Ashwagandha for Stress: What the Trials Measured, and What They Missed
Evidence Review · The Reading Room, Vol. 7 · 12 min read · What the trials measured, who they were actually run on, and the safety signal the shelf does not mention
Two 2025 meta-analyses agree it lowers a hormone and disagree about whether anyone feels different. The one trial run on adults your age found no effect on stress at all.
It is on the shelf in every pharmacy now, generally beside the magnesium, generally with a word like calm or adapt on the front. The proposition is modest and appealing: a plant, used for a very long time, that takes the edge off. Two capsules, roughly the price of a coffee a day, no prescription and no appointment where anyone tells you your results are normal.
This review examines what the trials support. It is a reading of published pharmacology and regulatory documents; we are not clinicians, the findings belong to the researchers named below, and only the plain-English rendering is ours.
The short answer: the evidence that ashwagandha lowers measured cortisol — the main hormone released under sustained stress — is reasonably consistent. The evidence that it makes people feel less stressed is not. Two systematic reviews published in 2025 reached opposite conclusions on that second point from overlapping literature.¹² That divergence has a name worth having, because it is what you are actually buying: the split result. The assay moves. The month does not. And the single trial conducted on adults of roughly the age reading this, mean age 54 to 55, found no reduction in stress at all — though fatigue did improve.³
- 15 trials, 873 participants: cortisol fell. A 2025 meta-analysis found a pooled reduction of −2.36 (95% CI −3.26 to −1.46, p < 0.0001).¹
- Two 2025 reviews, opposite verdicts on how people felt. One found perceived stress fell (−4.88, 95% CI −7.84 to −1.91).¹ The other is titled, in terms, “significant cortisol reduction but no effects on perceived stress”.²
- The trial in your age band found nothing on stress. 120 adults, mean age 54 to 55, 12 weeks: no reduction in stress. Fatigue did improve.³
- Quality of life did not move. In the same meta-analysis that found benefits on cortisol and stress, quality of life was not significantly changed (−1.76, 95% CI −5.61 to 2.09, p = 0.37).¹
- At least six documented liver-injury cases, at doses of 450 to 1,350 mg a day taken for one week to four months, plus regulatory action in Denmark and a 2024 French advisory.⁴⁵
The source for this review
Dr Adrian L. Lopresti, PhD
A clinical psychologist and researcher based in Western Australia, adjunct at Murdoch University and managing director of Clinical Research Australia. He runs randomised placebo-controlled trials of botanical extracts — saffron, curcumin, ashwagandha — and his group ran the one ashwagandha trial on adults of roughly the age reading this. He is worth naming precisely because the result was null on the outcome the product is sold for, and he published it that way.
What we read: Smith, Lopresti and Fairchild, Journal of Psychopharmacology (2023), alongside the two 2025 meta-analyses and the NIH Office of Dietary Supplements fact sheet.¹²³⁴
Where to follow his work: LinkedIn · ResearchGate · Kudos profile
Scope and method of this review
This review covers ashwagandha (Withania somnifera) taken orally by adults for stress, anxiety and related outcomes. Sources were selected for level of evidence: the two most recent systematic reviews with meta-analysis, the randomised trial whose sample is closest to this readership, and the US National Institutes of Health Office of Dietary Supplements fact sheet, which catalogues the individual trials and the safety reporting.
What this review does not cover. It does not evaluate any brand as a product, though it notes where a trial used a proprietary extract, because that turns out to matter. It does not cover ashwagandha for sleep, athletic performance, thyroid function or fertility, each of which has its own literature. It does not assess traditional use, which is a different kind of evidence answering a different question. It does not address use in pregnancy or breastfeeding, where the advice is straightforward and negative.⁴ And it makes no recommendation about whether you should take it, which is not a decision an evidence review is entitled to make.
The evidence table
| Source | Sample | Finding | Limitation |
|---|---|---|---|
| 2025 meta-analysis, BJPsych Open¹ | 15 trials, 873 adults | Cortisol −2.36 (95% CI −3.26 to −1.46); PSS −4.88 (−7.84 to −1.91); anxiety −3.52 at 8 weeks (−6.00 to −1.04) | Quality of life not significantly improved (p = 0.37); wide intervals on the stress outcome indicate substantial disagreement between trials |
| Albalawi, 2025² | Systematic review and meta-analysis | Cortisol reduced, perceived stress not. The review is titled for this split finding | Overlaps in source trials with the review above while reaching a different conclusion on the same outcome — the clearest signal that this literature is unstable |
| Smith, Lopresti & Fairchild, 2023³ | 120 adults, mean age 54–55, 12 weeks | No reduction in stress. Fatigue did improve | A single trial of one proprietary extract; a null result is not proof of no effect, only an absence of one here |
| 2021 systematic review (via NIH ODS)⁴ | 7 studies, 491 adults, all conducted in India | Reduced stress and anxiety significantly, over 6 to 8 weeks at 240–1,250 mg a day | Single-country evidence base; wide dose range; short duration; extraction and standardisation differ between studies |
| Björnsson et al., case series⁵ | Cases from Iceland and the US Drug-Induced Liver Injury Network | Liver injury following ashwagandha use, at ordinary retail doses | A case series cannot establish incidence — it shows the harm occurs, not how often. Denominator unknown |
Read the finding column downward. Consistent on the hormone, inconsistent on the experience, and null in the one sample close to your age.
If the load has been too heavy for too long and no capsule was ever going to shift that, the Quiet Audit is ten private minutes on the actual question. No programme, no advice, nobody sees your answers. Start here.
Key terms
Adaptogen — a marketing and traditional-medicine category for plants said to help the body resist stress. It is not a regulatory or pharmacological classification, and nothing follows from a substance being described as one.
Perceived Stress Scale (PSS) — a ten-item questionnaire measuring how unpredictable and overloaded life has felt in the last month. This is the outcome that matters to the person taking the capsule, as distinct from the one measured in saliva.
Standardised extract — a preparation processed to contain a fixed percentage of active compounds, here withanolides. Different proprietary extracts are not interchangeable, which is why a positive result for one is not a result for the bottle on your shelf.
Confidence interval (95% CI) — the range within which the true effect probably sits. If it crosses zero, the result is compatible with no effect at all. Watch the width: a wide interval means the trials disagreed considerably.
The gap between the hormone and the feeling
This is the whole story, and it is worth being precise about.
Almost every review agrees ashwagandha lowers measured cortisol. The 2025 BJPsych Open meta-analysis found a pooled reduction with a tight interval and a very small p value.¹ The 2025 Albalawi review agrees on cortisol.² Where they part company is what happens to the person.
The first review found perceived stress fell by 4.88 points — but look at the interval, which runs from −7.84 to −1.91.¹ That is wide. It means the underlying trials disagreed considerably about the size of the effect, and the pooled figure is an average across studies that were not telling the same story. The second review, working from overlapping literature, concluded there was no effect on perceived stress at all.² And in the same meta-analysis that found the stress benefit, quality of life did not move significantly.¹
Two systematic reviews published in the same year, drawing on much the same trials, disagreeing about whether anyone felt better. That is not a body of evidence converging on an answer.
Three structural reasons sit underneath it, all named by the NIH fact sheet itself.⁴ The trials are “most of them fairly small in size and of short duration”. They use different preparations with different extraction and standardisation processes, which the fact sheet says makes it difficult to identify which specific extracts have been studied. And the doses span an enormous range — from 120 mg to 1,250 mg of extract, and in one case 12,000 mg of whole root.
Add the geography. The 2021 review covers seven studies and 491 adults, all conducted in India, in a research culture where the substance is part of a traditional medical system.⁴ That is not a reason to dismiss the findings. It is a reason to want replication somewhere else — which is exactly what the trial in your age band provides, and it points the other way.
The saliva changed. The month did not.
The safety question
This section exists because the supplement aisle implies a safety profile that the regulatory record does not entirely support.
The NIH fact sheet documents at least six published cases of liver injury associated with ashwagandha, at doses of 450 to 1,350 mg per day, with onset between one week and four months of starting.⁴ A case series drawn from Iceland and the US Drug-Induced Liver Injury Network describes the pattern.⁵ Regulators have responded: the fact sheet records action in Denmark, and a 2024 statement from the French agency ANSES advising against use in pregnancy and breastfeeding and in people with endocrine disorders.⁴
Two things are true at once and both deserve stating plainly. Six documented cases against a very large number of users is a low absolute rate, and a case series cannot establish incidence because the denominator is unknown. But drug-induced liver injury is not a mild adverse event, the doses involved were ordinary retail doses rather than misuse, and the onset was within four months.
The practical reading is not that this is dangerous. It is that this is a pharmacologically active substance being sold in the same aisle as vitamin C, and it should be treated as the former.
How strong is this evidence?
Contested. Two systematic reviews published in the same year, drawing on overlapping trials, reached different conclusions about the outcome that matters most to the person taking it.
Reasonably established: that ashwagandha reduces measured cortisol across pooled trials,¹² and that it has been studied in genuine randomised, placebo-controlled designs rather than only in testimonial.
Not established: that it reliably reduces how stressed people feel; that any effect persists beyond the 6 to 12 weeks most trials ran; that results from one standardised extract apply to another; or that it improves quality of life, which was specifically not significant in the review that found the other benefits.¹
The limitation that matters most for the reader of this page. Almost none of this research was conducted on women in their late forties, fifties and early sixties. Where the mean age was closest — 54 to 55, in a trial of 120 adults over 12 weeks — the result on stress was null.³ That single null does not settle the question. It does mean the positive findings quoted on the packaging come predominantly from samples with mean ages in the twenties and thirties, and that the one attempt to test it on people your age did not replicate them.
Five things that look like this and are not this
The sections above protect the literature. This one is for the reader, and it is the section the packaging cannot write, because it invites you to consider that the thing you are medicating has a name and a test.
Five ordinary conditions produce exactly the state that sends a competent woman to the supplement aisle at fifty-two. Each is checkable, and four of the five are checkable in one appointment.
- Iron deficiency, with or without anaemia. Common in the years before the final period, and it produces breathlessness on stairs alongside the flatness. A ferritin result settles it.
- Hypothyroidism. Fatigue with cold intolerance, dry skin, weight change and slowed thinking. One blood test.
- Obstructive sleep apnoea. No supplement touches it. Waking unrefreshed after a full night is the signature, and someone else has usually noticed the breathing first.
- The perimenopausal sleep pattern. A fragmented second half of the night, most nights, with the exhaustion arriving before the day rather than at the end of it — eight hours in bed is not eight hours of sleep.
- Depression presenting as flatness rather than sadness. The distinguishing feature is not weight but responsiveness: depletion lifts a little on a clear weekend, anhedonia does not.
None of these is ruled out by a capsule and none is ruled in by one. They also stack, which is why a demanding year with untreated apnoea in it will defeat any supplement ever formulated.
Who else has assessed this
One trial is one trial, and one meta-analysis is one set of inclusion decisions. Three other groups have looked at the same question.
The 2025 BJPsych Open meta-analysis pooled 15 trials and 873 adults and found cortisol, perceived stress and anxiety all improved — with quality of life specifically not improved, and with a stress interval wide enough to show the trials were not agreeing.¹
A. A. Albalawi, publishing in Nutrition and Health the same year, reached the split verdict in the title: significant cortisol reduction, no effect on perceived stress.² Two reviews, one year, overlapping trials, opposite conclusions on the outcome that matters.
Helgi K. Björnsson (Google Scholar) and colleagues, working from Icelandic records and the US Drug-Induced Liver Injury Network, documented the liver-injury cases that regulators in Denmark and France subsequently acted on.⁵⁴ That is the part of the file the shelf does not carry.
The honest summary of the convergence: everyone agrees on the assay, nobody agrees on the person, and the safety file is thin but not empty.
What this actually changes
Not whether you take it. That is your decision and this review has no standing to make it.
What it changes is what you expect for the money. If the most likely outcome is a measurable hormone going down while your month feels identical, then the capsule is not competing with nothing — it is competing with the ferritin test you have not booked and the sleep review nobody has offered. Those have worse packaging and considerably better odds.
It also changes how you read the next bottle, and there will be a next one; the aisle refreshes roughly every eighteen months. The three questions in the one-move box below work on all of them.
When this belongs with a doctor
Ask a pharmacist or clinician before starting, not after, if any of the following apply: you take thyroid medication, sedatives, immunosuppressants, or medication for diabetes or blood pressure; you have a liver condition; you have an autoimmune or endocrine condition; you are pregnant or breastfeeding, where the advice is to avoid it.⁴ A pharmacist will check interactions in about four minutes and will not need an appointment.
Stop and seek medical advice the same day if you develop yellowing of the skin or the whites of the eyes, dark urine, pale stools, persistent nausea, unusual itching, or pain under the right ribs. These are the signs the reported liver-injury cases presented with, and they warrant prompt attention rather than a wait-and-see.
And the more common situation: exhaustion that has lasted months, with or without a supplement, is worth a conversation and a basic panel.
What to say at the counter, and what you will probably hear back
Two conversations, and the first one is free. Both short enough to read off a phone screen.
Short enough to read off a phone screen
At the pharmacy counter, say: “I am considering ashwagandha. Could you check it against what I already take, and tell me whether my liver history is a reason not to?”
If you hear “it’s natural, it’s fine”: “There are documented liver-injury cases and a Danish restriction. I would still like the interaction check.”
At the appointment, say: “I have been flat and exhausted for months and I have been buying supplements for it. I would rather check ferritin, thyroid and B12, and ask about sleep, before I buy another one.”
If what you actually want is an honest account of what the last six months have contained, before deciding whether the answer lives in a bottle, the Quiet Audit is ten private minutes on exactly that. Nobody sees your answers. Start here.
The one move
The one move
Before buying any supplement sold for stress, ask three questions in order. Which outcome improved — a measurement, or how people felt? Who was in the trial: mean age, sex, country? Is the extract in the study the extract in the bottle, named, with the withanolide percentage on the label?
Four minutes, before any money changes hands. Most of the aisle fails at question three.
If you are already taking it and it is helping, this review is not an instruction to stop. Placebo effects are real effects, twelve weeks is the outer edge of the studied duration, and the safety signal is small — check the interaction list with a pharmacist and get on with it.
Questions this review is asked
Does ashwagandha actually work for stress?
It reliably lowers measured cortisol in pooled trials.¹² Whether it changes how stressed people feel is contested: one 2025 meta-analysis of 15 trials and 873 adults found perceived stress fell, another 2025 review of overlapping literature found no effect on that outcome.¹²
How long does it take to work?
Most trials ran 6 to 12 weeks, with anxiety outcomes typically measured at 8 weeks.¹⁴ There is very little evidence about what happens beyond three months, because almost no trial has run longer than that.
Is ashwagandha safe for the liver?
At least six published cases of liver injury are documented, at doses of 450 to 1,350 mg a day with onset from one week to four months.⁴ A case series cannot establish how often this happens. Regulatory action in Denmark and a 2024 French advisory indicate the signal is being taken seriously.⁴
Does it work for women in their fifties?
The one trial with a mean age of 54 to 55 — 120 adults over 12 weeks — found no reduction in stress, though fatigue improved.³ Most positive trials had mean ages in the twenties and thirties. Applying those results to a woman of 52 is extrapolation.
Should I take it with my thyroid medication?
Ask a pharmacist first. Ashwagandha is pharmacologically active and interactions with thyroid, sedative, immunosuppressant and diabetes medication are plausible.⁴ This is a four-minute conversation at a counter and does not need an appointment.
Informational, not medical advice. Supplements interact with prescribed medication; a pharmacist or clinician is the right person to ask about your own situation.
Where to go next
In order, and each for a reason.
1. What a normal blood panel does and does not measure — because four of the five conditions above live there, and the panel is free.
2. Then what a cortisol test can and cannot tell you, since the assay this supplement moves is the same one those kits sell.
3. Then telling a life cause from a medical one, which is the question underneath the whole aisle.
The capsule was never the problem. It was just never going to be the answer either.
References
1. Effects of Ashwagandha Supplements on Cortisol, Stress, and Anxiety Levels in Adults: A Systematic Review and Meta-Analysis (2025). BJPsych Open. Read the meta-analysis
2. Albalawi, A. A. (2025). Dual impact of Ashwagandha: significant cortisol reduction but no effects on perceived stress — a systematic review and meta-analysis. Nutrition and Health. Read the review
3. Smith, S. J., Lopresti, A. L., & Fairchild, T. J. (2023). Exploring the efficacy and safety of a novel standardized ashwagandha (Withania somnifera) root extract in adults experiencing high stress and fatigue in a randomized, double-blind, placebo-controlled trial. Journal of Psychopharmacology. Read the trial
4. National Institutes of Health, Office of Dietary Supplements. Ashwagandha: Is it helpful for stress, anxiety, or sleep? Health Professional Fact Sheet. Read the fact sheet
5. Björnsson, H. K., et al. Liver injury due to ashwagandha: a case series from Iceland and the US Drug-Induced Liver Injury Network. Read the case series
On the researcher. Adrian Lopresti is a clinical psychologist and trials researcher, not your clinician, and nothing in his published work is medical advice for an individual. The 2023 trial tested a named commercial extract and the paper carries its own funding and competing-interest declarations, which are printed with it and worth reading at source. We flag that rather than hide it: a trial of a branded stress extract that reports no effect on stress is a harder result to dismiss, not an easier one.
Sourcing, disclosure and disclaimers
This is not medical advice. Blue Leaf Journal publishes general information for a general readership. Nothing here is a diagnosis, a treatment recommendation, or a substitute for care from a clinician who knows your history. Decisions about supplements and medication belong with your pharmacist or doctor.
We are not clinicians. Blue Leaf Journal is an independent publication. We read published research and translate it. The findings belong to the researchers and institutions named and linked above; the plain-English rendering is ours, and so is any error in it.
No affiliation and no endorsement. Blue Leaf Journal is not affiliated with Adrian Lopresti, Stephen Smith, Timothy Fairchild, Helgi Björnsson, Murdoch University, Clinical Research Australia, the National Institutes of Health, ANSES or any supplement manufacturer. None of them has reviewed, approved or endorsed this article, and none of them is responsible for it. They are cited because their published work is the evidence for what it says.
No commercial relationship. Nobody named above paid for or was paid for this coverage. There are no affiliate links, no sponsored placements and no gifted products in this article. We do not sell supplements, and we do not earn anything if you buy one.
How this was checked. Every figure above is drawn from the primary sources, each linked in the references, and can be verified there. Where two reviews disagree, both are reported rather than the more flattering one. Sources were checked on 27 August 2026. If you find something we have got wrong, write to miriamalderton@blueleafjournal.com and we will correct it and say that we did.
Miriam Alderton is Research Editor at Blue Leaf Journal. She reads the methods section first and the abstract last, and every figure in this piece is linked to its source above.
Written by Miriam Alderton, Research Editor, for Blue Leaf Journal. Updated: 27 August 2026.
