Best Practice & Research Clinical Endocrinology & Metabolism Vol. 19, No. 1, pp.

1732, 2005
doi:10.1016/j.beem.2004.11.004 available online at http://www.sciencedirect.com

2 Non-thyroid autoantibodies in autoimmune thyroid disease

Anthony P. Weetman*
Professor of Medicine Clinical Sciences Centre, University of Shefeld, Northern General Hospital, Shefeld S5 7AU, UK MD, DSc, FRCP

Autoimmune thyroid disease is frequently accompanied by other organ-specic and non-organspecic diseases, most likely because there is sharing of genetic and possibly environmental susceptibility factors. These associations are well recognized in the autoimmune polyglandular syndromes; autoimmune thyroid disease is one of the three major endocrinopathies in the type 2 syndrome and occurs in around 4% of type 1 patients. This review considers the frequency of disease-specic autoantibodies in patients with thyroid autoimmunity and briey examines the role of such antibodies in performing screening for the associated conditions. Recommendations are made for using such autoantibody tests in the setting of patients with autoimmune thyroid disorders, and also for the utility of screening for thyroid autoimmunity in patients with pernicious anaemia, Addisons disease, coeliac disease, primary biliary cirrhosis, myasthenia gravis, lymphocytic hypophysitis, systemic lupus erythematosus and rheumatoid arthritis. At present, however, there are no large-scale trials that have shown the costbenet ratio of autoantibody screening for autoimmunity screening, and clinicians must use individual judgement combined with heightened awareness to identify who to test. Key words: autoimmune hypothyroidism; Graves disease; autoantibodies; pernicious anaemia; Addisons disease; coeliac disease; primary biliary cirrhosis; myasthenia gravis; lymphocytic hypophysitis; systemic lupus erythematosus; rheumatoid arthritis.

In any medical textbook chapter on autoimmune thyroid disease, there will inevitably be a list of conditions which are more frequently detected in such patients than would be expected by chance. Typically, the list comprises most of the other organ-specic autoimmune disorders (with notable exceptions such as multiple sclerosis, despite the higher than expected prevalence of autoimmune thyroid disease in the rst-degree relatives of such patients).1 A varying number of non-organ-specic conditions usually gure in such lists as well, and there is generally a statement that such associations

* Tel.: C44 114 271 4160; Fax: C44 114 261 8775. E-mail address: a.p.weetman@shefeld.ac.uk 1521-690X/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved.

18 A. P. Weetman

reect common genetic and possibly environmental susceptibility factors at work, in concert with additional disease-specic factors that determine the clinical pattern. Such associations are most pronounced in the autoimmune polyglandular syndromes (APS). The type 1 syndrome is an autosomal recessive condition, now identied as being caused by mutations in the AIRE (autoimmune regulator) gene which encodes a transcription coactivator responsible for ensuring thymic presentation of autoantigens during development, in a way that normally allows autoreactive T cells to become tolerized. Only around 4% of such patients develop autoimmune thyroid disease, and this is never the rst manifestation of the syndrome.2 APS type 2 is much more common and is usually dened as the presence of two of the following disorders: Addisons disease, autoimmune thyroid disease and type 1 diabetes mellitus; however, some have reserved the name for those with Addisons disease plus one of the other two components.3 Typical features are show in Table 1. Patients, and often family members, with one of these syndromes are usually aware of the possibility that there is an increased risk of additional autoimmune problems and often ask for screening tests. It is far less clear whether there are any benets in screening for other autoimmune disorders in patients with apparently isolated autoimmune hypothyroidism or Graves disease. This clinical uncertainty is compounded by the commonplace observation that many such patients continue to have complaints despite the restoration of the euthyroid state, raising the anxiety that a subtle, coincident autoimmune disease may be present. The aim of this review is to focus on the frequency of non-thyroid autoantibodies in patients with autoimmune thyroid disease (Table 2), and at the same time to provide a brief analysis of the utility of such autoantibodies as markers on an underlying disorder. It is based on a literature search performed using PubMed and keywords autoimmune thyroid disease and either the associated disease or autoantibody. Diabetes and the related autoantibodies are considered in Chapters 7 and 8 and will not be dealt with here. Many papers describe isolated cases or families with multiple autoimmune diseases, or report disease associations rather autoantibody data. Moreover, the length of this review is such that it cannot include more than a small fraction of all the available studies; priority has been given to reviews of older studies and those studies which are large, recent, and have used appropriate control groups, as these have utilized the most contemporary autoantibody assays. Because autoimmune thyroid disease is so common compared to many of the other associated disorders (Table 2), the most informative studies are often those focussing on the associated disease rather than
Table 1. Typical features of autoimmune polyglandular syndrome type 2.3 Prevalence Inheritance Female:male ratio Peak age of onset Key endocrinopathies 1 in 20 000 Autosomal dominant; variable penetrance; strong association with HLA-B8, DR3 3:1 3040 years Autoimmune thyroid disease (75%) Type 1 diabetes (55%) Addisons disease (45%) Vitiligo (20%) Alopecia areata (6%) Pernicious anaemia (5%)

Other major associations

Non-thyroid autoantibodies in autoimmune thyroid disease 19

Table 2. Population prevalence and autoantibody patterns in diseases associated with autoimmune thyroid disease. Prevalence (%) Pernicious anaemia Addisons disease Lymphocytic hypophysitis Systemic lupus erythematosus Rheumatoid arthritis Primary biliary cirrhosis Coeliac disease Myasthenia gravis 0.130.15 0.005 !0.005 0.020.03 0.550.86 0.0040.01 0.11.0 0.0050.015 Key autoantibody HC/KC ATPase in gastric parietal cell 21-hydroxylase ? a-enolase Double-stranded DNA plus many other DNA and RNA components Rheumatoid factor; cyclic citrullinated peptide E2 component of pyruvate dehydrogenase complex Tissue transglutaminase Acetylcholine receptor

a series of patients with thyroid autoimmunity.4 However, it is not always clear how such patients are recruited, and biases may creep in which could alter the frequency of thyroid autoimmunity compared to an unselected series. Any recommendations made for clinical practice reect these limitations. At present there is no consensus on who to screen for associated autoimmune disease or at what frequency. Equally important, there has been no costbenet analysis of testing for autoantibodies in patients with autoimmune thyroid disease; in the end any increased testing in one area of practice will diminish the resource available in another area.

PARIETAL CELL (PC) AND INTRINSIC FACTOR (IF) ANTIBODIES Diagnostic use These antibodies are markers for pernicious anaemia, which is characterized by vitamin B12 deciency, autoimmune chronic gastritis affecting the fundus and body but not the antrum of the stomach, and a progressive course that may last 2030 years.5 In one survey of people over 60 years old, the prevalence of undiagnosed pernicious anaemia was 1.9%, although those of Latin, American or Asian origin are less commonly affected.6 The frequency of pernicious anaemia is twice as high in women. Gastric parietal cells (PCs), which are not present in the antral region, as well as the intrinsic factor (IF) they produce, are lost in pernicious anaemia. The gastric HC, KC-ATPase is the target autoantigen in the PC.7 This enzyme provokes the production of autoantibodies which can x complement in vitro but as with thyroid peroxidase (TPO) autoantibodies in autoimmune thyroid disease, there is still uncertainty as to any pathogenic role such antibodies play. In murine models of pernicious anaemia the key pathogenic agents are HC, KC-ATPase-specic CD4C T cells.5 PC antibodies are found in 7095% of patients with pernicious anaemia and 3060% of those with only atrophic gastritis.8 They also occur in around 30% of non-anaemic rst-degree relatives of patients with pernicious anaemia.5 There are two classes of antibodies against IF: blocking (or type I) antibodies recognize the vitamin B12 binding site of IF, thereby preventing the transport of the vitamin from the stomach to its absorption site in the terminal ileum, and binding

20 A. P. Weetman

(or type II) antibodies which do not interfere with vitamin B12 transport. They are found in around 70 (type I) and 34% (type II) of patients with pernicious anaemia (use of new assays may increase the frequency of type II antibodies) but they are rare in autoimmune gastritis without accompanying pernicious anaemia, therefore making them specic but insensitive markers of disease.8 It will be seen that neither PC nor IF antibodies are sufciently sensitive or specic for diagnosis of those with pernicious anaemia. When vitamin B12-deciency is suspected in general clinical practice, vitamin B12-measurement is the most logical rst step, supported by other biochemical tests such as raised homocysteine or methylmalonic acid levels: GP and IF antibodies provide ancillary information, whereas gastroscopy provides both diagnostic and causal information. Association with autoimmune thyroid disease In a review of three series of patients with Graves disease (nZ1333 in total), dating from 1962 to 1971, the prevalence of pernicious anaemia ranged from 1.7 to 2.3%, compared to 4.212.3% in three series of 490 patients with Hashimotos thyroiditis and autoimmune hypothyroidism.8 These gures exceed by more than 10-fold the weighted prevalence gure for pernicious anaemia in the USA in 1996, which was estimated to be 0.15%,9 and the closely matching gure of 0.13% in the UK derived from an extensive literature review.1 The prevalence of IF antibodies in the Graves group ranged from 2.6 to 4.7% and was from 5.0 to 6.7% in the Hashimoto and autoimmune hypothyroidism patients; PC antibodies were measured in only two studies and were found in 22% of Graves and 32% of Hashimoto patients.8 In 54 women with postpartum thyroiditis, PC antibodies were found in a third during pregnancy, rising 2- to 9-fold in 10 of the 18 in the postpartum period.10 Two of the antibody-positive women developed pernicious anaemia during a 5-year follow-up. Three recent series have added new insights. In 62 sequential patients with autoimmune thyroid disease (10 with Graves disease, 47 with Hashimotos thyroiditis, and ve with atrophic thyroiditis), 35% (nZ22) had hypergastrinaemia, in all of whom atrophic gastritis was conrmed at biopsy: 10 of these 22 also had pernicious anaemia.11 PC antibodies were found in 40% of the entire group, but these comprised only 15 (68%) of the 22 with atrophic gastritis; the remaining 10 PC-antibody-positive patients had normal gastroscopy and histology. Only seven of 22 patients with atrophic gastritis had simultaneous atrophic gastritis, pernicious anaemia and PC antibodies. In the second series, 63 randomly selected patients concordant for both type 1 diabetes and autoimmune thyroid disease were screened for pernicious anaemia using measurement of serum vitamin B12. One patient had already had this diagnosis made, and three further asymptomatic patients were found, all of whom were IF-antibody-positive.12 Although this prevalence of pernicious anaemia (6.3%) was suggested to be higher than that in type 1 diabetes alone (around 4%), the lack of a contemporary control group makes this difcult to evaluate fully. However, in 497 type 1 diabetes patients, PC antibodies were found in 20.9% (of whom 10.5% had pernicious anaemia), increasing with age, and TPO antibodies were found in signicantly more of the PC-antibodypositive subset, 33.6 versus 22.4% in the PC-antibody-negative group, supporting the increased strength of association in the presence of thyroid autoimmunity.13 Looked at the other way, autoimmune thyroid disease is particularly common in pernicious anaemia; for example, in 22 such patients, 50% had TPO antibodies and 13.6% had thyroglobulin antibodies, 23 times higher than in matched controls.14

Non-thyroid autoantibodies in autoimmune thyroid disease 21

These results support the idea that screening patients with pernicious anaemia for thyroid autoimmunity is probably worthwhile. It is also likely that those with type 1 diabetes mellitus and thyroid autoimmune disease are at increased risk of pernicious anaemia, and screening might be worth considering, although there is as yet no clear evidence that treatment of subclinical disease is benecial, nor is it clear whether measuring PC antibodies13, gastrin11 or vitamin B1212 is the best initial test. The low specicity of PC antibodies and the low sensitivity of IF antibodies mean that neither are good tools to screen for pernicious anaemia in cases of isolated autoimmune thyroid disease; vigilance, especially in the elderly or those with an anaemia11, leading to appropriate testing in those with clinical symptoms, is probably the best policy.

ADRENAL ANTIBODIES Diagnostic use This topic is fully covered in Chapter 6 and therefore only a brief summary is provided here. The key autoantigen for antibodies in Addisons disease is the adrenal-specic cytochrome 21-hydroxylase (21-OH); some patients also have autoantibodies against side-chain cleavage enzyme or 17a-hydroxylase, but these are most frequent in patients with APS type 1 and correspond to the cross-reactive steroid cell autoantibodies detected by immunouorescence.15 Concordance rates of only 80% have been reported, but provided care is taken with technique, immunouorescence methods to demonstrate the presence of adrenal cortex-specic autoantibodies can give equivalent results to immunoprecipitation assays using recombinant 21-OH. Using either method, around 80% of patients with autoimmune Addisons disease have detectable antibodies, whereas those with non-autoimmune Addisons diseases are antibody-negative.16 Subclinical Addisons disease is characterized by the presence of adrenal cortex or 21-OH antibodies without biochemical evidence of adrenal failure. Such a situation most commonly arises when patients with an organ-specic autoimmune disorder are tested for adrenal antibodies. Of 19 such patients studied using the cortisol response to CRH to assess subtle adrenal dysfunction, eight had an impaired response and two developed adrenal failure within 1 year.17 Thus, the measurement of adrenal or 21-OH antibodies provides a very useful test to determine the aetiology of primary adrenal failure, and the presence of these antibodies in organ-specic autoimmunity predicts a signicant risk of developing Addisons disease. Those with the highest antibody levels are at the greatest risk of progression.18 Autoimmune Addisons disease may occur in the absence of adrenal autoantibodies, and not all adrenal antibody-positive individuals progress to overt disease, although once found to have such antibodies, annual followup is indicated to assess adrenal reserve. Association with autoimmune thyroid disease An early survey of 88 children with autoimmune thyroid disease found that 2.3% had adrenal cortex antibodies, compared to none of 117 controls.19 More recently, a huge series of 4353 adult patients with autoimmune thyroid disease was reported, and although the individual thyroid disorders were not specied, 1.0% had adrenal cortex antibodies and 0.1% steroid cell antibodies.20 Twenty three patients with Graves disease and 18 patients with Hashimotos thyroiditis who had adrenal cortex antibodies

22 A. P. Weetman

were available for follow-up. All but two of these also had 21-OH antibodies. Four of the Graves patients and two with Hashimotos thyroiditis (one with additional type 1 diabetes) developed overt Addisons disease within 1.56 years. These results, and those in other patients with various organ-specic autoimmune diseases, indicate that adrenal cortex/21-OH antibodies are markers of a signicant (but modest) risk of progression to clinical Addisons disease, although longer-term follow-up is clearly indicated to assess life-long risk, and children with thyroid autoimmunity may have a higher risk than adults of developing adrenal antibodies. Looked at the other way, patients with autoimmune Addisons disease have a high prevalence of thyroid autoantibodies and autoimmune thyroid disease. A case ascertainment survey in the UK of 81 patients found that 20 (25%) had primary hypothyroidism and nine (11%) had Graves disease.21 Similar gures were reported in 125 patients from Poland, where 15 (12%) had thyrotoxicosis, 17 (14%) hypothyroidism and six (5%) subclinical hypothyroidism.22 In 61 Norwegian Addisons patients without a previous history of clinical thyroid disease, seven (11%) had TPO and/or TG antibodies.23 There is little doubt from these gures that annual screening of patients with autoimmune Addisons disease for autoimmune thyroid disease is worthwhile. There is no evidence to identify the best strategy. I perform TPO antibody measurement on all newly diagnosed patients to assess future risk, advising those who are antibody-positive with normal TSH levels about symptoms of thyroid disease and the relation of this to pregnancy. Thereafter, I advise all patients to have an annual measurement of TSH, On the other hand, there is little to support routine screening of patients with autoimmune thyroid disease for Addisons disease or to measure adrenal cortex/21-OH antibodies in this setting. Suggestive clinical symptoms of course warrant follow up; the rather slow tempo of disease evolution is somewhat reassuring in those at highest risk, although of course Addisons disease remains a life-threatening condition.

PITUITARY ANTIBODIES Diagnostic use Lymphocytic hypophysitis is a rare condition, almost certainly autoimmune in origin, in which the pituitary is inltrated by lymphocytes and transient or permanent hypopituitarism results.24 By 2003 over 200 cases had been reported, but it remains difcult to diagnose, and we have no clear picture yet of the entire spectrum of disease or how representative these cases are, as few centres have any extensive experience. Attempts to nd autoantibodies were prompted by the need to prove the putative autoimmune aetiology and to provide a tool that might spare patients biopsy, which has been the denitive diagnostic method to date. Variable results have been reported using complement consumption and immunouorescence assays.24 Apart from isolated ACTH deciency, in which antibodies are directed against a secretory granular protein which is neither POMC or ACTH25, the autoantigens involved have remained elusive, hampering efforts to dene the diagnostic utility of antibody testing. Immunoblotting studies have revealed that autoantibodies reactive with a 49 kDa pituitary protein could be detected in seven of 10 patients with biopsy-proven lymphocytic hypophysitis, but were also present in ve of 52 healthy controls.26 Microsequencing subsequently demonstrated that the protein was the ubiquitous

Non-thyroid autoantibodies in autoimmune thyroid disease 23

a-enolase.27 The demonstration of a-enolase antibodies (by immunoblotting) in 12 (63%) of 19 patients with Sheehans syndrome28 certainly proves that these antibodies are not specic markers for lymphocytic hypophysitis, and their sensitivity and utility in the diagnostic evaluation of this condition is similarly equivocal. Association with autoimmune thyroid disease One of the indications that lymphocytic hypophysitis is an autoimmune disease is its association with other known autoimmune disorders including thyroid disease.24 In one series of 25 patients with the probably separate condition of isolated ACTH deciency, hypothyroidism occurred in 14 (56%) and Graves disease in four (16%).29 The occurrence of autoimmune pituitary disease in autoimmune thyroid disease is inevitably rare, even if probably higher than expected. Antibodies to a-enolase were detected in four (19%) of 21 patients with Hashimotos thyroiditis and one (8%) of 12 Graves patients, compared to 10% of controls.26 However, using immunouorescence to detect pituitary antibodies, 12 (20%) of 59 patients with Graves disease and 22 (28%) of 80 patients with Hashimotos thyroiditis were positive, and four of the Hashimoto group with the highest antibody levels had isolated growth hormone deciency.30 These striking ndings have not yet been replicated, and it is premature to recommend such antibody testing until there has been replication and the development of robust assays, following autoantigen identication.

ANTINUCLEAR (AN) ANTIBODIES Primarily a diagnostic marker for systemic lupus erythematosus (SLE), the measurement of AN antibodies is primarily a tool for rheumatologists rather than endocrinologists, and the reader is referred elsewhere for guidelines on the use of this test.31,32 Sufce it to say here that the introduction of the human Hep-2 cell line as a routine substrate for measuring AN antibodies, instead of the heterogeneous tissues and cells used previously, has improved sensitivity and reduced uncertainty in the interpretation of this test. Nonetheless, low-titre AN antibodies are a frequent nding in subjects without signicant rheumatological disease and their frequency increases with age, leading to the stark statement that AN antibodies should not be tested in patients without clinical evidence of a rheumatological disorder.33 Association with autoimmune thyroid disease The increased prevalence of autoimmune thyroiditis in SLE is compelling evidence that the organ/non-organ-specic dichotomy of autoimmune diseases is more apparent than real. At least one reason for this association is that both disorders share HLA-DR3 as a susceptibility factor. In one series, 21 (51%) of 41 SLE patients had TPO and/or TG antibodies compared to 11 (27%) of 41 controls, and twice as many SLE patients had subclinical hypothyroidism.34 Similar gures for subclinical hypothyroidism have been reported subsequently, with thyroid antibodies appearing, as expected, more frequently in older patients.35,36 In another series of 300 SLE patients that had no matched controls, 5.7% had hypothyroidism, but only 1.7% had hyperthyroidism, the latter probably being only slightly in excess of the expected value.37 Such data support

24 A. P. Weetman

the view that patients with SLE should be examined for thyroid disease and that intermittent screening is may be warranted. However, there is no evidence to support the use of AN antibody testing in thyroid patients without suggestive features of a rheumatological disorder, given the frequent occurrence of these non-specic markers. For instance, AN antibodies were found in 26% of 50 patients with autoimmune thyroid disease, compared to 8% of 100 controls, and the prevalence of antibodies to smooth muscle and single-stranded DNA was also increased.38 However, none of these patients had antibodies to the more diseasespecic autoantigens, namely double-stranded DNA, extractable nuclear antigen, SS-A or SS-B. Others have also found an increased frequency of AN antibodies, but heavily dependent on assay method, in Graves disease and Hashimotos thyroiditis, and also not associated with rheumatological disease.39,40 Cardiolipin antibodies, found in the primary antiphospholipid syndrome as well as SLE, are another example of such nonspecic markers, in one series occurring in 55% of patients with autoimmune thyroid disease.41 Recently, it has been suggested that AN antibodies are more frequent in autoimmune hypothyroidism caused by atrophic thyroiditis rather than Hashimotos thyroiditis42; these interesting ndings require conrmation in a larger cohort of patients.

RHEUMATOID FACTOR (RF) Diagnostic use RF is an autoantibody that binds to the Fc portion of IgG and is found in around 80% of patients with rheumatoid arthritis (RA), particularly in those with the most aggressive disease.43 Rheumatoid factor (RF) is potentially pathogenic but only one of many such autoantibodies in RA; antibodies against citrulline-containing proteins may be more disease-specic, having negative predictive values of 75% and positive predictive values of up to 100%.44 However, normally measured RF is relatively non-specic and insensitive as a diagnostic or predictive tool: RF occurs in the presence of immune complex disorders, such as bacterial endocarditis, and in response to large concentrations of recall antigens. B cells from healthy individuals can be induced to produce low-afnity RF by EpsteinBarr (EB) virus transformation, and around 5% of healthy individuals have RF, a percentage that increases with age.33 As with AN antibodies, the recommendation is that RF should not be tested for in the absence of any features of a rheumatological disease.32,33 Association with autoimmune thyroid disease Several series have reported an increased frequency of autoimmune thyroid disease in RA. For instance, in a series of 58 multicase UK families with RA, 6% of the patients had thyroid disease, and 5% of the men and 15% of the women had TPO antibodies.45 In a controlled series of 101 RA patients from Greece, 12.9% had TPO antibodies, compared to 8.6% of controls46, and similar ndings have been made in Norway and Quebec.47,48 A recent series of 25 RA patients from Brazil found that 32% had TG and/ or TPO antibodies compared to 4% of 113 controls49, although in a similar series of 64 UK patients, the frequency of TPO antibodies was only 11%.36 Overall, these data support an association between RA and autoimmune thyroid disease, which is probably

Non-thyroid autoantibodies in autoimmune thyroid disease 25

not as strong as the association of the latter with SLE. Indeed, it seems likely that reports of an association between polymorphisms of the CTLA-4 gene and RA have been confounded by the association between RA and autoimmune thyroid disease strongly linked to this polymorphism.50 Given the high prevalence of RA in the general population (Table 2), it is debatable whether screening for autoimmune thyroid disease is warranted; screening with TSH levels alone may be inadequate, especially in patients taking corticosteroids.51 At the least, this association should be borne in mind in patients with RA who have new symptoms. On the other hand, the poor predictive value of RF antibodies does not justify their use in routine screening of patients with autoimmune thyroid disease.

ANTIMITOCHONDRIAL (AM) ANTIBODIES Diagnostic use AM antibodies react with the E2 components of the 2-oxo-acid dehydrogenase complexes in mitochondria, especially the E1a component of the pyruvate dehydrogenase complex, and are present in most patients with primary biliary cirrhosis (PBC).52 There is no evidence for a pathogenic role for AM antibodies. The prevalence of PBC has probably been underestimated, and the condition is found in around 0.01% of the population, especially women aged over 40.53 AM antibodies can be detected long before clinical disease develops, and their presence is strongly associated with progression to PBC. They are relatively disease-specic, although only a third of unselected patients with AM antibodies will actually have PBC at the time of detection.54 Up to 15% of PBC patients may be consistently AM antibody-negative, and even more if standard immunouorescence assays are used.55 Thus, the detection of AM antibodies, particularly by immunoblotting or enzyme-linked immunosorbent assay (ELISA), is a rather sensitive and specic marker for established PBC or a strong risk of developing this disorder. Association with autoimmune thyroid disease Uncontrolled studies have shown an excess of autoimmune thyroid disease in PBC; for instance, 13 (22%) of 58 patients with PBC had hypothyroidism in one series, 34% had TPO antibodies and 20% TG antibodies.56 More recently, the frequency of unspecied clinical autoimmune thyroid disease was reported to be 21% in 200 PBC patients, and 34% had thyroid autoantibodies; unlike RA, this study also showed that CTLA-4 polymorphisms were associated with PBC in the absence of autoimmune thyroid disease.57 In 351 sera selected for TPO antibody positivity, AM antibodies were found in seven (2%), six of whom were asymptomatic and four had normal liver function tests.58 This study has not been replicated, and it is questionable whether there are any advantages to picking up AM antibody-positive subjects amongst thyroid patients, since treatment for PBC is largely symptomatic and no disease-modifying agents exist which can be used in the asymptomatic phase.53 Given the high frequency of autoimmune thyroid disease in PBC, however, it would seem worthwhile screening such patients with TPO antibodies and TSH, as for patients with Addisons disease.

26 A. P. Weetman

TISSUE TRANSGLUTAMINASE (TTG) ANTIBODIES Diagnostic use Endomysial antibodies, detected by immunouorescence, are found in most patients with coeliac disease, having a sensitivity of 8598% and specicity of 97100%.59 TTG is the autoantigen recognized by these antibodies, and newer assays based on guinea pig or recombinant human TTG can be more sensitive but are less specic than endomysial antibodies. Tests for gliadin antibodies have moderate sensitivity but poor specicity. For example, in one recent series the sensitivity and specicity of endomysial antibodies in the diagnosis of coeliac disease was 90 and 98%, respectively; the corresponding values for TTG antibodies were 86 and 84%, and for gliadin antibodies 76 and 79%.60 Coeliac disease is more common than previously thought, as milder forms of the disease have come to be recognized.59 Using both endomysial and TTG antibody screening of 3654 Finnish students, with small bowel biopsy of those who were positive, 1.5% had a positive antibody test and 66% of these had coeliac disease on biopsy, giving a prevalence of one in 99.61 It is unclear whether screening for such a common disease is worthwhile; although there are long-term risks from untreated coeliac disease, exclusion of dietary gluten is difcult in an asymptomatic individual. Association with autoimmune thyroid disease The prevalence of autoimmune thyroid disease is increased in coeliac disease. For instance, in one series 20.5% of coeliac patients had this complication, compared to 11.2% of controls.62 Dermatitis herpetiformis is accompanied by usually asymptomatic coeliac diseases; in 115 such patients, thyroid antibodies were found in 48%, compared to 16% of controls, and 10% had thyroid dysfunction.63 Several studies have shown an increase in coeliac disease in thyroid autoimmunity (Table 3). Those with hypothyroidism who have undiagnosed coeliac disease, as shown by the presence of endomysial antibodies, have higher thyroxine requirements than patients who are antibody-negative.67 However, another study of 48 patients with Graves disease found only one with endomysial antibodies, in whom intestinal biopsy was normal and the antibodies disappeared after 3 years.68 How common such transient antibodies are in autoimmune thyroid disease is unclear. In summary, the frequency of autoimmune thyroid disease in coeliac disease and dermatitis herpetiformis is high enough to consider screening such patients routinely

Table 3. Frequency of coeliac disease or endomysial antibodies in autoimmune thyroid disease (AITD). AITD n Coeliac disease 172 83 220 Endomysial antibodies 169 1.2 1312 0.5 [67] 3.4 4.8 3.2 4000 249 250 0.3 0.4 0.4 [64] [65] [66] % Controls n % Reference

Non-thyroid autoantibodies in autoimmune thyroid disease 27

with TPO antibodies and TSH. The increased awareness of common, latent coeliac disease in the general population has made testing for this more common in patients with gastrointestinal or other unexplained symptoms. Although there have been recommendations for routine screening65,66, I suggest the data so far indicate that TTG/ endomysial antibody testing is best reserved for patients with thyroid disease who have any clinical features suggestive of coeliac disease, including osteoporosis and increasing or uctuating thyroid hormone dosage requirements.

ACETYLCHOLINE RECEPTOR (ACHR) ANTIBODIES Diagnostic use Myasthenia gravis is uncommon, with a prevalence of around ve per 100 0009, although some estimates are higher.1 Diagnosis usually depends on the clinical history, electromyography and the response to edrophonium, supported by the presence of AChR antibodies, which are found in around 85% of patients.69 Only 50% of patients with ocular myasthenia have AChR antibodies. The antibodies are undoubtedly pathogenic, causing symptoms after transplacental transfer or injection into mice. Seronegative myasthenia gravis is due to the presence of antibodies against musclespecic kinase or other autoantigens. Association with autoimmune thyroid disease The frequency of autoimmune thyroid disease is increased in myasthenia gravis. For instance, in 124 cases of myasthenia gravis studied in Belgrade from 1983 to 1992, thyroid autoimmune disease occurred in 7.3%70, similar to previously published gures from Denmark.71 Myasthenia associated with autoimmune thyroid disease has a milder clinical expression, preferential ocular involvement and hence a lower frequency of AChR antibodies.72 In this Italian study. the prevalence of autoimmune thyroid disease was 28.5% in the myasthenic patients: 7.7% with Graves disease, 10.9% with euthyroid Hashimotos thyroiditis and 9.9% with autoimmune hypothyroidism. Much lower gures have been reported recently from Austria, challenging the dogma of a strong association73, but the most recent survey, comprising 337 consecutive myasthenia patients from Athens, found that 39 (11.6%) had thyroid abnormalities, although unfortunately these were not detailed.74 It would be very difcult to demonstrate an increased frequency of myasthenia gravis in autoimmune thyroid disease, given its rarity. However, AChR antibodies may be detected in a small proportion (8%) of patients with thyroid-associated ophthalmopathy, in the absence of myasthenia gravis.75 This suggests a sharing of autoantigenic targets in the two disorders.76 However, there is no justication for testing for AChR antibodies in patients with any form of autoimmune thyroid disease unless there are clinical reasons for suspecting myasthenia gravis.

SUMMARY Autoimmune thyroid disease is frequently associated with organ-specic and nonorgan-specic autoimmune disorders, and less common associations than those

28 A. P. Weetman

Table 4. Summary of autoantibody frequency in patients with autoimmune thyroid disease and recommendations for screening such patients. Antibody Parietal cell Intrinsic factor Adrenal cortex/21-OH Pituitary/a-enolase Antinuclear Rheumatoid factor Mitochondrial Tissue transglutaminase/endomysial Acetycholine receptor
a b c

Prevalence (%) 2240 37 12 828 26 w5 2 12 !1c

Screening recommended Noa Not unless other features suggestive of pernicious anaemia No No No No No Nob No

Worthwhile in patients with concurrent type 1 diabetes. Perform in patients with osteoporosis or unusual thyroxine requirements. Eight percent in patients with Graves ophthalmopathy.

Table 5. Summary of TPO and TG autoantibody frequency in various conditions and recommendations for screening in these conditions. Condition Pernicious anaemia Addisons diseasea Lymphocytic hypophysitis Isolated ACTH deciency Systemic lupus erythematosus Rheumatoid arthritis Primary biliary cirrhosis Coeliac disease Myasthenia gravis
a b

Prevalence of TPOCTG antibodies (%) Up to 50 11 w10 O50 50 1132 34 2048 728

Screening recommended Yes Yes Yes Yes Nob No Yes Yes Yes

Not associated with APS type 1 or 2. More studies needed; may be worthwhile in older patients.

considered here are reviewed elsewhere.4 As a result there are increased frequencies of many non-thyroid autoantibodies in thyroid patients. It is important to recall that such autoantibodies are usually non-pathogenic and often mark a subclinical phase of disease. An overview of the data considered in this chapter is presented in Tables 4 and 5, along with tentative recommendations for screening. These recommendations take into account the frequency of each disease. No studies have yet been performed which show the costbenet of any of these recommendations. Some testing, such as in Addisons patients, seems so obvious as to never be challenged, while others are more contentious. However, as treatments for each of these conditions improve, and high throughput, inexpensive autoantibody assays become established, the situation will undoubtedly change; now is the time to be assembling long-term, large studies of outcome to assess the practicalities and benets of such screening.77

Non-thyroid autoantibodies in autoimmune thyroid disease 29

Practice points most of the diseases associated with autoimmune thyroid disease are uncommon, and antibody screening is not recommended an exception may be pernicious anaemia; a high index of suspicion or screening with gastric parietal cell antibodies is warranted in autoimmune thyroid disease, especially in combination with type 1 diabetes mellitus increasing or uctuating thyroxine requirements in hypothyroidism may indicate the presence of coeliac disease primary biliary cirrhosis is more common than usually thought, and antimitochondrial antibodies are common in asymptomatic patients routine testing of autoantibodies associated with rheumatological disorders is not recommended in the absence of signs or symptoms

Research agenda there is still a dearth of large surveys of autoimmune disease associations which are clearly free of biases and confounding factors no costbenet analyses of the value of screening for diseases associated with autoimmune thyroid disease have been performed it remains unclear how clinically useful autoantibody testing is in identifying individuals with a long prodromal phase to an autoimmune disorder development of rapid and inexpensive autoantibody assays based on protein arrays offers new potential to approach these issues the recently reported high prevalence of pituitary autoantibodies in autoimmune thyroid disease, accompanied by growth hormone deciency, requires replication

REFERENCES
1. Broadley SA, Deans J, Sawcer SJ et al. Autoimmune disease in rst-degree relatives of patients with multiple sclerosis. A UK survey. Brain 2000; 123: 11021111. 2. Peterson P, Pitkanen J, Sillanpaa N et al. Autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED): a model disease to study molecular aspects of endocrine autoimmunity. Clin Exp Immunol 2004; 135: 348357. *3. Dittmar M & Kahaly GJ. Polyglandular autoimmune syndromes: immunogenetics and long-term follow-up. Clin Endocrinol Metab 2003; 88: 29832992. *4. Jenkins RC & Weetman AP. Disease associations with autoimmune thyroid disease. Thyroid 2002; 12: 977988. 5. Toh B-H, Van Driel IR & Gleeson RA. Mechanisms of disease: pernicious anemia. N Engl J Med 1997; 337: 14411448. 6. Carmel R. Prevalence of undiagnosed pernicious anemia in the elderly. Arch Intern Med 1996; 156: 1097 1100. 7. Karlsson FA, Burman P, Loof L et al. Major parietal cell antigen in autoimmune gastritis with pernicious anemia is the acid-producing HC,KC-adenosine triphosphatase of the stomach. J Clin Invest 1988; 81: 475479.

30 A. P. Weetman 8. Burman P, Ma J-Y & Karlsson FA. Autoimmune gastritis and pernicious anaemia. In Weetman AP (ed.) Endocrine Autoimmunity and Associated Conditions. Dordrecht: Kluwer Academic Publishers, 1998, pp. 243267. *9. Jacobson DL, Gange SJ, Rose NR et al. Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol 1997; 84: 223243. 10. Burman P, Kampe O, Kraaz W et al. A study of autoimmune gastritis in the postpartum period and at a 5-year follow-up. Gastroenterology 1992; 103: 934942. *11. Centanni M, Marignani M, Gargano L, et al. Atrophic body gastritis in patients with autoimmune thyroid disease: an underdiagnosed association. Arch Intern Med 1999; 159:17261730. 12. Perros P, Singh RK, Ludlam CA et al. Prevalence of pernicious anaemia in patients with type 1 diabetes mellitus and autoimmune thyroid disease. Diabet Med 2000; 17: 749751. 13. De-Block CE, De-Leeuw IH & Van-Gaal LF. High prevalence of manifestations of gastric autoimmunity in parietal cell antibody-positive type 1 (insulin-dependent) diabetic patients. The Belgian Diabetes Registry. J Clin Endocrinol Metab 1999; 84: 40624067. 14. Ottesen M, Feldt-Rasmussen U, Andersen J et al. Thyroid function and autoimmunity in pernicious anemia before and during cyanocobalamin treatment. J Endocrinol Invest 1995; 18: 9197. *15. Betterle C, Dal-Pra C, Mantero F et al. Autoimmune adrenal insufciency and autoimmune polyendocrine syndromes: autoantibodies, autoantigens, and their applicability in diagnosis and disease prediction. Endocr Rev 2002; 23: 327364. 16. Betterle C, Volpato M, Pedini B et al. Adrenal-cortex autoantibodies and steroid-producing cells autoantibodies in patients with Addisons disease: comparison of immunouorescence and immunoprecipitation assays. J Clin Endocrinol Metab 1999; 84: 618622. 17. Boscaro M, Betterle C, Sonino N et al. Early adrenal hypofunction in patients with organ-specic autoantibodies and no clinical adrenal insufciency. J Clin Endocrinol Metab 1994; 79: 452455. 18. Laureti S, De-Bellis A, Muccitelli VI et al. Levels of adrenocortical autoantibodies correlate with the degree of adrenal dysfunction in subjects with preclinical Addisons disease. J Clin Endocrinol Metab 1998; 83: 35073511. 19. Bright GM, Blizzard RM, Kaiser DL et al. Organ-specic autoantibodies in children with common endocrine diseases. J Pediatr 1982; 100: 814. *20. Betterle C, Volpato M, Rees-Smith B et al. Adrenal cortex and steroid 21-hydroxylase autoantibodies in children with organ-specic autoimmune diseases: markers of high progression to clinical Addisons disease. J Clin Endocrinol Metab 1997; 82: 939942. 21. Kong MF & Jeffcoate W. Eighty-six cases of Addisons disease. Clin Endocrinol 1994; 41: 757761. 22. Kasperlik-Zaluska AA, Migdalska B, Czarnocka B et al. Association of Addisons disease with autoimmune disordersa long-term observation of 180 patients. Postgraduate Med J 1991; 67: 984987. 23. Myhre AG, Undlien DE, Lovas K et al. Autoimmune adrenocortical failure in Norway autoantibodies and human leukocyte antigen class II associations related to clinical features. J Clin Endocrinol Metab 2002; 87: 618623. 24. Bellastella A, Bizzarro A, Coronella C et al. Lymphocytic hypophysitis: a rare or underestimated disease? Eur J Endocrinol 2003; 149: 363376. 25. Sauter NP, Toni R, McLaughlin CD et al. Isolated adrenocorticotropin deciency associated with an autoantibody to a corticotroph antigen that is not adrenocorticotropin or other proopiomelanocortinderived peptides. J Clin Endocrinol Metab 1990; 70: 13911397. 26. Crock PA. Cytosolic autoantigens in lymphocytic hypophysitis. J Clin Endocrinol Metab 1998; 83: 609618. 27. ODwyer DT, Smith AI, Matthew ML et al. Identication of the 49-kDa autoantigen associated with lymphocytic hypophysitis as alpha-enolase. J Clin Endocrinol Metab 2002; 87: 752757. 28. Goswami R, Kochupillai N, Crock PA et al. Pituitary autoimmunity in patients with Sheehans syndrome. J Clin Endocrinol Metab 2002; 87: 41374141. 29. Kasperlik Za, Uska AA, Czarnocka B, Czech W et al. Secondary adrenal insufciency associated with autoimmune disorders: a report of twenty-ve cases. Clin Endocrinol 1998; 49: 779783. 30. De-Bellis A, Bizzarro A, Conte M et al. Antipituitary antibodies in adults with apparently idiopathic growth hormone deciency and in adults with autoimmune endocrine diseases. J Clin Endocrinol Metab 2003; 88: 650654. 31. Kavanaugh A, Tomar R, Reveille J et al. Guidelines for clinical use of the antinuclear antibody test and tests for specic autoantibodies to nuclear antigens. American College of Pathologists. Arch Pathol Lab Med 2000; 124: 7181.

Non-thyroid autoantibodies in autoimmune thyroid disease 31 *32. Tozzoli R, Bizzaro N, Tonutti E et al. Guidelines for the laboratory use of autoantibody tests in the diagnosis and monitoring of autoimmune rheumatic diseases. Am J Clin Pathol 2002; 117: 316324. 33. Kavanaugh AF. Evidence-based use of rheumatologic laboratory tests. Arch Intern Med 2004; 164: 109. 34. Weetman AP & Walport MJ. The association of autoimmune thyroiditis with systemic lupus erythematosus. Br J Rheumatol 1987; 26: 359361. 35. Vianna JL, Haga HJ, Asherson RA et al. A prospective evaluation of antithyroid antibody prevalence in 100 patients with systemic lupus erythematosus. J Rheumatol 1991; 18: 11931195. 36. Chan AT, Al-Saffar Z & Bucknall RC. Thyroid disease in systemic lupus erythematosus and rheumatoid arthritis. Rheumatology 2001; 40: 353354. *37. Pyne D & Isenberg DA. Autoimmune thyroid disease in systemic lupus erythematosus. Ann Rheum Dis 2002; 61: 7072. 38. Morita S, Arima T & Matsuda M. Prevalence of non-thyroid specic autoantibodies in autoimmune thyroid diseases. J Clin Endocrinol Metab 1995; 80: 12031206. 39. Petri M, Karlson EW, Cooper DS et al. Autoantibody tests in autoimmune thyroid disease: a case-control study. J Rheumatol 1991; 18: 15291531. 40. Loviselli A, Velluzzi F, Pala R et al. Circulating antibodies to DNA-related antigens in patients with autoimmune thyroid disorders. Autoimmunity 1992; 14: 3336. 41. Paggi A, Caccavo D, Ferri GM et al. Anti-cardiolipin antibodies in autoimmune thyroid diseases. Clin Endocrinol 1994; 40: 329333. 42. Zettinig G, Tanew A, Fischer G et al. Autoimmune diseases in vitiligo: do anti-nuclear antibodies decrease thyroid volume? Clin Exp Immunol 2003; 131: 347354. 43. Firestein GS. Evolving concepts of rheumatoid arthritis. Nature 2003; 423: 356361. 44. Nielen MM, van-Schaardenburg D, Reesink HW et al. Specic autoantibodies precede the symptoms of rheumatoid arthritis: a study of serial measurements in blood donors. Arthritis Rheum 2004; 50: 380386. 45. Silman AJ, Ollier WE & Bubel MA. Autoimmune thyroid disease and thyroid autoantibodies in rheumatoid arthritis patients and their families. Br J Rheumatol 1989; 28: 1821. 46. Andonopoulos AP, Siambi V, Makri M et al. Thyroid function and immune prole in rheumatoid arthritis. A controlled study. Clin Rheumatol 1996; 15: 599603. 47. Shiroky JB, Cohen M, Ballachey ML et al. Thyroid dysfunction in rheumatoid arthritis: a controlled prospective survey. Ann Rheum Dis 1993; 52: 454456. 48. Magnus JH, Birketvedt T & Haga HJ. A prospective evaluation of antithyroid antibody prevalence in 100 patients with rheumatoid arthritis. Scand J Rheumatol 1995; 24: 180182. 49. Innocencio RM, Romaldini JH & Ward LS. High prevalence of thyroid autoantibodies in systemic sclerosis and rheumatoid arthritis but not in the antiphospholipid syndrome. Clin Rheum 2003; 22: 494. 50. Vaidya B, Pearce SH, Charlton S et al. An association between the CTLA4 exon 1 polymorphism and early rheumatoid arthritis with autoimmune endocrinopathies. Rheumatology 2002; 41: 180183. 51. Ilias I, Mastorakos G, Mavrikakis M et al. Thyroid disease associated with rheumatoid arthritis is not adequately screened with a sensitive chemiluminescence thyrotrophin assay. Acta Medica Austriaca 1999; 26: 2628. 52. Palmer JM, Kirby JA & Jones DE. The immunology of primary biliary cirrhosis: the end of the beginning? Clin Exp Immunol 2002; 129: 191197. 53. Neuberger J. Primary biliary cirrhosis. Lancet 1997; 350: 875879. 54. Zuber MA & Recktenwald C. Clinical correlation of antimitochondrial antibodies. Eur J Med Res 2003; 8: 6170. 55. Muratori P, Muratori L, Gershwin ME et al. True antimitochondrial antibody-negative primary biliary cirrhosis, low sensitivity of the routine assays, or both? Clin Exp Immunol 2004; 135: 154158. 56. Elta GH, Seperskey RA, Goldberg MJ et al. Increased incidence of hypothyroidism in primary biliary cirrhosis. Dig Dis Sci 1983; 28: 971975. 57. Agarwal K, Jones DE, Daly AK et al. CTLA-4 gene polymorphism confers susceptibility to primary biliary cirrhosis. J Hepatol 2000; 32: 538541. 58. Vallas M, Monier JC, Tourniaire J et al. Prevalence of antimitochondrial antibodies in women with Hashimotos thyroiditis. Presse Med 1994; 23: 11171120. 59. Farrell RJ & Kelly CP. Celiac sprue. N Engl J Med 2002; 346: 180188.

32 A. P. Weetman 60. Johnston SD, McMillan SA, Collins JS et al. A comparison of antibodies to tissue transglutaminase with conventional serological tests in the diagnosis of coeliac disease. Eur J Gastroenterol Hepatol 2003; 15: 10011004. 61. Maki M, Mustalahti K, Kokkonen J et al. Prevalence of celiac disease among children in Finland. N Engl J Med 2003; 348: 25172524. 62. Sategna-Guidetti C, Bruno M, Mazza E et al. Autoimmune thyroid diseases and coeliac disease. Eur J Gastroenterol Hepatol 1998; 10: 927931. 63. Weetman AP, Burrin JM, Mackay D et al. The prevalence of thyroid autoantibodies in dermatitis herpetiformis. Br J Dermatol 1988; 118: 377383. *64. Berti I, Trevisiol C, Tommasini A et al. Usefulness of screening program for celiac disease in autoimmune thyroiditis. Dig Dis Sci 2000; 45: 403406. 65. Collin P, Salmi J, Hallstrom O et al. Autoimmune thyroid disorders and coeliac disease. Eur J Endocrinol 1994; 130: 137140. 66. Volta U, Ravaglia G, Granito A et al. Coeliac disease in patients with autoimmune thyroiditis. Digestion 2001; 64: 6165. 67. Jiskra J, Limanova Z, Vanickova Z et al. IgA and IgG antigliadin, IgA anti-tissue transglutaminase and antiendomysial antibodies in patients with autoimmune thyroid diseases and their relationship to thyroidal replacement therapy. Physiol Res 2003; 52: 7988. 68. Carroccio A, Custro N, Montalto G et al. Evidence of transient IgA anti-endomysial antibody positivity in a patient with Graves disease. Digestion 1999; 60: 8688. 69. Vincent A, Bowen J, Newsom-Davis J et al. Seronegative generalised myasthenia gravis: clinical features, antibodies, and their targets. Lancet 2003; 2: 99106. 70. Lavrni D, Jarebinski M, Rakocevi-Stojanovi V et al. Epidemiological and clinical characteristics of myasthenia gravis in Belgrade, Yugoslavia (19831992). Acta Neurol Scand 1999; 100: 168174. 71. Christensen PB, Jensen TS, Tsiropoulos I et al. Associated autoimmune diseases in myasthenia gravis. A population-based study. Acta Neurol Scand 1995; 91: 192195. 72. Marino M, Ricciardi R, Pinchera A et al. Mild clinical expression of myasthenia gravis associated with autoimmune thyroid diseases. J Clin Endocrinol Metab 1997; 82: 438443. 73. Weissel M. Mild clinical expression of myasthenia gravis associated with autoimmune thyroid disease. J Clin Endocrinol Metab 1997; 82: 39053906. 74. Potagas C, Dellatolas G, Tavernarakis A et al. Myasthenia gravis: changes observed in a 30-years retrospective clinical study of a hospital-based population. J Neurol 2004; 251: 116117. 75. Jacobson DM. Acetylcholine receptor antibodies in patients with Graves ophthalmopathy. J Neuroophthalmol 1995; 15: 166170. 76. Marino M, Barbesino G, Pinchera A et al. Increased frequency of euthyroid ophthalmopathy in patients with Graves disease associated with myasthenia gravis. Thyroid 2000; 10: 799802. *77. Scoeld RH. Autoantibodies as predictors of disease. Lancet 2004; 363: 15441546.