Ashwagandha: Stress Evidence and the Liver Flag

Ashwagandha is the most widely used Ayurvedic herb in the West, and its stress evidence is genuinely better than most of the shelf beside it — small randomized trials showing measurable cortisol reductions. The same literature now carries a second, quieter story: a growing case series of liver injuries serious enough to have drawn a regulator safety alert. This page holds both halves at once: a real candidate with a real caution label.

What Ashwagandha Is

Withania somnifera is a small shrub whose root has been used in Ayurvedic practice for centuries; "somnifera" nods to its traditional sleep associations. The modern supplement is a standardized root extract whose active family — withanolides — is thought to act on the stress-response system. Traditional Ayurvedic preparations used root powder by the gram; modern products are standardized extracts at 300–600 mg, and the two are different animals when case reports and trials disagree. Plant part is a live issue too: NIH notes products may contain root, leaf, or both, and leaf carries more than twenty times as much withaferin A, the withanolide reviews flag as the more toxic one.

The Stress Evidence

The trial most often cited — and most often copied into marketing — is Chandrasekhar and colleagues' randomized study of 64 chronically stressed adults (Indian Journal of Psychological Medicine, 2012). Over 60 days, the group taking 300 mg of a root-only KSM-66 extract twice daily reported large drops in perceived stress, and their average serum cortisol fell 27.9 percent — against 7.9 percent in the placebo group. The trial was double-blind and placebo-controlled, which is why it carries weight. The extract was supplied by its manufacturer, and the trial was small, brief, and in one population — which is why it carries the weight of a promising signal, not a settled conclusion.

Similar trials have followed, mostly in the 300–600 mg range of standardized extracts, mostly finding the same direction: modest reductions in stress scores, some cortisol movement, occasional anxiety and sleep improvements. Replication in the same direction counts for something real. But the base remains small trials, often industry-backed, rarely longer than two months — the profile of a second-tier candidate, which is exactly where this page files it. The stress system itself is covered honestly in the Cortisol 101 pillar, and the behavioral route — which carries no liver flag — is the Meditation protocol.

Sleep and Anxiety: The Smaller Signals

Marketing copies the stress findings onto sleep and anxiety claims, and the evidence there is thinner. Sleep trials exist — small ones, in older adults and in stressed workers, generally showing modest improvements in sleep-quality scores. Anxiety trials are smaller still. If ashwagandha interests you for sleep specifically, the honest comparison is against the behavioral alternatives: the sleep evidence hierarchy runs through stimulus control and CBT-I before any bottle, as When Sleep Won't Come lays out. Ashwagandha is at best a footnote in that hierarchy today.

A note on how these secondary claims grow: a stress trial measures a stress score, and no ashwagandha trial has measured anything downstream — sick days, doctor visits, anything you would call an outcome. Marketing that connects the cortisol arrow to "aging" or "adrenal health" is borrowing plausibility from the stress system's importance, not from trial data. The methodology page calls this the biomarker trap: moving a number is not the same as changing a life. Cortisol fell in these trials; what that buys you over years is exactly the question nobody has answered.

The Liver Flag

Since 2017, journals have accumulated case reports of liver injury in people whose only new exposure was an ashwagandha supplement. NIH's LiverTox database, which rates the herb a likelihood-B cause of clinically apparent liver injury, describes onset two to twelve weeks after starting it, usually with a cholestatic or mixed pattern, jaundice, and itching. A 2026 systematic review in the Egyptian Journal of Internal Medicine gathered 19 studies covering 26 patients: mostly young-to-middle-aged adults, taking ordinary doses for stress, sleep, or fertility. Nearly nine in ten developed jaundice, and RUCAM causality scoring rated the link probable or possible in most cases; most recovered after stopping. A 2026 Cureus scoping review of 13 studies and 25 patients found the same syndrome, plus one patient who needed a liver transplant and three deaths among people who already had cirrhosis — a group the review's authors and LiverTox both say should avoid the herb outright. And in February 2024, Australia's Therapeutic Goods Administration issued a safety alert after 12 reports of liver problems, four requiring hospitalisation.

The honest framing is two-sided. On one hand, national survey data put recent use at roughly 0.4 percent of US adults — on the order of a million people — against documented injuries numbering in the dozens: rare, idiosyncratic, and impossible to predict in advance. On the other, rare does not mean resolved: regulators logged further reports after the 2024 alert, the pattern recurred across countries, and "herbal" softens none of it. A supplement that occasionally hospitalizes healthy users has earned a caution label, whatever its stress numbers say.

The Watch-List Before You Take It

• 🩺 Liver disease or heavy alcohol use — the clearest avoid-group: the reported deaths were in people who already had cirrhosis, and LiverTox advises against the herb in cirrhosis or advanced chronic liver disease. Anyone with liver conditions should not take ashwagandha without a clinician's sign-off, and a suspected injury means never restarting the product.
• 🦋 Thyroid and autoimmune conditions — ashwagandha can push thyroid hormone output up: a placebo-controlled trial in subclinical hypothyroidism found it lowered TSH and raised T3 and T4, and a published case describes thyrotoxicosis in a woman with Hashimoto's. NIH advises against use in thyroid or autoimmune disorders, so anyone on thyroid medication, or with autoimmune thyroid disease, needs a clinician in the loop before use.
• 😴 Sedatives and CNS depressants — the herb's calming effects can stack with sleep aids, benzodiazepines, and alcohol; combining them is a pharmacist question.
• 🤰 Pregnancy and breastfeeding — avoid both; the abortifacient concern is traditional rather than trial-confirmed, and human safety data are absent.
• 💊 Other medications — NIH lists possible interactions with diabetes and blood-pressure medicines, immunosuppressants, sedatives, anticonvulsants, and thyroid hormone; in-vitro studies disagree on CYP3A4, and no clinical interaction is documented. Clear it with the prescriber.
• 🏷️ Product quality — root-only extracts from third-party-tested brands; leaf material carries far more of the withanolide withaferin A, which reviews flag as the more toxic one.

What the Trials Actually Used

The stress literature is a root-extract literature: 300–600 mg daily of standardized extract, typically for 8–12 weeks — a different preparation from raw root powder, and different branded extracts carry different withanolide profiles. Practical reading: if you and a clinician decide to trial it, use the form, dose, and duration the evidence actually tested, from a brand with third-party testing, and treat it as a time-boxed experiment. The Big Five framework exists to keep the supplement shelf short and evidence-anchored.

Where It Fits in a Real Plan

If the stress system is your target, the evidence ladder is not close. Exercise, sleep regularity, and time-boxed stress practices carry deeper literatures than any adaptogen — the Meditation protocol and the Cortisol 101 pillar are the honest first rungs. Ashwagandha sits lower, with a genuine signal: reasonable for a healthy adult to trial after the behavioral layers are in place, eyes open about the funding base of its evidence and the liver case series in the appendix of every positive trial. Anyone with liver, thyroid, or pregnancy considerations reads that section first, and the stop-symptoms callout is worth more than every marketing sentence combined.

Questions, Answered Briefly

• 🫀 "How likely is the liver injury?" Rare — dozens of cases against millions of users — but unpredictable and occasionally severe. Nobody can compute your individual risk; that asymmetry is the caution label.
• 🧪 "Should I get liver tests while taking it?" There is no trial-tested monitoring schedule; that question belongs to your clinician, and baseline labs are a reasonable ask.
• 🔁 "Is cycling it safer?" Unknown — the case reports do not follow a clean dose-or-duration pattern. Cycling is folk risk-management, not evidence.

Related Topics