Ashwagandha

Sold for stress, anxiety, sleep and testosterone. The trials report effects too large to believe, and it has caused liver injury.

At a glance

Our verdict	Implausibly large pooled effects, low certainty, and reports of liver injury.
Sold for	Stress, anxiety and sleep (claimed)
How sure the evidence is	Low, and a known harm
Cost a month	About $15 to $25

Written for adults in Canada. It explains the evidence; it is not advice for you personally. Last reviewed 20 October 2026.

Supplement Check

Is it right for you?

Already taking ashwagandha? It can injure the liver, change thyroid hormone and add to sedatives. Your check tells you whether to stop it, and what would help your stress or anxiety more. Our Supplement Check looks at:

Pregnancy: traditionally used to cause abortion
Breastfeeding
Liver disease
Overactive thyroid
Autoimmune disease
Every medicine you take

Coming soon. The Supplement Check and the Online Health Assessment open once our safety testing is complete.

$59 when it opens, paid once. Adults 18 and over, anywhere in Canada. While the service is new, Dr. French reads every report himself, and your report arrives within 3 working days. More about the Supplement Check

Tell me when it opens

What it is

Ashwagandha (Withania somnifera) is a herb used in Ayurvedic medicine. Products are extracts of the root, or of root and leaf, standardised to compounds called withanolides. Leaf-containing extracts are common and the withanolide content varies between products. It is one of the supplements that has newly become common in the United States national survey. [1]

Structure of Ashwagandha — Structure: PubChem (public domain).

How it acts

Proposed, not proven: an effect on the brain's calming messenger system (GABA) and on the stress-hormone system (the hypothalamic–pituitary–adrenal axis). This is mechanism, not evidence that it works.

Outline of a person showing where Ashwagandha acts in the body

What it is used for, and what the evidence says

It is sold for stress, anxiety, sleep, testosterone and fertility. Our review looked at the trials for anxiety and stress.

Use	What the trials found	How sure we are
Anxiety	Pooled trials report a fall in anxiety scores of 1.55 standard units. [2] That would be nearly three times the effect of talking therapy against placebo (0.56). The results varied far more between trials than chance allows (heterogeneity 83 to 94%), and most trials were small and sponsored by companies. An effect this large, from trials like these, is more likely to be bias than biology.	Low
Stress	Pooled stress scores fell by 1.75 standard units, with the same problems: small, mostly company-sponsored trials that disagree with each other. [2]	Low
Sleep, testosterone, fertility	It is sold for these. Our review did not grade trials for them.	Not reviewed

How sure we are: High means further research is very unlikely to change the answer; Moderate, it might; Low, it may well change; Very low, we are very unsure. These follow the GRADE method that guideline panels use.

Benefits

Uncertain. The reported falls in anxiety and stress scores look too large to be true. [2]
No benefit has been measured that lifestyle change or proven treatment does not provide.

Risks and side effects

Liver injury. Drug-induced liver injury of the type that blocks bile flow (cholestatic), starting 2 to 12 weeks after the first dose, in a case series from Iceland and the US Drug-Induced Liver Injury Network. [3]
Overactive thyroid. Case reports of thyrotoxicosis: it can raise thyroid hormone.
Gut upset.
Unreliable products. Leaf-containing extracts and variable withanolide content; Danish authorities have moved to restrict it in supplements.

Interactions with medicines

Our review found no measured effect on the liver enzymes that clear medicines. The concerns are added effects: on drowsiness, on thyroid hormone, on the immune system, and on blood sugar and blood pressure.

Medicine	What can happen	What to do
Sleeping pills, sedatives, alcohol	Added drowsiness	Do not take
Levothyroxine and other thyroid medicines	May raise thyroid hormone; the dose may need changing	Do not take; discuss with your doctor
Immunosuppressants (after a transplant, or for autoimmune disease)	Possible stimulation of the immune system (theoretical)	Do not take
Diabetes medicines, blood-pressure medicines	Added lowering of blood sugar or blood pressure has been reported	Do not take; discuss with your doctor

Who should not take it

Pregnancy: traditionally used to cause abortion. Do not take it.
Breastfeeding.
Liver disease.
Overactive thyroid.
Autoimmune disease.
Before surgery.
Anyone taking a medicine in the table above.
Under 18: outside the Supplement Check.

Dose and timing

We do not recommend a dose. For reference, trials used 300 to 600 mg of extract a day.

If you take it now: liver injury appeared 2 to 12 weeks after starting. [3] Yellowing of the skin or eyes, dark urine, pale stools or itching mean stop it and see a doctor promptly.

What we know about Canadians

No Canadian data on ashwagandha use in our review. What we know about the problems it is sold for: in 2022, 5.2% of Canadians aged 15 and older had generalised anxiety disorder, up from 2.6% in 2012, and 7.6% had a major depressive episode. [4] Counting any mood, anxiety or substance use disorder, it was 16 to 18%. [5, 4] The rise was steepest in young women aged 15 to 24, in whom generalised anxiety tripled. Fewer than half (48.8%) of people with a disorder spoke to a health professional about it. [4] Nearly half of Canadians (45.6%) took a supplement in 2015. [6]

Most at risk of anxiety: women 15 to 24, unemployed people, and people with substance use (half also have a mood or anxiety disorder). [5]

How sure are we: high. For the anxiety figures: national survey (Mental Health and Access to Care Survey 2022) with a structured diagnostic interview. No Canadian data on ashwagandha itself.

An honest comparison: anxiety

Effects are in standard units where the trials report them that way (a common scale that lets different questionnaires be compared). Exercise trials cannot be blinded, which inflates their effects. Cost a month in Canadian dollars.

Option	Effect	How sure we are	Cost a month (CAD)
Lavender oil (Silexan) 80 mg	34% more people had their anxiety score halved than on placebo (rate ratio 1.34); nearly all trials funded by the maker [10]	Moderate to Low	$30 to $40
Ashwagandha	Pooled effect −1.55 standard units: far larger than talking therapy, and too large to be believable; liver injury reported [2]	Low (low, and harm)	$15 to $25
Cognitive behavioural therapy (talking therapy)	About half a standard unit better than placebo on anxiety symptoms (g 0.56), across 41 placebo-controlled trials; large for generalised anxiety [7]	Moderate	Free provincial or online programmes where available; private psychologist about $200 a session
Exercise	A moderate fall in anxiety (median effect −0.42); trials could not be blinded [8]	Low	$0 to $21
Mindfulness programme (8-week mindfulness-based stress reduction)	About half a standard unit on stress (g about 0.5); moderate on anxiety [9]	Low	$0 (apps, free courses) to $40–60
SSRI or SNRI (escitalopram, sertraline, duloxetine)	About 2 to 3 points more than placebo on the doctor-rated anxiety scale (HAM-A) [11]	Moderate	$8 to $15

Read honestly: ashwagandha's pooled effect is the largest number in this table, and the least believable. Talking therapy has the strongest evidence, exercise and mindfulness cost little or nothing, and the medicines have known, modest effects. Ashwagandha adds a real risk of liver injury.

Why we don't recommend it

We do not sell ashwagandha and our check never recommends it. The trials report effects too large to be true, the certainty is low, and it has caused liver injury.

Already taking it? Discuss it with your doctor. Stop it now if you are pregnant, planning a pregnancy or breastfeeding, if you have liver or thyroid disease, or if you take a medicine in the interaction table. See a doctor promptly if your skin or eyes turn yellow. If anxiety is the reason you take it, talking therapy and regular exercise have better evidence; if your anxiety score is high, a doctor's assessment comes first (a screen is not a diagnosis).

The review ranked ashwagandha against talking therapy, exercise, mindfulness, lavender oil and the anxiety medicines, and Dr. French accepted its conclusion: not offered. An effect nearly three times that of talking therapy, from small company-sponsored trials that disagree with each other, is not believable; the true effect is unknown. Against that uncertain benefit sits a real harm: liver injury, starting weeks after the first dose, and case reports of an overactive thyroid. When the benefit is unknown and the risk is real, we do not sell it.