EPIDEMIOLOGIC APPROACHES TO THE STUDY OF TOXIC EPIDERMAL NECROLYSIS

The appropriate epidemiologic strategy for studying the etiology of toxic epidermal necrolysis is determined by the characteristics of the disease, particularly its rarity and the fact that it is caused by numerous drugs. Although information about drugs as risk factors can in principal be obtained from case reports and experimental studies, the former are subject to bias and the latter ace impractical because toxic epidermal necrolysis is so rare. Cohort studies are also impractical because of the rarity of the outcome. An automated database, even if based on a large population, can only yield valid results if it is used as the starting point for a case-control study that includes access to the subjects and to the medical records for information to confirm the diagnosis. A population-based case registry can provide a large enough and well-documented series of cases, but does not allow for the valid estimation of risks because it lacks a comparison series. This leaves a case-control study as the only strategy that is both practical and valid. An ongoing international case-control study of toxic epidermal necrolysis and Stevens-Johnson syndrome in relation to the use of drugs is described. Data collection has proceeded in France, Italy, Germany, and Portugal. The study in Germany is conducted within a population-based case registry, and the study in Portugal is also population based; this will allow for the estimation of absolute risks. Data on demographic factors and medical history, a detailed history of drug use in the month before hospital admission, and various other factors are collected by interview of the cases and hospital controls. Cases are confirmed in an independent review process in which the diagnoses, and classification a spectrum of Stevens-Johnson syndrome and toxic epidermal necrolysis, are determined without knowledge of drug use. As of June, 1993, 459 cases and 1299 controls have been enrolled. At the scheduled end of data collection in 1995, the projected totals are 691 cases and 1956 controls. These large numbers will allow for the detailed evaluation of even relatively uncommonly used drugs, for the evaluation of more commonly used drugs in relation to subtypes of toxic epidermal necrolysis/Stevens-Johnson syndrome, and for the comparison of results between countries.