Thyrotropin previously was known as thyroid stimulating hormone (TSH), and the abbreviation TSH is still used. Direct assay of TSH is now possible with commercially available kits that are as easy to use as those for T4 assay. Thyrotropin has a diurnal variation of 2 to 3 times baseline (literature range, 33%-600%), with highest levels occurring at about 10-11 P.M. (range, 6 P.M.-2 A.M.) and lowest levels at about 10A.M. (range, 8 A.M.-4 P.M.).

Results in thyroid disease. Serum TSH levels are elevated in about 95% of patients with myxedema due to primary thyroid disease, which comprises 95%-96% of hypothyroid patients. Serum TSH levels are low in most cases of secondary (pituitary or hypothalamic) myxedema (about 4% of hypothyroid patients). Some patients with secondary hypothyroidism have normal TSH values when one would expect low TSH levels. The pituitary of these patients is able to secrete a small amount of TSH, not enough to maintain normal T4 levels but enough to leave TSH values within reference range. The T3-RIA values in these patients may be decreased; but in some instances may be low normal, either because of preferential secretion of T3 rather than T4 by subnormal TSH stimulation or because of problems with kit sensitivity in low ranges. In primary hyperthyroidism, serum TSH is decreased; the percentage of patients with decreased values depends on the sensitivity of the particular kit used. Before 1985, most commercially available kits were poorly sensitive in the lower part of their range, and could not easily differentiate low values from zero or from lowormal values. Theoretically, in typical cases of hyperthyroidism the excess thyroid hormone produced causes the pituitary to completely stop production of TSH, reducing serum TSH to zero. The closer a TSH assay can approach zero in detecting TSH, the better it can differentiate between hyperthyroidism and certain other causes for decreased serum TSH (which usually produces a serum TSH value somewhere between lower reference limit and zero). Some nonhyperthyroid etiologies include partial pituitary insufficiency, early or mild hyperthyroidism in some patients, severe non thyroid illness in some patients (depending on the particular TSH assay kit), monitoring thyroid suppressive therapy, and dopamine or high-dose glucocorticoid therapy in some patients.

Therefore, about 1985, many manufacturers had begun modifying their TSH kits so as to increase sensitivity at the lower end of the TSH assay range in order to differentiate lesser decreases of TSH (less likely due to hyperthyroidism) from marked decreases (more likely due to hyperthyroidism). Manufacturers called the new TSH kits “high-sensitivity,” “ultra sensitive,” or “first, second, and third generation.” Unfortunately, these terms were used in different ways, the two most common being the theoretical lower limit the assay might achieve under the best experimental conditions and the “functional” lower limit the assay usually did achieve with a between-assay coefficient of variation less than 20%. Based on the most common usage in the literature, the theoretical lower limit of detection for first-generation assays could detect TSH as low as 0.3-0.1 mU/L (µU/ml); but their functional lower limit usually was no better than 0.3, frequently was no better than 0.5, and sometimes was 0.6-1.0. Second-generation assays theoretically can detect between 0.1-0.01 mU/L. Most have a functional lower limit between 0.07-0.04 mU/L. Generally, a functional ability to detect less than 0.1 mU/L qualifies the assay for second-generation or high-sensitivity (ultra sensitive) status. In the 1990s, a few third generation kits have been reported; these have a theoretical lower limit of detection between 0.01-0.005 mU/L and functional detection at least below 0.01 mU/L.

Drawbacks. Evaluations in the literature and my own experience have shown that not all TSH kits (either “standard” or ultra sensitive) perform equally well. Moreover, laboratory error may produce false normal or abnormal results. In addition, there are conditions other than hypothyroidism or hyperthyroidism that can increase TSH values or decrease TSH values (see box 1 and box 2).

Conditions That Increase Serum Thyroid-Stimulating Hormone Values

Lab error
Primary hypothyroidism
Synthroid therapy with insufficient dose; some patients
Lithium or amiodarone; some patients
Hashimoto’s thyroiditis in later stage; some patients
Large doses of inorganic iodide (e.g., SSKI)
Severe non thyroid illness in recovery phase; some patients
Iodine deficiency (moderate or severe)
Addison’s disease
TSH specimen drawn in evening (peak of diurnal variation)
Pituitary TSH-secreting tumor
Therapy of hypothyroidism (3-6 weeks after beginning therapy [range, 1-8 weeks]; sometimes longer when pretherapy TSH is over 100 µU/ml); some patients
Acute psychiatric illness; few patients
Peripheral resistance to T4 syndrome; some patients
Antibodies (e.g., HAMA) interfering with monoclonal sandwich method of TSH assay
Telepaque (Iopanic acid) and Oragrafin (Ipodate) x-ray contrast media; some patients
Amphetamines; some patients
High altitudes; some patients

Conditions That Decrease Serum Thyroid-Stimulating Hormone Values*

Lab error
T4/T3 toxicosis (diffuse or nodular etiology)
Excessive therapy for hypothyroidism
Active thyroiditis (subacute, painless, or early active Hashimoto’s disease); some patients
Multinodular goiter containing areas of autonomy; some patients
Severe non thyroid illness (esp. acute trauma, dopamine or glucocorticoid); some patients
T3 toxicosis
Pituitary insufficiency
Cushing’s syndrome (and some patients on high-dose glucocorticoid)
Jod-Basedow (iodine-induced) hyperthyroidism
TSH drawn 2-4 hours after Synthroid dose; few patients
Postpartem transient toxicosis
Factitious hyerthyroidism
Struma ovarii
Radioimmunoassay, surgery, or antithyroid drug therapy for hyperthyroidism; some patients, 4-6 weeks (range, 2 weeks-2 years) after the treatment
Interleukin-2 drugs (3%-6% of cases) or alpha-interferon therapy (1% of cases)
Hyperemesis gravidarum
Amiodarone therapy; some patients
*High sensitivity TSH method is assumed.

Some cases of elevated free T4 levels accompanied by TSH values that were inappropriately elevated (above the lower third of the TSH reference range) rather than depressed have been reported. These cases have been due to TSH-producing pituitary tumors, defective pituitary response to thyroid hormone levels, or peripheral tissue resistance to the effects of thyroid hormone (laboratory error in T4 or TSH assay also must be excluded). Some (a minority) of the current TSH kits (predominantly the solid-phase double-antibody technique) may cross-react with hCG if hCG is present in very large amounts. These TSH kits will indicate varying degrees of (false) TSH increase during pregnancy and in the rare syndrome of hyperthyroidism due to massive hCG production by a hydatidiform mole or choriocarcinoma. Likewise, some of the double-antibody kits using the “sandwich” technique with one or both of the antibodies being the monoclonal type may become falsely elevated due to interference by a heterophil-type antibody in the patient serum that reacts with mouse-derived antibodies. This occurs because monoclonal antibodies are usually produced by a fused (hybrid) cell containing a mouse spleen cell that produces the antibody combined with a myeloma tumor cell that reproduces the hybrid for long periods of time. These human antimouse antibodies (HAMAs) usually occur without known cause and in one study were found in 9% of blood donors. Some manufacturers have attempted to counteract or neutralize the effect of these antibodies, with a variable degree of success. Another antibody problem of a different type concerns artifactual interference in TSH assay procedure by anti-TSH antibodies in patients who had previous TSH injections and who developed antibodies against the injected material.

In severe non thyroid illness, it was soon noticed that a considerable number of the “ultra sensitive” TSH kits produced some abnormal results in clinically euthyroid patients. Previous “standard” TSH kits, although less sensitive in the lower range, generally produced normal TSH values in severe non thyroid illness. The percentage of abnormal results in the ultra sensitive kits varies between different kits. Most abnormality is reported in severe non thyroid illness, and is manifested by decreased TSH in about 15% of cases (range, 0%-72%) and elevated TSH in about 7% of cases (range, 0%-17%). When decreased in true hyperthyroidism, TSH usually is zero; whereas in non thyroid illness it is usually not that low (first generation TSH assays cannot make this distinction). Of those elevated, most (but not all) were less than twice the upper limit of the reference range. The decreased values return toward normal as the patient recovers; in some reports, most of the elevated values occurred in the recovery phase. This has been interpreted as pituitary inhibition during the acute phase of the illness and release of inhibition in the recovery phase

There is some controversy whether or not elevated TSH levels in some of these conditions (except adrenal insufficiency and heterophil or anti-TSH antibodies) represent a mild hypothyroid state with the pituitary forced to secrete more TSH to maintain clinical euthyroidism. The serum TSH level in primary hypothyroidism is usually elevated to values more than twice normal and frequently to more than 3 times normal, whereas TSH levels in the other conditions frequently are less than twice normal and in most cases are less than 3 times normal.