Developing predictive PK/PD models to optimise dosing regimens for antibiotics for gram-negative infections

Tran, H (2026) Developing predictive PK/PD models to optimise dosing regimens for antibiotics for gram-negative infections. Doctoral thesis, Liverpool John Moores University.

Preview

Text 2025tranphd.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (3MB) | Preview

Abstract

Pharmacokinetic/pharmacodynamic (PK/PD) modelling and simulation are increasingly recognised as powerful tools for optimising antibiotic dosing. This approach plays a critical role in enhancing the effectiveness of the older generation of antibiotics, preserving the efficacy of new generation agents and overcoming the increasing problem of antibiotic resistance. Yet, despite this potential, their use in clinical practice is still limited, especially in low- and middle-income countries. Here, a lack of patient data, the complexity of models, and the absence of supporting frameworks often result in suboptimal dosing and unfavourable outcomes. These issues become even more challenging when treating severe infections caused by multidrug-resistant Gram-negative bacteria (MDR GNB).
This thesis aims to provide practical solutions by developing population PK models from existing published data and integrating them with simplified PD models. Together, these models offer a framework to test dosing regimens that can be adapted to real-world clinical settings.
The work is presented in three phases. Phase I was a systematic review that identified PK/PD targets for antibiotics against MDR GNB, with the aim of determining which indices, whether EUCAST-recommended or more stringent thresholds, were most relevant for guiding dosing in clinical practice. Phase II examined current antibiotic use at two study sites: Cho Ray Hospital (Vietnam) and the Countess of Chester Hospital NHS Foundation Trust (UK), using retrospective data collected from both. Eligible data were analysed and simulations applied to assess the likelihood of achieving PK/PD targets across different MIC values. Phase III focused on meropenem and ceftazidime/avibactam, developing PK/PD model that incorporated assumptions of synergistic activity against KPC-producing Klebsiella pneumoniae. Simulations showed that no single regimen consistently achieved targets, underlining the limitations of empirical dosing.
Overall, this thesis provides a practical framework for dual antibiotic modelling and highlights the importance of precision, model-informed strategies to improve outcomes in MDR GNB infections.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	PK/PD, GNB, meropenem, ceftazidime/avibactam, modelling, simulations
Subjects:	R Medicine > RS Pharmacy and materia medica
Divisions:	Pharmacy and Biomolecular Sciences
Date of acceptance:	18 December 2025
Date of first compliant Open Access:	22 December 2025
Date Deposited:	22 Dec 2025 15:07
Last Modified:	22 Dec 2025 15:08
DOI or ID number:	10.24377/LJMU.t.00027616
Supervisors:	Madden, J and Penson, P
URI:	https://researchonline.ljmu.ac.uk/id/eprint/27616

View Item