Performance of Population Pharmacokinetic Models in Predicting Polymyxin B Exposures

Abstract

Polymyxin B is the last line of defense in treating multidrug-resistant gram-negative bacterial infections. Dosing of polymyxin B is currently based on total body weight, and a substantial intersubject variability has been reported. We evaluated the performance of different population pharmacokinetic models to predict polymyxin B exposures observed in individual patients. In a prospective observational study, standard dosing (mean 2.5 mg/kg daily) was administered in 13 adult patients. Serial blood samples were obtained at steady state, and plasma polymyxin B concentrations were determined by a validated liquid chromatography tandem mass spectrometry (LC-MS/MS) method. The best-fit estimates of clearance and daily doses were used to derive the observed area under the curve (AUC) in concentration-time profiles. For comparison, 5 different population pharmacokinetic models of polymyxin B were conditioned using patient-specific dosing and demographic (if applicable) variables to predict polymyxin B AUC of the same patient. The predictive performance of the models was assessed by the coefficient of correlation, bias, and precision. The correlations between observed and predicted AUC in all 5 models examined were poor (r² < 0.2). Nonetheless, the models were reasonable in capturing AUC variability in the patient population. Therapeutic drug monitoring currently remains the only viable approach to individualized dosing.

Keywords: area under the curve; pharmacokinetics; polymyxins; therapeutic drug monitoring.

Figure A1

An example of polymyxin B AUC correlation: predictions were based on the Sandri model [12] (highest correlation of the 5 models); r² = 0.197. The dashed line depicts the line of identity, y = x.

Figure 1

A typical pharmacokinetic profile: open circles depict observed data, and the dashed line represents the best-fit (one compartment) model; r² = 0.93.