Defining the unbound target exposure of teicoplanin, Part 2: Dose individualization through population pharmacokinetic modelling

Objectives The pharmacokinetic index of teicoplanin is best described by the area of the unbound concentration under the concentration-time curve (fAUC). In clinical practice, the total trough concentration is used for the purpose of therapeutic drug monitoring. An fAUC during 24 hours (fAUC24) of 75 mg h/L has been proposed as a target. This study aimed to develop a population pharmacokinetic model based on unbound teicoplanin concentrations to evaluate and improve teicoplanin dosing. Methods In this observational clinical trial, 30 patients admitted to the intensive care unit or haematology department were included. Dosing was at the discretion of the treating physicians. Samples were taken on day 2 (T = 0, 1, 2, and 4 hours) and day 5 (T = 0, 4, 8, 16, and 24 hours) of therapy. Teicoplanin pharmacokinetics were characterized by nonlinear mixed-effects modelling. The final model was used to simulate the probability of achieving an fAUC24 of 75–150 mg h/L under current dosing regimens, and to evaluate the potential improvement through model-informed precision dosing. Results In total, 188 samples were collected from 30 patients. The median full body weight was 80 kg [interquartile range (IQR): 71–90 kg], the median estimated glomerular filtration rate (eGFR) (Chronic Kidney Disease-Epidemiology Collaboration) was 88 mL/min [IQR: 46–117 mL/min], and the median serum albumin level was 19.5 g/L [IQR: 16–24 g/L]. Median unbound fraction was 21% [IQR: 16–23%]. A two-compartment pharmacokinetic model with eGFR as a covariate for unbound clearance and albumin as a covariate for Bmax best fitted the data. Simulations showed that the probability of target attainment in intensive care unit patients is 58%. Failure to achieve the target was predominantly observed in patients with elevated eGFR. Simulations showed that probability of target attainment can be augmented to 71% with improved dosing schedules and to 91% when the model is applied to model-informed precision dosing. Conclusions We developed a pharmacokinetic model capable of guiding teicoplanin dosing based on fAUC24. Current dosing schedules are suboptimal from a pharmacokinetic viewpoint. Model-informed precision dosing of unbound teicoplanin is advised to maximize target attainment.