A population-based study of Stevens-Johnson syndrome. Incidence and antecedent drug exposures.

To determine the incidence of Stevens-Johnson syndrome, a descriptive epidemiology study was performed using computerized Medicaid billing data from 1980 to 1984 from the states of Michigan, Minnesota, and Florida. The ratio of persons hospitalized with a discharge diagnosis of erythema multiforme (ICD-9-CM code 695.1) to persons with any claim for medical service was first used as an estimate of the incidence rate of the disease. Then, since the ICD-9-CM code for erythema multiforme includes other illnesses in addition to Stevens-Johnson syndrome and because these illnesses are frequently misdiagnosed, the information provided by a review of medical records for a subset of cases of erythema multiforme was used to determine the proportion of patients with true Stevens-Johnson syndrome. The incidence rates of Stevens-Johnson syndrome were 7.1 (6.1 to 8.2), 2.6 (1.6 to 4.0), and 6.8 (4.3 to 10.3) per million per year in each state, respectively. Penicillins, especially aminopenicillins, were frequently used in the 19 patients judged to be true cases of Stevens-Johnson syndrome. In conclusion, Stevens-Johnson syndrome is a uncommon condition. The excess risk of Stevens-Johnson syndrome due to any drug must, therefore, be very low.