|Home | About | Journals | Submit | Contact Us | Français|
Laboratory assessment of thyroid function is now often initiated with a low pre-test probability, by clinicians who may not have a detailed knowledge of current methodology or testing strategies. Skilled laboratory staff can significantly enhance the choice of appropriate tests and the accuracy of clinical response; such involvement requires both appropriate training and relevant information from the clinician. Measurement of the serum thyroid stimulating hormone (TSH) concentration with an assay of adequate sensitivity is now the cornerstone of thyroid function testing; for untreated populations at risk of primary thyroid dysfunction, a normal TSH concentration rules out an abnormality with a high degree of certainty. However, in several important situations, most notably pituitary abnormalities and early treatment of thyroid dysfunction, serum TSH can give a misleading indication of thyroid status. An abnormal TSH concentration alone is never an adequate basis for initiation of treatment, which should be based on the typical relationship between trophic and target gland hormones, based on serum TSH and an estimate of serum free thyroxine (T4). Six basic assumptions, some clinical, some laboratory-based, need to be considered, together with the relevant limiting conditions, for reliable use of this relationship. Current methods of free T4 estimation remain imperfect, especially during critical illness. Diagnostic approach differs significantly between initial diagnosis and follow-up of treated thyroid dysfunction. In some situations, serum triiodothyronine (T3) is also required, but serum T3 lacks sensitivity for diagnosis of hypothyroidism, and has poor specificity during non-thyroidal illness. Where assay results are anomalous, most atypical findings can be resolved by attention to the clinical context, without further investigation.
This review summarizes principles for the appropriate use of laboratory assays in the diagnosis and follow-up of thyroid disorders. Symptoms, physical signs, imaging techniques and cytological examination will not be considered in detail, although it is self-evident that laboratory results should be interpreted in this broader context. The recent monograph “Laboratory support for the diagnosis of thyroid disease” from the National Academy of Clinical Biochemistry, USA, should be consulted for detailed guidelines on the preparation, laboratory use and application of current thyroid assays.1 References for specific points in this review are cited at www.thyroidmanager.org.2
It is now well known that the presentations of thyrotoxicosis and hypothyroidism are so diverse (Table 1) that it is difficult to rule out these conditions clinically, or to make a conclusive diagnosis until the disorder is far advanced. Any of the presentations summarized in Table 1 is reason to seek confirmatory clinical features, and to measure serum TSH, using an assay sufficiently sensitive to clearly separate the suppressed values characteristic of thyrotoxicosis from the lower limit of the normal reference interval.
Apart from using laboratory tests when thyroid dysfunction is suspected, or in groups with an increased risk of thyroid dysfunction (Table 2), there are several situations where routine testing is appropriate. First, neonatal screening for congenital hypothyroidism is widely established. Second, recommendations from the American College of Physicians (Table 3) suggest that thyroid dysfunction is sufficiently common in women over 50 to justify routine testing at presentation for medical care (case-finding); the majority of abnormal findings in this group will identify subclinical rather than overt dysfunction (see below). Third, the finding of significant intellectual impairment in the offspring of women who were mildly hypothyroid early in pregnancy3 may justify routine testing of thyroid function, either before conception, or as early as possible in pregnancy.
The application of population reference intervals to individuals can obscure rather than clarify the diagnosis of minimal or mild thyroid dysfunction. Serial assessment of thyroid function shows that individuals remain close to a particular set-point for TSH and T4.4 Such studies suggest that potentially important variations from the individual set-point can still fall within the population norm. This issue is most relevant for serum TSH values in the upper tail of the logarithmically distributed “normal” range, especially in the assessment of optimal replacement therapy, or diagnosis of mild thyroid failure. Notably, TSH values of 2–4 mU/L, still within the reference interval, are associated with an increased prevalence of positive thyroid peroxidase (TPO) antibody.5 In practice, individualized interpretation of laboratory data remains difficult unless there is information on set-point prior to assessment for possible thyroid dysfunction.
Current laboratory techniques make the diagnosis and follow-up of thyroid disorders straightforward in the large majority of patients, but, in a small minority, problems due to assay artefacts or atypical clinical presentations can lead to misdiagnosis, inappropriate treatment, or unnecessary further investigation. Such cases emphasize the importance of continuing review of diagnostic methodology by careful clinical correlation.
The distinction between “overt” and “subclinical” hypothyroidism or thyrotoxicosis is based on whether an abnormal serum TSH concentration is associated with abnormal levels of the circulating thyroid hormones, T3 and T4, or whether serum TSH alone is abnormal. The terminology “mild thyroid failure” rather than “subclinical hypothyroidism” is gaining support, based on evidence that potentially important tissue abnormalities can occur during progressive thyroid failure before the serum T4 concentration becomes clearly subnormal. It should be noted that the more frequently thyroid function is tested in the absence of clinical features, the greater the proportion of results with serum TSH as the sole abnormality.
The terminology of thyroid antibody assays can be ambiguous. The terms microsomal and TPO antibody refer to the same moiety, which is the most sensitive marker of immune thyroid damage, usually associated with the lymphocytic infiltration that is most extreme in Hashimoto’s thyroiditis. The more sensitive and specific radioimmunoassay for TPO antibody has now superseded microsomal antibody techniques. Anti-thyroglobulin assays are less relevant for the diagnosis of immune thyroid disease, but are crucial for the valid interpretation of serum thyroglobulin assays, for example in the follow-up of differentiated thyroid cancer. Measurement of TSH receptor antibody (TRAb) identifies the probable causative agent in Graves’ disease, generally by measuring the extent to which a test serum inhibits binding of labelled TSH to a receptor preparation. Such assays do not distinguish between stimulatory and inhibitory activity.
The Whickham study, from an iodine replete region in Northern England, showed a prevalence of 1.9–2.7% overt thyrotoxicosis and 1.4–1.9% overt hypothyroidism in women, with progressive increase with age; prevalence in males was 10-fold lower.5 Estimates of subclinical hypothyroidism were 4–5 fold higher, with about 10% of women over 50 showing an increase in serum TSH, again with progressive increase with age. Further studies from the UK, USA and Australian data from the Busselton study suggest a similar prevalence.6,7
The 20 year Whickham follow-up showed that the likelihood of women developing hypothyroidism rose with age and was increased about 8-fold if either TPO antibody was positive or serum TSH increased in the initial study; this risk rose to almost 40-fold if both were abnormal. The likelihood of eventual hypothyroidism increased progressively for initial TSH values over 2 mU/L, values still well within the reference interval.
Thyroid dysfunction is also common when younger women are tested post-partum. A Perth study found abnormal thyroid function in 11.5% of women tested 6 months after delivery, with TSH values over 4.8 mU/L in 6%, of whom almost 90% had positive TPO antibody, indicating an autoimmune abnormality.8 In about half the untreated women with initial TSH elevation, TSH remained increased 30 months after delivery, consistent with other studies that show an increased prevalence of late hypothyroidism after postpartum dysfunction.
Findings from one region do not necessarily apply in other populations, because of ethnic differences or environmental variations such as iodine intake. For example, in Hong Kong, where iodine intake is marginally deficient, only 1.2% of Chinese women aged over 60 years had serum TSH values >5 mU/L, with a comparable prevalence of suppressed values indicating possible thyrotoxicosis. In general, hypothyroidism is more common with abundant iodine intake, with goitre and subclinical thyrotoxicosis more common with low iodine intake. Hence, optimal strategies for thyroid testing may vary between regions.
To place laboratory testing in perspective we need to consider the sensitivity and accuracy of clinical assessment for thyroid dysfunction. Studies of patients evaluated in primary care show that clinical acumen alone lacks sensitivity and specificity. In two Scandinavian studies of over 3000 unselected patients who were assessed by both clinical and laboratory criteria, a thyroid disorder was not suspected by primary care physicians in over 90% of those who tested positive, even when clinical features were apparent in retrospect.9,10 Further, in up to one-third of patients evaluated for suspected thyroid dysfunction by specialists, laboratory results led to revision of the clinical assessment.
There are some dissenting opinions on the relative value of clinical and laboratory evaluation of thyroid function. Some have expressed the view that the clinical criteria are being sidelined, while biochemical assessments are lacking in specificity. This point is generally made by considering TSH and free T4 measurements individually, rather than in the trophic hormone - target gland relationship that is the cornerstone of endocrine diagnosis (see below). Thyrotoxicosis and hypothyroidism can each have important consequences before the usual clinical features appear (Table 4). It may be no more valid to consider these diagnoses only when typical symptoms or signs appear, than to wait for the thirst and polyuria before considering the possibility of diabetes!
While laboratory tests facilitate early diagnosis before clinical features are obvious, increased sensitivity carries the price of decreased diagnostic specificity. It remains difficult to distinguish spurious results from those that indicate mild dysfunction, especially in the presence of associated illness, where abnormal TSH and free T4 results lack specificity.
All current methods of measuring TSH, free T4 or free T3 in serum, whether by immunoassay or immunometric techniques, are comparative, i.e. they depend on the assumption that the unknown sample and the assay standards are identical in all measured characteristics other than the concentration of analyte. A result will be spurious when this condition is not fulfilled, for example when a sample shows anomalous binding of tracer to serum proteins or antibodies, or non-specific interference, either with the system that separates bound from unbound tracer, or the assay signal. Heterophilic antibodies remain a potential cause of spurious assay results and there are currently no techniques that conclusively rule out this type of interference.
Secretion of TSH from the anterior pituitary is regulated by negative feedback from the serum free thyroid hormone concentrations. Immunometric TSH assays that use two antibodies against different epitopes of TSH show greatly improved assay sensitivity. Serum TSH can be precisely measured to at least 0.03 mU/L, so that the lowest concentrations in normal subjects are clearly distinguishable from those found in thyrotoxicosis. However, assay specificity is not perfect, and false-positive detectable serum TSH is still found in occasional patients with definite thyrotoxicosis. The serum TSH response to changes in serum free T4 is logarithmic; a two-fold change in free T4 induces inverse 10–100 fold changes in TSH. This feedback amplification of the serum TSH response as the serum free T4 increases or decreases, accounts for the fact that serum TSH can fall outside the reference interval several years before there is a diagnostic change in serum free T4.
Typical values for the lower reference limit for TSH are 0.3–0.5 mU/L, with upper limits of 4–5 mU/L, but the mean and median values are in the range 1–1.5 mU/L because of the logarithmic distribution. The terminology for subnormal serum TSH values needs to be clarified. Values associated with thyrotoxicosis, either overt or subclinical are suppressed, (ie <0.03 mU/L) and need to be distinguished from subnormal-detectable values in the range 0.05–0.4 mU/L that do not indicate thyrotoxicosis. Subnormal-detectable values are common in patients with goitre, only a few of whom develop thyrotoxicosis. Notably, during severe illness, serum TSH is often subnormal without indicating any persistent abnormality of thyroid function. Suppressed TSH values <0.03 mU/L can occur during critical illness without indicating any intrinsic abnormality of thyroid function and transient increases to above normal can occur during the recovery phase.
There have been many approaches to the estimation of free thyroid hormone concentrations in serum; some ingenious approaches have been of questionable validity.11 Some free T4 methods have been marketed before they have been rigorously assessed, so that unexpected interferences may only be noted after methods have been used for some time. Currently available assays compensate well when changes in total T4 and T3 are due to altered concentration of thyroid binding globulin (TBG), but no current method reflects the in vivo concentration of free hormone in undiluted serum. Equilibrium dialysis is often considered the reference method, but it is also subject to error, especially as a result of generation of fatty acids during sample incubation, and under-estimation of the effect of competitors that displace T4 and T3 from binding proteins in vivo (see below). Two-step methods that separate a fraction of the free T4 pool from the binding proteins before the assay step are generally least prone to analytical artefacts.
Numerous medications can displace T4 and T3 from TBG (Table 5), but it is technically difficult to get an accurate reflection of these effects with current free T4 methods that use diluted samples. Competitors are usually less protein-bound than T4 itself, so that with progressive dilution, the free concentration of competitor declines before the free T4 concentration. If T4, with a free fraction of about 1:4000 in undiluted serum, is compared with a drug that has a free fraction in serum of 1:50, progressive dissociation will sustain the free T4 concentration at 1:100 dilution, while the free drug concentration decreases markedly after a dilution of only 1:10. This difference leads to an under-estimate of free T4 after sample dilution, as demonstrated by comparing the T4-displacing effect of frusemide in three commercial free T4 assays; the effect of frusemide was least obvious in the method with highest sample dilution (Figure 1). Because of this effect, the apparent free T4 concentration in diluted samples will be an under-estimate in the presence of high therapeutic concentrations of drugs such as phenytoin, carbamazepine, frusemide, mefenamic acid (Ponstan) and salicylate.11
In contrast to such under-estimates of free T4, heparin has the opposite effect to increase the apparent free T4 concentration, due to an in vitro artefact of sample storage.12 In the presence of a normal serum albumin concentration, non-esterified fatty acid (NEFA) concentrations >3 mmol/L will increase free T4 by displacement from TBG, but such concentrations are uncommon in vivo. However, serum NEFA may increase to these levels during storage or incubation of samples from heparin-treated patients, as a result of heparin-induced lipase activity (Figure 2). This effect is accentuated if serum triglyceride concentrations are high, serum albumin concentration is low, or incubation at 37oC is prolonged; under these conditions doses of heparin as low as 10 units may produce this artefact, low molecular weight heparin preparations have a similar effect.
The fallibility of current free T4 methodology is demonstrated by a study of bias in nine commonly used commercial methods in relation to an equilibrium dialysis reference method,13see below. Hence it is absolutely essential, especially in pregnant women and in patients with an associated illness, to interpret results only in relation to a specific method.13 It is becoming clear that, in numerous clinical situations, free T4 estimation does not serve as a robust, reliable index of thyroid function. Despite the theoretical attraction of measuring the concentration of free or biologically active hormone, it remains uncertain whether current free T4 methodology is any improvement over an uncontentious measurement of total T4. The limitations of free T4 methodology are most evident where the diagnosis of thyroid dysfunction is clinically and analytically most difficult. For this reason, it is important that total T4 methods be retained for reference.
Whatever strategy is used for first-line testing, a sensitive serum TSH assay and an estimate of serum free T4 are both necessary for definitive assessment of thyroid status. As shown in Figure 3, the common types of thyroid dysfunction can be identified in a single sample from characteristic diagonal deviations in the normal free T4-TSH relationship.
The figure shows primary hypothyroidism due to target gland failure (high serum TSH, low free T4 : A), failure of TSH secretion (both low: B), autonomous or abnormally stimulated target gland function (high free T4, suppressed TSH: C), and primary excess of TSH, or thyroid hormone resistance (both high: D). Abnormal results that fall outside these areas suggest that some other factor has disturbed the TSH-free T4 relationship. (The link between components of other feedback systems can also be applied to the investigation of hypogonadism, glucocorticoid and mineralocorticoid abnormalities, hypoglycaemia, hypercalcaemia etc.)
The diagnostic validity of this relationship depends on a number of assumptions and limiting conditions (Table 6). It is notable that only the last three of these assumptions can be validated in the laboratory; the first three must be verified clinically. The first assumption, steady-state conditions, should be questioned when associated illness or medications perturb the pituitary-thyroid axis; the large difference between the half-lives of TSH (one hour) and T4 (one week) accounts for many transient non-diagnostic abnormalities, especially during critical illness. In several situations T3 as well as T4 is an important, or dominant, determinant of thyroid status. The relevance of serum free T3 measurement is summarized in Table 7. Note that serum free T3 is of little value in monitoring T3 treatment because of wide variations that depend on the interval between dosage and sampling.
Either serum TSH or free T4 can be used for initial screening and case finding, but TSH gives better first line testing than free T4, at slightly higher cost. Because thyroid gland abnormalities are 10–20 times more common than variations due to pituitary dysfunction, TSH changes can generally be regarded as giving an inverse reflection of thyroid status. An algorithm for the assessment of thyroid function based on initial measurement of TSH is shown in Figure 4. However, there are important situations in which TSH alone can give a misleading or ambiguous assessment of thyroid status (Table 8), despite the high negative predictive value of a normal serum TSH concentration in ruling out primary hypothyroidism or thyrotoxicosis.
Application of diagnostic strategy will differ depending on the test group, i.e. testing of untreated subjects in whom clinical features suggest thyroid dysfunction, screening or case finding in at risk groups, evaluation of the response to treatment, or assessment when associated illness or drug therapy are likely to complicate both clinical and laboratory assessment.
Assessment of untreated subjects now commonly begins with measurement of TSH alone, with free T4 and/or free T3 added only if TSH is abnormal, or if an abnormality of TSH secretion is suspected. According to this algorithm, free T4 should be measured to distinguish between overt and subclinical hypothyroidism when serum TSH is elevated, while a suppressed or subnormal TSH level should be followed by assay of both free T4 and free T3 to distinguish subclinical from overt thyrotoxicosis and to identify T3 toxicosis.
In patients with newly treated thyrotoxicosis, TSH may remain suppressed for several months after normalisation of serum free T4 and free T3; serious over-treatment may result if TSH alone is used for adjustment of antithyroid drug dosage. Furthermore, during drug treatment, thyrotoxicosis may persist due solely to T3 excess. Hence, reassessment of serum free T4 and free T3 levels is recommended after about 3–4 weeks drug treatment of thyrotoxicosis to allow appropriate dose adjustment. During long-term treatment, TSH generally gives a reliable guide to optimal drug dosage.
Similarly, during long-term replacement or suppressive therapy with T4, serum TSH is the best single index of appropriate dosage. Optimal replacement is generally reflected by a low-normal TSH value of about 1 mU/L, often with a slightly increased level of serum free T4. However, during the early phase of treatment of hypothyroidism, free T4 should also be measured because TSH may remain inappropriately elevated for many months after normalisation of T4 (Figure 5). In general, serum TSH should initially be checked after 2–3 months, but need only be checked annually after the first year of treatment.
During TSH suppressive therapy with T4, for example in the management of differentiated thyroid cancer, periodic assessment of free T4 and free T3 in addition to TSH, is appropriate to limit the degree of thyroid hormone excess, because over-treatment can have important adverse effects on the cardiovascular system and on bone density. In the treatment of hypothyroidism due to pituitary or hypothalamic disease, serum TSH is of no value in assessing T4 dosage, which should be judged from clinical response and serum free T4.
Especially in hospital practice, interpretation of thyroid function is often compromised by associated illness or by medications. There is a high prevalence of abnormal serum free T4 or TSH values in patients with acute medical or psychiatric illness, but when TSH and free T4 are considered together, as in Figure 3, few of these abnormalities indicate true thyroid dysfunction. Clinical assessment of thyroid status is difficult in the face of associated disease and some have advocated widespread laboratory testing. However, because of low specificity, opinion has moved away from routine testing during critical illness unless there is a clinical indication.
During any severe illness, one or more of the assumptions outlined in Table 6 may not be valid, for example when there are wide fluctuations from the steady state due to acute inhibition of TSH secretion or abnormally rapid T4 clearance. Serum free T4 estimates are prone to multiple method-dependent interferences, for example due to heparin (see above Figure 2). Measurement of total T4 continues to have a definite place in the assessment of potential difficulties with various free T4 estimates, particularly during critical illness, as demonstrated by a key study of severely ill euthyroid patients (Figure 6).14 Various free T4 assays gave widely discrepant method-dependent abnormal free T4 estimates, while the total T4 assay suggested that the majority of these patients remained euthyroid, with subnormal serum TSH, probably attributable to glucocorticoid treatment.14 Free T4 methods that were influenced by albumin binding of tracer tended to give subnormal estimates, while equilibrium dialysis and related methods appeared to be vulnerable to the heparin-NEFA artefact, with a trend towards high free T4 estimates.
There are important issues in selecting a panel of thyroid tests that will best serve a particular clinical population. For example, the multiple effects of severe illness on free T4 estimates may be of minor importance for a laboratory that serves predominantly ambulatory patients. A different assay profile will be required in a laboratory that needs to exclude thyroid dysfunction during critical illness or in pregnancy.
The multiple effects of medications on the pituitary-thyroid axis are summarized in Table 5. Effects on serum TSH are generally physiological, whereas most effects on free T4 estimates are methodological. Notably, lithium and iodine-rich compounds, in particular amiodarone, can cause thyroid function to become abnormal. Amiodarone is the most complex and difficult drug that affects thyroid status.15 There may be poor correlation between circulating thyroid hormone levels and clinical manifestations in amiodarone-induced thyroid dysfunction, because of interaction of this drug or its metabolites with thyroid hormone receptors. In iodine-replete regions the predominant amiodarone-induced thyroid abnormality is hypothyroidism, which is especially prevalent in those with associated autoimmune thyroiditis. Amiodarone causes two forms of thyrotoxicosis, one due directly to iodine excess and the other attributed to a unique type of thyroiditis. Benign euthyroid hyperthyroxinaemia occurs in up to 25% of treated patients, who show increased serum concentrations of free T4, with normal TSH and normal or low concentrations of free T3.
Lithium, used in the management of bipolar illness, has multiple effects on the pituitary-thyroid axis, the most important being inhibition of hormone release. Lithium can exacerbate, or may initiate autoimmune thyroid disease with development of goitre and eventual hypothyroidism; there are also some reports of lithium-induced thyrotoxicosis. Serum TSH, free T4 and free T3 assays generally give a true index of thyroid status during lithium treatment.
Phenytoin commonly results in subnormal serum total T4, with an apparent lowering of free T4, due predominantly to the dilution-related artefact described above, without the anticipated increase in TSH. Such findings are hard to distinguish from central hypothyroidism due to pituitary deficiency, but the major discrepancy is probably a methodological artefact related to underestimation of free T4 in diluted serum samples.16 The assessment of patients who have had pituitary surgery and are also taking phenytoin, remains very difficult.
In subclinical hypothyroidism, the presence of TPO antibodies indicates a 4–5-fold increase in the chance of developing overt hypothyroidism. The presence of this antibody also indicates an increased likelihood of postpartum thyroiditis or amiodarone-induced hypothyroidism. The finding of persistently positive thyrotropin receptor antibody (TRAb) is useful in indicating that apparent remission of Graves’ disease is unlikely to be sustained. TRAb measurement can also indicate the possibility of neonatal or intrauterine thyrotoxicosis in the infant of a mother with autoimmune thyroid disease and may also define the aetiology of atypical eye disease.
Serum thyroglobulin concentrations should always be interpreted in relation to the prevailing level of TSH, which is responsive to alterations in thyroid hormone dosage. In the long-term follow-up of differentiated thyroid cancer, an undetectable serum thyroglobulin concentration in the presence of high serum TSH indicates effective ablation of differentiated thyroid tissue, benign and malignant. Such a finding may justify less rigorous long-term T4-induced suppression of TSH. Thyroglobulin is undetectable in thyrotoxicosis factitia, and generally extremely high in subacute thyroiditis and in amiodarone-induced thyrotoxicosis due to thyroiditis.
There are still unsolved technical problems in the optimisation of thyroglobulin assays. Major issues relate to assay standardization, interference from endogenous anti-thyroglobulin antibodies and heterophilic antibodies, as well as inherent problems of assay sensitivity.
The diverse clinical presentations of thyroid dysfunction mandate laboratory requests from clinicians who may be unfamiliar with the interpretation of current assays, or with effects that interfere with these techniques. Comments from the laboratory can improve clinical response; the quality of this assistance depends on both the training and experience of the reporter and the available clinical information.
The competitive binding assays that are used for thyroid diagnosis were developed about 30 years ago with “in house” reagents. Optimal diagnostic technology was initially available only to specialists and results were often slow. Sophisticated standardised reagents and automated instrumentation (eg solid phase antibodies, magnetic separation systems, chemiluminescent detection systems) have now replaced these early methods; results are rapidly available to a wide range of practitioners. Non-specialist users of endocrine assays are most likely to benefit from laboratory-based assistance in the interpretation of results, but with assay automation, laboratorians have become more distant from the bedside. As clinicians receive less assistance, they provide progressively less relevant information and vice versa. Laboratorians, in turn may see results that are uninterpretable or ambiguous unless the relevant clinical information is available (Table 9). In these situations, the possibility of an assay artefact would need to be considered, but review of the clinical context often resolves an apparently anomalous result, so that assay validity is affirmed and unnecessary further investigation is avoided.
After review of the clinical context, the following steps are helpful in evaluating anomalous thyroid results:
Clinical feedback will remain a key aspect of quality assurance in laboratory testing. While assay precision or reproducibility can be evaluated from the laboratory, diagnostic accuracy requires clinical correlation.
The majority of thyroid diagnosis is now quite straightforward, but it is easy to underestimate problems that remain, since they represent only a small fraction of the total. Despite the elegance and ingenuity of current assay techniques, diagnostic inaccuracy of immunoassays still wastes substantial resources and studies of "cost effectiveness" do not evaluate the human costs that result from unnecessary further testing, false alarms and inappropriate management.19
Some thyroid disorders, notably papillary, follicular and medullary carcinomas require follow-up by tumour marker assays over many years. Changes in thyroglobulin and calcitonin assay methodology made without clinical consultation, in particularly changes made before the lower limits of detection have been critically defined, (i.e. apparent increase in sensitivity with loss of specificity), can give a false and at times disastrous impression of reactivation of disease.
Typical reference intervals for common thyroid assays are given in Table 10.