Abstract
Purpose: Optimal dosing of meropenem and piperacillin/tazobactam in critically ill patients receiving renal replacement therapy (RRT) is uncertain due to variable pharmacokinetics. We aimed to develop generalisable optimised dosing recommendations for these antibiotics.
Methods: Prospective, multinational pharmacokinetic study including patients requiring various forms of RRT. Independent population PK models were developed, externally validated and applied to perform Monte Carlo dosing simulations using Monolix and Simulx. We calculated the probability that these dosing regimens achieved standard and high therapeutic unbound antibiotic concentrations over 100% of the dosing interval for the treatment of Enterobacterales and Pseudomonas aeruginosa.
Results: We enrolled 300 patients from 22 intensive care units across 12 countries receiving continuous veno-venous haemodialysis (13.0%), haemofiltration (23.3%), haemodiafiltration (48.4%) or sustained low-efficiency dialysis (15.3%). Models were developed using data from 234 patients (8322 samples) and validated with 66 additional patients (560 samples). Predictive performance was high, with mean prediction errors of - 5.2% for meropenem and - 16.9% for piperacillin. Dosing simulations showed that meropenem and piperacillin/tazobactam dosing requirements were dependent on urine output and RRT intensity and duration (p < 0.05). In all scenarios, extended/continuous infusions led to a better achievement of effective concentrations with lower daily doses compared to short infusion. Dosing nomograms were developed to inform dosing for different RRT settings, urine outputs, and target concentrations.
Conclusion: RRT intensity and duration and urine output determine meropenem and piperacillin/tazobactam dosing requirements in critically ill patients receiving RRT. Extended/continuous infusions facilitate the attainment of effective concentrations.