Abstract
BACKGROUND AND OBJECTIVE: As a renally-eliminated beta-lactam antibiotic, estimated glomerular filtration rate (eGFR) is a central component of piperacillin/tazobactam pharmacokinetics. This study aimed to determine the optimal eGFR equation for inclusion in population pharmacokinetic models for piperacillin and tazobactam among critically ill adults. METHODS: The study included critically ill adults treated with piperacillin/tazobactam at a single academic medical center between 2018 and 2022. Excluded individuals had acute kidney injury, received kidney replacement therapy, or had extracorporeal membrane oxygenation at piperacillin/tazobactam initiation. Non-linear mixed effects population pharmacokinetic models using eGFR equations with creatinine, cystatin C, or both were developed and compared. RESULTS: Using 377 samples from 120 critically ill patients, we found that a two-compartment model with first-order elimination best fit the piperacillin data and a one-compartment model with first-order elimination best fit the tazobactam data. For piperacillin, the final population mean parameters (standard error) were 8.36 (0.55) L/h for CL and 12.96 (1.17) L for V1. The values for Q and V2 were fixed at 0.98 L/h and 18 L, respectively, due to low interindividual variation in these parameters. For tazobactam, the final population mean parameters were 8.12 (0.52) L/h for CL and 16.87 (1.44) L for V. Both final models identified eGFR(cystatinC) expressed in mL/min as the eGFR equation that best predicted drug clearance as a covariate. CONCLUSIONS: An eGFR equation that includes cystatin C improves the predictive performance of pharmacokinetic models for piperacillin/tazobactam in critically ill adults.