Skip to content
Cortisol guide

How to lower cortisol: what the studies measured, morning included.

Most advice treats cortisol as a number to push down. It is a rhythm: high after you wake, low by bedtime. Here is what the trials measured for each common habit, how big the effect was, and how far you can lean on it.

The short answer

Lower it at the right hour, not all day. Cortisol normally rises about 50% in the 30 minutes after waking and falls toward bedtime. The best-supported habits protect that shape: enough sleep (one 4-hour night raised next-evening cortisol 37%), regular exercise, an earlier last coffee, less alcohol, and time with people and outdoors. Meditation's effect is small in healthy adults. No habit treats a diagnosed condition. (Source: Wüst and colleagues, 2000; Leproult and colleagues, 1997.)

Read this first: when cortisol is not a habit question.

If you take steroid medicine by any route, do not stop it or cut the dose to lower your cortisol. That means tablets such as prednisone, and also inhalers, nasal sprays, skin creams and joint injections. A 2015 meta-analysis of 74 studies found adrenal insufficiency, meaning too little cortisol of your own, after steroid use by every route it examined, from 4.2% of nasal users to 52.2% after joint injections, and no route, dose or duration for which it could be ruled out. MedlinePlus lists abruptly stopping long-term steroid medicine as a cause of adrenal crisis. Any change goes through the clinician who prescribed it. More in coming off steroids.

Adrenal crisis is an emergency. Call 911 or your local emergency number. MedlinePlus lists its signs as severe weakness or fatigue, confusion or loss of consciousness, dizziness or light-headedness, nausea or vomiting, abdominal or flank pain, headache, high fever, low blood pressure, and a fast heart rate and breathing. You do not need every sign on this list to be in danger. Infection, severe dehydration and abruptly stopping long-term steroid medicine are among its triggers. If you have Addison's disease or hypopituitarism and cannot take your steroid medicine for any reason, vomiting included, MedlinePlus says to contact your health care provider, and to call 911 if crisis signs appear.

If you have features of too much cortisol, see a clinician before you try to change it with habits. The Endocrine Society guideline names the features that best discriminate Cushing's syndrome: easy bruising, a flushed red face, weakness in the muscles closest to the trunk (trouble rising from a chair, for example), and stretch marks that are reddish-purple and wider than 1 cm. It recommends testing people whose features are multiple and getting worse, or unusual for their age, such as osteoporosis or high blood pressure in a younger adult. If you already have a diagnosed cortisol or adrenal condition, your clinician sets the plan. Nothing on this page replaces it. The signs that send people looking are in signs of high cortisol and signs of low cortisol.

Should you try to lower cortisol at all?

Usually the aim is timing, not a lower number. Cortisol is supposed to be high after waking and low near bedtime. In their review of the daily rhythm, Adam and colleagues describe the typical day: levels surge an average of 50 to 60% in the 30 to 40 minutes after waking, drop quickly over the next few hours, then fall more slowly to a low around bedtime.

Lower is not automatically better. In a 2015 University of Colorado and Harvard study, 17 healthy adults lived on a laboratory schedule for 25 days that, for eight of them, put sleep out of step with the body clock. Their 24-hour cortisol went down, not up, while markers of inflammation rose. A night with no sleep, in the same study, pushed cortisol up. Both are disruptions, and they moved the number in opposite directions.

What researchers link with poorer health is a flat day. Adam's meta-analysis of 80 studies found flatter slopes from morning to evening associated with worse outcomes in 10 of 12 categories, including depression, fatigue, inflammation and mortality. The average association was modest (r = 0.147), and the authors discuss whether a flat slope is a marker of illness, a cause of it, or both.

So the better question is not "how do I lower my cortisol?" but "is it rising and falling at the right times?" That is also why a "cortisol detox" isn't a real thing: there is nothing to flush out. And if you arrived here from adrenal fatigue, the term is not a recognized diagnosis, though the exhaustion behind it is real.

A typical daily cortisol curve with a clear morning peak and a steady drop to a low at night.
The shape matters more than the level: a clear rise after waking and a steady fall to a low around bedtime.

How do you lower cortisol in the morning?

For most people, you don't want to. The morning rise, called the cortisol awakening response, is normal physiology. Pooling four studies of 509 adults, Wüst and colleagues found an average increase of about 50% within 30 minutes of waking. About 75% of people were "responders" whose levels clearly rose. Women's rise was the same size as men's, but their levels came down later.

Morning light makes the rise larger, not smaller. In a 1999 Amsterdam study, 14 healthy men spent the first hour after waking either in darkness or under 800 lux of light. Light added about 35% to the rise, and cortisol was higher 20 and 40 minutes after waking. The same light in the late evening had no effect. In a 2001 study of 8 men kept awake for 36 hours, switching from dim to bright light between 5 and 8 a.m. raised cortisol by more than 50% at once, while the same 3 hours of light from 1 to 4 p.m. did nothing.

Light studies do not all agree. Light has been reported to raise, lower or leave cortisol unchanged, and Jung and colleagues attribute the disagreement to intensity, duration and timing: in their study, 6.7 hours of very bright light during the biological night and morning lowered it. The common thread is timing. Morning daylight is advice about when: it strengthens the rise at the hour it is meant to happen.

Your morning coffee is probably not the lever. In a double-blind trial, people who had taken 300 or 600 mg of caffeine a day for five days showed no cortisol response at all to a 9 a.m. dose. The afternoon dose was a different story, covered in the caffeine section below.

What does a healthy morning rise look like, and what is a problem?

The typical pattern is a rise after waking, then a fall through the day to a low near bedtime. Problems with cortisol are diagnosed by clinicians with timed tests, not read from one morning. For suspected Cushing's syndrome, the Endocrine Society guideline recommends starting with one of: a urine cortisol test, a late-night saliva cortisol, or a dexamethasone suppression test, which checks whether a dose of the steroid dexamethasone suppresses the body's own cortisol. An abnormal result leads to an endocrinologist and a second test. For suspected adrenal insufficiency, MedlinePlus lists blood tests, including a low cortisol level, and a cosyntropin (ACTH) stimulation test.

Two things follow. A high reading soon after waking is taken at the hour cortisol is meant to be high, so on its own it tells you little. And a normal-looking morning does not rule anything out: if you have the features in the box at the top of this page, see a clinician whatever a home test shows. When to sample, and why the hour matters so much, is in the best time to test cortisol, and what laboratories call normal is in normal cortisol levels.

Which habits lower cortisol, and how strong is the evidence?

Every row below is a study or a pooled analysis we opened and read, with what it measured and in whom. The last column is our plain grading of how far you can lean on it, based on design, size and who was studied.

HabitWhat was measuredEffect on cortisolStrength
Enough sleepLeproult, 1997: young men, blood cortisol from 6 to 11 p.m. after a normal night, a 4-hour night or no sleep37% higher the next evening after a 4-hour night, 45% after none; the evening low arrived at least 1 hour laterModerate: controlled, small, men only
Regular sleep timingWright, 2015: 17 adults over 25 days; 8 had sleep moved out of step with their body clock24-hour cortisol fell rather than rose; inflammation markers roseWeak for this question: no trial we found isolates regular timing alone
Morning lightScheer, 1999: 14 men woke to 800 lux or darkness. Leproult, 2001: 8 men, bright light 5 to 8 a.m.Morning rise about 35% larger; an immediate rise of more than 50%. No effect in the evening or afternoonConsistent: light strengthens the morning rise rather than lowering it
Regular exerciseDe Nys, 2022: meta-analysis of 10 trials of activity programmes against controlsLower, standardized mean difference −0.37 (small to moderate)Moderate, by the authors' rating; mostly women with breast cancer
One hard sessionHill, 2008: 12 trained men, 30 minutes at 40%, 60% and 80% of maximal oxygen uptakeNo rise at 40%; +39.9% at 60%; +83.1% at 80%Moderate: controlled, small. A short rise after hard effort is normal
Caffeine timingLovallo, 2005: 96 adults, 250 mg at 9 a.m., 1 p.m. and 6 p.m. after 5 days at 0, 300 or 600 mg a dayDaily users: no response to the 9 a.m. dose; raised again from 1 to 7 p.m. after the 1 p.m. dose; back to control levels in the eveningModerate: double-blind crossover
Less alcoholBadrick, 2008: 2,693 men and 977 women, 6 saliva samples from waking to bedtimeMen: 3% higher per weekly unit; heavy drinkers: flatter fall across the day; heavy-drinking women: larger morning riseAssociation only: a single snapshot cannot show cause
MeditationPascoe, 2017: 45 trials against active controls. Koncz, 2021: 31 trials. Sanada, 2016: 5 trials, 190 healthy adultsLower overall (Pascoe); medium effect in blood only in people with a physical illness, small and not significant in saliva (Koncz); g = 0.41 in healthy adults (Sanada)Mixed: small in healthy people
Slow breathingMa, 2017: 40 adults, 20 sessions over 8 weeks at about 4 breaths a minuteLower midday saliva cortisol after training; no change in controlsWeak: one small pilot trial
Support from a friendHeinrichs, 2003: 37 men in a laboratory stress test, with or without their best friend beforehandSmaller cortisol response to the stress test with supportModerate for acute stress; laboratory, men only
Time in natureHunter, 2019: 36 city dwellers over 8 weeks. Antonelli, 2019: 22 forest studies, 8 pooledAn extra 21.3% an hour of fall beyond the normal daily decline; forest groups lower, but already lower beforehandWeak to moderate: no indoor-break comparison; expectation may play a role

Each row is detailed in the sections below and linked under References. "Strength" is our grading, not the authors', except where the row says otherwise.

Does more sleep lower cortisol?

Losing sleep raises it the next evening, the hours when it should be falling. Leproult and colleagues at the University of Chicago tracked blood cortisol in young men after a normal night (11 p.m. to 7 a.m.), a short night (4 to 8 a.m.) and no sleep at all. Between 6 and 11 p.m. the following day, cortisol was 37% higher after the short night and 45% higher after no sleep, and the quiet overnight phase started at least an hour late. The authors concluded that even partial sleep loss slows the system's recovery from its morning high.

Regularity is less tested than people assume. We did not find a controlled trial that changed sleep timing alone and measured cortisol. The closest evidence is the 2015 misalignment study above, where a schedule out of step with the body clock lowered 24-hour cortisol and raised inflammation, and the travel data in our guide to jet lag and cortisol, where flying across time zones leaves the morning peak on home time for days. A steady wake time is sound advice for the rhythm; its effect on cortisol by itself has not been measured. Shift workers face the hardest version of this, covered in night shifts and cortisol, and the full night-time picture is in cortisol and sleep.

Does exercise lower cortisol?

Regular activity programmes are linked with lower cortisol. A single hard session raises it. A 2022 meta-analysis from the University of Stirling pooled 10 trials of physical activity programmes against controls and found lower cortisol with a standardized mean difference of −0.37, a small-to-moderate effect, which the authors rated as moderate-certainty evidence. They warn against generalizing: most participants were women with breast cancer, few were men, and none were older adults.

Within one session there is a threshold. Hill and colleagues had 12 trained men exercise for 30 minutes at three intensities. Blood cortisol rose 39.9% at 60% of maximal oxygen uptake and 83.1% at 80%, did not rise significantly at 40%, and once the authors corrected for fluid shifts and time of day, the easy session lowered it. That short rise after hard effort is normal physiology. How long it lasts, what late sessions and lifting do, and how fitness changes it are in does exercise raise cortisol.

The popular add-ons fare worse. Our review of the heat studies found sauna lowers cortisol sometimes, not reliably. For cold water, no published study shows that cold plunging lowers cortisol, and the plunge carries real risks set out in that guide. And creatine did not change the cortisol response in controlled trials.

Does cutting caffeine lower cortisol?

If you drink coffee every day, the afternoon cup is the one still moving cortisol. Caffeine is a clear stimulus on its own: in a 1996 placebo-controlled study of 47 healthy young men at rest, 3.3 mg per kilogram of body weight (about 2 to 3 cups of coffee) raised cortisol to a peak 30% above placebo at 60 minutes.

Daily use changes that. In Lovallo's 2005 double-blind trial, 48 men and 48 women spent five days on 0, 300 or 600 mg of caffeine a day, then took 250 mg at 9 a.m., 1 p.m. and 6 p.m. After five days without caffeine, the test doses raised cortisol across the day. After five days of regular intake, the 9 a.m. dose produced no response, but cortisol was raised again between 1 and 7 p.m. after the 1 p.m. dose, then returned to control levels in the evening. The authors' summary: responses are reduced, not eliminated, in daily drinkers.

The practical reading for a daily drinker who wants to change one thing: move or drop the afternoon cup first. Alcohol, cold water, fasting and caffeine tolerance are set side by side in what moves your cortisol.

Does alcohol raise cortisol?

Heavier drinking is linked with higher cortisol and a flatter day. The Whitehall II study in the UK collected six saliva samples across a day, from waking to bedtime, from 2,693 men and 977 women. In men, cortisol was 3% higher for each unit of alcohol drunk per week. In heavy drinkers, cortisol fell less steeply across the day. Heavy-drinking women had a larger morning rise than moderate drinkers: 14.15 against 8.69 nmol/L.

Two limits apply. The findings were not explained by drinking on the day of sampling, which the authors read as a lasting change in heavy drinkers. And it is a single snapshot of a population, so it shows an association, not that cutting back will lower your cortisol.

Does meditation lower cortisol?

A little, on average, and most clearly in people who are ill or under heavy strain. Three meta-analyses agree on direction and disagree on size.

  • Pascoe and colleagues, 2017, pooled 45 randomized trials that compared meditation with an active control rather than with doing nothing. All forms analysed together reduced cortisol, and among the subtypes, focused-attention meditation did.
  • Koncz and colleagues, 2021, split 31 trials by sample. In 10 using blood, meditation had a significant medium effect, but only in patients with a physical illness. In 21 using saliva, the effect was small and not significant, with a marginal effect in people living in stressful situations.
  • Sanada and colleagues, 2016, found only 5 trials in healthy adults, 190 people in all. The pooled effect was moderately low (g = 0.41), and programmes with more sessions and more total hours showed larger effects.

A companion review by Pascoe's group of 42 trials found practices that include yoga postures associated with lower evening and waking cortisol than active controls, though the programmes varied widely. Slow breathing has thinner evidence: in a pilot trial of 40 healthy adults in Beijing, 20 sessions over 8 weeks at about 4 breaths a minute lowered midday saliva cortisol while the control group's did not change. The authors note they had no reference point for whether the drop mattered physiologically.

The fair summary: if meditation or yoga helps you feel calmer, that is reason enough to keep it. As a way to move cortisol in a healthy person, the measured effect is modest. How cortisol differs from the stress you feel is covered in cortisol and stress.

Do friends and time outdoors lower cortisol?

Company blunts the cortisol response to stress, and time in nature is linked with a faster daytime fall. In a 2003 study, 37 healthy men went through the Trier Social Stress Test, a standard laboratory stress test. Some were randomly assigned support from their best friend while they prepared. Support suppressed the saliva cortisol response. The lowest cortisol came with support plus a nasal spray of oxytocin, which was a research tool. The part anyone can use is the friend.

For nature, a University of Michigan study asked 36 city dwellers to spend at least 10 minutes outdoors somewhere that felt like contact with nature, at least three times a week for 8 weeks, at times and places they chose. Saliva taken before and after each visit showed cortisol falling an extra 21.3% an hour beyond the normal daytime decline of 11.7% an hour, with the best return per minute at 20 to 30 minutes. Their model predicts about a 10.6% extra drop after 30 minutes. There was no comparison with an equally long break indoors, and volunteers who enjoy time outside may respond more.

A 2019 meta-analysis of forest bathing reviewed 22 studies and pooled 8. Saliva cortisol was lower in forest groups than in city groups after the visit, but it was also lower before it, and the authors write that anticipated placebo effects can play an important role.

What doesn't lower cortisol, or isn't proven?

Most of what is sold on cortisol has the weakest evidence on this page. In our review of the trials, ashwagandha has the most, with modest drops in some of them; magnesium and L-theanine are weaker and act more on how stressed people feel than on the hormone. We go through the trials one supplement at a time in do cortisol supplements work, including safety and when to talk to a clinician first. The rest of the list:

  • Detoxes and "cortisol resets." There is nothing to flush out. See cortisol detox.
  • Cold plunging. No published study shows it lowers cortisol; noradrenaline is the hormone that moves.
  • Creatine. No consistent effect on cortisol in controlled trials.
  • Protocols for "adrenal fatigue." The condition they claim to treat is not a recognized diagnosis. See is adrenal fatigue real.

Why can't you tell whether a habit works without measuring?

Because the clock moves cortisol more than most habits do. In one study cited in Adam's review, time of day alone accounted for 72% of the variance in saliva cortisol. In the Michigan nature study, the normal daytime fall was 11.7% an hour. Test before a walk at 2 p.m. and after it at 3 p.m., and part of any drop is the rhythm, not the walk.

Single days are noisy too. Hellhammer and colleagues measured the morning rise on six consecutive days and found that a single day's rise was determined to a great extent by that day's circumstances and only a small part by stable traits. Getting a reliable read took two to six days of samples, depending on the measure. And in Wüst's data, about a quarter of people were not classed as responders at all.

So a fair test of a habit means the same times of day, several days before the change and several after, and one change at a time. Saliva kits can do this if you are strict about timing; what they can and cannot tell you is in at-home cortisol tests. None of them is a diagnosis, and none replaces testing ordered by a clinician.

The Auromone Curve is designed to read cortisol from sweat on your wrist through the day, so a week before a change could be set against a week after, against your own baseline rather than a population's. It is in development, and we publish no accuracy figure for it. How a cortisol wearable has to be shown to work, and where the Curve stands, is in how a cortisol wearable is validated. It is designed as a general wellness device: it does not diagnose or treat any condition and does not replace clinical testing.

How do you test one habit at a time?

  • Set a baseline. Sample at the same times for several days before changing anything. Hellhammer's data suggest two to six days for the morning rise.
  • Change one thing. Pick one row of the table and hold everything else steady: same wake time, same training, same coffee.
  • Give it time. The breathing and nature studies each lasted 8 weeks, and meditation programmes with more sessions showed larger effects.
  • Compare at the same hours. Same times, same days of the week. If nothing moved, that is an answer too.

If any sign from the box at the top of this page appears along the way, stop experimenting and see a clinician.

What to take from this

  1. Aim for shape, not a low number. A clear rise after waking and a steady fall to bedtime is the pattern; a flat day is what researchers link with poorer health.
  2. Leave the morning rise alone. It is normal, and morning light makes it larger. Short sleep raised cortisol the next evening, when it should be falling.
  3. Move regularly. Activity programmes lowered cortisol on average; hard sessions raise it briefly, which is normal.
  4. Cut the afternoon coffee before the morning one, drink less alcohol, and spend time with people and outdoors.
  5. Test one change at a time, at the same hours, and see a clinician, not a habit list, for the signs in the safety box.

References

This page is for general education and is not medical advice. The Auromone Curve is designed as a general wellness device, not a diagnostic, and does not lower cortisol, diagnose or treat any condition, or replace clinical testing. It is in development and not yet on sale.

Straight answers

How to lower cortisol FAQ

How do I lower cortisol in the morning?

Usually you shouldn't try. Cortisol normally rises about 50% in the 30 minutes after waking, and morning light makes that rise larger, which is expected. What matters is that it falls through the day. In daily coffee drinkers, a 9 a.m. dose produced no cortisol response after five days of regular intake, while an afternoon dose still did. (Source: Wüst and colleagues, 2000; Scheer and Buijs, 1999; Lovallo and colleagues, 2005.)

How can I lower cortisol naturally?

The best-supported habits are enough sleep, regular exercise, an earlier last coffee, less alcohol, and time with friends and outdoors. One 4-hour night raised next-evening cortisol by 37%, and a pooled analysis of 10 activity programmes found lower cortisol. Meditation's effect is small in healthy adults. This is general information, not medical advice. (Source: Leproult and colleagues, 1997; De Nys and colleagues, 2022.)

Does exercise lower cortisol?

On average, yes, as a regular habit: a meta-analysis of 10 trials found activity programmes lowered cortisol by a small-to-moderate amount. A single hard session raises it: 30 minutes at 60% of maximal oxygen uptake raised blood cortisol about 40%, and 80% raised it about 83%, while easy exercise did not. That short rise is normal. (Source: De Nys and colleagues, 2022; Hill and colleagues, 2008.)

Does caffeine raise cortisol?

Yes. In a placebo-controlled study of 47 young men, about 2 to 3 cups' worth raised cortisol 30% above placebo at 60 minutes. In daily drinkers the response is reduced but not eliminated: after five days of regular intake a morning dose had no effect, while an afternoon dose raised cortisol again into the early evening. (Source: Lovallo and colleagues, 1996 and 2005.)

How can I tell if a habit lowered my cortisol?

Not from one test. Time of day accounted for 72% of the variance in saliva cortisol in one study, and a single day's morning rise mostly reflects that day; two to six days of samples were needed for a reliable read. Sample at the same times for several days before and after one change. The Auromone Curve, in development, is designed to read cortisol through the day for this kind of comparison. (Source: Adam and colleagues, 2017; Hellhammer and colleagues, 2007.)

Your baseline, not a population's.

Every study on this page reports the average of a group, sampled at fixed hours. Whether your own evening fall moves when you change one habit is a different question. The Auromone Curve is designed to read cortisol from sweat on your wrist through the day, so you can compare the weeks before a change with the weeks after. It is a general wellness device in development and isn't on sale yet. Reserving is free.

Reserve