Impact on carbamazepine usage and cutaneous adverse reactions before and after the reimbursement of HLA-B*1502 genotyping in Taiwan, 2000-2017: A nationwide longitudinal study.

OBJECTIVE: The HLA-B*1502 allele is strongly associated with carbamazepine (CBZ)-induced Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN) in the Han Chinese population. This study investigated the impact of HLA-B*1502 screening on CBZ utilization and rates of severe cutaneous allergic reactions (SCARs) and SJS/TEN over time in Taiwan, where screening for HLA-B*1502 genotyping before prescribing CBZ was reimbursed in June 2010. METHODS: Using the Taiwan National Health Insurance Research Database, we analyzed 13 277 457 episodes of seeking treatment for epilepsy or neuralgia between 2000 and 2017. Episodes were categorized into quarters based on treatment time. Propensity score-based stabilized weighting (PSSW) ensured well-balanced covariates. The difference in 3-month SCAR and SJS/TEN rates between phase 2 (2011-2017) and phase 1 (2000-2009) was examined using a one-sample Z-test. Pearson correlation coefficients assessed the association between screening rate, the number of CBZ users and nonusers, and SCAR and SJS/TEN rates after HLA-B*1502 genotyping. RESULTS: CBZ prescriptions reduced from 7% (2000-2003) to 6% (2004-2010) and 4% (2011-2017). The screening rates of CBZ nonusers and CBZ users increased from 0%, .5% in 2011 to .8%, 16% in 2017, respectively. After PSSW, the mean 3-month SCAR incidence rates (per 10 000 episodes) significantly decreased from phase 1 to phase 2 for CBZ users (6.91 vs. 3.09, p < .0001) and nonusers (1.96 vs. 1.65, p < .0001). SJS/TEN incidence rates (per 10 000 episodes) significantly decreased from phase 1 to phase 2 for CBZ users (2.94 vs. 1.93, p < .0001) but not for nonusers (.71 vs. .74, p = .1492). In phase 2, SCAR incidence rates were significantly and negatively correlated with the screening rate for both CBZ users (r = -.38, p = .0342) and nonusers (r = -.80, p < .001). No significant correlation was found between SJS/TEN incidence rates and screening rates. SIGNIFICANCE: Recognizing HLA-B*1502 allele and avoiding CBZ therapy in HLA-B*1502-positive patients is critical for preventing CBZ-induced severe adverse events.