How to lower cortisol: what the evidence supports

The changes with the strongest evidence for lowering cortisol are habits rather than products: regular meditation or similar mental training, regular moderate exercise, enough sleep, drinking less alcohol and support from other people. In randomised trials, months of meditation training, a year of aerobic exercise and treatment for insomnia were each followed by lower hair cortisol, which reflects your average cortisol over months rather than at a single moment.¹²³ The effects were modest, took months to appear and were not found in every study.⁴ Supplements sold to lower cortisol have far weaker evidence, and some "adrenal" products have been found to contain hormones.⁵

What "lowering cortisol" should mean

Cortisol is not a toxin to be driven as low as possible. It helps control blood sugar, blood pressure, immune activity and your response to stress, and having too little of it, as in Addison's disease, causes fatigue, low blood pressure and weight loss.⁶ In healthy people it starts rising in the early hours, peaks within about three hours of waking and falls to its lowest point around midnight.⁷ Short rises with a deadline, a hard workout or a big meal are part of normal life.

What matters for health is cortisol that stays high for weeks or months. In a meta-analysis of 66 studies, hair cortisol was 22 per cent higher in people exposed to chronic stress, and 43 per cent higher when the stress was still going on.⁸ Higher hair cortisol has also been linked with a larger body mass index, waist-to-hip ratio and blood pressure, although those links are associations, not proof of cause.⁸ So the realistic goal is to bring down a long-term average, not to flatten every peak. For background, see what cortisol does and cortisol and stress.

Meditation and mental training

The clearest trial evidence comes from mental training. In the ReSource study, 362 adults took part in a randomised trial of meditation-based mental training, and hair samples were analysed for 227 of them.¹ Hair cortisol and cortisone (the inactive form of cortisol) fell steadily over the first three to six months of training, with no further fall by nine months. The fall was larger in people who practised more often, whichever kind of practice they did, and it was unrelated to how stressed people said they felt.¹

Larger reviews point the same way, with caveats. A meta-analysis of 45 randomised trials found that meditation lowered cortisol compared with active control activities, and that focused-attention meditation, such as concentrating on the breath, did so specifically.⁹ A second analysis, of 42 trials, linked yoga-based programmes with lower evening and waking cortisol.¹⁰ But a meta-analysis that split trials by sample type found a medium effect in 10 trials that used blood, seen only in people who were physically ill, and a small, non-significant effect in 21 trials that used saliva.¹¹ The practical message: regular practice seems to matter more than style, the benefit is clearer in people under strain, and it takes months.

Regular, moderate exercise

In a 12-month trial, 130 adults aged 26 to 58 were randomly assigned either to 150 minutes a week of moderate-to-vigorous aerobic exercise or to health information alone. Hair cortisol fell in the exercise group compared with the control group, although other stress and heart-related markers did not change consistently.²

The intensity matters. A single 30-minute session raises cortisol once it is hard enough: in one study cortisol rose by about 40 per cent at 60 per cent of maximal aerobic capacity (the most oxygen your body can use) and by 83 per cent at 80 per cent, while light exercise, once other influences were corrected for, lowered it.¹² Those short-term rises are normal. Heavy, sustained training is different: 304 amateur endurance athletes had higher hair cortisol than 70 non-athletes, and the more they trained, the higher it was.¹³ The evidence supports regular moderate activity rather than ever more and harder. See cortisol and exercise.

Sleep, and treating insomnia

Sleep and cortisol are tightly linked. Sleep normally damps the stress system down, while sudden changes to your sleep pattern, such as shift work or jet lag, switch it on and disturb the daily rhythm.⁷ In a laboratory study, a short night or a missed night raised cortisol the following evening by 37 and 45 per cent respectively, and the evening fall began at least an hour later than usual.¹⁴

Treating poor sleep may help. In a trial of 74 shift-working nurses with insomnia, a digital course of cognitive behavioural therapy for insomnia (CBT-I, a structured programme that changes sleep habits and unhelpful thoughts about sleep) improved insomnia markedly, and hair cortisol in the treated group fell by 1.8 pg/mg, or 44 per cent from the starting level.³ That fall was measured within the treated group, so it is a promising signal rather than proof. A GP can advise on CBT-I. See cortisol and sleep.

Alcohol, food, caffeine and other people

• Alcohol. In the Whitehall II study of 2,693 men and 977 women, men's salivary cortisol was 3 per cent higher for each extra weekly unit of alcohol, and women who drank heavily had a larger morning rise.¹⁵ In people with alcohol dependence, hair cortisol after a period of heavy drinking was three to four times that of people who had stopped drinking or of a control group, and those who had stopped did not differ from the controls.¹⁶ Cutting down is one of the better-supported changes.
• Crash diets. Three weeks on 1,200 kcal a day increased total cortisol output in 121 women.¹⁷ Strict calorie cutting seems to act as a stress in its own right. See cortisol and weight.
• Caffeine. After five days without caffeine, a dose of caffeine caused a clear rise in cortisol; drinking 300 to 600 mg a day reduced that response but did not remove it.¹⁸ Whether regular coffee drinking changes long-term cortisol does not appear to have been studied with hair tests. See caffeine, alcohol, food and cortisol.
• Other people. In a laboratory study, people who had support from a close friend before a public-speaking stress test had a smaller cortisol rise.¹⁹ Hair studies have not found a consistent link with how much support people say they have, so the effect on long-term levels is less clear.⁸

What the evidence does not support

Many supplements are sold for "cortisol balance" or "adrenal support", and the evidence behind them is thin. A systematic review of 52 controlled trials of 26 herbs and plant extracts judged their effect on the stress hormone system unclear, because the studies were so varied. The most consistent finding was lower morning cortisol with ashwagandha, and the authors concluded that more research was needed before firm conclusions could be drawn.²⁰

"Adrenal support" products raise a different worry. When researchers tested 12 such products bought over the counter, all contained thyroid hormone and most contained at least one steroid hormone.⁵ The Endocrine Society warns that many adrenal supplements have not been tested for safety, and that taking adrenal hormones you do not need can stop your own adrenal glands working for months and lead to a life-threatening adrenal crisis.²¹ "Adrenal fatigue", the condition these products are usually sold for, is not a recognised medical diagnosis, and no test can detect it.²¹²² See adrenal fatigue.

If you take a steroid medicine, do not stop it to "reset" your cortisol. Long-term steroid use, or stopping it suddenly, is the most common cause of abnormal cortisol levels, and any change should be agreed with your doctor.²³

How to tell whether it is working

How you feel is not a reliable guide. In the ReSource trial, the fall in hair cortisol was unrelated to changes in how stressed people said they were,¹ and across studies hair cortisol does not track self-reported stress consistently.⁸ A single blood or saliva test is not much better: in a classic study, about 75 per cent of the variation in a single saliva result reflected the situation on the day.²⁴

Hair is the measure trials use to see long-term change, because it is consistent. The same people's hair cortisol correlated at r = 0.84 from one month to the next, against 0.38 to 0.61 from week to week for saliva,²⁵ and a stable personal component explained 59 to 82 per cent of the variation in repeated hair samples.²⁶ That makes a change between two hair results more likely to reflect a real shift. Hair grows about 1 cm a month, so the 3 cm nearest the scalp cover roughly the last three months.²⁷ A sensible approach is one sample before you start and another three to six months later, from the same laboratory and with no new dye or heat treatments in between.

Expect variation. A review of 14 studies with 1,916 participants that used hair cortisol to judge psychological treatments, from relaxation and mental training to CBT and trauma therapy, found mixed results, and people who started higher sometimes responded better.⁴ A result that does not move is not proof that a change has done nothing for your health. See how often to test cortisol.

When high cortisol needs a doctor

Everything above is about everyday, stress-related cortisol. Very high cortisol from a medical cause is a different matter. Cushing's syndrome caused by the body itself, rather than by medicines, is rare: European studies report two to three cases per million people a year.²⁸ Its most common cause overall is long-term, high-dose steroid medicine.²⁹ Signs include more fat on the tummy, chest, upper back and neck with thin arms and legs, a red, round face, easy bruising, stretch marks and weakness at the tops of the arms and legs.²⁹ Lifestyle changes will not treat it.

If you have signs like these, see a GP. Doctors diagnose Cushing's syndrome with late-night saliva, 24-hour urine and dexamethasone suppression tests.²⁸ A hair test does not diagnose or rule out any condition. Tiredness and low mood that do not lift can also have other causes, such as depression or obstructive sleep apnoea, which are worth checking.²¹ See high cortisol symptoms and testing for Cushing's.