Ashwagandha: Which Extract, What Dose, and the Safety Signal Most Guides Skip
Ashwagandha has one of the deepest research files of any adaptogen on the shelf, and one of the messiest label conventions. Two bottles can both say "ashwagandha root extract, 500 mg" and still deliver very different compounds. Here is which extract to look for, what dose the trials used, and the liver signal most guides skip.
Why Ashwagandha Is Hard to Buy Well
The word "ashwagandha" on a label tells you almost nothing. It does not say which part of the plant was used, how concentrated the active fraction is, or whether the dose matches anything that was ever studied.
A root-only extract standardized to 5% withanolides and a root-and-leaf extract standardized to 35% are different preparations, not two strengths of one product. The evidence base is real but uneven, and reviews report effects ranging from modest to implausibly large. The same is true across the wider stress-support shelf, which is why reading the label and matching the dose matters more than what the front of the bottle promises.
The safety picture is more specific than most guides admit. Ashwagandha has a documented liver signal, a thyroid signal and an adrenal signal, and each one has numbers attached.
The Three Extracts, and What Actually Separates Them
Most standardized ashwagandha on the shelf traces back to one of three branded extracts. They differ in two ways that matter: which part of the plant goes in, and how much of the withanolide fraction is retained.
| Extract | Plant part | Standardization | Commonly studied dose |
|---|---|---|---|
| KSM-66 | Root only | At least 5% withanolides, full spectrum | 300 to 600 mg per day |
| Sensoril | Root plus leaf | At least 10% withanolide glycosides | 125 to 500 mg per day |
| Shoden | Root plus leaf | At least 35% withanolide glycosides | 60 to 240 mg per day |
Root only versus root plus leaf
This is the first thing to check, because it decides which body of research applies. Root-only extracts carry most of the stress, sleep, strength and testosterone research. Root-and-leaf extracts are more concentrated and are usually positioned as the calming option.
The plant part is the most useful piece of information on an ashwagandha label. It matters for safety too. The leading hypothesis for the rare liver cases points at withanone, a compound more abundant in leaf material.
Why the percentages are not comparable
A 35% withanolide glycoside product is not a stronger version of a 5% product. One is a hard concentrate, the other a full-spectrum root extract. Higher percentage means a different preparation, not a proven advantage in outcomes. One practical label detail: KSM-66 is made using a milk-based extraction process, and the manufacturer documents the extract as containing milk allergen, so check the allergen statement if you have a milk allergy.
What the Trials Actually Used
Doses in the research vary by goal, and the extract used in a study is not always the extract on the shelf. These ranges summarize the doses reported across the trial literature and published clinical dose guidance; the meta-analyses in the References anchor the sleep and anxiety figures quoted below.
| Goal | Dose used in trials | Extract type studied |
|---|---|---|
| General wellbeing | 250 to 500 mg per day | Root extract |
| Stress support | 225 to 400 mg per day, later trials 300 to 600 mg | Root extract, 5% withanolides |
| Sleep support | 600 mg per day or more | Root extract |
| Memory and cognition | 300 mg per day | Root extract |
| Muscle strength and recovery | 500 mg per day | Root extract |
| Immune support | 60 mg per day | Standardized extract |
| Male fertility | 300 mg twice daily | Root extract |
| Calming extract | 125 to 500 mg per day | Root plus leaf, 10% withanolide glycosides |
| Concentrated extract | 60 to 240 mg per day | Root plus leaf, 35% withanolide glycosides |
Three things stand out:
- 600 mg per day is the dose where the sleep data is strongest, and the benefit showed up most clearly at eight weeks or longer.
- 60 mg per day is the lowest dose we found in the reviewed literature, from the immune research.
- The dose range moves with the extract. 500 mg of a 35% concentrate is not 500 mg of a 5% root extract.
The head-to-head comparison worth knowing about
One study compares forms directly. It was a randomized, double-blind, four-way crossover bioavailability study in 16 healthy adults, run under fasting conditions. Each participant received four commercially standardized extracts: 35% and 10% withanolide glycosides, and 5% and 2.5% withanolides. Every product delivered the same 185 mg total withanolides.
Because the withanolide total was held equal, any differences came from the form of the extract rather than the dose. And there was a clear winner. The 35% withanolide glycoside extract absorbed far better than the other three, and the authors attributed that to its glycoside content, concluding that it supports preferential use of high-glycoside extracts. That finding comes with an important limit. Better absorption is not the same as better results. The clinical trials behind the sleep, stress and testosterone findings were run mostly on root-only extracts standardized to about 5% withanolides, so the form that absorbs best has not been shown to produce the strongest outcomes. Read the percentage on the label, and read which preparation the research behind it actually used.
The largest safety dataset
The most reassuring number here comes from a prospective, randomized, multi-center, double-blind, placebo-controlled safety study of 1,002 adults at 600 mg per day of root extract, in two divided doses, for eight weeks. There were no serious adverse events. Of 74 adverse events (7.4%), 46 occurred in the placebo group (9.2%) against 28 in the active group (5.6%).
The group taking ashwagandha reported fewer side effects than the group taking placebo. A separate pilot randomized trial in healthy men also found a Withania somnifera root extract well tolerated.
How Long It Takes
There are two timelines, and confusing them is the most common reason people quit too early.
The acute window is hours rather than days, and the bioavailability study above tracked plasma levels over a 24-hour period. The clinical window is 4 to 12 weeks.
What the sleep data says about timing
The cleanest quotable number here comes from a sleep meta-analysis of five randomized trials and 400 participants. Overall sleep improved with an effect size of -0.59 (95% CI -0.75 to -0.42, I2 = 62%). That is small but statistically significant, and it was more prominent in adults with insomnia.
Two conditions mattered. Benefit tracked with doses of 600 mg per day or more and with durations of eight weeks or more. For where ashwagandha sits against the alternatives, see which sleep supplement matches your sleep problem. If you take 300 mg for three weeks and feel nothing, you have not tested ashwagandha. You have tested a dose and a duration the sleep data does not support.
What the anxiety data says
A meta-analysis of five randomized trials with 254 participants found the Hamilton Anxiety Rating Scale reduced by a mean difference of -5.96 (95% CI -10.34 to -1.59). Modest, and consistent in size with the sleep findings.
The biomarker trap
One review found that supplementation significantly reduces cortisol with no effect on perceived stress. A lab marker can move while you feel exactly the same, so judge your response by how you sleep and how you handle your day. If cortisol is the marker you are tracking, which nutrients lower cortisol and at what dose covers the rest of that picture.
About those very large effect sizes
One 2026 dose-response meta-analysis of 22 randomized trials reports improvements in stress, depression and anxiety scores. The direction matches everything else here, but the pooled effect sizes are implausibly large for this literature, far bigger than the individual trials inside them show.
Reviews disagree sharply with each other, and the better-controlled sleep and anxiety reviews land on modest effects. When a review reports an effect much larger than the trials it pooled, trust the better-controlled numbers.
The Liver Signal, With the Actual Numbers
Most guides soften this into "rare liver issues, talk to your doctor." Here is what the case literature contains.
A 2026 scoping review identified 13 publications covering 25 patients with liver injury associated with ashwagandha. Injury typically developed after weeks of use, not days. Cholestatic or mixed biochemical patterns predominated, and jaundice and pruritus were commonly reported. Most patients recovered after stopping. Severe outcomes were rare but real: one case required a liver transplant, and three deaths occurred in patients who already had cirrhosis and decompensated. That is why the review advises anyone with advanced chronic liver disease to avoid ashwagandha.
How causality gets judged
Modern case reports use a structured tool to score whether a supplement likely caused an injury. One 2026 case report assessed ashwagandha liver injury with the updated 2016 RUCAM, and the scoping review above reports that causality assessments, when applied, generally classified ashwagandha as a probable or likely cause.
The mechanism is a hypothesis, not a fact
Laboratory work has shown that withanone, a withanolide compound, can damage DNA, and this has been proposed as a potential cause of liver toxicity in some people taking Withania somnifera products.
Treat that as a leading hypothesis, not settled science. Withanone is more abundant in leaf material, and analyses differ on whether leaf-derived material carries the risk. Nobody has proven the mechanism in people.
The counterweight matters too
Against 25 case patients, the 1,002-person controlled study above found no serious adverse events, and broader safety reviews agree that serious harm is uncommon at studied doses. One 2026 review of ashwagandha as a food additive does flag significant risk in pregnancy, in people with thyroid disorders, and in people with liver or kidney dysfunction.
The practical read
Serious liver injury appears to be rare, and the largest controlled dataset did not detect it. But the reported cases are real, and they cluster after weeks of use. If you notice yellowing of the eyes or skin, dark urine, itching, or discomfort in the upper right of your abdomen, stop and talk to a clinician.
The Thyroid Signal
A placebo-controlled eight-week study in people with bipolar disorder measured thyroid indices in a small subgroup and found that all three ashwagandha-treated participants had T4 increases from baseline, of 7%, 12% and 24%. Six of the seven placebo participants had T4 decreases over the same period.
That is three treated people, so treat it as a signal and not a proven effect. It lines up, though, with what larger reviews report.
A 2026 safety review of ashwagandha as a food additive found significant risk specifically in people with thyroid disorders. Clinical reviews also advise people with overactive, underactive or borderline thyroid function to avoid it, because the herb can raise thyroid activity.
If your thyroid is being monitored, or you take thyroid medication, talk to your prescriber before starting.
The Adrenal Signal Nobody Measures
The adaptogen story is that ashwagandha helps the body handle stress, and what adaptogens actually do in the body covers that broader category. The adrenal question is what that looks like over a long period.
There is a documented case. A 55-year-old postmenopausal woman took roughly 950 mg per day for over a year and presented with Cushingoid features and weight gain. Testing showed low basal cortisol and a suboptimal ACTH stimulation response, consistent with adrenal insufficiency. That is a single case, and it does not establish a general risk.
The more useful problem is what the cortisol reviews did not measure. A systematic review of nine studies, with durations from 30 to 112 days, found that supplementation decreases cortisol secretion with no significant adverse effects. That review states plainly that none of the studies evaluated the effect of cortisol reduction on adrenal function.
So the research can tell you cortisol went down. It cannot tell you the gland itself is fine. A separate 2025 review adds that cortisol dropped significantly while perceived stress did not change.
If you take a high dose long term, that is the pattern the case report describes. Keep the dose inside the ranges that were studied, watch for unexplained weight change or the features the case describes, and ask your clinician about check-ins.
Who Should Not Take Ashwagandha
Some of these are absolute and some are a conversation with your prescriber. The liver, thyroid, kidney and pregnancy cautions come from published safety reviews. The remaining items reflect long-standing clinical guidance for this herb rather than controlled trial data.
- Pregnant or breastfeeding. Reviews flag significant risk in pregnancy, and traditional safety guidance cites uterine contraction.
- Thyroid disorders, including overactive, underactive or borderline thyroid function.
- Liver disease, or any history of liver problems.
- Kidney dysfunction.
- Autoimmune conditions, including lupus, rheumatoid arthritis, type 1 diabetes and Hashimoto's.
- Prostate cancer.
- Upcoming surgery. Stop well before the date and tell your surgical team.
- Milk allergy, if you use an extract made with a milk-based process such as KSM-66.
- Anyone taking medication for diabetes, blood pressure, seizures, thyroid function or anxiety, or taking sedatives. Talk to your prescriber first.
- Anyone who feels more agitated rather than calmer. A paradoxical reaction is described in the literature.
How to Choose One
Working through these in order beats reading reviews.
- Check the plant part. Root only, or root plus leaf. That tells you which body of research applies.
- Check the standardization, not just the milligrams. A 5% root extract and a 35% root-and-leaf extract are different preparations, not different strengths of one.
- Match the dose to the extract. 300 to 600 mg for a 5% root extract. 125 to 500 mg for a 10% root-and-leaf extract. 60 to 240 mg for a 35% root-and-leaf extract.
- Match the dose to your goal. 600 mg per day if sleep is the goal. 250 to 500 mg per day for general wellbeing. 300 mg per day if memory is the goal, and the brain and memory support pillar covers the rest of that stack.
- Compare the study dose to the label dose. If the trial used 600 mg and the capsule gives 300 mg, the studied amount is two capsules, not one.
- Buy from a seller who can answer a formulation question. Practitioner-grade catalogs, including the lines the Agape team carries, print the extract name, the plant part and the standardization. Agape's professional-grade quality standards page explains how those lines get vetted.
- Start at the lower end, give it 4 to 12 weeks, and stop if something feels off.
The honest summary: ashwagandha has a modest evidence base, a well-documented safety profile at studied doses, and a small but genuine liver signal in a minority of users. Choosing well is mostly about matching the extract to the dose to the goal.
What We Recommend
Three options that print the extract, the plant part and the standardization on the label, which is the first thing this guide asks you to check.
XYMOGEN, Ashwagandha 35 - 60 Capsules
Root and leaf extract standardized to 35% withanolide glycosides, the concentrated form in the table above, for a healthy stress response and restful sleep.
$38.99
Pure Encapsulations, Ashwagandha 60 and 120 Capsules
A 500 mg root extract standardized to 2.5% withanolides, one capsule daily, for occasional physical and mental stress.
$29.00
Quicksilver Scientific, Adaptogenic Black Label 1.7 fl oz
A liquid blend of nineteen adaptogenic herbs for occasional stress, resilient energy and hormone balance.
$77.00
References
- Alsanie SA, Alhodieb FS, Askarpour M. Effects of ashwagandha (Withania somnifera) on mental health in adults: A systematic review and dose-response meta-analysis of randomized controlled trials. Complement Ther Med. 2026. PMID 41644067. https://pubmed.ncbi.nlm.nih.gov/41644067/
- Cheah KL, Norhayati MN, Husniati Yaacob L, Abdul Rahman R. Effect of Ashwagandha (Withania somnifera) extract on sleep: A systematic review and meta-analysis. PLoS One. 2021. PMID 34559859. https://pubmed.ncbi.nlm.nih.gov/34559859/
- Fatima K, Malik J, Muskan F, Raza G, Waseem A, Shahid H, Jaffery SF, Khan U, Zaheer MK, Shaikh Y, Rashid AM. Safety and efficacy of Withania somnifera for anxiety and insomnia: Systematic review and meta-analysis. Hum Psychopharmacol. 2024. PMID 39083548. https://pubmed.ncbi.nlm.nih.gov/39083548/
- Arumugam V, Vijayakumar V, Balakrishnan A, B Bhandari R, Boopalan D, Ponnurangam R, Sankaralingam Thirupathy V, Kuppusamy M. Effects of Ashwagandha (Withania Somnifera) on stress and anxiety: A systematic review and meta-analysis. Explore (NY). 2024. PMID 39348746. https://pubmed.ncbi.nlm.nih.gov/39348746/
- Marchi M, Grenzi P, Travascio A, Uberti D, De Micheli E, Quartaroli F, Laquatra G, Pingani L, Ferrari S, Galeazzi GM. The effect of Withania somnifera (Ashwagandha) on mental health symptoms in individuals with mental disorders: systematic review and meta-analysis. BJPsych Open. 2025. PMID 41140145. https://pubmed.ncbi.nlm.nih.gov/41140145/
- Rathi P, Kim SK. Randomized, Double-Blind, Crossover Study Comparing the Bioavailability of 4 Ashwagandha (Withania somnifera (L.) Dunal) Extracts in Healthy Adults Under Fasting Condition. Curr Ther Res Clin Exp. 2025. PMID 40792075. https://pubmed.ncbi.nlm.nih.gov/40792075/
- Pakhale K, Salve J, Ademola J, Parraca J, Langade D. Safety of 8-Week Administration With Ashwagandha (Withania somnifera) Root Extract in Adults With Stress and Anxiety: Findings From a Prospective, Randomized, Multi-Center, Double-Blinded, Placebo-Controlled Study. Phytother Res. 2026. PMID 41943502. https://pubmed.ncbi.nlm.nih.gov/41943502/
- Vaidya N, Agarwal R, Kshirsagar P, Ganu G, Nagore D, Mehta A, Vora A, Nair S. Safety and Tolerability of Withania somnifera Root Extract in Healthy Male Participants: A Pilot Randomized, Double-Blind, Placebo-Controlled Clinical Trial. Food Sci Nutr. 2026. PMID 41458530. https://pubmed.ncbi.nlm.nih.gov/41458530/
- Albalawi AA. Dual impact of Ashwagandha: Significant cortisol reduction but no effects on perceived stress - A systematic review and meta-analysis. Nutr Health. 2025. PMID 40746175. https://pubmed.ncbi.nlm.nih.gov/40746175/
- Della Porta M, Maier JA, Cazzola R. Effects of Withania somnifera on Cortisol Levels in Stressed Human Subjects: A Systematic Review. Nutrients. 2023. PMID 38140274. https://pubmed.ncbi.nlm.nih.gov/38140274/
- McIntyre D, Nguyen P, Kim Y, Meyer B, Salloum M. Ashwagandha (Withania somnifera)-Associated Liver Injury: A Scoping Review of Clinical Characteristics and Safety Considerations. Cureus. 2026. PMID 42367407. https://pubmed.ncbi.nlm.nih.gov/42367407/
- Perez Perez S, Zumaque-Valverde E, Saavedra Chacon MF, Arteta Cueto AA. Ashwagandha-Induced Herb-Induced Liver Injury Assessed With the Updated 2016 RUCAM. Case Reports Hepatol. 2026. PMID 42529606. https://pubmed.ncbi.nlm.nih.gov/42529606/
- Siddiqui S, Ahmed N, Goswami M, Chakrabarty A, Chowdhury G. DNA damage by Withanone as a potential cause of liver toxicity observed for herbal products of Withania somnifera (Ashwagandha). Curr Res Toxicol. 2021. PMID 34345852. https://pubmed.ncbi.nlm.nih.gov/34345852/
- Ronen Y, Ebert C, Tamim-Yecheskel BC, Zev S, Kantor O, Arbel HB. Comprehensive safety evaluation of Withania somnifera (Ashwagandha): an AI-driven meta-analysis and quantitative structure-activity relationship based toxicity assessment. Front Nutr. 2025. PMID 41368179. https://pubmed.ncbi.nlm.nih.gov/41368179/
- Kumar A, Sah BK, Ahmad F, Jamwal V, Sahu SK. Potential Adverse Effects of Ashwagandha: A Critical Review of Preclinical and Clinical Evidence. Phytother Res. 2026. PMID 41502355. https://pubmed.ncbi.nlm.nih.gov/41502355/
- Li L, Zhang X, Hao J, Tian X, Zhu Y, Yang Y, Qiu B, Wen J. Evaluating the Safety of Ashwagandha as a Food Additive: A Focus on Reproductive, Thyroid, and Immune System Effects. Phytother Res. 2026. PMID 41454545. https://pubmed.ncbi.nlm.nih.gov/41454545/
- Gannon JM, Forrest PE, Roy Chengappa KN. Subtle changes in thyroid indices during a placebo-controlled study of an extract of Withania somnifera in persons with bipolar disorder. J Ayurveda Integr Med. 2014. PMID 25624699. https://pubmed.ncbi.nlm.nih.gov/25624699/
- Javid J, Kunnuru SR, Anne B, Muskan S. From herbal hope to hormonal havoc: chronic ashwagandha use and HPA axis suppression. Endocrinol Diabetes Metab Case Rep. 2026. PMID 41553875. https://pubmed.ncbi.nlm.nih.gov/41553875/
Frequently Asked Questions
How much ashwagandha should I take per day?
It depends on the goal and the extract. General wellbeing research uses 250 to 500 mg per day. Sleep research shows the strongest effect at 600 mg per day or more, for eight weeks or longer. For stress, trials used 225 to 400 mg per day, with later studies at 300 to 600 mg. A root-and-leaf extract standardized to 35% withanolide glycosides was studied at 60 to 240 mg per day, and one at 10% withanolide glycosides at 125 to 500 mg per day.
What is the difference between KSM-66 and Sensoril?
The plant part and the standardization. KSM-66 is root only, standardized to at least 5% withanolides, studied at 300 to 600 mg per day. Sensoril is root plus leaf, standardized to at least 10% withanolide glycosides, studied at 125 to 500 mg per day, and usually positioned as the calming option. They are different preparations, not two strengths of the same one.
Is ashwagandha hard on your liver?
For most people, the controlled data says no. The largest safety study, in 1,002 adults at 600 mg per day for eight weeks, found no serious adverse events and fewer side effects than placebo. That said, a 2026 scoping review identified 13 publications covering 25 patients with liver injury linked to ashwagandha, mostly cholestatic or mixed patterns with jaundice and itching, typically after weeks of use. Most recovered after stopping. Stop and see a clinician if you notice yellowing of the eyes or skin, dark urine, itching, or upper right abdominal discomfort.
Why is ashwagandha not FDA-approved?
Because it is sold as a dietary supplement, not a drug. The FDA regulates supplements as a category of food, so they do not go through the pre-market approval process that medicines do. A supplement also cannot legally claim to diagnose, treat, cure or prevent a disease. That is a regulatory classification, not a verdict on whether ashwagandha works. It is the same reason you see structure and function language like "supports a healthy stress response" on the label instead of disease claims.
Does ashwagandha affect thyroid function?
There is a small signal. In one placebo-controlled eight-week study in people with bipolar disorder, all three ashwagandha-treated participants in a small thyroid subgroup had T4 increases from baseline, while most placebo participants had T4 decreases. That is a very small treated group, so treat it as a signal rather than a proven effect. It lines up, though, with safety reviews that flag significant risk in people with thyroid disorders.
Can I take ashwagandha every day?
The trials behind the stress, sleep and strength findings ran continuously for roughly 8 to 12 weeks. Cycling is a popular practice, but the clinical literature has not tested it, and there is no established long-term safety data beyond the durations that were studied. A documented adrenal case involved about 950 mg per day for over a year.
How long does ashwagandha take to work?
Expect two timelines. Any acute effect appears within hours of a dose, and the bioavailability study above tracked plasma levels over 24 hours. Clinical effects on sleep and stress measures took 4 to 12 weeks in the trials, and the sleep benefit was strongest at 600 mg per day for eight weeks or more.
Who should not take ashwagandha?
Pregnant or breastfeeding women, people with thyroid disorders, anyone with liver disease or kidney dysfunction, people with autoimmune conditions such as lupus, rheumatoid arthritis, type 1 diabetes or Hashimoto's, people with prostate cancer, anyone facing upcoming surgery, anyone with a milk allergy using a milk-processed extract, and anyone taking medication for diabetes, blood pressure, seizures, thyroid function or anxiety. Talk to your prescriber first in that last group.