Modelling Renal Function in Drug Pharmacokinetics

Abstract

Many drugs or their metabolites are eliminated from the body by the kidneys. Renal function is an important covariate to describe variability in drug clearance (CL) of renally eliminated drugs. This thesis explores the use of renal function in population pharmacokinetic (PK) and pharmacodynamic (PD) models, and its influence in PK study design for critically ill children. A model for renal function is presented, defined by the ratio of estimated glomerular filtration rate (GFR) to normal GFR. It provides a continuous variable from neonates to adults that accounts for size, maturation and body composition consistently. This model is applied to a population PD analysis investigating the effect of antibiotic exposure on renal function and population PK analyses of cefazolin and milrinone, two renally eliminated drugs. Cardiopulmonary bypass (CPB) is associated with large physiological disturbances and can impact drug PK through changes in renal function, haemodilution, haemofiltration and drug adsorption to the CPB device. Cefazolin concentrations were quantified in 50 patients (aged 3 days – 13 years) undergoing cardiac surgery supported by CPB and were used to estimate in vivo PK parameters. Antibiotic adsorption to the CPB device was quantified using mechanistic binding principles in an ex vivo study and the results incorporated into the in vivo population analysis. CPB reduced cefazolin CL by 33% [95% CI 14.0, 46.7] above what could be predicted by size, age and renal function. Data from 50 patients (aged 4 days to 16 years) were used to investigate milrinone PK after cardiac surgery. Milrinone CL increased by 66.0%, 95% CI [47.9, 107] over time after bypass, with a half-time of 11.4 h, 95% CI [9.3, 13.7]. This likely reflects a return to homeostasis, and milrinone’s PD effects of increasing cardiac output and renal perfusion. PK in paediatric populations involves a complex interplay between intrinsic and extrinsic factors (e.g., size, age, renal function, CPB). The model for renal function in this thesis predicts variability in renal drug elimination and thus is expected to improve dose individualisation. Its utility is demonstrated in population PK and PD models with a focus on antibiotics, CPB, paediatric populations and critical illness.