Aromatherapy

Does aromatherapy work, and where does the evidence stop?

Last reviewed

Amber dropper bottles, a lavender sprig, rosemary and a curl of orange peel on a linen cloth on a light oak table in warm light

Partly. Inhaled essential oils show small, real improvements in self-reported sleep and anxiety across several meta-analyses, and the reviewers who found those effects also describe the underlying trials as low quality and inconsistent. Nothing credible supports aromatherapy treating disease. Treat it as mood and routine, measured on questionnaires, not as medicine.

The question people are actually asking

The most common searches in this category are not "which oil for which ailment". They are `is aromatherapy pseudoscience` and `is aromatherapy scientifically proven`. People want to know whether they are being sold something.

Here is the answer we would give a friend.

What the trials do show

There is a real body of clinical work on inhaled essential oils, and it points in one direction: small positive effects on things people report about themselves.

  • A 2025 systematic review and meta-analysis of lavender in adults found a pooled effect on sleep quality of roughly −0.56 standardized mean difference — a moderate effect, statistically clear.
  • A meta-analysis in older adults found a larger pooled effect on sleep quality, around −1.02.
  • Reviews of aroma inhalation in hospital settings report improvements in sleep and anxiety scores.

These are not nothing. A questionnaire score is a legitimate outcome — how rested you feel is the thing you actually care about.

Where it stops, and it stops early

The same reviews are unusually blunt about their own limits:

  • Trial quality is low. Small samples, weak blinding, and enormous variation in what was given, at what dose, for how long.
  • Blinding a smell is close to impossible. Everyone knows whether the room smells of lavender, which is exactly the situation where expectation does the work.
  • Objective sleep measures rarely move. The improvement usually shows up on questionnaires, not on a laboratory readout.
  • Effects are for inhalation. Ingesting essential oils is a different question with a worse safety profile. Nothing here is meant for that.

So: real signal, weak evidence, honest uncertainty. Anyone telling you an oil treats a disease is ahead of the data by a wide margin, and anyone calling the whole field a placebo is behind it.

What follows for how you use it

If the effect is partly cue and partly chemistry, the practical advice is the same either way — and it is not what most shops recommend.

  • One note, not a blend. You cannot learn what works on you from a mixture someone else designed. A single note is one scent, one plant, nothing else in the bottle — and it is the only way to learn what a plant does to you.
  • Repeat it. The same smell at the same time is doing most of the work a routine can do.
  • Use less than you think. Three to five drops is a full room in an ultrasonic diffuser. Beyond that you are not getting more effect, only more oil in the air.
  • Change it when it stops registering. Olfactory adaptation is real and fast; a note you have run every night for a month will fade into the background. Rotating between two or three notes fixes it.

What we sell and why that shapes this page

We sell the devices, not the oil that goes in them. That still gives us a reason to be distrusted on this subject, so the standard we hold ourselves to is: every claim on this site either has a named source or is written as an opinion. If the honest version of an answer loses us a sale — as it does on oils around pets — the honest version is the one that goes up.

Read next

Every guide in this section

Questions people ask next

Is aromatherapy scientifically proven?

Partly, and only for narrow outcomes. Meta-analyses of inhaled essential oils find small-to-moderate improvements in self-reported sleep and anxiety, while describing the trials behind them as low quality. There is no credible evidence that aromatherapy treats disease.

Is it just placebo?

Probably not entirely, and it is hard to prove either way. You cannot blind someone to a smell, so expectation is baked into every trial. What survives that problem is a consistent direction of effect across many small studies.

Which oil is best for anxiety?

Lavender has the most trials behind it, which is not the same as being best for you. Reviews of aromatherapy for anxiety report improvements on rating scales with the same caveats about trial quality that apply to sleep.

How many drops should I use?

Three to five in an ultrasonic diffuser is a full room. In a tealight warmer, two or three into the water in the dish. More is not stronger for long — the nose stops registering a constant scent within about twenty minutes.

Do essential oils expire?

They degrade rather than spoil, mostly through oxidation. Three years is the shelf life most bottlers print on the label. Cap closed, out of sunlight, and amber glass does the rest.

Where these numbers come from

  • The Sleep-Enhancing Effect of Lavender Essential Oil in Adults, systematic review and meta-analysis (PubMed 40600743) — pooled SMD approximately −0.56
  • Effects of aromatherapy on sleep quality in older adults, meta-analysis (PMC11630975) — pooled SMD approximately −1.02
  • Systematic literature review and meta-analysis of aroma inhalation therapy on sleep problems (PMC7939222) — authors' own assessment of evidence quality
  • NCCIH, Anxiety and Complementary Health Approaches — summary of the evidence base for inhaled aromatherapy
  • Shelf-life figures stated by essential-oil bottlers on their own labels — typically three years

How we research and check these numbers