Is ashwagandha bad for your liver?

Disclosure up front: I run a supplement company, so I have a stake in how people see this category. It makes functional-mushroom products and doesn't sell ashwagandha. The full disclosure is at the bottom. I covered the stress evidence in Report 033. This one is only about safety.

The alarming evidence

The largest single-ingredient case series I found is by Cyriac A. Philips and colleagues, published in Hepatology Communications in 2023. Three Indian hospitals searched their records from January 2019 to December 2022 and found 23 patients with liver injury linked to ashwagandha products. The team then threw out 15 who had taken multi-herb formulas, other risky products, or had other possible causes. Eight remained.

Of those eight, six had severe injury, and three died. All three deaths were in patients who arrived with acute-on-chronic liver failure, meaning a sudden collapse on top of an already damaged liver.

Two details stand out to me. First, five of the eight had chronic liver disease, and for two of them it was only discovered while they were being worked up for this injury. They didn't know they were in the higher-risk group. Second, the team chemically analyzed five recovered products and found no adulterants or known liver toxins. One had trace lead, cadmium and arsenic. So in this series, "it was probably a contaminated bottle" is not an easy way out, though only five products could be tested.

This isn't a one-group finding. A series of five cases from Iceland and the U.S. Drug-Induced Liver Injury Network, published in 2020, described a similar picture: jaundice 2 to 12 weeks after starting, often with itching, in a cholestatic or mixed pattern. Cholestatic means bile flow is the main problem. Nobody in that series went into liver failure, and liver tests normalized in four of the five within 1 to 5 months (the fifth was lost to follow-up). Testing confirmed ashwagandha in the available products and found no other toxic compounds, though four of the five patients were also taking other supplements, and in one case rhodiola was a possible co-culprit. A 2026 systematic review pulled together 26 patients from 19 reports, with jaundice in 88.5%, and noted that severe cases needed intensive care or a transplant.

The reassuring evidence

Now the other side. Ketan Pakhale and colleagues published a randomized, double-blind safety trial in Phytotherapy Research in 2026. Across sites in India, the U.S., Australia, Portugal and Africa (about 80% from India), 1,002 adults with stress and anxiety took 600 mg a day of KSM-66 root extract or a placebo for 8 weeks.

There were no serious adverse events. Average AST and ALT, two standard liver enzymes, stayed normal and looked alike in both arms. At week 8, AST averaged 25.91 IU/L on ashwagandha and 26.33 on placebo. The authors report no clinically significant lab change in any participant, and fewer side effects on ashwagandha than placebo (5.6% vs 9.2%). The study was supported by Shri Kartikeya Pharma, the extract is made by Ixoreal Biomed, and the authors declare no conflicts of interest.

That's a big trial by supplement standards, and I give it real weight. It's strong evidence that this extract, at this dose, doesn't commonly stress the liver in healthy-ish adults over two months. If ashwagandha routinely damaged livers, this trial would have shown it.

Why both can be true

The two kinds of study answer different questions, and the trial's limits are worth spelling out.

It's too small for a rare event. There's a statistical rule of thumb called the rule of three: if nothing happens in N people, the true rate could still be as high as about 3 in N, with 95% confidence. For 498 people on ashwagandha, that's roughly 0.6%, or 1 in 170. So the trial rules out something common. It doesn't rule out something rare. By my arithmetic, if the true risk were 1 in 5,000, a trial this size would see zero cases about 90% of the time. Rare, idiosyncratic reactions are exactly how drug-induced liver injury tends to behave, and they often only surface after a product is on the market.

It enrolled the wrong people to find this. The trial excluded anyone with clinically significant metabolic or systemic disease, or a history of alcohol or substance abuse. The exclusion list doesn't name liver disease specifically, but it's hard to picture someone with cirrhosis getting through. That's the group where the deaths clustered.

It stopped at 8 weeks. That catches most of the reported window, but not all of it. In the Philips series, median time to symptoms was 41 days, ranging from 14 to 540. The authors list the short duration as a limitation.

It checked two enzymes. The paper reports AST and ALT as its liver assessment and calls them sufficient. Those are good markers of liver-cell damage. But the reported injuries are often cholestatic, where alkaline phosphatase and bilirubin are the more telling tests, and neither appears in the results. That's not a blind spot, since jaundice would show up as an adverse event and the Philips team describes at least mild enzyme rises in its patients. But it isn't the panel I'd pick to hunt this particular problem. Lab results were also available for 474 of the 498 people on ashwagandha, not all of them.

It tested one product. The trial used one standardized root extract. The Philips cases took homemade powder, herbal jam, detox tea, tablets and syrups, some at 10 to 20 grams a day. So you can't assume the series and the trial describe the same exposure. That cuts both ways, though: the published cases also include people taking ordinary label doses of commercial capsules, and the Philips review found overdose in only one earlier report.

The case reports have their own limit, and it's just as important. They have no denominator. Eight patients tells you this can happen, not how often. Three deaths in eight does not mean a 37.5% death rate among ashwagandha users. It means that among people sick enough to land in a liver unit and get written up, three died. The 2026 review says as much: the evidence can't estimate true incidence, and severe cases are more likely to get published. Causality is also a judgment call. The Philips team scored six cases "probable" and two "possible" on the standard RUCAM scale, and none "highly probable." No one has rechallenged these patients to prove it, nor should they.

Where the best neutral source lands

The NIH's LiverTox database, last updated for ashwagandha in December 2024, is the closest thing to a referee. It gives ashwagandha a likelihood score of B, a likely cause of clinically apparent liver injury. It says clinical trials haven't shown enzyme elevations, that the injury occurs but rarely, that most cases are mild to moderate and resolve within 1 to 4 months of stopping, and that fatal cases have been described, particularly with pre-existing liver disease. Its mechanism section is honest too: the cause isn't known, with withanolides the main suspects. Its advice on who should stay away is the plainest sentence in this whole literature: its use "should be avoided in patients with cirrhosis or advanced chronic liver disease."

How I'd read it

If you're a healthy adult taking a standardized extract at a normal dose, the best evidence says serious liver injury is unusual. I can't give you a number, and neither can anyone else. I'd also be skeptical of anyone who says "the trial proved it's safe" or "this herb destroys livers." Neither study can carry that claim.

If you have cirrhosis or any known liver disease, I'd take LiverTox at its word. The fatty liver a lot of people have and don't know about is a reason to mention the supplement to your doctor, and the Philips series is a reminder that some patients only learned about their liver disease in the hospital. This is a clinician conversation, not something to settle from a blog.

And regardless of who you are, know the signs. The reported cases typically show up as yellowing skin or eyes and intense itching, sometimes with nausea, fatigue or abdominal discomfort, usually within weeks of starting. If that happens, stop taking it and get seen promptly. Don't restart it: LiverTox specifically warns against rechallenge. In the cases I read, stopping was the main treatment, and most people without prior liver disease recovered.

The reusable lesson: a clean trial tells you a harm isn't common, and a scary case series tells you a harm is possible. To learn how often something rare happens, you need a very large denominator or real post-market surveillance, and for most supplements nobody has built either one.

Not medical advice. This is educational analysis, not a recommendation — a study is not a prescription. Talk to a qualified clinician before acting on anything you read here. Full disclaimer →

Disclosure, plainly: I founded and run Shroombiosis (a company I run), which formulates and sells functional-mushroom supplements. It doesn't sell ashwagandha, but I have a commercial stake in the supplement category generally. I have no relationship with the authors of these papers or with any ashwagandha maker. Nothing here is sponsored and no link earns a commission; here's the full policy. A recommendation with no stake at all: for performance nutrition, Die Tryin Co. is a fellow combat-veteran-owned brand I'm glad to point people to. I don't own it and earn nothing from the link.