Story behind the study
In the Summer of 2020, I had an interesting discussion with Dr Mazda Jenab (OncoMetabolomics Team, IARC-WHO) on the unresolved mechanisms explaining the relationships between high consumption of red and processed meats and colorectal cancer risk. We particularly discussed the suggested role of iron contained in various types of meats, which is still unsettled, as is the issue of possible sex-differences in the potential iron-cancer relationship. During our discussion, we considered that the unique scientific resources of the European Prospective Investigation into Cancer and Nutrition (EPIC) study (coordinated jointly by IARC and Imperial College London; https://epic.iarc.fr) could contribute additional robust data towards our questions surrounding the tumorigenic roles of this essential nutrient. We therefore decided to explore the colorectal cancer risk relationship with intake levels of total iron as well as iron specifically from meats (heme iron) or plants (non-heme iron) in men and women participants of the EPIC cohort.
Scientific rationale for the study
Colorectal cancer is one of the most common cancers globally, with over 1.9 million individuals diagnosed in 2020. Over the past decades, our understanding of risk factors for colorectal cancer has increasingly expanded, and we now know that higher consumption of red and processed meats increases one’s chance to develop a tumor in the colorectum.
As mentioned earlier, heme iron contained in meats has been hypothesized as a possible factor that promotes colorectal cancer occurrence. The rationale being that heme iron is a potent pro-oxidative agent and catalyzes the production of DNA-damaging free radicals. However, evidence from epidemiological studies in humans is unclear. In contrast to heme iron, non-heme iron contained in plant products, dairy and eggs is chemically different and doesn’t fuel the production of free radicals.
We decided to examine the associations between total dietary iron intake, and particularly heme and non-heme iron and the risk of colorectal cancer. We specifically conducted our analyses separately in men and women because they have different iron needs, intake patterns, absorption rates, turnover, and excretion in the body which could possibly lead to cancer risk differences.
How did we approach the question?
We used the data from EPIC, a large prospective cohort study with data from 10 European countries. Our data included 450,105 participants (318,680 women), recruited in 1990s and followed for over 14 years. At the recruitment of the participants, a panoply of information including anthropometry, lifestyle, and usual diet were collected. During the follow-up, 6,162 participants (3,511 women) developed colorectal cancer. Colorectal cancer is any malignant tumor that originates in the colon, i.e., from the cecum to the rectum.
To estimate total iron intake, we used iron content of all the food items consumed by an individual multiplied by the portions consumed. For the daily heme iron intake of each participant, we applied conversion factors to their amounts of meat consumed i.e., for beef, lamb, pork, processed meat, poultry, fish. We then extracted heme iron from total iron, to estimate daily non-heme iron intake.
We conducted statistical analyses using a type of time-event associated analysis called Cox regression models. We conducted multivariable-adjusted analysis which means that we excluded the effects of previously known risk factors for colorectal cancer such as smoking habits, alcohol intake, or socio-economic variables, in addition to factors that impede or enhance iron absorption and utilization, such as vitamin C, calcium, tea, or coffee. We additionally conducted a specific type of analysis called dose-response analysis to investigate whether the relationship between iron and colorectal cancer is linear, as well as a substitution analysis with the aim to model the risk of colorectal cancer if one decides to replace some heme iron with non-heme iron while maintaining the same level of total iron consumed.
What did we find?
In men, we did not observe any relationship between intake of total iron and colorectal cancer risk. However, higher intake of non-heme iron was associated with 20% lower colorectal cancer risk whereas no risk association was observed for heme iron. In contrast, in women we did not find any association between total, heme, or non-heme iron on colorectal cancer risk.
When we explored whether iron may be associated differently for colorectal tumors according to their location, we found that in men heme iron was associated with a higher risk of tumors that occurred in the rectum and the proximal colon (portion of the colon that ascends or is transverse) compared to the distal colon (colon that descends towards the rectum). There was no clear explanation of why the associations observed in men were stronger in the rectum and proximal colon. This difference according to anatomical location was not observed in women.
Our substitution models showed that, by replacing daily intake of 60 mg of cooked beef (1 mg of heme iron) with 30 g of boiled beans or two boiled eggs (1 mg of non-heme iron), colorectal cancer risk is reduced by 6 % in men.
Putting our findings in context
Iron deficiency is a public health issue in several countries. Our analysis focused on a European population where iron intake is relatively high. Therefore, our findings may not be generalizable, especially in settings where iron intake is low, and even more, with lower absorption due to lifestyle choices such as drinking tea during food time, repeated infections, or low intakes of vitamin C.
We found that dietary non-heme iron reduces colorectal cancer risk in men whereas heme iron intake showed a positive association restricted to the rectum and the proximal colon. Dietary iron, irrespective of the food sources, does not seem to influence colorectal cancer risk in women, possibly because of overall lower iron intake in women, or reproductive and hormonal factors. Our observations call for more studies, especially in other populations, exploring the mechanistic explanation of tumor site-specific associations.
Our study adds an additional layer to our general understanding of the complex relationship between dietary iron, and its heme and non-heme components and colorectal cancer risk in men and women.