Key takeaways
- Cortisol follows a strong diurnal rhythm, is released in pulses, and circulates mostly bound to carrier protein, so a single value is heavily condition-dependent.
- Morning serum drawn between 8 and 10 AM measures total cortisol at the peak against well-standardised ranges, and it is the right first test for excluding adrenal insufficiency or Cushing's.
- Three features of cortisol physiology make single-point measurement awkward.
- A single blood draw between 8 and 10am, sampled at the peak of the curve where the signal is largest and most reproducible.
- The waking and +30 pair capture the cortisol awakening response, the sharp rise that follows waking and is thought to reflect the anticipatory readiness of the axis rather than stress per se (Pruessner et al., Life Sci 1997).
Cortisol is the hardest common hormone to measure well, because it is the one that moves the most. It follows a daily curve, it pulses within that curve, it spikes in response to the needle taking the sample, and it is mostly bound to a carrier protein that changes independently of the hormone itself. A single number tells you where cortisol was at one moment under one set of conditions. Whether that is enough depends entirely on what question you are asking, and the two tests below answer genuinely different questions.
Why one cortisol value is a difficult thing to interpret
Three features of cortisol physiology make single-point measurement awkward.
It has a strong diurnal rhythm. Output peaks in the first half hour after waking and declines to a trough in the late evening. A value that is normal at 8am would be abnormal at 10pm. Time of day is not a detail on the requisition; it is half the result. Hormone diurnal rhythms covers how many other markers behave the same way.
It is pulsatile. Underneath the daily curve, cortisol is released in discrete bursts. Two draws twenty minutes apart in the same person on the same morning can differ substantially, and neither is wrong.
Most of it is bound. The large majority of circulating cortisol is attached to cortisol-binding globulin and albumin and is not biologically available. Serum assays measure the total; salivary assays measure the free fraction, because only unbound cortisol diffuses into saliva. That is why the two are not interchangeable units of the same thing — one measures the pool, the other the part acting on tissue. It also means anything that changes binding protein levels, notably oestrogen, moves serum cortisol without changing what any cell experiences.
What morning serum cortisol is for
A single blood draw between 8 and 10am, sampled at the peak of the curve where the signal is largest and most reproducible. Typical reference range around 5-25 µg/dL, method-dependent.
This is the correct first test when the question is whether the axis is capable of producing cortisol at all, or is producing far too much. Specifically:
- Adrenal insufficiency screening. A very low morning value in the right clinical context prompts confirmatory dynamic testing (Bornstein et al., J Clin Endocrinol Metab 2016).
- Cushing's screening, usually paired with an overnight dexamethasone suppression test rather than interpreted alone (Nieman et al., J Clin Endocrinol Metab 2008).
- Pituitary disease evaluation, where cortisol is one output among several being assessed.
- General inclusion on a hormone panel, as a coarse check that nothing dramatic is happening.
What it cannot do is describe the shape of the day. A person whose cortisol peaks appropriately in the morning and then stays high all evening will have a normal 8am serum value and a genuinely abnormal rhythm.
What the salivary curve is for
Four saliva samples: on waking, thirty minutes after waking, midday, and before bed. The waking and +30 pair capture the cortisol awakening response, the sharp rise that follows waking and is thought to reflect the anticipatory readiness of the axis rather than stress per se (Pruessner et al., Life Sci 1997). The midday and evening samples describe the slope of the decline.
The reason to run four samples rather than one is that the clinically interesting information in chronic stress is usually in the shape, not the level. The diurnal slope — how steeply cortisol falls from morning to night — associates with a range of mental and physical health outcomes across a large body of work, and a flatter slope is the consistent finding (Adam et al., Psychoneuroendocrinology 2017). Two people can produce identical total daily cortisol with completely different slopes and completely different symptoms.
The two tests side by side
| Test | Captures | Misses |
|---|---|---|
| AM serum | Peak level, total cortisol, well-standardised reference ranges | Diurnal pattern; free fraction; effects of binding protein changes |
| Salivary curve | Full rhythm, free cortisol, awakening response | Firm quantitative reference points; consistency across labs |
| 24h urinary free cortisol | Total daily output, integrated across the day | Pattern entirely; requires accurate collection |
| Late-night salivary cortisol | Loss of the evening trough, the earliest Cushing's signal | Everything else about the day |
Reading the patterns
- Normal curve — high on waking, higher again at +30 minutes, declining steadily through the day, low before bed.
- Flattened curve — the morning peak and evening trough converge. The commonest chronic-stress and burnout pattern, and the one most consistently linked to poor outcomes.
- Blunted awakening response — little or no rise in the first thirty minutes. Often accompanies the subjective experience of waking already exhausted.
- Elevated evening cortisol — the axis has not switched off. This is frequently the finding underneath difficulty falling asleep and the 3am waking pattern described in why you wake up at 3am.
- High across the board — earlier-stage chronic activation, and occasionally something that requires proper endocrine investigation rather than lifestyle advice.
- Low across the board — long-standing dysregulation, but this pattern must never be assumed before adrenal insufficiency has been properly excluded.
Where the curve is weaker than it is sold
This is the part most sources skip. The salivary curve is a useful descriptive instrument and a poor diagnostic one, for concrete reasons.
Reference ranges vary substantially between laboratories and assay methods, so a value flagged as abnormal by one provider may sit inside another's range. Day-to-day variability within the same person is high, meaning a single day's curve is a sample of one from a noisy distribution. And the awakening response is exquisitely sensitive to collection technique — a sample taken twenty minutes after waking rather than on waking, or after coffee, or after checking work email, produces a different curve for reasons that have nothing to do with the patient's physiology.
It also does not diagnose anything with a name. There is no cortisol pattern that establishes burnout, and "adrenal fatigue" is not a diagnosis the endocrine literature supports as a discrete entity — the adrenal glands are not exhausted. What the curve does is describe a regulatory state and give something concrete to re-measure. Burnout as a hormonal disorder covers what the rest of the picture usually looks like, and the HPA axis in depth covers the regulation itself.
The clinical pearl: morning serum cortisol answers "is the axis working?" and the salivary curve answers "is the axis regulated?" Those are different questions. Running the second without having answered the first is the most common sequencing error in this area, because a flat curve with a genuinely low output is a different problem with a different urgency.
Choosing between them
Morning serum is sufficient, and usually correct, for initial endocrine evaluation, for suspected Cushing's or adrenal insufficiency, and as a general inclusion on a broad hormone panel. It is cheap, standardised and interpretable.
The salivary curve adds real information when the presenting problem is chronic stress with a rhythm component: persistent fatigue with intact sleep duration, wired-and-tired evenings, waking unrefreshed, or a suspected burnout picture where the rest of the panel is drifting. It is also the more useful of the two for tracking response, because a flattened slope steepening over months is a visible, specific piece of feedback that a single morning value cannot provide.
Getting a usable sample either way
Both tests are easy to invalidate. For serum, the draw should fall inside the 8-10am window, ideally without a stressful commute immediately preceding it. For saliva, the waking sample means on waking — before getting up, before coffee, before the phone. Nothing to eat or drink beforehand, no brushing teeth immediately before a sample, no strenuous exercise on collection day, and a note of any medication affecting the axis, corticosteroids in any form included. Caffeine timing matters more here than people expect, as caffeine and cortisol sets out. A curve collected carelessly is worse than no curve, because it will still be interpreted.
Bottom line
Morning serum cortisol captures the peak of a total pool under well-standardised conditions, and it is the right first test for excluding real adrenal disease in either direction. The salivary curve captures the free fraction across the day and describes the rhythm, which is where the useful information sits in chronic stress and burnout. Neither diagnoses anything on its own, and the curve in particular is more variable and less standardised than its presentation usually admits. The sequencing that works is serum first to establish that the axis functions, curve second if the question is regulation, and both interpreted against symptoms by a physician rather than read off a reference range.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
