Don’t take our word for it. Here’s every fact on this site, with its evidence grade, its caveats, and a link to the source, so you can check us.
We start you on the strong ones: findings that have been replicated or come from one large, well-controlled study. Moderate and weak facts are here too, labeled for what they are. Use “Link to this fact” to share a single card.
Health effects / Other evidence
CTV quoted a Texas professor emerita (Dr. Claudia Miller) saying her research is finding biological causes make the immune system overreact in people with environmental sensitivities, and that solutions are cleaner, smoke-free homes.
Weak evidence: Researcher's summary of her own theory as paraphrased by CTV; not a study, and CTV also notes some researchers link symptoms to anxiety and depression. Contested.
Source: CTV News, April 2022 (via Smoke-Free Housing BC; redacted copy), CTV News (news article) Link to this fact
Health effects / Other evidence
A friend with MCS says the difficulty of separating the condition from housing barriers is that, with safe housing and accommodation, people "would choose to continue living," and that people may seek MAiD when years of attempts to improve conditions fail.
Weak evidence: Opinion of one person; no data; says nothing about any specific individual's reasons.
Source: Chemical Sensitivity Podcast ep. 101, 2026 Link to this fact
Health effects / Measured in people or real products
In the same Korean study, the dibutyl phthalate (DBP) breakdown product MNBP went with more omission and commission errors on a computer attention test, but not with teacher ratings; DEHP products showed the opposite pattern, so the two outcome types did not agree.
Weak evidence.
Caveats
Single cross-sectional study; one nominally significant DEHP result (MEOP omission errors) rests on an inconsistent table entry; no multiplicity correction across about 56 tested cells; one spot urine.
Health effects / Measured in people or real products
Of the low-molecular-weight metabolites, the metabolite of diethyl phthalate (MEP, the fragrance-solvent marker) was statistically flagged for several behavior scales in secondary models, but these were unadjusted for the other phthalates and shown without confidence intervals, so the study cannot say whether the fragrance phthalate itself drives the pattern.
Weak evidence.
Caveats
Table 4 reports only asterisks at p<=0.05; MEP in urine also tracks perfume use and co-occurs with DBP/DiBP metabolites; Ejaredar 2015 counts this and Miodovnik 2011 as one cohort; the 2018 MoBa cohort found MEP null for ADHD.
Source: Engel 2010, Environmental Health Perspectives Link to this fact
Health effects / Other evidence
The urinary marker of the fragrance-carrier phthalate DEP (MEP), at a median of about 403 ug/g creatinine, showed no link to lower cognitive or motor scores in the Spanish INMA cohort.
Weak evidence: Single cohort; MEP also marks other sources and the fragrance mixture; no product-use data.
Source: [[src-2015-prenatal-exposure-phthalates-neuropsychological-development-during-childhood|Gascon 2015]], International Journal of Hygiene and Environmental Health Link to this fact
Health effects / Measured in people or real products
Maternal thyroid hormones (TSH, T3, T4) and preterm birth did not explain the DEHP-ADHD association in this cohort (indirect effect odds ratios 1.00), so the mechanism remains unknown.
Weak evidence: Single mid-pregnancy measurements; U-shaped T4 relation not well handled; other phthalates and thyroid antibodies untested; strong untestable mediation assumptions.
Source: Engel 2018, Environmental Health Perspectives Link to this fact
Health effects / Other evidence
Only one study in the EPA review looked at a clinical ADHD diagnosis: in a Norwegian cohort of 802 children, higher mid-pregnancy urinary DEHP metabolites went with 1.5 times the odds of ADHD (CI 1.1 to 1.9), while DEP, DBP, DIBP and DINP showed no higher odds.
Weak evidence: One study, association only, single urine sample; DEHP is a plastics and food-contact phthalate, not a fragrance one; EPA still rated the DEHP-behavior evidence 'slight'.
Source: Radke 2020, Environment International Link to this fact
Health effects / Review of other studies
Across the review's studies, MEP (the breakdown product of the fragrance solvent DEP) shows small, inconsistent associations (e.g., +15% odds of language delay per doubling; null in several cohorts) and is a marker of mixed product use, not a measure of fragrance dose.
Weak evidence: MEP reflects many DEP sources; no study in the review separated fragrance; effect sizes small and often very low grade.
Source: Almeida-Toledano 2024, Science of The Total Environment Link to this fact
Health effects / Lab study (cells or chemistry)
A fragrance musk can look estrogen-like in one lab cell system and anti-estrogen-like in another: Galaxolide acted as a full agonist in breast-cancer-cell proliferation but as a blocker of one estrogen receptor subtype, with no agonism, in kidney-derived reporter cells.
Weak evidence: Two different studies and cell types; neither shows an effect in animals or people.
Source: Schreurs 2005, Toxicological Sciences Link to this fact
Health effects / Other evidence
The pump-blocking lasted at least 24 hours after a 2-hour musk exposure in mussel gill tissue, and the authors raise the possibility that low-toxicity chemicals like musks could make other pollutants more harmful.
Weak evidence.
Caveats
Mechanism hypothesis from one ex-vivo mussel study; co-toxicant uptake was not tested in this paper; human relevance is untested and the authors estimate human body burdens are 1-2 orders of magnitude below the effective concentrations.
Source: Luckenbach 2005, Environmental Health Perspectives Link to this fact
Health effects / Measured in people or real products
In the same newborn study, most Brazelton behavior scores showed no overall association with either the low- or high-molecular-weight phthalate sum, and the low-molecular-weight sum (which includes the fragrance-solvent marker MEP) went with slightly better motor scores in boys (about 0.09 points per log unit, p = 0.01).
Weak evidence: Exploratory sex-interaction result (p = 0.09), non-linear across quartiles; MEP was not analyzed separately; effect is in the favorable direction.
Source: Engel 2009 (Mount Sinai), NeuroToxicology Link to this fact
Health effects / Other evidence
The review reports a meta-analysis of 27 mostly Chinese cross-sectional studies of children drinking fluoride-rich water, with an average IQ difference of about 7 points, while the authors say the dose-response needs further characterization.
Weak evidence: Cross-sectional studies in high-fluoride areas, not studies of fluoridated water in the US or Europe; the authors themselves ask for better dose-response data.
Source: Grandjean & Landrigan 2014, Lancet Neurology, The Lancet Neurology Link to this fact
Health effects / Other evidence
The review reports that prenatal phthalate exposure has been linked to behavioral and attention problems, mainly in boys, and that bisphenol A in pregnancy was linked to behavior and executive-function problems at age 3, but it places phthalates and BPA in a 'suspected' tier and does not grade individual phthalates.
Weak evidence.
Caveats
Narrative summary of single cohorts; the later EPA graded review found most phthalate-outcome pairs slight or indeterminate; no phthalate is named and DEP, the fragrance phthalate, is not discussed.
Source: Grandjean & Landrigan 2014, Lancet Neurology, The Lancet Neurology Link to this fact
Health effects / Other evidence
In this cohort, prenatal DEHP metabolites were not associated with ADHD-type scores (DSM ADHD Problems beta 0.03, 95% CI -0.08 to 0.15); the DEHP signals were more somatic complaints in boys (beta 0.15, 0.03-0.28) and fewer anxiety symptoms in girls (beta -0.21, -0.35 to -0.06).
Weak evidence: Sex-stratified, exploratory, no multiple-comparison correction; DEHP is a plastics and food-contact phthalate, not a fragrance one.
Source: Kobrosly 2014, Environmental Health Perspectives Link to this fact
Health effects / Other evidence
Mild thyroid hormone shortage in early pregnancy is a human-cohort-supported risk marker for child ADHD-type symptoms, which makes thyroid disruption a plausible, but so far untested, route for chemicals; this study measured no chemicals.
Weak evidence: Plausibility only; the one formal test of thyroid mediation of a chemical (DEHP) and ADHD was null (Engel 2018), and no fragrance-related chemical was tested.
Source: Modesto 2015 (Generation R), JAMA Pediatrics Link to this fact
Health effects / Measured in people or real products
In the same Puerto Rico study, urinary MEP was not associated with thyroid hormones (free T4, free T3) in repeated-measure models.
Weak evidence: Small study (n = 106), wide confidence intervals (for example free T4 +2.7%, CI -2.7 to 7.8), so it cannot rule out an effect.
Source: Johns 2015, Reproductive Biology and Endocrinology Link to this fact
Health effects / Measured in people or real products
Metabolites of the plastics phthalate DEHP were not linked to maternal thyroid hormones across two visits in pregnancy, but were linked to about 8% lower free T4 at 24-28 weeks in a visit-specific analysis (95% CI -15.3% to -0.8%).
Weak evidence.
Caveats
Single small cohort (n = 81 at that visit), visit-specific result is one of many tests with no correction, interaction p = 0.048; DEHP is a plastics and food-contact phthalate, not a fragrance ingredient.
Source: Johns 2015, Reproductive Biology and Endocrinology Link to this fact
Health effects / Measured in people or real products
In the French EDEN boys cohort, prenatal triclosan, benzophenone-3 and paraben levels showed near-null or inconsistent links to behavior (triclosan IRRs about 1.01-1.02; benzophenone-3 and parabens slightly lower scores), and none were robust to multiple-testing correction.
Weak evidence: Authors call it exploratory and the first study of these chemicals; triclosan, benzophenone-3 and parabens are personal-care ingredients but not fragrance ingredients.
Source: Philippat 2017, Environmental Health Perspectives Link to this fact
Health effects / Measured in people or real products
Because only one urine sample was taken per mother, exposure to short-lived chemicals like BPA is probably underestimated, so true associations in such cohorts could be larger or smaller than reported; the authors' correction raised one BPA odds ratio from 1.4 to 5.4.
Weak evidence: Correction uses borrowed intraclass correlations and assumes classical error; not an observed effect size.
Source: Philippat 2017, Environmental Health Perspectives Link to this fact
Health effects / Other evidence
In CHAMACOS, prenatal sums of high-molecular-weight and DEHP phthalate metabolites went with lower IQ in boys and higher IQ in girls (interaction p = 0.01), but the boys' intervals included zero and the all-children results were null.
Weak evidence.
Caveats
Exploratory sex-stratified analysis; DEHP and other HMW phthalates are plastics and food-contact chemicals, not fragrance; the supplement carries more detail; the abstract and text differ slightly on one girls' CI.
Source: Hyland 2019 (CHAMACOS), Environmental Health Perspectives Link to this fact
Health effects / Other evidence
The DEHP-and-ADHD-behavior link in this adolescent study was seen in boys (RR 1.62, 1.38-1.91) but not girls (about 1.06), with a significant sex interaction (P = .002).
Weak evidence.
Caveats
Single cohort; other sex interactions imprecise; the text prints the girls' CI with a typo; prenatal cohorts disagree on which sex is affected, so the sex pattern is not established.
Health effects / Other evidence
The same study found no association of parabens, bisphenols (BPA, BPS, BPF), benzophenone-3, triclosan or triclocarban with ADHD-type behavior in 15-year-olds.
Weak evidence.
Caveats
Null in one small cross-sectional sample with one or two spot urines per teen (misclassification biases toward null); absence of association is not proof of safety; none of these is a fragrance ingredient.
Health effects / Measured in people or real products
In a 2020 Quebec birth cohort of 345 mother-baby pairs, babies with detectable methylparaben (a cosmetic and food preservative) in their first stool had about 2.3 times the odds of a physician ADHD diagnosis at age 6-7 after weighting for maternal and family factors (95% CI 1.45-3.76).
Weak evidence.
Caveats
Single small cohort; only 65 exposed and 33 ADHD cases; the unweighted (crude) odds ratio was 2.03 (0.92-4.51), not significant; exposure was detected vs not at 5 ng/g; only methylparaben was measured; observational, so not proof of cause. Methylparaben is not a fragrance ingredient.
Source: Baker 2020, Environment International Link to this fact
Health effects / Measured in people or real products
The study could not test whether maternal thyroid hormones carry the paraben-ADHD link: methylparaben was not associated with ADHD in the 123 women with thyroid data, and preterm birth and birthweight 'mediated' only 20% and 13% of the point estimate with intervals including 1.
Weak evidence: Underpowered subset; a different chemical from the DEHP null in Engel 2018; the authors themselves call the thyroid pattern only 'consistent with' mediation.
Source: Baker 2020, Environment International Link to this fact
Health effects / Review of other studies
A 2022 European scoping review that tallied other authors' reviews, without grading study quality, put lead, phthalates and BPA in its 'moderate to high' tier for association with ADHD in children.
Weak evidence.
Caveats
Counts review conclusions rather than appraising evidence; the authors state they did not evaluate evidence quality; heavy overlap between included reviews; contrast US EPA's graded review, which rated phthalate-ADHD evidence slight or indeterminate.
Source: Moore 2022 Link to this fact
Health effects / Measured in people or real products
The phthalate signal in this child-urine study came from DEHP metabolites (plastics and food), MHPP and MNBP; MEP, the urinary marker of the fragrance phthalate DEP, was in the mixture but stayed below the contributor threshold for every outcome.
Weak evidence: Single cohort; MEP is a non-specific DEP marker, and the exposure was one spot urine; absence of a contributor signal is not proof of no effect.
Source: Oh 2024 (CHARGE), Environmental Health Link to this fact
Health effects / Measured in people or real products
Among 43 urinary chemicals, only one single-chemical result survived false-discovery-rate correction: higher triclosan went with lower inattention scores (count ratio 0.89, 95% CI 0.83 to 0.96), the opposite direction to what a harm hypothesis predicts.
Weak evidence.
Caveats
Cross-sectional, one cohort, unexplained direction; the DEHP and MNBP single-chemical associations did not survive correction; rat work points the other way and the two are not directly comparable.
Source: Oh 2024 (CHARGE), Environmental Health Link to this fact
Health effects / Other evidence
The review states EDCs have been 'proven' to disrupt thyroid hormone synthesis and transport, but for bisphenols and phthalates its own human evidence is mixed and it cites no human thyroid dose-response; only perchlorate and thiocyanate have an established sodium-iodide-symporter mechanism.
Weak evidence: Mechanism wording outruns the human data; a judgment about the review, not a finding on chemicals.
Source: Street 2024, Front Endocrinol, Frontiers in Endocrinology Link to this fact
Health effects / Animal study
In rats given 50 mg/kg/day triclosan by mouth from pregnancy through 8 weeks of age, male offspring moved about 1.9 times faster in an open field, entered the open arms of a plus maze about 1.7 times as often (about 48% vs 28%) and scored lower on novel-object recognition than controls.
Weak evidence: Single lab, one dose, males only, 6 litters per group; values read from figures are approximate; behavior tests are proxies; animal result, not a human finding.
Source: Cui 2024, Ecotoxicology and Environmental Safety Link to this fact
Health effects / Animal study
The drug atomoxetine, which raises prefrontal dopamine and norepinephrine, suppressed the triclosan-exposed rats' hyperactivity and open-arm behavior, but pushed them below control activity, and it is not specific to dopamine, so the rescue does not prove dopamine deficiency was the cause.
Weak evidence: Single dose of atomoxetine (3 mg/kg), n = 6 litters; no ATX-only mechanistic controls beyond DA levels; norepinephrine effect not separated.
Source: Cui 2024, Ecotoxicology and Environmental Safety Link to this fact
Health effects / Animal study
The 50 mg/kg/day oral rat dose is roughly three orders of magnitude per kilogram above a single 4 mg human oral triclosan dose (about 0.06 mg/kg), and Swedish urine surveys found triclosan above detection in only about 37% of samples, so the rat result does not describe typical consumer exposure.
Weak evidence: Comparison is our own arithmetic, not the paper's; no allometric scaling; chronic developmental dosing differs from single dose; human matrices differ (urine, plasma).
Source: Cui 2024, Ecotoxicology and Environmental Safety Link to this fact
Health effects / Measured in people or real products
Higher prenatal pentachlorophenol (a wood preservative) was linked to lower cognitive scores only among children of non-smoking mothers (-2.17 points, 95% CI -3.83 to -0.51), not among children whose mothers had tobacco exposure.
Weak evidence: Exploratory interaction (p = 0.012) among 36 tests, no correction; only 113 non-smokers; direction opposite to the usual expectation; not a personal-care or fragrance chemical.
Source: Zhou 2025, Environmental Research Industry-funded Link to this fact
Health effects / Measured in people or real products
In 1,297 European children aged 6 to 11, higher urinary methyl-paraben and the phthalate metabolites MEHP, oh-MiNP and oxo-MiNP went with slightly more variable reaction times on a computer attention test (about 2 to 4% higher on the log scale for a 10th-to-90th percentile rise).
Weak evidence.
Caveats
Cross-sectional with a pooled two-sample urine reflecting about one day; 4 nominal hits among 21 chemicals with no multiple-comparison correction reported; oh-MiNP lower CI is 0.001; effects are small.
Source: Fabbri 2025 (HELIX), Environmental Research Link to this fact
Health effects / Measured in people or real products
In this child-urine study, MEP, the urinary marker of the fragrance phthalate DEP, was not associated with attention variability (log MC -0.001, 95% CI -0.025 to 0.023) or with cortisol, cortisone or corticosterone production.
Weak evidence: MEP is a non-specific DEP marker; one pooled two-day urine reflects about one day of exposure; a null here is not proof of no effect.
Source: Fabbri 2025 (HELIX), Environmental Research Link to this fact
Health effects / Measured in people or real products
In the same children, many chemicals (BPA, ethyl-paraben, MEHP, oh-MiNP, oxo-MiNP, methyl-paraben and some pesticide metabolites) went with higher urinary cortisol, cortisone and corticosterone production, with BPA largest (about 0.26 to 0.29 on the log scale).
Weak evidence.
Caveats
Cross-sectional; hormones from the night urine and chemicals from a different pooled sample; no multiplicity correction; mechanism (e.g. 11-beta-HSD2 inhibition) is from in-vitro work cited by the authors.
Source: Fabbri 2025 (HELIX), Environmental Research Link to this fact
Health effects / Measured in people or real products
The study found no overall link between urinary stress-hormone production and attention, but the sign differed by sex: higher cortisone production went with more variable reaction times in boys (0.068, 0.005 to 0.132) and the opposite in girls (-0.072, -0.154 to 0.011).
Weak evidence: Subgroup contrast from one cohort; the female estimate includes zero; no mediation analysis was done, so it does not show hormones carry the chemical effect.
Source: Fabbri 2025 (HELIX), Environmental Research Link to this fact
Health effects / Other evidence
A 2025 narrative review says BPA, PBDEs, PCBs, PFAS, phthalates, perchlorate, pesticides and metals can disturb thyroid hormone production, transport or metabolism, but much of that rests on lab, animal or small hormone-level studies rather than proof of harm to children.
Weak evidence: No systematic methods; mechanisms for phthalates cited from a PBDE review and a 1983 in vitro paper.
Source: Eraslan Aydemir & Sahin 2025, Turk J Med Sci, Turkish Journal of Medical Sciences Link to this fact
Health effects / Measured in people or real products
Among all children in this ECHO analysis, the only single-compound result with a confidence interval excluding zero was a lower sleep-impairment score (about 4.8 T-score points) in the middle methylparaben exposure tertile versus the lowest, with no dose-response in the top tertile.
Weak evidence.
Caveats
One of about 24 single-compound comparisons with no multiple-comparison correction; the paper's own Results wrongly describe it as 'per unit increase'; direction is opposite to the authors' hypothesis, so it is not evidence of harm or benefit.
Source: Geiger 2025, Frontiers in Pediatrics Link to this fact
Health effects / Other evidence
Urinary trans,trans-muconic acid (a benzene and sorbic-acid metabolite) was associated with higher malondialdehyde, a lipid-peroxidation marker, in early pregnancy (B 0.116, 95% CI 0.050 to 0.181, p .001).
Weak evidence.
Caveats
Single cohort, single spot urine; t,t-MA is not product-specific (benzene from smoke and traffic, sorbic acid from food), so it is not a consumer-product finding; no clinical outcome.
Source: Kim 2025, Environmental Analysis Health and Toxicology Link to this fact
Health effects / Measured in people or real products
The most heavily weighted compounds in the boys' externalizing signal were butylparaben, benzophenone-3 and MCNP, none of which is a fragrance ingredient.
Weak evidence: Weights (21%, 15%, 11%) are from the text only; the index is a statistical construct and weights are not independent effects.
Source: Nakiwala 2025 (ECHO), Environmental Research Link to this fact
Health effects / Measured in people or real products
In a small Chinese study of 42 pregnant women with mild underactive thyroid and 84 matched controls, the fragrance-solvent marker MEP was the only phthalate metabolite higher in the women with the condition (odds ratio 1.42, 95% CI 1.03-1.95), but three other pregnancy cohorts found no MEP-thyroid hormone link or only a marginal one.
Weak evidence: One small cross-sectional study; MEP also marks personal-care product use and no study measured fragrance; other cohorts split (one lower T4, others null).
Source: Gorini 2025, Toxics Link to this fact
Health effects / Measured in people or real products
The 2025 review reports the DEHP metabolite sum was linked to subclinical hypothyroidism in early pregnancy (odds ratio 1.92, 95% CI 1.01-3.78; 42 cases, 84 controls) and to papillary thyroid cancer (odds ratio 3.51, 95% CI 1.64-7.49; 111 matched pairs), but all thyroid-disease studies were small, single-center and measured exposure once.
Weak evidence: Cross-sectional or case-control, single-spot urine, reverse causation possible; DEHP is a food, packaging and medical plastics phthalate, not a fragrance one.
Source: Gorini 2025, Toxics Link to this fact
Health effects / Animal study
Rat experiments reviewed in 2025 show DEHP lowering thyroid peroxidase, the iodide transporter and thyroid-hormone transport protein and raising thyroid-hormone breakdown enzymes, but no human study has shown the same pathway, and the one mediation test of maternal thyroid hormones for DEHP and child ADHD found no mediation.
Weak evidence: Animal and in-vitro mechanisms; zebrafish MEHP data conflicted with rat data; Engel 2018 mediation test used single thyroid measures at about 17 weeks.
Source: Gorini 2025, Toxics Link to this fact
Health effects / Other evidence
In a 10-person crossover pilot, volunteers rated a white musk fragrance oil sweeter, more 'gorgeous' and more liked than no scent (like 48.9 to 67.4 of 100; sweet 18.4 to 53.5; gorgeous 14.7 to 49.8), but 'relaxing' and 'calming' ratings did not differ between scent and no scent.
Weak evidence.
Caveats
Single 10-person pilot (physiology n = 9), no sample-size calculation, many uncorrected endpoints; control was no scent only; product co-developed by an author's company (Nextday), no conflict declared. 21 VAS items uncorrected for multiple comparisons; the 'unscented' session had no odor, so participants could tell the sessions apart.
Source: Zennifa 2025, Scientific Reports Industry-funded Link to this fact
Health effects / Other evidence
In the same pilot, mood scores for tension, fatigue and total mood disturbance fell during the white musk session, but fatigue and confusion fell in the no-scent session too and no mood scale differed between the two sessions.
Weak evidence.
Caveats
Single 10-person pilot (physiology n = 9), no sample-size calculation, many uncorrected endpoints; control was no scent only; product co-developed by an author's company (Nextday), no conflict declared. Within-condition significance versus non-significance is not a between-condition difference (between-condition p 0.435 to 0.959).
Source: Zennifa 2025, Scientific Reports Industry-funded Link to this fact
Health effects / Other evidence
A small white musk crossover pilot found a lower heart-rate-variability stress ratio (LF/HF 2.02 vs 2.30, Bonferroni p = 0.024) and small EEG differences at two scalp sites (p = 0.032 and 0.048) during 10 minutes of the scent, versus no scent.
Weak evidence.
Caveats
Single 10-person pilot (physiology n = 9), no sample-size calculation, many uncorrected endpoints; control was no scent only; product co-developed by an author's company (Nextday), no conflict declared. One time point, 8 EEG channels, baseline imbalance visible in the O2 gamma data before the scent was introduced; clinical meaning unknown.
Source: Zennifa 2025, Scientific Reports Industry-funded Link to this fact
Health effects / Other evidence
Prenatal urinary benzophenones (UV filters such as BP-3) were associated with lower, not higher, odds of childhood asthma at age 5-9 (sum OR 0.77, 0.66-0.90; boys 0.66, 0.53-0.83), opposite to the endocrine-disruption concern.
Weak evidence.
Caveats
Unexplained; prior cohorts found null for BP-3 and asthma; authors suggest confounding by sunscreen use or maternal atopy, untested; parental asthma history heavily missing; no multiple-comparison correction.
Source: Miller 2025 (ECHO), Environment International Link to this fact
Health effects / Other evidence
The same research group found that bisphenol S in pregnancy went with LOWER eczema odds in girls in 2024 (extant assays) and HIGHER odds in girls in 2025 (one harmonized lab), so the BPS-eczema signal depends on how it was measured.
Weak evidence: Authors attribute the change to measurement methodology; both analyses overlap in cohorts, so this is not independent replication.
Source: Miller 2025 (ECHO), Environment International Link to this fact
Health effects / Review of other studies
A 2026 narrative review proposes that chemically different endocrine disruptors (bisphenols, phthalates, PCBs, pesticides, flame retardants, some metals) converge on shared brain mechanisms for learning and memory: calcium signaling, glutamate and GABA signaling, oxidative stress, epigenetic changes and glial inflammation.
Weak evidence: Narrative review with no grading or pooling; nearly all supporting mechanism data are rodent and cell-culture studies, not human.
Source: Jocsak 2026, Frontiers in Behavioral Neuroscience Link to this fact
Health effects / Review of other studies
In this review the most developed thyroid-to-brain chain is for PCBs: in rats, developmental PCB exposure lowers thyroxine and causes hearing and motor deficits, and thyroxine replacement attenuates them; for phthalates it cites only an exposure-thyroid association in preschool children, with no mediation test.
Weak evidence: Rat data for PCBs (not a fragrance ingredient); the phthalate thyroid link is one association and Engel 2018's null mediation result is not cited.
Source: Jocsak 2026, Frontiers in Behavioral Neuroscience Link to this fact
Health effects / Review of other studies
In the 2026 review, bisphenol A at 10-100 nM rapidly raised calcium and ERK signaling in hippocampal neurons in the lab, while at moderate doses in live rats it lowered spine density, so the review itself says estrogen-like effects on synapses depend on dose, region and context.
Weak evidence: Lab and rodent results summarized secondhand; BPA is not a fragrance ingredient.
Source: Jocsak 2026, Frontiers in Behavioral Neuroscience Link to this fact
Health effects / Review of other studies
The review states that human cohorts provide substantial evidence of adverse neurobehavioral outcomes after phthalate exposure, but cites studies without effect sizes or grading; the graded EPA review in this wiki rates 29 of 30 phthalate-outcome pairs slight or indeterminate.
Weak evidence: Review confidence outruns its own human evidence; its phthalate mechanisms are animal and in vitro, and it names no DEP or other fragrance exposure.
Source: Jocsak 2026, Frontiers in Behavioral Neuroscience Link to this fact
Health effects / Review of other studies
A 2026 review of lab and animal studies reports that linalool, the main scent molecule in lavender, acts on brain receptors (GABA-A and NMDA) and protected against Alzheimer's-type changes in mice and rats; it did not run new experiments.
Weak evidence: Doses were 25-350 mg/kg in animals and 0.06-5 mM in vitro, and mechanisms are unconfirmed by the authors' own account.
Source: Piccialli 2026, Phytotherapy Research Link to this fact
Health effects / Other evidence
The same review notes that oxidized and acetylated linalool derivatives lose the ability to boost the brain's GABA-A calming receptors, so aged or oxidized linalool is not the same active material.
Weak evidence: Reported second-hand; the review calls an oxidation product 'linalool hydroxide' and does not identify the hydroperoxides.
Source: Piccialli 2026, Phytotherapy Research Link to this fact
Health effects / Other evidence
In mice, inhaled linalool calmed females after repeated exposures but had the opposite, anxiety-raising effect in males unless exposure was brief, so effects of breathing it are not uniform.
Weak evidence: Mouse study cited second-hand; sex- and duration-dependence not replicated.
Source: Piccialli 2026, Phytotherapy Research Link to this fact
Health effects / Other evidence
In a rat pituitary cell test, chlorinated and brominated versions of butyl and pentyl parabens blocked thyroid-receptor signaling at low micromolar concentrations (IC50 about 10-23 uM for the dihalogenated forms, versus 54 uM for pentylparaben itself), and bromine was stronger than chlorine.
Weak evidence.
Caveats
Single-lab cell reporter assay with lab-synthesized derivatives; cytotoxicity and lipophilicity may contribute; the natural hormone and reference antagonist act at nanomolar levels; no human data.
Health effects / Other evidence
Chlorinated and brominated bisphenol F, unlike the parent compound, switched on thyroid-receptor signaling in the same cell test (about 2.3-2.7-fold at 25 uM versus about 2-fold for BPF), while fully halogenated forms did not.
Weak evidence: One cell line, EC50 not determinable, cytotoxicity limited the top dose; the paper cannot say what this means for people.
Health effects / Other evidence
In boys only, the pregnancy chemical mixture was associated with higher internalizing (anxiety/depression-type) scores, 3.59 points per quartile (95% CI 0.42 to 6.75), but the sex interaction was not significant (p about 0.69) and nothing was corrected for multiple comparisons.
Weak evidence: Subgroup estimate; weights driven by PCB180, DEHP metabolites and methyl paraben; not a fragrance finding.
Source: Puvvula 2026, Environ Int, Environment International Link to this fact
Health effects / Other evidence
In this mixture analysis the phthalate metabolite MEP, a marker of diethyl phthalate exposure, was a minor and sex-inconsistent contributor (positive weight 0.08 in boys, negative 0.17 in girls for working memory), so the study gives no support to a perfume-specific effect on child behavior.
Weak evidence: MEP also marks fragranced-product use; weights are unstable and the study did not test fragrance directly.
Source: Puvvula 2026, Environ Int, Environment International Link to this fact
Health effects / Other evidence
Among those teens, product use in the previous day (including cosmetics, creams and perfume or deodorant) did not statistically explain the higher phthalate and paraben levels in the ADHD group, although the ADHD group used more cosmetics and washes.
Weak evidence: 24-hour self-reported use; small n; supplementary mediation tables not available; does not show that products are not a source of exposure in general.
Source: Terzioğlu 2026, Psychiatry Investigation Link to this fact
Information, not medical advice. See also: myths we won’t tell you.