How to Lower Cortisol: What the Evidence Supports

Cortisol is a hormone your adrenal glands make, and it is meant to rise in the morning and fall at night. In trials, exercise and meditation programmes lowered it modestly, caffeine raised it in people not used to it, and ashwagandha lowered it in small trials. No study has shown that a lower reading improves health in healthy people.

Evidence at a glance

• Trials: nine pooled analyses of 7 to 46 studies each: exercise, meditation, yoga, lost sleep, fasting, live cultures and ashwagandha.
• Largest single trial: 141 adults with moderate stress, two ashwagandha extracts against placebo, 8 weeks.
• Strongest measured outcome: physical-activity programmes lowered cortisol across 10 studies (standardised mean difference -0.37), rated moderate certainty by the review.
• Not tested: whether lowering cortisol in healthy people improves any health outcome.

Sources checked on 13 October 2026.

What is cortisol, and what does high cortisol feel like?

Cortisol is a steroid hormone made by the adrenal glands, which sit above the kidneys, and researchers use it as a key marker of stress.¹ At the level of one person it is a loose marker: across 19 studies that sampled saliva during daily life (38,418 moments), negative emotion correlated with cortisol at r = 0.06, and in a review of 104 daily-life studies only 11 of the 25 that tested it found a direct link between an acute stressor and the cortisol reading.^{22, 23} It follows a daily rhythm, rising sharply soon after waking and falling through the day; tired but wired explains the morning rise.

Cortisol high enough to be a disease is Cushing's syndrome, a rare condition whose commonest cause is long-term, high-dose steroid medicine.² The NHS lists extra fat on the neck, upper back, chest and tummy with thinner arms and legs, a red, round face, easy bruising, stretch marks, weak upper arms and thighs, mood changes and irregular periods.² A stressful month is not Cushing's. What high and low cortisol mean, the NHS tests and who should be seen are on what is cortisol.

What lowers cortisol, according to trials?

Activity, meditation and yoga programmes lowered cortisol in pooled trials; one lost night of sleep made no clear difference; fasting and caffeine raised it. No trial here tested a quick fix, and nothing needs "flushing": in the caffeine trial, each rise faded by evening.³ "Cortisol detox" plans sell the same idea: cortisol is a hormone, not a toxin that builds up, and the advertising code's guidance is that a detox claim for a food is a health claim, acceptable only if it appears on the GB register, where no detox or cortisol claim exists.^{10, 24} The habits apply to everyone, although the physical-activity trials were mostly in women.

The "cortisol-lowering foods" lists online have nothing like this behind them: eggs and bananas, the two most named, have no pooled cortisol evidence (PubMed, 13 October 2026), and the clearest dietary finding beyond the table runs the other way, a glucose dose before an acute stressor raising the cortisol response across 21 studies of 1,216 people (d = 0.30).²⁵ The NHS stress busters start with being active and connecting with people, and say "Do not rely on alcohol, smoking and caffeine as your ways of coping".⁹ Sleep evidence is in the Wellgard sleep guide.

Do supplements lower cortisol?

No food supplement in Great Britain may claim to lower cortisol: no authorised health claim mentions it.¹⁰ What trials found for the ones people ask about most:

• Ashwagandha. Across 23 pooled randomised trials of 1,706 adults, cortisol was lower than on placebo (standardised mean difference -1.18), with results varying between trials.¹¹ A 2023 review of nine trials found lower cortisol; none looked at adrenal function or long-term effects.¹² A 2025 pooling of 488 people found cortisol lower than on placebo but perceived stress not significantly different, a reminder that a lower reading and feeling less stressed are two different measurements.²¹ The most-quoted trial, 64 stressed adults over 60 days, on an extract its manufacturer supplied, lowered serum cortisol more than placebo.¹³ In the 141-adult trial cortisol fell in every group, placebo included;¹⁴ a 12-week trial found no difference from placebo in perceived stress,¹⁵ and a 30-day trial found no group difference in stress scores.¹⁶ It must not be taken in pregnancy or while breastfeeding (who should not take ashwagandha). The detail is in the Wellgard ashwagandha guide; anxiety and memory results are in what ashwagandha does for the brain.
• Live cultures. Across 46 randomised trials of 3,516 people, cortisol was lower than in the control groups, a result the review rated low certainty.¹⁷
• Magnesium. One small analysis, made after a 24-week trial of 49 overweight adults and part-funded by a magnesium producer, found lower 24-hour urine cortisol on magnesium than on placebo.¹⁸

No trial has compared these with each other or with exercise on cortisol (PubMed, 13 October 2026), so none can be ranked. Combinations are covered in what can you take with ashwagandha.

A "cortisol blocker" means two different things. In medicine, a drug that reduces cortisol production is a prescription treatment for Cushing's syndrome: the NHS lists "medicine to control the amount of cortisol your adrenal glands produce" among its treatments.² A supplement sold as a "cortisol blocker" is a food supplement: no trial has tested a product sold under that name (PubMed, 3 November 2026), no food may claim to lower cortisol,¹⁰ and in July 2025 the Advertising Standards Authority upheld complaints against a drink whose advert said its magnesium reduced serum cortisol.²⁶ The ingredients in the drink ruled on, ashwagandha, magnesium and L-theanine, have their own trials: ashwagandha and magnesium above, L-theanine in what L-theanine does.

What is "cortisol belly", and can you get rid of it?

"Cortisol belly" is a social-media term, not a diagnosis. Belly fat has many causes, and a waistline is not a cortisol test. One measurement shows how loose the link is: across 122 cohorts of 26,527 people, long-term cortisol in hair correlated with body mass index at r = 0.10.²⁷ "Cortisol face" is the same kind of term; the NHS lists "a red, round face" among the signs of Cushing's.² The pattern the NHS links to too much cortisol is more specific: fat on the neck, upper back, chest and tummy with thin arms and legs, plus the other signs above.² No food supplement may claim to shift it. With no definition, it has never been a trial outcome; the ashwagandha trials measured cortisol, stress scores and, in a few, body weight, reported in the Wellgard ashwagandha guide. If the NHS pattern sounds familiar, see a GP.

Should you test your cortisol at home?

Cortisol tests diagnose disease. If a GP suspects Cushing's syndrome, they can arrange a blood, urine or saliva test, and a high result can lead to a hormone specialist.² For Addison's disease, a specialist checks blood before and after an injection.¹⁹ A reading moves with the hour, with caffeine³ and even on placebo,¹⁴ so a home result without symptoms has nothing to compare with. A one-day saliva curve is an unsteady measure in its own right: across 11 studies of 3,307 people, the day-to-day reliability of the five common diurnal cortisol measures ranged from poor to good, and the morning awakening response was the least reliable.²⁸ Taking thyroid medicine? Ashwagandha with levothyroxine covers whether to test cortisol first.

Can cortisol be too low?

Yes. In Addison's disease, a rare condition, the adrenal glands do not make enough hormones. The NHS lists extreme tiredness, darker patches of skin, loss of appetite, feeling sick, cravings for salty foods, dizziness on standing and muscle weakness.¹⁹ A common cause of a low reading is a steroid medicine such as prednisolone, which switches off the body's own production: across 12 studies of 1,170 adults who had taken a course of systemic steroids, 30% had adrenal insufficiency when first tested, and in a wider review it persisted in 15% of people retested three years after stopping.^{29, 30} The NHS says: "Do not stop taking your steroid medicine without talking to your doctor first."³¹ One UK case report describes a 37-year-old woman with very low cortisol and sodium who had been taking ashwagandha; she recovered with hospital treatment, and a single case cannot show the supplement caused it.²⁰ "Adrenal fatigue" is not a recognised diagnosis: no endocrinology society accepts it.³² Persistent tiredness is a GP matter, and the pattern behind the label is explained in tired but wired. One caution about "adrenal support" products: when a US laboratory analysed 12 of them bought over the counter, all 12 contained the thyroid hormone T3 and 3 contained the steroid budesonide, hidden ingredients the authors flagged as a potential risk.³³ A lower reading is not a goal in itself.

When should you see a GP about cortisol?

See a GP if you have the Cushing's symptoms above, especially if you take steroid medicine,² or think you might have Addison's disease.¹⁹

What we don't know

• No study has shown that lowering cortisol improves health in healthy people.
• No trial has compared a habit such as exercise with a supplement for cortisol.
• None of the nine ashwagandha trials in the 2023 review looked at adrenal function or long-term effects.¹²
• The physical-activity trials included few men and no older adults.¹

Frequently asked questions

What drink reduces cortisol?

No drink has an authorised claim to lower cortisol.¹⁰ The drink with the clearest data raises it: in a trial of 96 adults, caffeine raised cortisol after five days without it, and regular intake removed the morning rise but not the afternoon one.³ The "cortisol cocktail" or "adrenal cocktail", a social-media drink, has no trial: no study has given it to anyone (PubMed, 3 November 2026). On salt alone, in a pilot of 8 adults, a single high-salt meal raised cortisol only when it was eaten second,³⁴ and in a diet study of 28 adults a high-salt diet lowered morning cortisol in salt-resistant people and did not change it in salt-sensitive ones:³⁵ tiny, inconsistent, and no reason to add salt to anything. The NHS limit is 6 g of salt a day, about a level teaspoon.³⁶

Can magnesium lower cortisol?

One small analysis, made after a trial ended, suggests it might: 49 overweight adults had lower urine cortisol after 24 weeks on magnesium than on placebo, in a study part-funded by a magnesium producer.¹⁸ An earlier trial points the same way as a secondary result: in 46 older adults with insomnia, 500 mg of magnesium a day as the oxide for 8 weeks lowered serum cortisol against placebo alongside better sleep scores, cortisol a secondary measure there too.³⁷ Magnesium contributes to normal psychological function and to the reduction of tiredness and fatigue.¹⁰ Its sleep trials are in does magnesium help you sleep.

Which vitamin reduces cortisol?

No vitamin is authorised to lower cortisol. Pantothenic acid contributes to normal synthesis and metabolism of steroid hormones, vitamin D and some neurotransmitters, which describes its role in making hormones such as cortisol. Vitamins B6, B12 and C contribute to normal psychological function.¹⁰ No vitamin deficiency appears among the NHS causes of Cushing's syndrome.²

Is ashwagandha or L-theanine better for lowering cortisol?

No trial has compared them on cortisol (PubMed search, 13 October 2026), so neither can be ranked. Ashwagandha's cortisol results are pooled above; L-theanine's own trials are in what L-theanine does.

Does exercise raise cortisol?

A hard session raises it for an hour or two, and that is normal. Across 10 controlled studies (213 people) and 50 before-and-after groups (677 people) of a single high-intensity interval session, cortisol was higher immediately afterwards and at 30 and 60 minutes, below baseline at 120 and 180 minutes, and back at baseline by 24 hours.³⁸ Regular activity programmes lowered cortisol modestly in the pooled trials above.¹ No trial has tested the "cortisol release" routines, the breathing or shaking videos that go by that name online (PubMed, 3 November 2026).

Sources

1. De Nys L et al. Psychoneuroendocrinology 2022. Pooled analysis of 10 studies. PubMed 35777076
2. NHS. Cushing's syndrome. Page last reviewed 19 March 2025. nhs.uk/conditions/cushings-syndrome
3. Lovallo WR et al. Psychosomatic Medicine 2005. Randomised crossover trial, 96 adults, 4 weeks; US government-funded. PubMed 16204431
4. Pascoe MC et al. Journal of Psychiatric Research 2017. Pooled analysis of 45 randomised trials. PubMed 28863392
5. Koncz A et al. Health Psychology Review 2021. Pooled analysis of randomised trials: 10 with blood, 21 with saliva samples. PubMed 32635830
6. Pascoe MC et al. Psychoneuroendocrinology 2017. Pooled analysis of 42 randomised trials. PubMed 28963884
7. Chen Y et al. Endocrine Journal 2024. Pooled analysis of 24 studies, 21 of them crossover. PubMed 38777757
8. Nakamura Y et al. Stress 2016. Pooled analysis of 13 studies, 357 people. PubMed 26586092
9. NHS. 10 stress busters. Page last reviewed 26 June 2026. nhs.uk/mental-health/self-help/guides-tools-and-activities/tips-to-reduce-stress
10. Department of Health and Social Care. GB nutrition and health claims register, 19 May 2026. gov.uk/government/publications/great-britain-nutrition-and-health-claims-nhc-register
11. Fornalik M et al. Planta Medica 2026. Pooled analysis of 23 randomised trials, 1,706 adults. PubMed 41740946
12. Della Porta M et al. Nutrients 2023. Systematic review of 9 trials, 30 to 112 days. PubMed 38140274
13. Chandrasekhar K et al. Indian Journal of Psychological Medicine 2012. Randomised trial, 64 adults with chronic stress, 60 days; extract supplied by its manufacturer. PubMed 23439798
14. West RE et al. Journal of Medicine and Life 2026. Randomised three-arm trial, 141 adults randomised and 130 analysed, 8 weeks; the proprietary extract tested is a named maker's product; no funding statement; no conflicts declared. PubMed 41815853
15. Smith SJ et al. Journal of Psychopharmacology 2023. Randomised trial, 120 adults aged 40 to 75, 12 weeks; manufacturer-funded. PubMed 37740662
16. Remenapp A et al. Journal of Ayurveda and Integrative Medicine 2022. Randomised three-arm trial, 60 adults, 30 days; funded by the extract's maker. PubMed 34838432
17. Jain M et al. Nutrients 2024. Pooled analysis of 46 randomised trials, 3,516 people. PubMed 39458560
18. Schutten JC et al. Clinical Endocrinology 2021. Post-trial analysis of a randomised trial, 49 overweight adults, 24 weeks; part-funded by a magnesium producer. PubMed 33030273
19. NHS. Addison's disease. Page last reviewed 19 September 2025. nhs.uk/conditions/addisons-disease
20. Patel M et al. British Journal of Hospital Medicine 2024. Case report, one woman. PubMed 38941978
21. Albalawi AA. Nutrition and Health 2025. Pooled analysis of 7 cortisol and 6 perceived-stress trials, 488 people. PubMed 40746175
22. Joseph NT et al. Neuroscience and Biobehavioral Reviews 2021. Systematic review and meta-analysis of 22 daily-life studies; funding not stated on PubMed. PubMed 33662445
23. Weber J et al. PLoS One 2022. Systematic review of 104 daily-life studies; no competing interests declared. PubMed 35895674
24. Committee of Advertising Practice. Advice online: Detoxing (general). asa.org.uk/advice-online/detoxing-general
25. Kördel M et al. Neurobiology of Stress 2025. Systematic review and meta-analysis, 21 studies, 1,216 people; no competing interests declared. PubMed 41140644
26. Advertising Standards Authority. Ruling A25-1277097 on a website for a magnesium drink, three issues upheld, 16 July 2025: cortisol-reduction and calm claims were unauthorised health claims. Regulator ruling; search the reference on the asa.org.uk rulings page.
27. van der Valk E et al. Obesity Reviews 2022. Systematic review and meta-analysis, 146 cohorts, 34,342 people, of which 122 cohorts and 26,527 people in the body mass index pooling; no conflicts declared. PubMed 34811866
28. Norton SA et al. Comprehensive Psychoneuroendocrinology 2023. Meta-analysis of 11 studies, 3,307 people, plus two cohorts; no competing interests declared. PubMed 37635863
29. Bowley J et al. European Journal of Internal Medicine 2026. Systematic review and meta-analysis, 12 studies, 1,170 adults; one author reports a pharmaceutical research grant. PubMed 41521083
30. Joseph RM et al. Seminars in Arthritis and Rheumatism 2016. Systematic review, 73 studies; Medical Research Council, Arthritis Research UK and Wellcome Trust funded. PubMed 27105755
31. NHS. Steroid tablets. Page last reviewed 19 June 2025. nhs.uk/conditions/steroid-tablets
32. Cadegiani FA, Kater CE. BMC Endocrine Disorders 2016. Systematic review, 58 studies; no funding; an erratum of 16 November 2016 corrects one table cell. PubMed 27557747
33. Akturk HK et al. Mayo Clinic Proceedings 2018. Laboratory analysis of 12 products bought in 2016; funding not stated on PubMed. PubMed 29502560
34. Nowell A et al. European Journal of Nutrition 2024. Randomised crossover pilot, 8 adults; a research-council and heart-foundation early-career fellowship; no conflicts declared. PubMed 38409436
35. Kerstens MN et al. Journal of Clinical Endocrinology and Metabolism 2003. Dietary trial, 28 adults; research support from a source other than the US government listed on PubMed. PubMed 12970284
36. NHS. Salt: the facts. Page last reviewed 17 April 2023. nhs.uk/live-well/eat-well/food-types/salt-nutrition
37. Abbasi B et al. Journal of Research in Medical Sciences 2012. Randomised placebo-controlled trial, 46 adults, 8 weeks; funded by a university nutrition research institute; no conflicts declared. PubMed 23853635
38. Dote-Montero M et al. Scandinavian Journal of Medicine and Science in Sports 2021. Systematic review and meta-analysis, 10 controlled studies and 50 before-and-after groups; Spanish government grants. PubMed 34022085

Spotted something out of date? Email support@wellgard.co.uk with the page address and the sentence.