Mediterranean diet: what the evidence says and at what level of certainty
Mediterranean diet: what the evidence says and at what level of certainty, and which part of the pattern is what carries the result.
The Mediterranean diet is probably the most studied eating pattern in the world. That has an advantage and a drawback: there is a lot of evidence, and there are also many reviews of differing quality that reach different conclusions.
What an umbrella review goes over
An umbrella review is not a new study: it is a review that reviews systematic reviews and meta-analyses, in order to avoid counting the same trials twice. The one summarizing here the evidence on the Mediterranean diet and cardiovascular disease included 18 meta-analyses covering 238 randomized trials, with only 8% overlap between the primary studies [1].
The results, with their uncertainty
| Outcome | Comparison | GRADE certainty |
|---|---|---|
| Cardiovascular mortality | vs usual care | Low |
| Non-fatal heart attack | vs another active intervention | Not specified in the abstract |
In plain language: the Mediterranean diet was associated with less cardiovascular mortality compared with doing nothing, and with fewer non-fatal heart attacks compared with another active dietary intervention. In relative terms, the authors estimate it would reduce the risk of fatal cardiovascular outcomes by between 10% and 67%, and of non-fatal outcomes by between 21% and 70% [1].
Two nuances that are in the article itself and not in the news:
- The certainty of the evidence on mortality is low, according to GRADE [1]. That means the real effect is likely to be somewhere in that wide range, not that the central number is precise.
- The methodological quality of most of the reviews reviewed was low or critically low (16 out of 18, 84%) [1].
Why those numbers are so wide
For the effect range to go from a 10% reduction to a 65% one is not a statistical detail: it is the warning that we do not know how much of that effect is real. A range that wide does not allow you to claim that “the Mediterranean diet reduces cardiovascular mortality by 40%,” even if that number comes from
What a Mediterranean diet is in terms of food
Beyond the studies, the WHO describes the principles these kinds of patterns share: a variety of minimally processed or unprocessed foods, low in unhealthy fats, free sugars and sodium, and with plenty of whole grains, vegetables, fruit and legumes [2].
The IOM recognizes for the adult population an acceptable distribution range for carbohydrate and fat, and places within that range the emphasis on whole grains, vegetables, fruit and pulses (legumes) [3]. Its percentages are not reproduced here for the reason explained in the note below. And legumes have evidence of their own: a systematic review of legume consumption and cardiometabolic health found a significant inverse association between total legume consumption and the risk of cardiovascular disease and coronary heart disease [4].
An honest limit
One specific limitation of this article: the 2005 Dietary Reference Intakes is usually cited as the one setting the carbohydrate split as a percentage of energy [3]. We do not reproduce that figure because we could not open the report at its primary source (the National Academies viewer returns the page index, not the table) and this site does not publish figures it has not checked. The message of the article does not change: the Mediterranean diet is a pattern, and what the evidence supports is the structure (vegetables, fruit, legumes and whole grains, less processing), not a calorie split into percentages. If someone opens the source and confirms the table, the figure comes back with its citation. The Mediterranean diet is a pattern, not a specific food nor a restrictive diet: you can follow it at home, at work or in a restaurant, and it admits regional variants. What the evidence supports is the pattern as a whole, not any brand or any single ingredient.
This article describes what the systematic reviews say about the adult population (18 years and older), with and without established cardiovascular disease. If you have cardiovascular disease, you are pregnant, or you are on treatment, the conversation with your healthcare professional comes first.
Why “like in Italy” is not enough
A methodological detail that tends to get missed: in these trials, “Mediterranean diet” rarely means a single recipe. It is usually a pattern defined by proportions and by food frequency, assessed with questionnaires or scoring scales, and assigned to a group that receives counseling. That has two consequences:
- What works is probably the whole set, not one specific food: nobody has tested olive oil, legumes or fish separately.
- These trials are almost all about prevention or with limited follow-up, and the effect on mortality is measured in years, not weeks. A pattern that holds over years cannot be measured in a short trial, and that is precisely the limitation that runs through all the evidence on long-term diet.
That is why the practical recommendation that can be sustained is the one the guidelines already state: eat more vegetables, fruit, legumes and whole grains; less ultra-processed food, less salt and less saturated fat [2]. It is easier to maintain than “eating a Mediterranean diet” and it is the part of the advice with the strongest support.
What you take away
- An umbrella review of 18 meta-analyses and 238 trials finds less cardiovascular mortality and fewer heart attacks with the Mediterranean diet [1].
- But the certainty of that evidence is low and the quality of the reviews reviewed is generally low, so the exact size of the effect is not well established [1].
- The structural advice—legumes, whole grains, vegetables, fruit, little processing—is supported both by the public health guidelines [2][3] and by the specific evidence on legumes [4].
More from us: the Nutrition basics, with the figures and sources of the Nutrition section.
Related: what to eat to lose weight, which compares diet composition with calorie counting, and safe calorie deficit. multiplying two figures.
The authors say it plainly in their conclusion: better-quality reviews are needed [1]. And the effect was more marked in studies with populations that already had established cardiovascular disease, that is, in secondary prevention [1]. The authors assessed the quality of each review with AMSTAR-2 and extracted the certainty of the evidence with GRADE, the two standard tools in the field. Most of these teams do not stop at counting how many trials there are; they judge how much you can trust one.
Related articles
Sources for this article (4) 4 references
- Hareer LW, Lau YY, Mole F, Reidlinger DP (2025) The effectiveness of the Mediterranean Diet for primary and secondary prevention of cardiovascular disease: An umbrella review. Nutrition & Dietetics 82(2). See source (opens in a new tab) DOI: 10.1111/1747-0080.12891 Revisión paraguas (umbrella review) de la dieta mediterránea Licence: Artículo de acceso abierto (texto completo en PubMed Central, CC BY 4.0). Se cita con enlace. Consulted:
- Organización Mundial de la Salud (OMS) (2026) Healthy diet — hoja de datos. Organización Mundial de la Salud (OMS). See source (opens in a new tab) Hoja de datos de la OMS (recomendaciones de salud pública) Licence: OMS: el contenido de esta web está disponible bajo licencia CC BY-NC-SA 3.0 IGO. Se cita con enlace. Consulted:
- Institute of Medicine (IOM, National Academies) (2005) Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. The National Academies Press, Washington, DC. See source (opens in a new tab) Informe de consenso del Institute of Medicine (DRI de macronutrientes y fibra) Licence: Obra del Gobierno de los Estados Unidos (dominio público). Descarga gratuita en nap.edu. Consulted:
- Becerra-Tomás N, Papandreou C, Salas-Salvadó J, Bulló M, Viñas B, Estruch R, et al. (2019) Legume Consumption and Cardiometabolic Health. Advances in Nutrition 10(6). See source (opens in a new tab) DOI: 10.1093/advances/nmz003 Revisión narrativa con metaanálisis (consumo de legumbres y salud cardiometabólica) Licence: Artículo de revista disponible en PubMed Central, con derechos de autor (sin licencia libre). Se cita con enlace; no se reproduce texto salvo cita breve verificable. Consulted:
How to cite
Pan y Hierro (2026). Mediterranean diet: what the evidence says and at what level of certainty. https://panyhierro.com/en/blog/mediterranean-diet/. Updated October 12, 2026.
Citation copied