Evidence File 029 · Ashwagandha · Method v0.14 of 4 claims researched

AshwagandhaWithania somnifera root extract · 4 claims

The cortisol reduction is real. Whether people actually feel less stressed is where the meta-analyses stop agreeing. And one country has banned it.

4Claims scored

8Sources

~600 mgStudied dose

Jul 26Reviewed

Claim 01

Lowers cortisol

5.6/10

Emerging·Replicated · but cortisol is a marker, not a feeling

Multiple meta-analyses report significant cortisol reduction. A 2025 synthesis titled around exactly this point found significant cortisol reduction. A 2024 analysis of 9 RCTs and 558 participants found lowered cortisol alongside perceived stress and anxiety scores.

The consistency across independent syntheses is real. The question the next claim raises is whether it matters.

The working

6Quantity

5Quality

7Consistency

4Directness

(6 × 0.25) + (5 × 0.35) + (7 × 0.25) + (4 × 0.15) = 5.6
Directness scores 4. Serum cortisol is a biomarker several steps removed from what a person buying this actually wants.

MetaAlbalawi AA. Dual impact of Ashwagandha: significant cortisol reduction but no effects on perceived stress Nutrition and Health 2025;31(4):1395-1408 · doi 10.1177/02601060251363647
MetaArumugam V, et al. Effects of Ashwagandha on stress and anxiety: systematic review and meta-analysis Explore 2024;20(6):103062

Claim 02

Reduces perceived stress

5.2/10

Contested·Meta-analyses disagree on the outcome that matters

The 2025 meta-analysis is explicit in its own title: significant cortisol reduction, no effect on perceived stress. That is a biomarker-outcome dissociation of exactly the kind this format exists to surface.

Other syntheses disagree. A 2024 analysis of 9 RCTs found significant reduction on the Perceived Stress Scale and the Hamilton Anxiety Scale as well as cortisol. A 2025 analysis in BJPsych Open covering 15 studies and 873 patients examined the same question.

A recurring caveat across all of them is extract variability — different trials use materially different preparations, which limits what pooling them can tell you.

The working

6Quantity

5Quality

3Consistency

8Directness

(6 × 0.25) + (5 × 0.35) + (3 × 0.25) + (8 × 0.15) = 5.2
Consistency scores 3. The 2025 synthesis found the biomarker moved and the experience did not; other syntheses found both moved.

MetaAlbalawi AA. Cortisol reduction without perceived stress effect Nutrition and Health 2025;31(4):1395-1408
MetaBachour G, Samir A, Haddad S, et al. Effects of ashwagandha supplements on cortisol, stress and anxiety levels in adults BJPsych Open 2025 · doi 10.1192/bjo.2025.10136 · 15 studies, 873 patients

Claim 03

Improves sleep quality

6.0/10

Moderate·12 RCTs · 1,002 adults · small effect

A meta-analysis of randomised trials comparing ashwagandha capsules to placebo found small improvements in sleep quantity and quality, particularly in adults treated for more than 8 weeks. A further analysis of 12 randomised trials across 1,002 adults compared specific doses to placebo.

No serious adverse effects were reported in these trials, though reviewers note longer-term safety data are lacking. Reported side effects include diarrhoea, headache, sedation and nausea.

The working

6Quantity

5Quality

6Consistency

8Directness

(6 × 0.25) + (5 × 0.35) + (6 × 0.25) + (8 × 0.15) = 6.0
Better supported than the stress claim and more modestly promoted, which is an unusual combination.

MetaSleep quality meta-analyses summarised in Merck Manual Professional Edition, ashwagandha entry 12 RCTs, 1,002 adults · effect strongest beyond 8 weeks

Claim 04

Can cause liver injury

7.3/10

Supported·LiverTox category C · regulatory action taken

LiverTox classifies ashwagandha as a probable cause of clinically apparent liver injury — likelihood score C. A case series drawn from Iceland and the US Drug-Induced Liver Injury Network documented the pattern, and further cases have been assessed for causality using the updated RUCAM instrument.

The injury pattern is typically cholestatic. A published case describes a 22-year-old with no medical history developing severe itching, jaundice and liver injury roughly 30 hours after starting ashwagandha at the manufacturer’s recommended dose, having bought it after seeing a social media advertisement. Reported cases generally resolve after stopping.

Denmark has banned it. The French food safety agency ANSES issued a formal opinion in 2024 on the risks of Withania somnifera preparations in food supplements. This is the single most under-communicated fact about the most-hyped supplement on the market.

The working

6Quantity

7Quality

8Consistency

9Directness

(6 × 0.25) + (7 × 0.35) + (8 × 0.25) + (9 × 0.15) = 7.3
Scored as support for the harm claim. Case series with formal causality assessment, an international registry classification, and two national regulators acting on it.

CaseNIH Office of Dietary Supplements: Ashwagandha health professional fact sheet Collates LiverTox classification, DILIN case series, Denmark ban and ANSES 2024 opinion
CaseBjörnsson HK, Björnsson ES, Avula B, Khan IA, Jonasson JG, et al. Ashwagandha-induced liver injury: a case series from Iceland and the US Drug-Induced Liver Injury Network Liver Int 2020;40(4):825-829 · PMID 31991029
CaseAshwagandha-induced hepatic injury: a case report Cureus 2024 · doi 10.7759/cureus.71576 · onset at label dose

Dosage as studied

Most-studied regimen

~600 mg per day, extract

Trials vary widely by extract type, which the reviewers identify as a core limitation. Sleep benefits appeared mainly after 8 weeks.

Trial doses: Highly variable. Reviewers identify variability in the extracts studied as a major limitation on interpreting the whole literature.
Duration: Sleep improvements appeared particularly in adults treated for more than 8 weeks.
Regulatory status: Denmark has banned it. The French agency ANSES published a formal opinion in 2024 on risks of Withania somnifera preparations in food supplements.
Liver risk: LiverTox classifies ashwagandha as a probable cause of clinically apparent liver injury — likelihood score C.