What is the NOVA classification system?

NOVA is a food classification system introduced in 2009 that sorts foods by degree and purpose of industrial processing rather than by nutrient content. It's the basis for most "ultra-processed food" research and headlines, but 2025 systematic reviews found it isn't well suited to identifying which specific factor in a food actually drives harm.

Does ultra-processed food actually cause weight gain, or is it just correlated?

There's real causal evidence for at least one mechanism: a controlled 2019 NIH metabolic-ward study found participants eating an ultra-processed diet, even matched for calories and macronutrients to an unprocessed diet, ate about 500 more calories per day and gained weight -- while losing weight on the unprocessed diet. That's one well-controlled study pointing to a specific mechanism, not proof every UPF-linked health outcome shares the same cause.

Should I try to eliminate all ultra-processed food from my diet?

The evidence doesn't clearly support that as necessary or realistic for most people. A more targeted, evidence-aligned approach is prioritizing cuts to sugar-sweetened beverages and energy-dense, low-fiber, easy-to-overeat products, rather than treating every technically "ultra-processed" item -- like plain yogurt or canned beans -- as equally harmful.

Who should be careful about ultra-processed-food rules specifically?

Anyone with a history of disordered eating should avoid self-directed restrictive food-category rules, which carry a documented relapse risk, and should work with a treating clinician instead. People managing diabetes or another diet-sensitive condition should get specific guidance from their physician or a registered dietitian rather than applying a general wellness rule.

Guides · Metabolic & Fasting Research

Ultra-Processed Food: What the Evidence Actually Says

"Avoid all ultra-processed food" is easy to say and hard to define. Here's what large reviews actually show about UPFs and health, why a soda and a fortified whole-grain cereal don't belong in the same bucket, and what's still genuinely unknown.

6 min read · Updated August 4, 2026

The headline is real -- but it's less precise than it sounds

"Ultra-processed food" (UPF) is defined by the NOVA classification system, introduced in 2009, which sorts food by how industrially it was made rather than by its nutrient content -- packaged snacks, sodas, reconstituted meat products, and most breakfast cereals all land in the same "ultra-processed" bucket regardless of their sugar, fiber, or fortification profile. A 2024 umbrella review pooling 45 meta-analyses and roughly 10 million participants found what researchers called convincing evidence linking higher UPF intake to higher risk of cardiovascular-disease-related mortality, type 2 diabetes, and common mental disorders including anxiety. A separate 2022 dose-response meta-analysis found each additional 10% of daily calories from ultra-processed food was associated with a 15% increase in all-cause mortality risk. That's a real, large, repeatedly-replicated association -- not a fringe claim.

Where the popular version overreaches is treating "ultra-processed" as one uniform hazard. It isn't. Because NOVA classifies by processing method rather than nutrition, it groups a can of soda with a fortified whole-grain cereal or a plant-based protein product -- items with very different nutrient profiles get an identical label.

A 2024 umbrella review (45 meta-analyses, ~10 million participants) found convincing evidence linking higher UPF intake to cardiovascular-disease mortality, type 2 diabetes, and anxiety/common mental disorders.
A 2022 dose-response meta-analysis found each 10% increase in daily calories from UPF was associated with a 15% increase in all-cause mortality.
NOVA classifies by processing method, not nutrient content -- a soda and a fortified cereal both count as "ultra-processed" despite very different nutrition profiles.

Not all UPFs are equal -- and the "why" is still being worked out

2025 systematic reviews evaluating the NOVA system directly found substantial heterogeneity in health outcomes across UPF subtypes: sugar-sweetened beverages consistently show adverse associations, while fortified cereals and certain dairy products show neutral or even protective associations in the same body of research. Reviewers have concluded that current classification systems, NOVA included, aren't well suited to identifying which specific factor -- additives, energy density, the loss of fiber and structure during processing, or the hyper-palatability that drives overeating -- actually drives the harm.

That distinction matters for causation, not just semantics. Most of the large studies behind the striking mortality and disease numbers are observational: people who eat more UPF also tend to eat less fiber, less produce, and more added sugar and sodium overall, which makes it hard to isolate "processing itself" from "the overall diet quality that tends to come with it." The strongest causal (not just correlational) evidence comes from a smaller, tightly controlled 2019 NIH metabolic-ward study, where participants given an ultra-processed diet matched for calories, sugar, fat, fiber, and macronutrients to an unprocessed diet still ate about 500 more calories per day and gained weight on the UPF diet -- while losing weight on the unprocessed diet, with no calorie counting involved. That result points toward something about ultra-processed food itself (likely its energy density and eating speed) driving overeating, independent of nutrient content on paper -- but it's one study of a specific mechanism (calorie intake), not proof that every associated health outcome shares the same cause.

2025 systematic reviews found sugar-sweetened beverages consistently linked to harm, while fortified cereals and some dairy products showed neutral or protective associations -- "ultra-processed" is not one uniform risk category.
Most large UPF-and-mortality studies are observational; heavy UPF eaters also tend to eat less fiber and produce, which confounds isolating "processing" from "overall diet quality."
A controlled 2019 NIH metabolic-ward study found an ultra-processed diet, even matched for calories and macronutrients, led to ~500 more calories eaten per day and weight gain versus an unprocessed diet -- real causal evidence for one specific mechanism (overeating), not for every UPF-linked outcome.

The honest, practical takeaway

The evidence-backed version of this advice isn't "eliminate all ultra-processed food" -- for most people that's an unworkable, all-or-nothing rule built on a classification system its own reviewers say is too blunt. The better-supported approach is to weight attention toward the categories with the most consistent evidence of harm: sugar-sweetened beverages, and products that are energy-dense, low in fiber, and easy to overeat quickly (many packaged snacks and fast food fall here). A fortified whole-grain cereal, plain yogurt, or canned beans -- all technically "ultra-processed" under NOVA -- don't carry the same evidence of harm and don't need to be treated the same as soda. Reading labels for added sugar, sodium, and fiber content is a more targeted lever than sorting food by processing category alone.

The safety gate: this isn't a rule for everyone

Restrictive food-category rules -- "cut out all processed food" -- carry a documented relapse and preoccupation risk for anyone with a history of an eating disorder, and should not be self-directed without a treating clinician's input. People managing diabetes or another condition where carbohydrate composition and timing matter should get specific dietary guidance from their physician or a registered dietitian rather than a general wellness rule, since blood-sugar management depends on more than a food's processing category. None of this is a reason to ignore the evidence above -- it's a reason to apply it as a general weighting tool, not a rigid, self-directed elimination diet.

A quick honest recap

Large reviews (45 meta-analyses, ~10M participants) show convincing evidence linking higher ultra-processed food intake to cardiovascular mortality, type 2 diabetes, and anxiety/mental-disorder risk -- this part of the headline holds up.
But NOVA classifies by processing method, not nutrition -- a soda and a fortified cereal both count as "ultra-processed," and 2025 reviews found their health associations differ substantially.
Most of the strongest numbers are observational and confounded with overall diet quality; a controlled 2019 NIH study found causal evidence for one mechanism (UPF diets driving ~500 extra calories/day eaten) but not for every associated outcome.
The practical, evidence-aligned move is weighting attention toward sugar-sweetened beverages and low-fiber, easy-to-overeat products -- not treating every technically-"ultra-processed" food as equally harmful.

Frequently asked

Is all ultra-processed food bad for you?: No -- the evidence shows substantial variation within the category. Sugar-sweetened beverages are consistently linked to harm, while fortified cereals and some dairy products show neutral or even protective associations in the same research. The NOVA classification groups them together by processing method, not by nutrient content, which is a limitation reviewers have flagged directly.
What is the NOVA classification system?: NOVA is a food classification system introduced in 2009 that sorts foods by degree and purpose of industrial processing rather than by nutrient content. It's the basis for most "ultra-processed food" research and headlines, but 2025 systematic reviews found it isn't well suited to identifying which specific factor in a food actually drives harm.
Does ultra-processed food actually cause weight gain, or is it just correlated?: There's real causal evidence for at least one mechanism: a controlled 2019 NIH metabolic-ward study found participants eating an ultra-processed diet, even matched for calories and macronutrients to an unprocessed diet, ate about 500 more calories per day and gained weight -- while losing weight on the unprocessed diet. That's one well-controlled study pointing to a specific mechanism, not proof every UPF-linked health outcome shares the same cause.
Should I try to eliminate all ultra-processed food from my diet?: The evidence doesn't clearly support that as necessary or realistic for most people. A more targeted, evidence-aligned approach is prioritizing cuts to sugar-sweetened beverages and energy-dense, low-fiber, easy-to-overeat products, rather than treating every technically "ultra-processed" item -- like plain yogurt or canned beans -- as equally harmful.
Who should be careful about ultra-processed-food rules specifically?: Anyone with a history of disordered eating should avoid self-directed restrictive food-category rules, which carry a documented relapse risk, and should work with a treating clinician instead. People managing diabetes or another diet-sensitive condition should get specific guidance from their physician or a registered dietitian rather than applying a general wellness rule.

HWS Health Research Guide — free

Evidence-based health protocols — research-driven, not miracle claims. Self-help education, not medical advice; no guaranteed health outcome.

Send me the guide

Ready to put the research to work? Create a free account →

This guide is self-help educational research, not medical advice, and Health Wealth Stealth is not a healthcare provider. It does not create a doctor-patient relationship. Consult a licensed healthcare professional about your health.