This editorial appears in the June 2023 issue of the American Journal of Bioethics
Childress and coauthors present a case considering ECMO withdrawal over the objection of the conscious patient who is no longer a candidate for transplantation or other definitive therapeutic intervention. The authors consider various routes off the ECMO “bridge,” and they conclude none are ethically supportable for a conscious patient, so ECMO becomes the destination (ECMO-DT).
ECMO is extremely resource intensive in technology, cost, and personnel, and takes a physical toll on the patient (including risks of sepsis, stroke, and hemorrhage, among other burdens), an emotional toll on the family, and a toll in caregiver moral distress when the treatment is perceived to be of no benefit, or worse, to harm the patient.
If ECMO is considered, at least at the start, as a bridge, there are at least five potential paths forward associated with the bridge.
Futility/inappropriate Treatment
In a classic case of physiological medical futility, the intervention cannot, or is highly unlikely to restore or maintain the vital function at issue. Even in those situations of physiological futility when the patient or surrogates oppose the withholding or discontinuation of treatment, we lack consensus in the bioethics community on the ethical way forward. Some argue that interventions that are highly unlikely to work should still be provided to patients who themselves or their surrogates wish it because there might be a chance, however remote, that it could work, or it may be of such important symbolic value to the patient’s family, that even knowing it will not work, the family benefits knowing that “everything” has been done.
Patients supported by ECMO who are not candidates for further interventions, such as transplantation or ventricular assist devices, are essentially stranded, like Barney Clark, the first permanent artificial heart recipient in 1983, who was tethered to his hospital machine until he died 112 days later. In that case, the technology was understood to be a therapeutic but ultimately time-limited destination. Kirsch uses teleological arguments to reject the characterization of ECMO in any way as a “destination therapy”.
Issues inherent in ECMO discontinuation, where the intervention is “working,” in maintaining circulation are not readily resolved using futility policies focused on whether the intervention will restore or maintain vital function. The Extracorporeal Life Support Organization (ELSO) uses a broader definition of futility to encompass no hope for healthy survival with the triad of “severe brain damage, no heart or lung recovery, and no hope of organ replacement by VAD or transplant.” Kon and coauthors go even further: “ICU interventions should generally be considered inappropriate when there is no reasonable expectation that the patient will improve sufficiently to survive outside the acute care setting.”, and Kon supports that such a withdrawal from a conscious patient with procedural due process can be “consistent with good medical practice, ethical norms and the law”.
During the pandemic, allocation algorithms for ventilators and other resources that privileged those who might as a result live more years than others (longer-term life expectancy) were subject to opposition as unjust discrimination of the disabled and elderly. More widely accepted final allocation standards often rejected longer-term life expectancy and settled on either some minimal years of life expectancy, or the mere ability to survive to discharge from the hospital as the acceptable discriminatory criterion.
Just as survivability to discharge was an acceptable basis for allocation of resources during crisis standards implementation, recognizing that current treatment cannot achieve the medical goal of returning the patient to a level of health that permits survival outside of the acute care hospital might be set as a new criterion of futility, beyond physiological futility, at least as it pertains to ECMO as a destination, even without invoking crisis standards. Childress et al. offer reasons that would not countenance withdrawal of life sustaining ECMO from a conscious refusing patient. But should consciousness be a factor at all?
Permanent Lack of Consciousness
For some ethicists, consciousness, and in this case, capacity to interact, doesn’t change the evaluation. Each life is equally valuable and worthy of support no matter its neurological state, short of death. Why deny ECMO to the incapacitated? Or make judgments based on subjective evaluations?
A permanent lack of consciousness, such as unresponsive wakefulness state (UWS) would currently preclude an individual from candidacy for the limited resource of life-saving organ transplant, though requests for ventilatory support, and perhaps dialysis or cardiopulmonary resuscitation might be heeded by some clinicians. The exclusion from candidacy for scarce vital organ transplantation results not from devaluation of any life but maximizing the likelihood and longevity of successful vital organ transplantation among many deserving recipients. This categorization of candidates for transplantation—with fatal outcomes for many who do not receive the organs—has occurred over the past four decades without major objection from the ethics community or the public.
A permanent lack of consciousness, for instance from severe brain injury, is a contraindication to beginning ECMO. UWS needs to be confirmed and carefully differentiated from minimally conscious state (MCS). If a patient who was conscious at the start of ECMO, but progressed to UWS, the contraindication to beginning treatment would now become the contraindication to continuing it, and that contraindication could counter a claim of the fiduciary obligation of non-abandonment.
Mulaikal and coauthors support this assertion: “Patients on ECMO who cannot participate in goals of care discussions because of severe and irreversible neurological dysfunction and with no chance of survival to discharge represent the most extreme form of medical futility.” Batten and coauthors support linking lack of perception of benefits with the inability to survive outside the acute care setting.
Time Limitations and Non-escalation
Another path off the bridge might be an upfront stated limit of time that any patient would be supported on ECMO, no matter the level of consciousness. Such a time limit might initially be agreeable to an individual or family. Even if there is no legal ability to enforce such a “Ulysses-type” agreement, some individuals who had agreed to this limitation might better absorb the implications of the patient’s fate on ECMO and might voluntarily adhere to their agreement.
Another approach might be to use “non-escalation”: Instead of withdrawing treatments begun, do not escalate with new interventions, such as inotropes, and do not replace technologies such as a failed circuit or oxygenator unit. This would be done ideally with the consent of the patient or surrogate, who might prefer not to face a “time certain” of death with discontinuation of ECMO, but accept that some ECMO-related or other complication would cause death at an unknown time.
Courtwright argues that when medical contraindications develop once ECMO has begun, nonescalation could be involuntary, since the patient with those complications wouldn’t have been placed on ECMO in the first place and does not have an unfettered right to ECMO maintenance.
Allocation On A First Come, First Served Basis
A counterargument to any limitation of ventilators, ICU beds or ECMO is that there are relatively few requests for these treatments, so any costs in technology, funds, or moral distress of caregivers should be absorbed. Perhaps, ultimately, ECMO-DT will become an alternative standard of care.
During the height of the COVID-19 pandemic, patients who had access to ECMO beds on a first come, first served basis had a 57% survival, but once beds were filled, patients who were turned away had a 90% mortality. After extensive professional consideration and public engagement about allocation of PPE, ventilators, vaccines, and medications under crisis standards, the irony is that this real shortage of staffed ECMO beds without a formal declaration of crisis standards was resolved on a first come, first served basis of allocation with fatal results.
ECMO beds are still a scarce resource, due less to lack of technology than to scarcity of trained personnel. There is a real danger of ECMO units being overwhelmed again. If there is no rationale to stop ECMO when there is no other destination, the default will be to accommodate all those individuals for whom there might be any likelihood of success in reaching a destination. Once ECMO beds are filled, those who follow who might have been much more likely to have benefited will be turned away.
During the pandemic, the possibility was contemplated of discontinuing support of one patient in favor of another who would benefit significantly more. A distributive justice rationale might allow such an allocation of ECMO. Wilkinson and coauthors argue that if the unjustified treatment harms others (who cannot justly access it), it is justified to withdraw treatment.
Childress and coauthors assert, as well might many clinicians, that such a discontinuation for a conscious patient would be intolerable. The hoped for deliberative public judgment to proactively consider these issues even post-pandemic, still lies beyond the horizon.
Allocation with Variable Admission Criteria Based On Surging Demand
During the pandemic many ECMO units developed crisis standards for allocation of scarce ECMO beds, overriding the usual obligation to accept all patients who might possibly benefit from ECMO. Under crisis standards, only those who would be most likely to benefit from ECMO would be accepted. Crisis standards in most jurisdictions were never invoked, so these new criteria were not tested.
Surgical intervention includes a weighing of the likelihood of the success of the intervention, the risks of the intervention, and (under some circumstances) the resources available to provide it. At some point, a very low likelihood of success that would burden surgical resources, including personnel, routinely results in surgeons determining that a patient is “not a surgical candidate.” If the likelihood of success is very low, the ORs are filled, and the patient is not transferable, that lack of availability of resources might be the tipping point to a specific determination. Decisions about initiation of a resuscitation attempt and length of resuscitative effort might also take into account these factors in assessing candidacy for resuscitation.
In a similar way, as more ECMO units are filling (but not yet filled), those who set the criteria for admission to ECMO units might recalibrate their standards to incorporate for subsequent ECMO candidates a significantly higher likelihood to successfully survive off ECMO through transplantation or device implantation. This would result in a “sliding scale” of eligibility based on availability of staffed ECMO beds with higher hurdles for those who come later. Hutchinson and coauthors offer a step-like scale for allocating ECMO beds during periods of scarcity that could serve this way, intentionally omitting an option to withdraw “un-bridgeable” patients.
Piscitello and coauthors argue that during invocation of crisis standards, “it is ethically permissible to broaden exclusionary criteria to also withhold ECMO from patients who have a low likelihood of recovery, to maximize the overall number of lives saved”. The same rationale might be used short of invocation of crisis standards as ECMO units move from normal status to a “surge” status. No governmental or administrative body would be required to declare such a status, rather some percentage of beds filled would trigger more stringent eligibility standards.
This would still privilege those who get to the resource first and heighten the barrier to those who follow but would more likely prevent a discontinuation of ECMO over objections. Unfortunately, it also undermines the moral equivalence of withholding and withdrawal. If ECMO can never be ethically withdrawn over the objection of patient or family, then the scarce resource of ECMO, as ECMO beds fill, becomes even more dear and a sliding scale calibrated to availability of ECMO beds might be able to preserve the functioning of the system for periods of strain. An upfront barrier of heightened exclusionary criteria might be an ethically supportable way forward. As tragic as it seems, that path might be another ethical way to address a looming shortage of ECMO beds. It is not a new path off ECMO, but a steeper path toward it.
Without choosing an alternative path in response to ECMO-DT patients, at some point short of crisis standards invocation, we will have chosen by default Path #4, first come, first served. The ECMO bridge will fill, blocking the way for others who might have been able to successfully use it to reach their destination beyond. That default path, many have recognized, is not the best one to take.