Ordway Research Institute, Albany, NY 12208, USA.
Curr Pharm Biotechnol. 2011 Dec;12(12):2044-61. doi: 10.2174/138920111798808428.
Efficacious therapy is of utmost importance to save lives and prevent bacterial resistance in critically ill patients. This review summarizes pharmacokinetic (PK) and pharmacodynamic (PD) modeling methods to optimize clinical care of critically ill patients in empiric and individualized therapy. While these methods apply to all therapeutic areas, we focus on antibiotics to highlight important applications, as emergence of resistance is a significant problem. Nonparametric and parametric population PK modeling, multiple-model dosage design, Monte Carlo simulations, and Bayesian adaptive feedback control are the methods of choice to optimize therapy. Population PK can estimate between patient variability and account for potentially increased clearances and large volumes of distribution in critically ill patients. Once patient- specific PK data become available, target concentration intervention and adaptive feedback control algorithms can most precisely achieve target goals such as clinical cure of an infection or resistance prevention in stable and unstable patients with rapidly changing PK parameters. Many bacterial resistance mechanisms cause PK/PD targets for resistance prevention to be usually several-fold higher than targets for near-maximal killing. In vitro infection models such as the hollow fiber and one-compartment infection models allow one to study antibiotic-induced bacterial killing and emergence of resistance of mono- and combination therapies over clinically relevant treatment durations. Mechanism-based (and empirical) PK/PD modeling can incorporate effects of the immune system and allow one to design innovative dosage regimens and prospective validation studies. Mechanism-based modeling holds great promise to optimize mono- and combination therapy of anti-infectives and drugs from other therapeutic areas for critically ill patients.
有效的治疗对于拯救生命和预防危重病患者的细菌耐药性至关重要。本综述总结了药代动力学（PK）和药效动力学（PD）建模方法，以优化经验性和个体化治疗中危重病患者的临床护理。虽然这些方法适用于所有治疗领域，但我们专注于抗生素，以突出重要的应用，因为耐药性的出现是一个重大问题。非参数和参数群体 PK 建模、多模型剂量设计、蒙特卡罗模拟和贝叶斯自适应反馈控制是优化治疗的首选方法。群体 PK 可以估计患者间的变异性，并考虑到危重病患者可能增加的清除率和大体积分布。一旦获得患者特定的 PK 数据，目标浓度干预和自适应反馈控制算法可以最精确地实现目标，例如感染的临床治愈或稳定和不稳定患者的耐药性预防，这些患者的 PK 参数变化迅速。许多细菌耐药机制导致预防耐药性的 PK/PD 目标通常比接近最大杀菌作用的目标高几倍。中空纤维和单室感染模型等体外感染模型允许研究抗生素诱导的细菌杀伤和单药和联合治疗的耐药性在临床相关治疗时间内的出现。基于机制（和经验）的 PK/PD 建模可以纳入免疫系统的影响，并允许设计创新的剂量方案和前瞻性验证研究。基于机制的建模为优化抗传染病和其他治疗领域药物的单药和联合治疗提供了巨大的前景，以满足危重病患者的需求。
