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.
Children and pregnancy / Measured in people or real products
About half of children's diaper-rash creams (13 of 28) and a third of children's sunscreens (10 of 29) listed fragrance.
Weak evidence.
Caveats
Small category counts; label count only. Sunscreen share is market-specific: 64.9% (87/134) of Danish children's sunscreens and sun sprays listed fragrance or a named allergen (Botvid 2023), so 'a third of sunscreens' should not be generalized beyond Brazil; diaper creams were not reported separately in Denmark.
Source: Uber 2024, Jornal de Pediatria (Rio de Janeiro) Link to this fact
Children and pregnancy / Measured in people or real products
In 130 Danish and Finnish mothers, phthalate breakdown products were reported in essentially every breast milk sample (95-100% for each of the six measured) — but an independent, more contamination-controlled US study (Hines 2009) found the analogous class of monoester phthalate metabolites in under 10% of milk samples from a separate cohort, and declined to report its own milk hydrolytic-monoester data (which would include MEP) as likely contamination-inflated.
Weak evidence.
Caveats
1997-2001 samples; composite home-collected milk; phosphoric acid was not added at collection 'due to logistics' (this page's own Methods/limitations), a known contamination risk for hydrolytic monoesters. Hines 2009 (EPA/CDC, tighter contamination controls) found the analogous metabolite class in <10% of milk samples and withheld its own milk hydrolytic-monoester data on contamination grounds. US EPA systematic review (Radke 2018) separately excluded this study's milk data as a critically deficient exposure matrix.
Source: Main 2006, Environmental Health Perspectives Link to this fact
Children and pregnancy / Measured in people or real products
Baby boys whose mothers' milk contained more dibutyl phthalate had slightly lower free testosterone at 3 months (about 15% lower per 10-fold rise). For diethyl phthalate, the fragrance phthalate, the link was with higher SHBG, and its free-testosterone trend was not significant. These are correlations in 96 boys, not proof of harm.
Weak evidence.
Caveats
MBP–free T r = −0.22 (p = 0.033); mEP–free T r = −0.19 (p = 0.07, NS). US EPA review excluded this study (breast milk critically deficient as an exposure matrix for short-chain phthalates) and rated infant-testosterone evidence indeterminate (Lin 2011: MEP r = −0.10). 46% of Danish mothers used breast pumps.
Source: Main 2006, Environmental Health Perspectives Link to this fact
Children and pregnancy / Measured in people or real products
In one early, small US study, boys in the highest phthalate group had about 4 to 10 times the odds (3.8-10.2) of a shorter-than-expected anogenital distance than the lowest group, depending on the phthalate. A larger (N=753) follow-up by the same group found much smaller effect sizes for the one phthalate class (DEHP) that replicated, and did not test comparable categorical odds ratios.
Weak evidence.
Caveats
Single small cohort with very wide CIs (MEP OR 3.8, 1.03-13.9); US EPA rated this cohort's AGD data (extended n=106) low confidence. The DEP (MEP) part did not replicate in three larger cohorts, including the same group's own N=753 TIDES follow-up (Swan 2015), which also found smaller continuous effect sizes for the metabolites that did replicate (DEHP: beta -1.12 to -1.43 in TIDES vs -3.50 to -5.13 in this cohort, per Swan 2015's Discussion), consistent with lower current-day exposure. Quote as one small study's result only; prefer the EPA review or Swan 2015 for general claims.
Source: Swan 2005, Environmental Health Perspectives Link to this fact
Children and pregnancy / Self-reported survey
In a US survey, parents reported that 7 in 10 children aged 12 and under (219 of 312) had used makeup or body products such as face paint, glitter, nail polish or body spray, including more than half of children aged 3 or younger (48 of 84).
Weak evidence.
Caveats
207 self-selected parents recruited mostly online in NYC/Boston/Houston; not nationally representative; ~60% of raw responses excluded as spam/fraud. Partial external check: in a Brazilian clinic survey (Melo 2020) per-category color-cosmetic rates matched Medley's (lipstick 18.8% vs lip ≈21%, nail 22.4% vs ≈22%), but the 70% total is driven by body products (face paint, glitter, bath bombs) that Melo did not ask about, and Melo found make-up in only ≈3% of infants, so the under-3 figure reflects body products, not make-up.
Source: Medley 2023 Link to this fact
Children and pregnancy / Self-reported survey
Among US children who used makeup or body products, about 1 in 8 used them daily, about 1 in 5 wore them for 8 hours or more, and about 1 in 3 had swallowed some in the past year.
Weak evidence: Parent report for 217 child users; ingestion differed sharply by ethnicity (41% vs 11%), possibly a question-wording effect.
Source: Medley 2023 Link to this fact
Children and pregnancy / Self-reported survey
Most US children who use makeup also use products made for adults: only about a third (36%) used mostly children's products.
Weak evidence.
Caveats
Parent estimate of product share; small non-representative sample. Applies to makeup only: for cleansers, a Brazilian clinic survey (Melo 2020) found the majority used children's products (children's soap 56.5%, children's shampoo 65.9%; adult soap 37.3%, adult shampoo 27.2%, rising with age), so do not write 'most children use adult products' generally.
Source: Medley 2023 Link to this fact
Children and pregnancy / Self-reported survey
In a US survey, boys and girls who used makeup or body products used fragrance products (perfume, cologne, body spray) at the same rate, about 31%.
Weak evidence.
Caveats
Scent-purpose products only; cleansers excluded, so not a fragrance-exposure prevalence. Indirect support for sex parity: in Melo 2020 (Brazil, 182/276 perfume users, 126 girls) at least 56/150 boys (≥37%) used perfume, so perfume is not a girls-only product; but the Brazilian level (66% of all children) is 3x the US level (≈21%), so the ~31% figure must not be generalized.
Source: Medley 2023 Link to this fact
Children and pregnancy / Self-reported survey
The only population study of lavender and tea tree oils and hormone problems in children (556 US children) was too small to rule out a risk: regularly exposed children had 2.8 times the rate of endocrine diagnoses, with a 95% confidence interval from 0.35 to 22 times.
Weak evidence: Not statistically significant; parent-reported diagnoses; self-selected natural-product users; 9 events. Supports 'not ruled out', not 'causes'.
Source: Hawkins 2021, International Journal of Pediatrics and Adolescent Medicine Industry-funded Link to this fact
Children and pregnancy / Measured in people or real products
In 2007, US government (NIEHS) scientists described three boys aged 4–10 who developed breast tissue while using products containing lavender oil or lavender and tea tree oil (a homemade balm, hair gel and shampoo, soap and lotions); in each case it went away within months of stopping the products.
Weak evidence: 3 case reports, dechallenge only, no rechallenge, products not analyzed; causation contested (BASF guideline tests negative for linalool/linalyl acetate; Braunstein 2023 review).
Source: Henley 2007, New England Journal of Medicine Link to this fact
Children and pregnancy / Review of other studies
In the reported cases, children had been using lavender-scented products regularly (violet-water colognes, shampoos, soaps, baby bath, lotions) or sat beside a classroom diffuser, and breast growth went away after the products were stopped; early breast growth in children often goes away on its own, so this does not prove cause.
Weak evidence.
Caveats
Case reports with dechallenge only; no comparison group; at least one of the violet-water colognes declared and contained no essential oil by an industry analysis (Giroux & Orjubin 2020). Linklater 2015's own transient alpha-fetoprotein/prolactin elevation resolved on its own over the same 12 months, for an unrelated physiological reason (infant liver immaturity) the authors themselves cite — a concrete illustration, from inside this same case, that concurrent findings can normalize without being caused by the product change.
Source: Hawkins 2020, Complementary Therapies in Medicine Industry-funded Link to this fact
Children and pregnancy / Measured in people or real products
Four prepubertal children (three girls with early breast development and one boy with breast growth) whose doctors recorded continuous use of products they identified as containing lavender saw the breast tissue disappear within three to six months of stopping the product.
Weak evidence.
Caveats
Four uncontrolled case reports, dechallenge only, no product analysis by the authors and no chemical measured in any child; both conditions can resolve spontaneously. An essential-oil industry analysis later reported the boy's violet-water cologne declared and contained no essential oil (DEP, alpha-isomethyl ionone and azo dyes instead), and one girl's baby bath only 'very low' lavender oil (unquantified).
Source: Ramsey 2019, The Journal of Clinical Endocrinology & Metabolism Link to this fact
Children and pregnancy / Self-reported survey
At one Miami children's hospital, 16 of the 24 children seen over seven years for unexplained early breast growth had been using lavender-scented products; 14 of the 16 were Hispanic and the products were violet-water colognes of a kind often applied to babies.
Weak evidence.
Caveats
Clinic case tally with no unexposed comparison group and no use-prevalence denominator; a clinic aware of the lavender hypothesis will ask about lavender; one violet-water cologne from this clinic's case reports declared and contained no essential oil by an industry analysis. Diaz 2016 (2015, same hospital, corresponding author Alejandro Diaz also a co-author here) reported 3 boys with the same outcome in the years just before this tally's window; those 3 cases are almost certainly a subset of this 16/24 figure, not additional to it, so this figure and Diaz 2016 should never be added together.
Source: Ramsey 2019, The Journal of Clinical Endocrinology & Metabolism Link to this fact
Children and pregnancy / Measured in people or real products
One of the four children had never used a lavender product at all: she sat beside a classroom lavender oil diffuser that ran all day for a year, and her breast tissue regressed three months after the diffuser exposure stopped.
Weak evidence: Single case report, no air measurement was made in the classroom, and measured indoor terpene levels elsewhere are far below the concentrations that were active in the cell assays.
Source: Ramsey 2019, The Journal of Clinical Endocrinology & Metabolism Link to this fact
Children and pregnancy / Measured in people or real products
In the same 2018 NGO test, a children's shampoo had 24 unlabeled chemicals flagged for health hazards (17 on the fragrance-industry list), and a baby shampoo had 6.
Weak evidence: Counts, not amounts; single samples; not peer reviewed; not a risk ranking.
Source: BCPP 2018 Link to this fact
Children and pregnancy / Measured in people or real products
In a Danish–Finnish study (hormones measured in 96 of 130 baby boys), some phthalates in mothers' breast milk went with hormone changes at 3 months that point to weaker male-hormone action (higher SHBG; lower free testosterone for dibutyl phthalate only).
Weak evidence.
Caveats
Association; hormones in 96 boys; mEP (DEP) free-T link not significant (p = 0.07); 46% of Danish mothers used breast pumps; US EPA review excluded this study (milk critically deficient as exposure matrix) and rated infant testosterone indeterminate. Cite the primary, not the thesis, which inverted the direction.
Source: Main 2006 Link to this fact
Children and pregnancy / Self-reported survey
Children's lipstick, nail polish and eyeshadow use was about the same in a Brazilian hospital survey and a US parent survey (roughly 19-22%, 22% and 11-13%), but perfume use was three times higher in Brazil (66% vs about 21%).
Weak evidence.
Caveats
Our cross-study arithmetic (Medley per-user rates re-based to all 312 children); different definitions ('perfume' vs 'fragrance products primarily for children'), ages (<=14 vs <=12) and sampling.
Source: Melo 2020, Revista Paulista de Pediatria Link to this fact
Children and pregnancy / Self-reported survey
Every one of 13 Washington DC childcare facilities that reported health data had at least one child with physician-diagnosed asthma, with a typical (median) facility rate of about 10%.
Weak evidence: Center-Director-reported, not chart-confirmed; no exposure-outcome link tested; n=13.
Source: Quirós-Alcalá 2016, Environmental Research Link to this fact
Children and pregnancy / Measured in people or real products
Three more boys (ages 9-11) developed breast tissue while using lavender-labeled products, a shampoo or an 'agua de violetas' cologne sprayed on by a parent since early childhood; in each case it resolved within about 4 months of stopping the product.
Weak evidence.
Caveats
Three uncontrolled case reports, dechallenge only, no rechallenge; same hospital and a shared corresponding author (Alejandro Diaz) as Ramsey 2019's later clinic tally, so this is not an independent replication of the case count.
Source: Diaz 2016, Journal of Pediatric Endocrinology and Metabolism Link to this fact
Children and pregnancy / Measured in people or real products
A 14-month-old girl in Melbourne, Australia developed breast tissue while being given five different lavender-scented infant products every day (wipes, shampoo, body wash, cream and lotion); it went away after the products were stopped, and this case has no connection to the better-known United States case series.
Weak evidence.
Caveats
One case report (letter to the editor), dechallenge only, no rechallenge, no product testing; the exposure was identified only after the family saw a documentary about this exact link, not through routine clinical questioning.
Source: Linklater & Hewitt 2015, Journal of Paediatrics and Child Health Link to this fact
Fragrance is everywhere / Review of other studies
A 2022 review of 37 studies found that phthalates, parabens, aldehydes and aluminum salts are the contaminants most often reported in perfumes, colognes and related scented products.
Weak evidence: Non-quantitative review with factual slips; cite primary studies for numbers.
Source: Kazemi 2022, Journal of Environmental Health Science and Engineering Link to this fact
Fragrance is everywhere / Measured in people or real products
Pine-scent chemicals were also common in Korean household products: alpha-pinene from 19% and beta-myrcene from 12%.
Weak evidence: Headspace presence only; high detection threshold.
Source: Kwon 2007, Journal of Hazardous Materials Link to this fact
Fragrance is everywhere / Hypothesis
Immigrant women in the US most often work as house cleaners and make up 71% of manicurists; an expert job-exposure tool rates both jobs as probably exposed to fragrance chemicals, and house cleaning also to cleaning-product ingredients such as alkylphenols and antimicrobials.
Weak evidence: Exposure ratings from job descriptions (WORC), not measurements.
Source: Knox 2026, Journal of Exposure Science & Environmental Epidemiology Link to this fact
Fragrance is everywhere / Measured in people or real products
The same perfume does not come off the skin the same way every day: in women not on the pill, a fixed dose of vanilla on the forearm was rated about 3 times stronger on cycle day 22 than during menstruation.
Weak evidence: 18 women (9 per group), 25 male raters, no hormone confirmation, no unperfumed control, no statistics printed, storage time unreported.
Source: Chróst 2021, Journal of Cosmetic Dermatology Link to this fact
Fragrance is everywhere / Measured in people or real products
In a mannequin-head test of 16 UK day face moisturizers, the authors report that one application (0.45 g) delivered on average about 16 times more limonene-type terpenes than 24 hours of typical UK home air, and, for the 7 products whose ethanol reached the nose, about 340 times more ethanol.
Weak evidence.
Caveats
Primary: Yeoman 2022 (Indoor Air 32:e12948), Table 6. The abstract's 'up to' is wrong: ×16 and ×339 are means over products with a nonzero dose (14/16, 7/16); maxima ×37 and ×804; over all 16 the medians are ×9 and ×0. Absolute doses are small (≤1.2 mg limonene; below UK PHE limits). The paper's own breathing-zone traces (Fig. 2, limonene ≤0.16 mg/m³) imply doses ~15–500× lower than its tables (evidence base arithmetic), so the ratios are unverified. No people; unheated head, lab at 5 ACH; 'limonene' = all monoterpenes.
Source: Yeoman 2022 via Rádis-Baptista 2023, Journal of Xenobiotics Link to this fact
Fragrance is everywhere / Review of other studies
An NGO analysis (Plastic Soup Foundation, 2022) reported that 87% of products from major European cosmetic brands contained microplastic ingredients.
Weak evidence: NGO figure, unreferenced in the review, definition not stated; major European brands only.
Fragrance is everywhere / Measured in people or real products
In the same 33 breastfeeding women, urinary MEP (the metabolite of DEP, the fragrance phthalate) was on average about 57 times higher in women who "often" used nail polish than in women who never did (2,874 vs. 50.5 micrograms per liter).
Weak evidence.
Caveats
Small pilot study (n=33); frequency-group sizes not reported; perfume/cologne use was not asked about, so fragrance products cannot be separated from nail polish as a source; the difference was significant only on untransformed means, and the authors themselves suggest nail-polish users may also use more perfume and other DEP-containing products.
Source: Hines 2009, Environmental Health Perspectives Link to this fact
Green, natural and unscented / Measured in people or real products
Scented tampons carried a median of 692 ng/g of fragrance terpenes against 2.9 ng/g in unscented ones.
Weak evidence: Only 2 scented vs 20 unscented tampons (p=0.04).
Source: Lin 2020, Environment International Link to this fact
Green, natural and unscented / Government agency
Under EU/UK law, a 'free from [substance]' cosmetic claim about a substance already banned for all EU cosmetics would appear to fall under the rule barring claims that dress up mere legal compliance as a special product benefit.
Weak evidence.
Caveats
This is our own reading of the regulation's general Criterion 1(3) applied to 'free from' wording, not a quotation naming that claim type, and no enforcement decision confirming this application was found.
Source: Commission Regulation (EU) No 655/2013 Link to this fact
Health effects / Review of other studies
A 2022 review concluded that links between contaminants in perfumes and cancer, breathing problems and hormone disruption are 'probable' but still need research to prove.
Weak evidence: Evidence mostly from phthalate/paraben studies not perfume exposure.
Source: Kazemi 2022, Journal of Environmental Health Science and Engineering Link to this fact
Health effects / Lab study (cells or chemistry)
In lab tests on cancer-derived nerve cell lines, three unnamed commercial fragrances changed cell shape and structure after 72 hours and cut the share of cells carrying oxytocin and vasopressin receptors in some lines but not others.
Weak evidence: Cell lines only; fragrances unidentified; no replication; only one cell line per sex.
Source: Sealey 2015, Environmental Research Link to this fact
Health effects / Hypothesis
Silent Spring Institute researchers name house cleaners and nurses as priority jobs for cutting exposure to fragrance, cleaning and personal-care chemicals linked to breast cancer in lab and animal studies.
Weak evidence: No exposure or cancer data in these workers.
Source: Knox 2026, Journal of Exposure Science & Environmental Epidemiology Link to this fact
Health effects / Review of other studies
Many everyday chemicals, including some released by scented and cleaning products, act on the same irritant-sensor nerves in the nose and airways (TRPA1 and TRPV1), one reason very different products can cause similar symptoms.
Weak evidence: Mechanism largely from lab/animal studies; dose relevance to homes untested.
Source: Molot 2023, Neuroscience and Biobehavioral Reviews Link to this fact
Health effects / Animal study
In animal and lab studies, common scent chemicals such as linalool, limonene and eugenol act on brain signaling systems; what lifelong low-level exposure does in healthy people has never been studied.
Weak evidence: Mostly high-dose animal/in-vitro pharmacology; often protective effects in disease models.
Source: Pinkas 2017, Environmental Research Link to this fact
Health effects / Lab study (cells or chemistry)
In lab studies, inflammation lowers the threshold for this irritant receptor, a proposed reason scent reactions can be worse when airways are already inflamed.
Weak evidence: Mechanistic pathway only; not demonstrated in people exposed to products.
Source: Mihara & Shibamoto 2015, Allergy, Asthma & Clinical Immunology Link to this fact
Health effects / Measured in people or real products
Two German gas-works plumbers who repeatedly hand-poured the gas odorant tetrahydrothiophene in an unventilated room developed chronic rhinitis and lasting airway obstruction; neither had allergies on skin testing, and the accident insurer accepted both as occupational disease.
Weak evidence: Two case reports; exposure estimated, not measured; one worker smoked ~16 pack-years; industrial odorant, not fragrance.
Source: Baur & Bittner 2009, American Journal of Industrial Medicine Link to this fact
Health effects / Lab study (cells or chemistry)
Musk ketone does not damage DNA by itself, but in rat liver enzymes and human liver cells it made known carcinogens like benzo[a]pyrene more mutagenic.
Weak evidence: One laboratory's data (the review's senior author); human-cell LOEL 50 ng/ml is 12-70x above whole-milk levels.
Source: Schmeiser 2001, International Journal of Hygiene and Environmental Health Link to this fact
Health effects / Animal study
In two of the five products the airflow reduction kept worsening after the exposure ended.
Weak evidence.
Caveats
Mice, n=3-12 per time course; no human counterpart; mechanism unknown. An independent replication attempt (Tepper et al. 1995, carpets, full text now in our evidence base) found no significant carpet-related toxicity overall via individual-animal statistics, but it never measured airflow limitation or post-exposure persistence of any respiratory endpoint, so it does not directly test this finding either way.
Source: Anderson & Anderson 1998, Archives of Environmental Health Link to this fact
Health effects / Review of other studies
Frequent private use of household disinfectants (a composite of methods including hand-wash, machine, sponge and spray disinfection) was linked to close to a threefold increase in the odds of newly developing asthma in a single cross-sectional wave of young German adults (Weinmann 2017, OR 2.79, 95% CI 1.14-6.83); the evidence for household sprays alone driving that same result was already weak in that wave (OR 2.79, 95% CI 0.84-9.20), and a 10-year longitudinal follow-up of the same cohort (Pacheco Da Silva 2025) found no significant association between persistent spray-only use and current asthma or wheeze (point estimates at or below 1.0 for every outcome), with only a non-significant, attenuated OR of 1.68 (95% CI 0.48-5.88) for sustained use of multiple disinfectants/cleaning products together.
Weak evidence.
Caveats
Two analyses of the same German SOLAR cohort (Weinmann 2017's cross-sectional SOLAR II wave, and Pacheco Da Silva 2025's 10-year longitudinal follow-up adding SOLAR III with latent-class exposure profiles) now agree that spray-specific use is not significantly associated with adult asthma. The only positive signal across both papers is a composite 'persistent use of multiple DCPs' profile, which itself weakened from a significant OR 2.79 (single wave, simple frequency composite) to a non-significant OR 1.68 (10-year follow-up, latent-class model, wide CI 0.48-5.88). Both papers are single-cohort, self-reported-exposure studies from the same research group, so per the grading rule this stays below moderate. The follow-up nulls are underpowered (1,143 of the original 2,051; only ~8-12% in each persistent-use profile; CIs span roughly 0.5-6), so they weaken the claim without refuting it. Other cohorts not yet in our evidence base reported positive spray-asthma associations (ECRHS, Zock 2007; French EGEA, Pacheco Da Silva 2023); don't cite this cohort as evidence that sprays are safe.
Source: Salonen 2024, Environment International Link to this fact
Health effects / Review of other studies
Doctors have published detailed reports of 12 young children (8 boys, 4 girls) whose breasts began to grow while they used lavender or tea tree oil products such as colognes, baby bath soap or hair gel; in every case the growth went away within months of stopping.
Weak evidence: Case reports only; no rechallenge; this kind of breast growth often regresses on its own; the review itself concludes causation is not shown.
Source: Braunstein & Braunstein 2023, touchREVIEWS in Endocrinology Industry-funded Link to this fact
Health effects / Lab study (cells or chemistry)
In human adrenal cells, eight common fragrance terpenes — eucalyptol, geraniol, carvone, 3-carene, ocimene, both pinenes and dihydro-beta-ionone — cut production of the androgen precursor DHEA by 13–31%, and six others cut the activity of aromatase, the enzyme that makes estrogen, by 22–33%.
Weak evidence.
Caveats
Single 10 µM concentration, ~140–200× above what indoor air or ordinary product use could produce internally; the mass-spectrometry check was non-significant for eucalyptol (p = 0.0574) and null for dihydro-beta-ionone; the same 10 µM killed up to 50% of a prostate-cancer cell line by 48 h. Always state the concentration.
Health effects / Lab study (cells or chemistry)
In a lab-grown 3D model of human skin, polystyrene fragments up to 10 µm were cytotoxic and inflammatory in keratinocytes, but only the fragments 2 µm or smaller penetrated into the dermis (average maximum 4.7 µg) within an hour, switching on inflammation genes there.
Weak evidence.
Caveats
One 3D skin model study (Song 2024); independently read the same way by a second review (Menichetti et al. 2025, J Xenobiot), which also reports the <2 µm penetrating cutoff and notes that in ex vivo human and mouse skin from the same primary study, the cutoff for any penetration was stricter still, under 1 µm. Not intact living human skin; no real-use dose.
Health effects / Lab study (cells or chemistry)
Many perfumes contain strongly antibacterial compounds: activity against Gram-negative bacteria appeared at 50 billionths of a liter, and a single spray is about 2,000 times that amount.
Weak evidence: Plate assay with a marine bacterium; effect on the skin microbiome inferred by the authors, not tested.
Source: Morlock & Heil 2025, Journal of Chromatography A Link to this fact
Health effects / Measured in people or real products
In the same test, patients rated breathlessness and cough higher after perfume at the eyes (1.5 vs 0.4 and 0.9 vs 0.3 on a 0-3 scale), yet lung function, oxygen saturation and breathing rate did not change at all.
Weak evidence: Single-blind (scorers apparently not blinded), 3 of many symptoms analyzed, n=11; the airway symptoms did not replicate double-blind (Elberling 2006: 6/11 perfume vs 5/11 placebo).
Health effects / Lab study (cells or chemistry)
A small 2026 follow-up study found mutagenic and cell-killing compounds in 8 perfumes, consistent with an earlier 42-perfume finding from the same German lab.
Weak evidence.
Caveats
Same lab and platform as the earlier study, not independent; only 8 unnamed perfumes, unclear overlap with the 42; simulated liver metabolism was built for this paper but explicitly not tested on the perfumes; no antibacterial or hormone-related endpoints were run.
Source: Schmidtmann 2026, Analytical Chemistry Link to this fact
Health effects / Measured in people or real products
A small clinic study found that a real cologne challenge produced an objective, reproducible 18-58% drop in FEV1 (lung function) within minutes in asthma patients who already reported cologne sensitivity, blocked mostly by bronchodilators rather than cromolyn or blocking the sense of smell.
Weak evidence: N=4 evaluable patients, unblinded first exposure, cologne dose never measured, single cologne brand, all pre-selected for a cologne-sensitivity history, 1986, unreplicated.
Source: Shim & Williams 1986, The American Journal of Medicine Link to this fact
Health effects / Measured in people or real products
Pleasant fragrance smells can measurably calm the body (raising a marker of parasympathetic "rest and digest" activity by up to 47% for pine scent and 26% for citrus), but once that same citrus scent has reacted with ozone, the calming effect reverses into a small drop in that same marker.
Weak evidence.
Caveats
Each condition (pure pine, pure citrus, and the ozone-reacted citrus mixture) comes from a separate, small study, not one head-to-head design; the review's own authors suggest perceived unpleasantness of the reacted mixture, not necessarily toxicity, may explain the reversal.
Source: Wolkoff & Nielsen 2017, Environment International Link to this fact
Health effects / Self-reported survey
In a 10-year longitudinal follow-up of a German young-adult cohort (SOLAR, n=1,143), only participants who persistently used multiple disinfectants and cleaning products (sprays, disinfectant sprays, and nonspray disinfection methods) weekly at two assessments 10 years apart showed an elevated point estimate for current asthma (OR 1.68) and current wheeze (OR 1.71) compared with no weekly use; neither reached statistical significance (95% CI 0.48-5.88 and 0.75-3.90).
Weak evidence.
Caveats
Same underlying cohort as Weinmann 2017 (SOLAR 2 exposure data reused); small persistent-use subgroup (8.2% of n=1,143); wide confidence intervals; incident-asthma/wheeze estimates too imprecise for inference; authors draw no causal conclusion.
Health effects / Animal study
When the air freshener's emissions were combined with fresh latex-paint emissions, the mixture was not less toxic in mice than paint alone — the behavioral (FOB) score rose further and significantly, countering claims that air fresheners neutralize or mask other indoor pollutants.
Weak evidence: Single product, single co-pollutant, one dose combination, mouse behavioral scoring only; not tested with other pollutant classes or in humans.
Source: Anderson & Anderson 1997, Archives of Environmental Health Link to this fact
Health effects / Other evidence
California state scientists estimate that cosmetics may account for at least 4% (and plausibly much more, given wide uncertainty) of the PFOA-related chemicals measured in San Francisco Bay Area wastewater.
Weak evidence.
Caveats
The comparison spans a very wide range (roughly 4% to several hundred percent of the median measured loading) because of unreliable use-concentration inputs; this is an environmental-loading estimate, not a personal health-exposure estimate.
Source: Balan 2024, Environmental Science & Technology Link to this fact
Health effects / Animal study
In ex vivo real human and mouse skin (not a lab-grown skin model), plastic nanoparticles only penetrated when they were smaller than about 1 micrometer, and human skin let slightly more through than mouse skin at that size, likely because human hair follicles are larger.
Weak evidence: Single primary study (Song et al. 2024) as read by this review; ex vivo excised tissue, not living skin measured during real use; no real-use cosmetic dose.
Source: Menichetti et al. 2025, Journal of Xenobiotics Link to this fact
Health effects / Measured in people or real products
Almost half (44%) of patients in the founding irritable larynx syndrome series related their symptom onset to a viral illness, most often not a simple upper respiratory infection but herpes zoster, Epstein-Barr virus, viral meningitis, Lyme disease, or post-viral chronic fatigue syndrome.
Weak evidence.
Caveats
Retrospective patient-recalled association, not confirmed causally; n=39, one clinic; the paper's own abstract says "one third", which its own Results text ("almost half") and Table 4 (17/39=43.6%) contradict.
Health effects / Review of other studies
Citronella oil, marketed as a natural insect repellent, has been banned in the European Union and Canada since 2006, partly because it contains methyl eugenol, which the US National Toxicology Program considers 'reasonably anticipated to be a human carcinogen' based on animal studies.
Weak evidence.
Caveats
No human carcinogenicity data exist for methyl eugenol; this is a regulatory ban and an animal-based NTP classification relayed by a secondary review, not measured harm from using citronella products.
Source: Ramsey 2020 (YJBM), Yale Journal of Biology and Medicine Link to this fact
Health effects / Review of other studies
No published study has been found that tests whether quaternary ammonium compounds specifically from laundry fabric softeners or dryer sheets, as opposed to disinfectant or cosmetic-preservative exposure, cause asthma or skin sensitization in consumers.
Weak evidence: This is an evidence-gap statement, not a positive finding; the review also never discusses esterquats, the ester-linked chemistry that dominates modern fabric softeners, at all.
Source: Peyneau 2022, Frontiers in Toxicology Link to this fact
Hidden ingredients / Measured in people or real products
Acetaldehyde, a possible human carcinogen (IARC Group 2B), was released by 23 of 42 baby products and was never listed on any label.
Weak evidence: Not quantified; may partly form during 40 C incubation.
Source: Nematollahi 2018, Air Quality, Atmosphere & Health Link to this fact
Hidden ingredients / Review of other studies
As early as 2004, a letter in Environmental Health Perspectives warned that fragrances can draw on thousands of chemicals and that most fragranced products had never been tested for human toxicity.
Weak evidence.
Caveats
The '4,000' figure comes from a 1991 book (Ashford & Miller) and describes untested products, not chemicals; later counts run ~3,000. Date 2020, now read as primary, corroborates 'thousands' as the right order of magnitude for the RIFM inventory (~2,600 active + ~3,500 discontinued) and separately states many individual fragrance materials lack toxicity data for standard endpoints — closer to Curtis's 'never tested' claim than any inventory count — but that is about individual chemicals with read-across-fillable data gaps, not about untested products, and does not confirm Curtis's specific '>4,000' figure.
Source: Curtis 2004, Environmental Health Perspectives Link to this fact
Hidden ingredients / Journalism
Laboratory testing commissioned by US campaigners found 38 chemicals not listed on the label across 17 popular fragrances, 14 per product on average, including hormone-disrupting synthetic musks.
Weak evidence: NGO report reproduced in a review; cite Sarantis 2010 directly; primary not in our evidence base.
Source: Sealey 2016 (citing Sarantis 2010, Not So Sexy), Environment International Link to this fact
Hidden ingredients / Other evidence
A 2015 student lab analysis of 12 popular US perfumes and body sprays found 14-33 separate organic chemicals in each product (130 in all), while US law lets the label say just "fragrance".
Weak evidence: Identities are unconfirmed library matches and were never published; counts depend on the detection threshold. Pair with Steinemann 2011 rather than citing alone.
Source: Lee 2015, Young Scientist Journal (Vanderbilt University) Link to this fact
Hidden ingredients / Measured in people or real products
In a university lab test, 3 of 4 cheap generic or dollar-store plug-in air fresheners gave off petroleum-type solvents (alkanes, one with n-hexane) and the fourth was mostly ethanol; none of these adds scent or appears on the package, and the brand-name plug-ins tested showed neither.
Weak evidence: Student journal; one unit per product; identities not confirmed with standards; no amounts measured.
Source: Pino-Delgado 2021 (J Student Res), Journal of Student Research Link to this fact
Hidden ingredients / Government agency
The industry safety panel's own risk calculation for PFOA (a persistent 'forever chemical' associated with PTFE manufacture) in lipstick relies on a single chemical supplier's claim of under 25 parts per billion in the raw material before it is formulated into a product — no actual PTFE-containing lipstick was tested for PFOA.
Weak evidence.
Caveats
The <25 ppb figure and the resulting '755-fold below EPA's advisory' conclusion are worst-case arithmetic on an unverified, pre-formulation raw-material claim, not a measurement of any real cosmetic product.
Source: CIR 2018 Industry-funded Link to this fact
Hidden ingredients / Industry source
PTFE (Teflon) is reportedly used in some cosmetics at concentrations up to 30%, more than double the 13% leave-on maximum cited from the industry-funded Cosmetic Ingredient Review.
Weak evidence.
Caveats
The paper's own use-concentration data for most PFAS ingredients 'differed by orders of magnitude' across sources; the specific source and current-market relevance of the 30% figure could not be verified for this ingest, and it may reflect a patent-claimed rather than marketed formulation.
Source: Balan 2024, Environmental Science & Technology Link to this fact
Indoor air / Measured in people or real products
In a week-long study of 7 working adults in Leipzig, the air at home made up 42-73% of the airborne chemicals they breathed, more than the outdoor air or their workplaces.
Weak evidence: N=7, one week, 25 VOCs; home share includes infiltrated outdoor air.
Source: Gokhale 2008, Science of the Total Environment Link to this fact
Indoor air / Measured in people or real products
In a real shower test, limonene in the bathroom air rose to about 80, 150 and 380 ppb as the amount of shower product went up, roughly in proportion to the amount used.
Weak evidence: One participant, one bathroom, PTR-MS; limonene only; peaks last minutes.
Indoor air / Hypothesis
A computer model of a morning shower found fragrance terpenes from the products react with indoor ozone to make formaldehyde (about +4 ppb), other aldehydes, irritant PAN-type nitrates and a little fine particle matter, even with strong ventilation.
Weak evidence: Modeled (INDCM), not measured; increments are small and last about an hour; water-soluble products may dissolve in shower water, so these are upper limits.
Indoor air / Animal study
A continuously running essential-oil diffuser is now the exposure pattern with an animal result attached to it: continuous ambient lavender exposure advanced puberty in rats, while a larger once-daily nasal dose did not.
Weak evidence: Neither exposure was quantified, so the comparison shows a pattern, not a dose-response; with five animals per group the difference rests partly on one outlier.
Source: Kim & Lim 2022, Journal of Korean Medical Science Link to this fact
Indoor air / Instrument measurement
In a small closed test room, three short bursts of aerosol hairspray pushed fine-particle (PM2.5) readings to about 90 µg/m³, roughly 2.5 times the US 24-hour outdoor standard, though only for minutes.
Weak evidence: Uncalibrated consumer monitor; tiny room; short peak vs 24-h standard; primary spray aerosol, no fragrance characterized.
Indoor air / Measured in people or real products
A fresh, 10-minute-old mixture of limonene and ozone, at concentrations close to the high end measured in real homes and offices, made people's eyes blink significantly more often in a small controlled study, at roughly 2,000 times less limonene than pure limonene alone needs to irritate eyes.
Weak evidence.
Caveats
This is the paper's own LOEL (residual 0.23 mg/m3 limonene + 0.08 mg/m3 ozone, +17%, p=0.003) and was previously mis-cited on our evidence base to Klenø & Wolkoff 2004. Now verified as primary, but the same lab's own comparable re-test of Klenø & Wolkoff 2004 gave a 3.5x smaller effect (12% vs 42%) at nearly identical residual concentrations, and no dose-response held across the paper's own four LOPs mixtures (R2=0.7, ns).
Indoor air / Measured in people or real products
Methacrolein, a known limonene/isoprene ozonolysis product, produced a similarly-sized eye-blink-frequency increase on its own (+18% at 286 ppb) as the freshly reacted limonene-ozone mixture did, supporting the idea that it is reaction products, not parent limonene, that irritate eyes.
Weak evidence.
Caveats
Methacrolein was tested alone, not measured inside the LOPs mixtures, so this is an analogy of magnitude, not a demonstration that methacrolein specifically drives the limonene-ozone effect. Single small study, single lab.
Indoor air / Measured in people or real products
Whether an office air-conditioning filter was new or had been in service for about 10 months made no measurable difference to ozone or particle chemistry, or to occupants' reported symptoms, during ozone-limonene exposure.
Weak evidence.
Caveats
Compares only new vs. moderately used (MERV 13) filters at fixed ventilation/recirculation rates; does not test badly fouled filters or other filter grades, so it should not be read as showing filter replacement is never protective.
The industry / Journalism
The first press coverage of a 2011 study finding 133 undisclosed chemicals in 25 best-selling scented products, in the US government's environmental-health journal, quoted no manufacturer or industry group; the underlying numbers should be cited from the study itself (Steinemann 2011).
Weak evidence: Media-coverage note only; all product numbers belong to the primary src-2011-fragranced-consumer-products-chemicals-emitted-ingredients-unlisted.
Source: Potera 2011 (EHP news), Environmental Health Perspectives Link to this fact
The industry / Review of other studies
The fragrance industry's answer to autoxidation is to add antioxidants to linalool and linalool-rich natural extracts, but the Gothenburg researchers say that protection is hard to control.
Weak evidence: Industry practice via personal communication from a RIFM scientist (A. M. Api) in a RIFM co-funded paper; no data on effectiveness.
Source: Hagvall 2008, Contact Dermatitis Industry-funded Link to this fact
The industry / Animal study
Two common fragrance ingredients, hexyl and benzyl salicylate, may test positive in the mouse skin-allergy assay not because they are true human sensitizers, but because of a mouse-specific irritant/pharmacological effect, similar to a known false-positive detergent (sodium lauryl sulfate).
Weak evidence.
Caveats
A hypothesis proposed by structural-alert absence and analogy only; the paper offers no direct evidence (e.g., a biomarker distinguishing true haptenation from irritant-driven lymph node activity) and the underlying data are RIFM's own unpublished proprietary dataset.
Source: Roberts & Api 2018, Regulatory Toxicology and Pharmacology Industry-funded Link to this fact
The industry / Review of other studies
The detergent industry's own risk assessment for esterquat fabric softeners sets a margin of exposure of about 8,100 (NOAEL 300 mg/kg bw/day vs. an aggregate consumer exposure estimate of 36.9 micrograms/kg bw/day) and concludes esterquats in fabric conditioner are safe for consumer use.
Weak evidence.
Caveats
Every number comes from the industry's own unpublished, unnamed studies with no independent citation trail; the exposure model covers only dermal and minor oral routes (no inhalation route at all, despite fabric conditioner being used in heated dryers); no carcinogenicity study and no multigeneration reproductive study exist for esterquats — the 'safe' conclusion infers these endpoints are unnecessary from adjacent negative results rather than testing them directly.
Source: HERA Project 2008 (A.I.S.E./CEFIC) Industry-funded Link to this fact
Laws and loopholes / Journalism
As of 2010, the US FTC's Green Guides on environmental marketing claims had not been updated since 1998.
Weak evidence: Outdated: the Green Guides were revised in 2012 (and an update was proposed in 2022); do not use as a current fact.
Source: Dahl 2010, Environmental Health Perspectives Link to this fact
Laws and loopholes / Measured in people or real products
Musk ambrette, banned in the EU for years, was still detected in the blood of 2 of 100 young Austrians tested, and testing the specific lotion and perfume one of them used did not find musk ambrette in the product itself, so the source of the banned chemical could not be pinned down.
Weak evidence.
Caveats
N=2; the tested products were negative for musk ambrette, so contamination, an unlisted supply-chain source, or a different untested product cannot be distinguished; no source identified.
Source: Hutter 2009, Science of the Total Environment Link to this fact
Laws and loopholes / Government agency
Allergy UK's product-endorsement page does not disclose who pays for testing or endorsement; it sits under a business-facing URL and ends in a manufacturer submission form, consistent with (but not confirmation of) a manufacturer-initiated, fee-based process.
Weak evidence: Fee/payer structure is inferred from page layout, not stated by the source; do not cite this page for a fee claim.
Source: Allergy UK 2026 Link to this fact
Skin allergy / Measured in people or real products
Most patients who reacted to oxidized scent chemicals (58%) also had other fragrance allergies or a history of reacting to fragranced products.
Weak evidence: 31 positive patients.
Skin allergy / Measured in people or real products
People without skin disease reacted to oxidized limonene and linalool as strongly as dermatitis patients.
Weak evidence: Small groups (9 controls per comparison after exclusions).
Source: Bråred Christensson 2009, Contact Dermatitis Industry-funded Link to this fact
Skin allergy / Lab study (cells or chemistry)
In lab tests on human skin cells, the oxidized forms of limonene and linalool reshaped the cells' fats in the same way as a classic potent skin allergen (DNFB), a pattern consistent with oxidative stress.
Weak evidence: One cell line, one dose.
Skin allergy / Lab study (cells or chemistry)
In lab tests on reconstructed human skin, oxidized linalool broke down into reactive radicals that bind to skin proteins, which helps explain why linalool becomes an allergen only after it reacts with air.
Weak evidence: Skin model and model thiols; mechanism, not prevalence.
Source: Kuresepi 2020, Chemical Research in Toxicology Link to this fact
Skin allergy / Lab study (cells or chemistry)
In lab tests on human skin cells, the air-oxidized forms of limonene and linalool (the forms that cause fragrance allergy) damaged the fats in cell membranes at concentrations from 100 µM, with oxidized limonene strongest; a possible second route to skin inflammation not yet shown in people.
Weak evidence: Cell line and liposomes only; added iron; real skin doses unknown.
Skin allergy / Measured in people or real products
Once someone is sensitized, patch tests can trigger reactions at low ppm concentrations of oxidized terpene — but the specific 'as low as 1 ppm' figure is for Δ³-carene in older, narratively-reviewed occupational turpentine literature; for the two terpenes that dominate consumer exposure, modern peer-reviewed dose-response studies put the threshold one to two orders of magnitude higher (limonene hydroperoxides: ~24 ppm single patch application, ~140 ppm repeated realistic exposure, ED₁₀ ≈85 ppm; linalool hydroperoxides: ~560 ppm repeated realistic exposure).
Weak evidence.
Caveats
The 1 ppm figure applies to Δ³-carene in historical occupational patch testing, not to limonene or linalool, and its own study design (occlusion, participant numbers) is not given in the citing review. No limonene- or linalool-specific study supports reactions at ~1 ppm.
Source: Karlberg & Lepoittevin 2021, Contact Dermatitis Link to this fact
Skin allergy / Measured in people or real products
All four patients already allergic to oxidized linalool reacted to air-exposed lavender oil on patch testing.
Weak evidence: N=4 pre-selected sensitized patients; fresh oil not tested alongside; elicitation, not sensitization.
Source: Hagvall 2008, Contact Dermatitis Industry-funded Link to this fact
What you can do / Measured in people or real products
Childcare rooms with more fresh-air exchange had lower levels of several of the chemicals measured.
Weak evidence: Bivariate correlations, 34 rooms.
What you can do / Measured in people or real products
Keep scented products closed and avoid old, long-opened ones: limonene and linalool became skin irritants after weeks of air exposure.
Weak evidence.
Caveats
Extrapolated from lab-oxidized neat material (10-20 weeks air exposure); product storage not tested by this paper. Two independent industry market surveys have now directly tested the specific mechanism this advice assumes (in-bottle oxidation with age) and both found it absent: Kern 2014 (39 aged fine fragrances + 29 antiperspirants, 9 months accelerated storage) and Natsch 2019 (104 products spanning fine fragrances, antiperspirants, creams and lotions, including two fine fragrances tracked 3-6 years at room temperature) both found no progressive hydroperoxide accumulation with age, and consumer-product doses 2-4 orders of magnitude below elicitation/sensitization thresholds. Both studies are industry-funded (IDEA consortium: Givaudan, Firmenich, IFF, IFRA) and not independently replicated by a non-industry group, and neither tests essential oils or raw materials (separately known to accumulate hydroperoxide) or every product category. Moved to claims-to-avoid because the specific mechanistic basis for this advice (in-container autoxidation) is now negative in the only two real-product tests that exist, even though the underlying autoxidation chemistry itself, in neat/essential oils, is not disputed. Kern 2014 and Natsch 2019 are both industry studies (Givaudan / IDEA consortium) with Natsch as a co-author on both, so they are not independent replications of each other.
Source: Bråred Christensson 2009, Contact Dermatitis Industry-funded Link to this fact
What you can do / Review of other studies
Old, long-opened scented products are more allergenic than fresh ones, because the scent chemicals oxidize on contact with air.
Weak evidence.
Caveats
Time course is primary for lavender oil (Hagvall 2008: LLNA EC3 36% fresh -> 23% at 10 weeks -> 4.4% at 45 weeks in the same batch), but in neat oil in open lab flasks, not formulated products as stored. Kern 2014 first closed part of this gap (39 fine fragrances + 29 antiperspirants, no accumulation over 9 months accelerated storage); Natsch 2019 extends this to 104 products across more categories (adding limonene, creams, lotions) and to multi-year real-time storage (two fine fragrances tracked 3-6 years at room temperature, no accumulation), plus tests products donated by allergic patients themselves (also negative for hydroperoxide). Downgraded to avoid because the specific claim — that aging/opening itself makes a product more allergenic via in-container oxidation — has now failed two independent, broader industry market surveys with no non-industry evidence on the other side; creams, lotions and fine fragrances/antiperspirants are covered, though essential oils, raw materials and other categories remain untested and are not covered by this downgrade. Kern 2014 and Natsch 2019 are both industry studies (Givaudan / IDEA consortium) with Natsch as a co-author on both, so they are not independent replications of each other.
Source: Karlberg & Lepoittevin 2021, Contact Dermatitis Link to this fact
What you can do / Self-reported survey
People affected by fragrance say policies work only with purchasing rules for fragrance-free, low-emission products, staff education, better ventilation, and consequences for repeat violations, enforced like smoke-free rules.
Weak evidence: Participants' and authors' proposals; not tested as an intervention.
Source: Roy 2026 Link to this fact
What you can do / Hypothesis
Air a room or clean it instead of scenting it: non-chemical alternatives such as opening windows and better hygiene are the first step in Klaschka's ten-step hierarchy for reducing fragrance exposure.
Weak evidence: Expert recommendation; not tested in the paper.
Source: Klaschka 2020, Environmental Sciences Europe Link to this fact
What you can do / Review of other studies
If a child develops unexplained breast tissue, tell the doctor about every scented product the child uses or is near, including colognes, baby washes, hair products and diffusers; stopping them is what the case reports did.
Weak evidence: Practical advice drawn from case reports, not a tested intervention.
Source: Hawkins 2020, Complementary Therapies in Medicine Industry-funded Link to this fact
What you can do / Self-reported survey
People who always or usually used a healthy-product app, read ingredient labels or looked for eco-friendly labels used hair, skin and beauty products with somewhat lower EWG hazard ratings, by up to half a point on a 10-point scale.
Weak evidence: Cross-sectional; the rating is built from the same labels, so it cannot see undisclosed fragrance chemicals; no exposure measured.
Source: Barrett 2025, Journal of Exposure Science & Environmental Epidemiology Link to this fact
What you can do / Measured in people or real products
Using less shower product and running the bathroom fan both lower the fragrance-terpene peak: emissions scaled with product amount, and halving ventilation raised the modeled limonene peak from about 375 to 495 ppb.
Weak evidence: Amount-scaling from one person's shower; ventilation effect is modeled; no health outcome measured.
What you can do / Test-chamber measurement
Keeping indoor ozone low (no ozone-emitting 'air purifiers' or ionizers) limits both skin-oil chemistry and fragrance-terpene chemistry indoors, because ozone starts both chains.
Weak evidence: This paper makes no recommendation; the ozone-generator advice comes from Singer 2006, and the ventilation/ozone trade-off is our inference.
What you can do / Review of other studies
If you want to avoid persistent fluorinated ingredients, check cosmetic labels for 'PTFE', 'polytetrafluoroethylene', or names containing 'perfluoro-' or 'polyfluoro-'.
Weak evidence.
Caveats
Namazkar 2024 found INCI names for PFAS ingredients can themselves be ambiguous or incorrect (an apparent isomer mismatch for a silane; an undisclosed shorter chain-length range for a phosphate ester), and found unlabeled residual PFCA impurities in nearly every PFAS-containing cosmetic tested (5 of 6 products), which were far more dermally bioaccessible (43-76%) than the labeled ingredient itself (negligible bioaccessibility). Label-reading is a reasonable screen for intentionally added, correctly-named PFAS but does not detect the unlabeled residual impurities this paper shows can be present and bioaccessible regardless of which PFAS is named.
Source: Jomaa 2024, Journal of the Netherlands Society of Toxicology (JNST) Industry-funded Link to this fact
What you can do / Lab study (cells or chemistry)
Only 6 of 42 perfumes tested were comparatively low in hazardous compounds across ten lab tests, which suggests lower-hazard perfumes can be made.
Weak evidence: Products not named, so consumers cannot identify the cleaner ones; in vitro only.
Source: Morlock & Heil 2025, Journal of Chromatography A Link to this fact
What you can do / Instrument measurement
A filter-only HEPA air purifier kept fine-particle readings near zero after hair and nail products were used together in a small closed room.
Weak evidence: Flow-to-volume ratio far above any real home or salon; VOC removal by the carbon filter not credible for acetone/formaldehyde.
What you can do / Review of other studies
Fragrance-free policies are easier to make work when they start with a needs assessment and an inventory of the cleaning and personal products in use, involve employees in drafting, and agree how violations are handled before launch.
Weak evidence: Lessons from unnamed organizations; no outcome data.
What you can do / Review of other studies
Close fragranced products after use and discard them about a year after opening to limit exposure to oxidized allergens.
Weak evidence.
Caveats
Expert advice; no study tests whether it reduces dermatitis. The assumed mechanism (in-bottle oxidation during storage) has now failed two industry market surveys rather than one: Kern 2014 (fine fragrances, antiperspirants, 9 months accelerated storage) and Natsch 2019 (104 products including creams and lotions, two fine fragrances tracked 3-6 years at room temperature) both found no hydroperoxide accumulation with age or storage time. The advice may still be reasonable for essential oils and raw materials (which do carry hydroperoxide, e.g. lavandin ~220 ug/g), just not via this specific in-bottle mechanism for finished consumer products. Both studies are industry-funded (IDEA consortium: Givaudan, Firmenich, IFF, IFRA) and not independently replicated, and say nothing about essential oils, raw materials or product categories outside fine fragrances/antiperspirants/creams/lotions. Kern 2014 and Natsch 2019 are both industry studies (Givaudan / IDEA consortium) with Natsch as a co-author on both, so they are not independent replications of each other.
Source: Pastor-Nieto 2021 (Klaschka 2010), Current Treatment Options in Allergy Link to this fact
What you can do / Measured in people or real products
Moisturizers that listed alcohol or 'alcohol denat.' near the top of the ingredient list delivered far more ethanol to a mannequin's nose; products with no listed alcohol gave almost none.
Weak evidence: 16 products, qualitative (Fig. 4, no statistic); ethanol hidden inside 'parfum' is not visible on labels; absolute ethanol doses are disputed.
What you can do / Measured in people or real products
'Agua de violetas' colognes marketed to Hispanic families, including preparations made for babies, are a specific consumer product worth asking about if a young child has unexplained breast growth.
Weak evidence: Based on 2 of 3 case reports from one clinic; not every bottle sold under this name contains lavender oil.
Source: Diaz 2016, Journal of Pediatric Endocrinology and Metabolism Link to this fact
What you can do / Hypothesis
If you already notice fragrance-related symptoms, do not expect awareness alone to change your exposure — most Germans who link their own symptoms to fragrance still use fragranced products at or above the average rate; choosing fragrance-free products and spaces is the step that actually reduces exposure.
Weak evidence: Author's recommendation, drawn from the self-exposure paradox in this survey; not itself tested as an intervention.
Source: Klaschka 2020a, Environmental Sciences Europe Link to this fact
What you can do / Review of other studies
If households switch from fragranced to fragrance-free laundry detergent, modeling suggests dryer-vent limonene emissions across a whole city or state could drop substantially — an estimated 1.58 tons per year for metropolitan Melbourne, or 8.66 tons per year for the state of California.
Weak evidence.
Caveats
This is a modeled extrapolation, not a measurement: it multiplies a mean vent concentration from just 2-6 households (one grab sample each) by census household counts and usage-rate assumptions, with no confidence interval or sensitivity analysis anywhere in the calculation; the Melbourne input is dominated by a single outlier household roughly 11 times the other three.
Source: Goodman, Nematollahi & Steinemann 2020/2021, Air Quality, Atmosphere & Health Link to this fact
Information, not medical advice. See also: myths we won’t tell you.