Key words: Ethics: medical. Resuscitation: orders.
In the past 5 yr, many articles [1-9],* have been published on the topic of surgical patients who have do-not-resuscitate (DNR) orders. Some recommend negotiated reconsideration with the patient (or surrogate** before anesthesia. [2-7]The argument for this recommendation is based on the principle of respect for patient autonomy. Others recommend automatic suspension of DNR orders in the perioperative period. [8,9],*
Three arguments are given to support suspension of DNR orders during surgery. First, consent for surgery and anesthesia implies consent at least for those resuscitative measures intrinsic to the administration of an anesthetic and is, therefore, inconsistent with a DNR order. Second, results of resuscitation in the operating room (OR) are better than resuscitation outside the OR ; therefore, the usual considerations regarding DNR orders should not automatically apply to the patient having a surgical procedure. Finally, anesthetics cause cardiovascular and respiratory instability. Deprived of the ability to reverse these physiologic changes, the anesthesiologist may favor less anesthesia and greater physiologic stability, thereby depriving the patient of the full benefits of anesthesia, especially sufficient analgesia and amnesia. 
Part, but not all, of the disagreement is related to interpretation of the term "resuscitation." According to some, resuscitation during anesthesia refers only to closed chest cardiac compression and electrical countershock. To others, it means any procedure used in resuscitation, including tracheal intubation, assisted ventilation, or the use of vasoactive drugs, procedures that are also part of routine anesthetics.
The American Society of Anesthesiologists (ASA) has issued ethical guidelines (Figure 1) for the anesthesia care of DNR patients that recommend negotiated reconsideration of the DNR order and accession to the patient's wishes. The guidelines appear to use the broader definition of resuscitation.
This article describes one institution's policy on DNR orders in the OR, and critiques the ASA guidelines.
Policy on Do-not-resuscitate Orders in the Operating Room at Scott & White Hospital and Clinic
At Scott & White Clinic and Memorial Hospital in Temple, Texas, our policy is a compromise between the extremes of automatic suspension of DNR orders and complete autonomy of the patient. Consequently, it is subject to criticism from both sides.
Our policy requires discussion of each patient's DNR status between an anesthesiologist and the patient. This discussion occurs while obtaining informed consent for anesthesia. The anesthesiologist offers the patient two choices described on a clarification form (Figure 2). One choice is to maintain the DNR order, with the understanding that "clinical events thought to be temporary and readily reversible will be treated and that standard anesthesia procedures will be used." This intentional vagueness allows the anesthesiologist to practice unhindered by restrictions more appropriate for non-OR settings, where life-supporting procedures are not commonplace. It may be reasonable for a patient outside of the surgical suite to refuse tracheal intubation, mechanical ventilation, oxygen therapy, or vasopressors; refusing these interventions in the OR may not be reasonable, particularly when they are integral parts of the anesthetic management. Twenty senior staff anesthesiologists were unable to agree completely on what restrictions would be acceptable.
The other choice is to suspend the DNR orders during the time the patient is in the OR and postanesthesia care unit. This is the default choice when surgery is considered emergent enough to proceed in the case of an incompetent patient whose surrogate is not available.
The process of completing the clarification form is intended to ensure that the patient and anesthesiologist discuss the patient's overall goals regarding the surgical procedure rather than the use of specific interventions. A patient may refuse tracheal intubation when the real goal is not to be maintained on mechanical ventilation for a prolonged period. Whereas some authors recommend discussion of the risks, complications, and alternatives for all possible therapeutic modalities and intraoperative interventions, we chose to focus on clarification of patients' goals. We believe this goals-oriented approach appropriately places responsibility with the physician to counsel patients regarding their condition and prognosis and to select techniques and therapies appropriate to attain the agreed-on goals.
Critique of the ASA's Ethical Guidelines for the Anesthesia Care of Patients with Do-not-resuscitate Orders or Other Directives that Limit Care
I have two major criticisms of the ASA guidelines for the care of DNR patients in the OR: they are not feasible in our practice and they appear to make respect for autonomy an absolute, rather than a prima facie, principle. These arguments are based on my interpretation of the ASA guidelines and the discussion of the guidelines by Fine and Jackson.***
Beauchamp and Childress state that the implementation of moral principles must take into consideration feasibility, efficiency, and cultural pluralism. In our practice, the physician who obtains consent for anesthesia is often not involved with the intra- and postoperative anesthetic management. Other duties of the anesthesiologists, schedule changes, and cases not finishing as scheduled make it impossible to guarantee the continuity of care necessary to completely implement the ASA guidelines. Knowledge of other practices leads me to believe our logistical problems are not unique.
Our anesthesiologists' attitudes range from those who would not treat hypotension or dysrhythmias caused by induction of anesthesia, if that were the patient's wish, to those who would not administer an anesthetic without being able to treat easily correctable iatrogenic problems. All agree with the general concept of DNR orders. The disagreement is in the interpretation of what constitutes resuscitation and exactly where the line is drawn in a given clinical situation. Resuscitation may refer only to closed chest cardiac compression and electrical countershock, or it may include treatment of problems leading up to a cardiac arrest and even the treatment of drug-induced respiratory arrest. Without the compromise, we could never have reached an agreement that would allow one person to negotiate with a patient and make a commitment for one or more associates involved in the intra- and postoperative anesthesia care.
Ideally, the negotiation with the patient would involve the entire operating team; once again, an ideal that is not feasible. For logistical reasons, it was agreed that the clarification form would be completed by an anesthesiologist while obtaining consent for anesthesia. This seemed logical because we are often the cause of intraoperative "clinical events," and we are always the primary resuscitators.
Physicians are moral agents with autonomy. The ASA guidelines recognize this and allow for physicians to withdraw from unacceptable situations. During nights and weekends, our anesthesiologist on call is expected to provide anesthesia care for all presenting patients. Each anesthesiologist has a unique combination of strengths and weaknesses, preferences and dislikes, and levels of comfort for given situations. By focusing our discussions on goals, rather than specific interventions, we have kept such unacceptable situations to a minimum.
Therefore, we have adopted a policy that is a modification of the ASA guidelines (section X of the ASA guidelines), a compromise that is feasible and superior to our previous policy of automatic suspension of DNR orders in the OR.
In their discussion of the ASA guidelines, Fine and Jackson state: "The premier and fundamental principle of bioethics is that of respect for the autonomous choice of the patient" (p. 46). Although they mention nonmaleficence and beneficence and admit that anesthesiologists are also moral agents, they appear to consider respect for autonomy as an absolute, rather than a prima facie, principle. Prima facie principles are obligations to be followed unless they are in conflict with an equal or stronger obligation. An absolute principle-there can be only one-takes precedence over any other principle or rule when principles or rules conflict. Beauchamp and Childress, who developed the "four principle" theory of bioethics (autonomy, nonmaleficence, beneficence, and justice), reject not only absolute principles, but also any hierarchical or lexical ordering of the principles.
Several problems exist with making respect for patient autonomy "the premier and fundamental principle of bioethics," particularly in the practice of anesthesiology, and Fine and Jackson allude to some of them.
First, autonomy (self rule) is limited by our relationships with others; autonomy is never absolute.
Second, the patient is not the only moral agent involved. In addition to the patient are the members of the anesthesiology team, the surgical team, and postanesthesia care unit personnel. All moral agents have duties and rights. The perioperative period is not a situation where the patient is the only agent with rights. Whereas only patients can determine their ultimate goals, they must often trust their caregivers to help them achieve those goals.
Third, patients in the OR are rarely autonomous; usually, they are under the influence of drugs that render them incapable of acting intentionally and with understanding. The anesthesiologist, then, becomes the patient's surrogate to determine what action is in the patient's best interest. It is impossible to anticipate and discuss every possible situation that may occur during anesthesia; consequently, the patient must rely on the anesthesiologist to act on his or her behalf. The patient can set general goals (e.g., not to be maintained for a prolonged period on a ventilator or not to be kept alive in a persistent vegetative state) without demanding to have surgery for a bowel obstruction without the use of an endotracheal tube. When a situation arises that has not been discussed, the physician can fall back on the principle of proportionality, in which the benefits and burdens of a particular action are weighed. When an action carries great benefit and little risk of burden, the physician is obligated to take that action in the nonautonomous patient.
Fourth, the ASA uses a diversity of cultures as a reason for the guidelines (section I), yet there is evidence that the emphasis on respect for autonomy is a product of Western civilization and that discussions of prognosis and complications may actually be harmful to patients from other cultures. [14,15]
Fifth, strict enforcement of a DNR policy in the OR may lead to undertreatment of the patient. A patient who refuses vasoactive drugs may not receive adequate anesthesia, for fear that the predictable fall in blood pressure, if untreated, may cause harm. The patient who refuses tracheal intubation may, likewise, receive an inadequate anesthetic.
Finally, the rationale for the ASA guidelines appears to ignore the principles of nonmaleficence and beneficence. In addition, there is no consideration of other approaches to bioethics, such as virtue-based ethics, casuistry, and the ethics of care. The guidelines also appear not to emphasize the fiduciary nature of the doctor-patient relationship or the basic goals of medicine. Those goals, elaborated by Jonsen, Siegler, and Winslade, are :
- promotion of health and prevention of disease
- relief of symptoms, pain, and suffering
- cure of disease
- prevention of untimely death
- improvement of functional status or maintenance of compromised status
- education and counseling of patients regarding their condition and its prognosis, and
- avoiding harm to the patient in the course of care
In the OR, when any of the four principles of Beauchamp and Childress or any of the goals of medicine conflict, the burden may be placed on the physician to resolve the conflicts. How this is accomplished will largely depend on the virtues of the individual physician involved and the specific circumstances of each case, including the patient's declared goals. I fully support respect for patient autonomy as a prima facie principle as elaborated by Beauchamp and Childress. However, relying solely on respect for patient autonomy and ignoring other principles and approaches seems detrimental to the purpose of the guidelines, which is to improve patient care.
The author thanks Professor James F. Childress, Dr. Robert D. Truog, and Dr. T. Keller Matthews for helpful suggestions, and Janice E. Hudgins for assistance with manuscript preparation.
*Kang WB: Putting people to sleep: Do not resuscitate orders in the operating room. JD thesis, 1992 (personal communications).
**In this paper, whenever "patient" is referred to, that term will include "surrogate" if the patient is not competent or autonomous.
***Fine PG, Jackson SH: Do not resuscitate in the operating room: More than rights and wrongs. Am J Anesthesiol 1995; 22:46-51.