Granuloma annulare and necrobiosis lipoidica tissue reactions as a manifestation of systemic disease.

Granuloma annulare (GA) and necrobiosis lipoidica (NL) are generally considered to be idiopathic cutaneous palisading granulomatous dermatitides. There are sporadic reports of such lesions occurring in patients with coexistent systemic diseases other than diabetes mellitus. Having encountered 49 patients whose skin biopsies showed GA or NL lesions in the setting of extracutaneous disease, the authors set out to assess their clinical and histopathological findings to determine if any parameters were predictive of underlying systemic disease. Fifty-two skin biopsies from 49 patients having either GA or NL in whom there was a clinical history of an associated systemic disease were analyzed by light microscopy. The main systemic disease associations were rheumatologic, endocrine, hematologic, infectious, and inflammatory bowel diseases, ANCA positive vasculitic syndromes, and sarcoidosis. The clinical and histomorphological features were compared with those of a control group of patients whose skin biopsies showed GA or NL and in whom there was no history of extracutaneous disease. For the systemic disease group, patients were selected either retrospectively or prospectively from 160,000 cases accessioned in a 24-month period in the dermatopathology databases of Pathology Services, Inc (Cambridge, MA) and Central Medical Laboratories (Winnipeg, Canada). All systemic disease cases from the former service were analyzed blindly by the second author and from the latter service were analyzed blindly by the first author. Patients in the control group were obtained retrospectively from the Pathology Services Inc. database by the authors. The location of the lesions was atypical in 30 of 34 biopsies from systemic disease patients with a GA tissue reaction versus 10 of 22 biopsies of GA in the control group (P = .001). Six of 18 biopsies from patients with NL tissue reactions in the systemic disease group showed an atypical location, versus only 1 of 9 biopsies of NL from the control group (P = .19). The clinical diagnostic considerations were much broader in the systemic disease group versus the control group and included vasculitis, panniculitis, and connective tissue diseases including morphea in the former. In 22 of 34 GA biopsies and 16 of 18 NL biopsies from the systemic disease group, an active vasculopathy of leukocytoclastic, granulomatous, or thrombogenic subtypes was demonstrable. None of the GA or NL biopsies from the control group showed a similar active vasculopathy. An active vasculopathy was predictive of systemic disease in patients having either a GA-like or an NL-like tissue reaction (P < .001). Fifteen of 34 GA and 7 of 18 NL biopsies in the systemic diseases group showed extravascular neutrophilia in contrast to 3 of 22 GA (P = .02) biopsies and 2 of 9 NL (P = .33) biopsies in the control group. The finding of an active vasculopathy in a skin biopsy specimen showing a GA- or NL-like tissue reaction, particularly in the setting of an atypical clinical presentation both with respect to the location and appearance of lesions, should prompt consideration of an underlying systemic disease, as should extravascular neutrophilia in a skin biopsy showing a GA-like tissue reaction.