Thyroglobulin and thyroperoxidase
autoantibody testing: clinical value and methodological aspects

Contents

Thyroglobulin antibodies (Tg Abs) are circulating immunoglobulins directed against different epitopes of the thyroglobulin molecule. Thyroid microsomal antibodies (TPO Abs) are circulating immunoglobulins directed against a component of the smooth endoplasmic reticulum of thyroid cells. Recently, this microsomal antigen was found to be identical or at least to contain as main component thyroid peroxidase (TPO) (1). Detectable levels of Tg Abs and/or TPO Abs are mainly associated with thyroid autoimmune disorders and with thyroid cancers but low concentrations are also found in a significant percentage of the normal population.

While they represent a rather disparate group, thyroid autoimmune diseases are probably caused by similar but distinct disorders of the normal immune system. These chronic degenerative diseases may cause either primary hypothyroidism or primary hyperthyroidism (see Table I). In mild forms, compensation mechanisms at the pituitary level are often able to normalize the hormonal secretion of the gland. This, for example, would explain the euthyroid goiter observed in some patients with Graves' disease. Hashimoto's thyroiditis is usually associated with hypothyroidism but transient hyperthyroidism or euthyroidism are not rare at early stages.
   

In all countries where there is no iodine deficiency, autoimmune disorders are believed to be the leading cause of thyroid diseases, from hyperthyroidism to hypothyroidism. Reliable epidemiologic data have been difficult to establish. The numbers depend upon patient recruitment (hospital patients, blood donors, general population) and the criteria for selection (clinical evidence of thyroid disorders; abnormal TSH, T4 or T3 concentrations; presence of thyroid autoantibodies). One of the most carefully controlled studies remains the Wickham survey performed in the UK by Tunbridge et al. in 1977 (2).

Later studies agree reasonably well with Tunbridge's findings and indicate that established hyperthyroidism is found in 2.5% of the female population and 0.25% of the male population. More than 85% of these cases are due to Graves' disease. New cases affect 3 in 1000 women each year, mostly in their fourth or fifth decades.

Overt hypothyroidism is mainly due to Hashimoto's thyroiditis and is found in 1.4% of the female population with an annual incidence of newly diagnosed cases of 2-3 per 1000 women (mean age at diagnosis: 57 years).

Considerable efforts have also been made to evaluate the incidence of asymptomatic autoimmune diseases. Some studies are based on post mortem histological examination of thyroid glands but most recent surveys are based on the detection of thyroglobulin and microsomal antibodies. Here, the numbers are more elusive because they largely depend upon the cut-off selected for positive results and upon the sensitivity of laboratory tests. However, all reports agree on a much higher incidence of positive Tg and TPO Ab results in women, specially after 50, than in men.

From a practical point of view, the current impossibility to identify population groups at high risk of developing overt autoimmune thyroid disorders speaks clearly against mass-screening programs. Many individuals with thyroid antibodies reach an advanced age without ever developing clinical signs of thyroid disease. The heterogeneity of the antibodies detected further complicates the issues: some of them are probably not associated with thyroid disorders.

Several researchers have attempted to identify genetic factors linked to thyroid autoimmune diseases. Early studies of twins indicated that about half of monozygotic pairs are concordant for hyperthyroidism in contrast with less than 5% of dizygotic pairs. Other groups found that hyperthyroidism was more frequent in sisters of affected patients of Graves' disease. More recently, several studies attempted to identify HLA antigens closely associated to Graves' disease. Several authors found weak but significant associations with HLA-B8 in whites, with HLA-Bw35 in Japanese and a slightly stronger association with HLA-Bw3 and HLA-DR3 in whites. In general, however, all these associations point out towards a multifactorial etiology and are not sufficient to define at-risk populations.

In Hashimoto's thyroiditis, no clear association with any HLA antigen has been found but thyroid autoantibodies have been found in 50% of asymptomatic relatives of patients with lymphocytic thyroiditis. The disease occurs frequently in the fathers and brothers of patients (as well as in mothers and sisters) although the female: male ratio is 5-20 to 1 in unselected patients. Like in Graves' disease, there is an inherited tendency toward Hashimoto's disease but the exact mechanisms of transmission remain unknown and the same conclusion applies to myxedema.

Recent advances in the field of autoimmunity at large made it logical to investigate whether thyroid diseases could be related to systemic diseases. And indeed, Silman found a significantly higher frequency of thyroid dysfunction together with Tg and TPO Abs in patients with rheumatoid arthritis and their families (3) but no link to the HLA-DR3 or B8 antigens. With the availability of more sensitive testing methods (ELISA) for thyroglobulin and microsomal antibodies, similar results have been obtained with patients with myasthenia gravis, diabetes mellitus, SLE, biliary cirrhosis, mixed connective tissue disease and Sjögren's syndrome (4). Associations with Addison's disease (Schmidt's syndrome) and with pernicious anemia are well-documented (5) and these patients show in the laboratory detectable levels of anti-adrenal antibodies or anti-gastric parietal cell antibodies. TPO Abs are found to be more sensitive than Tg Abs in these studies. TPO Abs are detected in 50-90 % of autoimmune patients depending upon their selection: with or without clinical signs of a thyroid dysfunction. While, again, there is no clear-cut mechanism to explain these associations, complex genetic factors are thought to play a role and there might be a common genetic link.

From a practical point of view, it makes sense to include TPO Ab   detection in the laboratory workup of patients with non-thyroidal immune diseases, even in the absence of clinical symptoms and to perform a complete thyroid exploration of all patients testing positive (Table II).
  

   

Patients with diagnosed autoimmune diseases
	- insulin-dependent diabetes
	- pernicious anemia
	- myasthenia gravis
	- SLE
	- biliary cirrhosis
	- Addison's disease
	- mixed connective tissue disease
	- rheumatoid arthritis
Relatives (including males) of patients with Hashimoto's disease

Table II: Patients for which the risk of developing thyroid autoimmune disorders justifies TPO Ab testing in the absence of clinical and/or biological signs of thyroid dysfunction.

Several researchers have attempted to identify the mechanism triggering thyroid autoantibody production. All these studies have, until now, failed to provide a satisfactory answer. At this stage, it is likely that more than one mechanism is involved and this would be consistent with antibody heterogeneity. Different groups have suggested that there could be an antigenic stimulation at the initial step of antibody production. For example, it has been proposed that a viral infection such as subacute thyroiditis could induce occult antigenic changes within the thyroid. This is rather unlikely in Graves' or Hashimoto's diseases where a vast majority of cases arise without any recent history of thyroid infection. Similarly, almost all patients with subacute thyroiditis recover and do not develop thyroid antibodies (6).

Recent advances in molecular biology have failed to demonstrate antigenic alterations of the Tg molecule in thyroid autoimmune diseases. However, the antigenic stimulation hypothesis may be valid, at least in part, for thyroid cancers. In these patients, thyroglobulin release in the bloodstream is accelerated and iodinated forms, normally confined to the gland, are secreted.

Another hypothesis is that a disorder in immunoregulation is at the origin of both Graves' and Hashimoto's diseases. Some studies found alterations in the number of natural killer cells in patients with untreated Graves' disease but no anomaly in patients with Hashimoto's thyroiditis. Several groups have focused on the suppressor arm of the immune system, T lymphocytes (7). Early studies were contradictory because their authors were trying to substantiate a generalized defect of Ts cells. However, rather than a decrease in the number of T lymphocytes or in their overall activity, it seems that organ or antigen specific defects have to be considered. How these defects appear is another problem. Jansson et al. (8) have found lymphocytotoxic antibodies specific for CD8+ cells in patients with autoimmune thyroid diseases. These antibodies would in turn, reduce the number of Ts cells and lead to autoantibody overproduction. This diminution of CD8+ cells is observed in some patients, but not all. Other possibilities include biochemical abnormalities inherent to the Ts cells themselves. TSH is also thought to stimulate TPO Ab production in thyroid cells trough complex pathways involving cyclic AMP (9).

While Ts cells seem to play a significant role in thyroid autoimmunity, much work remains to be done to elucidate the long chain of reactions leading to antibody production and to relate it to genetic factors.

Much work has been done on antibodies binding the TSH receptor in Graves' disease patients. The stimulating (and sometimes) blocking effects of these antibodies, their role in the development of hyperthyroidism are well established but fall outside the scope of this review.

In contrast, no pathological role has been found, until now, for thyroglobulin antibodies. For thyroperoxidase antibodies, the situation is unclear. These antibodies are complement fixing and it has been thought that they were implicated in the thyroid cell damage characteristic of autoimmune thyroiditis. TPO antibodies are able to inhibit TPO enzymatic activity in vitro. Thyroid peroxidase has been extensively studied by Taurog et al. (10) and is known to catalyze iodination of tyrosine and oxidative coupling of two diiodotyrosine residues in thyroglobulin to form thyroxin. It is reasonable to think that blocking this mechanism will induce thyroid disorders but does not explain the cell destruction process characteristic of thyroid autoimmunity. While there is no firm evidence against a cytotoxic role of TPO Abs, other antibodies are certainly involved in cell destruction, via the complement pathway (11, 12). These IgG are not well characterized and are found in patients with or without Tg and TPO antibodies.

Thyroglobulin antibodies

Delespesse et al. demonstrated that Tg Abs can be present in the serum of patients with proven thyroid disorders as IgG, IgA or IgM. However, IgA represent an uncommon finding with only low concentrations in positive samples, IgM are more frequent but again at very low concentrations compared to IgG (13). This explains why most testing methods detect selectively IgG antibodies. Other groups have shown that these IgG antibodies belong to any of the IgG subclasses (14).

Considerable work has been done on these IgG antibodies, not for the assay of Tg Abs but in order to improve assays of serum thyroglobulin. In the clinical laboratory, the determination of thyroglobulin represents an extremely valuable marker for monitoring patients with differentiated thyroid cancer after surgery. Since the availability of Tg RIA or IRMA, it has been recognized that thyroglobulin autoantibodies do interfere in these assays and produce artificially low (IRMA) or high (RIA) Tg concentrations. Tg Abs are not present in all sera from thyroid cancer patients but do not constitute a rare finding, firstly because a thyroid autoimmune disease may coexist with a thyroid cancer and secondly because thyroid cancers themselves may trigger Tg Ab production. A simple recovery experiment on the patient sample will indicate if there is an interference from Tg Abs or not. Unfortunately, when the interference is present, there is no practical possibility to use the Tg assay in a reliable manner.

Thyroglobulin is an extremely complex molecule (MW 660 000) and is composed of two polypeptide chains and a carbohydrate moiety representing 10 % of the molecular weight. Such a molecule has obviously several antigenic determinants and several attempts have been made to raise monoclonal antibodies to Tg which would bind the molecule at epitopes different from those recognized by autoantibodies (15). About 40 antigenic sites have been identified and only 4 to 6 are recognized by circulating autoantibodies found in Graves' disease or in Hashimoto's thyroiditis. There are now examples of assays free of interferences from Tg antibodies in autoimmune thyroid diseases. Unfortunately, the situation is more complex in thyroid cancers where the heterogeneity of Tg Abs is much greater. A possible explanation of this phenomenon would be the heterogeneity of thyroglobulin in cancer. Until now, no antibody combination has been able to remove totally the risk of Tg Ab interference in Tg measurement for cancer patients. From a practical point of view, it remains necessary to conduct recovery experiments on all samples where Tg has to be measured for cancer follow-up or to restrict the recovery test to samples found positive in a sensitive Tg Ab test.

Thyroperoxidase autoantibodies

Studies on the heterogeneity of TPO Abs have led to observations similar to those obtained with Tg Abs; they can exist as IgM or IgA but predominantly as IgG. IgG subclasses 1 and 2 are domiment. These antibodies are clearly polyclonal and up to 7 epitopes have been identifed on human TPO , mainly at the catalytic site of the enzyme (16). Recent studies on TPO (or TPO) Abs using monoclonal antibodies and animal models have led to two interesting findings:

cross-reactivity between antibodies to thyroglobulin and thyroperoxidase antibodies: since the availability of the first sensitive radioimmune assays for thyroid autoantibodies, it was recognized that there could be some degree of cross-reaction of Tg Abs in TPO Ab assays. This was thought to be due to contamination of microsomal antigens by thyroglobulin during the purification process from human thyroid tissue. Some commercial assays for TPO Abs even included human Tg in the assay buffer to prevent this cross-reaction. Using stringent purification procedures, Ruf et al. (17) obtained TPO preparations free from Tg. Still, the cross-reaction persisted and they demonstrated that in some patients, there is a group of IgG antibodies able to recognize both Tg and TPO. This sub-population binds different epitopes on Tg and on TPO, is polyclonal and display a higher affinity for Tg than for TPO: whether Tg and TPO share epitopes or are these antibodies specific for closely related epitopes remains unclear. There is also cross-reaction of TPO antibodies with myeloperoxidase.

differences between TPO Abs in thyroid diseases and in other immune diseases or in healthy subjects: several groups have demonstrated that TPO antibodies in thyroid autoimmune diseases are able to inhibit TPO enzymatic activity, using the guaiacol assay (18). Kohno et al. have recently investigated TPO Ab inhibiting capacity (19) in healthy subjects and in SLE patients (a population where TPO antibodies are more frequent by a ratio 2:1 compared to the normal population). They found that the inhibiting capacity was much lower in SLE or in healthy subjects with anti-TPO antibodies, compared to patients with thyroid disorders and these differences cannot be explained by the difference in anti-TPO concentrations found in the different groups. This would support the hypothesis that anti-thyroid antibodies in non-thyroid disorders are produced by different immunological mechanisms and may be released to "scavenge" thyroid antigens from dead thyroid cells and to prevent the development of thyroid autoimmunity.

Several methods have been used to detect Tg and TPO Abs since the tanned red cell hemagglutination technique described by Boyden in 1951 for Tg Abs. They include immunofluorescence, radioimmunoassay and ELISA. Immunofluorescence and RIA are gradually abandoned for practical reasons while both passive hemagglutination assays and ELISA are widely used in routine. Independently of the intrinsic quality of commercial assays, it seems that ELISA techniques are more suitable for quantitative determinations and may reflect more accurately the response of the autoimmune process to treatment (20).

There has been some debate regarding the immunoglobulin classes to be detected in these assays. Most radioimmunoassays use 125I-Protein A as radioligand and are able to detect IgM and all IgG antibodies except IgG 3. In contrast, most ELISA methods measure selectively IgG antibodies. This may explain some discrepancies between the results obtained with different tests.

Another controversial issue has been the cross-reaction of Tg Abs in assays for TPO Abs. Originally, some assays were using microsomal preparations contaminated with Tg so that artefactual cross-reactions were indeed present. This can be prevented by adding large amounts of Tg to the incubation buffer as it is done for passive hemagglutination tests and for some RIA assays or by using highly purified microsomal antigen preparations as it is done for ELISA techniques. In fact, the issue is of academic interest but has little clinical relevance since thyroperoxidase antibodies are present much more frequently than Tg antibodies while Tg antibodies are rarely detected in the absence of TPO Abs (21). Moreover, as seen previously, a certain degree of cross-reaction has to be expected due to the fact that Tg and TPO share some antigenic determinants or have epitopes with extremely close structures.

The interference of circulating Tg in Tg Ab determination has also been observed in several assays. It may be partially removed by heat inactivation at 56°C for 30 min., a prerequisite for passive hemagglutination assays. In ELISA tests, endogenous Tg will lower the response but the risk of a false negative is limited to very high concentrations of serum thyroglobulin, a very rare occurrence except in thyroid cancers where quantitative autoantibody measurements are of little clinical relevance. The massive release of Tg and TPO observed during subtotal thyroidectomy also causes a marked decrease in Tg and TPO Ab concentrations (22).

Finally, major difficulties have been encountered by most groups who attempted to correlate numerically test results obtained with different methods. Besides what has been previously mentioned, there is here a clear unitage problem. All passive hemagglutination kits provide results in titers while most ELISA's (or RIA's) express results in arbitrary or international units. There is an obvious interest in standardizing all test results in international units. This would certainly harmonize the results and ease comparative studies but would not solve all problems. The two currently available reference preparations (anti-Tg Abs; MRC 1st international 65/93; anti-TM/TPO Abs: serum 66/387) are pools of pathological sera and the heterogeneity of the antibodies that they contain may influence assay calibration.

A last observation made by Lukinac suggests that some passive hemagglutination tests are subject to an interference by an unknown compound present in the serum from patients with chronic renal failure and partially removable by hemodialysis (24).

Graves' disease (syn: Basedow's disease, autoimmune thyrotoxicosis, exophtalmic goiter) is a form of chronic hyperthyroidism characterized by one or a combination of the following clinical signs:

• hyperthyroidism with diffuse enlargement of the thyroid gland,

• infiltrative ophtalmopathy (50 % of cases),

• infiltrative dermopathy or pretibial myxedema (1-2 % of cases).

Common symptoms include nervousness and/or tremor, weight loss (usually in face of an increased appetite), palpitations, heat intolerance and excessive perspiration, emotional lability, muscle weakness and hyperdefecation.

Overt Graves' disease has obvious symptoms and is easily diagnosed when thyrotoxicosis is accompanied by proptosis. Laboratory tests will concentrate on the evaluation of the thyroid function in order to normalize it with antithyroid drugs (thionamides), 131-iodine or surgery. Thyroid autoantibody testing has limited clinical value in these cases, except for TSI Abs in women in gestational age. Tg Abs are present in about 25 % of cases while TPO Abs are present in 90 % of cases so that combined Tg/TPO Ab testing will identify virtually all cases.

However, when eye signs are not obvious, several non-thyroid disorders may cause symptoms close to those of thyrotoxicosis:

	some psychiatric patients may present with fatigue, weight loss and palpitations,
	insulin-dependent diabetics often presents with weight loss and an increased appetite,
	patients with a pheochromocytoma will show emotional lability and weight loss despite a normal appetite.

In all these patients the absence of thyroid antibodies combined with normal thyroid tests will help to rule out Graves' disease.

It should also be noted that Graves' disease without exophtalmos is less frequent than found by clinical observation. CT scan and other imaging techniques have shown that about 65 % of patients with Graves' disease without eye signs have evidence of ocular muscle involvement.

While exophtalmos has been described in Hashimoto's disease (with or without hypothyroidism), thyroid eye disease occurs mainly in Graves' disease and is already described by Robert Graves himself in 1835 and by von Basedow in 1840. The etiology of these disorders is poorly understood so that, from a practical and clinical point of view, it still makes sense to distinguish between dysthyroid ocular signs and infiltrative ophtalmopathy, even if they might have a common origin.

Dysthyroid ocular signs

The prominent stare of Graves' disease is secondary to lid retraction. This sign is present in up to 90 % of all patients with Graves' disease. There might be also upper eyelid retraction and eyelid lag, but the vision is not significantly impaired and does not justify a specific treatment. In contrast with infiltrative ophtalmopathy, this condition may create problems of differential diagnosis. Almost invariably, Tg and/or TPO Abs will be present, sometimes at low levels, while thyroid function tests might be normal or reveal mild hyperthyroidism. Thyroid antibody testing do represent here a sensitive indicator and will allow to exclude other causes of eyelid changes such as orbital tumors.

Infiltrative ophtalmopathy

Clinically, this is a serious condition covering stage II (soft tissue involvement) till VI (sight loss) of the modified "NO SPECS" classification of eye changes of Graves' disease (American Academy of Ophtalmology). The fat content of extra-ocular muscles is abnormally high (orbital cellulitis) and the muscle volume may be increased to 8-10 times the normal. This proliferation causes the eye to protrude out of the orbital cavity (uni- or bilateral proptosis). Periorbital edema is caused by an inflammatory infiltrate by lymphocytes in the extra-ocular muscles and muscle necrosis occurs at a later stage. The progression can be fast and may require aggressive therapy. The autoimmune etiology is reasonably well established and plasmapheresis, cyclosporine A and/or corticotherapy can bring significant improvement. However, the association with thyroid autoimmunity is still mysterious.

Infiltrative ophtalmopathy may accompany or follow hyperthyroidism but in 10-15% of cases, it appears before hyperthyroidism or even in the absence of hyperthyroidism. The frequency of these cases of "euthyroid ophtalmic Graves' disease" also called "endocrine exophtalmos" has probably been overestimaded in early studies: more sensitive thyroid and antibody tests will show that there is usually an occult autoimmune thyroid disease but not always. Reports of an association of elevated thyroglobulin and Tg Abs in these cases have not been confirmed. Other reports have indicated that circulating antibodies against a retro-orbital muscle antigen are present in infiltrative ophtalmopathy (25). On limited numbers of patients, it has been observed that these antibodies accompany thyroid autoantibodies when there is a thyroid involvement but can also represent the sole marker of an autoimmune disorder. Like in Schmidt's syndrome, it might be that there is a common deregulation of the autoimmune system leading to the production of distinct antibodies, rather than a direct relationship between the thyroid and the ocular involvement.

Thyroid antibody testing is therefore of little interest for infiltrative ophtalmopathy and of no value for euthyroid Graves' exophtalmos (26).

Autoimmune destruction of the thyroid gland accounts for 95% of primary hypothyroidism, of which Hashimoto's thyroiditis and idiopathic myxedema represent the majority of cases.

Hashimoto's disease (syn. lymphocytic thyroiditis, struma lymphomatosa) is a chronic from of hypothyroidism caused by an inflammatory infiltration and tissue damage localized in the thyroid. Its main symptoms are the opposite of those seen in Graves' disease: dry coarse skin, swelling of the hands, face and extremities, cold intolerance, decreased sweating, hoarse voice, modest weight gain with anorexia, constipation ... Weakness, fatigue, lethargy and goiter are also present. The autoimmune origin is well established with Tg Abs and TPO Abs present in 85-100% of cases.

The diagnosis of Hashimoto's disease is normally not difficult but two conditions deserve a special mention:

Early hypothyroidism is, in principle, easy to diagnose in the laboratory, specially with the recent progresses made for free T4 and TSH testing. Nevertheless, numerous studies have shown that Hashimoto's disease is largely underdiagnosed, probably because the non-specificity of signs and symptoms does not immediately suggest hypothyroidism as their cause. The presenting symptom is usually an asymptomatic goiter, of minimal or moderate size. The patient may complain from a mild discomfort in the neck but most of the time, the initial finding occurs incidentally, during the palpation of the gland. At this stage, only 20 % of patients have biological signs of hypothyroidism. It is therefore advisable to include thyroid antibody testing in the laboratory initial exploration of a suspected hypothyroid patient.

In some cases, Tg and/or TPO Abs will be found together with normal or borderline TSH and free T4. These patients will have to be monitored closely since autoimmune thyroiditis may develop very insidiously. If antibody levels remain persistently high, the criteria to initiate treatment will be the TSH response to TRH. As soon as this response is impaired, treatment has to be considered.

In contrast, antibody testing, if negative, will rule out autoimmune thyroiditis for the numerous cases of non-toxic goiters where the cause of the thyroid enlargement must be identified (dietary goitrogens or goitrogenic drugs and chemicals).

Hyperthyroidism may also appear transiently, at the initial stage of Hashimoto's disease and histological studies have even revealed some cases of coexistence within the same gland of Hashimoto's and Graves' diseases lesions.

Juvenile lymphocytic thyroiditis is the most common cause of acquired hypothyroidism in children after 6 years (in the absence of endemic deficiency of the dietary iodine intake). It has been diagnosed in 1.3% of school children (11-18 years) in a US survey. Histological findings may be somewhat confusing with less or reduced lymphocytic infiltration compared to the adult form. TPO Abs are usually moderately elevated and will help to confirm the autoimmune origin of the disorder.

In this variant of Hashimoto's disease, fibrosis, which is often detectable in the classical form, predominates over lymphocytic infiltration. Goiter is present and glandular tissue is destroyed to a large extent so that hypothyroidism is usually easy to detect both clinically and in the laboratory. Tg and TPO Abs are present in 90-100% of cases and TPO Abs levels are extremely high. The disease is rare and the female : male ratio is 5 to 1.

Thought to be either a variant or the end-stage of Hashimoto's disease, primary myxedema is characterized by a mucinous edema. The skin throughout the body shows an accumulation of glycosaminoglycans in the dermis (the same observation, but localized, is made in the pretibial myxedema associated with Grave's disease). At the thyroid level, fibrous tissue replaces the normal glandular tissue, leading to a deprivation of thyroid hormones and to an enlargement of the pituitary with excessive TSH secretion. The florid nature of this disorder makes its diagnosis easy. Tg Abs are present in 70% of cases and TPO Abs are even more frequent.

Schmidt's syndrome has originally been described as the association of classic (non tuberculous) Addison's disease and primary hypothyroidism. It is now thought that Schmidt's syndrome is a form of multiple endocrine insufficiency where genetic factors play a significant role. Schmidt's syndrome has been observed in patients with diabetes mellitus, primary ovarian failure and idiopathic hypoparathyroidism. Circulating thyroid and adrenal autoantibodies are found together in the serum of these patients. The histologic lesions of the adrenal and of the thyroid gland are clearly autoimmune but, as in most autoimmune disorders, it is still unclear whether these circulating antibodies play a role in the pathogenesis of Schmidt's syndrome or are by-product of a primary autoimmune lesion within the glands.

Tg and TPO Ab circulating levels respond to therapeutic agents acting on the immunologic system or to the removal of their target antigens.

In Hashimoto's disease, replacement therapy with T3 and T4 to correct hypothyroidism or long-term suppressive treatment with anti-thyroid drugs to normalize hyperthyroidism will decrease antibody titers. Tg Abs levels decrease more slowly than TPO Abs.

In Graves' disease, classical anti-thyroid drugs such as methimazole inhibit hyper-secretion of T4 and T3, but may also act as immunosuppressors since levels of TPO Abs and TS Abs (if present) do usually decrease significantly.

When subtotal thyroidectomy is preferred to anti-thyroid treatment, antibody titers also decline steadily. Since there is surgical removal of the lymphocytes infiltrating the gland, this phenomenon is quite logical but also confirms that antibodies are produced within the gland and subsequently released in the circulation.

More puzzling is the effect of radiation therapy. After 131I treatment, serum antithyroglobulin and antithyroperoxidase Abs rise sharply, to many times the pre-treatment level. This phenomenon is transitory and levels usually return to lower but still pathological levels after a few months. A possible explanation would be that irradiation further depresses the suppressor T cell activity or that it increases antibody production by releasing infiltrated lymphocytes in the blood stream (28).

Since all forms of treatment for Graves' disease (anti-thyroid drugs, 131I or subthyroidectomy) have pros and cons, several authors have evaluated the potential interest of measuring Tg and TPO Ab levels before treatment to select the most appropriate therapy or to evaluate the chances of natural remission. Unfortunately, all studies available to date indicate that antibody testing are not a deciding factor.

Corticosteroids are normally not used to treat thyroid autoimmunity. In Hashimoto's disease, they are able to lower antibody titers and to reduce the goiter but are not suitable for long term treatment. In Graves' disease, a large dose of prednisone suppresses TSH stimulating antibodies and may improve exophtalmos but does not induce a complete remission so that this course of treatment is uncommon.

Hypothyroidism

Hypothyroidism in pregnancy is a relatively uncommon finding. If the criteria for selection is the concentration of T4 (free and total), low levels are found in 0.3 to 0.7% of pregnant women compared to 0.6 to 1.4% in the overall female population. This is not surprising considering that untreated hypothyroid women are subfertile or infertile. However, if the criteria for selection is an elevated serum TSH, the proportion increases significantly, up to 2.5% in a recent study of 2000 pregnant women. Among these women, a large majority of subjects present with compensated thyroid diseases and only 0.3% have very high TSH levels together with low T4. In both subgroups, autoimmunity is the most common etiologic factor since thyroid autoantibodies are detectable in 58% of the subjects in the moderately high TSH group and in 90% of the group with a very high TSH (29). In case of non-compensated hypothyroidism, the most frequent observations are Hashimoto's thyroiditis and atrophic asymptomatic thyroiditis.

Laboratory aspects: the antibodies most frequently detectable are TPO autoantibodies. All authors have noticed that TPO Abs are primarily found at the beginning of pregnancy so that ideal testing time is around weeks 15-18. During pregnancy, antibody concentrations usually decrease and they may even become undetectable in late pregnancy to rebound after delivery.

Thyroperoxidase antibodies do not represent per se a major threat for the mother or for the fetus although maternofetal transmission of the antibodies is clearly demonstrated but without any evidence of fetal cytoxicity. The therapeutical approach will consist in a substitutive therapy for the mother in case of low serum T4 to reduce the risk of abortion. This treatment does normally not affect TPO Ab concentrations.

Hyperthyroidism

Hyperthyroidism is thought to affect 0.05 to 0.2% of all pregnant women, the most frequent cause being Graves' disease but Hashimoto's thyroiditis in its hypermetabolic phase is also observed.

In pregnant women with Graves' disease, thyroglobulin, thyroperoxidase and TSH receptor autoantibodies are present, mainly during the first trimester. Since in most cases, the diagnosis of Graves' disease is known before the pregnancy, testing for thyroid antibodies in these cases is of limited value (with the exception of TBI Abs which cross the placental barrier and may induce the same hyperthyroidism in the fetus) and the laboratory workup will concentrate on TSH and thyroid hormones.

While, in general, there is an improvement of Graves' disease, severe hyperthyroidism can also appear, usually in the first trimester. In these cases, testing for autoantibodies will be of interest to differentiate between Graves' disease, Hashimoto's thyroiditis (hypermeta-bolic phase) or thyroid dysfunctions without autoimmune origin.

Postpartum thyroid dysfunction and thyroid antibodies

Postpartum thyroiditis, first described in 1976 by Amino, is now recognized as a clinical entity and affects 9 to 12% of all pregnant women. However, only half of the affected parturients show clear signs of thyroid dysfunction, either transient hypothyroidism or transient hyperthyroidism leading to hypothyroidism. These troubles generally occur during the first 6 months after delivery and regress spontaneously within the next 6 months. Symptoms are vague so that the condition goes usually unnoticed. Thyroxin administration is requested in rare cases of permanent hypothyroidism.

In the laboratory, serum TSH is the best marker to judge the evolution of these patients. However, Jansson et al. (30) observed that detectable thyroperoxidase antibodies during the first quarter of pregnancy is strongly associated with the subsequent development of postpartum thyroiditis. This finding should encourage testing for TPO Abs in early pregnancy since recurrent thyroiditis have been described in successive pregnancies.

Amiodarone is a drug widely used for long-term treatment of cardiac arrhythmia and angina pectoris. This drug contains 37.2% of iodine and its elimination is slow (half-life: 5-100 days). It is considered that approximately 9 mg of free iodine is released every day by the metabolism of a daily 300 mg intake. Amiodarone is also known to inhibit T4 and reverse T3 deiodination leading to a decrease in serum T3, an increase in serum reverse T3 and, in most cases, to some elevation of serum T4 (free and total). These phenomenons have to be expected in all patients and are acceptable within certain limits. However, amiodarone is also known to induce in certain patients overt hyper- or hypothyroidism.

Hyperthyroidism occurs in 10% of patients residing in areas where iodine intake is low or subnormal (such as continental Western Europe); amiodarone-induced hyperthyroidism is a serious complication because it often resists to conventional antithyroid drug therapy and amiodarone has to be discontinued immediately. It is best documented by an increase in serum T3.

Hypothyroidism occurs in about 20% of patients living in areas where iodine intake is sufficient; it is easily treated with L-thyroxine administration and it is best documented by an elevated TSH (or reverse T3). It has been found retrospectively that many patients who developed induced hypothyroidism had anti-microsomal antibodies and it is well accepted that Hashimoto's disease is a risk factor for the development of iodine-induced hypothyroidism.

As a general rule, careful thyroid examination and assays of TSH and thyroid hormones should be performed before initiating amiodarone treatment. Adding a test for the detection of TPO Abs will detect individuals at risk of exa-cerbation of an undiagnosed autoimmune disorder. Faced to low but positive results, gradual increases in the amiodarone posology under close monitoring of thyroid function will prevent severe complications. The same remarks apply by other iodine-containing drugs, such as benziodarone and radiology contrast agents.

The determination of Tg and TPO antibodies in differentiated thyroid cancers has no intrinsic clinical value. Tg Ab determination is of interest, as previously discussed, to validate Tg determinations. Thyroid autoantibodies are often detected in these patients, mainly, as expected in females and may represent either a reaction of the immune system against the tumor cells or the coexistence of a thyroid autoimmune disorder and of a carcinoma.

In a retrospective study of 9237 surgical specimens, Hirabayashi and Lindsay observed that 27% female patients (vs 6% of males) had evidence of Hashimoto's disease both inside and outside the tumor (32). Nevertheless, it has never been possible to associate Graves' or Hashimoto's disease with an increased incidence in thyroid carcinoma.

Tg and TPO Abs are routinely assayed since years. Besides specific clinical indications, summarized in Table III, they are not, at this stage of knowledge, of particular significance. Low concentrations in asymptomatic subjects, specially elderly women, are not justifying any treatment.

All laboratory tests detect an heterogeneous population of antibodies, some of them probably not involved in the development of thyroid autoimmune disorders. Even when they are involved, it is still unclear whether these antibodies play an active role in the pathogenesis of thyroid autoimmunity or if they simply reflect the development of these diseases. However, since both antibodies are directed against molecular entities playing a key-role in the physiological secretion of thyroid hormones, it is not a surprise that their presence at high concentrations is invariably associated with thyroid disorders.

In the view of most clinicians, assessment of TSH and thyroid hormone circulating levels are more important than thyroid antibody determination. This is quite logical since thyroid autoimmunity can easily be treated but not cured by supplementation of the deficient hormonal secretion or by antithyroid drugs. However, if new treatments become available to cure or to treat the cause i.e. a deregulation of the immune system instead of treating the effects, the situation may entirely change.

   

Clinical situation	Interest of Tg and TPO Ab determination and interpretation
Clinical suspicion of Graves' disease without thyroid or ocular signs	Absence of Tg and TPO Abs together with normal thyroid function tests rule out Graves' disease and suggest other causes (diabetes, pheochromocytoma,...).
Eyelid changes suggestive of Graves' disease	Subnormal or elevated Tg and/or TPO Abs suggest further thyroid function tests and point out towards thyroid eye disease. Negative findings suggest further exploration (orbital tumors).
Suspicion of hypothyroidism in the adult with thyroid enlargement	Elevated Tg and/or TPO Abs point towards Hashimoto's disease even if other thyroid tests are normal. Negative findings suggest other goitrogenic causes (chemicals, drugs, de Quervain's thyroiditis).
Clinical suspicion of hypothyroidism in the adult without thyroid enlargement	Elevated Tg and/or TPO Abs suggest regular monitoring of thyroid function; negative findings suggest a misdiagnosis and other explorations (e.g. renal disease).
Suspicion of hypothyroidism in school children	Moderate level of TPO Abs confirm juvenile lymphocytic thyroiditis.
Pregnancy - asymptomatic	Elevated TPO Abs represent a risk factor for post partum thyroiditis. The test has to be performed in early pregnancy.
Pregnancy - overt hyperthyroidism in patients without previous thyroid dysfunction (rare)	Elevated TPO and Tg Abs levels establish the autoimmune origin of the disease.
Thyroid cancers	Elevated Tg Abs are common in differentiated cancer. They have no clinical value but they suggest caution in the interpretation of serum thyroglobulin determinations.
Amiodarone treatment	Elevated TPO Abs before treatment suggests a risk of developing iodine-induced hypothyroidism so that the test should be included in the initial laboratory workup of these patients.

Table III: Interest of Tg and TPO Ab determinations.

References