Food Quality & Dietary Patterns

Meta-analysis of 4 RCTs (10,054 participants, 2–7 year follow-up) finding the Mediterranean diet reduced major adverse cardiovascular events by 48% vs. control diets (OR 0.52), making it the strongest RCT-level dietary pattern evidence for cardiovascular protection.

Massive meta-analysis of 87 studies encompassing 1.4 million participants confirming Mediterranean diet adherence reduces risk of coronary heart disease, atrial fibrillation, cerebrovascular disease, hypertension, and CVD mortality — validating whole-food dietary patterns as the evidence-based benchmark.

Meta-analysis of 28 studies (679,259 participants) finding high Mediterranean diet adherence reduces all-cause mortality by 23% and cardiovascular mortality by 27% in adults over 60 — demonstrating long-term survival benefits from sustained dietary pattern quality.

Controlled feeding study showing that increasing protein from 15% to 30% of calories while keeping total calories matched reduced spontaneous caloric intake by 441 kcal/day and produced 4.9 kg weight loss over 12 weeks — directly validating the app's protein targets as an appetite management tool.

Comprehensive review establishing the opposing roles of leptin (long-term satiety signalling) and ghrelin (short-term hunger hormone) in energy balance regulation, and that protein suppresses ghrelin more effectively than carbohydrates or fat per calorie — validating protein as the most satiety-efficient macronutrient for calorie management.

Meta-analysis of ≥12-week RCTs finding time-restricted eating, lower meal frequency, and front-loading calories earlier in the day are each associated with small but significant reductions in body weight, BMI, and waist circumference — supporting the app's meal timing and nutrient timing features.

Umbrella review confirming higher dietary fiber intake is consistently associated with reduced risk of CVD, cancer, type 2 diabetes, and obesity across multiple meta-analyses — with mechanisms including slowed glucose absorption, increased satiety, microbiome SCFA production, and reduced inflammation. Validates fiber tracking as a top-tier health metric.

Meta-analysis of 64 RCTs (2,099 participants) confirming dietary fiber — especially fructans and galacto-oligosaccharides — significantly increases Bifidobacterium and Lactobacillus species and fecal short-chain fatty acids, linking fiber tracking to microbiome and metabolic health outcomes.

Dose-response meta-analysis of 21 cohort studies establishing that each 5-unit increase in glycemic index was associated with 8% higher type 2 diabetes risk — validating why carbohydrate quality, not just quantity, matters for metabolic health tracking in the app.

Foundational meta-analysis establishing that regular SSB consumption significantly increases risk of metabolic syndrome (RR 1.20) and type 2 diabetes (RR 1.26), providing the earliest strong meta-analytic case for tracking and limiting added sugar in a fitness context.

Large-scale meta-analysis confirming SSB intake increased obesity risk by 17%, type 2 diabetes by 20%, coronary heart disease by 17%, and stroke by 10% — directly supporting sugar tracking as a meaningful health metric within RobustHealth's nutrition logging.

Meta-analysis of RCTs finding omega-3 supplementation significantly reduced IL-6, TNF-α, and CRP concentrations following exercise-induced muscle damage in healthy individuals — establishing omega-3 EPA/DHA as a priority recovery nutrient with direct relevance to training load management.

Meta-analysis showing omega-3 supplementation significantly reduced CK, LDH, and myoglobin — all biomarkers of exercise-induced muscle damage — in healthy individuals, validating omega-3 tracking within the app's micronutrient module as directly relevant to recovery.

Systematic review of RCTs concluding that omega-3 supplementation consistently reduces inflammatory markers and muscle damage biomarkers following exercise, with emerging evidence for preserved muscle mass and strength — directly validating omega-3 tracking as a recovery tool in RobustHealth's micronutrient system.

Mice fed carboxymethylcellulose (CMC) or polysorbate-80 at doses equivalent to human additive exposure showed bacterial encroachment into the intestinal mucosa, altered microbiota composition, low-grade inflammation, and metabolic syndrome. Germ-free transplant experiments confirmed the microbiota changes were necessary and sufficient to produce both inflammation and metabolic dysfunction — establishing the mechanism behind emulsifier toxicity.

In a double-blind controlled feeding study, 16 healthy adults on a diet containing 15 g/day of CMC showed reduced gut microbial diversity, decreased short-chain fatty acids in stool, increased abdominal discomfort, and in two subjects, bacteria encroaching into the normally sterile mucus layer. No such changes appeared in the control group eating identical food without CMC — the first human RCT confirming gut-disrupting effects of a common emulsifier at realistic dietary doses.

Twenty commonly used dietary emulsifiers were tested directly against human gut microbiota; the majority — including CMC, polysorbate-80, and carrageenan — altered microbiota composition and gene expression in a manner expected to promote intestinal inflammation, while soy lecithin had minimal impact. This broad-spectrum screen is the key citation differentiating which emulsifiers warrant red versus yellow flags in the ingredient safety score.

Overweight males receiving 250 mg/day carrageenan in a two-week crossover design showed reduced whole-body and hepatic insulin sensitivity, increased intestinal permeability, and elevated inflammatory markers (CRP, IL-6) — the first human RCT demonstrating that carrageenan at food-additive doses induces insulin resistance and gut permeability effects in humans.

Non-caloric artificial sweeteners (saccharin, sucralose, aspartame) induced glucose intolerance in mice through microbiota disruption — the effect was abolished with antibiotics and transferable to germ-free mice via microbiota transplant, confirming the mechanism. In a parallel human cohort, higher artificial sweetener consumption correlated with glucose intolerance and altered microbiota, establishing microbiome-mediated metabolic dysfunction as the mechanism behind sweetener-related harm.

Sucralose supplementation in rats disrupted gut microbiota composition, promoted systemic inflammation via lipopolysaccharide translocation, impaired glucose tolerance, and increased insulin resistance — providing mechanistic evidence for sucralose's metabolic effects independent of caloric intake.

This review synthesises the mechanistic evidence that non-caloric sweeteners alter host-microbiota interactions despite containing no calories, with saccharin showing the strongest effect on glucose metabolism. The authors call for re-evaluation of sweetener safety assumptions and highlight the key limitation: human microbiome response is highly individual, which explains conflicting trial results and supports a precautionary flagging approach rather than outright banning.

The most comprehensive synthesis of both RCT and cohort evidence on non-nutritive sweeteners. Key finding: RCTs showed no significant effect on BMI, glucose, blood pressure, or insulin at realistic doses. Prospective cohort studies, however, associated higher sweetener consumption with modestly increased BMI, hypertension, and metabolic syndrome. This RCT-vs-cohort divergence — the core of the conflict — most likely reflects reverse causation (people already heavier use more sweeteners) in observational data rather than a true causal harm at typical doses.

Meta-analysis of 15 RCTs finding that replacing caloric beverages with low-calorie sweetened versions significantly reduced body weight (−0.80 kg), BMI, fat mass, and waist circumference. Compared with water controls, the difference was not significant — meaning sweeteners are no better than water, but the evidence does not support the claim that they cause weight gain or metabolic harm when used as a sugar replacement in controlled conditions.

Phytic acid (InsP6) is a potent inhibitor of iron and zinc absorption from plant-based diets, with molar phytate:iron and phytate:zinc ratios used to predict bioavailability. The European Food Safety Authority now sets adult zinc requirements across four phytate-intake tiers because of this inhibition, and traditional processing techniques such as soaking, germination, and fermentation reduce phytate via phytase hydrolysis — directly justifying the lower bioavailability multiplier applied to grain and legume foods in the micronutrient score.

Iron and zinc absorption from legume-based diets is demonstrably poor despite high mineral content, primarily because phytate co-precipitates with iron and zinc in the gut while polyphenols in legumes further compound this by binding iron. Enzymatic phytate degradation through soaking, germination, and fermentation effectively removes these inhibitors and restores mineral bioavailability — whereas legumes consumed raw or minimally processed retain full anti-nutrient activity, validating the distinction in food matrix scoring.

Adding phytic acid to white bread at levels equivalent to whole-meal bread reduced fractional magnesium absorption from 32.5% to 13.0% at high doses and 24.0% at lower doses — a statistically significant, dose-dependent relationship established with stable isotope tracers. This controlled human trial confirms that phytate suppresses absorption of multiple minerals (not just iron and zinc) by up to 60%, quantifying the anti-nutrient penalty applied to grain-legume foods in the bioavailability matrix.

This landmark review catalogues polyphenol content across major food groups — fruits, vegetables, cereals, legumes, chocolate, tea, wine — and systematically examines how food matrix, processing, and cooking affect bioavailability, finding that the most abundant dietary polyphenols are not always the best absorbed and that structural features profoundly alter uptake. It establishes the scientific framework for using polyphenol content as a food quality dimension by demonstrating measurable, food-specific variation in antioxidant compound delivery.

Pooling seven prospective cohort studies covering 178,657 adults, higher dietary polyphenol intake was associated with a statistically significant 7% reduction in all-cause mortality (HR 0.93, 95% CI 0.91–0.95). This is the most current meta-analysis establishing that total dietary polyphenol exposure — measurable from food composition databases — predicts survival outcomes at the population level, providing the outcome-level evidence justifying polyphenol content as a distinct food quality dimension.

Across 187,382 health professionals, greater consumption of polyphenol-rich whole fruits — especially blueberries, grapes, and apples — was significantly associated with lower type 2 diabetes risk (blueberries HR 0.74 per 3 servings/week), while fruit juice consumption was associated with increased risk. The contrast demonstrates that the polyphenol-rich food matrix of whole fruits confers protective metabolic effects distinct from their sugar content, supporting antioxidant content as a quality marker that differentiates whole foods from processed equivalents.

The authors reviewed approximately 6,500 peer-reviewed articles and identified 45 food parameters each scored for their effect on six inflammatory biomarkers (IL-1β, IL-4, IL-6, IL-10, TNF-α, CRP), creating an individual-level Dietary Inflammatory Index (DII) ranging from −8.87 (maximally anti-inflammatory) to +7.98 (maximally pro-inflammatory), validated against global food consumption data from 11 countries. This is the foundational paper establishing that dietary inflammatory potential is a scientifically defined, individually calculable dimension of diet quality — directly supporting an inflammatory axis in food scoring.

A meta-analysis of 14 studies found individuals in the highest DII category (most pro-inflammatory diet) had a 36% higher risk of cardiovascular events and mortality compared to those in the lowest category, with each one-unit DII increase associated with an 8% rise in CVD risk. This outcome-level meta-analysis connects dietary inflammatory potential directly to hard clinical endpoints, validating inflammatory potential as a meaningful food quality dimension across diverse populations.

Western diets now contain omega-6 to omega-3 ratios of 15:1–17:1, far removed from the approximately 1:1 ratio of ancestral human diets; excessive omega-6 promotes pro-inflammatory eicosanoid production while high omega-3 intake suppresses it, with specific therapeutic ratios shown for different conditions (e.g., 4:1 ratio associated with 70% reduction in cardiovascular mortality). This foundational review establishes the omega-6:omega-3 ratio as a quantifiable, food-scoreable determinant of inflammatory potential, supporting fatty acid balance as a scoring metric alongside saturated fat penalties.

Across 30 RCTs and 38 cohort studies, reducing free/added sugar intake in adults produced a mean weight loss of 0.80 kg, while isocaloric replacement of sugars with other carbohydrates showed no weight change (0.04 kg), confirming that the effect is attributable to the specific metabolic properties of free sugars rather than calorie displacement alone. This WHO-commissioned meta-analysis is the primary evidentiary basis for the WHO's free sugar guideline (<10% of energy) and directly validates penalising the sugar/carb ratio in the macro balance score.

This umbrella review of 73 meta-analyses identified harmful associations between dietary sugar and 45 distinct health outcomes including obesity, type 2 diabetes, cardiovascular disease, depression, dental caries, and multiple cancers, with the evidence strongest for sugar-sweetened beverages. The authors recommend limiting added sugars to under 25 g/day — providing broad scientific validation for the sugar-penalty logic in the macro balance pillar and dose thresholds that can directly inform penalty curve calibration.

A meta-analysis of 51 controlled trials showed that excess energy from fructose-containing sugars — particularly from sugar-sweetened beverages — significantly increases intrahepatic lipid content (liver fat), a key driver of metabolic syndrome and insulin resistance; this effect was source-dependent, with SSBs showing the strongest impact. The finding that fructose-containing sugars promote liver fat accumulation beyond their calorie content provides the mechanistic link between high sugar/carb ratios and metabolic disease, strengthening the case for the sugar ratio penalty.

This dose-response meta-analysis of 45 prospective studies found that consuming 90 g/day of whole grains (three servings) was associated with a 19% lower risk of cardiovascular disease, 15% lower cancer risk, and 17% lower all-cause mortality compared to no whole grain intake — with clear dose-response relationships. Refined grains showed no such protective associations, directly establishing the health gap between whole and refined grain products that underpins the ingredient safety penalty for enriched and white flours.

In 291,988 men and 197,623 women followed for 5 years, higher whole grain intake was significantly associated with lower colorectal cancer risk (HR 0.79 for highest vs lowest quintile), whereas refined grain intake was not protective. This large prospective study isolates the grain quality distinction — not grain consumption per se — as the driver of disease risk, validating the scoring system's differentiation between whole grain and refined grain foods.

In a 12-week randomised crossover trial in hyperinsulinemic adults, a whole grain diet improved insulin sensitivity and reduced BMI, fasting insulin, and blood pressure compared to a refined grain diet of identical calorie and macronutrient content. The finding that two diets matched for calories and macros produce different metabolic outcomes provides direct human trial evidence that grain quality — beyond fibre content alone — has independent metabolic effects, justifying a quality distinction in the food scoring system.

In this prospective cohort study of 74,698 Japanese adults, dietary nitrate from vegetable sources was associated with significantly lower cardiovascular disease mortality, whereas processed meat intake (the predominant source of added nitrite) showed no protective effect and trended toward harm. This study is key evidence that the food matrix — not nitrate/nitrite per se — determines health impact, supporting the scoring system's approach of flagging sodium nitrite as an additive (harmful context) while not penalising whole vegetables for their natural nitrate content.

This review establishes that vegetables (especially beetroot, spinach, rocket, and lettuce) are the dominant source of dietary nitrate, providing ~80% of intake, and that the concomitant presence of vitamin C and polyphenols in vegetables inhibits nitrosamine formation — the carcinogenic conversion pathway. This is the mechanistic explanation for why vegetable nitrate and processed-meat nitrite produce opposite health outcomes, informing the nuanced scoring approach of flagging the additive E250/E251 while leaving naturally nitrate-rich vegetables unpenalised.

The IARC Monographs Working Group classified processed meat as a Group 1 carcinogen (sufficient evidence in humans) and red meat as Group 2A (probably carcinogenic), with nitrite/N-nitroso compounds cited as one of the key mechanistic contributors alongside haem iron and heterocyclic amines. This classification by the world's leading cancer research agency provides the authoritative scientific basis for the red-flag status of sodium nitrite and sodium nitrate as food additives in the ingredient safety score.

In three prospective cohort studies totalling over 4,000 cardiac patients, elevated plasma erythritol was independently associated with major adverse cardiovascular events (MACE) including heart attack and stroke over 3 years. In vitro and animal experiments showed erythritol directly enhanced platelet aggregation and thrombosis — the first mechanistic evidence linking an approved sugar alcohol to cardiovascular risk at levels achievable from a single erythritol-sweetened food or beverage, directly supporting the yellow-flag status for erythritol in the ingredient safety score.

A metabolomics study of college freshmen found that erythritol was among the strongest metabolic predictors of fat mass gain over the academic year — and that erythritol is endogenously produced by the pentose phosphate pathway from glucose, not solely from dietary intake. This adds a layer of complexity: erythritol is both an exogenous food additive and an endogenous metabolic byproduct, suggesting elevated plasma levels may reflect both dietary exposure and metabolic dysfunction, complicating interpretation of the Witkowski 2023 findings but not diminishing the caution.

This landmark review synthesised RCT and cohort data establishing that industrial trans fatty acids (iTFAs) from partial hydrogenation raise LDL cholesterol, lower HDL cholesterol, increase triglycerides, promote systemic inflammation, and impair endothelial function — producing a uniquely atherogenic lipid profile worse than any natural dietary fat. Each 2% of energy from iTFAs increases coronary heart disease risk by approximately 23%, making partially hydrogenated oils the most harmful fat type per gram in the food supply, and directly supporting a red-flag status for any ingredient listing "partially hydrogenated" oils.

This systematic review and meta-analysis of 73 observational studies found that trans fat intake — but not saturated fat intake — was consistently and significantly associated with higher all-cause mortality, cardiovascular disease mortality, and total CHD events. The dose-response relationship for iTFAs was steep, with no safe lower threshold identified, providing the meta-analytic evidence base for WHOs 2018 REPLACE action plan calling for complete elimination of partially hydrogenated oils from the global food supply.

Meta-analysis of 8 cohort studies separating industrial TFAs (from partial hydrogenation) from natural ruminant TFAs (from dairy and beef): industrial TFAs were associated with a 21% higher coronary heart disease risk per 2% energy increment, while ruminant TFAs showed no significant association with CHD risk. This important distinction confirms that the harm is specific to the industrial manufacturing process — partial hydrogenation — not to the trans-fat chemical structure per se, and supports flagging "partially hydrogenated" ingredients specifically rather than all trans fat sources.

This review by nephrology specialists documents that inorganic phosphate salts (E338–E452 range: phosphoric acid, sodium phosphates, potassium phosphates, polyphosphates) used as preservatives, emulsifiers, and leavening agents in processed meats, cheese, cola drinks, and bakery products are absorbed at 80–100% efficiency — compared to 40–60% for organic phosphate in whole foods. In healthy people, excess phosphate is excreted renally; in those with even mildly reduced kidney function (which is undetected in most adults), phosphate retention drives hyperphosphataemia, vascular calcification, and accelerated cardiovascular disease. The authors call for mandatory labelling of inorganic phosphate additives separately from naturally occurring phosphorus.

Surveying ingredient labels of 2,394 top-selling grocery items, 44% contained phosphate additives — with the highest prevalence in processed meats (93%), dried food mixes (82%), packaged convenience foods (73%), and carbonated soft drinks (69%). Since Nutrition Facts panels do not distinguish additive phosphorus from natural food phosphorus, consumers following phosphate-restricted diets are systematically misled. This study quantifies the hidden phosphate load in common ultra-processed foods, providing empirical justification for flagging phosphate salt ingredients in processed food scoring.

This CKD dietary management review establishes a critical hierarchy: inorganic phosphate additives in processed foods are nearly 100% absorbed versus ~50% from animal protein versus ~30% from plant-based phytate-bound phosphorus, making food source and processing status the most important determinant of phosphate bioavailability. For the general population, this bioavailability differential means ultra-processed foods deliver substantially more absorbable phosphate per gram than whole food equivalents — directly informing the ultra-processed food bioavailability penalty and justifying specific flags for phosphate salt additives.

After reviewing all available toxicological evidence, EFSA concluded that titanium dioxide (E171) — a white colorant in confectionery, chewing gum, medications, and bakery products — can no longer be considered safe as a food additive. The key concern was genotoxicity: EFSA could not exclude the potential for DNA damage following oral ingestion of TiO₂ nanoparticles, which are present in food-grade E171 and accumulate in the body over time. This opinion directly led to the EU ban on E171 in food from August 2022 — providing the regulatory endpoint backing a red-flag status for any ingredient listing titanium dioxide or CI 77891.

Rats fed food-grade TiO₂ (E171) at doses within human dietary exposure ranges for 100 days showed intestinal immune homeostasis disruption, increased colonic mucus secretion, initiation of preneoplastic lesions, and promotion of aberrant crypt foci — precursors to colorectal cancer. These findings were not caused by bulk titanium oxide but specifically by the nanoparticle fraction of commercial E171, which penetrates the intestinal mucosa and accumulates in lymphoid tissue, providing the mechanistic evidence underlying EFSA's genotoxicity concern.

Exposure to TiO₂ nanoparticles significantly worsened colitis in a mouse model and activated the NLRP3 inflammasome — the innate immune signalling complex responsible for IL-1β and IL-18 release and a key driver of chronic intestinal inflammation. Critically, nanoparticle-sized TiO₂ (but not micro-sized TiO₂) activated this pathway, identifying the inflammatory mechanism as size-dependent and explaining why food-grade E171, which contains a significant nanoparticle fraction, poses a gut inflammation risk not seen with coarser industrial titanium dioxide. This study is particularly relevant for individuals with IBD risk.

In a double-blind RCT of 297 children, two separate mixes of synthetic food dyes (including tartrazine/E102, sunset yellow/E110, carmoisine/E122, quinoline yellow/E104, and allura red/E129) combined with the preservative sodium benzoate significantly increased hyperactivity scores in both 3-year-olds and 8/9-year-olds compared to placebo, with effect sizes detectable in the general population — not only in children with ADHD. This study prompted the European Food Safety Authority to require warning labels on food containing the "Southampton Six" dyes, and the UK FSA recommended their voluntary removal from food.

This review of double-blind placebo-controlled trials concluded that artificial food colors do worsen ADHD symptoms in children who already have ADHD (mean effect size ~0.4), and produce measurable hyperactivity effects in the general child population, though the effect is smaller. The review distinguishes neurotoxic mechanisms (direct nervous system effects) from behavioural sensitisation, notes that the effect is independent of parental expectations, and points out that elimination of artificial colors is a safe, evidence-based dietary modification that produces clinically meaningful benefit in a significant subset of children — with no downside risk.

This comprehensive toxicology review of nine FDA-approved synthetic dyes found that several have inadequate safety evidence for human consumption: Red 3 (erythrosine) is a known thyroid carcinogen in animals; Red 40, Yellow 5, and Yellow 6 are contaminated with cancer-causing benzidine and other aromatic amines; and multiple dyes cause hypersensitivity reactions. The review concludes that the collective evidence warrants removing most synthetic dyes from the food supply — establishing that ingredient safety concerns extend beyond neurobehavioral effects to potential carcinogenicity, supporting red-flag status for synthetic dye ingredients (FD&C Red 40, Yellow 5, Yellow 6, Blue 1, Blue 2).