Therapeutic drug monitoring of polymyxin B cerebrospinal fluid concentrations in patients with carbapenem-resistant Gram-negative bacteria-induced central nervous system infection

Abstract Objectives Central nervous system (CNS) infections caused by carbapenem-resistant Gram-negative bacteria (CR-GNB) present a major health and economic burden worldwide. This multicentre prospective study aimed to assess the feasibility and usefulness of CSF therapeutic drug monitoring (TDM) after intrathecal/intraventricular administration of polymyxin B in patients with CNS infections. Methods Forty-two patients treated with intrathecal/intraventricular administration of polymyxin B against CR-GNB-induced CNS infections were enrolled. CSF trough level (Cmin) was collected beginning on Day 2 post-polymyxin B initiation and thereafter. The primary outcomes were clinical cure and 28-day all-cause mortality. Results All patients started with intrathecal/intraventricular administration of polymyxin B at a dose of 5 g/day, corresponding to a median CSF Cmin of 2.93 mg/L (range, 0.21–25.74 mg/L). Clinical cure was 71.4%, and the median CSF Cmin of this group was higher than that of clinical failure group [3.31 (IQR, 1.73–5.62) mg/L versus 2.25 (IQR, 1.09–4.12) mg/L; P = 0.011]. In addition, with MICs ≤ 0.5 mg/L, maintaining polymyxin B CSF Cmin above 2.0 mg/L showed a higher clinical cure rate (P = 0.041). The 28-day all-cause mortality rate was 31.0% and had no association with CSF Cmin. Conclusions After intrathecal/intraventricular administration of polymyxin B, CSF concentrations fluctuated considerably inter- and intra-individual. Polymyxin B CSF Cmin above 2.0 mg/L was associated with clinical cure when MICs were ≤ 0.5 mg/L, and the feasibility of TDM warrants additional clinical studies.