Date of Award

2023

Document Type

Dissertation

Degree Name

Doctor of Pharmaceutical Sciences - Pharmaceutics

Department

Pharmaceutical Sciences

First Advisor

David Taft

Committee Chair and Members

David Taft, Chair

Kenneth Morris

Rutesh Dave

James Turong

Keywords

Beta-lactam antibiotics, Critical care illy sepsis patient, PBPK modeling, Pharmacokenitcs

Abstract

Sepsis leads to a continuous pathophysiologic process that dysregulates host response which potentially causes significant organ damage and even death, especially respiratory, renal and cardiac failure. Changes in the clearance and volume of distribution are typically result in pharmacokinetic variabilities and often make it challenging for clinicians to appropriately dose beta-lactam antibiotics. Therapeutic drug monitoring (TDM) is necessary to further optimize beta-lactam pharmacokinetic/pharmacodynamic (PK/PD) parameters to improve clinical outcomes and reduce therapeutic failure.

Cefepime (CEF) and meropenem (MEM) primarily excreted by the kidney, and altered the kidney function of the ICU patients would affect the clearance of the medications. On the other hand, the volume of distribution is usually affected by differences in plasma protein binding and body composition in the septic population. For body composition, the following parameters are considered to contribute the changes in volume of distribution: total body water, extracellular water, intracellular water, total body fat and total body protein. In this research, a mechanistic approach to characterize CEF and MEM absorption, distribution, metabolism and excretion based on the anatomy and physiology of the human body in conjunction with physicochemical properties of the medication and information is needed. PBPK modeling is a good tool to use for this purpose. This dissertation research demonstrated how PBPK modeling and simulation using Simcyp® is a useful tool to predict the pharmacokinetics of cefepime and meropenem. The PBPK model developed in this investigation was designed to predict systemic exposure in the targeted population. Moreover, the PBPK modeling approach in this study was tested in a real-world setting by predicting drug exposure in individual patients when limited plasma sampling was performed.

The new and validated serum assay method with ultra-performance liquid chromatography- photodiode array detector (UPLC-PDA) was developed and used to analyze patients’ blood samples at low concentration levels as an alternative to more resource and time intensive LCMS methods, which can also be used to test the other antibiotics.

A critical care septic population was developed based on the altered subcutaneous/intramuscular fractions and use to capture the changes in the drug distribution induced by the tissue composition changes under the sepsis situation. The model was further modified by adding different physiologic perturbations in ICU sepsis patients. Among these, impaired renal function and a decreased fraction of cardiac output are the most common seen situations happened to the ICU sepsis patients. On the other hand, some ICU sepsis patients have been obese, which causes an increased GFR.

Verification of the developed critical care renally impaired sepsis population was conducted as part of this research. The developed population included altered body composition fractions, modified kidney size for a renal insufficiency function, and decreased cardiac output to create the final renally impaired sepsis population. PBPK model simulated systemic profiles were compared with published clinical data obtained from sepsis patients with impaired kidney function for the probe medication, vancomycin. The calculated fold errors associated with the AUC for both single and multiple doses indicate a good predictability of the model in the developed critical care septic population.

Likewise, the cefepime and meropenem PBPK models developed for healthy volunteers captured observed plasma concentrations from multiple publications evaluating different dose regimens of the medications. The observed PK profiles fitted well with model simulation. Especially, the model simulated elimination phase and Cmin very well, which are helpful to predict the Ctrough if requires. Cmax and area under curve (AUC) are also comparable with publication reported values. Prediction errors calculated for PK parameters suggest good predictability of two models in this population. The developed models can be used for the predictions of the target populations.

The developed and verified PBPK models were then used to predict the plasma exposure of cefepime and meropenem when empiric-dosing regimens were administrated to critically ill sepsis patients in a clinical setting conducted at The Brooklyn Hospital Center (TBHC). The simulations preformed for individual ICU septic patients based on their demographic data were compared with clinical data collected from TBHC. This study enrolled 15 patients including 8 patients treated with CEF, 7 patients treated with MEM. The developed PBPK model captured observed plasma concentrations well from 7 renally impaired sepsis patients. The remaining subjects were obese patients, who all had an increased GFR but the different performance in volume distribution. The simulations for the majority of the subjects were acceptable. For the subjects for which the developed model was not predictive, an extremely small volume distribution in these patients could be a contributing factor.

Overall, this dissertation research highlights the potential of a PBPK modeling approach to evaluate the PK and PK/PD targets in critically ill septic patients treated with beta-lactam antibiotics. Once verified, the model can be used to further optimize empiric dosing of medications in a critical care population with the goal of delivering safe and effective drug therapy in these patients.

Download

Share

COinS