Antineutrophil cytoplasmic antibody (ANCA)-associated autoimmune diseases induced by antithyroid drugs: comparison with idiopathic ANCA vasculitides

Antineutrophil cytoplasmic antibodies (ANCA) specific for proteinase 3 (PR3) and myeloperoxidase (MPO) are associated with necrotizing vasculitides, especially Wegener's granulomatosis (WG), microscopic polyangiitis (MPA) and idiopathic crescentic glomerulonephritis [1]. Pathogenesis of ANCA-associated idiopathic systemic vasculitides (ISV) is not well understood, but it has been shown that ANCA-activated neutrophils contribute to oxidative and proteolytic damage of blood vessels [2]. Cytoplasmic PR3-ANCA has high specificity (99%) for the newly diagnosed WG [3]. Perinuclear MPO-ANCA is a good serological marker for MPA, but it can also be found in patients with systemic lupus erythematosus (SLE), rheumatoid arthritis, drug-induced vasculitides (DIV), etc [4].

ANCA-associated ISV are rare and their annual incidence is approximately 9.5 per million (in Germany) [3]. Although WG and MPA belong to the ISV group, they can be triggered by some chemicals, viral and bacterial infections and certain drugs, among which antithyroid drugs (ATDs) are very common [5]. Propylthiouracil (PTU) and methimazole (MM) may induce ANCA-positive vasculitides [6].

The clinical and serological profiles of idiopathic and drug-induced autoimmune diseases (DIDs) can be very similar. Contrary to idiopathic vasculitides, DIDs have a milder course and often do not necessitate cytotoxic drug therapy [5]. Pathogenesis and clinical/serological characteristics of ANCA-associated diseases triggered by ATD have not been sufficiently investigated. In a retrospective study, we compared data from idiopathic and ATD-induced ANCA-positive patients.

In the ATD-treated group, only 1/16 patients developed ARF (SNGN with crescents) with severe pulmonary-renal syndrome; 2/16 patients developed hematuria and non-nephrotic range proteinuria. Only one female patient treated with MM had nodular lung infiltration pathohistologically verified as granulomatous vasculitis.

In contrast to patients with ISV, ATD-associated ANCA-positive patients had skin manifestations more frequently, especially urticaria-like vasculitis and urticaria (Table 2). Patients with urticaria-like vasculitis and purpura histologically had leucocytoclastic vasculitis. A pregnant patient with poorly controlled GD had extensive cutaneous necroses and hemorrhagic bullae, while a patient with pluriglandular autoimmune syndrome had lower leg ulcerations. Digital gangrenes and mouth ulcerations were noted only in PR3-ANCA-positive ISV patients.

None of the ATD-treated patients manifested central nervous system effects, while 1/16 patients had predominantly peripheral sensory polyneuropathy that regressed upon drug withdrawal.

In the ISV group, at the beginning of the disease, there was no correlation between cANCA or pANCA titers and BVAS (p > 0.05). Also, in drug-induced MPA patients (Table 5), there was no correlation between pANCA titers and BVAS (p > 0.05).

All patients with ISV were treated with immunosuppressive therapy. In ATD-treated patients, only 6/16 patients were treated with immunosuppressive therapy (Table 5). Although clinical remission was achieved in all 16 ATD-treated patients after 6 months, ANCA titers slowly declined and only finally disappeared after 2 years. Also, clinical remission preceded the disappearance of organ-nonspecific antibodies. Anti-DNA, AHA and cryoglobulins disappeared in 6 months, while aCL and anti-β2 GP I concentrations declined, but persisted for up to 1.5 years. Clinical remission preceded the disappearance of ANCA in 6/16 patients who were treated with immunosuppressive drugs.

In patients with ISV, induction of clinical remission correlates with significant ANCA titers reduction (Table 5), and this was observed in 46/56 individuals who were not resistant to standard immunosuppressive therapy. The standard regimen could not induce clinical remission and ANCA titer reduction in 10/56 patients, necessitating introduction of additional immunosuppressive therapy including plasmapheresis in three and intravenous immunoglobulins (2 g/kg) in seven patients, respectively. No patient in the ATD-treated group had to be treated with additional immunosuppressive therapy.

In the ISV group, chronic renal failure was statistically more frequent (p < 0.01) than in ATD-treated ANCA-positive patients (Table 5). There were no differences in terminal renal failure (p > 0.05) between the groups. In 2/16 ATD-treated patients, the drugs (MM 20 mg/day and PTU 250 mg/day, respectively) were not discontinued in time, and they developed terminal renal failure and chronic renal failure, respectively. Both patients had ATD-induced MPA during a relapse of GD. ATD-associated LLD is characterized by a milder course and better prognosis than ATD-associated MPA (Table 5).

In the ISV group, lethal outcome was more frequent (p = 0.059) than in the DID group. In the ISV group, 6/56 patients died during the first episode, while out of the remaining 50, 47 had at least one relapse. In the ATD-treated group there were no disease relapses.

All the patients with ATD-induced disease are now in remission without any therapy, whilst the 37/43 patients with ISV need to receive maintenance therapy (p < 0.001).

After the ATD withdrawal, hyperthyroid patients were treated with radioactive iodine or subtotal thyroidectomy. In a pregnant woman with GD, potassium perchlorate was prescribed at the lowest dose, keeping the patient mildly hyperthyroid. A healthy child was born normally 3 months later.

Standard clinical and serologic criteria for discrimination of ANCA-associated ISV from DID are inadequate at the initial presentation. The histopathology can be inconclusive, since both entities are characterized by leucocytoclasia and fibrinoid necrosis of the blood vessels [14]. ANCA-positive DID can be 'pauci-immune' but also with strong vascular immune deposits [1].

In this 11-year retrospective study, we compared clinical and serological data from idiopathic and ATD-induced ANCA-positive patients. A higher number of younger females in the ATD-treated group was found (Table 1), because hyperthyroidism is predominantly a disease of this sex and age group [9]. On the other hand, WG and MPA are predominantly diagnosed in the older population [1]. ANCA-positive ATD-induced diseases occur more frequently in GD than in HT [6], which we also found (Table 1).

All ATDs can induce asymptomatic production of ANCA, but PTU most often induces ANCA-associated MPA or LLD [15]. In our study, apart from long-term ATD use [16], the repeated use of ATDs was also a risk factor for developing the disease. As PTU is more commonly associated with ANCA-positive DID, MM may be the preferred medication if hyperthyroidism is combined with other autoimmune diseases [17]. At the molecular level, PTU and MM are based on thionamide, and the substitution of one with the other should be undertaken with caution. On the other hand, patients should be promptly brought to euthyroid state. Since MM is contraindicated in pregnancy, PTU-induced disease in pregnancy demands cautious use of iodides [18]. This enabled a successful delivery in one of our patients.

The analysis of the clinical parameters (Table 2) showed that the first symptoms (arthralgias and myalgias) of ANCA-positive ISV and ATD-treated patients were similar. Shorter duration of the symptoms before diagnosis (Table 1) in ATD-treated patients can be explained by a higher percentage of early skin manifestations. Although arthralgias, myalgias and weight loss can be seen in hyperthyroidism, temperature peaks over 38.5°C, and renal and pulmonal involvement are unusual [9]. Sera et al. [19] followed 21 MPO-ANCA-positive patients treated with PTU: 12 had no symptoms but nine patients complained of myalgia, arthralgia or common cold-like symptoms after the appearance of MPO-ANCA.

ANCA-associated pulmonary-renal syndrome is the most severe presentation of necrotizing vasculitis. In ATD-treated patients, renal lesions may be mild but also fulminant, including development of ARF [20]. One MM-treated patient developed drug-induced WG, which is a very rare form of DID [21]. After discontinuing ATD, our patient was treated orally with cyclophosphamide and prednisone for 3 months, while Pillinger et al. induced remission of WG only by discontinuing PTU [21].

Cutaneous changes in ANCA-positive ATD-treated patients can vary enormously from maculopapular exanthema to pyoderma gangrenosum, cutaneous ulcerations and vesiculo-bullous lesions typical for bullous SLE [22, 23]. Of our patients, 7/16 had urticaria and urticaria-like vasculitis while two had extensive cutaneous necrosis (Table 2). Urticaria-like vasculitis is frequently seen in patients with hypo-complementemic vasculitis and SLE. Gangrene and cutaneous necroses are poor signs of PR3-ANCA-positive ISV [24]. However, cutaneous necroses in our two ATD-treated ANCA-positive patients were not associated with a worse prognosis.

An analysis of the serologic profile of the ATD-treated patient group showed simultaneus presence of various autoantibodies (Table 3). In patients with GD and HT, even before starting ATD, different ANA were found [25], as well as different organ-specific autoantibodies.

On the other hand, most patients with GD and HT who were not treated with ATD do not have ANCA [19]. In the PTU-treated group, 25% of the patients had MPO-ANCA, while in the MM group only 3.4% had MPO-ANCA [15]. These findings suggest that the altered immune environment associated with GD and HT is not sufficient to develop ANCA, but treatment with ATD induces ANCA production.

The titer of pANCA was much higher in the ATD-treated group than in the ISV group, while concentrations of MPO-ANCA did not differ (Table 3). We suppose that patients taking ATD have ANCA specific to elastase and lactoferrin, which was described in PTU-induced vasculitis and LLD [26]. PTU is also associated with simultaneous production of MPO-ANCA and PR3-ANCA [26], which however, we did not see in any of our patients. It was shown that WG patients during ATD therapy changed PR3-ANCA to MPO-ANCA with flares of vasculitis [27].

The pathogenic role of ANCA remains controversial in ISV and DID. Because not all ANCA-positive patients develop the disease, some authors are considering, along with ATD, an infection as an additional trigger in inducing autoimmune diseases [20]. In our series of ATD-associated ANCA-positive patients, there were no clinical or laboratory parameters for infection. However, an unnoticed mild viral infection might induce priming (TNF-alfa) of neutrophils and translocation of PR3 and MPO to the cell surface [2, 28].

Some authors are considering the possibility of cross-over epitopes between thyroid peroxidase and MPO [29]. We showed that only some MPO-ANCA-positive patients have antimicrosomal antibodies (Table 3). Besides 44% homology in amino acid sequences, there is no cross-reactivity between thyroid peroxidase and MPO [30].

On the other hand, it was shown that PTU accumulates in neutrophils, binds to MPO and induces production of cytotoxic products [31, 32]. It has been suggested that some drugs could induce neutrophil apoptosis [33]. Moreover, neutrophil apoptosis, in the absence of priming, is associated with translocation of ANCA antigens to the cell surface [34]. Simultaneous presence of various autoantibodies (anti-DNA, AHA, anti-SSA, aCL and anti-β2 GP I) suggest that apoptotic blebs of neutrophils could be a source of immunogens in our patients with ATD-induced LLD [35, 36].

ATD-induced ANCA-positive diseases are divided in two groups (Table 4). The first group (4/16) consisted of PR3- or MPO-ANCA-positive drug-induced WG or MPA which resemble ISV. The second group (12/16) consisted of MPO-ANCA-positive patients with LLD. We speculate that different clearance of apoptotic materials by macrophages might determine type of autoimmune response [37]. PTU-induced LLD was described in a patient with HLA DR9 haplotype whose twin sister had SLE [38].

Our ATD-associated LLD patients, despite high concentration of MPO-ANCA, had mild clinical courses. IgM anti-MPO antibody-secreting lymphocytes are present in the peripheral repertoire of lupus-prone mice but rarely differentiate into IgG-producing cells [39]. We can suppose that ATDs induce production of the potentially pathogenic MPO-ANCA IgG₄ subclass [40]. MPO has the characteristics of a SLE-related antigen [39]. MPO-ANCA in SLE is most often found in the context of secondary APS [41], for which there were no criteria in our patients. Anti-β2 GP I antibodies are typical markers for APS [13] and have not been reported in ATD-induced ANCA-positive diseases. These data may have an impact on the approach to the patients since thrombotic events can be expected in case of an infection or a surgical procedure.

Simultaneous presence of cryoglobulins, aCL and anti-β2 GP I antibodies [42] in our two MPO-ANCA-positive patients with LLD synergistically contributed to development of cutaneous necrosis. However, we can assume that in patients with ISV, absolute or relative deficiency of α1 AT (Table 3), which is a specific PR3 inhibitor, could play a role in necrosis. The clinical picture of ISV depends on alotypic polymorphism of the Fcγ receptors, the isotype of PR3-ANCA or MPO-ANCA, and avidity as well as functionality of α1 AT [2]. Bearing in mind all these factors, it is clear that the ANCA titer in our patient with ISV did not correlate with BVAS (p > 0.05). Higher CR-P values in patients with ISV (Table 3) point to more intensive necrosis and inflammation than in indviduals with ATD-induced diseases.

Generally, ISV patients (Table 5) have a more severe course, demand more intensive immunosuppressive therapy, and have frequent relapses which we have not observed in ANCA-positive ATD-treated patients.

Therefore, patients with GD and HT should be carefully followed and monitored, not only for their thyroid status but also for the serious complications of ATD therapy. Patients with positive ANCA without evidence of vasculitis should also be carefully followed. The onset of urticaria-like vasculitis, erythrocyturia or pulmonary complaints demand immediate discontinuation of ATD.

Our results show that ATDs are frequent exogenous trigger factors of ANCA-associated autoimmune diseases. In comparison with drug-induced disease, ISV patients had a severe course with more frequent renal, pulmonal and neurologic manifestations. In contrast to patients with ISV, ATD-induced ANCA-positive patients had skin manifestations more frequently, especially urticaria-like vasculitis. ATD can induce two clinically and serologically different types of diseases. Simultaneous presence of various autoantibodies characterizes ATD-induced MPO-ANCA-positive LLD with good prognosis. On the other hand, ATD-induced MPO or PR3-ANCA-positive life-threatening MPA or WG resemble ISV and were not associated with organ-nonspecific autoantibodies. Different serological profiles could help in the differential diagnosis and adequate therapeutic approach to ANCA-positive ATD-treated patients with symptoms of systemic disease.