SPECIAL REPORT
Do Not Resuscitate (DNR) Orders During Surgery: Ethical
Foundations for Institutional Policies in the United States
Judith 0. Margolis, MD*, Brian J. McGrath, MD*, Peter S. Kussin, MDt, and
Debra A. Schwinn, MD*+
Departments of *Anesthesiology, tMedicine (Pulmonary), and $Pharmacology, Duke University Medical Center, Durham,
North Carolina
D
o Not Resuscitate (DNR) orders are established part of DNR orders. Most consider resuscitation, in the
by competent patients or appropriate surro- context of anesthesia, to be the administration of chest
gates to provide a mechanism for withholding compressions or cardioversion (7). Inhaled and intra-
specific resuscitative therapies in the event of a car- venous anesthetics can cause myocardial depression,
diopulmonary arrest. It is important for health care vasodilation, and cardiac dysrhythmias. Respiratory
institutions to develop policies to deal with DNR or- and circulatory depression may be produced or exac-
ders in the setting of anesthesia and surgery as it is erbated by opioids and sedatives. Local anesthesia
estimated that 15% of patients with DNR orders un- with conscious sedation may reduce but not eliminate
dergo a surgical procedure (1) and the frequency of the risk of cardiac or respiratory depression. The sur-
DNR orders is increasing as the population becomes gical procedure itself may lead to cardiovascular or
better informed about the Patient Self-Determination respiratory decompensation. These iatrogenic physio-
Act and advanced directives (2). Several surveys have logic abnormalities are superimposed on patients with
documented confusion on the part of individual prac- significant and often multiple organ system abnormal-
titioners and have demonstrated that only a small ities. Thus, the patient with DNR orders undergoing a
percentage of institutions have specific policies re- surgical procedure is likely to be at increased risk to
garding perioperative DNR orders (3-6). This lack of suffer cardiac or respiratory decompensation relative
consistency stems in part from the fact that surgery to the nonoperative patient. Conversely, patients sus-
and anesthesia routinely involve physiologic stresses taining a cardiac or respiratory arrest in the periop-
and suppression of vital signs far different from those erative period may also be more likely to suffer from
experienced outside the operating room. The unique transient insults (e.g., anesthetic-induced airway ob-
aspects of anesthesia create potential practical and struction), that are more responsive to resuscitative
ethical barriers to the implementation of perioperative efforts (8,9). A detailed discussion with patients, in-
DNR orders. In this paper, the complex medial and cluding prognostic information about cardiopulmo-
ethical issues surrounding the perioperative DNR or- nary resuscitation will help clarify their wishes (10).
ders are discussed, and the rationale for the recently This discussion should ideally occur at the time the
introduced American Society of Anesthesiologists DNR order is initiated, and include education about
(ASA) guidelines is highlighted. the ramifications of anesthesia.
Most of the surgery performed in patients with Ethical medical decision making involves the priori-
DNR orders is palliative and designed to improve tization of what are often conflicting ethical principles
patient comfort or simplify care. For many practitio- and practical considerations. When analyzing periop-
ners the definition of anesthesia includes routine sup- erative DNR orders, four ethical concepts are imme-
pression of respiration and performance of procedures diately important. The first principle is nonmalefi-
such as endotracheal intubation and volume or drug cence, epitomized by the Hippocratic dictum to “first
infusion, which in other settings are often refused as do no harm.” The second concept is beneficence, or to
perform a good, moral act for the patient. Third is the
This work was supported in part by National Institutes of Health principle of patient autonomy, which has emerged as
grant HL 02943 to D.A.S. a predominant principle in ethical medical decision
Accepted for publication November 30, 1994. making. The fourth principle is distributive justice
Address correspondence and reprint requests to Judith Margolis,
MD, Department of Anesthesiology, Box 3094, Duke University which allows society to provide medical resources to
Medical Center, Durham, NC 27710. those best able to benefit from them. Although justice
01995 by the International Anesthesia Research Society
806 Anesth Analg 1995;80:806-9 0003~2999/95/$5.00
ANESTH ANALG SPECIAL REPORT MARGOLIS ET AL. 807
1995:80:806-9 DO NOT RESUSCITATE ORDERS
has not traditionally been invoked in our society when B. Informed Suspension of the DNR
making individual patient decisions, it seems clear Orders During the Perioperative Period
that this will change as health care policy changes.
These basic principles may be competing or comple- The principle difference under this policy is that pa-
mentary in the various situations and institutional tients with DNR orders must be informed of periop-
solutions to perioperative DNR orders. We will now erative changes to their resuscitation status. Steps are
evaluate four possible hospital policies for the man- clearly taken to inform the patient (or surrogate) of the
agement of perioperative DNR orders in light of these risks of intraoperative cardiopulmonary arrest and
ethical principles. that in the event of an arrest full resuscitative efforts
would be used. Such policy risks being interpreted (by
the public) as a defensive institutional posture (12). A
positive point of this approach is that patient auton-
A. Automatic Suspension of DNR omy is better protected since patients have the option
Orders in the Perioperative Period of refusing a proposed procedure and seeking care at
an institution with policies more aligned with their
Automatic suspension of DNR orders during a surgi-
viewpoint(s). However, if surgery cannot be per-
cal procedure and for an arbitrary period postopera-
formed elsewhere, the patient may be denied care
tively is the most straightforward policy to invoke
unless he or she conforms to hospital policy. From
within a hospital. Since all patients and situations are
a beneficence standpoint, this policy is superior to
treated alike, the institution avoids the potential of that of not informing patients as it necessitates a
wrongful death law suits. During the perioperative discussion.
period it may be impossible to distinguish between a
cardiac arrest resulting from administration of a med-
ication, performance of an invasive procedure, or from C. Continuing DNR Orders Except for
natural progression of a patient’s primary disease. Airway Intervention
Nonmaleficence can be invoked to support that DNR
is incompatible with surgery and anesthesia. Some Some institutions honor perioperative DNR orders
practitioners consider every arrest occurring during with the specific exception of intubation and mechan-
ical ventilation. The rationale for this approach is that
anesthesia to be potentially reversible. However, re-
airway interventions constitute an intrinsic part of
cent data from the Study to Understand Prognoses
anesthesia and cannot be withheld when providing
and Preferences for Outcomes and Risks of Treatment
anesthetic care (7). Positive aspects of this approach
CXJPPORT) suggests that patients with DNR orders include freedom to perform anesthesia per routine by
who have a cardiac arrest in the operating room will the anesthesiologist and initiation of discussions with
not leave the hospital even if resuscitated (11). and education of the patient regarding differences
Strengths of automatic suspension of DNR orders between respiratory and cardiac arrest. Negative as-
during the perioperative period (with or without pa- pects of this policy include compromise of patient
tient permission) include gaining time to determine autonomy by excluding airway interventions from
actual cause(s) of arrest, and allowing the patient the any perioperative limits on resuscitation. For some
possibility of returning to baseline physiologic func- patients, placement of an endotracheal tube and pos-
tion after surgery. This approach also protects physi- sible reliance on mechanical ventilation in the postop-
cians who believe that willingly allowing a patient to erative period is the component of resuscitation most
die during surgery is a breach of the principle to do no feared; many patients are explicit in this regard in
harm. However, honoring the principle of patient au- advance directives. It is also possible that patients will
tonomy is in conflict with a policy of automatic sus- misunderstand the policy (especially in the absence of
pension of perioperative DNR orders. This policy ef- appropriate education), leading to potential conflict
fectively removes the patient from a decision-making between patient and clinician.
role, even if he or she is willing to accept the risk of
operative mortality. In fact, at least one hospital has
D. Continuing all DNR Orders During
been sued for negligence and battery related to the
performance of cardiopulmonary resuscitation on a
the Perioperative Period
patient with DNR orders (5). Another negative aspect Some institutions allow DNR orders to remain in ef-
of this policy is that withdrawal of care (a separate fect in the perioperative period. This policy is based on
issue from DNR) needs to be considered in some informed discussions with the patient (or surrogate)
patients if a persistent vegetative state results from which includes information regarding the risk of in-
resuscitation. traoperative cardiac or respiratory arrest tailored to
808 SPECIAL REPORT MARGOLIS ET AL. ANESTH ANALG
DO NOT RESUSCITATE ORDERS 1995;80:806-9
the contemplated anesthetic technique. Patient prefer- the informed Jehovah’s Witness to refuse transfusion
ences are determined regarding general anesthesia in and risk perioperative death. The DNR patient, like
the event that planned sedation or regional anesthetic the Jehovah’s Witness, has refused specific interven-
fails, with time limitations discussed for intubation tions. In refusing treatment neither one directly in-
and mechanical ventilation. Any therapies may be tends to die in the operating room but accepts death
withheld including chest compressions, electrical should it occur (12). Regarding the principle of dis-
countershock, vasoactive drug therapy, endotracheal tributive justice, a resuscitation policy that allows pa-
intubation, and mechanical ventilation. However, dur- tients with DNR orders to refuse resuscitation in the
ing perioperative discussions it must be made clear to perioperative period would be expected to conserve
patients that a planned sedation or regional anesthetic operating room and intensive care resources in a pop-
may fail, necessitating general anesthesia to complete ulation with a limited likelihood to benefit from these
the surgical procedure. The decision to refuse intraop- procedures.
erative resuscitation does not necessarily exclude
maximal therapeutic efforts for complications short of Recommendations
complete cardiac or respiratory arrest and could
conceivably allow for therapies such as opioid recep- We have two recommendations. The first is that each
tor antagonists, fluid therapy, pressors, or mask department of anesthesiology should have a policy
ventilation. regarding DNR orders in the perioperative period.
This policy upholds patient autonomy to the great- The second recommendation is that, of the choices we
est degree, with the freedom from unwanted resusci- have presented, we favor continuing all DNR orders
tative procedures. Precedence is found in the care of during the perioperative period after discussing this
terminally ill patients with DNR status on hospital with the patient (Option D, above). Although some
wards who receive parenteral pain medications and believe that conflicts between the ethical, medical, and
sedatives in large doses to achieve adequate analgesia practical considerations cannot be resolved, recent de-
regardless of the physiologic side effects. Improved velopments have shown that uniform policies are fea-
communication between clinicians and their patients sible. In October 1993 the ASA formulated their “Eth-
results. Preoperative discussions must include educa- ical Guidelines for the Anesthesia Care of Patients and
tion regarding the anesthetic state and establishment Do Not Resuscitate Orders or Other Directives that
of an understanding of the patients’ wishes within this Limit Treatment” (14). This policy acknowledges that
context. Patients with DNR orders may alter those automatically suspending DNR orders or other direc-
orders during the perioperative period once the tives that limit treatment prior to procedures involv-
unique aspects of anesthesia and surgery are ex- ing anesthetic care may not sufficiently address a pa-
plained clearly (9,13). tient’s rights to self-determination in a responsible and
Weaknesses of continuing DNR orders into the ethical manner. The ASA guidelines respect an in-
perioperative period lie predominantly within the formed suspension of DNR orders during the periop-
confines of the health practitioner and hospital. As erative period if explicitly discussed with the patient
mentioned above, some anesthesiologists think that or surrogate. The policy allows airway management
withholding resuscitation in the volatile physiologic and other treatment options, or will honor periopera-
setting of anesthetic care constitutes a violation of the tive DNR orders if the patient so states. Patient auton-
principle of nonmaleficence. Under these circum- omy is upheld to the greatest degree with this ap-
stances, the practitioner should have the option of proach. Professional integrity is maintained such that
transferring care of a patient to another anesthesiolo- in cases of moral conflict “the anesthesiologist should
gist. From an institutional perspective, this policy withdraw in a nonjudgmental fashion, providing an
could result in litigation, particularly if preoperative alternative for care in a timely fashion.” Distributive
discussions did not take place, were inadequately doc- justice is served in that an open discussion of options,
umented, or if families did not fully understand the resources, and outcomes should ensue with the pa-
patient’s wishes. Hence, preoperative discussions tient and family or proxy. The American College of
(which should include anesthesia and surgery attend- Surgeons has recently adopted similar guidelines (14).
ing doctors, as well as the primary care physician These statements provide important groundwork
when possible) are crucial in clarification of existing from which each hospital can develop policies to ad-
DNR orders. dress the issue of perioperative DNR orders.
Policies may stipulate the suspension of DNR or-
ders in emergencies when informed consent is impos- References
sible. A comparison can be made between the DNR
1. La Puma J, Silverstein MD, Stocking CB, et al. Life-sustaining
patient and the Jehovah’s Witness patient undergoing treatment: a prospective study of patients with DNR orders in a
surgery. Institutions have adopted policies that allow teaching hospital. Arch Intern Med 1988;148:2193-8.
ANESTH ANALG SPECIAL REPORT MARGOLIS ET AL. 809
1995;80:806-9 DO NOT RESUSCITATE ORDERS
2. Omnibus Budget Reconciliation Act of 1990. Title IV, Section 9. Cohen CB, Cohen PJ. Do-not-resuscitate orders in the operating
4206. Congressional Record, Oct. 26 1990;136:H12456-7. room. N Engl J Med 1991;325:1879-82.
3. Margolis JO, McGrath BJ, Kussin ES, Schwinn DA. Perioperative 10. Murphy DJ, Burrows D, Santilli S, et al. The influence of the
do not resuscitate (DNR) orders: a survey of major institutions. probability of survival on patients’ preferences regarding car-
Anesthesiology 1994;81:A1311. diopulmonary resuscitation. N Engl J Med 1994;330:545-9.
4. Clemency MV, Thompson NJ. “Do not resuscitate” (DNR) or- 11. Wenger NS, Greengold NL, Oye RK, et al. Do not resuscitate
ders and the anesthesiologist: a survey. Anesth Analg 1993;76: orders in the operating room. JAMA 1995. In press.
394-401. 12. Walker RM. DNR in the OR: resuscitation as an operative risk.
5. Troug RD. “Do-not-resuscitate” orders during anesthesia and JAMA 1991;266:2407-12.
surgery. Anesthesiology 1991;74:606-8.
13. Keffer MJ, Keffer HL. Do-not-resuscitate in the operating room:
6. Franklin CM, Rothenberg DM. Do-not-resuscitate orders in the
moral obligations of anesthesiologists. Anesth Analg 1992;74:
presurgical patient. J Clin Anesth 1992;4:181-4.
901-5.
7. Truog RD. What does “resuscitate” mean in a do-not-resuscitate
(DNR) order? Anesth Analg (letter) 1993;76:206. 14. Fine PG. DNR in the OR-anesthesiologists, medical ethics and
8. Peatfield RC, Sillett RW, Taylor D, McNicol MW. Survival after guidelines. ASA Newsletter 1994;58:10-4.
cardiac arrest in hospital. Lancet 1977;1:1223-5.