Deliberate self‐poisoning with isotretinoin

I cm and contained clear fluid. There were no mucosal lesions and the child was otherwise healthy. New lesions developed in groups on erythematous skin and in crops over succeeding days. No systemic treatment was given and the condition resolved after 3 months. There has been no recurrence to date. One month before admission he had been unwell with headache, nausea and abdominal pain but no bowel upset. He had received at that time a io-day course of co-trimoxazole. His full blood count and the serum immunoglobulins were normal. The blister fluid was examined by electron microscopy for viral particles with a negative result and there was no evidence of bacteria in the blister fluid either by Gram's stain or on culture. The histopathology was of a subepidermal bulla containing fibrin and with numerous neutrophils and a few eosinophils. Direct immunofluorescence of the biopsy was negative for immunoglobulins but there was a patchy distribution of complement and fibrin at the basement membrane. The complement fixation titre for adenovirus on i April 1976 was i: 16 rising to i :256 by 28 April 1976. It has been recognized that patients with buUous disorders, including that of chronic buUous disease of childhood, may have a prodromal illness' and an infective factor has been suggested.^ Chronic bullous dermatosis of childhood is characterized by the presence of linear IgA deposits at the dermo-epidermal junction-* and there are a number of other distinctive bullous disorders occurring in children.* Adenovirus infections are common and self limiting.^ In recent years adenovirus has been implicated in the pathogenesis of coeliac disease* and a sequence homology between A-gliadin and a structural protein of adenovirus has been reported.' It would be of interest to know if other centres have found an association between the occurrence of adenovirus infection and the development of chronic blistering dermatoses in children.