Total and Free Thyroid Hormone in Serum

3.1 Serum TSH as the initial test of thyroid function

Either serum TSH or free T4 can be used as the initial test of thyroid function, but TSH appears to give better first-line discrimination at slightly higher cost ( 90). A normal serum TSH concentration has high negative predictive value in ruling out primary thyroid disease (2), and this assay has become increasingly used as the single initial test of thyroid function (137), with further assays done routinely if serum TSH is normal. If serum TSH is increased, free T4 is measured on the same sample to distinguish between overt and subclinical hypothyroidism (figure 5 ). In subclinical hypothyroidism, measurement of thyroid peroxidase antibody is also appropriate, because it gives a further indication of the likelihood of progression to overt hypothyroidism (38). If serum TSH is suppressed, i.e.< 0.05 mU/l, both free T4 and free T3 should be assessed to distinguish between overt thyrotoxicosis, T3–toxicosis and subclinical thyrotoxicosis. The interpretation of subnormal TSH values is influenced by the functional sensitivity of the particular assay (see Chapter ). Subnormal detectable serum TSH levels in the range 0.05-0.4 mU/l merit follow-up, but there is currently no consensus as to whether such values merit the designation “ subclinical thyrotoxicosis” (24c, 24d, 40a). Notably, such values appear to be more likely to return towards normal than to progress to overt thyrotoxicosis ( 65).

The rationale for using TSH alone as a first-line test of thyroid function rests on the assumptions that primary, or thyroprivic hypothyroidism is much more common than central or secondary hypothyroidism and on the fact that serum TSH falls outside the reference range early in the natural history of thyrotoxicosis or progressive thyroid failure. The major weaknesses of this approach are first, the likelihood of missing secondary hypothyroidism (in which the concentration of immunoreactive serum TSH is frequently normal rather than low), and on the frequency of abnormal TSH concentrations in the absence of thyroid dysfunction, especially in patients with an associated illness.

Beckett and Toft (137a) have pointed out the adverse consequences of missing secondary hypothyroidism due to pituitary failure, on the basis of normal serum TSH. The diagnosis of this disorder is notoriously difficult, with diverse presentations to a wide range of practitioners who may not be attuned to the key clinical features. He estimated that as many as 1500 cases per year of central hypothyroidism could be missed in the UK if the “TSH first” policy were followed inflexibly. This potentially serious diagnostic problem was shown by Wardle et al (78) who reported an analysis of 56,000 requests for thyroid function testing over 12 months from a population of 471,000. Serum TSH was normal in association with subnormal total and free T4 in 15 patients who, on further investigation, had probable hypopituitarism that would have been missed by assessment of serum TSH alone. Cost benefit analyses would therefore need to balance the savings from not measuring serum T4 routinely, against the cost of further investigation and medical care for the missed diagnoses, in addition to some allowance for burden of suffering and potential litigation. As advances in technology reduce the unit cost of measuring T4, the arguments against the “TSH first” approach will probably become more compelling.

Table 5 summarizes the situations in which TSH alone can give a false indication of thyroid status. The sensitivity of TSH-based testing strategy is seriously impaired when TSH is normal in the presence of clinically important thyroid dysfunction.

Table 5: Situations in which serum TSH alone can give a false or uncertain indication of thyroid status.
Condition	TSH	fT4	fT3
Primary abnormality of TSH secretion
Pituitary-hypothalamic abnormality	L-N	L
Extremely premature infants	L-N	L	L
Central TSH excess	N-H	H	H
Thyrotoxicosis
T3 toxicosis	U	N	H
Subclinical	U	N	N
Early Treatment	U	H-N-L	H-N-L
TSH assay artefact	L-N-H	H	H
Hypothyroidism
Subclinical	H	N
Early Treatment	H	L-N
Thyroid hormone resistance	N-H	H	H
Medications
Dopamine	L	N	N
Glucocorticoids	L	N	L-N
Amiodarone (acute)	H	N-H	L
N: normal; L: low; H: high; U: undetectable.

Estimation of free T4, or total T4 linked to a measurement of the thyroid hormone binding ratio, should now be standard whenever TSH is abnormal, or in situations where TSH alone is known to give an inaccurate indication of thyroid status. Assays for total T4 and T3 in unextracted serum include a reagent such as 8-anilinonaphthalene sulfonic acid that blocks T4 and T3 binding to serum proteins, so that total hormone is available for competition with the assay antibody. Assays for free T4 or T3 omit this blocking reagent and use a wide variety of manoevres to isolate a moiety that reflects the free hormone concentration. The theoretical basis, practical utility, and validity of the many different approaches to the estimation of serum free T4 and T3, have been considered in detail (83, 138) (see Chapter 6A).

Two key assumptions in any method of free T4 estimation are (i) that the dissociation of bound hormone with sample dilution is similar in samples and standards, and (ii) that samples and standards show identical protein binding of the assay tracer. If either of these conditions is breached, the assay is likely to give inaccurate results. Serum free T4 and free T3 can be estimated either by two-step methods that separate a fraction of the free hormone pool from the binding proteins before the assay incubation, or by one-step methods in which the free hormone concentration is measured in the presence of binding proteins (83). Many of the one-step methods become invalid when the sample and standard differ in their binding of assay tracer, but the two-step methods are less prone to non-specific artefacts. Almost all techniques of estimating free T4 give a useful correction for moderate variations in serum TBG concentration, but no method can yet accommodate extreme variations of serum TBG, qualitative or quantitative albumin abnormalities, and the effect of circulating competitors for T4 binding to TBG (83).

a.	In suspected thyrotoxicosis with suppressed TSH and normal serum T4, to identify T3-thyrotoxicosis and distinguish this entity from subclinical thyrotoxicosis.
b.	During antithyroid drug therapy to identify persistent T3 excess, despite normal or low serum T4 values (139).
c.	For diagnosis of amiodarone-induced thyrotoxicosis, which should not be based on T4 excess alone because of the frequency of euthyroid hyperthyroxinemia during amiodarone treatment (106)
Serum T3 measurements may also be useful:
d.	For estimation of the serum T3-T4 ratio. A high ratio (>0.024 on a molar basis or >20 calculated as ng/µg) that persists during antithyroid drug treatment suggests that patients with hyperthyroid Graves' disease are unlikely to achieve remission (140). This ratio usually is lower in iodide-induced thyrotoxicosis (141) or thyrotoxicosis caused by thyroiditis (142) than in Graves' disease.
e.	To detect early recurrence of thyrotoxicosis after cessation of antithyroid drug therapy.
f.	To establish the extent of T3 excess during high-dose replacement or suppressive therapy with T4 (143), or after an intentional T4 overdose.

Low serum T3 concentrations have little specificity or sensitivity for the diagnosis of hypothyroidism. Many patients with nonthyroidal illness have low values, and the serum T3 concentration can remain in the reference range until hypothyroidism is severe (144). Serum T3 values are usually interpreted together with the levels of T4 (table 6).

Measurement of serum T3 has also been advocated without measurement of T4 for population screening studies as a follow-up to the finding of a suppressed TSH, because of its presumed sensitivity for the detection of thyrotoxicosis (5). It is likely that variations related to age (121) and associated illness (145) will limit its value as a first line test.

3.4 Indications for TRH Testing

The development of high-sensitivity TSH assays has almost eliminated the need for TRH testing in clinical practice. With intact hypothalamic-pituitary function, there is a close correlation between the TSH response to TRH and the basal level of serum TSH, when measured by a highly sensitive assay (146). Hence, TRH testing now offers little diagnostic advantage over accurate measurement of the basal serum TSH concentration in the detection of thyrotoxicosis (146). However, measurement of serum TSH 20 to 30 minutes after intravenous injection of 200 to 500 µg TRH remains useful for several purposes:

This iodinated 660 kDa dimeric glycoprotein, which is the backbone for thyroid hormone biosynthesis, is normally detectable in serum and is released in excess in a wide variety of situations where thyroid tissue is overactive, inflamed or proliferating. Undetectable serum levels suggest absence or suppression of thyroid tissue.

a.	Follow-up of treatment for differentiated thyroid cancer to identify or rule out the presence of residual thyroid tissue, whether in normal or metastatic sites. An undetectable serum thyroglobulin concentration in the presence of high TSH is presumptive evidence that differentiated thyroid tissue has been ablated (150). In contrast, the level of serum thyroglobulin that persists when serum TSH is suppressed, can give a useful index of tumor burden (150). Concurrent measurement of thyroglobulin antibody is required for reliable interpretation of serum thyroglobulin values (see below).
b.	Investigation of atypical thyrotoxicosis, where there is a suspicion of thyrotoxicosis factitia in which serum thyroglobulin is undetectable (139).
c.	Assessment of the activity of inflammatory thyroiditis, eg subacute thyroiditis, or that due to amiodarone (106).
d.	It has been reported that assay of thyroglobulin on the needle washes after biopsy of extrathyroidal neck masses is useful in identifying metastatic thyroid tissue (150a)
e.	Serum thyroglobulin may serve as a sensitive, if non-specific marker of iodine deficiency or ineffective synthesis of thyroid hormone (150b)

Because elevation of serum thyroglobulin concentration occurs in a wide range of thyroid disorders, interpretation of results always requires a knowledge of (i) the clinical context, (ii) the serum TSH concentration and (iii) whether thyroglobulin antibodies are present. For optimal thyroglobulin assay, it may be preferable to use radioimmunoassay for samples that are antibody-positive, with use of the more sensitive immunometric techniques to confirm absence of functioning tissue in antibody-negative samples (77).

Thyroglobulin is amongst the most difficult serum assays in current routine diagnostic use. As detailed elsewhere (77) and in chapter 6A , potential problems are:

a.	Interference in assay separation systems from thyroglobulin antibodies present in the patient serum, that can lead to an underestimation of serum concentration, particularly in immunometric assays.
b.	Interference from "hook effects" that lead to underestimation of high concentrations.
c.	Inconsistent standardization, leading to difficulty in serial interpretation of long-term follow-up data.

3.6 Thyroglobulin antibody (TgAb)

Assays for TgAb have evolved from tissue immunofluorescence, through agglutination and radioimmunoassay techniques to immunometric assays. Standardization remains difficult and sensitivity is variable (77). (see Chapter 6A).

a.	As an essential component of the interpretation of assays for serum thyroglobulin, to establish whether endogenous antibodies could be responsible for spuriously low thyroglobulin values by interfering with assay separation, particularly in immunometric assays.
b.	As secondary follow-up criterion in differentiated thyroid cancer, where a progressive decline in antibody concentration may follow successful ablation (150). There is currently no established role for measurement of TgAb in the assessment of autoimmune thyroid disease, because of the greater sensitivity and specificity of thyroid peroxidase antibody.

This entity, previously termed antimicrosomal antibody, is closely linked with autoimmune thyroid disease, in particular Hashimoto's thyroiditis (151). Measurement of thyroid peroxidase antibody is indicated as follows:

a.	To identify an autoimmune cause for primary hypothyroidism.
b.	In individuals with the TSH excess of mild thyroid failure, in whom a positive TPOAb indicates an approximately two-fold increase in risk of progression to overt hypothyroidism (38). A strongly positive result in the presence of mild thyroid failure may influence the decision to commence replacement (89).
c.	Prior to treatment with medications that may precipitate hypothyroidism in TPOAb -positive individuals, eg amiodarone (106), lithium (11,152), interferon alpha (12a,153).
d.	In the assessment of women with euthyroid goitre, in order to evaluate the risk of future hypothyroidism, at a stage before there is any increase in serum TSH (38).
e.	In early pregnancy, or in women intending to become pregnant, as a predictor of the potential for intrapartum hypothyroidism, or postpartum thyroid dysfunction (154).
Possible additional indications include:
f.	Routine measurement before or during pregnancy, if there is confirmation of recent studies that suggest increased fetal loss in mothers with mild TSH excess (23). This indication would become compelling if fetal loss can be alleviated by thyroid hormone treatment.
g.	In perimenopausal women in whom TPOAb positivity may predispose to depression by an effect that appears to be independent of thyroid function and menopausal status (155).

3.8 TSH Receptor Antibody (TRAb)

Antibodies that interact with the TSH receptor can be measured either by bioassay, or competitive binding techniques (101). Bioassays, that use thyroid cells of human or animal origin, or cells with transfected TSH receptor, generally depend on tissue production of cAMP for quantitation. These assays allow a distinction to be made between antibodies with thyroid stimulating (TSAb) and blocking (TBAb) activity. Radioreceptor (thyroid binding inhibitor immunoglobulin, TBII) assays that measure competition by circulating immunoglobulin for specific binding of labelled TSH, are more widely available, but do not distinguish between blocking and stimulating activity. A competitive binding assay using a recombinant human TSH receptor, showed 98-99% positivity in active Graves' disease, with positive results in <1% of subjects with non autoimmune thyroid diseases (156), but some assays still in current use do not share these impressive performance characteristics. Indications for the measurement of TRAb vary in different practice environments, but are clearly applicable to the following situations:

a.	During pregnancy in women with active or previous autoimmune thyroid disease, to assess the risk of neonatal thyroid dysfunction due to transplacental passage of TRAb Guidelines from the European Thyroid Association (157) suggest that TRAb measurements should be made early in pregnancy in women who have received previous ablative treatment, and in the last trimester in women receiving drug treatment for active Graves' disease. Antibody measurement during pregnancy was not regarded as necessary for Graves' disease was in remission (157).
b.	In the differential diagnosis of atypical thyrotoxicosis that may be due to Graves' disease.
c.	In atypical eye disease that may be due to thyroid ophthalmopathy.
d.	In assessing whether an apparent remission of Graves' disease is likely to be sustained, or whether relapse should be anticipated (158).
e.	To assess the chance of achieving a drug-induced remission of Graves' disease (158).

3.9 TSH alpha subunit

Assay of the TSH alpha subunit is indicated where thyrotoxicosis appears to be the result of central autonomous TSH excess to distinguish this entity from thyroid hormone resistance (109, 135). Most TSH producing pituitary adenomas show an increase in alpha subunit (135), whereas levels generally remain normal in thyroid hormone resistance (109). Values are also high in postmenopausal women, in men with hypogonadism and in gonadotrophin-producing pituitary tumors, because both thyrotrophs and gonadotrophs secrete this subunit.

3.10 Serum reverse T3

This assay does not have a general diagnostic role (159), despite previous suggestions that it might be useful in distinguishing true hypothyroidism from the hypothyroxinemia of severe illness.

Table 6. Relationship between serum T4 and T3 in various disorders*
	Serum T4
Serum T3	Low	Normal	High
Low	Severe hypothyroidism TBG deficiency # Severe nonthyroidal illness Euthyroid hypothyroxinemia	Nonthyroidal illness Medications Fetus Restricted nutrition	Thyrotoxicosis with severe nonthyroidal illness Amiodarone
Normal	Iodine deficiency T3 treatment Hypothyroidism		T4 treatment Euthyroid hyperthyroxinemia Thyrotoxicosis with nonthyroidal illness T4 binding autoantibodies
High	Iodine deficiency T3 treatment Antithyroid drugs	T3 toxicosis T3 binding autoantibodies	Thyrotoxicosis Excess T4 ingestion Hormone resistance TBG excess^#
* Excludes short term changes related to initiation or cessation of therapy # Effect on total hormone concentration only; free hormone remains normal.