In the last decade much information has been accumulated about the pathogenesis of TA at the genetic level. Two different types have been described. A somatic mutation in the gene for the alpha polypeptide of the G protein (Gs alpha) resulting in inhibition of GTPase activity, and a somatic mutation in the TSH-receptor (TSH-R) gene, both leading to constitutive activation of adenylate cyclase (CA) activity and to autonomous hyperfunctioning thyroid adenomas. The majority are mutations in the TSH receptor (10) An analysis of 33 TA from 31 patients for the presence of somatic mutations in the TSH-R gene disclosed in 27 TA (82%), activating mutations at 12 different residues or locations. These mutations cause low-level chronic activation of the TSH-R independent of TSH levels. In addition, in 2 TA (6%), mutations were detected in the Gs alpha gene. Only in four TA were no mutations observed (10a). This study thus shows that in the majority of Belgian patients TA are caused by mutations of the gene of the TSH-R. In a study from Japan, investigating mutations in the TSH-R gene in the "hot spots", i.e. coding for the third cytoplasmic loop and the sixth transmembrane segment, a mutation was found only in 1/38 AFTN. This mutation however did not display any functional abnormality (10b). A study from Austria analyzing the same hot spots reported a frequency of 3 out of 27 AFTN (18%), with mutations in these areas (10c). No mutations in the gene of the TSH-R or Gs-alpha were detected in 15 toxic thyroid adenomas in an Italian study,(10ca), but in six out of seven AFTN in Brazilian patients, mutations were detected in the TSH-R (10ca1). Levels of stimulatory and inhibitory G protein-alpha subunits were studied in toxic adenoma with or without TSH-R gene mutations together with basal and TSH-stimulated CA activities. Expression of both protein subunits was increased independent of the presence or absence of TSH-R gene mutations. Also no correlation was present between basal and TSH-stimulated CA activity and Gs protein subunit levels.(10cb) The authors concluded that mutational activation of the cAMP cascade is not sufficient to generate toxic adenoma and that the pathogenesis of AFTN is probably much more complex. Furthermore it was shown that CA concentrations in AFTN with mutant or wildtype TSH-R or Gs alpha were not different, but that total phosphodiesterase (PDE) activity was higher in the mutant AFTN's, primarily due to induction of PDE4D. This intracellular feedback mechanism may have an impact on the phenotypic expression of mutant AFTN's (10cb1). Derwahl (10f) states that, in addition to mutational changes in the genes of the TSH-R and Gs alpha protein causing TA, the natural heterogeneity of thyrocytes responding to these mutations may play a role in the phenotype expression. This heterogeneity may affect the degree of (hyper) function and histomorphology in the sense of monoclonal or polyclonal proliferation.

Despite different results in the early years it is now recognized that constitutively activating TSH receptor mutations, and with a lower prevalence, mutations of the Gs-alpha, play a principal role in the pathogenesis AFTN (10) For those monoclonal TA where no mutations are found in the TSH-R or Gs-alpha unit, probably other somatic mutations are involved (10fc). Although some believe that iodine deficiency may increase mutation rate and functional expression of autonomy (10 fa) in AFTN, others (10fb) consider the fundamental process of goitrogenesis in (multi)nodular goiter as independent from ID but operating through mechanisms innate to the hereditary and acquired heterogeneity among the thyrocytes themselves. However,superimposed iodine deficiency may shift clinical expression to younger ages. Van Sande et al (10d), hypothesized that, as 16 different activating mutations were identified in the TSH-R gene, that this receptor is in a constrained conformation in its wild-type form. They subjected CHO-K1 cells or COS-7 cells transfected with the human TSH receptor, to mild trypsin treatment, and observed increased cyclic AMP formation. The effect was also observed with the dog TSH receptor, but not with the MSH or LH receptor. The action of trypsin removes or destroys residues 354-359 of the extra cellular domain of the TSH-R. The results of their study support the hypothesis that the C-terminal portion of the large extracellular domain plays a role in the maintenance of this constraint. Further studies from the same group indicated that the first and second extracellular loops contribute to the silencing of the unliganded receptor. This study also showed that both the cyclic AMP and the inositol-phosphates pathways may be activated by TSH-R gene mutations. These mutations are distributed over the first and second extracellular loops, the third intracellular loop, and the third, sixth and seventh transmembrane segments 10e. They also report that TA have a high level of Na+/iodide symporter gene expression, a high thyroperoxidase mRNA and protein content, and a low H2O2 generation. Inositol uptake was also increased but inositolphosphates were not increased. TA secreted more thyroid hormone than the quiescent surrounding tissue. Other characteristics of TA were, increased cycling of thyrocytes as compared to normal surrounding tissue, little apoptosis, and low expression of early immediate genes (10ea). An important study by Fuhrer et al (10eb) shows that a panel of different activating TSHR mutations cause different functional and morphological responses in vitro in rat and human primary thyrocytes. Their data suggest that different biological properties of the TSHR mutants may result in different in vivo phenotypes(10ec)

The prefix "toxic" defines the adenoma from the functional point of view. It is clinically and scintigraphically a single nodule in combination with biochemical changes and clinical signs of thyroid hormone overproduction. Before the availability of sensitive diagnostic tests, particularly the sensitive TSH assay, it was often necessary to establish the diagnosis of a toxic nodule by administration of T3 or T4. Uptake of radioactive iodide or pertechnetate in the nodule would not be suppressed by this maneuver. In addition, administration of exogenous TSH could cause uptake of isotope in the surrounding previously suppressed thyroid tissue. Nowadays the presence of elevated serum free T4 and T3 levels, and suppressed serum TSH, in combination with radionuclide uptake only in the nodule on scintiscan, ( Fig.13-1 )is sufficient for the diagnosis. Hyperfunction of a remnant of thyroid tissue, for instance after thyroidectomy or after thyroiditis, is not excluded by these tests but is very rare. To discriminate between these two possibilities, the presence of a thyroidectomy scar, careful history taking, and measurement of circulating thyroid autoantibodies may be helpful.

Figure 1. Hot nodule in right lobe of thyroid. Note that uptake of radioactivity in the contralateral lobe is suppressed.

The functional development of a toxic adenoma of the thyroid is from a "warm" nodule via a "hot" nodule to a "toxic" nodule. A warm nodule is defined as a nodule that cannot scintigraphically be distinguished from the normal surrounding thyroidal tissue, whereas a "hot" nodule has more activity than the surrounding tissue. A nodule that is "hot" on scintiscan may not actually overproduce hormone in which case serum TSH is still in the normal range and extranodular thyroid tissue is visible on the thyroid scan. The difference between a "hot" and a "toxic" nodule is that in the latter TSH is suppressed and there is typically no activity visible on the scan outside the nodule. It is unusual that an autonomously functioning thyroid nodule is toxic when the diameter is less than 3 cm. (13,14)

Approximately 1 in 10, to 1 in 20 solitary nodules present with hyperthyroidism. This varies from country to country, being more common in Europe than in the USA. (13,15)The problem is 5 times more common in women than men. In a group of 349 patients with autonomous functioning thyroid nodules (AFTN) Hamburger found 18% to be toxic. The proportion of cases of AFTN with hyperthyroidism was 33% in males, but only 17% in females. 13 Forty-eight percent of euthyroid subjects were 40 years or older, whereas 73% of hyperthyroid patients fell in this age group. Of the group of patients 60 years or older, 57% were toxic, whereas this was the case only in 13% of patients younger than 60 years. Two-hundred-thirteen patients had nodules up to 2.5 cm in diameter and of these only 1.9% were toxic. In the remaining 136 patients with larger nodules, 42.6% were toxic. Of the 266 patients who were younger than 40 years only 19.5% had nodules 3 cm in diameter or larger, whereas in the older patients this figure was 45.9%. These data underline the fact that larger AFTN occur in the older age group and that at the time of presentation, nodules less than 2.5 cm in diameter are rarely toxic. The frequency of toxic adenoma in patients referred for thyrotoxicosis may vary considerably in different geographical areas. Percentages between 1.5 and 44.5 have been reported (see Table 13-1). In Europe a female to male ratio of 5: 1 was reported for toxic autonomous nodules (16).

A study from Malmö compared the mean annual incidence of TA between 1988 and 1990 (4.8 per 100,000) with that between 1970 and 1974. No difference in incidence was found (16a)

Autonomously functioning nodules may stay the same size, grow, degenerate or become gradually toxic. In one series 10% of patients followed for 6 years became toxic. Toxicity may develop independent of age, but is much more common in nodules over 3 cm in diameter (up to 20%) (13). By sonography the critical volume at which hyperthyroidism will occur, was 16 ml(17). Changes in nodule size as observed in 159 patients followed between 1 - 15 years (13) are shown in Table 13-2. An increase in size was seen in only 10%. Four percent decreased in size. Loss of function was seen in 4 patients because of degeneration. Eight percent developed overt thyrotoxicosis in a mean follow-up of 3 to 5 years. Three percent developed borderline hyperthyroidism (Table 13-3).

On macroscopic examination, a solitary toxic nodule is surrounded by normal thyroid tissue that is functionally suppressed. Microscopically one would expect, on the basis of current ideas about pathogenesis, to see the picture of a true adenoma with uniform thyroid follicles and without characteristics of malignancy. Studer et al. (1,2)found that hot nodules in so-called multinodular goiter, appear occasionally to be monoclonal in function ( as has also been shown to be the case on the basis of somatic point mutations in the TSH-R gene (see Chapt. 17), but usually are composed of heterogeneous follicles of different sizes and different capacity to take up radioactive iodine (Fig. 13-2). Although such heterogeneity has been identified in multinodular goiters, it is apparent that also in the isolated toxic adenoma multifunctional follicles appear to be present (Fig. 13-3, below). This multifocal histological picture however should not a priori be interpreted as suggestive of heterogeneous origin. Thus it is quite possible that, as in multinodular goiter (see Ch. 17), although of monoclonal origin, secondary acquisition of new inheritable qualities by replicating cells, renders the function and the histological picture of the follicles heterogeneous. Old hemorrhage, sometimes reflected by calcification, may also be present in toxic adenomas. On microscopic examination of a thyroid with a clinically single toxic adenoma, autonomously functioning micronodules are seen in the extranodular tissue. This finding is in agreement with the thesis of Studer et al. (1,2) that the true adenoma is one end of a large spectrum of thyroid nodules growing from single follicular cells or tiny cell families each replicating with an individual growth rate, whereas the grossly abnormal multinodular goiter is situated at the other end of the scale.

Figure 2. Autoradiograph of a hot nodule illustrating areas with different capacity of uptake of radioactive iodine (taken from ref. 1, with permission).

Figure 3. Uniform nature of cells in a nodule formed by proliferation of only one or a few clones of epithelial cells (taken from ref. 2, with permission).

Patients with toxic adenomas present with a lump in the neck with/or without symptoms compatible with thyrotoxicosis. The symptoms of thyrotoxicosis are not different from those with other causes of thyrotoxicosis except that characteristics of Graves' disease, such as ophthalmic disease, pretibial myxedema and acropachy, are not present. Coincidental occurrence of the two diseases may be seen (18). Patients with toxic adenoma tend to be older than those with Graves' disease and onset of thyrotoxicosis is generally slow. Many patients are aware of having a lump in the neck for some years, obviously being non-toxic for a long time. Nodules hardly ever are of such size that mechanical symptoms are present other than a slight discomfort during swallowing. Patients sometimes visit their doctor with cosmetic reasons as the primary complaint.

When clinical symptoms of thyrotoxicosis are evident and a single nodule within the thyroid area is clearly felt with little surrounding tissue on either side, the finding of a suppressed serum TSH value theoretically is sufficient to establish the diagnosis of a toxic nodule. However, to ascertain the severity of thyroid overactivity, measurement of a serum (free) T4 (and when normal of serum (free) T3) is important. It is wise to obtain a thyroid scan, preferably using 123-I. The scan of a toxic nodule will show prominent uptake in the nodule with little or no uptake in surrounding thyroid tissue, and will show appreciably less uptake in the surrounding tissue in the case of a hot nodule. As TcO4- is not organified in thyroid tissue (in contrast to radioactive iodine), discrepancies are occasionally seen. (Fig. 13-4) Rare nodules may be hot with TcO4-, while being cold on scanning using radioactive iodine. In these situations the nodule should be considered cold and therefore of malignant potential. It is therefore advised that if a nodule shows prominent activity with technetium, the scan should be repeated using I123 (19). This is not necessary however, if TSH is suppressed, indicating hyperthyroidism.

Figure 4. Nodule in isthmus of the thyroid which is "hot" on the sodium pertechnetate Tc 99m scan (left) and "cold" on the I131 scan (right).

The presence of carcinoma in a toxic or hot nodule is rare. Horst et al. (16) found no thyroid malignancy in a study of 306 patients with AFTN. Sander et al. (20) reviewed the literature on occurrence of carcinoma in a solitary hyperfunctioning nodule. They also concluded that the incidence of malignancy in a hot thyroid nodule is exceedingly low. Isolated cases of carcinoma development in a hot nodule have however been reported. (21,22) A study from the United States (23) reported the occurrence of 3 carcinomas in 30 consecutive patients operated for solitary hot nodules. (However some authors suggest that occurrence of carcinoma in AFTN may be more than coincidental. (24) It is generally felt that the presence of autonomous function is a reassuring characteristic with regard to the possible presence of thyroid carcinoma.

In a thyrotoxic patient with a solitary hot thyroid nodule in which the surrounding tissue is devoid of substantial uptake, the possibility of the presence of Graves' disease in remnant thyroid tissue (e.g. after previous thyroidectomy or in thyroid dysgenesis) is remote. The ultimate proof that surrounding tissue is suppressed but present can only be obtained after administration of r-TSH and subsequent scanning of the thyroid. However, this is virtually never required to establish the diagnosis. Ultrasound will often reveal a small contralateral thyroid lobe as well as the dominant nodule, but again adds little to the evaluation. It is of little use to perform fine needle aspiration in patients with toxic nodules because cytopathologic differentiation between adenoma and thyroid carcinoma is difficult, if not impossible. This can lead to a high proportion of false positive cytological diagnoses of follicular carcinoma. When a hot nodule is present, but there is still (be it less) uptake of isotope in surrounding tissue and serum TSH is within normal limits, autonomous function of the nodule can be proven by administration of T4 (200 ug/day for 14 days). After this a repeat thyroid scan should show that the surrounding tissue is inactive because of suppression of serum TSH. However, this procedure has no practical consequences and is therefore unnecessary in clinical practice.

As discussed in the previous section, the occurrence of malignancy in a hot or toxic nodule is rare. In the case of a hot nodule no active treatment is necessary. The majority of patients remain euthyroid. Clinical observation and serum TSH measurements at intervals of 6 to 12 months are usually sufficient. There are anecdotal observations that sometimes, probably due to a hemorrhage in the adenoma, spontaneous resolution of the nodule occurs or the nodule becomes cold on scan. Those nodules that grow in size and/or lead to overt thyrotoxicosis should be treated since thyrotoxicosis is generally permanent. Long-term treatment of a toxic nodule with antithyroid drugs is useless, as relapse will almost invariably occur after discontinuation of medication. Three definitive forms of treatment are available, i.e. nodulectomy, treatment with radioactive iodine and percutaneous ethanol injection into the nodule. There are advantages and disadvantages to these approaches.

The advantages of nodulectomy are rapid and permanent control of hyperthyroidism with a very low operative complication rate. Usually the patient is treated preoperatively by antithyroid drugs or, if mild thyrotoxicosis is present, by beta-blocking agents. The incidence of hypothyroidism after operation is low, but surprisingly not zero. Thus, in a series of 60 patients operated for autonomously functioning thyroid nodules( 4), 6.6% became hypothyroid after operation. Two of these patients had received in the past either therapeutic doses of 131I or long term treatment with antithyroid drugs (25). In another series of patients, also treated surgically by unilateral lobectomy for toxic adenoma, 5 out of 35 became hypothyroid, though in 3 it was only temporary. It is difficult to see how patients become permanently hypothyroid even after hemilobectomy. No information however is available about the presence of circulating thyroid autoantibodies and macroscopic status of the contralateral lobe in these cases (15). Generally it is believed that long-term suppression of the thyroid gland does not lead to permanent inactivation after suppression is relieved (Fig. 13-5). It seems likely that coexistence of another thyroid disease was the culprit. The disadvantages of surgery are the risks of general surgery and the expense as well as the residual scar.

Figure 5. Thyroid scan before (left) and after (right) nodulectomy.

Administration of 131-I is also a successful mode of treatment for patients with toxic adenoma (Fig. 13-6, below). The prevalence of hypothyroidism after treatment with radioactive iodine is reported to be absent or low in most publications. Ratcliffe et al. (26) report no hypothyroidism in 48 patients at 6 months after therapy. Also at the same time after therapy, Ross et al. (27) report no hypothyroidism in 45 treated patients. However, when the observation period after treatment is longer, hypothyroidism may be documented. In one report (28) 23 patients were followed 4 - 16.5 years after treatment. Eight patients (36%) had become hypothyroid. The incidence of hypothyroidism was not related to nodule size, the level of thyroid function at therapy, or the total dose of 131-I administered. In 54% of patients nodules were still palpable. In a similar study (29), 126 patients with hot nodules showed an incidence of hypothyroidism of 4.8%, 10 years after 131-I treatment. No relationship was found between the development of hypothyroidism, the size of the nodule, or the total amount of administered radioactivity. Hypothyroidism occurred in 9.7% of patients with an euthyroid hot nodule given 131I, and in only 1.5% of patients with a toxic adenoma. When antithyroglobulin and/or antithyroid microsomal antibodies were present, the prevalence after 10 years was 18.0% versus 1.4% in antibody negative patients. However after a follow-up of 10 years reported even a percentage of hypothyroidism of 40%. From these results it seems that longer follow-up periods may uncover hypothyroidism and the prevalence of this may be related to the presence of thyroid auto- antibodies and not so much to the size of the thyroid nodule and the 131I dose administered. One possible cause of hypothyroidism is found when patients are rendered euthyroid by treatment with antithyroid drugs before radioactive iodine. In this situation the TSH rises and suppressed normal thyroid tissue resumes uptake of radioactive iodine, and therefore is damaged by the isotope. Some authors administer T4 for two weeks , prior to 131-I treatment to be sure that the normal surrounding thyroid tissue is suppressed. It is also possible that high doses of 131I in the nodule provide enough radiation to the surrounding tissue that its function is seriously damaged. Last but not least, late development of hypothyroidism may also be a consequence of the damaging effects of humoral thyroid-autoantibodies triggered by 131I treatment. The usual mean doses of 131-I administered for toxic nodules vary between 296 and 1077 MBq. About 5% of patients develop Graves' hyperthyroidism after treatment with 131-I as has been described as well in patients similarly treated for toxic- or euthyroid multinodular goiter, due to the production of TSHR-antibodies in genetically susceptible subjects. 131-I doses of this order give thousands of rads to the normal tissue surrounding the nodule Both procedures are safe. This author has a preference for 131-I treatment in general.

Figure 6. Thyroid with toxic nodule before (A) and after (B) treatment with 131I.

A more recent development in the treatment of AFTN, used as an alternative to surgery or 131-I treatment, is percutaneous ethanol injection into the nodule under sonographic guidance. The results are good. Euthyroidism is achieved in around 85% of patients when assessed 12 months (30) or 2.5 years (31) after treatment. Cure is virtually 100% in pre-toxic adenomas. Usually injections are repeated 2-12 times at weekly intervals. The treatment is generally well tolerated with few side effects. Results of ethanol injection in relatively large AFTN, (diameter 3 to 4 cm or > 40 ml resp.) (31a,31a1) are also good especially when causing only subclinical hyperthyroidism. This form of treatment does not increase operative risk or histologic assessment when subsequently necessary (31b). Results after a follow-up of 56 months remain favorable (31b1). Laser photo-coagulation of AFTN is the most recently introduced therapy, but experience is very limited (31b2,3). So far this treatment seems to be effective and well tolerated.

The prognosis for the autonomous hot nodule is that most patients remain euthyroid and, as stated, clinical observation is sufficient. Whatever treatment is chosen in the case of a toxic nodule, ethanol injection, surgery or radioactive iodine or laser, most patients become and remain euthyroid after treatment. Long term treatment with antithyroid drugs is not indicated. After treatment, serum TSH measurements at yearly intervals are necessary to detect those patients, especially with circulating thyroid autoantibodies, that will eventually develop hypothyroidism. The occurrence of malignancy in an autonomous or toxic nodule is very rare, but the enlargement of a nodule after 131-I would raise this possibility.

Although the disease was earlier argued to be a mild form of subacute (De Quervain's) thyroiditis, there is now overwhelming evidence that it should be categorized as a lymphocytic thyroiditis. (34-40) There is an association with other autoimmune diseases supporting the concept that this form of thyroiditis is an autoimmune disease (43) No significant association has been found with viral infection. There is a significant association with HLA genotype DR3. Postpartum thyroiditis (see below) is considered to be identical to silent thyroiditis from the phenomenological point of view. (44) Here an association is found with HLA types DR3 and DR5 (45).

On histological examination follicles are disrupted and infiltrated predominantly by lymphocytes and by plasma cells. The infiltration is diffuse and/or focal, sometimes with the formation of lymphoid follicles. The follicular cells have a variable appearance. They can be cuboidal or columnar when stimulated by TSH. Some of the hypertrophied follicular cells have an oxyphilic cytoplasm and are therefore termed Hurthle or Askanazy cells. Thyroid tissue obtained during the hypothyroid or early recovery phase may show regenerating follicles with little colloid. At times persistent mild lymphocytic thyroiditis is seen, but the tissue may also return to normal in others. Extensive fibrosis may be present. Occasionally multinucleated giant cells, so characteristic of subacute thyroiditis, are observed. The histologic picture of postpartum thyroiditis is identical.

A review (35) compiled the reported clinical manifestations between 1971 and 1980 of 112 patients with 122 episodes of silent thyroiditis. Sixty-eight were female, their mean age +/- SD was 32.4 +/- 18.5 yr. while the males were 24.9 +/- 8.2 yr old. In none of the 122 episodes of silent thyroiditis was thyroidal pain present. Recurrences were uncommon. The presenting symptoms in 52 episodes of thyrotoxicosis are summarized in Table 13-4. These symptoms are similar to those found from other causes of thyrotoxicosis and were mild to severe. The duration of the thyrotoxic phase was variable but for the most part lasted less than one year. Mean duration was 3.6 +/- 2.0 (SD) months (range 1 - 12.5 months). Symptoms began 2.5 +/- 2.2 months preceding the initial evaluation. This period is shorter than is usually seen with Graves' disease and certainly in multinodular toxic goiter. Exophthalmos and pretibial myxedema were absent, although symptoms such as lid lag due to increased sympathetic tone were present. In one patient the consistency of the thyroid was reported to be soft, but most authors described the gland as firm. Forty three percent of patients had thyroid enlargement, which was generally symmetrical and enlargement was in most instances mild. Nodularity of the thyroid was uncommon. The clinical course of the disease often follows 4 sequential stages (Fig. 13-7, below), thyrotoxicosis - euthyroidism - hypothyroidism, and euthyroidism. Fifty-seven out of 112 patients became euthyroid and did not develop clinical hypothyroidism. After a brief period of euthyroidism, transient biochemical hypothyroidism developed in 17 patients. In 32 patients clinical hypothyroidism was present. This was temporary in 24 patients but 8 required thyroid hormone substitution (36). Development of Graves' disease, after painless thyroiditis has been documented and TSHR-antibodies have been found in these patients. (36a)

Figure 7. Schematic representation of the four phases of silent thyroiditis (taken from ref. 35, with permission).

During the first phase of the disease, discharge of hormone from the inflamed thyroid results in increases in serum T4, T3 and a decrease in serum TSH. At that time there is no uptake of radioactive iodine in the thyroid (Fig. 13-7). When the diagnosis of thyrotoxicosis factitia is suspected, estimation of serum thyroglobulin levels is useful. During ingestion of T4, little or no thyroglobulin is present whereas serum Tg levels are elevated in silent thyroiditis. In 17 out of 71 patients with silent thyroiditis, moderate elevations of antithyroglobulin antibodies were also present by tanned red cell hemagglutination assay (35). Antimicrosomal antibodies were examined in 53 patients using the complement fixation test or by microsomal fluorescence. Using the former technique 22 patients had positive antibodies, and by the latter 4 out of 7 were positive (35). The recently developed RIA for human antithyroglobulin antibodies has greater sensitivity. In a small series of 7 patients with silent thyroiditis, all were positive using this RIA (46). Indicators of inflammation were not useful. The white blood cell count is generally normal. In 53 episodes, 34 had elevated erythrocyte sedimentation rate (ESR), but it was greater than 40 mm/hr in only 8 (35). This contrasts to the typical marked elevation of ESR in patients with subacute granulomatous thyroiditis and helps to differentiate the two conditions. During the acute phase urinary iodine excretion is high normal to elevated and after resolution of the thyroiditis reduced to 1/3 of its original value (33) As the first phase progresses, T4 and T3 decline into the euthyroid range (second phase) and reach subnormal levels in the hypothyroid range (third phase) in 40% of patients (Fig. 13-7, above). The erythrocyte sedimentation rate, if elevated, decreases gradually while thyroglobulin levels decrease as well. After the third phase, patients gradually enter the euthyroid phase, heralded by an increase in thyroid hormone levels and resumption of thyroidal radioactive iodine uptake. Uptake may temporarily rebound above normal before returning to normal values. Serum TSH starts to elevate at the end of the hypothyroid phase. The hypothyroid phase may last several months. In 26 episodes, patients became euthyroid after a mean period of 62 months, after the onset of the hyperthyroid symptoms. Thyrotrophin levels may increase during the recovery phase, and can remain elevated for many months. The delayed increase of TSH (i.e. at the end of the hypothyroid phase) is due to TSH suppression during the thyrotoxic phase. This phenomenon is also usually seen after withdrawal of thyroid hormone in patients treated with supraphysiological doses. The delay time ranges between 2 and 5 weeks (47). Permanent hypothyroidism occurs in about 7% of patients with silent thyroiditis, but "cured" patients may ultimately become permanently hypothyroid (see following section). A weak positive correlation was reported between the echo level at the onset of thyrotoxicosis and the lowest T3 level during the clinical course (p less than 0.05) (47a).

As thyrotoxicosis in silent thyroiditis is usually mild, treatment to relieve toxic symptoms is often not necessary. If needed, ß adrenergic blocking agents can be administered. It is not useful to give antithyroid drugs because thyrotoxicosis is not the result of increased thyroid hormone synthesis, but of discharge of thyroid hormone from the thyroid gland due to the inflammatory process. The effect of propylthyouracil or iopanoic acid to block peripheral T4 to T3 conversion may be of some clinical benefit. If more serious thyrotoxicosis is present, administration of anti-inflammatory drugs may be of benefit. In this case prednisone can be administered in dosages ranging between 40 and 60 mg per day, usually resulting in rapid decrease of the inflammation (48). After 1 to 2 weeks the dose can be tapered by 7.5 to 10 mg per week. In the case of relapsing thyroiditis it is rarely necessary to perform a subtotal thyroidectomy (40). As an alternative, "thyroidectomy" may be induced by administration of radioactive iodine during a remission. After the thyrotoxic phase many patients become temporarily hypothyroid (see clinical course). Often no thyroid substitution is necessary during this period. If it is, the dose should not completely compensate for the hypothyroidism since a slight TSH elevation will facilitate thyroid recovery. Only a small proportion of patients remain permanently hypothyroid and then full substitution with L-thyroxine is necessary, keeping serum TSH within the normal range. Patients apparently fully recovered from silent thyroiditis may ultimately develop thyroid failure. In a series of 54 patients, Nikolai et al. (49) reported that in about half of patients permanent hypothyroidism ensued. This is in contrast to subacute thyroiditis, which is followed in almost all patients by permanent recovery. Therefore patients with painless thyroiditis should be followed thereafter at yearly intervals with appropriate testing.