Autoimmune diseases in vitiligo: do anti-nuclear antibodies decrease thyroid volume?

Abstract

An increased prevalence of autoimmune thyroiditis (AT) in vitiligo patients is well known. The aim of this study was firstly, to evaluate the clinical course of patients with both vitiligo and AT and secondly, to identify additional autoimmune disorders affecting the thyroid gland in a large cohort of vitiligo patients. We analysed a study group of 106 vitiligo patients and 38 controls. A detailed thyroid examination including sonography was performed in all study participants. In addition, the study participants were HLA typed and screened for various autoimmune disorders.

AT was significantly more frequent in vitiligo patients than in controls (21%versus 3%; P < 0·01). In 12 of the 22 patients with AT, vitiligo was the initial disease preceding AT by 4–35 years. In the other 10 patients with AT, both vitiligo and AT were diagnosed within one year. There were two individuals with diabetes mellitus type 1 and a single patient with Addison's disease. Anti-nuclear antibody (ANA), anti-smooth muscle cell antibody, and parietal cell antibody levels occurred with a similar frequency in patients and controls. In all vitiligo patients with both elevated ANA levels and AT (n = 6), the atrophic but not the goitrous variant was diagnosed. These vitiligo patients with both AT and elevated ANA levels had a significantly smaller thyroid volume compared to the vitiligo patients with AT whose ANA levels were normal (6·7 ± 4·5 ml versus 13·4 ± 9·1 ml, respectively; P < 0·05). The same was found in the entire study group: Thyroid volume of all vitiligo patients (with or without concomitant AT) was significantly smaller in the presence of ANA (6·9 ± 5·3 versus 10·5 ± 5·9 ml, espectively; P < 0·05). However, this phenomenon was not observed in the control group. There was a trend for a decreased frequency of HLA-DR3 (6·7%versus 23%) in our study group, but after correction for the number of comparisons, no HLA-allele was statistically significant associated neither with vitiligo nor with multiple autoimmune diseases in our patient sample. Our findings suggest that AT is the most frequent autoimmune disease associated with vitiligo. In our patients, AT presented simultaneously or after the onset of vitiligo but not before. Elevated ANA levels were associated with the atrophic variant of AT and may affect the volume of the thyroid gland, and there was no statistically significant association with the HLA system.

Introduction

Vitiligo is a disorder of pigmentation characterized by a selective destruction of melanocytes that affects approximately 0·5–2% of the population [1]. Today, an autoimmune aetiology is generally accepted [2] and possible genetic factors seem to play an important role in the pathogenesis of this disease. Vitiligo has been reported to be associated with a variety of autoimmune disorders such as adrenal insufficiency, pernicious anaemia, diabetes mellitus, and thyroid disease [3, 4, 5, 6, 7, 8]. All but one of the previous studies investigating thyroid disease in vitiligo patients diagnosed autoimmune thyroid disease only by the findings of abnormal thyroid hormones or elevated thyroid antibodies. Thyroid sonography was part of only one single former study [8]. To date, it is not known whether autoimmune thyroiditis (AT) in vitiligo patients is atrophic or goitrous and which of these two variants are associated with other autoimmune disorders. In addition, the chronological order of vitiligo and AT in patients with both diseases is unknown.

The aim of this study was to analyse a large sample of vitiligo patients to evaluate the clinical course of patients with both AT and vitiligo and to identify additional autoimmune disorders affecting the thyroid gland in these patients.

Patients And Methods

Study group

We investigated a total number of 106 patients (mean age 39 ± 18 years; range 6–80 years) who had been referred for treatment to a specialized unit at the Department of Dermatology. The diagnosis of vitiligo was made by an experienced dermatologist on the basis of the patients history and the typical clinical features consisting of discrete, well circumscribed, depigmented macules. The study group included 64 women (mean age 40 ± 18 years, range 6–80 years) and 42 men (mean age 38 ± 18 years, 9–76 years). 34 patients reported a positive family history. The disease was active within the last year in 88 patients and stable or slightly regressive in 18 patients. 93 patients had generalized vitiligo, 9 patients had the acrofacial type (macules on the hands, feet and face), and 4 patients the segmental type (macules involving a unilateral segment of the body) of vitiligo. All patients gave informed consent to participate in the study.

Control group

The control group consisted of 38 matched individuals (hospital employees performing their yearly medical check-up at our department due to sanitary regulation; the two eldest individuals were relatives of them). One male had AT, but the other controls had no history of autoimmune disorders. Because of ethical reasons, no individuals younger than 18 years were recruited, and for some older patients, the recruitment of a control person was not possible. Mean age of the control group was 40 ± 12 years (range 18–70 years). The control group included 21 women (mean age 38 ± 11 years, range 22–66 years) and 17 men (mean age 41 ± 13 years, range 18–70 years).

Examinations

All patients were investigated at the thyroid outpatient ward. They had a detailed history-taking of previous thyroid disease or other autoimmune disorders. Blood samples were taken, and a physical examination and thyroid sonography were performed. Immulite™ 2000 Third Generation TSH, Free T3, Free T4, Anti-TPO Ab, and Anti-TG-Ab immunometric assays (DPC Diagnostic Products Corporation, Los Angeles, CA) were used for the quantitative measurement of human thyrotropin (TSH), free circulating triiodthyronine (fT3), free circulating thyroxine (fT4), autoantibodies to anti-thyroid peroxidase (TPO-Ab) and autoantibodies to thyroglobulin (Tg-Ab) in serum. TSH receptor autoantibody (TRAB) levels were measured using the DYNOtest™ TRAK human assay (Brahms Diagnostica, Berlin, Germany).

Subclinical hypothyroidism was diagnosed on the basis of a raised TSH and a normal T3 and T4 value, the diagnosis of overt hypothyroidism required a low T3 and T4 value as well. The diagnosis of subclinical hyperthyroidism was based on a low TSH and a normal T3 and T4 value; overt hyperthyroidism was diagnosed by a low TSH and a raised T3 and T4 value. In our laboratory, the following normal ranges are established: TSH 0·1–4 µU/ml, fT4 0·6–1·8 ng/dl, fT3 1·8–4·2 pg/ml, TG-Ab <160 U/ml and TPO-Ab <70 U/ml. TRAB values> 2U/l were classified as positive. The interassay variation of all thyroid hormone assays was below 10%.

Thyroid sonography was performed in most of the patients with a Siemens Sonoline Prima diagnostic ultrasound system using a linear 7·5 MHz transducer by a single experienced physician (GZ) to avoid interobserver variance. In few cases, a GE Logiq 400 (GE Medical Systems, Milwaukee, WI, USA) was used. The sonographer was blinded to the serological findings prior to the measurement of thyroid volume. The gain settings of the ultrasound scanner were adjusted so that the lumina of the carotid artery and internal jugular vein were free of echoes, and the surrounding strap muscles would appear echo-weak in comparison to the echo-strong bright grey of normal thyroid tissue. Thyroid echogenicity was established by comparing the brightness of the echoes from the thyroid with those from the surrounding neck muscles and connective tissue [9]. Thyroid volume was calculated following the formula of Brunn et al. [10]. A thyroid volume>18 ml in females and>25 ml in males was classified as goitre. For children, Austrian normal values were used [11]. AT was classified as goitrous, when the thyroid volume was increased, and atrophic when the thyroid volume was normal or atrophic, respectively. All study participants with a history of thyroidectomy were excluded from thyroid volume analysis. For comparison of thyroid volume between study and control group, only data from vitiligo patients older than 18 years were used. Scintigraphy with Tc99m pertechnetate was performed only in patients with sonographic abnormalities.

Anti-nuclear antibodies (ANA) were determined by immunofluorescence staining on Hep-2 sections (Sanofi-Pasteur, Redmond, WA, USA), and antibodies to double stranded DNA (dsDNA-Ab) were analysed by a radioimmune assay (Ortho-Clinical Diagnostics, Amersham, UK). ANA detected after 1 : 80 dilution were classified as positive. Anti-smooth muscle antibodies (ASMA), anti-mitochondrial antibodies (AMA), and gastric parietal cell antibodies (PCA) were evaluated by immunofluorescenc staining on murine gastric/renal sections. ASMA, AMA, and PCA were classified as positive when detectable in 1 : 160, 1 : 40, and 1 : 80 dilution (or higher), respectively.

For serum cortisol measurement, blood samples were taken from all patients between 08.30 and 10.00h. and evaluated with a time-resolved fluoroimmunoassay (AutoDELFIA, PerkinElmer Life Sciences, Wallac Oy, Turku, Finland). The normal range in our laboratory is 8·9–26 µg/dl in the morning, interassay variation is <6%. Individual patients with multiple autoimmune diseases or serum cortisol levels <3 µg/dl were reexamined for adrenocorticotropic hormone (ACTH) level measurement. Blood samples were taken between 08.30 and 10.00h and stored immediatly in ice. The plasma was analysed with a chemoluminescence immunoassay (Nichols Advantage, Nichols Institute Diagnostics, San Juan Capistrano, CA, USA), the normal value for ACTH in our laboratory is 9–52 pg/ml, and interassay variation is <5%.

Haemoglobin A1c (HbA1c) was measured by high pressure liquid chromatography (Hi-Auto HA-8140, Arkray Inc., Kyoto, Japan). Inter-assay variation in our laboratory is <1·5% and the normal values are from 4·0 to 6·0%.

HLA typing

Serological HLA typing was performed according to the National Institute of Health microlymphocytotoxicity test [12].

Statistics

Results are given as mean ± standard deviation (range). Mann–Whitney-U-test was used for nonparametric comparison of the patient group (n = 106) with the control group (n = 38). χ² test was used to compare categorical variables and Spearman coeficcient of correlation was used when appropriate. A P-value <0·05 was considered significant.

The frequency of HLA specificities found in patients were compared in 2 × 2 contingency tables with values obtained from healthy Caucasoids [13]. For estimation of significance the χ² test or, where appropriate, Fisher's exact test were applied. The probability values obtained were multiplied with the number of comparisons (n = 67) in order to avoid type I errors.

Results

Thyroid disease

History of previous thyroid disease

Sixteen of the study patients (15%) had thyroid medication when they entered the study. Twelve individuals had levothyroxine for treatment of hypothyroidism, and four patients had levothyroxine treatment after thyroidectomy for nodular goitre. Three patients had a history of thyreostatic treatment, with remission a few years thereafter. In two further females, subclinical hypothyroidism was diagnosed and in one euthyroid female, elevated levels of thyroid antibodies were known. A family history of AT was noted in 14 of the study patients. Table 1 details the findings.

Table 1.

Thyroid state and history of additional diseases with known or suspected autoimmune aetiology: detailed information of all vitiligo patients with abnormal findings

Patient no.	Sex	Age (years)	Vitiligo (years)	Evidence of AT, additional autoimmune diseases, or family history of AT	Abnormal thyroid state	Elevated ab levels	Sonographic pattern	Positive autoantibodies	HLA typing
1	f	6	1	sclerodermia					n. d.
2	m	9	0·2					ANA	A1, 2; B44, 57; Cw2, 6; DR4, 13;
3	f	9	0·5	mother has AT		TG-Ab		ANA	A2, 11; B15, 61; Cw: n.d. DR15, 4;
4	m	10	2·75	mother and grandmother have AT				ANA	A1, 25; B57, 18; Cw6; DR15, 9;
5	f	13	0·3	type 1 diabetes for 6 years					A2, 3; B7, 37; Cw6, 7; DR9, 10;
6	f	14	0·75	AT (goitrous) for 2 months	euthyroid (T4 treatment)	TPO-Ab	hypo	PCA	A2; B27; Cw2, 4; DR2, 7;
7	f	14	12	mother has both vitiligo and AT					A3, 32; B5, 63; Cw1, 7; DR6;
8	m	16	1					ANA	A2, 24; B5, 35; Cw4; DR8, 9;
9	m	18	1			Tg-Ab		ANA	A3; B15, 41; Cw3; DR4, 7;
10	f	20	8	mother and grandmother have AT					A2, 29; B44, 57; Cw6;
11	f	22	6	AT (atrophic) diagnosed during study	subclin. hypo	TPO-Ab			A1, 28; B8, 53; Cw4, 7; DR6;
12	f	24	19					ANA	A2, 31; B44, 15; Cw4, 5; DR4, 7;
13	f	27	3	in vitro fertilization, no clear evidence of AT	euthyroid (T4 treatment)				A2, 66; B18, 41; Cw6; DR6;
14	m	29	9	Crohn's disease					A3, 29; B41, 55; Cw1; DR6;
15	f	29	21	mother and grandmother have AT					A1, 28; B35; Cw4; DR4, 11;
16	m	30	7	alopecia					A11, 24; B7, 55; Cw3, 7; DR15, 11;
17	f	31	23					PCA	A2; B15, 18; Cw4, 7; DR11, 12;
18	f	32	10	AT (atrophic) diagnosed during study		TPO-Ab	hypo		A2, 3; B44; Cw4, 5; DR4, 7;
19	f	32	13	mother has AT					A3, 24; B18, 35; Cw4, 7; DR1, 11;
20	m	32	1·5					ASMA	A2, 34; B7, 14; Cw7, 8; DR1, 15;
21	f	32	5	mother has vitiligo and AT				ANA, PCA	A3, 26; B5, 38; DR6, 12;
22	f	33	4	AT (atrophic) for 4 years	subclin. hypo (T4 treatment)	TPO-Ab	hypo	ANA	A26, 28; B38, 35; Cw: n.d. DR4, 15;
23	m	33	10					ANA	A2, 31; B7, 44; Cw5, 7; DR2;
24	m	34	9	antibody deficiancy syndrome for 1 years					A3; B7, 8; Cw7; DR15, 3;
25	m	35	4	diabetes mellitus (newly diagnosed)				ANA	A30; B44, 70; Cw7; DR7, 13;
26	f	36	2·5	AT (atrophic) for 3 years		TPO-Ab, Tg-Ab	hypo		A24, 31; B57, 60; Cw3, 6; DR4, 6;
27	f	36	12	subtotal thyroidectomy 9 years ago	euthyroid (T4 treatment)			ANA, PCA	HLA-A, B, C: n.d. DR1, 3;
28	f	39	25					ANA	A2, 29; B44, 50; Cw2, 7; DR2;
29	m	40	6	sister has AT					A28, 32; B44, 60; Cw3, 7;
30	f	41	12	AT (atrophic) for 8 years	euthyroid (T4 treatment)	TPO-Ab, Tg-Ab	hypo	PCA	A2; B7, 27; Cw2, 7; DR2, 4;
31	f	43	15	AT (atrophic) for 15 years, thyreostatic treatment with 27		TPO-Ab, Tg-Ab, TRAB	hypo	PCA	A2, 3; B7, 44; Cw5, 7; DR15, 13;
32	m	43	1	AT (atrophic) diagnosed during study	subclin. hypo	TPO-Ab, Tg-Ab	hypo	ASMA	A2, 3; B35, 37; Cw4, 6; DR4, 11;
33	m	44	39	AT (atrophic) for 13 years, thyreostatic trestment with 40		TPO-Ab	hypo		A2, 3; B38, 60; Cw3; DR6;
34	f	45	5	sister has AT					A2; B35, 60; Cw3; DR6;
35	f	46	27	AT (atrophic) for 2 years, Addison's disease for 25 years	euthyroid (T4 treatment)	TPO-Ab	hypo	ANA	A1, 3; B7, 8; Cw7; DR3, 13;
36	f	46	5	sister and daughter have AT		Tg-Ab			A24, 34; B18, 38; Cw7; DR12;
37	f	47	20					PCA	A2, 24; B5, 13; Cw6;
38	m	49	2	AT (atrophic) for 2 years	euthyroid (T4 treatment)	TPO-Ab			A2, 3; B7, 15; Cw7; DR1, 12;
39	f	51	37	AT, myxoedema 20 years ago; mother and sister have AT	subcl. hyper (T4 treatment)	TPO-Ab	hypo		A2; B7, 44; Cw7; DR16, 13;
40	f	51	36	AT (goitrous) for 2 years	euthyroid (T4 treatment)	TPO-Ab	hypo	PCA	A2, 3; B5, 35; Cw1, 4; DR1, 6;
41	f	52	18	AT (atrophic) for 18 years, thyreostatic treatment with 32		TPO-Ab, Tg-Ab		ANA, PCA	A1, 3; B56, 17; Cw1, 7; DR1, 8;
42	m	52	6	AT (atrophic) for 6 years		TPO-Ab, Tg-Ab	hypo	ANA	A2, 25; B44, 18; Cw5; DR4, 13;
43	f	54	10	AT (atrophic) for 1 years, brother has AT	euthyroid (T4 treatment)	TPO-Ab	hypo	ANA	A24, 32; B5, 27; Cw2; DR6;
44	f	54	41					PCA	A2; B35, 44; Cw4, 5; DR7, 13;
45	m	55	4	mother has AT					A2, 31; B7, 44; Cw5, 7; DR11, 6;
46	m	56	10	AT (atrophic) for 3 years		TPO-Ab, Tg-Ab, TRAB		PCA	A2; B18, 44; Cw5, 7; DR4, 8;
47	f	58	19	AT (atrophic) for 1 years, mother has AT	subclin.hypo	TPO-Ab	hypo		A3; B7, 27; Cw2, 7; DR15, 7;
48	f	58	20	AT (atrophic), diagnosed at study		TPO-Ab, Tg-Ab	hypo	ANA, PCA	A2, 11; B18, 38; Cw7; DR4, 11;
49	f	59	10					ANA, PCA	A11, 32; fB37, 61; C59w2, 6; D10R15, 7;
50	f	60	40		euthyroid (T4 treatment)			PCA	A2, 28; B14, 60; Cw3, 8; DR8;
51	f	63	4	AT (goitrous) for 4 years	euthyroid (T4 treatment)	TPO-Ab, Tg-Ab	hypo		A2; B7, 27; Cw2, 7; DR2, 7;
52	f	66	0·5	AT (atrophic) for 6 months		TPO-Ab	hypo		A3, 30; B44, 13; Cw6, 7; DR2, 7;
53	m	75	0·2	psoriasis					A3, 11; B15, 37; Cw3, 6; DR11, 13;
54	f	77	5	subtotal thyroidectomy 19 years ago	euthyroid (T4 treatment)			ANA	A2, 28; B44, 56; Cw1, 7; DR4, 11;
55	f	80	25	AT (goitrous) for 5 years	euthyroid (T4 treatment)	TPO-Ab, Tg-Ab			A2, 11; B7, 44; Cw7; DR15, 13;

AT, autoimmune thyroiditis; subclin hypo, subclinical hypothyroidism; subclin. hyper, subclinical hyperthyroidism; ab, antibodies; TPO-Ab, autoantibodies to anti-thyroid peroxidase; TG-Ab, autoantibodies to thyroglobulin; TRAB, TSH receptor autoantibody; hypo, hypoechogenic thyroid tissue; n.d., HLA typing not possible due to ethical reasons (patient 1) or ambiguous serological reactions.

Thyroid hormones

Subclinical hypothyroidism was diagnosed in four patients (TSH range 5·19–18·8 µU/ml), who all had elevated levels of thyroid antibodies. Two patients had subclinical hyperthyroidism: In a single patient, toxic nodular goitre was diagnosed whereas the other patient was taking too much thyroid hormone, overreplaced for hypothyroidism. All study patients had normal fT3 and fT4 levels. Thyroid hormones in the control group were normal.

Thyroid autoantibodies

Elevated levels of TPO-Ab were found in 22 study patients (21%; mean age 47 ± 15 years, range 14–80 years,16 females) and their values of TPO-Ab ranged between 79 and 1181 U/ml (mean 368 ± 269 µU/ml). Nineteen patients (18%) had elevated levels of Tg-Ab (mean age 46 ± 13 years, range 18–80 years,13 females) with a range between 177 and 3760 µU/ml (mean 809 ± 1079 µU/ml). Most of the patients with elevated levels of Tg-Ab had also elevated levels of TPO-Ab. Elevated TRAB values (3·1 and 4·1 U/l, respectively) were found in two patients, who had both elevated Tg-Ab and TPO-Ab levels.

In the controls, the male with known AT showed both elevated levels of TPO-Ab and Tg-Ab, and the female with multinodular goitre had also elevated levels of Tg-Ab. In all other controls Tg-Ab and TPO-Ab levels were normal (P < 0·01), and all controls had TRAB levels within the normal range.

Thyroid sonography

Hypoechogenic thyroid tissue was seen in 17 patients (16%) who all had elevated levels of thyroid autoantibodies. In the control group, only the individual with elevated levels of thyroid autoantibodies showed a hypoechogenicity (P < 0·01). The patients older than 18 years without history of thyroidectomy had a mean thyroid volume of 11 ± 6 ml (range 1–35 ml). Goitre was diagnosed in five cases (four patients had AT, and one a ‘cold’ thyroid nodule). Thyroid volume was not associated with length of levothyroxine therapy, and there was no association between thyroid atrophy and TRAB levels.

Mean thyroid volume of the controls was 12 ± 6 ml (range 4–32 ml); four controls had diffuse or nodular goitre with TPO-Ab levels within the normal range. Thyroid volume did not differ significantly between the study patients and the controls (P = 0·47),

Autoimmune thyroid disease

AT diagnosed on the basis of elevated levels of TPO-Ab accompanied by elevated levels of Tg-Ab, abnormal TSH, hypoechogenity, or a history of previous thyroid dysfunction was present in 22 patients (21%). AT was atrophic in 18 patients, whereas in 4 individuals, the hypertrophic variant of Hashimoto's autoimmune thyroiditis was diagnosed. The control group comprised a single patient with AT (P < 0·01). Active Graves’ disease was diagnosed in none of our patients.

Time of onset of thyroid disease and vitiligo

In our patient series, clinical presentation of thyroid disease did not precede the diagnosis of vitiligo: Both vitiligo and AT were diagnosed in 10/22 patients within one year, and the time since initial diagnosis ranged between one and 18 years before they entered the study. In 12/22 patients, vitiligo was the initial manifestation and AT was diagnosed after a median of 17 years (range 4–34 years). The time of onset of thyroid disease and vitiligo is listed in Table 1 in detail.

Other autoimmune disorders

Adrenal insuffiency

Mean serum morning cortisol level of the study group was 11 ± 5 µg/dl. Three individuals had decreased cortisol levels, but their ACTH value within the normal range, and all other ACTH measurements were also normal. A 46-year-old-female with generalized vitiligo since the age of 19 had a history of Addison's disease for 25 years, and two years ago, the atrophic variant of Hashimoto's thyroiditis was diagnosed. All controls had normal serum cortisol levels. Mean serum cortisol of the control group was slightly higher (13 ± 5 µg/dl), but the difference was not statistically significant.

Insulin dependent diabetes mellitus type I

The cohort included a 13-year-old boy with a history of insulin dependent diabetes mellitus (IDDM) since the age of seven years, and a 35-year-old man, in whom diabetes mellitus was newly diagnosed.

PCA levels

Elevated levels of PCA were found in 15 patients (13%). No study participant with increased levels of PCA had pernicious anaemia. Two controls (5%) had elevated PCA values. Although elevated PCA levels were found more frequently in the study group. the difference was not statistically significant (P = 0·15). There was no association between increased PCA levels and the presence of AT, and the presence of PCA did not affect thyroid volume.

ASMA and AMA levels

Two patients and one control had ASMA levels ≥1 : 160. None of the patients had detectable AMA values, which were therefore not measured in controls.

ANA levels

Eighteen patients (17%) had elevated ANA levels ≥1 : 80 of whom six had detectable dsDNA-Ab. The rate of ANA positive patients did not differ significantly from the controls in whom 4 individuals (10%) had ANA levels ≥1 : 80 including two with detectable dsDNA-Ab (P = 0·35). In all patients with both elevated ANA levels and AT (n = 6), the atrophic but not the goitrous variant was diagnosed. The vitiligo patients with AT and elevated ANA levels had a significantly smaller thyroid volume compared to the vitiligo patients with AT whose ANA levels were normal (6·7 ± 4·5 ml and 13·4 ± 9·1 ml, respectively; P < 0·05). Regarding the entire cohort of vitiligo patients (with or without concomitant AT), the same phenomenon was observed (Fig. 1): Thyroid volume was significantly smaller in the presence of ANA (6·9 ± 5·3 and 10·5 ± 5·9 ml, respectively; P < 0·05). In the control group, this phenomenon was not observed. Thyroid volume was higher in the four ANA positive individuals, but the difference was not statistically significant (P = 0·1). The presence of detectable dsDNA-Ab had no statistically significant effect on the thyroid volume.

Fig. 1.

The impact of ANA on the thyroid volume in all vitiligo patients.

HLA-typing

The distribution of the HLA types was compared between patients and a group of 100 healthy controls [13]. The most obvious finding was that the DR3 antigen was less frequent in the vitiligo patient group when compared to controls (6·7% versus 23%, χ² = 9·5, P < 0·0025). After correction for the number of comparisons (n = 67), however, we did not observe any statistically significant differences in the frequencies of HLA types in our patient sample.

Discussion

Several authors demonstrated an association between vitiligo and AT [3, 4, 5, 6, 7, 8,14]. It was already in 1941, when Robert suggested that vitiligo might be connected with an increased activity of the thyroid gland [15]: He noted a distinct rise of the basal metabolism in 10 out of 20 vitiligo patients tested. In addition, he presented a detailed overview of 28 case reports, case series and observational studies since 1887. All these papers reported vitiligo in patients with Graves’ disease or in patients with increased basal metabolism, patients with family histories of both Graves’ disease and vitiligo, or patients with both myxoedema and vitiligo [15]. In the age of ‘evidence based medicine’, these anecdotal observations have been confirmed. Several authors reported a significantly increased prevalence of autoimmune thyroid disease in vitiligo patients. Table 2 presents these studies in detail. In addition, there is also a study reporting a significantly increased prevalence of vitiligo in patients with autoimmune thyroid disease compared to patients with nonautoimmune thyroid disease [14].

Table 2.

Studies investigating the frequency of concomitant autoimmune disorders in vitiligo

Reference	n	Thyroid Ab (%)	PCA (%)	ANA (%)	Adrenal Ab (%)	ASMA (%)	AMA (%)
[3] Cunliffe et al.* (1968)	56	30* (P < 0.05)	5.7* (P < 0.004)	–	–	–	–
[4] Grimes et al. (1983)	70	11.4 (P < 0.05)	2.9 (n.s.)	2.9 (n.s.)	–	2.9 (n.s.)	–
[5] Betterle et al. (1985)	373	18.4 (P < 0.05)	10.4 (P < 0.001)	–	1.36 (P < 0.05)	–	–
[6] Hann et al. (1983)	226	7.1 (P < 0.01)	3.5 (n.s.)	12.4(P < 0.01)	–	25.7 (P < 0.01)	–
[7] Schallreuter et al. (1994)	205	20.0 (P < 0.05)	16.2	–	2.7	–	–
[8] Hegedüs et al. (1994)	35	26 (P < 0.05)	–	–	–	–	–
This study (2002)	106	21 (P < 0.05)	13 (n.s.)	17 (n.s.)	–	2 (n.s.)	0

^*Cunliff reported thyroid disease and pernicious anaemia. Thyroid Ab, thyroid autoantibodies; PCA, parietal cell antibodies; ANA, anti-nuclear antibodies; Adrenal Ab, adrenal autoantibodies; ASMA, anti-smooth muscle antibodies; AMA, anti-mitochondrial antibodies.

All but one of the previous studies diagnosed autoimmune thyroid disease only by findings of abnormal thyroid hormones or elevated thyroid antibodies (anti-thyroid microsomal antibdies or later TPO-Ab). Only Hegedüs et al. [8] included sonography and a detailed clinical examination in the analysis of his 35 vitiligo patients. He reported a markedly higher frequency (43% in the study group versus 20% in the controls) including also nonautoimmune forms of all sorts of thyroid disease. Regarding autoimmune thyroid disease, the authors reported similar results to ours: Nine of their 35 vitiligo patients (26%) had increased TPO-Ab levels (seeTable 1). In accordance to all previous studies, we also found a significantly increased prevalence of AT in vitiligo. Up to now it is not known, whether vitiligo or AT is the initial manifestation, and there is no way in a nonprospective study of detecting subclinical thyroid disease prior to the onset of vitiligo. In our series, clinical presentation of thyroid disease did not precede the diagnosis of vitiligo, and it is to note that most thyroid disease diagnosed these days is detected by testing patients with very vague symptoms or subclinical disease.

The cause of vitiligo has been remaining controversial over decades. Today, an autoimmune aetiology is generally accepted, which was primarily suggested by the frequent findings of concomitant other autoimmune diseases or antibodies in vitiligo patients. Several authors investigated the prevalence of various autoantibodies in vitiligo patients [1,3–3]. The diversity of their results is shown in Table 2.

In our patient series, no statistically significant association between vitiligo and Addison's disease or diabetes mellitus type 1 was noted. However, these diseases are much rarer than AT, and one case of Addisons's disease and two individuals with type 1 diabetes are far above the reported frequencies in healthy populations. We also identified additional patients with autoimmune disorders such as sclerodermia or alopecia (see Table 1), but due to the low frequencies, statistical significance could not be reached.

Our study group of 106 vitiligo patients as well as the subgroup of patients with both vitiligo and AT had a significantly reduced thyroid volume when increased ANA levels were present. An association between ANA and the atrophic variant of AT has not been previously reported. Several investigators reported a cytotoxic effect of anti-DNA antibodies in vitro[16,17]. ANA are directed against various components of the nucleus, they have the abibility to penetrate into the living cell [18], and there are indications that the penetration of autoantibodies occurs also in vivo and may lead to apoptotic cell death [19,20]. Alarcon-Segovia et al., who had reported intracellular penetration of anti-ribonucleoprotein antibodies for the first time [21], reviewed various papers reporting the penetration of autoantibodies in vivo in detail [18]. We are not aware of any previous reports dealing with the impact of ANA on the thyroid gland, but our findings suggest a possible independent role of ANA in inducing thyroid cell destruction in vivo. However, increased levels of ANA could also be the consequence of thyroid damage, and there is also the possibility that ANA may simply be a marker for more active disease and do not necessarily play a role in pathogenesis.

Several authors believe that atrophic and goitrous variant are a single entity with goitrous presentation initially and atrophic changes later on. Others consider atrophic and goitrous thyroiditis as two entirely separate entities with different pathophysiology [22]. Several reports support this concept, which is further strengthened by the association between ANA and the atrophic variant of AT. Genetic differences exist between both variants of AT [23,24] and we have previously demonstrated an association between dermatitis herpetiformis with the atrophic, but not with the hypertrophic variant of AT [25]. Cytotoxic thyroid antibodies are associated with the atrophic variant of AT and seem to favour the development of the atrophic rather than the hypertrophic variant of the disease [26].

Our cohort included no patients with active Graves’ disease. Two individuals had slightly elevated levels of TRAB but were classified as TRAB positive AT [27]. Noteworthy, this finding is in contrast to previous reports which identified Graves’ disease in up to 35%[8].

Several endocrine autoimmune components of polyglandular autoimmune syndrome type 2 share a common genetic background and could be linked to the HLA system [28,29]. Polyglandular autoimmune syndrome type 1, however, does not show association to specific HLA haplotypes. This disease has been assigned to a locus on chromosome 21q22.3 [30] and till today, a total of 16 different mutations have been observed in this locus [31]. HLA loci have not shown any disease association in this genome wide linkage study [30]. The association of autoimmune thyroid disease with HLA genes has received much attention and in several populations of different ethnic background, an association between the HLA system and Graves’ disease was reported [32–34]. In AT, the association with the HLA system is weaker and studies with those patients indicate that the HLA genes make only a small contribution to the overall genetic susceptibility to autoimmune thyroid disease [23,35]. Several genes outside the HLA system have been found to be associated with AT [35–37]. In our patient sample, there was a trend towards a decrease of the DR3 phenotype frequency which is in accordance with a previous report [38]. However, no HLA-allele was statistically significantly associated neither with vitiligo nor with multiple autoimmune diseases in our patient sample.

Conclusions

Our findings suggest that AT is the most frequent autoimmune disease associated with vitiligo. In our patients, AT occurred more frequently than in controls, and presented simultaneously or after the onset of vitiligo, but not before. Elevated ANA levels were associated with the atrophic variant of AT and may affect the volume of the thyroid gland, and there was no statistically significant association to the HLA system.