Erythema multiforme (EM) represents an uncommon, acute inflammatory, reactive mucocutaneous disorder [1-3]. Etiologic agents of EM include infections (particularly herpes simplex virus, mycoplasma pneumonia, Epstein-Barr virus and histoplasmosis) and medications [1-3]. EM direct immunofluorescent (DIF) has been described as nonspecific, but may show deposition of Complement/C3 and fibrin along the dermal/epidermal junction, as well as IgM, Complement/C3 and fibrin around upper dermal blood vessels [1-3]. Here we report an EM case that seemed to be triggered by multiple medications; the histologic and immunologic studies demonstrated a complex immune response, including anti-neutrophil cytoplasmic antibodies (c-ANCAs). We also noted cellular debris within the blisters, including fragments of melanocytes and keratinocytes.

A 68 year old African American female presented with recurrent, flaccid bullae; erosions, crusts, scales, hyperpigmented areas and erythematous circular lesions on her face, ears, neck, eyelids, genital areas, and hands and legs. The patient was taking multiple medications simultaneously, including Lasix^®, Omeprazol^®, Omega 3, spironolactone, Norvasc^®, gentamicin, hydroxyzine, atenolol, Lisinopril^®, Zoloft^®, eye drops, clonidine, Tylenol^®, Bumex, Procrit^®, levothyroxine, and minoxidil. Biopsies for hematoxylin and eosin (H&E) examination, DIF and immunohistochemistry (IHC) analysis were performed. A diagnosis of EM was established. The physician then began slowly removing each medication, trying to determine which ones could be triggering the disease; however, she was unable to determine which one(s) represented the exact triggering factor(s). The patient was prescribed Lidex^® cream at 0.5%; Tacrolimus^® at 0.1%; prednisone, 50 milligrams a day for two weeks, with a slow weekly decrease every week; and Allegra^® at 180 milligrams daily. After three weeks of treatment, the rash cleared.

Our H&E tissue processing and staining were performed as previously described [4,5]. Our case report was exempted from Institutional Review Board review and research patient consent.

The immune response in situ in EM has been studied, but remains incompletely characterized [1-3]. One group of authors studied several cases of EM, using H&E and DIF and reported positivity in 88% of forty-one skin biopsy specimens from thirty-four patients, for the presence of cytoid bodies, or fluorescent keratinocytes in the epidermis [6]. These authors noted that the cytoid bodies stained most often with IgM (homogeneously) or with Complement/C3 (speckled) [6]. Our findings are in agreement with them, as well as with other authors who tested colloid bodies: IgG, IgA, IgM, C3 and fibrinogen could be demonstrated by DIF [7]. We noted both colloid bodies and cytoid bodies using H&E (both represent eosinophilic hyaline ovoid bodies which are often found in the subepidermal papillary regions, or sometimes in the epidermis); and by DIF, primarily in the epidermis.

In addition, we observed some “large round bodies” exclusively inside the blisters that were much larger than cytoid bodies and/or colloid bodies. To our knowledge these “large round bodies” have never been previously described, and we speculate that these may be indicative of the fact that our patient was taking so many medications, possibly causing more significant tissue damage. Using double color IHC staining and multiple fluorochome DIF techniques, [8,9] we determined that the “large round bodies” as well as the cytoid and colloid bodies contained amalgamated fragments of neutrophils, keratinocytes and melanocytes, plus positive staining for CD4, CD68, fibrinogen, Complement/C3, C4, c-ANCAs and IgA. In addition, we observed a mixed immune response including T cells, antibodies, complement in the upper dermal inflammatory infiltrate. The immune response was seen against the BMZ, dermal vessels and eccrine glands.

Several authors have previously reported the presence of a specific T cell immune response in EM; that cell-mediated immunity appears to be responsible for the destruction of epithelial cells via CD8 and macrophages, and that the dermis displays an influx of CD4 positive lymphocytes in response to herpes virus infections [10]. No viral immune response was seen at least in situ in this case. Contrary to what has been described in many reports in the literature, HLA-DP, DQ, DR antigen stained weakly positive; our staining was primarily observed on inflammatory cells around the upper dermal vessels, and was negative on the vessels themselves.

Multiple medications may also trigger vasculitis without clinical EM, and ANCAs may present in other conditions [5-9]. ANCAs may be found in granulomatosis with polyangiitis, microscopic polyangiitis, Churg-Strauss syndrome and levamisole-adulterated cocaine intoxication. C-ANCAs have been described primarily to be of IgG isotype, directed against antigens in the cytoplasms of neutrophils. However, in previous reports we have described c-ANCAs presenting via other immunoglobulin isotypes [9]. In EM presenting with c-ANCAs and vasculitis, is also important to evaluate the patient for renal, pulmonary, joint and muscular function.

We found evidence of an amalgamation of multiple cells and their debris in cytoid bodies, colloid bodies and “large round bodies”. These findings may represent the results of defragmentation and amalgamation of multiple cells including those involved in the immune response to EM; in addition, the findings suggest a brisk immune response. The immune response was directed inside the blisters, on epidermal cell junctions, and against dermal vessels and eccrine glands. Overall, our results may be indicative of a more complex immune response than previously documented in EM, involving cytotoxic T cells, antigen presenting cells, neutrophils and c-ANCAs.