Introduction
Upper urinary tract infections (UTI) are associated with critical illness and significant mortality. Meropenem is efficacious in treating UTI and has a systemic and urinary minimum inhibitory concentrations (MIC) breakpoint of 1 mg/L for Enterobacterales and 2 mg/L for Pseudomonas aeruginosa. As the fraction of meropenem excreted in the urine decreases with declining renal function, we applied the mechanistic kidney model in physiologically based pharmacokinetic (PBPK) simulation to predict the concentration time profiles of meropenem in plasma and urine within the kidney medullary collecting duct.
Methods
All modelling was performed with SimcypTM version 23 using published in vitro and clinical data derived from healthy adult volunteers, elderly and renally impaired non-infected and infected, critically ill populations. Organic anionic transporter (OAT-3) intrinsic uptake clearance (CLint,OAT-3) was estimated via in vitro-in vivo extrapolation while P-glycoprotein (P-gp) intrinsic efflux clearance (CLint,P-gp) and renal dihydropeptidase (DHP) metabolic clearance (CLint,DHP) were estimated by optimization of the mechanistic kidney model to recapitulate observed PK parameters to meet the 2-fold error criterion and study specific acceptance limits. Global sensitivity analysis was conducted and parameters total sensitivity index of > 0.1 were considered to be significant and optimized. Local sensitivity analyses was done to estimate the covariate values. Standard intermittent infusion doses were simulated to estimate the probabilities of target attainment (PTA) of meropenem concentrations above the MIC for at least 40% and 100% of the dosing interval (40%fT≥MIC=1, 40%fT≥MIC=2, 100%fT≥MIC=1 and 100%fT≥MIC=2 respectively) in 5000 virtual individuals each for the elderly, mild, moderate and severe renal impaired populations.
Results
The observed PK parameters were successfully recapitulated in 17 out of 21 studies with all predictions within the 2-fold error. CL,int,DHP was lower in the models of elderly, moderate and severe renal impaired (3.5, 4.0, and 2.0 L/h respectively) when compared to healthy volunteers and mild renal impaired (4.5 L/h). CLint,OAT-3 followed similar trends in elderly, moderate and severe renal impaired compared to healthy volunteers and mild renal impaired (0.07, 0.02, 1.548 × 10-3 ,0.1548 µL/min/106 proximal tubular cells respectively). CLint,P-gp remained at 0.012 µL/min/106 proximal tubular cells for all models and was not optimized as the total sensitivity index was < 0.0001, implying the clearance of meropenem was insensitive towards the change in CLint,P-gp. PTA of 100%fT>MIC=1 and 100%fT>MIC=2 in plasma was less than 50% in the mild and moderate renal impaired and 89-98% for severe renal impaired populations. It was less than 50% in European and Chinese geriatrics respectively. The PTA of 40%fT>MIC=1, 40%fT>MIC=2 in plasma and urine and 100%fT≥MIC=1,100%fT≥MIC=2 in urine was at least 99%.
Conclusion
The intrinsic clearance values of kidney OAT-3 and DHP were reduced in the elderly and moderate and severe renal impairment. Individuals with urinary tract and bloodstream infections are at risk of subtherapeutic levels when adopting a target of 100%fT>MIC=1 and 100%fT>MIC=2. Therapeutic drug monitoring of plasma concentrations could be considered for elderly patients and patients with mild and moderate renal impairment.