Human Trials and Meta-Analyses of Foods: What Exists and Limits
What human trials and meta-analyses of foods and eating patterns exist, with real examples, and the limits that decide how much weight a pooled result deserves.
Content last updated . This page has not been medically reviewed.
Safety first. Supplements can cause side effects and interact with medicines. Read mushroom safety and interactions before you use any mushroom product.
The question
When a headline says "a meta-analysis of trials shows", what has actually been pooled, and how far can the result be trusted? This page uses real examples from the sources we read. It is information, not advice about what to eat.How the sources define the terms
- Clinical trial. A study in which participants "are assigned to receive one or more interventions" so researchers can evaluate effects on health-related outcomes. [1]
- Randomized trial. Treatments are "selected by a random process". [1]
- Meta-analysis. A method that combines the results of several studies into one estimate. The Cochrane saturated-fat review, for example, used random-effects meta-analyses and only counted randomized trials lasting at least 24 months. [2]
Food trials come in several kinds. Researchers may provide all the food, give counseling or advice, or add one food to the usual diet. The Cochrane saturated-fat review included trials with interventions ranging "from providing all food to advice on reducing saturated fat". [2] Those are different tests, and a pooled number mixes them.
Examples of what exists
| Question | What the source found | Strength of evidence (as the source rated it) | Population and form studied |
|---|---|---|---|
| Does eating less saturated fat change heart events? | Combined cardiovascular events were reduced by 17% (risk ratio 0.83); little or no effect on all-cause mortality. [2] | Moderate quality for both results, per the review's GRADE rating | 15 trials, 56,675 adults; whole diets, not single foods |
| Does a Mediterranean-style diet prevent heart disease? | Modest benefits for some risk factors in primary prevention; uncertainty remains for clinical events. [3] | Low or moderate for the risk-factor findings | 30 trials, 12,461 people; advice, food provision or both |
| Do oat products change blood lipids? | Oat-based products may reduce total and LDL cholesterol, with little effect on triglycerides, HDL and major cardiovascular events. [4] | Moderate certainty for the cholesterol results; overall risk of bias in the trials was high or of some concern | 17 trials, 1,731 people with dyslipidemia; beta-glucan products, oat bran and whole oats |
| Do whole oats and isolated beta-glucan differ? | Both oat and isolated beta-glucan interventions improved lipid profiles in the pooled analysis. [5] | Not rated in the abstract | 28 studies, 1,494 people; blinded trials of oats or isolated beta-glucan against placebo or control |
| Does a MIND-style eating pattern protect thinking? | No significant difference from a control diet in cognition or brain MRI over 3 years. [6] | One randomized trial | 604 older adults with a family history of dementia, both groups on mild calorie restriction |
| Do "superfoods" help metabolic syndrome measures? | Only limited evidence, because results were inconsistent or trials were few. [7] | Limited, per the review | 113 trials across 17 foods |
The two oat reviews show why the form matters. One pooled beta-glucan products, bran and whole oats together, while the other analyzed whole oats and isolated beta-glucan separately. [4] [5] A result for an extract or isolated compound does not transfer to the food. See Oats and Beta-glucan in foods.
What a meta-analysis can and cannot fix
- It cannot fix weak trials. The oat review found the risk of bias in the included trials was high or raised some concerns, because of inadequate randomization, allocation concealment or analysis. [4]
- It cannot make unlike things alike. The saturated-fat review combined a variety of interventions and reported heterogeneity (I² of 67% for the main result). [2]
- It reflects what was measured. In the Cochrane review of Mediterranean-style diets, the trials in one primary-prevention comparison did not report clinical endpoints, only risk factors. [3]
- It inherits problems from the trials. The PREDIMED trial was retracted and re-analyzed after randomization concerns at two of 11 sites; the Cochrane review rated its evidence on deaths as low quality. [3]
- Absence of trials is not absence of effect. The review of superfood trials found no controlled trials on metabolic syndrome measures for açaí, hemp seed or bee pollen (searched to April 2017). [7]
Food, extract and supplement are different
Trials of an isolated compound, extract or supplement are not trials of the food, and the reverse is true as well. When you see a pooled result, check which form went into it. See Functional foods vs supplements vs medicine.Frequently asked questions
Is a meta-analysis the best evidence?
It can be, when it pools good randomized trials. The Cochrane saturated-fat review states its inclusion rules and rated the certainty of evidence with GRADE. [2] But pooled results are only as good as the trials, and the oat review found the risk of bias in its pooled trials was high or raised some concerns. [4]
Why do two reviews of the same food give different numbers?
They may include different forms (whole oats, bran, isolated beta-glucan), different people and different comparators. [5] [4]
Do food trials measure illness or blood markers?
Both exist. The Mediterranean-diet Cochrane review reports risk factors such as blood pressure and cholesterol, and separately clinical endpoints such as deaths and strokes. [3]
What happens when a trial turns out to have problems?
Authors can withdraw and re-analyze it, as the PREDIMED investigators did after finding protocol deviations. [8]
Where do I find what trials show for one food?
See the food's own page, such as Walnuts, or the Research by food table.
More details: evidence, limits and cautions
The details behind this page: cautions, limits of the evidence and where sources disagree. Open any note to read it.
Limits of this page
- We read the abstracts of these reviews, not the full texts, so subgroup details are not covered.
- These are examples, not a complete list of food trials. Proper evidence not found: we searched PubMed for a single source that scores the trial evidence of all foods on one scale, and did not find one.
Safety and who should ask a clinician
References
- National Library of Medicine. Publication Types (MeSH). Read 2026-10-22.
- Hooper L et al. Reduction in saturated fat intake for cardiovascular disease. Cochrane Database Syst Rev 2020. PMID 32827219. Read 2026-10-22 (abstract).
- Rees K et al. Mediterranean-style diet for the primary and secondary prevention of cardiovascular disease. Cochrane Database Syst Rev 2019. PMID 30864165. Read 2026-10-22 (abstract).
- Li A et al. Efficacy of oats in dyslipidemia: a systematic review and meta-analysis. Food Funct 2024. PMID 38441173. Read 2026-10-22 (abstract).
- Separate effects of whole oats and isolated beta-glucan on lipid profile: a systematic review and meta-analysis of randomized controlled trials. Clin Nutr ESPEN 2023. PMID 36657917. Read 2026-10-22 (abstract).
- Barnes LL et al. Trial of the MIND diet for prevention of cognitive decline in older persons. N Engl J Med 2023. PMID 37466280. Read 2026-10-22 (abstract).
- van den Driessche JJ, Plat J, Mensink RP. Effects of superfoods on risk factors of metabolic syndrome: a systematic review of human intervention trials. Food Funct 2018. PMID 29557436. Read 2026-10-22 (abstract).
- Estruch R et al. Primary prevention of cardiovascular disease with a Mediterranean diet supplemented with extra-virgin olive oil or nuts. N Engl J Med 2018 (republished, revised analysis). PMID 29897866. Read 2026-10-22 (abstract).