Key takeaways
- Cortisol is a curve, and chronic stress usually flattens it rather than raising the average.
- A single morning serum draw rules out severe adrenal disease and cannot detect a disordered slope.
- Salivary cortisol measures the free, bioactive fraction, which removes the binding-globulin confound that affects serum.
- The four-point curve is the right test for chronic stress, burnout, unexplained fatigue and sleep-maintenance problems.
- Corticosteroids, oestrogen-containing contraception, shift work and acute illness all make a cortisol result uninterpretable.
"My cortisol was normal" is one of the most misleading sentences in medicine, and it is almost always said in good faith. What it usually means is that one blood draw, taken at the moment cortisol is naturally highest, came back inside a wide reference range. That test is good at what it was designed for — finding severe disease — and close to useless for what most adults are asking, which is whether years of stress have distorted the daily rhythm. Order the wrong test and you get a reassuring result that explains nothing.
Cortisol is a curve, not a number
Cortisol is secreted in pulses under a circadian programme: it rises in the last hours of sleep, peaks about 30 minutes after waking — the cortisol awakening response — declines across the day, bottoms out around midnight, then climbs again (Pruessner et al., Life Sci 1997). That shape is the signal — what tells the liver when to release glucose, the immune system when to damp down, and the brain when to be alert.
Chronic stress does not usually raise the whole curve. It flattens it: the morning rise blunts, the evening fails to fall, and the daily total can end up close to normal while the pattern is clearly disordered. Flatter diurnal slopes have been associated with worse health outcomes across a wide range of conditions (Adam et al., Psychoneuroendocrinology 2017). A single measurement is structurally incapable of detecting a slope — which is the mechanism behind the most common frustration here: someone wired at night and flattened in the morning gets a normal 8 AM cortisol and is told nothing is wrong. The test could not have found what they have.
Morning serum cortisol
What it measures: total cortisol in blood — both the fraction bound to cortisol-binding globulin and the small free fraction that is actually active — from a single draw, ideally between 7 and 9 AM.
Best for: ruling out severe adrenal insufficiency (Addison's disease) and, in combination with other tests, gross overproduction. If the morning value is clearly low, that is a finding worth acting on immediately.
Limitations: one point, no slope. It measures total rather than free cortisol, so anything that changes binding globulin — oestrogen-containing contraception, pregnancy — shifts the number without changing the amount of hormone doing work. And "morning" means relative to your waking, not the clock: a shift worker drawn at 8 AM is sampled at an arbitrary point on their own curve.
Typical target when drawn 7-9 AM: 10-18 µg/dL.
The four-point salivary curve
What it measures: free cortisol in saliva at four points — on waking, about 30 minutes after waking, around noon, and at bedtime. Some panels add a late-afternoon sample.
Why saliva: it reflects the free, bioactive fraction rather than the bound total, removing the binding-globulin confound. It needs no needle, so samples come at the times that matter rather than when a phlebotomy room is open — including the waking sample, which no blood test will ever capture.
Best for: chronic stress, burnout, unexplained fatigue with a normal morning serum, sleep-maintenance problems, and any question about HPA axis rhythm rather than magnitude.
Patterns to read:
- Healthy: clear rise from waking to +30, steady decline through the day, low at bedtime
- Chronic stress or burnout: flattened slope, evening value failing to drop
- Blunted awakening response: little or no rise at +30, often with exhaustion on waking
- Elevated throughout, no nadir: uncommon, and a reason to investigate pathological overproduction rather than stress
Where it goes wrong: almost entirely in collection. The +30 sample must be 30 minutes after waking, not after getting out of bed an hour later; eating, brushing teeth, caffeine and smoking before a sample all distort it; an unusual sleep or work day makes the whole set uninterpretable. Protocol adherence is the single biggest determinant of whether the result means anything (cortisol awakening response sampling and adherence). Collect on an ordinary day.
Hair cortisol
What it measures: cortisol incorporated into the hair shaft as it grows — a retrospective average over roughly the preceding 90 days (Stalder & Kirschbaum, Brain Behav Immun 2012).
Best for: chronic exposure over months, and for looking back at a period you can no longer sample. Immune to the day-of-collection problems that afflict saliva.
Limitations: no curve, no recent changes; affected by hair treatment, washing and growth rate; limited availability. It answers "how much, over the last season" and nothing else.
24-hour urinary free cortisol
What it measures: total free cortisol excreted across a full day — an integrated measure of production rather than a snapshot.
Best for: suspected Cushing's syndrome, alongside late-night salivary cortisol and a dexamethasone suppression test (Nieman et al., JCEM 2008). Late-night salivary cortisol has become a workhorse for this question, because loss of the midnight nadir is one of the earliest signs (late-night salivary cortisol accuracy).
Limitations: difficult to do correctly — one missed void invalidates the collection — and it gives total output with no information about shape.
Which test, when
| The question | The right test |
|---|---|
| Ruling out severe adrenal insufficiency | AM serum cortisol plus ACTH |
| Chronic stress or burnout pattern | 4-point salivary curve |
| Waking exhausted despite adequate sleep | 4-point salivary, focused on the awakening response |
| Wired at night, cannot switch off | 4-point salivary, focused on the evening value |
| Cumulative exposure over months | Hair cortisol |
| Suspected Cushing's syndrome | 24-hour urinary free cortisol, late-night salivary, dexamethasone suppression |
| Routine screening alongside a hormone panel | AM serum cortisol |
What to sort out before ordering anything
Cortisol testing is easy to waste. Several things move the result enough to make interpretation meaningless, and all are worth clearing first.
- Corticosteroids of any kind — oral, inhaled, topical, or a recent joint injection — suppress endogenous cortisol and invalidate the test
- Oestrogen-containing contraception raises binding globulin and therefore total serum cortisol, without raising the free fraction
- Shift work or a disrupted schedule shifts the whole curve; sample relative to your own waking, on a typical day
- Acute illness, surgery or an unusually stressful week raise cortisol appropriately — the system working, not failing
- Caffeine timing, which affects the awakening response — see caffeine and cortisol
The clinical pearl: "my cortisol is normal" nearly always means one morning serum draw was inside a wide range. That result rules out severe disease. It says nothing about a flattened diurnal slope, which is the pattern most adults asking about cortisol actually have. Match the test to the question before drawing anything.
What to do with the result
A cortisol pattern is a description, not a diagnosis, and no medication fixes a flattened slope. What it does is redirect attention, which is worth a great deal.
A blunted morning rise with an elevated evening value points at sleep timing, light exposure, evening stimulation and chronic load — and at whether something else is driving the stress response, such as untreated sleep apnoea, under-eating or overtraining. High evening cortisol also explains a very common symptom: waking at 3 AM feeling alert, which the 3 AM wake-up covers.
Read it alongside the hormones it interacts with. Sustained cortisol signalling suppresses the gonadal axis and shifts thyroid conversion, so a disordered curve rarely sits alone — see the HPA axis in depth, reverse T3 in chronic stress, and sleep, cortisol and recovery.
Bottom line
Morning serum cortisol is a competent test for severe adrenal disease and a poor test for most of what people order it for. The four-point salivary curve measures the free, active hormone at the times that carry the information, and it is the right choice for chronic stress, burnout, unexplained fatigue and sleep-maintenance problems — provided the samples are collected properly on an ordinary day. Hair cortisol answers a cumulative question; 24-hour urine belongs to the Cushing's workup. Decide what you are asking before you decide what to draw, and clear the confounders first, because the most expensive cortisol test is the one that had to be repeated. The 60-second assessment is a reasonable start if you want a physician choosing.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
