We appreciate the interest in our report shown by Dr Moncada and coworkers. We agree with their view that convincing data have been gathered in support of an abnormal T-cell response in toxic epidermal necrolysis (TEN), which correctly raises the issue of the place of immunosupressive therapy in the acute phase of this disease. This issue, a keystone in the controversy surrounding the treatment of TEN, is raised by the fear of infectious complications when corticosteroids are used and the reports of depressed lymphocyte numbers and function in the peripheral blood of these patients.1 Our observation of high lymphocyte counts in the blister fluid of patients with TEN suggests to us that peripheral blood immunodeficiency may be partially related to skin compartmentalization of immunocompetent lymphocytes and their final loss through extensive bullous effusions. The early administration and the short duration of high-dose systemic corticosteroid therapy, as used