Ashwagandha for Stress and Cortisol: What the Trials Show

Ashwagandha (Withania somnifera) has gone from a centuries-old Ayurvedic herb to one of the fastest-growing ingredients in US stress-relief supplements, and unlike some of the ingredients we’ve covered on this site, it actually has a reasonably consistent body of small human trials behind the stress claim. That doesn’t mean the picture is simple — the trial base is thin by the standards of a mainstream medication, dominated by short studies, and there’s a real, recently documented gap between what it does to a stress hormone and what people actually report feeling.

Here’s what we found, including a safety issue — rare liver injury — that gets far less attention than the stress claim does.

The short version

• The best-quality synthesis of the evidence, a 2022 meta-analysis pooling 12 trials in 1,002 people, found ashwagandha significantly reduced both stress scores (SMD −1.75) and anxiety scores (SMD −1.55) compared to placebo — large effect sizes by conventional statistical standards.
• But NCCIH, reviewing the same broad literature, describes the evidence as reasonably supportive for stress, while calling the evidence on anxiety specifically “unclear.” Those are different claims, and the marketing usually doesn’t distinguish them.
• A newer 2025 meta-analysis of 8 trials found something worth taking seriously: ashwagandha produced a statistically significant drop in measured cortisol, but no significant improvement in how stressed people said they felt on a standard perceived-stress questionnaire. A hormone moving in the right direction didn’t reliably translate into the subjective outcome the supplement is sold on.
• Several of the most-cited trials use a specific patented extract (KSM-66), manufactured and commercially promoted by the ingredient’s own supplier — a funding pattern worth knowing about, which we were only able to partially verify ourselves.
• Ashwagandha carries a real, if rare, safety signal that’s easy to miss in stress-relief marketing: documented cases of clinically apparent liver injury, including a small number of fatal cases and liver transplants, tracked by the NIH’s own drug-induced liver injury registry.

Our evidence rating is Moderate for stress as a broad outcome, but drops toward Limited for anxiety treated as its own distinct claim — see the ratings section for why we’re not collapsing these into one line.

What the trials actually measured

Most modern ashwagandha trials for stress use a standardized root extract, dosed once or twice daily, over 8 to 12 weeks, in adults who are healthy but self-report elevated stress — not people with a diagnosed anxiety disorder. That population detail matters for reading the results honestly.

The pooled evidence

A 2022 systematic review and meta-analysis by Akhgarjand and colleagues, published in Phytotherapy Research, pooled 12 randomized controlled trials covering 1,002 participants aged 25 to 48. Compared to placebo, ashwagandha supplementation was associated with a significant reduction in stress scores (standardized mean difference −1.75, 95% CI: −2.29 to −1.22) and anxiety scores (SMD −1.55, 95% CI: −2.37 to −0.74), both reported at p = .005. Those are large effect sizes as statistics go — but the same paper reported very high heterogeneity between trials (I² in the 83–94% range), meaning the individual studies varied a great deal in design, population, and result, which tempers how much confidence a single pooled number deserves.

NCCIH’s own review of the broader literature lands in a similar but more cautious place: it states research shows some ashwagandha preparations “may be effective for insomnia and stress,” while separately noting the evidence for anxiety specifically “is unclear.” That’s a meaningful distinction the Akhgarjand pooled anxiety number tends to flatten — a systematic review can show a significant pooled effect on an anxiety scale while the underlying evidence for anxiety as a clinical claim still doesn’t clear the bar a regulatory-adjacent body considers convincing.

The complication we want to be direct about

A newer 2025 systematic review and meta-analysis (Albalawi et al.), pooling 8 randomized trials in 488 adults, measured both cortisol — a bodily stress-hormone marker — and scores on the Perceived Stress Scale (PSS), a standard self-report questionnaire. Ashwagandha produced a statistically significant reduction in cortisol (roughly −1.16 µg/dL, 95% CI: −1.64 to −0.69, p < 0.001). But the same analysis found no significant effect on PSS scores (SMD −0.355, 95% CI: −1.188 to 0.47, p = 0.40) — people’s own reported sense of stress didn’t move in a way that cleared statistical significance, even though their cortisol did.

That’s a real tension worth sitting with rather than resolving in the supplement’s favor: a hormone marker moving favorably is not the same thing as a person feeling less stressed, and the second is what someone buying this product actually wants. We’re flagging it plainly rather than leading with the more flattering 2022 pooled-anxiety number and leaving this one for a footnote.

Where the money comes from — a disclosure we could only partly verify

Several of the most frequently cited trials in this space, including the foundational 2012 study by Chandrasekhar, Kapoor and Anishetty in the Indian Journal of Psychological Medicine (600 mg/day of a high-concentration extract, reducing both stress/anxiety scores and serum cortisol versus placebo), use KSM-66 — a specific patented ashwagandha extract manufactured and commercially promoted by its supplier, Ixoreal Biomed.

Per our Editorial Policy, we disclose funding and commercial ties in the research we cite. In this case, we were not able to independently confirm the funding and conflict-of-interest statement in the primary 2012 paper itself — the full text was not accessible to us during this research pass. What we can say with confidence: KSM-66 is a branded ingredient with an active commercial marketing operation built substantially around this and similar trials, which is a reason for extra scrutiny regardless of the specific funding line in any one paper. We’re naming that limitation rather than either asserting a funding conflict we haven’t verified or omitting the concern entirely.

Safety: the liver injury signal that gets little attention

This is the part of the ashwagandha story that rarely makes it into stress-relief marketing.

The NIH’s LiverTox database — a clinical reference on drug- and supplement-induced liver injury maintained by the National Institute of Diabetes and Digestive and Kidney Diseases — rates ashwagandha at Likelihood Score B: a likely cause of clinically apparent liver injury, based on an accumulating case-report literature rather than controlled-trial data (since trials are generally too small and short to catch a rare event). The typical pattern: liver injury presenting 2 to 12 weeks after starting the supplement, usually cholestatic or mixed in pattern, with jaundice and itching. Most cases resolve within one to five months of stopping the product, but LiverTox documents rare fatal cases and at least one instance requiring emergency liver transplantation, disproportionately in people with pre-existing liver disease. A 2023 case series review found roughly 20+ published cases at that point, still a small fraction of ashwagandha’s overall use, but a real and non-theoretical risk.

NCCIH’s own fact sheet corroborates the broad strokes: ashwagandha “may be safe” short-term (up to about 3 months), with insufficient data on long-term safety, and explicitly lists rare liver injury among the known concerns — alongside drowsiness, stomach upset, diarrhea, and vomiting as more common, milder effects.

Thyroid effects are a separate, documented concern. LiverTox’s bibliography includes a published case of thyrotoxicosis (an overactive thyroid) in a previously healthy woman that developed after starting, and then increasing the dose of, an ashwagandha preparation — resolving on its own after she stopped. NCCIH’s fact sheet independently advises against ashwagandha for people with thyroid disorders, autoimmune conditions, or before surgery, and flags interactions with diabetes medications, blood pressure medications, immunosuppressants, sedatives, and anticonvulsants.

Pregnancy and hormone-sensitive conditions. NCCIH states ashwagandha should be avoided during pregnancy and while breastfeeding. Because it may raise testosterone levels, NCCIH also advises against its use by people with hormone-sensitive prostate cancer.

What we could not check

• We did not verify the funding disclosure of the foundational KSM-66 trial (Chandrasekhar 2012) — flagged plainly above rather than assumed either way.
• We did not independently read the full text of either meta-analysis (Akhgarjand 2022, Albalawi 2025) — the effect sizes and confidence intervals above are drawn from the papers’ own published abstracts, cross-checked across multiple citations of each, not from a full-text read of methods and individual trial data.
• We could not verify a specific dose-response claim. Some secondary sources describing the Akhgarjand 2022 meta-regression cited an anxiety dose-response figure that looked implausibly high compared to any individual trial’s actual dosing; rather than risk repeating a garbled number, we’ve left it out of this piece entirely.
• We did not test any product. We have no independent data on what’s actually in any specific ashwagandha product sold today, including whether it matches its labeled extract type or concentration.
• We did not evaluate long-term safety. Both NCCIH and LiverTox note that data beyond roughly 3 months of continuous use is limited.

Who should talk to someone first

Given the documented, if rare, liver injury signal, anyone with existing liver disease — including cirrhosis or chronic hepatitis — should talk to a doctor before using ashwagandha, and LiverTox explicitly notes it should be avoided in people with cirrhosis or advanced chronic liver disease. The same applies to anyone with a thyroid disorder or autoimmune condition, anyone scheduled for surgery, anyone taking diabetes medication, blood pressure medication, immunosuppressants, sedatives, or anticonvulsants, and anyone who is pregnant, breastfeeding, or has hormone-sensitive prostate cancer.

If new, unexplained fatigue, itching, dark urine, or yellowing of the skin or eyes appears while taking ashwagandha, LiverTox’s own case reports suggest that’s a reason to stop the product and seek medical attention rather than wait it out.

Nothing here is medical advice, and no one on our team is a clinician. That’s stated plainly on our About page.

Our rating, and why

For stress as a broad outcome: Moderate. Under our evidence scale, this reflects a real, if heterogeneous, body of randomized trial evidence — 12 trials and over 1,000 participants in the largest pooled analysis, with a consistent direction of effect across most individual studies, even though the trials vary considerably in quality and the pooled statistics show high heterogeneity.

For anxiety treated as a distinct, separate claim: Limited. NCCIH’s own assessment — “unclear” — is more conservative than the pooled statistical significance in the broader meta-analyses would suggest on its own, and we’re deferring to that more cautious read rather than the more flattering summary statistic, consistent with this site’s standing rule against rating inflation.

On the cortisol-versus-felt-stress question specifically: we’re treating the 2025 finding (cortisol down, perceived stress unchanged) as an open, unresolved tension rather than folding it into either rating — it’s exactly the kind of result that should make a reader skeptical of any product claim that leans only on “lowers cortisol” language, since lowering cortisol did not, in that analysis, reliably make people feel less stressed.

If the research develops further — particularly longer trials, or work that resolves the cortisol/perceived-stress gap — this rating will change, and we’ll say so on this page when it does.

Sources

1. National Center for Complementary and Integrative Health (NCCIH). Ashwagandha: Usefulness and Safety. Last updated March 2023. https://www.nccih.nih.gov/health/ashwagandha
2. LiverTox: Clinical and Research Information on Drug-Induced Liver Injury. Ashwagandha. National Institute of Diabetes and Digestive and Kidney Diseases, NIH. Last updated December 3, 2024. https://www.ncbi.nlm.nih.gov/books/NBK548536/
3. Akhgarjand C, Asoudeh F, Bagheri A, et al. Does Ashwagandha supplementation have a beneficial effect on the management of anxiety and stress? A systematic review and meta-analysis of randomized controlled trials. Phytotherapy Research. 2022;36(11):4115-4124. https://onlinelibrary.wiley.com/doi/10.1002/ptr.7598
4. Albalawi AA, et al. Dual impact of Ashwagandha: Significant cortisol reduction but no effects on perceived stress — A systematic review and meta-analysis. 2025. https://journals.sagepub.com/doi/abs/10.1177/02601060251363647
5. Chandrasekhar K, Kapoor J, Anishetty S. A prospective, randomized double-blind, placebo-controlled study of safety and efficacy of a high-concentration full-spectrum extract of ashwagandha root in reducing stress and anxiety in adults. Indian Journal of Psychological Medicine. 2012;34(3):255-262.
6. Björnsson HK, Björnsson ES, Avula B, et al. Ashwagandha as a cause for liver injury. Liver International. 2020;40:2035-2036. (Cited via LiverTox, source 2.)
7. Philips CA, Valsan A, Theruvath AH, et al. Ashwagandha-induced liver injury — A case series from India and literature review. Hepatology Communications. 2023;7:e0270. (Cited via LiverTox, source 2.)
8. van der Hooft CS, Hoekstra A, Winter A, de Smet PA, Stricker BH. [Thyrotoxicosis following the use of ashwagandha]. Ned Tijdschr Geneeskd. 2005;149:2637-2638. (Cited via LiverTox, source 2; original in Dutch.)