3.3.5 Causes and laboratory investigations of thyrotoxicosis

The term thyrotoxicosis identifies the clinical syndrome caused by elevated circulating thyroid hormones of all sources, while the term hyperthyroidism includes only the disorders due to an increased secretion of hormones by the thyroid gland. Hyperthyroidism is the most frequent cause of thyrotoxicosis. Destructive processes involving the thyroid gland that induce unregulated discharge of preformed thyroid hormones (destructive thyrotoxicosis) and circulating thyroid hormone of extrathyroidal origin (exogenous or extrathyroidal thyrotoxicosis) are less common causes of thyrotoxicosis. Although careful history taking and physical examination often allows a diagnosis of thyrotoxicosis to be made, laboratory confirmation by measurement of thyroid-stimulating hormone (TSH) and thyroid hormone is always needed. Once thyrotoxicosis is confirmed, laboratory testing and thyroid imaging are required to identify the cause of thyrotoxicosis.

From a clinical standpoint it is useful to classify the different causes of thyrotoxicosis according to their pathogenic mechanisms. A practical classification is outlined in Table 3.3.5.1 and distinguishes the forms of thyrotoxicosis in two broad syndromes of thyroidal and of nonthyroidal origin. The first group, generally the most frequent, can be further divided into forms associated with thyroid hormone hypersecretion (hyperthyroidism) and forms characterized by the release of preformed hormones, secondary to destructive processes (destructive thyrotoxicosis). The second group includes a heterogeneous group of rare disorders in which the thyroid gland is not the primary source of thyroid hormone. The most useful test in differentiating hyperthyroidism from the other causes of thyrotoxicosis is thyroidal radioactive iodine uptake (RAIU), which is elevated or high to normal in hyperthyroidism and very low in destructive thyrotoxicosis and in thyrotoxicosis of nonthyroidal origin.

Graves’ disease is the most frequent cause of hyperthyroidism, accounting for more than 70% of cases in iodine-sufficient countries, where its prevalence may be as high as 2% in women (1). In Graves’ disease, hyperthyroidism is caused by an autoimmune reaction to the thyroid, leading to the production of autoantibodies to the TSH receptor (TSHR autoantibodies) (2). These antibodies mimic the action of TSH in stimulating the TSH receptor on thyroid follicular cells (TSHR-S autoantibodies). Since there is no feedback of thyroid hormone on the production of TSHR autoantibodies, uncontrolled stimulation of the receptor causes growth of the thyroid gland and excessive production and release of thyroid hormone, ultimately leading to hyperthyroidism.

TSHR autoantibody epitopes comprise discontinuous sequences of the polypeptide chain that are contiguous in the folded protein under native conditions (reviewed by Rapoport et al. (2)). Evidence that the shed TSHR ectodomain (primarily the A subunit) is the primary antigen driving affinity maturation of TSHR-S autoantibodies is mounting (3). TSHR autoantibodies with TSH binding inhibiting but not TSHR stimulating activity have been demonstrated in serum of individuals with no evidence of autoimmune thyroid diseases (4).

As the effect of TSHR-S autoantibodies is exerted on all follicular cells, a diffusely enlarged thyroid is the hallmark of the disease, but in some cases thyroid nodules can develop as a consequence of either a long-standing disease or of a pre-existing nodular goitre. Graves’ ophthalmopathy and, rarely, pretibial myxoedema are other typical physical findings. The clinical manifestations of Graves’ ophthalmopathy have been described elsewhere in this text. On careful physical examination, 30–45% of patients with Graves’ disease have some signs of Graves’ ophthalmopathy (5) and, when studied with refined imaging techniques, suggestive findings can be observed in up to 70% of cases. When obviously present, Graves’ ophthalmopathy is extremely useful in supporting Graves’ disease as the cause of thyrotoxicosis. Pretibial myxoedema is a peculiar skin manifestation of Graves’ disease characterized by oedema, inflammation, and lymphocytic infiltration localized to the pretibial dermis. It is only rarely observed in Graves’ disease, but almost never observed without it. Therefore, while its presence can confirm the diagnosis, its absence is by no means an exclusion criterion.

Toxic adenoma and multinodular toxic goitre are frequent causes of hyperthyroidism, especially in iodine-deficient areas. Toxic adenomas are benign isolated follicular tumours that synthesize thyroid hormones independently of TSH stimulation. One or more autonomous adenomas may develop in an otherwise normal thyroid. Unregulated thyroid hormone secretion first suppresses pituitary TSH secretion and eventually leads to overt hyperthyroidism. Because of TSH suppression, the extranodular thyroid tissue becomes functionally quiescent and may undergo some degree of atrophy. The incidence of toxic adenoma has been estimated at about 5/100 000 per year in Sweden (6). Toxic adenomas account for about 10% of the cases of thyrotoxicosis and are more common in areas of mild or overt iodine deficiency than in iodine-sufficient areas (7). Toxic adenomas tend to occur in the aged population and are more frequent in women than in men. The natural history of toxic adenoma is characterized by a slow growth over many years and a change, through different stages, of its functional properties. In the early phases, the amount of secreted thyroid hormones is not sufficient to completely suppress TSH secretion (partial autonomy) and the function of the extranodular tissue. With further growth of the nodule, TSH suppression becomes complete, while circulating thyroid hormones are in the upper range of normal values (complete autonomy). Eventually, overt thyrotoxicosis ensues, with frankly elevated thyroid hormone levels. The rate of progression is quite slow and in a large follow-up study, only 14 out of 159 autonomous nodules developed overt hyperthyroidism in 6 years, the risk being higher for adenomas more than 3 cm in size. Somatic mutations of the TSHR gene, which cause amino acid changes leading to constitutive activation of the TSHR, are the cause of 20–80% of toxic adenomas, while the rate of mutations of the Gsα-protein range from 8% to 75% (8). Both kinds of mutations cause permanent activation of the TSHR intracellular signalling pathway in the absence of TSH.

Multinodular toxic goitre is also often detected in iodine-deficient countries, in which accounts for up to 60% of cases of thyrotoxicosis (7). Epidemiological studies have clearly shown that multinodular toxic goitre represents the long-term outcome of many long-standing endemic goitres and is more common in older people and in women (7). The prevalence of multinodular toxic goitre in iodine-deficient areas has been reduced by iodine prophylaxis. The same somatic activating mutations of the TSHR demonstrated in toxic adenoma have been observed in toxic multinodular goitre (9, 10). However, in many nodules neither TSHR nor Gsα-protein mutations have been observed (8). The natural history of multinodular toxic goitre is similar to that of toxic adenoma, with the slow formation of multiple autonomously functioning nodular areas in the setting of an overall nodular goitre. The only known mechanism inducing subclinical hyperthyroidism and overt thyrotoxicosis is the administration of excessive amounts of iodine (11). Because of the slow progression through several degrees of thyrotoxicosis and because of their advanced age, patients with multinodular toxic goitre may report few symptoms.

TSH secretion by a benign pituitary adenoma, which is characterized by a partial or complete loss of the feedback regulation by thyroid hormones (central hyperthyroidism), causes a sustained stimulation of the thyroid gland, with the subsequent development of goitre and hyperthyroidism. TSH-secreting adenomas are rare, account for less than 1% of all pituitary adenomas, and have an estimated prevalence of about one case per million. However, its prevalence seems to have increased, probably as a consequence of the introduction of ultrasensitive assays for TSH measurement that enable an earlier detection of an inappropriate secretion of TSH (caused by a TSH-secreting adenoma or by resistance to thyroid hormones) (see below). TSH-secreting adenoma has a similar frequency in the two sexes and can be an expression of the multiple endocrine neoplasia type 1 syndrome (MEN 1). Most of the TSH-secreting tumours are macroadenomas (>10 mm), present with extrasellar extension, and show invasiveness into the surrounding tissues. Cosecretion of growth hormone and/or prolactin occurs in 25% of cases. TSH-secreting pituitary carcinomas and ectopic TSH-secreting adenomas are exceedingly rare.

The severity of thyrotoxicosis is extremely variable, ranging from slight to very high elevations of thyroid hormones. As a consequence, the patients may report no or few symptoms, or present as overtly thyrotoxic. On physical examination, a diffuse goitre is often felt, but multinodular goitres can also be observed, especially in long-standing diseases. Visual field defects can be rarely observed with large adenomas. When growth hormone and prolactin are cosecreted, acromegaly and galactorrhoea are present. Since most TSH-secreting adenomas are macroadenomas, involvement of other pituitary hormones and/or of the optic chiasm are common.

Human chorionic gonadotropin (hCG) is secreted in large amounts by placental tissue in normal pregnancy and also by trophoblastic tumours (hydatidiform mole and choriocarcinoma). Due to its partial homology with TSH, hCG can act as a weak TSH agonist and, when present in large amounts in the bloodstream, can overstimulate the thyroid gland and induce hyperthyroidism.

Hyperthyroidism may ensue in patients with a hydatidiform mole or a choriocarcinoma as a consequence of the large quantities of hCG produced by the tumour. Although thyrotoxicosis is common in trophoblastic tumours, clinically overt thyrotoxicosis is observed in a minority (10%) of patients in whom extraordinarily high levels of hCG (>3 000 000 IU/l) are found. The routine use of ultrasonography during pregnancy has led to earlier diagnosis of hydatidiform mole, when the tumour mass is smaller and the thyrotoxicosis less likely.

Hyperemesis gravidarum is a poorly understood complication of early pregnancy, characterized by prominent nausea and vomiting, weight loss, ketosis, and electrolyte abnormalities. It occurs in 1.5% of pregnancies. In 25–75% of pregnancies, increased levels of thyroid hormones have been reported, which correlate with serum hCG concentrations. A clinically evident thyrotoxicosis (termed gestational hyperthyroidism) is rare (12).

One family with an inherited hyperthyroidism that is exclusively associated with pregnancy has been reported (13). Hyperthyroidism was caused by a mutation in the TSHR gene causing increased responsiveness to hCG. Hyperthyroidism only manifests during pregnancy and recurs every time an affected woman becomes pregnant. The genetic defect does not have an effect in nonpregnant women and in men.

TSHR-S autoantibodies in the serum of mothers with Graves’ disease can cross the placenta and cause fetal and neonatal hyperthyroidism through direct stimulation of the fetal thyroid. The disease can be very severe and is characterized by tachycardia, jaundice, heart failure, and failure to thrive. A goitre is usually present. The disease is transient and resolves within 3 months after birth, when TSHR-S autoantibodies disappear.

After the first report of congenital hyperthyroidism caused by a germline de novo activating mutation of the TSHR gene (14), only a few cases of nonautoimmune neonatal hyperthyroidism, resulting from mutations of the TSHR, inherited as an autosomal dominant trait or arisen de novo, have been described (15, 16). The diagnosis should be suspected when a neonate presents with severe hyperthyroidism and goitre and the mother has no history of Graves’ disease.

Familial hyperthyroidism was described in 1994 (17). Autosomal dominant activating germline mutations of the TSHR gene are the cause of hyperthyroidism in this case, but since the effect of the mutation is mild, hyperthyroidism and goitre only develops in adults. Patients with the same mutation of the TSHR gene may present with wide phenotypic variability, with respect to the age of onset, severity of hyperthyroidism, and goitre (18). However, germline mutations of the TSHR gene have been reported to be uncommon in juvenile thyrotoxicosis (19). Cases of congenital hyperthyroidism from mutations of the Gsα-protein have also been reported and are associated with the McCune–Albright syndrome.

Subacute thyroiditis (also called granulomatous, giant cell, or de Quervain’s thyroiditis) is an inflammatory disorder of the thyroid, most probably of viral origin, which may last for weeks or months, predominates in females, and peaks in the fourth and fifth decades (see Chapter 3.2.7).

Painless thyroiditis (sporadic or silent thyroiditis), an autoimmune thyroid disorder, is characterized by a transient phase of thyrotoxicosis, similar to subacute thyroiditis, in the absence of neck pain and general symptoms. A goitre can develop or a pre-existing goitre can enlarge. At histology, a lymphocytic chronic infiltration closely resembling that of Hashimoto’s thyroiditis is usually found. The incidence of painless thyroiditis has been reported to vary from less than 5% to 23% of all cases of thyrotoxicosis and is more prevalent in the third to the sixth decades of life. The female to male ratio is 2:1. In some cases, painless thyroiditis is precipitated by radiotherapy, iodine, and treatments with lithium, interleukin-2, interferon-α, and granulocyte colony stimulating factor. HLA-DR3 and HLA-DR5 are more common in patients with painless thyroiditis. Circulating thyroid autoantibodies are detected in the vast majority of cases. The thyrotoxicosis is usually mild and self-limited and can be followed by transient hypothyroidism. Progression to spontaneous permanent hypothyroidism is observed in as many as 20% of patients in the long-term follow-up. Sometimes clinical findings suggestive of painless thyroiditis are found in the presence of positive TSHR autoantibodies and high uptake, with goitre and rapid evolution in hypothyroidism. These cases may be classified as a variant of Graves’ disease with predominant cytotoxic aspects, quickly leading to a clinical picture of Hashimoto’s thyroiditis. Painless thyroiditis is seldom associated with Graves’ ophthalmopathy.

Postpartum thyroiditis (PPT) is the painless thyroiditis that occurs in susceptible women within 12 months after delivery. It develops in 5–10% of pregnancies and is more common in women with type 1 diabetes. The presence of serum thyroid peroxidase (TPO) and thyroglobulin autoantibodies before or during the onset of the disease and the association with some HLA haplotypes (HLA-A26, -BW46, -DR3, -DR4, and -DR5) support the autoimmune pathogenesis of PPT (see Chapter 3.4.6).

Rarely, destructive thyrotoxicosis can be precipitated by anterior neck injuries. Thyrotoxic crises following thyroid surgery, a frequent complication in the early days of thyroid surgery, have now become extremely rare with the optimal use of antithyroid drugs and with the refinement of surgical procedures. Thyrotoxicosis may transiently worsen or recur in patients with Graves’ disease, toxic adenomas, and multinodular toxic goitre who are treated with radio-iodine. Two mechanisms are responsible for this phenomenon: (1) ongoing thyroid hyperfunction before radio-iodine fully takes effect and (2) radiation-induced thyroid destruction.

Iodine deficiency increases thyrocyte proliferation and mutation rates, inducing the development of multifocal autonomous growth and cell clones harbouring activating mutations of the TSHR. Some of these nodules maintain the ability to store iodine and can become autonomous causing thyrotoxicosis after iodine excess or even iodine supplementation. Because a pre-existing thyroid autonomy is required for the development of these disorders, iodine-induced thyrotoxicosis is far more prevalent in elderly patients and in areas of iodine deficiency (Box 3.3.5.1). A transient, unavoidable increase in the prevalence of mild thyrotoxicosis has been well documented in iodine-deficient countries soon after carrying out iodine supplementation programmes with physiological doses of iodine. Another predisposing condition is euthyroid or latent Graves’ disease or Graves’ disease in remission. In individual thyrotoxic patients, iodine contamination may be caused by a variety of medications and diagnostics, including lipid-soluble contrast media, disinfectants, and drugs, and some foods containing large amounts of iodine (Box 3.3.5.2).

Among drugs, the antiarrhythmic amiodarone deserves a special mention because of the dual mechanism by which it can cause thyrotoxicosis. One tablet of 200 mg amiodarone contains approximately 75 mg organic iodide and will release about 8 mg free iodine, a tremendous amount when compared with the daily recommended dose of 200 μg. In iodine-deficient regions, where some elderly people have nodular thyroid autonomy, and in patients with euthyroid Graves’ disease, this amount of iodine can precipitate hyperthyroidism (type 1 amiodarone-induced thyrotoxicosis) (20). However, in patients with no underlying thyroid disease, amiodarone can be directly cytotoxic to thyroid follicular cells in vitro (21) and can cause a form of subacute thyroiditis in vivo with the release of preformed hormones (type 2 amiodarone-induced thyrotoxicosis). Distinction between the two forms is useful for the appropriate treatment (20). Unfortunately, many patients present with a mixed form of thyrotoxicosis. The incidence of amiodarone-induced thyrotoxicosis ranges from 1% to 32%, being higher in regions with low iodine intake than in iodine-sufficient areas (20).

The term thyrotoxicosis factitia describes the voluntary excessive ingestion of thyroid hormone preparations with the purpose of mimicking thyrotoxicosis (Latin factitius = fake). However, the term has been widely applied to all forms of thyrotoxicosis due to the ingestion of thyroid hormones. True thyrotoxicosis factitia is most often observed in women with psychiatric disturbances who have access to thyroid medication, e.g. health professionals or people with relatives treated with thyroid hormones. Very often thyroid hormone is taken for weight reduction or to receive medical attention. Denial of thyroid hormone consumption may be extreme in these patients and the diagnosis is rarely obtained at history taking. Accidental or suicidal ingestion of large amounts of thyroid hormone has been also described, but this can usually be diagnosed by history alone. Sometimes, thyroid hormone is inadvertently taken as a component of ‘herbal’ or ‘alternative’ medications, usually for weight reduction. Finally, accidental grinding of cattle thyroids in hamburger meat has been reported as the cause of an outbreak of thyrotoxicosis among hamburger consumers in the USA (22).

Struma ovarii is a teratoma of the ovary that differentiates into thyroid cells. It comprises about 3% of ovarian teratomas, is bilateral in 10% of cases, and malignant in 10%. Thyrotoxicosis occurs in 10% of cases.

Differentiated thyroid carcinoma, even when metastatic and with large tumour burdens, does not usually produce relevant amounts of thyroid hormones. Very rarely, however, thyroid carcinomas of the follicular histotype, extensively metastatic to bone, may cause thyrotoxicosis. Coexistent TSHR-S autoantibodies are an extremely rare cause of thyrotoxicosis in patients with metastatic thyroid cancer.

The mainstay of the diagnosis of thyrotoxicosis is measurement of serum thyroid hormones and TSH. Because of its high correlation with free thyroxine (T₄), TSH is the single most useful test in confirming the presence of thyrotoxicosis (23). The current immunoassays are very sensitive and can measure TSH levels well below the normal range, with a functional sensitivity (TSH concentration at which the response of the assay has a coefficient of variation of 20%) of less than 0.02 mU/l. Since pituitary TSH secretion is tightly down-regulated by thyroid hormone level, TSH is undetectable in most cases of thyrotoxicosis. The only remarkable exceptions are TSH-secreting adenomas, in which a high or inappropriately normal TSH level is found in spite of overt thyrotoxicosis. Because of the sensitivity of the assay, low (less than 0.4 mU/l) but detectable TSH levels can be found. These levels are encountered in subclinical thyrotoxicosis and in other conditions, such as nonthyroidal illnesses and endogenous or exogenous corticosteroid excess (Box 3.3.5.3). TSH is a heterogeneous molecule and different TSH isoforms circulate in the blood and are present in pituitary extracts used for assay standardization (24). Although current methods have eliminated cross-reactivity with other glycoprotein hormones, they may detect different epitopes of abnormal TSH isoforms secreted by some euthyroid individuals and some patients with pituitary diseases. Very rarely, the presence in the serum of antimouse immunoglobulin antibodies may interfere in the TSH assay, causing falsely elevated TSH levels.

Measurement of serum thyroid hormone levels is mandatory in all patients with suspected thyrotoxicosis for a proper evaluation of a low TSH level and for an estimation of the severity of the disease. The active form of the hormones in serum is the very small amount of freely circulating T₄ and triiodothyronine (T₃), which can enter cells, interacting with the specific receptors. Total T₄ and total T₃ can be easily and inexpensively measured by radioimmunoassay, but their levels are influenced by the levels of binding protein levels, which vary in healthy people and may change in several conditions (25). Thus total thyroid hormone levels may not parallel those of free thyroid hormones, and their measurement is nowadays considered less useful in the evaluation of thyrotoxicosis (24, 26). Measurements of free thyroid hormone levels, although not completely exempt from flaws, are more satisfactory, since they provide a more accurate measurement of the active hormone (27).

In iodine-sufficient countries a single measurement of free T₄ is sufficient to confirm or reject the suspicion of thyrotoxicosis and, after TSH measurement, this is the test most often used in North America for thyroid function screening (28). In contrast, in iodine-deficient countries, a significant proportion of hyperthyroid patients (up to 12%) may have elevated free T₃ and normal free T₄ levels, a condition termed T₃ toxicosis. Conversely, free T₄ can be falsely elevated in conditions causing reduced peripheral conversion of T₄ to T₃, such as amiodarone treatment. In our practice, when thyrotoxicosis is suspected, we initially assess both free T₄ and free T₃ levels together with TSH, with little additional expense, in order to obtain a complete assessment of the thyroid function status.

Occasionally, a low or undetectable TSH level and normal free thyroid hormone levels are detected at routine thyroid function testing or in patients complaining of mild thyrotoxic symptoms. This condition is termed ‘subclinical thyrotoxicosis’ or ‘subclinical hyperthyroidism’ (29). This name is based on the recognition that even subtle variations in thyroid hormone levels can have a large effect on TSH secretion. In this respect, a low TSH level would be the first manifestation of a pending or subtle hyperthyroidism. However, the definition is somewhat unsatisfactory, since a subnormal TSH level can also be found in many nonthyroidal conditions, in the absence of true thyrotoxicosis, e.g. corticosteroid treatment, psychiatric and severe nonthyroidal illnesses, pregnancy, and others (Box 3.3.5.3). When biochemical findings suggest subclinical thyrotoxicosis, all of these conditions should be ruled out. Because of their high sensitivity, TSH tests now available can distinguish between partially (0.1–0.4 mU/l) and completely suppressed (<0.1 mU/l) TSH values.

In many cases, history and physical examination can readily identify the cause of thyrotoxicosis. However, in many other situations, a careful differential diagnosis is needed in order to establish an aetiological diagnosis. Classically, RAIU has represented a mainstay of the differential diagnosis of thyrotoxicosis. RAIU is easily performed by administering a minimal (tracer) dose of radioactive iodine and then measuring the per cent of administered radioactivity accumulated in the neck. In iodine-sufficient countries the upper limit of RAIU, 24 h after the administration of the tracer, is around 25%, while it may reach 40% in areas with mild to moderate iodine deficiency (30). Whenever excessive active formation of thyroid hormone takes place in the thyroid gland, RAIU is increased, since the thyroidal machinery for iodine trapping and organification is activated. Therefore, a high RAIU readily identifies true hyperthyroidism (i.e. with thyroid hyperfunction). In contrast, thyrotoxicosis with a low RAIU indicates either thyroidal destruction, with release of preformed hormone, or an extrathyroidal source of thyroid hormone. In thyroid destruction, the damaged follicular cells transiently lose their capability of iodine trapping, while when exogenous hormones are administered in excess, the suppression of the pituitary secretion of TSH causes shutting-off of the trapping capacity of follicular cells. The only exception to this rule is iodine-induced thyrotoxicosis, in which a low RAIU can be observed because of dilution of the tracer dose in the large body pool of iodine, in spite of true hyperthyroidism.

Nowadays, RAIU is not universally performed in the initial assessment of a thyrotoxic patient and a vast array of laboratory and imaging techniques have provided excellent tools for accurately identifying the cause of thyrotoxicosis without the information provided by RAIU. RAIU is still useful in difficult cases to broadly define forms of thyrotoxicosis according to their pathogenesis in order to proceed methodically with adjunctive diagnostic tools.

Since the cause of Graves’ disease hyperthyroidism is uncontrolled thyroid gland stimulation by circulating TSHR autoantibodies, their detection in the serum of thyrotoxic patients is particularly useful in establishing the diagnosis of Graves’ disease. Serum TSHR-S autoantibodies can be measured by different methods (2, 31). They were originally detected with in vivo bioassays, and later by in vitro systems. The most common tests assess the displacement of labelled TSH or TSHR autoantibodies from the TSHR by the immunoglobulin fraction of patients’ sera. These methods are termed TSH-binding inhibition (TBI) tests and do not distinguish between TSHR-S autoantibodies and TSHR blocking (TSHR-B) autoantibodies, which can also be detected in thyroid autoimmune disorders (2). TSHR-S autoantibodies can be tested in cellular systems carrying a functional TSH receptor, detecting the release of cAMP in the culture medium upon challenge with serum or purified immunoglobulins (TSHR-S autoantibodies assay) (2, 32). In a modification of the assay, TSHR-B autoantibodies can be detected as well (33). Since TSHR-S autoantibodies are properly the cause of hyperthyroidism in Graves’ disease, their assay should be considered the gold standard in the diagnosis of Graves’ disease. Unfortunately, the assay is quite expensive and requires cell-culture capabilities, making it available only to research centres. For clinical purposes, the TBI assays are most often used. By using the latest generation of assays, positive TBI tests are found in 75–95% of patients, with a high specificity (99%) (34). A TBI test is strictly needed in a minority of cases of Graves’ disease in which the clinical picture is unclear, e.g. in the differential diagnosis of hyperemesis gravidarum, in the nodular variants of Graves’ disease that must be differentiated from toxic nodular goitre, in patients with exophthalmos without thyrotoxicosis (euthyroid Graves’ disease) (5), and in pregnant women with Graves’ disease. The presence of high levels of TSHR autoantibodies at the end of antithyroid drug therapy has a high positive predictive value and specificity for relapse of hyperthyroidism but a low negative predictive value and sensitivity (35).

Whereas TSHR-S autoantibodies interact mainly with the N-terminal components of the ectodomain, TSHR-B autoantibodies interact to a greater extent with the C-terminus and to a lesser extent with the N-terminus and the midregion of the TSHR (2, 36). Accordingly, immunization of mice with the N-terminal component of the TSHR or with the TSHR holoreceptor generated preferentially TSHR-S and TSHR-B autoantibodies, respectively (37). Whereas the epitope(s) for TSHR-S autoantibodies are partially sterically hindered on the holoreceptor by the plasma membrane, those for TSHR-B autoantibodies are fully accessible (38).

TPO autoantibodies can be found by commercial radioimmunoassays in up to 90% of patients with untreated Graves’ disease (39), while Tg autoantibodies are less frequently positive, in about 50–80% of patients. Both autoantibodies, however, are also present in other forms of thyroid autoimmune disorders, some of which may cause thyrotoxicosis, such as postpartum thyroiditis and silent thyroiditis. A relatively high percentage (up to 25%) of positive thyroid autoantibodies tests is also found in normal individuals, especially women (40). Thus, TPO and Tg autoantibody tests do not establish the diagnosis of Graves’ disease as the cause of thyrotoxicosis, but may be useful as complementary tests in confirming the presence of thyroid autoimmunity. The view that the binding of Tg autoantibodies from patients with autoimmune thyroid diseases is restricted to a few epitopic regions on Tg has been recently confirmed (41, 42).

The finding of autoantibodies cross-reacting with thyroglobulin and TPO in patients with autoimmune thyroid diseases, which suggested a role for cross-reactivity of the B-cell response to Tg and TPO (43), has not been confirmed (44). Thyroid autoantibody production requires the presence of thyroid autoantigens, as indicated by their disappearance after total thyroid ablation obtained by thyroidectomy plus ¹³¹I treatment (45).

Megalin (gp330) binds Tg with high affinity and participates in its transcytosis across thyroid cells (46). Autoantibodies to megalin were detected in 50% of patients with chronic autoimmune thyroiditis and in some patients with Graves’ disease and thyroid carcinoma, but not in normal individuals (47). A role of the sodium-iodide symporter as autoantigen in thyroid autoimmunity has been proposed by some authors but excluded by others (48, 49).

In untreated hyperthyroid Graves’ disease patients, a high value of RAIU at 24 h is always found. As a distinctive feature, in some cases the 3- or 6-h value can be even higher than the 24–h value, as an expression of an extremely high iodine turnover. The test is very useful for ruling out transient thyrotoxicosis due to Hashitoxicosis or painless or subacute thyroiditis, factitious thyrotoxicosis, and type 2 amiodarone-induced thyrotoxicosis (20).

Thyroid imaging with radioisotopes can be performed with radio-iodine at the time RAIU is carried out or by using ^99mTc pertechnetate. Thyroid scanning in Graves’ disease is useful only when coexisting nodules are detected by palpation and their functional status needs to be evaluated.

The ultrasonographic appearance of the thyroid gland undergoes typical changes during Graves’ disease hyperthyroidism. Because of the reduction in the colloid content and of the lymphocytic infiltrate, the gland becomes diffusely hypoechoic (50). A similar pattern is also observed in chronic goitrous thyroiditis and, when diffuse, indicates the presence of thyroid autoimmunity (51). Therefore thyroid ultrasound scanning can be useful in confirming the suspicion of thyroid autoimmunity, during the evaluation of thyrotoxicosis. Marked hypoechogenicity at the end of antithyroid drug therapy may predict recurrence of thyrotoxicosis (52). As an adjunctive value, thyroid ultrasound scanning also allows an accurate measurement of the goitre size (53), information that is important in the choice of the most appropriate treatment. Finally thyroid ultrasound scanning accurately distinguishes true thyroid nodules from the lobulations that can be occasionally felt at palpation in Graves’ disease glands. The information provided by thyroid ultrasound examination is therefore quite useful in the initial evaluation of the Graves’ disease patients, although not strictly needed from a diagnostic standpoint.

The measurement of blood flow to the thyroid gland by colour flow Doppler ultrasonography has been also used in Graves’ disease patients. In untreated Graves’ disease, the colour flow Doppler pattern is characterized by markedly increased signals with a patchy distribution (54, 55). Colour flow Doppler studies of the thyroid gland can therefore be used in the same way as RAIU in distinguishing Graves’ disease from other forms of thyrotoxicosis, e.g. amiodarone-induced destructive thyrotoxicosis (56) or subacute thyroiditis and, possibly, painless thyroiditis.

When a single nodule is palpated in the thyroid of a patient being evaluated for thyrotoxicosis, the presence of a toxic adenoma must always be suspected. In confirming the diagnosis of thyrotoxicosis, it is important to measure both free T₄ and free T₃ levels, since T₃ toxicosis is distinctly frequent in toxic adenomas (26). A blunted nocturnal TSH surge may be an early indicator of progression to hyperthyroidism in patients who are still euthyroid on baseline testing (57). ^99mTc or radio-iodine thyroid scanning is extremely helpful in confirming the diagnosis, yielding typical findings. The nodule will appear ‘warm’, with the extranodular thyroid tissue clearly visible, when partial autonomy is present. In this case, parallel thyroid function tests will show a low but detectable TSH, and thyroid hormone levels in the upper part of the normal range. Only the nodule is visible on the scan when TSH is completely suppressed, e.g. in case of complete autonomy or of overt thyrotoxicosis. Ultrasound scanning of the neck provides no direct diagnostic information on the functional property of the nodule, but it is useful in detecting coexisting cold nodules and accurately defining the size of the nodule. Preliminary reports have shown a distinctive colour flow Doppler pattern in autonomously functioning thyroid nodules, characterized by an increased blood flow in the nodular tissue, in good correlation with radionuclide scans. However, the technique is not able to distinguish benign from malignant nodules (58) and is therefore of limited value. Ultrasound elastography has showed high sensitivity and specificity in differentiating benign from malignant thyroid nodules (59). Fine-needle aspiration biopsy is recommended in the initial evaluation of every solitary thyroid nodule, but often provides undetermined (follicular) neoplasm in hot nodules. The risk of malignancy in hot nodules is extremely low, although occasionally reported. Therefore, in the presence of a low TSH, fine-needle aspiration is only needed when coexisting nodules detected by palpation or ultrasonography are cold at radionuclide scanning.

Further imaging, such as neck radiographs, barium swallow, and CT scans, may be needed in selected patients with large nodules in order to evaluate the presence of significant tracheal and/or oesophageal compression. It is important to remember that CT scan, when done with this purpose, should always be performed without the administration of iodinated contrast media, since these may worsen thyrotoxicosis or precipitate it in the presence of partially autonomous nodules.

The same range of thyroid function test alterations described in toxic adenomas can be observed in toxic multinodular goitre, from a subnormal TSH level to an undetectable TSH level with frankly elevated thyroid hormone levels. The diagnosis of toxic multinodular goitre can often be suspected on history and physical findings. Thyroid radionuclide scanning is quite useful in identifying and mapping autonomous nodules and distinguishing them from other coexistent cold nodules. Scanning is also useful as an adjunct to TSHR autoantibody measurement in distinguishing true toxic multinodular goitre from Graves’ disease hyperthyroidism which develops on a pre-existing nontoxic multinodular goitre. RAIU is always elevated, unless iodine overload is present, but sometimes is not necessary for establishing the diagnosis. Thyroid ultrasonography is also useful to measure goitre size and, in association with radionuclide scanning images, to identify cold nodules amenable to fine-needle aspiration biopsy. Fine-needle biopsy should be performed in any palpable dominant nodule that is cold at scan.

The presence of a TSH-secreting adenoma should be suspected when a detectable TSH level in the presence of clearly elevated circulating thyroid hormone levels (inappropriate secretion of TSH) is found. The first step in the evaluation of inappropriate secretion of TSH is making sure that artefacts in the measurement of TSH or thyroid hormone levels are not the cause of the laboratory findings. Falsely elevated TSH levels can be observed occasionally when heterophilic antibodies are present in the patient’s serum. These antibodies are antimouse immunoglobulins that bind both the solid-phase and the labelled mouse antibodies employed in most TSH immunoradiometric assays, causing bridging between the two and therefore mimicking the presence of TSH (24). The problem can be overcome by incubating the patient’s serum with mouse immunoglobulins before TSH testing, thus precipitating the heterophilic antibody (60). The most recent TSH commercial assays contain these antibodies in their incubation buffers, making this problem quite rare nowadays. Falsely elevated free T₄ and free T₃ levels must also be excluded in the preliminary evaluation of suspected inappropriate secretion of TSH. Mild spurious elevations of free T₄ and free T₃ can occasionally be found in the presence of thyroid hormone autoantibodies, genetic or drug-induced alterations of thyroxine-binding globulin, and in nonthyroidal illnesses (27). The two-step methods for measurement of free thyroid hormones may be useful to rule out these conditions. Once these artefacts have been excluded, extensive laboratory testing is required to clarify the cause of inappropriate secretion of TSH. True inappropriate secretion of TSH is observed in two conditions: (1) TSH-secreting pituitary adenoma and (2) resistance to thyroid hormone. In theory, only TSH-secreting adenomas cause true and symptomatic hyperthyroidism and therefore should be considered in the differential diagnosis (see below).

The syndrome of resistance to thyroid hormone is caused by a relative insensitivity of the thyroid hormone receptor to the action of its ligand. Therefore, higher thyroid hormone concentrations are needed to down-regulate TSH secretion. In most patients, the defect is due to mutations of thyroid hormone β-receptor gene and is inherited in a dominant autosomal fashion (61). Less common are de novo mutations of thyroid hormone β-receptor gene (22%) or resistance in the absence of thyroid hormone β-receptor gene mutations (7%). As a consequence of the defect, the pituitary set point for TSH suppression is set at a higher level of circulating thyroid hormone, i.e. a higher level of thyroid hormone is required for TSH suppression. Since the same abnormality is present at the tissue level, higher thyroid hormone levels are also required to exert normal peripheral thyroid hormone actions and the patient is therefore only biochemically hyperthyroid. The clinical picture is, however, more complex because some patients with resistance to thyroid hormone present with mild symptoms suggestive of thyrotoxicosis, especially tachycardia. As an explanation, the existence of a distinct syndrome of selective pituitary resistance to thyroid hormone has been proposed, in which normal tissue effects of thyroid hormone are present, in spite of insufficient TSH suppression, causing true peripheral thyrotoxicosis (62). The observation that similar thyroid hormone receptor abnormalities have been found in patients with the generalized and pituitary form of the disease and the absence of clinical features clearly distinguishing the two disorders, however, challenges this view and rather suggests that resistance to thyroid hormone encompasses a spectrum of manifestations, due to variable expression of the defect in different tissues (62). As in TSH-secreting adenomas, sustained TSH stimulation leads to the development of goitre, mostly diffuse, but some distinctive clinical features can be found, such as skeletal abnormalities and hearing defects (61).

When the suspicion of inappropriate secretion of TSH is confirmed, the presence of a TSH-secreting adenoma must be differentiated from resistance to thyroid hormone. Because of the overlapping clinical presentation and because no single test accurately allows clear-cut differentiation between the two conditions, extensive baseline and dynamic laboratory testing is usually required.

A number of tests have been used to confirm the presence of thyrotoxicosis at the tissue level. Nocturnal heart rate, Achilles reflexometry, and other indirect measures of peripheral thyroid hormone actions have been used for this purpose (63), but are cumbersome and not sensitive enough, especially when only mild elevations of thyroid hormones are present. The presence of thyrotoxicosis at the tissue level can be documented by measuring a variety of biochemical markers of thyrotoxicosis such as sex hormone-binding globulin, alkaline phosphatase, cholesterol, and creatine phosphokinase (63). Unfortunately, these parameters are quite nonspecific and may be elevated (or reduced) in a number of other conditions. At variance with normal pituitary, TSH-secreting adenomas secrete the α-subunit of TSH in molar excess with respect to TSH. A serum α-subunit/TSH ratio of more than 1 is observed in approximately 90% of patients with TSH-secreting adenoma (64). High ratios can also be observed in postmenopausal women and even in normal individuals, making this test alone unable to establish the diagnosis. Growth hormone, insulin growth factor-I, and prolactin serum measurements are useful, since about 30% of TSH-secreting adenomas cosecrete growth hormone and prolactin.

Dynamic testing aims at the demonstration of the unresponsiveness of TSH-secreting adenomas to normal stimuli. In most (92%) TSH-secreting tumours, the TSH level fails to increase in response to a standard thyrotropin-releasing hormone (TRH) stimulation test, while a normal or increased response is observed in resistance to thyroid hormone (64). A diagnostic protocol to test the response of pituitary TSH to exogenous T₃ is also used. T₃ is administered orally and the dose is increased every 3 days, starting from 50 μg/daily and increasing to 200 μg/daily (65). Before every increase, basal and TRH-stimulated TSH is measured, together with peripheral markers of thyroid hormone action (65). In TSH-secreting adenomas, only partial or no suppression of TSH secretion is observed, while complete or partial suppression is observed in resistance to thyroid hormone. Alternatively, 80–100 μg T₃ can be administered for 8–10 days. Using this protocol, complete TSH suppression is obtained in normal individuals, while no changes or slight reduction in TSH levels are observed in all patients with TSH-secreting adenomas. In contrast, clear-cut reductions of TSH levels are observed in resistance to thyroid hormone patients (64). The test is contraindicated in elderly patients and in patients with arrhythmias and/or coronary artery disease. Available tests for the differential diagnosis of the syndrome of inappropriate secretion of TSH are given in Table 3.3.5.2.

Pituitary imaging is very important in confirming the diagnosis. Ninety per cent of TSH-secreting adenomas are more than 1 cm in diameter at diagnosis and therefore easily detected at pituitary MRI scanning (64). In addition, radiolabelled-octreotide pituitary scintigraphy can be useful in detecting small tumours (66), although it can be positive in other types of pituitary tumours.

The presence of a trophoblastic tumour should be suspected when thyrotoxicosis is found in an amenorrhoeic woman, especially when a palpable abdominal mass is found. The diagnosis is readily confirmed by the finding of extremely high circulating hCG levels and of a pelvic mass at ultrasonography. In trophoblastic tumours, serum hCG levels usually exceed 200 U/ml, whereas the peak concentration for normal pregnant women is 100 U/ml.

The diagnosis of thyrotoxicosis during hyperemesis gravidarum can be particularly challenging and it is one of exclusion. Because of weight loss and malnutrition, free T₃ levels may be disproportionately low or even normal in comparison with free T₄ levels, due to a reduced peripheral conversion of T₄ to T₃. The TSH level is often low during early normal pregnancy, but seldom undetectable, as it is in true thyrotoxicosis. The only distinctive laboratory feature is an inappropriately high hCG level, but large overlap with normal pregnancies exists. Therefore, the diagnosis of thyrotoxicosis in hyperemesis gravidarum relies mainly on the clinical picture and on appropriate exclusion of other more common forms of hyperthyroidism by appropriate testing. It is important to remember that RAIU is absolutely contraindicated in pregnancy, as is any other in vivo radioisotopic procedure.

Mothers with a past or present history of Graves’ disease should be carefully monitored throughout pregnancy. Fetuses of mothers who have been previously treated with radio-iodine or surgery are at high risk because they lack the protective effect of antithyroid drugs administered to the mother. The presence of a fetal heart rate above 160 beats/min, in the absence of other fetal abnormalities, is suggestive of fetal hyperthyroidism. The persistence of high levels of TSHR autoantibodies in the maternal serum by the end of pregnancy, when the transplacental passage is maximal, is a predictor of hyperthyroidism in the neonate. Fetal cord blood sampling has been performed to diagnose fetal hyperthyroidism, but it is a risky procedure and is not generally recommended. In contrast, it is very useful to test neonatal cord blood at the time of delivery for thyroid function tests and TSHR autoantibodies. When the mother has been treated with high-dose antithyroid drugs, the neonate should be retested 10 days after birth, since transplacental passage of methimazole or propylthiouracil may initially mask hyperthyroidism.

Neonatal hyperthyroidism in the absence of a maternal history of Graves’ disease and with negative TSHR autoantibodies should raise the suspicion of nonautoimmune congenital hyperthyroidism. Familiar hyperthyroidism should be suspected when relatives are affected and serum TSHR autoantibodies are absent. In both types of hyperthyroidism, sequencing of the TSHR gene is required to confirm the diagnosis.

Excessive iodine consumption should be always suspected when hyperthyroidism abruptly appears in a patient with a history of nodular thyroid disease. A careful history often identifies the source of iodine and all patients should be asked about recent consumption of any of the compounds listed in Box 3.3.5.2. With the exception of type 2 amiodarone-induced thyrotoxicosis, RAIU is usually low in thyrotoxic patients with heavy iodine contamination, but it is almost never suppressed, a feature that allows distinction from subacute and painless thyroiditis. The iodine/creatinine urinary ratio is, however, the gold standard in confirming iodine contamination and will be high in all cases.

When a history of taking amiodarone is elicited in a thyrotoxic patient, further testing is required to distinguish between the type 1 and type 2 forms of amiodarone-induced thyrotoxicosis, since treatment may be radically different (20). Type 1 (nondestructive) amiodarone-induced thyrotoxicosis differs little from other forms of iodine-induced thyrotoxicosis and an underlying thyroid disease such as Graves’ disease or nodular thyroid disease is usually detected with the appropriate diagnostic tools. Accordingly, RAIU is usually low, but not suppressed and may be normal or increased. In contrast, in type 2 (destructive) amiodarone-induced thyrotoxicosis, RAIU is always low or suppressed and often no clear underlying thyroid disorder can be identified. High circulating interleukin-6 levels have been proposed as a useful marker of thyroid tissue destruction. In type 2 amiodarone-induced thyrotoxicosis, colour flow Doppler ultrasonography shows a distinctive absence of vascularization in the gland (56).

Classically, subacute, painless, and postpartum thyroiditis are characterized by a low (<1%) RAIU during the thyrotoxic phase. This test alone, in the presence of a suggestive clinical presentation allows the diagnosis in almost all cases. Serum T₄ concentration is disproportionately elevated compared with T₃ concentration, reflecting the preferential release of T₄ from the injured thyroid. In subacute thyroiditis, a very high (always >50 mm/h and often >100 mm/h) ESR is a distinctive diagnostic feature. C-reactive protein is also elevated and a mild leucocytosis is often observed. High titres of TPO and Tg autoantibodies are usually found in postpartum and painless thyroiditis, as a marker of prominent thyroid autoimmunity, while only weakly and transiently positive tests are occasionally found in subacute thyroiditis. Ultrasonographic findings are generally characterized by patchy areas of hypoechogenicity in subacute thyroiditis, while a more diffuse hypoechoic pattern, closely resembling Hashimoto’s thyroiditis is found in postpartum and painless thyroiditis. The colour flow Doppler pattern shows reduced vascularity in the three disorders. Occasionally, and especially when patients are first seen in the recovery or hypothyroid phase, a more subtle picture can emerge from testing with a low but not nil RAIU, and with only mild elevations of ESR, making the differential diagnosis more difficult.

An extrathyroidal source of thyroid hormone should be suspected when more frequent causes of low RAIU thyrotoxicosis have been ruled out. When factitia thyrotoxicosis is suspected, a serum Tg measurement can be extremely useful in confirming the diagnosis, since this disorder represents the only condition in which thyrotoxicosis is associated with an undetectable Tg level (67). At the time of Tg measurement, however, it is important to test the patient’s serum for Tg autoantibodies, since these may cause falsely low Tg levels. Given the high prevalence of thyroid nodularities in the general population, especially in iodine-deficient areas, it is also useful to perform ultrasound scanning of the neck, since, in the presence of thyroid nodules, Tg may be elevated in spite of the ingestion of exogenous thyroid hormone. Colour flow Doppler ultrasonography shows hypervascularity of the thyroid in Graves’ disease and in toxic nodular goitre, and hypovascularity in factitious thyrotoxicosis.

The suspicion of struma ovarii can be confirmed at the time of RAIU, simply by scanning the pelvic area with the probe. The presence of functional thyroid tissue is demonstrated by a significantly increased uptake of iodine in the ovarian region. Further imaging (CT or ultrasound scan) will confirm the presence of an ovarian mass. The levels of CA 125 are elevated in both malignant and benign tumours.

When the source of thyroid hormone is metastatic thyroid follicular cancer, the presence of the latter is usually evident from the history. Since all patients with differentiated thyroid cancer after thyroidectomy take l-thyroxine in TSH-suppressive doses, thyroid function tests should be repeated after withdrawal of the medication in order to rule out iatrogenic thyrotoxicosis. Confirmation is obtained with whole body radio-iodine scanning that will show multiple foci of uptake in several skeletal regions.