Ashwagandha: the plant behind the stress and sleep claims

Ashwagandha has a way of turning up on every shelf you browse: capsules for stress, powders for sleep, gummies for "balance", each with a slightly different plant name on the back. Before one of them goes in the basket, it's worth knowing what the plant is, what the trials actually tested, and why Denmark, France and the Netherlands have all had something to say about it.

Does it help, and what should you know first?

A bit of both. Small human trials, mostly short and mostly built around a handful of branded root extracts, report modest improvements in self-rated stress, anxiety and sleep. The same plant has also been linked to rare but serious liver injury in published case reports. France's ANSES advises specific groups to avoid it, while the Dutch institute RIVM advises everyone against it, pregnancy above all. Denmark does not allow it to be sold as food, and the Dutch rules are still being worked out (the Netherlands section below has the dates).

Sleep is the part of the evidence people ask about most, so the trials come first. Before that, the plant itself, because "ashwagandha" on a label covers more than one thing.

The plant, the parts and the extracts

Ashwagandha is Withania somnifera (L.) Dunal, a member of the nightshade family. RIVM, the Dutch national institute for public health, describes more than 3,000 years of use in Ayurvedic medicine (RIVM report 2024-0029). The species name "somnifera" is Latin for sleep-inducing, which makes a lovely label and settles nothing: a plant name is not a trial result.

What matters for supplements is which part of the plant is in the capsule and how it was processed:

• Root powder is ground root, with whatever the root happens to contain.
• Root extract is a concentrate, often standardised to a percentage of withanolides, the plant's steroid-like compounds. Doses in trials are almost always extract doses.
• Root and leaf extracts exist too. The Shoden extract used in several trials is described in the sleep meta-analysis below as an extract of both root and leaf (Cheah et al. 2021). KSM-66, used in most of the sleep trials, is described by its maker as a root-only extract (Langade et al. 2019).

The parts differ chemically. France's food safety agency ANSES notes large differences in withanolide and alkaloid content between different organs of the plant, and that supplements said to contain root have, in the literature, turned out to contain leaf or other plants (ANSES opinion, 19 April 2024, translated from French). So "ashwagandha" on a label tells you less than it seems to.

"Adaptogen": a word with a Soviet pedigree and no EU definition

Nearly every product page calls ashwagandha an adaptogen. Where does the word come from?

The term goes back to the Soviet pharmacologist N. V. Lazarev in 1947, who used it for substances claimed to raise "non-specific" resistance to stress. His colleague Israel Brekhman later set out the classic criteria: an adaptogen is almost non-toxic, works non-specifically against a wide range of stressors, and has a normalising effect on body systems. That history comes from the European Medicines Agency's herbal committee, which published a reflection paper on the adaptogenic concept in 2008 (EMA reflection paper EMEA/HMPC/102655/2007).

The same paper adds that the principle of adaptogenic action needs further clarification, and that "the term is not accepted in pharmacological and clinical terminology that is commonly used in the EU." It also notes that the molecular targets of adaptogens are still unknown. Its example herbs were Eleutherococcus, ginseng, schisandra and rhodiola; ashwagandha was not among those named.

So "adaptogen" works as a browsing category: it groups herbs sold for stress and tiredness. It is not a regulatory status and it doesn't mean a herb has been shown safe. The first classic criterion, "almost non-toxic", is exactly what the safety debate below puts to the test.

What the trials tested: five studies, a few extracts

Most human evidence for ashwagandha and sleep comes from a small family of trials. The table lists the ones that matter most, with the preparation and who paid or supplied it.

• Trial: Langade et al. 2019, Cureus; Who took part: 60 adults with insomnia and anxiety (2:1 to extract or placebo); Extract and dose: KSM-66 root extract, 300 mg twice daily; Length: 10 weeks; What was reported: Time to fall asleep on actigraphy was about 29 versus 34 minutes at week 10 (placebo improved too); better sleep efficiency and sleep quality ratings; Funding or supply: Extract supplied by its manufacturer; authors declared no competing interests
• Trial: Langade et al. 2021, J Ethnopharmacol; Who took part: 80 adults, half healthy and half with insomnia; Extract and dose: KSM-66 root extract, listed as 600 mg a day in a later meta-analysis; Length: 8 weeks; What was reported: Better sleep onset and sleep scores; larger gains in the insomnia group; Funding or supply: Funding not stated in the abstract
• Trial: Deshpande et al. 2020, Sleep Med; Who took part: 150 healthy adults with non-restorative sleep (144 finished); Extract and dose: Shoden extract, 120 mg once daily; Length: 6 weeks; What was reported: Self-reported sleep quality up 72% versus 29% with placebo; Funding or supply: A later meta-analysis lists it as partly sponsored by the manufacturer
• Trial: Lopresti et al. 2019, Medicine; Who took part: 60 healthy adults with high self-reported stress; Extract and dose: Shoden extract, 240 mg once daily; Length: 60 days; What was reported: Anxiety score (HAM-A) lower than placebo (p = 0.040); a combined mood scale (depression, anxiety and stress) was close but not significant; Funding or supply: Funded by Arjuna Natural Ltd; one author had past funding from the company
• Trial: Movva et al. 2026, Clocks & Sleep; Who took part: 200 adults aged 18 to 50, four groups; Extract and dose: KSM-66 root extract 300 mg twice daily, melatonin 3 mg, both, or placebo; Length: 8 weeks; What was reported: Sleep onset time fell significantly in the treatment groups; the combination was the most consistent; Funding or supply: No external funding; KSM-66 sourced from its maker

Read side by side, a few things stand out.

The sleep effect is modest, and placebos do real work. In the 2019 insomnia trial, sleep onset improved in both groups, and the gap between them was a few minutes. That's worth knowing, and it's not a transformation.

The trials are small and cluster together. A 2021 meta-analysis of five placebo-controlled sleep trials found what its authors call a small but significant overall effect (standardised mean difference −0.59, 95% CI −0.75 to −0.42), stronger for people with insomnia, doses of 600 mg a day or more and treatment of eight weeks or longer. All five trials were run in India, four used KSM-66 and one used Shoden, and the results were quite uneven between trials (I² = 62%) (Cheah et al. 2021, PLOS ONE). A 2024 meta-analysis of five trials in people with anxiety or insomnia (254 participants) found better scores on anxiety and several sleep measures, but with extremely high heterogeneity (I² = 98% for anxiety), and the authors themselves call for larger studies (Fatima et al. 2024, Hum Psychopharmacol).

Stress and anxiety have a larger evidence pool than sleep. A 2024 meta-analysis of nine trials (558 people, searches to January 2022) reported lower perceived stress, lower anxiety scores and lower serum cortisol than placebo, with mild to moderate adverse events in four trials (Arumugam et al. 2024, Explore). A 2025 meta-analysis of 14 trials in people with diagnosed mental disorders (360 on ashwagandha, 353 controls) found improvements in anxiety, stress and sleep quality, but flagged a small number of studies, large differences between them and outlier effects. It found no significant difference in safety or tolerability against comparators (Marchi et al. 2025, BJPsych Open). The people in these trials were not typical shoppers, and the reviews don't say ashwagandha treats a medical condition.

Every dose is a dose of a specific extract. "600 mg" on a capsule could mean 600 mg of a standardised extract, or 600 mg of root powder, which is a different thing. The trials say nothing about how an unstandardised powder behaves.

Hormones move too. A 2026 meta-analysis of 23 trials (1,706 people) found lower cortisol with ashwagandha, a small rise in thyroxine (T4) but no change in TSH or T3, and higher testosterone in men but not in women (Fornalik et al. 2026, Planta Medica). Those are real biological effects, which is also why anyone with a thyroid or hormone condition should be careful.

What the shelf label skips: liver, thyroid and who should sit this out

The trials and the case reports tell different stories, and both are real.

Trials. A 2026 systematic review of 23 randomised trials of standardised, single-ingredient root extracts in healthy adults (2,317 participants, 125 to 600 mg a day, up to 180 days) reported liver, kidney, blood and hormone markers staying within normal ranges and no serious adverse events attributed to the extract. Two of its authors disclosed consulting for the supplement industry, and the review itself notes there are few data beyond 180 days (Coope et al. 2026, Pharmaceuticals).

Case reports. Outside the trials, doctors have published cases of liver injury after ashwagandha products:

• An Icelandic and US series of five patients with jaundice and itching after 2 to 12 weeks of use. Liver tests normalised within one to five months in four of them. One was lost to follow-up (Björnsson et al. 2020, Liver Int).
• An Indian series of eight patients with single-ingredient products. Five already had chronic liver disease. Three of those developed acute-on-chronic liver failure and died. Chemical analysis of the products found no adulteration or contamination (Philips et al. 2023, Hepatol Commun).
• A 2026 scoping review of 13 publications covering 25 patients described a pattern of delayed, often cholestatic liver injury that usually recovered after stopping. One patient needed a liver transplant, and three people who already had cirrhosis died after their liver function worsened (McIntyre et al. 2026, Cureus).

Case reports can't show how often this happens, or prove cause. The Dutch RIVM assessment makes the same point: it counted nine published case reports (seven liver, one thyrotoxicosis, one adrenal suppression) and four notifications of liver injury to the Dutch pharmacovigilance centre Lareb. It judged that most people are probably not at risk, because the clinical trials did not show these problems at comparable doses and durations. But no one knows who the sensitive individuals are, and the authors could not set a safe intake level (RIVM 2024).

Specific groups. Here is what the agencies say, in their own terms:

• Pregnancy and breastfeeding. RIVM advises everyone against use, especially in pregnancy, noting that the herb has been traditionally used to induce abortion and that no suitable reproductive studies cover early pregnancy (RIVM). ANSES advises against it in pregnancy, while breastfeeding, and under 18.
• Thyroid. ANSES and RIVM both list thyroid effects. Memorial Sloan Kettering's herb monograph notes several case reports of thyrotoxicosis (MSK).
• Liver and heart conditions. ANSES advises people with liver or heart conditions, hyperandrogenism (high androgen levels), or hormone disorders to avoid ashwagandha supplements.
• Sedatives. ANSES notes drowsiness in clinical trials and advises against combining it with central nervous system depressants, or using it before driving or operating machinery. MSK advises that people on benzodiazepines, anticonvulsants or barbiturates likely avoid it.
• Immunosuppressants and other medicines. We found no direct interaction study with immunosuppressive drugs. RIVM reports a larger immune response in some trials and animal immune effects, and MSK says ashwagandha may trigger or worsen autoimmune disease. ANSES describes in-vitro interaction signals as heterogeneous but enough to call for caution. RIVM, citing a drug-information database, also mentions diabetes drugs and thyroid hormones. Anyone on prescription medicine should check with a pharmacist or doctor; our guide to supplement and drug interactions explains how.

Denmark said no; the Netherlands is mid-argument

Rules for botanicals in food supplements are not harmonised across the EU, as ANSES points out, so ashwagandha is treated differently depending on where you live.

Denmark. The Danish Veterinary and Food Administration states that it is illegal to sell ashwagandha as food in Denmark, including the root, and the ban covers supplements and similar products. It relies on a 2020 risk assessment from the Technical University of Denmark (DTU) (Fødevarestyrelsen page, read 8 October 2026). According to RIVM and ANSES, DTU concluded that, from the available data, it could not establish an intake without risk: it pointed to effects on thyroid and sex hormones in animals and humans. The DTU report is in Danish, so we're relaying it as RIVM and ANSES summarise it. Press coverage dates the Danish ban to April 2023, and industry groups have criticised the assessment.

France and Germany. ANSES (opinion dated 19 April 2024) did not ban it. It could not define a risk-free intake either, and it advises the groups above to abstain. It also records that Denmark considers the plant should not be used in food, that Poland sets a maximum dose for the root, and that Germany's BfR recommended putting the root on the EU's list of substances under scrutiny (list C of Regulation 1925/2006).

EFSA. We found no EFSA safety assessment of ashwagandha and no authorised EU health claim for it. EFSA's evaluation of botanical health claims has been on hold since 2010 (NutritionInsight, June 2023).

The Netherlands. Here is the situation as of 8 October 2026, in the order it unfolded:

• March 2024. RIVM, commissioned by the health ministry (VWS), advised against using herbal preparations containing Withania somnifera, and above all in pregnancy (RIVM news, 5 March 2024). The report found that supplement doses on the Dutch market were in the same range as doses in the case reports.
• Today. Ashwagandha is not on the banned or restricted list in the annex to the Dutch Warenwetbesluit Kruidenpreparaten (version in force since 1 July 2022, read 8 October 2026). Products are sold.
• Early 2025. The ministry put a draft regulation for food supplements and herbal preparations out for consultation, a draft that rests on RIVM's assessments, including the one on Withania (Food Law Consult, April 2025).
• March 2026. The Dutch consultancy Nutri-Sana reported that the ministry had decided to ban ashwagandha from 1 January 2027, with a sales deadline of 1 July 2027. The same article said the industry association NPN had shared new safety data and that RIVM would assess it (Nutri-Sana, 25 March 2026).
• 17 September 2026. Nutri-Sana then reported that VWS had adjusted its proposal. The root would stay allowed "for now", a ban on other plant parts was still foreseen, and VWS had asked RIVM to reassess the root. In the meantime a mandatory warning was proposed for root products: not for use in pregnancy, breastfeeding or with liver or thyroid problems, plus a note that ashwagandha can in very rare cases cause liver problems. The article added that the change to the rules was not yet in force, and that VWS expected to decide on measures again towards the end of 2027 (Nutri-Sana, 17 September 2026).
• 6 October 2026. NutraIngredients reported that the Netherlands had notified the draft to the EU, with comments open until 7 December 2026. If adopted, the rules could apply as early as the end of 2026, followed by a six-month transition. It says the proposal would ban ashwagandha, and the part of the article we saw mentions no exception for the root (NutraIngredients).

The September and October reports pull in different directions, and both come from trade outlets. We found no ministry text or EU notification record to settle which one reflects the notified draft, so the final shape is unknown. Check the Rijksoverheid and the EU's TRIS notification database for current rules; this is a dated snapshot, not legal advice.

Reading an ashwagandha label without a degree in botany

A worked example, with no real product involved. Imagine two labels:

• Label A: "Ashwagandha root extract 300 mg (5% withanolides), 2 capsules daily."
• Label B: "Ashwagandha 600 mg per serving."

Label A names the part (root), says it is an extract and gives a standardisation. You can compare it to a trial: 300 mg twice a day of a root extract is the dose in the KSM-66 sleep trials above, although the trials used a specific branded extract, so a similar label is not the same product. Label B doesn't say whether it is powder or extract, which plant part, or whether there is any standardisation. The RIVM report shows how wide the range is: the supplements it reviewed gave between 50 mg and 2 g of extract a day, or 1.3 to 2.5 g of root powder.

Questions worth asking of any label:

1. Which part? Root, leaf or both. ANSES has flagged leaf adulteration of root products, and root and leaf are different chemically.
2. Extract or powder? And if extract, what is it standardised to?
3. Is the dose per day, per capsule, or of raw root equivalent?
4. Does the product repeat the trial it cites? Studies of one extract do not carry over to another.
5. Does the pack carry a warning for pregnancy, thyroid or liver conditions? Not every pack does yet.

Our guide to reading supplement labels covers the general method.

So, nightcap or no nightcap?

Ashwagandha has a better evidence base for stress and sleep than many herbs in its category, and a more serious safety file than most of them. The trials are small and short, much of the work was supplied or funded by extract makers, and meta-analyses agree on direction but not on size. For most healthy adults in trial conditions the picture is probably benign. A small number of people, nobody knows which, have developed liver injury, thyroid effects are plausible, and pregnancy is a clear "avoid" from RIVM and ANSES alike.

If sleep is the problem, behaviour matters at least as much as botanicals, and our sleep hygiene guide is a good place to begin. A long-running sleep problem is worth taking to a doctor, ideally before adding a supplement.

Where to read next

Ashwagandha has plenty of company in the sleep and adaptogen aisles:

• Sleep supplements: ingredients before the promise, for what the label lists, what studies say each ingredient does for sleep and what NL and EU rules allow.
• Sleep herbs explained, for what the research on valerian, passionflower, chamomile and lavender covers and where the cautions sit.
• Melatonin: timing, labels and the questions to ask, for the amounts and timings that studies and regulators report and what to ask of a label.
• Reishi and sleep: what the research shows, for reishi's calming reputation set against the evidence.
• Functional mushrooms and adaptogens: guide hub, for the six popular ingredients, each with its own guide, and a place to start.

Written by Smart Supplements with AI assistance from the sources linked in the text. It has not been reviewed by a clinician.

Last updated: 12 October 2026

This article is for informational purposes only and is not intended as medical advice. Always consult a healthcare professional before starting any new supplement, especially if you take prescription medication.