Ashwagandha and Cortisol: Why Extract Matters More Than the Herb

Ashwagandha is one of the most heavily marketed nervous-system supplements on the market, and — unusually for a herb with that much marketing behind it — the human trial evidence for stress and cortisol is genuinely strong. It's also an ingredient where the specific extract matters far more than the name on the front of the bottle. Here's what the research actually shows, why the picture is more nuanced than "ashwagandha lowers cortisol," and where the form we currently sell falls short of the standard this article argues for.

What the trials show

A 2021 systematic review pooled seven randomized, placebo-controlled human trials covering 491 adults, all using ashwagandha for six to eight weeks.¹ Across those trials, ashwagandha significantly reduced self-reported stress and anxiety levels compared to placebo, and — importantly, because this is an objective, measurable biomarker rather than a self-report — significantly reduced serum cortisol levels compared to placebo. A separate randomized, placebo-controlled trial measuring stress-relieving and pharmacological effects of a high-concentration ashwagandha extract found consistent results using validated stress-scale and cortisol measurements.² A further randomized, double-blind, placebo-controlled trial in healthy, stressed adults extended the picture to cognitive function, finding measurable improvements on tasks related to attention and processing speed alongside the stress reduction.³

A separate 2021 meta-analysis of five trials (372 adults, healthy or with insomnia) found a small but statistically significant improvement in sleep, using KSM-66 root extract at 250–600mg/day or Shoden root-and-leaf extract at 120mg/day over six to twelve weeks.¹ The reviewers noted the benefit was more pronounced at 600mg/day, with at least eight weeks of use, and in people who actually had insomnia rather than healthy sleepers — which is the useful way to read it: the sleep effect is real, modest, and concentrated in the people with a problem to fix.

Now the limitation that belongs in the same breath as those numbers, and which most ashwagandha write-ups leave out. All seven stress trials and all five sleep trials were conducted in India, where the branded extracts are manufactured.¹ That is not a reason to dismiss the findings. It is a reason to hold them a little more loosely than the effect sizes alone suggest: an evidence base concentrated in one country, around a small number of commercially branded ingredients, is exactly the shape that tends to shrink when independent groups elsewhere try to replicate it. Trials of branded extracts are also frequently funded or supplied by the extract manufacturer, so the funding statement is worth reading before the abstract. We flag this pattern because it recurs across the category — the same issue shows up in the rhodiola-versus-ashwagandha comparison, and cortisol itself is a messier marker than the headlines allow, which we've unpacked in the cortisol story.

Ashwagandha is one of the few adaptogens where the cortisol drop shows up in blood tests, not just on a stress questionnaire.

The part that gets lost in translation: extract matters

Here's the detail that most ashwagandha marketing glosses over. Of the seven trials in that 2021 review, three used a specific standardized root extract called KSM-66, two used a different standardized root-and-leaf extract (Sensoril/Shoden), and one used plain whole-root powder at a much higher dose (12,000mg/day, roughly equivalent to 6,000mg of powder) to get a comparable effect.¹ Doses studied ranged from 240mg to 1,250mg/day of extract, and the review specifically noted that benefits were more consistent at 500-600mg/day than at lower doses. In other words: "ashwagandha" on a label is not one thing. A capsule of generic root powder at an arbitrary dose is not equivalent to the standardized, tested extracts that produced these results — the plant species is the same, but the concentration of active withanolides, and therefore the dose that was actually studied, is not.

The practical upshot: check the label for "KSM-66" or "Sensoril/Shoden" by name and a dose in the 500-600mg/day range. If a product just says "ashwagandha root powder" with no standardization and no withanolide percentage, it isn't the thing the trials tested — and the effect size may not transfer. This is also why comparing two "ashwagandha" products by price per capsule is close to meaningless; a 500mg capsule of a standardized 5% withanolide extract and a 500mg capsule of raw root powder are not delivering the same active dose, even though the label looks identical at a glance. This is the same trap we walk through in how to read a supplement label, and the reason a certificate of analysis is the only way to confirm a withanolide percentage rather than take it on trust.

Testosterone and male reproductive health: a newer, narrower line of evidence

Beyond stress and sleep, a smaller but growing body of research has looked specifically at KSM-66 and male reproductive hormones. A randomized, double-blind, placebo-controlled trial published in Health Science Reports in 2022 gave 50 men reporting low sexual desire (average age around 35) either 300mg of KSM-66 ashwagandha root extract twice daily (600mg/day total) or placebo for eight weeks.⁴ That inclusion criterion matters and we got it wrong the first time: these were not men picked off the street, they were men who had presented with a specific complaint, and a hormone response in that group does not automatically transfer to men without it.

The reported results: a 17% increase in serum testosterone in the ashwagandha group versus 2% on placebo, and a 40% improvement in self-reported sexual-function scores (DISF-M) versus 25% on placebo.⁶ Prolactin didn't change in either group, which matters because it suggests the effect isn't being driven by some unrelated hormonal disruption. But read how each outcome was reported, because they are not the same strength of claim. The DISF-M improvement was conventionally significant (mean difference −9.8; 95% CI, −10.73 to −8.87; p<0.05).⁴ The testosterone comparison — the number that gets screenshotted — was expressed as an 88.5% probability of superiority over placebo, which is not the same thing as p<0.05 and is a good deal short of it.⁶ An 88.5% probability of superiority means roughly a one-in-nine chance the true direction of the effect is the other way. That is a signal worth following up. It is not a finding to build a testosterone claim on, and we aren't going to.

One trial of this size is a meaningful data point, not proof of a durable, population-wide testosterone effect — it needs replication in larger and more diverse samples, in men who weren't selected for a sexual-health complaint, before it belongs in a marketing headline. But it's consistent with the broader pattern in this ingredient's research: extract-specific, dose-specific effects that at least show up on objective bloodwork rather than only on a mood questionnaire. If testosterone is the thing you're actually trying to move, the evidence hierarchy puts sleep and training well above any capsule — which is the argument in men's brain health: start with sleep, not testosterone.

The safety side: rare but real liver case reports

No responsible discussion of ashwagandha is complete without addressing liver safety, because the case-report literature here is real and has grown over the past several years. The NIH's LiverTox database — a clinical reference on drug- and supplement-induced liver injury — documents roughly two dozen published cases of clinically apparent liver injury attributed to ashwagandha as of its most recent review, and assigns it a likelihood score of B — "a likely cause of clinically apparent liver injury."⁵ The NIH Office of Dietary Supplements describes five of those cases in detail, at daily doses from 450mg to 1,350mg taken for anywhere between one week and four months — note that the low end of that range sits inside the dose this article recommends looking for.¹ The typical pattern is a cholestatic or mixed-pattern liver injury with jaundice, appearing two to twelve weeks after starting the supplement and usually resolving within one to four months of stopping it, though rare cases involving people with pre-existing liver disease have been more severe. This is not a reason for alarm at a population level — millions of doses have been consumed against a few dozen documented cases — but it is a reason to skip ashwagandha (or check with a doctor first) if you have any existing liver condition, drink heavily, or take other medications processed by the liver, and to stop and seek care if you notice unusual fatigue, dark urine, or yellowing skin or eyes while taking it. It's also worth noting the causality caveat LiverTox itself puts on this: because commercial herbal preparations are often mixtures, and can be mislabelled or contain undeclared ingredients, it is not always clear that the reported cases were caused by ashwagandha at all. In several of them, though, the actual product the patient took was tested and found to contain ashwagandha with no contaminants — which is the detail that stops this being dismissible as a purely adulteration story and points at the herb itself.⁵ Either way, a chain-of-custody certificate of analysis isn't just a purity nicety — it's part of what makes a liver-injury signal interpretable at all, rather than a mystery of "which ingredient in this blend did it."

Comparing the standardized extracts

Where Proco's ashwagandha actually stands

This started as a straightforward sourcing problem, not an evidence problem. Ashwagandha's evidence is about as strong as any ingredient we cover — multiple randomized, placebo-controlled trials, consistent effect sizes, and an objective biomarker (cortisol) backing up the subjective stress-scale data, with the India concentration and the branded-extract funding pattern noted above as the standing caveat — and it fits our category (nervous-system and stress support) outright. But that strong evidence is specific to the standardized extracts actually used in the trials (KSM-66 or Sensoril/Shoden), not to unstandardised ashwagandha root generally — a sourcing distinction, not an evidence-bar one.

Ashwagandha is in Sleep Support today. The label lists it as "Ashwagandha (Withania somnifera) (root)" inside a 14-ingredient botanical blend with a combined weight — no branded extract named, no withanolide percentage, and no individual milligram amount. So by this article's own test we cannot tell you it is the thing the trials tested, and we're not going to imply otherwise by calling it a "standardized extract" or by printing a dose we can't verify. That is a real gap between our stated standard and what's on the shelf, and naming it is the point of this section.

What that means in practice: it stays in Sleep Support for now while per-ingredient disclosure is worked through with the manufacturer; moving it to a named, studied extract (KSM-66 or Sensoril/Shoden) at a disclosed dose is on the roadmap and has not shipped. We do not sell ashwagandha as a standalone product. Given the liver case reports above, once dosing is disclosed we'll also be direct on the label about who should talk to a doctor first — and in the meantime, if stress is the problem you're actually trying to solve, the interventions with the strongest evidence aren't capsules at all: those are in managing stress without medication, and the narrower question of what magnesium can and can't do for the stress response is in stress, cortisol and the nervous system.

References

1. "Ashwagandha: Is it helpful for stress, anxiety, or sleep?" Office of Dietary Supplements, National Institutes of Health. ods.od.nih.gov
2. "An investigation into the stress-relieving and pharmacological actions of an ashwagandha (Withania somnifera) extract: A randomized, double-blind, placebo-controlled study." PubMed, 2019. pubmed.ncbi.nlm.nih.gov
3. "Efficacy and Safety of Ashwagandha Root Extract on Cognitive Functions in Healthy, Stressed Adults: A Randomized, Double-Blind, Placebo-Controlled Study." PubMed, 2021. pubmed.ncbi.nlm.nih.gov
4. "Effect of standardized root extract of ashwagandha (Withania somnifera) on well-being and sexual performance in adult males: A randomized controlled trial." Health Science Reports / PubMed, 2022. pubmed.ncbi.nlm.nih.gov
5. "Ashwagandha." LiverTox: Clinical and Research Information on Drug-Induced Liver Injury, NIH National Institute of Diabetes and Digestive and Kidney Diseases. ncbi.nlm.nih.gov
6. "Ashwagandha root extract may boost testosterone levels, sexual function in men" — reporting of the Chauhan 2022 trial's percentage changes and 88.5% probability of superiority. NutraIngredients, 18 August 2022. nutraingredients.com

This article is for informational purposes only and has not been evaluated by the FDA. It is not intended to diagnose, treat, cure, or prevent any disease. Speak with a healthcare provider before starting any new supplement, especially if you take medication or have an existing health condition.

The general version of the standard this article is measuring ashwagandha against — and how often an ingredient with good press fails it — is in why we don't sell "brain blends".

Every ingredient we sell should be dosed and disclosed at the level its supporting human trials actually used — and ashwagandha in Sleep Support is not there yet: it is an unstandardised root ingredient inside a combined blend, and moving it to a named, studied extract at a disclosed dose is work in progress rather than a claim we can make today. Explore the current range at procohq.com.