ASHWAGANDHA AFTER 50: CORTISOL MOVES. PERCEIVED STRESS MAY NOT.

Ashwagandha has one of the most consistent findings in the supplement literature: it lowers measured cortisol. It also has a finding that gets quoted far less often — in a 2025 meta-analysis, that cortisol drop did not come with any significant change in how stressed people said they felt. And the trial closest to this site's readers, overweight men aged 40 to 70, found no significant between-group difference in fatigue or vigor at all.

What the Cortisol Data Actually Shows

Two recent syntheses agree on the biomarker.

Albalawi, publishing in Nutrition and Health in 2025 (31(4):1395–1408), pooled seven randomized trials on cortisol and six on perceived stress, 488 participants total, at oral doses of 250 mg/day or more for at least two weeks. Cortisol fell by 1.16 µg/dL (95% CI −1.64 to −0.69, P < 0.001).

Fornalik and colleagues, in Planta Medica in 2026 (92(8):790–805), ran a broader hormonal review — 23 randomized placebo-controlled trials, 1,706 patients, protocol pre-registered on PROSPERO. Cortisol again fell (SMD −1.18, P < 0.04).

So the biomarker moves. The question is what that buys.

The Finding That Doesn't Get Quoted

In the same Albalawi analysis, perceived stress — measured on the Perceived Stress Scale, the thing a person actually experiences — did not significantly change: SMD −0.355, 95% CI −1.188 to 0.47, P = 0.40. The confidence interval crosses zero comfortably. Heterogeneity was moderate (I² = 50.9%), and the author notes that risk of bias was generally moderate, with some trials lacking allocation concealment.

The author's own framing is that the disconnect between the cortisol result and the PSS result is the thing worth studying, not something to explain away.

Other syntheses land differently, and that disagreement is part of the honest picture. Arumugam and colleagues in Explore (2024;20(6):103062) pooled nine RCTs and 558 patients and did find significant reductions on the Perceived Stress Scale (MD −4.72, 95% CI −8.45 to −0.99), the Hamilton Anxiety Scale (MD −2.19, −3.83 to −0.55), and serum cortisol (MD −2.58, −4.99 to −0.16). Four of the included studies reported mild to moderate adverse events.

And Akhgarjand and colleagues in Phytotherapy Research (2022;36(11):4115–4124) pooled 12 papers and 1,002 participants, finding large reductions in anxiety (SMD −1.55) and stress (SMD −1.75) — while reporting heterogeneity of I² = 93.8% and 83.1% respectively, and rating the certainty of the evidence as low for both outcomes.

Different reviewers, overlapping trials, different conclusions on the subjective outcome, with the authors themselves flagging low certainty and very high heterogeneity. That is what the state of the evidence looks like.

Strength and Muscle: Real Numbers, Young Subjects

The strength literature is built on a small number of trials that get cited constantly.

The most-cited is Wankhede and colleagues, Journal of the International Society of Sports Nutrition 2015;12:43 — 57 men aged 18 to 50 with little resistance-training experience, 300 mg of root extract twice daily for eight weeks alongside a training program. The ashwagandha group gained more on bench press 1RM (+46.0 kg, 95% CI 36.6–55.5, versus +26.4 kg, 19.5–33.3; P = 0.001) and leg extension (+14.5 versus +9.8 kg; P = 0.04), more arm cross-sectional area (+8.6 versus +5.3 cm²; P = 0.01) and chest circumference (+3.3 versus +1.4 cm; P < 0.001), more testosterone (+96.2 versus +18.0 ng/dL; P = 0.004), and a greater drop in body fat percentage (3.5% versus 1.5%; P = 0.03).

Those are substantial numbers. Note the population: untrained men, upper age limit 50.

Zhu and colleagues pooled the wider literature in Frontiers in Pharmacology (2026;17:1799467) — 20 studies, 1,249 participants, GRADE-assessed. Muscle strength improved (SMD 0.58, 95% CI 0.12–1.04, P = 0.013) with substantial heterogeneity (I² = 56.9%). Testosterone rose modestly (SMD 0.33, 0.13–0.54). Body weight and body fat percentage did not change.

The subgroup analysis is the interesting part: the strength effect was significant in physically active individuals (SMD 1.03, 0.56–1.49, I² = 0%) but not in untrained or special populations. The same review found improvements in memory (SMD 0.52), attention and processing speed (0.29), and executive function (−0.42, sign reflecting the scale direction).

A 2026 trial in Nutrition and Health by Ferreira and colleagues gave 600 mg/day for eight days to 31 male handball players (mean age 20.3 ± 3.1). Handgrip strength improved versus placebo (d = 0.33, P = 0.011). Knee flexion and extension showed no consistent effect, and perceived soreness, fatigue, and recovery did not differ between groups.

The One Trial Closest to This Audience

Lopresti, Drummond and Smith ran a 16-week randomized, double-blind, placebo-controlled crossover trial in overweight men aged 40 to 70 with mild fatigue, published in the American Journal of Men's Health (2019;13(2)). Fifty-seven enrolled, 43 completed all 16 weeks. The extract delivered 21 mg of withanolide glycosides daily for eight weeks.

What moved: DHEA-S rose 18% more than placebo (P = .005) and testosterone 14.7% more (P = .010).

What did not move: fatigue, vigor, and sexual and psychological well-being all improved over time — in both arms — with no statistically significant between-group differences. Cortisol and estradiol showed no significant between-group difference either.

The authors call for larger trials to substantiate the hormone findings. The supplement manufacturer is named in the paper's own disclosure as not having been involved in study design, analysis, or writing — which is a disclosure worth reading rather than skipping.

This is the closest thing in the literature to a trial in this readership, and on the outcomes a 50-plus reader would care about most — feeling less tired, feeling more capable — it is a null.

The Age Gap, Stated Plainly

The Akhgarjand meta-analysis reports the pooled age range across its 12 trials as 25 to 48 years. Wankhede's strength trial capped at 50. The handball trial averaged 20. Zhu's subgroup found the strength signal in physically active people specifically, and not in the untrained or special-population groups.

After searching PubMed across older-adult, elderly, aging and sarcopenia terms, I could not find a randomized trial of ashwagandha in trained adults over 50 measuring strength, power, or endurance performance. The Lopresti trial in men 40 to 70 is the nearest, and it was not a training study.

The Thyroid Question — Relevant If You Take Levothyroxine

Fornalik's hormonal review found no effect on TSH or T3, but a modest increase in T4 (MD 0.61 µg/dL, P = 0.02). Testosterone rose in men (MD 57.43 ng/dL) but not in women (5.09 ng/dL), a statistically significant difference between sexes.

Separately, Sharma and colleagues ran an eight-week randomized placebo-controlled trial of 600 mg/day in 50 people with subclinical hypothyroidism, ages 18 to 50 (Journal of Alternative and Complementary Medicine 2018;24(3):243–248). TSH, T3 and T4 all shifted significantly versus placebo.

Thyroid hormone is a titrated prescription. A supplement that shifts thyroid indices is a supplement that interacts with that titration. That is a conversation with the prescriber who set the dose, not a decision to make from a blog post.

The Safety Signal

Philips and colleagues published a case series in Hepatology Communications (2023;7(10):e0270) covering 23 patients with liver injury attributed to ashwagandha across Indian centers between January 2019 and December 2022, including eight cases from single-ingredient formulations with other causes reasonably excluded. Cholestatic hepatitis was the commonest presentation. Five patients had underlying chronic liver disease; three presented with acute-on-chronic liver failure, and all three died on follow-up. In the others, injury was prolonged but self-limiting. Chemical analysis of the retrieved formulations found only natural phytochemicals — no adulteration or contamination, meaning the injury was attributed to the herb itself rather than to something added to it.

This is a case series, not a randomized safety trial, and case series cannot establish how often something happens. What it does establish is that the cases exist and that pre-existing liver disease appeared in the worst outcomes.

The wider context: Halegoua-DeMarzio and Navarro, writing in Liver International (2025;45(3):e16071), report that herbal and dietary supplement–induced liver injury now accounts for 20% of drug-induced liver injury cases in the U.S. Drug-Induced Liver Injury Network, and list ashwagandha among the major implicated agents alongside anabolic steroids, green tea extract, garcinia cambogia, kratom and turmeric.

Denmark Banned It. That Story Has Two Sides.

The Technical University of Denmark's National Food Institute carried out a risk assessment for the Danish Veterinary and Food Administration, and Denmark prohibited ashwagandha in food supplements in 2023, on the stated grounds that a safe intake level could not be established from the available data — with concerns raised about effects on thyroid and sex hormones and about reproductive safety.

The assessment drew substantial published criticism from researchers and industry, on the grounds that it leaned heavily on animal studies, used a selective literature base, and was not peer-reviewed.

Both of those things are true at once. One EU member state's regulator concluded it could not set a safe dose; other regulators have not followed. That is where it sits.

Where This Leaves It

Reported at the strength the evidence actually has:

• Cortisol. Consistently reduced across two independent 2025–2026 syntheses covering 488 and 1,706 participants. This is the most robust finding.
• Perceived stress. Genuinely contested. One 2025 meta-analysis null; two others positive; the ones reporting benefit also report low certainty and heterogeneity above 80%.
• Strength. A real pooled effect (SMD 0.58), concentrated in physically active people, driven by trials in men under 50.
• Adults over 50. No training trial found. The one trial in men 40–70 moved hormones and did not move fatigue or vigor versus placebo.
• Thyroid. Modest T4 increase pooled; a dedicated subclinical-hypothyroid trial shifted all three indices.
• Liver. Documented case series including three deaths in patients with pre-existing chronic liver disease; listed among major agents in the U.S. DILI network.

Doses used in the trials above ranged from 250 to 600 mg/day of standardized root extract, most commonly 600 mg/day or 300 mg twice daily. That is a description of what was studied, not a recommendation.

Ashwagandha appears to shift thyroid indices and has documented case reports of liver injury. If you take thyroid medication, have any liver condition, take sedatives or immunosuppressants, or are pregnant or trying to become pregnant, this is a conversation to have with your physician or pharmacist before starting anything — not a decision to make from a blog post.

Battle Hard. — Will Power