Sick patients, starved circuits and sticky drugs: understanding pharmacokinetics during extracorporeal membrane oxygenation to optimise drug dosing and improve patient outcomes

Abstract

Listen

Translate article

Extracorporeal membrane oxygenation (ECMO) is the final option for patients with acute severe cardiac and/or respiratory failure that is unresponsive to conventional management. ECMO plays a supportive role and its success relies on optimal drug therapy to reverse the underlying disease process and to prevent or treat complications. Vital drugs may be sequestered and/or degraded in the ECMO circuit resulting in altered pharmacokinetics (PK). This may be further complicated by the PK changes that occur in the presence of critical illness and may lead to therapeutic failure and/or drug toxicity. However, much of the currently available PK data is from neonates receiving ECMO and may not be extrapolated to adult patients given the developmental and physiologic differences. Equally, such PK alterations are complex and are challenging to investigate in a critically ill patient on ECMO. The aim of this research was to investigate each of the circuit, drug and critical illness factors affecting PK during ECMO in adult patients. A combination of linear and non-linear mixed effects modelling, compartmental and population methods and dosing simulations was utilised to characterise antibiotic, sedative and analgesic PK. The incremental research plan comprised of; (i) ex vivo experiments for drug stability testing in human blood; (ii) ex vivo drug disposition studies in ECMO circuits primed with fresh whole human blood; (iii) PK studies in healthy and critically ill ovine models of ECMO with appropriate non-ECMO controls and; (iv) an international multi-centre clinical population PK study in critically ill adults on ECMO. The introductory chapter (Chapter 1) outlines the origins of this research based on clinical observations, hypothesis generating preliminary data (Chapter 1.1), summarises the available extracorporeal life support therapies (Chapter 1.2) and provides a clinical context to PK alterations induced by ECMO (Chapter 1.3). This is followed by a literature review on PK alterations during ECMO (Chapter 2). Thus, an understanding of the clinical problem, available extracorporeal therapies, and identifying the gaps in literature (Chapter 3) facilitated the development of a comprehensive research plan (Chapter 4.1). Chapter 4.2 describes the methodology for the development of ex vivo and ovine models of ECMO. The protocol for the international multi-centre population PK study that aims to develop dosing guidelines for 18 key antibiotic, sedative and analgesic drugs is presented in Chapter 4.3. Ex vivo studies identified drug stability, lipophilicity (Chapter 5.1) and protein binding (Chapter 5.2) as the key drug factors determining their sequestration and /or degradation in ECMO circuits. An increase in peripheral volume of distribution (Vd) of the highly protein-bound, lipophilic drug midazolam was identified in both healthy and critically ill sheep on ECMO (Chapter 6.1) providing further in vivo evidence of circuit drug sequestration. PK of eight antibiotic drugs that exhibited a wide range of lipophilicity and protein binding was characterised in healthy sheep controls and healthy and critically ill sheep on ECMO (Chapter 6.2). This experiment identified that ECMO may variably influence Vd and decrease clearance (CL) in the presence of critical illness, based on lipophilicity and protein binding characteristics and these properties may be used to predict their PK during ECMO. Population PK modelling of meropenem in critically ill patients on ECMO (with or without renal replacement therapy) was performed and compared with data from non ECMO controls in a matched-cohort study (Chapter 7). This study provided preliminary evidence that in the absence of significant circuit sequestration for hydrophilic and less protein-bound drugs such as meropenem, an increased Vd (critical illness and ECMO factors) may be to an extent countered by a decreased CL (from ECMO and kidney injury) minimising the net influence ECMO has on PK. The closing chapter (Chapter 8) discusses key findings of this research and lays a roadmap for future work. The overarching finding of this research is that drug factors lipophilicity and protein binding play a significant role in altering PK during ECMO and drug factors may be used to predict PK of antibiotic and sedative drugs. The combined effect of critical illness and ECMO may be limited for hydrophilic and less protein-bound drugs compared with lipophilic and more protein-bound drugs. The ongoing clinical population PK study will provide robust dosing guidelines for prescription of commonly used antibiotics, sedative and analgesic drugs during ECMO. Future research should focus on optimising the interactions between drug, device and the disease. While it may be challenging to alter drug physicochemistry, whilst maintaining safety and efficacy, further refinements to ECMO circuitry may be considered. Equally, the impact of ECMO on pathophysiology, especially on hepatic and renal organ systems which are responsible for metabolism and excretion of most drugs need to be better understood. Building pharmacodynamic (PD) models based on animal and clinical studies will add another dimension of PK/PD during ECMO as the PD alterations induced by ECMO, if at all, are yet to be investigated.