Supplement
Ashwagandha
Adaptogen · Withania somnifera / KSM-66
Ashwagandha is the best-known adaptogen; meta-analyses show lower stress and cortisol scores over eight weeks. Quality of life did not improve in the same data, and the German Federal Institute for Risk Assessment warns of liver damage.
What ashwagandha is
Ashwagandha (Withania somnifera) comes from Ayurveda and is marketed as an adaptogen, that is, as an agent meant to improve adaptation to stress. Studies use standardized extracts such as KSM-66 or Sensoril.
The extracts are heterogeneous: different manufacturing processes yield different mixtures of compounds, which is why results cannot simply be transferred to another product. Only a few weeks have been studied; there are no data on long-term use.
What was measured – and what was not
The studies record scores on anxiety and stress scales as well as cortisol levels. Cortisol is a surrogate marker, that is, a lab value that is only meant to reflect a process. Quality of life, by contrast, is the outcome that directly matters to the people affected.
This very separation is revealing: scale scores and the lab value move, quality of life does not – both in the same analysis.
What is well supported
- The largest study reached its primary endpoint. Pakhale 2026 randomized 1002 people (498 versus 504), multicenter and double-blind, to 600 mg per day or placebo for eight weeks. The primary endpoint was safety, and it was reached: 5.6 percent adverse events under the active product versus 9.2 percent under placebo, 74 events in total, none serious. So fewer events occurred under ashwagandha than under placebo.
- Anxiety and stress: significant effects in the meta-analysis. The systematic review by Bachour 2025 analyzed 15 randomized studies with 873 adults. After eight weeks, the mean difference on the Hamilton Anxiety Rating Scale was −3.52 (95% CI −6.00 to −1.04; p=0.0053) and on the Perceived Stress Scale −4.88 (−7.84 to −1.91; p=0.0013). Both confidence intervals lie entirely below zero.
- The cortisol finding is the clearest single value. In the same analysis, cortisol fell by −2.36 (95% CI −3.26 to −1.46) at p less than 0.0001. The narrow interval points to a result that is consistent across studies and provides the biochemical anchor for the scale scores.
- Common side effects have been recorded in a large cohort. Liver, kidney and blood values remained unchanged in the 1002 participants over eight weeks. For common events this is a solid basis; rare individual reactions fundamentally cannot be captured by a cohort of this size.
What the studies show
Meta-analysis: stress and cortisol better, quality of life not
A systematic review with meta-analysis of 15 randomized studies in 873 adults with stress or anxiety (Bachour 2025) found after eight weeks a mean difference of −3.52 on the Hamilton Anxiety Rating Scale (95 percent confidence interval −6.00 to −1.04; p=0.0053), −4.88 on the Perceived Stress Scale (−7.84 to −1.91; p=0.0013) and −2.36 for cortisol (−3.26 to −1.46; p less than 0.0001). For quality of life, the same analysis found no effect: −1.76 (−5.61 to 2.09; p=0.3698). The paper explicitly does not show that a benefit exists beyond eight weeks or that it occurs in a clinically diagnosed anxiety disorder.
Largest safety study: unremarkable over eight weeks
In a multicenter, double-blind study (Pakhale 2026), 1002 people (498 versus 504) received 600 mg per day or placebo for eight weeks. The primary endpoint was safety, and it was reached: 5.6 percent adverse events under the active product versus 9.2 percent under placebo, 74 events in total, none serious. Liver, kidney and blood values remained unchanged. The study's funding could not be verified.
Case report: jaundice after three days
A case report (Perez Perez 2026) describes a 27-year-old woman who developed jaundice after three days of intake. The liver tissue sample showed a cholestatic pattern, that is, a disturbance of bile flow. The RUCAM score, a standardized assessment scheme for the link between a drug and liver injury, was 7, which corresponds to the category probable.
Where the data stop
- No improvement in quality of life. In the same meta-analysis, quality of life remained unchanged; the confidence interval clearly includes zero. If you are aiming for a better everyday experience, you will find no evidence for it.
- No evidence beyond eight weeks. The data end where the studies end. They say nothing about the common long-term use.
- No evidence in clinically diagnosed anxiety disorder. The studies looked at adults with stress or anxiety symptoms, not at conditions requiring treatment.
- Rare liver damage has not been ruled out. A study with 1002 participants over eight weeks cannot, mathematically, rule out events that occur less often than roughly once per 500 people. The reported cases of liver damage fall exactly within this range.
Status, approval and legal
In communication 039/2024 of 2024-09-10, the German Federal Institute for Risk Assessment warns of liver damage up to acute liver failure in connection with ashwagandha products. It advises children, pregnant and breastfeeding women, and people with liver disease against taking them.
The authority's addendum is essential: no safe reference value can be derived from the data – there is no amount for which safety has been established. This text therefore contains no dosage recommendation.
Safety
The two findings only appear to contradict each other. The large study shows that among 1002 people over eight weeks, no common and no serious side effects occurred. The case report and the authority's warning, by contrast, concern rare but serious liver reactions. Such events systematically remain invisible in studies of this size – not because they are ruled out, but because the number of participants is not sufficient.
The case report also shows that a liver reaction can occur after only three days and is therefore not tied to long-term intake. Jaundice, dark urine, persistent nausea or upper abdominal pain call for a medical evaluation. This text does not replace medical advice.
BK-Score Thin human evidence
| Human evidence | 5 | |
|---|---|---|
| Mechanism | 4 | |
| Safety data | 5 | |
| Hype gap | 4 | |
| Track record of use | 8 |
There are several RCTs on stress and cortisol, mostly small, short and close to manufacturers. The mechanism remains vague. Important point for the safety data: reports of liver damage have occupied several authorities.
The score rates the state of knowledge, not the substance. “Safety data 9” means well studied – not harmless. “Track record of use 9” means used long and widely – that is not proof of efficacy.
Subjective assessment by Biohacking Kompakt based on published scoring rules – not a scientific rating and not a medical recommendation. Rules and all ratings (German)
Frequently asked questions about ashwagandha
Does ashwagandha work against stress?
On rating scales, yes; in everyday experience, not demonstrably. A meta-analysis of 15 studies with 873 adults found lower scores for anxiety, perceived stress and cortisol after eight weeks. Quality of life did not improve in the same analysis. The difference between a scale score and a noticeable benefit is the decisive point here.
Why does quality of life not improve?
That is an open question, but it is the most striking finding in the data. Anxiety and stress scales and cortisol levels capture narrowly defined measures; quality of life captures overall experience. If only the former moves, it remains unclear whether the change matters to the people affected at all. So far, only the change in measured values is established.
Is ashwagandha harmful to the liver?
The German Federal Institute for Risk Assessment warns of liver damage up to acute liver failure and does not derive a safe reference value. A case report describes a 27-year-old woman with jaundice after only three days of intake, with a cholestatic pattern in the tissue sample. Such reactions are rare but serious and not tied to long-term intake.
But wasn't the large safety study unremarkable?
It was, and that still does not rule out rare liver damage. In the study, 1002 people received the active product or placebo for eight weeks; adverse events occurred in 5.6 versus 9.2 percent, none of them serious. Events that occur less often than roughly once per 500 people cannot, mathematically, be captured by a study of this size.
Who should avoid ashwagandha?
The German Federal Institute for Risk Assessment advises children, pregnant and breastfeeding women, and people with liver disease against ashwagandha products. No safe reference value could be derived from the data. This page gives no dosage or usage recommendation and does not replace medical advice; if you have existing conditions, the decision belongs with a physician.
How long has ashwagandha been studied?
The data extend to eight weeks. All studies analyzed fall within this range, and the meta-analysis explicitly states that a benefit beyond eight weeks has not been shown. For the common long-term use over months or years, there are neither efficacy nor safety data from controlled studies.
The podcast episode (in German)
Episode 41
Ashwagandha: the stress herb fact-checked
The podcast by Paul Höser (Episode 41) · with Paul & Paula. Three thousand years of Ayurveda meet more than sixty human studies: almost 28 % less cortisol (Chandrasekhar, Indian J Psychol Med 2012), better sleep (Langade, Cureus 2019), double the strength gain in the bench press (Wankhede, JISSN 2015), testosterone and libido data, less stress eating. Plus the honest side: the BfR liver warning, thyroid caution, sensible cycles and extract quality (KSM-66, Sensoril). Information only – not medical advice, no dosage or usage recommendation.
Related
- Works together withMagnesium
- Works together withRhodiola rosea
- Same categoryBacopa monnieri
- Same categoryPanax ginseng
- Also for hormones and stressTongkat ali
- Also for cortisol and sleepPhosphatidylserine (PS)
Sources
- Bachour 2025, BJPsych Open – meta-analysis on stress and anxiety, no effect on quality of life (PMC12242034)
- Pakhale 2026, Phytotherapy Research – largest safety study, n=1002 (PMID 41943502)
- Perez Perez 2026, Case Reports in Hepatology – case report of liver injury (PMID 42529606)
- BfR communication 039/2024 of 2024-09-10 – warning about liver damage
Open in the database – with search, filters and comparison (German app)
Information only, not medical advice and not a usage or dosage recommendation. Prescription-only and unapproved substances belong in the hands of a physician. Last updated: 2026-09-13.