Ashwagandha: The Real Evidence for Stress and Cortisol — Plus the Liver Risk Nobody’s Talking About

Ashwagandha (Withania somnifera) is a nightshade-family root used in Ayurvedic medicine for centuries and now one of the best-selling herbal supplements in the world, marketed as a fix for stress, sleep, testosterone, and gym performance. The strongest independent human evidence — a 2025 meta-analysis of 15 randomized trials — supports a real but modest reduction in perceived stress and cortisol at 300–600 mg/day of standardized extract over 8 weeks; sleep evidence is similarly real but modest; testosterone and strength claims rest almost entirely on trials funded by the branded-extract manufacturers whose product was being tested. The part most articles skip: ashwagandha has accumulated enough liver-injury case reports since 2017 to earn a "likely hepatotoxic" classification from the U.S. National Institutes of Health, has been banned as a food-supplement ingredient in Denmark since 2023, and carries a real, case-report-documented risk of thyroid hormone spikes and serotonin syndrome when combined with common antidepressants. This article grades every major claim against independent, funding-traced, human-only evidence — see the full breakdown and sourcing at the NIH LiverTox entry on Withania somnifera.

Table of contents

• Evidence summary
• What Ashwagandha is
• All forms and types of Ashwagandha
• How Ashwagandha works
• What works and what does not
• Benefits with evidence grades
• Risks and all side effects
• All interactions
• Who should avoid Ashwagandha
• Dosage and how to take Ashwagandha
• Infographics
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

What Ashwagandha is

Withania somnifera, commonly called ashwagandha, Indian ginseng, or winter cherry, is a small shrub in the nightshade (Solanaceae) family native to India, the Middle East, and parts of Africa. Its root — and, in some modern extracts, its leaf — has been used in Ayurvedic medicine as a "rasayana" (rejuvenative tonic) for centuries, traditionally prescribed for weakness, low libido, insomnia, and general vitality. Today it is one of the top-selling herbal supplements in the United States and Europe, marketed heavily under an "adaptogen" umbrella term that implies it helps the body resist stress of any kind (NIH Office of Dietary Supplements).

The root contains a class of steroidal lactones called withanolides — withaferin A and withanolide A are the most studied — along with alkaloids, saponins, and other phytochemicals. Withanolide content is the primary lever manufacturers use to differentiate "standardized" extracts from raw root powder, and it ranges from under 1% in plain root powder to over 30% in the most concentrated proprietary extracts (NIH ODS).

Modern clinical interest in ashwagandha concentrates on four areas: stress and cortisol regulation, sleep quality, anxiety symptoms, and — more speculatively and more commercially — testosterone, athletic performance, and cognition. Independent human-trial evidence is meaningfully stronger for the first two than the last two, a distinction this article treats as central rather than a footnote.

All forms and grades

Ashwagandha supplements are not interchangeable just because the label says the same milligram number. The active-compound concentration, plant part used, and extraction method all differ by brand, and almost all clinical trial evidence attaches to a specific named extract rather than "ashwagandha" as a generic ingredient.

Because withanolide content and plant part vary so much between products, a dose comparison only means something within the same named extract. A 600 mg dose of 1.5%-withanolide root powder delivers roughly 9 mg of withanolides, while 600 mg of a 35%-withanolide extract delivers roughly 210 mg — a 20-fold difference in the compound class assumed to drive the effect (NIH ODS).

How it works

The leading human-relevant hypothesis is that ashwagandha modulates the hypothalamic-pituitary-adrenal (HPA) axis, the hormonal signaling loop that governs the body's cortisol stress response. Human trials consistently measure lower morning and/or overall serum cortisol after ashwagandha supplementation compared with placebo, which is consistent with HPA-axis dampening (Bachour et al. 2025; Salve et al. 2019).

A second proposed mechanism — that withanolides act on GABA-A receptors similarly to benzodiazepine-class sedatives, producing an anti-anxiety and sleep-promoting effect — is largely a preclinical and mechanistic story. Independent human-trial evidence does not directly confirm GABA-receptor binding in people; this mechanism is included here for context only and should not be read as proven in humans. Per the evidence standard this article follows, mechanistic claims resting on non-human data are flagged rather than presented as settled science.

A third proposed mechanism relevant to the reproductive-hormone claims involves modest changes in luteinizing hormone (LH) and dehydroepiandrosterone sulfate (DHEA-S), both measured directly in small human RCTs (Ambiye et al. 2013; Lopresti et al. 2019). The magnitude and reproducibility of this pathway outside small, often industry-funded trials remains unclear.

Absorption and bioavailability data in humans are thin. Most trials dose orally, once or twice daily, for 8–12 weeks — the duration threshold above which the sleep-quality evidence becomes more consistent (Cheah et al. 2021). There is no independent human pharmacokinetic dataset establishing a validated bioavailability percentage for any of the major branded extracts, which makes cross-brand mg-for-mg comparisons unreliable beyond the withanolide-percentage math above.

Why "adaptogen" is a marketing category, not a clinical diagnosis

The term "adaptogen" originated in mid-20th-century Soviet pharmacology to describe substances proposed to non-specifically increase resistance to stress across biological, chemical, and physical stressors. It has never been formally recognized as a pharmacological or regulatory category by the FDA, EMA, or any major drug regulator — there is no standardized test a substance must pass to be labeled an adaptogen, and no dose-response relationship required to be established for the label to be applied. This matters for ashwagandha specifically because "adaptogen" branding often implies a broader, vaguer benefit ("helps your body handle stress in general") than what the actual RCT evidence supports (a specific, measurable reduction in cortisol and perceived stress at specific doses and durations). Readers should treat "adaptogen" as a category label taken from traditional and early Soviet-era pharmacology, not as a stand-in for "clinically proven to reduce stress broadly."

How ashwagandha compares with other stress-focused options with independent human evidence

Ashwagandha is not the only supplement with independent human-trial support for stress and sleep outcomes, and it is worth situating it against the two most commonly compared alternatives.

Magnesium. Independent human trials support magnesium supplementation for improving sleep quality in people with documented magnesium insufficiency or age-related sleep changes, and for reducing subjective anxiety in some populations, though the anxiety evidence base is smaller and more heterogeneous than ashwagandha's. Unlike ashwagandha, magnesium does not carry a documented liver-injury signal, though very high doses can cause diarrhea and, in people with kidney impairment, dangerous magnesium accumulation.

L-theanine. Independent human RCTs support L-theanine (an amino acid found in tea) for reducing subjective stress and anxiety, often with a faster onset (single-dose effects have been measured) than ashwagandha's multi-week dosing requirement. L-theanine's independent human safety record is also cleaner — no comparable liver-injury or thyroid-disruption signal has emerged in the pharmacovigilance literature at typical doses.

Neither comparison is meant to declare a universal "winner" — the point is that ashwagandha's cortisol/stress evidence is real but sits alongside other options with arguably cleaner independent safety profiles, and a reader whose primary goal is stress relief without the liver/thyroid risk profile has legitimate independently supported alternatives to weigh.

What the withanolide-content variability actually means for a shopper

Because U.S. dietary supplements are not required to undergo FDA premarket testing for potency or purity, third-party testing organizations have periodically found inconsistencies between labeled and actual withanolide content in commercial ashwagandha products — a general dietary-supplement industry problem, not unique to ashwagandha, but one that compounds the difficulty of matching a store-bought product to the specific extract and dose used in a given clinical trial. Readers who want their purchase to actually resemble what was tested should look for three things on the label: the specific named extract (e.g., "KSM-66," "Sensoril," "Shoden," or an equivalent named proprietary extract) rather than an unbranded "ashwagandha extract," a disclosed withanolide percentage, and — where available — third-party testing certification (such as USP Verified or NSF Certified for Sport) confirming the label matches the contents.

What works and what does not

Stress and cortisol: the best-supported claim

This is where ashwagandha's evidence base is strongest. A 2025 meta-analysis pooling 15 randomized controlled trials and 873 participants found that ashwagandha supplementation significantly reduced serum cortisol at 8 weeks (mean difference −2.36, 95% CI −3.26 to −1.46, p<0.0001) and reduced Perceived Stress Scale scores at 8 weeks (mean difference −4.88, 95% CI −7.84 to −1.91, p=0.0013) compared with placebo (Bachour et al. 2025, BJPsych Open). Two of the most frequently cited individual trials underpinning this pooled result explicitly reported no funding and no conflicts of interest: a 60-day, 300 mg twice-daily trial found cortisol fell 27.9% versus 7.9% with placebo alongside a 44% drop in Perceived Stress Scale scores (Chandrasekhar et al. 2012), and a separate 60-participant, dose-ranging trial replicated the cortisol-lowering effect at both 250 mg and 600 mg daily doses (Salve et al. 2019).

The important nuance most marketing skips: a separate 2025 quantitative review of the same trial literature found the cortisol-lowering effect held up consistently across studies, but the effect on subjective "I feel less stressed" questionnaire scores was not statistically significant once correction methods were applied across the pooled studies. In plain terms — ashwagandha reliably moves a blood biomarker, but the evidence that people subjectively feel calmer is less consistent than the biomarker data alone suggests. This is a real, meaningful distinction rather than a technicality: a hormone change without a matching subjective improvement is a weaker basis for a wellness claim than both moving together.

Quality caveats apply even to the strongest evidence here. The headline 2025 meta-analysis was published as a conference-abstract supplement reviewed by the Royal College of Psychiatrists' Academic Faculty, which is a lighter editorial process than a standard peer-reviewed journal publication — it is not on the same evidentiary footing as a Cochrane review. Most of the underlying RCTs were also conducted in India, in relatively small samples (typically 40–80 participants), over 6–12 weeks, which limits how confidently the results generalize to other populations, ages, and baseline stress levels.

Anxiety symptoms

Anxiety-specific rating scales (Hamilton Anxiety Rating Scale, HAM-A; State-Trait Anxiety Inventory) also improved in the pooled 2025 analysis, with a mean HAM-A reduction of −3.52 at 8 weeks (95% CI −6.00 to −1.04, p=0.0053) compared with placebo (Bachour et al. 2025). The unfunded Chandrasekhar 2012 trial reported comparable reductions in Hamilton scores at 300 mg twice daily over 60 days (Chandrasekhar et al. 2012). These trials generally enrolled people with self-reported or mild-to-moderate stress/anxiety rather than individuals with a clinical generalized anxiety disorder (GAD) diagnosis, so the evidence supports ashwagandha as an adjunct for everyday anxiety symptoms rather than as a validated treatment for diagnosed anxiety disorders. A 2022 provisional guideline from the World Federation of Societies of Biological Psychiatry and the Canadian Network for Mood and Anxiety Treatments (WFSBP/CANMAT) task force lists ashwagandha as a provisional adjunct option at 300–600 mg/day of a 5%-withanolide extract for GAD, explicitly flagged as provisional and calling for more data before stronger recommendations are made.

Sleep quality

A 2021 meta-analysis pooling 5 randomized trials and roughly 400 participants — unfunded, with no author conflicts of interest declared — found an overall standardized mean difference of −0.59 (95% CI −0.75 to −0.42, p<0.001) favoring ashwagandha for sleep-quality outcomes (Cheah et al. 2021, PLOS ONE). The effect was substantially larger in participants who already had diagnosed insomnia (SMD −0.84) than in generally healthy adults without a sleep disorder (SMD −0.63), and larger at doses of 600 mg/day or higher and durations of 8 weeks or longer (SMD −0.68–0.69) than at lower doses or shorter durations (SMD −0.37–0.44). Notably, when the analysis isolated the widely used Pittsburgh Sleep Quality Index (PSQI) specifically, the result was not statistically significant (SMD −0.90, confidence interval crossing zero, p=0.150) — meaning the overall positive pooled result is being driven partly by other sleep-measurement tools, not uniformly by the single most standard sleep questionnaire. One of the five pooled trials was part-sponsored by a supplement manufacturer, which the review authors themselves disclosed.

Testosterone, libido, and fertility: the industry-funded claim

This is the claim most in need of a funding-conflict warning label. A widely cited 2013 pilot trial found that 675 mg/day of a high-concentration root extract over 12 weeks in 46 men with low sperm counts (oligospermia) raised testosterone by roughly 17%, sperm concentration by 167%, semen volume by 53%, and sperm motility by 57% (Ambiye et al. 2013). The paper does not disclose external funding, but the extract used closely mirrors the specifications later branded and heavily marketed by an Indian ashwagandha-extract manufacturer, and the trial has not been independently replicated at this size or effect magnitude outside India.

A separate, explicitly manufacturer-funded 2019 crossover trial in 57 healthy (not infertile) men found the Shoden branded extract raised testosterone by 14.7% and DHEA-S by 18% relative to placebo — but found no significant difference in cortisol between groups, which is notable given that cortisol-lowering is ashwagandha's best-supported effect elsewhere in the literature (Lopresti et al. 2019). This trial was funded directly by Arjuna Natural Ltd, the manufacturer of the Shoden extract being tested — a direct financial conflict of interest that the paper discloses. The NIH's own synthesis of this literature describes the testosterone and sperm-quality evidence as "limited," not proven (NIH ODS).

No large, independently funded, multi-site RCT has replicated a testosterone-raising effect in a general (non-infertile) population. The trials that exist are small (46–57 participants), short (8–12 weeks), and either funded by the extract manufacturer or use a product that closely matches a manufacturer's specific branded formulation without disclosing the funding relationship.

Strength, muscle, and exercise performance: another industry-funded claim

A 2021 systematic review pooling 13 studies and 615 participants reported large effects on strength/power (Cohen's d = 0.68) and cardiorespiratory fitness (d = 1.37–1.85, depending on how outlier trials were handled) favoring ashwagandha supplementation during resistance training programs. This review has a serious, disclosed conflict of interest: its lead author is a science product manager for a company that sells ashwagandha-containing supplements and sits on the advisory board of a raw-material supplier to the supplement industry, and a co-author is a paid industry consultant who has conducted industry-sponsored ashwagandha trials. Per the independent-evidence standard this article follows, a review authored by someone employed by a company that profits from positive findings is treated as conflicted evidence, not primary support for a benefit claim — regardless of how the individual trials within it were funded.

Setting the conflicted review aside, individual trials within that pooled literature are themselves small (typically 40–60 participants), short (8 weeks), and conducted primarily in resistance-trained young men in India, again raising generalizability questions. No independently funded, adequately powered RCT isolates ashwagandha's effect on strength or muscle mass outside this conflicted literature base.

Thyroid function: a genuine double-edged sword

A single RCT of 50 people (46 completed) with subclinical hypothyroidism found that 600 mg/day of ashwagandha root extract over 8 weeks significantly improved thyroid-stimulating hormone (TSH), T3, and T4 levels compared with placebo (p<0.001, p=0.0031, and p=0.0096 respectively). This trial's funding was not disclosed in the published paper, which is itself a transparency gap that should lower confidence in the result regardless of its statistical significance.

The flip side, largely absent from consumer marketing, is that the same herb has been linked to thyrotoxicosis (excess thyroid hormone, i.e., overactive thyroid symptoms) in independent case reports, including cases at doses as high as 1,950 mg/day over roughly two months. The NIH's synthesis of the literature explicitly lists thyrotoxicosis as a reported adverse effect (NIH ODS), and the Dutch national health institute's 2024 risk assessment documented at least one confirmed thyrotoxicosis case and one adrenal-suppression case among Netherlands pharmacovigilance reports (RIVM 2024). Denmark's food safety authority cited thyroid disruption as one of its stated concerns when banning ashwagandha food supplements.

The mechanistic tension is real, not a contradiction to be waved away: a substance capable of nudging an underactive thyroid toward normal can, in a different person or at a different dose, push thyroid hormone levels too high. This is precisely why ashwagandha is considered contraindicated in people with hyperthyroidism or Graves' disease, and why anyone on thyroid medication should not self-supplement without medical supervision.

Cognition and memory

The cognitive-enhancement claim is the weakest of ashwagandha's major marketing angles under an independent-evidence standard. Available human trials are small, short, inconsistently funded-disclosed, and often measure cognition as a secondary rather than primary outcome. Independent research-synthesis organizations reviewing this literature, including groups focused specifically on cognitive aging, describe human evidence for cognitive benefit as "very limited." No independently funded RCT of adequate size and duration has established a reliable cognitive-enhancement effect in a general population. Given the state of the evidence, this article's position is: independent human-trial evidence is insufficient to conclude that ashwagandha meaningfully improves memory or cognition in healthy adults.

Immune function and general "adaptogen" claims

The broad claim that ashwagandha is an "adaptogen" that helps the body handle stress of any kind — physical, chemical, biological — is a category label, not itself a testable clinical endpoint. The only specific, well-measured outcomes with meaningful independent human RCT support are cortisol reduction and sleep quality, covered above. Broader immune-modulation claims rest primarily on non-human laboratory and animal research and are excluded from this article's evidence base under the human-trials-only standard; independent human-trial evidence is insufficient to conclude that ashwagandha meaningfully modulates immune function in generally healthy people.

Benefits by claim

Risks and all side effects

No supplement is side-effect-free, and ashwagandha is not an exception. The evidence below is drawn from randomized trial adverse-event reporting, independent case reports, and government pharmacovigilance data — not manufacturer safety claims.

Common, generally mild side effects (seen in RCTs, usually comparable in frequency to placebo)

• Gastrointestinal upset — nausea, diarrhea, vomiting, abdominal pain, loose stools
• Drowsiness and sedation, particularly at higher doses or combined with other CNS depressants
• Headache
• Dizziness
• Nasal congestion

These effects are documented across the human RCT literature summarized by the NIH Office of Dietary Supplements and are generally described as mild and self-limiting (NIH ODS).

Rare but serious side effects

Drug-induced liver injury. This is the safety signal most consumer-facing ashwagandha content omits entirely. The NIH's LiverTox database — a joint project of the National Institute of Diabetes and Digestive and Kidney Diseases — assigns ashwagandha a Likelihood Score of B, meaning it is classified as a likely cause of clinically apparent liver injury based on accumulated case reports (LiverTox, NIH NIDDK, updated December 2024). Case reports documented in the LiverTox entry and in an independent, NIH-funded case series describe a cholestatic or mixed injury pattern, typical onset 2–12 weeks after starting the supplement, presenting classically with jaundice and itching (pruritus), usually resolving within 1–5 months of stopping the supplement — but with rare fatal cases and at least one instance requiring liver transplantation, concentrated in patients who had pre-existing liver disease (Björnsson et al. 2020). The Björnsson case series, funded through the NIH Drug-Induced Liver Injury Network with no industry funding or author conflicts of interest, documented five such cases with detailed laboratory values and causality assessments, several sourced from Iceland. Australia's Therapeutic Goods Administration separately logged 12 liver-injury reports through February 2024, with 7 assessed as possibly caused by ashwagandha (TGA safety alert, 2024), and the Netherlands' RIVM documented 4 liver-injury notifications through its national side-effect reporting center alongside 9 total human case reports reviewed (RIVM, January 2024).

Thyrotoxicosis (excess thyroid hormone). Independent case reports describe thyrotoxicosis developing in people taking ashwagandha, including at doses up to 1,950 mg/day over about two months, and RIVM documented at least one confirmed Netherlands case (NIH ODS; RIVM 2024).

Adrenal suppression. At least one documented case from the Netherlands pharmacovigilance system links ashwagandha use to adrenal suppression, consistent with the herb's HPA-axis-modulating mechanism working in an unwanted direction in a susceptible individual (RIVM 2024).

Serotonin syndrome (single case report, combined with an SSRI). A 2025 case report published in Neurology describes a 22-year-old woman on long-term escitalopram (a commonly prescribed SSRI antidepressant) who developed serotonin syndrome — limb myoclonus, eyelid flutter, tachycardia, fever, vomiting, dilated pupils, and cardiac arrhythmia — after consuming both a high-dose ashwagandha supplement (600 mg) and ashwagandha tea (roughly 1,520 mg) within two days. She was treated with benzodiazepines and supportive care and recovered. All study authors disclosed no funding and no conflicts of interest. This is a single case, so causality cannot be proven with certainty, but the case is well-documented with a plausible mechanism (theorized serotonergic/GABAergic activity layered onto an existing SSRI) and was significant enough to warrant peer-reviewed publication in a major neurology journal.

Severe gastrointestinal events. Australia's TGA safety alert also recorded 16 cases of severe gastrointestinal reactions requiring hospitalization associated with ashwagandha-containing products (TGA, 2024).

No independent, adequately powered long-term (multi-year) human safety trial exists for ashwagandha. Nearly all RCT safety data covers 6–12 week windows. Long-term safety with daily use over years is, honestly, unknown — this is a genuine evidence gap, not a reassurance.

All interactions

No confirmed, independent human interaction data exists for ashwagandha with anticoagulants/antiplatelets (warfarin, DOACs, aspirin, clopidogrel), statins, PPIs/antacids, or antiepileptics specifically. Absence of data is a safety gap, not evidence of safety — treat any combination with prescription medication as something to discuss with a pharmacist or physician rather than assume is fine by default.

Who should avoid Ashwagandha

• Pregnant people. Every regulator that has reviewed ashwagandha in the last several years advises against use in pregnancy. Ashwagandha has a traditional ethnobotanical use as an abortifacient, and RIVM's 2024 review found no adequate human reproductive-toxicity data exists to rule out risk (RIVM 2024). Germany's BfR explicitly lists pregnant and breastfeeding women among the groups advised to avoid it entirely (BfR, September 2024).
• Breastfeeding people. No adequate independent human safety data exists; treated the same as pregnancy by German and Dutch regulators.
• Anyone with pre-existing liver disease. This is the population in which the rare fatal and transplant-requiring liver injury cases concentrated (Björnsson et al. 2020; TGA 2024).
• Anyone with hyperthyroidism or Graves' disease. Documented thyrotoxicosis case reports make this a clear contraindication (NIH ODS).
• People on thyroid medication. Risk of additive/unpredictable thyroid hormone changes; requires medical supervision if combined.
• People with autoimmune conditions or on immunosuppressant medication. Theoretical immune-stimulating activity could counteract treatment; human evidence is sparse but the mechanistic concern is taken seriously by NIH and NCCIH.
• Anyone taking sedatives, benzodiazepines, or with an alcohol use pattern involving frequent heavy drinking. Additive CNS depression risk.
• Anyone on SSRIs/SNRIs or other serotonergic medications, especially at high ashwagandha doses. A documented, peer-reviewed serotonin syndrome case exists; while a single case, the underlying mechanism is plausible.
• People with hormone-sensitive cancers (e.g., prostate cancer) or on hormone therapy. Theoretical testosterone/DHEA-S-elevating activity warrants avoidance without specialist guidance.
• Anyone scheduled for surgery. Stop at least two weeks beforehand due to anesthesia-interaction risk (NCCIH).
• Children. Germany's BfR explicitly advises against use in children given the absence of pediatric safety data (BfR 2024).

The regulatory story in detail: how six jurisdictions reached similar conclusions independently

What makes ashwagandha's safety story unusual is not any single case report — it's that multiple government bodies in different countries, working from different national pharmacovigilance datasets, independently converged on the same concerns within a roughly four-year window. That pattern is harder to dismiss as coincidence or media exaggeration than a single flagged case would be.

Denmark moved first, and moved hardest. The Technical University of Denmark (DTU) conducted a formal risk assessment in 2020 examining ashwagandha's safety profile against the food-supplement regulatory framework. The assessment cited concerns across four domains: thyroid function disruption, reproductive and sex-hormone effects, central-nervous-system effects (sedation, mood alteration), and immune-system modulation. It also traced ashwagandha's traditional use as an abortifacient through an ethnobotanical citation chain — a use documented in traditional pharmacopeia references rather than modern clinical data, but treated by DTU as a legitimate reproductive-safety red flag given the absence of adequate modern human reproductive-toxicity data to rule the concern out. The Danish Veterinary and Food Administration (DVFA) acted on this assessment by banning ashwagandha as an ingredient in food supplements in 2023, a status that as of this writing remains in force. This makes Denmark the only major Western jurisdiction with an outright supplement-category ban, rather than a caution or warning label.

An academic rebuttal to the Danish position was published in 2024, arguing the ban rested on outdated or overly cautious risk assessment. This article treats that rebuttal as a conflicted source rather than a neutral scientific correction: one of its co-authors holds an honorary association with a pro-ashwagandha industry advocacy organization, and another holds an honorary government Ayurveda research professorship, both of which create a incentive to defend the ingredient's reputation. The existence of a pushback paper does not, on its own, invalidate DTU's original concerns — it simply means the debate is contested, with real financial and institutional interests on the pro-ashwagandha side of that contest.

The Netherlands built its case on hard pharmacovigilance numbers. RIVM's January 2024 risk assessment is notable for being unusually granular: it reviewed 9 individual human case reports (including 1 thyrotoxicosis case, 1 adrenal-suppression case, and 7 liver-injury cases) plus 4 separate liver-injury notifications logged through the Netherlands' national side-effect reporting center, Lareb. RIVM explicitly stated that no safe daily intake level (a health-based guidance value) could be established from the available data, and recommended against use generally, with particular emphasis on avoiding it during pregnancy given the absence of adequate reproductive-safety data.

Germany's BfR took a more targeted, population-specific approach. Rather than issuing a blanket caution, the Federal Institute for Risk Assessment's September 2024 communication named specific groups who should avoid ashwagandha entirely: children (due to a complete absence of pediatric safety data), pregnant and breastfeeding women (due to the reproductive-safety data gap), and people with pre-existing liver disease (due to the concentration of severe and fatal liver-injury cases in this population). For the general adult population without these risk factors, BfR's guidance was more measured — general caution rather than an outright recommendation to avoid.

Australia's TGA safety alert is the most detailed public accounting of case volume. The February 2024 alert documented 12 liver-injury reports through Australia's adverse-event reporting system, with 7 of those cases assessed by TGA reviewers as possibly caused by ashwagandha after ruling out other contributing medications or conditions in at least 4 of those cases. Separately, TGA logged 16 cases of severe gastrointestinal reactions serious enough to require hospitalization. Worth noting: Australia's food-safety regulator (FSANZ) had already classified ashwagandha as a prohibited "novel food" for use in the food category specifically — a separate and stricter classification than the therapeutic-goods/supplement category the TGA alert addresses, reflecting two different regulatory bodies reaching caution from two different angles.

The European Union opened a bloc-wide review rather than acting country by country. In June 2024, the Heads of (European Food Safety) Agencies initiated an Article 8 procedure — a formal EU mechanism for coordinating safety assessment of a substance across member states — and listed ashwagandha root among the "critical substances" under active review. This is a slower, more deliberative process than a single country's ban, but it signals that Denmark's concerns were taken seriously enough at the EU level to warrant a coordinated look rather than being dismissed as one country's outlier position. The UK's Food Standards Agency ran a parallel, public Call for Evidence in mid-2024, inviting researchers, manufacturers, and the public to submit safety data — a process that, as of this writing, has not concluded with a final UK-wide determination.

France's ANSES has also weighed in with a narrower, population-specific caution, recommending against ashwagandha use specifically in pregnant and breastfeeding women and in people with endocrine (hormone-system) disorders — a position that lines up closely with Germany's and the Netherlands' population-specific concerns even though it was reached through a separate national process.

The United States remains the outlier — not because the underlying safety signals are different, but because the regulatory framework is different. Ashwagandha is sold in the U.S. as a dietary supplement under the Dietary Supplement Health and Education Act (DSHEA), which does not require premarket safety or efficacy review the way pharmaceutical drugs do. The FDA can act against a specific supplement only after receiving evidence of harm significant enough to meet its intervention threshold — a reactive rather than proactive model. The NIH's Office of Dietary Supplements and National Center for Complementary and Integrative Health (NCCIH) are the primary U.S. government bodies independently synthesizing the science, and both institutions' fact sheets already reflect the liver-injury, thyroid, and interaction concerns documented above — but neither issues consumer-facing bans or mandatory warning labels the way DVFA, BfR, or TGA can.

Debunking the commercial noise

Four claims recur constantly in ashwagandha marketing copy, and each one collapses or needs heavy qualification once traced back to its actual evidence source.

"Clinically proven testosterone booster." This claim traces to two trials: a small, unreplicated 2013 pilot in infertile men using an extract that closely matches (but does not explicitly disclose funding from) a later heavily marketed branded product, and a 2019 crossover trial explicitly funded by the manufacturer of the extract being tested. Neither meets the bar of independently replicated, adequately powered evidence. Worse, the manufacturer-funded trial found no significant cortisol benefit — undercutting the idea that "boosting testosterone" and "reducing stress" are two effects of the same well-established mechanism; they may not even show up together in the same funded trial.

"Ashwagandha won't affect your liver — it's a natural, gentle herb." This is directly contradicted by the NIH's own LiverTox classification (Likelihood Score B — a likely cause of clinically apparent liver injury), by an NIH-funded case series documenting five detailed cases including fatalities and a liver transplant, and by independent pharmacovigilance data from the Netherlands and Australia. "Natural" is not a safety category, and ashwagandha's own case-report literature is one of the clearer illustrations of why that distinction matters.

"Thousands of years of Ayurvedic use proves it's safe." Traditional use documents tolerability in the specific forms, doses, and durations historically used — typically lower-concentration root preparations, often combined with other herbs, in variable individual doses. It does not substitute for modern randomized controlled trial evidence on the concentrated, high-withanolide, isolated extracts sold today at doses and durations that may not resemble traditional use at all. Traditional use is also the same evidence base that documents ashwagandha's abortifacient reputation — an example of how "traditional" is not synonymous with "harmless."

"Our extract's clinical dose is X mg — take exactly this much." Branded-extract marketing sometimes cites a "clinically studied dose" that doesn't match the dose used in the specific trial being referenced, or blends results across trials that used different formulations, withanolide percentages, or participant populations. Because withanolide concentration varies enormously between raw powder (under 1%) and the most concentrated branded extracts (up to 35%), a dose claim is only meaningful when tied to a specific, named, withanolide-percentage-disclosed product — a generic "clinically proven 600 mg" claim divorced from which extract was tested is not verifiable.

Dosage and how to take Ashwagandha

There is no single, universally validated ashwagandha dose. What follows are the dose ranges actually used in the human trials this article relies on — not manufacturer marketing claims.

• Stress and cortisol: 300 mg twice daily (600 mg/day total) of a standardized root extract (≥5% withanolides), for at least 8 weeks, is the most common effective dose across the strongest trials (Chandrasekhar et al. 2012; Bachour et al. 2025).
• Anxiety: Same 300–600 mg/day range; the WFSBP/CANMAT 2022 task force provisionally suggests 300–600 mg/day of a 5%-withanolide extract as a GAD adjunct, explicitly flagged as a provisional, not firm, recommendation.
• Sleep: ≥600 mg/day for ≥8 weeks shows the most consistent effect size in the pooled sleep meta-analysis; lower doses and shorter durations produced smaller, less reliable effects (Cheah et al. 2021).
• High-concentration branded extracts (e.g., Shoden-type, up to 35% withanolides): Trials used considerably lower milligram doses (120–240 mg/day) to deliver a comparable withanolide amount — do not assume a "low mg number" means a weak dose if the withanolide percentage is high.
• Traditional root powder doses (3,000–6,000 mg/day) reflect Ayurvedic tradition, not modern RCT evidence — this dose range has not been validated in quality-controlled human trials the way standardized extracts have.

No regulator has established a tolerable upper intake level (UL) for ashwagandha, and the U.S. FDA does not review dietary supplements for safety or effectiveness before they go to market under the Dietary Supplement Health and Education Act (DSHEA) framework (NIH ODS). "No UL exists" should be read as a data gap, not a green light for high doses — if anything, the liver-injury and thyrotoxicosis case reports above occurred in some cases at doses well above typical trial ranges (up to 1,950 mg/day of extract in one thyrotoxicosis case).

Most trials dosed ashwagandha with food, generally split into a morning and evening dose or taken once in the evening for sleep-focused use. There is no independent human evidence establishing a definitively superior time of day for stress/anxiety use; evening dosing is more commonly used in the sleep-focused trials given the drowsiness side effect profile.

Infographics with full text versions

Infographic 1: Evidence strength by claim

Infographic 2: Global regulatory status, 2020–2026

Related research

Ashwagandha's best-supported use case — lowering cortisol and easing everyday stress and anxiety symptoms — sits within a broader picture of stress management that extends past any single supplement. For a fuller, evidence-graded look at stress, anxiety, and depression prevention approaches, see the Pure City Research Stress, Anxiety, and Depression Prevention Guide.

Because ashwagandha's sleep-quality evidence is strongest in people who already have diagnosed insomnia, readers focused primarily on sleep may get more value from a broader, sleep-specific evidence review — see the Pure City Research Sleep Prevention Guide.

Given the documented interaction risk between ashwagandha and blood-pressure medication, readers managing hypertension alongside supplement use should also review the Pure City Research Blood Pressure Prevention Guide before combining the two.

Because ashwagandha has shown fasting-glucose-lowering effects in some trials — and therefore carries an additive hypoglycemia risk with antidiabetic medication — readers managing blood sugar should also review the Pure City Research Diabetes Prevention Guide.

Frequently asked questions

Sources and funding notes

Every source below was screened for funding, author conflicts of interest, and country of origin, following an independent-evidence standard that relies exclusively on human-trial data. Industry-funded and conflicted sources are included only to explain what this article does not rely on for benefit claims.

Animal and non-human evidence excluded

The following animal studies were encountered during research and would likely have been cited if this article permitted non-human evidence. They are excluded under the independent-evidence standard this article follows and do not support any claim made above.

In-vitro evidence used

None. Research for this article identified in-vitro work (for example, mouse hippocampal HT-22 cell line studies paired with mouse in-vivo depression models), but none met this article's narrow allowance for in-vitro evidence — full replication of a human biological context using human cells at human-relevant doses, used only when no human-trial evidence exists. Because meaningful human RCT and human case-report evidence exists for every claim covered in this article, no in-vitro evidence was needed or used.