Nutrient profiling
The good, the bad, and the ultra-processed
Madam
In his recent invited commentary, Carlos Monteiro proposes a classification of foods based on the type and intensity of food processing(Reference Monteiro1). In particular, he identifies a category of ‘ultra-processed foods’, the consumption of which should be avoided to prevent disease and enhance well-being. In his fairly provocative title, he states that ‘The issue is not food, nor nutrients, so much as processing’.
Yet, we eat foods and we need nutrients, which is why we need efficient food-based approaches to meeting nutrient requirements. The evidence for a link between nutrition and health has prompted many countries to design food-based dietary guidelines(2). However, the implementation of these recommendations may be impaired by their imprecision(Reference Ferguson, Darmon, Briend and Premachandra3). Indeed, they are based on wide food categories, not on individual foods in the form actually bought by consumers. In that sense, those guidelines are wrongly called ‘food’-based dietary guidelines, because they do not provide recommendations on individual foods, but on categories of foods, the definition of which is very imprecise. As a result, clear recommendations on foods composed of more than one food category, such as mixed dishes and snacks, are lacking.
Moreover, food category-based recommendations are useless when it comes to choosing between two foods that have the same selling name but different ingredient and nutrient compositions and different prices. However, stigmatising a category as ‘ultra-processed foods’ will not help to overcome these limitations, because the classification Dr Monteiro proposes also lacks precision, and is therefore unlikely to be useful and operational.
Given the actual complexity of the food supply, we urgently need guidelines that present a real guarantee of optimal nutrition. Nutrient profiling systems, by providing clear information on the nutritional quality of individual foods, and explicit recommendations on the consumption of these foods, could be the missing link between nutrient-based recommendations and food category-based recommendations.
Initially intended for consumer protection and the regulation of health and nutrition claims in Europe(4), nutrient profiles can be used for different purposes, including food labelling, marketing controls, taxation/subsidies policies or product reformulation. Some of them could also be used for nutrition education and information.
This is the case with the SAIN,LIM nutrient profiling system proposed by the French Food Safety Agency(5). This system is not based on the idea that there are good foods and bad foods, but on the notion that all foods may present positive and negative aspects for health.
The SAIN,LIM system provides factual information rather than a global judgement. The positive aspects are estimated through the SAIN (score of nutritional adequacy of individual foods, calculated as the mean percentage nutrient adequacy per 100 kcal), and the negative ones through the LIM (score of nutrients whose intakes should be limited, calculated as the mean percentage of maximal recommended values for salt (as sodium), saturated fatty acids and added sugars per 100 g).
Each food can be represented on a graph (SAIN = y axis and LIM = x axis) and, by defining threshold values for both scores, each food can be classified into one of four possible classes. With this system, most unprocessed and unrefined foods fall into class 1 (i.e. the most favourable nutrient profile: high SAIN, low LIM); whereas most energy-dense nutrient-poor foods fall into class 4. Because there is no compensation between the two scores, and because artificially added nutrients are not taken into account when calculating the SAIN score, using this system or a similar one should encourage the formulation of food products that are low in energy, fat, sugar and salt, and also rich in essential nutrients and other beneficial micro-constituents naturally present in foods.
Unfortunately, the nutrient profiling system that is going to be enforced at the European level to control health and nutrition claims(6) does not present such advantages. It will likely induce the development of products that, in order to ‘pass’ the system, will be moderately loaded with fat, sugar and/or salt, and in order to have something to claim, may be artificially fortified with vitamins, minerals or other ingredients considered as positive.
Clearly, promoting the consumption of such foods, by authorising them to display nutrition and health claims, will not help people to balance their diets. On that point, I fully agree with Dr Monteiro when he says that ‘ “premium” ultra-processed foods are not a solution’. But well-done nutrient profile systems could provide a rigorous approach to overcome these drawbacks by helping to make the difference between foods that really contribute to healthy eating and foods that will instead induce nutrient inadequacy(Reference Darmon, Vieux, Maillot, Volatier and Martin7).
