Beyond the Do-not-resuscitate Order: An Expanded Approach to Decision-making Regarding Cardiopulmonary Resuscitation in Older Surgical Patients

Default options are events or conditions that will be set in place if an alternative is not actively chosen.²  Because they often do not involve an explicit decision-making process, default options exert unique power over behavior in a variety of contexts, and have been the target of ethical scrutiny in healthcare settings.^2–4  Appropriateness of default options in health care depends on both preserving patient autonomy by allowing individuals to deviate from the default without significant barriers and satisfying the best-interest standard, i.e., promoting the reasonably assumed best interest of the patients affected (since the default is affecting care without preferences being elicited).^2–4,12  Use of defaults is discouraged “when no single decision is clearly optimal for a majority of people.”²  Ensuring that patients can choose an alternative is particularly important when “patients’ health states and preferences may influence the propriety of the default.”⁴  While certain defaults are unavoidable (e.g., options for treatment must be presented in some order, and this ordering can be changed but not abolished), others represent assumptions that can be abandoned or clarified via an explicit decision-making process (e.g., re-establishing code status on admission to the hospital instead of assuming patients are “full code” by default).

For example, in analyzing the once widely employed practice of automatically reversing DNR orders before surgery, Truog outlined the complexity of decisions about whether to perform CPR in patients with preexisting directives limiting treatment.¹³  He concluded that automatic reversal should be modified in favor of “the traditional medical practice of responding individually and compassionately to the unique needs of each patient.”¹³  When reconsidered using frameworks outlining ethical use of default options, automatic reversal failed to preserve freedom of choice (because patients were not regularly informed or consulted regarding reversal of their code status) and did not reliably satisfy the best-interest standard (due to its inflexibility and lack of nuance regarding individual patient circumstances and preferences).

ASA guidelines were revised accordingly in 1993, and since 2001 have included the “goal-directed approach” to reviewing and reconsidering perioperative DNR orders articulated by Truog et al. in 1999.¹⁴  These analyses and revisions reflect a trajectory of enhanced shared decision-making and patient-centered care in patients with DNR orders or other directives limiting treatment, but the practice of treating all other patients as “full code” by default raises other ethical concerns.^14–16  First, the current approach does not acknowledge freedom of choice in patients without DNR orders. The guidelines do not recommend informing patients that a decision about code status is even being made, let alone enabling them to deviate from the default (fig. 1). Second, there is mounting evidence that a growing number of patients are at risk for poor outcomes after CPR, and use of “full code” as a default option in this population may not consistently satisfy the best-interest standard.^6,7,17–21 

The American Medical Association (Chicago, Illinois) Code of Medical Ethics outlines the following as relevant considerations in applying the best-interest standard: pain and suffering associated with an intervention, degree of/potential for benefit, impairments that may result from the intervention, and expected quality of life.¹²  Resuscitation for perioperative cardiac arrest is generally believed to satisfy the best-interest standard because of its relatively high likelihood of success and the degree of benefit.^13,18  This reasoning is sound for most patients, but may be less applicable to the subset of elderly patients who face a less favorable trajectory not only after surgery and anesthesia but also after cardiac arrest and CPR.

A recent study of outcomes for nearly 7,000 perioperative cardiac arrests found an 11- and 32-fold increase in 30-day mortality for patients with ASA Physical Status IV and V, respectively (compared to ASA I and II patients).⁶  These findings are consistent with a previous investigation of perioperative CPR outcomes demonstrating survival to discharge of 17% for ASA IV and 8% for ASA V patients.⁷  That study also identified age and impaired functional status as important predictors of mortality. Survival to discharge in patients age greater than 85 yr was only 14%. Patients without functional impairment had a 25% rate of survival to discharge, compared to 15% in partially dependent patients and 11% in totally dependent patients.⁷ 

The risk/benefit ratio of perioperative CPR may be less favorable in high-risk elderly patients for several reasons. In older patients with significant comorbidities, cardiac arrest in the operating room is more likely to be related to the burden of underlying conditions, and thus might be less reversible compared to arrests caused by anesthetic or surgical complications.¹⁷  Such patients are also more likely to be clinically frail, a state of diminished physiologic reserves that predisposes to worse outcomes in a variety of settings.⁵  Although outcome data for frail patients undergoing perioperative CPR are lacking, frail patients undergoing CPR for in-hospital cardiac arrest have a dismal prognosis. Survival to discharge ranges from 0 to 4.8% in recent studies (compared to 26% and 31% in non-frail patients).^19–21  Quality of life and functional outcomes in those who survive are in need of further study.

Outcomes in frail patients after perioperative CPR may be better than these data suggest, but the consistently unfavorable trajectory for frail patients in other contexts is nonetheless revealing. It suggests that their course is shaped by factors beyond the clinical circumstances surrounding cardiac arrest, including vulnerability to complications of resuscitation itself. Significant injuries are common even in non-frail patients who undergo CPR (e.g., rib fractures, sternal fracture, pulmonary contusion, pneumothorax), and traumatic rib fractures carry a poor prognosis for elderly patients even outside the context of cardiac arrest.^22,23  Outcomes of mechanical ventilation in high-risk patients are also relevant, since respiratory complications and need for ventilator support are common in this setting.⁷  Patients with clinical frailty who require mechanical ventilation are more likely to experience in-hospital mortality, extubation failure, need for tracheostomy, and discharge to long-term care.²⁴ 

Because high-risk patients are less likely to survive and more likely to suffer complications and changes in quality of life after perioperative CPR, it is apparent that using “full code” as a default option does not reliably satisfy the best-interest standard. The available evidence does not necessarily justify viewing perioperative CPR as “inappropriate” in high-risk patients, but it does suggest that its appropriateness depends on patient-centered consideration of risks and benefits. Many high-risk patients have a DNR order or other directive limiting treatment in place, so clarifying code status is already standard practice in these cases. However, well-described deficits in advance care planning reveal that a large subset of high-risk patients proceed to surgery without having such discussions.²⁵ 

Limitations of code status documentation are also relevant. Although a patient may be “full code” in the medical record, this status may be “unconfirmed,” or “presumed.” In such cases, it is possible either that no discussion of code status has occurred, or that a patient’s documented preferences have not been integrated into their medical record. These possibilities are particularly concerning given litigation over the last decade around failure of medical practitioners to adhere to patients’ documented preferences to forgo life-sustaining therapies. Although courts have yet to rule in favor of plaintiffs’ claims of “wrongful life,” several lawsuits have resulted in out-of-court settlements and administrative penalties.²⁶  Review of these legal proceedings is beyond the scope of the current discussion, but such cases highlight the importance of proactive efforts to minimize the likelihood of delivering unwanted care.

Since using “full code” as a default option neither consistently satisfies the best-interest standard nor preserves freedom of choice in high-risk surgical patients, it should not be assumed that all elements of resuscitation are concordant with high-risk patients’ goals and preferences (fig. 2). When caring for patients age 75 yr or greater, anesthesiologists should be attuned to documentation or evidence of conditions that indicate vulnerability to complications and poor outcomes after CPR, including age 85 or greater, ASA Physical Status IV or greater, functional impairment, and frailty. Whenever possible, the presence of these conditions should trigger discussion with patients (or their surrogates in cases of decisional incapacity) to clarify preferences regarding perioperative resuscitation. More detailed, patient-centered outcome data may justify expansion of the proposed framework to include younger patients with serious illness in the future.

Ideally, the anesthesiologist’s discussion regarding perioperative resuscitation in high-risk patients would occur in collaboration with the surgeon/proceduralist and, if applicable and practicable, another physician with a close relationship to the patient (e.g., primary care physician, geriatrician, or other medical specialist). Use of the goal-directed approach (i.e., limitations on attempts at resuscitation with respect to specific procedures or patient’s goals) might be particularly useful in generating a plan tailored to high-risk patients’ specific concerns and vulnerabilities.^1,14,27  In many cases, the discussion may be brief (e.g., when confirming a status of “full code” that reflects a previous code status discussion that remains relevant). This scenario may become increasingly common with implementation of American College of Surgeons (Chicago, Illinois) standards for geriatric surgery, which recommend that surgeons confirm code status in all patients age 75 yr or greater and clarify preferences for other life-sustaining therapies in patients with planned admission to the intensive care unit (fig. 3).¹¹  Circumstances will sometimes prevent providers from clarifying code status in high-risk patients, particularly in the setting of emergency surgery where time constraints, patient incapacity, and unavailability of surrogates are often insurmountable barriers. Consistent with current practice, “full code” should remain the default in such situations.

This proposed approach has several advantages over current guidelines. First, it expands shared decision-making regarding perioperative CPR to include high-risk patients irrespective of existing code status, thereby promoting freedom of choice and goal-concordant care in a high-risk patient population.²  Second, it is compatible and synergistic with new geriatric surgery standards from the American College of Surgeons that will call on anesthesiologists and consultants to collaborate in tailoring perioperative care to elderly patients’ vulnerabilities and preferences (fig. 3).¹¹  Third, its implementation is realistically scalable, serving to practically enhance institutions’ adherence to new American College of Surgeons geriatric surgery standards and other screening recommendations including those from the American Geriatrics Society and the Society for Perioperative Assessment and Quality Improvement.^8–11 

Expanding decision-making to include high-risk patients irrespective of existing code status raises several issues in need of further discussion, including the question of how high-risk patients should be defined and identified. Research in this area is expanding, and refinement of criteria that define “high-risk” will be necessary based on emerging evidence and multidisciplinary collaboration. While ASA status and age are easily established, there are dozens of tools to measure frailty.^5,28  A recent systematic review and meta-analysis of 70 studies measuring 35 frailty instruments found that the Clinical Frailty Scale was most effective in predicting postoperative mortality and was the easiest to use.²⁸  Its accuracy in predicting outcomes after perioperative CPR is unknown, but the Clinical Frailty Scale score has been associated with increased mortality after resuscitation for in-hospital cardiac arrest and unfavorable outcomes after mechanical ventilation.^19,20,24  Based on its demonstrated practicability and association with related outcomes, the Clinical Frailty Scale is a logical choice for screening in this context.²⁸  However, consistent with societal recommendations, the proposed approach emphasizes the importance of screening without dictating use of a particular screening tool.^11,29  These initiatives face significant barriers, and in many cases it will continue to fall upon individual providers to identify high-risk patients via review of the medical record and preoperative evaluation.³⁰  Although it should not generally be considered a substitute for formal screening, clinical impression can be important in identifying high-risk patients when more detailed evaluations are not feasible due to limited time, resources, or infrastructure.³¹ 

The proposed framework is inherently interdisciplinary. Its implementation will therefore require buy-in from multiple stakeholders, including surgeons, proceduralists, and anesthesiologists, as well as geriatricians and palliative care specialists. Although surgeons’ opposition to limitations on aggressiveness of care is sometimes cited as a barrier to code status discussions, newly established American College of Surgeons standards reflect a heightened commitment to goal-concordant care and geriatric risk stratification.^10,11  Some stakeholders will likely object to the time commitment associated with engaging patients and families in discussions about perioperative resuscitation. This is an important concern that highlights the need for mechanisms to identify high-risk patients early so they can be referred to geriatricians and/or palliative care specialists for formal evaluation and high-quality discussions regarding goals and care preferences.^11,29,32  Referring all high-risk patients, however, would risk overwhelming already strained geriatrics and palliative care services and may not be an option in many settings.^33,34  It will therefore become increasingly important to develop an understanding of which high-risk patients will benefit most from geriatric or palliative care consultation, and to establish mechanisms by which surgeons and anesthesiologists can develop competency in discussing code status and performing elements of geriatric risk assessment.

The trajectory of perioperative medicine is one of increasing collaboration and integration of multidisciplinary expertise. Adoption of the proposed approach to decision-making regarding perioperative CPR is therefore not the task of the individual anesthesiologist. Rather, implementation must begin with discussion among perioperative physicians regarding limitations of existing guidelines when applied to high-risk patients. Such discussion may serve as a basis for revision of ASA guidelines, integration with ongoing efforts from the American College of Surgeons (fig. 3), and initiatives at the institutional level to facilitate identification and management of high-risk geriatric surgical populations.^1,8–11,29  New American College of Surgeons standards for geriatric surgery face similar barriers and present an opportunity for interdisciplinary collaboration to create sensible hospital policies and infrastructure.^8–10,29  Best practices regarding the content and documentation of code status discussions and developing/amending hospital policies regarding approach to DNR orders in the perioperative setting have been well-described elsewhere, and these remain broadly applicable.^1,14–16 

Because there is a growing, identifiable subset of surgical patients who are less likely to survive and more likely to suffer complications and changes in quality of life after perioperative CPR, it is no longer appropriate to assume as a default that all elements of resuscitation are concordant with high-risk patients’ goals and preferences. Instead, anesthesiologists should implement an expanded approach to perioperative decision-making regarding CPR that is focused on patient-specific preferences and vulnerabilities. Doing so will take time and a multidisciplinary effort, but is a necessary step toward realizing “the traditional medical practice of responding individually and compassionately to the unique needs of each patient.”¹³ 

Support was provided solely from institutional and/or departmental sources.

Dr. Bernacki is a board member of the American Academy of Hospice and Palliative Medicine (Chicago, Illinois). The other authors declare no competing interests.