Marco Albrecht (1), Stephan Schaller (1)
(1) esqLABS GmbH
Objectives: Fluid therapy is a common treatment in intensive care units (ICU). This has implications on the absorption, distribution, metabolism, and excretion (ADME) properties in critically ill patients, leading to inadequate treatment decisions at intensive care units (1). Some critical factors, which are responsive to fluid therapy and have an effect on drug ADME properties, are organ perfusion, vascular compartment size, and renal function. Also, the impact of fluid therapy on electrolytes and acid-base-balance, while known, is rarely considered in PBPK models even though it likely impacts compound distribution. While current PBPK software has a good representation of the steady-state in physiology, useful for most use-cases in drug development and regulatory authorization, the dynamic nature of rapidly changing physiology in intensive care patients requires a likewise dynamic PBPK software to assess the pharmacology of drugs in the ICU. Consequently, a modeling and simulation (M&S) framework is developed to predict the effects of fluid resuscitation on electrolyte distribution, acid-base balance, and volume shifts in ICU-related use cases. In an exemplary case study, a sensitivity analysis is performed to assess the impact of fluid resuscitation on the pharmacokinetics (PK) of dexamethasone. With bedside information on the conducted fluid therapies or a blood test for electrolyte balance, the shifts in ADME properties are calculated to predict changes in drug PK within the intensive care setting. The ultimate objective is the development of a model-based framework for decision support for drug therapy within the ICU.
Methods: To calculate volume shifts, the fluid balance model of Wolf, originally implemented in VisSim (2; 3), is used. The model has been re-implemented in R and then linked to the open-source and freely available PBPK software platform PK-Sim and MoBi to translate changes in physiology to the PBPK model parameter base. The Wolf model depends on anthropometric equations and has four compartments representing erythrocyte, plasma, interstitial, and cell volume. Equations for electrolytes, mass balance, and electrophysiology govern the volume balance across these compartments. To describe the balance between plasma and interstitial fluid, we use the revised Starling equation, which considers the endothelial glycocalyx layer and integrates with the PBPK model through the two-pore model (4), a formalism for endothelial transport of molecules, by changing endothelial pore radii and lymph flow rates (5; 6).
Results: The implemented R-model of fluid resuscitation has been validated on data occurring in the series of papers of Matthew B Wolf as well as from publications on volume therapy (2; 3; 4). This encompasses data on acid-base balance, hyperglycemia-induced hyponatremia, and hematocrit. The impact of the resulting changes in organ physiology on dexamethasone PK has been assessed and quantified to provide a prototype for further qualification studies.
Conclusions: The integration of electrolyte and fluid balance models such as the Wolf model and the revised Starling equation to a PBPK framework has proven an essential step to provide an integrated system for on-line PK predictions at an ICU ward. As a next step, the model shall be used as a digital twin to inform personalized drug dosing as a crucial component within control algorithms in medical devices at the bedside. The extended PBPK framework will integrate data as they occur at the bedside, extrapolate treatment consequences, and visualize both physiological and pharmacological key states for risk assessment. In the long term, the platform will be extended towards a companion diagnostics and risk assessment platform and assess common physiological disbalances for ICU related drugs, to alert for potential safety issues based on the state of a patient, and help guide treatment decisions to meet higher safety standards in intensive care.
References:
[1] Owen, Emily J, Gibson, Gabrielle A. and Buckman, Sara A. Surg. Infect. (2018) Pharmacokinetics and Pharmacodynamics of Antimicrobials in Critically Ill Patients.
[2] Wolf, Matthew B. J. Crit. Care (2017) Hyperglycemia-induced hyponatremia: Reevaluation of the Na+ correction factor.
[3] Wolf, Matthew B. Am. J. Physiol. Renal Physiol. (2013) Whole body acid-base and fluid-electrolyte balance: a mathematical model.
[4] Levick, J. Rodney and Michel, C. Charles. Cardiovasc. Res. (2010) Microvascular fluid exchange and the revised Starling principle.
[5] Rippe, Bengt and Haraldsson, Börje. Physiol. Rev. (1994) Transport of Marcomolecules Across Microvascular Walls: The Two-Pore Theory.
[6] Niederalt, Christoph, et al. J Pharmacokinet Pharmacodyn (2018) A generic whole body physiologically based pharmacokinetic modelfor therapeutic proteins in PK-Sims.
Reference: PAGE () Abstr 9433 [www.page-meeting.org/?abstract=9433]
Poster: Methodology - New Modelling Approaches