Explore Supplements Subscribe
The VerdictMODERATE CONVICTIONVerdict Score 78

"Processed" is too vague — sugary drinks and processed meats are the real problem, not every food in a packet.

Tonight, look at the packaged foods in your kitchen. Ask yourself two questions: "Does this have added sugar or is it a sweetened drink?" and "Is this processed meat?" If yes to either, that's the actual high-risk category — not every food in a wrapper.

  1. In the most tightly controlled study ever done, people eating ultra-processed food ate 500 extra calories per day — even when the meals were matched for protein, fat, carbs, and fiber. They ate faster and their fullness hormones didn't fire properly.
  2. When researchers broke NOVA Group 4 into sub-categories, industrial whole-grain breads and plant-based alternatives carried zero elevated risk. Sugary drinks and processed meats carried almost all of it.
  3. When overall diet quality was accounted for, ultra-processed food alone showed no consistent death-rate signal. It's what UPFs displace in your diet — not the processing itself — that drives much of the harm.

Think of the NOVA classification like a hotel star rating that lumps a roadside motel and a boutique hotel into the same category because they both have beds. A bag of chips and a high-fiber whole-grain bread are both "ultra-processed" — but one overrides your hunger signals while the other feeds the good bacteria in your gut. The label tells you almost nothing about the actual risk. You have to look at what's inside, not just the classification on the outside.

SH
Dr. Seth Holbrook, DPT — Doctor of Physical Therapy • Coach to 300+ clients
I built The Verdict to cut through recycled health advice and show what the evidence actually supports.
Partially Correct

What "Processed" Actually Means for Health

The ultra-processed food panic is partly justified — but the label hides more than it reveals. Here's what the evidence actually targets.

Conviction: Moderate

Tonight, look at the packaged foods in your kitchen. Ask two questions: "Is this a sugary drink?" and "Is this processed meat?" If yes to either, that's where the actual risk lives.

Large-scale cohort data shows almost all elevated risk from ultra-processed food concentrates in these two sub-categories — not in every food that comes in a wrapper.

Takes 2 minutes. Just open the fridge and read labels.

The Verdict

"Processed" is too vague — sugary drinks and processed meats are the real problem, not every food in a packet.

Think of the "ultra-processed" label like a hotel review system that lumps a roadside motel and a boutique hotel into the same category because they both have beds. A bag of chips and a high-fiber whole-grain bread are both "ultra-processed" — but one overrides your hunger signals while the other feeds the good bacteria in your gut. The label tells you almost nothing about the actual risk. You have to look at what's inside, not just the classification.

  1. In the most controlled study ever done, people eating ultra-processed food spontaneously ate 500 extra calories per day — even when meals were matched for protein, fat, carbs, and fiber. Their fullness signals simply didn't fire.
  2. When researchers broke the "ultra-processed" category apart, whole-grain breads and plant-based alternatives carried zero elevated risk. Sugary drinks and processed meats carried almost all of it.
  3. When overall diet quality was accounted for, ultra-processed food on its own showed no consistent death-rate signal — it's what UPFs push out of your diet that drives much of the harm.

Want the full evidence? Keep scrolling

What Most People Think

The "Processed = Bad" Frame

The Common Belief

"Processed equals bad" is the dominant frame. People avoid ultra-processed foods because the word itself sounds dangerous — chemically altered, nutritionally stripped, body-confusing. If it comes in a wrapper with an ingredient list you can't pronounce, it must be harming you.

Where It Leads

The logical conclusion: clean eating is mandatory for health, IIFYM (flexible dieting) is risky, and food processing itself is an independent toxin. The advice that follows is usually blanket restriction of anything that doesn't look like it came straight from a farm.

What's Missing

This frame treats a label as a risk factor. It doesn't ask: which specific processed foods carry the risk? Is it the processing itself, or what happens to the food's structure? Or is it simply that ultra-processed food pushes better food off your plate?

What the Evidence Actually Shows

Five Findings That Reshape the Picture

Ultra-processed food overrides your fullness signals — even when macros are matched STRONG

Hall et al. 2019 ran the gold-standard test: a metabolic ward study (n=20) where both diets were matched for calories, protein, fat, carbs, sugar, sodium, and fiber. The only difference was whether the food was ultra-processed or not.

Result: people on the ultra-processed diet spontaneously ate 508 extra calories per day. They ate about 17 calories per minute faster. They gained 0.9 kg in two weeks. The group eating whole foods lost 0.9 kg.

This is the most important study in this domain. It rules out nutrient composition as the driver. Something about the food's structure — the way it's been processed — short-circuits fullness.

The mechanism: food that doesn't need chewing bypasses your hunger brakes STRONG

Here's what's actually happening. When food structure is degraded by processing, it requires minimal chewing. You eat faster. But your fullness hormones — the signals that tell your brain "stop eating" — take time to kick in. By the time they fire, you've already overeaten.

On top of that, intact fiber normally ferments in your gut and produces short-chain fatty acids. These fatty acids trigger longer-lasting fullness signals through receptors in your gut lining. Processed food strips this fiber structure, so those secondary fullness signals never arrive.

The "ultra-processed" label is too broad to be useful STRONG

Cordova 2023 studied 266,666 people across 7 European countries for 11 years. When they broke "ultra-processed" into sub-categories, the picture changed dramatically.

Industrial whole-grain breads? Hazard ratio 0.97 — no elevated risk. Plant-based alternatives? Same: 0.97. Sugary drinks and processed meats? That's where the signal lived, with hazard ratios around 1.09. A bag of chips and a whole-grain loaf are not the same thing, even though NOVA calls them both Group 4.

The death-rate link is real, but partly disappears when diet quality is controlled MODERATE

A 2025 meta-analysis of 1.1 million people (173,000 deaths) found a 10% increase in mortality risk for every 10% increase in ultra-processed food calories. That sounds alarming.

But a separate 30-year study of 114,064 people (Fang 2024, published in the BMJ) found that when they adjusted for overall diet quality, the ultra-processed food signal substantially weakened. Within each diet quality bracket, ultra-processed food alone showed no consistent mortality signal.

Translation: ultra-processed food harms you partly because it pushes better food off your plate — not purely because processing itself is toxic.

Causal disease pathways confirmed — but only for specific foods MODERATE

Using a technique called Mendelian randomization (which uses genetics to establish cause-and-effect), researchers found that high processed meat intake causally increases the risk of Crohn's disease (OR = 3.7). They also found that DHA (an omega-3 fat) deficiency mediates 17% of the ultra-processed food connection to Crohn's — suggesting UPFs harm you partly by displacing omega-3-rich foods.

These are genuine causal signals. But they're for specific foods causing specific diseases — not for "ultra-processed food" as a general category causing general harm.

The Debate

Is Ultra-Processed Food Independently Deadly?

The Core Disagreement

Liang 2025 meta-analysis (N=1.1M, 173K deaths)

Real dose-response relationship: every 10% increase in ultra-processed food calories raises mortality risk by 10%. The signal is consistent across multiple large cohorts. Ultra-processed food is independently harmful at a population level.

VS

Fang 2024, BMJ (N=114,064, 30+ years)

When you adjust for overall diet quality, the ultra-processed food mortality signal substantially weakens. Within each diet quality bracket, ultra-processed food alone shows no consistent mortality signal. The harm is confounded by poor diet.

Both sides are partly right. Ultra-processed food harm is real — but it's smaller than headlines suggest, and concentrated in specific sub-categories (sugary drinks, processed meats), not spread evenly across everything NOVA calls "Group 4." Much of the mortality signal is really a proxy for eating a poor diet overall.

Honest Limitations

What We Still Don't Know

The Gold-Standard Study Has a Small Sample

In the lab: Hall 2019 metabolic ward RCT is the most controlled study ever done on this topic. Near-definitive evidence that ultra-processed food overrides fullness signals.
In reality: Only 20 participants. Mean age 31. Mean BMI 27. A Japanese replication showed an even larger effect (+813 kcal/day) but with only 9 people. We're highly confident in the mechanism — but a large free-living study hasn't confirmed it yet.
More Conservative

Sub-Category Analysis Is Observational

In the lab: Cordova 2023 EPIC cohort (N=266,666) broke ultra-processed food into sub-categories and found risk concentrated in sugary drinks and processed meats.
In reality: This is observational cohort data. It can't prove that whole-grain bread is safe or that sugary drinks are the causal agent. People who eat industrial whole-grain bread may differ from soda drinkers in dozens of unmeasured ways.
More Conservative

Causal Pathways Are Disease-Specific

In the lab: Mendelian randomization confirmed DHA deficiency as a causal mediator of the ultra-processed food to Crohn's disease connection.
In reality: This causal signal is for one specific nutrient deficiency (DHA) and one specific disease (Crohn's). Generalising it to all ultra-processed food associations would be premature.
More Conservative

The Practical Takeaway

What to Actually Do About It

  1. Target the sub-categories, not the label. Sugary drinks and processed meats are where the evidence converges. Industrial whole-grain breads, fortified cereals, and plant-based alternatives don't carry the same risk. Stop treating a whole-grain bread the same as a bag of chips just because they're both "ultra-processed."
  2. If you're tracking calories, treat ultra-processed food as a fullness problem. It's not just about the numbers on the label. Ultra-processed food actively undermines your ability to control how much you eat — you eat faster, your fullness signals fire slower, and you overshoot your target. Even when the macros look identical on paper.
  3. Flexible dieting works — if you can actually enforce it. In a controlled setting where someone else measures your food, ultra-processed food harm is minimised when macros are matched. But in real life, ultra-processed food undermines the self-regulation you need to hit those targets in the first place. Honesty about your own tracking accuracy matters here.
  4. Don't fear high-fiber packaged foods. High-fiber cereals and whole-grain commercial breads contain intact or partially intact fiber. They feed the bacteria in your gut that produce protective compounds. They carry neutral-to-protective signals in population studies. Blanket-restricting these is throwing out good food for a bad reason.

The Nuance

What the Simple Answer Misses

The Classification Itself Is Broken

The NOVA classification has real problems as a research tool. A bag of chips and a fortified whole-grain cereal are both NOVA Group 4, but they're mechanistically different — one degrades your food structure and overrides fullness, the other provides intact fiber and micronutrients.

Research that treats them identically produces noisy, attenuated risk estimates. The actionable signal doesn't live at the NOVA Group 4 level — it lives in the sub-categories.

Displacement, Not Toxicity

Some of the harm attributed to ultra-processed food is actually nutrient displacement. Ultra-processed foods push omega-3-rich fish, vegetables, and whole foods off your plate. The resulting deficiencies — like the DHA gap that mediates 17% of the Crohn's disease connection — drive specific disease pathways.

Fix the displacement (eat the good stuff alongside the processed stuff) and you reduce that mechanism. The processing isn't poisoning you — it's crowding out what your body actually needs.

The Evidence Gap

The Hall 2019 metabolic ward RCT is the strongest piece of evidence in this domain. But it studied 20 people with an average age of 31, in a controlled ward. A Japanese replication with 9 people showed an even bigger effect (+813 calories per day), which is directionally consistent but below strict sample size thresholds.

We have high confidence in the mechanism. What we don't yet have is a large-scale, free-living study confirming it plays out the same way when people go about their normal lives, with all the messiness that entails.

Moderate Conviction

What would change this: A large-scale free-living RCT (n>200) with full nutritional quality compensation that either eliminates or confirms the mortality signal independently of diet quality. Sub-category Mendelian randomization at scale — beyond Crohn's — would sharpen specific food-level risk estimates and could move individual sub-categories to HIGH conviction.

Sources

Key References

  1. Hall KD et al. (2019). Cell Metabolism. Metabolic ward RCT, n=20. Ultra-processed vs unprocessed diets matched for calories, macros, sugar, sodium, and fiber. UPF group spontaneously ate +508 kcal/day and gained 0.9 kg in 2 weeks. Unprocessed group lost 0.9 kg.
  2. Cordova R et al. (2023). Lancet Regional Health Europe. EPIC cohort, N=266,666, 7 European countries, 11-year follow-up. Broke ultra-processed food into sub-categories. Industrial whole-grain breads (HR=0.97) and plant-based alternatives (HR=0.97) carry no elevated risk. SSBs and processed meats: HR ~1.09.
  3. Fang Z et al. (2024). BMJ. Nurses' Health Study + HPFS, N=114,064, 30+ years follow-up. Adjusting for AHEI dietary quality score substantially attenuated the ultra-processed food mortality association. Within each AHEI quartile, UPF alone showed no consistent signal.
  4. Liang S et al. (2025). Systematic Reviews meta-analysis. N=1,148,387 participants, 173,107 deaths. Dose-response analysis: HR 1.10 per 10% increment in ultra-processed food caloric share.
  5. Wang et al. (2026). MedRxiv (ONE-IBD Study). Mendelian randomization, UK Biobank n=91,306 validation. DHA deficiency mediates 17.1% of UPF-Crohn's association. Processed meat causally increases Crohn's risk (OR=3.696).

Want coaching that actually uses this evidence? SLH Fit — evidence-based physique coaching.

Verdict Score

How strong is the evidence for the claims in this review? Higher = more confidence the claims are supported. This does not measure how large the effect is or how important it is compared with other levers.

78 Mixed evidence
80–100Strong evidence
60–79Mixed but supportive ◀
40–59Uncertain
0–39Weak support

Where this sits — Nutrition That Actually Matters

Approximate contribution to this goal, based on effect sizes from intervention research. These are practical estimates, not exact causal percentages.

Leverage confidence: Moderate

Total Caloric Balance
~45%
Diet Quality (Whole Foods > Processed) ←
~25%
Adequate Protein
~15%
Fiber Intake
~10%
Meal Timing
~5%
Specific Oil Avoidance (Seed Oils)
<1%
Individual Food Demonization
<1%
Foundation

Reality Check

Contribution: ~25% of the outcome
Bigger levers: Total Caloric Balance
Time investment: Daily

Get weekly verdicts — no fluff, just evidence

Conviction-scored health research in your inbox. What works, what doesn't, and what the studies actually measured.

Subscribe free

Related free research

Metabolic Health
Seed Oils — Evidence vs Controversy
 Metabolic Health
Organic Vs Conventional Is The Premium Worth It
 Metabolic Health
Post-Workout Nutrition — What Matters, What Doesn't

There are 424 more inside

Conviction-scored verdicts on supplements, nutrition, training, physio, and recovery.

Explore all Get weekly verdicts