Why hormone tests mislead: timing, binding proteins and assay interference
- Principle
A hormone result can be wrong even when the laboratory made no mistake. Knowing the usual traps spares patients needless scans and treatment.
What it is
Hormone results depend on biology (pulsatile and daily rhythms, binding proteins, acute illness, pregnancy, age), on medicines, and on the assay itself. Most hormones are measured by immunoassays, which can be disturbed by high-dose biotin supplements, by heterophile antibodies in the patient's blood, by large inactive hormone complexes such as macroprolactin, and by the «hook effect» when a hormone is so high that it saturates the test. When a result does not fit the patient, suspect the test before the patient.
What to know
- When: cortisol is highest in the early morning and lowest near midnight; testosterone is highest in the morning; sex hormones vary with the menstrual cycle.
- Bound to what: most thyroxine, cortisol and testosterone travel on binding proteins. Pregnancy and oestrogens raise them (total levels go up); obesity and insulin resistance lower SHBG (total testosterone goes down) without changing the free hormone.
- Acute illness lowers T3 and sometimes TSH (non-thyroidal illness); thyroid tests in an acutely ill patient are best avoided unless thyroid disease is suspected.
- Biotin supplements, sold for hair and nails, interfere with streptavidin–biotin assays and can produce a pattern that mimics Graves disease.
- Heterophile antibodies give falsely high or low results that do not fit the patient and do not dilute in a straight line.
- Macroprolactin, a large prolactin–antibody complex, raises measured prolactin without symptoms; it is detected by polyethylene glycol precipitation.
- The hook effect: in a giant prolactinoma the true value is so high that the assay reports only a modest rise unless the sample is diluted.
- Medicines that change endocrine tests include glucocorticoids, opioids, dopamine blockers, amiodarone, lithium and heparin.
Step by step
- Ask whether the result fits the patient; if it does, act on it, and if it does not, pause.
- Check the timing: time of day, cycle day, fasting state, days since an acute illness.
- Check the binding proteins: pregnancy, oestrogens, liver disease, nephrotic syndrome, obesity.
- Review medicines and supplements, especially biotin, glucocorticoids, opioids and dopamine blockers.
- Talk to the laboratory: repeat on another platform, dilute, use blocking reagents or test for macroprolactin.
- Repeat the test before imaging or treating.
Common traps and how to unmask them
| Trap | What you see | How to unmask it |
|---|---|---|
| Biotin interference | Odd results, often a pattern that looks like hyperthyroidism | Stop biotin for a few days and repeat |
| Heterophile antibodies | A result that does not fit the patient | Repeat on another platform or with blocking reagents |
| Macroprolactin | High prolactin in a person without symptoms | Polyethylene glycol precipitation |
| Hook effect | Large sellar mass with only a mild prolactin rise | Repeat after serial dilution |
| High binding proteins | High total T4 or cortisol in pregnancy or on oestrogens | Use free hormone and pregnancy-specific ranges |
| Non-thyroidal illness | Low T3, sometimes low TSH, in a very sick patient | Recheck after recovery |
In clinical practice
- Ask every patient about supplements before thyroid tests; biotin is often not mentioned unless asked.
- Reference ranges are usually adult ranges; children and pregnant women have their own.
- A single low morning testosterone or cortisol is repeated before a deficiency is diagnosed.
Next step / Initial management
- When a result is discordant, call the laboratory before ordering imaging.
- Repeat borderline results under standard conditions (morning, fasting, off interfering drugs where safe).
- For a high prolactin, check pregnancy, medicines, TSH and macroprolactin before an MRI.
- Refer to endocrinology when results stay discordant after repeating them or when a dynamic test is needed.
Common traps
- A high prolactin calls for an MRI straight away: first exclude pregnancy, dopamine-blocking drugs, hypothyroidism and macroprolactin.
- Thyroid results that look like Graves disease are always real: biotin interference can mimic them, and Graves disease must not be diagnosed from it.
- A modest prolactin rules out a giant prolactinoma: the hook effect can hide it, and missing it sends the patient to surgery that a tablet could have avoided.
- Abnormal thyroid tests in intensive care need thyroid treatment: non-thyroidal illness lowers T3 and sometimes TSH without thyroid disease; tests are rechecked after recovery.
- Testosterone can be measured at any time of day: an afternoon sample makes healthy men look deficient.
Test yourself
Think of the answer, then tap the question to see it.
A man taking a hair supplement has a low TSH and high free T4 but no symptoms. What do you suspect and what do you do?
Biotin interference with the immunoassay. Stop the supplement for a few days, repeat the tests and let the laboratory know.
A 4-cm sellar mass comes with only a mildly raised prolactin. Which laboratory request can change the treatment?
Serial dilution of the sample to exclude the hook effect: a giant prolactinoma is treated first with a dopamine agonist, not surgery.
Why does pregnancy raise total T4 without hyperthyroidism?
Oestrogen raises thyroxine-binding globulin, so more T4 circulates bound; the free hormone, which is what acts, need not change.
Elsewhere in DurgiPedia
Connections
Educational content; it does not replace a medical assessment and is not a clinical instruction. For confusion, persistent vomiting with diabetes, fainting or very low blood pressure in someone who takes or has stopped glucocorticoids, go to an emergency department at once.
Sources: Laboratory Assessment of Thyroid Function (StatPearls (NCBI Bookshelf), 2025); Hyperprolactinemia (StatPearls (NCBI Bookshelf), 2025); Euthyroid Sick Syndrome (StatPearls (NCBI Bookshelf), 2022); Biotin (StatPearls (NCBI Bookshelf), 2023).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.
