Viberg Anders, Cars Otto, Karlsson Mats O, Jönsson Siv
AstraZeneca R&D Södertälje, Clinical Pharmacology & DMPK, SE-151 85 Södertälje, Sweden.
J Clin Pharmacol. 2008 Nov;48(11):1270-81. doi: 10.1177/0091270008320923.
An approach for estimation of dosing strategies based on data-derived models and assessment of the risk associated with deviation from the treatment target is presented. The work is illustrated by establishing a dosing strategy to be used for a priori individualization on the basis of renal function for the antibiotic cefuroxime. Treatment involved exposing patients to concentrations above the minimum inhibitory concentration (MIC) for 50% of the dosing interval. The risk (penalty) function incorporated both deviations from the target and the use of excess amount of drug. Dosing strategies were estimated for a target population by minimizing the risk function. The population was characterized by a population pharmacokinetic model, and distributions of CLcr and body weight were reflective of the target group. The estimated dosing strategies were assessed by evaluating population distributions of (1) percentage of dosing interval with concentrations above MIC, (2) time of drug exposure below MIC, and (3) drug administered in excess to reach the target. These distributions were generated using wild-type MIC distributions for Escherichia coli and Streptococcus pneumoniae. The authors illustrate how benefits and risks of drug treatment can be weighed quantitatively in decision-based risk functions and subsequently used in the estimation of drug dosing.
本文提出了一种基于数据驱动模型估计给药策略以及评估与偏离治疗目标相关风险的方法。通过建立一种基于肾功能对头孢呋辛进行先验个体化的给药策略来说明这项工作。治疗包括使患者在50%的给药间隔内暴露于高于最低抑菌浓度（MIC）的浓度。风险（惩罚）函数既纳入了与目标的偏差，也纳入了过量使用药物的情况。通过最小化风险函数为目标人群估计给药策略。该人群由群体药代动力学模型表征，肌酐清除率（CLcr）和体重的分布反映了目标群体的特征。通过评估以下各项的群体分布来评估估计的给药策略：（1）浓度高于MIC的给药间隔百分比，（2）药物暴露于低于MIC的时间，以及（3）为达到目标而过量给药的情况。这些分布是使用大肠杆菌和肺炎链球菌的野生型MIC分布生成的。作者说明了如何在基于决策的风险函数中定量权衡药物治疗的益处和风险，并随后将其用于估计药物剂量。
