Tumour necrosis factor (TNF)-α is an important mediator of inflammation and has been reported to have potent disease-modifying effects in several chronic inflammatory diseases, including rheumatoid arthritis (RA) and spondyloarthropathies (SpA).1 2 In the last years TNF-α antagonists have been approved for the treatment of the above-mentioned chronic inflammatory diseases.3^–6 ‘ Despite a good overall efficacy profile and an acceptable toxicity profile they can induce a number of adverse effects, including autoimmune phenomena.7 Various autoimmune diseases have been reported after anti-TNF-α therapy, such as psoriasis,8 skin vasculitis,9 systemic lupus erythematosus10 and others.11

Because the use of TNF inhibitors is becoming more widespread, a greater understanding of possible side-effects is being recognised. When dealing with patients on these agents all physicians should be aware of possible adverse events and the potential development of such complications. Thus, we describe here nine patients who developed granuloma annulare (GA) skin lesions during the administration of anti-TNF-α therapy.

MATERIALS AND METHODS

In the last 7 years 199 patients with RA and 127 with SpA have been treated with infliximab, adalimumab and etanercept. We analysed all the patients who developed skin lesions suggestive of GA. Cutaneous diagnosis was confirmed by skin biopsy and histopathological examination. Medical diseases, concomitant immunosuppressive drugs, age of affected patients, type of reaction (pattern, localisation), progress of cutaneous lesions, and therapeutic options were reviewed.

RESULTS

From a cohort of 199 RA patients treated with anti-TNF agents (150 women, 49 men, mean age of 62.5 (SD 10.6) years and mean disease duration of 12.8 (5.7) years), nine cases with GA were identified. All patients responded well to anti-TNF agents. One hundred and twenty-three were treated with infliximab, 59 with adalimumab and 17 with etanercept. In addition, from a cohort of 127 patients with SpA (56 women, 71 men with a mean age of 45.7 (9.5) years, and mean disease duration of 10.6 (4.9) years) who were treated with infliximab (89 patients) and etanercept (38 patients), none developed GA. Adalimumab was not administered to SpA patients because there was no indication at that time for the treatment of SpA. None of these nine patients had clinical evidence of GA before or at the commencement of anti-TNF treatment. Moreover, none had disorders such as diabetes mellitus, tuberculosis, thyroid diseases, malignancies, or had or developed infections or other conditions such as trauma or receiving medications associated with GA.12 13 The main clinical characteristics of our patients are shown in table 1. There were eight women and one man, the GA was developed during the first year of treatment in seven patients and after the first year in two patients. All had a long-standing RA, were refractory to disease-modifying antirheumatic drugs and eight of nine were seropositive of IgM rheumatoid factor. The most common form of GA was the generalised type affecting the upper extremities. All cases were seen by a consultant dermatologist and were histopathologically confirmed. Owing to the extent of the skin lesions, the treatment was stopped in two patients, while all patients were treated with local corticosteroids with good results.

Representative cases

Patient 1

A 60-year-old woman with a long-standing seropositive RA refractory to leflunomide and methotrexate was treated with adalimumab 40 mg every 2 weeks subcutaneously. She responded well to adalimumab treatment, but after 3 months of treatment she developed erythematous skin eruptions covering her fingers, hands and forearms. The clinical diagnosis was consistent of generalised GA skin lesions (fig 1), which was confirmed with skin biopsy (fig 2). Owing to the extent and widespread of the skin lesions, adalimumab was stopped and the patient was treated with local corticosteroids with excellent results after 4 weeks of therapy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Generalised form of granuloma annulare. Widespread symmetric erythematous skin eruptions covering the fingers, hand and forearms bilaterally. Informed consent was received for the publication of this image.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Palisading granuloma: a central core of degenerate collagen surrounded by a radially arranged infiltrate of histiocytes and lymphocytes (haematoxylin and eosin×100) (A). Diffuse infiltrates of histocytes and lymphocytes in the dermis around of small foci degenerated collagen (haematoxylin and eosin×100) (B). Immunohistological picture showing the presence of CD3+ T lymphocytes (× 200) (C).

Patient 2

A 33-year-old woman with long-standing seropositive RA was treated with infliximab (3 mg/kg) in a loading dose of 0, 2 and 8 weeks and every 8 weeks thereafter. She was refractory to methotrexate and prednisone and responded well to infliximab therapy. Nine months after the initiation of treatment, she developed symmetric erythematous eruptions covering both fingers and hands. The clinical diagnosis was consistent with generalised type of GA, which was confirmed by skin biopsy. Owing to the extent of the eruptions, infliximab was discontinued. The patient was treated with local corticosteroids with the resolution of the eruptions after 3 weeks.

Patient 3

A 36-year-old woman with seropositive RA and disease duration of 4.7 years, refractory to methotrexate and prednisone, was treated with etanercept 25 mg twice per week subcutaneously. The patient responded well to etanercept treatment. Eight months later she developed erythematous skin eruptions affecting both fingers and hands. The clinical picture and the histological findings were compatible with GA. Local corticosteroids resulted in a good clinical response and etanercept was continued.

DISCUSSION

GA is a skin disorder that occurs in children and adults. Most patients are under 30 years of age, and women are affected twice as often as men. GA has been reported in association with a number of clinical conditions (eg, trauma, use of various medications), chronic diseases (eg, diabetes mellitus, tuberculosis, thyroid diseases, malignancy) and infectious diseases (eg, human immunodeficiency virus, Epstein–Barr virus, hepatitis C).12 13 Five clinical recognised variants of GA have been reported. There are the localised, generalised, subcutaneous, perforating and patch types. The most common is present on the extensor surfaces of the hand, arms, forearms, legs and feet.12 13

We report nine patients with GA during anti-TNF therapy in RA patients. In fact these patients were drawn from a cohort of 199 RA patients in whom 123 have been treated with infliximab, 59 with adalimumab and 17 with etanercept. From these nine patients who developed GA, two were treated with infliximab (1.6%), six were treated with adalimumab (10.2%) and one was treated with etanercept (5.9%). None of them had clinical evidence of GA before or at the time of initiation of anti-TNF therapy and none had or developed diseases or conditions associated with GA. All patients but one were seropositive for IgM rheumatoid factor. The occurrence of GA in patients with RA treated with adalimumab is higher compared with those treated with infliximab. Although this difference between the two monoclonal antibodies is difficult to explain a some simple explanation could be that despite the fact that both compounds have a similar mode of action, they differ in the way they are administered, and their pharmacokinetics and half-life. Of interest is the fact that from a cohort of 127 patients with SpA who were treated with infliximab and etanercept in our institution, none developed such skin reactions. The reason for this is still unknown. However, applying a χ² test between patients with RA and SpA under anti-TNF-α therapy, no differences were noted. This implies that GA development is a class-effect and not disease specific. Another possible explanation is the younger age and the male preponderance in patients with SpA.

Although the prevalence of GA may by high among patients with chronic diseases, its occurrence in 4.5% of our patients who received anti-TNF treatment for RA clearly exceeds the prevalence that might be expected by chance, especially within such a short period of observation. To our knowledge, only one case report of GA induced by anti-TNF agents has been reported so far.14 The exact cause of GA is unknown. Because of the histopathology and T cell subtypes associated with the lesion, a delayed type of hypersensitivity reaction is suspected, but the inciting antigen has not been recognised.13 The underlying pathophysiological mechanism responsible for the development of GA after anti-TNF therapy remain elusive. Several lines of evidence suggest that under certain conditions, anti-TNF treatment promotes the activation of autoreactive T cells leading to tissue damage via autoimmune mechanisms.2 This is exemplified by the common induction of antinuclear antibodies, anti-dsDNA, anticardiolipin antibodies in patients receiving anti-TNF therapy, and the occurrence of many autoimmune syndromes, such as psoriasiform eruptions, cutaneous and systemic lupus erythematosus, skin vasculitis, etc.7^–11 On the other hand, based on the critical role of TNF-α in granuloma formation some investigators successfully treated a patient with disseminated recalcitrant GA using infliximab.15

Our series of TNF-α antagonist-induced GA together with the previous case, strongly supports a link between TNF-α inhibition and GA development in some patients. Our observation may add to the discussion about new autoimmune phenomena induced by anti-TNF-α therapy. Further studies in collaboration with dermatologists are needed to identify risk factors for TNF antagonist-induced GA.