With medicine’s increasing capabilities, situations arise in which possible interventions will not serve the patient’s or the family’s goals of care. Under such circumstances, the decision to forgo treatment, even potentially lifesaving treatment, may be appropriate. This chapter discusses one means of limiting treatment: do-not-attempt-resuscitation (DNAR) orders. It describes the historical development of DNAR orders, outlines the process of writing such orders, and explains why they should generally not be written unilaterally. The related issues of family presence during resuscitation and DNAR orders in the operating room and outside the hospital are also reviewed.

Do-not-resuscitate (DNR) orders developed out of the recognition that cardiopulmonary resuscitation (CPR) lacks efficacy in certain patient populations and that a formal process of advance planning was needed. Although modern CPR was initially developed for patients suffering anesthesia-induced cardiac arrest, it became the standard of care for cardiac arrest in hospitalized patients regardless of their underlying diagnoses. Experience, however, demonstrated that the effects of CPR were often transient. In some institutions, covert decision-making processes evolved to withhold or limit resuscitation efforts. Hospitals developed DNR policies in the 1970s to address the need for both a decision-making process and a means to communicate these decisions.¹

There are limited data in the pediatric literature regarding the efficacy of CPR in hospitalized patients. In reviewing the literature it is important to focus on patient-centered outcomes. A retrospective review of data from the National Registry of Cardiopulmonary Resuscitation (NRCPR) found that 52% of children who experienced an in-hospital cardiac arrest resuscitation (pulseless cardiac arrest requiring chest compressions, defibrillation, or both, that elicited an emergency resuscitation response and resulted in a resuscitation record) had return of spontaneous circulation for greater than 20 minutes and 27% survived to hospital discharge. Of the children who survived to discharge, 65% had a good neurological outcome defined as a Pediatric Cerebral Performance Category of normal functioning, mild disability, or moderate disability.²

A retrospective review of in-hospital cardiac arrests in 15 children’s hospitals within the Pediatric Emergency Care Applied Research Network (PECARN) found that 48.7% of patients between 1 day and 18 years of age who received greater than 1 minute of chest compressions and had return of circulation for at least 20 consecutive minutes survived to hospital discharge.³ This compares to 51% in the NRCPR study.² Using different criteria, the PECARN investigators found a higher rate of good neurological outcomes; among survivors who had prearrest and discharge Pediatric Cerebral Performance Category scores available, 94.3% had discharge scores of normal or mild disability or no change in score.³

Information regarding the efficacy of CPR within specific diagnostic categories is even more limited. The PECARN study found pre-existing hematologic, oncologic, or immunologic disorders and pre-existing genetic or metabolic disorders were associated with increased hospital mortality.³

Based on the limited efficacy of CPR, some authors prefer the term do-not-attempt-resuscitation (DNAR).⁴ Reinforcing that many resuscitation attempts are not successful is important because families may have false high expectations of the efficacy of CPR based on television’s depiction of CPR.^5,6 More recently, the term allow natural death (AND) has been proposed. Proponents argue that DNR sounds cold and cruel, and that AND is warmer and more comforting.⁷ A single study reports that nursing students and other college students, but not nurses, are statistically more likely to endorse an AND order than a DNR order in response to a near-death scenario.⁸ Critics of this proposed change argue that unlike DNAR, AND does not convey which specific procedures will be withheld. For example, it is not clear whether an AND order precludes therapeutic treatments such as vasopressors and antibiotics.^9,10

Even if CPR might be effective, there may be situations in which it is ethical to withhold it. For example, in terminally ill patients, it may only prolong the dying process. More controversially, the patient’s quality of life may be so diminished that it need not be maintained.

The Patient Self-Determination Act, which became effective in 1991, requires most health care institutions to ask adult patients on admission whether they have advance directives; however, there are few mechanisms for parents or guardians to express their wishes for their children or for adolescents to express their wishes should they lose medical decision-making capacity.^11,12 Hospitalists should review the goals of treatment—curative, uncertain, or primarily comfort—with the families of children with special healthcare needs (Table 183-1). The patient’s primary care provider may be a valuable resource, especially if he or she has a longstanding or close relationship with the patient and family and has discussed treatment goals with them. Discussing the goals of treatment is particularly important when significant changes occur in the child’s clinical condition. If there is a possibility of cardiorespiratory arrest, especially if there is a low likelihood of survival or survival would not be in the child’s best interest, the hospitalist should discuss the risks and benefits of CPR, and a DNR order should be considered. This difficult topic can be broached by asking, “What is your hope for your child?”

DNR orders generally prohibit CPR regardless of the cause of the arrest. If the DNR order is based on a low likelihood of survival and if different causes of arrest have different survival rates, this should be discussed.¹³ For example, a family might desire intubation and mechanical ventilation for respiratory depression that is a side effect of anticonvulsants.

Hospitalists should be cautious about permitting families to pick and choose components of CPR. For example, administering medications while withholding chest compressions lacks a pathophysiologic rationale; the medications cannot be effective if they are not circulated. It may, however, be reasonable to specify a limited duration of resuscitation efforts.