PREDIMED trial

PREDIMED is the Spanish trial behind most claims that a Mediterranean diet prevents heart attacks and strokes, and one of the few large diet trials that counted real events rather than blood markers. Its roughly 30% reduction survived a retraction and re-analysis, but it is no longer a clean randomised trial, and in a major 2024 review of the Mediterranean diet it was the only randomised trial.

Read PREDIMED as good, not clean, evidence that the whole Mediterranean pattern, with generous olive oil or nuts, lowers cardiovascular events in people at high risk. It does not show that olive oil, nuts or polyphenols work on their own.

What the trial did

PREDIMED (Prevención con Dieta Mediterránea) assigned 7,447 Spanish adults at high cardiovascular risk to one of three diets: a Mediterranean diet with extra-virgin olive oil supplied, the same diet with nuts supplied, or a low-fat control diet. The olive-oil protocol called for at least four tablespoons a day, about 50 g.

Both Mediterranean arms had roughly 30% fewer heart attacks, strokes and cardiovascular deaths than the low-fat arm.^[1] This was a prevention trial: participants had not yet had a heart attack or stroke.

The retraction, and what survived it

The original 2013 paper was retracted and republished in 2018 after the investigators found protocol deviations: for 1,588 of the 7,447 participants, about one in five, group assignment was known or suspected to have broken protocol. Members of the same household were enrolled together without being randomised, one of 11 sites assigned some participants without randomising, and another used its randomisation tables inconsistently.

The re-analysis, including one that left those participants out, found about the same benefit, so the roughly 30% held. But the trial can no longer be read as a clean comparison of individually randomised people; treat it as good but imperfect evidence.

Who paid for it

The funding was public, from the Spanish health ministry's research institute. The olive oil and nuts were donated by producers — Spanish olive-oil, almond and hazelnut firms and the California Walnut Commission — none of which, the paper states, had any role in the design, analysis or reporting. A real conflict of interest to know about, but not industry funding of the trial itself.

Why one trial carries so much weight

Randomised evidence for the Mediterranean diet can look broader than it is. A 2024 review of 28 studies in older adults included two randomised studies, and both were PREDIMED.^[2] The rest of the case is observational.

The trial also tested a whole diet against a low-fat diet, so it cannot say which part did the work. A later analysis of its participants tried to pin it on polyphenols, plant compounds found in many foods including extra-virgin olive oil: people in the top fifth of total polyphenol intake had 37% lower mortality than the bottom fifth.^[3] Three things weaken it. It compared people by what they reported eating, not by what they were randomised to; it rested on 327 deaths with no clear dose-response trend; and the link held only in the low-fat control arm, not in either Mediterranean arm. The wider olive-oil evidence is on Olive oil.

The follow-up trial: PREDIMED-Plus

PREDIMED-Plus tests something different: a calorie-reduced Mediterranean diet plus physical activity and behavioural support, against Mediterranean-diet advice alone. In its first year, in a subsample of 626 adults with overweight and metabolic syndrome, the intervention group lost 2.5 kg more on average, and 33.7% against 11.9% lost at least 5% of their weight.^[4] The weight loss belongs to the whole programme, not the diet alone.

Over a median of six years, the intervention group had 31% fewer new cases of type 2 diabetes, 9.5% against 12.0%.^[5] Because both groups were advised to eat Mediterranean-style, this shows the value of eating less and moving more on top of the diet, not of the diet itself. The trial's main result — heart attacks, strokes and cardiovascular deaths — has not yet been published.

What to do

Take the ~30% as the best single estimate for the Mediterranean pattern, in people at high cardiovascular risk, with the retraction in mind.
Eat the whole pattern, including the olive oil or nuts the trial supplied, rather than adding one ingredient to an otherwise unchanged diet.
If you need to lose weight, add fewer calories and more activity to it — that, not the diet alone, is what PREDIMED-Plus tested. The full pattern is on Dietary patterns.