Optimising tazobactam dosing in critically ill adult patients without renal replacement therapy: a population pharmacokinetic analysis of individual patient data pooled from multiple clinical studies

Background Critically ill patients exhibit altered β-lactam pharmacokinetics (PK), yet β-lactamase inhibitor exposure is often overlooked during dose optimisation. This study characterised the PK of tazobactam in critically ill patients and defined optimal dosing regimens that achieve different pharmacokinetic/pharmacodynamic (PK/PD) targets. Methods Tazobactam concentration-time data from nine clinical studies were analysed using nonlinear mixedeffects modelling with Monolix. One- and two-compartment models with linear elimination were evaluated, and covariates were identified. The probability of target attainment at steady state was evaluated across multiple dosing regimens for various PK/PD targets at estimated glomerular filtration rates (eGFR) of 40, 80, and 140 mL/ min. Results A total of 1771 plasma concentration-time data points for tazobactam from 421 patients were included in the analysis. A two-compartment model with linear elimination adequately described tazobactam PK. eGFR, weight and co-administered antimicrobial (piperacillin or ceftolozane) were identified as factors influencing tazobactam clearance. Standard dosing regimens achieved the conservative target of 20% fT> 1 mg/L at 80 mL/min and 140 mL/min. However, aggressive targets (50%, 85%, and 100% fT> 2 mg/L; 100% fT> 4 mg/L), which may be relevant for high-level β-lactamase expression, were rarely met in patients with an eGFR of 140 mL/min. Achieving these targets required higher tazobactam doses administered through continuous infusion. Conclusion Standard tazobactam dosing often failed to meet the aggressive PK/PD targets at higher eGFR, indicating a substantial risk of underexposure in critically ill patients with augmented renal clearance, supporting the need for dose optimisation, particularly in scenarios of high-level β-lactamase expression.