Abstract
An 81-year-old woman with a history of non-alcoholic cirrhosis, refractory ascites, and previous variceal bleeding was admitted to the intensive care unit (ICU) with septic shock from spontaneous bacterial peritonitis. She arrived from the emergency department on moderate vasopressor support with norepinephrine and vasopressin. She had no family with her, but was alert enough to participate in the history being obtained. She reported that her quality of life had been gradually decreasing over the past year. She had been admitted to the ICU with septic shock and gastrointestinal bleeding two times over the past 6 months. As part of your routine ICU admission process, you approach her to discuss code status, including her preferences for cardiopulmonary resuscitation (CPR) in the case of in-hospital cardiac arrest. She asked you to do everything to help her get better – and that seeing her six grandchildren grow up is her greatest joy in life.
Case
An 81-year-old woman with a history of non-alcoholic cirrhosis, refractory ascites, and previous variceal bleeding was admitted to the intensive care unit (ICU) with septic shock from spontaneous bacterial peritonitis. She arrived from the emergency department on moderate vasopressor support with norepinephrine and vasopressin. She had no family with her, but was alert enough to participate in the history being obtained. She reported that her quality of life had been gradually decreasing over the past year. She had been admitted to the ICU with septic shock and gastrointestinal bleeding two times over the past 6 months. As part of your routine ICU admission process, you approach her to discuss code status, including her preferences for cardiopulmonary resuscitation (CPR) in the case of in-hospital cardiac arrest. She asked you to do everything to help her get better – and that seeing her six grandchildren grow up is her greatest joy in life.
Even though in-hospital cardiac arrest is a relatively rare event, determining and documenting orders for CPR before the onset of cardiac arrest is important for hospitalized patients and their medical teams. Once cardiac arrest ensues, unresponsive patients are unable to participate in medical decision-making, and delays in initiating CPR may lead to poorer outcomes.1 “Code status” conversations are dialogues between patients, surrogates, and clinicians in which preferences for CPR are ascertained. As a result of code status conversations, patients may be “full code,” which entails receiving CPR should they suffer cardiac arrest, or “do-not-resuscitate” (DNR), meaning that no CPR would be performed if the patient were to suffer cardiac arrest. Other names for DNR include do-not-attempt-resuscitation or allow natural death. Decisions regarding CPR may be integrated with decisions to allow versus limit other treatments, such as the need for intubation and mechanical ventilation, hemodialysis, or vasopressor support. Sometimes, patients are unilaterally assigned a code status by the medical team (and not involved in the decision-making), although this is less common in the United States. Also, in the United States, if patient or surrogate preferences are unclear, the default code status is full code and CPR is administered.
The decision to receive CPR is often considered to be a preference-sensitive decision (meaning patients and surrogates have to be involved in the decision-making process). Shared decision-making is one recommended method to collaboratively engage patients, surrogates, and clinicians in the decision-making for preference sensitive decisions.2 This chapter examines the incidence and outcomes of cardiac arrest and the challenges with shared decision-making for DNR, as well as methods and interventions to improve decision-making.
5.1 Incidence and Outcomes of Patients with In-Hospital Cardiac Arrest
A knowledge of the incidence and outcomes of in-hospital cardiac arrest (as well as the likelihood that an individual patient might experience such outcomes) is often, but not necessarily always, helpful to inform high-quality shared decision-making regarding code status.
5.1.1 Incidence
Although approximately 290,000 patients experience in-hospital cardiac arrest annually in the United States, the overall incidence is rare.3 The mean cardiac arrest event rates per 1000 inpatient bed-days are estimated to be 0.58 (for all hospitalized patients), 0.34 (for ICU patients), 0.11 (for monitored ward patients), and 0.13 (for unmonitored ward patients).4 Approximately 20% of patients with in-hospital cardiac arrest have shockable rhythms (ventricular fibrillation and pulseless ventricular tachycardia).5
5.1.2 Hospital Survival
Among adult patients who experience in-hospital cardiac arrest, approximately 25% survive to hospital discharge5 according to 2017 data from the American Heart Association’s Get With The Guidelines-Resuscitation registry, which includes more than 400 hospitals in the United States.6 Although there is marked variability in hospital survival (0%–42% in published studies worldwide),7 most larger studies in the past 10 years report survival rates of approximately 15%–25%.8 The rate of hospital survival for patients with shockable rhythms is approximately 45%, compared with 20% for patients with nonshockable rhythms,8 and early defibrillation is associated with improved survival.9
Over the past 20 years, the rate of hospital survival for patients with in-hospital cardiac arrest has increased from approximately 17% in 2000 to 25% in 2017.5, 8, 10 Survival for in-hospital cardiac arrest is higher than survival for out-of-hospital cardiac arrest, where approximately 10% of adults survive to hospital discharge.5 Several factors associated with poor prognosis include demographic elements, such as older age, non-Caucasian race, and residence in a skilled nursing facility, as well as particular comorbidities, such as renal failure, hepatic dysfunction, acute stroke, and immunodeficiencies.11,12
5.1.3 Long-term Survival
Although the hospital survival rate is estimated to be approximately 20% to 25%, the 1-year survival rate is only estimated to be approximately 13% according to a 2018 systematic review with more than 1 million patients.13 Among older adults (≥65 years old) in the United States who survived to hospital discharge, 59% were alive at 1 year and 50% were alive at 2 years.14
5.1.4 Quality of Life
Among patients who survived to hospital discharge, approximately 80% to 85% survived with an acceptable neurologic outcome at hospital discharge (Cerebral Performance Category 1 or 2; please refer Chapter 10, “Hypoxic-Ischemic Brain Injury After Cardiac Arrest”)5, 14 and approximately 50% were directly discharged from the hospital to home as compared with a rehabilitation center, skilled nursing facility, hospice, or another hospital.15 Among older adults (≥65 years old) in the United States who survived to hospital discharge, 34% had not been readmitted to the hospital at 1 year, and 24% had not been readmitted to the hospital at 2 years.14 In one study assessing the quality of life of patients after cardiac arrest, 75% of survivors were noted to be independent after hospital discharge, 17% were cognitively impaired, and 16% had depressive symptoms. These measures were worse compared with a control group of other elderly individuals, but better than that of a reference group of patients with stroke.16
Nearly all patients with in-hospital cardiac arrest who do not receive CPR die. Nevertheless, patients who die without CPR are often perceived to have a higher quality, less traumatic death with less pain and less distress.
5.2 Predicting the Outcome of Hospitalized Patients
For individual patients at the bedside, predicting the likelihood of survival with an acceptable quality of life in the case of in-hospital cardiac arrest is complex. The likelihood of a good outcome depends on several factors, including patient baseline characteristics, the reversibility of the cause of cardiac arrest, the ability of a hospital system to detect and treat the cause of the cardiac arrest, and the duration of resuscitation. Although published outcome data (see Section 1.2) provides baseline estimates of outcomes for large cohorts of patients, applying these estimates to individual patient situations is fraught with challenges.17
Prediction models have been developed to try to estimate individual patient outcomes after CPR for in-hospital cardiac arrest. One such model is The Good Outcome Following Attempted Resuscitation score (available: https://www.gofarcalc.com/). This scoring system estimates the likelihood of survival to hospital discharge with good neurologic status (Cerebral Performance Category of 1) of adult patients who receive CPR for in-hospital cardiac arrest. The likelihood of survival with a good neurologic status is based on 13 factors present before cardiac arrest.18 The prediction model was derived and validated in a cohort of 51,240 patients with cardiac arrest in more than 400 hospitals in the United States (Get With the Guidelines-Resuscitation cohort).18 The Good Outcome Following Attempted Resuscitation calculator was also validated externally in several populations.19, 20 It has also been adapted for use in decision aids.21
Hospitalized patients tend to overestimate (up to three to four times) their chances of survival in the event of in-hospital cardiac arrest,22 possibly in part owing to media portrayals of successful recovery.23,24
5.3 Challenges Associated with DNR Decision-Making
No matter how many statistics patients have when trying to decide whether or not they want to be resuscitated, the decision remains a personal one, and one of the hardest patients and their families will make each hospital admission. As difficult as this choice is, the conversations preceding the decision are also challenging for physicians. This is especially true in the ICU, where there is usually no preexisting relationship between the patient and the physician and time is limited. Because of this circumstance, there are many ways this conversation can go poorly.
One of the challenges associated with deciding on DNR status is that often people who choose to be DNR are also mistakenly assumed to be patients who should not receive other treatments by members of the health-care team. Patients with a DNR order are often incorrectly considered as someone who would not want to be intubated, not want to be dialyzed, not want to be admitted to the ICU, and sometimes even not want basic treatment like antibiotics. These are distinct entities and, although the outcomes for cardiac arrest are quite poor, the outcomes for patients with respiratory failure or acute kidney injury are often better. Thus, for many people who are DNR, they may feel it is appropriate to receive a trial of other treatments, such as mechanical ventilation or hemodialysis.
Another difficulty associated with the DNR conversation in the ICU is that, even under the best of circumstances, it does not really occur at the right time or with the right person. Ideally, the conversation would happen in the outpatient setting, with an entire visit set aside for the discussion with a physician well-known to the patient, and involving both the patient and the patient’s family. Unfortunately, in the ICU, this is almost never the case. This critical discussion usually occurs early in the morning in a busy unit. It almost always involves a physician who is unfamiliar to the patient and family and probably takes place within the first 10 minutes of meeting this physician. Just as important, the patient is often critically ill and not able to fully participate in the conversation. It may also take place with the most inexperienced person on the team; at academic teaching hospitals, these conversations often fall to the resident physicians. This practice can be problematic, because junior physicians may not be trained at having these difficult conversations skillfully and so may make the mistake of using medical jargon and/or not attempting to elicit the patient’s decision-making preferences. It is not uncommon for junior physicians to awkwardly elicit code status by asking: “If your heart stopped, would you want us to restart it by pounding on your chest?”
In the ICU setting, patients may not be able to fully participate in the code status discussion. This circumstance leaves the decision to a surrogate decision maker, who may not know what their loved one would want in the event of a cardiac arrest. When prior planning has occurred, it may be well-informed and apply to the patient’s current circumstance. When prior planning has not occurred, the surrogate may not be able to act in the “best interest” of the patient, as is expected for an appropriate surrogate decision maker.
5.4 How DNR Shared Decision-Making Can Go Well
Although it is recommended that shared decision-making be used to help patients make preference-sensitive decisions, there is no one recommended technique to use in shared decision-making for code status.2 Different methods of shared decision-making can be used based on the different circumstances of patients.25 Most often, a successful shared decision-making discussion involves putting the patient’s overall prognosis into context with their current and previously outlined goals (if they exist). Individualized factors such as performance status, recent quality of life, prognosis for both acute and chronic issues, and pertinent and applicable national resuscitation data should be applied to allow the patient to make an educated decision.2 Table 5.1 lists possible frameworks for engaging in shared decision-making for CPR versus no CPR.
Table 5.1. Possible frameworks for engaging in decision-making for CPR
Related posts:
Chapter 2 – How Much Does the Family Want to Be Involved in Decision-Making?
Chapter 4 – Communication Skills for Critical Care Family Meetings
Chapter 6 – The Do-Not-Intubate Order
Chapter 14 – Potentially Inappropriate Treatment and Conscientious Objection
Chapter 20 – Brain Death Discussions
Chapter 13 – Severe Traumatic Spinal Cord Injury
Full access? Get Clinical Tree
