For more than a decade, the phrase “ultra-processed food” has functioned as nutritional shorthand for danger, a dietary villain blamed for rising rates of obesity, diabetes, and heart disease. The public message has been simple: If it is ultra-processed, avoid it. But science is rarely that simple.

My colleagues and I recently reviewed the growing body of research examining ultra-processed foods, not as a single monolithic category, but as distinct subgroups with markedly different metabolic effects. What we found challenges one of the most widely accepted assumptions in modern nutrition policy: Processing alone is not the primary driver of disease risk.

Some ultra-processed foods are consistently associated with harm. Others, surprisingly, are linked to neutral or even beneficial outcomes. Across large prospective cohorts and multinational datasets, a clear pattern emerges. Three ultra-processed food categories are repeatedly associated with increased cardiometabolic risk:

• Processed meats
• Sugar-sweetened and artificially sweetened beverages
• Fats, spreads, and sauces

These foods are metabolically problematic for well-established reasons. Processed meats deliver saturated fat, sodium, nitrates, and heme iron, a combination that promotes inflammation, oxidative stress, and insulin resistance. Sugary beverages overwhelm metabolic pathways with rapidly absorbable calories. Even artificially sweetened drinks, long viewed as benign substitutes, show associations with cardiometabolic risk.

But several ultra-processed foods, including breads, breakfast cereals, and whole-grain products, are actually associated with lower risk of diabetes and cardiovascular disease. Plant-based meat alternatives, often criticized for processing intensity, are also linked to reduced diabetes risk when they replace animal products. Randomized clinical trials reinforce this pattern: Substituting plant-based alternatives for animal foods improves weight, insulin sensitivity, and lipid profiles, regardless of processing level. Why? Because fiber, phytochemicals, unsaturated fats, and lower heme-iron exposure fundamentally alter metabolic signaling. The body responds to nutrient composition, not manufacturing technique.

The NOVA classification system, widely used to define ultra-processed foods, was designed to categorize processing intensity, not biological effect. That distinction matters. When ultra-processed foods are treated as nutritionally interchangeable, policy risks unintended consequences: discouraging fortified whole-grain foods, undermining plant-based dietary transitions, and conflating processed plant staples with processed meats. The result is public confusion, and diluted dietary guidance.

Nutrition policy must evolve beyond processing absolutism.

• First, differentiate ultra-processed foods by biological source.
• Second, incentivize high-fiber processed plant foods such as fortified cereals, whole-grain breads, and plant proteins.
• Third, align dietary and climate policy to support plant-based substitutions, even processed ones, that advance both metabolic and environmental health.

The question is not whether a food is processed. The question is what it is made from, and what it replaces. If we fail to make that distinction, we risk misdirecting both policy and public behavior at a time when nutritional precision has never mattered more.

Hana Kahleova is an endocrinologist.