Abstract

Background Several ongoing clinical trials are evaluating high-dose rifampicin (up to 35 mg/kg) for tuberculous meningitis (TBM). However, rifampicin pharmacokinetics at higher doses is not fully characterized, particularly in cerebrospinal fluid (CSF), the site of TBM disease. Methods In a randomized controlled trial, adults with HIV-associated TBM were assigned to experimental arms of high-dose rifampicin (oral, 35 mg/kg; intravenous, 20 mg/kg) plus linezolid, with or without aspirin, or a control arm that received the standard of care with 10 mg/kg of oral rifampicin. Rifampicin concentrations, including the unbound fraction, were measured on plasma samples, and CSF was collected on days 3 and 28 of study enrollment. Data were analyzed by nonlinear mixed effects modeling. Results In total, 400 plasma and 44 CSF rifampicin concentrations from 48 participants were used for model development. The median (range) age and weight were 39 years (25–78) and 60 kg (30–107). Rifampicin pharmacokinetics was best described by a 2-compartment disposition model with first-order transit oral absorption and elimination via saturable hepatic extraction. Typical clearance values for the standard dose for days 3 and 28 were 33.1 and 41.4 L/h, respectively; high-dose values were 46.1 and 70.2 L/h. The CSF-plasma ratio was approximately 6% and the equilibration half-life was 3.2 hours. Simulated standard-dose rifampicin did not reach CSF concentrations above the critical concentration for Mycobacterium tuberculosis. Conclusions CSF penetration with standard-dose rifampicin is low. Our findings support continued evaluation of high-dose rifampicin for TBM treatment.