Advances in medical management, surgical techniques, and mechanical circulatory support (MCS) for pediatric patients with congestive heart failure continue to improve patient outcomes.
Indications for the use of MCS continue to evolve.
Patient size, expected duration of support, and goals of support (i.e., bridge to recovery vs. bridge to transplant) must be considered in the choice of an MCS device.
MCS is a lifesaving therapeutic option for patients with advanced heart failure. Device options in children are expanding but remain limited due to size constraints.
In children, most short-term MCS continues to be achieved with extracorporeal life support.
Small patients (<3 kg), renal failure, and the need for biventricular support greatly increase the risk of death during support.
A multidisciplinary team approach to the selection of the appropriate mechanical circulatory device is critical to a successful outcome.
Pediatric heart failure
Heart failure (HF) is the final pathway of many pathophysiologic states affecting cardiovascular performance leading to inadequate cardiac output (CO) and decreased end-organ perfusion with diminished oxygen delivery to the vital organs. These include states of altered preload, afterload, contractility, and abnormal heart rate (HR) or rhythm. The pathophysiologic syndrome of HF is the result of a complex interplay among circulatory, neurohormonal, and molecular abnormalities. The etiology of HF in children differs from that of the adult and includes both primary cardiac and noncardiac causes, with the largest disease burden being due to congenital heart malformations and cardiomyopathies. HF in children leads to characteristic signs and symptoms—such as poor growth, feeding difficulties, respiratory distress, exercise intolerance, and fatigue. The pathophysiology of HF includes volume overload (e.g., left-to-right shunting, semilunar valve regurgitation), pressure overload (e.g., congenital aortic stenosis), or a combination of both as seen in patients with complex congenital heart disease.
The primary goal in the management of HF is to ensure adequate tissue oxygen delivery, as an imbalance between delivery and consumption results in organ dysfunction, failure, and cardiogenic shock. Acute decompensated HF (ADHF) is a common final pathway for children with congenital or acquired heart disease. They can present with failed palliation of congenital heart disease, acquired cardiomyopathies, or acute exacerbations of chronic HF. Initial therapies to treat ADHF target providing respiratory support; decreasing metabolic demands (i.e., work of breathing); optimizing preload, afterload, and contractility; and optimizing HR and rhythm. Despite advances in pharmacologic therapies beyond inotropes, diuretics, and antiarrhythmic agents, a group of patients will continue to deteriorate or require excessive cardiopulmonary support to maintain adequate CO. This group should be evaluated early for mechanical circulatory support (MCS) before additional end-organ dysfunction or irreversible damage occurs. Several innovative types of MCS are being created and brought into clinical care while extended indications and management strategies are currently being developed for older devices. Simultaneously, extracorporeal life support (ECLS) continues to improve and remains the most commonly deployed form of MCS for infants and children.
Low cardiac output syndrome
Low cardiac output syndrome (LCOS) describes a clinical state with a specific profile of biochemical markers in which there is inadequate systemic oxygen delivery (D o 2 ) to meet the metabolic demands (oxygen consumption [V o 2 ]) of the patient. The condition has been recognized since the 1960s; numerous studies have documented the predicted changes in physiologic parameters. , LCOS has been an active area of research, and multiple reviews of current therapies are available to guide the intensivist. LCOS is frequently seen in myocarditis, cardiomyopathies, with prolonged bradyarrhythmias or tachyarrhythmias, and commonly after complex pediatric cardiac surgery. Postoperative physiologic changes secondary to cardiopulmonary bypass, myocardial ischemia during aortic cross-clamping, cardioplegia, residual uncorrected lesions (i.e., aortic arch obstruction), ventriculotomy, changes in the loading conditions to the myocardium, and dysrhythmias may all contribute to the development of LCOS. A variety of proinflammatory triggers are activated during cardiopulmonary bypass as a result of blood contact with foreign surfaces, ischemia, reperfusion, oxygen free radicals, tissue trauma, and temperature fluctuations. These complex inflammatory responses include complement activation, cytokine release, leukocyte and platelet activation, and the expression of adhesion molecules. LCOS has been reported to affect up to 25% of infants and children following cardiac surgery, typically occurs between 6 and 18 hours after cardiac surgery, and results in a longer intensive care and hospital stay with increased morbidity and mortality. It is associated with elevated systemic vascular resistance and pulmonary vascular resistance (PVR), impaired myocardial function, dysrhythmias, and capillary leak. When unrecognized or inadequately treated, LCOS can result in irreversible end-organ failure, cardiac arrest, and death. Prevention, early recognition, and optimal treatment are essential to ameliorate or reverse its course.
The concepts of D o 2 and V o 2 are of fundamental importance in managing critically ill children. Systemic oxygen delivery (D o 2 ) is defined as the amount of O 2 delivered to peripheral tissues each minute and is determined by CO and the oxygen content of arterial blood (Ca o 2 ). CO is defined as the product of HR and stroke volume (SV), and the arterial oxygen content (Ca o 2 ) is determined by hemoglobin (Hb) concentration, the arterial oxygen saturation (Sa o 2 ), and the partial pressure of oxygen in the arterial blood (Pa o 2 ). Lastly, V o 2 can be measured by indirect calorimetry, or calculated using the Fick equation as the arterial venous difference in oxygen content (Ca o 2 − Cv o 2 ) multiplied by the CO:
Achieving a positive balance between O 2 supply and demand is essential and can be accomplished by decreasing V o 2 or increasing D o 2 . V o 2 is determined by tissue metabolism and is increased during periods of increased muscular activity (i.e., seizures, exercise), infection, fever, or with increased levels of circulating catecholamines. Normally, the ratio between D o 2 and V o 2 is 4:1; this ratio is high enough that cellular respiration is not supply dependent and V o 2 is mainly a function of tissue O 2 demands. When V o 2 increases based on increased metabolic demands, D o 2 increases accordingly. Therefore, consumption drives delivery. Inability to maintain the normal D o 2 :V o 2 ratio is initially compensated by increased O 2 extraction. However, when the rate of V o 2 exceeds D o 2 , anaerobic metabolism begins.
Early studies in neonates with transposition of the great arteries (TGA) who underwent the arterial switch procedure documented the development of low CO 6 to 12 hours after surgery. These findings led to a placebo-controlled intervention study whereby milrinone was randomly administered (bolus followed by continuous infusion) in a low (0.25 μg/kg per minute) versus high (0.75 μg/kg per minute) dose to infants and children after cardiopulmonary bypass (CPB), in addition to the required inotropic agents needed to separate from CPB. In those patients who received high-dose prophylactic milrinone, the development of LCOS in the first 36 hours after surgery was decreased to 12% compared with 26% in the placebo group. Recent Cochrane reviews have failed to demonstrate a consistent improvement in the treatment of cardiogenic shock in infants and children after cardiac surgery with either milrinone or levosimendan, a calcium-sensitizing agent. ,
Although a complete assessment of D o 2 and V o 2 in critically ill infants and children is extremely challenging, numerous hemodynamic and biochemical biomarkers can be readily obtained to help guide the bedside clinician. Mixed venous oxygen saturation (SvO 2 ) or central venous oxygen saturation (Scv o 2 ), arterial lactate, and near-infrared spectroscopy (NIRS) are important clinical parameters that can be serially/continuously measured in patients at risk for LCOS.
Sv o 2 and Scv o 2 are a reflection of the total body D o 2 /V o 2 ratio and can be important metrics to follow in critically ill patients. , An elevated lactate level on admission (>4.5 mmol/L) and rising at 0.9 mmol/L per hour postoperatively is associated with major adverse events, including death, in infants after cardiac surgery. Furthermore, lower Sv o 2 may increase the predictive power of elevated arterial lactate levels for mortality after pediatric cardiac surgery. ,
It is not uncommon to observe a disparity between direct measurements of D o 2 and CO and estimates based on physical examination and interpretation of conventional laboratory and hemodynamic parameters. Compensatory increases in systemic vascular resistance maintain arterial blood pressure as CO decreases and central venous pressures may not correlate well with ventricular filling (preload). , Femoral venous catheters are routinely placed in neonates and children, providing an additional factor (intraabdominal pressure) in the accurate interpretation of these continuous variables. The effects of positive-pressure ventilation on these measurements are well described. ,
Hemodynamic stability in the early postoperative period after Norwood palliation depends on an adequate total CO from the single ventricle and a balanced pulmonary-to-systemic blood flow ratio (Qp:Qs). Systemic oxygen saturation (Sa o 2 ) alone is a poor predictor of Qp:Qs because of variability in systemic venous oxygen saturation (Sv o 2 ) and pulmonary vein saturation (Sp vo 2 ) after the Norwood operation. Monitoring and optimizing Sv o 2 have been shown to improve outcomes in pediatric patients at risk for developing shock, including hypoplastic left heart syndrome. , Intermittent Sv o 2 monitoring can be obtained to assess oxygen transport balance, but the presence of intracardiac shunts (i.e., left-to-right shunting of pulmonary venous blood across the atrial septum) near the site of sampling and the need for repeated blood sampling limits its universal use. It has been reported that when Sv o 2 monitoring is used in the postoperative care of critically ill neonates, significantly fewer adverse events are encountered. Significant decreases in Sv o 2 can occur without appreciable changes in Sa o 2 , blood pressure, or HR. When low Sv o 2 is recognized, increased inodilator support and measures that decrease metabolic demands (i.e., sedation, neuromuscular blockade) often successfully return this metric toward the normal range. In contrast, ventilator and inspired gas adjustments have less effect on correcting Sv o 2 . The critical O 2 extraction ratio is defined by the onset of shock and ranges from 50% to 60%. ,
NIRS is a noninvasive technique used to monitor tissue oxygenation and perfusion. NIRS was initially used in the operating room to monitor brain oxygenation during cardiopulmonary bypass but has expanded as a useful and frequent monitoring technique in the intensive care unit (ICU). Cerebral NIRS noninvasively assesses cerebral tissue oxygenation and relies on the relative lucency of biological tissue to near-infrared light where oxy- and deoxyhemoglobin have distinct absorption spectra. , The oximeter monitors the nonpulsatile signal reflecting the microcirculation where 75% to 85% of the blood volume is venous. Thus, the NIRS-derived oxygen saturation is an indicator of oxygen extraction for the region of brain beneath the optode. There is a good correlation between cerebral oxygen saturations, jugular bulb, and superior vena cava saturations. , Cerebral oxygen saturations are closely and inversely correlated with the oxygen extraction ratio in neonates following the Norwood procedure. Strategies measuring NIRS in both cerebral and somatic regions provide better estimates of outcome. , ,
The pulse contour CO (Pulse index Continuous Cardiac Output [PiCCO]) measurement system allows continuous hemodynamic monitoring using a large (femoral or axillary) artery catheter and a central venous catheter. This technology uses intermittent transpulmonary thermodilution and pulse contour analysis. With the use of specific algorithms, various parameters such as CO, extravascular lung water (EVLW), global end-diastolic volume (GEDV), pulse pressure variation (PPV), and stroke volume variation (SVV) can be obtained. PiCCO can assist in preload assessment in a volumetric manner. The PiCCO system may give inaccurate measurements in patients with arrhythmias, rapid temperature changes, intra/extracardiac shunts, aortic aneurysm, aortic stenosis, pneumonectomy, pulmonary embolism, and extracorporeal circulation. , ,
The USCOM ultrasonic CO monitor (USCOM Pty Ltd.) is a noninvasive device that determines intermittent CO by continuous-wave Doppler ultrasound and estimates of the normal aortic valve annulus based on the weight or height of the patient. Systemic vascular resistance can then be calculated when an invasive or noninvasive blood pressure is obtained simultaneously. Although approved by the US Food and Drug Administration (FDA) for use in children, pediatric data are still limited.
Specific treatments to improve cardiac function
HF is a clinical and pathophysiologic syndrome that results from ventricular dysfunction and volume or pressure overload, either alone or in combination. In contrast to adult patients, tremendous heterogeneity exists between the etiology and pathophysiology of HF in pediatric patients. There are significant barriers in applying adult HF data to children owing to factors related to developmental cardiac physiology. Despite these differences, treatment strategies for pediatric patients have often followed the recommendations from large, randomized multicenter trials in adult patients or “consensus” opinions based on single “best” institution clinical practices due to a lack of randomized, multicenter pediatric clinical trials. With the expanded authority and mandates of the FDA Pediatric Advisory Committee and the origins of the National Institutes of Health Collaborative Pediatric Critical Care Research Network and the new Pediatric Cardiac Critical Care Consortium (PC ), the future is brighter for high-quality clinical trials in children. ,
Management of chronic congestive HF with digitalis formed the basis of early therapy (1970s) despite its narrow therapeutic index for safety. Volume overload is common, and loop and/or thiazide diuretics ± the addition of the potassium-sparing antimineralocorticoid, spironolactone are still also commonly used today, although drug compounding for children and potassium monitoring is problematic. Patients with advanced pulmonary hypertension, right ventricular (RV) failure, or restrictive left ventricle (LV) physiology present a unique challenge, as CO may worsen in face of the rapid reduction in LVSV that can occur with aggressive diuresis. Although angiotensin-converting enzyme (ACE) inhibitors exert favorable effects on cardiac remodeling and survival in adults with congestive HF, their role in children is less clear as randomized placebo-controlled trials in children evaluating the effect of ACE inhibitors in single-ventricle patients have failed to demonstrate a beneficial effect.
Selective β-blockers, such as metoprolol, are a mainstay in adult cardiac patients, with some utility in children. , The nonselective β-blocker/α 1 -blocker, carvedilol, has been used commonly in pediatric HF treatment in recent years. Most pediatric uses of these drugs have been extrapolated from the positive effects seen in adult trials that demonstrated a reduction in mortality and risk of hospitalization. Single-center trials have demonstrated both improved ejection fraction with the use of carvedilol in children awaiting heart transplantation with dilated cardiomyopathy and a delayed time to transplantation or death. , However, the most recent and largest multicenter, randomized, double-blind placebo-controlled trial of carvedilol in children and adolescents with symptomatic systolic HF did not demonstrate an improved survival benefit. A recent Cochrane systematic review of randomized controlled trials investigating the effect of β-blockers in pediatric HF concluded that there are not enough data to recommend or discourage the use of these drugs in pediatric HF. ,
A current review from the PC database found that 6% of pediatric patients admitted to a cardiac intensive care unit (CICU) were diagnosed and treated for ADHF and that these patients had multiple comorbidities, high mortality rates, and frequent readmissions, particularly those with congenital heart disease (CHD). In this multicenter data collection, the median age at admission was 0.93 years and 57% of the cohort had CHD. A total of 88% received vasoactive infusions while 59% required mechanical ventilation. Common complications were arrhythmias (19%), cardiac arrest (10%), sepsis (7%), and acute renal failure requiring dialysis (3%). The median length of stay was 7.9 days and the readmission rate was 22%. Overall, CICU mortality was 15% but 19% in those with CHD, compared with 11% in those without CHD. Independent risk factors associated with CICU mortality included age less than 30 days, CHD, vasoactive infusions, ventricular tachycardia, mechanical ventilation, sepsis, pulmonary hypertension, extracorporeal membrane oxygenation (ECMO), and cardiac arrest.
For hospitalized children with ADHF and/or LCOS, traditional first-line management is to use continuous infusions of catecholamines or their analogs (i.e., dopamine, epinephrine, norepinephrine, or dobutamine) to increase cardiac contractility (β 1 ). However, when used in high doses these agents can be deleterious to global and myocardial V o 2 , induce myocardial cell apoptosis, and lead to increased mortality. In these patients, it is important to integrate diuretic, inotropic, and vasodilating therapy in conjunction with careful monitoring of hemodynamic parameters and end-organ perfusion.
Milrinone, a phospodiesterase-3 inhibitor, has emerged as an important inodilating agent, which is now widely used in children following open-heart surgery after a landmark study showed that the prophylactic use of milrinone was associated with a decreased likelihood of LCOS in children following open-heart surgery in a dose-dependent fashion. This benefit is thought to result from both improved myocardial contractility and pulmonary and systemic vasodilatory effects. Milrinone reduces RV and LV afterload through systemic and pulmonary vasodilation and improves diastolic relaxation (lusitropy) of the myocardium through its enhanced cyclic adenosine monophosphate–dependent diastolic reuptake of calcium.
Sildenafil, an oral or intravenous drug that inhibits cyclic guanosine monophosphate–specific phosphodiesterase type 5 causing smooth muscle relaxation, has proven to be beneficial in neonates with severe pulmonary hypertension and as an adjunct in weaning these patients from inhaled nitric oxide.
Levosimendan, a newer drug, is a calcium sensitizer that enhances the contractility of the ventricle by binding to cardiac troponin C. In addition, the drug acts as a vasodilator by opening ATP-sensitive potassium channels in the vascular smooth muscle, resulting in decreases in systemic and pulmonary vascular resistance. The improvement in myocardial performance is accomplished without an increase in intracellular calcium, thus providing a cardioprotective effect. Finally, an active metabolite with a half-life of ≈3 days prolongs the duration of action for this continuously infused medication. , A recent prospective observational study in 110 patients with a median age of 346.5 days (4 days–19.6 years) undergoing cardiac surgery reported the safety of levosimendan in all age groups and categories of congenital heart disease, demonstrating optimization of CO with a low incidence of arrhythmias. Levosimendan was started at the initiation of the rewarming phase of bypass and continued for 48 hours. Despite these encouraging studies, a Cochrane review in 2017 concluded that the “current level of evidence is insufficient to judge whether prophylactic levosimendan prevents LCOS and mortality in pediatric patients undergoing surgery for congenital heart disease.” The authors concluded that “no significant differences have been detected between levosimendan and standard inotrope treatments in this setting.”
Fenoldopam, a selective dopamine-1 receptor partial agonist, causes systemic vasodilation and increased renal blood flow with improved renal function in adults. Many studies have examined fenoldopam as a possible therapeutic agent capable of preventing the onset of postoperative acute kidney injury (AKI). Two most recent meta-analyses concluded that although fenoldopam decreases the incidence of postoperative AKI, it does not reduce the need for renal replacement therapy or mortality. , Data on the use of fenoldopam in pediatric patients are sparse, with one retrospective study demonstrating a significant improvement in diuresis in neonates after CPB with the addition of fenoldopam to conventional diuretic therapy, whereas another prospective trial failed to demonstrate a beneficial effect. Ricci et al. demonstrated in a small randomized controlled trial that high-dose fenoldopam decreased urine neutrophil-gelatinase-associated lipocalin (NGAL) and cystatin C, suggesting a renoprotective effect. At this time, there is no strong evidence for the routine use of fenoldopam to prevent AKI after pediatric cardiac surgery.
Broad treatment strategies
Supportive care for ADHF begins with strategies aimed at improving the specific components of D o 2 and V o 2 listed previously ( Fig. 28.1 ). This includes standard critical care therapies primarily focused on noncardiac support, such as intubation and mechanical support for respiratory insufficiency, temperature control, red cell transfusion for significant anemia, and control of pain and agitation with multimodal analgesia and anxiolysis. Mechanical respiratory support decreases metabolic demands (Vo 2 ) by decreasing the work of breathing, reduces LV afterload (wall stress), and can improve Sa o 2 by the administration of O 2 and positive end-expiratory pressure (PEEP), thus increasing Ca o 2 . However, endotracheal intubation can be risky in decompensated patients with HF, as the induction agents, laryngoscopy, and conversion to positive-pressure ventilation can induce cardiac arrest. The benefits of noninvasive positive-pressure ventilation (i.e., bilevel positive airway pressure) have not been proven, and the risks remain significant, particularly with regard to delay in timing of intubation and the need for additional sedation in children.
Ca o 2 can also be increased by transfusing packed red blood cells to increase hemoglobin concentrations in significant anemia. Research in the PICU has not demonstrated a negative impact on patient outcomes when a restrictive transfusion practice was embraced, though the study did not examine complex or high-risk congenital cardiac patients or those with unrepaired or palliated cyanotic cardiac defects where a relative polycythemia is normally maintained to compensate for the decrease in Sa o 2 . Clearly, numerous significant deleterious effects on the recipient occur after allogenic blood transfusion, and the risk of donor-directed blood can be greater. In patients who are euvolemic, hypervolemia with worsening pulmonary edema must be avoided when the decision to transfuse red cells is made. ,
CPB results in numerous significant neurohormonal perturbations in neonates, infants, and children after cardiac surgery. A pattern of sick euthyroid syndrome has been identified in pediatric patients after CPB. Several studies have demonstrated acute salutary effects of preoperative thyroid prophylaxis or postoperative therapy with triiodothyronine resulting in increases in specific hemodynamic parameters, improved diuresis, and a decreased need for additional cardiopulmonary support without significant adverse effects , although better outcomes have yet to be shown.
Critical illness–related corticosteroid insufficiency (CIRCI) has been demonstrated imprecisely in clinical trials in neonates, infants, and children after cardiac surgery. Stress-dose hydrocortisone replacement similarly has resulted in positive acute hemodynamic improvements and, yet again, this has not translated into improved survival or decrease in morbidities in these patients. All of these studies are confounded by the use of differing adrenal corticosteroids in varying doses used in the preoperative/intraoperative anesthetic/perfusion protocols. The STRESS trial is an ongoing randomized double-blind placebo-controlled trial evaluating the safety and efficacy of single-dose intraoperative methylprednisolone 30 mg/kg versus placebo in infants (<1 year of age) undergoing cardiac surgery with CPB.
A multicenter randomized controlled study evaluating two levels of glycemic control was stopped prematurely by the data safety monitoring committee owing to futility in reaching significance and apparent harm (hypoglycemia and hospital-acquired infections) in the low-glucose (80–110 mg/dL) compared with the high-glucose target (150–180 mg/dL) groups. Interestingly, pediatric cardiac surgery patients were excluded from enrollment. Despite the results of this large multicenter pediatric trial, the physiology and pathophysiology of hyperglycemia in the stress response to critical illness remains significant, and other well-controlled studies point to differing conclusions based on the details of the management strategy. , Finally, a Pediatric Cardiac Intensive Care Society’s consensus statement could not recommend any hormonal replacement/monitor strategies (thyroid, corticosteroid, or insulin) owing to the lack of robust double-blinded randomized clinical trials.
Terlipressin, a synthetic long-acting analog of vasopressin with a higher affinity for vascular V 1 -receptors than vasopressin, has been used in adult patients to treat extremely low CO, but its application in children is limited. In postoperative pediatric cardiac patients, terlipressin demonstrated an improvement in respiratory, hemodynamic, and renal indices in refractory LCOS. Although these results are encouraging, further investigation with prospective randomized trials is needed to provide data regarding the efficacy and safety of terlipressin in infants and children.
Inhaled nitric oxide (iNO) is a potent short-acting vasodilator resulting in numerous beneficial cardiopulmonary effects on the pulmonary, systemic, and coronary circulations. Its use is standard for severe, reversible pulmonary arterial hypertension (PAH) in the neonate and in children with congenital heart disease. It is also useful in the treatment of RV failure. A variety of other drugs (e.g., nitroprusside) and amino acids (arginine, citrulline) are precursors to NO and thus increase its circulating levels. Pediatric patients with severe PAH can also be treated with intravenous prostacyclin analogues, such as epoprostenol or treprostinil, , or continuously inhaled epoprostenol. In studies in neonates with persistent pulmonary hypertension , and critically ill noncardiac pediatric patients, , this therapy has proven to be equally efficacious as iNO and, in extreme cases, they can be used synergistically in a child.
Mechanical circulatory support in pediatric patients
Medical therapy for HF has improved survival and quality of life, although a large number of patients still require advance treatment with intravenous inotropic support, mechanical ventilation, and/or heart transplantation. MCS should be considered for children with decompensated HF who cannot be stabilized with medical therapy alone. Rapid advances in the field of MCS have dramatically changed the management of children with end-stage HF with emphasis on timely evaluation of ventricular assist devices (VADs) to preserve or recover end-organ function. When medical treatment is maximized or ineffective, patients should be considered for MCS for temporary support until heart function recovers, as a bridge to heart transplantation, or as a destination therapy (life with permanent VAD).
MCS through a variety of devices can be used to treat right, left, or biventricular failure. This is accomplished with devices that are either extracorporeal (outside the human body), paracorporeal (partial within and outside the human body), or intracorporeal or implantable (residing completely within the human body). MCS devices can provide either pulsatile flow or continuous flow depending on the specific design. MCS can also be provided to the left ventricle alone with an intraaortic balloon pump or it can replace the entire function of the heart with a total artificial heart. Finally, cardiac and pulmonary support can be provided with the addition of a membrane oxygenator in specific types of ECLS devices. Venoarterial (VA) ECLS provides cardiopulmonary support, whereas venovenous (VV) ECLS provides only pulmonary support for severe respiratory failure.
Despite improvements in operative techniques, management of CPB, and myocardial protection, myocardial dysfunction and failure can occur after surgery for CHD with involvement of the left and/or right ventricles. Approximately 5% of children undergoing cardiac surgery require MCS. The patient’s underlying pathophysiology, size, and expected length of support dictate which technique is most suitable. In children, most MCS continues to be achieved with ECLS. Since ECLS and VADs have advantages and disadvantages, determining the most appropriate modality for the individual patient requires significant experience and expertise.
Extracorporeal life support
ECLS is the most utilized form of short-term MCS for pediatric patients with decompensated HF unresponsive to medical therapy. Adaptations and simplification of the traditional CPB circuit have resulted in the standard VA ECLS circuit ( Fig. 28.2 ) through either extrathoracic (i.e., carotid artery/jugular vein or femoral vessels) or transthoracic (right atrium/ascending aorta) placement after median sternotomy. The venous cannula allows for drainage of blood from the patient by a negative pressure created by the servo-regulated centrifugal pump or, previously, a roller-head that propels blood through the remainder of the system at high pressure. Gas exchange (O 2 addition and CO 2 removal) occurs next through an artificial lung (oxygenator comprised of hollow fibers) where countercurrent sweep gas composed of O 2 /air mixture titrated to a specific fraction of inspired oxygen (F io 2 ) passes external to the blood phase. Blood temperature is controlled by a heat exchanger before blood is returned to the body through a major artery. Additional components include ports for infusion of medications, arterial and venous pressure monitors, Sv o 2 and/or blood gas analyzers, and flow and bubble detectors. There can be placement of a hemofilter for fluid control or a dialysis circuit (venovenous, arteriovenous, arterial-arterial) and, frequently, a bridge connecting the venous and arterial sides of the circuit is used during trials off ECLS or in case of a circuit emergency. Individual centers customize their ECMO circuits to serve their patient population by creating a less complicated circuit that is easier to manage with fewer connectors to reduce the number of potential sites for blood stasis. ,
VA ECLS provides biventricular and pulmonary support. It is important to note that complete cardiac bypass cannot be provided by VA ECLS owing to incomplete capture and drainage of the cardiac systemic venous return by the venous cannula. Routine cannulation for VA ECLS occurs via one of three vascular access points: (1) The transcervical approach places the venous cannula in the right atrium (RA) via the right internal jugular vein and the arterial cannula in the transverse aortic arch via the right carotid artery. (2) The transthoracic approach results in direct cannulation of the RA and ascending aorta through a median sternotomy for patients who either cannot be weaned from cardiopulmonary bypass or require MCS in the immediate postoperative period. (3) Femoral artery and vein cannulation for adolescents and adults uses longer cannulae to reach the RA and descending aorta. A combination of these approaches with more than one venous cannulation site can be employed when very high flows are required. The transcervical and transthoracic approaches are the preferred methods for small children. Fig. 28.3 is a chest radiograph obtained for evaluation of cannula placement in a patient supported by VA ECLS.
VV ECLS provides pulmonary support without cardiac support. In VV ECLS, a single double-lumen cannula ( Fig. 28.4 ) is placed in the RA through a transcervical approach. This cannula provides both inflow and outflow for the circuit. VV ECLS can improve RV function as a result of oxygenated, pH-balanced blood flowing to the lungs, thus decreasing PVR and right heart afterload. Although some patients experience improvement in overall cardiac function with VV ECLS, it might not be sufficient or sustained. Therefore, when significant myocardial dysfunction exists, VA ECLS is the preferred modality.
Extracorporeal membrane oxygenation indications and contraindications
It is widely accepted that all patients considered for MCS should have either a reversible physiologic process or should be a candidate for bridge to transplant or destination therapy. The severity of organ dysfunction and estimated time frame for recovery of cardiac and other organ failure both are used in determining optimal device selection. Several studies emphasize the importance of early institution of ECLS before a prolonged period of low CO results in multiorgan dysfunction. Appropriate patient selection is vital to maximize survival.
Myocarditis and extracorporeal life support
The clinical course of myocarditis is variable, with some patients presenting with subclinical disease, others with an indolent course progressing to a dilated cardiomyopathy, and a distinct subset of patients with fulminant disease. Without MCS, patients with rapidly progressive disease had expected survival rates of only 25% to 50%. Patients with acute fulminant myocarditis have worse short- and long-term outcomes when there are decreased LV ejection, ventricular arrhythmias, and/or LCOS. With aggressive utilization of ECLS as a bridge to transplantation or recovery, survival rates for patients are now reported to be as high as 90%. It has been shown that the institution of MCS can normalize ventricular geometry, cellular composition, metabolism, and, ultimately, function—a phenomenon referred to as reverse remodeling . This process is thought to improve ventricular dysfunction because of favorable influences on the neurohormonal cardiovascular milieu and ventricular unloading. For patients with end-stage dilated cardiomyopathy secondary to myocarditis, the use of MCS without transplantation has resulted in survival rates as high as 67% to 80%.
Failure to wean from CPB occurs in approximately 1% to 3.2% of pediatric congenital cardiac surgery cases. Individual institutions have reported survival rates to hospital discharge between 32% and 54% for pediatric patients who require MCS after cardiac surgery. The use of MCS is currently widely accepted to support vital organs while allowing for myocardial recovery. It is imperative that these patients be evaluated for the presence of residual cardiac defects that may be causing or worsening cardiovascular collapse. Intraoperative transesophageal echocardiography (TEE), transthoracic echocardiography (TTE), and diagnostic cardiac catheterization may identify the need for reoperation or interventional cardiac catheterization to improve the patient’s hemodynamic status. In one large center, ∼2% of cardiac surgeries resulted in an early postoperative catheterization, ∼50% diagnostic and ∼50% interventional, and there was no procedural mortality. Importantly, 30% of those undergoing catheterization required reoperation. Balloon valvuloplasty, angioplasty of aortic arch obstructions, device closure of residual septal defects, coil occlusion of aortopulmonary collateral vessels, or atrial septostomy all may be crucial interventions to improve the patient’s hemodynamic state and allow for separation from MCS. Untreated and significant residual cardiac defects have been shown to be almost universally fatal for patients requiring MCS. The use of MCS for postcardiotomy support in neonates and infants with single-ventricle physiology is technically more complex with worse outcomes. , Thus, previously this could have been considered a relative contraindication to MCS. , The 2020 Extracorporeal Life Support Organization (ELSO) Registry (2015–20) reports a 44% survival rate for ECLS use in neonates with hypoplastic left heart ( eTables 28.1 and 28.2 ).
|Total Runs||Average Run Time (h)||Longest Run Time (h)||Survived||% Survived|
|Congenital heart defect||963||144||1463||464||48|
|Total Runs||Average Run Time (h)||Survived (n)||Survived (%)|
|Hypoplastic left heart||427||136||192||44|
|Cyanotic increased pulmonary blood flow||64||177||25||39|
|Cyanotic decreased pulmonary blood flow||299||1271||39||58|
Extracorporeal cardiopulmonary resuscitation
Survival from pediatric cardiac arrest has improved tremendously over the last 20 years. However, half of the children who suffered in-hospital cardiac arrest do not survive to hospital discharge and those who do survive have significant morbidity. Although neurocognitive outcome has improved for survivors of cardiac arrest, the duration of cardiopulmonary resuscitation (CPR) remains inversely proportional to survival. The American Heart Association guidelines for in-hospital pediatric cardiac arrest now recommend consideration of extracorporeal CPR (ECPR) during CPR if the conditions leading to the arrest are likely to be reversible or amenable to heart transplantation. One of the most important aspects of the decision to institute ECPR is adequate patient selection. There is sufficient data to support the use of ECPR in children with cardiac disease who suffer an in-hospital cardiac arrest. However, wide variability in patient selection and the ability to institute ECLS in a timely fashion (ideally, <60 minutes) has led to variable success, with survival rates ranging from 0% to 100%. , Thus, some large centers have created systems for rapid-deployment ECLS using a team that is immediately available to cannulate with a preprimed circuit. , , Children with CHD have increased risk for an in-hospital cardiac arrest. Their survival to discharge after ECPR has been reported to be up to 50%, yet single-ventricle patients suffer more cardiac arrests and have the worst outcome. , , , Reduced survival of ECPR has been associated with lower postcannulation pH, higher lactic acid, and end-organ injury. Post–cardiac arrest management (temperature control, target flow rates, target perfusion pressure) is beyond the scope of this chapter but constitutes an extremely important factor that can impact outcome. The January 2020 ELSO Registry for neonatal and pediatric patients supported with ECLS following cardiac arrest (ECPR) reported a survival to hospital discharge of 46% for neonates and 37% for children.
Bridge to transplantation
Although heart transplantation is the treatment of choice for end-stage myocardial failure, many children die every year waiting for a suitable organ to become available. As a result, MCS is now increasingly used as a bridge to heart transplantation. , The most common indications for MCS as a bridge to transplant include cardiac failure due to CHD, cardiomyopathy, and graft rejection after heart transplantation. Children with myocardial failure secondary to fulminant myocarditis also demonstrate increased short-term survival when treated with MCS and transplantation. , Complications associated with ECLS generally limit the duration of support to approximately 2 weeks, but as long as complications precluding transplant have not developed, no arbitrary cutoff for duration of ECLS has been determined.
An analysis of the United States Scientific Registry of Transplant Recipients demonstrated that waitlist mortality varied as much as 10-fold based on recipient factors. Over the last decade, the proportion of pediatric heart transplant candidates with CHD increased from 48% in 2008 to 62.2% in 2018. From the pediatric heart transplantations performed in 2018, 7.8% were in children younger than 10 years. The overall percentage of candidates with VADs at the time of listing increased from 11.8% in 2008 to 32.6% in 2018. Recipient characteristics associated with waitlist mortality included ECMO, mechanical ventilation, listing status 1A, CHD, dialysis support, and non-white race. Waitlist mortality for infants was significantly higher than for older patients ( Fig. 28.5 ). Although waitlist mortality remains high, VAD support has been an important factor in reducing waitlist mortality in small children. An analysis of the United Network of Organ Sharing (UNOS) database comparing the pre-VAD (1999–2004) and post-VAD (2005–15) eras reported more than 50% reduction in waitlist mortality in the post-VAD era, supporting the importance of centers of excellence that can provide these state-of-the-art therapies.
A study merging UNOS and Pediatric Health Information Systems databases reported that children with CHD, particularly single-ventricle conditions, require substantially greater hospital resource utilization and have significantly worse outcomes during the first year after heart transplantation compared with other indications, emphasizing the importance of modifying patient risk factors with interventions such as pretransplantation conditioning with VAD support and cardiac rehabilitation.
MCS can be lifesaving for pediatric patients with malignant dysrhythmias unresponsive to pharmacotherapy. A subset of patients with acute fulminant myocarditis will present with refractory LCOS and ventricular tachycardia or high-degree heart block that further compromises end-organ perfusion. In this scenario, MCS may be used to bridge patients to recovery following initiation of antiarrhythmic agents and/or electrophysiologic mapping with ablation therapy or transplantation. Several potentially lethal arrhythmias—such as supraventricular tachycardia, junctional ectopic tachycardia, ventricular tachycardia, or torsades de pointes—can occur congenitally, can be acquired in the perioperative period, or manifest as a result of the ingestion of toxic substances or medications. Again, ECLS can allow time for resolution of the dysrhythmia with aggressive medical treatment and the subsequent recovery of cardiac function.
Although individual institutions may have variations to this list, the number of absolute contraindications to ECLS continues to decrease. Irreversible cardiac failure without the option for transplantation, irreversible lung failure, severe neurologic dysfunction, grade III/IV intracranial hemorrhage, uncontrolled bleeding, and lethal congenital anomalies/genetic syndromes are considered contraindications. Each center should evaluate carefully the decision to provide ECLS to these patients. Relative contraindications include prematurity (<34 weeks), weight less than 2 kg, and grade II intraventricular hemorrhage. Cardiac arrest is not a contraindication if rapid effective CPR was initiated, the underlying etiology is deemed reversible, and time to cannulation is less than 60 minutes. The challenge remains determining what is irreversible , shortened life span , poor prognosis , or severe , as these definitions have changed over time and with differing expectations for quality of life if survival is achieved. Technical limitations are an inability to obtain vascular access secondary to thrombosis, abnormal anatomy, or prior surgery.
Critical care management during extracorporeal life support
Patients who require MCS are critically ill by definition, often with multiple end-organ dysfunction. Thus it is not surprising that complications occur with a greater frequency for this group than for other critically ill patients. Basic management principles are discussed in this section.
When assessing the hemodynamic state of a patient, one must consider to what degree conventional support (mechanical ventilation and vasoactive agents) is contributing to cardiac and pulmonary function versus the ECLS system. The amount of flow provided by the ECLS system is measured through ultrasonic flow probes on the high-pressure side of the circuit beyond where any shunts (i.e., high-pressure to low-pressure connection through a hemofilter) may occur. An approximation of flow generated by the pump can be calculated from the product of the revolutions per minute (RPMs) and circuit tubing diameter. For most cardiac ECLS patients, flow is initiated at 120 to 150 mL/kg per minute. However, it is important to note that flow should be adjusted to fit the physiologic needs of each patient. Patients with septic shock may require higher flows, approaching 200 mL/kg per minute, to support their metabolic needs assuming that venous drainage is sufficient to yield such flows. A precise measurement of the patient’s intrinsic CO (i.e., the amount of blood flow that is not passed through the ECLS circuit) is unobtainable. However, indirect evaluation of the patient’s CO is possible through an assessment of arterial systolic blood pressure, pulse pressure, HR, organ perfusion, SvO 2 , and lactic acid levels on a given flow rate. Comparisons of these variables over time allow the clinician to make decisions regarding the adequacy of circulatory support or the readiness to wean from support. An echocardiogram can also provide valuable information about cardiac filling and function and guide therapy, particularly when weaning from ECLS or during a trial off ECLS. Sv o 2 is measured in the venous return portion of the circuit, with a goal of 65% to 80%. However, if a left-to-right shunt exists, such as a left atrial vent, the Sv o 2 will be falsely elevated, particularly if the patient is receiving high F io 2 and PEEP. Serial lactate measurements are often helpful to aid in the assessment of global end-organ perfusion. Elevated lactate levels may occur in patients with ongoing hepatic dysfunction, sepsis, low CO, and end-organ hypoperfusion. A rough approximation of the relative contributions of the patient and circuit pulmonary parameters is possible by analyzing serial patient and circuit blood gas assays (pH, partial pressure of arterial carbon dioxide, Pa o 2 , Sa o 2 ) while taking into consideration the mechanical ventilation settings (including F io 2 and PEEP), circuit flow rate, gas sweep rate, and circuit O 2 concentration. Finally, shunts within the patient or the circuit must be taken into consideration. Examples of right-to-left patient shunts include those with a patent ductus arteriosus, atrial septal defect, or ventricular septal defect in the setting of severe PAH. Circuit left-to-right shunts are generally limited to a left atrial vent, open bridge, the arteriovenous hemodialysis filter, and in vivo continuous arterial blood gas devices. Total blood flow for the patient on MCS requires the addition of ECLS circuit flow with the patient’s native CO minus any shunt within the system as a whole.
Hemodynamic compromise often continues despite MCS. Low-dose inotropes or inodilators can aid cardiac contractility to augment native CO and reduce afterload to both the right and left ventricles. The most commonly used agents are dopamine (3–5 μg/kg per minute), epinephrine (0.03–0.05 µg/kg/min), dobutamine (5 µg/kg per minute), or milrinone (0.25–0.75 μg/kg per minute). Recall that while these are the typical modest range of doses in children, the presence of ECLS reduces the plasma level of these agents owing to the marked increase in the volume of distribution and circulating blood volume while on ECLS. The use of catecholamines in high doses is detrimental to cardiac recovery and should be avoided by increasing circuit flow to provide adequate CO. Likewise, the complete removal of any inotropic support is typically not suggested while on ECLS either, particularly when cardiac stun (lack of opening of the aortic valve during systole, which leaves the arterial line flat from only ECLS flow) is present. Tachydysrhythmias and pulseless electrical activity requiring cardiopulmonary resuscitation occurs in 3% of neonatal and pediatric ECLS runs. Pharmacologic/electrical cardioversion of any dysrhythmia should be attempted emergently even while on ECLS to prevent further deterioration in cardiac perfusion and function. Cardiac pacing can also be used to optimize CO.
Hypovolemia is a common occurrence during MCS for a variety of reasons. Inadequate venous drainage secondary to cannula malposition, cardiac tamponade, tension pneumothorax, or hemothorax may occur and generally results in hypotension requiring immediate correction. Ongoing evaporative losses from the oxygenator and bleeding secondary to coagulopathy can contribute to hypovolemia. Initiation of ECLS activates a host of inflammatory mediators, resulting in capillary leak and hypovolemia. Finally, attempts at mobilizing the large amount of “third-spaced” fluid with either diuretics, hemofiltration, or through drains in pleural/peritoneal cavities can quickly cause either inadequate CO from the patient or inadequate venous drainage to the ECLS circuit.
Hypertension secondary to neurohormonal dysregulation or pain/agitation is one of the most common and unavoidable cardiovascular complications of ECLS. Although hypertension has not been demonstrated to negatively impact patient survival, its presence can worsen bleeding or further impair cardiac function and thus should be promptly addressed.
Cardiac stun , a term that describes reversible global dyskinesia of the ventricle, was coined by Braunwauld and Kloner in 1982. Reversible cardiac dysfunction that results in the lack of antegrade LV ejection during systole resembles electromechanical dissociation, which has been observed frequently in patients following the initiation of VA ECLS. Excluding conditions of physiologic tamponade from thoracic issues, blood, or air, an infant on ECLS should have sufficient cardiac function to generate a minimal pulse pressure of 10 mm Hg. Evaluations of patients who experience cardiac stun upon initiation of ECLS defined this condition as the absence of aortic valve opening during systole, equalizing of the patient preductal and postmembrane circuit Pa o 2 , and absence of pulse pressure in the aorta. The etiology of cardiac stun is multifactorial and hypothesized to be the sequelae of acute ischemia followed by reperfusion or severe electrolyte disturbances. The incidence of stun on ECLS is 5% to 12% in neonates and, when present and prolonged (>24 hours), results in a significant increase in mortality for these patients. Stun typically occurs during the initiation of bypass in patients who were hypoxic, hypercarbic, acidotic, and suffered a cardiac arrest prior to ECLS. Important factors in trying to minimize cardiac stun upon the initiation of ECLS include correcting the pH and ionized calcium levels in the circuit and infusing calcium chloride to the patient upon commencing ECLS, followed by close monitoring of arterial blood gases and electrolytes with rapid correction of abnormalities.
Echocardiography and cardiac catheterization
Initial diagnostic and hemodynamic assessment should be attempted by transthoracic echocardiography, but imaging windows may be severely limited resulting in insufficient information. TEE may improve diagnostic accuracy but may not be possible secondary to bleeding risks or patient size. Even if adequate imaging is obtained, the specific hemodynamic state of the patient on ECLS must be taken into account when interpreting these studies. Invariably, the loading conditions of the heart are markedly altered during ECLS. The echocardiographer should actively interface with the ECLS team (to potentially modify flows, add volume, modify mechanical ventilator support, etc.) in order to obtain the most comprehensive assessment of cardiac function to inform clinical decisions regarding the continuation or removal of further ECLS. Cardiac catheterization can be a useful tool for select patients who fail to wean from ECLS. The specific loading conditions of the heart must be considered in the context of the acquired data in order to make sound clinical decisions. Therapeutic interventions that might be performed in the catheterization lab include balloon or blade atrial septostomy to alleviate left atrial hypertension, balloon valvuloplasty or angioplasty of vascular obstructions, device closure of residual atrial or ventricular shunts, and/or coil embolization of aortopulmonary shunts. Correction of these types of residual defects in the catheterization laboratory or surgical correction in the operating room may be required to allow for separation from MCS. For patients with significant LV failure, it may be necessary to decompress the left heart to prevent or reverse pulmonary edema or hemorrhage, decrease mitral regurgitation and, importantly, improve coronary perfusion to increase the chances of myocardial recovery. In this scenario, venting of the LV occurs through surgical placement of a cannula in the left atrial appendage connected to the venous drainage to the ECLS circuit. This is accomplished in the operating room/ICU during or after transthoracic ECLS cannulation or in the catheterization laboratory through creation of an interatrial connection via balloon or blade septostomy. This procedure can markedly improve LV function and increase the chances of survival. ,
Lower survival rates are universally found in this subset of patients, which may be attributed to an imbalance of the systemic and pulmonary circulations, volume burden to the single ventricle after complex palliative surgery, compromised single-ventricle function (particularly with RV morphology), and/or impaired coronary perfusion to the systemic ventricle when a systemic-to-pulmonary artery (PA) connection (i.e., modified Blalock-Taussig [BT] shunt) results in diastolic runoff from the aorta. Despite these challenges, larger centers continue to report improved outcomes with the accumulation of experience and application of innovative strategies. For example, initial efforts to balance the systemic and pulmonary circulations on ECLS included either completely or partially occluding the aortopulmonary shunt, which has now been demonstrated to increase mortality. The use of smaller-size BT shunts or the use of the Sano modification (RV to PA nonvalved conduit for stage one hypoplastic left heart palliation) has contributed to decreasing the recirculation that would otherwise occur. Of note, higher ECLS with flows approaching 200 mL/kg per minute may be required to provide adequate systemic and pulmonary support when the shunt is left open. ,
ECLS requires meticulous management of hemostasis to limit patient morbidities. Hemorrhagic and thrombotic complications are a major concern for patients during ECLS, particularly after cardiac surgery. Bleeding can manifest at surgical sites (arterial/venous cannulation sites, surgical repair sites [atriotomy, ventriculotomy, aortotomy sites], sternal incision, indwelling catheter sites, etc.) or may be masked in areas such as the thorax, intracranial vault, or the gastrointestinal tract. A meta-analysis of observational studies, including 1763 patients on VA ECMO, reported a 33% incidence of bleeding that was mostly correlated to the heparin monitoring strategy. Prevention strategies that target reduction of hematologic complications focus on maintenance of the hemostatic regulatory mechanisms as close to normal as possible. Patients placed on ECLS following cardiac surgery represent a unique population at risk for hemorrhagic management since they often have multiple surgical sites, dilutional coagulopathy, and abnormal coagulation patterns. Apart from single-center experience, no well-defined consensus or protocol is available for pediatric and neonatal ECLS. The patient’s age, diagnosis, clinical status in conjunction with the specific details of the ECLS device, and, finally, the flow through the circuit will dictate which anticoagulation strategy should be employed. The most commonly used agent for anticoagulation on ECLS is continuously infused unfractionated heparin (UH). Most centers measure platelet counts, hematocrit, prothrombin time (PT), partial thromboplastin time (PTT), fibrinogen, specific factor levels (i.e., anti-factor Xa, antithrombin III [ATIII]), quantitative heparin levels, and activated clotting time (ACT). The ACT is the most commonly measured test for coagulation, as it can be performed quickly at the bedside, though other coagulation tests are available at the point of care (i.e., PTT). , Despite this advantage, ACT results vary markedly based on the technique used and are a nonspecific measure of coagulation because they measure the total time for a clot to form after ex vivo activation. The ACT is a global composite of the different individual components of coagulation; thus, more specific tests (listed earlier) must be used in concert to ascertain which specific component in the coagulation cascade is affected. No single test is currently used to guide the anticoagulation regimen; rather, a panel of tests must be obtained and analyzed in concert. Increasingly, many centers have reported on the effective use of thromboelastography (TEG) in determining patient coagulation status. When “adequately anticoagulated on a continuous unfractionated heparin infusion,” the target ACT should be between 180 and 220 seconds when the ECLS pump is flowing at full flow (i.e., >150 mL/kg per minute). This goal can be decreased to 160 seconds on full flow if significant bleeding is present, particularly in the immediate postoperative period. UH is metabolized via two mechanisms: at low doses, via a saturable mechanism representing clearance by the reticuloendothelial system and endothelial cells to which heparin binds with high affinity, and at high doses, via a nonsaturable mechanism represented by renal excretion. Thus, close monitoring during ECLS must be followed when a patient’s urinary output is oscillating between oliguria and polyuria.
Newer treatment strategies include the use of additional agents such as ATIII in either bolus or infusion form to correct abnormalities in the clotting cascade. ATIII is an α 2 -glycoprotein serine protease inhibitor that inactivates a number of enzymes from the coagulation system, including the activated forms of factors II, VII, IX, X, XI, and XII. Replacement of ATIII is controversial, with some centers replacing it only when the level is low (<30%) and heparin infusion rates are increasing without a concomitant increase in ACT or PTT, , while other centers maintain levels near 100%.
Bivalirudin is a direct thrombin inhibitor that works independent of antithrombin on both circulating and clot-bound thrombin. It is currently approved for use during percutaneous coronary intervention and heparin-induced thrombocytopenia. It has a quick onset of action and a short half-life. Bivalirudin infusions have been reported in a small number of pediatric patients on ECLS. At our institution, we have successfully used bivalirudin in patients with suspected heparin-induced thrombocytopenia, heparin resistance, risk of major bleeding, and/or evidence of thrombosis while on heparin. Ranucci et al. demonstrated that the use of bivalirudin in ECMO was associated with less total blood loss and fewer transfusion needs in postcardiotomy patients requiring ECMO. There is wide variability on the dosing strategies, but at our institution we start the infusion at 0.15 to 0.20 mg/kg per hour without the administration of a bolus and titrate to goal PTT 60 to 80 seconds. Anticoagulation with Coumadin, clopidogrel, low-molecular-weight heparin, and aspirin individually or in combination can be considered with certain MCS devices other than ECLS.
Heparin-induced thrombocytopenia (HIT) is a relatively rare but serious complication of heparin administration caused by antibodies binding to a complex of heparin and platelet factor 4 that leads to large-vessel thrombosis and increased mortality. A drop in platelet count by more than 50% of the highest previous value should raise suspicion of HIT and trigger investigation. Treatment includes discontinuing heparin administration and initiating direct thrombin inhibitor therapy (i.e., argatroban) if continued anticoagulation is needed. , The lysine analogs tranexamic acid and ε-aminocaproic acid are antifibrinolytic agents that have been shown to reduce bleeding in ECLS patients undergoing surgical procedures. However, prophylactic administration failed to reduce the incidence of intracranial hemorrhage in neonates. ,
Profound abnormalities in many components of the coagulation cascade commonly occur in postoperative cardiac patients on ECLS. With this in mind, an attempt to normalize components of coagulation not impacted by heparin is important. Platelets are consumed at surgical bleeding sites, sequestered by the membrane oxygenator; thus, transfusions are required to maintain counts greater than 100,000/mm 3 in bleeding patients or patients at high risk. However, lower transfusion thresholds can be set to avoid excessive exposure to blood products in the nonbleeding patient. Administration of fresh-frozen plasma to broadly increase multiple coagulation factors activity is also common. When hypofibrinogenemia occurs, cryoprecipitate infusion can be used owing to the high concentrations of fibrinogen in the low volume of the cryoprecipitate unit. The availability of plasma protein concentrate allows administration of highly concentrated factors in a substantially reduced volume, decreasing the volume burden associated with conventional therapy. Although rare, the application of heparin-free ECLS has been reported. ,
Ventilator management during ECLS remains controversial. Although data exist to guide clinicians regarding prevention of barotrauma, volutrauma, and O 2 toxicity for mechanically ventilated patients with acute respiratory distress syndrome in general, there is a lack of controlled clinical trials and consensus for pediatric patients on ECLS. Lung collapse strategies used for respiratory support on ECLS are not used by most cardiac centers. Goals continue to target the prevention of atelectasis with utilization of appropriate PEEP in order to maximize oxygenation of nonbypassed blood returning to the left atrium, which then is ejected by the LV to perfuse the coronary arteries. In addition, providing modest levels of ventilator support can be achieved with either a pressure- or volume-limited mode of ventilation targeting a delivered tidal volume of ∼6 mL/kg. Respiratory rates between 10 and 25 are set depending on the age and rest strategy being employed and the degree that the patient’s lungs are required for gas exchange. Optimizing Pa o 2 to the patient and circuit by blending F io 2 to keep the F io 2 below 0.6 reduces free radical formation and O 2 toxicity, providing adequate oxygenation to decrease pulmonary hypertension and optimize myocardial D o 2 . Chest radiographs are routinely performed to assess and guide strategies to optimize lung volume so that volutrauma and barotrauma are avoided.
Fluid, nutrition, and renal
Fluid overload and electrolyte disturbances such as high or low serum levels of potassium, calcium, magnesium, and phosphorus are common and need correction. Most cardiac patients on ECLS receive total parenteral nutrition owing to the increased risk of gastrointestinal complications in patients with cardiac defects, umbilical artery catheters, poor perfusion, or other bowel abnormalities. A select subset may tolerate trophic enteral feedings; whenever possible, even a small amount of enteral feeds should be provided. Diuretics—commonly, furosemide—are often employed to provide optimal fluid balance in patients with significant capillary leak and fluid retention. Optimization of fluid status is essential to weaning and eventual separation from ECLS. For anuric or oliguric patients, early placement of an in-line hemofilter into the ECLS circuit with or without countercurrent dialysate or a complete continuous renal replacement therapy device is recommended. Still, the management decisions that balance the use and timing of diuretics versus hemofiltration are center specific. The indications for initiation of dialysis are the same as for other critically ill patients with renal failure. Multiple retrospective reviews of patients supported by ECLS have found that fluid overload and AKI are risk factors for increased mortality.
Analgesia and sedation
Adequate analgesia and sedation are essential for both safety and comfort and to decrease metabolic demands in patients with circulatory compromise in the early postoperative recovery period. While an opiate and a benzodiazepine class drug have been used historically, a multimodal approach is increasingly preferred, which could include other agents such as dexmedetomidine, ketamine, or propofol. Current oxygenators can bind many drugs, including opioids, depending on the lipophilic and protein-binding qualities of the specific drug. Thus dosing amounts are usually significantly higher while on ECLS. , In this study, ECMO circuits were set up using Quadrox-iD pediatric oxygenators primed with whole blood to represent a 5-kg patient. Hydromorphone and fentanyl were injected (also, mycophenolate and tacrolimus) and serial blood samples were taken over 12 hours. In this ex vivo model, hydromorphone hydrochloride was not as significantly sequestered compared with fentanyl (mycophenolic acid and tacrolimus serum concentrations were stable). Benzodiazepine infusions may be used with cautious monitoring, as toxicity from propylene glycol (lorazepam), other solvents (midazolam), or long-acting toxic metabolites (diazepam) have been reported in critically ill neonates, infants, and children. Recent studies are finding an increase in delirium when benzodiazepines are used in critically ill pediatric patients; thus, some centers are beginning to limit their administration. Neuromuscular blockade should largely be avoided to limit the development of critical care neuromyopathy, allow for regular evaluation of the central nervous system, and limit soft-tissue fluid accumulation. Central nervous system infarcts, hemorrhage, or seizures are all known complications of ECLS. For infants with an open fontanel, a daily head ultrasound should be performed early in the course of treatment and with any change in clinical neurologic status. For older patients, a significant change in their neurologic status has a high likelihood of heralding major intracranial pathology, which needs to be promptly diagnosed by computed tomography in order to guide treatment and determine patient viability.
Several centers have reported a new approach in which older patients are supported without the need of continuous analgesia and sedation and without the need of mechanical ventilation. This awake ECLS modality has been used as bridge to recovery, bridge to VAD, or bridge to transplantation. ,
It is not surprising that patients on ECLS are at a high risk of developing nosocomial bloodstream infections (BSIs). Identified risk factors include the duration of ECLS, open versus closed chest cannulation, the presence of central venous lines, and undergoing a major procedure prior to or while on ECLS. It appears that older patients on ECLS for respiratory failure may be at higher risk of healthcare-associated infection than neonates or cardiac patients. BSIs during ECLS for both pediatric cardiac patients post-CPB and neonates with cardiac or respiratory failure have been associated with a poor outcome. The diagnosis of sepsis is difficult in patients supported with ECLS. Although variable degrees of leukopenia have been documented for neonates supported with ECLS, an increase in phagocytosis and intracellular killing by neutrophils also occurs. Temperature is controlled by the circuit’s heat exchanger; thus, infection is generally not manifested by fever in these patients. Hypotension or thrombocytopenia may occur for a variety of reasons. In view of this observation, the standard of care in many ECLS centers has been to perform routine surveillance cultures and provide prophylactic antibiotics. However, management strategies to limit infectious risks continue to evolve. Owing to lack of a proven benefit and concerns regarding the long-term impact of broad-spectrum antibiotics on local bacterial resistance profiles, an increasing number of centers now perform daily blood cultures without the routine use of prophylactic antibiotics. Additional retrospective data may suggest that routine surveillance cultures may not be warranted. Further investigation is needed to determine the impact of prophylactic antibiotic use on the incidence of BSI, local antimicrobial flora, length of stay, survival, and cost.
Crucial situations exist for patients supported with ECLS that require intrahospital transport. This can include mobilization from the intensive care unit to a variety of locations, such as the catheterization laboratory, radiology suite, or the operating theater. Reluctance to perform diagnostic or therapeutic interventions is often driven by fear of potentially disastrous complications during transport, yet these fears are largely unfounded. Guidelines designed to promote the establishment of an organized, efficient transport process supported by appropriate equipment and personnel have been recognized and are increasingly used in hospitals. Intrahospital transport for patients on ECLS is a labor-intensive process that should be approached in a coordinated effort with specific focus on the preparatory phase, the transfer phase, and the posttransport phase. Simulation practice with the various teams is now standard practice, resulting in safe intrahospital transport for patients on ECLS.
Ventricular assist devices
The two main types of VADs are pulsatile pumps and continuous-flow pumps. Although the initial types of VADs were pulsatile, continuous-flow pumps have been gaining popularity worldwide both for adult and pediatric patients ( eTable 28.3 ) because of their decreased incidence of thromboembolic complications, smaller size, and ability to discharge patients home. Most VADs share similar basic principles, and based on the pump-patient interface, devices can be classified as intracorporeal or paracorporeal. Cannulation depends on the type of support required. The right atrium (systemic venous drainage) and PA (arterial return) are cannulated for a right ventricular assist device (RVAD), and the LV (pulmonary venous drainage) and aorta (arterial return) are cannulated for a left ventricular assist device (LVAD). A combination of both right and left ventricular assist devices is termed a biventricular assist device (BiVAD). For children with complex CHD, including single-ventricle physiology, placement of the inflow and outflow cannulas can be varied and complex. The pump is connected to a controller and power supply and other monitoring sensors. A comparison of ECLS and VADs is listed in eTable 28.4 .