Not a clinic day goes by without multiple patients asking me what they should eat, both while on treatment for their cancer and during the survivorship period. If you Google diet and cancer, you are informed that there are 207 million results.¹ Such an association seems logical; we all grew up hearing the phase “you are what you eat,” a phrase attributed to French politician and gastronome Jean Anthelme Brillat-Savarin, who wrote in 1826, “Dis-moi ce que tu manges, je te dirai ce que tu es [Tell me what you eat and I will tell you what you are].”² Despite the persistent belief for the past nearly two centuries, there remain limited, consistent data on most dietary factors and many disease, including cancers.

Studies of red meat and processed meat have been a rare example of fairly consistent results showing an association between increased intake and risk of developing colorectal cancer.^3,4 In 2007, The World Cancer Fund and American Institute of Cancer Research‘s expert panel reported that there was convincing evidence that red meat and processed meat increased the risk of developing colorectal cancer.⁵ Their meta-analyses found that consumption of red meat led to a 1.43 (95% CI, 1.05 to 1.94) increased risk of colorectal cancer per times per week consumed and 1.29 (95% CI, 1.04 to 1.60) per 100 g/d. Similarly, consumption of processed meat was associated with a 1.21 (95% CI, 1.04 to 1.42) increased risk per 50 g/d.⁴ Given these consistent findings, it seems reasonable to test whether consumption of red or processed meat affected patients who already have a diagnosis of colorectal cancer.

In this issue of Journal of Clinical Oncology, McCullough et al⁶ report on a cohort of 2,315 subjects who developed colorectal cancer while participating in the Cancer Prevention Study-II Nutrition Cohort. They report that the quantity of red and processed red meat consumed before the diagnosis of colorectal cancer was associated with all-cause but not colorectal cancer–associated mortality. Though not statistically significant, the data suggest that this increase was due to cardiovascular-associated mortality. However, the consumption of red and processed meat after diagnosis was not associated with either end point. Curiously, those with the highest consumption of red and processed meat consistently before and after diagnosis did have an increased risk of colorectal cancer–associated mortality. This seemingly highest risk group (high intake before and after diagnosis) did not have a statistically higher risk of overall or cardiovascular-associated mortality.

Several issues are worthy of consideration in interpreting the McCullough study.⁶ First, why did an exposure that convincingly increases the risk of developing colorectal cancer not affect the natural history of the disease once it developed?^7,8 Although the exact mechanism of action for red and processed meat increasing colorectal cancer development is not known, several plausible biologic mechanisms have been proposed. Red and processed meats cooked at high temperatures contain heterocyclic amines, which are carcinogenic. A second mechanism involves endogenous formation in the gastrointestinal tract of N-nitroso compounds from the heme in red meat, many of which are carcinogenic. In addition, nitrites or nitrates added to meat for preservation could increase exogenous exposure to nitrosamines, N-nitroso compounds, and their precursors. All these proposed mechanisms lead to carcinogenic effects on the mucosa of the bowel, leading to mutational effects on those cells and potential for abnormal growth and cancer formation. The risk of recurrence for patients with nonmetastatic colorectal cancer is related to the growth of micrometastatic disease, already spread before detection and treatment of the primary lesion. Thus, local carcinogenic effects will be less significant to colorectal cancer survivors’ outside risk of forming new primary tumors. In contrast, recent studies on diet and colorectal cancer survivorship have demonstrated association with cancer recurrence, specifically as a result of high intakes of Western-pattern diet and glycemic load.^9,10 Both exposures are proposed to be associated with disease recurrence by increasing insulin and insulin-like growth factors, which affect cell growth, proliferation, and metastatic potential, thereby influencing the growth of micrometastatic disease in colorectal cancer survivors.¹¹

Another consideration in survivorship studies is what recommendations can be made to the patient at hand. When a patient is diagnosed with cancer, they want recommendations on what they can do now to help their chances of cure and/or survival. If an exposure before diagnosis is associated with an outcome, but not the exposure after diagnosis, there is not necessarily a recommendation that can be made to a patient. However, such data may suggest a factor that influenced the biology of the tumor that developed. Another question is whether more favorable or less favorable biology influenced by the exposure can be affected by what the patient does after diagnosis. One consideration could be whether a worse prognosis tumor as a result of dietary exposure should influence treatment given to the patient (ie, the factor having prognostic and predictive value). Certainly, studies of diet and lifestyle to date are far from being able to lead to such conclusions. However, it is possible such a mechanism might explain the curious finding that colorectal cancer–associated mortality was influenced by high consumption of red and processed meat both before and after diagnosis, but not in either time frame only. Specifically, patients with high consumption of red and processed meat before diagnosis should try to decrease intake after diagnosis because those who maintained high levels had a higher risk of recurrence. Such a conclusion is purely speculative on the basis of the data in this article and would need other cohorts to further clarify.

Finally, the study by McCullough et al suggests that some risk factors for colorectal cancer also increase risk for other diseases, and thus colorectal cancer patients will often have comorbidities that influence survival. Thus, the current study does remind clinicians that, although one cannot influence exposures before diagnosis, management of comorbidities is important in the care of colorectal cancer survivors to improve survival.

In conclusion, studies of host factors and cancer survival require us to consider whether the results can be of utility to our patients. First, if the exposure after diagnosis influences outcomes, one should consider whether the strength of the evidence justifies making recommendations to alter diet or lifestyle, for instance. Although a randomized controlled trial would be ideal to address this question, changing diet and lifestyle behaviors in the number of patients needed to have statistical powers remains a challenge. Because studies of diet and lifestyle in colorectal cancer survivors are all observational to date, one needs to consider potential biases and confounding. Second, it is important to understand whether an exposure affects cancer recurrence, survival, or both. Although both end points are important in survivorship care, they may have different management implications. Finally, although a message that prediagnosis diet influences outcomes may seem to have limited utility for a patient when they develop cancer, it furthers the strength of the recommendation for people to maintain a healthy diet and lifestyle throughout their life to maximize the health benefits.

Source: JCO