Revisiting Drug Dosage Optimisation for Renal Impairment Patients in the Era of Model-Based Precision Dosing

Abstract

Abstract Background: Renal impairment (RI) is a challenging disease as many patients take, on average, over ten daily medications. This requires care when dosing to obtain the optimal treatment outcome and to avoid disease progression. Drug labels are usually the source of dosage recommendations in RI. However, many drugs approved by the Food and Drug Administrationand marketed lack explicit dosing for RI. Usually, dedicated RI studies provide data at approval to complete the section on dosage recommendations in RI. From 2014 to 2019, many agents lacked a label for RI and especially, for severe renal impairment. Advanced technology and the utilisation of physiologically-based pharmacokinetic (PBPK) modelling and simulation have helped to address such issues. However, these models require detailed refinement and validation. Objectives: The objectives are to explore protein binding alterations in RI, to develop and validate a model to predict the fraction unbound (fu) in varying degrees of renal impairment, and to predict unbound clearance from the predicted fu protein levels. Method: A literature search was performed to collect data on the measurement of albumin and alpha-1-acid glycoprotein at varying degrees of RI. These data were used to develop a model that predicts levels of protein, considering glomerular filtration rate as the independent variable. To predict the fu, data were collected on the measurement of the fu at varying degrees of RI. The fu was predicted by scaling from healthy volunteers and accounting for changes in protein levels. Results: The fu predicted for albumin substrate was superior to that predicted for alpha-1-acid glycoprotein substrate as the majority of data were within a 2-fold error. Even though the model fails to capture the magnitude of change observed in the fu at varying degrees of RI, it seems to predict unbound clearance well since the majority of the predictions fall within a 2-fold error. Conclusion: This thesis has provided a deeper insight into the changes in protein levels in RI and for the first time, explored predicting the fu from these changes. This understanding is considered to be the first step towards improved refinement and to assist in developing more complex scenarios to account for affinity and more than one protein binding function, amongst other factors, which can be subsequently implemented in PBPK modelling.

Date of Award	31 Dec 2022
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Amin Rostami-Hochaghan (Supervisor), Adam Darwich (Supervisor) & Daniel Scotcher (Supervisor)