Read this first. Some of these are not harmless experiments.
This page reports what researchers measured. It is not a protocol, and nothing here is a suggestion to try any of it. Four of the five inputs below have put people in hospital, and the mechanisms are worth 60 seconds of your attention before you read a single number.
Cold water can stop your heart, and being fit does not protect you. Sudden immersion triggers the cold shock response: an involuntary gasp, uncontrollable hyperventilation, and a sympathetic tachycardia. If your face is in the water when the gasp happens, you inhale water. Shattock and Tipton (2012) report that arrhythmias occur in about 2% of healthy volunteers during free-breathing head-out immersion, and in 62 to 82% of young, fit participants during submersion with breath-holding, which they attribute to "autonomic conflict" between the cold shock and diving responses. They also report that nearly 30% of victims of otherwise-unexplained drowning carried cardiac ion-channel mutations, that 67% of drownings occur in strong swimmers, and that 55% of those happen within 3 metres of safety.
- Being young, fit, and a strong swimmer does not rule out the arrhythmia. Neither does having done it before without incident. The channelopathies that make people vulnerable are usually silent until the moment they are not. Talk to a doctor first if you have any heart condition, an arrhythmia, high blood pressure, or a family history of sudden cardiac death, unexplained drowning, or fainting.
- Getting out is not the end of the exposure. Core temperature keeps falling after immersion (afterdrop), and rewarming is when peripheral vessels reopen. Feeling fine as you climb out is not evidence that you are fine. Never do this alone, and never in open water.
If you have diabetes of any type, do not fast to run an experiment. If you take insulin, a sulfonylurea (glyburide, glipizide, glimepiride) or a meglitinide, these drugs keep lowering blood glucose whether or not you eat, and hypoglycaemia is the immediate risk (Grajower and Horne, 2019), whose guidance is that people with diabetes should fast only with a clinician's help and a medication plan. And do not rely on feeling it coming: hypoglycaemia unawareness means some people get no warning symptoms at all. Low blood sugar is not the only way fasting hurts you, and it is not the most deceptive one. If you take an SGLT2 inhibitor (empagliflozin, dapagliflozin, canagliflozin, ertugliflozin, often sold as Jardiance, Farxiga or Invokana, and increasingly prescribed for heart failure and kidney disease as well as diabetes), fasting is a recognised trigger for euglycaemic diabetic ketoacidosis, which the FDA has warned about since 2015 and which Grajower and Horne address by advising these drugs be skipped on a fasting day. The danger is that your blood glucose looks normal. A reassuring fingerstick does not rule it out. Nausea, vomiting, abdominal pain, deep or rapid breathing, or fruity-smelling breath while fasting on one of these drugs is an emergency: stop fasting, eat, and seek urgent medical care. If you have a history of an eating disorder, or you are pregnant, fasting content is not neutral information and this page is not the right tool.
After unaccustomed or very hard exercise, dark urine is an emergency, not a rest day. Rhabdomyolysis is muscle breakdown that can cause kidney failure, and dangerous potassium rises that can stop the heart. Its classic signs are severe muscle pain out of proportion to the session, marked weakness, swelling, and tea- or cola-coloured urine, usually 24 to 72 hours later. Go to an emergency department, do not "hydrate and see." And the filter cuts the other way: StatPearls notes the full triad is seen in less than half of patients and dark urine in only 30 to 40%. Clear urine does not rule out rhabdomyolysis.
Separately, chest pain or pressure, unusual breathlessness, or fainting or near-fainting during or just after exercise is a stop-now emergency, not a same-day appointment. Stop exercising immediately and call 911 or your local emergency number. MedlinePlus says to call 911 for sudden crushing, squeezing, tightening or pressure in the chest; fainting during exertion is a recognised red flag for cardiac outflow obstruction and exercise-induced arrhythmia, which are the causes of sudden cardiac death in otherwise healthy, fit people. Do not drive yourself, do not "cool down and see," and do not train again until a clinician has cleared you. Do not file any of it under being out of shape.
With alcohol, the dangerous direction is stopping. If you drink heavily every day, quitting abruptly can cause withdrawal seizures and delirium tremens, which MedlinePlus calls a medical emergency. The two have different clocks, and this matters: withdrawal seizures are most common in the first 12 to 48 hours after the last drink, while delirium tremens most often appears at 48 to 96 hours and sometimes as late as 7 to 10 days. Being inside the first day is not being in the clear. Do not cut alcohol out cold to see what your cortisol does. Get medical advice first. And if you have already stopped and you are shaking, sweating, confused, agitated, seeing or hearing things, or your heart is racing: go to the emergency room or call 911 now. Do not wait to see whether it passes. MedlinePlus describes delirium tremens as especially common in people drinking heavily for months; treat that as a description of who is most at risk, not as a threshold below which you are safe.
And if your own cortisol production is suppressed, none of this is a game. If you take steroid medication by any route, including inhaled, topical, nasal and joint-injected steroids, or you have adrenal insufficiency, then cold, hard exercise, illness and fasting are physical stressors that increase the cortisol your body needs, which is exactly what you may not be able to produce. That is the mechanism of an adrenal crisis. Read what happens to your cortisol when you come off steroids and the signs of low cortisol, and never stop or reduce a steroid on your own.
Finally, if you are here because you feel wrong and you are hunting for the input that explains it, this page will not find it. A list of things that move cortisol in a laboratory is not a diagnosis and cannot become one. Take that to a doctor.
Does caffeine raise cortisol?
Yes, and the best-measured figure is about 30%. Lovallo and colleagues (1996) gave 47 healthy young men an oral dose of 3.3 mg/kg, which the paper describes as the equivalent of 2 to 3 cups of coffee, in a double-blind placebo-controlled crossover. ACTH was significantly elevated from 30 to 180 minutes and cortisol from 60 to 120 minutes. Peak increases relative to placebo, both at 60 minutes:
- Cortisol: +30% (+2.7 µg/dL)
- ACTH: +33% (+5.2 pg/mL), the pituitary signal that tells the adrenal glands to release cortisol
The authors' conclusion is worth quoting because it is narrower than the headline usually is: caffeine "can activate important components of the pituitary-adrenocortical response in humans during the resting state." That is a real finding about a real hormone. It is also a finding about one dose, in men who had abstained from caffeine, sitting still. Almost nobody drinks coffee under those conditions.
Where the "50%" number came from: nowhere. A figure of roughly 50% circulates widely online, often attached to a Lovallo citation. No Lovallo paper reports it. If you see a number next to a reference, the reference is a promise that somebody read the paper, and in this case a lot of people did not.
Why does the same coffee stop working?
Because you build tolerance to the morning dose, and not to the rest of the day. This is the most useful thing in the caffeine literature and it is almost never reported.
Lovallo and colleagues (2005) took 96 healthy adults, had them abstain from dietary caffeine for 5 days while taking capsules totalling 0, 300 or 600 mg per day, then challenged them on day 6 with 250 mg at 9:00 AM, 1:00 PM and 6:00 PM, sampling saliva 8 times from 7:30 AM to 7:00 PM. What they found:
| Prior 5 days | Response to the 9:00 AM dose | Response to the 1:00 PM dose |
|---|---|---|
| Placebo (abstinent) | Cortisol significantly higher at all times of day | Elevated |
| 300 mg/day | Abolished. Not significantly above the placebo week | Returns. Elevated from 1:00 to 2:00 PM onward |
| 600 mg/day | Abolished | Returns |
The paper's conclusion, verbatim: "Cortisol responses to caffeine are reduced, but not eliminated, in healthy young men and women who consume caffeine on a daily basis."
Read that as a habitual coffee drinker and it says something specific. Your first cup is the one you have gone deaf to. Your 4:00 PM cup is not. Add the pharmacology: the mean plasma half-life of caffeine in healthy adults is about 5 hours, with a range from 1.5 to 9.5 hours, and that spread is driven by genetics, pregnancy, oral contraceptives, and smoking. A slow metaboliser's afternoon coffee is still substantially present at midnight, which is one of the reasons the same drink lands so differently on different people, and it is a live issue for how cortisol and sleep interact.
One honest complication. A companion study from the same group (Lovallo, 2006) found that in the same 96 people, "caffeine did not acutely elevate cortisol levels in the absence of stress," although repeated doses across the day did raise cortisol, and caffeine taken before exercise elevated cortisol in both men and women. Two papers, one lab, results that do not line up neatly. That is not a flaw in the research. It is the finding.
How much caffeine is a lot
EFSA's 2015 Scientific Opinion concludes that habitual intakes up to 400 mg per day (about 5.7 mg/kg of body weight) and single doses up to 200 mg (about 3 mg/kg) do not raise safety concerns for healthy adults, and that up to 200 mg per day does not raise safety concerns for the fetus in pregnancy. For children and adolescents, EFSA proposes a safety level of 3 mg/kg of body weight per day, reasoning that caffeine clearance in that group is at least as fast as in adults.
Those are population figures. They are not a personal permission slip: sensitivity, medications, heart conditions and how fast you clear caffeine all move the line, and highly concentrated caffeine powders have killed people who measured them by eye.
What do alcohol, cold water, exercise and fasting do to cortisol?
Every row below is a controlled measurement with a comparison condition, and every one of them is linked in the references. Note how little the popular version of each one survives contact with the numbers.
| Input | What was actually done | What happened to cortisol |
|---|---|---|
| Caffeine, abstinent | 3.3 mg/kg oral, 47 men, at rest, vs placebo | Peak +30% at 60 min (Lovallo, 1996) |
| Caffeine, daily drinker | 250 mg at 9:00 AM after 5 days at 300 or 600 mg/day, 96 adults | Morning response abolished. Afternoon response persists (Lovallo, 2005) |
| Exercise | 30 min cycling at 40%, 60%, 80% VO2max, 12 moderately trained men | -6.6% at rest, +5.7% at 40%, +39.9% at 60%, +83.1% at 80% (Hill, 2008) |
| Cold water, long | 1 h head-out immersion at 14°C, young men | Noradrenaline +530%, dopamine +250%, metabolic rate +350%. Cortisol only "tended to decrease" (Šrámek, 2000) |
| Cold water, short | 5 min ice bath at 8 to 12°C, 12 adults, morning and evening | Noradrenaline +127% (AM) and +144% (PM). Cortisol did not respond significantly to the ice bath at either timepoint, apart from a rise in the morning between 5 and 30 minutes post-immersion (Scientific Reports, 2025) |
| Alcohol | 0.8 g/kg (about 4 drinks), 32 heavy vs 23 light drinkers | Response attenuated in heavy drinkers on the falling limb of blood alcohol (King, 2006) |
| Fasting | 5 days, water only, 8 men, blood sampled every 5 min for 24 h | 24-hour cortisol production 1.8-fold higher. Daily peak delayed into the afternoon (Bergendahl, 1996) |
Cold plunges: the hormone that moves is not the one you were promised
Cold water immersion is sold on cortisol. Cortisol is the one thing in the studies that does not clearly move. Šrámek and colleagues (2000) put young men through 1-hour head-out immersions at 32°C, 20°C and 14°C. At 14°C, noradrenaline rose 530% and dopamine 250%, metabolic rate rose 350%, and cortisol merely "tended to decrease." A 2025 crossover in Scientific Reports put 12 adults into an 8 to 12°C ice bath for 5 minutes, in the morning and again in the evening: noradrenaline rose 127% and 144% respectively, and cortisol did not respond significantly to the ice bath at either timepoint, apart from a rise in the morning between 5 and 30 minutes post-immersion.
The 2025 PLOS One systematic review of cold-water immersion (11 randomised trials, 3,177 participants) did not even meta-analyse cortisol. What it did find on stress was oddly shaped: no significant change immediately after or 1 hour after immersion, a significant reduction at 12 hours, and nothing again at 24 or 48 hours. The same review states that it provides no dedicated safety analysis and that future studies need to examine safety and participant experience. We are carrying that caveat across rather than quietly dropping it, because the safety content in the box above is the part of the cold-water literature nobody puts in a thumbnail.
Exercise: there is a threshold, and easy sessions are below it
Hill and colleagues (2008) had 12 moderately trained men cycle for 30 minutes at 40%, 60% and 80% of VO2max, plus a resting control. Percent change in cortisol from before to after: -6.6% resting, +5.7% at 40%, +39.9% at 60%, +83.1% at 80%. Their conclusion is that a threshold of roughly 60% of VO2max or greater is what elicits a rise, and that low-intensity exercise does not.
Cortisol's job in that moment is a normal one: it mobilises glucose so the working muscle has fuel. What we are not going to do is tell you what your own post-session number means, because a cortisol reading is a measurement and not a verdict, and nobody can interpret it for you from a web page. If you want the honest version of the "high cortisol" story, it is on the signs people associate with high cortisol and on what the evidence says about lowering it.
Fasting: a slow, large, and reshaping effect
A note before the numbers: the study below is a supervised 5-day water-only fast in a clinical research unit, and it is not something to reproduce. Multi-day fasting risks dangerous electrolyte shifts, and refeeding syndrome, in which resuming food drops blood phosphate and can trigger a fatal arrhythmia, strikes at the moment you think you are finished. Bergendahl and colleagues (1996) sampled blood every 5 minutes for 24 hours in 8 men, on a fed control day and on the fifth day of a water-only fast. Fasting produced a 1.8-fold increase in 24-hour cortisol production, driven by a 1.6-fold increase in the mass of cortisol released per secretory burst rather than by more frequent bursts. It also moved the daily peak: maximal cortisol occurred between 09:30 and 13:34 when fed, and between 11:16 and 16:12 when fasted.
That last detail is the one that matters. Fasting did not only raise cortisol. It changed the shape and the timing of the day, which is not something a number can express, and is the reason the morning cortisol rise and the daily slope are the things endocrinologists actually look at.
Alcohol: the response shrinks in the people who drink the most
King and colleagues (2006) gave 0.0, 0.4 and 0.8 g/kg of alcohol (roughly 0, 2 and 4 drinks) to 32 heavy social drinkers and 23 light social drinkers, double-blind, sampling saliva at baseline and 15, 45, 105 and 165 minutes afterwards. At the 0.8 g/kg dose, heavy drinkers showed an attenuated cortisol response compared with light drinkers during the declining phase of blood alcohol (group by dose by time interaction, p < 0.005). The same blunting pattern that caffeine shows in daily consumers shows up here in the people who drink most.
Separately, and importantly: heavy or chronic drinking is a recognized cause of genuinely raised cortisol and cushingoid features, which is covered in the piece on what cortisol does and does not do to your face. Acute dosing in a lab and chronic drinking in a life are different questions, and only one of them belongs in a conversation with a doctor.
One thing the dosing above is not: a licence. The 0.8 g/kg figure describes what a research team gave supervised volunteers, and it is not a target. There is no safe amount of alcohol in pregnancy, and no cortisol question makes one exist.
Why one cortisol test cannot tell you what any of this did
Because every result on this page is a difference between two matched states, and cortisol's own daily swing is bigger than most of the effects being measured.
Look at the ice-bath study again, at a number nobody quotes from it. The same 12 people, before doing anything at all, had resting cortisol of 179 ± 108 pg/mL in the morning and 91 ± 59 pg/mL in the evening (p = 0.013). Their cortisol roughly halved across the day on its own. Now imagine sampling once before your morning coffee and once in the afternoon and attributing the difference to the coffee.
This is exactly why the studies here are built the way they are, and it is worth naming what they all share:
- A placebo and a control day. Lovallo, Hill and King all ran a comparison condition in the same people, because there is no other way to know what the day would have done by itself.
- Samples fixed to the clock. Lovallo sampled 8 times between 7:30 AM and 7:00 PM. Bergendahl sampled every 5 minutes for 24 hours. The peak Lovallo found sits at 60 minutes, which a test taken at 30 or at 180 minutes would have missed or halved.
- Repetition. A single pair of readings, in one person, on one day, is a data point with no error bar and no control. It cannot distinguish "the coffee did that" from "it was 9:00 AM."
None of that is an argument for or against any product. It is a description of how cortisol research is done, and it is the reason that the honest answer to "what does coffee do to me" is that a group average from 47 abstinent men in 1996 does not contain it. That is a limit of the evidence, and no consumer device removes it. If you are weighing up the testing options that exist today, they are laid out in the guide to at-home cortisol tests and in what a "normal" cortisol level actually means.
What this page refuses to give you
There is an obvious ending for an article like this: a protocol. Cut coffee after 2:00 PM, plunge on Tuesdays, fast on Thursdays, watch the number fall. We are not writing that, and not out of caution.
The studies above measured group averages under laboratory control, in specific people, at specific doses, at specific hours. They do not license a rule about you, and any page that converts them into one has quietly replaced evidence with a workout plan. Worse, the two inputs with the most confident internet protocols (cold water and fasting) are the two carrying real risk of death and hospitalisation, and the confidence is exactly what makes them dangerous. What we will say is what the research says, and then stop.
This guide is for general education only. The Auromone Curve is a general wellness device, not a diagnostic, and does not replace medical advice or clinical testing. Nothing on this page is a recommendation to fast, drink, enter cold water, or train at any intensity. Talk to a healthcare provider before changing anything, especially if you take prescribed medication.
References
- Lovallo WR, Al'Absi M, Blick K, Whitsett TL, Wilson MF. Stress-like adrenocorticotropin responses to caffeine in young healthy men. Pharmacol Biochem Behav, 1996. (n = 47; 3.3 mg/kg, "equivalent to 2-3 cups of coffee"; peak vs placebo at 60 min: ACTH +33% / +5.2 pg/mL, cortisol +30% / +2.7 µg/dL.)
- Lovallo WR, Whitsett TL, al'Absi M, Sung BH, Vincent AS, Wilson MF. Caffeine stimulation of cortisol secretion across the waking hours in relation to caffeine intake levels. Psychosom Med, 2005. (n = 96; the 9:00 AM response is abolished by 300 or 600 mg/day for 5 days, the 1:00 PM response is not.)
- Lovallo WR, Farag NH, Vincent AS, Thomas TL, Wilson MF. Cortisol responses to mental stress, exercise, and meals following caffeine intake in men and women. Pharmacol Biochem Behav, 2006. ("Caffeine did not acutely elevate cortisol levels in the absence of stress"; caffeine before exercise did.)
- Hill EE, Zack E, Battaglini C, Viru M, Viru A, Hackney AC. Exercise and circulating cortisol levels: the intensity threshold effect. J Endocrinol Invest, 2008. (n = 12; -6.6% rest, +5.7% at 40%, +39.9% at 60%, +83.1% at 80% VO2max.)
- Šrámek P, Šimečková M, Janský L, Šavlíková J, Vybíral S. Human physiological responses to immersion into water of different temperatures. Eur J Appl Physiol, 2000. (1 h at 14°C: noradrenaline +530%, dopamine +250%, metabolic rate +350%; cortisol "tended to decrease".)
- Effects of time-of-day on the noradrenaline, adrenaline, cortisol and blood lipidome response to an ice bath. Scientific Reports, 2025. (n = 12; 8 to 12°C for 5 min; noradrenaline +127% AM and +144% PM; cortisol did not respond significantly to the ice bath at either timepoint, apart from a rise in the morning between 5 and 30 minutes post-immersion; resting cortisol 179 ± 108 pg/mL AM vs 91 ± 59 pg/mL PM, p = 0.013.)
- Effects of cold-water immersion on health and wellbeing: a systematic review and meta-analysis. PLOS One, 2025. (11 RCTs, 3,177 participants; stress reduced at 12 h post-immersion only; no meta-analysis of cortisol; the authors state safety was not analysed and needs study.)
- Shattock MJ, Tipton MJ. "Autonomic conflict": a different way to die during cold water immersion? J Physiol, 2012. (Source for the cold shock response, the 2% vs 62-82% arrhythmia figures, ion-channel mutations in unexplained drowning, and 67% of drownings occurring in strong swimmers.)
- King A, Munisamy G, de Wit H, Lin S. Attenuated cortisol response to alcohol in heavy social drinkers. Int J Psychophysiol, 2006. (0.4 and 0.8 g/kg; 32 heavy vs 23 light drinkers; group x dose x time interaction p < 0.005.)
- Bergendahl M, Vance ML, Iranmanesh A, Thorner MO, Veldhuis JD. Fasting as a metabolic stress paradigm selectively amplifies cortisol secretory burst mass and delays the time of maximal nyctohemeral cortisol concentrations in healthy men. J Clin Endocrinol Metab, 1996. (n = 8; 5-day fast; 1.8-fold rise in 24 h production, 1.6-fold rise in burst mass, peak delayed from 09:30-13:34 to 11:16-16:12.)
- Grajower MM, Horne BD. Clinical management of intermittent fasting in patients with diabetes mellitus. Nutrients, 2019. (Insulin and sulfonylureas: high hypoglycaemia risk; meglitinides moderate; SGLT2 inhibitors to be skipped on a fast day; fasting should be done with a clinician's guidance; hypoglycaemia unawareness raises the stakes. The paper also advises against fasting in pregnancy and lactation, young children, frail older adults, immunodeficiency, eating disorders and dementia.)
- StatPearls (NCBI Bookshelf). Rhabdomyolysis. (The classic triad "seen in less than one-half of patients"; dark urine in 30-40%.)
- MedlinePlus, U.S. National Library of Medicine. Delirium tremens. ("Delirium tremens is a medical emergency." Withdrawal seizures most common 12 to 48 hours after the last drink; delirium tremens most often 48 to 96 hours, sometimes 7 to 10 days. "Go to the emergency room or call 911 or the local emergency number if you have symptoms.")
- MedlinePlus, U.S. National Library of Medicine. Chest pain. (Call 911 for sudden crushing, squeezing, tightening or pressure in the chest.)
- European Food Safety Authority. Caffeine (Scientific Opinion, 2015). (400 mg/day habitual and 200 mg single doses do not raise safety concerns for healthy adults; 200 mg/day in pregnancy; 3 mg/kg bw/day proposed for children and adolescents.)
- Institute of Medicine, Pharmacology of Caffeine (NCBI Bookshelf). ("The mean half-life of caffeine in plasma of healthy individuals is about 5 hours. However, caffeine's elimination half-life may range between 1.5 and 9.5 hours.")
