Ashwagandha: What the Research Actually Says Before You Try It

A 20-year-old man in Germany's LiverTox database started ashwagandha for "troubled thoughts." Within four weeks his bilirubin had climbed to 17.3 mg/dL and his skin had turned yellow. He recovered after stopping the supplement, five months later. That case sits alongside more than 20 others collected by the U.S. National Institutes of Health, and it is the other half of the ashwagandha story that rarely makes it onto a product page next to the calming stock photo.

Ashwagandha (Withania somnifera) is one of the best-selling adaptogens on the market, promoted for stress, sleep, testosterone, and general resilience. Some of that promotion is backed by real trial data. Some of it is marketing copy stretched well past what the evidence supports. This guide separates the two, using the U.S. National Center for Complementary and Integrative Health (NCCIH), the NIH's LiverTox registry, New Zealand's Medsafe, and the published randomized trials themselves, so a shopper can make an informed decision instead of a hopeful one.

Key Takeaways

• Ashwagandha has real trial support for two things: sleep and self-reported stress or anxiety symptoms. Evidence for testosterone, athletic performance, cognition, and most other uses is thin or mixed, according to NCCIH.
• The most-cited stress trial (Chandrasekhar et al., 2012) used 300 mg of a full-spectrum root extract twice daily for 60 days and found a statistically significant drop in Perceived Stress Scale scores and serum cortisol versus placebo.
• Rare but real liver injury has been reported in more than 20 published cases, usually appearing two to twelve weeks after starting the supplement, according to the NIH LiverTox registry updated December 2024.
• Ashwagandha is not recommended during pregnancy or breastfeeding, before surgery, or for people with autoimmune or thyroid disorders, per NCCIH's health professional fact sheet.
• In New Zealand, ashwagandha supplements are not pre-approved by a regulator before sale. Medsafe places legal responsibility for label accuracy on the company selling the product, not on a government reviewer.
• "Withania somnifera," "Indian ginseng," and "ashwagandha" refer to the same plant; the extract type, standardization method, and dose vary widely between brands and matter more than the plant name on the front label.

What Ashwagandha Actually Is

Ashwagandha is a low evergreen shrub native to India, parts of Africa, and the Middle East, and it has been used in Ayurvedic medicine for centuries. The part typically sold as a supplement is the root, sometimes combined with leaf material, standardized to a group of steroidal compounds called withanolides. Different manufacturers standardize to different withanolide percentages, use different solvent extraction methods, and land on different daily doses, which is why "ashwagandha" as a category covers products that are not interchangeable in how they were tested.

NCCIH describes it plainly: many clinical trials exist, but a lot of them are small and use different preparations, which makes it hard to generalize results from one branded extract to every bottle labeled "ashwagandha" on a shelf.

What the Evidence Actually Shows: Stress and Sleep

The trial most frequently cited by supplement marketers is Chandrasekhar, Kapoor, and Anishetty's 2012 study in the Indian Journal of Psychological Medicine. It enrolled 64 adults with a history of chronic stress in a single-center, randomized, double-blind, placebo-controlled design. The treatment group took 300 mg of a high-concentration full-spectrum ashwagandha root extract twice a day for 60 days. By day 60, the treatment group showed a statistically significant reduction (P<0.0001) across every stress-assessment scale used in the study, plus a significant drop in serum cortisol (P=0.0006) compared to placebo. Adverse effects were mild and similar between groups, with no serious events reported.

A second trial, Salve et al. 2019, published in the journal Medicine, ran a 60-day randomized, double-blind, placebo-controlled study in 60 stressed adults, using 240 mg per day of a different standardized extract (Shoden). The ashwagandha group had a statistically significant reduction on the Hamilton Anxiety Rating Scale (P=.040), a near-significant reduction on the DASS-21 stress subscale (P=.096), and significantly greater reductions in morning cortisol (P<.001) and DHEA-S (P=.004) than placebo. All participants completed the trial with no adverse events reported. The authors were careful to flag their own limitation: the sample was small, from one clinical and cultural population, and larger trials at varying doses are still needed.

Ashwagandha may be safe when taken in the short term (up to 3 months). There is not enough information to allow conclusions about its long-term safety to be reached. — U.S. National Center for Complementary and Integrative Health, "Ashwagandha: Usefulness and Safety" fact sheet, last updated March 2023

For sleep specifically, NCCIH points to a 2021 systematic review and meta-analysis by Cheah and colleagues in PLoS One that pooled five studies on ashwagandha and sleep, all conducted in India. NCCIH's own summary of that body of research states that some ashwagandha preparations may be effective for insomnia, making sleep the outcome with the most consistent supporting data compared to the more mixed picture for anxiety specifically. That distinction matters for anyone shopping by claim on a label: "helps you sleep" currently has firmer backing than "cures anxiety."

Testosterone and Fertility: Limited, Not Proven

NCCIH notes there is some limited evidence that taking ashwagandha for two to four months may increase testosterone levels and sperm quality in certain populations, but the agency stops short of calling this established. A separate systematic review (Smith et al. 2021, Advances in Nutrition) examined herbs and testosterone concentrations broadly and found the picture across studies inconsistent. Anyone buying ashwagandha specifically for fertility or hormone support should treat this as an open question worth discussing with a doctor rather than a settled outcome worth building a routine around.

Extract Types, Blends, and How Ashwagandha Compares

Not every ashwagandha capsule is made the same way, and the differences go beyond marketing language. Root-only extracts, root-and-leaf blends, and water extracts each carry a different concentration of withanolides, the steroidal lactones researchers believe are behind most of ashwagandha's biological activity. The two branded extracts behind the strongest trial data, KSM-66 and Shoden (marketed as Sensoril in some regions), use different extraction processes and land on different withanolide percentages, which is part of why their studied doses differ (300 mg twice daily for one, 240 mg once daily for the other). A supplement that lists only "ashwagandha root extract 500 mg" without a standardization percentage or a named extract type is not necessarily lower quality, but it is harder to compare directly against a specific published trial, because the active compound concentration is unknown from the label alone.

This is also where "proprietary blend" language becomes a genuine obstacle to label literacy. When a product combines ashwagandha with several other adaptogens or botanicals under one combined milligram total, there is no way to know from the label how much ashwagandha the blend actually contains. NCCIH's underlying trial data, and the LiverTox case reports, are both about ashwagandha taken largely on its own at a known, disclosed dose. A blended product's real-world effect and risk profile cannot be assumed to match either body of evidence.

Shoppers often encounter ashwagandha alongside rhodiola, holy basil (tulsi), and Asian ginseng on the same "stress support" shelf, and it helps to know what separates them. Ashwagandha's clinical trial base, small as it is, specifically measured cortisol and validated anxiety and stress scales over 60-day periods, which is a more standardized outcome measure than most of what exists for rhodiola or holy basil. That does not make ashwagandha automatically superior; it means ashwagandha currently has more comparable, dose-specific randomized data attached to a smaller number of well-defined outcomes (sleep, self-reported stress, cortisol), while the evidence for the other botanicals tends to be thinner or measured differently between studies, making cross-comparison harder. None of these botanicals share ashwagandha's specific liver injury case report pattern in the same volume, which is worth knowing if liver health is a person's primary concern when choosing among adaptogens.

Dose, Duration, and Reading Past the Front Label

Both major stress and sleep trials measured outcomes at or after 60 days of continuous use, not after a few days or a single week. NCCIH's own sleep summary notes that ashwagandha "seemed most helpful when taken at doses of 600 mg per day or more for at least 8 weeks" in the pooled sleep data it reviewed. That timeline matters for expectation-setting: a shopper who takes ashwagandha for five days and stops because they feel nothing is not testing it the way the underlying trials tested it, and a shopper who keeps taking it for a year without checking in with a doctor is going well past the roughly three-month window NCCIH says has actual safety data behind it. The liver injury case reports in LiverTox span a wide range of exposure times, from four weeks to more than a year, which means duration alone does not predict risk, but it also means there is no dose or timeline that has been shown to be risk-free indefinitely.

A front label calling a product "clinically studied" or "clinically proven" does not tell you which clinical study, at what dose, using which extract. The two trials above used 240 mg and 300 mg twice daily of specific standardized extracts, taken for a full 60 days before the endpoints were measured. A product delivering a different total daily dose, a different extract standardization, or intended for shorter use is not automatically backed by those same numbers, even if its label cites "clinical research" in general terms. Checking the supplement facts panel for the actual milligram amount per serving, and comparing it to the dose used in a specific published trial, is the only way to know whether a product's claim and its label match.

Side Effects Most People Experience

NCCIH and LiverTox both describe the common, non-serious side effect profile the same way: drowsiness, stomach upset, diarrhea, and vomiting in some users, generally attributed to direct irritation of the intestinal lining at higher doses. These effects are described as uncommon rather than rare, and they typically resolve on their own or with a lower dose. They are the most likely reason someone stops taking ashwagandha, far more often than anything serious.

The Liver Injury Reports, In Detail

This is the part of the ashwagandha conversation that gets skipped most often in marketing copy. The NIH's LiverTox database, last updated December 3, 2024, states that ashwagandha "has not been implicated in causing serum enzyme elevations during therapy but recently has been implicated in cases of clinically apparent liver injury." LiverTox catalogs individual published case reports, and the pattern across them is fairly consistent:

• A 32-year-old woman developed thyrotoxicosis symptoms six weeks after starting ashwagandha and shortly after increasing her dose; her thyroid tests normalized after stopping (van der Hooft et al., 2005).
• A 20-year-old man developed jaundice with prolonged intrahepatic cholestasis a month after increasing his ashwagandha dose; jaundice persisted more than two months before resolving (Inagaki et al., 2017).
• Björnsson et al. (2020) reported five cases of liver injury from Iceland and the U.S., in adults aged 21 to 62, with symptoms and jaundice appearing two to twelve weeks after starting the supplement. All five resolved within one to five months, with no deaths and no chronic injury.
• A 39-year-old woman developed nausea and vomiting six weeks after starting a low dose (154 mg on alternate days) for anxiety, with bilirubin rising from 9.0 to 19.1 mg/dL before slowly recovering with treatment (Ireland et al., 2021).
• A case series from three hospitals in India (Philips et al., 2023) described eight patients with ashwagandha-related liver injury between 2019 and 2022. Five of the eight had pre-existing chronic liver disease, and three of those five developed acute-on-chronic liver failure and died. The remaining cases resolved, though often after prolonged cholestasis.
• A 2021 systematic review of herb-induced liver injury (Ballotin et al.) identified roughly 936 cases tied to 79 different herbal products across the literature; ashwagandha accounted for about 9 cases, or roughly 1% of the total, making it an uncommon but recorded cause relative to more frequently implicated herbs like green tea extract and kava.

Who Should Not Take It Without Medical Advice

NCCIH's health professional fact sheet lists specific groups who should avoid ashwagandha or use it only under medical supervision:

• People who are pregnant or breastfeeding, where the agency states ashwagandha should be avoided.
• People preparing for surgery, because of possible interaction with anesthesia and sedation.
• People with autoimmune or thyroid disorders, given documented cases of thyroid hormone disruption.
• People taking medications for diabetes or high blood pressure, immunosuppressants, sedatives, anticonvulsants, or thyroid hormone medication, due to documented interaction potential.
• People with hormone-sensitive prostate cancer, because ashwagandha may raise testosterone levels.
• Anyone with existing liver disease, given the concentration of severe outcomes in that specific patient group within the published case reports.

New Zealand's Regulatory Reality and What to Check Before Buying

Ashwagandha supplements sold in New Zealand fall under the Dietary Supplements Regulations 1985, administered by Medsafe under the Food Act 2014. As with other dietary supplements, there is no government pre-approval step before a product reaches a shelf. Responsibility for the accuracy of dosage, extract type, and any claims on the label sits with the company selling the product, not with a regulator checking it first. That is not unique to ashwagandha, but it means the burden of matching a label's dose to a specific clinical trial falls on the shopper, not on an approval stamp.

That makes a short buying checklist worth running through before adding ashwagandha to a routine:

• What is the total daily dose in milligrams, and does it match a dose actually used in a published trial (240 mg or 300 mg twice daily are the two studied here)?
• Is the extract standardized, and to what percentage of withanolides? Products that don't disclose standardization make it harder to compare against trial data.
• Do you have any liver condition, thyroid disorder, autoimmune condition, or upcoming surgery? Any of these is a reason to talk to a doctor first rather than start on your own.
• Are you pregnant, breastfeeding, or taking medication for diabetes, blood pressure, or thyroid function? Each of these interacts with the safety profile described by NCCIH.
• How long do you plan to take it? The safety data NCCIH cites covers roughly three months of use; long-term safety has not been established either way.

Limitations in the Current Evidence

Both trial types discussed here, stress-reduction and sleep, come from small sample sizes (60 to 64 participants) run at single centers, mostly in India, using specific standardized extracts that are not universal across brands. Liver injury case reports document what happened to individual patients but cannot establish how common liver injury actually is across the full population of ashwagandha users, since most people who take a supplement without incident never appear in a case report. NCCIH is explicit that long-term safety beyond about three months has not been established in either direction. None of this means the trial results or the case reports are wrong; it means both should be read as early, real, but incomplete evidence rather than a final verdict.

Publication bias is also worth naming directly. Liver injury case reports get published precisely because something unusual happened; a person who took the same dose for the same length of time with no problem has no reason to appear in a medical journal. That skew does not make the case reports meaningless (they still represent confirmed, documented harm), but it does mean the true rate of liver injury among all ashwagandha users cannot be calculated from case report counts alone. Similarly, the two headline stress and sleep trials were funded or supported by the companies that manufacture the extracts studied (KSM-66 and Shoden), a common and disclosed arrangement in supplement research that does not invalidate the results but is worth knowing when weighing how independent the evidence base is.

What This Means for Shoppers

Ashwagandha is not a fringe supplement with no data behind it, and it is not a risk-free wellness staple either. The sleep and stress trials are real, peer-reviewed, and specific about dose. The liver injury reports are also real, published in a government-maintained database, and specific about timing. A shopper who checks the milligram dose against a studied amount, avoids ashwagandha if they fall into one of NCCIH's caution groups, and tells their doctor before combining it with existing medication or an existing liver condition is applying the same label-literacy standard that should apply to any supplement, not treating ashwagandha as an exception in either direction.

Sources: U.S. National Center for Complementary and Integrative Health (NCCIH), "Ashwagandha: Usefulness and Safety," last updated March 2023; NIH LiverTox, "Ashwagandha," last updated December 3, 2024; Chandrasekhar, Kapoor, and Anishetty, Indian Journal of Psychological Medicine, 2012; Salve et al., Medicine, 2019; New Zealand Medsafe, "Regulation of Dietary Supplements," revised October 15, 2024; Philips et al., Hepatology Communications, 2023.

Sources

www.nccih.nih.gov

www.ncbi.nlm.nih.gov

pubmed.ncbi.nlm.nih.gov

pubmed.ncbi.nlm.nih.gov

www.medsafe.govt.nz

pmc.ncbi.nlm.nih.gov

This article is for general education and does not replace advice from a qualified healthcare professional.