Risk of Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis After Initiation of Antiseizure Medication: A Danish Nationwide Cohort Study.

Abstract

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are serious cutaneous adverse reactions associated with several antiseizure medications, particularly lamotrigine and carbamazepine. We assessed the 90-day absolute risk of SJS/TEN after initiation of antiseizure medications in a nationwide Danish cohort and compared medication-specific risks with the general population background risk. All Danish residents initiating any of 25 antiseizure medications between 1995 and 2024 were included. Initiations and incident SJS/TEN diagnoses were identified through Danish registries. For each medication, we calculated the 90-day risk of SJS/TEN (cases per 100,000 initiators) with 95% CIs. Among cases, time from initiation to diagnosis was summarized using cumulative percentages, and 1-year mortality was assessed. General-population SJS/TEN cases were identified to estimate the 90-day background risk. Among 1,388,397 antiseizure medication initiations, the median age was 55.3 years (interquartile range [IQR] 40.2-69.9), and 56% were in female individuals. Within 90 days after initiation, 83 SJS/TEN cases occurred, most within the first weeks. One-year mortality among cases was 17%. Overall, these 83 cases accounted for 7% of all incident SJS/TEN cases in Denmark during the study period. The highest observed absolute risk was for lamotrigine (29.31 per 100,000 initiators; 95% CI 21.46-39.09), followed by carbamazepine (16.66; 95% CI 8.32-29.80), phenobarbital (15.84; 95% CI 5.14-36.96), oxcarbazepine (11.25; 95% CI 3.65-26.25), and valproic acid (7.31; 95% CI 2.68-15.91). Risks were lower for gabapentinoids, including pregabalin (2.26; 95% CI 0.83-4.92) and gabapentin (1.24; 95% CI 0.50-2.56). No SJS/TEN cases were observed for several newer antiseizure medications, although limited exposure for some precludes firm conclusions. The estimated 90-day background risk in the general population was 0.17 per 100,000 individuals (95% CI 0.16-0.18). This nationwide Danish study provides updated, population-based absolute risk estimates of SJS/TEN within 90 days of initiating 25 antiseizure medications in routine care and places these risks in the context of the general-population background risk. These estimates update and extend older evidence and provide a contemporary reference for SJS/TEN risk in antiseizure pharmacotherapy, including newer agents. The findings may also support risk communication around antiseizure medication initiation.