八种英夫利西单抗群体药动学模型在荷兰儿童炎症性肠病队列中的预测性能
该回顾性研究利用70名荷兰儿童炎症性肠病患者的443个英夫利西单抗浓度,评估八种群体药动学模型的预测性能,发现Fasanmade模型最适合整体数据集,其最大后验预测谷浓度≥5 mg/L的敏感性和特异性分别为83.5%和80%,AUC为0.870。
要点
- Fasanmade模型在荷兰儿童IBD队列中预测性能最佳,可用于临床。
- 多种模型预测性能可接受,模型引导的精准给药有助于维持目标谷浓度。
- Fasanmade模型预测谷浓度≥5 mg/L的AUC为0.870,敏感性和特异性分别为83.5%和80%。
核心结论
- 在荷兰儿童炎症性肠病队列中,多种英夫利西单抗群体药动学模型预测性能可接受,Fasanmade模型最适合临床使用。
研究设计与样本
- 回顾性研究,纳入70名荷兰儿童炎症性肠病患者,共443个英夫利西单抗浓度,评估八种群体药动学模型。
主要发现
- Fasanmade模型对总数据集最合适,治疗药物监测前后相对偏倚分别为-20.7%和11.2%,相对均方根误差分别为84.1%和51.6%。
- Fasanmade模型最大后验预测下一英夫利西单抗谷浓度≥5 mg/L的敏感性和特异性分别为83.5%和80%,AUC为0.870。
- 各模型间差异较小,多个模型被认为适合临床使用。
关键数值与推荐
- 谷浓度目标为≥5 mg/L。
- Fasanmade模型治疗药物监测前相对偏倚-20.7%,相对均方根误差84.1%;治疗后相对偏倚11.2%,相对均方根误差51.6%。
- Fasanmade模型预测谷浓度≥5 mg/L的敏感性83.5%,特异性80%,AUC 0.870。
临床怎么用
- 模型引导的精准给药可帮助维持英夫利西单抗谷浓度在目标范围内。
- 临床可选用Fasanmade模型预测儿童IBD患者英夫利西单抗剂量需求。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND AND OBJECTIVE: Efficacy of infliximab in children with inflammatory bowel disease can be enhanced when serum concentrations are measured and further dosing is adjusted to achieve and maintain a target concentration. Use of a population pharmacokinetic model may help to predict an individual's infliximab dose requirement. The aim of this study was to evaluate the predictive performance of available infliximab population pharmacokinetic models in an independent cohort of Dutch children with inflammatory bowel disease. METHODS: In this retrospective study, we used data of 70 children with inflammatory bowel disease (443 infliximab concentrations) to evaluate eight models that focused on infliximab pharmacokinetic models in individuals with inflammatory bowel disease, preferably aged ≤ 18 years. Predictive performance was evaluated with prior predictions (based solely on patient-specific covariates) and posterior predictions (based on covariates and infliximab trough concentrations). Model accuracy and precision were calculated with relative bias and relative root mean square error and we determined the classification accuracy at the trough concentration target of ≥ 5 mg/L. RESULTS: The population pharmacokinetic model by Fasanmade was identified to be most appropriate for the total dataset (relative bias before/after therapeutic drug monitoring: -20.7%/11.2% and relative root mean square error before/after therapeutic drug monitoring: 84.1%/51.6%), although differences between models were small and several were deemed suitable for clinical use. For the Fasanmade model, sensitivity and specificity for maximum posterior predictions for the next infliximab trough concentration to be ≥ 5 mg/L were respectively 83.5% and 80% with an area under the receiver operating characteristic curve of 0.870. CONCLUSIONS: In our paediatric cohort, various models provided acceptable predictive performance, with the Fasanmade model deemed most suitable for clinical use. Model-informed precision dosing can therefore be expected to help to maintain infliximab trough concentrations in the target range.
引用
- Bevers NC, Keizer RJ, Wong DR, et al. Performance of Eight Infliximab Population Pharmacokinetic Models in a Cohort of Dutch Children with Inflammatory Bowel Disease. Clin Pharmacokinet. 2024. PMID: 38488984. doi:10.1007/s40262-024-01354-7
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。