Cardiac arrest and cardiopulmonary resuscitation in pediatric patients with cardiac disease: a narrative review

Abstract

Children with cardiac disease are at a higher risk of cardiac arrest as compared to healthy children. Delivering adequate cardiopulmonary resuscitation (CPR) can be challenging due to anatomic characteristics, risk profiles, and physiologies. We aimed to review the physiological aspects of resuscitation in different cardiac physiologies, summarize the current recommendations, provide un update of current literature, and highlight knowledge gaps to guide research efforts. We specifically reviewed current knowledge on resuscitation strategies for high-risk categories of patients including patients with single-ventricle physiology, right-sided lesions, right ventricle restrictive physiology, left-sided lesions, myocarditis, cardiomyopathy, pulmonary arterial hypertension, and arrhythmias. Cardiac arrest occurs in about 1% of hospitalized children with cardiac disease, and in 5% of those admitted to an intensive care unit. Mortality after cardiac arrest in this population remains high, ranging from 30% to 65%. The neurologic outcome varies widely among studies, with a favorable neurologic outcome at discharge observed in 64%-95% of the survivors. Risk factors for cardiac arrest and associated mortality include younger age, lower weight, prematurity, genetic syndrome, single-ventricle physiology, arrhythmias, pulmonary arterial hypertension, comorbidities, mechanical ventilation preceding cardiac arrest, surgical complexity, higher vasoactive-inotropic score, and factors related to resources and institutional characteristics. Recent data suggest that Extracorporeal membrane oxygenation CPR (ECPR) may be a valid strategy in centers with expertise. Overall, knowledge on resuscitation strategies based on physiology remains limited, with a crucial need for further research in this field. Collaborative and interprofessional studies are highly needed to improve care and outcomes for this high-risk population.

Introduction

Children with cardiac disease are at higher risk of cardiac arrest (cardiac arrest) compared to healthy children [1, 2] as a result of their propensity for myocardial dysfunction, arrhythmias, and hemodynamic instability. In particular, children with congenital heart disease (CHD) often have abnormal circulatory physiology including intracardiac shunts, outflow tract obstruction, and common mixing that often leads to abnormal pressure or volume loading conditions, cyanosis, or congestive heart failure. Taken together, these factors can increase the risk of cardiac arrest, especially in the post-operative period after surgical correction or palliation. Additionally, cardiorespiratory interactions have a greater detrimental impact on the hemodynamics of critically ill children with CHD as compared to healthy children, especially when supporting them with invasive positive pressure ventilation in the post-operative period.

The anatomical and physiological substrates of CHD can also influence the effectiveness and success of the resuscitation strategy itself, especially in single ventricle (SV) patients and in neonates [2]. The past few decades have been marked by an intense effort to investigate cardiac arrest prevention and resuscitation strategies in children with cardiac disease. In 2010, the American Heart Association officially identified the pediatric cardiac patient as a specific high-risk patient for cardiac arrest in its resuscitation guidelines, with particular reference to the SV patient [3]. In the 2015 updates, the American Heart Association advised consideration of venoarterial (VA) Extracorporeal Membrane Oxygenation (ECMO) as part of the resuscitation strategy in cardiac patients – either in the pre-arrest phase or during CPR (ECPR) - when expertise and equipment are available [4]. In 2018, these considerations were similarly included in an American Heart Association Resuscitation Statement entirely dedicated to children with cardiac disease [2].

With this narrative review, we aim to review the physiological aspects of resuscitation in different cardiac physiologies, summarize the current recommendations in the field of cardiac arrest in children with cardiac disease, provide un update of current literature, and highlight important knowledge gaps to help guide new research efforts.

Epidemiology and risk factors

A recent study involving a total of 3,739 hospitals in 38 states participating in the Kids’ Inpatient Database showed that CPR occurred in about 1% of the hospitalized children with cardiovascular disease, which corresponds to a 13-fold higher risk of cardiac arrest compared to children without cardiac disease (odds ratio [OR] 13.8, 95% confidence interval [CI] 12.8–15.0) [1]. A recent meta-analysis estimated that about 5% (95% CI 4-6%) of children with a cardiac disease admitted to a pediatric or cardiac intensive care unit (ICU) experienced at least one episode of cardiac arrest during their admission [5]. The higher estimate in the latter study may have resulted from the type of population included (i.e. children receiving higher level of care in the ICU, compared to all hospitalized children).

Table 1 summarizes the main studies addressing the epidemiology and risk factors for cardiac arrest in this population according to hospital setting. Common risk factors included younger age, lower weight and prematurity, genetic syndrome, univentricular physiology, arrhythmias, pulmonary arterial hypertension, renal failure, sepsis, seizures, mechanical ventilation or ECMO before cardiac arrest, recent complex cardiac surgery (Society of Thoracic Surgeons mortality category 4-5), and factors related to available resources and hospital characteristics [1, 5-11]. In the cardiac catheterization laboratory, younger patients and those undergoing interventional procedures were at highest risk [12]. Similarly, in the operating room, neonatal age increased the likelihood of intra-operative CPR events [13]. Knowledge of risk factors can guide a priori identification of patients at the highest risk, so that contingency planning may be put in place in case of clinical deterioration.

Table 1:

Studies addressing risk factors for cardiac arrest in pediatric patients with cardiac disease according to setting.

Author, year	Study design, Setting and period	Population	Exclusion criteria	N patients with CA/N total (%)	Definition of cardiac arrest	Predictors / Risk factors for CA
Hospitalized patients
Lowry AW, 2013	Retrospective analysis of prospective data, KID Registry, Multicenter (38 USA States) 2000,2003,2006	Hospitalized cardiac patients	Not defined	3709/ 498610 (0.7)	ICD-9 procedure code	Multivariable model: RISK: age<1yr OR 2.1 (1.8-2.5), Children’s hospital OR 1.4 (1.2-1.7), small hospital OR 1.3 (1.01-1.6), heart failure OR 2.3 (2- 2.6), PAH OR 2.0 (1.7-2.3), SV OR 2.4 (CI 2-2.8), cardiomyopathy OR 2.1 (1.8-2.6), heart transplant (prior) OR 1.4 (1.01-1.9), acute pericarditis OR 2.4 (1.7-3.2), coronary artery disease OR 2.8 (2-3.8), myocarditis OR 2.7 (2.0 −3.5), bacterial endocarditis OR 1.9 (1.2-2.9). PROTECTIVE: CHD OR 0.6 (0.5-0.8).
Gupta P, 2014 Resuscitation	Retrospective analysis of prospective data, VPS (NACHIRI) Registry, Multicenter (108 USA Centers), 2009-2013	Hospitalized cardiac patients with at least 1 episode of CA	Patients with “altered code status”	2182/ 2182 (100)	Any event characterized by either pulselessness or critically compromised perfusion treated with external chest compression and/or defibrillation	Univariate analysis: Center volume: OR 1.13 (1.07-1.19). Categorical analysis: low volume center OR 0.67 (0.55-0.82), low-medium volume center OR 0.73 (0.65-0.82), high-medium volume center 0.69 (0.61-0.77), reference: high volume center. Multivariable model: NS
Intensive care unit setting
Alten et al., 37 2017	Retrospective analysis of prospective data, PC4 Registry, Multicenter (23 USA Centers), 2014-2016	P-CICU pediatric patients (medical and surgical)	None	492/ 15,908 CICU encounters (3.1) ***	Cardiopulmonary arrest requiring chest compressions and/or defibrillation for pulseless VT or acute respiratory compromise requiring emergency assisted ventilation leading to cardiopulmonary arrest requiring chest compressions and/or defibrillation	Multivariable model: For SURGICAL patients: premature neonate OR 5.04 (2.98-8.54), term neonate OR 3.77 (2.54-5.60), infant OR 2.48 (1.69-3.63), underweight OR 1.56 (1.17 - 2.08), any chromosomal abnormality/ syndrome OR 1.36 (1.04-1.78), any STS preop. risk factor OR 2.14 (1.68-2.74), STS mortality category 4 or 5 OR 3.92 (2.94-5.22). For MEDICAL patients: premature neonate OR 3.15 (1.54-5.37), medical condition OR 2.20 (1.56-3.34), acute heart failure OR 2.23 (1.47-3.19), lactate>3 mmol/L within 2 hours of P-CICU admission OR 3.00 (1.86-4.86), MV 1hr post P-CICU admission OR 2.61 (1.70-3.82).
Gupta et al., 10 2016 Resuscitation	Retrospective analysis of prospective data, VPS (NACHRI) Registry, Multicenter (62 USA Centers) 2009-2014	P-CICU pediatric patients with CHD post cardiac surgery	ICU readmission, lack of surgical documentation, surgical closure of isolated PDA or surgery not listed in STS-EACTS	736/ 26,909 (2.7)	Any event characterized by either pulselessness or critically compromised perfusion treated with external chest compression and/or defibrillation	Multivariable model: RISK: younger age OR 0.73 (CI 0.56-0.96), female OR 1.18 (CI 1.01-1.38), development disorder OR 1.71 (CI 1.16-2.51),high complexity operations OR 1.81 (CI 1.51-2.16), MV before surgery OR 2.79 (CI 2.33-3.35), higher PIM-2 score OR 1.28 (CI 1.20-1.36), SV anatomy OR 1.3 (CI 1.08-1.57), PAH OR 1.8 (CI 1.4-2.3), acute lung injury OR 1.50 (CI 1.27-1.77), renal failure OR 2.92 (CI 2.29-3.71), chylothorax OR 1.65 (CI 1.11-2.47), arrhythmia OR 2.69 (CI 2.29-3.16), seizures OR 3.60 (CI 2.82,−4.59), brain hemorrhage OR 2.13 (CI 1.27-3.57), MV after surgery OR 1.52 (CI 1.07-2.16), hemodialysis catheter in place OR 1.98 (CI 1.13-3.46). PROTECTIVE: younger Age (>28d,<1years) OR 0.73 (CI 0.56-0.96), higher weight OR 0.73 (CI 0.88-0.00), arterial line OR 0.58 (CI 0.35-0.96), attending intensivist OR 0.35 (CI 0.26-0.47)
Gupta et al., 8 2014 Ann Thorac Surg	Retrospective analysis of prospective data, STS-CHSD Registry, Multicenter (97 USA Centers), 2007-2012	P-CICU pediatric patients with CHD post cardiac surgery	Surgery not classified into one of the STS-EACTS Mortality Categories, missing outcome data	1843/ 70,270 (2.6)	Cessation of effective cardiac mechanical function	Univariate analysis: Female sex (p=0.003), lower age (p<0.0001), lower weight(p<0.0001), prematurity (p<0.0001), congenital disorders (p<0.0001), preop. LOS (p<0.0001), preop. MV (p<0.0001), preop. sepsis (p<0.0001), preop. shock (p<0.0001), preop. RI (p<0.0001), preop. CPR (p<0.0001), CPB time(p<0.0001), previous cardiothoracic surgery (p<0.0001), STS mortality high risk(p<0.0001), STS morbidity high risk (p<0.0001). Multivariable model: NS
Hansen et al., 43 2011	Case-control, Single-center (Edmonton, Canada) 1996-2005	NICU patients post cardiac surgery with CPB, ≤6 weeks of age. Cases: at least 1 CPR event, Controls: no CPR events	Cardiac surgery not requiring CPB, patients having CPR preoperatively or in the operating room	29 CA (cases) (of 343 patients post cardiac surgery) (8.5)	Not defined	Univariate analysis: Lower birth weight (−0.57; 95%CI, −0.84, −0.31 kg) and gestational age (−1.5; 95%CI, −2.64, −0.40 weeks), longer preop. ventilator days (4.1; 95%CI, 1.0, 7.2), worse postop. day 1 peak lactate (4.1; 95%CI, 2.3, 5.9 mmol/L), postop. day 1 base deficit (−2.9; 95%CI,−5.4,−0.3), postop. day 1 pH (−0.04; 95%CI,−0.08,−0.01), and inotrope score (11.6; 95%CI, 3.3, 22.4)
Suominen et al., 3 2001	Case control Single-center (Helsinki, Finland), 1990-1994	P-CICU pediatric patients with CHD post cardiac surgery. Cases: at least 1 CA; Controls 1: with DHCA, no CA; Controls 2: without DHCA, no CA	Patients who only received resuscitation drugs or MV, or who had received CPR in the operating theater	82 CA (48 CPR, 44 resuscitation not attempted) (of 1,115 post cardiac surgery patients) (7.3)	Absence of consciousness, apnea, and lack of palpable pulses in major arteries	Univariate analysis: Younger age (p=0.04), SV (p<0.01), preop. MV (p=0-03), PGE1 (p<0.001), preop. inotropic support (p=0.04), longer mean aortic-cross-clamp time (p<0.0001), longer CPB time (p=0,0002), longer DHCA time (P =0,0002), higher inotropic support during surgery (p<0.0001) and higher postop. inotropic support (p=0.002)
Cardiac catheterization laboratory setting
Odegard KC, 2014	Retrospective, CAS and CA Registry, Single-Center (Boston, USA), 2004-2009	Patients with CHD undergoing cardiac catheterization	Not defined	70 / 7289 (1.0)	Sudden cessation of cardiopulmonary circulation or ventilation requiring external chest compressions for resuscitation	Univariate analysis: Interventional procedures (p<0.001), younger age (p<0.001)
Operating room setting
Odegard KC, 2007	Retrospective, Single-center (Boston, USA), 2000-2005	Patients with CHD undergoing cardiac surgery	Not defined	40/ 5213 (0.8)	Any event requiring external chest compression or internal cardiac massage with or without cardioversion	Univariate analysis: Neonatal age (p<0.001)

CA: cardiac arrest; CHD: congenital heart disease; CPB: cardiopulmonary bypass; CPR: cardiopulmonary resuscitation; DHCA: deep hypothermic circulatory arrest; MV: mechanical ventilation; NA: not applicable; NS: not significant; NICU: neonatal intensive care unit; OR: odds ratio; PAH: pulmonary arterial hypertension; P-CICU: pediatric cardiac intensive care unit; PGE1: prostaglandin E1; PICU: pediatric intensive care unit; Preop.: preoperative; Postop.: postoperative; STS: Society of Thoracic Surgeon; SV: single ventricle.

Pre-resuscitation measures and resuscitation peculiarities based on physiology

Single ventricle physiology

Patient with SV physiology have the highest risk of cardiac arrest. For example, patients with hypoplastic left heart syndrome undergoing the Stage 1 palliation (i.e. Norwood operation), which entails reconstruction of the aorta to provide adequate systemic flow, removal of atrial restriction, and establishment of a source of pulmonary blood flow with a right ventricle (RV) to pulmonary artery or a modified Blalock Taussig Thomas shunt, have parallel systemic (Qs) and pulmonary (Qp) circulations. These parallel circulations may be imbalanced, resulting in a higher Qp than Qs, especially in the late neonatal and early infant periods [14]. Additional risks include low cardiac output syndrome in the early postoperative period, during which myocardial work and oxygen demand are increased, and occlusion of the shunt, the sole source of pulmonary blood flow. Each of these conditions increases the risk of rapid cardiovascular collapse with the need for resuscitation [2-4]. In this cohort, the rate of in-hospital cardiac arrest has been described to be up to 12.7% in the acute postoperative period [9], with a mortality or transplant rate of 31% at 12 months [15]. Interestingly, the incidence of cardiac arrest was significantly lower in the presence of a Sano shunt compared to a BTT shunt [15]. Risk of cardiac arrest also increase in the presence of a genetic syndrome, ventricular dysfunction, or significant atrioventricular valve regurgitation (AVVR) [16-18]. Given this unique physiology, resuscitation in these patients represents a challenge. Main interventions that may help prevent a cardiac arrest include: (1) early recognition and treatment of low cardiac output syndrome: correction of acidosis, reduction of metabolic demand with sedation and paralysis, adoption of a low mean airway pressure ventilation strategy, initiation of inotropic support, maintenance of an open chest, and early deployment of VA-ECMO support when there is significant instability or inadequate response to medical management; (2) careful balancing of the two circulations through manipulation of pulmonary and systemic vascular resistance, with avoidance of hyperventilation and hyperoxygenation or use of systemic vasodilators [19], and (3) early recognition and treatment of shunt obstruction, with anticoagulation or interventional procedures [2, 18]. In the event of cardiac arrest, it is important to consider chest compressions in these patients will provide flow to parallel circulations, with the majority of flow directed to the lower vascular resistance district (generally the lungs) and consequently decreased flow to other vital organs, including the brain [2]. Adequate chest recoil and target of low mean airway pressure ventilation can improve filling of the preload-dependent SV and the consequent compression related output (Figure 1) [2].

Figure 1.

Pre-arrest phase characteristics and cardiopulmonary resuscitation strategies according to cardiac physiology. CPR: cardiopulmonary resuscitation.

The second stage of palliation is a superior cavopulmonary anastomosis (more commonly the bidirectional Glenn), which facilitates passive pulmonary blood flow from the superior vena cava to the pulmonary arteries and volume unloading of the SV. Completion of palliative single ventricle circulation for many patients includes the Fontan operation, in which all the systemic return (i.e., inferior vena cava) is baffled to the pulmonary circulation, placing the two circulations in series. These patients are preload dependent, meaning that preload determines pulmonary blood flow, pulmonary venous return to the SV, and ultimately systemic cardiac output. Pulmonary blood flow is highly impacted by pulmonary vascular resistance and the transpulmonary gradient. The common atrial pressure can be affected by AVVR, end-diastolic pressure, and atrioventricular dissynchrony. Therefore, hypovolemia, increased pulmonary vascular resistance, elevated common atrial pressure, ventricular dysfunction, and arrhythmias can result in inadequate pulmonary blood flow and ultimately compromise cardiac output, increasing the risk of cardiac arrest. In the prearrest phase, these patients may benefit from inotropic support, afterload reduction, and gentle positive pressure ventilation (when mechanical ventilation is needed) [2, 20, 21]. In particular, ventilatory strategies that target mild hypercarbia and the minimal mean airway pressure necessary to maintain functional residual capacity can be useful to increase cerebral and systemic arterial oxygenation [2, 22]. Negative pressure ventilation has been shown to have some benefit, although it is not available in many centers [23]. Given their in-series circulations, chest compression for both Glenn and Fontan circulations will augment systemic flow but not pulmonary blood flow. This reduction of pulmonary blood flow will limit oxygenation and preload to the SV, resulting in low stroke volume and thus cardiac output with chest compressions. Optimal compression with adequate chest recoil to allow preload and pulmonary blood flow are therefore fundamental (Figure 1), although outcomes often remain poor even when high-quality resuscitation is performed [2, 9, 24].

Right-sided lesions and right ventricle restrictive physiology

Patients undergoing correction of right sided lesions (i.e. pulmonary stenosis, tetralogy of Fallot, double-outlet RV with pulmonary stenosis, truncus arteriosus) are at high risk for RV systolic and diastolic dysfunction, with consequent increased risk of cardiac arrest. Patients at highest risk are mostly those who have been exposed to prolonged abnormal pressure or volume loading conditions, and those with residual lesions. Restrictive RV physiology, characterized by doppler demonstration of persistent antegrade diastolic blood flow into the pulmonary artery in late diastole [25], is frequently seen postoperatively. A poorly compliant RV is associated with elevated RV end diastolic pressure, and systemic venous hypertension. Patients are therefore preload dependent, and lack of adequate preload reduces RV stroke volume, left ventricle (LV) preload, and cardiac output [26, 27]. Alternatively, fluid overload can be deleterious, and especially in the setting of capillary leak related to cardiopulmonary bypass [28]. This population is also at a high risk of developing arrhythmias, such as supraventricular tachycardia, junctional ectopic tachycardia, and ventricular tachycardias; the decrease in cardiac output due to the arrhythmias per se or due to the atrioventricular dissynchrony in a dysfunctional RV can lead to rapid collapse and cardiac arrest. These patients may benefit from gentle fluid administration to maintain preload, low mean airway pressure ventilation strategies [29, 30], preservation of atrioventricular synchrony, initiation of inotropic agents to support RV systolic dysfunction, and administration of pulmonary vasodilators to decrease RV afterload [26]. In patients with RV restrictive physiology, the presence of an atrial septal communication allows right to left shunting which may be helpful to reduce the RV wall stress and preserve LV preload. Additionally, maintaining an open sternum for the first few postoperative days may improve hemodynamics by reducing strain on the RV.

If cardiac arrest occurs, chest compressions may inadequately fill a restrictive RV with diastolic dysfunction or lead to obstruction of the RVOT. Furthermore, chest compressions can exacerbate residual pulmonary regurgitation and worsen cardiac output. High-quality CPR with adequate chest recoil to optimize RV filling and avoid excessive ventilation to maintain low RV afterload are key factors for successful resuscitation in these patients (Figure 1). Patients may also need their chest reopened urgently to enable open chest cardiac massage and improve hemodynamics [2].

Left-sided lesions

Both severe mitral stenosis or regurgitation can cause left atrial hypertension and pulmonary arterial hypertension. In the postoperative period, the reactivity of the previously hypertensive pulmonary bed – exacerbated by the inflammation secondary to the cardiopulmonary bypass - can precipitate pulmonary arterial hypertension crises. Additionally, given the abnormal LV loading conditions, these patients are at higher risk for postoperative low cardiac output syndrome and cardiac arrest [2, 31, 32]. Prevention of arrest includes close monitoring of cardiac output, left atrial pressure, and arrhythmias. Afterload reduction, inotropic support, controlled positive pressure ventilation, and arrhythmia control remain key principles for the management of these patients.

Neonates with critical aortic stenosis present with fixed obstruction and a hypertrophied or dilated LV with decreased contractility, especially when fibroelastosis is present [33, 34]. Elevated LV end diastolic pressure can translate into left atrial hypertension and pulmonary edema. These patients will benefit from a prostaglandin infusion to maintain systemic perfusion and percutaneous transcatheter balloon valvuloplasty. Postintervention monitoring of cardiac output using clinical and laboratory parameters such us heart rate, peripheral perfusion assessment, urine output, and serum lactate, as well as support with inotropes are important [35]. Persistence of low cardiac output syndrome despite adequate treatment of critical stenosis should prompt the assessment of adequacy of the left-sided structure to support a biventricular circulation, with possible conversion to a univentricular approach [33, 34].

Cardiac output generated by chest compressions for severe mitral stenosis or mitral regurgitation is limited by the elevated left atrial pressure and pulmonary vascular resistance, which limit effective pulmonary blood flow and ultimately systemic cardiac output . Additionally, a stiff or dysfunctional LV may impair filling. In cases of significant aortic valve disease, cardiac output generated during chest compressions is reduced due to obstruction. In all these scenarios, high-quality CPR with adequate compression depth and recoil is crucial (Figure 1) and, when ineffective, early consideration of VA-ECMO may offer the best chance of survival.

Myocarditis and cardiomyopathy

Decompensated heart failure secondary to a baseline myocardial disease – such as myocarditis or cardiomyopathy - can often lead to cardiac arrest [2, 36, 37]. In patients with chronic dysfunction, an intercurrent illness or procedural sedation may be enough to further decrease the cardiac output and induce significant clinical deterioration. This can manifest as insufficient cardiac output, arrhythmias, and cardiac arrest. Conversely, patients with fulminant myocarditis may present with relatively preserved systolic function and absence of cardiomegaly on chest radiography but will have rapid deterioration. Although the probability of myocardial recovery is generally good in children with fulminant myocarditis, prevention of cardiac arrest and support of any rapidly evolving myocardial dysfunction is imperative, including early initiation of mechanical circulatory support to allow full myocardial recovery [38]. Careful arrhythmia monitoring, avoidance of electrolyte abnormalities or acidosis, end-organ function monitoring, and careful attention to an excessive or inadequate preload status are fundamental for the prevention of cardiac arrest in these populations. In hypertrophic cardiomyopathy, any reduction of preload should be avoided. In the pre-arrest phase, patients may benefit from inotropic support, and positive pressure ventilation. Finally, VA-ECMO should be considered early in case of refractoriness to conventional medical therapies. [36-38].

Unique challenges in CPR in these patients include the dramatic reduction in coronary blood flow and myocardial perfusion with cardiac arrest in the setting of an already dysfunctional LV, with decreased likelihood of return of spontaneous circulation (ROSC). Additionally, in a patient with hypertrophic cardiomyopathy, adequate ventricular preload and coronary perfusion pressure to support cardiac output and myocardial perfusion may be compromised. Management of tachyarrhythmias can also be extremely challenging in these patients [36, 37]. Once cardiac arrest occurs, VA-ECMO should be considered for early institution [36, 37].

Pulmonary Hypertension

Elvated pulmonary vascular resistance and pulmonary arterial hypertension is an underappreciated cause of morbidity and mortality in children with congenital and acquired heart disease. An estimated 2% to 5% of all pediatric patients following cardiac surgery develop pulmonary arterial hypertension, and up to 5% experience pulmonary hypertensive crisis postoperatively, especially following atrioventricular septal defect repair (14%) and in patients with trisomy 21 (10%) [39-41]. The in-hospital mortality rate among patients with pulmonary arterial hypertension crises is as high as 20% [40]. Pulmonary arterial hypertension crises can be precipitated by various stimuli (pain, anxiety, endotracheal tube suctioning, hypoxia, and acidosis) and can rapidly lead to acute RV failure, acute LV preload reduction, and cardiac arrest. Medical interventions to prevent and treat pulmonary arterial hypertension crises include administration of analgesia, sedation, muscle relaxants, and avoidance or correction of hypoxia, hypercarbia, and metabolic acidosis. The use of inhaled pulmonary vasodilators (e.g. oxygen supplementation and inhaled nitric oxide) and prostacyclin analogs in the acute phase, as well as the use of phosphodiesterase-5 inhibitors (e.g. sildenafil) and endothelin receptor antagonists (e.g. bosentan) chronically, represent the current therapeutic approach. Inotropic and vasoactive pharmacotherapies can be used to support RV function and maintain adequate coronary perfusion pressure to avoid ischemia secondary to systemic hypotension. In severe refractory cases to medical therapy, pulmonary artery decompression procedures, such as an atrial septostomy or placement of a Potts shunt, can significantly reduce the risk of pulmonary arterial hypertension crises [39, 40, 42, 43].

Conventional resuscitation and medications are often unsuccessful at restoring pulmonary blood flow, and thus cardiac output. Correcting possible triggers for pulmonary arterial hypertension crises and avoidance of hyperventilation are crucial. In cases refractory to high-quality CPR, rapid mobilization of ECMO may allow for the best chance of survival by providing a bridge to heart/lung transplantation or spontaneous recovery of RV function.

Arrhythmias

Patients with cardiac disease are at higher risk of conduction anomalies and arrhythmias compared to the general population [2]. Children with CHD after cardiac surgery are at heightened risk of complete heart block, which is generally not well tolerated especially in the setting of low cardiac output syndrome. Temporary pacing is used to restore atrioventricular synchrony in some cases, and implantation of permanent pacemaker is considered for patients with heart block who fail to recover sinus rhythm within the first postoperative week to 10 days [44]. Supraventricular tachycardia (SVT) is the most common tachyarrhythmia in children, and it may be even more common in patients with CHD given presence of atrial suture lines or accessory pathways [45]. Prolonged episodes of SVT can cause deterioration of cardiac function, with onset of congestive heart failure and collapse. Synchronized cardioversion or adenosine should be rapidly considered in hemodynamically unstable patients. Adenosine has been shown to be an effective therapy for SVT; other treatments for SVT refractory to adenosine include procainamide, esmolol, or amiodarone [45, 46]. Another arrhythmia with the potential to induce rapid circulatory failure and cardiac arrest is junctional ectopic tachycardia (JET), an automatic rhythm that originates from the atrioventricular node or high in the His-Purkinje system most commonly observed in the early postoperative period (up to 8% of children after cardiac surgery) [47-49]. Overdrive pacing, anti-arrhythmic therapies such as procainamide, and strategies to decrease oxygen demand using sedation and analgesia, and temperature control have been demonstrated to be effective treatments [50]. Ventricular arrhythmias including torsade de pointes in patients with long QT syndrome, ventricular tachycardia (VT) and ventricular fibrillation (VF) are rare in children. Initial resuscitation is targeted to eitiology, including magnesium sulfate for torsade de pointes, lidocaine or amiodarone for monomorphic VT, while direct current cardioversion should be considered early for pre-arrest polymorphic VT [2, 4]. Finally, lidocaine can be considered in pediatric patients with VF/pulseless VT in-hospital cardiac arrest, which seems to be associated with an increased likelihood of ROSC [51].

ECMO and ECPR

Venoarterial ECMO can provide mechanical circulatory support during and after resuscitation in children who experienced cardiac arrest refractory to conventional medical therapies or CPR. The use of ECMO in the form of ECPR is becoming frequent, both in surgical and medical cardiac patients, specifically for in-hospital cardiac arrest [2, 52-54]. Recent pooled data showed that in centers with ECMO expertise, 22% (95%CI: 14-33%) of pediatric patients with cardiac disease underwent ECPR [5]. In a nationwide study comparing cardiac patients who did or did not undergo ECPR, Lasa et al. demonstrated that patients receiving ECPR had higher odds of survival to discharge (OR 2.80; CI 2.13-3.69) and survival with favorable neurological outcome (OR 2.64; CI 1.91-3.64) than patients who received CPR only [55]. This association persisted when analyzed by propensity score-matched cohorts [55]. Overall, the use of extracorporeal strategies to support CPR has been rapidly adopted in many centers and its utilization is growing consistently as the ECMO expertise increases worldwide.

Mortality after cardiac arrest and neurologic outcome

Despite an overall improvement of the survival rate after in-hospital cardiac arrest in the general pediatric population in the last decade (3-fold improvement), the mortality rate for patients with cardiac disease who experienced cardiac arrest remains high (30% to 65%) [2, 9, 64, 56-63]. Among studies on children admitted to an ICU who experienced a cardiac arrest, the pooled in-hospital mortality was recently calculated at 51% (95%CI: 42-59%) [5]. Mortality also remains high among patients rescued with ECPR, with reported rates from 44% to 65% [63, 65-70]. Table 2 summarizes studies addressing the mortality rates and risk factors for mortality in pediatric cardiac patients following cardiac arrest. Main reported risk factors are univentricular physiology, comorbidities, higher vasoactive-inotropic score, longer CPR, cardiac arrest during the weekend, limited nurse experience, and – in post-surgical patients - surgical complexity (Society of Thoracic Surgeons [STS] mortality category 4-5, or highest Risk Adjustment for Congenital Heart Surgery-1 [RACHS-1] score). Admission to a CICU was reported to decrease the risk of mortality [8, 9, 72-74, 10, 58-60, 62-64, 71].

Table 2:

Studies addressing risk factors for mortality after cardiac arrest in pediatric patients with cardiac disease according to setting.

Author, year	Study design, Setting and period	Population	Exclusion criteria	Definition of cardiac arrest	N patients with CA/N total (%)	Outcome measures		Predictors/ risk factor for CA associated Mortality
						Short term mortality No. (%)	Late mortality No. (%)
Hospitalized patients
Gupta P et al., 2016 PCCM	Retrospective analysis of prospective data, GWTG-R (AHA), Multicenter (157 USA Centers) 2000-2010	Hospitalized cardiac patients with at least 1 episode of CA	DNR, out-of-hospital CA, newborns in delivery room, NICU patients, CA resolved with implanted defibrillator	Pulseless in-hospital CA requiring chest compressions >1 min	1889/1889 (100)	No-ROSC**: 563 (29.8) (362 deaths, 201 EPCR) At 24h: 739 (39.1) At discharge: 929 (49.2%)	NA	Univariate analysis: Recurrent arrest (p<0.001) Multivariable model: NS
Lowry et al., 2013	Retrospective analysis of prospective data, KID Registry, Multicenter (38 USA States) 2000,2003,2006	Hospitalized cardiac patients	Not defined	ICD-9 procedure code	3709/498610 (0.7)	At discharge 2083 (56.2)	NA	Multivariable model: RISK: SV OR 1.7 (1.2-2.6). PROTECTIVE: Age<1 yr OR 0.7 (0.6-0.9), cardiac surgery prior to CPR OR 0.6 (0.5-0.8)
Gupta et al., 2014 Resuscitation	Retrospective analysis of prospective data, VPS (NACHIRI) Registry, Multicenter (108 USA Centers), 2009-2013	Hospitalized cardiac patients with at least 1 episode of CA	Patients with “altered code status”	Any event characterized by either pulselessness or critically compromised perfusion treated with external chest compression and/or defibrillation	2182/ 2182 (100)	At discharge: 34 (95%CI 27-44) per 100 cardiac admissions	NA	Multivariable model: NS
Ortmann et al., 2011	Retrospective analysis of prospective data, GWTG-R (AHA), Multicenter (265 USA Centers), 2000-2008 Overlap data	Hospitalized cardiac patients with at least 1 episode of CA (medical and surgical)	DNR, out-of-hospital, NICU, newborn in the delivery room, obstetrics patient, shock by an implanted defibrillator	Cessation of cardiac mechanical activity with the absence of palpable central pulse, apnea, and unresponsiveness	1214/ 1214 (100)	No-ROSC** 632 (52.1) (481 death, 151 ECPR) At 24h 594 (48.9) At discharge 821 (67.6)	NA	Multivariable model (Outcome: survival): RISK for cardiac-surgical: renal failure OR 0.1 (CI 0.03, 0.3), heart failure OR 0.5 (CI 0.3, 0.8), beds n<300 OR 0.4 (CI 0.2, 0.9), teaching hospital, OR 0.3 (CI 0.09, 0.8), longer CPR, OR 0.6 (CI 0.5, 0.7). PROTECTIVE for cardiac-surgical: Age 1month-1year OR 2.7 (CI 1.6, 4.4), Age 1year-8year , OR 2.6 (CI 1.4, 4.9), ECPR OR 2.5 (CI 1.3, 4.5). RISK for cardiac medical: CA in the Emergency Department, OR 0.3 (CI 0.1, 0.6), metabolic/electrolyte abnormality OR 0.4 (CI 0.1, 0.96), atropine OR 0.4 (CI 0.2, 0.7), longer CPR duration OR 0.7 (CI 0.6, 0.8). PROTECTIVE for cardiac-medical: arrhythmia OR 2.6 (CI 1.5, 4.3), airways compromise OR 8 (CI 2.5, 26), ECPR OR 3.8 (CI 1.4, 5.8).
Ramamoorthy et al., 2010	Retrospective analysis of prospective data, POCA Registry, Multicenter (79 USA Centers), 1994-2005	Hospitalized cardiac patients with at least 1 episode of anesthesia related CA	CAs in the pediatric and neonatal intensive care units or on the ward	Administration of chest compressions or death	127/ 127 (100)	CA-related death (no time defined): 42 (33.1) ECPR 7/68 – no outcome defined	NA	Univariate analysis: Unrepaired lesions compared with palliated or completely repaired lesions (p=0.006), longer total duration of resuscitation (p=0.001), larger number of drugs (p=0.046), and more rounds of drugs (p=0.038)
Intensive care unit setting
Dagan et al., 12 2019	Retrospective, Single-center (Melbourne, Australia), 2007-2016	P-CICU patients post cardiac surgery	Children with medical cardiac conditions, children who suffered CA following procedures as cardiac catheterization, CA prior to cardiac surgery, DNR	Cessation of cardiac mechanical activity requiring cardiac massage for ≥1 min	211/ 4983 (4.3)	At discharge 64 (30.1)	NA	Univariate analysis: Younger age (p<0.001), male (p=0.001), lower weight (p<0.001), prematurity (p<0.001), chromosomal/genetic syndrome (p<0.001), need for ECMO/VAD (p<0.001), higher RACHS-1 category (p<0.001)
Dhillon et al., 17 2018	Retrospective, Single-center (Texas, USA), 2011-2016	P-CICU patients who experienced at least 1 CA	Multiple events in the same patient, events with incomplete documentation, CA outside the CICU	CPR ≥ 2 min	90 (of 150 events over 5,947 unique admissions) (150/5,947=2.5%)	No-ROSC** 41 (46.0) (18 deaths, 23 ECPR) At 24h 25 (27.8) At discharge 49 (54.4)	NA	Univariate analysis: No epinephrine infusion pre-CA (p=0.02 for CHD medical patients, p=0.03 for surgical patients), no arterial line pre-CA (p=0.02 for surgical patients), longer CA duration (p=0.02 for surgical patients), higher number of epinephrine doses (p<0.01 for surgical patients)
Yates et al., 16 2019	Prospective, Multicenter (PICqCPR study, USA centers, CPCCRN network), 2013-2016	PICU or P-CICU patients (medical and surgical) with invasive arterial BP monitoring line prior and during CPR	Patients for which first compression was not captured on the waveform data, or compression start and stop could not be determined	CPR for at least 1 min	113/ 113 (100)	No ROSC** 72 (63.7) (39 deaths, 33 ECPR) At discharge 56 (49.6)	NA	Univariate analysis: Diastolic BP ≥25 mmHg for infants or ≥30 mmHg for children (cohort surgical patients only, p=0.018)
Gupta et al., 10 2016 Resuscitation	Retrospective analysis of prospective data, VPS (NACHRI) Registry, Multicenter (62 USA Centers) 2009-2014	P-CICU patients with CHD post cardiac surgery	ICU readmission, lack of surgical documentation, surgical closure of isolated PDA or surgery not listed in STS-EACTS	Any event characterized by either pulselessness or critically compromised perfusion treated with external chest compression and/or defibrillation	736/ 26,909 (2.7)	At discharge 229 (31.1)	NA	Multivariable model: RISK: ECMO OR 3.04 (CI 2.02, 4.57), SV anatomy OR 1.60 (CI 1.04, 2.46), renal failure OR 2.78 (CI 1.70, 4.54), brain hemorrhage OR 3.09 (CI 1.10, 8.62), hemodialysis catheter in place OR 3.42 (CI 1.05, 11.15). PROTECTIVE: younger age (<28days) OR 0.47 (CI 0.28, 0.81), presence of Cardiac PICU OR 0.48 (CI 0.25, 0.92)
Gupta et al., 8 2014 Ann Thorac Surg	Retrospective analysis of prospective data, STS-CHSD Registry, Multicenter (97 USA Centers), 2007-2012	P-CICU patients with CHD post cardiac surgery	Surgery not classified into one of the STS-EACTS Mortality Categories, missing outcome	Cessation of effective cardiac mechanical function	1843/ 70,270 (2.6)	At discharge 910 (49.4)	NA	Multivariable model: Low volume centers (<150 case/y) OR 2.0 (1.52-2.63), low-medium volume centers (150-250 case/y) OR 1.39 (1.09-1.77), STS mortality category 1-3 in low and in medium volume centers (OR 2.29 (1.19-4.41) and 1.88 (1.12-3.18)); STS mortality category 4-5 in low and medium-low volume centers (OR 2.0 (1.37-2.9) and 1.41 (1.03-1.94)).
Ahmadi et al., 40 2013*	Single-center (Tehran, Iran), 2001-2002	P-CICU patients <7 years of age, post cardiac surgery	Not defined	Not defined	59/529 (11.2)	At discharge 37 (62.7)	NA	Univariate analysis: Lower mean arterial BP before the CA (p=0.04)
Gaies et al., 7 2012	Retrospective, Single-center (Ann Arbor, USA) 2006-2008	P-CICU patients with at least 1 episode of CA	Not defined	Event requiring active chest compressions for any duration	102 (of 2,230 P-CICU admission) (4.6)	No-ROSC ** 27 (16.5) (17 death, 10 ECPR) At discharge 53 (52.0)	NA	Multivariable model: Arrest during weekend OR 4.4 (1.2-15.5), experience of primary nurse <1yr OR 9.4 (1.6-55.0), VIS>=20 OR 6.4 (1.8-22.9)
Hansen et al., 43 2011	Case-control, Single-center (Edmonton, Canada) 1996-2005	NICU patients post cardiac surgery with CPB, ≤6 weeks of age. Cases: at least 1 CPR event, Controls: no CPR events	Cardiac surgery not requiring CPB, patients having CPR preoperatively or in the operating room	Not defined	29 CA (cases) (of 343 patients post cardiac surgery) (8.5)	No-ROSC ** 17 (8 death, 9 ECPR) (58.6)	At 1 month 11 (37.9) At 2 years 17 (58.6)	Multivariable model on all cohort, not on patients with CA only: Minutes of chest compression OR 1.04 (CI 1.01, 1.06)
Parra et al., 2 2000	Retrospective, Single-center (Miami, USA), 1995-1997	P-CICU patients with at least 1 episode of CA	DNR patients	Cessation of circulation and respiration that required CPR for>2 mins	32 (38 events) / 32 (100)	No-ROSC** 18 (56.2) (14 deaths, 4 ECPR) At discharge: 18 (56.2)	At 6 months: 21 (65.6)	Univariate analysis: NS
Rhodes et al., 1 1999	Retrospective, Single-center (New York, USA), 1994-1998	P-CICU patients with CHD and age <12months post-cardiac surgery	Not defined	Chest compressions or the absence of a palpable spontaneous pulse that was not resolved with only airway intervention	34/ 575 (5.9)	No-ROSC 11 (32.4) At discharge: 20 (58.8)	At 6 months 20 (58.8) At follow-up (median 21 months) 21 (61.8)	Univariate analysis: Lower pre-arrest MAP (p=0.0003), Lower arterial pH (p<0.02), Higher epinephrine doses (p<0.001), Higher bicarbonate dose (p=0.005), Longer CPR duration p<0.001)

^**for the purpose of this table, ECPR was considered as no return to spontaneous circulation (no-ROSC).

BP: blood pressure; CA: cardiac arrest; CHD: congenital heart disease; CPB: cardiopulmonary bypass; CPR: cardiopulmonary resuscitation; DNR: do not resuscitate; ECPR: ECMO- cardiopulmonary resuscitation; MAP: mean arterial pressure; MV: mechanical ventilation; NA: not applicable; NS: not significant; NICU: neonatal intensive care unit; OR: odds ratio; PAH: pulmonary arterial hypertension; P-CICU: pediatric cardiac intensive care unit; PICU: pediatric intensive care unit; Preop.: preoperative; Postop.: postoperative; RACHS-1: Risk Adjustment for Congenital Heart Surgery-1; ROSC: return of spontaneous circulation; STS: Society of Thoracic Surgeon; SV: single ventricle; VAD: ventricular assist device; VIS: vasopressor inotropic score.

Neurologic outcome after cardiac arrest varies widely among studies. A favorable neurologic outcome at discharge, generally defined as a pediatric cerebral performance category (PCPC) of 1-3 [75], has been reported in a variable range from 64% to 95% of survivors [55, 76-79]. Neurologic dysfunction following ECPR remains high: among survivors of ECPR in the THAPCA trial, about 29% of children <6 years of age experienced persistent severe cognitive deficits and 40% experienced at least moderate neurologic injury at 12 months [76], with similar results reported from other studies [63, 65-67, 77, 80].

Research gaps and future directions

Research gaps in the field of resuscitation in children with acquired and congenital heart disease are numerous. As highlighted in this review, the neonatal and pediatric cardiac population have unique risk factors, for which specific resuscitation strategies are crucial. A recent large multicenter collaborative study has shown that a low-technology cardiac arrest prevention bundle consisting in a structured risk assessment performed by the care team by physiology was effective in reducing the aggregate cardiac arrest rate by 30% compared with control hospitals [81]. Education efforts and further quality improvement studies are needed worldwide to confirm these results and improve outcomes. Researchers are also investigating dedicated resuscitation techniques for SV patients, who have the highest risk of cardiac arrest [82, 83]; efforts should continue in that direction. Overall, specific resuscitation strategies – with respect to chest compressions, ventilation strategies, and pharmacologic approaches - should be investigated for children with different physiologies. Further studies are also needed to investigate the role of specific drugs for which the level of evidence is low, such as bicarbonate and calcium [84, 85]. Collaborative studies and consensus are needed to better define, predict, and manage low cardiac output syndrome. Furthermore, given that current knowledge on pulmonary arterial hypertension, myocarditis, and cardiomyopathies are still mainly based on adult data, dedicated pediatric studies are warranted. Finally, researchers should continue to investigate the role of ECMO as resuscitation strategy, particularly regarding the quality of CPR prior to and during cannulation, timing of deployment, and type of circuits used in this setting.

Conclusions

Children with cardiac disease are at higher risk of cardiac arrest as compared to healthy children. Main risk factors include neonatal age, genetic syndrome, SV physiology, arrhythmias, pulmonary arterial hypertension, comorbidities, ECMO support before cardiac arrest, and recent complex cardiac surgery. Preventive and resuscitation strategies strictly depend on physiology. Outcomes following cardiac arrest in this population remain poor and prevention is pivotal. Knowledge on physiology-based resuscitation strategies and its importance should be widely spread and its impact investigated in dedicated studies. Collaborative and interprofessional studies are needed to address research the substantial knowledge gaps in this field to improve care and outcomes for children with cardiac disease.

What is known/what is new.

What is known

• Children with cardiac disease are at high risk of cardiac arrest, and cardiopulmonary resuscitation may be challenging due to unique characteristics and different physiologies.

• Mortality after cardiac arrest remains high and neurologic outcomes suboptimal.

What is new

• We reviewed the unique resuscitation challenges, current knowledge, and recommendations for different cardiac physiologies.

• We highlighted knowledge gaps to guide research efforts aimed to improve care and outcomes in this high-risk population.

Abbreviations list:

CHD

congenital heart disease

CI

confidence interval

CPR

cardiopulmonary resuscitation

ECMO

extracorporeal membrane oxygenation

ECPR

extracorporeal cardiopulmonary resuscitation

ICU

intensive care unit

LV

left ventricle

OR

odds ratio

Qp

pulmonary flow

Qs

systemic flow

ROSC

return to spontaneous circulation

RV

right ventricle

SV

single ventricle