| アブストラクト | Triamcinolone acetonide (TA) is a widely used glucocorticoid for inflammatory diseases, including arthritis. Although intra-articular TA injection provides effective symptom relief, its safety profile in real-world clinical practice remains incompletely characterized. Adverse event reports associated with intra-articular TA injection for arthritis were retrieved from the FAERS database (2004-Q1 2026) and the JADER database (2004-Q4 2025). Disproportionality analyses were performed using four signal detection algorithms, including ROR, PRR, BCPNN, and MGPS. Subgroup and sensitivity analyses were conducted to evaluate signal heterogeneity and robustness. Time-to-onset patterns were assessed using Weibull analysis. Endocrine-related adverse events were further evaluated using multivariable logistic regression and machine learning classification models with SHAP-based interpretation. A total of 738 FAERS reports involving intra-articular TA were identified, covering 27 system organ classes. Multiple established and potential safety signals were detected, including adrenal insufficiency, hyperglycemia, osteonecrosis, bacterial arthritis, postmenopausal hemorrhage, and Cushing's syndrome. Additional potential safety signals included type 2 diabetes mellitus, flushing, erythema multiforme, pemphigus, and septic shock. The strongest disproportionality signals were observed for injection site calcification, hypothalamic-pituitary-adrenal axis suppression, injection site atrophy, and Cushing's syndrome. Most adverse events occurred within 30 days after exposure (84.55%), with a median onset time of 5 days (IQR: 2-14). External signal consistency assessment using 112 JADER reports identified eight signals consistent with FAERS findings. Multivariable logistic regression showed that age, sex, therapy type, and reporter type were associated with endocrine-related adverse event reporting. Machine learning models demonstrated good classification performance, with random forest achieving the highest test set AUC (0.864). SHAP analysis identified demographic and treatment-related variables as important contributors to model classification. This study provides a comprehensive real-world pharmacovigilance assessment of intra-articular TA and identifies both established and potential safety signals. Integration of conventional signal detection, regression analysis, and interpretable machine learning approaches may facilitate systematic characterization of TA-associated adverse event patterns and support future safety monitoring. |
|---|
| ジャーナル名 | Naunyn-Schmiedeberg's archives of pharmacology |
|---|
| Pubmed追加日 | 2026/11/3 |
|---|
| 投稿者 | Yang, Yang; Xun, Jialin; Cui, Zhiwei; Song, Ping; Chen, Huayou; Wu, Xiaoling |
|---|
| 組織名 | Department of Obstetrics and Gynecology, The Second Affiliated Hospital of Xi'an;Jiaotong University, Xi'an, Shaanxi, 710004, China.;Tangdu Hospital, Air Force Military Medical University, Xi'an, China.;Department of Obstetrics and Gynecology, The First Affiliated Hospital of Xi'an;Jiaotong University, Xi'an, Shaanxi, China.;Department of Human Anatomy, Histology and Embryology, School of Basic Medical;Sciences, Xi'an Jiaotong University Health Science Center, Xi'an, Shaanxi, China.;Department of Hand and Foot Microsurgery, The Second Affiliated Hospital of;Hainan Medical University, Haikou, Hainan, China.;Jiaotong University, Xi'an, Shaanxi, 710004, China. wxlxjtu07@stu.xjtu.edu.cn. |
|---|
| Pubmed リンク | https://www.ncbi.nlm.nih.gov/pubmed/42972126/ |
|---|
