Same healthy-eating guidance, two versions: the less processed one lost more weight

Almost everything known about ultra-processed food and health comes from observational studies: people who eat more of it have worse outcomes, and the arguments are about what else is going on. A trial published in Nature Medicine on 4 August tested something narrower and more answerable — whether the same national dietary guidance, followed with mostly ultra-processed foods or with mostly minimally processed ones, produces different results [s1].

The answer was yes, by a small margin.

The design

Fifty-five adults in England took part, all with a body mass index of at least 25 and below 40, and all habitually getting at least 50% of their daily calories from ultra-processed food [s1]. Each was provided with two eight-week diets in a randomised 2 × 2 crossover: one built from minimally processed food, one from ultra-processed food, both following the UK Eatwell Guide, and both eaten ad libitum — participants ate as much or as little as they wanted [s1].

Twenty-eight were randomised to the minimally processed diet first and 27 to the ultra-processed diet first; 50 participants made up the intention-to-treat sample [s1]. The primary outcome was the within-participant difference in percentage weight change between the two diets from baseline to week 8, and participants were blinded to it [s1].

The design point worth dwelling on is that both arms followed healthy-eating guidance. This is not a test of a junk-food diet against a wholefood diet. It is a test of whether processing itself matters once the nutritional profile has been held to the same national standard.

What happened

Both diets produced weight loss. On the minimally processed diet, mean percentage weight change was −2.06% (95% CI, −2.99 to −1.13); on the ultra-processed diet it was −1.05% (95% CI, −1.98 to −0.13) [s1].

The difference between them was −1.01 percentage points (95% CI, −1.87 to −0.14; P = 0.024), with a Cohen's d of −0.48 (95% CI, −0.91 to −0.06) [s1]. Mild gastrointestinal adverse events were common on both diets [s1].

The authors' conclusion is that the findings indicate greater weight loss on the minimally processed diet, and that dietary guidance on food processing is needed in addition to existing recommendations [s1].

How to size that result

A one-percentage-point difference in body weight over eight weeks is small on any reading. The confidence interval reaches almost to zero at its upper bound (−0.14) [s1], which means the true difference could be very small indeed.

The effect size, Cohen's d of −0.48, sits in the range conventionally described as small-to-moderate, and its interval likewise reaches close to zero [s1].

Set against that: both diets caused weight loss without calorie restriction, in people who habitually got more than half their calories from ultra-processed food [s1]. Simply following the Eatwell Guide moved the needle regardless of processing. The processing question is a second-order effect layered on a first-order one.

Fifty-five participants is a small trial, and eight weeks per arm is short. It is also, as the authors note, the first trial to assess the health impact of ultra-processed food consumption within the context of national dietary guidelines [s1] — which is to say the evidence base it belongs to is very new.

The scale of the exposure

The trial's relevance depends on how much ultra-processed food people actually eat, and a National Center for Health Statistics data brief published on 7 August gives the current US figures [s2]. Using National Cancer Institute food categorisations linked to NHANES dietary recall data, it reports that during August 2021 to August 2023, youth aged 1 to 18 got 61.9% of their calories from ultra-processed foods and adults aged 19 and over got 53.0% [s2].

The top sources overlap almost entirely across ages: sandwiches (including burgers), sweet bakery products, savoury snacks and sweetened beverages were four of the top five categories for both youth and adults [s2].

One trend in the brief cuts against the general direction of the conversation. Between 2013-2014 and August 2021-August 2023, mean calories from ultra-processed foods among adults decreased [s2]. The brief does not report a corresponding decrease among youth.

What is still unresolved

The trial cannot say why the minimally processed diet produced more weight loss. Candidate explanations — energy density, eating rate, satiety signalling, food matrix effects — are not distinguished by this design, and the paper does not claim to distinguish them.

Nor does it establish anything about cardiometabolic outcomes over a meaningful horizon. Sixteen weeks of feeding, split across two arms, is a window in which weight moves and long-term disease risk does not become visible.

What the trial does establish is narrow and worth stating precisely: within one country's healthy-eating guidance, eaten freely, the less processed version produced modestly more weight loss than the more processed version in a small group of adults with overweight or obesity who normally eat a lot of ultra-processed food. Whether that justifies adding processing to national guidance is a policy judgement the trial informs rather than settles.

Sources

• Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial — Nature Medicine, 2025-08-04
• Ultra-processed Food Consumption in Youth and Adults: United States, August 2021-August 2023 — NCHS Data Brief, National Center for Health Statistics, 2025-08-07