Risk of toxic epidermal necrolysis and Stevens-Johnson syndrome associated with benzodiazepines: a population-based...

We aim to estimate the incidence of Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) among tetrazepam users and compare it with benzodiazepine users in a Spanish primary care database (BIFAP). The incidence in the general population (GenPop) and among phenytoin new users (as a positive control) was also estimated.

We identified a cohort of GenPop free of SJS/TEN (N = 3,155,364). Cohort entry was the date after 1 year of register with the physician during 2001-2011. No age restrictions were applied. Patients were followed from entry up to the first of the following: a record of SJS/TEN (potential cases), death, end of information, or December 2011. History of potential cases were manually reviewed blinded to exposure and considered "probable" when diagnosed in referral reports. Three cohorts of patients newly prescribed with benzodiazepines (N = 531,813), tetrazepam (N = 343,568), or phenytoin (N = 4993) were extracted from the GenPop cohort. Incidence rate (cases per million person-years (py)) for the GenPop and cumulative incidence (per million new users) during the first 9 weeks after each drug prescription were computed.

In the GenPop, 48 probable cases (38 SJS and 10 TEN) were identified (3.21/million py; 3.37 in men and 2.94 in women). In the benzodiazepines cohort, 2 probable TEN cases was identified (3.76/mill.). In the tetrazepam cohort, 1 probable SJS/TEN case was identified (2.91/mill.). In the phenytoin cohort, 4 probable cases (2 SJS and 2 TEN) were identified (801.12/mill.).

The incidence of SJS/TEN in tetrazepam users was very rare and similar to benzodiazepines users. The incidence in the GenPop and among users of phenytoin agreed with the literature.