Counts describe this library, not the strength of evidence.
How to read this evidence
Comparators explain many apparent contradictions. The 2022 trial tested adding a window to calorie restriction; the 2026 synthesis compared timing and duration across trials. Neither supports assuming the shortest window is best. The corrected PREDIMED report also illustrates why publication history matters.
Among 49,261 Swedish women followed for an average of 20.4 years, high adherence to a Mediterranean diet was linked to a lower risk of clinically diagnosed depression than low adherence (HR 0.82, 95% CI 0.71-0.94). Diet was reported once in 1991/92, and as an observational finding it cannot show that the diet itself prevents depression.
The international journal of behavioral nutrition and physical activity
Stool metagenomics in 56 pairs of people with irritable bowel syndrome (IBS) and household controls found that 50% of IBS cases had a 'pathogenic-like' microbiome (IBS-P) and that this group responded better clinically to 4 weeks of a low FODMAP diet (p = 0.02). The diet follow-up covered only 41 pairs, with no comparison group on a normal diet, so the result is a hypothesis for further testing.
In 33 observational studies, time in range rose by 3.75% (95% CI 2.56 to 4.92) during lockdown in people with type 1 diabetes, while type 2 diabetes showed a short-term, non-significant drift upward in HbA1c (+0.14; 95% CI -0.13 to 0.40). Type 2 evidence came from only 8 studies, and HbA1c in type 1 did not change significantly.
In national survey data on 20 905 children and 39 757 adults, school food showed the biggest gain, with poor-quality food falling from 55.6% to 24.4% (P < .001 for trend), while food from other sources worsened. Diet is based on one-time recalls and food source was not linked to health outcomes, so this describes food quality trends and not health effects.
Among 15,249 US adults, healthier scores on five diet indices were associated with frailty, and with lower 8-year mortality, for example HEI-2015 hazard ratio 0.93 per 10 points (95% CI 0.89-0.97). These are observational links from a one-time dietary assessment, so diet may be partly a marker of other differences between people.
In a crossover trial of 26 people with Alzheimer's disease, a 12-week modified ketogenic diet raised daily-function scores (ADCS-ADL +3.13 points, P = 0.0067) and quality-of-life scores (QOL-AD +3.37, P = 0.023) versus usual diet, but not cognition (ACE-III +2.12, P = 0.24). The trial was very small, and the abstract does not report any long-term follow-up.
In 74 people with fatty liver (NAFLD), 12 weeks of a 5:2 diet or a low-carb high-fat diet reduced liver fat by 6.1% and 7.2% in absolute terms, versus 3.6% with general lifestyle advice. The trial was open-label and small, so it cannot tell us whether the benefit lasts beyond 12 weeks.
Pooling 12 papers, a low-FODMAP diet lowered irritable bowel syndrome severity more than control diets (standardized mean difference -0.66, 95% CI -0.88 to -0.44) and raised quality-of-life scores. The trials were few and varied in design, and the abstract leaves long-term nutrition and gut-microbiome effects unanswered.
Mendelian randomization using genetic predisposition to type 2 diabetes and higher HbA1c found associations with higher risk of any ischemic stroke, large artery stroke and small vessel stroke, plus markers of brain small-vessel disease and atrophy. The abstract gives no effect sizes, and the method relies on assumptions about genetic variants that cannot be fully verified.
In an 8-week pilot trial of adults with obesity, 14:10 time restricted eating plus a commercial weight program cut body weight by 8.5% versus 7.1% with 12:12 (difference -1.4%, 95% CI -2.7 to -0.2). Only 60 of 78 participants finished, and the comparator also lost weight, so the added effect is small and uncertain.
Pooling before-and-after changes in people with type 2 diabetes, a ketogenic diet was followed by lower fasting glucose (down 1.29 mmol/L, 95% CI -1.78 to -0.79) and lower HbA1c (down 1.07, 95% CI -1.37 to -0.78). Without control groups in the pooled estimates, the abstract cannot separate the diet's effect from other changes over time.
In a 16-week randomized trial of 244 overweight adults in Washington, DC, body weight fell by 5.9 kg (95% CI 5.0-6.7 kg; P < .001) on a low-fat vegan diet, with improved insulin sensitivity and less liver and muscle fat; none of these changed significantly in the control group. Participants knew their group and were mostly women (87%), so results may not apply to everyone.
In 805 coronary heart disease patients randomized to a Mediterranean or low-fat diet, flow-mediated dilation (a test of blood vessel function) rose 3.83% on the Mediterranean diet versus 1.16% on low-fat, a difference of 2.63% (95% CI 1.89-3.40, p = 0.011). This was a secondary, exploratory analysis of a larger trial, so false positives are possible.
An umbrella review of 60 systematic reviews and meta-analyses found that Mediterranean and DASH diets, whole grains, yogurt and olive oil were linked to lower type 2 diabetes risk, while high glycemic load diets, red and processed meat and sweetened drinks were linked to higher risk. The summary was qualitative, with no pooled numbers given in the abstract.
This narrative review concludes that diet, obesity, stress and mental disorders such as depression and posttraumatic stress disorder (PTSD) are linked in probably two-way ways, but causal pathways have not been established. It reports no pooled numbers and does not say how studies were chosen, so it maps ideas rather than measuring effects.
Across 117 studies with 3,202,496 participants, the highest versus lowest Mediterranean diet adherence was linked to lower cancer mortality (RR 0.87, 95% CI 0.82-0.92) and lower risk of several site-specific cancers. Most comparisons were rated low or very low certainty, and observational studies cannot show that the diet itself caused the difference.
In 14 overweight women with polycystic ovary syndrome (PCOS), 12 weeks on a ketogenic Mediterranean diet with phytoextracts was followed by a 9.43 kg drop in body weight and improvements in insulin, cholesterol and sex hormones. There was no control group and only 14 women, so the diet's own effect is unproven.
Among 475,581 UK Biobank adults followed an average of 5.7 years, eating 76 g/day of red and processed meat versus 21 g/day was linked to 20% higher colorectal cancer risk (95% CI 4-37), and each extra 10 g/day of alcohol to 8% higher risk. Diet came from a short food-frequency questionnaire, so these are associations that may be affected by diet reporting error.