Thyroid function panel: TSH, free T4 and free T3

A thyroid function panel asks one question: is the hypothalamic-pituitary-thyroid (HPT) axis underactive, normal or overactive, and, if abnormal, is the problem in the thyroid or above it? The three routine blood tests are thyroid-stimulating hormone (TSH, also called thyrotropin), free thyroxine (free T4, FT4) and free triiodothyronine (free T3, FT3). TSH comes from the anterior pituitary; free T4 and free T3 come from the thyroid. Each test is a different piece of the feedback loop, and reading them together is what gives the panel its power ^[1].

TSH is the screening test for most patients. Most laboratories run TSH first and add FT4 automatically if TSH is outside the reference range; children, pregnancy, known or suspected pituitary disease and amiodarone treatment are common reasons to measure FT4 (and FT3) from the start ^[6]. NICE advises measuring TSH alone in adults when secondary (pituitary) dysfunction is not suspected, then FT4 in the same sample if TSH is above the reference range, and FT4 with FT3 if TSH is below the reference range ^[2]. When pituitary disease is suspected, both TSH and FT4 are measured together because a low or normal TSH can no longer be taken to mean a normal thyroid ^[2].

Only thyroid hormones that are not bound to protein are biologically active, and most of the circulating hormone is protein-bound. Estrogen, pregnancy, opiates, androgens and glucocorticoids change binding-protein levels and therefore total hormone levels, while the free fractions are what the tissues actually see; this is why free T4 and free T3 have replaced total T4 and total T3 in routine practice ^[5].

TSH (thyrotropin). TSH is a pituitary glycoprotein that drives the thyroid to make and release hormone. The HPT axis works on negative feedback: as free T4 and free T3 rise, TSH falls, and as they fall, TSH rises. This relationship is close to log-linear, so a small change in thyroid hormone produces a large change in TSH. In one healthy person thyroid-hormone levels stay remarkably constant around a personal "set point", so TSH is a sensitive early indicator of change long before free T4 becomes clearly abnormal ^{[1] [5]}. TSH reaches its lowest level in the late afternoon and peaks between midnight and 4 a.m.; this matters for timed tests ^[5].

Free T4 (FT4). Free T4 is the main secretory product of the thyroid. It is measured directly when possible, although most laboratories use automated immunoassays that depend on method and on binding proteins ^[5]. Free T4 alone falls in about the last quarter of the normal range in central (pituitary or hypothalamic) hypothyroidism, and it is the test that must be used when TSH cannot be trusted ^[4].

Free T3 (FT3). Free T3 is the more active hormone. Most of it is not made in the thyroid at all but is produced in peripheral tissues by deiodination of free T4. Because of this biology FT3 is more sensitive than FT4 for hyperthyroidism and is low early in critical illness, but it is poor at detecting hypothyroidism and can be misleading in non-thyroidal illness ^[5]. In iodine-sufficient adults the thyroid contributes only a small fraction of circulating T3, so FT3 can be normal or only slightly low even when FT4 and TSH are clearly abnormal in hypothyroidism.

Thyroid antibodies. These are not hormones and are often added to the panel when the pattern does not fit a simple primary disorder:

• Anti-thyroid peroxidase (anti-TPO) antibodies are found in most patients with Hashimoto thyroiditis and in many with Graves disease, and they predict progression from subclinical to overt hypothyroidism ^[5].
• TSH receptor antibodies (TRAb) include stimulating, blocking and neutral forms. Stimulating TRAb is the cause of Graves disease and is measured either as TSH-binding inhibition or as thyroid-stimulating immunoglobulin. A positive TRAb helps to separate thyrotoxicosis from hyperthyroidism and to predict fetal or neonatal thyroid disease in pregnancy ^{[2] [5]}.
• Anti-thyroglobulin antibodies (anti-Tg) are present in a proportion of thyroid autoimmune disease and in roughly one tenth of the general population. They are critical when thyroglobulin is used as a tumour marker in differentiated thyroid cancer, because even modest titres make the thyroglobulin assay unreliable ^[5].

Reference intervals are not universal. They depend on the method, the population used to set them, age, sex, iodine status and the units of the local laboratory, so the laboratory's own interval always takes priority and all thyroid results should be read against it ^{[5] [6]}. The values below are example intervals from one European laboratory using a Roche Elecsys platform; they are a teaching illustration, not your laboratory's interval ^[6].

A few points about intervals that every clinical result depends on:

• TSH has a right-skewed, roughly log-normal distribution, so the upper limit is set by the 97.5th percentile. European laboratories often use 0.4 to 4.0 mIU/L, while the NHANES III disease-free, TPO antibody-negative population had an upper limit of 4.5 mIU/L ^[5].
• TSH rises with age, both because of altered TSH isoforms and a gradual reset of the set point, and upper limits move from about 3.75 mIU/L at 40 years to about 5.0 mIU/L at 90 years ^[5].
• Pregnancy (2011 ATA guidance): TSH is lower throughout pregnancy, and the ATA standard 2011 recommendation is trimester-specific intervals of 0.1 to 2.5 mIU/L in the first, 0.2 to 3.0 in the second and 0.3 to 3.0 in the third, or your laboratory's own method-specific trimester ranges; free T4 intervals also shift and are method-dependent in pregnancy ^[3].
• One laboratory example (Roche) gives first/second/third trimester TSH intervals of 0.3-4.6, 0.4-4.1 and 0.2-3.2 mU/L ^[6], showing how much ranges still vary.

Causes that sit behind the common patterns:

• Levothyroxine. Too much raises free T4 (and T3) and suppresses TSH; too little or malabsorption leaves TSH high and free T4 low. Macro-TSH and interference can also keep TSH high despite apparent over-replacement ^[1].
• Non-thyroidal illness ("sick euthyroid"). In critical illness free T3 falls first, within 24 hours; free T4 then falls; TSH can be low, normal or mildly raised and may transiently exceed the normal range on recovery. These changes resolve as the illness resolves and can look like thyroid disease if read without the clinical context ^[5].
• Pregnancy. Human chorionic gonadotrophin (hCG) stimulates the TSH receptor; TSH falls in the first trimester, total hormone levels rise through increased thyroxine-binding globulin, and the gland enlarges. Only trimester-specific intervals make sense of the results ^[3].
• Drugs. Amiodarone blocks type I deiodinase, so FT4 often rises and FT3 falls initially; in iodine-sufficient regions it commonly causes hypothyroidism (especially with anti-TPO), and in iodine-deficient regions thyrotoxicosis is more common, with a median onset of about 3.5 months for type 1 and 30 months for type 2 disease ^[5]. Lithium and tyrosine kinase inhibitors push toward primary hypothyroidism; dopamine agonists, glucocorticoids, somatostatin analogues and metformin suppress TSH; phenytoin, salicylates, high-dose intravenous frusemide and heparin can raise free hormone readings through carrier-protein displacement ^[5]. Biotin (100 to 300 mg daily for metabolic disorders, or 1 to 10 mg tablets bought for hair or nails) can produce a spurious low TSH with a spurious high FT4 on streptavidin-biotin immunoassays, falsely suggesting hyperthyroidism; routine multivitamins contain 1 to 3 mg and are usually harmless ^{[2] [5]}. ^[1]
• Central defects and confounders. Glucocorticoids, dopamine, cocaine, metformin, anti-epileptics and anti-psychotics can all lower TSH secretion, and many drugs can create biochemical pictures that resemble central hypothyroidism ^[4].
• Iodine. Insufficient iodine is the commonest cause of hypothyroidism worldwide in the iodine-deficient setting; in iodine-replete regions autoimmune thyroiditis is the dominant cause ^[5].

The panel is read with TSH first, then free T4, then free T3, always with the clinical state in mind. The common named patterns are ^[1]:

• Primary hyperthyroidism (overt): TSH less than 0.03 mIU/L (undetectable on most ultrasensitive assays), raised free T4 and raised free T3 ^[1].
• Subclinical hyperthyroidism: low or fully suppressed TSH with free T4 and free T3 still inside the reference range. Full suppression carries the stronger evidence for atrial fibrillation and osteoporosis, especially in older and postmenopausal patients ^[1].
• Primary hypothyroidism (overt): raised TSH with low free T4 (and usually low free T3) ^[1].
• Subclinical hypothyroidism: raised TSH with free T4 still inside the reference range. TPO antibody positivity predicts progression and guides treatment ^{[1] [5]}. NICE considers levothyroxine for adults with subclinical hypothyroidism when TSH is 10 mIU/L or higher on two occasions 3 months apart, and a trial for younger symptomatic adults when TSH is above range but below 10 ^[2].
• Central (secondary) hypothyroidism: low or inappropriately normal TSH with low free T4 (and low free T3). TSH cannot be used as the thyroid marker here; the full pituitary profile is mandatory because cortisol deficiency must be excluded before starting therapy ^{[1] [4]}.
• Sick euthyroid syndrome (non-thyroidal illness): usually low or low-normal TSH with low or normal free T4 and low free T3 early in the illness. It resolves with recovery and becomes the diagnosis only after primary and central causes have been excluded ^{[1] [5]}.
• Resistance to thyroid hormone or a TSH-secreting pituitary adenoma: free T4 (and often free T3) raised with TSH that is not suppressed, or even raised. Once drug and assay interference are excluded this is rare, and referral to an endocrine centre is required to separate the two ^{[1] [5]}.
• Thyroid hormone resistance (RTH beta): a rare autosomal dominant syndrome with raised thyroid hormones and unsuppressed TSH; it is in the same differential ^[5].
• Familial dysalbuminaemic hyperthyroxinaemia (FDH): a rare albumin variant with higher affinity for T4, giving raised total and higher-than-expected free T4 readings on some assays, usually with a normal TSH and no thyrotoxicosis ^[5].

Every sheet below is an illustrative sample invented for teaching. None of it is a real patient or a real laboratory report.

Sample sheet 1 (normal adult male, 34 years, asymptomatic). Illustrative sample only.

Step-by-step interpretation:

1. TSH 2.10 mIU/L sits inside the local reference interval (0.3 to 4.2) ^[6]. ^[5]
2. With a normal TSH, a single screening TSH result is sufficient in an adult without suspected pituitary disease ^{[2] [6]}.
3. Free T4 14.6 pmol/L and free T3 4.8 pmol/L also sit inside their intervals, so there is no discordance ^[6].
4. The clinical state is asymptomatic; no symptoms, antibody tests or further thyroid investigation are required on these results ^[1]. Conclusion: this is a normal thyroid function panel in an euthyroid adult; no further action is needed, but retest if new symptoms develop ^[2].

Sample sheet 2 (48-year-old woman, tired and gaining weight). Illustrative sample only.

Step-by-step interpretation:

1. TSH 18.4 mIU/L is well above 4.2 and free T4 7.2 pmol/L is below 11; this is the raised-TSH, low-free-T4 pattern of primary hypothyroidism ^[1].
2. Free T3 3.0 pmol/L is only just below its interval, which is typical: FT3 lags and is insensitive for hypothyroidism ^[5].
3. Raised anti-TPO (162 mIU/L, above the local cut-off of less than 35) supports autoimmune thyroiditis as the likely cause ^[5].
4. This is not a central defect, because TSH is clearly raised rather than inappropriately low; pituitary testing is not required for this pattern ^[1].
5. Subclinical forms and non-thyroidal illness have been excluded because free T4 is low, not normal ^[1]. Conclusion: this is overt primary (autoimmune) hypothyroidism; standard practice is to start levothyroxine as first-line treatment and recheck TSH every few months until it stabilises, then annually ^[2].

Sample sheet 3 (55-year-old, weight loss, palpitations). Illustrative sample only.

Step-by-step interpretation:

1. TSH is fully suppressed (less than 0.1 mIU/L) while both free T4 and free T3 are above their intervals: the low-TSH, high-hormone pattern of primary hyperthyroidism ^[1].
2. Free T3 is disproportionately high, which is common in T3-toxicosis and supports a thyroidal source rather than exogenous T4 alone ^[1].
3. A positive TRAb confirms Graves disease as the cause of thyrotoxicosis rather than toxic nodular goitre or transient thyroiditis ^{[2] [5]}.
4. Non-thyroidal illness and exogenous hormone have been excluded by the clinical context and by the persistently high TRAb; assay interference is unlikely because both FT4 and FT3 are raised together with a suppressed TSH ^[1]. Conclusion: this is thyrotoxicosis due to Graves disease; specialist treatment (antithyroid drugs, radioactive iodine or surgery) will be decided by the endocrine team according to the 2019 NICE guideline ^[2].

• TSH alone is not enough when the pituitary is in doubt. In known or suspected hypothalamic or pituitary disease, a low or even normal TSH must not be read as a normal thyroid; free T4 has to be measured and the result followed, with a full pituitary profile ^{[1] [4]}.
• TSH can lag behind treatment. After starting levothyroxine or antithyroid drugs, free T4 responds faster than TSH, so the dose is titrated against free T4 in the early weeks and against TSH later ^[5].
• Non-thyroidal illness mimics disease. Critically ill patients regularly show low free T3 first, then low free T4, and TSH changes late in recovery. Testing during acute illness, unless thyroid disease is strongly suspected, produces misleading results; the changes resolve as the patient recovers ^[5].
• Biotin interference is real and dose-dependent. Doses above 100 mg daily can create a false low TSH with a false high FT4; patients taking it should be identified before testing, and where interference is suspected the sample should be re-run on a biotin-free method or pre-treated with streptavidin ^[5]. NICE lists high-consumption biotin as a documented cause of falsely high or low test results ^[2].
• Heparin and other binding competitors. Heparin releases free fatty acids that displace T4 and T3 from thyroxine-binding globulin, raising free but not total hormone levels; TSH is usually normal and the patient is not thyrotoxic ^[1].
• Amiodarone changes all three analytes. Early amiodarone use transiently raises FT4, lowers FT3 and raises TSH; later it causes either hypothyroidism or thyrotoxicosis depending on iodine status and the type of thyroid dysfunction ^[5].
• Pregnancy needs trimester-specific intervals. Non-pregnant ranges misclassify many first-trimester samples; if your laboratory does not have its own pregnancy ranges, the 2011 ATA fallback of 0.1 to 2.5, 0.2 to 3.0 and 0.3 to 3.0 mIU/L is a starting point ^[3].
• Antibody results are not thyroid hormones. A raised TPO antibody only predicts risk and confirms autoimmunity; it does not by itself make a patient hypothyroid, and it should not be repeated in adults with primary hypothyroidism ^{[2] [5]}.
• Assay interference should not be dismissed. Heterophilic antibodies, autoantibodies to thyroid hormones, macro-TSH, familial dysalbuminaemic hyperthyroxinaemia and Tg antibodies can all make the panel look wrong. Sending the sample to another laboratory using the same platform may not clear the discrepancy, so discuss it with the laboratory ^{[1] [5]}.
• Older patients have a different set point. Mild TSH elevations between 4 and 10 mIU/L in an older TPO-negative patient can be watched rather than treated ^[5].

• The HPT axis is a negative-feedback loop: rising free T4 and free T3 suppress TSH, and falling free T4/free T3 raise TSH; a small hormone change makes a large TSH change ^{[1] [5]}.
• TSH is the usual screening test; FT4 is added if TSH is outside the range, and FT4 with FT3 if TSH is low ^{[2] [6]}.
• Read results against the local reference range, not a textbook number, because they differ by method, age, sex and iodine status ^{[5] [6]}.
• Primary hypothyroidism: high TSH, low FT4. Subclinical hypothyroidism: high TSH, normal FT4, with TPO positivity predicting progression ^{[1] [5]}.
• Primary hyperthyroidism: fully suppressed TSH with raised FT4 and/or raised FT3 ^[1].
• Central (pituitary or hypothalamic) hypothyroidism: low or inappropriately normal TSH with low FT4, which is why TSH alone misleads ^{[1] [4]}.
• Sick euthyroid syndrome is common in acute illness and mimics both hypo- and hyperthyroidism; free T3 is the first to fall ^[5].
• Pregnancy changes every value; use trimester-specific intervals ^[3].
• Drugs, especially amiodarone, lithium, heparin and biotin, and non-thyroidal illness all distort the panel; the medication and supplement history is part of reading the result ^{[1] [2] [5]}.
• Antibodies explain aetiology (Hashimoto, Graves) and modify interpretation (TPO predicts progression; TRAb confirms Graves; anti-Tg interferes with thyroglobulin assays) ^{[2] [5]}.
• Resistance to thyroid hormone and TSH-secreting adenomas show raised FT4 with non-suppressed TSH and need endocrine referral ^{[1] [5]}.
• Whenever FT4 is low with an inappropriately low TSH, exclude central hypothyroidism and cortisol deficiency before starting thyroxine ^{[1] [4]}.