“A report on childhood health from the federal Make America Healthy Again (MAHA) commission released in May 2025 mentions the concept of harm associated with UPF dozens of times, signaling the possible inclusion of this category in the forthcoming Dietary Guidelines for Americans, 2025–2030. The available evidence, however, suggests that a focus in national nutrition policy on reducing UPF consumption won’t make Americans any healthier.
The term “ultraprocessed food” [UPF] was adopted as part of a food-classification system that was introduced in a 2009 commentary by Carlos Monteiro and subsequently named “Nova.” Initially including three groups of food, Nova was updated several years later to include four groups, with foods being categorized “according to the extent and purpose of industrial processing” that occurred before sale. Group 1 contains unprocessed or minimally processed foods (e.g., foods that have undergone the removal of inedible parts, roasting, freezing, drying, crushing, or pasteurization). Group 2 consists of foods with processed ingredients used in traditional culinary practices (e.g., sugar, flour, fats, or salt). Group 3 includes traditionally processed or homemade foods prepared by adding products from group 2 to those in group 1 (e.g., some types of cheese, smoked fish, fruit preserved in syrup, pickled vegetables, or freshly baked breads). Group 4 consists of UPF, “formulations of ingredients, mostly of exclusive industrial use…often obtained from a few high-yield plant foods” and animal products and at least one characteristic item, “either food substances never or rarely used in kitchens, or classes of additives whose function is to make the final product palatable or more appealing.” [..]
Nova provides, in effect, an operational definition of “junk food” — one not based solely on amounts of sugar, fats, and salt — which could facilitate the incorporation of this notion into nutrition policy. Its focus on processing offers an implicit armistice in the decades-long “diet war,” which has pitted proponents of various nutrient-based diets (e.g., low-fat or low-carbohydrate diets) against one another. But a shift away from nutrient-based criteria to a classification system based on a single, composite metric risks imprecision, a critical consideration for public health translation.
Processing of most carbohydrate-containing foods rapidly accelerates digestion, boosting glycemic and insulinemic responses. Minimally processed grains (e.g., wheatberries) have consistently been shown in both observational and interventional studies to have health benefits as compared with highly processed grain products (e.g., white bread). Similarly, whole fruits are healthier than fruit juice and especially healthier than sugar (the main source of calories in most fruits). In contrast, processing of proteins and fats has relatively little effect on the rate of digestion and on metabolic responses. For example, health effects don’t differ substantially between steak and ground beef, soybeans and tofu, olives and olive oil, or peanuts and peanut butter. [..]
Concerns about Nova’s practicality and robustness have been noted repeatedly in the scientific literature. In one large cohort study, overall UPF consumption was linked to an elevated incidence of cardiovascular disease, but the direction of the association depended on the food type (e.g., the association was positive for sugar-sweetened beverages and negative for savory snacks). In another cohort study, the association between UPF consumption and mortality also varied according to food type. Moreover, within each quartile of diet quality, as measured by a nutrient-based index, UPF consumption wasn’t consistently associated with total or cause-specific mortality. The authors of that study concluded that diet quality had a greater effect than UPF consumption on mortality. Nonetheless, in some studies, associations between UPF consumption and disease have been at least partially independent of specific nutrients.
Evidence related to UPF consumption derives nearly exclusively from observational studies, which have a high risk of bias because of confounding and misclassification and cannot prove causation. An umbrella review graded the quality of most meta-analyses examining associations between UPF consumption and health outcomes as “low” or “critically low.”
Few randomized, controlled trials of UPF consumption have been conducted, and none have assessed long-term outcomes. In one highly cited crossover study, the 20 participants consumed about 500 more calories per day when they were fed UPF (e.g., a cheeseburger and fries) than when they consumed unprocessed food (e.g., salmon and green beans), but this effect weakened progressively during the 2-week study, and there was no washout period between phases. Differences in the energy density of solid foods or other factors known to influence short-term eating behavior, rather than any newly identified effects of UPF, may explain the findings.
The U.S. 2025 Dietary Guidelines Advisory Committee graded the evidence linking UPF consumption to obesity as “limited,” while noting that analysis was difficult “largely because of the lack of clear definition of” UPF. In a request for information published on July 25, 2025, the Food and Drug Administration and the Department of Agriculture noted concerns “about the full ability of UPF classification systems to accurately capture the characteristics of UPFs that may impact health.” A U.K. parliamentary committee declared that the country’s “broken food system” had contributed to “an epidemic of unhealthy diets, obesity and diet-related disease.” Nevertheless, it cautioned that “the Nova classification of UPFs lacks sufficient precision to be suitable for the characterisation or regulation of individual foods.”
The MAHA movement’s rush to enact policy reforms in response to what some people consider an overreliance on weight-loss drugs and procedures such as bariatric surgery to manage diet-related chronic disease in the United States is understandable. But adoption of a national nutrition policy focused on Nova — and on UPF, in particular — could do more harm than good, by misguiding consumers and the food industry and imposing costs and burdens on the public (see table). Indeed, some UPFs can be healthy (e.g., certain varieties of commercial tofu, olive-oil–based dressing, and dark chocolate), and some less-processed foods can be unhealthy (e.g., potato products, white rice, and unrefined cane sugar). Furthermore, because they are typically convenient and relatively inexpensive, healthy UPFs could play an important role in supporting population health, especially among low-income groups.
Nova may represent an important conceptual advance in long-standing efforts to reduce the burden of chronic disease caused by modern, commodity-based diets of highly processed foods. Nonetheless, it is unclear whether the Nova classification system — which considers processing, additives, and preparation venue with the use of a single metric — can be codified at a national level to support public health. No government commission or expert team has conducted a comprehensive, independent review to determine whether Nova is more robust than other food-classification systems, whether inclusion of nutrient-based criteria would enhance its precision, and whether implementing any such food-classification system would be more effective than simply advising people to avoid junk food. Critically, no clinical trial has shown that a focus on reducing UPF consumption would prevent diet-related chronic disease. Pending actionable experimental data, it would be prudent to focus nutrition-policy reform on the processed foods for which there is already strong scientific evidence of harm — for example, refined grains, added sugar, and other highly processed carbohydrates.”
Full perspective, DS Ludwig, New England Journal of Medicine, 2025.9.3