A Diet Trial Was Stopped Early Because the Control Group Was Doing Measurably Worse
PREDIMED tested a Mediterranean diet with added olive oil or nuts against low-fat advice. Diabetes incidence was 6.9 per cent with olive oil against 8.8 per cent on the low-fat arm, and the follow-up trial did considerably better.
Outspoken Digest Nutrition Desk
Saturday, August 15, 2026/3 min read
Nutrition science has a credibility problem, and it is mostly deserved. Most of what the public hears comes from observational studies that follow people who chose their own diets, which means the finding that broccoli eaters live longer may be telling you about broccoli or about the sort of person who eats broccoli.
PREDIMED is one of the exceptions, which is why it carries so much weight. It randomised people, gave them food, and followed what happened.
What did PREDIMED test?
Participants at high cardiovascular risk were randomly assigned to one of three groups: a Mediterranean diet supplemented with free extra virgin olive oil, a Mediterranean diet supplemented with mixed nuts at around 30 grams a day, or advice on a low-fat diet as the control.
Providing the olive oil and nuts free of charge is the detail that makes the trial work. Most diet trials fail because people do not follow the diet, and the two intervention groups here had a strong practical reason to comply.
In the diabetes analysis, 273 participants developed type 2 diabetes over the follow-up: 6.9 per cent of the Mediterranean diet with olive oil group, 7.4 per cent of the Mediterranean diet with nuts group, and 8.8 per cent of the low-fat control group.
In the earlier PREDIMED-Reus substudy, which followed 418 non-diabetic participants aged 55 to 80 over a median of four years, the pattern was the same.
Is that difference actually meaningful?
It is modest per person and substantial per population, and both framings are honest.
Going from 8.8 to 6.9 per cent is roughly a fifth fewer new cases of a lifelong condition, achieved by changing which fats people ate rather than by restriction. Nobody counted a calorie. Nobody was told to lose weight.
The follow-up trial, PREDIMED-Plus, pushed further by combining an energy-reduced Mediterranean diet with increased physical activity, and reported a 31 per cent lower relative risk of incident diabetes over six years compared with a traditional unrestricted Mediterranean diet, in people with metabolic syndrome and overweight or obesity.
That is the more useful number for most readers, and it carries a clear message: the Mediterranean pattern is good, and the Mediterranean pattern with portion control and movement is considerably better.
What is actually in this diet?
Less exotic than the name suggests, and much of it is already familiar in this region.
- Olive oil as the main fat, used generously rather than sparingly. This is the component that most distinguishes it from low-fat advice.
- Vegetables, legumes and fruit in quantity. Lentils, chickpeas and beans are central rather than incidental.
- Nuts, in a meaningful daily amount rather than as an occasional garnish.
- Fish regularly, poultry moderately, red and processed meat rarely.
- Whole grains rather than refined, and very little in the way of commercial baked goods and sweetened drinks.
Note what is absent. No carbohydrate ban, no fasting protocol, no elimination of food groups, and no branded products. It is a pattern rather than a regimen, which is why adherence in the trials was achievable and why the head-to-head diet comparisons keep converging on the person rather than the macro.
Does it translate to the Gulf?
Better than most imported dietary advice, with two adjustments.
The eastern Mediterranean is not a foreign cuisine here. Olive oil, lentils, chickpeas, burghul, leafy salads, yoghurt, grilled fish and fresh herbs are already regional staples. Fattoush, tabbouleh, mujaddara, moutabal and grilled fish are Mediterranean diet dishes under different names.
What has changed is the ratio. The traditional pattern has been progressively displaced by refined carbohydrate, fried food and sweetened drinks, and the health consequence is visible in the region's diabetes prevalence figures. This is less a foreign intervention than a return.
The two adjustments are portion size, since PREDIMED-Plus suggests that is where the additional benefit lives, and bread, since the quantity of refined flatbread eaten alongside these dishes can undo a good deal of the rest. Our piece on bread and flour swaps deals with that directly.
Why does this diet keep winning?
Probably because it is the only one people keep doing.
The most striking thing about the trial evidence is not the mechanism. It is adherence. People stayed on this pattern for years, in ordinary life, with families and social occasions and no particular sense of deprivation.
A diet with a slightly smaller theoretical benefit that is followed for a decade beats an optimal one abandoned in March. On that criterion this is not merely the best-evidenced dietary pattern for diabetes prevention. It is the most realistic one, and those two things are rarely the same.
Published in The Outspoken Digest
Editorial desk
Outspoken Digest Nutrition DeskDiet, food science and the evidence behind claims about what we eat.
Newsletter
The Digest, in your inbox
One edition, sent when it is ready. No noise, and your address is never passed on.
Read Next
More Nutrition →
Psyllium Husk Is the Dullest Thing in the Supplement Aisle and Almost the Only One That Works
Aug 20, 2026/5 min read
Sugar-Free Does Not Mean Carbohydrate-Free, and the Biscuit Knows That Even If You Do Not
Aug 19, 2026/3 min read
Eating the Same Meal in a Different Order Cut Post-Meal Glucose by More Than a Third
Aug 18, 2026/3 min read
A Protein Bar Is Ultra-Processed. New Trial Data Says That May Not Be the Point
Aug 17, 2026/3 min read