Aim
Build a physiologic based pharmacokinetic (PBPK) model for adult patients with ECMO that underpins the development of evidence-based guidelines for prescribing antimicrobial drugs during ECMO.
Background
Extracorporeal membrane oxygenation (ECMO) is an advanced life support system which allows for prolonged cardiopulmonary support in patients with life-threatening respiratory or cardiac failure(Bartlett, 2005; Bartlett & Gattinoni, 2010). It can be used as a bridge to either lung or heart transplantation, or to a long-term ventricular assist device. There are 2 types of ECMO – Venoarterial (VA) or Venovenous (VV) as shown in figure 1(Shekar, Fraser, Smith, & Roberts, 2012); in both the blood is drained from the venous system and oxygenated outside the body. In VV ECMO the blood is returned to the venous system; this is driven by the patients’ own heart function, hence maintaining pulsatile blood flow, and provides respiratory support only. While in VA ECMO the blood is returned to the arterial system bypassing both the heart and the lungs so it can be used for supporting both respiratory and cardiac failure.
Critically ill patients will often have a number of organ dysfunctions requiring support including mechanical ventilation and renal replacement therapy. These therapies together with the patients’ underlying pathophysiological conditions, are known to affect the pharmacokinetics (PK) of many drugs, such as the volume of distribution (Vd) and clearance (Cl)(Mousavi, Levcovich, & Mojtahedzadeh, 2011; Shekar, Fraser, et al., 2012). Where Vd refers to the volume of plasma in which the total amount of drug in the body would be required to be dissolved in order to reflect the drug concentration attained in the plasma; and Cl is the volume of plasma cleared of drug per unit time by the processes of metabolism and excretion.
ECMO is considered a supportive therapy, whilst drug therapy is used to treat the underlying disease and minimize complications.
| Our hypothesis: ECMO significantly alters the PK of antimicrobial drugs, thereby contributing to an elevated risk of therapeutic failure, drug toxicity and emergence of microbial resistance in critically ill adult patients on ECMO. |
Essentially, ECMO may alter the PK in a number of ways which include sequestration of the drug by the ECMO circuit(Shekar, Roberts, McDonald, et al., 2012; Wildschut, Ahsman, Allegaert, Mathot, & Tibboel, 2010), increasing the Vd, alteration in renal and liver blood flow, altered plasma protein binding and decreased drug elimination. The impact of these PK changes can lead to either therapeutic failure or toxicity. The physicochemical properties of the drug will determine to what extent its PK is altered; in that those lipophilic drugs are significantly sequestered in the circuit while hydrophilic drugs are significantly affected by hemodilution. Hence, the ECMO system itself may induce PK changes in a critically ill patient who already exhibits altered PK; the independent effects of ECMO add to the complexities in this group and are difficult to quantify.
There are a number of PK studies in patients with ECMO; the majority were performed in neonates and showed significant changes in the PK of antibiotic, sedative and analgesic drugs. These results cannot be extrapolated to adults as the physiologic processes that affect absorption, distribution, metabolism and excretion have not fully developed and are thus different to those in adults.
While the use of diuretics, anticoagulation, sedative, inotrope and vasopressor agents can be titrated to a measurable pharmacodynamic (PD) end-point, the use of anti-infectives cannot. Hence the application of PK principles in both selecting the appropriate anti-infective and its dosage regimen is important in optimizing outcome. Changes in antibiotic PK can lead to either therapeutic failure or toxicity in the patient; however, the potential emergence of resistant bacteria or fungi has wider implications.
There are a few case studies or small studies in adult patients with ECMO which have investigated vancomycin (Mulla & Pooboni, 2005), oseltamavir (Lemaitre et al., 2012; Lemaitre et al., 2010; Mulla et al., 2013), voriconazole (Ruiz et al., 2009; Spriet et al., 2009) and caspofungin(Ruiz et al, 2009); the conclusions are preliminary and in the case of vancomycin and caspofungin. They conclude with a recommendation of tailored dosing using therapeutic drug monitoring, however, not all of these drugs have their blood levels routinely measured e.g. caspofungin.
Currently there are no guidelines for dosing of drugs in patients with ECMO. The use of standard dosing in these critically ill patients can be sub-optimal, and ideally dosing should be individualized through the use of therapeutic drug monitoring.
There is a multi-centre, open label, descriptive PK study currently recruiting – ASAP ECMO: Antibiotic, Sedative and Analgesic Pharmacokinetics during Extracorporeal Membrane Oxygenation(Shekar, Roberts, Welch, et al., 2012). The authors will collaborate and recruit into this study concurrently to the research proposed by the author. This research will add to the evidence base for dosing in this group of patients. The PBPK model for an adult ECMO patient will be invaluable in predicting how these patients will handle other drugs in the future.
Anti-infectives are commonly administered to critically ill patients on the intensive care unit, including those with ECMO. The antifungals: voriconazole, posaconazole and caspofungin are within the first- and second-line treatment guidelines at the Royal Brompton & Harefield NHS Trust. Hence they are chosen as the first group of drugs to be studied to determine changes in their PK, which could lead to robust dosing guidelines in this group of patients.
Physiologically Based Pharmacokinetic (PBPK) modelling offers a mathematical modelling framework for integrating mechanistic data on absorption, distribution, metabolism and excretion to predict the time course of synthetic or natural chemical substances in humans(H. M. Jones, Parrott, Jorga, & Lave, 2006; H.M. Jones & Rowland-Yeo, 2013). These models are built using compartments or ‘building blocks’ which are connected by flow rates which mimic the circulating blood system; an example is illustrated schematically as follows (fig.2)(H.M. Jones & Rowland-Yeo, 2013)
Ultimately, these models can be used to predict the plasma concentration of a drug following intravenous or oral administration; hence they determine a dose in a disease population from that in a healthy volunteer if the model has the relevant physiological properties of this population incorporated(H. M. Jones, Mayawala, & Poulin, 2013).
Hence to develop a simulated PBPK model of a critically ill patient on ECMO support (as a separate compartment) the role of ECMO as an independent variable in a critically ill patient with altered PK can be quantified. The interplay with renal replacement therapies and ECMO can also be studied; in addition the differences between the two types of ECMO ie VV and VA can be defined.
Our proposed study will provide critical PK data to guide clinicians in administering “the right dose of the right drug, at the right time” during ECMO.
Phase 1 (currently in process) will determine the degree of sequestration of the commonly used anti-infectives posaconazole, voriconazole and caspofungin in the ECMO circuit.
An ECMO circuit will be set up as per the standard protocol for adult patients on ECMO, and will be maintained at a physiological temperature. For each drug (voriconazole, posaconazole and caspofungin):
- The drugs will be sourced from commercial stock.
- The drug will be added to the blood reservoir simulating actual drug administration in an adult patient. A baseline sample will be taken prior to the addition of the drug, followed by serial samples at 2, 5, 15, 30, 60, 120 and 360 minutes 12 and 24 hours after drug addition from a post-oxygenator site.
- The drugs will be quantified by mass spectrometry; the theoretical maximum concentration can be calculated by dividing the administered dose by the volume of the ECMO circuit. The percentage recovered can then be determined at the different time points and plotted as a concentration vs time graph. The lipophilicity or Log P (or the partition coefficient between 1-octanol and water) values for the individual drugs are derived from the University of Alberta Drugbank website. The correlation between the recovery rates and the drugs Log P value can be calculated using linear regression.
- 4 ex-vivo experiment runs per drug to determine the degree of mean loss of drug over 24 hour period
- Samples will be analysed at an accredited lab
The results of this phase will feed into the development of the PBPK model – this commenced in October 2016.
Phase 2 will build a simulated Physiologic Based Pharmacokinetic (PBPK) model for a critically ill patient with ECMO; the role of ECMO as an independent variable in a critically ill patient with altered PK can be quantified.
Mechanistic data obtained from our exvivo circuit studies (phase 1) will inform our PK models for patients. Initially we will identify the most important drug properties, ECMO and patient factors that affect the PK of our study drugs. These drug properties include lipophilicity, degree of ionization at physiological pH, molecular size, plasma protein binding, and stability; ECMO variables include type of ECMO [VA vs. VV], age of circuit components (such as pump, oxygenator and tubing), and priming of the device; and patient factors include critical illness, haemodilution, bleeding and transfusion, organ dysfunction, and Renal Replacement Therapy (RRT). We will then incorporate these effects into the PBPK models to be developed in this project.
- Choice of drug: Physicochemical properties of drugs along with patient and microbiological factors are key factors when prescribing for patients on ECMO. For example, midazolam and fentanyl (Fig 4) may not be ideal for ECMO patients as they are highly sequestered in the circuit.
- Dose, mode of administration (bolus vs. infusions) and dosing intervals: Adjusting the method of administration can minimise the effect of PK variability and hence will be considered during our guideline development.
- Dose adjustments for RRT during ECMO: No drug dosing guidelines during concomitant RRT and ECMO exist. ppreliminary data clearly showed that significant adjustments of the standard dosing may be required for some drugs.
Phase 3 will build the population PK models of the commonly used anti-infectives posaconazole, voriconazole and caspofungin and validate the PBPK model from Phase 2.
Following the results of the phases described above the validated PBPK and population PK models for an adult patient on ECMO can be employed to underpin the development of evidence-based guidelines for prescribing antimicrobial drugs during ECMO.
Dissemination of research findings
No immediate commercial benefit is foreseen with this research project; however, the knowledge gained from this research project will be disseminated at a number of National and International meetings such as the International Society of Heart & Lung Transplantation (ISHLT) annual meeting and published in peer-review journals.
The clinical transplant, critical care and pharmacy multidisciplinary teams at RBHFT will be kept up-to-date of research progress via monthly/quarterly meetings. There are a number of forums to present updates such as the transplant and critical care teaching sessions which attract a multidisciplinary audience, in addition to newsletters.
Patients and carers will be informed of the research via the RBHFT and Transplant unit website.
Results will be presented at the annual AHP, HCS and Nursing Poster day which is open to staff and the public to view and involves competitive selection with prizes for the top research projects.
Expected project outputs
While the use of diuretics, anticoagulation, sedative, inotrope and vasopressor agents can be titrated to a measurable pharmacodynamic (PD) end-point, the use of anti-infectives cannot.
Hence the application of PK principles in both selecting the appropriate anti-infective and its dosage regimen is important in optimizing outcome. Changes in antibiotic PK can lead to either therapeutic failure or toxicity in the patient; however, the potential emergence of resistant bacteria or fungi has wider implications. Currently there are no guidelines for dosing of drugs in patients with ECMO. The use of standard dosing in these critically ill patients can be sub- optimal, and ideally dosing should be individualized through the use of therapeutic drug monitoring.
The antifungals: voriconazole, posaconazole and caspofungin are within the first- and second-line treatment guidelines at RBHFT. Hence they are chosen as the first group of drugs to be studied to determine changes in their PK, which could lead to robust dosing guidelines in this group of patients. This research will add to the evidence base for dosing in this group of patients. The PBPK model for an adult ECMO patient will be invaluable in predicting how these patients will handle other drugs in the future.
Being a consultant pharmacist in transplantation & VADs at RBHFT, supported by the unit director, Dr Andre Simon, Dr Anna Reed Consultant in Respiratory and Transplant medicine and experienced advisors such as Dr Jason Roberts who have extensive research experience and reputations, I will be able to alter clinical practice from this research by implementing changes to patient management, ensuring that anti-infective dosing are the most effective for each individual. In doing this, I will help individuals have confidence in their anti-infective treatment regime.
