Why You Should Take Studies Linking Certain Foods to Cancer with a Grain of Salt

We've all seen this scenario play out: a study is published suggesting that eating [insert food name here] increases your risk of developing cancer.  Then a few months later another study comes out, probably in just as reputable a journal, saying that the evidence suggests no such link. The same thing happens with studies claiming that eating a particular food reduces cancer risks.

So what should we believe?

According to a new study that examined 35 years' worth of studies like those above, we should think twice before believing most of them.

Researchers from the Stanford Prevention Research Center and Harvard Medical School selected the first 50 ingredients they found in randomly chosen cookbook recipes, including meats, fish, vegetables, dairy products, bread and spices.

The researchers then ran each ingredient through a medical journal database to find studies linking how much of the ingredient people consumed to their risk for some type of cancer. For 40 out of the 50 ingredients there were a total of 264 such studies; 103 studies suggested the ingredient was tied to an increased risk of cancer, and 88 to a decreased risk.

The average effect shown in each study was roughly a doubling or a halving of cancer risk for any given ingredient.  The usual suspects like fatty meats and sugar were in the "doubling cancer risk" category, and foods including onions, celery and green tea were in the "halving cancer risk" bucket.

The problem, the researchers found, is that the evidence presented in any one study was usually weak. In larger reviews that included multiple studies, the links between particular foods and cancer risk were typically much smaller or nonexistent.

Dr. John Ioannidis from the Stanford Prevention Research Center in California, who worked on the analysis, told Reuters Health:"We have seen a very large number of studies, just too many studies, suggesting that they had identified associations with specific food ingredients with cancer risk. People get scared or they think that they should change their lives and make big decisions, and then things get refuted very quickly."

This does not mean, however, that eating a lot of certain kinds of foods won't elevate or decrease your cancer risk -- it means that any one study claiming such a link is probably not telling the whole story.  Comprehensive reviews of findings over time present a more accurate story, but that's not usually how science is reported. Doing so requires researchers, like those who conducted this study, to sift through the clutter and find convincing trends, if they exist.

More of this "meta analysis" of studies is needed for several reasons, not the least of which is to prevent the public from becoming jaded about the over-hyping of individual, headline grabbing studies.

The study, "Is everything we eat associated with cancer? A systematic cookbook review" was published in the American Journal of Clinical Nutrition.

H/T: Reuters Health

Posted on January 8, 2013 .