Does Ashwagandha Actually Reduce Stress?

Ashwagandha has something most supplements on this site lack: a body of randomised, placebo-controlled trials that consistently report it reduces stress and anxiety. It also has something most supplements lack in the other direction: a growing case literature of liver injury serious enough to include transplants, and a national ban.

The trials are the reason it is the most searched supplement of 2026. The size of the effects they report is the reason to slow down. One 2025 meta-analysis of 22 studies reports that ashwagandha improves stress with a standardised mean difference of minus 5.88. For scale, an effect size of 0.8 is conventionally called large, and the best-evidenced antidepressants produce effects around 0.3. A pooled effect seven times larger than a large effect is not a finding about ashwagandha; it is a signal that something is wrong with the studies being pooled, and the reviewers who say so are the ones worth reading.

What the trials find, and why the numbers are strange

A 2022 meta-analysis in Phytotherapy Research set out because, in the authors’ words, clinical trial studies revealed conflicting results on the effect of ashwagandha extract on anxiety and stress. Pooling 12 randomised trials and about 1,000 participants, it found significant reductions in both, with effect sizes around minus 1.5 for anxiety and minus 1.75 for stress. It then rated the certainty of that evidence as low, with heterogeneity, the statistical measure of how much the studies disagree with each other, at 84 to 94 per cent. Anything above 75 per cent is conventionally called substantial.

A 2024 meta-analysis of nine trials and 558 patients found a mean reduction of 4.72 points on the Perceived Stress Scale and 2.19 on a standard anxiety scale, with confidence intervals that in both cases came close to zero. A 2025 analysis in BJPsych Open pooled 15 studies and 873 patients with similar conclusions and no external funding.

And then the outlier. A 2025 dose-response meta-analysis of 22 studies reports stress reduced by a standardised mean difference of minus 5.88, depression by minus 5.68 and anxiety by minus 6.87. Those figures are not credible as effects of a herbal supplement on mood, and they are almost certainly an artefact of how the pooled studies were designed and reported: small samples, short durations, single research groups, and outcome measures that move a great deal with expectation.

The most useful reading of the whole literature comes from a 2025 review that separated the two outcomes. Its title says it: significant cortisol reduction but no effects on perceived stress. Ashwagandha lowered the stress hormone in the blood; it did not change how stressed people said they felt. The authors describe that disconnect as highlighting the need for longer treatment and broader populations, and note that gaps remain regarding long-term safety, dosing and effectiveness across diverse populations.

A biomarker moving while the symptom does not is a pattern this site has met before. It was the story in multivitamins and energy and it sits alongside creatine and the brain, where the effect was real only in a subgroup. Whether a lower cortisol reading matters if you feel no different is a question the trials do not answer.

Who ran the trials

Section 4 of this site’s standards requires disclosure of funding that could bear on a finding. Most of the foundational ashwagandha trials, including the 2012 study most often cited for cortisol reduction, used extracts supplied by the manufacturers of branded formulations, and several were run in India by groups with ongoing manufacturer relationships. The two most recent meta-analyses that reached positive conclusions do state they had no specific funding, which is worth crediting. But the studies they pool are the same small, manufacturer-supplied trials, and pooling does not launder the design of what goes in.

The liver

This is where ashwagandha differs from the other supplements on this site, and it is the reason this article exists.

The United States National Institutes of Health’s LiverTox database, updated December 2024, states that ashwagandha has recently been implicated in cases of clinically apparent liver injury. Its review of the literature counts 23 reported cases, with typical onset of cholestatic hepatitis after a month of use, most resolving spontaneously, but with occasional instances of fatal liver failure or chronic injury.

A case series from India, published in Hepatology Communications in 2023, describes ashwagandha-related liver injury presenting as cholestatic hepatitis and states it can lead to acute-on-chronic liver failure with high mortality in those with pre-existing liver disease. A separate series from Iceland and the United States Drug-Induced Liver Injury Network was published in 2020, and a 2022 report in the American Journal of Therapeutics describes acute liver failure requiring transplantation. A 2026 systematic review of case reports collects the literature and received no funding.

Two things make this hard to weigh. Case reports cannot establish a rate, and against millions of users a few dozen documented injuries is a low one. Defenders point out that across more than a thousand trial participants no serious liver injury was recorded, which is true and also what you would expect when the rare event has not had enough exposures to appear. And some cases may involve contaminated or mislabelled products rather than the herb, a real problem in an unregulated market that does not help a buyer choose.

Denmark banned ashwagandha supplements on the basis of a risk assessment by its technical university. A response in an Ayurvedic journal argues the assessment overrelied on animal studies and selectively reviewed the literature, and that it ignored the herb’s long history of traditional use. That is a fair description of a regulatory dispute, and a reader should know the ban exists and is contested.

Where this belongs with a clinician

Anyone with existing liver disease should not take ashwagandha, on the evidence above. Anyone taking it who develops jaundice, dark urine, pale stools, itching or unusual fatigue should stop and see a doctor promptly, because those are the presenting signs in the case reports. It interacts with sedatives, thyroid medication and immunosuppressants, and is generally advised against in pregnancy.

Persistent stress or anxiety that affects sleep, work or relationships is a reason to see a doctor or therapist, and the treatments with strong evidence are not herbal. If you are in crisis, contact your local emergency services or a crisis line now.

This article names no dose, no form and no product. The trials used different extracts at different doses, which is part of why they disagree, and this site does not resolve that by picking one.

The shortfall pattern does not apply here

This differs from magnesium and sleep and low vitamin D and tiredness, where the recurring finding was that effects appear in people with a deficiency to correct. Nobody is deficient in ashwagandha. The claim is pharmacological: that a plant compound acts on the stress response. The evidence for that is a stack of small manufacturer-supplied trials reporting effects larger than any known drug, one careful review finding the hormone moves while the feeling does not, and a safety signal that is rare, real and occasionally severe.

Common questions

Does ashwagandha work for stress and anxiety?

The randomised trials say yes, but the numbers they produce are the problem. A 2022 meta-analysis of 12 trials found significant reductions but rated the certainty of evidence as low with heterogeneity of 84 to 94 per cent. A 2025 analysis reports a stress effect of minus 5.88, which is implausibly large for any intervention. A separate 2025 review found ashwagandha lowered cortisol without changing perceived stress. The signal exists and the studies behind it are small, short and mostly manufacturer-supplied.

Is ashwagandha bad for your liver?

Rarely, and sometimes seriously. The NIH’s LiverTox database counts 23 reported cases of clinically apparent liver injury, typically cholestatic hepatitis about a month into use, mostly resolving but occasionally fatal or requiring transplantation. An Indian case series warns of acute-on-chronic liver failure with high mortality in people with pre-existing liver disease. Denmark has banned the supplement; that decision is contested. Anyone with liver disease should avoid it and anyone developing jaundice, dark urine or itching should stop and see a doctor.

Why are the effect sizes so large?

Because the pooled trials share features that inflate effects: small samples, short durations, outcome scales that move readily with expectation, and a small number of research groups, several with manufacturer relationships. Heterogeneity above 80 per cent means the trials disagree substantially with each other, and reviewers rate the certainty of evidence as low. An effect size several times larger than large is a sign to distrust the pooling, not to believe the herb.

Does ashwagandha lower cortisol?

Across trials, yes, and a 2025 review confirmed it with moderate heterogeneity and no publication bias detected. The same review found no effect on perceived stress. Whether a lower cortisol reading matters when you feel no different is unresolved, and the authors call for longer studies in broader populations. Cortisol is also a poor single measure of stress, since it varies with time of day, sleep and exercise.

Real trials, unreliable numbers, a real risk

Ashwagandha is not a placebo with good marketing. It is a compound that does something measurable in trials, and the trials are too small, too short and too close to the companies that sell it for the size of that something to be trusted.

The liver signal is the part that separates it from the rest of this site’s supplement coverage. It is rare. It is also documented, occasionally severe, and the reason a herb sold for calm now carries a warning in a national liver database. If you take it, know the signs. If you have liver disease, do not.