Li Hongmei, He Xiaojun, Chen Cui, Ni Qiao, Ni Linghao, Zhou Jiawei, Peng Bin
Department of Health Statistics, College of Public Health, Chongqing Medical University, Chongqing 401331, China.
Pharmaceuticals (Basel). 2025 Jul 22;18(8):1084. doi: 10.3390/ph18081084.
Pulmonary arterial hypertension (PAH) is a progressive and life-threatening disease. Adverse events (AEs) related to its drug treatment seriously damaged the patient's health. This study aims to clarify the causal relationship between PAH drugs and these AEs by combining pharmacovigilance signal detection with the Bayesian causal network model. Patient data were obtained from the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS), covering reports from 2013 to 2023. In accordance with standard pharmacovigilance methodologies, disproportionality analysis was performed to detect signals. Target drugs were selected based on the following criteria: number of reports (a) ≥ 3, proportional reporting ratio (PRR) ≥ 2, and chi-square (χ) ≥ 4. Bayesian causal network models were then constructed to estimate causal relationships. The do-calculus and adjustment formula were applied to calculate the causal effects between drugs and AEs. Signal detection revealed that Ambrisentan, Bosentan, and Iloprost were associated with serious AEs, including death, dyspnea, pneumonia, and edema. For Ambrisentan, the top-ranked adverse drug events (ADEs) based on average causal effect (ACE) were peripheral swelling (ACE = 0.032) and anemia (ACE = 0.021). For Iloprost, the most prominent ADE was hyperthyroidism (ACE = 0.048). This study quantifies causal drug-event relationships in PAH using Bayesian causal networks. The findings offer valuable evidence regarding the clinical safety of PAH medications, thereby improving patient health outcomes.
肺动脉高压（PAH）是一种进行性且危及生命的疾病。与其药物治疗相关的不良事件（AE）严重损害了患者的健康。本研究旨在通过将药物警戒信号检测与贝叶斯因果网络模型相结合，阐明PAH药物与这些不良事件之间的因果关系。患者数据来自美国食品药品监督管理局（FDA）不良事件报告系统（FAERS），涵盖2013年至2023年的报告。按照标准的药物警戒方法进行不成比例分析以检测信号。根据以下标准选择目标药物：报告数量（a）≥3、比例报告率（PRR）≥2且卡方（χ）≥4。然后构建贝叶斯因果网络模型来估计因果关系。应用干预演算和调整公式来计算药物与不良事件之间的因果效应。信号检测显示，安立生坦、波生坦和伊洛前列素与严重不良事件相关，包括死亡、呼吸困难、肺炎和水肿。对于安立生坦，基于平均因果效应（ACE）排名靠前的药物不良事件（ADE）是外周水肿（ACE = 0.032）和贫血（ACE = 0.021）。对于伊洛前列素，最突出的ADE是甲状腺功能亢进（ACE = 0.048）。本研究使用贝叶斯因果网络对PAH中药物 - 事件的因果关系进行了量化。这些发现为PAH药物的临床安全性提供了有价值的证据，从而改善患者的健康结局。
