A comparison of IgA positive and IgA negative dapsone responsive dermatoses.

Abstract

A study of thirty-three patients with a clinical diagnosis of dermatitis herpetiformis (DH) referred to our DH clinic over the las 11 years is reported. Twenty-six were referred by other consultant dermatologists. The diagnosis had been made by the clinical features and response of the rash to dapsone. Seventeen patients were found to have IgA in the uninvolved skin (IgA positive) and in sixteen no IgA was found (IgA negative). The duration of the rash prior to referral to the DH clinic was 3 months to 19 years (mean 5.0 years) for the IgA negative patients and 2 months to 22 years (mean 5.2 years) for the IgA positive group. The length of follow-up was 3 months to 11 years (mean 5.0 years) for the IgA negative, and 2--11 years (mean 5.6 years) for the IgA positive group. During follow-up the rash cleared completely and required no treatment in seven of the sixteen IgA negative patients. Thirteen of these sixteen patients no longer required dapsone, but six patients were receiving alternative treatment. In the three patients still taking dapsone IgA has not been found on subsequent biopsy. Of the seventeen IgA positive patients only three were able to stop dapsone during follow-up and in these three the IgA was still detected in the skin. Small intestinal mucosa was abnormal in eight of eleven IgA positive patients, but was normal in all thirteen IgA negative patients in whom jejunal biopsies were performed. An alternative diagnosis to DH has subsequently been made in thirteen of the sixteen IgA negative patients. Although the significance of IgA in the skin in DH is not known it appears to be part of the disease process. Patients who have a rash suggestive of DH and which is dapsone responsive, but in whom IgA is not found in the uninvolved skin, usually turn out to have a dermatosis other than dermatitis herpetiformis. Referral to a unit with expertise in immunofluorescence techniques of skin biopsies would appear to be helpful.