Is hair cortisol testing accurate? What the validation evidence shows

Yes, for what it is designed to measure. A hair test is the most accurate way to measure your long-term cortisol from a single sample, and far more consistent from one test to the next than blood, saliva or urine. In a study where adults collected saliva three times a day for 30 days, the hair nearest their scalp tracked the 30-day average at r = 0.61, and the same people's hair results correlated at r = 0.84 from one month to the next.¹ (r is a correlation: 1 means two measures move in perfect step, 0 means no relationship.) For diagnosing Cushing's syndrome, research puts its accuracy similar to the standard first-line tests, not better.² But "accurate" needs unpacking, because there is no perfect yardstick for cortisol over months. This guide sets out the evidence and its limits.

What "accurate" means for a long-term measure

A test is normally judged against a reference standard, a more definitive measurement of the same thing. For the amount of cortisol your body produced over three months, no such standard exists. Nobody can sample blood continuously for months, and the closest substitutes, many saliva samples or repeated 24-hour urine collections, are noisy in their own right.¹³

So researchers test hair in four ways, and this guide takes them in turn:

• Agreement. Does hair track repeated saliva and urine measurements over the same period?
• Consistency. Does the same person get a similar result when tested again?
• Behaviour. Does hair cortisol rise where cortisol is known to rise, for example in late pregnancy and long-running stress?
• Clinical accuracy. Does it separate people with Cushing's syndrome, a rare disease of excess cortisol, from people without it?

Laboratory precision, how closely a method repeats its own measurements, is a fifth and more technical question. It is covered in how hair cortisol testing works.

Does hair agree with repeated saliva and urine tests?

Moderately, and most closely with saliva averaged over weeks. These are the main studies that compared hair with repeated measurements over the same period:

Three things stand out. First, hair agrees with the right thing. It matched total daytime cortisol output, but not the morning awakening response or the slope of the daily curve, in both the 30-day study and the 2023 study.¹⁶ That fits hair as a measure of how much cortisol you produce overall, rather than of its daily pattern.

Second, agreement with urine is weaker, and in two small studies it was not statistically significant.¹³ Third, part of the shortfall lies with the yardsticks. In the 2011 study, saliva correlated with itself at only r = 0.514 from one week to the next, and not significantly over two weeks (r = 0.187).⁴ A measure cannot agree closely with a comparison that does not agree with itself.

Hair also moves with saliva over longer spans. In 21 pregnant women, hair cortisol and daily salivary output correlated throughout pregnancy and rose together as it progressed.⁷ And a 2017 meta-analysis of 66 studies found positive associations between hair cortisol and short-term salivary measures.⁸

How consistent is hair from one test to the next?

This is where hair is strongest. In the 30-day study, the same people's hair cortisol correlated at r = 0.84 from one month to the next, against 0.38 to 0.61 week to week for averaged saliva and 0.59 to 0.68 for 24-hour urine.¹ In two further studies, one sampling 45 people a year apart and another sampling 64 people three times at two-month intervals, repeated hair measurements correlated at r = 0.68 to 0.79, and a stable personal component explained 59 to 82 per cent of the variation.⁹ With one LC-MS/MS method, the variation within the same person was below 15 per cent.¹⁰

Compare a single saliva sample. In a classic study of saliva taken at 1 pm on occasions six weeks apart, about 75 per cent of the variation reflected the situation on the day and only about 21 per cent the person's stable level.¹¹ That is the core of the case for hair: a single blood, saliva or urine result mostly tells you about the day it was taken, while a single hair result mostly tells you about you.

Does hair cortisol behave the way cortisol should?

A valid long-term measure should rise where cortisol is known to rise, and hair does. In the first study of hair segments, new mothers had twice the hair cortisol of other women in the segment that grew during the third trimester, when cortisol naturally climbs.¹² In people taking hydrocortisone, which is chemically identical to cortisol, hair cortisol was higher and rose with the dose.¹³ And in the 2017 meta-analysis, covering 10,289 people, groups exposed to chronic stress had 22 per cent higher hair cortisol overall, rising to 43 per cent when the stress was still going on.⁸

The same meta-analysis found no consistent link with how stressed people said they felt, or with mood disorders.⁸ Self-reports and hormones measure different things, so that is not in itself a failure of the test. But it does mean hair cortisol cannot diagnose depression, anxiety or burnout in an individual.

How accurate is hair for Cushing's syndrome?

Cushing's syndrome is the main setting where hair has been tested against a confirmed diagnosis. Three studies give the picture:

• In 14 confirmed patients and 96 controls, an ELISA hair test had 86 per cent sensitivity (the share of patients it picked up) and 98 per cent specificity (the share of people without the condition it correctly cleared).¹⁴
• In 43 patients, 35 people in whom Cushing's had been suspected and ruled out, and 174 healthy controls, a cut-off of 31.1 pg/mg gave 93 per cent sensitivity and 90 per cent specificity.² The authors concluded that a single hair sample offers diagnostic accuracy similar to current first-line tests.²
• Using LC-MS/MS in 89 untreated patients and 295 community controls, hair cortisol had 81 per cent sensitivity and 88 per cent specificity.¹⁵ Hair cortisone, a closely related hormone, did slightly better, at 87 and 90 per cent.¹⁵

For comparison, published studies of late-night salivary cortisol, one of the standard first tests, report 92 to 100 per cent sensitivity and 93 to 100 per cent specificity.¹⁶ So "similar to" is the fair description, not "better than". Hair also has a clear blind spot. In 67 people with adrenal incidentalomas, small adrenal growths found by chance on a scan, it could not detect mild autonomous cortisol secretion, and the authors judged it unsuitable for that purpose.¹⁷

Hair is not part of the diagnostic guidelines for Cushing's syndrome, which rely on late-night saliva, 24-hour urine and dexamethasone suppression tests.¹⁶¹⁸ A hair result cannot diagnose or rule out the condition. See testing for Cushing's syndrome.

Where the evidence runs out

• No single normal range. Healthy values differ more than tenfold between methods: 0.84 to 8.07 pg/mg in an LC-MS/MS reference study, 17.7 to 153.2 pg/mg in an early immunoassay study.¹⁹⁵ A result can only be read against its own laboratory's range.
• Things other than cortisol can shift the result. Hair dye was linked with 12.5 per cent lower values in the Whitehall II study, and 40 hours of sunlight cut hair cortisol by a median of 54 per cent in an experiment.²⁰²¹ See testing dyed or treated hair.
• It misses the most recent weeks. The centimetre nearest the scalp reflects hair formed about 1.3 to 2.2 months before the cut.²²
• It is still a research and specialist tool. A 2024 expert review in Endocrine Reviews said hair cortisol still needs validating outside research settings.²³

The verdict

Taken together, the evidence supports a clear but bounded claim. If your question is what your cortisol has been doing over the past few months, a hair test is the most accurate measure of long-term cortisol you can get from a single sample. It reflects the average of weeks of saliva sampling, it gives much the same answer when repeated, and it is not affected by the time of day it is taken or by a bout of sweating.¹⁹²⁴ Agreement with other tests is moderate rather than close, partly because those tests are so variable themselves.

Blood, saliva and urine answer different questions: what your cortisol is doing at a particular moment, and, in a doctor's hands, whether dynamic tests show disease. The comparison guides for hair vs blood, hair vs saliva and hair vs urine go into each.