With increasing multidrug resistance coupled to a poor development pipeline, clinicians are exploring antimicrobial combinations to improve treatment outcomes. In vitro hollow-fiber infection model (HFIM) is employed to simulate human in vivo drug clearance and investigate pharmacodynamic synergism of antibiotics. Our overarching aim was to optimize the HFIM-based pharmacokinetic (PK) assay by using rifampicin and polymyxin B as probe drugs. An ultrapressure liquid chromatography tandem mass spectrometry method was validated for the quantification of rifampicin and polymyxin B components. In vitro profiling studies demonstrated that the experimental PK profiles of polymyxin B monotherapy were well correlated with the human population PK data while monotherapy with rifampicin failed to achieve the expected maximum plasma concentration. Chemical stability studies confirmed polymyxin B was stable in broth at 37 °C up to 12 h while rifampicin was unstable under the same conditions over 12 and 80 h. The calculated mean clearance of rifampicin due to chemical degradation was 0.098 ml/min accounting for 12.2 % of its clinical total clearance (CL = 0.8 ml/min) based on population PK data. Our novel finding reinforces the importance to optimize HFIM-based PK assay by performing chemical stability study so as to account for potential discrepancy between experimental and population PK profiles of antimicrobial agents.