According to the NOVA classification developed at the University of São Paulo, an ultra-processed food is an industrial formulation containing at least one ingredient that does not exist in a domestic kitchen: hydrogenated oils, high-fructose corn syrup, xanthan gum, polysorbate 80, artificial flavorings. The distinction concerns the production process rather than the nutritional profile. The most recent epidemiological data cited in the source show that every 10% increase in daily calories from ultra-processed foods is associated with a 10% increase in all-cause mortality risk.

The term “ultra-processed” has entered the everyday food vocabulary in recent years with the same speed that “gluten-free” did a decade ago. The result is similar: everyone uses it, few people understand it, and interpretations proliferate somewhere between alarmism and total denial.

It is worth understanding properly. Not so you become afraid of what you eat, but so you have a compass when you are standing in front of a supermarket shelf and do not know where to start.

NOVA: where the term comes from

“Ultra-processed” is not a journalistic invention. It comes from a food-classification system called NOVA, developed by researcher Carlos Monteiro and his group at the University of São Paulo from 2009 onward and updated over the years. NOVA divides foods into four groups according to the degree of industrial processing — not according to their nutritional profile, but according to the production process.

The criterion defining group 4 is specific: these are industrial formulations containing at least one ingredient or additive you would never normally find in a home kitchen. Hydrogenated oils, protein hydrolysates, high-fructose corn syrup, xanthan gum, artificial flavors, emulsifiers such as polysorbate 80. The point is not to judge whether these ingredients are dangerous — it is to recognize that they indicate a production process far removed from home preparation.

What the epidemiological data say

Research on ultra-processed foods has exploded over the past decade. The broadest review cited in the source — published in the BMJ in 2024 and covering almost ten million participants — identified direct associations between high consumption of NOVA 4 foods and 32 of the 45 health outcomes analyzed. All-cause mortality, cardiovascular disease, type 2 diabetes, some cancers, gut health and mental disorders were among them.

The concrete numbers cited in the source: every 10% increase in daily calories from ultra-processed foods is associated with a 10% increase in all-cause mortality risk. Each additional daily serving corresponds to a 4% increase in cardiovascular risk.

It is important to say honestly that these are observational studies, not controlled experiments. Nutritional epidemiology measures associations, not direct causation. We cannot say with certainty whether ultra-processed foods themselves cause those outcomes or whether they are a marker for other factors — an overall poor diet, sedentary lifestyle, lower income. The scientific community still debates this point, and anyone who tells you the question is definitively closed in either direction is simplifying it.

What we can say is that the association is robust, replicated in different settings and consistent enough to justify attention.

The limitations of the classification

NOVA has structural weaknesses worth knowing, if only so you do not use it like a machete.

The first is that it treats every NOVA 4 food alike regardless of nutritional profile. Packaged whole-wheat bread with an emulsifier and industrial potato chips end up in the same group. That is methodologically debatable, and some of the most serious criticism of NOVA starts here.

The second is that classification can be ambiguous around borderline cases. The same yogurt may be classified as NOVA 1, 3 or 4 depending on how ingredients are interpreted. Reliability can vary between observers.

The third is that NOVA measures the presence of additives, not their quantity. A product with a trace of thickener and one loaded with technological additives receive the same label.

That said, an imperfect tool that gives you direction can be more useful than having no tool at all.

Italy — and why this concerns us too

Italy is one of the Western countries with the lowest prevalence of ultra-processed foods in the diet. NOVA 4 foods account for about 23% of Italians’ daily calories in the figures cited by the source — compared with around 60% in the United States and 50% in the United Kingdom.

The problem is the trend. In 2005 the Italian figure was around 12%. In fifteen years it had almost doubled. And the products driving that growth are not exotic imports: snack cakes, packaged biscuits, salty snacks, sugary drinks and ready meals. Foods that have already been established on Italian tables for decades.

The examples that surprise people

One useful feature of NOVA is that it challenges some familiar intuitions.

Products often perceived as healthy that can be NOVA 4 include flavored fruit yogurt with thickeners and flavorings, industrial whole-wheat bread with emulsifiers, “high-fiber” breakfast cereals with sweeteners and colorings, soy-based plant burgers with methylcellulose and flavorings, and industrial plant drinks with gellan gum and lecithins. Organic certification does not automatically prevent a product from being NOVA 4.

Products that feel industrial but can be NOVA 3 include canned tuna in oil, Parmigiano Reggiano, canned peeled tomatoes without additives and artisan sourdough bread without additives. Cured meats also require an important distinction because they do not all belong to the same category.

DOP dry-cured ham made without additives — meat and salt, nothing else — is NOVA 3. Cooked ham, mortadella, frankfurters and most industrial cured meats containing nitrites, preservatives and flavorings are NOVA 4. This distinction also matters for health: processed meats in the technical sense — meats transformed through industrial curing, smoking or chemical preservatives such as nitrites — are classified by IARC as Group 1 carcinogens for colorectal cancer, based on more than 800 epidemiological studies. That does not mean they have the same potency as cigarette smoking: Group 1 describes the certainty of the evidence, not the size of the risk. The risk is dose-dependent — every 50 grams of processed meat per day is associated with an 18% increase in colorectal-cancer risk — and is primarily relevant to habitual, prolonged consumption. An occasional slice of mortadella is not a sentence. But it is information worth knowing, especially if industrial cured meats are on the table every day.

The practical rule

There is no precisely defined safe threshold, but the data cited in the source suggest that risks become statistically relevant around 20–25% of daily calories from NOVA 4 foods. Italians are already around that level on average.

The 80/20 rule — which Memoria Mediterranea uses as a broader principle — works well here too. If 80% of what you eat is recognizable, made from ingredients that exist in nature and that you could use in your own kitchen, the remaining 20% of more elaborate products does not undermine the whole pattern. The point is not to ban snack cakes or never buy a ready meal. It is not to make them the foundation of your diet.

The simplest practical tool is the ingredient list: if you see substances you would never use while cooking — perhaps names you cannot even pronounce — you are probably looking at a NOVA 4 food. It is not a verdict. It is information.