Ashwagandha for anxiety: what the clinical trials actually show

What ashwagandha actually is

Ashwagandha (Withania somnifera) is a shrub used in Ayurvedic medicine for thousands of years. The part that carries the activity is the root, and the active fraction is a family of steroidal lactones called withanolides. Modern supplements are sold as standardized root extracts — the two dominant commercial ones are KSM-66 and Sensoril — and that word, standardized, is doing a lot of work. A bargain-bin raw powder, a full-spectrum root extract, and a leaf-heavy extract are not the same thing once they hit your bloodstream, and the trials were run on the standardized versions, not the cheapest tub on the shelf.

Anxiety and stress are where ashwagandha’s largest evidence base sits, which is why it gets sold as a calm-down adaptogen in the first place. This piece narrows the lens to one question: does it reduce anxiety specifically, and what do the trials that measured anxiety actually report? If your question is the broader stress-and-cortisol story, we have a separate fully cited read on ashwagandha for cortisol and stress, and if it is sleep, on ashwagandha for sleep. The rest of the anxiety and mood coverage puts it next to the other options.

Why it touches anxiety at all

The plain-language version: ashwagandha’s best-replicated effect is that it turns down the body’s stress output. Anxiety, for a lot of people, is that stress signal stuck in the “on” position — a racing, keyed-up arousal that does not switch off. Ashwagandha appears to lower that signal rather than sedate you over the top of it, which fits the slow, build-over-weeks pattern the trials report instead of an instant hit.

Underneath that, two mechanisms are usually invoked, and this is the section where the technical terms belong. The first is the stress axis — the hypothalamic-pituitary-adrenal (HPA) axis, the loop that ends in the adrenal glands releasing cortisol, the body’s main stress hormone. Ashwagandha’s most consistent biological finding is a reduction in cortisol: the Chandrasekhar 2012 trial measured a 27.9% drop in serum cortisol versus 7.9% on placebo (p = 0.002), and the Lopresti 2019 trial found a 23% morning-cortisol reduction versus a slight rise on placebo (p < 0.001).¹² The second is the inhibitory GABA system — gamma-aminobutyric acid, the brain’s main “slow down” signal. Preclinical work suggests withanolides have GABA-mimetic activity, which would explain a calming effect, but that line is largely animal-derived, so treat it as mechanism-plausible, not trial-proven in humans. The honest summary: the signal it pulls is mostly a downshift in the stress axis, with a plausible but less-proven nudge to the brain’s calming system on top.

The anxiety trials, read honestly

Start with the pooled picture, because it washes out the noise of any single small study. A 2025 systematic review and meta-analysis in BJPsych Open gathered 15 randomized controlled trials totaling 873 adults with stress and/or anxiety.⁴ On the Hamilton Anxiety Rating Scale (HAM-A) — the standard clinician-rated anxiety measure — ashwagandha beat placebo, and the effect grew with time: a pooled mean difference of about −1.55 points early on, widening to −3.52 points (95% confidence interval −6.00 to −1.04, p = 0.0053) by eight weeks.⁴ Perceived stress fell too — about −4.9 points on the Perceived Stress Scale at eight weeks — alongside the cortisol reduction.⁴ That is a consistent, statistically significant anxiolytic signal across a real body of trials.

Underneath the pooled number sit the individual RCTs, and they tell the same story. The foundational one is Chandrasekhar 2012: 64 chronically stressed adults randomized to 300 mg of a full-spectrum KSM-66 root extract twice daily or placebo for 60 days.¹ The ashwagandha group’s Perceived Stress Scale scores fell 44.0% versus 5.5% on placebo (p < 0.0001) — and, notably, this is one of the few trials with no declared industry funding.¹ Lopresti 2019 used a different standardized extract (Shoden, 240 mg once daily) in 60 stressed adults over 60 days and found a 41% HAM-A reduction versus 24% on placebo (p = 0.040).² Langade 2019 randomized 58 people with insomnia and anxiety to 300 mg of KSM-66 twice daily or placebo for ten weeks; HAM-A fell significantly more on ashwagandha (from 23.6 to 18.5, versus 23.4 to 21.5 on placebo; p = 0.002).³

Now the part the marketing leaves out, and it is not a footnote. These trials are small — mostly 40 to 80 people. They are short — six to ten weeks, with essentially no data past a couple of months. They use different extracts at different doses, which is exactly why the 2025 meta-analysis and earlier reviews flag substantial heterogeneity: the trials do not all agree, and pooling chemically different products papers over real differences. And a large share are funded or supplied by the extract manufacturers — Lopresti’s extract came from Arjuna Natural; Langade’s KSM-66 was supplied by Ixoreal Biomed.²³ Industry funding does not automatically make a result wrong, but it is a known source of bias, and it means independent, well-powered, long-duration replication is the missing ingredient. The fair verdict: this is one of the more promising natural options for anxiety, on moderate-quality evidence — not a proven drug-equivalent.

Dose, timing, and the extract problem

We do not hand out prescriptive doses here, but the trial protocol is worth stating plainly because it is remarkably consistent. Across the anxiety and stress studies, the schedule was a standardized root extract at 300 mg, taken once or twice daily — 300 to 600 mg total — for six to ten weeks.¹²³ Unlike a benzodiazepine, this is not a take-it-when-anxious rescue dose; the benefit accrued over weeks of steady daily intake, consistent with a stress-axis effect rather than an acute sedative one. If you want to reason about how long a daily compound takes to reach steady levels in the body, our half-life tool walks through the logic.

Here is the catch the marketing glosses over. “Ashwagandha” is not one product. KSM-66 is a root-only extract standardized one way; Sensoril is a root-and-leaf extract standardized to a higher withanolide content; Shoden is a third. They have different chemical fingerprints, different studied doses, and different intended profiles. A trial that validates one does not automatically validate the others, and none of them validates an unstandardized powder. When a label says “clinically studied,” the fair question is always: studied as which extract, at what dose, for how long — because the answer changes whether the evidence applies to what is in your hand.

Ashwagandha turns the stress signal down rather than sedating you over the top of it. That is why the effect builds over weeks instead of hitting in an hour — and why it is not a rescue pill for a panic spike.

The safety flags the labels skip

This is where the honest version of ashwagandha diverges hardest from the wellness-aisle version, and it matters more for an anxiety supplement than for occasional use, because managing anxiety means daily, chronic dosing — the exact use case where these signals show up.

The liver. The clearest concern is herb-induced liver injury. The NIH’s LiverTox database rates ashwagandha a likely cause of clinically apparent liver injury, with roughly two dozen reported cases.⁵ The pattern is consistent: a cholestatic or mixed injury with jaundice, showing up two to twelve weeks after starting — exactly the window a daily anxiety user would be in — usually resolving after stopping, but with rare cases of fatal liver failure or transplantation, especially in people with preexisting liver disease.⁵ A 2023 review applied the updated RUCAM causality method — the standard tool for attributing liver damage to a specific agent — and scored new ashwagandha cases “probable,” noting a prior case that ended in liver transplantation.⁶ This is uncommon, but it is real and it is documented.

The thyroid. Ashwagandha can raise circulating thyroid hormone levels. The cleanest documented signal is a published case of thyrotoxicosis — a dangerous excess of thyroid hormone — in an otherwise healthy woman taking ashwagandha; her labs and symptoms resolved after she stopped.⁷ For most people this is a non-event, but if you have a thyroid condition or take thyroid medication, ashwagandha is not neutral.

Pregnancy and drug interactions. Ashwagandha is traditionally considered an abortifacient and is contraindicated in pregnancy — a hard stop, not a caution.⁸ Because it can have a sedative and immune-modulating effect, it can also stack with or interfere with sedatives and benzodiazepines, thyroid medication, and immunosuppressant drugs.⁸ If you are on any of those, this is a conversation with your prescriber, not a self-experiment.

Grey areas — including the medication question

“It works as well as anxiety meds.” This is the claim to be most careful with. There is no robust, well-powered human head-to-head trial comparing ashwagandha against an SSRI or a benzodiazepine for an anxiety disorder. The “comparable to medication” line traces mostly to animal models — rodent maze studies showing ashwagandha matching lorazepam — and to small, underpowered human work. Animal anxiolytic data and a placebo-controlled symptom-score reduction are real, but they are not the same as proving equivalence to a prescription drug in people with a diagnosis. Treat “as good as medication” as unproven.

Anxiety is a diagnosis, not a vibe. The trials enrolled people with stress and subclinical-to-moderate anxiety, not, for the most part, people with severe diagnosed anxiety disorders. Clinical anxiety frequently rides alongside depression, trauma, or a medical driver, and the evidence-based first-line treatments are therapy (notably cognitive behavioral therapy) and, where indicated, prescription medication. If anxiety is disrupting your life most days for weeks, that is a clinician conversation — ashwagandha is not a substitute for one.

Extract non-equivalence. Because KSM-66, Sensoril, and Shoden are chemically distinct, the evidence base is fragmented across products that are not interchangeable. “Ashwagandha reduces anxiety” is really “these particular extracts, at these doses, in these populations, reduced anxiety scores.” That is a meaningfully narrower claim.

Where it fits: a tiered view

It helps to place ashwagandha honestly on a spectrum of how settled the evidence is and who it is for.

Foundational — fix the inputs first. No supplement out-competes the basics for everyday anxiety: sleep, regular movement, daylight, limiting caffeine and alcohol, and addressing the actual stressors. If those are a mess, that is the higher-yield lever, every time — and most of it is free.

Research-curious — the targeted trial-of-one. If your foundations are solid and you have ongoing, stress-driven, keyed-up anxiety, ashwagandha is one of the better-evidenced botanical options, with several RCTs and meta-analyses behind it.⁴ Judge it over the six-to-ten-week window the trials used, with a known standardized extract, and only if you fall into none of the safety-flag groups.

Experimental — treating it as a fix for a diagnosed disorder. Using ashwagandha to self-manage clinical anxiety, or leaning on it indefinitely in place of treatment, is the weakest-supported and riskiest use — long-term safety data simply are not there, and the liver signal grows more relevant the longer you dose.⁵

What we still don’t know

No long-term data. The trials run weeks, not years. For a compound with a documented liver signal, the absence of long-duration daily-use safety data is the gap that matters most — and it is why indefinite use sits in the experimental tier.⁵

No independent, well-powered replication. The effect is consistent but the trials are small and frequently industry-affiliated. A large, independent, multi-site RCT in people with a formal anxiety-disorder diagnosis is the study the field is missing.⁴

No head-to-head against standard treatment. Until ashwagandha is tested directly against an SSRI or therapy in diagnosed anxiety, “comparable to medication” stays a marketing claim, not a finding.

Population gaps. The evidence skews toward general stressed adults. There is little to nothing on adolescents, older adults, or pregnant and breastfeeding people — for whom the answer is already “avoid.”

What this article is not saying

This is not “ashwagandha doesn’t work.” The trials and the 2025 meta-analysis point the same direction: a real, statistically significant reduction in anxiety scores versus placebo, at 300 to 600 mg over six to ten weeks.¹⁴ Dismissing it outright is as wrong as overselling it.

This is not “ashwagandha will cure your anxiety.” The effects are moderate, the trials are small and short and partly affiliated, the extracts are not interchangeable, and there is no large independent long-term replication and no head-to-head against real treatment. It is a slow-building, stress-mediated nudge — not a sedative, and not a substitute for care.

And this is not “ashwagandha is risk-free.” It carries documented liver, thyroid, pregnancy, and drug-interaction concerns that a daily habit makes more relevant, not less. The point of this piece is to tell you what the anxiety trials show and where they stop — and what to weigh before you make it a daily thing. If your question is really cortisol and stress, read the companion piece on ashwagandha for cortisol and stress; if it is sleep, the sleep read.