Ultra-Processed Food: What the Evidence Says About Health Risks

Beyond “Just Eat Real Food”

The conversation around processed food has historically been vague — eat more “whole foods,” avoid “junk food.” But in the last decade, researchers have developed a more precise framework for classifying foods by their degree of processing, and the evidence linking ultra-processed food (UPF) to poor health outcomes has become difficult to ignore.

The NOVA Classification System

Developed by Carlos Monteiro’s group at the University of São Paulo, NOVA classifies all foods into four groups:

Group 1 — Unprocessed or minimally processed: Fresh fruits, vegetables, eggs, meat, legumes, nuts, plain milk, water.

Group 2 — Processed culinary ingredients: Oils, butter, sugar, salt, flour — used in cooking but not consumed alone.

Group 3 — Processed foods: Canned vegetables, artisanal bread, cheese, cured meats — recognizable foods with added Group 2 ingredients.

Group 4 — Ultra-processed foods: Industrial formulations with five or more ingredients, typically including substances not used in home cooking (high-fructose corn syrup, hydrogenated oils, emulsifiers, flavor enhancers, colorings). Think packaged snacks, instant noodles, soft drinks, reconstituted meat products, and most fast food.

In high-income countries, UPF now accounts for 50-60% of total caloric intake. In the US and UK, it’s even higher among younger demographics.

The Hall 2019 Study: The Smoking Gun

Kevin Hall’s 2019 metabolic ward study at the NIH is arguably the most important nutrition study of the decade. Twenty adults lived in a metabolic ward for four weeks, eating either an ultra-processed or unprocessed diet for two weeks, then switching. Both diets were matched for calories, macronutrients, sugar, fat, sodium, and fiber offered.

The key finding: On the ultra-processed diet, participants spontaneously ate approximately 500 more calories per day and gained about 1 kg in two weeks. On the unprocessed diet, they lost about 1 kg. This happened despite both diets being equally available and palatable-rated.

This study is powerful because it’s a randomized crossover trial in a controlled environment, eliminating confounding variables that plague observational research. Something about ultra-processed food drives overconsumption beyond its macronutrient composition.

Why Ultra-Processed Food Causes Overconsumption

Several mechanisms likely work together:

Eating Speed

Participants in Hall’s study ate the ultra-processed meals significantly faster — about 50 calories more per minute. UPF is typically softer, requires less chewing, and breaks down faster in the mouth. This outpaces satiety signaling, which takes about 20 minutes to register.

Hyperpalatability

UPF is engineered to combine fat, sugar, salt, and texture in combinations rarely found in nature. This combination activates reward circuits more intensely than whole foods, driving desire to eat beyond homeostatic hunger.

Nutrient Displacement

When UPF dominates the diet, it displaces nutrient-dense foods. The protein leverage hypothesis suggests that humans eat until protein needs are met — if food is low in protein relative to calories (as much UPF is), total caloric intake increases to reach adequate protein.

Gut Signaling Disruption

Emerging evidence suggests that emulsifiers and other additives in UPF may alter gut barrier function and microbiome composition, potentially disrupting gut-brain satiety signaling. This research is still early-stage but mechanistically plausible.

Epidemiological Evidence

Beyond the Hall study, large prospective cohorts consistently link UPF consumption to:

• Obesity: Dose-response relationship across multiple populations
• Type 2 diabetes: Each 10% increase in UPF calories associated with 15% higher diabetes risk. Our analysis of sugar and type 2 diabetes explores this pathway.
• Cardiovascular disease: Including trans fats and heart disease — though most countries have now banned industrial trans fats
• Cancer: The NutriNet-Santé cohort found associations with overall cancer, breast cancer, and colorectal cancer. Processed meat and cancer risk is one of the most established links.
• All-cause mortality: Multiple cohorts show 15-30% increased mortality risk in highest UPF consumers

The limitation of observational data is always confounding — people who eat more UPF may also smoke more, exercise less, and have lower socioeconomic status. But the consistency across populations, dose-response relationships, and biological plausibility strengthen the case.

What Actually Counts as Ultra-Processed?

This is where NOVA gets tricky. Some examples that surprise people:

Ultra-processed (Group 4): Most commercial breads, breakfast cereals, flavored yogurts, protein bars, plant-based meat alternatives, instant soups, soft drinks, most packaged snacks

Not ultra-processed: Traditional sourdough bread, plain Greek yogurt, home-cooked meals with basic ingredients, cheese, olive oil, butter

The distinction isn’t about health halos — it’s about industrial processes and ingredient lists. A food with 15 ingredients including methylcellulose and maltodextrin is UPF regardless of whether it’s marketed as “natural” or “organic.”

Practical Approaches

Eliminating all UPF is unrealistic for most people. A more practical framework:

1. Cook more meals from basic ingredients — even simple ones
2. Read ingredient lists — if you see substances you wouldn’t find in a kitchen, it’s likely UPF
3. Focus on displacement — adding more whole foods naturally reduces UPF share
4. Don’t moralize food — UPF isn’t “evil,” it’s just poorly suited to our satiety mechanisms
5. Prioritize protein and fiber — these dampen the overconsumption effect

Key Takeaway

Ultra-processed food isn’t dangerous because of any single ingredient — it’s dangerous because it systematically bypasses our satiety mechanisms, leading to chronic overconsumption. The Hall study demonstrated this causally. Reducing UPF from 60% to 30-40% of calories while increasing whole food intake is probably one of the highest-leverage dietary changes most people can make.