Ultra-processed foods: from additives to packaging, but do we really know everything?

In recent years, the debate on ultra-processed foods (UPF) has taken centre stage in scientific research and communication to the public. However, the idea that 'we know everything' about UPF nowadays seems excessive, starting with the very definition of 'ultra-processed' on which there is no unanimous agreement. The concept proposed by the NOVA system has been modified several times over time and is based on criteria that mix technological aspects (the "process") with elements linked to the presence of so-called "non-household" ingredients or even the number of ingredients, the presence of "cosmetic" additives ("formulation") or the use of "sophisticated packaging", extremely simplified aspects that have little scientifically robust and that do not allow a consistent distinction to be made between very different foods.

This is compounded by a substantial methodological difficulty: there is still no agreed and validated way of assessing and quantifying UPF consumption. Different epidemiological studies use inconsistent indicators (e.g. grams, percentage of energy, frequency of consumption, number of servings), making it difficult to compare results and interpret the real impact on health.

From the point of view of scientific evidence, the current literature is dominated by observational studies, which show associations between high UPF consumption and chronic disease risk, but do not allow causal links to be established. Furthermore, more recent analyses have shown that, when UPF are divided into subcategories, certain groups (such as ready-to-eat cereals or plant-based foods) may show neutral or even protective effects. If the negative effect were attributable to 'ultra-processing' alone, one would expect a uniform impact on all categories, which does not seem to be the case. Controlled interventional studies are still limited and often of short duration.

These elements suggest that the nutritional composition and overall dietary context carry much more weight than the "UPF" classification suggests. Moreover, all these limitations outlined so far make it premature to consider incorporating the UPF concept known to date into nutrition policies or population guidelines. Furthermore, studies of associations between UPF consumption and health outcomes are increasing, which emphasise the importance of analysing the contribution of specific sub-categories of UPF and not of the whole class, demonstrating that simplification can be misleading.

Together with the Universities of Florence and Teramo, we are conducting an investigation, dubbed Promenade, to assess whether within a Mediterranean dietary model, the inclusion of foods classified as so-called UPF may have an impact on cardiometabolic, microbiota and other health markers.