Tea tree oil is a mixture of about a hundred compounds. The one doing most of the work is terpinen-4-ol, which disrupts the outer membrane of bacteria and fungi so the cell leaks and cannot maintain itself. Australian standards require tea tree oil to contain at least 30% terpinen-4-ol and no more than 15% of a related compound, 1,8-cineole, which is the more irritating fraction. Bottles vary, and a cheap oil may not meet that specification.
The gap between laboratory activity and real results is the important part. Killing Malassezia or acne bacteria in a dish is easy; getting enough oil into a hair follicle or under scale on a scalp, at a concentration the surrounding skin will tolerate, is not. That is why the trial results are consistently modest rather than absent — the mechanism is real and the delivery is the limitation. There is a second catch in the chemistry: when tea tree oil is exposed to air and light its terpenes oxidise into compounds that are considerably more allergenic than the fresh oil, so an old bottle is not just weaker, it is more likely to cause a reaction.
The important thing about tea tree oil's side effects is that they are more common than its reputation as a gentle natural remedy suggests.
Do not use it if you have reacted to tea tree oil before, or to another essential oil in the myrtle family. Sensitisation is permanent and cross-reactions are common. Check labels for "melaleuca oil," which is the same ingredient under a different name.
Do not use it on broken, weeping or eczematous skin, or on skin that is already inflamed. Damaged skin absorbs more and sensitises more readily, and people with eczema are more likely to become allergic to it in the first place.
Do not use leave-on tea tree products on infants and young children, and do not use it on a child's face at all. Keep the bottle out of reach — swallowing it is genuinely dangerous.
Do not put it in the ear canal, in the eyes or in the mouth, and do not swallow it in any amount.
In pregnancy and breastfeeding there is no evidence of harm from ordinary use of a properly diluted product, and there is also very little evidence at all. Given that better-studied alternatives exist for everything tea tree oil is used for, that is the easier choice.
Do not use it as the only treatment for anything that matters — an infected wound, a spreading rash, a nail infection you want cured, or moderate to severe acne. The evidence does not support it in that role.
Acne, 5% gel. Two weeks: little visible change. Six weeks: fewer inflamed spots, and this is roughly where the trials began to show a difference. Twelve weeks: the full effect, which in the head-to-head study was a real reduction in lesion counts — but smaller than 5% benzoyl peroxide achieved, and reached more slowly.
Dandruff and seborrheic dermatitis, 5% shampoo. Itch usually eases in one to two weeks, flaking at three to four. The trial ran four weeks and showed a moderate improvement over placebo. If nothing has changed by six weeks, switch to a proper antifungal shampoo.
Athlete's foot. Four weeks of a strong solution improved symptoms in trials but cured the infection in a minority. A standard pharmacy antifungal cream costs about the same and clears it far more reliably.
Nail fungus. Six months of daily use produced some improvement in appearance and poor rates of actually eradicating the fungus. If you want the nail cured, this is not the route.
Skin tags, warts, cysts and moles. Never. There is no evidence for any of these, and no plausible way a surface oil reaches inside a closed sac under the skin.
When you stop. Whatever was being suppressed returns at the same rate as after stopping any other topical — a few weeks for acne, a few weeks for scalp flaking.