The Mediterranean diet's brain benefits mostly vanish in randomised trials

Observational studies have long suggested that people who eat a Mediterranean diet have better cognition and perhaps a lower risk of dementia. But people who eat that way tend to differ in many other respects, so only randomised trials can show whether the diet itself changes the brain. Until this review, nobody had systematically gathered the randomised evidence on cognition and on brain morphology and function. The authors also asked a subtler question: how faithful were the trial diets to the traditional Mediterranean pattern they claim to be testing?

The reviewers searched nine databases from inception to December 2017 for randomised controlled trials comparing a Mediterranean diet against an alternate diet, with cognitive or brain morphology and function outcomes. Five trials qualified, covering 1888 participants, and between them they reported 66 individual cognitive tests plus one brain function outcome. The prescribed Mediterranean diets varied considerably from study to study, particularly in their quantitative food advice. Data limitations meant the results could not be pooled into a meta-analysis, so findings were assessed trial by trial.

Only 8 of the 66 individual cognitive outcomes, or 12.1%, significantly favoured the Mediterranean diet, with effect sizes ranging from 0.32 to 1.66, while 2 outcomes favoured the control diets. Of 8 domain composite scores from 2 studies, 3 from the PREDIMED trial were significant, covering memory, frontal and global function, with effect sizes from 0.39 to 1.29. In the same trial, the risk of low plasma brain-derived neurotrophic factor was reduced by 78% at 3 years on the Mediterranean diet, and a genotype analysis at a second site produced mixed results. There was no benefit for incident cognitive impairment or dementia.

This review complicates a popular story: the observational signal for the Mediterranean diet and brain health is strong, but the randomised evidence is mostly nonsignificant, with small effects. The picture is not simply negative, because the most robustly designed study showed significant improvements in cognitive domain composites, which the authors say warrants additional research. The wide variation in what each trial actually prescribed as a Mediterranean diet also makes the studies hard to compare. Five trials remain a thin base for a question this large.

If you follow a Mediterranean pattern partly for your brain, nothing here suggests it harms you, and the strongest trial offers genuine encouragement on memory and overall function. What the evidence does not yet support is confidence that the diet prevents cognitive decline or dementia, since the trials found no benefit on those outcomes. This is a reason to hold the brain claims more loosely than the headlines do, while recognising that the trial evidence is still too thin and inconsistent for any final verdict.