Every fact on this site, with its evidence grade, its caveats, and a link to the source.
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 / Review of other studies
A 2025 systematic review of 18 human studies found bisphenol A, phthalates, and the UV filter octocrylene positively associated with polycystic ovary syndrome (PCOS) in most (though not all) of the studies that tested them.
Moderate evidence.
Caveats
Observational case-control/cross-sectional studies with biomonitoring, not causal evidence; BPA alone had 8-10 positive studies against 2-4 null studies depending on how analogs are counted; most individual chemicals rest on only 1-5 studies.
Health effects / Measured in people or real products
A 2025 global disease-burden model estimates that DEHP exposure from plastics was linked to about 356,000 cardiovascular deaths worldwide in 2018 (about 13.5% of CV deaths among 55-64-year-olds), and about 10.5 million years of life lost.
Moderate evidence.
Caveats
Built on a single hazard ratio (1.10, 95% CI 1.03-1.19) from one earlier US NHANES-linked mortality study by the same senior author, extrapolated to 200 countries; the authors explicitly state the model 'cannot make claims about causality'.
Health effects / Measured in people or real products
The estimated DEHP-attributable cardiovascular mortality burden falls hardest on the Middle East, South Asia, East Asia, and the Pacific, which together account for about 73% of all estimated DEHP-attributable deaths, while the USA and Europe show the lowest attributable share (about 10% and 8%).
Moderate evidence.
Caveats
Regional exposure estimates for Africa, Asia-EPA, Asia-MESA, Latin America, and Australia came from a meta-analysis/imputation model rather than direct national biomonitoring, so their precision is lower than the USA/Canada/Europe estimates, which used real biomonitoring surveys.
Health effects / Measured in people or real products
Countries with large aging populations and expanding plastics industries — India, China, and Indonesia — bear the largest absolute numbers of DEHP-attributable cardiovascular deaths in 2018 (about 104,000, 61,000 and 20,000 respectively).
Moderate evidence: Modeled estimate built on country-level population and mortality-rate data, not a direct count of deaths caused by DEHP.
Health effects / Other evidence
In a 2004 FDA laboratory study, two of three chemicals that penetrated into skin and stayed there after 24 hours (a self-tanning ingredient and a hair-dye ingredient) were shown, using an extended 72-hour test, not to actually enter the bloodstream in any meaningful additional amount.
Moderate evidence: Single FDA lab, in vitro diffusion-cell system (not in vivo humans); findings are chemical-specific, not a general rule about skin-retained chemicals.
Source: Yourick 2004, Toxicology and Applied Pharmacology Link to this fact
Health effects / Lab study (cells or chemistry)
In FDA's own lab tests, only about 0.1% of the diethanolamine (DEA) applied in real shampoo and hair-dye products reached the receptor fluid (the in vitro proxy for systemic absorption) after a normal 5-30 minute use; 96-98% of the DEA that penetrated the skin stayed lodged there instead.
Moderate evidence: Single-lab, in vitro, excised human skin diffusion-cell model; not a human in vivo measurement; n=2 products per class.
Source: Kraeling 2004, Food and Chemical Toxicology Link to this fact
Health effects / Lab study (cells or chemistry)
With body lotion left on skin for 24 hours, up to about 15% of applied DEA penetrated the skin, and 3 consecutive days of lotion use built DEA up in skin to nearly 30% of the applied dose, but systemic absorption into the receptor fluid still stayed under 1% each day.
Moderate evidence.
Caveats
Single-lab, in vitro; the authors argue this reservoir does not go on to reach systemic circulation, but that specific claim is itself unreplicated (see C348, where the same lab's own musk xylol data shows the opposite behavior).
Source: Kraeling 2004, Food and Chemical Toxicology Link to this fact
Health effects / Government agency
Air fresheners accounted for only 0.58% of all 2.4 million chemical exposures reported to US poison control centers in 2005, and 95% of the resulting injuries were minor; deodorizers (the category containing air fresheners) ranked 20th of 23 product categories for pediatric poison-control exposures.
Moderate evidence.
Caveats
Poison-control call data undercounts chronic, low-dose, or misattributed exposures and cannot establish long-term health risk; EPA itself notes it has no way to verify whether asthma or other symptoms are correctly attributed to air fresheners by callers.
Source: EPA 2007, citing AAPCC 2005 Link to this fact
Health effects / Other evidence
In a 2011 in vitro study on human cadaver abdominal skin, undiluted propylene glycol crossed the skin at a measured steady-state rate of about 97.6 micrograms per square centimeter per hour, and dipropylene glycol crossed about 2.5 times more slowly, at about 39.3 micrograms per square centimeter per hour.
Moderate evidence.
Caveats
Industry-funded (Cefic Propylene Oxide and Glycols Sector Group; 3 of 5 authors employed by propylene-glycol manufacturers Dow, LyondellBasell, Shell); undiluted 'infinite dose' design meant to represent worst-case continuous exposure, not typical product use; cadaver skin, single lab, unreplicated.
Source: Fasano 2011, Toxicology in Vitro Industry-funded Link to this fact
Health effects / Government agency
IARC classifies benzophenone (CAS 119-61-9) as Group 2B, possibly carcinogenic to humans, based on sufficient evidence in experimental animals: dose-related liver tumors (hepatocellular adenoma, and rare hepatoblastoma) in male mice, histiocytic sarcoma in female mice, renal tubule adenoma in male rats, and mononuclear-cell leukemia in rats, from two 2-year NTP feed studies; no human cancer data exist on benzophenone.
Moderate evidence.
Caveats
Classification rests on animal bioassays only; IARC itself states no human data were available, and calls the mechanistic evidence for tumor induction 'weak' even while it cannot exclude human relevance.
Source: IARC 2013 Link to this fact
Health effects / Government agency
Benzophenone's oxidative metabolite 4-hydroxybenzophenone, which is estrogenic in vitro (MCF-7 cells, ERalpha binding) and in vivo (rat uterotrophic assay), can also be formed by ultraviolet or sunlight irradiation of benzophenone itself, not only by enzymatic metabolism; the Working Group noted this matters because benzophenone has been used as a UV filter in cosmetics.
Moderate evidence.
Caveats
Photochemical formation and estrogenic activity of the photoproduct are drawn from cited primary studies (Hayashi et al. 2006, Nakagawa & Tayama 2001/2002) summarized in the monograph, not independently re-verified here; this describes the parent compound's own photoproduct, not a sunscreen-derivative UV filter's photoproduct.
Source: IARC 2013 Link to this fact
Health effects / Other evidence
Across 11-13 published studies from 1943-2015 (human volunteers, animals, and excised skin), mineral oil and petrolatum ingredients in cosmetics consistently stayed in the outermost skin layer (the stratum corneum), with at most a few percent reaching deeper skin layers and none detected in blood in any study.
Moderate evidence.
Caveats
This is a single industry-funded narrative review re-analyzing pre-existing, methodologically heterogeneous studies (no PRISMA protocol, no independent replication of the underlying data); one study in the set (Stamatas 2008) found petrolatum penetrating deeper into the stratum corneum (30 micrometers) than any other, still short of the living epidermis.
Source: Petry et al. 2017, Toxicology Letters Industry-funded Link to this fact
Health effects / Review of other studies
A 1950s-70s antiseptic called hexachlorophene showed how a chemical that binds tightly to the skin's outer dead layer can be both protective and dangerous: routine hexachlorophene baths cut newborn staph-infection admissions from about 6.5% to 0.6%, but when 18 French children were accidentally exposed to 6% hexachlorophene talc in 1972, four died and others developed permanent neurological damage from the same skin-binding property letting the chemical build up to toxic levels.
Moderate evidence.
Caveats
Historical case-series data as reported in a 2018 review citing 1950s-70s primary sources; the review does not re-verify the original studies' methods, and the toxic exposure was accidental overapplication (6% in talc powder to skin, some broken), not typical consumer use.
Source: Li, Cary & Maibach 2018, Archives of Dermatological Research Industry-funded Link to this fact
Health effects / Review of other studies
How long a chemical lingers in the outer layer of skin after being applied depends mainly on three things: its chemical structure (ionic/charged chemicals bind skin proteins more strongly), how hydrated the skin is, and how well the chemical dissolves in the product's carrier liquid — and, in one specific dataset (a set of six chemical warfare agent simulants), skin binding rose in step with how fat-soluble (lipophilic) the chemical was.
Moderate evidence.
Caveats
General mechanism synthesized from mixed antimicrobial, sunscreen, and warfare-agent-simulant studies (not fragrance ingredients); the lipophilicity relationship comes from a single dataset of six chemicals with no fragrance or cosmetic ingredients among them.
Source: Li, Cary & Maibach 2018, Archives of Dermatological Research Industry-funded Link to this fact
Health effects / Measured in people or real products
IARC found no convincing evidence that styrene exposure causes lung cancer or other solid tumors in humans; the strongest human cancer evidence for styrene concerns lymphohematopoietic malignancies, especially AML and T-cell lymphoma, in reinforced-plastics-industry cohorts.
Moderate evidence.
Caveats
Based on occupational cohort and case-control studies; effect estimates were often small with wide confidence intervals, and confounding by co-exposures (1,3-butadiene, benzene) could not be ruled out.
Source: IARC 2019 (Vol. 121) Link to this fact
Health effects / Animal study
Styrene-7,8-oxide, styrene's main metabolite, was separately classified by IARC in 2019 as Group 2A ("probably carcinogenic to humans") based on sufficient evidence in animals and strong mechanistic evidence that it is an electrophile which reacts directly with DNA and is genotoxic, even though human cancer evidence for styrene-7,8-oxide itself was rated "inadequate."
Moderate evidence.
Caveats
Human cancer evidence for styrene-7,8-oxide specifically (as opposed to styrene) was inadequate; the classification rests on animal bioassay and mechanistic/genotoxicity evidence, not human epidemiology.
Source: IARC 2019 (Vol. 121) Link to this fact
Health effects / Lab study (cells or chemistry)
Studies reviewed in the cosmetics industry's own 2021 benzophenone safety report found benzophenone-3 estrogenic in MCF-7 breast cancer cells (EC50 1.56-3.73 uM) and estrogenic/antiandrogenic in a yeast reporter assay (EC50 6.44 uM estrogen, 10.2 uM antiandrogen), and found benzophenone-1 and benzophenone-2 estrogen- and androgen-receptor-active in ovarian and prostate cancer cell lines; the industry panel's own Discussion and Conclusion do not address this data when reaching a 'safe' determination.
Moderate evidence.
Caveats
The yeast-reporter EC50s (6.44/10.2 uM) are a different assay system from the MDA-kb2 mammalian-cell IC50 (4.98 uM) reported in Schlumpf 2004; do not conflate the two. The MCF-7 EC50 of 3.73 uM corroborates Schlumpf 2004's own number via the same research group's 2001 precursor study, not independent replication.
Source: CIR 2021 Benzophenones Industry-funded Link to this fact
Health effects / Government agency
A study of 423 patients cited in the cosmetics industry's own 2021 benzophenone safety report found a positive association between women's urinary benzophenone-3 levels and the incidence of Hirschsprung's disease, with the cited authors noting that sunscreen use can expose a fetus to benzophenone-3 during the weeks 5-12 gestational window in which neural crest cell migration (implicated in the disease) occurs; the industry panel's Discussion and Conclusion do not address this study.
Moderate evidence: Single case-control study (China); association, not established causation; reported by CIR without further discussion or rebuttal in its Discussion/Conclusion sections.
Source: CIR 2021 Benzophenones Industry-funded Link to this fact
Health effects / Review of other studies
The Endocrine Society's April 2025 Position Statement defines an endocrine-disrupting chemical as "an exogenous chemical, or mixture of chemicals, that can interfere with any aspect of hormone action," and states that because non-monotonic dose responses are common, it cannot be assumed there are non-zero thresholds below which EDC exposure is safe.
Moderate evidence.
Caveats
A professional society position, not a government/regulatory finding; the Society's own definition omits the WHO/IPCS requirement that the interference actually cause adverse health effects.
Source: Endocrine Society 2025 Link to this fact
Health effects / Self-reported survey
More than half of Canadians with MCS surveyed (52%) said their current housing exposed them to mold or other symptom-triggering conditions, and 15% (18/119) fell into the paper's homeless category: moving frequently, living in a vehicle or tent, or staying with friends or family.
Moderate evidence.
Caveats
Self-reported, n=119, advocacy-recruited convenience sample; the abstract states 17% homeless but Table 3 gives 18/119 (15%), a discrepancy the paper does not resolve; 'unsafe housing' is a broad self-defined category, not independently verified.
Source: Diallo 2026 Link to this fact
Health effects / Other evidence
The Endocrine Society's definition of an endocrine-disrupting chemical covers interference with any aspect of hormone action, including hormone synthesis, not only interference at hormone receptors.
Moderate evidence.
Caveats
This is the definition as characterized in Sharma 2024's discussion of its own steroidogenesis-enzyme findings, not a direct quote from the Endocrine Society's own primary scientific statement (e.g., the 2009 or 2015 Endocrine Society Scientific Statements on EDCs, which are not themselves in our evidence base). It matches the widely used consensus phrasing but should ideally be corroborated against the primary statement if a stronger source is later ingested.
Health effects / Review of other studies
Formaldehyde, released by household cleaning products both directly and via reaction with indoor ozone, is described as a human carcinogen as well as a respiratory irritant.
Moderate evidence.
Caveats
This is a background descriptive statement in an emissions-chamber study, not new primary carcinogenicity data; it does not itself cite IARC or NTP classifications by name. Corroborated separately in our evidence base by US EPA's own indoor-air VOC guidance, which states plainly that benzene is 'a known human carcinogen' (src-2026-epa-voc-impact-indoor-air-quality) — so both halves of the video's statement (benzene and formaldehyde) are independently backed by evidence base source pages, though not yet as a combined graded claim.
Source: Rossignol 2013, Atmospheric Environment Link to this fact
Health effects / Government agency
The WHO/UNEP 'State of the Science of Endocrine Disrupting Chemicals — 2012' report's Concluding Remarks state verbatim that endocrine disruptors are 'a global threat that needs to be resolved.'
Moderate evidence.
Caveats
IMPORTANT PROVENANCE NOTE: this exact sentence ('This is a global threat that needs to be resolved,' p.26, Section 14 'Concluding remarks') was verified directly against the primary report PDF already held in the vault (raw/papers/processed/2013-who-unep-state-science-endocrine-disrupting-chemicals.pdf, via pdftotext), NOT against the curated evidence base/sources markdown summary page, which paraphrases this report extensively but does not itself quote this specific sentence. A curator should add it to the source page's Key Findings before treating this as fully 'in our evidence base.' It is an expert-consensus statement of concern, not a new empirical finding.
Source: WHO/UNEP 2012 Link to this fact
Health effects / Measured in people or real products
BPA was detected in 93% of urine samples tested in the US NHANES national biomonitoring survey.
Moderate evidence.
Caveats
Relayed by a 2021 review citing NHANES 2003-04, so the specific figure is now dated (over two decades old); the review does not report BPA levels in blood, breast milk, or spinal fluid — that part of the video's claim is not addressed by this or any other source currently in our evidence base, though it is independently plausible (BPA in human serum/blood and breast milk is documented elsewhere in the broader literature, just not yet captured here).
Source: Khalid & Abdollahi 2021, Iranian Journal of Pharmaceutical Research Link to this fact
Health effects / Measured in people or real products
People who wear the same fragrance or live in a continuously scented home adapt to it and stop noticing it, even though people around them still can: in a controlled home-exposure study, two weeks of daily exposure to a single odorant elevated detection thresholds specifically for that odorant, and most subjects did not return to normal sensitivity until about three weeks after the odorant was removed.
Moderate evidence.
Caveats
Single primary study, n=10 (Dalton and Wysocki 1996, unpublished, described in this 2000 review); no independent replication of the specific 3-week recovery figure, though the general adaptation mechanism is independently replicated across many other studies cited in the same review.
Source: Dalton 2000, Chemical Senses Industry-funded Link to this fact
Health effects / Measured in people or real products
Adaptation to a scent is specific to that scent, not a general loss of smell: in the same home-exposure study, sensitivity to a control odorant that was never part of the exposure stayed normal throughout, so someone who has gone nose-blind to their own perfume can still smell other things fine.
Moderate evidence.
Caveats
Same single small study (n=10) as the companion claim above; the specificity finding is corroborated qualitatively by an independent occupational study (acetone-exposed workers, no change to control odorants) already in our evidence base.
Source: Dalton 2000, Chemical Senses Industry-funded Link to this fact
Health effects / Measured in people or real products
People with multiple chemical sensitivity do not appear to have a more sensitive sense of smell at baseline; the evidence instead points to their brains habituating less than other people's brains do to a constant low-level chemical exposure.
Moderate evidence.
Caveats
Rests on one small ERP study (Andersson 2009, 21 vs 17 participants) that this review cites accurately for this specific finding; does not show whether reduced habituation makes exposure more hazardous, only that perception differs.
Source: Pellegrino 2017, Physiology & Behavior Industry-funded Link to this fact
Health effects / Measured in people or real products
In an FDA-run randomized trial of 24 healthy adults applying commercial sunscreens under maximal labeled use (2 mg/cm2 over 75% of the body, 4 times a day, for 4 days), oxybenzone reached a geometric mean peak blood concentration of 169-210 ng/mL, roughly 340 to 420 times the FDA's own 0.5 ng/mL threshold for requiring additional cancer and reproductive-toxicology studies, and was still above 20 ng/mL a full three days after the last application.
Moderate evidence.
Caveats
Single small feasibility trial (n=6/product), not designed to compare formulations, skin types or ages; maximal-use protocol, not typical daily use; conducted indoors with no real sun, heat or humidity exposure.
Source: Matta 2019, JAMA Link to this fact
Health effects / Measured in people or real products
All four tested sunscreen active ingredients — avobenzone, oxybenzone, octocrylene, and ecamsule — exceeded the FDA's 0.5 ng/mL systemic-absorption threshold on day 1 of a maximal-use trial, with blood levels rising further after repeated daily application (day 4 vs day 1), consistent with drug accumulation, and terminal half-lives ranging from about 1 day (oxybenzone) to over 3 days (octocrylene).
Moderate evidence.
Caveats
Exceeding the threshold is a regulatory trigger for more toxicology testing, not a finding that the ingredients are unsafe; the paper explicitly states the clinical significance of these levels is unknown.
Source: Matta 2019, JAMA Link to this fact
Health effects / Government agency
A 2024 CIR draft stated there is broad scientific consensus that airborne particles with an aerodynamic diameter greater than 10 micron essentially cannot penetrate to the lung's alveolar region, so particles below 10 micron (and especially below about 4 micron) are the ones capable of reaching the deep lung.
Moderate evidence: This is background physical/physiological consensus reported in a draft document, not a new finding of this document itself.
Source: CIR draft (Zhu) 2024 Industry-funded Link to this fact
Health effects / Measured in people or real products
In a controlled 2-week study, everyone who lived with a single pleasant fragrance running continuously in their bedroom developed a measurable loss of smell sensitivity specific to that exact scent, shown by elevated forced-choice detection thresholds in the lab.
Moderate evidence.
Caveats
Single lab, n=8, all female University of Pennsylvania students; sensitivity to an unrelated control odorant was unaffected, confirming the effect is odorant-specific rather than a general dulling of smell.
Source: Dalton & Wysocki 1996, Perception & Psychophysics Link to this fact
Health effects / Measured in people or real products
After a 2-week home fragrance exposure ended, only half of the people studied had fully regained their normal smell sensitivity to that fragrance two weeks later, and one person's sensitivity had not begun to recover at all.
Moderate evidence: N=8; the paper's own general discussion reports a preliminary, apparently never separately published follow-up suggesting recovery can take up to about 4 weeks for some people.
Source: Dalton & Wysocki 1996, Perception & Psychophysics Link to this fact
Health effects / Measured in people or real products
By the end of 2 weeks of constant home exposure, many people in this study reported they could no longer smell the fragrance in their home at all, even though they had not actually lost the ability to detect it at any concentration, and unexposed visitors could still smell it readily.
Moderate evidence: Self-report of 'no longer smelling it' is a questionnaire measure, not a threshold measurement; n=8.
Source: Dalton & Wysocki 1996, Perception & Psychophysics Link to this fact
Health effects / Other evidence
For the synthetic musk fragrance ingredient musk xylol, chemical still sitting in human skin 24 hours after a single application did not stay put: when an FDA lab extended testing to a full week, about three-quarters of the 24-hour skin reservoir had disappeared, while roughly 14-16% of the original applied dose had moved into the fluid representing the bloodstream side of the skin barrier.
Moderate evidence: Single FDA lab, in vitro excised human-skin model with only two human subjects in the extended-study cohort; not a living-human in vivo measurement.
Source: Hood 1996, Food and Chemical Toxicology Link to this fact
Health effects / Other evidence
Hairless guinea pig skin is substantially more permeable than human skin to musk xylol, especially for how much reaches circulation: in vitro, over half of the absorbed dose in guinea pig skin reached the receptor fluid within 24 hours, versus under 5% in human skin, even though both skin types absorbed a similar total amount into the skin itself.
Moderate evidence: Single FDA lab, in vitro excised-skin model; relevant to how much animal skin-absorption data can be extrapolated to humans for fragrance ingredients.
Source: Hood 1996, Food and Chemical Toxicology Link to this fact
Health effects / Other evidence
Intravenous CGRP alone, with no other trigger, caused migraine-grade headaches in people who get migraines.
Moderate evidence.
Caveats
Small single-group trial (n=9 analyzable of 12 enrolled), not independently replicated; 2 of 12 subjects were excluded for severe hypotension, meaning the dose used was close to the maximum tolerated.
Health effects / Other evidence
In the trial, all 9 analyzable patients given IV CGRP developed a delayed headache (median 5 hours later), and 3 of 9 met full diagnostic criteria for migraine without aura, versus 1 of 9 and 0 of 9 respectively on placebo.
Moderate evidence.
Caveats
Small sample; delayed-migraine difference (3/9 vs 0/9) was not itself statistically significant (P=0.21) even though the broader delayed-headache effect (9/9 vs 1/9) was (P=0.0004).
Health effects / Measured in people or real products
In a controlled human study, a single oral dose of triclosan equivalent to swallowing about 1.3 g of triclosan toothpaste was absorbed rapidly (peak blood levels within 1-3 hours) and cleared with a half-life of roughly 19-21 hours, with a median of 54% of the dose excreted in urine within 4 days.
Moderate evidence.
Caveats
Single lab, single small cohort (n=10); individual variability in urinary excretion was large (range 24-83% of the dose within 4 days), and one subject's data were confounded by accidental post-exposure triclosan-toothpaste use.
Source: Sandborgh-Englund et al. 2006, Journal of Toxicology and Environmental Health, Part A Link to this fact
Health effects / Measured in people or real products
Everyday background triclosan exposure (from sources other than labeled personal-care products, such as food, textiles, and footwear) is large and variable enough that, in this study, it obscured any clear difference between people who recently used triclosan-labeled toothpaste/soap and those who hadn't, once both groups had stopped using such products for 6 days.
Moderate evidence.
Caveats
Qualitative observation (no statistical test reported for this specific comparison); small sample (n=10, 5 vs 5); the authors themselves attribute the pattern to incomplete exposure monitoring rather than concluding product use has no effect.
Source: Sandborgh-Englund et al. 2006, Journal of Toxicology and Environmental Health, Part A Link to this fact
Health effects / Measured in people or real products
86% (12 of 14) of air fresheners NRDC tested in 2007 contained at least one of five phthalate esters (DBP, DEP, DIBP, DMP, DIHP), at levels from trace amounts up to 7,300 ppm of diethyl phthalate (DEP) in one product.
Moderate evidence: Single commercial lab, one sample per product, not peer reviewed; the report itself calls the sampling 'far from comprehensive' and not independently replicated since 2007.
Source: NRDC 2007, Clearing the Air Link to this fact
Health effects / Measured in people or real products
In a sleep-lab study of 36 healthy adults, brief puffs of scent delivered during sleep (lavender, vanillin, vetiver oil, or ammonium sulfide) did not increase nighttime arousals or awakenings compared to clean-air baseline, regardless of whether the smell was pleasant or unpleasant.
Moderate evidence.
Caveats
Single night per participant, n=6-15 per odorant, healthy young adults (23-36y) screened for normal sleep; does not address sleep apnea patients, the population the study was ultimately motivated by.
Health effects / Measured in people or real products
One unpleasant, mildly trigeminal odor (vetiver oil) significantly reduced the frequency of arousals during sleep compared to a no-odor baseline, rather than increasing it.
Moderate evidence: N=7 with complete data for this comparison; single study, not replicated.
Health effects / Measured in people or real products
Brief odor puffs during sleep caused a small, consistent breath-level change lasting about 30 seconds (less inhalation, more exhalation), present for pleasant and unpleasant, trigeminal and purely olfactory odorants alike, and consistent across light, deep, and REM sleep.
Moderate evidence.
Caveats
Effect measured via sensitive breath-by-breath pneumotachography on a single night; magnitude ~25-30% change in inhale/exhale ratio on the first breath, fading by the sixth breath; not shown to affect overall sleep quality, architecture, or next-day outcomes.
Health effects / Lab study (cells or chemistry)
In vitro, extracts of the submicrometer (0.1-1.0 um) fraction of temple incense-burning particles stimulated human coronary artery endothelial cells to produce about 2-fold more interleukin-6 and endothelin-1 than blank filters, and significantly more than the larger (1.0-10 um) or smaller (<0.1 um) particle fractions from the same temple, at 50 ug/mL for 4 hours.
Moderate evidence: Cell-culture evidence only; no human subject was exposed or measured. A commercial endothelial cell line, not cells from any worshipper, was used.
Source: Lin 2012, Science of The Total Environment Link to this fact
Health effects / Lab study (cells or chemistry)
The same submicrometer (0.1-1.0 um) temple particle fraction that raised inflammatory markers also significantly reduced nitric oxide synthesis in human coronary artery endothelial cells, while the larger and smaller particle fractions from the same temple instead increased nitric oxide relative to blank filters.
Moderate evidence.
Caveats
Cell-culture evidence only, single temple, single exposure duration (4h) and dose (50 ug/mL); the paper's own discussion notes this NO-reduction direction conflicts with an older in vitro study (Kang and Cheng 1997) that found PAH exposure increased endothelial NO via NOS activation.
Source: Lin 2012, Science of The Total Environment Link to this fact
Health effects / Lab study (cells or chemistry)
Within the submicrometer (PM1.0-0.1) temple particle fraction, naphthalene, acenaphthylene, acenaphthene, and anthracene content were each strongly correlated with the degree of nitric oxide reduction in exposed human coronary artery endothelial cells (Pearson r = -0.90 to -0.99, p<0.05); no PAH correlated significantly with IL-6 or endothelin-1.
Moderate evidence.
Caveats
Correlation within one dataset of 13 samples from one temple, not a dose-response or causal manipulation of individual PAHs; the paper's own text also names fluorene as a fifth significant correlate, but its Table 2 does not mark fluorene's r=-0.84 as statistically significant — an internal inconsistency in the primary.
Source: Lin 2012, Science of The Total Environment Link to this fact
Health effects / Other evidence
In a British birth cohort followed from birth to age 42, people who had ever worked as hairdressers, barbers, or beauticians had 88% higher odds of adult onset asthma than office workers (OR 1.88, 95% CI 1.24-2.85), adjusted for sex, smoking, childhood social class, and childhood hay fever.
Moderate evidence: Occupational, not fragrance-specific: exposure is the job title, not a measured product or chemical; single UK cohort, not yet independently replicated.
Health effects / Other evidence
Four cleaning occupations (unspecified cleaners, domestic helpers and cleaners, helpers/cleaners in offices and hotels, and hand-launderers) were associated with adult onset asthma in this cohort, with odds ratios from 1.58 to 2.26; the offices/hotels cleaning code (OR 1.82, 95% CI 1.34-2.48) and home-based personal care workers (OR 1.94, 95% CI 1.40-2.69) remained significant after correcting for testing 61 occupations.
Moderate evidence: Job-title-level exposure, not product-level; the lowest estimate (unspecified cleaners, OR 1.58) did not reach significance on its own (p=0.078).
Health effects / Other evidence
Using a job-exposure matrix built specifically for asthma research (ASJEM), lifetime exposure to low-molecular-weight reactive cleaning or disinfecting products was associated with 67% higher odds of adult onset asthma (OR 1.67, 95% CI 1.26-2.22), the single exposure category that remained significant after correcting for 18 tested high-risk exposures (q<0.001).
Moderate evidence.
Caveats
The ASJEM exposure category is an expert judgment covering an unspecified mix of cleaning/disinfecting chemicals, not a named ingredient; single-cohort, not yet independently replicated.
Health effects / Other evidence
About 16% of adult onset asthma in this British cohort (95% CI 3.8-27.1%) was estimated to be attributable to occupational chemical exposure of any kind, a population-attributable-fraction estimate comparable in size to the fraction attributable to ever having smoked (11.2%, 95% CI 2.6-19.1%).
Moderate evidence.
Caveats
This is the PAF for all occupational exposure combined (farming, printing, metalworking, textiles, cleaning, etc.), not a cleaning-specific or fragrance-specific estimate; the outcome is self-reported 'ever had asthma,' not spirometry-confirmed.
Health effects / Other evidence
People who went on to develop adult onset asthma spent less time working in the jobs linked to asthma than people who did not develop asthma (e.g., home-based personal care workers: 1.88 years vs 3.18 years, p=0.016; domestic helpers and cleaners: 1.08 years vs 3.14 years, p=0.022), consistent with a 'healthy worker effect' in which people leave jobs that cause or worsen their disease.
Moderate evidence: This pattern supports the cohort's occupational associations being real rather than reverse-caused, but does not rule out all residual confounding.
Health effects / Self-reported survey
In a Brazilian headache-clinic study of 200 diagnosed migraine patients, 70% (140/200) reported that an odor — most often perfume (75.7% of those affected) — had triggered a headache at least once in their life, with pain typically beginning within about 10 to 25 minutes of exposure.
Moderate evidence.
Caveats
Retrospective lifetime recall at a single referral clinic in Teresina, Brazil (2011); non-random convenience sample; no prospective diary, blinded challenge, or control for premonitory-phase symptoms that can be mistaken for triggers.
Health effects / Self-reported survey
None of 200 tension-type headache patients at the same clinic reported an odor-triggered headache, versus 70% of migraine patients — making odor-triggering 100% specific (but only 70% sensitive) for migraine over tension-type headache in this sample.
Moderate evidence: Single-clinic sample; a diagnostic-discrimination statistic, not evidence that odors cause migraine specifically versus being recalled more by migraineurs generally.
Health effects / Self-reported survey
In this clinic sample, migraine patients were overwhelmingly female (182/200, 91%) while tension-type headache patients were overwhelmingly male (163/200, 82%), consistent with migraine's well-established female predominance.
Moderate evidence.
Caveats
Single non-random clinic sample; this is a diagnosed-migraine clinical cohort, not a general-population fragrance-sensitivity survey, so it speaks to migraine's known sex skew rather than to whether fragrance reactivity itself is sex-neutral.
Health effects / Self-reported survey
In a retrospective cohort of 1,010 tertiary-care migraine patients, premonitory symptoms (onset 2 or more hours before the headache) occurred in 38.5% (389/1,010), the most frequent being a tense neck, phonophobia, and difficulty concentrating; osmophobia as a premonitory symptom occurred in roughly a fifth of patients and was significantly more common in migraine with aura than without (p<0.05).
Moderate evidence.
Caveats
Retrospective design subject to recall bias; tertiary headache center sample likely more severely affected than the general migraine population; abstract states 38.9%, Results/Discussion body text states 38.5% — an internal inconsistency in the primary.
Source: Schulte 2015, The Journal of Headache and Pain Link to this fact
Health effects / Self-reported survey
In a cross-sectional study of 113 diagnosed migraineurs in Brazil, 95.5% reported osmophobia (headache worsening on exposure to an odor during a crisis) — the highest osmophobia prevalence reported in this literature to date.
Moderate evidence: Single-center, cross-sectional, self-report only, unvalidated 65-item questionnaire, no healthy-control comparison group.
Source: Fornazieri 2016, Headache: The Journal of Head and Face Pain Link to this fact
Health effects / Self-reported survey
Among the same 113 migraineurs, 90.2% had migraine crises triggered by odors, and of those, perfume was by far the most common odor trigger (95.1%), followed by cleaning products (81.3%), cigarette smoke (71.5%) and motor vehicle exhaust (70.5%).
Moderate evidence: Self-report, unvalidated questionnaire, single clinical sample, no provocation testing of the named odors.
Source: Fornazieri 2016, Headache: The Journal of Head and Face Pain Link to this fact
Health effects / Self-reported survey
Perfume as a migraine trigger was significantly more common in migraineurs without auras than with auras (P = .02), while physical exercise as a trigger was more common in those with auras (P = .01); no other olfactory symptom (osmophobia, odor-triggered crisis, interictal hypersensitivity, hallucinations, phantosmia, cacosmia/euosmia) differed between the two groups (all P > .4).
Moderate evidence: Single-center cross-sectional sample; aura classification self-reported, not independently verified.
Source: Fornazieri 2016, Headache: The Journal of Head and Face Pain Link to this fact
Health effects / Self-reported survey
Across online-panel surveys described by the author as nationally representative of US, Australian, and UK adults, 4.3% report a medical diagnosis of autism, an autism spectrum disorder, or both, with no significant difference in prevalence among the three countries.
Moderate evidence: Self-reported yes/no doctor-diagnosis item, no clinical confirmation; cross-sectional online panel Online-panel samples (see C207 on opt-in panel effects).
Source: Steinemann 2018, Air Quality, Atmosphere & Health Link to this fact
Health effects / Self-reported survey
83.7% of self-reported autistic adults in a US/AU/UK survey reported one or more adverse health effects from fragranced products, versus 29.5% of non-autistic adults (prevalence odds ratio 12.05, 95% CI 7.66-18.95).
Moderate evidence.
Caveats
Self-report, no clinical confirmation of autism/ASD diagnosis or of health effects; cross-sectional; studies effects among people already diagnosed autistic, not causation of autism.
Source: Steinemann 2018, Air Quality, Atmosphere & Health Link to this fact
Information, not medical advice. See also: myths we won’t tell you.