As inflation affects household budgets more than ever, other rising figures have also become worrying, particularly the obesity epidemic. A new survey by INSERM reveals that a growing number of French people are obese or overweight. This is in spite of public health initiatives such as the flawed Nutriscore food labelling system. How can we account for the lack of results, despite the resources allocated to this system? Are we relying on outdated nutritional concepts?
Rising obesity rates in France
A new study conducted by The League against Obesity through INSERM found that 47.3% of the 10,000 adults surveyed in the study were overweight or obese – a figure that represents one person out of every two in France. According to the authors, this proportion has doubled since 1997. Although it is important not to confuse being overweight and obesity, the figures remain alarming. Being overweight is a risk factor while obesity, characterised by a Body Mass Index (BMI) of over 30, is a recognised disease.
These results raise the question: how is the situation worsening when French public health authorities have put so much effort into communication campaigns, nutritional labelling, and the development and application of the Nutriscore scheme? While not mandatory, more and more products are displaying the traffic light-coloured, A to E scheme. Was this system tested before its implementation? Did France look at similar systems in other countries? Could the system be too simplistic? These questions have become crucial as the debate on food labels rages in Brussels. It would be unscientific to generalise a system without knowing its effectiveness.
Nutriscore system imposing its own rules
Nutritional advice follows some essential rules: it should promote a balanced diet based on moderate food intake and composed of a variety of food groups. The Nutriscore system is a five-letter rating (A to E in decreasing value) that characterises the nutritional level of 100g or 100ml of a food. It encourages consumers to prefer products rated A or B. But does it actually encourage healthy nutrition?
The Nutriscore scheme was designed to help consumers make healthier choices and to combat obesity by reducing the caloric density of food. This approach is similar to various nutritional labelling approaches trialled unsuccessfully for over forty years in the United States. These approaches did not reduce the weight gain of the average American citizen since metabolic syndrome is the result of multiple imbalances. It cannot be controlled by a simplistic algorithm.
Nutriscore’s promoters aimed to provide clear, easily understood information for consumers to inform them about the excess sugar, salt, and fat in foods. This is intended to help them avoid high-calorie foods and to simplify changes in consumption habits.
The algorithm’s calculation is quite simple. On one hand, it limits foods and ingredients which are unhealthy when consumed in large quantities, such as calories, saturated fatty acids, sugars, and salt. And, on the other hand, it promotes healthy foods, based on the amount of fibre and protein, and the presence of vegetables, fruits, nuts and olive oil. Fats, cheeses, and drinks have their own specific weighting in the algorithm. It’s also important to point out one little-known detail: the Nutriscore is in no way adapted to foods for children under the age of three.
Why are we struggling to see the results?
The picture painted by the INSERM survey should concern the promoters of the Nutriscore scheme. The labelling system, which has been designed to fight against obesity, seems to have had no impact. In fact, according to the survey, the situation seems to be getting worse. This should lead us to a closer examination of the algorithm.
First of all, it should be noted that the Nutriscore algorithm does not take into account the micronutrient composition of foods, such as vitamins or minerals like calcium.
Secondly, the caloric density of foods is based on a gross portion size that favours foods with high water content over processed foods with less moisture. For example, the total calorie intake of 100ml of semi-skimmed milk (rated A) has nothing to do with 100g of cheese (rated D at best), even though the fat profile of the two products is very similar. It is obvious that the food intake will be different for the two products, but the colour code does not take this into account.
Finally, it is important to remember that healthy eating is not determined by the content of certain components in a given food product. It is determined by the average composition of a consumer’s diet and lifestyle, including the frequency of food intake and the quality of the food they choose. But it gets worse the closer one looks at the Nutriscore system. The algorithm belies a concerning bias.
An outdated concept: evidence from rapeseed oil and saturated fatty acids
The methodological generalities mentioned are significant, but they say nothing about the “science behind the algorithm”. The main problem with the Nutriscore label is that it relies on an outdated understanding of lipids that disregards all the scientific advances of the last thirty years. This can be shown through two examples: rapeseed oil and saturated fatty acids.
The favourable weighting the algorithm gives to rapeseed oil is emblematic: the high rating is essentially based on the low content of saturated fatty acids (around 7%) in rapeseed oil, which is accurate but extremely reductive (1). During the 1990s, experts agreed on the intake of the two polyunsaturated families: 5% of total calorie intake for the linoleic family and 1% for the linolenic family, although this threshold is not always met. Unfortunately, Nutriscore does not consider these advances in our knowledge, and the score it assigns to rapeseed oil does not reflect this factor.
One of the other problematic aspects of the Nutriscore is its negative view of saturated fatty acids. They are only harmful when there is a deficiency in the two types of polyunsaturated fatty acids just mentioned. In fact, the most harmful fatty acids are trans fatty acids, which often come from artificially hydrogenated oils. There are many studies that examine the effects of saturated fatty acids, which have very different functions depending on the length of their carbon chain.
As early as 2010, a meta-analysis of 21 studies concluded that there was no link between the intake of saturated fatty acids (SFA) in dairy products and the risk of cardiovascular disease. At present, there are several meta-analyses showing an inverse correlation between consumption of various dairy products and heart attacks, strokes, type 2 diabetes, and metabolic syndromes.
We refer the reader to our paper on the nutritional makeup of the fatty acids found in cow milk that was recently published in the IAA journal. The paper provides for more details on the physiological properties of SFAs according to their chain length (2) (3).
Lastly, we would like to reflect on the role of myristic acid given the poor rating of dairy products by Nutriscore.
Multiple physiological roles and the complexity of nutrition: the case of myristic acid in milk
Dr. François Mendy conducted a series of studies at the request of the French Dairy Interbranch Organization (CNIEL). The studies confirmed the positive role of myristic acid on cholesterol (HDL). Myristic acid is derived mainly from mammalian milk. The studies also found an interesting increase in DHA (decosahexaenoic acid).
These studies also showed an improvement in the fluidity and plasticity of red blood cell membranes, which facilitated increased circulation in the fine arterioles of the brain. Lack of fluidity is a major problem of aging and can increase the risk of a stroke. In addition, myristic acid combines with many proteins, which allows for synergies in the metabolism of unsaturated fatty acids.
This quantity of scientific data, published in top-level peer-reviewed journals (4), leads to significant questions about a rating algorithm that gives undue weight to the saturated fatty acid content and penalises cheeses with low ratings.
We are greatly concerned about these issues because they show a lack of awareness of the latest scientific developments. Ultimately, we must recognize that the Nutriscore system is just an oversimplification. It is also worrisome because it implies a patronizing view of our society, where only the elite can access knowledge. This approach ignores the fact that obesity has multiple causes: imbalances due to a sedentary lifestyle, the lack of regular physical activity including walking, and the social aspects of eating. The proportion of responsibility attributed to food is unfairly reductive, emphasizing excess calorie intake, despite the causes of metabolic syndrome being related to several physiological imbalances.
Once again, we face technocratic ambitions that require little knowledge and public action. We should not be surprised that these efforts fail to reduce the epidemic of overweight and obesity that afflicts our society. Ultimately, we wonder why no serious impact studies on the subject have been launched. Who knows whether we might discover that the effect of Nutriscore is the opposite of what was expected?
(1) Between 1985 and 1989, in my role as director of the oilseed inter-professional association (ONIDOL), I took part in the rehabilitation of rapeseed oil and fatty acids of the linolenic family (known as n-3 or omega 3). The recommended intake is 1% of the total bioavailable calorie intake, i.e., in position 2 (or median) of the triglyceride. Rapeseed oil contains 10% linolenic acid, two-thirds of which are in the middle position of the triglycerides.
(2) We review the short fatty acids (butyric (C4:0) and caproic (C6:0)), the medium fatty acids (caprylic (C8:0), capric (C10:0), and lauric (C12:0)) and the myristic (C14:0) and palmitic (C16:0) acids.
(3) See also in the magazine Paysans et Société: “L’intérêt nutritionnel de la matière grasse du lait et des fromages”.
(4) We refer the reader to our study published by the IAA for details on the analyses of myristic, palmitic, and stearic acids.
This post is also available in: FR (FR)