Don’t take our word for it. Here’s every fact on this site, with its evidence grade, its caveats, and a link to the source, so you can check us.
We start you on the strong ones: findings that have been replicated or come from one large, well-controlled study. Moderate and weak facts are here too, labeled for what they are. Use “Link to this fact” to share a single card.
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 / Self-reported survey
Perfume is most often sprayed on the neck and throat, then the arms — close to the nose and mouth — so people concerned about fragrance inhalation might consider applying it lower on the body or on clothing instead.
Weak evidence: Based on where 326-340 already-regular perfume users in one US industry survey said they applied product, not on any measurement of how application site changes inhaled dose.
Source: Loretz 2006, Food and Chemical Toxicology Industry-funded Link to this fact
What you can do / Self-reported survey
Migraineurs who notice perfume, paint/varnish, gasoline, or bleach smells before their attacks may want to keep a trigger diary, since these were the most common self-reported odor triggers in one clinic sample.
Weak evidence.
Caveats
Rests on a single non-random clinic sample (n=200, one Brazilian city, 2011); a personal pattern should be confirmed prospectively, not assumed from one study, and may reflect premonitory-phase symptom misattribution rather than a true external trigger.
What you can do / Self-reported survey
Migraineurs who notice their attacks are triggered by specific odors (most commonly perfume, cleaning products, cigarette smoke, or vehicle exhaust in this sample) may reasonably consider reducing exposure to those specific triggers as a low-cost self-management step.
Weak evidence: Based on self-reported association in one clinical sample, not a controlled-avoidance or provocation trial; benefit of avoidance was not tested.
Source: Fornazieri 2016, Headache: The Journal of Head and Face Pain Link to this fact
What you can do / Measured in people or real products
In 21 patients whose hydroperoxide allergy was judged certainly relevant, the products blamed were mostly rinse-off products such as soap and shampoo (15 of 21), then leave-on cosmetics and creams (9 of 21), plus detergents, deodorants and perfumes.
Weak evidence.
Caveats
Small subgroup from one clinic; product type is the patient's and dermatologist's identification of products containing limonene or linalool on the label, not confirmed hydroperoxide content.
Source: Dittmar & Schuttelaar 2019, Contact Dermatitis Link to this fact
What you can do / Other evidence
In a lab model, perfume ingredients rubbed from one cotton swatch onto a clean one in as little as 10 seconds, and were still transferring when the perfume had dried on the first fabric for 48 hours.
Weak evidence.
Caveats
Six pure compounds at about 50 times below typical perfume strength, cotton to cotton, crockmeter rubbing under a 9 N load, n = 3; one 48 h sample; forensic bench study, not a health or exposure study.
What you can do / Other evidence
Women who used anti-aging facial cream in a Mexican study had about 2.5 times the median monoethyl phthalate of nonusers (178 vs 71 ug/g creatinine); perfume users showed a similar-sized but statistically uncertain difference.
Weak evidence: Only 11 and 19 users, unadjusted rank-sum tests with about 80 product-metabolite comparisons, so a chance finding is plausible; perfume's own MEP difference was not significant.
Source: Romero-Franco 2011, Environment International Link to this fact
What you can do / Other evidence
In the perfume rat study, oral vitamin E (40 mg/kg/day) lessened many of the changes (blood pressure, heart rate, serum MDA, CYP1A1/CYP1A2 induction) without normalizing creatinine, troponin or PO2.
Weak evidence.
Caveats
Rat oral dose, five animals per group, no human test; the printed serum MDA for the vitamin E group is identical to the control value, suggesting a transcription error; statistical letters show partial effects only.
Source: Al-Sowayan 2026, Frontiers in Pharmacology Link to this fact
Information, not medical advice. See also: myths we won’t tell you.