|Home | About | Journals | Submit | Contact Us | Français|
Most North American hospitals require that cardiopulmonary resuscitation (CPR) be attempted unless a do not attempt resuscitation (DNAR) order has been completed with the consent of the patient or the patient's surrogate. For patients with advanced, progressive cancer, there can be tension in DNAR discussions between patients and families who desire “everything” to be done and the evidence-based opinion of their health professionals that CPR would be futile. Here we consider the relationship between health professionals' obligations to prevent harm from futile interventions and to respect their patients' autonomy.
A 39-year-old woman has recurrent squamous cell cervical cancer metastatic to the pelvis, periaortic, and mediastinal nodes, and directly invasive of the sternum. Despite multiple conventional and phase II chemotherapy approaches, her disease shows rapid progression on the most recent computed tomography scan, now replacing one quarter of the sternum, with new rib and liver metastases. She is admitted with severe pain and constipation, and is presently on fentanyl patches and oral long-acting and intermediate morphine. Her pain level is a constant 8 to 9 out of 10. She is in the fetal position and moaning. The attending physician has been reluctant to discuss the possibility of a DNAR order, as neither the patient nor her husband have been receptive to her attempts to discuss other end-of-life care issues. The pain service attending wants the patient to reach a decision about resuscitation before inserting an epidural catheter. The house staff also wants to expedite a resolution. They are concerned that if resuscitation were attempted, the patient's sternum and rib lesions would fracture and cause even more pain. They want to talk to the patient even if the attending will not. The patient's husband begs them not to approach the topic of resuscitation with the patient because “if you tell her she is going to die, then she will die.”
CPR is the default response in most hospitals unless orders exist to forego CPR. However, policies about resuscitation can vary among institutions and can be interpreted differently. Some focus solely on full cardiopulmonary collapse; others call out decisions about vasoactive drugs, intubation, mechanical ventilation, antibiotics, and fluid resuscitation; others specifically include withholding all of these interventions. Efforts to clarify the scope of hospital and governmental resuscitation policies improve and focus decisions about end-of-life care. For example, the physician's orders for life-sustaining treatment form used in Oregon for both inpatients and for emergency medical technicians responding to outside calls separates decisions about resuscitation for complete cardiopulmonary arrest from a menu of choices for management of other acute events. Patients who opt for DNAR orders may feel reassured that a broad array of palliative care and pain management options is available to them.1,2
The efficacy of resuscitation in hospital for patients with cancer is difficult to assess, given the variety of definitions for what constitutes resuscitation and the heterogeneity of the patients. However, a recent meta-analysis suggests that the overall survival to discharge after CPR for patients with metastatic cancer is 5.6% (ie, a 94.3% death rate) against a background of 15% for unselected in-hospital arrests.3 Quality of life at discharge was not assessed in this meta-analysis. Furthermore, it overestimates outcomes as a variety of procedures were included, and a broader selection of patients than those with end-stage cancer were included.
Beneficence is the ethical principle that holds health professionals responsible for assuring that patients receive potential benefit from the health care delivered.4 Physicians are also ethically obligated to prevent harm to patients, and are not ethically obligated to provide interventions that create more harm than benefit. The only goal of medicine to be achieved by resuscitation of patients with progressive cancer is prevention of untimely death. This raises the question of whether a patient's death would truly be “untimely,” and what this means when viewed in light of the potential for increased harm (nonmaleficence) to survivors from CPR. In the case described in the vignette, fractures are likely to occur if the patient survives CPR and would substantially add to her pain.
CPR, when thought to be futile, is best explored from both quantitative (at least 94.3% death before discharge) and qualitative (increase in pain and suffering) perspectives.5 Given both components of futility, CPR would seem to have little to offer most patients at the end of life, including the patient described previously.
Patients may feel determined to pursue CPR because unlike health professionals, they overestimate the success rate of CPR and the quality of life experienced by survivors after CPR.6 They may believe that resuscitation will return a patient to his or her present state of health, and allow for more time for interaction with friends and family. Often patients protect families from the reality of their impending death, and families are concerned about preserving the time they might have with the patient. Everyone involved is reluctant to engage in an honest discussion of the medical nature of resuscitation and the potential for benefit or harm. It is important to note that while a spouse or other family member may not want such issues discussed with the patient, the physician's obligation is to the patient first.
CPR discussions, however, are not the same as discussions about therapeutic interventions. CPR discussions are intended to prevent a futile and potentially harmful intervention, not to initiate an intervention intended to provide benefit. Physicians can reframe discussions about resuscitation by providing patients with information that conveys the importance of avoiding harmful interventions, and helps the patient understand this basis for the recommendation that resuscitation should not be attempted in their case. This approach is more true to physicians' ethical obligations, and empowers patients to understand their prognosis and make appropriate choices about end-of-life care. This approach also supports patient autonomy for decisions about other limited-benefit supportive care by addressing resuscitation separately from other interventions. As noted above, policies like physician's orders for life-sustaining treatment allow physicians to facilitate the prevention of futile resuscitation in a manner that is distinct from management options for other acute events. Clinicians may discuss providing fluids, blood transfusions, or antibiotics as well as relieving specific symptoms such as pain or nausea with analgesics or antibiotics. Chemotherapy, radiotherapy, or surgery with palliative intent would not be precluded by a DNAR order which identified resuscitation as futile in this setting.7
In the context of this vignette, the initial discussion with the patient by the attending oncologist should revolve around what can be done to control her major symptoms and pain, as well as the implications of her disease progression through multiple treatments and the futility of further anticancer treatment.
DNAR orders are best discussed by the clinician who has an ongoing relationship with the patient as part of shifting the goal from treatments aimed at cancer cure to those which relieve symptoms. Understanding the perception of the patient about the meaning of this discussion and assuring her that she will still have the support of her oncologist during end-of-life care is critical.7 With the patient's permission, it would be beneficial to include her husband in this discussion. Framed in an appropriate context, the oncologist should find it easier to discuss the futility and burden of resuscitation as a part of her ongoing obligation to benefit patients and protect them from harm.
The authors greatly acknowledge the ASCO Ethics Committee for their contributions to this article.