Does lavender actually help you sleep?

In this trial, yes, but indirectly. A standardised lavender oil capsule improved sleep quality significantly more than placebo over ten weeks (p = 0.002). The mediation analysis then showed that 98.4% of that improvement travelled through the drop in anxiety, and only 1.6% was a direct effect on sleep — a share that could not be told apart from chance.

Does lavender make you drowsy during the day?

The paper's starting point is that this lavender preparation improves sleep without sedation, which is why the result was worth explaining. That is the practical difference from most drugs recommended for anxiety and insomnia, where the daytime dullness and the sleep are two sides of one effect.

What dose of lavender was used?

80 mg a day of Silexan, a proprietary standardised essential oil of Lavandula angustifolia taken as a capsule, for ten weeks. This is a purified, measured, swallowed preparation and not the same thing as the plant, an aroma or a topical oil.

Does a lavender sachet or a pillow spray do the same thing?

Nothing in this study says so. What was tested was a cleaned, standardised lavender extract sealed in a capsule and taken every day at a measured dose. A sachet in a drawer was not tested here, and the result cannot be carried across to it.

Will lavender help if I sleep badly but I am not anxious?

This study cannot answer that. The patients were being treated for anxiety, and anxiety was the entire road from the capsule to the sleep. If a person has no anxiety to remove, the mechanism the paper describes has nothing to work on.

Who paid for the study?

One of the three authors, Sandra Schläfke, works for Dr. Willmar Schwabe GmbH & Co. KG, the company that makes the lavender preparation Silexan. That does not make the numbers wrong, and it is worth knowing when you read them.

Lavender improves sleep by removing anxiety, not by sedating you

In short

A 2019 re-analysis of a ten-week randomised trial in 212 patients asked why a lavender oil capsule improves sleep. The preparation was Silexan, a standardised Lavandula angustifolia oil, at 80 mg a day against placebo.
98.4% of its effect on sleep was explained by its effect on anxiety. The remaining 1.6% — a direct effect on sleep — could not be told apart from chance (p = 0.958).
Anxiety separated from placebo at week 2, sleep only at week 6. A four-week gap is what a knock-on effect looks like, not a direct one.
So it improves sleep without sedation because it is not acting on sleep at all. It removes the anxiety, and sleep returns on its own.
It was tested in people being treated for anxiety. Nothing here says what it does for somebody who sleeps badly and is not anxious.

Many people keep lavender in a drawer. A small bag next to the socks, a bar of soap, a wind chime from a gift shop. The internet is full of advice to put three drops on your pillow. Most of us read that as a harmless superstition and stop thinking about it after four seconds.

In 2019 a group of researchers published a paper in the Journal of Psychiatric Research. The first author works at the psychiatric hospital of the University of Zurich.

They were not writing about dried flowers. They had 212 patients, ten weeks of treatment, a placebo group and strict statistics. By the end of the paper they had taken away the main selling point of ordinary sleeping pills.

The deal you make with a sleeping pill

If a doctor has ever prescribed you something for sleep, you know the trade. The drug produces sedation, and that dullness is what pushes you into sleep. Most medicines recommended for anxiety and insomnia have a sedating effect, and that effect is normally believed to be the reason sleep gets better.

Sedation is not a flaw that chemists are still working on. It is part of how the result is produced. The problem is that sedation does not repair sleep. It mutes the nervous system. Along with the insomnia it takes away clear thinking and morning energy, and over time it asks for a larger dose.

So there is an obvious question. Can sleep come back without the dullness?

Back to the drawer

The answer had been sitting next to the socks the whole time — but it is not about the little bag, and not about drops on a pillow.

Take a standardised extract of Lavandula angustifolia, seal it in a capsule and take it every day like a medicine. At that point the folklore stops and clinical research starts. The preparation used in these trials is called Silexan.

Clinical trials had already shown two things about it. It reduces anxiety and it improves sleep, both against placebo. And it does this without sedation.

In pharmacology that looked like a contradiction. Sleep was supposed to improve because of the dullness. Here there was no daytime dullness and sleep improved anyway. One of those two beliefs had to be wrong, and somebody had to find out which one.

Two possible explanations, and only two

The researchers did not run a new trial. They went back to data that already existed and asked it a question it had not been collected to answer.

Lavender acts on sleep directly. Some mechanism nobody has described yet improves sleep on its own, without going through anxiety. If this is true, the perfect sleeping pill has been found: sleep with no sedation.
Lavender does not act on sleep at all. It removes anxiety. And a person who has stopped being anxious simply sleeps, the way anyone sleeps once they have calmed down. Good sleep would then not be an effect of the drug. It would be an effect of an effect.

A doctor with a stack of questionnaires cannot tell these apart. In both cases a patient is sitting there saying they sleep better now. From the outside it is the same picture.

What they actually did

For ten weeks, half of the 212 patients took 80 mg a day of the lavender oil preparation, and half took a placebo. Neither the patients nor the doctors knew who got which. These were patients being treated for anxiety, and disturbed sleep was one of their symptoms.

Two things were measured:

Anxiety, with the Hamilton Anxiety Scale (HAMA).
Sleep quality, with the Pittsburgh Sleep Quality Index (PSQI).

After ten weeks both were clearly better than placebo: anxiety at p < 0.001 and sleep at p = 0.002. The capsule worked on both fronts. That still did not say which of the two explanations was right.

So they stopped looking at how much the scores changed and looked at when.

What was measured	Scale	When it separated from placebo
Anxiety	Hamilton Anxiety Scale (HAMA)	week 2
Sleep quality	Pittsburgh Sleep Quality Index (PSQI)	week 6

Four weeks apart. If lavender acted on sleep directly, there would be no reason for sleep to arrive a month late. A direct effect arrives with the substance. The thing that arrives late is the consequence, because it has to wait for its cause.

The arithmetic

To settle it properly the authors ran a mediation analysis. They took the whole improvement in sleep and split it into two parts: the part that travelled through the drop in anxiety, and the part that acted on sleep directly.

The split was almost total:

98.4% of the effect of the lavender preparation on sleep was explained by what it did to anxiety.
1.6% was left for a direct effect on sleep. The direct effect had a p-value of 0.958 — nothing that can be told apart from chance.

The second explanation won. This is not a sleeping pill. It removed the anxiety, and sleep came back on its own.

Why the boring answer is the better one

The first explanation looked like the exciting one, and many people would have bet on it. But the trouble with ordinary sleeping pills is not that they fail. They work. The trouble is the price: sleep always comes bundled with daytime dullness.

The dull-sounding answer removes that price completely. If normal sleep is simply what happens once a person is calm, then nobody has to buy the dullness to get it. The capsule improves sleep almost entirely through its effect on anxiety rather than through sedation.

What this does not say

Four limits, and they matter as much as the result.

This is one re-analysis of one trial. 212 patients, ten weeks. It is a good study, not a final answer.
It was tested on people with anxiety. Nothing here says what happens to someone who sleeps badly and is not anxious. If anxiety is the whole road from the capsule to the sleep, then a person without anxiety has no road.
One of the co-authors works for the company that makes the preparation. That does not make the numbers wrong, and it is worth knowing.
The bag in the drawer is still just a bag in the drawer. What was tested was a cleaned, standardised extract, sealed in a capsule and taken daily for ten weeks at a measured dose. Nobody tested the sachet.

So the folk advice turned out to be half right, and not for the reason people assume. Something in that plant really does help people sleep. But only after it has been purified, standardised and swallowed every day for ten weeks. And only because it takes the anxiety away first.

Source: Seifritz E, Schläfke S, Holsboer-Trachsler E. Beneficial effects of Silexan on sleep are mediated by its anxiolytic effect. Journal of Psychiatric Research, 2019;115:69–74. PMID 31121394