目的 挖掘美国食品药品管理局（FDA）不良事件报告系统（FAERS）数据库中达普司他相关药品不良事件信号，为临床安全用药提供参考。方法 提取2020年第2季度—2026年第2季度达普司他为首要怀疑药物的不良事件报告，采用报告比值比（ROR）、比例报告比（PRR）、贝叶斯置信传播神经网络（BCPNN）及多项伽玛-泊松缩减法（MGPS）进行信号挖掘，以ROR为主分析方法，其余3种方法进行交叉验证，并采用Cohen’s Kappa和Spearman秩相关评估方法间一致性；进一步按严重报告进行分层敏感性分析。结果 共纳入1 435例以达普司他为首要怀疑药物的病例，涉及3 795例次不良事件。报告频数较高的SOC包括各类损伤、中毒及操作并发症、各类神经系统疾病及良性、恶性及性质不明的肿瘤（包括囊状和息肉状）。4种方法共同检出111个阳性首选术语（PT）；按照预先规定的治疗背景、适应证及非临床事件排除规则进一步筛选后，信号主要涉及心血管（血栓栓塞）、代谢营养及血液系统，报告例次数排名前3位的信号为吞咽困难、心力衰竭、消瘦。ROR值居前3位为大脑微梗死、消瘦、中枢性甲状腺功能减退。ROR、PRR与BCPNN的阳性判定一致性较高，MGPS与其他方法的一致性较低至一般。结论 达普司他检测到涉及多个系统的不良事件信号，其中若干心脑血管/血栓、营养状态及血液学事件值得进一步评价。

Objective To identify signals related to adverse drug events associated with daprodustat from FAERS database, providing reference for safe clinical use. Methods Adverse event reports in which daprodustat was the primary suspect drug were extracted from the second quarter of 2020 to the second quarter of 2026. Signal detection was performed using ROR, PRR, BCPNN, and MGPS. ROR was used as the primary analytical method, while the other three methods were applied for cross-validation. Consistency among methods was assessed using Cohen’s Kappa and Spearman rank correlation. Subgroup sensitivity analyses were further conducted based on severity of reported events. Results A total of 1 435 cases where daprodustat was the primary suspected drug were included, involving 3 795 adverse events. The most frequently reported SOC included various injuries, poisoning and operational complications, various neurological diseases, benign, malignant and unknown nature tumors (including cystic and polypoid). The four methods jointly detected 111 positive primary terms (PTs), after further screening according to the predefined treatment background, indications and non-clinical event exclusion rules, the signals mainly involved cardiovascular (thromboembolism), metabolic-nutritional and hematological systems. The top three signals in terms of reported case numbers were dysphagia, heart failure and emaciation. The ROR values ranked the top 3 as cerebral microinfarction, emaciation, and central hypothyroidism. The positive determination consistency of ROR, PRR and BCPNN was relatively high, while the consistency between MGPS and other methods was lower to general. Conclusion Daprodustat detected adverse event signals involving multiple systems. Among them, several cardiovascular and cerebrovascular/thrombotic, nutritional status, and hematological events are worthy of further evaluation.

R978

海南省药物警戒中心关于药品安全性主动监测模式研究课题