Flavour of the Week

Are All Foods Linked to Cancer – What to Believe?

With articles like those above, one begins to wonder… is there a food out there that *hasn’t* been linked to cancer? Turns out you wouldn’t be the only person asking this question! A recent article published in the American Journal of Clinical Nutrition (1) addresses the question:

“Is everything we eat associated with cancer?”

“Nutritional epidemiology” includes studies seeking to identify connections between dietary factors and health outcomes. There are thousands of these studies published each year and a significant number of them look at the connection between diet and cancer. Often these studies get picked up by the media and either misinterpreted or the connection is blown out of proportion as compared to what the actual statistics suggest. This is no surprise as cancer is a hot topic, likely due to the fact that it is now the leading cause of death in Canada (2).

Ideally, a randomized clinical trial (RCT), whereby a dietary factor is tested for its cancer-promoting or cancer-reducing effects, would provide much better evidence than simply an association (see our previous post on Correlation vs. Causation). However, these trials are very challenging to conduct in humans, and would be considered unethical in many cases. In the instances where these randomized trials have been undertaken, the results are often equivocal – some show no change despite previous epidemiological studies showing strong associations (3); in some cases an increase in cancer risk may be found in an RCT where a decrease was expected (4).

Jonathan Schoenfeld and John Ioannidis (1) randomly selected 50 ingredients using The Boston Cooking-School Cook Book (5), which was originally published in 1896, to determine what percentage of the ingredients had since been linked to cancer in the scientific literature. The ingredients included meats, spices, grains, coffee, wine, potatoes, vegetables, milk, and fruits. Of the 50 ingredients selected, 80% had at least one study that looked at the specific food ingredient and its relation to cancer risk. They also noted the few ingredients that currently have no clinical data linking them to cancer risk are those less common in the diet (such as bay leaf, thyme, vanilla, cloves, hickory, etc.); many of these less-common foods, however, have been investigated with respect to cancer risk in animal studies (1).

Of the ingredient-related studies looking at cancer risk, Schoenfeld and Ioannidis took a closer look at the studies’ results and found:

  • 39% found that an ingredient was linked to an increased risk for malignancy
  • 33% showed a decreased risk with higher intakes of the ingredient
  • 5% found a borderline effect
  • Only 23% found no statistical evidence to of either an increased or decreased risk

Importantly, however, the effect was weak or only nominally significant in 75% of the studies that found an increased risk and in 76% of studies that found a decreased risk (1) – translation: < 20% of studies showed a strong link between an ingredient and cancer risk, whether an increased or decreased risk. What does this mean in layman’s terms? For the majority of food ingredients in our diet, it is possible to find a study linking them to cancer risk BUT in reality the overall effect in most studies is very small.

In addition to overblown results, it has also been suggested that the dissemination of scientific evidence may be biased towards positive findings (including those “false positives”); it has been demonstrated that “negative” results are much less likely to get published and publicized in the media (6-7) – negative results may not be exciting, but they are still valid! As a result, study findings that are *just barely significant* may be widely spread and overemphasized (1). Could this be the case in studies that link diet to cancer?

What can we do to increase the quality of information and the risk assessment? Well, Schoenfeld and Ioannidis did find that meta-analyses (studies that combine the results of several different studies on the same topic), were less likely to show an effect and may be more accurate. Interestingly, a link between a specific food ingredient and cancer became less noticeable as more data was collected (1).

No Baloney’s advice. Don’t get too excited (or worried!) when you read the latest study connecting a particular food ingredient to cancer. Wait for supporting scientific evidence and check out the original research to evaluate the actual size of the correlation or relative risk. Over-inflated media reports risk demonizing foods that may not be unhealthy or valuing foods that may not actually have a significant health benefit.

There are a few points on diet and cancer that have come up time and again and may be worth taking note of:

  1. Fruits and vegetables are a good bet for lowering your risk of cancer and a variety of other chronic diseases and there is really no downside to getting your recommended number of servings.
  2. Processed meats have been linked to an increased risk and really offer no nutritional upside; we’d recommend omitting them from your diet or at least limiting your intake.

If you are feeling overwhelmed, keep a sense of proportion and don’t let the “everything is linked to cancer” argument discourage your healthy eating plan.


  1. Schoenfeld JD, Ioannidis JPA. Is everything we eat associated with cancer? A systematic cookbook review. Am J Clin Nutr 2013; 97:127-34.
  2. Statistics Canada. Leading causes of death; 2012. Retrieved January 15, 2013 from http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/hlth36a-eng.htm
  3. Gaziano JM, et al. Vitamins E and C in the prevention of prostate and total cancer in men: the Physicians’ Health Study II randomized controlled trial. JAMA 2009; 301:52-62.
  4. Klein EA, et al. Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 2011; 306:1549-56.
  5. Farmer FM. The Boston cooking-school cookbook. Boston, MA: Little, Brown and Company, 1896.
  6. Easterbrook PJ et al. Publication bias in clinical research. Lancet 1991; 337:867-72.
  7. Kyzas PA, et al. Almost all articles on cancer prognostic markers report statistically significant results. Eur J Cancer 2007; 43:2559-79.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s