Dietary patterns and risk of colorectal cancer in Tehran Province: a case–control study

In 2008, colorectal cancer (CRC) was estimated to be the third and second most commonly diagnosed cancer in the world with 10.0% and 9.4% of total cancer cases among males and females, respectively [1, 2]. Although the rates of CRC in Africa and Asia are lower as compared to CRC rates reported in Europe, North America, New Zealand and Australia [3], the former tend to increase with migration and westernization. In Asia, higher rates of CRC are observed in Japan and China [4]. Among Iranians, CRC is the third and fourth most commonly diagnosed malignancy in males and females, respectively. In the last few decades, the incidence and mortality rates of CRC have increased markedly in Iran [5, 6].

Family history, unfavourable socioeconomic conditions, tobacco smoking, alcohol consumption, obesity, physical inactivity, cooking techniques and various aspects of diet have been associated with the risk of CRC [7, 8]. A number of studies have assessed the relationship between diet and risk of CRC, but they mainly focused on the intake of single food items or nutrients [9, 10]. Although such analyses are valuable, they give inadequate information on diet and disease relationship. As people consume meals that consist of a complex combination of variety of nutrients and non-nutrients that are interactive or synergistic with each other, it is difficult to isolate the effect of a single nutrient or food on health [11]. Examination of patterns of dietary intake (multiple dietary components) rather than focusing on the individual nutrients is recommended as a complementary and suitable approach for clarifying relationships between diet and health [11, 12]. Most of the existing data on cancers and dietary patterns were collected from developed countries, with almost two-thirds of these surveys conducted in either North America or Europe [13]. There is a lack of published data on this topic in Iran as a developing country. The aim of this study was to identify dietary patterns and its association with the risk of CRC in Tehran, Iran.

Table 2 shows the socio -demographic and lifestyle characteristics of the71 cases and 142 controls. By frequency matched design, age and gender were similar in controls and cases. The mean age of respondents at diagnosis of colorectal cancer (CRC) was 59.9 years for men and 55.6 years for women.There were no statistically significant differences between case and control groups by educational level, occupation, income, physical activity (MET. h/day), cooking methodsor smoking status. However, cases were more likely than controls to have a family history of CRC in the first (p=0.004) and second degree relatives (p=0.003). Of the four main types of non-steroidal anti-inflammatory drugs (NSAIDs) which were commonly used by respondents, only aspirin (p= 0.012) and acetaminophen (p= 0.049) were significantly different between case and control groups. Compared to the cases, controls were more likely to take mineral supplements (25.3% versus 11.3%; p=0.017). The mean body mass index (BMI) was slightly higher in the case than in the control group, but it did not differ statistically. Cases had significantly higher energy intake than the controls (p=0.045).

Factor analysis revealed two dietary patterns and the factor loadings for each dietary pattern are presented in Table 3. Food groups with absolute factor loadings > 0.20 were considered as having significant contribution to the pattern. These two dietary patterns explained 22.20% of the total variance in food intake. The first pattern with high loadings for fruits, vegetables, liquid oil, olive, fish, yoghurt drink, whole grains, carrot, low-fat dairy products and nuts was labelled “ Healthy “ dietary pattern. This pattern was significantly related to respondents who were non-married, had diploma or higher degree, self-employed, with higher income, use acetaminophen, vitamins and commonly consume raw/ fresh vege. The second pattern which loaded heavily on sugar, processed and red meat, animal butter, refined cereals, tea, pickles, solid oil, mayonnaise, soft drink, legumes, sweets and desserts, and snacks was named “Western” dietary pattern. This pattern was related to being married, current smoker and not using vitamins and aspirin and with high energy intake.

The odds ratio and their 95% confidence interval for CRC by the mean of dietary pattern scores are presented in Table 4. After adjusting for family history of CRC in first and second-degree relative, energy intake, cooking of vegetables, aspirin, acetaminophen and mineral supplement use, the Healthy dietary pattern was associated with a decreased risk of colorectal cancer. Those having “high” healthy dietary pattern (high consumption of healthy foods) had lower risk of CRC (OR= 0.227, 95% CI=0.108–0.478). An increased risk of colorectal cancer was associated with “high” Western dietary pattern in that those consuming more western foods had higher risk of CRC (OR=2.616; 95% CI =1.361–5.030).

In this case–control study of Iranian adults in Tehran, Iran, a Western dietary pattern (WDP) was associated with a greater risk of CRC while a Healthy dietary pattern (HDP) conferred a protective effect against CRC risk after adjustment of socio-demographic and lifestyle factors. The findings suggest that the associations between these dietary patterns and risk of CRC are independent of other risk factors of CRC.

Although consuming certain food items such as solid oil, sweets and sugar that are representative of WDP might explain the relationship with CRC, its mechanism remains unknown. The most likely mechanism is through overweight and obesity, which are important risk factors for CRC [25]. In a cross-sectional study in Iran,the adoption of a Western pattern, which was characterized by a higher intake of butter, high-fat dairy products, sweets and desserts, hydrogenated fats, soft drinks, potatoes, pizza, red and processed meat was positively related to obesity [26]. Murtaughand co-workers [27] found the animal protein pattern was associated with a more than threefold increased risk of obesity, while the prudent or healthy dietary pattern showed a 29% decreased risk of obesity among Hispanic women.

The CRC and dietary pattern relationship observed in our study was comparable to those identified in previous studies. In a study among the US population [12], the Prudent pattern, characterized by a high consumption of vegetables, fruits, whole grains, poultry, fish, and legumes, was inversely associated with the risk of CRC while the Western pattern (high intake of sweets and desserts, red and processed meats, refined grains and french fries) was positively associated with this malignancy. Other studies have also reported that a dietary pattern characterized by low intake of fried foods, alcohol and processed meat or high intake of fruits, vegetables, yams, cereals, legumes and low fat dairy products lowered the risk of CRC [28‐30].

In the present study, fruits, vegetables and whole grains were related to HDP. These food groups have been shown to have a protective effect against CRC through various mechanisms [31, 32]. They are good sources of various micronutrients and other compounds including minerals, vitamins A, C, E, carotenoids, selenium, flavonoids, and fibres, which display overlapping and complementary mechanisms of action, such as binding and dilution of carcinogens and anti-oxidant effects [31, 33]. Higher intake of dietary fibre and whole grains may bind and dilute potential carcinogens in stool and speed up their transit through the intestines [32, 33]. These actions may consequently change the colonic flora to an environment that is protective against CRC.

It has been hypothesized that high fat intake, through its consequent high caloric intake and obesity, may potentially cause hormonal changes that can over motivate the re-generation of colonic endothelial cells and persuade the growth of polyps and adenomas [34, 35]. It may also promote CRC risk through its altered immunologic responses and insulin resistance [12, 34]. Several studies have reported that it is the source of fat as well as the type of fat that is associated with the occurrence of CRC [36, 37]. For example, the protective role of olives against CRC could be due to the type of polyunsaturated fatty acids (PUFAs) and the existence of some micro-nutrients like vitamin E, poly phenols and various anti-oxidant components [12, 38, 39]. In this study, consumption of olive oil and liquid oil was related to HDP while the consumption of solid oil and animal butter was related to WDP.

A number of studies [10, 40] showed that the intake of milk and milk products is negatively associated with CRC risk. The protective effect of milk and dairy products can be attributed to the existence of some micronutrients such as riboflavin, calcium, vitamin B12and vitamin D. This food group also contains other protective substances such as butyric acid, linoleic acid, sphingomyelin and probiotic that can reduce cancer risk [41]. The relationship between dairy products and CRC may also be influenced by the fat content. While high fat dairy products increased the risk of CRC [39, 42], low fat dairy products lowered the risk of CRC [43, 44]. Similar to other studies [39, 42], the present study found that high intake of low-fat dairy products contributed to HDP.

Fish consumption has been recognized in dietary patterns as a protective factor for CRC [12, 39]. This inverse association could be attributed to the high intake of n-3 polyunsaturated fatty acids and vitamins D and A from fish. In the present study, fish was a component in HDP. The traditional Iranian diet does not entail high consumption of fish due to the high prices and/or seasonal availability [45]. However, Tehran has a significantly larger fish consuming population compared to other provinces of Iran [46].

In some studies, tea consumption is a risk factor for CRC [47, 48] while in others [12, 39], an inverse association was observed between tea intake and CRC risk. The inconsistent finding of tea consumption might be related to constipation and the inverse association with secretion of bile acids [49]. In the present study, tea had a high factor loading for Western (unhealthy) dietary pattern. In Iran, black tea is the most popular drink that is commonly served with sugar or cube sugar. This study also showed that sugar group (e.g. sugar, sugar cube, candy, sugar candy, tahini) had the highest factor loading for WDP as compared to other food groups in this pattern.

Similar to other studies [38, 42] our study suggested that a high consumption of red or processed meat elevates the risk of CRC. A meta-analysis of 19 cohort studies also revealed a positive association between high intake of red meat with CRC risk [50], This effect can be attributed to the meat content of saturated fatty acids and heme iron as well as the formation of N-nitroso-compounds (NOCs), polycyclic aromatic hydrocarbons (PAHs), heterocyclic amines (HCAs) and acryl amide as a result of higher temperatures during cooking [7, 51]. The relationship between meat and cancer risk could depend on the cooking or preservation method of meats [51].

Surprisingly, we also found that intake of legumes was observed in WDP [37]. The findings on legumes and CRC risk are mixed with Fung et al., [12] reported that high intake of legumes in the prudent pattern decreased the risk of CRC while Kesse and colleagues [38] failed to show any significant relationship between legumes and CRC risk. At present, we do not understand the positive association between legumes and CRC risk but it may be due to the preparation method of legume dishes that could diminish the favorable effect of legumes. In Iran, important legumes such as lentil, chickpea, pea and different varieties of bean are commonly prepared with rice, meat and vegetables [52]. A popular traditional lentil soup served for breakfast is usually prepared with animal butter and fried onion. Red bean and chickpea are used in ghorme sabzie and ghymeh khoresht, respectively. In ghorme sabzie, fried vegetable (using solid oil) and red meat are common ingredients while in ghymeh khoresht, fried chickpea is mixed with red meat, fried onion and potatoes [52]. It should be noted that in this study, solid oil, animal butter and red meat groups had higher factor loadings in WDP.

This study has several limitations. First, the possibility of selection bias cannot be avoided in retrospective case–control studies. The present study minimized this problem by matching the cases and controls by age and sex. In addition, both groups were distributed similarly by hospital status [29]. There is also a possibility that individuals with colorectal cancer would recall their diets differently than controls as a result of their disease status [53]. As cases were selected not more than 6 months of cancer diagnosis and during the interview, cases were reminded to report foods consumed prior to diagnosis, these could minimize the possibility of cases reporting post-diagnosis. Also, our findings that Western dietary pattern was associated with increased risk of CRC and Healthy dietary pattern reduced the risk of CRC could further corroborate the non-difference in diet recall between cases and controls. Second, this study could not assess the long term effects of risk factors on the incidence of colorectal cancer. A cohort or longitudinal study is a better alternate for the development of chronic diseases such cancer and can determine the associations with socioeconomic, lifestyle behaviours and anthropometry. Third, the different stages of factor analysis utilized in this study included the grouping of food items, determining the number of factors, Eigenvalue, method of rotation and labelling of factors. These interpretations were subjective as they were based on the decision of the researcher guided by previous published methods of analysing dietary patterns. Finally, the sample size of the current study was relatively small and the study was only conducted on people living in Tehran city. These could limit the generalization of study findings to the entire Iranian population. Despite these limitations, this is the first study in Tehran province in the northeast of Iran to examine the relationship of major dietary patterns and colorectal cancer risk among middle age adults.

The present study suggested that a diet characterized by high consumption of fruits, vegetables, liquid oil, olive, carrot, fish, yoghurt drink, whole grains, low-fat dairy products and nuts (labelled as healthy diet) was associated with lower risk of colorectal cancer, while a diet with high intake of sugar, processed and red meat, animal butter, refined cereals, tea, pickles, solid oil, mayonnaise, soft drink, legumes, sweets and desserts and snacks might increase the risk of colorectal cancer. The study also supported the importance of using dietary pattern method to investigate the compound relationship between diet and colorectal cancer. This diet-disease relationship can be used for developing interventions that aim to promote healthy eating for the prevention of chronic diseases, particularly colorectal cancer in the Iranian population.