Food and household ingredients, explained through research — with concerns, context, and clear limits.
Search across all 26 ingredient research entries.
Transparency = Honesty
A complete ingredient list gives you clear information about what is inside a product. Terms like “fragrance” or “proprietary blend” do not name every ingredient. Our listing standard requires full ingredient disclosure.
What the evidence tells us
These papers examine specific chemicals, concentrations, and exposure conditions. A substance’s ability to cause harm is not the same as the risk from using a particular product. Our screening choices are not proof that every excluded ingredient is harmful in every use.
The studies do not verify the sample products in this directory or establish that listed products are risk-free. Summaries are educational, not medical advice.
Food ingredients we screen for
These are the 15 ingredient groups we research in packaged foods, reviewed against human studies, laboratory research, and public safety assessments. Some overlap: canola is a seed oil, and many gums, colours, and preservatives have E numbers.
Here, “screen for” means examine the evidence, not automatically exclude. We examine why ingredients are questioned: human findings, possible biological mechanisms, regulatory restrictions, and research funding. We distinguish a measured effect from a hypothesis and a human trial from a cell or animal experiment. These research topics do not add new exclusions or change existing catalog decisions.
A hard-to-pronounce name is not a safety test. Natural and artificial ingredients both need evidence. An acceptable daily intake is an estimate for long-term exposure, not a sharp boundary between safe and poisonous; “no numerical limit needed” does not mean unlimited amounts.
Sources reviewed 2026-10-27 · Original source years shown in each entry · Educational information, not medical advice.
Research context · Not a blanket exclusion
Stabilisers
Carrageenan (E407), carboxymethylcellulose or CMC (E466), and gellan gum (E418) help foods stay smooth instead of separating. They are different ingredients, not one health-risk category.
Potential human concerns
CMC has human feeding evidence of altered gut bacteria, reduced microbial metabolites, and modestly greater stomach discomfort. Animal studies of CMC and polysorbate 80 found inflammation and metabolic effects. Carrageenan has cell, animal, and small human studies raising intestinal and inflammation questions; short human trials have not given uniform results. Food-grade carrageenan is not the same as poligeenan, a degraded material not authorized for food. These findings cannot be assigned to gellan gum merely because it also stabilizes food.
Randomized Controlled-Feeding Study of Dietary Emulsifier Carboxymethylcellulose Reveals Detrimental Impacts on the Gut Microbiota and Metabolome
Chassaing et al. · Gastroenterology · 2022
Controlled feeding trial in 16 healthy adults; seven received 15 g CMC daily for 11 days
In plain language
Adults eating CMC had modestly more discomfort after meals and changes in their gut bacteria and the substances those bacteria made. Two people had bacteria move closer to the gut's protective mucus layer.
Conclusion
CMC has a directly measured human microbiome signal worth taking seriously. Carrageenan research includes concern signals and small studies with different results. Neither a regulator's permission nor a single experiment settles long-term effects for all stabilisers.
What this doesn’t tell us
Very few people, a short study, and one relatively substantial dose. Gut-bacteria changes are not automatically disease. EFSA's reassuring 2018 cellulose assessment predates this trial. Carrageenan and gellan require their own evidence.
Who funded the research?
CMC feeding trial, 2022
Declared support included US NIH, the Max Planck Society, and public/university/foundation support for participating laboratories. The authors state that funders had no role in design, data collection, analysis, interpretation, or writing. Dr Lewis disclosed consulting/advisory relationships with pharmaceutical companies and Nestle Health Science, and research funding from Nestle Health Science and pharmaceutical companies. The remaining authors disclosed no conflicts.
The short duration and small sample remain limitations regardless of sponsor.
Abbott Nutrition and consultancy affiliations are documented. Marinalg International confirms industry support for the piglet study.
A potential industry conflict deserves disclosure. A 28-day animal study cannot settle chronic human effects; sponsorship alone does not invalidate its findings.
The authors declare no potential conflicts. Funding: Juhani Aho Medical Research Foundation, Folkhälsan Research Center, Wilhelm and Else Stockmann Foundation, and Novo Nordisk Foundation.
Foundation support is not the same as sponsorship by the commercial company Novo Nordisk. This remains a small short study.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Controlled human feeding study of CMCPMID 34774538 · 16 healthy adults; 15 g/day in seven participants for 11 days
In plain language
CMC changed gut bacteria and their products. Two participants had bacteria move closer to the protective mucus layer.
Conclusion
A short human study found measurable effects, not just a laboratory hypothesis.
What this doesn’t tell us
High experimental intake, few participants, and no proof of chronic intestinal disease.
Read the PubMed recordCarrageenan and relapse in ulcerative colitisPMID 28447072 · Very small randomized pilot in people with disease in remission
In plain language
Three relapses were reported with carrageenan capsules and none with placebo while participants followed a carrageenan-free diet.
Conclusion
A possible concern for a susceptible population that needs larger confirmation.
What this doesn’t tell us
Twelve participants supplied questionnaire data; recruitment, attrition, adherence and the underlying disease limit generalization to healthy consumers.
Read the PubMed recordCarrageenan in quiescent ulcerative colitis: short randomized pilotPMID 37920677 · Small short-term human trial
In plain language
The pilot did not demonstrate a clear pro-inflammatory effect under its tested conditions.
Conclusion
Human findings are not uniform.
What this doesn’t tell us
Limited power and duration do not establish universal or long-term safety.
Read the PubMed recordCarrageenan and insulin sensitivity: 2024 crossover trialPMID 39593091 · 20 men; 500 mg/day for two weeks
In plain language
Overall insulin sensitivity did not differ significantly. Permeability increased and BMI-dependent metabolic/inflammation-related findings were reported.
Conclusion
A measured permeability signal with a null overall primary insulin-sensitivity outcome.
What this doesn’t tell us
Short male-only trial; subgroup findings are not proof of diabetes or a uniform effect.
Polysorbate 80 (E433), lecithin (E322), and mono- and diglycerides (E471) help oil and water mix, like keeping a dressing from separating.
Potential human concerns
Mouse experiments found that CMC and polysorbate 80 changed gut bacteria and promoted inflammation and metabolic effects. A large French cohort linked higher intake of some emulsifiers with cardiovascular disease and cancer; the associations involved particular additives, not every emulsifier. Lecithin's source proteins may matter for allergies. Mono- and diglycerides need their own evidence and impurity controls.
Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome
Chassaing et al. · Nature · 2015
Mouse experiment involving polysorbate 80 and CMC
In plain language
Mice exposed to these two emulsifiers developed changes in gut bacteria. Some had low-level inflammation and metabolic changes; mice already prone to intestinal disease developed worse inflammation.
Conclusion
Experimental mechanisms and human observational associations justify closer scrutiny of particular emulsifiers. They do not prove that an ordinary serving causes inflammation, cardiovascular disease, or cancer, or that lecithin behaves like polysorbate 80.
What this doesn’t tell us
Mouse dosing and direct cell exposure differ from eating a meal. Cohort diets are not randomly assigned: other dietary patterns, measurement error, and lifestyle can affect the results despite statistical adjustment. Association does not establish cause. The existing small CMC trial is human evidence about CMC, not every emulsifier.
Who funded the research?
NutriNet cancer cohort, 2024
French public research/health institutions, ERC, INCa, French Ministry of Health, university support, and the Bettencourt-Schueller Foundation are disclosed. Authors declared no competing interests and no funder role.
Funding disclosure does not remove residual confounding. This paper and the CVD analysis use the same cohort, not independent replications.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
A powder made by breaking starch into smaller pieces. Ordinary maltodextrin is a quickly digested carbohydrate used for texture or as a carrier for other ingredients. Resistant maltodextrin is a different, fibre-like ingredient.
Potential human concerns
Digestible maltodextrin can raise blood glucose quickly. In laboratory experiments it helped Crohn's-associated bacteria attach to intestinal cells and form sticky communities. Cell and mouse infection experiments found weakened antibacterial responses and increased Salmonella survival or colonization. These are concrete experimental concerns, not evidence that ordinary dietary intake causes Crohn's disease or infections in people.
Crohn's Disease-Associated Adherent-Invasive Escherichia coli Adhesion Is Enhanced by Exposure to the Ubiquitous Dietary Polysaccharide Maltodextrin
Nickerson & McDonald · PLOS ONE · 2012
Laboratory experiments with bacteria and cells, not a human feeding trial
In plain language
In lab tests, maltodextrin helped certain bacteria form sticky communities and attach more readily. These bacteria have been studied in connection with Crohn's disease.
Conclusion
Maltodextrin deserves research attention for bacterial adhesion and intestinal defenses as well as its rapidly digested carbohydrate content. The published cell and mouse findings do not establish human disease causation or a harmful food serving size.
What this doesn’t tell us
The 2015 article discussing these findings is an addendum, not an independent human trial. The original experiments do not establish inevitable mucus thinning or a universal glycemic-index value. Digestible and resistant maltodextrins differ. Dose and meal formulation matter.
Who funded the research?
Nickerson studies, 2012 and 2014
Declared NIH grants, HHMI Med into Grad support, and Gerald and Nancy Goldberg support. The original papers declare no competing interests.
These disclosures do not remove cell/mouse model limitations. The earlier allegation of Corn Refiners Association sponsorship was not verified and is not used here.
Sunflower, safflower, cottonseed, and grapeseed oils provide unsaturated fats, including omega-6 linoleic acid. High-oleic sunflower and safflower varieties have a different fat balance.
Potential human concerns
Linoleic acid can react with oxygen to form oxidized fats, including HODEs. Related breakdown products such as 4-HNE can affect cells in experiments. Human feeding studies have found changes in some oxidation measures, and one soy-oil trial found increased oxidized LDL and Lp-PLA2, an inflammation-related enzyme. These are specific research findings, not proof that every seed oil causes chronic inflammation. Oxidation, common blood inflammation markers, and diagnosed disease are different outcomes.
Impact of 8-week linoleic acid intake in soy oil on Lp-PLA2 activity in healthy adults.
Kim et al. · Lipids in Health and Disease · 2017
Eight-week randomized dietary trial in 150 adults
In plain language
Researchers assigned adults to three diets: replace some soy oil with an apple, keep usual intake, or replace some rice with soy-oil capsules. Compared with the lower-intake group, the higher-intake group had greater increases in oxidized LDL and apoB, a protein on certain cholesterol-carrying particles. Its Lp-PLA2 enzyme activity also increased more than in the other groups.
Conclusion
This trial gives a concrete reason to investigate oxidation and inflammation-related pathways, rather than judging an oil only by whether it lowers cholesterol. Other cited studies show that changing dietary fats can change oxidation measurements, and lowering linoleic acid can lower some oxidized metabolites. They do not establish that all seed oils cause disease. A separate review of 15 human trials generally found no increase in common inflammation markers with higher linoleic acid intake.
What this doesn’t tell us
The soy-oil trial changed foods as well as fat intake, lasted only eight weeks, and measured biomarkers rather than heart attacks or chronic disease. Cell experiments expose cells directly to breakdown products; they do not test eating an oil. Observational disease studies cannot establish which came first. Historical heart-disease trials used particular diets and populations, and their results are not uniform. The amount, oil composition, processing, and cooking conditions matter; these papers do not establish a harmful serving size or rank cold-pressed, refined, olive, avocado, and coconut oils. Food research remains separate from Crunchify's admission exclusions.
Who funded the research?
Underlying oil studies and reviews
The complete funding and author-conflict statements were not independently verified in this review.
Do not interpret missing information as independence or as evidence of industry influence.
Impact of 8-week linoleic acid intake in soy oil on Lp-PLA2 activity in healthy adultsPMID 28503188 · Randomized human dietary trial
In plain language
In 150 adults, higher soy-oil intake increased plasma linoleic acid and Lp-PLA2 activity; oxidized LDL and apoB increased more than in the lower-intake group.
Conclusion
Soy-oil intake affected these particular biomarkers under the tested conditions.
What this doesn’t tell us
Eight weeks, food substitutions differed between groups, and no clinical disease outcomes were measured.
Read the PubMed recordEffects of oleate-rich and linoleate-rich diets on the susceptibility of low density lipoprotein to oxidative modification in mildly hypercholesterolemic subjectsPMID 8432867 · Human feeding study with laboratory LDL tests
In plain language
After eight weeks, LDL taken from people eating an oleate-rich diet was less easily oxidized in laboratory tests than LDL from the linoleate-rich group.
Conclusion
Dietary fat composition can change how readily isolated LDL oxidizes.
What this doesn’t tell us
Copper- and cell-induced oxidation in a laboratory is not a measurement of heart attacks or ordinary oxidation inside a person's arteries.
Read the PubMed recordEffects of diets rich in monounsaturated fatty acids on plasma lipoproteins—the Jerusalem Nutrition Study: high MUFAs vs high PUFAsPMID 2008870 · 24-week crossover feeding study in 26 men
In plain language
Both diets lowered cholesterol. The polyunsaturated-fat diet lowered it somewhat more, but showed greater lipid-peroxidation susceptibility; the monounsaturated-fat diet showed lower susceptibility.
Conclusion
Cholesterol and oxidation measurements can move in different directions.
What this doesn’t tell us
Small study, biochemical outcomes, and no test of long-term disease or every individual oil.
Read the PubMed recordDietary intakes of polyunsaturated fatty acids and indices of oxidative stress in human volunteersPMID 10452406 · Four-week dietary periods in 10 healthy men
In plain language
A higher-PUFA diet increased oxidized glutathione in blood and urinary TBARS, two oxidation-related measures. Several other blood measures did not change, and the lower-PUFA diet increased total cholesterol.
Conclusion
Some, but not all, oxidation measures changed with PUFA intake.
What this doesn’t tell us
Very small, short study. TBARS is not a specific test for one chemical or a diagnosis of inflammation; the results do not establish disease risk.
Read the PubMed recordStrong increase in hydroxy fatty acids derived from linoleic acid in human low density lipoproteins of atherosclerotic patientsPMID 9488997 · Observational comparison of 17 patients and 19 volunteers
In plain language
People with atherosclerosis had more 9-HODE, a linoleic-acid oxidation product, in their LDL. Age also affected the measured levels.
Conclusion
An oxidation product was associated with atherosclerosis in this small sample.
What this doesn’t tell us
The study did not assign diets or identify dietary seed oils as the cause. Existing disease may affect oxidation; association does not show direction of cause.
Read the PubMed recordLowering dietary linoleic acid reduces bioactive oxidized linoleic acid metabolites in humansPMID 22959954 · 12-week dietary intervention in people with chronic headaches
In plain language
Lowering dietary linoleic acid reduced circulating linoleic acid and several oxidized derivatives, including HODEs and oxoODEs.
Conclusion
Diet can change levels of these oxidized metabolites in people.
What this doesn’t tell us
This report did not establish that lowering the metabolites prevents heart disease, diabetes, liver disease, or cancer. This PMID was supplied twice and is listed once.
Read the PubMed recordCorn Oil in Treatment of Ischaemic Heart DiseasePMID 14288105 · Small historical randomized secondary-prevention trial
In plain language
The trial compared corn oil, olive oil, and a control regimen in people who already had heart disease. Corn oil lowered cholesterol but did not demonstrate a clinical benefit; the authors did not recommend it under the trial's conditions.
Conclusion
The tested corn-oil regimen did not demonstrate benefit for these patients.
What this doesn’t tell us
Small historical trial with adherence and tolerability issues. It did not establish oxidation or inflammation as the cause, or a harmful effect for every seed oil. PubMed has no abstract; conclusions were checked against the original scanned report.
Read the PubMed recordLow fat, low cholesterol diet in secondary prevention of coronary heart diseasePMID 727035 · Original Sydney trial report; 458 men with coronary disease
In plain language
The original report described slightly better survival in the group eating less polyunsaturated fat and more saturated fat, but reported no significant relationship between dietary factors and survival in its multivariate analysis.
Conclusion
The original report did not establish a significant independent dietary effect on survival in that analysis.
What this doesn’t tell us
This is the original 1978 report, not the later recovered-data reanalysis. Participants had already changed diet, smoking, or weight; it did not test oxidation as a mechanism. PubMed has no abstract; the publisher's summary was checked, not an accessible full text.
Read the PubMed recordTest of effect of lipid lowering by diet on cardiovascular risk. The Minnesota Coronary SurveyPMID 2643423 · Randomized institutional dietary trial; 9,057 participants
In plain language
Replacing much of the saturated fat with polyunsaturated fat lowered cholesterol. The original report found no difference overall in cardiovascular events, cardiovascular deaths, or total mortality.
Conclusion
Lower cholesterol did not translate into an overall clinical benefit in this trial's original report.
What this doesn’t tell us
Average time on the diets was about a year despite the longer trial period. Institutional populations and multiple dietary changes limit generalization. This PMID is not a report that the intervention significantly increased overall mortality.
Read the PubMed recordThe role of oxidized lipid species in insulin resistance and NASH in childrenPMID 36263326 · Narrative review, not a new feeding trial
In plain language
The authors review possible roles for oxidized products of linoleic and arachidonic acids in pediatric fatty liver disease and insulin resistance.
Conclusion
Oxidized fats are plausible pathways for further investigation in these conditions.
What this doesn’t tell us
A review of mechanisms and associations does not demonstrate that eating seed oils causes these illnesses in children.
Read the PubMed recordOxidized fatty acids: A potential pathogenic link between fatty liver and type 2 diabetes in obese adolescents?PMID 23815500 · Observational study of 80 adolescents with obesity
In plain language
Some oxidized fatty acids were associated with liver-injury markers and lower insulin secretion. Adolescents with type 2 diabetes had higher levels than those with normal or impaired glucose tolerance.
Conclusion
The findings support a hypothesis connecting oxidized fats with liver injury and diabetes.
What this doesn’t tell us
No dietary seed-oil intervention; disease, obesity, genetics, or other factors could affect the metabolites. The study cannot establish dietary causation.
Read the PubMed record4-hydroxynonenal causes impairment of human subcutaneous adipogenesis and induction of adipocyte insulin resistancePMID 28088621 · Human tissue sampling and cultured-cell experiments
In plain language
Researchers used tissue from 18 women with obesity and treated cultured precursor fat cells with 4-HNE. Treatment impaired cell growth, fat-cell development, and insulin-related signaling.
Conclusion
Direct 4-HNE exposure can disrupt these cell functions under laboratory conditions.
What this doesn’t tell us
Participants did not eat assigned oils. Direct cell exposure cannot tell us which dietary oil, dose, or cooking method produces a harmful human exposure. 4-HNE is not unique to one oil.
Read the PubMed recordRole of physiological levels of 4-hydroxynonenal on adipocyte biology: implications for obesity and metabolic syndromePMID 23025469 · Experiment in cultured mouse fat cells
In plain language
Direct acute and repeated 4-HNE exposure increased oxidative stress and disrupted fat-cell development and function in a mouse cell line.
Conclusion
This supports a possible cellular mechanism involving an oxidation product.
What this doesn’t tell us
Not a human feeding study or proof that dietary seed oils cause obesity, metabolic syndrome, or insulin resistance.
Oil from the fruit of an oil palm. It contains more saturated fat than many liquid vegetable oils. Environmental sourcing questions are separate from human health evidence.
Potential human concerns
Higher saturated-fat intake can raise LDL cholesterol, a heart-disease risk factor. High-temperature refining can also create glycidyl and MCPD esters; levels depend on processing, and other refined oils can contain them too.
Process contaminants in vegetable oils and foods
EFSA · 2016 risk-assessment summary
Regulatory review of toxicology and measured food-contaminant exposure
In plain language
EFSA reviewed unwanted substances formed when oils are refined. Palm oils had the highest levels in the data it assessed. Experimental evidence raised concerns about glycidol, which can damage DNA.
Conclusion
Contaminants need control and saturated-fat intake deserves attention. This is not proof that eating palm oil causes cancer in humans.
What this doesn’t tell us
Historical measurements are not today's result for every brand. Much contaminant hazard evidence comes from animal or laboratory work. Risk depends on actual exposure. EFSA updated its 3-MCPD assessment in 2018.
Who funded the research?
Underlying oil studies and reviews
The complete funding and author-conflict statements were not independently verified in this review.
Do not interpret missing information as independence or as evidence of industry influence.
Canola is a seed oil made from low-erucic-acid rapeseed. It contains mostly unsaturated fats, including linoleic acid, an omega-6 fat. Its fat balance differs from corn, soybean, and ordinary sunflower oils. Liquid canola is not the same as partially hydrogenated oil.
Potential human concerns
Like other fats containing linoleic acid, canola can form oxidized fats when it reacts with oxygen. Some breakdown products can disrupt cells and participate in inflammation-related pathways in experiments. The linked soy-oil trial found increases in oxidized LDL and Lp-PLA2, an inflammation-related enzyme; it did not test canola. Repeated heating and storage conditions can affect oil quality, but these papers do not establish that eating canola causes inflammation in people. Refining and solvent residues require separate evidence about the actual product.
The effects of canola oil on cardiovascular risk factors: A systematic review and meta-analysis with dose-response analysis of controlled clinical trials
Amiri et al. · Nutrition, Metabolism and Cardiovascular Diseases · 2020
Systematic review and meta-analysis covering 42 controlled-trial articles
In plain language
Across controlled trials, canola generally lowered LDL cholesterol, especially compared with saturated fats. That finding answers a cholesterol question, not every question about oxidation, inflammation, or long-term health. The separate seed-oil studies examine different oils, changes in oxidation measurements, and effects of breakdown products on cells; they are not direct canola tests.
Conclusion
Crunchify examines oxidation and inflammation-related research alongside cholesterol findings. Oxidized fats are a legitimate research concern, but the supplied papers do not show that ordinary canola consumption causes chronic inflammation. Evidence about one oil or a directly exposed cell cannot establish the same effect from eating another oil. A review of 15 human trials generally found no increase in common inflammation markers with higher linoleic acid intake.
What this doesn’t tell us
Cholesterol, oxidized fats, inflammation markers, and diagnosed disease are different outcomes. Lower LDL alone does not prove fewer heart attacks; an oxidation marker alone does not prove inflammation or illness. The supplied historical trials and cell studies cannot determine canola-specific disease risk or a harmful serving size. These studies did not compare every refined and cold-pressed oil or test every product's solvent residues. EFSA's 2024 technical-hexane report identified data gaps, not demonstrated harm from normal canola consumption. High-erucic rapeseed findings should not be transferred directly to low-erucic canola. This educational entry does not change catalog admission rules.
Who funded the research?
Underlying oil studies and reviews
The complete funding and author-conflict statements were not independently verified in this review.
Do not interpret missing information as independence or as evidence of industry influence.
Silicon dioxide (E551) and calcium silicate (E552) keep powders from sticking together. They need separate assessments despite doing the same job.
Potential human concerns
Food-grade silicon dioxide raises questions about small-particle behavior. A 2024 cell and mouse study reported disrupted oral tolerance and worsened gluten-related pathology in a susceptible mouse model. EFSA's separate 2024 assessment concluded no safety concern at reported uses while retaining small-particle uncertainties. Calcium silicate has separate data gaps. Titanium dioxide is a food colour, not an interchangeable anti-caking agent; it is covered in the E-number entry.
Re-evaluation of silicon dioxide (E 551) for infant foods and follow-up for all population groups
EFSA · 2024
Updated regulatory toxicology and exposure assessment, including small-particle considerations
In plain language
EFSA reviewed food-grade silicon dioxide, including how its particles behave, and found no safety concern at the uses and amounts reported. This updates important uncertainties in the earlier review.
Conclusion
There is experimental silica evidence worth examining alongside EFSA's reassuring exposure assessment. The mouse study does not establish that E551 causes human coeliac disease or allergies. No numerical ADI does not mean either proven harm or unlimited permission.
What this doesn’t tell us
Genetically susceptible or sensitized mice differ from the general human population. EFSA used a margin-of-exposure approach and reported adequate margins for the scenarios assessed, while noting particle dispersion/characterization limitations. Food-grade amorphous silica is not inhaled crystalline-silica dust.
Who funded the research?
Silica oral-tolerance study, 2024
The full text identifies ANR and Canadian Institutes of Health Research support.
Complete grant and author-conflict verification was not completed; this does not establish the absence of other interests.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
An E number is an additive identifier, not a danger score. Vitamin C is E300, lecithin E322, and pectin E440. Some additives are natural and some are made synthetically.
Potential human concerns
Authorization can change when new toxicology appears. Titanium dioxide (E171), a colour, lost EU food authorization because EFSA could not rule out DNA-damage concerns. Rat experiments reported immune changes and preneoplastic colon lesions. These concerns attach to titanium dioxide, not to every ingredient carrying an E number.
Food additives
World Health Organization · 2023
Public guidance explaining substance-specific safety assessment
In plain language
Additives do different jobs, such as preventing spoilage or keeping texture consistent. Safety reviews consider the exact substance, toxicology, and how much people are expected to eat.
Conclusion
E numbers are identifiers, not danger scores. E171 demonstrates why named ingredients deserve review and why historical authorization is not a permanent guarantee. Its EU removal does not establish that E300 vitamin C or E440 pectin shares its hazard.
What this doesn’t tell us
This is guidance, not a trial of every additive. Authorization is not a promise of zero risk at every dose, and assessments can change. An acceptable daily intake is a long-term exposure estimate, not a sharp line between safe and poisonous.
Who funded the research?
Bettini titanium dioxide experiment
The complete funding and author-conflict statements were not independently verified in this review.
Do not interpret missing information as independence or as evidence of industry influence.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
MSG (E621) adds a savoury taste. Glutamate also occurs naturally in foods such as tomatoes and cheese.
Potential human concerns
A Chinese prospective cohort linked higher MSG intake with later overweight; observational findings cannot establish MSG as the cause. High-dose animal neurotoxicity research used conditions unlike ordinary mixed meals. Controlled challenges found short-lived symptoms in some self-described sensitive people after large doses without food, but responses were poorly reproducible. EFSA set a group intake guideline and identified possible exposure exceedances in some populations.
Questions and Answers on Monosodium glutamate (MSG)
US FDA · 2012
Regulatory safety explanation drawing on expert review and controlled studies
In plain language
FDA considers MSG generally recognized as safe. Blinded studies have not consistently reproduced people's self-reported reactions, though some may react to a large amount taken without a meal.
Conclusion
Weight-related associations, acute symptom responses, and the EFSA exposure assessment warrant transparent, dose-specific discussion. They do not establish that ordinary MSG food use causes brain injury, chronic inflammation, obesity, or cancer in every consumer.
What this doesn’t tell us
Diet and lifestyle can confound a cohort association. Injected or very large animal doses cannot be treated as ordinary oral human exposure. Acute challenge studies do not establish lifetime safety. EFSA's group ADI is 30 mg/kg/day expressed as glutamic acid for E620–625 combined; it is a long-term guideline, not a legal serving limit or a count of all protein-bound glutamate.
Who funded the research?
He et al. 2011 cohort
NIH grants supported the study and cohort; pilot work had university and Chinese CDC support. Authors reported no conflicts.
A disclosed public funder does not resolve confounding or dietary measurement error.
The companion report explicitly states support by a grant from the International Glutamate Technical Committee (IGTC).
This is a potential industry-related conflict. It is the same multicenter research, not an independent replication; the separate JACI article’s complete disclosure was not independently verified.
The opinion lists Ajinomoto-submitted dossiers/studies and an IGTC literature submission.
Submitted data are not proof that these organizations financed EFSA or controlled the opinion. Individual panel interests and every underlying study’s funding were not audited.
MSG intake and incidence of overweight in Chinese adultsPMID 21471280 · Prospective observational cohort
In plain language
Higher estimated MSG consumption was associated with greater incidence of overweight during follow-up.
Conclusion
A human association worth investigating, not a proven causal effect.
What this doesn’t tell us
Intake measurement and residual confounding remain possible; no random dietary assignment.
Read the PubMed recordMulticenter double-blind MSG challengePMID 11080723 · Repeated crossover challenges in 130 self-selected sensitive volunteers
In plain language
Five grams without food produced more initial symptom responses, but repeated responses were inconsistent and persistent reactions with food were not demonstrated.
Conclusion
Some acute symptoms can occur under bolus-challenge conditions.
What this doesn’t tell us
Selected participants, attrition and tiny late-stage samples; not a chronic disease trial.
Sodium benzoate, potassium sorbate, sulphites, and nitrites slow spoilage. They are different chemicals with different benefits and concerns, not one risk category.
Potential human concerns
Sulphites can provoke reactions in susceptible people, especially some people with asthma. Nitrites can contribute to nitrosamine formation, while also helping prevent botulism. Benzoate plus vitamin C can form benzene in some beverage formulations, depending on conditions. A direct-cell sorbate study reported positive DNA/chromosome endpoints in some assays; those exposures are not the same as eating food. BHA is an antioxidant and potassium bromate a flour-treatment agent, not the same preservative.
Sulfites: safety concern for high consumers, but data lacking
EFSA · 2022
Updated regulatory assessment using animal evidence and dietary exposure estimates
In plain language
EFSA found too little evidence to keep its earlier temporary daily-intake limit for sulphites and flagged possible concern for people who eat large amounts. This did not demonstrate neurological injury in people.
Conclusion
Specific preservative and processing concerns deserve attention without combining distinct hazards. Benzoate is not benzene; formation is not inevitable. Direct-cell sorbate findings must be weighed against metabolism and regulatory toxicology. Ingredient, dose, formulation, susceptibility and preservation benefits all matter.
What this doesn’t tell us
An exposure estimate above a guideline is not proof someone was harmed. Processed-meat cancer evidence concerns the food as eaten, not proof that nitrite alone explains the risk. Older assessments do not measure a particular current product.
Who funded the research?
Sorbate cell assay and benzene model study
The complete funding and author-conflict statements were not independently verified in this review.
Do not interpret missing information as independence or as evidence of industry influence.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Colours change a food's appearance. Specific dyes have different evidence and rules; food dyes are not the same as textile dyes in the household evidence below.
Potential human concerns
The Southampton trial found small increases in aggregate hyperactivity scores for some age/mixture comparisons after drinks containing dyes plus sodium benzoate. These effects deserve attention, but the mixtures do not identify which ingredient was responsible. Red No. 3 has a separate high-dose male-rat cancer history and US legal withdrawal; it is not one of the Southampton six.
Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community
McCann et al. · The Lancet · 2007
Randomized double-blind placebo-controlled crossover trial in 297 children
In plain language
Drinks contained sodium benzoate plus one of two dye mixtures. Some comparisons showed a small increase in activity/attention-related scores, varying by age, mixture and analysis. This supports concern about specific mixtures rather than proof that all dyes cause ADHD.
Conclusion
Human behavior signals underpin a warning for six specified EU colours. They do not establish identical effects for each dye. Red No. 3 withdrawal is a separate legal action based on animal carcinogenicity; FDA describes that mechanism as rat-specific.
What this doesn’t tell us
Short-term scores are not a clinical ADHD diagnosis. Mixture designs cannot separate benzoate from individual dyes. EFSA’s 2008 assessment considered the evidence insufficient to change individual ADIs on this trial alone; that is not proof there was no effect. US food reformulation for Red No. 3 is due January 15, 2027, with transition provisions; it is incorrect to describe every currently permitted dye as already banned.
Who funded the research?
Southampton child-behavior trial
UK Food Standards Agency funding. The original paper states the sponsor had no role in design, collection, analysis, interpretation or writing.
The complete author-conflict statement was not independently verified. Public funding does not remove mixture-attribution or statistical limitations.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
A broad label for substances that give food a particular taste. FDA permits some flavours to be named collectively; natural flavour labels can hide individual identities too.
Potential human concerns
Some named synthetic flavouring substances have raised animal-toxicology concerns. FDA removed six after animal cancer findings triggered a statutory prohibition, while concluding their very low intended-use exposure did not pose a public-health risk. A collective flavour label does not disclose the actual substances; that is a transparency concern, not proof every flavouring causes cancer.
FDA Removes 7 Synthetic Flavoring Substances from Food Additives List
US FDA · 2018
Regulatory decision explaining animal hazards and very low intended-use exposure
In plain language
FDA removed six synthetic flavourings because animal cancer findings triggered a legal rule, and another because industry no longer used it. FDA also concluded that the six did not pose a public-health risk at their very low intended-use amounts.
Conclusion
The substance and exposure matter. Artificial flavouring is not one toxic chemical, and natural does not automatically mean safer. Legal permission to use a collective label is different from Crunchify's full-disclosure standard.
What this doesn’t tell us
This decision is not a human cancer study or a review of every flavour. Broad labels support neither a blanket harm claim nor a universal safety guarantee.
Who funded the research?
FDA 2018 flavouring decision
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Xanthan gum (E415) thickens food and keeps mixtures together. It is made with the help of bacteria during fermentation.
Potential human concerns
Premature infants developed severe necrotizing enterocolitis after receiving a commercial xanthan-containing feed thickener in case reports and FDA adverse-event reviews. This is a serious product/population-specific signal, not proof that every xanthan food use has that effect. A separate mechanistic study found gut bacteria capable of digesting xanthan and changes in colonized mice; a microbiome change is not automatically damage. Larger amounts may cause digestive symptoms.
Re-evaluation of xanthan gum (E 415) as a food additive
EFSA · 2017
Regulatory review including toxicology and human tolerance studies
In plain language
EFSA found no safety concern at assessed food-use levels and did not consider a numerical daily-intake limit necessary. Larger supplemental amounts affected stool output or bowel frequency in tolerance studies.
Conclusion
Infant feed-thickener reports must not be dismissed or generalized to ordinary adult foods. Use infant medical thickeners only with professional guidance. Xanthan’s interaction with gut microbes deserves study; the cited microbiome work does not establish chronic human inflammation.
What this doesn’t tell us
Case reports have no randomized control and cannot establish that xanthan alone caused NEC or identify its mechanism. Prematurity and medical feeding differ from adult dietary exposure. The 2022 microbiome work includes human-derived microbial samples and mouse experiments, not a human disease trial. EFSA’s food-use review remains reassuring for assessed uses, with separate infant assessments.
Who funded the research?
Xanthan microbiome study, 2022
Acknowledgments identify US NIH, the Research Council of Norway, University of Michigan gastrointestinal research support and US Department of Energy facility support. Five authors disclosed an enzyme-related patent application (PCT/US2021/050494); Martens disclosed consulting/advisory service for Novome Biotechnologies. Other authors declared no competing interests.
Patent and consulting interests are relevant context, not proof that microbial findings are false. The study is not a human disease trial.
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Late-onset NEC following a xanthan-containing thickening agentPMID 22575248 · FDA review of 22 exposed infant cases
In plain language
Adverse-event reports suggested an association and an unusual illness pattern after commercial thickener use.
Conclusion
A serious safety signal in a vulnerable infant population.
What this doesn’t tell us
Not randomized; does not prove xanthan alone was causal or establish an adult-food risk.
Read the PubMed recordNEC in premature infants receiving SimplyThickPMID 22289705 · Three clinical case reports
In plain language
Three premature infants developed late-onset colonic NEC after receiving thickened feeds.
Conclusion
Supports caution about untested medical uses in premature infants.
What this doesn’t tell us
Few cases, no control group, and no proven mechanism.
Read the PubMed recordMechanistic insights into xanthan consumption by gut microbiotaPMID 35365790 · Human-derived microbiome analyses and colonized-mouse experiments
In plain language
Identified bacterial xanthan digestion and a cross-feeding pathway; feeding xanthan expanded relevant bacteria in colonized mice.
Conclusion
Shows microbial adaptation, not that human gut changes are inherently harmful.
What this doesn’t tell us
Not an adult human feeding trial measuring inflammation or clinical disease.
Guar gum (E412) comes from guar beans. It thickens foods, and gut bacteria can break it down much like some dietary fibres.
Potential human concerns
Guar can cause gas, bloating and abdominal discomfort at larger intakes. Dry concentrated gum can swell; improper supplement use and difficulty swallowing require separate attention. Native guar, partially hydrolyzed guar and normal food formulations are not interchangeable. Infant and special-medical feeding need their own assessments rather than an automatic extension of adult food-use conclusions.
Re-evaluation of guar gum (E 412) as a food additive
EFSA · 2017
Regulatory toxicology and human tolerance review
In plain language
EFSA found no safety concern for the general population at assessed uses and did not require a numerical daily-intake limit. It recommended attention to abdominal discomfort in infants and young children and noted gaps for certain medical foods.
Conclusion
Individual digestive tolerance matters, but guar gum in ordinary foods is not established as broadly harmful.
What this doesn’t tell us
Food servings, native guar, and partially hydrolysed guar supplements differ. A general-population assessment does not automatically answer questions about infant feeding or every person's tolerance.
Who funded the research?
EFSA guar assessment
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Opens the original source; full text may require access.
Research context · Not a blanket exclusion
Thickeners
Modified starch, pectin (E440), and agar (E406) help foods hold their shape or become thicker. Pectin is found in fruit; agar comes from seaweed. Modified starch does not necessarily mean genetically modified starch.
Potential human concerns
Large quantities can cause digestive discomfort. The digestibility and source of a starch matter for glucose response and allergy/coeliac labeling; pectin and agar are not equivalent to rapidly digested starch. Gel-forming ingredients can create a physical choking concern in specific confectionery formats, which differs from a toxicological effect after swallowing. Ordinary food and infant medical formulations need separate assessment.
Re-evaluation of modified starches as food additives
EFSA · 2017
Regulatory review of multiple specific modified starches
In plain language
EFSA found no safety concern at reported general-population food uses for the modified starches it assessed. Separate assessments of pectin and agar also found no safety concern at assessed uses.
Conclusion
The exact thickener and amount matter. Being used to change texture is not evidence that a food ingredient damages the gut.
What this doesn’t tell us
These are not trials showing every thickener has identical effects. Food-use reassurance does not mean unlimited concentrated supplements or infant medical feeds are appropriate. Agar exposure data covered only a small share of authorized food categories.
Who funded the research?
EFSA modified starches, pectin and agar assessments
This is an institutional regulatory assessment, not a clinical trial. It evaluates published research and may include industry-submitted data.
Individual expert interests and the sponsorship of every underlying study were not independently audited. Submission of data is not proof that industry funded or controlled the assessment.
Permission depends on the exact ingredient, food category, and amount. An EU acceptable daily intake is a scientific exposure guideline, not a legal per-serving limit. Authorization is not a health guarantee; a restriction is not proof of human illness.
Sources reviewed October 27, 2026. Rules and transition dates can change. Infant formula and foods for special medical purposes have separate rules.
CMC (E466) and gellan gum (E418)EU: Authorized for specified uses
European Union
Authorized in specified food categories under additive rules; this is not blanket permission in every infant food.
United States
Specified technological food uses are permitted under applicable ingredient rules.
Different substances need separate assessments. CMC findings cannot establish gellan harm.
Authorized for specified uses in general foods. Ordinary infant/follow-on formula must be distinguished from infant foods for special medical purposes, which have specific authorized uses.
United States
Permitted under 21 CFR 172.620, subject to its specifications and technological-use conditions.
A claim that it is banned from every EU infant medical food is incorrect. Food-grade carrageenan and poligeenan are not interchangeable.
Used as an ordinary food ingredient, not a single E-number additive category. Annex II of Regulation 1169/2011 expressly exempts wheat-based maltodextrins from wheat-allergen labeling, subject to its qualification.
United States
Affirmed GRAS under 21 CFR 184.1444 when specifications and good manufacturing practice are met.
The earlier blanket claim that wheat-derived maltodextrin always requires an EU wheat-allergen warning was incorrect.
Ordinary edible oil use is allowed subject to food composition, contaminant and processing requirements; these oils are not a single E-number category.
United States
Ordinary edible oil use is permitted subject to applicable food standards. Partially hydrogenated oil has separate rules.
Oil identity, refining contaminants and heating conditions are separate from blanket claims about seed oils.
Ordinary edible oil use is allowed subject to food composition, contaminant and processing requirements; these oils are not a single E-number category.
United States
Ordinary edible oil use is permitted subject to applicable food standards. Partially hydrogenated oil has separate rules.
Oil identity, refining contaminants and heating conditions are separate from blanket claims about seed oils.
Rapeseed oil / canola oilEU: Authorized for specified uses
European Union
Ordinary edible oil use is allowed subject to food composition, contaminant and processing requirements; these oils are not a single E-number category.
United States
Ordinary edible oil use is permitted subject to applicable food standards. Partially hydrogenated oil has separate rules.
Oil identity, refining contaminants and heating conditions are separate from blanket claims about seed oils.
Silicon dioxide (E551)EU: Authorized for specified uses
European Union
Authorized for specified uses. EFSA 2024 found no safety concern at reported use levels; it retained uncertainties and used margins of exposure rather than a numerical ADI.
United States
21 CFR 172.480 allows anticaking use where effective, no more than necessary, generally up to 2% by food weight.
A positive mouse oral-tolerance study and a reassuring regulatory exposure assessment answer different questions.
Regulation 2022/63 withdrew food authorization, with transition provisions for foods made under the previous rules before August 7, 2022. This is not a blanket ban in medicines or cosmetics.
United States
Federal food-colouring authorization continues, generally up to 1% by food weight; FDA lists the repeal petition as under review.
EFSA could not rule out genotoxicity. Withdrawal is not a finding that it causes human cancer.
MSG (E621) and glutamates (E620–625)EU: Authorized for specified uses
European Union
Authorized with food-category conditions. EFSA 2017 set a group ADI of 30 mg/kg body weight/day as glutamic acid; modeled exposure can exceed it in some groups.
United States
FDA considers intended MSG use GRAS and acknowledges possible transient symptoms after large doses without food.
The ADI is not a universal legal food concentration or acute poisoning threshold.
Sodium benzoate (E211) and potassium sorbate (E202)EU: Authorized for specified uses
European Union
Authorized under category-specific limits. EFSA’s 2019 group sorbate ADI is 11 mg/kg/day as sorbic acid for E200/E202.
United States
Specified uses permitted under good manufacturing practice; sodium benzoate 21 CFR 184.1733 and potassium sorbate 21 CFR 182.3640.
Potential benzene formation is a formulation issue, not a blanket benzoate ban. 5 ppb is a water benchmark, not a universal statutory soft-drink ceiling.
Erythrosine / Red No. 3 (E127)EU: Specific restrictions
European Union
Narrowly authorized, principally for specified cherry products.
United States
FDA revoked authorization; food reformulation deadline January 15, 2027, and ingested-drug deadline January 18, 2028. Applicable transition provisions allow certain previously manufactured products to remain on sale.
FDA invoked the Delaney Clause after male-rat cancer findings and states that available evidence does not establish human cancer risk at relevant exposure.
Artificial flavouring as a categoryEU: Authorized for specified uses
European Union
Regulated through Regulation 1334/2008, the Union list and substance-specific restrictions; there is no single ban on artificial flavouring.
United States
Individual uses are regulated. FDA removed six synthetic flavouring uses in 2018 under the Delaney Clause and delisted styrene separately due to abandonment.
The exact current EU list entries of all six US-delisted substances were not fully audited. Do not infer a matching EU ban or authorization.
Authorized for specified food uses, including some special medical foods subject to conditions. Gums are restricted in jelly mini-cups because of choking risk; a blanket EU infant-medical-food ban is incorrect.
United States
Food uses permitted under 21 CFR 172.695. Commercial infant feed-thickener safety is a separate clinical issue.
Ordinary infant formula, special medical food and commercial thickener use are different categories.
Authorized in specified general foods. Regulation 2026/196 deletes E412 from infant formula (13.1.1) and infant special-medical-food (13.1.5.1) entries, generally applying from August 18, 2026 with transition provisions. A narrow 13.1.5.2 medical use is time-limited until April 27, 2027.
United States
Food uses permitted under 21 CFR 184.1339; dry concentrated supplement use is not the same as a food serving.
The complete authentic annex and transition text could not be retrieved in this review; these changes were cross-checked against a reproduction, not a full legal sign-off. Older general-food assessments do not establish current infant-formula permission.
Who pays for a study can matter. Each food entry identifies the disclosures we could check, including manufacturer support, public funding, and information we could not verify. We apply the same scrutiny to research reporting concerns and research offering reassurance.
Industry support is a potential conflict, not proof that a result is false. University affiliation is not proof of independence. Regulatory assessments may use industry-submitted data as well as published research; the assessors and the underlying studies are not the same. Missing disclosures are labeled unknown, not independent.
GRAS reviews: a documented independence concern
A 2013 audit of 451 voluntarily submitted US GRAS notices from 1997–2012 found that assessments came from manufacturer employees (22.4%), manufacturer-selected consultants (13.3%), or manufacturer/consultant-selected panels (64.3%). None used a standing expert panel selected by an independent third party. These are documented financial relationships, not evidence that 99% of additives are harmful.
Neltner et al. · JAMA Internal Medicine · PMID 23925593. Pew supported the research; the complete individual disclosure audit was not completed. This historical sample does not describe every current approval. FDA’s 2026 mandatory-notification initiative is a proposal, not an implemented repeal of current GRAS rules.
A 2020 documentary study examined 15,078 pages obtained through freedom-of-information requests and identified examples of ILSI promoting industry-aligned research and messages. That supports scrutiny of particular activities, not a claim that every ILSI-linked paper is false or every additive review was manipulated.
Steele, Ruskin & Stuckler · Public Health Nutrition · PMID 32416734. The Laura and John Arnold Foundation funded the work. Ruskin disclosed employment by US Right to Know and its donors, including Organic Consumers Association and Dr. Bronner’s; Stuckler disclosed ERC/Wellcome support. A selected documentary record is not a comprehensive audit of additive toxicity.
This is a targeted literature and source-disclosure review, not an exhaustive search of all PubMed records or a completed audit of every underlying study. Sources, designs, doses, and limitations are provided so the evidence can be checked.
Household ingredients we screen for
Our existing exclusion rules for cleaning, laundry, personal-care, and baby products. HOCl is included as allowed-ingredient context, not an exclusion.
Ingredient context · Not automatically excluded
Hypochlorous acid (HOCl)
Dilute, properly formulated HOCl can be suitable for label-directed household surface use. It is not automatically harmful, but it is not harmless in every use. Concentration, pH, freshness, other ingredients, and how exposure occurs all matter. This evidence does not certify any listed brand or product.
HOCl and sodium hypochlorite bleach share related chlorine chemistry, but a dilute HOCl product is not the same formulation as concentrated household bleach. The fact that immune cells produce HOCl does not prove that every external exposure is safe. Crunchify’s bleach exclusion does not automatically exclude HOCl.
Use limitations
Use a surface cleaner only for its labeled purpose. Do not drink it, use it as skincare or wound treatment, or intentionally inhale its mist.
Do not extrapolate surface-use evidence to room fogging or nebulizing. EPA has not evaluated unlisted application methods; fogging requires explicit product-label authorization.
Surface disinfection with hypochlorous acid: Phi6 laboratory carrier study
Hamilton et al. · Food and Environmental Virology · 2023
In plain language
Researchers sprayed a surrogate virus dried onto steel and vinyl with 500 or 1,000 ppm HOCl for 5, 30, or 60 seconds. The reduction in infectious virus depended on concentration, contact time, surface, and surrounding material.
Conclusion
HOCl has antimicrobial activity, but the ingredient name alone does not establish a product’s performance.
What this doesn’t tell us
Controlled laboratory surfaces and a surrogate virus are not household transmission or human safety tests. These experimental concentrations and times are not directions for a catalog product.
Effect of stabilized hypochlorous acid on re-epithelialization and bacterial bioburden in acute wounds
Burian et al. · Acta Dermato-Venereologica · 2022
In plain language
Twenty healthy adults had small standardized wounds treated with a medical solution containing 0.016% HOCl and 0.25% acetic acid, or saline. The study reported improved early healing and lower bacterial counts with the tested formulation.
Conclusion
This offers limited support for this particular medical formulation’s short-term tolerability, not permission to apply surface cleaners to people.
What this doesn’t tell us
Small adult sample, short follow-up, manufacturer involvement, and a combined formulation. It does not establish chronic, child, inhalation, or household-cleaner safety; effects cannot be assigned to HOCl alone.
Sources reviewed 2026-10-22 · Educational information, not medical advice.
Household exclusion rules
Sources reviewed October 2026 · Original publication years shown below
Parabens
Parabens (methylparaben, propylparaben, butylparaben and similar) are preservatives used in some lotions, wipes, and cleaners. We screen out any ingredient named as a paraben.
Concentrations of parabens in human breast tumours
Darbre et al. · Journal of Applied Toxicology · 2004
Small laboratory measurement study of 20 tissue samples
In plain language
Researchers measured parabens in samples of breast tumour tissue and found them in most samples, mostly methylparaben, in a form suggesting they had been absorbed largely unchanged.
Conclusion
Parabens from everyday products can be found intact in human tissue, which supports further study of how they are absorbed and stored.
What this doesn’t tell us
Only 20 samples, with no comparison to healthy tissue. The study does not show that parabens cause cancer or identify which products were the source.
Opens the original source; full text may require access.
Phthalates
A family of chemicals studied for possible effects on hormone-related development.
Decrease in anogenital distance among male infants with prenatal phthalate exposure
Swan et al. · Environmental Health Perspectives · 2005
Observational study of mothers and young boys
In plain language
Researchers measured phthalate breakdown products in mothers’ urine during pregnancy and compared them with physical measurements in their sons. Higher levels of several phthalates were associated with a shorter distance between the anus and genitals, a measurement used in research on reproductive development.
Conclusion
The findings support further research into whether exposure before birth affects male reproductive development; an association alone does not prove cause and effect.
What this doesn’t tell us
This was a relatively small observational study, not a household-cleaner test. It did not identify cleaners as the source of exposure or establish a child’s future fertility.
Opens the original source; full text may require access.
Chlorine bleach
Our exclusion covers chlorine bleach, not every chlorine-containing substance. This paper concerns accidental chlorine-gas releases; it does not assess dilute hypochlorous acid (HOCl) cleaners.
Civilian exposure to chlorine gas: A systematic review
Govier & Coulson · Toxicology Letters · 2018
Review of 36 papers covering 1,566 exposed people
In plain language
Researchers brought together reports of civilian chlorine-gas incidents. Cough, breathlessness, and a sore throat were common; among people with follow-up information, 90% were reported to have recovered fully.
Conclusion
Accidental chlorine-gas exposure can cause acute irritation of the airways, eyes, and throat; many reported patients recovered.
What this doesn’t tell us
Follow-up was available for only 60% of people, and exposure conditions varied. The review does not measure risk from ordinary household-cleaner use or from every chlorine-containing compound.
Opens the original source; full text may require access.
Quats (ammonium compounds)
Quaternary ammonium compounds are used in some disinfectants. The study here tests one of them in freshwater organisms.
Toxicity and genotoxicity of the quaternary ammonium compound benzalkonium chloride (BAC) using Daphnia magna and Ceriodaphnia dubia as model systems
Lavorgna et al. · Environmental Pollution · 2016
Laboratory study in two freshwater crustacean species
In plain language
Researchers exposed two types of tiny freshwater animals to benzalkonium chloride, a quat used in disinfectants. At the tested concentrations, they found toxic effects and evidence of DNA damage.
Conclusion
This particular quat affected aquatic organisms under controlled laboratory conditions.
What this doesn’t tell us
The paper studied one quat, not all ammonium compounds. It does not show human health effects, prove persistence in waterways, or measure harm from the concentrations found at a particular real-world location.
Opens the original source; full text may require access.
Sulfates and SLS
Epsom salt (magnesium sulfate) is chemically a sulfate, but it is admissible under Crunchify’s ingredient screen. Other sulfates and SLS/SLES remain excluded. Sodium lauryl sulfate (SLS) is a cleansing ingredient. This paper examines skin irritation from SLS, not all chemicals called sulfates.
Sodium lauryl sulphate for irritant patch testing—a dose-response study using bioengineering methods for determination of skin irritation
Agner & Serup · Journal of Investigative Dermatology · 1990
Controlled skin patch tests in 12 volunteers
In plain language
Researchers tested four SLS concentrations on volunteers’ skin and measured responses after 24 and 48 hours. Higher concentrations produced greater irritation, including increased water loss through the skin.
Conclusion
Skin irritation increased with SLS concentration under these patch-test conditions.
What this doesn’t tell us
The study was small, and prolonged patch exposure differs from brief washing and rinsing. It does not establish effects for every sulfate, eye irritation, or risk from every finished product.
Opens the original source; full text may require access.
Synthetic dyes
Some dyes can trigger contact allergy. The available study here concerns textile dyes, so its relevance to cleaner colorants is limited.
Patch Testing to a Textile Dye Mix by the International Contact Dermatitis Research Group
Isaksson et al. · Dermatitis · 2015
Patch-testing study of 2,493 dermatitis patients at nine clinics
In plain language
Researchers tested a mix of eight textile dyes on patients being evaluated for dermatitis. Ninety patients (3.6%) reacted to the mix, and ordinary baseline allergy tests would have missed some dye allergies.
Conclusion
Testing for textile dye allergy can identify reactions that standard testing misses in some dermatitis patients.
What this doesn’t tell us
These were patients with dermatitis, not the general population. Textile dyes are not the same as dyes in cleaners; this paper does not demonstrate allergy risk from household-cleaner colorants or all synthetic dyes.
Steinemann et al. · Environmental Impact Assessment Review · 2011
Laboratory emissions analysis of 25 fragranced consumer products
In plain language
Researchers tested the air above fragranced cleaning supplies, laundry products, personal-care products, and air fresheners. They detected 133 different volatile chemicals across the products; only one of those chemicals was listed on any product label.
Conclusion
The labels did not identify most of the chemicals detected in emissions from the tested products, illustrating a disclosure gap.
What this doesn’t tell us
The paper measured emissions and labeling, not people’s exposure or health outcomes. Emitted chemicals are not necessarily intentionally added ingredients, and the study explicitly makes no claims about risks from product use. Its findings reflect the products and labeling rules at the time.
Opens the original source; full text may require access.
Sodium hydroxide
A strong alkali used in some oven cleaners. Concentration and the way a person is exposed matter when interpreting the evidence.
Toxicity resulting from exposure to oven cleaners as reported to the UK National Poisons Information Service (NPIS) from 2009 to 2015
Day et al. · Clinical Toxicology · 2017
Review of poison-service enquiries involving 780 patients
In plain language
Researchers reviewed reports of swallowing, inhaling, or getting oven cleaners on skin or in eyes. Most products contained sodium hydroxide and/or potassium hydroxide, and some exposures involved burns or other symptoms.
Conclusion
Reported exposures to strong-alkali oven cleaners included corrosive injuries, showing why exposure conditions matter.
What this doesn’t tell us
These were enquiries about mixed cleaning products, not tests of pure sodium hydroxide. The records cannot assign each injury to one ingredient or estimate risk during ordinary, correctly directed use.
Opens the original source; full text may require access.
Sulfuric acid
A strong acid that can be corrosive at high concentrations. This paper concerns ingestion of battery acid, not everyday cleaning.
Battery acid burns of the upper gastro-intestinal tract
Wormald & Wilson · Clinical Otolaryngology and Allied Sciences · 1993
Retrospective case series of 18 battery-acid ingestions
In plain language
Doctors reviewed 18 patients who swallowed battery acid containing 30% sulfuric acid. Injuries included burns in the upper digestive tract, and four patients needed surgery to restore function.
Conclusion
Swallowing concentrated sulfuric acid caused serious corrosive injuries in this case series.
What this doesn’t tell us
This is a small study of concentrated battery-acid ingestion. It does not establish effects from dilute solutions, skin contact, inhalation, or normal household-cleaner use.
Opens the original source; full text may require access.
Microplastics and plastic-derived materials
Tiny plastic particles can come from synthetic materials such as polyester, polypropylene, and polyethylene — including the fabrics used in some baby wipes and diapers. Our screening asks what a wipe or diaper is actually made of, not just what lotion is on it.
Occurrence of Polyethylene Terephthalate and Polycarbonate Microplastics in Infant and Adult Feces
Laboratory analysis of fecal samples from 6 infants and 10 adults
In plain language
Researchers measured two common microplastics in stool samples from infants and adults in New York. Infant samples contained markedly higher concentrations of PET (polyester) microplastic than adult samples, which the authors suggest may relate to babies’ contact with plastic-containing products such as bottles, toys, and synthetic textiles.
Conclusion
Infants can carry measurable amounts of microplastic in their bodies, and everyday plastic-containing products are plausible sources worth studying further.
What this doesn’t tell us
This was a very small pilot study that measured presence, not health effects. It did not test wipes or diapers specifically, did not prove any product caused harm, and cannot say what a safe level is. It supports asking what baby products are made of — it does not condemn any single material.
