Flucloxacillin dosing in critically ill patients with hypoalbuminaemia: special emphasis on unbound pharmacokinetics

Abstract

Objectives: To describe the total and unbound plasma concentration-time profiles for highly protein-bound flucloxacillin (95%-97% protein binding) in critically ill patients with hypoalbuminaemia and without severe renal dysfunction, and to use population pharmacokinetic modelling and Monte Carlo simulations to assess the probability of target attainment against an MIC distribution.

Patients and methods: Ten patients with hypoalbuminaemia and receiving flucloxacillin as part of therapy were enrolled. Sixty-seven total, 67 unbound plasma and 10 urine samples were collected and analysed. Population pharmacokinetic modelling of unbound plasma data and Monte Carlo simulations were then undertaken with NONMEM. Non-compartmental pharmacokinetic analysis was performed for total plasma concentrations.

Results: Total flucloxacillin V was increased in critically ill patients with hypoalbuminaemia 2-fold compared with healthy volunteer data. Unbound flucloxacillin concentrations after 2 g bolus fell below 1 mg/L 4 h after the end of the infusion, providing evidence that standard dosing would be insufficient for the treatment of methicillin-susceptible Staphylococcus aureus (MSSA) (MIC = 2 mg/L). Monte Carlo simulations suggest that continuous infusion of 8 g/24 h flucloxacillin would enable 100% successful attainment of the pharmacodynamic target, 50% fT( > MIC). For more aggressive targets (4-5x MIC for 100% fT( > MIC)), continuous infusion of higher doses (i.e. 12 g/24 h) would be required.

Conclusions: Administration of standard doses by intermittent bolus is likely to result in underdosing, and continuous infusion of higher doses is more likely to achieve pharmacokinetic-pharmacodynamic targets for the treatment of infections caused by the most common wild type of MSSA. Our data emphasize the importance of using unbound concentrations for determining dosage regimens for highly bound antibiotics.