Tuberculous meningitis (TBM) is the most devastating form of TB, resulting in death or neurological disability in up to 50% of patients affected. Treatment is similar to that of pulmonary TB, despite poor cerebrospinal fluid (CSF) penetration of the cornerstone anti-TB drug rifampicin. Considering TBM pathology, it is critical that optimal drug concentrations are reached in the meninges, brain and/or the surrounding CSF. These type of data are difficult to collect in TBM patients. This review aims to identify and describe a preclinical model representative for human TBM which can provide the indispensable data needed for future pharmacological characterization and prioritization of new TBM regimens in the clinical setting. We reviewed existing literature on treatment of TBM in preclinical models: only eight articles, all animal studies, could be identified. None of the animal models completely recapitulated human disease and in most of the animal studies key pharmacokinetic data were missing, making the comparison with human exposure and CNS distribution, and the study of pharmacokinetic-pharmacodynamic relationships impossible. Another 18 articles were identified using other bacteria to induce meningitis with treatment including anti-TB drugs (predominantly rifampicin, moxifloxacin and levofloxacin). Of these articles the pharmacokinetics, i.e. plasma exposure and CSF:plasma ratios, of TB drugs in meningitis could be evaluated. Exposures (except for levofloxacin) agreed with human exposures and also most CSF:plasma ratios agreed with ratios in humans. Considering the lack of an ideal preclinical pharmacological TBM model, we suggest a combination of 1. basic physicochemical drug data combined with 2. in vitro pharmacokinetic and efficacy data, 3. an animal model with adequate pharmacokinetic sampling, microdialysis or imaging of drug distribution, all as a base for 4. physiologically based pharmacokinetic (PBPK) modelling to predict response to TB drugs in treatment of TBM.

中文翻译：

---

优化结核性脑膜炎治疗的临床前模型——系统评价

结核性脑膜炎 (TBM) 是最具破坏性的结核病，导致多达 50% 的患者死亡或神经功能障碍。尽管基石抗结核药物利福平的脑脊液 (CSF) 渗透性较差，但治疗方法与肺结核相似。考虑到 TBM 病理，在脑膜、大脑和/或周围 CSF 中达到最佳药物浓度是至关重要的。这些类型的数据很难在 TBM 患者中收集。本综述旨在确定和描述代表人类 TBM 的临床前模型，该模型可为未来临床环境中新 TBM 方案的药理学表征和优先级排序提供必要的数据。我们回顾了关于在临床前模型中治疗 TBM 的现有文献：只有八篇文章，所有动物研究，可以识别。没有动物模型完全概括了人类疾病，并且在大多数动物研究中缺少关键的药代动力学数据，这使得与人体暴露和 CNS 分布的比较以及药代动力学-药效学关系的研究变得不可能。另外 18 篇文章被鉴定为使用其他细菌诱发脑膜炎，治疗包括抗结核药物（主要是利福平、莫西沙星和左氧氟沙星）。在这些文章中，可以评估脑膜炎中结核病药物的药代动力学，即血浆暴露和脑脊液：血浆比率。暴露量（左氧氟沙星除外）与人体暴露量一致，而且大多数脑脊液：血浆比率也与人体中的比率一致。考虑到缺乏理想的临床前药理学 TBM 模型，我们建议结合 1。