In-Hospital Cardiac Arrest

Abstract

IMPORTANCE

In-hospital cardiac arrest is common and associated with a high mortality rate. Despite this, in-hospital cardiac arrest has received little attention compared with other high-risk cardiovascular conditions, such as stroke, myocardial infarction, and out-of-hospital cardiac arrest.

OBSERVATIONS

In-hospital cardiac arrest occurs in over 290 000 adults each year in the United States. Cohort data from the United States indicate that the mean age of patients with in-hospital cardiac arrest is 66 years, 58% are men, and the presenting rhythm is most often (81%) nonshockable (ie, asystole or pulseless electrical activity). The cause of the cardiac arrest is most often cardiac (50%-60%), followed by respiratory insufficiency (15%-40%). Efforts to prevent in-hospital cardiac arrest require both a system for identifying deteriorating patients and an appropriate interventional response (eg, rapid response teams). The key elements of treatment during cardiac arrest include chest compressions, ventilation, early defibrillation, when applicable, and immediate attention to potentially reversible causes, such as hyperkalemia or hypoxia. There is limited evidence to support more advanced treatments. Post–cardiac arrest care is focused on identification and treatment of the underlying cause, hemodynamic and respiratory support, and potentially employing neuroprotective strategies (eg, targeted temperature management). Although multiple individual factors are associated with outcomes (eg, age, initial rhythm, duration of the cardiac arrest), a multifaceted approach considering both potential for neurological recovery and ongoing multiorgan failure is warranted for prognostication and clinical decision-making in the Post–cardiac arrest period. Withdrawal of care in the absence of definite prognostic signs both during and after cardiac arrest should be avoided. Hospitals are encouraged to participate in national quality-improvement initiatives.

CONCLUSIONS AND RELEVANCE

An estimated 290 000 in-hospital cardiac arrests occur each year in the United States. However, there is limited evidence to support clinical decision making. An increased awareness with regard to optimizing clinical care and new research might improve outcomes.

In-hospital cardiac arrest is an acute event that can potentially affect any hospitalized patient. For the purposes of clinical care, research, and guideline development, in-hospital cardiac arrest (as opposed to death without resuscitation) is most commonly defined as the loss of circulation prompting resuscitation with chest compressions, defibrillation, or both.

Traditionally, in-hospital cardiac arrest has been viewed as a condition with such poor outcomes that resuscitation may not even be warranted. Although outcomes remain poor, recent data suggest improvement over the past 2 decades.^1,2 One reason for this improvement might be an increased awareness of the influence that clinical management can have on outcomes in patients with in-hospital cardiac arrest and cardiac arrest in general. Despite this increased interest, in-hospital cardiac arrest remains a somewhat neglected condition compared with out-of-hospital cardiac arrest and other cardiovascular conditions, such as stroke and myocardial infarction. For example, in a systematic review of all randomized clinical cardiac arrest trials (n = 92) involving at least 50 patients from 1995 to 2014, only 4 (4%) exclusively involved patients with in-hospital cardiac arrest.³ Although guidelines for in- and out-of-hospital cardiac arrest are almost identical,^4,5 there are important differences between the conditions that warrant consideration (Table 1).

Table 1.

Comparison of Out-of-Hospital and In-Hospital Cardiac Arrest

	In-Hospital Cardiac arrest	Out-of-Hospital Cardiac Arrest
Incidence	290 000 per year in the United States	350 000 per year in the United Statesa
Patient characteristics	Mean age: 66 y Approximately 60% men	Median age: 65 y Approximately 60% men
Presenting rhythm	Often nonshockable (approximately 80%)	Often nonshockable (approximately 80%)
Cause	Primarily cardiac and respiratory	Primarily cardiac
Prevention	Potentially possible with recognition of deterioration and early intervention	Often impossible given the lack of pre-cardiac arrest monitoring
Timing of basic life support	Often instantaneously	Variable depending on bystander involvement
Timing of advanced life support drugs	Within 5 to 10 min	On average, approximately 20 min after the onset of cardiac arrest
Airway management	Approximately one-third of patients already intubated (eg, intensive care unit patients); often performed by physicians	Often performed by clinicians (eg, paramedics) with variable experience in advanced airway management
Drugs	Limited evidence; epinephrine and amiodarone recommended	Some evidence; epinephrine and amiodarone recommended
Post-cardiac arrest treatment	Limited evidence; supportive care and targeted temperature management recommended	Some evidence; supportive care and targeted temperature management recommended
Prognostication	Limited evidence; focuses on both neurological status and organ failure	Some evidence; focuses on neurological status
Survival to discharge	Approximately 25%	10% to 12%

^aAssessed by emergency medical services but not necessarily treated.²

In this review, we discuss adult in-hospital cardiac arrest, including epidemiology, causes, management during and after cardiac arrest, characteristics related to outcomes, prognostication, and quality improvement. There are relatively few randomized clinical trials of patients with in-hospital cardiac arrest (Table 2). Therefore, much of the current knowledge is based on observational studies primarily from large registries, extrapolation of results from trials of out-of-hospital cardiac arrest, and expert opinion.

Table 2.

Randomized Trials Including In-Hospital Cardiac Arrest Patients Between 1988 and 2018

Study	Country	Sites	Years of Inclusion	Main Inclusion Criteria	Main Exclusion Criteria	Patients	Intervention	Control	Main Findings
Brain Resuscitation Clinical Trial II Study Group,6 1991a	Multiple	24	1984-1989	IHCA, age ≥12 y, ROSC, comatose	Terminal illness, central nervous system disease, atrial fibrillation, women of childbearing age	184	Lidoflazine (calcium channel blocker)	Placebo	No difference in survival at 6 mo (12% vs 15%)
Stiell et al,7 1992a	Canada	2	1989-1992	IHCA, age ≥16 y	Terminal illness, acute trauma, in the operating or recovery rooms of hospitals	315	High-dose epinephrine (7 mg; max 5 doses)	Standard-dose epinephrine (1 mg)	No difference in ROSC (25% vs 31%) or survival to hospital discharge (5% vs 7%)
Sack et al,8 1992	United States	1	1989-1890	IHCA, age ≥18 y	Traumatic or respiratory cause, pregnancy, abdominal aortic aneurysm	103	Interposed abdominal counterpulsation CPR	Standard CPR	Higher ROSC (60% vs 25%), survival to hospital discharge (25% vs 7%), and survival to hospital discharge neurologically intact (17% vs 6%) in the intervention group
Sack et al,9 1992	United States	1	1990-1991	IHCA, age ≥18 y, initial nonshockable rhythm	Traumatic or respiratory cause, pregnancy, abdominal aortic aneurysm, recent abdominal surgery, prolonged intubation	143	Interposed abdominal counterpulsation CPR	Standard CPR	Higher ROSC (49% vs 28%) and survival at 24 h (33% vs 13%) in the intervention group. Few survived to hospital discharge
Cohen et al,10 1993	United States	1	1992-1993	IHCA, age ≥18 y, witnessed cardiac arrest	Traumatic cause, inability to achieve endotracheal intubation within 15 min	62	Active compression-decompression CPR	Standard CPR	Higher ROSC (62% vs 30%) and survival at 24 h (45% vs 9%) in the intervention group. Few survived to hospital discharge
Lipman et al,11 1993	South Africa	1	1990-1991	IHCA in the ICU, age ≥18 y, witnessed cardiac arrest, asystole	NA	37	High-dose epinephrine (10 mg)	Standard-dose epinephrine (1 mg)	No difference in ROSC (68% vs 66%) or survival at 24 h (21% vs 31%). Few patients survived to hospital discharge
Tucker et al,12 1994	United States	1	1992-1993	IHCA, age ≥18 y	Traumatic or respiratory cause, pregnancy	53	Active compression-decompression CPR	Standard CPR	Higher ROSC (60% vs 32%) and survival at 24 h (48% vs 21%) in the intervention group. No difference in survival to hospital discharge (24% vs 11%) or neurologically intact at hospital discharge (20% vs 11%)
Woodhouse et al,13 1995	Australia	1	1989-1992	IHCA or OHCA with ongoing CPR in the ED	Noncardiac cause	194 (109 with IHCA)	Epinephrine (10 mg; max 2 doses)	Placebo (max 2 doses)	No difference in ROSC (10% vs 7%). No patients survived to hospital discharge
Patrick et al,14 1995	Canada	1	1990-1992	Witnessed IHCA or OHCA with ongoing CPR in the ED	Traumatic cause, severe hypothermia/hemorrhage	145 (91 with IHCA)	Methoxamine (40 mg per dose)	Epinephrine (2 mg per dose)	No difference in ROSC (42% vs 53%), 24-h survival (30% vs 37%), or survival to hospital discharge (12% vs 15%)
Stiell et al,15 1996a	Canada	5	1993-1995	IHCA, age ≥16 y	Terminal illness; acute trauma or exsanguination; recent sternotomy; arrest in operating, recovery, or delivery room	773	Active compression-decompression CPR	Standard CPR	No difference in survival at 1 h (35% vs 35%) or at hospital discharge (10% vs 11%)
Thel et al,16 1997	United States	1	1993-1996	IHCA, age ≥18 y	Arrest in emergency, operating, or recovery room; clinical indication for magnesium; signs of irreversible death	156	Magnesium (2 g bolus + 8 g infusion over 24 h)	Placebo	No difference in ROSC (54% vs 60%), survival at 24 h (43% vs 50%), or at hospital discharge (21% vs 21%)
Stiell et al,17 2001	Canada	3	1997-1998	IHCA, age ≥16 y	Terminal illness; traumatic injury prior to admission; arrest due to exsanguination; arrest in operating, recovery, or delivery room	200	Vasopressin (40 IU; max 1 dose)	Epinephrine (1 mg)	No difference in survival at 1 h (39% vs 35%), 24 h (26% vs 24%), hospital discharge (12% vs 14%), or 30 d (13% vs 14%)
Mentzelopoulos et al,18 2009	Greece	1	2006-2007	IHCA, age ≥18 y	Terminal illness, exsanguination, treatment with IV corticosteroids prior to the cardiac arrest	100	Vasopressin (20 IU; max 5 doses) + methylprednisolone (40 mg) + hydrocortisone (300 mg/ 24 h) if in shock after cardiac arrest	Placebo	Higher ROSC (81% vs 52%) and survival to hospital discharge (19% vs 4%) in the intervention group
Weidman et al,19 2010	United States	1	2007-2008	IHCA, age ≥18 y	Arrest in emergency or operating room	98	Immersive simulation training of residents	Standard training	No difference in CPR quality between groups
Pittl et al,20 2013	Germany	1	2008-2009	IHCA or OHCA, age ≥18 y, ROSC, cardiac origin, comatose	Pregnancy, coagulation disorder, cardiogenic shock, terminal illness	80b	Invasive cooling	Surface cooling	No difference in survival to hospital discharge (62% vs 54%) or with a good neurological outcome (36% vs 36%). More bleeding complications with invasive cooling
Mentzelopoulos et al,21 2013	Greece	3	2008-2010	IHCA, age ≥18 y	Terminal illness, exsanguination, treatment with intravenous corticosteroids prior to the cardiac arrest	268	Vasopressin (20 IU; max 5 doses) + methylprednisolone (40 mg) + hydrocortisone (300 mg/ 24 h) if in shock after cardiac arrest	Placebo	Higher ROSC (84% vs 66%) and survival to hospital discharge with a good neurological outcome (14% vs 5%) in the intervention group
Vahedian-Azimi et al,22 2016	Iran	4	2014	IHCA in the ICU, age ≥18 y	NA	83	CPR with feedback devise	Standard CPR	Higher CPR quality and ROSC (72% vs 35%) in the intervention group
Eastwood et al,23 2016	Australia and New Zealand	4	2012-2014	IHCA or OHCA, age ≥18 y, mechanically ventilated	Traumatic cause, pregnancy, imminent death, raised intracranial pressure/bleeding, severe chronic airflow limitation, severe metabolic acidosis	86 (16 with IHCA	Mild hypercapnia after cardiac arrest (PaCO2 50-55 mm Hg) for 24 h	Normocapnia after cardiac arrest (PaCO2 35-45 mm Hg) for 24 h	Lower NSE values in the intervention group. No difference in survival to hospital discharge (74% vs 63%) and good neurological outcome at 6 mo (59% vs 46%)
Movahedi et al,24 2016	Iran	1	2014	IHCA, endotracheal intubation, age 18-85 y, arrest in hospital ward or ED	Traumatic cause, pregnancy, abdominal surgery in the past 2 weeks, history of abdominal aortic aneurysm, coagulopathy, ascites, active gastrointestinal bleeding, pulmonary embolism	83	Interposed abdominal counterpulsation CPR	Standard CPR	Higher end-tidal carbon dioxide in the intervention group. No difference in ROSC (60% vs 53%) or survival at 24 h (38% vs 38%)
Anantharaman et al,25 2017	Singapore	4	2005-2008	IHCA or OHCA with ongoing CPR in the ED, age ≥21 y, initial shockable rhythm	Traumatic cause, pregnancy	235 (89 with IHCA)	Escalating high-energy shocks (200-300-360 J)	Low-energy shocks (150-150 - 150 J)	No difference in first shock success (67% vs 64%), ROSC (55% vs 55%), or 30-d survival (21% vs 28%)
Koster et al,26 2017a	The Netherlands	1	2006-2014	IHCA, age ≥18 y	Traumatic cause	199	Mechanical chest compressions (AutoPulse or LUCAS)	Manuel chest compressions	No difference in serious or life-threatening visceral damage (9% vs 8%)
Zhang et al,27 2017	China	50	2012-2015	IHCA, age ≥18 y	Pregnancy, malignancy, HIV, arrest caused by brain/liver/lung disease, end-stage heart disease, Shenfu (a traditional Chinese medication including ginseng and aconite) allergy	978	Shenfu (200 mL/d for 14 d)	Standard care	Higher 28-d (43% vs 30%) and 90-d survival (40% vs 26%) in the intervention group
Look et al28 2018	Singapore	1	2006-2014	IHCA or OHCA, age 18-80 y	Traumatic cause, hemodynamically unstable, pregnancy, poor premorbid status	45 (7 with IHCA)	Invasive cooling	Surface cooling	No difference in 30-d survival (48% vs 32%) or good neurological outcome (30% vs 23%)

Abbreviations: CPR, cardiopulmonary resuscitation; ED, emergency department; IHCA, in-hospital cardiac arrest; ICU, intensive care unit; NA, not applicable; NSE: neuron-specific enolase; OHCA, out-of-hospital cardiac arrest; PaCO₂, partial pressure of carbon dioxide; ROSC, return of spontaneous circulation.

^aThe trial also included patients with OHCA, but those patients are not included in this table.

^bNumber of IHCAs not reported.

Methods

This review was based on a series of informal searches of PubMed addressing each relevant topic, including incidence and outcomes, causes, prevention, treatment, and prognostication of in-hospital cardiac arrest. In addition, we performed a systematic search to identify all randomized trials including patients with in-hospital cardiac arrest published within the past 30 years. Additional details regarding the systematic search are provided in Supplement 1.

Results

Incidence and Outcomes

The global incidence of in-hospital cardiac arrest in adults has not been well described,^29,30 and the majority of data are derived from the American Heart Association’s Get With The Guidelines-Resuscitation (GWTG-R) registry³¹ and the National Cardiac Arrest Audit from the Resuscitation Council (UK) and the Intensive Care National Audit and Research Centre.³² Based on GWTG-R data from 2003 to 2007, the estimated incidence of in-hospital cardiac arrests in the United States was 211 000 annually, or roughly 6 to 7 cardiac arrests per 1000 admissions.^30,31 Data from 2008 to 2017 showed the incidence increased to 292 000 annually, or 9 to 10 in-hospital cardiac arrests per 1000 admissions.³³ In contrast, an incidence of 1.6 in-hospital cardiac arrests per 1000 admissions in the United Kingdom from 2011-2013 was estimated using data from the UK National Cardiac Arrest Audit.³²

Based on data from the GWTG-R registry, the mean age of patients with in-hospital cardiac arrest in the United States is 66 years, 58% are men, and the presenting rhythm is most often (81%) nonshockable (ie, asystole or pulseless electrical activity). Approximately half of in-hospital cardiac arrests occur in wards, with the remaining half occurring in other locations, such as intensive care units and operating rooms.^2,32,34

In a review from 2007, survival (most commonly to hospital discharge) varied from 0% to 42% between studies, although most larger studies reported survival around 20%.²⁹ Survival has been increasing over the last 2 decades (Figure 1)^1,2 and, in 2017, survival to hospital discharge was 25% in the GWTG-R registry. Among patients alive at hospital discharge, 85% were discharged with a favorable neurological outcome (cerebral performance category 1 or 2).² Data from 2011 to 2013 indicate 18% survival to hospital discharge in the United Kingdom³² while, in Denmark and Sweden, 30-day survival is approximately 30% in contemporary national registries.^35,36 A 2018 systematic review that included more than 1 million in-hospital cardiac arrests and 39 studies from 1992 to 2016 found an overall 1-year survival rate of 13%, with large between-study variability and an increase in 1-year survival over time.³⁷ In elderly (aged ≥65 years) patients with in-hospital cardiac arrest in the United States who survived to hospital discharge, 59% were alive after 1 year and 34% had not been readmitted to a hospital.³⁸

Figure 1. Survival After In-Hospital Cardiac Arrest, 2000 to 2017.

Based on data from the Get With The Guidelines-Resuscitation registry on all adult in-hospital cardiac arrests from 2000 to 2017 The dotted lines represent 95% CIs. Shockable rhythms include ventricular fibrillation and pulseless ventricular tachycardia. Nonshockable rhythms include asystole and pulseless electrical activity. Adapted from Benjamin et al.²

The variability between countries in both incidence of and survival after in-hospital cardiac arrest likely reflects differences in (1) the definitions used to identify in-hospital cardiac arrest, (2) the proportion of cardiac arrests captured by various registries, (3) the patient populations, (4) country-specific culture surrounding cardiopulmonary resuscitation (CPR), do-not-resuscitate orders, and withdrawal of care, and (5) treatment during and after cardiac arrest. Comparisons between countries or registries should therefore be performed carefully.

Causes of Cardiac Arrest

Historically, the etiologies of cardiac arrest have been dichotomized as cardiac or noncardiac. Because patients with no obvious cause are generally classified as cardiac, and because discrepancies often exist between clinical and postmortem diagnoses, the causes of cardiac arrest are often uncertain. In general, cardiac causes of cardiac arrest, such as myocardial infarction, arrhythmia, or heart failure, are most frequent, with a prevalence of approximately 50% to 60%. Respiratory insufficiency is the second most common cause (15%-40%).^34,39,40 The median admission duration prior to cardiac arrest is 1 to 2 days, with a higher prevalence of respiratory insufficiency as the cause of cardiac arrest with longer duration of preceding hospitalization.^32,36,40 Neurological causes of cardiac arrest are rare in the in-hospital setting.^36,41

Identifying the cause of cardiac arrest serves several purposes. During cardiac arrest, resuscitation guidelines emphasize that potential reversible causes should be identified, which are categorized into 4 h’s and t’s (Box).⁴² Although not all of these categories (eg, hypothermia) are applicable in the in-hospital setting, the majority of in-hospital cardiac arrests can be categorized using this approach.⁴³ Identifying the cause of cardiac arrest could improve outcomes.⁴⁴ Identification of the cause of cardiac arrest also has implications if return of spontaneous circulation is achieved, because Post–cardiac arrest organ dysfunction is partly dependent on the underlying cause, and post–cardiac arrest treatment should be tailored accordingly (see “Treatment After Cardiac Arrest”). Additional research is needed to create a useful and more precise framework for the classification of causes of cardiac arrest. Such a framework might not only improve research efforts but may eventually improve care provided during and after cardiac arrest.

Box. Potential Reversible Causes of Cardiac Arrest.

h's

Hypokalemia/hyperkalemiaa

Hypothermia

Hypovolemia

Hypoxia

t’s

Tamponade

Tension pneumothorax

Thrombosis (coronary or pulmonary)

Toxins

^aCan include other metabolic alterations such as severe acidosis.

Prevention of cardiac arrest will be best achieved by addressing the mechanisms underlying the cause of the cardiac arrest. For example, the prescription of QT interval–prolonging drugs during hospital admission can lead to arrhythmias, while the prescription of opioids or sedatives may lead to respiratory insufficiency ^45,46 Another potentially preventable cause of cardiac arrest is sepsis. The prevalence of preexisting sepsis in patients with in-hospital cardiac arrest varies across studies, ranging from 13% to 27%.^34,47 Organ failure from sepsis contributes to multiple potential causes of arrest, including circulatory failure, respiratory insufficiency, and metabolic derangements.

Prevention of Cardiac Arrest

Contrary to the out-of-hospital setting, in-hospital cardiac arrest facilitates observation of a patient’s clinical condition prior to cardiac arrest. Clinical deterioration is common prior to in-hospital cardiac arrest,⁴⁸ and many in-hospital cardiac arrests are considered preventable or avoidable on retrospective review.⁴⁹ Prevention has therefore been added as the first link in the Chain of Survival for in-hospital cardiac arrest in the 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care.⁵⁰ Key elements of success include identification of at-risk patients combined with early interventions to prevent deterioration to cardiac arrest. Such identification might occur through the use of early warning systems triggered by specific vital sign abnormalities, a scoring system based on multiple criteria, or by staff concern. However, current prediction models often lack optimal sensitivity and/or specificity to identify at-risk patients and may be limited by differences in the logistical structure of individual hospital systems.⁵¹ Further, although interventions in terms of rapid response teams are generally supported by the literature, the evidence is weak because there are few rigorous randomized clinical trials and heterogeneity between hospital systems.⁵² Despite these limitations, individual hospitals should create a 2-part system for preventing cardiac arrest that includes (1) identifying at-risk or deteriorating patients (requiring relevant education, monitoring, and recognition) and (2) creating appropriate interventional responses (eg, rapid response teams).

Treatment During Cardiac Arrest

Chest compressions, ventilation, and early defibrillation, when applicable, are the cornerstones of cardiac arrest treatment.^4,5 Early initiation of CPR is associated with improved outcomes for both out-of-hospital⁵³ and in-hospital⁵⁴ cardiac arrest. CPR training for all hospital personnel has, therefore, been mandatory in many hospital systems for decades, facilitating the rapid identification and management of cardiac arrest prior to the arrival of the cardiac arrest team. Quality of chest compressions and of CPR in general have been associated with better outcomes in patients with cardiac arrest.⁵⁵ Optimization of CPR quality is therefore a priority.

Although only approximately 20% of patients with in-hospital cardiac arrest have an initial shockable rhythm, rapid defibrillation is associated with improved outcomes for these patients.⁵⁶ It is not clear whether automated (compared with manual) external defibrillators, which have been associated with markedly improved outcomes in the out-of-hospital setting,⁵⁷ are of any benefit in the in-hospital setting.⁵⁸

Data supporting the efficacy of medications during in-hospital cardiac arrest are sparse. Current guidelines recommend the use of epinephrine and amiodarone, both of which improve short-term outcomes in out-of-hospital cardiac arrest, but there is limited evidence supporting substantial neurological improvement when these medications are used.^59,60 Given the differences between in-hospital and out-of-hospital cardiac arrest (Table 1), especially the much earlier administration of drugs in the in-hospital setting, it is unclear whether findings from studies of out-of-hospital cardiac arrest apply to in-hospital cardiac arrest. For in-hospital events, early administration of epinephrine in patients with a nonshockable rhythm is associated with better outcomes.⁶¹ In contrast, early epinephrine for patients with shockable rhythms is associated with worse outcomes.⁶² The combination of vasopressin and methylprednisolone during in-hospital cardiac arrest has been tested in 2 small randomized clinical trials, with promising results.^18,21 This combination of drugs is not recommended in the US or European guidelines because of insufficient evidence to support their use (see eTable 1 in Supplement 1 for an overview of ongoing randomized clinical trials).^42,63

Airway management is a key component of advanced life support during cardiac arrest. Endotracheal intubation has traditionally been considered the preferred approach to ensure adequate ventilation and oxygenation, but emerging evidence in both out-of-hospital^64-66 and in-hospital⁶⁷ cardiac arrest suggests that alternative approaches, such as bag-valve-mask ventilation or supraglottic airways, may be equally or even more effective. How to optimize ventilation and oxygenation likely depends on the specific clinical conditions of the patient with cardiac arrest.

Extracorporeal circulation during CPR is being assessed for use in cardiac arrest and it may be useful for in-hospital events and in certain patient populations, such as in patients who have recently undergone cardiac surgery. Some reports have shown good outcomes from extracorporeal circulation during CPR but, overall, the evidence of benefit is scarce and limited by the observational nature of the studies.⁶⁸ Extracorporeal circulation during CPR may help carefully selected patients, but balance must be achieved between efficient resource use and clinical benefits.

Treatment After Cardiac Arrest

Management in the Post–cardiac arrest period generally focuses on the precipitating cause, hemodynamic and respiratory support, and neuroprotective care. Conditions before and during cardiac arrest determine the severity of the Post–cardiac arrest syndrome (see “Characteristics Related to Outcomes”) and the need for various interventions. For example, a patient admitted with acute coronary syndrome who develops ventricular fibrillation that is rapidly treated with defibrillation may be cognitively intact and require treatment specific for the cardiac condition without need of neuroprotection. Conversely, a patient who has a prolonged cardiac arrest and significant ischemia-reperfusion injury may have multiorgan injury and require numerous treatments. One of the more distinguishing components of in-hospital cardiac arrest is that the event may result from progressively worsening underlying disease, whereas out-of-hospital arrest is often sudden and unpredictable.

Despite some controversy, targeted temperature management at 32°C to 36°C for at least 24 hours after cardiac arrest remains the primary neuroprotective approach following out-of-hospital cardiac arrest.^69,70 To our knowledge, no randomized trials have evaluated targeted temperature management following in-hospital cardiac arrest. Outcomes data are limited to observational studies and extrapolation from the out-of-hospital investigations. The largest observational study of targeted temperature management in patients with in-hospital cardiac arrest (N = 26 183) found that it was associated with worse overall outcomes.⁷¹ However, this study did not consider coma status, which could have biased the results against the intervention because targeted temperature management is only used in comatose patients.⁷² While future studies are needed to better evaluate targeted temperature management for patients with in-hospital cardiac arrest (eTable 1 in Supplement 1), the current American Heart Association recommendation is to provide targeted temperature management for at least 24 hours.⁷⁰

Other proposed neuroprotective strategies for in-hospital cardiac arrest supported by indirect observational data include minimizing supplemental oxygen therapy when oxygen levels are adequate and maintaining normal carbon dioxide levels.^73,74 Observational studies that examined the association between the maintenance of low tidal-volume ventilation and outcomes in patients with out-of-hospital cardiac arrest⁷⁵ and in-hospital cardiac arrest⁷⁶ have shown conflicting results. While further studies are necessary, one important element to note is that patients who experienced cardiac arrest rarely die from refractory respiratory failure,^77,78 potentially limiting the effectiveness of ventilator-based interventions.

Hemodynamic management is essential, but no specific differences have been defined in the management of patients after cardiac arrest compared with other critically ill patients. Preexisting disease, underlying diagnoses, and myocardial stunning from ischemia reperfusion all contribute to the hemodynamic profile of patients after cardiac arrest. Targeted temperature management appears to have some effects on hemodynamics and data from a small 2002 trial suggested that patients receiving targeted temperature management had lower heart rates, increased systematic vascular resistance, and slightly decreased cardiac outputs.⁷⁹ A 2015 analysis of patients cooled to 33°C vs 36°C reported higher vasopressor dosages and lactate levels for patients maintained at 33°C.⁸⁰ Consequently, patients can be maintained with higher temperatures (ie, 36°C) if there is a concern that targeted temperature management negatively influences hemodynamics.⁸¹ Targeted temperature management at 33°C should probably be avoided in patients with sepsis and septic shock⁸² or bacterial meningitis,⁸³ because recent trials suggest worse outcomes in these patient groups with targeted temperature management.

Characteristics Related to Outcomes

Many patient and event characteristics are associated with the clinical outcomes of in-hospital cardiac arrest. Some of these characteristics cannot be modified, such as age, gender, and preexisting conditions, while other characteristics, such as time to drug administration and monitoring, are modifiable and the subject of quality-improvement efforts. While many risk factors, such as advanced age, are predictive of clinical outcomes, none can individually be used to estimate prognosis following cardiac arrest (see “Prognostication”).

Increased age is associated with decreased survival following cardiac arrest in most studies, especially for patients older than 70 years.⁸⁴ The association with outcomes of other demographic features, such as sex and race, is less clear. Although the incidence of in-hospital cardiac arrest is higher among men, men and women have similar clinical outcomes,⁸⁵ although women of child-bearing age (15 to 44 years) may have better outcomes compared with men of the same age.⁸⁶ Studies that investigated the relationship between race and outcomes have found black and Hispanic patients to have lower rates of neurological recovery and survival following in-hospital cardiac arrest compared with white patients.^47,87 Data from the GWTG-R registry have shown that racial disparities in outcomes have narrowed over time, with a reported absolute survival difference between black and white patients of 4.5% in 2000 and 1.8% in 2014.⁸⁷ Differing distribution of risk factors⁸⁸ or variability in patient- and hospital-level care during and after cardiac arrest may explain why these racial differences exist.

The presence of preexisting medical and surgical conditions is strongly associated with outcomes following in-hospital cardiac arrest. For example, malignancy, sepsis, poor functional status prior to the cardiac arrest, pneumonia, hypotension, renal dysfunction, and hepatic dysfunction have been identified as significant predictors of poor survival.^84,89 Conversely, acute myocardial infarction causing an in-hospital cardiac arrest is associated with increased survival compared with cardiac arrest not caused by myocardial infarction.⁹⁰

Factors related to early detection of cardiac arrest, such as the event being witnessed^47,91 or occurring in a monitored location,^34,47,84 are associated with improved outcomes. However, the association between monitoring/location and outcomes^34,47,84,91 is complex given the different case-mix of patients in various locations.

Two of the factors most strongly associated with outcomes are the presenting rhythm and the duration of the cardiac arrest.^84,92,93 Patients with a shockable rhythm have 2 to 3 times higher survival to hospital discharge compared with patients with a nonshockable rhythm (Figure 1).^2,92 While this difference might reflect the potential for more effective treatment in the shockable group (ie, defibrillation), part of the difference is likely also explained by differences in patient characteristics and preexisting conditions, which may influence the presenting rhythm. The chance of 30-day survival markedly decreases with increasing duration of CPR.⁹³

Prognostication

During Cardiac Arrest

Deciding when to stop CPR during cardiac arrest remains challenging, with limited guidance in contemporary guidelines.^94,95 Although longer duration of resuscitation is associated with worse outcomes, survival with good neurological outcome is possible with prolonged CPR.⁹³ Additionally, Goldberger et al⁹⁶ found that hospitals where CPR is performed for longer durations have better outcomes, suggesting that average CPR duration may be too short at some hospitals.

Several observational studies have reported that cardiac standstill on point-of-care cardiac ultrasonography is associated with very low likelihood of survival.⁹⁷ However, there are significant concerns with interrater variability in image interpretation and potential interference of ultrasonography with CPR.⁹⁸ End-tidal carbon dioxide values less than 10 mm Hg after 20 minutes of CPR are also strongly associated with poor outcomes.⁹⁹ However, studies regarding point-of-care cardiac ultrasonography and end-tidal carbon dioxide should be interpreted carefully because blinding was rarely performed and the findings described might reflect a self-fulfilling prophecy where resuscitation is terminated based on these specific findings. Neither cardiac ultrasonography nor end-tidal carbon dioxide should be used in isolation, but might be considered together with other factors, such as the initial rhythm and the duration of the cardiac arrest (see “Characteristics Related to Outcomes”).

After Cardiac Arrest

Current guidelines on prognostication are based on literature describing out-of-hospital cardiac arrest and focus exclusively on neurological status.⁷⁰ While two-thirds of patients with out-of-hospital cardiac arrest who survive to intensive care unit admission die of neurological causes, neurologic death only occurs in one-fourth of patients with in-hospital cardiac arrest, for whom multiorgan dysfunction drives mortality.⁷⁸ Because in-hospital cardiac arrest is most often witnessed, times from cardiac arrest to initiation of CPR and to return of spontaneous circulation are shorter than in out-of-hospital cardiac arrest, which may contribute to the lower rate of neurological injury. These differences leave clinicians without clear guidance regarding how to best prognosticate outcomes (ie, survival and survival with good neurological/functional recovery) following in-hospital cardiac arrest.

Chan et al attempted to address this issue by developing a scoring system for estimation of survival to hospital discharge with a favorable neurological outcome in patients with return of spontaneous circulation after in-hospital cardiac arrest (Figure 2; Supplement 2).⁸⁴ The final score, which has been externally validated,¹⁰⁰ is based on 11 parameters available immediately after the cardiac arrest. The score successfully categorized patients with varying chances of hospital survival.⁸⁴ Such a scoring system can inform conversations about goals of care when patients survive the initial cardiac arrest but remain seriously ill. However, the score is not able to accurately identify many patients with a very low/no chance of survival⁸⁴ and should therefore not be used in isolation for decisions regarding withdrawal of care.

Figure 2. The Cardiac Arrest Survival Postresuscitation In-hospital (CASPRI) Score.

Reprinted from Chan et al,⁸⁴ where a detailed description of the score's interpretation is presented. Scores of 0-4 are associated with 83% survival, 15-19 are associated with 23% survival, and 30-34 are associated with 2% survival. CPC indicates cerebral performance score; VF/VT, ventricular fibrillation or ventricular tachycardia.

Treatment recommendations for patients with in-hospital cardiac arrest with presumed severe neurological impairment following the arrest are based on studies of patients with out-of-hospital cardiac arrest.⁷⁰ A key recommendation is that neuroprognostication should not be performed too early, especially when targeted temperature management is used, because slower metabolism of sedatives and neuromuscular blockade may occur. Several studies have found that delayed awakening (>48 hours after cessation of sedation) is common in patients after cardiac arrest.¹⁰¹ In a study of patients still in a coma 7 days after cardiac arrest, 22% obtained a favorable neurological outcome at 6 months.¹⁰² The American Heart Association recommends that neuroprognostication based on physical examination findings should be deferred until at least 72 hours after return of spontaneous circulation or rewarming (if targeted temperature management is used), and often longer if effects of sedation or neuromuscular blockade may still be present.⁷⁰

The clinical examination findings most predictive of poor outcomes are absent pupillary light reflexes and absent corneal reflexes after 72 hours, as well as status myoclonus (continuous, prolonged, and generalized myoclonus) within 72 to 120 hours of the cardiac arrest. Extensor motor responses and intermittent myoclonus are less predictive of outcomes and should not be relied upon. On electroencephalography, persistent burst suppression after rewarming and nonreactivity predict poor outcomes, as does the absence of somatosensory evoked potentials at 72 hours. Cerebral edema on early head computed tomographic imaging and restricted diffusion on magnetic resonance imaging at 2 to 6 days after return of spontaneous circulation are potentially helpful in predicting outcomes, but cannot preclude a favorable neurological outcome. Biomarkers for the degree of neurological injury (eg, neuron-specific enolase and S100 calcium-binding protein B) do not adequately predict outcomes by themselves, but may be helpful when considered with other clinical features that predict outcomes.^70,103

Little evidence exists for any one tool for prognostication after in-hospital cardiac arrest. A multifaceted approach that assesses the neurological prognosis combined with ongoing organ failure is prefered.^102,104-106 Information regarding neurological prognosis and organ failure can be considered along with patient and cardiac arrest characteristics to support clinical decision making.

Quality Improvement

Organizations have proposed strategies and quality-improvement initiatives to improve outcomes after in-hospital cardiac arrest. For example, the GWTG-R registry currently tracks a number of quality-improvement measures, including the proportion of cardiac arrests that are monitored or witnessed, the time to relevant intervention (defibrillation for shockable rhythms and epinephrine for nonshockable rhythms), and confirmation of correct airway placement.¹⁰⁷ Observational studies have shown that longer duration of participation in the GWTG-R registry is associated with both improved quality of care¹⁰⁸ and increased return of spontaneous circulation.¹⁰⁹ Specific adherence to performance measures is also associated with improvement in outcomes.¹¹⁰ Other aspects of cardiac arrest care, such as training, monitoring, and cardiac arrest team composition, are also potential targets for quality improvement. To monitor and improve quality of care for patients with in-hospital cardiac arrest, hospitals are encouraged to participate in national registries (eTable 2 in Supplement 1).

Limitations

The aim of this review was to provide an overview of in-hospital cardiac arrest. As such, many topics are only briefly discussed and others have been omitted. We did not perform comprehensive systematic reviews to address all topics and, therefore, relevant studies might have inadvertently been omitted. The data and recommendations provided herein are limited by the available data and, in some cases, reflect expert opinion.

Conclusions

An estimated 290 000 in-hospital cardiac arrests occur each year in the United States. However, there is limited evidence to support clinical decision making. An increased awareness with regard to optimizing clinical care and new research might improve outcomes.