To the Editor.—
In the January 1990 issue of the Archives, two articles were of great interest to us, both concerning the incidence of toxic epidermal necrolysis.1,2 The study by Roujeau and colleagues was a nationwide surveillance in France covering the period 1981 through 1985. The incidence was estimated from two different sets of data (direct inquiry to dermatologists and mortality data) as 1.2 and 1.3 cases per million per year, respectively. The other study utilized the data system at Group Health Cooperative of Puget Sound, Seattle, Wash, covering about 260 000 people, and considered all hospitilization for toxic epidermal necrolysis, erythema multiforme, and Stevens-Johnson syndrome in that area from 1972 through 1984. The incidence of toxic epidermal necrolysis was estimated at about 0.5 per million per year.In January 1989, we started a case-control surveillance of Stevens-Johnson syndrome and toxic epidermal necrolysis covering 33 dermatological services (hospital and