It’s no surprise that a paper published in the prestigious British Medical Journal on the link between food preservatives and cancer received a lot of media attention. After all, it’s not uncommon for food to contain preservatives, and cancer is scary. But how much should you furrow your eyebrows? It depends on whether you look at the headlines screaming that preservatives increase cancer rates by 16%, or dig deeper into the data that produced these numbers. Let’s dig in!
The data described in the paper comes from the French NutriNet-Sante study, which has followed around 105,000 people since 2009, including 4,226 people who were ultimately diagnosed with cancer. On three days every six months, participants completed a detailed dietary questionnaire to estimate the preservative content of the foods they consumed from a publicly available food composition database. Such calculations involve significant uncertainties. For example, different producers may use different amounts of sorbic acid in their jams, as well as other preservatives such as nitrites in ham and sulfites in wine.
People are also notoriously unreliable when it comes to accurately reporting the foods they consume, especially the amounts. And of course, while such observational studies may establish an “association,” they cannot prove causation. Although attempts can be made to correct for confounding factors such as smoking, activity level, and weight, there are other factors that can skew the results. For example, foods that contain preservatives may also contain dyes, emulsifiers, sweeteners, flavor enhancers, and chemicals such as acrylamide and polycyclic hydrocarbons that are formed during cooking. Their presence may obscure the relationship between preservatives and cancer.
Putting these uncertainties aside, let’s look at how the study authors came up with the 16% increase in overall cancers. The subjects were divided into three groups depending on the amount of preservatives they were believed to be consuming: “low consumers,” “medium consumers,” and “high consumers.” They then calculated the proportion of people in each group who were diagnosed with cancer to see if that was associated with preservative intake.
Of the 43,586 participants in the low consumer group, 1,288, or 3%, were diagnosed with cancer. In the higher group, it was 1,339 of 38,503, or 3.5%. So where does the 16% increase in cancer come from? 3.5 is 16% greater than 3. But what a 0.5% difference in cancer incidence actually means is that for every 200 people who switch from a diet high in preservatives to a diet low in preservatives, one case of cancer may be avoided. Similarly, for every specific cancer for which an increase was calculated, such as a 26% increase in breast cancer with potassium sorbate and a 32% increase in prostate cancer with sodium nitrite, between 0 and 3 cancer cases could be avoided if 100 people switched to a low-preservative diet. But are preservatives really the culprit, or could a diet high in preservatives simply be a marker of a diet high in fat, sugar, and salt?Such diets are associated with an increased risk of cancer.
However, the possibility is not free from the dangers of preservatives, as the combination of nitrites and benzoates can cause carcinogenic effects. Nitrites, which are widely used in processed meats, can be converted in the body to nitrosamines, which are recognized carcinogens. Under certain conditions, benzoates can release benzene, another known carcinogen. However, there is no independent evidence from laboratory or animal studies that other preservatives can cause cancer. That’s why the results of the NutriNet study, in which researchers linked 11 preservatives to cancer, are so interesting. That curiosity extended to their calculation that acetic acid, the main ingredient in vinegar, increased cancer rates by 12%. If acetic acid were truly carcinogenic, the epidemiological link between consumption of vinegar and pickled foods should have been established long ago.
Another issue with this publication is the reliability of the data as judged by the reported “confidence intervals”. This is an estimate of the range of values that includes the value of interest. For acetic acid, this value has been reported to be between 1.01 and 1.25, meaning the increased risk may be as little as 1% or as much as 25%. Such a large confidence interval means that the results should be discounted considerably.
Additionally, for many of the preservatives included in the study, the confidence intervals for the cancer association ranged from 0.91 to 1.10 for calcium propionate. This essentially renders the results meaningless, as calcium propionate intake can reduce cancer risk by as much as 9% or increase it by as much as 10%.
One final point is important. It is unrealistic to focus on risks without considering the benefits. Food producers don’t add preservatives on a whim. They add them to prevent disease. Nitrite prevents potentially fatal botulism, benzoate suppresses yeast and bacteria, sorbate prevents mold, fungi, and yeast, and propionate prevents mold from taking root in bread. It is impossible to know the exact effect of preservatives because the “diseases prevented” cannot be calculated. What is clear is that food poisoning has decreased dramatically since preservatives were introduced. No more worrying about your bread going moldy after a day or getting botulism from eating a hot dog.
So what can we take from this paper linking food preservatives to cancer? Reporting risks in percentage terms without mentioning absolute numbers is misleading, and ignoring the potential benefits of preservatives is negligent. The size of the confidence interval often casts doubt on the data.
On the positive side, this paper does not prove that preservatives cause cancer, but the data collected suggests that eating foods containing preservatives slightly increases cancer. Of course, the presence of preservatives may simply indicate a food high in salt, fat, or sugar. Nevertheless, the authors’ advice that public health policies should be strengthened to make fresh, seasonal, home-grown products available and affordable to consumers is sound. But did it take 20 authors a lot of time and money to reach that conclusion? Given the vast number of variables in this study, that thought makes me frown deeply.
