Ashwagandha Benefits and Side Effects, Trial by Trial

Ashwagandha, botanically Withania somnifera, is a root used in Ayurveda and sold across Europe as a food supplement. The ashwagandha benefits reported in published trials are narrow and specific: lower scores on stress questionnaires, lower serum cortisol, shorter sleep onset latency, larger strength gains during resistance training. None of those is an authorised health claim under Regulation 1924/2006. Four national food safety bodies have assessed the plant since 2020 and none could set a safe intake level. In Denmark it is unlawful in food.

That last sentence is the part most benefit pages leave out, so this one starts there. Below you get what each trial measured, in whom, at what dose and for how long; the European legal position country by country with dates; the side effects that are actually documented; the liver case reports that produced them; and the interactions and exclusions that decide whether this belongs in your cupboard at all.

What ashwagandha is, and what a withanolide figure means

The ashwagandha plant, Withania somnifera, is a shrub in the nightshade family. The root is the part used in classical Ayurvedic preparation, and root extract is what almost every European supplement claims to contain. In German it is Schlafbeere, in older English sources winter cherry.

Its two main groups of secondary metabolites are steroidal lactones called withanolides, and tropane and piperidine alkaloids. ANSES notes in its 2024 opinion that the amounts of both vary enormously between plant organs, qualitatively as well as quantitatively, which is why linking any single compound to any reported effect is still not possible.

That variability has a practical consequence. ANSES records published cases of supplements sold as root or root extract being adulterated with leaf material from the same plant, which is alkaloid-rich, or with other species entirely. The BfR makes the same point from the other direction: because plant part and extraction method both change the composition, it is not currently possible to name the substances and doses responsible for the effects reported in studies.

So a withanolide percentage on a label tells you what was standardised, not what the rest of the extract contains. Poland is the only European country that has fixed numbers to it: a 2020 resolution caps the root at 3g per 24 hours and withanolides at 10mg per recommended daily portion, and requires the marketer to hold a specification proving the withanolide content. The European Medicines Agency reached a different conclusion again. Its herbal committee could not establish a monograph for Withaniae somniferae radix, because the extract specifications in the available data were judged insufficient against pharmaceutical requirements.

If you are choosing a product, the extract specification is the thing to ask about, not the milligram number on the front. Dose and standardisation get their own treatment in ashwagandha dosage and timing.

How we read a claim about ashwagandha

We are a shop, not a laboratory and not a clinic. What we can do is apply the same seven checks to every study before repeating anything from it, and tell you when a study fails them.

Population. Who was actually enrolled. Most ashwagandha trials recruit adults already reporting stress or insomnia, at a single centre, and almost all are run in India. RIVM says plainly that it is unknown whether those results apply to healthy people or to people from other regions.

Outcome measured. A questionnaire score is not a clinical outcome. Perceived Stress Scale, Hamilton Anxiety, Pittsburgh Sleep Quality Index and the Menopause Rating Scale are self-reported instruments. Serum cortisol and sleep onset latency are objective. We name which was used.

Dose and preparation. The trials cluster on 300mg twice daily or 600mg once daily of a root extract, but the extracts differ between trials and are rarely characterised well enough to compare.

Duration. Eight weeks is the modal trial length and sixty days the next most common. Harvard Health notes that data beyond three months are limited, and the European assessments say the same about long-term use.

Independence and size. Sample sizes of 50 to 100 are typical, several trials share authors, and several were sponsored by extract manufacturers. Small, related, single-country trials support a signal, not a conclusion.

Whether harms were collected. This is the check that matters most here. The BfR states that the human studies were designed to look for benefits and did not systematically record side effects. RIVM confirms that in none of the 28 trials it reviewed were adverse effects statistically different from placebo.

What a regulator concluded. A national risk assessment reads the same literature with a different question. Four have now been published and they agree with each other more than they agree with the marketing.

Where ashwagandha stands in Europe, country by country

The positions below are taken from the assessments themselves, not from summaries of them.

Country	Position on ashwagandha in food supplements	Assessment behind it	Date
Denmark	Unlawful to sell as food, supplements included	DTU: no threshold identifiable below which the root is free of risk	2020 assessment
Germany	Legal, but the federal risk body advises against taking it	BfR Mitteilung 039/2024, following its 2012 assessment	10 September 2024
France	Legal and listed without restriction, with an official caution	ANSES opinion, saisine 2021-SA-0077	19 April 2024
Netherlands	Legal, RIVM advises the whole population not to use it	RIVM letter report 2024-0029	2024
Poland	Legal with binding limits and label warnings	National resolution: 3g root per 24 hours, 10mg withanolides	2020
Italy and Belgium	Permitted, whole plant, no specific restriction	National plant lists, Belgium requires product notification	2017 and 2021 lists
United Kingdom	Legal, under active review	FSA call for evidence, referred to the Committee on Toxicity	July to September 2024
Switzerland	Unresolved, see the note below	No Swiss assessment found	not applicable
European Union	Recommended for an Article 8 procedure under Regulation 1925/2006	Heads of Food Safety Agencies working group, first report	6 June 2024

Three of those rows need expanding.

The Danish position is the hardest. The Danish Veterinary and Food Administration states that products containing ashwagandha are unlawful to sell as food in Denmark, and points consumers away from buying them online from other markets. The DTU assessment behind it cites effects on sex hormones and reproduction, on metabolism, on immune function and on the central nervous system, and concluded that no health-based guidance value could be derived.

The German position is not a ban and is easy to misread as one. The BfR advises children, pregnant and breastfeeding women, and anyone with current or previous liver disease not to take ashwagandha preparations, and advises the rest of the general population to be restrained about it. Its stated reason is not proof of harm but the absence of data: it could not derive a safe intake amount either. The BfR had already recommended in 2012 that the root be added to Part C of Annex III of Regulation 1925/2006, the list of substances under European scrutiny.

The European row is the one to watch. The Heads of Food Safety Agencies working group on food supplements, in its first report dated 6 June 2024, put Withania somnifera among thirteen prioritised substances and grouped it with those showing possible carcinogenic, mutagenic or reprotoxic properties. An Article 8 procedure can end in prohibition, in restriction, or in four years of Union scrutiny. It has not been executed.

What this means for a European seller is blunt. Denmark is currently closed to any ashwagandha product, whatever the country of dispatch. Poland is open only with the dose ceiling and the label warnings. Everywhere else in the table is open today and under review.

The liver case reports that produced all of this

The serious side effects attributed to ashwagandha are few, published, and worth reading in full. Every one of the four assessments turns on the same small body of case reports, so it is worth reading them rather than the headline.

The core publication is Björnsson and colleagues (2020) in Liver International, a series of five patients with liver injury attributed to ashwagandha-containing supplements. Three were collected in Iceland during 2017 and 2018 and two came from the US Drug-Induced Liver Injury Network in 2016. Three were male, mean age 43, range 21 to 62. All five developed jaundice, with nausea, lethargy, itching and abdominal discomfort, after a latency of 2 to 12 weeks. The injury was cholestatic or mixed. Itching and raised bilirubin lasted 5 to 20 weeks.

Now the parts that get dropped. No patient developed hepatic failure. Liver tests normalised within one to five months in four of them, and one was lost to follow-up. Chemical analysis confirmed ashwagandha in the available supplements and found no other toxic compound. None of the five was taking a prescription medicine known to injure the liver. Four were taking additional supplements, and in one case rhodiola was a possible cause alongside ashwagandha.

RIVM pooled the wider literature and found nine case reports in total: seven of liver injury, one of thyrotoxicosis and one of suppressed adrenal function, at daily exposures of 77 to 1350mg of extract. The authors of those reports judged the role of the plant definite in one case, highly likely in two, probable in two and possible in one; in three the likelihood was not formally scored. The Netherlands Pharmacovigilance Centre separately logged four notifications of liver injury.

The confounder RIVM could not resolve is the interesting one. The doses in the case reports overlap the doses used in clinical trials, and the trial durations overlap the time to onset, yet no trial detected these injuries. RIVM checked whether the supplements were impure: in five liver cases the products were analysed and no known toxic compound or trace metal was found, and only one case had a second ingredient that might plausibly have been involved. Its conclusion is that some individuals are unusually sensitive, and that nobody knows which individuals those are.

Set against the volume of ashwagandha sold in Europe, this is rare. It is also unpredictable, not screenable in advance, and serious enough that four regulators changed position over it. Both halves of that are true and this page will not print one without the other.

What the trials actually measured

People ask what ashwagandha does to your body. The honest answer is that in trials it moved these specific numbers, in these specific people, over these specific weeks.

Stress and cortisol

Chandrasekhar and colleagues (2012) randomised 64 adults with a history of chronic stress to 300mg of a full-spectrum root extract twice daily or placebo for 60 days, and reported lower scores on the stress scales and lower serum cortisol in the extract group. It is a single-centre trial of 64 people and it establishes nothing about anyone outside it.

Akhgarjand and colleagues (2022) pooled 12 randomised trials with 1,002 participants in Phytotherapy Research and found lower anxiety and stress scores against placebo. Two figures from that paper matter more than the effect size: heterogeneity was 93.8 per cent for anxiety, and the authors rated the certainty of the evidence as low for both outcomes. That is a signal in need of better trials, not a settled result.

RIVM adds a framing the marketing never uses. It notes that reduced cortisol and raised thyroid hormone levels were treated as positive findings by the study authors, but that from a toxicological perspective they are considered adverse.

Sleep

Langade and colleagues (2021), in the Journal of Ethnopharmacology, ran an eight-week trial in 80 people, 40 healthy and 40 with insomnia, and reported improved sleep onset latency and sleep efficiency, with larger changes in the insomnia group. Cheah and colleagues (2021) pooled five trials with 400 participants in PLOS ONE and found a small but statistically significant effect on overall sleep, more prominent in diagnosed insomnia, at doses of 600mg per day or more and durations of eight weeks or more. Their own conclusion notes that safety data are limited and that long-term use is unassessed.

Exercise and body composition

Wankhede and colleagues (2015), in the Journal of the International Society of Sports Nutrition, gave 57 untrained men aged 18 to 50 either 300mg of root extract twice daily or a starch placebo alongside eight weeks of resistance training. The extract group gained more on bench press one-rep max, 46.0kg against 26.4kg, and on leg extension, 14.5kg against 9.8kg, plus more arm and chest size, a larger rise in serum testosterone and a larger fall in body fat percentage. Eight weeks, 57 beginners, one centre, and no follow-up.

On body weight, which is a common German-language query, Choudhary and colleagues (2017) gave 52 adults under chronic stress 300mg twice daily for eight weeks and reported changes in perceived stress, food craving scores, body weight and body mass index. Fifty-two people, eight weeks, largely self-reported eating measures.

Trials in women

Two pilots exist and both are worth naming precisely. Dongre and colleagues (2015) randomised 50 healthy women to 300mg twice daily or placebo for eight weeks and reported higher Female Sexual Function Index scores. Gopal and colleagues (2021), in the Journal of Obstetrics and Gynaecology Research, gave 100 perimenopausal women the same dose for eight weeks and reported lower Menopause Rating Scale scores, along with higher serum estradiol and lower FSH and LH.

Both are single-centre Indian pilots, and the endocrine movement in the second is exactly the kind of finding the regulators flag rather than celebrate. What neither trial did is enrol women who were pregnant, breastfeeding, or being treated for a thyroid condition. That is not an oversight. It is the exclusion list.

Interactions, and the thyroid question

The thyroid interaction is a genuine contraindication rather than a marketing caveat, and it has evidence on both sides of it.

Sharma and colleagues (2018) gave 600mg of root extract daily for eight weeks to 50 people with subclinical hypothyroidism, defined by a TSH of 4.5 to 10 microIU per litre, and reported lower TSH with higher T3 and T4 against placebo. Read that as a demonstration that the extract moves thyroid hormones, because it has a mirror image. Van der Hooft and colleagues (2005) reported a 32-year-old woman in the Netherlands who developed thyrotoxicosis while taking ashwagandha capsules for fatigue, on no other medication; her symptoms and laboratory values resolved after she stopped.

An ingredient that shifts thyroid indices in one direction in one population can shift them in a direction you did not want in another. That is why every European assessment names thyroid disease in its exclusion list.

Combination	What the assessments say	Source
Thyroid medication, or thyroid disease	Abstain. Endocrine effects including thyrotoxicosis are the named reason	ANSES 2024, BfR 2024
Sedatives, sleeping tablets, antiepileptics	Avoid. Poland requires this as a label warning	ANSES 2024, Poland 2020
Alcohol, driving, operating machinery	ANSES advises against use here and alongside any central nervous system depressant	ANSES 2024
Immunosuppressants	Seek medical advice before combining	BfR 2024
Antidiabetic medicines	Seek medical advice first, effects on blood sugar are reported	BfR 2024
Antihypertensive medicines	Seek medical advice before combining	BfR 2024

One question that recurs in German-language search is whether ashwagandha suits people with rheumatic disease. No trial establishes anything for rheumatic conditions, and because the assessments flag effects on immune function and name immunosuppressants as an interaction, this is a conversation for a rheumatologist rather than for a supplement page.

What is not established, and who should not take it

No safe intake level exists for ashwagandha in Europe. Not one of the four assessments could derive a health-based guidance value, and Poland’s 3g and 10mg figures are a risk management decision, not a toxicological threshold.

No health claim for ashwagandha is authorised in the European Union. Claims for botanicals have sat on hold in the Commission register since 2010, which means anyone printing a benefit on a label is doing so without authorisation rather than with it.

Long-term use is unstudied. The modal trial is eight weeks. Nobody has published on what daily use for a year does.

Do not take ashwagandha if you are pregnant or breastfeeding. ANSES cites a traditional use as an abortifacient and RIVM says the risk to an unborn child cannot be adequately assessed. Do not give it to anyone under eighteen. Do not take it if you have thyroid, liver, cardiac or other endocrine disease, or a history of liver disease. If you take prescribed medication of any kind, speak to a doctor or pharmacist before starting, and stop and seek medical advice if you develop nausea, unusual tiredness, itching, or yellowing of the skin or eyes.

Our own ashwagandha and shilajit capsules carry 250mg of ashwagandha alongside 250mg of shilajit extract per capsule, and every line above applies to them exactly as it applies to any other ashwagandha product on the market.

What actually decides this

Three things decide it, and none of them is a benefit list. First, where you live: in Denmark the question is closed, in Poland it comes with a dose ceiling, everywhere else in Europe it is legal and under review. Second, whether you are on the exclusion list, which is longer than most pages admit and includes conditions people do not think of as relevant. Third, whether a small change in a questionnaire score, from mostly eight-week single-country trials, is worth an unquantified and unpredictable liver risk to you personally. That is a judgement, and it is yours.

If your honest answer is no, it costs nothing to skip. If it is yes, buy something with a stated extract specification, start at the low end of the range, and tell your doctor you are taking it. For how the evidence for shilajit compares, which is a different substance with a different literature, start with shilajit benefits, and for where ashwagandha sits among the other adaptogens by evidence grade, what are adaptogens sorts them.