Sedation Strategies for Extracorporeal Membrane Oxygenation Support

Adkins, Kandis L.

doi: 10.1097/MAT.0000000000000536
Invited Commentaries

From the Department of Cardiovascular and Thoracic Surgery, University of Louisville School of Medicine, Louisville, Kentucky.

Submitted for consideration January 2017; accepted for publication in revised form January 2017.

Disclosures: The author has no conflicts of interest to report.

Correspondence: Kandis L. Adkins, Department of Cardiovascular and Thoracic Surgery, University of Louisville School of Medicine, 201 Abraham Flexner Way, Suite 1200, Louisville, KY. Email:

Article Outline

To live is to suffer, to survive is to find some meaning in the suffering.

Friedrich Nietzsche

My mission in life is not merely to survive, but to thrive; and to do so with some passion, some compassion, some humor, and some style.

Maya Angelou

Unfortunately, there exists no perfect sedative to fit all patients supported with extracorporeal membrane oxygenation (ECMO). Mainly, this is due to the wide variation in clinical indication for ECMO support, mode of support, individual patient temperament, and the unpredictable time to potential heart and/or lung recovery or acceptable organ procurement. An ideal sedative would be short-acting and easily titratable, provide dose-dependent hypnosis and amnesia and cause minimal change in hemodynamic and respiratory parameters. It would also have low reliance on renal and hepatic function for metabolism and excretion. No single agent meets these conditions for the wide variety of patients on ECMO support, but most conditions can be met with a combination of medications.

It is imperative to establish goals that the multidisciplinary care team can agree to support before designing a sedation regimen. This patient-specific strategy anticipates future interventions such as renal replacement therapy or bedside procedures, the need for spontaneous breathing trials or assessment of neurologic function, as well as participation in passive or active therapy. According to the Extracorporeal Life Support Organization (ELSO) guidelines, ECMO is indicated when the mortality risk from respiratory failure is at least 80% or, in the case of cardiogenic shock, when hypoperfusion is refractory to maximal medical therapy. It may be necessary to limit all responsiveness to control hemodynamic stability and promote recovery, as in the case of postcardiotomy cardiogenic shock. Once signs of improvement are identified, altering the Richmond Agitation-Sedation Scale (RASS) goal may be appropriate. Alternatively, in the case of isolated refractory acute respiratory failure requiring venovenous ECMO (vvECMO), wakefulness, physical therapy, and even extubation may be tolerated with minimal to no sedation. Research thus far has mostly combined the wide array of patients as individual center experience is limited. But with ECMO use increasing at experienced centers and the number of centers growing and participating in collaborative research, comparisons between similar groups can be studied in the future.

When designing a sedation strategy for an ECMO patient, utilizing a multimodal approach is ideal. Typically, using a combination of analgesics, anxiolytics/hypnotics, and antipsychotics will provide a safe environment while minimizing agitation, adverse reactions, and toxicity. Acute and chronic pain, as well as physical dependence on opiates or benzodiazepines commonly go unrecognized when patients are sedated. Negligence here can lead to withdrawal, adding to hemodynamic instability and making RASS goals difficult to achieve. Special attention must be given to the tolerant patient. By including opiates and/or benzodiazepines, withdrawal and poorly-controlled pain will be minimized, if not avoided altogether. As many centers have experienced and ex vivo studies have confirmed, a greater requirement for lipophilic medications, such as fentanyl, midazolam, and propofol should be expected due to sequestration and increased volume of distribution from added volume within the circuit.1 When the gastrointestinal tract is functioning, enteral medications should be included to promote stability while titrating shorter-acting intravenous medications. Using enteral medications may also lessen the effect of sequestration on drug dosing requirements. Regular assessment of sedation level and titration of short-acting medications will help avoid oversedation and drug accumulation as lower doses are needed. Further studies are needed to better define the changes in pharmacokinetics of enteral and intravenous drugs while on ECMO.

Verkoyen et al2 explore a novel approach, comparing inhaled isoflurane with propofol sedation in ECMO patients with parameters of overall outcome, time on ECMO, time sedated, ventilator settings, and incidence of delirium. As expected, and supported by this retrospective study, survival, duration of ECMO support, and length of stay (LOS) are independent of sedation strategy. Likely, the association resides with etiology of organ failure, time to ECMO support, and time to recovery or transplantation. The study also compares rates of spontaneous ventilation (bilevel positive airway pressure or continuous positive airway pressure tolerance) between the groups but does not equate spontaneous ventilation with any clinical benefit. It is the efficiency of gas exchange, not the act of inhaling and exhaling, which determines recovery and liberation from ECMO. Furthermore, it is unfair to compare the two as, by definition, inhaled anesthetics maintain minute ventilation by causing an increase in respiratory rate and small decrease in tidal volume, whereas propofol is a potent respiratory depressant readily causing apnea by reducing respiratory rate and tidal volume. This effect is augmented in the company of opiates but especially sufentanil, used by the authors for analgesia, with its added risk of profound chest wall rigidity. This group identifies another key point: in patients with respiratory failure so profound that even passive carbon dioxide exchange is limited, reliability of volatile gas exchange is questionable. Additionally, if lung failure includes severely reduced compliance refractory to bronchodilation, alveolar concentrations of inhaled agents may be insufficient, incapable of achieving desired RASS goal. This study includes patients supported on venoarterial extracorporeal membrane oxygenation, vvECMO and interventional lung assist and, again, these patients likely had varying sedation needs given the nature of heart or lung failure.

Although inhaled anesthetic gases offer the added benefits of bronchodilation, preservation of minute ventilation and cardioprotection over the intravenous and enteral sedatives, there are several factors that limit utilization in the intensive care unit. First, there exists little data on the prolonged exposure to inhaled volatile gases, including stability, safety, and potential risks to patients and staff members. Moreover, current usage is limited to the operating room setting, with limited exception, under the direct supervision of an anesthesiologist or nurse anesthetist. Outside of these training programs, there exists no formal instruction to establish competence for a critical care nurse handling volatile anesthetic gases. Similarly, a nonanesthesiologist intensive care physician cannot be expected to supervise an inhaled anesthetic, especially while simultaneously caring for other critically ill patients. Even with a dedicated specialist, the resources needed to provide this service seem cost-prohibitive and further investigation is certainly needed. With new technology available aimed at simplifying administration and monitoring of volatile drugs, further studies are warranted to prove applicability and feasibility in the intensive care unit, including confirmation of scavenging and disposal of waste. Finally, the inhaled volatile agents are not free of hemodynamic consequences. They, too, can lower systemic vascular resistance and cause hypotension and tachycardia. Granted, at higher doses than reported by Verkoyen et al,2 isoflurane can be pungent, and airway irritation would be detrimental in respiratory failure patients. There also exists the potentially catastrophic adverse reaction, malignant hyperthermia, and this small risk may outweigh any potential benefit.

Low-dose infusion of propofol offers many benefits when a part of a balanced sedative plan. It produces reliable hypnosis, is easily titratable (thanks to a relatively low context-sensitive half-life) and usage is not typically limited by organ dysfunction. Prolonged infusion of propofol, however, carries the risk of producing hypertriglyceridemia, propofol infusion syndrome, and pronounced hypotension which limits its clinical usefulness in the critically ill patient. Beyond the potential adverse reactions, concern has been expressed for detrimental interactions between propofol and the ECMO oxygenator. This goes beyond previously proven sequestration of drug in tubing, but involves a reduction in the lifespan of membrane oxygenators, suspected occlusion due to the high lipophilicity of propofol.3 In a well-designed retrospective analysis, Hohlfelder et al4 examine this issue and conclude that propofol infusion does not decrease oxygenator lifespan. They actually observed a longer oxygenator lifespan in patients on propofol but recognized many limitations of the study including small sample size, the retrospective nature, nonrandomization, and unmatched cohorts based on duration of ECMO support. As propofol was not randomly prescribed, selection bias is assumed and causality cannot be inferred to conclude prolonged oxygenator lifespan with propofol use. Fewer patients with cardiogenic shock received propofol, likely due to pre-existing hypotension and heart failure. Factors related to hemodynamic instability, such as transfusion requirement and limitations on anticoagulation, could account for the difference seen in duration of therapy and oxygenator lifespan. The authors questioned if longer duration of propofol infusion could shorten oxygenator lifespan, but this may be difficult to ascertain as longer duration of ECMO therapy itself is a risk factor for oxygenator failure. Future studies should aim to compare patients with similar indications who were supported longer than the manufacturer’s oxygenator lifespan estimate, in this case, 14 days.

In conclusion, several factors must be considered when selecting a sedation strategy including etiology of organ failure, patient’s mental and hemodynamic stability, and expected endpoint to ECMO course. Understanding all sedatives have the ability to cause harm; a multimodal approach combining enteral and parenteral analgesics, amnestics, and antipsychotics can minimize potential side effects and adverse reactions while creating an ideal environment for healing. Due to the complex nature and management requirements of inhaled volatile agents, they have a very limited role in prolonged sedation in the intensive care unit. Propofol, despite its many benefits, comes with added risks, namely profound hypotension, that can be avoided with the traditional opiate-benzodiazepine-antipsychotic combination. Finally, further investigation comparing goals of sedation while on ECMO, irrespective of specific agents used, would be useful, especially as collaborative, multicenter trials capable of pooling similar patient groups. Simultaneously, studies exploring pharmacokinetics and the interaction of ECMO circuitry with commonly used medications will provide better direction on ideal sedation strategies.

Back to Top | Article Outline


1. Shekar K, Roberts JA, Mcdonald CI, et al. Sequestration of drugs in the circuit may lead to therapeutic failure during extracorporeal membrane oxygenation. Crit Care. 2012.16: R194.
2. Verkoyen K, Schildhauer TA, Strauch JT, Swol J. The effects of propofol and isoflurane sedation on the outcomes of surgical patients receiving extracorporeal membrane oxygenation. ASAIO J. 2017.63: 174–178.
3. Lemaitre F, Hasni N, Leprince P, et al. Propofol, midazolam, vancomycin and cyclosporine therapeutic drug monitoring in extracorporeal membrane oxygenation circuits primed with whole human blood. Critical Care. 2015.19: 40.
4. Hohlfelder B, Szumita PM, Lagambina S, Weinhouse G, Degrado JR. Safety of propofol for oxygenator exchange in extracorporeal membrane oxygenation. ASAIO J. 2017.63: 179–184.
Copyright © 2017 by the American Society for Artificial Internal Organs