Normal cortisol levels by time of day and age

There is no single normal cortisol level. The number depends on the time of day, the type of sample (blood, saliva, urine or hair), the lab’s method and the units on the report. A value that is ordinary at 8 a.m. can be high at midnight.

This page gives the ranges that published sources print, names each source and says where no source gives a figure. It does not tell you what your own result means. For that, use the range on your lab report and take the result to a clinician.

Cortisol AM and PM

A morning result and a later one, because cortisol changes through the day. Follow the times on your lab order.

$129

See this test at HealthLabs

Price checked November 2026. Not available in ND, RI, NY, NJ, MD and HI. We earn a commission on sales.

Why there is no one normal level

A reference range is built from the results of a group of healthy people. MedlinePlus says labs often use different methods, so they often use different ranges, and you can’t compare results between labs. It also says a test can have separate ranges for different groups, and that healthy people sometimes fall outside the range (MedlinePlus).

Cortisol shows this clearly. Each lab sets its own performance for its cortisol immunoassay, and the review cites a study in which five automated immunoassays showed a bias of 1.08 to 1.36 against a mass spectrometry method (Endocrine Reviews, 2024). The same review says there is a lack of assay-specific normal data for specific groups of people. So treat any chart, including this one, as a rough guide.

Units matter too. US labs usually report blood cortisol in µg/dL, and many other labs use nmol/L. Multiply µg/dL by about 27.6 to get nmol/L, or use the cortisol units converter.

Normal cortisol levels by time of day (blood)

Cortisol starts to rise in the early hours, peaks in the morning and falls through the day. A 2023 laboratory review says the lowest point is shortly after going to sleep, around midnight (Diagnostics, 2023).

Published blood ranges differ by source:

Source	Morning	Afternoon
Review in Int J Mol Sci (2025), for total serum cortisol	3 to 23 µg/dL at 8 a.m.	3 to 13 µg/dL at 4 p.m.

That is one source. Labs and methods differ, so treat it as a rough guide and use the range printed on your report.

The 2025 review took its blood, saliva and adult urine (under 100) figures from a laboratory reference textbook and its 36 to 137 urine figure from a 2018 journal article, so the two urine figures differ. They are published figures, not ranges from one lab’s method.

Midnight and other times. None of the sources used for this page prints a serum range for midnight or for other clock times, so none is shown. For the late evening, clinicians use saliva instead (below).

Saliva, urine and hair ranges

Other samples measure cortisol differently, and their numbers are not comparable with blood.

Sample	Time or period	Published range	Source
Saliva	7 to 9 a.m.	100 to 750 ng/dL	Int J Mol Sci, 2025
Saliva	3 to 5 p.m.	Less than 401 ng/dL	Int J Mol Sci, 2025
Saliva	11 p.m. to midnight	Less than 100 ng/dL	Int J Mol Sci, 2025
Urine	Whole 24 hours (free cortisol)	36 to 137 µg per 24 hours	Int J Mol Sci, 2025 (text)
Urine	Whole 24 hours, adults	Less than 100 µg per 24 hours	Int J Mol Sci, 2025 (table)
Hair	3 cm of hair nearest the scalp	0.84 to 8.07 pg/mg	J Endocrinol Invest, 2026
Hair	Not specified in the review	40 to 128 pg/mL	Int J Mol Sci, 2025 (different unit and method)

How to read this table:

• Saliva holds only the free form of cortisol, so its numbers are far smaller than blood, and ng/dL is a thousand times smaller than µg/dL (Int J Mol Sci, 2025). A review of cortisol assays says saliva work is limited by the lack of a single validated reference range and poorly standardized assays (Ann Clin Biochem, 2017, via PubMed).
• Urine cut-offs depend on the method. Immunoassays can over-measure urinary cortisol, so equivalent cut-offs need to be method specific (Diagnostics, 2023).
• Hair is a research measure. The 0.84 to 8.07 pg/mg figure comes from 256 healthy adults in Northern Italy tested by mass spectrometry (J Endocrinol Invest, 2026). The authors call it the first population-based reference interval of its kind and a benchmark for research. The 2025 review quotes a different figure in a different unit, so there is no accepted standard.

See saliva, urine and hair tests for what each is used for.

Normal cortisol levels by age

Age changes the daily pattern. A 2024 review says that with age the daily swing gets smaller, the early-morning rise comes earlier and the lowest point is higher. It adds that age-specific cut-offs are typically not used in clinical practice (Endocrine Reviews, 2024). A 2025 review describes older adults as having a flatter daily slope and higher evening cortisol (Int J Mol Sci, 2025).

For children, no approved source used here prints age-banded ranges, so none is shown. The daily rhythm is likely present in all infants from about 9 months old (Diagnostics, 2023). Children’s labs publish their own ranges, so use the one on the report.

Differences between women and men are on a separate page: normal cortisol for women and men.

What moves a level without a disease

MedlinePlus says an abnormal result doesn’t always mean a condition that needs treatment. The most common cause of abnormal cortisol is taking steroid medicines for a long time or suddenly stopping them. Stress, pregnancy, exercise, hot and cold temperatures and some medicines, such as birth control pills, can also change a result. The time of the sample matters, and so does the carrier protein that binds most cortisol, which changes total cortisol without changing the active free part (Endocrine Reviews, 2024). Our page on free and total cortisol goes further.

How to use a range

1. Match the time. Compare a morning value with a morning range, not a later one.
2. Match the units. Check µg/dL against nmol/L.
3. Use your own report’s range. It matches the method your lab used.
4. Use the same lab to follow a trend. MedlinePlus advises this, since results from different labs can’t be compared.
5. Don’t read a range as a diagnosis. A value outside a range may or may not be a problem, and a value inside one is not proof of good health.

If a result falls outside the range, that is a reason to talk to a clinician. They may repeat the test or add others, such as ACTH. See testing for high cortisol and testing for low cortisol. If you take a steroid medicine, don’t stop or change it. Ask your prescriber.

To get a blood pair at the right times, see AM and PM timing.