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 / Self-reported survey
In a nationally representative survey of 1,136 US adults, 34.7% said fragranced products such as air fresheners, cleaners, and perfume gave them health problems, most often breathing difficulties, eye/nose irritation, or migraine.
Moderate evidence.
Caveats
Self-report, any symptom from any product; not a diagnosis; adults 18-65 in an opt-in online panel; same US dataset as the 2018 asthmatics, 2018 MCS, and 2019 pooled papers (not replications); 62.7% of the affected also report chemical sensitivity or MCS, but 17.8% of people with neither are still affected.
Source: Steinemann 2016, Air Quality, Atmosphere & Health Link to this fact
Health effects / Self-reported survey
In a national survey of 1,098 Australian adults, more than half of those with asthma or an asthma-like condition (55.6%) said fragranced products made them ill, against about a quarter (23.9%) of people without asthma.
Moderate evidence.
Caveats
Self-report; unadjusted; 'asthma-like condition' undefined (it is 47.9% of MCS respondents in the MCS re-cut); same dataset as the AU arm of Steinemann 2019 and Steinemann 2018 (MCS, AU).
Source: Steinemann 2018, Air Quality, Atmosphere & Health Link to this fact
Health effects / Self-reported survey
About 1 in 4 Australian adults with asthma (24.0%) said a fragranced product had triggered an asthma attack; air fresheners were the most common trigger (14.1% of all asthmatics).
Moderate evidence: Self-report; consistent with US and Danish samples (~34-43% airway symptoms).
Source: Steinemann 2018, Air Quality, Atmosphere & Health Link to this fact
Health effects / Review of other studies
Cleaning products were the suspected exposure in 236 of 1,915 (12%) confirmed work-related asthma cases across California, Massachusetts, Michigan, and New Jersey (1993-1997); in California, janitors and cleaners had roughly double the confirmed-case rate of all other occupations combined (4.3 vs 2.3 per 100,000 workers).
Moderate evidence.
Caveats
Now traced to its primary source (Rosenman 2003), not a secondary citation. But the underlying surveillance system's own authors say it misses 53-87% of true cases (Michigan capture-recapture) and that up to 37% of confirmed cases in such systems may not be truly work-related (Malo et al., cited by Rosenman); 'cleaning products' is a broad, mostly unspecified chemical category (36% of named agents never identified further than 'cleaning compound'), not fragrance-specific (limonene only 2/300 exposure counts).
Source: Nazaroff & Weschler 2004, Atmospheric Environment Link to this fact
Health effects / Self-reported survey
In a US national survey, about 1 in 5 adults (20.4%) said air fresheners gave them health problems such as breathing difficulty, migraine, or asthma attacks.
Moderate evidence: Restated from Steinemann 2016 survey; cite the primary; self-report.
Source: Steinemann 2017, Building and Environment Link to this fact
Health effects / Self-reported survey
In a nationally representative survey of 1,100 UK adults, 27.8% said fragranced products such as air fresheners, cleaners, or perfume gave them health problems, including migraines (8.4%) and asthma attacks (6.8%).
Moderate evidence: Self-report, any symptom; same data as Steinemann 2019 UK arm.
Source: Steinemann 2018, Air Quality, Atmosphere & Health Link to this fact
Health effects / Review of other studies
In a small double-blind, randomized lab test (Ternesten-Hasséus 2002, the primary study behind this review's capsaicin-challenge claim), 12 people with multiple chemical sensitivity coughed significantly more than 12 matched healthy controls at every capsaicin dose tested (P<0.05 to P<0.005), a sign that their airway irritant nerves are oversensitive.
Moderate evidence.
Caveats
Now confirmed by reading the primary source directly rather than only via this review's secondhand paraphrase: n=12 patients vs 12 matched controls, single Gothenburg clinic, MCS diagnosed only by Cullen's clinical criteria (no independent biomarker), and not independently replicated outside this same research group (which has authored nearly every paper in this specific evidence line).
Source: Molot 2023, Neuroscience and Biobehavioral Reviews Link to this fact
Health effects / Measured in people or real products
Estimated cancer risks were below one in a million for most feminine hygiene product types; washes (3.3 in a million) and sprays/powders (2.1 in a million) were just above that level under worst-case assumptions.
Moderate evidence: Modeled upper-bound scenario.
Source: Lin 2020, Environment International Link to this fact
Health effects / Lab study (cells or chemistry)
In lab studies, common flavor and fragrance chemicals act directly on TRPA1, the nerve receptor that senses irritants such as mustard oil: cinnamon's cinnamaldehyde and thyme's thymol switch it on, while camphor and eucalyptus oil block it.
Moderate evidence: Cell and animal pharmacology; the review's framing is therapeutic, not hazard.
Source: Mihara & Shibamoto 2015, Allergy, Asthma & Clinical Immunology Link to this fact
Health effects / Measured in people or real products
In a double-blind lab test where people could not smell it, perfume vapor piped to one eye caused more eye irritation than clean air in 10 of 42 people, including 2 healthy volunteers: perfume can irritate the eyes without being smelled.
Moderate evidence.
Caveats
N=42; mostly self-rated eye symptoms; patient vs control difference P=0.07; only 2 objective tearing events. Consistent with the earlier single-blind Millqvist 1999 (eye irritation 1.1 vs 0.2 on 0-3, n=11), a different group, perfume and patient type. Das-Munshi, Rubin & Wessely 2006's systematic review classifies both single-blind Millqvist perfume studies (1996, 1999) as insufficiently blinded (nose clips only, no true olfactory mask), so they should not be counted as strong independent corroboration; Elberling's own double-blind result stands on its own primary data, which is why this stays at moderate rather than strong.
Health effects / Measured in people or real products
The eye irritation from perfume vapor was not an allergy: people patch-test allergic to the very same perfume were no more likely to react than those who were not.
Moderate evidence.
Caveats
21 patients, 10 patch-test positive; small sample. A related 2007 study from the same group, clinic, and challenge perfume found the same pattern in a different assay: basophil histamine release to perfume was not associated with contact-allergy patch-test status either (Table 4, all P > 0.05, n=17 patients). This is modest corroboration from the same research program (same lab/clinic/perfume), not independent replication, so the grade is unchanged.
Health effects / Measured in people or real products
In a Toronto clinic for work-related breathing problems, about 1 in 10 referred workers had an irritable larynx (throat) syndrome rather than asthma; perfumes and cleaning agents, bleaches and air fresheners were among the named triggers.
Moderate evidence.
Caveats
Single referral clinic, retrospective, n=30; perfume named by 5/30 and cleaning agents/fresheners by 4/30; 10% of referrals, not of workers. Morrison 1999, the founding case series behind the diagnostic criteria Hoy applied, independently found perfume was a named trigger in only 26% (10/39) of its own ILS cohort, behind reflux (72%) and other airborne irritants (51%) — the syndrome is not fragrance-specific at its point of origin either.
Health effects / Measured in people or real products
About 80% of workers with scent- and irritant-triggered irritable larynx syndrome were women, versus 24–52% of workers with occupational asthma at the same clinic.
Moderate evidence: N=30, one clinic; referral and care-seeking may differ by sex.
Health effects / Measured in people or real products
Scent-triggered throat and breathing reactions can mimic asthma but do not respond to asthma inhalers; almost two-thirds of these patients had been prescribed asthma medication anyway.
Moderate evidence.
Caveats
Clinic series; non-response to asthma drugs is the authors' clinical statement, not a trial. Independently corroborated by a different clinic and decade: Millqvist & Lowhagen 1996 (Sweden, 1996) found 7/9 (78%) of its scent-triggered, methacholine-negative patients had been prescribed beta2-agonist and inhaled corticosteroids with 'no demonstrated — but only a subjectively doubtful — effect,' and directly demonstrated the 'mimics asthma' half objectively: perfume provoked significantly more symptoms than saline (P<0.01) with zero change in PEF, O2 saturation, heart rate, respiratory rate, or expired CO2. But like Hoy 2010, the 'does not respond to inhalers' half remains a subjective clinical judgment embedded in a case series in both papers, not a controlled medication trial, so per the grading rule (a claim with any unverified part stays below strong) the grade stays at moderate rather than moving to strong.
Health effects / Measured in people or real products
An older, unblinded human study (Stewart 1970) reported eye, nose, and throat irritation from PGME solvent vapor at 250 ppm in 23 subjects across five experimental runs, and odor tolerance developing within 25 minutes in 6 subjects at 100 ppm; a later double-blind study with objective eye measurements (Emmen 2003) found no objective eye effects at up to 150 ppm and only very slight subjective eye discomfort at 150 ppm, but did not test 250 ppm.
Moderate evidence.
Caveats
Stewart 1970 (read directly) used 23 subjects across five experimental runs at ~230-250 ppm for 1-7 hours (not 'two subjects' as this claim previously stated; the 2-subject figure applies only to a separate rising-concentration ramp experiment used for the 50-75 ppm odor-objectionable and 1000-2050 ppm severity findings) and 6 subjects at 100 ppm, with eye/nose/throat irritation developing in 20 of 23 subjects within 45-60 minutes at 250 ppm. The design remains genuinely unblinded (no placebo condition or masking anywhere in the paper) and reports no objective ocular measurements (redness, corneal thickness, blink counts) — only visual acuity, coordination, neurological, and brake reaction-time tests, all normal throughout at and below 250 ppm. One author (Torkelson) was a Dow Chemical employee and the test material was Dow-supplied ('Dowtherm 209'), a more direct link than Emmen 2003's Dow-supplied sample alone; since the finding (irritation at 250 ppm) runs counter to the sponsor's commercial interest, the COI does not obviously bias this particular result toward understatement. Emmen 2003 did not test 250 ppm, so it does not directly refute the finding, only the confidence one can place in an unblinded, non-objective design; per our grading rule, a single unblinded single-lab study without objective measures cannot reach 'strong' even with n=23.
Source: CIR 2008 (citing Stewart 1970), International Journal of Toxicology Industry-funded Link to this fact
Health effects / Animal study
Citronellol, a rose-scented fragrance ingredient present in about half of household products, crosses the blood-brain barrier: brain-to-blood ratios were 0.5-0.8 in fish, up to 0.6 in mice, and 0.77 in a laboratory barrier built from human brain cells.
Moderate evidence: BBB crossing is measured and dose-independent as a property, but the doses used were far above consumer exposure; routes were waterborne and oral, not inhalation or skin.
Source: Kim 2024, Journal of Hazardous Materials Link to this fact
Health effects / Animal study
At high doses in zebrafish and mice, citronellol pushed tryptophan metabolism toward 3-hydroxykynurenine, a neurotoxic metabolite, and produced brain inflammation, roughly double the oxidative stress, a five-fold rise in dying brain cells, a leaky blood-brain barrier, and anxiety-like behavior.
Moderate evidence: Zebrafish 2-8 mg/L in water; mice 345-3450 mg/kg by gavage (top dose = an LD50); 10^5-10^6 times estimated human exposure; male mice only.
Source: Kim 2024, Journal of Hazardous Materials Link to this fact
Health effects / Animal study
Citronellol's effect on the brain reverses with dose: protective in a rat Parkinson's-disease model at 25 mg/kg oral (given as a month-long pretreatment before each day's toxin dose, not as treatment of existing damage), damaging at 345 mg/kg and above in a different paper, species and route.
Moderate evidence.
Caveats
Both dose points are single-dose, single-model animal work. The protective study (Jayaraj 2022) has now been read directly rather than via Kim 2024's citation and confirms the 25 mg/kg number exactly, but that dose is itself 10^4-10^5x above estimated human citronellol exposure (1,600-8,000x after allometric human-equivalent-dose scaling) — so neither the protective nor the toxic limb has been shown near a human-relevant dose, and the crossover point between 25 and 345 mg/kg remains unmeasured.
Source: Kim 2024, Journal of Hazardous Materials Link to this fact
Health effects / Review of other studies
Salicylate fragrance chemicals pass through skin and are converted in the body to salicylic acid, the compound aspirin becomes: for methyl salicylate (wintergreen), human studies found roughly 12-30% of an applied dose absorbed.
Moderate evidence.
Caveats
Human absorption data are almost all for methyl salicylate, often under occlusion at pharmaceutical concentrations; the review quotes the range three different ways (2-43%, 12-30.7%, 11.8-30.7%).
Source: RIFM Expert Panel / Belsito 2007, Food and Chemical Toxicology Industry-funded Link to this fact
Health effects / Animal study
Mice breathing the vapor of five ordinary commercial colognes for one hour showed airway irritation and a 30-35% drop in mid-expiratory airflow, despite never having been exposed to any allergen or asthma trigger.
Moderate evidence.
Caveats
Mouse study; vapor load never measured and far above household exposure; classical RD50 endpoint was not significant; unreplicated outside the authors' own laboratory. Their 1997 air-freshener study (same two authors, same commercial lab, same computerized ASTM E-981 method, different product) found a matching ~35% peak airflow fall, which is within-lab consistency, not independent replication, and that paper's own 'real-use exposure' claim does not survive comparison with measured indoor VOC data. An independent EPA/ManTech attempt to replicate this lab's method (Tepper et al. 1995, carpets, full text now in our evidence base) found no significant carpet-related toxicity by individual-animal statistics and attributed the lab's positive-looking signals to restrained-exposure procedural stress present even in air-only controls, but it never measured the airflow-limitation endpoint this claim rests on, so it bears on the lab's method credibility generally without directly retesting this specific finding.
Source: Anderson & Anderson 1998, Archives of Environmental Health Link to this fact
Health effects / Measured in people or real products
Over 20 years (1993–2012), California's work-related asthma surveillance confirmed 270 cases linked to perfume or air fresheners at work, 3.8% of all confirmed cases; perfume was the 9th most common workplace asthma exposure.
Moderate evidence: Surveillance count of physician-diagnosed, reported cases; misses >=2/3 of cases; no objective testing required; excludes fragranced cleaners.
Health effects / Measured in people or real products
94% of workers with fragrance-related work asthma in California were women, mostly in offices, health care, and schools, and most cases came from perfume worn by others.
Moderate evidence: Authors attribute the sex skew to occupation and exposure, not susceptibility; care-seeking bias possible.
Health effects / Self-reported survey
Many people with chemical sensitivities delay or avoid medical care because clinics and hospitals are heavily scented; healthcare was named as one of the least accessible settings (31 of 60).
Moderate evidence: Qualitative, self-selected Canadian sample; frequency of care avoidance not quantified.
Source: Roy 2026 Link to this fact
Health effects / Self-reported survey
In a nationally representative German survey, half of the people who get sick from fragranced products (51.1%) had been kept from going somewhere because of fragrance, and 43.8% had avoided public toilets with air fresheners or scented products.
Moderate evidence: Self-report; Germany 2019, n=219 fragrance-sensitive of 1,102; same data as Steinemann & Klaschka 2019.
Source: Klaschka 2020, Environmental Sciences Europe Link to this fact
Health effects / Review of other studies
Fragrance can cause skin allergy without touching your own products: from a partner's toiletries (consort dermatitis), from soap residue carried on the hands to the eyelids, and through the air, as in a small case series of a ride-share driver and four passengers whose airborne facial/eyelid dermatitis (all patch-test positive for fragrance mix I and/or balsam of Peru) was linked to car fragrance diffusers, with the driver's rash clearing when his diffuser was removed.
Moderate evidence.
Caveats
Mechanisms from case reports and a systematic review of consort ACD cited by the review; the airborne example is now confirmed against its primary (Perper 2017, now ingested): actually 5 total patients (1 driver + 4 passengers), not just the driver, published in a Contact Dermatitis section the journal states is exempt from mandatory peer review, with no chemical/air testing, no controls, and diffuser-removal resolution documented only for the driver. Kept at moderate, not raised to strong, per the single-small-uncontrolled-study rule; no prevalence data.
Health effects / Lab study (cells or chemistry)
In lab tests on human cells, both lavender oil and tea tree oil switched on estrogen-receptor signaling (about half the maximum effect of the body's own estradiol) and blocked testosterone-type (androgen-receptor) signaling by 41–52%.
Moderate evidence.
Caveats
Cell assays with whole commercial oils at ~0.005–0.025% vol/vol; no data that such concentrations reach a child's body; linalool and linalyl acetate alone were inactive in later guideline assays (industry).
Source: Henley 2007, New England Journal of Medicine Link to this fact
Health effects / Lab study (cells or chemistry)
Lavender oil and tea tree oil activated the human estrogen receptor in cell assays to about 17-18% of the response to estradiol, and blocked testosterone signaling more completely than flutamide, a prescription anti-androgen drug.
Moderate evidence.
Caveats
Human cell lines, oils at 0.005-0.025% v/v; percentages are relative to 1 nM estradiol or 1 nM testosterone in transfected HepG2 and MDA-kb2 cells; Henley 2007 reported ~50% for the same oils in a different cell system.
Source: Ramsey 2019, The Journal of Clinical Endocrinology & Metabolism Link to this fact
Health effects / Review of other studies
If a young child has unexplained early breast development, it is reasonable to ask about lavender and tea tree products, including room diffusers, and try stopping them – which is what the study's authors recommend to doctors.
Moderate evidence: This is the authors' clinical recommendation, not a proven treatment; they explicitly state they are not recommending that anyone avoid these products generally.
Source: Ramsey 2019, The Journal of Clinical Endocrinology & Metabolism Link to this fact
Health effects / Self-reported survey
Of all personal care product types, perfumes and colognes had the highest hazard ratings in the Environmental Working Group's Skin Deep database, and adults aged 18–39 used the highest-rated ones (average 7.2 of 10, vs 5.9 for those 60 and over).
Moderate evidence: EWG hazard ratings from listed ingredients, not measured toxicity or exposure; 39% of perfumes used could not be rated.
Source: Barrett 2025, Journal of Exposure Science & Environmental Epidemiology Link to this fact
Health effects / Self-reported survey
Black women in a US university sample were twice as likely as White women to have recently used a hair product rated high-hazard by EWG (relative risk 1.99, 95% CI 1.37–2.89).
Moderate evidence.
Caveats
Average hair-product scores did not differ by race; hazard rating, not exposure; 53 Black women. Helm 2018 (GC/MS of 18 hair products used by/marketed to Black women) found somewhat higher DEP and newly-detected nonylphenol ethoxylates vs. a general-market baseline, which is consistent with elevated exposure but is measured chemistry with no White-women comparison group in that study, not a replication of this RR — see C100 extension. Chan 2023 (retail-shelf audit, 14,019 hair products, 50 Boston stores) found stores in lower-income Roxbury and Mission Hill ~2.3x as likely to stock an EWG high-hazard hair product as higher-income Beacon Hill — a structurally distinct, supply-side corroboration (store stocking, not individual use; only 2 of 6 neighborhood comparisons significant; race and SES confounded in that study's own design) — see second C100 extension.
Source: Barrett 2025, Journal of Exposure Science & Environmental Epidemiology Link to this fact
Health effects / Lab study (cells or chemistry)
Essential-oil terpenes can interfere with the enzymes that BUILD hormones, not just the receptors hormones act on — a route that the standard estrogen- and androgen-receptor safety screens are not designed to detect.
Moderate evidence: Mechanistic point only. The measured effects were weak (13–33%) and all at a single 10 µM with no dose–response.
Health effects / Lab study (cells or chemistry)
The antibacterial triclosan, once in about 45% of US soaps, reacts to UVB at an intensity matched to real midday sunlight by generating free radicals that damaged human skin cells (HaCaT keratinocytes) in the laboratory, dose-dependently and only when triclosan and radiation were combined.
Moderate evidence.
Caveats
Single lab, single immortalized human keratinocyte cell line (HaCaT), not primary human skin or an in vivo model; triclosan concentration (5 ug/ml) not validated against measured human skin-surface levels; not replicated.
Source: Khalid & Abdollahi 2021, Iranian Journal of Pharmaceutical Research Link to this fact
Health effects / Review of other studies
A 2024 safety review by an industry toxicologist states there are no data on whether PTFE passes through the skin, how the body handles it, or whether it affects hormones, and that it does not biodegrade.
Moderate evidence.
Caveats
CIR's own 2018 primary safety report (now ingested, src-2018-safety-assessment-polyfluorinated-polymers-cosmetics) confirms in its own words that no dermal penetration or ADME data exist for PTFE, directly corroborating that part of this claim from the industry panel itself, not just the review's paraphrase. However, the CIR primary never once mentions endocrine disruption or aquatic toxicity anywhere in its 58 references — those two items in Jomaa's abstract are not sourced to CIR and are best read as the review's own added framing (general PFAS-family concerns), not a CIR-specific finding for PTFE. Absence of data is not evidence of harm; tested endpoints (irritation, sensitization, genotoxicity) were negative.
Source: Jomaa 2024, Journal of the Netherlands Society of Toxicology (JNST) Industry-funded Link to this fact
Health effects / Lab study (cells or chemistry)
In a 2025 lab screen of 42 everyday perfumes, almost all contained compounds that damaged bacterial DNA, killed cells, blocked the hormone-making enzyme aromatase, or acted on human estrogen or androgen receptors in test organisms, regardless of price or whether the perfume was marketed to men or women.
Moderate evidence.
Caveats
In vitro (bacteria, yeast, isolated enzyme on a chromatography plate); 42 unnamed perfumes from one German campus; no compound identified; no human exposure measured. A 2026 same-lab follow-up (Schmidtmann 2026) confirms only the DNA-damage/cytotoxicity part (bacterial Ames, not the same assay) in a smaller, different, unnamed 8-perfume set, still without metabolic (S9) activation for perfumes and without retesting the hormone-active endpoints — same-lab corroboration, not independent replication.
Source: Morlock & Heil 2025, Journal of Chromatography A Link to this fact
Health effects / Animal study
Two fragrance musks, AETT and musk ambrette, damaged the nervous system of rats in industry-sponsored studies and were withdrawn; musk ambrette was also barred from skin products because it caused allergic reactions in sunlight.
Moderate evidence.
Caveats
Both primaries now read directly: Spencer 1979 (AETT) is a small, single-lab, single-route dietary study (N=16); Spencer 1984 (musk ambrette) is a large, dual-route, dose-ranging study (N=220) that alone would meet the bar for 'strong'. Because the claim bundles both compounds and the AETT sub-claim rests on one small single-lab study, the combined claim stays at moderate per the grading rule (one small study keeps a claim at moderate; a claim with any weaker bundled part stays below strong). Musk ambrette's photosensitization ground for its skin-product ban is independently documented (Cronin 1984, Giovinazzo 1980) and distinct from the neurotoxicity finding.
Source: Ford 1994 (RIFM) Industry-funded Link to this fact
Health effects / Measured in people or real products
Sunlight can turn some fragrance chemicals into skin allergens: the now-banned perfume ingredient 7-methoxycoumarin made 11 of 25 volunteers photoallergic, and in people already sensitized, 0.001% plus UVA light triggered a reaction.
Moderate evidence: Artificial UVA at 20 J/cm2 (~1 h summer sun) for the 0.001% figure; predictive photomaximization test; ingredient banned in EU cosmetics since 2000.
Source: SCCNFP 2000 Link to this fact
Health effects / Lab study (cells or chemistry)
Two sunscreen ingredients, oxybenzone (benzophenone-3) and homosalate, blocked the male hormone receptor in human cells, at roughly the potency of the pesticide metabolite p,p'-DDE.
Moderate evidence.
Caveats
MDA-kb2 reporter cells, IC50 4.98 and 5.57 uM against 0.1 nM DHT; cell assay only, no animal or human anti-androgen endpoint in this paper; data from the group's own Ma 2003. The cosmetics industry's own 2021 CIR benzophenone safety report cites the same research group's 2001 precursor study (matching MCF-7 EC50 of 3.73 uM for benzophenone-3) but its Discussion and Conclusion never weigh this antiandrogenic/estrogenic data when concluding benzophenone-3 is 'safe as used' — industry review of the data is not the same as industry rebuttal of it.
Health effects / Self-reported survey
In a national survey of 1,098 Australian adults, 6.5% said a doctor or health professional had told them they have multiple chemical sensitivities, and 18.9% said they are unusually sensitive to everyday chemicals.
Moderate evidence: Single self-reported 'ever told' item, online panel, adults 18-65; a South Australian population survey found 1% diagnosed; same dataset as the AU fragrance reports.
Source: Steinemann 2018, Preventive Medicine Reports Link to this fact
Health effects / Self-reported survey
Among Australians who said they had been diagnosed with multiple chemical sensitivities, 91.5% reported health problems from fragranced products such as air fresheners, scented cleaners, and other people's perfume.
Moderate evidence: N = 71; self-report; partly built into the definition.
Source: Steinemann 2018, Preventive Medicine Reports Link to this fact
Health effects / Self-reported survey
In a 2016 US national survey, 86% of people who said a doctor had diagnosed them with multiple chemical sensitivity reported health problems from fragranced products, most often breathing problems (50%), migraines (47%) and eye/nose symptoms (47%).
Moderate evidence: Self-report; 'diagnosed' = told by any health professional; opt-in online panel; same dataset as Steinemann 2016.
Source: Steinemann 2018 (JOEM), Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Self-reported survey
In the same 2016 US survey, about 1 in 4 adults (25.9%) said they were unusually sensitive to everyday chemicals such as cleaning products, perfume, and detergents, and 81% of them reported health problems from fragranced products.
Moderate evidence: Self-report; adults 18-65 in an opt-in online panel; earlier US telephone surveys gave 11-16%.
Source: Steinemann 2018 (JOEM), Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Measured in people or real products
Some people whose breathing problems are set off by scents do not have asthma or allergy: in a Swedish clinic study, 11 such patients had normal lung function, negative methacholine and allergy tests, and asthma medicines had given most of them little relief.
Moderate evidence: Clinic referrals selected for scent-triggered complaints; n=11; describes a phenotype, not its prevalence.
Health effects / Measured in people or real products
In a small lab test where patients could not smell it, perfume released near their eyes caused more eye irritation than a saline dummy (1.1 vs 0.2 on a 0-3 scale); a later double-blind study found the same eye effect.
Moderate evidence: Single-blind, n=11, self-rated; corroborated double-blind by Elberling 2006 (8/21 patients vs 2/21 controls, P=0.07).
Health effects / Review of other studies
A systematic review of workplace irritants found that people with pre-existing airway sensitivity (atopy plus non-specific bronchial hyper-responsiveness) can develop new-onset asthma from chronic, repeated exposure to an irritant at concentrations below the official occupational exposure limit.
Moderate evidence.
Caveats
General occupational-irritant literature (solvents, isocyanates, dusts, gases), not fragrance-specific; the below-OEL pattern is narrative synthesis over 3 older sources plus compiled case-series examples, not an independently SIGN 2+-graded finding.
Source: Baur, Bakehe & Vellguth 2012, Journal of Occupational Medicine and Toxicology Link to this fact
Health effects / Measured in people or real products
In a small double-blind lab test, 12 people with multiple chemical sensitivity coughed significantly more than 12 healthy matched controls every time they inhaled capsaicin (the compound that makes chili peppers hot), and the difference grew larger at higher doses.
Moderate evidence: N=12 vs 12, single Gothenburg clinic, MCS diagnosed only by Cullen's clinical criteria (no independent biomarker), not independently replicated outside this research group.
Source: Ternesten-Hasséus 2002, Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Measured in people or real products
The extra coughing in multiple chemical sensitivity patients was not explained by asthma or classic allergy: patients tested negative on standard allergy skin-prick and methacholine airway-reactivity tests, and lung function (FEV1) did not change after capsaicin, pointing to an oversensitive airway nerve response rather than an allergic or asthmatic reaction.
Moderate evidence.
Caveats
Small single-center sample; several patients used asthma/allergy medication for symptom control (withheld before testing), so baseline airway status over time is not fully independent of treatment.
Source: Ternesten-Hasséus 2002, Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Measured in people or real products
In a nationally representative sample of 2,432 US women aged 20-49, women who douched frequently (2+ times/month) had 81% higher blood levels of 1,4-dichlorobenzene than women who never douched, with a dose-response trend across frequency categories (p=0.04).
Moderate evidence: Cross-sectional NHANES biomarker data; wide CI (2% to 221%); mothballs and toilet deodorizers are DCB's dominant known sources and could not be ruled out as confounders.
Source: Ding 2020, Journal of Women's Health Link to this fact
Health effects / Measured in people or real products
Using feminine powder in the past month was linked to 36% higher blood ethylbenzene in a nationally representative NHANES sample of reproductive-aged women.
Moderate evidence: Single VOC, modest effect size; alternative sources (e.g. talc packaging) not ruled out; no dose-response could be tested (frequency not collected).
Source: Ding 2020, Journal of Women's Health Link to this fact
Health effects / Measured in people or real products
Non-Hispanic Black women douched far more often than white women (39% vs 13% in the past month) and had significantly higher blood 1,4-dichlorobenzene, a possible contributor to racial differences in chemical exposure.
Moderate evidence: Association, not a demonstrated causal pathway; unadjusted racial comparison for the biomarker figure.
Source: Ding 2020, Journal of Women's Health Link to this fact
Health effects / Animal study
In a rat model of Parkinson's disease, oral citronellol (25 mg/kg/day) given as a month-long pretreatment before each day's dose of the neurotoxin rotenone reduced brain oxidative stress, inflammatory cytokines, glial activation, and the loss of dopamine-producing neurons.
Moderate evidence.
Caveats
Single dose, single species, prevention paradigm (given before, not after, toxin exposure); dose is 10,000-50,000x above estimated human citronellol exposure and 1,600-8,000x above even after allometric human-equivalent-dose scaling; this is our directly-read primary source for the citronellol 'protective at 25 mg/kg' claim, previously cited only second-hand.
Health effects / Measured in people or real products
Cleaning products were the suspected exposure in 236 of 1,915 (12%) confirmed work-related asthma cases across California, Massachusetts, Michigan, and New Jersey, 1993-1997; most were new-onset asthma (80%) and about a fifth were consistent with reactive airways dysfunction syndrome, an irritant-induced, non-allergic asthma type.
Moderate evidence.
Caveats
This is a primary surveillance source, not a population prevalence estimate: the authors' own capture-recapture analysis found the surveillance system misses 53-87% of true work-related asthma cases in Michigan, and up to 37% of confirmed cases in such systems may not be truly work-related (Malo et al., cited by the authors). Fragrance/limonene was a minor named agent (2 of 300 exposure counts); bleach, acids, disinfectants, and quaternary ammonium compounds dominate.
Source: Rosenman 2003, Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Measured in people or real products
In California, janitors and cleaners had roughly double the confirmed work-related asthma rate of all other occupations combined (4.3 vs 2.3 per 100,000 workers).
Moderate evidence.
Caveats
Single-state comparison using SENSOR case counts against 1990 census occupation denominators; not adjusted for other confounders; occupations with the highest overall work-related-asthma rates (firefighters, correctional officers) were not cleaning-related, so elevated rate alone does not isolate cleaning-product causation.
Source: Rosenman 2003, Journal of Occupational and Environmental Medicine Link to this fact
Health effects / Self-reported survey
In a 1986 clinic and hospital survey of 60 asthma patients, 72% said perfume or cologne worsened their asthma, and two asthma patients who could not smell at all still reacted to perfume worn by others nearby.
Moderate evidence.
Caveats
Small, single-center, unblinded self-report, not selected for representativeness; higher than the ~34-43% airway-symptom rate in later representative population surveys, consistent with a clinic ceiling.
Source: Shim & Williams 1986, The American Journal of Medicine Link to this fact
Health effects / Self-reported survey
In a German cohort of 1,695 young adults (19-24) followed from childhood, frequent private use of household disinfectants (a composite of methods including hand-wash, machine, sponge, and spray disinfection) was associated with close to a threefold increase in the odds of newly developing asthma, compared with people who didn't use disinfectants (OR 2.79, 95% CI 1.14-6.83).
Moderate evidence.
Caveats
Single self-report cohort (SOLAR II); physician-diagnosis-plus-medication-use outcome definition; no causal conclusion can be drawn per the authors; a sensitivity analysis subgroup did not reproduce this exact result.
Source: Weinmann 2017, Occupational and Environmental Medicine Link to this fact
Health effects / Animal study
AETT, a musk fragrance once common in soaps, deodorants, and cosmetics, caused progressive nerve-fiber demyelination and irreversible pigment build-up in the brains of rats fed the compound for 12-20 weeks; the fragrance industry voluntarily withdrew it from cosmetics in 1978 on the strength of its own data.
Moderate evidence: Small single-lab study (N=16 rats), one route (dietary), one dose pair (25/50 mg/kg/day); AETT was withdrawn from cosmetics in 1978 and is not a current fragrance ingredient.
Source: Spencer 1979, Science Industry-funded Link to this fact
Health effects / Measured in people or real products
The fragrance-linked phthalate DEP was found in 78% of the hair products tested, and cyclosiloxanes, parabens, and DEP were found at the highest concentrations, including one anti-frizz product that was 38-46% cyclosiloxane by weight.
Moderate evidence: N=18 products; concentration in the product is not the same as absorbed dose or exposure.
Health effects / Measured in people or real products
Nitrate radicals alone, generated from ozone and nitrogen dioxide with no terpene present at all, raised human eye-blink frequency 21% — comparable in size to the effect of isoprene ozonolysis products — suggesting that some of the eye-irritating power of 'scented product plus ozone' indoor chemistry may come from reactive radicals in general, not from limonene specifically.
Moderate evidence.
Caveats
Single small study; the specific irritant chemical species were not identified or measured, only the percent of ozone chemically consumed. Not yet tested by chemical fractionation in any study in this evidence base.
Health effects / Review of other studies
Odor and fragrance exposure reliably distracts people and slows some cognitive tasks (reaction time, lexical decisions), independent of any toxic effect.
Moderate evidence: Effect sizes are small and inconsistent across tasks, and persistence beyond a single exposure session is not established.
Source: Wolkoff & Nielsen 2017, Environment International Link to this fact
Health effects / Animal study
Mice breathing the vapor of one commercial solid air freshener for one hour showed dose-dependent airway irritation and, at higher doses, a ~35% fall in mid-expiratory airflow — the same magnitude reported for colognes in the same lab's companion study.
Moderate evidence.
Caveats
Same commercial testing lab and authors as the cologne study (Anderson Laboratories); this is within-lab consistency across product classes, not independent replication; FOB blinding status for this study is not clearly stated; exposure measured only as product mass and bulk FID TVOC, not a validated breathing-zone concentration.
Source: Anderson & Anderson 1997, Archives of Environmental Health Link to this fact
Health effects / Measured in people or real products
Using the identical question and instrument in the same Swedish town, adults report being bothered by strong odors about twice as often as teenagers (33% vs 15.6%), and objectively confirmed sensory hyperreactivity is about seven times more common in adults (6.3% vs 0.9%), suggesting chemical sensitivity often develops with age rather than being present from childhood.
Moderate evidence.
Caveats
Cross-sectional age comparison across two separate studies (not a longitudinal cohort); the adult comparator (Johansson 2005/2006) is cited secondhand and is not yet a primary source in our evidence base.
Source: Andersson 2008, International Journal of Hygiene and Environmental Health Link to this fact
Information, not medical advice. See also: myths we won’t tell you.