Role of food processing on human health and current limitations

The importance of macro- and micronutrients (such as vitamins and minerals) for human health is well known: while their role in nutritional deficiency context is undisputed (Coppens 2020), evidence concerning their supplementation for disease prevention through industrially produced supplements has often shown quite disappointing results and raised questions as to their benefit (Ronis et al. 2018). Growing evidence suggests that the food matrix and the level of food processing may also play a role in the relation between diet and health, aside from the nutritional content. The most common classification used so far is the NOVA classification, which includes four groups (unprocessed/minimally, culinary processed, processed and ultra-processed foods (UPFs)) based on the level of processing (Martínez Steele et al. 2016). There are consistent findings suggesting that consumption of UPFs are associated with various negative health effects, including higher risk of obesity (Askari et al. 2020), cardio-metabolic disorders and mortality (Pagliai et al. 2021). Moreover, new evidence on their potential association with mental disorders is emerging (Askari et al. 2020). Interestingly, a recent study showed that adjustments to limit the confounding effect of the nutritional quality of the diet just maintain the retrieved associations between UPF consumption and mortality risk (Bonaccio et al. 2022), suggesting that factors other than “nutritional” would play a role on the potential effects of UPF on human health.

The study of Rosignoli da Conceição et al. (2023) published in the present issue showed that consumption of unprocessed foods was inversely related to obesity features but also associated with higher probability of cardiovascular events. The study was conducted on more than 2000 participants from the Brazilian Cardioprotective Nutritional Program Trial (BALANCE Program Trial). Unprocessed and minimally processed foods were associated with lower prevalence of elevated waist circumference and overweight, but the authors also found a concerning rise in probability of cardiovascular events. The authors explain that contrary to the expectations, higher consumption of unprocessed and minimally processed foods was related to higher intake of trans fats and cholesterol due to their major dietary sources, such as beef, poultry and other traditional food preparations common in Brazil. There is extensive evidence from the scientific literature supporting the benefits of fruit and vegetable (Angelino et al. 2019), legumes (Martini et al. 2021) and whole grain (Tieri et al. 2020) consumption against cardiovascular disease (CVD), as well as a substantial increased risk associated with excess consumption of meat (Grosso et al. 2022). Although results from a meta-analysis of nationally representative samples show that higher intake of NOVA group four UPFs is correlated with lower nutritional quality of the overall diet (Martini et al. 2021), the use of this categorisation for unprocessed and minimally processed foods in relation to health outcomes is rarely explored and largely unknown.

While current evidence is underlying an association between individuals consuming more UPFs and risk of various diseases, there are several limitations of the current classification (and its variations over time) that needs to be considered. Most studies fail to report which food groups actually contribute to the energy share of UPFs in their samples and, more importantly, how these foods are distributed across categories of UPF consumption. Since most of such foods are by definition energy-rich, it is likely that a large part of high consumers would assume high caloric “discretionary” foods (defined as foods which are not essential for human health), which would be highly represented by so-called “junk foods”, while those identified as consuming less UPFs would probably consume more nutritionally adequate foods (even though ultra-processed). Since all effect sizes retrieved in epidemiological studies are relative measures (odds ratios, relative risks, etc.), it may be possible that these categorisations are in fact comparing patterns of dietary choices rather than the level of processing of the foods included in such dietary patterns. This is one crucial point of the overall matter: the taxonomy used is not adequately characterising the exposure, putting in relation heterogeneous exposures but attributing them all to the level of processing. Several studies correctly adjust for diet quality measures or proxies, but this approach does not necessarily prevent the potential confounding or modifier effect.

Another methodological limitation yet related to the identification of the exposure is the impossibility, from a biological and mechanistic point of view, to compare the results, yet significant, from countries where consumption of UPF is very high to those where it has been reported to be low. In fact, countries where the energy share of UPFs reaches up to 80% of the diet would have a completely different dietary profile than those countries with much lower consumption: for instance, a high UPF country will have a considerably higher prevalence of sugary-sweetened beverage consumption than low UPF countries, because the physical characteristics of beverages could promote higher intakes at different day occasions (Robinson et al. 2013, 2014). In contrast, countries with relatively low UPF consumption have most likely a wider range of contributors to UPF intake, resulting in completely different dietary patterns and, consequently, mechanisms of action towards health. Even considering the non-nutritional factors, the situation remains ambiguous: assuming the detrimental effects on health would be driven by “some” additives, the net amount of consumption would be considerably different between low and high UPF countries, as well as their different intake across categories of level of UPF consumption would be unmeasurable being unrelated to energy intake (i.e. individuals falling into a low category of UPF intake can still consume more quantities of additives derived from low-calorie or low-weight UPFs as opposed as in higher categories of UPF intakes). Finally, from the ecological point of view, such different amounts in exposure to UPFs should result in an equivalent burden of related non-communicable disease outcomes across countries, which in fact are not supported by the current estimates.

In summary, while the potential effects of additives present in UPFs are hypothesised to play a role in human health, the nutritional domain should also be taken into consideration. Unprocessed and minimally processed foods may still be part of unhealthy dietary patterns and provide a nutritionally inadequate contribution to the diet. While current evidence concurs with the recommendation of limiting the consumption of UPFs, a preference for a healthy and nutritionally adequate dietary pattern in line with the scientific literature on the risks and benefits associated with the major food groups should be yet kept in mind.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

None.