Food scientists suggest that lifestyle factors may be more responsible for adverse health outcomes than ultra-processed food consumption, challenging current public health recommendations.
Perspective: Ultra-processed foods and health: are we correctly interpreting the available evidence? Image Credit: Rimma Bondarenko / Shutterstock
In a recent perspective piece published in the European Journal of Clinical Nutrition, researchers discussed epidemiological studies on the effects of ultra-processed food (UPF) consumption on human health.
They conclude that there is insufficient evidence that ultra-processing specifically, and not other confounding factors, is causing observed adverse outcomes.
Confounding factors and proxy behaviors
Several recent studies have found associations between UPF consumption and adverse health outcomes, including increased risks of obesity, overweight, heart disease, type-2 diabetes, and cancer. Researchers have also found that individuals consuming the highest amount of UPFs may be exposed to a higher risk of mortality from all causes.
Results of studies showing correlations between UPF consumption and poor health have often been interpreted causally, that is, that UPF consumption is causing poor health outcomes. However, the studies are primarily observational, so causality cannot be inferred, and confounding factors may play a role.
In these studies, researchers must also assume that the measurements of food intakes are precise and accurate, that the composition of the food is known and can be characterized quantitatively, and that the results are not affected by the storage, cooking, and preparation of the food.
A step in the right direction is a recent study that focused on the contribution of emulsifiers, finding that some (but not all) were linked with the probability of developing some cancers and overall cancer risk. However, the researchers caution that extending these findings to all emulsifiers is incorrect, as many were not associated with cancer risk.
The importance of sub-group analyses
One issue is that studies may combine different groups of UPFs into one category and that many published papers do not explicitly mention sub-group analyses. However, even in a single study, different sub-groups may show significantly different relationships with health outcomes.
For example, while some UPFs (including sauces, margarine, and foods containing ultra-processed fats) may be associated with a higher risk of diabetes, cancer, or heart disease, there are indications that others, such as cereals or ultra-processed breads, may be protective.
Another study found that the relationship between heart disease and UPF consumption is driven by increased consumption of ultra-processed sauces, condiments, and fats and a lower fiber intake. An analysis of the link between death from colorectal cancer and UPF intake found that it was attributable to ultra-processed ice cream and sherbet consumption.
Similarly, a study linking mortality to UPF consumption found that most of this relationship came from ultra-processed drinks. While breast cancer incidence has been linked to sugary products, it is not associated with sweetened beverages. No studies to date have found associations between adverse health outcomes and ultra-processed vegetables or fruit. The lack of sub-group analyses in many studies means that some UPFs may be unfairly grouped with those that pose more significant health risks, further complicating public health messaging.
Regional differences and UPF consumption
Thus, experts say that while people who consume the highest amount of UPFs may face the most overall health risks, this may be attributable to a small set of UPFs rather than the ultra-processing itself. For example, certain animal-based products were established as being unhealthy long before the current UPF classification came into being.
The observed relationships may be because UPF consumption is often a proxy for non-optimal lifestyles or diets. That is, people with poorer diets are also more likely to engage in unhealthy habits such as smoking.
Importantly, the findings also show significant regional heterogeneity. On average, people in Italy consume fewer UPFs; there, mortality from all causes significantly rises for people whose UPF consumption contributes to approximately 24% of their energy intake.
Findings from countries where people consume more UPFs, such as Great Britain, are different. Mortality only increases after UPF consumption rises to greater than 40% of the overall energy intake. This regional variation suggests that other factors, such as lifestyle and dietary patterns, may play a greater role than UPF consumption itself.
Notably, studies have found no risk of increased mortality associated with UPF intake among low-income American populations who were extremely reliant on ultra-processed items. Since it is improbable that UPFs in some countries are healthier than others, this further supports the idea that confounding factors, such as broader dietary habits or socioeconomic status, may be driving the observed differences.
Conclusions
Strengthening the current evidence requires presenting results for all groups of UPFs and not combining them into one category. Studies must also investigate each step of the food preparation process and specific compounds that could contribute to observed adverse health outcomes.
If ultra-processing is not the culprit driving poor health outcomes, then public health recommendations that seek to limit the consumption of UPFs will not improve health outcomes. Instead, they could reduce access to nutrition for low-income populations who are reliant on these products.
The observational evidence from current studies can be strengthened by using experimental frameworks that explore the mechanistic and causal underpinnings of these relationships. More experimental research is needed to investigate the causality between UPF consumption and health outcomes directly, which observational studies alone cannot establish. Focus should also be given to UPFs that may be ‘nutritionally beneficial’ and protective against some diseases or adverse health outcomes.
Journal reference:
Ultra-processed foods and health: are we correctly interpreting the available evidence? Visioli, F., Del Rio, D., Fogliano, V., Marangoni, F., Ricci, C., Poli, A. European Journal of Clinical Nutrition (2024). DOI: 10.1038/s41430-024-01515-8, https://www.nature.com/articles/s41430-024-01515-8
