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.
This scoping review of 66 articles (screened from 17,613) found that AI dietary tools fall into image-based and motion-sensor-based types and are described as quicker and less prone to recall bias than traditional methods. The abstract gives no accuracy figures, and the authors themselves say more studies are needed to test accuracy and efficacy.
This narrative review concludes that machine learning, wearables, chatbots and image-recognition tools are changing dietary assessment, personalized nutrition and disease management. The abstract gives no study counts or accuracy figures, and its wording is promotional, so it cannot show how well any tool works.
Among 19,451 US adults, each standard deviation rise in the C-reactive protein-triglyceride glucose index (CTI) was linked to 1.357 times the risk of existing coronary heart disease (CHD). CHD was self-reported by questionnaire at one time point, so the index cannot show future prediction.
In 311,892 European adults followed for an average of 10.9 years, each 10% increment of daily food weight from ultra-processed food (UPF) was associated with 17% higher incident type 2 diabetes (95% CI 1.14-1.19). Diet was reported once at baseline and the UPF sub-groups behaved differently, so the effect of UPF is not uniform and not proven causal.
Among 19 188 US adults (kidney stone prevalence 9.6%), people without vigorous recreational activity showed a rising trend of kidney stones with more sitting, but the difference was not significant. A cross-sectional survey with self-reported sitting cannot show that sitting causes stones.
International journal of surgery (London, England)
In 2019 poor diet was linked to 1.74 million stroke deaths and 5.52 million years lived with disability worldwide; since 1990 counts rose (25.96% and 74.76%) while age-standardized rates fell (42.29% and 11.34%). These are modelled Global Burden of Disease estimates that attribute stroke to diet statistically, not measured individual cases.
Among 5297 US adults in NHANES, each higher unit of body roundness index (BRI) was associated with higher odds of gallstones (OR 1.16, 95% CI 1.12-1.20), and the top quartile had odds about 3 times the bottom (OR 3.04). This is a one-time survey, so it cannot show that a rounder body shape causes gallstones.
In 405 Chinese adults with overweight or obesity and early type 2 diabetes, 16 weeks of a 5:2 meal-replacement diet cut HbA1c by 1.9%, more than metformin (1.6%) or empagliflozin (1.5%). The trial was open-label, lasted only 16 weeks with an 8-week follow-up, and 332 participants completed treatment.
Across 37 studies (17 observational, 20 interventional), following a Mediterranean diet was associated with higher gut microbiota diversity in 10 observational and 14 interventional studies, plus better glycemic control and lower fat mass. The review is narrative with no pooled estimates, and the microbiota findings are mostly associations.
A meta-analysis of 21 studies with 65,955 participants found higher Mediterranean diet adherence was associated with lower dementia risk (pooled OR 0.89, 95% CI 0.84-0.94), and lower Alzheimer disease risk (OR 0.73, 95% CI 0.62-0.85). Heterogeneity between studies was high (I2 up to 89.70 in cohort studies), and the included studies are observational.
In 226 adolescents aged 12-19 from two private schools in Nepal, those who received 12 weeks of nutrition education scored 1.80 points higher on nutrition knowledge (95% CI 1.11-2.49) than the comparison school. With only two schools and no random assignment, differences between the schools could explain part of the gap.
Among 33,859 US adults, the median Planetary Health Diet Index score rose 4.2 points, from 62.7 in 2003-2004 to 66.9 in 2017-2018 on a 0-140 scale, and higher scores went with a lower probability of inadequate iron, fiber and potassium. Diet came from 2 days of recall in a cross-sectional survey, so it shows trends and associations, not what changing diet would do.
Among 12,402 US adults in NHANES 2011-2016, a higher systemic immune-inflammation index (SII) was associated with higher odds of metabolic syndrome (OR 1.33; 95% CI 1.14-1.55). It is a one-time survey snapshot, so it cannot show that inflammation causes metabolic syndrome, and the link was inverted-U shaped.
Among 166,377 UK Biobank participants followed for a median 9.46 years, each additional unit of the dietary inflammatory index (DII) was linked to a 4.6% higher incidence of all-cause dementia (HR 1.046); 1372 people developed dementia. Diet was assessed from 24-hour recall questionnaires, and as an observational cohort the study cannot show that changing diet changes dementia risk.
In a network meta-analysis of 42 randomized trials with 4809 people with type 2 diabetes, ketogenic, low-carbohydrate and low-fat diets significantly lowered HbA1c versus a control diet, and ketogenic, Mediterranean, moderate-carbohydrate and low glycemic index diets ranked well overall. The authors say high-quality, long-term studies are still needed to strengthen the ketogenic diet's credibility.
Across 30 randomised trials with 1,827 adults with type 2 diabetes, probiotics lowered fasting blood glucose (SMD -0.331, 95% CI -0.424 to -0.238) and HbA1c (SMD -0.421, 95% CI -0.584 to -0.258) compared with placebo. These are standardised effects with no clinical units given in the abstract, so the real-world size of the change in HbA1c is not stated.
In this review of 17 meta-analyses covering 68 trials, 51 of 115 associations (44%) were statistically significant, and only four were backed by high-quality evidence: lower triglycerides (2), fewer seizures (1) and higher LDL cholesterol (1). The trials were short (median follow-up 13 weeks) and small (median 42 participants), so long-term effects on heart events or survival remain unknown.
Among 171,544 UK Biobank participants followed for a median of 10.2 years, those with the most pro-inflammatory diets had a higher risk of severe non-alcoholic fatty liver disease (NAFLD) than those with the most anti-inflammatory diets (HR 1.19, 95% CI 1.03 to 1.38). The link is modest, and the diet was scored from food reports, so it cannot show that changing diet lowers risk.