Ashwagandha: Benefits, Evidence, Dosage & Safety

Ashwagandha may modestly reduce stress and improve sleep over 8–10 weeks, but the evidence is low-certainty: 12 small randomized trials with substantial heterogeneity (Akhgarjand et al., 2022). Doses of 300–600 mg/day of root extract were studied. Short-term use up to three months appears reasonably safe; long-term data do not exist.

TL;DR

I have read the ashwagandha trials end to end, and here is where I land. In one 64-person trial, stressed adults taking two capsules of ashwagandha root extract daily for sixty days came out with lower stress scores and lower blood cortisol than the placebo group. A separate 10-week trial found people fell asleep faster and slept better on it. Those are real results. The rest is caveat: the 2022 meta-analysis that pooled the stress trials rated the certainty of evidence low for stress and anxiety (Akhgarjand et al.), short-term use up to about three months looks reasonably safe, and anything longer is genuinely unstudied. Nobody knows what this herb does over years. Anyone who says otherwise is guessing.

What it is

The name tells you most of what you need to know, which I love. Ashwagandha is roughly "smell of the horse" in Sanskrit, a nod to the root's pungent odor and to the old claim that it lends something horse-like: strength, stamina, staying power. Botanically it is Withania somnifera (L.) Dunal, a small woody shrub in the nightshade family, growing wild across the Indian subcontinent, the Middle East, and parts of Africa (PMC review).

Here is the part that matters practically, and where I see people go wrong: both the tradition and the clinical trials use the dried root. Leaves show up in some preparations too, but the root is the main event (PMC review). When a trial says "ashwagandha worked," it means the root. Keep that in mind every time you read a label.

The compounds researchers keep circling back to are the withanolides, steroidal lactones by chemistry. NCCIH notes they have been associated with anti-inflammatory and antioxidant effects. Associated with, not proven to cause. I keep that distinction in my head whenever I read a marketing page (NCCIH).

Traditional use

In Ayurveda, ashwagandha is a rasayana, usually translated "rejuvenative." Accurate, but bloodless next to the traditional pitch, which was bigger: longevity, resilience under pressure, stamina, recovery from exhaustion (PMC review).

The official definition is precise, and I like precise. The Ayurvedic Pharmacopoeia of India (Part I, Volume I) says the drug "asvagandha" is the dried mature roots of Withania somnifera Dunal, roots specifically, not leaves, not a whole-plant powder (Tier 1: API Pt. I Vol. I, Government of India; institutional holding: Wellcome Collection).

The Ministry of Ayush calls it "a cornerstone of Ayurveda for centuries," and they have published expert-reviewed assessments of where the safety and efficacy evidence actually stands (Ministry of Ayush press statement, CCRAS; PIB release).

Go back far enough, to the Samhita-period texts, and you find it prescribed for joint inflammation, nerve disorders, epilepsy, tumors, heart disease, ulcers, usually as root powder or paste (PMC review of traditional uses). It turns up in Unani and Siddha practice too (PMC review).

All of that is traditional claim. I want to be explicit about what that means: centuries of use counts for something, but it is not a clinical trial. The trials are next, and they are a separate question entirely.

What modern research says

Stress

This is where the evidence is strongest. Which, I should say immediately, is not the same as saying it is strong. Let me show you what is actually there, and then tell you what I think of it.

The trial everyone cites is Chandrasekhar and colleagues, published in 2012: sixty-four chronically stressed adults, randomized, double-blind, placebo-controlled, taking two capsules daily of a high-concentration full-spectrum root extract at 300 mg each, for sixty days. Stress-scale scores fell significantly against placebo (P<0.0001). Serum cortisol fell too. Side effects were no worse than what the placebo group reported (Chandrasekhar et al., Indian Journal of Psychological Medicine).

A good result. Now the caveat, because there is always one and it usually matters more than the result: the extract was KSM-66. That is a branded product from Ixoreal Biomed, standardized to at least 5% withanolides, tested at a single center (systematic-review tabulation; paper). One company's extract. One lab. One group of researchers. What that tells you about the bottle on a store shelf is, frankly, an open question, and I would not let anyone hand-wave past it.

What does the wider literature say? A 2022 systematic review with dose-response meta-analysis pooled 12 randomized trials, 1,002 participants total, in Phytotherapy Research. The headline finding: ashwagandha reduced stress and anxiety across the pooled data. But the authors rated the certainty of evidence low for both outcomes. Substantial heterogeneity between studies, different extracts, different doses, different populations, and they called for better trials (Akhgarjand et al., 2022). Twelve trials and the verdict is still "low certainty." That tells you something about the state of this field.

On dose, the meta-analysis favored 300-600 mg per day for stress. The anxiety data were messier, with a wider effective range.

NCCIH lands in the same neighborhood: some ashwagandha preparations may help with stress; for anxiety, the evidence is unclear (NCCIH). My read, after going through all of it: a real signal, wrapped in real uncertainty. I hold both at once.

Sleep

Smaller evidence base here, and I will not dress it up. The trial worth knowing about is Langade et al., 2019: sixty patients, double-blind, placebo-controlled, ten weeks on the same KSM-66 root extract at 300 mg twice daily. That is the branded KSM-66 extract from Ixoreal Biomed, the same one as the stress trial. The asterisk stays.

What changed: sleep onset latency shortened (p=0.019), sleep efficiency improved, and Pittsburgh Sleep Quality Index scores improved (p=0.002) (Langade et al., Cureus 2019). NCCIH agrees that some preparations may help with insomnia (NCCIH).

Sixty people, ten weeks, a genuine signal, and still a long way from conclusive. I would not build a sleep protocol on one trial.

Testosterone and male fertility

NCCIH describes "some limited evidence" that two to four months of ashwagandha might raise testosterone and improve sperm quality (NCCIH).

Count the qualifiers in that sentence with me: some. Limited. Might. The trials are small, the findings are preliminary, and this is not a basis for treatment claims. If someone sells you ashwagandha as a testosterone therapy, they are out ahead of the data, and I would keep my wallet closed.

What the evidence does not support (yet)

NCCIH is blunt about the rest of the list, so I will be too: the evidence is not there for asthma, athletic performance, cognitive function, diabetes, menopause, or female infertility, and the COVID-19 evidence is not high-quality (NCCIH).

Stress, sleep, and the preliminary hormone findings: that is where the data lives. Everything beyond that is running ahead of what has actually been shown.

Dosage

There is no established dose. Anyone who tells you otherwise is simplifying, and I would ask what they are selling.

Preparations vary enormously, root powder, standardized extracts, different withanolide percentages, so a single number would mislead anyway. What I can tell you, because it is the only dosing with evidence behind it, is what the trials actually used.

Both the stress and sleep trials used 300 mg of KSM-66, the branded high-concentration full-spectrum root extract from Ixoreal Biomed standardized to at least 5% withanolides, twice daily. That is 600 mg per day total, for 60 days to 10 weeks (Chandrasekhar et al.; Langade et al.).

Read that carefully, because it names a specific branded extract at a specific standardization. A retail product with "ashwagandha" on the label is not the trial extract. Do not assume they match. I certainly don't.

NCCIH notes the herb has been studied for short-term use up to about 3 months. Beyond that, the data thin out fast (NCCIH).

Safety & contraindications

NCCIH's bottom line is measured, and mine matches it: possibly safe for short-term use (up to 3 months); not enough data to say about long-term use.

What has been reported: drowsiness, stomach upset, diarrhea, vomiting. And rarer but worth knowing, there are case reports of liver injury linked to ashwagandha supplements (NCCIH). Rare does not mean never.

Who should steer clear: anyone pregnant or breastfeeding, who should avoid it entirely. It is not recommended before surgery, nor for people with autoimmune conditions or thyroid disorders. I would not negotiate with any of those.

Interactions are the part people underestimate, in my experience. Possible interactions with diabetes medications, blood pressure medications, immunosuppressants, sedatives, anti-seizure drugs, and thyroid hormone medications (NCCIH). And because of the possible testosterone effect, anyone with hormone-sensitive prostate cancer should avoid it (NCCIH).

My rule of thumb: if you take prescription medicines, treat ashwagandha as a check-with-your-clinician-first herb rather than a try-it-and-see one.

FAQs

Does ashwagandha help with stress?

The best-supported claim on this page, and I still want you to hear the qualifications. In a 2012 trial, 64 stressed adults took 300 mg of root extract twice daily for 60 days, and both stress-scale scores and serum cortisol dropped versus placebo (Chandrasekhar et al.). A 2022 meta-analysis of randomized trials found stress and anxiety reductions across the pooled data (Akhgarjand et al.). NCCIH's read: some preparations may work for stress, while the anxiety evidence stays unclear (NCCIH). That is about as good as it gets for this herb, and it is still "may."

Does ashwagandha help with sleep?

Maybe. That is the whole honest answer, and I would rather give you three words than a paragraph of fog. A 10-week randomized trial in 60 patients found 300 mg twice daily improved how fast people fell asleep, how efficiently they slept, and their overall sleep quality scores versus placebo (Langade et al.). NCCIH notes some preparations may help insomnia (NCCIH). Small, short trials are all that exist so far. I would not promise anyone better sleep on this evidence.

Does ashwagandha lower cortisol?

More than one trial says yes, though both used branded extracts. In the 2012 Chandrasekhar trial, serum cortisol fell significantly more in the KSM-66 group than in the placebo group over 60 days (Chandrasekhar et al.). In a separate 2024 randomized trial, the Shoden extract at 60 mg and 120 mg daily drove morning serum cortisol down 66% and 67% respectively, against a 2% change on placebo (Mishra et al., Heliyon 2024). That one used a root-and-leaf extract, not the root-only KSM-66. Two small trials in stressed adults is a finding, not a general principle, so resist the blanket version of this claim. I do.

Is ashwagandha safe?

Up to about three months, trials suggest it is reasonably safe for most healthy adults, with mostly mild side effects (NCCIH). The caveats are real, and I take them seriously: rare cases of liver injury exist, avoid it in pregnancy and breastfeeding, and take the interaction list seriously if you use thyroid, diabetes, blood pressure, sedative, or immunosuppressant medications (NCCIH). "Reasonably safe" is not "safe for everyone." It never is.

Can I take ashwagandha with thyroid medication?

This one deserves real caution rather than self-experimentation. NCCIH specifically flags interactions with thyroid hormone medications and advises against use in thyroid disorders (NCCIH). Take it to your clinician. I am not going to soften that.

How long does ashwagandha take to work?

The trials only tell us about 8 to 10 weeks of daily use, which is when the stress and sleep benefits showed up (Chandrasekhar et al.; Langade et al.). Whether a single dose does anything, or how quickly effects begin, was never tested. The trials were not designed to answer that, so I will not invent an answer.

Voices

The classical record. The Charak Samhita's own reference table places ashwagandha twice in its opening pharmacological catalog: in the Brumhaniya mahakashaya, the group of bulk- and strength-promoting herbs (Cha.Sa. Sutra Sthana 4/9(2)), and in the Balya mahakashaya, the strength-tonic group (Cha.Sa. Sutra Sthana 4/9(7)). It also appears as an ingredient in the Vajikarana Ghrita of the vajikarana chapter (Cha.Sa. Chikitsa Sthana 2/1/34). The Bhavaprakasha Nighantu lists it as Balya, "promotes strength." These are classical attributions, not trials. They tell you what the tradition claimed, not what has been proven (Charak Samhita Online, Ashwagandha reference table).

A clinician's endorsement, with the asterisk. "Ashwagandha has long been used in Ayurvedic medicine to increase energy, improve overall health and reduce inflammation, pain and anxiety," said Dr. Yufang Lin, an integrative medicine specialist at the Cleveland Clinic. Her framing of the stress story is mechanistic and unromantic: chronic threat keeps the stress response chronically engaged, and that is precisely the pattern the trials were measuring (PhillyVoice).

A vaidya's dissent from the hype. Vaidya Mishra of Shaka Vansiya Ayurveda, writing in his published newsletter, is blunt about limits: the herb should not be used by high-pitta body types or "in the presence of amavisha, reactive hot toxins in the body," because its heating potency adds to existing heat. His piece is literally titled "How NOT to Take It," a reminder that traditional use comes with traditional cautions (SV Ayurveda newsletter).

What "adaptogen" even means. Barbara Sorkin, Ph.D., director of Botanical Research Centers at the NIH Office of Dietary Supplements, defined the term plainly in an interview: an adaptogen is "loosely defined as a compound or product that increases the ability of a person to resist, adapt or become resilient in nonspecific ways to biological, physical or chemical stressors." Note the word loosely. Even the NIH's own botanical research chief treats the label as approximate (Record Gazette).

Sources

The stress trial is Chandrasekhar and colleagues' 2012 randomized, double-blind, placebo-controlled study of a high-concentration full-spectrum root extract, published in the Indian Journal of Psychological Medicine (full text at PMC3573577, DOI: 10.4103/0253-7176.106022). I cite it first because it is the trial everything else leans on, and the one most people overstate. The sleep trial is Langade et al.'s 2019 double-blind RCT of ashwagandha root extract in insomnia and anxiety, published in Cureus (doi:10.7759/cureus.5797). The pooled picture comes from Akhgarjand et al.'s 2022 systematic review and dose-response meta-analysis of 12 randomized trials in Phytotherapy Research (doi:10.1002/ptr.7598), the paper I point skeptics to, because its "low certainty" verdict keeps everyone honest. The second cortisol trial is Mishra et al.'s 2024 randomized, double-blind, placebo-controlled study of the Shoden extract (60 mg and 120 mg daily, standardized to 35% withanolide glycosides) in 60 adults with elevated stress and anxiety, published in Heliyon (PMC11402924).

NCCIH's "Ashwagandha: Usefulness and Safety" page, updated March 2023, is the backbone for the safety statements and the evidence summaries throughout (nccih.nih.gov). When I write "the evidence is unclear," that page is usually why. The traditional-use background draws on two PMC reviews: a broad review of ashwagandha and well-being (PMC12252077) and one on traditional uses (PMC10633886). The pharmacopoeial definition of the drug is via the Wellcome Collection's institutional holding of the Ayurvedic Pharmacopoeia of India (wellcomecollection.org); the Ministry of Ayush position is documented in its CCRAS press statement (PDF) and the accompanying PIB release (pib.gov.in).

Disclaimer

This article is for educational purposes only and is not medical advice. Herbal products can interact with medications and are not suitable for everyone. Talk to a qualified healthcare practitioner before use, especially if you are pregnant, breastfeeding, managing a medical condition, or taking prescription medicines.