The findings, published online by the British Medical Journal on February 14, are based on NutriNet-Santé, an ongoing population-based cohort study launched in France in 2009. Analysis of results was led by researchers from Paris’ Sorbonne University. The paper, which has not yet been peer-reviewed, sparked a rash of national press attacks on overly processed food products.
In response, Dr Ian Johnson, nutrition researcher and emeritus fellow, Quadram Institute Bioscience, formerly the Institute of Food Research, said: “The problem is that the definition of ultra-processed foods they have used is so broad and poorly defined that it is impossible to decide exactly what, if any, causal connections have been observed.”
The four NOVA food groups
The scientists used a classification system known as NOVA, developed by researchers in Brazil, which groups food into four categories.
The four categories
- raw and minimally-processed food, including nuts and seeds, fresh fruit and vegetables, eggs and milk;
- processed ingredients used in cooking, such as sugar, food oils and butter;
- processed foods, such as salted, cured or smoked meats, packaged vegetables, canned fish and meat and cheese;
- ultra-processed foods, or industrial formulations of five or more ingredients including those exclusively used in this group, such as colours, flavours and processing aids; products such as carbonated drinks, crisps and snacks, breakfast cereals and ready meals are given as examples.
The paper claimed a 10% increase in the proportion of ultra-processed foods in the diet was associated with a 12% increase in overall cancer risk and an 11% increase in breast cancer risk.
Those working on the results claimed lifestyle factors, such as exercise, obesity levels, smoking and alcohol consumption were accounted for as part of the reporting by the 105,000 participants. Medical records and relative amounts of foods and drinks consumed were also tracked.
Tom Sanders, professor emeritus of nutrition and dietetics, King’s College London, stated: “However, the high-consumers did differ in several aspects that may have contributed to the risk associated with ultra-processed food intake.
“For example, the participants who consumed a lot (33.3%) of these ultra-processed foods compared with those who consumed very little (about 18.7%) were more likely to be current cigarette smokers (20.2% vs 16.9%), physically inactive (24.7 vs 20.9%) and more likely to be taking oral contraceptives (30.8% vs. 22.0%).”
Referring to the use of the term ‘ultra-processed’, Sanders stated: “From a nutritional standpoint, this classification seems arbitrary and based on the premise that food produced industrially has a different nutritional and chemical composition from that produced in the home or by artisans. This is not the case.”
He also pointed out that the study observed no link between the consumption of processed meat such as mechanically-recovered and nitrite-treated meat and the development of cancer. That was despite the fact that other recent research had made that link. Such meat would not fit into the paper’s definition of ultra-processed food.
The researchers proposed four theories as to why consumption of ultra-processed foods could be linked to higher cancer risks.
The four theories
- they are richer in energy, sodium, fats and sugar and poorer in fibre and micronutrient content;
- they often contain a cocktail of additives, consumption of which may be connected to cancer development;
- the use of heat treatments in the production of such foodstuffs is linked to the formation of contaminants such as acrylamide, which research has associated with increased cancer risk;
- studies have suggested the use of chemical endocrine disruptors such as bisphenol A in the packaging of ultra-processed food could be linked in increased cancer risk.