon cancer and the intake of saturated fat, whether assessed on the basis of g per day or as a percentage of the caloric intake, was observed from a cohort study investigated 7,074 middle aged men of Japanese ancestry living in Hawaii during a 15-year follow-up (Pietinen et al., 1999). In 1994, Giovannucci et al. (1994) reported that intakes of total fat, saturated fat and animal fat were not related to risk of colon cancer. However, an elevated risk of colon cancer was associated with red meat intake (relative risk (RR), 1.71; 95% CI 1.15–2.55 between high and low quintiles; P = 0.005 for trend) (Giovannucci et al., 1994). Similarly, Goldbohm et al. (1994) found out that the increased risk of colon cancer was mainly contributed by processed meat consumption especially sausages rather than dietary fat intake. In 1996, other two prospective studies indicated no association between dietary fat and CRC risk was observed but with positive findings for age, alcohol intake, pack-years of cigarette smoking and sausages consumption (Gaard et al., 1996; Chyou et al., 1996). The New York University women’s study (Kato et al., 1997) also showed that there was no overall positive or inverse association of colorectal cancer risk with intakes of total or subclasses fat, whereas there was an inverse association with total protein. Neither was the intakes of total calories, carbohydrate, or dietary fiber (Kato et al., 1997). In 2001, Terry et al. (2001) examined total fat, fat types and specific fatty acids in relation to colorectal cancer risk in a population-based prospective cohort of 61,463 Swedish women. The result indicated that intake of total, saturated, monounsaturated and polyunsaturated fats were not associated with colorectal cancer at any sites, no matter proximal or distal colon. The ratio of saturated fat:polyunsaturated fat, docosahexaenoic acid (DHA)+ eicosapentaenoic acid (EPA):linoleic acid, ω-3:ω-6 fatty acids showed similar lack of association respectively (Terry et al., 2001). Similarly, Brink et al. (2004) found that the intake of total, saturated and monounsaturated fat was not significantly associated with CRC in a Netherlands cohort study. However, the difference from the Swedish women cohort was that high intake of dietary polyunsaturated fat and, specifically, linoleic acid was associated with an increased risk of K-RAS mutated colon cancer and it was also associated with G>A transitions and G>T or G>C transversions in the colon (Brink et al., 2004). Later, in 2007, after 7.3 years of follow-up of the Netherlands cohort study, linoleic acid, the most abundant polyunsaturated fatty acid in the diet, was still associated with increased risk of colon tumors with only a KRAS mutation and no additional truncating Apc mutation or lack of MLH1 expression (Weijenberg et al., 2007). The association of fat intake and risk of CRC was also investigated in Asian population. In a prospective study (Butler et al., 2009) composed of 62,321 Singapore Chinese, a dose-dependent, positive association between saturated fat and localized colorectal cancer (Dukes A or B) was detected among women. Marine n-3 polyunsaturated fatty acid (PUFA) intake was associated with increased risk of advanced disease (Dukes C or D), in both men and women (Butler et al., 2009). However, inconsistent conclusion was drawn among Caucasian. Song et al. (2015, 2016) reported that high marine ω-3 PUFA intake was inversely associated with risk of colorectal cancer with high-level, but not low-level, FOXP3+ T-cell density, and microsatellite instability (MSI)-high CRC but not microsatellite stability (MSS) tumors in the nurses’ health study and health professionals follow-up cohort, suggesting a potential role of ω-3 PUFAs in cancer immunoprevention through modulation of regulatory T cells and DNA mismatch repair. In 2017, they reported that high marine ω-3 PUFA intake after CRC diagnosis is associated with lower risk of CRC-specific mortality, indicating that increasing consumption of marine ω-3 PUFAs after diagnosis may confer additional benefits to patients with CRC (Song et al., 2017). Thus, we could conclude that human studies on association of ω-3 PUFA and CRC risk are with inconsistent results and more studies are warranted to draw a definitive conclusion.
A meta-analysis performed in 1997 based on thirteen case-control studies showed the absence of impact of total dietary fat on the risk of CRC (Howe et al., 1997). Considering the potential recall and selection bias, prospective cohort studies would be more reliable than case-control studies to examine the influence of dietary fat on the risk of CRC. In 2011, a meta-analysis based on 13 prospective cohort studies that investigated the association between dietary fat intake and risk of CRC suggested that dietary fat intake might not be associated with the increased risk of CRC (Liu et al., 2011). Taking all the abundant evidence together, a definitive conclusion still could not be made, and thus, more evidence on association of dietary fat or specific types of fat and risk of CRC are needed to clarify the conclusion on this issue.
Protein
On the basis of global epidemiological and scientific studies, evidence suggests that the risk of CRC is increased by meat consumption (O’Keefe, 2016). The two recent systematic reviews and meta-analysis published in 2011 analyzed available prospective studies of red meat and CRC, and suggested that a high intake of red and processed meat significantly increased the risk of CRC (Alexander et al., 2011; Chan, 2011). A meta-analysis (Chan, 2011) of observational studies on red and processed meat intake and risk of colorectal adenomas was conducted and found that the summary RR of colorectal cancer for the highest versus the lowest intake was 1.22 (95% CI= 1.11–1.34) and the RR for every 100 g/day increase was 1.14 (95% CI= 1.04–1.24). These indicated that high intake of red and processed meat was associated with significantly increased risk of CRC (Chan, 2011). Based on numerous epidemiological and animal studies, the International Agency for Research on Cancer (IARC) concluded that the evidence on the carcinogenicity of red and processed meat was sufficient to classify processed meat as “carcinogenic to humans” (Group 1) and red meat as “probably carcinogenic to humans” (Group 2A) in 2015 (Domingo and Nadal, 2016).
The possible mechanisms behind the positive findings on red, processed meat and CRC may include the followings. Firstly, mutagens/carcinogens like heterocyclic amines (HCA) and polycyclic aromatic hydrocarbons (PAH) are produced during the excessive cooking procedure. HCA was shown to be highly mutagenic both in vitro and in vivo experiments (Nagao, 1999; Okochi et al., 1999; Okonogi et al., 1997; Nagao et al., 1997; Burnouf et al., 2001). Rodents were found to develop multiple cancer lesions with alterations in genes including Apc, β-actinand Ha-ras, in different organs including breast, prostate, colon and liver, when fed with diet added with HCAs, which cause carcinogenic effect by producing DNA adducts through the formation of N-C bonds at guanine bases (Sugimura et al., 2004; Matsukura et al., 1981; Ohgaki et al., 1984a, b, 1986, 1987; Fujita et al., 1999; Esumi et al., 1989; Takayama et al. 1984a, b, 1985a, b; Tamano et al., 1994; Kato et al., 1988, 1989; Ito et al., 1991; Shirai et al., 1997; Ochiai et al., 2002). PAHs are metabolized to cause DNA damage through covalently binding to DNA (Phillips, 1983; Phillips and Grover, 1994). Thus, if the DNA damage is not repaired correctly, mutations will be induced and may initiate the development of CRC. However, the evidence of epidemiological studies investigating the interaction between red meat and HAC, PAH intake, and development of CRC was still inconsistent (Sinha et al., 2005a, b; Gunter et al., 2005; Tabatabaei et al., 2010; Cross et al., 2010; Ferrucci et al., 2009; Shin et al., 2007; Burnett-Hartman et al., 2011). Secondly, high consumption of heme iron in red meat may lead to the formation of carcinogenic N-nitroso compounds and lipid peroxidation, which both contribute to the development of CRC. Four large prospective studies showed that a high intake of heme iron was associated with a higher risk of CRC (Cross et al., 2010; Lee et al., 2004; Larsson et al., 2005; Ferrucci et al., 2012). However, other two studies failed to establish the association between heme intake and risk of CRC (Kabat et al., 2007; Zhang et al., 2011). It might be explained by the different concentration of heme iron in different red meats and then, the risk of CRC varies (Egeberg et al., 2013).
Despite the consensus on the association of red and processed meat and risk of CRC, no positive association has been presented regarding to various sources of protein except for red and processed meat. A systemic review and meta-analysis (Carr et al., 2016) analyzed meat subtypes and their association with CRC, and reported that pork intake had no overall association with CRC risk, neither in meta-analysis of cohort nor of case-control studies. Poultry consumption was consistently not associated with increased risk of CRC, its sub sites or its precursors (Carr et al., 2016). In a meta-analysis published in 2012, Wu et al. (2012) revealed that fish consumption was inversely associated with colorectal cancer. Furthermore, in 2014, another meta-analysis investigated the association of fish intake and risk of gastrointestinal (GI) cancers, and reduced risk was observed in CRC, esophageal cancer and hepatocellular cancers, suggesting that fish consumption might reduce total GI cancer incidence (Yu et al., 2014). In terms of dairy products, although several epidemiological studies indicate that milk intake might play a role in protecting against CRC due to the high level of calcium in dairy products, which is consistently agreed, there are still other studies reporting that no obvious association was found (Giovannucci et al., 1994; Kesse et al., 2005; Bostick et al., 1993; McCullough et al., 2003b; Lin et al., 2005; Jarvinen et al., 2001a; Kampman et al., 1994; Cho et al., 2004; Terry et al., 2002; Kearney et al., 1996). In 2014, a meta-analysis (Ralston et al., 2014) indicated an overall inverse risk of colon cancer of 0.74 (95% CI 0.60–0.91) in men who consumed non-fermented milk. No association was found of consumption of non-fermented milk and rectal cancer in men or non-fermented milk and CRC in women. No protective positive or negative association was found between intake of solid cheese or fermented milk and CRC (Ralston et al., 2014).
In summary, there is convincing evidence that red meat or processed meat increases risk of CRC. The evidence for poultry and fish is inconsistent while milk consumption may protect against CRC (Table
1).
Table 1Dietary factors that influence CRC risk