Yan Desheng, Ju Gehang, Liu Xin, Shao Qing, Zhang Yan, Wang Na, Yan Keyu
Department of Pharmacy, Xi'an Mental Health Center, Xi'an, Shaanxi, 710100, People's Republic of China.
Department of Clinical Pharmacology, Xiangya Hospital, Central South University, Changsha, Hunan, 410000, People's Republic of China.
Drug Des Devel Ther. 2024 Dec 27;18:6345-6358. doi: 10.2147/DDDT.S469149. eCollection 2024.
This study aimed to evaluate the predictive performance of published amisulpride population pharmacokinetic (PopPK) models in schizophrenia patients with an external data set and establish remedial dosing regimens for nonadherent amisulpride-treated patients.
A systematic search was conducted on PubMed, Embase, and Web of Science to identify PopPK models for evaluation. The evaluation process involved analyzing 390 serum concentration samples obtained from 361 Chinese adult inpatients diagnosed with schizophrenia. Model predictability was evaluated by prediction-based and simulation-based diagnostics. Based on validation results, a modified PopPK model was constructed to characterize amisulpride pharmacokinetic in our patients. Monte Carlo simulation was employed to investigate non-adherence scenarios and the impact of subsequently administered remedial regimens.
In the five assessed published models, four included trough concentrations from schizophrenia patients, and one combined single-dose data from healthy older adults and trough concentrations from older adults with Alzheimer's disease. The PE for population and individual predictions ranged from -92.89% to 27.02% and -24.82% to 4.04%, respectively. In the simulation-based diagnostics, the NPDE results indicated noticeable bias in all models. Therefore, a modified one-compartment model, with estimated creatinine clearance(eCLcr) as covariates on the apparent clearance (CL/F) of amisulpride, was developed. For delays in medication dosing, if the delay is within 12 hours, take half the missed dose right away, then resume the normal schedule; if the delay is up to 24 hours, just continue with the regular dosing schedule.
Existing published models lack the necessary reliability for cross-center application. Future prospective studies are required to assess our model before integrating it into clinical practice. Model-based simulations provided a rational approach to propose remedial strategies for delayed or missed doses.
本研究旨在利用外部数据集评估已发表的氨磺必利群体药代动力学（PopPK）模型在精神分裂症患者中的预测性能，并为未坚持服用氨磺必利治疗的患者制定补救给药方案。
在PubMed、Embase和Web of Science上进行系统检索，以确定用于评估的PopPK模型。评估过程包括分析从361名诊断为精神分裂症的中国成年住院患者中获得的390份血清浓度样本。通过基于预测和基于模拟的诊断方法评估模型的可预测性。基于验证结果，构建了一个改良的PopPK模型，以表征氨磺必利在我们患者中的药代动力学特征。采用蒙特卡洛模拟研究不依从情况以及随后给予的补救方案的影响。
在评估的五个已发表模型中, 四个纳入了精神分裂症患者的谷浓度，一个结合了健康老年人的单剂量数据和阿尔茨海默病老年人的谷浓度。群体预测和个体预测的预测误差（PE）分别为-92.89%至27.02%和-24.82%至4.04%。在基于模拟的诊断中，归一化预测分布误差（NPDE）结果表明所有模型均存在明显偏差。因此，开发了一个改良的单室模型，将估算的肌酐清除率（eCLcr）作为氨磺必利表观清除率（CL/F）的协变量。对于给药延迟，如果延迟时间在12小时内，立即服用错过剂量的一半，然后恢复正常给药计划；如果延迟时间长达24小时，则继续按常规给药计划给药。
现有的已发表模型缺乏跨中心应用所需的可靠性。在将我们的模型整合到临床实践之前，需要未来的前瞻性研究对其进行评估。基于模型的模拟为提出延迟或漏服剂量的补救策略提供了一种合理的方法。
