스타틴 관련 근육이상반응 보고와 관련된 요인 식별을 위한 머신러닝 기반 클러스터링 분석

초록

: We aimed to identify factors associated with adverse event (AE) reports in statin-associated muscle symptoms (SAMS) using hierarchical clustering of patients in the Korea Institute of Drug Safety and Risk Management - Korea Adverse Event Reporting System database (KIDS-KAERS DB) (2105A0027). Methods: To explore the characteristics and risk factors of SAMS reports, we analysed the KIDS KAERS DB from 2016 to 2020. We included reports with a causality category level of “possible” or higher. Hierarchical clustering analysis was used to identify distinctive patterns within the dataset, with a particular focus on variables such as sex, age, statin type, contraindicated drugs and concomitant drugs. The reporting characteristics were described according to the cluster. Results: Four clusters of AE reports were distinguished by hierarchical clustering: atorvastatin- and rosuvastatin associated AE (cluster 1), pitavastatin- and simvastatin-associated AE (cluster 2), rosuvastatin-associated AE (cluster 3), and atorvastatin-associated AE (cluster 4). Cluster 1 had a relatively higher proportion of men (57 cases, 50.9%) and a higher mean age (64.8 years) than the other clusters. Concomitant drug use was more common in cluster 1 (56 cases, 50.0%) than in other clusters (33.5%–46.2%), and all serious AEs were observed in cluster 1. Conclusion: Using hierarchical clustering, we found four distinct clusters based on SAMS report characteristics. Our findings further emphasize that patients prescribed statins, especially elderly male patients taking rosuvastatin and atorvastatin concomitantly with other medications, should be closely monitored for the development of rhabdomyolysis. (PeRM 2024;16:29-39)

키워드