Ashwagandha
Withania somnifera

Ashwagandha (Withania somnifera) is a plant from Ayurvedic medicine that has been used for more than two thousand years to manage stress and improve rest. In the shop you will find it as a supplement, usually in capsules (the powder tastes very bad).

It is worth knowing this from the outset: ashwagandha is neither a sedative nor an antidepressant, it does not act within hours, and it has no effect from a single one-off dose. Clinically measurable changes appear between the fourth and eighth week of continuous use, and quality studies cover treatments of up to three months.

The name has history. Ashwa means “horse” in Sanskrit and gandha is “smell”: the fresh root smells intensely, and the Ayurvedic tradition adds that the plant confers the strength and vitality of a horse. The Latin epithet somnifera means “sleep-inducing”. In common usage it is also known as “Indian ginseng” or “winter cherry” — popular nicknames that imply no botanical kinship with those plants. It already appears in the Charaka Samhita, the founding treatise of Ayurveda written more than two thousand years ago, classified as Rasayana — the category of tonic plants associated with longevity and immunity.

The agents responsible for the clinical effects are the withanolides. Not all withanolides are alike, nor are they distributed equally between root and leaves, and that explains why two supplements with the same total percentage can have different effect profiles. The section on extracts and commercial brands develops this question in detail.

Chemically, withanolides are lipophilic C28 steroidal lactones structurally related to ergosterols. The reference phytochemical characterization is that of Mirjalili et al., 2009 (Molecules), which identifies more than forty variants with distinct pharmacological profiles — notably withaferin A (more bioactive and more cytotoxic, abundant in leaves), withanolide A (more abundant in the root, neuroprotective profile), and the glycowithanolides (variants with added sugars, which appear to be absorbed better orally).

The raw berry is toxic, and commercial extracts are made exclusively from the root or from processed leaves — never from the whole fruit. Botanically the plant belongs to the family Solanaceae — the same as tomato, potato, eggplant, pepper, tobacco and belladonna; if you have a known sensitivity to nightshades, it is worth keeping in mind, although the clinical evidence on this is anecdotal.

The solid evidence for ashwagandha is concentrated in three areas: it serves to reduce perceived stress and cortisol levels, it improves sleep quality, and it provides modest support for recovery after exercise. For cognition, longevity and athletic performance there are promising but still preliminary data.

Stress and subclinical anxiety. This is the indication with the most solid scientific basis. In people with chronic stress but no psychiatric diagnosis, ashwagandha reduces morning cortisol by between 23% and 33% and improves perceived-stress scores by between 30% and 44% versus placebo.

Sleep. In people with non-restorative sleep or mild insomnia, ashwagandha reduces the time it takes to fall asleep, increases total sleep time and improves the subjective sense of rest. It is not a sedative: it does not make you fall asleep quickly after a one-off dose. What it does is improve sleep architecture over the weeks.

Memory and attention. There is a single rigorous trial in adults with mild cognitive impairment showing improvements in immediate memory, attention and processing speed. Extrapolation to people without declared cognitive impairment is not validated, and there is no trial in Alzheimer’s or in frank dementia.

Added to this is an effect on the GABAergic system, the brain’s main inhibitory system. Some minor withanolides act on the same receptor that benzodiazepines act on, although with much lower intensity. This explains the calming effect and the improvement in sleep. Unlike benzodiazepines, no tolerance or dependence has been documented to date, although studies lasting longer than three months are scarce.

There are also other secondary mechanisms — effects on serotonin and dopamine in preclinical models, antioxidant action, bidirectional immunomodulation and an effect on the thyroid that raises T3 and T4 while lowering TSH. The mechanistic detail of these effects is developed in the absorption and synergies section; the clinical consequence of the thyroid effect is developed in the contraindications section, because it determines who should not take it.

The clinical evidence in stress and anxiety rests on three converging pivotal trials: Chandrasekhar 2012, Lopresti 2019 and Salve 2019. Together they establish a consistent pattern — reductions in morning cortisol between 23% and 33% and improvements on the Perceived Stress Scale between 30% and 44% — and they establish the market’s reference dose (600 mg/day of KSM-66, superior to 250 mg/day in the only available dose-response RCT).

In physical performance the evidence diverges by domain. For cardiorespiratory endurance, Choudhary 2015 documented a VO₂max improvement of 5.67 versus 1.86 mL·kg⁻¹·min⁻¹ for placebo at 12 weeks with 600 mg/day of KSM-66 in healthy athletic adults. For strength, the reference trial is Wankhede 2015 — exclusively in untrained young men. The review by Namysł and colleagues (2026) concludes that ashwagandha does not significantly raise testosterone in women.

Ashwagandha is intended for adults living with chronic stress, sleep problems or mild anxiety, who are looking for natural support backed by reasonable evidence.

It also has two secondary profiles with more preliminary but significant evidence: physically active people seeking support for recovery after exercise, and adults who notice mild cognitive complaints — occasional forgetfulness, difficulty concentrating — without any medical diagnosis involved.

The main profile is well defined by the evidence: an adult with self-reported chronic stress, sleep problems or subclinical anxiety, with no psychiatric diagnosis involved.

The athletic profile deserves nuance. The evidence covers two distinct profiles. The first is the untrained young man — sedentary, starting a strength program for the first time. There the gains in strength, muscle mass and testosterone are clear, but not extrapolable to trained men, to women or to elite athletes. The second profile is the physically active adult of either sex seeking support for recovery. If you already train regularly, its likely usefulness is in recovery, not in pushing performance.

The cognitive profile is the narrowest of the three. The only rigorous trial available (Choudhary 2017) was done in adults with clinically diagnosed mild cognitive impairment, mean age 50. If you are 40 and notice occasional forgetfulness but have no clinical diagnosis, ashwagandha may help through its effect on stress and sleep — two factors that affect cognition indirectly — rather than through a direct nootropic action.

The vast majority of pivotal clinical trials have been done in adult Indian populations, with small sample sizes (n=40 to 80), durations of 8 to 12 weeks and a heterogeneous mix of sex ratios. Extrapolation to European or American populations is reasonable but not formally validated. Three profiles are under-studied: adolescence (there is no RCT in those under 18), very advanced age (trials do not usually go beyond 65) and women of childbearing age with near-term pregnancy plans. The interaction with hormonal contraceptives is an explicit evidence gap: no RCT has evaluated it directly.

Unlike other supplements where the dose is universal, with ashwagandha the dose depends on the commercial extract you are taking. The three brands with scientific backing (KSM-66, Sensoril and Shoden) use different concentrations of the active compounds, so the effective doses are also different. The practical rule is to follow the recommended dose of the brand on your bottle; if your product does not name any identifiable brand, it is wise to be wary of its quality.

Clinically measurable effects appear from the fourth week of continuous use and their peak usually occurs in the eighth week. Quality studies cover treatments of six to twelve weeks. Beyond three months, the available evidence becomes scarce.

The equivalence between doses of different brands is not mathematical. That Shoden works at 60 mg and KSM-66 at 600 mg does not mean Shoden is “ten times more potent” — it means the composition is different. Shoden concentrates up to 35% glycowithanolides (the withanolides with added sugars, which appear to be absorbed better orally), whereas KSM-66 contains at least 5% total withanolides. What makes no sense is to take the milligrams from a KSM-66 trial and apply them to a bottle of Shoden, or vice versa.

For use focused on stress and sleep, a common practice among integrative clinicians — without an RCT formally validating it — is to start with two weeks at a low dose to assess tolerance (300 mg of KSM-66 or 60 mg of Shoden), move up to a standard dose (600 mg of KSM-66 or 120–240 mg of Shoden) for six to eight weeks, assess the response, and decide whether to continue another four weeks or stop. This stepwise approach mirrors the typical pattern of rigorous clinical trials.

Ashwagandha is absorbed better when you take it with food containing fat. The active compounds — the withanolides — are fatty molecules, they do not dissolve in water. That is why the Ayurvedic tradition combined it with warm milk and clarified butter, a preparation called golden milk. In modern practice no such ceremony is needed: it is enough to take it with the main meal of the day, or with a whole-milk yogurt, a handful of nuts or a glass of milk.

There are two supplements that combine well with ashwagandha in the context of sleep and anxiety: magnesium glycinate (which also acts on the GABAergic system) and L-theanine (an amino acid from green tea with a calming daytime effect), although the combination is not validated by formal clinical trials.

Withanolides are lipophilic steroidal lactones, structurally related to the sterols your body routinely processes with dietary fats. That lipophilicity explains two things: why absorption improves with fat in the meal, and why several withanolides manage to cross the blood-brain barrier. This is the clinical reason the acute effect is slight and the benefits accumulate over weeks: it is not a drug with a one-off response curve, it is a sustained-action biological modulator.

The synergies with a mechanistic basis are three. Magnesium glycinate for sleep: both ashwagandha and glycinate modulate the GABAergic system, by slightly different but compatible routes. L-theanine for daytime anxiety: it acts on GABA and glutamate and has a fast, gentle effect profile that complements ashwagandha’s more sustained action. Rhodiola rosea as a complementary adaptogen for cases with predominant fatigue — although there is a published case of hepatotoxicity with the two in prolonged use. More combined adaptogens is not necessarily better: each one adds hepatic load and complicates the attribution of effects or adverse effects.

For use focused on cognition and focus, some people combine ashwagandha with bacopa monnieri or with moderate caffeine. There are no rigorous comparative trials on these combinations. Bacopa has its own body of evidence for memory in older people and shares an antioxidant mechanism with ashwagandha. Caffeine acts on a completely different axis (adenosine), with no known antagonism. They are reasonable but exploratory combinations, not validated protocols.

The pharmacokinetic study by Kim, Venkatesan and Rathi (2023) introduces a critical observation that has not yet been independently replicated: in a double-blind crossover with 16 healthy adults, the AUC₀-inf of total withanolides per mg of extract was about 281 times higher with Shoden (35% glycowithanolides) than with a standard 2.5% extract. The implication is that two products labeled “ashwagandha 500 mg” can have radically different plasma profiles depending on the type of extract. A 2025 follow-up study (Rathi and Kim, Current Therapeutic Research) extends the comparison to four extracts and confirms that the glycowithanolides govern, not the total content. Limitations: small sample, authors tied to Arjuna Natural (manufacturer of Shoden), pending independent replication.

On secondary pharmacodynamic mechanisms: the thyroid registers one of the best-documented and clinically relevant effects. Sharma 2018 showed T3 +41.5%, T4 +19.6% and TSH −17.5% with 600 mg/day for 8 weeks in subclinical hypothyroidism. This same effect, useful in that context under medical supervision, is clearly harmful in hyperthyroidism — which explains the firm contraindication in the final section.

Unlike other supplements where a gram is a gram, with ashwagandha the extract matters more than the amount. Two bottles that say “ashwagandha 500 mg” can contain radically different concentrations of active compounds depending on how the plant has been processed. That is why the industry developed standardized extracts — preparations with a guaranteed percentage of withanolides, validated in clinical trials. There are three with real scientific backing.

On the label of a good product you should find five things: the name of the brand, the percentage of withanolides and the measurement method (usually HPLC), the part of the plant used, the batch and expiry date, and the exact dose per capsule.

There is also a group of third-party quality seals that certify the purity of the final product: USP Verified, NSF Certified for Sport, Informed-Sport and organic certifications. With ashwagandha these seals matter more than with other supplements because imported Ayurvedic roots have a documented history of contamination with lead and other heavy metals. An analysis in JAMA (2008) found that roughly 20% of imported Ayurvedic herbal products contained heavy metals above safety limits. In Spain these seals are uncommon, so their absence should not be a reason for distrust.

A cross-cutting warning about the body of evidence: practically all the pivotal clinical trials are funded by, or receive the extract from, the manufacturer. Chandrasekhar 2012 declared “no conflict of interest” but the extract was supplied by Ixoreal Biomed. Lopresti 2019 is funded directly by Arjuna Natural. Deshpande 2020 has authorship partly internal to Arjuna. Salve 2019 with KSM-66 received the extract donated by Ixoreal.

This does not invalidate the results — the methodological design is sound, the data are published and the findings are consistent with the Akhgarjand 2022 meta-analysis. But it does demand editorial transparency and caution in the reading. Independent replication is scarce: academically independent studies such as Coope 2025 in female soccer players do exist, but they are the exception, not the norm. This explains why the foundational effect sizes tend to moderate in meta-analyses and partial replications.

Ashwagandha has a notably solid base of clinical evidence for a medicinal herb. More than thirty clinical trials in humans have evaluated its effects, many of them randomized, double-blind and placebo-controlled. The consensus concentrates on three conclusions: the effect on perceived stress and cortisol is consistent and reproducible, the effect on sleep quality is clear but more modest, and the effects on physical performance and cognition are promising but limited by the composition of the samples.

A warning worth having from the start: the vast majority of pivotal trials are funded by the manufacturers of the commercial extracts. This does not invalidate the results, but it conditions how the body of evidence should be read. The development of this cross-cutting warning is in the extracts and brands section.

Stress and cortisol

Three pivotal trials converge on a consistent pattern: Chandrasekhar 2012, Lopresti 2019 and Salve 2019. The meta-analysis by Akhgarjand and colleagues (2022) covering 12 trials and 1,002 participants confirms the effect directions with an SMD of −1.55 for anxiety and −1.75 for stress, but warns of high heterogeneity (I² of 83 to 94%) and low GRADE certainty: the direction of the effect is supported, but the exact magnitude is hard to pin down.

Sleep

Two main trials: Langade 2019 (KSM-66, 10 weeks, objective actigraphy) and Deshpande 2020 (Shoden, 6 weeks, n=150). The meta-analysis by Cheah and colleagues (2021) covering 5 trials and 400 participants reports an SMD of −0.59 on sleep quality with larger effects in diagnosed insomnia, doses at or above 600 mg/day and durations longer than 8 weeks.

Physical performance

For cardiorespiratory endurance, Choudhary and colleagues (2015) documented, with KSM-66 600 mg/day for 12 weeks, a VO₂max improvement of 5.67 versus 1.86 mL·kg⁻¹·min⁻¹ for placebo. For strength, the reference trial is Wankhede 2015 — done exclusively in untrained young men, with the full study card at the advanced level. The review by Namysł and colleagues (2026) explicitly concludes that ashwagandha does not significantly raise testosterone in women.

Cognition

The evidence is clearly weaker than in the previous domains. There are no trials in Alzheimer’s or in frank dementia. The popular claim that “ashwagandha lengthens telomeres” is not supported by clinical evidence in humans: it comes from an in vitro study in cancerous HeLa cells funded by Ixoreal, not extrapolable.

“It raises testosterone in women.” This is a popular myth in marketing contexts, especially around libido and energy. The review by Namysł and colleagues (2026) concludes that ashwagandha does not raise testosterone in women to a clinically significant degree. What has been documented in perimenopausal women (Gopukumar 2021) is an increase in estradiol, not testosterone — a completely different hormonal effect aimed at other symptoms.

“It cures depression.” It neither cures nor treats it. What there is, is preliminary evidence of mood improvement in people with stress or subclinical anxiety. Ashwagandha is not an antidepressant, it does not act on the same systems as SSRIs and SNRIs, and it should not replace the medical treatment of a diagnosed depressive disorder.

A final reflection on the concept of adaptogen. The category has no regulatory recognition: neither the US FDA nor the European EFSA use it as a defined category. The three original criteria of Brekhman 1969 — increasing nonspecific resistance to stress, bidirectional physiological normalization, absence of significant toxicity — are useful as functional descriptors, but they do not constitute a formal pharmacological category. When it is claimed that “ashwagandha is an adaptogen”, what is being used is a descriptor with an ethnobotanical basis and a research history, not a regulatory classification with objective, verifiable criteria.

The optimal duration. Almost all rigorous clinical trials cover treatments of 6 to 12 weeks. There are no double-blind controlled trials lasting longer than 6 months. Studies show that the biomarkers tend to return partially to baseline after stopping, without rebound. The usual practice in integrative care is 8 to 12 weeks with assessment and monitoring if it is prolonged (liver panel and TSH).

The need to cycle is not demonstrated. The traditional Ayurvedic recommendation of “eight weeks on, two weeks off” is repeated in much popular literature, but it does not come from any clinical trial that has validated it. We do not know whether the body develops tolerance with continuous use.

The sex-differential dose is not characterized. Almost all pivotal trials have used the same dose range without prior stratification by sex. We know that men and women respond differently in some domains (testosterone rises in untrained young men; in women it does not), but we do not know whether the optimal dose for reducing cortisol or improving sleep should be different between the sexes.

The interaction with hormonal contraceptives is an explicit gap. No trial has evaluated this interaction directly. Since ashwagandha modulates FSH and LH in perimenopausal women, there is a theoretical possibility that it interferes with contraceptives that act on the same axis. In integrative clinical practice the usual recommendation is to discuss it with your gynecologist, especially if there are changes in the menstrual pattern during use.

Clinical thyroid. Sharma 2018 showed clinically useful effects in subclinical hypothyroidism (T3 +41.5%, T4 +19.6%, TSH −17.5%). But cases of iatrogenic thyrotoxicosis with ashwagandha are documented (Burrage 2022; Kang 2019). It is the same property — it raises thyroid hormone — but with opposite consequences depending on the context. As a treatment for subclinical hypothyroidism it is an interesting possibility, but it requires medical supervision with TSH monitoring, and it is a firm contraindication in hyperthyroidism.

Alzheimer’s and frank dementia. There are no rigorous clinical trials in patients with Alzheimer’s disease or with other established dementias. There is abundant preclinical evidence in animal models, but the translation to humans is not demonstrated. What we have is Choudhary 2017 in mild cognitive impairment. From there to frank dementia is a leap that current evidence does not allow.

The methodological heterogeneity of the body of evidence is another important gap. The Akhgarjand 2022 meta-analysis reported I² of 83 to 94% — very high levels that indicate the studies are not measuring exactly the same thing in the same populations. This does not invalidate the general direction of the effect, but it does limit precise quantitative claims. Until a new generation of RCTs with harmonized protocols exists (same extract, same doses, same instruments, same durations), the aggregate figures will carry a margin of uncertainty greater than desirable.

Ashwagandha is generally well tolerated in healthy adults, but it is not for everyone. There is a group of clinical situations in which it should not be taken, and another group in which it requires prior medical consultation. If you fit any of these profiles, this section is the most important in the whole page.

Consult your doctor before starting if you take any chronic medication (antihypertensives, antidiabetics, anxiolytics, antidepressants, thyroid hormone, anticoagulants), you have an autoimmune disease, you are being treated for a hormone-sensitive cancer, you are going to undergo surgery in the next two weeks, or you are seeking pregnancy in the short term.

Pregnancy. The contraindication is absolute and the current reason is the absence of safety data, rather than demonstrated toxicity. A recent review (Tallon, Koturbash and Blum, 2025) has questioned the soundness of the classic datum on abortifacient action, noting that it was based on the aerial parts of the plant, not the root. It is worth reading in context: the lead author signs from a food-law consultancy and the review is presented openly as a response to the regulatory threat of EU Article 8, so its skeptical reading of the classic datum should be weighed with that orientation in mind. Even so, in the absence of trials in pregnant humans, the prudent stance remains an absolute contraindication. If you are planning pregnancy, it is wise to stop at least a month before trying to conceive.

Liver disease and the hepatic question. The published cases are rare against millions of consumers, but the pattern is documented: latency of 2 to 12 weeks, usually cholestatic presentation, complete resolution in 1 to 5 months after stopping. For prolonged use beyond three months, integrative clinical practice recommends a baseline transaminase check (AST, ALT, GGT) before starting and a repeat at 8–12 weeks.

Hyperthyroidism. Ashwagandha raises T3 and T4 and lowers TSH (Sharma 2018: T3 +41.5%, T4 +19.6%, TSH −17.5%). This can be clinically useful in subclinical hypothyroidism under medical supervision, but it is clearly harmful in any form of hyperthyroidism. If you take levothyroxine for established hypothyroidism, it is not flatly contraindicated, but it requires supervision: the levothyroxine dose might need adjustment and there must be TSH monitoring when starting treatment.

Hepatotoxicity — clinical detail. The best-documented series are Björnsson and colleagues (2020, Liver International) with 5 cases in Iceland and the DILIN network. The presentation patterns are three: cholestatic (the most frequent, with resolution in 1–5 months), hepatocellular (more variable latency) and, in rare cases, fulminant liver failure in patients with prior liver disease. Complete resolution after stopping is the rule. The exact incidence is unknown because it depends on population denominators that are not well established.

Thyrotoxicosis — clinical detail. The published cases (Burrage 2022, Cureus; Kang 2019, Annals of Internal Medicine) show a consistent pattern: a patient with previously normal thyroid function or well-controlled hypothyroidism, start of ashwagandha, development of hyperthyroidism symptoms in the following 4–12 weeks (palpitations, heat intolerance, weight loss, tremor). Suppressed TSH, elevated free T4. Complete resolution after stopping the supplement within weeks to months.